Wood and Cellulosic Chemistry, Second Edition Revised and Expanded

WOOD AND CELLULOSIC CHEMISTRY second edition, revised and expanded

edited by

David N.-S. Hon Clemson University Clemson, South Carolina

Nobuo Shiraishi Kyoto University Kyoto, Japan

M A R C E L

MARCEL DEKKER, INC. U E K K E R

NEWYORK BASEL

Library of Congress Cataloging-in-Publication Data

Woodandcellulosicchemistry / editedbyDavidN.-S.Hon, Nobuo Shiraishi.-2nded.,rev.and expanded. p. cm. Includes index. ISBN 0-8247-0024-4 (alk. paper) 1. Cellulose. 2.Wood-Chemistry. I. Hon,David N.-S. 11. Shiraishi,Nobuo. QD323.W662000 572'.56682"dc21 00-060

This book is printed

o n acid-free paper.

Headquarters

Marcel Dekker, Inc. 270 Madison Avenue, New York, NY 1 0016 tel: 2 12-696-9000; fax: 2 12-685-4540 Eastern Hemisphere Distribution

Marcel Dekker AG Hutgasse 4, Postfach 812, CH-4001 Basel, Switzerland tel: 41-61-261-8482; fax: 41-61-261-8896 World Wide Web

http://www.dckker.com The publisher offers discounts on this book when ordered in bulk quantities. For n1ore infornmation. write to Special Sales/Professional Marketing at the headquarters address above. Copyright 0 2001 by MarcelDekker,Inc.

All RightsReserved.

Neithcr this hook nor any part nuy be reproduced o r transmitted in any form or by any means. electronic o r mechanical, including photocopying. microlilnling. and recording. o r by any information storage and retrieval systcm. without permission in writing from the publisher. Current printing (last digit): 1 0 9 8 7 6 5 4 3 2 1 PRINTED IN THE UNITED STATES OF AMERICA

Preface

Life and its surroundings are constantly changing within our dynamic world. As we stride into the new millennium, information technology and biotechnology continue to flourish. Rapid economic expansion, social development, and high demands for shelter, clothing, energy, and food for our overpopulated world have resulted in a desperate need for new and yet functional materials to support society’s infrastructure. Wood or lignocellulosic-based materials have made a significant contribution to the quality of living for human beings. With new developments in wood chemistry, scientists are confident that wood will continue to play an important role in fulfilling the needs of human beings. Over the past decade, the trend of emphasizingbio-basedtechnologieshasbeen observed worldwide. In February 1998, a long-term development project, PlanVCrop-based Renewable Resources 2020, was implemented among the U.S. Department of Agriculture, U.S. Department of Energy, and many U.S. companies, agricultural associations, and universities. The aim of the project was to obtain novel chemicals from plant- and crop-based renewable resources in order to widen the usage of crops, the yield of which has been significantly increased through bio-technological advancements. The recent movement of producing foods by means of genetically manipulated seeds should enhance the effectiveness of this project. Before the start of this project-which is considered the future of the petrochemicalindustry-majorchemicalcompanies in the UnitedStates,suchasDow Chemical,Dupont, and Monsanto,havebeenchanging their strategies in research and development.Theyhavestrengthened their bio-basedresearch field, trying to yield as many chemicals as possible from biomass. They are developing production technologies for ethanol, sorbitol, lysine, tryptophane, citric acid, lactic acid, poly(lactic acid), erythritol, 1,3-propanediol,etc.,frombiomass.Furthermore, in August of 1998 PresidentClinton issued an executive order, “Developing and Promoting Biobased Products and Bioenergy,” to further the development of a comprehensive national strategy that includes research. development, and private sector incentives to stimulate the creation and early adoption of technology needed to make bio-based products and bio-energy cost-competitive in national and international markets. Also. there has been research in so-called “green chemistry.” In this new methodology. biomass is the recommended MW material. Thc importance of wood and cellulose rescarch is thus rccognizcd. iii

iv

Preface

Since the publication of' the first edition of this book, considerablc advancement i n various fields ofwood chemistry has been made, as can be attested by many scientific publications in addition to well-attended international conferences. We contacted the contributors to the first edition, soliciting their opinions on revising and updating the book, and we received tremendous support from them as well as the publisher. Unfortunately, and inevitably, several authorswereunableto participate, buttheyrccomrnended their successors. Although most of the chapters in this new edition carry the same titles as those i n the previous edition, they have all been extensively revised and updated. In addition, this edition includes several new chapters representing important threads in the total fabric of wood chemistry. These new chapters cover the subjects of chemical synthesis of cellulose, preservation of wood, preservation of waterlogged wood, biodegradable polymers from lignocellulosics, recycling of wood and fiber products, and pulping chemistry. As editors, we feel fortunate to have been able to recruit some of the best talent in the field to this endeavor. We thank the contributors for their efforts. Any praise for the content should be addressed to them, and comments and criticisms to us will be welcome.

David N.-S. Hon NoDuo Shiruishi

Contents

1.

Ultrastructure and Formation of Wood Cell Wall

1

Minoru Fujittr trrlcl Hirnshi Hcrmdrr

2.

ChemicalComposition and Distribution

SI

Shirr) Strkrr

3.

Structure o f Cellulose: Recent Developments in Its Characterization

83

Frrrtlittrktr Hot-ii

4.

Chemistry of Lignin 109 Akirn Srrkrrkihrrr-rrcrrlrl Yoshillit-o Strrlo

S. Chemistry of Cell Wall Polysaccharides

175

Ttrtltrslli lsllii rrrltl Kuxrt1tr.w Shirr1i:rr

6.

Chemistry of Extractives Toshitrki U r ~ r t w ~ ~ r

7.

Chemistry of Bark

2 13

243

Kokki Sakoi

8. ChemicalCharacterization o f Wood and Its Components

275

Jrrirtw Htrexr cult1Jucrrlittr Frret-

9.

Color

IO.

3x5 NoDrryrr Mitlcwlrr.cr

~ t n dDiscoloration

D m i t l N . -S. Horl

trrld

Chemical Degradation

443

Krcrrl-ZotrgLrri

1 I. Weathering and Photochemistry o f Wood IltrlGcl N.-S. Hot1

S I3

V

vi

Contents

12. Microbial, Enzymatic, and Biomimetic Degradation of Lignin in Relation to Bioremediation 547 Rrkqfumi Huttnri und Mikio Shimadu 13. Chemical Modification of Wood

573

Misato Nothoto

14. Chemical Modification of Cellulose599 Akirn Isogai

15.

ChemicalSynthesis of Cellulose627 Furniaki Nukatsubo

16. Wood Plasticization

655

Nohuo Shiruishi

17. Wood-Polymer Composites

701

Hirnshi Mizunztrchi

18.

Adhesion and Adhesives

733

Hiroslli Mizurturc.hi

19. Pressure-SensitiveAdhesives and Forest Products765 Hiroshi Mizunlcrchi

20. 21.

Wood-InorganicCompositesas Shiro Suku Preservation of Wood

Prepared by the Sol-Gel Process

795

D u r r d D . NicAolcrs

22.

Preservation of Waterlogged Wood

807

David N.-S. Hot1

23.

Biodegradable Plastics from Lignocellulosics Muriko Yr)shioku m c l Nohuo Shirtrishi

74.

Recycling o f Wood and Fiber Products849 Tcrkcrrlori Arirrrn

25.

Pulping Chemistry 859 Giirn11 Gelle~rstcclt

827

781

Contributors

TakanoriArima Department of Biomaterial Sciences, Graduate School and Life Sciences, The University of Tokyo, Tokyo, Japan

of Agricultural

Jaime Baeza Departamento de Quimica, Facultad de Ciencias, Universidad de Concepcicin, Concepcicin, Chile Juanita Freer Departamento de Quimica, Facultad de Ciencias, Universidad deConcepcicin, Concepcicin, Chile Minoru Fujita Division of Forest and BiomaterialsScience,Graduate culture, Kyoto University, Kyoto, Japan Goran Gellerstedt Department of Pulp and PaperChemistry Institute of Technology, Stockholm, Sweden

School of Agri-

and Technology, Royal

Hiroshi Harada Division of Forest and Biomaterials Science, Graduate School riculture, Kyoto University, Kyoto, Japan

of Ag-

TakefumiHattori

Wood Research Institute,Kyoto University, Kyoto, Japan

David N.-S. Hon Carolina

School of Nature Resources, Clernson University, Clemson,South

FumitakaHorii

Institute for Chemical Research,

Kyoto University, Kyoto, Japan

TadashiIshii Division of Bio-Resources Technology, Forestry and Forest Products Research Institute, Ibaraki, Japan Akira Isogai Department of Biomaterial Science, The University of Tokyo, Tokyo, Japan Yuan-Zong Lai Faculty of Paper Science and Engineering, SUNY College of Environmental Science and Forestry, Syracuse, New York vii

viii

Contributors

NobuyaMinemura

Hokkaido Forest Products Research Institute,Hokkaido, Japan

Hiroshi Mizumachi

Professor Emeritus, The University of Tokyo. Tokyo, Japan

FumiakiNakatsubo Division of Forest and BionlaterialsScience,GraduateSchool Agriculture, Kyoto University. Kyoto, Japan Darrel D. Nicholas State, Mississippi MisatoNorimoto

Forest Products Laboratory, Mississippi State University, Mississippi

Wood Research Institute, Kyoto University, Kyoto. Japan

Shiro Saka Department of Socio-Environmental Energy Science,GraduateSchool Energy Science, Kyoto University, Kyoto, Japan KokkiSakai

of

of

Faculty of Agriculture, Kyushu University. Fukuoka, Japan

AkiraSakakibara Laboratory o f Wood Chemistry. Research Group of Bioorganic Chemistry, Division of Applied Bioscience, Hokkaido University. Sapporo. Japan Yoshihiro Sano Laboratory of Wood Chemistry. Research Group of Bioorganic Chemistry, Division of Applied Bioscience, Hokkaido University, Sapporo, Japan Mikio Shimada Wood Research Institute, Kyoto University, Kyoto, Japan Kazumasa Shimizu Division of Wood Chemistry, Forestry and Forest Products Research Institute. Ibaraki, Japan Nobuo Shiraishi Division of Forest and Biomaterials Science. Graduatc riculture. Kyoto University, Kyoto. Japan

School of Ag-

Toshiaki Umezawa Wood Research Institute. Kyoto University, Kyoto. Japan Mariko Yoshioka Division of Forest and Biolnaterials Science, Graduate School of Agriculture, Kyoto University, Kyoto, Japan

Ultrastructure and Formation of Wood Cell Wall Minoru Fujita and Hiroshi Harada Kyoto University, Kyoto, Japan

1.

A.

GENERAL STRUCTURE OFWOOD AND WOOD CELLS Wood

SoftwoodandHardwood In introduction it should be understood that the term “wood” refers to the secondary xylem formed by cell division in the vascular cambium of both gymnosperms (softwoods) and angiosperms(hardwoods). and especially i n Ginkgo. Similarsecondary xylem may be produced by plants of different form and structure, such as vines and shrubs, the xylem of which may be an important resource of pulping material. The structure and formation of the secondary xylem are discussed in this chapter. Both softwoods and hardwoods are widely distributed on earth, from tropical to arctic regions.The xylem of those species present in moderate-temperate to arcticregions is characterized by distinct growth rings, in which some anatomical differences can be noted. In the softwoods consisting mainly of tracheids (approximately 90% of wood volume), the latewood (summer wood) can be distinguished from the earlywood (spring wood) by its smaller radial dimensions and thickercells walls. Theseanatomicaldifferenccsare reflected in the higher density of the latewood compared with the earlywood. In softwoods growing i n tropical or warm areas, growth rings cannot be distinguished due to the indistinct boundary between earlywood and latewood. As with the softwoods, hardwoods are also present i n tropical to arctic regions. In colderregions,hardwoodspeciesaredeciduous, whereas i n tropical regions, they are predominantly evergreen and their growth rings are difficult to recognize. The macroscopic characteristics of hardwoods are reflected in the distribution and number of different ccll typessuch as vessels (pores),parenchyma, and fibers. Although fibers may account for only 25% of wood volume, in some cases, for hardwood, it may be as high as 50-70%. In contrast to the tracheid as the main cell in softwoods,hardwoods have a variety of cells. Some deciduous hardwoods such a s oak or elm have very large vessels concentrated at the beginning of annual rings. Suchwoodsare called “ring porous wood,”whereas otherdeciduousspecies and almost all evergreen hardwoods in which the vesselsare evenly dispersed over the annual ring are callcd “diffuse porous wood.” The above dis1.

1

Harada 2

and

Fujita

tinctions represent extremes and there are many intermediate arrangements of the vessels. Variations in arrangements of these vessels with other xylem tissues such as parenchyma are reflected in the “figure” and “grain” of the wood itself when it is cut from the tree. The physical properties of wood such as density also result from such arrangements of the cells.

2. Sapwood and Heartwood When a tree stem is cut transversely, a portion of “heartwood” can be seen frequently as a dark-colored zone near the center of the stem. This portion is surrounded by a lightcolored peripheral zone called “sapwood.” The sapwood or at least the outer part of the stem conducts water throughthe tissue where the water is transpired, and mineral nutrients are also carried with water from the roots into the wood. In addition, the sapwood has living parenchyma tissue, which often plays some physiological role such as the storage of starch or fat. From this point of view, the sapwood is considered an active xylem tissue. In contrast to sapwood, heartwood is dead xylem. As the tree matures, all parenchyma cells of the sapwood die, and other typesof cells such as tracheids or fibers become occluded with pigment composed of polyphenols and flavanoids supplied mainly from the ray parenchyma. The bordered pits of gymnosperms become aspirated, whereas the vessels are blocked by tyloses or gum in angiosperms. Thus, heartwood does not participate in water conduction. Although the conducting and physiological functions are lost in heartwood. the durability of wood against rot or insect decay is remarkably improved due to an addition of such pigments. Moreover, these pigments confer a variety of beautiful colors on wood. 3. Reaction Wood Reaction woods that appear on branches or a leaning stem by any force such as a landslide or snowfall have a peculiar nature. Once reaction wood is formed as a biological response, the living tree tries topreserve the original position of its stemorbranches.For the practical use of woods, the reaction woods have not been appreciated very much because of their different characteristics fromnormalwood in both a physical and a chemical sense. The occurrence and nature of reaction woods contrast quite a bit between softwood and hardwood. In softwood trees, the reaction wood forms at the lower side of a leaning stem or branches, where the compression stress reacts on the xylem. Therefore, this reaction wood is generally called “compression wood.” compression woodis heavy and appears dark brown on account of its highly lignified tracheid walls (see Section II), which seem to adapt to compression stress. Thus, compression wood is easily distinguished from normal wood by its dark color. The cambial activity at the lower position of a leaning stemorbranchacceleratesveryquicklyanddevelops a widercompressionareathan normal wood on the opposite side. Through the accumulation of compression wood tracheids over many years, a leaning stem will return gradually to the vertical position. The annual rings of such a stem, however, are conspicuously eccentric. On the contrary, reaction wood in many species of hardwoods is formed at the upper side of a leaning stem or branches where the xylem loads the tensile stress. Therefore, such reaction woods are called “tension wood.” Fibers of tension wood have a slightly lignified cell wall (see Section 11) that is adapted to the tensile stress just like a bowstring. It is not so easy to distinguish this area from a normal one on account of its slightly pale tone, in comparison to the case of compression wood.

Formation Ultrastructure and

Wall

of Cell

3

In fact, the occurrence of both reaction woods is averytroublesomeproblem in wood utilization. These reaction woods, however, are interesting material for the examination of wood structure and formation, as will be noted often in the following sections.

B. Wood Cells Wood cells are produced in the vascular cambium from two types of meristematic cells: the fusiform initial and the ray initial (Fig. l ) . Since cells derived fromthe fusiform initials that are upright in the stem occupy a major part of xylem, woods show remarkable anisotropism. The principal functions of xylem tissue are water conduction from roots to shoots, the mechanical support of a huge tree body, and a physiological role such as the storage of starch. Although these functions are common in both softwoods and hardwoods, the xylem of the latter is more evolved than that of the former, being adapted to each function.

softwood

hardwood

pits

%I

fusiform initials

a

&#

xial parenchyma cells

.B@'

ray tracheid FIGURE 1 hardwood.

axial parenchyma c e l l s

ray \ parenchyma ray cell initials p

Shapes of major wood cells fromthefusiformandrayinitialsinsoftwoodand

Harada 4

and

Fujita

I n softwoodsand Ginkgo, tracheids, beingmajorcells of xylem, are considered relatively underevolvedbecausetheyhavebothconductiveandmechanical properties. Bordered pits, the occurrence of which define a cell as a tracheid, are very important to the regulation of water flow. On the other hand, cell wall thickness is related directly to the strength of tracheids. The earlywood tracheids, therefore, seem to be well adapted to the conducting function whereas the latewood tracheids are loaded with the mechanical property, judging from their peculiar shapes. On the earlywood tracheids, well-developed pit pairs are distributed abundantly between the neighboring tracheids, and the cell walls of latewood tracheids are very thick. Only a small number of fusiform cells are subdivided into strand cells by horizontal partitions and compose an axial parenchyma. These parenchymatous cells survive in the sapwood for many years, being different from the tracheid, in which the protoplast is lost soon after differentiation (see Section III), and are part of some physiological functions. In somegenera of Pinaceae, axial resin canalssurrounded by epitherial cells are constructed. The occurrence and structure of resin canals are often used in the identification of softwoods, although the volume of such resin canals is very slight in wood. Ray cells are derived from the ray initials and elongated radially. A series of these ray cells make a ray parenchyma. Needless to say, these parenchyma cells are alive in the sapwood and are tied to the storage of nutrients such as starch or fat and also the transportation of some metabolites between the phloem and the heartwood. As a result, they must be related to the secretion of heartwood substance into the tracheids. Also, in some genera of Pinaceae, radial resin canals surrounded by epitherial cells are formed in many ray tissues, and more ray tracheids occur in the ray tissues. Hardwood xylem can be characterized by the development of vessel elements and wood fibers specialized for water conduction and the mechanical property, respectively. The vessel elements construct a very long and thick tube, namely, a vessel, being joined vertically with one another by a perforation that has a more developed style compared with the bordered pit pairs between tracheids. The occurrence of perforation distinguishes the vessel elements from the tracheids. Wood fibers elongate remarkably and possess very thick cell walls. The most developed type of cell, having simple pits (see Section II), is called libriform wood fiber. On the other hand, there are some intermediating cells from the tracheids to the vessel elements or wood fibers, i.e., vascular tracheids, vascentric tracheids, andfiber tracheids. The fiber tracheids are often included in the categoryof wood fibers. because there is no need to separate them from the libriform wood fibers in the practical use of wood. Vessel elements, wood fibers, and various types of tracheids in the hardwoods lose their protoplast just after the development of their secondary wall. However. in some hardwood species specialized wood fibers that remain alive for several years and often store starch grains are formed; they are called “living wood fibers.” Axial parenchyma cells, which are dispersed on the transverse section of softwoods, are clustered at the vessel periphery or form a group that is often linked tangentially. Resin canals that are surrounded by epitherial cells are formed in many genera of Dipterocarpaceae and a few Leguminosae. Ray parenchyma cells sometimes aggregate and develop a so-called broad ray. The broad rays make a peculiar figure on a board. especially on the radial surface, as observed in oak or beech. Cells contained i n the ray also vary in their anatomical features. Some of them are upright or square at the marginal position. These variations are used for the identitication of hardwoods [ l ] . Both axial and ray parenchyma cells are apparently concerned with physiological functions-for instance, the storage of nutrients or heartwood

Formation Ultrastructure and

Wall

of Cell

5

formation. Radial resin canals or latex tubes are formed in the ray tissue of some tropical hardwoods.

II.

ULTRASTRUCTURE OF WOOD CELL WALL

Wood is a natural composite material and a chemical complex of cellulose, lignin, hemicelluloses, and extractives [2]. Cellulose is the framework substance, comprising 40-50% of wood in the form of cellulose microfibrils, whereas hemicelluloses are the matrix substances present between cellulose microfibrils. Lignin, on the other hand, is the encrusting substance solidifying the cell wall associated with the matrix substances. The significance of lignin as the encrusting substance can be demonstrated by examination of the lignin skeleton created by the acid removal of carbohydrates (Fig. 2). The roles of these three chemical substances in the cell wall are compared to those of the constructing materials in the structures made from the reinforced concretein which cellulose, lignin, and hemicelluloses correspond, respectively, to the iron core, cement, and buffering material to improve their bonding.

A.

Cellulose Microfibrils

The crystalline nature of cellulose in wood has been demonstrated by studies with X-ray diffractometry and polarization microscopy. This crystalline nature was also confirmed by the electron diffraction patterns of the secondary walls of wood cells in selected areas [3]. Figure 3a isatransmissionelectronmicrograph of a longitudinalsection of latewood tracheids of Pinus densifloru, showing the intercellular layer (I), and the S, and S, layers. The electron diffraction diagram is of a selected area in S2 (Fig. 3b), which is represented by a small circle. The (101), (loi), and (002) of the equatorial reflections and (040) of

i

FIGURE 2 Electron micrograph of ultrathin transverse section of earlywood tracheids from Pinus densgora, showing thedistribution of lignin inthe cell wall, which was skeletonized using the hydrofluoric acid technique.

6

Fujita and Harada

FIGURE 3 (a) Electron micrograph of ultrathin longitudinal section of tension wood fibers from Pinus densiporu. (b) The corresponding diffraction diagram taken from the encircled area.

Formation Ultrastructure and

Wall

of Cell

7

the meridional reflection can be seen. It should be noted that crystallographic planes are based on the Meyer and Misch (1937) model of the unit cell of cellulose I, i n which the h axis (the fiber axis) is vertical. I t iswell known that in the wood cell wall, celluloseexists in the form ofthin threads with an indefinite length. Such threads are called cellulose microfibrils, and they play an important role in the chemical, physical, and mechanical properties of the wood. The greenalga, Kdorzia. which is oneform of Chlorophyceae,hasbeenstudied intensively by microscopists and crystallographers as an excellent material for the ultrastructural study of cellulose microfibril. Why then is W o t z i a used for the study of the cellulose microfibril of the wood cell wall? Because the cell walls of Valonia are unlignified, their microfibrils are readily isolated. Furthermore, as described later, Vrtlonicc microfibrils are approximately 20 nm in width, which is about five times larger than those of wood, and they are highly crystallized. However, the difference between algal microfibrils such as those of Vhlorli~zand ordinary ones produced by the higher plants also must be stressed. One of the differences is the selectively uniplaner orientation of algal microfibrils, that is, the ( 101) plane facing the cell surface, while cellulose microfibrils of higher plants are randomly oriented, although both microfibrils are laid along the cell surface i n their longitudinal direction [3,4]. The other is the crystallographic heterogeneity in algal microtibrils as detected by NMR [ S ] , and a triclinic system mixed with an ordinary monoclinic system was detected by electron diffraction [6]. The interface between these systems is not yet shown, although the former amounts to about 50%.

1. Dimensions of the Cellulose Microfibril As described above, it is clearly demonstrated through electron microscopy that the cellulose molecular chains are organized into strands as cellulose microfibrils. Figure 4 shows transmission electron micrographs of disintegrated cellulose microfibrils negatively stained withuranyl acetate. Figures .la and 4b, respectively, show the microfibrils of klonicr tnc~cmphyscrcell wall and the holocellulose of Pirlus drnsijur-a. A discrepancy in the size of the crystalline region of cellulose, obtained by X-ray diffractometry and electron microscopy, led to differing concepts as to the molecular organization of microfibrils. Frey-Wyssling 171 regarded the microfibril itself as being made up of a number of crystallites, each of which was separated by a paracrystalline region and later termed“elementary fibril” by Frey-WysslingandMuhlethaler 181. The term “elementary fibril” is therefore applied to the smallest cellulosic strand. Muhlethaler [ 10,l 11 applied this term to the cellulose fibril with a diameter of approximately 3.5 nm, using the negative-contrastpreparationtechnique for electron microscopy.Preston and Cronshaw [91,on the otherhand,considered the microfibril tohaveasinglecore of cellulose crystallite surrounded by a paracrystalline region. The width of cellulose microfibrils is reported to vary in different cellulose materials [ 121. For instance, as shown i n Fig. 4, Vrrlorzia cellulose microfibrils, being about 20 nm wide, are much larger than those of wood holocellulose. Shown in Table 1 are the crystallite size and microfibril width for several cellulose materials [ 131. The crystallite size was estimated with Scherrer’s equation at the reflection (002) or (101) of X-ray diffractometry, whereas the microfibril widthsweremeasured directly from the electron micrographs. The width range and mode width are also included in this table. It should be noted that the size of crystallites varies in different sources of cellulose materials, for results from both X-ray diffractometry and electron microscopy. According to Heyn [ 141, the negative stain can penetrate only the regions accessible to water. Thus, the translucent parts seen on the electron micrographs correspond to the

8

Fujita and Harada

FIGURE 4 Electron micrographs of the cellulose microfibrils of Vuloniu mucrophysa (a) and of Pinus densifloru holocellulose (b) (disintegration, negatively stained with uranyl acetate), showing the difference of cellulose microfibril width between wood and Vuloniu.

Formation Ultrastructure and

Wall

of Cell

9

TABLE 1 Crystallite Size and Microfibril Width

Crystallite Microfibril size"width Samples Pinus dens$ora

2.02 002)

-

Untreated 2.76Holocellulose

(2.5)b

Populus euramericana

4.1

layer Gelatinous Normal wood Valonia 15-30

(002)

2.2 (002) 14.3 11.9 (101)

(20.0)b

'Reflection examined. hModewldth. Source: Ref. 12.

crystalline regions of cellulose. Therefore, the difference in the microfibril width must be ascribed to that in the size of cellulose crystallites. In addition, the values obtained are not always equal to the 3.5 nm in elementary fibrils proposed by Muhlethaler [ l l ] .

2. Cross-Sectional View of Cellulose Microfibrils Figures 5a and 5b are similar electron micrographs of the ultrathin cross section of cellulose microfibrils from Valonia macrophysa and the gelatinous layer of Populus euramericana tension wood fiber. These were obtained by means of diffraction contrast in the bright-field mode foran epoxy resin-embedded section. This technique reveals a crystalline region as a dark zone dueto electron diffraction. Thus, cellulose microfibrils have a highly

FIGURE5 Electron micrograph of ultrathin transverse section of cellulose microfibrils (diffraction contrast in the bright field mode), showing their cross-sectional views: (a) from Valonia macrophysa; (b) from G layer of Populus eurarnericana.

Harada10

and

Fujlta

crystalline nature. It is interesting to note in Fig. 5a that a Vuloniu microfibril does not have any subunits corresponding to the elementary fibrils [13,17]. Additionally, cellulose microfibrils appear to be almost square in their cross section in both wood and Vuloniu [15-171.

3. Crystalline Structure of Cellulose Microfibrils Figure6shows Vuloniu macrophysu microfibrilsmechanicallydisintegratedwithacid, taken by diffraction contrast in the bright-field mode. Cellulose microfibrils can be seen as the dark areas, again indicating the highly crystalline structureof cellulose. However, the internal crystalline ultrastructure of cellulose microfibrils is not revealed by electron microscopic techniques such as negative staining and diffraction contrast, because lattice imagesof cellulose microfibrils are not obtained. The most important reason is that cellulose microfibrils are damaged by the electron beam and their crystalline nature is destroyed by irradiation under normal photographing conditions. Recently, the crystalline ultrastructure of cellulose microfibrils in Vuloniu macroby a specially developed technique for taking highphysu cell wall has been revealed resolution lattice images [15,16]. Figure 7 is an example of the lattice fringe substructures from disintegrated cellulose microfibrils. This micrograph shows the lattice image of 0.60 nm, corresponding to thatof the (101) plane. The lattice spacingof 0.60 nm is also shown in the electron and optical diffraction patterns. The lattice lines are observed regularly at about 20 nm width across the cellulose microfibril and are also visible along its length for more than 50 nm without any disruption. Figure 8 shows images of the cross section of cellulose microfibrils obtained using ultrathin sections. Lattice lines at0.60, 0.54, and 0.39 nm are visible in this figure. Therefore, a single microfibril is indicated as the individual crystal. Accordingly, it is suggested that the crystal line subunits as 3.5 nm elementary fibril and periodicity in its length does not exist inside the cellulosemicrofibril. Unfortunately, lattice images of cellulose microfibrils have not yet been taken in wood cellulose, since wood cellulose has low crystallinity and the size of the cellulose

FIGURE6 Electron micrograph of cellulose microfibrils fromVuloniu mucrophysu (disintegration, diffraction contrast in the bright-field mode), showing the crystalline nature of cellulose microfibrils.

Formation Ultrastructure and

Wall

of Cell

11

FIGURE 7 Lattice image of a disintegrated cellulose microfibril of Valonia mcrophysa, showing the lattice spacing of 0.60 nm.

FIGURE 8 Lattice images of the cross-sectional face of cellulose microfibrils from Valonia macrophysa, showing the lattice spacings of 0.60, 0.54, and 0.39 nm, respectively.

Fujita and Harada

12

microfibril is rather smaller compared with that of Valonia. In the near future, beam damage at room temperature against wood cellulose microfibrils would be reduced at least 10 times with cryo-electron microscopy. The cellulose microfibrils of the gelatinous layer of poplar (Populus eurntnictrna) tension wood in disintegrated samples are found to have many kinks, suggesting that the cellulose microfibril is highly crystalline [13]. However, the cellulose microfibrils of the gelatinous layer, about 100 nm in length prepared by ultramicrotome, become shorter than their original length upon hydrolysis [ 131. As a result, the crystalline regions in the cellulose microfibril of wood cell wall are thought to havesomecrystallinedislocations caused by chain ends [ 181. The cellulose microfibrils consist of a core crystalline region of cellulose surrounded by paracrystalline cellulose and short-chain hemicellulose. Lignin encases them and binds them into a rigid structure of wood cell wall.

B. Cell Wall Layers and Lamellae At the first step of differentiation of a woody cell, the living protoplasm produces a primary wall (P) that can be extensively increased in its surface as the cell develops. The substance between the primary walls of adjacent cells is called the intercellular layer (I) or the middle lamella. Since it is difficult to distinguish the region between the I layer and the P wall in the mature cell wall, the termcompoundmiddlelamella (CM) is generally used to designate the combined I layer and the two adjacent P walls. After the enlargement of the cell ceases. the cell wall layers are formed by the apposition of wall substances onto the inside of the primary wall. These wall layers are called the secondary wall (S) Although the primary wall andsecondarywallare classified by the ontogenetic process of plant cells, actual layered structures have been examined by the orientation of cellulose microfibrils. As a result the concept of lamellae, which are composed ofvery thin layers of only one or two cellulose microfibril width, is introduced. Cell walls were thickened by the appositional supply of these lamellae from the protoplast, so cellulosic interlamellae bridges are not accepted in the concept. The lamnella structure on the secondary wall is interesting in both physical and chemical properties of wood. Kerrer and Goring proposed a composite model with hemicelluloses and lignin [IS]. Although it is very intelligent, actual microfibril orientation on a lamella may fluctuate more [2O,2 I ] .

1.

Tracheids and Fibers

Figures 9 and I O are polarized photomicrographs at crosscd polars of transverse sections of tracheids and fibers, respectively. Both reveal the three-layered structure of the cell wall due to the differences in the orientation of cellulose microfibrils. According to the concept of Kerr and Bailey [22], normal wood cell wall consists of P and S walls, and the S wall is composed of a relatively narrowor thin outer layer (S,), an inner layer (S3), and a relatively thick middle layer (S?). However, the P wall cannot be distinguished in the figure due to the strong birefringence of the S , layer adjacent to the P wall. The S , and S, layers appear bright in the photographs, whereas the S, layer is at total extinction. That the birefringence of the S, layer occurs to a lesser degree than that of S , in the fibers of F q u s crewtcl indicates the poordevelopment of the S,. Despitesubsequentextensive studies with electron microscopy, the concept and terminology described above are still commonly accepted. Figure 1 1 is an electron micrograph ofan ultrathin transverse section from Cty7tomer-iajapotzicn, stained with silver protenate. It shows the intercellular layer (I), different

Ultrastructureand Formationof Cell Wall

13

I

FIGURE 9 Polarized-lightphotomicrograph of transverse section from earlywood tracheids of Pinus c/ensiforu, showing thethree-layeredstructure of the cell wall due to the birefringence of cellulose microfibrils.

layers of the secondary wall (S), and the warty layer (W) in an earlywood tracheid. The same layering structure from an earlywood tracheid of Pinus densiflot-a is shown more clearly in a longitudinal section that was skeletonized by the hydrofluoric acid technique (Fig. 12). Figure 16, (pg. 18). shows the texture of the P wall diagrammatically. The microfibril orientation in the primary wall was interpreted by the multinet growth hypothesis proposed by Roelofsen [23] and supported for the differentiating conifer tracheids by Wardrop[24].

FIGURE 10 Polarized-lightphotomicrograph of transverse section from wood fibers of Fugus crenatcr, showing the same structures as in Fig. 9.

14

Fujita and Harada

FIGURE 11 Electronmicrograph of ultrathintransverse section of an earlywood tracheidfrom Cryptomeria japonica, showing I, S,, S*, S1,and W (warty layer) at the final differentiating stage of the cell wall.

FIGURE 12 Electronmicrograph of ultrathinlongitudinal section of earlywood tracheidsfrom Pinus densifom (skeletonized cell wall with the hydrofluoric acid method), showing the I, P, S,, Sl, and S3 of the cell wall.

Formation Ultrastructure and

Wall of Cell

15

In the multinet hypothesis, the microfibrils are first deposited transversely to the cell axis and passively shifted longitudinally during cell extension. From an opposite viewpoint, an orderedfibrilhypothesiswasproposed by Roland et al. [25] in order to interpret the crossed polylamellated structure in the primary wall of parenchyma cells. According to this hypothesis, whether the orientation of microfibrils becomes transverse, oblique, or longitudinal is determined at the time of deposition of cell wall and may not be changed thereafter. Recently, Fujii et al. [26] proposed a modified multinet hypothesis of microfibrils orientation in the primary wall. The difference between this conceptand Roelofsen’s theoryisthattheshift of microfibrilorientation during cell extension is made in the individual lamella and each lamella becomes thin on the outer surface of the P wall due to extension. The three layers of the secondary wall, designated S,,S2, and S3,are organized in a plywood type of construction. The S , or S3,with a large microfibril angle to the cell axis, is designated as a flat helix, and the S2,with a small angle, as a steep helix (see Fig. 16). It is also shown that the layers themselves are of lamellae of microfibrils with varying amounts of shift in orientation, visible in the transmission electron micrograph. The S , is composed of several lamellae with alternating S and Z helices of microfibril orientation [28,29], and this structure in the S , is termed “crossed fibrilar texture” [28]. Figure 13 is

FIGURE 13 Electronmicrograph of theradialinnersurfacein a differentiatingtracheidfrom Pirzus densgoru (direct carbon replica), showing the microfibrillar orientation of the newly deposited microlamella crossing that of the underlying microlamella in S,.

Harada16

and

Fujita

a transmission electron micrograph of a replica of the inner surface of the differentiating early wood tracheid of Pinus densgora forming the S,, showing the criss-crossed texture of the microfibril orientation in the two different lamellae. The middle layer of the secondary wall (S,) is the thickest within the layers of the secondary wall. Therefore, the S, contributes most to the bulk of the cell wall material and is a compact region in which a high degree of parallelism of microfibrils exists. The S, isathinlayer of flathelices of microfibrilorientationasseenin S,. As opposed to the highly oriented S,, the S, is loosely textured. The S,, birefringent to a somewhat lesser degree than theS, in wood fiber, shows that this layer is poorly developed. Althoughthe S2 exhibitsamicrofibrillarorientationwithsteephelices,there are transition lamellae on its inner and outer surfaces. Several lamellae in these regions show agradualshift of microfibrilanglesbetween S, and S, andbetween S, and SJ [30]. However, the gradual shift of microfibril angles is more abrupt between S, and S, than between S, and S,. The transition lamellae in the secondary wall are not detected in TEM micrographs of ultrathin sections, since this lamella is relatively thin compared with the S, and S,. The method for evaluating microfibril angles in the secondary wall of wood cells was proposed by Yamanaka [31]. Figure 14 is a "EM micrograph of a transversely oblique section of an earlywood tracheid from Pinus dens.ijZora (stained with KMnO,). The curve through black dots shows the changes of the microfibrillar angle with respect

F FIGURE 14 Electron micrograph of ultrathin oblique section of an earlywood tracheid in Pinus densijffloraand microfibril angles in the secondary wall: top, I. bottom, lumen side.

Formation Ultrastructure and

of Cell Wall

17

to the tracheid axis from the top S, to the bottom S,. The horizontal line in the upper part of the figure shows the angles of microfibrillar orientation, the symbols (-) and (+), respectively, referring to Z and S helices. The gradual changes of the microfibrillar angles from S, to S, and from S, to S, are shown there. The helical cellulose microfibril orientationin the S, is typically demonstrated in Xray diagrams of wood [32]. The arcs at 0.39, 0.54, and 0.60 nm in the X-ray diagram of wood show that the cellulose crystallites (microfibrils) lie in a helix around each wood fiber or tracheid. The microfibril orientations are Z helices in the S, and S helices in the S,, although S, is the crossed arrangement of S and Z helices. Preston [32] suggests that the structure with various microfibril angles in the secondary wall passes through only one cycle, but this may be the brief duration of wall thickening in higher-plant cell walls compared with that in algae. Roland and Mosiniak [33] presented a diagram regarding the changesof cellulose microfibril angles in the case of a secondary wall of tracheids and wood fibers (Fig. 15). Figure 15 illustrates the case between the S, and S, layer. The change of microfibril angle is regular and continuous between the S , and S, layers,butitstopsduringthe deposition of the S , layer. Afterwards the change of microfibril angle reopens toward S, layer deposition. The texture of cellulose microfibrils in the P and S walls of softwood tracheids and hardwood fibers is shown as a schematic diagram in Fig. 16. The thin primary wall (P) consists of a loose aggregation of microfibrils oriented more or less axially to the cell axis on the outer surface. The S, layer is a flat helix but with crossed structure, whereas the S, layer is a steep helix and the S, layer is a flat helix. There are intermediate layers: the S,,, present between the S , and S, layers; and the S,,, between the S, and S, layers. The spiral thickening is the ridge of microfibrils that exist on the inner surface of the S, layer. The spiral thickening is considered part of the S, layer becauseof its continuity with the S, layer and parallel arrangement to the S, layer microfibrils.

FIGURE 15 Schematic diagram of the change of microfibrilorientationfrom three-layered structure of the cell wall. (From Ref. 32.)

S, and Sz in the

Harada18

and

FuJlta

FIGURE 16 Schematic diagram of the microfibril orientation in the primary wall and different layers of the secondary wall from tracheids and fibers: Po,PI;outer and inner parts of the primary wall; SlzrS23, intermediate layers between S , and S , and between S2 and S,, respectively.

The warty layer is one of the major structural features of wood cells found by electron microscopy [34]. It was first foundin softwood tracheids and laterin the tracheids, vessels, and wood fibers of hardwoods (see Fig. 11). The major chemical constituents of warts arereported to be lignin and hemicelluloses according to examination by component removal treatment of ultrathin wood sections [35]. The warts are believed to arise from the extra wall materials and remains of cytoplasm that are deposited on the S, layer through the plasma membrane [36,37]. The warty layer is not found in all softwoods and hardwoods [30,38]. Parham and Baird [39] have pointed out that the appearance of warts in wood has a phylogenetic trend. Softwood tracheids and primitive hardwood cells nearly always have warts, but as the cell types become more advanced or specialized, they become wart-free.

2. Vessels The texture of cellulose microfibrils in the walls of specialized cells such as vessel elements and parenchyma cells cannot be readily described as in softwood tracheids and hardwood fibers. A concept of standardized cell wall organization in vessel elements was, however, represented by Kishi et al. [40,41]. The microfibrils in the primary wall extend straight and are arranged parallel to one another within one lamella, and the wall consists of three parts, P-outer, P-middle, and P-inner, each showing a different microfibril orientation. The microfibrils are oriented transversely with respect to the vessel axis in the Pouter and are oriented at random in the P-middle. The P-inner consists of a crossed polylamellatedstructure. It isalsoreportedfromthe examination of vessel elements from nearly 30 Japanese hardwoods with polarizing and electron microscopy that the layered structure of the secondary wall can be classified into three categories: the typical threelayered structure, an unlayered structure, and a multilayered structure. The typical threelayered structure consists of S,, S2,and S3 similar to those of softwood tracheids and hardwood fibers, although the S, and S3 layers are thicker than those of tracheids and

Formation Ultrastructure and

Wall of Cell

19

wood fibers. The unlayered structure has only microfibrils, with the orientation of a flat helix. The multilayered structure has more than four layers, in which microfibril angles to the vessel axis change. This type of structure contains in some cases the so-called bowshaped pattern. Figure 17 isa TEM micrograph of the transverse sectionof Cinnamomum camphora and shows the microfibril angle and helix in the part of the bow-shaped pattern appearing on the vessel wall of the multilayered type of structure. As shown in Fig. 17, the pattern results from the progressive changes of microfibrillar orientation in the wall from 90" to 0" and from 0" to 90".

3. Parenchyma Cells In spite of the fact that parenchyma cells had been generally considered to have only primary wall, thoseof wood are reported sometimes to develop secondary wall, in addition to complicated primary wall. It is evident from recent studies that ray and axial parenchyma cells in both softwoods and hardwoods have variations or complexities in their wall structure that are not observed in the cell walls of tracheids and wood fibers. In softwoods, the cell wall structure of the ray parenchyma cells was divided into five categories by Fujikawa and Ishida [42]. However, as shown in Fig. 18, it is fundamentally classified into two types; the firsttype consists of the primary walland protective

b

90 FIGURE 17 Electron micrograph of ultrathin oblique sectionof a vessel wall stained withKMnO., from Cinnamomum camphora, showing a bow-shaped pattern (a) and the microfibril angles and helices (b).

20

Fujita and Harada

L-

i

I l

I I

P

S1

i

S2

1

I

FIGURE 18 Schematicdiagram of themicrofibrilorientation in the cell wall of softwood ray parenchyma cell: (a) the first type; (b) the second type. (From Ref. 41.)

layer (Fig. 18a), and the second type consists of the primary wall, secondary wall, and protective layer (Fig. 18b) [42]. However, the protective layer and a random arrangement of microfibrils is omitted in this figure. The P, appears with microfibrils of almost parallel orientation to the ray cell axis, the P, with the network appearance of microfibrils, and the P3 with several crossed polylamellate at microfibrillar angles of 30-60". It is interesting to note that the ray parenchymacell wall in thediploxylem of Pinus develops in two stages: that is, the primary wall and inner protective layer are fornled in the sapwood, and just before the heartwood is developed, the secondary wall and protective layer are deposited. In the axial parenchyma cells of softwood, the cell wall texture is very similar to that of ray parenchyma cells, except that the microfibrils are arranged in a flat helix with respect to the cell axis in the P , . In hardwoods, the primary wall ofray parenchyma cells has the so-called polylamellated structure proposed by Chafe and Chauret 1431. It was pointed out by Chafe and Chauret [43] that an isotropic layer and protective layer characterize the layered structure of the secondary wall of xylem parenchyma cells in hardwoods. According to examinations of thechemicalcomponents of these two layers using aseries of treatments on serial ultrathin sections, both a protective layer and an isotropic layer are rich in hemicelluloses and contain some pectic substances and cellulose microfibrils, but they have little lignin at the first stage of their developing process and become lignin-rich after the deposition of the inner secondary walls on them [44]. Consequently, both layers are considered the

Ultrastructureand Formation Wall of Cell

21

same in their origin and are called “amorphous layer” by Fujii et al. [M].Figure 19 shows electron micrographs of transverse sections by ray parenchyma cell from Tiliu juponicu; Fig. 19a shows cell walls skeletonized with hydrofluoric acid, while Fig. 19b shows sodiumchloride-treatedcellwalls.Blackzonesshow an amorphous layer indicating the presence of much lignin (Fig. 19a), but these disappear through delignification as seen in Fig.19b. It has been reportedbyFujii et al. [45] fromtheexamination of ray and axial parenchyma cell walls from about 50 species of Japanese hardwoods that the secondary wall is composed of a lignified cellulosic layer (CL) and an amorphous layer (AL) and that the cell wall structure can be classified into three types according to the presenceand organization of these two kinds of layers. Figure 20 is a schematic diagram of the cell wall organization of hardwood ray parenchyma cells: (1) 3CL-type, (2) 3CL+AL-type, (3) 3CL+AL+IL-type. CL refers to the lignified cellulosic layer that is similar to the ordinary wood cell wall, whereas ICL refers to the lignified cellulosic layer inside the amorphous layer (AL). The 3CL-type wall structuremay be considered thestandard structure of parenchyma cells of hardwoods, whereas the 3CL+AL-type wall structure occurs in cells that have extensive pit contact with vessels.

FIGURE 19 Electron micrographs of ultrathin cross section of the ray parenchyma cell from Tilia japonica, showing the amorphous layer (AL) of the secondary wall: (a) delignified cell wall; (b) cell wall skeletonized using hydrofluoric acid treatment.

Harada22

and 3CL

1-q

AL

Fujita

ICL

.: .............,..,..........:..-

x:

(a)

(C)

FIGURE 20 Schematic diagram of the cell wall organization of hardwood ray parenchyma cell, showing three types of wall structure: (a) 3CL; (b) 3CL AL,(c) 3CL + AL ICL.

+

4.

+

Reaction Wood Cell Wall

As described above (see Section I), softwood reaction wood is called compression wood and hardwood reaction wood is called tension wood. Figure 21 is a polarizing micrograph of compression wood tracheids from Pinus densijlora, and it demonstratesthat the S, layer present in normal wood tracheids is lacking. This is clearly shown in an electron micrograph of a cross section of a compression wood tracheid from Pinus densijioru (Fig. 22), and the presence of deep spiral checks in the S, layer is also revealed. The microfibrillar orientation of the S, layer is nearly 45",

Harada24

and

Fujita

cellulose of the G layer is highly crystalline. Its microfibrils are oriented parallel to the longitudinal axis of the fiber and the G layer is easily separated from the remainderof the fiber wall. Another structural feature of the tension wood fiber wall is that the G layer deposits on any one of the normal three secondary wall layers, S,, S2,and S,. The secondary wall of the tension wood fiber consists of three types, that is, S, G , S, S, G , and S, Sz + S, + G, depending on the wood species or part within a stem. Consequently, the G layer is called “the S, layer” when we refer to the S,, Sz,and S, layers.

+

C.

+

+

+

Sculpturing of the Wood Cell Wall

Cellulosic fibers such as cotton, ramie,and jute are relatively simple, smooth-walled composites of lamellae, but in wood the cell walls are almost invariably interrupted by gaps (pits) and sculpturing features.

1. Pit Structure Pits are gaps in the secondary wall of wood cells. There are two types of pits: bordered pits and simple pits. Generally, pits are present as pairs between two adjacent cells: bordered pit pairs, simple, and half-bordered pit pairs. In softwood, the pit border region of the cell wall is composed of border thickening (BT), S,, S2,and S, from the outer part of the cell wall as shown in Fig. 24. The presence of BT and thicker S, are features of the pit border wall. The microfibrils circle at theBT and sweep around thepit at the individual layers S,, Sz,and S,. In softwood bordered pit pairs, many species show a thickening at the center of the pit membrane. The torusis suspended from fine cellulosic strands to form a margin around the torus as shown in Fig. 25. The margin consists of an open net of

FIGURE 24 Schematic diagram of pit border organization in bordered pits of softwood tracheids. BT, initial pit border.

Formation Ultrastructure and

of Cell Wall

25

FIGURE 25 Electronmicrograph of the surface of pitmembranefrom Cryptomeria japonica (direct carbon replica), showing the pit membrane structure. T, torus; M, margo.

radially oriented microfibrils superimposed on an unoriented primary wall network, and it extends from the torus to the pit border. The torus is generally convex lens-shaped in cross section. On the other hand, the torus is seldom thickened in other cases. The former is true in species of the Pinaceae and Sciadopityaceae families, and the latter case involves species of Ginkgoaceae, Taxaceae, Chephalotaxaceae, Cupressaceae, Podocarpaceae, and Araucariaceae. The pit membrane of a half-bordered pit pair between tracheids and ray or axial parenchyma cells is quite thick. There is no torus in the center of the pit membrane, and no openings can be seen even at high magnification with an electron microscope. The central feature of the membrane structure of simple pit pairs in the interparenchymatous pits is the presence of plasmodesmatal pores. In hardwoods, the cell wallof the pit border consists of BT, P, S,, S*, and S3in tracheids and fiber tracheids of hardwood, like softwood tracheids. However, the pit border of vessels lacks not only BT but also S, in some parts of the pit border region [47]. The pit membrane of the bordered,half-bordered, and simple pit pairs in hardwoods is equal in thickness, exhibiting the primarywall texture, and there is usually no evidence of a torus. However, the presence of a torus in the intervessel pit membrane is reported in several species of hardwoods [48]. The pit membrane of simple pit pairs has plasmodesmatal pores as seen in softwoods.

Fujita and Harada

26

2. Vesture Pits In hardwoods, the pit chamber and pit apertures that are decorated by outgrowths of wall material are known as vestured pits. The outer growths of vestured pits are constructed chemically of lignin, hemicelluloses, and a little pectin [35].The shape and size of the outgrowths of vestured pits are variable. The development of vestured pit outgrowths is regarded as similar to that of warts. 111.

GENERAL DEVELOPMENT OF WOOD AND WOOD CELLS

A.

Vascular Cambium and Cambial Activity

One of the characteristic features ofa tree is the formationof the vascularcambium cylindrically surroundingastem,branches,and roots. The vascularcambiumproduces xylem inward and phloem outward. This sequence allows a tree to make itself a huge body. The cylindrical vascular cambium occurs through a series of developing meristem, namely, the apical meristem, the occurrence of procambium in the ground meristem, the growth of the vascular bundle, and the connection of intrafasicular cambium by the development of interfasicular cambium. The vascular cambium is composed of two types of meristematic cells. One is the fusiform initial occupying the major part of meristematic cells, and the other is the ray initial. Through their active cell division, parts of xylem and phloem are produced. However, since their activity in cell division is to a great extent affected by the season and weather, the result is the formation of annual rings in temperate regions. These initials must also multiply themselves on the tangential plane according to the increment of stem diameter. These two types of cell divisions can be distinguished by the direction of the division. The former division is defined as “periclinal division,” and the latter is called “anticlinal division” (see Figs. 26 and 27). Periclinal division is the mostimportant in view of woodformationandthus is discussed in detail. Cell division of the initial is extremelyrapid in spring. Moreover, several derivative cells (xylem mother cells) just inside the initial also have the ability to multiply through periclinal division. It is practically impossible to determine the true initial cell among these dividing cells. Therefore, just for convenience, a group of these cells is consideredcambialcellsand their area is called the cambialzone. In softwoods, the fusiform cells derived from the cambial zone differentiate directly into the tracheids except in onlyafewcasesinvolving the formation of the parenchymastrand,whereasthey differentiate into vessel elements, wood fibers, and several types of tracheids and parenchyma cells in hardwoods. Carnbial activity and the derivative differentiation are very important sequences in the growth of trees, environmental preservation of forests, and production of wood as a biomaterial. That is, they are the major sink of organic substances which are synthesized on leaves by CO, fixation, and then the major source of other life activities such as insects and also human beings. A detailed review of the vascular cambium has been published by Larson [491.

B.

Differentiation of Wood Cells

The tern1 “differentiation” has several meanings in the fieldof biology. I n this chapter, the term will be applied to the restricted case of the process of cell devclopment from the just-forming state i n the meristematic tissue to the mature state at which it is accomplished.

P

..

R.

l!

I f

FIGURE 26 (a) Light micrograph around the cambial zone ( C ) ,phloem (Ph), and enlarging xylem (E) from a transverse section of Robinia pseudoacacia. Most fusiform cambial cells are undergoing periclinal division, except for a trace experiencing anticlinal division (arrow).(b) Electron micrograph of fusiform and ray cambial cells. (c) Cytoplasmic feature of enlarging cells. 27

28

Fujlta and Harada

FIGURE 26 Continued

For instance, the differentiation of tracheids implies their maturing process from birth at the cambial zone to death after the secondary wall formation, by which both water-conducting functions and mechanical properties are given to the tracheid. The method of differentiation of parenchyma cells is quite different from that of tracheids, because they have only the primary wall or an underdeveloped secondary wall on the primary wall. They may be already functioning at thecambial zoneand have the ability to redifferentiate. Therefore, their differentiation is not addressed here. First of all, the differentiationof softwood tracheids, which is the most basic process of wood cell formation, will be discussed in detail. When a specimen block around the cambial zone is taken from the stem of a living tree and then a transverse section is observed under a light microscope, it is noticed that cells are piling up on a radial row from the mature phloem to the mature xylem through the cambial zone (Fig. 27a). The differentiating zoneof tracheids is located between the cambial zone and the mature xylem area. If the whole life of a particular tracheid from birth to death could be traced in situ in a tree stem, the tracheid differentiation would be clearly elucidated. However, it is really impossible to do so because the cells must be fixed with some reagent to preserve their cytoplasmic structure. Regrettably, their dynamic cell actions evolve into static phase by fixation. Therefore, the differentiating process of a tracheid must be deduced from the static cell structure of a series of differentiating tracheids. From this point of view, the differentiating zone of earlywood tracheids is favored for the precise examinationof their differentiation. In the spring, the production of tracheids from the cambial zone is very

D

3

Q

n 0

7

3

s

0

3

FIGURE 27 (a) Light micrograph of the cambial zone (C) and the derivative tracheids in five differentiating stages (RE, S,, S,, S3, and F) between phloem (Ph) and mature xylem (MX)from Cryprorneriujuponicu. (b) Enlarged view of S, depositing cells.

h)

W

Harada30

and

Fujlta

FIGURE 27 Continued

constant and, as a result, a series of differentiating tracheids is lined up in an orderly fashion along a radialrow from the just-formed stage to the mature stage. This series can be considered a good substitute for the life story of a tracheid, and since it is possible to trace the series using many microscopic techniques, the differentiating processof a tracheid can be grasped dynamically by tracing these differentiating tracheids along radial rows (Fig. 27a). Thedifferentiation of tracheids will be separatedintoseveral developing stages. Tracheids are pushed out in an inward direction from the cambial zone so as to begin enlargement. In the case of tracheids, the enlargement proceeds mainly in the radial direction, whereas enlargement in the tangential and longitudinal directions is very slight. Therefore, it may be appropriate to call this stage the radial enlarging (RE) stage. This fact results in the thinner radial walls of tracheids and the reorientation of cellulose microfibrils that may occur during the extension of the wall. The tracheid in this stage is composed of primary wall similar to the wallof cambial stage (C) cells. The thinned wall is recovered by the supplement of new wall materials on the inner surface. The extended wall is so fragile that it is often damaged and tom off during sampling of a specimen block from a living stem. After the enlargement of cell size, tracheids thicken secondary wall layers with the formation of the S,, S?, and S, layers. These stages are performed by the active deposition of cellulose microfibrils. However, the outermost region of the cell wall, including the intercellular layer, the cell comers, and the primary wall, is lignified during the S, stage. This lignification, which will be called “intercellular layer W i g n i fication,” may play an important role in stabilizing the cell size and conjugating the differentiating cells with one another. This I-lignification is accomplished in the middle phase of the S2 stage. Hemicelluloses are also supplied just after the deposition of cellulose

Formation Ultrastructure and

of Cell Wall

31

microfibrils (see Section IV). The secondary wall, which is still porous and flexible after the deposition of hemicelluloses, is encrusted with lignin and becomes very rigid. The lignification of the secondary wall, which willbe called “S-lignification” in contrast to “I-lignification,” is the most active after the S, stage,namely, in thefinal (F) stage of differentiation, although its initiation can be detected already during the S, stage. In this F stage some decorative elements such as warts or helical thickenings are added on the inner surface of the wall. After the wall layers develop, tracheids lose their cytoplasm by autolysis. Amorphous substances that have embedded the pit membrane also dissolve enzymatically sometime in the F stage. Tracheid differentiation is completed as this point and water conduction is achieved in the mature xylem (MX). The differentiation of vessel elements is characterized by enormousexpansion in both the radial and tangential directions. Although the developing stages of tracheids cannotbe applied directly to those of vessel elementsdue to a different secondary wall structure, the relationship of enlargement to secondary wall thickening and lignification is consideredsimilar to the sequence of tracheid differentiation. Needless to say, the formation of perforation pores is completed by the disappearance of the membrane itself, apart from the removal of only an embedding substance in the bordered pit pairs. On the contrary, the differentiation of wood fibers is characterized by the remarkable elongation in cell length that occurs at cell tips [50], and the other properties of differentiation are quite similar to those of tracheids. In hardwood, although the differentiation of both vessel elements and wood fibers proceeds simultaneously, vessel elements differentiate faster than wood fibers. How long a wood cell needs for its differentiation is also an important question. The time requirement for differentiation has been deduced by several methods, but the results are conflicting. A detailed timerequirementwascalculated for young trees of several softwoods by means of periodic inclinations for the internal date marking on the xylem. By these markings and the cell numbers contained in each differentiating stage, a time requirementofabout three weeks for passingthrough the five developingstages of a tracheid (RE, S , , S?, S,, and F) was calculated [51].

C. Cytology of Wood Cells Cambial cells, the differentiating cells of the tracheid, vessel element, and wood fiber, and also living parenchyma cells possess protoplast in their cell lumens. The most peculiar cytoplasmic structure of the fusiformcambialcells is the existence of ahuge central vacuole (CV) (Fig. 26b). This vacuole is maintained during the differentiation of tracheids (Fig. 27b), vessel elements,and wood fibers, whereas the cytoplasmicregion(Cy) is restricted to the very narrow area between the plasma membrane (Pm) and the vacuole membrane tonoplast (T) (Fig. 26c). On the contrary, the ray cambial cells and their derivative parenchyma cells are full of cytoplasm in their cell lumen, although several smaller vacuoles sometimes occur (Fig. 26b). In the axial parenchyma cells formed by the redivision of a young fusiform derivative, the central vacuole becomes small, and the cytoplasmic area expands in the reverse way. In spite of the cambial zone and differentiating xylem existing under the circumstance of very high pressure between the bark and mature xylem, the cellscontained in this areahaveonlya thin wall. Althoughvacuolation is generally considered a symptom of cell decay, the conspicuous vacuolation of these cells is supposedtoplay a very important role in the sustainment of their cell shapeunder presure. The enlargement of cell volume also depends on the turgor pressure of the vac-

32

Fujita and Harada

uole. In fact, it can be pointed out by arealneasurements that the vacuole is the best developed of the cells at the RE stage, when cells are just expanding (Fig. 28a). A nucleous is located around the central position of a fusiform cell in the longitudinal direction [SO], but on the transverse plane it is still pushed to one side of the cell lumen by vacuolation (Fig. 26c). In the cytoplasm, ordinary cell organelles such as Golgi bodies (Go),rough and smooth endoplasmic reticula (r-ER and S-ER). mitochondria (M), plastids (P), small vesicles (v), ribosomes, microtubules, and so on, are contained in a very narrow cytoplasmic region, although the occurrence of these cell organelles except microtubules between plasma membrane (Pm) and tonoplast (T) is not so abundant during the difl’erentiation of tracheids, wood tibers, or vessel elements. On the contrary, the cytoplasm of differentiating ray and axial parenchyma cells is crowded with many cell organelles. Especially, starch grains in the plastids and lipid droplets are very abundant, and r-ER are also well developed, whereas microtubules are very scarce. Ray cambial cells and mature ray cells arc almost identical to the differentiating parenchyma cells in their cytoplasmic features(Fig.26b). However, the number and size of starchgrains and lipid droplets contained in mature parenchyma cells change during a year 152-541. Thecytoplasmicfeatures of tracheidschange 21 little both i n quality and quantity according to their differentiation. The increase and decrease of the cytoplasmic area and its constituents of cell organelles were revealed by the combined use of light microscopy (Fig. 28a) and electron microscopy (Fig. 2%) on the differentiating zones of normal and compression woods of Cqptmtwricr juponicu. Areas of cell outline (A,,,,,,,;,,),cytoplasmicsurface (A,,,,,,,,,),and central vacuole (A,,,,,,,,,) were measured on an enlarged light micrograph of the transversesections o f differentiating tracheids using a digitized system connected to a computer (Fig. 28a). The nxasuretnent was performedalongthedifferentiatingtracheids, which were numbered from the initiation of the S , stage, and about 30 radial rows were surveyed. These radial rows of tracheids were sectioned at random in their longitudinal direction, so that the average value of tracheids of the same cell number reflects the makeup of volume i n each region. Areas of cell wall (A,v,J and cytoplasm (A,,.,,,,,,,,,,,) can be calculated by finding the remainder between those of the cell outline, the cytoplasmic surface. and the central vacuole, respectively. These values are diagramed in Figs. 29 and 30. I t should be noted that the cytoplasmic volume of both the normal and compression woods has two peaks during tracheid differentiation. The earlier peak i n both cases is at the intermediating phase from the S , stage to the S, stage, whereas the later one is located just prior to the initiation of the S , or F stage. On the contrary, during the S2 thickening, the cytoplasm is rather poor. Following this, proportions i n RE, S , , early S?, middle S,, late S,, and S, stages (Fig. 29) show the relative constituents of major cell organelles surveyed by electron microscopy (Fig. 2%). The general change in these cell organelles can be grasped by inultiplying the relative value by the total area of cytoplasm diagramed in Fig. 29. In addition to the changes in these cell organelles, the plasma membrane, important to the transportation of materials in and out of the cytoplasm, is always observed during tracheid differentiation and disappears after the development of the cell wall. The cytoplasmic features of differentiating wood fibers and vessel elements are also similar to those of tracheid differentiation, although the vacuolation of vessel elements is Inore extreme. In some species, such as acer or black locust, the living wood tibers are formed during the later period of a growing season. Their protoplast remains after the development of a cell wall and stores many starch grains in the cytoplasm for several years. Therefore, the mature xylem i n the sapwood is composed of ray and axial parenchymacells and sometimes the living wood fibers as the cells have aprotoplast.The

Formation Ultrastructure and

of Cell Wall

33

8 5

.. B

l I

FIGURE 28 Light (a) and electron (b) micrographs of transverse sections of S? depositing tracheids from Cryptomeria japonica. Areas of some cell organelles, for instance, Golgi bodies (Go), were measured with electron micrographs such as those shown in (b).

Fujita and Harada

9

B

r 6 S 4

3 2 1

0

0

c e l l number

C

RE

SI

s?r

Sam

Sar

S3

F

FIGURE 29 Changes of cytoplasmic volume during the differentiation of normal wood tracheids and proportions of some cell organelles at the stages of RE, S , , early S2, middle S,, late S?, S,, and F in Cryptonzeriu jqoniccr (see Figs. 28a and 28b).

FIGURE 30 A change of cytoplasmic volume during the differentiation of compression wood (see Fig. 28a).

Formation Ultrastructure and

Wall

of Cell

35

cytoplasmic features of these living cells are affected by the season, and also some of them seem to be specialized in their cell shape and cytoplasm. That is, the cells surrounding a vessel, particularly those directly contacted, become envelope-shaped and are very rich in Golgi bodies, r- and S-ERs, ribosomes, and mitochondria common to cells of active phase. On the other hand, their storage function seems to decay. These vessel-associated parenchyma cells are shown to concern the transportation of materials with vessel lumens [52,53] and also the formation of tyloses or gum that plugs the vessel lumen [55-571.

IV.

FORMATION OF WOOD CELL WALL

There is no doubt that cell walls are formed by the actions of cell organelles contained in each cell, even though some precursors of wall materials such as sugars may be supplied by the intercellular transport system.Therefore, cell wallformation is realized by the careful observation of cytoplasm that is undergoing cell wall development. It is also very important to select proper plant materials for precise examination of cell wall formation, becausegeneral plant cellsbearmanyphysiologicalfunctions in addition to cell wall formation. Moreover, the cell wall is composed of several types of chemical materials that are supposed to be metabolized by different cell actions, and their deposition on the wall may overlap. These complicated factorsare the major reason that the formation mechanism of plant cell walls has not yet been explained clearly, in spite of many investigations. Differentiating wood cells such as tracheids are very useful materials from this point of view. That is, they construct a very thick secondary wall, of which the ultrastructure and chemical components have been examined in detail, and the general sequence of cell wall formation can be traced through the series of differentiating cells along a radial row. Besides, the cell organellespossessed by thesecells are concernedonlywith cell wall formation, except vacuolation for the turgor pressure. In addition, if the depositing phase of individual wall materials such as cellulose, lignin, or hemicelluloses can be detected separately in differentiation, the relationship of cell organelleswith the metabolism of those materials would be grasped more clearly.

A.

CelluloseMicrofibrilDeposition

Cellulose is the mostbasic cell wall material in thewhole plant and it constructsthe framework structure of cell walls in the form of crystalline microfibrils as mentioned in Section 11. The formation has been studied using various plant cells from lower plants such as fungi or algae, to higher plants. Plasma membranes located just inside developing cell walls seem to be the most important cell organelles in relation to cellulose microfibril deposition. Although cross-sectional structures composed of unit membraneshadbeen observed by ordinary electron microscopic methods such as chemical fixation and ultrathin sectioning, faceviewsalongthemembranebecamepossiblewith the development of freeze-fracture or etching methods coupled with replication. Small particles on the outer surface of plasma membranes had been reported in various plant cells. The epoch-making discovery, however, was the characteristic assembly of granules located in the interior of the plasma membrane and revealed on the fractured surface in green algae such as Oocystis, Myclusteriu, or Vcloniu [58-601. Interesting structures have been reported using mainly single or naked cells such as algae [62], actobacteria [63,64], and cotton fiber [61], which can be frozen rapidly.

Harada36

and

Fujita

Small granule assemblies at the tips of microfibrils are called the terminal complex. These granules are considered to be the enzyme for the polymerization of cellulose molecules and their alignment i n the conlplex in relation to crystallization [64]. In underevolvcd algae such as vrtlonicr, the assemblies are large and linear, corresponding to their thick microfibrils [62]. On the other hand, evolved plants have small groups called rosettes [6 11. Thus, the form of the complex is considered to be related to the shape of the cellulose microfibrils and also to the evolution of plants. The polymerization and crystallization of cellulose microfibrils have been surveyed in detail using AcetoDactor- . x - y l i t z ~ ( ~which ~/, produces a thin cellulosic thread [6S] and has various mutants (661. The sequence of cellulose synthesis described above has not been traced in differentiating wood cells yet, because the freeze-fracture method is difficult to apply to them. However, cellulosic frameworks of wood cell walls are supposed to be constructed by a similar way. perhaps by a rosette. Although the freeze-fracture method isvery effective for visualizing characteristic structures such as the terminal complex on the membrane, the overall structure of differentiating wood cells depositing cellulose microfibrils must be examined by ordinary sectioning methods. Especially in wood cells depositing secondary wall layers, a control lnechanism for microfibrillar orientation is a very interesting viewpoint. Also, as the cellulose deposition is accompanied by the synthesis and accumulation of hernicclluloses and lignin, actions of various cell organelles must be traced in detail. Hence, the deposition phase of cellulose microfibrils i n differentiating tracheids can be traced in both normal and reaction woods. The phase can be detected by the increment of cell wall thickness (671, by means of autoradiography [68-70] (Fig. 3 l ) , and by chemical analysis of selectively collected rnaterials in some developing stagesof tracheids [ 7 I 731 and wood fibers [74](Fig. 32). The results obtained by these methodsshow that cellulose microtibrils are supplied to the wall mainly in the early and middle phases of the S, and S2 deposition stages. In addition to these deposition stages of cellulose microfibrils, most noticeable were the deposition of the G layer in the tension wood fibers and the S, thickening stage in the compression wood tracheids. This stage is composed of the deposition of cellulose microfibrils and is followed by the depositing stages of hemicelluloses and lignin. Compared with other differentiating stages of tracheids and wood fibers, the cytoplasm of cells forming the G layer isvery poor in its activity due to the fact that the region between the plasma membrane and the tonoplast is very narrow and cell organclles are rare there (Fig. 3%) 1751. The exceptionally abundant cell organelle in the cytoplasm is microtubules (MT). They are regularly distributed just inside the plasmamembrane (Pm), keeping a constant space of approximately 8 nm to the inner membrane and also between themselves (Figs. 33a and 33b). They are exactly oriented parallel to the depositing cellulose microfibrils in the stages of the G layer as well as the S , and S2 layers (Fig. 33b). The diameter of the microtubules is approximately 23 nm, and their numbers increase up to 20 per I p m of the plasma membrane, as calculated by their transverse direction. This abundant distribution results in the covering of about 40% of the cytoplasmic surface (Fig. 33b). A feasible link between the microtubules and the inner layer of plasma membrane is also discernible (Fig. 3321).These characteristics strongly suggest that microtubules and plasma membrane comprise the outermost complex of cytoplasm. On the other hand, there are only traces of Golgi bodies, S- and r-ERs, and the vesicles derived from them in the cytoplasm, in spite of the very active synthesis of cellulose microfibrils in this phase of the cell. On the contrary, in the beginning of the S, thickening stage of compression wood tracheids, in which cellulose microfibrils are supplied to the wall at 45" to the cell axis,

Formation Ultrastructure and

of Cell Wall

37

FIGURE 31 Serial light microscopic autoradiographs of “before section treatment” (a) and “after sectiontreatments”(b)withsodiumchloriteandhot 1.3% H,SO, from the differentiating compression wood tracheids in Cryptomeria japonica administered with 3H-glucose. Silver grains in (b) show the specific incorporation of radioactivity only on the inner surface of S , and S2 thickening tracheids, which reflects the deposition of cellulose by way of “apposition.” Removed activity can be detected in the intercellular layer of cells in the S , stage (arrows) and in the preexisting secondary wall of cells in the late S2stage (cells marked by an asterisk)by a comparison between (a) and (b) that implies lignin and hemicelluloses are supplied to the wall by wall of “intussusception.”

Harada38

and

Fujita

FIGURE32 Electron microscopic autoradiographs showing the incorporation of ‘H-phenylalanine in the transitional cell from S , to Sz,namely, in the stage of I-lignification (a), and from S , to S, in the stage of S-lignification in Cryptomeria japonica. Radioactivity can be observed around the intercellular layer and also within the Golgi bodies and vesicles in (a). In (b), vesicular inclusion is supplied to the wall by exocytosis (arrows) and radioactivity is often detected in such vesicles and the secondary wall.

cytoplasm isvery dense andwide. A similar complex between the microtubules and plasma membrane, however, is still observed [76]; microtubules are very abundant beneath the plasma membrane, having a link to it (Fig. 34a). The direction of microtubules is also in this case parallel with depositing cellulose microfibrils. Various cell organelles in the cytoplasm, such as Golgi bodies or ER, were shown to be involved in the synthesisof lignin precursors for the succeeding lignin deposition into the S, layer (see the next section). The characteristic appearance of microtubules is always applied to the cells depositing other wall layers such as S , , S2,and S3 without exception. It is interesting to note that the reorientation of microtubules precedes that of depositing cellulose microfibrils in the transition from the completion of a wall layer to the initiation of the next layer (Fig. 35). Moreover, the treatment of colchicine in which microtubule construction is obstructed by the formation of conjugation with microtubule protein “tubulin” resulted in a remarkable disturbance of depositing cellulose microfibrils (Fig. 36) [77,78]. These results clearly show that although microtubules present inside the plasma membrane cannot readily synthesize long and rigid microfibrils, these are involved in a great extent in the control of depositing cellulose microfibril orientation.

Formation Ultrastructure and

of Cell Wall

39

FIGURE 32 Continued

B. LigninDeposition Lignin is a very important cell wall component, particularly in wood cells, for the enhancement of the physical properties of cell walls and also for sealing the wall from prevention of waterleaks.Fortunately,lignincan be detected under an ordinary light microscope with the use of several stains such as a Wiesner reagent and Maule color reaction. Ultraviolet microscopy is especially useful for the quantitative analysis of lignin distribution in the cell wall [79] and more useful for studying the types of lignins present in the cell wall [80,81]. Needless to say, transmission electron microscopy coupled with potassium permanganate staining [82] or hydrofluoric acid treatment [83], electron probe microanalysis [84], and autoradiography [85-901 are also very useful for the observation of lignin from various points of view. The lignification of tracheid walls is generally known to last for a long period, from the S, stage to theF stage [79]. During this period, the lignification starts at the cell comer, spreads into the intercellular layer, and extends centripetally to the secondarywall. Lignin deposition, however, should be examined more closely in relation to the deposition of cellulose and hemicelluloses on each wall layer. This was attempted in the differentiating tracheids of compression wood, which were convenient for separating thelignification of the I region and S region because of their conspicuously highly lignified secondary wall, especiallyattheouterregion of the S, layer [67]. It has been shownthatthelignin deposition can be separated into two lignification stages, namely, I- and S-lignification. The former is active only during the early stage of secondary wall thickening, mainly at

40

Fujita and Harada

FIGURE 33 Electron micrographs of the cytoplasm-cell wall region of a transverse section from a tension wood fiber of Populus euramericana depositmg G layer (a) and of an obliquely sliced section from a S,-depositing fiber (b).

the S , stage, and is soon finished. The shape of I-lignification seems to stop the enlargement of cell size and adheres firmly between neighboring cells. On the other hand, the latter proceeds mainly after the development of a secondary wall framework, even though it begins at the middle phase of S , thickening. At any rate, lignin precursors permeate deeply into the cellulose microfibril framework of both primary and secondary walls and accumulate by way of “intussusception.” These two types of lignification were also applied in the differentiationof normal wood tracheids[86,91,92]. Moreover, when the speed

Formation Ultrastructure and .

.

Wall

of Cell

41

..... .,v*

(b) FIGURE 33 Continued

of lignin accumulation and the distribution of peroxidase were compared between the two regions [86], I-lignin seemed to be richer in the “condensed-type lignin” caused by the bulk polymerization than the S-lignin. This would be so because lignification proceeds with the higher content of lignin monomers and peroxidasein a rather large space without microfibrils. This assumption was confirmed by selectively labeled precursors coupled with light microscopic autoradiography [89,90]. When the lignifying cells are observed from the viewpoint of cytology, the cytoplasm is wider and denser than that of cells depositing cellulose microfibrils. Especially in the compression wood tracheids, an enormous amountof lignin precursors must be synthesized

Harada42

and

Fujita

Q t

p

FIGURE 34 Electron micrographs of 45"-inclined sections from compression wood tracheids in Cryptomeria japonica. (a) Shows the cytoplasmic feature in the cell just beginning S , deposition. At the cytoplasm-cell wall region (enlarged view), the distribution of microtubules (MT) is similar to that in Fig. 33a, although the cytoplasm is full of cell organelles, especially Golgi bodies. (b)

Shows the huge ridges and cavities in the of Golgi vesicles.

S , layer and poor cytoplasm after the active exmytocis

in their cytoplasm and thentransported from the cytoplasm to the wall. The area of cytoplasm becomes wider at the S , stage and also at the transition from Szto F (Fig. 30), where the cytoplasm becomes rich in Golgi bodies (Go) and ER (Fig. 34a). Although small vesicles (v) are produced mainly from Golgi bodies, they do not move to the cytoplasmic surfaceyet. These small vesicles increasein number and grow larger, occupying

43

Ultrastructure and Formation of Cell Wall

I

FIGURE 34 Continued

the largest part of the cytoplasm during the following late phase of the S, stage. S-lignification at the F stage is characterized by the active fusion of the well-developed vesicles to the plasma membrane and by the release of the vesicle inclusion to the wall area, namely, exocytocis. The cytoplasmic area resultsin the formationof an empty region after lignification (Fig. 34b). This sequence indicates thatlignin precursors are synthesized and stored in the vesicles that havebeen derived from Golgi bodies and on occasion from ER. In fact, the process was proven by autoradiography using tritiated lignin precursors [ S S ] . These sequences were also examined inboth lignifications at the I and S regions of normal woodtracheids(Figs. 32aand32b) [86]. InS-lignification, S-ER seems to be related to the lignification in addition to the Golgi bodies (Fig. 29), whereas I-lignification

Harada44

and

Fujita

;S1

FIGURE 35 Electron micrograph of an obliquely sliced cytoplasm-cell wall region from a tension wood fiber of Populus euramericana, which is just traveling from S2to the G layer. Many microtubules (MT) are oriented parallel to the fiber axis, although several (large arrowheads) are still oblique. Fine striations in the cell wall (small arrowheads) show the deposition of the S layer, and cellulose microfibrils of the G layer cannot be detected yet.

is performed mainly by the action of Golgi vesicles, similar to the case of compression wood tracheids. The lignin of compression wood tracheids is generally known to be rich in the condensed-type lignin. The cytoplasmic features of these lignifying cells seem to be consistent with the types of lignin suggested by Takabe et al. [86]. That is, the I- and S-lignins of compression woods and also the I-lignin of normal wood are metabolized mainly by Golgi bodies and the derivative vesicles, being richin the condensed-type lignin,

Formation Ultrastructure and

of Cell Wall

45

FIGURE 36 Scanningelectronmicrographs of theinnersurface of the developing S, layer of Crytomeria compression wood tracheids after incubatlon withoutcolchicine (a) and with colchicine (b). (a) Shows remarkably developed ridges and regularly depositing cellulose microfibrils parallel with the ridges, whereas (b) shows the disturbed deposition of them.

whereas the S-lignin of normal wood is synthesized through the cooperation of Golgi bodies, S-ER, and their vesicles, resulting in noncondensed-type lignin. It is interesting to note that the cytoplasmic regions becomewider, corresponding to both lignifications in the I- and S-regions in both normal and compression woods (Figs. 29 and 30). In addition, the peak of the S-lignification of compression wood is bigger than that of normal wood. The tendency of these peaks is to respond to the absolute amount of lignin that will be supplied to the separate wall regions. The precursors of lignin are most likely synthesized in the cytoplasm and stored temporarily, and then released from the cytoplasm to the wall, whereas the cytoplasm may possibly be rather narrow during the active depositing phase of cellulose microfibrils.

C.

HemicelluloseDeposition

Examination of hemicellulose deposition is divided into two groups. One covers the microscopic observations [68-70,931; the other is the chemical analyses of the tissues or cell walls collected selectively [71-741. Although the use of microscopy is a prerequisite for the observation of the microlevel localization of hemicelluloses, the specific staining method has not been improved enough to be applied on each hemicellulose and even mixed ones, being difficult to distinguish from cellulose or lignin. If any specific radio-

Harada46

and

Fujita

active precursor of each hemicellulose can be applied to the differentiating xylem, autoradiography will provide invaluable information on the deposition of hemicelluloses. A similar effect, although applied only to the total hemicelluloses, was achieved by a combined technique involving autoradiography and the removal of hemicelluloses from the tissue [93] or sections [53] that had been administered with tritiated glucose as the general source of cell wall materials. It becomesclear by these methods that hemicelluloses, although not so deeply as in the cases of lignin, accumulate in the preexisting framework of cellulose microfibrils by way of “intussusception” (Fig. 31). The depositing phase also intermediates the deposition of cellulose microfibrils with lignin accumulation. The deposition of each hemicellulose must be traced by the sugar analyses of tracheids or wood fibers selectively collected from the differentiating zone according to their development. This technique was achieved qualitatively by Meier et al. [7 l ] and improved quantitatively by Takabe et al. [73]. Judging from the content of polysaccharides in wood cell walls and the sugar constituents of hemicelluloses, glucose, mannose, xylose, arabinose, and galactose are, respectively, reflected in cellulose, glucomannan, arabino-gluconoxylan, and galactoglucomannan. As shown in Fig. 37a, mannose is supplied to the wall just after the cellulose microfibril deposition, followed by the deposition of xylose in the tracheid differentiation. In the case of wood fibers (Fig. 37b), xylose deposition follows directly the deposition of cellulose microfibrils. On the contrary, galactose and arabinose seemto be supplied to the walltogether in both stages of I- and S-lignification. The disagreement between the depositing manners of xylose and arabinose, namely, arabinose showing more affinity for lignin than xylose, may suggest that a chain of arabino-glucurono-xylan is not polymerized at one time but separately. That is, in concert with lignification, arabinose may be added, possibly as a side chain of the backbone of xylan already deposited.

I 0’

1

.

.

.

.

.

.

.

.

a

’

.

L

1 2 3 4 5 6 7 8 9 1 0 1 1 1 2 f rdctlonnumber

i r d c t l o nn u m b e r

Glucose

4 2 - 9

Hannole-A-,

rylor~-A-r

Ardblnose

- 0 - 1

GdIdctore”.--.

FIGURE 37 Depositionofpolysaccharidestocell wall duringdifferentiation ofnormalwood tracheids in Cryptomeria japonica (a) and of normal wood fibers in Juglans sieboldicrna (b).

Ultrastructure and Formation of Cell Wall

47

According to this speculation, galactose is also supplied to the glucomannose chain. On the other hand, chemical analyses of wood offer the evidence that groups of galactose, arabinose, and 4-0-methylglucuronic acid are combined directly between the polysaccharide and lignin in the so-called lignin-carbohydrate complex [941. These lines of evidence strongly suggest that the sugar groups forming branches of hemicelluloses are the ignition site of lignin accumulation. When the depositing periods and types of cell wall materials are coordinated with one another, the wood cell wall is concluded to develop by the following four processes: the appositional deposition of cellulose microfibrils on the preexisting wall, resulting in the construction of a framework of cell walls; the supply of hemicellulose main chains around the cellulose microfibrils and the reinforcement of the framework; the addition of hemicellulose branching chains such as galactose or arabinose;lignin accumulation starting on the branch and encrusting almost all spaces between the framework. The cytoplasmic relation to hemicellulose deposition has remained uncertain at many points. However, in contrast to the case of cellulose microfibrils that are synthesized at the surface of cytoplasm, the precursors of hemicelluloses are surely metabolized in the cytoplasm, judging from the combined observations from autoradiography and chemical treatments 1691.

REFERENCES D. A.Kribs. Bot. Gn:.. Y4:547 (193.5). H. A. Core, W.A. C M . and A. C. Day, Wood Structure r r t d Idmtijiccrtiorz. 2nd ed.. Syracuse Univ. Press. New York, p. 30 ( I 979). 3. J. F. Revol and D. A. 1. Goring. Wood Sei.. 14: 120 (1982). 4. J. Sugiyama. H. Chanzy, and J. F. Revol. Plmtcr, /Y3:260 (1994). S . H. Yamamoto and F. Horii. M~rcronlolrcrrlc.~, 26: 13 13( 1993). 6. J. Sugiyama, R. Vuong. and H. Chanzy. Mrrcror?loleclt/c.s,24:4168 ( 1991). I. 2.

7. A. Frey-Wyssling. Scirrm,, //Y:80 (1954). 8. A. hey-Wyssling and K. Muhlethalcr. Mrrkrortlol. Clwrn., 62:25 (1963). 9. R. D. Preston and J. Cronshaw. Nature, 18/:248 (1958). I O . K. Miihlethaler. Bioh. Z. S c h ~ ~ i :For.stwreitf.. . 30:55 (1960). I I . K. Miihlethnler. i n Ccllltlrrr Ultrtrstrrrcturc~of Wooc!\ Plat1t.s (W.A. C W . Jr., ed.).Syracuse Univ.Press,NewYork. p. 191 ( 1965). 12. V. Bnlnshov and R . D. Preston, Ncrtrrw. 176:64 ( 1955). 13. H. Haradn and T. Goto, i n Ce//lrlo.sc, t r t d Other Ntrt/trrr/ PnlynrcJr .Sy.stcr/r.s (R. M. Brown. Jr.. ed.). Plenum Publishing, New York. p. 383 (1982). 14. A. N. J. Heyn. J . Cell Biol.. 2Y: I81 (1966). I S . J. Sugiyama. H. Haroda. Y. Fujiyoshi. and N. Uyeda. Pltrr~ttr.Ihh: I61 ( 1985). 16. J. Sugiyomn. H. Harada. Y. Fujiyoshi. and N. Uycdo. Moku:tri Gtrkktrishi. 31:6 I ( 1985). 17. J. F. Revol. J . Mrrtrt: Sci. Lett.. 4 : 1347 ( 1985). 18. K. Muhlethaler. J. P o / w w r Sci.. C2N:305 (1969). 19. A. J . Kcrr ancl D. A. I. Goring. Wood Sci.. Y: I36 ( 1977). 20. K.Rue1 and F. Barnoud, ISWPC Stockholm. I : I I (1981). 31. Y. Kataoka. S. Saiki. and M. Fujita. Mokrr:rri Gnkkctishi, 38:327 (1992). 32. T. Kerr and 1. W. Bailey. ./. A r d d Arborotrr/r~,/5:327 ( 1934). 33. P. A. Roelotken. i n A d 1 w r x ~ c . sirr Bo/trrlictr/ Rc,srtrrch. Vol. 2 (R. D. Preston. d . ) . Academic Prcss. L,ondon a n c l NcwYork. p. 69 ( 1959). 24. A. B. W d r o p . A r ~ s t r d ..I. Hot.. h:299 ( 1958). 25. J. C. Rolnnd, B. Vam. and D. Reis. J . Cc,// Sci.. / Y : 1 3 9 (1975). 36. T. Fujii.Pl1.D.thesis.KyotoUniversity.Kyoto. Japan. 1981.

48

Fujita and Harada

27. 28. 29. 30. 31. 32.

H.Saiki. Mokuzui Gcrkkniski. /6:237 (1970). A.B.Wardrop. Hol;forschurlg, //:l02 (1964). Y. Imatnura,H.Harada,andH.Saiki, Bull. Kyoto Univ. For(>st,s,44: I83 ( 1972). H.Harada. Y. Miyazaki,andT.Wakashima, Bull. Go~vr.Forest Exp. Stu., 104:I ( 1958). K.Yamanaka,M.S.thesis,KyotoUniversity,Kyoto.Japan. 1969. R. D.Preston, The. Physictrl Biology of’ PlLmt Cell Walls. Chapman & Hall,London.p. 302

33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

SO. 5 1. 52.

53. 54. SS.

56. 57.

58. 59. 60. 61. 62.

63. 64.

65.

( 1 974). J . C.RolandandM.Mosiniak, IAWA BM//.,4:15 (1983). K. Kobayashiand N. Utsumi,unpublishedresults, 1951. N.Mori.M.Fjita,H.Saiki,andH.Harada, Bull. Kyoto U n i ~Forc..st.s,55:299 (1983). J. Cronshaw. Protoplasnl, 60:233 (1965). K.Takiya.H.Harada, and H. Saiki, Bull. Kyoto U I I ~ IForests, ! 48: 187 (1976). J. Ohtani, Bull. College Exp. Forests Hokkaido u/1j\J.. 33:4()7 (1979). R . A.Parhamand W. M.Baird, Wood Sei. Techrd., & l (1974). K.Kishi,H.Harada.andH.Saiki, Bull. Kyoto Unit,. Forc..sr,s.4Y:122 (1977). K.Kishi,H.Harada,andH.Saiki. Mok~cztriGtrkkaishi. 25521 (1979). S. Fujikawaand S. Ishida, Mokuzni Gakknishi, 2/:445 (1975). S. C. Chafeand G. Chauret, Protoplasm. ;YO:129 (1974). T. Fujii.H.Harada.andH.Saiki, Mok~rzniG~lkk.st.s, IJ. SO: 183 (1978). M.Fujita. Y. Shoji, and H. Harada, H u l l . Kyoto Urli1: Forrsts, 4Y: I16 (1977). N. Shibata,M.Fujita,H.Saiki.andH.Harada. &dl. Kyoto U I I ~ Forc..sts. IJ. SO: 174 (1978). R. M.Brown.Jr..and D. Montezinos, Proc. N d . Accrd Sci. USA. 73: l43 (1976). T.ltohand R. M.Brown,Jr.. Plarrtcr. /60:372 (1984). D.Montezinos,in The Cytoskeletou i r ~Plrrrlt Gron,th ctrld 1 l e w l o p r ) w ~ t(C. W. Lloid, ed.). Academic Press, London, p. 147 (1982). J. H. N. Willison, J. Appl. Po1yr11c.rSyrnp., 3791 (1983). A.M. C. Emons, i n Bio.syntlw.sis m r l BiotlrRrcr~lrtiorl (fCellulo.se (C. H.HaiglcrandP. J. Weimer, eds.). Marcel Dekker, New York. p. 7 I ( 1991). K. Zanr, .l. Cell Rio/.. A0773(1979). C. H.HaiglerandM.Benziman,in Cellrrlow C ~ I Other I ~ NLrturrrl Polyrr~crSy.sterr~.s(R. M. Brown, Jr.. ed.). Plenum Press, p. 273 (1982). C. H. Haigler, in Bio.syrrt/w.si.s c r r d Hioclrgrcrdrrtiorl o f Cdlulosc~(C. H. Haigler and P. J. Weimer,

eds.). Marcel Dekker, New York, p. 99 ( 1991). 66. S. Kt~ga,S. Takagi.and R. M.Brown, Jr., Polyrr~er.343291 (1993). 67. M.Fujita,H.Saiki,andH.Harada, Moklrzai Gakktrishi, 24:158 (1978). 68. M.Fujitaand H. Harada. Mokuzcri Gnkkaishi,24:435 (1978). 69. M.Fujita. K. Takabe.andH.Harada. Mokuzri Gdkoixhi. 27337 (1981). 70. K.Takabe.M.Fujita,H.Harada,andH.Saiki. Mokuzcri Gokknishi. -?(l:103 (1984). C. B.Wilkie. Hol~fi)rsckurlg,13: l77 (1959). 7 1. H.MeierandK. 72. H.Meier,in Bio.syr~rhe.si.sc m 1 BioLI~.grcr~Irriorlof’ Wood Corr~porz~r~ts (T. Higuchi. ed.). Academic Press. New York

(1985).

Ultrastructureand Formation of Cell Wall

73. 74. 7s. 76. 77. 78. 79. 80. 81. 82. 83.

84. 85. 86. 87. 88.

89. 90. 91. 92. 93. 94.

49

K. Takabe, M. Fujita, H. Harada. and H. Saiki. Mok~czoi GNkkaishi, 29:183 (1983). K. Takabe, M. Fujita, K. Tanaka, and H. Harada, Bull. Kyoto CJniv. Forests. .56:234 ( 1984). M. Fujita, H. Saiki, and H. Harada, Mokuzui Gakkaishi,20:147 (1974).

M. Fujita, H. Saiki, and H. Harada, Moktczai Gnkkaishi, 24:355 (1978). J. D. Pickett-Heaps, D e ~ dBiol.. /5:206 ( 1967). M.FijitaandH.Harada.“Colchicinetreatmentandmicrolibrilorientationindifferentiating compression wood tracheids,” Proc.. 27th A n r l d Meetirrg of Jup,crrwsc~Wood Resectr-chSociety. p. 311 (1977). A. B. Wardrop, TAPPI, 40:22S (l9S7). B. J . Fcrgus and D. A. I . Goring, H o l ~ f o r . s c l ~ ~ c24: r ~ 118 g , (1970). A. Yoshinaga, M. Fujita, and S. Saiki, Mokuzcli Gnkknishi, 38:629 (1992). M. Mauer and D. Fengel, Hol;for.schurlg. 44:4S3 (1990). I . H. Clark, TAPPI, 45:310 (1962). S. Sakaand D. A. I. Gorgin.in Biosy1lthesi.s crrltl Riorlesmrltrtiorl of’ Wood Cort~pot~c,r~t.s (T. Higuchi. ed.), Academic Press, New York, p. S 1 (1983. M. Fujita and H. Harada, Mokuzai Gakknishi, 25:89 (1979). K. Takabe, M. Fujita. and H. Harada. Mokuzcti Gtrkknishi, 31:613 (1985). N. Terashima. K. Fukushima, and K. Takabe, Holifi)r.sc~hwzg,40(Suppl.): 101 ( 1986). K. Fukushima and N. Terashima, J . Wood Cllern. T e c h o l . , /0:413 (1990). N. Terashima and K. Fukushima. Wood Sci. fi~cl~nol., 22:2S9 (1988). K. Fukushima and N. Terashima, Hol;fi~rsc.hur~g. 45:87 (1991). M. Fujita, K. Takabe, and H. Harada, ISWPC (Tsukuba), !:l4 (1983). K. Takabe, M. Fujita, H. Harada, and H. Saiki, Mokuzai Gczkkaishi, 27:813 (1981). P. M. Ray, J . Cell Biol.. 3.5:660 ( 1967). D. Fengel and G. Wegener, Wood, Walter de Gruyter. Berlin, 1984.

This Page Intentionally Left Blank

Chemical Composition and Distribution Shiro Saka Kyoto University, Kyoto, Japan

1.

INTRODUCTION

Wood is a complex of natural polymer substances: cellulose, hemicelluloses, and lignin. These polymer substances are not uniformly distributed within the wood cell walland their concentrations change from one morphological region to another. In order to understand the physical and chemical properties of wood, it is essential to study the topochemistry of these polymer substances. This chapter, therefore, deals with chemical composition andits distribution in normal and reaction woodsfrombothsoftwoodandhardwood species.

II. GENERAL FEATURES OF WOOD CELLS Woody plants have several different types of cells in bothsoftwoodsandhardwoods. However, the anatomy of softwoods is less complex than that of hardwoods. In softwoods, the principal types of cells are tracheid and parenchyma, whereas those in hardwoods are fiber, vessel, and parenchyma.Sincesoftwood tracheids andhardwood fibers constitute the majority of wood cells, they contribute in a major way to the physical and chemical properties of wood. The cell wall organization of typical softwood tracheids or hardwood fibers [ 1,2] is described in Fig. 1 . Basically, the cell wall consists of the primary (P) and secondary (S) wall layers. The P layer is formed during the surface growth of the cell wall, and the S layer is formed during the thickening of the cell wall. This layer is composed of three sublayers termed S , , Sz, and S3, based on differences in microfibril orientation. A layer called middle lamella (ML) is located between adjacent cells. Since it is difficult to differentiate the ML from the two P walls on either side, the term compound middle lamella (CML). which encompasses the ML and the two adjacent P wall layers, is frequently used (Fig. 2a). The pattern of the cell wall organization in reaction wood is somewhat different from that of normal wood [31. Compression wood, developed on the lower side of a leaning softwood stem or branch, lacks the S, layer but contains an extra layer of lignin [%(L)] located between the S , and S, layers (Fig. 2b). Tension wood, formed on the upper side of a leaning hardwood stem or branch, often lacks one or more of the three secondary 51

52

Saka

FIGURE 1 The gross structure of a typical softwood tracheid or hardwoodfiber.(Courtesy Prof. Emer. R. J. Thomas, North Carolina State University, Raleigh, NC.)

of

wall layers. Instead, the gelatinous layer (G layer) is usually deposited adjacent to the cell lumen (Fig. 2c). The G layer contains little or no lignin and consists mainly of cellulose microfibrils oriented parallel to the fiber axis.

111.

CHEMICAL COMPOSITION OF WOOD

A.

Normal Wood and Reaction Wood

The chemical constituents of wood are well known,and a numberof authors have provided an excellent review of this work [4-81. The major cell wall constituents are cellulose, hemicelluloses, and lignin. Other polymeric constituents, present in lesser and often varying quantities, are starch, pectin, and ash for the extractive-free wood. Tables 1 and 2 show comparisons of the chemical composition made by Time11 [9] for five hardwoodsand five softwoods, respectively. Although the cellulosecontent is more or less the same (43 -t 2%) for both groups, the hardwoods contain less lignin. The lignin content of hardwoods is usually in the range of 18-25%, whereas that of softwoods varies between 25% and 35%. However, tropical hardwoods can exceed the lignin content of many softwoods. The structure of lignin is different between these two groups: softwood lignins are composed mostly of guaiacyl units, whereas hardwood lignins consist of syringyl and guaiacyl moieties [ 101. The hemicelluloses found in these groupsvary both in structure and quality, as shown in Fig. 3. The predominant hardwood hemicellulose is a partly acetylated, glucuronoxylan (O-acetyl-4-O-methylglucuronoxylan),accounting for 20-35%, whereas softwoods containglucuronoarabinoxylan (arabino-4-0-methylglucuronoxylan)in therange of 10%. Hardwoods contain only a small quantityof glucomannan. In softwoods, however, a partly acetylated galactoglucomannan (0-acetylgalactoglucomannan) makes up as much as 18%. In additiontothesemajorcellwall components, pecticmaterialsandstarch are included in much smaller quantities inboth softwoods and hardwoods. Ash usually makes up between 0.1% and 0.5% of wood, but tropical species often exceed this range. Wood

FIGURE 2 (a) Cross section of brominated normal wood tracheids in Douglas fir [Pseudotsuga rnenziesii (Mirb.) Franco]. Transmission electron micrograph of ultrathin section. Dark zones indi-

cate the higher lignin concentration. (b) Cross section of compression wood tracheids in eastern white pine (Pinus strobus L.). Ultraviolet micrograph of thin section (upper portion) shows high concentration of lignin in the S,(L) layer indicated by an arrow. Polarized light micrograph (lower portion) shows a lack of the S3 layer.

53

Saka

54

4 FIGURE 2 Continued. (c) Cross section of tension wood fibers in Enoki [Celtis sinensis Pers vat. japonica (Planch) Nakai]. Transmission electron micrograph of KMn0,-stained ultrathin section.

Note a nonstained G layer deposited following the stained Prof. Emer. H. Harada, Kyoto University, Kyoto, Japan.)

S2 and S3 layers. (Courtesy of the late

in heartwood also contains varying quantitiesof extractives that are always more abundant than sapwood. The chemical composition of reaction wood differs from that of normal wood. Table 3 shows a comparison made by Timell [ 1l ] of the average chemical composition of normal and compression woods of many conifers. Pronounced compression wood contains, on average, 39% lignin and 30% cellulose, compared to 30% and 42% for normal wood,

TABLE 1 Chemical Composition of Wood from Five Hardwoods"

Cell wall constituent Cellulose Lignin Glucuronoxylan Glucomannan Pectin, starch, ash, etc. "All values in percent of Source: Ref. 9.

Acer rubrum

Betula papyrifera

Fagus grandifolia

Populus tremuloides

Ulmus americana

45

42

45

48

51

24

19 35 3

22 26 3 4

21

24

25 4 2

24

extractive-free wood.

1

19 3 4

4 2

"

omposition Chemical TABLE 2

55

Chemical Composition of Wood from Five Softwoods"

constituent wallCell Cellulose 27 Lignin Glucuronoarabinoxylan Galactoglucomannan Pectin, starch, ash, etc.

A hies hulsarnen

42 41 29 9 18 2

Picea glauca

Pinus strobus

41

41 29 9 18

13 18 1

3

Tsuga canudensis

33 7 16 3

TI1uja occidentalis 41 31 14 12

2

"All values in percent of extractive-free wood. Source: Ref. 9.

respectively. The content of galactoglucomannan is only 9%, half that in normal wood. The amount of xylan, on the other hand,is the same in thetwotissues.Compression woodcontains 2% of a 1,3-linked glucanand 10% of a galactan,bothpresentinonly trace amounts in normal wood. With regard to chemical composition of tension wood, the most characteristic feature is that it contains less lignin andxylan described as pentosan, but has much more cellulose

Saka

56

TABLE 3 Average Chemical Composition of Normaland Compression Woods of Softwoods"

Cell wall constituent Lignin Cellulose Galactoglucomannan 1.3-Glucan Galactan Glucuronoarabinoxylan Other polysaccharides

Normal wood Compression wood

30

39

42

30 9 2 IO

18

Trace Trace 8 2

8

2

"All values in percent of extractive-free wood. Source: Ref. I 1.

and galactose residues than normal woods (Table 4) [ 121. Furthermore, higher ash and uronicacidcontentsarereportedfortension woodthanin the side woodof Japanese beech (Fugus crenuru Blume) [ 131. The higher cellulose content in tension wood is due to the presence of a gelatinous layer, which is often quite thick and unlignified (Fig. 2c).

B. Tracheids and Ray Cells The tracheids in softwood or fibers in hardwood constitute more than 80% of the cells found in wood; thus, the global analyses reflect moreor less the composition of these types of cells. However, the chemical composition of ray cells may not be inferred from that of the whole wood. The data in Table 5 reveal a comparison made by Hoffmann and Timell [ 141 between the ray cells and tracheids from red pine (Pinus resinosu Ait.). The defibrated samples after delignification by acid chlorite were subjected to the separation of tracheids and ray cells through screening, followed by subsequent analysis of sugar residues. As reported in the literature [ 151, the ray cells contain more lignin. somewhat less cellulose, only half as much galactoglucomannan, and the same amounts of xylan and pectin as do the tracheids. A small amount of a 1,3-linked glucan also present in the ray cells appears to be absent

Chemical Composition of NormalandTension Woods from Euccclyptus gonioc.rr1y.r"

TABLE 4

Cell wall wood Tension wood Normal constituent 13.8 Lignin 57.3 Cellulose Pentosan Acetyl 7.4 Galactose residue

29.5 44.0

1s. I

11.0

3.0 2.5

I .9

"All values in percent of extroctlve-free wood. Solrrcr: Ref. 12.

omposition Chemical

l

57

Normal wood

wood" Colnpresslon

Tracheids cells Ray Tracheids Ray cells Cell

(%)

(%)

(c/o)

(c/o)

Lignin Cellulose Galactoglucomannon 1,3-Glucan Tracc Galactan Glucuronoarabinoxylan Pectin Other polysaccharides

40

28 42 20 -

40 35 II 2

40 30

Trace

IO

10 I

7 I

I

1

35 9 2

Trace 11

2 I

8 1 1

9 2

"From Ref. 14. hFrom Ref. 18.

in the tracheids. Overall, the results in Table 5 are in good accord with those by Perilii [ 161 and Perilii and Heitto [ 171. However, these authors indicated more xylan in ray cells than normal tracheids i n Scots pine (Pirzus sylvesfris L.). The lower content of xylan in Table 5 is due possibly to partial removal during chlorite delignification. Also included in Table 5 are the chemical compositions of tracheids and ray cells from red pine compression wood [ 181. Although compression wood tracheids have a different chemicalcompositionfromnormalwood tracheids, the ray cells in compression wood are chemicallyindistinguishablefromthose in normalwood.Comparedwith the compression wood tracheids, the ray cells have the same contents of galactoglucomannan, 1,3-glucan, and xylan. However, the galactan typical of grosscompressionwood is missing.

C. Earlywood and Latewood Meier [ 191 has studied the effect of the cell wall thickness on polysaccharide content by comparing earlywood and latewood from normal Scots pine(Pirzus sylvesfris L.). As shown in Table 6, the latewood contains more glucomannan and less glucuronoarabinoxylan than the earlywood. Since the proportion of the latewood tracheid S I layer to the whole wood is greater than that of earlywood, the observed differences are due mainly to the thicker S2 layer in latewood tracheids. I t may therefore be concluded that the tracheid S I layer hasmoreglucomannanand less glucuronoarabinoxylanthan doothermorphological regions, which agrees reasonably well with the results of Whiting and Goring 1201 for the secondary wall and middle lamella fractions of black spruce (Picecr rnnriarzcr Mill.). Also shown in Table 6 are the results of compression wood balsam fir [Abies ~ N I sanlecr (L.) Mill] 1211. Although earlywood and latewood have the same content of lignin [22],thepolysaccharidecomposition is different betweenthesetwo tissues. It canbe speculated in a similar way that the tracheid S, layer in compression wood contains more celluloseandgalactoglucomannanbut less galactan,arabinan,andxylan that doother morphological regions of compression wood.

58

Saka

TABLE 6 Polysaccharide Composition of Earlywood and Latewood from Normal Scots Pine and Balsam Fir Compression Wood

Balsam fir compression wood’ Polysaccharide

pine”

Scots

Normal

Latewood Earlywood Latewood Earlywood (%l

Cellulose 56.2 Galactan 3.1 Glucomannan Arabinan Glucuronoarabinoxylan

(%)

(%)

56.7

45.0

3.4 20.3

19.0 16.0

50.4 15.0 18.7’

1.o 18.6

(%)

24.8 1.8

14.1

0.9

0.6

19.1

15.3

“From Ref. 19. hFronl Ref. 2 I . ‘Values as galactoglucomannan.

IV.

DISTRIBUTION OF POLYSACCHARIDES

A.

Introduction

The distribution of cellulose is probably the easiest to study at the various morphological regions of wood.Onepossiblemethodinvolves the useofholocellulose after desired poststaining with heavy metals such as uranyl acetate [23-2.51 and lead citrate [26]. Although the orientation of the cellulose microfibrils is quite different in the various cell wall layers, cellulose is quite evenlydistributedthroughout the secondary wall. In the primary wall, however, microfibrils are rather loosely and randomly arranged (Fig. l), so the concentration of cellulose in the primary wall may be lower than that of the secondary wall. Unlikecellulose, a study of the distribution of hemicellulose is difficult. This is because histochemical techniques are generally nonspecific and frequently unreliable. In order to overcome these difficulties, BouteljeandHollmark [27] introduced the use of interference microscopycombinedwithenzymatic treatment. Sinner et al. 128,291 used electron microscopy to study the enzymatic degradation of the cell wall components by xylanases, mannanase, and avicellase for delignified spruce [Picea abies (L.) Karst.] and beech ( F c q p s sylvaticcr L.). It was found that xylan concentration is rather high in the S , and S, layers for both woods. Hoffmann and Parameswaran [30] made another attempt to study the polysaccharide distribution in spruce tracheids through oxidation of polysaccharides with heavy metal. Subsequent electron microscopic observations indicatedthe highest concentration of hemicelluloses in the S , layer. Awano et al. [31] have recently applied immunoelectron microscopy to studying the distribution of glucuronoxylan in buna (Fagus crencrtu Blume). An extensive study in the future will provide useful information on its distribution. The distribution of polysaccharides also has been studied by examination of holocelluloseskeletonsafterremoval of lignin with acid chlorite [21],atechniquefurther refined by Fujii et al. 1321by using ultrathin sections. An electron micrograph of the holocellulose skeleton from the compression wood of tamarack [Lcrrix laricim (Du Roi) K. Kochj is shown in Fig. 4. Forcomparison, Fig. 5 shows a micrograph ofa lignin skeleton of thc same wood. Although the presence of the residual lignin and some removal

omposition Chemical

59

FIGURE 4 Holocellulose skeleton of two tracheids in compression wood of tamarack [Lark lurkina (Du Roi) K. Koch]. Note the low concentration of polysaccharides inthe S,(L) layerand absence of substances in the middle lamella region. Transmission electron micrograph of a cross section. (Courtesy of Prof. Emer. W. A. C M , Jr., State University of New York, Syracuse, W.)

of polysaccharides may obscure the data, overall the results obtained by this method are in good agreement with the holocellulose distribution inferred from the lignin skeleton seen in Figs. 4 and 5 [33]. Some other methods also have been proposed; Parameswaran and Liese [34] have given an excellent review of these studies. For the localization of pectin, Albersheim et al. [35] have used the hydroxylamineiron method developed by McCready et al. [36,37] for the quantitative measurement of pectin. With this method, Parameswaran and Liese [34] have found a homogeneous distribution of pectin across the secondary wall. The middle lamella also was found to be highest in concentration. The use of ruthenium red and alcian blue also is proposed for staining pectin substances [38-401. For quantitative determinationof the polysaccharide distribution, the microdissection technique has often been used. One of the oldest is Bailey’s work in 1936 [41] for the pentosan content determinationof the middle lamella in Douglas fir. In 1959, Meier [42,43] adopted a similar technique for hardwood fibers (Betula verrucosa Ehrh.) and softwood tracheids (Pinus sylvestris L. and Picea abies Karst.) at different stages of development

60

Saka

FIGURE 5 Lignin skeleton of threetracheidsin compression wood of tamarack [ L a r i x luricina (Du Roi) K. Koch]. Note the high concentration of lignin in the S,(L) layer as well as in the middle lamella region. Transmission electron micrograph of a cross section. (Courtesy of Prof. Emer. W. A. CBtC, Jr., State University of New York, Syracuse, NY.)

that were microscopically distinguished, isolated, and subsequently subjected to microanalysis for sugar residues. From a knowledge of the chemical composition of different polysaccharides in wood, the contentsof polysaccharides at various morphological regions could be calculated. Although some doubt exists as to additional deposition of polysaccharides during the later stage of the secondary wall thickening, the technique developed by Meier remains applicable [44]. Later, Norberg and Meier [45] isolated the gelatinous layer(G layer) in tension wood fibers (Fig. 2c)by using ultrasonic treatment. Subsequent analysisindicated that it contains 98.5% glucose and 1.4% xylose, suggesting thepure cellulosic nature of the G layer. Luce [46] determined the radial variationin the content of hemicelluloses in softwood tracheids by a chemical peelingtechnique.Burkeetal. [47] measured the sugar content of the polysaccharides in the primary walls of a suspension-cultured Douglas fir. In 1981, Hardell and Westermark [48] have developed a method for peeling layers of the cell wall from a slightly delignified single tracheid of Norway spruce [Picea abies (L.) Karst.]. They reported only small differences in the relative amounts of polysaccha-

omposition Chemical

61

rides between the compound middle lamella and the secondary wall, a finding that is not in agreement with Meier's results [42]. It appears that a treatment of slight delignification may cause a partial dissolution as well as redistribution of hemicelluloses. The arabinose and galactose contents for the compound middle lamella were foundto be 7.3% and 7.6%, values that are considerably lower than those from nonlignified wood [6]. More recently, Whitinget al. [49]developedanothermethod of preparing wood tissue fractions from the compound middle lamella and secondary wall of black spruce (Piceu nzariana Mill.) by taking advantage of the difference in density ( p ) between lignin ( p = 1.4 g/mL)andpolysaccharide ( p zz 1.5 g/mL).Themost significant finding after analyses of carbohydrates for these wood tissue fractions [20] was that the concentrations of celluloseandglucomannan are smaller in the middlelamellathan in the secondary wall, whereas the concentrations of other polysaccharides are more or less the same in both the secondary wall and the middle lamella regions. Compared to the previous methods by Meier [42] or Hardell and Westermark [48], the method of Whiting et al. [49] is more reliable, due to only the physical treatment of specimens without introducing any chemical changes.

B.

Distribution of Polysaccharides in Normal Wood

Whiting and Goring [20] conducted carbohydrate analyses of fractions of tissue from the middle lamella and secondary wall of black spruce tracheids. Figure 6 shows the relationships between polysaccharide content and lignin content for the various tissue fractions. Each fraction is amixtureof the secondary wall, primary wall, andmiddlelamella in varying proportions. The fraction with a lignin content of 22% is from the secondary wall tissue, whereas the extrapolated results to a composition at which the cellulose content becomes zero would represent the polysaccharide composition of the true middle lamella with a lignin content of 70%. Figure 6, therefore, shows that the middle lamella contains less cellulose and glucomannan but more galactan and arabinan than the secondary wall.

t

U

0.2

0.3

0.4

0.5

0.6

0.7

FIGURE 6 Polysaccharide content versus the lignin content for thevarious black spruce (Picecc tnnriat~nMill.). (From Ref. 20.)

tissue fractions of

Saka

62 TABLE 7

l481

Relative Polysaccharide Percentages of the Secondary Wall Tissue

Polysaccharide Cellulose Glucomannan Glucuronoarabinoxylan 12.8 Galactan

WhitingGoring and [201

Meier Hardell Westermark and [421 63.0

1.1

0.0

Arabinan

58.1

60.0

20.8

23.7

14.3

10.7

4.8 2.0

4.1 1.5

However, the concentration of glucuronoarabinoxylan is essentially the same in both morphological regions. For the secondary wall tissue, Table 7 shows a comparison made by Whiting and Goring [20] of the relative polysaccharide percentages measured by Meier [42] and Hardell and Westermark [48] on Norway spruce (Picea abies Karst.) and by Whiting and Goring [20] on black spruce (Picea nzariarza Mill.). The values of Meier were calculated using a proportion of 90% secondary wall and 10% middle lamella for the whole wood [50].It is apparent that the agreement between the results obtained by three investigators is good, particularly between the data by Hardell and Westermark and those by Whiting and Goring. It istherefore likely that in the tracheid secondary wall thecontents of hemicelluloses decrease in thefollowingorder:glucomannan,glucuronoarabinoxylan,galactan, and arabinan. A comparison of the relative polysaccharide percentages of the middle lamella-rich fractions is shown in Table 8 for the same three investigators. The results of Whiting and Goring are the composition at the 70% lignin content in Fig. 6, which would be representative of the true middle lamella. For comparison, data from tissue fraction with 39% lignin (Fig. 6) are also included. This fraction includes part of the secondary and primary wall tissues, as well as the middle lamella fraction. In contrast to the excellent agreement for the secondary wall fractions (Table 7), the results on the middle lamella are at variance with each other. This is because the middle lamella fraction is most difficult to prepare in a pure state. It should be noted that the cellulose content of 50.3% for Hardell and Westermark (Table 8) is not much different from the value of 58.1 in Table 7 for the secondary wall fraction. Additionally, the results by Hardell and Westermark are i n good agreement with the data from the tissue fraction of 39%)lignin content by Whiting and Goring. Thus,

TABLE 8

Relativc Polysaccharide Percentages of Middle Lamella-Rich Fractions Hardell and Westermark Meier ~421

Polysaccharidc

(-)

50.3 Ccllulose Glucomannan

33.4 7.9 13.0 16.4 29.3

13.3 Glucuronoarnbinox~lat~ Galactan Arnbinan

Whiting and Goring [20]

148I (4I % lignin)

(70%. lignin) 0.0

50.8

22.6

12.5

21.6

15.4

(39% lignin)

37.5 7.6 6.2

29.2 20.8

7.2 5.0

Composition Chemical

63

a sample collected by Hardell and Westermark must be to some extent contaminated with the secondary wall fractions. It is of interest to note that both the results of Meier for the compound middle lamella and of Whiting and Goring for the true middle lamella show glucomannan to be the lowest among hemicelluloses in the middle lamella region. For overall trendsof the carbohydrate distribution across the cell wall, Meier [42,43] has indicated that, although the cellulose content is very low in the ML (middle lamella) and P (primary wall) regions, arabinan is almost completely confined to M P regions, and galactan is almost completely confined to M P S , regions. Glucomannan, however, increases from M + P to the S , layer in softwoods and remains at a rather constant low level in hardwoods. Glucuronoxylan in hardwoods has a higher concentration in the secondary wall than in the M P. Figure 7 shows the distribution of polysaccharides across the woodcellwall of Cryptomeria tracheids obtained by Takabe [U] through the technique of Meier [42]. Interestingly, cellulose is rich in the middle of the S , layer, whereas hemicelluloses of glucuronoarabinoxylan and galactoglucomannan are abundant in the S , and outer parts of the S, and S3 layers. The warty layer (W) is composed mainly of galactoglucomannan.

+

+ +

+

C.

Distribution of Polysaccharides in ReactionWood

C6t6 et al. [21] also used the technique of Meier [42] to study the polysaccharide distribution in compression wood tracheids of balsam fir [Abies balsamea (L.)Mill.]. Figure 8 shows the results obtained. It should be noted that the glactoglucomannan, arabinan, and xylan are homogeneously distributed across the secondary wall, whereas a higher concentration of galactan was found in the outer regionof the secondary wall. This fact was later confirmed by Larson [5 1,521. The content of cellulose is, on the other hand, higher at the inner portion of the cell wall. For the compound middle lamella (P M P), the high

+

+

:.:.:.:.:* ....:.:\.:\..:.., >>: ....... .:.:.:.:.:................. ....... ,...... :., ... ..... .A.

............... .....:..

...

... ...

m Glucuronoarabrnoxylan 0 blactoglucomannan 0 tcllulorc

FIGURE 7 The distribution of polysaccharides across the wood cell wall of tracheids in Cryptomeria japonica D. Don. (From Ref. 44.)

64

Saka

Percent

GALACTAN

ARABINAN

CELLULOSE

ARAEINOGLUCURONOXYLAN

GALACTOGLUCOMANNAN

FIGURE 8 Graphical representation of the distribution of polysaccharides in compression wood tracheids of balsam fir. (From Ref. 21.)

content of arabinan and galactan would be due to high pectin content. It is reported that chemical composition of the primary wall is the same in normal and compression woods [5 1,521.

V.

DISTRIBUTIONOF LIGNIN

A.

Introduction

Unlike polysaccharides, a number of reliable methods can be used to study the distribution of lignin in wood. One of the oldest procedures is selective staining, followed by study under the light microscope[53]. Although some doubt exists as to this specificity for lignin [54,55], potassiumpermanganatestaining [56] has been usedextensivelyforstudying lignin distribution by electron microscopy [57-601. Also reported were studies by electron microscopy of lignin skeletons (Fig.5 ) created by the carbohydrate removal by brown-rot fungi [61] or concentrated hydrofluoric acid [32,62-641. Although some alteration of the lignin through condensation may result and the possible presence of residual carbohydrates may obscure the data, overall the results obtained by this method are in reasonable agreement with those from potassium permanganate staining [59]. Although the above methods are useful in elucidating the presence of lignin in the various morphological regions of wood, they can provide only qualitative evaluation of thelignindistributionacrossthecellwall. For quantitativevisualization of thelignin distribution, ultraviolet (UV) microscopy with thin sections of wood has provided good results. This method was initiated by Lange [65], who estimated the weightconcentration of lignin to be, respectively, 16% and 73% for the secondary wall and compound middle lamella of Norway spruce tracheids. This result was in excellent agreement with Bailey’s value of 71% for the Douglas fir middle lamella fractions obtained by a direct analytical method [66]. Previous to Bailey’s work, Ritter [67] had concluded that approximately75% of the lignin in wood is located in the middle lamella, with the other 25% being located in the secondary wall. Apparently, a distinct difference exists between the results by LangeBailey [65,66] and Ritter [67]. However, considerable confusion has appeared in the lit-

Composition Chemical

65

erature. Some of this confusion could be due to the use of the symbol o/o to denote both the percentage fraction of the total wood lignin contained in a particular morphological region and the lignin content of that region. Therefore, g/g. i.e., g of lignin/g of cell wall substance. is used in this chapter to denote lignin concentration. The symbol o/o is then reserved for the proportion of total lignin in a particular morphological region. Later. Goring and co-workers [68-701 refined the UV microscopy method through a preparation of the thin section (0.5 p m ) to avoid errors caused by nonparallel illumination. Goring et al. then determinedquantitatively the distribution of lignin in wood [50,70-751 and proved that the result by Lange-Bailey[65,66] was correct.They also proved that the conclusiondrawn by Berlyn and Mark [76] is correct. that the middle lamella region can contain at most 40% of the total lignin in wood due to its small volume fraction of wood. In addition to these, they discovered that different lignins occur in different types of cells and different cell wall regions of wood [72-741. More recently, through the use of UV microscopy, Yang and Goring [77.78] have found that the secondary wall lignin of softwoodscontains twice as many phenolic groupsas the middlelamella.This finding was later confirmed by Whiting and Goring [79] from a study of the secondary wall and middle lamella fractions. Of other methods for the quantitative assay of the lignin distribution, Lange and Kjaer [SO] proposed the use of interference microscopy, and Boutelje [811 later refined this technique.More recently, Saka et al. [82-861 developeda new techniqueforthe quantitative determination of the lignin distribution in wood. The method involves a specific bromination for lignin in a nonaqueous system (CHCI,). Bromine concentrations in the various morphological regions of wood are then determined by electron microscopy (TEM or SEM) coupled with energy-dispersive X-ray analysis (EDXA). By knowing the lignin reactivity toward bromination, the distribution of lignin can be determined for various morphological regions of wood. Figure 9 shows the direct comparison made between two techniques of UV microscopy and EDXA measurement in bromination [S61 over the

1 .oo Earlywood

Latewood

I

0

.-C .-

l

SECONDARY WALL

0

I

I

I

15

10

5

I

I

1

5

10

Cell number FIGURE 9 Variation of lignin concentrationsacrosstheearlywood/latewood boundary of black spruce measured by U V microscopy ( 0 ) and the EDXA technique ( 0 ) .(From Ref. 86.)

Saka

66

earlywood/latewood boundary of black spruce (Piceu r~zuriur~a Mill.). It is quite apparent that the agreement between the results obtained by the two methods is good. With this EDXA technique, another method has also been developed by Westennark et al. [87] and Eriksson et al. [SS], based on a mercurization of lignin, followed by determination of mercury concentration in different morphological regions of wood.

B.

Distribution of Lignin in Softwoods

Table 9 shows the distribution of lignin in tracheids of black spruce (Picea marianu Mill.) as determined by UV microscopy [50]. The results show that the lignin concentration in the secondary wall (S) is considerably lower than that in the middle lamella (ML or ML,,). However, the secondary wall makes up a muchlarger proportion of the total tissue volume. Thus, the majority of the lignin is located in the secondary wall. Furthermore, the lignin is uniformly distributed across the secondary wall in black spruce tracheids, as seen in Fig. IO. For more detailed information, the distribution of lignin in the xylem of Douglas fir [Pseudotsuga tnenziesii (Mirb.) Franco] is given in Table 10 [74]. It should be noted that the distribution of lignin in the various morphological regions of the tracheids is basically the same as that shown for black spruce in Table 9. For the ray parenchyma secondary wall, the lignin concentration is higher than that for the tracheid secondary wall but lower than that for the middle lamella. However, the secondary wall of the tracheids does not differ much from that of ray tracheids in its lignin Concentration. Table I I shows the distribution of lignin in loblolly pine (Pinus ruedcl L.) tracheids as determined by bromination coupled with SEM-EDXA [85]. One of the advantages of this techniquecomparedwith UV microscopy is the ability to study the S,, S,, and S, layers in the secondary wall as a separate entity. Such resolution is often difficult with UV microscopy. It is interesting to note that the lignin concentration in the S? layer is lower than that in either the S , or S, layer. The line profile of the bromine X-rays in Fig. 1 1 showssuch differences clearly. FukazawaandImagawa[89]havealsoreporteda similar finding of high UV absorbance near the lumen/wall interface for juvenile wood tracheids of Japanese fir (Abies suchalinensis Fr. Schm.). A comparison of Tables 9-1 1 shows that, minor differences not withstanding, the trends in the distribution of lignin in the tracheids of the three softwoods are similar. For the ray parenchymacellsconstitutingabout 5% of the total xylem tissue in softwoods, Harada and Wardrop[ 151 have reported a lignin content of 0.44 g/gin Japanese

TABLE 9 The Distribution of Lignin in Black Spruce Tracheids UV Microscopy

as

Determined by

~~

Lignin

Tissue Wood

Morphological volume region

(%)

(g/g) conc. (% of total)

~~~

Earlywood Latewood

S

87

72

0.23

ML ML,,

9 4 94 4 2

16

0.50 0.85

S

ML ML,, Source:

Ref. SO.

12 82 IO 8

0.22 0.60 1 .oo

67

Composition Chemical

FIGURE 10 UVphotomicrographtakenat 240 nm of theearlywoodtracheidwallsinblack spruce. The densitometer tracing was conducted along the dotted line. (Courtesy of Prof. Emer. D. A. 1. Goring, University of Toronto, Toronto, Canada.)

[66]obtained cedar (CryptorneriujuponicuD. Don). By a microdisection technique, Bailey a value of 0.41 g/g for the segregated ray parenchyma cells of Douglas fir. Fergus et al. [50] also determined byUV microscopy a lignin concentration of 0.40 g/g forblack spruce. These results by a variety of methods are in good agreement with the data shownin Table 10 for Douglas fir earlywood parenchyma cells. Interestingly, the ray parenchyma cellsin softwoods possess significantly higher lignin contents than the whole wood.

TABLE 10 The Distribution of Lignin in Douglas Fir Xylem as Determined by W Microscopy

Lignin

Tracheid

racheid

Tissue Wood

Earlywood ray ray Latewood ray ray Source: Ref. 74.

Morphological volume region S Tracheid ML Tracheid ML, Paren. S Tracheid S S

Tracheid ML Tracheid ML,, Paren. S Tracheid S

(%) (g/g)

conc.(% total)

74 10 4 8 4

58

0.25

18

0.56 0.83

90

78 10

0.40 0.28 0.23 0.6

6

0.9

4

-

4 2 3 1

11

10 3

2

Saka

60

TABLE 11 The Distribution of Lignin In Loblolly Pine Tracheids as Determined by Bromination with SEM-EDXA

Lignin

Tissue

wood Earlywood

Morphological volume region SI S2

S, ML

ML, Latewood

SI S2 S,

ML ML,

. (glg)

(%) conc total)(70of

13 60

9 12 6 6 80

5 6 3

12

0.25

44

0.20

9

0.28

21

0.49 0.64 0.23 0.18

14 6 63 6 14

0.25

0.5 1

11

0.78

Source: Ref. 85.

Regarding the distribution of lignin in the compression wood of softwoods, Timell [ 1l] has given an excellent review. As observed in an electron micrograph shown in Fig. 5 of the lignin skeleton from the compression wood of tamarack [Lark luricina (Du Roi) K. Koch], the S, layer appears to have a slightly lower lignin concentration than the inner S,. However,aringpresent in the S , layer [&(L)] reveals a high lignin concentration about equal to that in the middle lamella. Table 12 shows acomparisonmadebyTimell[l13 of theligninconcentrations determined by Wood and Goring 1741 of Douglas fir and by Fukazawa [90] of Japanese fir (Abies sachalinensis Fr. Schm.). Although the lignin content in Japanese fir is lower in most of the morphological regions, the overall trends are basically the same.

C

FIGURE 11 Scanning electron micrograph (a) of brominated latewood trachids in loblolly pine (0.5-pm section). The distribution map (b) of Br-L X-rays was taken of the same area as the scanning electron micrograph. The distribution of bromine (c) was taken along the line across the double cell wall. (From Ref. 83.)

69

Chemical Composition and Distribution TABLE 12 TheDistribution of LignininCompression Wood Tracheids of Douglas Fir and Japanese Fir

Morphological Douglasregion

Lignin concentration (%)

49

29 42 26 49

75

65

40 54 36

"From Ref. 14. hFrom Ref. 90.

C.

Distribution of Lignin in Hardwoods

Hardwood lignins consist mainly of guaiacyl and syringyl residues, and its ratio seems to change from one morphological region to another. Fergus and Goring [72,73] attempted to determine the distribution of lignin in white birch (Betula papyriferem Marsh.) byUV spectral analysis. The syringyl and guaiacyl residues have, however, markedly different UV absorptivities. Thus, it is essential to know its exact ratio before the lignin concentration in a particular morphologicalregion is computedfrom the UV microscopy. In the 1980s Saka et al. 191,921 developed a new method to compute the ratio of guaiacyl and syringyl residues at the various morphological regions by combining UV microscopy with bromination-EDXA (UV-EDXA). This could be used to determine lignin distribution in hardwoods. Shown in Table 13 is the ratio of guaiacylkyringyl residues in various morphological regions of white birch wood as determined by the UV-EDXA technique 1911. For com-

TABLE 13 Distribution of GuaiacylandSyringylResidues White Birch

in Lignin i n

Guaiacy1:syringyl

DXA Morphological omination" with region

uv analysish spectral Syringyl Guaiacyl Syringyl 5

0

Guaiacyl 50:50 5050

"From Ref. 9 1. 'From Ref. 73. 'Fiber/fiber. dFiber/vesscl. 'Fibedray. 'Ray/ray.

70

Saka

parison, the results obtained by Fergus and Goring [73] through UV spectral analysis are also included. It is indicated by both methods that the fiber secondary wall (S2) contains predominantly syringyl residues, whereas the vessel secondary wall (S?) consists mostly of guaiacyl residues. The study by UV-EDXA [91] revealed that the ray parenchyma cell contains about equalproportions of guaiacylandsyringylresidues in lignin. However,apredominant amount of syringyl-type lignin was found by UV spectral analysis [73], as in the fiber secondary wall. For the cell corner middle lamella (ML,,), 80-100% of the lignin was found to be guaiacyl residue, with the remaining 0-20% being syringyl residue by the UV-EDXA technique [91]. This result is not in agreement with the data by UV spectral analysis [73]. However, it supports the later suggestion of Musha and Goring [75]that the middle lamella lignin consists entirely of guaiacyl residues. Itis therefore apparent in Table 13 that, in hardwoods, the ratio of guaiacyland syringyl residues in lignin varies in different morphological regions. These findings have been supported by several investigators; Wolter et al. [93] have shown that the vessels in aspen callus cultures contain a pure guaiacyl lignin. Kirk et al. [94] found that the fungal degradation of lignin in birch wood was consistent withthe presence of syringyl-rich lignin in the fiber walls. Furthermore,Yamasakiet al. [95] isolated syringyl-rich lignin from several hardwoods. Hardell et al. [96] fractionated birch wood to determine the syringyl and guaiacyl ratio, and indicated that lignins in both the middle lamella and vessel secondary wall are rich in guaiacyl units, whereas the ratio of syringyl/guaiacyl residues is high in the fiber and ray cell. Cho et al. [97] studied the filmlike substance isolated from the fines of birch in which a high proportion of the compound middle lamella was recognized.This material wasfound to possessalow ratio of syringyl to guaiacyl units. Terashima et al. [98] administered 'H-labeled guaiacyl and syringyl model compounds to magnolia shoots and determined their location in the growing cell wall by microautoradiography. They found that the vessel wall, cell comer, and compound middle lamella were lignified by the deposition of guaiacyl-type lignin, and the fiber wall was composed of syringyl-guaiacyl lignin. Recently, UV and visible-light microscopic spectrophotometry have been combined with the Maule color reaction for detecting syringyl lignin by Yoshinaga [99], and this method has been extended to taxonomic studies of the distribution of hardwood lignins [ 100- 1021. Table 14 shows the distribution of lignin in white birch wood as determined by UVthe results obtainedearlier by UV microscopy 1721 are EDXA [91].Forcomparison, included. For the fiber secondary wall, the lignin concentration in the S, layer is slightly lower than in either the S , or S2 layer. However, its difference is so small that the lignin may be considered to be distributed uniformly across the secondary wall. The vessel walls also reveal a uniform distribution of lignin, but the concentration is about 1.9 times higher than that of the fiber walls, which in turn is higher than that of ray parenchyma cells. The cell comer middle lamella (ML,,) associated with fibers and vessels has the highest lignin concentration. In spite of sufficient analytical resolution by the EDXA system, the middle lamella between cell corner areas (ML) was 10-30% lower in concentration than the cell corner middle lamella (ML,,). It is of interest to note that the lignin concentration in the middle lamella regions of hardwoods is lower than that of softwoods, as seen in Tables 9 and 14. A comparison of the data made between UV-EDXA and UV microscopy techniques indicates that lignin concentrations in fiber and vessel secondary walls are in agreement

Composition Chemical

71

TABLE 14 The Distribution of Lignin in WhiteBirch

Element

Lignin concentration (g/g)

TissueMorphological volume region

(%)

11.4 58.5 3.5 5.2 2.4 1.6 4.3 2.3 0.8 =O

8.0 2.0 =O =O

UV-EDXA

uv only”

0.14

-

0.14 0.12 0.36 0.45 0.26 0.26 0.27 0.40 0.58

0.16

0.12

0.22

0.38 0.47 0.41

-

0.34 0.72 -

0.22 -

0.35 -

“From Ref. 9 1. hCalculated using xylem lignin content of 0.199 g/g; from Ref. 72. “Fiberlfiber. ‘Fiber/vessel. ‘Fiberlray. ‘Raylray.

with each other. However, the lignin concentration in the ray parenchyma cells byUVEDXA[91] is nearly half aslow as the dataobtained by UV microscopyalone[72]. Although the middle lamella between two cell corners (ML) of fibers and vessels revealed similar values by these two techniques, the concentration in the cell corner middle lamella (ML,,) was lower by the UV-EDXA technique [91]. The observed discrepancies are due probably to the uncertainty in estimating the guaiacyl/syringyl ratio, as the analysis is made by UV microscopy alone.

VI.

DISTRIBUTION OF INORGANIC CONSTITUENTS

A fair amount of information is available on the inorganic constituents of wood [1031091 and bark [ 110,ll l]. In woods from temperate zones, elements other than carbon, hydrogen, oxygen, and nitrogen make up between 0. l % and 0.5% of the weight of wood [ 1 12,1131, whereas those from tropical regions make up to 5% [ 1 141. This proportion, although small, contains a wide variety of elements. For example, spectrographic analysis of grand fir [l031 revealed as many as 32 elements (Table 15). In many cases, alkali and alkali earth elements such as Ca, Mg, and K make up about 80% of the total inorganic constituents [l 151. These elements probably occur in wood as salts, e.g., oxalates, carbonates, and sulfates [ 1161, or inorganic moiety bound to the cell wall components such as carboxyl groups of pectic materials [ 115,117,1181. Some of the inorganicelementspresent in wood are essential forwoodgrowth, whereas others are not necessarily required. Metalic elements are often absorbed into the tree through the root system and are transported to all areas within the growing tree [ 1031.

Sa ka

72

TABLE 15 Classification, Function, and Approximate Level of Occurrence of Elements Found in Wood of Grand Fir (ppm of dry weight)

Essential Major

Constituent C

Ca

754

0 H N P S

K Mg Na

865

c1

Si

171 23 -

B Mn

Fe

0.9 19.3 2.6

Ag 0.23 AI 5.4 Ba 20.2

MO

0.005

CO

Cu Zn

2.5 0.9

Cr Ni Pb Rb Sr

0.0 1 0.05 0.1 1 0.12 2.0 10.2

Ti

0.11

Au

0.04

0.02 Ga In La

0.03

Li 0.003 Sn

0.13

0.04

v

0.001

Zr

0.002

Source: Ref. 103.

For seven species, Young and Guinn [ 1091 have determined the distribution of 12 inorganic elements in various tissue areas of a tree such as the roots, bark, wood, and leaves. The results indicated that both total ash content and concentration of each element vary significantly within and between the species. Therefore, unlike major cell wall components such as cellulose and lignin, the content of inorganic constituents varies to a great extent with the environmental conditions under which the tree has grown [ 105,1131. Little has been published regarding the morphological distribution of elements in the cell [ 1 19- 1221. By microincineration, Lange [ 1 191 found that mineral constituents of Swedish spruce are deposited predominantly in the compound middle lamella. Zicherman and Thomas [ 1201 also have pointed out that careful ashing of microtome sections of loblolly pine (Pinus r n e h L.), followed by electron microscopic observations, gives an ash residue distributed throughout the cell wall and concentrated in the compound middle lamella and S, layer. Wultsch [ 1211 stated that manganese is concentrated in ray cells, and Bergstrom [ 1221 reported that the phosphoruscontent is highest in the cambiumandadjacentxylem portions. Saka and Goring [ 1151 have studied the distribution of inorganic constituents from the pith to the outer ring of black spruce (Picea nznriann Mill.) by means of TEM-EDXA. The TEM-EDXA technique is a useful tool fordetectinganyelementaboveneonand recently above boron in the periodic table. Figure 12 shows seven morphological regions of the tracheids, ray tracheids, andrayparenchymacells investigated. Thedarkcircle indicates the location of the analysis and its diameter corresponds to the resolution of analysis (400 nm). Detected were 15 different elements, such as Na, Mg, AI, S, Cl, K, Ca, Cr, Fe, Ni, Cu, Zn, and Pb, above neon in the periodic table. The secondary walls of tracheids, ray tracheids, and ray parenchyma cells usually contain detectable concentrations of only four elements: sulfur, chlorine, potassium, and calcium. In contrast, almost all the elements were found to be localized and concentrated in the torus and half-bordered pit membrane regions (Fig. 13). The total content of inorganic constituents decreased in the order of torus (2%) > half-bordered pit membrane (1%) > middle lamella (0.4%) > ray parenchyma cell wall (0.3%) > tracheid secondary wall (0.1-0.15%). The total content of inorganic constituents was higherin earlywood than latewoodfor any of the morphological

omposition Chemical

and Distribution

73

FIGURE 12 Transmission electron micrographs of a cross section of black spruce showing the seven different morphological regions. All micrographs were takenat the same magnification. S , = secondary wall of the tracheid CC = cell comer middle lamella surrounded by tracheids TT = tours in an intertracheid pit pair SR = secondary wall of the ray parenchyma cell M = a half-bordered pit membrane between ray parenchyma cell and tracheid SRT = secondary wall of the ray tracheid TRT= torus in an intertracheid pit pair between ray tracheid and tracheid

74

Saka

FIGURE 13 EDXA spectra from the tracheid secondary wall and tracheid torus in black spruce. (From Ref. 115.)

regions studied. This is probably because the earlywood tracheids that have large lumens and abundant pits are the major water-conducting tissues, whereas thick-walled latewood tracheids with fewer pits may act as a physical or mechanical support for the wood. Bailey and Reeve [l231 have recently used imaging microprobe secondary ion mass spectrometry (SIMS) to determine the distribution of the trace elements in black spruce (Picea rnariana Mill.). This imaging microprobe S N S technique is a powerful tool for detecting inorganic elements with high spatial resolution and high sensitivity. Their overall findings correlatewell with results from the TEM-EDXA studyby Saka and Goring[ 1151. However, due to its higher sensitivity compared with the EDXA technique, the distribution of the elements within the cell wall could be more clearly demonstrated. Figure 14 is one example in which some elements are visualized and concentrated in the middle lamella region. Recently, Saka and Mimori [l241 have studied the distribution of inorganic constituents of Japanese birch wood (Betula platyphylla Sukatchev var. Japonika Hara) by the SEM-EDXA technique with thin sections. Figure 15 shows six morphological regions of the fibers, vessels, and ray parenchyma cells investigated. The dark circle corresponds to the resolution of analysis (800 nm). Detected were 11 different elements: Na, Mg, Al, Si, P, S, Cl, K, Ca, Fe, and Zn. The secondary walls of wood fibers, vessels, and ray parenchyma cells usually contained detectable concentrations of three elements, S, Cl, and Ca, while, in the amorphous layer of ray parenchyma cell and pit membrane between vessel and ray parenchyma cell, almost all of the detected elements were found to be localized and concentrated (Fig. 16). The total content of inorganic constituents decreased in the order amorphous layer (0.68%) > fiber middle lamella (0.54%) > vessel middle lamella

Composition Chemical

75

FIGURE 14 Ion image and its intensity for Ca. Fe, and Mn from a tangential section of a double cell wall of black spruce heartwood. (Courtesy of Prof. D. W. Reeve, Universityof Toronto, Toronto, Ontario, Canada.)

> fibersecondary wall (0.14%) > vessel (0.48%) > ray parenchema cell wall (0.15%) secondary wall (0.10%).This observed trend is basically the same as found in black spruce by Saka and Goring [115]. VII.

CELL WALL ORGANIZATION

In the previous sections, current knowledge of the distribution of cell wall constituents was described. In this section, therefore, how these constituents construct and organize the cell wall structure is discussed. In wood cell walls, cellulose acts as the structural framework in the formof cellulose microfibrils, while hemicellulose is the matrix substance present between these microfibrils. Lignin, on the other hand, is the encrusting substance binding the wood cells together and giving rigidity to the cell wall. Generally, the S2 layer increases with increasing wall thickness, whereas the S , and S , remain fairly constant. Because of its greater thickness, the S2 layer is largely responsible for the physical and mechanical properties of the cell walls. Figure 17 shows the relationship for softwoods between the lignin content and microfibrillar angle (e) in the tracheid S2 layer determined by the X-ray diffraction method. Since the majority of the lignin in softwoods is in the tracheid S , layer [50], the whole lignin content of wood must be closely correlated to the lignin concentration in the S , layer of the tracheid. Thus, from Fig. 17, the lignin concentration in theS , layer increases

76

Saka

c

FIGURE 15 Scanning electron micrographs of a transverse section of the Japanese birch wood showing the six different morphological regions considered in this study. Fs = secondarywall of thewoodfiber F,, = cell comer middle lamella surrounded by wood fibers Vs = secondarywall of thevessel VML= cell comer middle lamella surrounded by vessel and wood fibers Rs = secondarywall of therayparenchymacell RA, = amorphous layer in the ray parenchyma cell

with increasing microfibrillar angleof the tracheid S2 layer [ 1251. The biosynthetic origin

of this relationship is not known. However, it does suggest that, in order to construct the enforced plywood type of structure shown in Fig. 1, the three major chemical constituents of wood mutually interact and strengthen each other to make up a natural supercomposite material. Figure 18 shows such an ultrastructural arrangement of cellulose microfibrils, hemicellulose, and lignin in wood cell walls as proposed by Harada and CBtC [126]; around the core of cellulose microfibrils, paracrystalline regions of cellulose are thought to exist, which are associated with hemicellulose and lignin. Lignin encases them and binds them into the rigid structure of the wood cell wall. At the molecular level of arrangement of the chemical composition, the presence of a chemical bond between lignin and carbohydrate has been proved to be a lignin-carbohydrate complex (LCC) [l271 which is considered to be a compatibilizer-like substance localized at the interface between hydrophobic macromolecules of lignin and hydrophilic carbohydrates, by enhancing the physical and mechanical properties of wood [128].

77

Chemical Composition and Distribution

I

.

D-

4

I

I

FIGURE 16 Scanning electron micrograph (a) of a cross section of Japanese birch. The arrow shows the location of the EDXA analysis at the pit membrane between vessel and ray parenchyma cells from which the EDXA spectrum (b) was obtained.

45

2

40

Y

c

c

2 35 S

c *g 30 3

25 20 0

10 5 020 4 0 3 0 Microfibrillar angle Cel

FIGURE 17 Relationshipbetweenthemicrofibrillarangle lignin content of wood. (From Ref. 125.)

(e) inthetracheid

S2 layerandthe

78

Saka

FIGURE 18 Schematic diagram of the ultrastructural arrangement of a cellulose microfibril (Mf), hemicellulose (H), and lignin (L) in the wood cell wall. (From Ref. 126.)

VIII.

CONCLUDINGREMARKS

Knowledge of the chemical composition of woodis essential for studying the physical and chemical properties of wood. However, it can provide nothing but the average of the cell wall constituents. For a better understanding of wood properties, more detailed information is required about their distribution across the wood cell wall. However, in spite of a variety of methods proposed, all the methods have drawbacks and thus some discrepancy exists among investigators. A good method for resolving such discrepancies would be to separatevarioustypes of tissues physicallywithoutintroducinganychemicalchanges [49,96]. Analysis of the separated tissues could then provide definitive information on the distribution of the cell wall constituents at the various morphological regions of wood.

REFERENCES 1.

2. 3.

4. 5.

6.

7. 8. 9. 10. 11.

12. 13. 14.

A.B.Wardropand D. E. Bland, in Biochemistry of Wood (K. Kratzland G. Billek, eds.), Pergamon Press, London, p. 92 (1959). H. Harada, Mokuzai Gakkaishi, 30:513 (1984). F.F.P. Kollmann and W.A. C M , Jr., Principles of Wood Science and Technology, Vol. I, Solid Wood, Springer-Verlag, Berlin, p. 43 (1968). B. L. Browning, The Chemistty of Wood, Wiley-Interscience, New York, p. 57 (1963). T. E. Timell, in Cellular Ultrastructure of Woody Plants (W. A. CBtC, Jr., ed.), Syracuse Univ. Press, New York, p. 127 (1965). H. Meier, in Biosynthesis and Biodegradation of Wood Components (T. Higuchi, ed.), Academic Press, New York, p. 43 (1985). F.F.P. Kollmann and W. A. CBtC, Jr., Principles of Wood Science and Technology, Vol. I, Solid Wood, Springer-Verlag, Berlin, p. 55 (1968). W. A. C M , Jr., in Recent Advances in Phytochemistry, Vol. 11 (F. A.Loewusand V. C. Runeckles, eds.), Plenum Press, New York, p. 1 (1977). T. E. Timell, Wood Sci. Technol., 1:45 (1967). H. Higuchi, KASEAA, 13:206 (1975). T. E. Timell, Wood Sci. Technol., 16233 (1982). G. Schwerin, Holdorsch., 12:43 (1958). M. Fujii, J. Azuma, F. Tanaka, A. Kato, and T. Koshijima, Wood Res., 68:8 (1982). G. C. Hoffmann and T. E. Timell, Tappi, 55:733 (1972).

Chemical Composition and Distribution

79

15. H. Harada and A.B.Wardrop, MokuzaiGakknishi, 6:34 (1960). 16. 0. Perila, J. Pol.ymer: Sci., 51:19(1961). 17. 0. Perila and P. Heitto, Suomen Kemistilehti, B32:76 (1959). 18. G. C. Hoffmann and T. E. Timell, Tappi, 55:871 (1972). 19. H.Meier, Pure Appl. Chem., 5:37 (1962). 20. P. Whiting and D. A. I. Goring, Post-Grad. Res. Lab. Rep. 241, PPRICAN, Quebec, Canada (1981). 21. W. A. C8t6, Jr., N. P. Kutscha, B. W. Simson, and T. E. Timell, Tnppi, 51:33 (1968). 22. W. A. C M , Jr., A. C. Day, N. P. Kutscha, and T. E. Timell, Holz&rsch., 21: 180 (1967). 23. R. B. Hanna and W. A. C8t6, Jr., Cytobiologie, 1 0 1 0 2 (1974). 24. T. Goto, Ph.D. thesis, Dept. Wood Science Technology, Kyoto Univ., Kyoto, Japan (1976). 25. A. N. J. Heyn, Tippi, 60:l59 (1977). 26. G. Cox and B. Juniper, J. Microscopy, 97343 (1973). 27. J. B. Boutelje and B. H. Hollmark, Hol$orsch., 26:76 (1972). 28. V. M. Sinner, N. Parameswaran, H. H. Dietrichs, and W. Liese, Hol$orsch., 2 7 3 6 (1973). 29. M. N. Sinner, N. Parameswaran, N. Yamazaki, W. Liese, and H. H. Dietrichs, Appl. Polymer: Symp., 28:993 (1 976). 30. P. Hoffmannand N. Parameswarm, Holdorsch., 30:62 (1976). 3 1. T. Awano, K. Takabe, and M. Fujita, Abstr: 46th Annual Meeting of the Japun Wood Resenrch Society, p. 32 (1996). 32. T. Fujii, H. Harada, and H. Saiki, MokuzniGakknishi, 27:149 (1981). 33. W. A. C M , Jr., A. C. Day, and T. E. Timell, Wood Sci. Technol., 2 : I3 (1968). 34. N. Parameswaran and W. Liese, Holz.Roh.-Werkst.. 40:145 (1982). 35. P. Albersheim, K. Muhlethaler,and A. Frey-Wyssling. J. Biophys. Biochem. Cyrol., 8501 ( 1 960). 36. R. M. McCready and R. M. Reeve, Agric. Food Chem., 3:260 (1955). 37. M. Gee, R. M. Reeve, and R. M. McCready, Agric. Food Chem., 7 3 4 (1959). 38. E. M. Barmicheva and M. F. Danilova. Bo?.Zh., %':l278 (1973). 39. P. M. Colombo and N. Rascio, J . Ultrustruct. Res., 60:135 (1977). 40. Y. Czaninski, Biol. Cellulnire, 35:97(1979). . R389 (1936). 41.A. J. Bailey, lnd. Eng. Chem., A I Z U ~Ed., 42. H. Meier, J. Polymrc Sci.. 5 / :1 1 (1961 ). 43. H. Meierand K. C. B.Wilkie, Hol
80

Saka

61. H. Meier, Holz Roh-Werkst., 13:323 (1955). 62. I. B. Sachs, 1. T. Clark, and J. C. Pew, J . Polymer. Sei. Part C, 2:203 (1963). 63. A. L. K. Bentum, W. A. CGtC, Jr., A. C. Day, and T. E. Timell, Wood Sei. Technol., 3:218 ( 1969). 64. R. A. Parham and W. A. CGtC, Jr., Wood Sei. Technol., 5:49 (1971). 65. P.W. Lange, Svensk Papperstidn., 57525 (1954). 66. A. J. Bailey, Ind. Eng. Chem., Anal. Ed., 8 5 2 (1936). 67. G. J. Ritter, Ind. Eng. Chem., 17:1 194 (1925). 68. J . A. N. Scott, A. R. Procter, B. J. Fergus, and D. A. I. Goring, WoodSei.Technol., 3:73 ( 1969). 69. J. A. N. Scott and D. A. I. Goring, Wood Sei. Technol., 4:237 (1970). 70. J. R. Wood and D. A. 1. Goring, J. Microsc., 100:105 ( 1 974). 71. J . A. N.Scott and D. A. I. Goring, Cell. Chenz. Technol., 4:83 (1970). 72. B. J. Fergus and D. A. 1. Goring, Hollforsch., 24: I 18 ( 1970). 73. B. J. Fergus and D. A. I. Goring, Hollforsch., 24: 1 13 (1970). 74. J. R. Wood and D. A. I. Goring, Pulp Paper Mag. Can.,72:T95 ( 1971 ). 15. Y. Musha and D. A. 1. Goring, Wood Sci. Technol.. 9 4 5 (1975). 76. G. P. Berlyn and R. E. Mark, Forest Prod. J., 16:140 (1965). 77. J. M. Yang and D. A. I. Goring, Pulp Paper Can. Trans., 4:2 (1978). 78. J. M. Yang and D. A. I. Goring, Can. J. Chem., 58:2411 (1980). 79. P. Whiting and D. A. I. Goring, Paperi j a Puu, 10592 (1982). 80. P.W. Lange and A. Kjaer, Norsk Skogind, 11:425 (1957). 81. J. B. Boutelje, Svensk Papperstidn., 75:683 (1972). 82. S. Saka, R. J. Thomas, and J. S. Gratzl, Tcppi, 6l:73 (1978). 83. S. Saka, R. J. Thomas, and J. S. Gratzl, Proc. ISWPC, Stockholm, Sweden, Vol. I , SPCI Rep. 38, p. 35 (1981). 84. S. Saka and R. J. Thomas, Wood Sei. Technol. 16:1 ( I 982). 85. S. Saka and R. J. Thomas, Wood Sci. Technol., 16:167 (1982). 86. S. Saka, P. Whiting, K. Fukazawa, and D. A. I. Goring, Wood Sci. Technol., 16:269 (1982). 87. U. Westermark, 0. Lidbrandt, and 1. Eriksson, Wood Sci. Technol.. 22:243 (1988). 88. I. Eriksson, 0. Lidbrandt, and U. Westermark, Wood Sei. Echnol., 22:25 1 ( 1988). 89. K. Fukazawa and H.Imagawa, Wood Sei. Technol., 15:45 ( 1981). 90. K. Fukazawa, Res. Bull. Coil. Exp. Forests Hokkaido Univ., 31:87 (1974). 91. S. Saka and D. A. I. Goring, Hollforsch., 42: 149 ( 1988). 92. S, Saka, S. Hosoya, F. G. T. St-Germain, and D. A. I. Goring, Holiforsch., 42:79 ( 1988). 93. K. E. Wolter, J. M. Harkin, and T. K. Kirk, Physiol. Plartt, 31: 140 (1974). 94. T. K. Kirk, H.-m. Chang, and L. F. Lorenz. Wood Sei. Technol., 9 8 1 (1975). 95. T. Yamasaki, K . Hata, and T. Higuchi, Hollforsch.. 32:44 (1978). 96. H.-L. Hardell, G . J. Leary, M. Stoll, and U. Westermark, Svensk PqJperstidn.,83:71 (1980). 97. N. S. Cho, J. Y. Lee, G. Meshitsuka, and J. Nakano. Mokuwi Gakkcrishi, 26:527 (1980). 98. N. Terashima, K. Fukushima. and K.Takabe, Hol;for.sch., .CO(Suppl.):lOl (1986). 99. A. Yoshinaga. Ph.D. thesis, Dept. Wood Science Technology, Kyoto Univ., Kyoto,Japan ( 1995). 100. K. Takabe, S. Miyauchi. R. Tsunoda, and K. Fukazawa. IAWA Bull., I Z . ~ . 1, 3 : 105 ( 1992). 1 0 1 . J. Wu. K. Fukazawa. and J. Ohotani, Hol&rseh., 46:181 (1992). 102. Y. Watanabe and K. Fukazawa, Res. Bull. Hokknido Uni1: Forests, 50:349 ( 1 993). 103. E. L. Ellis, in Cellulrr Ultrastructure cf Woo& Pltrrrts (W. A. CGtC, Jr., ed.), Syracuse Univ. Press, New York, p. 18 1 ( 1965). 104. R. F. Dyer, Tech. Bull. No. 27, Univ. Maine, Maine Agric. Exp. Sta. (1967). 105. E. L. Ellis, Forest Prod. J., 12:271 (1962). 106. W. L. Galligan, H. Stern, and P. Hohenschuh, Forest Prod. J., 15:185 (1965). 107. H. E. Young and P. W. Carpenter, Tech. Bull. No. 28, Univ. Maine. Maine Agric. Exp. Sta. ( 1967).

omposition Chemical 108.

109. 1 IO. 111. 112.

113.

114. 1 15.

116. 117. 118.

119. 120.

121. 122. 123.

124. 125.

126. 127. 128.

81

H, E. Young, P. N. Carpenter. and R. A. Altenberger. Tech. Bull. No. 20, Univ. Maine, Maine Agric. Exp. Sta. (1965). H. E. Young and V. P. Guinn, Rrppi, 49:190 (1966). M. L. Harder and D. W. Einspahr, Tappi. 63:110 (1980). H . E. Young. Forest Prod. J., 2 / 5 6 (1971). A. J . Panshin and C. de Zceuw, Te.srbook of Wood Techrwlogy, 3rd ed., Vol. l , McGraw-Hill, New York, p. 73 (1970). B. L. Browning. The Chenlisfry of Wood, Wiley-Interscience, New York, p. 355 (1963). J. Savard, J. Nicolle, and A. M. Andre, Analyse chimique des bois tropicaux, Centre Technique Forestier Tropical. Nogent-sur-Marne (1960). S. Saka and D. A. I. Goring, Mokuzcri Gakkcrishi, 29:648 (1983). E. T. Choong, G. Abdullah, and J. Kowalczuk, LSU Wood Uti1i:utiorl Notes No. 24, (1976). B. E. Cutter, E. A. McGinnes, Jr.. and D. H. McKown, Wood Fiber, 12:72 (1980). H. Wazny and J. Wazny. Hol: Rol~-Werkst.,22:299 (1964). P. W. Lange. Pulp. Ptrper Mcrg. Carl., 59:2 10 (1958). J. B. Zicherman and R . J. Thomas, firppi, 54:1727 (1971). F. Wultsch, P ~ p e rFc11x. 4: 128 ( 1944). H. Bergstriim, S ~ w z s kPcrpper.stidn.. 62:160 (1959). J. H. E. Bailey and D. W. Reeve, J . Pulp Prrper Sci., 20:J83 (1994). S. Saka and R. Mimori, M o k w a i Gakknishi, 40:88 (1994). S. Saka and M. Tsuji, Cellulose Clwm Teclmol.. 2/:225 (1987). H. Harada and W. A. C6t6, Jr.. B i o s p t h e s i s and Biodegradation of Wood Co,nponer~f.s, p. 20, Academic Press, New York (1985). T. Koshijima, Wood Res. Tech. Nores, 19:1 1 (1984). S. Takase, N. Shiraishi. and M. Takahama, Wood Processing and Utilizcrtion, p. 243, Wiley, Chichester. U.K. (1989).

This Page Intentionally Left Blank

Structure of Cellulose: Recent Developments in Its Characterization Furnitaka Horii Kyoto University, Kyoto, Japan

1.

INTRODUCTION

The first Laue X-ray diffraction photographs of native cellulose, such as ramie and bamboo, were taken by Nishikawa and Ono in 1913 [l]. Since then a lot of effort has been madetocharacterize the crystal structure of nativecellulose,mainlyusing the X-ray diffraction technique. Nevertheless, the detailed structure has not been clarified as yet, and the reason for the difficulty has only recently been understood reasonably. Although it was assumed in the traditional analyses that native cellulose was composed of single homogeneous crystals, this assumption has been found not to be fulfilled in most specimens from different native sources: native cellulose crystals have been confirmed to be composites of at least two types of crystal allomorphs. This chapter deals with the process of theconfirmation of such a new crystal structuremodelfornativecellulosesomewhat historically. Further recent developments relating to this model of crystallization are also described by focusing attention on cellulose biogenesis by a bacterium. For other progress in characterization of cellulosic materials, refer to recent review articles [2-41.

II. COMPOSITE CRYSTAL MODEL Since the first observation of CP/MAS I3C NMR spectra of native celluloses in 1980 [5,6], it has been recognized that there are great differences in multiplicities of the Cl and C4 resonance lines between two groups, the bacterial-Vrtlorziu group and the cotton-ramie group. Figure 1 shows CP/MAS I3C NMR spectra of the crystalline components of these native celluloses [7,8], which were selectively measured by using longer "C spin-lattice relaxation times (TIC)of the components[9,10]. Asis readily seen,cotton and ramie celluloses give almost the same crystalline spectrum, whereas bacterial and klonia celluloses produce a different type of crystalline spectrum. The most prominent features of these two types of crystalline spectra appear in the C1 resonances; a predominant doublet with a weak central singlet is observed for cotton and ramie celluloses, while a triplet composed of an enhanced singlet and a minor doublet appears for bacterial and Valonia celluloses. These fine splittings, including those in the C4 and C6 lines, should be due to 83

a4

Horii

120

IIIIIIII,,I,,,II,,,,1,,,~I,,,,l,,,,I,,,,I,,,,I,,,,I,,I,I,,,,I,,,,I, 110 10 0 90 80 70 60 50

P m

from TMS

FIGURE 1 CP/MAS "C NMR spectra of the crystalline components included in differcnt native celluloses: (a) cotton; (h) ramie; (c) bacterial; (d) Vcrlorzicc celluloses 171.

the existence of carbon nuclei in magnetically nonequivalent states. It is therefore suggested that there must be significant differences in conformations around the 1,4-glycosidic linkage and the C5-C6 bond, hydrogen bonding, or molecular packing between the two groups. Nevertheless, it should be noted that the relative intensities of the C 1 and C4 lines are not described i n terms of small whole numbers, as would be expected if they arose fro111 different sites within a single unit cell. Similar structural differences between these two groups were also suggested by mea-

Structure of Cellulose

a5

surements of IR spectra about 40 years ago [ 1 l]; the absorption bands due to stretching of OH and CH groups were markedly different between the bacterial-Vulonia group and the cotton-ramie group. A separate electron diffraction study [ 121 also proposed that the so-called eight-chain unit cell was appropriate to interpret the diffraction pattern observed for Vuloniu cellulose, whereas the diffraction spots for ramie and cotton celluloses were usually indexed in terms of the Meyer-Misch-type two-chain unit cell. In the 1970s more detailed structural analyses were also carried out by wide-angle X-ray diffractometry [ 13151. However, these important efforts did not lead in a straightforward way to the idea that native cellulose crystals are composed not of single pure crystals but of a mixture of crystals with different crystal forms. In 1984,AtallaandVanderHart [ 16,171 proposed by their careful elucidation of CPMAS l3C NMR spectra of different native celluloses that native cellulose crystals are composites of two allomorphs that are referred to as cellulose I, and cellulose I,. This proposal, which should be estimated as an epoch-making contribution to the enhancement ofnew effortstoward the structural characterization of nativecellulose,wasmade by showing that the individual resonance lines were obtained by appropriate linear combinations of the spectra of the regenerated cellulose I that contains almost the pure I, form and bacterial cellulose that is rich in the I, form. Figure 2 shows schematically the spectra of the C, and C, carbons thus obtained for celluloses 1, and I, [16-181. The I,, form has singlets for the C l and C6 carbons and a doublet for the C4 carbon, whereas the I, form has doublets for the C l , C4, and C6 carbons. The multiplicities in the I, and I, spectra are reasonably interpreted by the structural inequivalences in single unit cells, in contrast to those in the whole spectra as described in Fig. 1. However, the novel structural model thus proposed, which is hereafter referred to as the composite crystal model, was not readily established, because the relative intensities of the C l and C4 triplets could not be fully interpreted by this model for different native celluloses [7], suggestingthe existence of some exceptions depending onthe native sources of celluloses. For example, Fig. 3 shows the results of lineshape analysis of the C4 resonance lines for native celluloses with higher contents of cellulose I,, [ 191. The most upfield line, which would be described by a single Lorentzian curve according to the composite crystal model, are found in this case to be composed of two Lorentzians with different intensities for these cellulose samples. In such a situation we need additional experimental evidence to support the composite crystal model.

111.

CRYSTAL TRANSFORMATION FROM CELLULOSE I, TO CELLULOSE I,

There maybe someways to confirm the composite crystal modelexperimentally.For example, one way is to collect native celluloses having more widely different I,, and I, contents, particularly to find native cellulose with much higher I,, content or the pure I,, form. Since this may be awfully time-consuming work, we tried to find a route to induce the crystal transformationbetweencelluloses I,, and I,. The fourfollowingcaseswere found to induce the crystal transformation from cellulose I,, to I,: l . Annealing at hightemperatureswith saturated steam [20] 2 . Annealing at hightemperatures in alkaline aqueous solutions [21] 3 . Solid-stateregenerationfromcellulose triacetate I [22] 4. Solid-stateregenerationfromcellulose 111, [23]

Horii

c1

c4

cellulose I a I

I

I

I

I

I

I I

I I

I I

1.j1 Ill I

cellulose t

I I

I I

i

I

h! I

cellulose I g

The first treatment was developed originally from the steam explosion of wood, by carefully elucidating the experimental fact that CP/MAS "C NMR spectra of Japanese birch and cypress were greatly changed before and after steam explosion at 255°C [24]. However, drastic degradation of cellulose occurs with saturated steam at 255°C when the samples are not wrapped with glass fiber sheets. Such degradation Inay be induced by the effect of hydrogen ions that are produced by the much higher level of the dissociation of water at higher temperatures. Considering this situation, influences of the pH values of aqueous nledia were investigated at different temperatures. Finally, annealing was found

87

Structure of Cellulose

,

92

90

.

1

R8

.

,

X6

1

92

,

I

90

.

I

88

,

l

X6

i

,

92

I

,

I

.

I

90

86 ppm from TMS

FIGURE 3 Lineshape analyses of the C4 resonance lines for Valonia mncrophyscr (a), Cladophonr (b), and bacterial (c) celluloses [19].

to be most effectively carried out in a 0. I N NaOH aqueous solution at 260-280°C without significant degradation [21]. Figure 4 shows CP/MAS I3C NMR spectra of Valotzicc macrophysa cellulose and its annealed samples at 220-280°C for 30 min in the 0.1 N NaOH aqueous solution [ 171. The enhanced central line of the C l triplet of the intact sample, which is one of the features for Itr-rich samples, is greatly reduced in intensity with increasing annealing temperature, whereas the doublet at both sides is concomitantly increased in intensity. The C4 triplet is also transformed into a doublet as a result of the reduction of the most downfield line and the concomitant increase of the most upfield line. The C6 resonance line is also a triplet with minor separation, and almost the same change in lineshape as the C l triplet is induced by the annealing. Finally, the C I , C4, and C6 resonance lines are all doublets, which is in good accord with the spectrum proposed for cellulose I, as shown in Fig. 2 . All the C l and C4 resonance lines shown i n Fig. 4 were successfully interpreted in terms of the linear combination of the spectra for celluloses I,, and I, shown in Fig. 2 [ l 81. The mass fraction of cellulose I,, that was determined by the lineshape analysis of the C l and C4 resonances 181 is plotted against the annealing temperature in Fig. 5. Here, the results obtained by annealing under high pressure are also shown. I t is clearly seen that almost all cellulose I,, is transformed into cellulose I, by the annealing, because there is almost no change in the total degree of crystallinity in this treatment. In addition, the appearance of the microfibrils also undergoes no change, indicating almost perfect crystal transformation occurring in this system. Organic solvents such as ethylene glycol, ethyl alcohol. and monoglyme and helium gas are also effective in inducing the crystal transformation from cellulose I,, to l,,, but the efficiency is not so high compared to the cases of saturated steam and alkalineaqueous solution (251. In the cases of generation from cellulose triacetate I and cellulose 111, there were significant decreases i n crystallinity and apparent splaying of the microfibril structure (22,231. Electron diffraction analysis also confirmed the crystal transformation from cellulose l,, to I,. Figure 6 shows electron diffraction diagrams of delaminated Vrrlor~icrrt~crcrq~hyscr fragments before and after annealing in the 0.1 N NaOH aqueous solution [26].The diffraction diagram of the original sample. which is identical with the patterns previously published. has triclinic character, particularly on the third layer line. In Fig. 7 is shown the densitometer traces of the third-layer lines for the samples annealed at different temperatures. The reflections occurring at the left-hand side of the meridian have different intensities from those at the right-hand side for the original Krlorzitr cellulose. Such triclinic features are drastically changed with increasingannealingtemperature. Finally. the dif-

88

Horii

c1 C 2.3.5

l r

Y

100 60

80 ppm from T M S

FIGURE 4

CPA4AS I3C NMR spectraof k / o r l i < l macrophysa celluloseannealed at different temperatures in 0.1 N NaOHaqueoussolution:(a)original; (b) 220°C; ( c ) 240°C; (d) 260°C; (e) 280°C [ 181.

Structure of Cellulose 0.

1

1

1

89 1

1

I

I

1

I

t 1

0 Annealing temperature /'C

FIGURE5 The mass fraction of cellulose I, as a function of annealing temperature when annealed in 0.1 N NaOH aqueous solution: (A) under saturated steam pressure; (0)under 5 kbar.

fraction diagram becomes completely symmetric for the sample annealed at 260°C, indicating monoclinic character. It was also found that all spots shownin Fig. 6C are indexed with a two-chain P2, unit cell with a = 0.792 nm, b = 0.822 nm, c = 1.036 nm, and y = 97.3", whilethecorrespondingspotsshown in Fig. 6Aareindexedwithatwo-chain triclinic unit cell with a = 0.954 nm, b = 0.825 nm, c = 1.036 nm, a = 90°, p = 57.0", y = 96.6". It is therefore concluded that cellulosesI, and I, should be assigned to the twochaintriclinicandmonoclinicphases,respectively.However, the more stable form of cellulose I, can be assigned to the one-chain triclinic phase, as described later. Similar structural changes by annealing are also recognized by FT-IR spectroscopy. Figure 8 shows the FT-IR spectrum of Rhizoclonium cellulose before and after annealing in the 0.1 N NaOH aqueous solution [27]. Two absorption bands are clearly observed at 3240 and 3270 cm" in the OH stretching region and also at 750 and 710 cm" in the CH2 rocking region for the original sample, but the former bands almost disappear after annealing. This fact clearly indicates that the former bands are assignable to cellulose I,, whereas the latter bands can be ascribed to cellulose I,. Moreover, it may be suggested

FIGURE 6 Electron diffraction diagrams of Vuloniu cellulose before andafter its annealing: A, original; B, annealed at 240°C; C, annealed at 260°C [26].

90

Horii mer i d i a n

I

l

I

I

0

.2

I

.4

R FIGURE 7 Densitometer traces of the third layers of the electron diffraction diagrams for Valonia cellulose annealed at different temperatures: (a) control; (b) 220°C; (c) 240°C; (d) 260°C.

L

4000

3600 3200 2800 2400

wavenumber (cm-') FIGURE 8

1000

800

600

400

wavenumber (cm-')

FT-IR spectra o f original (A) and annealed (B) Khi:oc.lorliltr?! cellulose 127)

Structure of Cellulose

91

that the hydrogenbondingassociatedwith the CH,OH group is significantly different between the two allomorphs, in good accord with the finding by Raman spectroscopy [28]. Using the bands in the CH, rocking region, we developed a more convenient method to determine the mass fraction of cellulose I, or I, compared to the solid-state "C NMR method described above. In this case lineshape analyses of FT-IR spectra and CPMAS I3C NMR spectra were carried out at the same time for Valonia and bacterial celluloses annealed at different temperatures. Then the following equation was obtained between the mass fraction f F of cellulose I, determined by FT-IR and the mass fraction f t"" estimated by solid-state "C NMR [29]:

ft"" = 2.55f:

- 0.32

r"")

of cellulose Since this equation is a sort of calibration curve, the mass fraction f , (=f I, can be determined by FT-IR spectroscopy using a specimen of the order of milligrams.

IV.

DISTRIBUTION OF CELLULOSES I, AND I, IN NATURE

Figure 9 shows mass fractions of cellulose I, of representative native celluloses, which were determined by C P M A S I3C NMR or FT-IR spectroscopy. Marine algal and bacterial cellulosesarefound to be rich in cellulose I,; the average fraction is about0.63.For example, its value is 0.64 for Valonia rnacrophysa, 0.60 for Valonia aegurropilu, 0.67 for Chaetornorpha, and0.65 for Cladophora. In the case of bacterial cellulose, the mass fraction of cellulose I, depends on strains and culture temperature, ranging from 0.64 to 0.71. Careful purification with aqueous alkaline solution will reduce the content of cellulose I, by several percent for bacterial cellulose. On the other hand, cellulose-forming cell walls of higher plants such as cotton and ramie are rich in cellulose I,, the mass fraction being about 0.8. When these native celluloses, including algal and bacterial celluloses, undergo annealing at high temperatures, their I, fractions all increase up to about 0.9, but there is still some contribution from the I, phase. Since such a minor contribution cannot be detected by electron diffractometry, the size of the crystallites of the residual I, form must be significantly small. In nature, however,almostpurecrystals of the I, formcanbeobtainedfrom tunicate cellulose

valonia, bacterial

I-

-

L-

cotton, ramie annealed

tunicate

FIGURE 9 Distribution of the mass fraction of cellulose I,, in nature [ 191.

Horii

92

[21,30]. In contrast, it is still impossible to obtain the pure l,, form in nature and also by any artificial method at present. As for the characterization of woods, there is normally a difficulty in exact I3C NMR measurements of the crystalline component because of the low crystallinity and the coexistence of hemicelluloses and lignin. However, our recent CP/MAS "C NMR analysis has clearly revealed that normally lignified woodcellulose in Populus muxomowiczii, which belongs to hard woods, is cotton-ramie type, the mass fraction of cellulose I, being estimated to be about 0.8 1311. This result seems to be in conflict with our preliminary results for Japanese cypress and birch [24], in which cellulose I,, was assumed to be rather dominant. Recently, Newman measured separately the CP/MAS "C NMR spectrum of the cellulose component as a result of the removal of the contributions from hemicelluloses and lignin by using the difference in ' H spin-lattice relaxation time TlpHin the rotating frame for different hardwoods and softwoods [32]. Since the C1 and C6 resonance lines still seemed to contain unidentified contributions, the relative peak intensities of the most downfield and upfield lines of the C4 triplet were used to estimate the relative proportions of the I,, and I, forms. It was then concluded that the I,, fraction for softwoods was at almost the same level as for Vulonia-bacterialcelluloses,while the fractions for hardwoodsweresimilar to those for the cotton-ramie group. This conclusion may suggest that the crystallization of cellulose in woods will be affected in the presence of hemicelluloses and lignin, possibly resulting in the difference in fractions of I, and I, forms. A possible stress-induced crystallization of cellulose I,, which may be also induced in woods by the coexistence of hemicelluloses and lignin in the hybrid composites, will be described in the case of bacterial cellulose in a later section.

V.

CRYSTAL STRUCTURE OF CELLULOSES I, AND I,

As described above, the preliminary analysis of electron diffraction diagrams of celluloses I,, and I, were performed for delaminated fragments of Vulonia nzucrophysa. A more detailed microdiffraction analysis L331 was carried outalong single microfibrils with the widths of 25-35 nm that were separated from cell walls for Microdicryon, which is also a greenmarine alga. Two series of spot electron diffractogramswerealmostobtained independently, on different specimen areas with dimensions of about 50 nm, as shown in Figs. 10 and 1 1 . The diffraction spotsobserved in Fig. 10 are aligned in lines that are markedly inclined with respect to the cellulose chain axis (the long axisof the microfibril). I n all, 27independentreflections of this serieswereobtained, and all of themcanbe indexed using a one-chain P1 triclinic unit cell with cz = 0.674 nm, I? = 0.593 nm, c (chain axis) = 1.036 nm, a = 117", p = 1 I3', and y = 81'. This unit cell is found to be a singlechain version of the unit cell proposed by Sarko and Muggli [ 131. Another series of single-crystal diffraction diagramswereobserved for the same microfibrils, as shown in Fig. 1 I . In this case all the diffraction spotshaveorthogonal symmetry, indicating the monoclinic character. In fact, 38 independent diffraction spots observed in a l l as reflections of this series can be indexed in tcrms of a two-chain P2, unit cell with ( I = 0.801 nm, 0 = 0.817 n m , c (chain axis and unique monoclinic axis) = 1.036 nm. a = p = 90", and y = 97.3'. This unit cell is also found to be the same as unit cells [ 14,15,26,34-761 previously reported except for minor differences i n parameters ( I and y. The densities calculated for the triclinic and monoclinic unit cells are 1.582 and 1.599 g/cm', respectively. The significant increase in density for the monoclinic unit cell suggcsts the highcr thermodynamic stability of this unit cell compared to the triclinic unit

, -

8

. B

c

Ti4 0

@

O

110

O

FIGURE 10 A series of spot electron diffractograms presenting the triclinic features, which were obtained for the single microfibril of Microdictyon cellulose. (From Ref. 33, with permission of the American Chemical Society,Washington, DC.)

P 1

A

0 B

0

.

0

.

0

. . . 0

.

.

0

FIGURE 11 A series of spot electron diffractogramspresenting the monoclinic features, which were obtained for the single microfibril of Microdicryon cellulose. (From Ref. 33, with permission of the American Chemical Society, Washington, DC.)

I

-. P

Structure of Cellulose

95

cell, in good accord with the crystal transformation from cellulose I,, to I, as described above. Figure 12 shows the crystal structure models proposed for the one-chain triclinic and two-chain monoclinic phases on the basis of the microdiffraction analysis described above [33]. The monoclinic unit cell is the so-called Meyer-Misch-type cell, and the central chain in the cell is shifted downward by c/4 with respect to the comer chains. Moreover, this chain is rotated by 7.4” relative to the (200) plane, while there is no such rotation for the comer chains in the corresponding planes [33]. In contrast, when the central chain and the corner chains at the right side are shifted upward respectively by cl4 and 2d4 with respect to the corner chains at theleft side, one type of two-chaintriclinic unit cell can be obtained. Moreover, when the central chain is identical with the comer chains without any rotation with respect to the (200) plane, then another type of one-chain triclinic unit cell can be defined. This is the case for another allomorph of cellulose, cellulose I,, as shown in Fig. 12. As for the relative position of neighboring chains in the (1-10) plane of the one-chain triclinic unit cell, there are two possibilities, “parallel up” and“paralleldown.”Here, “up” or “down” indicates that the z coordinate of 0 5 is larger or smaller than the coordinate of C5, respectively. A molecular dynamic simulation in the crystalline environment was carried out by using the structure models of celluloses I, and I, shown in Fig. 12 as starting structures [37]. The program used was GROMOS 87 equipped with the appropriate force field. The triclinic phase was simulated under the periodic boundary condition for 4 X 6 X 3 (a X b X c ) unit cells with 2016 atoms, while the monoclinic phase was simulated for 3 X 3 X 3 unit cells with 15 12 atoms. It was found that the I, phase is energetically lower by 8.7 kJ mol-’ cellobiose-’ than the I, phase, in agreement with the almost complete crystal transformation from cellulose I,, to I, at higher temperatures [20,21]. Such higher stability in the I, form is due to intraplanar electrostatic interactions, particularly in the (200) plane. Moreover, the rotation of the chain in the (200) plane increases from 7.4” to I l S ” , resulting

Monoclinic Two-Chain Triclinic One-Chain

FIGURE 12 Crystal structure models of celluloses I, and I, assignable to the one-chain triclinic and two chain monoclinic crystals, respectively [33].

Horii

96

in significant changes in radial distribution functions and hydrogen bonding patterns, in addition to the reduction in energies. Structural features found for the triclinic and monoclinic phases were also qualitatively related to their spectroscopic features in FT-IR and CP/MAS "C NMR spectra. Similar molecular dynamics simulations were also performed using the CHARMM molecular modeling program, with the PARM 20 parameter set to characterize the differences in molecularmobilityandhydrogen-bondformation-breakage for the two allomorphs [38]. As a result of the simulation during 75 PS at 20°C after reaching the equilibrium state, it was found that the torsion angles 4 and IC, fluctuate with time on the order of 230" around the initial values for the I,, phase, whereas such fluctuations are only of the order of 10" for the I,. The time fluctuation of the chain center along the c axis was also considerably higher for the I, form than for the I, form. On the basis of these results, a break-slip model was proposed for the crystal transformation from cellulose I, to I,. In this model, it is suggested that the transformation is initiated by heat-induced torsional rotations of the CH,OH and OH groups accompanied by hydrogen-bond breakage. Cellulose chains are then subjected to rotation around the molecular chain axis and sliding along the axis, resulting in conversion to the I, form. For further reliable elucidation of the crystal structureforthesecelluloseallomorphs,however,parameters used in these simulations should be improved by comparison with experimental results.

VI.

CRYSTALLIZATION PROCESS OF NATIVE CELLULOSE

Even after the establishment of the composite crystal model for native cellulose, there still remains a question to be answered: how are celluloses I,, and I, crystallized in nature? The recent electron microdiffraction analysis described above also revealed that the I,, and I, phases are alternatively locatedwitha periodicity of about 100 nm along the single microfibril for Microdictyorz [33]. In the case of bacterial cellulose, such a periodic structure cannot be observed and almost one series of diffraction spots assignable to the I,, phase is obtained along the single microfibril [39,40]. Since the mass fraction of cellulose I, is 0.37, this phase must be located in the thin central core area along the microfibril, as will be described later. These facts suggest that the mode of distribution of the two allomorphs in single microfibrils may change from sample to sample and thus the crystallization of the allomorphs also depends on the conditions for the production of different cellulose samples.When the hemicellulosesand ligin coexistduring the crystallization process, effects of these materials should be well evaluated (411. In this situation, it will be very important to investigate the crystallization process in each system producing probably different composite crystals in nature. Here the case of the bacterial cellulose system is described in some detail, because this system has been considerablyinvestigated hitherto 142-441.

A.

Formation of the Normal Ribbon Assembly

It is well known that a gel-like pellicle of cellulose with a high water content is produced on the surface ofan incubation medium when a Gram-negative bacterium called AcetnDucter .ryylinum is cultured in anaqueousmediumcontainingacarbonsourcesuchas glucose at about 30°C. Such a macroscopic cellulose material is organized as a result of high-orderedaggregation of the normalribbonassemblies that are synthesized by the

Structure of Cellulose

97

FIGURE 13 Transmission electron micrographs of the negatively stained twisting ribbon assembly (a) and splayed microfibrils (b) produced in the presence of 1.0% CMC [19].

individual Acetobacter xylinum. Figure 13a shows a transmission electron micrograph of such a negatively stained normal ribbon assembly produced from a single bacterial cell. In Fig. 14 is shown schematically the process of the formation of the ribbon assembly from the bacterial cell on the basis of a large number of publications [42-441. More than several tens of cellulose-synthesizing sites are located in the cytoplasmic membrane, being parallel to the longitudinal cell axis. Cellulose synthetases in the respective sites produce 12-16 cellulose chains and extrude them into the culture medium as thin fibrils with a width of about 1.5 nm (often called subelementary fibrils) through small pores in the outer membrane. These subelementary fibrils aggregate with each other to form microfibrils, and the microfibrils furtheraggregate to produce the ribbon assembly with width 40-60 nm. The crystallization of cellulose will be induced during the different processesof the aggregations, because the subelementary fibrils may be too thin to be crystallized. In fact, subelementary fibrils have been confirmed to be noncrystalline for a specimen obtained by incubation in the presence of fluorescent brightening agents that prevent the aggregation of subelementary fibrils into microfibrils[42,43]. Since theparallel orientationof cellulose chains is already realized in the subelementary fibril, somewhat local-level rearrangements

Horii

98

shouldbeenoughtoinduce crystallization by aggregationinto the microfibril mainly through hydrogen bonding. Therefore, physicochemical factors affecting the aggregation process into microfibrils and the ribbon assembly will be associated with the crystallization of celluloses I, and I,. Of many possible factors, we first examined effects of the addition of different water-soluble polymers into the incubation medium on the formation of the two allomorphs, because carboxymethyl cellulose sodium salt and xyloglucan are known to interrupt the formation of the normal ribbon assembly [42-441. Low-molecular-weight compounds such as fluorescent brightening agents and direct dyes are also very effective in producing different types of fibrillar structures of cellulose. However, the crystallization into celluloses I, and I, is also highly interrupted in most cases [42-471.

B.

Effects of Polymeric Additives

Stationary cultures of Acerobacter xylinum were grown in Hestrin-Schramm’s medium at 28°C in the presence of carboxymethyl cellulose sodiumsalt (CMC), pea xyloglucan(XG), poly(viny1alcohol) (PVA), or poly(ethy1eneglycol) (PEG) [26,48]. Figure 15 shows CP/MAS I3C NMR spectra for bacterial celluloses cultured in the presence of the different polymeric additives [48]. The central line of the C l triplet decreases remarkably in intensity for bacterial celluloseincubated in the presence of 2.5 wt% CMC or 2 wt% XG, indicating the preferable crystallization of cellulose I, under such conditions. The increase in mass fraction of cellulose I, is also confirmed from the increase in intensity of the most upfield line in the C4 triplet and the concomitant decrease in intensity of the most downfield line. These changes were more clearly observed in the spectra of crystalline componentsrecorded selectively [48]. In contrast, almostnochange in relative intensity is observed in the C l and C4 triplets for the samples cultured in the presence of 20 wt% PVA or 25 wt% PEG. Similar addition effects were also reported for glucomannan as well as XG, and a marked decrease in crystallinity was found in the case of the former polysaccharide [49]. In Fig. 16 the mass fraction of cellulose I, is plotted against the concentration of CMC for samples with different degreesofpolymerization (DP) and different degrees of substitution (DS) [29]. Here, the mass fraction of the I, form was determined by the FT-IR methodbasedon Eq. (1). As is clearly seen in this figure, the mass fraction of cellulose I, decreases markedly with increasing CMC concentration. The most prominent decrease is observed for CMC with DP = 80 and DS = 0.57, suggesting the existence of an optimal DP and DS for the effect of CMC on the crystallization of the two allomorphs. Since there is almost no change in the degree of crystallinity in this system, CMC really promotes the preferable crystallization of cellulose I,, possibly as a result of the suppression of the crystallization of cellulose I,. Similar reduction in the mass fraction of the I, form was also confirmed in the case of XG, but the extent of the decrease was not so prominent even for the sample with the optimal molecular weight compared to the case of CMC [29]. As described above, CMC and XG interrupt the aggregation of microfibrils into the normal ribbon assembly, probably by the adsorption of CMC or XG on the surface of the microfibrils through hydrogen bonding. We have also confirmed such interruption of the aggregation by transmission electron microscopyasshown in Fig. 13bandfound that these splayed microfibrils were remarkably reduced in average diameter with increasing concentration of CMC or XG. Moreover, it has been finally clarified that there exists a simple linear correlationship with the mass fraction of cellulose I, and the average size of microfibrils for bacterial celluloses incubated in the presence of CMC, XG, and methyl

99

Structure of Cellulose

C

J

l

l

l

l

~

l

l

l

l

l

f

l

l

l

l

l

l

l

l

l

~

100

~

l

f

I

I

~

I

~

~

~

1

1

~

80

~

I

1

~

~

~

I

I

~

I

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

60 ppm f r o m T M S

FIGURE 15 CP/MAS "C NMR spectra for bacterial celluloses cultured in the presence of different polymeric additives: (a) control; (b) 2.0% XG; (c) 2.5% CMC; (d) 25% PVA; (e) 10% PEG [481.

[

1

1

1

1

l

*

~

Horii

100

1 0

0.5

1.o

1.5

Concentration of CMC l wt%

FIGURE 16 Mass fraction of cellulose I,, in bacterial cellulose cultured in the presence of CMC with various degrees of polymerization (DP) and degrees of substitution (DS) versus concentration of CMC: ( 0 ) control; ( 0 ) DP = 630, DS = 0.65; (A)DP = 80, DS = 0.57; ( 0 ) DP = 40, DS = 0.57; ( 0 ) DP = 80, DS = 1.43 [29].

cellulose [40,50]. It was therefore suggested that cellulose I,, is preferably crystallized in larger-size microfibrils, whereas cellulose I, is dominantly produced in thinner microfibrils. The mechanism of the preferential crystallization of celluloses I, and I, in microfibrils with different sizes will be discussed after the description of the crystallization of cellulose I1 in the bacterial cellulose system.

C.

Crystallization of Cellulose II in Nature

Cellulose 11, which is another representative crystal form of cellulose, is usually obtained by crystallization from a solution or by regeneration from alkali cellulose after mercerization. In the bacterial cellulose system the so-called native band material is sometimes observed in an irregular region where the normal ribbon assemblyisdisrupted by the formation of this material [42,44]. It has recently been revealed that such band material is composed of cellulose 11, by using a spontaneous variant of Acrtobucter xylinum that produces the band material as a main product under the standard culture condition [51,521. More recently, we have found that Acefobacter q l i n u m , which is collected from smooth colonies on a solidified Hestrin-Schramm medium, preferentially produces the native band material in incubationat 4"C, whereas the same bacterium produces the normal ribbon assembly at 28°C as usual [53].This indicates that the crystallization of cellulose I or I1 evidently depends on the culture temperature in this bacterial system. It is therefore very important to investigate this system in detail to clarify the mechanism of the crystallization

Structure of Cellulose

101

of celluloses I, and I, as well as cellulose II. Here, experimental results at the initial stage are briefly described, because investigation has just started with this system. Figure 17 shows transmission electron micrographs of the negatively stained band materialspreparedbyincubationfor 3 h in Hestrin-Schramm’s medium at 4°C on an electron microscope grid [53]. As is clearly observed in Fig. 17a, the band material is extruded out perpendicularly to the long axis. This band material seems to be composed of small irregular granules loosely linked to form strandlike structures in lying the direction of band extrusion, as pointed out in the previous paper [51]. However, when the micrograph is enlarged by 7.5 times as shown in Fig. 17b, most strands seem to be composed of irregularly coiled or folded thin fibrils. Since each strand is basically extruded from each pore on the surface of the bacterium in agreement with the previous finding [51], each strand will correspond to the subelementary fibril with an irregular coiled or folded appearance. In some areas parallel orientation of several units of the subelementary fibril are observed, but almost no discrete electron diffraction due to the crystalline entities could be detected for this band material by selected-area electron diffractometry. In contrast, three discrete diffractions are obtained on the equatorial line for the band material detached from the bacterium cell, as shown in Fig. 18 [53]. These three diffractions are well indexed as (1-10). (110), and (020) by using the unit cell of cellulose 11 [13], in good accord with the previous result [51]. Although they are considerably arced, molecular chain axes are confirmed to be almost perpendicular to the extrusion direction of the band bacterial. It is therefore concluded that cellulose 11 is partly formed in the native band material, but it is very important to obtain more informationabout the detailed structure of the band material, including the existence of folding structureof subelementary fibrils, which was suggested previously [51,52], to propose the mechanism of the crystallization of cellulose 11. In this bacterial system, another important finding is that the band material and the normalribbonassemblyaresequentiallyproducedwhentheincubationtemperatureis changed between 4°C and 28°C [53]. Figure 19 shows the drastic change in the cellulose production from the band material to the normal ribbon assembly as observed by electron microscopy. In this case a TEM grid with a culture drop containing Acetobacter xylinum that had been kept at 4°C for 5 h was transferred into an incubator regulated at 28°C and the culture was carried out there for 7 min. It is clearly found that the band material,

FIGURE 17 Transmission electron micrographs of the negatively stained native band material that was cultured at 4°C [53].

102

Horii

FIGURE 18 Selected-area electron diffraction pattern of the native band material detached from the bacterial cell. Three equatorial reflections are ascribed to those of (l-lo), (110). and (020) planes of cellulose II [53].

which was formed at 4"C, is linked sequentially to the normal ribbon assembly that is produced at 28°C. It is very difficult at present to describe how the production of the band material or the normal ribbon assembly is controlled in this system. However, the movement of the bacterium may be associated with the formation of these structures. Previous lightmicroscopic observation [42,44,54] revealed that the bacterial cell rotates around its longitudinal axis as it is propelled forward by the elongating ribbon during the production of the ribbon assembly. We also observed similar specific movements of the cells in the culture medium by a differential-interference light microscope with a TV system [55]. In

FIGURE 19 Sequential production of thenormalribbon assembly andthebandmaterial successive 28OC/4"C incubation of Acerobactor xylinum from the smooth colony [53].

by the

Structure of Cellulose

103

contrast, their movements are not specifically similar to the well-known irregular movements of Escherichia coli, when the cellulose is not produced. More careful observation of the movements of Acetohacter xylinum is necessary at 4 and 28°C to elucidate the processes of the different fibrillar structures.

D. Hypotheses on the Crystallization Mechanismof Celluloses I, and I, First it should be pointed out that the normal ribbon assembly shown in Fig. 13a is always twisted with a periodicity of about 1 p m around the longitudinal axis in one direction, possibly as a right-handed helix. Since such twisting is also observed for a long ribbon or intertwined ribbons, the twisting force must be induced not from the terminal of the ribbonbutfrom the cell sideduring or after the formation of the ribbonassembly. A previous report has suggested that the twist of the ribbon might be attributed to some property of cellulose molecules or to their interaction at the cell surface [43]. However, we can also see similar overall twisting for splayed microfibrils obtained by incubation in the presence of CMC, as shown in Fig. 13b. The latter type of twisting is not due to any cause attributed to the individual splayed microfibrils, particularly to cellulose molecules, but it could be produced simplyby the rotation of the bacterial cell around the longitudinal axis, which is usually observed by light microscopy as described above. It is therefore reasonable to assume that the overall rotation of Acefobacter xylinum also induces the twist of the normal ribbon assembly. During the production of the normal ribbon assembly the bacterial cell also moves translationally along the twisting ribbon axis together with the rotation. These cooperative movements will promote the aggregation of subelementary fibrils into microfibrils and furtherformation of the twisting ribbonassembly. In the incubation at 4"C, where the native band material is produced instead of the ribbon assembly, such specific movement may be hindered for some reason unknown at present, although subelementary fibrils are still allowed to be extruded into the culture medium. The next important problem is the possible effect of the rotational motion of the bacterial cell on the crystallization of celluloses I, and I,. Recently we have elucidated the effect of twisting a thin plate like the ribbon assembly. Figure 20 shows the process of twisting for a thin deformable plate PQRS around its longitudinal axis A. If the side PQ is rotated by angle 8 around the A axis under the fixation of the side SR, the longitudinal lengths PS and QR are extended to P'S and Q'R, respectively, by such twisting. When the plate is assumed to be deformed according to Hooke's law, the stress F produced by the twisting will be expressed as

Here, E is the Young's modulus of the plate along the longitudinal direction, r is the halfwidth of the plate, and 8 is the twisting angle around the A axis, which is defined for a given length L of the plate. Equation (2) clearly indicates that shear stress F, which is proportional to r2, is induced along the longitudinal direction of the plate, when the plate undergoes twisting around the long axis. The normal ribbon assembly is not a homogeneous thin plate as shown in Fig. 20, but the constituting microfibrils with somewhat different sizes are closely connected with each other, probably through hydrogen bonding. Moreover, the twist of the ribbon will be produced during the aggregation of the microfibrils. Nevertheless, when the normal ribbon assemblyundergoes twisting around the longitudinal axis before crystallization, shear

104

Horii

A P'

'

I

-r

FIGURE 20

-

Schematic diagram for twisting of a thin plate [19].

stress may be induced along the longitudinal axis in proportion to the squares of the half width of the ribbon. Under such shear stress, cellulose I,, will be preferentially produced through so-called stress-induced crystallization, because the orientation of the molecular chains in the I,, form seems to be organized by the shear stress along the chain axis. That is, the cellulose chains in a given ( 1 10) plane are shifted upward or downward by c/4 against the corresponding chains in the upward or downward successive (110) planes, as shown in Fig. 12. In addition, cellulose I, will be crystallized mainly in the regions along both edges of the ribbon assembly, because the shear stress is higher in those regions. In contrast, cellulose chains are packed in a more stable form in the unit cell for cellulose I,, asshown in Fig. 12. This type of crystal may beformedunder less shear stress or without shear stress, which may be referred to as stress-jree crystcrllization. According to the structural model shown in Fig. 20, cellulose 1, will be crystallized in the central core region because the shear stress is less or free in this region. When the ribbon assembly is partly disordered or splayed into somewhat thinner microfibrils, such a part may be also composed of cellulose 1, because the twisting induces much less shear stress for disordered or smaller-size microfibrils. The dependence of the mass fraction of cellulose I, on the size of microfibrils described in the previous section is also well interpreted in terms of the structural model shown in Fig. 20. In the case of Microdicfyon cellulose, in which 1, and I, crystals appear alternately along a single microfibril as describedabove [331, another crystallization mechanism should be proposed: the so-called two-step orientationcrystallization [ 191. Before the initialization of the crystallization, the cellulose chains are highly oriented in each microfibril just as in the liquid crystalline state, but they are not fully extended to the molecular chain length. Therefore, after the crystallization of cellulose I,, under some different stress (stress-induced ctystallization), the residual part of such chains left in the noncrystalline state will be relaxed in length as a result of the full extension of the crystalline chains. This willlead to the relaxation of the stress, andthen anothertype of crystallization,

Structure of Cellulose

105

stress-free crystallization, may occur to form cellulose I, in the central part of the noncrystalline region. Thissuccessivetwo-step orientation crystallization will producethe alternate I, and I, crystals along the single microfibril. Finally, it should be noted that cellulose I can be crystallized in vitro after a cellulasecatalyzedpolymerizationofP-cellobiosyl fluoride substratemonomer in acetonitrile/ acetate buffer [56].Under the normal condition employing unpurified cellulases, cellulose I1 is crystallized in a reaction medium after the polymerization as expected in analogy of the crystallization of cellulose from the solution [57]. In contrast to this fact, the cellulose I allomorph was confirmed by selected-area electron diffractometry for fibrous materials produced by using substantially purified cellulaseenzyme,although the discrimination between the I, and I, forms was not achieved [56].Such an interesting result may be due to the parallel orientation of extended cellulose chains with the same polarity, which is assumedto be organized as a result of a micellaraggregation of the partially purified enzyme and the substrate in the nonaqueous/aqueous solvent system. More detailed investigations of this system will also contribute to the exact interpretation of the crystallization mechanism of native cellulose.

VII. CONCLUDINGREMARKS As described above, the composite crystal model has been established in native cellulose, although there may be minor exceptions depending on the sources of cellulose. Nevertheless, the detailed structure, such as the chain conformation, hydrogen bonding, and molecular packing, has not yet been clarified in both allomorphs because precise structural analyses of the intensities of X-ray or electron diffraction diagrams have not been performed systematically so far. The situation is more serious for cellulose I,, because no pure I, specimen is available yet either in nature or by an artificial method. As a better way, the analysis will be made for specimens with higher contentsof the I,, form, assuming the linear combination of diffraction intensities of the I, and I, forms. In that case it should be taken account that the I, form is frequently subjected to some modification depending on the source of cellulose as shown in Fig. 3. As for the crystallization processes of celluloses I, and I, as well as cellulose I1 in nature, more detailed observations are necessary for the microfibril or subelementary microfibril structure at the level of several nanometers to several tens of nanometers by transmissionelectronmicroscopyandalso for the overall movementsof Acerobacter xylinurn by high-performance light microscopy. Information about the enzymatic degradation, acid hydrolysis, alkaline treatments, etc., for microfibrils, which is not described here because of space limitations, will be also very helpful for understanding the structures of microfibrils andsubelementary microfibrils. Through these investigations ourunderstanding of the structure of native cellulose will be greatly advanced in the next 5 or 10 years.

ACKNOWLEDGMENTS

The author thanks Dr. Asako Hirai and Hiroyuki Yamamoto for their constant cooperation throughout the work described here. He is also grateful to Prof. Masaki Tsuji and Prof. Junji Sugiyama for their kind cooperative contributions to structural analyses by electron microscopy.

106

Horii

REFERENCES 1 . S. Nishikawaand S. Ono, Proc. Tokyo Math.-Phys. Soc., 7131 (1913). 2. K. Okamura, in Wood and Cellulosic Chemistry (D. N.-S. Hon and N . Shiraishi, eds.), Marcel Dekker, New York and Basel, p. 89 (1991). 3. L. M. J. Kroon-Batenburg, and J. Kroon, Carbohydrate in Europe, p. 15 (1994). 4. A.C.O’Sullivan, Cellulose, 4: 173 (1997). 5 . R. H. Atalla, J. C. Cast, D. W. Sindorf, V. J . Bartuska, and G. E. Maciel, J . Ant.Chern. Soc., 102:3249 (1980). 6. W. L. Earland D. L. VanderHart, J. Am. Chem. Soc., /02:3251 (1980). 7. F. Horii, A. Hirai, and R. Kitamaru, Macromolecules, 20:2117 (1987). 8. F. Horii, in Nuclear Magnetic Resonance in Agriculture (Pfeffer, P. E., Gerasimowicz, W. V., eds.), CRC Press, Boca Raton, FL, chap. 10 (1989). 9. F. Horii, A. Hirai, and R. Kitamaru, J. Carbohydr.Chem., 3:641 (1984). 10. A. Hirai, F. Horii,andR.Kitamaru, CelluloseChem.Technol., 24:703 (1990). 11. H. J. Marrinan and J. Mann, J. PolymerSci., XXI:301 (1956). 12. G. Honjo and M.Watanabe, Macromolecules, /81:326 (1958). 13. A. Sarko and R. Muggi, Macromolecules, 7486 (1974). 14. K. H. Gardnerand J. Blackwell, Biopolymers, 13:1975(1974). 15. C. Woodcock and A. Sarko, Macromolecules, 13:1183(1980). 16. R. H. Atalla and D. L. VanderHart, Science. 223:283 (1984). 17. D. L. VanderHart andR. H. Atalla, Macromolecules, /7:1465 (1984). 18. H. Yamamoto and F. Horii, Macromolecules, 26: 13 13 ( 1993). 19. F. Horii, H. Yamamoto, and A. Hirai, Mucromol. Symp., 20: l97 (1997). 20. F. Horii, H. Yamamoto, R. Kitarnaru, M. Tanahashi, and T. Higuchi, Macron~olecules,20:2946 (1987). 21. H. Yamamoto, F. Horii, and H. Odani, Macromolecules, 22:4130 (1989). 22. A. Hirai, F. Horii, and R . Kitamaru, Macromolecules, 20:1440(1987). 23. H. Chanzy, B. Henrissat, M. Vincendon, S. F. Tanner, and P. S. Belton, Cnrbohydr. Res., 160: 1 (1987). 24. M. Tanahashi, T. Goto, F. Horii, A. Hirai, and T. Higuchi, Mokuzai Gakkaishi, 35:654 (1989). 25. E. M.Debzi, H. Chanzy, J. Sugiyama, P. Tekely, and G. Excoffier, Macromolecules, 246816 (1991). 26. J. Sugiyama, T. Okano, H. Yamamoto, and F. Horii, Macromolecules, 23:3196 (1990). 27. J. Sugiyama, J. Persson, and H. Chanzy, Macromolecules, 24:2461(1991). 28. J. H.Wiley and R. H. Atalla, in The Structure of Cellulose (R. H. Atalla, ed.), ACS Symp. Ser. 340, American Chemical Society, Washington, DC, p. 151 (1987). 29. H. Yamamoto, F. Horii, and A.Hirai, Cellulose, 3:229 (1996). 30. P. S. Beleon, S. F. Tanner, N. Cartier, and H. Chanzy, Macromolecules, 22:1615 (1989). 31. M.Wada, T. Okano, J. Sugiyama, and F. Horii, Cellulose, 2:223 (1995). 32. R.H.Newman, J . WoodChem.Technol., 14:451(1994). 33. J. Sugiyama, R. Vuong, and H. Chanzy, Macromolecules, 24:4168 (1991). 34. K. H. Meyerand L. Misch, Helv. Chim. Acta, 11534 (1937). 35. D. P. Miller and A. Li, in Cellulose and Wood, Chemistty and Technology (C. Schuerch, ed.), Wiley, New York, p. 139 (1989). 36. R. P. Millaneand T. V. Narasaiah, CelluloseandWood:Chetnistty und Technology (C. Schuerch, ed.), Wiley, New York, p. 39 (1989). 37. A. P. Heiner, J. Sugiyama, and 0. Teleman, Curbohydr. Res., 273:207 (1995). 38. B. J. Hardy, and A. Sarko, Polymer; 37:1833(1996). 39. F. Horii, A. Hirai, H. Yamamoto, and J. Sugiyama, Preprinfs of the3rdAnnual Meeting of the Cellulose Society of Japan, published by the Cellulose Society of Japan (% Institute for Chemical Research, Kyoto University), p. 1 1 (1996). 40. A. Hirai, F. Horii, M. Tsuji, J . Sugiyama, and H. Yamamoto, Proc. 1st Int. Symp. Inst.Chem. Res., Kyoto Univ. (ICRIS ’96), p. 132 ( 1 996).

Structure of Cellulose

107

41. Y. Kataoka and T. Kondo, Macromolecules, 29:6356 (1996). 42. C. H. Haigler and M. Benziman, in Cellulose and OtherNatural Polymer Systems. Biogenesis, Structure and Degradation (R. M. Brown, Jr., ed.), Plenum Press, New York-London, p. 273 (1982). 43. C. H. Haigler and H. Chanzy, in Cellulose and Wood: Chemistry andTechnology (C. Schuerch, ed.), Wiley, New York, p. 493 (1989). 44. C. H. Haigler, inBiosynthesis and Biodegradation of Cellulose (C. H. Haigler and P.J. Weimer, eds.), Marcel Dekker, New York-Basel-Hong Kong, p. 99 (1991). 45. A. Kai and H. Kitamura, Bull. Chem. Soc. Japan, 58:286 (1985). 46. A. Kai, F. Horii, and A. Hirai, Mucromol. Chem., Rapid Commun., 12:15 (1991). 47. A. Kai, P. Xu, F. Horii, and S. Hu, Polymer; 35:75 (1994). 48. H. Yamamoto and F. Horii, Cellulose, 1:57 (1994). 49. J. M. Hackney, R. H. Atalla, and D. L. VanderHart, Int. Bio. Mucromol., 16:215 (1994). 50. A. Hirai, H. Yamamoto, M. Tsuji, and F. Horii, Proc. '94 Cellulose R&D, Cellulose Society of Japan, p. 41 (1994). 51. S. Kuga, S. Takagi, and R. M. Brown, Jr., Polymer; 34:3293 (1993). 52. H. Sibazaki, S. Kuga, F. Onabe, and R. M. Brown., Jr., Polymer; 36:4971 (1995). 53. A. Hirai, M. Tsuji, and F. Horii, Cellulose, 4:239 (1997); more detailed results will be published elsewhere. 54. R. M. Brown, Jr., and D. Montezinos, Proc. Natl. Acad. Sci. USA, 73:143 (1976). 55. A. Hirai and F. Horii, to be published. 56. J. H. Lee, R. M. Brown, Jr., S. Kuga, S. Shoda, and S. Kobayashi, Proc. Natl. Acad. Sci. USA, 91:7425 (1994). 57. S. Kobayashi, K. Kashiwa, T. Kawasaki, and S. Shoda, J. Am. Chem. Soc., 113:3079 (1991).

This Page Intentionally Left Blank

Chemistry of Lignin Akira Sakakibara and Yoshihiro Sano Hokkaido University, Sapporo, lapan

1.

INTRODUCTION

Lignin exists as one of the essential wood components, ranging in amount from 10% to 30%. It is thought that lignin is a polymer formed by the enzymatic dehydrogenation of phenylpropanesfollowed by radical coupling.Softwood lignin is composedmainly of guaiacyl units originating from the predominant precursor, rrans-coniferyl alcohol (l), while hardwood lignin is composed of both guaiacyl andsyringyl units derived from transconiferyl (L) and trans-sinapyl (2) alcohols, respectively. Grass lignin contains p-hydroxyphenyl units derived from trans-p-coumaryl alcohol (3),besides units originating from the foregoing two precursors. However, strictly speaking, almost all plants consist more or less of all three units, namely, guaiacyl, syringyl, and p-hydroxyphenyl moieties. Lignin has no optical activity, in contrast to other compounds, because the radicals formed by enzymatic dehydrogenation couple with one another at random to give the lignin polymer. (OH

OH

(OH

OH

con OH

Lignin is the mostcomplexpolymeramong naturally occurringhigh-molecularweight materials, and investigations devoted to the elucidation of its structure have been under way for a long period of time. The presence ofmany complex carbon-to-carbon linkages between the units makes it difficult to degrade the polymer to low-molecularweight fragments. Furthermore,it has not yet been possible to isolate all parts of the lignin completely from plant tissues without engendering structural changes. [Bjorkman prepared milled wood lignin (MWL) that hasundergone little change,butalthough it is a very useful preparation, the yields are at most 50% of Klason lignin.] These characteristics make it hard to elucidate the chemical structure of lignin. The biosynthesis of lignin in vitro, worked out by Freudenberg and co-workers, provided guidance in approaching this 109

110 San0

and

Sakakibara

problem. Information on the degradation products from protolignin could provide direct evidence about lignin structure. For this purpose, the following procedures for bringing about lignin degradation are effective: catalytic hydrogenolysis, controlled hydrolysis, and degradation by thioacetolysis or thiacidolysis followed by reduction. As complementary procedures, oxidations of lignin with KMnO,, nitrobenzene under alkaline conditions, and acidolysis are available. The quantitative data for various functional groups and linkage types in protolignins are also essential to understanding lignin structure. Ultraviolet (UV), infrared (IR), and nuclear magnetic resonance (NMR) spectroscopic techniques, particularly when used in conjunction with chemical modification, have contributed to estimating the frequencies of functional groups and linkage types. Some lignin structural models could be proposed from these combined studies, but they do not represent a strict molecular structure such as those for other natural polymers such as cellulose and proteins. This may be unavoidable, however, in view of the biogenetic mechanism governing the formation of lignin.

II. SPECTROSCOPY

A.

Ultraviolet(UV)Spectra

Ligninshows a strongabsorptionspectrum in the UV region, becauseof its aromatic nature. The lignin spectrum of a typical softwood in Fig. 1 [ l ] has two maxima at 205 and 280 nm, shoulders at 230 and 330 to 340 nm, and a minimum at 260 nm. In general, softwood lignin shows a maximum at 280 to 285 nm, and hardwood lignin at 274 to 276 nm [ l ] . As the method is verysimple, UV spectrophotometric investigations are used extensively to characterize lignin preparations. Aulin-Erdtman [2-51 applied UV spectra to estimate effectively the amounts of certain functional groups, especially the phenolic hydroxyls of lignin, by means of differ-

200

250

300

350

400

450 nm

FIGURE 1 UV spectra of pine and beech lignins and

A E curve for pine lignin. (From Ref.

1.)

Chemistry of Lignin

111

entia1 measurements.Both the ionizedhydroxylandaldehydegroupscauseamarked bathochromic spectral shift. The phenolic hydroxyl content in polymers can be determined by comparing the adsorptionspectra in neutral andalkalinesolutions.Understrongly alkaline conditions, the absorptivity in the wavelength region of maximum adsorption for the phenolate ion is increased. The magnitude of this increase can be used quantitatively to determine the amount of phenolic hydroxyl groups. With this A&, method, Aulin-Erdtman [2]estimated the phenolicgroupcontentforsolublederivativesofspruce lignin (Picea abies) andBrauns’ lignin (BL) from Piceamariana and Tsugaheterophylla, in addition to DHP (a synthetic dehydrogenation polymer) by comparison with appropriate models. Further, phenylcoumaran and coniferyl aldehyde structures could be characterized spectrophotometrically [3].More closely related model compounds, such as coniferyl (L), sinapyl and dehydrodiconiferyl (M)alcohols, were also investigated with this A&, method [4]. From studies of about 40 p-substituted compounds, a series of regular absorptiondifferenceswerefoundbetweenthem [ 5 ] . A marked effect fromo-methoxyl groups on the planarity of 2,2-dihydroxy-biphenyls was observed. Biphenyl compounds showed a higher degree of coplanarity in dioxane, in which hydrogen bonds exist between the phenolic hydroxyls and solvent. The mono-ol showed a smaller interplanar angle in hexane, due to the presence of a seven-membered 0-H-O-H-OCH3 “ring” than in an acidic dioxane solution. where OH-dioxane bonds dominate.

(z),

(I)

OH

OH W

0 ,

.bCH3

H“

Pew [6] found that C-4-substituted guaiacyl and related compounds with unconjugatedsidechainsshow similar andalmost identical curves, as shown in Fig. 2 . These model compounds haveultraviolet absorption maxima at 280 nm, but unlike that for lignin, the curve tills abruptly to zero at 300 nm and nearly to zero at 250 nm. The filling in of the spectral trough at 250 nm in lignin can be explained by the presence of biphenyllinked units (Fig.2). Until then, biphenyl units in lignin molecules had generallybeen assumed to be minor contributors, butPew [7 I presumed that coniferous lignin may contain considerable biphenyl-linked units. A simpleand rapid method for the determination of phenolichydroxylgroups in lignin preparations was developed by Goldschmid [8] based on the A&, method. The phenol content of the sample is calculated from the absorptivity maximum of the resulting difference curve, and the molar absorptivity maximum of model phenols is determined i n the samemanner. The 300-nm maximum of the differencecurves is characteristic of phenolic hydroxyl groups without con.jugation. The method is suitable for routine appli-

112

Sakakibara and San0

1.0

0.8

E 0.6 m

-2 S1

2

0.4

0.2

0

300 320340360400

230240250260270280290

A-mp

FIGURE 2 UV spectra of spruce cellulytic enzyme lignin and of C4substituted unconjugated guaiacyl compounds. (From Ref. 6.) cation to technical softwood lignin preparations. Phenylcoumaran (g) (Arna 281 nm) undergoes dehydration during acidolysis to give a phenylcoumarone structure which has a strong absorption band at maximum 310 nm. By studying the spectral curves of these compounds, Adlerand Lundquist [9] estimated the contentof phenylcoumaran units in lignin. Schdning and Johansson [lo] studied the W absorption of lignin from pulp waste liquor and concluded that acid-soluble lignin components should be determined at 205 nm, because the absorption maximum at 280 nm is influenced by degradation products from carbohydrates, such as furfural. It was found that wood samples from spruce and pine contain 0.2%acid-soluble lignin, from birch and eucalyptus 3-4%, from wheat straw 2%, and from bamboo 1.5%. Klason lignins, especially those from hardwood, must be corrected for the acid-soluble lignin content. Wegener et al. [l11 found that hexafluoropropanol is an excellent solvent for UV and IR spectroscopy of lignins, because the absorption maximum of water-insoluble samples near 200 nm can be recorded exactly and evaluated quantitatively, due to the high UV transmittanceproperties of thissolventwithoutinterferencefromthedegradation products of polysaccharides.

(B),

B.

Infrared(IR) Spectra

Infrared (R) spectroscopy has been used often for the characterization of lignin because the technique is simpleand the sample to be studied does not need to be dissolved in any solvents and is required only in very small quantity. Qpical IR spectra of soft- and hardwoods are shown in Fig. 3. Various lignin preparations can be easily compared using this technique. Assignment of various absorption bands of lignins in IR spectra has been stud-

113

Chemistry of Lignin

I

I so0

I

I

I

1

I

1600

1100

120c

1000

800

Wavenurnber (cm")

FIGURE 3 FTIR spectra of MWLs from oak, birch, and spruce. Legend: oak (Q~erc-uscrispcln Blurne), birch (Berrrla plntyphylln Sukatchev var. jcrponicn Hara), and spruce (Pinus glehnii Mast.).

ied by a number of lignin investigators, using a variety of methods. The major absorption band frequencies and the most probable assignment of each band in guaiacyl and guaiacylsyringyl lignins are shown in Table I , which has been somewhat supplemented from that summarized by Hergert 1 1 21. Kolboe and Ellefsen [l31 used IR spectroscopy as an independent method for estimating lignin content and coumaran groups. The absorption at 1515 cm" was chosen for the determination of lignin, because this region is assigned to aromatic skeletal vibrations. The lignin content was estimated by the difference spectrum between the original wood and holocellulose at ISIS cm-', giving 28-29%. Thatis in agreement with the value generally accepted. Further, the absorption band at 1495 cm" was assigned to the coutnaran ring, and it was estimated to contribute about 5% of the total phenylpropane units. Sarkanen et al. 1141 comparcd the spectra of specifically deuterated guaiacyl and syringyl models with thosc of undeuterated lignin models, to give several new band assignments. From these deuteration studies, the hitherto unidentified bands at 1450 to 1420 cm" were considered to be associated with the ring-stretching modes strongly coupled with the CH in-plane deformation similar to the l SO0 c m - ' band. The 1340- l380 and 1250- l l SO cm" bandswere usually assigned to phenolic hydroxyl groups. The latter 1250- I 150 cm I band may be described as an "0-H in-plane deformation with ring-stretching character," and the 1340- 1380 cm ' band a s having a "ring-stretching with 0 - H bending character." The 1240 cm ' band scctncd to have the most pronounced methoxyl character,

and

114

Sakakibara

San0

TABLE 1 Assignment of Infrared Absorption Bands in Mildly Prepared Wood Lignins Band (cm"')"

Hardwood

2880

1595 1505

Softwood lignin

3450-3400 3425-3400 2940 2920 2875-2850 2820 1715-1710 1715 1675-1660 1650-1630 sh 1605 1515-1510 1495 1470- 1460 1430 1370 1270 1230 1 l40

970 915 860

Assignment stretching0-H C-H

1470- 1460 1425 1370- 1365 1330- 1325 1235-1230 1275

1085 1035

1030

970 855 815 750-770 sh

Carbonyl stretching in unconjugated and ketone conjugated carboxylic groups Carbonyl stretching in conjugated ketone groups Carbonyl stretching in y-lactone skeletalAromatic vibrations

1675-1660

1145 1130 1085

stretching methylene ingroups and methyl

Coumaran ring C-H deformations (asymmetric) Aromatic skeletal vibrations C-H deformations (symmetric) Syringyl ring breathing with C-0 stretching Guaiacyl ring breathing with C-0 stretching

C-H inplane deformation in guaiacyl C-H inplane deformation in syringyl C-0 deformation in secondary alcohol and aliphatic ether C-H in-plane deformation in C-0 guaiacyl, and deformation in primary alcohol (trans) deformation out-of-plane =CH Aromatic C-H out-of-plane deformation sh

750 sh

"sh: shoulder. Source: Ref. 12.

because on methylation, the intensity of the 1240 cm" band increases at the expense of the 1275 cm" band. MWLs from the sapwood of several wood species have been characterized by analytical and spectral methods [15]. The results indicate that the basic lignin structure of conifers is almost the same, and the differences between wood species seem to be the differences in the identity and amount of ester groups ( l 7 15 cm- ') related to various lignins. The IR spectra of lignins isolated from whole wood samples that include heartwoodshowedan extra absorptionband at 1630 cm", whichwas totally absent in the MWL from sapwood. Also, the IR spectrum of larch lignin showed a weak peak from a conjugated carbonyl group at 1670 cm-'. These extra bandsthat may originate from polyphenols were considered to be indicative of the chemical modification of lignin during the heartwood formation process. It was also demonstrated that the infrared spectrum is useful

Chemistry of Lignin

115

asan indicator of the ratio of syringylpropane to guaiacylpropane units in heartwood lignins [ 161.

C.

Nuclear Magnetic Resonance (NMR) Spectra

The magnetic resonance spectrophotometric technique is very useful to lignin chemistry, because it gives a variety of information on the lignin molecule that cannot be obtained by ordinary chemical analysis. Especially as quantitative data about linkage types obtained by chemical methods are incomplete, this technique will supplement such deficiencies.

1. Proton Magnetic Resonance (‘H NMR) Spectra Ludwig et al. [ 17,181 were the first to obtain ’H NMR spectra of lignin and its model compounds,usingdeuterochloroformasasolvent(Fig. 4). The chemical shifts of the NMR signals from protons in various model compounds, including biphenyls, p-0-4 dilignols phenylcoumarans (E),and pinoresinol (g), were determined. The NMR of dehydrodiconiferyl alcohol indicated a cis configuration in its furan ring, and the diequatorial configuration of pinoresinol (g) was confirmed. Thereafter, the NMR spectra of acetylated dioxane lignin, MWL, and BL were interpreted using the results from lignin model compounds [18]. It was found that semiquantitative estimates on the free benzylic hydroxyls, aliphatic and aromatic hydroxyls, and total aliphatic hydrogens in the lignin preparationswere possible. Moreover, it wasfound that NMRspectroscopy affords a unique method for estimating the degrees of condensation in lignin preparations.

(z),

(m)

0

0

f ;

1

A HCCl3

HMD

I

8.48 7.81 7.50 5.18 5.74 6.28

8 FIGURE4 Ref. 18.)

7

9.95

4

5

6

7

8

9

10

6

‘H NMR spectrum of acetylated spruce MWL (60 MHz, solvent: chloroform-d). (From

116 San0

and

Sakakibara

The presence of highly shielded protons in the range 6 1 SS-0.38 is noticeable. These protons are clearly due to methyl or methylene groups that are not attached directly to oxygen functions, carbonylgroups,aromaticsystems,orotherdeshieldinggroups,but their structural origin seems to be hydrocarbon contaminant because lignin does not include these groups. Lenz [l91 studied ‘H NMR spectra of bothunderivatizedandacetylated lignins, using various deuterated solvents. Twelve lignin preparations, including MWL, kraft and soda lignins, and dioxane lignin, were examined. It was found that alkali and acidolysis lignin preparations from both hard- and softwoods showed marked differences in the degree of condensation of aromatic rings, phenolic and aliphatic hydroxyl groups, and the number of highly shielded aliphatic protons. Spruce MWL gave a proton distribution close to that found by Ludwig [IS]. Bland and Sternhell [20] obtained from ‘H NMR spectra estimates of the fraction of protons attached directly to aromatic nuclei in some lignin preparations from Pinus radiata and Eucalyptus regnans. They estimated the frequencies of “condensed” units in the lignin molecule; those in methanol and acetylated lignin from E. regnans and methanol lignin from l? radiata are 0.60, 0.66, and 0.71/OCH3, respectively. However, it should be noted that the condensed units here are aromatic rings in which at least one carbon atom, not necessarily the 5-carbon alone, is linked directly to another carbon atom outside the ring. According to Morohoshi et al. [21], NMR spectra indicate that the degrees of condensed units in lignins from normal wood and compression wood of Abies sachalinensis are 0.48 and 0.79, respectively. Also,from the results of the other analytical data, the structure of compression wood lignin is generally composed of more condensed units. Horisaki et al. [22] estimated the frequencies of “condensed units” in lignins using methoxyl content, the ratio of guaiacyl and syringyl rings, and proton numbers assigned to aromatic rings, ethylene groups, and benzyl alcohol groups obtained by the NMR (500 MHz) of acetylated lignins and their hydrogenation derivative-containing internal standard (p-nitrobenzaldehyde) for quantitative analysis. They foundthat the contents of condensed units and benzyl alcohol groups are 0.35 and 0.29/CCJin spruce MWL, and 0.30 and 0.451 C, in birch MWL. The formylprotons in the lignin moleculecould not bedetectedwithlower-frequency (60-MHz) NMR spectrometers [ 17-19]. Later, Lundquist and Olsson [23] studied the formyl groups in spruce lignin at 270 MHz and found that the signals in the range 6 9.6- 10.0 were essentially due to the protons of aldehyde and vanillin units. By integration, the amount of coniferyl aldehyde units was determined to be 4%. MWLs from birch 1241 and spruce [25] were further studied. The spectra are shown in Fig. 5. The use of a 270MHz instrument and new techniques have greatly increased the amount of available structural information. Signals in the spectra of acetylated birch and spruce lignins are summarized in Table 2. The peak at 6 308 of birch lignin was attributed to H,, in p-p structures, and integration suggests that around 0.05/C,,C3 are involved in @-pstructures. The peak at 6 5.44 was assigned to H in p-5 structures or noncyclic benzyl aryl ethers, and 6 4.60 (H,) and 6.01 (HJ were essentially attributed to p-0-4 structures. Integration of the 6.01 peaksuggested that 40-50% of the units are attached to an adjacent unit by a p-0-4 linkage. In the case of spruce MWL, the peak at 6 2.28 (aromatic acetate) was found to correspond to 0.26 phenolic groups/C,J (although certain phenol groups in biphenyl structures are not included in this estimate [ 17,181. From the peak at a 6 2.62 attributed to H,, in p-p structures, 0.01 to 0.02/C,C, were estimated for dihydrofuran or open-type p-p units. Later, however,from the results of studies usingdeuterioacetonesolutions [261, higher values of 0.02 to 0.03/C,C3 were obtained. The integral of the 6 5.49 peak, H,, i n

Chemistry of Lignin

117

FIGURE 5 Spectral range 6 8.0- 10.5 (400 MHz, solvent DMSO-d,) of MWL from spruce. Notes: the assignments of the peaks are indicated in the figure (the peak at 6 8.14 is due to an identified contaminant). The dashed lines indicate the baselines used in the connection with quantitative estimates. (From Ref. 27.)

p-5 structures, corresponds to about 0.1 l/C,C,. p - 0 - 4 substructures were estimated to be 0.3 to O.5/C,C3 by integration of the 6.06 (H,) peak from p-0-4units. Li andLundquist[27]studiedphenolicgroups in ligninsanalyzed by the NMR spectrometry at 400 and 500 MHz using DMSO-d, and 300 K as shown in Fig. 6. Under the conditions the majority of the signals from protons in phenolic hydroxyl groups are found at 6 8.0-9.3, and those in most carbonyl-conjugated phenols at >9.3. They estimated that the number of phenolic groups in spruce MWL is 0.24/C9, of which about20% originatefromphenolicgroups (6 8.1 -8.4) in biphenyl and diary1 ethersubstructures. Spectra of MWL from birch exhibited separate peaks for phenols in guaiacyl units (6 8.59.1) and syringyl units (6 8.1-8.2), though the latter peaks overlap with those for biphenyl substructures. The total number of phenolic groups was estimated as 0.18/C,, in which the proportion in guaiacyl and syringyl units is 3:2. Ede et al. [28] studied 'H-'H COSY and J-resolved spectra of acetylated MWL from spruce,concluding that the results confirm thepresence of most of the known lignin structural units, but show a-0-4 aryl ether, p - l , and @-punits to be less significant in lignin structures than were previously thought because of no existence of each cross-peak.

2. Carbon-l3 Magnetic Resonance (I3CNMR) Spectra Ltidemann and Nimz [29-311 were the first tostudy "C NMR spectra of lignins. The chemical shifts of various carbons in lignin model compounds were assigned, and also the effect on the chemical shifts of the aromatic carbon atoms from methoxyls ortho to the 4-phenolic hydroxyl group was studied [29]. When a methoxyl group is introduced into an aromatic ring, the substituted carbon shifts about 32 ppm to a lower field, but on the other hand, both the ortho- and para-carbon atoms shift to a higher field.

118

Sakakibara and San0

TABLE 2 Assignments of Signals in the 'H NMR Spectra of Acetylated Birch and Spruce MWLs in Chloroform-d

S unitslppm 1.26 2.01, 1.95, 2.02 2.13 2.28-2.29 2.94 3.08 3.76-3.8 I 4.18, 4.27-4.28 4.39 4.43 4.60 4.65 4.70 5.44 5.49 6.0 1-6.06 6.93-6.94 7.4 1 7.50 7.53 9.64 9.84-9.86

contaminantHydrocarbon Aliphatic acetate Aliphatic acetate (including aromatic acetate in biphenyl structure) acetate Aromatic Unknown absent peak (the is in acetate the of lignin reduced with NaBH,) Hp in p-p structures Protons in methoxyl groups H, in several structures H, primarily in p - 0 - 4 (erythro) and p-5 structures H,structures in several H, in p - 0 - 4 structures (birch) H,, in p - 0 - 4 structures (spruce) including methylene protons in cinnamyl alcohol units H, in p-p structures (birch) including methylene protons in cinnamyl alcohol units H, in p-5 structures noncyclic and benzyl ethers aryl (birch) H,, in p-5 structures noncyclic and benzyl aryl ethers and H,, in aryloxypropiophenones (spruce) H,, in p - 0 - 4 and p-l structures, and vinyl protons Aromatic and vinyl protons Aromatic protons in benzaldehyde units, vinyl protons carbon the on atom adjacent to aromatic rings in cinnamaldehyde units (spruce) Aromatic protons located ortho to carbonyl groups (birch) Aromatic carbonyl protons groups located (spruce) to ortho protons Formyl in cinnamaldehyde units Formyl protons in benzaldehyde units

On the basis of the model compounds, the various signals of the "C NMR spectra of beech and spruce MWLs were assigned [29]. The spectra of DHP, spruce and beech ligninsareshown in Figs. 7 and 8, respectively. The numbering 1-40 of the peaksis convenient for comparing these spectra. The assignment o f . these carbon atoms is summarized in Table 3. These chemicalshiftscan be divided roughly intothreereg' 'Ions: carbonylcarbonsappearat 6 200- 160 (with respect t o TMS as a standard), Cl -C6 aromatic carbons, C,, and C, of double bonds on the side chain, at 6 160- 1 0 0 , and other saturated side chains C,,, C,,, and C, at 6 90 to 20. Methoxyl carbons always appear in the narrow range of 6 56.3 t 0.2. From these peak assignments, it was concluded that the content of phenylcoumaran units i n spruce lignin is much more than that in beech lignin and presumed that the content of dioxabicyclo-octane units [pinoresinol (3) and/ or syringaresinol (g)] of beech lignin may be more than that of sprucc lignin. The "C NMRspectra of thedehydrogenationpolymers."Zulauf"-and"Zutropf"-DHPs.were compared with thosc o f sprucelignin 1301, indicating that the conditions of formation

119

Chemistry of Lignin

I

6

9

8

7

6

5

4

3

I

I

2

1

0

FIGURE 6 ' H NMR spectra (270 MHz) of acetylated MWLs from birch and spruce wood. (From Refs. 24 and 25.)

affect the constitution of DHPs. Spruce lignin differs from Zulauf-DHP in a lower content of pinoresinol units and cinnamyl alcohol groups, and in a greater content of P-aryl ether units, a-carbonyl groups, and etherified guaiacyl residues. Nimz and Ltidemann 1311 also investigatedacetylatedlignins and DHPs. Acetylation makes it not only easy to assign signals, but also has the advantage of increasing the solubility of lignin preparations (Fig. 9). Gagnaire and Robert (321 studied a DHP polymer model of lignin that was synthesized by enzymatic dehydrogenation of coniferyl alcohol enriched to 90% in "C on the benzylic position. Gated proton decoupling and selective proton irradiation were used to facilitate assignment of the difficult "C, signals of the DHP. C,, atoms of various structural units, such as vanillin, vanillic acid,coniferylalcohol,cinnamaldehyde. @ - S , p-p, and p-0-4 dilignol units, and the C,, involved in benzyl etherbonds in p-0-4 dilignolstructures, were assigned. Obst and Ralph [33] have tried to determine the relative syringyl/guaiacyl ratios for hardwood lignins, and the following rcsults wereobtained. The syringyUguaiacyl integrated peak area ratio for red oak fiber MWL is I .69, and that for white birch fiber MWL is 2.28. Lapicrer and Monties 1341 estimatedeasily the ratios of syringyl/guaiacylfor hardwood lignins using the signal intensities of C-2 plus C-6 for each from the conventional NMR spectra.

120

Sakakibara and San0 3

19

6

FIGURE 7

33

a

5

24

29

33 34

L 4a

2

FIGURE 8

34

"C NMR spectrum of spruce MWL. (From Ref. 29.)

3 38

2a

22

"C NMR spectrum of birch MWL. (From Ref. 29.)

40

Chemistry of Lignin

121

TABLE 3 "C-Chemical Shifts (6) in ppmfrom TMS and Relative Intensities Spruce MWL (Fig. 7) and Beech MWL (Fig. 9)

of Acetylated

~

Beech, Signal no. 1

2 2a 3 3a 3b

4 4a 5

Spruce, PPm (intensity) 194.9 192.3 179.5 171.4 (159) 170.5 (95) 169.5 (34) 162.1 153.5

6 7 8 10

15 1.4 ( 5 5 )

11

140.8 137.7 (15) 136.7

12 13 14 15 16

17

149.1 148.3 145.4

134.1 (14) 132.4 (26) 129.3 (17)

18 19 20 21 22 23 24 24a 25a 25 26 27 28 29

123.6 (98) 120.7 ( 122) 118.8 (86) 1 16.5 112.9 (133)

30a 30 30b

76.5 75.5 (48) 74.8 (53) 73.3 (25) 72.6 (18) 70.0 (27) 68.3 ( 19) 66.2 (56)

31a 31

32a 32b 32

33a

90.1 88.5 86.3 85.4 83.6 80.7 (60)

PPm (intensity) 195.2 (161) 192.6 178.8 171.7 (161) 171.0 (161) 170.0 162.3 158.7 153.8 (197) 152.0 148.7 144.9 140.4 137.9 136.4 ( 1 9) 136.0 (28) 133.8 (34) 132.6 129.3 123.3 120.5 (31) 118.8 116.5 112.4 (35) 106.9 (45) 105.0 (222) 101.4 88.7 86.7 83. I 81.4 (157)

Assignment a-CO and y-CH0 in cinnamaldehyde a-CHO CO in primary acetoxyl CO in secondary acetoxyl CO in aromatic acetoxyl B(3). C3/5, C(3/5), D M , D(3/5), C,, in cinnamaldehyde A(3), A4, B4 B3 A3 C-4 in B-ring in cyclic p-5 (dehydro diconiferylalcohol acetate) A(4), B( 11, D( 1 ) D4 c 4 , C( 1) D1 BI, Cl A l , C/? in cinnamaldehyde C(4). D(4), C-l in cyclic p-5, C-2/6 in 17hydroxyphenyl ring 4 5 ) . B(5) A6,B6 AS, A(6), B(6) B5 A2, A@) C(2/6), D(2/6) C2/6, D2/6 C,, in cyclic p-S C,, in p-p (pinoresinol acetate)

73.3 (42) 72.6 (48)

C, in GOA (guaiacylglycerol-P-arylether acetate). C,,,,, in a$-diarylether C,, in p- l ( 1,2-disyringylpropane-1 ,3-diol acetate) C,, in GOA C,, in GOA (diasteromer) C, in open p-/?(dibenzyltetrahydrofuran) C, i n p-p

66.0 64.3 ( I 16)

C, in cyclic p-5 and cinnamyl alcohol acetate C, in p-l and a./?-diaryl ether

77.1 (36) 75.5 ( I 19)

55.4

Sakakibara and San0

122

TABLE 3

Continued

~~~~~~~

~~

Signal no. 33 34 35 51.4 35a 36

~

~

~

Spruce, PPm (intensity)

Beech, PPm (intensity)

64.1 (133)

41 .O

6 3 . 3 (197) 5 6 . 4 (255) 55.4 51.3 41.2

20.5 (255)

20.5 (255)

(255)

56.4

40

~

(11)

6

Assignment C, in GOA OCH3 C,, in 0-0 C,in p-1 and p-5 C,, in open p-p CH, in acetoxyl

I

A :R,=H.

0~~

R1 OR

R~=Ac B : R,=H, R H k y l C : Rl=OMc, RFAC D : R I S M C , Rl-

(R = alkyl or Ac) Source:

Ref. 34.

A

A

Ace t y 1a t e d SPRUCE MWL "quant i t a t i v e "

170

150

Acetylated SPRUCE MWL

"routine"

170

150

100

Chemistry of Lignin TABLE 4

123

Formulas forMilledWoodLignins

Wood species

Reference formula Methoxy-free C, formula

Spruce

Beech

111.

ANALYSES OF STRUCTURAL ELEMENTS IN LIGNIN

A.

Formula of Milled Wood Lignin (MWL)

Some representative formulas of milled wood lignins from soft- and hardwoods are shown in Table 4 135-371. These values vary according to the source and nature of the lignin preparations, for instance, wood age, preparation conditions of MWL, carbohydrate contamination, accuracy of determination, and so forth. The differences between species of hardwoods are especially remarkable, arising from the original differences in the nature of the protolignins. In Table 4, the extents of dehydrogenation and additional water per phenylpropane unit without the methoxyl group are shown. The oxygen atoms, excluding the two inherent in the p-hydroxycinnamyl alcohol moieties, are those from water molecules added to the quinonemethide and are shown in parentheses.

B.

End Groups with Unlinked Side Chains

As the lignin macromolecule is thought to be formed by the dehydrogenation of cinnamyl alcohols (1-2), it is probable that unsaturatedsidechains(-CH=CH-CH20H) are retained asendgroups. It is wellknown that Freudenberget al. [36] obtainedmany dilignols and trilignols with cinnamyl alcohol side chains(-CH=CH-CH,OH) and two dilignols with cinnamaldehyde side chains (-CH=CH-CHO) in the enzymatic dehydrogenationproductsfrom coniferyl alcohol (L). Similar lignols withthese side chains were also isolated by the mild hydrolysis of protolignins [38,39]. Treatment with phloroglucinol-HC1, known as the Wiesner reaction, is a typical color reaction of lignin, forming a violet cationic chromophore (Fig. 10). The reaction mechanism involving the coniferyl

OR (6) R= H or alkyl

(z,

FIGURE 10 Colorreaction of coniferyl alcohol groups.

I

124 San0

and

Sakakibara

aldehyde function was elucidated by Adler et al. [40]. From this color reaction, Adler et al. [42] estimated coniferyl aldehyde groups to be 2-2.5% in spruce lignin and 3-4% in BL of western hemlock. Later, however, the A&,, method indicated that 3-4% [42] and 3% [45] cinnamyl alcohol groups are present in spruce MWL. The existence of a small amount of coniferyl alcohol groups i n the lignin molecules was demonstrated with a color reaction exploited by Lindgren and Mikawa [44]. Coniferyl alcohol (1)and its 4-0-methyl ethers (8) react nitrosodimethylnitrile after tosylation, giving the p-dimethylaminoanilide of styrylglyoxalnitrile (y), which isredin color (maximum 475 nm), via intermediate (Fig. 11). Coniferyl alcohol groups in spruce lignin were estimated to have the same content as cinnanlyaldehyde moieties (2%) by this color reaction. The presence of cinnamic acid-type side chains in wood lignins is negligible, even to a lesser extent than in glass lignins [45,46]. Glycerolsidechains in lignin molecules are still beingdebated,buttwo p-0-4 dilignols with glycerol side chainswere isolated by hydrolysis with dioxaneandwater [47,48], extraction with water [49], and also by treatment with metallic sodium in liquid ammonia [50].A model experimentsubstantiated the fact that the glycerol side chain could not beformedfroma p-0-4 structure by hydrolysisunder neutral or acidic conditions [51]. Higuchi et al. [52] detected small amounts (0.03-0.6%) of three arylglycerols in the enzymatic dehydrogenation mixture of p-hydroxycinnamyl alcohols. This finding supports the contention that glycerol side chains can exist in the native lignin macromolecules, but the presence in lignins is rather insignificant because of formation of only small of formaldehyde by periodate oxidation 153,541. Dilignolsbearingo-propanolsidechains (-CH,-CH,CH,OH) havebeen isolated from the dioxane-water hydrolysate of protolignin [55]. The aglycones of glycosides (11)and (g) that were isolated as extractives from scotch pine (Pinus .syhwrris) were optically active [56]. On the other hand, the hydrolysis product is not optically active, suggesting that it may originate from lignin. The NMR spectra of MWLs show unknown proton signals in the higher-field region (6 0.38-1.58), and Ludwiget al. [IS] estimated the content of correspondingproton content to be 0.2-O.4/C,C3. The protons in this range may correspond to those of y-methyl and methylene. Compound (11) may represent one of these side chains. Presumably, some disproportionation may occur, or a reducing mechanism similar to that in the case of wood extractives may be operative during biogenesis.

(m)

(m)

HC

l

-0 I

HC

R7 O

O M e

R3

Cu1,

R,=H. R:=H

,Rs=OMe

Chemistry of Lignin

bo"

125

FH20H

OMe OMe

r I

OMe

aniline

OMe

OMe

0

Ce,pAmethylanin&lidc

of

styrylglyoxylnimlc ,red (mm.4751x11)

FIGURE 11

Color reaction of coniferyl alcohol groups.

known. Adler and Marton [59] determined carbonyl groups spectrophotometrically through the reduction of various aldehydes and ketones in guaiacylpropane structures with sodium borohydride in alkaline solution. The Acr changes in the adsorption spectra were treated by means of borohydride reduction curves, where the carbonyl compounds are reduced to the corresponding alcohols with different rates. Then, the A&,. curves of methylated and unmethylated spruce MWLs were qualitatively and quantitatively analyzed by Comparison with the A&, curves of modelcompounds. The results allowed the values for various carbonyl contents per OCH, in MWL to be estimated as shown in Fig. 12. In addition to these carbonyl groups (in total, 0.09-0.1 I/OCH,), unconjugated carbonyls may exist by as much as the same amount. The types of the unconjugated carbonyl groups in the lignin molecule have not been elucidated completely as yet, but some of this is due to glyceraldehyde-P-aryl ether substructure which the side chains have been eliminated. Examples of and As,, curves are shown in Figs. 13 and14.

D. Phenolic and Aliphatic Hydroxyl Groups The free phenolic groups have been quantitatively determined by various methods, such as the A&, method of Aulin-Erdtman as described in the section on UV spectra, conductometric or potentiometric titration in aqueous [60,61] or nonaqueous [62,63] solutions, reaction with dinitrofluorobenzene [64], reaction with sodium periodate [65], methylation with diazomethane [66], NMR spectra [18,19,27], etc. Adler and Hernestam [65] treated guaiacyl model compounds (Q) with sodium periodate, giving the corresponding o-quinones (@), which were determined quantitatively

C.!'& OMe

OH

OMe

/o

' OMe

OH

I

n

m

4.01/0CH3

0.0310CH3

O.OI/OCH,

FIGURE 12 Various carbonylgroupsin

&

' OMe

carbonyl

/o N 0.05-O.WOCH3

V 0.09-0.11/OCH3

lignin. (From Ref. 59.)

unconjugated

126

Sakakibara and San0

& x10 - 3

NaBH4 NaOH

0.01N

0 (b)

I

250

I

300

350 my

FIGURE 13 (a) Absorption curve of a-guaiacoxy-p-oxy-propioveratroneand the reduced one with NaBH,: (bj AF, curve.

by spectrophotometry (Fig. IS). From these results, it was estimated that 30% of all guaiacy1 units have freephenolichydroxyls.Thestructural units with afree phenolic group react with I-nitrosonaphthol to give compounds with a maximum absorption at SOS nm. Okay [67] has determined free phenolic hydroxyl groups in lignins by the application of this reaction. However, the values obtained by these methods do n o t always agree with each other, because they are not equally effective with all types of phenolic hydroxyls. For instancc, the AE, method cannot be used for determining hindered phenolic hydroxyls. It has been claimed that the potentiometric titration with sodium colamine in ethylenediamine could be used todetermine all freephenolicgroups (621. but i n fact, the values obtained are somewhat larger (0.33-0.34/OCH, in spruce MWL) than they should be because of the cleavage of benzyl aryl ethers. Furthermore, the oxidation with periodate

127

Chemistry of Lignin

I

4,O

UV absorption (neutral)

log E.

3.0

2.0

1 .o

250

300

350

1,m p

400

FIGURE 14 UV absorption in neutral solution and ionization A&-curves: a, a ' , untreated MWL; b, b', hydrogenated MWL; c, c', NaBH,-reduced and subsequently hydrogenated MWL.

[65] is applicable only to guaiacyl lignins (0.30/OCH3 in spruce MWL). The 'H NMR approach is simple and can be used to determine the protons of phenolic acetoxyl groups in acetylated lignins, but the accuracy is not satisfactory (0.27/OCH3 [ 191 and 0.29/OCH3 [ 181 in spruce MWL). Furthermore, Li and Lundquist [27] estimated by the NMR spectrometry of lignins at 400 and 500 MHz using DMSO-d, that phenolic groups in spruce MWL are 0.24/Cy, of which about 20% originate from phenolic groups in biphenyl and diary1 ether substructures, and those in birch MWL are 0. I WCy, in which the proportion in guaiacyl and syringyl units is 3:2. Robert and Brunow [68] have estimated the phenolic hydroxyl groups in MWL by "C NMR. Chang et al. [69] have determined the phenolic hydroxyls in cellulolytic enzyme lignins fromsweetgumandspruceandobtainedsomewhatlowervaluesthanthose in

.Q OH u1)

-Q N~OJ

OMe

+

CH30H

0

0 W

FIGURE 15 Oxidation of guaiacyl nucleus with sodium periodate.

nces

e

Sakakibara and San0

128

TABLE 5 Phenolic Hydroxyl Groups of Spruce MWLs Method Phenol-OH/OCH,

0.29, 0.27 0.27, 0.24 0.20 0.30 0.33-0.34 0.33 0.15-0.20

'H NMR NMR 'H "C NMR Periodate Titration Aminolysis

t18, 191

W , 271 L681 ~651

[621 [701

"Cellulolytic enzyme lignin

correspondingMWLs, 0.09-O.13/C,C3 for sweetgumand 0.15-O.20/C6C, for spruce. MAnsson1701 has developed a new method for the determination of phenolic hydroxyls in lignins by acetylation and subsequent selective aminolysis with pyrrolidine. The phenolic hydroxylgroupcontentsobtained so f a r are summarized in Table 5. Robertand Brunow 1681 also estimated the different types of hydroxyl groups in lignin preparations usinga "C NMR pulse sequence that involves gated proton decoupling. The carboxyl carbons in acetylated samples give signals that allow three different types of hydroxyl groups to be distinguished. The results obtained are shown in Table 6. Recently, the evaluation of 'H-, '.'C-, "P-NMR [71], FTIR (721, and wet chemical methods has been made to determine the contents of total hydroxyl, phenolic, and aliphatic hydroxyl groups in lignins [73], reporting that FTIR or 'H-NMR spectroscopy is recommended for routine determination of phenolic OH/aliphatic OH ratios, and that "P-NMR affords the determination of the stereochemical configuration of p-0-4 linkages and the phenolic OH contents due to G and S units.

E. Condensed Units Potassium nitrosodisulphonate, known as Fremy's salt, oxidizes p-substituted phenols to o-quinones. For example, 4-propylguaiacol is oxidized to methoxy-5-propyl-o-quinone (16) (Fig. 16) 1741. The guaiacyl units (B) possessing an unsubstituted 5-position are called "uncondensed," and the units that carry C-C or ether bonds at this position are called "condensed." Adler and Lundquist applied this oxidation to estimate uncondensed units in lignin. The o-quinone (16)formed can be quantitatively determined by means of spectrophotometry. Hereby, it was found that 0.15-0.1 8 units per methoxyl in MWL were uncondensed phenolic units, corresponding to 50-60% of the 0.30 phenolic

(e)

TABLE 6 Numbers of Hydroxyl Groups per OCH, Calculated from the Intensities of Acetyl Signals in Acetylated Samples and Those Obtained by Acetylation and Aminolysis OH Sample DHP Spruce MWL Suruce MWL

OH

Total Phenolic Secondary Primary OH OH

1.32 0.78 l .34 1.26

0.350.73

0.16 0.3 1

0.20 0.33

1681 1681 1701

129

Chemistry of Lignin

c-c-c

c-c-c OMe

OH

0

ui,

0

FIGURE 16 Determination of condensed units by oxidation of guaiacyl nucleus with Fremy’s Salt.

units present; i.e., 0.12-0.15 per methoxyl were presumed to be condensed units from the difference, corresponding to 40-50%. However, as the oxidation with Fremy’s salt can onlybeapplied to the units witha free phenolichydroxyl, no information is obtained about etherified units. By means of ‘H NMR spectroscopy, about 45-50% condensed units [18,21] and less amounts of them (35% and 30% for spruce and birch MWLs) [22] have been estimated. The latter were calculated by combination of the amounts of OCH, determined by the bromine method and of aromatic protons by ‘H NMR of hydrogenated MWLs using p-nitrobenzaldehyde as an internal standard.

F. p-5 LinkedUnits P-5 linkage units are represented by the phenylcoumaran structure (B).Dehydrodiconiferyl alcohol was isolated by the 0.5% HCl-methanol treatment of spruce wood by Freudenberg et al. [75], and phenylcoumaran lignols were isolated through dioxane-water hydrolysis [76,135]. Adler et al. [77] found that dihydrodehydrodiconiferyl alcohol (g) and its phenyl methyl ether are converted to the corresponding phenylcoumarone (B)in 90% yield after 20 h of heating in 0.2 M HCI (Fig. 17). The phenylcoumarone with its stilbenoid conjugationshowsastrongabsorption at 310 nm,and the phenoliccompoundgivesa Asi maximum at 338 nm. This conversion provides characteristic difference curves. Quantitative evaluation of the spectra indicated that the spruce MWL contained 0.1 1 dimeric structures per methoxyl that can be converted into phenylcoumarone moieties on acidolysis. When the MWL was methylated with diazomethane, the number of phenylcoumarone moieties formed on acidolysis was about 0.08. This means that the MWL contains about 0.03 P-5-condensedbut not ring-closeddimeric units (E), which,however,are converted into phenylcoumarone moieties closure on acidolysis. Therefore, the number of phenylcoumaran moieties (18) in the MWL is calculated to be 0.08 per methoxyl.

(B)

G.

Benzyl Alcohols and Benzyl Ethers

Benzyl alcohols and benzyl ethers are two of the most important functional groups in the lignin molecule for various reactions involving pulping. The determination of these groups is therefore especially significant. Various benzyl alcohols, dilignols, and tri- and tetralignols have been isolated by mild hydrolysis and hydrogenolysis, as will be described later (see Section W). However, noncyclic benzyl ethers are very difficult to isolate from the degradation products of lignin because of the labile nature of these ether linkages. Adler and Gierer C781 treated lignin with methanolic hydrochloric acid and concluded that the total amount of benzyl alcohol and noncyclic benzyl ether was about 0.43/OCH3 in spruce MWL, becausecyclicbenzylethers are not methylated.Benzylalcohol units withfree

130

Sakakibara and Sano CH2CH2CH20H

‘acidolysif

OMe

*

(O.2N Ha in dioxanc-

H20.2Oh.rdlux)

8-

OMe I

c-

0

OMe

OMe OH

OH 0

uz)

c-c-c

c-c-c

OMe

I

$L0

CHOR

OH

OMe OH

@OM. OH

FIGURE 17 Determination of p-5 linkedunitsby

acidolysis.

phenolic hydroxyls ( 2 0 ) have been determined with the quinone monochloroimide (2) color reaction by Gierer [79,80] (Fig. 18). Adler et al. [81,82] found that p-alkoxybenzyl alcohol (3) is oxidized to the corresponding aryl ketone ( 2 5 ) withdichlorodicyan-pquinone (2), which is reduced to (Fig. 19).The a-ketone (g) formed was determined spectrophotometrically, and a value of 0.16/OCH3 was obtained. Thus, p-hydroxybenzyl alcohol units constitute 0.05/OCH3, esterified benzyl alcohol 0.10/OCH3, and benzyl ether (except cyclic ether) 0.06/OCH3. Higuchi et al. [83] estimated a-aryl ethers to be 0.070.09/C,C3 from the results of the acidolysis of MWLs of bamboo, beech, and Thuja stundishii. Freudenberg et al. [84] estimated these groups by cleaving the benzyl ethers with

(B)

OMe OH

OMe

0

quimemonochlorimidc

indophenol

A mal. 6x)nm

FIGURE 18 Reaction of quinonemonochlorimide

(U)with p-hydroxybenzyl

alcohol groups (g).

z

131

Chemistry of Lignin R

NC

OMe

0

OMe

OH

OMe

W

0

Cl

OMe (21)

dichlomdicyan-pquinonc

FIGURE 19 Reaction of dichlorodicyan-p-quinone (g) with p-alkoxybenzyl alcohol groups (g).

sodium colamine in ethylene diamine solution. Besides this reaction, cleavage with methanolic hydrochloric acid after methylation and sulfonation were also applied [85].From the results, it was estimated that p-hydroxybenzyl aryl ethers amounted to about 0.041 OCH, and p-alkoxybenzylaryl ethers 0.06-0.09/OCH3. These results are summarized in Table 7. The benzyl alcohol units have also been determined semiquantitatively from NMR spectra as 0.33/OCH, [ 181, 0.32 [19], 0.31 [68], or 0.29 [22] in spruce MWL and 0.45 [22] in birch MWL. Gagnaire and Robert [32] have estimated the benzyl alcohol content to be 0.31/OCH3 from the "C NMR of DHP.

H. p-0-4 LinkedUnits Arylglycerol-p-aryl ether units (g) and (g) belong to the most important substructures in lignin molecules. It is well known that Hibbert's ketones are formed from these linkages in lignin during alcoholysis (Fig. 20). Adler et al. [86] found that acidolysis liberated about 0.3 phenolicOWOCH, in MWL andreportedasimilarnumberofterminalC-methyl groups characteristic of the side chains of Hibbert's ketones (31-34). The results indicated that p-0-4 units (2-2) may be 25-30% of all phenylpropanes. From 'H NMR spectra, Lundquist estimated later that 40-50% of birch lignin units [24] and 30-50% of spruce lignin units [25] are attached to an adjacent unit by a p-0-4 linkage. Miksche et al. have studied the oxidation products of lignins and have estimated the content of p-0-4 units in birchandspruceMWLstobe0.62 [87] and 0.49-O.51/C,C3 [88], respectively, by multistage oxidations.

TABLE 7 BenzylAlcoholandBenzylEtherContents Spruce Lignin

in

Hydroxyl ~~

p-Hydroxybenzyl alcohol p-Alkoxybenzyl alcohol p-Hydroxybenzyl ether ether p-Alkoxybenzyl

0.06

U311

0.05

V91

0.10 0.02 0.04

L811 l811 1841

0.06 0.06-0.09

P11 [g41

132 Sano

and

Sakakibara

cH20pb I

CH2OI I ?leo

I

H?--0

HC HC2 0-p oO

HC-0-C

HC-OH

I

b

C

-

O

b

II

CH

OMe

OMe OH

0

0

0

H,Oo

OMe

'

OH

0 0.30/C&

OH

Formation of Hibbert's ketones by acidolysis of p-0-4 units.

FIGURE 20

1.

p-l LinkedUnits

Nimz [S91 first isolated diarylpropanediols (g) from the degradation products of spruce and beech protolignins by mild hydrolysis. The mechanism of formation of the p-1 linked units (g) has been proposed by Lundquist and Miksche [90] as shown in Fig. 2 1. Phenoxy radicals (2)and maycoupletogive the cyclohexadienone (E),which maybe cleaved, giving the p- 1 linked compound (B) and glyceraldehyde-2-aryl ether (S). Lundquist and Miksche estimated the content of this aldehyde (0.3%) from the yield of methylglyoxal during acidolysis of lignin.

(x)

F

d

O

H

8,.""'

R1

J.

BiphenylStructures

Aulin-Erdtman [91] found that the mostobvious effect ofincreasing the pH of lignin solutions was a higherabsorption in the UV spectrumabove 300 nm.Thisabsorption band is characteristic of biphenyl structures. The number of hydroxy-biphenyl units in

133

Chemistry of Lignin CH2OH I

H?-OAr

HC-OAr

*

I

0CHOH

$

-

~ H O HOMe %,H'

I

I

H?

HrH

CH20H

CH20H

OMe +

-

I

CH

OMe

Q 0O M e

0

0

up,

H20

Hys;H

KMe I

+

OH ue,

FIGURE 21

CH0 I

HC-OAr I

CH20H

0

Formation mechanism of diarylpropanol on biosynthesis of lignin. (From Ref. 90.)

black spruce BL was estimated using a twice-reduced difference curve at 325-340 nm. The results obtainedindicated that there are 0.05/OCH3 biphenyl units in BL. Pew [7] estimated a still higher value, 0.25/C6C,, for spruce MWL. Miksche et al. estimated values of 0.045 for birch lignin [87] and 0.095-0.11/C,C3 for spruce lignin [88] from oxidation product yields. Nimz [38] estimated 0.O23/C,C3 for beech lignin. Between these estimated values there are considerable differences. From the permanganate oxidation products [92], in additionto the mainbiphenyl unit, 5-5 5-6 (g), 5-1 (Q), and 6-6 (g) type biphenyl units are presumed to be present, but have not been isolated from the products of hydrogenolysis and hydrolysis, indicating that they may give rise to minor products.

(e),

COOH

Me0

OMe

OMe OMe

K. 4-0-5 Linked Units Freudenberg and Chen [92,93] first isolated 4-0-5 (E,%) and 1-0-4 (g,%) type diphenyl ether compounds by permanganate oxidation, and then Larsson and Miksche [94] isolated two 4-0-5 type oxidation products (S,@) Nimz et al. [95] isolated a guaiacyl-syringyl dilignol with a 4-0-5 linkage (168) by treating beech lignin with thioacetic acid, followed by reduction with Raney nickel. Yasuda et al. [96] isolated a guaiacyl dilignol with a 4-

and

134

Sakakibara

San0

(m)

0-5 linkage from the hydrogenolysis products of larch compression wood lignin. Miksche et al. estimated the frequency of 4-0-5 linked units from the yield of permanganate oxidation products to be 0.035-0.04 for spruce lignin [88] and 0.065/C6C3for birch lignin [87]. On the other hand, Nimz et al. [95] reported 0.O15/C,C3 for beech lignin.

L.

p-p LinkedUnits

p-p type structures are involved in lignans represented by pinoresinol (44) and syringaresinol Of unitslinked in this way, (@ wasfound to be formed by enzymatic dehydrogenation of coniferyl alcohol (L) [97]. However, with the exception of the finding that pinoresinol is detected in the products formed by room-temperature methanolysis from spruce protolignin [75], isolation of that compound from lignin degradation products has failed. Syringaresinol,on the otherhand,was isolated from the productsof the mild hydrolysis of beech wood by Nimz [98]. Omori et al. [99] also isolated syringaresinol and episyringaresinol from Fraxinus mandshurica by hydrolysis with dioxane and water. Further, a dioxa-bicyclo-octane composed of guaiacyl and syringyl units (S) [48], dimethoxylariciresinal [ 1001, and a trilignol involving the syringaresinol moiety [99] were isolated, as will be described later. Nimz [95] also isolated compounds with the isolariciresinol ring andwith the tetrahydrofuran ring substituted byguaiacyland syringyl units In general, p-p linked units are involved much more in hardwood lignin than softwood lignin. Lundquist [25] estimated a low content for pinoresinol units (0.02-0.03/OCH3) in spruce MWL from its 'H NMR spectrum. Miksche et al. also estimated 0.03-0.05/C,C3 p-@units for birch lignin [87] and 0.O2/C,C3 forspruce lignin [88].OgiyamaandKondo [ 1011 estimated the content of pinoresinol structures as 0.05-0.10/OCH3 for softwood lignin from the yield of the di-y-lactone formed by nitric acid oxidation.

(e).

(m).

R2

HO$..(

OMe

H0

M. p-6 and p-2 Linked Units p-6 and p-2 linked units are presumed to exist in the lignin molecule, as metahemipinic (p) and hemipinic ( 7 8 ) acids were detected in the products of permanganate oxidation [92]. This substructure was first isolated as a dimeric phenylpropane from the hydrogenolysisproducts of spruce lignin bySudo et al. [102].Yasudaet al. [l031 isolated p-6 linkedphenylisochroman (E) from larch lignin by hydrogenolysis. These p-6 linked units may energetically form more easily an a - 0 - y linked phenylisochroman ring than an open structure, but it is not clear whether the p-6 linked units should always exist as a closed ring or not. Miksche et al. estimated the content of p-2 and p-6 linked units to be 0.015-0.025/C6C3 for birch lignin [87]and0.025-0.03/C6C3forspruce lignin 1881, respectively.

135

Chemistry of Lignin

N. Other Linkage Units Freudenberg et al. [93] isolated anaromaticacidinvolvinga 1-0-4 linkagefrom the permanganate oxidation products of lignin, suggesting the existence of 1-0-4 linked lignols (E,%), but such a compound has not been isolated as a phenyl-propanoid as yet. Nimz [95] isolated cyclolignan-bearing p-p and a-6 linkages from beech lignin by thioacetic acid degradation. This structure was postulated by Freudenberg [92] as a precursor of the benzene polycarboxylic acid that was found among the permanganate oxidation products.Nimzalso isolated a-p typecompounds (E) andtetrahydrofurandilignols involving 'y-0-7and p-p linkages Pew et al. [7] have suggested that the diphenyl ether 1-0-4 linkage is formed by side-chain displacement after radical coupling at C, and C,, leading, for example, to an intermediate ( S )in the formation of dioxepin (*), as shown in Fig. 22.

(x)

(m).

W.

DEGRADATION

A.

Oxidation

1. Alkaline Nitrobenzene

Freudenberg [l041 was the first to report that lignin provides a high yield of vanillin (E) by the alkalinenitrobenzeneoxidation. The yield of 20-28% fromspruce lignin proved the aromatic nature of lignin. Later, Leopold [ 1051 studied these oxidation products from spruce wood in detail, also deducing the presence of p-hydroxyphenyl units in softwood lignin (Table 8). Leopold [ 1061 and Pew [ 1071 demonstrated that the side-chain structure has considerable effect on the yields of oxidationproductsfrommodelcompounds. Units having side chains substituted with a hydroxyl at the a-position and vinyltype guaiacyl units give high yields of vanillin, and those with a-carbonyl groups increase the yield of vanillic acid (S). Units bearing alkyl substituents atthe 0-position to phenolic hydroxyl exhibit the highest resistance to oxidation. Substituents at the a-carbon of the side chains and cyclic ether structures such as pinoresinol are also difficult to oxidize. The products obtained by alkaline nitrobenzene oxidation are summarized in Fig. 23. Brink et al. [108,109] investigated the oxidation products (methylated) of white firin detail and reported various compounds besides the products.

(E)

MeO Me0 OH

OH

MeO

OMe OH

OH

m FIGURE 22

W

A mechanism of formation for diphenyl ethers. (From Ref.

0 dioxepin

7.)

OH

136

Sakakibara andSano

TABLE 8 NitrobenzeneOxidationProducts from Spruce Wood Compound

Yield (%)

Vanillin ( 5 4 ) p-Hydroxybenzaldehyde Syringaldehyde ( g ) Dehydrodivanillin ( g ) Vanillic acid (g) acid Syringic 5-Formylvanillic acid (63) 5-Carboxbvanillin (9) Dehydrodivanillic acid Acetoguaiacone (g)

(S)

(a)

(E)

27.5 0.25 0.06 0.80 4.8 0.02 0.1 1.2 0.03 0.05

Source: Ref. 105.

PermanganateOxidation Freudenderg et al. [92,93] heated spruce lignin or spruce wood with 70% aqueous potassium hydroxide in order to bring about hydrolytic cleavage of ether linkages and subsequently protected the phenolic groups liberated by methylation. Permanganate oxidation of the methylated products at pH 6-7 gave veratric acid ( 7 4 ) in a yield of about 8% of the lignin and minor amounts of isohemipinic and dehydrodiveratric acids (g,%). Furthermore, they isolated and identified 19 methoxy-substitutedbenzenecarboxylic acids.

2.

CH0

COOH

R1

R2

F2

71

M

e

O

m

O

M

OH Rl=H, R2=OMe (U) R,=RZ=OMe (12, Rl=Rz=H

0 Rl=CHO R2=OMe 0 R]=COOH, R2=OMe W Rl=CHzOH. RZ=OMe I

R l

OH R=H (hl) R=CHzOH

161)R=COCH3

c66)R=CHzCH2CH20H (42) R=CH2COCH3

FIGURE 23 Alkaline nitrobenzene oxidation products of lignin.

e

Chemistry of Lignin

137

Miksche and co-workers [ 1 10,ll l] found that considerably higher yields of the aromatic carboxylic acids were obtained if the oxidation was carried out by a mixture of sodium periodate and permanganate in aqueous t-butanol with sodium hydroxide at 82°C. Since the product mixture contained appreciable amounts of phenylglyoxylic acids, however, the latter acids were degraded to the corresponding benzoic acids in a secondary oxidation step, consisting of brief treatment with alkaline hydrogen peroxide. The latter method gave well-reproducible results as shown in Table 9. If wood or isolated lignin were methylated and oxidized, the resulting aromatic acids reflected the units in lignin which carried a free phenolic hydroxyl group. Preheating with alkali convertednonphenolic units intophenolic. This is, andfrom the increase in the yields of aromatic acids, therefore, the proportion of etherified units could be estimated [94]. Alkaline cleavage of the ether linkages was performed under conditions for kraft cooking or oxidation with alkaline cupric oxide. The mixture of benzoic acids was finally methylated, and the resulting mixture of methyl esters was assigned by gas chromatography and a combination of gas chromatography and mass spectrometry for quantitative estimation and structural identification, respectively. Table 9 shows the yields of the major acids obtained on degradation of a methylated spruce MWL and of the same lignin which had been subjected to ether cleavage prior to methylation [ 1 1 l]. The results obtained with methylated kraft lignin prepared from spruce wood meal is also included in Table 9, indicating the good reproducibility of the method, and spruce MWL is structurally very similar to the lignin in the wood. Also, pretreatment with NaOH/CuO gave considerably higher yields of most of the aromatic acids than pretreatment under kraft cooking conditions. A total of 40 aromatic acids has been identified in the reaction mixture obtained on oxidation of methylatedspruceMWLwhich had notbeensubjected to ether cleavage [ 112,1131. From spruce lignin, hemipinic acid biphenyls (4J-g,@-@), diphenyl ethers (87-90), andbenzenepolycarboxylic acids weredetectedbesides veratric isohemipinic (E), and metahemipinic acids as summarized in Fig. 24. More abundant aromatic acids were obtained from softwoods were (T),(g),and (@-g), and

(z), (z), (z,2),

TABLE 9 Yields of Methyl Esters (in mg/100 mg of lignin) Obtained by Oxidation of Methylated Spruce Lignin

MWL methylated

2.0

0.75 0.70.25

11.2

1.1

I S5

MWL treated with 2 M NaOH/CuO and methylated

29.8

5

0.7

1.1

2.1

5.0

MWL treated with kraft cooking conditions and methylated

21.4

0.5

21.3

0.5

3.6

I .75

Kraft lignin from wood meal methylated

Source:

0.45 Ref. 167.

0.6

5.93.4

I .65

6.0

(B),

138

Sakakibara and San0

RH

RQ C O O H

OMe R2

R1

COOH

OMe

(HL) R=H

0 R=OMe 0 R=COOH

COOH

COOH

I

Me0

Me0 R=H @&)R=OMe

OMe

Q

OMe

R

OMe

OMe

COOH OMe

@ Rl=Rz=H &l (81) RI-+Me ,Rz=H (86) RI=RflMe

OMe (8e) R=H

@Q) R=OMe

Me0 OMe

ceu

COOH

COOHA I

Me0 V

O

koM ‘’ ’ COOH

I

OMe M

eMe0

COOH COOH

Me0

OMe

/

OMe

OMe

OMe

0

OMe

W

OMe

W COOH

COOH COOH Me0 Me0 OMe

0 0

FIGURE 24

OMe OMe

0

OMe

Permanganate oxidation products of lignin.

(z,z),

those from hardwoods were (E),(Q), and (86-83). Other aromatic acids appear in small amounts (around 0.1 % of the lignin) or in traces (>O. 1%). The diaryl ether (g) and the biphenyl acids ( g )being , monocarboxylic acids, originated from substructures from which one of the side chains has been detached. The tricarboxylic acid (3) is one of the examples indicating a mixed radical coupling (coniferyl and p-coumaryl alcohols) and, analogously, the diaryl ether ( g ) is derived from a substructure formed from a sinapyl and a coniferylradical. The trimethoxylated ring in the trace constitutents were remarkable.

Chemistry of Lignin

139

The methoxyhydroquinone ring in these acids may be have been formed by reduction of a methoxy-p-quinone moiety which seems to be the hydrolysis product of a 2,4-cyclohexadienone diary1 ketal structure such as (3) being formed by dehydrogenative coupling [ 1131. It points to the presence in lignin of a methoxyhydroquinone.

B.

Hydrogenolysis

Previously, the hydrogenolysis of lignin wasstudiedtoproducechemicalsandalsoto obtain structural information. Recently, protolignins have been subjected again to catalytic

YHOH YHOH

R R=OMe R=H

OH Cez, R I = R 2 = b = H or OH R3=H or OMe

OH

0

Me0Q O M e OH

OH

CH2 CH20H

OMe MeO’

0

OMe (ep) R=H or OMe OCH2CH2OH

MeOQ

0

O M e OH

umz, CH2R

YH2OH

I

y

HC-

2

H&”O“CH I

MeO OH OH

OMe OMe

cuL1)R=H or OH

OMe

OH UM) R=H

or OH

OMe

MeO

OH

W

FIGURE 25 Products obtained by mild catalytic hydrogenolysis of protolignins with copper chromium oxide at about 240°C.

140 San0

and

Sakakibara

Me0 OH OH

R3

uQ6)R,=R2=H or OH R3=H or OMe

OMe

Q O M e OH

R=H or CHIOH

oc'o"-

y

m6

HOH27

2

7H2

H?-

CH2

OMe OH

OH

Me0

OC'O,

?H2

H?-

YH

CH2 /

OMe

/

OMe

OH

uu OH

FIGURE 25

W

Continued

hydrogenolysis under conditions at about 240°C with copper chromium oxide, leading to the isolation of various dimeric and trimeric compounds (94-111) besides a substantial amount of monomers [102,103,114-117,1211. The lignols isolated are summarized in Fig. 25. Hydrogenolysis cleaves most of aryl-alkyl ethers, but a few of these linkages remain intact as seen in compounds (95 - R, = OH, 96). Compound (96) - has an a-hydroxyl that

HC-0

141

Chemistry of Lignin R

I CH II CH

CH20H I

l CHOH

HOH27 Q O M e HOH2C HC-0 I CHOH

I

R

$$,

I

HC-

0

OMe

OH R=H or OMe

OH

0 Rl=CHCHCHzOH Rz=H

G OH O M e

R$

OMe

u1z) R,=CHCHCHO

R2=OMe

0 R,=CHzCH?CHzOH Rz=OMe

FIGURE 26

Products obtained by mild hydrolysis of protolignins with aqueous dioxane at 180°C.

142 San0

and

Sakakibara

(m)

(m)

has quite exceptionally remained unaffected. Compounds [ 102,1141 and [ 1031 should give metahemipinic acid (2) by permanganate oxidation of methylated lignins. Compound was isolated from the lignin of larch compression wood, but its occurrence in normalwood lignin isalsoprobable. Itcanbeconsidered that compounds (101,102) could be derived by reductive cleavage from pinoresinol and (E), respectively. However, a model experiment indicated that the alkyl-alkyl ether of the tetrahydrofuran ring is very stable toward hydrogenolysis, and the starting material (g) was almost completely recovered [ 1151. This fact suggests that compounds (E) may not be derived from compound after all, and these linkagepatterns may exist independently in lignin molecules. Compound (g) was isolated from the hydrogenolysis products of hardwood protolignin [ l 161, indicating that the alkyl-alkyl ether is fairly stable against the reductive cleavage. Compound (96)contains a 7-0-4 linkage [ 1171 so far not known. It is, however, very probable that a p-0-4 linkage can be enzymatically rearranged to a y0 - 4 linkage. The existence of 7-0-4 linked units waspostulatedatan early stage by Freudenbeng [ 1191 in 1933 without experimental support, but later he abandoned the idea. Subsequently, in 1960, Brauns et al. [l201 proposed the same substructure. However, this linkage pattern may be a minor one in the lignin macromolecule. Compound (F), which was isolated from hardwood protolignin [ 1211, has a heterocyclein the molecule involving a 4,5-dihydroxy-3-methoxyphenylmoiety. Compounds ( I 10,111) are trilignols with a y-lactone. They have no optical activity. The IR spectra of MWLs show a small shoulder at 1760 cm”, indicating the existence of y-lactones. The facts indicate that p-hydroxy cinnamic acids are also involved in the radical coupling scheme after enzymatic dehydrogenation. Nimz [95] cleaved protolignin from beech wood with thioacetic acid using boron trifluoride as a catalyst followed by reductionwithRaney nickel toproducemanydimericcompounds.Theproductswere somewhat different from those formed by hydrogenolysis.

(m)

(e)

(e)

C.

MildHydrolysis

1. Hydrolysis withWater

Nimz [125-1291 percolated extractive-free wood powder with water at 100°C for several weeks (“mild hydrolysis”), showing that beech wood loses about 40% of lignin, while only 20% of lignin in spruce wood goes into solution. From hydrolysis products of spruce were isolated eight dilignols, two diastereoisomeric trilignols, and one tetralignol: guaiacylglycerol-p-coniferyl ether [ 1261, a trilignol involving p-0-4 and p-1 linkages (123)[127], a tetralignol involving one 6-1 and two p-0-4 linkages (125)[ 1271, guaiacylglycerol-p-guaiacylglycerol ether [ 1261, and guaiacylglycerol-p-coniferylaldehyde (M)[ 1281, and from beech were isolated syringaresinol (g) [98] and three diarylpropanediols (115)[89,123].

(e) (e)

2. Hydrolysis with Dioxane and Water Sakakibara et al. [ 130-1391 found that 40-60% of lignin can be dissolved by treating wood powder with a dioxane and water (1: 1) mixture at 180°C. The many degradation products are almost the same as those obtained by Nimz (Fig. 26). Arylglycerol-p-aryl ether (114) [48], three diarylpropanediols [ 133,134,551,thephenylcoumarans (1 17,122) [76] and [ S ] , syringaresinol(45) [99], adilignol with an a-carbonyl group (g) [loo], the trilignols (122)[135], (123)[133], (124)[100,136], and the C6-C3-C3 lactone (Fig. 27) [ 1371 were isolated and identified, besides considerable amounts

(m)

(m)

(x)

143

Chemistry of Lignin

OH OH

M e o ~ o M e

+

H2?/"CH MeoQoMeI

Me0QOMe OH

OH

0

OH OH M e O A O M e

M e o ~ o M e

CH20H I

'CH

H2qA0"CH II

I

(b) R

6 0

HC$/C" I

CH20H

HCI

I

CH I

CH l

Me0

OMe+ Me0

OMe

OMe

0

OMe

0

/

0

J +H20 OMe H o H q G O H HOHC I

R Q OH O M

OMe e

R=H or OMe FIGURE 27

Proposedformation

(g). (From Ref.

+

Meo6

H2?C ,"H

I

HC-

CH

I

I

O+C-~-CH2 UZL)

mechanism of compound for biogenesis of substructure unit

137.)

of monolignols. A compound (M)with an w-propanol side chain supports that the existence of such reduced sidechains in the lignin structure is probable,considering the highly shielded signals in their NMR spectra, as discussed before 18,191. The formation of cornpound (E), which has eliminated an aromatic ring, may be explained i n two ways: ( a ) ring closure during hydrolysis or (b) displacement of the diarylpropanediol dur-

144

Sakakibara and Sano

+ OH

OH

OMe OMe

(m FIGURE 28 DegradationofcY,P-diarylethers (3 and aqueous dioxane at 180°C for 20 min. (From Ref. 139.)

g) by “mild

hydrolysis”using 50%

ing coupling ofradicals (Fig. 28) formed eitherin wood in situ or secondarly by homolysis of phenolic P-ethers in lignin, as shown in Fig. 29 [137,141]. As described before, the degradation products from hydrogenolysis and hydrolysis provide much important information about lignin structure and supportthe theory of lignin formation by enzymatic dehydrogenation of cinnamyl alcohols. Hydrolysis mainly cleaves the a-ethers of side chains in the lignin molecule in spite of poor model experiments, but nonphenolic P-ether cleavage occurs hardly at all [58,139]. The model experiments with p-hydroxyarylglycerol-a$-diary1 and -&aryl ethers [ 138- 1411 have indicated that homolysis occurs to a slight extent during hydrolysis, resulting in subsequent coupling of the radicals formed. But Sakakibara [l651 has stated that the formation of artificial products from protolignin by “mild hydrolysis” is negligible because of the following fact: Dehydrodiconiferyl alcohol (M)and pinoresinol and dehydrodiguaiacylpropane (106, R,=OCH,, R,=R,=H) that were found in the model experiment [138-1401 could not bedetectedeven in traces in the hydrolysatefromspruceprotolignins in spite of the proposed mechanism of compound shown in Fig. 27. The homolytic degradation of phenolic p-0-4 linkages will be summarized in the following section because of getting many informations on homolytic cleavage of phenolic P-ethers under the conditions similar to those for “mild hydrolysis” and “catalytic hydrogenolysis” of protolignins.

(e)

(m)

3. Homolysis With the aim of finding a method which would degrade lignin without involving simultaneous condensation reactions, Nimz [ 125- 1291 and Sakakibara [ 130- 1391 subjected extractive-freesoftwoodandhardwoodto“mildhydrolysis”under neutral or slightly acidic conditions followed by a percolation with hot water and a mild hydrolysis with

145

Chemistry of Lignin CH20H

CH20H

l

HCI

H A - 0 9

+

QOMe 0.

OMe

Rb

OH

0 Ra

uz&, (24.0%)

CH20H l CH II

m+RaorRb OMe 0-

OH

RC

(1) (2.6%)

OH .OMe

CHsCH-CH20H

CKOH

I

+

2xRa

CH20H H { - 0 9

FH20H j_\

OMe

Ra+Rc

-

OMe

+

H$-0 Q O M e Q O M e

Ra + Rb + RC

-

0 '

w(3.656)

polymerichydrolysisprodUCts of

FIGURE 29 Homolytic degradation of guaiacylglycerol-P-guaiacylether (g) by "mild hydrolysis" using 50% a q ~ ~ e o dioxane us (pH 3.54) at 180°C for 20 min (%: yield of the amount of starting material used). (From Ref. 140.)

146 San0

and

Sakakibara

aqueous dioxane at 1 80”C,respectively. Separately, they have isolated and identified many monolignol-to-trilignol hydrolysis products in small amounts, of which most compounds are identical to each other. The hydrolysis products due to cleavage of the benzyl aryl ether linkages have been thought to give valuable information on the end groups in lignin and the biosynthesis process of lignin, since it is probable that in the “mild hydrolysis” the hydrolytic cleavage of benzyl aryl ether linkages in lignin is the main reaction in spite of deficient experimental evidence. To obtain definitive information for the reaction mechanism of lignin under the conditions of “mild hydrolysis,” some lignin model compounds were subjected to mild hydrolysis. The hydrolysisofmodelcompounds(126,127)forphenolicandnonphenolic a$-diary1 ethers were carried out in 50% aqueous dioxane at 125-180°C for 20- 120 min [138-1411. The phenolic a-aryl ether bond in (126)was cleavaged by 41% and 83% at 140°C for 20 min and 120 min, and by 100% at 180°C for 20 min, and the nonphenolic a-aryl ether bond in by only 39% at 180°C for 20 min [139]. The latter was more resistant to mildhydrolysisthan the former. The nonphenolicP-ethercompoundwas recovered quantitatively from the reaction mixture of (g). New dimeric and trimeric compounds were obtained from the reaction mixture of (3) [138]. Guaiacylglycerol-P-guaiacyl ether was subjected to “mild hydrolysis” by two procedures, that is, 50% aqueous dioxane (pH 3.54) at 180°C for 20 min, and also water (pH 3.54) at 110°C for 48 h in place of the percolation procedures. Thin-layer chromatograms of reaction products obtained from the former were completelyidentical with those from the latter, indicating that the cleavage of the P-ether according to the two procedures proceeds by the same mechanism [140]. From the reaction mixture obtained by the former procedure of the starting material(24.0%),coniferylalcohol (1)(2.6%),pinoresinol (0.8%), 1,2-diguaiacyl1,3-propanediol (E) (1.4%), dehydrodiconiferyl alcohol (4.5%), and two trimeric compounds (1 29,130) were isolated and identified besides substantial amounts of unknown polymerizedmaterials [l401 as illustrated in Fig. 29. The compound (3.8%) was composed of phenylcoumaran and P-ether moieties, whereas (3.6%) had two P-aryl ether links. Gel filtration of the reaction mixtureshowed that their molecularweights increase with increasing reaction time, demonstrating that the degradation and polymerization of lignins take place simultaneously by “mild hydrolysis.” The polymerizationtookplaceduring the heat-up time up to 18O”C, and became predominent for 120 min to form more stabilized polylignols in large quantity. Both the relative absorbance of the “mild hydrolysis” products at A,,, about 280 nm for UV (neutral), and at A,,, about 300 nm for ionization differential spectrum ( A s i ) increase with increasing the reaction time, respectively, when measured immediately after the hydrolysis. In 348 h after the hydrolysis both of them decrease strikingly, and the absorbance at A,,,,,, about 255 nm for UV (neutral) increased reversely compared to every one when measured immediately. The UV spectrum for the reaction products, which was measured 348 h after “mild hydrolysis” of for 120 min, were very similar to that in spruce MWL except for slightly high absorbance at A 360 nm for A&,, due to phenolic conjugated carbonyl groups. The results obtained by UV analyses show that more reactive sites such as quinonemethides are present among the “mild hydrolysis” products of the model compounds, MWL, and wood-in-situ lignin [140]. When 1-guaiacyl-2-guaiacoxy- 1 -propene-3-01 (131) was subjected to mild hydrolysis with 50% aqueous dioxane at 180°C for 20 min, Hibbert’s monomers (3 1-34) were formed in addition to the starting material (15.6%) (Fig. 30). This implies that phenolic arylgly-

(m)

(m)

(e) (m)

(m),

(B)

(m)

(m)

II CH I

CHO OMe Ribkrt’s ketones

QOMe OH

(3.l-U (Total about 16%)

COOH

+ I

OH

(L31) (15.6%)

FIGURE 30 Degradation of I-guaiacyl-2-guaiacoxy- 1-propene (131)by “mild hydrolysis” using 50% aqueous dioxane (pH 3.54) at 180°C for 20 min [%: yield of the amount of (=)I. (From Ref. 140.)

148 Sano

and

Sakakibara

cerol-P-ethers in lignin are never degraded via corresponding enol-ethers under the “mild hydrolysis” conditions [ 1401. From the “mild hydrolysis” products of syringylglycerol-P-syringylether (133)by aqueous dioxane [ 14 l], syringol ( 1 4.5%), syringaldehyde (2)(1.1 YO),sinapyl aldehyde (134)( I . I %), 1-syringyl-2-syringoxy-3-hydroxypropanone-1 (136)(0.9%), D,L-syringaresinol (g) (5.6%), 2-formyl-3-hydroxymethyl- 1,4-bis-syringyl- 1,3-butadiene (3) (1.3%), and 1,2-syringyl- 1,3-propanediol (7.1 %) were isolated besides the starting material (33.3%) and polymeric compounds as shown in Fig. 3 1 . Sinapyl alcohol (2) and syringylglycerol-P-sinapylether which were not detected amongthe products, might be too thermolabile to exist in the reaction mixture, even if they were formed during the reaction process. The thermolabilities of (2,137) are apparent from the fact that they are synthesized only by reduction with LiAIH, at -30°C. The product patterns shown by “mild hydrolysis” of the two @-ether models and 134)by two procedures with aqueous dioxane at 180°C for 20 min and with water at 110°C for 48 h suggest that the reaction imitates the dehydrogenative formation of lignols and higher polymers from coniferyl alcohol and sinapyl alcohol (2). Accordingly, it is proposed that under the “mild hydrolysis” conditions phenolic arylglycerol-@aryl ether bonds may be subjected to homolytic degradation followed by polymerization as illustrated in Figs. 29 and 3 1. Phenolic @-aryl ethers as end groups in lignin lose a molecule of water to give the corresponding quinone-methides (QM), followed by homolytic cleavage of Paryl ether bonds with the formation of a p-hydroxycinnamyl alcohol and phenoxy radicals (Ra and Rb), whereas a radical transfer reaction between the former radical (Ra) and the phenolic hydroxyl groups yields a corresponding p-hydroxylcinnamyl alcohol (1, 2, or 3) and a new phenoxy radical (RC).“Endwise”addition of ap-hydroxycinnamylalcohol radical (Ra) to the latter phenoxy radical (RC)produces the trilignols in Fig. 29 and highermolecular lignols. Nimz et al. found that guaiacylglycerol-P-dihydroconiferylether gave dihydroconiferyl alcohol (41%) and dihydrodehydrodiconiferyl alcohol (phenylcoumaran, 32%), in addition to unknown products, when heated in water for 7 days [142]. Later, guaiacylglycerol-P-vanillyl ether (W)gave a phenylcoumaran coniferyl alcohol (l), vanillyl alcohol (g), and unknown condensed products as precipitates in addition to the starting material (35%) in yields of about 10, 18, 20, and 15%, respectively, when heated with water at 130°C for 4 h, and the yield of condensed products after 20 h at 130°C reachedabout30% in the absenceof starting material(Fig.32).Fromthese results Nimz concluded that a homolytic cleavage of the phenolic P-0-4 linkage occurs with water in neutral solution at 130°C [ 1431. In order to investigate the extent of homolytic reactions under the conditions of “mild hydrolysis,” Westtermark et al. [ 1441 subjected guaiacylglycerol-P-guaiacyl ether (128) and spruce wood to “mild hydrolysis” using dioxane:buffer ( I : 1) at different pH (3.69.6) and various temperatures ( I 30- 180°C). The main products obtained by the heating of spruce wood at 180°C were coniferyl alcohol (L), vanillin and coniferyl aldehyde ( 5 ) in order of peak area. The addition of a catalytic amountof FeZ+considerably increased the yield of (L) without influence on the rate of degradation. The amount of (l)formed was found to be strongly dependent on the temperatures; that is, the amount formed at 130°C was only one-fifth of that at 180°C. Its yield reached a maximum after about 100 millat 130”C, and a set of two maxima after 60 minand 120 min at 160°C as well as 180°C. then slowly decreased, showing that the coniferyl alcohol emerges from at least two different sources, and the formation and decomposition of (l)take place during the ‘‘mild hydrolysis.” The maximum amount of at 180°C exceeds the amount correspond-

(m)

(m),

”

(m

(L)

(m)

(x)

(E)

(m),

(z),

(L)

(L)

149

Chemistry of Lignin 7H20H OMe

HF-oa CHOH OMe

7H2OH OMe -H*O

0

H ? - 0 9 CH OMe

e Me0G O MOMe OH

OMe

7H20H

-6 '?H

+I[.

+M

Me0

0

0

OH

e O Q O M x e O Q O M e 0. OH

Meo60Me M e o ~ o M e

CH20H CH I

2XRa"+

HC-?H

l

I

CH2OH CH Me0O 0O M e

"t

(I4 S%)

Kb

7H2OH CH II

.CH

H+CH HC-o-kH,

Me0G O M e

FIGURE 31 Homolyticdegradation of syringylglycerol-P-syringyl ether (133)by "mild hydrolysis" using 50% aqueous dioxane (pH 3.54) at 180°C for 20 min (%: yield of the amount of (g)]. (From Ref. 141.)

150

Sakakibara and San0

...goMe CH

YH20H CH qH2OH

Ra+Rc

H-C

L 1

Q

CHo OMe

Me0

Ra+Rb

c-c

$H I

-

$oM;o Me0 OH

+ I I,O

Me0

OMe

0

Me0 OH U)(7 1%)

FIGURE 31

Continued

ing to the total content of coniferyl alcohol end groups (about 2%) in softwood lignin. This means that a considerable part of the coniferyl alcohol must originate from sources other than a-ether end group substructures in lignin. The stability of coniferyl alcohol (1) toward “mild hydrolysis” in 50% dioxane (pH 7) at 130°C demonstrated that coniferyl alcohol is degraded or polymerized fairly rapidly even at 130°C. And guaiacylglycerol-Pguaiacyl ether ( 128), which was degraded by 80% in the same medium after 40 min at 1 8O”C, a f f o r d e d 0 at a 20% mole yield. No Hibbert’s ketones (31-34) could be detected in the reaction mixtures even when about 50% of ( 128) was degraded in 50% dioxane at a pH as low as 2.7 and 160”C, indicating that a c i d o 5 i s does not occur by “mild hydrolysis.” Based on the above results, Lundquist et al. [ 1441 concluded that the use of “mild hydrolysis”asa tool in lignin analysis leads to erroneous and confusing results if the homolytic cleavage of the @-ethers is neglected. It is well-known that phenolic P-ethers in lignin are prone to undergo reactions via ionic intermediates under acidic and basic conditions. However, it has been reported recently that the P-etherbond i n syringylglycerol-P-syringylether (133)is subjected to homolysis under conditions similar to those in a soda cook [ 1451. Furthermore, Sipila et al.[l461 allowed syringylglycerol-P-syringylether to stand in aqueousdioxane solutions at pHof 4-7 at roomtemperatureandfoundsyringaresinol (G),P-l (E: R,=R2=OCH,) and 2,6-dimethoxy-p-quinones, giving further evidence for the existence of homolytic P-aryl ether cleavage in delignification of hardwoods. Sano [ 147,1481 subjected extractive-free oak wood meal, guaiacylglycerol-P-guaiacy1 ether and its 4-0-methyl ether to “solvolysis” in p-cresol-water ( 1 : 1 ) at 180°C for 30 min in order to explain the delignification mechanism of wood lignin by solvolysis pulping with aqueous phenol at elevated temperatures. Fromreaction products of the wood, six compounds (143-148) were isolated and identified, as the same cresolated compounds

(m)

(m),

151

Chemistry of Lignin

KMe OMe

+

Q O M e

OH

20 h

OH

OH

+

W (35%)

condensed products (30%)

FIGURE 32 Degradation of phenolic P-ethers (3 and 141)by “mild hydrolysis” with H 2 0 at 100°C or 130°C (%: yield of the amount of (139)or (E)]. (From Ref. 142.)

with guaiacyl and syringyl rings also obtained, respectively, from the reaction products of guaiacylglycerol-P-guaiacyl ether (128)and sinapyl alcohol (2) treated under the same conditions. The reaction mechanism by phenolysis has been illustrated asfollows(Fig. 33). The cleavage of phenolic @-ethers in lignin proceeds via homolysis to form two freeradicals, which are trapped by the phenols as a solvent and/or other active intermediates to give p-cinnamyl alcohols (1-3) from end groups and new phenol groups in lignin. The latter are further cleavaged in reaction sequences of the “peeling” type to lead the extensivedepolymerization of lignin.p-Hydroxycinnamylalcohols (1-3) aredehydrated to extended quinonemethides (E), which are condensed with the phenols by ionic reactions. The resulting resonance-stabilized phenolated compound are oxidized with the phenol radicals and/or lignin radicals to the corresponding radicals, followed by radical coupling. Interestingly, it may be noted that no syringaresinol (G)was detected among the solvolysis products of oak wood in spite of the inactive compound toward the solvolysis reaction [ 147,1481. Lin et al. [ 1501 recently reported that is subjected to homolysis under conditions for the liquefaction of wood with phenol at elevated temperatures. Thus, homolytic cleavages of phenolic P-arylethers in wood lignin can occur at elevated temperatures in many technical processes, for example, high-yield pulping and steam hydrolysis [ 1491, and in thedegradationproceduresfor structural studies under

(m)

(m)

152 San0

and

Sakakibara

coupling Radical coupling Radical Products

Products

W

FIGURE 33 Homolytic degradation of (3) by “solvolysis” using p-creso1:water:acetic acid (9: 1:O.l) at 180°C for 2 h. (From Ref. 147.)

neutral conditions at pH 2-9 and elevated temperatures, for example, mild hydrolysis and hydrogenolysis to yield p-hydroxycinnamyl alcohols and their radicals, and end phenoxy1 radicals in lignin, followed by formation of condensed polymers and chromophores, which influence the brightness and brightness stability of wood and high-yield pulps, and organosolv pulpings. The mild catalytic hydrogenolysis of protolignins has been studied extensively to obtain structural information. Sakakibara et al. [99-1171 subjected protolignins to catalytic hydrogenolysis in 60-90% dioxane containing a catalyst, copper chromium oxide,at 220240°C for 1 h. The hydrogenolysis of protolignins appears to proceed as follows: protolignins are depolymerized and solubilized in aqueous dioxane at 220-240°C higher than those for “mild hydrolysis,” then subjected to catalytic hydrogenation and hydrogenolysis to form “stabilized” hydrogenolysis products. The reaction mechanism for the solubilization and hydrogenolysis of protolignin needs clarification in details to apply the catalytic hydrogenolysis as a meaningful tool in the elucidation of lignin structures. On treatment of 1,l-diphenyl-2-picrylhydrazinewith aqueous dioxane at 180°C for 30 min, the hydrazyl radical is formed in large amounts, but not at below 140°C [ 1381, indicating that phenolic hydroxyl groups may be converted to phenoxy radicals at elevated temperatures.

153

Chemistry of Lignin

OH

-

CH II

OMe

’

O O C H 3 0

FIGURE 33 Continued

D. Acidolysis, Thioacetolysis, and Thioacidolysis

1. Acidolysis with 90% Aqueous Dioxane Containing HCl Since it had been found that refluxing of wood with 90% aqueous dioxane containing 0.2 M HCI, results in the formation of an ether-soluble oil in addition to a high-molecular lignin product, this treatment, “acidolysis,” was subsequently applied both to model compounds and to lignin preparations instead of ethanlysis. Upon 4 h acidolysis of guaiacylglycerol-P-guaiacyl ether (g), the P-ether linkage was cleaved, guaiacol being released, and furthermore, m-hydroxyl-guaiacylacetone (31) could be isolated in a yield of 53%. The latter was slowly further converted, yielding t h e isomeric ketols (3 1,32, I5 1 total yield: 15%), as well as small amounts of ketones in addition to only 3 S % of unchanged starting material as shown in Fig. 20 [IS l]. The acidolytic cleavage of the P-ether linkage in (128)is assumed to proceed via a benzylium ion and an enol ether (E), which is susceptible to acid hydrolysis, followed by formation of monolignols (Hibbert’s ketones 3 1 -34,15 1 ) with carbonyl groups. When spruce Bjokman lignin (MWL) was subjected to the acidolysis treatment under the same conditions, the low-molecular portion of the resulting mixture could be resolved by gel filtration into fractions containing monomeric. dimeric, and oligomeric compounds, respectively. As shown in Fig. 34, in the monolneric fraction the same ketones a s those

(e)

(2,s)

Sakakibara and Sano

154

HCO

l

y

2

H

+

+C,

um

Q O M e

0

I

OH

OH W

1

CH20H I

CH3 I

CH3

I

CH3

CH3

F="

c=o

6. 6 6 6 6 c=o

c=o l

-

1

I

I

I

H OH

OMe OH

OMe

' OMe

OH 0

U

m

0

CH20H

HCO

HS;0

c=o

HC

HC

I

I

66 I

1l

OH

OH

uiz,

0

W

II

OH 0

FIGURE 34 Monomeric products obtained by acidolysis of spruce MWL. (From Ref. 124.)

formed from model compounds (128) were detected, the predominating ketol (31)being obtained in yields of S-6% of the lignin. In addition, the presence of small amounts of homovanillin and formaldehyde (E) was demonstrated. The side reaction can be regarded as areverse Prins reaction of the benzyliumion intermediate. Theseresults constitute clear evidence of the substructures of the guaiacylglycerol-P-aryl ether type. The monomer fraction contained small amounts of ketol coniferyl aldehyde Q), and p-coumaraldehyde (E). In a similar monomeric fraction obtained from the acidolysis of birch MWL, a number of syringyl analogs were detected in addition to most of the compounds shown in Fig. 34 [ 1521. The yields of the syringyl monomers were higher than those of the guaiacylmonomers, although the ratio of syringyVguaiacyl is about 1 : 1 in birch. This is due to the fact that some of the guaiacyl units are linked to an adjacent unit by S - S , p-S, and S-0-4 bonds. which cannot occur in syringyl units. From the dimeric fraction obtainedfromspruce lignin, compounds (45,154- 160) were isolated and identified as summarized in Fig. 35. With the exception of trace constituents ( l S5,160), the dimeric fraction of birch lignin gave the same guaiacyl compounds,

(x)

(m),

"

CHpOH

CHpOH I c=0

I

c=o I

?H

?"

I

qH2

H3C,

+HC OM II.

Ff +OMe

6 GOMe

7-0

OH OMe

QOMe OH

OH

OH

ui8)

OH

(m)

HC-CH3

I

OMe

(rn)

0

'CH2

'

HpC' HCI

Me0

w

UIZ)

AH I

0 OH

Q O M e OH

(m

HC-YH I HpC-, /CH O I

(41,

Me0Q O M e OH

FIGURE 35 Dimeric products obtained by acidolysis of spruce MWL. (From Ref. 124.)

OH u.33

156

Sakakibara and San0

the corresponding syringyl analogs with one or two syringyl nuclei, and, furthermore, the stereoisomeric compounds D,L-syringaresinol ( S ) and D,L-epi-syringaresinol (E). The phenylcoumarone (154)and the stilbene (155)originate from a lignin substructure as shown in Fig. 36. The phenylcoumarone (%), which is formed in much higher yield than the latter, has a characteristic and very strong UV absorption 11531. This permitted its quantitative estimation, which indicated that about 10% of the C, units in the spruce lignin are connected to anadjacent unit byan a-0-4 aswellas a p-5 linkage, giving rise to a phenylcoumaran system. The acidolytic conversion of a hydroxylmethylsubstituted phenykoumdraninto a methyl-substitutedphenylcoumarone is readily explained by a sequence of ring opening, allylic rearrangement, and recyclization. The dimeric compounds (156-159)all exhibit only one side chain per two guaiacyl residues. It was postulated that these compounds could arise from a 1,2-diguaiacyl-1,3propanediol substructure (E)incorporated into lignin by acid-hydrolyzable linkages (Fig. 37). A plausible mechanism for their formation has been presented [ 1541. Compound (E) and related compoundscarryingone or twosyringyl nuclei were isolated from“mild hydrolysis” products of softwood and hardwood, respectively, [89,129,133,134], and (3) was also detected among the low-molecular productsof coniferyl alcohol dehydrogenation. The biogenesis of the lignin substructure (E)can be visualized as shown in Fig. 21. A p-l coupling between a coniferyl alcohol radical and the radical of a p-hydroxybenzyl alcohol end group forms (E)and a glyceraldehyde-2-aryl ether group Experimental evidence for the presence of glyceraldehyde end groups of type ( 4 0 ) was provided by the detection of pyruvaldehyde (150,methylglyoxal) in the acidolysis mixtures from spruce and birch MWL. The mechanism of the acidolytic formation of pyruvaldehyde is presented in Fig. 38. Colorimetric determination of the aldehyde formed on acidolysis of spruce and birch MWL, as well as of model compounds (g), indicated that only about 2% of the C, units of lignin were bound to glyceraldehyde as shown by formula (S). The results of ‘H-NMR studies on the aldehyde groups present in spruce MWL point to a similar value for the amount of glyceraldehyde groups. This figure, however, appears very low in view of the fair yields of degradation products of the p-1 type obtained by thioacetolysis of

(c).

s? r 6

2. a

a

L

Q O M e OH

w

OH (1121

acidolysis

Y O M e OH

w

(From Ref. 154.) 1,3-diolunits (2). FIGURE 37 Acidolysis of 1,2-diarylpropane-

m

I

Sakakibara andSan0

158

c-c-c

FIGURE 38 Acidolysis of glyceraldehyde-2-aryl ether units

(S).(From Ref. 90.)

beech wood. It has been proposed by Sarkanen that the unconjugated carbonyl groups in spruceMWL,whichamount to about 10 carbonylgroupsper 1 0 0 C, units, mightbe regarded as being present in glyceraldehyde groups. This proposal, however, does notfind support in the analytical results mentioned above. In Fig. 39, the substructures which have been disclosed by the acidolysis procedure are summarized. Of these structures, the arylglycerol-P-aryl ether structure undoubtedly is the most abundant one. As already mentioned, C, units which are linked to an adjacent unit by forming a phenylcoumaran system occur in spruce lignin in an amount of about 10%.D,L-syringaresinol ( g )and its stereoisomer D,L-episyringaresinol have also been found among the acidolysis products from birch. The corresponding guaiacyl compounds (S), however, could not be found in spruce or birch acidolysis mixtures. If pinoresinol (44) structures are present in lignin in

(m)

t; C,

C

c:

0’

C

I G O M e H2COH HC-0

MeoQoMe H,C*O-?H

Chemistry of Lignin

159

appreciable amounts, one must assume that they are linked to adjacent units by acid-stable bonds, i,e, 5-5 and 5-0-4 bonds, to an unexpectedly great extent. The "NMR spectrum of spruce lignin also indicates a very low content of pinoresinol structures. However, experimental evidence in favor of the occurrence of such structures in conifer lignin has been presented. A further acidolysis product exhibiting coupling, namely, D,L-divanillyltetrahydrofuran (E), was isolated in small amounts from spruce lignin acidolysis. The compound differs from the other dimeric acidolysis products in possessing a lower degree of oxidation. Degradation of beech wood by thioacidolysis afforded the syringyl analog. Its formation seems to involve an oxido-reduction process, but it remains open what substructures (@) in lignins originated from [154].

(e)

2. Thioacetolysis and Thioacidolysis Thioacidolysis causes cleavage of P-0-4 bonds, and brings a more deep-ground fragmentation of the lignin than acidolysis [95,155]. As much as 91% of the lignin of beech wood and 77% of the lignin of spruce wood were degraded to mixtures of monomeric to tetrameric products. The principle of the three-step degradation method has been formulated by Nimz as shown in Fig. 40. Treatment of wood with thioacetic acid and boron trifluoride converts the arylglycerol-P-aryl ether unit (40.1) via the benzylium ion ( 4 0 . 2 ) into the Sbenzyl thioacetate (=). Subsequent saponification with 2 N NaOH at 60°C gives a benzyl thiolate ion (40.4) which loses the P-aryloxy group by nucleophilic attack of the neighboring thiolate ion on the P-carbon atom to give an episulfide The latter dimerizes to dithianes or polymerizes to thioethers. In a final step, treatment with Raney nickel and alkali at 115°C removes the sulfur and yields the reduced phenolic reaction products. The 20 dimers obtained from beech wood are shown in Fig. 41. Most of the bond typesexhibited by these dimers are identical withthoserevealed by otherdegradation

(a).

cn,on l

I

I HCOAr 13 F,

R'

CH20COCHs

CHPOH

HCOAr Cl I$X)SI I

NaOl I

OMe

R'

OR (40.2)

OR (4.3)

CHZOH

CH2R

l

I

HqSo

R'

OMe OH

OH R=H or OMe K'=llorOlI (40.7)

FIGURE 40 Degradation of lignin by thioacetolysis with thioacetic acid.

(From Ref. 95.)

160

Sakakibara and San0

&Q'16 OH

R

OMe OH

H $ oOH O M e

R OMe / OMe OH

R=H or OMe

OH R=H or OMe

CH3

CH3

HC

I

Me0

OMeMeO OMe OH OH

OH

R

OMe Me0

OMe

OH

\

R=H or OMe , R'=H2or =O

/

u6i, (0.5%)

CH3

I

CH3

I

C

Me0

OMe

R

OHOH R=H R=OMe or OH

c166)(0.45%)

FIGURE41 95.)

R=OCH3.

R=OMe (142) (0.4%)

OH

OH R=H or OH

(0.1%)

0 (0.3%)

Dilignols obtained from beech protolignin by thioacetolysis (% of lignin). (From Ref.

Chemistry of Lignin

161

methods, especially acidolysis and oxidative degradation cited above. The a-p-junction (165) found in some compounds (0.5%) was assumed by Nimz and Das [95] to be present i n e e c h lignin, although it cannot be the result of a dehydrogenation. Its formation and that of some related structures is assumed to be due to a proton-catalyzed polymerization of coniferyl alcohol (1)or coniferyl alcohol end groups caused by the natural acidity of the cell sap. However, we must also point to the possibility mentioned by Nimz [ 1801 that (1)andsinapylalcohol (2) may arise during the degradation process, analogous to the formation of (L) as an intermediate in kraft cooking of spruce lignin. The heating with alkali in the last degradation step may cause Michael-type additions of the p-C atom of coniferyl alcohol to the a-position of quinonemethide structures whichalsocanbe assumed to be intermediates. The a-p-linked products may accordingly be artifacts. On the basis of the yields of crude and pure degradation products, Nimz has calculated the frequencies of the various bond types in beech lignin and has also proposed a structural scheme for this lignin (Fig. 46). Recently,a new acid degradationmethod, thioacidolysis (solvolysis in dioxaneethanethiol with boron trifluoride etherate) has been studied by means of a reproducible and mild routine procedure to obtain detailed structural information about lignin [ 1561591. The acid degradation of lignin was composed of two consecutive thioacidolysis and desulfuration of thioacidolysis products over Raney nickel as illustrated in Fig. 42, which is similar in principle to that proposed for thioacetolysis (Fig. 40). However, thioacidolysis is carried out using a few milligrams of sample in dioxane at 100°C for 4 h instead of thioacetolysis performed at 20°C for 1 week. Reaction conditions for the former seem to be more advantageous as a tool for structural studies of lignin than those for the latter. The thioethylated monolignols and the desulfurated dilignols of spruce MWL and wood are shown in Fig. 38 [1591. Among the thioethylated monomers from spruce MWL, the compounds (170,171), whichwereformedfromuncondensed p - 0 - 4 linked units, were obtained in a total yield of 9 3 % of the monolignols. They reflect the higher content of uncondensed p - 0 - 4 linkages in the MWL products and from coniferyl alcohol end groups were obtained in a total yield of 6.6% based on the monomers. The compounds and orginating from coniferyl aldehyde end groups and from dihydroconiferyl alcoholendgroups, respectively, weredetected in trace amounts.The yields of main dilignols obtained from spruce MWL and wood are shown in Fig. 43 and characterize the various types of condensed linkages in softwood lignin. From their yields, it can be concluded that p-S, 5-5, and p-1 linkages are present as major types of condensed interunit linkages in softwood lignin. In addition, the dimerswithdiphenylether (181)(4-0-S), phenylisocoumaran (p-5), and tetrahydrofuran (M)(p-p) structures, of which the structures were assigned only from their fragmentation patterns, were detected in trace amounts, and their linkages seem to be minor types among condensed linkages. However, the latter two dilignols have been assigned to be compounds with different structures in two papers reported by the same authors [ 159.1, indicating that they need to be isolated and identified. The total area of GC peaks due to the dilignols assigned accounts for more than 90% of all the peaks corresponding to dilignols in the chromatograms of spruce MWL and spruce protolignin. The total amount of the dilignols is about 30 mole% of the main thioacidolysis monomers obtained from both of the lignins. The relative importance of condensed interunit linkagesimplied by the mole ratio of the thioacidolysis dilignols is almostequal between spruce MWL and sprucewood lignin in situ. The yields of compounds which were characterized as monolignolsanddilignols among the thioacidolysis mixture were only 40-S0% of lignin [ 1593, which may reflect the limitation of thioacidolysis results to the degradable part of lignin.

(m)

(m) (m)

(m) (m),

162 San0

and

Sakakibara

H2COH HCSEt

I l

H2yH HCORp

HCOR2 HCOR2

I

@,R1

F 6

BF3 E(SH

_ I )

OMe

OR

H2COH

OMe9F3

H2COH

OR

I

I

COR2

H27OH

OH

HpCOH

-

I

I

6z; 6E":;61 OR

It

OH

!$p

I I

'OMe OR

H

OR

OR

OR

OH

OH

FIGURE 42 Reaction mechanism of p-0-4 substructure units by thioacidolysis and subsequently desulfuration with Raney nickel. (From Ref. 156.)

In order to use the interesting thioacidolysis method as a routine procedure for the characterization of the total structure of lignins, it is necessary that the thioacidolysis productscontaining trilignols to oligolignols be clearly characterizedandacid-derived condensation by thioacidolysis clarified.

V.

STRUCTURAL MODELS FOR LIGNINS

A.

Frequencies of Functional Groups and Typical Linkage Types in Lignins

Summarized in Fig. 44 are the main types of lignin structural units which were obtained from the various lignin degradation products described above. Frequencies of the functional groups and the typical linkage units in spruce and birchMWL, and beech protolignin are collected in Tables 10 and 11, respectively. The most important linkage types in the lignin molecule are p-0-4(B) and then p-5 (g),5-5 (E), p-1 (C), and a-0-4 (A). The pp linked units are represented by pinoresinol (g), syringaresinol (g), and dibenzyltetrahydrofurans F(c) (160 and Pinoresinol substructure (@-p)may be rather minor in softwood lignin, but syringaresinol (45) - is abundant in hardwood lignin. Though the com-

x).

Me

dsEt do"

Chemistry of Lignin

&Et

OMe

Et$

SEt Et$EtS

163

SEt

OMe

OH OH

OH

0

0

OH OH

OH

uze,

0

uz2,

W

R

Me0

OMe OH

5-5 series

p1 series

4-0-3 series

Meor i

H0

Me0

H O H z C e Me

R2

\

p5OH \ / series OMe OH

OMe HOH2CfMe p3OH \ senes

p5 series

CH3 R=H.

p5 senes

or CH20H

u8L,

OMe

OH

p p series

Rl=CH3 or CH20H R2=H 01 CH3

u&+M=33%)

FIGURE 43 ThioethylatedmonomersandRaneynickel-desulfurateddimersobtainedbythioacidolysis of spruce MWL (% of dimers). (From Ref. 158.)

(M)

pound was obtained as one of the important substructures next to A, B, C, D, and E typesfromacidolysisand thioacidolysis products of spruce MWL, it remainsopen whether is a substructure in lignin or acid-condensation products of (101:R=OH). P-Aryloxyglyceraldehyde units B(a) have been estimated by acidolysis and 'H-NMR. 40-5-Diphenyl ether and 0-6 units have been confirmed by not only oxidation but also catalytic hydrogenolysis. By the latter, frequencies of bond types has not been estimated. The yields of oxidation products were used in the estimation of the bond types in the MWLs, though certain assumptions, e.g., regarding the actual and theoretical yields of oxidation products, are involved [ 1611. Naturally, the frequency values in Tables 10 and

(m)

164

Sakakibara and San0

F

8-Q

E - 0 - 0 I

6" 6

6

C

A

B

"8 F'

go

F F

C-

FC

l

I

F

E

F

I

?

C I

D

F C?-

C

FC

F

cI

$

I

C

G

H

0 G(traces)

F F

FC

6-6 I

I

FIGURE 44 Typical linkageunitsin

lignin.

11 are to be regarded as approximate rather than accurate, although mostofthem are fairly reproducible. Some uncertainty is attached to the values (0.02-0.15) given for bond type C in spruce MWL in spite of differing estimation methods. Sarkanen [ 1721 has stated that p-1 units are main substructures in endwise lignin for the middlelamellaregion rather than in bulk lignin for the cell wall. However, Lapierre et al. [l581 have reported that p-

Chemistry of Lignin

165

TABLE 10 Functional Group and Structural Unit of Spruce

MWL

Functional Groups units

Functional groups and structural Aliphatic OH 168,701 Phenolic OH [82,701 Total carbonyl

0.93 0.33

c,,=o

Unconjugated C=O Ar-CH=CH=CHO Ar-CH=CH-CH20H Phenolic C,-OH Nonphenolic C,-OH

C&

References

1.09, 0.26, 0.20 0.06-0.07 0.10 0.03-0.04

1591 [S91 1591 1411

0.03

1441

0.05-0.06 0.15, 0.10

176.781 [59,781

~~

A: a-0-4 (open)

Phenolic Nonphenolic B: p - 0 - 4

c: p-5 Noncyclic D: p-l E: 5-5, 5-6 F: p-B

Pinoresinol units G: 4-0-5, 4-0-1

0.12, 0.07, 0.06-0.08 0.04, 0.02 0.05-0.09, 0.06 0.49-0.5 1, 0.50 (0.25-0.30, 0.3-0.5)'' 0.02b 0.14, 0.9-0.12 0.03 0.15, 0.02, 0.07 0.19-0.22, 0.10-0.1 1 0.13 0.05- 1 .O, 0.02-0.03 0.07-0.08, traces

"Except displaced side-chain unlts. "Arylglyceraldehyde-P-arylether. Source: Ref. 167.

1 units are almost as frequent as in spruce in-situ lignin and basis of the results obtained by thioacidolysis.

MWL preparations on the

B. Structural Models for Softwood Freudenberg [ 1631 attempted to constructa structural formula for softwood lignin, utilizing theknowledgeobtainedfromtheenzymaticdehydrogenation of coniferylalcohol. The formula, which was composed of 18 units, was later modified several times 1164- 1661. Adler [ 1671 has collected theprominent substructures of spruce ligninin a structural model comprising 16 C& units. Glasser [ 1681 has proposed a structuralmodelbasedon 81 phenylpropane units that was constructed by computer simulation. A structural model of softwood lignin consisting of 28 units was proposed by Sakakibara [ 1691 as shown in Fig. 45. Alternative units are indicated in brackets against letter (b). The linkage patterns between thephenylpropaneunitsarebased mainly on results obtained by hydrolysis and hydrogenolysis. The formula of the models for spruce lignin are calculated as C,H,,,~,O2,,,(OCH,),,,,, which agrees well with those of spruce MWL shown in Table 4. The formula is a tentative one and is constructed only from information obtained so far, omittingunits,theexistence of which isat present uncertain.Onlypart of theactual

and Sano

166 TABLE 11

StructuralUnits(per

1 0 0 C&, units) of Birch MWL and Beech Lignin

Beech [95]

Birch [ 1671 Units A: a-0-4 (open) B: p - 0 - 4 34-39

Total

G

S

6 60 2 7

22-28

p-0-4"

c: p-l D: p-5 E: 5-5 F: P-p

Total

6 4.5

2.3

4.5 3

P-B p-P and L Y - ~ ~ G: 4-0-5, 4-0-1 H: C(a)-2, C(a)-6 I: a-P

1 1-1.5

5.5 0.5- 1

6.5 1 .5-2.5

65b

15 6

5 2 0.5 1.5

2.5

"In glyceraldehyde-P-aryl ether. hA + B. 'In dibenzyltetrahydrofuran units. 'In tetralin units.

number of these units has been arbitrarily selected because of the lack of adequate quantitative data.

C. Structural Models for Hardwood Lignin Nimz [l701 proposed a constitutional scheme for beech lignin on the basis of the results from mild hydrolysis and thioacetic acid degradation of beech lignin (Fig. 46). This structural model consists of 25 phenylpropane units containing 14 guaiacyl, 10 syringyl, and one p-hydroxyphenyl moiety, of which six units can to some extent be replaced by the dilignol units enclosed in the brackets. The models give a representative section from a beech lignin molecule 10-20 times larger, in which the 10 different bond types are randomly distributed. Glyceraldehyde-2-aryl ether units B (a) are not detected on degradation of beech lignin, but their presence is evidenced as a counterpart by the occurrence of p1 dilignol units (C) as shown in Fig. 21. The formula of this structural model is calculated as C,H,,,,O,,,(OCH,),,,,, which is close to the formula shown in Fig. 4. Furthermore, the I3CNMR spectrum calculated for the proposed structure was compared with that observed for beech lignin.

D. Heterogeneity of Protolignin These structural modelsareonlyaverage pictures for lignin structures that may have different chemical configurations in the different morphological regions of the cell wall. Fergus and Goring [ 17 I ] indicated varying amounts ofsyringyl- and guaiacylpropaneunits in the various cell wall layers and middle lamella of birch wood by means of UV spectrophotometry. Matsukura et al. [ 1721 showed by oxidation and alcoholysis of spruce wood that the lignin polymer has a heterogeneous structure. Furthermore, it has been pointed out that lignin in the middle lamella region may possess more of the nature of an endwise

167

Chemistry of Lignin

H$?-0“

bH W O M e

FHOH FHOH CH20HI

FIGURE 45 A structure model for softwood lignin.

CH20H

168

Sakakibara and San0

7HpOH HV-

-0

OMe Me0 MeO MeO

MeO

OMe

OMe

Me0 OMe Me0 MeO

OMe -0

FIGURE 46

0-

-0

OH

A structure model for beechwood lignin. (From Ref. 170.)

polymer than that permeating the polysaccharide matrix in the S2 layer [ 1731. Compression wood contains not only more lignin but also more condensed-type units than normal wood lignin [21]. The reason may be explained by the fact that the outer S, layer of compression wood is highly lignified, whereas the middle lamella is not completely lignified [ 1741. The controversysurroundinglignin-carbohydratecomplexes(LCC)hasbeendebated for a long time. In spite of numerous studies, the question of whether the association between lignin andcarbohydrates is physicalorchemical in naturehas not yetbeen resolved. However, the evidence for chemicalbondsbetweenthemhasbeengrowing gradually. For instance, enzymatic hydrolysis of LCC preparations gives concentrated LCC

Chemistry of Lignin

169

fractions that indicate the existence of covalent bonds between lignin and carbohydrates, because enzymes do not cleave lignin-carbohydrate linkages [ 175- 1801. The structure of protolignin must take into consideration the presence of LCC linkages. Some typical forms of LCC linkages have been suggestedby Freudenberg et al., who isolated phenylpropanecane sugar compounds from the mixture arising from the simultaneous enzymatic dehydrogenation of coniferyl alcohol and cane sugar [ 181, l 821. Finally, a word should be mentioned about Brauns lignin (BL), which Brauns [ 1831 first isolated from black spruce by extraction with ethanol in 1939. Since then, BL has been consideredto be natural lignin and has been used for several basic studies of lignin. However, it is not clear whether BL is a true lignin or not, as Freudenberg [ 1841 pointed out that BL may be a fraction of resinous material. Recently, various new lignans have been isolated [ 1601. They consist of dimeric, trimeric, and tetrameric phenylpropanes that are very similar to the lignols from degradation product mixtures, as already mentioned, except for optical activity and some other details. The existence of monomeric, oligomeric, and polymeric phenylpropanes in the lignan fraction suggests that these constitute a continuous spectrum of lignans. In conclusion, BL can be considered a polymeric fraction of lignans and not a true lignin.

VI.

OUTLOOK

The concept of lignin as dehydrogenation polymer of p-hydroxycinnamylalcohols is now well established. The efforts to clarity the structures of the different types of lignin have resulted in a detailed picture of the various modes in which the C& units are linked together in the lignin polymer. Whereas there is good agreement regarding the frequency of the predominant type of linkage, that is, arylglycerol-p-aryl ether substructures, there is even now some uncertainty regarding the proportions of some linkage units, such as noncyclic cy-aryl ether (A), p-! (C), @-p(E), andothersoccurring in minoramounts. “Mild hydrolysis” and “catalytic hydrogenolysis,” which has been used as the conventional degradation methods to characterize the structures of protolignin, appear to lead to erroneous and confusing results by homolytic cleavage of phenolic @-aryl ether linkages and subsequent secondaryradical couplings, so the elucidation of their reaction mechanism of protolignin will be required. Acid-catalyzed degradation, such as acidolysis and thioacidolysis, which give rise to self-condensation of lignin, will deviate to a certain degree froma useful tool toanalyze total linkages of phenylpropane units in lignin. We will require much continued effort to analyze total linkage units in at least MWL and also unchanged lignin in wood. Lignin, which is the most abundant natural polymer next to cellulose, is produced toabout 80 milliontonsandburnedtorecoverkraft-pulpingchemicalsandtomake energies for pulping and papermaking. Regenerated wood biomass will have to be applied to the saving andor substitution of petroleum oil for both energy and chemicals in the future for the sustainable development of the world without environmental pollution. Although wood cellulose is the most predominant pulp material, it will be utilized together with other polysaccharides and old paper as raw materials for about 95% of petrochemicals. The polysaccharides may be converted easily to various chemicals, but the immense problem of finding industrial applications for lignin remains a great challenge to wood chemists. Continued efforts will be required to finish the conversion of wood biomass, that is, wood biomass is separated into pulp, hemicellulosic sugars, lignin, and extractives

170 San0

by a novel process with less amounts wood chemicals of high value.

and

of energy and environmental

Sakakibara

pollution to use as

REFERENCES T. Terashima, in Chemistry of Lignin (J. Nakano, ed.), Yumi Publishers, Tokyo, p. 70 (1978) (in Japanese). 2.G.Aulin-Erdtman, Svensk Pupperstidn., 55:745-749(1952). 3.G.Aulin-Erdtman, SvenslcPapperstidn., 56:91-101(1953). 4.G.Aulin-ErdtmanandL.Heghom, Svensk Puppersridn., 60:671-681(1957). 5. G. Aulin-Erdtman and R. Sanden, Acra Chem. Scand., 22: 1187- 1209 (1968). 6. J. C.Pew, J . Org. Chem., 28:1048-1054(1963). 7.J.C.Pew, Nature. 193:250-252(1962). 8. 0. Goldschmid, Anal. Chem., 26:1421-1423(1954). 9. E. Adler and K. Lundquist, Actu Chem. Scand., 17: 13-26 (1963). 10.A.G.SchdningandC.Johansson, SvenskPupperstidn., 68:607-613(1965). 11.G.Wegener,M.Przyklenk,andD.Fengel, Holdorsch., 37303-307 (1983). 12. H. L. Hergert, in Lignins (K. V. Sarkanen and C. H. Ludwid, eds.), Wiley-Interscience, New York, pp. 267-297 (1971). 13. S. Kolboeand 0.Ellefsen, Tuppi, 45:163-166(1962). 14.K.V. Sarkanen,H.-M.Chang,andB.Ericsson, Tuppi, 50572-575 (1967). 15. V. Sarkanen, H.-M. Chang, and G. G. Allan, Tuppi, 50583-587 (1967). 16. K. V. Sarkanen, H.-M. Chang, and G.G. Allan, Tappi, 50587-590 (1967). 17. C . H. Ludwig, B. J. Nist, and J. L. McCarthy, J. Am. Chem. Soc., 80:1186-l196 (1964). 18. C. H. Ludwig, B. J. Nist, and J. L. McCarthy, J. Am. Chem. Soc., 8 0 1 196- 1202 ( 1964). 19. B. L.Lenz, Tuppi, 51511-519(1968). 20. D. E.Blandand S. Sternhell, Ausrral. J . Chern., 18:401-410(1965). 21.N.MorohoshiandA.Sakakibara, Mokuzui Gakkaishi, 17:393-399(1971). 22. K. Horisaki, K. Shimatani, Y. Sano, and T. Sasaya, in 37th Lignin Symp., Kyoto, Japan, pp. 125- 128 (1992). 23. K. Lundquistand T. Olsson, Acrn Chem. Scond., B31:788-792(1977). 24. K. Lundquist, Actu Chem. Scand., B33:27-30(1979). 25.K.Lundquist, Actu Chem. Scand., B34:21-26(1980). 26. K. Lundquist, A. Paterson, and L. Ramsey, Acta Chem. Scand., B37:734-736(1983). 27. S. LiandK.Lundquist, Nord. Pulp Paper Res. J., 9:191-195 (1994). 28. R. M.Ede,G.Brunow,L.K.Simola,and J. Lemmetyinen, Holdorsch., 44:95-101(1990). 29.H.-D.LtidemannandH.Nimz, Mulcrornol. Chem., 175:2409-2422 (1974). 30.H.Nimz, I. Mogharab,andH.-D.Ltidemann, Makromol.Chern., 175:2563-2575(1974). 31. H.NimzandH.-D.Ltidemann, Hol$orsch., 30:33-40(1976). 32.D.GaganireandD.Robert, Makromol.Chem., 178:1477-1495(1977). 33. J. R. ObstandJ.Ralph, Holdorsch., 37297-302 (1983). 34.C.Lapien-eandB.Monties, Holdorsch., 39:367-368(1985). 35.A.Bjdrkmanand B . Person, Svensk Pupperstidn., 60:158-169(1957). 36. K. Freudenberg and A. C. Neish, in Molecular Biochemistry and Biophysics, Vol. 2, SpringerVerlag,Berlin-Heidelberg,p.113(1968). 37. D. Fengel,G.Wegener,and J. Feckel, Holdorsch., 3551-57 (1981). 38.H.Nimz, Holgorsch., 20:105-109(1966). 39.A.Sakakibara, Rec. Adv. Phytochem., 11:117-139(1977). 40. E. Adler,K.J.Bjorkvist,and S. Haggroth, Acta Chem. Scund., 293-94 (1948). 41. E. AdlerandL. R. Ellmer, Acra Chem. Scand., 2339-840 (1948). 42. E. Adlerand J. Marton, ActaChem. Scand., 15:357-369(1961). 43. J. Martonand E. Adler, Acra Chem. Scand., 15:370-383(1961). 1.

Chemistry of Lignin 44. 45. 46. 47. 48. 49.

50. 5I. 52. 53. 54.

55. 56. 57.

58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79.

80. 81. 82. 83. 84. 85. 86. 87.

88. 89. 90. 91. 92.

171

B. 0. Lindgren and H. Mikawa, Acta Chem.Scand., 11:826-835 (1957). T. Higuchi, Y. Ito, and I. Kawamura, Phytochemistry, 6:875-881 (1967). H.A. Stafford, PlantPhysiol., 37643-649 (1962). Y. San0 and A. Sakakibara, Mokuzai Gakkaishi, 16:81-86 (1970). S. Omori and A. Sakakibara, MokuzaiGakkaishi, 20:388-395 (1974). H.Nimz, Chern. Ber., 98:3153-3159 (1965). A.Yamaguchi, MokuzaiGakkaishi, 19:185-193 (1973). A. Sakakibara, H. Takeyama, and N. Morohoshi, Holdorsch., 20:45-47 (1966). T. Higuchi, H. Nakatsubo, and Y. Ikeda, Holdorsch., 28: 189-192 (1974). E. Adler and S. Yllner, Acta Chem. Scand., 7:570 (1953). K. Lundquist and R. Lungren, Acta Chem. Scand., 26:2005-2023(1972). M. Aoyama and A. Sakakibara, Mokuzai Gakkaishi, 24:422-423 (1978). T. Popoff and 0. Theander, Phytochemistry, 142065-2066 (1975). J. Gierer and S. S6derberg. Acta Chem. Scand., 13:127-137 (1959). J. Marton, E. Adler, and K. I. Person, Acta Chem. Scand., 15:384-392 (1961). E. Adler and J. Marton, Acta Chem. Scand., 13:75-96 (1959). Y. Hachihama, K. Nira, and T. Kyogoku, Kogyo Kagaku Zasshi, 47209-215 (1944). H. Mikawa, K. Sato,C.Takasaki,and K. Ebisawa, Bull.Chem. Soc. Japan, 28:653-660 ( I 955). K. Freudenberg and K. Dall, Naturwiss., 22:606-607 (1955). K.V. Sarkanen and C. Schuerch, Anal. Chem., 2 7 1245-1250 (1955). K. Freudenberg,, in Fortschritte d. Chem. org. Naturstoffe, XI, Zechrneister, Springer-Verlag, Berlin-Heidelberg, pp. 43-82 (1954). E. Adler and S. Hernestam, Acta Chem. Scand., 19:319-334 (1955). A. Bjorkmanand B. Person, Svensk Papperstidn., 60:285-292 (1957). A. Okay, Hul~ursch.,24:172-175 (1970). D. R. Robert and G. Brunow, Holdorsch., 3835-90 (1984). H.-M. Change, E. B. Cowling, and W. Brown, Holdorsch., 29:153-159 (1975). P. Miinsson, Hol$orsch., 37143-146 (1983). D. S. Argyropoulos, J. WoodChem.Technol., 14:65-82 (1994). 0. Faix, C. Grunwiind, and 0. Beinhoff, Hol$orsch., 46:425-432 (1992). 0. Faix, D. S. Argyropoulos, R. Danielle, and N. Vincent, Holdorsch., 48:387-394 (1994). E. Adler and K. Lundquist, Acta Chem. Scand., 15:223-224 (1961). K. Freudenberg, C.-L. Chen, J. M. Harkin, H. Nimz, and H. Renner, Chem. Commun., 224 ( 1965). M. Aoyama and A. Sakakibara, Mokuzai Gakkaishi, 22591-592 (1976). E. Adler, S. Delin, and K. Lundquist, Acta Chem. Scand., 13:2149-2150 (1959). E. Adler and J. Gierer, Acta Chem. Scand., 9534-93 (1955). J. Gierer, Acta Chern. Scand., 8:1319-1331 (1954). J.Gierer, Chem.Ber., 80257-262 (1956). E. Adler, H.-D. Becker, T. Ishihara, and A. Stamvik, Holdorsch., 20:3-11 (1966). H.-D. Becker and E. Adler, Acta Chem. Scand., 15:218-219 (1961). T. Higuchi, M. Tanahashi, and F. Nakatsubo, Wood Res. (Kyoto), 54:9-18 (1972). K. Freudenberg, J. M. Harkin, and H. K. Werner, Chem. Ber, 97:909-920 (1964). E. Adler, G. E. Miksche, and B. Johanson, Holtforsch., 22:171-174 (1968). E. Adler, J. M. Pepper, and E. Eriksoo, Ind. Eng. Chem, 49: 1391-1392 (1957). S. Larsson and G. E. Miksche, Acta Chem. Scand., 25:647-662 (1971). M. Erickson, S. Larsson, and G. E. Miksche, Acta Chem. Scand., 27:903-914 (1973). H.Nimz, Chem. Ber, 98:3160-3164 (1965). K.Lundquist, G. E. Miksche, and I. Berndtson, Tetrahedron. Lett., 46:4587-4591 (1967). G. Aulin-Erdtman and L. Haghom, Svensk Papperstidn., 61:187-210 (1958). C.-L. Chen,InauguralDissertation,UniversitatHeidelberg,Heidelberg,Germany, pp. 1-75 ( 1962).

172 San0

93. 94. 95. 96. 97. 98. 99. 100. 101.

102. 103.

104. 105.

106. 107. 108.

and

Sakakibara

K. FreudenbergandC.-LChen, Chem. Ber., 100:3683-3688(1967). S. Larsson and G. E. Miksche, Acru Chem Scund., 25:673-679 (1971). H. Nimzand K. Das, Chem. Bes, 1042359-2380(1971). S. Yasuda and A. Sakakibara, MokuzaiGukknishi, 23:383-387 (1977). K. Freudenbergand D. Rasenak, Clzem. Res, 86:755-758 (1953). H. Nimz and H. Gaber, Chem. Bes. 98:538-539(1965). S. Omori and A. Sakakibara, MokuzaiGakkuishi, 17464-467 (1971). S. Omori and A. Sakakibara, Mokuzni Gakkaishi, 21:170-176 (1975). K. Ogiyama and T. Kondo, Mokuzai Gnkknishi. f4:416-420 ( 1968). K. Sudoand A. Sakakibara, MokuzaiGakknishi, 20:396-401(1974). S. Yasuda and A. Sakakibara, MokuzniGukkaishi, 22:606-612(1976). K. Freudenberg, Z. Angew. Chem., 52362-366(1939). B. Leopold, Acrn Chetn. Scnnd., 6:38-48 (1952). B. Leopold and J. L. Malmstrom, Actcl Chem. Scund., S:936-940 (1951). J. C.Pew, J . Am. Chern. Soc.. 77:2831-2833(1955). D. L. Brink, Y. T. Wu, H. P. Naveau, J. G. Bicho, and M. M. Merriman, nippi, 55:719-721

( 1972). 109. H. P. Naveau, Y. T. Wu, D. L.Brink, M.M. Merriman,and J. G. Bicho, 72ppi. SS:13561361 (1972). 110. S. Larsson and G. E. Miksche,Acta Chern. Scand., 2/:1970-1971 (1967). I 1 1. M. Eriksson, S. Larsson, and G. E. Miksche, Actu Chem. Scancl., 2 7 127-140 (1973). 112. S. Larsson and G. E. Miksche, Actu Chern. Scand., 233337-3351(1969). 113. S. Larsson and G. E. Miksche, Actu Chem. Scatzd., 26:20-31 (1972). 114. S. Yasuda and A. Sakakibara, Mokuzui Gakkuishi, 23: 114-117 (1977). 1 15. M. Matsukura and A. Sakakibara, Mokuzni Gakknishi, 19:17 1-176 (1973). 116. K. Sudo. B. H. Hwang, and A. Sakakibara, MokuzaiGakkaishi. 24424-425 (1978). 117. B. H. Hwang, A. Sakakibara, and M. Miki, Hol7forsch., 35:229-232 (1981). 118. K. Miki, V. Renganathan, and M. H. Gold, Uiochenzistt-y, 25:4790-4796 (1986). 119. K. Freudenberg, in Rrnnin, Cellulose, Lignin, Springer-Verlag,Berlin, p. 133 (1933). 120. E. E. Brauns and D. A. Brauns, in The Chemistry of Lignin, Academic Press, New York, p. 626 ( 1960). 121. B. H. Hwang and A. Sakakibara, Hol
Chemistry of Lignin

173

143. H. Nimz, U. Tschirner, and M. Roth, in 1983 Int. Symp. on Wood and Pulping Chernistv, Tsukuba, Japan, Vol. I , p. 90 ( I 983). 144. U. Westermark, B. Samuelsson, and K. Lundquist, Res. Chem. Intermed., 21:343-352 (1995). 145. K. M. J. Barrow, R. E. Ede,and I. D. Suckling, in 7th Int. Syntp. on Wood and Pulping Chemistry, Vol. I , pp. 73-81 (1993). 146. J . Sipil,G. Brunow, and P. Tunninen, in 7th Int. Symp. on Wood ctnd Pulping Chemistry, Vol. 3, p. 12 ( 1 993). 147. Y. Sano and A. Sakakibara, Mokuzai Gakkaishi, 31: 109-1 I8 (1984). 148. Y. Sano, Mokuzai Gakkaishi, 35:813-819(1989). 149. M.Bardet, D. Robert,and K. Lundquist, Svensk Papperstdn., 6:61 (1985). 150. L. Lin, M. Yoshioka, Y. Yao, and N. Shiraishi, in Abstr. 46th Annual Meeting of the Japan Wood Research Society, Kumamoto, Japan, p. 278 (1996). l 5 l. K. Lundquist and R. Lundtgren, Acta Chern. Scand., 26:2005-2023 (1972). 152. K. Lundquist, Acta Chern. Scand., 27:2597(1973). 153. E. Adler and K. Lundquist, Acta Chenz. Scand., 17:13-26 (1963). 154. K. Lundquist and G. E. Miksche. Tetrcthedron Lett., 2131 (1965). 155. H. Nimz, Chem. Ber., 102:799-810(1969). 156. C. Lapierre, B. Monties, and C. Rolando, J . Wood Chem.Technol., 5:277-292 (1985). 157. C. Lapierre and B. Monties, Holiforsch., 40:47-50(1986). 158. C. Lapierre, B.Pollet, and B. Monties, Holiforsch., 4 5 - 6 8 (1991). 159. N. Terashima,R. H. Atalla, S. A. Ralph, L. L.Landucci, C. Lapierre,and B. Monties, Holiforsch., 50:9- 14 (1 996). 160. A. Sakakibara, T. Sasaya, K. Miki, and H. Takahashi, Holzj-orsch., 41:1-11 (1987). 161. M. Erickson, S. Larsson, and G. E. Miksche, ActaChem. Scund., 27:903-914 (1973). 162. S. Larsson and G. E. Miksche, Acta Chern. Scand., 25:647-662 (1971). 163. K. Freudenberg, Holiforsch., 18:3-9(1964). 164. K. Freudenberg, Hokforsch., 18:166- 168 (1 964). 165. K. Freudenberg, Science, 148595-600(1965). 166. K. Freudenberg and A. C. Neish, in Molecular Biology, Biochemistry and Biophysics, Vol. 2, Springer-Verlag, Berlin-Heidelberg, p. 103 (1968). 167. E. Adler, Wood Sci. Technol., /1:169-218(1977). 168. W. Glasser, in Pulp and Paper, 3rd ed., Vol I (J. P. Casey, ed.), Wiley, New York, pp. 39I l l (1980). 169. A. Sakakibara, Wood Sci.Technol., 1489- 1 0 0 (1980). 170. H. Nimz, Angew. Chem., 13313-321(1974). 171. B. J. Fergus and D. A. 1. Goring, Post-Grad. Res. Rep. No. 12, McGill University, Montreal, Quebec, Canada (1968). 172. M. Matsukura and A. Sakakibara, Mokuzai Gakkaishi, 15:35-39 (1969). 173. K. V. Sarkanen, in Lignins (K. V. Sarkanenand C. H. Ludwid, eds.), Wiley-Interscience, New York, p. 154 ( 1 97 1 ). 174. W. A. Cote, B. W. Simon, and T. E. Timmell, Svensk Papperstidn., 69547-558 (1966). 175. K. P. Kringstad andC. W. Chang, Tappi, 52:2382-2385(1969). 176. F. Yaku, Y. Yamada, and T. Koshijima, Holiforsch., 30: 148-156 (1976). 177. 0. Eriksson and B. 0. Lindgren, Svensk Papperstidn., 80:59-63(1977). 178. R. Simonson, Svensk Papperstidn., 74:153-165(1971). 179. D. FengelandM. Przyklenk. Svensk Papperstidn., 78:617-620 (1975). 180. 0. Eriksson and D. A. I. Goring, Wood Sci.Technol., 14:267-279 (1980). 18I . K. Freudenberg and G. Grion, Chrrn. Ber., 92: 1355- 1363 (1959). 182. K. FreudenbergandJ. M. Harkin, Chem. Ber, 93:2814-1819(1960). 183. E. F. Brauns, J . Am.Chem. Soc., 61:2120-2127(1939). 184. K. Freudenberg, Angew. Chenz.,68:84-92(1956).

This Page Intentionally Left Blank

Chemistry of Cell Wall Polysaccharides Tadashi lshii and Kazumasa Shimizu Forestry and Forest Products Research Institute, Ibaraki, Japan

1.

INTRODUCTION

Cellulose and hemicellulose are major cell wall polysaccharides in woody plants which consist mainlyof lignified secondaryxylem.Theirchemicalstructureshavebeenwell characterized [ 1-41. Recent studies have revealed important biological and physiological functions of the primary cell wall polysaccharides in plant growth and regulation [5-71. Significant differences exist in polysaccharide composition of the primary cell walls and the lignified secondary walls. In this chapter, the structure and functions of cell wall polysaccharides, particularly pectin and hemicellulose, in growing tissue are discussed.

II. PRIMARY CELL WALL COMPONENTS The primary cell wall consists of two phases, a microfibrillar phase and a matrix phase. As shown in Table I , the primary cell walls consist of pectin, hemicellulose, cellulose, and a small amount of glycoproteins and phenolics. Polysaccharide compositions of primary cell walls are very similar in hardwoods(angiosperms)andsoftwoods(gymnosperms) (Table 2) [8], although the secondary wall polysaccharide compositions of these two divisions are different [4]. Traditionally, polysaccharides are classified as follows: pectins are polysaccharides extracted from cell walls by hot water, ammonium oxalate, weak acid, or chelating reagents; hemicelluloses are not extracted by weak acids but are by relatively strong alkali. The wall residue remaining after alkali extraction is mainly cellulose. This classification based on extraction procedure is convenient to use, but it is not always appropriate. For instance, pectic polysaccharides can be defined as galacturonic acid-containing polysaccharides [9]. Using this structural definition, pectic polysaccharides are extracted with both chelating reagents and with alkali. The a-cellulose fraction also contains residual galacturonic acid residues. The subdivision of matrix polysaccharides into pectin, hemicellulose, and cellulose is, however, useful (see Section IV).

175

lshii and Shimizu

176

TABLE 1

Primary Wall Components

Phase

Components

Microfibrillar Matrix"

Xylan

Cellulose (p-1,4-glucan) Pectins

Rhamnogalacturonan I Arabinan Galactan Arabinogalactan I Homogalacturonan Rhamnogalacturonan I1

Hemicelluloses Glucomannan Mannan Galactomannan Glucuronomannan Xyloglucan Callose (p-1.3-glucan) p- 1,3-, p- 1,4-glucan Arabinogalactan I1

ensin

Proteins

Phenolics

acid

Arabinogalactan protein Others, including enzymes Ferulic Others, e.g., coumaric acid, truxillic acid

"NB: Not all these matrix components are found in all cell walls. Source: Ref. 19.

111.

MONOSACCHARIDE RESIDUESCOMMONLY FOUND IN PLANT CELL WALLS

The most commonly occurring sugars found in cell wall polysaccharides are shown in Fig. 1. The carbonyl (aldehydo or keto) carbons of reducing sugars form acetal or ketal linkages, called glycosidic linkages. A glycosidically linked sugar is called a glycosyl residue; e.g., 4-linked glucosyl residues are glucosyl residues glycosidically linked at C- 1 and also have another glycosyl residue attached to them at C-4. In this chapter the following abbreviations are used: Glc = glucose; Gal = galactose; Man = mannose; Xyl = xylose; Api = apiose; Ara = arabinose; Rha = rhamnose; Fuc = fucose; GlcA = glucuronic acid; GalA = galacturonic acid; AceA = aceric acid; Kdo = 3-deoxy-~-manno-octulosonic acid; Dha = 3-deoxy-~-lyxo-2-heptulosanic acid; p = pyranose ring form; f = furanose ring form.

IV.

PURIFICATION OF PRIMARY CELL WALLS AND ISOLATION

OF POLYSACCHARIDES Xylem differentiating zones of woody plants, commonly called cambial tissues, contain primarywallsandmiddle lamella. Collection of cambial tissues for chemical analysis requires many young trees and much time-consuming work. The cambial tissues of aspen (Populus tretnuloides), basswood (Tilia umericarzn) [ 10- 141, and sugi (Cryptotneria jap o n i c a ) [ 151 were isolated and their cell wall polysaccharides were well characterized.

a 9

TABLE 2

Major Polysaccharides of Primary Cell Walls a

Dicots Suspension-cultured sycamore cell"

0'

z u)

Gymnosperms Aspen Cambial tissueh

Suspension-cultured Douglas-fir cells'

Sigi cambial tissue"

wt % of cell wall

Polymer

0

z ij: (P

u)

Cellulose Hemicellulose

Pectin

"From "From 'From "From

Ref. Ref. Ref. Ref.

5. 10.

8. 15.

Xyoglucan Xylan Glucomannan Homogalacturonan Rhamnogalacturonan-I Rhamnogalacturonan-I1

30 25 5 15 15 5

22 6 11 1

147

23 15 2

j2Z

35 18 5 8

178

lshii and Shimizu

bH OH p-D-Glucopyranose

bH

OH a-L-Rhamnopyranose

a-L-Fucopyranose

p-D-Xylopyranose

p-D-Mannopyranose

OH a-L-Arablnofuranose

CYOH

COOH

COOH

OH

I

OH p-D-Galactopyranose

64 CHpOH

OHOH

p-D-Aplofuranose

I

I

OH a-D-Galactopyranosyluronicactd

'0

OH p-D-Glucopyranosyl-uro nic acid

CH2OH

I

COOH

HO-C

H

OH 2-keto3-deoxy-D-manno2-octulosonic acld (Kdo)

"

O

H COOH

3-dWxY-D-l~~o-2-hep tulosaric acid (Dha)

OH p-L-Aceric acld

FIGURE 1 Structure of monosaccharides commonly found in the cell wall.

Suspension-cultured cells are also a good source for preparing primary cell walls. Albersheim and his co-workers [ 161 have used suspension-cultured sycamore (Acer pseudoplaranus) cells for studying the structures and functions of primary cell walls. Suspensioncultured cells of woody plants can be easily obtained from young leaves, seedlings, and cambial tissues. Procedures for preparation of cell walls from suspension-cultured cells (Fig. 2 ) and isolation and fractionation of polysaccharides are described byYork et al. [ 161, while alternative procedures for preparing cell wall materials from fresh tissues have

Chemistry of Cell Wall Polysaccharides

h

3

aJ

.e

5 c Ir

M

0

0

2 v)

v

8

B

3 M

aJ

v)

Y

:a 2&

179

lshii and Shimizu

180

been proposed by Selvendran and O’Neill [ 171 (Fig. 3). The sequential extraction of pectin and hemicellulose from cell wall material is shown in Figs. 4 and 5, respectively. These approaches gave an enormous amount of information on the structure of individual polysaccharides, but they also have their limitations. One notable disadvantage is that certain bonds in the wall are destroyed in order to extract its components [ 18,191, which may make it difficult to observe important covalent bonds of the polysaccharides. This drawback might be partly overcome by comparing different extraction methods and especially by comparing chemical and enzymatic extraction methods. A second limitation is that cell wallpolysaccharidesareoftenheterogeneouswithrespectto their primary structure and molecular size. Extraction and purification of polysaccharides may result in only partial recovery of a particular type of molecule, and the recovered material may not be representative of the whole. This ultimately leads to a distorted picture of the overall structure of the component. Despite the many limitations of studying cell wall polysaccharides, extremely powerful tools for analyzing small amounts of complex carbohydrates are currently available. Theseinclude capillary gas-liquidchromatography-massspectrometry (capillary GCMS), liquid chromatography-mass spectrometry (LC-MS), fast-atom-bombardment mass

Fresh tissue Homogenize in 1.5%(w/v) SDS (2.5 fresh weight) containing 5 mM Na,S20s at 5-10°C until homogeous Filter and wash (x 2 ) in 0.5% (w/v) SDS containing 3 mM dm-3Na2SzOS

t

Supernatant: containing cell contents and watersoluble cell wall components

Residue

+

Ball mill at 1°C in 0.5%(w/v) SDS containing 3 mM dm-3Na,S,05 Centrifuge andwash in water (x 2 )

1

+

Supernatant:containing water-soluble pectic substances

Residue

(1) Extract with 90%(v/v) DMSO to remove starch (2) Wash in distilled water, and dialyse

Purified cell wall material (Yield, 0.5-2%(w/w) of fresh material)

FIGURE 3 Procedure f o r (he preparation of cell wall material from fresh tissue. (From R d . 17.)

181

Chemistry of Cell Wall Polysaccharides

Cell wall material (1 g) CDTA-(1) Stir with 100 mLCDTA 50 mM, pH 6.5 for 5 h at 20°C Centrifuge and wash lx with distilledH 2 0 Add washingsto supernatant (SIN) Residue

SIN

filter through glassfiber paper, dialyse, concentrate under reduced pressurebelow 40"C, dialyse and concentrate CDTA-( 1) soluble

I

CDTA42) Re-extract withCDTA as above for 2 h Centrifuge and wash as above

t Residue

SIN

Filter, dialyseand concentrate CDTA-(2) soluble Na2C03-(1) Stir with 100 mL 50 mM Na2C03+ 2 0 mM NaBH, (0,-free unde Ar) for 2 0 h at 1'C Centrifuge and wash withHzO

t

Residue

UN

Na2C03-(2) Re-extract with 50 mM Na2C03 as above for 2 h at 20°C Centrifuge and wash

Extractwithalkali(Fig.

FIGURE 4

5)

Extraction of pectic polysaccharides

1

Filter, adjustpH to 5 with acetic acid, dialyse and concentrate Na2C03-(l)-(cold)-soluble

SIN Filter,etc.,asabove Na2C03-(2)-(room temperature)-soluble from cell wall material. (From Ref. 17.)

lshii and Shimizu

182

Depectinated CWM

Stir with 75-100 mL 02-free 0.5M KOH + 10 mMNaBH, under Ar for 2 h at 20°C Filter under suction

v

Residue

Filtrate, adjust pH to 5 with acetic acid, dialyse and concentrate 0.5 M KOH-soluble Stir with 75-100 mL 02-free 1 M KOH + 10 mMNaBH, under Ar for 2 h at 20°C Filter under suction

v Residue

Filtrate, adjust pH to 5 with acetic acid, dialyse and concentrate 1 M KOH-soluble

Extract with 75-100 mL 0,-free 4 M KOH + 10 mMNaBH, under Ar for 2 h at 20'C

l+

Filter under suction

t

Residue a-Cellulose

Filtrate, treat as above

4 M KOH-soluble

FIGURE 5

Extraction of hemicellulose from depectinated cell wall material.

(From Ref. 17.)

spectrometry (FAB-MS), electrospray ionization mass spectrometry (ESI-MS), matrix-associated laser-desorption time-of-flight mass spectrometry (MALDI-TOF-MS), and highresolution nuclear magnetic resonance (NMR) spectroscopy. Purified enzymes for selectively hydrolyzingglycosyllinkages are also available. These tools provide much information regarding the structure and function of cell wall polysaccharides, even though there are always problems associated with isolating complex polysaccharides and elucidating their exact structural features.

Polysaccharides Chemistry Wall of Cell

183

Several methods are available for the analysis of intact cell walls. Solid-state NMR and Raman spectroscopy give useful information. Immunocytochemical study using goldlabeled antibodies and wall-degrading enzyme also yields information about the distribution of cell wall components.

V.

STRUCTURAL ANALYSIS OF POLYSACCHARIDES

Polysaccharides are characterized by glycosyl composition and glycosyl-linkage analyses. Figure 6 is a flow chart for the sequencing of the neutral and acidic polysaccharides.

A.

MonosaccharideComposition

1. Acid Hydrolysis Plant cell wall polysaccharides contain at least 12 monosaccharides which are linked by a variety of glycosyl linkages. Theirsusceptibility to acid hydrolysis varies [20]. Therefore, a single hydrolysisprocedurecannotprovidequantitativehydrolysis of everyglycosyl linkage. Furthermore, the monosaccharides released by acid have different acid stabilities. Neutral noncellulosic polysaccharides can be hydrolyzed quantitatively using 1 M H,SO, for 2.5 h at 100°C, or 2 M trifluoroacetic acid (TFA) for 2 h at 120°C [21]. The cellulose Polysaccharide Glycosyl composition Analysis Glycosyl linkage analysis

NeutraVAcidic

I

I

I

Partially degrade

Carboxyl-reduce

I

Methylate

I

I

(chemiclally, enzymatically) Oligosaccharides

Gel permeation Ion-exchange HPLC Purified oligosaccharides

'"C-NMR

FIGURE 6

NeutraiIAcidic

Methylate I

Carboxyl-reduce I

I

Partially degrade Reduce Realkylate

Alkylatedoligoglycosyl

alditol

l

I

Reversed-phase HPLC

Hnarysls

Flow chart lor sequencing polysaccharides. (From Ref. 17.)

184

lshii and Shimizu

is first degraded with 72% H,SO, for 2 h at 20°C and then hydrolyzed completely by 1 M acid for 4 h at 100°C [ 171. TFA is more convenient than H,SO, because it is volatile and gives higher recovery of neutral and acidic sugar residues than H,SO,. Ketoses are very acid-labile sugars and degrade during hydrolysis of polysaccharides. Kdo and Dha can be released by 1 M acetic acid at 40°C for 16 h and 2 M TFA at 120°C for 5 min, respectively [9]. Kdo and Dha are specific components of rhamnogalacturonan-I1 (RG-II), which are present in all higher plants [9].

2. AlditolAcetatesMethod Neutral sugars released by acid hydrolysis are reduced with sodium borohydride (NaBH,), acetylated, and analyzed as alditol acetates by gas-liquid chromatography (GLC) [ 161. The alditol acetates are well separated from each other by GLC and determined quantitatively (Fig. 7).

3. TrimethylsilylEthers of MethylGlycosides Polysaccharides that contain acidic sugarresidues are subjected to methanolysisand methyl glycosides of neutral sugars and also methyl ester of uronic acid residues, which are converted into trimethylsilyl (TMS) derivatives. TMS derivatives can be analyzed by GLC with capillary column [l61 (Fig. 8).

B. Glycosyl-LinkageAnalysis Methylation analysis gives information about the glycosyl linkages of the polysaccharides. The procedure includes permethylation of all free hydroxyl groups, hydrolysis, reduction of partially methylated monosaccharides, and acetylation. The Hakomori method, using sodium dimethylsulfinyl anion (dimsyl anion) in dimethylsulfoxide, is commonly used for permethylation [22]. Potassium dimsyl anion is more easily prepared [23]. Blakeney and Stone [24] prepared the lithium carbanion from butyllithium, which is available in a much purer form than the alkali metal hydrides, thus providing a clearer chromatogram. Ciucanu and Kerek [25] reported a new methylation method in which powdered sodium hydroxide and iodomethane are added to carbohydrate in dimethylsulfoxide. Methylation with sodium hydroxide and iodomethane in SOz-diethylamino-dimethylsulfoxide is useful for cellulose and lignified samples [26]. Acidic polysaccharides have to be reduced either before or after methylation to analyze their uronic acids as partially methylated alditol acetates (PMAAs). The uronic acids are reduced with a deuterated reagent so that they can be distinguished by mass spectrometry. Water-soluble polysaccharides whose carboxyl acids have been activated with carbodiimide can be reduced with sodium borohydride and subjected to methylation analysis. Alternatively, the methylated polysaccharides are readily reduced by refluxing with lithium aluminumdeutride in dichloromethaneether [17]. York et al. reported that lithium triethylborodeutride in tetrahydrofuran is highly effective onmethyl esterified methylated polysaccharides [ 161. Methylated polysaccharides are hydrolyzed with 2 M TFA for I h at 120"C, reducedwithsodiumborodeutride,andacetylated.PMAAs are separatedon capillary GLC columns. PMAAs can be easily identified from the pattern of fragment ions produced by electron-impactmassspectrometry[27].ThePMAAsderivedfrom arabinoxylan are shown in Fig. 9.

185

Chemistry of Cell Wall Polysaccharides

l 4

x

X

I

I

I

I

I

1

I

Time (min) FIGURE 7

GLC of alditol acetates of neutral sugars.

VI.

MATRIX POLYSACCHARIDES

A.

Pectic Polysaccharides

The primary cell walls of dicots have a relatively high content (-35%) of pectic polysaccharides. The primary wall and middle lamella are rich in pectic polysaccharides. The most characteristic glycosyl residue of pectic polysaccharides is galacturonosyl residue. Rhamnosyl, arabinosyl. and galactosyl residues are also present. The rhamnosyl residues are closely associated with galacturonosyl residues in integral components of the polysaccharides. Arabinosyl and galactosyl residuesarecomponents of arabinan and galactan, which are covalently attached as side chains to a rhamnogalacturonan backbone 191. Pectin is covalently linked to phenols 1281, and may also be linked to protein 128,291 and lignin [301. Three pectic polysaccharides have been isolated and characterized. These are homogalacturonan,rhamnogalacturonan I (RC-I), and rhamnogalacturonan I1 (RG-11). Hornogalacturonan, oligogalacturonides [degree of polymerization (DP), between I and 31, R C I, and RG-I1 were solubilized from suspension-culturedsycamore cell walls with CY-l,

lshii and Shimizu

186

0

f

FIGURE 8

C

E

GLC of TMS derivatives of neutral and acidic sugars.

(From Ref. 16.)

4-endo-polygalacturonase (EPC) treatment. These results suggest that homogalacturonan is covalently attached to RG-I and RG-I1 in vivo [9].

1. Homogalacturonan Homogalacturonan is a homopolymer of a- 1,4-linked galacturonic acid (Fig. 10). Oligogalacturonides having DP 2- 12 were prepared from cell walls of soybean by partial acid hydrolysis and isolated by ion-exchange chromatography [3l]. Recently, oligogalacturonides (DP 5-20) were separated by high-performance ion-exchange chromatography on a Carbo Pac PA1 column (Dionex) (Fig. 11). The carboxyl groupsof many of the galacturonosyl residues of pectic polysaccharides in cell walls are partly esterified [32], but the distribution of the methyl esters is unknown. Calacturonosyl residues are partially acetylated at 0-3 position [33]. Homogalacturonanshave structural roles in the plant cell walls, and oligogalacturonidesderivedfromthehomogalacturonanhave biological activities [34],including elicitation of defense responses, influence on plant growth and development, and promotion of rapid changes in ion flux (Table 3). 2. Rhamnogalacturonan I (RG-I) RG-I was solubilized from suspension-cultured sycamore cell walls after treatment with EPG[9].Glycosylcomposition analysis of RC-Ishowed that arabinose,galactose,galacturonic acid, and rhamnose are its major monosaccharide constituents (Table 4). Small amounts of fucoseare present. Glycosyl-linkageanalysis,followingreduction of the methyl esterified galacturonic acid residue, established that RC-I is a complex, branched

187

Chemistry of Cell Wall Polysaccharides

a

$

"

o

~

o

9

OH

OH

o \

~

G J + Me0

OH

OH

J

o

Q

Methylation

g

o

\

COOMe

MeO

M

OMe

OMe Reductbn of uronate AcM hydrolysls

CHDOAc

CHDOAc

M&$"" AcO

AIS{-::

CH20Me CH20Ac 1-Araf

t

Reduction Acetylation

CHDOAC

M e 0 i o M e OAC

OAc CH20Ac

CH20AC 3,4-Xylp

CHDOAc

+XYlP

FIGURE 9 Partiallymethylatedalditolacetates

24-XYlP

from arabinoxylan.

188

lshii and Shimizu

-+4)-a-~GalpA-(l+4)-a-~GalpA-(l-+4)-a-~-GalpA-(l+4)-a-~-GalpA-(l+4)-a-~GalpA-(l-+ FIGURE 10 Structure of homogalacturonan. Galacturonic acid residues are partly esterified.

pectic polysaccharide (Table 5 ) composed of the repeating disaccharide +4)-cu-D-GalpA(1+2)-cw-~-Rhap-(1+ (Fig. 12-1). Typically, 50% of the 2-linked rhamnosyl residues are submitted at 0-4, but the pattern of side-chain substituents along the backbone is not clear. Calacturonosyl residues of RC-I are acetylated on 0 - 2 and 0 - 3 135-371. Acetylated RCI oligomers were obtained from the walls of bamboo shoot by Driselase hydrolysis (37). RC-I has a DP of about 2000, even after extraction from the walls by EPG treatment 191. RC-I has a number of different side chains, attachedto the 0 - 4 position of rhamnosyl residues. Lithium degradation of the galacturonosyl residues in RC-I released oligoglycosy1 side chains attached to 0 - 4 of 2-linked rhamnosyl residues (Fig. 12-2) 191. The cell walls of suspension-cultured Douglas fir (Pseudorsugu menziesii) cells [8] and the cambial tissue of sugi 1381 have a RG-I with a structure very similar to sycamore RC-I. Pectic polysaccharides extracted from various tissues and plants contain RC-I with a variety of side chains. Their structures are remarkably conserved. It may be that cells produce RCI with different side chains depending on ages and difference in tissues. The RC-I may have some physiological functions in cells, but there is little evidence for this notion. Some walls [Chenopodiaceae, e.g., spinach (Spinncia oleraceu and sugar beet (Betu ~xdgaris)] contain esterified phenolics. Several feruloyl oligosaccharides wereisolated from enzymatic hydrolyzates of spinach suspension-cultured cell walls 139,401, spinach leaves [41], and sugar beet pulp [40]. They were identified to be 0-(2-O-trtrn.s-feruloyl-a-~-Ar~fl(1-+5)-~-Acaf (Fig. 13-1), 0-(6-O-rrc~ns-feruloyl-~-~-Galp)-( 1+ 4 ) - ~ - C a l p (Fig. 13-2, 0a-~-Araf-(1+3) 0-(2-O-rlun.s-feruloyl-~-~-Ar~f)-( l + 5 ) - ~ - A r ~ f(Fig. 13-3), and 5 - 0 rrc/n.s-feruloyl-c~-~-Araf-( 1 +3)-O-p-u-Xylp-( 1+4)-~-Xylp (Fig. 13-4). These feruloyl res-

540-

480420 -

- 360al

U7

6

3006

0

f

240-

I

U

1800

12060 -

0-

B

l

11

Polysaccharides Chemistry Wall of Cell TABLE 3 Biological Activities of

189 a-l ,4-Oligogalacturonides

Activity

dp“

Molar conc.h

Plant

Plant defense responses Induction of

phytoalexins

Induction of proteinase Inhibitors ligninInduction of

of

8-13 (12)’ 9-13 (13) ND* 2-20 20 8-11 (11) ND

-

-

5 1 0 - 4 10-5

10”

ofInduction P-1,3-glucanase ND Induction ND chitinase Induction of isoperoxidases ND Inhibition of hypersensitive response ND Elicitation NDof necrosis

Soybean Castor bean Parsley Tomato Tomato Cucumber Castor bean Parsley Tobacco Castor bean Tobacco Cowpea

Development and growth Inhibition of auxin-induced elongation Regulation of TCL morphogenesis: flower formation Induction of ethylene

Pea stem

>8 10-14 (12-14)

10”

>8 22

-

Enhancement of cell expansion and 100 separation Rapid responses at the plasma membrane and cell surface Efflux of K’ and influx of Ca’+ Rapid depolarization of plasma membrane and ND Induction of HzO,oxidative burst Enhancement of in-vitro phosphorylation of 34-kDa protein

Tobacco Tomato Pear Soybean

Tobacco Tobacco

12-15 1-7 and 10-20

Soybean 14-20

-I 0~-7

Tomato

‘Dp (degree of polymerization) range of oligogalacturonldes that show the designated biological activity. hOrder-of-magnitude estimatton of the concentratton of oligogalacturonldes that give the half-maximum biological response. The concentration is included only where purified oligogalacturonides are assayed. ‘Numbers in parentheses represent dp of most active oligogalactouronide. dND = dp of active oligogalacturonides not determined. Source: Ref. 34.

iduesareattached to arabinanandgalactan that are sidechains of RC-I in pectic polysaccharides.They may beinvolved in the cross-linking of pectin in cell wall (see Section VI).

3. Rhamnogalacturonan I1 (RG-11) RC-I1 is a polysaccharide composed of a 1,4-linked a-D-galacturonic acid backbone with both keto-sugars (i.e., ketoses) and aldehyde sugars (i.e., aldoses) in its oligosaccharide

lshii and Shimizu

190

Glycosyl Residue Composition of RG-I from the Walls of Suspension-Cultured Sycamore, and Douglas Fir

TABLE 4

Glycosyl Douglas Sycamore‘ Sycamoreb residue‘‘

fir”

Rha

16

9

9

Fuc

2

1 35 2 43 0 10

4 30

Ara XY 1

Gal Glc GalA

4

32 0 31 0

21

6 26

“Values expressed as mol%. hEndo-polygalacturonase-solubilizedRC-I. ‘NazCO,-solubilized RG-I. “LiCl-solubilized RC-l. Source: Ref. 9.

side chains. RG-I1 was for the first time isolated from the walls of suspension-cultured sycamore cells with EPG hydrolysis followedby size-exclusion chromatography [23]. The polysaccharide is composed of about 11 glycosyl residues and has extremely complicated glycosyl-linkage compositions (Tables 6 and 7) [9]. Aceric acid (3-C-carboxyl-S-deoxy-~xylose) (Fig. 1 ) was identified for the first time in nature as a component of RG-I1 [42]. Kdo and Dha (Fig. l), which are acid-labile molecules, were also identified as integral components of RG-I1 [9]. Partial acid hydrolysis of the RG-I1 gave several oligosaccharides (Fig. 14) [9,43]. Organization in RG-I1 of the oligosaccharides was elucidated by sophisticated sugar linkage analyses. Partial acid hydrolysis of methylated, carboxyl-reduced. and remethylated RG-11, in combination with selective deuteriomethyl labeling of those hydroxyl groups exposed by the partial acid hydrolyses, showed the points of attachment of oligosaccharides to the a-1A-linked galacturonosyl residues [9,44] (Fig. 15). The exact galacturonosyl residues in the homogalacturonan (Fig. 14-1) to which oligosaccharides 1, 2, 3, and 4 (Fig.14-2-5) are attachedhave not beendetermined. RG-I1is present in Douglas fir [S], sugi [38], rice (Oyvzra sativa) [4S], onion (Allium cepa] [46], kiwi fruit (Actinidin deliciosa) [47], bamboo shoot [48], and all other higher plants that have been examined [S]. Very recently a borate-RG-I1 complex was isolated from radish (Rq3hanu.s sativus) roots 1491, sugar beet pulp [SO], sycamore [SI], red wine [ S I ] , and bamboo shoot 1521. Boron (B) is knownto be an essential micronutrient for all higher plants [ S ] . Borondeficiencysymptoms first appear at growingpointsand are characterized by cell wall abnormalities (541. The finding that B selectively binds RG-I1 in pectin to form a crosslink (Fig. 16) implies that B-RG-I1 complex plays an essential role in cell wall architecture and cell formation. A pectic polysaccharide containing galacturonosyl and xylosyl residues (xylogalacturonan) (Fig. 17) was isolated from mountain pine pollen 1.551 and modified hairy regions of apple [ 5 6 ] . Xylogalacturonan was solubilized when the hairy region of polysaccharides was treated with rhamnogalacturonan hydrolase, which cleaves the backbone of RG-I but does not cleave xylogalacturonan. This result suggests that xylogalacturonan and RG-I is connected covalently in the wall.

Polysaccharides Chemistry Wall of Cell

191

Glycosyl-Linkage Composition of RC-I Isolated from the Cell Walls of Sycamore, Douglas Fir, and Maize

TABLE 5

48%

Glycosyl linkage" Rha

Fuc Ara

Sycamore Percent branched rhamnosyl residues fir

T

0

2 4 2,3 2,4 239

7.8

T

0.6 1.4

0.6

3,4

0

0

T

9.5 2.2 2.2 11.2

2.2

2 3 5

23 3s XYI Gal

1 .O

T

3.5 2.0

4 2.4

0 0

T

2 3 4 6 2,4 2.6 3.4 3,6 4.6 GlcA GalA

0 0 8.O

3.0 15.2 2.8 1.6 8.O

T T

4 2,4 3,4

6.3 0.6 2.7 8.4 7.5 6.3 1.2 0 l .2 2.4

1.1

0

0.9 3.4 l .4 1.1

0.7 0

0.5

1.3 0.6 0 10.6 2.0 2.4 0 2.8 0.4 2.6 8.0 3.2 5.9 0.3 0

0

0

8.1

6.6 2.9 3.9 11.3 8.1 6.6 2.5 1.4 6.5 0.8

1.o

2.2 2.9 4.4 0

0.2 0.4 0.6 1.5

0

0

0

1.6 15.2

4.4 30.6 0.2

0.6 12.5 0.2 1.9

1.o 0

1.1

Maize

0 1.1 0 t

0.4

2.8

11.8 2.1

0.5

t 0.5

13.4 t

17.2 19.8 6.8 7.4 t t t 15. I

0.8 4.2 4.9

1.o

15.1 0

5.7 0 0

0 0

14.9 0.8 2.5 13.0 l .4 6.1 1.1 1 .O 1.1

14.2 0.2 3.5 4.8 4.8 14.2 1.9

1.2

0.7

1.1 1.1

2.6 0

1.6

0 0' 0 0 0

0' 0 0 0

"Values expressed as mol%. Source: Ref. 9.

4.

Arabinans

Arabinans have been isolated from the tissues and cell walls of many plants [7]. It still remains unknown whether arabinans exist in growing tissue as separate homopolymer, or as covalently linked side chains of RG-I. Arabinans are highly branched molecules composed of a-lS-linked arabinofuranosyl residues that are more frequently substituted at 03 than at 0-2 (Fig. 18). Sycamore RG-I contains arabinosyl oligosaccharides (DP 2-20) attached to 0-4 of the 2-linked rhamnosyl residues in the backbone [36,58].

192

lshii and Shimizu

(1) +4)-a-~-GalpA(1-.2)-a-L-Rhap-(l--4)-a-D-GalpA(l-r2)-a-L-Rhap(l-

a-L-Fucp-( 1-2)-p-D-Galp( 1-r4)-p-D-Galp(l-r4)-Rhamnitol

a-L-Araf-(1+5)-a-L-Araf-( 1-+2)-a-L-Araf-(l-r3)-p-~-GaIp(1-.4)-Rhamnitol

Araf-[Araq,,-Rhamnitol

Galp-[Galp],,-Rhamnitol

FIGURE 12 Structure of rhamnogalacturonan-l: ( 1 ) backbone structure and (2) several side chains.

5. Galactans Some primary cell walls contain p-IP-linked galactans (Fig. 19) [7]. Galactose-containing oligosaccharidesattached to 0 - 4 of 2,4-linkedrhamnosylresidueswere isolated from sycamore RG-I [59,60] and tobacco RG-I [61].

Arabinogalactan Two types of arabinogalactan were isolated from plants (Fig. 20) [7]. Arabinogalactan I is a polysaccharide composed of a p- 1,4-linked galactosyl backbone that is substituted at 0-3 with short a-i,S-linked arabinosyl side chains. Arabinogalactan I1 is found in gymnosperms, especially in larches 1621. It is a highly branched polysaccharide containing p3, p-6-, and P-3,6-linked galactose with various amounts of arabinosyl, galactouronosyl, andglucuronosyl residues. Suspension-cultured plant cellshavebeenfound to secrete arabinogalactan protein into the culture medium 191. The polysaccharide portion is very similar to arabinogalactan 11. This polysaccharide may be covalently attached to hydroxyproline-rich proteins [62,63]. 6.

B. Xyloglucan Xyloglucan is the principal hemicellulose of the primary cell walls of dicotyledonous plants. It was first isolated and characterized from tamarind (7hmarindu.s indicu) seeds. The polysaccharidesformabluecomplexwith iodine. Thestructureandfunction of primary cell wall xyloglucan have been reviewed extensively [7,64,65]. The basic structure of this cell wall polysaccharide consists of a backbone of p-1,4-linkedD-ghCOSYl residues, with D-xylosyl side chains a-linked to 0 - 6 of some of the glucosyl residues. Some of the xylosyl side chains are extended by the addition of D - G a l j , ~ - F u c + 2 - a - ~ - G a l + to 0-

193

Chemistry of Cell Wall Polysaccharides

H.OH

' * O H OCH3

c=cH

b Lo (4)

FIGURE 13 Feruloyl oligosaccharides isolated from spinach and sugar beet cell walls.

194

lshii and Shimizu

TABLE 6 Glycosyl-Residue Composition of RC-I1 Isolated from

Different Plant Sources

RiceSycamore residue" Glycosyl 3.7 4.1

14.5 29.3

Rha Fuc 4.9 2MeFuc 5.3 Ara 2MeXyl 7.3 Apiose 9.0 Gal 12.3 GlcA GalA 26.7 Aceric acid Kdo Dha

2.8 3.5 10.0 4.8 12.2 9.0 3.2 31.2 3.5 3.5 3.5

15.3 4.1

10.0

10.2

6.3

6.7

+ + +

+ + +

"Values expressed as mol%. Source: Ref. 9.

TABLE 7

Glycosyl-Linkage Composition of RC-I1 from Different Plant

Sources" ~

~

~

ycamore linkagebGlycosyl Rha

T'

2 5.9 3 3.72 3

3.1 Fuc 5.52MeFuc Apiose Ara

5.3 9.0 5.8

GlcA

Tf

TP 2, T

2MeXyl Gal 5.6 GalA

T

3'

T

10.0

2,4 3,4 10.2 T 4 3.13,4 2.82,4 2,3,4

8.6

6.6 t

5 .O t

5.7 4.5 4.8 10.9 6.1 0 5 .O

4.5 5.24.9 6.5 0

10.3 6.08.8 7.3 4.6 1 .S

6.3

0

4.2 7.3 6.8 5.5 0

1.7 6.3

10.2 6.3 4.9 0 4.1 2.8

7.9 7.6 4.2

2.7 6.3

1.5

6.7

"5-LinkedKdo and 5-linked Dha are also present in these preparations and account for -5% of the material. hValues expressed as mol%. Source: Ref. 9.

Chemistry of Cell Wall Polysaccharides

195

a-D-GalpA-(l~4)-(a-D-GalpA]5.,-(1-4)-~-GalpA (1)

a-D-Galp 1

1

&

4

3 ~-D-GalpA-(l+4)-a-~-Fu~p(l+4)-p-~-Rhap(l+d)-Apif 3 2 2

t

t

1 2Me a-D-Xylp

a-D-GalpA

1

(2)

2Me a-L-Fucp 1

t 2 a-D-Galp(l-+2)-p-~-Acef-(l+3)-p-~-Rhap(l+3')-Apif

4

t 1 a-L-Arap 2

t 1

a-L-Rhap

(3) a-L-Rhap(l+5)-D-KDOp (4)

p-L-Araf-(l-.5)-D-DHAp (5)

FIGURE 14 Oligosaccharides releasedfromrhamnogalacturonan I1 bypartialacid

hydrolysis.

2. Structure of xyloglucan oligosaccharides and their nomenclature are shown in Fig. 21 [66]. In some tissues, the xyloglucan has been proposed to be composed of a repeating nonasaccharide unit (XLLG, Fig. 22-l), or of alternating nonasaccharide and heptasaccharide units (XXXG,Fig. 22-2) [7].Xyloglucan isolated from sycamore extracellular polysaccharides has acetyl groups [67,68], attached to the 2-linked P-galactosyl residues of thenonasaccharidesubunit. These P-galactosylresiduesweremono-0-acetylated and di-0-acetylatedat 0-6, 0-4, and 0 - 3 at degrees of 55-60%, 15-20%, and 2025%, respectively. Sycamore xyloglucan also contains arabinose residues. An arabinosecontainingheptadecasaccharidewas isolated fromsycamoreextracellularxyloglucan (XXFGAXXG, Fig. 22) [68]. The heptadecasaccharide was a combination of nona- and heptasaccharide components. An arabinosyl residue was glycosidically linked at 0-2 at

9

c

Q

196

i

4:-

.-.

v

4 P t

d

'c

h t

- 2n h + ci

h

n

'c

-2

U

dai

P b Q

lshii and Shimizu

8

i

0

cc

-L-Araf

1 97

Chemistry of Cell Wall Polysaccharides

+4)-a-~GalpA-(l+4)-a-~GalpA-(l-t4)-a-DGalpA-(l+4)-a-~GalpA-(l+4)-a-~GalpA-(l+

3

3

t

t

1

1

P-DXYlP

FIGURE 17 Structure of xylogalacturonan.

~5)-a-~-AraF(l+5)-a-~-AraF(1-15)-a-L-Araf(1+5)-a-~-AraF(l-15)-a-~-AraF(1+5)-a-~-Araf-(l~

3

3

2

t

1‘

t

1 a-L-Araf

1

1

FIGURE 18 Structure of arabinan.

+4)-p”alp-(1~4)-p-D-Galp-(l-t4)-p-PGalp-(l~4)-p-D-Galp(l+4)-(3-D-Galp-(1-14)-~-D-Galp-~l+ 3 3

t 1 a-L-Araf-(l+5)-a-L-Araf-(l+5)-a-L-Araf

FIGURE 19 Structure of galactan.

t 1 a-L-Araf

L-Araf

lshii and Shimizu

198

Arabinogalactan I

~3)-~-r>Galp(l~3)-~r>Galp(l-13)-~-r>Galp(l~3)-~-r>Galp~l~3~-~-r>Galp~l-13~-(3-~Galp(l~3)-

6

6

6

t

t

t

1 S-c-Galp

1 a-~-Araf-(l-13)-B-r>Galp

1 a-L-Araf-(l-13)-(3-r>.Galp

6

6

6

t

7

7

1

1

1 a-L-AraC(1-13)-(3-r>Galp

a-L-Araf a-~-AraF(1+3)-p-r>Galp

6

6

t

7

1 a-L-AraF(1+3)-p-r>.Galp 6

1

t 1 a-L-Araf Arabinogalactan II

FIGURE 20

Structures ofarabinogalactan I and arabinogalactan 11.

the nonreducing-endglucosylresidue of the heptasaccharidecomponent of theheptadecasaccharide. Xyloglucan oligosaccharides containing from 17 to 20 glycosyl residues were isolated andcharacterized[69-7 l]. Extensive structural characterization of xyloglucan oligosaccharides from various cell walls led to a ‘H-NMR database that allows the complex signals of xyloglucanderivedfromvarious plant species to be assignedwith relative ease [72,73]. The xyloglucans in Solanaceae species have unusual structures. For example,xyloglucansfrom Nicotiniatabacum and Solanumtuberosum do notcontain fucose, but have arabinose instead [64,74]. Xyloglucans in monocot cell walls have fewer substituted xyloseresiduesthan in dicotxyloglucans. Some monocotxyloglucanshave galactosyl-containing side chains like dicot xyloglucans. The presence of fucosyl-containing side chains has not been clearly established in monocot xyloglucans [64]. Monocot xyloglucan contains ester-linked ferulic acid residues [75]. A feruloyl xyloglucan disaccharidewas isolated frombambooshoot cell walls[76](Fig.23).This indicates that xyloglucan may have diferuloyl groups that cross-link xyloglucan-polysaccharides networks in the wall. Xyloglucan has a structural role in plant cell walls. Some or most of the cell wall xyloglucan is hydrogen bonded to cellulose (Fig. 24), as strongalkali is required to extract xyloglucan from cell walls and from pure cellulose in vitro [64]. Xyloglucan binds rapidly and strongly to cellulose in vitro. The binding of xyloglucan to cellulose fibers in the cell

3 (P

Abbreviations Structure Glc+Glc-tGlc-tGlc

t

Xyl

Old

New

XG10

XLFG

xt I xt I

t XG9

XXFG

x

x71 Gal

Fuc

Fuc

FG

XG9n

XLLG

t

t

Xyl

Gal Gal

t

Xyl

t

t

~

XXLG

fGalI

t

t

t xI

? Gal

t

t

Xyl Xyl

f

Xyl

Fuc

Xyl ~

XXXGXXXG

Xyl

t

XLXG

7 1 '

x

Xyl Xyl

t

G1c-t Glc-tG b+G Ic-t Gk+G Ic-tG Ic+G Ic Xyl Xyl

7 Gal

Glc+Glc+G k+Glc Xyl

t

J

XG8

t xI

Xyl

t

Ara

Glc+Gb+Glc-tGb

2-

s

3 v) P) c) c)

8

Xyl Xyl

7 7

s

e5

GIc+G Ic-tG Ic-t GIc-Gc-tG I c j GIc-t Glc

t

.z

3

1'

t t xI xI

2.

%

8

GlcjGlc

I 7 Gal

Glc-tGlc-tGlc+Glc Xyl

XXG

t

t

1

t

t

Xyl

Xyl Xyl

t

Xyl

t

New

XG7-01 XXXGol

Glc+Glc+Glc

GIc+G Ic-t GIc-tGlc Xyl

Gb+Glc+Glc+Glucitol

t

Fuc

t

Old

Structure

Xyl Xyl

f 7 Gal Gal t

7

Abbreviations

~

FIGURE 21 Abbreviated nomenclature of some xyloglucan oligosaccharides. (From Ref. 66.)

XXFGAXXG

lshii and Shimizu

200 a-L-Fucp 1

(1)

1 2 @D-Galp 1 I

I

FIGURE 22 Structure of xyloglucan repeating units of (1) XXFG and (2) XXXG.

wall would probably limit the self-association of cellulose fibers and might provide sites for cross-linking of cellulose fibers [64]. Xyloglucan appears to havea regulatory function as well asstructuralfunctions [7,64,65]. Xyloglucan oligosaccharides (XXFG) inhibited auxin-induced elongation of pea stem segments at about 10” M [78,79]. Related oligosaccharides lacking a fucosyl residue (XXLG and XXXG) are ineffectual. Xyloglucan oligosaccharides can promote the elongation of pea stem segments in the absence of 2,4-D [80].

/

OCH,

OH

OH

FIGURE 23 Structure of a feruloyl xyloglucan disaccharide.

201

Chemistry of Cell Wall Polysaccharides

f

1,11

Cellulose Microfibril FIGURE 24

Hydrogen bonds between cellulose microfibril

and xyloglucan. (From Ref. 64.)

C . Xylans Xylans are major hemicelluloses in the primary cell wall of monocots and are found in smaller amounts in the primary cell walls of dicots. Secondary walls of dicotyledonous plants contain a significant amount of xylan. The basic structure of xylan has been reviewed [4]. Xylans have a backbone of /3-1,4-linked xylose residues (Fig. 25). The backbone is substituted by varioussidechainsattachedthrough 0 - 2 or 0-3 of the xylosyl residues. Terminal arabinofuranosyl residues are usually attached to 0-3 of the 4-linked xylosyl residues. The backbone is substituted by a-linked 4-O-methyl-/3-~-glucopyranosyl uronic acid on 0 - 2 of xylosyl residues and acetyl esters on 0 - 2 or 0-3. The degree of side-chain substitution determines the degree of solubility of the xylan and its ability to bind to cellulose. Xylans having a high degree of side chains are more water soluble and bind less tightly to cellulose, whereas molecules with fewer side chains are less water soluble and bind to cellulose tightly. Primary cell walls of gramineous monocots contain esterified ferulic and p-coumaric acids [S l]. Three feruloyl and two p-coumaroyl arabinoxylan oligosaccharides were isolated from bamboo shoot cell walls (Fig. 26) [76,82-841. Feruloylation and p-coumaroylation occur at 0 - 5 of the arabinofuranosyl side chain of xylan. It has been hypothesized that the feruloyl esters are subjected to peroxidase-catalyzed coupling (Fig. 27) to yield a diferuloyl group, thereby cross-linking the xylan molecules. Such “lateral” cross-linking of polysaccharides could have profound effects on the physical properties of the cell wall and thus on its ability to grow and to resist enzymatic digestion [28,81]. Feruloyl arabinoxylan oligosaccharides appear to have regulatory functions just like xyloglucan oligosaccharides. The feruloyl oligosaccharides inhibited auxin-stimulated and gibberellin-inducedelongationgrowth of rice cells[85,86]. An arabinoxylanoligosaccharide has no inhibitory effect. Ferulic acid itself has weak inhibitory effect. These results indicate that the feruloyl substituent of feruloyl oligosaccharides is necessary for the inhibitory effect, but the glycosyl portion of feruloyl oligosaccharide is also important for increasing this inhibitory activity.

202

lshii and Shimizu

UH

OH

I

OH

H *' CH,O

0

A-

(3)

FIGURE 26 Structure of feruloyl and p-coumaroyl arabinoxylan oligosaccharides obtained from bamboo shoot.

D. p-1,3- and p-lY4-Glucan (P-Glucans) The p-glucans are important cell wall components in monocots. The P-glucans consist almost entirely of p-1,3- and P-1,4-linked D-glucopyranosyl residues. The ratio of p-1,3top-1,4-links is between 1:2 and 1:3 [5,30].The usualarrangementoflinkages is for single 1,3-linked residuestoseparatesequences of two, three, orfour 1,4-linked residues.

203

Chemistry of Cell Wall Polysaccharides

OH

FIGURE 26

Continued

E. Glucomannans Glucomannans are major hemicelluloses of the secondary cell walls of gymnosperms, as well as being a minor component of angiosperm secondary walls. This has beenwell summarized in the reviews [2-41.

VII.

A.

CROSS-LINKS BETWEEN CELL WALL POLYMERS

Covalent Linkages

1. GlycosidicLinkage There is Some evidence for covalent bonding between xylan and pectin. Kat0 and Nevins isolated arabinoxylan-rhamnogalacturonancomplex from maize cell walls [881.

h)

0 P

Polysaccharide

I Polysaccharide I

P 2H20

H202

Peroxidase H&O

Ferulate

Diferulate

OH

0

I

i ‘0

Polysaccharide

FIGURE 27

Formation of diferulic acid cross-link between polysaccharides. nl

3 P

Chemistry of Cell Wall Polysaccharides TABLE 8

Some Possible Cross-Links Between

205 Wall Polymers Cleaving reagents stated)(aq. unless

mple Possible Cross-link

(a) Covalent Glycosidic

Ester (uronoyl) Borate di-ester Ester (uronoyl, etc.)

Phenolic coupling (1) Ether (2) Biphenyl

Ether Disulfide (b)Noncovalent Hydrogen bond

Ionic bond

Arabinogalactan .RC-I Xylan. xyloglucan Xylan. RC Pectin. cellulose RC-II'RG-I1 Feruloyl .pectin Feruloyl .arabinoxylan Feruloyl . xyloglucan ( 1) Extensin. (Tyr-Tyr) . extensin (2) Pectinpectin Xyloglucan * xyloglucan Arabinoxylan arabinoxylan PS. (ether). feruloyl-PS Cystine (R-SS-R') Hemicellulose. cellulose

Extensin. pectin

Calcium bridge

Hydrophobic interaction van der Waals bonds Lectin bond

Gelling of pectin (also involves H-bonds) (Many) Lectin. PS

Endoglycanases, hot acid, dry HF

Esterases, Na,CO,, NaOH, MeOH/NaOMe Weak acid Esterases, NaOH, MeOWNaOMe

NaCIOz a pH 4 and 70°C (diferulate also cleaved by NaOH, etc.)

NaC102?BBr,? Dithiothreitol, mercaptoethanol MMNO, KOH (urea, guanidinium thiocyanate, and heat are not very effective) Salts (LaCI, > CaClz > NaCI), acids, alkalis Chelating reagents, e.g., EGTA, CDTA, EDTA, oxalate, hexametaphosphate; low pH Organic solvents (Reagents that change molecular conformation?) Sugar hapten; denaturation

Abbr-evrattorls: Ara, arabinose; Fer, ferulate;Gal,galactose;Glc,glucose, Me, methyl;MeOH,methanol; NaOMe, sodium methoxlde; PS, polysaccharide; RG-I, rhamnogalacturonan-I; Rha, rhamnose; Tyr, tyrosine. Source: Modified from Ref. 28.

2. Diferuloyl Cross-Link The occurrence of ferulic and p-coumaric acids ester-linked to arabinoxylans in grasses [87], to pectic polysaccharides in spinach [39-411 and sugar beet [40], and to xyloglucan in bambooshoot [76] is well characterizedasdescribed in the previous section. The possibility of covalent linkages betweenesterified ferulic acid on wall polysaccharides was first proposed in 1971 by Geissman and Neukon [89]. The ferulic acid residues on feruloyl arabinoxylan from wheat flour have been cross-linked with peroxidase and hydrogen peroxidase to make a gel. This demonstrated that a dehydrogenative coupling between two esterified ferulic acid residues on arabinoxylan to form dehydrodiferulic acid had occurred (Fig. 27). Sugar beet pectin that contains ferulic acid esterified to arabinose and galactose residues [40] also make gel following peroxidase-catalyzed, oxidative cross-linking [90].

lshii and Shimizu

206

The phenolic coupling has been invoked to explain termination of cell expansion. Small amounts of diferulate have been detected by alkaline hydrolysis of cell walls. The goal of detecting an oligosaccharide fragment, cross-linked by a diferuloyl bridge, was achieved in 1991 [91]. The linkage group was isolated and characterized from bamboo shoot arabinoxylan, providing definitive evidence for the existence of diferuloyl ester cross-link (Fig. 28). Ralph et al. [92] isolated a series of ferulic acid dehydrodimers in addition to 5-5 coupled dehydrodiferulate from saponified grass cell walls (Fig. 29). These dehydrodimers (8-5, 8-0-4, and 8-8) also are involved in cross-linking of polysaccharides in cell walls. Isolation ofwallfragmentscontainingthesedehydrodimers is animportant challenge. Other possibilities for dimerization of phenolic acid substituents of polysaccharides exist. A series of homo- and heterocyclodimers of the cyclobutane type, formed by headto-tail or head-to-head association of ester-linked p-coumaric acid and ferulic acids, were isolated [93] (Fig. 30). A similar peroxidase-catalyzed cross-link may occur between tyrosine residues of extensin (Fig. 31). The phenolic ether linkage of isotyrosine is known to form intramolecularly within extensin, and may also occur intermolecularly [28].

3. Ester Bonds Cold Na2C03,which hydrolyzes ester bonds but does not cause p elimination-degradation, solubilized pectin that is not solubilized by chelatingreagent or EPC [9]. Theseester bonds may be methyl esterified galacturonosyl residues, diferuloyl bridges, or ester bonds between uronic acids and neutral sugars. Although intermolecular ester bonds have been proposed [94], their identification has not yet been achieved.

4. Borate Diol-Diester Cross-Linkage Borate cross-links two RC-I1 molecules in pectic polysaccharide to form an RG-I1 dimer (Fig. 16). This cross-link is extremely acid-labile. Treatment of the borate-RG-I1 complex with 0.5 N HCl at room temperature for 30 min cleaved the borate ester linkage. The borate cross-linkage might play an important role in connecting pectin networks in the cell wall and may be involved in the acid-induced elongation of cells during growth.

"

O

H

c=C-C-OH 11

5 ) - a - ~ - A r a f -(1+3 ) -p-D-xylp- ( 1 4) -D-xylp

0

YCO

FIGURE 28 A diferuloyl arabinoxylan hexasaccharide isolated from bamboo shoot cell walls.

Chemistry of Cell Wall Polysaccharides

HOv

207

0

“ O v O OH

M S 0

FIGURE 29

B.

Structure of dehydrodimers of ferulates.

NoncovalentLinkages

As cell wall polysaccharides are polyhydroxylic, many hydrogen bonds form in the walls. Multiple hydrogen bondsare present within cellulose microfibrils. Hydrogen bondsare probably responsible for the incorporation of xylan, xyloglucan, and glucomannan into the cell wall. The hydrogen bonding between cellulose and xyloglucan, and between cellulose and xylan, has been demonstrated in vitro. Ionicbonds will be formed in the case of polymers that contain charged groups. Homogalacturonan and 4-O-methylglucuronoxylan have negative charges, while extensin has a positive charge. Ionic binding may occur between these charged polymers in the wall. Negatively charged galacturonic acid residues in homogalacturonan and RG-I can form cross-linking with Ca’+ to form an “egg-box” [95]. About 15-20 contiguousga-

lshii and Shimizu

208

OH

H0

FIGURE 30 Structure of cyclodimer of ferulic acid.

lacturonic acid residues are needed in each chain to make a stable complex. Methyl esterification in homogalacturonan and rhamnosyl residues in RG-I interrupt the concerted binding.Furthermore, acetylation occurs at 0 - 3 of galacturonicacidresidues in homogalacturonan [333 and at 0-2 and 0-3 of galacturonic acid residues of RG-I, respectively [35-371. Therefore,egg-boxcross-linkage is likely to be limited in the wall, although isolated pectin and pectin in jam and jellies are known to give rise to gel-like structures in vitro.

VIII.

CELL WALL MODEL OF GROWING PLANT CELL

The cell wall has a variety of components that assemble to form an extremely complicated structure. Several wall models have been proposed. An early cell wall model was proposed by Albersheim and co-workers [96]. This model contained covalent links between xyloglucan and RG-I and hydrogen bonds between cellulose and xyloglucan, leadingto indirect cross-linking of cellulose microfibrils through a series of hydrogen bonds and covalent bonds in the matrix. Some of the details of this model have been disproven by subsequent chemical analysis [ 181. Capita and Gibeaut [97] and McCann and Robert [98] have proposed structural models for primary cell walls basedon this new information. It is probably impossible to describe all cell wall components and their interactions with a single and simple model. However, the existence of two principal polysaccharide networks in the growing cell wall have been proposed: a tension-resistant load-bearing cellulose/xyloglucan network and a compression-resistant pectic polysaccharide network.

IX.

CONCLUDING REMARKS

Thischapter briefly summarizespresentknowledgeof the pectic polysaccharidesand hemicelluloses. Primary cell walls commonly contain cellulose, xyloglucan, arabinoxylan, homogalacturonan, RG-I, and RG-11. These six polysaccharides account for all or nearly all of the primary wall polysaccharides, and their primary structures have been well conserved among species. The six polysaccharides are to some extent cross-linkedby covalent and noncovalent bonds, making up a complicated macromolecular network in the primary

Chemistry of Cell Wall Polysaccharides

+

e e,

e, D

0

rc

209

lshii and Shimizu

210

walls. This chapter also discusses the functions of primary cell walls. The recent discovery that oligosaccharide fragments derived from cell wall polymers in growing tissues can act as potent and specific regulators of gene expression is of importance to the biochemistry of all living systems [see refs. 34,99,100].

REFERENCES 1. 2. 3. 4.

5. 6. 7. 8. 9. 10. 11. 12. 13. 14. 15. 16. 17. 18. 19.

G. 0. Aspinal, Advan. Carbohyd,: Chem., 14: 429 (1959). T. E. Timell, Advan. Carbohyd,: Chem., 19: 247 (1964). T. E. Timell, Advan. Carbohydr: Chem., 20: 409 (1965). K. Shimizu, in Wood and Cellulosic Chemistry (D. N.-S. Hon and N. Shiraishi, eds.), Marcel Dekker, New York, p. 177 (1991). A. G.Darvill, M. McNeil, P. Albersheim, and D. P. Delmer, in The Biochemistry ofplants, 1 (N. E. Tolbert, ed.), Academic Press, New York, p. 91 (1980). P. Albersheim and A. G. Darvill, Sci. Am., 253(3):58 (1985). M. McNeil, A. G. Darvill, S. C. Fry, and P. Albersheim, Ann. Rev. Biochem., 53: 625 (1984). J. R. Thomas, M. McNeil, A. G. Darvill, and P. Albersheim, Plant Physiol., 83: 659 (1987). M. A. O’Neill, A. G. Darvill, and P. Albersheim, Methods Plant Biochem., 2: 415 (1990). B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 39 (1978). B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 51 (1978). B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 63 (1978). B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 79 (1978). B. W. Simson and T. E. Timell, Cellulose Chem. Technol., 12: 137 (1978). Y. Edashige, T. Ishii, T. Hiroi, and T. Fujii, Hol~orchung,49: 197 (1995). W. S. York, A. G. Darvill, M. McNeil, T. T. Stevenson, and P. Albersheim, Methods Erzzymol., 118: 3 (1986). R. R. Selvendran and M. A. O’Neill, Methods Biochem. Anal., 118: 3 (1986). P. M. Dey and K. Brinson, Advan. Carbohyd,: Chem. Biochem., 42: 265 (1987). C. Brettand K. Weldron,in Topics in PlantPhysiology, 2, Unwin Hyman, London, p. 4 ( 1990).

20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34.

35. 36.

G. 0. Aspinal, in Techniques of Chemistry, 4 (K. W. Bentley and G. W. Kirly, eds.), WileyInterscience, New York, p. 379 (1973). P. Albersheim, D. J. Nevins, P. D. English, and A. Karr, Carbohyd,: Res., 5: 340 (1967). S. Hakomori, J. Biochem. (Tokyo), 55: 205 (1964). A. G. Darvill, M. McNeil, and P. Albersheim, Plant Physiol., 62: 418 (1978). A. B. Blakeney and B. A. Stone, Carbohyd,: Res., 140: 3 19 (1985). I. Ciucanu and F. Kerek, Carbohydr: Res., 133: 105 (1984). A. Isogai, A. Ishizu, J. Nakano, S. Eda, and K. Kato, Carbohyd,: Res., 138: 99 (1985). P.-E. Jansson, L. Kenne, H. Liedgren, B. Lindberg, and J. Lonngren, Chern. Commun. Univ. Stockholm, No. 8 ( 1976). S. C . Fry, Ann. Rev. Plant Physiol., 37: 165 (1986). X. Qi, B. X. Behrens, P. R. West, and A. J. Mort, Plant Physiol., 108: 169 1 ( 1995). A. Bacic, P. J. Harris, and B. A. Stone, in Biochemistry of Plunts, 13 (J. Preiss, ed.), Academic Press, San Diego, CA, p. 297 (1988). E. Nothnagel, M. McNeil, and P. Albersheim, Plunt Physiol., 71: 916 (1983). G. 0. Aspinal, Polysaccharides, Pergaman Press, Oxford, U.K. (1970). T. Ishii, Plant Physiol., 113: 1265 (1997). A. Darvill, C. Auger, C. Bergmann, R. W. Carlson, J.-J. Cheng, S. Eberhard, M. G. Hahn, V.-M. L6, V. Marfi, B. Meyer, D. Mohnen, M. A. O’Neill, M. D. Spiro, H. van Halbeek, W. S. York, and P. Albersheim, Glycobiology,2: 181 (1992). P. Komalarilas and A. Mort, Curbohydr: Res., 189, 261 (1989). P. Lerouge, M. O’Neill, A. G. Darvill, and P. Albersheim. Cnrbohyd,: Res., 243, 359 (1993).

Chemistry of Cell Wall Polysaccharides

37. 38. 39. 40. 41. 42.

211

T. Ishii, Mokuzui Gakkuishi, 41, 561 (1995). Y. Edashige and T. Ishii, Phytochemistry, 42, 61 1 (1996). S. C . Fry, Biochem. J., 203, 493 (1982). T. Ishii, Plant Cell Physiol., 35, 701 (1994). T.Ishii and T. Tobita, Carbohyd,: Res., 248, 179 (1993). M. W. Spellman, M. McNeil, A. G. Darvill, P. Albersheim, and K. Henrick, Carbohydr. Res.,

122, 115(1983). 43. A. J . Whitcome, M. A. O’Neill, W. Steffan, P. Albersheim,and A. G. Darvill, Curbohyd,: Res., 271, 15 (1995). 44. V. Puvanesarajah, A. G. Darvill, and P. Albersheim, Curbohyd,: Res., 218, 21 1 (1991). 45. J. T. Thomas, A. G. Darvill, and P. Albersheim, Curbohydr. Res., 185, 261 (1989). 46. T. T. Stevenson, A. G. Darvill, and P. Albersheim, Curbohyd,: Res., 182, 207 (1988). J. Brasch, Curbohyd,: Res., 209, 191 (1991). 47. R. J . Regwell, L. D. Melton, and D. 48. T. Ishii, Mokuzui Gukkaishi, 41, 669 (1995). 49. M. Kobayashi, T. Matoh, and J. Azuma, Plant Physiol., 110, 1017 (1996). 50. T. Ishii and T. Matsunaga, Curbohydr. Res., 284, 1 (1 996). P. Pellerin, T. Doco, A. G. Darvill, and P. Alber51. M. A. O’Neill, D. Warrenfeltz, K. Kates, sheim, J. Biol. Chem., 271, 22923 (1996). Phytochemistry, 44, 243 (1997). 52. S. Kaneko, T. Ishii, and T. Matsunaga, 53. B. J. Shelp, in Boron and Its Role in Crop Production (U. C. Gupta, ed.), CRC Press, Boca Raton, FL, p. 58 ( 1992). 54. W. D. Loomos and R. W. Durst, Bio Factor, 3, 229 (1992). 55. H. 0. Bouveng, Acta Chem. Scand., 19, 953 (1965). G. J. Voragen, Carbohyd,: Res., 279, 265 56. H. A.Schols,E. J. Bakx,D.Schipper,andA. (1995). 57. I. Eriksson, R. Andersson, E. Westerlund, R. Andersson, and P. Aman, Curbohyd,: Res., 281, 161 (1996). G. Darvill,and P. Albersheim, Curbohydr: Res., 243, 373 58. P. Lerouge,M.A.O’Neill,A. ( 1 993). 59. M. McNeil, A. G. Aman, and P. Albersheim, Plant Physiol., 70, 1586 (1982). 60. J. M. Lau, M. McNeil, A. G. Darvill, and P. Albersheim, Curbohyd,: Res., 168, 245 (1988). 61. S. Eda, K. Miyabe, Y. Akiyama, A. Ohnishi, and K. Kato, Curbohyd,: Res., 158, 205 (1986). 62. A. M. Stephen, in The Polysaccharides, 2 (G. 0. Aspinal, ed.), Academic Press, New York, p. 97 (1980). 63. H. Du, A. E. Clarke, and A. Bacic, Trends Cell Biol., 6, 413 (1996). 64. T. Hayashi, AIUIU.Rev. Plant Physiol. Plant Mol. Biol., 40, 139 (1989). 65. S. C. Fry, J. Exp. Bot., 40,1 ( 1989). 66. S. C.Fry, W. S. York,P.Albersheim,A. G. Darvill, T. Hayashi, J.-P. Joseleau, Y. Kato, E. P. Lorences, G. A. Maclachan, M. McNeil, A. J. Mort, J . S. Reid, H. U. Seitz, R. R. Selvendran, A. G. J. Voragen, and A. R. White, Physiol. Plant, 89: 1 (1993). 67. W. S. York, J . E. Oates, H. van Halbeek, A. G. Darvill, and P. Albersheim, Carbohyd,: Res., 173, 113 (1988). 68. L. L. Kiefer. W. S. York, A. G. Darvill, and P. Albersheim, Phytochemistry, 28: 2105 (1989). 69. L. L. Kiefer. W. S. York, P. Albersheim, and A. G. Darvill, Carbohyd,: Res., 197, 137 (1990). 70. M. Hisamatsu, G. Impallomeni, W. S. York,P.Albersheim,andA. G. Darvill, Curbohyd,: Res., 211, 117 (1991). 71. M. Hisamatsu, W. S. York. A. G. Darvill, and P. Albersheim, Carbohyd,: Res., 227.45 (1992). 72. W. S. York,H.vanHalbeek,A. G. Darvill,and P. Albersheim. Carbohyd,:Res., 200. 9 (1990). 73. W. S. York, G. Impallomeni,M.Hisamatsu, P. Albersheim,andA. G. Darvill, Carbohyd,: Res., 267, 79 ( 1995). 74. W. S. York. V. S. K.Kolli, R. Orlando, P. Albersheim, and A. G. Darvill, Cur-bohydr. Res., 285. 99 (1996).

212

lshii and Shimizu

75. 76. 77. 78. 79. 80. 81.

T. Ishii, J . R. Thomas, and T. Hiroi. Phytocher,~i.st~~, 29: 1999 (1990). T. Ishii and T. Hiroi, Curbohyrlr. Res., 206, 297 (1990). S. Aldington, G. J. McDougall, and S. C. Fry, Plant Cell, u r d Enr~iromnent,14: 625 (1991). W. S. York, A. G. Darvill, and P. Albersheim, Plant Physiol., 75: 295 (1984). G. J. McDougall and S. C . Fry, Plmta, 175: 412 (1988). G. J . McDougall and S. C . Fry, Plant Physiol., 93: 1042 (1990). J. Ralph and R. F. Helm, in Foruge Cell Wull Structure und Digestibility (H. G. Jung, D. R. Buxton, R. D. Hatfield, and J. Ralph, eds.), American Society of Agronomy, Madison, WI, p. 201 (1993). T. Ishii and T. Hiroi, Crrrbohydr: Res., 196: 175 (1990). T. Ishii, Phytochetnistry, 30: 23 17 (1991). T. Ishii, Mokrczai Cakkuishi, 42: S15 (1996). T. Ishii and H. Saka, Pltrnt Cell Physiol., 33: 321 (1992). T. Ishii and T. Nishijima, Plant Cell Physiol., 36: 1447 (1995). K. Iiyama, T. B.-T. Lam, and B. A. Stone, Plunt Physiol., 104: 3 IS (1994). Y. Kato and D. J. Nevins, Curbohydr: Res., 227: 315 (1992). T. Gerssman and H. Neukom, Helv. Chern. Acttr, 54: I108 (1971). F. M. Rombouts, J. F. Thibault, and C. Mercier, French Patent Appl. 83,07,208; Chern. Abstt;

82. 83. 84. 85.

86. 87. 88. 89. 90.

102: 60792u (1983). 91. T. Ishii, Curbohydr. Res., 21 9: IS ( 1 991). 92. J. Ralph. S. Quideau. J . H. Grabber, and R. D. Hatfield, J . Chern. Soc. Perkin Truns.. I : 3485 ( 1994). 93. W. H. Morrison, 111. R. D. Hartley, and D. S. Himmelsbach. J. Agric. Food Chem., 40: 766 ( 1992). 94. J. A. Brown and S. C. Fry, Plont Physiol., 103, 993 (1993). 95. D. A. Powell, E. R. Morris, M. J. Gidley, and D. A. Rees, J . Mol. Biol.. 155: 5 17 (1982). 96. P. Albersheim, M . Rev. Bioc.hern., 16: 127 (1978). 97. N. C. Carptia and D. M. Gibeaut, Plont J., 3: 1 (1993). 98. M. C . McCann and K. Roberts, J . Exp. Bot., 45. 1683 ( 1994). 99. S. A. Ldington and S. C. Fry, A h . Bot. Res., 19: I (1993). 100. F. CBte and M. G. Hahn, Plunt Mol. R i o / . , 26: 1379 ( 1994).

Chemistry of Extractives Toshiaki Umezawa Kyoto University, Kyoto, Japan

1.

INTRODUCTION

Extractives are the wood constituents which can be extracted with neutral solvents. They are obtained by extracting wood meal with organic solvents or water or by steam distillation, and some are obtained as exudates from wounded trees. The amountofextractives is small,generallyupto 5-10% in the woodin the temperate zone. However, in some tropical woods relatively high amounts of extractives are found [l]. Amongwood species, differences of chemical structures of three major cell wall components, cellulose, hemicellulose, and lignin, are few. However, a great diversity in extractive composition is found throughout wood species. Although the extractives are low in concentration compared with those of the cell wall polymers, this fraction characterizes eachwoodspecieschemically.Mostcomponents of woodextractives are classified as secondary metabolites, and the distribution of specific compounds is restricted in certain wood species. This feature provides the basis of chemotaxonomy of woody plants. Furthermore, individual compounds are often found in specific tissues of individual trees, and their amounts can vary from season to season even in the same tissue. Many phenolic compounds are accumulated in heartwood, whereas they are found only in trace amounts in the corresponding sapwood. Extractives are the predominant contributors to woodcolor, fragrance, and durability. Extractives also influencethe pulping, drying, adhesion, hygroscopicity, and acoustic properties of wood.Manyextractiveshave specific biological activities, andvariouswoods have been used as sources of crude drugs and medicines for centuries. Recent progress of the biosynthesis of secondary metabolites of woody plants including heartwood components [2] has suggested the possibility of biotechnological control of their biosynthesis. This may lead to biotechnological production of biologically active extractives, and to elucidating molecular mechanisms of heartwood formation. Progress in the chemistry and biochemistry of natural products including wood extractives has been reviewed in a number of articles, such as in Natural Product Reports. Thisjournalincludesregularreviews of the relevant literature publishedduringwelldefined periods with respect to individual topics in the fields.

213

214

Umezawa

II. LIGNANS, NEOLIGNANS, AND RELATED COMPOUNDS Lignans and neolignans are phenylpropanoids that occur in many plants including softwoods, hardwoods, and medicinal plants [3-51. The term “lignan” was introduced by Haworth to describe a group of phenylpropanoid dimers, where the phenylpropane units were linked by the central carbon (C8) of their side chains [6]. Gottlieb coined “neolignan” for compounds containing two phenylpropane units that are linked otherwise than C8-C8’ [7]. Later, neolignans were redefined as the dimers of allyl- or propenylphenyls, while lignans were regarded as the dimers of cinnamyl alcohols [8]. However, in this review Haworth’s definition of lignans [6] and Gottlieb’s former definition of neolignans [7] will be used, because the original definitions are being applied by most researchers. Tri- and tetramers of phenylpropanoid units are referred to as sesqui- and dilignans, respectively. Lignans are classified into several subgroups:dibenzylbutanes,dibenzylbutyrolactones, furans, furofurans, aryltetralins, arylnaphthalenes, and dibenzocyclooctadienes [3,5] (Fig. I). Lignans often occur as glycosides. Examples of neolignans are shown in Fig. 2. Chemical structures of lignans and neolignans are similar to that of lignins. Thus (+)-pinoresin01 (5) and the corresponding (-)-enantiomer are typical lignans, while this structure is also involved in a lignin molecule as a substructure (i.e., pinoresinol or p-p’ substructure) [9]. Also the structures of neolignans are similar to lignin substructures of p-0-4 (arylglycerol p-aryl ether), 5-5’ (biphenyl), and p-5 (phenylcoumaran) types [IO]. In spite of the structural similarities, lignins differ sharply from lignans and neolignans in terms of optical activity. The former is optically inactive, while the latter is optically active, suggesting the difference in stereochemical mechanisms in their biosyntheses [9]. Following the first example of an in-vitro enantioselective formation of an optically pure lignan, (-)-secoisolariciresinol (l),with an enzyme preparation from Forsythia inin lignan biosynthesis termedia [ 1 l], much investigation has been done for enzyme systems of Forsythia spp. [9,12,13]. Figure 3a shows the conversion of coniferyl alcohol (14) into lignans with Forsythia enzymes. Recently, cDNA cloning of pinoresinol/lariciresinol reductase which catalyzes reduction of (+)-pinoresin01 (5) and (+)-lariciresinol (4) to (+)lariciresinol (4) and (-)-secoisolariciresinol (l),respectively, has been reported [ 141. In addition, enantioselective coupling of two coniferyl alcohol radicals giving rise to (+)pinoresinol (S) was reported [15]. The coupling was highly enantioselective only in the presence of dirigent protein isolated from Forsythia plant. Formation of (+)-secoisolariciresinol (lS), the oppositeenantiomerto the one in Forsythia spp.,with an enzyme preparation from Arcfiurn lappa petioles, has been reported (Fig. 3b) [16]. On the other hand, an enzyme preparation from seeds of A. lappa was recently found to catalyze enantioselective formation of (-)-secoisolariciresinol (1) from coniferyl alcohol (14), indicating that two enzyme preparations from different organs of a single plant species can catalyze the selective formation of different enantiomers of a lignan [ 171. The structural similarity with lignins suggests that neolignans maybe synthesized by enantioselective coupling of two phenolic phenylpropane units. Recently, two examples of enantioselectiveformation of neolignansfrom coniferyl alcohol (14) werereported [18,191. Lignanshavesuch biological activities asantitumor[podophyllotoxin (7) and steganacin (9)], antimitotic [podophyllotoxin (7)],antioxidant (nordihydroguaiaretic acid andsesaminol), and antiviral [podophyllotoxin (7) againstcytomegalovirusandherpes simplex l-virus, and (-)-arctigenin (3) against human immunodeficiency virus], etc. [3-

215

Chemistry of Extractives

Furan

Dibenzylbutane Dibenzylbutyrolactone

OH

H0 OCH3 -Secoisolariciresinol (-)

1

R=E (-) "atairesinol 2 Re83 (-) -Arctigenin 3

Aryltetralin

E'urofuran

OCH3

OCH3 R=E (+) -Pinoresin01 5

Podophyllotoxin 7

R=OC83 (+) -Syringaresinol 6

DibenzocycloGlycoside octadiene

Steganacin 9 FIGURE 1

Arctiin

10

Examples of lignans.

OCH3

Kadsurenone 11 FIGURE 2

(+) -Lariciresinol 4

Magnololl2 Eonokiol13

Examplcs o f neolignans.

Diphyllin

8

Umezawa

216

%" Coniferyl alcohol 14

(+) -Pinoresinol 5

(+) -Larici-

I

J

(-) -Secoisolariciresinol resinol 1

(-) "atai2

(-)

resinol 4

--ct

igenin 3

(b)

H3CO H O T : :

OCH3 Coniferyl OH

alcohol 14

\

0

OH

O C H ~ (+) -Secoisolariciresinoll5

FIGURE 3 Enzymatic lignan formation.

5,201. Antagonism toward the platelet-activating factor (veraguensin) and inhibitory activities toward certain enzymes have also been detected in many lignans [3-5,201. Among the biologically active lignans, antitumor podophyllotoxin (7)has attracted particular interest [see also Section XLC]. Mammalian lignans are known to be produced from plant lignans such as secoisolariciresinol diglucoside by the action of bacterial flora in the colon of human or animals. The mammalian lignans and their precursor secoisolariciresinol diglucoside have a protective effect on the promotion stage of mammary tumorigenesis [21], and are receiving widespread interest. A neolignan, kadsurenone (ll),has antagonistic activity to platelet-activating factor [22].Honokiol (13) is antibacterial and antifungal, and magnolol (12) has antimicrobial and muscle-relaxant activities [23,241.

111.

NORLIGNANS

Norlignans are diphenylpentane (C,-C,)-(Cz-C,) compounds, and typical examples are shown in Fig. 4. Although the biosynthetic sequences for norlignans have not been elucidated, the structures are seemingly composed of the phenylpropane (C,-C,) and phen-

217

Chemistry of Extractives

OH

Hinokiresinol 16

Sequirin (Sequirin-C) 18

Agatharesinol 17

Sugiresinol Hydroxysugiresinol Cryptoresinol (Sequirin-B) 20

21

( Sequirin-A) 19

HdHaWO H0 OH

Pueroside-A OH 22

H O f/l o H

cis-Hinokiresinol 23

\3

OH

Yateresinol Sequirin-D nyl)-2-cyclo-pentene24 25 FIGURE 4

2,3-Bis(p-hydroxy-phe l-one 26

Examples of norlignans.

ylethane (C,-C,) units connected via C8-C7' [e.g., hinokiresinol (16)] and in some cases C8-C8' [e.g., yateresinol (24)] and C9-C8' [e.g., sequirin-D (25)](Fig. 4). Most norlignans have been isolated from softwoods belonging to Cupressaceae, Taxodiaceae, and Araucariaceae [25-271, while some norlignans occur in herbaceous plants. cis-Hinokiresinol (23) was isolated from Anen7arrl1et~a crspI7ndt.loide.s (Liliaceae) 1281, and two norlignan glycosides, pueroside-A (22) and -B, were isolated from PuercrriLI lohtc1 (Leguminosae) [29] (Fig. 4). No concrete experimental evidence has been reported for the origin of the two aromatic nuclei of a norlignan molecule. Based on structural considerations, several possible schemes have been proposed for the biosynthesis of norlignans, involving coupling of two phenylpropane units followed by loss of one carbon atom 125,301. Norlignans are of particular interest with respect to heartwood coloration. The coloration of heartwood of Japanese cedar (Cty>tornc~ria japonica) is due to polymerization

Umezawa

218

of norlignans, e.g., hydroxysugiresinol (20) and sequirin-C (18) [31,32], andthat of hinoki cypress (Charnaecyparis obtusa) is related to hinokiresinol (16) [33]. Inhibitory activity on cyclic adenosine-3’,5’-monophosphatephosphodiesterase was reported for cis-hinoki resinol (23) [28]. Sugiresinol (19) and hydroxysugiresinol (20) have inhibitory effect on polymerization of methyl methacrylate [34].

IV.

FLAVONOIDS

Flavonoids are diphenylpropane (C,-C&,) compoundswhicharecomposedof the C,-C, (phenylpropane) fragment derived from the shikimate-cinnamate pathway and the C , fragmentderivedfrommalonyl-CoA (40). Flavonoids are classified into flavanones, flavones, chalcones, dihydroflavonols (flavanonols), flavonols, aurones, flavan-3-01s (catechins), flavan-3,4-diols (leucoanthocyanidins),anthocyanidins, isoflavonoids, andneoflavonoids (Fig. 5). The term “flavonoids” in the strict sense is sometimes applied to those except for isoflavonoidsandneoflavonoids.Proanthocyanidins(condensedtannins) are oligomers and polymers of polyhydroxyflavan-3-01 units [35] (see also Section IX). Typical examples of flavonoids are shown in Fig. 5. Flavonoids occur widely in the plant kingdom, and presentwidelyor specifically in barks, heartwoods, flowers, fruits, seeds, roots, etc. Flavonoids reported to occur in wood or bark are listed in Ref. [36], and the chemistry of flavonoids is reviewed in a number of works [36-411. Chalcone is biosynthesizedfromcinnamoyl-CoAs [especially, p-coumaroyl-CoA (39)], which is formed via the shikimate-cinnamatepathway,and three moleculesof malonyl-CoA (40). The transformation is catalyzed by chalcone synthase as shown in Fig. 6.6’-Deoxychalcone [e.g., isoliquiritigenin (30)] is likewisesynthesizedfrommalonylCoA (40) and 4-coumaroyl CoA (39) by chalcone synthase in coaction with a NADPHdependent polyketide reductase. The chalcones are the immediate precursors for all flavonoids (in the strict sense). Thechalcones are converted to flavanones, flavones, dihydroflavonols, flavonols, leucoanthocyanidins, anthocyanidins and their glycosides (anthocyanins), catechins, and aurones [42,43]. The enzymes which are responsible for these conversions and the genes encoding these enzymes have been well characterized [42,44481. Formation of isoflavonoidsinvolves the rearrangement of aphenylgroupof the flavanoneskeleton[42,46].On the otherhand, little is knownaboutbiosynthesisof neoflavonoids. Flavonoids have various biological activities 137,381. Thus, symbiotic nitrogen-fixing bacteria recognize flavonoids as signals for the activation of their nodulation genes [49]. Isoflavonoids are the major structural class of phytoalexins in legumes [49-511. Anthocyanins occur as flower pigments [52,53]. Green teas contain significant amounts of catechins which have an antioxidant activity [54]. The effects of flavonoids on mammalian biology are reviewed [55,56]. Flavonoids seem to protect plants from ultraviolet-induced injury 1571. Manyflavonoids are biosynthesized in response to externalstresses, e.g., ultraviolet light, microbial attack, and physical injury. Hence, the flavonoid biosynthesis is a metabolic event which is suitable to investigate stress-gene expression relation. Heartwood formation which does not occur in herbaceous plants is one of the metabolic events specific to woody plants, but little is known about its biochemical mechanisms. However, this metabolic event involves or is accompanied by deposition of significant amounts of secondary metabolites, so-called heartwood extractives suchas flavonoid, stilbene, lignan, norlignan, etc. Seasonal changes and site specificity of chalcone synthase activity have been examined in relation to heartwood flavonoid synthesis [58,59].

Chemistry of Extractives

219

Qfy

Flavanone

Cbalcone

Flavone

\

0

H OH 0 Sakuranetin 27

Dihydroflavonol (Flavanonol)

\

I OI

S

OH 0 Apigenin 28

H

O

\

R

S

0

R=on, Chalconaringenin 29 R=H, Isoliquiritigenin 30

Aurone

Flavonol

OH 0

H o e : : I

'

I

OH

OH 0

Dihydrokaempferol 31

Flavan-3-01 (Catechin)

Quercetin 32

OH

0

Sulfuretin 33

Flavan-3,I-diol (Leucoanthocyanidin) Anthocyanidin

OH

OH OH

H \o d : :

H o e OHo H

OH OH

(+)-Catechin 34

OH OH

Leucodelphinidin 35

Isoflavonoid

Neoflavonoid

% Daidzein 37

FIGURE 5 Examples of flavonoids.

OH Pelargonidin 36

Umezawa

220

p-Coumaroyl-CoA 39

Flavonoids

HOOC>SCoA

0 Malonyl-CoA 40 FIGURE 6 Biosynthesis of flavonoids and stilbenes.

Many genes of the enzymes involved in flavonoid synthesis have been cloned, and mechanisms of the gene expression have been investigated intensively [47,48]. Thus, flavonoids are one of the best-understood groups of plant secondary metabolites, especially in terms of biosynthesis. It should be noted that extensive genetic information available about flavonoids, especially anthocyanin pigments of flowers, has accelerated significantly the biosynthetic studies in this field.

STILBENES

V.

Historically, the termstilbene referred to compoundspossessing the 1,2-diphenylethene structure,butnowadays the newlydiscoveredbibenzylsandphenanthrenes,whichare composed of C,-C,-C, skeleton, are also involved in this group. Stilbenes occur in the Pinaceae, Moraceae, Betulaceae, Leguminosae, etc. [60]. Typical examples of this class are shown in Fig. 7. Stilbenes are elaborated from CoA esters of cinnamic acids, and there is a similarity in the biosynthesis of stilbenes with that of flavonoids (Fig. 6). Stilbene synthases catalyze

Resveratrol

Pinosylvin

Elydxangeic acid

42

I

OH

Lunularin

Batatasin I

45

FIGURE 7 Examples

or stilbenes.

Chemistry of Extractives

221

condensation of CoA estersof cinnamic acids [e.g., cinnamoyl-CoA and p-coumaroyl-CoA (39)J with three molecules of malonyl-CoA (40), as in chalcone synthesis catalyzed by chalcone synthase [2,61]. However, the cyclization of polyketide moiety of the C& polyketocarboxylic acid (41) occurs in a different wayfrom that by CHS to give rise to stilbenes (Fig. 6); resveratrol (42) and pinosylvin (43) are formed with elimination of one carbon atom (Fig. 6),while hydrangeic acid (44) is formed without the elimination [61,62]. Although pinosylvin (43) and pinosylvin monomethyl ether occur in sound heartwood of Pinus spp., they are formed as a response to stress such as fungal infections or UV light [601. Hence, the role of stilbene in decay resistance and induction of stilbene synthesis has attracted much attention [60]. A stilbene synthase from UV-stressed seedlings of Pinus sylvestris has been purified and characterized [63], while a stilbene synthase gene from grapevine (Vitis vintfera) was transferred to tobacco, and the regenerated plants were found to display increased resistance to Botrytis cinerea [64]. Pinosylvin (43) and pinosylvin monomethyl ether are also known as inhibitors of sulfite pulp cooking [65].

DIARYLHEPTANOIDS

VI.

Diarylheptanoids are composed of two phenyl rings connected with a C , carbon chain (Fig. 8). Many of this type of compound are isolated from plants belonging to the Betulaceae and Zingiberaceae. Besides these two families, the occurrence of diarylheptanoids in the following species was also reported: Centrolobium spp. (Leguminosae), Myrica spp. (Myricaceae), A c e r spp. (Aceraceae), and Garuga spp. (Burseraceae). Recently the chemistry and biological activity of diarylheptanoids have been reviewed [66,67].

OH OH

0 0

Ilannokinol 47

Curcumin 48

OH

H0

P l a t y p h y l l o s i d a 49 0

"'OH H 0 OH

Asadanin 50 FIGURE 8

Acerogenin

Examples of diarylheptanoids.

A 51

Urnezawa

222

Two different results of feeding experiments have been reported regarding biosynthetic precursors of diarylheptanoids. One suggested that one aromatic ring of curcumin (48) is derived from a cinnamate unit and the other from acetate (or malonate), based on administration of [ I4C]acetate, [I4C]malonate, and [14C]phenylalanine to Curcuma longa [68]. Other studies suggestedthat diarylheptanoids, acerogeninA (51), and platyphylloside (49) were derived from two phenylpropane units and one acetate (or malonate)unit, based on administration of [“C]cinnamate, [14C]phenylalanine, [ I4C]acetate, and [ 14C]malonate to Acer nikoense [acerogenin A (51)] [69], and of [14C]cinnamate and [I4C]malonate to Betula platyphylla [platyphylloside (49)] [70].

VII.

ISOPRENOIDS

lsoprenoids is the generic name of compounds composed of isoprene (C,H,) units connected linearly or cyclically. Isoprenoids consist of terpenoids, steroids, and tropolones. Since the chemistry of terpenoids, steroids, and tropolones has been developed independently in spite of their close relationship in biosynthesis, the three classes are usually treated separately. Terpenoids are divided into monoterpenes (C,,,), sesquiterpenes (C,,), diterpenes (C?”), sesterterpenes(C,,), triterpenes (C,,,), tetraterpenes (C,,), and polyterpenes (C,,,), depending on the number of the constituent isoprene (C,) units. Each subclass of terpenoids is further classified into many groups of different carbon skeletons. The terpenoid compounds are generally elaborated via the mevalonate pathway as outlined in Fig. 9. Although the mevalonate pathway has generally been accepted, a novel pathway concerning the early steps of isoprenoid biosynthesis toward isopentenyl pyrophosphate (56) has recently been demonstrated [71]. The novel pathway, which involves the condensation of a triose phosphate with activated acetaldehyde, has been characterized in several different bacteria [71,72]. In addition,Eisenreichetal.haveshown that the taxane carbon skeleton is not of mevalonate origin in Taxus chinensis [73] (see also Section X1.C). Isoprenoids are the largest group among plant secondary metabolites, and occur in a huge number of plants including woody plants. It is beyond the scope of this book to list all the plants producing this class of compounds and to describe their biosynthetic schemes. Comprehensive lists of the compounds, their biosynthesis, and biological activities are summarized elsewhere [2,37,74-771.

A.

Terpenoids

Specific fragrances of different woods are usually due to the composition of monoterpenes and volatile sesquiterpenes. They can be easily separated from wood by steam distillation, and the oily substance obtained is called “essential oil.” Turpentine, essential oil from Pinus spp., is obtained by steam distillation of exudates from pine trees (oleoresin); the residue is gumrosin,which is composedmainly of diterpene acids (rosin acids), e.g., abietic acid (91). Turpentines obtained from pine wood and those recovered from kraft pulp waste liquor are called wood turpentine and sulfate turpentine, respectively. Rosins are used for sizing of papers. Monoterpenes are derivedfromgeranylpyrophosphate (58). They are subdivided into acyclic and cyclic monoterpenes (Fig. IO) a-Pinene (64) and @-pinene (65) are major components of turpentine.

223

Chemistry of Extractives

y SCoA

y SCOA Acetyl-coA

Acetyl-coA 52

52

Ki,SCOA Acetoacetyl-coA 0 ,o

53

3-Eydro~y-3-methylglutaryl-CoA 54

Mevalonic

- JJPP

acid55

pyrophospxate 1 56

1

Isopenten

uopp

"-l l

Dimethylallyl pyrophosphate 57

I

OPP

Monoterpenoids

M Geran

1 pyropiosphate 58

Sesquiterpenoids Squalene \ TriFarnesyl pyrophosphate 59 terpenoids

'

.

Diterpenoids \ Tetraterpenoids c

G e r m lgeranyl 'Phytoene pyropxosphate 60

Sesterterpenoids Getan lfarnesyl pyropiosphate 61

Polyterpenoids FIGURE 9 Generalscheme of terpenoid biosynthesis. PP: pyrophosphategroup.

Sesquiterpenes are derived fromfarnesyl pyrophosphate (59), and constitute the largest class of terpenoids [74]. Some 120 distinct skeletal types of sesquiterpenes are known [74]. Figure 11 depicts the important sesquiterpene skeletal types from acyclic (e.g., farnesane) to tricyclic (e.g., thujopsane), which are often encountered as wood constituents. Qpical examples of each type of sesquiterpenes are also shown under the corresponding skeletal types in Fig. 11. Diterpenes are derived from geranylgeranyl pyrophosphate (60), and some 130 distinct skeletal types are reported [74]. Figure 12 shows the typical diterpene skeletal types

Urnezawa

224

P-Myrcene

(-) -Citronellol

62

63

(-) 64

-a-Pinene

(-) +-Pinene

65

(- ) -Limonene

1,B-Cineol

67

68

(+) -Camphor 66

FIGURE 10 Examples of monoterpenes.

with corresponding examples. A phytane diterpene, plaunotol (87), has anti-ulcer activities [78]. ent-Gibberellane diterpenes, gibberellins, are important plant hormones. Occurrence of sesterterpenes in higher plants is highly limited. Triterpenes are elaborated from squalene formed via tail-to-tail coupling (coupling between pyrophosphate ends) of two farnesyl pyrophosphate units 121. Two major classes of nonsteroidal triterpenes are tetracyclic and pentacyclic [74]. Examples of typical skeletal types of this class are shown in Fig. 13: lanostane, dammarane,euphane,limonoids [tetranor (C,,) compounds], quassinoids (mainly Czo compounds), lupane, and oleanane. Figure 13 also shows examples of compounds belonging to each skeletal type. Oleananes often occur as aglycons (sapogenins) of saponins [76].

B. Steroids Steroids, which are also derived from squalene, are compounds with cyclopentanoperhydrophenanthrene skeleton and their congeners elaborated from them. The basic structure of steroids is shown in Fig. 14. Positions on the same side as the angular methyl (18- and 19-CHJ are denoted as p, and those on the opposite side are denoted as a. Substituents on the 8, 9, 10, 13, 14, and 17 positions of steroids are projected to Sp, 9a, lop, 13P, 14a, and 17P positions, respectively. p-Sitosterol (or sitosterol) (108) (Fig. 14) is widely distributed in the plant kingdom. a- and P-Ecdysones isolated from silk worm (Bombyx rnori) have steroid skeletons, and are well known as molting hormones. Compounds with similar structure to the ecdysones were isolated from plants (e.g., Podocarpus and Taxus spp.), and are referred to as phytoecdysones [e.g., ponasterone A (109)] (Fig. 14) [79].

C. Tropolones Tropolones are nonbenzenoidaromaticcompoundshaving a seven-memberedenolone structure. They occur in Cupressaceae plants and exhibit antimicrobial activity. Examples

225

Chemistry of Extractives

B i s a b o lFaanren e s a n e

trans-Farnesol 69

(+) -P-Bisabolene 70

(+) -Juvabione

71

Cuparane

Caryophyllane Germacrane

Germacrone (+) -Costunolide 72 73

+-Caryophyllene 74 (-)

(+) -Cuparena

75

Eudesmane

OH

P-Eudesmol 76

(-) -CryptoT-Cadinol meridiol 78 77

FIGURE 11 Examples of sesquiterpenes.

of this class are shown in Fig. 15a. Hinokitiol(P-thujaplicin) (111) was isolated from Charnaecypuris ruiwunmsis by Nozoe [80,81]. a,P, and y-Thujaplicins, (110), (lll),and (112), were isolated from Thuju plicufu by Erdtman [82]. Tropolones are composed of 10 or 15 carbon atoms, and they have been regarded as mevalonate origin, i.e., a subclass of isoprenoids. Recently this has been supported by feedingexperiments.[“C]Mevalonic acid wasfound to beincorporatedinto hinokitiol (111) in suspensioncultures of Cupressus lusitunicu, suggesting that hinokitiol (111) is elaborated via the mevalonate pathway [83].

226

Umezawa

Guaiane

(-)

-Guaiol

79

Aramadendrane

Himachalane

(+) -Aromadendrene

(- 1 -a-Himachalene

80

81

Acorane

Chamigrane

Cedrane

Thuj o p s a n e

q y-Acora-

P-Chamigrene

diene

83

a-Cedrene 84

Thujopsene 85

82

FIGURE 11 Continued

On the other hand, the tropolone structure and aromatic ring of colchicine (114) (Fig. 15b),which is analkaloidfrom Colchicum autumnale, arederivedfromtyrosineand phenylalanine, respectively [84]. Thus colchicine (114) is a phenylpropanoid compound butnot an isoprenoid. In addition,tropolonesoccurring in microorganismsarebiosynthesized via the acetate-malonate pathway [85].

VIII.

QUINONES

Various types of quinones occur in many plant families, and most of them are benzoquinones, naphthoquinones, or anthraquinones. Most of the quinones found in nature are p quinones, but o-quinones also exist [86]. Spica1 examples of this class are shown in Fig. 16. Quinones are biosynthesized via various pathways, i.e., the shikimate, the mevalonate, and the acetate-malonate (polyketide) pathways [2]. Quinones are pigments and have various biological activities. Juglone (116), which occurs in black walnut (Juglans nigra), is skin-irritating [87-891. This compound is also well known as an allelochemical [90]. Tectoquinone (117) and related compounds have strongantitermite activity [91]. Mansonone A (118) and its congenercause allergies [89,92].

227

Chemistry of Extractives

Phytane

LJwwLL OH

m

HO-

OH

Plaunotol87

Phytol 86

Labdane

(+) -transCommunic acid 88

Pimarane,

Isopimarane

(-) -Sandaracopimaric acid

(+) -Pimaric acid 89

90

Nagilactone

Abietane

& p!?

COOH

(-) -Abietic acid

@ H 0

(+) -FerruginolInumaki-Nagilactone 92 lactone A 94

91

A

93

FIGURE 12 Examples of diterpenes.

IX. TANNINS Tannins are water-soluble phenolic compounds having molecular weights between500 and 3000. Besides giving the usual phenolic reactions, they have special properties such as the ability to precipitate alkaloids, gelatin, and other proteins [93,94]. This class of compounds also has high astringency, and gives blue or green coloration with femc chloride. Tanninsare distributed widely in wood, bark, andleaves of many plants. Tannins are classified into hydrolyzable and condensed tannins.

Umezawa

228

ent -Kaurane

ent-Gibberellane

Ginkgolide

COOH (-)

-Kaurene

95

Gibberellin

A19

Ginkgolide B 97

96

Taxane

10-Deacetylbaccatin I11

Taxol 99

98

FIGURE 12 Continued

Hydrolyzable tannins (Fig. 17a) are esters of an aliphatic poly01 and phenolic acids (Fig. 17b), and can be hydrolyzed into the components. As shown in Fig. 17a, galloyl, hexahydroxydiphenoyl, and depside galloyl groups are esterified to the polyol, generally D-glucose. Hydrolyzable tannins that give gallic acid (120) by hydrolysis are referred to as gallotannins, while compounds that afford ellagic acid (122) are referred to as ellagitannins. Hexahydroxydiphenic acid (121) is lactonized to give rise to ellagic acid (122) in the hydrolysis. Some ellagitannins [e.g., casuarinin (119)] have C-glycosidic structures. Two different pathways have been proposed for the biosynthesis of the phenolic unit, gallic acid (120). One is P-oxidation of the side chain of cinnamates to give gallic acid (120), while the other is direct conversion of 3-dehydroshikimate to gallic acid (120) [95]. Recently, it has been shown that formation of gallic acid (120) via cinnamic acids can be ruled out as a major pathway in the fungus Pllycornyce.7 blakesleeanus and in young leaves of Rhus typhirza, and gallic acid (120) is probably formed from an early intermediate of the shikimate pathway, most probably 3-dehydroshikimate [96]. In the pathways to hydrolyzable tannins, the first specific intermediate is P-glucogallin ( l-O-galloyl-P-D-glucose), which is formed from UDP-glucose and gallic acid (120). Then, P-glucogallin is

229

Chemistry of Extractives

Lanostane

Dammarane

Euphane

H3CO"

Abieslactone 100

Limonoid

Dammarenediol I 101

Quassinoid

Euphol 102

Lupane

Oleanane

$$

,8

H0

0

H

OH

Cedrelone 103

Quassin 104 -

P-Amyrin

Betulin ( R S E , O E ) 105

Betulinic

107

acid

(R=COOH) 106

FIGURE 13 Examples of triterpenes.

converted to pentagalloylglucose and gallotannins. Enzymes which catalyze these conversions from gallic acid (120) to gallotannins have been studied intensively by Gross and co-workers, summarized in the review by Gross [97]. Condensedtannins(proanthocyanidins)(Fig.17c)areoligomers and polymersof polyhydroxyflavan-3-01 units [35]. The repeating unit is connected through C4-C6 or C4-C8 bonds. 3-Hydroxyl groups of condensed tannins are often galloylated (e.g., tannins of Diospyros kaki) [98]. Condensation of flavan-3,4-diols and flavan-3-01s may give rise to proanthocyanidins (condensed tannins), but possible mechanisms for the process and their enzymology are still unknown. In spite of their difference in the basic structures, hydrolyzable and condensed tannins have a similarity in that they have many phenolic units and therefore are often called plant polyphenols [94]. Besides the phenolic nature, tannins have the following general characteristics: antioxidant and radical-scavenging activities and the ability to complex with

Umezawa

230

5P-Steroid 5a-Steroid

p-sitosterol

Ponasterone A

108

109

FIGURE 14 Examples of steroids. R: alkyl group.

a-Thujaplicin 110

Einokitiol (P-Thujaplicin)

y-Thu japlicin 112

111

Nootkatin 113

H3coq H&O

H3C0

CH3

.

0 OCH3

Colchicine 114 FIGURE 15 Examples of tropolones.

231

Chemistry of Extractives

OCH3

H3C0

0

0 118

0

J'uglone

2,C-Dimethoxyp-benzoquinone 117 16

TectoquinoneMansonone

A

1

115

FIGURE 16 Examples of quinones.

R=

yo

CO

/OR

OH

RO&&&%' OR Hydrolysable

tannin -

OH OH Hexahydroxydiphenoyl group

Depside galloyl group

OH

l

.n 119

(b)

Gallic acid

120

Haxahydroxydipbenic acid

OH Ellagic acid

122

121

FIGURE 17 (a): Hydrolyzableandcondensedtannins. of (c): hydrolyzable tannins.

(b): phenolic acids whicharecomponents

Umezawa

232

metal ions and with other molecules such as proteins, polysaccharides, and alkaloids [54]. These properties underlie their biological activities as well as their industrial applications. Tannins exhibit various biological and pharmacological activities, e.g., bacteriocidal action, inhibition of HIV replication, as well as astringency [37,54]. Inhibitory effects of condensed tannins on the activities of Streptococcus sobrins glucosyltransferases involved in dental caries formation were also reported [99]. Recently, the role of plant phenolics in diets is attracting considerable interest, since epidemiological investigations suggested that the consumption of beverages containing plant phenolics (such as tannins and flavonoids), in particular, green tea and red wines, reduced the risk of certain degenerative diseases [54,100]. Recent study demonstratedthat the iron-chelating properties of polyphenols contribute to limit the growth of microorganisms [ l o l l . The ability of tannins to complex proteins has long been utilized for tanning agents in leather manufacturing,while that to complexmetal ions hasbeenappliedtodyeing [37,54].Condensed tannins, especiallywattle (ormimosa) tannins, havebeenused to produceadhesives for wood-basedmaterialssuchas particle board and plywood.The tannin adhesives have been successfully commercialized especially in South Africa, Australia, and New Zealand [ 102- 1041. Many reviews of chemical properties, biological significance, and commercial significance of tannins are available in a recent book [105]. Mechanisms of tannin-protein complexation will be described briefly in Section X1.B.

X.

OTHER COMPOUNDS

Besides the compoundsmentionedabove,sugars, triglycerides andwaxes,monomeric aromatic compounds and phenols, alkaloids, etc., occur as extractives in woody plants.

A.

Sugars

D-Glucose and D-fructose are found, along with sucrose, in the sapwood of woody plants, while L-arabinose is found in heartwood [ 1061. Larches (Larix spp.) contain significant amounts (IO-25%) of a water-extractable arabinogalactan [ 1071.

B. Glyceridesand Waxes Glycerides are esters of glycerol and long-chain fatty acids. Among the glycerides, triesters (triglycerides) are dominant. Triglycerides of pine woods cause pitch trouble in ground wood pulping. Waxes are complex mixtures of aliphatic compounds, and a majority of these compounds are wax esters composed of fatty acids and fatty alcohols, hydrocarbons and derivatives, long-chain fatty acids, and long-chain fatty alcohols [ 1081. C.

MonomericAromaticCompounds

Phenylpropanoid monomers are distributed widely in plants, including woody plants. Coniferin (124) and syringin (125) (Fig. 18), aglycons of which are precursors of lignins and lignans, were isolated from many woody plants [109,110]. A soil bacterium, Agrohacteriumtumefaciens, can initiate the neoplastic disease called crown gall on dicotyledonous plants. Tumor-inducing plasmid of the bacterium is used to construct a useful vector to produce transgenic plants. Virulence gene expression

233

Chemistry of Extractives

OH R=E

OH

Coni feryl alcohol 14

R=H Coniferin 124

R=OCH3

Sinapyl alcohol 123

Anethole 126

R=OCE3

Syringin 125

Eugenol 127

Umbelliferone 130

Safrole 128

Cinnamaldehyde 129

Aesculetin 131

R Urushiol 132 R=ClSH25-31 FIGURE 18 Examples of monomeric aromatic compounds.

of the bacterium is activated by coniferyl alcohol (14) and sinapyl alcohol (123) [ l I l ] (Fig. 18) as well as acetosyringone (3’,5’-dimethoxy-4’-hydroxyacetophenone) [ 1 121. Various phenylpropanoid monomers are components of essential oils, for example (Fig. 18), anethole (126) (star aniseed oil), eugenol (127) (clove oil), safrole (128) (sassafras oil), and cinnamaldehyde (129) (cassia oil), which have been used for spice and perfume. Coumarins constitute another class of phenylpropanoids which have 2H- 1 -benzopyran-2-one structure. They are distributed widely in plants, particularly Umbelliferae and Rutaceae [ 113- 1 151. Two examples, umbelliferone (130) and aesculetin (131), are shown in Fig. 18. Biological activities of coumarins are reviewed in the literature [37]. Urushiol (132) (Fig. 18) is a major component of urushi (Japanese lacquer), milky exudate from urushi (Rhus vernicijlua), which has been used to produce japan (Japanese lacquer ware) [ 116,1171.

234

Urnezawa

0 "+,/

OCH3

N

H

/

Berberine 133

H% H3COys

Quinine 134

9 H '

HH0

0 OH

-tine

,I

OCH3 OCH3

135

H 3 c 0H 3 ~ C: O H" w

0 OCH3

Yohimbine Strychnine Reserpine

OCH3 OCH3

136

Camptothecin FIGURE 19 Examples of alkaloids.

D. Alkaloids Although the occurrence of alkaloids is much less than other wood extractives, most of them are of considerable interest due to their biological activity. Examples of alkaloids from woody plants are as follows [37,118- 1201 (Fig. 19): berberine (133) from Phellodendron arnurense, which is antibacterial; quinine (134) from Cinchona pubescens, which is employed for the treatment of malaria; emetine (135) from Cephaelisipecacuanha, which is anamebicide;yohimbine (136) from Pausinystalia yohimbe, which is a selective inhibitor of the presynaptic a-2-adrenergic receptors; strychnine (137) from S t r y c h n o s n u - v o m i c a , a very toxic alkaloid; reserpine (138) from Rauwolfia serpentina, which is antihypertensive; and camptothecin (139) from Camprotheca acurninata. which is antitumor.

XI.

CONTRIBUTION OF EXTRACTIVES TO THE PROPERTIES OF WOOD AND UTILIZATIONOF WOOD EXTRACTIVES

Extractives of wood influence various properties of wood, e.g., color, fragrance, and durability. Some extractives have injurious effects on human health[89]. Troubles in pulping processes and adhesion in production of wood-based materials are sometimes due to extractives. They are described in detail in a book [ 1211 and outlined byKaiin the first edition of this book [122]. Therefore, several topics in these fields are described briefly in this section.

Chemistry of Extractives

A.

235

AcousticProperties

Recently, acoustic properties and internal friction (loss tangent, tan S) of several woods and a cane have been observed to be strongly influenced by extractives. The term tan S is an indication of decrement of vibration of solid materials, and materials with lower tan S exhibit higher sound radiation. Methanol extraction of heartwood specimens of western red cedar (Thuja plicata), which is usedfor the top plate of the guitar, increased tan S values by 15.3-36.9% [ 123,1241. The same effect of methanol extraction was observed for rosewood (Dalbergia spp.), black cherry(Prunus serorina), and padauk (Pterocarpus indicus),whereas no effects of heartwood contents or methanol extractives were detected in bubinga (Guibourtia demeusei) [ 123,1241. Pernambuco (Guilandinaechinata syn. Caesalpiniaechinata) andcane (Arundo donax) have been used for violin bows and reeds of woodwind instruments, respectively. Water extracts from pernambuco reduced tan S value [ 1251, while in the case of air-dried cane, water extracts increased tan S value [126]. It was suggested that the water extracts of the cane consisted mainly of oligoglucans [126].

B. Tannin-ProteinComplexation The ability to complex with proteins is one of the most important features of tannins, and the mechanisms for tannin-protein complexation have long been studied. These studies clarified effects of environmental factors (such as pH, temperature, and ionic strength) on the process, and the following three principal features of tannin structure and properties which are important in the complexation with protein have been established: molecular size, conformational flexibility, and water solubility of the tannin. [54,94,127-1291. The precipitation ability increases with an increase of the degree of polymerization of condensed tannin [130], while in the case of hydrolyzable tannin this ability is enhanced with the addition of galloyl ester group [ 13 l]. Selective interaction of the condensed tannin from Sorghum bicolor with various proteins was also investigated, and it was reported that proline-rich and flexible proteins have high affinity for the polyphenol [ 1321. As for the mode of interaction between tannins and proteins, involvement of hydrogen bonding and hydrophobic interaction was suggested [ 1271. Although the relative importance of these two types of interactions remains uncertain, the initial association of protein-polyphenol by hydrophobic interaction followed by reinforcement of the association by hydrogen bonding was proposed [ 1271. Since tannins have complicated structures, synthetic tannin models which have definite structures are useful to elucidate mechanisms of tannin-protein complexation. Based onexperimentswithaseries of synthesizedcondensedtanninmodels, the distribution pattern of the phenolic hydroxyl group in the tannin molecule, but not the existence of odihydroxyphenyl groups, was found to be important for higher protein-precipitating capacity [ 133,1341. Two mechanisms were proposed for tannin-protein co-precipitation [94,127]. One is a “cross-linking mechanism,” in which one tannin molecule binds more than two protein molecules simultaneously to form an aggregate to be precipitated. The other is a “twostage precipitation mechanism,” which consistsof initial complexation of tannin molecules to a protein molecule to form a complex, followed by aggregation of the complexes to give precipitates. Recently, a series of hydrolyzable tannin models has been synthesized, and studies withthe models suggested that the two-stage mechanism is involved in tanninprotein co-precipitation [129,135,136].

Umezawa

236

C. Antitumor Taxol and Podophyllotoxin Recently, phytochemical studies of yew trees (Taxus spp.) have been developed exponentially [ 1371. This is due mainly to the plants’ producing taxane diterpene, taxol (99), which has strong antitumor activity. Taxol (99) is one of the most promising anticancer drugs and has been marketed under the name of TaxoP. Paclitaxel is the generic name for TaxoP. Because of great familiarity with the word taxol, however, it is used in this review in lieu of paclitaxel. Taxol (99) is present only in trace amounts in Taxus spp. (e.g., 0.0001-0.069% from 7: brevifolia bark) [ 1381. The sources of taxol (99) and related taxaneswerereviewed [ 1391. Interestingly, Taxomyces andreanae,an endophytic fungus of 7: brevifolia, produces taxol (99) andbaccatin 111 whenincubated in a semisynthetic liquid medium,although the quantities detected are very low (24-50 ng of taxolk) [140]. The limited production of taxol (99) in nature as well as the projected needs for the compound and its unique chemical structure have provided a very challenging target for syntheticorganicchemists.Recently,two total syntheses of this compoundhavebeen reported independently [ 141-1431. The production of taxol (99) through semisynthesis fromthemore readily available 10-deacetylbaccatin 111 (98) is nowwellestablished [ 144,1451. The biosynthesis of taxol (99) and its production by cultured cells have also received much interest. Recently,taxadienesynthasecatalyzing the cyclization of the universal diterpene precursor, geranylgeranylpyrophosphate (60), to the taxanesystem [taxa4(5), 1 l( 12)-diene] in a single enzymatic step was purified and characterized [ 1461. Very recently, Eisenreich et al. showed that the taxane ring system is not biosynthesized via mevalonate [73]. Production of taxol (99) by cultured cells is reviewed in the literature [120,145,147]. An aryltetralin lignan podophyllotoxin (7)has strong antitumor and antimitotic activity [4,5,20,148]. Since the lignan possesses severe gastrointestinal toxicity, semisynthetic derivatives of podophyllotoxin (7)have been developed to avoid the toxicity. Etoposide, which isone of the derivatives,hasbeenusedsuccessfully for cancerchemotherapy [5,148]. Podophyllotoxin (7)and its congeners occur in root and rhizome of Podophyllum spp. as well as other plants including woody plants, e.g., callus culture of Callitris drummondii (Cupressaceae) and leaves of Juniperus spp. (Cupressaceae) [20,148]. These lignans are elaborated from matairesinol [5,9,149,150].

D.

Pitch Trouble in Pulp and Paper Making

Wood extractiveshavecaused technical andeconomic pitch problems in the pulpand paper industry. Recently,twotypes of biochemicaltreatmentshavebeendeveloped to alleviate the problems. One employs enzymes, and the other uses a microorganism. Japanese red pine (Pinus densijora) isan important raw material for groundwood pulp. The wood contains significant amounts of resinousmaterialswhichcause pitch trouble in groundwood pulp process. The trouble can be partly avoided by seasoning the logs. Recently, lipase-catalyzed hydrolysis was foundto reduce triglycerides in the resinous materials to 70%, when the enzyme was added to groundwood pulp at 9000 U/kg. This treatment reduced the pitch deposits remarkably, and allowed the use of unseasoned logs up to 50% of total wood supply without pitch troubles [ 15 1,1521. Theother is a biological approach to decreasingextractivesfrom wood prior to pulping using the fungus Ophiostoma piliferum. A 2-week treatment of pine chips (50%

Chemistry of Extractives

237

Pinus taedcl and 50% Pinus virginiuna) with the fungus reduced ether-extractable pitch components by 22% compared with chips seasoned naturally for 2 weeks. GC-MS analysis of pitch components in fresh chips and in fungus-treated chips showed significant declines in the concentrations of triglycerides, fatty acids, and resin acids [ 1531.

REFERENCES 1. N. Migita, Y. Yonezawa, and T. Kondo (eds.), Wood Chemistry (in Japanese), Kyoritsu Shuppan, Tokyo ( 1 968). 2. T. Higuchi, Biochemistry and Molecular Biology of Wood, Springer-Verlag, Berlin (1997). 3. C. B. S. Rao (ed.), Chemistry of lignans, AndhraUniversityPress,AndhraPradesh,India ( 1978). 4. W. D. MacRaeandG.

5.

6. 7. 8. 9. IO. 1I.

H. N.Towers, Phytochemistry,23: 1207 (1984). D. C. Ayres and J. D. Loike. Lignans:Chemical,BiologicalandClinicalProperties, Cambridge University Press, Cambridge, U.K. (1990). R.D.Haworth, Ann. Rep. frog. Chem., 33: 266 (1936). 0. R.Gottlieb, Phytochernistry, 11, IS37 (1972). 0. R. Gottlieb, Fortscht:Chem. Org. Nuturst., 35, 1 (1978). T. Umezawa, in Biochetnistr?:andMolecularBiology of Wood, (T. Higuchi, ed.), SpringerVerlag, Berlin, pp. 181 - 194 ( 1997). 0. R. Gottlieb, in Chemistry of lignuns (C. B. S. Rao, ed.), Andhra University Press, Andhra Pradesh, India, pp. 277-305 ( I 978). T. Umezawa,L. B. Davin,and N. G.Lewis, Biochem.Biophys.Res.Cotnmun.. 171: 1008

(1990). 12. A. Chu, A. Dinkova, L. B. Davin, D. L. Bedgar, and N. G. Lewis, J . B i d . Chern., 268: 27026 ( 1993). 13. P. W. Park,H.-B.Wang,L. B. Davin, and N. G. Lewis, Tetrahedron Lett., 35: 4731 (1994). 14. A. T. Dinkova-Kostova, D. R. Gang, L.B. Davin, D. L.Bedgar, A. Chu, and N. G. Lewis, J. B i d . Chern., 271: 29473 (1996). IS.

16. 17. 18. 19. 20. 21. 22. 23. 24. 25.

26. 27. 28. 29. 30. 3I .

L. B. Davin, H.-B. Wang, A. L. Crowell, D. L.Bedgar, D. M. Martin, S. Sarkanen, and N. G. Lewis, Sciencc~,275: 362 (1997). T. Umezawa and M. Shimada, Biosci.Biotechnol.Biochem.. 60: 736 (1996). S. Suzuki, T. Umezawa,andM.Shimada, Biosci.Biofechnol.Biochem., 62: 1468(1998). J . D. Orr and D. G. Lynn, Plant Physiol., 98: 343 (1 992). T. Katayama and Y. Kado, J . Wood Sci., 44: 244 ( 1998). T. Umezawa, MokuzaiCakkaishi. 42: 91 1 (1996). L. U. Thompson, M. M. Seidl, S. E. Rickard, L. J. Orcheson, and H. H. S. Fong, Nutrition and Cancer, 26: 159 ( I 996). T. Y. Shen, S.-B. Hwang, M. N. Chang, T. W. Doebber, M.-H. T. Lam, M. S. Wu, X. Wang, G. Q. Han, and R. Z . Li, Proc. N r d . Acad. Sci. USA, 82: 672 (1985). A. M.Clark, F. S. El-Feraly, and W.-S. Li, J . Pharmaceut. Sci., 70: 951 (1981). K. Watanabe,H.Watanabe, Y. Goto, M. Yamaguchi, N. Yamamoto, and K. Hagino, Plantu Med., 49: 103 (1983). D. A. Whiting, Naturcrl Prod.Rep., 4 : 499(1987). D. A.Whiting, NaturalProd. Rep., 7 349(1990). R. S. Ward, Nuturd Prod. Rep., 12: 183 (1995). T. Nikaido, T. Ohmoto, H. Noguchi, T. Kinoshita, H. Saitoh, and U. Sankawa, Plantcl Med., 43: 18 (1981). J. Kinjo, J. Furusawa,and T. Nohara, Tetrahedron L e f t . , 26: 6101 (1985). A. J . Birch and A. J. Liepa, in Chemistr~y ofLignans (C. B. S. Rao, ed.), Andhra University Press, Andhra Pradesh, India, pp. 307-327 (1978). Y. Kai, H. Kuroda, and F. Teratani, Mokuzai Gakkaishi. 18: 3 1 S ( I 972).

238

Umezawa

32. K. Takahashi, Mokuzai Gakkaishi, 2 7 654 (198 l ) . 33. H. Ohashi, H. Hayashi, M. Yamada, and M. Yasue, Res. Bull. Fac. Agr. Gifu. Univ., 52: 131 (1987). 34. H. Funakoshi, K. Nobashi, and T. Yokota, Mokuzai Gakkaishi, 24: 141 (1978). 35. L. J. Porter, in Methods in Plant Biochemistry, Vol. l . Plant Phenolics (J. B. Harborne, ed.), Academic Press, London, pp. 389-419 (1989). 36. J. B. Harborne, in Natural Products of Woody Plants I, (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 533-570, (1989). 37. J. Bruneton, Pharmacognosy, Phytochemistry, Medicinal Plants (C. K. Hatton, transl.), Lavoisier, Paris (1995). 38. J. B. Harborne (ed.), The Flavonoids: Advances in Research Since 1986, Chapman & Hall, London (1994). 39. J. B. Harborne, T. J. Mabry, and H. Mabry (eds.), The Flavonoids, Chapman & Hall, London (1 975). 40. J. B. Harborne and T. J. Mabry (eds.), The Flavonoids: Advances in Research, Chapman & Hall, London (1982). 41. J. B. Harborne (ed.), The Flavonoids: Advances in Research Since 1980, Chapman & Hail, London (1988). 42. W. HellerandG.Forkmann,in The Flavonoids: Advances in Research Since 1986 (J. B. Harborne, ed.), Chapman & Hall, London, pp. 499-535 (1994). 43. K.HahlbrockandH.Grisebach,in The Flavonoids (J. B.Harborne, T. J. Mabry,andH. Mabry, eds.), Chapman & Hall, London, pp. 866-915 (1975). 44. R. K. Ibrahim and L. Varin, in Methods in Plant Biochemistry, Vol. 9, Enzymes of Secondary Metabolism (P. J. Lea, ed.), Academic Press, London, pp. 99- 131 ( 1 993). Plant Mol.Biol., 24: 743 45. F. Sparvoli,C.Martin,A.Scienza,G.Gavazzi,andC.Tonelli, (1994). 46. G. Forkmann, in Polyphenolic Phenomena (A. Scalbert, ed.), INRA, Paris, pp. 65-71 (1993). 47. G. Forkmann, in The Flavonoids: Advances in Research Since 1986 (J. B. Harborne, ed.), Chapman & Hall, London, pp. 537-564 (1994). 48. B.W. Shirley, W. L. Kubasek, G. Storz, E. Bruggemann, M. Koornneef, F. M. Ausubel, and H. M. Goodman, The Plant l.,8: 659 (1995). Recent Advances in 49. R. A. Dixon, C, A. Maxwell, W. Ni, A. Oommen, and N. L. Paiva, in Phytochemistry,Vol. 28, Genetic Engineering of PlantSecondary Metabolism (B. E. Ellis, G. W. Kuroki, and H. A. Stafford, eds.), Plenum Press, New York, pp. 153-178 (1994). 50. C. A. Williams and J. B. Harborne, in Methods in Plant Biochemistry, Vol. I , Plant Phenolics (J. B. Harborne, ed.), Academic Press, London, pp. 421-449 (1989). 51. P.M. Dewick, in The Flavonoids: Advances in Research (J. B. Harborne and T. J. Mabry, eds.), Chapman & Hall, London, pp. 535-640 (1982). 52. K. Hayashi, Plant Pigments, an Introduction to Research and Experiments (enlarged ed.) (in Japanese), Yokendo, Tokyo (1988). 53. T. Goto and T. Kondo, Angew. Chem. Int. Ed. Engl., 30: 17 (1991). 54. E. Haslam, J. Natural Prod., 59: 205 (1996). 55. E. Middleton, Jr., and C. Kandaswami, in The Flavonoids: Advances in Research Since 1986 (J. B. Harborne, ed.), Chapman & Hall, London, pp. 619-652 (1994). 56. T. Okuda, in Polyphenolic Phenomena (A. Scalbert, ed.), INRA, Paris, pp. 221-235 (1993). 57. T,P. Jungblut, J.-P. Schnitzler, W. Heller, N. Hertkorn, J. W. Metzger, W. Szymczak, and H. Sandermann, Jr., Angew. Chem. Int. Ed. Engl., 34: 312 (1995). H. Ziegler, Trees, 5 : 203 (1991). 58. E. A. Magel, A. Drouet, A. C. Claudot, and 59. E. Magel and B. Hubner, Bot. Acta, 110: 314 (1997). 60. T. Norin, in Natural Products of Woody Plants I (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 512-533 (1989). 61. H. Kindl, in Biosynthesis and Biodegradation of Wood Components (T. Higuchi, ed.), Academic Press, Orlando, FL, pp. 349-377 (1985).

Chemistry of Extractives

239

62. H. Kindl, Hoppe-Seylerh Z. Physiol. Chem., 352: 767 (1971). 63. R. Gehlert, A. Schoppner, and H. Kindl, Mol. Plant-Microbe Interact., 3: 444 (1990). 64. R. Hain, H.-J. Reif, E. Krause, R. Langebartels, H. Kindl, B. Vornam, W. Wiese, E. Schmelzer, P. H. Schreier, R. H. Stocker, and K. Stenzel, Nature, 361: 153 (1993). 65. H. Erdtman, Tappi, 32: 346 (1949). 66. P. Cleason, P. Tuchinda, and V. Reutrakul, J . Indian Chem. Soc., 71: 509 (1994). 67. G . M. Keserii and M. N6grBdi, in Studies in Natural Products Chemistry. Vol. 17 (Atta-urRahman, ed.), Elsevier, Amsterdam, pp. 357-394 (1995). 68. P. J. Roughley and D. A. Whiting, Tetrahedron Lett., 1971: 3741. 69. T. Inoue, N. Kenmochi, N. Furukawa, and M. Fujita, Phytochemistry, 26: 1409 (1987). 70. N. Watanabe, Y. Uraki, and Y. Sano, “Biosynthesis of Diarylheptanoids in Betulaceae”, in 71. 72. 73. 74. 75. 76.

Abstracts of the 40th Anniversary Conference of the Japan Wood Research Society, Tokyo, Japan, p. 420 ( 1 995). M. Rohrner, M. Knani, P. Simonin, B. Sutter, and H. Sahm, Biochem. J., 295: 517 (1993). M. Rohmer, M. Seemann, S. Horbach, S. Bringer-Meyer, and H. Sahm, J. Am. Chem. Soc., 118: 2564 (1996). W. Eisenreich, B. Menhard, P. J. Hylands, M. H. Zenk, and A. Bacher, Proc. Natl. Acad. Sci. USA, 93: 643 1 ( 1996). S. Dev, in Natural Products of Woody Plants I1 (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 691-807 (1989). W. R. Nes, in Natural Products of Woody Plants I1 (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 808-842 ( 1 989). J. D. Connolly and R. A. Hill, Dictionary of Terpenoids, Vols. 1-3, Chapman & Hall, London

(1991). 77. R. Croteau and M. A. Johnson, in Biosynthesis and Biodegradation of Wood Components (T. Higuchi, ed.), Academic Press, Orlando, FL, pp. 379-439 (1985). 78. S. Kobayashi, C . Ishibashi,A.Morita, H. Masuda,andA.Ogiso, Pharmacometrics (Oyo kkuri), 24: 599 (1982). 79. K. Nakanishi, M. Koreeda, S. Sakai, M. L. Chang, and H. Y. Hsu, Chem. Commun., 1966: 915. 80. T. Nozoe, Bull. Chem. Soc. Jap., 11: 295 ( 1936). 81. T. Nozoe, Fortschr Chem. Org. Nuturst., 13: 232 (1956). 82. H. Erdtman, Progr. Org. Chem.. 1: 22 (1952). 83. K. Sakai, T. Yamaguchi, and R. Itose. Mokuzai Gukkaishi, 43: 696 (1997). 84. A. R.Battersby, T. A. Dobson, D. M. Foulkes, and R. B. Herbert, J. Chem. Soc. Perkin 1, 1730 (1972). 85. M. C. O’Sullivan and J. M. Schwab, Bioorg. Chem., 23: 131 (1995). 86. A. J.J. van den Berg andR. P. Labadie, in Methods in P h l t Biocherni,ytl;y, Vol. I (J. B. Harborne. ed.), Academic Press, London, pp. 451-491 (1989). 87. E. P. Swan, in Nrrtural Products sf Woody P1unt.s I1 (J. W. Rowe, ed.). Springer-Verlag, Berlin, pp. 931-952 ( 1989). 88. J. W. Rowe and A. H. Conner. Generul Tech. Rep. Forest Pmd. h b . Mudison, 18: 22 ( 1 979). 89. B. Hausen. Woods Injurious to Humun Health: A M a n ~ ~ ndel , Gruyter, Berlin ( 1981). 90. E. L. Rice, Allelopcrthy. 2nd ed.. Academic Press. Orlando, FL (1984). 91. W. Sandermann and M. H. Simatupang, Hol: u l s Roh- und Werkstoflff; 24: 190 (1966). 92. N. Tanaka. M. Yasue, and H . Imamura. Tetrahedron Lett.. 1966: 2767. 93. T. Swain and E. C. Bate-Smith, in Cortlparative Biochemistry, Vol. 3 (M. Florkin and H. S. Mason, eds.). Academic Press, New York, pp. 755-809 (1962). 94 E. Haslam, Plnnt Po1yphenol.s. Cambridge University Press, Cambridge, U.K. (1989). 95. E.Haslamand Y. Cai. Nuturd Prod. Rep., 11: 41 ( 1994). 96. I. Werner, A. Bacher, and W. Eisenreich. J. Riol. Chem., 272: 25474 (1997). 97. G . G. Gross. in Methods i n Plunt Bioc.k~.rlri.str?: W . 9, Enzymes of Seconrltrt:\. Metcrl~olisrn (P. J. Lea, ed.). Acadcmic Press, London. pp. 25-43 (1993).

240

Umezawa

98. T. Matsuo and S. Ito, Agric. Biol. Chem., 42: 1637 (1978). 99. T. Mitsunaga, I. Abe, M. Kontani, H. Ono, and T. Tanaka, J. Wood Chem. Technol., 1 7 327 ( 1997). 100. A. L. Waterhouse, Chem. Ind.. 1995: 338. 101. I. Mila, A. Scalbert, and D. Expert, Phytochemistry, 42: 155 1 (1996). 102. A. Pizzi, in Polyphenolic Phenomena (A. Scarbert, ed.), INRA, Paris, pp. 267-274 (1993). 103. Y. Yazaki, Proc. 18th Symp. on WoodAdhesion, Tokyo,pp. 36-48 (1997). 104. H. L. Hergert, in AdhesivesfromRenewableResources,ACS Synzp. Ser: 385 (R. W. Hemingway and A. H. Conner, eds.), American Chemical Society, Washington, DC, pp. 155-171 (1989). 105. R. W. Hemingway and P. E. Laks (eds.), Plant Po/yphenols, Plenum Press, New York (1992). 106. J. N.BeMiller,in NaturalProducts of WoodyPlants I , (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 155-1 78 (1989). 107. T. E. Timell, Wood Sci. Technol., I : 45(1967). 108. P. E. Kolattukudy and K. E. Espelie, in NaturalProducts of WoodyPlants I (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 304-367 (1989). 109. M. Terazawa, H. Okuyama, and M. Miyake, Mokuzui Gakkuishi, 30: 322 (1984). 110. M. Terazawa and M, Miyake, Mokuzai Gakkaishi, 30: 329 (1984). 1 1 1. P. A. Spencer and G. H. N. Towers, Phytochemistry, 2 7 2781 (1988). 112. S. E. Stachel, E. Messens, M. van Montagu, and P. Zambryski, Nature, 318: 624 (1985). 113. R. D. H.Murray, Forrschr: Chem.Org.Nuturst., 35: 199 (1978). 114. R. D. H. Murray, Nuturul Prod.Rep., 12: 477, (1995). 115.

116. 117. 1 1 8.

119. 120. 12 I . 122. 123. 124. 125. 126. 127. 128. 129. 130. 13 I .

132. 133.

134. 135. 136. 137. 138.

0. Theander and L. N. Lundgren, in Natural Products of Woody Plunts I (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 369-399 (1989). W. E. Hillis, in NaturalProducts o f WoodyPlants I (J. W. Rowe,ed.), Springer-Verlag, Berlin, pp. I - 13 ( 1989). W. E. Hillis, Heartrt'ood and Tree Exudates, Springer-Verlag.Berlin, pp. 59-60 (1987). C. W. W. Beecher, N. R. Farnsworth, and C. Gyllenhaal, in Nuturul Products of Woody Plants II (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 1059-1 164 (1989). S. I. Sakai, N. Aimi, E. Yamanaka, and K. Yamaguchi, in Natural Products of Woody Plunts I (J. W. Rowe, ed.), Springer-Verlag, Berlin, pp. 200-257 (1989). S. Tachibana, Mokuzai Gakkaishi, 41: 967(1995). J. W. Rowe (ed.), Natrtrd Products of Woody Plants. I and 11, Springer-Verlag, Berlin (1989). Y. Kai, in Wood andCellulosicChemistry (D.N.-S. Hon and N. Shiraishi,eds.),Marcel Dekker, New York, pp. 2 15-255 ( 199I ). H. Yano, Holiforsch.,48: 491 (1994). H. Yano, H. Kajita and K. Minato, J . Acous?. Soc. Am., Y6: 3380 (1994). M. Matsunaga, M. Sugiyama, K. Minato, and M. Norimoto, Holiforsch., "TO: 51 1 (1996). E. Obataya, T. Umezawa, F. Nakatsubo, and M. Norimoto. Holz,forsch., 53: 63 (1990). C. M. Spencer, Y. Cai, R. Martin, S. H. Gaffney, P. N. Goulding, D. Magnolato, T. H. Lilley, and E. Haslam, Phytochemistry, 27: 2397 (1988). G. Luck. H. Liao, N. J . Murray, H. R. Grimmer, E. E. Warminski, M. P. Williamson, T. H. Lilley, and E. Haslam, Phytochemistry, 37: 357 ( 1994). H. Kawamotoand F. Nakatsubo, Recent Res. Devel. Phytochem., I : 515(1997). E. C.Bate-Smith, Phytochemistry, 12: 907(1973). E. Haslam, Biochem. J.. 139: 285 (1974). A. E. Hagerman and L. G. Butler, J. B i d . Chenl., 256: 4494(1981). H. Kawamoto, F. Nakatsubo,and K. Murakami, J. Wood Chenz. Technol.. 1 0 59 (1990). H. Kawamoto, F. Nakatsubo,andK.Murakami, J. Wood Chem. Technol.. l o : 401 (1990). H. Kawamoto, F. Nakatsubo, and K. Murakami. Ph?,toc/lerni.stt:v,40: 1503 ( 1995). H. Kawamoto, F. Nakatsubo,and K. Murakami, Phytochemistry, 41: 1427 (1996). G.Appendino, Nrtturul Prod.Rep.,12: 349 (1995). N. Vidensek, P. Lim, A. Campbell, and c. Carlson, J . Nuturul Prod., 53: 1609 (1990).

Chemistry of Extractives

241

139. D. G. I. Kingston, Phcrrnlctcol. Tl~er:.52: 1 (1991). 140. A. Stierle, G. Strobel, and D. Stierle, Science. 260: 214 (1993). 141. R. A. Holton, C. Somoza. H,-B. Kim, F. Liang, R. J. Biediger, P. D. Boatman, M. Shindo, 142. 143. 144. 145. 146. 147. 148. 149. 150.

151.

152. 1 53.

C.C.Smith, S. Kim. H. Nadizadeh, Y. Suzuki,C. Tao, P. Vu, S. Tang, P. Zhang, K. K. Murthi, L. N. Gentile, and J. H. Liu, J. Am. Chem. Soc., 116: 1597 (1994). R. A. Holton, H.-B. Kim, C. Somoza, F. Liang, R. J. Biediger, P. D. Boatman, M. Shindo, C.C.Smith, S. Kim. H. Nadizadeh, Y. Suzuki,C. Tao, P. Vu, S. Tang, P. Zhang, K. K. Murthi, L. N. Gentile, and J. H. Liu. J. Am. Chern. Soc., 116: 1599 (1994). K. C. Nicolaou, Z. Yang, J . J. Liu, H. Ueno, P. G. Nantermet, R. K. Guy, C. F. Claiborne, J. Renaud, E. A. Couladouros, K. Paulvannan, and E. J. Sorensen, Nrrture, 367: 630 (1994). D. G . I . Kingston, in filxurw AnticwlcerAgents:Basic Science crnd CurrentStatus,ACS Syrnp. Ser: 583 (G. I. Georg. T. T. Chen, I. Ojima, and D. M. Vyas, eds.), American Chemical Society, Washington, DC, pp. 203-216 ( 1995). M. Suffness, in fir\-nr~eArzticcrrlcer Agents: Basic Science crud CurrentStatus,ACS Syrnp. Ser: S83 (G. I. Gcorg, T. T. Chen, 1. Ojima,and D.M. Vyas, eds.),AmericanChemical Society, Washington, DC, pp. 1-17 (1995). M. Hezari, N. G. Lewis, and R. Croteau, Arch. Biochem. Biophys., 322: 437 (1995). S. Furusaki, M. Seki, and M. Nakajima, Biosci. Ir~d.,52: 896 (1994). D. L. Sackett, Phtrrmrcol. Tlzer., 59: 163 (1993). P. M. Dewick, in Studies i n Naturol Products Chenlistry, Vol. S. Structure Elucidutiorl (Pcrrt B) (Atta-ur-Rahman, ed.), Elsevier, Amsterdam, pp. 459-503 ( 1 989). A. J. Broomhead. M. M. A. Rahman, P. M. Dewick. D. E. Jackson. and J. A. Lucas. P h y tocherrlistry, 30: I489 ( I99 I ). Y. Fujita, H. Awaji, M. Matsukura, and K. Hata, J q m z Tappi J . , 4.5: 905 (1991). Y.Fujita, H. Awaji, H. Taneda, M . Matsukura, K. Hata, H. Shimoto, M. Sharp, H. Sakaguchi, and K. Gibson, firppi J.. 74: 117 (1992). T. S. Brush, R. L. Farrell, and C. Ho. R ~ p p J., i 77: 155 (1994).

This Page Intentionally Left Blank

7

Chemistry of Bark Kokki Sakai Kyushu University, Fukuoka, Japan

1.

INTRODUCTION

Tree bark usually refers to all tissues external to and surrounding the vascular cambium. It occupies a much smaller volume than the wood of a mature tree stem because fewer bark cellsare produced than wood cells and also because the outermost bark cellsare continuously discarded in most tree species, while wood cells are retained and thus accumulate as the tree grows. In spite of its small volume, the bark plays important roles in a living tree. Tree barks often have developed complex anatomy and/or chemical compositions in order to manifest or maintain three main functions: (1) nutrient transport from the leaves to the rest of the tree, (2) protecting the sensitiveinnercambiumfromdesiccation,and (3) shielding from the environment as the primary defense of the tree against wildfire, mechanicalinjuriescaused by heavy wind, and attacks by phytopathogens,phytophagous insects, larger animals, and so on. The objective of this chapter is to provide the new information about the chemical composition and utilization of tree barks, mainly since 1986. The basic knowledge of bark chemistry and important findings in this field up until 1985 were well stated by Laks in the first edition of this book [ 11.

II. THE FORMATION AND ANATOMY OF TREE BARKS The formation and anatomy of bark are described very briefly, as quite detailed discussions of them were presented by Laks [ l ] . The epidermis. the cortex, and the primary phloem are produced during longitudinal growth from the apical meristem located at the apices of growing roots and branches. The secondary phloem and periderm are formed during radial growth of the tree. Accordingly, the primary phloem remains in only the outermost region of the bark of young trees. In most tree species the outer bark cracks and peels off as the tree grows, due to the successive formation of periderms within the bark and growth of the xylem. Thus the mature bark consists mainly of secondary phloem and periderm, a group of tissues including the phellogen or cork cambium, the phelloderm, and the phellem or cork tissue. As the sec243

Sakai

244

ondary phloem thickness, new periderm is formed within the phloem and any cells to the exterior of the periderm soon die. Most tree barks, therefore, have two zones, the inner bark that contains some living cells and the outer bark or rhytidome that does not contain any living cells. These regions are sometimes compared to the sapwood and heartwood in the xylem, respectively. Recently, Trockenbrodt [2] surveyed and discussed the terminology used in bark anatomy, andsuggestedterms for the tissue zones as illustrated in Fig. 1 . In the barkchemistry field, however, the terms “outer bark” and “inner bark” have been used for “rhytidome” and “secondary phloem up to the last-formed living secondary phloem,” respectively.

111.

CHEMICAL COMPOSITIONAND NONEXTRACTABLE COMPONENTS OF TREE BARKS

As described above, bark consists of the inner bark and outer bark zones. It is therefore preferable to determine the chemical composition accuratelyfor these two zones separately. However, chemical analyses have often been made with the whole bark because separation between the zones is sometimes cumbersome and time consuming, and because for most applications the whole bark is to be utilized. It has been known that chemical analyses by standard methods for wood often give erroneous results due to the presence and variability of suberin and high-molecular-weight tannins which are rarely or never found in wood. Surprisingly, some erroneous data were published evenin 1991 [ 3 ] .Preliminary extractions or corrections for these substances are very important for the accurate analysis of the bark of trees. Some of the general and comprehensive reviews onthe chemical composition of tree barks were referred to by Laks [ l ] . More recent examples of bark analysis are listed in Table 1 to show how different the chemical compositions of the tree bark are from those of wood. In general, bark contains much more extractives, slightly less lignin, and smaller amounts of holocellulose as compared to wood of the same tree. Nonextractable components in the bark consist of polysaccharides (cellulose, hemicellulose,and pectic substances),phenolicpolymers (lignin andhigh-molecular-weight tannins), and cross-linked polyesters (suberin and cutin). Significant parts of hemicellulose

/-

rhytidomc

secondary phlocm u p to thc last fonncd pc’idclnd living sccondaryphlocm calnbiunl xylcm FIGURE 1 Suggested tcrms for the tissue zones resulting from rhytidomc formation. Rcproduccd with permission of IAWA.

Chemistry of Bark

245

TABLE 1 Some BarkAnalyticalResults

(9%of DryMaterials)

Components Species

Extract. Lignin Ash

Holocellul. Cellulose Pentosan

1 % NaOH soh. Ref.

Eucalyptus glohulrts" 0.3

B kh

5 18.6 7.9 2.0

Wd' Populus Hybrid

1

NE388d B kh Wd' Pinus pirznster

28.7

B kh19.1 I? radiata Bkh Wd' I? sylvestris

Bkh Wd' Snlix roridrr

Bkh S. rkinuycrnngi Bkh

8.0

22.3

41.8 5.6

14.1 16.7

43.2 50. I 42.9 84.5

19.6 20.5

30.6 22.5

23.2 44.0 22.7

0.5 "2q.f -26 27.2 1.8

31.4


12.6 8.2

25.7 14.8

59.3 12.5

I1.1F

7.4

54.2"

10.4F

7.9

53.5"

"Twelve-year-old trees. hBark. 'Wood. "Four-year-old trees. 'Estimated by the solid-state "C-NMR method. 'Tannin. 'Alcohol-benzene extract. hAfter extraction with alcohol-benzene

and pectins can be removed during the extraction of tannins with alkaline solutions. This solubility also causes the difficulty in accurate analyses of the bark components. Recent studies on nonextractables in the bark are very limited. A lipidic biopolymer, suberin, is the most important structural component of the cork cell walls, which play the physiologically important role of thermal and hydric insulation in bark tissues. Suberin composition of Cedr-us libani bark was studied, and it was noted that a,w-dibasic acids and w-hydroxy-monobasic acids were the two dominated by hexadecane- I , 16-dioic acid and 16-hydroxyhexadecanoic acid as the major constituents in the suberin monomers of the bark of this species [ 1 l ] . Besides these acids, smaller amounts of long-chain primary alcohols, alkanoic acids, and ferulic acid were observed in C. libani bark, as are often observed with other species. Thus, the composition of the suberin monomer of the C. libnni bark was quite different from that of spruce and pine, where unsaturated C,,-acids (hydroxyoctadecenoic acid and octadecenedioic acid) were major monomers [ l ] . There were trials to isolate suberin from the bark without chemical action. Perra et al. extracted lignin and suberin from bark of Fngus sylvaticcr milled in a vibratory ball

246

Sakai

mill by extraction withaqueousdioxaneand purified each fraction. Treatment of the suberin fraction with BF,-methanol provided methyl esters of ferulic acid in addition to ordinary suberin monomers [12]. Preextracted corks from Quercus suber were thoroughly depolymerized by using a sodium methoxide-catalyzed methanolysis to detect 5.2% glycerol, 48.0% suberinic fatty acids and minor amounts of ferulic acid. A very mild depolymerization using calcium oxide-treated methanol solubilized only 2.0% of the cork material, of which43.8%were aliphatic acids, 2.1% werel-alkanols,and32.1%were monoacylglycerols. Based on GC-MS analysis of these acids and acylglycerols, Garqa and Pereira proposed that suberin is a glyceryl-based polymer and that its insoluble character is given, at least in part, by the cross-linking of dicarboxylic fatty acids with glycerol [ 131. In an alkaline extract of Pinus radiata bark, a,w-dibasic acids were found, suggesting that partial saponification of cutin or suberin occurred in the bark during the extraction [ 141. Dioxane-HC1 lignins of woods and barks of P. sylvestris and P. caribaea were studied comparatively. In both species the bark lignin showed smaller methoxyl contents and lower yields of vanillin upon nitrobenzene oxidation than wood lignin of the same species [9,15]. Thesebark lignin preparationsmusthavebeencontaminatedwith parts of condensed tannins which were resistant to extractions with neutral solvents and aqueous alkali.

IV.BARKEXTRACTIVES

Generally, bark contains much larger quantities of extractives than wood of the same tree, as seen from Table 1. Bark chemical components can be roughly fractionated by a sequence of extractions starting with a nonpolar solvent, such as petroleum ether, in many instancesfollowed by diethyl ether, ethanol, and water. Lesspolarcompoundssuchas waxes, resin, lipids, higher fatty acids, phytosterols, and terpenes are extracted mainly with non- or less polar solvents. On the other hand, ethanol and water dissolve relatively polar substances, such as flavonoids, phenolics, their glycosides, condensed tannins, sugars, etc. Acetone-water mixture is the preferred solvent for extraction of condensed tannins [ 15b]. Ethyl acetate is sometimes used for extraction of low-molecular-weight tannins or oligomeric proanthocyanidins. Aqueous NaOH solutions (e.g., 1%) have been often used for extraction of polymericcondensed tannins. However, attention should be given to the alkaline rearrangement resulting in partial or total rearrangement of tannins of the procyanidin class to “phenolic acid,” which have less phloroglucinol functionality [ 161. Saponification of ester bonds in cutin and suberin can also occur in aqueous alkali [ 141. The yields of sequential extracts from barks were compiled as tables by Laks [ I ] for various softwood tree species: Pseudotsuga menziesii, Pinus tueda, P. echinattr, l? virginiana, P. elliottii, P. sylvestris, P. brutia, l? btuksiana, P. corltortu, P. nigru, P. hulepensis, l? radiata,Piceueizgelnmnni, P. abies,Larixkaempferi, L. eurolepsis, L. deciducr,Abies balsurrtea, and Tsuga canadensis. Also compiled are the yields from barks of some hardwood species: Bruguierugymrtwrrhi:a.Rhizoplzera stylosa, Fugus sylvutica, Betula verI ’ U C O S ~ I ,B. allegl~uniensis,B. papyrifera, A1rur.s rubra, Populus tremuloides, Quercus d b a , and Q. rubra. Extractives yields recently published for barks of some species are listed in Table 2. The extractives content can vary according to a number of factors, including environmental, genetic, and seasonal variations. Seasonal fluctuation of extractive contents was investigated with barksof Quercus rohur; Piceacrbies, and Pirius.sdvestris grown in Germany [ 171. No season-related differences were observed in the amount of cyclohexane extract. Amounts of ethanol extract, NaOH cxtract,and water extract increased to different

rk

Chemistry of Bark TABLE 2

247

Yieldsof Extractives from Tree Barks Solvent Total

Hot

Species Hexane Benzene Ethanol water Ether 2.2-3.4 10.5-12.6 20.0-25.6 0.8-1.0 2.0-6.6 1.8-2.9h 3.2-5.1 0.5-0.8

Pinus pinastrr glohulus Eucalyptus Qurrcus rohrrr

Inner Outer bark radiata Pinus Cedrus

13.6

1 .9’

2.7

1.Od23.8‘ 2.9’

3 3

10.5

6.3 1 2.9h

1% NaOH

extractives Ref. 41.8

I61

-

6.3-8.5

[4]

14.0

40.0

[l71 1171

-

36.6

1181

[l91

“Hexane extract + benzene extract. hBenzene-ether. ‘Cycrohexane. “Toluene. ‘Ethyl acetate. ‘Acetone extract + MeOH extract. ’Petroleum ether. “Acetone.

extents from spring to winter. Of the three extracts, an increase in ethanol extract content was the most extensive. The extractives yield also depends on the aging between felling the tree and carrying out the analysis. Tree age influences the quantity and quality of bark extractives. In freshly cut P o ~ ~ u l u s t r m u r l o i d e sthe , higher up the tree,thegreater the amounts of bark extractives, probably due to the fact that the more inner bark regions are towards the top of the tree [ 201. Young Pinus rtrdiata bark (12 years old) provided extractives in a significantly higher yield (32.9%) than that (29.4%) from 30-year-old bark of the same species [2l]. In the case of Rhizoplzoru mucmmta, however, barks of older trees gave higher extractives and tannin contents [22].

A.

Lipophilic and Terpenoid Extractives

Waxes from Douglas fir bark were used for many purposes in the United States for several years from the late 1970s to the early 1980s. Thus,considerableknowledge has been accumulated on conifer bark waxes. Laks [ l ] cited such studies and listed major components of hydrolyzed and nonhydrolyzed bark waxes appearing in the literature until 1985. Free fatty acids, resin acids, and alcohols found in tree barks were also shown in a table. Little work has been performed on the lipophilic constituents of bark since1986. Analyses of sterol and wax esters in the “hexane wax’’ from Douglas fir (Psrudorsugcr ttzm:ie.sii) bark has revealed that the sterol ester was composed of sitosterol and campesterol esterified t o any or a l l of the fatty acids, C,,-, C,,-, C,,-, C,<,-.C?,!-, C??-, and C?,saturated acids. and oleic acid. The wax ester afforded, upon saponification, I-docosanol and I-tetrncosanol in addition to the above-mentioned fatty acids [23]. An ester mixture composed of such alcohols esterified to ferulic acid was isolated from bark hexane wax of the same species 1241. This mixture of a great number of individual esters is likely to account in part for the softness and low melting point of Douglas fir wax, ;IS Lavnr ct a l . I23 1 suggested.

Sakai

248

In petroleumetherextractivesof Cedrus libani bark, fatty acids, resin acids, and unsaponifiables occurred in similar amounts (35.4%, 34.3%, and 30.3%, respectively) [ 191. Compositionof the extractives after saponification is shown in Table 3. The saturated acidswerepredominantly fatty acids,similartobarkwaxes of manyotherconiferous species. Isopimaric acid was the principal constituent, while pimaric acid was not detected in the resin acids of the C. libani bark extractives. Fatty alcohols formed a dominating group in the saponifiable fraction, and tetracosanol was the main constituent of unsaponifiables. Normal paraffins from C,, to C,,) occurred in small amounts. Undec- IO-en-2-one and tridec- 12-en-2-one were identified as the major and minor components of the essential oil of the fresh bark of the Litseaelliptica tree, which is known for its termite resistance and repellent properties [3]. Examples of some mono-, di-, sesqui-, and triterpenes, and the species in which they occur, were shown by Laks [l]. Mono- and sesquiterpenes were recently identified in the essential oil of Magnolia oflcinalis bark, which has been used as a remedy for flatulent dyspepsia, cough, and asthma in China. By analysis with gas chromatography/mass spectrometry fitted with accurate mass analysis, 93.8% of the compounds detected in the oil were identified [25]. Thepredominantconstituentsweresesquiterpenes,P-eudesmol 17.4%, cadinol 14.670, and guaiol 8.7%. Their structures are shown in Fig. 2 . Glucosylated monoterpenic and sesquiterpenic tropolones wereisolated from the bark of Italian cypress, Cupressus senzpervirens, in response to infection by the fungus, Diplodia p i m a f. sp. cupressi. These tropolone glucosides inhibited in-vitro germination of spores

TABLE 3 Composition (%) ofPetroleumEtherExtractives from the Bark of Cedrus Iihani After Saponification [ 191

35.4 Resin acids Fatty acids Sandaracopimaric 0.5 Myristic Levopimaric 2.4 Palmitic Plustric 14-Methylhexadecanoic 0.9 0.4 Isopimaric Stearic Abietic 7.6 Oleic Dehydroabietic 1.1 5,9-Octadecadienoic Neoabietic 3.4 Linoleic Abietatrienic 0.9 Pinolenic Linolenic 0.5 Unsaponifiables Alkanes (C,?-C,,,) Arachidic I .3 Fatty alcohols 1 I , 14-Eicosadienoic I .2 Stearyl Behenic 3.6 Nonadesyl 7.1 Lignoceric Arachidyl Hexacosanoic 2.0 Behenyl Others 2.5 Lignoceryl Ceryl Phytosterols Campesterol Campestanol p-Sitosterol p-Sitostanol Others

34.3 1.0 0.9 6.3 9.8 4.9 5.7 5.3 0.4 30.3 1.S 18.3 0.4 0.9 I .3

2.1 2.2 11.4 8.4 2.6 0.2 5.1 0.5 2.1

249

Chemistry of Bark

a-Cadinol FIGURE 2

I)-Eudcsmol

Guaiol

Structures of cadinol, eudesmol, andguaiol

of various phytopathogenic fungi such asD.Pinecr, Seireciium cardinale [26]. A compound, (5aRY,6R*,9R*,9aS)-4-cinnarnoyl-3,6-dihydroxy1 -methoxy-6-methyl-9-( 1-methylethyl)5a,6,7,8,9,9a-hexahydrodibenzofuran(Fig.3),which is biosynthesizedprobablyfroma menthane type monoterpene and a chalcone, was isolated from the bark of Lindercr umhellatcr. This compound exhibited potent inhibitory activity on melanin biosynthesis [27]. Taxol, a unique and complex diterpenoid containing the taxane ring system, a rare four-memberedoxetane ring andestersidechains(Fig. 4), hasbeen one of the most extensivelystudiedditerpenoids in this decade. It was isolated as the majorcytotoxic principle in extracts from the bark of Pacific yew, R x u s brev$olin, by Wani et al. [28], but clinical trials were delayed. Phase I trials in human cancer commenced in 1983 and the great promise of taxol was confirmed. It has become an important new cancer chemotherapeutic agent, having significant activity in drug-refractory ovarian cancer. Taxol was approved for treatment of this disease by the U.S. Food and Drug Administration in 1992 [29]. Chemicalsynthesis[30],semisynthesisfromprecursors[31],production by endophytic fungi [32], and productionby cell cultures of Taxus spp. [33] have been studied very actively, since taxol is contained in different parts of Tuxus trees only in small amounts(Table 4) [34] and the trees grow quite slowly in general.Two new taxane diterpenoids (Fig. 4) were isolated from the barks of Taxus yunnanensis recently [35]. Diterpenoids with a daphnane skeleton (Fig. 5 ) were isolated from the bark of Wikstroemia retusa, which is used locally in the Ryukyu Islands in Japan for the preparation of traditional paper that can be stored for long periods of time, probably due to constituents against insect damage [36]. Daphnane diterpenoids have biological activities, e.g., antileukaemic, piscicidal, and antifertility. Three new pterocarpane-type diterpenes, margolone, margolonone, and isomargolonone, shown i n Fig. 6, were isolated from the stem bark of Azadirachta inclica (Meliaceae),

250

Sakai AcO

FIGURE 4

RA

Structures of taxol and some taxane diterpenoids obtained from bark of Erxus spp.

TABLE 4 Contents of Taxol in Different Parts of Trrrus Trees 1341 Plant Species 7irxus hrevifolicr

7: bncc.nrtr

7: rnrdin

7: ctrspicltrtcr

Taxol material Bark Roots Wood Wood with bark Branches Leaveslneedles Twigs Seedlings Stem Twigs Leaf Stem Twigs Leaf Twigs

(%)

0.015

0.004 0.0006 0.0003 0.00 17 0.00 1S 0.0012 0.0058 0.00 1 0.0006 0.003 0.002 0.009 1.002 0.0006

251

Chemistry of Bark R‘

R

R=HZ,R’=Mc,R”=COOH Margolonc R=O, R=Mc, R”=COOH Margolonone lsomargolononc R=O, R’=COOH, R”=Me FIGURE 6 Diterpenoids from the stem bark of Azcldirachfn indica.

which is regarded as a bitter tonic, astringent, and as being useful in fever, thirst, nausea, vomiting, and skin diseases [37]. These diterpenoids possessed different antibacterial activities against Klebsiella oxytoca, Staphylococcus epidermidis, and Serratia lutea. Supercritical carbon dioxide was used to extract cis-abienol efficiently from the bark of Abies saccharinensis [38]. This extract has a tobacco odor and oviposition-deterrent activity to some aphids and worms. Betulinand related triterpenes with the lupaneskeletonarecommon in barks of Betula species. Theirstructures are shown in Fig.7. However, their contents are very different from species to species, as shown in Table 5 [ 1,39-411. White-barked birches, B. verrucose in Europeand B.plutiphyllcr var. japonica in Japan,contained betulin to about 25% of their outer bark weight. On the other hand, much less betulin was observed in the barks of dark-barked birches, B. ermani in Japan and B. nigrll in North America. Especially in the latter species, betulin and lupeol were no longer the predominant triterpene components, and 3-O-acetyloleanolic acid predominated [40]. Triterpene contents in inner bark and root bark were very much smaller than those in outer bark. It is noteworthy that all the triterpenes isolated were caffeoylated in the root bark of B. ermcmii [41]. Betulin oligosaccharides were enzymatically synthesized by using cyclodextrin glycosyl transferase withan aim to increase its biological activity [42].Unsaponifiables in the barks of Fclgus crenata, Cyclobanopsis glauca, and Castanopsiscuspiciata wereclaimedtobe responsible for “pitch trouble” during kraft pulping process, as lots of pitch specks were observed on pulp sheets prepared from wood chips contaminated with their barks. These barks have larger contents of unsaponifiables than wood 1431.

B.

Low-Molecular-WeightPhenolics

Laks [ l ] dealt withlow-molecularphenolicextractivesincluding flavonoids, lignans, phenylpropanoids,andothersimplephenolicsfound i n tree barks until 1985and cited somereviewsandbooksconcerningthem. In this section onlymore recent results are referred to. Phenolic compounds are present in relatively large amounts in coniferous tree barks. Thebarks of somehardwoodspecies,such as Q w m l s , Euctrlyptus, Accrcicl. and Soli.v, also contain large amounts of phenolic extractives, as exemplified in Table 6, while some of otherhardwoodspecies, Betulcl spp., for instance, containonly negligible amounts of phenolic extractives in their outer barks. Itis likely that phenolic constituents i n the

252

Sakai

CH,OH CH 2OH CH0 CO2H CO2Mc CO ?H C02Mc CO 2H CH3 CH 3 CH3 CH0

AC:

CH,C&

FIGURE 7 Structures of' triterpenes found in Betlrlcr barks.

inner bark and root bark are largely glycosylated so that they can be translocated in the phloem sap.

1. Monoaryl Compounds, Lignans, andRelated Products A variety of phenolic compounds has been found in bark, as described by Laks i n the first edition of this volume I l]. Common natural products with a phenylnlethane (C,-C,)skel-

etonare the hydroxylated benzoic acid and benzaldehydederivatives which have been found in many plant materials including barks of many softwood and hardwood species. A recent example in this regard is an HPLC evaluation of Elrcalyfrrs glohulus bark, which

Chemistry of Bark

253

TABLE 5 Triterpenoid Composition in the Barkof Betula spp.

Composition (% of the bark) B. vrrrrcosa

[I1

Compound (structures in Fig. 7) Betulin (=Beturinol) (1) 3-0-caffeoylbetulin (2) Betulinic aldehyde (3) Betulinic acid (4) + methyl betulinoate (5) 3-0-Acetylbetulinic acid (6) Methyl 3-0-acetylbetulinoate (7) 3-0-caffeoylbetulinic acid (8) Lupeol (9) 3-0-caffeoyllupeol (10) Lupenone (11) Lup-20(29)-ene-3-one-28-al(12) Oleanolic acid (13)

3-0-Acetyloleanolic acid (14) 3-0-Caffeoyloleanolic acid (15) 3-P-Acetoxyolean-I 1-oxo- 12-ene28-oic acid (16) Monogynol (17) 3-0-Caffeoyldammarendiol I1 (18)

Others

0

B. phtyphylla var. jnponicc~ B. nigra ~ 9 1 l401

B. ermanii [41]

0

0

0

I”

26.20

0.134 0.067

3.515

-

0.030

-

-

-

-

-

0.011 -

1.34 -

-

0.03 0.96

0.010 0.006 -

-

-

-

0.01 1

1.oo 0.14 -

0.038

0.07

-

R

0.062

0.129 0.462 0.015 0.02 1

-

“0,1, and R denote outer bark, inner bark, and root bark, respectively. hIncludes methyl oleanolate and 3-0-acetyloleanolic acid.

revealed the presence of gallic, protocatechuic, vanillic, and ellagic acids, and protocatechuic aldehyde, together with some flavonoids and ellagitannins [46].Some ellagic acid derivatives, methylated and/or glycosylated, were isolated from the stem bark of Diplopanax stachvanthus [47]. The bitter-tasting bark of Prunus gra-yancl contains phenylethanoid (C,-C,) glucosides and their caffeate esters, together with a tannin-related 3,4,S-trimethoxybenzoyl-glucoside. The latter two had a strong bitter taste [48].Arylpropanoids (C,-C,) biologically synthesized from the amino acid phenylalanine are very common in the plant kingdom and are constituents of the naturally important phenolic polymers, lignins and condensed tannins. Arylbutanoid (C,-C,) glycosidesareknowntobe present in the inner bark of Betula pendula 1491 and in the stem bark of A c e r nicoense [SO]. Bioactive styryl-lactones, goniofufuranone, goniopypyranone, and 8-acetylgoniotriol (Fig. 8) were isolated from the stem bark of Goniothalarnus giganteus [Sl]. Thesecanberegarded as phenylheptanoids (C,-C,). Isorhapontin, one of the stilbenes (C(,-C2-C,),was found to be present in very high content (13%) in the bark of Picea engelmanni. This content was dramatically different

I

Sakai

254 TABLE 6

Contents of Phenolic Extractives in Barks 70% acetone extracts

Total phenolics

Plant name

(%)

(%)

S. roricln S. kinrcyanagi 13.7

34.9 30.0 31.0 27.2

17.2

13.2 S. gilgiana S. grncilistda 12.9 Run1e.r hastotus Acncicrniloticcc 21.3 A. auriculiformis A. fcrrnesiana 18.7 A. leucophloecl A. ferruginecc4.6 A. torta 4.6 A. ctresicr 6.6 A. srtrldnr A. dedbcrta 14.0 A. larronutn 15.3 Bruguiern gymnorrhiza ' 13.1 B. gytlrrorrhizrc h Ceriops tcrgrcl"45.2 Rhizopharc~apiccclcrtcr 19.3 '' R. rnucro~ntrtrr" 29.4 R. stylosoh 26.2 35.2 Xylocccrpcs grcrncrtutn 17.9 Sorrtrernticr cclhcr"

Tannin content

(a)

(%l

Ref.

17.7

9.7 7.7 9.0 5.6

[101

15.0

-

16.8 11.8 10.4

14.4 9.9

33.8 25.8 35.5 25 .o 13.3 19.6 24.7

1451 - [45 -

17.2 13.2

3.O

9.5

21.7 36.9 22.8 24.8 45.8 28.4 41.8 32.6 35.9 24.2

Total flavanols

t451 13.9

t651

-

9.9 30.7 25

[651

21.9 25

[ 101

[101 [ 101

1441 19.2 15.8 15.4 12.8 3.7 4.1 6.2 2.6 13.3 12.3

12.0 8.8 35.4 15.1 (651 24.4 22.2 27.6

10.0

15.5

13.8

t451

1451 1451

1 1451 [451

1451 1451 1651 1651

12.4

.o

1651

.o

t651 1651

"Grown in Thailand. hGrowninJapan.

from those in other softwood barks from five genera, suggesting isorhapontin to be a good taxonomic marker [S2]. Phenyl propane dimers (C&,-C$,) linked at the P-positions are referred to as lignans, and ones linked at a position different from the P are sometimes called neolignans. A book of lignans in general [S31 and a review of lignans from woody plants [S41 was published recently. In the root bark of A c c r n / / ~ r ~ ~ ~ r rsrn/ico.su.s, nc~x an empiricaloriental medicinal plant. five lignans, two coumarins, and three other phenolic compounds were

Gonyofufuronc FIGURE 8

Goniopypyronc

Structures o f biouctive styryl-lactoncs.

X-Acctylconiotriol

Chemistry of Bark

255

identifed. Seven of these compounds were glycosides and one was caffeate. The predominant constituent was syringaresinol diglucoside 1551. A neolignan rhamnoside and a lignan xyloside were isolated from the inner bark of Ostrya japonica [57]. Betula ermanii inner bark contains two lignan glycosides and a dicaffeoylated lignan [41]. Contents of lignans were compared between the wood and bark of Magnolia kobus var. borealis, which has been used as a material for Oriental medicine blends [58].The bark contained much larger amounts of lignans, especially permethylated lignans, magnolin, and yangambin. The bark of Ocotca catharinensis contains I O neolignans(average yield 0.07%) of thebicyclo[3.2. lloctanoid and the hydrobenzofuranoid structural types [59]. These neolignans are also present in much larger contents in the bark than in wood of the same species. A series of diarylheptanoid (C,-C,-C,) glycosides together with the aglycon, platyphyllone, was identified in the inner bark of Betula pendula [49], and a diarylheptanoid glycoside platyphylloside was found to be present at high levels in B. pendula (20-60 mg/g dry bark) but at low levels in B. pubescens (
2. Flavonoids and Proanthocyanidins Barks of almost all tree species contain flavonoids with carbon skeleton C,-C&, (Fig. 9). The first edition [ l ] dealt with somecommon flavonoids includingquercetin,dihydro-

R5

R5

Catechin(CAT)

2R :3s

H

Epicatcchin (ECT)

2R:3R

OH

H OH OH OH OH H

ent-Catechin (entCAT)*

2S:3R

OH

H

Galloatechin (GCT)

2R : 3 s

OH

Fisetinidol (FIS)

2R :3s

H H

Robinctinidol (ROB)

2R :3s

H

OH OH OH OH OH OH

OH OH OH OH OH OH OH OH OH

H H

H OH OH OH OH

OH OH OH OH OH

H

FIGURE 9 Basic structure, numbering system, and configuration offlavonoids flavan-3-01s). *Use the right structure for ent-flavan-3-01s.

(exemplifiedfor

I

Sakai

256

quercetin,myricetin,and their glycosides found in the bark. After that, some excellent reviewsandbooksonflavonoids and related compoundswerepublished[15b,66,67], though they are not confined to the bark components. The flavan-3-01s catechin and epicatechin are very commonly present in the bark, probably as precursors to and by-products of the biosynthesis of condensed tannins. Flavan-3-01 glycosides were thought to be rare. However, some were recently found in the inner bark of different species, for example, catechin 3-O-a-~-rhamnopyranosidein Quercus miyagii bark [68]. catechin 3-o-p-D-ghcopyranoside and 3-O-[au-~-rhamnopyranosyl-( 1 +6)-P-~-gulucopyranosyl]-catechin in Quercus marilandica bark 1691, and catechin 7-O-~-D-glucopyranosidein the inner barks of Betula pubescens 1701 and Pseudotsugu menziessii [71]. The latter also contained epicatechin 7-O-p-D-glucopyranoside, catechin 4'-O-P-~-glucopyranoside, and 3'-O-methylepicatechin 7-0-P-D-glucopyranoside as flavan-3-01 glycosides [71]. The position 3 of the flavanols is sometimes galloylated, and secretagogue activity of 3-0-galloyl-epicatechin was reported [72]. Unstable gallates of catechin esterified at the 3',4' and/or 7 positions were recently found in Acacia gerrardi bark (731. Gallic acid, catechin, and epicatechin showed inhibition of chemiluminescence production by the activated human polymorphonuclear leukocytes, which is a measure of oxidative burst [74]. There has been an expolsive growth in the understanding of the chemistry of plant proanthocyanidins over the past two decades, due to the development of techniques in chromatographic isolation andspectrometric structural analysis, including 'H- and "CNMR and MS [66,75].Manyproanthocyanidindimersand trimers and sometetramers have been isolated from barks of many species in recent years. The nomenclature system commonly accepted for the proanthocyanidins is outlined as follows 1761: (1) The fundamental structural units are defined in terms of the monomeric flavan-3-01s. The names catechin, epicatechin, fisetinidol, etc., are reserved for units with 2R absolute configuration. Flavan-3-01 units with the 2s configuration are distinguished by the "ent" (namely, enantio) prefix. ( 2 ) The interflavonoid bond at C-4 is indicated by an arrow and its configuration by the ap nomenclature in parentheses, as (4p -+8). Structures of some proanthocyanidin oligomers are shown in Fig. 10. Their flavanol units are often linked by (4a -+ 8) bond in 2R: 3 s units, such as catechin, or (4p + 8) in 2R:3R units, such as epicatechin. In some cases there are ( 4 a + 6) or (4p + 6) bonds, too. Some of them have 0-acyl (galloyl usually) or 0-glycosyl bond at position 3. Basedon the structuresof proanthocyanidin oligomers. structures of polymeric proanthocyanidins or condensed tannins have been constructed. Some of the proanthocyanidin oligomers have physiological or biological activities. The bark of Psrudotsuga menziessii contains an epicatechin tetramer (compound 5 in Fig. 10) that is a potential inhibitor of rat brain protein kinase C as well as bovine heart cyclic AMP-dependent protein kinase [77]. From the bark of Byrsonima crassifoliu, which hsa beenusedmedicinally by the MixeIndians to treat gastrointestinal disorders and skin infections, a proanthocyanidin trimer (4 in Fig. 10) and related ent-epicatechin-based oligomers were isolated as guided by a nematodicidal activity assay [78]. The stem bark of Stqphonodendron adstringens, a medicinal plant grown in Brazil, contains a variety of 3-0-acylated prodelphinidins (epigallocatechin-based proanthocyanidins) [79]. Antitermite activity was observed with a low-molecular-weight proanthocyanidin fraction from Acacia mearnsii bark, consisting of dimeric and trimeric prorobinetinidins [SO].

C.

HydrolyzableTannins

In general, galloyl and the related hexahydroxydiphenoylesters of glucosehavebeen referred to as gallotannins and ellagitannins, respectively, because they are easily hydro-

257

Chemistry of Bark

H0

H0

OH

OH

OH

3

1 R=H 2 R=Gk

H0 H0

OH OH

OH

4

NO.

5 plant spccics

Structure many

Ref.

General

l

CAT-(4a+8)-CAT

2

ECT-(4R+8)-CAT

Gcncral

3

~-O-G/C-CAT-(~U.+X)-CAT

Quercus mutihndicu

many

[69]

~-~-G~I-ECT-(~B+~)-~-~-G~I-C~~-ECT-(~U.+X)cruss4~~,iu ,7Kl cnt-ECT

5

ECT(4lJ+8)-ECT-(4l~+X)-ECT-(4~~8)-ECT

Pseudotsugu menziesii [ 771

FIGURE 10 Structures of proanthocyanidin oligomers recently identitied in tree barks. C A T catechin, E C T epicatacchin, ent-ECT: ent-epicatechin, Glc: P-D-glucopyranosyl, Gal: galloyl.

258

Sakai

lyzed by acid to give gallic acid and ellagic acid. A recent book and a review of Haslam [81,82] gave an extensive discussion of hydrolyzable tannins. Description in this section is confined to bark tannins, although tremendous numbers of hydrolyzable tannins have been isolated during the past decade from various materials, including fruits, leaves, and wood parts of different plant species [84]. Recently, Nonaka [83] suggested that some of the hydrolyzable tannins can be called “complex tannins” based on their structural features, which have a flavonoid, flavanol, or proanthocyanidin moiety in addition to the more common structural constituents, polyols (mostly D-glucose) and the gallic acid derivatives. Quite a few novel gallotannins, ellagitannins, and complex tannins have been found and structure-determined in the barks of Quercus acutissima [85,89], Q. miyagii [85,89], Q. stenophylla [85,89], Q. rnongolicu var. grosseserrata [85,89,90,86], Q. denata [89], Q. petraea [91], Castanea crenata [85], Castanopsis cuspitada var. sieboldii [89], Cercidiphyllurn japonica [92], Mallotus japonicus [93,94,96], M. philippinensis [96], Pterocatya streptera [95], Eucalyptus viminalis [95], Hammamelis virginicrna [88], Anogreissus acurninata var. lanceolata [87], Psidium guajuva [97], and Platycarya strobilacea [98]. A few examples of hydrolyzable tannins are shown in Fig. 11. It is noteworthy that the configurations at the glucose C-l position of representative C-glycosidic ellagitannins, 1S for castalagin and 1R for vescalagin [99], were revised to the opposite configurations, 1R and lS, respectively, on the basis of NMR spectroscopy [ 1001. Accordingly, the configurations at this position must be revised for all C-glycosidic tannins known before 1990, including ones in Haslam’s book [81] and stenophyllanin B shown in the first edition [l].

D. Condensed Tannins Condensed tannins distribute more widely than hydrolyzable tannins in the plant kingdom and are mostly polymeric proanthocyanidins with structures closely related to the oligomeric structures depicted in Fig. 10. Somebooksandreviewsdealingwithcondensed tannins have been published recently [ 15b,81,101- 1031. The principal types of condensed tannins and variations in their hydroxylation patterns of the A- and B-rings are shown in Fig. 12. Their chemical characteristics are strongly influenced by hydroxylation patterns of the A-ring; thus procyanidins and prodelphinidins with a phloroglucinol type A-ring tend to react more easily than profisetinidins and prorobinetinidins having the resorcinolic A-ring, with both nucleophilic and electrophilic reagents. Molecular weight or degree of polymerization (DP) of these compounds also affects their nature; thus condensed tannins of low DP are soluble in water and some polar solvents, whereas high-DP tannins are not extractable in neutral solvents. Large parts of polyphenols in the conifer barks are solubilized by alkaline extraction following neutral solvent extractions. The extracts formerly called “phenolic acids” are now known to be artifacts produced from rearrangements in high-DP procyanidins in the presence of alkali [ 1,161. In this connection, Yazaki et al. [l041 reported the detrimental effect ofalkalinetreatmentof the extractsfrom Pinus radiata barkon their Stiasny values, which is a reliable estimate of the polyflavanoids reactive to formaldehyde. Characteristics of condensed tannins from the barks of Pinus taeda, a number of Japanese tree species, Acacia mearnsii, andsomeotherconiferousspecieshavebeen studied as described in the first edition [l]. More recently, the structure of procyanidintype polymers of Douglas fir (Pseudotsuga menziessii) inner bark was studied using degradation with phloroglucinol as an analytical method. The results suggest that the config-

259

Chemistry of Bark H0

OH

H0 H0 H0 H0

OH

OH

Castalagin Miricu esculentu [61] Quercus petrueu [ 011

Acutissimin A (a furavano-cllagitannin) Quercul ucutissimu, Q. miyugii, Q. stenophyllu, Q. tnongolicu var. grossesenutu, Custuneu crenutu [ 8 5 ] ,Q.petrueu (911, andhogeissus ucurninutu var. lunceolutu [ 871

OH

OH

OH

Mongolicanin (a procyadino-cllagitannin) Quereus tnongolicu var. gro.s.serrutu [ 861 Q.petrueu [ 91 1 FIGURE 11

l-O-(4-Hydroxybcnzoyl)-,2',5-di-0galloyl-IJ-D-hamamclofuranosc Hutnutne1i.s virginiunu [ 881

Hydrolyzabletanninsisolatedfrom tree barks.

uration of the extender units is almost exclusively 2,3-cis, while the terminal units are mixed, with 2 , 3 4 7 slightly predominating. The (4+8) interflavonoid bond predominates over the (4+6) bond by a 4: 1 ratio [ 1051. Condensed tannins from the barks of a number of angiosperm species have been characterized. Ohara et al. [80] studied the acetone-water extracts from the bark of Acacia mearnsii (52.7% of dry bark) and reconfirmed the condensed tannins of this bark to be mainly 2,3-trans-prorobinetinidinswith a relatively small average DP of 5 (poly dispersity

260

Sakai 3'

5

5P C Propelargonidin Procyanidin Prodelphinidin Proguibourtinidin Profisetinidin Prorobinetinidin

4

pdttcm

Hydroxylation 3,4',5,7 3,3',4',5,7 3,3',4',5,5',7 3,4',7 3,3',4',7 3.3'.4'.5'.7

FIGURE 12 Principal types of condensed tannins and their hydroxylation patterns.

1.36),containingsmallamountsofprocyanidinorprodelphinidinchain-extender units. The angular structure was suggested by the coexistence of angular trimeric proanthocyanidins in the same bark extracts. No distinctive difference in molecular size distributions was observed between the A. nzearnsii bark tannins from China and South Africa [ 1061. Barks of Quercus species seem to contain both hydrolyzable tannins and condensed tannins [85,89,91,107,108J. Tannins are supposed to be responsible for the therapeutic effects of oak (Q. petraeu) bark, which is used against hemorrhoids, chilblains, mouth sores, and indigestion in Europe. The angular or branched structure, with average DP of 6.1 and a procyanidin-to-prodelphinidin ratio of 6:4, wasproposedfor the condensedtannins of Q. perruea bark [107]. The bark of Q. fulcatu, the most important species of red oak in the forests of the southern United States, is a rich source of quercitrin (quercetin-3-rhamnoside). It contains procyanidin-type condensed tannins made up predominantly of 2,3-cis extender units and terminated with 2,3-truns catechin units, together with a flow concentration of catechin and three major dimeric procyanidins. The tannins contain only small amounts of 2,3-trans-procyanidin extender units and only traces of prodelphinidin units [ 1081. The condensed tannins of Salix spp. bark have comparatively large DPs for angiosperm bark tannins. Those present in S. roridu bark consist of prodelphinidin and procyanidin types of flavanol units with an average DP of about 8 and with 2,3-truns and 2,3-cis configurations in the extender units and 2,3-truns in the bottom or terminal units. Prodelphinidin units are more abundant than procyanidin units [IO]. In the case of S. petsusu bark, the condensed tannins, with average DP about IO, are composed of both procyanidinandprodelphinidin types, with the formerdominating.Extender units in the tannins have largely 2,3-cis and a small proportions of 2,3-rruns configurations [ 1091. The bark of S. sieboldiann, a shrub common in Japan, contains a homologous series of procyanidindimersand trimers esterified with 1-hydroxy-6-oxo-2-cyclohexene carboxylic acid at position 3 of the catechin bottom units [ 1 lo]. Three oligomeric proanthocyanidin fractions, with average DPs of 4, 5-6, and 6-7, respectively, were found in a blood-red sap from the slashed bark of Croton lechleri, used by South Americans for the treatment

Chemistry of Bark

26 1

of numerous illnesses and diseases. These oligomers were prodelphinidins consisting of different proportions of 2,3-cis and -trans extender units and contained small amounts of procyanidin units [ 11l]. Polymeric anthocyanidins that inactivated cholera toxin were isolated from the bark of Guazuma ulmifolia, which is used by the Mixe Indians to treat diarrhea [ 1121. The bioactive compounds, with average DP ranging from 14.4 to 32.0, consistedmainlyof(-)-epicatechin units whichwereconnected by (4+8) bondsand, less frequently, by (4+6) bonds. Inhibition of cholera toxin by the condensed tannins from the G . ulmifdicr barkincreasedwithDPandconformation flexibility of the tannin molecule. The bark of Bruguiera gymnorrhiza, a commercially important mangrove species in the estuaries of Indonesia,containscondensedtanninswhichhaveprodelphinidinand procyanidin types and predominantly 2,3-cis configuration in the chain-extender units and 2,3-rrans stereochemistry in the bottom units. From this tannin preparation 3-O-a-~-rhamnopyranosyl-catechin-(4a+2)-phloroglucinol was formed upon acid-catalyzed cleavage in the presence of phloroglucinol, thus providing evidence for covalently bonded glycoside moieties in the chain-extender units of mangrove bark tannins [ 1 131. It is now recognized that glycosylation of proanthocyanidins in plants isnot exceptionally rare, even though proanthocyanidin glycosides are not commonly found in most plants. These structural elucidations of condensed tannins have been performed by development of different techniques including a procedure combining their chemical degradation with 'H-NMR spectroscopy [ 1 1 l], high-temperature 'H-NMR spectra of their methyl ether acetate derivatives [ 1141, negative-ion fast-atom bombardment spectrometry [ 1 151171, and 'H- and "C-NMR spectroscopy [ I 1 8,1191, as well as the standard analytical method of thiolysis.

E.

Biological Activityof Bark Components

The barks of many species have been used as materials of traditional medicine. Perhaps the most promising medicinal compound from bark is taxol from the Tcu-us sp. (Fig. 4), an important cancer chemotherapeutic agent with a unique diterpenoid structure, as described above. Some examples of such medicinal barks which are now used in China and Japan [ 120,121] are shown in Table 7. The compounds identified in each bark extract also are included in the table, although the extracts are in fact very complex mixtures and isolated compounds often do not account for all the efficacies ofwhole extracts. Acompound composed of monoterpene and chalcone moieties (Fig. 3) was isolated from the bark of Lindera umbellato and significantly inhibited melanin biosynthesis of cultured B- 16 melanomacellswithoutcausingany cytotoxicity in the culturedcells or skin irritation in guinea pigs, suggesting its potentials in use asahumanskin-whiteningagentand/ora remedy for the disturbances in pigmentation[27].Goniopypyrone(Fig. 8), abioactive styryl-lactone isolated from the stem bark of Goniothalamus giganteus, showed nonselective EDSo values of about 0.7 pg/mL in each of three human tumor cell lines [51]. Syringaresinol diglucoside, the predominant constituent in the root bark of Acanrhopanax senticosus, protected rats to a significant extend from fatigue induced by chronic swimming stress [ S ] . Infusions of the bark of Cassia abbreviuta are used to treat blackwater fever, abdominal pain, and toothache in Africa. Guibourtinidin dimers were isolated from this bark [ 1221. Forlines et al. [ 1231 introduced the use of polyphenols by Native Americans on the west coast of Washington State's Olympic Peninsula. They have made materials and med-

TABLE 7

Barks that Have Been Used as Sources of Oriental Medicine (120,121]

Species

Parts"

Lycium chinense Fraxinus japonica Acanthopanax giraldii. A. gracilistylus, A. spinosum Punica granatum Rhus vernicijua Melia azedarach Phellodendron chinense, P. amurense Albizzia julibrissin

rb b b, rb

Antiphlogistic, antipyretic, roborant Antiphlogistic, stypsis Dermatitis, analgesic, roborant

b, rb e b b b

Antiparasitic Emmenagogue, vermicidal, antitassive Antiflatuent, vermicidal (tapeworm) Stomachic, antiflatuent, stypsis Roborant, analeptic, antitussive, analgesic, external use for contusion and fracture Roborant, analgesic, sedative Antiphlogistic, antipyretic, sedative, hematocatharsis Aromaticstomachic, carminative, stypsis, antinausant, hidrosis, antipyretic, dedative Stypsis, diuresis, expectoration

Eucommia rclmoides Paeonia suflruticosa

b rb

Cinnamomum cassia, C. loureirii

b

Magnolia ohovata, M . ofJicinalis

b

Morus alba, M . bombycis

rb

Efficacy

Antiplogistic diuresis, lasative, antitussive, expectoration

Compounds identified Betaine Aesculin and other coumarins Lignans including syringaresinol diglucosideb Pelletierine and ellagitannins UrusioIs Sugiol, nimbiol, nimbin, rnelianone Berberine, palmatine, and other alkaloids Acacic acid and saponin Gutta-percha Paeoniflorin Essential oil containing cinnamaldehyde, cinnamyl acetate, etc. Monoterpenes, P-eudesmol, magnolol, and magnocurarine Scopoletin, urnberlliferone, moran A

"b, rb, and e denote bark, root bark, and exudate, respectively. hFound in A. senficosus.

k?

X

m.

Chemistry of Bark

263

icines for centuries from a widevariety of forest plants, including the bark of Alnus rubra, Tsuga heterophylla, Thuju plicata, Berberis nervosa, Rhamnus purshiana, Oplopanux horspp., Salix spp., and Taxus brevifolia. ridum, Sambucus racemosu, Pyrus fusca, Prunus Many of the medicinally used plants were rich in procyanidins and associated compounds. Biological activities of some oligomeric and polymeric proanthocyanidins isolated from barks were also described in the sections on “Flavonoids and Proanthocyanidins” and on “Condensed Tannins.” The antifungal activity of bark components has been known and the activity has been traditionally utilized in fermentation of Philippinesugarcanewine.Apolyphenol component obtained from an aqueous extractof the samac (Macharanga grandqolia)bark, which gave catechin, cyanidin, delphinidin, and sugars upon acid hydrolysis, inhibited the growth of lactic acid bacteria which caused deterioration of the fermentation mixture [ 1241. In certain communities of southern Nigeria, the pharmacologically active aqueous extract of the bark of Sucoglottis gabonensis is commonly used as an additive to palm wine. The exact biological role of this extract in the beverage maturation process is not clear, but it exhibits antioxidant properties and at least one constituent, bergenin, is an inhibitor of yeast alcohol dehydrogenase [ 1251. It is interesting that hot water extracts from the inner barks of three coniferous species (Cryptomeria japonica, Chamaecyparis obtusa, and P i n u s d e n s g o r a ) showed almost no inhibition of the mycelial growth of all the edible mushroom fungi which are wood decay basidiomycetes, while these extracts exhibited great inhibition of two fungi that are pathogenic to the mushroom fungi [ 1261. Acetone extracts of a bark-compost possessed strong antifungal activity against Fusarium oxysporum, Helminthosporium sigmoideum, Gibberella zeae, etc., fungi of interest to plant pathologists, but the extracts were inactive against yeast and procaryotic organisms [ 126bl. Mori et al. also observed that all the bark extracts from 21 conifers belongingtosevenfamilies inhibited the growth of plant pathogenic fungi more effectively than that of wood decay fungi [ 1271. The feature of the bark being degraded by wood decay fungi is perhaps of important significance in the mass cycle in nature, where dead trees are biologically mineralized by organisms including insects, earthworms, fungi, and bacteria. On the basis of this property of bark components, biodegradable polyurethane foams were prepared from commercial wattle tannn [ 1281, the barks of Acacia mearnsii and Ctyptomeria japonica[ 129,1301. The foams produced were gradually degraded by both white-rot and brown-rot fungi. However, biocidal activity of condensed tannins against wood decay fungi was intensified by complexing with copper(I1) ions. The best wood-preservative effect was observed with a dual treatment using a sulfited bark extract first, followed by a CuCl, treatment. This method yielded wood with greater resistance to decay than wood treated with pentachlorophenol [ I3 1 1. Condensed tannin from several tree barks exhibited enzymeinhibitory activity against plasmin, thrombin, papain, and trypsin. A tannin from Cedrus deodara bark showed anti-plant-viral activities against tobacco mosaic virus, potato virus X, and cucumber green mottle mosaicvirus [ 1321. Antimicrobial properties of tannins, not only from bark but also from wood, leaves, fruits, and roots of both woody and herbaceous plants, were reviewed by Scalbert [132b]. Bark extracts influence animals physiologically, too. 3-0-Acetyloleanolic acid, the predominant triterpene component in Betula nigra outer bark, exhibited significant antifeedant activity against the Colorado potato beetle, Leptinotarso decemlinoata, an agriculturally important insect pest [40]. Some flavonoids and cerebrosides were isolated from Quercusdentata,Eucalyptusrubida, and Prunusjamasakura as repellent compounds against the blue mussel, Mytilus edulis, one of the gregarious fouling organismsthat cause

Sakai

264

serious problems to ships and fishing nets [ 1331. Tannins in the bark extract of Norway spruce (Picea abies) as a model for debarking waste water were shown to be responsible for acute and subacute toxicity in carp. Oxidative polymerization of tannins in the extract into high-molecular-weight polymers abolished the aquatic toxicity [ 1341. The wastewater from debarking plants is highly toxic to anaerobic microorganisms also, especially methane-forming bacteria. It was shown that tannin, estimated by adsorption to insoluble polyvinylpyrrolidone, caused most of the methanogenic toxicity observed in conifer bark extract [135], and they can be mostly detoxified by alkaline autoxidation [136], hydrogen peroxide oxidation [ 1371, and enzymatic polymerization with phenol oxidases followed by flocculation [ 1381. The detoxified wastewatercan be treated using conventional anaerobic methods [ 138,1391. Swan[l401 reviewed health hazardsincluding toxic, allergenic, and carcinogenic properties associated with extractives from woody plants.

V.

UTILIZATION OF TREE BARKS

Only recent developments in the technologies for adhesives production from bark tannins and other recent topics of bark utilization are dealt with here. Laks [ l ] reviewed extensively the utilization of tree barks as a source of adhesives, pharmaceuticals and biocides, and other utilizations in the first edition.

A.

Adhesives

Condensed tannins from the bark of Acacia mearnsii (black wattle) tree have been successfully utilized for adhesive purpose for over two decades in South Africa. The wattle tannin-based adhesives have been progressively displacingsynthetic phenol-formaldehyde resins, in countries where these tannins are produced, and are consumed by the manufacture of particleboard, plywood, glulam, fingerjointing, and cardboard. Condensed tannins from other species have been intensively investigated for adhesive applications, but they could not be exploited as readily as wattle tannins. Pilot production of additives from the Tsuga hererophylla bark extract in the United States was not cost competitive with petrochemical phenol. Commercial-scale production of adhesives for particleboard from Pinus rcldiata bark in New Zealand was unfortunately discontinued in the early 1980s. A comprehensive description of the situation prior to 1985 was presented in the first edition [l]. The worldwide production of wattle (A. nzearnsii bark) and quebracho (Schinopsis spp. heartwood) tannins was 150,000 tons per year in the early 1990s, of which not more than20-30% are available for adhesiveapplication, the rest beingreserved for their traditional leather market [141]. In China, plantations of the tree A. nzecrrnsii have been rapidly increasing, and tannin production was expected to increase from 100 tons per year in 1989 to nearly 2000 tons per year in 1994 [142]. Zhao et al. have studied to develop production technologies for wattle tannin adhesives which would be usable for exteriorgrade plywood under Chinese factory conditions, which requires a wide range of closed assembly time from 30 min to 16 h [ 106,142- 1441. The polyflavanoid contents of Chinese wattle tannins evaluated as Stiasny values were higher and their molecular weight distributions were very similar, compared to those of South African commercial tannin [ 1061. Adhesives were successfully formulated from Chinese commercial wattle tannin fortified with phenol-formaldehyde or phenol-urea-formaldehyde resins as cross-linking agents, instead of paraformaldehyde, whichis relatively expensive in China. Factory trials strongly

Chemistry of Bark

265

indicated that the adhesives would be extremely applicable to the Chinese industry manufacturing exterior-grade plywood. Recently, Santana et al. reported that the bark of A. mearnsii was liquefied in phenol in the presence of sulfuric acid catalyst [ 1451. The resulting solution was reacted with formalin in a basic solution to yield a resol resin that, under the best conditions, performed similarly to the commercially available phenol-formaldehyde resins in plywood adhesion. Studies of the adhesive preparation from the bark have been performed with other hardwood species, including mangrove trees, Khaya ivorensis and Avicennia alba [1461, and Moroccan afforestation trees, Eucalyptus astringents, E. Sideroxylon, and Acacia decurrens [147]. However, commercially feasible technologies for production of adhesives from conifer bark tannins, if developed, would be accepted more widelyin the world, as Coniferous trees have a very large plantation area, and wood industries in different countries are using huge amounts of them for timber production and producing large quantities of conifer bark every year as by-products. Conifer bark tannins will be supplied more stably than hardwood bark tannins if their utilization is established. Therefore, serious and intensive work has been done with bark tannins of different coniferous species, including Picea abies [21,148,149], ktrixleptoides [150], Pinussylvestris [21,148,149], P. halepensis [151,152], P. pinaster [21,153-1561, P. radiata [14,157-1651, P. caribaea [21], P. elliottii [21], and the U.S. southern pines [166-1681. Interest has been increasing in the utilization of pine tannins as materials for adhesives in such countries as Australia, Chile, and the United States. However, difficulties have been encountered in the utilization of tannin extracts from softwood barks, primarily due to low extractable yields, excessive viscosity, and much faster reactivity toward formaldehyde, as stated in the first edition [ l]. Yazaki et al. reported that the extraction of P. radiata bark using a four-stage squeeze extraction provided a high extractives yield (approximately 30%), despite a ratio of 1 bark to 3 solvent used (by mass) [ 1591. The combined extract had viscosities greater than 8000 mPa. S at 40% solid content at 25°C. However, when part of the extracts was sulfited with sodium metabisulfite under reflux for 2 h and combined with the rest of the extracts, the viscosity of the combined extracts was suitable for the formulation of adhesives which provided high-quality glue bonds [ 1621. The gluing properties of the adhesivesderivedfromextracts of differentbarkspecies appeared to be dependent on their contents of formaldehyde-reactive polyflavanoids as indicated by their Stiasny values, with a value of 65% being the minimum for producing a high-quality adhesive by this method [21]. The viscosity modification observed with the sulfite treatment is based on the findings of Foo et al. that sulfonation of 5J-dihydroxyproanthocyanidins involves interflavanoid bond cleavage formingprocyanidin-4-sulfonates and not formation of polymeric sulfonates as had originally been thought [ 1691. Furthermore, sulfonates (Y to a phloroglucinol ring, namely, procyanidin-4-sulfonates, are good leaving groups at ambient temperature and pH greater than 8.0. Consequently, under tYPical adhesive formulation conditions, the sulfonic acid groups on tannin derivatives from conifer barks will be displaced, resulting in water-insoluble polymer [ 1701. Kreibich and Hemingway have studied ways to develop tannin-based adhesives and recently reported and reviewed that tannins extracted with sodium sulfite solution from the U.S. southern pine bark were able to replace about 50% of phenol-resorcinol-formaldehyde resin in cold-setting wood adhesives. Bonds in laminates exceeded the requirements of the American standards for dry-sheer and vacuum-pressure cold-water soaktests [ 166,1671. Part of the resorcinol in adhesives made of resorcinol, formaldehyde, and a

I

Sakai

266

styrene-butadiene-vinyl pyridine terpolymer latex for bonding nylon cord to rubber could be replaced with purified loblolly pine tannin [171]. In Chile a company began in the late 1980s to extract a range of midly sulfited and nonsulfited tannins from Pinus rudiata and f? insignis mixed bark, to serve the requirements of the local leather industry as well as to supply tannins to other applications. Based on laboratory formulations using commercially produced extracts, Leyser and Pizzi [ 1571 developed a new “honeymoon” fast-set system that is a two-component system in which one part is a conventional phenol-resorcinol-formaldehyde resin with the pH adjusted to a value greater than 13 (pot life = indefinite), and the other part is the sulfited pine tannin extract containing excess paraformaldehyde (pot life 5 h). The result of the fingerjointing and glulam industrial trial using the system satisfied the relevant international standards specifications. More recently, mill trials for application of basically the same honeymoon cold-set adhesive system were conducted for structural end joints in six different mills in North America. The phenol-resorcinol-formaldehyde/tannin honeymoon system was capable of producing cold-set, fully exterior-grade end joints at mill production rates [ 1681. However, yield of the pine tannin, industrially extracted in Chile, was comparatively low (less than 20%). The phlobaphenesformationand precipitation during sulfite/water extraction of pine tannin from pine bark was minimized by blocking tannin self-condensation by the addition of smallamounts of astrongnucleophilesuchasphloroglucinol. mphenylenediamine, and urea, the latter due to its low cost for industrial application. The yield of pine tannidurea extracts increased from 19% to 25% in industrial scale, and the extracts proved to give good thermosetting wood adhesives for panel products [ 1581. Pinetannin-basedadhesives for exteriorparticleboardhave been obtained by the reaction of polymeric MD1 (4,4’-diphenylmethane diisocyanate) with mildly sulfited pine tannin extract [141]. This type of adhesive from wattle tannin was already developed in the early 1980s [l]. Pinetanningave better results because of a faster reaction of the phloroglucinolicA-ringwithformaldehyde, resulting in formation of methylolgroups which in turn easily react with isocyanates to form urethane bonds [141]. The characteristics of the particleboards obtained industrially by using MDI-fortified pine tannin adhesives satisfied well the requirements of exterior-grade waterproof particleboards. Carbohydrates in the bark extract might react also with the MD1 to form urethanes, as pointed out by Laks [l]. Pizzi et al. reported that autocondensation of polyflavanoid tannins is accelerated by Lewis acids and cellulose. Based on this observation, particleboards of excellent internal bond strengths were obtained using a pine bark tannin at pH 10.2 and higher [172]. This afforded the possibility to prepare interior-grade wood binders, presenting no formaldehyde emission at all. In this connection, a possibility of the spruce bark reacting with formaldehyde during manufacture of particleboards and acting as a formaldehyde scavenger was pointed out [ 1731.

B.

Other Use of Bark

Bast fibers from the inner barks of some trees, including Edgeworthin papyriferu, Broussnetia papyrijiera, B. kazinoki, and Wikstroemiasikokiarza, have been long used for production of traditional paper in various Asian countries. Their pulping or maceration have been done traditionally by alkaline cooking. Recently, a biochemical or enzyme pulping process was proposed. the enzymatic fiberization of bast of the different plant species in the biochemical pulping process under alkaline conditions proceeded through the concerted action of endo-pectate lyase and endo-pectin lyase fromsoft-rot Erwinia carotovara, where

Chemistry of Bark

267

the former enzyme was the primary agent for the fiberization and the latter plays a supplementary but indispensable role [ 174,1751. A technology was developed for papermaking from the innerbarkof Cryptomeriajaponica (Japanesecedar),whichhas the largest plantation area in Japaneseforests [176]. The innerandouterbarks are separated by treatment with a hot ammonium oxalate solution that dissolves pectic substances. The bast fiber pulp prepared from alkaline cookingof the separated inner bark is being used suitably for specialty purposes such as calligraphy, artistic wrapping, paper arts, etc., after blending with other bast fiber pulps and/or staining in different colors. Acetic acid pulping of the Pinus pinaster bark after tannin extraction was studied, but properties of the resultant pulp were not evaluated [ 1771. It is not likely to be feasible to make ordinary paper from fibers contained in bark. Enzymatic saccharification of bark polysaccharides has been studied to obtain sugars and ethanol. Alkali-extracted Pinus pinaster bark was enzymatically hydrolyzed with a mixture of cellulase and P-glucosidase to produce sugars, but their yield was less than 10% of theoretical. The sugar yield increased to 75% by successive treatments, extraction with NaOH (15 min), and delignification with acid chlorite (7 h) prior to the enzymatic saccharification [ 1781. Delignification with hydrogen peroxide andacetic acid had a similar effect on the sugar yields [179]. These delignification treatments, however, are not economically feasible. In the case of Picea ezoensis bark, NaOH treatment effectively increased the glucose yield to 40-45% of holocellulose content [ 1801. Glucose in the saccharification mixture was fermented to accumulate ethanol, and 1 kg of the NaOH-treated bark was expected to convert to 140 g of ethanol. The barks of three poplar trees (Populus tremuloides, P. maximowiczii X trichocarpa, and P. trichocarpa X deltoides) were susceptible to dilute sulfuric acid pretreatment, and relatively high levels of enzymatic digestibility of cellulose were observed after the pretreatment [ 18 l]. However, the barks from sweetgum (Liquidambar styraciflua)[ l 8 1 ] and Pinus pinaster [ 1821 were unresponsive to acid prehydrolysis in terms of enzymatic digestibility. Hemicellulosewhichwashydrothermallyextracted from beech bark at approximately 200°C was saccharified enzymatically with high yields. However, the tannins had first to be removed at temperatures of 120- 140"C, as they have an inhibitory effect on xylanases [ 1831. Oil-absorbent mats were very simply prepared from barks of Cryptomeria japonica and Chamaecyparisohtusa. Barkswerecrushedinto fibrous fragments of 1-3 cm in length, treated with water-repellent emulsion, blended with polyester fibers ( 1 cm length) as binders, molded, and heat-treated at 130°C. The mats, consisting of 90% bark and 10% polyester, had a density of 0.07 g/cm3 and absorbed fuel oil to more than 3 g/g of mats [ 1841. The mat may be used to remove mineral and vegetable oils in the effluents from factories, gas stations, kitchens, etc. Porous spherical tannin resin was prepared from Acacia mearnsii tannin by reaction with formaldehyde in nonpolar media, polybutene, under stirring at 60°C [ 1851. The spherical resin, with 139 m2/g of surface area and 0.5- 1.O mm of diameter, adsorbed heavy-metal ions such as Cr(VI), Cd(II), Cu(II), and Fe(J1) [1851871. Adsorption of Cu(I1) ion to the resin was shown to be a physical process [186]. Biodegradablepolyurethanefoamswereprepared by the reaction ofdiisocyanate with poly01 mixtures consisting of commercial wattle tannin or bark, a synthetic polyol, a siliconesurfactant, catalysts, and wateras a foamingagent[128-1301.Someof the foams obtained had densities of less than 0.04 gkm' and thermal conductivities as small as that of commercial polyurethane foams. It is suggested that, based on a model reaction with catechin, isocyanate groups react preferentially with phenolic hydroxyl groups in the B-ring of condensed tannins in the barks to form a urethane bond [ 1881.

Sakai

268

REFERENCES 1.

2. 3. 4. 5.

6. 7. 8. 9. 10. 11.

12. 13. 14. 15. 15b. 16.

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

P. E. Laks, in Wood and Cellulosic Chernistn (D. N. S. Hon and N. Shiraishi, eds.), Marcel Dekker, New York, p. 257 (1991), and literature cited therein. M. Trockenbrodt, IAWA Bull., 11: 141-166 (1990). C. M. Chen and J. K. Pan, HolzJorsch., 45: 1 55- I59 ( 199l ) . H. Pereira, Wood Fiber Sci., 20: 82-90 (1988). P. R. Blankenhorm, T. W. Bauersox, C. H. Strauss, G. L. Stimely, L. R. Stover, and M. L. Dicola, Wood Fiber Sci., 20: 74-81 (1988). G. Vazquez, G. Antorrena, and J. C. Parajo, Wood Sci. Technol., 21: 65-74 (1987). G . J. Leary, K. R. Morgan, and R. H. Newman, Appita, 40: 181-184 (1987). E. Sjostrom, Wood Chenzistv, Academic Press, New York (1981). R. Solar, I. Melcer, and F. Kacik, Cell. Chenz. Technol., 22: 39-52 (1988). S. Ohara and K. Yanagi, Mokuzai Gakkaishi, 41: 406-413 (1995). H. Hafizoglu and M. Reunanen, Holz$orsch., 41: 261-263 (1987). B. Perra, J.-P. Haluk, and M. Metche, HolzJorsch., 47 486-490 (1993). J. G a r p and H. Pereira, HolzJorsch., 51: 225-234 (1997). Y. Yazaki and T. Aung, Holz$orsch., 42: 357-360 (1988). R. Solar, F. Kacik, and I. Melcer, Holz Roll Werkst., 50: 291-294 (1992). L. J. Proter, in NaturalProducts of Woody Plunrs. I (J. W. Rowe, ed.), Springer-Verlag, New York, pp. 651 -690 (1989). P. E. Laks, R. W. Hemingway,and A. H. Conner, J. Chenl. Soc., Perkin Truns. I : 18751881 (1987); P. E. Laks and R. W. Hemingway, HolzJorsch., 41: 287-292 (1987). A. Burmester and W. Kieslich, Holz Roh Werkst., 4 4 : 419-422 (1986). T. D. Lomax, K. L. Mackie, R.Meder,M. Croucher,and R. J. Burton, J . Wood Chern. Technol., 14: 539-561(1994). H. Hafizoglu and B. Holmbom, Ho/z$orsch., 41: 73-77 (1987). N. Dunlop-Jones, J. Huan, and L. H. Allen, J. Pulp Paper Sci., 1 7 560-566 (1991). Y. Yazaki and P. J. Collins, Holz Roh Werkst., 52: 307-3 10 (1994). M. N. Nurulhuda. L. T. Chew, N. M. Y. Mohd, and R. M. A. Abdul, Holz Roh Werkst., 48:

38 1-383 (1990). 23. 24. 25. 26.

M. L. Laver and H. H.-L. Fang, Wood Fiber Sci., 18: 553-564 (1986). M. L. Laver and H. L. Fang, J . Agric. Food Chern., 3 7 114- 1 1 6 ( I 989). Q. L. Pu, L. K. Pannell, and X.-D. Ji, Planta Med., 56: 129- 130 ( 1 990). Z . Madar. H. H. Gottlieb, M. Cojocaru, J. Riov, Z. Solel, and A. Sztejnberg, Phytocheniwy.

56: 129-130 (1990). 27. Y. Mimaki, A. Kameyama, Y. Sashida, Y. Miyata.andA. Fujii, Chern.Pharnz. Bull, 43: 893-895 (1995). 28. M. C. Wani, H. L. Taylor, M. E. Wall, P. Coggon, and A. T. McPhail, J. Am. Chenz. Soc., Y3: 2325-2327 (1971). 29. S. B. Horwitz, Nurure, 367: 593-594 (1994). 30. K. C. Nicolau et al., Nuture, 367: 630-634 (1994). 31. J. C. Dennis, Correa, and A. E. Greene, J. Org. Chern., 56: 6939 (199 I ) . 32. A. Stierle, G. Strobel, and D. Stierle, Science, 260: 214-2 16 (1993). 33. S. Tachibana, E. Watanabe, Y. Itoh, and T.Oki, Mokuzai Gukknishi, 40: 1254-1258 (1994),

and literature cited therein.

34. N. Vidensek, J . Naturul Prod., 53: 1609- 16 I O ( 1990). 35. G. Chengwei,C.Zhangyu,andC. Yushu, Abstracts2nd Int. S y n p . Chern. Utili:. Tree Extructives, pp. 188- 189 ( 1995). 36. 1. Ara, B. S. Siddiqui, S. Faizi, and S. Siddiqui, J. Chenz. Soc.. Prrkin Truns. 1, 343-345 ( 1989). 37. S. Yaga, K. Kinjo, H. Hayashi, N. Matsuo, F. Abe, and T. Yamauchi, Phytochemistr)!, 32: 141-143 (1993).

Chemistry of Bark

269

38. T. Ohira and M. Yatagai, Mokuzai Gakkaishi, 40: 75 1-757 ( l 994). 39. S. Ohara, M. Yatagai, and Y. Hayashi, Mokuzai Gakkaishi, 32: 226-273 (1986). 40. Y. Hua,M. D. Bentley, B. J. Cole, Y. K. D. Murra,and A. R.Alford, J. Wood Chetn. Technol., 11: 503-516 (1991). 41. H. Fuchino, T. Satoh,and N. Tanaka, Chern. Pharrn. Bull, 43: 1937-1942 (1995). 42. S. Ohara and S. Hishiyama, Mokuzrri Gakkaishi, 40: 444-45 1 ( I 994). 43. Z. Abe, Jpn. Tappi J . (in Japanese), 47: 392-397 (1993). 44. H. P. S. Makkar, B. Singh, S. K. Vats, and R. P. Sood, BioresourceTechnol., 45: 69-71 ( I 993). 45. S. R. Devi and M. N. V. Prasad, BioresourceTechnol., 36: 189-192 (1991). 46. E. Conde, E. Cadahia, M. C. Garcia-Vallejo, and F. Tomas Barberan, Wood Fiber Sci., 2 7 379-385 ( 1995). 47. D. D. Khac, S. Tran-Van, A. M. Campos, J.-Y. Lallemand, and M. Fetizon, Phytochetnistry, 29: 251-256 (1989). 48. H. Shimomura, Y. Sashida, and T. Adachi, Phytochemistry, 26: 249-251 (1987). 49. E. Smite, L.N. Lundgren, and R. Anderson, Phytochemistry, 32: 365-369 (1993). 50. M. Nagi, M.Kubo, K. Takahashi, M.Fujita, and T. Inoue, Chetn. Pkann. Bull., 31: 1923 ( 1983). 51.

52.

53. 54. 55.

X. Fang, J. E. Anderson,C.Chang, P. E. Fanwick,and J. L.McLaughin, J. Chem. Soc., PerkinTrans. I , 1655-1661 (1990). T. L. Lowary and G. N. Richards, J. Wood Chem. Technol., 9: 333-339 (1989). D. C. Ayres and J. D. Loike, Lignans Chemistry: Biological and Clinical Properties, Cambridge University Press, Cambridge, U.K. (1990). 0. R. Gottlieb and M. Yoshida, in Naturcrl Products of Woody Plants, I (J. W. Rowe, ed.), Springer-Verlag, New York, pp. 439-5 11 (1989). S. Nishibe, H. Kinoshita, H. Takeda, andG.Okano, Chetn. Pllrrrm. Bull, 38: 1763-1765

( 1990). 56. E. Hayashi, Y. Kojima, andM.Terazawa, MokuzaiGakkuishi, 41: 925-931 (1995). 57. E. Hayashi, Y. Kojima, and M. Terazawa, MokuzaiGakkaishi, 41: 1029-1034 (1995). 58. Y.-G. Kim, S. Ozawa. Y. Sano, and T. Sasaya, Res. Bull. HokkaidoUniv.Forest, 53: 2943 ( 1996). 59. M. Ishige, M . Motidome, M. Yshida,and 0. R. Gottlieb, Phytochemistry, 30: 4121-4128 (1991). 60. L. N. Lundgren, H. Pan. 0. Theander, H. Eriksson, U. Johansson,andM.Sveningssong, Cm. J. Forest Res., 25: 1097-1 102 (1995). 61. D. Sun, Z. Zhao, H. Wong, and L. Y. Foo, Phytochetnistp, 27: 579-583 (1988). 62. Y. Ito, Y. Hayashi, and A. Kato, MokuzaiGakkaishi, 41: 694-698 (1995). 63. K. Ishimaru, G. Nonaka, and I. Nishioka, Phytockemistry,26: 1147-1152 (1987). 64. R. Saijo, G. Nonaka, and I. Nishioka, Phytochetnistty, 28: 2443-2446(1989). 65. M. Higaki, N. Kamimura, T. Suzuki, and T. Iijima, Mokuzai Gakkaishi, 36: 738-746 (1990). 66. R. W. Hemingway, in Nuturd Products of WoodyPlants, I ( J . W. Rowe, ed.), SpringerVerlag, New York. pp. 57 1-65 1 (1989). 67. J. B. Harbone (ed.), The Flovonoids, Advances i n Resenrch Since 1986, Chapman & Hall, London ( 1994). 68. K. Ishimaru, G. Nonaka, and I. Nishioka, Phytochetnistry, 26: I 167- I I70 (1987). 69. Y. S. Bae, J. F. W. Burger, J. P. Steynberg, D. Ferreira, and R. W. Hemingway, Phytochenz;.SI?, 35: 473-478 (1994). 70. H. Pan and L. N. Lundgren, Phytochemistry, 36: 79-83 ( 1 994). 71. L.Y. Fooand J.Karchesy, Phyfochernistty, 28: 1237-1240 (1989). 72. J. G. Peralta, A. Zarzuelo, R. Busson, C. Cobbaert, and P. Witte, Planta Med., 58: 174-175 ( 1992). 73. E. MaIan and D. H. Piennar, Phytochemistry, 26: 2049-205 1 ( 1 987). 74. J. M. van derNat, W. G. van derSluis, L. Hart,H. V. Dijk,R. P. Labadie, K. T. D. der Silva, and P. L. Labadie, Plnnta Med., 5 7 65-68 (1991).

270

75 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87.

Sakai L. J. Porter,in The Flavonoids,Advances in Research Since 1986 (J. B. Harbone,ed.), Chapman & Hall, London, pp. 23-55 (1994). R. W. Hemingway, in Chemistry and Sign$cance of Condensed Tannins (R. W. Hemingway and J . J . Karchesy, eds.), Plenum Press, New York, pp. 83-107 (1989). G. M. Polya and L. Y. Foo, Phytochemistry, 35: 1399-1405 (1994). F. Geiss, M. Heinrich, D. Hunkler, and H. Rimpler, Phytochemistry, 39: 635-643 (1995). J. P. De Mello, F. Petereit, and F. Nahrstedt, Phytochemistry, 41: 807-813 (1996). S. Ohara, K. Suzuki, and T. Ohira, Mokuzai Gakkaishi, 40: 1363-1374 (1994). E. Haslam, Plant polyphenols, Cambridge University Press, Cambridge, U.K. (1989). E. Haslam, in Natural Products qf Woody Plants, I (J. W. Rowe, ed.), Springer-Verlag, New York, pp. 399-438 (1989). G. Nonaka, Pure Appl. Chem., 61: 357-360 (1989). T. Okuda, T. Yoshida, and T. Hatano, Prog. Chem. Natural Prod., 66: 1-1 17 (1995). K. Ishimaru, G. Nonaka, and I. Nishioka, Chenl. Phann. Bull., 35: 602-610 (1987). K. Ishimaru, M. Ishimatsu, G. Nonaka, K. Mihashi, Y. Iwase, and I. Nishioka, Chent. Pharnz. Bull., 36: 3319-3327 (1988). T.-C. Lin, T. Tanaka, G. Nonaka, 1. Nishioka, and T. J. Young, Chem. Pharm.Bull., 39:

1144-1147 (1991). 88. C. Haberland and H. Kolodziej, Plantu Med., 60: 464-466 (1994). 89. G. Nonaka, K. Ishimaru, K. Mihashi, Y. Iwase, M. Ageta, and I. Nishioka, Chem. Pharm. Bull., 36: 857-869 (1988). 90. K. Ishimaru, M. Ishimatsu, G. Nonaka, K. Mihashi, Y. Iwase, and I. Nishioka, Chem. Pharm. Bull., 36: 3312-3318 (1988). 91. M. Konig, E. Scholz, R. Hartmann, W. Lehmann, and H. Rimpler, J. Natural Prod., 5 7 1411-1415 (1994). 92. G. Nonaka, M. Ishimatsu, M. Ageta, andI. Nishioka, Chem. Pharm. Bull., 37: 50-53 (1989). 93. R. Saijo, G. Nonaka, and I. Nishioka, Phytochemistry, 29: 267-270 (1990). 94. R. Saijo, G. Nonaka, and I. Nishioka, Chern. Pharm. Bull., 3 7 2063-2070 (1989). 95. G. Nonaka, K. Ishimaru, R. Azuma, M. Ishimatsu,and I. Nishioka, Chem.Pharm. Bull., 37: 207 1-2077 (1989). 96. R. Saijo, G. Nonaka, I. Nishioka, 1.-S. Chen, and T.-H. Hwang, Chem. Pharm. Bull., 37: 2940-2947 (1 989). 97. T. Tanaka, N. Ishida, M. Ishimatsu, G. Nonaka, and 1. Nishioka, Chem. Pharm. Bull., 40: 2092-2098 ( 1 992). 98. T. Tanaka, S. Kirihara, G. Nonaka,and I. Nishioka, Chem. Phann. Bull., 41: 1708-1716 (1993). 99. W. Mayer, H.Seits, J. C.Jochims, K. Schauerte,and G. Schilling, Justus Liebigs Ann. Chem., 751: 60 (1971). 100. G.Nonaka, T. Sakai, T. Tanaka, K. Mihasi, and I. Nishioka, Chem. Pharm. Bull., 38: 215 12156 (1990). 101. L. J. Porter and R. W. Hemingway, in Natural Products of Woody Plants, I1 (J. W. Rowe, ed.), Springer-Verlag, New York, pp. 988-1027 (1989). 102. R. W. Hemingwayand J. J. Karchesy(eds.), ChemistryandSigniJicance of Condensed Tannins, Plenum Press, New York ( 1 989). 103. R. W. Hemingway and P. E. Laks (eds.), Plant Polyphenols, Synthesis, Property and Sigrtijcance, Plenum Press, New York (1994). 104. Y. Yazaki and T. Aung, Holzjorsch., 43: 281-282 (1989). 105. L. Y. Foo and J. J. Karchesy, Phytochemistry, 28: 3185-3190 (1989). 106. G. Zheng, Y. Lin, and Y. Yazaki, Hol$orsch., 42: 407-408 (1988). 107. E. Pallenbacch, E. Scholz, M, Koenig, and H. Rimpler, PIanta Med., 59: 264-268 (1993). 108. S. Ohara and R. W. Hemingway, Holz$orsch., 43: 149-154 (1989). 109. S. Ohara and H. Hujimori, Mokuzai Gakkaishi, 42: 618-623 (1996). 1 IO. F. Hsu, G. Nonaka, and I. Nishioka, Phytochemistry, 24: 2089-2092 (1985).

Chemistry of Bark

271

Y. Cai, F. J. Evans, M. F. Roberts, J. D. Phillipson, M. H. Zenk, and Y.Y. Gleba, Phytochemistry, 30: 2033-2040 (1991). 112. M. H&, M. Heinrich, and H. Rimpler, Phytochemistry, 42: 109-1 19 (1996). 113. S. Achmadi, G. Syahbirin, E. T. Choong, and R. W. Hemingway, Phytochemistry, 35: 217-

111.

219 (1994). 114. H. Kolodziej, Phytochemistry, 28: 3487-3492 (1989). J, J. Karchesy, L. Y. Foo, R. W. Hemingway, D. F. Barofsky, and E. Barofsky, Wood Fiber Sci., 21: 155-162 (1989). 3 7 1748-1750 116. R. Isobe, T. Tanaka, G. Nonaka,and I. Nishioka, Chem.Pharm.Bull.,

115.

( 1 989). 117. D.F. Brofsky, in Chemistry and Significance of Condensed Tannins (R.W. Hemingway and J. J. Karchesy, eds.), Plenum Press, New York, pp. 175-195 (1989). 118. D. Ferreira and E. V. Brandt, in Chemistry and Significance of Condensed Tannins (R. W. Hemingway and J. J. Karchesy, eds.), Plenum Press, New York, pp. 153-174 (1989). 119. D. Thompson and A. Pizzi, J. Appl. Polymer Sci., 55: 107-1 12 (1995). 120. S. Ohmura, Medicinal Plants f o r Pharmacognosy in China and Japan (in Japanese), Hirokawa Shoten, Tokyo (1978). 121. T. Kariyone, Japanese and Chinese Pharmacognos.y (in Japanese), Hirokawa Shoten, Tokyo (1971). 122. E. Malan, E. Swinny, D. Ferreira, and P. Stemberg, Phytochemistry, 41: 1209-121 3 (1996). J. Karchesy,in Plnnt Polyphenols, 123. D. R.Forlines, T. Tavenner, J. C. S. Malan,andJ.

124. 125. 126. 126b. 127.

Synthesis. Property and Significance (R. W. Hemingway and P. E. Laks, eds.), Plenum Press, New York, pp. 767-782 ( 1 994). K. Mura, Y. Iitoi, and W. Tanimura, Hakko Kougaku (in Japanese), 65: 399-404 (1987). Z. S. Okoye and G. E. Neal, Food Chem. Toxicol., 26: 679-689 (1988). M. Samejima and T. Yoshimoto, Mokuzni Gukkaishi, 30: 41 3-416 (1984). H. Kai, T. Ueda. and M. Sakaguchi, Soil B i d . Biochern., 22: 983-986 (1990). M. Mori, M. Aoyama, S. Doi. A. Kanetoshi, and A. Hayashi, Holz Roh Werkst.. 53: 81-82

( 1995). 128. J.-J. Ge and K. Sakai, Mokuzni Gakkuishi, 39: 801-806 (1993). 129. J.-J. Ge and K. Sakai, Mokuzai Gakkuishi, 42: 87-94 (1996). 130. Y. Nakashima, J.-J. Ge, and K. Sakai, Mokuzai Gakkaishi, 42: 1105-1112 (1996). 131. P. E. Laks. P. A. McKaig, and R. W. Hemingway, Hol;for.sch., 42: 299-306 (1988). 132. J. Zhang, K. Takahashi, Y. Kono. Y. Suzuki, et al., J . Pesticide Sci., 15: 585-591 (1990). I32b. A. Scalbert, Phytochernistty, 30: 3875-3883 (1991). 133. N. Yamashita, H. Itoh, K. Skata, A. Yagi, H. h a , and K. Ina, Agric. Biol. Chern., 53: 13831385 ( 1 989). 134. J. H. M. Temmink.J.A.Field,J.C. Van Haastrecht, and R. C. M. Merkelbacha. Wnter RES..23: 34 1-344 ( 1 989).

135. J . A. Field. M. J . H. Lcyendeckcrs, R. S. Alvarez. G. Lettinga, and L. H. A. Habets, Wtrter Sci. Technol., 20: 2 19-240 ( 1988). 136. J.A. Field, G. Lettinga. andL. H. A. Habets. J. Chern. Techrrol. Biorechnol.. 49: 15-33 ( 1990). 137. J. A. Field, R. Sierra-Alvarez, G. Lettinga, and L. H. A. Habets, J . Chern. Techno/.,49: 35S3 ( 1990).

138. M. Savolainen and K. Jokinen. in Ligrlocellltlosics, Science. Technolop. De\dqnnent ~ l n d Use (J. F. Kennedy, G. 0. Phillips. and P. A. Williams, eds.). EllisHorwood. New York. pp. 393-404 (1992). 139. J.Field, M. J. H. Lcycndeckers. R. Sierra-Alvarez. L. H. A. Habets.and A. G. Lettinga. Biotechnol. Bioeng.. 37: 247-255 ( 199l ) . 140. E. P. Swan. in Noturd Products o f Woocly Pltrnts, I / (J. W. Rowe, ed.), Springer-Verlag, New York. pp. 931-952 (1989). 141. A. Pizzi. E. P. von Leyser. J. Valenzuela. and J. G. Clark. Hol;fif~rsc/~.. 47: 168- l74 (1993).

Sakai

272

142. L. Zhao, B. Cao, F. Wang, and Y. Yazaki, Holz Roh Werkst., 52: 1 13-1 18 ( 1994). 143. L. Zhao, B. Cao, F. Wang, and Y. Yazaki, Holz Roh Werkst., 53: 117-122 (1995). l#. L. Zhao, B. Cao, F. Wang, and Y. Yazaki, Holz Roh Werkst., 54: 89-91 (1996). 145. M. A. E. Santana, M. G. D. Baumann, and A. H. Conner, J . Wood Chem. Technol., 16: 1-

19 (1996). 146. 147. 148. 149. 150.

E. A. Taiwo and R. A. Ogunbodede, Wood Sci. Technol., 29: 103-108 (1995). M. Fechtal and B. Riedl, Holzforsch., 4 7 349-357 (1993). B. Dix and R. Marutzky, Holz Rolz Werkst., 45: 457-463 (1987). B. Dix and R. Marutzky, Holz Roh Werkst., 46: 19-25 (1988). R. Takano, Research Report of Biomass Conversion Program, No. 29:67-76 (1991) (in Japanese). 151. A. Grigoriou, E. Voulgaridis, and C. Passialis, Holdorsch. Holzvenuert., 39: 9- I 1 (1987). 152. A. Grigoriou, Holz Roh Werkst., 48: 377-380 (1990). 153. G. Vazquez, G. Antorrena, J. Parajo, and J. L. Francisco, Holz Roh Werkst., 4 7 491-494 ( 1989). 154. G. Vazquez, G. Antorrena, J. L. Francisco, and J. Gonzalez, HolzRohWerkst., 50: 253256 ( 1992). 155. G. Vazquez, G . Antorrena, J . L. Francisco, M. C. Arias, and J . Gonzalez, H012 Roh Werkst., 51: 221-224 (1993). 156. G. Vazquez, G. Antorrena, J. Gonzalez, and J. C. Alvarez, HolzRohWerkst., 54: 93-97 ( 1996). 157. E. von Leyser and A. Pizzi, Holz Roh Werkst., 49: 25-29 (1990). 158. V. J. Sealy-Fisher and A. Pizzi, Holz Roh Werkst., 50: 212-220 (1992). 159. A. Pizzi and A. Stephanou, J. Appl. Polymer Sei., 51: 2109-2124 (1994). 160. A. Pizzi and A. Stephanou, J. Appl. Pol.ymer Sei., 51: 2125-2130 (1994). 161. Y. Yazaki and P. J. Collins, Holz Roh Werkst., 52: 185- 190 (1994). 162. Y. Yazaki and P. J. Collins, Holz Roh Werkst., 52: 241-243 ( 1994). 163. S. Kim and D. E. Mainwaring, J. Appl. Polymer Sci., 56: 905-913 ( 1995). 164. S. Kim and D. E. Mainwaring, J. Appl. Polymer Sei., 56: 915-924 (1995). 165. S. Kim and D. E. Mainwaring, Holdorsch., 50:42-48 (1996). 166. R. E. Kreibich and R. W. Herningway, Forest Prod. J., 37(2): 43-46 (1987). 167. R. E. Kreibich, in Chemistry and Significance of Condensed Tannins (R.W. Herningway and J. J. Karchesy, eds.), Plenum Press, New York, pp. 4.57-478 (1989). 168. R. E. Kreibich, R. W. Hemingway, and W. T. Nearn, ForestProd. J . , 43(7/8):45-48 (1993). 169. R. W. Hemingway, in Chemistry und Sign$cance .f Conder~sedTannins (R.W. Herningway and J. J. Karchesy, eds.), Plenum Press, New York, pp. 265-283 (1989). 170. G. W. McGraw, P. E. Laks, and R. W. Hemingway, J . WoodChem. Technol., 8: 91-109 (1988). 171 K. H. Chung and G. P. Hamed, in Chemistrv and Signifcance of Condetmed Tcnnirzs (R. W. Hemingway and J. J. Karchesy, eds.), Plenum Press, New York, pp. 479-502 (1989). 172. A. Pizzi, N. Meikleham, and A. Stephanou, J. Appl. Polymer Sci., 55: 929-933 ( I 995). 173. B. Prasetya and E. Roffael, Holz Roh Werkst., 49: 341-344 (1991). 174. H. Tanabe and Y. Kobayashi, Agric. Biol. Chem.. 49: 3595 (1985). 175. H. Tanabe and Y. Kobayashi, Agric. Biol. Chem., 51: 2845-2846 (1987). 176. T. Itoh, Modern Foresh (in Japanese), no. 7 48 (1993). 177. G. Vazquez, G. Antorrena, and J. Gonzalez, Wood Sci. Technol, 28: 403-408 (1994).

178. G. Vazquez, J. Parajo, G. Antorrena, P. Thonart, and M. Paquot, WoodSci.Techrlol.,

21:

167-178 (1987). 179. D. V. Penas, M. A. L. Yusty, and J . C. P. Linares, J. Chem. Technol. Biotechrwl., 54: 6374 (1992). 180.

M. Kuwahara, T. Sawada, Y. Asada, Y. Nakamura, M . Uchikoshi, and M . Komatsu, Mokuzcli Gakkaishi, 34: 75-81 (1988).

Chemistry of Bark

273

181. R . Torget, M. E. Himmel, and K . Grohmann, Bioresour. Technol., 35: 239-246 (1991). 182. G. Vazquez, G. Antorrena, J. C. Parajo, and X. L. Francisco, Wood Sci. Technol., 22: 219225 (1988). 183. E. Walch, A. Zemann, F. Schinner, G. Bonn, and 0. Bobleter, Bioresou,: Technol., 3 9 173I77 ( 1992). 184. T. Sakuragawa and T. Miyamoto, Mokuzrti Kogyo, 51: 266-269 (1996) (in Japanese). 18.5. H. Yamaguchi, M. Higuchi. and i. Sakata, J. Appl. Polywzrr Sci., 45: 1455- 1462 ( I 992). 186. H. Yamaguchi, R. Higasida, M. Higuchi, and I. Sakata, J . Appl. Polymer Sci., 45: 14631472 ( 1992). 187. H. Yamaguchi, Y. Iura, M. Higuchi, and I. Sakata, Moklczai Gakkaishi, 37: 815-820 (1991). 188. J.-J. Ge and K. Sakai, Mokuzrri Grrkknishi, 42: 417-426 (1996).

This Page Intentionally Left Blank

Chemical Characterization of Wood and Its Components Jaime Baeza and Juanita Freer Universidad de Concepcidn, Concepcidn, Chile

1.

INTRODUCTION

Wood analysis comprises the determination of the wood components as well as the isolation, purification, and characterization of the woodconstituents. Wood is chemically heterogeneous and its components can be divided into two groups: structural components of high molecular weight (cellulose, polyoses or hemicelluloses, and lignin), which are the major cell wall components; and nonstructural components of low molecular weight (extractives and inorganic compounds). The macromolecular components are not uniformly distributed in wood cells, and their concentration changes from one morphological region to another. Therefore, knowledge about the distribution of chemical components in the cell walls is of great importance to understanding the properties of wood. The chemical composition of wood varies from species to species. In general, hardwoods contain more hemicellulose than softwoods but less lignin. Figure 1 shows the typical composition of hardwoods and softwoods [ I ] . There are different types ofwood analysis. One may consider only the main cell wall components, holocellulose (cellulose and polyoses), lignin, extractives, and ash; on the other hand, a very detailed analysis may include functional groups (e.g., acetyl groups), individual units of the different components (e.g., sugar pattern), different types and frequency of linkages (e.g., p-0-4 linkage in lignin), etc. Usually, wood is analyzed by the separation of the different components, but there are serious difficulties in achieving selective isolations. The separation is never complete and leads to structural changes,predominantly in the lignin. An array of classical, wet chemicalproceduresandagrowingnumber of instrumentalmethodsare available for analysis of wood. The wet chemical methods permit the acquisition of data on the gross composition of wood.and they require the separation ofwood into macroscopicwood components (e.g.. lignin, holoccllulose, etc.). This is the reason it is always necessary to report the isolation techniques used. On the otherhand, the instrumentalmethods are conducive to higher specificity and convenience of wood analysis. Initially, chemical specificity was achieved in a macroscopic scale; for example, by chromatographic methods the separation and determination of the individual sugars can be obtained after hydrolysis of a wood sample.More recently, techniquessuch a s ultraviolet microscopy, electron mi275

276

Baeza and Freer

50 40 m

gc

30-

g

20-

W

a

10 -

0

-L

cellulose

lignin

polyoses

extractives"

FIGURE 1 Typical composition of softwoods and hardwoods. common beech (hardwood). *CH,Cl, followed by C2H50H.

D,

Norway spruce (softwood); m,

croscopy coupled with X-ray analysis, and infrared spectrometry have permitted descriptions of the distribution of chemical constituents in wood and fiber walls. The methods of wood analysis are more or less standardized. Detailed descriptions of the analysis of wood are giving in the specialist literature[2-81. The CPPA (Technical Section, Canadian Pulp and Paper Association, Montreal,PQ, Canada), TAPPI (Technical Association of the Pulp and Paper Industry, Atlanta, GA), ASTh4 (American Society of Testing and Materials, Philadelphia, PA) have issued new or revised test methods for the analysis of pulp and paper materials. There are excellentreviews covering analytical techniques for wood and its components [9-151. The purpose of this chapter is to inform the reader about some of the methods available for the chemical analysis of wood. The analysis of the constituents of wood, according to the scheme shown in Fig. 2, will be described.

II. SUMMATIVE ANALYSIS The objective of the summative analysis is toaccount for all of a sample. The summative analysis of wood is based on the isolation and identification of certain groups of wood components and therefore does not deal with the determination of chemical uniformity of substances. These groups are mainly cellulose, hemicelluloses, lignin, and extractives, and often other designations are added, e.g., holocellulose, a-cellulose, or sometimes terms originated from the method used (such as kraft lignin, Cross and Beavan cellulose). In evaluating the results of the summative analysis, it is important to take into consideration the methods used. To obtain a complete mass balance of a wood sample, no constituents may be overlooked or repeated. Nothing should be determined by difference. Summative analysis will typically account for most of the wood components. Examples of summative analysis of wood composition have been reported elsewhere [16-211. Summations may be taken in several ways, and some examples of different types of wood analyses are given in Table 1. Valuesof 98-101 are generallyacceptable,butfrequentlyvaluesdeficientor in excess by about 10% are obtained [ 14,16,19,20]. Possible reasons for failure to achieve a complete mass balance have been reviewed by Browning [20]. Generally, the summative analysis is corrected via normalization, giving the same factor of error to all the components, independent of the amount in which the different components are present and in the analytical technique used. The possibility of undetected or unknown compounds in

277

Chemical Characterizationof Wood

EXTRACTION

I

I

I

EXTRACTIVES

I SOLUBLE IN ORGANIC SOLMNTS

I EXTRACTIVEFREE WOOD

I

I

DELlGNlFlCATlON

SOLUBLE IN WATER

FIGURE 2 The separation of wood components.

sample wood isignored.Kaarand Brink [21]developed a summativeanalysisscheme (bomb/HPLCsummativeanalysismethod)thatprovides a completeaccounting of the starting material without normalization. This method utilizes sealed vessels to allow the retention of the volatiles during the high-temperature stage of hydrolysis, with the hydrolyzate analyzed by HPLC. There were no significant unidentified peaks in any of the HPLCchromatograms.Therangeof mass balancedeterminations for thegroup of 10 wood specimens on unextracted and extractive-free bases was 98.43-99.63% and 98.2199.60%, respectively. The factors that could be responsible for the discrepancy from 100%

TABLE 1 SummativeAnalysis;Examples of Wood Analysis

A Extractives Lignin Holocellulose Ash Xylan

B

Extractives Lignin 0-Cellulose Hemicelluloses Galactan groups Acetyl Ash

C Extractives Lignin Glucan Mannan Arabinan Uronic anhydride Ash

D

Extractives Lignin Glucan Mannan Galactan Xylan Arabinan 4-0-Me glucuronic ac. Glucuronic ac. Galacturonic ac. Acetyl groups Protein Ash

Freer 278

and

Baeza

in the mass balance were extensively discussed. These factors include: the exact contribution of galacturonic acid in the wood; the methanol formed from the demethoxylation of polysaccharides and lignin; the loss of water in the lignin condensation/dehydration relations; the wood components included in the polysaccharides and/or lignin moieties that exist in such small quantities that they are not detected by the analytical methods utilized; and residual water that remains bonded in the wood after “oven drying” which is released in the hydrolytic stage and becomes part of the hydrolytic solution. Clearly, the contribution of any of these factors is small. The complex nature of wood seriously complicates the quantification of these contributions. The improvement of the methods and techniques of analysis could result in a mass balance close to 100%.

111.

SAMPLING AND PREPARATION OF SAMPLES FOR ANALYSIS

The kind of sampling and sample preparation depends on many factors and on the aim of the analysis. Thus, the magnitude of the sampling needed for general characterization of aspecies is quite different than for the evaluation of trees in a specified stand. It is important to ensure that representative samples are collected which are free from outside contamination, and properly preserved. No analysis is better than the sample on which is based. However, for comparison of techniques and methods the only requirement is that the sample be uniform. Astandardizedsamplingprocedure is given in TAPPIStandardT-257cm-85 [2]. The procedure given is appropriate for wood in all forms, i.e., logs, chips, or sawdust. A probability sampling plan and an economic or engineered sampling plan are described. A detailed discussion of sampling and preparation of samples is given by Browning [22]. Wood for chemical analysis, after air drying, must be milled to achieve complete penetration by reagents and to ensure uniform reactions. Heating, preparation of very fine and dusty material, and regrinding coarse material must be avoided. Samples are screened and normally material passed through a 0.40-mm (40-mesh) sieve and retained on a 60mesh sieve. The selected fraction should represent, if not the entire amount of material, at least 90-95% of the original sample. Theextractiveshould be removedbeforeany chemicalanalysis,exceptwhere the extraction process and subsequentwashingcould interfere with the analysis.Aprocedureforfurtherpreparation of wood for chemical analysis that has been sampled in accordance with TAPPI 257 is provided in TAPPI Test Method T264 0111-88 121. Neutral solvents, ethanol and benzene, are employed to obtain extractive-free wood, removing material which is not part of the wood substance or which may interfere with subsequent analysis. Moisture determination is included. Related methods are ASTM D1 105 (ANS),“Preparation of Extractive-FreeWood” 131; and CPPA G.31P. “Preparation of Wood for Chemical Analysis” [4]. It is dangerous to include benzene in the solvent mixture to extract wood. due to its carcinogenic properties for which it has been long banned. Accordingly, any contact with the skin or inhalation of benzene vapor must be avoided. Extreme safety precautions must be taken in carrying out the above procedure. Gloves, good ventilation, and a chemical fume hood must be used. A mixture of ethanolandtoluenewasfound to remove the samematerials fro111 wood as ethanol-benzene 1231. However, the mixture of ethanol and toluene does not boil and reflux at a constant temperature and rate. Wood samples collected for later analysis must remain moist and cold-stored. Samples should not be oven-dried to avoid changes in reactivity and lost of volatiles. After

279

Chemical Characterizationof Wood

Extraction Klason Lignin

~ioacido~ysis

Extractive-free wood

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

Lignin

IO mg

(160

r

Delignification Lignin in holocellulose

Klason

Hobcellulose (80 m@

2

.m

TFA hydro]ysis

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

................I ..............

+

20 mg

mmm

m

mm.

I.

Neutralsugar ratio r................:

m

:

i

Dissolution by alkali a-cellulose (20

Nitration ................................... 'I IO mg

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

DP of cellulose

i

Calculation (holocellulose - a-celulose) W

Hemicellulose

FIGURE 3 Experimentalsmall-scalemethod.(From Ref. 24.)

airdrying, most chemicalproperties of wood do not changeunderadequatestorage conditions. Recently, a small-scale method for the determination of wood components such as extractives, lignin, a-cellulose, and hemicellulose has been published [24]. This method is also used in the sample preparation for structural analysis of each component (Fig. 3 ) .

IV.

DETERMINATION OF WATER CONTENT

Water is a natural constituent of all parts of a living tree. In green wood, water is commonly about 5 0 % of the total weight. When the tree dies or a log is processed into lumber, chips, etc., the wood loses some of its moisture to the surrounding atmosphere. However, some water will remain within the structure of the cell even after wood has been manufactured into lumber, particle, veneer, or fiber product. The amount of residual water depends on the cxtent of drying and the environmental conditions. The wood-water system is very important i n many tields of wood technology. The physical and mechanical properties, resistance to biological deterioration, and dimensional stability of the products are affected by the amount of water present. Chemicalanalysis of wood is altnost always performed on air-dried samples, but results are reported on a moisture-free basis. A moisture determination must therefore be r ~ for ~ nalmost every sample submitted for analysis. The amount of water in wood is expressed in two ways:

Baeza and Freer

280

1.

The moisture content (MC) is expressed as a percent of with water). Thus: %MC =

2.

the total weight (wood

weight with water - OD weight x 100 weight with water

The wet weight base is generally used in the pulp and paper industry and when wood is used as fuel. Themoisturecontent (MC*) is definedas the weightofwaterexpressedas percentage of the moisture-free or oven-dry (OD) weight of wood. Thus: %MC* =

weight with water - OD weight x 100 OD weight

Because the denominator is the dry weight, the moisture content calculated in this way canbeover 100%. Thismethodof calculating moisturecontent is generally accepted as standard for wood-based materials such as lumber, plywood, particleboard, and fiberboard. Some of the methods used for moisture determination are described below. Oven-Dry Method. One of the mostcommonmethods of determining MC is to weigh about 2 g of the wet sample and oven-dry it at 105 2 3°C until the weight does not change by more than 0.002 g following a l-h heating period. The details of this ovendry method are described in TAPPI Test MethodT264 om-88 [2]. The major disadvantages of the oven-dry method are that it is a destructive method, it is time-consuming, and some volatile components other than water can be driven off during drying. Electrical Moisture Detectors Method. TheMC can be determined by use of electrical moisture detectors, whichhave the advantage of being relatively simpleand direct. There are a variety of electrical meters. The resistance moisture meter, used for lumber, indicates the moisture based on the linear relationship between the logarithm of electrical resistance and the moisturecontent of woodfrom 7% to the fiber saturation points [25].Some electric meters are based on the effect that water behaves as a capacitor when placed in a high-frequency field [26]. However, the dielectric methods have been described as unsatisfactory for solids such as moist wood. The most common moisture balance consists essentially of an infrared lamp mounted above the pan of a top-loading electronic balance. The energy input and drying times are controlled by the operator. Moisture Determination by Solvent Distillation. Water inwoodmay be determined by distillation carried out with solvents such as toluene, xylene, and other waterinsoluble solvents. Details of this method by toluene distillation are given in TAPPI Standard Method T208 om-S9 [2]. The material is boiled in toluene. The resultant water vapor is condensed, collected in a distilling reservoir consisting of a graduated trap (Stark and Dean or other type), and measured. This method is specially useful for determination of moisture content in wood because volatile substances such as turpentine and resins do not interfere, giving a better measure of true water content than that obtained from oven drying. Moisture Determination by Karl Fischer Method. Karl Fischer titration is one of the most sensitive techniques available for the actual measurement of very small quantities of water in different matrices. ASTM Standards E 203 and D 1348 [3] describe the procedure for determining the moisture content in cellulose and have been applied successfully for wood [27]. The Karl Fischer reagent is a solution of iodine, sulfur dioxide, and pyridine, usually in methanol as asolvent.Methanol may bereplaced by methyl cellosolve, dioxane, or acetic acid, but pyridine (Py) is essential. It reacts with water as shown below:

Chemical Characterization of Wood

H,O

281

+ I, + SO, + Py (excess) + 2Py.HI + Py*SO, Py.SO, + CH,OH + 2PyHSOaCH.j

The general procedure is as follows: the wood is ground and extracted with methanol to displace the water. The water in methanol is titrated with the Karl Fisherreagent. Titrations can be performed best by potentiometric endpoint determinations. The results obtained for this method are not affected by volatile substances other than water in the sample. Moisture Determination by Nuclear Magnetic Resonance. Special physical methods such as infrared, nuclear magnetic resonance, attenuation of p and y radiations and neutron moderation are also applicable to MC determination i n wood and pulp. Proton nuclear magnetic resonance ('H NMR) can provide more detailed and quantitative information on water in wood than any other method. Information on both macroscopic and microscopiclevels can beobtained. It has been demonstrated that the 'H NMR signals of solid wood and water are separable, so the MC of wood samples can be determined by NMR.The MC has been measured in wood samples utilizing wide-line NMR spectrometers 128,291. Wide-line NMR arethose in which the bandwidth of the source of the lines is large enough that the fine structure due to chemical environment is obscured, so only a single peak is associated with each species. These spectra are useful for quantitative analysis. The MC in wood relative to a known standard has been measured by pulsed NMR techniques 1301. These techniques provide a convenient approach to the study of the liquid adsorbed on or trapped in the solid matrix and the experimental observables include dynamics as well as structural information. The relationship between the water content of wood and spin-spin relaxation time (T,) has been studied [31-331. Menon et al. [34,35] determined the MC by NMR in samples of western red cedar and Douglas tir. The values of MC determined by NMR were related to those values determined by the oven-dried method. Furthermore, on the basis of T2 relaxation times, the water could be separated into signals for the bound water i n the cell walls and that in the cell lumens. Thus, NMR images have been obtained for these different water reservoirs. Araujo et a l . [36]carried out studies of water in wood by using NMR relaxation techniques and NMR relaxation selective imaging techniques. The distribution of water in sapwood, heartwood, and juvenile wood as well as two rehydrated heartwood samples of white spruce were analyzed. Spectra of T, for white spruce show separate peaks corresponding to the water in different environments.Thearea under each peak corresponds to the amount of moisture i n a particular environment, and the T1 value indicates the nature of the environment.The shape of the Tz spectrum of water in lumen reflects thc radius distribution of water-tilled cell lumens in wood sample. The NMR techniques for characterization of water in wood are very useful in the investigation of a wood drying process.

V.

WOOD POLYSACCHARIDES

Wood contains sevcral polysaccharides. of which cellulose and hemicelluloses are the most abundant. The polysaccharides of wood are built up of a relatively limited number of sugar residues, mainly D-glucose, D-xylose, D-mannose, D-galactose, L-arabinose, 4-0-methylD-glucuronic acid, D-galacturonic acid, and glucuronic acid. Less common sugar units arc L-rhamnose, L-galactose. L-fucose. and 0-methylated neutral sugars.

Freer 282

and

Baeza

Cellulose, hemicelluloses, and lignin exist in wood as an interpenetrating system. It is generally difficult, therefore, to isolate pure components from wood. In spite of this. several methods of separation have been devised that can yield considerable information about the chemical composition of wood samples.

A.

Cellulose

Cellulose is the mostabundantorganic material on earth. It is the chief constituent of wood and cotton. Cellulose is a long-chain polymer of P-D-glucose in the pyranose form, linked together by 1,4-glycosidic bonds to form cellobioseunits that are the repeating units in the cellulose chain (Fig. 4). Though relatively pure cellulose is produced industrially as microcrystalline cellulose powder from bleached wood pulp by acid hydrolysis, its molecular weight is far less than that of native cellulose of wood, and it still contains trace amounts of xylose and mannose [W.

1. Sample Preparation and Determination of Cellulose There are various methods for isolation and determination of cellulose. The isolation methods are based on its insolubility in water, organic solvents, alkaline solutions and its relative resistance to oxidizing agents and susceptibility to hydrolysis by acids. Wood cellulose is isolated in laboratories usually from holocellulose by extraction of hemicellulose with strong alkali solutions. U . Cellulose .from Holocellulose Prepumtions. Holocellulose is defined as the water-insoluble carbohydrate portion including cellulose and hemicelluloses and none of the lignins. However, preparations of holocellulose always involve some loss of carbohydrates and retention of lignin. Holocellulose is obtained from wood by using different delignification methods, applying either strong oxidizing agents or acidic or basic solutions at high temperatures. It is usually prepared by removing the lignin of ground, extractive-free wood with chlorine gas, an acid solution of sodium chlorite, or peracetic acid. The chlorination method was first proposed by Ritter and Kurth [38]. Several modifications to the chlorine holocellulose method have been proposed to minimize the degradation of carbohydrates [39-421. The chlorination method is especially useful for the study of hardwoods.Small amounts of acetyl groups originally present in wood are lost in this procedure 1411, especially if alkaline interactions are employed to assist delignification. Details of the chlorination method are provided i n TAPPI, Useful Method 249 1431, and by Browning [44]. Samples of air-dried wood meal extracted successively with ethanolhenzene, ethanol, and hot water are filtered on a medium fritted glass crucible and washed withhot and cold water, and then chlorinated in the crucible by passing chlorine gas. The chlorinated lignin is removed using adequate solvents. The chlorite method was originally described by Jayme 1451. Modifcations to the chlorite procedure for holocellulose preparation have been made [18,45-53). The procedure of delignification using NaCIO,-acetic acid has been described by Browning (541. In the standard procedure the preextracted samples of wood are treated with acid solutions of sodium chlorite (pH 4) for 3-5 h at 70-80°C. The effective components of the delignifying solution are chlorine dioxide, chlorine. and chlorate. The peracetic acid method. first described by Poljak [SS],and further developed by other authors [56-59]. has been carried out using dilute peracetic acid.

Chemical Characterizationof Wood

I

l

0

0

0

0

I

0

& I

0

\

' C

U

283

Freer284

and

Baeza

Considerable controversy exists about the relative effectiveness of the chlorination, peroxyacetic acid, and chlorite techniques of cellulose preparations [ 18,50,59-61]. It is likely that all processes can yield roughly equivalent cellulose components if precautions are taken. However, it must be considered that the delignification depends on many factors (e.g., wood species, residual lignin, holocellulosecontent). The chlorination process is likely to be less destructive of the hemicellulose components, and the chlorite technique more convenient for preparing large quantities of hemicelluloses and is the only process for delignifying complete woody structures [ 6 I 1. The hemicelluloses may be separated from cellulose by extraction of holocelluloses with 24% aqueous sodium hydroxide containing 4% boric acid. As an example, one laboratory procedure that has been used for softwood fractionation is shown in Fig. 5 . b. Cellulose .from the Deterrnitzntiorl of Monosacchuritles Compositiorz. The cellulose content in a wood sample can be analyzed by a chromatographic technique to estimate the monosaccharide composition after hydrolysis. Cellulose is assumed to be equal to the total glucan (glucose content X 162/180) less the glucan associated with the glucomannans andgalactoglucomannans in the hemicelluloses[62-681.However,sugar ratios in the hemicellulosepolymers mayvary among species, individual trees, locations within the tree, and as a function of storage of the sample prior to analysis. The average ratio of mannose and glucose units is about 1 S-2.1 in most of the hardwoods, and that of mannose, glucose,andgalactose is about 3:l:l in softwoodmannans. The ratios found in various woods are summarized by Fengel and Wegener 1681. The hydrolysis procedures and the chromatographic techniques used in the analysis are discussed in the section on wood and pulp sugar analysis. c. Other Cellulose Prqxzrntiom. The cellulose of wood can be isolated directly as a more or less crude preparation by the chlorination method of Cross and Bevan [69].The isolation of cellulose preparation comprises alternate chlorination and extraction with a hot, aqueous sodium sulfite solution. The product obtained for the general procedure of Cross and Bevan consists largely of cellulose, but also contains considerable quantities of hemicelluloses. This preparation is designated Cross L I B ~e v m cellulose to avoid confusion. Details o f . the method are provided by Browning 1701. Other common cellulose determinations on pulp andwood are a-cellulose, @ - W lulose, and y-cellulose determinations. a-Cellulose is defined as the residue that is insoluble i n a strong sodium hydroxide solution when the treatment is carried out under specified conditions. The portion which is soluble in the alkaline medium but precipitable from the neutralized solution is called P-cellulose, and the portion which remains soluble corresponds to y-cellulose. In general, the a-cellulose indicates undegraded high-molecularweight cellulosc content; p- and y-celluloses represent degraded cellulose and hemicelluloses. The separation of these fractions is an empirical procedure and is a rapid quality control that is widely used to evaluate pulps. Before these determinations can be performed onwood or unbleached pulps, the material mustbe delignified. A revised TAPPI Test Method procedure was made in 1993, and is described in T-203 om 93 [2]. The general procedure comprises the consecutive extraction of the delignified sample with 17.5% and 9.45% sodium hydroxide solution at 25°C. The fraction consisting of the P- and y-celluloses, is determined volumetrically by oxidation with potassium dichromate (first oxidation). The a-ccllulose is calculated a s the undissolved fraction by the difference between the total delignified specimen (100%) andthe dissolved fraction determined in the tirst oxidation. The y-cellulose is determined by titration with potassiumdichromate of the solution alter precipitation of the @-cellulose (second oxidation). The @-cellulose is found

Chemical Characterization of Wood

285

I

Freer 286

and

Baeza

by the differencebetween the first andsecondoxidations.Asimilarprocedurefor the determination of these different fractions is also given in CPPA G.29 [4]. d. Residual Lignin and Hemicelluloses in Cellulose Preparations. Different methods have been implemented to determine the residual lignin and hemicelluloses in wood cellulose. Residual lignin and hemicelluloses in spruce and beech a-cellulose were analyzed by using the permethylation method followed by acid hydrolysis and GC-MS technique (711. Samples of a-cellulose were dissolved in SO,-diethylamine (DEA)-dimethylsulfoxide(DMSO)andpermethylated in one step using methyl iodide and powdered sodium hydroxide to the degree that no hydroxyl groups were detected by IR. The methylateda-cellulosesampleswere fractionated basedonmolecularweights,followed by acid hydrolyzing and analyzed by GC-MS. Small amounts of residual lignin and sugars originating from hemicelluloses were detected even in the highest-molecular-weight fraction of permethylated a-cellulose.

2. Characterization of Cellulose a. Viscosity, Molecular Weight (Molecular Mass) and Molecular Weight Distribu-

Thecellulosesynthesized in naturehas a certain degree of polydispersitywhich influences its physical properties. Viscosities andor molecular weight distribution (MWD) of cellulose are important parameters for all of its end uses. The polymeric properties of cellulose are sometimes studied in solution, and based on the solution properties, conclusions concerning the average molecular weight, polidespersity, and chain configuration are drawn.However, its isolation fromwood, extraction, and dissolution often degrade the polymer, resulting in structural changes, and the separation is never complete [72]. The length of the molecules forming a cellulose fiber is very important for the internal cohesion of the fiber, which influences its physical and chemical behavior, such as reactivity and accessibility L73J. Knowledge of the MWD of cellulose is an essential requirement, particularly for its processing and industrial utilization. For molecular studies, the quantitative isolation of pure and undegraded species in concerned is required. However, in wood cellulose a series of problems caused by anatomical and chemical phenomena make it impossible to isolate undegraded cellulose as a homopolymeric substance. Problems in isolation of cellulose from wood for the determination of its molecular properties have been reviewed by van Zyl [74] and by Korner et al. 1751. Solvents of Cellulose. To measure the molecularweight,MWD, viscosity, and other properties of cellulose, it is necessary to dissolve it. Due to the highlyordered structure of cellulose, resulting from the formation of fibrils and microfibrils via inter- and intramolecular hydrogen bonding, cellulose is not soluble in common solvents [76J. It has been found that cellulose could be dissolved in strong acids such as hydrochloric, sulfuric. and phosphoric acids [ 771, but its solubilization requires high concentration of the mineral acid,andseveredegradation of the cellulosemolecule and, possibly, substitution will occur. Different solvent systems for cellulose dissolution have been developed over the years 178-1071. Cellulose is soluble in no simplesolvents.There are metal-amine (or ammonia)systems,ammoniumthiocyanate-ammoniumsystems,amineoxidesystems, and dimethylacetamide-lithiumchloride(DMAC-LiCI)systems.Cellulosedissolves in these solvents without previous derivatization, but none of these methods is straightforward and may suffer problems of color, oxidative instability, and degradation of dissolved cellulose. Table 2 showssome of the solvents for cellulose withsome of their relevant characteristics. Much attention has been given to the investigation of ccllulose derivatives tion.

TABLE 2

Solvents of Cellulose ~~

Solvent

Designation

Formula”

Cuprammonium hydroxide

Cuoxan

Cu(NH;),(OH)Z

Cupriethy lenediamine hydroxide

Cuen, CED

[Cu(en)&OH),

Triethylenediamine cobalt hydroxide Triethylenediamine nickel hydroxide Triethylenediamine zinc hydroxide

Cooxen

[CO(~~)~I(OH)~

Nioxen

[Ni(en)21(OH)Z

Zinconxen

IZn(en)21(OH),

Cadoxen

[Cd(en)Z1(OH)Z

EWNN or FeTNa

[(C4H,06)3FelNa,

Triethylenediamine cadmium hydroxide Iron-tartaric acid-sodium complex solutions Methylmorpholine-Noxide

MMNO

Dimethylacetamide-LiC1

DMAC-LiCI

Liquid ammoniaammonium salt

NIi,(l)/salt/water

Dimethylsulfoxide-paaformaldehyde

DMSORF

‘(en): -NH-CH,-CH,-NH-.

General characteristics Good solvating properties, extensive oxidative degradation rather unstable, clear, colored (blue) [891 Good solvating properties, extensive oxidative degradation rather unstable on storage, clear, colored (blue) [89] Good solvating properties, extensive oxidative degradation colored (claret) [891 Good solvating properties, extensive oxidative degradation colored (violet) [89] Questionable solvating power. stable only at low temperatures, slight oxidative degradation, colorless [89] Good solvating properties, slight oxidative degradation, clear, stable, and colorless [89] Good solvating properties, slight oxidative degradation, extremely high salt concentration, colored (green) 1891 Good solvent. M.P. 74°C. To use at low temperature it is necessary to dilute with organic solvents (such as DMSO) [95]. At low water contents, the cellulose is readily dissolved by the MMNO-water system [96]. Nondegrading experimental conditions (i.e., low temperature and short time) [96] Good solvent. It is necessary to activate the cellulose with water. The solutions are stable at room temperature for a long time [97] Only very narrow range of NH,/salt/HzO ratios can be employed to get acceptable cellulose dissolution [97] Good solvent. Stable, clear, colorless [lo61

Refs. [79,801

(79,s 1-83]

[79-841

[79,851 179,861 [79,87-891 [79,90-931 194-971

[97- 101 ]

[97,102]

[ 105-1071

Freer 288

and

Baeza

(nitro, acetyl, or carbanil), usually as an indirect means to study cellulose. Cellulose derivatives are soluble in common solvents, such as THE acetone, etc. The carboxymethyl derivative of cellulose has been used in aqueous GPC [ 1 OS]. The conversion of cellulose into derivatives in some cases is not entirely reliable, and the results obtained must be considereda reflection of the natureof the derivative itself and onlyas a secondary description of the original cellulose sample. There are solvents, such as the amine oxides [96], which are capable of dissolving both cellulose and its derivatives. The aqueous systems usually dissolve cellulose only after extensive swelling of the fibers; the diffusion into the cellulose lattice conduces to either decrystallization or the formation of specific crystalline inclusion complexes. Fast swelling leads to rapid dissolution. Dissolution in nonaqueous solvents occurs more readily; the fibers seem to explode into numerous spindle-like fragments, which are rapidly dissolved. In general, the metal-amine solvents present good dissolving power, and the relative stability is dependent on the nature of the metal. The dissolving power is dependent on the metal and the reacting base concentrations. In these types of solvents the complexforming metal ions are partly linked to the cellulose molecules. High concentrations of both metal and base are required, limiting the use of the resulting solutions essentially to molecular weight and viscosity determinations. For the determination of molecular weight, the stability of the solutions of cellulose is the fundamental criterion. Solution in cuoxan and cuen, which have been considered good solvents for cellulose, are easily autooxidized and must be used under an atmosphere of nitrogen or hydrogen. Sometimes antioxidants are added andor the solutions are prepared before using. The oxidative degradation of cellulose solution in cadoxen is nearly negligible even for the highest-molecular-weight samples. The rate of degradation in cuen is very much greater than in EWNN and in this is slightly greater than in cadoxen, but still insignificant. The general characteristics of the metal-amine solvents are included in Table 2 . In general, the metal-amine solvents are not convenient for osmotic pressure and dialysis experiments because the membranes are not resistant to them, nor for light-scattering measurement, primarily because they are very strongly light-absorbing in the region normally used. However, cadoxen solutions of underivatized cellulose prepared by a modified procedure have been used for osmotic pressure and dialysis experiments by using ordinary cellulose membranes and also for light-scattering measurements [891. Formeasurementofmolecularweight by the osmoticpressure or light-scattering method, cellulose derivatives (nitrate or acetate) are usually used. Cellulose is solvated in strongly polar, aprotic solvents, such as N,N-dimethylformamide (DMF) or N,N-dimethylacetamide (DMAC) after introduction of a certain amount of dinitrogen tetroxide (NzO,) or nitrosyl chloride (NOCI). A highly esterified cellulose nitrite ester is formed and the cellulose is dissolved in the N,N-dialkylacylamides, which likely act as proton acceptors. The cellulose nitrite is an unstable compound and unmodified cellulose will be regenerated easily by contact with protic solvents. The solubilization of cellulose in nonderivatizing solvent systems, such as DMACLiCI, amine oxide, and liquid ammonidammonium salt systems, occurs only in selected ranges of composition, and the condition of dissolution becomes critical. The NH,/sal/H,O ratios that can be used to obtain acceptable cellulose dissolution are in a very narrow range, and usually outside the conditions normally used for swelling. N-methylmorpholine N-oxide ("NO), dimethylethanolamine N-oxide (DMEAO), and a mixture with DMF or DMSO in the presence of different amounts of water have been used as solvents for cellulose derivatives. The amine oxide/water systems may either dissolve or only swell cellulose depending on the water concentration. This is of critical

racterization Chemical

of Wood

289

importance: up to a certain critical amount (19% for MMNO and 14% for DMEAO, and intermediatevalues for the mixtures of bothamineoxides), the liquids are unable to dissolve cellulose. The dissolving power of the liquid increases by removing more water from the amine oxide system, the dissolution of cellulose in amine oxide-water systems occurs only at low water concentrations. Small amounts of water are necessary to lower the melting point of the amine oxide system and consequently to obtain dissolution of cellulose at lower temperatures, avoiding its degradation. Apparently, a certain amount of water is needed to reactivate and reopen the internal areas of cellulose pores that have been closed or deactivated by drying during the pulping process. To be able to work at low temperature (near room temperature), it is necessary to dilute the solutions with an organic solvent (such as DMSO or DMF). The amounts of the organic solvent that may beadded to cellulosesolutions in MMNO/waterareconsiderablebefore the cellulose precipitates. For "NO, the active dissolving part of the molecule is the N - 0 appendage, which forms up to two H bonds with hydroxylated compounds. MMNO in the presence of 13.5% of water (corresponding to the monohydrate) dissolves the cellulose easily. When 2 moles of water (23.5% of water) are present to interact with the N - 0 bond there is no further driving force for NMMO to interact with the cellulose OH groups. In the case of DMEAO, a similar rationale has been developed, where the N - 0 group is able to form only one hydrogenbondwith an external OH due to which it may adoptacyclicconformation stabilized by the H bond. In the dissolution of cellulosewithLiCI-DMACsysteman initial activation with water to induce swelling of the fibers is required. A subsequent solvent exchange procedure, in which the water is replaced with DMAC, must be realized. Complete removal of water is necessary to achieve thorough penetration of the solvent into the sample and total dissolution. The LiCl forms a complex with dimethylacetamide, releasing Cl-, which acts as a base toward the cellulose hydroxyl group hydrogens. A possible mechanism of interaction of cellulose with LiCl and DMAC is shown in Fig. 6. Other salts of Li, such as LiBr, LINO,, etc., do not work. The solutions of cellulose in the DMACLiCI system are very stable at room temperature for a long time. Viscosity of Cellulose Solutions. The molecularweight of macromoleculessuch as cellulose can be determined by absolute and relative methods. The average molecular weight of cellulose is commonly obtained from the determination of viscosity of a solution

FIGURE 6 Solvated complex of cellulose, LiCI, and DMAC. (From Ref. 97.)

Baeza and Freer

290

of cellulose in an appropriate solvent. To determine the degree of polymerization of a cellulosefromviscometricmeasurement,varioussolventsandtypes of viscometer are used. The usual solvents are cuen, cadoxen, and cuoxam [89,109]. Other solvent systems that have been used are MNNO/H,O/DMSO [95] and DMAC-LiCI [ 1031. More recently, DMSO/PF has been proposed for viscosity determinations [ 1061. The presence of a macromolecular solute increases the viscosity of a solution over that of the pure solvent. The effect is large even at low concentrations. Most directly related to the nature of the individual solute molecules is the intrinsic viscosity [V], which has the effect of macromolecule intermolecular interaction removed by the extrapolation to infinite dilution. [v]= limit qSJC when C -) 0, and it is obtainedfrom the curve qJC = f(0forazeroconcentration.The specific viscosity is calculatedfrom the formula = (q - q,,)/q",where q and q , are the viscosities of the solution and the solvent, respectively. Several theoretical studies have tried to give a relationship between viscosity and concentration. The formula of Huggins [ 1101 is the most commonly used: VSP 7 = [q]+ k'[q]'C

L

where k' is the Huggins constant. The intrinsic viscosity ofmacromolecularsolutionsand related through the Mark Houwink formula:

the molecularweight are

[v] = KMu = K'(DP)" where, at a given temperature, K and a are constants for a specific polymer-solvent system [ 1I l]. The degree of polymerization (DP) is expressed as the average molecular weight ( M ) divided by 162 (where 162 is the weight of an anhydro-glucose unit). The viscosity method is a relative method, so the values of K and a must be calculated from [q]values of samples whose molecular weights are already known by absolute methods. A plot of log[q] versus log M usually yields a straight line of slope a and intercept log K. Values for the constants K and a for cellulose in various solvent systems are given i n Table 3. Molecular Weight (Molecular Mass) and Molecular Weight Distribution. Due to the polydisperse nature of cellulose, M depends on the method usedin its determination. Table 4 shows the types of average molecular weights (molecular mass), definition, and measurement methods. The number average molecular weight (M,,) of cellulose can be measured using osmometry or by determining the number of reducing end groups. The weight-average molecular weight M,
aracterization Chemical

of Wood

291

TABLE 3 Constants for the Mark-Houwink Equation [TI = KM" = K'(DP)o as Determined for Cellulose and Cellulose Derivatives in Several Solvents

system Solvent

K' (dL/g)

K (dL/g)

Cellulose 6.8 X 10" 9.8 X lo-'

Cuoxan Cuen Cadoxen EWNN NH3/NH4SCN MMNO/H?O/DMSO [ 1/1.5 (w/w)]

3.38

X

3.85

X

10" 10

3.15

X

10.'

6.86

X

IO-'

6.6

X 10"

19.9 X 10"

a

Method

0.9 0.9 0.75 0.76 0.93 1.01 0.95 0.79

SD SD SD SD, LS GPCNISC SD VISC VISC

0.84

GPC/VISC

1.0

VISC VISC VISC

Ref.

Cellulose tricarbanilate (CTC) 5.3 x 10P

THF

Cellulose nitrate 5 x IO" 11 x 13 X IO-.'

Acetone Ethyl acetate

1 .0

I .0

"SD, sedimentation-diffusion: LS. light scattering; VISC, viscosity: GPC, gel permeation chromatography

TABLE 4

Types of Average Molecular Weight and Methods of Measurement

Type

2, 2,n,

Number-average molecular weight

tLM, M,, = -

Weight-average molecular weight

-

Osmotic pressure; reducing end group

x, x,

scattering Light

"1

M,v= -

ILM,

Z-average molecular weight Viscosity-average molecular weighth

Viscosity -

51,

\

are the number of molecules of molecular weight M,. for cellulose.

ha= 1

/

Baeza and Freer

292

nitrate was 1.9 (M,$. = 2.7 X IO', by light scattering) [ 1201. For Novacel K wood pulp the M,,JM,, was found to be 2.72 with DP,,. = 870 and DP,, = 320, while for Tyrecell N wood pulp the values were 2.69, DP,,. = 1385, and DP,, = 430 [ 1211. For MWD determinations, size-exclusion chromatography (SEC), such as GPC and gel filtration chromatography (GFC), has gained wide acceptance as a preferred method. SEC is a special form of liquid chromatography; it is an entropically controlled separation technique that depends on the relative size or hydrodynamic volume of a macromolecule withrespectto the size andshape of the pores of the packing.GFC is referred to as aqueousSEC at less than 1 kPa,whilesystemsusinghighpressuresandnonaqueous solvents are referred to as GPC. With high-performance SEC (HPSEC), the relative MWD can be obtained. If the column is calibrated with known molecular weight samples and/ or using molecular weight detectors, average molecular weights values can be obtained. In other words, in conventional SEC the raw data represent the elution volume distribution of the polymersample by weight,and a transformation of it tomolecularweight is required. Cael et al. [ 1221 and Lauriol et al. [ 1231 used a double detection system consisting of a low-angle laser light-scattering photometer (LALLS) and UV to determine the MWD of cellulose which had converted into tricarbanilate. SEC-LS detectors have been used to characterize polysaccharides [ 1241. Triple detector systems LS andviscosity detectors, both combined with SEC, were used to study carbohydrates [ 1251. SEC-FTIR has been reported for the characterization of cellulose esters [ 1261. The capability of SEC has been improved considerably by development of online viscometers and light-scattering detectors which have greatly extended the usefulness of SEC [ 1241. The appropriate analytical techniques are generally dependent on the dissolution of the polymer. Two aspects must be considered in MWD of cellulose: 1.

2.

Due to the insolubility ofcellulose in the organicsolventscommonly used as an eluting phase in GPC, it is necessary to convertcellulose into stable and soluble derivatives. It is necessary to select standards to obtainameaningful calibration curve so that data from the GPC chromatograms can be converted into MWD curves to obtain true average molecular weights.

The calibration procedure has been generally achieved by using the Mark-Houwink coefficients ( K and a ) derived from both the polymer under study and narrow-distribution polystyrene standards. Theaveragemolecularweightsand MWD obtainedfromsuch calibration methods depend on the correctness of K and a for the polymer/solvent pair. Different efforts have been realized to obtain more realistic calibration curves. In the initial application of GPC to study cellulose, it was converted into cellulose nitrate. Some examples are given below. Segal [ 109,119,127]studied the MWD of nitrated cottonand wood cellulose by analyzing the chromatograms of the nitrate derivatives in THF solutions, using polystyrene calibration curves. The values for the average molecular weights were markedly greater than those measured viscometrically. The authors consider that the extensive solvation of the cellulose nitrate molecule by THF may account for these results. However, the results of Meyerhoff and Havanovics [ 1281, by comparing the integral distribution obtained from the precipitation fractionation of cellulose nitrate and the GPC chromatograms in THF, gave lower average degrees of polymerization than those obtained by polystyrene calibration curves applied to cellulose nitrate. Huang and Jenkins [ 1211, using a calibration curve based on cellulose nitrate, showed that the average molecular weight was in good agree-

aracterization Chemical

of Wood

293

mentwith the molecularweightobtained by viscometryandosmometry. The cellulose samples for the calibration were prepared by ( I ) precipitation fractionation of cellulose nitrate and (2) gamma-ray irradiation of wood cellulose. They compared theseresults with those obtained by using polystyrene standards and concluded that for a given molecular weight, cellulose nitrate eluted at a lower elution volume during the GPC separation process than polystyrene. This was caused by the relative stiffness of the cellulose nitrate chain because of the glucosidic linkage. The use of nitrate has some disadvantages [ 1291. The nitration procedure can cause significant chain scission if proper precautions are not taken, thereby causing the MWD of the nitrate to differ from that of the underivatized cellulose, and the stability of the nitrate may be limited, and consequently, a considerable variability exists in the degree of substitution. A wide range of polydispersities has been described on what appear to be rather similar starting materials. As Cael et al. [ I221 pointed out, the method of calibration is critical. The initial difference between nitrates and carbanilates basedDP,,.for the nitrate/ acetone system could be reconciled by adjustments of Mark-Houwink coefficients, suggesting that chain scission really does not occur during nitration. A series of cellulose nitrates have been prepared and evaluated by SEC [130]. Eremeeva et al. [ 13l ] studied non-size exclusion effects of nitrocellulose using THF as the mobile phase and silanized silica packing. The substitution heterogeneity and the presence of ionic groups in nitrocellulose influenced the SEC behavior of nitrocellulose. Addition of acetic acid to the mobile phase suppressed nonexclusion effects and led to the validity of universal calibration between nitrocellulose and polystyrene. HPSEC of cellulose tricarbanilate (CTC)has emerged as a useful techniquefor obtaining MWD of cellulose samples [ 114,122- 125,129,132- 1391. The usual procedure for the preparation of CTC derivatives is to react the cellulose with phenyl isocyanate in pyridine as the solvent and catalyst as well. Afterremoving the unreactedphenyl isocyanate, the CTC is isolated by precipitation in a nonsolvent (e.g., into methanol). The CTC is a stable derivative and the fully trisubstituted product may be readily obtained. The carbanilation of cellulose to obtain soluble samples for SEC was discussed by Saake et al. [140]. SEC of the tricarbanilate derivative of cellulose has been applied to study the effects of xylanase pretreatments on the MWD of hardwood and softwood kraft pulps 11351. The carbanilate procedure was similar to that used by Schroeder and Haigh [137], butwas done at a microscale and SEC was performed on a series of four Ultrastyragel columns (IO', IO5, linear, IO") held at 35°C with THF (1 mL/min) as the eluting solvent, using a UV detector ( h = 278 nm). A universalcalibration curve was used, withthe Mark-Houwink coefficient published by Wood et al. [ 1381. The reproducibility of the technique allows the indication, with reasonable accuracy, of the effects of enzyme-action, i.e., xylan removal, and the integrity of the cellulose component. In most of the studies on CTC the universal calibration procedure using polystyrene standards has been applied 1129,137- 1391. Vidal et al. [ 1391 evaluated the use of universal calibration for cellulose tricarbanilates using known Mark-Houwink constants for the cellulose derivatives. Absolute molecular weights were obtained by using LALLS detector, and hence the construction of an absoluteMWDwasmade[122-123,1401. The GPCI LALLS technique has been applied to tetrahydrofuran (THF) solutions of CTCs prepared from cellulose having a wide range of molecular weights [ 1401. Direct GPC measurements of cellulose in solution would avoid many of the problems associated with derivatization. As was indicated above, due to the highly ordered structure and the strong inter- andintramolecularhydrogenbonding,cellulose is notsoluble in

294

Baeza and Freer

common solvents. Some methods have been devised for the determination of nonderivatized cellulose [98-99,1411. Kennedy et al. [98] reported the MWDs of cellulose samples dissolved in DMAC-LiCl results. They were determined by SEC using a poly(styrenedivinylbenzene)GPCcolumnwith a wide fractionation range (10' to IO7 polystyrene MW).Thesame solventsystemwasused as the mobilephase. The methodinvolves activation of the cellulose with waterto induce the swelling of the fibers, solvent exchange, and complexation with LiCI. The amount and the length of the dissolution time required depends on the cellulose sample. Birch and spruce cellulose samples required less LiCl than cotton linters. The chromatographic traces of the samples show bimodal distributions for the softwood and hardwood samples with a peak eluting very close to the exclusion volume of the column and the other within the fractionation range. Two sharp peaks were obtained for the birch hardwood sample, whereas for the softwood sample the peaks have relatively higher dispersities. Timpa [99] appliedthe universal calibration concept to obtain MWDs for cotton cellulose, corn, and wheat starch flours and avocado cell wall polymers. Heused the samesolventsystem(DMAC-LiCl) to characterize them by SEC,witha refractive index and viscometry. Applications of SEC to the characterization of cellulose and cellulose derivatives were reviewed [ 142- 1441. b. OtherChcwacterizationProperties. Variousinstrumentaltechniqueshavebeen used to characterize cellulose. These include infrared (IR) spectrometry, Raman spectroscopy, NMR, and X-ray and electron diffraction. In the solid state a regular system ofthe hydrogen bonds between cellulose molecules results in an ordered system with crystal-like properties. There is also a noncrystalline or amorphous portion in cellulose. The degree of crystallinity (crystallinity index), which is a measure of the crystalline portion in a cellulose sample, depends on the origin of the cellulose. The crystallinity index can be determined by X-ray diffraction [145,146], and by IR using ratios of certain absorption bands [ 147,1481. The values of crystallinity index for wood pulps range from 60% to 70%. Wood celluloses tend to be less crystalline than other celluloses. therefore less amenable to study by diffraction methods. Also, wood celluloses are highly sensitive to electron beam damage at doses far lower than those required for usual high-resolution electron microscopy [149]. On the other hand, Valonia cellulose is highly crystalline and resistant to radiation. It is a widely accepted standard for natural cellulose. A considerable number of studies on cellulose structure have used Valonia cellulose [150- 1521. Variation in the physical structure of cellulose has been observed according to its sources and developmental stage. This variation, which includes differences in microfibril crystallographic orientations, degree of polymerization, crystallite size, pattern of glucanchain hydrogen bonding, and glucan chain polarity, has made it difficult to determine the basic crystalline structure of cellulose. The most common crystalline form (allomorph) of native cellulose is cellulose I, which is metastable and can be irreversibly converted into another crystalline state. TWO distinct forms of cellulose I (I, and Io) have been reported by Atalla and VanderHart [ 1531 based on CP-MAS I3C NMR evidences. Morerecently, the electron diffraction experiments of Sugiyama et al. [ 1541 demonstrated the coexistence of these two phases (I, and Ia) as regions of different crystalline structure rather than alternative chains along the cell wall. Cellulose I, and I, differonly in their patterns of hydrogenbonding; their molecular conformations are identical [155,156]. The I, and I, allomorphs are most easily distinguishable via the C4 resonanceof certain solid-state I3C-NMR and by OH stretching regions of the Raman spectra [155,157]. Electron diffraction patterns are consistent with

aracterization Chemical

of Wood

295

triclinic and monoclinic unit cells in these two forms [156]. The native celluloses can be classified into two types: algal-bacterial and cotton-ramie-wood types, which are referred to as celluloses I,, and I,, respectively [ 1581. The cellulose I1 polymorph is formed from cellulose I by mercerization or by precipitation from solution, and it is the most stable allomorph known. These two are the most common polymorphs. Other forms (cellulose 111, IV, and the so called cellulose x) have been reported. Transformation of cellulose into its various lattice modification is summarized by Fengel and Wegener [ 1591. Each cellulose allomorphcanbe identified by its characteristic X-rayor electron diffraction pattern [160,161]. Infrared techniques have been used to analyze the crystal structure of cellulose and to study cellulose reactions. IR in solid-state techniques can be used to distinguish among cellulosesamplesfromvariousorigins.Theassignmentof the IR absorptionbands of cellulose is basically derived from that of the glucose unit. Studies of the polarized infrared spectrum [ 162- 1661, the effect of deuteration [ 1631, and the correlation of the bands or groups of bands with the bands of chemically related compounds [ 167,1681 have given information for a complete assignment of the absorption bands. Table 5 shows the relative intensity, polarization, and assignment of the main absorption bands of celluloses I and I1 [ 1691. The early work on the vibrational spectra of cellulose was summarized by Blackwell [170]. The information obtained from conventional IR was limited by lack of resolution of bands in the spectra.With the introduction of the Fourier-transformspectrometers (ITIR), new methods for measurement and evaluationaredeveloped. Theincrease in sensitivity resulting from the combination of FTIR and computer techniques has greatly enhanced the usefulness of this technique. Improved resolution can be obtained by deconvoluting the spectra [ 17 1,1721, curve-fitting [ I731 or second derivatization [ 174- 1771. FTIR spectra of cellulose, wood and other major components are shown in Fig. 7. Marriman and Mann [ 1781 were the first to point out the difference between spectra obtained from Valonia and bacterial celluloses, in which a band at 3242 cm” appears, and those of tunicin, ramie, or cotton, in which this band is absent. They suggested that the difference was “either due to a different crystal structure in the two cases or to a difference in degree of perfection of crystals having the same basic molecular arrangement.” Liang and Marchessault [ 1641 also observed an intense band at 3245 cm” exclusively in Valonia and bacterial celluloses. More recently, Michell [l761 studied a series of native celluloses by using a second-derivativetechnique,supporting the hypothesis of Marriman and Mann that the native celluloses possess two different crystal structures. Transformation between phases of cellulose by some treatments has been monitored by IR. The conversion of I, phase to the most stable I, by an annealing treatment has been reported by Sugiyama et al. [161]. It was demonstrated by combined infrared and electron diffraction techniques that some absorption bands in the IR are specific to the I, (3240 and 750 cm”), while others correspond to I, cellulose allomorph (3270 and 710 cm-’). Michell [l771 analyzed the changes in structure of a eucalyptus wood cellulose by treating it with solutions of NaOH ranging from 0 to 20%. Better resolution of the bands by using second-derivative FTIR spectroscopy can be used in the conversion of native cellulose into mercerized and into an amorphous form. It can be used to confirm known changes in band intensities and to discover more subtle changes. The number of bands in the second-derivative spectra of cellulose sensitive to mercerization is far greater than the four bands (at 1428, 1111, 990, and 893 cm-’) which were identified earlier by normal spectra reported by McKenzie and Higgins [ 1791.

Baeza and Freer

296

TABLE 5 Frequency (cm ")

Infrared Assignments of Cellulose [l691 Relative intensity Polarization

Assignment Infrared Assignment of Cellulose Peaks I

Ca

650

m

895 965 988

W

By Tsuboi" OH out-of-plane deformation

W

m

100s

S

1030

1069 1078 1 l06 1117

vs vs vs vs vs

CO and CC stretching and CH, rocking

1 l61

S

1204 I232 1249 1275

W

1310 1335 1365 I426 1446 1630

m m m m m

1650

m m vw vw vw

Adsorbed HzO

W

CH2 symmetric stretching CH, stretching CHZ antisymmetric stretching OH stretching OH stretching By Liang et aLh

1720 Ca 2340 Ca 2500 Ca 2700 285 1 2907 2967 Ca 3300 Ca 3400 663 -700 -740 -800 895 985 1000 1015 1035

W

W

CH2 wagging CH and OH deformation

m CH deformation CH in-plane deformation CH deformation CH, symmetric bend OH in-plane deformation

111

m W

vs vs

m m sd sd

OH out-of-plane bend

W

CH2 rocking? Ring breathing (p) Antisymmetric out-of-phase stretching

sd sd m

CO stretching

S

1058

S

11 I O 1 l25

S

1162

S

Antisymmetric in-phase ring stretching

sd Antisymmetric bridge oxygen stretching

aracterization Chemical TABLE 5

of Wood

297

Continued

Frequency Relative (cm-Polarization ') intensity

1205 I235 1250 1282 1317 1336 I358 1374 1430 I455 l635 2853 2873 2897 2910 2914

-

Assignment OH in-plane bending

W

W W

CH bending CH2 wagging CH in-plane bending

m m m m m m sd

CH bending CH2 bending CH in-plane bending Absorbed H,O CH, symmetric stretching

-

CH stretching CH, antisymmetric stretching CH stretching

2945 2970 3245 3275 3305 3350 3405

OH stretching (intermolecular hydrogen bonds in 101 plane) OH stretching (intermolecular hydrogen bonds) OH stretching (intermolecular hydrogen in 101 plane) Infrared Assignment of Cellulose Peaks I1

650 700 760 800 892 965 996

S

1005

S

1020 1035 1060 I078 I l07 I133 I200 1225 I273 1277 1315 1335 I365 1375 1416

S

sh sh sh m sh

OH out-of-plane bending

II I1 I I

II

rocking? CH, Ring breathing (B) C ,frequency group

S

CO stretching

S

vs vs 5 S

m m

Antisymmetric in-phase ring stretching Antisymmetric bridge oxygen OH in-plane bending?

W

m W W

m m W

CH bending CH2 wagging OH in-plane bending CH bending CH2 bending

Freer 298

and

Baeza

TABLE 5 Continued

Frequency (cm"') sh

1440 1470 1635 2850 2874 289 1 2904 2933 2955 2968 298 1 3175 3305 3350 3447 3488

Relative intensity Polarization

Assignment

II sh W

sh S S

m W W W

F S

S

vs S

I

II?

$1

OH in-plane bending H,O Adsorbed CH, symmetric stretching CH stretching

L

I?

"l I1

:I)

antisymmetric CH2 stretching CH stretching OH stretching (intermolecular hydrogen

bonding)

I I

"l II

OH stretching (intermolecular hydrogen bonding)

"From Ref. 163. hFrom Refs. 165 and 166.

Deconvolution of the IR spectra of cellulose and cellulose derivatives, particularly in the range of OH stretching vibrations, provides detailed evidence on crystallinity, crystal modification, and degree of substitution. Much better separation of the bands in an IR spectrum after deconvolution permits improvement in the recognition of minimal differences. Distinction between cellulose I and cellulose I1 was done [ 1801 by attributing certain bands to the crystal modifications independent of the crystallinity. Deconvolution of the range 3200-3700 cm-' reveals distinctions between celluloseI and cellulose I1 by different band positions. The disappearance of some bands (3350 and 3418 cm") and the shift of the band at 3466 cm" to 3438 cm" are observed during the change from crystalline to amorphouscellulose.Certainbandsareassumedtobesuitable for a measurement of crystallinity. The influence of water on the intensities and positions of the bands in deconvoluted spectra of cellulose was analyzed [ l8 l]. It was concluded from the deconvoluted spectra of the OH and CH2 ranges of a cellulose-water system that the positions remain constant except for very high water content. The intensity of the various bands is influenced by water in varying degrees. 'H and I3C NMR spectroscopy are invaluable for primarystructuredetermination and for conformational analysis of polysaccharides. A considerable number of reports have appeared recently [ 182- 1901. The hydroxyl resonances observed by an NMR analysis of carbohydrates, using water-suppression techniques in supercooled aqueous solutions, provide information about chemical shifts and coupling constants that are not available from more traditional studies by D,O [ 191,1921. Modification of crystallinity and crystalline structure of cellulose (of Acetobacter xylinurn) has been demonstrated in celluloses isolated in the presence of hemicelluloses. "C-NMR has been used to analyze the alteration in crystallinity and shift in relation to the amounts of I, and I, [193]. The relative intensities of C4 and C6 peaks associated

299

Chemical Characterizationof Wood 0.90

a

h

/-

0 0

a a

!

\

I

I

Frequency (cm-') FIGURE 7

FTIR spectra of E. regnas wood and its major components. (From Ref. 174.)

with the ordered regions (sharper resonances) and those broader on the upfield side of the crystalline counterparts resonating downfield (associated to disordered regions) allow the crystallinity to be determined. The solid-state transformations (I(, + I,) of Valonia cellulose crystal by an alkaline hydrothermal treatment have been analyzed by comparing the diffraction diagrams before and after hydrothermal annealing, which shows substantial modifications [161]. This transformation has also been observed by I3C-NMR [194,195]. The spectrum of annealed Va-

Freer 300

and

Baeza

lonia cellulose shows a well-resolved multiplicity of two for the signals of the carbon atoms at C 1, C4, and C6 of the glucose moieties, indicating only two glucose residues that are magnetically nonequivalent in the unit cell. In the spectra of the original Valonia samples, at least three signals were observed for these carbons. The relevant signals in the I3C-NMR spectra that permit distinguishing between I,, and I, are well resolved in spectra of algal or bacterial celluloses, but poorly resolved in the spectra of wood or pulp. Studies of combinations of "C-NMR with the analysis of principal components 11961 and/or improved resolution-enhancement function [ 1971 have been reported. By using the latter approach. the ratio IJI, is estimated from the ratio of heights of the peaks at 90.2 and 88.5 ppm assigned to C4 in I,, and I, crystallite interiors, respectively. Values of 1.8, 0.8,and0.4havebeenreported for softwood timber, thermomechanical pulps, and Kraft pulp, respectively [197,198]. The effect of temperature, time, and alkali concentration in the conversionbetweencrystallineformsofcellulose during pulping has been reported [1981.

B.

Hemicellulose or Polyoses

Hemicelluloses constitute 20-30% of wood. They are found predominantly in the primary and secondary cell walls. A smaller amount occurs in the middle lamella. They differ from cellulose by containing various sugar units, with much shorter chains, and by branching of the chain molecules. They are soluble in alkali, and some hardwood polyoses are even soluble in water. Hemicelluloses are more reactive than cellulose. Hemicelluloses are generally classified according to the types of sugar residues present. Thus, xylan is a polymer of ~-xylosylresidues and mannan a polymer of D-mannosyl residues (homoglycans). However, the most frequently found hemicelluloses contain two to four, and rarely five or six, different sugar types (heteroglycans). Hemicelluloses in softwoods and hardwoods differ not only in the percentages of total hemicelluloses but also in the percentages of individual polyoses and their composition. In softwoods, the major hemicelluloses are partially acetylated galactoglucomannans, and a small amount of arabino-4-0-methyglucuronoxylans.In hardwoods, the predominant hemicelluloses present are 0-acetyl-4-0-methyglucuronoxylans (acetylation of the xylose groups is about 70% on C-2 or C-3) with a small proportion of glucomannans. Partial chemical structures of polyoses are shown in Fig. 8.

1. Isolation and Determination of Hemicelluloses Duringgeneral wood analysis,hemicellulosesmustbeisolated.Theycan be extracted directly from wood but most commonly from holocellulose. Extraction from holocellulose rather than from wood results in a more complete removal of hemicelluloses with less contamination from lignin. However, during the holocellulose preparation some soluble hemicelluloses may be lost. They may be oxidized and degraded by hydrolytic reactions [39,199,200]. These problems can be avoided by extracting hemicelluloses directly from wood. However, they do not represent the total polyoses of wood and must be purified to remove lignin. These polyoses isolated are especially convenient for structural analysis. A large portion of hardwood xylans can be extracted in considerable yields from extractivefree wood directly by aqueous solutions of potassium hydroxide [66]. The yields depend on the wood [53,201,202]. Only the water-soluble arabinogalactan, when present, is readily extracted from softwoods. A significant portion of the main polyoses in softwoods can be extracted only after delignification.

CH20H COOH

0 OH

OH

FIGURE 8 softwood.

1

OH

OAC

CH,OH

CH20H

I

OH

Partial chemical structures of hemicelluloses: (a) 0-acetyl-4-0-methylglucuronoxylanfrom hardwood; (b) 0-acetyl-galactoglucomannan from

302

Baeza and Freer

Hemicelluloses are commonly isolated from holocellulose by extraction with aqueous alkaline solutions. Potassium, sodium, and lithium hydroxides exhibit similar abilities to remove xylose-containing polymers from a chlorite holocellulose [203]. Aqueous solutions of potassium and sodium hydroxide are the most extensively used agents for extraction of hemicelluloses. However, potassium hydroxide solutions are used widely because the potassium acetate formed in the neutralization with acetic acid is more soluble in ethanol, which is used to precipitate the isolated hemicelluloses [204]. Sodium hydroxide solution hasbeenfound to be more effective in extracting the resistant mannose-containing hemicelluloses. A typical isolation procedure [49,204] includes a two-step extraction. A general procedure is as follows: the holocellulose is transferred to an Erlenmeyer flask. The air is displaced by passing nitrogen gas through the flask. Potassium hydroxide solution (5%) is added while shaking andthe mixture is stored at 20°C for 2 h. After filtering and washing the residue successively with 5% potassium hydroxide and water, the hemicelluloses are precipitated by the addition of ethanol after neutralization with acetic acid (hemicellulose A).Theresidue is extractedwith24%potassiumhydroxide, filtered, andwashedwith potassium hydroxide (24%), water, and acetic acid (10%). The hemicelluloses are precipitated by additionofethanol to the filtrate andwashed(hemicellulose B). The hemicelluloses A and B are allowed to settle, the supernatant liquor is removed, and the precipitates are centrifuged with ethanol to remove water and finally with ether to remove the ethanolandthen dried. The sum of the two fractions does not represent the total content of hemicelluloses of the sample. Some, such as pentosan, are lost during delignification, some residual polyoses may remain in the a-cellulose, and not all the dissolved polyoses may precipitate out from the alcoholic solutions [205]. Fengel [206] applied 5% and 17.5% NaOH consecutively for the determination of softwood polyoses. The content of polyoses corresponds tothe sum of the polysaccharides isolated [49]. The content of pentosan, uronic acids, and acetyl groups were directly determined in water-extracted wood by Smelstorius and Stewart [205]. Hemicellulose purification may be achieved by fractional precipitation (by successive extraction with a series of solvents or with complexing agents) or by chromatographic techniques. Fractional precipitation gives good results only if the difference in the solubility of the hemicelluloses or derivatives is large. Fractional precipitation with ethanol has been widely used for the purification of hemicelluloses [207-2091. Gradual addition of ethanol to dilute aqueous solutions at near-neutral pH allows the isolation of pure polysaccharides. Separation at low pH has been carried out, usually at low temperature, to avoid or at least decrease acid hydrolysis, and the separations should be conducted rapidly. DMSO was found to be a good solvent for hemicelluloses, and this solvent can also be used as a rather selective precipitating agent for hemicelluloses [210]. The naturally acetylated xylan has been isolated by extraction with DMSO or hot water from holocellulose preswollen with liquid ammonia [210,21l]. Then, extraction with aqueous potassium or sodium hydroxide is carried out in the presence of borate to permit the removal of glucomannans [212]. A high percentage of acyl groups was found in the DMSO extract from birch holocellulose. Asodiumhydroxide solution is better thanapotassiumhydroxide solution as a solvent for glucomannans. Therefore, extraction with potassium hydroxide solution effects a greater separation between xylans andthe less soluble glucomannans. Addition of sodium borate to the alkali facilitates the dissolution of galactoglucomannans and glucomannans. Borate acts by complexing with the cis hydroxyl groups. The effect of cations (Li+, Na+,

aracterization Chemical

of Wood

303

and K') andborateon the extraction of pulpshasbeenreported [213]. Southernpine pulps were extracted with Li, Na, and K hydroxides at several concentrations from 0.5 to 4.0 molal. The extraction of xylans and glucomannans depends on the ability of alkaline solutions to swell the cellulose structure and also on their ability to form polyanions with the two hemicelloses (carboxilate groups of xylan, the borate complex with glucomannans, and the alcholates with both at high alkali concentrations). Various alkali concentrationsfrom2% to 24%havebeenused for hemicellulose extractions.Alkali extraction can result in manychanges in the polysaccharides 12141. Alkali extractions have the disadvantages of removing acetyl groups [215,216], a significant proportion of the hexuronic acid residues in 4-0-methylglucuronoxylans are cleaved [217], and alkaline degradation occurs at the reducing ends of the polysaccharides [218]. These disadvantages can be avoided by using dilute alkaline extraction under nitrogen to lower oxidative and alkaline degradations, addition of iodine ion to increase the stability of the polysaccharides during the alkaline extraction [219], reduction of the aldehydic end groups with borohydride to reduce the alkaline degradation [218], and successive treatments with increasing alkali concentrations to remove soluble hemicelluloses in a sequential fashion in order to avoidexposing the moresolublehemicellulosestohigh alkali concentrations. An effective alternative procedure for extracting hemicelluloses from softwoods is to add barium hydroxide to the holocellulose to block dissolution of mannose-containing polysaccharides [220]. The glucomannan is separated from contaminating xylan by precipitating it with barium hydroxide [221]. The addition of barium hydroxide at several points in a separation scheme followed by acidification with acetic acid and precipitation with ethanol yields glucomannan and galactoglucomannan fractions 162,222,2231. Xylans are then readily extracted with 10% aqueous potassium hydroxide. After the removal of barium ions, sequential extraction with I % and 15% sodium hydroxide removes galactoglucomannan and glucomannan, respectively. Xylans and glucomannans dissolved in saturated aqueous alkaline earth halides may be fractionated with iodine according to their extent of branching by precipitating the less highly branched components as a blue complex [224]. An alkaline solution of copper(I1) salts can be added to a hemicellulose solution to precipitate those containing significant amounts of D-mannosyl or D-xylosyl units [60]. A Fehling solution precipitates hemicellulose complexes enriched in xylose and impoverished in arabinose, galactose, and glucose [225]. Copper(I1) acetate can also be used to separate neutral polysaccharides from acidic polysaccharides 12261. Quaternaryammonium salts [227-2301suchascetyltrimethylammoniumbromide or cetyltrimethylammonium chloride have been used for fractionation of acid polysaccharides, and it has been shown that the method can be applied to neutral polysaccharides [230,231 1. Purification of hemicellulose extracts has been carried out by using paper chromatography and column chromatography. Among the column packings are alumina, carbon, cellulose. diethylaminethyl (DEAE) cellulose, and synthetic ion-exchange resins [232]. In one application [2331, the hemicellulose was first extracted from a spruce holocellulose with potassium hydroxide solution. After neutralization, a portion of the extract was applied to a DEAE cellulose column in the acetate form, and eluted with solutions of NaOH (increased concentration from 0.1 to I N). Most fractions contained principally arabinoxylan, and some galactoglucomannan was found i n two fractions. Unfortunately, elution of the column with alkali was accompanied by dissolution of the ion exchanger.

304

Baeza and Freer

Separation of polysaccharides by molecularweight is carried out by using GPC [234]. The most common column materials are Sephadex [235] or polyacrylamide gels [236,237]. If only the total content of hemicelluloses is required, isolation of the hemicellulose is not necessary. The total content of polyoses of a sample can be determined by sugar analysis of total hydrolysis of the polysaccharides. The hydrolysisproceduresand the techniques used in the sugar analysis are discussed in the section on analysis of wood carbohydrates. The hemicellulose content of wood celluloses and pulps can be determined from the alkaline extracts without isolation of the hemicelluloses from the solution by HPSEC on a Separon S Hema 1000 column using 0.5 M NaOH as eluent. The calibration graphs are obtained with known concentrations of the corresponding hemicelluloses versus chromatographic peak area. Hemicellulose content obtained by HPSEC is in good agreement with the hemicellulose content determined by standard analytical methods [238]. a.OtherPreparations of Hemicelluloses. Anothercommonmethod for determination of hemicelluloses in woodand/orpulps isthe alkali solubility ( 1 % sodiumhydroxide solubility; alkali solubility of pulp at 25°C). Wood or pulp is extracted with a hot 1% sodium hydroxide solution for 1 h. The loss in weight is determined and calculated as percent of solubility. Some readily soluble polyoses are extracted together with degraded cellulose (TAPPI Standard T212-om-88 [2]; ASTM D 1109-56 [3]). This value could indicate the degree of fungal decay or the degradation by heat, light, oxidation, etc. The solubility of pulp indicates the extent of cellulose degradation during pulping and bleaching and has been related to strength and other properties of pulp [239]. Another important value for bleached and delignified pulps is the alkali solubility in IO%, 18%, and 21.5% NaOH (S,,,, SIX, and S21 under defined conditions (TAPPI T 235 cm-85 [2]; DIN 54356 [7]; I S 0 Standard 692-1974 [6]). Pulp is extracted with sodium hydroxide solution of IO%, 18%, and 21.5% at 25°C for 1 hr. The dissolved carbohydrates are determined by oxidation with potassium dichromate. The solubilities of a pulp in alkali provide information about the degradation of cellulose and on a loss or retention of hemicelluloses during pulping and bleaching processes.

2. Characterization of Hemicelluloses Once a hemicellulose has been purified, its structure can be determined. The sample is characterized by determination of constituent sugarsandsugar acids, specific rotation, linkage and sequences, functional groups, molecular weight, and MWD. a. Am1y.si.s of Wood Curbohydrates. In the structural analysis of a hemicellulose it is necessary first to determine the kind and quantity of glycosyl units present. Hemicelluloses are commonly hydrolyzed by mineral acids, formic acid [24O], or trifluoroacetic acid[241,242].Thesubsequentdeterminations of the componentmonosaccharidesare normally carried out by chromatographic techniques. Methodologies for analysis of wood sugars have undergone rapid advances in recent years. Older methods applied to determine carbohydrate composition of wood and pulp are paper or thin-layer chromatography. These techniques were replaced by GC following derivatization, which offers the advantage that it can easily be connected to MS. CationexchangeHPLCcoupledwith refractive indexdetectionwasfound tobe simplerand faster. Recently,anion-exchangeHPLCwithpulsedamperometricdetection(PAD)was found to have the advantages of greater sensibility and easier sample preparation. Also, capillary electrophoresis has been shown to be a valuable tool in analyzing carbohydrates.

racterization Chemical

of Wood

305

A useful source of references for analytical protocols for carbohydrates is given in Cnrhohyclrnte Atzdysis, n Prtrcticul Approoclz, edited by Chaplin and Kennedy [243]. A review summarizing advances for the broader aspects of anion-exchange chromatography of carbohydrates was presented by Lee 12441. The determination of the sugar units comprising a polysaccharide involves hydrolysis. This is carried out by acid or enzymatic reactions. The application of acids, particularly sulfuric acid, is usual [245-247). In general, in a procedure described by Saeman et al. [246], the sample is treated with 72% sulfuric acid solution for 1 h at 30°C and then with4% sulfuric acid for1h at 120°C. The useof77%sulfuricacid in the primary hydrolysis 12481 andsecondaryhydrolysisunder reflux in 3% sulfuric acid [249]have been proposed. Acid Hydrolysis. The splitting rate varies between different glycosidic bonds. The use of sulfuric acid as a catalyst in the hydrolysis of polysaccharides under the conditions of total hydrolysis of cellulose to glucose may destroy some sugars, especially pentoses and uronic acids. The reaction of monosaccharides with the acid decreases their yields. On the other hand, some linkages in xylan, namely, those between 4-0-methylglucuronic acid and xylose, are quite resistant to acid hydrolysis and may survive conventional acid hydrolysis. The loss during hydrolysis varies among investigators using similar procedures [246,250]. A possible source of variation may be the temperature of the secondary hydrolysis, which is carried out in an autoclave. The sugar loss may be less by using 3% H,SO, (w/w) in the secondary hydrolysis 12491. The optimum conditions for acid-catalyzed hydrolysis differ from sample to sample, and also the optima reaction conditions for one carbohydratecomponentdo not give the optimum yield for anotherone in the same sample. Therefore, it is necessary to optimize the conditions of hydrolysis for each sample. Trifluoroacetic acid (TFA) (pK,, = 0.23) has been proposed as a better reagent for the hydrolysis of polysaccharides than sulfuric acid [241,242]. TFA preserves the monosaccharides produced during the hydrolysis and the acid can be completely removed by evaporation. TFAis a good cellulose solvent, which makes it possible to hydrolyze cellulose in solution. Several methods have been developed depending on the sample to be hydrolyzed; i.e., the hydrolysiswith TFA can be welladjustedto the specificstarting material [25 I ] . Hemicelluloses can be hydrolyzed with diluted TFA by refluxing in 2N/ TFA for 1 h. Cellulose, pulp, and wood require treatment with concentrated TFA in homogeneous solution, and the procedure depends on the lignin content [252]. TFA works well with soluble polysaccharides, but there is a risk of incomplete hydrolysis in the case of lignified materials. EnzymaticHydrolysis. Analytical acid hydrolysis is alwaysacompromisebetween incomplete hydrolysis or sugar destruction. For this reason enzymatic hydrolysis, which is carried out under mild conditions. is an alternative. By using suitable mixtures of cellulases and/or hemicellulases, pulps, holocelluloses, and hemicelluloses can easily be hydrolyzed. However, enzymes are unable to degrade native wood efficiently. Tenkanen et al. 12531 developed an enzyme-aided analytical method for chemical characterization of kraft pulp. A mixture of commercial enzyme preparations containing cellulolytic and hemicellulolytic enzymes was used. The anlounts of sugar analyzed after enzymatic hydrolysiswasalmostequal to that after acid hydrolysis. The main advantage of using enzymes is that it enables the quantitative analysis of acid-labile sugars such as hexenuronic acid, which are destroycd in acid hydrolysis. Rydlung and Dahlman [254] reported capillary zone electrophoresis analysis of chemically (TFA) and enzymatically (xylanase) hydrolyzed spruce wood xylan. Samples of enzymatically dissolved and partially degraded

306

Baeza and Freer

xylan from unbleached hardwood and softwood were characterized with respect to their content of mono- and oligosaccharides. Carbohydrate Determination. TAPPITestMethodT249cm-85[2] or other equivalent methods (e.g., ASTM D 1915 [3]) are accepted procedures for the analysis of wood and pulp carbohydrates. The determination of the carbohydrate composition is carried out by gas chromatography of the carbohydrates as alditol acetates [255,256]. The TAPPI method involves hydrolysis with sulfuric acid of extractive-free samples using a two-step technique. Inositol is added as internal standard and the solution is neutralized with Ba(OH),, and then the sugars are reduced to alditols with NaBH,. The alditols are acetylated with acetic anhydride using sulfuric acid as the catalyst, and the alditol acetates are precipitated in ice water and extracted with methylene chloride and analyzed by gas chromatography. The analysis by GCusing the alditol acetates is time-consumingand involvesnumeroussteps,andeachstep requires skilled care if the procedure is to be quantitative. The alternative trimethylsilyl or trimethylsilyloxime derivatives are easier to prepare [257], but the chromatography is complicated by isomerization [258]. Liquid chromatography has been also used for carbohydrate analysis and offers the advantage of not requiring derivatization to chromatography. Several low-pressure liquid chromatography systems have been developed for analysis of carbohydrates [259-2621. HPLC has been widely used for the analysis of carbohydrates. Various methods, including aminoalkyl bonded silica [263-2651, reverse-phase C,, [266], and ion-exchange columns [249,252,267-2691, have been reported. Ligand exchange on calcium and lead cation-exchange columns [267] coupled with refractive index detection and water as eluent have been widely used for analysis of wood sugars. The sugar content in pulp mill effluents, TFA-hydrolyzed pulp and wood samples, by using lead-based columns and refractive index detection, was reported by Paice et al. [252]. The nature of the sugar-metal ion interaction was reviewed by Angyal [269]. The monosaccharides commonly found in wood and wood pulp hydrolyzates can be separated by using anion-exchange resins. The separation is based on the weakly acidic properties of carbohydrates in alkaline conditions. At high pH, neutral and acidic carbohydrates are partially or completelyionizedand thus retained on the column.Column selectivity is changed using a NaOH gradient. Anion-exchange HPLC (HPAE) with pulsed amperometric detection (PAD) was found to have the advantages of greater sensitivity and easier sample preparation [270]. PAD is based on oxidation of the carbohydrates in multistep potential waveforms (E,) applied to Au electrodes (sensitivity for PDA of carbohydrates is larger at Au than Pt electrodes [27 11) in miniature flow-through cells. During the first potential, E,, the signal from carbohydrates is measured, and the second and third potentials (E? and E3) correspond to oxidative and reductive cleaning pulses, respectively, and permit the maintenance of the electrode activity. This technique is applicable to monosaccharides, disaccharides, and oligosaccharides. The analysis of wood and wood pulp hydrolyzates by HPAE-PAD in combination with acid or enzymatic hydrolysis has been carried out by various authors 1253,270,2722781. HPAE-PAD provides rapid and versatile methods for carbohydrate analysis. It is a sensitive method of high selectivity and specificity. Sample preparation is minimal, without neutralization or monosaccharidesderivatization;onlygroups that containoxidizable groups are detected. It is possible to effectively separate and analyze quantitatively trace amounts of monosaccharides released from hemicellulose hydrolysis in the presence of as much as 95 times the amount of D-glucose from cellulose hydrolysis. HPAE-PAD chromatograms are shown in Fig. 9.

307

Chemical Characterizationof Wood 5

1. Fucosa Deoxyribose 2. 25 3. Arabinose 25 . Galactose 25 . Glucose 25 16. Xylose 25

3

1A

8

J

5

0

15

10

TIME.rnin

L

20

(b)

1. Arabinose

2. Galactose 3

1

I 0

155

I 10

5

I

3. Glucose 4. Xylose 5.Mannose

I 20

TIME, rnin

FIGURE9 HPAE-PAD chromatograms: (a) common monosaccharides; (b) hydrolyzed wood pulp. (From Ref. 274.)

The degradation of sugar during the acid hydrolysis and the effects of other parameters of sample preparation on the measurement of wood sugars by anion-exchange HPLC using PAD were investigated 12771. Unhydrolyzed and hydrolyzed sugar standards, and wood meal (Pinus t m d a ) and wood meal spiked with a standard before hydrolysis were analyzed. Loss of sugar in hydrolyzed standard ranged from 6.4% for arabinose to lS.796 for mannose. In wood samples to which a standard was added before hydrolysis, the results were very close to the sum of wood samples and hydrolyzed standard analyzed separately, except in the case of galactose, which was about 4.7 lower i n the combined sample. These results indicate that the effect of hydrolysis on sugars is probably the same in wood. Thus, the other wood components apparently do not protect sugars or enhance their degradation during the hydrolysis. Threechrornatographic methods for the quantitativedetermination of wood sugar were compared for the ease of their operation and accuracy [270]. They were: ( I ) borate complex anion-exchange chrolllatography-reducing sugars were detected by postcolumn derivatizntion with Cu-bicinchoninLlte (0.33 mL/min) measuring the absorbance at 546 n m ;

Baeza and Freer

308

(2) anion-exchangechromatography in NaOHmedium-PADdetector;and (3) high-performance thin-layer chromatography (HPTLC). Lowest standard deviation was found for borate-HPLC. It seems the borate-HPLC is the most reliable method, although it is seldom usedandseems to havebeenreplaced by NaOH-HPLC. The authors consider that the accuracy of the latter method is overestimated, and that of the HPTLC is underestimated. They concluded that the three chromatographic methods evaluated need improvements in reproducibility. Capillary zone electrophoresis (CZE) has been found to be a useful technique for the determination of neutral and acidic wood-derived oligosaccharides. A basic scheme of this technique and an electropherogram of a mixture of 6-aminoquinoline derivatives of monosaccharides are shown in Fig. 10. This technique in general is fast and offers high sensitivity. Capillary electrophoresis is performed in buffer-filled capillaries by applying a high voltage over the ends of capillary. The separation is obtained by a differential migration of the charged species. By using alkaline borate as a buffer solution, neutral saccharides become negatively chargedby formation of complex with borate. The application of CZE to the analysis of wood hemicelluloses of different origin, in combination with chemical or enzymatic hydrolysis of the samples, was reported by Rydlund and Dahlman [254]. The methodused comprises the following steps: ( I ) derivatization of the saccharides with an UV-derivatization reagent (6-aminoquinoline), (2) CZE separation of the resulting derivatives in an alkaline borate buffer system, and (3) highly sensitive UV detection and quantification of the derivatizedsaccharides.Uronic acids and acidic oligosaccharides containing 4-0-methylglucuronic acid units were also separated using the same buffers system.

Buffer vials

[O - 30 kV1

V

7

9

i3

1'5

Min

FIGURE 10 CZE technique: (a) basic scheme of a CZE instrunlent: ( b ) electrophcrograln of 6aminoquinoline derivatives o f monosaccharidcs. (From Ref. 254.)

racterization Chemical

of Wood

309

The determination of the component monosaccharides in hemicelluloses by precolumn derivatization of carbohydrates with p-aminobenzoic acid and the separation by borate complexation by means of CZE have been carried out by Huber et al. [280]. On-column UV monitoring at 285 nm allowed the detection of aldoses and uronic acid in the lower femtomole range. Other derivatizing agentsthat have been used are ethyl p-aminobenzoate and p-aminoabenzonitrile. Uronic Acids. Uronicacid units are not readily hydrolyzedandremainlinked to anothersugar as an aldobiuromic acid. The mostcommonone is 2-0-(4-0-methyl-~glucopyranosyluronic acid)-D-xylose. In softwoods the uronic acid-xylose ratio is about 1:5, while in hardwoods 1: 10 [66]. Little galactouronic acid would be expected in holocellulose preparation, because it is present in pectic material which is removed in the delignification. The determination of uronic acid in a wood sample or isolated hemicellulose is based on decarboxylation with strong mineral acid [281,282]. After removalof furfural and water, by passing a stream of nitrogen carrier through suitable traps, the evolved CO, is adsorbed in a matrix such as ascarite or Ba(OH), solution. The CO, is determined by the weight gain of the ascarite or by conductivity versus time measurementof the solution of Ba(OH),. The CO, amount, after correction for nonuronic CO?, represents the uronic acid content in the sample. Colorimetric determinations have been widely used to determine hexuronic acids. The carbazole sulfuric acid method, developed by Dische [283], involves the reaction of carbazole with the unstable acid-hydrolyzed derivatives of hexuronic acids. It takes about 2 h to develop full color; when using harmine the color is developed immediately, but it gives anomalously high results [284]. Modification of the carbazole sulfuric acid method of Dische has been proposed [285]. Scott [286] developed a method for determination of hexuronic acid by using 3,5-dimethylphenol as a selective reagent for 5-formyl-2-furancarboxilic acid (5FF), which is formedfrom the reaction of several uronicacids with concentrated sulfuric acid. Ground wood samples moistened with ethanol are dissolved in 72% sulfuric acid by heating at 50°C for I O min, and then diluted with water to a volume estimated to contain 20-80 pg/mL. A 0.125-mL sample of this solution is mixed with the same volume of 2% NaCl and 2 mL of concentrated sulfuric acid are added. The mixture is heated at 70°C for I O min and then cooled to room temperature, and then 0.1 mL of 3,5-dimethylphenol solution (0.1% p/v in glacial acetic acid) is added. After 10-15 min, the absorbance is read at 450 and 400 nm against water as a reference. Galacturonic acid is used as a standard. D-Galacturonic and 4-O-methy~-~-ghcuronic acids could be measured separately from D-g~ucuronic acid by adding H,BO,. Uronic acids in the hydrolyzates can be analyzed by paper or thin-layer chromatography, by ion-exchange chromatography, or by gas-liquid chromatography. For GC analysis the uronic acid has to be converted into volatile derivatives, such as trimethylsilyl esters, alditoacetates, or aldonitriles acetates [287]. Analyses of uronic acids havebeen carried out by HPLC.Anion-exchangechromatography systems have been developed [288,289]. HPLC analysis of uronic acid and oligogalacturonic acids after enzymatic hydrolysis was studied by Voragen et al. [290]. A different approach, when uronic acid units are present, is to reduce them to glycosyl units before hydrolysis of the polysaccharide. This can be done by treating the polysaccharide with a water-soluble carbodiimide and then with sodium borohydride [2911. The reduction of carboxyl groups of uronic acid not only facilitates the hydrolysis but also can be used to identify the type of uronic acid and the type of glycosidic bond [292].

310

Baeza and Freer

Pentosans. Pentosans content corresponds to the total amount of pentosans without the determination of the individual sugar components. An effective pentosan determination should be obtained by the sum of the amount of pentose sugar determined by chromatographic techniques in the hydrolyzate. A rapid and simple method for determining total pentosans is based on the conversion of pentosan into furfural by boiling them in 3.85 M HCI. Furfural is collected in the distillateanddetermined by gravimetric,volumetric, colorimetric, or spectrometric methods [2931. Direct determination of pentosans in a water extract is described by Smeltorius and Stewart 12051. Details of the procedure are given in TAPPI Standard T223 cm-84 [2] and ASTM Standard D 1787 [3]. AcetylGroups. The principal acyl groups in wood are the acetyls, andmuch smaller amounts of formyl groups. The amount of acetyl groups usually are determined by transesterification reactions. The acetyl group is converted to the corresponding ester, normally methyl acetate, which is removed by distillation and the amount is determined by the alkali consumed in the saponification. Details of the methods employedto determine acetyl groups are given by Browning 12941. Direct determination of the content of acetyl groups in water-extractedwoodhasbeendescribed by SmelstoriusandStewart [205]. Acetyl content has been determined by GC by measuring the acetic acid content in wood hydrolyzed. b. Linkages and Sequence. The natureof the linkagesbetween the units and the sequence in the hemicelluloses is determined by proceduresinvolvingmethylation,periodate oxidation, partial hydrolysis, acetolysis, degradation by p-elimination, and analysis by nuclear magnetic resonance. Methylation. Briefly, methylation analysis involvescompletemethylation of the hemicellulose sample 12951, following by hydrolysis of the methylated product, and subsequent analysis for GC and MS, after derivatization to the partially methylated alditolacetate or in some cases to peracetylated aldonitriles [296]. With some polysaccharides it may be necessary to use special techniques of hydrolysis, since the methylated derivatives are not soluble in mineral acids. A preliminary short hydrolysis of the methylated polysaccharide in 90%formicacid to disrupt aggregates of the methylatedpolysaccharides yields aproductwhich is more easily hydrolyzed by mineralacids.Thefreehydroxyl groups in the methylated sugars represent the positions at which the sugars were involved in glycosidicbonds in the polymer. If onlyone free hydroxylgroup is present in the methylatedmonosaccharide, it musthavebeenanonreducingterminalglycoside unit. Methylated sugars with more than one free hydroxyl group must have been linked within the chain structure or served as the reducing end unit. A methylated sugar with three free hydroxyl groups must be from a branch point in the polysaccharide. Methylation analysis does not give information about the anomeric configuration of the glycosidic linkages nor about the sequence of the monosaccharidesresidues in the polysaccharide. The latter determination must be done by other methods. For example, partial hydrolysis of totally methylated polysaccharides provides information about the position at which the linkage occurred i n the original molecule. The classical methylation procedures involves treatment of the polysaccharide with methyl sulfate and potassium hydroxide [297.298] or with methyl iodide and silver oxide [299-30l], which have been replaced by the Hakomori procedure 12951. In this method, the polysaccharide is treated with sodium lnethylsulfillyllnethide (dimisyl sodium), which is generatedfromsodiumhydride in DMSO,followed by reaction with methyl iodide. Procedures giving improved yield of methylated products and cleaner products have been reported [302,303] in which dimisyl potassium (prepared either from potassiumhydride or potassium t-butoxide) instead of dimisyl sodium is used. The most useful criterion for

Characterization Chemical

of Wood

311

the complete methylation is the absence of a hydroxyl peak near 3500 cm" in the IR spectrum. Details of the Hakomori procedure are given elsewhere [304,3051. Modifications to the Hakomori method have been proposed [306,307]. Some polysaccharides that show strong resistance to permethylation by the Hakomori procedure may be completely methylated with a mixture of DMSO and 1,1,3,3-tetramethylurea in a 1 :1 ratio 13071. If uronic acids are present, the complete permethylation is more difficult and may yield secondary products. Methylation is facilitated by reducing the carboxyl groups of the original polysaccharide or the fully methylated polysaccharides to hydroxyl groups. It has been recommended the following steps: first an initial methylation, forming a uronic acid methyl ester, then the reduction of this ester with lithium aluminum hydride to a hydroxyl group, and finally a remethylation [305]. In the case of alkali-labile polysaccharides it is advisable to prereduce with borohydride or borodeuteride to avoid "peeling" reactions [308]. The methylation techniques have been reviewed by Rauvala et al. 13081. Periodate Oxidation. Periodate oxidation of polysaccharides is usedin the structural characterization andmonosaccharidesequencedetermination of polysaccharides [309]. Carbohydrate residues having glycol groups are oxidized to dialdehydes (Fig. 1 l ) , but if the residues contain hydroxyl groups in three adjacent carbons, formic acid can also be produced. In addition to glycol units, oxidation occurs in a-hydroxyaldehydes, a-hydroxyketones, and a-hydroxyhemiacetals. Formaldehyde is produced from the oxidation of primary hydroxyl groups, and formic acid and aldehyde or carbinol units between two other hydroxyl-bearing carbons [3 lo]. The oxidation is conducted in a dilute solution of sodium metaperiodate in the dark at low temperature and at pH 3-3.5. The periodate consumed, formaldehyde and formic acid produced, and the characterization of the oxidized polymer provide information on the molecular structure, nature of the end groups, and point of linkages between the constituents. Additional information may be obtained by applying the Smith degradation sequence,whichcomprisesaborohydridereductionof the periodateoxidationproducts followed by a mild acid hydrolysis of the polymer (Fig. 12) 131 l]. Normally, when a sugar residue of a polysaccharide is cleaved by periodate and reduced, it corresponds to an acetal which is acid-sensitive. When a sugar unit that survives oxidation is joined to a cleaved unit, it appears as a glycoside that is relatively stable to acid. Due to the difference of stability between true acetals and glycosides, the products of periodate oxidation reduction, and mild hydrolysis are low-molecular-weight glycosides, glycoaldehyde, and polyalcohol. Analysis by GUMS after conversion to volatile derivatives provides information on the structure of the parent polysaccharide 131l ] .

'l0dH

CH,OH

'lot

CH,OH

NalO,

CH0 OHC OH

FIGURE 11 Pcriodatc oxidation.

COH

31 2

'lot

Baeza and Freer

CH,OH

CH,OH o j R z

oO j Rz

' l o <

NaBH, ____)

CH0 OHC

CH,OH HOCH,

CH,OH mild acid

1

HCOH l

hydrolysis

+

HCOR,

CH0 I

+

R,OH

+

R,OH

CH,OH

I

CH,OH

CH,OH acid

l

HCOH I

hydrolysis

I

+

CH0 I

CH,OH

CH,OH

FIGURE 12 Smithdegradation.

Lead tetraacetate also cleaves glycols, similarly to periodates. The reaction is usually conducted in acetic acid because lead tetraacetate decomposes in water. Methyl sulfoxide can be also used as solvent [3 121. The formic acid produced in the reaction is further oxidized to carbon dioxide by lead tetraacetate. Lead acetate has been used to differentiate between cis and trans hydroxyl groups in glucomannan [313]. Another analytical scheme for identifying periodate oxidation products of polysaccharides is knownas the Barrydegradationandhasbeendescribed by Lindberget al. [314]. In the Barry degradation the periodate-oxidized products are reacted with phenylhydrazine in dilute acids, followed by hydrolysis to release the phenylhydrazone-containing units. Some polysaccharides can be degraded by the Barry method with a sequential removal of monosaccharides residues. Then the sequence of carbohydrates residues in the polysaccharide chain can be determined. Partial Hydrolysis. Partial hydrolysis, by either acids or enzymes, gives a mixture of monosaccharidesandoligosaccharides.Glycosidicchainlinkagescanbeestablished through the carbohydratefragments.Rates of monosaccharideproductionalsoprovide information on the location of single sugar units as branches and the nature of their ring forms. The rate constants for the acid hydrolysis of the glycosidic bond in polysaccharides vary greatly depending on the monosaccharide composition, positions and configuration of the glycosidic linkages, and the ring form. This different resistance of the glycosidic linkage can be advantageous. For example, the resistance of the glycosidic linkages of the uronic acids makes it possible to obtain, by acid hydrolysis, a good yield of aldobiouronic acids (separated easily) and its glycosidic linkage can be determined after reduction of the carboxyl group to produce the neutral polysaccharide, which can be analyzed. Uronic acid

Characterization Chemical

of Wood

313

residues may be introduced in some polysaccharides by oxidation of the primary hydroxyl groups with oxygen in the presence of platinum. It is also possible to introduce acid-labile linkages in polysaccharides by chemical modifications. Partial hydrolysis of a fully methylated polysaccharide and the investigation of the products, even when the hydrolysis is not very selective, give valuable information [315]. In the procedure proposed by Albersheim et al. [316,3171, the glycosyl sequence analysis is determined as follows: the carbohydrates are methylated by a modification of the Hakomori procedure and, by partial hydrolysis, converted into a complex mixture of partially methylated oligosaccharides. The mixture is reduced with sodium borodeuteride, and free hydroxyl groups are labeled by ethylation. The alkylated mixtureis fractionated by reversephase LC on anactadecylcolumn.Theanomericconfiguration of the glycosyl of the fractionated peralkylatedoligosaccharide is determined by 'H-NMRspectroscopy. The oligosaccharides in the different fractions are analyzed by GUMS after conversion to the alditol-acetal derivatives. The information so obtained is pieced together to determine the structure of the carbohydrate. Theenzymaticmethodcan be coupledwithacidhydrolysistechniquetoobtain additional structural information. Enzymatic hydrolysis gives valuable information on the structure and residue sequences of the fragments and the polysaccharide. Acetolysis. Acetolysis of polysaccharides results in the complete acetylation of free hydroxyl groups of the polysaccharide and the selective cleavage of the glycosidic bonds. Acetolysis involves treatment of the polysaccharide in either acetic anhydride or with a mixture of acetic anhydride, acetic acid, and sulfuric acid, generally in the ratio 10: IO: 1 . In this mixture the carbohydrates are first acetylated and then hydrolyzed. Due to differences in the rate constants for the cleavage of glycosidic bonds, the fragmentation of the polysaccharide occurs preferentially at specific glycosidic bonds ( 3141. Acetolysis of polysaccharides can be used as a complementary method to conventional acid hydrolysis. &Elimination. Carbohydrate residues with substituent groups in the p-position to electron-withdrawing groups, such as carbonyl or carboxylic ester, undergo p-elimination reactions upon treatmentwith a base[215,314].The substituent groups,whichcanbe eliminated, include alkoxyl, hydroxyl, and glycosyl groups. The presence of a hydrogen atom in the a-position to the withdrawing groups is necessary. The p-elimination reactions have been especially useful in the structural determination of carbohydrates, particularly those containing uronic acid residues. The carboxyl group is methylated and then treated with base, resulting in elimination of the substituent at C4, leaving an acid-labile, unsaturateduronic acid 13181. Subsequently,mild acid hydrolysisdegrades the uronicacid residue and releases the substituent at C l . Both types of moieties, from C4 and C l , can be isolated andused for further structural analysis. The remainingpolysaccharides are methylated and identified by GC/MS. This analysis permits the determination of the position of attachment of the original uronic acid. Applications of p-elimination to the structural determination of polysaccharides have been reviewed by Lindberg et al. [314]. Nuclear Magnetic Resonance. 'H and "C-NMR spectroscopy are invaluable techniques for primary structure determination and conformational analysis of polysaccharides. Reviews of NMR analysis of polysaccharides have appeared [319,320]. 'H-NMR has been used for quantitative estimation of specific functional groups such as 0-mcthyl and 0-acetyl substituents and for identifying anomcric configuration of the glycosidic linkages i n polysaccharides 1321,3221. IZC-NMR provides information on anomeric configuration but also on other aspects of polysaccharide structure such as monosacharide composition, the monosaccharide sequence, and the conformation of the polysaccharide [323,324].

314

Freer

and

Baeza

c. Deternlinatiort of Molecular Weight and Molecular Weight Distribution. Hemicellulosesgenerallyoccurwith a normal distribution of molecularweight.Hemicelluloses present low DP values compared to those of cellulose.Hemicellulosesfrom hardwoods have DPs of 150-200, and softwoods have DPs of 50-300. The methods of determination of molecular weight and MWD were discussed above. The DP of neutral polysaccharides may be calculated by using the method developed by Yamaguchi et al. 1325-3271 and improved by Tanaka [328]. The DP is calculated from the ratio of monosaccharides to the alditol, after hydrolysis of the polysaccharide in which the reducing end residue is reduced to an alditol residue. The numbermolecularweights of xylansamplesweredetermined by isothermal distillation [218]. Benzene was used as the solvent for the methylated derivatives, and 1,3dioxanwasused for the acetylatedpolysaccharides. This technique,described by Gee 13291, is basedon the rate of distillation of the solvent in the solution. This dynamic method requires calibration. The straightforward way to determine molecular weight of hemicelluloses is by viscometry in anappropriatesolvent.Cuenhasbeencommonlyusedas the solvent for viscometry determination of hemicelluloses 1330-3351. Sears et al. 13311 reported cuen DP data on different hemicellulose fractions separated from holocellulose, sulfite paper pulp, and kraft paper pulp from black spruce and balsamfir. The [v] values were converted into DP using the following equations: for xylans, DP, = 208[77]"'" and for glucomannans and galactoglucomannans, DP, = 3 5 9 [ ~ ] ' . ~ ' ~ DMSO has been shown to bea good solvent for hemicelluloses. Determinationof viscosity in DMSO-water (4:l) containing 0.05 MNaCl of acidic glucanhasbeen carried out 13361. DP values were also obtained by determining [v] of the nitrate derivatives in ethyl acetate at 25°C by the formula DP = 75[7]. The viscometric nitrate DP value scale was based on a calibration curve made with standards of cellulose oligomer saccharides of varying sizes [ 118,331-3361. Osmometry molecular weights have been obtained for both derivatized and nonderivatized hemicelluloses. The solvents used for derivatized hemicelluloses were acetone and n-butyl acetate (nitrate) [336], toluene (methyl) [336], and chloroform-ethanol (9:l) (methyl and acetate) [201,337]. Water-soluble hemicelluloses are run without derivatization; water-DMF (20:80) 13381 and 0.2 M aqueous NaCl solutions 13391 were used as the solvents. GPC analysis of hemicelluloses have been performed on soft gel column packing, which cannot withstand high pressures. Cross-linked dextran resins and water or sodium acetate buffers as eluent have been used to determinate the MWD of xyloglucan 13401, arabinogalactan [338,34 1 I, and 4-0-methylglucuronoxylan [ 341 1. Cross-linked polyamide packing columns have been used for water-soluble galactoglucomannan from loblolly pine 13421 and arabinogalactan from acacia 13431 and larch 13441. Studies of the HPSEC elution behavior of nitrate derivatives of hemicelluloses isolated from holocellulose, sulfite paper pulp, and kraft paperpulp of blackspruce and balsam fir were carried out I33 l ] . THF was used as the solvent, and four columns packed with Styragel (pore sizes of 1 x IO", 3 X 1 Of', I X IO5, and 3 X 10'' A) were used. The calibration was done with cellulose oligomer saccharides (cellobiosc. cellotriose, etc.),and

Characterization Chemical

of Wood

315

therefore the results are suitable only for comparative purposes and do not represent the true molecularity. The elution behavior of 4-O-methylglucuronoxylan,isolated from birch, has been studied using a Separon HEMA 1000 column and DMSO and DMFA as the mobile phases [345].The polyelectrolyte effectsweresuppressed by the addition of acetic acid and lithium bromide (0.03 M). Fractions of the same polymer, characterized by viscometry in cadoxen, were used for column calibration. Thismethodproved useful to estimate the MWD of xylans rapidly. HPSEC for the determination of MWD of wood hemicelluloses, using the same type of column with 0.5 M NaOH as eluent, has been investigated by Eremeeva and Bykova [238]. Nonexclusioneffects were observed on the Separon S HEMA 1000 column in 0.5 M NaOH. In this method, the analysis of the alkaline extracts from wood pulps was performed without previousisolation of the hemicelluloses of the solution, shortening the analysis timeandminimizing the changes in the hemicelluloses due to sample preparation. Only extraction with basic solutions and filtration are required. Both a series of dextrans and xylan fractions were used to calibrate the column. It was shown that the universal calibration between dextran and xylans is valid.

VI.

LIGNIN

Different techniques have been reported to characterize lignins. Lignin is the second most abundant biopolymer. It consists of an aromatic system composedof phenyl propane units. Figure 13 shows a typical representation of a lignin model [346]. An important characteristic of this natural polymer is the presence of different functionalgroupssuch as phenolics,methoxyls, aliphatic alcohols, aldehydes, ketones, and ethers [347]. This last group together with carbon-carbon bonds are present in the formation of the polymeric bonds. The frequency of thesegroupsand the molecularweight are important in the characterization of the lignin. The lignin composition depends on different factors. Lignins from different plants differ appreciably in their structure, and it is possible to distinguish, for example, softwood lignin (“guaiacyl lignin” with coniferyl alcohol as main polymer unit) and hardwood lignin (“guaiacyl-syringyl lignin” with coniferyl and sinapyl alcohols as main copolymer units). Morphological locations from which lignin is isolated are also important, as shown in Table 6, in which the guaiacyl-syringyl ratios and the distribution in lignins in white birch, determined by UV-EDXA technique, are given [348]. Another important factor is the method used to isolate the lignin, due to the appreciable changes that occur during the isolation procedure. The chemical structure of lignin affects the reactivity ofwood during industrial processes, such as pulping and bleaching, and also the final fiber and the possible new applications of technical lignins. For this reason, the characterization of lignin has been an important concern and many research groups are working in the development of different analytical techniques to obtain more structural information. A large number of papers has been the result of this activity.

A.

Isolation of Lignin

The analysis may be carried out using lignins isolated by different methods or directly from wood or pulp, without previous isolation. The method used in the isolation of the lignin is a very important aspect to be considered from the analytical point of view, and

Baeza and Freer

316 H2COtI

4 I

HC-O-

l

ICH2OHI H3CO

I

HC-OH2COH H2:OH

HC=O

I

I

I

HC

HCOH

4 II

H$O

H 3 C 0 p " '

H2COH I

0-CH

HOH2C"C"C=O

OCH3

I

H~;N'\CH

HC-0

I

HCOH

0-

l

OCH3 OCH,

OCH3

H,CO

OH

OH [U-C1

FIGURE 13 Structural model of spruce lignin. (From Ref. 346.)

for that reason it is necessary to specifythe type of lignin.The different lignins from wood are: Milled wood lignin (MWL),orBjorkmanlignin, isolated with tolueneafter ball milling 13491. I t has been postulated that it is a lignin with a chemical structure close to native lignin. Brauns native lignin (BNL), isolated by extraction with ethanol, following by precipitation with ether [350]. Klason lignin. isolated from extracted wood meal after treating with cold concentrated sulfuric acid (TAPPI Standard Method T-222 om 8 3 ) 121. Cellulolytic-enzyme lignin (CEL), isolated from ground wood, using an enzymatic hydrolysis [ 35 1,3521. This is a slow procedure, but the lignin suffers less structural alteration than MWL lignin. Swelled-enzyme lignin (SEL), in which wood nleal is swelled in an organic solvent before treatment with a cellulase 13531. Organosolv lignin, extracted with different solvents or as side product in an organosolv pulping process [354-3581.

317

Chemical Characterizationof Wood TABLE 6 Ratios of Guaiacyl and Syringyl Residues and White Birch [348]

Distribution in Lignins in

~~~

Element

Tissue Morphological region

S,

Fiber

14

12:88S?

Vessel

Ray Parenchyma

S, ML ML,,(f/f)* S, S? S, ML MLcc(f/v)h S

Guaiacyl-syringyl ratio

volume (%)

11.4 58.4 3.5 5.2 2.4 l .6 0.26 4.3 2.3

-0 -0

0.14

-

0.12 0.36 0.45 0.26

-

88: 12 -

80:20 49:5 1

2.0 ML,,(f/r)‘ MLcc(r/r)d

-

91:9

0.8 -0 8.0

-

1oo:o 0.4

Lignin concentration (g/&

88: 12

0.27 0.40 0.58

0.12 0.38 0.47 1

“Fiberltiber. hFiber/vessel. ‘Fiberhay. “Raylray.

Kraft lignin, obtained during the kraft (sulfate) process. This is the most abundant industrial lignin [359]. Lignosulfonate, obtained in a sulfite pulping procedure [360]. Details of lignin isolation procedureshavebeendescribed by Browning [361]. A critical review of methods of isolation of lignin has been published by Lai and Sarkanen [362]. The subject has been also discussed by Lundquist [363], Fengel and Wegener [364], Brauns [365], and Brauns and Brauns [366].

B.

Characterization of Lignin

The chemical characterization of lignins is carried out using: Degradative procedures, in which the lignin is depolymerized through chemical reactions, followed by the identification of the low-molecular-weight degradation products. Nondegradativeprocedures, in which the polymer is characterizedwithoutdegradation. Spectroscopicmethods are the mostcommonnondegradative procedures.

1. Degradative Procedures a. Acid~l~~.~i.~-Etlranol~~.~i.~. In acidolysis, the lignin is refluxed i n a mixtureofdioxane-hydrochloric acid (0.2 N) 9:l (v/v). The a- and P-aryl ether linkages are cleavage due to protonation of the oxygen with the formation of a good leaving group. After com-

318

OH

Baeza and Freer

OH

OH

FIGURE 14 Monomericphenolsformedduringacidolysis-ethanolysis CH,CH,). (From Ref. 367.)

( R , = Hor OCH3; Rz =

plete degradation, a mixture of monomeric phenols (Hibbert's ketones) is detected (Fig. 14), which are characteristic of the arylglycerol-p-aryl ether structure of lignin. The softwood lignins generate the guaiacylpropane monomers (R, = H). In addition, the hardwood lignins generate the syringyl monomers (R, = OCH,). If the hydrolysis occurs in ethanol as a solvent, the ethoxy group will be present (Rz = CH,CH,). These compounds may be analyzed by GC as trimethylsilyl derivatives [367-3701. b. Thioaciddysis. Thioacidolysis is an efficient proceduretocleave arylglycerolP-aryl ether bonds in lignins. The solvolysis is carried out in dioxane-ethanethiol (9:l v/v), 0.2 M boron trifluoride etherate. The mixture of recoveredcompounds is mainly thioethyl phenylpropane adducts (TPP) [371], which provide information about the arylglycerol aryl ether units (Fig. 15). The monomers are separatedusingchromatographic techniques. The characterization is done by using UV, IR, 'H-NMR, "C-NMR, and highresolution MS after the monomers are converted into trimethylsilyl derivatives. A more recently modified procedure includes a second step of desulfuration using Raney nickel. The lignins are also premethylated with diazomethane. The thioacidolysis of CH,N,-methylated lignin samples isolated from pine compression and poplar woods allowed the study of the drastic structural differences between the samples [372]. Reaction Mechanism. Lapierre et al. [37 1,373-3751 have investigated the reaction using different models. They proposed a mechanism in which the boron trifluoride acts as a hard Lewis acid and the thiol as a soft nucleophile. The mechanism (Fig. 16) includes two steps. (a) The hydroxyl group at C,, is converted to oxonium cation (1) and then by a substitution reaction the thioethyl derivative is formed (2). (b) Formation of the oxonium cation at C, (3) occurs, and finally the formation of the di- and tri-ethyl derivatives (4)

Ho+cHR,

R ,H ,c--- ,~c--

OCH, FIGURE 15 Thioacidolysisproducts.(From Ref, 37 I .) G ~ t i ~ Ser-ie.s: y l I -(4-hydroxy-3-mcthoxyphenyl)- 1,2.3-(tris-thioethyI) propane (R, = SC,H,: R? = H); Syr-ir~gyISeries: 1-(4-hydroxy-3.5din1ethoxyphenyl)- I .2.3-(tris-thioelhyl) propane ( R , = SC,H,; R? = OCH,).

31 9

Chemical Characterization of Wood Y H2COH

R

H2COH

H2COH

4

Et2 “BF,

R2

(-ROH) R*

R2

bR

H2COH

H,CSEt

L

H SEt

l

l

HCSEt

HCSEt

4R2 EtSH

R,

OH

R2

(-ROH)

I

OR

(5)

(4)

OR

(3)

FIGURE 16 Mechanism of lignin thioacidolysis. (From Refs. 371, 373, and 374.)

and (5). This last step is important because it avoids the condensation reaction present in the acidolysis procedure. When R, is different from R2,50% of each isomer, erythro and threo, are obtained. The total and the relative amounts of guaiacyl and syringyl groups give information about the phenylpropane units involved in arylglycerol-p-aryl ether bonding pattern. Table 7 shows the results reported by Lapierre et al. [375]. These results indicate that lignins of woody angiosperms contain twice as many units of these types as lignins from woody gymnosperms. c. Nitrobenzene Oxidation ( N O ) . In this reaction, whichwasintroduced by Freudenberget al. [376], the lignin is treated withnitrobenzene in alkalinemedium ( 2 M NaOH) at elevated temperature (170- 180°C). The products have been identified by GCMS after the residue is silylated (Fig. 17). Under these conditions, softwoods give vanillin (main product) (17a) and vanillic acid (minor product)(17b). On the other hand, hardwood lignins, in addition to these two oxidation products, give syringaldehyde (17c), and the corresponding acid, syringic acid (17d) [377,378]. Reaction Mechanism. It is not clear which is the mechanisminvolved in this reaction. Chang and Allan [379] proposed a two-electron transfer process with a quinonemethide intermediate. Iiyama and Lam [380] also suggested a two-electron mechanism, but Schultz et al. [381,382] concluded that the oxidation involves a one-electron transfer through a free-radical process.

Baeza and Freer

320 TABLE 7 Monomers Released from the Thioacidolysis of Extractive Free Walls 13751 [H

+ G + S]”

[WG/SIh

Angiosperm woods Luurelia phillipiana

Cottonwood (Populus delroides) Poplar (Populus euroamericana) Oak (Quercus robur) Birch (Betula verrucosa) Nothofagus dombeyi

Gymnosperm woods Spruce (Picea abies) Pine (Pinus pinaster) (Compression wood)

1a60 I950 2310 1970 2490 2355

-166144 -144156 -137163 -132168 -122178 -l 14/86

1230 1140

I a1a21t

219a1t

“Micromoles per gram of klason lignin. hRelative distribution. t = trace.

cl. PermanganateOxidation. This is a specific oxidativedegradationmethod for the analysis of the phenyl nuclei and the linkagebetween the monomers.Thismethod wasproposed by Freudenberget al. [376], and later modified by Larson and Miksche [383]. The modified procedure includes first an alkylation step using dimethyl sulfate or diethyl sulfate to protect the free phenolic hydroxyl groups, followed by two oxidation steps, the first with potassium permanganate and the second with alkaline hydrogen peroxide. The degrative products are identified by GC, after methylation of the carboxylic acids with diazomethane. To obtain more detailed structural information, an alkaline solution of cupricoxide is used in combinationwithpermanganateoxidation.Figure 18 shows the principal products that have been detected by this method 1384,3851. e.PeriodateOxidation. Thismethodwas first described by Adler 1386,3871 and later by Lai et al. [388-3901 to estimate the phenylhydroxyl group content in wood lignin in situ, for both softwood and hardwood lignins. In this method the wood meal is treated with sodium periodate and the methanol formed analyzed by GC. The methanol is formed when the aromatic ring is oxidized to an orthoquinone (Fig. 19). Lai et al. [391,392]used this technique in combinationwith the phenylnucleus exchange and nitrobenzene oxidation reactions to estimate the distribution of phenyl hydroxyl groups in uncondensed lignin structures. In a related method, Ni et al. [393] described a relatively fast method to convert the methoxyl groups in lignin in methanol, using elemental chlorine. They defined the “meth-

OH

FIGURE 17 Nitrobenzeneoxidationproducts.(From Ref. 377.) a) R, = H, R? = CHO; b) R , = H, R, = COOH; C ) R , = OCH3, R? = CHO; d) R , = OCH,, R2 = COOH.

321

Chemical Characterizationof Wood a) R,=R,=R,=H b) R,=R,=H, R,=COOH c) R,=R,=H,R,=COOH d) R,=R,=COOH.R,=H e) R,=OCH,,R,=R,=H 9 R,=OCH3, R,=COOH,R,=H g) R,=OCH,,R,=H.R,=COOH h) R,=R,=OCH, R,=H

&

R1

OCH,

COOH

H3CW

(11

I

O OCH, OCH,

C

H

,

(2)

+0cH3q0cH3 COOH

R

COOH

I

FOOH

COOH

H,CO f 1 0 C H 3

0

I

OCH, OCH, a) R=H b) R=OCH,

(4)

COOH

I

COOH

I

COOH

OCH, H,CO W

O

C

H

OCH,

H,CO Q-QOCH, \

OCH,

a) R=H b) R=OCH,

,

OCH, OCH,

(5)

FIGURE 18 Permanganateoxidationproducts.(From

Ref. 384.)

anol number" as the methanol concentration in units of mg/L produced during 5 min of chlorination at 25"C, 1% consistency, and an initial chlorine concentration of 3.0 g/L. Ozonation. Ozone iswell known asa chemicalreagent in the elucidation of chemical structures. The main characteristic is its ability to react with the aromatic ring, giving complementary information to that obtained by other oxidative degradation methods, suchaspermanganateandnitrobenzeneoxidationmethods.Importantinformation

FIGURE 19 Methanol formation. (From Ref. 390.) R = lignin side chain; R, = H, OCH3, or lignin unit.

Freer 322

and

Baeza

about the behavior of the different structural units present in the lignin during the ozonation has been obtained from the study of lignin model compounds, Sarkanen et al. [394] gave a list of monomeric and dimeric lignin model compounds that have been ozonated. Matsumoto et al. E3951 described the products obtained by complete degradation of the aromatic rings present in the lignin, with the aliphatic carboxylic acids the main products formed (Fig. 20). Several lignins (MWL, klason lignin, dioxane lignin, thiolignin, and soda lignin) were reduced with sodium borohydride, then treated with ozone, followed by saponification. Theproductswereanalyzed by "C-NMRandGC.Erythronic, threonic, glyceric, and glycolic acids were detected. The ratio for the first two acids, which are derived from erythro and threo isomers of arylglycerol-p-arylether type structures, was almost 1 :1 for MWL and wood meal. Morerecently, it was reported [396] that when wood meal is subjected to ozonation the erythronic and threonic acids are predominant, but not in the case of unbleached kraft pulp. Taneda et al. [397] used this reaction to determine the relative abundance of the steric structures of lignin side chain. The ratio of erythro and threo isomers of the p-0-4 structure was determined in model compounds and in the residue lignin during the kraft process. The results showed that the E R ratio decreased during cooking. g. NucleusExchangeReaction(NE). This reaction, developed by Funaoka et al. [398-4051, consists of the treatment of lignin with boron trifluoride in an excess of phenol, causing a selective cleavage of C,-C, linkages of phenylpropane and forming methylene linkages of diphenylmethane type of structural units. The lignin aromatic units that are displaced by the phenol are extracted with ether from the reaction mixture and estimated by GC as trimethylsilyl derivatives. From model compounds studies, they found that all bonds lignin units are cleaved with the exception of diphenyl ether and biphenyl linkages. Reaction Mechanism. This is three-step reaction: (1) formation of the diphenylmethane structure by phenolization at the a-position of the lignin side chain, (2) a nucleus exchange step that involves displacement of the phenyl nuclei of lignin for phenol, and (3) demethylation of the methoxyl groups of the released phenols (Fig. 21). As shown in Fig. 22, in the case of softwood lignins the uncondensed guaiacyl units are converted first to guaiacol (22a), and this compound is partially 0-demethylated to give catechol (22b). In addition to these two phenols, the hardwood lignins give 1,3-0dimethyl pyrogallol (22c) from the syringyl unit, which by subsequent demethylation gives 1 -0-methylpyrogallol (22d) and pyrogallol (22e). This method in combination with the nitrobenzene oxidation process has been used to evaluate the amount of noncondensed, condensed, and diphenylmethane units of guaiacy1 and syringyl in residual lignins during the pulping process [406,407]. Chan et al. [408], more recently, indicated that the nucleus exchange reaction does not allow accurate measurement of aromatic units in lignin. They found erroneously high syringyl contents of lignins when lignins from eucalyptus woods andkraft pulps were studied. They also found, from studies of the NE reaction using different models, that diphenyl ether and biphenyl compoundsgavecathecol as a reaction product.Theyconcluded that the amounts of noncondensed structures and the ratio of syringyl and guaicyl nuclei in lignin by the NE reaction and the amounts of diphenyl moieties determined by NE-nitrobenzene oxidation in modified lignins is likely to give incorrect results. h.Pyrolysis-GasChromatography-MassSpectrometry(Py-GC-MS). This analytical techniqueincludespyrolysis (Py) to generate volatile degradationproducts, GC to separate the degradation fragments, and MS as a detection system. The identification of the products can also be performed using the normal flame ionization detector (FID), and Fourier-transform infrared (FTIR). Several pyrolyzers are commercially available, such as

Chemical Characterizationof Wood

0

E

2 tL,

C

m

.-0

S 0

W

K

3

PU

323

324 I

0

0--0--0

0 I 0

I

Baeza and Freer

Chemical Characterizationof Wood

325

OH

FIGURE 22 Products of NE reaction of lignins. (From Ref. 398.) a) R, = OCH2, Rz = H; b) R , = OH, Rz = H; c ) R , = OCH,, Rz = OCH,; d) R, = OCH,, R? = OH; e ) R, = OH, Rz = OH.

JHP-3 model Japan Analytical IndustryCO;Chemical Data System Probe; PYROLA, Pyrol AB, Lund, Sweden; Fischer, Germany. This is an effective method for the characterization of polymers in a short time, with simplepreparation (no special pretreatment), high sensitivity, andusingsmallsamples. The application includes the analysis of the isolated lignin from the wood or pulp and also the lignin without previous isolation (total lignin). Some good examples of Py-GCMS are the following, reported in the literature. Obst [409] used this technique to pyrolyze wood and classify lignins as either guaiacy1 type or syringyl-guaiacyl type. He was also able to isolate vessel elements and identified the type of lignin. The pyrograms of the woods and the milled wood lignins are clearly different, and the products that were identified allow the distinction between the two types of lignins. Figure 23 shows the pyrograms of milled wood and the MWL lignin for loblolly pine and white oak. This technique permits quick and clear separation of guaiacyl and syringyl-guaiacyl lignins. Kuroda et al. [4 IO] compared pyrolysis products of sugi wood and the lignins isolated by different procedures. As an example, the pyrogram of sugi wood is showed in Fig. 24. The compounds identified are the 4-p-hydroxyphenyl and guaiacyl types. These authors also studiedthe pyrolysis of milled wood, alcoholbisulfite, hydrochloric acid, kraft, and klason lignins. In all cases, they observed that the lignin preparation methods greatly affected the product profiles and the total yield products. The yields werealsostronglydependent on the number of condensed units in the lignins. Analysis of chlorolignin residues was done to obtain information after treatment of wood chips with sodium chlorite. For this purpose, Pouwels and Boon [411] identified chlorinated methoxyphenol, guaiacol, and syringol derivatives in the pyrolyzate of the xylan fraction isolated from beech. Characterizationofdissolvedorganicsubstanceswas done on anewsprintwhite water system (thermomechanical pulp, TMP) [412]. The chemical structure of the dissolved polysaccharides and of the hydrophilic lignin were established. A comparison between the lignin and sugar composition of pulp, long fibers, fines, and paper were also done. Previously, Sjostrom and Reunanen[4 131 used this technique for characterization of water-soluble organic substance isolated from spruce groundwood pulps. The syringyl/guaiacyl (S/G) ratios were determined in hardwood kraft pulps by Tanaka et al. [414]. The relationship between S/G and cooking time shows similar behavior to the ratios of syringaldehyde to vanillin (S/V) by nitrobenzene oxidation. The authors also analyzed the pulp after bleaching with chlorine.

326

t

-

ti 9

m

-I

-

r

l iL

1

lunm UOI

puyd

h

0; 0 W

0

cr

m (v

W

a

N

0

Chemical Characterizationof Wood

C V -

0

U

P

-S

r

327

Freer328

and

Baeza

2. Nondegradative Procedures a. Ultraviolet Spectroscopy (UV). This is the most simple and classical analytical method for lignin characterization. The spectrum of a softwood lignin (“guaiacyl lignin”) shows a maximum at 280 nm (band B), in agreement with the spectrum obtained when the benzene ring is substituted by hydroxyl or methoxyl groups, shifting the secondary maximum from 254 nm to 270-280 nm, accompanied by a five- to sixfold increase in intensity. Due to the presence of the syringyl structure in the hardwood lignins (“guaicylsyringyl lignins”), the maximum appears at 274-276 nm (Fig. 25). Two extra bands are present in typical softwood lignins, a shoulder at 230 nm (band E2) and a sharp peak at 200-210 nm (band E,). The B-band absorptivity values for softwoods are in the range of 18-21 L/g.cm, while for hardwoods the values are much lower (12- 14 L/g.cm). This band is usually used in qualitative and quantitative UV determination of lignins [4164191. Difference spectra ( A s curves) have been used to help in the interpretation of the UV spectra. This method has been used to study the different types of chromophores in softwood and also in the case of more complicated hardwood lignins. However, to obtain more information about the structure of the lignins it is initially necessary to study the spectra of model compounds [420,421]. The ionization difference spectrum, Asi, obtained by difference between the absorptivities in alkaline solution (ionized) and in neutral or acid solution (nonionized) (Fig. 25), represents the absorption of the ionizablephenolicgroups,and the phenolichydroxyl content can be estimated by comparison with standards. The spectrum obtainedafter treatment with reducing agents such as sodium borohydride resulted in a decrease in the intensity of the band assigned to free phenolic groups conjugated to reducible groups such as carbonyl, carbon double bonds (300-400 nm) [422,423].

-----

Red Pine MWL Beech MWL

n

200

300

400

Wavelength (nm) FIGURE 25

UV and A s , spectra red pine and beech MWLs. (From Ref. 415.)

aracterization Chemical

of Wood

329

The hydrogenationdifferencespectrum, A€,,, is obtained as the difference of the spectrumof the lignin hydrogenated by using a mild catalyst and that of the original sample. Marton and Adler [424], using this technique, determined the content of cinnamaldehyde units in a milled wood lignin. The double-bond content has been determined by using a AE,,curve of a reduced sample with sodium borohydride [425]. The a-carbonyl groups, when present, can be determined by NaBH, reduction difference spectra, Asr. This approach is based on the spectral changes occurring upon borohydride reduction of the a-keto groups, causing a decrease in the absorption at 260280 nm [426,427]. One important aspect to consider is the solvent used to run the spectra. Due to solvent effects, the spectrum of an organic compound would be modified. Some lignins, such as the lignosulfonates, are soluble in water, but for those that are insoluble, it is possible to use different organic solvents, such as dimethylformamide,ethanol,2-methoxyethanol, dioxane,dimethylsulfoxide, pyridine, dichloroethane,cellosolve,orhexafluoropropanol. The last one presents ideal UV transmission, being a convenient solvent for the determination of lignin content without interference from polysaccharides or their potential degradation products by using the intense maximum between 200 and 205 nm, permitting an estimation of the content of lignin in small samples [428] (Fig. 26). Acetyl bromide is a good solvent for wood, holocellulose, and pulp samples, forming the basis for the spectroscopic determination of lignin. In this method, the samples are dissolved in a mixture of 25% of acetyl bromide in glacial acetic acid at 70°C and the absorbance measured at 280 nm, which is proportional to the lignin content of the sample [422-4241. An improved method,using acetyl bromide-containingperchloric acid, hasbeenapplied to analyze lignin in samples of woodandwood pulp. Thismixtureallows a faster andcomplete dissolution, and coarse samples can be used [429-4311. b. Infrared Spectroscopy (IR). Two vibrational spectroscopy regions, near infrared (near IR) between 10,000 and 4000 cm-', and mid infrared (mid IR) in the region of 4000-450 cm", have been used in the analysis ofwood and some of the applications will be discussed.

2.1

1 .E5

1 : MWL beech

2: MWL spruce 3: Organosolv lignin .-0 p 1.2 .k 0.9

spruce

0.6 0.3 190

FIGURE 26

2 i0

230

270 290 Wavelength (nm)

250

3io

330

350

UV spectra of lignins measuredinhexafluoropropanol (HFP). (From Ref. 428.)

Freer 330

and

Baeza

Infrared is an easy way to obtain relevant information in the analysis of lignin. At first, dispersive infrared instruments were used, but this technique presents limitations of selectivity and detection. This situation was improved with the appearance of the FTIR spectrometer. A complete description of FTIR techniques, including instrumentation, advantages of FTIR, and its applications, is given by Perkins in three articles in the Journal of Chemical Education [432-4341. Also, Michell [ 1741 published a clear explanation of the difference between the Fourier transform technique and dispersive IR spectroscopy. In the FTIR technique the sensitivity is improved with the use of the Michelson interferometer, whichallows the continuousdetectionof all of the transmittedenergy simultaneously. The discovery of the fast Fourier transform algorithm (FIT) and improvements in computer software that allowed more advanced data analyses give FTIR a further advantage. Some of the characteristics of this techniques are: short analysis time, high sensitivity, high linear range for quantitative work, stability, ease of use, convenient data handling, and increase in resolution by using the deconvolution technique or derivative spectroscopy. Several optical techniques has been developed to collect the diffuse reflected radiation, suchas diffuse reflectance infrared Fouriertransform(DRIFT)spectroscopyand attenuated total reflection (ATR). Drift spectroscopy is a method that involves minimal sample preparation and is useful for powder samples and also for polymers, fibers, and films. ATR is a contact sampling method involving a crystal with high refractive index and low IR absorption. Mid Infrared. Various applications of FTIR in wood chemistry have been reported, but only some typical examples will be given. Faix [435], in an interesting paper, classified lignins according to the FTIR spectra obtained from more than 100 milled wood lignins. The lignins were differentiated according to their basic units: guaiacylpropane (G), syringylpropane (S), and 4-hydroxyphenylpropane (H). The band assignments for milled wood lignin in spruce (G), beech (GS), and bamboo (HGS) are shown in Table 8. The G type lignins have a typical maximum band at 1140 cm", and those of type GS show a maximum between 1128 and I 125 cm-'. A few percent of S units in a lignin is enough to change the absorption maximum from 1140 to a wavenumber below 1128 cm". The band assignments in Table 8 agree with those given by Schultz and Glasser [436] They did a quantitative structural analysis of lignin by diffuse Fourier-transform infrared spectrometry,obtaininganempirical relationship which permits good predictability of structural features, including phenolic hydroxyl content, methoxyl content, aromatic hydrogen content, hydrolysis ratio, and condensed ratio. A rapid FTIR methodfor quantification of phenolic hydroxyl groupsin isolated lignin and in spent liquors by using acetylated samples of various origins and chemical composition was reported by Wegener and Strobe1 [437]. The method is based on the evaluation of the phenolic ester band (about 1765 cm-') and the aliphatic ester band (about 1745 cm"), normalizing the spectra by setting the aromatic band (1510 cm") to 100%. The relation of the two ester bands (1765/1745) was found to be a suitable index for the quantitative evaluation of the phenolic OH groups. Resolution of the two maxima is possible by using either the deconvolution technique, the first derivative, or extension of the spectra in the region 1700-1800 cm". Good correlation between the IR results and the values obtained by the time-consuming aminolysis method was found. FTIR has proved to be useful for analysis of chemical changes during pulping and bleaching processes. Michell [438] demonstrated a good relationship between the spectroscopic and analytical data for both the kappa number and the yield of the pulp during the cook. The kappa numbers of pulpsweredeterminedfrom the IR spectra of the spent

aracterization Chemical Wood TABLE 8

331

of

Band Assignments MWL Lignins of Spruce, Beech, and Bamboo

Range of maxima" Bamboo Beech Spruce

Maxima" Peak assignment

3428 3002 2942 2879 2840 1709

1738- 1709

3412 3000 2937 2879 2840 1722

3460 3000 2940 2880 2840 1735

1675-1655

1663

I658

1605- 1593 1515-1505 1470-1460 1430- 1422

1596 1510 1464 1423

1593 1505 1462 1422

1370- 1365

1367

1367

1330- 1325 1270- 1266 1230- 1221 1166 1140

1326 1269 1221

1329 1266 1227

1329 1267 1229 1166

1 l26

1127

1032

3460-34 I2 3000-2842

[435]

OH stretch CH stretch in CH3 and CH2 groups

C=O stretch in unconjugated ketone, carbonyl, and ester groups

C=O stretch on conjugated p-substituted aryl 1601 1511 1462 1423

1140

1086

1086

1030-1035

1032

1033

925-915 858-853

919 858

925

835-834

835

832-8 17

817

"Wavenumbers in cm

834

ketones Aromatic skeletal vibration plus C=O stretch Aromatic skeletal vibration CH deformation in CH, and CH2 Aromatic skeletal vibration combined with CH in-plane deformation Aliphatic in CH stretch in CH,, not in OMe; phenol OH S-ring plus G-ring condensed G-ring plus C=O stretch C-C plus C-0 plus C=O stretch C=O in ester group; typical for HGS lignins Aromatic C-H in-plane deformation; typical for G units G-condensed etherified, typical for S units; plus secondary alcohols plus C=O stretch C-0 deformation in secondary alcohols and aliphatic ethers C-0 deformation in primary alcohols; plus C=O stretch (unconj.); plus aromatic C-H in-plane deformation CH out-of-plane; aromatic C-H out-of-plane in positions 2, 5 , and 6 of G units C-H out-of-plane in positions 2 and 6 of S, and in all positions of H units C-H out-of-plane in positions 2, 5, and 6 of G units

'

liquors from kraft pulping of Eucalyptus sieberi, obtaining a precise relationship (linear plot with R' = 0.99) between integrated band intensity (1 118 cm") and kappa number. Previously, the same author studied the chemical changes in woods during soda and soda antraquinone processes [439]. Faix et al. [440] evaluated the continuous process control of pulping by FTIR. The kappa numbers and the yields of pulps obtained in ASAM and kraft AQ processes were determined.

Freer332

and

Baeza

Due to the complexity of wood it is difficult to assign each peak to a single component, and interpretation of isolated bands in wood FTIR is misleading [ 174,4411. For this reason the use of multivariate analysis (MVDA) have been frequently used to correlate FTIR and near-infrared reflectance spectra data with changes in structure or wood composition during chemical or physical processes. Multivariate data analysis is the process of learning how to combine data from numerous channels to overcome selectivity problems, gain new insights, and allow automatic detection. It has been shown to be useful for the evaluation of experimental data, especially for data containing correlated variables. By usingMVDAtoevaluatespectroscopicdata, it is possible to select useful datato interpret andmodelhighlycomplex spectra withoverlappingpeaks.Usingthese techniques, several papers have reported rapid methods for estimating the concentrations in wood or pulp components [442-4451. It is necessary to have a set of samples of known data, which are used to model the relationship between some quality of the sample and the absorbance band. This step is called “calibration” and is done by using chemometric software, which provides methods such as partial least-square regression (PLS), principalcomponents analysis (PCA), and principal-componentsregression (PCR)[442,446,447]. After the model is tested by using another set of known data (validation step), it is possible to predict new, unknown values. Another lignin behavior that has been studied using FTIR is the photo-induced color reversion of bleached and unbleached stone groundwood pulp [448]. Changes in the chemical structure were monitored after exposures to UV irradiation. The change in the lignin content was followed using intensity of the peakat1509 cm”, and the change in the carbohydrate content followed at 1060 cm”. Also, a study of the yellowing of Eucalyptus reganns from cold-soda pulp has been done [449]. Michell [450] studied the reaction of inorganic agents with wood, particularly those that are used to protect the wood surface, such as chromium trioxide. Due to the reaction with the aromatic rings of the lignins, changes at 1505 and 1595 cm” were observed. Near-Infrared Spectrometry (Near-IR). The absorption bands in this region can be assigned to overtone and combination vibrations, primarily of OH, NH, and CH functionalities. In general, visual interpretations are more difficult because of the nature of the bands. The spectral bands are often overlapped, and differentiation between similar materials appears less definitive than in the mid-IR region. In spite of these apparent disadvantages, the near-IR region provides useful qualitative information, especially with the use of computer-assisted data analysis techniques. There are also other advantages, such as the use of cells of glass or quartz, insensitivity to water, and lower signal-noise ratio compared with mid-IR. Changes in the bands of near-IR might be used to measure differences in wood chemistry such as occur in different tree zones, between woods of different species, between woods grown on different sites, etc. Near-IR has been utilized to determine the content of lignin and cellulose, and also toestimatethe fiber orientation, moisturecontent,kappanumber,and the pulp yield. Recently,Brunneret al. [451]analyzed 90 samples of 12 different species,and it was possible to distinguish between samples of a given wood species of different origins. For this purpose, a chemometric software was used. Michell and Schimleck [452] investigated the origins of the bands in the near-IR spectra of Eucnlypfus glohulus woods. Different approaches were used, including comparisons of bands in the near-IR with those in the near-IR of its major components (cellulose, glucuronoxylan, MWL, and hot water extractives). Figure 27 shows the second-derivative near-IR spectra of E. glohulus wood, and of some of the different isolated components-i.e., those removed by treatment of the wood with chlorite and with weak alkali-and also variations in intensity of the major bands

333

Chemical Characterization of Wood

0 0

0 0

r

r

m

m

0 0

t

0 0

z

z

0

I 0

O m N

Wavelength ( m ) FIGURE 27 Second-derivative near-IRspectra of E. globulus, cellulose, xylan, lignin, and extractives. (From Ref. 452.)

in the mid-IR spectra and their connections with near-IR bands and particular chemical components in wood.Many partial correlationswerefound,showing a highdegree of intercorrelation between the bands and confirming that the near-IR arose from the combination of several fundamental bands. Recently [452], the use of diffuse reflectance, near-IR, and multivariate evaluation have been applied to determine yield, kappa number, lignin, glucose, xylose, and uronic acid during akraft pulping process. This method has the advantage of being fast. Principalcomponents analysis (PCA) has also been used in conjunction with near-IR to discriminate between woods from pines and eucalyptus, between woods from different eucalyptus species, between woods from different provenance, and between woods fromthe same species of eucalyptus on different sites [453].

Freer334

and

Baeza

Infrared Microspectroscopy. This technique allows the determination of the molecular composition of microscopic particles. This is a valuable microanalytical technique, mainly for the analysis of contaminants in the pulp and paper industry. In Table 9 are given some of the results obtained by Sommer and Katon [454]. This technique provides molecular information on samples whose size is of the order of the analytical wavelengths being employed, and the coupling of a research optical microscope to the spectrometer allows morphological examination of the sample. c. NuclearMagneticResonanceSpectroscopy(NMR). NMR is one of the most valuable techniques in the elucidation of the structure of organic compounds, including wood components. The four most important NMR methods in lignin chemistryand in organic chemistry in general are 'H, "C, "P, and "F. All of these nuclei have spin quantum numbers (I) of 'h. 'H-NMR Spectroscopy. 'H-NMR spectra are usually run by using acetylated lignins, generally in deuterochloroform as a solvent. To avoid the problems of derivatized lignins, the spectra can be carried out in solvents such as DMSO-d6, trifluoroacetic acid, and dioxa-d,-D,O. The main signals in the 'H-NMR spectra of an acetylated lignin are given in Table 10. The total number of hydroxyl groups in lignins has frequently been determined by 'H-NMR spectroscopy of their acetate derivativesby using a method developedby Ludwig et al. [456]. This method is the basis of several applications [457-4581. Aliphatic acetoxyl and aromatic acetoxyl signals permit the estimationof the total number of hydroxyl groups in lignin, but it is difficult to obtain accurate values of the number of phenolic groups on the basis of the 6 = 2.3 peak, because the aromatic and aliphatic OAc signals overlap [459]. Besides, one should bear in mind that the aromatic acetoxy groups in 5-5 bonded dimeric units appear in the range of aliphatic acetoxy groups and are not considered in the estimation of the phenolic groups. To overcome these problems, the spectra of nonderivatized lignins in DMSO-& solutions may be used. Under these conditions, the majority of the protons corresponding to phenolic hydroxyl groups are found in the spectral range 6 = 8.0-9.3. In Fig. 28 the NMR spectrum (400 MHz) for MWL from spruce in the spectral range 8.0-10.5 is given. The assignments of the peaks were made by using model compounds. I3C-NMR Spectroscopy. I3C-NMR spectroscopy has several advantages compared with 'H-NMR [460]. "C-NMR became a useful technique after the introduction of Fourier

TABLE 9 Mixture Components, Identifying Adsorptions, and Most Probable Source [454]

source probableMost absorptions Identifying Components Kaolin 3695, 3619, 1031, 915 and 938, cm-' Calcite 1450, 874, and 710 cm" Aragonite 1496 and 855 cm-' cm"10 11 CaSO, Cellulose 3368, 1161, and 1033 cm" Oxalatesalt1620and1318 cm-' Defoamer 3305, 2950, 2922, 2866, 2854, 1635, making) paper cm" 1366 and 1462, Polyethylene 1719, 1339, 1257, 1253, and 714 cm" Paper machine (wire material) Terephthalate

Additive Additive/environmental Additive/environmental Processing interaction Raw materials Raw materials (wood) Additive (pulp preparation and/or

335

Chemical Characterizationof Wood TABLE 10 Assignment of the Main Signal of the 'H-NMR Spectra of

Acetylated Lignins [4551 Chemical shift range (6 Type in ppm) 9.58-9.86 7.23-7.90 6.25-7.23

5.75-6.25 5.20-5.75 4.50-5.20 3.95-4.50 3.55-3.95 2.50-3.55 2.20-2.50 H 1.50-2.20 H

of hydrogens

Ar-CH=CH-CH0 CH0 in Ar-H in Ar-COR Ar-H in Ar-R H-a in Ar-CH=CH-CH0 H-p in Ar-CH=CH-CH0 H-a in Ar-CH=CH-CH,OAc H-a with a-0-Ac in p-0-4 and p-1 H-p in Ar-CH=CH-CH,OAc H-a with a-0-Ac in p-5 (0.09/C9) H-a with a-0-Ac in p-0-4 and p-l (O.OS/C,) H-p in p-0-4 H-y in Ar-CH=CH-CH,OAc H-a in p-p H-y in p-0-4, p-5, p-l, and p-p (3.39/C,)Ar-0-CH,, H-p in 0-5 (O.O9/C,) H-y in p-p (O.O4/C,) H-p in @ - l , p-p, others and in Ar-OAc except for 5-5 unit in Aliph-OAc Ar-OAc and 5-5in units

1.10-1.50

0.75-1.10

( I ) it is possible to obtain transforms (FT). Some of the advantages are the following: information about all the carbon skeleton; (2) there is more resolution over a much wider chemical shift range and less overlap of signals (the I3C-NMR range is about 200 ppm, compared with only 12 ppm for 'H-NMR); ( 3 ) spin-spin coupling between carbons almost does not exist. The I3C-NMR spectrum of lignin can be divided into three main segments: ( I ) 200I65 ppm, carbonyl bonds; (2) 165- 100 ppm, aromatic and olefinic carbons; ( 3 ) 100- I O ppm, aliphatic carbon atoms. The assignments have been made by using I3C-NMR lignin model spectra. Table 1 I summarizes the chemical shift data [461] for 'C-NMR of a lignin from poplar, using DMSO-d, as solvent, and in Fig. 29 appears a typical routine '.?2-NMR spectrum [462]. Like 'H-NMR and IR, '7C-NMR have been used to distinguish the origin of lignins. Nimz and co-workers 14631 studied the structural differences using acetylated lignins.Takingintoaccount that therearethreecharacteristicsignals in thesesamples (81.1, 75.5, and 63.8 ppm), corresponding to carbons atoms a, p, y in p - 0 - 4 structures, it is possible to distinguish the aromatic nuclei occurring in lignin guaiacyl (G), syringyl (S), and p-hydroxylphenylpropane (H). Several experiments that have employed "C-NMR have been useful in the elucidation of lignin structure, beside simple experiments without irradiation of the protons. One of these is distortionless enhancement by polarization transfer (DEPT), which simplifies the spectra. DEPT is a one-dimensional pulse sequence, involving the spin-echo phenomenon sequence, which allows separate recording of the NMR signals for the CH,, CH,,

w w

Q)

C C

I

C

FIGURE 28

H-NMR of MWL from spruce (400-MHz; solvent DMSO-d,). (From Ref. 459.)

aracterization Chemical

of Wood

337

Chemical Shifts and Assignments of "C-NMR Signals for Poplar Lignin Samples (Solvent DMSO-d,,, T = 323 K ) [46 I ]

TABLE l 1

Signal I

2 3 4 5

6 7 8 9 10 11

12 13 14 IS 16 17 18

19 20 21 22 23-24 2s 26 27 28 29 30 31 32 33 34 3.5 36 37

6 (ppmnMS)

Assignments

171.6 168- 167 163.4 152.7- 152.3 149.6- 149.2 147.6-146.9 145.6 139.1-138.1 13s- 134 133- 132 131.6-131.3 129.5 I22 119.4-1 19.2 115.8-115.6 111.8-111.1 107.8-107.3 106.8-104.5 103.6 102-101 07.6-97.3

C=O acetyl in xylan C=O in benzonic acid C-4 in H ne C-3/C-5 in S p-0-4 e C-3 in G e C-3 in G p-0-4 ne; C-3/C-5 in S p-0-4 ne C-4 in G p-0-4 ne C-l in S p-0-4 and G in p-0-4 e C-4 in S p-0-4 e and ne C-l in G p-0-4 ne; C-5/C-5' in 5-5' units C-2/C-6 H units C-a and C-p in vinylic structures C-l in H units C-6 in G units C-5 in G units, C-3, C-5 in H units C-2 in G units C-6 in S units with C=O C-2/C-6 in S units C-2/C-6 in S p-p X-l in xylose units X,,: anomerin carbon in xylose units X<.: anomerin carbon in xylose units C-p in S and G p-0-4; C-LYin p-p C-dC-p in p-0-4 and a-0-4 units: C-4 in 4-OMeGlu X-2 in xylose units linked X-4 in xylose units X-3 in xylose units X-2 in xylose units C-a in p-0-4 units C-y in p-p; C-2 in 4-OMeGlu C-S in 4-OMeGlu C-5 in reducing xylose units C-y in p-S and p-0-4 with a C=O; X-5 in xylose units C-y in S and G p-0-4 Aromatic OMe in S and G units C-p i n p-p and p-5 units

92

87-84 81 76.9-76.0 75.5-75.0 74.7-74.0 72.7 72.2 71.8-71.0 70.1-69.5 65.7-65.5 63.3 60.2 56 53.4-52.7

S = syringyl units; G = gunincyl unlts: H = p-coulnaryl units: c = in etherified struciures: nc structures: X , = xylose units: 4-OMcClu = residue of methyl glucuronic acid.

= in nonethcrifcd

and CH groups. Another clear advantage is a large signal intensity enhance~nent due to the polarization transfer. This polarization transfer relies on a spin population interchange fromthehigh-sensitivitynucleus, 'H, to the less sensitivenucleus, "C, to enhancethe latter. In Fig. 30 is shown a DEPT edited "C-NMR spectrum of birch lignin 14641. This technique was also used to obtain more information about kraft lignin obtained during the alkaline delignification of hardwood (poplar) in a flow reactor [461 I andmilled lignin obtained from softwood ( Pirzus .sylvc..str-is) 14651.

338

r

-In

0

2

-0

In

-0

og

n . .-c

' E

-0

Baeza and Freer

-3 -3 -8 N

0 N -In

Chemical Characterizationof Wood

Baeza and Freer

340

The application of "C-NMR to solid samples gives spectra with weak. broad, and poorlyresolved signals. Theweak signals havebeenimprovedusing cross-polarization pulsesequences(CP)and the broadsignalsusingmagic-anglespinning(MAS)[466]. Figure 31 shows "C CP/MAS NMR spectra for different lignins [467]. A "delayed-contact" pulse sequence was used to separate I3CCP/MAS-NMR spectra into subspectraof kraft pulp components. The method exploits differences in rotatingframe relaxation time constants for cellulosic and noncellulosic domains within the sample. Lignin contents were estimatedfor the subspectra, and good agreement was found between these values and those determined by klason lignin. The author concluded that this technique is useful in chemical analysis in pulping processes [468]. Different morphological fractions of spruce woods were analyzed usingI3C CP/MAS NMR spectra [468]. Spectra for the whole wood and five morphologically different fractions (compression, ray cells, cambial, middle-lamella particles, and "fines") were analyzed. Differences in the relative amounts of cellulose, lignin, hemicelluloses, and protein were observed. The fractions also present different degrees of cellulose cristallinity; the "fines" fraction showed the lower value, suggesting that it is derived mostly from the primary wall. Two-Dimensional (2D) NMR Methods. The use of 2D-NMR spectroscopy gives additional information, including connectivity (structural information) and coupling constants (stereochemical information) [470]. A brief review of 2D NMR has been published by Williams and King [471]. It has been shown that it is possible, by the combined use of homo- and heteronuclear correlative techniques, such as homonuclear Hartmann-Hahn

C ~

. 250

-

"

-

~

'

200

~

"

i

"

150

"

l

"

"

100

l

"

-

~

50

l

'

~

-

-

0

FIGURE 31 I7C-CP/MAS NMR spectrafor ligninfromsouthernpine (b) and (c) kraft lignins after different cook times. (From Ref. 467.)

i

- 50

PPm

wood: ( a ) Klasonlignin;

aracterization Chemical

of Wood

341

(HOHAHA) [472] or TOCSY, to determine H-H connectivity and heteronuclear multiple quantum coherence (HMQC) [473]in the determination of H-C connectivity. These methods unambiguously determine the presence or absence of interunit structures that have beenpreviouslycharacterized by degradativemethodsandone-dimensional NMR. As indicated by Ede and Kilpelainen [474], these methodsare particularly sensitive and permit the acquisition of information in a reasonable time, using small lignin samples. As examples, some of the results obtained with these techniques are given below. H-H COSY experiments and J-resolved 2D-NMR spectra of acetylated lignins and synthetic lignins were recorded and the presence of arylglycerol and 3-arylpropanol was confirmed. The percent of a-O-4-aryl ether and p-C-1 linkages were determined, each accounting for no more than 2% of the phenylpropane units, confirming that these linkages are less frequently present in the lignin than was previously thought. In Fig. 32 appears the COSY spectrum of acetylated lignin and in Table 12 the assignments of the peaks [470]. Information concerning the heterogeneous distribution of p-1 structure in different MWLs from Pinus radiata sapwood (dioxane-water, acetone-water, and acetic anhydride lignins) was obtained from the HOHAHA spectra of the acetylated MWL samples under identical conditions[475,476]. It hasbeenshown that there is extremely good agreement between the substituted model compound sidechain and the topology of the HOHAHAcorrelationsfromacetylated

F1

8 FIGURE 32

7

6

5

4

3

2

COSY spectrum of acetylatedspruce MWL. (From Ref. 470.)

and

342

Baeza

Freer

TABLE 12 Assignments of COSYCross-Peaks [470]

Integral"

F

sc

DHP

MWL

A

7.5417.34

0.5

-b

0.7

B

7.3416.52 6.871632 6.5216.09 6.1414.62 5.9914.63 5.9215.44 5.4513.65 4.7914.03 4.4313.13 4.20/3.00 3.9811 .81 2.521 I .S1 1 S217.34

6.6 9.8 3.3 3.3d 0.5 3.3 2.3

3.2 25.3 5.9

Peak

C D E E' F G

H I J

K L M

4.7d -

2.1 -

-

-

1 .o

0.9 2.4 1.9

6.6 6.6

-

-

5.2

0.4 2.3 0.5 2.6* 0.3 1.9 1 .o 0.5 2.6 2.6 -

Assignment protons Aromatic Adjacent to carbonyls alP:4 ?' dP:3 PI y:3 alp:1 dP:6 alP:2 ? ? PI y:9 ylP:5 aIulp:5 PI y:4

"Normalized to OCH, = 1000. hNot observed. 'Unknown. "Uncertain due to close proximity of E. 'Folded and not integrable.

MWL [477]. Figure 33 shows the HOHAHA spectrum of the side-chain region of acetylated dioxane-water MWL. The cross-peak in region A, arising from H,,-H, magnetization transfer in p-0-4 structures, permits the conclusion that the levels of p-0-4 were approximately equal for each lignin. The correlations from the other structures were also of similar intensity among the three samples. However, there are significant variations in region B, which encompasses the chemical shift ranges of the side-chain proton from p-l structures. The higher content of p - l units was found in the acid anhydride MWL, being somewhat less in the acetone-water lignin and significantly higher than that of the dioxane-water lignin, suggesting that the p-1 units are not evenly distributed within the cell wall. Acetylated residual lignin from unbleached kraft pulp, isolated by cellulase treatment. wasstudiedusingHOHAHAexperiments in combinationwitha selective cleavagewithpivaloyl iodide. It wasconcluded that: ( I ) the residual lignin obtained from beech unbleached kraft pulp still contained p-0-4 and resinoltypechains; and (2) glycosidicbondsbetween lignin andcarbohydrates are present in the residual lignin [478]. The use of HOHAHAandHQMCexperimentsshowed that these techniques are sensitive, rapid, and unambiguous probe for the presence or absence of noncyclic benzyl aryl ether (a-0-4) structures in soluble lignin samples. The limits of detection by 2D NMR techniques is < 0.3 structure of a-0-4 structures per 100 C9 units. Other techniques that have beenused to determine a-0-4 are the acidolysis and ID NMR techniques, both beingambiguous [479]. In the

59 61

I .

J

49

4.7

45

A 49

63

_c--

47

49

45

47

45

45

5.0 5.5

J *

6.5

3

3

35

34

3.3

-55

@

-60

P 1

3.5 34

33

I

B

~-.- . , 3.5 3 4 3.3

J

FIGURE 33 Side-chain region of HOHAHA spectrum of acetylated dioxane/water (Pinus radiafa) and expansion of the regions A and B for acetylated MWLs: (a) dioxane-water: (b) acetone-water; (c) acetic anhydride. (From Ref. 476.)

344

Baeza and Freer

first case the cleavage of any other ether-linked phenol will give an overestimation of the a - 0 - 4 structures. In the case of I D NMR it is not possible to do a correct chemical shift assignment.

3'P-NMR Spectroscopy. "H-NMRspectroscopyhasbeen used intensively in lignin chemistry but presents some limitations, such as the limited range of chemical shifts, poor spectral resolution, and extensive signal overlapping. To overcome these disadvantages, ?IP-NMR spectroscopy has been used. The labile protons of residues of phenols, alcohols, aldehydes, sugars, and carboxylic acids react with 1,3,2-dioxaphospholanyl chloride (Fig. 34) [480]. This method presents several advantages: ( 1 ) rapid and quantitative reaction; (2) a sharp, single "P signal is obtained; (3) phenols, alcohols, and simple carboxylic acids give "P signals in separate regions; (4) chemical shift is sensitive to the chemical environment of the phosphitylated center; (5) it is possible to obtain details about the stereochemistry i n both lignin modelcompoundsand isolated lignins [481].The syringyl, guaiacyl, p-hydroxyphenyl-free phenolic groups, primary groups, and secondary groups can be quantitatively determined by this technique 14821. The analysis of a set of lignins by this method has been done, andthe results obtained for the total hydroxyl contents compared with those obtained by using 'H-NMR and wet methods. The results appear in Table 13, and a typical 31P-NMR is shown in Fig. 35 14831. Structural changes during the pulping process have been studied by using this technique. The ortho-quinone contentin mechanical pulp can be evaluated [484].The hydroxyl content during the kraft process was monitored using "P-NMR and the results were correlated with "C-NMR. Also it was possible to establish the stereospecificity of the degradation reaction 14851.In a similar study, dissolved lignin isolated from the three-stage formic/peroxyformic acid (Milox) pulping process was studied using 31P-NMRand oxidative degradation 14861. Quantitative"P-NMRallowed the determinationof guaiacyl, syringyl, total condensed free phenolic OH groups, carboxylic acids, and the erythro and threo forms of the a-OH present in the p-0-4 units. In addition, oxidative degradation provided information about the fate of specific condensed structures such as biphenyls, biphenyl ethers, and C5- and C6-substituted guaiacyl units. It was possible to obtain information on the rate and the topochemistry of the scission reactions of arylglycerol-paryl ethers and the condensation reactions that occur in the lignin. The cleavage of the p0 - 4 aryl ether linkages within the lignin was approximated 45% during the first stage and considerably intensified in the second stage, with a concomitant increase in the total phenolic hydroxyl content. A significant amount ( 1 6%) of fhreo-arylglycerol-~-aryl ethers seems to remain after the third stage. The guaiacyl phenolic units predominated within the solubilized lignin after the first stage. This, together with the fact that guaiacyl units are less reactive than syringyl units towardperoxyacids, indicates that a topochemical effect is operating during this stage, being a greater accessibility toward the guaicayl-rich middle lamella lignin. The second stage seems to preferentially cleave the syringyl-rich

ROH

+

CCp/

'1

\O

PyndinelCDCI, w

25 "C

R-0"P

'1

+

HCI \O

(1)

FIGURE 34 Derivatization reaction. (From Ref. 480.) Where R = residues of phenols, alcohols, aldehydes, sugars, carboxylic acids.

3n,

3. c)

EL

3 e

TABLE 13 Functional Group Distributions Derived by Quantitative "P-NMR Analysis for Lignins" (Figures in Parentheses are Averages Obtained During the International Round Robin Analytical Effort) [483]

Lignin sample Steam explosion (aspen) Steam explosion (yellow poplar) Ball-milled enzyme (cottonwood) Alcell'" organosolv (mixed hardwoods) Indulin"' AT kraft (mixed softwoods)

-COOH 0.04

Syringyl -OH

0.14

0.27

0.00

0.15

0.24

0.00

0.04

0.05

0.06

0.25

0.45

0.06

0.2 1

0.00

Total phenolic -OH

0.4Zh (0.45) 0.4gh (0.59) 0.15' (0.18) 0.48 (0.59) 0.57* (0.67)

2

-.N4

z

Lignin functional group (mol/C,)

Guaiacyl -OH

m

Alpha-OH in p-0-4 structures Primary -OH

Eritrho

Threo

0.44

0.24

0.13

Total P-0-4-OH

0.37

2 Total hydroxyl content

9

6

1.27 ( 1.26)

0.36

0.12

0.10

0.22

0.52

0.39

0.14

0.53

0.32

0.06

0.10

0.16

0.4 1

0.06

0.08

0. I4

1.06 ( I .20) 1.20 ( 1.47) I .24 (1.20) 1.18 ( I .23)

"Average of four quantitative experiments. hlncludes condensed biphenolic structures 0.09 mol/C,,. 'Includes condensed biphenolic structures. 0.04 mol/C,, and p-hydroxyphenyl structures 0.02 mol/C,. 'Ilncludes condensed biphenolic structures 0.36 mol/C,.

w

P

ul

Baeza and Freer

346

136

138

126

128 134 130

132

PPm FIGURE 35 Quantitative"P-NMRspectrum phosphitylated. (From Ref. 483.)

of steam explosion lignin produced from aspen.

lignin structures of the secondary wall, but the high syringyl content of stage I1 lignin may be the reason the guaiacyl units participate more extensively in condensation reactions. Condensationreactions were found to bepredominant in thesecondformic acid digestion stage, due to the reactivity of the benzylic carbon and the formation of interand intramolecular C,-C, and C,-Cc, carbon-carbon bonds. The results obtained by "PNMR have an excellentquantitativecorrelation with thosefromoxidativedegradation using potassium permanganate and hydrogen peroxide. JiangandArgyropoulos [487] used "P-NMR to quantified parcr-hydroxyphenyls, catechols, guaiacyls, and phenols bearing C, and C, substituents, after a Mannich reaction of softwood kraft lignins and models with piperidine. 19 FNMR Spectroscopy. Barrelle [488-4901 developeda method using "F-NMR for the quantitative determination of OH groups (hydroxylic, phenolic, and carboxylic). The fluorobenzoate lignin and the lignin model compound derivatives are obtained using 2- or 4-fluorobenzyl chlorides or the respective anhydrides. This technique permits one to distinguish between a guaiacyl lignin and a guaiacyl-syringil lignin, to determine the syryngyVguaiacyl ratio of structural units with phenolic groups, and to approximately determine the (a-C=O) content. In Table 14 appear the chemical shifts for lignin model compounds and in Fig. 36 the spectrum of a fluorobenzylatedorganosolvlignin.Barrelleassignedthesepeaks to ketocompounds,G-compounds, and S-compounds,aftercomparison with the spectrum obtained from a mixture of eight model compounds 14901. cl. Molecular Wkight m d Molecular Weight Distribution. Lignin is a polymer with a wide range of molecular weights and when it is removed from wood, the original value of the molecular weight is affected. There are a multiplicity of isolation procedures, giving lignins with different characteristics. MWLs are considered as preparations i n which minimal changes have occurred, but these depend on the material and the milling procedure 149 1.4931. Various techniques are available for the determination of MW and MWD of lignin samples, as described before for cellulose and hemicelluloses: SEC, viscometry. osmometry, light scnttcring, and ultracelltrifllgation. SEC has been recently and greatly cnriched

TABLE 14

a-(C=O) compounds I 2

3

AS G-S(C=O)

6"

FBzl-0-G-CHO 25341 25 07

FBzl-0-S-CHO FBzI-O-G-CO-CH,

} j

0.27

25.25

4

FBLI-O-S-CO-CH~

25.02

5

FBLI-O-G-CO-CH~OG

25.28

6

FB/I-O-S-CO-CH20G

7

FBz~-O-G-CO-CHOG

25.08

CH,OH FBzl-0-S-CO-CHOG

I

CH,OH "Two values for two diastereoisometric fonns:

25.20

i

G = &OM"

a-(C-OR)

compounds

S

A6 a-(C=O)-a-(C-OR)

FBzI-O-G-CH~OH FBZI-O-G--CH,OG FBzI-O-G-CH~OS FBzI-O-S-CH,OH

24.77 24.92 24.93 24.61

0.57 0.42 0.41

FBzI-O-G-CHOH-CH1

24.60

0.46 0.65

FBzl-0-S-CHOH-CH,

24.58

0.44

FBzl-0-G-CHOH-CHZOG FBzI-O-G-CHOG-CH~OG FBzI-O-S-CHOH-CH~OG

24.89 24.85 24.65

0.39 0.43 0.43

FBz~-O-G-CHOH-CHOG

24.87"

0.53

24.90 24.65"

0.50 0.55

24.62

0.58

0.23

0.20

25.40

I

8

3m

Assignment of Signals in the '"F-NMR Spectra of Model Compounds [490]

I

0.20

CH,OH FBzl-0-S-CHOH-CHOG

I

CHzOH

s:''fle

z

D

rr

G

2!

s s d 0

P

348

Baeza and Freer

Correlation: A:a-ketocompounds;B:G-compounds;C:S-compounds (with a-COR) FIGURE 36

"'F-NMR spectrum of

B

fluorobenzylated organosolv lignin. (From Ref. 490.)

by the advent of the real-time differential viscometer (DV) and LALLS photometer. However, due to the characteristics of lignins in solution, mainly their low viscosity, the behavior during the determination of molecular weights is quite different from that of polysaccharides or synthetic polymers 14931. A general review with 66 referenceson the use of SEC in thedetermination of MWD in lignin derivatives was given by Himmel et al. [4941. MWD of lignin preparations by SEC with viscometric detectors and ultracentrifuge sedimentation equilibrium analysis is illustrated. A reliable method for estimating MW and MWD requires a suitable solvent, in which the aggregation (solute-solute interactions), solute-solvent, and solute-packing material interactions are minimized, in addition to solubilizing lignins over a wide range of mo-

racterization Chemical

of Wood

349

lecular weights. Many lignins are not soluble in suitable organic solvents or water, but lignin derivatives are soluble in solvents such as THF and DMF. By derivatization, the adsorption and association caused by hydroxyl groups are diminished. Acetylated [492,495-5011, propylated [502], and sylanated [495] derivatives have been used. SEC permeation chromatograms of MWL have shown multimodal [491,503,504] or single symmetric distributions [354,492], depending on the experimental conditions. AS a rule, SEC/THF of acetylated lignins does not show exceptional irregularities. In DMF and DMF-THF mixes, multimodal elution behavior was observed, which is indicative of associative phenomena. These can be eliminated by the addition of lithium chloride to the DMF [503,504]. The effects of the associative/dissociative processes in nonaqueous andl or aqueous mediaon the MW and MWD of kraft lignins [505-5071 and organosolv lignins [504] have been discussed. The early publications dealing with MW and MWD of lignins using GPC employed soft gel dextran columns [508,509]. More recently, cross-linked gels were employed. Several studies have been conducted on GPC on dextran gels columns using DMF, DMSO, or dioxane-water mixtures as solvents [508-5151 and on agarose gels (Sepharose CL) [354,5 131. Polystyrene-divinylbenzene copolymer gels columns (e.g.. p-Styragel, Waters Associates; p-Spherogel, Beckman Instruments; PLgel columns, Polymer Labs) have been used to conductHPSEC of lignins allowingconvenientrecording of MWD[354, 496,498,501,5171. Column calibration has been carried out by using dehydrogenation polymers (DHP) of coniferyl alcohol [509]. Himmel et al. [499] have demonstrated that a series of commercially available standards as well as low-molecular-weight lignins all fit universal calibration. FaixandBeinhoff[500]discussed SEC column calibration withpolystyrene standards, lignin fractions,and lignin-like modelcompounds.TheMWdeterminations calibrated by lignins can deviate from those obtainedby polystyrene calibration depending on the polydispersity and MW. This can be explained mainly by the fact that lignin molecules present a spherical shape, being more densely packed in solution than the flexible chains of polystyrene. Also, as can be expected, due to the chemical structure, a higher adsorption affinity of lignin to the gel contributes to longer retention times at the Same nominal MW as the polystyrene. By derivatization, the adsorption and association caused by hydroxyl groups are diminished. In Table 15, average molecular weights of different lignins for the underivatized and derivatized samples are shown [495].

TABLE 15 AverageMolecularWeightsandPolydispersities

of LigninSamples 1495)

PS compounds calibration calibration Model

Sample“ Birch (EXWL)

M,,M,, U

970

a

1,440

U

2,160 770 850 1,340 460

a

1 ,o I

S

Poplar U (EXWL) a ’i

Pine (kraft)

O

1,120

M,,

/M,,

4.880 1,180 5.03 9,22,000 10 6.38 12.110 2,530 5 .60 2,280 930 2.97 3,330 1.300 3.89 5 ,200 1,690 3.88 2,330 540 5.07 7,590 1,520 7.54 7.480 1,480 6.69

“Key: U. underivatized; a, acetylated;

S,

sylylatcd

M,,

M,.

M,. IM.j

6,640 7,370 9,350 3,000 3,250 4,600 3,090 6,240 6, I 50

5.64 3.68 3.70 3.23 2.50 2.72 5.70 4.09 4.17

I

Baeza and Freer

350

HPSECofacetylated alkaline-extracted, steam-explodedaspen lignin (AESEAL), MWL from aspen (MWAL), and organosolv black cottonwood lignins (OSBCL) on styrene-divinylbenzenecopolymergels as a function of theorganic elution systemwere investigated. Neat and mixed THF and DMF solvents, and DMF in the presence of 0.1 M LiBr were assayed. Polystyrene, Igepal polymers, and lignin model compounds were used as calibration standards [498]. In THF the elution profile obtained is unimodal, but in the mixed solvent (THF/DMF) and in DMF a multimodal distribution can be distinguished for all acetylated lignin samples investigated. The associative effects in solvent systems containing DMF were found to be a function of the history and nature of the lignin sample. The addition of 0.1 M LiBr to DMF brings the shape of the elution profiles of OSBCL and MWAL close to those observed in THF, but AESEAL still exhibits multimodal behavior, indicating that the IigninAignin associative interaction still persists. Similar behavior was observed in THF/DMF mixtures. Deconvolution of the elution profile of OSBCL was carried out to the lower-molecular-weight portion of the chromatogram. After deconvolution, nine well-resolved peaks were obtained at 164, 209, 263, 31 6, 372, 426, S 12, 602, and 728. The three first peaks were attributed to monomers, the next four to dimers, and the last two to mixture of trimers. HPSEC/DV has proved to be a reliable and convenient method for absolute molecular weight determination of lignin derivatives. Glasser et al. [50 11 report results obtained for several commercial and semicommercial lignins from hardwood, softwood, and cane bagasse, isolated by the kraft or organosolv pulping, or by steam explosion/autohydrolysis. The acetylated lignins were dissolved in THF and polystyrene molecular weight standards wereusedforcalibration.Absolutemolecularweightvaluesobtained by GPC/DV of hydroxypropylatedligninswerecomparedwith the number-averagemolecularweights obtained by vapor-phase osmometry (VPO), verifying the validity ofthe universal calibration [SISI. HPSEC/DV of fractions from preparative GPC of aspen acetylated lignins and unfractionated samples were analyzed by universal calibration (4991. The summation of the MWD of the individual fractions lead to values of MWD similar to those found for the unfractionated parent sample. MWD by GPCLALLS has been also employed to avoid calibration problems [519], but corrections for fluorescence, light absorption and polarization complicate these results. Number-average molecular weights of organosolv lignins were determined by vaporphaseosmometry (VPO), using THF as a solventandbenzil(diphenylethanedione)as calibration standard. The MWD were obtained by HPSEC, using THF as solvent and two different column sets: Plgel (10 pm, 30 cm length, 7.8 in. I.D.) 10" + SO0 + 100 (1) and 10' + 10' + S00 100 (11), using in both cases polystyrene standards and ethyl benzene for calibration [520).The MWD obtained from both column sets were different mainly in the higher-molecular-weight region, consequently M,!.were also different. The M,, values obtained by VPO are i n agreement with those by HPSEC with both column sets. VPO,LALLS,andSECmethods for weightdeterminationwere investigated and their application to lignins discussed 15191. M,,, M,,., and MWD were determined by VPO, LALLS,andSEC, respectively, for dioxane lignin (sprucewood), alkali lignin (black cottonwood). and organosolv lignins (black cottonwood, hornbeam chips). The solvent and temperature effects 011 M,, values were determined for different fractions of spruce dioxane lignin and black cotton alkali lignin. Nonsignificant differences were observed, indicating that the associative phenomenon is not relevant i n the case of these typical samples. The M,, weredeterminedbefore and after acetylation of spruce alkali lignin fractions.The observed values after acetylation were only slightly higher than those expected. M,, values for organosolv lignin fractions and alkali lignin fractions fromblackcottonwoodwere

+

aracterization Chemical

of Wood

351

determined by LALLS, obtaining values in the range 1,500-74,000 and 4,700-55,000, respectively. Determination of MWD of kraft lignin and lignin sulfonates have been carried out [521]. For kraft lignins, columns of Sephadex 50 were used and the elution was performed with 0.5 M NaOH solutions and lignin sulfonates on elution through Sephadex G-50, G75, and Sephacryl S-300 using water or 0.5 M NaCl buffered at pH 8 as the eluent. Protein and lignin sulfonate fractions with known molecular masses were used to calibrate the columns. Lignosulfonates and kraft lignin have been fractionated according to their polarities by reversed-phase liquid chromatography [522]. High-molar-mass lignosulfonates and kraft lignin are fractionated on the basis of molar mass, with the highest-molar-mass compounds eluted last. Kraft lignin was fractionated into hydrophilic (fraction I) and less hydrophilic (fractions 11-IV) compounds (Fig. 37a). Thefraction I contains low-molecularmass compounds and a small amount of high-molar-mass lignin derivatives. These derivatives are polar and some are bound to carbohydrates. The molecular masses of the compounds present in the hydrophobic fractions increase in the order 11,111, and IV (Fig. 37b). Strongly hydrophilic lignin-carbohydrate compounds can be separated from virtually carbohydrate-free lignin. LignosulfonateswereanalyzedusingaqueousHPSECwith TSK

90 120 Retention timdmin

0.8-

0.2

0 I

0.4 I

5000

I

3000

0.6

1.4 0.8 1.2 1.0

Relative retention volume I I 1500 loo0 Molar mass

FIGURE 37 (a) Fractionation of pinekraftlignin by preparativereversed-phaseliquidchromatography. (b) Molecular mass distribution of pine kraft lignin. (From Ref. 522.)

and

352

Baeza

Freer

G3000SWcolumns,andcombinedHPSECandpyrolysis-gaschromatographic-mass spectrometric study of lignosulfonates in pulp mill effluents was carried out [523]. The MWDs show distinct differences between the various lignosulfonate samples and can be used to characterize structural modifications. Sodium lignosulphonate standard, prepared under mild conditions,is a relatively polydispersepolymer with a large proportion of preserved phenylpropaneunits. Lignosulphonates discharged by pulpmills are more monodisperse macromolecules, showing lower PD values than the standard, and they also present modifications to a greater extent. Macromolecular characteristics of alkali lignins from western hemlock wood were reported by Dolk et al. [524]. Wood thin platelets of uniform thickness (0.4 mm), extracted exhaustively with amixture of ethanol-benzene (1:2 v/v), were delignified with 1.0 N aqueous sodium hydroxide solution, and the dissolved lignins were fractionated into acidinsoluble (AIL) and acid-soluble lignin (ASL) fractions. For the AIL, M,, and M,? were determined by VPO and LALLS, respectively, and for ASL, the two MWs were estimated by SEC. For AIL, the M,, values are relatively small, but they show a steady increase as the delignification process advances; values of 874 and 1804 were obtained after 10 and 420 min reaction time. With M,c values a more striking increase occurs (2,318 and 20,685 after 10 and 420 min reaction time), giving rise to a significant increase of the polydispersity of the lignins as the delignification proceeds. The values of M,, and M,, for the total ASL material using the SEC pattern with column calibration were around 400 and 630, respectively. The SEC pattern for ASL shows a number of peaks, andthe values were assigned to the modes of these peaks (230, 440, 700, and 1020) suggest the presence of monomeric, dimeric, trimeric, and some larger fragments of lignin. The condensation of a lignin has been determined by analyzing the molecular weight distribution of the lignin thioacidolysis products [525].The MWD can be used to give a measure of reactive to unreactive or condensed bonds in the lignin. A higher proportion of unreactive linkages in the lignin conduces to higher-molecular-weight fractions in the thioacidolysis products. The MWD was analyzed by HPSEC on polystyrene columns with THF as eluent. There was no relationship between MW of the thioacidolysis products and the MW of the starting lignin. The LALLS methodology and the correction procedure for optical effects (fluorescence, absorption, and optical anisotropy) of lignin solution for the determination of MW of kraft lignin havebeendescribed by DongandFricke[526].Based on the absolute molecular weight characterized with LALLS, the Kuhn-Mark-Howuwink-Sakutara (KMHS) equation was developed, providing the KMHS constants for the kraft lignin. In Table 16, KMHS parameters of kraft lignins in DMF and 0.5 N NaOH are shown [527]. Based on these results, it was considered that the lignin molecules in solution are approximately spherical and only slightly solvated.

TABLE 16 Kuhn-Mark-Howuwink-Sakutara (KMHS) of Kraft Lignin (5271 Solvent

DMF DMF 0.5 N NaOH

Temp. (K)

3 18.2 0.13 350.7 303.2

K

2.5 100 1.8895 0.23 0.5 165

(Y

0.1 1

Chemical Characterizationof Wood VII.

353

EXTRACTIVES

The extraneous components of wood are substances which are not considered as essential structural parts of the cell wall or middle lamella. Unlike cellulose, hemicelluloses, and lignin, the extraneous components are nonpolymeric (except pectins and condensed tannins) andmay be separated from the insoluble cell wall materials by their solubility in water or organic solvents. They cover a wide range of chemical compounds even though they generally represent only a small part of the wood. Because most of the extraneous compounds are commonly isolated from wood by solvent extraction, they are called extractives. Strictly speaking, the two terms are not synonymous. However, in most cases the distinction between extractives and extraneous materials is academic. Knowledge of the composition and amount of extractives in wood is of great interest. Many differences in the properties of woods are determined by the composition of the extractives. Many woods contain extractives which aretoxic to fungi, bacteria, and termites [528,529]. Other extractives can add color and odor to wood, accent the grain pattern, and enhance strength properties [530]. On the other hand, extractives can cause some negative or undesirable properties. For example, the presence of extractives results in corrosion of metals in contactwithwood[531], inhibition of setting of concrete, glue, and finishes [532], contribution to color reversion in pulps, pitch problems during papermaking [533], etc.Furthermore,theextractiveshave industrial importance.Forexample, tall oil and turpentine have been used traditionally in cosmetics, paints, and varnishes, and have been proposed as a source for diesel fuel, and energy for steam and electric power generation [5341. The composition and the amount of the extractives are dependent on the wood species, within and among trees, tree age, and the environmental conditions under which they grew. Details of the composition of the extractives are given in the literature [531,5355381. The extractives are sometimes characterized into chemical classes that have a direct influence on the pulping process, namely, saponifiables andunsaponifiables [539-5421. Saponifiables (fatty acids, resin acids, some steryl esters, and glycerides) are considered to becompounds that formsolublesoapsunder alkaline conditions.Unsaponifiables (waxes, some steryl esters, diterpene alcohol and aldehydes, sterols, triterpene alcohols, and fatty alcohols) do not form soaps and have a tendency to deposit and cause pitch problems [543-5451. The techniques of analysis of extractives involve the isolation of components (extraction, distillation of volatiles, chemicalorchromatographicseparation)and analysis (GC, LC, G C M S , NMR, IR, etc.). A large variety of analytical techniques are used in the analysis of extractives, but the methods vary depending, among other things, on the type of information required from the analysis. No single solvent is capable of removing all the substances considered as extractives, and no single sequence is applicable to all woods. Different schemes of separation and sequences can be found in the literature [546,547]. For example, according to the scheme of Kurth [548,549], successive steam distillation and extractions with ether, ethanol, and water remove different types of extractives. The general scheme outlined by Kurth is of great value as a general guide. Figure 38 gives an overview of the groups of extractives from an analytical standpoint,withexamples of subgroupsandindividualcompounds [SO]. Many other extraction sequences have been employed [g], the selection depending on the safety, reproducibility, and the desired extractives of the material being examined. In the isolation of extractives, normally halogenated compounds (mutagenic compounds)

I

Baeza and Freer

racterization Chemical

of Wood

355

and aromatic compounds, especially benzene which is carcinogenic, are used as solvents. Laboratories are advised to avoid the use of these solvents. Acetone has been found to be asuitablesolventfor the analysis of extractives in wood, pulp, andpapersamples [5,542,55 1-5541. To reduce the hazards associated with the use of large amounts of potentially harmful organic solvents, together with costs and environmental dangers of wastedisposaland emission of the solvent into the environment during sample concentration, some alternative methods of extraction havebeen applied. Areview of some of the modem analytical methods that can be used in the analysis of extractives from wood and pulp is available [555].Two broad aspects are discussed, sample isolation procedures and analytical procedures. Furthermore, to reduce or eliminate the use of toxic organic solvents, the new proceduresdiscussed are simpler than the traditional methods,whichincludemultistep procedures. Among the techniques that have been used for extraction of wood and wood products are Soxtec extraction, gas-phase extraction, which includes headspace sampling and supercritical fluid extraction (SFE), and sorbent extraction, being solid-phase extraction (SPE), the most commonly used sorbent extraction technique [556]. The Soxtec method is based on Soxhlet and Goldfisch extractions.It consists of three steps: (1) boiling-initial extraction, in which the sample is completely immersed in the boiling solvent; (2) rinsing-condensed solvent washes last traces of soluble matter from the sample; and (3) solvent recovery-solvent is evaporated, condensed, and collected. A scheme of the operation and features of Soxtec extraction is shown in Fig. 39. By using Soxtec extraction, the extraction canbeperformedfasterand the solventvolumes are about 3.5 times less than the traditional solvent extractions using Soxhlet extractors[5555571. However, the Soxtec values tend to be lower than those of Soxhlet [558,559], due probably to an inefficient washing during the rinsing stage. This problem can be overcome by performing a second extraction of the sample with fresh solvent [558]. Solid-phase extraction (SPE) methods also require only small quantities of organic solventand are rapid[553,560,561]. The sorbent is packed in disposablecolumnsor cartridges. SPE is especially suited for sample preparation of diverse compounds such as extractives in deposit and wood resin in pulp. Supercritical fluid extraction (SFE) [563] has recently been used in the separation of extractives in wood and woodproducts[552,562].Because supercritical fluids possess bothgaslikemass transfer and liquidlike solvatingcharacteristics, SFE is an attractive solvent-free sample preparation technique. It is rapid and simple, but it requires heavy equipment for on-site field analysis. The SFE extraction method has Some distinct advantages over others: thermally unstable compounds are undamaged, extraction times can be short, and nontoxic solvents can be used [555].

A.

VolatileMaterials

The methods for determination of volatile materials (collectively called essential oils or volatile oils) are based on steam distillation. These components include cyclic hydrocarbons(terpenesand terpenoids), aliphatic hydrocarbons,phenols,alcohols, ethers, aldehydes, and lactones [564]. In general, the amount of essential oils in hardwoods is negligible, but they are present in considerable amounts in softwoods. The essential oils of pines are called turpentine, which consists primarily of monoterpenes. The total amount of volatile oils can be found by loss of weight of the wood sample after steam distillation [565]. Thismethod is not suitable when the amount of volatile

356

Baeza and Freer

aracterization Chemical

of Wood

357

materials is small. The volatile oils can also be determined after recovering by condensation techniques or the adsorption of vapor on charcoal or other suitable materials The water-insoluble volatile oils can be determined by using distillation equipment with a suitably calibrated trap. The wood is distilled until no more oil comes over, and the volume is read directly. The amount of volatile oils can also be determined by extraction of a team distillate with ethyl ether. Fresh wood meal is steam-distilled, normally for 2 h, although the time can be extended 3-4 h to improve the recovery of high-boiling-point compounds. The distillate is extracted with ether. The organic solution is dried and the solvent is removed to yield the volatile oils. Details of the procedure are given elsewhere [567,568]. Distillation may be conducted with a caustic solution in order to prevent degradation and isomerization [569,570]. The basic solution may also contain ethylene glycol [569]. To minimize bumping during distillation, ground wood is placed in a cheesecloth bag that is placed on a supporting metal screen in a resin kettle [571]. The yield of volatile oilshasbeendetermined by gaschromatographyusing an internal standard 15721. A known amount of tetradecane as an internal standard was added to the ground wood sample prior to distillation. Yield was determined by GC recording the proportion of tetradecane to terpene components in a portion of the distillate. With this method the turpentine yields are about 5 % greater than those obtained by the volumetric procedure. The characterization of the volatile oils is commonly determined by GC and GCMS [570,573,574]. Extraction of volatile compoundswith supercritical carbondioxideandhot-water distillation was conducted for coniferous woods, and the extracts were analyzed by GC and GC-MS 15621. The yields by SFE for 30 min were greater than those by hot-water distillation for 8 h. For example, by SFE of western red cedar and Douglas fir, the yields were0.61% and 0.95%(0.d.b.)(at 300 kgf/cm' and 40°C), respectively, while by hotwater distillation the yields were 0.12% and 0.07% (o.d.b.), i.e., 5.6 and 13.6 times more, respectively. The yields by SFE are dependent on the conditions and time of extraction. For example, the yields of extractives for 30 min were about 80% of those for 90 min. The effect of pressure on the yield is given in Fig. 40. The components of SFE extracts and essential oils by hot-water extraction were similar for some species, but they were quite different for western red cedar and Douglas fir woods. Headspace volatiles emitted from extracts of SFE using carbon dioxide and essential oils by hot-water distillation from seven species of woods were collected and analyzed by GC. They were compared with those of woods [575]. The composition of the three headspace(volatilesfromwood,SFEextracts,and essential oils)weredifferentfromeach other, and in general, a-pinene and/or P-pinene were the main compounds of each headspace volatiles. Another method for monitoring volatiles involved FT-IR measurements. Emission of terpenesfromchip piles wasdetermined[576].Themethodusesapolymer film wind tunnel and measures a-pinene, P-pinene, and y-3-carene on a semicontinuous basis with a detection limit of 1 mg/m3. The total hydrocarbonswasmonitored simultaneoLlsly by GC-FID.

B.

Extractives Soluble in Organic Solvents

The extractives that are soluble in organic solvents include resins and fatty acids and their esters,waxes,unsaponifiablesubstances,coloring matter, etc.Resin acids are tricyclic

358

Baeza and Freer 4.0

A

\

\

3.0

n

S

W

in

9 Q) 2.0

F

Hinoki 1.o

sugl Hinokiasunaro Alaska cedar W.redcedar I

I

I

300

200

100

Pressure (kgf/cm2) FIGURE 40 Effect of extraction pressures on yields (40°C.30 min). (From Ref. 562.)

diterpenoids and occur naturally in conifers. The major acid is often dehydroabietic acid (DHAA), although which acid is predominant depends on tissue type, age, and species. Several of the resin acids have two conjugated double bonds, and it is these which appear to be least resistant to chemical degradation. In softwood species such as Pinus radiara, free resin acids may comprise up to halfof the extractable organic compounds present [577], although the amount present varies with the age of the trees and the conditions under which they grow [578]. Anumber of standardprocedureshavebeenpublishedwhich are very similar in principle and procedure [579]. These include the TAPPI Standard, Method T-204 om-SS [2] and ASTM D 1107, D 1108, and D 1794 [3]. The general procedure is as follows. A sample of air-dried wood sawdust is weighed in an extraction thimble and placed in a Soxhlet extraction apparatus. The extraction is carried out with a suitable solvent for 4-8 h, having solvent siphons over at least 6 times per hour. The flask is removed from the apparatus and the solvent is partially evaporated to reduce the volume to 20-25 mL. The extract is transferred to a tared weighting disk by washing it with small amounts of fresh solvent. The solvent is evaporated to neardryness. The dish and contents are dried in an oven for 1 h at 105°C (1 15°C for ethanoltoluene), cooled in a desiccator, and weighed. The sample can be also dried to a constant weight in a vacuum oven at 60°C.

aracterization Chemical

of Wood

359

Maximum amounts of extractives are removed with ethanol-benzene ( 1 :2) (TAPPI Standard,MethodT-204om-88[2]).Extractionwithethanol-tolueneprovidessimilar amounts of materials as from ethanol-benzene, but has poor reproducibility. Dichloromethane extraction gives lower amounts of extractives. Aliphatic and aromatic hydrocarbonsolventstend not to extract allof the oleoresins[571].Ethylether is apreferred solvent for the study of pine extractives [571]. Volatile substances are largely lost during dryingof the extract. Hence,the extractives obtained from the above procedure correspond to nonvolatile extractives. Different solvent systems may be employed. A number of schemes for separating nonvolatileextractives into groups of components having similar properties havebeen outlined [9,579,580]. For example, a scheme for determining amounts of unsaponifiables and free and combined fatty and resin acids is shown in Fig. 41. Acetone has been found to be a suitable solvent for the analysis of extractives in wood pulp, and paper samples. A scheme for the identification and quantification of the components in the acetone extractives of wood and bark samples has been described by SitholC etal.[554].Theschemehasbeenapplied to aspenand involves: (1) sample

I

ETHANOL- BENZENE EXTRACTIVES

l

Carefully concentrate to dryness Extract with ether

a Neutrals

I

I

1

I

Add water-acldified Extract with ether

I

+ Glycerol

Unsaponifiables Extract withdil.Aq. NaOH

Selective esterification

FIGURE 41

Separation of nonvolatile extractives. (From Ref. 9.)

Freer 360

and

Baeza

preparation; (2) solvent extraction with acetone to determine the amount of extractives (freeze-dried samples were extracted with acetone in a Soxtec extraction apparatus, then the remaining soluble extract in the sample and thimble walls were rinsed with solvent, and the extractdriedusinganitrogenstream);(3) fractionation of the extractivesinto weak acids (pK,, > 5), strong acids (pK,, < 5 ) , and neutrals by ion-exchange chromatography using a DEAE-Sephadex ion-exchange procedure developed by Zinkel [560,561]; (4) derivatization and GC analysis using short columns to separate the components in the extractives (methylated); and (5) identification and quantitation of the extractives components by using a spreadsheet program. Figure 42 showsa chromatogram for a fresh wood sample. The peaks were assigned by using individual standards ranging from fatty acids to triglycerides. The use of a short column permits the elution of high-boiling-point fractions and fatty acid-methylatedesters in a reasonabletime(30min)andwithmuch better resolution than is obtained with packed columns. Acetone extracts of white spruce and trembling aspen woods and a series of kraft pulps, were analyzed by GC [544]. Chemical changes in wood resin during pulping performed over a range of initial effective alkali (EA) from 11 .O to 44.9 g/L, and the implications of the results in deresination and pitch control were discussed. The ratio of resin acids and wood fats (fatty acids and glycerides) to resin (sterol and steryl esters) is of great importance in the deresination of the pulps. For a complete dissolution the ratios werefound to be 2.6: 1 and 1.4: 1 for spruce and aspen pulps, respectively. Below this value, deresination is not complete and some insoluble components are not washed out of the pulp. Consequently, it can be predicted that poorer deresination will result in the aspen sample than in the spruce. More frequent occurrence of pitch problemsfrom the kraft pulping of aspen due to relatively higher amounts of neutral and unsaponifiable materials can be generally observed. With EA values in the range 25.0-40.0 g/L for spruce and 30.0-44.9 g/L for aspen, over 70% of the total acetone-extractable was removed during pulping, while at EA values below25.0 g/L and 30.0 g/L for spruce and aspen,respectively, the percentage of extractsincreased rapidly. This increasemight be due to incomplete removal of the lignin degradation products during cooking, which are later extracted with the other acetone-soluble compounds. Under the normal cooking conditions (EA concentration of -40 g& forspruceand -30 g/L for aspen), the glycerides are practically completely saponified into soaps. The concentrations of fatty acids, resin acids, and steryl esters in spruce, and fatty acids and steryl ester in aspen, are reduced compared to the original values in the woods. The steryl esters content is decreased considerably by their hydrolysis to sterols. At lower initial EA concentration, the total amount of extractives remaining in the pulps increases rapidly. The glycerides and steryl ester increase in both species. The content of fatty and resin acids in spruce, and fatty acids in aspen, increases towards its original value in the wood. An analysis of the acetone extractives of fresh trembling aspen (Populus tremuloides Michx.) wood was reported by Dunlop-Jones et al. [542]. Freeze-dried samples were extracted for 18 h with acetone. The dried acetone was dissolved in a diethyl ether-methanol-water mixture (89: 10:1 ) and fractionated into weak acids, strong acids, and neutrals, using a DEAE-Sephadex column. The neutral fraction was saponified and fractionated into saponifiables (acids) and neutral (unsaponifiable) using an ion-exchange column. The free acids and combined saponified fatty acids were analyzed as to their methyl esters by GC. Theunsaponifiableswere silylated with a mixture of N,O-bis(trimethylsily1)-trifluoroacetamide(BSTFA)andtrimethylchlorosilane (TMCS), andthenanalyzed by GC. The neutral componentswere identified by comparisonwithknownpurecompounds or by mass spectrometry. The ratio of saponifiables to unsaponifiables found for the fresh aspen

v)

r 0

>

W'

05

o.201

fD

3.

c)

m c U

0.151

0 0

P

cn

z

0

a cn W

E

/-

I I 0

8

I 5

I

I 10

I

I

15

I

I

20

RETENTION TIME, min FIGURE 42 Chromatogram (GC) of an acetone extraction of aspen wood. (From Ref. 554.)

I

I

25

r

1

30

8

362

Baeza and Freer

was about 2: 1, and this low value is a reason for its tendency to give pitch problems. Other workers [580,581] have obtained higher ratios, but the amount of seasoning of the woods analyzed has not been discussed. Seasoned wood chips exhibit fewer pitch problems, since during the storage, extractives undergo volatilization, enzymatic hydrolysis, and air oxidation, but long storage times promote microbial deterioration [582,583]. The storage effect is greater in chips than in roundwood. SPE was used to separate and quantify lipid classes in acetone extracts of wood and pulp [553]. A method for rapidly separating acetone wood or pulp extractives into five different classes was described. The acetone extracts were absorbed onto an aminopropylphase column and the recovery and quantification of different classes were carried out by eluting with a sequence of solvents. The samples were further analyzedby GC and HPLC. The method was applicable to both hardwoods and softwoods. Separation of the different lipid classes was highly reproducible. The five classes are (1) fatty acids (FA) and/or resin acids (RA), (2) steryl esters (SE)/waxes (W), (3) triglycerides (TG), (4) monoglycerides (MG), and (5) sterols (S)/diglycerides (DG)/fatty alcohols (Fal) (Fig. 43). The separation takes about 2 h per sample, with a recovery rate of 95-99%. This method is presented as an alternative to the traditional one used in the wood and paper industries, in which acidic compounds are separated from neutral by ion exchange. Neutral lipids are hydrolyzed and finally analyzed by GC [565,566]. Analysis of oriental beech (Fugus orientalis) wood fatty acids by supercritical acetone extraction (30 min, 240°C, 6.0-6.5 Kpa) was carried out and the results were compared with Soxhlet extraction [552]. While the yield of the Soxhlet extract was 2.5470, the yield of the SFE extract was 9.55% (dry wood basis). The fatty acids present in the extracts were separated by chemical and chromatographic methods and analyzed by GCMS. Among the fatty acids, from both Soxhlet and SFE extracts, linoleic acid was the major constituent, followed by linolenic and palmitic acids. Palmitic acid appears to be the main saturated fatty acid. The proportion of the dienoic fatty acid is lower in the SFE extracts due to the high temperature used in the extraction. The main resin and fatty acids (RAFA) of Pinus elliottii Engelm. were characterized before and after pulping (cooking liquor, methano1:water 80:20) [584]. The RAFA were saponifiedand/ormethylatedandcharacterized by GC andGC-MS.Determination of RAFA isolated from wood, pulping liquor, and pulp were carried out without fractionation of the extracts before the GCanalysis.The total resin acidcontentdoes not undergo quantitative changes after pulping, but instead qualitative modifications occur as a result of eitherisomerizationand/oroxidation reactions, leadingtoformation of abietic and dehydroabietic acids as the preferred end products. Selective removal of the extractives from the black liquor was obtained by using a liquid-liquid extraction with diethyl ether without affecting the solvent composition of the black liquor. A rapid spectrophotometric procedure for the determination of total resin and fatty acids (RAFA) in pulp and paper matrices was developed [585], which can be applied to wood chips, whitewaters and effluents. The method involves the following steps: ( 1 ) extraction of the fatty and resin acids from the matrix, (2) complex the free fatty and resin acids with copper(I1) ions to yield blue complexes, (3) extraction of the complexes with a solvent, and (4) measuring of the absorbance at 680 nm. The calibration curves were obtained using oleic acid as the standard. Compounds containing resin and fatty acids, such as glycerides, and metals salts, do not react with cupric ion, but the bound RAFAs can be determined by the difference between the total RFA value obtained before and after the hydrolysis of the extractives. This method is simple and can be used as a rapid procedure to estimate the amount of RAFAs in extractives, yielding results which agree with

363

Chemical Characterizationof Wood

Extract (TG, SE, FA, RA DG, S , MG, W, Fal)

4

+ Elh; j

CHC:,:?

Column 1

R A -

R A -

FA DG MG

S Fa1

tj, H +

H

W SE TG

Fraction A

Fraction B

Dry down Hexane

Dry down Ethyl acetate

(Save) PE

Fraction D

RA FA (Save)

Fa1

Fraction C

Dry down Ethyl acetate

& 6!: 0 (Save) TG

Fraction E

(Save)

Fraction F

(Save) MG

Fraction G

Legend: DG= diglycerides; FA= fatty acids;Fal= fatty alcohols;MG= monoglycerides; RA= resin acids; S= sterols; SE= steryl esters;TG= triglycerids;W= waxes. FIGURE 43 Elutionsequenceforseparating wood and pulp (SPE). (From Ref. 553.)

andisolatinglipid

classes from acetone extracts of

Freer364

and

Baeza

those obtained by chromatographic procedures. This method has a higher limit of detection and requires much larger volumesofsamplesthanthose used in chromatographic procedures. Wood extractives released during pulping or as by-products of the pulping process have been reviewed by SitholC [lo]. There is abundant literature on the analysis of by-products of the kraft pulping process. Proton NMR has been used to determine DHAA and total resin acids in rosin [586]. Asimple, rapid, andaccuratemethod for the quantitativedetermination of DHAA in commercially disproportionated rosin acids was developed and tested [587]. The method entails converting the acid into its methyl ester derivative before analysis by capillary GC with a FID detector, using methyl stereate as the internal standard. The method can also be applied to the quantification of DHAA in rosin or other rosin derivatives. The analysis of pulp mill process water, sediments, and fishes is very important from an environmental point of view. Investigations have been centered fundamentally on the study of organochloride compounds (which originate in the bleaching process with elementalchlorineorchlorinedioxide),butsomemethodshavebeendeveloped for the determination of wood extractives. Resin acids are toxins to fish at very low concentrations ( 1 -2 mg/L) [588,589]. An analytical procedure for the rapid determination of fatty acids, resin acids, and triterpenoic components in pulp mill process water was published by Backa et al. [590]. The extractives were isolated from alkaline aqueous samples using reversed-phase methods. The highest yields of extraction were achieved when an octadecyl phase, C,,, was used (C,-C,, reversed-phase chain lengths were tested). The cations of the adsorbed acid salts are exchanged in situ for quaternary ammonium ions. The acids are methylated by thermal decomposition of the quaternary salts in the injector of the gas chromatograph. The analysis scheme for determining extractives is: sample application + reversed-phase adsorption + rinsing with alkali -+ ion pairing -+ elution + pyrolytic methylation and gas chromatographic separation. A chromatogram of extractives isolated from a kraft black liquor is shown in Fig. 44. Resin acids in effluents, river waters, and sediments from a paper mill from Australia were determined by HPLC and GC (FID and MS) [591]. The resin acids in effluent were extracted by passage through a C,, cartridge at pH 9 and determined by HPLC and GC as their 7-methoxycoumarin-4-yl and 7-acetoxycoumarin-4-yl esters. The sediments were extracted with acetone, the extracts were dried, dissolved in water at pH 11 (KOH), and loaded onto a C,, column prepared as for the water extraction. The resin acid, eluted with acetone, was derivatized with diazomethane and analyzed by GC. The results confirmed the presence of resin acids derivedfrom the paper mills. The major resin acids in the effluent, water, and sediment samples were dehydroabietic, palustric, abietic, and pimaric acids. Smaller amounts of isopimaric, neoabietic, and sanderacopimaric acids were also found. The study concluded that dehydroabietic acid could be used as a tracer for organic matter derived from the paper mill. Wood extractives often cause problems in pulping and papermaking. Not only are they responsible for the formation of pitch deposits in the process system or in the pulp, they may affect the quality of the product.Successfulpapermaking often requires that compounds which cause pitch problems be removed from the system or otherwise neutralized. Rapid, accurate, andsensitivemethods of analysis are neededto classify and quantify the resin components. Knowledge of the extractives of the wood is important for the implications for deresination and pitch control. Methods for separating and identifying components of wood pitch have been reported [543-545,592-5951. They can be deter-

Chemical Characterizationof Wood

Fatty acids

-

365

Resin acids

I

Triterpenoids

r I

l 10

0

20

Time, min FIGURE 44

A gas chromatogram of extractives (Kraft black liquor). (From Ref. 590.)

mined by pyrolysis, gas chromatography/mass spectrometry with on-column methylation of the extractive components. This technique permits distinguishing woodresin from other non-wood resin extractives [596]. Detailed analysis of some of the organicallysoluble fractions of the deposits indicated that they are mostly neutral and unsaponifiable materials, with only a small percentage of resin and fatty acids being present. Extractives in papermaking process waters were determined by size-exclusion chromatography/tandem mass spectrometry ( S E C N S N S ) [597]. Also detected were a lignan and its fragmentation by chemical ionization MS/MS mode. The M S N S technique involves coupling one mass spectrometerto a second. The first spectrometer serves to isolate the molecular ions of various components of a mixture; these ions are introduced one at a time into the second mass spectrometer, where they are fragmented to give a series of mass spectra, one for each molecular ion produced in the first.

C.

PhenolicExtractives

Phenolic compounds possess free phenolic functional groups in their structure. Extractable phenolic compounds from wood, bark, and foliage range in complexity from simple phenolics (for example, vanillin) to polymeric condensed tannins. In some woods the amount is small. It is usually high in barks and foliage. Thephenolicextractivesaresoluble in the morepolarsolventssuchasacetone, alcohol, or water, and they are also soluble in aqueoussodiumhydroxideandsodium carbonate solutions. Some of the compounds of low molecular weight appear among the volatile components. The phenolic compounds constitute a heterogeneous class of compounds, which may be divided into the following groups [598].

366

Baeza and Freer

1. Hydrolyzable tannins, which are not very common in wood. 2. Flavonoids, which are polyphenols with a C,C,C, carbon skeleton. Typical representatives ofthisclass are 5,7-dihydroxyflavoneanddihydroquercetin. The polymeric flavonoids are called condensed tannins. 3. Lignans, which are formed by oxidative coupling of two phenylpropane units, e.g., pinoresinol and syringaresinol. 4. Stilbene derivatives, which are very reactive compounds, e.g., pinosylvin, present in Pinus species. 5. Tropolones,which are characterized as anunsaturatedseven-membercarbon ring, typically present in cedars.Forexample, 0-,p-, and y-thujaplicin have been isolated from western red cedar heartwood. The phenolic extractives are normally isolated from dried, ground, preextractedwood with petroleum ether [599,600] or benzene [601] to remove resinous materials. They are also extractable withethanol[602-6051,acetone[606],acetone-water[599,600,602605,6071, water [608], and dilute caustic solutions. SFE with acetone, THF, dioxane, and toluene have also been used [609]. Resolution of the complex mixture of phenolic substances obtained by total extractives or fractions obtained after selective separations requires the application of different techniques, such as selective solubility, chromatography, ionexchange,countercurrentliquid-liquidextractions,andformation of derivatives or reaction products [610]. Paper chromatography and thin-layer chromatography (TLC) have been widely used to monitor separations. After isolation of the phenolic compounds, they are characterized by various analytical techniques: GC [600,609,61 l], MS [5991, UV, IR, and NMR [612].

D. Water-Soluble Components The water-soluble substances are those which are dissolved by cold or hot water. They include water-soluble carbohydrates, some organic acids, many of the phenolic materials, and some inorganic constituents. Some of these are also soluble in organic solvents, and the amount obtained in the water fractions depends on the previous solvent extractions. Subsequent separations are necessary before identification. The isolation and characterization of the different fractions are given by Browning [613]. Extractable carbohydrates of wood have been analyzed by different very well established techniques, such as "C-NMR, HPLC, and GC. These were used to characterize the extractable carbohydrates for Norway spruce trees growing in different SO,-polluted sites [614]. Samples representingtypical growth periods were selected. Each of them was milled to a particle size of 0.25 mm, with cooling to avoid thermal decomposition, dried to a moisture content of 13% relative to the dry weight, and stored at -20°C. Samples were extracted with cold water, 1 g in 20 mL and 1 g in 100 mL for HPLC and NMR analysis, respectively. For GC analysis, the samples were treated with TFA (10 mg of wood with 500 p L of 2M TFA) for 4 h at 100°C. The hydrolyzates were dried, washed twice with methanol, reduced overnight with NaBH4. After neutralization, drying and washing, the samples were neutralyzed, acetylated, and analyzed for GC using 2-deoxy-glucose as an internal standard. All the samples contained carbohydrates composed of glucose, mannose, galactose, xylose, and arabinose, and small amounts of rhamose. A higher content of xylose andarabinosewasfound in the heartwoodregionthan in sapwoodofeach tree. The average content of both glucose and fructose was found to be about 3% higher in the

haracterization Chemical

of Wood

367

highly stressed trees, and the content of free D-fructose was found to be higher than that of free D-glucose in trees from the less damaged forest site.

VIII.

INORGANIC MATERIAL

The inorganic part of wood is analyzed as ash by incineration of the organic matter. The inorganicmaterialscomprisefrom0.1% to 0.5% ofoven-dryweight of wood in the temperature zones and up to 3-4% in tropical woods. There are numerous reports on the ash content of various woods [615]. The main components found in wood ash are potassium, calcium, magnesium, sodium, iron, silica, sulfate, phosphate, chloride, and carbonate. Trace levels of many other elements have been detected. The amounts and types of inorganic components depend on the soil in which the tree has grown, the fertilizers used, air pollution, and on preservatives applied to lumber. The components are not distributed uniformly through the tissue.

A.

Determination of Ash

The determination of ash is always accomplished by incineration to remove the organic matter. Some loss of volatile components, such as alkali metal chlorides and ammonia salts, may occur. Procedures for ash determination in wood are given by Browning [616], TAPPI Test Methods T 211 om-93 [2], ASTM D 1102 [3], and CPPA G. 10 [4]. All of these utilize different temperatures. The value specified in the procedure for ash determination in wood, pulp, paper, and paperboard given by TAPPI Test Method T 2 11 om 93 [2] is 525°C. However, the user must specify the temperature used in order to present accurate results and desired information. Combustion at 900°C is useful when an understanding of the noncellulosic materials present in the sample is required (TAPPI T 41 3 t21). More reproducible and somewhat higher values are obtained from the sulfate ash determination. The inorganic salts are converted to nonvolatile sulfates by adding sulfuric acid before the ignition is completed. The general procedure is as follows: wood is heated at a low temperature until most of the volatile materials are removed and a carbonaceous residue remains. Some drops of 50% sulfuric acid are added, and the crucible is heated until excess sulfuric acid is fumed off, and the ignition is completed at 700-800°C [616].

B.

Determination of Elements

In studies related to the growth, metabolism, and feeding of trees, it is important to know the concentration of a number of elements in various parts of the trees. Analysis for trace and major elements in solid samples generally requires decomposition of the organic matter followed by dissolution to give a solution for subsequent analytical determination. The decomposition may be achieved by various classical or recently developed procedures. each of which has particular advantages and disadvantages [617]. The analysis of adigestedsamplegenerally is carried out by atomicabsorption spectrometry (AA), inductive coupled plasma emission spectrometry (ICP-AES), electrochemical, colorimetric. gravimetric methods. and more recently, ion chromatography [61 S]. There are twogeneralprocedures for destroyingorganic matter: dryashingandwet digestion.

368

Baeza and Freer

Dry ashing is simple, but some elements are lost in the process. The ash from wood is used directly for analysis by arc emissionspectrography,or it is dissolved in dilute hydrochloric acid or nitric acid for the analysis, generally by ICP-AES or AA. An alternative to the muffle furnace is the low-temperature plasma. Ash content of loblolly pine was found to be higher when it was determined by the plasma method (0.335%) than that obtained by the muffle procedure (0.250%) [6191. Wet digestion is carried out by using mixtures of acids, including nitric and perchloric acid [620], as well as perchloric, nitric, and sulfuric acids [621]. Hydrogen peroxide has beenincluded in the digestionmixturetoavoid the useofperchloricacid[622-625]. Bomb digestion with hydrogen peroxide provides a relatively rapid means of decomposition that assures a complete recovery of elements in wood samples [626]. Microwave ovenshavebeen used for aciddigestion of manytypesof solids [627].Digestion by microwave has proved to be a rapid and reliable method for plant tissues [628,629], which may be applicable to wood samples. The digestion procedure depends on the matrix and the elements to be determined. Samples are digested with either nitric acid, hydrochloric acid, fluorhydric acid, or hydrogen peroxide in closed Teflon PFA vessels in a microwave oven. Organic matter can also be destroyed by combustion in a Schoniger flask or Parr bomb [630]. This technique is time-consuming, and only small samples can be burned. Loss of volatile elements is prevented because these devices are closed. Chlorine may be lost during ashing and wet digestions. Therefore, samples containing organically bound Cl are usually burned in a Schoniger flask or oxygen bomb [631]. The chloride formed is usually analyzed by potentiometric titration with silver nitrate. Nitrogen content is usually determined by Kjeldahl digestion, and sulfur in solution as sulfate is analyzed gravimetrically by precipitation with barium, or by ion chromatography [6 1 81. Analysis of tree leaves, bark, and wood by sequential ICP-AES for Ca, Mg, K, Na, P, Mn, Fe, Al, B, Cu, and Zn was carried out [6321. Samples were shredded and air-dried before being powdered and dried at 105°C. A 5-g sample of this material was ashed at 500°C. The residue was extracted with 5 mL of I : 1 hydrochloric acid, filtered and diluted with water up to SO mL. Leaf and bark extracts were diluted, while wood samples did not need additional dilution. The solutions were analyzed byICP. The choice of the line to be used for the analytical measurement was essentially determined by the sensitivity of the line, or lack thereof, and by the presence of spectral interference. After decomposition of hydrogen peroxide, which is used to treat wood, and electrothermal vaporization, analysis with ICP-AES of wood samples of red spruce and sugar maple were carried out [626]. Dried wood samples were decomposed in a bomb made of Teflon with 50% hydrogen peroxide and heated in an oven at 125°C for 4 h. The element concentrations were obtained sequentially by electrothermal vaporization ICP-AES using S-pL sample aliquots. Due to the high volatility of mercury, the commonlyuseddigestionmethods are susceptible to mercury loss, requiring special wet techniques. The loss of mercury was virtually eliminated by using aqua regia in the digestion of pulp and paperboard samples [633]. This technique is also useful for wood. The measurement of mercury was done by cold vapor atomic absorption. Neutron activation analysis (NAA) is extremely sensitive and accurate and there are no requirements for the destruction of the organic matter 16341. A discussion of the use of NAA in the analysis ofwood samples is provided by Meyer and Langwig [63S], and data for different species are available [636,6371. For this

aracterization Chemical

of Wood

369

techniquesimultaneousmultielement analysis canbeachieved by using solid samples (wood, sawdust, ash). The specific locations of elements in intact samples may be determined by an X-ray analysis attaching to a scanning electron microscope 16381.By imaging-microprobe secondary-ionmassspectrometry (SIMS), the spatial distribution of the elements may be determined. The spatial distribution of trace elements in jack pine, Pinus husksiarzn (lamb), by SIMS was determined [639]. Trace elements were found to be concentrated in specific morphologicalfeatures,namely, the torus, middle lamella, cell comers, and ray parenchyma wall. The samples of jack pine were examined for Ca, Mn, Cu, and Znby NAA and/or ICP-AES and for Fe, K, AI, Cl, Mg, Sr, and Cr by ICP-AES. Although differences in environmental conditions during the growth of a tree can result in large variations in the concentration of trace elements, the values obtained from this study are in agreement, within experimental error, with values obtained from the bulk inorganic content of jack pine [640,641]. Saka and Goring [642] studied the distribution of inorganic constituents of black spruce (Picea muriarza Mill.) by means of transmission scanning microscopy coupled with energy disperse X-ray analysis (TEM-EDXA) and detected 14 elements (Na, Mg, AI, Si, S, Cl, K, Ca, Cr, Fe, Ni, Cu, Zn, and Pb). Almost all of these elementsdetectedwere foundtobeconcentrated in the torus andhalf-bordered pit membrane regions. In the secondary walls of tracheids, ray tracheids, and ray parenchyma cells, only S, Cl, K, and Ca were detected. Saka and Mimori [643], by scanning microscopy coupled with EDXA (SEM-EDXA), determined the distribution ofinorganicconstituents in Japanese birch wood (Befuloplatyphyllu Sukatchev var. japonica Hara). Six morphological regions of the wood fibers, vessels, and ray parenchyma cells were investigated, and up to 11 different elements (Na, Mg, AI, Si, P, S, Cl, K, Ca, Fe, and Zn) were detected. The secondary walls of wood fibers, vessels, ray tracheids, and ray parenchymacellsusuallycontainonly detectable concentrationsof S, Cl, and Ca. In contrast, almost all of these elements detected were found to be localized and concentrated in the amorphous layers of ray parenchyma cells and pit membranes between vessels and ray parenchyma cells. The content of inorganic constituents determined by SEM-EDXA is in good agreement with the results obtained from ash residues of wood by bulk analysis.

REFERENCES 1.

2. 3. 4. 5. 6. 7. 8. 9. IO. 1 1.

12.

E. Sjiistriirn. Wood C/wruisfryF/rrldrrr,rc.rlrtr/.sm d Applicnfior1.s. 2nd ed., Academic Press. New York. p. 249 ( I 993). TAPPI 7?sf Mc.fhod.s, Technical Association of the Pulp and Paper Industry, Atlanta, GA. ASTM Sffmf[rrd.s,American Society for Testing and Materials. Philadelphia, PA. CPPA S/mckrrd.s. Canadian Pulp and Paper Association, Montreal. Quebec, Canada. Appiftc Stant/trrd.s, Australian Pulp and Paper Industry Technical Association. I S 0 Sfcrrrhrds, International Organization for Standardization. DIN-Norrrwrl, Deutsches lnstitut fur Normung. B. L. Browning. M e t h o d s of Wood Chcw~isrry.Interscience, New York (1967). D. B. Easty and N. S. Thompson. Fiber Sci. Technol. Ser., l / : 49 (1991). B. B. Sithole. Arltrl. Cllcwl.. 67: 87R (1995). D. Fengel and G. Wegener. Wood Chc~r,ri.sfr;y,U//rur.strrcc.trcrc,h'e~rc/ior~.s, Walter de GrLlyter, Berlin. chap. 3 (1989). S. Y. Lin and C. W. Dence (cds.). M r f h o t l s i r l Ligrlirl C'/wrrli.stry, Springer-Verlag, New York ( 1992).

Freer 370

and

Baeza

13.H.H.Nimz,in Proc. 8th Int. Symp. on Wood and Pulping Chentistry, Helsinki, Vol. 1, p. 1 (1995). 14. W. E. Moore and D. B. Johnson, Procedures f o r the Chemical Analysis of Wood and Wood Products, US. Forest Products Lab. (1967). 15. N.Morohoshi,in D. N.-S.HonandN.Shiraishi(eds.), Woodand Cellulosic Chemistry, Marcel Dekker, New York, chap. 8 (1991). 16. J. A.Smelstorius, Holdorsch. 28: 67 (1974). 17. D. Fengeland D. Grosser, Holz Roh-Werkstofl 33: 32(1975). 18.G.Wegener, Das Papier; 28: 478(1974). 19. C. M. Stewart. J. F. Melvin, S. H. Tham, and E. Zerdoner, Cellulose Chem. Techrzol. 7: 371 (1973). 20. B. L. Browning, Methods of Wood Chemistry, Interscience,NewYork,p.12(1967). 21. W. E. Kaarand D. L.Brink, J. WoodChem.Technol., 11: 479(1991). 22. B. L. Browning, Methods of Wood Chernistr?,, Interscience, New York, p. 45 (1967). 23. M. Goetzler, TAPPI, 65: 149(1982). 24.T.Ona,T.Sonoda,M.Shibata,andK.Fukazawa, TAPPI J., 78: 121 (1995). 25. R.T.Lin, ForestProd. J., 1 7 54 (1967). 26. R. T. Lin, ForestProd. J., 17: 61 (1967). 27.G.LohseandH.H.Dietrichs, Holz Roh-Werkstofl 30: 468(1972). 28.A.J.Nanassy, Wood Sci., S: 187(1973). 29.A. J. Nanassy, Wood Sci., 9: 104(1976). 30. A. R. Sharp, M. T. Riggin, R. Kaiser, and M. H. Schneider, Wood Fiber Sci., 10: 74 (1978). 3 1. E. Hsi, R. Hossfeld, and R. G. Bryant, J . Colloid Interface Sci., 62: 389 ( l 977). 32. E. Brosio, F. Conti, C. Lintas,and S. Sykora, J. FoodTechno/., 13: 107(1978). 33. J. E. CarlesandA.M.Scallan, J. Appl. Polytner Sci., 17: 1855(1973). 34. R. S. Menon, A. L. MacKay, J. R.T.Hailey,M.Bloom,A. E. Burgess,and J. S. Swanson, J. Appl. Polymer Sci., 33: 1 141 (1987). 35. R. S. Menon, A. L. MacKay, S. G. Flibotte, and J. R. T. Hailey, J . Mag. Res., 82: 205 ( 1 989). 36. C. D. Araujo,A.L.MacKay,J.R.T.Hailey,K. P. Whittall, and H. Le, Wood Sci. Technol., 26: 101 (1992). 37. A. Isogai,A.Ishizu, J. Nakano, S. Eda,andK.Kato. Carbohydr. Res., 138: 99 (1985). 38.G.J.Ritterand E. F. Kurth, lnd. Eng. Chetn.,25: 1250(1933). 39. T. E. Timelland E. C. Jahn, Svensk Papperstid., 54: 831(1951). 40.T. E. Timell, Pulp Paper Mag. Can., 60: T26(1959). 41.G. W. Holmesand E. F. Kurth, TAPPI, 42: 837(1959). 42. E. F. KurthandA. A. Swelim, TAPPI, 46: 591(1963). 43. Useful Method 249, in TAPPI Useful Methods, Technical Association of the Pulp and Paper Industry, Atlanta, CA. 44.B.L.Browning, The Chemistry of Wood, Interscience,NewYork,p.389 (1967). 45.G.Jayme, Cellulose Chern., 20: 43(1942). 46.G.JaymeandG.Schwab, Papierfabr.. 40: 147 (1942). 47. L. E. Wise, Ind. Eng. Clwm., Anal. Ed., 17: 63 ( 1945). 48. E. L. Lovell, Ind. Eng. Chenl., 3 7 1034 (1945). 49. L. E. Wise,M.Murphy,andA.A.D'Addieco, P q w r Trarlt. J.. 122: 35 (1946). 5 0 . N. S. Thompsonand 0. A.Kaustinen, TAPPI, 47: 157 (1964). 51. N. S. Thompsonand 0. A.Kaustinen, TAPPI, 53: 1502 (1970). 52. B. L. Browning, The Clwrnistry of Wood, Interscience,NewYork.p.396 ( 1967). 53. D. Fengel, H. Uqar,and G. Wegener, Dtrs Papier. 33: 233 (1979). 54. B. L. Browning, Methor1.s qf Wood Chtvnistty, Vol.2.Wiley-Interscience.NewYork.p.394 ( 1967). 5 5 . A. Poljak. Angen: Chcw~..AGO: 45 (1948). 56. J. S. AlbrechtandG.A.Nicholls. P q m i Puu. 58: 49 (1976). 57. H. Haas. W. Schoch,and U. Str6le, D(/s Ptrpier; Y: 469 (1955).

Chemical Characterizationof Wood

371

58. B. Leopold, TAPPI, 44: 230 (1961). 59. B. Leopold, TAPPI, 44: 232 (1961). 60. M. Kono, K. Sakai,and T. Kondo, J. Jpn. TAPPI, 19: 27 (1965). N. S . Thompson, Holiforsch., 22: 124 (1968). T. E. Timell, TAPPI, 44: 88(1961). T. E. Timell, Svensk Pupperstid., 63: 472(1960). T. E. Timell, TAPPI. 43: 844(1960). A. J. Mian and T. E. Timell, Svensk Pupper.stirl., 63: 884 (1960). T. E. Timell, Wood Sci. Technol., I : 45(1967). J.Janson, Prlperi Puu, 52: 323(1970). D. Fengel and G . Wegener, Wood Chemistry, Ultrrrstructure Reactions, Walter de Gruyter, Berlin, p. 115 (1989). 69. C. F. Cross and E. J. Bevan, J. Chertz. Soc., 3 : 666 ( 1880). 70. B. L. Browning, The Chemistry of Wood, Interscience,New York, p. 403 (1967). 71. A. Isogai, A.Ishizu,and J. Nakano, Holif
61. 62. 63. 64. 65. 66. 67. 68.

( 1993).

B. DalbeandA. Peguy. Cdltrlosc~Chwr. f i T h t ? O / . , 24: 327 (1990). H. Chanzy. P. Noe. M. Poillct. and P. Smith, J . Appl. Polyttwr Sci. Syttrp., 37: 239 ( 1983). 97. A. F. Turbac. TAPPI J.. 67: 94 ( 1984). 98. J. F. Kcnnedy, Z. S . Rivera, C. A. White, L. L. Lloyd.and F.P. Warner, Cellulose C ' h e t n . 7 i ~ h n o l .24: . 3 19 ( 1YYo). 99. J. D. Timpa. i n T. Provder. H. G. Barth, and M. W. Urban (eds.). Cl~rott~rtto~rurphir. Chtrrtrc.tcri:rrtiorr c$' P o l ~ r t r o s :Hypherrtrtcvl ctrrtl Mr(lti(littfl,tf,si[~tfrr/ f i ~ c h t ~ i t p r rAdvanccs ~.~. in Chcmistry Series 347. American Chcmical Society. Washington. DC, p. 141 (1995).

95. 96.

Freer 372

and

Baeza

100. A. El. Kafrawy, J. Appl. Polymer Sci., 27: 2435 (1982). 101. C. L. McCormick, P. A. Callais, and B. H. Hutchinson, Mucrornolecrcle.~,18: 2394 (1985). 102. P. C. Scherer, J . Anz.Cl7em. Soc., 53:4009 (193 l ) . 103. S. M. Hudson and J. A. Cuculo, J . Polymer Sri., Pol~vnerChetn. Ed., 18: 3469 (1980). 104. S. M. Hudson and J. A. Cuculo, J. PoI.vmer Sci., Polynzer Chenl. Ed., 20: 499 (1982). 1 05. D. C. Johnson, M. D. Nicholson, and F. C. Haigh, Appl. Polymer Swnp., 28: 931 (1976). 106. H. A. Swenson, Appl. Polynzer Syrnp., 28: 945 (1976). 107. T. J. Baker, L. R. Schroeder, and D. C. Johnson, Carbohydr. Res., 6 7 C4 (1978). 108. K. E. Almin, K. E. Eriksson, and B. A. Pettersson, J. Appl. Polymer Sei., 16: 2583 (1972). 109. L. Segal, Polwzer Lett., 4: 1011 (1966). 1 IO. M. L. Huggins, J. Am. Chern. Soc., 64: 2716 (1942). 1 1 1 . M. Houwink, J . Prokt. Chenz.. 157: 1940. 112. M. Marx, Makrotnol. Chern., 16: 157 (1955). 113. W. Brown and R. Wikstron, Eur. Polymer J.. I : 1 (1965). 114. J. Daiihelka, I. Kiissler, and V. BohAEkovri, J. Polyrner Sci., Polyn~erChenl. Ed., 14: 287

(1976). S. Claesson, N. Bergmann. and G. Jayme, Svensk Pupperstidn., 62: 141 (1959). A. M. Holtzer, H. Benoit, and P. Doty, J. Phys. Chern., 58: 624 ( 1954). H. Sihtola. B. Kyrklund, L. Laamanen. and I. Palenius, Paperi Plur, 4a: 225 (1963). W. J. Alexander and R. L. Mitchell, Anal. Chen7. 21: 1497 (1949). L. Segal, J. Pdynzer Sci. C , 21: 267 (1968). 120. D. A. I. Goring, Pure Appl. Chern., 5: 233 (1962). 121. R. Y. M. Huang and R. G. Jenkins, TAPPI, 52: 1503 (1969). 122. J. J. Cael, D. J. Cietek, and F. J. Kolpak. J . Appl. Polymer Sei.: Appl. Polymer .Svtnp., 37:

1 15.

116. 117. 118. 119.

509 (1983). 123. 124. 125. 126. 127. 128.

129. 130. 131. 132. 133. 134. 135. 136. 137. 138.

139. 140.

J. M. Lauriol, P. Froment, F. Pla, and A. Robert, Holzforsch., 41: 109 (1987). A. Huber, Mukromol. Chern., 61: 2248 (1992). D. L. Williams, J. A. Pretus, I. W. Browder, J . Liq. Chr~~rnrrtogr:, 15: 2297 (1992). G. W. Saunders, Diss. Abstr., Int. E., SI: 2324 (1990). L. Segal, J. D. Timpa, and J. I. Wadsworth, J. Po1yrnc.r Sei. A - / , X : 8:25 (1970). G. Meyerhoff and S. Javanovic, Pdymer Lett. 5: 495 (1967). L. Valtasaari and K. Saarela, Puperi PMI, 5 7 5 (1975). 0. Soubelet, M. A. Presst, M.Figini, and M. Angew., Mukrotnol. Sei. Chertz., 175: 1 17 (1991). T. E. Eremeeva, T. 0. Bykova, and V. S. Gromov. J . C/~romcrtogr:. 522: 67 (1990). T. Rantanen, P. Flrm, and J. Sundquist, P c p r i Pull, 68: 634 ( 1986). T. Rantanen and J. Sundquist, TAPPl J., 70: 109 (1987). R. Evans, R . H. Wearne, and A. F. A. Wallis, J . Appl. Po/ynzer Sci., 3 7 3291 ( 1989). D. Miller, D. Senior, and R. Sutcliffe, J . Wood Chern. 7kchrzo/., 11: 23 (1991). A. E. El. Ashmawy. J. Daiihelka, and I. Kossler, Swrz.sk Puppc~r.wtirl.,16: 603 (1974). L. R. Schroeder and F. C. Haigh, TAPPI, 62: 103 (1979). B. F. Wood, A. H. Conner, and C. G. Hill, J . Appl. P o l y n w Sei.. 32: 3703 (1986). P.F. Vidal, N. Basora, R. P. Overend, and E. Chornet, J . Appl. Po!\wc>r Sci.. 42: 1659 (1991). B. Saake, R. Patt. J. PUIS, K. J. Linow, and B. Phillip, Mdrorno/. Chcnz. Makron~ol.Synp.,

61: 219 (1992). 141. W. Schwald and 0. Bobletcr, J . Appl. Pulyrner Sei.. 35: 1937 ( 1988). 142. L. L. Lloyd, F.P. Warner, J. F. Kennedy, and C. A. White, in J. F. Kennedy, G. 0. Phillips, G. Owain, and

P. A. Williams (eds.),

Wood C ~ ~ / / ~ t / o .Ellis s e , Horwood, Chichester. U.K., p.

203 ( 1987). 143. 144. 145. 146. 147.

K. Kamide and M. Saito. A h Polynwr Sci.. 83: 1 (1987). A. H.Conner. Chronztrtogt: Sci. SCK, 69: 33 1 (1995). 0. Ant-Wuorinen and A. Visappii. Ptrpevi Puo, 47: 3 1 1 ( 1965). H. Knollc and G. Jayme. Dus Ptrpior; IY: 106 (1965). M. L. Nelson and R. T. O’Connor, J . App/. Po/yrnc~rSci., IS: 131 1 ( I 964).

Chemical Characterizationof Wood

373

148. L. Ferrlis and P. Pa&, Cellulose Chem. Technol., 11: 633 (1977). 149. J. Sugiyama, H. Harada,and H. Saiki, 1nt. J. B i d . Mucromol., 9: 122 (1987). 150. A. Bourret,H. Chanzy,and R. Lazaro, Biopolpers, 11: 893(1972). 151. K. H. Gardnerand J. Blackwell, Biopolymers, 13: 1975 (1974). 152. W. Claffeyand J. Blackwell, Biopolymers, 15: 1903(1976). 153. R. H. Atalla and D. L. VanderHart. Science, 223: 283(1984). 154. J. Sugiyama, R.Vuong, and H. Chanzy, Macromolecules, 24:4168 (I991). 155. J. H. Wiley and R. H. Atalla, in R. H. Atalla (ed.), Strucfure of Cellulose, ACS Symp. Ser. 340, American Chemical Society, Washington, DC, p. 151 (1987). 156. J. Sugiyarna. J. Persson, and H. Chanzy, Macromolecules, 24:2461(1991). 157. D. L. VanderHart and R. H. Atalla, Mucrornolecules, 17: 1465 (1984). 158. F. Horii. A. Hirai, and R. Kitarnaru, Mrrcrotnolecules, 20:2117 (1987). 159. D. Fengel andG. Wegener, Wood Chemistry, Ultrastructure Reactions, Walter deGruyter, Berlin, p. 87 (1989). 160. A. Sarko, TAPPI. 61:59(1978). 161. J. Sugiyarna, T. Okano, H. Yamamoto, and F. Horii, Macromolecdes, 23: 3196 (1990). 162. S. E. Darrnon and K. M. Rudall, Disc. Faraday Soc., 9: 251 (1950). 163. M.Tsuboi, J. Polymer Sci., 25: 159 (1957). 164. C. Y. Liang and R. H. Marchessault, J . Polymer Sci., 37: 385 (1959). 165. C. Y. Liang and R. H. Marchessault, J. Polytner Sci., 39: 269 (1959). 166. R. H. Marchessault and C. Y. Liang, J. Polymer Sci., 43: 71 (1960). 167. H. G. Higgins, C. M. Stewart. and K. J. Harrington, J. Polymer Sci., 51: 59 (1961). 168. K. H. Bassett, C. Y. Liang, and R. H. Marchessault, J . Polymer Sci. A, 1 : 1687 (1963). 169. D. N.-S. Hon and N. Shiraishi (eds.), Wood u r d Cellulosic Chernistty, MarcelDekker.New York, p. 341 (1991). 170. J. Blackwell, in J. C. Arthur (ed.), Cellulose Chemistry and Technology, American Chemical Society. Symp. Ser. 48, Washington, DC. p. 206 (1977). 171. D. Fengel and M. Ludwing, D m Papier, 45: 45 (1991). 172. D. Fengel, Dus Papier, 45: 97 (1991). 173. L. G. Braznik, R. G. Zhbankov, T. A. Tsetsokho,and A. M.Shishko, Zh. Prikl. Spektrosk, 53: 633 ( 1991). 174. A. J. Michell, Appitcr. 41: 375 (1988). 175. A. J. Michell, Curlmhydr. Res., 173: 185 (1988). 176. A. J. Michell, Curbohpdr. Res., 197 53(1990). 177. A. J. Michell, Cellulose Chern. Techno/., 27: 3(1993). 178. H. J . Marrinanand J . Mann, J . Polymer Sci. 21: 301 (1956). 179. A. W. McKenzieand H. G. Higgins, Svensk Puppersridn., 61: 893 (1958). 180. D. Fengel, Holzjiirsch., 46: 283(1992). 181. D.Fengel, Ho1;forsch.. 47: 103(1993). 182. D. V. Whittaker. L. A. S. Parolis, and H. Parolis, Curbohydr. Res., 253: 247 (1994). 183. G. Grbnberg, U. Nilsson. K. Block, and G. Magnusson, Carbohydr. Res., 257: 35 ( 1994). 184. J . Sano, N.Ikushima, K. Takada, and T. Shoji, Cnrbohydr. Res., 261: 133 (1994). 185. T. J . Rutherford, C. Jones, D. B. Davies, and A. C. Elliott. C~trbohylt:Res., 265: 79 (1994). 186. M. Matulov6, R. Toffanin, L. Navarini, R. Gilli, S. Paoletti, and A. Ceslro, Carho/Zydt:Res..

265: 167 ( 1994). 187. S. M. A. Homlbeck, P. A. Petillo, and L. E. Lerner, Riocherni.stc\; 33: 1426 (1994). 188. M. Bruix, J . JimCnez-Barbero,and P. Cronet, Curld~ycir.Res., 273: 157 (1995). 189. R. P. Gorshkova, V.V. Isakov, E. N. Kalmykova,and Y. S. Ovodov. Curbohydr: Res., 268: 249 (1995). 190. A. J . DeBruyn, J . Curbohydr: Chem., 14: 135 (1995). 191. L. Poppc and H. van Halbeek. N u t u r d Sfr’uct. Bid., 1: 215 (1994). 192. B.Adarns and L. E. Lerner, M q n . Reson. Chern., 32:608 (1994). 193. J . M. Hackney. R. H. Atalla, and D. L. VanderHart, I n t . J . B i d . Mucrotnol., 16: 215 (1994).

Freer 374

and

Baeza

194. F. Horii,H. Yamamoto, R. Kitamaru,M.Tanahashi,and T. Higuchi, Macromolecules, 2 0 2946 (1987). 195. H. Yamamoto, F. Horii, and H. Odani, Macromolecules, 22: 4130 (1989). 196. H. J. Lennholm, T. Larsson, and T. Iversen, Curbohyd,: Res., 261: 119 (1994). 197. R. H. Newman, J . Wood Chem.Technol., 14: 451 (1994). 198. R. H. Newman and J. A. Hemmingson, in Proc. 8th Int, S.ymp. on Wood and Pulping Chemist?, Helsinki, Vol. 3, p. 519 (1995). 199. A. Jeanes and H. S. Isbell, J. Res. Natl. Bu,: Std. 27: 125 (1941). 200. E. S. Becker, J. K. Hamilton, and W. E. Lucke, TAPPI, 48: 60 ( 1965). 201. J. K. N. Jones, C. B. Purves, and T. E. Timell, Can. J. Chem., 39: 1059 (1961). 202. K. Kurschenerand S. Karicsonyi, Holdorsch., 15: 107 (1961). 203. J. K. HamiltonandG.R. Quimby, TAPPI, 40: 781 (1957). 204. B. L. Browning, Methods of Wood Chemistry, Interscience, New York, p. 567 (1967). 205. J. A. Smelstorius and C. M. Stewart, Hol$orsch., 28: 204 (1974). 206. D. Fengel, Das Papier, 34: 428(1980). 207. R. L. Whistler and G. E. Lauterbach, Arch. Biochem. Biophys., 7 7 62 (1958). 208. R. L. Whistler and J. L. Sannella, Meth.Curbohyd,:Chem., 5: 34 (1965). 209. R. E. Gramera and R. L. Whistler, Arch.Biochem.Biophys.. 101: 75 (1963). 210. E. Hagglund, B. Lindberg,andJ.MCPherson, Acta Chetn. Sccmd., 10: 1160 (1956). 21 I . P. D. Cafferty, C. P. J. Glaudemans. R. Coalson, and R. H. Marchessault, Svensk Puperstid.. 6 7 845 (1964). 212. J. K.N. Jones, L. E. Wise, and J. P. Jappe, TAPPI, 39: 139 (1956). 213. R. W. Scott, J. Appl. Pol.vmer Sci.. 38: 907(1989). 214. R. L. Whistler and J. N. Be Miller, Adv. Curhohyd,: Chern., 13: 289 (1958). 215. G. Katz, TAPPI, 48:34 (1965). 216. T. Matsuo and T. Mizuno, Ag,: B i d . Chenl., 38: 465 ( 1974). 217. M. H. Johanssonand 0. Samuelson, Wood Sci. Technol., 11: 251 (1977). 2 18. G. 0. Aspinall, C. T. Greenwood, and R. L. Sturgeon, J. Chern. Soc., 3667 (1961). 219. J. L. Minorand N. Sanyer, TAPPI, 5 7 109 (1974). 220. A. BeClik, R. J . Conca, J. K. Hamilton, and E. V. Partlow, TAPPI, 50: 78 ( 1967). 221. C . Schuerch, in B. L. Browning (ed.), The Chemistry of Wood, Wiley-Interscience. New York, p. 191 (1963). 222. D. L. Brinkand A. A. Pohlman, TAPPI, 55: 380 (1972). 223. G. G. S. Dutton, B. J. Joseleau, and P. E. Reid. TAPPI. 56: 168 ( 1973). 224. B. D. E.Gaillard, N. S. Thompson, and A. J. Morak, Curbohydt: Res., 11: 509 (1969). 225. S. K. Chanda, E. L. Hirst, J. K. N. Jones, and E. G. V. Percival, J . Chem. SOC . . 1289 ( 1950). 226. A. J. Erskine and J. K. N. Jones, Curl. J . Chem.. 34: 821 ( 1956). 227. J. E. Scott, Chem. 1nd. (Lond.), 168 (1955). 228. J . E. Scott, Meth. Rioc.hern. Anal.. 8: 145 ( 1960). 229. B. L. Browning, The Chemi.stt~~ of Wood, Interscience, New York. p. 58 I ( 1967). 230. H. 0. Bouvengand B. Lindberg, A c m Chew. Scund., 12: 1977 (1958). 23 I . E.L.Hirst, D. A. Rees, and N. G. Richardson, Biochemistry, 25: 13 ( 1970). 232. J. D. Blake and G. N. Richards, Curbohydt: Res., 17: 253 ( 1971). 233. D.Fengel and M. Przyklenk, S w n s k Puppersfid., 78: 17 (1975). 234. S. C.Churms. A h . Curbohyd,: Chcrn. Biochem.. 25: 13 ( 1970). 235. K. Kringstad and 0. Ellefsen. Das Pupicv; /K: 583 ( 1964). 236. S. C. Churms and A. M. Stcphen. C~rhohytlr.Res., 21: 91 (1972). 237. A. Heyraud and M. Rinaudo. J. Chromtrtogt:. 166: 149 (1978). 238. T. E. Erenlecva and T. 0. Bykova. ./. Chrornrltogr: 639: IS9 (1993). 239. 0. E. Anderson, Pupcr Trcrtlr. J., 104: 42 ( 1937). 240. K.-S. Jiang and T. E. Timell. Ce/lulo.se Che~n.7tc~hnol..6: 493 ( 1972). 24 I . P. Albersheim, D. J. Nevins. P. D. English, and A. Karl. C~~rl?(~h.vr/t: K r s . . S: 340 ( 1967). 242. D. Fcngel. M. Przyklenk. and G. Wegener. P q ~ i c 3~0:: 240 ( 1976).

racterization Chemical

of Wood

375

243. M. F. Chaplin and J. F. Kennedy (eds.) Carbohydrate Analysis, a Practical Approach. IRL Press, Oxford-Washington, DC (1986). 244. Y. C. Lee, Anal. Biochern, 189: l5 1 (1990). 245. G. W. Monier-Williams, J . Chem. Soc., 119: 803 (1929). 246. J. F. Saeman, W. E. Moore, R. L. Mitchell, and M. A. Millett, TAPPI, 37: 336 (1954). 247. G. Jayme and H. Knolle, Z. Anal. Chem., 178: 84 ( 1960). 248. J. E. Jeffery, E. V. Partlow, and W. J. Polglase, Anal. Chem., 32: 1774 (1960). 249. M. L. Laver, D. F. Root, F. Shafizadeh, and J. C. Lowe, TAPPI, 50: 618 (1967). 250. R. C. Pettersen, V. H. Schwandt, and M. J. Effland, J . Chrornatogr. Sci., 22: 478 (1984). 251. D. Fengel and G. Wegener, in R. D. Brown, Jr. and L. Jurasek (eds.), Hydrolysis of Cellulose: Mechanisms of Enzymaticand Acid Catalysis, Adv. Chem. Ser. 181, American Chemical Society, Washington, DC, p. 145 (1979). TAPPI, 65: 103 (1982). 252. M. G. Paice, L. Jurasek, and M. Desrochers, 253. M. Tenkanen, T. Hausalo, M. Siika-aho, J. Buchert, and L. Viikari, in Proc. 8th Int. Svmp. on Wood and Pulping Chemistty, Helsinki, Vol. 2, p. l89 (1995). Chemistry, 254. A. Rydlund and 0. Dahlman, in Proc.8th Int. Svrnp. on WoodandPulping Helsinki, Vol. 2, p. 159 (1995). 255. E. P. Crowell and B. B. Burnett, Anul. Chenz., 39: 121 (1967). 256. L. G. Borchardt and C. V. Piper, TAPPI, 53: 257 (1970). 257. Pierre Handbook and Catalogue, 1979-1980, pp. 173 and 182, Methods 2 and 21. 258. K. M. Brobst and C. E. Lott, Jr., C e r e d Chem., 43: 35 (1966). 259. L.-[. Larsson and 0. Samuelson, Svensk Papperstid., 70: S71 (1967). 260. M. Sinner, M. H. Simatupang, and H. H. Dietrichs, J . Wood Sci. Technol., 9: 307 (1975). 261. M. Sinner and J. Puls, J. ChromrrtogI:, 156: 197 (1978). 262. M. R . Ladisch, A. L. Huebner, and G. T. Tsao, J . Chromutogr., 147: 185 (1978). 263. A. D. Jones, I. W. Bums, S. G. Sellings, and J. A. Cox, J. Chrornatogc, 144: 169 (1977). 264. R. Schwarzenbach, J . Chromatogr., 1 1 7 206 (1976). 265. R. T. Yang, L. P. Milligan, and G. W. Mathison, J. Chromntogr., 209: 316 (1981). 266. A. Heyraud and M. Rinaudo, Liy. Chromatogr., 3: 72 I (1 980). 267. H. D. Scobell, K. M. Brobst, and E. M. Steele, Cereal Chem., 54: 905 ( I 977). 268. F. E. Wentz, A. D. Marcy, and M. J. Gray, J. Chromatogr. Sci., 20: 349 (1982). 269. S. J. Angyal. Chenl. Soc. Rev.. 9: 415 (1980). 270. R. C. Pettersen and V. H. Schwandt, J . Wood Chem. Technol., 11: 495 (1991). 271. G. G. Neuburger and D. C. Johnson. A n d . Chem., 59: 203 (1987). 272. T. Hausalo, Proc. 8th M . S w p . Wood trnd Pulping Chemistry, Helsinki, Vol. 3, p. I3 1 ( 1995). 273. D. A. Mead. L. A. Larew, W. R. Lacourse, and D. C. Johnson, in P. Jandik and R. M. Cassidy (eds.), Advunces i r l I o n Chror,latogrtrphy. Vol. 1. p. 13 (1989). 274. W. T. Edwards. C. A. Pohl, and R. Rubin. TAPPI J., 70: 138 (1987). 275. M. L. Lavcr and K. P. Wilson, TAPPl J., 76: IS5 (1993). 276. M. Suzuki, R. Sakamoto, and T. Aoyagi, TAPPI J.. 78: 174 (1995). 277. J . J. Worrall and K. M. Anderson. J. Woocl Chem. Technol., 13: 429 (1993). 278. L. Van Nifterik. J. Xu, J. L. Laurent. and J. Mathieu. J. Chrornutogf:,640: 335 (1993). 279. J . Puls, T. D. Glawischnig.A.Herrmann,A.Borchmann, andB. Saake. in Proc. 8 t h I n t . Syrnp. on Wood and Pulpir~~g Chcwistry, Helsinki. Vol. I , p. S03 ( 1 995). 280. C. Huher. E. Grill, P. Oefner, and 0. Bobleter, Fre.senirr.s’J . A n d . Chcrn., 348: 82.5 ( 1994). 281. B . L.Browning. TAPPI. 32: l19 (1949). 282. M. Bylund and A. Donetzhuher. S ~ v n s kPuppr,v~id.,71: SOS (1968). 283. Z. Dische. J. Biol. Chcvn.. 167: 1 89 ( 1947). 284. R. W. Scott, W. E. Moore, M. J. Effland. and M. A. Millett. Antrl. R i o c ~ h c ~ n r21: . . 68 ( 1967). 285. T. Bittcr and H. M. Muir. A/?tr/.BiochenI., 4: 330 ( 1962). 2x6. R. W. Scott, Anrrl. Chcwr.. S/: 936 (1979). 2x7. T. M. Jones and P. Albcrsheinl. Pltrnt Physiol., 3Y: 926 (1972). 288.

S. Johnson and 0. Samuelson. Anul. Chirn. Actu. 36: 1 ( 1966).

I

376

Baeza and Freer

289. M. J. Spiro, Ann/. Biochem., 82: 348 (1977). 290. A. G. J. Voragen, H. A. Schols,J. A. de Vries, and W. Pilnik, J. Chrornntogr., 244: 327 ( 1982). 291. R. L. Taylor and H. E. Conrad, Biochern, / l : 1383 (1972). 292. J. H.Pazur, D. J. Dropkin, K. L. Dreher, L. S. Forsberg, and C. S. Lowman, Arch. Biochem. Biophys., 176: 257 ( 1976). 293. B. L. Browning, The Chemisfry of Wood, Interscience, New York, p. 615 (1967). 294. B. L. Browning, The Chemistry of Wood, Interscience, New York, p. 653 (1967). 295. S. Hakomori, J. Biochern. (Tokyo), 55: 205 (1964). 296. H. Bjorndal. C. G. Hellerqvist, B. Lindberg, and S. Svensson, Angew. Chern. Int. Ed., 9: 610 (1970). 297. W. N. Hawrorth, J. Chern. Soc., 107: 8 (1915). 298. H. A. Hampton, W. N. Haworth, and E. L. Hirst, J. Chem. Soc., 136: 1739 (1929). 299. T. Purdie and J. C. Irvine, J. Chetn. Soc., 83: 1021 (1903). 300. T. Purdie and J. C. Irvine, J. Chem. Soc., 85: 1049 (1904). 301. R. Kuhn, H. Trischmann, and I. Low, Angew. Chem., 67: 32 (1955). 302. J. Finne, T. Krusius, and H. Rauvala, Carhohyrlr. Res., 80:336 (1980). 303. L. R. Phillips and B. A. Fraser, Curbohydr. Res., YO: 149 (1981). 304. P. A. Sandford and H. E. Conrad, Biochemistty, 5: I508 ( 1 966). 305. J. H. Pazur, in M. F. Chaplin and J. F. Kennedy (eds.), Cnrbohydrcrre Andysis, n Practical Approach, IRL Press, Oxford, p. 67 (1987). 306. I. Ciucanu and F. Kerck, Carhohvdr. Res., 13: 209 (1984). 307. T. Narui, K. Takahashi, M. Kobayashi. and S. Shibata, Curbohydr. Res., 103: 293 (1982). 308. H. Rauvala,J.Finne, T. Krusius, J. Karkkainen,and J. Jarnefelt, Adv. Cclrbohvdr. Chem. Biochem., 38: 389 (1981). 309. G. W. Hay, B. A. Lewis, and F. Smith, in R. L. Whistler (ed.), Methods in Carbohvdrute Chemist?, Academic Press, New York, Vol. 5 , p. 347 (1977). 310. J. M. Bobbitt, A ~ ICnrbohyir. : Res., 11: 1 (1956). 311. 1. J.Goldstein,G. W. Hay, B. A. Lewis,and F. Smith, Meth.Cnrbohydr:Chetn., 5: 361 ( 1965). 3 12. C . T. Bishop, Meth. Curhohydc Chern.. 6: 350 (1972). 313. J. M. Vaughan and E. E. Dickey, J. Org. Chem., 29: 715 (1964). 3 14. B. Lindberg, J. Lonngren. and S. Svensson, Adv. Carhohydr. Chern. Biochem., 31: 185 (1975). 315. B. Lindberg, K. Samuelsson, and W. Nimmich, Cctrbohydr. Res., 30: 63 (1973). 3 16. B. S. Valent, A. G. Darvill, M. McNeil, B. K. Robertsen, and P. Albersheim, Crlrbohydr: Res., 79: 165 (1980). 317. T. J. Waeghe, A. G. Darvill, M. McNeil, and P. Albersheim, Carhohydr. Res.,123: 281 (1983). 3 18. J. Kiss, Ad\!. Curbohydr. Chetn. Biochem., 29: 229 (1974). Strlccturrrl Atrn1s.si.s cf Oli319. J. Dabrowski, in W. R. Croasmun and R. M. K. Carlson (eds.), 2nd ed., VCH, New York ( 1 994). ~qoscrcchrwidesrrnd Pol~~.strcc~harirle.s, 320. H. van Halbeek. Mefh. Enzyrnol., 230: 132 ( 1994). 321. T. Usui, T. Toriyama, and T. Mizuno, Agric. B i d . Chern., 43: 603 (1979). 322. P. A. J. Gorin, Arlv. Curbohydr. Chern. Biocherrz., 38: 13 (198 1 ). 323. A. S. Perlin and G. K. Hamer, ACS Syrnp. Ser., 103: 123 (1979). 324. F. R. Seymour, ACS Sytnp. Sex, 103: 27 (1979). 32s. H. Yamaguchi, S. Inamura, and K. Makino, J. Biochern. (Tokyo), 79: 299 (1976). 326. H. Yamaguchi and K. Makino, J. Biochem. (Tokyo), 81: 563 (1977). 327. H. Yamaguchi and K. Okamoto, J. Biochem. (Tokyo), 82: S I I (1977). 328. M. Tanaka, Curbohydr. Res., 88: 1 ( 1981 ). ~ ~ 36: 1162 (1940). 329. G . Gee, Trtrrw. F u r a d r ~Soc.. 330. T. E. Timell, .l. Am. Chrrn. Soc., 82: 521 1 (1960). 33 1 K. D. Sears, W. J. Alexander, 0. Goldschmid, and J. K. Hamilton, TAPPI, 61: 105 (1978). 332. K. B. Laffend and H. A. Swenson. TAPPZ J., 5 / : I18 (1968).

aracterization Chemical

333. 334. 335. 336. 337. 338. 339. 340. 341. 342. 343. 344. 345.

of Wood

377

J. K. Hamilton, P w e Appl. Chern., 5: 197 (1962). P. E. Gardner and M. Y. Chang, Dissolving Pulp Conference, Atlanta, C A (1973). R. L. Casebier and J. K. Hamilton, TAPPI, 48: 664 (1965). G. C. Hoffmannand T. E. Timell, Svensk Papperstid., 75: 135 (1972). A. J. Mian and T. E. Timell, Can. J . Chem., 38: 15 1 I ( 1960). B. W. Simson, W. A. CBtt, Jr., and T. E. Timell. Svensk Papperstid., 71: 699 (1968). B. L. Browning, The Chemistq oj' Wood, Interscience, New York, p. 703 (1967). B. W. Simsonand T. E.Timell. Ce/lrtlo.se Chern. Technol., 12: 51 (1978). B. V. Ettlingand M. F. Adams, TAPPI. 51: 116 (1968). G. P. BelueandG.D.McGinnis, J . Chron~atogr..9 7 25 (1974). S. C. Churms, E. H. Merrifield, and A. M. Stephen, CLrrbohydr. Res.. 55: 3 (1977). S. C. Churms, E. H. Merrifield, and A. M. Stephen, Cmbohydr. Res.. 64: C1 (1978). T. E. Eremeeva and 0. E. Khinoverova, Cellulose Ckem. 72chnol., 24: 439(1990) (in Russian); Chenl. Abstr. 118, 17 1256x ( 1 993). 346. E. Adler, Wood Sci. Technol., / I : 169 (1977). 347. G. F. Zakis, Functional Analysis of Lignin and Their Derivtrtives, TAPPI Press, Atlanta, GA ( 1 994). 348. S. Saka and D. A. 1. Goring, Hol$orsch., 42: 149 (1988). 349. A.Bjorkmanand B. Person, Svensk Przp/wrstid., 60: 158 (1957). 350. F. E. Brauns, The Chemistry of Lignin, Academic Press, New York, p. 51 (1952). 351. J. C. Pew and P. Weyna. TAPPI, 45: 247 (1962). 352. H. M. Chang, E. B.Cowling, W. Brown, E. Adler, and G. E. Miksche, Ho1lf)rch.. 29: 153 ( l 975). 353. J.-Y. Chen, Y. Shimizu, M. Takai, and J. Hayashi, Wood Sci. Technol.. 29: 295 (1995). 354. N. N. Erismann, J . Freer, J. Baeza, and N. DurBn, Bioresource 7khnol. 47, 247 (1994). 355. P. Benarand U. Schuchardt, Cellu/o.se Chem. Technol., 28: 435(1994). 356. B. Hording, K. Poppius,and J. Sundquist, Hol;forsch., 45: 109 (1991). 357. L.Laamanenand K. Poppius, Paper crnd Timber, 143 (1988). 358. J. H. Lora and S. Aziz, TAPPI J., 68:94 (1985). 359. S. Y. Lin, in S. Linand C. W. Dence (eds.), Methods i n Lignin Chenisrry, Springer-Verlag, Heidelberg, p. 75 ( 1992). 360. 1. A. Pearl, TAPPI, 52: 1253(1969). 361. B. L. Brownlng, Methods qf Wood Chemistry, Wiley Interscience. New York, p. 660 (1967). 362. Y.-Z. Lai and K. V. Sarkanen, in K. V. Sarkanen and C. H. Ludwig (eds.), Lignins-Occurrem?. Fonnntion, Structure trnd Rractions. Wiley, New York, p. 16.5 ( 1971 ). 363. K. Lundquist, in S. Lin and C. W. Dence(eds.), Methods i r l Lignirl Chemistry, SpringerVerlag. Heidelberg. p. 65 (1992). 364. D. Fengel and G. Wcgcner, Wood C/~c~rni.stry Ultrtrsrructure r r n d Retrctions, Walter de Gruyter, Berlin, p. 49 (1989). 365. F. E. Bmuns. The Chemistry ($'Lignin. Suppl. Vol.. Academic Press. Ncw York. pp. 49. 742 ( 1952). 366. F. E. Brauns and D. A. Brauns, The Chcwlistry of Lignin, Suppl. Vol., Academic Press, New York. p. 62 (1960). 367. C. Lapierre, C . Rolando. and B. Monties, Hol~orsch.. 3 7 189 (1983). 368. E. Adler, B. 0. Lindgren.and U. SaedPn. Stwnsk Ptrpperstid.. 55: 245(19.52). 369. E.Adler, J. M. Pepper. and E. Eriksoo. h / . E q . Chcwl.. 4Y: 1391 ( 1957). 370. G. Gellcrstedt and E. L. Lindfors. C. Lapierre, and B. Monties. Slvnsk f'cqqwrstid., 87: R61 ( 1984). 371. C. Lapierre, B.Monties. andC.Rolando, H o 1 : j k s c ~ h . .40: 47 (1986). 372. C.Lapierre and C . Rolando. Ho/;fi~r.sc~h.. 42: I (1988). 373. C. Lapierre. B. Montics. and C. Rolando, Hol;fbr:sch., 40: I13 (1986). 374. C. Lapierre, B. Monties,C.Rolando. in Pro(.. 4th h r . Syrnp. 011 Wood Pulping Chrrni.str;v. Paris. Vol. 2. p. 431 (1987).

Freer 378

and

Baeza

375. C. Lapierre, B. Pollet, and C. Rolando, in Proc. 8th h t . Symp. on Wood Pulping Chemistry, Helsinki, Vol. 2, p. 131 (1995). 376. K. Freudenberg, A. Janson, E. Knopf, and A. Haag, Chem. Ber.. 69: 1415 (1936). 377. F. D. Chan, K. Nguyen, and A. F. A. Wallis, J. Wood Chern. Techno/., 15: 329 (1995). 378. K. Y. Wu and D. L. Brink, J. Agric. Food Chern., 25: 692 ( 1 977). 379. H. M. Chang and G. G. Allan, in K. V. Sarkanen and C. H. Ludwig (eds.), Lignin, Occurrerzce, Formation, Structure crnd Kecrction, Wiley-Interscience, New York, p. 433 ( 1 97 l). 380. K. liyama and T. B. T. Lam, J. Sri. Food Agric., 51: 481 (1990). 381. T. P. Schultz, T. H. Fisher, and S. M. Dershem, J. Org. Chenz., 52: 279 (1987). 382. S. M. Dershem, T. H. Fisher, S . Johnson, and T. P. Schultz, Ho/@rsch., 42: 163 (1988). 383. S. Larson and G. E. Miksche, Acta Chem. Scrmd., 25: 647 (1971). 384. W. G. Glasser and N. Morohoshi, TAPPI, 62: 101 (1979). 385. N. Mohroshi and W. G. Glasser, Wood Sci. Techno/., 13: 165 (1979). 386. E. Adler, I. Falkehag, and B. Smith, Acta Clwrn. Scand., 16: 529 (1962). 387. E. Adler, S. Hernestam, and I. WalldCn, Svensk Pupperstid., 61: 641 (1958). 388. Y.-Z. Lai, in S. Lin and C. W. Dence (eds.), Methods in Lignin Chemistry, Springer-Verlag, Heidelberg, p. 423 ( 1 992). 389. Y.-Z. Lai and X.-P. Guo, Wood S c i Techno/., 25: 467 (1991). 390. Y.-Z. Lai, X.-P. Guo, and W. Situ, J. Wood Chem. Technol.. I O : 365 (1990). 391. Y.-Z. Lai and M. Funaoka, Holiforsch., 4 7 333 (1993). 392. Y.-Z. Lai and M. Funaoka, J. Wood Chenz. Technol., 13: 43 (1993). 393. Y, Ni, G. J. Kubes, and A. R. P. van Heiningen, J. Pu/p Puper Sci., 16: 583 (1990). 394. K. V. Sarkanen, A. Islam, and C. D. Anderson, in S. Lin and C. W. Dence (eds.). Methods in Lignin Chemistrv, Springer-Verlag, Heidelberg, p. 387 ( 1 992). 395. Y. Matsumoto, A. Ishizu, and J. Nakano, Holiforsch., 40: 81 (1986). 396. Y. MatsumotoandG.Meshitsuka, in Pro(.. 8th I n t . SynqJ. on Wood Pulping Chenzistty, Helsinki, Vol. 1, p. 557 (1995). 397. H. Taneda, N. Habu, and J. Nakano. Ho/;for.sch.. 43: 187 (1989). 398. M. Funaoka, I. Abe. and V. L. Chiang, in S. Lin and C. W. Dence (eds.), Methods i n Lignin Chcmi.st~?t,Springer-Verlag, Heidelberg, p. 369 ( 1992). 399. M. Funaoka and I. Abe, Mokuxri Grrkktrishi, 24: 256 ( 1978). 400. M. Funaoka and 1. Abe, M o k u x i Gakkrrishi, 28: 529 ( 1 982). 40 I . M. Funaoka and I. Abe, Mokuzrri Gnkknishi, 28: 621 ( I 982). 402. M. Funaoka and 1. Abe. Mokuzui Grtkknishi, 28: 718 (1982). 403. M. Funaoka and 1. Abe. Mokumi G a k k a i s h i , 3 / : 67 1 (1985). 404. M. Funaoka and I. Abe, Mokuzc~iGakknishi, 32: 47 (1986). 405. M. Funaoka, M. Shibata, and 1. Abe, Ho/&rsc/z., 44: 357 (1990). 406. V. L. Chiang and M. Funaoka, Ho/;for.sch., 44: 309 (1990). 407. V. L. Chiang and M. Funaoka, Ho/;fi)rsch.,44: 147 (1990). 408. F. D. Chaw K. L. Nguyen. and A. F. A. Wallis. J. Wood Chcwr. 7 k ~ h n o l . .1 5 : 473 ( 1995). 409. J. R. Obst, J . W(~ ot/Chem. Techno/..S: 377 (1983). 410. K. Kuroda. A. Yamaguchi, and K. Sakai, Mokuztri Gakkrrishi, 40: 987 (1994). 411. A. D. Pouwels and J. J. Boon, J . Wood Chem. T c . r ~ h n o / . . 7 : 197 ( 1987). 412. M. Klecn and K. Lindstriim. Nord. Pu//J P q w r K o s . J.. 2: 1 I 1 (1994). 413. J. Sjostrom and M. Reunanen. J . A n d . AI?/>/.Pyro/ysis, 17: 305 ( 1990). 414.

415. 416. 417. 41 8.

419.

J. Tanaka. A. lzumi. H. Ohi. K. Kurodn. and A. Yamaguchi. in Proc.. 8th I n / . Symp. on Wr~ocl Pulping Chc~rnistr:v.Helsinki. Vol. 3, p. 143 (1995). D. N.-S. Hon and N. Shiraishi (eds.). Wood trnd Cdlrrlosic Chc~rrri.st~:v, Marcel Dekkcr. New York. p. 317 (1991). D. B. Johnson. W. E. Moore, and L. C. Zank, TAPPI, 44: 793 (1961). A. G. Schiining and G. Johansson. S ~ ~ c w sPupperstid., k 68:607 ( 1965). J. Marton. 7APPl. 50: 335 (1967). P. A. Ahlgrun and D. A. I . Giiring. Ctrn. .l. Chc)trr..49: 1272 ( l97 I ).

Chemical Characterizationof Wood

420. 421. 422. 423. 424. 425. 426. 427. 428. 429. 430. 431. 432. 433. 434. 435. 436. 437. 438. 439. 440. 441. 442. 443.

379

G. Aulin-Erdtman, Svensk Papperstid., 56: 91, 287 (1953). G. Aulin-Erdtman and R. SandCn, Acta Chem. Scand., 22: 1187 (1968). K.Freudenbeg and K. Dall, Naturwiss., 42: 606(1955). A. S. Wexler, Anal. Chem., 36: 213 ( I 964). J. Marton and E. Adler, Acta Chem. Scand., 15: 370 (1961 ). E. Adler, Paperi Puu, 43: 634(1961). E.Adlerand J. Marton, Acta Chern. Scand., 13: 75 (1959). K. V. Sarkanen, H.-M. Chang, and G. G. Allan, TAPPI, 50: 583 (1967). G. Wegener, M. Przyklenk, and D. Fengel, Holzforsch., 3 7 303(1983). K. Iiyama and A. F. A. Wallis, Wood Sci. Technol., 22: 271 (1988). K. Iiyama and A. F. A. Wallis, Appita, 41: 442 (1988). K. Iiyama and A. F. A. Wallis, Hollforsch., 43: 309 (1989). W. D. Perkins, J. Chem.Ed., 63: AS (1986). W. D. Perkins, J. Chem. Ed., 64: A269 (1987). W. D.Perkins, J. Chem. Ed., 64: A296 (1987). 0 . Faix, Hokjorsch., 45, Suppl.: 21 (1991). T. P. Schultz and W. G. Glasser, Hollforsch., 40, Suppl.: 37 (1986). G.WegenerandC.Strobel, Second Meeting of the Interncrtionnl ExpertCouncil on the Chemistry of Vegercrble Resources (COCVER), St. Petersburg, Russia, p. 121 (1992). A. J.Michell, TAPPI J., 73: 235 (1990). A. J.Michell, Cellulose Chem. Technol., 22: 105 (1988). 0. Faix, H.-L. Schubert, and R. Patt, in Proc. 5th Symp. on Wood Pulping Chemisft?, Raleigh, NC, p. 1 (1989). N. L. Owen and D. W. Thomas, Appl. Spectrosc. 43: 451 (1989). A. J. Michell, Appitn, 47: 29 (1 994). R. J. Olsson, P. Tomani, M. Karlsson, T. Joseffson, K. Sjoberg, and C . Bjiirklund, TAPPZ J . , 78: 158 (1995).

444. 445. 446. 447. 448. 449. 450. 45 l . 452. 453. 454. 455. 456. 457. 458. 459. 460. 461. 462. 463. 464. 465.

T. P. Schultz, M. C. Templeton, and G. D. McGinnis, Anal. Chetn. 57: 2867 (1985). A. Ferraz. J. Baeza, J. Rodriguez, and J. Freer, Llioresource Technol., 74: 201 (2000). E. V. Thomas, Anal.Chem., 66: 795 A (1994). S. T. Balke, P. Cheung,L.Jeng, R.Lew, and T. H. Mourey. J . Appl.Polymer Sci. Appl. Polymer Symp., 48: 259 (1991). F. Kimura, T. Kimura, and D. G. Gray, Hol$orsch., 46: 529 (1992). A. J. Michell, P. J. Nelson. and C. W. J. Chin, Appita, 42: 443 (1989). A. J. Michell, Wood Sci. Technol., 2 7 69(1993). M. Brunner, R. Eugster, E. Trenka, and L. Bergamin-Strotz, Holdorsch., 50: 130 (1996). A. J. Michell and L. R.Schimleck. Appita, 49: 23 (1996). L. R. Schimleck, A. J. Michell. and P. Vinden, Appittr, 4Y: 319 (1996). A. J. Sornmer and J. E. Katon, TAPPI Proc. Pulping Confircwce. p. 543 (1993). C.-L. Chen and D. Robert, in W. A. Wood and S. T. Kellog (eds.). Mothotls in En:ynology, Academic Press, New York, Vol. 161, p. 156 ( 19x8). C. H. Ludwig. B. J. Nist,and J. L. McCarthy. J . Am. Chrm. Soc... H6: I196 (1964). 0. Faix, C . Grunwald. and 0. Beinhoff, Hol;/or.sc./~., 46: 425 (1992). K. Lundquist, Nord.PulpPaper Res. J., 7 4 (1992). S. Li and K. Lundquist, Nord. Pulp Ptrprr Res. J., 9: 191 ( 1994). D. A. Skoogand J. J. Leary, Principles of' lnstr~rmenfalAntrlysis. HarcourtBrace College. Orlando, FL ( 1992). A.Labidi, D. Robert. and F. Pla. Hol,;fhr.sc/~., 47: 133 (1993). L. Chcnand D. Robert. inW. A. Wood and S. T. Kellog (eds.). Methods i n El,:ymolosqy. Academic Press, New York. Vol. I6 1. p. 16I ( 1988). H. H. Nimz, D. Robert, 0. Faix. and M. Ncmr, Ho/;fi)r,sc~h.. 35: I6 ( 198 1 ). M. Bnrdet. M.-F. Foray. and D. Robert. M d r o r n o l . Chem., lS6: 149.5 ( 1985). G. Gellerstedt and D. Robert, Acto Chwl. Sctrntl. H . 41: S41 ( 1987).

I

380 466. 467. 468. 469. 470. 47 1. 472. 473. 474. 475. 476. 477. 478. 479. 480. 481. 482. 483. 484. 485. 486. 487. 488. 489. 490. 491. 492. 493. 494. 495, 496. 497. 498. 499

500 50I 502. 503. 504. 505.

506.

Baeza and Freer G. J. Leary and R. H. Newman, in C. Dence and S. Lin (eds.), Lignin Cl?ernistry, SpringerVerlag, Berlin, p. 147 (1992). J. F. Haw and G. E. Maciel, Anal. Clre~n.,56: I323 (1984). G . J. Leary, K. R. Morgan, and R. H. Newman, Appita, 40: 181 (1987). G. J. Leary, K. R. Morgan, and R. H. Newman. Hol;forsch., 40: 221 ( 1986). R. M. Ede, G. Brunow, L. K. Simola, and J. Lemmetyinen. Holdorscl?.,44:95 (1990). K. R. Williams and R. W. King, Anal. Cl?em.,67: AI 25 (1990). A. Bax and D. G. Davis, J. Magn. Reson., 65: 355 (1985). M. F. Summers, L. Marzilli, and A. Bax, J. Am. Chern. Soc.., 108: 4285 (1986). R. M. Ede and Y. Kipellinen, Res. Chern. Interrned., 21: 313 (1996). R. M. Ede, J. Ralph. K. M. Torr, and B. S. W. Dawson, Ho/;forsch., 50: 161 (1996). I. Kilpellinen, J. Sipill, G. Brunow, K. Lundquist, and R. M. Ede. J. Agric. Food Clwnl., 42: 2970 ( I 994). R. M. Ede and J. Ralph, in Proc. 8th I n t . Synlp. 011 Wood crnd Prrlpirlg Cl?c~nri.stry. Helsinki, Vol. 3, p. 97 (1995). N. Fukagawa, G. Meshitsuka, and A. Ishizu, J. Wood Ckern. Tecllr~ol.,12: 425 (1992). R. M. Edeand 1. Kilpellinen, in Proc. 8th h1t. S y n p . o n Wood rrrzd Pulpirlg Cl?ernistry, Helsinki. Vol. I , p. 487 (1995). D. S. Argyropoulos, H. I. Bolker, C. Heitner. and Y. Archipov, Hol$orsck, 47: 5 0 (1993). D. S. Argyropoulos, H. 1. Bolker, C. Heitner, and Y. Archipov, J . Wood Chenl. fi~chrlol.,13: 1 x7 ( 1993). D. S. Argyropoulos, J. Wood cI?L'1?1. E~h~lol., 14: 45 (1994). D. S. Argyropoulos, J . Wood Clrern. Techrwl., 14: 65 (1994). D. S. Argyropoulos and C. Heitner, Hol2;forsch., 48. Suppl.: I12 (1994). P. Malkavaara and R. A l h , in Proc. 8/11 h i t . Syrnp. on Wood c r n d Prrlping Clwrnistty, Helsinki. Vol. 3, p. 103 (1995). D. S. Argyropoulos, Y. J. Sun, M. Mazur. B. Hortling, and K. Poppius-Levlin. Nord. Pulp Puper Res. J., 10: 68 ( 1995). Z.-H. Jiang and D. S. Argyropoulos, in Proc. 8th I11t. Symp. on Wood m d Pulping Cherni.st~y, Helsinki. Vol. I , p. 5 1 1 (1995). M. Barrelle, J . Wood Cl~rrn.Tecknol., 12: 413 (1992). M. Barrelle, Hol&rsc/1.. 47: 261 (1993). M.Barrellc, J . Wood Chenl. Tech~lol..15: 179 (1995). G . Wegener and D. Fengel, Wood Sci. Tecllr~ol.,/ I : 133 (1977). 0. Faix, W. Lange, and E. C. Salud. Hol;for.sc.l1., 35: 3 (1981). D. A. I. Giiring, in K. V. Sarkanen and C. H. Ludwig (eds.), Lignins, Occ~rrrcr~cc~, Formr/ioII, Structur-e c r n d Rerrctiorls, Wiley-Intersciencc, New York. p. 695 ( 1 97 1 ). M. E. Himnlel. J. Mlynar. and S. Sarkancn, Clzmnltrtogr: Sci. SL'I:, 6 9 353 (1995). J. Pellinen and M. Salkinoja-Salonen, J . C h r o ~ ~ ~ r 328: ~to~ 299 ~ ~( :1985). . 0 . Faix. W. Langc. and 0. Beinhoff, Hol;forsch.. 34: 174 (1980). W. Langc, W. Schwcers. and 0. Beinhoff, H o l ~ ~ r . s c h35: . , 119 (1981). H. L. Chum, D. K. Johnson, M. P. Tucker, and M. E. Himmel. Hol;fi)r.sch., 41: 97 (1987). M. E. Himmel, K. Tatsumoto, K. K. Oh, K. Grohmann. D. K. Johnson, and H. L. Chum. in W. G. Glasser and S. Sarkanen (eds.), Lig11i11 Properties ( r n d Mtrtc.rirr1.s. ACS Symp. Ser. 397, chap. 6, p. X2 ( 1989). 0. Faix and 0. Beinhoff, Hol$)r:sch.. 46: 355 ( 1992). W. G. Glasser. V. DavC. and C. E. Frazier, J. Wood Che111. k h r ? o l . . 13: 545 (1993). E. J . Siochi. T. C. Ward. M. A. Haney, and B . Mahn. M~lcro~ilolecrr/r.s. 23: 1420 (1990). S. Sarkanen. D. C. Tellcr. J . Hall. and J . L. McCarthy. Mrrc.r~)111olec./r/(~.s. I d : 426 (1981). W. J. Connors. S. Sarkanen, and J. L. McCarthy, Hol;/i)r.sch.. 34: X0 (1980). S. Rudatin. Y. L. Scn. and D. L. Woerncr, in W. G. Glasser and S. arka an en (cds.). Li,qlfill PropvticJs m d Mtrterids, ACS Sytnp. Scr. 397. chap. 11. p. 144 ( 1989). S. Dutta. T. M. Carver, Jr.. and S. Sarkancn. i n W. G. Glasser and S. Sarkanen (eds.). 12ig~tifl P I - O ~ C ~ It - r~ d; CM .NS~ c I . ~ u / sACS , Symp. Ser. 397. chap. 12, p. I 55 ( 1989).

aracterization Chemical Wood

of

381

507. S. Sarkanen, D. C. Teller, C. R. Stevens.and J. L. McCarthy, Macromolecules, 17: 2588 ( 1984). 508. E. Adler and B. WesslCn, Acta Chem. Scand., 18: 1314 (1964). 509. T. N. Soundararajanand M. Wayman, J. Polymer Sci. C, 30: 521 (1970). 510. W. J. Connors, HolzJorsch., 32: 145 (1978). 51 1. W. J. Connors, L. F. Lorenz, and T. K. Kirk, Holzforsch., 32: 106 (1978). 512. R. Concin, E. Burtscher, and 0. Bobleter, Holzforsch., 35: 279 (1981). 5 13. T. K. Kirk, W. Brown, and E. B. Cowling, Biopolymers, 7 I35 ( I 969). 514. T. I. Obiagaand M.Wayman, J. Appl. Polymer Sci., 18: 1943(1974). 515. S. Sarkanen,D.C. Teller, E.Abramowski,and J. L. McCarthy, Macromolecules,15: 1098 (1982). 516. A.Huttermann, Holzj?orsch., 32: 108 (1978). 517. 0. Faix, W. Lange,and G. Besold, Hol;forsch., 35: 137 (1981). 5 18. E. J. Siochi,M. A. Haney, W. Mahn,and T. C. Ward, in W. G.Glasserand S. Sarkanen (eds.), Lignin Properties and Materials, ACS Symp. Ser. 397, Chap. 7, p. 100 ( 1 989). 519. P. Foment and F. Pla, in W. G. Glasser and S. Sarkanen (eds.), Lignin Properties and Materials, ACS Symp. ser. 397, chap. 10, p. 134 (1989). 520. D. T. Balogh, A. A. S. Curvelo, and R. A. M. C. De Groote, in J. F. Kennedy, G. 0. Phillips, 52 1. 522. 523. 524. 525. 526. 527. 528. 529. 530. 53 1 .

532. 533.

534.

Ellis and P. A. Williams (eds.), CelluloseChernisttyBiochemistryandMaterialsAspects, Horwood, London, U.K., p. 279 ( 1993). K. Forss, R. Kokkonen, and P.-E. Sigfors, in W. G. Glasser and S. Sarkanen (eds.), Lignin Properties and Materials, ACS Symp. Ser. 397, chap. 9, p. 124 (1989). K. Forss.R. Kokkonen, and P.-E. Sigfors, in W. G. Glasser and S. Sarkanen (eds.), Lignin Properties and Materials, ACS Symp. Ser. 397, chap. 13, p. 177 (1989). E. R. E. van derHage, W. M. G. M. van Loon, J. J. Boon, H. Lingeman,and U. A. Th. Brinkman, J. Chromutogc, 634: 263 (1993). M. Dolk, F. Pla, J. F. Yan, and J. L. McCarthy, Macromolecules. 19: 1464 (1986). 1. D. Suckling, M. F. Pasco, B. Hortling, and J. Sundquist, Holifbrsch., 48: 501 (1994). D. Dong and A. L.Fricke, J . Appl. Polymer: Sci., 50: 1131 (1993). D. Dongand A. L. Fricke. Polymer; 36: 2075 (1995). J. Bauch, 0. Schmidt, W. E. Hillis, and Y. Yazaki. Hol~$orsch..S I : I (1977). Y. Yazaki, Holyorsch.,36: 249 (1982). W. W. Wilcox and D. D. Piirto, Wood and Fiber; 7: 240 (1976). J. W. Rowe and A. H. Conner, Extractives in Eastern Hardwoods. A Review, General Tech. Rep. FPL. 18, U.S. Foresst Products Laboratory, Madison, W1 (1979). 1. Abe and K. Ono, J. J p n . Wood Res. Soc., 26: 686 (1980). G. Venkoba Rao.N. V. S. R. Murthy, and G. K. Srinivasan, TAPPI, 64: 96 (1981). E. S. Lipinsky, D. Anson, J . R. Longanbach,andM. Murphy, J . Am.OilChem. Soc., 62:

940 ( 1985). 535.

536. 537. 538. 539. 540. 541. 542. 543. 544. 545. 546. 547.

N. Levitin. Pulp PaperMag. Carl., 71(T361):81 (1970). W. E.Hillis and M. Sunlimoto, in J. W. Rouse (ed.), Naturd Products c$ W o o ~ l yPlants I , I I . Springer-Verlag, New York (1989). R. A. Abramovitch and 0. A. Koleoso. Can. J . Chem., 44: 2913 ( 1966). S. A. Rydholm, Pulping Processes, Interscience, New York (1965). A. Assarsson and I. Croon, S w n s k Papperstid., 66: 876 ( 1 963). A. Assarsson, 1. Croon, and A. Donetzhuber, Svensk Prqpvxtid. 66: 940 ( 1963). H. M. Nugent.L. H. Allen, and H. I. Bolker, Trrrns.Tech. Sec. CPAA, 3 : 103 ( 1977). N. Dunlop-Jones, H. Jialing. and L. H. Allen, J. Pulp Paper Sci., 17: 560 (1991). P. J. Allison, Paper SoutherAfrica, 8: 16 (1988). M. Douekand L. H.Allen, TAPPI, 61: 47 (1978). R. R. Affleck and R . G. Ryan. Pulp Puper Mag. Crm., 70(T563): 107 ( 1969). B. L. Browning, TheClzcwzistry of Wood, Interscience, New York, p. 75 (1967). R. G. Rickey, J. K. Hamilton, and H. L. Hergert, Wood Fiber; 6 : 200 (1974).

I

Baeza and Freer

382

548. E. F. Kurth, Ind. Eng. Chem., Anal. Ed., 11: 203 (1939). 549. B. L. Browning, The Chemistry of Wood, Interscience, New York, p. 119 (1967). 550. D. Fengel and G. Wegener, WoodChemistry,UltrastructureReactions, Walter de Gruyter, Berlin, p. 33 (1989). 551. D. W. Reeve and A. B. McKague, J. Pulp Paper Sci., l 7 5115 (1991). 552. A. Demirbag, Wood Sci. Technol., 25: 365 (1991). 553. T. Chen, B. Breuil, S. Carrikre, and J. V. Hatton, TAPPI J.. 7 7 235 (1994). 554. B. B. SitholC, J. L. Sullivan, and L. H. Allen, Hol7,forsch.. 46: 409 (1992). 555. B. B. SitholC, Appita, 45: 260 (1992). 556. Z. Zhang, M. J. Yang, and J. Pawliszyn, Anal. Chem., 66: 844A (1994). 557. H. Halvarson and 0. Hult, TAPPI J.. 66: 105 (1983). 558. B. B. Sitlholt, P. Vollstaedt, and L. H. Allen, TAPPI J., 74: 187 (1991). 559. E. J. Nurthen, B. V. McCleary, P. L. Milthorpe, and J. W. Whitworth, Anal. Chem., 58: 448 (1986). 560. D. F. Zinkel, TAPPI, 58: 109 (1975). 561. D. F. Zinkel, J . Wood Chem. Technol., 3: 131 (1983). 562. F. Terauchi, T. Ohira, M. Yatagai, T. Ohgama, H. Aoki, and T. Suzuki, Mokuzai Gakkaishi, 39: 1421 (1993). 563. F. Bright and M. E. P. M. McNally (eds.), Supercriticul Fluid Technology Theoretical and Applied Approaches in Analytical Chemistry, ACS Symp. Ser. 488 (1992). of Wood, Interscience, New York, p. 91 (1967). 564. B. L. Browning, The Chernist~):,, 565. A. W. Schorger, Ind. Eng. Chem., 9: 556 (1917). 566. M. Myer, B. Hausmann, N. Zelmann, and K. Kratzl, Proc. TAPPI Research and Development Division CO$, Technical Association cf the Pulp undPcrper Industry, Atlanta, GA, p. 155 (1984). 567. A. Sato and E. von Rudloff, Can J. Chem., 42: 635 (1964). 568. D. F. Zinkel, in M. H. Esser (ed.), Proc. 6th Annual Lightwood Research C o n j , Southeast Forest Experiment Satation, Asheville, NC, p. 147 (1979). 569. J. Drew, J. Russell, and H. W. Bajak, Sulfate Turpentine Recovery, Pulp Chemical Association, New York (1971). 570. J. Drew and G. D. Pylant, Jr., TAPPI, 49: 430 (1966). 571. D. F. Zinkel and C.R. McKibben, in M. H. Esser (ed.), Proc. 5th Annual Lightwood Research C o n j , Southeast Forest Experiment Satation, Asheville, NC, p. 133 (1978). 572. S. V. Kossuth and J . W. Munson, TAPPI, 64: 174 (1981). 573. E. Zavarin, Y. Wong, and D. F. Zinkel, in M. H. Esser (ed.), Proc 5th AnnuolLightw.ood Resetrrch Cotrf.. Southeast Forest Experiment Satation, Asheville, NC, p. 19 (1978). 574. R. P. Adanis. M. Granat, L. R. Hogge, and E. Rudloff, J . Chromotogt: Sci.. 17: 75 (1979). 575. F. Terauchi, T. Ohira, M. Yatagai. T. Ohgama, H. Aoki, and T. Susuki, Mokuzai Gakkcrishi, 39: 1431 (1993). U. Svedberg, Norrl. Pulp Papt,r RPS.J., 7: 155 576. H. Axelsson. C. Bostrom, D. Cooper, and (1 992). 577. R. C. McDonald and L. J . Porter. N.Z. J . Sci., 12: 352 (1969). 578. J. M. Uprichard and J. A. Lloyd, N.Z. J. Forest Sci.. IO: 55 I (1980). 579. I . D. Suckling and R. M. Ede, Appitcc. 43: 77 (1990). 580. B. L. Browning and L. 0. Bublitz. TAPPI, 36: 418 (1953). 581. H. Wienhaus. Chcwr. Trchnol. (Berlin), 5 : 25 (1953). 582. J. P. Casey (cd.), P d p und P q w r . 3rd ed., Wiley. New York, Vol. I , pp. 126. 424. 487 (1980). A. M. Mimms, M. J . Kocurek. J. A. Pyatte, and E. E. Wright (eds.), Krqf? Pulping, A Cornpilnrion of N o r c ~ s TAPPI . Press. Atlanta, CA. p. 2 1 ( I 993). 584. A. Quinde and L. Paszner, Ho/;fi)r.sch. 46: 5 I3 ( 1992). 585. B. B. SitholC. TAPPI J.. 76: 123 (1993). 586. M. Yang. Frw.ui Hlrtr.une, 21: I148 (1993). 587. M. J . Brites. A. Guerreiro, B. Gigante. and M. J . Marcelo-Curto, J. Chrotncrto~I:.641: 199

583.

(1993).

Chemical Characterizationof Wood

383

588. I. A. Rogers, Pulp Paper Mag. Can. 74(T303): 11 1 (1973). M.Castrkn, T. Nakari, B. Snickars-Nikinmaa, H. Bister, and E. 589. A. Oikari,B.E.Lonn, Virtanen, Water Res., 1 7 81 (1983). 590. S. Backa, A. Brolin, and N.-0. Nilvebrant, TAPPI J., 72: 139 (1989). 591. J. K. Volkman and D. G. Hodsworth, J. Chromatogx, 643: 209 (1993). 592. F. Alvarado, K. Lindstrom, J. Nordin, and F. Osterberg, Nord. Pulp Paper Res., 7 : 37 (1992). 593. C. J. Biermann and M.-K. Lee, TAPPI J., 73: 127 (1990). 594. J . Sjostrom, R. Btldenlid, and M. A. Norborg, Holtforsch., 4 7 446 (1993). 595. B. Charrier, M. Marques, and J. P. Haluk, Holtforsch., 46: 87 (1992). 596. H. L. Hardell, J. Anal. Appl. Pyrolysis, 27: 73 (1993). 597. Y. Ohtani, Appita, 46: 39 (1993). 598. E. Sjostrom, Wood Chemistql: Fundamentals and Applications, Academic Press, New York, p. 92 (1981). 599. R. Ekman, Holtforsch., 3 : 79 (1976). 600. C.-L. Chen, H.-M. Chang, and E. B. Cowling, Phytochemistry, 15: 547 (1976). 601. T. W. Pearson, G. S. Kriz, Jr., and J. Taylor, Wood Sri. Technol., 10: 93 (1977). 602. I. A. Pearl and S. F. Darling, Phytochemistry, 9: 1277 (1970). 603. K. Miki, K. Ito, and T. Sasaya, J. Jpn. Wood Res. Soc., 25: 665 (1979). 604. R. W. Spencer and E. T. Choong, Holtforsch., 31: 25 (1977). 605. M. Samejima and T. Yoshimoto, J. Jpn. Wood Res. Soc., 25: 671 (1979). 606. T. Popoff and 0. Theander, J. Appl. Polymer Sri.: Appl. Polymer Symp., 2 8 1341 (1976). 607. J. J. Karchesy, P. M. Loveland, M. L. Laver,D. F. Barofsky, and E. Barofsky, Phytochemistry, 15: 2009 (1976). 608. Y. Yazaki and W. E. Hillis, Holtforsch., 31: 20 (1977). 609. A. Calimli and A. Okay, HolZforsch., 32: 7 (1978). 610. B. L. Browning, The Chemistry of Wood, Interscience, New York, p. 223 (1967). 611. E. L. Johnson and A. J. Cserjesi, J. Chromatogs, 107 388 (1975). 612. S. V. Fonseca, L. T. Nielsen, and E. A. Ruveda, Phytochemistry, 18: 227 (1979). 613. B. L. Browning, The Chemistry of Wood, Interscience, New York, p. 253 ( 1 967). 614. A. Hofinger, B. Hinterstoisser, and S. Hofer, in Proc. 8th Int. Symp. on Wood and Pulping Chemistry, Helsinki, p. 121 (1995). 615. D. Fengel and G. Wegener, WoodChemistry,UltrastructureReactions, Walter de Gruyter, Berlin, p. 56 (1989). 616. B. L. Browning, The Chemistry of Wood, Interscience, New York, p. 88 (1967). 617. B. Sansoni and V. K. Panday, in S. Facchetti (ed.), Analytical Techniques,for Heavy Metals in Biological Fluids, Elsevier, Amsterdam, p. 91 (1982). 618. H. Small, T. S. Stevens, and W. C. Bauman, Anal. Chem., 4 7 1801 (1975). 619. J. B. Zicherman and R. J. Thomas, Holzforsch., 26: 150 (1972). 620. Useful Method 243, in TAPPI Useful Methods, Technical Association of the pulp and Paper 621. 622. 623. 624. 625. 626. 627. 628. 629. 630. 631. 632.

Industry, Atlanta, GA. G. D. Christian, Anal. Chem., 41: 24A (1969). M. T. Friend, C. A. Smith, and D. Wishart, Atomic Absorption Newslert., 16: 46 (1977). J. P. Price, TAPPI, 54: 1497 (1971). N. M. Arafat and W. A. Glooschenko, Analyst, 106: 174 (1981). T438 cm-90, in TAPPI Standard Methods, TechnicalAssociation of thePulpandPaper Industry, Atlanta, CA. H. Matusiewicz and R. M. Barnes, Anal. Chem., 5 7 406 (1985). S. Kokot, G. King, H. R. Keller, and D. L. Massart, Anal. Chim. Acta, 259: 267 (1992). J. Nieuwenhuize and C. H. Poley-Vos, Atomic Spectrosc. 10: 148 (1989). L. H. J. Lajunen, J. Piispanen, and E. Saari, Atomic Spectrosc., 13: 127 (1992). R. Lammi, Pnperi Puu, 63: 605 ( 1 98 1). E. King and 0. Schalin, Paperi Puu, 5 7 209 (1975). A. A.Verbeek, Spectrochim.Acta,39B: 599 (1984).

384

633. 634. 635. 636. 637. 638. 639. 640. 641. 642. 643.

Baeza and Freer D. C. Lee and C. W. Laufmann, Anal. Chem., 43: 1127 (1971). J. Marton and T. Marton, TAPPI, 55: 1614 (1972). J. A. Meyer and J. E. Langwig, Wood Sci., 5: 270 (1973). C. A. Osterhaus. J. E. Langwig, and J. A. Meyer, Wood Sci., 8: 370 (1975). H. E. Young and V. P. Guinn, TAPPI, 49: 190 (1966). R. A. Parham, Paperi Puu, 55: 959 (1973). J. H. E. Bailey and D. W. Reeve, Pulping Conf.. Book 3, San Diego, CA, p. 1233 (1994). M. L. Harder and D. W. Einspahr, TAPPI, 63: 1 10 (1980). A. Wong, Pulp & Paper Can., 84: 38 (1983). S. Saka and D. A. I. Goring, Mokuzai Gakkaishi, 29: 648 (1983). S. Saka and R. Mimori, Mokuzai Gakkaishi, 40: 88 (1994).

Color and Discoloration David N.-S. Hon Clernson University, Clemson, South Carolina

Nobuya Minemura Hokkaido Forest Products Research Institute, Hokkaido, Japan

1.

INTRODUCTION

Wood is an excellent material to absorb and reflect light. This physical interaction produces wood whose color may range from almost white, as in the sapwood of many species, to almost black, as in the heartwood of black ebony. The color characteristics depend on the chemical components of wood that interact with light. Hence, the reaction of wood components to light, heat, and chemicals will change the color of wood. Extensive studies and observations have shown that most, if not all, wood species of commercial importance, and in particular those used for furniture, paneling, and decks, are prone to discolor with age. Discoloration occurs both indoors and outdoors. Manyfactors and elements participate in the discoloration of wood. In this chapter, the major factors playing a role in discoloration, as well as the methods of removing and avoiding discoloration, are discussed. II. COLOR OF WOOD

The color of wood varies with wood species. In this section, a general concept of color, the coloration of wood, and characteristics of the color of wood species are discussed.

A.

Mechanisms of Coloration

Isaac Newton, the English physicist/mathematician, said, “Rays are not colored.” Color is recognized only when a rayof light enters the eye and is absorbed in the retina by light-sensitive receptor cells called cones and rods. Visible light, which produces the visual sense for human eyes, is part of an electromagnetic wave. Its wavelength ranges from 380 to 780 nm, as shown in Fig. 1. Ultraviolet (UV) light is at the lower end and infrared (IR) light at the upper end. Visible sensitivity varies with wavelength. In a dark place, a wavelength of 500 nm can be seen. In a bright place, the wavelength must be increased to 550 nm before the human eye can distinguish it. UV light does not reach the retina because it is absorbed into the cornea or crystalline lens; IR light reaches the retina but is not registered. 385

Minemura 386

and

Hon

Cosmic ray

Waveforradio

4

Sho-ave

-

X ray

Wave U e l e v i s l o n

Vacuum UV

radar Wave for

c,

uv

IR

"

Visible t",

I 1 5 00

400

P ' Db P: Purple

Y' Yellow

'

B

I

600

G

Db. Deep blue 0 : Orange ,VR

' Y '

0

0 : Blue R: Red

700 (nm) I

R G . Green

FIGURE 1 A portion of theelectromagneticspectrumshowingtherelationship region to other types of radiation.

of the visible

When visible light strikes an object, if all the light is reflected, we recognize the object's color as white. In contrast, when all the light is absorbed, we recognize the color as black. Most materials absorb certain wavelengths and reflect the rest. The reflected light is recognized as acolor, which is dependent onthe composition and amount of the reflected light. For example, the reflection of wavelengths longer than 590 nm produces an orange color. Absorption of light by a material excites its electrons. Generally, electrons are in the lowest energy state or ground state. If adequate energy is absorbed by the electron from outside, the electron will transit toahigherenergystate, or excited state. Light is an aggregate of photons that have energy, so depending on its wavelength, it can provide the energy necessary for electron excitation and transition. Betweentheenergyofaphoton ( E , kcal/mol) and thewavelength (A, nm), the following equation can be derived:

E=

2.86

X

10'

A

Figure 2 shows the relationship between energy and wavelength. The shorter the wavelength, the higher the energy will be The electron of an unsaturated bond (e.g., )C=C(, )C=O, )C=NH, -N=N-) can transfer easily to an excited state with a small amount of energy. In molecules containing many unsaturated groupings that are all conjugated, the molecular orbitals containing the electrons in the system will extend over these groups. The resulting high degree of delocalization of the electrons means that the energy required for a transition decreases. For example, one unit of )C=C( absorbs light at 190 nm, but @-carotene, in which I 1 units overlap, absorbs the light at 520 nm to give red. An atomic group having rr electron, such as an unsaturated bond, is called a chromophore. An atomic group having isolated electron pairs, such as " O H , " C O O H , and "OR, is called anauxochrome.Auxo-

387

Color and Discoloration (kcall

160' 150

~

140 -

130. 120 110-

F

100-

E, 5

80-

6

60.

W

go-

c

70-

50. 40.

30

200

400

600

800

Wavelength

FIGURE 2

Relationship between light energy quanta and wavelength.

chromes assist the action of chromophores by intensifying the coloration or enabling the absorption of light having a longer wavelength.

B.

Representation and Determinationof Color

The reflection curve in the visible region most accurately represents the color of a material. However, a representation with a numbered value is often useful. The numerical representation of color can be derived by two methods. One way is based on a comparison with a color specimen in which various colors are classified and numbered. Another way is based on trichromatic quality, which means that any color can be made by mixing three other colors. Color specification systems such as XYZ, Lab, L*a*b*, and UVW have been used to determine trichromatic quality [ 1,2]. There are two methods of determining color mechanically: by determining the percent of spectral reflectance and by reading tristimulus values directly. A spectrophotometer and standard white plate of magnesium oxide or magnesium carbonate are used for the formermethod.Thewhitenessof the white plate is considered to be 100%. Relative spectral distribution to it is shown by the reflectance curve in the visible region. The proportion of up-and-down areas of the curves relates to lightness. The upper part of the curve means high lightness. A photoelectric colorimeter is used for photoelectric tristimulus colorimetry. A test specimen is irradiated with a xenon light, the reflected light is collected with an integrative globe that leads to the XYZ light receiver, and it is converted to electric current by a photocell to indicate the numerical value.

I

Minemura 388

and

C. Characteristics

Hon

of Wood Color

Wood absorbs visible light. Consequently, we see a wood’s color as red, brown, yellow, and so on. Thesurfaceofwood is not uniform like metal; it is composed of cells of various sizes. Various cell volumes and the difference of components cause delicate differences of color even on the same wood surface.

Coloring Substances of Wood The main structural materials in wood are cellulose, hemicellulose, and lignin. Cellulose and hemicellulose do not absorb visible light, Native lignins that are isolated with minimum chemical or physical changes are pale yellow. In coniferous wood, lignin color can be attributed to phenyl-substituted benzoquinone and dehydrogenative co-polymers of coniferyl aldehyde. In wood, it is assumed that lignin is incorporated into a cellulose matrix and absorbs wavelengths below 500 nm [3]. Moreover, many woods absorb light beyond 500 nm due to the presence of phenolic substances such as flavonoids, stilbene, lignan, tannin, and quinone. In Fig. 3, spectral reflectance curves of woods are shown [ S ] . The lightnesses of these woods are different from each other. Numerical values of the color by the Lab specification system are alsoshown in the figure. Allof the woodsshown absorb light beyond 500 nm, and darker-colored wood absorbs more light. Ordinary sapwood has a lighter color than heartwood. The transition of sapwood into heartwood is accompanied by the loss of its physiological activity and the formation of various organic substances with darker color. When darker-colored woods such as rosewood and ebony are extracted with organic solvents, the extracted solution colors strongly, as shown in Table 1. Mostcoloredmaterials are presumablyhigh-molecular-weightpigmentswhich are insoluble in solvent [56], while the existence of colored materials for low-molecular1.

L

Q @ @ @

White birch Japanese larch Mizunara Black walnut

78.1 63.0 48.8 36.2

a 2.0 11.3 8.9 5.7

0 400

500

600

700 (nm)

Wavelength

FIGURE 3 Spectral reflectance curves and numerical color values for some woods.

b 16.3 21.5

16. 2 8.6

389

Color and Discoloration TABLE 1 Color of Dark-ColoredVeneer

Determined Before and After Extraction with Acetone

Ebony Before After Rosewood Before After

L

a

b

2.6 19.2 19.9

2.2 2.0

2.6

7.7 37.0 44.0

7.9 9.4

8.9

weightpigmentsuch known.

as mansonon F[4]and4-methoxy-dalbergion

[57] arealsowell

2. Physical Factors Affecting Wood Color a. Irradiating Direction of Light. When light is irradiated on the surface of wood, one part is reflected directly and the other part enters cells having voids and pigmentsthat absorb some wavelengths of the entering light. The light that is not absorbed in the cell is emitted again through scattering, reflection, and transmission. We recognize the unabsorbed light as the wood color. Wood cells are slender in shape and arranged in layers in one direction. Therefore, the wood color will be slightly different according to the irradiating direction of the light. Figure 4 illustrates the change in color, as shown by a Hunter Lab system, when light is irradiated at various angles toward the fiber direction of wood at an incidence angle of 45" [5]. Lightness is lowest when the direction of the progress of an incidence light is in accordance with the direction of the wood fiber. Lightness is highest when the an incident light crosses the wood fiber at a right angle. The behaviors of a and b shown in Figs. 4a and 4b are completely contrary to the behavior of lightness. Saturation is, therefore, lowest when the light meets wood fiber at right angles. Values a and b show the same tendency of increase and decrease. This means that the hue does not change. When the direction of the incident light meets the fiber direction at right angles, the quantity of the light that reflects and scatters on the surface without penetration into a cell might become larger, causing lightness to rise and saturation to become lower. b. Moisture Content. Unseasoned wood contains a significant amount of free water in its cells. When the inside of a cell is filled with water, light is transmitted deep into the cell but is scattered slightly in the cell wall. This wood color is called wetting color. As shown in Table 2, lightness is lower in unseasoned wood than in seasoned wood. The wetting color of seasoned wood is similar to the color of wood coated with a clear paint. c. Roughness of theSurjiace. If the wood surface is not even, the reflectance and scattering of light on the surface become larger, causing the lightness to rise, as shown in Table 3.

3. Distribution Sphere of Wood Color The distribution sphere of wood color for about 100 wood species, which are frequently used in wood-working industries, is shown in Fig. 5 by use of the Lab specification system.

Hon and Minemura

390

70

..Q .Q

.. ..

.. .. .. ..

68-i :

E -U

.

.-c 6 6 - ;

0

,. ..

..

.

.. .

..

..

0

0

.

:

.

0

180

360 (deg )

Angle

a

.

'2

60

180

1 0

360

180

(deg)

360 (de91

Angle FIGURE 4 Relationship between chromaticity index a and angle ofirradiatedlight. when light is irradiated parallel to grain; a and b are chromaticity indices.

TABLE 2 Color of Todomatsu Sapwood Determined Before and After Drying

Green wood with moisture content of 175% Wood with moisture content of 13% dried at room temperature

L

a

b

68.9

5.3

20.8

80.9

1.4

16.0

TABLE 3 Color of Yew Determined After Planing and Grinding ~~

L

a

h

53.6 58.9

18.0 15.7

22.9 22.2

~

After planing Grinding with sandpaper #320

Angle is 0

391

Color and Discoloration 30 0

OD

35

o

0

25 o

0

o

0

€jb

0

0

20

30 0

0

0

0

15

25

b

€P

8

B,

Ip

20

0

0 0 0

oo

0

0 0

10

0 0

0

0

8 0

I

15

5

W 0

0

W O

FIGURE 5

10

0

0

5

a

10

15

20

Distribution sphere of color of various woods.

Allof the woods distribution are in positive sphere of chromaticness indices a and b. Numerical values of lightness range from 20 to 85 [5]. In Figs. 6 and 7, the correlation of three color factors by the Munsell specification system is shown [55,58]. When reddish extent becomes large, Munsell value decreases. When yellowish extent becomes large, Munsell value increases. Saturation is in proportion to Munsell value for typical tropical woods.

Wood Color and Its Use Color affects human feelings in various ways. Wood is a natural material and its color seems compatible with human life. As compared with the color of plastic and concrete, that of wood conveys peace of mind and a feeling of natural gentleness. On the earth, forest resourceshavealwaysabounded,andwoodhasbeenusedwidelysinceancient timesasamaterial in construction, farming tools, furniture,carving,and so on. Most often, wood has been used because of its pleasant color as well as its warmth, hardness, and strength. a. White Wood. The image that white projects is purity, freshness, and sacredness. White woods include poplar, mangashinoro, white lauan, igem, yellow cypress, fir, white birch, shinanoki, and the sapwood of hinoki. These woods are used in building construction, obsequies, chopsticks, toothpicks, wood shavings, and so on. Because of hinoki's fragrance and excellent durability, its sapwood is considered the best construction material for shrines and palaces. White wood can be changed to any color by dyeing. White woods with high penetration are therefore used widely as basewood materials for dyeing. Because white wood shows dirt, it is normally used in areas that people do not touch. b. Red Wood. As it appears in autumn leaves or flowers, red is a passionate color. Because red harmonizes well with green, it works well as the exterior color of a house. The red heartwoods of sugi and hinoki have the highest value as building materials in

4.

392

a

-*

K

k

r-

0

U

U

r-

I- g

. r

d C

3

0 0

d .-m

P)

c -U C d

Hon and Minemura

393

Color and Discoloration A

AA

2

1

2

3

4

5

6

7

8

Saturation FIGURE 7

Relationship between Munsell value and saturation of typical tropical woods in Asia.

Japanese-style houses and are widely used as ceiling board, wall paneling, and posts. Red birch is popular for furniture and interior doors. Rosewood and Chinese quince are widely used as decorative materials in house construction, as well as for carving, instruments, high-quality art objects, and so on. c.Yellowish-BrownWood. Yellow and orange colors convey warm, homelike feelings. Teak and keyaki are widely used as building materials and for furniture. Tsuge is used for stamps and combs. Keyaki, which has a peculiar pattern, is the most precious wood. d.Light Brown Wood. Mizunara is a light-colored, expensive material. Mizunara and nire are used for furniture as well as window frames, building construction, highquality goods, and so on. Trend in color furniture has been considered an indication of the economic climate. For example, if light-colored woods such as mizunara are popular, it implies economic recovery. In contrast, if dark colors such as deep red are popular, it suggests an economic recession. e.BlackWood. Blackconveys the impressionoforderandcalm.Ebony is used for carvings, furniture, decorative materials in home construction, Buddhist family altars, and so on. 5 Contrast of Color in Sapwood and Heartwood. The color of heartwood and sapwood differs in manywoods.Thisdifference is often used decoratively. Forexample, decorative pillars of ebony display the brown color of heartwood and the white-yellow color of sapwood; decorative pillars of hinoki utilize the natural contrast of red and white. Such a contrast also is emphasized in fancy goods, souvenirs, cufflinks, and so on.

111.

DISCOLORATION OF WOOD

A.

Types of Discoloration

As wood is a biological material, it is decomposed by microorganisms and reacts chem-

ically when it comes into contact with substances such

as metal ions, acid, and alkali.

Hon and Minemura

394

Because wood is porous, water-soluble substances or salts are often deposited in its voids during the course of growth or after logging. Such deposits can change the wood’s color. Except in the caseofdecay by microorganisms,discoloration does not meanan accompanying loss of wood strength and is usually limited to the surface layer. Because wood color is an important factor that strongly affects value, discoloration is a serious problem from the viewpoint of commercial worth. Table 4 summarizes the characteristics of discoloration according to the factors that influence the change in color.

B. Characteristicsof Various Types of Wood Discoloration 1. Discoloration by Light A newspaper left in a window turns yellow in a few days. If a calendar is hung on a wall of natural wood, the part covered by the calendar retains its original color, whereas the uncovered part changes color. Photo-induced discoloration is consideredundesirable in manycases.Discoloration is alsoafactor in the manufactureofhigh-yieldpulps that contain a high amount of lignin. In this section, the characteristics of photo-induced discoloration of various wood species and methods for preventing it are described. a. Discoloration Behavior of Various Woods under Light Irradiation Classification of thePattern of Photo-Induced Discoloration. Every wood changes color with light irradiation, but the rate and course of change varies with wood species. Tables 5-8 show the quantity of photo-induced discoloration and the declining rate of whiteness of 100 commercially used wood species [ 5 ] . These results are obtained from the accelerating test by use of a carbon arc light as the light source. The measured values are classified according to the difference in discoloration with elapsed time. The color differences are calculated with the Lab system on the basis of the original color before light irradiation. A large color difference means a large amount of discoloration. Lowering rates of whiteness are calculated by dividing the difference in whiteness after light irradiation by the original whiteness before light irradiation. An increase in the lowering rate of whiteness indicates darkening; a decrease indicates lightening. As shown in Tables 5-8, discoloration is classified into five patterns in an elapsed time of 100 h: darkening only, darkening and then fading, darkening-fading-darkening, fading only, and fading and then darkening. The quantity of photo-induced discoloration after 100 h of exposure in most woods is beyond A E = 3, which is the limiting value that can be distinguished by the naked eye. The value of the most intensive discoloring wood is AE = 25. The relationship between the chromaticness index a of the original wood color and the quantity of photoinduced discoloration is shown in Fig. 8. It is well known that the magnitude of the a value has a positive correlation with the extent of redness. As illustrated in Fig. 8, woods with a white color show significant darkening. Many softwoods continue darkening with light irradiation, and many tropical woods discolor with a mixture of darkening and fading. In woods an intensive discoloration at the initial stage often is attributed to extractives. Change of Hue or Saturation. The discoloration of the above 100 wood species is classified by the change in a. The elapsed changes of hue andsaturation are also outlined as follows [ 5 ] . l.

Group showing an increase in a with light irradiation. Many woods in this group increase in saturation and discolor toward orange. As illustrated in Fig. 9, the value of b shows a small decrease at an initial state of irradiation, followed by a significant increase. It decreases again after 50 h of exposure. This tendency

TABLE 4 Characteristics of Wood Stain Classification After logging

-

Addition of source of stain from the outside

Cause of stain ~

Biological source

Propagation of microorganism

Blue stain

Chemical source

Bonding of metal ion Bonding of acid Bonding of alkali Heating Irradiation of light Metal ion Enzyme Resin Imperfect pruning Deposition of substance

Iron stain Reddish discoloring of zelkova Adhesion of cement Sticker mark Discoloration by sunlight Blackish discoloring of sugi Red discoloration of alder Exudation of resin Brown stripes Existence of specks

Physical source Immanence of source of stain In shade

Example

~~~~

Hon and Minemura

396

TABLE 5 Quantity of Photo-Induced Discoloration and Declining Rate of Whiteness of Woods After Exposure to Carbon Arc Light After exposure for 100 h

Species Yellow cypress Sitka spruce Aspen Douglas fir Lawson cypress Noble fir Listwennitza, Larix dahurica Corean pine Red cedar Western hemlock Red oak White lauan Champaka New Guinea basswood Amberoi White cheesewood Tetrameles Manggasinoro Evodia Ipoh Celtis Ramin Canarium Sterculia Kedondong Antiaris Agathis Japanese poplar Todomatsu, Abies sachalinensis Shinanoki, Tilia japonica Shirakaba, Betula phatyphylla var. japonica Ezomatsu, Picea jezoensis Plane Hinoki, Charnaecyparis obrusa Buna fagus crenata Japanese yew Hiba, Thujopsis dolabrata Japanese larch Magnolia Painted maple Zelkova Camphor tree Sen, Kcdopanax pictum Japanese red birch Radiation on wood surface: 4032 cal/crn'.

Color difference (AE)

Declining rate of whiteness

23.4 21.7 21.6 19.1 17.2 16.1 15.2 14.8 14.4 11.5 7.7 21.1 20.8 19.8 19.7 16.9 16.1 13.5 13.3 12.3 12.0 10.9 10.1 7.2 6.7 6.3 5.4 24.7 23.8 21.8 21.3 21.2 20.3 18.5 18.3 17.5 15.2 14.7 13.7 11.2 9.8 9.0 6.2 5.5

32.9 19.3 27.5 30.3 25.9 26.2 23.2 24.5 27.4 18.2 13.8 28.4 32.6 26.8 26.1 22.6 26.2 22.4 17.8 15.4 15.9 17.6 17.4 7.4 10.6 9.3 9.8 31.4 33.7 27.8 26.7 30.2 23.5 29.5 25.1 20.1 20.8 25.7 22.8 19.4 14.3 13.3 9.0 10.5

(%)

397

Color and Discoloration TABLE 6 Quantity of Photo-Induced Discoloration and Declining Rate of Whiteness of Woods That Change from Dark to Faded and to Dark After Exposure to Carbon Arc Light After exposure for 100 h rateDeclining Color difference Species Redwood Jongkong Teak Santiria Yellow hardwood Terminalia Spondias Miwa mahogany Elaeocarpus Myristika Trichadenia Box wood Aogatsura, Cercidiphyllumjrcponicurn (dark) Formosan cypress Kiri, Paulowina tornentosa Kihada, Phellodendron arnurense Sugi, Cryptomeria japonica Swamp ash Higatsura, Cercidiphyllurnjnponicurn (pale) Japanese alder Values in parenthesesindicatehigherdiscolorationwithin determined after exposure of 1 0 0 h.

(

of whiteness

W

9.5 13.0 (13.6) 12.9 9.4 8.3 7.3 7.3 6.3 (6.7) 5.7 5.4 (7.1) 4.6 12.0 11.5 8.5 8.2 7.4 (7.9) 4.8 4.7 4.7 (5.1) 4.0

15.3 23.3 20.9 16.3 10.0

13.3 11.3 8.9 1.7 5.9 4.0 16.9 21.3 15.6 7.2 6.9 9.0 1.5

8.1 1.1

1 0 0 h of exposure,comparedtothevalue

often is observed in white wood, accompanying a high degree of discoloration. Spruce, Douglas fir, and todomatsu belong to this group. This group also contains woods that have only a decreasing value of b as well as no change in this value. At the final stage of light irradiation, woods in this group are deep red in color. 2. Group slightly showing changes in a . This group generally evidences significant fading. The value of b increases greatly and the color changes toward yellow. Rosewood and walnut belong to this group. 3. Group showing a decrease in a. Some woods in this group also show a decrease in the value of b and are nearly achromatic in color. Japanese yew and Chinese quince behave like this. In this case, lightness decreases and whiteness declines as the irradiation time increases. Otherwoods in this groupshow an initial decrease in h, followed by an increase. Many tropical woods with a dark color belong to this group. 4. Groupshowingrepetition of increaseanddecrease in thevalue of a. Many woods in this group show a slight photo-induced discoloration. The woods in this group also often show repetition of increase and decrease in the value of h.

Change of Lightness and Whiteness. Woods that alternately show darkening and fading have, in many cases, slight photo-induced discoloration. The quantity of photo-

Hon and Minemura

398 TABLE 7 Quantity of Photo-Induced Discoloration and Declining Rate

of Whiteness of Woods That Change from Dark to Faded After Exposure to Carbon Arc Light

After exposure for 100 h Color difference Species Melapi African mahogany Silkwood Nato Rosewood Pterocarpus Andes rose Artocarpus Kapur Calophyllum Malas Red lauan Rengas Kossipo Sapele Taun American walnut Eugenia Dao Zebra wood Sloanea Eurasian teak Kingiodendron Ebony Dracontomelon Maniltoa Yamaguwa, Morus bornb.ycia Shiurizakura, Prunus ssiori Locust tree Mountain cherry Mizunara, Quercus crispuln Japanese walnut Japanese hophornbeam

(

W

21.6 (23.6) 15.3 (15.9) 11.4 (12.6) 1 1.4 (15.5) 10.4 (9.5) 9.9 (8.2) 9.6 9.4 (10.3) 9.3 (10.4) 9.3 (10.6) 8.4 (8.8) 7.5 (8.0) 5.9 5.7 (4.5) 5.7 (6.9) 5.6 (6.4) 5.3 5.0 (7.7) 4.4 (5.0) 4.2 (4.5) 3.7 (7.1) 2.1(4.0) 4.0 (3.1) 3.5 (0.9) 2.1 (4.0) 2.0 (3.0) 19.9 (22.8) 12.9 (13.8) 11.7 (13.0) 8.7 (10.3) 5.0 (7.5)

Declining rate of whiteness (96)

30.0 28.8 18.8

17.3 12.2 4.7 -27.5 13.1 13.7 12.6 1.1

13.5 -5.2. 1.6 5.1 5.1 - 5.6

5.5 (1.8)

1.7 6.4 5.2 2.6 1.9 -6.3 - 15.6 -2.5 -5.1 26.7 21.4 16.3 13.3 8.6 -5.3

2.2 (4.9)

-0.5

Values in parentheses indicate greatest photo-Induced discoloration within exposure

of 1 0 0 h.

induced discoloration is calculated based on the total difference of the lightness and chromaticity indices before and after light irradiation. As lightness generally changes more than chromaticity indices, lightness and the quantity of photo-induced discoloration have a strong mutual relationship. Whiteness is also calculated from the sum of the lightness and chromaticity indices. As the lightness of wood is higher than chromaticity indices, whitenessandlightnesscorrelate in mostcases. In the case of asmallchange in the chromaticity indices, the declining rate of whiteness and the quantity of the photo-induced

399

Color and Discoloration TABLE 8 Quantity of Photo-Induced Discoloration and Declining Rate Whiteness of Woods That Fade After Exposure to Carbon Arc Light

of

After 100 h of exposure Declining rateColor difference Species

(

Fading only Rosewood Indian rosewood Fading-darkening Teijsmanniodendron Nire, Ulmus davidiann davidiana var. japonica

26

8.2 5.2

-9.7 -1.1

0

. A

8

18 -

A

0 0

0 0

0

1614 -

t L

12 -

5

10

V

- 30.3 -50.8

0

20 -

L

15.9 12.7

0

22 -

2

of whiteness (%Io)

-

24 -

%

W

0

0 0

0

00 A

B 0 0

A

o

-o

a

0

6 4 -

d o

0

0

02

AAA A 0

A

2 -

A

ta

0

0

8 -

A

A'

0 " " " " " ' 0 2 L 6 8 18 10 161412 a ChangeInduced by exposure 0 ; Darkening (0) , b ; D + Lightening (L) 0 ; D+L+D. '

FIGURE8 Relationship between chromaticity index LI and quantity of photo-induced discoloration for various woods during exposure t o a carbon arc light for 100 h.

400

Hon and Minemura

26 -

24

b

-

22-

/

20

-

l8 L 2


Numerals in the figure show exposure (hr) to carbon arc light.

!ime

I

4

6

0

10

12

14

a FIGURE 9 Change of chromaticity indices ( a and h ) of a todomatsu during exposure to a carbon arc lamp for 100 h.

discoloration also have a mutual correlation. In the 100 wood species described earlier, 10% haveahighervalue of whitenessafter 100 h of light exposureascompared to unirradiated wood. Photo-Induced Discoloration of Sapwood and Heartwood. Comparedwith heartwood, sapwood has a pale color. The photo-induced discoloration of woods that have a distinct contrast in color betweenheartwoodandsapwoodshows that the sapwood usually keeps darkening, even if the heartwood changes from darkening to fading in color. The quantity of the photo-induced discoloration, therefore, is higher in sapwood than in heartwood. The discoloration pattern is similar to that illustrated in Fig. 9. It changes in the direction of red with high saturation. Photo-Induced Discoloration WhenLeft Indoors for a Long Time. Timber of 50 wood species were left indoors for 1800 days and their photo-induced discoloration determined [59]. As shown in Fig. 10, the color of the timbers changed in various directions within 780 days, but after 980 days all the woods changed only in the yellowish direction. In Fig. 11 the change of photo-induced discoloration of skinanoki, as a function of the change of irradiated wavelength, is shown. Photo-induced discoloration becomes large when the wood is irradiated with light in the near-UV region. When the wavelength changes from short to long. hue changes are as follows: yellow + red + purple + blue. b. WavelengthsParticipatinginPhoto-InducedDiscolorution. When 75 kinds of commercial wood were exposed to light, 62% of the woods discolored with UV light and 28% of the woods discolored with visible light [6]. Figure 12 shows the amount of photoinduced discoloration when the sapwood and heartwood of karamatsu were covered with various glass filters andthenexposed to light [7]. In Fig. 12, a positive valuemeans darkening and a negative value means fading. It is clear that the heartwood of karamatsu discolorsstronglywith UV light and slightly with visible light. In the photo-induced discoloration of sapwood, light over 390 nm gives rise to lightening, and light under 390 nm brings about darkening. The wavelength range for lightening (or bleaching) is consideredtobebetween 390 and 580 nm. This range for lightening is also discernible in a photo-irradiated newspaper 181.

401

Color and Discoloration b* b' I

35 -

30 -

25 -

20 -

15-980th

day

lot

FIGURE 10 Direction of photo-induced discoloration of 50 popular wood species whenleftindoors for 1800 days: from 0 day to day 780; right, from day 980 to day 1800.

kfi.

310 nm

I

I

I

I

I

I

a'

FIGURE 11 Direction of photo-induceddiscoloration of shinanokiwhenirradiatedwithvarious wavelength lights (numbers indicate wavelength of irradiation).

402

-5

Hon and Minemura

t

u

t

I

UV Non

LOO

500

I

I

600

700 (nm)

Lowerlimit

of transmittedwavelength

UV ’ Filterwhtch transmitsultraviolet only, Non. Filter was not used, A : Heartwood , 0 A 0: SapwoodExposure time . e 0 , 25 hr, A A , 100 hr 0 , 200 h r .

,

FIGURE 12 Discoloration of a Japanese larch covered with various light filters when exposed to

xenon light.

Rosewood shows typical fading with light irradiation. The influence of various wavelengths on the fading is shown in Fig. 13 [ 5 ] .The fading becomes stronger when UV light with shorter wavelengths is used. The previously described range for lightening does not fortify fading in this case. Figures14and 15 show the rates of increaseanddecrease in the reflectance of irradiated karamatsu which was covered with various filters during exposure, against unexposed wood [7]. When a filter was not used or when only UV light was used, heartwood and sapwood changed toward redin color. The change in reflectance shows the largest decrease at 410 nm, likely due to the formation of a quinonoid structure [ 91. Judging from the irradiation wavelengths and the pattern of the reflectance curves of karamatsu heartwood, it appears that wavelengths between 300 and 390 nm cause discoloration to red, 390 to 580 nm cause discoloration to yellow, and over 580 nm results in scarcely any change. Wavelengths of 390 to 590 nm also cause lightening of heartwood. In the sapwood of karamatsu, light of 300 and 390 nm cause yellowing, 390 to 580 nm cause lightening, and over 580 nm cause no discoloration. That hue and saturation differ with wavelengths probably means that the reaction occurring in wood during photoinduced discoloration isnot simple. Details of the photochemistry and photoxidation related to discoloration can be found in Chapter 11. c. Wood Cor?lpound.s Relutecl toPhoto-Induced Discolorcrtion. Most woodcompounds related to photo-induced discoloration are high-molecular-weightcompounds which are insoluble in solvent [56],although a few low-molecular-weight compounds are known. In western hemlock sapwood, five lignans. one neolignan, and three minor con-

403

Color and Discoloration

-5 C

= D .c

-10

-

----"""_

-b

0 0

-0

-15 -0 ~~

0

50

25

100

75

(W

Exposuretimetocarbonarclight Transmittedwavelength

,

x

, Infrared only

,

0 ; Over

660 run , A ; Over 630 nm.

.

Over 500 nm , v , Over 530 nm , B Over 480 nm. A , Over 430 nm , Over 370 nm , A , Over 360 nm , o ;Over 320 nm, 0 , Ultravidet only,

0 ,

I

v . No fillerwasused.

FIGURE 13 Photo-induced discoloration of a rosewood covered with various light filters.

stituents have been isolated as the causing materials for photo-induced discoloration [60]. Almost all ofthemhave a quaicylringstructureandanoxygenationstructureat the neighboring a-position on the aromatic ring. Concerning the photo-induced discoloration of western red cedar, the participation of the causative compounds are plicatic acid, plicatinaphtol, and plicatinaphtalene in a rough ratio of 5:2:2 1611. In rosewood, 4-methoxydalbergione has been isolated as the causative material [57]. d. Restoration of Sound Color to Discolorated Wood. One method for restoring sound color is todecomposethesurfacechromophoricstructuresofthesystem with bleaching chemicals such as hydrogen peroxide or sodium chlorite. Another method is to sand the surface with sandpaper or a plane. Surface treatment is effective because discol-

E

o

EY

-20

c

0,

Q

-l0

.-c

-30

g

-40

C

2

V

-50

500 Wavelength

600

700 (nm)

FIGURE 14 Change in percent reflectance o f a karamatsu heartwood that was covercd with light filtersandthenexposed to a xenonlight.Transmittedwavelength: 0 : Ultraviolet only (300-415 n m ) , 0 : Over 430 nm. 0 : Over S80 nm. A: N o lilter was used. Light source: Xenon light.

404

Hon and Minernura

("IO)

2C 30

-

2

20 -

G

10

?!

-

0 c a, Y -10 Q a -20 c ._

+

0 Mizunara

0 Sawagwurnl

0

5 V

-40

0 Kri A Bund

-50

400 600

500 Wavelength

700 (nm)

FIGURE 15 Change in percent reflectance of a karamatsusapwood that was covered with light filters and then exposed to a xenon light.

oration is usually limited to the surface layer. For a faded surface, painting with dye or pigment is also effective. e. ControllingPhoto-InducedDiscolorution. Although the photo-induced discoloration of wood can add to an impression of dignity and age, for most applications discoloration is regarded as an unfavorable reaction. Various mechanisms have been proposed with respect to photo-induced discoloration. On the basis of the accepted mechanisms, one or acombination of thefollowingmethods may beadopted to prevent photo-induced discoloration:

1. Cutting off UV light 2. Modifying the light-absorbingstructures 3. Destroyingthestructuresparticipating in discoloration 4. Eliminating oxygen or capturing the singlet oxygen 5. Scavengingfreeradicals 6. Extractingtheprecursors of discoloration The method selected must not damage the original wood color. It must be easy to carry out, low in cost, and safe. Cutting off UV Light. One way to prevent photo-induced discoloration is to coat the wood surface with asubstance that absorbs UV light, i.e., to use a UV absorber. Normally, commercially available UV absorbers are colorless or pale yellow and absorb light below 400 nm. Some UV absorbers change their structure by absorbing light and releasing the absorbed energy as heat to regain its original structure without degradation. About 35% of 70 commercial wood speciesdiscolored by light canbecontrolled by coating with UV absorbers [6]. The general treatment is to coat the wood surface with organic solvents or paint in which the UV absorber dissolves. Coating with paint is ideal because a film of UV absorber forms on the wood surface, preventing degradative UV light from reaching the surface. However, if the film is peeled off, the protection disappears. A better method is to use a UV absorber with functional groups that will react with wood. For example, coating with 2-hydroxy-4-(2,3-epoxypropoxy) benzophenone. accompanied by heating under pressure,

Color and

405

results in high photostability [ 101. When 2-hydroxy-4(3-n1ethacryloxy-2-hydroxypropoxy) benzophenone, in which an acryl group and UV absorber are bonded, is used, it reacts with wood components and has the same effect as paint. This treated wood also has high photostability [ I 1 I. Titanium oxide and zinc oxide effectively cut off UV light. They also cut about 60% of visible light. When used on white-colored wood with a clear grain, these materials have achieved considerable controlling effects. A microfine powder must be suspended into an organic solvent before coating. For the purpose of preserving, packing, and transporting wood products, the use of a film sheet containing UV absorbers or pigment is desirable. Wood usedoutdoors takes in sunlight and rain. As a result, celluloseand lignin decompose. The surface becomes rough and fine sands or dirts can be deposited on it. To prevent deterioration, an oil-stain coating that contains a light-resistant fine pigment and a water-repellent such as paraffin is effective. It is assumed that the pigment diminishes the penetration of light onto the wood surface and controls the photodeterioration. Veneer which is impregnated with the mixture of polycarbonate resin and toluene diisocyanate has less transparent, high wood feeling and high weather-proofing properties for exterior use [62]. Modifying the Light-Absorbing Structure. Since the a-carbonyl conjugated carbon-carbon double bond and phenolic hydroxyl groups are the principal chromophoric groups in wood, they can be modified to reduce discoloration. The treatment of the phenolic hydroxylgroup by acetylation, methylation,andbenzoylationdecreasesphoto-induced discoloration [12].Acetylation is effective only for control at an early stage. Prolonged irradiation givesphenoxy radicals as a result of degradative reactions such as deacetylation and subsequently forms chromophoric groups [63]. A purple pigment that has been isolated as the causative material of photo-induced discoloration in black walnut has a carbonyl group and a chelated hydroxy group. To prevent the discoloration of this wood, the use of 3,5-dinitrobenzoylation is most effective [64]. If diazomethane is used in methylation, it changes the a-CO group to the oxirane structure [13]. When the a-CO of MWL is reduced byNaBH,, the quantity of photo-induced discoloration reduces to one-fourth that of the untreated wood. Furthermore, when the conjugated double bond is hydrogenated in the presence of a catalyst, no light-induceddiscoloration is observed [ 141. For the purpose of chelate formation with a phenolic hydroxyl group, a coating of ferric ion [ 151 and chromium ion [ 161 is effective. However, the wood color becomes dark after coating. Coating with semicarbazide to modify carbonyl groups has a large controlling effect on the initial discoloration of many woods, as shown in Table 9 [17]. The relationship between the amount of coating and the controlling effect for discoloration is shown in Fig. 16. The favorableamount of coating is 5- I O g/m2. A higher controlling effect is obtainable by combining semicarbazide with titanium oxide. Because semicarbazide also has a high capture effect forformaldehyde [ 181, using it onplywoodorparticleboard bonded with urea resin can control both photo-induced discoloration and the release of formaldehyde. Because this chemical reacts with resin components, in addition to reducing photo-induced discoloration, the control of resin exudation is also recognized for Douglas fir [ 191. In some cases, lightness rises slightly immediately after coating with this chemical. This may be due to the partial rupture of the conjugated system that takes place when woodcolors.Alkylderivatives of semicarbazideandthiosemicarbazidehave B similar effect. Using a polyurethane top coat on a wood surface coated with semicarbazide causes noproblems [ 5 ] . Besidessemicarbazideand its derivatives, coatingwith the following

406

Hon and Minemura

TABLE 9 Quantity of Photo-Induced Discoloration of Woods Coated with Sernicarbazide

Species

Coated

Shiurizakura Nato Japanese yew Douglas fir Todornatsu Sitka spruce

Japanese larch Pterocarpus Redwood Japanese walnut Hinoki Painted maple Corean pine Noble fir Zelkova Ezornatsu Sugi Red lauan Western hemlock Kiri

Japanese red birch

Uncoated 15.5 15.5 15.0 12.0 10.4 10.3

5.0

3.8 5.9 3.2 4.0 5.3 3.2 4.0 3.1 5.2 2.3 3.8 3.6 2.1 3.2 3.4 4.2

10.0

9.3 9.1 9.0 7.9 7.8 7.6 7.5 7.4 7.2 7.0 6.2 6.2 5.1 5.0

Coatweight: I 1 glm’. Irradiation: Fade meter with carbon arc light

1 .S 2.5

2.6 I .5 for 10 h.

reducing chemicals is somewhat effective for initial control: Na,SO,, NaHSO,, Na,S20,, ( N H M O , , (NH,),S2O3, ascorbic acid, etc. For some tropical woods, control is achieved in several ways [65]. Sodium chlorite treatment gives the best result for white meranti, while acetylation is best for selangan batu and sepetir. Hydrogen peroxide treatment and methylation give the best result for yellow meranti and nyatoh, respectively. Coating of chitosan onto Douglas fir wood leads to a decrease of photo-induced discoloration [66]. Destroyingthe Structures Participating in Discoloration. Destroying the functional groups and precursors of discoloration is also a method of preventing discoloration. Acetylationcombined with oxidativebleaching and treatment with NaBH, areknown [2O,2 1 1. Wood coated with polyethylene glycol (PEG) becomes white with light irradiation, as shown in Table I O [22]. Figure 17 shows the behavior of photo-induced discoloration of manggashinoro coated with PEG. As the amount of coating increases, the values of n and b diminish,implyingdiscoloration in the direction of achromatic color. Figure 17 shows the increases of lightness and whiteness, and Fig. 18 shows the effects of discoloration for various woods. This treatment works well on light-colored wood such as manggashinoro, ezomatsu, and Douglas fir, but has negative effects on dark-colored wood such as walnut and rosewood, because it turns such woods white. Regardless of the size of the molecule, PEG has controlling effects on discoloration. The coated wood, however, has a tendency to be sticky and look wet because of the hygroscopic nature of PEG, especially if a large amount of low-molecular-weight PEG is used. Although wood bleached with an oxidativebleachingagentsuch as hydrogen peroxideorsodiumchloritebecomesdark

407

Color and Discoloration

0

50

25 Exposure time to

75 carbon arc light

100 (hr)

FIGURE 16 Photo-induceddiscoloration of a Japanese larchcoated with variousamounts semicarbazide. Coat weight of semicarbazide (g/m2): 0 ; 0,A; 1.7, 0 ; 3.7, 0 ; 7.3, A; 11.

of

when irradiated with light, a coating of PEG on bleached wood has a good controlling effect on discoloration [23]. When newsprint made of softwood is coated with PEG and irradiated with light, it gives lower levels of photo-induced discoloration [24], lower alkaline extract, and lower initial radical formation than the untreated form. As shown in Fig. 19, PEG-coated filter papers impregnated with various phenolic substances gives very low quantities of photo-induced discoloration [25]. For whitening of wood coated with PEG, the mechanism was elucidated 1551. When the terminal hydroxyl groups of PEG are methylated or methacrylated, the effect is the same as with unmodified PEG [ 5 ] . When saturated hydrocarbon is replaced by hydroxyl groups, the same controlling effect is recognized. This suggests that the hydroxyl group plays an important role. Polypropyleneglycol (di-ol type) exhibitsasimilarcontrolling effect. Pulp which is saturated with alcohol is bleached when it is irradiated with nearultraviolet in the presence of oxygen [67]. This is presumably due to the formation of ahydroxyhydroperoxides from alcohol. The structure of polyoxymethylene is similar to that of PEG. It is well known that, with light irradiation, this compound splits the main chain to generate free radicals -CH,O. and .CH20- [26]. Based on the above-described facts, the following mechanisms are considered. When wood coated with PEG is irradiated with light, radicals are generated by two different routes. One route is by abstraction of hydrogen in the ethylene oxide chain by the excited a-carbonyl group. The other is photolysis of PEG. The radicals being generated react with oxygen in air to give peroxy radicals. These reactions are summarized below:

@ I

C=O

I

I + h v + C=O” l

408

Hon and Minemura

TABLE 10 White of Wood Coated with Polyethylene Glycol Determined After Exposure to Carbon Arc Light for 100 h ~~

Species Yellow cypress Lawson cypress Western hemlock Listwennitza Noble fir Sitka spruce Corean pine Douglas fir Red cedar Ipoh Amberoi Celtis Newginea basswood White cheesewood White lauan Manggasinoro Melapi Rosewood Plane Shirakaba Shinanoki Buna Japanese poplar Painted maple Kihada Hiba Ezomatsu Todomatsu Hinoki Yamaguwa Shiurizakura Japanese larch Japanese walnut

Uncoated

Coated

48.2 44.5 44.3 44.3 43.6 43.5 41.6 39.6 34.0 53.6 48.7 48.2 48.2 48.2 44.8 44.6 44.0 42.4 59.0 53.2 52.4 50.7 50.3 46.4 46.0 43.3 43.2 42.6 42.5 40.9 38.6 38.4 33.8

51.7 49.8 46.1 45.2 54.2 50.0 44.0 44.0 45.4 57.2 52.3 53.1 50.6 49.8 46.5 55.8 55.5 42.2 64.8 63.6 61.2 60.5 54.4 48.9 53.4 49.4 50.7 50.8 55.6 42.4 46.1 43.7 41.9

Coat weight of PEG: 1 I g/m'.

0-0. -CH?-CH-O-

I + 0, + -CH2-CH-0-

@ -CH2-CH2-O-CH2-CH2-O-

+ h u 4 -CH2-CH2"O*

+ .CH2-CH2-O.CH2-CH,-O-

+ 0 2 + .00-CH2-CH2-O-

These peroxy radicals will destroy the coloring structure or its precursor. PEG does not absorb sunlight, but it does associate with phenolic compounds. Therefore, in wood coated with PEG, there is the possibility that a complex is formed between PEG and penols or

409

Color and Discoloration

0

25

50

Exposuretimeto

75 carbon arc light

100 (hr)

FIGURE 17 Photo-induced discoloration of woods coated with polyethylcne glycol after exposure to a carbon arc lamp for 100 h. Coat weight of PEG (gln?): 0 ; 0, 0 ; 6 , A: I 1, X ; 22. Numerals in the figure mean exposure time.

in combination with lignin that absorbs light. Use of a peroxide such as benzoperoxide PEG shows a higher whitening effect in some cases [25].In the use of wood coated with PEG, additional coating with paint is necessary. The adhesivestrength of polyurethane film on wood coated with PEG is high even after light irradiation. It is assumed that a urethane bond is generated due to the reaction of PEG with isocyanate. Instead of paint, a coating of wax may be used 1271. PEG is also effective for whitening of yellowed paper [ 5 5 ] . Eliminating Oxygen or Capturing the Singlet Oxygen. If light penetrates a surface isolated from oxygen, discoloration does not occur. Wood-plastic complexes show lower rates of discoloration when exposed to light. However, since void space in a wood cell is filled with plastic, the diffused reflection of light vanishes and the following phenomena occur: appearance of wet color, decline in lightness, and a rise in saturation level. A singletoxygenquencher traps the excitedenergy of IO2 that actsasacatalyst duringphotoreaction.p-Carotene retards the light decomposition of lignin model substances [28].Nickel complex and 1,4-diazabicyclo[2,2,2]octanearealso well known as singlet oxygen quenchers. Many quenchers are colored substances and do not regenerate after quenching. Hence, applications of single oxygen quenchers to wood are limited. Scavenging Free Radicals. A coatingcompound that has activehydrogens such as phenolic derivatives and phenolic amines can be used to captureradicals.However, since it is not reclaimed after use and often crystallizes when used in a thick coat, its use in wood is limited. Extracting Solvents. Extraction can remove discoloration that has been caused by solvent-soluble components. This has been reported for pencil cedar [29]and rengas (30) extracted with methanol. Incomplete extraction can cause an accumulation of the discolored components on the surface, due to migration from the inner wood when the solvent evaporates.

Hon and Minemura

410

0

Color difference A € 5 10 15 20 25

Manggasinoro Hinoki Melapi Larch Douglas fir Redcedar Ezomatsu Hiba Todomatsu Shiurizakura Yamaguwa Kihada Walnut Rose wood Coat weight of from the top

PEG (g/ma) : , 0, 6 , 11, 22

FIGURE 18 photo-induced discoloration of woods coated with polyethylene glycol after exposLIre t o a carbon arc lamp for 100 h.

2. Discoloration by Iron During the woodworking process, a black stain often appears on the surface of veneer 01' lumber. This stain usually is caused by a chemical reaction between iron ions and wood components. Generally, iron stain is seen i n heartwood rather than sapwood. Such iron stains account for about 70% of the discoloration problems i n the wood industry. Causes of iron stain, methods of stain removal, and staining prcvention methods are discussed in this section.

41 1

Color and Discoloration Color difference AE 1

p -cumaric acid

p-Hydroxyphenylacetic acid

10

5

0 .

,

.

1

,

'

"

'

1

L"-"""

c

--------

FIGURE 19 Effect of polyethylene glycol on photo-induced discoloration of filter papers coated with phenolic substances. Coat weight (g/m') of PEG 1000 to 1 g of model compounds: ----, 0; -, 5.

( I . Occurrence o j Iron St& in Wood Processing a d Prmtical Use ?f Wood Products Examples in Veneer Production. The first step in veneer production is to boil the log. Boiling softensthelog so that it can be easily sliced with a rotary lathe. In this

process, the cross section and crack in the log often become black. This means that iron ions in boiling water penetrate into the wood and react with wood components. Iron ions can be derived from industrial water, the boiling iron vat, steam pipes, and mud containing iron. Veneer is produced in two ways: by a rotary lathe and by a slicer. When a boiled log is cut with a rotary lathe, a stain occurs from the knife. The knife edge of the slicer is very thin and can break off easily when it comes into contact with a hard part of the wood such as a knot. If the broken parts scatter on the surface, black spots are created. On the other hand, if the broken places form sharp corners and press the surface locally. a linear stain will occur. In producing laminated veneer lumber(LVL) from a smalldiameter log, the log is peeled until the diameter of the log is reduced to several centimeters. During this process. the log is supported by a rotating wheel with a side-driver system.Since the wheel contains iron. the marks it makes on the surface of the wheel may become black. If the slicer and pipes are cool, the vapor from the boiled log condenses on the surfaces and can fall onto the veneer surface. If this condensed water contains iron, the veneer surface will become black.

412

Minemura

and

Hon

In transporting the yeneer from the slicer to the dryer, iron stains often occur when the veneer touches the metal fittings of the joints of the carrying belt. The veneer dryer usually is made of iron. Therefore, the wet veneer often is stained when it touches the dryer. If the slicer process is used to produce veneer, the same phenomena are observed as when a rotary lathe is used. The thin veneer made from flitch is about 0.2 mm thick. This veneer is dried by hanging it at room temperature. Black marks often occur when supporting iron hooks are used. ExamplesinPlywoodManufacturing. Plywood is made by pressing with a hot press. The plate of the press is made of iron, so an iron ion is produced when the moisture in the glue layer evaporates during pressing and touches the hot plate. When this iron ion transfers with water onto the surface of the plywood, an iron stain is produced. If the vessel used in mixing gluing components (adhesive, water. filler, and hardener) is made of iron, the surface coated with glue can be discolored. Examples During Sawing. When green lumber is bound with steel belts for transportation, the wood may become black where it touches the belt. Prior to transporting, greentimber is treated withamold-proofingsolution. This water-soluble, antimold agent is usually prepared by diluting the chemicals with water and placing them in an iron vat. In this case, iron dissolves into the solution. When the lumber is immersed into this solution and dried. the surface of the lumber often becomes black. Round poles are made by shaving long green logs withan exclusive shaving machine. Small guide rollers are fitted into the machine to prevent the logs from bending. Because the rollers are made of iron, linear stains often are formed where the wood touches the rollers. This usually occurs at high temperatures and in heartwood. Examples of Other Woodworking Processes. In the manufacture of carved wood, wood is cooked with hot water and fixed with a molding flask. A black stain occurs when the wood contacts bolts fastening the flask. When the lumber is dried, stains occur when the wood makes contact with carrying metal fittings that contain iron or with condensed water from the ceiling in the drying room. Water-soluble adhesives such as polyvinyl acetate or urea resin are used in manufacturing laminated wood. If the instruments for adhesion (the vessel for preparing glue, the spreading apparatus, the pressing tool, etc.) contain iron, the wood surface where the glue is applied often becomes black. When furniture is painted, sealing material is applied to the wood surface prior to painting. Red sealing material often contains iron oxide, so surfacescoatedwith it turn black. Red putty is oftenused to fill cracksorcavities in plywood.When the surface of the plywood is laminatedwithadecorative veneer, the veneer on the filling often becomes black. Occurrenceof Stain inFinishedWoodProducts. Woods used asexterior wall panels or in fences form black stains around nails because iron from the nails dissolves into rain water and reacts with the wood. A similar phenomenon is observed in wooden entry doors equipped with iron fittings. In wood flooring, iron plates are fitted vertically under the flooring and are inlayed into green concrete. The water in concrete dissolves the iron and penetrates into the flooring to form iron stains. In many cases, the stain does not reach the surface because of the thickness of the flooring. Wood plates often are used as chopping boards. Black marks can occur when a wet kitchen knife is left on a board. EvaluationofIron Stain. Iron stains are very similar to stainscaused by black mold. In order to remove the stain, rapid judgment of whether or not the stain has been

Color

413

caused by iron is required. Iron stain usually occurs in heartwood. The stain is flat, not swollen like a mold stain. Small iron pieces may be seen in the center of the stain. When the stain is coated with 5% oxalic acid solution for several minutes, if it fades, the stain has been caused by iron. b. Chemical Factors Influencing the Occurrence of Iron Stain Concentration of Iron Solution. Oak can be stained with 0.0001% of iron [31]. Thirty-three wood species of decorative veneers widely used in furniture or interiors can be stained with 0.00005% of iron (Table 1 1 ) [32]. Sixty percent of the tested wood was stained at one-tenth of this concentration, whereas hinoki and kiri were stained at onefiftieth of this concentration. Tannin Content. Iron ions are widelyused in the qualitative analyses of various phenolic substances because such ions react easily with them to produce coloring substances. In wood there are many phenolic substances, such as lignin and tannin. Gallotannin, catechin tannin [33], and gallic acid [34] are recognized as sources of iron stains. The relationship betweenthetannincontentand iron stain produced in 16woodspecies is summarized in Table 12 [35]. Woods that contain high levels of tannin show a significant decline in lightness. Generally, the degree of the stain is higher in heartwood that contains a large amount of tannin. pH. The pH values of woods vary with species, as shown in Table 12. Woods that have low pH values seem to stain easily. The relationship between the pHof wood and staining with ferric chloride solution has been examined in about 55 species of Japanese woods and 54 species of tropical woods [36]. In Japanese wood, the amount of staining generally decreases as the pH value increases, but in tropical woods this tendency is not clearly recognized. The relationship between the pH of the iron solution and the amount of staining is shown in Fig. 20. Itis clear that the staining is greatest at a pHof 4 and lowest at a pH of 7 [37]. Moisture Content. Nailingdriedlumberdoes not cause a change in color, but nailing greenlumbercancause a black stain around the nail. The lower limit of the moisture content of wood that causes an iron stain is at the fiber saturation point [33]. Table 13 shows the relationship between humidity and the occurrence of iron stains [37]. This experiment used veneers prepared at various humidities. Iron powder was sprayed on the veneersand left undervarioushumidity conditions. At 100% relative humidity (RH), all woods stained, but below 95% RH staining occurred only in Douglas fir. The equilibrium moisture content at 100% RH is the fiber saturation point, which is 22-35% at toom temperature.

TABLE 11 Minima Concentrations of Ferric Chloride Solution That Caused Staining Various Woods

of

Concentration (%) walnut,

Black bead Japanese tree Buna, zebra wood, horse chestnut, kihada, camphor tree, dao, nire. mahogany, manggashinoro Sugi. mizunara, kihada, sawagurumi (Ptrrocctryn rlwifolin), shinanoki. zelkova. akamatsu (Pinus densifom), koa, bubinga, mountain cherry, magnolia. Japanese pear, Douglas fir, palosapis, Lawson cypress, swamp ash, painted maple, Japanese chestnut, sen, Japanese red birch Hinoki, kiri

0.0000s 0.00001 0.00000s

0.000001

41 TABLE 12 Relationship Among Tannin Content, pH, and Decreasing Rate Woods Treated with 0.1% Ferric Chloride Solution

Tannin pH content

(%l

Species Softwood Douglas fir Sugi Lawson cypress Akamatsu Hinoki Hardwood Sawagurumi Mizunara Painted maple Black walnut Buna Kiri Japanese red birch Manggashinoro Magnolia Swamp ash Teak

(heartwood) (heartwood) (sapwood) (heartwood) (sapwood) (heartwood) (heartwood)

0.3 0.3

(heartwood) (heartwood) (sapwood) (heartwood) (heartwood) (heartwood) (heartwood) (heartwood) (heartwood) (heartwood) (sapwood) (heartwood) (heartwood)

2.1 5.6 1.2 0.6 2.0 0.4 0.6 0.3

0.1

0.2 0. l 0.1 0.1

0.2

0.4 0.2 0.2 0.4

of Lightness of

Decreasing rate of lightness (%)

of wood

5 min

4 days

3.75 6.05 5.40 4.35 5.OO 4.55 5.30

50.9 38.4 25.5 28.2 43.5 20.4 23.1

86.4 57.3 53.4 63.9 65.7 69.9 59.2

4.20 4.65 5.10 4.75 4.70 5.60 4.80 4.60 3.95 5.05 5.60 5.40 5.00

66.2 68.0 70.0 58.2 51.6 40.4 42.4

77.0 79.9 76.0 73.7 58.8 77.0 54.7

32.1

50.1

20.8 21.2 14.7 26.0 4.9

47.8 39.7 38.4 34.6 16.8

Test specimens were soaked in 0.1% ferric chloride solution for S rnin or 4 days.

Oxygen. The amount o n staining is less in a nitrogen atmosphere than in air. When a test specimenwastakenout of anitrogenatmosphereand leftin air, the degree of staining became the same as that of the specimen left in air from the beginning 1371. This indicates that oxygen is necessary to accelerate staining. c. Physiccrl Fcrctors I~lfluerrcit~g theOccurrence of Irotl Stnirl Time. Because iron staining is a chemical reaction, it is influenced by temperature and time. The timenecessary for iron staining to occur is generally 3 min at ordinary temperature and 1 min at high temperatures [37]. However, the time is affected by such factors as the method of contact and the wood species. Table 14 shows the time needed for iron staining to occur for two methods of contact in 33 species widely used as decorative wood [32].On contact with iron solution, all the tested woodsshowedstainswithin 45 S, and 40% of the tested woodshoweda stain immediately. Iron powder takes two to three times longer to stain than does iron in solution. Theorderoftimesneededfor staining in the tested wood species is the same regardless of the method of contact. Woods that show strong stains seem to stain quickly. The relationship between contact time with iron and quantity of stain is shown in Figs. 21 and 22. At the initial stage of contact, the quantity of the stain increases in proportion to the contact time, but it does not change after a certain amount of time has passed [37].

415

Color and Discoloration 0.1% Ferricchloridesolution

v

2.0 3.0 4.0

1

5.0 6.0 7.0 8.0 9.0 10.0

0.001% solution chloride Ferrlc

.P .c

0

40

v)

rd

S

0-lo.

2:O

3:O

4:O

5:O

6:O

710 810 910 10.0

PH

FIGURE 20 Relationship between pH of ferric chloride solutions and quantity of stain of some woods immersed in the solution for 2 days.

Temperature. The time for staining to occur after wood comes into contact with iron diminishes as the temperature increases. When the temperature increases, the times needed for staining with iron powder that are used in Table 14 are shown in Table 15 [32]. At temperatures between 85 and 95"C, all tested woods quickly showed stain. When an iron solution was used, the degree of staining increased in proportion to temperature (Fig. 23). The degree of staining showed the same properties when wet veneer was put into contact with an iron plate of various temperatures, as shown in Fig. 24 [37]. Light. Because iron stains also occur in dark places, light probably does not participate in the stain reaction [37]. d. Removal of Iron Stain. There are two ways to remove iron stains: decoloring by means of a chemical reaction and sanding with a planer or sander. Removal by sanding is effective when the stain is limited to the surface and is small in size. However, iron stains in woodworking occur mainly in wood with high moisture content and, therefore, are often large and deep. Such stains must be removed by chemicals. Removal with Chemicals. It is well known that a coating of oxalic acid solution is effective in removing iron stains. Oxalic acid has a higher decoloring ability than sulfuric acid or hydrochloric acid [32]. In decoloring iron stains in mizunara, phosphorous acid, hypophosphorous acid, and phosphoric acid are also effective, as shown in Fig. 25 [38]. The order of decoloring ability of these chemicals is oxalic acid > hypophosphorous acid

TABLE 13 Relationship Between Moisture Content of Wood and Occurrence of Iron Stain -

Relative humidity (76) 95

100

PH Species Sugi Hinoki Douglas fir Mizunara Sawagururni Nire Kiri Japanese red birch

of wood

6.05 5.30 3.75 4.65 4.20 6.70 4.80 4.60

Moisture content Stain

+

+ + + + + + +

90

Moisture content

(%)

Stain

23.4 25.8 26.3 26.9 24.6 25.7 23.2 29.3

-

+

-

-

-

85

Moisture content

(%)

Stain

16.7 18.6 17.1 19.7 16.6 18.8 16.5 20.2

-

+

-

-

(%)

Stain

15.6 16.7 15.2 18.7 14.9 16.9 15.9

-

16.5

After iron powder was scattered on the test specimens, they were conditioned under various humidities for 20 days. f : Iron stain occurred. -: Iron stain did not occur.

-

-

-

80

Moisture content (96) 12.9 13.9 13.8 16.1 12.9 14.7 13.5 14.1

Stain -

-

-

-

Moisture content (96) 11.5 12.1 12.3 11.9 11.4 11.9 11.5 11.3

I 0 3

m S

P

E 3

8

3

s

Color

7

TABLE 14 Time Required for the Iron Stain to Occur in Woods Treated with Time

(S)

Iron

Species

Dripping i n t o l o/c ferric chloride solution Mizunara, sawagurumi, Japanese chestnut, black walnut, Japanese red birch, 0-5 horse chestnut, painted maple, shinanoki. bubinga, dao, mahogany, koa 5-15 Kiri, hinoki, Douglas fir. Lawsoncypress,akamatsu,Japanese pear, kihada, mountain cherry, magnolia, zebra wood, Japanese bead tree 15-25 Manggasinoro, buna, sen, zelkova, palosapis, nire, teak 25-35 Swamp ash, sugi 35-45 Camphor tree Scuttcving of iron ponder 5-15 Mizunara. koa, Japanese chestnut, sawagurumi, painted maple 15-25 Bubinga, buna 25-35 Black walnut,sendan, horse chestnut, dao. shinanoki, mahogany, sen,Japanese red birch, kiri, akamatsu, Douglas fir, mountain cherry 35-45 Swamp ash, sugi, kihada, zelkova, magnolia 45-55 Camphor tree, hinoki 55-65 Zebra wood, nire 65-75 Japanese pear 75-85 Manggashinoro Lawson cypress 95- 105 Palosapis 105-115 115-125 Teak Test specimens for scattering iron powder were

previously Immersed in water for S min.

FIGURE 21 Relationship between immersion time in 0.01 % ferric chloride solution and quantity of stain of some woods. (For key, see Fig. 20.)

°C

418

Hon and Minemura

0 5 20 LO 60 Contact time

120 (min)

FIGURE 22 Relationship between contact time of some wet woods onto iron plates and quantity of stain. (For key, see Fig. 20.)

> phosphorous acid > pyrophosphoric acid = orthophosphoric acid. Chelate chemicals such as a disodium salt of ethylenediaminetetraacetic acid (EDTA-2NA) can be used on light stains. This chemical has a lower decoloring ability than acid, but adding acid to this chemical increases its decoloring ability [32]. When woods decolored with the acid described in Fig. 25 are left under sunlight, only the wood decolored with oxalic acid becomes dark again, as shown in Fig. 26 [38]. Ultraviolet light has a significant influence on this restaining, but visible light also causes black staining, as shown in Fig. 27. Irradiation under nitrogen atmosphere yields only a small amount of restaining, as shown in Fig. 28. It is clear that oxygen has a great influence on restaining [38]. Controlling the Restaining of Woods Treated with Oxalic Acid. A small amount of oxalic acid can decolorize iron stains and given them almost the same color as the sound wood. However, the wood has a tendency to restain. Washing the treated wood with fresh water controls the restaining. Because oxalic acid is an acidic chemical, incomplete washing leaves a red color on the surface involved. After washing, drying is required. When treated wood is washed with water, the oxalic acid occasionally elutes more quickly

TABLE 15 Effect of Temperature on Time Required for the

Iron Stain

to Occur Necessary time

(S)

Species Douglas fir Sawagurumi Buna Japanese red birch

30-35 0-3

25-30 40-45

Zelkova

Painted maple Iron powder was scattered

10-15 15-20

0-3

10-15

on the test specimen.

3-8 0-3 0-3

0-3 0-3

3-8

0-3 0-3

3-8 0-3

419

Color and Discoloration

FIGURE 23 Relationshipbetweentemperature stain of some woods immersedinthesolutionfor

of 0.01% ferricchloridesolutionandquantity 1 h. (For key, see Fig. 20.)

of

than the iron ions and the surface becomes black again. If a stain must be removed from the surface of fabricated furniture or thick lumber, washing cannot be used; therefore, coating with a chemical that has the ability to prevent restaining and red discoloration is needed. Chemicals that have this ability are dihydrogen phosphate, hypophosphite, hydrogen phosphite, and hexametaphosphate [38]. With respect to price, operator safety, degree of discoloration, etc., sodium dihydrogen phosphate (NaH2P0,) is the best chemical to use. This chemical is used in an aqueous

Hon and Minemura

420

20 Phosphorous acid Hypophosphorous acid

8

2 %

Unstained

5 5 P S o d i u m fluoride "@Benzenesulfonic

50 L

acid " 0 1 2 3 4 5 6 7 a

FIGURE 25 Color of iron-stained mizunara decolorized with various chemicals.

solution. Thick coating causes crystallization on the surface after drying; therefore, the coat weight of the chemical must be below 10 &m'. Wood decolorized with oxalic acid together with NaH'PO, displays the same behavior of photo-induced discoloration as does sound wood, as shown in Fig. 29. The chemical solution can be applied with a spreader, sponge, or brush, but a brush with iron wire should not be used. An iron-free vessel made with plastic or glass must be used to dissolve the chemical. Painting Performance of Wood Decolorized with Oxalic Acid and NaH,PO,. Both oxalic acid and NaH'PO, remain on the wood surface after decoloring. Becausedecolorized wood is usually used in furnitureorinteriorwood,thesurfaceis often painted. We might think that the existence of a chemical layer between the wood surface and the paint could cause such problems as weakening of the bonding strength of the paint film or discoloration of the paint. However, painting with a commercial polyurethane indicates no such problems, as shown in Fig. 30 and Table 16 [38]. Mechanism of Occurrence and Prevention of Iron Stain. Water, iron, and phenolic substances are necessary for staining to occur. When iron dissolves in water, iron ions are produced. Since wood is an acidic substance, it accelerates ionization. Iron ions react with the hydroxylgroups in phenolicsubstances in wood and formdeepblack substances (iron stain). This reaction is accelerated by oxygen.

0

1

2

3

Coat weight

4

5 (g/ma)

FIGURE 26 Quantity of photo-induceddiscoloration of iron-stainedmizunaradecolorized with various acids when exposed to a carbon arc light for 100 h. 0 ; Oxalic acid, 0 ; Hypophosphorous acid, X ; Phosphorous acid. 0 ; Pyrophosphoric acid, A; Orthophosphoric acid.

421

Color and Discoloration

""VI

0

h

10

25

50

100 (hr)

Exposure time to carbonarclight

When oxalic acid is applied to an iron stain, it removes iron fromthe stain and makes ferrous oxalate. This product is superior to the bonding strength between the phenolic hydroxyl group and iron ions. Ferrous oxalate is pale yellow, but this color does not show in wood when it is present in very small quantities. It is thought that ferrous oxalate has low photostability and decomposes easily when it absorbs UV light. If a phenolic substance is present, the regenerated iron ion can react with it and form a black substance. However, if phosphoric ions are present, the iron ions might react with them before the phenolic substances and form ferrous phosphate. Ferrous phosphate is very stable under light. These results are summarized as follows:

Iron

PI

water

oxalic acid

iron ion

ferrous oxalate phenolic substance

phenolic substances black complex >phenolic and wood substances (PI)

of iron ion

NaH,PO, ferrous phosphate substances phenolic hv oxalic acid

" +

NaH,PO,is a weak acidic substance that keeps the woodsurface in weak acidic condition after treatment with oxalic acid. It is possible to use phosphoric acid instead of oxalic acid. In such a case, however, the surface acidity is high. A weak acidic substance such as NaH,PO, must be used simultaneously to reduce the surface acidity. For some woodswith light stains, it is possibletoremove iron stains with NaH,PO, alone.The bonding strengths of iron ions and phenolic substances probably differ according to the kinds of phenolic substances involved.

422

Hon and Minernura

Color difference 0

Inthe

air

In N, gas

AE

10

20

'/////////////////////////=I

b

FIGURE 28 Quantity of photo-induced discoloration of iron-stained mizunara decolorized with oxalic acid when exposed to sunlight with an intensity of 400 mW min/cm2 in the UV region.

e. Prevention of IronStain. Because removing iron stains requires a great investment of labor, time, and chemicals, preventing the stain represents a better use of such resources. The main methods of preventing stains are as follows:

1. Preventingcontactwithiron-containingsubstances 2.Capturing iron ions 3. Controlling iron ionization 4. Usingasubstitute

These prevention methods are described here according to the wood processing involved. Prevention During Veneer Manufacturing. The mud that has adhered to the surface of a log or flitch must be removed carefully prior to boiling, and the vat for boiling must be made of stainless steel or concrete. Steam pipes madeof stainless steel or titanium must be used. If pipes made of iron are used, they must be coated sufficiently. If the water used contains iron ions, a chelating agent such as EDTA-2Na or a weak acidic phosphate

20 -

b

-

15-

e" -e

Unstained

Oxalicacld Oxalic acld lSodiumdihydrogenphosphate figure (Numerals In the mean exposure time.) e--.-.

10 L

5

4

loo

l

10

15

a FIGURE 29 Photo-induced discoloration of iron-stained zelkova decolorized with chemicals when exposed to carbon arc light for 100 h.

423

Color and Discoloration

0

Unstained

A

Oxalic acid

0 Oxalic

acid phosphate

0

+

Sodium dhydrogen-

50 75 Exposure time to carbon arc iight 25

100 (hr)

FIGURE 30 Quantity of photo-induceddiscoloration of iron-stainedmizunarathatwas ized with chemicals, painted with polyurethane, and exposed for 100 h.

TABLE 16 BondStrength of PolyurethaneFilms on IronStained Woods That Were Decorated with Chemicals used

rol)

Chemicals Water Oxalic acid Oxalic acid

(kgf/cm2)

+ dihydrogenphosphate

11.5 (36) 6.8 (48) 10.2 (64)

Numerical values in parentheses show the percentage of delamination area between wood and film.

decolor-

Minemura 424

and

Hon

such as NaH,PO, should be added. If EDTA-2Na is used, it should be added in the amount of 0.5 g per gram of iron ion, as shown in Fig. 31 [32]. If water is dark brown in color due to the contamination of iron ions, the addition of aluminum potassium sulfate or a high-molecular-weight coagulant can remove the iron ions. If there is a possibility that the knife can break when a rotary lathe is used, the sliced veneer should be immersed in an EDTA-2Na solution. If condensed water drops from the surface of cold machinesin winter, the machines should be heated before use or the surface should be coated with paint. In the manufacture of laminated veneer lumber with a sidedriving system, gears made of stainless steel should be used. If veneer touches metal fittings fastened to the carrying belt when green veneer is carried to the dryer, the fittings should be coated or covered with vinyl tape. When a dryer is used, the wire netting on which the green veneer is placed should be made with stainless steel and the temperature of the hot wind should be kept above 140°C. High-temperature air brings on rapid vaporization of water from the veneer and does not produce a stain. High temperature also helps control deterioration of the exhaust pipe, because the moisture-containing acid component of the wood does not condense on the pipe. Prevention During Plywood Manufacturing. Because adhesives used in plywood production are water-soluble, it is desirable to useenameledironware or vessels made with stainless steel or plastic in preparing the glue. If a press is used in the gluing process, analuminum plate or duralumin plate withgoodheatconductivityshould be inserted between the plywood and hot plate to prevent staining. Prevention During Working of Lumber. To prevent stains where the surface of green lumber touches steel belts during packing, a piece of wood or cardboard should be inserted between the belt and the lumber. When packing a small amount of lumber, plastic tape should be used. If an iron vessel is used for mold-proofing wood with water-soluble chemicals, a vessel made from a thin plate of polyvinylchloride or stainless steel should be placed inside the iron vessel. A wooden vessel covered with thick polyethylene sheets can be used instead of an iron vessel. During the process of laminating wood, both the vessel used to prepare glue and the instrument for coating should be iron-free

VI Q,

20

x

EDTA.4Na No added

0

Q,

0 C

X

ul fd

5-10 ' ' ' -0.01 0.3 0.5 1.0 Concentration

8

L

1.5 0.01 0.3 0.5 1.0 ("'0) Concentration

1.5 ( %)

FIGURE 31 Relationship between concentration of EDTA solution and quantity of stain. Wood specimens were immersed in 1000 cm' of 0.01% ferric chloride solutionwith 10 cm7 of EDTA solution of various concentrations.

Color and

425

Stain Control in House Construction. When flooring is fixed on unset concrete, metal fittings containing less iron should be used. Nails made with stainless steel or brass should be used to nail wood on fences or outdoor wallboard.If iron nails are used, colored nails are preferred. The nail heads should also be coated. Metal fittings used for wooden entry doors or windows should be made with copper, aluminum, or stainless steel. Stain ControlDuring Furniture Manufacturing. If areddishsealer is required in the sealing process, a sealer that contains no ferric oxide should be used. If a sealer containing ferric oxide is used, EDTA-2Na should be added. 3. Discoloration by Acid Acid stain is caused by acid interacting with wood. Acidic substances are not used very often in woodworking processes. Therefore, acid stains do not occur very frequently. a.Occurrence of AcidStuin in WoodworkingProcesses. Aminoalkyd resin is widely used as an abrasive-resistant and inexpensive paint for wood coating. This paint is mixed with two liquors immediately before use. One of the liquors is a hardener such as paratoluenesulfonic acid. Wood paintedwithexcesshardenerand left in sunlight often turns red. Zelkova is usually used as a thin veneer because it is expensive but has a fine grain. It is stained easily with iron, and the surface of the flitch becomes black. Therefore, the flitch is often dipped into a solution of oxalic acid prior to slicing. When sliced veneer is glued to the base wood, the edge that contacts the solution of oxalic acid often turns red. Urea formaldehyde resin is moderately waterproof and is used widely as an inexpensive wood adhesive.Ammoniumchloride is added to it immediatelybefore use. This chemical reacts slowly with the formaldehyde in the resin and produces hydrochloric acid. On rare occasions, a plywood surface with this resin on it may turn red. b. Factors Affecting the Occurrence of Acid Stain pH. The acids involved in acid staining during woodworking are hydrochloric acid and oxalic acid. The extent of stains that occur when several woods are immersed into acid solutions with various pH values and then left indoors is shown in Fig. 32 [39]. In the pH range from 5 to 2, all wood species show weak stains that are not recognizable to the naked eye. However, at pH below 1.5, strong red or reddish-purple stains occur. Light. Figure 33 shows the change in color of four kinds of wood after immersion into a solution of oxalic acid with a pHof I , whichwere left to dry in a dark place, indoors or under the light of a mercury lamp [39]. Under the mercury lamp, a maximum color change was obtained after 5 min of exposure. With indoor exposure, 5 days were required to obtain the same extent of staining. When stored in a dark place, a slight change of color occurred. However, when the immersed specimens were exposed to indoor light, the colorchanged rapidly to reach the sameextent of staining as the indoor-exposed specimen. From this, it is evident that discoloration caused by acid stain is accelerated by UV light. Oxygen. When wood is immersed in a solution with a pH of 1 and then irradiated with a mercury lamp under nitrogen atmosphere, it shows the same extent of staining as wood irradiated in air [39]. This clearly indicates that oxygen does not participate in the stain’s development. Wood Extractives. Table 17 shows the relationship betweentannincontent and stain [39]. Softwood shows high staining, which might be due to the existence of condensed tannin. Catechol tannin causes acid stains. Because bubinga and koa contain leucoanthocianin, which easily turns red with acid, this substance might be the cause of the stains in both woods. When woods that show a significant stain are treated with acid after

426

Hon and Minemura

o\

15[

0 ;Sugi 0 ;Akamatsu

A ; Manggasinoro

-

0

1.0 2.0 3.0 4.0

5.0

PH FIGURE 32 RelationshipbetweenpH of oxalic acid solution usedforimmersion color difference after l-week exposure under indirect sunlight.

of woods and

hot-water extraction, they do not discolor. This indicates that the source of the acid stain is not lignin but a phenolic extractive. c. Removal of Acid Stain. If the stain is limited to the very top surface, it can be removed by planing or by sanding with sandpaper. For deeper stains, destruction with a bleaching agent or neutralization with alkali is effective to a certain degree. The use of sodium chlorite is desirable because it acts under acid conditions. Other bleaching agents such as hydrogen peroxide or sodium hypochlorite can also be used. Sodium bicarbonate or calcium carbonate can be used for neutralization. d. Prevention of Acid Stain. If paints or adhesives that harden with acid are used, the amount of hardener added should be kept to a minimum. To produce adhesion with

Indirect sunlight (thinline:darkplace)

Mercury lamp lot

.

0

10 20 Exposure time

30 (day)

0

----x-------

5 (hr)

10

15

time Exposure

FIGURE 33 Relationship between exposure time and color difference after soaking in oxalic acid solution at pH 1 . 0 : Sugi; 0 : Akamatsu; A: Manggasinoro; X: Buna.

427

Color and Discoloration TABLE 17 Sensitivity of Wood Species to AcidStain

Stain grade

Species

Akamatsu Strong Buna Bubinga Koa Painted maple Hinoki Sugi Douglas fir Medium Lawson

Weak

cypress Japanese red birch Manggasinoro Magnolia Sawagurumi Black walnut Japanese chestnut Kiri Swamp ash Mizunara Teak

Tannin content (%)

Color difference ( A E )

0.1 0.4 -

15.3 13.3

0.6

8.9 13.8 10.0

0.3

10.6 6.3

0.3

10.9

0.2 0.3 0.2 0.4

7.4

2.1 2.0

4.7 3.1

-

1.9 5.O

0.1

0.6

6.3 5.9

6.3

5.6

3.1 3.1

0.4

1.5

0.2

heat in plywood production, penetration of acid into the veneer should be prevented by raising the viscosity of the glue, diminishing the coating amount of the glue, lowering the moisture content of the veneer, and lowering the pressure and temperature. When oxalic acid is used to remove iron stains, sufficient washingwithwater or the addition of NaH,P04 is required.

4. Discoloration by Alkali Alkali stain is the discoloration caused by the reaction of alkali chemicals with wood. This stain is observedmoreoftenduring the useofwoodproductsthan in woodworking processes. a.Examples of Occurrence of theStain. Freshconcrete is strongly alkali. When wood contacts it in the presence of water, an alkali stain often occurs. Flooring is often bonded on concrete. If water overflows on the floor and reaches the concrete layer, the water in concrete becomes alkaline and penetrates into the flooring to form a brownish alkali stain. When an excess amount of alkaline water penetrates, even the surface of the flooring becomes discolored. A plate made with calcium silicate is used often as a flame-retardant board. On the plate, a decorative veneer with fine grain is often laminated and used in the interior field. The base plate is inorganic and alkali. When glue is coated on the base plate, the alkaline substances in the plate dissolve and react with the veneer to form a brownish stain. b. Factors Affecting the Occurrence of the Stain pH. Figure 34 shows the extentofalkalinestains in fourkinds of woodwhen immersed in solutions with various pH levels of calcium oxide or sodium hydroxide [40].

Hon and Minemura

428 Calciumhydroxid

30

$ 25

0 l

(L,

Sodium hydroxld

t

I

X

0 ,Sawagurumi X 0; Buna

/

"20

C

Q,

Z .U

15

-0b 10 V

5 9.0

1ao

11.0 12.0 13.0

FIGURE 34 Relationship between pH of an alkaline solution and quantity of stain when immersed in the solution for 5 min and then dried in air in a dark place.

Under pH 11.4, little staining occurs, but beyond this pH rapid discoloration is observed. The color differs according to pH. For example, the color of sugi is reddish-brown up to pH 12.5, but beyond this pH it becomes bluish. Light. The effect of light on stains is shown in Fig. 35 [41]. Woods immersed in a solution of pH 12 (in Fig. 34) were left in a dark place, indoors, and under a mercury lamp. The wood left in a dark place retained its original color at immersion, whereas the wood left indoors faded in color and the wood under a mercury lamp showed stronger fading. These results show that light is not required for staining to occur. The opposite is true with acid stains. Oxygen. Whenwoodimmersed in an alkali solution isleftin a nitrogenatmosphere, the extent of discoloration is less than in air. When wood in a nitrogen atmosphere is taken out and left in the air, it discolors to the same extent as wood left in the air from

2ot d" e"--,

Indirectsunlight (thin line : dark place)

Mercurylamp

-0

FIGURE 35 Relationship between exposuretimeand color differenceaftersoaking hydroxide solution at pH 12. 0 : Sawagurumi; 0 : B u m ; A: Sugi: X : Douglas fir.

in calcium

Color

429

the beginning. This means that the production of an alkali stain requires oxygen and the colored substance is formed by oxidative polymerization [40]. Wood Extractives. Table 18 shows the relationship betweentannincontentand amount of staining. In the 16 wood species tested, woods with high tannin content had more of a tendency to stain [41]. It has been found that if sufficient extraction with hot water is completedbefore the wood is immersed in analkalinesolution, the stain is scarcely recognized. Only a small yellowish-ocher stain occurs above pH 13. From these results, it can be surmised that the alkali stain is due mostly to the water-soluble phenolic components. Lignin also participates in discoloration under alkali conditions. c. Removal ofAlkali Stain. Most alkali stains occurring in a short period of time can be removed with bleaching agents. The stain on the surface of concrete blocks with alkalinecementdescribedearliercanbe easily removedwithhypochloritesolution, as shown in Table 19 [42]. Concreteblocks are highly alkali-resistant, but notvery acidresistant. Because the surface is rough and has a lot of voids, it is very difficult if not impossible to dissolve the colored substances of the stain with an alkaline solution. Hypochlorite is an alkaline bleaching agent;its solution can be used to decompose the colored substances without damaging the block. Stains on the top layers of wood surfaces can be removed by planing or sanding. Stains at a certain depth that cannotbecompletelyremoved by these methods can be bleached with a bleaching agent or coating with dilute acid. d. Prevention qf AlkaliStain. Plywood that does not discolorwith the alkali of cement or plywood coated with alkali-resistant paint can be used to frame concrete for hardening. When a decorative veneer is laminated on the alkaline inorganic plates, the use of an adhesive film is desirable. When a water-soluble adhesive is used, the following methods should be considered: increasing glue viscosity, diminishing water, lowering the moisture content of veneer, decreasing pressing time and temperature, coating the plate with an alkali-sealing paint, and so on.

5. Discoloration by Microorganisms Approximately half of wood components are carbohydrates, and in green wood they contain a moderate volume of water. When green wood is left under certain conditions, microorganisms propagate on the wood. This often is accompanied by the discoloration or lowering of wood strength. The microorganisms that cause discoloration are bacteria, mold, and basidiomycetes. Mold discolors the surface of wood but does not diminish its strength. Basidiomycetes cause a decline in strength. Among the basidiomycetes are brown-rot fungi and white-rot fungi. The former mainly decompose cellulose and hemicellulose, whereas the latter also decompose lignin. Bacteria may occur in stored wood when it is immersed in or sprayed with water. Discoloration is due to the pigments of the microorganism or coloring compounds produced by the reaction of the woodcomponentswith the secretions of the microorganism. a. ExcImples of Occurrence ?f Stains in Woodworking Processes or Wood Products. When fresh green lumber is stacked on a warm day under high humidity, many colonies of mold with various colors can grow on the surface of the wood overnight. When sliced veneer is transported without drying, discoloration caused by fungi and bacteria can occur on the surface. When logs are piled outdoors for a long period of time, brownish discol-

Minemura 430

Hon

TABLE 18 Sensitivity ofWoodSpeciestoAlkalineStain

Color Stain grade

gurumi

ash

and

Strong

Medium

Tannin content (%)

(

m

Mizunara Black walnut Buna Douglas fir Sugi Mizunara Painted maple

(heartwood) (heartwood) (heartwood) (heartwood) (heartwood) (heartwood) (sapwood) (heartwood)

2.1 5.6 2.0 0.4 0.3 0.3 1.2 0.6

25.6 11.9 9.5 20.9 15.2 15.3 18.2 16.0

Kiri Lawson cypress Lawson cypress Japanese red birch Manggasinoro Akamatsu Teak

(heartwood) (sapwood) (heartwood) (heartwood) (heartwood) (heartwood) (heartwood)

0.6 0.1 0.2 0.3 0.2 0.1 0.4

9.3 12.7 4.1 7.7 5.8 3.3 3.6

Magnolia Magnolia Hinoki Sugi

(heartwood) (sapwood) (heartwood) (heartwood) (sapwood)

0.2 0.2 0.4 0.1 0.1

2.9 1.5 1.6 8.2 3.4

Swamp Weak

oration may occur in the cross sections of both sapwood and heartwood. This stain is caused by basidiomycetes. A blue stain is observed only in sapwood and does not bring about a decline in strength. Ilomba wood often discolors after felling to give a reddish-brown color. This wood is normallyallsapwoodandcontains substrates which are suitablefor the growthof bacterial [68].Some bacterials propagated on the lumber produce ammonia as a metabolite and form colored substances by the reaction of the components with ammonia [43]. For wood components which are responsible for the discoloration, (+)-catechin and (-)-epi-

TABLE 19 Decoloring of Alkaline-Induced Color Substances of Wood on Concrete Block

Effect of

oring chemical Coated sol.(CIO)? 2% Ca 5% NaClO sol. 15% H 2 0 2sol. (pH 10) 10% NaOH sol. 0:

Excellent.

A: Common. X:

Poor.

0

0

A X

Color

431

catechin are confirmed [69]. In the brown-stained region of hemlock, dark-pigmented fungi are predominant.Thesefungiinducebrown discoloration in the sapwood.Browning is accompanied by an increase in pH from 5 to 7, a decrease in total soluble phenols, and oxidation of phenols such as catechin. Intensive discoloration occurs at pH 7, and oxygen is indispensable for the development of the discoloration [70-721. Brownish discoloration in beech wood is caused by bacterials which produce ammonia to give a pH 7.3 [73]. Yellow discoloration of oak heartwood is caused by mold fungus. It is assumed that metabolic compounds of the fungus react with hydrolyzable tannins and give yellow substances [74]. From the blue-stain fungi, the dark coloring pigments have been isolated. They are classed with the group of melanins and are associated with carbohydrates and proteinaceous components [75]. Concerning pink stains of angiosperm and gymnosperm woods caused by fungi, a red pigment has been isolated and identified as 5,8-dihydroxy-2,7-dimethoxy1,4-naphthalenedione [76]. Some injurious insects penetrate tropical woods to the inside. Such insects often carrymicroorganisms.Forexample, in aplacewhere Limnoria lives, there is awhite corpse of Ambrosia beetle. When the insect moves in the tangential direction in wood, the discoloration is dappled in the radial section and striped in the tangential section. Larvae of the sugi bark borer feed on the wood of living sugi trees and induce discoloration [77]. In the discoloredsapwoodand in the reaction zone of the soundsapwood-discolored sapwood boundary, potassium and magnesium begin to accumulate within one year, and calcium within two years. Discolored sapwood has a greater cation-exchange capacity. b. FactorsAffectingtheOccurrence cf Stains. Thefollowinggrowth factors are essential for the propagation of microorganisms: water, air (oxygen), moderate warmth, and nutrients. Wood itself is a nutrient. Generalgrowingconditionsare3-40°C, 90% relative humidity, and 20- 150% wood moisture content. c.Removal of Stains. Stainscaused by moldcan beremoved by planingor by coating with a bleaching agent. Because stains causedby basidiomycetes often occur deep in wood, they cannot be removed completely. In such cases, it is effective to immerse the wood in a bleaching agent. For example, the brownish stain on a shina log that is caused by invasion of basidiomycetes can be removed by immersing the log in a dilute solution of sodium hypochlorite for several hours [44]. For removing fungal stain of ponderosa pine sapwood, 2% hydrogen peroxide solution with sodium hydroxide and sodium silicate as a buffer give a good result [78]. Comparedwithchemical fungicides, biological control is generallybenign to the environment.Concerning the biological control of sapstain fungi,metabolitesobtained from two fungi were examined on stained pine veneer disks and it was found that they remove sapstain and kill existing fungal growth [79]. d. PreventingStains Addition of GrowthInhibitors. To preventstains, it is effective to adhere preservatives or antimold agents to wood. This can be done by coating, spraying, immersing, and pressure impregnation. The chemical should be selected on the basis of low toxicity and slight color. Organic compounds containing tin or iodine are soluble in organic solvents, and solutions of these materials have good permeability to wood. For prevention of mold growth during drying, pretreatment of green lumber with propionic acid is recommended [74]. As a preventing chemical for brown stain i n hemi fir, a quaternary ammonium compound, didecyldimethylammonium chloride, is effective [SO]. The preservative

Minemura 432

and

Hon

treatmentmustbedoneassoonaspossible after sawing. Prior to outdoor storing, the treated wood should be kept away from sunlight and rain for at least one day. Controlling Growth. Stain preventioncanalsobeachieved by minimizing the growth factors described earlier. To reduce moisture, green fresh wood should be piled in a location with sufficient ventilation and transferred quickly to the seasoning process. To reduce the amount of air, logs should be stored in water or sprayed with water to cut off oxygen. Logs can be stored in snow and covered with sawdust or plastic foam to reduce the temperature. Sliced decorative veneer should be kept at a low temperature or dried with high frequency. The drying must be completed with bundling in order to prevent splitting. Lumber should be dried as soon as possible after sawing. Obstructing the Penetration of Microorganisms. Covering the cross section of the log with preservatives will prevent the penetration of microorganisms [45]. When wood is coated with preservatives, recoating with elastic paint such as polyurethane is even more effective. It is also important to keep the working place clean. Decayed wood and wood wastes must always be removed. It also is desirable to irradiate with UV light at night and to spray fungicides regularly.

6. Discoloration by Enzymes Variousenzymes in wood participate in manymetabolismsystems.Someenzymes are active even after logging. In sawing or veneering, when fresh green wood makes contact with oxygen, these enzymes often take part in discoloring the surface of the wood. U . Exurnples of Stains Occurring in Woodworking Processes. Alder generates reddish-orange discoloration immediately after felling. This discoloration is caused by interaction of catechol oxidase and hirsutoside, which is a xyloside of diarylheptanoid containing two catecholic nuclei [81]. Shina is widely used for plywood production. When veneer sliced with a rotary slicer is left without drying for several hours, the surface often becomes yellow. When sliced walnut veneer is allowed to stand in the same manner, it becomes black. The fresh green lumber of todomatsu become yellow. Kiri wood is widely used for furniture in Japan. This wood changes color to dark brown when sawn immediately after felling. This discoloration might be caused by catalytic oxidation with oxidase. Caffeic acid sugar esters have beenisolated as the compounds responsible for discoloration [82,83]. Peroxidase consists of heat-labile and heat-resistant enzymes. The activity of the latter enzyme occupies about 12% of the total [84]. Concerning brown stain in sapwood of Douglas fir, its enzymatic extract showed two pH optima for activity (pH 5.5 and 8.0) and highest activity at 35°C. It showed also highest activity for (-)-epicatechin and dihydroquercetin [85]. When beech wood chip is stored outdoors, it changes color to deep brown after a few days. As beechwoodcontains sufficient amounts of activeperoxidaseandmalate dehydrogenase,phenoxy radicals are formed in the lignin andsubsequently new chormophores are formed [86]. b. FuctorsAffectingDiscolorution by Enzymes. Moistureandhumidity significantly affect discoloration. For discoloration to occur, the surrounding humidity must be about 100%. Temperaturealsoinfluences discoloration, with discoloration occurring slowly below 20°C. Phenolic substances in wood might oxidate to colored substances by means of enzymes in the wood and oxygen in the air. c. Stuin Removul. Yellow stains can often be removed by bleaching with hydrogen peroxideor extraction with hot water. Because the sapwood of todomatsu is usedfor chopsticks, extraction with hot water is recommended.

hemical

Color

433

d. StainPrevetztiot?. In order to prevent the stain, it is necessary to create an environment in which the enzyme does not act. An enzyme is a protein and undergoes an irreversible change when it is heated or comes into contact with some chemical substances. To remove the yellow discoloration of shina and the sapwood of todomatsu, immersion of the wood into boiling water for half a minute is effective [46]. Radiation with microwaves in an oven also works. These treatments are recommended for woods used with foods. When treating with heat, a rapid rise of temperature is required. When a lotof wood is immersed in a small quantity of hot water, the temperature falls rapidly and can reach a temperature suitable for enzyme action. For prevention of brown stain of Douglas fir sapwood, steaming it to 212°F is recommended [87]. Coating with various chemicals is also effective. As shown in Table 20, coating with dilute acids, sulfites, and EDTA-2Na is valid [46]. Sulfites may act as reducing substances. EDTA-2Na may react with metal ions that are an essential part of a co-enzyme or react with the phenolic substances in wood. Optimum enzyme action occurs at weakly acidic pH. Therefore, the pH in the surface of the woodcan be loweredwithoutan acid stain occurring,or the pH canbe raised without an alkali stain occurring. Coating or immersing with a solution of dilute acid or carbonate is effective in changing the pH value. For white pine, immersion in a solution of sodium carbonate or sodium borate with a pH value of 10 is effective [47].

TABLE 20 Effect of Chemicals o n the Control of Orange Stain of Shim Coat weight (g/m') Coated Hydrochloric acid Sulfuric acid Phosphoric acid Hypophosphorous acid Nitric acid Boric acid Formic acid Acetic acid Oxalic acid Ascorbic acid Benzensulfonic acid Semicarbazide hydrochlorlde Sodium bisulfite Sodium sulfite Sodium hypophosphite Sodium nitrite Fornlaldehyde Urea Thiourea EDTA. disodium salt

0.1

I

5

A

e

e

0

0

A A A

0

X

X

X

X

e A A

X

X

X

A

0

e

X

A

0

X

0

X 0

A

A

0

0

A

A

0

X

X

X

X

X

X

X

X

0

X

X

X

X

X

0

A

0

0

a

e 0

Hon and Minemura

434

Forpreventionofdiscoloration, the removal of the causativecompounds is also effective. To prevent discoloration of kiri wood, sufficient natural seasoning of sawn timber has been conducted traditionally. During this seasoning, the compounds responsible for color change dissolve into rain water and are removed. As a more effective and rapid preventive method, impregnation of timbers in cold water and subsequent treatment with warm moisture are also recommended [88]. Immersion into urea solution is also suggested [511. Quick drying under good ventilation or storage at low temperatures is also valid. Discoloration by enzymes sometimes is encouraged in the wood industry. For example, in the manufacture of decorative walnut veneer, sliced green veneer is left until a blackbrown color develops, and the veneer is then heated with a roller press to stop the enzyme action [48].

7. Discoloration by Nonmicrobial Oxidation with or Without Heating It is well known that woods discolor when they are subjected to high temperatures. Even without being subjected to heat, however, some woods still change their color readily by oxidative reactions that arenotaccompanied by microbiologicalorenzymatic actions. These discolorations are described in this section. U . Churucteristics of Discolorution. When fresh sawedlumber is dried at ahigh temperature, the wood changes color. The color differs according to the wood species and drying temperature. It may be yellow, brown, red, gray, etc. Wood left at high temperatures for long periods of time usually becomes brown. Discolorationduringdryingincreasesas the temperatureandhumidityincrease. Hardwood generally discolors at a lower temperature than softwood. Heat discoloration of several woods is as follows [33,49]: Red color Maple (above 50°C and 65% RH) Beech (above 50°C and 65% RH) Brown color Oak (above 80°C and 65% RH) Sugar pine (above 65°C and 65% RH) Walnut alder (steaming) Spruce fir (above 90°C) The discoloration in artificial seasoning of todomatsu is shown i n Table 2 1 . Discoloration increases with the rise of temperature and the prolongation of time. The formation

TABLE 21

Color of Todomatsu Dried at High Temperature Drying condition

ying Wet-bulb Dry-bulb temperature temperature ("C) 100-1

IO

100-1 10 100

Color

("C) 100

100-86 100-80

ti me (h)

48.5 24.5 48.0

L

57. I

(1

63.7

11.9 9.0

62. I

6.9

h

26.6 33.4 27.8

Color

435

of colored substances from a phenolic compound oxidized with air and the formation of dark materials from hydrolysis of hemicellulose have been considered the causes of discoloration. When the material causing discoloration is water-soluble, these materials rise to the surface and accumulate there, discoloring the wood [49]. It is assumed that many substances that change color with heat also discolor with enzyme action. Brown discoloration oftenoccurs in Europeanoakwoodduring kiln drying.This discoloration occurs at a wood moisture content between 30% and 60%, kiln temperature above 25”C, and relative humidity of about 70% [89,90]. In brown stain of oak, colored polyphenolic polymers and complex esters, hexahydroxydiphenolesters, are found in larger amounts than in nonstained wood [91]. This discoloration is presumably due to oxidative coupling of compounds related to gallic acid. Discoloration during kiln drying may be the result of hydrolysis and oxidative transformation of ellagitannins 1921. Concerning sticker stain in sugar maple, chemical analysis of stained sapwood has been conducted [93]. The amount of acetone-water-soluble materials in the stained part was less than in the clear part. This suggests that phenolic extractives accumulated under the sticker stain during drying and then oxidized to insoluble polyphenolic compounds. Scopoletin was isolated from both the stained and unstained parts as the phenolic compound produced during drying. As anatomical characteristics of brown stain after kiln drying in hemlock, the stain exists in sapwood, particularly in the earlywood, and is recognized mainly in longitudinal tracheids [94]. From a study of the correlation between loss of brightness in mechanical pulp and storage time of western hemlock chip, it is suggested that d-catechin polymerizes oxidatively to give an insoluble polymer and to cause the brightness loss [95]. The heartwoodcolor of sugi is classified into three types:normalreddish-brown color type, black color type, and color-changeable type. The latter occurs when the wood is left at room temperature; the surface color changes from reddish brown to black in 30 min after sawing. This phenomenon is observed in wood grown on tree farms. This type of stain cannot be controlled by oxalic acid; therefore, it is not an iron stain. The characteristics of this phenomenon are as follows [96,97]: The stain occurs either in the dark or under light. Atmospheric oxygen is necessary. In nitrogen, discoloration does not occur. When the color-changeable sugi is extracted with water preliminarily, it gives no more black color. The water extract contains a potassium hydrogen carbonate(0.4% w/w), which keeps the wood weakly alkaline. When normal reddish-brown sugi wood is immersed into awater solution of KHC03, it changescolor to black. So, this inorganicchemical is recognized as one of the causative materials of the black discoloration. The characteristics of black sugi which is alreadyblack when standing in the forest, before logging, have been examined. The black sugi has high moisture content and alkalinity, and contains K’, Na’ , and HCO, [%l. The black substances are presumably a polymer of water-soluble norlignans such as segurin-C and plicatinaphthol [99]. Heartwood of murasakitagayasan changes rapidly i n color from brownish yellow to dark purple after sawing. For this discoloration, oxygen is indispensable. Light and water acceleratc the reaction [ 1001. As the substancewhich contributes to the discoloration, 7,3’,4’-triacctoxy-6’-n~cthoxyisoHav-3-enehas been isolated [ 1 0 1 1. This discoloration might be caused by the autooxidation of this compound to give the quinonoid structure. h. Rrr~~o\w/ of ~ i . s c ~ o / o r t r r i o rWhether ~. sawed lumber has been artitically seasoned or not. its surface is rough. When the lumber is used as an interior wood o r for furniture, its surfacc is usually planed. B ~ C ~ Lheat I S Cdiscoloration is mostly limited to the surface, a sound surface should reappcar after planing to ;l thickness of 2 mm.

436

Hon and Minemura

The yellow discoloration of todomatsu can be removed by immersing the stain in boiling water. Heat discoloration of a relatively light color can be removed by oxidative decomposition with a bleaching agent. c. Preventing Discoloration. A drying process generally consists of a natural seasoningandsubsequently an artificial seasoning. It is desirable to conductan artificial seasoning after enough natural seasoning in order to control discoloration. When sawed green lumber of todomatsu must be dried immediately after cutting, the drying must be done below 80% RH and at 50°C until the wood reaches the fiber saturation point. On drying, it is important to insert enough sticker to prevent close contact and dampness. Drying must be done as soon as possible afterlogging. Low-molecular-weight sugars or amino acids increase as time passes. These substances can cause discoloration [49]. As a method of preventing brown discoloration of oak, drying under a vacuum and superheated steam are effective [ 1021. Coating the wood with antioxidant, reducing agent, acid, and so on, is also an effective method to avoid discoloration. A solution witha concentration of 5-6% ammonia, ammonium carbamate, and zinc oxide is effective in preventing the brownish discoloration of white pine [50]. Sawn timber from water-stored oak logs develops gray stain in the sapwood sooner than does freshly cut logs. For prevention of this stain, a 5-min dip in a 5% sodium bisulfite solution is recommended for sawntimberfrom freshly cut logs, and 10% sodium bisulfite solution forsawntimber from water stored logs [ 1031. For logs stored in water for more than 3-4 weeks, however, this chemical does not give complete prevention. For prevention of nonmicrobial discoloration, methyl bromide fumigation has been tested [104]. This is effective for red alder, but not for western hemlock. This treatment causes rapid death and modification of living parenchyma cells. The effectiveness of coating shiurizakura with semicarbazide for heat discoloration is shown in Fig. 36 [5]. To prevent blackish discoloration of sugi, coating with acidic and chelating chemicals such as phosphoric acid, oxalic acid, nitric acid, formic acid, EDTA-disodium salt, etc., is effective [ 5 5 ] .

8. Discoloration with Exudation of Resin When resin in wood exudes to the surface, the color of the surface changes. This phenomenon is often observedondecorative thin veneer, on lumber that hasbeendried inadequately, and even on furniture or paneling.

W

Q

437

Color and Discoloration TABLE 22 Color of Hinoki Determined Before and After

Removal of Resin L

b

n

21.4 10.4 59.6 Before removal 8.5 62.2 After removal with methanol

19.4

a. Characteristics of Discoloration. Certain woods have resin canals in the direction of the stemand radiation. Such canals open onto the surface of the lumber after sawing.Lightpale resin exudes on the surface in softwood. This resin is a mixture of terpenoids with various boiling points. Exuded resin becomes hard with volatilization of substances with a lowboiling point and then exhibitsa wet color. There is a brown resinous material in the tracheid of mizunara and ash. b. Removal of Discoloration. Discolorationcanberemovedthroughphysical or chemical methods. The resin of hinoki dissolves well in alcoholic solvents such as methanol and ethanol. When only a small amount of resin is exuded, wiping the surface with a cloth impregnated with the solvent is recommended. Resin from thick lumber can be removed in the same way.If a lot of resin has exuded from thin veneer, immersing the veneer in a solvent is an effective method. The colors of hinoki before and after elution with methanol are shown in Table 22. The increase of lightness and loweringof saturation, as well as the disappearance of the wet color, are due to the removal of resin. Besides alcohol, methyl ethyl ketone and acetone also can be used. The resin of karamatsu dissolves well in hexane and trichloroethylene. Karamatsu is used for furniture or as a decorative material in interiors and often exudes resin during use. This phenomenon is mainly observed when the wood is used at high temperatures. To remove the resin, scrape off asmuch resin as possible, thenwipe it withacloth impregnated with a suitable solvent. After removal, coating with polyurethaneis desirable. This film is somewhat effective in controlling exudation of the resin. To remove the resinous products packed in the tracheid of mizunara or ash, it is effective to immerse the wood in a 1 % solution of polyethylene oxide or nonionic surfactant and then keep it at 80°C for half an hour [52].Table 23 presents data on the color before and after resin is removed. Lightness increases considerably after removal. c. Preventing Discoloration. Because the resin of softwood is light-colored, the use of an artificial seasoning condition that vaporizes substances with low boiling points is desirable. The following example is a typical procedure. At the beginning of seasoning, lumber is dried at 100% RH and 90°C. After that, the temperature is kept at 50°C and then raised by steps until 80°C is reached.

TABLE 23 Color of Mizunara Determined Before and After Removal of Resin L

8.6 Before removal After removal with 18.3 1% P E 0 solution 6.5

45.0 59.5

n

b

16.1

Hon and Minemura

438

9. Discoloration in a Standing Tree Stains in standing trees include spotted stains caused by deposit of inorganic or organic substances in tracheids, infestation of insects, or imperfect pruning. a. Troubles with Speck. Whenaninorganicmaterialabsorbed by the roots or an organicmaterialsynthesized in wood is depositedaswhite or yellow-brownish fillers, these materials appear on the surface of the sawn lumber or veneer as a speck. Silica. Silica is often contained in tropical wood and is recognized as white grain in tracheids, rays, and axial parenchyma, and so on. It hastens the abrasion of a saw blade in sawing. The solution of hydrogen fluoride dissolves silica, but cannot be used as a removal agent because of its poisonous properties and tendency to discolor wood. Unfortunately, there is no effective way to remove silica. Calcium Oxalate and Calcium Carbonate. Calcium compounds appear as white crystalline substances in tracheids. Calcium carbonate reacts with dilute hydrochloric acid to form carbon dioxide and water-soluble calcium chloride. Calcium oxalate dissolves in hydrochloricacid.Bothcompounds, therefore, canberemoved by coatingwithdilute hydrochloric acid. Washing or wiping with water must be done sufficiently after removal. When the surface is below pH 2 after treatment with water, it should be coated with a dilute solution of sodium carbonate until the surface becomes weakly acidic. Isoflavone. Eurasian teakwood is used widely in furniture. White spots or lines are often found in the tracheid of this wood. These are a mixture of isoflavones which consists of afrormosin and biochanin-A. As the melting points of these substancesare below 200"C, they can be removed when the wood is left in a hot press heated to 200°C [53]. Fisetin. There are yellow substances in the tracheid of melbau.Thesesubstances form spots or lines on the surface of lumber after sawing. This colored material contains fisetin, robinetin, and quercetin. As the melting point of these substances is above 300"C, removal during the heating process is impossible. These substances react with boric acid to form a water-soluble chelate compound, so when the wood is immersed in 2% boric acid solution for several hours, the stain can be removed [54]. The Rest. In planted teak, a lot of small bluish blackspots are frequently seen. These spots appear particularly in the heartwood like an annual ring at a frontier near to the sapwood, so they might be a dehydrotechtol [ 1051. For removal, the following methods are suggested [ S ] .The bluish-black color disappears immediately upon contact with organic solvents. However, it reappears again with larger area after evaporation of the solvent. Finishing with transparent paint with thinner is effective for disappearance of bluish black spots and maintenance of the characteristic color of teak. As shown in Table 24, sufficient immersion into organic solvents brings an increase of lightness and decrease of reddish component to lose the peculiar color of teak. When heated at high temperature, the bluish-black spots melt to give pale brown color and disappear.

TABLE 24 Color of Teak with Bluish-Black Spots Before and After Sufficient Immersion i n Solvent

Before extraction After extraction with ethanol 3.3 After extraction with acetone Normal unstained teak

33.3 63.6 64.2 55.2

-

0.4

2.8 9.3

7.0 18.9 19.8 22.2

Color

439

10. Stains Caused by Adhesives in Woodworking Processes There are many uses of adhesives in woodworking processes. Adhesives can be colored or can react with other substance and then discolor. These substances often stain wood. In this section, some examples are given and their control discussed. a. Stair1 Caused by Oozing of Adhesives. Adhesives containing phenolic substances (phenol-formaldehyde resin, resorcinol resin, tannin resin, etc.) have a dark red or reddishbrown color. When they are used asadhesives in plywood, they exude to the surface through the tracheid of athin-surface veneer. When the color of the surface veneer is white or light, the color of the adhesive layer reaches the surface and darkens the color of the surface veneer even if there is no exudation. To prevent such oozing, the following methods are effective: raising the glue viscosity, decreasing the moisture content of the veneer, reducing the pressing timeortemperature, and increasing the thickness of the veneer. In order to prevent reflection of the color of an adhesive, it is possible to mix a white pigment such as titanium oxideintotheadhesives without influencing bonding strength. b. Stains Caused b y Reactions with Adhesive Components. Vinyl urethane adhesive is composed of phenyl isocyanate and vinyl polymers. These two substances are mixed immediately before use. Thisisocyanate often makes a colored substance with tannin. When it is used for the adhesion of mizunara, the adhesive layer becomes gray. To prevent this, an aliphatic isocyanate should be used. c. Stciins Ccrusecl b y Whterproof Asphalt. The back side of flooring that contacts concrete directly should be coated with asphalt to prevent permeation of the water of the concrete. When much water penetratesinto the concrete (e.g., in a flood), the pressure caused by water evaporation acts on the asphalt coating layer of the flooring. Flooring of mizunara has a tracheid of large diameter, and this pressure can cause the asphalt to reach the surface through the tracheid. As a result, a stain composed of black specks may appear on the surface. To prevent this stain, an alkali- and waterproof sheet should be bonded on the back of the flooring rather than using asphalt coating, or the sheets may be placed on the concrete before setting the flooring. d. Stcrins During A.s.senzhIy with ( I Dowel. In the manufacture of furniture. assembly with a dowel is often practiced. The dowcl is coated with vinyl acetate adhesive and put into the opening. In this case, excess adhesive will squeeze out from the opening. When the excess adhesive is wiped off with a wet cloth, the wiped mark is often noticeable. This is caused by the tine fibers on the wood surface. This fiber is forced down at planing, but it stands up when water is absorbed.This stain can be removed by grinding with abrasive paper.

REFERENCES

Minemura 440

and

Hon

R. S. Williams, J. Appl. Polymer Sci., 28:2093(1983). D. N.-S. Hon. S.-T. Chang, and W. C. Feist, J. Appl. Polymer Sei., 30: 1429 (1985). 12. K. Kringstad, Tappi, 52:l070 ( 1969). 13. J. Gierer and S. Y. Lin, Svensk Pqerstidn., 75:233 (1972). 14. S. Y. Lin and K. P. Kringstad, Eppi, 53:l675 ( 1970). 15. K. Umehara and N.Minemura, J. Hokkaido Forest Res. I n s f . , 300:13 (1977). 16. W. C . Feist, USDA ForestServ. Res. Paper FPL 339, 1979. 17. N. Minemura, J. Hokkaido Forest Res. Inst., 311:18 (1977). 18. S. Imura and N.Minemura, J. Hokkaido Forest Res. Inst., 305:l (1977). 19. Y. Kai and M. Kawamura, Mokuzai Gakkaishi, 31:766 (1985). 20. V. Loras, Pulp Paper Mug. Can., T49 ( 1 968). 21. M. Takahashi, Abstracts of Papers Presented at the 26th Annu. Meeting Japan Wood Research Society, Shizuoka, pp. 3 18-3 19 ( 1 976). 22. N. Minemura and K. Umehara, J . Hokkaiclo Forest Res. Inst., 3/5:1 (1978). 23. N.Minemura, Japan Finishing, /7( 12):179 (1978). 24. N. Minemura, Mokuzoi Gakkuishi, 24587 (1978). 25. N. Minemuraand K. Umehara,Paper presentedatthe ACS/CSJChemicalCong.,No. 94, Honolulu.Hawaii(April 1979). 26. B. Ranby andJ. F. Rabek, Photode~rcldrltiotl.Photo-Oxidution und Photo.st~~bilizatio~l of’ Polynzers, Wiley, New York, p. 210 ( 1975). 27. N. Minemura, K. Umehara, and M. Sato, J. Hokkaido Forest Res. hsr., 380:11 (1983). 28. G. Gellerstedtand E.-L. Petterson, Svensk Puperstidn., 80:lS (1977). 29. E. A.McGinnes. Jr., Wood Sei., 7:270(1975). 30. T. Yoshimoto and M. Samejima, Mokuzui Gakkaishi, 23:601 (1977). 3 I . W. Sandermann and M. Luthgens, Holz Roh- Werkst., 11:435 (1953). 32. K. Takenami, Mokulai Kogyo (Wood I n d . ) , 24:263(1969). 33. F. Kollmann, R. Keylwerth, and H. Kubler. Holz Roh-Werkst., 9382 (1951). 34. T. Kondo, H. [to, and M. Suda, Nippon Nogeikagaku Kaishi, 30:28 1 ( 1956). 35. K. Takenami, Mokrczai Grrkkaishi, 10:22(1964). 36. T. Goto and H. Onishi, Bull. Shitnane Agric. Uni\j., 15(A-2):80(1967). 37. K. Takenami, Mokuzai K o g ~ o(Wood Ind.), 24:210 ( 1969). 38. N. Minemuraand K. Umehara,Abstracts of Papers Presentedatthe 12th Annu.Meeting Hokkaido Branch Japan Wood Research Society, Asahikawa, Hokkaido, pp. 59-62 (1980). 39. K . Takenami, Mokuxri Gukkaishi, 11:41 (1965). 40. K. Takenami, Mokuzai Kogvo (Wood Ind.), 2 4 3 14 ( l 969). 41. K. Takenami, Mokrczni Gukknishi, 11:47 (1965). 42. N. Minemura, Abstracts of PapersPresentedatthe 13th Annu. Meeting on Chemical Treatment of Wood by the Japan Wood Research Society, Tsukuba, Ibaragi, pp. 1-8 ( 1983). 43. J. Bauch, 0. Schmidt, Y. Yazaki, and M. Starck, Holqjorsch.. 39:249 (1985). 44. H . Kawakami, Annu. Rep. Hokkrrido Forest Prod. Res. I n s t . , 1980-198/, p. 1 I . 45. A. Nunomura. Abstracts of Papers Presented at the 16th Annu. Meeting on Studies of Forest Technology, Sapporo, Japan. pp. 325-326 (1976). 46. N. Minemura, Mokuzai Kogya (Wood Ind.), 38:363 (1983). 47. H. A. Hulme, Forest Prod. J., 25:38 (1975). 48. T. Yoshimoto. Ki no Hnnushi (Storcs 0 1 1 Wood). Outsuki. Tokyo, p. 60 (1983). 49. M. A.Millett, Forest Prod. J.. 2:232 (1952). 50. J. K. Shields, R. L. Desai, and M. R. Clarke, Forest Prod. J., 23:28 (1973). 5 1. K. Makino. Y. Kobayashi. T. Matsuura, and T. Ousako. Rep. Ind. Arts Inst. Hiroskirnu Pm-

IO. I 1.

,fecttcre, 9:24

(1980).

52. K.Umehara. Annu. Rep. Hokknido Forest Prod. RES. I n s t . , 1978-1979. p. 1 I . 53. H. Imamura, Y. Tanno, and T. Takahashi, M o k u x i Gukktrishi, 14:295 (1968). 54. H. Imamura, H. Fushiki, S. Ishihara,and H. Ohashi, Res. Bull. Fuc. Agric. G@ utli\?.3-7:

99 (1972).

Color

441

442

Hon and Minemura

100. R. Kondo, T. Mitsunaga, and H. Imamura, Mokuzai Gokkaishi, 32:462 (1986). 101. T. Mitsunaga, R. Kondo, and H. Imamura, MokuzaiGakknishi, 33:239 (1987).

102. B. Charrier and J. P. Haluk, Holz Roh- ur~d Wrrkst., 50:433 (1992). 103. P. G . Forsyth and T. L. Amburgey, Forvsr Prod. J., 42(4):59 (1992). 104. B. Kreber, E. L. Schmidt, and T. Byrne, ForestProd. J.. 44(10):63 (1994). 105. W. Sandermannand H. H. Dietrichs, Hol;forsch., 13:137 (1959).

10 Chemical Degradation Yuan-Zong Lai SUNY College of Environmental Science and Forestry, Syracuse, New York

1.

INTRODUCTION

Wood and other lignocellulosic materials are labile to a wide variety of chemical changes. These transformations, depending on the conditions of reaction environment, may vary from an undesirable discoloration (Chapter 9) to a selective breakdown of the major cell wallcomponents [ 1-31. In the chemical utilization of these lignocellulosic substrates, lignin usually plays a negative role, and must be modified, partially degraded, or totally removed, depending on the end uses of the final products. The commercial pulping and bleaching operations 141 generally are very nonselective, being accompanied by a significant degradation of the polysaccharide components. For example, the yield of lignin-free softwood pulp for the most widely used kraft process is only about 44%, as compared to a theoretical 67% for pine [ 5 ] .Thus, a great technical challenge for the paper industry is how to improve the delignification selectivity or carbohydrate stabilization. The fundamental chemistry of the degradation of isolated polysaccharide and lignin samples as well as related model compounds is now reasonably well understood [l-31. The detailed kinetics of reactions involving wood components in situ, however, are still not fully clarified, and are complicated by their heterogeneous nature across the cell wall [6,7], the possible role of lignin-carbohydrate complex (LCC) or linkages [ 1 -3,8.9], and the pore structure of the cell wall matrix. Theoretically, accessibility is a significant factor affecting the degradation behavior of wood polymers in situ, and its significance varies with the nature of chemical environments. This revised chapter largely retains the original format [21 for easy reference, and discusses the chemistry and controlling factors in the degradation of cellulose, hemicellulose, and lignin under acidic, alkaline, and oxidative conditions.

11.

REACTIVESITES

A.

Polysaccharides

The major functional units of wood polysaccharides are reducing end groups, glycosidic linkages, and hydroxyl groups. The reactivity of these units, however. varies considerably among the cellulosc and hemicellulosecomponents,contributing largely to their differences in supramolecular and chemical structures. 443

Lai

444

1.

ReducingEndgroup

Allnatut-al polysaccharide molecules contain a reducing end group which, being hemiacetal i n nature, is partially converted to an open-chain aldehyde function i n solution. This functional group can he reduced and oxidized to a n alditol and aldonic acid moiety, respectively. Also, the anomeric hydroxyl group (at the C1 position). being the nlost acidic [ I O ] . can be selectively etherified [ I I]. Reduction with sodium borohydride is often used for quantitative estimation of the reducing end-group contenl [13_]. The reported contents of g l ~ ~ c o smannose. c, and xylose end groups in wood 1 1 31 are gcnerally consistcnt with the molecular tnasscs established for cellulose. glucomannan, and xylan. Regarding the accessibility of reducing end groups. it was reported by Gentile et a l . [ 141 for a libroushydrocellulosesample based on the assumption that an amorphous cellulose was totally accessible. The latter sample was prepared by regenerating ;I cellulose solution i n a dimethyl sulfoxide (DMSO)-parafortnaldehyde (PF) solventsystem [ 121. Rcducing end groups were clctertnined by reduction with tritiated sodium borohydride i n dilute alkalis. Approximately 12%. of the reducing end groups i n the fibrous cellulose were shown to be inaccessible t o the borohydride treatment. Interestingly, a large difference i n hydroxyl accessibility between ;I native and ;I regenerated cellulosesample ( S 1 versus 99%) was previously indicated by the deuteration method [ IS]. Since the penetration of reagent into thc crystallites would be negligible under the mild conditions used (an 0.25 M borohydride solution a t ambient temperature). it appears that the concentration of reducing end gt-oups is significantly higher i n the amorphouscomponent than in thc crystallites. Reducing end groups i n alkalis undergo readily a series of the so-called Lobt-y de Bruyn-Alberda van Ekenstein transfonnations [ 1 S ) . and play a dominant rolc i n the a l kaline degradation of polysaccharides.

2. Glycosidic Linkages The glycosidic bonds. being acetal i n nature, are hydrolyzable under acidic, alkaline. and oxidative conditions. Acid hydrolysis proceeds very readily and forms the basis of :I saccharification proccss, whereas the alkaline cleavage reaction requires more drastic conditions. This hydrolytic reaction i n general weakens the mechmical properties of wood and fibers.

3. Hydroxyl Groups The intcrunits of ccllulosc and hemicelluloses contain one primary hydroxyl group for each :mydro-hcxose u n i t and two secondaryhydroxylsfor each anhydro-hcxosc and -pentose u n i t . Thew hydroxyl g r o ~ p saresusceptible to oxidation, and the resulting a l dehyde or keto g r o ~ ~may p initiate further degradation reactions. such ;IS dehydration and cleavage o f glycosidic linkages. Among the three hydroxyl groups, the ?-OH group is the most acidic 1 1 1,16- 191. and this has been gencrolly attributed t o an activating effect o f the nnomeric center. Hearne et a l . 1201 observed t h a t methyl P-D-ribopylanoside was more acidic t h m methyl P-D-xylopyranoside, and they differ only i n the C3 conformation. Thus. the acidity of the ?-OH group is also likely influenced by other factors such ;IS the hydrogen bonding system. The hydroxyl reactivities in a heterogeneous system are further affected by the Xcessibility factor. and thc reported data on cellulose have been shown t o vary considerably with the type and the conditions of reactions used. Inaccessibility may arise from either a

Chemical

region being inaccessible to a reagent or a other units.

B.

Degradation 445

functional group being hydrogen-bonded

to

Lignin

Lignin occurring in plants is well known for its variability or heterogeneity in terms of both morphological distribution and chemical characteristics. Significant variations have been observed between juvenile and mature woodlignins; among the normal, compression, and tension wood lignins; and for lignins in different morphological regions. Our present understanding of lignin structure has been obtained largely from analysis of milled wood lignin (MWL) preparations [9], which are usually obtained at less than 50% yield. The millingprocessused in MWLpreparation is knowntoinducesomechemicalchanges, notably an increase in the phenolic hydroxyl group content [21,22]. Additionally, MWL has been shown to originate mainly from the secondary wall lignin [8,23-251. Thus, the extent to which MWL may represent the lignin in situ requires further evaluation. Although the approximate contents of major lignin linkages are now reasonably well understood, the chemical structure of lignin, unlike that of cellulose or the hemicelluloses, cannot be defined precisely. Sincecarbon-carbonlinkages are generallyvery resistant tochemical attack, the degradation or fragmentation of lignin is limited largely to cleavages of ether units at the a - and P-positions. The nature of these hydrolyzable units and other functional groups having a significant impact on the reactivity of lignin is outlined below.

1. HydrolyzableEtherLinkages The hydrolyzable ether units in lignin are the P-aryl, a-aryl, and a-alkyl ether linkages (Fig. I ) . As summarized in Table I , the P-aryl ether based on phenyl propane (C,) units constitutes approximately 50% and 60% of spruce and birch MWL, respectively [21], and is present as two isomers. Proton and I3C NMR analysis indicates that spruce lignin contains about equal proportions of the erythro and threo forms, whereas the erythro form dominates in birch lignin [26,27]. Spruce MWL was reported to contain 6-9% of acid-labile (presumably noncyclic a-aryl) ether units by a mildly acidic hydrolysis reaction [28], but an appreciably lower value (<3%) by a 2-D NMR analysis [29]. The latter method gave a higher value ( 5 % ) for a birch MWL. Also, Lai and Guo [30] obtained a higher value for aspen (6%) than for spruce (4%) wood lignin when measured in situ by a selective acid hydrolysis. On the other hand, spruce MWL was shown to have a higher content of the cyclic a-aryl ether (as in p-5 units) than birch MWL (9-12% versus 6%) [31]. Recently, Brunow et al. [32] suggested the possible presence of dibenzodioxocin unit [(4) in Fig. l ] in lignin. Additionally, lignin may contain some a-ethers linked to carbohydrate [ 1 -3,8,9,33]. Thesebenzyl alkyl ethers, like a-arylethers, are susceptible to acid hydrolysis, but at considerably lower rates. 2. PhenolicHydroxylGroups The phenolichydroxylgroup is one of the mostimportant functionalities affecting the physical and chemical properties of lignin polymers. It plays a prominent role in commercial delignification processes by virtue of its ability to promote alkali-catalyzed cleavages of interunitary ether linkages and the oxidative degradation of lignin [ 1 -3,8,34,35] as well as lignin modification reactions [36]. Reported content of this functional group in

446

Lai

OCH3

OR

Q-

OCH3

OR

1

2

R,= aryl

I

OCH3

H3C0

CH -0

0

0

CHzOH CH3

OCH3

OR

OH 4

3

FIGURE 1 Hydrolyzablelinkagesin

TABLE 1

lignin.

Proportions of HydrolyznbleLinkages i u Milled Wood Lignins

____

Percentage of intermonomeric linkages Type o f linkage P-Aryl ether a-Aryl ether Noncyclic" Cyclic

I) Structure (Fig. Birch

Spruce

(la)

48

60

(Ih). (2) (3)

<3

S

9-12

6

Chemical Degradation

447

spruce MWL preparations showed considerable variability (18-33% of C, units) as summarized by Lai 1221. Significantly lowervalues(10-13%)wereobserved for softwood lignin in situ, while higher values (55-70%) were obtained for soda and kraft lignin 1371. The phenolic hydroxyl group content was generally lower in hardwood than in softwood lignins, andarange of 9-14%wasreported for MWLand cellulolytic enzyme lignins (CEL) preparedfromsweetgum [38]. Laiand Guo[39] showed that hardwood lignins in situ displayed significant variation among different species in the content of this functional group, which decreased with an increase in the proportion of syringyl units in the wood lignin. Conceivably, this is caused by increase in etherified syringyl-type p-04 units resulting from the increase in syringlypropane units in the lignin.

3. Aliphatic Hydroxyl Groups Lignin contains two major types of aliphatic hydroxyl groups, located at the y- and a positions of the side chains. The latter type, being a benzyl alcohol, is very reactive and, like the phenolic hydroxyl group, plays a dominant role in lignin reactions. The amount of benzyl alcohol groups in spruce lignin is about 16 per 100 C, units 1211.

Uncondensed Units The uncondensed units of lignin, in general, may be defined as those units with positions at C2, C3, C5, and C6 being free orsubstituted only by a methoxyl group. Reported values for softwood lignin varied slightly with the analytic method used: 50-60% by oxidation with potassium nitrosodisulfonate (Fremy’s salt) 1401; 50-55% by ‘H NMR [41,42]; and 45-57% by nucleusexchange reactions [43].Thus,softwood lignin containsapproximately an equal proportion of uncondensed and condensed units. Hardwood lignins containingsyringyl units, as expected,have a highcontent of uncondensed units, and a value of 83%wasreported for sweetgum lignin by nucleus exchange reactions 1431.

4.

5. Unsaturated Groups Lignin contains some unsaturated groups, mainlyas coniferyl alcohol and coniferaldehyde end groups. Carbonyl groups may also occur as a-keto or nonconjugated units. The latter type is probably associated mostly with a detached side chain such as a glyceraldehyde group.

6. Ester Groups Ligninfrom certain species may contain a significant amount of estergroups [60,61]. Brauns lignin preparation from aspen was shown to contain more than 7% of p-hydroxybenzoic ester groups [44]. Also, grass lignins contain significant amounts of p-coumaric acid and ferulic acid moieties, which were reportedly esterified mainly at the y-hydroxyl group (80%) with the remainder at the @-position [45-471. Additionally, the wester functionwould also include the possible lignin linkages connected to the 4-0-methyl-glucuronic acid unit of xylan molecule, as summarized by Lai [2]. Obst estimated [481 that approximately IO-20% of lignin-carbohydrate linkages in aspen lignin were labile to mild alkali treatments (0.1 M NaOH at 20°C for 90 min), and these were assumed to be present as an a-ester type.

I

Lai

448

Methoxyl Groups Methoxygroupcontent often serves as an indication of the approximateproportion of different phenylpropane units in the lignin. Methoxyl groups are relatively resistant to both acidic andalkaline hydrolysis. They are, however,hydrolyzablewithconcentratedhydriodic acid [491, hydrosulfide and methyl mercapitide ions [SO], or sulfite ions [SI].

7.

8. Accessibility Lignin appears to be amorphous, occurring in plant tissues and in isolated samples, and, like celluloseandhemicelluloses,ithas a hightendency to formhydrogenbondings. Michell [52]concluded, from infrared analysis of MWL samples and related lignin model compounds, that all detectable hydroxyl groups in lignin were involved in hydrogen bond formation. Both a- and phenolic hydroxyl groups appeared to be involved preferentially in intramolecular hydrogen bonding. Similarly, crystal p-0-4 lignin model dimers were shown to contain a variety of intra- and intermolecular hydrogen bonds 1531. The morphology or fine structure of lignin, unlike that of wood polysaccharides, has not been studied comprehensively [ 1-3,7,8]. C.

Lignin-CarbohydrateComplexes

The concept of lignin-carbohydrate complexes (LCC) has been derived largely from the difficulty encountered in separating the carbohydrate impurity from a lignin preparation, or vice versa. The accumulated evidence strongly supports the existence of chemical linkages between the lignin and polysaccharides components in the cell wall matrix [ l -3,7-9,21,33]. Three major possible types of lignin-polysaccharide linkages are the benzyl ester, benzyl ether, and glycosidic bonds (Fig. 2 ) . The reactivity of these units varies considerably with their chemical structures and the reaction environments. For example, in alkali media, the ester type is hydrolyzed readily, whereas the benzyl ether ofan etherified unit is fairly stable even under alkaline pulping conditions.

1. Ester Linkages (5) The evidence for the occurrence of ester linkages is based mainly on the observation that upon mild alkali treatments of MWL or isolated LCC samples [48,54-59], a significant amount of residual carbohydrates, consisting mostly of xylan, was released. A release of carbohydrates at SO% and 90% for a spruce and a birch MWL sample, respectively. was reportedupontreatmentwith 0.05 MNaOH at ambienttemperatureovernight 154,551. These results are generally interpreted in terms of saponification of an ester-type L-C linkage, probably associated with the 4-0-methylglucuronic acid units of xylan. The presence of ferulate-polysaccharide esters was also demonstrated recently in grasses by NMR studies [60,61].

2. Ether Linkages (6) The portion of LCC resistant to mildly alkaline hydrolysis is generally ascribed to the presence of benzyl ether linkages between the lignin and polysaccharide components. The nature of these alkali-stable linkages has been obtained mostly from analyzing the sugar residues following the typical methylation, Smith degradation, and acid hydrolysis techniques. Reported data indicate that the ether-type linkages could involve all types of wood

Chemical Degradation

r

\

U ,

0 X

P

-6

D

rc 0

b) M

g

D

a

449

Lai

450

polysaccharides, including xylan [59,62,63], galactoglucomannan [59], and even cellulose [59,63]. Concerning the location of sugar units linked to lignin, both the primary and secondary hydroxyl groups are probably involved. Eriksson et al. [59] reported that the participation of arabinose unit (in softwood xylan) occurs mainly at the C2 and C3 positions, whereas the galactose unit of galactoglucomannanprobablyinvolves the C3 hydroxyl group. Minor [63] concluded that the primary hydroxyl groups of galactose and arabinose units are actively involved in lignin-carbohydrate bond formation, while the xylose units involves the C3 position. It was reported that the benzyl ether-type L-C linkage can be selectively cleaved by oxidation with 2,3-dichloro-5,6-dicyanobenzoquinone(DDQ) [33] or by pivalolyl iodide [77]. Also, some LCCs were shown to have a strong tendency to form micelles or aggregates in aqueous solution [33].

3. Glycosidic Linkages (7) The nature of possibleglycosidiclinkagesbetween lignin and polysaccharideshas not been thoroughly investigated. Koshijima et al. [64] analyzed an LCC sample from a pine MWL preparation by the methylation and acid hydrolysis methods. Among other products, 2,3.4,6-tetra-O-methyl-D-xylose, 2,3,5-tri-O-methyl-L-arabinose, and2.3,4-tri-O-methylD-xylose were obtained, suggesting that these sugar units are bound glycosidically to the lignin. The presence of aryl-glycosidic linkages was also suggested for a beechwood LCC sample [65], based on the observation that new phenolic hydroxyl groups were released upon alkaline hydrolysis ( 1 N NaOH at 90°C for 13 h).

111.

ACID-CATALYZED REACTIONS OF WOOD

Acid-catalyzed hydrolysis of glycosidic linkages in polysaccharides and the cleavage of a- and p-aryl ether bonds in lignin are the primary degradation reactions that occur when lignocellulosic materials are placed in an acidic environment. These hydrolytic reactions are often accompanied by further chemical transformations, including dehydration, degradation, and condensation reactions. The nature of these reactions and their implications in biomass utilization are discussed in this section.

A.

Polysaccharides

The acidic degradation of polysaccharides is affected by their physicalstructures, the conformation of sugar constituents, and the nature of an acidic medium.

1. Fundamental Aspects (1. Acid Hydrolysis. Figure 3 illustrates one of the generally accepted mechanisms for the acid hydrolysis of glycosidic linkages [2,66-691. It involves an initial protonation of the glycosidic oxygen followed by decomposition of the resulting conjugate acid (9). The rate-controlling step is likely in the formation of the carbonium-oxonium ion (lo), which may exist in a half-chain conformation (11). Table 2 summarizes the relative hydrolysis rates [70] and activation energy [68,71] for a series of methyl pyranosides pertinent to wood polysaccharides. In general, the panomers react faster than the corresponding a-forms, with an exception of the L-arabi-

451

Chemical Degradation

Ho&F

CH20H

CH20H

=

H0

H0

OH 8

-.+l l+.+ Q&?!!+

l1

CH2OH

H0

11

CH20H slow

H@ 0 @ - I s

other Glucose + products

H20

OH 9

10

FIGURE 3 Acid-catalyzed hydrolysis of glucopyranosides. (From Ref. 68.)

noside. Also, the glycosidic linkages of nonglucose units are generally more reactive than the glucoside. For the p-series, the relative hydrolysis rate increases in the order glucoside ( l ) , mannoside ( 3 ) , galactoside and xyloside (4.8), and rhamnoside (IO). It is evident from Table 2 that the presence of a carboxylic group at the C5 of the methyl glucoside (glucuronoside) reduced the hydrolysis rate by about 50-70%. A larger reduction (97%) wasobservedbetween the twoglycosidiclinkages of cellobioseand cellobiouronic acid [72]. Thereduced reactivity of these glucuronosides had beenattributed to both inductive 173-761 and conformational [71,78-80] effects. It is of interest that the presence of a C6 aldehyde group increased the rate by a factor of 20 to 70 [81]. Also, the two glucosidic linkages of cellotriose behaved differently [82]. The linkage at the nonreducing end was about 50% more reactive than the one at the reducing end, and it was comparable to that of cellobiose. Additionally, the reactivity of glycosidiclinkages is profoundlyinfluenced by the ring size of sugar units. The aldofuranosides, because of their more strained structures,

TABLE 2 Relative Rate and Activation Energy for Acid Hydrolysis of Methyl Pyranosides with 0.5 M Acid at 60-90°C Relative rate''

AEh (kcalhol)

~~

Methyl pyranosides of D-glucose 5.7 D-mannose D-galactose D-xylose9.1 L-rhamnose 9.0 L-arabinose D-glucuronic 0.62 Acid

P-Anomer a-Anomer

P-Anomer a-Anomer

1.9

1

2.4 5.2 4.5 8.3 13.1 0.47

"From Ref. 70 using 0.5 M HCI. "From Refs. 68 and 71 using 0.5 M H,SO,.

32.0 9.2 33.9 19.0 29.3

35.1 34.7 34.0 33.4

32.5

-

-

30.2

-

Lai

452

are hydrolyzed much faster than the corresponding pyranosides [83], by a factor of 100 between the methyl a-D-galactof~lranosideand methyl a-D-galactopyranoside [ 10l ] . The energy of activation for the hydrolysis reaction is very similar among the neutral methyl pyranosides (32-35 kcal/mol), and is lower for the glucuronosides (Table 2). 6 . Dehydrution m c l Condensation Reuctiorls. Although the immediatehydrolysis products of polysaccharides,mono-andoligosaccharides, are reasonablystableunder mildly acidic conditions, they can undergo further dehydration, fragmentation, and condensation reactions to yield a variety of nonsugar products including furan [841 and phenolic [85-871 compounds. Furan Compounds. A good yield of 2-furaldehyde (14) (88% of the theoretical) is generally obtained by distillation of a xylose solution (12) in concentrated hydrochloric acid (12% HCI) (Fig. 4). This reaction constitutes the classical method for determination of the pentosan content [88], and can be carried out in dilute acid (around 0.1 M H2S04) at elevated temperatures (100-280°C) as well [89]. Glucuronic acid, upon heating with acid, also forms 2-furaldehyde, but at a much lower yield than from xylose. Hexoses (13) are more stable than pentoses under acidic conditions, and the major dehydrationproductsare 5-(hydroxymethyl)-2-furaldehyde (HMF) (15), levulinic acid (16), and polymeric materials. The yield of HMF from glucose is generally low (10-20%), and can be increased by raising the pH of the reaction mixture. A high yield of HMF (45%) was obtained by heating glucose in a pyridine-phosphoric acid system at 200238°C for 20 min [90]. HMF polymerizes readily underacidicconditionsandcanbe further degraded to yield levulinic acid (16) and formic acid. Phenolic Compounds. Popoff and Theander [85-871 isolated a variety of phenolic compounds (Fig. 5 ) from mildly acidic treatments of monosaccharides (pH 4 at 96°C for 48h).Compound (19) was the majorphenolicproductproducedfrom the reaction of hexoses (at 3.4% yield). The yield of phenolic products is generally higher from pentoses and hexuronic acids than from hexoses. Phenolic compounds were also formed from the acidic degradation of erythrose and dihydroxyacetone, indicating the complexity of the chemical transformations involved.

2.

Cellulose

a. Overall Process. The heterogeneous degradation of cellulose is characterized by two distinct phases, an initial rapid reaction followed by a slow process [9 1-94], and may be monitored either by weight loss equivalent to the formation of soluble materials (Fig. 6) [95], or by DP changes in the cellulose residue (Fig.7) [91]. The initial phase associated with a small cellulose dissolution (7%) and a sharpDP reductionhasbeengenerally attributed to degradation in the easily accessible region. This observation was often used to indicate the percent amorphous component of a cellulose sample. The results, however, varied significantly withhydrolytic conditions, e.g., acidconcentrationandtemperature ~911. The slow degradation stage was characterized by relatively small DP changes, especially after reaching the so-called leveling-off DP (LODP). Nelson and Tripp [93] observed that LODP was not appreciably affected by hydrolytic conditions, and is characteristics of a cellulose structure [92-941. Both the initial and slow degradation processes are i n agreement with pseudo-first-order kinetics. The initial hydrolysis rate of amorphous components was generally one to two orders of magnitude greater than that of crystallites [92,951.

453

Chemical Degradation

12

CH2OH @ € € O H Y - 3H20 H0 OH 13

14

products Condensation

I

CH2

1

HOHzC

I

acHo CH2

HC02H

+

l 0 c= I

0

CH3

15

l0

FIGURE 4 Acid-catalyzed formation of 2-furaldehyde, 5-hydroxymethyl 2-furaldehyde, and ulinic acid from monosaccharides.

lev-

b. Influence of PhysicalStructure. The hydrolyticbehavior of cellulose is much influenced by its physical structure and lateral order [92-1041. Wood cellulose was hydrolyzedtwice as quickly as cotton [94]. Hydrolysis rate wasnoticeablyincreased by physical or chemical pretreatments including ball milling, mercerization, ammonia treatment, and regeneration. The relative effects of pretreatment, however, depend onthe source of cellulose. Hill et al. [96,97] examined the influence of mercerizationonhydrolytic reactions for several cellulose samples, which were prehydrolyzed to remove the readily hydrolyzable components. The residues were then subjected to isothermal hydrolysis with 2% H,SO,in the 150- 180°C range. As shown in Table 3, mercerization increased the hydrolysis rate of cotton (by 40%) and of ramie (by 7%) cellulose, whereas the opposite

R

17

R = H, OH, CH3, CO2H or COCH3

19

0

R

OH

0

OH 18

R = CH3 or C H 0

20

FIGURE 5 Typicalphenolic compounds formed from acid-catalyzed dehydration and condensation of sugars. (From Refs. 85-87.)

454

Lai

a

o

16

24

32 40 48

Time, h

FIGURE 6 Rate of weight loss in the acidic degradation of a-cellulose in 2 M and 4 M HCI at 90°C. (From Ref. 95.)

> 1500

a

n

1000

500

0

L

L

0

60

120 180

240

300

Reaction lime, h

FIGURE 7 Rate of DP reduction in the acidic degradation of cellulose in 1 M HCI at 50°C. (Data from Ref. 91.)

TABLE 3

Influence of Mercerization on the Hydrolysis Rate Constant ( k ) of Hydrocelluloses (obtained after boiling with 20.2% HCI for 2 min) in 2% H,SO, at 150°C

Mercerized

99

Hydrocellulose Sanlple

DP,,.

k . 10' min- '

Dl',,.

Cotton Ramie Linen a-Cellulose Rayon

162 182

1.5 1.5 2.1

80 95 102

4.4

45

Source: From Ref. '97.

159

-

17

k . 10' min" 2.1 l .6 1.5 3.3 17.8

nc.

Chemical

Degradation 455

effect was observed for linen and a-cellulose samples, which showed an approximate 30% reduction. It is evident that the rayon sample had the highest reactivity. Based on kinetic analysis, it was concluded that the end-attack model proposed by Sharples [98,99] can be applied only to a cellulose I1 structure and not to a cellulose I crystallite. Also, a slightly lower activation energy was observed for cellulose I samples (40.7 kcal/mol) than for cellulose I1 samples (43.0 kcal/mol). Thus, the conformation of cellulose appears to be an important factor affecting its reactivity, and possibly the hydrolytic mechanism as well. c. Reaction Conditions. The degradation pattern of cellulose depends considerably on the hydrolytic conditions, especially those factors that affect the swelling of crystallites. Acid Concentration. The hydrolysis rate of cellulose generally increases with increasing acid concentration (Fig. 6). The hydrolysis of cellulose in dilute acid at elevated temperatures often resulted in glucose yields not exceeding 60% [105]. A complete cellulose hydrolysis for quantitative analysis requires an initial treatmentwithstrong acid (72% H2SOJ)to effect dissolution, followed by boiling in dilute acid (3-4%) [106]. When usingconcentrated acids, e.g., 51% HISO, [107], 12% HCl [108,109], or 70% H,PO, [ 1lo], the hydrolysis process was considerably enhanced, approaching that of a homogeneous reaction. Acidconcentrationalsoappearstohaveaninfluenceon the apparent activation energy of the hydrolytic process. As summarized in Table 4, hydrolysis conducted in dilute acid at elevated temperatures ( > l 50°C) was associated witha rather high activation energy. A similar tendency was noted previously by Nelson [92]. Most of the valuesobtained below 100°C or using concentrated acids were close tothat of model glucoside, cellotriose, or methyl P-D-glucopyranoside. Solvents. The acid-catalyzed degradation of cellulose depends considerably on the nature of the solvent [ 102,1131. The addition of ethanol, propanol, or methyl ethyl ketone accelerates the degradation process, including the formation of nonglucoseproducts, whereas dimethyl sulfoxide (DMSO) has a negative effect. The solvent effect has been explained in terms of it affecting the hydronium ion reactivity [ 1131, or a r e h a t i o n of

TABLE 4

Variation of Hydrolytic Reactions in Activation Energy ~

Acid Sample Corn stover Douglas fir Cellulose I Mercerized Hollocellulose Cotton Mercerized cotton Decrystallized cotton Viscose rayon Cotton Cotton Regenerated cellulose Cellulose Cellotriose Methyl P-D-glucopyranoside

E,, Ref. (kcal/mol) 0.5- 1 .S% HZSO, 0.4- 1.6% HISO,

2% HZSO, 2%H,SO, 2% H2S0, 6 MHCI 6 M HCI 6 M HCI 6 M HCI 1 MHCI 0.1 M H2S0, 1 M HCI 51% HISO, 5 1% H2S0, 5 1 % H,SO,

155-240 170- 190 150- 170 150- 170 1 SO- 170 80808080-

100 100 100 100

30-50 50-70 30-50 18-30 18-30 18-30

45 43 41 43 38

111 1 05

33

97 97 I12 92 92 92 92

30

100

34 35 31

27 27 30 29 29

I00 100 1 07

96 96

456

Lai

structural stress in the cellulose [ 1021. Thus, nonswelling media may help preserve the structural stress of cellulose and enhance its hydrolytic degradation. In addition, cellulose can be effectively depolymerized in ethylene glycol at high temperatures(200-240°C) [ 1141 whileminimizing the oxidativedegradation reaction. Also, the depolymerized residues were shown to have a high accessibility toward cellulase in enzyme-catalyzed hydrolysis.

3. Hemicelluloses Hemicelluloses are amorphous materials and also contain a variety of nonglucose units [6,115- 1171. The nonglucose units, because of their different ring structures and hydroxyl configurations, generally have higher reactivity than the glucose residue, and often can be selectively removed from lignocellulosic substrates. Consistent with the behavior of simple glycosides (Table 2), the relative-hydrolysis rate of p-( 1 +4)-linked polysaccharides in a homogeneous system [ 118,1191 increased in the order cellulose (1) < mannan (2-2.5) < xylan (3.5-4) < galactan (4-5). A heterogeneous hydrolysis of these polysaccharides displayed even greater variations in reactivity, following an order cellulose (1) < mannan (60) < xylan (60-80) < galactan (300). This further demonstrates the important role of accessibility in acidic degradation reactions. Acetyl groups presents in hardwood xylans and softwood galactoglucomannans are also hydrolyzable by acid, especially at elevated temperatures. The acetic acid released contributes significantly to the acidity of reaction media. a. Xylan. The presence of uronic acid groups has a profound impact on the hydrolysis of xylans, as it reduces appreciably the hydrolysis rate of glycosidic linkages (Table 2). Thus, high yields of aldobiuronic acid dimers were generally obtained upon partial hydrolysis of xylans 16.1171. Also, the initial hydrolysis rate of various hardwood xylans (with water at 170°C) was closely related to their uronic acid contents [120]. A higher stability of softwood xylans compared to that of hardwood xylans in sulfite pulping may be partly attributed to their higher uronic acid group contents [6]. The acetyl groups in hardwoodxylanswerefoundto exhibit remarkable stability under the relatively drastic conditions of acid sulfite cooking [ 1211. They were reportedly, in steamtreatmentsof birch wood,more stable than the 4-0-methyl-glucuronosyl unit 11221. On the other hand, the arabinofuranosyl linkage in softwood xylans is very labile, and can be selectively hydrolyzed in dilute H2S0, (0.05 M at 97°C for 3 h) [ 1231 or oxalic acid (0.01 M at 100°C for 1.5 h) 11241. b. Glucornannan. The a-D-( 1+6)-galactosidiclinkage in galactoglucomannans is very labile, andcanbe selectively hydrolyzed in dilute oxalic acid (0.05 M) at100°C [125]. Its high reactivity 1125,1261, however, cannot be satisfactorily explained in terms of the behavior of simple glycosides (Table 2). Interestingly, the alkali-induced deacetylation of glucomannans increased its resistance toward acid hydrolysis, as evidenced in acid sulfite cook [ 127,1281. This was presumably caused by the deacetylated glucomannan being adsorbed onto the cellulose or partially crystallized.

B. Lignin 1. General Aspects Figure 8 illustrates the general types of acidic degradation reactions for a lignin model trimer (21b) containing both a- and @-etherunits [21,34,129,130]. The reaction is initiated by protonation of the benzyl oxygen, followedby a-ether elimination of the corresponding

457

Chemical Degradation GI-CH -C-CH, bH 8

CH2OH CHzOH

c 4

G,-C-CH .CH,

I

6

7 2

OCH,

OR

OR

21

OH

a: RI = H b: R,= a r y l

26ketones Hibberts

25

OR

30

@H2$

B

OCH,

H,CO OH

OH

31

t

T

+

OCH,

CH0

I

OR

28a

27

!ab

OCH)

FIGURE 8

Acidic degradations of a- and P-ether units.

phenol or alcohol to give the benzylic carbonium ion intermediate (23), which may undergo three major competing processes. Pathway A leads to the formation of the C&-enol ether (24), which readily undergoes acid-catalyzed hydrolysis to give the a-P-keto1 (25), and then Hibbert's ketones (26). Pathway B involves a carbon-carbon bond cleavage between the P- and y-positions via

Lai

458

reverse Prins reaction to give formaldehyde and the C,C,-enol ether (27), which is degraded slowly to homovanillin (28a). Pathway C consists of an intermolecular electrophilic addition of benzylic carbonium ion to another aromatic unit. giving mainly the a-6-diphenylmethane (DPM) unit (30) and some a-5 condensed structure. Additionally, formaldehyde generated from reaction B may condense with two aromatic units to form another DPM-type condensed structure (31). These condensation reactions have been well established in lignin model reactions [ 13I , 1321. Table 5 illustrates the relative hydrolysis rate of lignin model compound reactions (with 0.2 M HCI in 9: 1 dioxane-H20 solution at 50°C) reported by Johansson and Miksche [ 1331.It is evident that both the a - and P-aryl ether hydrolyses were enhanced by the presence of a phenolic hydroxyl group. Also, the a-aryl ether was much more reactive than the P-aryl ether, roughly by factors of 25 and 65 for the phenolic and etherified units, respectively. Reactivity of the a-ether units varies with their chemical nature [ 1361. Leary and Sawtell [l341 showed that a p-hydroxybenzyl a-aryl ether was about 400 times more reactive than a p-hydroxybenzyl alkyl ether.

2. a-ArylEther Model compound studies, especially by Meshgini and Sarkanen [ 1361, indicate that the overall a-aryl ether hydrolysis (pathway A, Fig. 8) was significantly affected by the nature of the benzyl and a-ether groups. As shown in Table 6, benzyl units (ring A) of the syringyl type, as compared to that of a guaiacyl or p-methoybenzyl moiety, reduced the hydrolysis rate, andhad a higher activation energy. The reaction was also retarded by the presence of a @-aryl ether unit. On the other hand, a syringyl moiety on the a-ether unit (R, group) had a positive effect. The activation energyvariedfrom 19 to 28 k c a h o l , and appears to be related to the stability of the ether linkage. Also, solvents play anoticeable role in the a-aryl ether hydrolysis 134,1361. The rate generally increases with increasing solvent polarity or a decrease in the proportion of organic solvent (dioxane or ethanol) in aqueous systems.

TABLE 5 Relative Rates for Hydrolysis of a- and P-Ether Lignin Model Compounds in 0.2 M HCI A q u e o u s Dioxane at 50°C Linkages Structure (Fig. ~~

~

I)

Relative rate

~

P - A r y l ether

Nonphenolic Phenolic a-Aryl ether Nonphenolic Phenolic

(la), R = CH,

(la), R = H (lb), R = CH, (lb), R = H

I 12

6.5 305

a-Alkyl

Phenolic” “Estimated from the data Source: Ref. 133.

G-O-(CH2)3“G in Ref. 134; G = guaiacyl.

0.8

Chemical Degradation

459

TABLE 6 RelativeRatesandActivationEnergies for the Hydrolysis of Nonphenolic a-Aryl Ether Linkages with 0.2 N HCI in 47.5% Aqueous Ethanol at 50°C Compound (39)rate

Relative

A E (kcal/mol)

Formula"

H,CO@-CH,-0-@CH, H,CO@-CH,-0-@CH, H,CO@-CH,-0-@CH,

H,co@-cH~-o-@ H,co@-cH~-o-@ H,CO@-CH,-0-@ H,CO@-CH-0-@CH,

1

30 30

20 33 3

12

24.5 21.7 23.7 21.3 18.9 21.9 22.8

I

h

O-@CH2 H,CO@-CH-O-@CH,

0.4

28.2

.'G = guaincyl; @ = syringyl: @ = p-hydroxybenzyl. Sortrce: Ref. 136.

3. /?-Aryl Ether The reactivity of P-aryl ether linkages, like that of a-aryl ethers, is substantially enhanced by a phenolic hydroxyl group [133,137-1411, and is influenced by their structures and the reaction conditions. a. Reaction Mechanism. The major degradation pathway of P-ether units shown in Fig. 8 is generally accepted as proceeding through an ionic mechanism under acid-catalyzed conditions [ 142- 1461. Under typical acidolysis conditions (0.2 M HCl in 9: 1 dioxane-water, 4 h at lOO"C), the ether cleavage reaction (pathway A) predominates, yielding Hibbert's ketones (26). These reactions have been used extensively in lignin analysis [21,144]. However, different mechanisms appear to be involved for hydrolytic reactions conducted in the absence of an acid catalyst. The phenolic P-aryl model compounds (32) and (33), when treated with water at 100- 130°C [ 138,1391 or with SO% aqueous dioxane at 180°C [ 147,1481 gave a variety of transformation products, including dimers of the P-S (35), P-l, and P-P types, and other condensation products (Fig. 9). Formation of these coupling products was explainedin terms of a radical mechanism by Sano et al. [147,148]. It should be noted that the etherified P-aryl ether dimer of (32) was unreactive in aqueous solution at temperatures below 130°C. Also, the phenolic Paryl ether dimer of the syringyl type, unlike the guaiacyl dimer, was reactive even under steam treatment conditions [140], and gave complicated products in aqueous dioxane at 180°C [ 1481. b. Solvent. Solventshave a significant influence on the overall degradationof Paryl ethers. The hydrolysis reaction conducted in an aqueous solution was enhanced by the addition of dioxane, and unlike the a-aryl ether hydrolysis [ 1361, was retarded in the presence of ethanol [ 1371. These organic solvents were shown to favor the ChC2enol ether formation (reaction B in Fig. S), especially at elevated temperature [l451 as illustrated in Fig. 10. Thus, the reduced ether hydrolysis in ethanol solution can be partially explained

Lai

460 OCH3

Q

RI

H20 130°C. 6 h

.

Condensation

c

products

OCH3

OCH3

OH

OH

(12.5%)

OH

32

34

(2 1.5%)

35

RI= CH20H (61.2%)

RI= CH =CH-C&OH

R1= CH20H (16%)

RI= CH=CH-C&OH (12.5%)

(6.9%)

CH3

Q OOi C H 3

OH

33

1

G- CH =CH -CH20H (2.6%)

38

G - p (p-5)G- CH=CH-CH20H H2O- Dioxane 1 80°C, 20 min

(4.5%) (1.4%)

+

v CHzOH

J 0

36

Condensation products

(65.9%)

QOCH3

0

37

39

W",

FIGURE 9 Hydrolysis products of phenolic guaiacylglycerol P-aryl ether in water at 130°C for 6 h (from Ref. 160), and in 50% aqueous dioxane at 180°C for 20 min (from Ref. 147).

in terms of increased formation of C&-enol ethers which, as noted earlier, are relatively resistant to acid hydrolysis. Hoo et al. [ 1451 studied the kinetics of acidic hydrolysis of P-aryl ether dimers in 50%aqueousethanolcontaining0.2 M HCl, and obtainedasimilaractivationenergy (-36 kcal/mol) for both the phenolic and etherified types. This value is substantially higher than that of the a-aryl ethers (Table 6) [ 1361.

461

Chemical Degradation

1.200r

l

I

8 I

i

m 17OoC 0

0.900

14OT} E ~ HHP:

f

1709c}Dioxane:H20

O 140°c 0 17OoC, Pure H 2 0

I

I

l

/

(0.002M HCI) /

l

/

Vol% Organic Solvent in HzO

FIGURE 10 The effect of solventcompositionand reaction temperatureontheethercleavage ( k , ) and enol ether formation (k,3)from acidic treatments of erythro-veratrylglycerol &(2-methoxyphenyl) ether. (From Ref. 145.)

Acidictreatments of @-aryletherdimers in concentratedorganicacids,however, resulted in only limited ether cleavages. These studies include using a 85% formic acid at reflux temperature [149,150] and a 75-90% acetic acid at 160- 180°C [ 1511.

4. Carbon-Carbon Linkages Acidic cleavages of the carbon-carbon linkages in lignin are limited mainly to the bond between the p- and y-carbon atoms, as indicated in reaction B (Fig. 8) for a p-0-4dimer with a simultaneous release of formaldehyde. Similar reactions may also occur with a p1 (40) or p-5 (42) type units (Fig. 11) [ 129,1521. These reactions do not contribute significantly to the formation of low-molecular-weight lignin products.Formaldehyde released, however, may participate in the lignin condensation reactions. Under acidolysis conditions [ 1.521, the formaldehyde yield from lignin model dimers decreased in the order p-1 (15%) > p - 5 (9%) > p-0-4 (3%). The main product from a p5 dimer (dihydrodehydrodiconiferyl alcohol) (42) was a phenyl courmarone (43b) (75%). The latter product has a characteristic UV absorption, and is often used in quantitative estimation of the p-S units [ 153,1541.

5. Lignin-Carbohydrate Complex Among the three possible types of lignin-carbohydrate (L-C) bonds (Fig. 2), the ester (5) and glycoside types (7) are probably more labile to acid hydrolysis than the ether type (6). Model compound reactions [ 1351 indicate that the benzyl ethers of vanillyl methyl

462

Lai

HCHO

OCH3

OCH3 OCH3 42

H 0 OCH3

OCH3

-

OCH3

43a

43b

FIGURE 11 Formation of formaldehyde from the acidic degradation of p-1 and p-5 lignin model dimers. (From Ref. 144.)

ether (44a) and methyl 4-0-veratryl a-D-glucopyranoside (44b) were substantially more stable than the glycosidic linkage of methyl a-D-glucopyranoside (44d) (Table 7). The reactivity of benzyl ethers was also significantly enhanced by the presence of a phenolic hydroxyl group [ 1331. Judging from the behavior of model compounds (Tables 5 and 7), LCC of the aether type, if present in lignin, can only be hydrolyzed slowly [135]. Also, the etherified units are more stable to acid hydrolysis than the glycosidic linkages.

6. Condensation Reactions Lignins are known to undergo condensation reactions even under mildly acidic treatments [9]. This is attributed mainly to the high reactivity of the benzyl hydroxyl groups. Three major types of lignin condensation processes are possible based on the lignin model compound reactions. a. Phenolation. This type of intermolecular condensations occurs between the benzyl carbon and another aromatic nucleus, mainly at the 6-position (30) (80%) (Fig. 8) plus some at the 5-position [131,155-1601. The condensation reaction varies with the nature of the phenyl units and reaction conditions. Yasuda and Ota [ 1601 reported that syringyl nuclei condensed more readily than guaiacyl nuclei upon reaction with vanillyl alcohol in 5% sulfuric acid at 100°C. The formation of benzyl chloride was observed upon treatment of p-0-4 dimers in hydrochloric acid, and this would reduce condensation at the benzyl position. b. Formaldehyde Addition. The condensation of aromatic units with formaldehyde results in the formationofmethylenecross-links (31) withpossiblysome1,3-dioxane derivatives [159]. Acidic hydroxymethylation occurs largely at the C, or C, position of aromatic nuclei, which may be phenolic or etherified. The initial hydroxymethylation for syringyl units was faster than for guaiacyl units, and was promoted by the presence of a phenolic hydroxyl group, whereas the subsequent cross-linking reaction was facilitated by an increase in acid concentration and reaction temperature [162]. c. Intramolecular Type. Yasuda et al. [ 1631 identified a phenylcoumaran-type condensation product (46) in the acidic treatment (5% H,SO, at 100°C) of a p-ether dimer

463

Chemical Degradation

Relative Hydrolysis Rates of BenzylEthersandMethyl Glucoside in 0.1 M HCI at 75°C

TABLE 7

Compounds rate

(44)

Relative

Structure

cH3of)--”cn,-ocH3 -

A

1

C H30

B

1 (3343

OH

CH20H

D

20 OCH3

OH

(45) (Fig. 12). Thisintramolecularcondensation was shown to be dominant in an 85% formic acid treatment [ 149,1501,whereas it was practically insignificant for acid-catalyzed (0.2 M HCI) reactions i n 50% aqueous ethanol at 135°C [ 1371.

7. Lignin In Situ The overall degradation of lignin, like model compound reactions, depends considerably on the acidic environment. In aqueous media, lignin condensation reactions dominate and lead to the formation of acid-insolubleresidues.This principle servesas the basis for quantitative determination of lignin content in plant materials [ 1641. Lignin condensation

45

R=HorCH3

FIGURE 12

46

An acid-catalyzed intl.nmolecular condcnsation reaction. (From Ref. 163.)

464

Lai

reactions, however, can be minimized by using mildly acidic conditions in the presence of organic solvents, or nucleophiles. a. Atyl-Ether Cleavages. Lai and Guo [ 130,1651 determined the acid-catalyzed hydrolysis of aryl ether linkages in wood lignin, which was evaluated in terms of the phenolic hydroxyl group generated. As indicated in Fig. 13, temperature had a significant influence on the aryl ether hydrolysis reaction. The low-temperature reaction (<65"C) displays two distinct phases, notably in the case of spruce wood lignin. The rapid phase was likely attributable to the hydrolysis of very reactive noncyclic a-aryl ether or possibly the dibenzodioxocinlinkages (4) (Fig. 1) [32], whereas the slowphasewas likely associated with the P-aryl ether hydrolysis. Accordingly, the highly reactive a-aryl ether units (presumably the noncyclic type) were determined to be 4% and 6% of C,, units for spruce and aspen lignin, respectively [30]. It is also evident that the aspen lignin contained a high proportion of P-aryl ether units with hydrolysis rates substantially higher than those of the spruce lignin. Nimz [ 166,1671 subjected wood to water percolation at 100°C for several weeks. Approximately 20% and 40% of the spruce and beech wood lignin, respectively, became soluble, and contained a variety of lignin oligomers. Sakakibara [ 1681 obtained similar results upon lignin hydrolysis in 50% aqueous dioxane at 180°C. These soluble products were assumed to arise mainly from the cleavage of a-aryl ether units. b. Condensation. The extent of lignin condensation depends on the acid used, and is oftenmeasuredbyareduction in the yield of simplealdehydes upon nitrobenzene oxidation [8,169]. It is generally higher in sulfuric acid than in hydrochloric acid solution [9,170,17l], probably because of the chloride ions being able to form a stable benzyl chloride adduct [ 1611. Similarly, nucleophilic species, such as bisulfite and thioglycolic acid, are known to reduce lignin condensation reactions in dilute acid by being able to undergo sulfonation or formation of thioglycolic acid derivatives [ 1461. The sulfonation reaction [SI] is essentially the chemistry of delignification in commercial sulfite pulping [ 1721. Also, the extent of lignin condensation reaction varies with the nature of aromatic units, beinghigher for the guaiacyl than for the syringyl units, asobservedduring the initial hydrolysis of sweetgum wood in concentrated hydrochloric acid [173j.

C. Applications The chemical analysis and utilization of lignocellulosic components require a quantitative separation or a selective degradation of individual components, which, however,is difficult to achieve, especially for lignin [ 1741. even under an ideal laboratory condition. Several analytical and practical processes have been developed based on controlled degradation of the individual cell wall components [2].

1. Acidolysis Under typical acidolysis conditions (with 0.2 M hydrogen chloride in a 9: 1 dioxane-water mixture at 100°C for 4 h) [21,144], lignin was depolymerized through a- and P-aryl ether cleavages to givesolubleproducts.The yieldof monomericanddimericproductswas substantially higher from birch (30%) than from spruce ( 1 7%) lignin. The aryl ether hydrolysis reaction can be further enhanced by solvolysis in a dioxane-ethanethiol solution containing boron trifluoride ethereate (thioacidolysis) [ 1751. These low-molecular-weight products provide valuable information about lignin structure.

465

Chemical Degradation

t60

0 2

4

6

8

Reaction T m e . h

85-c

IO

12

Lai

466

2. Acid-Sulfite Pulping Conventional calcium-based acid sulfite pulping [4] uses an acidic liquor (pH = 1.5-2) containing 6% total and 1% combined SO,, and is conducted in the range 120-140°C for 5-20 h. The mechanism of delignification [34,51,172] is attributed mainly to the hydrolysis of benzyl ether linkages (Table 5 and Fig. 8) and the sulfonation reactions. Wood polysaccharides, especially hemicelluloses, are very susceptible to dissolution and hydrolytic degradation reactions. Table 8 illustrates typical compositionof sulfite pulpsfromspruceandbirchas compared to that of the original wood [5]. It is evident that the hemicellulose loss was very substantial, notably in the case of birchxylan (83%). A large proportion of the dissolvedpolysaccharideswas in the formofmonosaccharides [ 1761. The acid sulfite process, because of the problem of species limitation, chemical recovery, and weaker sheet properties, had been largely replaced by the kraft process. 3. Prehydrolysis Hemicellulosesfrombiomasscanbe preferentially removed by hydrolysis in water (at 170°C) [177,178], in dilute acid (0.1 M HCl at 120°C) [120], or in concentrated acid (2030% HCI at 40°C) [ 1791. These prehydrolysis conditions resulted in little dissolution of the cellulose and lignin components, and may be used in connection with kraft pulping to produce dissolving pulps [4] or with wood saccharification. To preserve pulp quality for the production of dissolving pulp and other applications [ 180- 1821, the prehydrolysis is normally carried out in water at 170°C for about 2 h withthe organic acid released (mainly acetic acid) acting as catalysts. On the other hand, prehydrolysis for wood saccharification emphasizes the quality of the hydrolyzates. In terms of a xylose production, prehydrolysis with dilute acid (0.4% H2S0,) gave better results than steam hydrolysis alone [ 1771. The xylan removal from hardwood generally displayed an initial rapid phase followed by a slow reaction [ 120,1831. Thus, xylan contains fractions of different reactivity. Reported activation energy for the initial-phase reaction varied from 28.2 kcal/mol based on wood solubilization in HCl (0.56.0 M) [ 1841 to 22.8 kcal/mol based on a xylose formation in 30% HCl solution [ 1791.

Autohydrolysis Autohydrolysis is basically a steam hydrolysis process conducted at elevated temperatures (175-220°C) [ 1851. The bulk of hemicelluloses become solubilized, while lignin can be

4.

TABLE 8 Typical Composition of Softwood and Hardwood SulfitePulps Percentage of original wood Components Pulp yield Lignin Cellulose Glucon1annan Xylan Pitch Other

Birch IO0 27 41

52 2 41

18

S 4

8 2 4

100

0.5

20 40 3 30 3

-

4

49 2 40 1

S

I -

Chemical Degradation

467

extracted from the residue by organic solvent or alkali. After a short hydrolysis of aspen at 215°C (for 4 min) or at 195°C (for 25 min), about 90% of the lignin became soluble in a 90% aqueous dioxane solution. The lignin dissolution resulted largely from cleavages of the a-aryl ether plus some P-ether units [ 186- 1881 and had an activation energy (29.3 kcal/mol), similar to that of acid hydrolysis for an etherified syringyl a-aryl ether model dimer (28.2 kcal/mol) 11361. At a given temperature, the extractability of lignin as a function of time went through a maximum [ 1851, indicating that lignin condensation is a controlling factor [187-191]. Phenolic compounds, especially 2-naphtho1, were reported to be effective in preventing lignin repolymerization [ 19 1,1921.Thus, hardwood under proper autohydrolysis treatment followed by solvent extraction can be separated roughly into its three major components. The autohydrolysis process is generally unsatisfactory for softwood with respect to delignification. Also, the pulp produced is not attractive for papermaking, because of its poor sheet properties caused by the hydrolytic degradation of cellulose [ 1931. The latter degradation, however, can be reduced slightly by continuous removal of the steam condensate during autohydrolysis [ 1941.

5. Steam Explosion Process The steam explosion process can be considered a modification of the autohydrolysis conditions. The original Iotech process [l951 involves a steam hydrolysis of wood materials at 234°C and 600 psi for about 1 min, followed by sudden decompression to atmospheric pressure. The combined chemical and mechanical action resulted in extensive depolymerization of the polysaccharide and lignin components. A high proportionof P-ether linkages, in addition to labile a m y l ether units, were hydrolyzed [196-2031, and a value of 60% was observed for a eucalyptus wood sample [ 1991. Based on the nature of soluble lignin isolated from the steam explosion of aspen, the degradation of P-0-4 ether structures likely involves both ionic and radical mechanisms, and results in a significant formation of the C&-enol ethers (Fig. 8) and unsaturated structures (Fig. 1 l ) . The exploded materials were shown to have high accessibility toward enzymes [ 195,2041, and have been used for the production of ruminant feed and fermentation substrates. Also, they can be separated roughly into hemicellulose (soluble in water), lignin (by alcohol extraction), and cellulose residues [205,206] for further utilization as a source of liquid fuel and chemical feedstock. 6. Organosolv Delignification The principle of acid-catalyzed organosolv pulping is basically a controlled steam hydrolysis i n the presence of organic solvents 1207-21 I ]. Among various pulping processes reported, the Alcell process, usingaqueousethanolunderautocatalyzcdconditions, is currently the only one in a pilot-scale operation [212,213]. In recent years, solvent-assistcd alkaline delignification (organosolv pulping) has also received much attention, notably in sodacooking with an acidic methanolpretreatment (Organocell), and acombined alkaline sulfite-anthraquinone(AQ)-methanol (ASAM) system [ 2131. In acid-catalyzed organosolv processes. the partially degraded lignin and hemicellulose components are solubilized simultaneously, and thc delignification selectivity varies with the nature of the solvent and catalyst used ;IS well ;IS their concentrations. Although ethanol. methanol, and sulfuric acid are most commonly used i n laboratory studies, delig-

Lai

468

nitication with organic acids, e.g.. acetic 12 141 and formic [ 149- 1 S l ] acids. also has received much attention. Available data [30,165,2 15-2 171 suggest that the a- and p-ether units cleavages arc very important to acidicdelipnilication. Also. the hydrolysis of L-C linkages may bc involved 12071. because isolated organosolvlignins generally have a low carbohydrate contcnt. Lai and Mun [ 2 16,7 171 recently reported the significance of methanol and aryl-ether hydrolysis i n the acidic delignitication of aspen wood. As indicated i n Fig. 14, the acidic system without methanol resulted i n preferential retnoval of the carbohydrate components. For the two mcthanolic systems, the autocatalyzed process had ;I slightly higher delignification selectivity than the acid-catalyzed one. Figure 1 S illustrates the dependence of acidic delignification on aryl-ether clcavagc reactions as indicated by the l’ormation of phenolic hydroxyl groups. A nearly linear re-

1 0 3

c

0 ’= a

60

.-V

L

20

0

I

Chemical

Degradation 469

lationship was observed initially, and extended to approximately 30% and 60% delignification for the watersystemand the twomethanolsystems, respectively. The reduced impact of aryl-ether hydrolysis on delignification observed in the water system was likely caused by lignin condensation reactions 12151. 7.

WoodSaccharification

The hydrolysis of cellulose in the wood for glucose production (saccharification) is usually performed on prehydrolyzed residues with low hemicellulose content. Cellulose, because of its physical structure and crystalline nature, is relatively resistant to acid hydrolysis. This physical inaccessibility presents a formidable task for the commercial production of glucose from ligno-cellulosic substrates. The glucose yield varies considerably with hydrolytic conditions. With dilute acid (0.1 - 1.6% HISO,) at hightemperatures(around 200"C), the glucose yield was usually less than 60% [ 105,220,22 I], because of the formation of nonsugar products. In general, the net glucose yield favored a short reaction duration at high temperatures; e.g., a 54% yield was obtained with 0.4% acid at 260°C for 27 S [2201. With concentrated acid at low temperatures (around 40"C), cellulose is extensively swelled and eventually dissolved, and the hydrolysis proceeds rapidly to give glucose in nearly quantitative yield [ 1061. The use of hydrochloric acid up to 16 M in the 20-50°C range was also reported [log]. The reactivity of the hydrolysis lignin residue depends on the type of acid used. Lignin obtained from hydrolysis in hydrochloric acid, as compared to sulfuric acid, was shown to be less condensed [ 170,17 l], and may be more attractive for by-product applications. Additionally, there are nonconventional methods that use organic acids or solvents. The useof trifluoroacetic acid[222],hydrogen fluoride [223,224],amixture of acetic acid-acetic anhydride-dimethylformamide-sulfuric acid [225], or aqueous acetone with acid catalyst [2261 has been explored. The organosolv delignification of hardwood was reported to have an activation energy of 19.2 kcalhnol(forblackcottonwood in aqueousmethanolcontaining 0.05 M H2S0,) [218], or 16.2- 19.8 kcalhol (for birch in 60%aqueousethanol)[219].These values are slightly lower than those for the hydrolysis of a-aryl ethers dimers (18-28 kcal/mol) (Table 6) [ 1361, and substantially lower compared to that of P-aryl ether compounds (36 kcalhol) [ 1371. This variation may suggest that diffusion is a significant factor in the acidic organosolv delignification process.

IV.

ALKALI-CATALYZED REACTIONS OF WOOD

The primary hydrolytic degradation of wood components in alkaline media, like that under acidic conditions, involves the cleavage of hydrolyzable linkages in lignin ( a - and /?-aryl ethers) and of glycosidic bonds in polysaccharides. However, the alkaline and acidic degradations proceed through distinctly different mechanisms. The major loss of polysaccharides from the alkaline degradation process is caused by endwise depolymerization reactions (peeling),leading to the formation of carboxylic acid derivatives.Thealkaline degradation of lignin plays a dominant role in the utilization of lignocellulosic components, and constitutes the fundamental chemistry of the alkaline pulping or kraft process, whereas the alkaline degradation reactions of polysaccharides cause undesirable losses in pulp yield and necessitate the use of an excess of alkali to neutralize the acidic degradation products.

Lai

470

A.

Polysaccharides

The chemistry of the alkaline degradation of polysaccharides and related model compounds has been extensively studied and reviewed [ 1-4,130,2271.

1. Endwise Degradation (Peeling) a. Mechanism. Reducing end groups play a key role in the alkaline degradation of polysaccharides by being able to undergo a series of so-called Lobby de Bruyn-Alberda van Ekenstein transformations [ 151, leading to the so-called peeling or endwise depolymerization reaction. The peeling reaction, as indicated in Fig. 16 for cellulose and other 1,4-linked polysaccharides, is initiated by an enolization of reducing end groups to form enediol intermediatesincluding (48) and (49) [ 15,2281. Theintermediate (49) thenundergoesa peliminationprocess resulting in adetachment of endgroupsfrom the cellulosechain (reaction A). The peeled end unit may proceed by either a benzylic acid rearrangement to yield isosaccharinic acid (51) or a cleavage between the C3 and C4 linkage followed by benzylic acid rearrangement to give lactic acid (57). Other fragmentation patterns are also possible, as reflected in the detection of products such as formic and glycolic acid [lS]. The average amount of acidic products produced per glucose unit degraded is fairly constant,beingapproximately 1.5 mol regardless of the reaction temperatureor alkalinity [229-2311. The remaining cellulose chain contains anew reducing end group, which can proceed by the same peeling process repeatedly until a stable end group is formed. Johansson and Samuelson [2321 determined the composition of stable acidic endgroupsformedfrom alkali treatment of cellulose at 170"C,and the majortypesweremetasaccharinic (53) (71%) and 2-C methylglyceric (55) (23%) plus small amounts of aldonic acid (6%). Formations of the first two end groups, as illustrated in Fig. 16, were likely initiated from a p-elimination of the C3 hydroxyl (reaction C) and from a cleavage of the C4-CS linkage (reaction B). respectively. Interestingly, the alkaline treatment of xylose and glucose with dilute alkali (0.63 M NaOH at 96°C for 4 h under nitrogen), like that under acid conditions [85-871 (Fig. 5 ) , also producedsmallamounts of cyclicenolsandphenoliccompounds(Fig. 17) [233]. These products likely resulted from a series of fragmentation, dehydration, and condensation reactions. h. Corztrollirlg Fcrctors. The overall peelingprocesscontrolled by twocompeting reactions (peeling and stopping) is significantly affected by the nature of the substrate and reaction environments, especially the type and concentration of alkali used. Accessibility. In a heterogeneous system, the submicroscopic structure o f cellulose exerts a dominating influence on the termination process o f the endwise degradation reaction. Many reports 1230,234-2381 indicate that when hydrocellulose was treated with alkalis (7% NaOH at 100- 120"C), not all the cellulose chains were terminated by stable acidic end groups, and stable residues still contained noticeable amounts of reducing end groups 114,236-2381. This phenomenon, especially important in low-temperature reactions, was ascribed to ;I physical stopping process 12351, when a degrading cnd reached a crystalline region inaccessible to the alkali. Available data (235,251,2521 also indicated that the number of pccled-off glUCOSC units for each reducing end group. either existing in hydrocellulose or formcd by alk aI'tne hydrolysis of glycosidic linkages in situ, was nearly independent of reaction temperature

471

Chemical Degradation

H2COH

R 0e

o

-

-

RO-

OH

t

47

OH 51

H0

OH 48

0 50

49

I

R&H=0 OH

0 OH

52

[:gyH -

54

56

1 1

l

H2COH

-

I C02H

R 04 7 c 0 2 H OH 53

I

R0

CHOH OH 55

I

CH3 57

FIGURE 16 Endwise depolymerizotioll (peeling) process of cellulose.

or alkalinity below the mercerizing strength (<2 M NaOH). An average of 68 and 40 was observed for a native and mercerized cellulose, respectively. As anticipated, the peeling process wasconsiderably more extensive for soluble polysaccharides.e.g.,amylose [2291 orglucolnannans 12391, because of the lack of it physical stopping process. It was observed that mlylose could be totally degraded by a peeling reaction alone in dilute alkali treatments at low temperatures (2291. Saccharide Composition. The peeling reaction of wood polysaccharides is affectcd by their- composition, a s reflected i n the behavior of disaccharides in alkalis 1240.241 1. Thc disaccharide reactivity increases in the order mannobiose < ccllobiose <

472

Lai

CH3 I

c=o

R &OH OH R=H

17

R

QOH OH 18 R= CH3, CH0

OH 58

CH3 l

OH 59

FIGURE 17 Typical phenoliccompoundsformedfrom sation of monosaccharides. (From Ref. 233.)

base-catalyzed dehydrationandconden-

xylobiose. The high stability of mannobiose may be attributed to its isomerization to a fructose moiety being a slow reaction [242]. The formation of stable acidic end groups is significantly higher for xylobiose than for other disaccharides, andgenerallydecreases with increasing reaction temperature. Alkali Concentration. Thepeelingandchemicalstopping reactions forhomogeneous degradation of amylose [229] were shown to be consistent with the formation of mono- and dianions from reducing end groups as reactive intermediates, respectively. The rate of peeling increased with hydroxyl ion concentration up to 0.1 M alkali, remaining constant thereafter; whereas the rate of termination reaction continued to increase beyond this point, leveling off finally to a constant value at approximately 1.5 M alkali. The influence of alkali concentration in a heterogeneous degradation of cellulose is somewhatcomplicated by the fact that it can affect bothphysical accessibility of the substrates and relative rates of the peeling and chemical stopping reactions. Lai and Ontto [243] observed that the extent of peeling of hydrocellulose at 120°C increased with alkali concentration up to approximately 6 M and decreased sharply thereafter. The initial enhanceddegradationwasascribedtoincreased accessibility of the cellulose,while the reduced degradation in the high-alkalinity region was caused by increased formation of a stable acidic end group. Q p e of Base. DivalentcationssuchasCa” [230] and Sr’+ [234] are known to retard the peeling process. When hydrocellulose was treated with a concentrated strontium hydroxidesolution,most of the degradingchainswereterminated to stable acidic end groups [234]. The observed effects may be rationalized in terms of cation stabilization of a dianionic intermediate. Also, the peeling of hydrocellulose in mild alkali solutions (pH 9- 1 I ) was significantly reduced by the addition of ammonia, and to a lesser degree by borate 12441.

Chemical

Degradation 473

Temperature. The activation energy of the peeling reaction (21.2kcal/mol)for amylose in ahomogeneoussystem[229]was slightly higherthan that ofachemical stopping reaction (19.3 kcal/mol). Thus, the extent of amylose degradation, like that of disaccharides [240], was significantly increased at high temperatures. On the other hand, the peeling of hydrocellulose in a heterogeneous system [235] had a substantially lower activation energy than the chemical stopping reaction (24.6 versus 32.2 kcal/mol). However, the expected positive effects of increasingtemperatures on cellulose stabilization were not observed, and were probably counteracted by an accompanying increase in cellulose accessibility at high temperatures. Additives. The peeling reaction can be reduced or prevented by chemical modification of the reducingendgroups with additives[235,245].Reducingagentssuchas sodiumborohydride[245]andhydrogen sulfide [246,247]wereeffective in converting reducing end groups to alkali-stable alditol units. Similarly, reducing end groups could be convertedintostable acidic endgroups by a variety of oxidationagents,includingan alkali-oxygen system [248-25 l], anthraquinone (AQ) derivatives [252-2571, and polysulfide 1258,2591. Among these agents, only O,, polysulfide, and AQ have some practical significance. 2.

Cleavage of Glycosidic Linkages

a. Mechanism. Figure 18 illustrates apredominantmechanismgenerallyaccepted for the alkaline cleavage of the P-D-glycosidic linkage, which has beenlargely established forsimpleP-D-glucopyranosides[67,260-2651. An initial rapidequilibrium-controlled process likely involves ionization of a C2-hydroxyl (60). In the rate-determining step, the C2 hydroxyl anion (61) undergoes an intramolecular displacement process to yield a 1,2anhydride intermediate (62). The latter anhydride may be hydrolyzed to yield a reducing end group or transformed into ;I 1,6-anhydroderivative (63). The latter reaction provides a facile preparation of 1,6-anhydro-P-D-glucopyranose[levoglucosan (63)] [266]. Studies with 1,5-anhydrocellobiitoI [267,268] indicate that some oxygen-aglycon bond cleavages (around 10%)(B-B' in Fig. 18) occurred, in addition to the predominant glycosyl-oxygen bond cleavage (A-A'). The oxygen-aglycon cleavage was recently reported to be dominant for a conformationally rigid cellulose model, a 4,6-O-benzylidene of 1,5-anhydrocellobiitol [269], and was suggested to involve an initial ring-opening mechanism. Thus, the nature on the alkaline cleavages of rigid cellulosemoleculesneedsto be further clari fied. h. Cotltrolling Fuctors. The alkaline cleavage of glycosidiclinkages, like peeling, is affected by their saccharide compositions and chemical environments. Saccharide Variation. Consistentwith the proposedmechanism, the alkaline hydrolysis reactions are significantly influenced by the type of glycosides, as illustrated in

CH2OH

e CH20H o -A " N

OH

H

CH2-0

CH2OH

B

B

0

K

"

OH

0

R

k

-OR

0-

-+

OH

___)

H OH

-

H

0

H

OH

Lai

474

Table 9 for a variety of methyl pyranosides pertinent to wood polysaccharides. The reactions (conducted in 2.5 M NaOH at 170°C) were reported mainly by Janson and Lindberg [270]. All the relative hydrolysis rates were based on methyl a-D-glucopyranoside being unity. It is evident that the P-glycosides of D-glucose, D-galactose, and D-xylose residues were substantially more reactive than the respective a-anomers, whereas the reverse was true for D-mannoside and L-arabinoside. These variations, however, are consistent with the contention that glycosides with the aglyconic and the C2 hydroxyl groups being in a trans position react faster than the corresponding cis anomer [270]. Also, high reactivity was observed for a glucuronoside under alkaline conditions [271]. In wood polysaccharides, the bulk of glycosidic linkages are present in a P-pyranoside form except for the galactose, arabinose, and glucuronic acid residues, which are in a-form. Also, arabinose units, unlike others, are present as a furanosidic structure. Thus, in a homogeneous system, the relative reactivity of various glycosidic linkages present in the wood would be expected to increase in the order of saccharide moieties: galactose (1) < mannose ( 1 . 1 ) < glucose (2.5) < xylose (5.8) < arabinose (32) < glucuronic acid (280). This order ofreactivity is distinctly different from that of acid hydrolysis (Table 2), notably for the high stability of the galactose side groups and the high reactivity of the glucuronosides [271] under alkaline conditions. Alkali Concentration. Hydroxyl-ionconcentrationwasshown to significantly influence the hydrolysis rate for a variety of simple glycosides [260,265,267,272,273]. The alkaline cleavage reaction as illustrated in Fig. 18 [260,272] can be described kinetically i n terms of proceeding via anionic intermediates according to Eqs. ( I ) and (2): GlcOR

+ OH

K

= H,O

+ GlcOR

K1

GlcORk,h*

+ products

k,K[OH-] = 1 + Kl0H-I

where GlcOR is a glycoside, GlcOR- is an anionic intermediate such as (61) (Fig. 18), K

TABLE 9 Relative Alkaline Hydrolysis Rates in 10% NaOH at 170°C

of Methyl Pyranosides Relative rate

Methyl pyranosides of

P-Anomer a-Anomcr

Chemical

Degradation 475

is an equilibrium constant between a neutral and an ionized glycoside, and k , is a specific rate constant in the conversion of anionic intermediates into products. As confirmed experimentally for simple glycosides [272], the influence of hydroxylion concentration on the overall reaction rate (kobr)can be described by Eq. (3). Accordingly, the hydrolysis rate increased initially with the hydroxyl-ion concentration but levels off to a constant value at higher concentrations. Temperature. The activation energy of alkaline hydrolysis for various neutral glucosides conducted in 10% NaOH was very similar (36-38 kcal/mol) [260]. These values are substantially higher than that of the peeling reaction (21.2-24.6 kcal/mol) [229,235].

3. Cellulose The alkaline degradation of cellulose is often encountered in alkaline pulping as well as in hot-alkali refinement of dissolving pulps. a. OverallProcess. Figure 19 illustrates the overall degradationprocess of cellulose in alkali at elevated temperatures (150- 180°C) [237,274,275]. The existing reducing end groups will rapidly initiate the endwise degradation (peeling) ( k , ) .This primary peeling process, however, is generally negligible for cellulosebecause of its high DP (around 10,000). The alkaline hydrolysis of glycosidic linkages (k,,) generates new reducing end groups which give rise to a similar peeling process (k?), resulting in a loss of about 65 glucose units [274,275]. Both the peeling and glycosidic cleavage reactions of cellulose, like a homogeneous reaction of simple glycosides [272] and amylose [229], have been shown to conform with pseudo-first-order kinetics [235,274,275]. The peeling process was relatively rapid in the 150-190°C range and roughly by a factor 10' times faster than the alkaline hydrolysis reaction, according to Lai and Sarkanen's estimation [275]. Although the alkaline hydrolysis of cellulose is a relatively slow reaction resulting in small weight loss, its impact on pulp viscosity in alkaline pulping is appreciable. The activation energy for the alkaline

I

k , ( Rapid )

GG-G-G-G-G-G-G-G,y

1

l

"G-G-Gs

G : Anhydroglucose units

Gr: Reducing end groups

+

R,

k h ( Slow )

k2 ( Rapid )

+ X, 2 n:Acidicdegradationproducts CS: Stable end groups

FIGURE 19 Alkaline degradation of cellulose at elevated temperature. (From Ref. 275.)

Lai

476

hydrolysis of cotton cellulose (36 kcal/mol) [275] was similar to those reported for simple glycosides 1260,2761. A substantially higher activation energy (43 kcal/mol) has been reported for the cleavage of wood cellulose in kraft pulping of sprucewood as measured by a reduction in pulp viscosity [277]. h. Physical Structure. The overall alkalinedegradation of cellulose is profoundly influenced by its morphological structure, which has a significant impact on the physical accessibility and possibly the reaction mechanism as well. As noted earlier, the supramolecular structure of cellulose exerts a dominating influence on termination of the peeling process, as the degrading ends may be stabilized when they reach regions inaccessible to alkali. At low temperature, the majority of degrading chains terminate with a normal reducing end group at the crystalline-amorphous transition region. The observed, approximately constant, degradable chain lengths in native cotton cellulose, as Lai and Sarkanen suggested 12751, must be attributed to the submicroscopic structure of microfibrils or more specifically to the average length of accessible segments of the cellulose molecules. For mercerized cellulose,the degradable chain length was lower ascompared to native cellulose (40 versus 68), and this may beascribedtoashorter accessible segment of the mercerized cellulose molecules. After alkaline degradation, the residue from mercerized cellulose had a higher content of stable acidic end groups than that of native cellulose 1237,2381. Thus, it appears that the fringe area of ordered regions in the cellulose I1 type are inaccessible to the peeling reaction but conversionto stable acidicendgroups may occur, while in corresponding areas of the native cellulose both reactions are impeded. Also, both alkaline peeling and chemical stopping reactions occurred more rapidly in the amorphous region of a regenerated hydrocellulose than in the disordered regions of a fibrous hydrocellulose 1141. In glycosidic cleavage reactions conducted in the range 150- 190"C, the hydrolysis constant for mercerized cellulose was approximately 70% higher than that of native cellulose [275].This suggests that the number of accessible glycosidic linkages in mercerized cellulose is higher by a factor of 1.7 than in native cellulose. In addition, Gentile et al. 1141 detected the alkaline hydrolysis reaction for amorphous hydrocellulose under relatively mild conditions ( 1 M NaOH, 60-80°C). They suggested that the disordered regions associated with cellulose I and I1 polymorphs had different degrees of structural order and reactivity. Thus, molecular conformation, along with molecular mobility and accessibility, appears to influence the alkaline susceptibility of glycosidic linkages.

4.

Hemicelluloses

Hemicelluloses in alkali are susceptible to both physical changes and chemical reactions includingswelling,dissolution, saponification, reprecipitation. peeling, and glycosidic cleavage reactions [4].Alkali-induceddeacctylationandhydrolysis of the uronicacid group of xylan proceed readily under alkaline pulping conditions, and contribute significantly to its redeposition or adsorption onto the fibers. Consistent with the simple glycoside reaction (Table 9), the galactose side chain in galactoglucomannans is fairly resistant to alkaline hydrolysis. Glucomannans were generallyless stable than wood xylansin alk,A I'me pulping 141, although this trend was not reflected in the reactivity of model glycosides (Table 9). U . Xyltrrls. Thepeelingprocess of xylans is basically similar to thatof cellulose (Fig. 16), but it can be retarded chemically by their unique structures. Extended treatment of a rye-four arabinoxylan with dilute sodium hydroxide at room temperatures resulted in only a 29% reduction i n tnolecular mass 12781. The peeled xylose units were converted

Chemical

Degradation 477

to qdo-isosaccharinic acid plus several other acids. Chemical stopping was also shown to involve the formation of xylo-metasaccharinic acid end group. As anticipated, the polysaccharide after reductionwithsodiumborohydridewascompletely stable to alkaline peeling. The moderate alkaline stability of xylans can be partly attributed to their unique endgroup arrangements, and the presence of 4-0-methyl-glucuronic acid groups. Johansson and Samuelson [279,280] showed that the peeling of xylan molecules was retarded when a galactouronic end group was substituted at the C2 position with a rhamnose unit, or a xylose end group containing a 4-0-methyl-glucuronic acid unit at the C2 position. The former rhamnose unit, stable in dilute alkali at low temperatures, was labile to elimination at moderate temperature (95°C). Similarly, the retardation effect of 4-0-methylglucuronic acid groups was appreciably reduced at elevated temperatures, because these groups are unstable under alkaline pulping conditions [238,281-2851 and maybe degraded according to the scheme shown in Fig. 20 [279,282,283]. It involves a series of @-elimination reactions, typical for uronic acidcontaining carbohydrates [286]. The key step is a base-catalyzed demethoxylation of (64) to give the hexene-uronic intermediate (65), which after another p-elimination releases the anticipated acidic moiety (67). The 4-deoxyhex-4-enuronic acid group (65) was recently reported to be present in pine kraft pulps[287].Thisproposedmechanism is consistentwiththedegradation of glucuronic acid groups in xylan as a reaction dependent on alkali concentration [283]. Softwood xylans, in addition to having a higher content of 4-0-methyl-glucuronic acid groups than hardwood xylans, contain some arabinose units attached to the C3 position of xylose residues. The latter substitution pattern also induces the chemical stopping reactionand certainly contributestohigher alkali stability of softwoodxylans [4,284J. Similarly, the beneficial effects of arabinose groups are probably diminished at elevated pulping temperatures, because they are extensively removed from the xylans [239,284,288J. Regarding the alkaline hydrolysis of xylans, the relative reactivity of glycosidic linkages, based on methyl glycosides reactions (Table 9), decreases in the order glucuronoside (280) < arabinofuranoside (32) < xyloside (5.8). The high reactivity ofglucuronosidic linkages was attributed largely to the presence of a carboxylic function, which is capable of initiating a series of p-elimination and transformation reactions similar to those shown in Fig. 20.

Lai

470

b. Glucomannans. Glucomannans are less stable thanxylansandundergoamore extensive endwise depolymerization process [239,289-2911, partly because of the lack of substituents at the C2 or C3 position, which would somewhat retard the peeling reaction. The extent of glucomannan degradation at 100°C varied withthe nature of polysaccharides [291], being higher for pine galactoglucomannans (57%) than for spruce glucomannans (47%). The peeling process of isolated glucomannans was similarto that of hemicelluloses in the wood [292], and of amylose [229], with activation energy being in the range 20.224.5 kcal/mol. Similarly, alkaline treatments of glucomannans at elevated temperatures resulted in extensivedegradation[288,290]. The galactosesidegroup in galactoglucomannans,as indicated in the reactions of methyl glycosides (Table 9), was fairly resistant to alkaline hydrolysis [284,288]. Niemela and Sjostrom [293] identified about 30 hydroxy-monocarboxylic acids from alkali treatments of mannan, and their compositions were influenced by the presence of AQ additives.

B.

Lignins

The alkaline degradation of lignins, like that under acidic conditions, involves mainly the cleavage of a- and P-aryl ether linkages, which also may be accompanied by condensation reactions. The nature of these chemical reactions, derived mostly from lignin model compound reactions, hasbeenextensivelyreviewed, [ I -3,21,34,35.146], but it is still not entirely clarified in the degradation of wood lignin, especially with regard to condensation processes.

1. General Aspects Figure 21 illustrates a general scheme for the alkaline degradation of hydrolyzable ether units, which are present in both phenolic (68) and etherified (75) types. The reaction of phenolic units is initiated by a phenoxide ion (69) to yield a quinonemethide intermediate (QM) (70) with elimination of the P-ether unit as R,OH. The QM may participate in several reactions depending on the alkali environment. In soda cook, it undergoes mainly a carbon-carbon bond cleavage of the P - y linkage to yield formaldehyde and C&-enol ether (74) (reaction C). The latter enol ether may undergo further degradation slowly [294,29S]. Thus. soda cooking of phenolic P-aryl ether units resulted in only limited ether cleavages (reaction A) 1296,3021. In kraft liquor, the QM reacts readily with hydrosulfide ion and the resulting adduct (71) then undergoes an intramolecular displacement process leading to the P-ether cleavage (reaction B). On the other hand, alkaline cleavage of the etherified P-aryl ether unit (75) proceeds directly through an intramolecular displacement mechanism (reaction D). Under kraft cooking conditions, the P-aryl ether cleavage of phenolic type could be 12 to SO times faster t h a n that ofan etherified type, depending on the hydroxide- and sultide-ion concentrations (3031. Lignin condensation reactions may include the formation a-S-diphenylmethane (DPM) (79) and a-carbohydrate ether linkage (81), derived possibly from the quinonerncthide (70) (2 1,34.304-308] or the epoxide (77) (309.3101 intermediate. Formaldehyde release from reaction C may participate in the formation of a 5.5’-DPM u n i t (80) [3OXl, while coniferyl alcoholproducedfrom reaction B may involve in the formation o f a P-y-linked condensed unit (82) (306,3071. The overall degr;ldation of ether units. asrevealed by lignin model reactions, is profoundly influenced by their chemical structures and reaction conditions.

479

Chemical Degradation

CH2OH I CH CH

CH3

-G ____)

OCH,

OH 72

71

t SH-

73

HCHO OCH,

o,,

OCH,

0

OH

74

70

Condensation Products

)$4

Go-CH

+

H,CO

CH-OG

II

OCH,

0-

OH

80

79 OCH, CHzOH

t OCH, OCH,

OH

81

75

lI

H ' OCH, CHzOH kH- O 6HO-

CHzOH &H

D -G

Q -

+ X I

OH-

OCH,

om,

76

FIGURE 21

6 CH2OH

o:hH

b CH3

$ OCH, OCH,

om,

77

Generalscheme for thealkalinedegradation of

U-

andP-etherunits.

78

Lai

480

2. a-Ether The alkaline cleavage of phenolic wether structures (68) is generallyaccepted as proceeding through the formation of quinonemethide intermediates (70),and is significantly affected by the nature of the ether group. This cleavage reaction occurs quite readily for an a-aryl ether unit under mild alkali treatment ( I M NaOH at 25°C) [31 I], whereas the wether linkage of a lignin-carbohydrate model was shown to be stable under the same conditions [ 1351. Concerning the reactivity of a-aryl ethers of etherified units, they were generally shown to be stable in alkali even at elevated temperatures ( 2 M NaOH at 170°C) [312], although some a-aryl ether models were recently reported to be hydrolyzed slowly [3 131.

3. P-Aryl Ether The alkaline cleavage of P-aryl ether linkages generally proceeds through an intramolecular displacement (by a neighboring-group participation) process, with distinctly different mechanisms for the phenolic and the etherified types. The reactivity varies appreciably, with the P-etherconformationbeinghigher for the erythrothan for the threoisomer 1314,3 151. The nature of the phenyl units also has a noticeable effect on the degradation process, being especially facilitated by the syringyl nuclei [316-3181. a. PhenolicUnit Mechanism. The dominant reaction of phenolic P-ether models in soda liquor, as noted earlier, is the formation of the C6C,-enol ether (74) (Table 10) resulting from elimination of the y-carbinol group as formaldehyde. Theether cleavage, being a minor process (12-33%), is generallythought to involve an ionic mechanism[296,300,305] via qui-

TABLE 10 Effect of Alkali, Sodium Sulfide, and AQ on the Alkaline Degradation Products of Phenolic Guaiacylglycerol-P-Guaiacyl Ether (33)

AQ NazS System ~~~~~~

Soda

NaOH (M)

(%)

Temp. Enol ("C)

Time (min)

145

105 140

Guaiacol

ether"

Ref.

70 70 68

L2961

N.D.h

12981 12991 [300] L2961 12991 12981 [300] L2971 12991 12991

~

0.1 1 .0 1 .0

2.0 2.0 2.0 Kraft

(M)

0.1

0.88 0.88 0.13

140

-

0.05 0.04 0.13

Soda-AQ

1 .o

-

Kraft-AQ

2.0 0.88

-

0.04

120 I20 120 90 120 120 I20 160 200 1120 Od 1120 0"

160 I70 170 170 145 170 170 170 160 170 170

"See compound (74) in Fig. 21. hNot determined. c Percentage substrate as 1,4,11,12-1etrahydroanquinone. "As reduced AQ (AHQ).

60

12 15 13161 18 l2971 20 19 33 64 44 50 85 68 43 S2

52 N.D. 15

4 N.D. N.D.

3 25 4

481

Chemical Degradation CH20 I

OCH3

A

0 33

36

OCH3 OH 74

Q0 OCH3

CH: I CH

7H20H

I

C HI

79

FIGURE 22

83

84

Proposed alkaline cleavages of phenolic guaiacylglycerol P-guaiacyl ether.

nonemethide intermediates (reaction B) or the epoxide pathway (reaction C ) (Fig. 22). The possible involvement of a radical mechanism also has been suggested for the syringyl Pether units [3 16-3 181. The observation that the cleavage reaction was facilitated in a highalkali concentration (80% cleavages in 4 M NaOH) [296],andfor the erythroisomer [314], however, suggest that the epoxide pathway is a viable reaction. Structural Effects. The reactivity of phenolic P-ether units in soda liquor is considerably affected by their chemical characteristics, and is higher for the erythro than for the threo isomer [316]. Also, the syringyl P-ether model dimer displayed distinctly different behavior than the guaiacyl dimer, notably an unusually high ether cleavage (Table 11) [3 171, whereas the enol ether formation is largely suppressed. As illustrated in Fig. 23 for the degradation products of a syringyl dimer (85) (in 1 M NaOH at 140°C for 180 min), the yield of enol ether was quite low, being 5% [316] as compared to approximate 70% for a guaiacyldimer. The syringylP-ethercleavage reaction wasreported to be relatively unaffected by the hydroxyl-ion concentration in the range 0.1- 1 .O M [318]. Additives. The cleavage of phenolic P-ether linkages was greatly facilitated by a variety of additives, notably sodium sulfide [298-300,3 19-3221 and anthraquinone (AQ) derivatives[322,323-3251. The enhancedcleavage is usually explained in terms of a nucleophilic addition to the QM intermediate followed by an intramolecular fragmentation of the adducts (71)(Fig. 21) and (91) (Fig. 24) to cleave the P-ether group. In the case of AQ, an alternative mechanism based on electron transfer reactions also has been suggested [325,326]. Other types of additives effective in enhancing the P-ether

O

Lai

482

TABLE 11 Influence of Syringyl Units on Relative P-Ether Cleavages in Soda (0.1 M NaOH) and Kraft (0.1 M NaOH + 0.015 M Na,S) at 130°C Estimated at the Half-Life Reaction Relative P-Ether Cleavage, % P-4

Kraft

G G-l

25 52 80 14

S G-I S S-I G GI1 S GI1 S95SI1

84 92 90 90 74

6 80

S<JUKY:Ref. 317.

GG-Series

SG-Series

SS-Series OCH3

R

R !Hdo+cH3

HO-&H

H

HO-&H

OCH3 OH

GG-I: R=H GG-II: R=CH 20H

$$

HO-CH

OCH3

OH SG-I: R=H SG-II: R=CH 20H

S S 4 R=H SS-II: R=CH 20H

cleavageincludesimplephenols(xylenols)[327,328],ascorbic acid, reducingsugars [329,330], metal ion complexes of meso-tetra (p-sulfophenyl) porphyrin [33 l], anthrone derivatives [302,332], and methyl sulfide [333].The chemical effect of sodium sulfide and AQ is a subject of continued interest, because of their significance in commercial pulping [4,227,334,3351. Recent reports [336,337] indicate that the phenolic P-aryl ether cleavage was further enhancedby combined additives, e.g., anthrahydroquinone (AHQ) plus sulfide, sulfite, or alcohol. Alkali and Solvent. An enhancedalkalinecleavage of phenolicP-ethermodels wasobserved in an aqueousdimethylsulfoxide(DMSO), or a DMSO-potassium rerrbutoxide system [338]. The latter system (at 75°C for 20 min) was reportedly even more effective than a kraft liquor (NaOH + Na,S) at 170°C for 2 h (81% versus 70% of Pether cleavage). It appears that different mechanisms are involved when a DMSO-butoxide solution is used. The use of sodium hydride or sodium methoxide in DMSO was less effective. On the other hand, a methanol-sodium methoxide system selectively accelerated the enol formation. Recently, methanol addition was shown to enhance the P-ether cleavage in a sodaAHQ system, whereas it had no effect on a soda or kraft system [337]. These suggested that organic solvents are probably able to affect the redox behaviors of AHQ species, and thus improve the electron transfer reaction. Reaction Kinetics. The activation energy reported for the cleavage of several phenolic guaiacyl P-ether dimers in dilute alkali (0.1 -0.5 M) [296,314] appears to be slightly

483

Chemical Degradation OCH3 CH20H CI H - O O C H 3 &:cCH3

-

IMNaoH

140T 180 min

H3C0

OH L

86

85

H3CO$CH_

R

R = CH3 (75%)

R=H R =CH0 R =COCg

87

OH

(5%) 88

89

FIGURE 23 Alkaline degradation products of a phenolic syringylglycerol P-syringyl ether. (From Ref. 316.)

influenced by the side-chain structure (Table 12). The presence of a y-methyl group resulted in a slightly lower activation energy than that of a glycol-type unit (26-28 versus 31 kcal/mol) in a soda cook. This difference, however, was not evident in the kraft cook, and a distinctly higher value (33 kcal/mol) was reported for the threo isomer. b. Nonphenolic Units. Alkaline cleavage of nonphenolicP-etherstructuresgenerally proceeds through an intramolecular displacement process and yields the epoxide intermediate (77)(Fig. 2 I ) [312,315,3411. This reaction is facilitated by an ionized hydroxyl group at the a- (76) or the y-position.Consistent with the stereochemicalrequirement, the erythro isomer was more reactive than the corresponding threo unit by a factor of 4 [315]. Also, the cleavage reaction was not affected by the addition of sodium sulfide [339] or AQ [342], and was appreciably influenced by alkali concentration, the nature of the solvent, and structural modification. Alkali Concentration. As a base-catalyzedreaction via anionicintermediates [260], the cleavage of nonphenolic P-ethers (75) increases initially with the hydroxyl-ion concentration and levels off to a constant value at a certain alkali concentration. Limited data [315,339] indicate that the cleavage reaction is directly proportional to the hydroxylion concentration within the range of alkali used (0.1- 1 M). An estimation based on these data using Eq. (3) indicates that a linear relationship between the reaction rate and alkali concentration extends to about 4 M alkali.

404

Lai

I

+

H0 p&o -

0 36

90a

Carbohydrate Oxidation

&

91

l 0

-

QOC", 0-

1

yH20H CH 1 l

CH

+ G 00 O C H j

90

73

FIGURE 24

A proposed anthraquinone catalyzed alkaline cleavage of phenolic P-aryl ether linkages. (From Ref. 34.)

TABLE 12 Activation Energy for Alkaline Cleavages of Phenolic and Etherified Lignin

P-Aryl

Ether Model Dimers [OH-] Dimers

Structure (92)

R, = R2 = H

a

R, = H, R? = CH, Erythro

b

Threo R , = CH,, RZ= H R, = CH3, R? = H R , = R2 CH,, = CH20H

C

d e

OCH3

OR, 92

15h-1 M

M

0.5 0.5

0.08

0.1 0.2 0.1 0.2

0.5 1.o 1 .o

-

0. l -

0.1

AE kcal/mol 31.3 28.9

28.0 28.7 26.1 33.3 33.5 28.7 31.6

Ref.

H

Chemical Degradation

485

Solvent and Alkali. Limited data on the P-ether model reaction indicate that the effect of organic solvents on the alkaline cleavage reaction varies with the side-chain unit [343]. In the case of veratrylglycerol-P-guaiacylether (75), the cleavage rate was slightly reduced by using dioxane as a co-solvent, but was unaffected by methyl cellulose. Interestingly, the cleavage reaction was enhanced in a monoethanolamine (MEA)-sodium hydroxide solution 13421. For veratryglycol-P-guaiacol ether, the ether cleavage was suppressedby the addition of methyl cellosolve andespecially dioxane 13431. Also, the reaction was greatly facilitated in a DMSO-potassium tert-butoxide system[338];and64%cleavagewasobtained at 75°C in 30 min. Structural Effects. The hydrolysis of nonphenolicP-etherlinkages (75) canbe greatly improved by modification of the a- or y-hydroxyl group [328,339,344-3471 (Fig. 25). The presence of an a-carbonyl or a p-tolysulfone group was reported to give the highest enhancement, and the alkaline hydrolysis was observed even under ambient conditions [345]. A possible mechanism, as indicated in Eq. (4) (Fig. 25),may involve a series of @-elimination, a hydrosulfide-ion addition to the enone intermediate (94), and an intramolecular fragmentation of the keto thiol intermediate (95) to cleave the P-ether linkage [344]. Similarly, the P-ether bond cleavage was facilitated by an a-sulfonate [345] (98), an a-2,4-xylenol (102) 13451, or a y-p-toluenesulfinyl 13471 group. The chemical effect of a 2,4-xylenol group (102) on the P-ether cleavage may be accounted for by it being able to initiate an intramolecular displacement process [Eq. (6)] [327,328].

93

94

95

96

97

I

R 98

99

100

101

CH20H

102

FIGURE 25 Alkalinecleavagesof vetratrylglycerol-P-aryl ethercontainingan sulfonate, or a-2,4-xylenol group. (From Refs. 327, 328, 344, and 354-356.)

a-carbonyl, a-

486

Lai

Kinetics. In contrasttopureisomers,alkalinecleavage of anerythroandthreo mixture of the p-ether model did not follow simple first-order kinetics [343], because the erythro isomer was preferentially degraded. Reported activation energies for the cleavage of unmodified etherified p-ether linkages 1315,339,3401 were in the range of 29-34 kcall mol (Table 12). A significantly lower value (23 kcal/mol) was observed for a modified pether dimer containing an a-sulfonate group [346]. 4. Carbon-Carbon Bond Cleavage Alkaline cleavages of the lignin carbon-carbon linkages, like those under acidic conditions, are largely restricted to the p- andy-carbons of phenolic units, resulting in the formation of formaldehyde and enol ether as indicated in reaction C (Fig. 21) for a p-04 dimer. A similar degradation of the phenolic p-5 (104) [348-3511, p-1 (107) [349,352], and @-p(109) [353] units yielded the stilbene-type structures (106), (108), and (111) (Fig. 26). The rate-determining step is likely the splitting off of the y-methylol group from the quinonemethide intermediates. Moreover, the overall reaction hasa relatively low activation energy (18.5 kcal/mol) [350]. Another type of carbon-carbon bond cleavage was observed for p-0-4 unit containing an a-sulfonate group (98), which upon alkaline hydrolysis yielded vanillin (101), resulting froman a-p cleavage linkage [Fig. 25, Eq. ( 5 ) ] [354-3561.Reactions of this type are important in the production of vanillin from lignosulfonates [357].

5. Condensation The nature of alkaline condensation reactions [34,307,358-3611 has been extensively studied using lignin model compounds. There are four major types: conjugate addition, formaldehyde addition, epoxide addition, and aldol reaction. One of the key factors in these reactions is the formation of carbanions, which has been reported for several aliphatic carboxylic acids and simple phenols at 170°C [362]. a. ConjugateAddition. Theadditionofnucleophilesderivedfromphenolicor enolic units to a QM (quinonemethide) or an extended QM (from coniferyl alcohol) intermediate leading to the formation of stable carbon-carbon or ether linkages has been welldemonstrated[21,34,304-3081.Thetypesofcondensed units formedinclude a-5 DPM (79), a-carbohydrate ether (81) [363], and p-y linkages from dimerization of coniferyl alcohol (82) [306,307] (Fig. 21). Also, the formationof a - l (114) and a-a (116) units (Fig. 27) was reported after alkaline treatment of a ðer dimer (112) (in 0.5 N NaOH at 140°C) [305]. The participation of an enediol intermediate arising from the reducing end groups of carbohydrates has also been suggested [300]. These reactions clearly indicate that a wide variety of carbanions may be generated from lignin- and carbohydrate-derived products and are involved in the condensation process. The conjugate-type condensation was reported to be suppressed in the presence of additives, e.g., AHQ [364,365], and could be inhibited by steric factors in a two-phase system [366]. b. Formaldehyde Addition. The alkali-induced addition of formaldehyde to simple phenols occurs mainly at the C5 position via a Lederer-Manasse reaction, leading to the formationof5,5’-DPM (80) (Fig. 21).Thiscondensation reaction is wellestablished [308,367,368], and may involve formaldehyde generated in situ from the reaction of lignin and related dimers [353,358,369,370]. Also, the formaldehyde-type condensation was retarded in the presence of sulfide or AQ [371].

Chemical Degradation

0 X

0

lo

zQ

v)

z

8 F

0

U -U -U

0

+

8X

0

+

487

488

Lai

OH 112

113

114

I

OH 115

116

G=

+OH

OCH,

FIGURE 27 An example of the formation of a-l and a-a' condensed units. (From Ref. 305.)

c. Epoxide Addition. It has been shown that alkaline treatment of an etherified pether dimer (117) in the presence of methyl @-D glucoside yielded the formation of a benzyl ether linked to a glucosyl unit (119) 13091, indicating the involvement of an epoxide intermediate (118) (Fig. 28). The extent of this condensation was alsoreduced in the presence of sulfide or AQ additives. d.AldolCondensation. Figure29 illustrates twotypes of aldol condensations, whichweredetectedunderalkalinepulpingconditions.Equation (7) showsa reaction between two modified side-chain units [349], whereas Eq. (8) may represent a possible reaction between two partially degraded carbohydrate and lignin molecules [372].

-OH,A v

- QOCH3 OCH,

-

OCH, ( + isomers)

117

118

119

FIGURE 28 Formation of lignin-carbohydratelinkageduringthecleavage of etherified p-aryl ether unit in the presence of methyl p-D-glucopyranoside. a: R = H; b: R = CH,OH (erythro- and threo- forms). (From Ref. 309.)

Chemical Degradation

489

120

122

123

124

FIGURE 29 Examples of aldol condensationreactionsof products. (From Refs. 349 and 372.)

C.

lignin andcarbohydratedegradation

AlkalineDelignification

The kraft process, using mainly sodium hydroxide (-0.8 M) plus sodium sulfide (-0.2 M) at elevatedtemperatures ( 160- 180°C) for 1-2h, is the mostwidelyused in the production of chemical pulps for papermaking and cellulose applications [4,227,358]. Unfortunately, suchalkalineconditionsrequired for delignification are alsoconductiveto polysaccharide degradation.

1. Overall Process The kraft delignification process is known to consist of three distinct stages: the initial, bulk, and residual phases [373-377). They differ considerably in delignification selectivity, as illustrated in Fig. 30 for a relative change of polysaccharide and lignin content during a kraft cooking of spruce and beech wood [378]. The residual phase, associated with the removal of the last 10% of the lignin, is accompanied by severe cellulose degradation. In

e Beech

P

OSpruce

40

I

0

l , l , l , l , l , L IO 15 20 2 5 30

5

Residual Ltgnin (%of Wood)

FIGURE 30 Relative changes of the polysaccharide and lignin content showing theinitial (ID), bulk (BD), and residual (RD) phases of kraft delignification process. (From Ref. 378.)

Lai

490

commercial practice, pulping is usually terminated at the transition between the bulk and residual phases to maintain fiber quality. Table 13 illustrates the composition of typical kraft pulps from pine and birch as compared to the original wood [ 5 ] . The yield of lignin-free pulp is quite low, e.g., 44% compared to a theoretical 67% for pine. The loss in pulp yield is largely a consequence of the alkaline degradation of polysaccharides (see Section IV.A), being higher for glucomannan (70-80%) than for xylan (40-50%). The complex mixture of nonvolatile hydroxyacids plus formic and acetic acid produced consumed the major portion of alkali charge (around 63%) in kraft pulping [4]. Improved methods for enhanced delignification and carbohydrate stabilization are still a technical challenge for the pulp and paperindustry. a. Initial Phase. The initial delignification occursduring the heating-upperiodto about 15OoC, and is generally lower for softwood (l8-24%) [288,292,380,381] than for hardwood (25-30%) [373,380,382]. The carbohydrate loss, however, is severe, amounting to about 50% and 60% of the total yield loss encountered in the kraft pulping of softwood and hardwood. respectively. Delignification. In softwood, the initial delignification (ID) was further shownto consist of two distinguishable subphases: rapid (ID,) and slow (ID,) [292,380]. In the ID, subphase, the delignification reaction was affected by the chipmoisturecontent[375]. However, the reaction was independent of chemical concentrations within the ranges 0.21.5 M for NaOH and 0.1-0.4 M for NazS 13751, and has a low activation energy ( I O - 12 kcal/mol)[292,375];whereas the reaction in the ID, subphasedependedon the alkali concentration [380] and has a slightly higher activation energy (17.5 kcal/mol) [292]. In the case of hardwood(aspen), the IDz subphasewas not discernible, and the ID was practically unaffected by increasing alkali concentrationabove the 0.1 M level 13801. Although the ID reaction is generally thought of as being a diffusion-controlled process [375], the proposedcleavages of some reactive a-arylether bonds[292,303,359]was supported by a recent report [379]. Carbohydrates Loss. The pulp yield loss encountered in the IDstage is largely limited to the hemicellulose components, and is noticeably higher and faster for hardwood and for softwood [380]. Also, in the case of softwood, two subphases were discernible for carbohydrate loss in the ID stage [292,380]. These variations may be attributed to the loss of xylan and glucomannans, caused mainly by the physical dissolution and the peeling process, respectively. The activation energy for the latter reactions is likely in the range 21.2-24.5 kcal/mol, as observed for amylose [229] and isolated glucomannan 12911.

TABLE 13 Typical Compositions of Softwood and Hardwood Kraft Pulp Percent of original wood Component

34 0.5

Pulp yield Lignin Cellulose Glucomannan Xylan Pitch Other Source: Ref. 5 .

Birch

I 00

27 40

39

17 8

3

4 5

47 3 35 4 5 0.5 -

20

53 2

3 30

1 16

100

4

-

Chemical

Degradation 491

b. Bulk Phase. The bulk delignification (BD), which occurs at temperatures above 150°C andremovesan additional 70% of the lignin, involvesmainly the cleavage of etherified P-aryl ether units [303,359]. It was recently shown [379] to be closely related to the aryl ether cleavage reactions. The delignification process follows pseudo-first-order kinetics [373,376,383-3861, and was shown to have activation energies in the range 3235 kcal/mol [383-3861, similarto that of lignin modeldimers [315,339]. In addition, preferential cleavage of the more reactive erythro isomer (Section IV.B.3.b) was observed in the early BD phase [387,388]. This is consistent with the reactivity of different isomeric P-aryl ether dimers [3 151. Thecarbohydrate loss during the bulkphase is causedmainly by the secondary peeling resulting from the alkaline cleavage of glycoside linkages including cellulose (Section 1V.A). c. Residual Phase. The residual phase, which begins at about 90% delignification, is characterized by a slow delignification coupled with a rapid carbohydrate-degradation reaction (Fig. 30). The chemical nature of residual lignin causing its resistance to degradation is still not entirely clear, although it is known to contain few intact P - 0 - 4 ether structures [389], a highproportion of enol ether units [389-3911, and a highphenolic hydroxyl group content [392]. The residual delignification was reportedly influenced appreciably only by the hydroxide ion concentration and pulping temperature, and had a higher activation energy than the bulk phase (30 versus 35 kcal/mol) [393]. Also, it was relatively unaffected by the sulfide concentration [376,393] and the aryl ether cleavage reaction [379]. These observations are consistent with a contention that the cleavage of carbon-carbon linkages is important to the residual-phase delignification [359].

2. Aryl-Ether Cleavages The significance of aryl-ether cleavages in alkaline delignification has been well recognized [303,359], and is clearly illustrated in a gradual increase of phenolic hydroxyl groups in both dissolved [394,395] and residual pulp [379,394] lignin. Also, the dissolved lignin has a considerablyhighercontent of this functionalgroupthanthe residual lignin [394,396-3981.

LaiandKuo [379] recently reported that the impact of aryl ethercleavages on alkaline delignification, like delignification selectivity (Fig. 30), displayed three discernible phases as shown in Fig. 31 for a kraft cook of Norway spruce. As indicated, initial delignification was associated with the hydrolysis of few very reactive aryl ether units, whereas the bulk delignification extending to an approximate 90-95% level was correlated directly to the aryl ether cleavage reaction. On the other hand, the ether hydrolysis has relatively little impact on the residual delignification.

3. Role of Lignin Condensation Despite extensive efforts,the significance of lignin condensation in alkaline delignification is still not entirely understood. One of the major difficulties encountered in residual lignin analysis is the lack of a selective technique capable of revealing different types of condensed units, especially the diphenylmethane (DPM) types suspected of being generated during the pulping stage. Because of the limitation in analytical techniques, studies using different procedures often have led to inconsistent conclusions. Lignin analysis based on the permanganate oxidation technique [398-4001 generally indicates that lignin-lignin condensation does not appear to be significant in kraft pulping.

492

S

80

U) .-(

W

Q

20

:

Lai

c

150°C 160°C

+

170°C

-

0

0

/

I

100

Phenolic hydroxyl

I

200

I

I

300

I

I 400

g r o u p , mmo1/100 g lignin

FIGURE 31 Influence of temperature in kraft pulping of Norway spruce wood on percent delignification as related to the phenolic hydroxyl group content of pulp residual lignin. (From Ref. 379.)

By contrast, an opposite conclusion was obtained froma combined nitrobenzene oxidation and phenyl nucleus exchange analysis [401,402]. The latter technique, although its validity in analyzing the DPM condensed units has been questioned [403], seems to indicate that the residual lignin contains an appreciable amount of unknown condensed structures. Available data suggest that the residual lignin has a rather high content of condensed structure, and its characteristics cannot be accommodated by a mere accumulation of those less reactive condensed units in the wood lignin. Interestingly, a recent "C NMR study was able to detect the presence of 5,5'-DPM units in the residual lignin isolated from kraft pulps [404]. The reported content per aromatic unit increased from 0.11 to 0.17 as the pulp kappa number went from 28 to 13.

4. Role of Lignin-Carbohydrate Complex There is considerable indirect evidence to support the presence of lignin-carbohydrate (L-C) linkages in kraft pulps [396,397,405-4091. These alkali-stable (L-C) ether linkages are probably of an etherified type. They may originate in the wood lignin (Section II.C.2) or formed during the pulping stage (Section IV.B.5). Figure 32 illustrates that an etherified benzyl ether linkage (125) is hydrolyzable only after a prior cleavage of the p-0-4 ether bond. In the latter process, the intramolecular displacement reaction (A), assisted by a yhydroxideion(Section IV.B.3.b), is likely to proceed very slowly[315]. Similarly, the cleavage of an a-carbohydrate bond (reaction B) is expected to be a slow process as well. Additionally, some isolated LCCs were reported to have a strong tendency to form micelles or aggregates in aqueous solutions [33]. Thus, the presence of alkali-stable L-C bondswould certainly increase the degree of cross-linking of lignin polymers, and is expected to have a significant influence on their reactivity and solubility [410,41 l].

493

Chemical Degradation CH20I CH-OR, I CH-OR,

o-CH2 \AH I CH-OR,

CH20H -0AH I CH-OR,

CH20H I 0-CH \&H

B H,CO

H3C0 OCH3

H3C0

125

RI=Lignin units, R2 =Carbohydrate units

FIGURE32 A proposed alkaline cleavage of the lignin-carbohydrate ether linkageof an etherified

type.

V.OXIDATIVE

DEGRADATION REACTIONS

The oxidativedegradationofpolysaccharidesand lignin is usuallyencounteredduring pulp bleaching processes [4]. The detailed chemistry of bleaching reactions, based largely on model compound experiments, has been extensively summarized [35,361,412], notably in a recentcomprehensivemonograph[413].Theoxidative delignification reactions in general are complex and nonselective. In this section, only the major types of oxidative reactions will be briefly illustrated.

A.

Polysaccharides

The oxidation of polysaccharides, besides a glycol cleavage, may be discussed in terms of reactions at the reducing end group, at the hydroxyl group, and at the anomeric (acetal) position, and as illustrated in Fig. 33. The tendency of these oxidations depends considerably on the nature of the oxidants and the hydroxyl configuration.

1. Oxidation Types a.ReducingEndGroups. The reducingendgroups of polysaccharides are very reactive and can be readily oxidized to a gluconic acid moiety by a variety of agents [ 161, including bromine and hypohalites. In alkaline solutions, oxygen [414-4181 and AQ (anthraquinone) [419-4211 degrade the cellulose end groups to a mixture of gluconic (129), mannonic, arabinonic, and erythronic acid moieties. These aldonic end groups are generally resistant to the peeling reaction, and thus reduce the yield loss in kraft pulping [335] and in O2 bleaching [422]. h. Glycol-Cleavage Oxidation. Oxidants such as periodate and lead tetraacetate are effective in the cleavage of a,@-diol units, giving the dialdehyde derivative (131). This oxidation isvery specific and has been used extensively in both structural analysis and modification of polysaccharides [3,6,423]. A cis diol unit is generally more reactive than a trans diol. Thus, the cleavage of a cis diol in the mannose residues proceeds faster than that of a trans diol in the glucose or xylose residues. c. Hydroxyl Oxidation. The selectivity of hydroxyl oxidation, as shown for simple glycosides, varies significantly with the type of oxidants and substrates. Oxidant. Platinum black was shown having a high selectivity in oxidation of the 6-OHgroup of methyl a- and P-D-glucopyranosides to the correspondinguronic acid

494

Lai

CH,OH

CH20H A

R 2 K C 0 2 H

OH

R2

OH

128

129

131

c _R2i

o OH R 1 132

"e"' OH

R2

133

CH20H @"l

__I

1

R2

OH 130

135

l CH20H E

-

1

QR'

R2

137

CH,OH 4

CO,H OH 128

129

FIGURE 33 Major types of cellulose oxidation.

Further Degration

Chemical Degradation

495

derivatives, which were obtained at 87% and 68% yields, respectively [424]. A slightly lower preference was observed upon oxidation with nitrogen dioxide, which also gave the 2-keto, 3-keto, and 4-keto derivatives [425]. Reaction with chromic acid [426] or Fentons's reagent (Fe" plus H20,) resulted in rather nonspecific oxidations [427]. On the other hand, the initial oxidation of methyl P-D-glycoside in oxygen-alkali [428-4311 or peroxide-alkali [429] solutions occurred mainly at the 2-OH or 3-OH group. The resulting ketoglucosides could be further oxidized to the 2,3-dicarbonyl intermediate (133, leading to the formation of methyl 2-carboxy-P-D-pentofuranoside derivative (136). Different Glycosides. For 0,-alkali oxidation (in 0.5 MNaOH at 120°C for30 h), the reactivity of methyl glycosides (as percent oxidation) increased in the order pxyloside (12%) < a-glucoside (18%) < p-glucoside (20%) < a-mannoside (28%) [431]. Also, 1,5-anhydroribitol was more reactive than 1,5-anhydroxylitol [432], whereas higher reactivity was observed for methyl P-D-ribopyranoside than for methyl p-D-xylopyranoside [20]. Thus, glycosides containing an axial hydroxyl group seem to promote the oxidative degradation process. Also, the reactivity of methyl glycosides toward an aqueous bromine oxidation [433] maybe related to their hydroxylconformation.Asindicated in Table 14, methyl a-Dglucopyranoside yielded mainly the 2-keto and 4-keto derivatives, whereas oxidation of the corresponding@-glucosidewas less specific and resulted in additional formation of the 3-keto compound. The a-anomerof manno- and galactopyranoside, containing anaxial hydroxyl group at the C2 and C4 positions, respectively, gave mainly the corresponding 2-keto and 4-keto products. The p-anomer of these two glycosides also gave a noticeable amount of the 3-keto derivative. Steric factors were assumed to be a contributing factor in determining the specificity of the hydroxyl oxidation. cl. Anomer-ic Site. Anotheroxidationtype is direct attack ontheanomericcarbon (Cl), resulting in glycosidic cleavages and the formation of a gluconic acid moiety. These oxidative degradations occurred in chlorination [434,435], and were reported to be a dominant mechanism in ozonation [436-4401.

2. Cellulose Controlled oxidation of cellulose provides a means to prepare useful cellulose derivatives [423,441], but it is generally difficult to achieve good selectivity. Most of the cellulose

TABLE 14

Brominc Oxidation of Methyl Pyranosides a t pH 7" Yield of keto ('h)

Methyl glycoside Type Gluoside Mannoside Galactoside

Anomer

c,

C1

C,

CY

17

0

P

C)

26 17 -

35 24

P

II 29 24

-

-

cy

-

31

34

P

-

IC)

36

CY

-

6 IO

material Unreacted

70

44

496

Lai

degradations encountered during pulp bleaching [412,413] are very undesirable, resulting in the loss of fiber strength. a. Controlled Oxidation. A classical example is in the aging of viscose to obtain a proper pulp viscosity for making cellulose xanthate and other products [4,442]. An alkali0, system is normally used to induce the oxidative cleavage of glycosidic linkages. Other controlled oxidations are related to the preparation of oxycellulose. Glycol Cleavage. Oxidation with periodic acid selectively converts cellulose to the 2,3-dialdehyde derivatives (131)(reaction B of Fig. 33), which are useful as intermediates for preparing nitrogen-containing products [443]. The initial periodate oxidation of cellulose, like other chemical reactions, is largely limited to the readily accessible component, and has been used to indicate the accessibility of cellulose substrates [444,445]. Hydroxyl Oxidation. Amongvariousoxidants,nitrogendioxidewas relatively more selective in oxidation of the 6-OH group to yield an oxycellulose containing 87.5% of the glucuronic acid moiety and about 6% of the 2-keto and 3-keto groups [446]. Oxidation with DMSO-acetic anhydride (AC,O) gave a good yield of the C6-aldehyde (4762%) and a significant amount of the C2- or C3-keto groups [447]. However, oxidation with DMSO-AC20 in a DMSO-PF solvent system gave selective oxidation at the 3-OH group [448]. This was suggested as being attributed to a reversible formation of hydroxymethyl and polyoxymethyleneol groups at the C2 and C6 positions. b. UncontrolledOxidation. Oxidativedepolymerizationofcelluloseandhemicellulase is the most undesirable side reaction encountered in pulp bleaching [412,413], notably with alkali-oxygen, peroxide, and ozone. It involves mainly an oxidative cleavage of glycosidic linkages resulting in a reduction of pulp viscosity, which may lead to a loss of fiber strength. The cleavage reaction may result from direct oxidation (reaction F, Fig. 3 3 ) , e.g., with ozone, or may be promoted by a ketol group present in the polymer. Ketols formed at the C2, C3, or C6 position are known to readily undergo an alkali-induced palkoxy elimination reaction [422,449-45 I]. Figure 34 illustrates one of the most common depolymerization reactions facilitated by a C2-keto group. Such a process, like a peeling reaction, can proceed under mild alkali conditions. For 0, and peroxide bleaching conducted in an alkali medium, the ketol formation and the p-elimination process can occur in a successive manner. For bleaching in an acidicmedium, the keto group formed also initiates a similar p-elimination process during a subsequent alkaline stage.

@oeoR -Re

CH2OH

CH2OH

CH2OH

R

.OH

OH

OH

0 138

128

+

@OR OH

OH 139

Chemical Degradation

497

Acidic Conditions. Common bleaching agents used under acidic conditions [4 12,4131 include chlorine, chlorine dioxide, hydrogen peroxide, and ozone. Among these oxidants, chlorine dioxide is the most selective delignification agent and reacts only slowly with polysaccharides, whereas peroxide and ozone are rather nonselective and can degrade cellulose quite extensively depending on reaction conditions. Initial degradation includes oxidation of the 2-OH, 3-OH, or 6-OH group to a carbonyl or carboxyl group (Fig. 33). The ketol-containing glycosidic linkages, as noted earlier, are very labile and easily cleaved when exposed to alkali media according to a P-alkoxy elimination mechanism. Also, direct oxidative cleavage of the glycosidic linkage (reaction F, Fig. 33) is possible, especially i n ozonation. The oxidative degradation reaction has been generally shown to involve a radical mechanism, and is enhanced by the presence of transition metal ions. Carves [ 1 IO] determined the influence of oxidizingagents in the degradation of cellulose dissolved i n phosphoric acid. Chlorine was found to induce some carbonyl and carboxyl group formation and had a minor effect on the cleavage of glycosidic linkages. The addition of hydrogen peroxide plus ferric sulfate considerably enhanced the glycosidic cleavage, which was shown to be associated with a very low activation energy ( 1 0.8 kc& mol). Alkaline Conditions. Commonbleachingagentsemployed under alkalineconditions include hypochlorite, peroxide-alkali, and oxygen-alkali systems. Hypochlorite oxidation 14521 is rather nonselective. and can occur at any hydroxyl group including the C l position (Fig. 33), whereasoxidation in an oxygen-alkalisystem 1453,4541 occurs preferentially at the 2-OH or 3-OH group. The resulting keto units can undergo either a P-alkoxy elimination leading to a glycosidic cleavage or further oxidation to yield a 2,3diketointermediate (135) (Fig. 33). Alkali-inducedbenzylic acid rearrangement of the latter intermediate then gives the furanosidic acid unit (136). The peroxide-alkali oxidation of cellulose is generally thought of as being similar to that of the oxygen-alkali system. The oxidative cleavage of glycosidic linkages under alkali-o? or -peroxide bleaching conditions can be considerably suppressed by the addition of additives 1455-4581, notably magnesium salts. Recently, the oxygen-alkali and cobalt-hydrogen peroxide-induced oxidation of cellulose were shown tobe significantly influenced by its morphology [459]. Amorphous cellulose was degraded more rapidly than kraft pulp or highly crystalline cellulose samples. Thus, accessibility also plays a dominant role in the oxidative degradation of cellulose. 3. Hemicellulose The oxidative degradation of hemicelluloses is expected to be similar to that of cellulose in reaction type, but it will be more extensive because of the relatively high accessibility. The oxidation of birch xylans with hypochlorite resulted in the formation of a variety acids including aldonic acid end groups, uronic acid, formic acid, and glycolic acids 14601. The degradationreactions of corn cob [ 4611 and birth [462] xylans in oxygen-alkali treatments were shown to be very similar to that of cellulose.

B. Lignin Theoxidative degradation of lignin involvesextensivecleavages of the side-chain and phenyl units, and has been discussed extensively by Gierer [35,360,361], notably on the development of a general concept. This section is limited to a brief outline of the major types of lignin oxidation and 0, delignification. These reactions, unlike delignification i n

Lai

498

pulping, are characterized by the cleavages of carbon-carbon linkages and the formation of acidic groups from ring degradations.

ReactionTypes Most of the bleaching systems are not selective and can achieve a wide range of reactions, 1.

which are outlined in Figs. 35 and 36. U . Substitution Reactions. The substitution reactions commonlyobservedwith chlorine and nitric acid C4631 include electrophilic substitution at anunsubstituted ring position (141) and at the C1 position with a side-chaindisplacement (142). The latter displacement reaction depends on the nature of the benzyl unit, and is possible with a benzyl alcohol unit. b. Side-ChainCleavages. Theoxidativecleavageof lignin side-chain units is largely confined to the a-p linkage of ring-conjugated structures containing an a-carbonyl (143) or an ethylenic (145) group. These functional units may be generated during the pulping stage, e.g., stilbene types, or may be formed in situ during the bleaching process. Both unsaturated units are very reactive under alkali-0, or peroxide bleaching conditions, whereas the ethylenic units are also rapidly degraded in ozonation. c. Ring Degradation. The oxidation of aromatic nuclei to dicarboxylic acids (Fig. 36) represents one of the most distinct features of bleaching actions, and has been observed in nearly all the common oxidants, including 03,peracetic acid, CIO?, Cl,, and alkaline0, and-peroxide. The ring openingbetween the C3-C4 bondto yield muconic acid

5CH0

0 OR 140

*

xQ

+

OH

142

I

l

CR

QOCH3

OH

141

H-C-R

143

X

A

X=CPNO?

OCH,

+

RI CHOH

RI CHOH

144

I

O=FR

499

Chemical Degradation

I

1

OH 147

OH

0

148

149

1

I B ___)

C

CO,H 150

151

v

R

P

HO

OH 154 152

153

FIGURE 36 Typical oxidative degradation of lignin phenyl units.

derivatives (151) may proceed directly (reaction B), or may be preceded by a demethoxylation (148) or an ortho-quinoneformation (149). Additionally, ring degradationwith peracetic acid may be initiated by a hydroxylation (152) followed by oxidation to maleic derivatives (154) 1464,4651. This pathway is probably also applicable to an acidic hydrogen peroxide system. 2.

0, Delignification

In the current trend of producingbleachablechemical pulps, 0, bleachinghasbecome nearly a standard process following kraft pulping, but it can only readily remove approximately 50% of the pulp residual lignin without causing excessive degradation of the tiber [466]. Despite extensive efforts, the nature of the residual lignins, especially those resistant to 0, delignitication, is still not well established. Lignin model compound reactions reveal that phenolic and enolic structures are the major reactive sites under 0, bleaching conditions [35,360,3611, andtheir chemical nature has a considerableinfluence on the reactivity 14671. The biphenyl unit is generally accepted as being relatively resistant to degradation [468]. Recent reports [469,470] indicate that the phenolic lignin DPM (diphenylmethane) model dimers are quite reactive in alkali-0, solution, resulting in the formation of monomeric and oxidative coupling products, as illustrated in Fig. 37 for an a-5 dimer (155) (4691. However, a preliminary analysis on kraft pulp lignins s e e m to suggest thatthey contain some DPM-like structures that are resistant to degradation in 0, bleaching 14711. These stable “diphenylmethune” units were suggested to be probably o f an etherified type.

500

Lai

OH

OCH,

NaOH

OCH, _,,

HOH,C

OH

OH

155

CH,

CH,O OH

156

157

i

i

OCH, OH

OH 158

159

FIGURE 37 Productsidentifiedfromthealkali-oxygentreatment ofan a-S lignin diphenylmethane model dimer (155) in 0.1 N NaOH at 55°C for 1 h. (From Ref. 469.)

VI.

TOPOCHEMISTRYOFDELlGNlFlCATlON

The topochemistry of delignification refers to the relative rates of lignin being removed from various morphological regions. It has an appreciable impact on the overall pulping and bleaching efficiency. If the lignin component in the middle lamella (ML) and cell corner (CC) can be selectivelyremoved, this will greatly facilitate the fiber separation process, and thus improve the pulping efficiency. Also, the morphological distribution of residual lignin will have a significant bearing on the bleachability and papermaking properties of pulp fibers. This subject has been studied by various groups, among which Goring and co-workers deserve special mention [472]. They reported, in kraft cooking of spruce wood [473], that the initial delignification occurred preferentially i n the secondary wall (SW); and at about 50% delignification, lignin in the ML and CC regions began to dissolve faster than the SW lignin, leaving most of the residual lignin in the SW (Fig. 38). A smaller topochemical effect was observed in sulfite cooking, whereas none was found in chlorite treatments [473]. The bulk of reported data[472,475-4781 generally support a contention that the nature and extent of topochemical effects vary significantly among different pulping systems. They are, however, not entirely consistent even for a similar process obtained by different authors. For example, Giidda [479] observed an opposite linding to that of Procter et a l . [473] in kraft cooks. The initial kraft delignification of pine wood was shown to be faster in the ML than in the SW region. The reported contrary results, although they were obtained with different species (pine versus spruce), may be largely attributed to variation in the experimental conditions used. This contention is supported by an interesting finding of Westermark and Samuelson

501

Chemical Degradation

-

I

I

I

I

I

I

I

I

I

I

I

I

-

0 20406080 0 2 0 4 0 6 0 8 0 020406080w3o DELlGNlFlCATlON OF W O D

FIGURE 38 Plot of percentlignin removedfromthemiddlelamella(ML)andsecondary (SW) on pulping of black spruce as determined microscopically. (From Ref. 474.)

wall

[480]. They observed that the ML tissue was delignified more rapidly than the whole wood by using a highliquor-to-wood ratio (40:l), whereas the oppositewasfound in a low liquor-to-wood ratio cook (4: 1). Similarly, as discussed by Chen et al., the reported topochemistry of acid sulfite and organosolv delignification [477,478] was not entirely consistent. Most reports indicate that the ML lignin was chemically attacked by the liquor at the early stage of a cook. Also, the initial removal of the ML lignin was either faster or comparable to that of the SW lignin. In contrast, based on studies of isolated tissue, the SW lignin was clearly more responsive to acidic degradation than the ML lignin. These seemingly conflicting findings can only be reconciled as the topochemical delignification being the result of an interplay among the morphological, physical, and chemical properties of the cell wall matrix. This subject merits further investigation.

REFERENCES I. 2.

3. 4. S. 6. 7. 8. 9. IO. I 1. 12.

13.

D. Fengel and G. Wegener, Wood-Chemistry,Ultrastrrrcture, Rerrctions. Walter de Gruyter, Berlin (1984). Y.-Z. Lai, in WoodandCellulosic Chenzistry (D. N.-S.Hon and N. Shiraishi, eds.), Marcel Dekker, New York, p. 455 (1990). E. Sjostrom, Wood Chenzistry-Furzdcrrnentnls rrtzd Applicatiom, 2nded.,AcademicPress, New York, London ( 1993). S. A. Rydholm, Pulping Processes, Wiley-Interscience, New York (1965). E. Sjostrom, Tappi, 60(9): 15 1 ( 1 977). T. E. Timell. A h . Cmhhydr: Chern., 20:409(1965). D. A. I. Goring, in Lignin: Properties and Muterinls (W. G. Glasser and S. Sarkanen, eds.), ACS Symp. Ser. No. 397, American Chemical Society, Washington, DC, p. 2 (1989). C.-L. Chen, in Wood Structure and Cornpositions (M. Lewin and I. S. Goldstein, eds.), Marcel Dekker, New York, p. 183 ( 1991). Y.-Z. Lai and K. V. Sarkanen, in Lignin (K. V. SarkanenandC.Ludwig,eds.), WileyInterscience, New York, p. 165 ( I97 1 ). J. A. Rendleman, Jr., A h . Chern. & K , No. l 1 7 5 1 (1973). A. H. Haines. A h . Carboh>vlt: Chenz., 33: I 1 (1976). L. R.Schroeder, V. M. Gentile,andR. H. Atalla, J. Wood Chenz. Techno/.. 6(1):1 (1986). 0. Samuelson, Proc. l Y K l h r . S w p . WoodPulping Cham.. SPCI.Stockholm, vol. 2, p. 78 (1981).

502 14.

15. 16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32.

33.

34.

35. 36. 37.

38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49.

SO. 51. 52. 53.

Lai V. M. Gentile, L. R. Schroeder, and R. H. Atalla, in The Structure r$Ce//u/ose (R. H. Atalla, ed.),ACSSymp. Ser. No. 340, AmericanChemical Society,Washington, DC, p. 272 (1987). J. C. Speck, Jr., A h . Carbohyd,: Chem., 1 3 5 3 (1958). J. W. Green, in The Curbohydrutes (W. Pigrnan and D. Horton, eds.), Academic Press, New York, vol. IA, p. 1101 (1980). J. M. Sugihara, Adv. Carbohydr. Chem., 8:1 (1953). R. W. Lenz, J. Am. Chern. Soc., 82: I82 (1960). J. A. Rendleman, A h . Carbohyd,: Chem., 21:209 (1966). D. 0. Hearne, N. S. Thompson, and L. R. Schroeder, J . Wood Chern. T e c h ~ ~ o11(3):307 /., (1991). E. Adler, Wood Sei. Techno/., / l :169 (1977). Y.-Z. Lai, in Methods in Lignin Chemistry (S. Y. Lin and C. W. Dence, eds.),Springer-Verlag, Berlin, Heidelberg, p. 423 (1992). N. Terashima, K. Fukushima, and T. Imai, Ho/
Chemical Degradation

503

54. K. Lundquist,R.Sirnonson, and K. Tingsvik, Sverlsk fcrp/wr.stih., "44 (1983). S S . K. Lundquist.R. Simonson, and K. Tingsvik, Svensk Papperstidn., 83:452 (1980). 56. J.-P. Joseleau and C. Gancet, Svensk Pcrpperstidr1., 84(15):R123 (1981). 57. D. E. Bland andM.Menshun, A p p i t ~ 2/:17 ~, (1967). 58. J. A. Smelstorius, Hokforsch., 28(3):101(1974). 59. 0 . Eriksson, D. A. 1. Goring, and B. 0. Lindgren, Wood Sci. Techrlol.. 14:267 (1980). 60. J. Ralph,J.H.Grabber,and R.H. Hatified. Crrrbohydr. Res., 275:167 (1995). 61. J. Ralph, Proc. 9th h r . S y n p Wood Pr4lpirzg Chen1istr:v. Montreal. p. PL2 (1997). 62. 0 . Erikssonand B. 0. Lindgren, Sverz.sk Pcrpperstic/n., 80(2):59(1977). 63. J. L. Minor. J. Wood Chern. Techtlol., 2( 1):l (1982). 64. T. Koshijima, F. Yaku, and R.Tanaka, Appl. Po/yr?zerSynlp.. 28:1025 (1976). 65. B. KoSikovi. J. PolEin, and D. Joniak, Holdorsch., 31(6):191(1977). 66. B. Capon, Chern. Re)].,69407 (1969). 67. J. N.BeMiller. Adu Carbohyel~Ckem, 22:25(1967). 68. T. E. Timell, C m . J. Chern.. 42:1456 (1964). 69. W. G. Overend. in The Cr1rhohydrcrte.s (W.Pigmanand D. Horton.eds.),Academic Press. New York, p. 279 (1972). 70. M. S. Feather and J. F. Harris. J . Org. Cherx. 30: 1 53 ( 1965). 7 I . T. E. Timell, C ~ P I II nI d. . 503 ( 1964). 72. I. Johansson, B. Lindberg, and 0. Theander, Acter Chenz. Sccrtltl., 172019 (1963). 73. R. H. Marchessault and B. G. R h b y . S1wlsk Pquperstitlrz., 62:230(1959). 74. J. Nakanoand B. RSnby, Sverrsk P q p m t i d n . , 65:29 (1962). 75. L. K. Semke, N. S. Thompson. and D. G. Williams, .l. Org. Cherrz., 29: 1041 ( 1964). 76. E.Dyer, C. P. J. Glandemans, M. J. Koch,and R. H. Marchessault. J . Chern. Soc., 336 ( 1962). 77. N. Fukagawa. G. Meshitsuka. and A. Ishizu. J . Wood Chrrrl. Techrzol., 12:425 (1992). 78. T. E. Timell. W. Entermann. F. Spencer. and E. J. Slotes, C m . J. Chetn., 43:2296 (1965). 79. M. D. Saundersand T. E.Timell. Ctrrbohyd,: Res., 5:453 (1967). 80. D. B. Easty. J . 0 t ; y . Cherrl., 272102 (1962). 8 1. 0 . Theander. Actcc Chern. Sctrrzd.. IN: I297 (1964). 82. M. S . Feather and J. F. Harris. J . A m C h w l . Soc.. 895661 ( 1967). 8 3 . J. N.BeMiller. Ad\,. Ctrrbohytlr: Chem., 22:25 ( 1967). 84. M. S. Feather and J. F. Harris, Ad\: Cnrbohydr: C / I ~ ~ 2N: ? I .161 , ( 1973). 85. T. Popoff and 0. Theander, Ct/rhOh~til:R e s . . 22: I35 (1972). 86. T. Popoff and 0. Thcander. Actcr C h w . S c m t l . . B30:397 (1976). 87. T. Popoff and 0. Theander. Acter C/~errr.S c m d . , B.30:705 ( 1976). 88.

B. L. Browning,

M~~thods of

Wood Chc~rrristr~. Wiley-Interscience, New York. vol. 11. p. 589

( 1967).

89. D. F. Root. J. F. Sacman. J. F. Harris, and W. K. Ncill, Forest Prod. ./.. Y: 158 ( 1959). 90. M. L.Mednick. J. Org. C / I ~ W27:398 I., ( 1962). 9 I . J.-M. Lauriol, J. Comtat. P. Froment. F. Pla. and A. Robert. Ho/;/i)r.st./r., J/(.3):165 (1987). 92. M. L.Nelson. J. Po/yrrwr Sci.. / K 3 5 I ( 1960). 93. M. L. Nelson and V. W. Tripp. J . Po/yrrlc,r Sci., /0:577 (1953). 94. M. A. Millett. W. E. Moorc. and J. F. Sacman. 1 1 1 d Eng. CIIPI~I.. 4(,(7):1493 (1954). 95. E. H. Daruwalla and R. T. Sheet. fi,.rri/e Res. J.. 32942 (1962). 96. B. F. Wood. A. H. Conner. and C. G. Hill. Jr.. J . App/. Po/yrrler Sc.i., 37:1373 (1989). 97. C.-H. Lin. A. H. Conner. and C. G. Hill. Jr.. J. A/>/]/.Po/yrrwr Sci., 42: 417 ( 1991 ). 98. A. Shorples. Trctr~s.Fcrrrrtkcy Soc.. 53: 1003 (1957). 99. A. Sharpies. Trctr~s.E'crrcrrkcy Soc... 5 4 3 I3 ( 1978). 1 0 0 . E. H. Daruwallu and M. G. Narsian, 7lrppi, JY(3): 106 ( 1996). 1 0 1 . R. H. Atalla. in Hydrolysis of' Ccllrrlosc~:Mec.hccrli.srrl of G~,-yr~lcrtic. trnd Acid C'tr/tr/y,si,s(R. D. Brown and L. Jurasek. cds.). Adv. Chem. Ser. No. 181, American Chemical Society. Washington. DC. p. SS (1979).

504

Lai

102. B. Philipp, V. Jacopian, F. Loth, W. Hirst, and G. Schulz, in Hydrolysis of Cellulose: Mecha n i s m of Enzyrnutic ccnd Acid Cuta1ysi.s (R. D. Brown and L. Jurasek, e&.), Adv. Chem. Ser. No. 18 I , American Chemical Society, Washington, DC, p. 127 (1979). 103. A. Sharpies, in Cellulose ancl Cellulose Derivutives (N. M. Bikales andL. Segal, e&.) WileyInterscience, New York, part V, p. 991 (1971). 104. M. A. Millett and M. J. Effland, in H.ydrolysis of Cellulose (R. D. Brown and L. Jurasek, eds.), Adv. Chem. Ser. No. 181, American Chemical Society, Washington, DC, p. 7 1 (1979). 105. J. F. Saeman, Ind. Eng. Chenl., 37( 1):43 (1945). 106. C. J. Biermann, Adv. Crrrbohydr. Clzem. Biochenz., 46:25 1 ( I 988). 107. K. Freudenberg and C. Blomqvist, Be,: Drsch. Chem., B68:2070 (1935). 108. I. S. Goldstein, H. Pereira, J. L. Pittman, B. A. Strouse, and F. P. Scaringelli, Biotech. Bioeng. Synzp., 13:17 (1983). 109. E Bayat-Makooi and I. S. Goldstein, in Cellulose and Its Derivatives (J. F. Kennedy, G. 0. Phillips, D. J. Wedlock, and P. A. Williams, eds.), Ellis Horwood, Chichester, U.K., p. 135 ( 1 985). 1 IO. K.Garves, Holzforsch., 4 7 149 (1993). 111. N. Bhandari, D. G. Macdonald, and N. N. Bakhshi, Biotechnol. Bioeng., 26:320 (1984). 112. D.H. Foster and A. B. Wardrop, Austrczl. J. Sci. Res., A4:412 (1951). 113. K. Garves, Cell. Chem. Technol., 18:3 (1984). 114. J. Bouchard, G. Garnier, P. Vidal, E. Chornet, and R. P. Overend, Wood Sci. Techno[., 24: 1S9 ( 1990). 1 15. K. C. B. Wilkie, Adv. Carbohydr. Chern. Biochem., 36:215 (1979). 116. K. Shimizu, in Wood and Cellulosic Chemist~y(D. N.-S. Hon and N. Shiraishi, eds.), Marcel Dekker, New York, p. 177 (1990). 117. T. E. Timell, Wood Sci. Technol., /:45 (1967). 118. J. N. BeMiller, Adv. Carbohyd. Chetn., 22:25 (1967). 119. J. N. BeMiller, in Starch Chernistr~and Technology (R. L. Whistler and E. F. Paschall, eds.), Academic Press. New York, vol. I , p. 495 ( 1 965). 120. E. Springer and L. L. Zoh, Tnppi, 51(5):214 (1968). 121. G. E. Annergren and I. Croon, Svensk Pnpperstidn., 64:618 (1961). 122. H. E. Korte, W. Offermann, and J. PUIS,Hol&r.sch., 45:419 (1991). 123. S. K. Banerjee and T. E. Timell, Tappi 43: 10 (1960). 124. K. Lundquist, R. Simonson, and K. Tingsvik, Svensk Papperstidn.. 86:R44 (1983). 125. J. K. Hamilton, E. V. Partlow, and N. S. Thompson, J. Am. Chern. Soc., 82:451 (1960). 126. T. E. Timell. Tuppi, 45:734 (1962). 127. G. E. Annergren, I. Croon, B. F. Enstrom, and S. A. Rydholm, Svensk Papperstidn., 64386 (1961). 128. I. Croon, Svensk Pupperstidn., 66: 1 (1963). 129. K. Lundquist, Appl. Poljvner. Sci., 28: 1393 ( 1976). Mcrtericr1.s (D. N.-S. Hon, ed.), Marcel 130. Y.-Z. Lai, in Chelniccrl Modifccrtion of Ligr~ocell~rlo.sic~ Dekker, New York, p. 35 (1996). 131. J. M. Harkin. Aclv. Chem. Set:. 59:65 (1966). 132. T. Ito, N. Terashima, and S. Yasuda, Mokuzai Gakktrishi, 27(6):484 (1981). 133. B. Johansson and G. E. Miksche. Actu Chern. Scarzd.. 26:289 (1972). 134. G. J. Leary and D. A. Sawtell, Hol
Chemical Degradation

141. 142. 143. 144. 14.5. 146. 147. 148. 149. 150. 1.51.

505

A. Sakakibara, H. Takeyama, and N. Morohoshi, Ho/
E. Adler, K. Lundquist. and G. E. Miksche, Adv. C/lern. Soc., 59322 (1966). K. Lundquist and R. Lundgren, Acta Ckem. Sccrnd., 26:2005 (1972). K. Lundquist. App/. Po/yner Sei., 28:1393 (1976). L. H. Hoo. K. V. Sarkanen, and C. H. Anderson, J . Wood Chern. Techrd., 3(2):223 (1983). A. F. A. Wallis, in Lignirzs (K. V. Sarkanen and C. H. Ludwig, eds.),Wiley-Interscience, New York, p. 345 ( 197 1 ). Y. Sano, Mokuzai Gnkknishi, 21(9):508 (1975). Y. Sano, T. Yasuda, and A. Sakakibara. Mokrrzcri Gcrkkaishi. 23( 10):487( I 977). R. M. Ede and G. Brounow, Nordic Pulp Pcrper Res. J., 3(3):119 ( 1988). R. M. Ede and G. Brounow. Ho/;fursch., 43(5):317 (1989). J . L. Davis, F. Nakatsubo, K. Murakami, and T. Umezawa, M o k u z l i Gakkuishi, 33(6):478

(1987). 152. K. Lundquist and L. Ericsson, Acto CJ2enr. Scnnd., 24:368I ( l 970). . /5:669 (1981). 153. K. Lundquist, Cell. C h e r ~ Techr~ol., 154. E. Adler and K. Lundquist. Acto Chern. Scnr~d.,17:13 (1963). l.. (1997). 155. K. Shimada, S. Hosoya. and T. Ikeda, J . Wood C h m . f i ~ h r ~ o 1757 34( 1):1 1 (1980). 156. K. Kratzl and M. Oburger. Ho/,;fhr.sc~h.. 157. K. Kratzl and M. Oburger, Holifiwsch.. " ( 6 ) : 191 (1980). 23(3):84 (1969). 158. H. Nimz. Holifibrsc~l~., 159. T. Ito, N. Tcrashima. and S. Yasuda, Moklcztri Gakknishi, 27(6):484 (1981). 160. S. Yasuda and K. Ota, Mokuzai Gnkkcrishi, 32( 1):51 (1986). hi, (1982). 161. S. Yasuda, N. Terashima, and H. Kaneko. M o k t u r i G ~ k k ~ ~ i s28(9):.570 162. G.H. van der Klashorst and H. F. Straws, J . Po/yrner Sei. A: P o / w ~ e rCherrl., 24:2143 ( 1986).

163. S. Yasuda, K. Hayashi, T. Ito. and N. Tcrashima, Mokuztri Gnkknishi. 27(6):478 (1981). 164. C. W. Dence, i n Methods irl Lignirl Chemistry (S. Y. Lin and C. W. Dence, eds.), SpringerVerlag. Berlin, Heidclberg, p. 33 (1992). 165. Y.-Z. Lai and X.-P. Guo, Proc. h f . S y r p Wood Pulping Chcwzistt;y, Melbourne. vol. l , p.

199 (1991). 166. H. Nimz, Ho/;fi)rsck.,20(4):105 (1966). 167. H. Nimz, Arlgcw: Chcw~.,/3(.5):313 (1974). 168. A. Sakakibara. Wood Sei. Techrzol.. 14:89 (1980). 169. H.-M. Chang and G. G. Allan, i n Ligr1irl.s (K. V. Sarkanen and C. H. Ludwig. eds.). WileyInterscience. New York, p. 433 ( 197 1 ). 170. M. H. Winston,C.-L.Chen. J. S. Gratzl, and I. S. Goldstein, Ho/;fiw,sc/l., 40(Suppl.):45 ( 1 986). 171. D. K. Sharma and 1. S. Goldstein. J . Wood C h c w ~f i ~ h r ~ o l .l0:379 . ( 1 990). 172. D. W. Glcnnie. in Li'ynir1.s (K. V. Sarkanen and C. H. Ludwig. cds.). Wiley-Interscience. New York. p. 597 ( I97 I ). 173. J. Papadopoulos. C.-L. Chen. and 1. Goldstein. Ho/;fi)rsch.. 35(6):283 (1981). 174. S. Y. Lin and c. W. Dencc (eds.). Merhotls irl Lignirl Chcwi.str;y, Springer-Verlag. Berlin, Heidelbcrg (1992). 17.5. C. Rolando. B. Monties. and C. Lapierre. i n Mcthotls i r l Ligr~irrC/rcwri,str;\s (S. Y. Lin and C. W. Dence. eds.). Springer-Verlag. Berlin. Heidelberg. p. 334 (1992). 176. J. M. MacLeod. J . Wood Clwrn. i ' k l t r d . , 2(2):207 ( 1 982). 177. E. L. Springer and J. F. Harris. S l w d Po/ppc,r.sticln., R5( I S):RIS2 ( 1982). 178. R. W. Scott, T. H. Wegner. and J . F. Harris. J . Wood Chrrr~.7 k h m l . 3(3):245 (1983). 179. J. Papadopoulos, C.-L. Chcn. and I . S. Goldstein, J. A/)/)/.Po/yrtwr Sci., 37:63 1 ( 1983). 1x0. R. L. Cascbier. J. K. Hamilton. and H. L. Hcrgert. n~ppi, 52( 12):2369 ( I 969). 181. G. C. April. R. Bharoocha. J. Sheng. and S. Hansen, f i l p p i , 65(2):41 (1982). 182. T. Mnrttaln, E. Hiinnincn. B. Arhippainen. and M. Nikula, Proc. E ~ p p Pulpirlg i Conj:. Dcnvcr. p. 401 (19x1).

506 183. 184. 185. 186. 187. 188. 189. 190. 191. 192. 193. 194. 195. 196.

Lai

E. L. Springer, J. F. Harris, and W. K. Neill, Tnppi, 46(9):551 (1963). E. L. Springer, 7hppi, 49(3):102 (1966). J . H. Loraand M. Wayman, Tappi, 61(6):47 (1978). K. Sudo, K. Shimizu,and K. Sakurai, Ho/&rsck., 39(5):281 (1985). M. G. S. Chua and M. Wayman, Can. J. Chern., 57:1141 (1979). M. Wayman and M. G. S. Chua, Can. J. Cherx, 572599 (1979). J. H. LoraandM.Wayman, Can. J. Cl~em..58:669 (1980). M. Wayman and J. H. Lora, J. Appl. Polyrner Sei., 25:2187 (1980). M. Wayman and J. H. Lora, Tcrpl,i, 61(6):57 (1978). M. Wayman and J. H. Lora, Tappi, 62(9): 113 ( 1979). J. H. Lora and M. Wayman, Tuppi, 61(12):88 (1978). C. J. Biermann, T. P. Schultz, and G. D. McGinnis, J . Wood Chern. Techrlol., 4(1):1 I (1984). E. A. Delong, Canadian Patent 1,096,374, Iotech, Ontario, Canada ( I98 I ). R. H. Marchessault, S. Coulombe, H. Morikawa,and D. Roberts, Carl. J. C/~ern., W2372

( 1982). 197. M. Bardetand D. R. Robert, Sverlsk Pupper.sridn., 88(6):R61 (1985). 198. J. A. Hemmingson, J. Wood Chenz. Techno/.. 3(3):289 (1983). 199. J. A.Hemmingson, J. Wood Chem. Technol.. 5(4):513 (1985). 200. J. A. Hemmingson, J. Wood Chern. Teckrlo/.. 6( I ): 1 13 ( 1986). K. L. West, and J. N. Saddler, J . Wood Chett~.Techlo/., 53: 201. K. L. Mackie, H. H. Brownell, 405 ( 1985). 202. H. L. Chum, M. Ratcliff, H. A. Schroeder, and D. W. Sopher, J . Wood C/~ern. Technol., 4 4 ) : 505 ( 1 984). 203. S. V. Lonikar, N. Shiraishi, and T. Yokota, J. Wood Chern. Tech~lol.,5 ( 1 ) : 1 I 1 (1985). 204. G. Michalowicz, B. Toussaint, and M. R. Vignon, Holdorscl?..45: 175 ( 199 1 ). 205. R. H. Marchessault, S. Coulombe, T. Hanai, and H. Morikawa, PLC//? Puper Mug. C m . Trarus., 6:TR52 ( 1980). 206. A. F. A. Wallis and R. H. Wearne, Appiru, 38(6):432 (1985). 207. K. V. Sarkanen. in Progress i n Biornass Corlver.siorl (K. V. Sarkanen and D. A. Tillman, eds.), Academic Press, New York, vol. 2, p. 127 ( 1980). 208. L. PasznerandH. J. Cho. Tcppi J., 72:135 (1989). 209. K. V. Sarkanen, Ttrppi J., 7-?:215 (1990). 2 IO. Proc. Sol1~7tPltlpirlg Syrtzp., Tappi Press, Atlanta (1992). 2 1 1. T. J. McDonough, j%ppi J., 76: 186 ( 1993). 212. J. H. Lora and E. K. Pye. Proc. S o h v r l t Pul/~i17gSyrnp., Tappi Press, Atlanta, p. 27 ( 1992). 213. P. Stockburger. Tappi J., 76:71, 1993. ., I ( 199 1 ). 2 14. J. L. Davies and R. A. Young, H o / ~ ? ) r s c h4S(Suppl.):6 215. D. K . Gallagher. H. L.Hergert, M. Cronlund.andL. L. Landucci. Proc. hit. Syrrl]~.Wood Pu//~ir7gC/lcwr.. Raleigh, vol. 1 , p. 709 (1989). 216. Y.-Z. Lai and S.-P. Mun. Proc. Inr.Syn7p. Wood P l r l p i r l g Chew.. Beijing. vol. 1. p . 266 (1993). 217. Y.-Z. Lai and S.-P. Mun, Hol;fi~r.sch.,48:203 (1994). 218. S. Tirtowidjojo, K. V. Sarkanen. F. Pia, and J. L. McCarthy. Hol;fi~rsc./r.,42: l77 (1988). 2 19. G. C. Goyal and J. H. Lora. Proc. Inr. Sy~np.Wood Pdpirrg Clrc~m,Melbourne. vol. I , p. 205 (1991 ). 220. J . F. Saernan. ACS Syrnp. Set:, /44:185 (1980). 221. A. H. Comer. B. F. Wood. C. G. Hill, Jr.. and J. F. Harris. J. Wood Chc.rr1. 7i>chrrol.. .'j(4): 461 (1985). 222. D. Fengel and G. Wegcner, Ad\*. Cl1cw1.Ser:. /N/:145 (1079). 223. J. J. Smith, D. T. A. Lamport,M.C. Hawley. and S. Sclke, J . Appl. Po/yrrwr Sci.. 37:641 (198.3). C. Pedersen. J . A/>/)/. 1'01yn11~1: Sci.. .?7:653 224. J. Dct'aye. A. Gndelle. J. Papndopoulos,and (1983). 225. K. Garvcs. /Id\: Cl~orr.Ser:. /H/: I59 (1979).

Chemical

Degradation 507

226. L. Paszner, A. A. Quinde.and M. Meshgini, Int. Symp. Wood Pul/in
3 (1989). 228. W. Pigman and E. F. L. J. Anet, in The Crcrbohydrate Chenlistty and Biocherni.str~v(W. Pigman and D. Horton, eds.), Academic Press, New York, vol. IA, p. 165 (1972). 229. Y.-Z. Lai and K. V. Sarkanen, J. Polymer Sci., Port C, 28:15(1969). 230. G.MachellandG. N. Richards, Tappi, 4/:12 (1958). 23 1. 0. Sarnuelson and A. Wennerblorn, Svensk Pappersticfn., 57:827 ( I 954). 232. M. H. Johanssonand 0. Samuelson, Cczrbohydt: Res., 34:33(1974). 233. I. ForsskHhl, T. Popoff, and 0. Theander, Carbohydr. Res., 48:13 (1976). 234. R. L. Colbran and G. F. Davidson, J . TextileInst., 52:T73 (1961). 235. D. W. Haas. B. F. Hrutfiord, and K. V. Sarkanen, J . App/. Polyrner Sci., //:S87 (1967). 236. K. Christofferson and 0. Samuelson, Svensk Papperstidn., 63(20):729(1960). 237. K. Christofferson and 0. Samuelson, Svensk Pqqwrstidn., 65( 13):517 (1962). 238. U. Albertsson and 0. Sarnuelson, Svensk Papperstidn., 6324): 1001 (1962). 239. H. Meier, Svensk Papperstirh.. 65( l6):589(1962). 240. B. Lindberg, 0. Theander, and J.-E. Uddegard, Svensk Pq~perstidn.,69( 10):360 (1966). 241. R. Malinenand E. Sjostriim, Paper; Puu, 565395 (1974). 242. Y.-Z. Lai, Carlmhydr: Res., 28: I 54 ( 1 973). 243. Y.-Z. Lai and D. E. Ontto, J . Appl. Polymer Sci.. 233219 (1979). 244. V. L. Chiang and K. V. Sarkanen, J . Wood Chetn. Techno/., 5(2):203 ( 1966). 245. D. W. Clayton and L. M. Marraccini, Svensk Pqqwrstidn., 69(9):31I (1966). 246. A. R. Procter, Cell. Clwrn. Technol., 4:269(1970). 247. V. L. Chiang and K. V. Sarkanen, J . Wood Ckem. Technol., 4( 1 ): 1 (1984). 248. F. L. A. Arbin, L. R. Schroeder, N. S. Thompson, and E. W. Malcolm, Cc4l. Chetn. Techno/.. 15:523 (1981). 249. 0. Sarnuelson and L. Stolpe. 7hppi, 52(9):1709 ( 1969). 250. H. Kolrnodin and 0. Samuelson, Svensk Pcpper.stidn., 73(4):93 (1970). 25 I . T. Vuorinen and E. Sjostrom, J. Wood Chern. Technol., 2(2): I29 ( 1982). 252. L.Liiwendahl and 0. Samuelson, Actcc Chern. Scand. B3353 1 ( 1 979). 253. L. Lowendahland 0. Samuelson, 7kppi 61(2):19(1978). 254. U. Carlson and 0. Sarnuelson, Svemk Pqqx~r.~tidn., NO( I7):549 ( 1979). 255. K. Ruohoand E.Sjiistrom. %ppi. 6(7):87(1978). 256. T. Vuorinen. Corbolrytlr. Res.. IIh:61 (1983). 257. T. Vuorinen. Cccrhohydc Res., 14/:307 (1985). 258. P. Ahlgren, A. Ishizu, I. Szabo, and 0. Theander, S w n s k I ’ l / / ~ / ) ~ t . . ~ / ; ~ 7/(9):355 ltl., ( 1968). 259. B. Alfredsson and 0. Samuelson, S ~ w ~ P~cpprstidn., sk 72( 1):361 (1969). 260. Y.-Z. Lai. Pro(.. 1081 Int. Sy~np.Wood Pulping Cl~c~tn., SPCI, Stockholm, vol. 2. p. 26 (1981). 261. W. G.Overend, in The Ccr~hol~ytlrtc~~~.s-CChc~r~~i.stry trntl Bioc~herni,s/~~v (W. Pigman andD. Horton, eds.). Academic Press, New York. vol. I A , p. 279 (1972). 262. C. E. Ballou, Ctrrbohydr: Chert/.. Y:SY ( 1954). 26.3. R. C. Gasman and D. C. Johnson, J . Org. Chern.. -?I: I830 ( 1966). 264. C. M. McCloskcy and G. H. Colcman. J . Orq. Chern., I O : 184 ( 1985). 265. Y.-Z. Lai and D. E. Ontto. Ctrrbohytlt: K t ~ s . .75:sI (1979). 266. G. H. Coleman, M ~ t h Crrrholtyrlt: . Clwrn.. 2:397 ( 1963). 267. R. E. Brandon. L. R. Schroedcr, and D. C. Johnson. ACS Sytnp. Set:. /0:125 (1975). 268. D. A. Blythe and L. R. Schroeder, J . Wood Clw~tt.7 k h 1 1 o / . , 5(3):3 l3 ( 1985). 269.

270.

D. R. Dimmcl and R. M. Knylor. Proc. I n t . Syrrrp. Wood f / / / p i n g Cher~r..Helsinki. vol. I , p. 309 (lY95). J . Janson and B. Lindbcrg. A t ~ t rClwrn. S c m d . . 14(9):205I (1960).

508

Lai

271. 272. 273. 274. 275. 276. 277.

J. H. Robins and J. W. Green, Trppi S2(7):l346 (1969). Y.-Z. Lai, Curbohydr: Res., 2457 (1972). E. V. Best and J. W. Green, i'?rppi, .52(7):1321(1969). 0. Franzonand 0. Samuelson, Sverrsk Pupperstidn.. 60(23):872(1957). Y.-Z. Lai and K. V. Sarkanen. Cell. C/zenr. Techrrol.. 1517 (1967). R. D. Brooksand N. S. Thompson, Tuppi, 49(8):362(1966). G. J. Kubes, B. I . Fleming, J. M. MacLeod, and H. I. Bolker, J . Wood Cherrr. T e c h r r o l . , 3(3): 313 (1983). G. 0 . Aspinall, C. T. Greenwood, and R. J . Sturgeon, J . Chern. Soc., 3667 (1961 ). M. H. Johanssonand 0. Samuelson, Svensk Pupperstidrr., 80519 (1977). M. H. Johanssonand 0. Sarnuelson, Wood Sci. T w h r r o l . , 11:251 (1977). R.Simonson. Sverrsk Pupperstidn., 74(21):691(1971). M. H. Johanssonand 0. Samuelson, Curbnhyd,: Res.. 54:295 (1977). D. W. Clayton, Svrrlsk Pupperstidrr., 66(4):1 I5 (1963). J.-A.Hanssonand N. Hartler, Swrrsk Puppr~rstidn.,7/(9):358 (1968). 1. Croon and B. Enstrom, S ~ r r . s kPapperstidrl.. 65( 16):595 (1962). J. Kiss, A h . Ctrrbohydr: Chem., 29:229 ( I 974). J. Buchert, A. Teleman, V. Harjunpai, M. Tenkanen, L. Viikari,and T. Vuorinen, E/p/~i J.,

278. 279. 280.

281. 282. 283.

284. 285. 286. 287.

78(21 ): 125 ( 1995). 288. R. Aurell andN. Hartler, S w r w k Pnppersfirlrr., 68(3):.59(1965). 289. R. L. Casebier and J. K. Hamilton, Rrppi, 48( 11):664 (1965). 290. J.-A. Hanssonand N. Hartler, Hol$orsch., 24(2):54 (1970). 291. R. A. Young and L. Liss, Cell. Clmm Techol., 12399 (1978). 292. R. Kondoand K. V. Sarkanen, Ho/;fi~rsch.,38( 1):31 (1984). 293. K. Niemelii and E. SjGstrGm, Ho/;fur.sch., 40:9 (1986). 294. D. R.Dinlmel andL. F. Bovee, J . Wood Chem TecArrol., 13( l4):S83 (1993). 295. D. R.Dimmel and L. F. Bovee, 1. Wood Cherrl. Techno/.,6:535 (1986). 296. J. Gierer and S. Ljunggren, Svcwsk Pupperstkh., 83( I 7 ) : m (1979). 297. T. Yaguchi, S . Hosoya. J. Nakano, A. Satoh. Y. Nomura, and M. Nakamura. M o k ~ ~ Grrkri knishi. 25(3):239 (1979). 298. E. Adler, I . Falkehag, J . Marton, and H. Halvarson. Actrr Chern. Scnr~l..18(5):1313 (1964). 299. J. R. Obst. L. L. Landucci,and N. Sanyer, firppi 62(1):55(1979). 300. J. Gierer. B. Lenz. and N.-H. Wallin, Actcr Chern. Scmd., 18:1469 ( 1964). 301. J. R. Obst. i?/~pi, 64( 10):89 (I981 ). 302. G. Brunow and K. Poppius. Acto Clrerrr. Sr~rrrd.,B36:377 ( 1982). 303. S. Ljunggren, S I W IPuppcwtidrr., .~~ 83( 13):363 ( I 980). 304. H. Taneda.J. Nakano, S. Hosoya,and H.-M. Chang. .l. Wood Cherrr. Echrlol., 7(4):485 ( 1987).

B. Johansson and G. Miksche. Actu Chcw~.S r ~ r r ~ d23:924 . (1969). G. Brunow and G. Miksche. Appl. Pnlyrrwr Syrvp., 28:l 155 (1976). 307. J. Gierer, F. Imsgard. and 1. Pettersson. App/. P o / w l r r S w r p . , 28: 1 I95 ( 1976). 308. S . Yasuda.B.-H. Yoon. and N . Terashimo, Mokrrvri Gtrkktri.slri, 26(6):421 (1980). 309. J. Gierer and S. WHnnstrGm. Ho/;fi)r.sch.,40:347 ( 1986). 310. 1. Gicrer. 1. Norin. and S. Wiinnstriim. Hn/~fi)r.srh.. 41:79 (1987). 3 I 1. E. Adler. G. E. Miksche. and B. Johansson. Hol;fi~r.sch..22(6):17I ( 1968). 3 12. J. Gierer and I. Noren. Actcl CIIPIII. Sr~rrrcl..16:17I3 ( 1962). 313. H. I. Bolkcr. A. M. Frutcau de Laclos. and S. A. Mirshokratc. J . Prrlp Ptrpc,r. Sci.. /7(6):Jl94

305. 306.

(1991).

G. E. Miksche, Ar.tcl Chern. Scurltl.. 26:4137 (1972). G . E. Miksche. Ac.trr Chcwr. S r ~ r r r d . , 2633275 (1972). 3 16. G. E. Mikschc. Actrr Chorrr. Scurrtl.. 27: I355 (1973). 3 17. R.Kondo. Y. Tsutsumi, and H. Imamura. Ho/,7for:sc~h..41(7):83 ( 1987). 318. Y.Tsutsumi. R. Kondo, and H. Imatnura, J . Wood C/rerrr. k h r r o l . , l.?(1 ):25 ( 1993).

314. 315.

Chemical

Degradation509

319. J. Gierer and L.-.&. Smedman, A r m Chem. Scand., 19:I103 (1965). 320. J. Gierer, I. Pettersson, L.-.&. Smedman,and

I. Wennberg, Acta Chem. Scand., 272083

( 1 973).

321. 322. 323. 324. 325. 326.

G. Brunow, T. Ilus, and G. Miksche, Acta Chem. Scand., 26:I 117 (1972). B. N. BrogdonandD. R. Dimmel, J. Wood Chem. Technol., /6(3):261 (1996). J. Gierer, 0. Lindeberg, and I. Noren, Holzj3rsch., 33(6):213(1979). L. L. Landucci, Tappi, 63(7):95(1980). D. R. Dimmel, J. Wood Chem. Technol., 5(1):1 (1985). D. R. Dimmel, L. F. Perry, P. D. Palasz, and H. L. Chum, J. Wood Chem. Technnl., 5( 1):15

(1985). 327. 328. 329. 330. 33 1 . 332. 333. 334.

J. Giererand I. Pettersson, Can. J. Chern., 5.5593 (1977). J. Gierer and S. Ljunggren, Svensk Pupperstidn., 86(9):R100 (1983). T. J. Fullerton and L. J. Wright, Tcppi, 67(3):78(1984). T. J. Fullerton and A. L. Wilkins, J. Wood Chetn. Technol., 5(2):189 (1985). L. J. Wright and T. J. Fullerton, J. Wood Chetn. Technol., 4( l):6l (1984). K. Poppius. Acta Chenz. Scund., B38:611 (1984). S. Ohara, G. Meshitsuka, and J. Nakano, MokuzaiGakkuishi, 29(9):61I (1983). H. H. Holton, PulpPuper Mag. Can., 78:T218(1977). 335. G. J. Kubes, B. I. Fleming, J. M.Macleod,and H. I. Bolker, Wood Sei. Techtzol., 14:207 ( 1980). 336. B.N.Brogdon andD. R. Dimmel, J . Wood Chern. Technol.. 16(3):285(1996). 337. B. N. Brogdonand D. R. Dimmel, J. WoodChetn.Technol., 16(3):297(1996). 338. T. J. Fullerton, Svensk Pupperstidn.. 78(6):224(197.5). 339. J. Giererand S. Ljunggren, Svensk Pappersridn., 83(3):71(1979). 340. T. F. Hubbard, T. P. Schultz, and T. H. Fisher, Holzjiorsch., 46315 (1992). 341. J. Giererand I. Nor&. Acrn Chern. Scand., /6:1976(1962). 342. J. R. Obstand N. Sanyer, Tappi, 63(7):11I (1980). 343. J. R. Obst, Holzjbrsch., 37(1):23(1983). 344. J. Gierer, S. Ljunggren, P. Ljungquist, and I. Nor& Svetzsk pup per st id^^., 83(3):75 (1980). 345. S. Ljunggren, Proc. 1981 h r . Svmp. WoodPulpingChern., Stockholm, vol. 5 , p. 83 (1981). 346. S. Ljunggren, P. 0. Ljungquist, and U . Wenger, Acta Chem. Scund., B37313 (1983). 347. Y. Matsumato, A. Ishizu, and J. Nakano, MokuzaiGnkkuishi. 26(12):806 (1980). 348. E. Adler, J. Marton,and I. Falkehag, Acta Chetn. Scund., /8(5):1311 (1964). 349. S. 1. Falkehag, J. Marton, and E. Adler, Adv. Chon. Ser., 5975 (1966). 350. G.Miksche, Acra Chern. Scand.. 263260 (1972). 351. B.-Y. Yoon, N. Terashima. and S. Yasuda. Mokuzui Gakkuishi, 27(4):31I (1981). 352. S. Yasuda, Mokuzui Cakkaishi, 31(2): 119 (1985). 353. J. Giererand L.-.&. Smedman, Acru Chetn. Scund., 25:1461 (1971). 354. K. Kratzl, Paperi PLIII,11:643 (1961). 355. K. Kratzl, E. Risnyovszky-Schiifer, P. Claus. and E. Wittmann. Holzjiorsch., 20(1):21 (1966). 356. K. Kratzl and J. Spona, Hol$~r.sch., 20(1):27 (1966). 357. D. W. Goheen, in Lignins (K. V. Sarkanen and C. H. Ludwig, eds.), Wiley-Interscience, New York, p. 797 ( 197 1 ).

358.

359. 360. 361. 362. 363. 364. 365. 366.

J. Marton, in Lignins (K. V. SarkanenandC. H. Ludwig,eds.), Wiley-Interscience,New York, p. 639 (1971). J. Gierer, WoodSei.Techno/., 14:241 (1980). J. Gierer, Hol;for.sch.. 44(5):387 (lY90). J. Gierer, Hol;for.sch., 44(6):395 (1990). D. A. Smith and D. R. Dimmel, J . Wood Chem. Technol., 4( 1):7S (1984). S. Ohara, S. Hosoya, and J. Nakano. Mokuzui Gnkknishi, 26(6):408(1980). D. R. Dimmel, D. Shepard, and T. A. Brown, J. Wood Chetn. T e c h n o l . , l(2):123 (1981). R . Kondo and J. L. McCarthy, J. Wood Chern. T e c h n o l . , 5( 1):37 ( 1985). R. A. Barkhau, E. W. Malcolm, and D. R. Dimmel, J . Wood Chern. Technol., /4( l ) : 17 (1994).

510

Lai

367. J. Marton, T. Marton, S. I. Falkehag, and E. Adler, Adv. Chem. Se!:, 59:125 (1966). 368. K. H. Ekman, E&, 48(7):398(1965). 369. N.Terashima, H. Araki, and N.Suganuma, MokuzaiGakkaishi, 23(7):343 (1977). 370. H. Araki, Y.Tomimura, and N.Terashima, Mokuzai Gakkaishi, 26(2):102 (1980). 37 I . T. J. Fullerton, J . Wood C h e m Technol., 7(4):441 (1987). 372. J. Gierer and W. Wannstrom, Holdorsch., 38(4): 181 (1984). 373. P. J. Kleppe, Tuppi, 53(1):35 (1970). 374. P. Axeglrd, S. Norden, and A.Teder, Svensk Papperstidn., 8/(4):97 (1978). 375. L. Olm and G. Tistad, Svensk Papperstidn., 82(15):458(1979). 376. A.TederandL. Olm. Puperi PULI,63(4a):315(1981). 377. P. Axegard and J.-E.Wiken, Svensk Pappersticln., %:R178 (1983). 378. V. MaSura, Cell. Chem. Technol., /R421 (1984). 379. Y.-2. Lai and C.-H. Kuo, Ahstr: 40th Anniverscuy Cor$ of the Japnrz Woocl Research SocieV, Tokyo, p. 392 ( 1995). 380. Y.-2. Lai, A. R. Czerkies, and I. L. Shiau, Appl. Polymer S.ynzp., 37:943 (1983). 38 I . J . R. Obst. Toppi, 68(2):100 ( 1985). 382. R. Aurell, SIvnsk Prrpperstidn.. 67(2):43(1964). 383. H. Wilder and E. J. Daleski, Jr., Tappi, 48(5):293(1965). 384. T. N. Kleinert, fifppi,49(2):53 (1966). 385. S. Lemon and A. Teder, Svensk P q p x s f i d n . ,7h( 11):407 (1973). 386. G. Wilson and A. R.Procter. Pulp Paper Mrrg. C m . , 7/:T483 (1970). 387. Y. Matsunloto, A. Ishizu and J. Nakano, Hol$or.sc/t., 4O(Suppl.):81 (1986). 388. Z.-H. Jiang and D. S. Argyropoulos, J. Pulp PuperSci.. 20(7):J183 (1994). 389. G. Gellerstedt and E. L.Lindfors, Sverzsk Papperstih., 87(9):R6I (1984). 390. D. R. Robert,M. Bardet. G. Gellerstedt,andE.-L. Lindfors. J . Wood Clzeru. Techno/., 4 ( 3 ) : 239 (1984). 391. A. M.Croucher,J. A. Lloyd, K. Nishi. M. F. Pasco,and I . D. Suckling, Pro(,. Int. Synrp. Wood Pulping Chen~..Helsinki, vol. I , p. 201 ( 1995). 392. Y.2.Lai. S.-P. Mun, S.-G. Luo, H.-T. Chen, M. Ghazy, H. Xu. and J. E. Jiang. Ho/;for.sch., 49319 (199s). 393. C. T. Lindgren and M. E. Lindstrom, J . Pulp Puper Sci., 22(8):J290 (1966). 394. G. Gellerstedt and E.-L. Lindfors, S V C I I SPnppersfidn., ~ 87( 15):R 1 15 ( 1984). 395. E. Evstgneycv. H. Maiyorova. and A. Platonov, 7kppi J.. 76(5):l77 (1992). 396. T. Yamasaki. S. Hosoya.C.-L.Chen, J. S. Gratzl, andH.-M.Chang, Proc. 1981 Int. S J w p Wood Pulping Chew.. SPCI. Stockholm, vol. 2, p. 34 (I981). 397. G. Gellerstedt and E.-L. Lindfors, Pro(.. h r . Pulp Blenching Cor$. Stockholm. vol. I . p. 78 (1991). 398. G. Gellerstcdt. K. Gustafsson. and R. A. Northey, Nordic Pulp Ptrper Res. J.. 3(2):87(1988). 399. G. Gellerstedt and E.-L. Lindfors, Ho1;fimch.. 38(3):15I (1984). 400. G. Gellerstedt and K. Gustafsson. J. Wood C / w m Trchrzol.. 7( 1):65 (1987). 401. V. L. Chiang and M. Funaoka. Hd;forsch., 42(6):385(1988). 402. V. L. Chiangand M. Funaoka. H ( J / ~ ) I x .44(2):147 ~.. (1990). 403. F. D. Chan. K. L. Nguyen, and A. F. A. Wallis. J . Wood Che/u. 7 k h r w l . . 15(3):329 ( 1995). 404. P. M. Froass. A. J. Ragauskas. T. J. McDonough, and J. E. Jiang, Pro(.. h i . P d p Rlmchirfg CO$. Washington.DC.vol. 1. p.163 (1996). 405. B . Wortling. E. Turunen. and J. Sundquist, Nordic Pulp P q w r Res. J.. 3 : 144 (1992). 406. T. Iversen and S. Wiinnstriim, Ho/;forsch, 40: 19 (1986). 407. A. Isogai. A. Ishizu. and J. Naknno, .I. Wood Chenf. 7 k h 1 ? 0 / . . 7:463 ( I 987). 408. U. Westermark and K. Gustafsson. Ho1;forsc.h.. 48(Suppl.):146 (1994). 409. 0. Karlsson and U. Westermark. Proc.. Eppi Pltlpir~gCon/:.San Diego. v o l . I , p . I ( 1994). 410. J. Sundquist and T. Rantanen, Ptrperi Puu. 65( I I ):733 (1983). 41 I . J. Sundquist. T. Rantanen. and L. Hovi, P a p 4 P ~ I 66(3):253 ~I. (1984). 412. R. P. Singh (cd.), The Bleccching o f P d p , 3rd ed.. Tappi. Atlanta (1979).

Chemical

511

413. C. W. Dence and D. W. Reeve (eds.), Pulp Bleaching: Principles n t d Prccctice, Tappi Press, Atlanta (1996). 414. 0. Samuelson and L. Thode. Tappi, 52( 1 ):99 (1969). 415. 0. Samuelson and L. Stolpe, Tappi, 52(9):1709 (1969). 416. R. Malinen and E. Sjostrorn, Pap. Puu, 54:45 1 ( 1 973). 417. R. Malinen and E. Sjostrom, Pap. Puu, 55547 (1973). 41 8. T. Vuorinen and E. Sjostrom, J . Wood Chern. Technol., 2 : 129 ( I 982). 419. L. Lowendahl and 0. Sarnuelson, Acta Chern. Scand., B33:531 ( I 979). 420. U. Carlson and 0. Samuelson, Svensk Pqyxrstidn., 82:48 (1979). 421. T. Vuorinen, J . Wood Chen~.Technol., 13(1):97 (1993). 422. J. S. Gratzl, Tappi Oxygen Delig~~i'ccctior~Symp. Notes, Tappi Press, Atlanta, p. 1 (1990). 423. M. Yalpani (ed.), Polysaccharides: Synthesis, ModiJcarions, and Stnccture/Proj?erties Relcrtions, Elsevier, Amsterdam (1988). 424. S. A. Barker, E. J. Bourne, and M. Stacey, Chem. Ind. (London), 970 (1951). 425. A. Assarsson and 0. Theander, Acta Chern. Scar~d..1 8 5 5 3 (1964). 426. A. Assarsson and 0. Theander, Acta Chern. Scand., 183727 (1964). 427. A. N. de Belder, B. Lindberg, and 0. Theander, Acta C h e n ~Scanrl., 17: 1012 (1963). 428. B. Ericsson, B. Lindgren, 0. Theander, and G. Petersson, Carbohyd~:Res.. 23:323 (1972). 429. H. Kolrnodin, Carbohydr. Res., 34:227 (1974). 430. B . Ericsson and R. Malinen, Cell. Chern. Technol., 8:327 (1974). 431. R. M. Malinen and E. Sjostrom, Cell. Chenz. Technol., 9:231 (1975). 432. E. C. Millard, L. R. Schroeder, and N. S. Thompson, Carhohydr: Res., 56:259 ( 1 977). 433. 0. Larm, E. Scholander, and 0. Theander, Carbohydr: Res., 49:69 (1976). 434. P. S . Fredricks, B. 0. Lindgren, and 0. Theander, Svensk Papperstidn., 74597 (I971 ). 435. J. Gierer, Holiforsch., 44(6):395 (1990). 436. A. A. Katai and C. Schuerch, J . Polymer Sei. Part A - l , 4:2683 (1966). 437. P. Angibeaud, J. Defaye, and A. Cadelle, in Cellulose and Its Deri1wtive.s (J. F. Kennedy, G. 0. Phillips, D. J. Wedlock, and P. A. Williams,eds.),Ellis

Horwood, Chichester, U.K.,

p.

161 (1985). 438. G. Y. Pan, C.-L. Chen, H.-M. Chang, and J. S. Gratzl, Proc. b ~ tS. y n p . Wood Pulpiug Chern., SPCI Stockholm, vol. 2, p. 132 (1981). 439. G. Y. Pan, C.-L. Chen, J. S. Gratzl, and H.-M. Chang, Res. Chern. Intenned.. 21:205 (1995). 440. T. Kishimoto and F. Nakatsubo, H o l ~ o r s c h .50:372 , (1996). 441. T. P. Nevell, in Cellulose Chetnistly und Its Applications (T. P. Nevell and S. H. Zeronian, eds.), Ellis Horwood, Chichester, U.K., p. 243 (1985). 442. E. E. Treiber, in Cellulose Chemistry and I t s Applications (T. P. Nevell and S. H. Zeronian, eds.), Ellis Horwood, Chichester, U.K., p. 455 (1985). 443. E. Maekawa and T. Koshijima, in Cellulose: Strueturd and Functionul Aspects (J. F. Kennedy, G. 0. Phillips,and P. A.Williams,eds.), Ellis Horwood,Chichester, U.K., p. 337 (1989). 444. E. R. Cousins, A. L. Bullock, C. H. Mack. and S. Rowland, Textile Res. J., 34:953 (1964). 445. R. Jeffries, J. G. Roberts, and R. N. Robinson, Textile Res. J., 38:234 (1968). 446. T. J. Painter, Carbohydr. Res., 55:95 (1977). 447. S. L. Snyder, T. L. Vigo, and C. M. Welch, Carhohydr: Res., 34:91 (1974). 448. C. Bosso, J. Defaye, A. Cadelle, C. C. Wong, and C. Pederson, J . Chern. Soc., Perkin Trms. I : 1579 (1982). 449. 0. Samuelson, Svensk Pcrpperstidn., 83(8):205 (1980). 450. 0.Theander, in Chemistry of Delignijication with Oxygen, Ozone, and Peroxides ( J . S. Gratzl, J. Nakano, and R. P. Singh, eds.), Uni, Tokyo, p. 43 (1980). 45 1, E. Sjostrom, Paperi Puu, 63(6-7):438 ( 1 98 1). 452. I. Norstedt and 0. Samuelson, Svensk PLrpperstidn., 68565 (1965). 453. R. Malinen, Paperi Puu, 57(4a):193 (1975). 454. E. Sjostrom, Paperi Puu, 63:438 (1981).

512

Lai

455. E. Sjostrom, in Chernistry of Deligni’cation with Oxygen, Ozone, and Peroxides (J. S. Gratzl, J. Nakano, and R. P. Singh, eds.), Uni, Tokyo, p. 61 (1980). 456. A. Robert and A. Viallet, ATIP Re\]., 25:237 (1971). 457. A. Robert and C. De Choudens, ATIP Rev., 36(6):332 (1982). 458. P. J.-M. Lauriol, P. Froment, F. Pla, and A. Robert, Hol7forsch., 41:215 ( 1 987). 459. D. R. Dimmel, Proc. 1nt. Pulp Bleaching Confi, Vancouver, p. 107 (1994). 460. S.-l. Anderson and 0. Samuelson, Svensk Pupperstidn., 81:79 (1978). 461. H. Kolmodin and 0. Samuelson, Svensk Papperstidn., 74:301 (1971). 462. H. Kolmodin and 0. Samuelson, Svensk Papperstidn., 76:71 (1973). 463. C. W. Dence. in Lignins (K. V. Sarkanen and C. H. Ludwig, eds.), Wiley-Interscience, New York, p. 373 (1971). 464. Y.-Z. Lai and K. V. Sarkanen, Tappi, 5 / ( 10):449 (1968). 465. D. C.Johnson, in Chemistry of Deligni’cation withOxygen, Ozone and Perosides (J. S. Gratzl, J . Nakano, and R. P. Singh, eds.), Uni, Tokyo, p. 21 7 (1980). 466. T. J. McDonough. in Pulp Blecrching: Principles and Practice (C. W. Dence and D. W. Reeve, eds.), Tappi Press, Atlanta, p. 213 (1996). 467. E. Johansson and S. Ljunggren, Pmc. Int. S ~ w p Wood . Pulping Chern., Beijing. p. 180 (1993). 468. G. Gellerstedt, K. Gustafsson, and E.-L. Lindfors, Nordic Pulp Paper Res. J., 3:14 (1986). 469. H. Xu, S. Omori, and Y.-Z. Lai, Hol$or.sch., 49323 (1995). 470. H.Xu and Y.-Z. Lai, Proc. Tappi Pulping Cor$, Nashville, TN, p. 197 (1996). 471. Y.-Z. Lai, M. Funaoka, and H.-T. Chen., Hol+r.sch., 48(4):355 (1994). 472. D. A. 1. Goring, Appita, 38(1):3 1 (1985). 473. A. R. Procter, W. Q. Yean, and D. A. I. Goring, Pulp Paper Mug. C m . , 68(9):T445 (1967). 474. J. R. Wood, P. A. Ahrgren, and D. A. I. Goring, Svensk Pupperstidn.,75(1):15 (1972). 475. T. E. Timell, Compression Wood in Gymnosperons, Springer-Verlag, New York, vol. I , p. 433 ( 1986). 476. L. Paszner and N. C. Behera, Hol$orsch., 43:159 (1989). 477. H.-T. Chen, Y.-Z. Lai, and Y.-C. Ku, Q. J. Chin. Forest., 26( I ):93 ( 1 993). 478. H.-T. Chen, Y.-C. Ku, and Y.-Z. Lai, Q. J. Chin. Forest.. 27(3):105 (1994). 479. L. Gadda, Pciperi Puu. 63:793 ( 1981 ). 480. U. Westermark and B . Samuelson. Holzjiwsch., 4O(Suppl.): I39 (1986).

11 Weathering and Photochemistry of Wood David N.-S. Hon Clernson University, Clemson, South Carolina

1.

INTRODUCTION

Wood, a naturally occurring polymer composite composed of cellulose, hemicelluloses, lignin, and extractives, is the most versatile and widely used structural engineering material for indoor and outdoor applications. Wood holds a special place in our culture because of its impressive range of attractive qualities, such as esthetic appeal, low density, low thermal expansion, and desirable mechanical strength. Annual timber harvest and use in the United States are comparable in weight to all cement, steel, aluminum, and plastic products combined. Specifically, wood is very important in the construction field, with about 40% of all manufacturedlumberandplywood utilized for this purpose.Morethan 40% ofall exterior siding used in construction is made from wood and wood-based products, and more than 1.2 billion board feet of wood are used as siding annually on 1.5 million new houses. Wood is beautiful and durable. It is warm to the touchand is easyto process. Nonetheless, because of its biological nature and esthetic qualities, unprotected wood is susceptible to weathering, photooxidative degradation, and acid precipitation. Although the weathering of wood depends on many environmental factors, such as solar radiation (ultraviolet, visible, and infrared light), moisture (dew, rain, snow, and humidity), temperature, oxygen, and air pollutants, it is generally accepted that only a relatively narrow band of the electromagnetic spectrum of sunlight, i.e., ultraviolet light, is responsible for the primary photochemical process in the weathering or oxidative degradation of wood. Because of its esthetic appeal, wood is a good light absorber. It interacts with all ranges of electromagnetic energy, including fluorescent light and sunlight [ I ] . The interaction of photons withpolymericcompounds distributed at the woodsurfaceinvolvessomeexceedingly complex chemistry and physics. From a chemical point of view, it is not surprising that all of the wood chemical components-namely, cellulose, hemicelluloses, lignin, and extractive-are susceptible to degradation by sunlight or ultraviolet light. The consequence of this photoreaction normally leads to drastic changes in wood’s appearance, i.e., discoloration, loss of gloss and lightness, roughening and checking of surfaces, and the destruction of mechanical and physical properties [2,3]. In recent years, growth of the wood products industry has been accompanied by a significant expansion in use of wood in outdoor applications. This movehasbrought considerable attention to the importance of weathering and light-induced chemical reac513

Hon

514

tions, which determine the usefulness of wood products. Moreover, the deleterious effects of acid rain on wood buildings and wood-based products have come to the fore in the past three decades and are the subject of continuing investigation. There is an evident need for enhancing the resistance of wood to weathering and photodegradation, increasing its durability as a substrate and extending the service life of wood [4]. In responseto technological demands. the field of weathering, photochemistry, and photophysics of wood has, during the last three decades, moved from a largely empirical body of accumulated practical knowledge to an increasingly sophisticated science employing the most advanced techniques of physics and chemistry. As a result, our understanding of weathering and photodegradation is not insignificant. For a comprehensive view of wood photodegradation, a number of specialty articles are available [5-71. However, for the purposes of this chapter, it is instructive to consider the general features of weathering, photodegradation, and photooxidation of wood and its chemical components. The degradative effects of air pollutants and acid rain on the wood surface quality are also considered.

II. LIGHT-ABSORPTION CHARACTERISTICSOF WOOD CONSTITUENTS: CELLULOSE, HEMICELLULOSE, LIGNIN, AND EXTRACTIVES

Wood exhibits beautiful color because its chemical constituents eitherreflect, scatter, transmit, or absorb light. Unfortunately, because of its light-absorption properties, a small part of this absorbed energy may well be specific enough to trigger undesired photophysical and photochemicalprocesses.Suchphotochemicalprocessescaneventuallychange the chemical, physical, optical, and mechanical properties of wood surfaces. In fact, the quantum energies associated with light at the short ultraviolet end of the sunlight spectrum are more than sufficient to break many of the chemical bonds present in wood constituents, namely, cellulose, hemicelluloses, and lignin (Fig. l ) . For example, carbon-carbon, carbon-oxygen, and carbon-hydrogen bonds that form the basic backbone skeleton of holocellulose and lignin might be expected to degrade with light of quantum energy at or 120

110

-

h

.

32 100-

-E

90 -

Y 80W C

W

-

z

70

CS

60: 50L 200

' 250

'

I

I

l

1

I

350 400 Wavelength (nm) 300

I

L

450

FIGURE 1 Approximate bond energy of chemical bonds in woods.

Photochemistry and Weathering

of Wood

515

above the value associated with the bond strength. Hence, the absorption of light is a topic of theoretical and practical interest to wood scientists and technologists. Wood is a composite material containing intrinsic and extrinsic color that possesses functional groups or systems, i.e., chromophores, which will absorb light. The chromophoric groups or systems may control the course of photoreactions by acting as donors or acceptors in energy-transfer processes. Chain impurities of wood polymers may play a special role a s weak links or energy sinks and thereby control the act of chain scission in photoreaction of wood. An inspection of the ultraviolet (UV) absorption curves of wood, cellulose, hemicelluloses, and lignin (Fig. 2 ) reveals that cellulose absorbs light stronglybetween200 and 300 nm, with a tail of absorption extending to 400 nm. Lignin and polyphenols, i.e., extractives, absorb light strongly below 200 nm and have a strong peak at 280 nm, with absorption down through the visible region. The combination of the U V absorption curves of cellulose and lignin makes up the absorption curve of wood. According to Norrstrom [S], lignin contributes 80-95%, the carbohydrates 5-20%, and the extractives about 2% of the absorption coefficient. Pure cellulose is not a good light absorber. The absorption by cellulose may be due to the presence of a carbonyl group that is accidentally introduced into cellulose molecules during its isolation and purification. In addition to the carbonyl group, it has been suggested that the absorption chromophores in cellulose also can be contributed by acetal 19-1 I] or ketonic carbonyl [ 121 groups at the C l position of the nonreducing glucose unit. Because of the structural similarity, the light-absorption characteristics of hemicellulose should resemble those of cellulose. Unlike cellulose, lignin is a good light absorber. The absorption occurs at chromophoric structural elements within the molecular network of lignin. Hon and Glasser 1131 have classified the potential chromophoric system as follows:

1.

Chromophoric functional groups:phenolichydroxylgroups,doublebonds, bonyl groups, etc. 2.Chromophoricsystems:quinones,quinonemethides,biphenyls,etc.

1.0

0

FIGURE 2

250

300 350 Wavelength

400 km)

450

480

Ultravioletspectra of (a) wood, (b) lignin. and (c) cellulose.

car-

Hon

516

3.

4.

5.

Leuco chromophoric systems: methylenequinones, phenanthraenequinones, phenylnaphthalenediones, bimethylenequainones, etc. Intermediates:free radicals Complexes:chelatestructureswithmetalions

Because of lignin’s predominant light-absorption properties, it absorbs more light, resulting in more degradation than cellulose. Moreover, because of lignin’s phenolic structures, the photon energy absorbed by cellulose is likely to delocalize and transfer to lignin. Hence, the presence of lignin will protect cellulose from photodegradation to some extent [ 141, although lignin itself is consequently discolored and degraded. The protective effect of lignin also has been observed by Kleinert [ 151. He explained that this phenomenon is due to the high absorption and the strong capability of autoxidation of lignin. Nonetheless, all of the components in wood are capable of absorbing enough ultraviolet and visible light to trigger photochemical reactions, leading ultimately to changes in esthetic. physical, chemical, and mechanical properties. Because of the wide range of chromophoric groups or systems distributed at wood surface components, light cannot easily penetrate into wood. Furthermore, the absorption of light at the wood surface triggers photochemical reactions that result in rapid discoloration that forms a shielding layer to avoid further light penetration. Essentially, discoloration of wood by light is a superficial phenomenon. The dark-brown surface layers of ponderosa pine and redwood that are affected by light extend only 0.5-2.5 mm into the wood [ 16,171. As photooxidation progresses, most woods change to a grayish color, but only to a depth of about 0.10-0.25 mm. Visible (400-750 nm) light, as measured spectrophotometrically, can penetrate into wood asfar as 2540 p m [ 171. The gray wood surface layer was reported to be 125 p m thick; beneath the gray layer was a brown layer from 508 to 2540 p m thick. Moreimportant,thesecolorchanges are the consequences of photochemical reactions. The use of UV light-transmission techniques to measure free-radical chain reactions that are always involved in the penetration of light through radial and tangential sections of different woods as a functionof thickness revealed that UV light cannot penetrate deeper than 75 pm. Visible light, on the other hand, can penetrate up to 200 p m into wood [ 181. Although visible light can penetrate deeper in wood, its energy (400-700 nm) is insufficient to cleave chemical bonds in any of the wood constituents because the energy is less than 70 k c a h o l (see Fig. l). Consequently, the brown color formed beneath a depth of 508-2540 p m could not be caused by visible light. It must be due to a sequential freeradical chain reaction. It has been suggested that the aromatic moieties of wood components at wood surfaces initially absorb UV light, and that an energy-transfer process from molecule to molecule dissipates the excess energy to create new free radicals [ 181. The use of electron spin resonance (ESR) techniques to monitor free radicals generated underneath different layers ofwood confirmed that primaryfree radicals weregenerated underneath 80 p m of exposed wood [ 181. This fact also had been substantiated by studying the changes in infrared (IR) absorption of carbonyl groups and lignin of wood exposed to UV light underneath different layers of wood [191. The energy-transfer processes betweenelectronically excited groupsat the outerlayer of the wood surface and another group underneath the wood surface thereby account for the photoinduced discoloration of wood underneath the surface, which absorbs practically no UV light. Furthermore, free radicals generated by light are high in energy and tend to undergo chain reactions to stabilize their parent radicals. Consequently, new free radicals formed in this way may migrate deeper into wood and cause discoloration reactions.

517

Weathering and Photochemistryof Wood

111.

ULTRAVIOLET LIGHT-INDUCED FREE-RADICAL REACTIONS

Becausehomolyticbond dissociation is the usual pathway in chemicaldeactivation of excited states, free radicals are almost ubiquitous in photochemical processes. It is therefore imperative to discuss the nature and characteristics of free radicals generated in wood in order to obtain a complete picture of the mechanisms of photodegradation and photooxidation.

A.

Free-Radical Characteristics in PhotoirradiatedWood Surface

Wood does not contain any intrinsic free radicals [20]. However, wood, wood fiber composites, and isolated lignin do contain certain amountsof stable free radicals that are detectable by ESR spectroscopy [21,22]. Although a trace amount of free radicals is detectable from freshly cut wood in the presence of oxygen, most stable free radicals are generated in wood and its processed products during mechanical preparation as well as in wood being exposed to electromagnetic irradiation [23-251. Free radicals can be detected from cellulose, hemicelluloses, and lignin under the action of ultraviolet light (see subsequent sections). The formation of free radicals is a sign of initiative degradation of the polymer. ESR studies revealed that wood interacts readily with sunlight, fluorescent light, and artificial UV light to produce free radicals, either in the presence of air or in vacuo [ l ] (Fig.3).Generally, free radicals generated in vacuohave a relatively longlifetime compared to those generated in the presence of air. Higher amounts of free radicals were generated in vacuo than in air. Oxygen is a mandatory element to activate the wood surface Storage Tlme (hours)

0

330

Light off

Lighton

-.-P c

24

270

48

12

96

I.Vac, control 2.Vac, fluorescent lamp 3.Air. control 4.Air, fluorescent lamp

a-

L

e

210 23 >

c 1 -

In

c

0)

E: 150 m

c m .v, % W .c

-2 m

30 1

240

180 60

l

120

IrradiationTime(min)

FIGURE 3 ESR signal intensity (recorded at storage time at ambient temperature.

77 K) of wood as a function of irradiation time and

Hon

518

for promoting free-radical formation. When woodis exposed to a weaker light such as fluorescent light at ambient conditions,the free radicals generated are active toward oxygen and other active gases such as sulfur dioxide and nitric oxide. They will be discussed in Section 1II.E.

B.

Free-Radical Reactions in Cellulose and Hemicellulose

The light sensitivity of cellulose has been recognized for more than a century. In 1883, Witz showed that the photodegradation of cellulose is chemical in nature [26]. Since then, many investigations have been conducted to obtain information about the effect of ultraviolet light oncellulose[27],forsuch reactions are of commercial significance. Free radicals are believed to be important intermediates involved in photodegradation. Understanding the free-radical species generated can provide important clues about the entire degradation mechanisms. The use of ESR spectroscopy has been found useful in elucidating primary radicals formed in cellulose during photoirradiation. Intensive ESR studies have been performed by Hon, and most of the free radicals formed have been successfully identified [28]. The rate of photodegradation for cellulose and hemicellulose depends markedly on the intensity and energy distribution of the light. Pure cellulose is not influenced in vacuo by the irradiation of light longer than 340 nm [29], and cellulose degradation by light is confined to a narrow band of the electromagnetic spectrum. However, in the presence of air (mainly oxygen), cellulose degradation may take place at a slower rate when exposed to light of wavelengths longer than 340 nm. Kleinert [30,31] produced detectable amounts of free radicals during ultraviolet irradiation ofpulpcellulose in vacuoandunderoxygen.Whencellulose is subjected to sunlight, the glycosidic linkages are cleaved, which causes a loss of strength and degree of polymerization [29]. The formation of free radicals located due to the chain scission at the C-l and C-4 positions can be detected by ESR spectrophotometry [29]. Discoloration and formation of hydroperoxide on exposed surfaces can be recognized easily. Kleinert [30] produced detectable amounts of free radicals during irradiation of pulp cellulose in vacuo, and under oxygen and nitrogen atmospheres. The rate and type of free-radical formation in cellulose are heavily dependent on the irradiation energy, i.e., wavelength. The formation mechanisms of various free radicals by different wavelengths deduced from the extensive ESR studies are summarized in Figs. 4, 5, and 6. Whencellulose is exposedto light withwavelengthslonger than 340 nm in

FIGURE 4

Free-radical formation in cellulose irradiated with ultraviolet light of h > 340 nm.

519

Weathering and Photochemistryof Wood

\

t

\.m* .;eyoHm H0

OH

H&OH

\

FIGURE 5 Free-radical formation in cellulose irradiated with ultraviolet light of A > 280 nm.

vacuum, no free-radical formation is noticed. However, in the presence of oxygen, alkoxy and carbon radicals due to the cleavage of glycosidic bonds are noticed. With wavelengths longer than 280 nm, in addition to chain scission, dehydrogenation takes place, preferentially at the C-l and C-5 positions. Dehydroxymethylation due to the cleavage of the C-5/C-6 side chain of cellulose is observed when cellulose is exposed to light longer than 254 nm. The formation of carbon radicals, alkoxy radicals, formyl radicals, and hydrogen atoms in cellulose irradiated with various light sources can be detectedby ESR [28]. The degreeof degradation with different light sources can be evaluated by the change of viscosity, loss of degree of polymerization, and weight loss [9]. In general,alkoxy radicals generated in cellulose are stable comparedtocarbon radicals. The carbon radicals readily undergosecondarytermination reactions. Carbon

\

CHSOH

j .CH0

t

H,

FIGURE 6 Free-radical formation in cellulose irradiated with ultraviolet light of A > 254 nm.

Hon

520

radicals in vacuo have an affinity for recombination and hydrogen abstraction to stabilize themselves in the presence of oxygen, and they are transformed rapidly into hydroperoxide radicals to build up hydroperoxide. This rapid oxygenation reaction is further accelerated when excited oxygen is presented [32]. The involvement of singlet oxygen in oxidation of cellulose has been a question of considerable debate [33,34], although several scientists have suggested that singlet oxygen is involved [35-371. Although cellulose is not sensitive to ultraviolet light of wavelengths longer than 340 nm, the presence of metal ions [37], particularly ferric ions [38], dyes [39-451, and many sensitizers [46-491, promotes free-radical formation even when cellulose is exposed to light longerthan 340 nm [37,38]. In addition to wavelengthsand sensitizers, other factors that have significant effectson free-radical formationanddegradation rate are morphology [50] and humidity and wetness [51].

C.

Free-Radical Reactions in Lignin

The conventional lignin model gives a broad picture of the reactive groups available in native lignin that make it an excellent light absorber. Lignin has an absorption peak at 280 nm, with its tail extending to over 400 nm (see Fig. 2). The reactive groups available in lignin consist of ethers of various types, primary and secondary hydroxyl groups, carbonyl groups, and carboxyl groups. There also exist a number of aromatic and phenolic sites and activated locations capable of interacting with light to initiate free-radical chain reactions. Because of the complexity of the lignin structure, identifying the free-radical sites formed is extremely difficult. However, with careful selection of model compounds, detailed study of photoinduced free radicals has been possible [52,81]. Lignin is sensitive to light with wavelengths shorter than 350 nm. Significant color buildup or formation of chromophoric groups is recognized. Based on a study using mill wood lignin and various lignin model compounds, it is clear that the phenolic hydroxyl group is an important source that reacts with light rapidly to produce a phenolic radical, which in turn transforms into o- and p-quinonoid structures by demethylation or by cleavage of the side chain, as shownin Fig. 7. Lignin is not degraded by light with wavelengths longer than 350 nm. Photobleaching or whitening of lignin can be observed when it is exposed to light with wavelengths longer than 400 nm. It also is observed that carbon-carbonbondsadjacentto a-carbonyl groups are photodissociated via the Norris type I reaction (Fig. 8). However, the Norris type I reaction does not occur efficiently in those structures with ether bonds adjacent to the a-carbonyl group. In this case, the a-carbonyl group appears to absorb light effectively and transfer such energy to the P-aryl ether linkage that leads to the cleavage of the ether bond to generatephenolicandcarbon radicals, asshown in Fig. 9. Moreover,structureswith benzoyl alcohol groups are not susceptible to photodissociation except when metal ions or other sensitizers are presented.

D. Hydroperoxidation in PhotoirradiatedWood Surface Autooxidation of wood surfaces is a very slow process. However, the rate of oxidation can be accelerated by ultraviolet light, heat, and metal ions [53]. Surface modification by ultraviolet light is also manifested by the ultimate formation of oxygenated species such ascarbonylandcarboxylgroupsaccompanied by discoloration [54].Ahydroperoxide group which appears at the early state of oxidation is the key functional group that leads to a clear understanding of the primary mechanism of oxidation. Diffuse reflectance spec-

Weathering and Photochemistry of Wood

521

l

I HC-OH

$0CH3

I HC-OH

0

\

HC-OH

L +OCH3

I

-6

OCH,

0

6

OCH,

0 FIGURE7 Formation of o - and p-quinonoid structures during ultraviolet light irradiation of lignin.

troscopy coupled with Fourier transform infrared spectrophotometry (DRIFT) can be used to detect hydroperoxide without any sample preparation or damaging the oxidized surfaces [55].The appearance of a doublet absorption peak at 3550 cm", due to the hydroperoxide, can be detected from the tangential section of a southern yellow pine, irradiated with lights of A > 223 and h > 300 nm. More hydroperoxide is detectable from the specimens irra-

i=O CEO

-0

-0

QOCH,

-0 FIGURE 8

Norristype I photoreactionin lignin.

522

Hon Energytransferredhere

here

-0

"c.I

4

0

"0

FIGURE 9 Dissociation of the @-aryl etherlinkage in lignin by an energy-transferprocess via excited a-carbonyl group.

diated with the latter light source. Competitive reactions between hydroperoxide formation and decay revealed that at the initial 90 days of irradiation with the light of A > 300 nm, the rate of formation exceeded the rate of degradation. When the wood is irradiated with light of A > 223 nm, most of the hydroperoxide is generated and converted simultaneously into carbonyl group. This chemical conversion is also observed from the specimen irradiated above 65°C. ESR can also be used to monitor the formation and decay of hydroperoxide radical. A typical ESR spectrum of peroxy radicals in photoirradiated wood is shown in Fig. 10. The peroxy radicals seek to complete their unsatisfied valences, which may be done by abstracting a proton from a nearby molecule to form a hydroperoxide. The hydroperoxide is relatively unstable toward heat and light, and is usually transformed into a new chromophoricgroupsuchasacarbonyl or carboxylicgroup.Thehydroperoxideimpurities

FIGURE 10 A typical ESR spectrum of peroxy radicals in photo-irradiated southern yellow pine.

ochemistry andWeathering

of Wood

523

generated at woodsurfacescanbedetermined by spectrophotometrictechniquesusing iodometric and triphenylphosphine methods [56].

1. Mechanism of Hydroperoxidation Formation and Decomposition a. Kinetic of Initiated Oxidation. It has been known for a long time that chemical reaction between atmospheric oxygen and organic components at wood surfaces at ambient temperature is a very slow process. However, the rate can be enhanced by metal ions and light. In common with other radical chain reactions in polymers, photooxidation of wood surfaces can be divided into three separate processes: initiation, propagation, and termination, as indicated below:

hv

+ H*

(1)

+ 0, -% RO;

(2)

Initiation:

RH”+ R.

Propagation:

R.

RO; Termination:

RO;

+ R H & ROOH + R k + R05 a non-radicalproduct

R.

+ ROi

R.

+ R - ”% non-radical product

product non-radical

(3) (4) (5)

(6)

when RH represents chemical components, such as cellulose, hemicelluloses, and lignin, at the wood surface. From these mechanisms, the rate of oxidation can be illustrated as follows:

d[o’l - k,[R.][O2] ”

dt

Using the usual steady-state assumptions, the rate of chain initiation can be illustrated as follows:

+

+

R, = k,l[ROO*]2 2krl2[R*][ROO.] kf2[R.]’

(8)

If k,,’ is equal to (kflkf2)’”,Eq. (8) can be derivatized into Eq. (9): RI = (k,,[ROO*]+ k,JR-])’

(9)

At the initiation stage, if R * or ROO. is the only product or both are formed, the rate of oxidation can be illustrated in Eqs. (lo), ( 1 l), and (12), respectively.

524

Hon

If the system is rich in oxygen, then the rate of oxidation becomes

If the system is low in oxygen, then the rate of oxidation becomes l12 "

dt

b. Formation Mechanisms. When wood is irradiated with ultraviolet light, free radicals are generated at the surfaces due to the dehydrogenation, dehydroxylation, dehyroxymethylation, demethoxylation. and chain scission that occurred in cellulose, hemicellulose, and lignin distributed at the wood surface [Eq. ( l ) ] [52].The presence of oxygen in the system provides the opportunity for oxygen molecules to react with free radicals in wood to generate hydroperoxy radicals [Eq. ( 2 ) ] ,which in turn abstract protons to produce hydroperoxides [Eq. (3)]. This can be seen from the transformation of a multiplet signal of ESR, due to the various carbon radicals, to an asymmetric singlet signal, due to the hydroperoxy radicals. Singlet oxygen, resulting from the interaction of ultraviolet light and molecular oxygen, and its subsequent attack on wood surfaces, has been proved as another possible initiation route for hydroperoxide formation [571.

E.

Reactions of Gas Pollutants in PhotoirradiatedWood Surface and Its Components

Industrial pollutants are playing an ever-increasing role in our environment. Sulfur dioxide (SO,) and nitric oxide (NO) are important gases in air pollution. Due to their high reactivity, these gases are likely to have ill effects on the surface quality of wood when it is exposed to them [58,59]. Sulfur dioxide and nitric oxide accommodate unpaired electrons; hence they possess a paramagneticproperty that canbedetected by ESR. Since these gases are free radicals in nature, they will react with photo-induced free radicals in wood, which have been shown to be quite active toward oxygen to undergo hydroperoxidation. As discussed earlier, when wood or its chemical components, namely, cellulose, hemicelluloses, and lignin, is irradiated with light, various forms of free radicals are generated. In short,phenoxy and carbon radicals weregenerated in cellulose,hemicelluloses,and lignin. While most of the phenoxy radicals are stable at room temperature, carbon and alkoxy radicals are unstable at that temperature. Nitric oxide readily reacted with carbon and alkoxy radicals to produce nonradical nitroso and nitrite products, respectively. However, only a portion of carbon radicals react with sulfur dioxide to produce sulfonyl radicals. Alkoxy radicals are reactive with sulfur dioxide to produce sulfite radicals. Sulfonyl and sulfite radicals are unstable at room temperature and terminate rapidly to form sulfinic acid and sulfonate ester, respectively. Phenoxy radicals are inert toward sulfur dioxide and nitric oxide. BasedonsystematicESRstudies, the mechanisms of interaction betweensulfur dioxide and nitric oxide and photoinduced free radicals in wood have been elucidated [59]. When wood is irradiated with ultraviolet light, various free radicals are formed: Wood

-+ P-O(a).

+ P-O(s)* + R-0- + R(A)- + R(B).

(15)

where P-O(a)., P-O(s)., R-Os, R(A)., and R(B). are active phenoxy radicals, stable phenoxy radicals, alkoxy radical, group A carbon radicals, and group B carbon radicals, respectively. At 25"C, P-O(a) ., R(A) ., and R(B). decay rapidly and stabilize, whereas

hemistry andWeathering

P-O(s) dioxide,

of Wood

remainsstable.Whenphoto-inducedwood

525

free radicals are exposed to sulfur

+ SO, -+ no reaction + SO, -+ no reaction

P-O(s). P-O(a).

(16)

(17)

+ SO, -+ R-SO,. (formationofsulfonyl R(B). + SO, -+ no reaction

R(A).

R-0-

+ SO, -+ R-0-SO,.

(formation of

radical)

sulfite radical)

(20)

Sulfonyl and sulfite radicals are unstable at 25°C. They are likely to convert into sulfinic acid and sulfonate ester. When photo-induced wood free radicals are exposed to nitric oxide,

P-O(s). P-O(a)

1

+ NO -+ no reaction + NO "+ no reaction

+ NO "+ R-NO R(B)- + NO -+ R-NO + NO -+ R-0-NO R-0.

R(A)

(formation of nitroso group) (formation of nitroso group) (24) (formation of nitrite ester group) (25)

The R(B) group is more reactive with nitric oxide than sulfur dioxide. F. Effect of Water and Moisture on the Formation and Stability of Free Radicals

Wood used outdoors is exposed to the influence of water whether in the form of airborne humidity or as rain or dew. Water is considered to be a critical element in wood's photodegradability. Because water is a polar liquid, it readily penetrates and swells the wood cell walls. Water molecules may interact with free radicals generated by light. ESR studies [60] showed that when wood wasexposed to fluorescent light, the intensity, which is directly proportionalto the free-radical concentration(either in vacuoorair), initially increased as the moisture content increased from 0 to 3.2%, and reached a peak at 6.3%. At 15.9% moisture content, a significant decrease in intensity, i.e., decrease in free-radical concentration, was observed. At 31.4% moisturecontent,only a weakESR signal was detected (Fig. 1 l ) . From the stereotopochemistry point of view, it has been suggested that the principal role of water is to facilitate light penetration into the accessible regions and to open up the nonaccessible regions for light penetration. Thus, more free radicals are generated in these regions. The excess water molecules present probably trap free radicals to form a wood free-radical/water complex. The moisture content in wood cellulose itself also has exhibited a similar effect during photoirradiation. The presence of moisture in the range of 5-7% in photo-irradiated cellulose leads to a significant decrease in the ESR signal intensity, and when the moisture content increases beyond this critical range, the ESR signal intensities again increase (Fig. 12) [SO].

W . PARTICIPATION OF SINGLET OXYGEN IN THE PHOTOIRRADIATION PROCESS

In addition to sunlight and water, oxygenmolecules are among the mostubiquitous in nature. They play a unique role in many photophysical and photochemical processes. As

526

Hon

3.0

r

0% m.c.3.2%m.c.

6.3%m.c. i10.5%m.c. ;15.9%m.c.i 31.4%m.c.

L

c

Earlywood

FIGURE 11 Comparison of ESR relativesignalintensities (carriedout in a vacuum) of free radicals in earlywoodwithdifferentmoisturecontents. Key: SYP, southernyellowpine;WRC, western red cedar; DF, Douglas fir; RW, redwood.

discussed earlier, oxygen is an important element in promoting free-radical formation, and possibly that peroxide impurity is formed due to the interaction of free radicals and oxygen molecules.However, the rate of oxidation of mostpolymers is usuallyverysmall at ambient conditions without radiation. The acceleration of the reaction rate by electromagnetic energy may be due to the generation of excited oxygen species. Considerable evidenceexists that many photooxidation reactions involve the low-lyingsingletstateof oxygen ('A# and as intermediates [61]. The participation of singlet oxygen during the photoirradiation process has been reported [57]. Wood is a polymer blend containing cellulose, hemicellulose, lignin, and extractives. Thesewoodcomponentscontain internal chemicalentitiessuchascarbonyl,carboxyl, aldehyde, phenolic hydroxyl, and unsaturated double bonds, and external entities such as wax, fat, and metal ions. The absorption of light energy by these components may bring themtoanexcited triplet state that transfers the energy to triplet ground-stateoxygen

'x+)

527

Weathering and Photochemistryof Wood

.,

10

50

40

30

20

Moisture Content (“h) FIGURE 12 Relationship between ESR relative signal intensity and moisture content of cellulose irradiated with a high-pressure mercury lamp at 77 K for 60 min.

molecules to create singlet oxygen. The participation of singlet oxygen in the photooxidation ofwood wasevidenced by usingsinglet-oxygengeneratorsandsinglet-oxygen quenchersduring irradiation [57].Iodometrystudiesrevealed that hydroperoxidewas formed in wood photo-irradiated in the presence of oxygen. The formation rate of hydroperoxide at the wood surface increased when singlet-oxygen generators suchas rose bengal solutions were added to the wood prior to irradiation (Fig. 13). Peroxide radicals involved in the interimweredetected by an ESR spectrophotometer, i.e., anasymmetric single signal of peroxy radicals with an average g value of 2.021 (g,,= 2.034; g, = 2.007) was

Wood (control) Wood

wood

lrradlated In N,

1

lrrodlated In air

Wood lrradlated

In oxygen

. . . . . .

Wood nl th Rose Bengal Hood

wood with

nlth Rose Bengal Irradloted InN,

1

Rose Benpal Irradiated In air

Wood nlth Rose Bengal lrradlated In oxygen 1

0 0.4

0.2

.

1

.

0.6 0.8 O~tlcalDensltY at 360 M

1

1.0

.

1.2

FIGURE 13 Effect o f oxygen and rose bengal on the rate of peroxide formation in wood photoirrndiated for 24 h.

528

Hon

Wood (control1 Wood irradiated In air Wood wlth DABCO irrodlated in alr

I

Wood wl th Rose Bengal

IWoodwlthRoseBengalirradiated

in air

1

Wood wlth Rose Bengal andTEM Irradlated I n air

0

Wood wlth Rose Bengal and DABCO lrradlatedIn a i r 0.2 0.4 0.6 0.8 1.0 Optical Densltv at 360 nm

1

1.2

FIGURE 14 Inhibiting effect of DABCO and triethylamine o n peroxide formation in wood photoirradiated for 24 h.

detected. On the other hand, when singlet quenchers, such as triethylamine or 1,4-diazobicyclo[2,2,2]-octane (DABCO), wereusedunder identical experimentconditions, the hydroperoxide content was reduced in some cases, even in the presence of rose bengal (Fig. 14). This evidence supports the theory that singlet oxygen is formed during photoirradiation and that it interacts rapidly with free radicals of wood to produce hydroperoxides. Due to its instability against heat and light, the hydroperoxide decomposes rapidly under ambient conditions to create chromophoric groups, such as carbonyl and carboxyl groups. These groups contribute to the discoloration of wood surfaces.

V.

EFFECT OF ACID RAIN ON WOOD SURFACE QUALITY

The deleterious effects of acidrain on lakes, aquatic ecosystems, vegetation, forests, buildings, and artifacts have come to the fore in the past three decades and are the subjects of continuing investigation [62]. Wood materials offer an impressive range of attractive properties and in many of their applications they are exposed to the outdoorenvironment. Hence, they are subject to sunlight, weathering, and acid precipitation. Their increase use in outdoor applications has resulted in the needtounderstandweathering reactions involving acid rain. Although ultraviolet light is a major element in degrading wood polymers, it is to be expected that for lignocellulosic biopolymers, for which hydrolytic breakdown can be an important modeof deterioration 1631, acid rain will have a catalytic effect. In Section 1II.E we discussed that wood interacts readily with sulfur dioxide, one of the principal elements in acid rain, to trigger a wholeseries of free-radical reactions, especially in the presence of ultraviolet light. These reactions lead ultimately to discoloration and loss of surface integrity. Many factors are involved in wood deterioration in the polluted environment. The absorption of UV light and acid rain can lead to chemical changes on wood surfaces, and deterioration of tensile strength was proved experimentally. When wood surfaces are exposed to ultraviolet light, carbonyl-group content increased and lignin content decreased simultaneously. These changes are accelerated whenthey are also exposed to acid rain,

529

Weathering and Photochemistryof Wood TABLE 1 EffectofAcid Ram on Carbonyl-Group Content in Wood at 65°C and 6S% Relative Humidity in the Absence of UV Light Acid concentration (%)

Exposure time (h)

3.7 0 6 12 24 36 3.8 48

0.074 0 0.009 0.002 3.7.l 1

-

4.2

5.6I

4.2

-

-

3.9

-

I

"FTIR absorption peak r a t i o : 1735 c m '1895 c n - I.

i.e., a dilute sulfuric acid solution, especially at 65°C and 65% relative humidity. FTlR studies demonstrated that carbonyl groups are generated on photo-irradiated wood surfaces at ambient temperature, and the rate of increase in carbonyl groups is enhanced i n the presence of acid. When wood was exposed to an elevated temperature, i.e., 65°C and 65% relative humidity, its chemicalcomponents did not showa noticeable degradation. The carbonylcontent of untreated wood remained almostconstantas the exposuretime increased. The same is true for lignin content. In the presence of acid under such conditions, the carbonylgroupincreases slightly as a function of acid concentration (Table 1). No significant change in lignin is observed (Table 2). The effect of acid and U V light at 65°C and at 65% relative humidity on carbonyl-group and lignin content is significant (Table 3). When wood was exposed to UV light in the absence of acid at such conditions, it was noticed that carbonyl-groupcontent increased almost sixfold after 48 h of irradiation. About 80% loss of lignin content was observed (Table 4). Therate of degradation is further enhanced in the presence of acid. The higher the acid concentration, the more carbonyl group is found. Although lignin is resistant to acid attack, significant reduction of lignin content is observed under such conditions. The increase i n carbonylgroup and reduction of lignin content at wood surfaces being exposed to acid and U V light at ambicnt and at high temperature and high humidity

TABLE 2

Effect of Acid Rain on Lignin in Wood at 65°C and 6S% Relative Humidity i n thc Absence of U V Light Acid concentration (S)

Exposure time ( h )

4.4 0 3

6 3.8

5.5 12 24 5.8 36 48

0

0.003- 0.074 0.009

4.4" 4.3

.c>

3.8 4.2 4. I 3.9

3.8 3.8

-

-

-

5.2

-

5.0

Hon

530

TABLE 3 EffectofAcid Rain on Carbonyl-Group Content in Wood at 65°C and 65% Relative Humidity in the Presence of UV Light

Exposure time (h)

Acid concentration (%) 0.074 0 0.009 0.002

0

3.7"

6 12 24 36 48

8.2

3.7 9.8 16.0 19.0 19.3 23.0

3.7 11.7 16.5 20.4 21.3 23.7

"FTIR absorptionpeakratio: 1735 cm '1895 cm

3.7 11.9 19.2 22.8 24.4 25.4

I

signaled that chemical reactions take place in the polymeric components of wood, i.e., cellulose, hemicelluloses, and lignin. Since holocellulose is sensitive to acid hydrolysis, there is little doubt that the strength of wood will be reduced due to the chain scission reaction. The change in color of wood from pale yellow to dark brown and the embrittlement of wood specimens clearly indicate severe degradation of wood quality. Experimental results showed that oxidative degradation of wood surface is initiated by UV irradiation. However, photooxidation by itself appeared not to be a contributor to the loss of tensile strength. Acid rain appeared to be the major culprit contributing to the loss of strength [80]. Effects of acid rain on tensile strength at ambient temperature and at 65°C are shown in Figs. 15 and 16. As discussed earlier, the photo-induced free radicals in wood are capable of reacting with sulfur dioxide to produce various oxidized products. They, in turn. further influence the color and surface properties. The presence of UV light may also accelerate oxidation of sulfur dioxide to sulfur trioxide, which will subsequently react with water to produce sulfuricacid.Sulfuricacid may also form by othercatalyticprocessesinvolvingsulfur dioxide, water, and a catalyst present in the atmosphere. Hence, three plausible mechanisms that can lead to surface degradation of wood are worth considering. They are summarized below.

TABLE 4 Effect of Acid Rain on Lignin in Wood at 65°C and 65% Relative Humidity in the Presence of U V Light

Acid concentration (%I)

Exposure time (h)

0

6 12 24 36 48

0

4.44.4" I .8

0.002

0.009

4.4 2.1 I .3

1.2 I .o

0.8

0.9 0.X

"FTIR absorption peakratio: IS07 cn"/Xc)S c m

2.0 I .4 I .3 I .2 1.2 I.

0.074 4.5 2.5

1.7 I .S

I .3 I .3

531

Weathering and Photochemistryof Wood

"

t

d

IO

Exposure

40

30

20

Time

50

(H)

FIGURE 15 Effect of acid rain on tensile strength of wood at ambient temperature in the presence of UV light: (a) UV-irradiated without acid; (b) UV-irradiated with 0.002% acid; (c) UV-irradiated with 0.009% acid; (d) UV-irradiated with 0.074% acid.

100

-

0

y 80 v

C

.-0 60 W t

Q

a 40 c c

0 C

al

2 20

m

0

10

20 Exposure

30

40

50

Time (H)

FIGURE 16 Eftect o f acid rain 011 tensile strength of wood at 65°C: ( ; I ) control at 65°C: ( b ) UVirradintctl withoutacid: ( c ) UV-irradiated with 0.002% acid: (cl) UV-irradiatedwith 0.009%' acid: ( c ) UV-irrxliatcd with 0.074% acid.

Hon

532

1. 11.

+

Wood h u + woodfreeradicals Wood free radicals SO, + oxidized products

+ hu + so, + O? -3 so,

+ H,O H,SO, H,SO, + wood + degradation products 111. SO, + O2 + catalyst -+ [SO,. H 2 0 ] SO,

[S02-HrO]+ H2S0,

H,SO,

+ wood -+

degradation products

VI.

CHANGES IN PHYSICAL AND CHEMICAL PROPERTIES

A.

Discoloration

The effect of light on the color of wood is to cause it to fade or darken and to bring about changes in tone. Extensive studies and observations have shown that most, if not all, wood species of commercial importance are prone to discoloration with age. The rate of discoloration is usually related to the intensity of light and its wavelength, and also depends on the species of wood. Light of normal intensity tends to promote darkening generally, while that of high intensity usually causes fading, though often after an initial period of darkening. Pale-colored wood such as pine, oak,birch,beech, and sycamore generally respond to radiation below 400 nm, while the coloredspecies that absorblight in the visible spectrum also are sensitive to certain wavelengths i n the visible region. Discoloration is influenced by such factorsastemperature, water, and atmosphere.Thussome woods weathered by sunlight become red-brown; and in the presence of water, some woods become gray [64]. When wood is exposed to ultraviolet light fora relatively short time, changes in reflectance and color are readily observed. When birch and redwood are exposed to ultraviolet light, they darken during the first several hours in the atmospheres of air, oxygen, nitrogen, and argon. Upon continuing the UV exposure,the wood samples in air and oxygen stop darkening and become lighter, while those in nitrogen and argon continue to darken. The decreases in reflectance and color during 480 days of UV irradiation of several wood species in air are shown in Figs. 17 and 18, respectively. It is clear that, in addition to the change in reflectance, all wood species exposed to ultraviolet light changed in color from pale yellow to brown and to gray after 180 days of exposure. Changes in wood color reflect chemical changes in wood during ultraviolet irradiation. As discussed earlier, cellulose has a fair degree of resistance to photooxidation and is not known to discolor appreciably in ordinary light. Lignin, on the other hand, is more susceptible to photooxidation and readily undergoes structural changes i n ultraviolet light to generate chromophoric groups. The extent of lignin degradation has been analyzed in weathered wood from the brown underlayer to the outer gray layer of wood, as shown in Table S. The details of discoloration are discussed further in Chapter 9.

B.

MicroscopicChanges

Microscopicchangesaccompanythegross physical changes in wood during ultraviolet irradiatioll 13.65-671. The tirst sign of deterioration i n softwood surfaces is enlargemcnt

533

Weathering and Photochemistryof Wood

K

100

.-0 U

K

Southern Yellow Pine

Q) +

50

0

60

120

180

240 360 300

480

420

Weathering Time (Days)

FIGURE 17 Decrease in brightness of outdoorweathered wood.Key: 0, southern yellow pine; 0, Douglas fir.

M,

westernred

cedar;

0,

redwood;

loo 80

i

20

0

60

120 240

., .,

180

FIGURE 18 Change i n color of outdoor weathered wood. Key: wood;

Douglas fir;

420

360 300 Weathering Time (Days)

western red cedar.

0,

480

southern yellow pine;

D,

red-

Hon

534

TABLE 5 Change in Lignin Content of a Weathered Southern Pine-l0 Years Exposure

Centered portion Brown underlayer 24.2 Graylbrown underlayer Gray surface

A"

Bb

28.0 23.4 19.2 14.5

27.9 19.1 14.6

'Calculated from FlYR absorptionpeak at 1510 cm". hCalculatedfrom UV absorbance at 278 nm.

of apertures of bordered pits in radial walls of earlywood tracheids. Futo [68] observed that the degradation begins at a relatively low irradiation intensity with an attack on the compound middle lamellae. Higher intensities and longer exposure also degrade the secondary walls, as is made visible by the formation of cavernes. Elevated temperature intensifies the photolytic degradation process. The degradation of the wall substance during UV irradiation effects a contraction of the cell walls, resulting in microchecks along the compound middle lamellae and, particularly in latewood, along the border between S , and

FIGURE 19 Cross section of southern yellow pine (700X).

ochemistry andWeathering

of Wood

535

S , [69].Diagonal fissures that follow the fibril orientation of the Sz layer also have been observed. The apertures of the bordered pits in softwood were enlarged or ruptured by microchecks. The scanning electron microscope (SEM) is frequently used to study the breakdown of the structureof wood due toweathering. The surface texturesof woods from Norwegian stave churches and other wooden construction several hundred years old were studied by Borgin using an electron microscope [70,71]. Deterioration of wood surfaces after exposure to artificial W light was observed after wood was exposed for only 500 h [65]. Photodegradative effects on transverse, radial, and tangential surfaces of a typical southern yellow pine are described in the following sections.

1. Transverse Section The transverse sectionof southern yellow pine is normally quitesimple and homogeneous. Its axial system is essentially composed of wood tracheids, with only a relatively small number of parenchyma cells. An SEM micrograph of a transverse southern pine surface before exposure is shown in Fig. 19. A microtomed transverse wood surface was exposed to W light for 500 h. Surface deterioration of the exposed wood surface was observed readily from the SEM micrograph

FIGURE 20 Cross section of southern yellow pine exposed to UV light for 500 h (700X).

536

Hon

FIGURE 21 Cross section of southern yellow pine exposed to UV light for loo0 h (700X).

(Fig. 20). The cell walls were separated at the middle lamella zone. In the extreme case, the secondary wall almost collapsed. Roughening of the surfaces could be observed visually. Surface deterioration further developed when specimens were exposed for a total of lo00 h (Fig. 21). Bordered pits located at the tracheid walls were totally destroyed. The color of the exposed wood changed from pale yellow to light brown and then dark brown after 500 and 1000 h of W light exposure, respectively.

2. Radial Section Bordered pits in southern yellowpine could be observed at radial walls in both earlywood and latewood. Generally, bordered pits located in the earlywood were larger and more numerous than those in the latewood. Dpical SEM micrographs for half-bordered pitsand bordered pits at radial walls before W exposure are shown in Figs. 22 and 23. The first perceptible change in the anatomical structure of the radial sectionof southem yellow pine upon exposure appears to take place at the pits. After 500 h of UV exposure, half-bordered pits were damaged. Bordered pits also interacted with light, but to a lesser extent (Fig. 24). The bordered pits could still be recognized.In addition, checking and void formation in radial walls occasionally could be seen from the exposed specimen. After 1000 h of exposure, however, severe deterioration of the bordered pits was

chemistry andWeathering

of Wood

FIGURE 22 Half-bordered pit structures of southern yellow pine

537

on radial section (700X).

observed. The SEM micrograph (Fig. 25) shows that the apertures of bordered pits were enlarged to the limit of the pit chambers. The pit domes were destroyed completely. At the extreme, the deterioration also spread over the radial surface of the tracheid wall. Complete degradation of these cell walls would probably take place at a longer exposure time. Disappearance of bordered pits also has been observed in redwood exposed to UV light.

3. Tangential Section Bordered pits are rarely found in the tangential surfaces observed. SEM studies revealed that diagonal microchecks passing through bordered pits in tracheid cell walls were the most conspicuous anatomical change at the tangential section upon UV exposure. The narrow microchecks were oriented diagonally to the axis of the cell wall, thus indicating that microchecks occur at the fibril angles of the S2 cell wall (Figs. 26 and 27). Similar observations have been reported. The common appearance of the diagonal microchecks during UV exposure was suggested to be the resultof local concentrations of tensile stress at right angles to the fibril direction of the S , layer. Relatively wide diagonal checks were observed in the tangential section of tracheid walls of latewood.

Hon

538

FIGURE 23 Bordered pit structures of southern yellow pine on radial section (700X).

C.

ChemicalChanges

The consequences of photodegradation and photooxidation of wood are changes in chemical and physical properties. As discussed earlier, irradiated wood may exhibit a form of discoloration, loss of lightness, checking, cracking and rougheningof surfaces, damage of microstructure, and loss of weight. It is believed that these changes are caused by severe chemical modification of the structures of cellulose, hemicelluloses, and lignin. Over a century ago, Wiesner [72] reported that the intercellular substance of wood had been lostandtheremainingmembranes,consisting of chemicallypure or nearly chemically pure cellulose, were observed. As discussed in the earlier section on microscopic changes, it is clear that absorption of UV light by lignin in the middle lamella as well as in the secondary cell walls results in preferential lignin degradation. Most of the solubilized lignin degradation products are washed out by rain [54].Careful analyses of surface layers of a southern pine that had been weathered for about 10 years revealed that the top gray layer consistently exhibited a very low lignin content. The grayhrown layer had a higher lignin content than the outer gray layer but less than the brown underlayer and in the centered portion (see Table 5). Accordingly, it is obvious that ultraviolet light initiates significant modification to the wood polymeric system. From the ESCA study [73], the increase in signal intensities of carbon-oxygen bonds and oxygen-carbon-oxygen bonds (or unsaturated carbon-ox-

ochemistry andWeathering

of Wood

539

FIGURE 24 Deterioration of half-borderedpits and cell wall of southern yellow pine atradial section after exposure to UV light for 500 h (700X).

ygen bonds) and oxygen-to-carbon ratio, and the decrease in carbon-carbon and carbonhydrogen bonds of weathered and UV-irradiated wood surfaces, suggested that the wood surface was oxidized. The oxygen-to-carbon ratio data also revealed that weathered wood surface was rich in cellulose and poor in lignin. FTIR [55] studies showed the increase in carbonyl and hydroperoxide groups and the decrease in lignin content of UV-irradiated wood surfaces. On the whole, cellulose, hemicelluloses, and lignin are degraded as illustrated by increase in solubility and reducing power of cellulose and formation of carbonyl, carboxylic, hydroperoxide groups, quinone, and conjugated double bonds in lignin. The changes in carbonyl, carboxyl, and hydroperoxide groups in cellulose are shown in Table 6 [36]; and the changes in carbonyl and hydroperoxide groups in wood are shown in Table 7. During photoirradiation, in the initial stages up to 1 h, only CO, COz, Hz,and HzO were detected as gaseous products. At longer exposure times, however, methane, ethane and ethylene hydrocarbon gases were found. Figure28 shows the rateof formation of CO, COz, and H2 in a southern pine irradiated with ultraviolet light. Volatile products of formaldehyde, methanol,acetone,methylformate,acetaldehyde,propionaldehyde,vanillic acid, vanilin, and syringyl aldehyde also were found after longer exposures [57]. The photo-irradiated wood increased its solubility in water, benzene, alcohol, and alkaline aqueous solutions. Carbohydrates and phenolic compounds are detectable from

540

Hon

FIGURE 25 Deterioration of bordered pits and cell wall of southern yellow pine atradial section after exposure to UV light for lo00 h (700X).

the solutions. The contents of holocellulose, cellulose, and lignin in wood were decreased as a function of irradiation time (100 days) with different wavelengths. Table8 shows the results of such degradation. A study of the loss of weight by UV irradiation (A > 340-320 nm) was carried out by Futo [68,74]. He found that the loss of weight is highly influenced by the temperature and the irradiation energy. The loss of weight is much higher when wood is irradiated in the presence of water, which indicates that water-soluble products are formed in addition to gaseous and volatile products. The collection of water-soluble fragments was characterized using UVhisible spectroscopy by Hon and Chang [54]. They found that the lowmolecular-weight, water-soluble products are derived mostly from lignin. The degradation products contained carbonyl-conjugated phenolic hydroxyl groups and had a weight-average molecular weight of about 900 as determined by gel permeation chromatography. The degradation of cellulose under the influence of ultraviolet light is indicated by a decrease in strength and degree of polymerization, and an increase in alkali solubility and copper number. When a bleached softwood pulp was irradiated in vacuum and in oxygen for only 10 h, the DP of a-cellulose was reduced from 850 to about 380 and 260, respectively. The content of a-cellulosedecreasedfrom 88% toabout 50% and 40%, respectively [32]. Furthermore, ultraviolet light causes yellowing and browning and for-

Weathering and Photochemistryof Wood

541

M

l

"

FIGURE 26 Microchecks of cell wall of southern yellow pine at tangential section (earlywood) after exposure to UV light for 500 h (700X).

mation of carbonyl, carboxyl, and hydroperoxide groups along the cellulose chain, and a fragmentation of molecules to diversities of neutral and acidic nonvolatile, volatile, and gaseous products. Among the volatile degradation products of photo-irradiated cellulose are acetaldehyde, propionaldehyde, methyl formiate, acetone, methanol, ethanol, methane, and ethane [76]. Glucose, cellobiose, cellotriose, xylose, xylo-oligomers, arabinose, and 3-P-D-glucosido-D-arabinose wereidentified fromthesoluble degradation products [10,11,77]. The photodegradation of lignin also has been observed. The reduction of the methoxy1 content and the splitting of monomeric units were reported [15,75,78,79].

VII. CONCLUSIONS The deterioration of wood materials upon weathering involves a very complex reaction sequence. Penetration of W light into wood does not traverse deeper than 75pm. Nonetheless, wood surface reactions initiated or accelerated by light can be observed by discoloration, loss of brightness, and change in surface texture after artificial W light irradiation or long-term solar irradiation.

Hon

542

FIGURE 27 Microchecks of cell wall of southern yellow pine at tangential section (latewood) after exposure to UV light for 500 h (SOX).

TABLE 6 Formation of Carbonyl, Carboxyl, and Hydroperoxide Groups in Photo-Irradiated Cellulose" Hydroperoxide Carboxyl Carbonyl group Irradiation group time group (h)

(mmoV100 g)

0 3 5

10 15 20

0.1

(mmoV100 g)

(mmOV100g)

0.2

0.0

5.6 10.1

1.1

0.0 0.0

15.9

1.9 4.2

18.5

6.5

0.3 0.5

9.7

0.6

'Irradiated with a hlgh-pressure quartz mercury lamp (h > 253.7 nm) at amblent temperature.

543

Weathering and Photochemistryof Wood TABLE 7

Formation of Carbonyl and Hydroperoxide Groups Photo-Irradiated Wood

IrradiationCarbonyl time Hydroperoxide group group (mmoll100(h)

24

100 g)

~~

74

85

in

1.20

0

30 60 90 120

I .71 4.25

3.10

4.32 8.26

150

180 210 "Irradiated with a high-pressure quartz mercury-xenon compact arc lamp (A > 223 nm) at ambient temperature.

m 0) H

0

E

I

0

50

100 150 200 250

300

Irradiation Time (H) FIGURE 28 Rate of formation of (a) carbon dioxide, (b) carbon monoxide, and (c) hydrogen a southern yellow pine irradiated with ultraviolet light of A > 254 nm.

TABLE 8 Effect of Ultraviolet

Light on Holocellulose, Cellulose. Lignin, and Southern Pine Irradiated in Air for 100 Days

Chemical constituent wood Holocellulose Cellulose Lignin Extractives"

Extractives of

Content (%)

Unirradiated

68.1 45.3 27.8 10.5

in

A > 340 nm

A > 280 nm

A > 253 nm

58.4 43.7 22.1 16.4

56.8 41.5 14.6 27.8

34.5 32.7 8.6 40.2

"Obtained by a two-step extraction: first step, water extraction for 72 h; second step, benzynekthanol extraction for 72 h.

Hon

544

Various types of free-radical species, such as phenoxy, alkoxy, and carbon radicals, are readily generated in wood by light. Phenoxy radicals are quite stable at ambient ternperature, whereas alkoxy and carbon radicals decay rapidly at that temperature. Carbon radicals rapidly interact with oxygen to produce hydroperoxide impurities that are decomposed easily to produce chromophoric groups such as carbonyl and carboxyl groups. A competitive reaction between formation and destruction of hydroperoxide appears to be occurring during the photoirradiation, and the hydroperoxide is unstable above 65°C. The hydroperoxidation of wood surfaces can be readily detected by using the DRIFT technique without any sample preparation and destroying the surface. The ESR technique also providedvaluableinformation on the free-radical formationmechanism that explains the formation of hydroperoxide. The use of singlet-oxygen generators, such as rose bengal andmethylene blue, aswell as singlet-oxygenquenchers,such as 1,4-diazo-bicyclo[2.2.2]octane and triethylamine, suggests the participation of singlet oxygen as an effective intermediate in photooxidation reactions at the wood surface. Phenoxy radicals are inert toward sulfur dioxide and nitric oxide. All carbon radicals or alkoxy radicals are capable of reacting with nitric oxide to form nonradical products such as nitroso and nitrite groups. Somecarbon radicals are sensitivetoward sulfur dioxide to form sulfonyl and sulfite radicals andconvertedinto sulfinic acidandsulfonate ester. The presence of water in wood also influences the rate of free-radical formation. When moisture content in wood is increased from 0 to 6.3%. more free radicals are formed. Beyond this stage, the rate of radical decay increases. Infrared studies reveal that carbonyl groups are generated in cellulose and lignin. Water-soluble fractions of degraded wood exhibit characteristics of phenolic absorptions due to the loss of lignin. ESCA studies show that oxidized wood surfaces contain higher oxygen contents than unexposed wood surfaces. In addition to UV light, acid rain seems to be an important element contributing to deterioration of wood surface quality and tensile property of wood. Experimental results showed that while carbonyl group was generated on wood, lignin content simultaneously decreased when it was exposed to light; and this process was further enhanced when it was also exposed to acid rain. Either with or withoutultraviolet, at 65"C, wood deteriorated slightly faster than that at ambient temperature. Acid further accelerated the degradation process. Wood lost its tensile strength in the presence of acid both at ambient temperature and at 65°C.

REFERENCES 1. 2. 3. 4.

5. 6. 7. 8. 9. 10. 11.

D. N.-S. Hon and W. C. Feist, Wood t r n d Fibel; 12: 121 (1980). D. N.-S. Hon, J. Appl. Polymer Sci., Appl. Polymer Sywp., 3 7 845 (1983). H. Turkulin and J. Sell, Bericht 115/36 (1997). D. N.-S. Hon, Durability and Serviceability of Wood and Wood-based Products, presented at The New Tropical Timber Crops: Challenges in Processing and Utilization, International Tropical Wood Conference, June 17-20, 1997, Kuala Lumpur. Malaysia. D. Fengel andG. Wegener, Wood: Cl~en~istty, lJltrtrstrut~ture9 Recrctiorl, Walter de Gruyter, Berlin-New York ( 1984). T. Yoshinloto, Mokuztri Gtrkkniski, IS: 49 ( 1 972). W. C . Feist and D. N.-S. Hon, Ad\!. Cl~enz.SL'K,207: 401 (1984). H. Norrstrom, Sver~.Ptrpperstidn., 72: 25 (1969). N.-S. Hon, J . Polyrnt'r Sci., Polwwr Cllent. Ed., I S : 1347 (1975). A. Beelik and J. K. Hamilton, Pupel; 13: 77 (1959). A. Beelik and J. K. Hamilton, J . Org. Cllern., 26: 5074 ( 1961).

Weathering and Photochemistryof Wood

545

12. A. Box, J . Appl. Polyrner Sei.,16: 2567 ( 1972). D. N.-S.Hon and W. Glasser. Polymer-Ptrst. Teclznol. Eng.. 12: 159 (1979). 14. N.-S.Hon, J . Polymer Sei., Polyner Clzern. Ed., 13: 2641 (1975). 15. T. N.Kleinert, P q i e K 24: 207 (1970). 16. U.S.Forest Service Research Note, Wood Fini.shirt
D. N.-S. Hon, in Developnent in Polymer Degrccdrttion-3 (N. Grassie, ed.), Applied Science Publishers, Essex, U.K., chap. 8 (1982). 53. G. Scott, Atnzospheric Oxiclcrtiorz crnd Antio.ridatants, Elsevier, Amsterdam-London-New York 52.

(1965). 54. D. N.-S.Hon and S.-T. Chang, J. Polynzer Sei.. Polynler Chern. Ed., 22: 2227 (1984). 5 5 . D. N.-S. Hon and W. C. Feist, Wood Fiber Sci., 24: 448 (1992). 56. J. F. Rabek, ExperinzentulMethocls in Polynler Clzenristry, Wiley, New York, p. 625 (l98 l).

Hon

546

57. D. N.-S. Hon. S.-T. Chang, and W. C. Feist, Wood Sri. Trchnol., 16: 193 ( 1982). 58.

59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81.

D. N.-S. Hon and W. Y. Chao. in Cellulose and Wood: Chemistnland Technology (C. Schuerch, ed.), Wiley, New York, pp. 1037-1057 (1989). D. N.-S. Hon and W. C. Feist, Wood Fiber Sci., 25: 136 (1993). D. N.-S. Hon and W. C. Feist, Wood Sci., 14: 41 (1981). M. L. Kapalan and A. M. Trozzolo, in Singlet Oxygen (H. H. Wasserman and R. W. Murray, eds.). Academic Press, New York, chap. 11 ( 1 979). E. T. Paparozzi, in Materials Degradation Caused by Acid Rain (R. Baboian, ed.), American Chemical Society, Washington, DC, pp. 332-342 (1986). I. S. Goldstein, in Organic Chemicals from Bion~uss(I. S. Goldstein, ed.), CRC Press. Boca Raton, FL, pp. 101-124 (1981). T. N. Kleinert. Hol&r.sch Holzver-wert.. 22: 21 (1970). S.-T. Chang. D. N.-S. Hon, and W. C. Feist, Wood and Fiber, 14: 104 (1982). M.-I. Kuo and N. Hu, Hol@orsch., 45:347 (1991). L. Paajanen, Structural changes in primed Scots pine and Norway spruce during weathering, Maret: Srruct., 2 7 237 (1994). L. Futo, Holz Roh-Werksr., 32: 303 (1974). V. P. Miniutti. Paint Technol., 45: 27 (1973). K. Borgin, J. Microsc.. 92: 47 (1970). K. Borgin, J. Inst. Wood Sci., 5: 26 (1971). J. Wiesner, Akad. Wiss. Wien, 49: 61 (1864). D. N.-S. Hon, J. Appl. Polyrner Sci., 29: 2777 (1984). L. Futo. Holz Roh-Werkst, 34: 31 (1976). G . J. Leary, Zrppi, 50: 17 (1967). R. L. Desai and J. A. Shields, Mcrkrotnol. Chem., 122: 134 (1969). B. A. Gringras, D. Cooney, K. A. Jackson, and C. H. Bayley, Textile Res. J., 33: 1000 (1963). W. Sandermann and F. Schlumbom, Holz Roh-Werksr., 20: 245, 285 (1962). G. J. Leary, Tcrppi, 51: 257 (1968). D. N.-S. Hon, Woocl Fiber Sei., 26: 185 (1994). D. N.-S. Hon. J. WoodChem.Technol.,12: 179 (1992).

12 Microbial, Enzymatic, and Biomimetic Degradation of Lignin in Relation to Bioremediation Takefumi Hattori and Mikio Shimada Kyoto University, Kyoto, Japan

1.

INTRODUCTION

Next to cellulose, lignin is the most abundant renewable resource on Earth. Although it is an important structural component in wood, together with cellulose and hemicellulose, its utilization in isolated form as a chemical material has not yet been successfully achieved [ l ] . The main reason for this is the fact that lignin molecules lack stereoregularity and repeating units in the molecule, andthey aretooheterogeneousandcomplex.Current research interests focus instead on environmentally benign delignitication processes with biocatalytic systems,includingwhite-rotfungi, their enzymesystems, and biomimetic catalysts. Since the cellulase-less mutant of Phanerochoete chr~y~osporiumwassuccessfully isolated by Ander and Eriksson [ 1 a], more than 20 years have passed, and it is more than 10 years since discovery of lignin peroxidase (Lip) [2,3] and manganese peroxidase (MnP) [4] and proposal of the one-electron oxidation mechanism for the enzymatic reactions [571. However, it is deceptively simple to apply these bioligninolytic systems to industrial delignitication processes despite a great deal of recent work on microbial, enzymatic, and biomimetic degradation of lignin in relation to the roles of lignin-degrading enzymes of Lip, MnP, and laccase (Lac). This chapter gives a brief overview of current lignin biodegradation research, focusing on microbial, enzymatic, and biomimetic approaches to lignin biodegradation chemistry and related fields. Shimada and Higuchi [8] reviewed the biochemical approach to lignin biodegradation research in the first edition of this book, and the present authors have attempted to include as many recent findings as possible in this chapter. However, since it is beyondour ability to cover all current publications, readersshould refer to recent international proceedings [9- 121 and reviews [ 13- 191 to follow the trend of related research areas.

II. MICROBIALAPPROACH A microbial approach is basically important not only for application of fungal ligninolytic systems to biopulping, biobleaching of unbleached pulps [lo], bioremediation of waste 547

Shimada 548

and

Hattori

effluents, and bioconversion of agricultural lignocellulosic residues, but also for the basic study of physiological features of lignin decomposers in natural environments. Since the discovery of the cellulase-lessmutant, Phanemchaetechrysosporium hasacquired the most prominent status in lignin biodegradation research. However, it is still important to screen other lignin-degrading microbes with high ligninolytic activity [20,2 l ] . As to biomechanical pulping, Eriksson and Vallander reportedthat treatment of wood chips with cellulase-less mutant for 10 days could save electrical energy by about 30% [22], and several research teams have also reported that pretreatment of wood chips with wild-type white-rot fungi could succeed in reducing the energy requirement and also increase paper strength [23]. Similar results were also obtained by Akamatsu et al. [24]. To find a more appropriate fungus for this purpose, Ceriporiopsis subvertnisporu was selected out of 400 species of the ligninolytic fungi [20]. The biomechanical biopulping method has been refined [25-281. Nishida and his co-workers have succeeded in isolating a powerful ligninolytic fungus named IZU-154 [29]. It is noteworthy that they used only wood meal substrate for screening this fungus, which has, i n fact, a high selectivity for lignin decomposition during fungal treatment of the wood meals. Coarse beech mechanical pulp and wood meal were incubated with only the mycelia of IZU-154, Coriolus versiThe ligninolytic activity of fungus IZU-154 was color, and Phuneroc~huete chr~ao.sporiunz. muchgreater than that of the other fungi. Theyhave also reported that 7-dayfungal treatment of coarse hardwood mechanical pulp reduced the refining-energy requirement by aboutone-half to two-thirdsandimprovedpulp strength properties abouttwofold. Similar energy savings and improvement of pulp strength properties were achieved for softwood samples after a little longer period of treatment. As to biobleachingwork, this funguswasfound to bleachkraftpulps[30]. The bleaching activity may be due to the action of the MnP system [311, although this fungus remains to be identified. Paice et al. demonstrated direct biological bleaching of hardwood kraft pulp with the fungus Coriolus (Trametes) versicolor [32]. Kondo and Sakai and coworkers have isolated Phatlerochaete sordidu (YK-624 strain), which has strongkraft pulp bleaching activity [33]. They also demonstrated that the MnP system catalyzes pulp bleachingand that there is agood correlation between the level of MnP activity andfungal bleaching ability [34]. Decolorization of pulping waste liquors was first investigated by Fukuzumi et al. in Japan [35]. Quite recently, Pallerla and Chambers [36] reported the continuous decolorization and AOX reduction of bleach-plant effluents by freeandimmobilized Trumetes versicolor. They achieved successful results of about 89% continuous decolorization and about 70% AOX reduction. A number of papers have reported for fungal treatments of effluents from pulp mills 137-441. Bioremediation of environmentally hazardous compounds by use of wood-degrading fungi and bioreactor systems will also be important for removing carcinogenic pollutants from agricultural and industrial wastewatersbefore letting them flow into rivers and oceans. For detailed information readers should refer to the recent publications cited [43].

111.

ENZYMATIC APPROACH

As to the distribution of ligninolytic enzymes among white-rot fungi, recently, in addition to Phurlerochnete chrysosporium, several other white-rot fungi such as Ceriporiopsis suhvermisporu,Phlehinrudiatu,Phanerochuetesordidu,Bjerknnderu adustu, and Trumeres

Degradation Bioremediation Relation to ofinLignin

549

(Curiolus) versicolor alsohavebeenreceivingmuch attention i n studies of lignin biodegradation and biobleaching of unbleached pulps [ 21,45,46]. This section gives background on enzymatic systems involved in lignin biodegradation, caused primarilyby white-rot fungi. Future application problems also are discussed. The three types of ligninases, Lip,MnP, and Lac, have been focused on in lignin biodegradation in relation to the roles of white-rot fungi in natural wood-decay processes. Although the name“ligninase” for alignin-degradingenzymehas notyet been rigorously defined, it may be used as a general term for enzymes that catalyze the depolymerizationof lignin. The“ligninase” isolated fromaculture fluidof Phut~erochaere chry.so.sporiutn [47-SO] was shown to catalyze predominantly Ca-C@ bond cleavage of lignin model substrates, yielding aromatic aldehydes as major products as depicted in Fig. I . However,mechanistic studies on ligninase catalysishaverevealed that the enzyme should be called “peroxidase,” since it utilizes hydrogen peroxide to form compound 1 [5,51,52] with a high-valent form of oxo-iron porphyrin complex as a one-electron oxidizing agent [6,53]. The one-electron oxidation mechanism has beenrationalized to explain Ca-CP bond cleavage [S41 and the aromatic ring opening [55] of the lignin model substrates byLIP. However, clear-cut evidence for the substantial depolymerization of lignin has not been provided until quite recently [56],although partial enzymaticdegradationof synthetic lignins with MnP and Lip systems were previously reported [57-591.

OMe

L

CHOH

0

CH ”+”,

Me0 OMe OMe

OH

guaiacylglycol I

OMe

Me0

OMe

OMe

veratraldehyde benzaldehyde

FIGURE 1 Schematic model for Ca-Cp bondcleavage of lignin macromoleculecatalyzed by “ligninase”/HzOz, yielding aromatic aldehyde and phenylglycol products. (Modified scheme from Tien and Kirk [47].)

Shimada 550

A.

and

Hattori

Lignin Peroxidase

After purification andcharacterization, lignin peroxidase (Lip) isolated from Phaneroof 41,000-43,000 daltons chaete chrysosporiunz was found to have a molecular weight [48,6OJ, with an iron porphyrin IX as the prosthetic group [4,48,51,52,60]. The enzyme catalyzes oxidation of veratryl alcohol (VA) or 3,4-dimethoxybenzyl alcohol substrate to form a veratraldehyde product in the presence of hydrogen peroxide as shown in Fig. 2. The enzyme was found to have a broad substrate specificity to catalyze oxidations of a wide variety of substrates [48,51]. At present, the most convenient assay method for Lip is probably to use VA as a substrate instead of lignin to measure the increase in absorbance at 310 nm of veratraldehyde formed [48]. It is interesting to note that tobacco plant peroxidase can catalyze the oxidation of VA in the presence of calcium and magnesium ions and hydrogen peroxide with a pH optimum of 1.8 [61]. These dicationic metal ions plays an important role in stabilization and expression of activity under these extremely low-pH conditions. Similar stabilization of Lip by calcium ions was also observed [62]. Distribution of Lip among white-rot fungi seems to be restricted to certain consortia of white-rot fungi and it may serve to detect the Lip activity among others with low pH value around 2.0 for further survey of Lip activities [21,45,46]. 1.

The ReactionMechanism

n. C a - C P B o n d C l e a v a g e qf'P-Z Model Substrate. A one-electron oxidation mechanism has been proposed by several different laboratories [6,7,63,64] to explain the C a C p bond cleavage reactions of the side chain of lignin. Before ligninase had been identified as a hemoprotein, Shimada et al. [53] demonstrated the Ca-c@ bond cleavage of the p-l lignin model substrate under both aerobic and anaerobic conditions with biomimetic model catalysts of cytochrome P-450 and peroxidase, which were established to form compound I in the presence of iodosylbenzene or tert-butylhydroperoxide. They also established that the porphyrin catalyst used virtually mimicked lignin peroxidase i n may respects as described below. Alternatively, Kersten et al. [ 5 ] proved for the first time the ligninase-catalyzed formation of aryl cation radical of methoxybenzenesusing electron spin resonance (ESR) techniques, indicating that compound I in two-electron-deficient oxo-iron porphyrin complex stepwise abstracts two electrons from the aromatic moieties. Further, as shown in Fig. 3, Hammel et al. [63J reported that the aryl cation radical formed from diarylethane- I ,2-

CH20H

Veratryl alcohol

Veratraldehyde

FIGURE 2 Oxidation of 3,4-dimethoxybenzyl alcohol (veratryl alcohol) catalyzed by lignin peroxidase (LIP) to give rise to 3,4-dimethoxybenz aldehyde (veratraldehyde) [48]. The enzymic reaction rate can be determined by measurement of the absorption of veratraldehyde product formed at 310 nm at ambient temperature.

551

Degradationof Lignin in Relationto Bioremediation

H0

OCH, OCH3 OCH,

+i

CH0

Aerobic

02< OH

Anaerobic

7 L i p , ,, ?

4

?l-

H+,e-

FIGURE 3 Proposedschemeforaerobic and anaerobic degradation of 1-(3,4-dirnethoxyphenyl)2-phenylethane-1,2-diol catalyzed by Lip from P. chrysosporiunz [63].

diol underwent Ca-Cp bond cleavage, yieldingthe corresponding benzaldehyde products. Gold et al. [51,64] also reported the importance of compound I, which is reduced back to the native state by adding VA. Thus, the driving force for the C-C bond cleavage of the p-1 model substrate has been established to be the two-electron-deficient compound I of Lip. Hydrogen abstraction from the substrate is not necessary for the bond cleavage, but the one-electron transfer from ring A or B to form radical cations is the initial step of the homolytic or heterolytic C-C bond cleavage. Thus, mechanistic investigation showed that the primarily important active oxygen speciesformed is an oxenoidintermediate of compound I of Lip in high-valent iron porphyrin r-cation radical species rather than free active oxygen species such as singlet oxygen [65,66] and OH radical [67-691, as formerly proposed. The one-electron oxidation mechanism can also be applied to the enzymatic degradation of other dimeric and monomeric compounds as described below. b. CB-EtherCleavage of p - 0 - 4 ModelSubstrates. The reaction mechanismfor the side-chain cleavage of p-0-4 type linkage may be the most important to be clarified, because these aryl-ether bonds account for about 40% and 60% in coniferous and deciduous lignins, respectively. The enzymatic oxidation of p-0-4 substrates, similar to Fig. 3, yields the aromatic aldehyde product and smaller amounts of arylglycerol and other phenolic compounds after C-C bond cleavage [47-501. The mechanism for fungal degradations of arylglycerol-P-aryletherdimer model substrates that were reported to yield the products corresponding to guaiacol, arylglycerol,

I

Shimada 552

and

Hattori

guaicoxyethanol, and benzyl alcohol derivatives have long been problematic in the early work [70-731, because of a paucity of firm evidence to conclude the mechanism until the above one-electron oxidation mechanism had been established. Alternatively, Umezawa and Higuchi clearly indicated that guaiacol and arylglycerol were produced via different pathways and these two products were not counterparts to each other in the degradation of the p - 0 - 4 model substrates in the culture of P. chrysosporium [74].

According to the one-electronoxidationmechanism,the initial step of the bond cleavages also seems to be the aryl cation radical formation [75] in ring A or B of the substrate by compound I of Lip. Thus, C a - C p bond cleavage of p - 0 - 4 model substrate caused @-ether bond cleavage to give rise to veratraldehyde, guaiacol, and glycoaldehyde [74-781 as one of the main degradations of a p - 0 - 4 model compound catalyzed by Lip (Fig. 4). Furthermore, a wide variety of cleavage reactions of p - 0 - 4 compound clarified by Umezawa et al. with "C,"O-labeled p - 0 - 4 substrates [78,79] were summarized previously [8]. Although the mechanism of enzymatic degradation of p - 0 - 4 model compounds is more complex than that of the p-l model, the results obtained indicate that the initial step of these diverging fragmentations is the one-electron transfer to form the aryl cation radical in ring A or B of the substrate, yielding both side-chain and ring-cleavage products. Lignin contains other dimeric substructures such as pinoresinol @-p),phenylcoumaran ( p - 3 ,and biphenyl (5-5) structures. The enzymatic degradation of these model compounds has not yet been investigated in detail, except for the oxidative breakdown of the biphenyl model compound [80,81]. c. Aromatic Ring Cleavage qf Lignin Model Substrates. Aromatic ring cleavage of lignin prior to its depolymerizationwas postulated by Chen et al. [82] from chemical analyses of decayed wood lignin, and ring-cleavageproductsformed from the p - 0 - 4 substrate were first reported by Umezawa et al. [83,84]. In this case also, a one-electron oxidationmechanismwasemployedtoexplaintheringopening reaction as shown in Fig. 5. Thus, the "direct" aromatic ring cleavage of nonphenolic p - 0 - 4 model compounds was demonstrated in an in-vitro system with Lip in the presence of hydrogen peroxide,

OMe Veratraldehyde

OMe Guaiacol

FIGURE 4 Proposedmechanisms for Ca-CP and P-ether bond cleavage of p-0-4modelcompound catalyzed by ligninase/HzO2/O2and the ligninolytic culture. The chemical structure of the substrate ( I ) is not exactly the same as the ones used in the original reports 175-781. 0 , ''0 atom; intermediates (2a), (2bj, ( 2 c ) , and glycolaldehyde ( 3 ) , are all hypothetical.Theoxygenation of ( 3 ) with labeleddioxygen is notyet proved,probablydue to the rapid exchangereaction with water.

T

0

c

m

X

0

U

0-

U

5

T

I,

Y

K

N

Y

K

N

I,II

n

J

L

N

Z

Y

n

o

LLL

o

r

Degradationof Lignin in Relation to Bioremediation

5

9 0,

5

m

0%

n

a

T

o

I1

U

m

X

1,

-2.

"

Y

M

.W

.-CC

M C L

.M

U

x

a

0

E

a In

553

554

Hattori and Shimada

and the p,y-cyclic carbonate, y-formate, and diesterof oxalic acid (a-arylglycerol-p-methyloxalate) were identified by Umezawa and Higuchi, who rationally explained the reaction mechanism for the ring cleavage reactions [84]. Alternatively, Hattori et al. [85,85a] successfully elucidated a new type of oxygenation mechanism for the ring cleavage of veratryl alcohol(VA); the unprecedented regiospecific oxygenation with either water or dioxygen was demonstrated to occur during the enzymatic ring rupture (Fig. 6). d. C a - C p Bond Cleavage of Lignin Carlwhydrate Complex Model Substrcrte. In spite of the importance of the enzymaticbreakdown of lignin carbohydratecomplex (LCC), it hasbeen little investigated. However,Tokimatsu et al. havesynthesizedfour stereochemically pure isomers of p - 0 - 4 LCC model compound [86] and confirmed that the LCC model substrates also undergo the Ca-Cp bond cleavage by the Lip system as a result of breakdown of the benzyl ether bondings between glucose and the p - 0 - 4 lignin model compound moiety [87] (Fig. 7).

2. Broad Substrate Specificity of Lip Ligninase is capable of catalyzing oxidation of other organic substrates that are not at all structurally related to the lignin molecule. One example of this phenomenon is oxidation of a-keto-y-methyl thiobutyric acid and methional in the presence of hydrogen peroxide to yield ethylene, as reported by Forney et al. [68]. Lip was found to catalyze oxidation of the dye called Poly R 188,89], which may conveniently allow the colorimetric assay of the ligninase activity from microbes. However, the ligninolytic activity of the isolated microbe must be tested finally with lignin substrates. Evidence for Lip-catalyzedbreakdownof the carcinogeniccompounds,including dioxin and polyaromatic hydrocarbons, suggests that bioremediation research with whiterot fungi may be promising in the future [90]. In view of the finding that many xenobiotic pollutants were decomposed by whiterot fungi,such as P h a n e r o c h a e t e c h r ~ ~ s o . s ~ ~ o r itheir ~ l m , ligninolytic system mayplay a primary role in degradation of xenobiotic compounds 1911. VeratrylAlcohol as a Possible Mediator A possible role of veratryl alcohol (VA) as a diffusible mediator in a ligninolytic system was first proposed by Harvey et al. [92]. The mediator concept is fascinating to explain the extremely broad substrate specificity of Lip and also in that hemoprotein Lip utilizes VA cation radical mediator like a “boomerang” to attack lignin polymer at a distance rather than to attack the insoluble lignin polymer substrate directly (Fig. 8). On the other hand, VA was proposed to play a role in preventing Lip from inactivation in the presence of hydrogen peroxide in the catalytic cycle [93]. Since VA cation radical species were detected directly by ESR spectrophotometric method under specified conditions [94], it has been proposed that VA cation radical possibly acts as an enzyme-bound redox mediator but not as a diffusible oxidant for LiPcatalyzed lignin or pollutant degradation [95]. They suggested that VA may act as a redox mediator for the indirect oxidation of compounds during the reaction. As shown in Fig. 9, Shimada et al. first discovered that Lip indirectly decomposed oxalate in the presence of VA and hydrogen peroxide [96]. At almost the same time, Kirk et al. also reported similar results [97]. They reported that VA oxidation was noncompetitively inhibited by oxalate present in the reaction mixture [96,97]. Similar noncompetitive 3.

I:

aJ

0

0 L O

0

I

Degradation of Lignin in Relation to Bioremediation

aJ

l =

X

; 4 20 I

555

556

Hattori and Shimada

CH0 OCH3

-

-

OEt

at 22Oc

OEt

Erythro-f o m or Threo-f o m

FIGURE 7 Lip-catalyzedcleavage of benzyl etherlinkage of thelignin model compounds 1871.

carbohydratecomplex

FIGURE 8 Proposed mechanisms for the degradation of lignin by veratryl alcohol cation radical as a redox mediator [92].

CH0

OCH3

@

veratraldehyde

OC

OCH3

VA’

YOOH

’COOH

FIGURE 9 Proposed mechanisms for the degradation of oxalate by veratryl alcohol cation radical [96]. The lower the level of oxalate in this reaction system, the more efficiently VA is oxidized to form cation radical, which is converted to veratraldehyde(route a) or attacks lignin and other organic compounds. The higher the level of oxalate, themoreefficiently scavenged is the cation radical intermediate back to VA.

Degradation Bioremediation Relation to ofinLignin

557

inhibition caused by cellobiose-quinone oxidoreductase was also reported by Samejima et al. [98]. Ma et al. successfully derived a novel kinetic equation to explain these unique noncompetitive inhibitions [99]. Interestingly, Khindaria et al. [l001 found that lignin peroxidase can oxidize manganese in the presence of VA, whose oxidation by Lip H' was inhibited by Mn2+present in the reaction system. Thus, if there is such a reducing substance i n a lignin-degrading site, the VA cation radicals and lignin molecule cation radicals are scavenged by these reductants. As a rule, it is important for white-rot fungi to reduce the extracellular level of such a reductant, including oxalate, in some way so that lignin may be decomposed more efficiently.

B. Manganese Peroxidase Manganese peroxidase (MnP) was first discovered by Kuwahara et al. [4]. Purification and characterization of MnP were described by Glen and Gold [ 1011 and Paszcsynski et al. [102]. Like Lip, MnP is a hemoprotein of about 46,000 daltons with protoporphyrin IX as a prosthetic group. This enzyme catalyzes oxidation of Mn'+ in the presence of hydrogen peroxide to form Mn7+ as a direct oxidant, which is reduced back to Mn" substrate levels after oxidation of lignin and other oxidizable organic substrates present (Fig. 10). Thus, manganese acts as a redox mediator like VA. However, Mn7'. is not strong enough to oxidize the nonphenolic lignin moiety. Nevertheless, in the presence of Tween 80 (as a source of unsaturated fatty acid) in the reaction mixture, nonphenolic lignin substructure was cleaved by MnP 11031. Similarly, polycyclic aromatic hydrocarbons such as phenanthrene was oxidized to yield diphenolic acid by MnP catalysis [104]. It appears that MnP is more widespread than Lip among white-rot fungi and plays a key role in the initial stages of lignin catabolism. At present, MnP has been considered enzyme with the most potential in the biobleaching of kraft pulps, because this enzyme system from Phanerochaefe chtysosporium has been shown to break down high-molecular-weight chlorolignin [ 1051 and because a good correlation between fungal biobleaching activity and MnP activity was also observed 132,1061. Thus, the MnPsystemsappear to bemoreimportantthan Lip for practical application [ 1071. Actually, Kondo et al. [l081 succeeded in brightening hardwood kraft pulp by 43 points by six treatments with the MnP system and alkaline extractions. However, the mechanism for the enzymatic bleaching reactions catalyzed by the MnP system is not clearly understood.

C . Laccase Laccase (EC 1.10.3.2), secreted as an extracellular enzyme from white-rot fungi[ 109,1101, is one of the copper-containing enzymes, catalyzing oxidation of p-diphenol to p-benzo-

Oxidized Lignin FIGURE 10 Proposed mechanisms for the oxidation of HzOz and Mn" 141.

of lignin catalyzed by MnP in the presence

Shimada

558

Hattori and

quinone in the presence of dioxygen. Laccase (Lac) has been called polyphenol oxidase, or phenolase, similar to urushiol oxidase isolated from the Japanese lacquer tree. Like the lignin peroxidase, Lac is not specific to organic substrates and oxidizes not only phenolic compounds such as lignin-related compounds, coniferyl alcohol, vanillic acid, and p-cresol, but also ascorbic acid, p-phenylenediamine [ 109,tl l], and 2,2'-azinobis(3-ethylbenzthiazoline-6-sulfonate) (ABTS) [ 11 21. However, it differs from tyrosinase (EC 1.14.8. l), which is a copper-containing o-diphenol oxidase with monooxygenase activity acting on tyrosine. Although the correlation between occurrence of Lac and white-rot fungi has been well known as the Bavendamm reaction since 1928 [ 1131, the direct participation of Lac in lignin biodegradation has been questioned, because the Lac system preferentially repolymerizes lignin rather than depolymerizes it. Furthermore, the enzyme is incapable of oxidizing nonphenolic moieties oflignin because of the low redox potential of the oxidized enzyme. However, Bourbonnais et al. first reported that Lac catalyzes both the preferential depolymerization of lignin [ 1141 and oxidation of the nonphenolic lignin model substrate, [115], i.e., VA, in the presence of ABTS(Fig. l l ) , whichmightserve as the electron carrier like VA or Mn(I1) involved in the Lip or MnP system, respectively. They found that the two Lac isozymes gave similar qualitative effects on kraft lignin and residual lignin in kraft pulp, with no evidence for a marked preference to depolymerization by either isozyme alone. However, the presence of the mediator ABTS prevented and reversed the polymerization of kraft lignin by either of the Lac isoforms. When the delignification of hardwood and softwood kraft pulps with the two isozymes and the mediator was compared, either Lac was found to be able to reduce the kappa number of pulp, but only in the presence of ABTS. Interestingly, Eggert et al. have recently found that a secondary metabolite, 3-hydroxyanthranilate, which was produced by the white-rot fungus Pycnoporus cinnabarinus, served as a mediator in Lac-catalyzeddegradation of a nonphenolic P-0-4-type lignin model compound and synthetic lignin [ 1161. They have further reported that laccase-less mutants of Pycnoporus cinnabarinus were greatly reduced in their ability to metabolize I4 C-ring-labeled DHP [ 1171. In fact, this white-rot fungus does not contain Lip nor MnP. Thus, it is clearly demonstrated that Lac is essential for lignin degradation by white-rot fungi which lack Lip and MnP. Alternatively, Marzullo et al. [ 1 1 81 reported that the cooperative actions of Lac and VANA oxidase systems prevented the repolymerization of phenolic compounds and reduced the molecular weight of lignosulfonates to a significant extent. In this system, the FAD-containing VA oxidase can prevent the recondensation of the phenolic reaction products, since the flavoprotein oxidase, like cellobiose-quinone oxidoreductase in Phanerochaete chrysosporiunz, is able to reduce laccase-generated quinoids and phenoxy radicals [ 1191. At present, no unifyingconceptwhich may explain a wide variety ofoxidative reactions catalyzed by Lip, MnP, andLachas yet been established, but it is becoming

0 2

x

H20

x

accaseox

~

B

~

s

o

x

~

L

i

g

n

i

n

mediation Oxidized Lignin

FIGURE 11 ABTS-mediated oxidation of lignin catalyzed by laccase [ 1141

Degradation Bioremediation Relation to ofinLignin

559

common that each of these three lignin-degrading enzymes seems to require its own redox mediator. And it would be interesting to find a possible occurrence of a common mediator for Lac systems among white-rot fungi, if any. As to industrial use of an enzyme system for pulping and kraft pulp bleaching, it is still challenging how to solve serious cost problems in practice. However, there are only one or two successful examples of introducing a soft enzyme catalyst into a hard pulping industry. The most successful one is the utilization of commercial lipase [ 1201 to eliminate the pitch problem in papermaking and xylanases [ 121,1221 for bleaching of unbleached pulps at the industrial and commercial level.

D.

Active Oxygen Species

1. Source of HydrogenPeroxide Since Lip and MnP requirehydrogenperoxideto initiate theone-electronoxidation,a source of hydrogen peroxide must be assured and the enzyme system should be extracellular. Although in early work enzyme systems such as glucose oxidaseand NADH oxidase were postulated as plausible candidates for supplying hydrogen peroxide, nowadays extracellular glyoxal oxidase (GLOX) [ 1231 and aromatic alcohol oxidase (AAO)[ 124- 1261 are receivingmuch attention. Both of themcatalyzeoxidation of the substrates in the presence of dioxygen, yielding hydrogen peroxide and other products. However, quite recently, Urzura et al. [ 126aI reported a new type of hydrogen peroxideproductioncatalyzed by MnP system in the presence of oxalate, glyoxylate, and dioxygen. It has been reported that during oxidative breakdown of p-0-4 model substrate, the cleavage product glycol aldehyde may act as the source of approximately 2 equivalent hydrogen peroxides required for the subsequent oxidation of Lip substrates [ 1231. Consequently, the cleavage of arylglycerol P-aryl ether structures by a ligninolytic system with Lip andGLOXrecycleshydrogenperoxide to supportsubsequentcleavage reactions. However, there may be other routes for production of hydrogen peroxide in the extracellular site: ( I ) autoxidation of reductants such as ferrous oxalate complex by dioxygen and (2) hydrogen peroxide secreted from an intracellular site to the extracellular site without decomposition by catalase. 2. Formation of Free Oxygen Radicals Oxygen oxidation of organic compounds involves in principle the three following modes: ( I ) activation of molecular oxygen; (2) activation of organic substrates, and (3) activation of both dioxygen and substrates. Either of these cases also is applied to both biological andchemicaloxidation of lignin and otherwoodcomponents. Inan early lignin biodegradation study, Shimada [ 1271 first proposed that radical processes are important as a general principle in both lignin polymerization(biosynthesis)and its depolymerization (biodegradation), pointing out an important concept of non(stere0)specificity of enzymes, including peroxidase and Lac together with free active oxygen species. As shown in Fig. 12, those active oxygen species are generated in the process of reductive oxygen metabolism or by light irradiation of dioxygen, with a sensitizer in the case of singlet dioxygen.Amongthem,hydroxyl radical species are the most reactive agents and have long been receiving much attention concerning fortuitous breakdown of biological components. Studies on Lip and MnP have revealed that not the free species

560

Hattori and Shimada

HO'

1

0 2

H++0; H++H02- 0;

It

H++0,

H++OH-

+

H'

FIGURE 12 Various forms of active oxygen species produced during reductive oxygen metabolism or light (hv) irradiation. The dashed arrow indicates a conceptual path for formation of the two-electron-deficient oxygen atom of ''oxene,"which is almost equivalent to the active oxygen species in high-valent oxo-ironporphyrincomplex(compound I of peroxidaseandcytochrome P-450).

but the enzyme-bound"oxene" in iron porphyrincomplex is an active form to initiate oxidation of lignin. However,quite recently, the activeoxygenspeciessuchassuperoxideanion and hydroxyl radical were demonstrated to be formed in the presence of oxalate in both LIP and MnP systems [ 128,1291 (Fig. 13). Although superoxide anion radicals are not reactive enough by themselves to decompose lignin [ 1301 but break down the cellulosic fiber [ 13 l], free-radical reaction mechanisms [68,132] involving hydroxyl radical species have been recommended. Then oxidation of oxalate, yielding formate radicals and consequently superoxide radicals in the presence of dioxygen, has received much attention as a source of reducing power of dioxygen to generate hydroxyl radicals ultimately. However, it is unlikely that such nonspecific reactions are major and even controlled for specific purpose. These free oxygen radical species may be important for bioremediation of pollutants in natural environments 119,1331. Alternatively, Wood [ 1341 has postulated that cellobiose dehydrogenase plays a key role in reducing Fe(II1) to Fe(II), which serves as a reductant of both dioxygen to form superoxide radicals in the presence of oxalate and hydrogen peroxide to generate hydroxyl radicals (Fig. 14). In this case, iron oxalate complex serves as a mediator. Thus, during white-rot wood decay processes, ligninolytic enzymes (Lip, MnP, and Lac) and cellobiose

FIGURE 13 Proposed mechanisms for the production of active oxygen species mediated by veratryl alcohol cation radical and oxalate in the LIP system 1451.

Degradationof Lignin in Relation to Bioremediation Cellodextrin

2Fe(III)-Oxalat

V

Lactone

561

20;-HzOz+Oz

1/

LFW)

ZFe(I1)-Oxalate

Cellobiose dehydrogenase

Autoxidation

FIGURE 14 Cellobiose dehydrogenase as an agcnt for hydroxyl radical production [ 1341.

dehydrogenase/cellobiosequinoneoxidoreductaseseem to degrade lignin and cellulose cooperatively [ 1351. Another possibility for the generation of active oxygen species is oxidation of oxalate in the presence of dioxygen by MnO,. which is deposited in white-rot wood decay sites as a result of disproportionation from Mn3+ produced by the MnP system; this may play an important role in generation of formatcradicals and consequentlysuperoxide anion radicals, which reductively dehalogenate carbon tetrachloride [ 1361, and other recalcitrants arc broken down by hydroxyl radicals concomitantly produced in this system (Fig. IS). Actually, reduction of MnO, by reducing compounds such as oxalate and phenolic acids derived from lignin is reported to be essentialforsupplyingmanganousions to plants through their recycling processes [ 1371. The oxalate/MnO, system under aerobic and anaerobic conditions is an interesting model system to appreciate such bioremediation processes from geochemical and ecological viewpoints.

W.

BlOMlMETlCAPPROACH

We reported that synthctic and natural iron-porphyrins mimicked the Lip enzyme catalyzing Ca-CP bond cleavage and ring-opening reactions of the nonphenolic lignin model substrates [S.?,1381, proposing the one-clectron oxidation mechanism for the cleavage rcaction (61. Since then, ligninase-mimetic systems have been receiving widespread interest from basic and practical viewpoints 1 1 39- 14 I 1. Recently, a wide variety of model systems with water-soluble synthetic metalloporphyrins, phthalocyanines, and other metallocomplexes have been reported for lignin deg-

Goo-

y cooc0,-+c0,"~

Mn (11)

1 .

Reactive Species

FIGURE 15 A proposcd role o f oxalate i n brcnking down ol'thc recalcitrant lignm with the MnO,/ oxalntc system undcr acrobic conditions.

Hattori and Shimada

562

radation. Among them, polyoxometalate (POM) complex has been focused on as quite a promising system in many respects for bleaching unbleached kraft pulps [ 1421.

A.

Mechanism for Lip-Mimetic Oxidation

Since the hemin catalyst simulates the essential features of Lip for C-C bond cleavage, D-retention, and oxygenation, the model system with hemin has been established as relevant to the ligninase system for elucidation of biochemical mechanisms of lignin degradation. In view of the high redox potential of the oxenoid species in oxo-iron porphyrin complex 11431 and the findings obtainedwith the biomimeticsystem, the one-electron transfer mechanism has been proposed (Fig. 16) [6] to explain the Ca-Cp bond cleavage of (A) concomitant with hydroxylation to form (C), in which formation of a cation radical (Y) by one-electron abstraction from the substrate (A) with the oxo-iron porphyrin complex (X) is the initial step of the reaction [6]. The cation radical (U)formed in ring A or B undergoes homolytic or heterolytic C-C bond cleavage, yielding anisaldehyde (B) as one product concomitant with deprotonation and hydroxymethylbenzyl radical as the counterpart (Z). The carbon-centered radical ( Z ) is then attacked by molecular oxygen at a diffusion-controlled rate [ 1441. Thus the dioxygen adduct or the related hydroperoxide intermediateformedmighteventuallybeconverted to the hydroxylatedproduct ( C ) by the reduction with the catalyst (route a in Fig. 16). Alternatively, the radical intermediate ( Z ) escaping from dioxygen attacks, particularly under anaerobic conditions and undergoing a second electron abstraction to form the benzyl cation and the subsequent hydroxylation with water orthe oxo-iron complex, whose oxygen atom might be exchanged with water [145,146], could explain the incorporation of hydroxylicoxygenfromwater(routeb in Fig. 16). Anotherhydroxylationmode is explained by route c, whereoxygen in the oxo-ironcomplex is transferred via an OH ligand, which was not exchanged with water [ 1471. The proposed radical process is noteworthy since neither hydrogen abstraction nor a 1,2-shift of hydrogen atom is involved in this hydroxylation. This is in sharp contrast to the usual radical reactions and the hydroxylations catalyzed by P-450 model systems 11481. Taking the evidence together, a one-electron oxidation mechanism [ 1391, proposed from the biomimeticmodel, rationally explains many interesting observed traits of Lip affecting p- 1 substrate. Likewise, Hattori et al. [ 1381 demonstrated the ring-opening reaction of VA substrate with hemin catalyst, which clearly mimicked Lip catalysis. One oxygen atom from either dioxygen or water was incorporated into the ring-opened product, and the possibility of participation of superoxide anion radical in the aromatic ring opening of VA proposed by Schoemaker et al. [ 1491 was eliminated with our ''0 experiment (Fig. 6).

B.

Metalloporphyrins for Ligninase Model Systems

Figure 17 shows the chemical structures of iron porphyrin and phthalocyanine complexes which have been used for modeling the lignin-degrading enzymes. Dolphin'sgroup [ 1501 synthesized sterically protected and water-solublemetalloporphyrins. They have reported that their iron-porphyrin catalysts mimicked Lip activity in the ring opening a s well as Ca-Cp bond cleavage of thc dimeric and monomeric lignin model compounds [ 15 I , 1521. It is noteworthy that, quite recently, Mansuy's group 153.154] has demonstrnted that similariron-porphyrinsalsocatalyzed thering opening of 1,2-dimethoxyarencs withan

X

cy

0 -

0

X

0

r

A

Degradation of Lignin in Relation to Bioremediation

I

.

0

2

563

564

Hattori and Shimada

FIGURE 17 Chemicalstructuresusedforoxidation of lignin. A. hemin; B, meso-tetra(2.6-dichloro-3-sulfonatopl~e~~yl)-P-porphinato iron (111); C. octacarboxyphthalocyanine: D. tetrasulionyl-

phthalocyanine.

electron-withdrawing group which are not oxidized by Lip. The results indicated that the muconic acid esters a s the ring-opening products are even selectively produced in the 3S% yield based on the substrates, which might be promising for further application to the ring openings of modified kraft lignin. They also succeeded in synthesizing binuclear [Mn (11) (bipyridyl) Fe(II1) (porphyrin)] complex as a MnP model which showed the first evidence for reversible formation of Mn (111). Alternatively, metallophthalocyanines (azaporphyrins) have also rcceived much attention as catalysts mimicking lignin-degrading enzymes. Zhu and Ford have reported that VA was oxidized in the presence of dioxygen or hydrogen peroxide in water [ 1 SS, 1561. Tanihara et al. [ 1571 reported oxidative cleavage of p-0-4 lignin nlodel compounds with water-soluble iron-octacarboxyphthalocyaninei n the presence of terr-butylhydroperoxide in the alkaline medium. Their phthalocyanine system simulated Lip activity in C a C p bond cleavagc,p-ethercleavage.ring-openingcleavagereactions. However, these water-soluble phthalocyanines were unstable and easily bleached and arc not useful for application of delignitication.

Degradationof Lignin in Relation to Bioremediation

C.

565

Other Model Systems

Instead of using the metalloporphyrin catalysts, Tung and Sawyer [l581 have found that bis(2,2-bipyridine) iron (11) and several similar complexes catalyzed oxidation of both VA and benzyl alcohol to form veratraldehyde and benzaldehyde, respectively. However, ligninolytic activity to cleave dimeric p-0-4 model compound was not confirmed. Alternatively, we have developed the MnP-mimetic system containing Mn(I1) and peracetic acid. The biomimetic system not only degraded Lip-resistant p-0-4 model substrate but also bleached kraft pulps [159,160]. As a result, a marked increase in brightness to 86 points has been successfully achieved [ 1611. However, this unconventional chlorinefreebleachingprocesscauseddecrease in depolymerization of cellulose.However, the cellulose degradation was significantly controlled by combination with the conventional method [ 161aI. Quite recently, Cui et al. [ 161b] have reported the hydrogen peroxide bleaching of pine pulp catalyzed by a binuclear Mn(III/IV) complex as a MnP-mimetic model system, which increased brightness without severe degradation of cellulose. Recently, we havefoundanother interesting unconventional reaction systemwith Mn(III), oxalate, and dimethyl sulfoxide (DMSO) [ 162,1631 and succeeded in cleaving the C a - C p and CB-ether bonds of the recalcitrant nonphenolic p-0-4 lignin model substrate which was not oxidized by Lip and MnP [75]. Since the cleavage reactions were dependent on either Mn(III), oxalate, or DMSO, the reaction mechanism is not simple enough to explain yet. However,oxalatemight firstbe oxidized by Mn(III),yielding formate radicals (carbon dioxide anion radicals) to reduce dioxygen to form superoxide anion radicals under aerobic conditions, and then a wide variety of active oxygen radicals are produced to attack lignin molecules, or formate radicals may be more directly involved in degradation of lignins under anaerobic conditions, because superoxide anion radicals are not reactive enough to break downthe a-keto-containing p-0-4 model substrates [ 1301, and no breakdown of the substrate occurred under 100% Oz. Alternatively, Hames and Kurek [ 1641 studied the effect of an MnO,/oxalate oxidation system on degradation of lignin in wood cell walls, and reported a strong decrease in the lignin content of P-0-4-linked guaiacyl and syringyl structures, the former being predominantly attacked. The main feature of this simple system is the very high specificity toward lignin oxidation, leaving the polysaccharide fraction almost untouched, although this system isnot ligninolytic because lignin wasmodified butnot depolymerizedinto lower-molecular-weight compounds.

D. Application of Biomimetic Systems Dolphin’s group [ 1501 used sterically protected and water-soluble metalloporphyrins for bleaching unbleached pulp. They reported that kraft pulp bleaching and effluent decolorization were clearly achieved with their “synthetic enzymes”; the reduction of the kappa numberfrom 21.1 to 13.8 or45% delignification wasachieved [ 1651. However,they reported that there are a number of problems to be solved for industrial application: ( I ) low yields of the porphyrin synthesis, (2) high costs of the catalyst and peroxide used, and (3) considerable damage of pulps during the bleaching, etc. Recently, Kurek et al. [l661 attempted to degrade isolated lignins by use of watersoluble hydrogen peroxide-resistant pentafluorophenyl porphyrin, and recognized that the spruce milled wood lignin underwent the Ca-Cp cleavage. They concluded that the ironporphyrin mimics the action of the fungal ligninolytic peroxidase but also is able to penetrate within the compact structure of wood and degrade lignin to a significant extent.

Hattori and Shimada

566

FIGURE 18 Proposedmechanismforthebleach and oxygen [ 1421.

of pulp catalyzed by polyoxometalate (POM)

Quite recently, Weinstock et al. [l421 have developed a promising pulp bleaching technology, based on the use of a polyoxometalate (POM) such as SiVW,,O,Oand oxygen. As the delignification reaction cycle shown in Fig. 18, the reaction system is reminiscent of the biomimetic system of Lac, which requires oxygen as oxidant. They claimed that the approach is unique by intrinsic design and implementation in that the versatile capabilities of a single POM species are utilized to facilitate multiple operations that sum to the selective conversion of wood pulp to paper and nontoxic products ( H 2 0 and CO,), using air as the oxidant and water as the only solvent.

E.

Bioremediation with Biomimetic Systems

Recently, bioremediation of xenobiotic pollutants with white-rot fungi has been receiving widespread interest, and such investigations were reviewed by Barr and Aust [ 1671. Sorokin et al. 11681 first reported the oxidativedechlorination and aromaticcycle cleavage of trichlorophenol by H,O, using iron tetrasulfophthalocyanine. Theyfurther found out that iron-tetrasulfophthalocyanine catalyzes the oxidation of polychlorinated phenols, including 2,4,6-trichlorophenoI as major pulp bleach effluent halogenated compound, to CO, as the ultimate degradation product in the presence of H,O,, although CO, accompanied the water-soluble major degradation products [ 1691. Since metallophthalocyanines are a cheap industrial dye product, they are readily available catalysts for oxidation of pollutants [ 1701. Thus, these catalytic systems may have a promising future for the oxidative removal of pollutants as an environmentally friendly biomimetic system.

V.

CONCLUDING REMARK

Since the first edition of this book was published in 1991, tremendous numbers of reports onmicrobial,enzymatic and biomimeticdegradation of lignin and seemingly unrelated compounds, including pollutants and xenotiotics, have been reported. Although the oneelectron oxidation mechanism for lignin degradation now allows us to understand more comprehensively oxidative degradation processes of lignin, it is still deceptively simple to apply, in practice, bioligninolytic systems to the pulp and paper industry. Nevertheless, new vital ligninolytic fungi were screened, investigated, and characterized. New aspects of radical reactions of lignin and other compounds catalyzed by manganese peroxidase and laccase in the presence of Tween 80 and ABTS, respectively, have been receiving much attention. Another intcresting aspect of free-radical reactions caused by ligninolytic systems has focused on the generation of superoxide anion radicals and hydroxyl radicals as the result of oxidative degradation of oxalate in the presence of dioxygen and ferric ions. These nonspecific radical species may play a key role i n degradation of lignocellu-

Degradation Bioremediation Relation to ofinLignin

567

losic components and in bioremediationofourpollutedenvironments by decomposing nonspecifically a wide variety of pollutants and xenobiotics. As to the biomimetic approach to lignin degradation, use of polyoxometalate complex may be promising for chlorine-free bleaching of unbleached kraft pulps, although water-soluble, stable, and cheap phthalocyanines may be more useful for removal of pollutants from industrial waste effluents in some ways. Consequently, studies of biodegradation of lignin are essential and becoming more and more important for understanding our ecosystem and preventing pollution problems on earth.

REFERENCES I. K. V. Sarkanen and C. H. Ludwig, Lignins: Occurrorlce. Fonncrtion, Structure t l l ~ dR ~ ( K tiom. Wiley-Interscience, New York, pp. 1-9 1 1 16 ( I97 1 ). I a. P. Ander and K. E. Eriksson. Svensk., 78: 643 (1975). 2. M. Tien and T. K. Kirk, Scirr~ce.221: 661 (1983). 3. J. K. Glenn, M. Morgan, M. B. Mayfield, M. Kuwahara. and M. H. Gold, Biochenr. Biophys. Res. Comnm., 114: 1077 (19x3). FEBS LcW., 160: 247 (1984). 4. M. Kuwahnra. J. K. Glenn, M. A. Morgan. and M. H. Gold. S . P. J. Kersten. M. Tien. B. Kalyanaraman. and T. K. Kirk. J. Bioi. C / I ~ I260: I I . 2609 (19x5). 6. T. Habe, M.Shimada. T. Okamoto, B. Panijpan,and T. Higuchi, J . C1zer11.Soc. Chern. c o l ~ l l l l L l l l . , 108.7: I323 (1985). 7. H. E. Schoemaker, P. J . Harvey, R. M. Bowen. and J. M. Palmer. FEBS Lett.. 183: 7 (1985). T. Higuchi, in Wood andCellulosicChemistry(D. N. S. Hon and N. X. M.Shimadaand 9.

IO. 1 I.

12. 13. 14.

IS. I 6.

17.

1x. I 9.

20. 21. 22.

13. 24. 25.

26.

Shimishi, eds.), Marcel Dekker, New York. p. SS7 (1991 ). T. K. Kirk and H . M. Chang. in Riotcxhrlology i n Pulp tuld Pnpc’r Mmzufc~ctwc(T. K. Kirk and H.-M. Chang. eds.). Butterworth-HeinetIlann, Boston. p. 666 (1990). M. Kuwahnra and M. Shirnada. in B i o t e c h o l o g y i r l Pulp m e / Ptrprr lnt/u.stry (M. Kuwahara and M. Shimada. eds.), Uni Publishers, Tokyo. p. 544 (1992). P. Ander. M. C. Ferreira, and J. C. Duarte, FEMS Microbiol. Rev. 13: 121 (1994). E. Srebotnik and K. Messner, B i o t e c ~ h ~ z o l oirlg ~Pulp m t l P q w r Imht.sr,y. Facultas Universitatverlag. Vienna, pp. 1-66 1 ( 1997). J. A. Buswell and E. Odier. CRC Ctit. Re\: Biorechrzol., 6 : 1 (1987). T. K. Kirk and R. L. Fnrrell, Annu. Ret: Microbiol., 41: 465 (1987). R. A. Blanchette. A m u . REI: Phytopcrthol.. 29: 38 I (199 l ) . B. Kurek, i n Pltrrlt Prrosidtrtr.sc~.s 1080-1900. lbpics t r n d Dc.tcrile.t/ Litc.rtrttrw 0 1 1 Mo/eclf/(lr; Hiochc~rrlic.crl,t r r l d Ph~.siologicc~/ Aspec~t.s(C. Panel, Th. Gaspar, and H. Greppin, e&.), University of Geneva. p. 139 ( 1992). M. H. Gold and M. Alic, Microhiol. Kc,\:, 57: 605 (199.3). M. Shimada. D. B. Ma, Y. Akamatsu. and T. Hattori, F E M S Microhiol. Rc.1:. 13: 285 (1994). D. P. Barr and S. D. Aust. Emiron. S c i . T~c~hrrol., 28: 78A (1994). L. Otjen. R. A. Blanchette. M. Eftland, and G. Leatham. Hol;fifor.sc.h.. 41: 343 ( 1987). A. Hatakka. FEMS Microbial. Ret:. I S : 25 ( 1994). K.-E. Eriksson and L.Vallander. i n Lignirr Biodc.Rrtrt/trtiorl: Mic~robio/ogy,C/lc,/ll;,ytl:\’ cl,l(/ Potcrltictl Appliccttiorl. Vol. 11 (T. K. Kirk, T. Higuchi, and H. M. Chang. e&.). CRC Press, Boca Raton, FL, p. 2 I3 ( 1980). S. S. Bar-Lev. T. K. Kirk, and H. M. Chang. nrpp;,65: 1 I (1982). I. Akamatsu. K. Yoshiharn, H. Kamishima. and T. Fujii, Moku:tri GtrMtr;,s/f;,-30: 697 ( 1984). G. F. Leatham. G. C. Myers, T. H. Wegner. and R. A. Blanchette. in Biotechtzo/o,yT ;!I P L ~ / / ) t r t d Pnpc.r Mtrr?ufirc.tlrrc (T. K. Kirk and H. M. Chang. eds.), Butterworth-Heinelnann, Bo+ ton. p. 193 ( 1990). M. Akhtar, M. C. Attridge.R. Blanchette, G. C. Nyers. M. B. Wall, M. S. Sykes, J. W, Koning, Jr.. R. R. Burgess. T. H. Wagner. and T. K.Kirk. in Bjotty./frlo/og? pit//) c l r l c /

Shimada 568

and Paper Munufactrrre (M.KuwaharaandM.Shimada,

Hattori eds.), Uni Publishers,Tokyo, p. 3

( 1992).

27. M. Akhtar, Holzforsch. 48: 194 (1994). 28. M. Akhtar, T. K.Kirk, and R. A. Blanchette, in Biotechnology in Pulp and Paper Manufacture (T. K. Kirk and H. M. Chang, eds.), Butterworth-Heinemann, Boston, p. 187 (1996). 29. T. Nishida, Y. Kashino, A. Mimura, and Y. Takahara, Mokutai Gakkaishi, 34: 530 (1989). 30.

K.Fujita,

R. Kondo, K. Sakai, Y. Kashino, T. Nishida,and

Y. Takahara, Tappi, 76: 81

( 1993).

31. N. Katagiri, Y. Tsutsumi, and T. Nishida, Appl. Environ. Microbiol., 61: 617 (1995). 32. M. G. Paice, L. Jurasek, C. Ho, R. Bourbonnais, and F. Archibald, Tappi, 72: 217 (1989). 33. Y. Kashino, T. Nishida, Y. Takahara,K.Fujita,R.Kondo,and K. Sakai, Tnppi, 76: 167 (1 993). 34. H. Hirai, R. Kondo, and K. Sakai, Mokuzai Gnkkaishi, 40: 980 (1994). 35. T. Fukuzumi, A. Nishida, K. Aoshima, and K. Minami, Mokuzai Gakkaishi, 23: 290 (1977). 36. S. Pallerla and R. P. Chambers, J. Environ. Sci. Health, A30: 423 (1995). 37. R. Lacknen, E. Srebotnik,and K. Messner, in Biotechnology in Pulp and Paper Industry (M. Kuwahara and K. Shimada, eds.), Uni Publishers, Tokyo, p. 45 (1992). 38. A. Hatakka, M. Valo, and P. Lankinen, in Biotechnology in Pulp and Paper Industry (M. Kuwahara and K. Shimada, eds.), Uni Publishers, Tokyo, p. 69 (1992). 39. H. Fukui, T. L. Presnell, T. W. Joyce, and H.-M. Chang, in Biotechnology in Pulp and Paper Industry (M. Kuwahara and K. Shimada, eds.), Uni Publishers, Tokyo, p. 75 (1992). 40. A. M. Calvo, J. L. Copa, J. B. Fermandez-Larrea. and A. E. Gonzalez, in Biotechnology in

41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

52. 53. 54. 55.

56. 57. 58.

Pulp and Puper Industty: Recent Advances in Applied and Fundumentul Resecrrch (E. Srebotnik and K. Messner, eds.), Facultas Universitatsverlag, Vienna. p. 253 ( 1 996). C.H. Kong,H. K. On,andC. H. Won,in Biotechnology i n Pulp and Paper Industry: Recent Advances i n Applied and Fundamental Reseurch (E. Srebotnik and K. Messner, eds.), Facultas Universitatsverlag, Vienna, p. 263 (1996). P. Manzanares, S. Fajardo, A. Cabanas, and C. Martin, in Biotechnology in Pulp and Paper Indust?: Recent Advances i n Applied und FundamentalResearch (E. Srebotnikand K. Messner, eds.), Facultas Universitatsverlag, Vienna, p. 267 ( 1996). J. D. Stahland S. D. Aust, in BiodegradutionofNitroaromatic Corrzpout1d.s (J. C. Spain, ed.), Plenum Press, New York, p. 1 17 (1995). J. Michels and G. Gottschalk, in Biodegradcrtion of Nitroaronrutic Conr1x)urrd.s (J. C. Spain, ed.), Plenum Press, New York, p. 135 (1995). E. De Jong, J. A. Field, and J. A. M. De Bont, FEMS Microbiol.Ret]., 13: 153 (1994). U. Tuor, K. Winterhalter, A. Fiechter, J. Biotechnol., 41: 1 (1995). M. Tien and T. K. Kirk, Science, 221: 661 (1983). M. Tien and T. K. Kirk, Proc. Natl. Acari. Sci.USA, 81: 2280(1984). J. K. Glenn,M. A. Morgan,M.B. Mayfield, M.Kuwahara,and M. H. Gold, Biochetn. Biophys.Res.Cotnmun., 114: 1077 (1983). M. H. Gold, M. Kuwahara.A.A.Chiu,andJ. K. Glenn, Arch. Biochertr. Bioplrys.. 234: 353 (1984). L. Anderson, V. Renganathan,A. A. Chiu. T. M. Lohr, andM. H. Gold, J. B i d . c/7eI?l., 260: 6080 ( 1985). D. Kuila, M. Tien. J. A. Fee, and M. R. Ondreas. Biochetnistry, 24: 3394 (1985). M. Shimada, T. Habe, T. Umezawa, T. Higuchi.and T. Okamoto. Bioche~n.Biophy.~. Res. Conrnzun., 122: 1247 (1984). K. E. Hammel, M. Tien. B. Kalyanaraman. and T. K. Kirk, J . B i d . Chen~..269: 8343 ( 1986). T. Umezawa, and T. Higuchi, FEBS Lett., 218: 255 ( I 987). K. E. Hammel, K. A. Jensen, Jr.. M. D. Mosuch, L. L. Landucci, M. Tien, and E. A. Pease, J. Biol. Chetn., 268: 12274 ( 1993). T. Umezawaand T. Higuchi, Mokuxi Gtrkknishi, 35: 1014 (1989). K. E. Hammel and M. A. Moen, E~zzynzeMicro/?. T e d ~ l d .1, 3 : 15 (1991).

Degradation Bioremediation Relation to ofinLignin

569

59. H. Wariishi,K. Valli, and M. H. Gold, Biochenz. Biophys. Res. Cornrnur7., 176: 269 (1991). 60. V. Renganathan, K. Miki. and M. H. Gold, Arch. Biochern. Biophys., 241: 304 (1985). 61. I. G. Gazarian, L. M.Lagrimini, S. J. George,andR. N. F. Thorneley, Biockenz. J., 320: 369 ( 1996). 62. T. L. Poulos, S. L. Edwards, H. Wariishi, and M. H. Gold, J. Biol. Chem., 268: 4429 (1993). 63. K. E. Hammel, M. Tien, B. Kalyanaraman, and T. K. Kirk, J. Biol. Chern., 260: 8348 ( 1985). 64. V. Renganathan, K. Miki,and M. H. Gold, Arch. Biochern. Biophys., 246: 155 (1986). 6.5. F. Nakatsubo, I. Reid, and T. K. Kirk, Biochern. Biophys. Res. Comrnurz., 102: 484 ( 198 1 ). 66. T. K. Kirk, F. Nakatsubo. and I. D. Reid, Biochen7. Biophys. Res. Cormzurz., 111: 200 (1983). 67. P. Hall, Erz,-?,rne Microhinl. Techno/., 2: 179 ( 1980). 68. L. J. Forney, C. A. Reddy,M. Tien, and S. D. Aust, J. B i d . Chenz.. 257 1 1455 (1982). 69. H. Kutsuki andM. H. Gold, Biochenl. Biophys. Res. Cornrnun.. 109: 320 (1982). 70. H. lshikawaand T. Oki, Mokrczni Gnkkaishi, 12: 101 (1966). 71. T. Fukuzumiand T. Shibamoto, Mokuzcri Gnkknishi, 11: 248 (1965). 72. T. Fukuzumi, H. Takatsuka, and K. Minami. Arck. Biochern. Biophys., 129: 396(1969). 73. A. Enoki, G . P. Goldsby, and M. H. Gold, Arch. Microbiol.. 129: 141 (1981). 74. T. Umezawa and T. Higuchi, FEMS Microbiol. Lett., 26: 123 ( 1 985). 75. T. K. Kirk, M. Tien, P. J. Kersten, M. D. Mozuch, and B. Kalyanaraman, Biochern. J.. 236: 279 (1 986). 76. K. Miki, V. Renganathan,M.H.Gold, Biochemistry, 25: 4790(1986). 77. T. K. Kirk and M. Tien, in Recent Advnrzces i n Ligrzirz Biodegrudotior7 Resecrrch (T. Higuchi, H.-M. Chang, and T. K. Kirk, eds.), Uni Publishers, Tokyo, p. 233 (1983). 78. T. Umezawa, Wood Res., No. 75: 21 (1985). 79. T. Umezawa and T. Higuchi, in Erzzyrnes irz Bion~cr.s.sCm~versiot~ (G. F. Leatham and M. E. Himmel, eds.), American Chemical Society, Washington, DC, p. 236 (1991). 80. T. Hattori and T. Higuchi, Mokuzcri Gnkknishi, 37: 542 (1991). 81. S. Yokota, T. Umezawa, T. Hattori, and T. Higuchi, Mokuzcri Gctkke~ishi,37: 644 (1991). 82. C. L. Chen, H.-M. Chang, and T. K. Kirk, J. Wooe/ Chem. Techr7ol., 3: 35 (1983). 83. T. Umezawaand T. Higuchi. FEBS Lett., 182: 257 (1985). 84. T. Umezawaand T. Higuchi, FEBS Lett.. 218: 255 (1987). 85. M. Shimada, T. Hattori, T. Umezawa, T. Higuchi,and K. Uzura. FEBS Lett., 221: 327 (1987). 85a. T. Hattori, Wood Res., No. 78: 15 (1992). 86. T. Tokimatsu. T. Umezawa. andM. Shimada, Hol;for.sch., 50: 156 (1996). 87. T. Tokimatsu. S. Miyata,S.-H. Ahn. T. Umezawa, T. Hattori,andM.Shimada, MokLczcri Gukkcrishi, 42: 173 ( 1996). 88. I. Chet, J. Trojanowski, and A. Hutterman, Microbiol. Lett., 29: 37 (1985). 89. J. K. GlennandM. H. Gold, Appl. Environ. Microbial.. 45: 1741 (1983). 90. D. K. Joshi and M. H. Gold, Biochernistn), 33: 10969 ( 1994). 91. J. A.Bumpusand S. D.Aust, Scierzc~,228: 14 (1985). 92. P. J. Harvey. H. E. Schoemaker, and J. M.Palmer: FEBS Lett., 195: 242 (1986). 93. K. Valli, H. Wariishi. and M. H. Gold, Biochemistry, 29: 8535 (1990). 94. A. Khindaria, T. Grover. and S. D. Aust, Bioc.h~rrzi.str~\;34: 6020 (1995). 95. A. Khindaria, 1. Yalnazaki, and S. D. Aust, Biocl~~wzi~try. 35: 6418 (1996). 96. Y. Akamatsu, D. B.Ma. T. Higuchi, and M. Shimada. FEBS Lett., 269: 261 (l9c)O). 97. J. L. Popp. B. Kalynaraman,and T.K. Kirk. Bioclwrnistn, 29: 10475 ( 1990). 98. M. Samejima and K. E. Eriksson, FEBS Let/.. 2Y2: 15 1 ( 1991). D. B. Ma. T. Hattori, Y. Akomatsu. M. Adachi. and M. Shimada, B i o s c i . H i o c ~ h c ~ r r Biorec4r. 56: I378 ( 1992). 100. A. Khindaria. D. P. Barr,and S. D. Aust. Uioc~lrc~rr7i.str~~. 34: 7773 ( 1995). 1 0 1 . M. H. Gold and J. K. Glenn, Metl7otl.s Erzzyno/., 161: 258 (1988). 102. R. L. Paszcsynski, R. L. Crawford.and V. B. Huynh. Mc,thoc/.s En:yrzol. 161: 264 (1088). 103. K. A.Jcnsen, Jr.. W. Bao, S. Kawai. E. Srebotnik,and K. E. Hummel. App/. Erl17irorz. Mic,robio/.. 62: 3679 ( 1996). 99.

no/.,

570

Shimada

Hattori and

104. M.A.MoenandK. E. Hammel. Appl. E I I I ~ ~ Micw)hiol. OII. 60: 1956 (1994). I 05. R.Lackner, E. Srebotnik,and K. Messsner. R i o c h w ~ .Biophys. RPS. C o t t ~ t t ~ u ~17H: r . , 1092 (1991).

106. R. Kondo. K . Kurashiki, and K. Sakai, AppI. E t ~ v i r o t Micro/?iol., ~. 60: 921 (1994). 107. F. S. Archibald. Appl. E t ~ ~ ~ i Microbiol., rm. SH: 3 101 ( 1992). 108. R. Kondo, K. Harazono, and K. Sakai. Appl. Etlvirotl. Microhiol.. 60: 4359 ( 1994). 109. G. Gahraeus and H. Ljungren, Riockerrt. Hiophys. Acttr., 46: 22 (1961). I IO. T. Nakamura, Hiochetn. Biopllys. A c ~ N .3 ,0 : 538 ( 1958). 111. A. I. Yaropolov. 0. V. Skorobogat'Ko. S. S. Vartanov,and S. D.Varfolotneyev. Appl. Biochertt. Bioteclrrlol., 4Y: 257 (1994). 112. E. Matsumura. E. Yamamoto. A. Numata, T. Kawano, T. Shin, and S. Murao, Agric. Biol. Chern., SO: 1355 (1986). 113. W. Bavendamm, Z. Pfltrrl:eukr:, 38: 257 ( 1 928). 114. R.Bourbonnais,M.G.Paice, 1. D. Reid. P. Lanthier.andM.Yaguchi, AppI. Emiron. Micwhiol.. 61: 1876 (1992). 115. R. Bourbonnais and M. G. Paice. FEBS Lett.. 267: 99 (1990). 116. C. Eggert, U. Temp. J. D.D.Dean.andK. E. L. Eriksson, FEHS Lett., 391: 144 ( 1996). 117. C. Eggert, U. Temp, and K. E. L. Eriksson. FEBS Lc~tt..4407 89 ( I 997). 118. L. Marzullo. R. Cannino, P. Giardina. and M. T. Santini. ./. Hiol. Chcw~..270: 3823 ( 1995). 119. M. Samcjima and K. E. L.Eriksson. Ertt: J. H i o d ~ m ~2202: ., 103 (1992). I 20. K. Hata. M. Matsukura. H. Taneda, and Y. Fujita, i n E t ~ : y t ~ c t v f i Prtlp ) r m t l Ptrpc~rP r o ~ ~ e s s i ~ ~ g (T. W. Jclferies and L. Viikari, cds.), ACS. Symp. Scr.. Washington. DC, p. 279 (1996). 121 L. Viikari.A.Kantelinen.J.Lundzuist,andM.Linko, FEMS Micwhiol. Re,:. 13: 335 ( 1994).

122. J. Buchert. T. Oksanen. J. Pere, M. Siika-aho. A. Suut-nakki. and L. Viikari. in Tric/rotIer)m J; Clioc.ltrc/irou (E. Harman and C. Kubicek. eds.). Vol. 2. pp. 343-363 ( 1998). 123. K. E. Hammel, M. D. Momch. K. A. Jenscn, Jr.. and P. J . Kersten. Rioc~/tc,rtti.sft~\..33: 13349 ( 1994). 124. P. Sannia, P. Limongi, E. Cocca. F. Buonocore. G. Nitti, and P. Giarclina. B i o c ~ l r i t t r .Biophys. Acttr. 107.1: 114 (1991). . . 603 (1992). 125. F. Guillen. A. T. Martinez, and M. J. Martinez. Errt: J . B i o c ~ h c ~ t r20Y: 126. H. G. Beckcr and A. P. Sinitsyn. H i o t c c l r r ~ o l .Lctt.. IS: 289 ( 1993). I26a. U. Urzura. P. J. Kersten. and R. Vicuna. Appl. ErI\-iro/r. Mict.ohio1.. 64: 68 ( 199X). Microbiology, C / r t ~ t t ~ i . tsr t/ rtc~/ ~PotctrtictlA/)l)lic~trtiotr.s. 127. M. Shimada. i n Ligtliu BiotI~~,~rtrtltrtiotr: Vol. 1 (T. K.Kirk. T. Higuchi.andH.-M.Chnng. eds.), CRCPress.BocaRaton. FL. p .

195 ( 1980).

Biophy.~..2YH: 480 128. D.B.Barr. M. M. Shah. T. A.Grover. ;und S. D.Aust. A t d r . Bioc~hc~ttr. ( 1992). 129. A.Khind:tria.T. A . Grover.and S. D. Aust. A/.rh. H i o c h c t t l . Riophys.. -314: 301 (1994). 130. R. Barkhau. J. Bastian. and N. S. Thornpson. 7irppi, 64: I I O (1985). 131. N . S. Thotnpson and H. M. Corbett, Ttrppi, 64: 126 ( 19x5). Microhiolo,q~:Chc~r~ri.vtt~\. rrtrrl / ' o t t , t ~ t i ~ r / A / ) / ) l i ( ' ( / t i o / ~ . s . 132. M. Shimadn. i n Ligrli~rBiorlc~,~/.rrcltrtiot~: Vol. I (T. K. Kirk. T.Higuchi.andH.-M. Chang. eds.), CRC Press.Boca R a t o n . FL, p. I95 ( 19x0). ~ ~ o u t / P(rpcr / t r d ~ t . s t / ; \(E. ~ Srcbotnik ancl K. Messncr. 133. M.Shimada. i n / ~ i o t c , t . / r r ~ o /it1o ~Pull) e h . ) . Faculto Univcrsistatverlag. Vicnna. p . 17 ( 1997). 134. P. M.Wood. FEMS Mic.rnhio1. KcI:. 1.3: 3 13 ( 1994). 135. P. Ander. FEMS Mic.rohiol. R1.l:. 13: 297 ( 1994). ., 887 ( 1993). 130. J. D. Shoh and S. D. ALISI.H i o c , h r t r , . Biophys. lic,s. C o r t ~ t t ~ r t t ~191: 137. A . T. Stone ond J . J . Morgan. E ~ ~ \ ~ i r oSIcti.. T v c . / ~ r r o / . . IS: 617 (1984). I3X. T. Hattori. M. Shimada. T. Umczawa. T. Higuchi. M. S. A. Leisola. ancl A. Fiechtcr. Agric,. / W . C ' / l ~ ~ t / I x. , : x79 ( 198%). 139. M. ShimaJn. T. Habc. T. Higuchi. T. Okamoto. ancl 13. l'anijpan. H o l ; f o r . v t ~ / I . , 41: 3-77 (19x7).

Degradation of Lignin in Relation to Bioremediation 140. 141. 142. 143. 144. 145. 146. 147. 148. 149. 150. 151.

152. 153.

154. I SS. 156. 157. 158. 159. 160. 161. 161a.

161b.

162. 163. 164. 165.

166.

571

M. Shimada, Mokuxri Gtrkkcrishi, 3 7 1 103 ( l99 1 ). B. Meunier, in Mrtcr/loporph~rir~.s i n Crrtcr1yticOxit1rrtion.s. p. I33 ( 1994). I. A. Weinstock, R. J. Atalla, R. S. Reiner, M. A. Moen, K. E. Hammel. C. J. Houtman, C. L. Hill, and M. K. Harrup, J . Mol. Crrtal., A. C/?em.,116: 59 (1997). C. K. Chang and D. Dolphin, in Bioor\qcrrric C/remi.stry.Vol. IV (E. E. Tamelen. ed.), Academic Press, New York, p. 37 (1978). B . Mailard, K. U. Ingold,and J. C. Scaiano, J . AI^. Chen~.Soc.. 105: 5095 (1983). J. T. Groves. R. C. Haushalter. M. Nakamura. T. E. Nemo, and B. J. Evans. .1. Am. Chern. Soc., 103: 2884 (1981). R.E. WhiteandM. B. McCarthy, J . Am. Chew. Soc., 106: 4922 (1984). J. A. Smegal. B. C. Schardt, and C. L. Hill. J . Anr. Chcwr. Soc., 105: 3510 (1983). C. K. ChangandM. S. Kuo, J . Am. Cherlr. Soc., 101: 3413 (1979). H.E. Schoemaker. P. J. Harvey, J. M. Palmer, and H. J. M. Bosman. i n Proc. Rio-Orqrrrlic Heterocycles (H. C . Van der Plas et al.. eds.), Elsvier, Amsterdam, pp. 297-302 (1986). D. Dolphin. T. Nakano, T. E.Maione, T. K. Kirk. and R. Farrell, in Lignirl EIrzyrric c r r d Micro1icrl Degrad[ctiorr (E. Odier. ed.), INRA-CBAI, Thiverval-Grignon. p. 157 ( 1987). F. Cui, D. Dolphin. T. Wijesekera, R. Farrell, and P. Skerker, in Biotcd~/~o/o,qy irr Pulp trt1t1 Pcrpclr Mmufacttrre (T. K. Kirk, and H.-M. Chang. eds.). Butterworth-Heinemann. Boston. p. 48 I (1990). F. Cui and D. Dolphin, C m . J . Cherrl., 73: 21 53 ( 1995). I. Artaud,H. Grenberg. K. Ben-Aziza.and D. Mansuy, J . Chrm. Soc.. C h ~ r r Cowuluu., ~. 1992: 1036 (1992). I. Artaud, K. Ben-Aziza,and D. Mansuy, J . Org. C ~ I U ~ 5N: I . 3373 , (1994). W. Shu and W. T. Ford, J . Mol. Cattrl.. 78: 367 ( 1993). K. W. Hampton and W. T. Ford. J. Mol. C~tctl.A. Chon.. 113: 167 ( 1996). Y.Tanihara. T. Hattori. H. Shirai. and M. Shimada, Wood Rc.~..No. 81: I 1 (1994). H. C. Tung and D. T. Sawyer, FEBS Lett.. 3 / 1 : 165 (1992). M. Nakagawa. M . Shimada.and T. Higuchi, in Proc. .<4t/1L i g t ~ Syrtlp.. i~ Nagoya. Japan, pp. 109-1 12 (1989). M. Shitnada, M . Nakagawa. T. Hattori. and T. Higuchi. Mokrr:tri Gukktri.shi. 3-5: X59 ( 1989). H. Asaoka. T. Shigcmoto. T. Higuchi. andM.Shimada, PatentNo. 1-38272. (1989). M. Shimada. T. Shigemoto, T. Hattori, M. Takano, K. Saitaka. and T. Higuchi. in Plnc.. loth ISWPC. Yokohama. Japan. pp. 1-562-1-505 ( 1999). Y. Cui, P. Puthson. C.-L. Chen, J. Gratzl. A. Kirkman. andR. Patt. in Pro(.. IOth ISWPC. Yokohama, Japan. pp. 1-256-1-26 1 ( 1999). T. Hattori. N. Kontlou,and M.Shimada. Mokrrzrri Gtrkkrrishi. 41: I176 (I995). N. Kondou. T. Hattori.and M. Shimada. i n Proc. Y t h ISWPC'. Montreal. Canada. p p . 491-49-4 ( 1997). B. R. Hames and B. Kurek. i n P m , . 9th ISWPC'. Montreal. Canada. pp, CS- 1 - G 5 5 ( 1097). P. S. Skerker, R. L . Farrcll. D. Dolphin, and F. Cui. in Riotc,c,l~r~o/o~q?. irr Ptrlp ( r u d /'(r/wr M~r~~r!/irc~trrrc (T. K. Kirk and H.-M. Chang. eds.). Butterwortll-Hcine~~~~~~ltl. Boston. p. 203 ( 1990). B. Kurek. I. Artatd. B. Pollet. C. Lapicrrc. and B. Monties. .1. A ~ r i c ,Footl . Clrcur.. 44: 1953

( 1996). 167. D. P. BNI- at1d S. D. Atlst. E/lt,i/.o/l. S(.;. fi,c./1/101.. 28: 7XA ( 1994). 168. A. Sorokin. J-L. Scris. and B. Meunicr. S c i c v r c c , . 268: I 163 ( 1995). 169. A. Sorkin. S. De Suzzoni-Dcmrtl, D. Poullain. J. P. Nocl, and B. Meunicr. J . A/)/. Ghoul. SOC. .. I I S : 7401 (1096). 170. A. Sorokin ancl B. Mcunicr. Clrcwr. Elrr: .1., 2: 1 I308 ( 1996).

This Page Intentionally Left Blank

Chemical Modification of Wood Misato Norimoto Kyoto University, Kyoto, lapan

1.

INTRODUCTION

Wood is strong along the grain for its weight, moderately hard, and easy to work. Its grain pattern, texture, and color are beautiful and give a natural feeling. Light reflecting from the wood surface does not glare and provides a mild stimulus to our eyes. As wood is highlyhygroscopicandhas large specific heat, the air in a roomsurrounded by it is moderatelyconditioned.These excellent properties of woodsas structural and interior finishing materials for residential construction result from wood structures ranging from macroscopic to microscopic and the molecular level. In many situations, however, the difficulties encountered in wood utilization are related to its dimensional instability to moisture, biodegradability, and flammability. In the case of softwoods, especially when used for flooring boards and furniture, hardness and abrasion resistance may also become a problem. Among these difficulties, the changes of wood properties caused by ambient humidity variations are of great importance. Dimensional stability has been evaluated by measuring dimensional changes in loadfree wood specimens. In actual practice, however, wood components are often subjected to humidity variations under mechanical force. This is the main factor causing an exceptionally high creep strain of structural members and tonal instability in wooden musical instruments. Generally speaking, any type of instability of wood caused by humidity variations originates in its high hygroscopicity. One of the main reasons for chemical modification of wood is to reduce its dimensional instability. Chemical modification of wood is defined as any chemical reaction between some reactive part of a wood cell wall component and a simple single chemical reagent that forms a covalent bond between the two components [ l ] . However, in this chapter, it is understood in a wide definition and includes the use of heat, steam, resins, or other chemicals. In most cases, theyaffect the amorphous components of the cell wall, the cell lumens sometimes being filled by resins or chemicals. Some modifications of wood reach the cores of cellulose microfibrils, destroying the crystalline structure and eliminating most of the composite structures of the wood. Consequently, the resulting material does not have any of the characteristic properties of wood. In compensation, it may be provided with other properties such as thermoplasticity 121. Chemical modification of wood as discussed in this chapter excludes such radical modifications. It may reduce some defects relative to wood utilization, enhance its properties. 573

Norimoto

574

and create new performance or functions, while keeping the bulk of the superior mechanical properties of wood. In this chapter, a classification of chemically modified woods will be proposed and the properties of typical modified woods will then be analyzed in relation to their structure.

II. STRUCTURE OF CHEMICALLY MODIFIED WOODS A.

Models for Chemically Modified Woods

Woodis composed of highly elongated cells whose walls have a complex, multilayered structure. In each layer, cellulose molecules are grouped together in long filaments called microfibrils embedded in a matrixcomposed of amorphoushemicellulosesand lignin. Wood is a cellular solid and, at the same time, a fiber-reinforced composite. In ordertostudyhowchemicaltreatmentschangethe properties of wood, it is important to have a clear understanding of the structural modifications of wood by the treatments. For this purpose, it is useful to classify the chemically modified woods according to adoublecriterion: (A) modification of the lumens(cellularlevel)and (B) modification of the cell wall substance (molecular level) [3-51. Figure 1 showsmodels for the chemicaltreatment of wood. A-l showsthecross section of a single cell of untreated wood. Most of the chemically modified woods can be classified into three categories at the cellular level. In A-2, the cell wall is modified with no deposit of a resin or any other product i n the lumen. In A-3, the cell wall is modified withadeposit on the internal face of the lumen. In A-4, the cell wallremainsalmost untreated, while a resin fills (partially or totally) the lumen. On the other hand, B-l shows the model at the molecular level of the untreated wood. The chain denoted by CI refers to the crystalline core of a cellulose microfibril in the middle layer of the secondary wall (Sz layer), but it must be understood in a wider definition. The chain may be any part of the lignocellulosic substance in the S? layer that is not affected by water. The two chains are

A-l A-3

A-2

A- 4

B4

55

FIGURE 1 Modelsforchemical modification of wood.a. Cellular level: A - l . untreated; A-2, treated cell wall with no chemical deposit in lumen; A-3, treated cell wall with deposit on internal face of lumen; A-4, untreated cell wall with filling of lumen. b. Molecular level: small open circle, OH group available for hydrogen bonding; small filled circle. substitution of OH group; large filled circle. bulking agent.

fication Chemical

of Wood

575

represented as two independent members for the sake of clarity, but in reality, they belong to a fully interconnected framework including not only the crystalline core of the microfibrils, but also some of the surrounding molecules that do not react to water. The chain has a water-reactive zone at its boundary, illustrated by an OH group ( h ) represented by a small open circle, which is available for hydrogen bonding with an adjacent OH group or with water. The ring ( c ) refers to the outer layer of the secondary wall (S, layer), which preventsmoisture-inducedexpansion(swelling) of the S2 layer fromexceedinga limit (hoop effect). Neighboring chains are linked to each other by the hydrogen bond. Water sorption on the OH group has a double effect. First, the change of matrix volume forces lateral displacement of the twochains ( a ) in the direction indicated by d. Second, it weakens the connection between the chains and facilitates slippage in the direction indicated by e, resulting from local shear stresses. At the macroscopic level, the first effect results in swelling or shrinkage and the second in creep or stress relaxation. The small filled circles in B-2 and B-3 indicate that the OH group has been substituted withchemicalbondingand isnot available anymore.The large filled circles in B-3 andB-4 indicate the bulking effect caused by theintroduction of large molecules between the constituents. A line shows a covalent bond. The cross-linking effect is thus pictured by an unbroken sequence of lines and small circles linking the two chains shown in B-2, preventing both lateral expansion ( 6 ) and molecular movement ( e ) . In B-3, the reactant establishes a stable bond on one side and a bulking effect occurs. If the reactant is hydrophilic, it suppresses hydrogen bonding on one side but simultaneously creates a new sorption site at the other. In B-4, no stable bond is established between the reactant and the constituents. If the bulking agent is extremely hydrophilic, it establishes extensive hydrogen bonding with the constituents. However, if the bulking agent is hydrophobic, it does not interact with the constituents or with water. The pyrolysis and oxidation of the wood constituents as well as the crystallization of cellulose reduce the water-reactive OH groups. These structural changes may be represented by B-5. Any type of chemical modification can be characterized by a combination of the two criteria (A) and (B).

B. ChemicalModificationTreatments This chapter deals with eight kinds of chemical treatments: vaporous formalization (F), acetylation (A), propylene oxide (PO) treatment, phenol-formaldehyde (PF) resin impregnation, polyethyleneglycol(PEG)impregnation,wood-methylmethacrylatecomposite (W), heating (H), and steaming (S). Formalization was done in a closed vessel with tetraoxane as well as paraformaldehyde and SO, (catalyst) at 120°C for various lengths of time. It is a reaction involving the formation of OCHz bridges between the OH groups of wood constituents by HCHO. A small molecular bridge is made while the chains are close to each other, so the structure of the resulting product is (A-2) (B-2). Acetylation was performed in neat acetic anhydride at 120°C for 10 h, followed by leaching for 10 h in boiling water and oven drying overnight. As the OH group is substituted with the hydrophobic and bulky OCOCH, group, the resulting structure belongs to (A-2) + (B-3). Propylene oxide treatment was done using PO and TEA (catalyst) inan autoclave at 120°C for 2 h. The addition of PO provides wood with the structure of (A-2) + (B-3). The OH group is substituted with the hydrophilic and bulky OCH2CH(OH)C2H, group.

+

I

Norimoto

576

PF (molecular weight: 300)resin impregnation was done by soaking wood specimens in aqueous solutions with increasing concentrations until completely saturated. After air drying, the impregnated specimens were cured. The treatment yields (A-3) + (B-4). PEG (molecular weight: 1000) impregnation was done by saturating wood specimens first in water and then in a 25% aqueous solution at room temperature for 12 h, followed by oven drying overnight. The structure results in (A-3) + (B-4). Wood-MMA composite specimens were prepared by impregnating oven-dried specimens with an MMA solution containing a catalyst. wrapping the specimens in aluminum foil, and curing the polymer at 80°C for 3 h. The structure of (A-3) + (B-3 or B-4) can be expected under experimental conditions, but (A-4) was obtained in this case. Heat treatment was performed at 120-220°C for various lengths of time by three methods: under molten metal, in the presence of air, and under evacuation of air. Heating wood in a confined atmosphere induces the formation of cohesive structures between the cell wall polymers, but at the sametimecauses their degradation.Thedegradation of water-reactive polysaccharides accompanies a decrease in their hygroscopicity. Steam treatment was performed at 120-220°C for various lengths of time by two methods: by introducingsteamintoa pressure-resisting container in which wetwood specimens were placed; and by heating air-dried specimens using a hot press equipped with an air-tight O-ring seal, resulting in steam treatment by the vaporization of moisture contained in the specimens [ S ] . Steaming wood is accompanied by degradation of the cell wall polymers, especially hemicelluloses, as well as the formation of cohesive structures of the cell wall polymers. The structure of heated or steamed wood may be represented by (A-2) + (B-2 and B-5). 111.

CHARACTERIZATION OF WOOD STABILITY

A.

ConventionalDimensionalStability

At the cell wall level, swelling of the S2 layer is allowed in its thickness direction, but is prevented from exceeding a limit by the hoop effect of the S , layer. The moisture content at this limit is considered to be the fiber saturation point. In B-l, the extent of swelling in the direction ( d ) from the oven-driedcondition to the fiber saturation pointcanbe represented by the clearance between the chain ( U ) and the hoop ( c ) . Dimensional stability provided by treatment is evaluated by the use of an antiswelling efficiency (ASE) given by ASE = ___ x 100 (%) S,, where S is the volumetric swelling of treated specimen between a dry state and a wet state, and S,, is the value measured for the untreated wood under the same experimental conditions. An ASE of 100% means that the specimen is perfectly stabilized, whereas an ASE of 0% means that the treatmenthas no effect at allon dimensional stability. A negative ASE has the opposite effect on stabilization. The experiment to evaluate dimensional stability by ASE involves measuring dimensional changes of load-free specimens. Therefore,ASEalonedoesnotgiveany indication of the modification of mechanical properties that involve the action of external force on the material. The formation of a bridge between adjacent OH groups represented by B-2 in Fig. 1 may reduce the swelling of the cell wall. In vaporous formalization, short inflexible

cation

Chemical

577

molecularbridges are formedwhile the chainsareclose to each other. Thistreatment irnproved dimensional stability for a very small weight increase, and ASEs of 80-90% could be obtained [7]. However, i n liquid-phase formalization, the reaction is made i n a swollen state, so that after drying the structure develops considerable potential looseness, resulting in almost no dimensional stability [ 3 ] . The bulking shown in B-3 and B-4 may also reduce the swelling of the cell wall. The introduced chemical or resin occupies the space between the chains ( ( I ) , so that the clearance between ( a ) and ( c ) is reduced. This bulking leads to a decrease in additional swelling by moistureadsorption. In acetylationbelonging to B-3, the hydrophilic OH group is substituted with the hydrophobic and bulky OCOCH, group. Consequently, hygroscopicity decreases and density increases depending on weight percent gain (WPG). The ASE increased with increasing WPG up to 20%, but became sluggish for higher WPG. ASE’s of 65-75%couldbe attained [8,9]. The ASE remained unchanged by soaking/ drying-cycle test [ l ] . The fiber saturation point of the acetylated wood was significantly reduced [ l ] . The dimensional instability i n the thickness direction of particleboards and veneer-faced particleboards made from acetylated particles and veneers could be greatly improved [ IO, I 1 1. Propylene oxide treatment also belongs to group B-3. The OH groups are substituted with hydrophilic and bulky groups, so that this treatment improves dimensional stability and, at the same time. yields high hygroscopicity.The ASE increased with increasing WPG up to 25-33% and showed a reduction above 33% WPG. This reduction was explained by check formation in the cell walls by the treatment 1121. A maximum ASE of 60-70% was attained at a WPG of 25-33%. In soaking/drying-cycle tests, the ASE decreased by the first soaking, but remained unchanged i n subsequent cycles [ 121. Low-molecular-weight PF resin treatment may belong to group A-3 + B-4. The WPG increased proportionally with increasing resin concentrations, but the bulking efficiency (BE), defined by the percent volume increase of oven-dried specimen after treatment and ASE, were not linearly related to resin concentrations and leveled off above 25% concentrations [ 13,141. This fact showed that only a limited amount of resin was in the cell wall acting as a bulking agent and there was excess resin in the lumen. Therefore, this treatment belongs to group A-3 in Fig. I . A resin concentration of about 25% achieved an ASE of 65% and a BE of 10% [13,14]. In PEG treatment, molecular weights should be lower than 1000 to allow penetration into the cell wall. This treatment results in B-4 and the introduced bulking agent is extremely hydrophilic. A maximum ASE of 80% could be obtained by this treatment [3,7]. Heating wood in a confined atmosphere may induce the formation of cohesive structures of the cell wall polymers, but at the same time causes chemical changes as well as degradation.The formation of cohesivestructureswasspeculated from the increase in crystallinity and dynamic Young’s modulus during the early stage of treatment [ 151. The decrease in hygroscopicity was considered to bedue to theformation of hydrophobic furfural polymer from hydrophilichemicelluloses [ 161. A maximum ASE of 50% was achieved by heat treatment [17). The structure may be B-5 or B-2. The treatment of A-4 with no modification of the cell wall substance obtained by resin impregnation can delay but not reduce the adsorption of water and consequential dimensional changes. On the other hand, the hydrophilic bulking in B-3 and B-4 may be used in the case of dimensional stability without any reduction in hygroscopicity being desired.

I

Norimoto

578

B.

Stability of Mechano-Sorptive Creep

The mechanical behavior of the matrix is influenced by water in two ways. First, when the matrix contains a large amount of water, it has a much higher viscosity than it does in the dry state. Second, the sorption process itself contributes further to the viscosity of the matrix. The arrival of water molecules at a sorption site, or their departure from it, causes a rearrangement of the surrounding molecules, and during the time needed for the rearrangement they can shift more easily to the more stable configurations implied by the forces acting on them. This transient effect of sorption explains the hygromechanical coupling in wood, or a sorption-induced viscosity that cannot be predicted by the direct effect of moistureor viscosity. This effect causes an exceptionallyhigh level of creep strain (mechano-sorptive creep [18]). It is well known that the creep of wood is greatly accelerated when humidity is changing, and that sometimes this leads to failure. Model B-l in Fig. 1 suggests that the basic phenomenaassociatedwithexpansion ( d ) and creep (e) are not connectedwitheach other, because they occur in different directions at the local level. In the case of longitudinal loading, the proportion of matrix involvement in the deformation process is determined by the average inclination of the microfibrillar angle relative to the fiber axis in the axial-circumferential plane. The model does not contain this information, but it is implied by the stress-induced strain being shown as a sheardeformation i n the direction e, which thus stands for the direction of local stresses supported by the matrix. Dimensional stability with respect to such stress-induced deformation as creep requires a reduction in matrix viscosity due to moisture sorption. The cross-linking of B-2 may efficiently prevent viscous deformation in the matrix. Bulking alone, however, does not necessarily provide this effect. Only hydrophobic bulking has a good chance of reducinghygroscopicity.Hydrophilicbulking,however,favorsmatrixhygroscopicityand may thus enhance the plasticizing action of water on the matrix, resulting in instability. The mechano-sorptive creep of spruce (Picea sirchensis) along the grain was measured in bending at 30°C underhumiditycyclingbetween29%and 86% RH [3].The initial loading was adjusted to obtain a stress level of 10 MPa in the dry state. The creep test lasted 2 days for each specimen, which was loaded first at 29% RH, kept for 10 min, moved to 86% RH for 24 h, and then returned to 29% RH for the remaining 24 h. To characterize the ability of a treatment to reduce mechano-sorptive creep, an anticreep efficiency (ACE) was defined by ACE =

A J , - AJ x loo(%) A Ju

where AJ is the increase of compliance (strain divided by stress) induced by a complete cycle of humidity, and A J c , is the value obtained from an untreated specimen. A specimen with 100% ACE has no creep at all. Negative ACE indicates mechano-sorptive instability. Figure 2 shows the relative deflection plotted against time for each chemical modification [ 3 ] .Increasing humidity induced a marked increase in deflection, and the following reduction in humidityinducedanotherincrease in deflection. After the specimenwas unloaded at 29% RH, only a slight recovery would occur over a long period as long as it remained at this humidity level. On the other hand, a high humiditylevel induced a marked recovery even though the recovery was slightly masked by the final equilibrium at 29%

RH. TheACES and ASEsobtained by variousmodifications are compared in Fig. 3. Formalization yielded a high ASE for a low weight-percent gain. It was also the most

579

Chemical Modification of Wood

U FF

0

1

.

2

3

2

4

3

O

0

1

2

4

3

4

PG G

.'" 0

0

U

l

I

I

I

1

2

3

4

W U 0

1

2

3

4

O

0

1

2

3

4

t (day)

PO e

PG

e

FIGURE 3 Relationship between anticreep efficiency (ACE)andantiswelling efficiency (ASE) for chemically modified woods. See legend to Fig. 2 for treatment abbreviations.

580

Norirnoto

efficient treatment regarding ACE [3,19,20]. The same results were obtained for the specimens treated with dialdehydessuchas glyoxal and glutaraldehyde, which form crosslinks between the OH groups [20]. In the case of treatments that induce cross-linking, good correlation could be expected between ASE and ACE [20]. Acetylation yielded high values of both ACE and ASE. Long-term creep tests conducted in an uncontrolled room showed a remarkable reduction in creep deformation after acetylation [ 191. The creep deformation under humidity changes for particleboards with the acetylated particles was also suppressed to a large extent. The retention of the modulus of elasticity (MOE), modulus of rupture (MOR), and internal-bond strength after the creep test increased with increases in the WPG of particles 1211. To stabilize piano tones under humidity changes, a pin block must have dimensional stability. Since the string is under great tension, any movement of the tuning pin causesa reduction in the pitch of the vibrating string. Measurement on a full-scale model of the string-sustaining part using acetylated and control wood pin blocks showed that changes in the resonant frequency of strings with humidity change for the acetylated blocks were much less than that for the control wood blocks 122,231. The comparison between acetylation and propylene oxide treatment is of special interest. These treatments have high ASEs. However, the most striking difference between these treatments is the opposite effect on mechano-sorptive creep. Acetylation yielded a positive ACE, but epoxide treatment yielded an extremely negative value. This meant that epoxide treatment stabilized stress-free wood but destabilized loaded wood. Although the reactant saturates the OH groups, it produces new ones during the reaction. Furthermore, as a bulking agent, the reactant increases the accessibility of water molecules to the waterreactive region. This results in more water sorption and more creep. Although most PF resin is not supposed to react with the constituents of the cell wall and thus should act as a pure bulking agent, a small proportion of linkages may be expected. The hydrophobic nature of the bulking agent resulted in a good ACE for a quite modest ASE. PEG impregnation yielded results similar to those of epoxide treatment, resulting in an excellent ASE but a negative ACE. The creep deformation of the PEG-treated wood was extremely large even at a constant humidity condition, because PEG acts as a plasticizer 1231. A low ASE of wood-MMA composite showed that only B small proportion of A " penetrated into the cell wall. Therefore, neither ASE nor ACE can be affected unless the cell wall is modified at the molecular level.

C.

Stability of Acoustic Properties

Wood has been used as a material for soundboards of musical instruments such as violins and pianos for a long time. Because wood is highly hygroscopic, the acoustic properties as well as the dimensions ofwood change during ambient humidity variations. Varnish applied to the soundboards of wooden musical instruments helps stabilize the dimensional and acoustic properties by delaying moisture absorption. However, excessive varnishing causes suppression of the sound level radiated from the soundboards. Chemical modification is another effective way to stabilize the acoustic properties of wood. The specific dynamic Young's modulus (E'ly: dynamic modulus divided by specific gravity) and dynamic loss tangent (tan 6) are important parameters of acoustic properties. The former is related to the sound velocity and the latter to sound absorption. The acoustic converting efficiency is defined as the ratio of the acoustic energy radiated from a beam to the energy

of Wood

ification Chemical

581

y)

given to vibrate the beam and is proportional to the square root of (E'y-l)l(tan 6 [24]. A large E ' l y and small tan 6 give a high converting efficiency which characterizes the acoustical quality of wood used for soundboards 125,261. The dynamic mechanical properties of Glehn's spruce (Picea glehnii) were measured at constant humidities of 0%, 35%, 60%, and 85% RH at 20°C by means of the free-free vibration method [27,28]. Figure 4 shows the effect of chemical treatment on the relationship between logarithms of tan 6 and E'ly [29]. A linear regression line indicated by a dotted line on each graph was obtained for the untreated wood using all of the values measured before the treatment. This linear relationship was found regardless of species [30-331, measuring direction [34], or moisture content [35]. Some comment should be made concerning this regression line. The E ' l y along the grain characterizes the average rigidity of the cell wall [33,36], whereas the tan 6 represents the relative amount of viscous strain to elastic strain and the participation of the matrix in the deformation process. The existence of a correlation between E ' l y and tan 6 for the untreated woodisby no means accidental. Both quantities depend mainly on the mean microfibrillar angle of the S, layer and very little on density [33]. As variations in both E ' l y and tan 6 originate from the same ultrastructural factors, they should not be independent of each other. When different viscous behavior is apparent between a treated specimen and the untreated reference. we do not know a priori whether the difference is due to the treatment or originates from the bad matching of the two specimens. However, we did confirm that it is not the latter case by observing that the point representing the modified wood data not only differs from that of the untreated wood but also lies far from the dotted line. According to this interpretation, a perfect correlation between E'ly and tan 6 at a given level of humidity corresponds to the ideal case where all structural parameters would be

'

-2.4 1.00.8

-2.0

1 '.

-2.21 -2.4 0.8

"'"

-2.0

1

I

1.2 1.6 1.4

I

\

A.'\\',

-A, $,' 1.6 1.01.4 1.2

FIGURE 4 Relationships between logarithms of dynamic loss tangent (tan S) and specific dynamic modulus ( E ' l y ) for chemically modified woods at various relative humidities. Dotted lines represent experimental correlation for untreated wood. See legend to Fig. 2 for abbreviations.

582

Norimoto

either constant or exactly correlated with the microfibrillar angle. The deviance of a given plot relative to the reference line is thus explained by the disturbing action of these other structural parameters. The tan 6 decreased as the humidity changed from 0% to 35% RH and then rose with increasing humidity. The E ' l y remained about the same from 0% to 35% RH and then decreased with increasing humidity. In an oven-dried state, molecular chains in the amorphous regions of the cell wall are unnaturally distorted. This causes a smaller E ' l y and a greater tan 6. With increasing moisture content from oven drying, the distorted molecular chains are rearranged. As a result, at 35% RH, a more stable state for the cell wall structure is obtained. Above 35% RH, water acts as a plasticizer that allows more molecular movement and the cohesive forces between molecules are decreased, resulting in decreasing E'ly and increasing tan 6. In vaporous formalization, cross-linking occurs in dry state of the cell wall and prevents swelling. This decreases the mobility of molecular chains, thus reducing tan 6 at all humidity levels. This treatment decreased tan S as much as 40-45% in the frequency range from 150 to 500 Hz [37]. The reduction was smaller at high frequencies and larger at low frequencies. The E ' l y at 65% RH remained unchanged in the grain direction and increased 10% in the direction perpendicular to the grain [37]. By a sensory evaluation test, a violin having a treated bridge was judged to be comprehensively better than the same violin having an untreated bridge [37]. Vaporous formalization is effective not only for improving the acoustic properties of wooden instruments but also for their dimensional changes resulting from humidity variations. The E ' l y and tan 6 of the acetylated wood decreased as the WPG increased [38]. The E ' l y decreased slightly, while the tan 6 increased greatly as the frequency increased, but no differences in frequency dependence were observed between the acetylated and untreated woods [38]. The reduction in E ' l y and tan 6 at 20% WPG was 10-15%. The hydrophobic OCOCH,groups were subjected to hydrogen atoms in the cell wall. This reduced the E'ly and tan S owing to the steric hindrance of chains imposed by the bulky groups. This effect was seen at all humidity levels. However, the humidity-induced changes of E'ly and tan 6 were greater than those of the formalized wood. Because all available OH groups were not acetylated, water still was able to bond with the cell wall polymers and act as a plasticizer. The acoustic properties of wood changes more in a nonequilibrium moisture condition than in an equilibrium moisture condition. However, acetylation greatly suppressed these changes, especially in the process of adsorption [23,35,38]. Acetylation was shown to stabilize tone quality under conditions of changing humidity [23,39]. Propylene oxide treatment (WPG; 22%) results in bonded cell wall bulking as in the case of acetylation,except that the introduced group is hydrophilic. This made a big difference as humidity increased. At 0% RH, the treated wood had a slightly lower E ' l y because of the increase in specific gravity and the tan 6 was lower thanthat for the untreated wood. At 35% RH, the tan 6 of the treated wood was almost the same as that for the untreated wood, but at 60% and 85% RH, it was much larger. This was due to the flexibility introduced into the cell wall polymers because hydrophilic ether allowed water to act as a plasticizer. The changes in both E'ly and tan 6 varied with average molecular weights of PF resin [40]. As the molecular weight of resin decreased, the E'ly increased and tan 6 decreased. This result was ascribed to the resin content maintained in the cell walls: the higher the molecular weight, the less the resin was maintained. Similar results were also obtained in treatment with MF resin with a low molecular weight. The treatment with low-molecular-weight PF resin (WPG; 45%) introduced a bulky and hydrophobic group

583

Chemical Wood Modification of

into the cell wall. Because of the presence of a rigid benzene ring in the resin backbone as well as reduced moisture sorption, molecular mobility was reduced, which lowered tan 6 for all humidity levels tested. PEG with a low molecular weight is a hydrophilic polymer that can enter the cell wall. This very flexible hydrophilic polymer swells and plasticizes the cell wall even at 0% RH, resulting in very large tan S. The hydrophilic nature of the cell wall does not change as the humidity increases because it does not matter whether PEG or water is acting as a plasticizer. When only the lumen was filled or coated with chemical, there was no deviation from the untreated reference regardless of humidity conditions. The treatment with MMA (WPG; 138%) indicates that a small amount of this polymer may enter the cell wall, but in general this treatment only fills the lumen space. The increase in specific gravity decreased E’ly. Water can still enter the cell wall and act as a plasticizer, so that the tan 6 increased with increasing humidity. From a practical point of view, it is of importance to predict the stability to changes in humidity for chemically modified wood with respect to E’ly and tan 6. To characterize how E ’l y and tan 6 change with increasing humidity conditions, and E‘ly stability (S,) and tan 6 stability (S,.) are defined by S,: =

A,

-

~

A,

B,

x loo(%),

AT - B,.

S,.= ___ x loo(%),

A,.

where A and B are the relative changes of specific modulus ( E ) or loss tangent ( T ) for the untreated and treated woods, respectively, A is the change of E ‘ l y or tan 6 between 35% and 85% RH, and E ’ ly and tan 6 are the values at 35% RH. Figure 5 shows the relationship between SE and ST for the chemically modified woods. The formalized wood, acetylated wood, and PF resin-impregnated wood had large positive values of S, and S,., while the epoxide- and PEG-treated woods had large negative

FIGURE5 Relationship between tan S stability (S.,) and E’ly stability (S,) for chemically modified woods. See legend to Fig. 2 for abbreviations.

Norimoto

584

values of S, and &.These results shows that the greatest stability to changes in acoustic properties with increasing RH is achieved with vaporous formalization, acetylation, and PF resin treatment.

IV.

FIXATIONOFCOMPRESSIVEDEFORMATION

A.

Mechanism of Drying Set

In coniferous woods and fast-growing woods with low density, surface hardness and abrasion resistance may become a problem, especially when these woods are used for flooring boards, furniture, and interior finishing materials. Softwood with useful surface properties can be obtained by compressing the wood in the direction perpendicular to the grain. Figure 6 shows the result of radial compression with loading-unloading cycles at 100°C for wet sugi (Cryptomeria japonica) [41-451. Almost complete recovery was observed provided the strain level of the first loading remained small within the linear range. The curve of the second loading was nearly superimposed on that of the first loading. As soon as the linear limit (yield point) was exceeded, the strain increased remarkably with increasing stress. After unloading, there remained a slight residual deformation, so that most of the strain was not plastic (apparent plasticity). Thesubsequent loading never followed the previous path. Instead, the curve had a lower initial slope, followed by a new plateau for a much lower yield stress, indicating a reduction in stiffness. As soon as the point before unloading was reached, the apparent plasticity started again, as if no unloading had occurred in between. As an interpretation for this plateau, it is suggested that the portions of wood with lowest density (earlywood) are crushed first, followed by the next in terms of density, and so on, until the portions of wood with highest density (latewood) are crushed as well [44]. When the cells collapsed so much that opposing cell walls touched at large strains, the curve showed a steep rise (densification) by the deformation of the cell wall itself. Eventually, the highest previous stress and strain were reached again simultaneously, so that the original curve corresponding to monotonous loading was reached exactly where it had been left previously. When a portion has been crushed to some level, it will be damaged and behave differently during subsequent loading. As far as the observations by scanning electron microscope (SEM) were concerned, no indications 4

3

2 -52 b

1

0

25

50

1

3

E (%)

FIGURE 6 Nominal stress (+strain ( E ) curve with increasing loading cycles in radial compression for wet sugi wood at 100°C. Solid and dotted lines represent loading and unloading, respectively. Numbers indicate unloading points.

585

Chemical Modification of Wood

of damages were observed in the cell wall [46]. However, the irreversibility of the deformation process suggests the occurrence of some structural changes inside the cell wall. Much the same results were obtained for wet specimens at 20°C. On the other hand, a large residual strain occurred after a loading beyond the yield point in an air-dried condition [45]. However, the strain was largely recovered through hygrothermal treatment. When a compressed wet specimen is dried under restraint, the stress gradually decreases until it has disappeared. The deformation is fixed in the deformed state (drying set). However, this drying set is only apparent, because it is almost recovered by boiling. Figure 7 shows the cross sections of the noncompressed specimen (A), the compressed specimen (B), and the specimen recovered by boiling [47]. The drying set results from the rehardening of the matrix upon cooling and drying [48]. The elevation of temperature in wet condition softens the matrix, and its two main constituents, hemicelluloses and lignin, shift from the glassy state to something near the rubbery state. On the other hand, the microfibrils remain inthe glassy state and are almost unaffected by moisture or heat. When a load is applied to the material, most of it is supported by the microfibrils at the local level. Softening of the matrix enables the relative displacement of the microfibrils in order that the whole framework deforms elastically to its deformation adjust to the local loading. As lignin is a slightly cross-linked high polymer, should be viewed as increased viscoelastic strain rather than plastic flow. The departure of water molecules due to drying induces the re-formation of hydrogen bonds between the molecules of the matrix constituents. Together with the temperature decrease, this process leads to return to the glassy state, where the elastic deformation of the microfibrils and the matrix are frozen. Accordingly, the set will not be recovered provided no resoftening of the matrix occurs. However, as soon as the matrix is softened again through remoistening and heating, most of the set is recovered due to the liberation of the energyelastic strain stored in the microfibrils and the entropy-elastic molecular movements in the matrix. The percentage of size reduction from the original state to the compressed state (compression set: CS) is calculated by

--

" -*-

-7

1 A

B

C

FIGURE 7 Cross sections of noncompressed specimen (A), compressed specimen (B), and specimen recovered by boiling (C) [47].

Norimoto

586

where T,, is the oven-dried thickness before compression and Tc is the oven-dried thickness after compression. The recovery of set (RS) is calculated by

where TK is the oven-dried thickness after recovery.

B.

Fixation by Cross-Linking

In order to utilize the compressed wood, permanent fixation of the compressed deformation is essential. One of the most effective ways of fixing deformation is the formation of crosslinkings between the wood components. Vaporous formalization is applicable to this purpose. Compressed sugi specimens (CS: 50%) were treated with formaldehyde vapor generated from paraformaldehyde and SO, in a dry condition at 135°C for 20 min. Figure 8 shows the relationship between RS and treating time 1491. The RS rapidly decreased with increasing treating time. It should be noted that the deformation was completely fixed by the treatment in about 5 min. The deformation was also completely fixed by the reaction with tetraoxane and SO, at 120°C for 2 h [SI]. If the mechanism of fixation by formalization arises from the formation of interchain covalent bonds, the deformation should be recovered by the scission of the cross-linkings. To confirm this, the formalized compressed specimen was soaked in a 10 N H,SO, solution. The result showed a remarkable recovery of set deformation [49]. In liquid-phase formalization. the RS decreased with increasing treating time, although complete fixation was not realized within the treating time of 250 min [49]. Formaldehyde treatment is also effective in improving the dimensional instability of medium-density fiberboard (MDF). The untreated MDF swelled infinitely in its thickness direction when boiled, whereas the treated MDF swelled only about 20% and recovered its original thickness when dried again [SO]. These results suggested that the dimensional stabilization of MDF is attributable to the formation of chemical bonds between fibers.

'0

20

40

60 80 100

1 li

t (min)

FIGURE 8 Relationship between recovery of set (RS) and treating time pressed sugi wood (compression set, CS: 50%).

(1)

for formalized com-

587

Chemical Modification of Wood

C.

Fixation by ResinTreatment

Almost complete fixation of deformation can be achieved by using PF resin that is polymerized during the deformation stage. Oven-dried sugi specimens were soaked in resin solutions of increasing concentrations. After air drying for 24 h, the impregnated specimens were slowly warmed over a 12-h period to 105°C and then compressed in the radial direction using a hot press. The treated specimens were soaked in water until saturated and their thickness measured. After drying at 40°C for 20 h and then at 105°C for 4 h, their thickness was again measured. This procedure was repeated. After the final soaking cycle, the specimens were placed in a boiling water bath for 2 h, measured, and then oven-dried. Figure 9 shows the results of soaking-drying cycle test for the compressed specimens (CS: 55%) treated with PF resin (molecular weight: ca. 200). The untreated specimens showed a large increase of RS during the cycle test, while the specimens soaked in 15% and 20% PF solutions effectively resisted swelling. The SEM micrographs showed no remaining PF resin in the cell lumen. The resin in the cell wall resulted in ASEs of 60-70% at about 20% WPG. Hardness of the products increased with increasing CS and resin concentrations. A resin concentration of 15% and a CS of 60% increased hardness by a factor of 3, compared to that of the noncompressed wood; and a 40% resin concentration, by a factor of 4. Therefore, larger values of hardness than those of high-density hardwoods can be obtained. Abrasion resistances of the specimens compressed to 60% were about 60% regardless of resin concentrations. Abrasion resistances of the products increased with increasing CS. The MOE increased with increasing resin concentrations. The MOR decreased in low concentrations, but increased above 25% concentration. The compressed specimens (CS: 54%) soaked in 2% and 4% MF (molecular weight: 380) solutions showed a large increase of RS during eight soaking-drying cycles. However, the specimens soaked in 8%, ls%, and 25% MF solutions effectively resisted swelling. After four boiling cycles, the specimens treated with 2% and 4% solutions recovered in almost the same way as the untreated specimens. Thespecimens treated with 25% solutions retained high dimensional stability. This method achieved a maximum BE of 5% and an ASE of 45%, showing that the chemical had not completely penetrated the cell wall. Hardness increased in proportion to concentration. Starting with a hardness of 0.24 MPa at 0% concentration, it increased to 0.48 MPa and 0.72 MPa at 15% and 25% concentrations, respectively, at a compression of 54%. A remarkable decrease in abrasion resistance with increasing concentrations was observed. Increases in MOE and MOR at a 25% concentration were about 10% and 18%, respectively [ 13,141. 100

80

- 60

2

2 40 20 n

" ~

~

~

~ Cycles

~

~

~

~

~

~

~

~

FIGURE 9 Soaking/drying-cycle test for compressed sugi wood (CS: 55%) treated with PF resin. D, oven-drying; W, soaking; B, boiling; RC, resin concentration.

~

~

Norimoto

588

D.Fixation

by Heating

Heating while under deformation is another effective way of fixing. After wet sugi, radiata pine (Pinus rdutl1), and albizia (Pamsrrienthes,fnlcatrr) specimens were irradiated with microwaves at 2.4 kW, they were compressed in the radial direction to about SO% of their original thickness (CS: 50%), followed by oven drying at 150°C. The compressed specimens were then heated at 160, 180, or 200°C for various lengths of time by three methods: under molten metal (MH), in the presence of air (CH), and under evacuation of air (VH). The treated specimens were soaked in water until saturated, placed in boiling water for 30 min, then oven dried. The RS increased in the order of CH, MH, and VH when compared at the same heating temperature and time. However, it was independent of heating method and wood species when compared at the same weight loss (WL)as shown in Fig. 10 [Sl]. The relationship of RS to WL was expressed by a following hyperbolic equation: RS =

53.02 WL + 0.579

-

11.6

0 5 WL 5 4.0

Almost complete fixation was achieved at about 4.0% WL. Conventional heating for 20 h at 180°C or 5 h at 200°C achieved complete fixation. The retentions of MOE and MOR at perfect fixation were about 88% and 78%, respectively. As for color changes in the L*-a*-b* color system, the decrease in shade was 28% and the color difference was 29% 1141.

E. Fixation by Steaming The fixation of compression set in wood can be also achieved by steaming. Wet sugi specimens irradiated with microwaves were compressed to about SO% CS in the radial direction and then oven dried under restraint at 105°C. The compressed specimens were fitted between stainless steel plates that were placed inside stainless side restraints. They were steamed in an autoclave at 140, 160, 180, or 200°C for various lengths of time. Figure 1 1 shows the effect of steam treatment on RS [52].As the temperature increased, the RS decreased. Almost no recovery of set was observed after only 1 min of steaming at 200°C or 8 min at 180°C. Hardness resulting from compression increased from 0.07

1

589

Chemical Modificationof Wood

140

t (mid

FIGURE 11 Relationship between RS and treating time ( t ) for compressed sugi wood (CS: 50%) steamed at 140, 160, 180, and 200°C.

MPa to 0.25 MPa, an increase of about three times. A small reduction in hardness was observed when steaming the compressed specimens. Figure 12 shows changes in MOR of the steamed noncompressed specimens [52]. Only a small change in MOR resulted from steaming specimens at either 180°C for 8 min (3.3%) or 200°C for 1 min (8.6%). Only a slightyellowing and darkeningoccurred in the specimenssteamed at 180°C for 8 min Compression and hygrothermal treatment can be done simultaneously using a hot press equipped with an air-tight O-ring seal. When a sugi specimen with 17% moisture content was compressed at 180"C, perfect fixation was achieved in 10 min. However, the fixation was insufficient for dry specimens, so that moisture in the wood was considered to act so as to fix the deformation [6]. To clarify the mechanism of fixation, crystallinity changes by steaming or heating of sugi powder were investigated. Steaming was conducted in an autoclave at 120-220°C for 10 min, whereas conventional heating was in an oven at the same temperature levels for 20 h [53].The treated powder was compressed into a disk for X-ray diffraction measurements. Diffraction pattern was obtained from 5" to 40" by the reflection method. The degree of crystallinity (DC) was calculated by the ratio of the area corresponding to the crystalline region to that of both crystalline and amorphous regions in diffractograms. The half-width of the (200) diffraction (HW) was determined to evaluate crystalline width: the larger HW is, the smaller crystallite size results. As shown in Fig. 13, the DC of steamed

-60

1

590

2o

Norimoto

120

140

160

180

200

220

TeC)

FIGURE 13 Relationship between degree of crystallinity (DC) and treating temperature for heat(*) and steam- ( 0 ) treated sugi wood powder.

powder increased with increasing temperature while that of heated powder decreased. The HW of heated powder increased with increasing temperature while that of steamed powder decreased. Significant changes in the bands assigned to the CO and COOH groups at 1736, 1719, and 1698 cm” in IR spectra were detected, which were more apparent in heat treatment than in steam treatment. The stresses stored in the matrix and microfibrils may be released, because hemicelluloses degrade by both treatments at 180°C [ 16,541. At the same time, the cross-linking reactions in the matrix and the crystallization of the microfibrils are likely possible. The results of X-ray diffraction and IR absorption measurements showed that the former mechanism was dominant in heat treatment and the latter in steam treatment. However, compressive deformation was also fixed to some extent by preheating or steaming. Therefore, it is reasonably supposed that the fixation of compressive deformation by heating resulted mostly from the release of the stresses stored in both microfibrils and matrix by the degradation of the cell wall polymers. On the other hand, the fixation by steaming is considered to occur probably because of the formation of cohesive structure in the cell wall as well as the relaxation of the stresses stored in both microfibrils and matrix.

V.

DYNAMICVISCOELASTICPROPERTIES

A.

ViscoelasticModeling

The dynamic mechanical properties of the cell wall of chemically modified woods were analyzed by using a model shown in Fig. 14 [ S ] , in which an amorphous isotropic matrix is disposed in parallel along the axis of cellulosic fibrils ( 1 direction) inclining at 8 to the grain direction of wood. The complex dynamic modulus of the model along the grain (E?) is expressed by

where ET and ET are the complex dynamic moduli in the 1 and 2 directions, G” is the complex shear modulus in the 1-2 plane, and p12 is Poisson’s ratio. If 8 is small enough to ensure sinJ@= 0, cos‘e = 1, sin’O.cos’8 = e’, and p12 is much smaller than the real part of E ? , the E’ and tan 6 of wood along the grain, as the first approximation, can be expressed by

Chemical Wood Modificationof

591

FIGURE 14 A model of the S, layer in the cell wall.A, side view; B, cross section;f, microfibril; m, amorphous isotropic matrix; 8, microfibrillar angle; 1, longitudinal direction of microfibril; 2, direction perpendicular to 1 direction.

where A is the volume fraction of the S, layer in the cell wall, y and yw are the specific gravities of wood and the cell wall, are the dynamic and loss moduli in the 1 direction, and G’ and G” are the dynamic shear and loss moduli in the 1-2 plane, respectively. E; and E:’ are expressed by

-

E: = WE, + (1 - W)E,,,

WE,

E’,’ = (1

- *)E,,, tan S,,,

where W is the volume fraction of the fibrils in the cell wall, E, is the dynamic modulus of the fibrils, and E,,, and tan S,,, are the dynamic modulus and loss tangent of the matrix, respectively. In the model, the fibrils with square cross section are embedded in the matrix, so that the fibrils and matrix are aligned partly inseries and partly in parallel tothe direction of shear force. According to the law of mixtures [56], G‘ and G’’can be expressed by tan S,,, The experimental values of E’, tan S, and y at 20°C and 60% RH for the untreated and chemically modified woods were fitted to the equations [27,28]. For the untreated wood, W = 0.5, yw = 1.45, A = 0.84, 8 = 0.09 rad, E, = 134 GPa, and E,,, = 2 GPa were used [33,36]. For the modified wood, W = 0.33-0.45 and yw = 0.92-1.29 were estimated from both weight gains and volume swellings. The values of E,,,and tan S,,, were quantified by eliminating the swelling contribution to the E‘ and tan S of five kinds of chemically modified woods, including untreated wood. The relationship between logarithms of tan S,,, and E,,,are shown in Fig. 15 [ S ] . Propylene oxide treatment is similar to acetylation in terms of conventional stabilization, but tends to increase the molecular mobility instead of reducing it because of the hydrophilic nature of the bulking agent. Although the two treatments induced the same amount of volume and weight increases, pronounced drops of tan S,,, implies a strong reduction of matrix

I

592 -1.3

-1.4

Norimoto

1

0

p~

-1.5 0

F

L"!

-l8 -1.9

0

0.1

0.2

0.3

04

log Ern

FIGURE 15 Relationship between logarithms of loss tangent (tan S,,,) and dynamic modulus (E,,,) for matrix substances of chemically modified woods. See legend to Fig. 2 for abbreviations.

mobility because of the hydrophobic nature of the bulking agent. PEG treatment is rather similar to propylene oxide treatment in terms of chemical modification of the cell wall, except that the hydrophilic nature of the molecule is even more pronounced. A slight decrease oftan S,,, in formalization impliessmaller mobility, which is what the crosslinking would be expected to produce, but a slight decrease in E,,, suggests some degradation of the matrix components.

B. Relaxation Process One of the useful ways of understanding the mechanical relaxation processes of chemically modified woods is through the temperature dependence of E' and tan S at fixed frequencies. The temperature variations of E' and tan 6 along the grain for five kinds of chemically modified woods as well as untreated wood were measured over a temperature range from - 150°C to 200°C with an automatic dynamic viscoelastometer (ORIENTEC Co. Ltd., Rheovibron: DDV-2SFP). The variations of E' and tan 6 with temperature at 11 Hz for the untreated (U) and formalized woods (F) in a dried condition are compared in Fig. 16 [57,58]. With respect to tan 6 for the untreated wood, four relaxation processes were detected within the temperature range examined. The tan 6 increased with increasing temperature above 100°C. This relaxation process. labeled a(,, shifted to lower temperatures with increasing moisture content and exhibited a clear peak at high moisture contents below 100°C [59,60]. This process was attributed to the micro-Brownian motions of the cell wall polymers in the noncrystalline region [S9]. A small tan 6 peak at about 70°C has been observed by many investigators, who assigned it to the local motions (torsional oscillations) of the cell wall polymers. However, its location remained unchanged in measurements atdifferentfrequencies. Besides, the peak was not observed in the absolutely dried specimens [ S i ' ] . It may be ascribed to the segmental motions of the cell wall polymers activated during water desorption. The loss peak at around -40"C, labeled P(,,was not detected i n the absolutely dried specimens. The same process was also observed in the dielectric measurements for specimens containing a small amount of water [61]. The results showed that this process was due to the rotational motion of the adsorbed water itself [61-64]. The tan 6 peak at around - I lO"C, labeled y(,, was also detected at the same temperature-frequency location in the dielectric measurements. The apparent activation energies obtained by both measurements showed the same value of 9.8 k c a l h o l [61]. This process has been reported by

Chemical Modification of Wood

0.001

-150 -100 -50

593

0 50 E ' l EO xx)

T ("C) FIGURE 16 Temperaturevariations of dynamicmodulus ( E ' ) andlosstangent treated (dotted line) and formalized (solid line) woods.

(tan S) for un-

many investigators, who assigned it to the motion of the CH,OH group of the cell wall polymers in the noncrystalline region [61,65-671. In formalization, three relaxation processes were observed in tan 6, labeled aI.,PI., and yI; in order of decreasing temperature at which they detected [57,58]. This treatment involves the cross-linking of chains by OCH, bridges between the OH groups of the cell wall polymers. The bridges may restrict a part of the micro-Brownian motions of the cell wall polymers to some extent, resulting in a decrease in tan 6 above 100°C. The PI. process is comparable to the Dl, process. The loss peak due to the motion of the OCH, group was observed at almost the same temperature location as that of the CH,OH group [68]. Therefore, the y,: process may involve both motions of the remaining CH20H groups and introduced OCHz groups. The variations of E' and tan 6 with temperature at 1 1 Hz for the acetylated wood are compared to that of the untreated wood in Fig. 17 159,601. The E' was lower than that of the untreated wood over the temperature range tested. Two relaxation processes, labeled a , and P;l, were recognized in tan 6. The OH groupsaresubstituted with the bulky OCOCH, groups, which reduces the cohesive forces between the main chains of the cell wall polymers and probably facilitates their backbone motions. This effect is similar to that produced by the addition of plasticizer and is usually known as internal plasticization. The marked decrease in E' and increase in tan 6 over 100°C are ascribed to the internal plasticization.Thus, the process is assigned to themicro-Brownian motions of the acetylated cell wall polymers. The loss peak comparable to the p(, process related to water desorption was not recognized in acetylated wood because of a reduction in hygroscopicity. As most of the OH groups i n the noncrystalline region are replaced by the OCOCH, groups at about 20% WPG, the motion of the OCOCH, group instead of the CH,OH group ought to occur. The peak location agreed well with that due to the motion of the OCOCH, group i n acetylcellulose 1691.

Norimoto

594

0.1

0

5 0.01

Y

0.001 -150-100 -50

50 T ?C> 0

m

1 5 0 200

The variations of E' and tan 6 with temperature at I 1 Hz lor the propylene oxidetreated wood areshown in Fig. 18 [57,58]. Thistreatment results in bonded cell-wall bulking as in the case of acetylation, except that the introduced group is hydrophilic. The E' was lower than those of the untreated samplesover the temperature range tested, especially above 100°C. Three relaxation processes, labeled a,,(,to 'y,.(,,were detected in

""-

kll """""

Chemical Wood Modification of

595

tan 6. As the introduced OCH,CH(OH)C,H5 group is bulkier and more flexible than the OCOCH, group. the cell wall polymers are much plasticized compared to those of acetylation and the a/.,,peak was observed within the temperature range tested. The treatment reduceshygroscopicity at low relative humidity levels 1291, so the &, processdue to adsorbed water became less distinct. The introduced side chain contains the OCH, group whose motion may be responsible for the y,.(, process as in the y/..process. The temperature dependence of E' and tan 6 at I 1 Hz for the wood-MMA composite is shown in Fig. 19. The E' was larger than that of the untreated wood within the temperature range tested. With respect to tan 6, five relaxation processes were observed, labeled aIV,PI,,, ylv, 6,,., and E,,. in order of decreasing temperature. A very low ASE shows that the cell wall remains untreated, while PMMA fills the lumen. Therefore, both relaxation processes observed in the untreated wood and PMMA are expected. Two relaxation processes assigned, respectively, to the micro-Brownian motion of the main chain and the motion of the OCOCH, side group were detected in PMMA 1701. These locations corresponded, respectively, exactly to the p,,. and yl,. processes. Thus. the remaining al,,, 6,,,, and cl,, processes can be ascribed to the same motions as those of the a(,to yI, process. Thetemperaturedependence of E' and tan 6 at 1 1 Hz for the PEG-impregnated wood is shown i n Fig. 20 [57,58]. The E' was larger than that of the untreated wood below O"C, but became lower above that temperature. Three peaks were observed in tan 6, labeled p,.(;,and y/.(;i n order of decreasingtemperature. I t is well known that PEG molecules with molecular weights of 1000 can penetrate into the cell wall [71,72] and act as a plasticizer, which leads to large reductions in the cohesive forces between the cell wall polymer molecules. The a/.(;was attributed to the micro-Brownian motions of the cell wall polymers plasticized with PEG molecules [60,72]. The tan 6 of the PEG-impregnated glass fibers had a peak between -50°C and 0°C. The apparent activation energy for this process estimated from the measurements at five different frequencies was about 40 kcal/mol, which suggested a mechanism involving chain backbonemotions of PEG

.

-

9

0.01

l

Norimoto

596

2 -51 (3 v

"""-

lo

W

5

FIGURE 20 Temperature variations of E' and tan 6 for untreated (dotted line) and polyethylene glycol-impregnated (solid line) woods.

molecules. On the other hand, the tan 6 increased remarkably above 20"C, which corresponds almost to the melting point of PEG. Therefore, this sharp increase in tan 6 may be related to the flow of PEG molecules. The p,.(; process was not recognized in PEG. PEG molecules in the cell wall plasticize the cell wall polymers, but at the same time their micro-Brownian motions must be restricted to some extent by the cell wall polymers. This may lead to a shift of the peak due to the micro-Brownian motions of PEG molecules toa higher temperature.The peak due to CHzOHgroup occurred about 20°C below the y,, peak temperature, and its value was rather large.

REFERENCES 1.

2. 3. 4.

S 6. 7. 8. 9. IO. II. 12.

R. M. Rowell, in Wood r r n d Ce//u/o.srChemistry (D. N.-S. Hon and N. Shiraishi, eds.), Marcel Dekker. New York and Basel, p. 703 (1991). N. Shiraishi, T. Aoki, M. Norimoto, and M. Okumura, C h e n ~ t c ~ h13: . 366 (1983). M. Norimoto, J . Gril, and R. M. Rowell, Wood Fiber Sci., 24: 25 (1992). M. Norimoto and J . Gril, in R w d Resenrrh U I I Wood m d W O O ~ - B ~ ~Mnrerids . S C > L ~ (N. Shiraish. H. Kajita, and M. Norimoto, eds.). Elsevier Applied Science, London and New York, p. 13.5 (1993). Modificrcriorl U/ o f Ligrzoce//[r/o.sicMrrfericcls (D. N.-S. Hon,ed.). M.Norimoto, in C / W I ? I ~ C Marcel Dekker. New York, Basel. and Hong Kong, p. 3 11 (1996). M. Inoue, N. Kadokawa, J. Nishio, and M. Norimoto. Wood Res. Techno/. Notes, No. 29: S4 ( 1993). R. M.Rowell and P. Konkol. FPL Gen. f i 4 . Rep. (USIIA). No. 55: 1 (1987). , 493 (1947). A. J. Stanm and H. Tarkow, J. Phys. Co//oir/C / ~ e m .31: M. Norimoto. Wood Res. 7 k ' / ? I l . Nures, No. 24: 13 ( 1988). R. M. Rowell. A.-M. Tillman. and R. Simonson. J . Wood C/7rw7. f i d t r r o l . , 15: 293 (1986). R. M. Rowell. Y. Imamura, S. Kawai. and M. Norimoto, Wood Fiber Sci.. 21: h7 (1989). R. M. Rowell and W. D. Ellis, FPL Res. Pup. (USDA), No. 451: 12 (1984).

Chemical Wood Modificationof

597

13. M. Inoue, S. Ogata, M. Nishikawa,Y. Otsuka, S. Kawai, and M. Norimoto,Mokuzai Gakkaishi, 39: 181 (1993). 14. M.Inoue, S. Ogata, S. Kawai,R.M.Rowell,and M. Norimoto, Wood FiberSci., 25: 404 (1993). 15. N. Hirai, N. Sobue, and I. Asano, MukuzuiGakknishi, 18: 535 (1972). 16. A. J . Stamm, Wood and Cellulose Science, Ronald Press, New York, p. 317 (1964). 17. M. Inoue and M, Norimoto, Wood Res. Technol. Notes, No. 27:31 (1991). 18. P. U. A. Grossman, Wood Sci. Tech., 10: 165 (1976). 19. M. Norimoto, J. Gril, K. Minato, K. Okarnura, J. Mukudai, and R. M. Rowell, Wood Ind., 42: 504 ( 1987). 20. R. Yasuda, K. Minato, and M. Norimoto, Wood Sci. Technol., 28: 209 (1994). 21. S. Takino, M. Norimoto, S. Kawai, and H. Sasai, Mokuzai Gakknishi, 35: 625 (1989). 22. A. H. Yano, J. Mukudai, and M. Norimoto, MokuzaiGnkknishi, 34: 94 (1988). 23. H. Yano, M. Norimoto, and R. M. Rowell, Wood and Fiber Sci., 25: 395 (1993). 24. C . Tanaka, T. Nakao, and T. Takahashi, Mokuzai Gakkaishi, 33: 8 I 1 (1987). 25. M. Norimoto, T. Ono, and Y. Watanabe, J. Soc. Rheol. Jpn.. 12: I 15 (1984). 26. M. Norimoto, MokuzaiGakknishi, 28: 407 (1982). 27. H. Akitsu, M. Norimoto, and T. Morooka, Mokuzai Gakknishi, 37: 590 (1991). 28. H. Akitsu, J. Gril, and M. Norimoto, Mokuzai Gakkaishi, 39: 258 (1993). 29. H. Akitsu, M. Norimoto, T. Morooka, and R. M. Rowell, Wood Fiber Sci., 25: 250 (1993). 30. T. Ono and M. Norimoto, Jpn. J . Appl. Phys., 22: 61 1 (1983). 3 1. T. Ono and M. Norimoto, Rheol. Acta, 23: 652 (1984). 32. M. Norimoto, J. Gril, and T. Sasaki, in Pruc. Eurupenn Scientific Colloquium on the Mechunicul Behnvior uf Wood, Bordeaux, France, p. 37 (1 988). 33. M. Norirnoto, F. Tanaka, T. Ohogama, and R. Ikimune, Wood Res. Technol. Notes, No. 22: 53 (1986). 34. T. Ono and M. Norimoto, Jyn. J. Appl. Phys., 24: 960 (1985). 35. T. Sasaki, M. Norimoto, T. Yamada, and R. M. Rowell, Mukuzui Gakkaishi, 34: 794 (1988). 36. M. Norimoto, T. Ohogama, T. Ono, and F. Tanaka, J. Soc. Rheol. Jpn., 9: I69 (198 I ) . 37. H. Yano and K. Minato, J . Acoust. Soc. Am., 92: 1222 (1992). 38. H. Yano, M. Norimoto, and T. Yamada, Mukuzui Gakkaishi, 32: 990 (1986). 39. T. Ono, Y. Katoh, and M. Norimoto, J . Acoust. Soc. Jpn., 9: 25 (1988). ~ . Am., 96: 3380 (1994). 40. H. Yano and K. Minato, J . A c ~ u s SOL.. 41. Y. Liu, M. Norimoto,and T. Morooka, MokuzniGakkni.shi, 39: I140 (1993). 42. M.Norimoto, MoklczuiGnkkcrisl~i, 39: 867 (1993). 43. M.Norimoto, Wood Res. Z~chnol.Nutes, No. 30: 1 (1994). 44. J. Gril and M. Norimoto, in Proc. COST-508 Wood Mechanics Wurkshop on Wood: Plasticity tend Danlccge, Ireland, p. I35 ( 1993). 45. Y. Liu, M. Norimoto. and T. Morooka, Wood Res. Technol. Nutes, No. 31: 44 (1995). 46. 1. Iida and M. Norimoto. and Y. Irnamura, Mokuzui Gnkknishi, 30: 354 (1984). 47. M.Inoue, T. Aoki, and G. Egawa, Wood Res. TeLhllol. Notes, No. 28: 59 (1992). 48. M. Norimoto and J. Gril, J. Micr-owntv Electromngn. Energy, 24: 203 (1989). 49. M. Inoue, K. Minato,and M. Norimoto, MokuzuiGcckkoishi, 40: 931 (1994). 50. K. Minato, N. Kubo, M. Norimoto, H. Sasaki. M. Sawada, and T. Yamamoto, Moku-ui Gnkkuishi. 38: 67 (1992). 51. W. Dwianto, M. Inoue, and M. Norimoto, Mukrrzcci Gukknishi, 43: 303 (1997). 52. M. Inoue. M. Norimoto. M. Tanahashi, and R. M. Rowell, Wood Fiber Sci., 25: 224 (1993). 53. W. Dwianto. F. Tanaka, M. Inouc, and M. Norimoto, Wood Res., No. 83:47 (1996). 54. H. E. Hsu, W. Schwald. J. Schwald, and J. A. Shields, Wood Sci. Technol.. 22: 281 (1988). 55. E. Obataya. M. Sugiyama.and M. Norimoto, Wood. Res, No. 83: 40 (1996). 56. M. Takayanagi. H. Harima, and Y. Iwata, Rep. Prog. Polymer- P l y . Jpn., 6: 1 13 ( 1963). 57. M. Sugiyama, E. Obataya, and M. Norimoto, Wood Res., No. 82, 31 (1995). Mokrrvti Grrkkoishi, 42: 1049 (1996). 58. M.SugiyamaandM.Norimoto.

598

Norimoto

H. Becker and D. Noack, Wood Sci. Techno/., 2: 213 ( 1968). T. Sadoh, Wood Sci. Techno/., 15: 57 (1981). E. Obataya, M. Yokoyama, and M. Norimoto, Mokuzui Gakkrcishi, 42: 243 (1996). M. Norimoto and T. Yamada. M o k u w i Grtkkaishi, 23: 99 (1977). G. Zhao, M. Norimoto, and T. Yamada, Mokuzrri Cukkrrishi. 36: 257 (1990). M. Norimoto and G. Zhao, Mokuzcti Gakknishi, 39: 249 (1993). G. P. Mikhailov, A. I. Arthyukhov, and V. A. Shevelev, Polwter Sci. USSR, 11: 628 (1969). M. Norimoto, Wood Res., No. 59/60: 106 (1976). M. Kimura and J. Nakdno, J . Polyrner Sci., Polymer Lett. Ed.. 14: 741 (1976). T. Morooka, M. Norimoto, T. Yamada, and N. Shiraishi. Wood Res., No. 72: 12 (1986). T. Morooka, M. Norimoto, T. Yamada. and N. Shiraishi, Wood Res., No. 69: 61 ( I 983). N. G. McCrum, B . E. Read, and G. Williams, Artelmtic crntl Dielecfric eflec'fs in Po/ytneric' Solids, Wiley, New York, p. 240 (1967). 71. H. Tarkow, W. C. Feist, and C. F. Southerland, Forest Prod. J.. 16: 61 ( 1966). 72. Y. Tominagd and T. Sadoh, Mok~rzniGukkrriski, 36:263 ( 1990).

59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70.

Chemical Modification of Cellulose Akira lsogai The University of Tokyo, Tokyo, Japan

1.

INTRODUCTION

Looking back on the history of the science and technologies of cellulosic materials, we realize that numerous technologies for modifying cellulose by chemical and mechanical methods have been developed. Some of these technologies have been practically applied to produce modified cellulosic materials for our daily and industrial necessities. Cellulosic materials are generally strong, hydrophilic, insoluble in water, stable to chemicals, safe to living bodies, reproducible, recyclable, and biodegradable. With these specific and advantageous characteristics of cellulose, “modification” techniques to reinforce these original properties or to add new functionalities to cellulose have been investigated, and have contributed to the development of cellulose science and technologies. Chemical treatments have been positioned at the center of the field of cellulose modification. The purposes of chemical modifications or derivatizations of cellulose are divided into three major groups: ( l ) to produce modified cellulosic materials having specific properties at commercial levels, ( 2 ) to characterize cellulosic materials at laboratory levels, and ( 3 ) to approach the nature of cellulose for scientific interests. Commercial cellulose derivatives having water solubility, organic-solvent solubility, ion-exchanging groups, or hydrophobic groups are used as aqueous thickeners, plastics, column-supporting materials for chromatography, and others. Cellulose carbamates and nitrates are used for determining molecular mass and molecular mass distribution of cellulose to evaluate cellulosic materials by size-exclusion chromatography. Furthermore, since there are still many unsolved and mysterious subjects in cellulose science, chemical modifications are sometimes used to study fundamental research subjects of cellulose, such as solid-state structures of cellulose (hydrogen bonding patterns, chain conformations, crystal and amorphous structures, etc.), interactions with other substances at molecular levels, and molecular dynamics in solution states. The typical modifications of cellulose are esterifications and etherifications at hydroxylgroups of cellulose. Most water-soluble and organicsolvent-solublecellulose derivatives are prepared by these substitution reactions, and drastic changes in the original properties of cellulose can usually be achieved by thesechemical modifications. Othersare ionic and radical grafting,acetalation,deoxyhalogenation,andoxidation. Since the usual cellulosic materials originating from wood and cotton pulps have aldehyde and carboxyl groups in quite small quantities, depending on the purity of the pulps, 599

lsogai

600

these minor groups are also target positions for chemical modifications. Figure l shows schematic representation of positions in the cellulose structure for chemical modifications. The methods for chemical modifications of cellulose are various, depending on the reactions. Cellulose esters such as cellulose acetate and nitrate are generally prepared from cellulose pulps with reagents in the presence of concentrated sulfuric acid as an acid catalyst. In the case of cellulose nitrate preparation, degree of substitution (DS) is controlled by water content in the reaction medium, the H,SO,-HNO, system, and the shape of the solid cellulose pulp is maintained during the nitration. Cellulose triacetate is prepared from solid cellulose pulp in a mixture of acetic anhydride, acetic acid, and sulfuric acid.The reaction mixture becomes transparent by dissolving celluloseacetate in the medium, as the DS values are increased by acetylation. Water is then added to the mixture to hydrolyze acetyl ester groups, in part for preparing cellulose diacetate, and this partial

Substitution reaction

Esterification Etherification Deoxyhalogenation

To carboxylic acid To aldehyde Acetalation

Oxidation

1

Acid hydrolysis Oxidative cleavage

H0 OH

OH

Etherification Deoxyhalogenation Acetalation

Reactions at

minor groups

Carboxyl group Aldehyde group

Esterification Amidation Reduction to alcohol Oxidation to carboxylic acid Reduction to alcohol

FIGURE 1 Positions in cellulosestructure for chemicalmodifications.

Chemical Modification of Cellulose

601

deacetylationproceeds under homogeneoussolution-stateconditions in the mixture. On the other hand, most cellulose ethers such as methyl-, ethyl-, carboxylmethyl-, hydrolyethyl-, and hydroxypropyl-celluloses are prepared from cellulose pulps through the alkalicellulosestage with or without i-propyl alcohol.Then, an etherifying reagent such as methyl chloride, ethyl chloride, monochloroacetic acid, ethylene oxide, or propylene oxide is added to the mixture for substitution reactions. Heterogeneous solid-liquid phase reactions are maintained during these etherifications of alkalicellulose. Degrees of substitution (DS) and distribution of substituents as well as molecular mass and molecular mass distribution strongly influence the properties of cellulose derivatives, and are significant factors for characterizing them. In the case of cellulose acetate, the controlling DS to a certain value i n the range 0.5-2.0 is difficult as long as conventional acetylation and the followingdeacetylationprocessesareadopted. DS values of cellulose ethers are controllable to some extent by selecting the etherification conditions. However, since the etherifications proceed heterogeneously to swollen alkali cellulose in solid-liquid phase,homogeneousdistribution of substituents is generally difficult to achieve. Here, the concept of distribution of substituents has the following different levels: ( I ) distribution of substituents among the three hydroxyl groups of one anhydroglucose residue, (2) that along one cellulose chain, (3) that among cellulose chains, and (4) that among fibrils or fibers (Fig. 2). Heterogeneous distribution of substituents leads to fluctuation of properties of the celluloseethers,eventhosehaving the same DS values. Furthermore, some new functionalities may be expected from cellulose ethers having homogeneous distribution of substituents. Since esterifications and etheritications are condensation reactions at hydroxyl groups of cellulose, the presence of water in the reaction medium generallycauseslower reaction efficiency by consumingreagentsto form byproducts. Therefore, preparation of cellulose ethers with high DS is difficult a s long as aqueous alkalicellulose systems are used. Based on this background, control of DS and distribution of substituents (or preparation of regioselectively substituted cellulose derivatives) have been extensively studied for cellulose derivatives from fundamental aspects. In some studies, homogeneous cellulose solutions were used as derivatization media, in place of heterogeneous solid-liquid phase reactions, in order to prepare regioselectively substitutedcellulosederivativesor those having high DS values. Various cellulosic materials have been investigated a s the starting cellulose samples for derivatizations (Table l ) . Softwood and hardwood bleached kraft pulps from low to high a-cellulose contents, linter cellulose, and bacterial cellulose have been used as native cellulose samples. Sometimes, as swollen or decrystallized cellulose samples, native celluloses pretreated with liquid ammonia, organic amines, or aqueous inorganic alkali (such as 15-20% ay. NaOH) were subjected to derivatizations. Since regenerated cellulose has generally lower crystallinity and higher accessibility to reagents than native celluloses, it was used in some reports as the starting cellulose material for derivatizations. Regenerated low-molecular-weight celluloses with degrees of polymerization (DP) of 7 and 15 can be prepared in good yields by dissolving native cellulose samples in about 85% phosphoric acid followed by pouring the solution into water or methanol [ l ] . These celluloses have highly crystalline cellulose I1 structure. Regenerated amorphous celluloses having stable amorphousstructuresevenunderaqueous media can be prepared by dissolving native cellulose samples with various DP values in SOz-ami~~e-dimethylsulfoxidesolvent systems followed by pouring the solution into water [2]. These low-DP crystalline celluloses and completely amorphous ones were used as cellulose samples for derivatization. Since commercial cellulose diacetate and some cellulose ethers are soluble in dimethylsulfoxide

602

lsogai

Among three OH groups

Along one cellulose chain

Among cellulose chains

Within one microfibril, and among microfibrils

Within one fiber, and among fibers FIGURE 2

Distribution of substituents at various levels.

and other organic solvents, these are also examined as the starting materials for derivatizations. From industrial and environmental points of view, the possibility of using lowera-cellulose-content pulps, especially for producing cellulose acetate, is a quite significant subject. Since detailed and systematic information about chemical modifications of cellulose have been already reviewed by Professor Ishizu in the first edition of this book [3], scientific and technical reports in this research field published after 1990 are supplemented in the following sections.

Chemical Modificationof Cellulose

603

TABLE 1 Cellulosic Materials and Cellulose Solutions Examined for Derivatizations in 1990- 1997 Native cellulose

Bleached softwood and hardwood chemical pulps Cotton linter with high a-cellulose content Bacterial cellulose Microcrystalline cellulose powder

Decrystallized cellulose

Native cellulose treated with liquid NH, Native cellulose treated with organic amine Native cellulose treated with aqueous alkali, including mercerized cellulose

Regenerated cellulose

Defibrated rayon (from cellulose xhantate solution) Cellulose regenerated from Cu(EDA),(OH), or Cu(NHA(OH), solution Porous regenerated cellulose beads for chromatography Amorphous cellulose regenerated from S02-amine-DMS0 solution Cellulose prepared by saponification of cellulose acetate Low-molecular-weight cellulose with DP of 7 or 15

Cellulose derivatives

Cellulose diacetate Cellulose ethers (MC, HPC, HEC, CMC, etc.) Methylolcellulose prepared from PF-DMSO system Triphenylmethylcellulose Cellulose p-toluenesulfonate Halodeoxycellulose Dialdehydecellulose

Cellulose solutions

LiCI-DMAc LiBr-DMAc N@-DMF Chloral-pyridine-DMF PF-DMSO SO,-amine-DMSO

~~~

~

EDA,ethylenediamine;MC,methylcellulose;HPC,hydroxypropylcellulose;HEC,hydroxyethylcellulose; CMC,carboxylmethylcelluloseNasalt; PF, paraformaldehyde; DMSO,dimethylsulfoxide;DMAc,N,N-dimethylacetamide; DMF, N,N-dimethylformamide.

II.

ESTERIFICATION

On the subject of esterification, new processes or methods for preparing conventional cellulose esters in laboratory and industrial levels, preparation of new cellulose derivatives, applications to new analytical methods, and others have been reported. From the industrial viewpoint, developments of new processes for producing cellulose diacetate (DS = 2.5) from bleached sulfite pulps with lower a-cellulose content are quite innovative. On the other hand, for fundamental aspects, the LiCl-DMAc cellulose solvent system has been widely used as a homogeneous esterification medium to prepare new cellulose derivatives

lsogai

604

or to control DS and distribution of substituents. Carbanilation or preparation of ccllulose carbamates is included in this section for convenience.

A.

New Acetylation Process

Cellulose diacetate is being manufactured in thc largest quantity among cellulose derivatives, and is used as fibers for cigarette filters, cloths, plastics, and others. In order to reduce the process energy in acetylation and to utilize the lowest-a-cellulose-content wood pulps possible, a new acetylation process has been developed [4,5]. This includes ( I ) an acetylation stage at high temperature under reduced pressure together with accurate computer control of the temperature, (2) a deacetylation stage at high temperature, and (3) successive flash evaporation stage for separation of concentrated acetic acid from the reaction mixture. Eventually, this process leads to huge reductions in process energy and amounts of reagents, and to great improvement in productivity. Furthermore, sulfite pulps with lower a-cellulose content for not only softwood but also hardwood have come to be applicable to cellulose diacetate production by the new process. When conventional acetylation is applied to softwood sulfite pulps with low a-cellulose content, a relatively large quantity of the acetone-insoluble fraction, which originates from hemicellulose acetate and causes serious problems in the spinning process. is formed into cellulose diacetate. On the other hand, thenew process can extremely reduce this fraction, because hemicellulose acetate becomes soluble in acetone by partial depolymerization under the acetylation process at higher temperature. The difference between the conventional and new acetylation processes is illustrated in Fig. 3 [4]. Ueda et al. studied detailed structures of the acetone-insoluble fraction of cellulose diacetate prepared from softwood sulfite pulp with low a-cellulose content, by size-exclusion chromatography and some analytical techniques [ S ] . The mechanisnl proposed for the formation of the acetone-insoluble fraction is connected with the interaction between cellulose and hemicellulose partly present in the native wood state. Namely, a part of cellulose and hemicellulose molecules are chemically and/or physically bound to each other in the original pulp and native wood as well, and acetylation of these cellulose molecules proceeds with hen~icellulosethereby entangled. Under these restraining conditions, cellulose triacetate having the crystal structure of cellulose triacetate I instead of 11 is formed, and is hard to deacetylate i n the following hydrolysis stage. Generally, the structure of cellulose triacetate I, which can be defined from X-ray diffraction patterns, is formed by solid-phase acetylation of cellulose using, for example, toluene as a poor solvent of cellulose triacetate. Hence, cellulose triacetate having low solubility in acetone forms a gel-like and acetoneinsoluble fraction coagulated with hemicellulose acetate at molecular levels. The new acetylation and deacetylation processes at high temperatures must release the cellulosehemicellulose interactions present in the pulp by partial degradation of hemicellulose, and thus the acetone-soluble fraction increases even lor sulfite pulps with lower a-cellulose content. Saka et al. 17-91 further studied some analyses ofthe insoluble fractions of cellulose triacetate prepared from softwood and hardwood bleached kraft pulps, and proposed some techniques to reduce the insoluble fraction by pretreatment of the pulps with an acetic acid-sulfuric acid mixture. At the present timc, bleached softwood and hardwood sulfite pulps with a-cellulose content from 98% down to about 93% have come to be utilized for cellulose acetate manufacturing at industrial levels by the above new acetylation process. Further innovative techniques, by which normal bleached kraft pulps can be used in acetylation, will hopefully be developed someday in future.

605

Chemical Modificationof Cellulose

Conventional process ~. ...............

,

.................. :ii .............. H,O i ~

/ ................

Dissolvlng pulp Lmter or pulp

:H

t

1

I

cellulose De-acetylation Crude diacetate

T

I

trlacetate

I

-

I ,

HP1

I oniatA;; process

purification process Separation and

i

H,S;

I

I

.......... ........., .....................

/

' ! A70 A;OH Ac,O production orocess

I

.............. .................. i DilutedAcOH j

~.................... . ........................................ . . . .....................

4

1 4

~

~

I

G,-

............ ........:

{ ACOHprocess recovery I

A

;

AcOH

L.....

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

I

New process .......................

:

L

>

I .................... Steam i

IChernc la1 l1

v Acetylation at

4 ...................

.......................................... . . . ................... :......H,SO, Ac,O AcOH .................c ....., ..................... L

~

t

i

Ac,O production

,

~

l

I

-.......... ........

iL.. .................. AcOH

process

I

FIGURE 3

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

New and conventionalacetylation

processes 141.

B. Other New Esterification Methods In the case of cellulose esters other than cellulose acetate, some were prepared by new methods at laboratory levels, often using cellulose solvents to prepare cellulose derivatives having various DS values or homogeneous distribution of substituents. Shimizu et al. found that cellulose esters having a wide range of DS values can be prepared with organic acid (or organic acid salt), pyridine, and p-toluenesulfonyl chloride (tosyl chloride: TsCI) with or without DMF [ 10,l l ] . In the case of acetylation of cellulose

606

lsogai

with sodium acetate, pyridine, and TsCl in DMF, cellulose acetates with DS 1.4-2.8 were obtained, depending on the reaction conditions. The effect of molar ratios of TsCl/AcONa on the DS values was investigated in detail. Tri-0-substituted cellulose benzoates having various substituents at the benzene ring were also prepared by a similar method without using DMF. The substituents at the benzene ring had little influence on the DS values obtained. In these systems, the organic acid anhydrides or organic acid chlorides seem to be formed in situ in the organic acid-TsCI-pyridine mixture, and thus these esterifying reagents do react with cellulose hydroxyl groups in the presence of pyridine [ 121. The LiCl-DMAc cellulose solvent system was often used for homogeneous esterifications of cellulose in solution states. When N,N-dicyclohexylcarbodiimide,organic acids (or organic acid anhydrides) and 4-pyrolidinopyridine were added to acellulose/LiClDMAc solution, cellulose esters having DS values lower than 2.5 were prepared [ 13,141. In the case of cinnamoylation of cellulose in the LiCl-DMAc system, cinnamoyl groups were introduced almost selectively into C6 hydroxyl groups at DS values lower than about 1 [ 151. Cellulose propionates with DS values lower than 1.76 were prepared using the LiCI-DMAc system and propionyl chlorideasacellulose solvent and an etherifying reagent. Solution-state properties including liquid crystalline behavior of the products were studied, and rigidity of cellulose propionate molecules in solution states decreased with an increase in the DS values [16]. When cellulose is dissolved in N,O,-DMF or N,O,DMSOsystems, all hydroxyl groups finally form unstable nitrite esters in the solution [ 171, where the ease of nitrosation among the hydroxyl groups is in the order of C6 > C2 > C3. On the other hand, when the cellulose solutions are poured into water, all nitrite ester groups are finally removed from cellulose by hydrolysis to produce regenerated cellulose, where the ease of denitrosation is of the order of C6 >> C3 > C2 [ 181. Some water-soluble polysaccharides having sulfate esters have blood anticoagulant activity and also sometimes an anti-HIV one in the medical field. Preparation of watersoluble cellulose sulfate esters has been tried from these points of view. as well as ion exchanging and other specific properties suchasCMC in aqueoussolutions.Since the conventional sulfonation of cellulose with S03-DMF complex leads to severedepolymerization of cellulose, even though the DS values reach up to about 2, a homogeneous solution-state sulfonation with various reagents was examined [191. The results obtained showed that SOzClz gavethe highest efficiency of sulfonation when N,O,-DMF was used as the cellulosesolvent.Thesulfate groups were distributed asC6-OH > C2-OH on sulfonation at 20°C and C2-OH = C6-OH at -20°C. Permethylation analysis of cellulose sulfate with DS 0.52- 1.37 was carried out for elucidating distribution of sulfate groups among the three hydroxyl groups of one anhydroglucose residue by gas chromatography and I3C-NMR 1201. Since p-toluenesulfonate ester groups (tosyl esters) can be introduced selectively into primary hydroxyl groups of carbohydrates under certain conditions and also can be subjected to the following deoxy-substitution reactions, preparations of cellulose tosyl esters have been examined for a long time. When microcrystalline cellulose was reacted with TsCl and pyridine under heterogeneous solid-liquid phase conditions, the corresponding esters with DS 2.3 were obtained [21]. Arai and Aoki reported preparation of cellulose tosylates with DS 0.79-0.92 under heterogeneous (i.e.,solid-liquidphase)conditions, and the cellulose tosylates obtained were further reacted with sodium bisulfite to prepare cellulose deoxysulfonic acid with DS 0.065-0.1 [22]. However, the products obtained still contained tosyl groups tosome extent. The LiCI-DMAc cellulose solvent system was introduced to the solution-state homogeneous tosylation of cellulose by Heinze et al. [23.24]. Their DS values were in the range 0.4-2.3, and were controllable by the reaction

dificationChemical

of Cellulose

607

conditions. Thermal properties of these cellulose tosylates and further esterifications were examined using the cellulose tosylates with pyridine, sodium acetate, and esterifying agents (organic acid anhydrides). Regioselective tosylation at C6 hydroxyl groups of cellulose does not seem to be achieved even by the homogeneous tosylation, and severe depolymerization may occur on cellulose chains during the reactions.

C. New Cellulose Esters New cellulose esters were prepared using new reagents and/or organic cellulose solvent systems in terms of both fundamental and practical aspects of cellulose modifications. Cellulose derivatives containing relatively long aliphatic chains such as ester groups were prepared with fatty acid chloride and pyridine under heterogeneous solid-liquid phase conditions, and their liquid crystalline properties were studied [2S]. When the LiCl-DMAc cellulose solvent system was used with p-toluenesulfonic acid and organic acid anhydrides, waxy cellulose esters having C12-C20 fatty acid ester groups in the DS range 2.8-2.9 were prepared. Liquid crystalline properties of the products and crystallization behavior among their long aliphatic side chains were studied [26]. lwata et al. prepared regioselectively substituted cellulose esters, 6-0-acetyl-2,3-di-0-propanoylcellulose and 6-0-propanoyl-2,3-di-O-acetylcellulose,from 6-O-tritylcellulose, which was prepared beforehand from cellulose with DP S7 using a cellulose/LiCl-DMAc solution. Their crystal structures were made clear by X-ray and electron-diffraction analyses [27-301. Although liquid crystalline properties of cellulose derivatives have not yet been applied at practical levels, the accumulation of such fundamental information contributes to our understanding of the nature of cellulose molecules in solid, mesophase, and liquid states. Fluorine-containing polymers have unique properties such as thermal resistance, water and oil repellency, small dielectricity, and clear piezoelectricity. From these aspects, introduction of fluorine-containing groups through ester linkages into cellulose has been investigated (Fig. 4). When cellulose was dissolved in trifluoroacetic acid at room temperature, trifluoroacetate ester groups were introduced almost regioselectively at C6 hydroxyl groups of cellulose [31]. On the other hand, cellulose trifluoroacetate esters with DS 1.5-2.1 and DP 170-800 were prepared by reactions of cellulose with a trifluoroacetic acid-trifluoroacetic anhydride mixture at 20-150°C for 4 h under high pressures [32]. The ester bonds are, however, unstable to moisture or water and hydrolyzed to some extent, when the cellulose trifluoroacetate esters are exposedto air atmosphere. Furthermore, depolymerization of cellulose is inevitable under such strong acid conditions, being especially influenced by water content in the reaction media. Cellulose was reacted with 4-perfluorononeyloxy phthalic anhydride and either pyridine or triethylamine as a base in a LiC1-DMAc cellulose solution, and the corresponding half-esters with DS < 2.1 were obtained [33]. Introduction of fluorine-containing groups into cyanoethylcellulose by esterification was also studied using fluorine-containing alkanoyl fluoride, and thermal fluidity. Physical properties of the obtained products (DS of fluorine-containing ester groups < O S ) were studied [34]. Perfluorooctanoate ester groups were introduced into hydroxypropylcellulose, and its unique liquid-crystalline properties were studied [3S]. Some cellulose triesters and tricarbamates have been practically applied as columnpacking materials for chromatographic enantiomer separation using the homochiral property of the cellulose backbone of the derivatives. This is one of the most exciting topics in these two decades for the subject of “functionalizations of cellulose.” Figure S shows representative cellulose derivatives used for this purpose [36].

608

lsogai

CF3COOH

Cellulose

Dissolution

U

(CF3CO)zO t CF3COOH

Cellulose

W

Cellulose-0-CCF, II

Dissolution

DS = 1.5-2.1

F

+

pyridine or triethylarnme

Cellulose

*

-q F&

perfluorononenyloxyphtharic anhydride

Cellulose-o

LiCI-DMAc DS < 2.1

cF3

COOH

1 .l ,2,2,3-pentafluoropropyloxy-

22-difluoroporpionly fluorlde t

pyridine or triethylamine

Cellulose

LiCI-DMAc

*

Cellulose-o-c,

0 I1 Fz Hz

/c, /c,

C

DS
pyridine or trlethylarnine

Cellulose

LiCI-DMAc

-

Cellulose-0-c,

0 II Fz Hz

/c, /c,

C F,

DS < 2.1

FIGURE 4

O

O

,CH~F C F2

/CF, C F2

Chemical structures of fluorine-containing cellulose esters.

m

o

c

,

Cellulose-0

3 Cellulose-0

y Cellulose-0O

>CH3 Cellulose-0

m

e

Cellulose-0

>Q

Cellulose-0

c

l

dification Chemical CellUlOSe

of

609

Other new celluloseesters were prepared fromcellulose or commercial cellulose derivatives with new reagents or methods. Reactions between cellulose andA'-oxazolinones in aPF-DMSOcellulose solution or in the presence of hydrazine gave unique cellulose esters 137,381. When CMC, ethylcellulose, and other commercial cellulose ethers were reacted with chloroformiate 2-hydroxymethyl methacrylate, new cellulose derivatives containing photosensitive groups with DS 0.07-0.165 were obtained. UV-light irradiation to the products led to disappearance of the double bonds, indicating that intra- and intermolecular cross-linking occurred in the product [39]. The behavior of photo-cross-linking of cellulose cinnamate and allylcellulose cinnamate was also studied [40]. Preparation of highly water-absorbable materials was attempted from cellulose sulfate by cross-linking with glutaraldehyde [41]. Sulfonation of CMC with CIS0,H or SO, in pyridine gave CMC sulfate with DS 0.4- 1.48 [42].

D. Analyses of Distribution of Substituents I n the case of cellulose esters with DS values lower than 3 and those containing substituents of more than two different groups (cellulose heteroesters) such as cellulose acetate butylate, distribution of substituents is one of the significant factors which influence the properties of cellulose esters for end use. 'H- and ',C-NMR have been used for obtaining the information about distribution of substituents among the three hydroxyl groups of one anhydroglucose residue, i.e.,C2-OH,C3-OH, and C6-OH. Kowsaka et al. determined distribution of sulfate groups of cellulose sulfate sodium salts by 'H- and '.'C-NMR, together with their correlation spectra which were prepared from cellulose with SO,-DMF complex 1431. In the case of cellulose acetate, the residual free hydroxyl groups were first esterified with propionic anhydride and pyridine, and the chloroform-solublecellulose acetate propionate obtained was subjected to I3C-NMR analysis for determining distribution of acetyl groups [44]. Cellulose heteroesters such as cellulose acetate butylate and 02-hydroxypropyl-0-methylcellulose acetate succinate were also analyzed by asimilar method L45.461.

E. Application of Cellulose Esterification to Analytical Methods Phenylcarbanilation, or preparation of tri-0-phenylcarbamated cellulose, has been used for conversion of cellulosic materials into tetrahydrofuran-soluble derivatives with phenylisocyanate and pyridine at about 70°C for 1-2 days, and their molecular mass and molecular mass distribution can be evaluated as phenylcarbamate derivatives by size-exclusion chromatography (SEC)[47]. For this purpose, depolymerization of cellulose must be avoided as much as possible. Evans et al. examined the optimum conditions for preparing tri-0-phenylcarbamated cellulose with less depolymerization, in terms of the addition of DMSO or DMF as a co-solvent or that of amines to the reaction media [48,49]. This phenylcarbanilation seems to be better than nitration of cellulose for SEC analysis, because H' ions are not formed during the former reaction. Evans proposed that depolymerization of cellulose occurring to some extent during the phenylcarbanilation is due to oxidative cleavage of the celluloseglycosidebond.This carbanilation requires relatively large amounts of cellulosic samples and the following isolation stage. Therefore, heterogeneous solid-liquid phase nitration with a mixture of nitric acid-phosphoric acid-phosphorous pentoxide has been sometimes applied to cellulosic samples in small quantity, such as bacterial cellulose obtained by cultivation under specific conditions, for determining molecular mass and molecular mass distribution of cellulose by SEC [50].

lsogai

610

F. Mechanisms of AKD and ASA Sizing of Cellulose The esterifications described so far in the above sections bring about drastic changes in cellulose properties by introducing ester groups, whose DS values are generally more than 0.5. On the otherhand, so-called reactive sizes have been practically used as wet-end chemicalsforaddingsuitable water repellency to hydrophilic paper sheets in alkaline papermaking systems, where CaC03 is present as a filler. Alkylketene dimers (AKD) and alkenylsuccinic anhydrides (ASA) are included in this category. These reactive sizes are added as cationic emulsions 0.1- 1.5 p m in diameter to paper stock. Paper sizing with the reactive sizes has the following characteristics. 1.

2. 3.

4. 5. 6.

7.

Addition of small amounts of these sizes to paper stock gives sufficient sizing to paper (generally 0.05-0.2% on dry weight of pulp). It takes generally within 1 min between the size addition stageto paper stock and the final rolling stage of dried paper. Even though a large amount of water is present in paper stock (l-2% pulp consistency), good sizing can be given to paper by the reactive sizes. Basic properties of cellulosic pulp fibers are unchanged by the size treatments; only surface modifications of cellulosic fibers occur. Neither isolation nor purification of the sized paper to remove by-products and others is required. Thus, the sizing of paper with the reactive sizes is a quite efficient modification method for surfaces of cellulosic materials. So far, the covalent bond formation between these reactive size molecules and hydroxyl groups of cellulose on pulp fiber surfaces have been believed to be essential for giving sizing features to paper so efficiently.

Figure 6 shows possible reactions occurring in paper sheets during the papermaking process. It is true that AKD and ASA form the corresponding esters with cellulose under nonaqueous conditions in the presence of a base catalyst. However, the most controversial question is whether the covalent bonds are really formed under such aqueous conditions as papermaking within a short time without reactions with water to form hydrolyzed byproducts. Thefollowing are the major reasons why covalent-bond formation has been supported by many researchers.

1. The reactive structures of AKD and ASA are necessary for efficient paper sizing; hydrolyzed products give no sizing effect at the same addition levels. 2. Sizingdegrees are unchanged evenafterSoxhletextraction of the AKD- and ASA-sized papers; sizingfeatures must be lost by the extraction, if the size components are present in the paper sheets without forming the covalent bonds with cellulose hydroxyl groups [51,521. 3. Some analytical studies such as FT-IR, "C-labeling of the size molecules, and solid-state "C-NMR analysis of the sized papers prepared thereby, indicated the presence of covalent bonds in the paper sheets [53,54]. In the case of usual esterifications of cellulose, additions of large amounts of reagents as well as careful control of water content in the reaction media are required. In contrast, since highly efficient esterifications can be achieved when AKD and ASA emulsions are used as reagents even in the presence of water, these reactions must be applicable also to other cellulosic materials as more efficient modification techniques, if the covalent bond formation is the true mechanism f o r the paper sizing. Detailed reexaminations of sizing

61 1

Chemical Modificationof Cellulose R-CHXH-CH-R*

*

I

Cellulose-OH

I

0

OH

0-Cellulose

ASA-cellulose half ester

CH, -CH o=c,

l F=O

o=y

R-CH=CH-CH-R*

I

l

R-CH=CH-CH-R* HO , (OH- )

ASA

y=o o=y OH

OH

ASAcid Cellulose-OH

R-CH,-C=O

I

i R- CH

I

-P

R-CH=F:

-c=o

R*

F=O 0- Cellulose AKD-cellulose P-keto ester

HO , (OH-)

AKD

R-CH,-fi-CH,-R*

0

CO,' CO,

Ketone

FIGURE 6 Possible reactions of AKD and ASA inpaper sheet.

mechanisms of paper by AKD and ASA were carried out using some analytical techniques, and the following conclusions were obtained 155-611. Synthesis of "C-labeled AKD and the following solid-state "C-NMR analysis of the handsheets prepared thereby revealed that nearly no covalent bonds are present in the AKD-sized handsheets, when resonance peaks were assigned correctly. Thus, the hydrolyzed products of the reactive sizes, i.e., ketones for AKD and alkenylsuccinic acid for ASA, must consequently contribute to sizing performance of paper. 2. Most size components, i.e., hydrolyzed products, are physically anchored to pulp fiber surfaces, and these components are extractable from fibers when the paper sheetsare defibrated at 60°Cin1% Tween 80, which is one of the nonionic surfactants. 3. The structures of AKD and ASA that are reactive with water are necessary for achieving homogeneously distributed hydrophobic sizingcomponents with smaller coagulants as possible on hydrophilic cellulosic pulp fiber surfaces by in-situ hydrolysis of the sizing components in paper. 4. In the pulpsuspensions,cationicsize emulsion particles are adsorbed on pulp fiber surfaces by forming ionic bonds between the cationic groups on the sizing emulsion surfaces and carboxyl groups on pulp fiber surfaces, which are present as minor groups of 0.02-0.08 mmol/g for bleached kraft pulps.

1.

612

lsogai

From the aspect of esterifications of cellulose, therefore, the sizing of paper with AKD or ASA is not included in the story of this chapter, “chemical modification of cellulose.” However, such efficient surface modification techniques of cellulose as paper sizing with reactive sizes may give some clues to developingnew techniques for modifying cellulosic materials efficiently with reagents that are reactive with “water” under usual conditions, where water is always present and has strong interactions with cellulose.

111.

ETHERIFICATION

Since ether linkages are more stable than ester linkages to acid and alkaline conditions, various cellulose ethers have been prepared for both practical and fundamental purposes. Especially in the case of commercial cellulose etherifications, control of DS values and distribution of substituents at various levels (Fig. 2) are quite important subjects for manufacturing, because generally etherifications proceed heterogeneously to solid cellulose pulps through swollen alkalicellulose.Thus,some modified processes to prepare commercial cellulose ethers having more homogeneous distribution of substituents have been developed at industrial levels. Also, regioselectively substituted cellulose ethers were prepared, sometimes using cellulose solvent systems or multistage reactions, to obtain fundamental information about cellulose ethers and cellulose itself. As to new cellulose ethers, hydrophobically modified water-soluble cellulose ethers have been found to have unique viscosity behavior in aqueous solutions.

A.New

CarboxymethylationProcess

Carboxymethylcellulose sodium salt(CMC) is produced in the largest quantity among water-soluble cellulose derivatives, and is utilized as a thickener or dispersant in various fields. Taguchi et al. studied the effect of distribution of substituents of CMC samples on stability of their solution properties, and a new carboxymethylation process of cellulose was developed on the basis of the results obtained thereby [62]. Distribution of substituents of CMC samples along one cellulosechain and that among cellulose chains were evaluated by cellulase degradation and electrophoresis, respectively. The results showed that CMC samples having more homogeneous distribution of substituents had higher stability of their solution properties to enzymes and salt concentrations, even at the same DS values and the same distribution of substituents among the three hydroxyl groups of one anhydroglucose residue. The new process had two key points for producing CMC having more homogeneous distribution of substituents: ( I ) the use of i-propyl monochloroacetate ester as the reagent instead of monochloroacetic acid or (2) separate additions of NaOH solutions to the reaction mixture. This process has been already utilized at practical levels. Other new carboxymethylation methods were reported for fundamental aspects. 60-triphenylmethylcellulose (tritylcellulose) was first prepared under homogeneous conditions using the LiCI-DMAc cellulose solvent system, and residual hydroxyl groups at C2 and C3 of tritylcellulose were partially or completely carboxymethylated.Carboxymethylcellulose partially substituted at C2 and C3 hydroxyl groups were obtained by detritylation of the carboxymethylated tritylcellulose with HCI.Liu et al. reported that these CMC samples became water-soluble at DS values more than 0.3. thus suggesting that distribution of carboxymethyl groups strongly influences the solubility behavior in water [63]. 2,3-Di-O-carboxymethylcellulosewas prepared from tritylcellulose by ( l ) repeated carboxymethylation with monochloroacetic acid sodium salt and powdered NaOH

Modification Chemical

of Cellulose

613

using a tritylcellulose/DMSO solution and (2) the following detritylation [64]. 6-0-Carboxymethylcellulose has not been prepared yet.

B. Other New Etherification Methods Some new etherification methods have been reported in order to prepare regioselectively substituted cellulose ethers and to clarify their solution properties or interactions with other materials. 2,3-Di-O-alkylcelluloseswere prepared from tritylcellulose dissolved in DMSO with the Corresponding alkyl iodide and powdered NaOH followed by detritylation of the 2,3-di-0-alkyl-6-0-tritylcellulosewith HCl [65]. Preparation of6-0-alkylcellulose required multistage reactions: ( l ) tritylation at C6-OH in LiCI-DMAc, (2) complete allylation of the tritylcellulose at C2-OH and C3-OH in DMSO, (3) HC1 gas treatment in CH,CI, for detritylation of the 2,2-di-O-allyl-6-O-tritylcellulose, (4) isomerization of the allyl groups in the 2,3-di-O-allylcellulose to l-propenyl ones with potassium t-butoxide in DMSO, ( 5 ) alkylation at C6-OH of the 2,3-di-O-propenylcellulosewith the corresponding alkyl iodide in DMSO, and (6) treatment with 0.1 N HCl in 90% methanol to remove the 1 -propenyl groups of the 2,3-0-propenyl-6-0-alkylcellulose[66]. Methylcellulose with DS 0.9-2.2 was prepared under homogeneous conditions with methylsulfinyl carbanion, which was formed from NaH and DMSO, using a LiCI-DMAc cellulose solution [67]. Residual hydroxyl groups at C2-OH and C3-OH in tritylcellulose were partially or completely methylated with powdered NaOH and methyl iodide in DMSO, and methylcellulose samples with DS 0.53-2 at C2 and C3 were obtained by detritylation of the methylated tritylcellulose. The minimum DS value required for complete dissolution of methylcellulose in water was investigated using these samples [68]. Cellulose was reacted with propylene oxide to prepare hydroxypropylcellulose in a homogeneous cellulose/LiCl-DMAc solution or in a heterogeneous alkalicelluloseli-propanol system. Their solubility in water, as well as liquid-crystalline behavior, were compared in terms of distribution of substituents. Hydroxypropylcellulose prepared under heterogeneous conditions had lower solubility in water, high viscosity and some surfaceactive properties, compared with that prepared under homogeneous conditions [69]. The following three methods for alkylation of all hydroxyl groups of cellulose to prepare tri0-alkylcellulose ethers by one step have been reported so far, and were used for permethylation analysis of cellulosic materials and regioselective etherifications:

I.

2. 3.

C.New

Alkylation of cellulose with the corresponding alkyl iodide or alkyl bromide and powdered NaOH in homogeneous cellulose/SO,-diethylamine-DMSO solutions [70,7 11 In-situ deacetylation and simultaneous alkylation of cellulose acetate in homogeneous DMSO solutions with the same reagents as the above 1721 Alkylation of regenerated cellulose dissolved in LiCI-DMAc with alkyl iodide and methylsulfinyl carbanion [73]

CelluloseEthers

Preparation of hydrophobically modified water-soluble cellulose ethers and characterization of their solution properties are significant current topics for cellulose ethers [74]. For example, commercially available hydrophobically modified hydroxyethylcellulose (HMHEC) is prepared from HEC by reactions with epoxides having long alkyl chains of C12C24. This HM-HEC contains l-2% hydrophobic groups, i.e., DS 0.01-0.03, and thus is

614

lsogai

Hydrophobic

FIGURE 7 1751.

Interaction between molecules of hydrophobically modified cellulose ethers in water

water-soluble and nonionic. The molecules of hydrophobically modified water-soluble cellulose ethers form network structures in water by hydrophobic association at more than about 0.2% concentration (Fig. 7) [75],and thus their solutions sometimes have extremely high viscosity compared with unmodified cellulose ethers (Fig.8) [74]. Since HM-cellulose ethers, including methylated and ethylated hydroxypropyl- and hydroxyethyl-celluloses, have both hydrophilic and hydrophobic structures in the molecules, they behave as sur-

15

I

l / / Hydrophoblcally modified 1 / hydroxyethylcellulose 1 / / l / / / / / / / / /

-

U! 1 0 Kl

a

v

.a m 0 0 v)

5

I

5

/

Normal

/ " 0 0

"

"

0

1

2

3

Concentration of polymer in water ("h)

FIGURE 8 Viscosity behavior of aqueoussolutions of hydrophobically modified hydroxyethylcellulose and normal hydroxyethylcellulose [74].

Modification Chemical

of Cellulose

615

factants in water and have some interactions with other compounds in water. Therefore, the solution properties of these HM-cellulose ethers have been studied in terms of temperatures, phase-separation behavior, and interactions with water, surfactants, or latex particles, which have hydrophobic sites on their surfaces [76-791. The unique solution properties due to HM-cellulose ethers may be further utilized in the future in various fields. Pentyl ether of hydroxypropylcellulose and its cross-linked gels were prepared in nonaqueous solutions, and their thermotropic liquid-crystalline behavior was studied on the basis of changes in the optical pitch by heating [80]. Also, liquid-crystalline properties of cellulose ethers having long alkyl chains (cellulose-0-C,,,H,,,,-OH:m = 12-24) were studied by Liu et al. [81]. Comb-shaped amphiphilic cellulose derivatives, 0-(2-hydroxy3-butoxypropyl)cellulose, with MS 0.4- 1.4 were prepared with butylglycidilether in celluloseLiC1-DMAc solutions. The substitution occurred at the three hydroxyl groups in the order C6 > C2 >> C3, and the products were soluble in water or DMSO, depending on their DS values [82]. When cellulose was reacted with p-methoxytriphenylmethyl chloride in a cellulose/LiCl-DMAc solution at 25-70°C for 4-96 h, the corresponding cellulose ether-like tritylcellulose with DS of about l was obtained. This cellulose ether is soluble in DMF, DMAc, and DMSO [83]. Amine group-containing cellulose ethers were prepared fromepoxide group-containing cellulosederivatives by reaction with hexamethylene diamine or polyethylene imine, and the products obtained had some antimicrobial activity [84]. Cellulose ethers containing Michler's ketone groups as substituents of DS 0.56 were prepared, and their photoregulable properties were investigated [ U ] . SUIfoethyl groups were introduced into cellulose by reacting cellulose with BrCH,CH,SO,Na and powdered NaOH in the cellulose/SO,-diethylamine-DMSO solution or by reacting celluloseacetate with the same reagents in the celluloseacetate/DMSO solution. The water-soluble products prepared by the latter method had some anti-HIV activity [86]. Further derivatizations of CMC have been studied in some reports. Etherifications of CMC with diethylaminoethyl chloride HCl salt or 3-chloro-2-hydroxypropyltrimethylammonium chloride gaveamphoteric cellulose derivativescontaining both carboxyl groups and either tertiary amine or quaternary amine groups. Size-exclusion chromatography of these amphoteric cellulose derivatives indicated that some hydrophobic interactions were present among the molecules in the diluted aqueous solutions [87,88]. Carboxyl groups of CMC were partly amidated by hexadecylamine in DMSO containing dicyclohexylcarbodiimide, and the products obtained were analyzed by I3C-NMR after acid hydrolysis [89]. CMC xanthate was prepared from CMC by the addition of NaOH and CS, in the aqueous CMC solutions, and it had some metal-adsorptive properties [90]. CMC is soluble in formic acid, forming CMC formyl ester with DS0.4-2.0[91].Some new cellulose ethers reported recently are illustrated in Fig. 9.

D. Analyses of Distribution of Substituents As described previously, distribution of substituents of cellulose ethers at various levels has great influence on their properties, and various analytical methods have been proposed. Although average distribution of substituents of CMC among the three hydroxyl groups of one anhydroglucose residue have been established by the 'H-NMR method using 50% D,SO,/D,O [92], more detailed information about distribution of substituents was obtained by the sequence permethylation-reductive degradation-acetylation-GC analysis [93] and "C-NMR analysis [94]. Distribution of substituentsalongonecellulose chain and that among cellulose chains for CMC were measured by the previously described methods in Section 1II.A [62]. Distribution of substituents along one cellulosechain in methylcellulose

616

lsogai NMe2

l

Cellulose - 0 OH

Cellulose- 0 - CH , 2,

-OH

(m = 12-24)

NMe,

-(

Cellulose- 0 CH2CH2 f OCH, Cellulose- 0 -(CH2CH2 fOCH2CH3 n

Cellulose- 0- CH2CHCH2-0

l

- CH2CH2CH2CH3

OH

Cellulose - 0 -CH2CHCH2-NH - CH2CH2CH2CH2CH2CH2 - NH,

I

OH Cellulose-0-CH2CH2-N

I

\CH2CH3

OCH2COONa

Cellulose-O-CH2CHCH2-N+-CH3

I

OCH2COONa OH

I

I

Hydroxyethylcellulose - 0 - CH2CH

I

OH Hydroxyethylcellulose- 0 -CH2CH2CH2CH2CH3

FIGURE 9 Chemical structures of new cellulose ethers.

was studied by partial acid hydrolysis of methylcellulose followed by GC and FAB-MS analyses of the hydrolyzates [95]. Distribution of substituents of hydroxyethylcellulose with high molarsubstitutionvalueswereanalyzed by permethylation, acid hydrolysis, acetylation, and then GC-MS method. Relative reactivity among C2-OH, C3-OH, C6-OH, and substituted ethoxyl-OH was evaluated for hydroxyethylcellulose with various DS values [96].

Chemical Modification of Cellulose

IV.

617

GRAFTING

Grafting is also one of the chemical modification methods of cellulose. In order to introduce fluorine-containing groups into cellulose by grafting, cellulose was reacted with perfluorooctylethylacrylate as a reagent and either ammonium persulfate or AIBN as a catalyst using acellulose/PF-DMSOorcellulose/LiCl-DMAc solution [97]. The grafting efficiency was evaluated for the products, and the PF-DMSO system with ammonium persulfate gave good results. On the other hand, aqueous and heterogeneous solid-liquid phase reactions gave poor efficiency in cellulose grafting. Thermal and chemical properties of the grafted products were studied from various aspects [98]. Grafting of cellulose acetate with N-vinylcarbazol or epoxide-containing cellulose derivatives with various vinyl monomers was also investigated [99,100]. When cellulose was reacted with diethylaminosulfur trifluoride (DAST) in a cellulose/LiCl-DMAc solution, a polysaccharide having C6-0-C1 branch structures of the cellulose backbone was obtained [ I O I 1.

V.

DEOXYHALOGENATION

Preparation of cellulose derivatives having C-X (X: halogen) groups have been studied often using nonaqueous cellulose solvent systems(Fig. lo), and in some reports these deoxyhalogenated cellulose derivatives were further subjected to substitution reactions to add functionalities to cellulose. Deoxychlorination of cellulose was carried out in a cellulose/LiCl-DMAc solution with N-chlorosuccinimide-triphenylphosphine (TPP). At the early stage the reaction occurred only at C6-OH, and then it also occurred at C3-OH with Walden inversion as the deoxychlorination proceeded. The maximum DS value was 1.86 by this chlorination [ 1021. When cellulose was reacted with sulfuryl chloride in a cellulose/LiCl-DMAc solution, deoxychlorination occurred at C6-OH and C3-OH, with Walden inversion, up to DS 1.8. In this case, however, sulfur-containing groups were also introduced into cellulose [103]. Since cellulose is soluble in the LiBr-DMAc system, deoxybromination was performed with N-bromosuccinimide-TPP in acellulose/LiBr-DMAc solution. The products obtained had deoxybromide groups only at C6 with DS 0.9 [104]. Although cellulose was reacted with tribomoimidazole-TPP under heterogeneous solid-liquid phase conditions, the products had deoxybromide groups with DS less than 0.6 [ 1051. On the other hand, when the homogeneous deoxybromination was adopted with the same reagents using a cellulose/LiBr-DMAc solution, DS values of deoxybromide groups reached up to 1.6 by reacting at C6-OH and C3-OH with Walden inversion [106]. Halodeoxycelluloses thus prepared were further subjected to nucleophilic substitution reactions with tiols. inorganic compounds, or long aliphatic amines [ 107- 1091. 6-Deoxyfluorocellulose acetate with DS less than 0.6 was prepared without depolymerization of cellulose by reacting cellulose acetate with diethylaminosulfur trifluoride (DAST) in dioxane [ 1 IO]. Regioselectively substituted 6-deoxyfluorocellulose with DS 0.9 was prepared through the following steps: ( I ) preparation of 6-0-tritylcellulose using the LiCI-DMAc cellulose solvent system, ( 2 ) acylation (benzoylation or acetylation) of the tritylcellulose to prepare 2,3-di-0-acyl-6-0-tritylcellulose, ( 3 ) HBr treatment for detritylation, (4) reaction with DAST to prepare 6-deoxyfluoro-2,3-di-O-acylcellulose, and (5) saponification of the acyl groups with methanolic sodium methoxide [ 11 11.

618

lsogai

Q +O

N-Cl

LiCI-DMAc

OH

P

CH2CI

%?

Cl DS < 1.86 OH

di-2 CH20H

CH2CI SOpCI, LiCI-DMAc

-Q 1 1 1 1 1 l l l l l

c &-p 6*o c2 OH

Cl

OH

DS < 1.8

N-Br

CH20H

+ e

p

CH@

O

LiBr-DMAc

OH

OH DS = 0.9

Br

CH20H

CH2Br

Br

H

*

LiBr-DMAc

OH

Br

OH

DS < 1.6

1) Tritylation 5) De-acylation in LiCI-DMAc in MeONdMeOH 2) Acylation with AcPOor BzpO

TH20H

HBr

OH

3) De-tritylation with F 4) Fluorinationwith F-S-NEt, I DAST F

CH2F

l

OH DS= 1

FIGURE 10 Deoxyhalogenation of cellulose.

VI.

OXIDATION

Many studies about cellulose oxidation have been done, concerning both chemical modifications of cellulose and oxidative bleaching of residual lignin in chemical pulps. Oxidation reactions applied to cellulose for chemical modifications in this decade are summarized in Fig. 11. Some oxidation reactions occur on cellulose selectively at particular

619

Chemical Modificationof Cellulose CH20H

C CH O

YHC

OH

I

I

I

NaBH,

I

NaCIO,

c2

CH20H

+ side reactions

fQ

N204in CHCI,

*a OH

OH CH20H pH 10-11

TEMPO-NaBr-NaCIO

OH

Pulp-CH0

OH

HCIO,, pH 4-5

*

Pulp-COOH

FIGURE 11 Oxidation of cellulose.

positions. The periodate oxidation is typically involved in this category, and has been used to prepare dialdehyde cellulose at laboratory levels. Generally, the oxidation requires severaldays at room temperature in the darktopreparedialdehydecellulose from solid cellulose samples, whose C2-C3 bonds are mostly cleaved, and thus depolymerization is inevitable during the periodate oxidation. When acelluloseRF-DMSO solution was poured into methanol, partly methyloylated cellulose was obtained as a precipitate. Since

I

lsogai

620

this methyloylcellulose was soluble in water, the periodate oxidation proceeded homogeneously in the aqueous solution, and almost completely oxidized dialdehyde cellulose was obtained within 20 h [ 1121. Since the aldehyde groups of dialdehyde cellulose form intraand intermolecular hemiacetal linkages, the isolated and dried dialdehyde cellulose, even having 100% oxidized structures at the C2-C3 bond, is insoluble in water. The dialdehyde cellulose was oxidized to the corresponding dicarboxyl cellulose with sodium chlorite, or reduced to the corresponding dialcohol cellulose with sodium borohydride. These oxidized and reduced products were soluble in water, and characterized from various aspects [ 1 121 1 81. Dialdehyde cellulose was further modified to the corresponding hydroxamic acid derivatives by three-step reactions, and their behavior to form metal complexes in water was investigated [ l 19,120). Periodate oxidation was also applied to cellulose beads for chromatography [ 12 1 1. It is well known that primary alcohol groups of cellulose are partly converted to carboxyl ones by oxidation with N20, in chloroform. However, side reactions are inevitable during this N20, oxidation. On the other hand, recently a new water-soluble reagent, 2,2,6,6-tetramethylpipelidine- I -oxy1 radical (TEMPO), has become commercially available, and TEMPO can oxidize primary alcohol groups of water-soluble polysaccharides such as starch to carboxyl ones in good yields and selectivity in the presence of a cooxidizing agent at pH 9- l l [ 1221. The TEMPO-NaBr-NaC10 system was first applied to native cellulose by Chang and Robyt [ 1231, although their method could not give watersoluble cello-uronic acid [ 1241. Then the TEMPO-mediated oxidation of cellulose samples under various conditions was examined to prepare water-soluble cello-uronic acid having high DP. Although only small amounts of carboxyl groups were introduced into native cellulose samples by this oxidation, water-soluble cello-urionic acid sodium salts were obtained quantitatively by using regenerated and mercerized celluloses as starting materials (Fig. 12) [ 1241. DP values of the cello-uronic acids thus prepared were greatly influenced by the oxidation conditions, and partial depolymerization occurred on cellulose chains by p-elimination under alkaline conditions. Since cello-uronic acids prepared by the TEMPO system regularly consist of the glucuronic acid repeating unit, differing from the conven-

-C,OONa

l

220 200

I

I

180

I

160

I

140 120

I

I

loo

I

80

I

60

ppm

FIGURE 12 "C-NMR spectrum of cello-uronic acid Na salt dissolved in D,O 11241.

odification Chemical

of Cellulose

621

tional water-soluble cellulose derivatives, they may have some unique solution properties or bioactivity.

VII.

CHEMICAL REACTIONS AT MINOR GROUPS

Cellulosic fibers usually contain carboxyl and aldehyde groups in quite small quantities, depending on the purity of the fibers. Usual bleached kraft pulps have carboxyl contents of 0.02-0.08 mmol/g and aldehyde contents of 0.01-0.03 mmol/g. Cotton linter pulps have much lower values. The carboxyl groups originate from those present in the original hemicellulose and/or those formed during pulping (especially anthraquinone-added kraft pulping) and bleaching. The C 1 groups at the reducing ends of cellulose and hemicellulose chains are accounted forasaldehydegroups.Carboxylgroups in cellulosic fibers play significant roles in fiber processing treatments: dyeing and surface treatments in fiber processing, and efficient adsorption of wet-end additives on pulp fibers at the wet end in papermaking. Aldehyde groups in cellulosic materials can be oxidized with ClO; to carboxylic acids, and thus their carboxyl contents are increased by this oxidation [60]. Since watersoluble carbodiimide, WSC: 1-ethyl-3-(3-dimethylaminopropyl)carbodiimideHCI salt, has become commercially available, carboxyl groups can be selectively converted to amides by WSC and compounds having primary amine groups under aqueous conditions at pH 4.75. More than 95% of the carboxyl groups in a bleached kraft pulp were converted into nonionic or cationicamides with WSC and the corresponding amine in aqueous pulp suspensions (Fig. 13) [57-60,1251. Here, the cationic pulps prepared by this method can be regarded as true cationic pulps, because the cationic groups are introduced into pulps by blocking the anionic carboxylgroups. On the other hand, the usual cationic pulps prepared from normal pulps by etherification with, for example, diethylaminoethyl chloride HCI salt at hydroxyl groups, should be called “amphoteric” pulps. Fundamental properties such as pulp freeness were unchanged by the amidation, and thus only slight chemical modifications of pulp fibers can be achieved by this amidation. However, when handsheets were prepared from the nonionic pulp, which was prepared from normal bleached kraft pulp by amidation with methylamine and WSC, effects of wet-end additives, such as sizes, wet-strength resins, and retention aids,decreased drastically. This is because retention values of these additives clearly drop by blocking the anionic carboxyl groups with nonionic methylamide groups in the pulp fibers. Thus, carboxyl groups in pulp fibers, even though their quantity is quite low, behave as essential retention sites for wet-end additives in paper stock by ionic interactions [57-60,1251.

VIII.

CONCLUSION

As described above, chemical modifications of cellulose have been studied extensively from both fundamental and practical points of view by many researchers, and so much information about fundamental properties of cellulose and cellulose derivatives has been accumulated. Some results have been applied practically at industrial levels. Since cellulose always has some strong interactions with water, large amounts of energy are required for complete removal of water from cellulosic material. Thus, substitution reactions occurring selectively at hydroxyl groups of cellulose instead of water (or OH” ions) must be significant to achieve efficient substitution reactions in cellulose.

622

lsogai

Pulp-COOM (M = Na, Ca, H)

-----

t

PuI~-COOH"

PuIP-COOCH~

Solvent exchange from water to ether CH2N2 in ether through methanol and n-hexane

Pulp-COOM (M = Na, Ca, H)

1

EtN=C=N(CH2)3NMe2*HCI (WSC), pH 4.75

/ N+H-Et

o=c

&N+H-Et

Pulp-coo-c,

Pulp-CONH-R NH(CH,),NMe,

RNHz, pH 4.75 R : -CH3 (nonionic) R : -CH2CH2NHMe2 (tertiary amine) R : -NHNHCOCH2N+Me3 Cl- (quaternaryamine)

R : -CH2(CH2)&H3 (hydrophobic)

(hydrophobic)

FIGURE 13 Methylation and amidations of carboxyl groups in bleached kraft pulp [125].

Preparations of regioselectively substituted or oxidized cellulose derivatives and their characterizations may bring about development of cellulose derivatives having new functionalities. The nonaqueous cellulose solvent systems have often been used to prepare cellulosederivatives in laboratory levels. For practical purposes, these multicomponent solvent systems, usually having high boiling points, are hard to utilize unless the cellulose derivatives obtained have quite unusual functionalities. If highly regioselective reactions to the three hydroxyl groups of one anhydroglucose residue are applicable to solid cellulosic materials in one step under aqueous conditions, they must be valuable for practical use. Not only drastic changes in the properties of cellulose by esterifications or etherifications but also some efficient surface modification techniques of cellulose, such as paper sizing, may also be utilized for other purposes in functionalizations of cellulosic materials.

REFERENCES 1. 2.

A. Isogai and M. Usuda, Mokuzai Gakkuishi, 37: 339 ( 1 99 1 ). A. Isogai and R. H. Atalla, J . Polymer Sci., Polymer Chem. Ed., 29: 113 (1991).

ificationChemical 3. 4. 5. 6. 7. 8. 9. IO. 11. 12. 13. 14. 15.

16. 17. 18. 19. 20. 21. 22. 23. 24. 25. 26. 27. 28. 29. 30. 31. 32. 33. 34. 35. 36. 37. 38. 39. 40. 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51.

of Cellulose

623

A. Ishizu, in Wood and Cellulosic Chenlistly (D. N.-S. Hon and N.Shiraishi, eds.). Marcel Dekker, New York, p. 757 (1991). H. Yabune, Cellulose Commun., 4: 114 (1997). S. Z. Rogovina, V. A. Zhorin, and N. S. Enikolpiau, J . Appl. Polymer Sci., 57: 439 (1994). K. Ueda, S. Saka, Y. Funaki, and S. Soejima, Mokuzai Gakkaishi. 34: 356 (1988). H. Matsumura and S. Saka, Mokuzai Gakkaishi, 38: 270 (1992). H. Matsumura and S. Saka, Mokuzui Gakkaishi, 38: 862 (1992). S. Saka and K. Takahashi, Preprints '97 Cellulose R&D, Nagoya, p. 44 (1997). Y. Shimizu, A. Nakayama, and J. Hayashi, Cellulose Chem. Technol., 25: 275 (1991). Y. Shimizu, J. Nakayama, and J. Hayashi, Sen 'i Gakkuishi, 49: 352 (1993). Y. Shimizu, J. Nakayama,and J. Hayashi. Preprints '94 Cellulo.se R&D. Tokyo, p. 165 ( I 994). G. Samaranayake and W. G. Glasser, Carbohydr. Polymers, 22: 1 (1993). G. Sarnaranayake and W. G. Glasser, Carbdzydr. Polymers, 22: 79 ( I 993). A. Ishizu, A. Isogai, M. Tomikawa, and J. Nakano, Mokuzai Gakkaishi, 3 7 829 ( I 99 I). E. Bianchi, E. Marsano, L. Ricco, and G. Conio, Carbohydr. Polymers, 34: 91 (1997). A. Isogai, A. Ishizu, and J. Nakano, Sen'i Gakkaishi, 42: T654 (1986). W. Wagenknecht, I. Nehls, and B. Philipp, Carbohydr. Res., 237: 21 1 (1992). W. Wagenknecht, I. Nehls, and B. Philipp, Carbohyd,: Res., 240: 245 (1993). M. Gohdes, P. Mischnick, and W. Wagenknecht, Carbollydr. Polymers, 33: 163 (1997). C. Roussel, C. Popescu, and L. Fabre, Carbohydr. Res., 282: 307 (1996). K. Arai and F. Aoki, Sen 'i Gakkaishi, 50: 5 I O ( l 994). T. Heinze, K. Rahn, M. Jaspers, and H. Berghmans, J . Appl. Polymer Sci., 60: 1891 (1996). T. Heinze, K. Rahn, M. Jaspers, and H. Berghmans, Macromol. Chem., 197 4207 (1996). T. Itoh, Y. Tsuji, H. Suzuki, T. Fukuda, and T. Miyamoto, Polymer J., 24: 641 (1992). J. E.Sealey, G.Samaranayake, J. G.Todd,and W. G. Glasser, J . Polyner Sci., Polymer Chem. Ed., 34: 1613 (1996). T. Iwata, J. Azuma, K. Okamura. M. Muramoto,and B. Chun, Carbohvdr. Res., 224: 277 ( 1992). T. Iwata, K. Okamura, J. Azuma, H. Chanzy, and F. Tanaka, Cellulose, l : 67 (1994). T. Iwata, K. Okarnura, J. Azuma, and F. Tanaka, Cellulose, 3: 91 ( 1996). T. Iwata, K. Okamura, J. Azuma, and F. Tanaka, Cellu/o.se. 3 : 107 (1996). M. Hasegawa, A. Isogai, F. Onabe, and M. Usada, J. Appl. Polymer Sci.. 45: 1857 (1992). T. Liebert, M . Schnabelrauch, D. Klemm, and U. Erler, Cellulose, l : 249 (1994). M. Muramoto, M. Yoshioka, and N. Shiraishi, Sen'i Gukkaishi. 46: 496 (1990). K. Goto, M. Yoshioka, and N. Shiraishi, Mokrczai Gakkaishi, 37: 57 (1991). F. Guittard, T. Yamagishi, A. Cambon, and P. Sixou, Macromolecules, 27: 6988 (1994). A. Ohnishi and T. Shibata, Cell~clow Conmun.,4 : 2 (1997). S. Ciovica, V. Sunel, and N. Asandei, Cellulose Clwm. Technol., 28: 58 1 (1994). S. Ciovica, V. Sunel, and N. Asandei, Cellulose Chetn. Technol., 28: 493 (1994). R. Pernikis and B. Lazdina, Cellulose Chern. Techno/., 30: 187 (1996). M. Karnath, J. Kincaid, and B. Mandal. J. Appl. Polytner Sci., 59: 45 (1996). K. Arai and H. Coda. Sen 'i Gakkuishi. 49: 482 (1993). S. Vogt, T. Heinze, K. Rottig, and D. Klemm, Carbohydr. Res., 266: 315 (1995). K. Kowsaka. K. Okajirna, and K. Kamide, Polymer J., 23: 823 (1991). Y . Tezuka and Y. Tsuchiya. C d m h y d r : Res., 273: 83 (1995). Y. Tezuka, C~wlx~hydr. Res., 241: 285 (1993). Y . Tezuka, K. Imai. M. Oshima, and K. Ito, Carbohydr. Res., 222: 255 (1992). L. R. Schroeder and F. C. Haigh. Ttrppi J., 62: 103 (1979). R. Evans, R. H. Wearne and A. F. A. Wallis, J . Appl. Polyrner Sci., 42: 8 13 ( 1991 ). R. Evans. R. H. Wearne and A. F. A. Wallis, J. Appl. Polymer S r i , 42: 821 ( l 991 ). H. Shibazaki, S. Kuga and T. Okano, Cellulose, 4: 75 (1997). T. Lindstrom and G. Soderberg, Nordic P~rlp Paper Res. J.. I : 26 (1986).

624 52. 53. 54. 55. 56. 57. 58. 59. 60. 61. 62. 63. 64. 65. 66. 67. 68. 69. 70. 71. 72. 73. 74. 75. 76. 77. 78. 79. 80. 81. 82. 83. 84. 85. 86. 87. 88. 89. 90. 91. 92. 93. 94. 95. 96. 97.

lsogai J. C.Robertsand D. N. Garner, Tuppi J., 68(4): 118 (1985). L. Odberg, T. Lindstrom, B. Liedberg,and J. Gustavsson, Tuppi J., 70(4): 135 (1987). K. J. Bottorf, Tappi J.. 77(4): 105 (1994). M. Nishiyama, A. Isogai, and F. Onabe, Sen ’i Gakkuishi, 52: 180 ( 1 997). M. Nishiyama,A. Isogai and F. Onabe, Sen’i Gukkaishi, 52: 195 (1997). A. Isogai,M. Nishiyamaand F. Onabe, Sen’iGakkaishi, 52: 195 (1997). A. Isogai, in TheFundamentals of Pupermaking Materials, Vol. 2 (C. F. Baker, ed.). Pira International, Leatherhead, p. 1047 (1997). A. Isogai, J. Pulp Paper Sci., 23: 5276 (1997). A. Isogai, C. Kitaoka, and F. Onabe, J. PulpPuper Sci.. 23: 5215 (1997). A.Isogai, J . Pulp Paper Sci., 25(7): 251-256 (1999). A. Taguchi, T. Oomiya, and K. Shimizu. Cellulose Commun., 2: 29 ( 1995). H. Q. Liu, L.-N. Zhang, A. Takaragi, and T. Miyamoto, Macrond. Chern. Rupid Cotnrn~tn.. 18: 92 1 ( 1997). T. Heinze and K. Rottig, Macromol.Chem.Rapid Cornmun.. /S: 31 1 (1994). T. KondoandD.G. Gray, Curbohydr.Res., 220: 173 (1991). T. Kondo, CurbohydsRes.. 238: 23 I (1993). M. Hirrien, J. Desbrieres, and M. Rinaudo, Curbohydr. Pol~vners, 31: 243 (1996). H. Q. Liu, L.N. Zhang, A. Takaragi, and T. Miyamoto. Cellulose, 4: 321 (1997). T. Fukuda, T. Sat0 and T. Miyamoto, Sen’i Gakkuishi, 48: 320 ( 1992). A. Isogai,A. Ishizu, J. Nakano, S. Eda, and K. Kato, Curbohydr. Res., 138: 99 (1985). A. Isogai, A. Ishizu, and J. Nakano, J . Appl. Po/ytner Sci., 31: 341 ( 1986). T. Kondo, A.Isogai, A. Ishizu, and J. Nakano, J . Appl. Polymer Sci., 34: 55 (1987). L. Petrus, D. G. Gray, and J. N.BeMiller, Curbohydr. Kes., 268: 319 (1995). R.Tanaka, APAS7: 20: 10 (1996). R. Tanaka, J. Meadows,G. 0. Phillips,and P. A. Williams, Curbohydr.Polymers, 12: 443 ( 1990). R. Tanaka, P. A. Williams. J. Meadows, and G. 0 . Phillips. Colloid Surfaces, 66: 63 (1992). R. Tanaka, T. Hatakeyama,and H. Hatakeyama, Mucromol.Chem.. 198: 883 (1997). K. Thuresson, B . Lindman,and B . Nystrom, J. Phys. Chern., 101: 6450 (1997). K. ThuressonandB.Lindman, J. Phys. Chenz., 101: 6460 (1997). T. Yamagishi and P. Sixou, Polymer; 36: 23 15 ( 1995). H.-Q. Liu.L.-N. Zhang, A. Takaragi, and T. Miyamoto, Macrwmol. Chetn. Rupid Conzrnun.. 18: 921 (1997). H. Nishimura, N. Donkai,and T. Miyamoto, Cellulose, 4: 89 (1997). J. A. C. Gomez, U. W. Erler, and D. 0. Klemm, Mucromol. Chern., IY7: 953 (1996). 1. Ikeda, N. Tanaka, and K. Suzuki, Sen ’ i Gakkaishi, 48:332 ( 1 992). K.Arai and Y. Kawabata, Macromol.Chem., 196: 2139 (1995). T. Ishiguro, S. Inoue, G. Meshitsuka, A. Ishizu, K. Murakami, and K. Watanabe, S m ’ i Gakkuishi, S I : 571 (1995). G.Z.Zheng, G. Meshitsuka,andA.Ishizu, J. Polymer Sci., PolymerPhys. Ed., 33: 867 (1995). G. Z. Zheng, G. Meshitsuka, and A. Ishizu, Pol.vnzer; 37:1629 (1996). D. Charpentiev,G. Mocanu,A. Carpov, S. Chapelle, L.Merle, and G . Muller, Cc~rl,ohydr: Po/ymer.s, 33: 177 (1997). H. Zeng, W. Li, and Z. Li, J . Appl. Polymer Sci., 54: 1989 ( 1994). T. Heinzeand U. Heinze, Mucromol.Chem. Rupid Commun.. 18: 1033 (1997). F. F.-L. Ho and D. W. Klosiewicz, A n d . Chetn., 52: 913 (1980). S. G. Zeller, G. W. Griesgraber, and G. R. Gray, Curbohydr.Res., 211: 41 (1991). A. Baarand W. M. Kulicke, Macromol.Chem., 19.5: 1483 (1994). P.W. Arisz, H. J. J. Kauw, and J. J . Boon, Ccwbohydr. Res., 271: Cl5 ( 1995). P. W. Arisz, J. A, Lomay. and J. J. Boon, Carbohydr: Res.. 243: 99 ( 1993). H. Matsui and N. Shiraishi, Mokuzui Gakkuishi, 39: I188 (1993).

Chemical Modification of Cellulose 98. 99. 100. 101.

102. 103. 104. 105. 106.

107. 108. 109. 1I O .

111. 112. 113. 114. I IS. 116. 117. 118. 119. 120. 121. 122. 123. 124. 125.

625

H. Matsui and N. Shiraishi, Mokrr:rri Gakkaishi, 39: I 194 (1993). S. Basu. A. Bhattacharyya. P. C. H. Mondal,and S. N. Bhattacharyya. J . Po/yrncpr Sci.. Polyrncv Chern. Ed., 32: 2251 (1994). 1. Ikeda, Y. Yamada. N. Tanaka. and K. Suzuki, Sen’i Gcrkkai.shi, 48: 157 ( 1992). C. E. Frazier, S. L. Wcndler. and W. G. Glasser, Ccrrbohydr. Po/ymrrs. 31: 11 ( l 996). K. Furuhata, H.-S. Chang, N. Aoki. and M. Sakamoto, Carbohvclr. Res., 230: IS1 (1992). S. Furubeppu, T. Kondo, and A. Ishizu. Srn’i Gcrkknishi, 4 7 592 (1991). K. Furuhata, K. Koganei, H.-S. Chang, N. Aoki,and M. Sakamoto, Cor-bohyclr: Res.. 230: I65 ( 1992). N. Aoki. S. Suzuki, K. Furuhata, and M. Sakamoto, Sen’i G d k r t i ~ h i 50: . 515 (1994). K.Furuhata. N. Aoki. S. Suzuki, M. Sakamoto, Y. Saegusa, and S. Nakamura, Crrrboh)~dr: P o / y n f ~ ~ r26: . s , 25 ( 1995). N. Aoki, K. Koganei. H.-S. Chang, K. Furuhata,andM.Sakamoto, C d ~ o h y d Po/yrnr~-.v. ~: 27: 13 (1995). K. Furuhata, H.-S. Chang, K. Koganei. and M. Sakamoto, Sen’i Gakknishi, 48: 602 (1992). S . Nakamura and N. Sanada, Scw ’ i G~zkkniski,53: 467 ( 1 997). N.Kasuya, K. liyama, and A. Ishim, Cc~rbohydr.Res., 229: 131 (1992). N. Kasuya, K. liyama. G. Meshitsuka, and A. Ishizu, Crrrbohyd~:Res., 260: 251 (1994). T. Morooka, M. Norimoto. and T. Yamada. J . AppL Po/ymer Sci., 38: 849 (1989). T. Morooka and M. Norimoto. Sen’i Gakkuiski. 47: 328 (1991). A. J. Varma and V. B. Chavan, Ccrrboh~dr.Po/ynler.s, 2 7 63 (1995). M. I . Popa, N. Aclcnci. and G. lonescu, Cd/do.sc. Chern. T e c h n o / . . 30: 33 ( 1996). A. J. Varma and V. B. Chavan. Cc4ldose. 2: 41 (1995). K. Tihlarik and M. Pasteka, Cellrtlose Chern. Techol.. 2 7 267 (1993). E. Maekawa and T. Koshijima, J. App/. Po/yrner Sci., 42: 169 (199 I ). E. Maekawa and T. Koshijima, J . A p p / . P d y e r Sci.. 40: 1601 (1990). E. Maekawa and T. Koshijima, J. AppL P o / w e r Sci., 43: 417 (1991). K. Pommerening, H. Rein, D. Bertram, and R. Muller, CarboAydr. Res., 233: 219 (1992). A. E. Nooy, A. C. Besemer, and H. Bekkum, C(rrbokydr: Res.. 269: 89 (1995). P. S. Chang and J . F. Robyt. J . Crrrbohylr. C/wrn.. 15: 819 ( 1 996). A. Isogai and Y. Kato. Ce[l~r/o,se. S: 1S3 (1998). T. Kitaoka, A. Isogai, and F. Onnbe. Nordic P I / / / >Prrper Res. J., 10: 253 ( 1 995).

This Page Intentionally Left Blank

15 Chemical Synthesis of Cellulose Fumiaki Nakatsubo Kyoto University, Kyoto, japan

I. INTRODUCTION Cellulose is the most abundant natural organic polymer existing as a main plant cell wall component and is important as a biodegradable and renewable organic resource [l]. The study of cellulose, therefore, has continued for more than 150 years. However, there are still several problems which should be solved: biosynthesis, crystal structure, chemical synthesis, regiospecific substitution reactions, structure-function relationships of their derivatives, and so on [ 1,2]. The synthesis of cellulose has been a very important but extremely difficult problem to solve, sinceSchlubach first tried the synthesis in 1941 [3]. Recently, Kobayashi and co-workers reported enzymic synthesis of cellulose [4]. Their synthetic method with cellulase is important and interesting as the first in-vitro synthesis using enzyme. The method, however, does not satisfy the recent molecular design of cellulose derivatives having special functions,because it may not enable us to regiospecifically introduce the special functional group into only the desired hydroxyl groups in the repeating pyranose units of cellulose. There are many functional cellulose derivatives: cellulose esters and ethers having liquid crystalline properties [5] and chiral recognition ability [6], sulfated cellulose with anticoagulant activity like a heparin [7], branched cellulosederivatives with antitumor activity [8], and so on. Much is still unknown about the relationship between the structure and properties, which derivatives are more functional or active among those substituted at 2,3- or 6-positions. For these studies and,furthermore,for molecular design of the advanced materials from cellulose, methods which make it possible to prepare cellulose derivatives with functional groups at special positions among 2,3,6-hydroxyl groups in the repeating pyranose unit of cellulose are indispensable. Polycondensation and ring-opening polymerization methods using glucose derivative as a starting monomer satisfy the above requirements [9], but all trials synthesizing cellulose by these methods have been unsuccessful. Husemann and Muller [lo] and Hirano [ I l l reported the condensation of 2,3,6glucose tricarbanilate with phosphorus pentoxide in a mixture of chlorofoddimethylsulfoxideto give a cellulose-like polymer containing branched polymer and about 1% phosphorus. Micheel et al. [ 121 and Uryu et al. [ 131 tried cationic polymerization of 1,4anhydro-2,3,6-tri-O-benzyl-a-~-glucose initiated with various Lewis acids, but stereoregular (1 +4)-P-D-glUCOpyranan was not obtained. Furthermore, Uryu et al. [ 141 reported 627

Nakatsubo

628

the first synthesis of cellulose-type glycopyranan, (1-4)-/3-~-ribopyrananby cationic ringopening polymerization of 1,4-anhydro-a-~-ribopyranose derivatives. However, their strategy is not applicable to the synthesis of cellulose although it is useful for the preparation of glucan with the same hydroxylation pattern as ribose. Very recently, Kochetkov described in his review [9cl that Malysheva synthesized completely stereoregular (l+4)-pglucan from 1,2-O-cyanoethylidene derivative only at high pressure, but the paper describing the method in detail has not appeared. Recently, the author and his co-workers succeeded in the first chemical synthesis of cellulose [ 151. In this chapter, the first chemical synthesis of cello-oligosaccharides and cellulosederivatives by both astepwisesynthetic method and a cationic ring-opening polymerization, and conversion to cellulose by removing their protective groups, are described. These synthetic methods are now drawing attention in the world [ 16).

II. SYNTHESIS OF CELLO-OLIGOSACCHARIDES Cello-oligosaccharides are generally defined asa series of saccharidescontaining from dimer, cellobiose, up to cellodecaose [ 171, and have an important role as a bridge between commercially available glucose and cellobiose with low molecular weight, and cellulose with high molecular weight. For the elucidation of chemical structure, and chemical and biological reactivity of a complex macromolecule, model compounds have played a significant role in the history of lignin chemistry 1181. A few cases use such model compounds in the field of cellulose research, because the chemical structure of cellulose is simple. We may, however, discover new merits again in using cellulose model compounds to clarify several problems which still exist after about 150 years of cellulose research as described above. Thus,cellooligosaccharides are quite useful as the model compounds. These cello-oligosaccharides may be obtained by the partial hydrolysis of cellulose and appear as single peaks on the chromatogram [191, but it is extremely difficult actually to isolate a suitable amount of these oligosaccharides for model experiments, especially in the case of higher-molecular-weight compounds. Consequently, chemical synthesis is the most promising method for obtaining suitable amounts of these pure oligosaccharides, when synthetic problems suchas selection of the protective groups of glucose and pglycosylation methods are solved. There appear to be a few papers on the synthesis of cello-oligosaccharides. Around 1930, Helferich et al. [20] and Freudenberg et al. [ 2 11 reported the synthesis of cellobiose. In I97 I , Hall et al. [22] reported erroneous synthesis of cellotriose, but Takeo et al. [231 succeeded in the first true synthesis of cellotriose. In 1980, Schmidt et al. 1241 developed a new glycosylation method called the imidate method, using trichloroimidoyl groups as a leaving group from the C,-anomeric position, and applied it to the first synthesisof cellotetraose 1251. Later, Takeo et al. [26] also reported cellotetraose synthesis, but their methods do not seem to apply for the synthesis of the higher-molecular-weight oligosaccharides. Thus, a new synthetic strategy completely different from their methods must be considered for the higher oligosaccharides (271.

A.

BasicSyntheticStrategy

In principle, the synthesis of cellulose seems tobe simple because only three synthetic problems need to be solved: regiospecific control and stereospecific control, and increasing the molecular weight (Fig. l ) .

629

Chemical Synthesis of Cellulose

Stereospecific control ( P -glucosidic bond)

n OH

U Regiospecific control (1,4-bond)

Regiospecific control means to control 1,4-bond formation, which could be done by the selection of 2,3,6-tri-O-substituted glucopyranose derivative as a starting material. stereospecific control means to control P-glucosidic bond formation, which could be accomplished by stereospecific P-glycosylation. Stereospecific control would be extremely difficult and a key problem. Thus, we may obtain useful information about stereospecific P-glycosylation from the synthesis of cello-oligosaccharides without having to deal with the molecular-weight problem. Two basic synthetic methods, i.e., linear and convergent synthetic methods, are conceivable for the cello-oligosaccharides and also for cellulose, as shown in Fig. 2 [28]. For these synthetic designs, the starting material, D-glucose, should have three kinds of protective groups, X, Y, and R, for the regulation of regiospecificity. Here, X and Y groups are “temporary” protective groups and the R group is a “persistent” protective group [29]. It is a prerequisite for the selection of these three protective groups that the Y and R groups or the X and R groups must not change upon the removal of the temporary X and Y groups, respectively. That is, it is most important that each of these temporary groups be removed independently without any influence on the other functional groups. In a linear synthetic route as shown in Fig. 2, after repeating two reactions, deprotection of the Y group and P-glycosylation, a series of cello-oligosaccharides whose degrees of polymerization (DPs) are 2, 3, 4, . . . could be theoretically obtained. On the other hand, in a convergent synthesis as shown in Fig. 2, after repeating a set of two reactions, deprotection and P-glycosylation, n times, we will theoretically obtain a cellooligosaccharide whose DP is 2 . The convergent synthesis is preferred for obtaining higherDP cello-oligosaccharides by minimum reaction steps [30]. Thus, the selection of the three kinds of protective groups and the P-glycosylation method are extremely important. The possible combinations of the protective groups are shown in Fig. 3 [31].

B.

Synthesis of Cello-Octaose by a Linear SyntheticMethod

Kawada et al. established the first synthetic route for cello-octaose acetate starting from allyl 2,3,6-tri-0-benzyl-4-(p-methoxylbenzyl)-~-~-glucopyranoside ( l )as shown in Fig. 4 [32].Here allyl, p-methoxybenzy (PMB), and benzyl (Bn) groups were selected as the X, Y, and R groups, respectively. ThePMBgroup of the starting compound (1) can be removed selectively by oxidative reaction conditions with cerium(1V) ammonium nitrate (CAN) in acetonitrile-H,O to afford glycosyl acceptor (2).The allyl group can be removed by the two reactions, migration of the double bond with Ko-‘Bu in DMSO and subsequent acid hydrolysis of the enol ether obtained with HCl in acetone to give a glycosyl donor

630

X

0

Nak:atsubo

631

Chemical Synthesis of Cellulose

OR CICHpCO-

J

1 -CHpCH=CHp (All)

-COCH,( Ac) R = -CO-C(CH& (Pi") -CH2C& (Bn)

[

FIGURE 3 Possible combination of X. Y, and R protective groups.

which is further converted to the activated compound, a-imidate (3). The glycosylation of the acceptor (2) with the donor (3) catalyzed by BF, etherate in CH,CIZ at -70°C gave the expectedcellobiosederivative (4) in 85% yield without any production of the a anomer, via a completely S,2 reaction mechanism as proposed by Schmidt et al. [33]. Repeating the set of two reactions, deprotection of the PMB group and subsequent P-glycosylation with donor (3), six times, finally affords a cello-octaose derivative (10) which is converted to an acetyl derivative after deprotection and subsequent acetylation. The cello-octaose acetate obtained was interestingly compared with the acetate from hydrolysis product of cellulose reported by Buchanan et al. 1341. Thus, the selection of both combinations of the protective groups, allyl, PMB, and Bn groups, and the P-glycosylation method (imidatemethod)werefound to be quite suitable for the linear synthesis of cello-oligosaccharides. However, it was revealed that selection of the protective groups is not suitable for the convergent synthesis, because of instability of the dimeric and tetrameric imidates: imidoylation of the donor derived from cellobiose (4) and cellotetraose (6) derivatives always gave a mixture consisting of a - and P-anomers which cannot be obtained as a pure compound, i.e., these imidates are very unstable, resulting in decomposition upon separation by silica gel TLC. Generally, an electron-donatingprotectivegroup, such as a Bn group,accelerates reactivity of the glycosyl donor, but an electron-withdrawing group such as the acyl group reduces it [ 3 5 ] .Consequently, it is expected that we must design suitable starting materials with both optimum reactivity and stability by a partial exchange of several Bn groups with acyl groups in compound (1). C.

Substituent Effects on P-Glycosylation and Selection of the Starting Material for the Convergent Synthesis

Takano et al. 1361 obtained surprising but very interesting experimental results, leading to success in the first chemical synthesis of cellulose as a consequence, on glycosylation with a-imidate (3), which was used as the most suitable glycosyl donor, giving only @-glycoside for the linear synthesis of cello-octaose derivative as described in Section 1I.B. As shown in Fig. 5, upon glycosylation with a-imidate (3), glycosyl acceptor (2) having benzyl groups gave the only expected P-glucoside in a high yield, but contrary to their expectation, glycosyl acceptor (11). having acetyl groups, gave a mixture consisting of a - (29%) (12) and P-glucoside ( 16%) in low yield. Then, Takano et al. systematically examined the effect of the protective groups of the glycosyl acceptor and donor on the glycosylation reaction in order to select the best protective group system. The results are summarized in Table 1 . It is clear from the results

632 Nakatsubo

633

Chemical Synthesis of Cellulose

”.

O , Bn

. s m 0 0 -4 OBn

OBn

OBn

OBn

no a-Glucoside

-0 HO,

--

-/

-

12 [a-Glucoside (29%)

-U


FIGURE 5


11

p-Glucoside (1 6%)

Substituent cffccts o n P-glycosylation.

that the 3-0-Bn group is indispensable for obtaining highly stereoselective P-glycosylation in high yield. The effect of the benzyl group introduced into the 6 - 0 position on the alp ratio seems to be small. These results agree well with those 011 the glycosylation ofNacetyl glucosamine derivatives with acetyl galactosyl bromide reported by Sinay 1371. Consequently, the positions in compound (1) i n which the protective groups can be replaced by acyl groups for the stabilization of a-imidate are limited only to the 2 - 0 and 6 - 0 positions for molecular design of the starting material for the convergent synthesis. Then, Nishimura et al. [ 38) considered the most suitable combination of the 2 - 0 and 6 - 0 protective groups of glycosyl acceptor and donor, as summarized in Table 2. Reactions 1-4 were conducted by the general glycosylation method. and reactions S-7 used a simple vacuum system (Fig. 6). A high-vacuum system is conveniently used

TABLE 1 Glycosylation of Glycosyl Acceptors with ct-hidate (3) Reaction

Yield (%)

Reaction

- Ac -Ac

-Bn

- Ac -Bn -Ac

-B11

-Bn

-0Bn

CCI,

2 2 I I

16.4 13.0 94.2 96.0

29.0 33.0

Nakatsubo

634

TABLE 2 Effects of 2 - 0 and 6 - 0 Substituents on P-Glycosylation

R2

Reaction no. R, 1

Donor R2 - AC -Ac -Piv -Piv -Piv -Piv -Bn -Bn

-Ac -Piv -Piv -Piv -Piv -Piv -Bn -Bn -Piv -Bn

2 3 4 S 6 7:' 8

Acceptor

- AC -AC -Ac -Piv

-Piv -Bn

0-glycoside yield (96)

-Ac -AC -AC -Piv -Piv -Piv -Bn -Bn

15 17

32 S1

85 60 ( ( ~ 2 0 ) 86 84

Stability of imidate

Q

8 0 Q Q 0 0 X

8,

"Reactionwascarried o u t bytheuse of asimplevacuumsystem. purified byboth TLC andcolumn purified only by column chromatography; X, purilicd only by crystallization. chromatography; 0,

for the synthesis of higher cello-oligosaccharides to escape the hydrolysis of the moisturesensitive imidate. A sing-opening polymerization, which must be carried out under completely anhydrous reaction conditions, is also conducted in the high-vacuum system as described in a later section. All of these reactions gave only P-glycosides, except for reaction 6. Thesedata indicate that the introduction of acryl (acetyl or pivaloyl) groups fairly stabilizes the unstable imidate to be purified by silica gel chromatography as expected, and that pivaloyl groups are more useful than acetyl groups for the P-glycosylation. Pivaloyl groups have a larger electron-donating effect than do acetyl groups [39], resulting in higher reactivity of the glycosyl acceptors and imidates. In addition, pivaloyl groups at the 0 - 2 position prevent the formation of by-products such as an orthoester because of the larger steric hindrance. OPiv PMB'

B 8 F o d OPiv 13

OBn 14

Thus, both allyl 3-O-benzyl-4-O-(p-methoxybenzyl)-2,6-di-O-pivaloyl-~-~-glucopyranoside (13) and allyl 3,6-di-0-benzyl-4-0-(p-methoxybenzyl)-2-O-pivaloyl-~-~-glucopyranoside (14) were found to be suitable starting materials for the convergent synthesis of cello-oligosaccharides because the imidates derived from these compounds have the appropriate stability upon preparation and give P-glucosides in high yields on the glycosylation. Compound (13) is more suitable for large-scale synthesis because it is more stable.

Chemical Synthesisof Cellulose

t

U

l

m 4

U

a a

a c u n a a

v1 N

0

a

S

U m

? G .-M S

W

v)

.-

635

636

Nakatsubo

D. Synthesis of Cello-Octaose by a Convergent Synthetic Method The convergent synthesis of cello-octaose from starting material (13) by Nishinwra et al. 1401 is shown in Fig. 7. The reactivity of the imidate decreases with an increase i n the degree of polymerization (DP). In fact, dimeric inlidate (17) was stable under the glycosylation conditions with monomeric inlidate using 0.15 eq. of BF,-etherate at -70°C to give P-glucoside with an 86% yield. Tetrameric imidate (21) was stable under the same reaction conditions as those of dimeric imidate (17). Consequently, the glycosylation of these higher imidates (17) and (21) needcd more drastic reaction conditions, such as increasing the amount of BF1-etherate used or raising the reaction temperature. However. under such drastic reaction conditions, the trouble was that sevcral side reactions, such as removal of the acid-sensitive PMB group, formation of a-fluoride [i.e., (IU)] by the attack of fluoride ion on the inlidate, and hydrolysis of the imidates, took place simultaneously. To depress these side reactions to a minimum, ( I ) a minimum amount of BF,-etherate should be used, ( 2 ) PMB groups should be replaced by other protective groups which are stablc under acidicconditions, and (3) anhydrousglycosylationconditions should be strictly achieved. These problems were solved by the selection of the stable acetyl group under acidic reaction conditions as a new Y group and by the use of high-vacuum apparatus, as shown in Fig. 6 , for the completion of the anhydrous reaction conditions. Here, selective deacetylations of (16) and (22), keeping pivaloyl groups, were carried out using diazabicyclo (5.4.0lundec-7-ene (DBU) in methanol at room temperature to afford glycosyl acceptors (20) and (24), respectively. Thus, glycosylation between dimeric glycosyl donor, imidate (18), and dimeric glycosyl acceptor (20) carricd o u t at - 15°C using 0.05 eq. of BF.>ethcrate gave the expected tetramer (22) with a 71% yield. Glycosylation between tetrameric donor (23) and acceptor (24) at 0°C using 0. 15 eq. of BF,-ethrate gave cello-octaose derivative (25) with a 95% yield. The cello-octaose derivative (25) was converted to cello-tetraose (27) by removing the protective groups. Thus, the first convergent synthesis of cello-octaose was established by reexamining the protective groups and by considering the reaction conditions.

E. Synthesis of Cello-Oligosaccharides up to Eicosaose Derivatives by a Stepwise Synthetic Method For the clongation of the carbohydrate chain with a minimum of reaction steps based on the same convergent synthetic method as that shown i n Fig. 7, cello-octaose derivatives (25) must beconvertedinto both glycosyl donor and acceptor. However, there wasa problem in the preparation of a-imidate from compound (25). That is. imidoylation reaction always gave P-inlidate as an oil in quantitative yield, which is a kinetically controlled product, but the product was an unstable compound. This compound decomposed within a few days at room temperature to afford the hydrolysis product. Thus. elongation of the carbohydrate chain from cello-octaose derivative (25) was carried out via a stcpwise synthetic routc as shown in Fig. 8 [41]. Here, tetrameric a-imidate (23) was used a s the sole glycosyl donor. Cello-octaosederivative (25) wasconvertedintothcglycosylacceptor (28). then glycosylated with tetrameric imidate (23), to afford the expected cello-dodecaose derivative (29) with a high yield. After repeating thc set of two reactions, removal of acetyl group and subsequentglycosylation with imidate (23) twice. cello-eicosaosederivative (33) was obtained.Compound (33) was convertedinto the acetyl derivativeafter the

Chemical Synthesis of Cellulose

637

638 Nakatsubo

Chemical Synthesis of Cellulose

639

removal of protective groups and subsequent acetylation. The ‘H-NMR spectrum of the synthesized acetyl cello-eicosaose derivative was found to be almost identical to that of authentic cellulose triacetate (CTA). Generally, in carbohydrate chemistry the term oligosaccharide refers to a substance having two to ten monosaccharide units 1171. According to this terminology, derivatives with degree of polymerization (DP) above 12 are not oligosaccharides, although they may not be macromolecules. Concerning this point, Kobayashi et al. regarded their synthesized product having a DP value of 22 as a polysaccharide, that is, “cellulose” [4]. In fact, at around this DP, several characteristics of cellulose such as crystal structure seem to appear [4]. Isogai et al. also regarded samples having DP values of 15 as low-molecular-weight cellulose [42]. Thus, the synthesized compounds (29)-(33) may be regarded as “cellulose,” more strictly as low-molecular-weight or medium-size cellulose. Here, the stepwise synthesis of cellulose was achieved for the first time. Interesting changes in several of the properties accompanying an increase in DP of a series of the synthesized medium-size molecular weight compounds were found. Peak patterns of both ‘H- and “C-NMR spectra were gradually simplified with an increase of DP, leading to peaks attributable only to the internal repeating units. Thespectrum of acetyl cello-eicosaose derivative was almost identical to that of CTA. Upon gel permeation chromatography (GPC) analysis, the plots of logarithm of molecular weight (log M> versus elution volume are clearly linear, as shown in Fig. 9. Segal reported measurement of the molecular weight of cellulose trinitrates by GPC, but the values obtained were fairly large in comparison with those obtained viscometrically [43]. In this investigation it was found that, at least with high-DPcello-oligosaccharidesor medium-size cellulose up to eicosaose, the molecular weight can be determined almost exactly using this calibration curve. According to Freudenberg [44], if the linkages in a polymer-homologous series are uniform, the plot of[M] n/n against ( n - I)/n yields a straight line. Here, [m n = molecular rotation and 11 = DP. Figure 10 shows a plot of molecular rotations versus DP. The Freudenberg theory was found to be applicable to a series of a-D-acetates of cello-oligosaccharides up to DP 7 [45]. For the synthesized cellulose derivatives,the theory appears to fit up to DP 16. The point for DP 20 deviates from this straight line in spite of uniformity of the glycosidic linkages. It may be assumed that a conformational change due to a polymer effect starts to appear at around cello-eicosaose. Kajiwara et al. proposed an extended zigzag structure with an inflection point at DP 20, by the Monte Carlo simulation method based on the molecular mechanics [46]. For the proof of this possibility, synthesis of cellulose derivatives above eicosaose and their properties have to be studied.

111.

SYNTHESIS OF CELLULOSE BY A RING-OPENING POLYMERIZATION

A.

Synthetic Characteristicsand Problems of Ring-Opening Polymerization

Let us consider synthesizing cello-eicosaose derivative by the two typical synthetic methods, stepwise (linear and convergent syntheses) and ring-opening polymerization. Linear and convergent synthetic methods need 40 and 18 reaction steps for the synthesis of celloeicosaose derivative from XYR-glucose derivative as shown in Figs. 4, 7 , and 8, respectively, although the product is single-dispersed.

a a C

0:

1 ooooc

~o~ys~yrene

10000

2 0,

0 -

1000

1 0 0 " " " " " "

13

14

15

16

17

Elution time (rnin)

18

19

100 13

14

15

16

17

18

19

20

Elution time (rnin)

FIGURE 9 GPC analyses of synthesized 2,6-di-O-pivaloyI-3-O-benzyl cellulose and cellulose acetate series. Column, Shodex GPC KF-802 + KF-803: solvent. THF ( 1 mL/rnin).

641

Chemical Synthesis of Cellulose

c c

\

-5000

-10000

-1

/4

I

I

I

I

I

0.5

0.6

0.7

0.8

0.9

1

(n-1 )/n

FIGURE 10 Relation between degree of polymerization and molecular rotation: n, degree of polymerization; IM] ! l . molecular rotation; 0.2,h-di-O-pivaloyl-3-O-benzyl cellulose series; (Y-Dacctatc of cello-oligosaccharides series rcported by Wolforrn and Dacons 1451.

*,

A*,, 36: (1-4)-a-P

37: (1+4)-p-P

On the other hand, a ring-opening polymerization requires only one reaction for the syntheses of cello-eicosaose and also polysaccharides from the starting monomer, although usually it gives a polydispersed polymer. Thus, ring-opening polymerization is very useful for the preparation of high-molecular-weight polysaccharides. This method, however, has to overcome two extremely difficult problems, regiospecificity and stereospecificity, at the same time during the polymcrization. That is, ring-opening polymerizations of 1,4-anhydro glucose (34) and glucose 1,2,4-orthoester (35) monomers usually yield four possible structural units, that is, 1,4-anhydro glucose monomer (34) usually gives the (1+4)-a (36) and (1+4)-p (37) -D-glucopyranosidic units and the ( I + S ) - a (38) and ( I "+S)-p (39) -Dglucofuranosidic units. And glucopyranose l ,2,4-orthoestermonomer (35) gives the

Nakatsubo

642

(1-+4)-a (36) and (1+4)-P (37)-~-glucopyranosidicunits, and (1 -+2)-a (40) and (1+2)P (41)-~-glucopyranosidicunits. That is, for synthesis of cellulose, glucan consisting of only (1-4)-P-~-glucopyranosidic unit in four possible units has to be made. Furthermore, the stereo- and regioselectivity must be extremely high during the polymerization. For example, 99% stereoselectivity yielding 99% P- and l% a-glucosides upon glycosylation is a satisfactory result in the stepwise synthesis, because the undesired a-anomer can be separated by chromatography after glycosylation. However, 99% stereoselectivity is not acceptable for the ring-opening polymerization. This is because the eicosaose derivative obtained by polymerization with 99% selectivity is theoretically a mixture consisting of 524,288 (= 2'2')"') molecular species containing 82.6% (= 0.99(20"))of a stereoregular cello-eicosaose derivative. This is why no one so far has succeeded in the synthesis of cellulose by ring-opening polymerization.

B.

Cationic Ring-Opening Polymerization of 1,4-AnhydroglucopyranoseDerivative

The synthetic approach to cellulose via ring-opening polymerization of 1,4-anhydro-2,3,6tri-0-benzyl-a-D-glucopyranosewas reported for the first time by Micheel et al. [l21 to yield a cellulose-like polymer. Uryu etal. [ 131 also tried polymerization of the same monomer, but obtained an unexpected stereoregular ( l +5)-a-D-glUCOfUranan, with stereochemistry expected on the basis of the antiperiplanar theory of Deslongchamps [47]. Since I ,4-anhydro-a-~-glucopyranose, which may also be regarded as l ,5-anhydro-P-D-glucofuranose,has two ring-opening modes, that is, 1,4- or 1,5-ring scission, there are four possible structural units in the polymer obtained, which are caused by the ring-opening modes and anomeric a- and @-configurationsas described in Section 1II.A. Ring-opening polymerization is affected by reaction conditions, and there is a possibility of achieving the chemical synthesis of cellulose by finding optimum reaction conditions. In Chapter 2, the substituent effects on the highly stereoselective glycosylation in the syntheses of cello-oligosaccharides are described. The benzyl group at 3-0 was indispensable to obtain P-linked glucosides in high yield, and the pivaloyl group introduced into the 2 - 0 led to P-glycosidic linkage by the P-side attack of glycosyl acceptor because of the neighboring-group participation. Thus, there is a good possibility of synthesizing the expected P-(1-+4)-glucan with both stereo- and regioselectivity by ring-opening polymerization utilizing such substituent effects. In fact, Ichikawa et al. [48] and Kobayashi et al. [49] reported the syntheses of (1+6)-~-~-galacto-oligosaccharidesby applying the neighboring-group participation of the 2-0-acyl group. However, there are no papers describing such substituent effects in ring-opening polymerization of 1,4-anhydro-a-~-glucopyranose derivatives. In this section, substituent effects on the ring-opening polymerization of 1,4-anhydroglucose derivatives toward the chemical synthesis of cellulose are described. In orderto study substituent effects on ring-opening polymerization of 1,Canhydroglucopyranose derivatives, Kamitakahara et al. [50] selected four starting monomers, 1,4-anhydro-2,3-di-O-benzyl-6-O-pivaloyl-a-~-glucopyranose (42), 1,4-anhydro-3-O-benzyl-2,6-di-O-pivaloyl-a-~-glucopyranose (43). 1,4-anhydro-3-O-benzyI-2,6-di-O-pivaloyla!-D-glUCOpyranOSe (44), and 1,4-anhydro-6-O-benzyl-2,3-di-O-pivaloyl-a-~-glucopyranose ( 4 9 , and polymerized them under several reaction conditions. The results are summarized in Table 3.

Chemical Synthesis of Cellulose

643

TABLE 3 Polymerization of 1,4-Anhydro-a-~-Glucopyranose Derivatives“ Yield Time Temp (“C)

Exp. Monomer Initiator no. 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18

19 20 21 22 23 24 25

42

30 30 - 30 0 20 20 20 -30 - 30 - 30 20 20 20 - 30 - 30 - 30 20 20 20 - 30 - 30 - 30 20 20 20 -

43

44

(%)

25 21 20 20 17 20 16 34 60 20

65h 66h 80h 82h 1Ih 5~6~ 62h 87h 78’ 100’ 74h 86h ca. 100 Trace 42’ 100 Trace 100 100 Trace 7d 7* 35* ca. 100’ 7*

15

20 1

239 136 20 240 18 1.5

26 17 17 24 16 16 ~

[a], (deg)

(h)

~

10-’MGKe

DP,,

+79.7 +72.9 +43.4 +88.9 +68. I +58.1 +27.2 -66.9 -69.3 -65.6 -5.8 - 19.0 -23.4

10.4 8.0 I .9 9.3 2.7 5.1 3.1 9.4 7.5 3.4 1.4 4.7 2.2

24.3 18.9 4.3 21.8 6.4 12.1 7.3 22.0‘ 17.6‘ 8.1 ‘ 3.3 11.0 5.2‘

-59.3 -57.9

10.0 6.3

23.4 ‘ 14.9’

15.7 -48.0

3.2 4.6

1 1 .of

I .4 2.0 1.9 2.5 1.6

3.3 4.8 4.4 6.2 3.7

-

-26.0

7.6

~-

“Initiator concentration, 5 mol%; solvent, CH,Cl,; monomer/solvent, 50 g/lOO mL. *Polymer was insoluble fractlon in n-hexane. ‘No unreacted monomer was detected. ‘Polymer was separated from unreacted monomer by TLC. ‘Number-averaged molecular weight was determined by GPC using polystyrene standards. ‘Stereoregular (1-1S)-P-o-glucofuranan derivative was given.

1. Substituent Effect on Molecular Weight of Polysaccharides Molecular weights ofpoly(42)s obtained by the polymerization of monomer (42), poly(43)s from (43), and poly(44)s from (44) decreased with an increase of reaction temperature, but those of poly(45)s from (45) were low under all reaction conditions tried. The polymerizability of four monomers is in the order (42) G (43) = (44) >> (45), as judged from the highest molecular weight obtained from these monomers (Table 3, experiments 1, 8, 15, and 24) and from the yield of polymer at -30°C.

Nakatsubo

644

Generally, the electron-donating benzyl group accelerates reactivity. but the electronwithdrawing acetyl group retards both the glycosylations and the polymerizations of anhydro-sugars. For example, Zachoval et al. [51 ] reported that I ,6-anhydro-P-D-glucopyranose triacetate was less reactive than 1,6-anhydro-P-1~-glucopyranose tribenzylether. Monomers (42) and (43), having two benzyl groups. expectedly afforded polymer with high molecular weight. Monomer(45), having one benzyl group, did not afford polymer under various reaction conditions (Table 3, experiments 20-25), but unexpectedly, monomer (44), having only one benzyl group, afforded polymer with high molecular weight. Consequently, the present results indicate that the important factors for accelerating polymerizability are not only the number of the substituent benzyl groups but also the positions where benzyl groups are attached. The comnlon substituent among monomers (42), (43), and (44) is a benzyl group at the 3 - 0 position: monomer (45) does not have a benzyl group at the 3 - 0 position. Thus, the benzyl group at the 3 - 0 position is indispensable for yielding glucan with high molecular weight. Comparing monomers (43) having a benzyl group at the 6-0 position and (44) having a pivaloyl group at the 6-0 position, there is little difference in molecular weight under optimum reaction conditions (Table 3, experiments 8 and 15). Consequently, the benzyl group at the 6 - 0 position did not have much effect on polymerizability. 2.

Substituent Effecton Stereoregularity of Polysaccharides

Substituent groups at 0 - 2 greatly affect specific rotation. as shown in Table 3. All poly(42)s are dextrorotatory. The polymerization of (42), having a benzyl group at the 2 - 0 position, gave a nonstereoregular polymer consisting mainly of ( 1 jS)-cY-glucofuranosidic units. On the other hand. all poly(43)s, poly(44)s, and poly(45)s are levorotatory. All these monomers have a pivaloyl group at the 2 - 0 position. The polymerization of (43) gave a stereoregular ( I ~ 5 ) - P - ~ - g l u c o f u r a n derivative an (Table 3, experiments 8- 10 and 13). Polymerization of (44) also gave a stereoregular ( 1 +5)-P-~-glucofurananderivative (Table 3, experiments IS, 16, and 19). Polymerization of (45) tended to give (l+S)-P-furanosidic units, but none of the conditions afforded a stereoregular poIy(45), although monomer (45) has the same 2-0-pivaloyl group as monomers(43) and (44) (Table 3, experiments20-25).Consequently, it turned out that the benzyl group at the 3-0 position is indispensable for obtaining a stereoregular polysaccharide. It is predicted that the favorable complexation of Lewis acids with both a C , and a C , oxygen tends to result in enhancement of polymerizability and stereoregularity, because the electron-donating 3-0-benzyl group raises the electron density of the C, oxygen and consequently elevates the coordination power with Lewis acids. The electron-withdrawing pivaloyl group at the 3-0 position, on the contrary, weakens the coordination power of the Lewis acids so that polymerizability and stereoregularity are lowered. The fact that both polymerization of (43), having a benzyl group at the 6 - 0 position, and that of (44), having pivaloyl group at the 6 - 0 position, gave stereoregular (I-+5)-PD-glucofuranan derivatives with almost the same under these optimum conditions (Table 3, experiments 8 and IS) indicates that the substituent group at the 6 - 0 position hardly affects either stereoregularity or polymerizability.

3. Importance of 3-0-Benzyl Group of 1,4-Anhydro-cw-~-Glucopyranose Derivative on Ring-Opening Polymerization Substituent groups at the 6-0 position did not remarkably affect stereoregularity or polymerizability. comparing results from monomers (43) and (44). It was confirmed that the

Chemical Synthesis of Cellulose

645

benzyl group at the 3 - 0 position has a special function for yielding ;I stereoregular polysaccharide with high molecular weight in a ring-opening polymerization [results from monomers (43) and (45)). It was reconfirmed that the presence of the pivaloyl group at the 2 - 0 position makes the polysaccharide take the P-configuration [results from monomers (42) and (44)]. Polysaccharides with high molecular weight tend to have high stereoregularity as shown in Table 3 (experiments 8, 9, 15, and 16). Consequently. both the pivaloyl group at the 2 - 0 position and the benzyl group at the 3-0 position are indispensable for yielding stereoregular ( 1 +5)-/3-~-glucofuranan derivatives with high molecular weight. Furthermore, polymerization of 1,4-anhydro-a-~-glucopyranose was found to always preferentially afford ( 1 +5)-~-glucofuranose units, not ( 1 ~4)-P-D-giucopyranoseunits. These results agreed with the cases of the ring-opening polymerization of 2,7-dioxabicyclo-[2.2.1]heptane [52] and that of 1,4-anhydro-2,3,6-tri-O-benzyl-a-~-glucopyranose [ 131. Thus, it is also concluded that the 1,4-anhydro-a-~-glucopyranoseskeleton is not suitable for yielding a ( 1 +4)-P-~-gIucopyranan, that is, a cellulose molecule. It was revealed from the results of the polyrnerizations of 1,4-anhydroglucose derivatives that at least one of the most important problems i n the synthesis of cellulose, i.e., stereospecificity giving a stereoregular P-glucosidic polymer, could be successfully solved by applying substituent effects of 2-O-pivaloyl and 3-O-benzyl groups obtained from synthetic studies of cellooligosaccharides.

C.

Cationic Ring-Opening Polymerization of Glucopyranose 1,2,4-Orthoester Derivatives

1.

First Chemical Synthesis of Cellulose by a Cationic Ring-Opening Polymerization of Glucose 1,2,4-Orthopivalate Derivative

As described in Section III.B, polymerization of l ,4-anhydro-a-~-glucopyranose (43) was found always to preferentially afford ( 1+5)-~-glucofuranose units [poIy(43), not ( l +4)P-D-glucopyranose units poIy(47)]. One strategy for realizing the highly regioselective 1,4-scission is to substitute a l ,4-ether bond of 1,4-anhydro-a-~-glucopyranose derivatives (43) for another, more reactive linkage such as that of orthoester derivative (47), as shown in Fig. 11. Several cationic ring-opening polymerizations of such tricyclic intramolecular orthoesters prepared from arabinose and xylose have been studied extensively by Bochkov, Kochetkov, and their co-workers 1531. but they neither considered the substituent effect on polymerization nor achieved a stereoregular polymer. N,N'-carbonyldiimidazoleis usually used for preparing cyclic carbonate [54). In our initial synthetic plan, if such a 1,4-cyclic carbonate derivative was obtained from compound (46) by treatment with N,N'-carbonyldiimidazole, the carbonate might also be used as a starting monomer for the polymerization, as expected in the polymerization of Ncarboxy-a-amino acid anhydride resulting in poly-a-amino acid [ S ] .Thus, compound (47) was treated with N,N'-carbonyldiimidazole and then the product was polymerized with triphenylcabenium tetrafluoroborate. Surprisingly, the polymerization product obtained as crystals was found to be a stereoregular (1+4)-P-~-glucopyranan [poly (47)] which was identified by 'H- and I3C-NMR spectra (Fig. 12). I t was an exciting and memorable day for us on the Saturday evening of July 25 in 1994. We had thought at that time that the starting monomer was a I,4-cyclic carbonate, but latcr the starting monomer was found to be an orthopivalate (47). Stereoregular poly(47) was converted into a cellulose triacetate (49) by deprotections and subsequent aectylation, and identified by 'H- and "C-NMR spectra as shown in Fig.

Nakatsubo

646

43 OPiv 46

t \

OBn BnO /

OR1 poly(43): R1 = Piv, R2 = R3 = Bn poly(43)': R1 = R2 = R3 = H

\

C

X"",, 47

poly(47): R, = Piv, R2 = R3 = Bn

48:R1 = Ac, R2 = R3 = Bn 49: R1 = R2 = R3 =Ac 50: R1 = R2 = R3

H

FIGURE 11 Synthetic route for cellulose by a cationic ring-opening polymerization. "p-TsOH/ benzene/reflux/55%,"PF,/toluene/-3O0C, 'N,N'-carbonyldiimidazole/benzene/reflux, 62.8%. "Ph,CBFJ CH2Cl,/r.t.

100

95

90

85

80

75

70

6 (PPW FIGURE 12 2D-NMR spectra of poIy(47) from C-H COSY experiment (CDCI? as solvent).

647

Chemical Synthesisof Cellulose C2-OAc

.

t

.-,OAc

I

.oO *- AcO OAC

Authentic CTA

SyntheticCTA

7.0

8.09.0

6.0

t

1

5.0

t

~ - t

4.0

3.0

2.0

1.0

80

60

2040

0.0

6 (ppm)

120 140 l60 180

100

6 (PPW FIGURE 13 solvent).

'H- and I3C-NMR spectra of authentic and synthetic cellulose triacetate (CDCI, as

13. Finally, the cellulose triacetate (49) thus obtained was converted into cellulose (50) by deacetylation with NaOCH, in THF/methanol. The IR spectrum and X-ray diagram (Fig. 14) of cellulose prepared in this way were completely identical with those of regenerated cellulose with the cellulose-I1 crystal structure. Thus, the first chemical synthesis of cellulose by a cationic ring-opening polymerization had been achieved [ 151.

Nakatsubo

10

20

30

40

28 (deg.) FIGURE 14 X-ray diffractograms of (A) Whatman cellulose CFI 1. (B) regenerated celluhe. and (C) synthetic cellulose.

Substituent Effects at the 3-0- and 6-0-Positions on Stereo- and Regioregularity of Polysaccharides For examining the substituent effects at the 3-0- and 6-0-positions on the cationic ringopening polymerization of glucose 1,2,4-orthopivalate, Kamitakahara et al. [56] selected three additional orthopivalates, (Sl), (52), (53), and polymerized them under several reaction conditions. The results are summarized in Table 4.

2.

The molecular weights of poly(47). poly(51), and poly(52) were almost equal, but those of poly(53) were low under all reaction conditions tried. The polymerizability of four Inonomers is in the order (47) (51) >> (53), as judged from the highest molecular weight and yield obtained from these monomers (Table 4, experiments 3, S , 8, and 12). It can be said that polymerizability of (53) could notbe compared exactly with that of (47), (Sl), and (52) because polymerization of (53) was conducted at lower monomer concentration than that of (47), (Sl), and (52), but the above-mentioned trend must be true and substantial. In cases of ring-opening polymerizations of I ,4-anhydro-c~-~-glucopyranose derivatives, the benzyl group at the 3-0 position is indispensable for yielding glucan with high molecular weight. Similarly, in the present cases of cY-D-glUcOpyranOSe 1,2,4-orthopivalate derivatives, the benzyl group at the 3-0 position is indispensable for realizing stereoregularity, butis dispensable for yielding glucan with high molecular weight: that is, monomer (52), having a pivaloyl group at the 3-0 position, polymerized well but without yielding stereoregularity. Monomers having one benzyl group and two pivaloyl groups (including an orthopivaloyl group), that is, monomers (51) and (52), had the same polymerizability as monomer (47), having two benzyl groups and one pivaloyl group. Monomer (53), however, had

Chemical Synthesis of Cellulose

649

TABLE 4 Polymerizations of a-D-Glucopyranose 1,2,4-Orthopivalate Derivatives" Exp. no. 1

2 3

[a]::, (deg)

Temp. ("C)

Time (h)

Yield

Ph,CBF, Ph,CBF, Ph,CBF,

-30 20

16 18 14

59 96 93

-30 0 20

96 22 2

21

- 1.4

51 60

-3.7 + 1.9 7.5

4. I 4.9 3.2

9.7 11.6

33 30 0 20

96 49 15

-26.8 - 12.6 -18.8

2.9 3.7 3.0

6.9

47 58 Trace Trace 1.4

-24.8

Monomer

Initiator

47 47 47 S1

0

4 5 6

51 S1

Ph7CBF4 Ph,CBF, Ph7CBF,

7 8 9

52 52 52

Ph,CBF, Ph,CBF, Ph,CBF,

-

10" I l" 12''

53 S3 S3

Ph,CBF., Ph,CBF, Ph,CBF,

-30 0 1720

97 96 17 3.5

(76)

6.9 -20.1 -32.9 -35.2

DP,,

10

2.9 3.8 4.5

8.9 10.5

x.x 7. I

"Initiator concentration, S mol%; solvent. CH2CI,: monomer/solv.. 50 g/IOO mL "Monomer/solv.. 25 gll00 mL.

remarkably low polymerizability among the four monomers. Consequently, at least one benzyl group is indispensable for yielding a polymer with high molecular weight. Polymerizations of (47) and (51), having 3-0-benzylgroups, gave stereoregular ( I +4)-p-D-glUCOpyranan derivatives (Table 4, experiments 1-6). However. polymerizations of (52) did afford a stereo-irregular polysaccharide consisting of ( 1 +2)-a-P, ( 1 +2)p-P, and ( 1 -94)-p-P units. The (l+2)-bond formation increased with an decrease in temperature. That is, while the probability of coordination of Ph,CBF., with a C., oxygen was equal to that with a C., oxygen at 0°C and at 20°C, the coordination with the C? oxygen took place in preference to that with the C, oxygen at -30°C: the product ratios were ( I +2)-a-P/( 1 +2)-p-P/( 1+4)-p-P = 1 : I :2 [Table 4, experiment 8 (0°C) and 9 (20"C)I and (l+2)-a-P/( 1+2)-p-P/( 1 +4)-p-P = 3:3:4 [Table 4, experiment 7 (-30"C)l. The products ratio was calculated from the peak areas of the anomeric carbons of the corresponding constitutional units. The results indicate that in the case of coordination of a catalyst with the C, oxygen, the probability of an a-side attack resulting in the formation of (l+2)-aP was equal to that of a @-side attack resulting in formation of (1+2)-p-P. After all, the benzyl group at the 3 - 0 position is indispensable for yielding a stereoregular cellulose derivative, i.e., a ( 1 -4)-P-~-glucopyranan derivative. This fact agreed with the case for yielding stereoregular ( 1 +5)-/3-~-glucofuranan derivatives from 1,4anhydro-a-glucopyranose derivatives.

3.

Substituent Effects of Orthoester Groups on the Cationic Ring-Opening Polymerization of Glucopyranose 1,2,4-Orthoesters Generally, an electron-donating substituent increases the reactivities of both glycosyl donors and acceptors, resulting in highly stereoselective glycosylation with high yield, but an electron-withdrawing substituent exhibits the oppositeeffect.This means that ether groups are superior to acyl groups. Thus, the pivaloyl groups with relatively small electronwithdrawing effects in acyl groups are highly effective as protective groups of sugar hy-

Nakatsubo

650

droxyl groups on the glycosylation. The electron-withdrawing abilities of acyl groups are associated with the pK,, values of the corresponding carboxylic acids. OBn

47: R=-C(CH& 54: R=-CHPCH~ 55: R=-CH3 56: R=-CBHS R

Hori et al. [57] selected three additional orthoesters, 1,2,4-orthopropionate (54), orthoacetate ( S ) , and orthobenzoate (56), as starting monomers for cationic ring-opening polymerization and prepared from propionic (pK, 4.88), acetic (pK, 4.76), and benzoic (pK,, 4.20) acids to investigate the electronic effects of the orthoester groups on the cationic ring-opening polymerization. The results were compared with that of orthopivalate (47) (pK,, of pivalic acid, 5.05), whose polymerization gave a completely stereoregular cellulose derivative as described previously. Polymerizations of 3,6-di-O-benzyl-a-~-glucopyranose 1,2,4-orthoester derivatives (54)-(56), were camed out under the same reaction conditions as those giving cellulose derivative with BE 19.3 in the polymerization of (47). That is, all polymerizations were carried out at 20°C in the presence of triphenylcarbenium tetrafluoroborate as an initiator, in 5 mol% of initiator concentration and in 100 g/100 mL of monomer concentration. The results are summarized in Table 5. Specific rotations of polys(54)-(56) were positive values, 1.56", 12.9", and 11.8", respectively, as compared with poly(47), a cellulose derivative which has a large negative specific rotation, -37.2". These specific rotation values suggest that not all polymers newly obtained from monomers (54)-(56) are stereoregular. In fact, all these polymers were found to be nonregioregular, consisting of ( 1 -2)-and ( 1 +4)-P units, although with P-glycosidic linkage, that is, with stereospecificity but without regiospecificity. These data indicate that there is no distinct relationship between the production of (1-+4)-p-Punit and the pK,, values of the carboxylic acid groups introduced at the 2-0positions of the monomers. Only the 2-0-pivaloyl group has a characteristic effect on the fate of ring-opening polymerization, resulting in the formation of stereoregular (1+4)-p-

+

+

+

TABLE 5 Polymerization of 3,6-O-dibenzyl-ar-~-Glucopyranose 1,2,4-OrthoesterDerivatives"

~~

47 54 55 56

4.88 4.20

2 6 I .5 1.5

62 56 72 94

-37.2 +67 IS6 + 77 12.9 81 +11.8

19.3 15.8 10.4 9.4

0 33 23 19

100

"Initiator, Ph,CBF,; initiator concentration, S mol%; solvent, CH,CI,; monomer/solvent, 100 g / l O mL; reaction

temperature, 20°C. hpK,,values are those of the carboxylic acid groups introduced into 2-0-pos~tionsof monomers. 'Polymer is insoluble fraction in chloroformh-hexane (ca. 1/S. v/v). 'Calculated from polystyrene standard. "Calculated from the anomeric peak ratios in "C-NMR spectra of polymers.

thesis

Chemical

of Cellulose

651

glucopyranan. 2-0-acyl groups affect highly stereoselective P-glucosidic bond formation, and, in addition, the pivaloyl group in the acyl groups also affects further highly regioselective, (1+4)-glycosidic bond formation, probably because of steric effects, not electronic effects. Orthoester derivatives (52) having a 3-0-pivaloyl group gave polymerized products consisting of almost the same amount of (1+2)-a- and (1+2)-P-P units in the case of (1+2)-glycosidic bond formation (Section III.C.2). However, the present polymerizations of orthoesters (54)-(56) having a 3-0-benzyl group gave only P-glucosidic linkages upon (1+2)-bond formation. Consequently, the benzyl group at the 3-0-position has a great electronic effect upon the highly stereoselective P-glucosidic bond formation in the polymerization of a-D-glucopyranose l ,2,4-orthoester derivatives. Thus, both the 3-0-benzyl group and the orthopivaloyl group are indispensable substituents for the synthesis of stereoregular ( 1+4)-P-~-glucopyrananderivatives, cellulose derivatives, in the ring-opening polymerization of a-D-glucopyranose 1,2,4-orthoester derivatives.

IV.

FUTUREPROSPECTS

In spite of all efforts at chemical synthesis of cellulose for about55 years since Schulubach first started synthetic study, no one reached the difficult goal. Recently, the author and his co-workers succeeded in the first chemical synthesis of cellulose. In this chapter, their story of search for synthesis for a period of 15 years was described briefly. In this story, extremely important findings are on the substituent effects on P-glycosylation. That is, both 2-0-pivaloyl and 3-0-benzyl groups were found to be indispensable for obtaining Pglycoside with high stereoselectivity, from the results obtained on both synthetic methods, i.e., stepwise synthesisgiving cello-eicosaose derivative (DP20) and cationic ring-opening polymerization giving cellulose derivative with DP about 20 (DP 45 in the recent experimental results). Furthermore, interesting results were obtained that the 2-0-pivaloyl group also participates in the highly regioselective 1,4-bond formation in ring-opening polymerization. However, the way that both stereo- and regioselectivity can be overcome by the effects of both 2-0-pivaloyl and 3-0-benzyl groups are still unknown, especially highly regioselective control by the 2-0-pivaloyl group on ring-opening polymerization of the orthoester derivatives, although the neighboring participation of the 2-0-pivaloyl group giving P-glycoside is well known. Furthermore, additional problems which cannot be understood are that the present ring-opening polymerization method developed for the synthesis of cellulose cannot be applied for the synthesis of (1 +4)-P-xylopyranan. The ring-opening polymerization of 3-0-benzyl-xylopyranose 1,2,4-orthoester always gives a polymer consisting of both (1+2) and (1+4)-P-xylopyranose units [58]. We expect that research and the solution of the aboveproblems in synthesis may lead to the logical molecular design for the synthesis of other natural polysaccharides. We have succeeded in the syntheses of stereoregular galactofuranan and arabinofuranan [59]. On the other hand, one of the future aims in cellulose application is development of functional derivatives with special functions. For this, preparation of highly regioselectively substituted cellulose derivatives is a key. At present, 6-0-substituted cellulose derivatives can be synthesized, but the cellulose derivatives substituted at only the 2-0- and 3-0-positions cannot be obtained. Cellulose derivatives (44) in Fig. 8 and poly(47) and poly(51) may be theoretically used as starting materials for the preparation of regioselec-

Nakatsubo

652

tively 2-0- and 3-0-substituted cellulose, because the 2-0- and 3-0-positions of these derivatives are protected with ester and ether groups, respectively, which can be deprotected independently. At present, we have succeeded in the synthesis of all kinds of pos1591. sibleregioselectivelymethylatedcellulosederivativesfrompoly(47)andpoly(51) The studies on the relationship between structure and function of these regioselectively substituted cellulose derivatives may enable us to design new functional cellulose derivatives in the future. Thus, the present chemical synthetic methods for cellulose are expected to apply to the solution of basic problems of cellulose chemistry and technology.

REFERENCES I.

2.

3. 4. S.

6.

7.

8.

9.

10. 11.

12.

13. 14.

IS.

See, for example, T. P. Nevell and S. H. Zeronian, in Crllulosc~Chenris/r;yemel I t s Applicertion.s (T. P. Nevell and S. H. Zeronian, eds.), Ellis Horwood-Wiley, NcwYork, p. IS (1985). See books on cellulose: (a) R. H. Atalla, in ACS S y n p Sex 340 (R. H. Atalla, ed.), American Chemical Society, Washington, DC, p. I (1987); (b) R. P. Millane et al.. In Cel/ulose ( r n d Woot/"Chet?listry c n ? d Tec.hr~ology,Proc. Tenth Cellulo.sc~ Cor$ (C. Schuerch. ed.). Wiley. New York, p. 39 ( 1989). H.M. Schlubach and L. Luhrs, Ann, 547: 73 (1941). S. Kobayashi. K. Kashiwa, T. Kawasaki, and S. Shoda. J. An?. Chetn. Soc., 113: 3079 (199 l ) . (a) P. Sixou ct al., in Cellulose-Str~~ctl~re, Modification crnd Hydro1y.si.s (R. A. Young and R. M. Rowell, eds.), Wiley, NewYork,p.203 (1986); (b) M. Siekmeyer et al.. in CelluloseStructure c u d Functioned Aspecrs (J. F. Kennedy, G. 0. Phillips, and P.A. Williams, cds.). Ellis Horwood, NewYork,p.34.5 (1989). (a) G. Hesse and R. Hagel, Chrotncr/ogrrrphia, 6: 277 (1973): (b) T. Shibata, I. Okamito, and K. Ishii, J. Lie/. Chromatog., Y: 313 (1986); (c) Y. Okamoto, M. Kawashima, and K. Hatada, J. A n . Chetn. Sac.. 106: S357 (1984); (d) T. Shibata, T. Sei, H. Nishimura, and K. Deguchi. Chror~~c~togrc~phia~ 24: SS2 ( I 987). (a) K. Kamide, K. Okajima, T. Matsui, and H. Kobayashi, Po/ytnc,r J.. 15: 309 (1983); (b) K. c ~ n c lFunctional Aspects (J. F. Kennedy. G. 0. Phillips, and Okajima, in Cel/L~lose-Stnlcrllr-r),se-Sfr~~ct~~re P.A. Williams, cds.). Ellis Horwood. New York, p. 439 ( 1989). (a) K. Matsuzaki. I. Yamamoto,and T. Sato. Makrornol. Chern.. 186: 449 (1985); (b) 1. Yamamoto, 1. Takayama. K. Hornma, T. Gonda, K. Matsuzaki, K. Hatanaka, T. Uryu, 0. Yoshida. H. Nakashirna, N. Yamamoto, Y. Kaneko; and T. Mimura, Carhohydr: Polymers, 14: S3 ( 1991). General reviews on polysaccharide synthesis: (a) A. F. Bochkov and G. E. Zaikov, Chemistt? of the 0-Glycosiclic Bond-Fornution m e 1 CIlwwge (C. Schuerch. trans., cd.). Pergamon 43: 2389 (1987); (c) Press. Oxford, U.K., p. 130 (1979); (b) N. K. Kochetkov, 7ktrc~hc~clron. N.K. Kochetkov. in Studies in Nutuml Products Chernistty, Vol. 14. Elsevier. Amsterdam. p. 20 I ( 1994). E. Husemann and G. J. M. Muller, Makrornol. Ch~tn..01: 212 ( 1966). S. Hirano, Agric. Biol. Chetn.. 37: 187 (1973). (a) F. Micheel, 0.-E. Brodde, and K. Reinkmg. LiehigsAnnu. Chcw~,124 (1974); (b) F. Micheel and 0.-E. Brodde, Liebigs Annu. Chcwr., 702 (1974); ( c ) F. Micheel and 0.-E. Bordde, Liehigs Annu. Chern., I 107 ( 1975). T. Uryu. C. Yarnaguchi. K. Morikawa, K. Terui, T. Kanai. and K. Matsuzaki. Mercror~?oleclrl~~s. /H: 599 (1985). (a) T. Uryu, K. Kitano, K. Ito, J . Yamanouchi,and K. Matsuzaki, Mclc,nrtlo/c~crrlt.s, 14: I (1981); (b) T. Uryu, J. ydmanouchi, T. Kato. S. Higuchi. and K. Matsuzaki, J . An?. Chenr. Soc.. 105: 6865 (1983). F. Nakatsubo, H. Knmitakahra. and M. Hori, J. Am.Chrrn. Soc.. 1fH: 1677 (1996).

Chemical Synthesis of Cellulose

653

16. (a) Inside R&A, 25: 1 (1996): (b) E ~ I Chern. : News, 27 May-2 June: 23 (1996); (c) Y. Nishiyama. Kngclku t o Kogyo, 50: 626 ( l 997) (in Japanese); (d) K. Matsuoka, Trends Gfyosi. Glycotecllrlol., 9: 441 (1996). 17. (a) R. W. Binkley, in Modern Cctrl>ohydrctte C l ~ m i s t r yMarcel , Dekker. New York, Basel, p. I (1988): (b) S. H. Pine, Organic Chernisfn, McGraw-Hill Int. Ed., New York, p. 758 (1987). 18. (a) K. V. Sarkancn and C. H. Ludwig, Lignins, Wiley-lnterscience,New York, 1971: (b) T. Higuchi, T. K. Kirk, and S. Shimada, in Biosynthesis ctnd Biorlegrndrrtion pf Wood Cornponenrs (T. Higuchi ed.), Academic Press, New York, p. 557 (1985). 19. M. Murayama. B. Chun, J. Azuma, and K. Okamura, Bull. Kyoto Unir. Fore.st.s. 59: 310 (1987). 20. B. Helferich and F. Bredereck, Brrichfe, 64: 241 1 ( 1931 ). 21. K . Freudenbergand W. Nagai. Uerichte. 66: 27 (1933). 22. (a) D. M. Hall and T. E. Lawler. Curhohydr. Res., 16: I (1971); (b) D. M. Hall, T. E. Lawler, and B. C. Childress, Crtrl~ohyrlr.Ras., 38: 359 (1974). 23. K. Takeo, T. Ysato. and T. Kuga, Cctrhohydr. Res., 93: 148 ( I981 ). 24. R.R. Schmidt, Arzgert: Chenl. In?. Ed. Engl., 19: 731 (1980). 25. R. R. SchmidtandMichel, Angcw Chern. In?. Ed. EngI., 21: 72 (1982). 26. K. Takeo, K. Okushio,and K. Fukuyama, Cctrhohydr. Res., 121: 163 (1983). 27. F. Nakatsubo, T. Takano, T. Kawada, H. Someya, T. Harada. H. Shiraki,and K. Murakami, Mem. Coll. Agric. Kyoto Unir.. 127: 37 (1985). 28. L. Velluz, J. Valls, and J. Mathieu, Angerc: Chern. h r . Ed. Ens/., 6: 778 (1967). 29. (a) J. M. Frcchet and C. Schuerch, J. A m Chenl. Soc., 93: 492 (1971): (b) I . J. Goldstein and T. L. Hullar, Ad,: Crrrhol~yrlc Chem. Biochenl., Vol. 21, Academic Press, New York, London, p. 43 I (1966). 30. S. Turner, The Design of' Organic Syntheses, Elsevier. Amsterdam-Oxford-New York (1976). 3 1. T. W. Greeneand P. G. M.Wuts, Protective Groups irl Organic Synthesis. 2nd c d . , Wiley, New York, 1990. 32. (a) T. Kawada, F. Nakatsubo,and K. Murakami, Mokrczcti Gakkaishi, 35: 14, 35: 21 (1989): (b) T. Kawada, F. Nakatsubo, andK. Murakami. Cdlrrlose Ckern. Techno/.. 24: 343 (1990); (c) T. Kawada, F. Nakatsubo. K. Murakami,and T. Sakuno, Mokuzcti Gukknishi, 37: 930 ( 1991): (d) T. Kawada, F. Nakatsubo, T. Umezawa. K. Murakami, and T. Sakuno. Mokuzrti Gltkkaishi, 40: 738 ( 1 994). 33. Schmidt, R . R., Angew. Chern. lnt. Ed. EngI.. 25: 212 (1986). 34. C. M. Buchanan, A. Hyatt, S. S. Kelly. and J. L. Little, Mctcromolc~cu1e.s.23: 3747 (1990). 35. Paulsen, H., Angew. C h m . Int. Ed. Ens/.,21: 155 (1982). 36. (a) T. Takano, F. Nakatsubo, and K. Murakami, Cellufosc~Chern. Technol., 22: 13.5 (1988); (b) T. Takano, Y. Harada, F. Nakatsubo, and K. Murakami, Cellrtlosr Ckern. 7 k c ~ h n o l . . 24: 333 ( 1990). 37. P. Sinay, Prwe Appl. Chem., 50: 1437 ( 1978). 38. T. Nishimura. F. Nakatsubo. and K. Murakami, Mokuztri Gctkknishi, 3 9 40 (1993). 39. H. KunzandA. Harreus,Liehigs Annu. Chem., I : 41 (1982). 40. (a) T. Nishimura, F. Nakatsubo, and K. Murakami, Mokuzai Cnkkaishi, 40: 44 (1994); (b) T. Nishimura and F. Nakatsubo, C~trbohydr.Res.. 294: 53 (1996). 41. (a) T. Nishimura and F. Nakatsubo. Tefruhedron L e f t . , 37: 9215 (1996); (b) T. Nishirnura and F. Nakatsubo, Cellttlose. 4: 109 ( 1 997). 42. A. Isogai and M.Usuda, Mokuzui Gctkktrishi, 3 7 339 (1991). 43. L. Segal. J . Polymer Sci., 134: 101I (1966). 44. K. Freudenberg, Zutnin, Ceftrlo.c.e,Lignin, Springer, Berlin, p. 104 ( I 966). 45. H. Wolfrom and J. C. Dacons. J . Am. Chem. Soc., 74: 5331 (1952). 46. (a) M. Mirnura, H. Urakawa, K. Kajiwara, S. Kitamura, and K. Takeo, Mncrolnol. S y n p . , 99: 43 (199.5); (b) K. Kajiwara, M. Mimura. S. Kitatnura,and K. Takeo. Cell. Comnrcn., 3: 18 (1 996) (in Japanese). 47. P. Deslongchamps, C. Mareau, D. Frehel, and P. Atlani, Cm. J . Chetn., SO: 3402 (1972). 48. H. Ichikawa, K. Kobayashi.and H. Sumitorno, C~trhohytlt:Res., 179: 315 (1988).

654

Nakatsubo

49. K. Kobayashi, T. Ishii, M. Okada, and C. Schuerch, Polymer J., 25: 49 (1993). 50. (a) H. Kamitakahara, F. Nakatsubo, and K. Murakami, Mokuzui Gnkkaishi, 40: 302 (1994); (b) H. Kamitakahara, F. Nakatsubo, and K. Murakami, Macromolecules. 27: 5937 (1994); (c) H. Kamitakahara and F. Nakatsubo, Macromolecules, 29: 1 1 19 (1996). 51. J. Zachoval and C. Schuerch, J. Am. Chm. Soc., 91: 1165 (1969). 52. H. K. Hall,Jr., F. DeBlauwe, L. J. Carr, V. S. Rao, and G. S. J. Reddy, J. PolymerSci., Polymer Symp.. 56: 101 (1976). 53. (a) N. K. Kochetkov, A. F. Bochkov, and I. G. Yazlovertsky, k v . Akad. Nauk SSSR, Se,: Khinz., 1972 (1966); (b) N . K.Kochetkov, A. Ya.Khorlin, A. F. Bochkov, and I. G. Yazlovertsky, Curbohydr. Res., 2: 84 (1966); (c) A. F. Bochkov, I. G. Yazlovertsky, and N. K. Kochetkov, 17s. Akad. Nuuk SSSR, Ser. Khim., 1812 (1968); (d) N. K. Kochetkov, A. F. Bochkov, and 1. G. Yazlovertsky, IZIJ.Akad. Nauk SSSR, Se,: Khim., 1818 (1968); (e) N. K. Kochetkov, A. F. Bochkov,and I. G . Ydzlovertsky, Curbohydr. Res., 9: 49 (1969); (f) A. F. Bochkov, V. N. Chernetsky,and N. K. Kochetkov, lzv. Akad. Nuuk SSSR, Se,: Khinz., 465 (1975); (g) A. F. Bochkov, I. V. Obruchnikov, and N. K. Kochetkov, Zh. Obshch. Khim., 42: 2766 (1972); (h) A. F. Bochkov, I. V. Obruchnikov, and N. K. Kochetkov, Zh. Obshch. Khinz., 44: 1197 (1974); (i) A. F. Bochkov, V. N. Chernetsky, and N. K. Kochetkov, Curhohydr. Res., 43: 35 (1975); ti) A. F. Bochkov and A. V. Rodionov, Izv. Akacl. Nauk SSSR, Sex Khim., 2789 ( 1 976). 54. (a) J. P. Kutney and A. H. Ratcliffe, Synthefic Commun., 5: 47 (1975); (b) S.-K. Kang, J.-H. Jeon, K.-S. Nam, C.-H. Park, and H.-W. Lee, Synthetic Commun., 24: 305 (1994). 55. E. Katchalski and M. Sela, in Advances in Prorein Chernistq (C. B. Antinsen, Jr., M. L. Anson, K. Bailey, and J. T. Edsall, eds.), Academic Press, New York, p. 243 (1985). 56. H. Kamitakahara, M. Hori, and F. Nakatsubo, Muc,rornolecules, 29: 6126 (1996). 57. M. Hori, H. Kamitakahara, and F. Nakatsubo, Macromolecules, 30: 2891 (1997). 58. M. Hori and F. Nakatsubo, unpublished data. 59. F. Nakatsubo, Proc. 10th ISWPC, Yokohama, Japan, 1999, Volume l , p. 20.

16 wood Plasticization Nobuo Shiraishi Kyoto University, Kyoto,

1.

lapan

INTRODUCTION

Methods for processing wood are very limited. While metals, plastics, and glass can be processed in the liquid phase at high temperatures, natural wood cannot. This difference is due to the lack of plasticity of wood, so that it cannot be melted, dissolved, or softened sufficiently for molding. Consequently, the scope for utilizing wood is restricted, which sometimes makes wood less valuable as a material. If plastic properties could be imparted to wood, it would become a more useful material. This need, and the desirability of using wood waste and renewable forest product resources better, have resulted in extensive new studies on wood chemicals, modified natural polymers, new pulping methods, and reconstituted wood products. Inherent in this work is the plasticization of wood by simple chemical processing [ 1-71.

II. THERMOPLASTICITYOF WOOD The name “plastics” is given to the numerous macromolecular organic materials that can be softened, melted, and molded by heat and pressure or by mechanical means, such as mixing and calendering. Wood, which is also composed of macromolecular organic materials, differs from the plastics in such basic properties as plasticity, thermoplasticity, and dissolubility in organic or aqueous solvents. In other words, wood lacks plasticity, which means that it cannot be softened sufficiently for molding, melted, or dissolved. A summary of the previous literature on thermoplasticities of wood reveals that only the phenomena up to the second-order transition and/or thermal softening have been investigated, but not the phenomena of thermal flow [&lo]. Wood is composed of 50-55% cellulose, 15-25% hemicellulose, and 20-30% lignin, with small quantities of ash and extractives. The main components make up an interwoven network in the cell walls and middle lamella. The minor components are mostly in cell lumens or special tissues such as resin canals, and are directly or indirectly related to the physiology of trees. Thus, the components that are directly related to the fundamental properties of wood, such as thermoplasticity, are considered to be the main ones, and such properties are 655

Shiraishi

656

formed not as the simple summation of properties of individual components but as the integration of these properties as a result of mutual interactions by these components. This has been observed in previous studies on thermoplasticity of wood. Lignin and hemicellulose are amorphous polymers and are much more thermoplastic than cellulose, which is a highly crystalline polymer. Goring's data [S1 concerning the thermoplasticity of wood components are typical examples of measurements of this kind. They show that, in a dry state, while lignin and hemicellulose have thermal softening temperatures around 127-235°C and 167-2 17"C, respectively, for cellulose the thermal softening temperature is around 231 -253°C. The variation in the value of the softening temperature found in each wood component is caused chiefly by the difference in the method of isolating the components from wood. On the other hand, the thermoplasticity of lignin, hen~icellulose, and cellulose can be increased by wetting the samples. When the water content is increased to 20%, lignin and hemicellulose show softening points around 72- 128°C and 54- 142"C, respectively, much lower than for the dry state. By contrast, cellulose shows a decrease of only about 6-9°C in a wet state. These differences indicate how water can act a s a plasticizer and how the softening temperature depends on the two distinct states, crystalline and amorphous, of the components. On the other hand, an examination of thermal softening of wood as a whole reveals that no thermal softening can be identified similar to that of individual components. Wood does not show thermal softening until it is heated to a much higher temperature.This suggests that interaction among the main wood components plays an important role. Goring [8] found an insignificant softening at temperatures over 200"C, and Chow [9] reported that thermal softening of wood begins at 180°C and reaches the maximum rate of softening at 380°C. Back et al. proved that the second-order transition point of hardboard made from wood fiber is around 230-350°C. Baldwin and Goring [ 101 found that steamed wood has a lower softening point, around 200°C. This may be because steaming reduces the interactions among the main components of wood, causing amorphous portions of the main components to soften independently of the cellulose. This implies that the thermoplasticity of wood is governed by cellulose. The thermal behavior of lignin and hemicellulose is restricted by interactions due to secondary intermolecular bonding with cellulose. The difference in the effect of sorbed water on thermal softening of each of the main components implies that the crystallinity of cellulose contributes greatly to the degree of thermoplasticity of wood.

111.

REASONS FOR LOW THERMOPLASTICITYOF WOOD

As described above, although wood shows thermal softening, it occurs only at temperature over 200°C. Thermal fluidity is not observed for wood. Thus, wood is known as a nonplastic material. The fact that wood is insensitive to heat and shows no thermal fluidity and plasticity can be attributed to the following reasons:

1. 2. 3.

Cellulose is a crystalline polymer with about 50-70% crystallinity. Lignin has a three-dimensional molecular structure with very high molecular weight. Chemical bonds are formed even between main components of wood, such as in lignin-carbohydrate complexes (LCCs).

Cellulose is a linear high polymer having glucose residue as a repeating unit. Glucose has three hydroxyl groups, which means that cellulose has the ability to form significant

Wood Plasticization

657

hydrogen bonding. The resulting high intermolecularforces, in addition to the regular structure of the polymer, result in its high degree of crystallinity. The crystalline melting point of cellulose is considerably above its decomposition temperature. Thus, no melting of cellulose can occur at temperatures that do not cause pyrolysis. Therefore, cellulose is a material with low thermoplasticity. However, the above properties of cellulose can be altered by converting cellulose into derivatives, which enables it to become more plastic. For example, cellulose nitrate, cellulose acetate, benzylcellulose, and so forth, are known as cellulose plastics. I n this sense, it may be easily postulated that if cellulose in wood could be derivatized in situ, it would show thermoplasticity up to melting. The idea of converting wood as a whole into plastic material did not appear until recently. This is natural when one considers that lignin has a three-dimensional gel structure. Furthermore, lignin has been proved to exist as a stereoscopic spongelike aggregate within the cell walls of wood. Cellulose is interwoven as fibril bundles through the network of this aggregatestructure of lignin, and the spacesare filled with hemicellulose. The crystalline portions of cellulose fibrils can be regarded as rigid junctions among the linear polymer chains, and in this sense it is considered that cellulose aggregates also form a three-dimensional network. The cell wall can be considered as an interpenetrating polymer network (IPN) with partial chemical interactions between the main components of wood, such as those in LCCs. All these factors make wood a thermally insensitive material.

IV. THERMOPLASTICIZATIONOF WOOD The preceding section suggests that it maybe difficult to change wood intoplastics; however, the author has recently found that wood can be converted into a thermally flowable material by chemical modifications such asesterification,etherification, and some other dcrivatizations 11 -7,l I , 121. Chemical modification does not necessarily require special techniques, and therefore conventional and simplemethods can be used for this purpose. This phenomenon can be explained most simply in terms of the internal plasticization of wood. That is, the internal plasticization of wood through chemical modification can producea meaningful change in fundamentalpropertiesincluding thermoplasticity. The degree of change in the plastic properties of wood is dependent on the molecular size of the substituentgroupsintroduced, the degree of substitution, as well as the reaction method.

A.

Large-Substituent Modification

Accordingly, the introduction of large substituent groups into wood can result in a chemically modified wood with high thermoplasticity. Since the change in thermoplasticity is significant in this case, early studies dealt with thermoplasticization of wood reacted with large groups. Plasticization of wood was first observed after esterification with a series of higher fatty acids in a nonaqueous cellulose solvent medium (N,O,-dimethylformamidc) [ I , 121. In this case, acylation of more than one-third of the available hydroxyl groups in the wood is sufficient for the products to show thcrmofluidity. The thermofluidity was proved with a thermomechanical analyzer, a flow tester, and scanning electron microscope (SEM) observations. Figure 1 shows the SEM of heat-treated birchwood meal and that of lauroylatcd wood meal [ 121. While the untreated wood meal which was heat-treated at 270°C (Fig.

658

Shiraishi

FIGURE 1 (a) Untreated birch sawdust and (b) lauroylated sawdust with 93% ester content, both heat-treated at 270°C. Observed under scanning electron microscope (SEM).

la) shows normal wood tissues such as parenchyma and wood fibers, the corresponding lauroylated wood meal (Fig. lb) reveals clear melting with the disappearance ofwood tissue. In the latter case, about one-third of the hydroxyl groups in the wood have been lauroylated. It has also been proved that this thermal flow of the modified wood is not caused by chemical degradation during chemical modification or thermal deterioration of the sample during molding under heat and pressure [ 1l]. Films molded by heat and pressure were crushed and reexamined by a thermomechanical analyzer to show that their thermodiagrams were not different from those of the samples prior to hot-press molding. The thermal softening behavior of samples first acylated and then completely saponified were also found to be essentially the same as that of untreated wood meal. It was thus shown that the acylated wood, obtained underspecial reaction conditions by using a nonaqueous cellulose solvent as a reaction medium, showed thermofluidity. Subsequently, studies were continued to determine whether wood can be plasticized in the same manner by more general methods of higher-aliphatic-acid acylation. Preparations of cellulose esters of a series of higher aliphatic acids are generally carried out by reaction with a trifluoroacetic anhydride-fatty acid mixture (TFAAmethod) or a fatty acid chloride-pyridine system (acid chloride method). We performed the former reaction under conditions of temperature 30-50°C and reaction time 0.5-24 h; and the latter under conditions of temperature lOO"C, reaction time 2-8 h with DMF as a solvent. Both methods yield acylated woods with thermofluidity. Typicalthermograms of the acylated woodmeasuredwith the thermomechanical analyzer are shown in Fig. 2 [13,14]. The figure compares the thermoplasticity of the

659

Wood Plasticization

I

1

I

I

0

0 lauroylaled wood

0

I

I

I

100

200

300

T

0 0 400

('C)

FIGURE 2 Thermomechanicalcurvesfor (a) untreated, (b) acetylated. (c) propionylated, and (d) lauroylated sawdust.

lauroylated and, as examples of lower fatty acid esters, the acetylated and propionylated wood meal, with that of untreated wood meal. Acylation in all the above cases was carried out by the TFAA method. Measurements were carried out by following the collapse of a column of powder sample under a constant load of 3 kgfkm' in a heated glass capillary tube in the temperature range 20-350°C at a programmed heating rate of l"C/min. The untreated wood meal shows no thermal softening until it is heated to a relatively high temperature (around 200°C) and seems to have a softening point around 270°C. Measurements covering carbonization temperature ranges prove that the untreated wood sample as a whole does not show any thermal flow. On the other hand, the lauroylated wood meal shows a sharp drop dueto the complete flowof the sample at 195°C. Itis alsorecognizable in the figure that acetylated and propionylated woods, both prepared by the TFAA method, can have complete flow state. At this stage, it should be clarified whether the thermal flow is due to the complete melting of all the components of the acylated wood or of only some of them, with others remaining in a solid phase and flowing together with the former. To examine this point as well as to make other observations, films were molded from the acylated wood meals under heat and pressure. An example of the results is shown in Fig. 3. Conditions for the hot-pressing in this case were 140"C, 150 kgfkm', and 2 min. Transparent films, as shown in the figure, were obtained, which did not show any solid residue when observed under an optical microscope. This suggests that the flow behavior observed in the thermomechanical analyzer is caused by the thermal flow of all the main components of the chemically modified wood. The complete melting of all the components of lauroylated wood was also confirmed by measurement with a flow tester. It was found that lauroylated wood could be easily extruded from the flow tester under a rather low pressure (10 kgfkm')

660

FIGURE 3 Lauroylatedsawdust (left) and film(right)madeup perature, 140°C; time, 2 min; pressure, 150 kg/cm2.

Shiraishi

of it. Molding conditions: tem-

from a nozzle with a diameter of 0.5 mm, when heated to about 150-200°C. If any one of the major acylated woodcomponents could not reacha flow state, then it was impossible to extrude the whole material through the nozzle. The conclusion that the thermal flow of the chemically modified wood accompanies the flow of all of its main components should be reasonable when the IPN structure of wood components described above is considered. In addition, the temperature actually appliedin the molding mentionedabove is much lower than the flow temperature obtained with a thermomechanical analyzer for the same acylated wood shown in Fig. 2. In connection with this, it was proved experimentally that this flow temperature decreased with the increasing load applied. In general, amorphous polymers, such as higher-fatty-acid esterified wood, incontrast to the crystalline polymers, do not have a definable melting point but do have an apparent melting point. This point varies depending on the conditions of measurement, such as the load pressure or the rate of heating. If different substituent groups are introduced into wood, the apparent melting point (flow temperature) will be different. Even if the same groups are substituted, the apparent melting points of the products may be different, depending on the method and conditions of the reactions. In this regard, and also as another example of plasticization of wood by introducing large substituent groups into wood, results of benzylation are shown in Fig. 4 W]. This figure shows how the extent of benzylation affects the thermoplasticity of the etherified wood. Only sample 4, which has an ether content of 43%, gives a transparent film. Other filmsfromtheetherifiedwoodof lower substitution comparedwiththat of sample 4 contain solid granules. These results agree wellwiththe other experimental data [l61 obtained with the thermomechanical analyzer. Benzylated wood with ether content less than 40% does not show complete flow, but that with ether content within the 40-50% rangerevealsanapparentmeltingpointofabout 300°C. Further increasein the ether content results in a steady decrease in the melting temperature, which finally reaches about 200°C with an ether content of more than 60%.

FIGURE 4 Benzylated wood films. Conditions for benzylation were as follows:

Sample 1 2 3 4

40%Sodium hydroxide aq. (ml/g wood)

Benzylchloride (ml/g wood)

3.5 3.5 3.5 7.0

3.6 3.6 3.6 7.2

Reaction time" (h)

aeactlon temperature 100°C.

B. Small-SubstituentModification Larger substituents are not easily introduced using practical chemical modification methods. Therefore, as the next stage of this study, the author has been examining the plasticization of wood by introducing smaller groups within wood [ 13,14,17]. One of the most practical chemical modifications of wood is acetylation. The thermoplasticity of acetylated wood was found to vary with the acetylation method adopted. The plasticity of acetylated wood preparedby the TFAA method is high and shows a clear melting phenomenon at about 300-320°C even under a low pressure of 3 kgf/cm2, as shown in Fig. 2. However, thisis an exceptional case, as acetylated wood samples prepared

Shiraishi

662

by other conventional procedures do not show thermal flow. For example, acetylated wood samples prepared by an acetic anhydride-acetic acid in the presence of catalyst (sulfuric acid or perchloric acid) did not show thermal flow under the above pressure. Techniques for converting thermally nonmeltable acetylated wood prepared by the conventional method into plastic material have been investigated.The first method involves partial saponification subsequent to acetylation, which enhances the plasticity of cellulose acetate within wood. Figure 5 shows the thermodiagram of peracetylated wood and that of partially saponified wood. Although the peracetylated wood does not show thermofluidity, complete thermal flow can be seen for the sample after partial saponification. Considering that cellulose triacetate is a polymer that can be partially crystallized and has an apparent melting temperature as high as 300°C, and that cellulose acetate with high thermoplasticity is a diacetate with a degree of substitution of around 2.4, we can understand the effect of saponification as found above. The second method is mixed esterification with other acyl groups such as butyryl or propionyl groups. As shown in Fig. 6 , thermofluidity is conferred by this kind of mixed acetyl-propionylation. Mixed esterification is known to reduce the crystallinity and enhance the plasticity of cellulose as compared to cellulose triacetate. The third method for obtaining thermofluidity in acetylated wood isby replacing acetic acid by trifluoroacetic acid (TFA) in the pretreatment step of acetylation [ 1S]. This

1.0

0

100 Temperature

200 ( O C

300

)

FIGURE 5 Effect of saponification on the thermoplasticity of acetylatedwood. N o aging: peracetylated wood: Aging: partially saponified acetylated wood (acetylated hy acetic anhydride-acetic acid-sulfuric acid system at 50°C for 3 h, followed hy aging at 70°C for 3 h after partial neutration and dilution).

663

Wood Plasticization

Untreated 0

BBaDpo

A.A. : 0 :

1:

2:

3:

L

0

100

200

300

FIGURE6 Thermoplasticity (differential thermomechanical diagrams) of untreated wood and ;teetyl-propionylated woods prepared with different molar ratios of acetic anhydride ( A . A . ) and propionic anhydride (P.A.). Ratio of numerical values i n the figure represents the molar ratio of A.A. and P.A. used i n the esterification reaction.

method is considered as a simulation of the TFAA method. which can be used to prepare thermaily flowable acetylated wood asdescribed previously. Figure 7 shows that this method of acetylation can also give thermofluidity to acetylatedwood.Whatis more interesting is that the acetylated wood obtained shows an apparent melting temperature of about 210°C. This temperature is almost 90°C lower than the apparent melting temperature of ordinal acctylated wood or cellulose triacetate. Actually, the apparent melting temperature of cellulose triacetate is about 300°C (Fig. 8). The figure shows the differential thermonlechanical diagrams of a series of fractionated

664

Shiraishi 3-

(

h

Ace.arhyd(O.26).TFA(L.6). 50°C

0.. 2 -

0..4 -

0. .6-

0.8-

1.10-

0

200

300

FIGURE7 Effect of pretreatnlent time on thermal softening and flow behavior of acetylated wood. Pretreatment: TFA (4.6 lnllg wood)-acetic anhydride (0.28 mllg wood). S O T , 2-6 h; acetylation: 50°C. 3 h.

cellulose acetates. The apparent melting temperature is shown as the peak at the higher temperature. Even the cellulose acetate with very low molecular weight (0.61 X 10') flows around 300°C. I n this connection, the findings of the previous figure that the acetylated wood showed apparent melting temperature of about 210°C are quite characteristic. The compatibility and/or the mutual plasticization among the acetylated wood components can be considered to account for the shift of the flow temperature. Then, plasticization of celluloseacetate with an interaction with acetylated lignin could bc postulated. In order to confirm this. variations i n the thermofluidity of acetylated wood were investigated in relation to a stepwise extraction of acetylated lignin from the sample. The results are shown in Fig. 9. The curve for the acetylated wood before methanol extractionshowsa flow at almost 200°C. Thecurvefor the sampledelignitied by the peracetic acid method behaves like cellulose triacetate. The thern~omechanical diagrams for the acetylated wood partially delignitied by methanol extraction appear between the above-mentioned two curves. The methanol extracts were characterized by IR and NMR, and were found to be acetylated lignin contaminated with a small amount of acetylated hemicellulose.

665

Wood Plasticization

Cellulose triacetate

k

666

Shiraishi

k

0.4

tI

0.6

o : Not ext. withMeOH 0

IA I

A

: 12 h MeOH ext

: 24 h MeOH ext. : 48 h MeOH ext.

0 . 8 t o : 7 2 h MeOHext.

t

m : 120 h MeCti ext.

x : Delignification

From these results, it can be concluded that acetylated wood that was prepared aftcr TFA pretreatment flows at a rather low temperature. and this is caused by the plasticization of cellulose acetate by acetylated lignin within the acetylated wood cell wall. The splitting of the benzylaryl ether and, less frequently, the existing ester bond in the lignin macromolecule by the action of trifluoroacetic acid could play a large role in making the acetylated wood prepared after TFA pretreatment flowable at a n unexpectedly low temperaturc. I t can also be said that the same effect makes acetylated wood prepared by the TFAA method thermally flowable. I n the latter case, however, acetylated and partially cleaved lignin was removed during the purification stage, i.e., during washing with methanol. Thus, acetylated wood prepared by the TFAA method cannot flowat around 200°C but reveals an apparent melting temperature of about 300°C as mentioned above. These findings, in addition to the above-mentioned finding that acetylated wood prepared by the traditional method does not show thertnofluidity, suggest that the lignin inherently suppresses thermal f o w i n acetylated wood if i t is not partly split.

Wood Plasticization

667

As a fourth method to confer thermofluidity on acetylated wood, the author examined the explosion of wood as a pretreatment for acetylation. In this case, the explosion pretreatment activates lignin so that self-condensation of the lignin occurs during acetylation in an acidic medium. This could be prevented by the addition of a nucleophilic reagent, such as P-naphthyl methyl ether, during the acetylation. Figure 10 shows that this method is also successful in offering thermally flowable acetylated wood. The fifth method of offering thermofluidity to acetylated wood is the additional use of external plasticization. By mixing appropriate plasticizers, synthetic polymers, or both, acetylated wood can be converted to thermally flowable materials. Figure 1 1 shows how effective the use of some external plasticizers is. That is, even though the acetylated wood pulp prepared by the perchloric acid catalyst method cannot be molded into a film, it can be brought to moldability by blending with an equal weight of PMMA or, more effectively, with PMMA and dimethylphthalate in a 5 : 3 : 2 weight ratio. As otherexamples in this fifth investigation, the casesfor allylation as well as carboxymethylation of wood [ 171 are described here.

0

0.2

0.4 Q

0.6

0.8

1.0

T(

O c

1

FIGURE 10 Thermofluidity (thermoplasticity) of the acetylated woodpreparedafter explosion pretreatment at 179°C for I O min. Acetylation was conducted at 50°C for 3 or 6 h in the presence of ;I small amount of P-naphthyl ethyl ether (0.1 m o l per mole of phenyl propane unit estimated to exist in exploded wood n x A ) .

668

Shiraishi

i

i

l

FIGURE 11 Appearance of moldings from acetylated wood (wood pulp), acetylated wood-polymethyl methacrylate (PMMA) [ 1:1 (wt)], and acetylated wood-PMMA-dimethyl phthalate (DMP) [5:3:2 (wt)] blends.

In the upper part of Fig. 12, untreated wood meals are shown on the left, allylated wood meals in the middle, and carboxymethylated wood meals on the right. In the lower part thecorresponding molded materials made by hot-pressing are shown in the same order. It can be seen that none of these could be molded into a transparent film. Examinations were carried out to see if these allylated or carboxymethylated wood meals could be converted into thermally flowable material by mixing with plasticizers, synthetic polymers, or both. Figure 13 shows the results for allylated wood. The molded allylated wood (left), a film molded after mixing 25% with dimethylphthalate (DMP) (middle), and a film molded after mixing 100% with polymethylmethacrylate (PMMA) and 25% with DMP (right) are shown. The spots found in the film mixed only with DMP are bubbles formed during hotpressing. The film on the right is transparent, showing the complete thermofluidity of the sample. These results also indicate that conventional allylation does not cause sufficient internal plasticization and hence cannot confer thermofluidity on wood. The addition of adequate external plasticizer can confer thermofluidity on the allylated wood. Similar results were obtained for carboxymethylated wood (CM wood) (Fig. 14). CM wood powder is shown on the left, and after molding in the middle. A transparent film could not be obtained in this case. CM wood was converted to a thermally flowable material, however, after mixing with resorcinol (right). In other experiments it was found that hydroxyethylated wood meal can be molded into a transparent film after mixing and kneading with a small quantity of water. In all of these examples, it should be pointed out that the appearance of chemically modified wood meal is not very altered from that of untreated wood meal, as can be seen in Figs. 12-14.

Wood Plasticization

669

FIGURE 12 Appearance of etherified wood meal and moldings. Upper: untreated wood (left), allylated wood (center), and carboxymethylated wood (right). Lower: moldings from untreated wood (left), allylated wood (center), and carboxymethylated wood (right).

I

i

FIGURE 13 Plasticization of allylated wood by blending with DMP and PMMA-DMP. Appearance of moldings of allylated wood (left), allylated wood-DMP blend (center), and allylated woodPMMA-DMP blend.

670

Shlraishi

FIGURE 14 Plasticization of carboxymethylated wood by blending with resorcinol. Appearance of carboxymethylated wood meal (left), and moldings of carboxymethylated wood (center) and its blend with resorcinol (right).

From the above observations andfindings, it can be said that, in principle, regardless of the molecular sizeof the substituent group introduced, wood can be given thermoplastic properties by internal plasticization, that is, chemical modification, supplemented, if necessary, by external plasticization [17]. The sixth method for conferring thermofluidity on acetylated wood is thatof grafting vinyl monomerssubsequenttothe acetylation [13,14]. Inthiscasetheresultant vinyl polymer acts as a plasticizer within wood. Figure 15 shows that acetylated wood prepared by the perchloric acid catalyst method does not reveal thermal flow, whereas subsequent grafting with styrene changes the acetylated wood to a thermally flowable material. This grafting method is similar in principal to the method supplemented by external plasticization with mixing with synthetic polymer. In that case it was not easy to find appropriatesyntheticpolymers as suitableplasticizersforacetylated wood. A suitable polymer-plasticizer should be one that is compatible with the polymer component of the acetylated wood at the molecular level. Compatibility between polymers, however, rarely occurs. To overcomethis difficulty, graftingtechniquescanbe used. Compatibility of acetylated wood components and synthetic polymers is attainable by the grafting technique. Even with a combination of acetylated wood and synthetic polymers, with which a homogeneous transparent film cannot be effected merely by blending, the use of grafting techniques can result in a transparent molded film. The method of grafting is also important. It is desirable to select a method by which grafting or polymerization of monomer compounds occurs homogeneously within acetylated wood to produce products with high compatibility. A representative experimental result (Fig. 16) shows how the grafting technique is superior to blending in yielding high compatibility among the components of

Wood Plasticization

671

O~

a

0.5

1 .o

FIGURE 15 Thermomechanical behavior of acetylated wood ( 0 ) and acetylated wood-polystyrene composite prepared by the y-ray-induced graft copolymerization in a pyridine medium (e). Conditions: total dose 2 Mrad; resultant weight increase 76.7%.

acetylated wood and synthetic polymers. The figure illustrates the temperature dependence of loss tangent obtained by a viscoelastic measurement for composites prepared by blending acetylated wood [acetylated radiata pine refiner ground pulp (RGP)] with PMMA with a kneading technique or by grafting MMA onto the acetylated wood. This figure shows that grafting results in forced compatibility of acetylated wood with PMMA. Although the blended material of acetylated wood and PMMA shows two T, peaks ascribable to the two constituting components, the corresponding grafted material reveals only one intermediate T, peak. This shows that graft products of PMMA onto acetylated wood com-

672

Shiraishi

ponents can be partially produced which can act as colnpatibilizers and which enable the total mixing and interaction of the acetylated wood and PMMA to be enhanced. On the other hand, some investigations have been advancing from the point of view of not only conferring but also enhancing the thermofluidity of chemically modified wood. Among them, the method of using TFA as a reagent for the pretreatment, to split selectively the parts of bonds combining lignin units [IS], a s well as methods chlorinating lignin's aromatic ring parts of chemically modified wood [ 19,201, are very effective and specific. The former method has already been explained i n detail as the third method of offering thermoplasticity to acetylated wood. It should be noted here that since a significant shift of the apparent melting temperature toward the lower temperature can be effected as mentioned previously, the former method is also useful as a method for enhancing the thermofluidity of modified wood. The latter method is concerned with a postchlorination of the chemically modified wood. Morita, Sakata, and co-workers have been investigating the effect of postchlorination of cyanoethylated wood on the thermofluidity of modified wood [ 19,201. The chlorination was conducted by soaking the modified wood in a delute aqueous chlorinesolution (0.1-0.2%), at a low temperature, for a short time (O'C, 5 min) under stirring, followed by washing with water and drying. Under these mild conditions of chlorination, chlorine is known to react selectively with the aromatic ring of lignin. The chlorination of lignin can be considered to cause ( I ) chlorine substitution reactions at the S - and 6-positions of lignin aromatic rings; (2) chlorination substitutions at the I-position of the aromatic ring,

Wood Plasticization

673

eliminating the propane side chain; (3) hydrolysis of ether bonds (demethylation of methoxy groups and hydrolysis of phenyl ether bonds); and (4) oxidation of the aromatic rings. Among these, the former two reactions are primary reactions, while the latter two are side reactions. As representative results of these studies, it has been shown that the flow temperature ofthe cyanothylated wood can be shifted downward by 100- 120°C (Fig. 17). At this point, decrystallization of the samplcs is known to proceed without degradation of the cyanoethyl cellulose fraction [19,20]. The lowering oftheflow temperature of cyanoethylated woodis thought to be attributable to a loosening of the lignin structure by chlorine substitution, action ofthe chlorinated lignin as an external plasticizer for cyanoethyl cellulosc within the modified wood (mutual interaction among modifiedwood components), a s well as decrystallization produced by the chlorination [X)].Lowering the flow temperature is considered not to be caused by the degradation or depolymerization of the components of the chemically modified wood. This is concluded from the observations that moderate chlorination, as described above, results in a tremendous drop in the flow temperature of the chemically modified wood, as well as the fact that the viscosity change of the aqueous solution of cyanoethyl cellulose before and after the chlorine treatment is very low (7%) 1201. Similar phenomena have been found with other chemically modified woods. For example, chlorination causes the lowering of the flow temperature of benzylated wood, and results in thermally flowable hydroxyethylated wood that does not show thermofluidity before the chlorination.

V.

APPLICATION OF THERMOPLASTICIZATIONOF WOOD

It has become possible to confer thermofluidity on wood, as described in the previous section, and based on these findings, several attempts have been made to apply the result to useful products. Examples are the preparation of films, sheets, and other moldings from chemically modified wood; preparation of three-dimensionally cured plasticlike woodboard; preparation of deep-drawable hardboard; application to hot-melt adhesives; surfacelayer plasticization of wood intended to develop surface densitification and emboss processings; and preparation of modified plywood, LVL (laminated veneer lumber), particle board with densified surface, flame resistance, decay resistance, and insect resistance. The preparation of films, sheets, and other moldings from chemically modified woods or their blends with other materials, mentioned above as the first example, has been attempted, and the mechanical as well as the viscoelastic properties of the parts of the moldings have been studied. The mechanical and other properties of the modified wood film depend on the type of substituent groups, the extent of substitution, and the method of modification. Figure 18 [21] shows a stress-strain curve for a film from caprylated wood. Curve a is a stressstrain curve for Hinoki cypress stretched in the fiber direction, and curve b is for the same wood stretched perpendicular to the fiber axis, both of which are shown as reference. Curve c is caprylated wood. In the case of the untreated wood specimens (a and b), the elongation at break is small ( 1 -2%). On the other hand, the elongation at break found for the caprylated wood is very large (as high as 100%). In this case, a yield point is recognizable, and the breaking strength is almost equal to that of Hinoki cypress perpendicular to the fiber direction.Thus, caprylated wood film can be said to exhibit elastomerlike mechanical properties.

674

Shiraishi

I

I

l

.-

E

d

l00 200 Temperature ( "C l FIGURE 1 7 Thermal softening and flow curves for CE-wood treated with chlorine.

The temperature dependences of dynamic modulus E' and loss tangent tan S for the caprylated wood film are shown in Fig. 19 [22]. For comparison, the same relation was obtained for an untreated Beihi cypress specimen (fiber direction) and is shown as a dashed line in the figure. The dynamic modulus E' for the untreated wood shows a small decrease within the sameorder of magnitude when the temperature increases from - 180°C to

(U

E V

\

m 400

-Y

" I 200

C I

I

I

I

20

40

60

80

E: (%) FIGURE 1 8 Stress (n)-strain ( E ) curve for caprylated wood film (c) and wood specimens (Hinoki) with two different edge cuttings (a and b).

Wood Plasticization

675 E'

-----------------------------""~~""""""""""""~""" """

-1

1010-

rt I

N

S

2 C z

-0.1

109-

E

"-"

iJl

108

"_"

- 0.01

107, -200

-100

0

100

200

T (C)

FIGURE 19 Temperaturedependenceof

E' and tan 6 forcaprylated

wood film and wood

specimen.

250°C. Although the dynamic modulus E' for caprylated wood film at - 170°C is very close to that for the untreated wood, it decreases with temperature and, especially, is dramatically lowered between -90 and -40°C and between 30 to 100°C. As a result, the E' value becomes three orders of magnitude lower than that of the untreated wood at 140°C. With respect to tan S , three relaxation processes are detectable within the experimental conditions. They are labeled as a, p, and y process in order of decreasing temperature. These processes were identified on the basis of the information obtained from a series of cellulose acetates. The a process is assigned to be due to a micro-Brownian motion of the main chain of the caprylated wood (the process corresponding to the glassrubber transition); the p process to the motion of the capryl side chain introduced, and the y process to the motion of a part of the side chain (the motion initiated by a minimum of three methylene groups in addition to the oxycarbonyl group of the side chain). Figure 19 shows that a large decrease in E' is caused by the motion initiated by the whole acyl side chain ( p process) and that the glass-rubber transition temperature of the caprylated wood is higher than room temperature. These results are interesting in relation to the results shown in Fig. 18. That is, the observation in Fig. 18 that the caprylated wood film behaves as though it is an elastomer at the measuring temperature (room temperature) can be directly explained by the effect of the micro-Brownian motion of the acyl side chain, and not by micro-Brownian motion of the main chain. As another example of the mechanical properties of a modified wood with large substituent groups, stress-strain curves for a benzylated wood film measured at various temperatures are shown in Fig. 20 [15]. The stress-strain curve varies widely with the temperature. It clearly shows a yield point when measured at 55 and 81°C while the film reveals brittleness at room temperature. The breaking strength of the film decreases with increasing temperature, and the breaking elongation increases rapidly with temperature up to around 1 17°C and then decreases beyond it.

676

Shiraishi

n

E ob)

FIGURE 20

Stress (u)-strain (c)curve for benzylated wood f i l m .

I n Fig. 2 I , the temperature dependence of dynamic modulus E' and loss tangent tan 6 for the benzylated wood film is shown [ 151. Compared with the corresponding results with a Beihi cypress specimen (fiber direction), results with the benzylated wood specimen are similar to those obtained with general amorphous synthetic polymers. Within its glassystate region, where the E' value is more than 10"' dynes/cm2, a secondary dispersion ( p process), accompanied by a very slight decrease in E', is recognized around -9O"C, and the tan 6 curve reveals a corresponding peak. This process is assigned to be due to the motion of the benzyl side chain introduced. The effect of the motion of the benzyl side chain on the decrease i n the E' value is considerably smaller compared to that caused by the motion of the capryl side chain mentioncd above. On the other hand, a primary dispersion ( a process) appears around 118°C in this benzylated wood. The findings in Fig. 21 explain the observations shown in Fig. 20;i.e., benzylated woodfilm behaves asa brittle and glassy polymer film at room temperature, and the breaking elongation increases with the measuring temperature up to around 118°C (the glass transition temperature) and then decreases beyond it. In Table I , mechanical properties of the films from certain benzylated woods are compared with those of several common synthetic polymers [ 161. It can be seen that there are no fundamental differences between the two. The physical properties of benzylated wood can be altered or, more desirably, enhanced by blending with synthetic polymers and/or low-molecular-weight plasticizers. Based on similarity in chemical structure between the benzylated wood and polystyrene, good compatibility can be expected between them. Thus, the related blending has been studied [23]. In Fig. 22, apparent melting points (T,) for the benzylated wood, the polystyrene, their blends in several ratios. as well as a typical chlorinated benzylated wood (Cl-BzW) are shown. From this figure, it is apparent that the thermoplasticity of benzylated

677

Wood Plasticization llT 1

lO$O

71-3 -

FIGURE 21 specimen.

l

I

l

100

0

100

Temperaturedependence

200

of E' and tan S for benzylatcd wood film and wood

TABLE 1 Comparison of Mechanical Properties of Benzylated Wood Films (C-l to C-7) with Those of Common Synthetic Polymers Sample plastics

c-1 C-3

c-4 c-7 Vinyl chloride resin (hard) Polystyrene A S resin Methacrylic resin Polypropylene Polypropylene (glass reinforcement) Cellulose acetate Ethylcellulose Polyurethane

Tensile strength (kg/cm')

Elongation (%)

Young's modulus (kg/cm')

300 24 1 320 337 420-530 350-530 630-840 490-770 300-390 420- 1020 130-630 140-560 320-590

5.5 2.5 3.2 4.7 40-80 1 .o-2.5 1 S-3.7 2.0- 10 200-700 2.0-3.6 6.0-70 5.0-40 100-650

10,057 15.051 14,477 14,140 25,000-42,000 28.000-42.000 28,000-39,000 27.000-32,000 1 I ,000- 16,000 32.000-63.000 5,000-28,000 7,000-2 1,000 700-25.000

Shiraishi

678

,

1010 614812

,

I

.

,

.

,

416 218 BzWPS (wt/wt)

.

0110

FIGURE 22 Apparent melting temperature (TJ for benzylated wood (BzW), polystyrene (PS), and their blends with various ratios. Cl-BzW should be referred to the text.

wood can be markedly enhanced. Especially, the samples obtained by blending more than 50% of polystyrene are known to have T , values almost equal to that of the polystyrene. In this connection, Morita and Sakata [l91 reported that the apparent melting temperature of the chemically modified wood can be considerably lowered by a weak chlorination aftertreatment, which gives the Cl-BzW mentioned above. Dynamic viscoelastic properties of the benzylated wood, the polystyrene, and a blend of the two in equal weight ratio were measured, and the obtained temperature dependencies of logarithmic decrement (aT)are shown in Fig. 23. In general, the comparison of this kind of viscoelastic relaxation curves is suitable for judging qualitatively the compatibility of a polymer blend. That is, in the case of the compatible polymer blend, the glass transition temperature (T,) appears as a single peak in a position proportional to the existing amounts of component polymers, whereas in the case of an incompatible polymer blend, the individual phase domains retain the glass transitions of their respective parent homopolymers. These viscoelastic relaxation data were correlated with electron microscopic observation results and the following have been pointed out. In the case of the compatible blend system, the component polymers were found to exist as domain mixtures, whose grain diameter was less than 15 nm (150 whereas in the case of the incompatible blends, they were often existing as individual phase domains whose diameters were larger than 100 nm. I n the case of the semicompatible blend, in which even though a number

W),

FIGURE23 Temperature dependencies of logarithmic decrement (a,)for benzylated wood (BzW) (H),polystyrene (PS) ( 0 ) . and their blend with BzWPS = 5/5 by weight (A).

Wood Plasticization

679

of T, peaks can be observed, the transitions are broadened and their temperatures become closer together, and the individual phase domains have grain diameters in the range 20100 nm. Consequently, the experimental results shown in Fig. 23 indicate that relatively significant molecular mixing takes place in the benzylated woodpolystyrene blend. When the transparency of the molded sheets from this benzylated woodpolystyrene blend is also taken into consideration, the two- or multicomponent polymer system is concluded to be a compatible polymer blend, and the individual phase domains have grain diameters of less than15 nm. Figure 24 shows the results of tensile tests obtained for films prepared from benzylated wood, polystyrene, and their blended mixtures. From this figure, it can be seen that the tensile properties obtained for these blending composites are not very good. That is, even though the benzylated woodpolystyrene blends are miscible, the interfacial bonding of the related individual phase domains is not enough for the corresponding molded sheets to have desirable mechanical properties. This result is not necessarily unusual, and in order to overcome this situation, compatibilizers can be used in various ways. In this case, a styrene-maleic anhydride copolymer (SMA: Arc0 Chemical, Inc., Dylark 232; maleic anhydride content 8 wt%) was tested in order to make clear whether it can work as a compatibilizer ornot when certain amounts of it are kneaded with the benzylated wood and polystyrene at a temperature as high as 160°C. The result is shown in Figs. 25 and 26, in which the amount of benzylated wood was fixed to 50 wt% and the mixing ratio of polystyrene to SMA, the total amount of which was 50 wt%, was changed. From the figures, it is seen that both the tensile strength and the breaking elongation take the maximum values when SMA is added at 5 wt% to the total weight of the blended composites. This result shows clearly that benzylated w o o d SMA graft copolymers, which have been produced by kneading [blending at high temperature (16O"C)], can act as a compatibilizer. In other words, due to the localized existence of thus-produced benzylated wood components/SMA graft copolymers at the interfaces between the benzylated wood and the polystyrene phase domains, as shown in Fig. 27, their interfacial bonding is considered to be improved, giving materials with excellent physical or mechanical properties. In Fig. 25, when the amounts of SMA added to polystyrene exceed 1096, the tensile strength of the composite films obtained shows a monotonic decrease. This is considered 15 14 13 12 11

515

1010

BzWPS (wtlwt)

FIGURE 24 Tensile properties for benzylated wood, polystyrene, andtheir blending mixtures. 0 , tensile strength (um1.,J; A, breaking elongation m, Young's modulus ( E ) .

Shiraishi

680

FIGURE 25 Polystyrene-to-SMAratiodependency of thetensilestrength of themoldedfilms prepared from their blends with benzylated wood (the benzylated wood content is fixed at SO wt%).

L

1010

, . , . , . , . 812

614 416 218 PSISMA (wtlwt)

0110

FIGURE 26 Polystyrene-to-SMA ratio dependency of the breaking elongation of the molded films prepared from their blends with benzylated wood (the benzylated wood content is fixed to 50 wt%).

SMA Chain

P

FIGURE 27 Schematic diagram showing the localized existence of benzylated wood components/ SMA graft copolymer at the interface between the benzylated wood and the polystyrene phase.

Wood Plasticization

681

to be caused not by the decline of the mechanical properties of products with increasing SMA content, but from the formation of crosslinking within the benzylated wood components phase, the amounts of which increase with the amount of SMA added. The formation of crosslinking during kneading or blending results in products with less moldable thermomechanical properties. Films can sometimes be molded directly even from modified wood with small-molecular-sized substituents. For example, acetylated wood prepared by the TFAA method and the TFA pretreatment method, as well as acetyl-butyrylated wood, can be molded into films. Table 2 is a n example of the experimental data. Even though the films can be molded from wood substituted by small groups, the films show brittleness. In order to improve the physical properties of these moldings and to enhance the thermofluidity and moldability of modified woods having small-molecular-sized substituent groups, these modified woods have also been blended and plasticized with plasticizers, synthetic polymers, or both. Graftcopolymerization with synthetic polymers has also been tried, and the resulting physical properties have been measured. Mechanical properties of films prepared after the external plasticization of allylated as well as carboxymethylated wood meals with synthetic polymers andor low-molecularweight plasticizers [ 171 are shown in Table 3. Corresponding data for each of the synthetic polymers used as the external plasticizer and for a benzylated wood film are also included for comparison. It is known that the films from the externally plasticized allylated as well as carboxymethylated wood reveal mechanical properties comparable to those of the corresponding synthetic polymers. Reinforcement because of the presence of modified wood can be recognized. It can be concluded that the physical properties of chemically modified wood alloyed with synthetic polymers with or without other plasticizers are dependent on the species of synthetic polymers and the composition of the alloys. Generally, external plasticization by graft-copolymerization results in greater enhancement of thermofluidity (moldability) as well as better physical properties of the moldings than those provided by mere mixing (blending). In the case of graft-copolymerization, the method of grafting also plays an important role. Wood-synthetic polymer composites with higher thermofluidity and better physical properties can be obtained when grafting techniques resulting in a uniform distribution of polymers within the wood cell wall are adopted (seeFigs. 16 and 28). As shown in Fig. 28, however, even though the graft-polymerization technique is adopted, two peaks assignable to the respective primary dispersion of acetylated wood and that of polymethyl methacrylate (PMMA) can be found in the tan S-versus-temperature curve of a graft product prepared by use ofan azobisisobutyronitrile initiator (AIBN catalyst

TABLE 2 Mechanical Properties of Films from Acetyl-Butyrylated Woods Prepared by TFAA Method Sample (acety1:butyryl)

0: 10 1 :9 3:7 1:l

Tensilc strength (kghn’)

Elongation

263 294 313 414

4.9 6.0 12.8 12.5

(%l

Young’s modulus (kgkm’) 1 8,600

8.840 7,720 10,700

Shiraishi

682

TABLE 3 Mechanical Properties of Films from Blends of Allylated Wood (AW) or Carboxymethylated Wood (CMW) with Synthetic Polymers and/or Plasticizers. Data on Synthetic Polymers and Benzylated Wood Are Also Shown as References

Sample AW:PE = 1:2, blend PE AW:PP = 1.2, blend PP AW:PMMA:DMP = 4:4: 1, blend PMMA

CMWResorsinol = 1: I , blend Benzylated wood

X

Tensile strength IO-" (dynedcm')

Elongation

9.22

14.6

(%)

0.41-3.82

20- 1000

15.9

3.8

2.94-3.82

200-700

13.1

4.1

4.80-6.23

3-10

8.73 2.77-4.06

28.2 1.6 1-9.49

Young's modulus (dynes/cm')

X 10""

2.43 0.096-1.3 4.35 1.10-1.55

3.28 -3.142.56 1.62-4.17

method). This result is quite different from that of Fig. 16, where a curve for a grafted product prepared by a vinyl polymerization without initiator in the presence of water is included. In this case, only one primary dispersion peak appears, suggesting a high compatibility at the molecular level among the components of the acetylated wood and PMMA. As already mentioned, when individual phase domains composing the blended composite have grain diameters less than 15 nm, the composite can exhibit one principal glass transition. Itis also known that the tensile strength and the elongation at the break of the molded films become larger when higher interfacial bondings among the individual phase domains are developed. Further addition of adequate plasticizers, such as dimethyl phthalate (DMP), di-2-ethylhexyl phthalate (DOP), tricresyl phosphate (TCP), etc., in suitable quantities, to chemically modified wood-synthetic polymer composites can substantially enhance the thermofluidity of the polymer alloys. More homogeneous films are thus obtainable by hot-pressing. From the standpoint of preparing molding materials with excellent properties from chemically modified wood, it is possible to make use of chemicals that can act as plasticizers for the modified wood as well as that can react with the components of the modified wood during kneading and hot-pressing. One example of this is the preparation of films from blends of acetylated wood-resorcinol paste with various synthetic polymers (polyvinyl acetate, ethyl acrylate, acrylonitril-butadiene rubber, etc.), to which a definite amount of formalin was added. The mixture was reacted under heat and kneading, and then molded into films by hot-pressing. Three-dimensionalcuringoccursduring molding. Physical properties are shown in Fig. 29. In this case, formalin was added in a quantity of 0.5 mol of formaldehyde for each mole of resorcinol. Although the mechanical properties of the films are largely dependent on the kind of synthetic polymers added, tensile strengths up to 770 kgf/cm2 could be obtained even within the small number of experimental trials. Another example of the use of chemicals as plasticizers as well as reacting agents for chemically modified wood is that of preparing three-dimensionally cured plasticlike

683

Wood Plasticization

PMMA+AcP

X.method N.method

F.method

c t!

A.method

Blend ( DMP)

Blend

-50

0

50

100 T ("C1

150

200

FIGURE 28 Temperature dependence of ay(= tan 6) for various PMMA-acetylated wood graft products prepared by different procedures and that for PMMA-acetylated wood blends. X.method, xanthate method; N.method, noncatalyzed grafting method; Emethod, redox method with Fenton's reagent; A.method, catalyzed method with AIBN; Blend (DMP), blending method using kneader in the presence of DMP; Blend, blending method using kneader.

wooden board, developed by Matsuda and Ueda [24]. The special feature of this method is the combined use of carboxyl group-bearing esterified wood, such as maleoylate wood (maleic acid half-esterified wood), phthaloylated wood (half-ester), and bisphenol A diglycidyl ether. In this case, a crosslinking reaction accompanied by plasticization occurs during the hot-press molding stage, resulting in new types of cured wood. Various molded boards having plasticlike appearance as shown in Fig. 30 were obtained. The color of the board depends on the species of the esterified wood. Most common are red-brown, yellowbrown, or black-brown colored boards. When the wood content of the boards falls to 6070%, materials with high water resistance can be obtained. The physical properties, such as strength, elongation, etc., of the boards are superior to those of conventional boards (fiberboard, particle board, etc.). Cured board from meleoylated wood is superior in blending strength, while boards from phthaloylated wood have excellent compressive strength

684

Shiraishi

l

aoo

Ac -Res - PVAc

7: 3 1 3 (0.5rnol HCHO)

56oo

U

\

0

x

m

m

E

I

v1

k-Res-NBRI

7:3:3

(0.5mol HCHO 1 L

0

20

40 Strain(%1

60

FIGURE 29 Stress-strain curves for the acetylated wood-resorcinol-formaldehyde-synthetic polymer blended and cured films.

(2026 kgfkm'), hardness, thermomoldability, and water resistance. Esterification of wood with melaic anhydride, or phthalic anhydride, is very simple and can be accomplished by heating without the use of solvents 1251. Maleic anhydride and phthalic anhydride are cheaper than acetic anhydride, and the use of a catalyst is not usually necessary. When a catalyst is required, sodium carbonate can be used with satisfactory results. An example of a three-dimensionally moldable fiberboard-in other words-deepdrawable fiberboard, is shown in Fig. 3 1 [26,27]. In this case, wood fibers were chemically modified to levels that made the product thermoplastic but not thernlofluid. As shown in the figure, when wood fibers are chemically modified to an appropriate extent by acetylation o r other lower-aliphatic-acid esterification, fiberboards that are highly moldable under hot-pressing can be obtained. Application of chemically modified wood, especially blends with synthetic polymers, to the film-type hot-melt adhesives has been studied. This application makes use of the thennofluidity of chemically modified wood, which can be enhanced by blending with appropriate synthetic polymers. An example of the results of this study is shown in Fig. 32. Butyrylated wood was blended with polyvinyl acetate (PVAc) in various blending ratios, and adhesiveness as hot-melt adhesive was estimated based on Japanese Industrial Standard (JIS) K 6852. The figure shows that the adhesive strength of the hot-melt adhesive can be enhanced by blending with the synthetic polymer (PVAc). These blended adhesives give stronger adhesion than PVAc film. The maximum compressive shear ad-

Wood Plasticization

685

i

FIGURE 30 Three-dimensionally cured wooden boards prepared by curing phthaloylated woodis 20 X bisphenol A diglycidyl ether blends. The maximum dimension of the board in the figure 20 X 0.7 cm.

FIGURE 31 Hot-pressdrawability of hardboard withadensity of approximately 1.0: ordinary hardboard (left) and hardboard prepared from acetylpropionylated wood pulp (right).

686

Shiraishi e e

e

e m

e

m

150

e

m

e

e

~~

0

0.3

0.5

0.7

1.o PVAc

Butylated Wood Composition

FIGURE 32 Relationbetweenblendingratio

for butylatedwood-PVAcsystemsandadhesion

strength.

hesive strength found in Fig. 32 is more than 150 kgf/cm2, which is 1.5 times higher than the JIS specification demands.

VI.

LIQUEFACTION AND DISSOLUTION OF CHEMICALLY MODIFIED WOOD

Chemically modified wood has been found to liquefy and/or dissolve in various neutral aqueous solvents, organic solvents, or organic solutions, depending on the characteristics of the modified wood [2-7,28-341. So far, three methods have been found for this purpose. The first trial of the liquefaction and/or dissolution of chemically modified wood was accomplished by treating chemically modified wood in the presence of organic or aqueous solvents, or solutions at various temperatures in the range 80-290°C [34]. One example used wood samples esterified with aseries of aliphatic acids which could be liquefied in benzyl ether, styrene oxide, phenol, resorcinol, benzaldehyde, aqueous phenol,

Wood Plasticization

687

chloroform-dioxane mixture, benzene-acetone mixture, etc., after treating at 200-270°C for 20-150 min. Examples of the results of the liquefaction are shown in Fig. 33. Benzyl ether solutions of a series of acylated woodsare shown. A series of aliphatic acid-acylated wood, with almost complete substitution, prepared by the TFAA method, were liquefied in benzyl ether by heating at 250°C for 20 min. All the samples except the acetylated wood were shown to be homogeneously liquefied. Even the acetylated wood was found to be completely liquefied in the solvent if it was treated at 270°C for 20 min. Carboxylated wood, allylated wood, and hydroxyethylated woodhave been found to dissolve or liquefy in phenol, resorcinol, or their aqueous solutions, formalin, etc., after allowing them to stand or with stimng at 170°C for 30-60 min [17]. Benzylated wood could be dissolved in DMSO when heated at 80°C. Another method for liquefaction makes use of solvolysis during the process [33,35]. By using conditions which allow phenolysis of part of the lignin, especially in the presence of an appropriate catalyst, the liquefaction ofchemically modified wood into phenols could be accomplished under milder conditions (80°C for 30-150 min). Allylated wood, methylated wood, ethylated wood, hydroxyethylated wood, acetylated wood, and others have been found to liquefy in polyhydric alcohols, such as 1,6-hexanediol, l,Cbutanediol, 1,2ethanediol, 1,2,3-propane triol (glycerol), and bisphenol A, using liquefaction conditions as just described above. Each of these materials caused partial alcoholysis of lignin macromolecules [22]. The liquefaction process can produce a pastelike solution with a high concentration of wood solute (70%). The third method of liquefaction or dissolution involves postchlorination [%l. When chemically modifiedwoods are chlorinated, their solubility in solvents is tremendously enhanced. For example, at room temperature, cyanoethylated wood can dissolve in o-cresol

FIGURE 33 Dissolution of fatty acid-acylated wood prepared by TFFA process in benzyl ether (C = number of carbon atoms in the acyl group). Dissolution conditions: 250°C and 20-40 min.

Shiraishi

688

by only 9.25%. However, once chlorinated, it can dissolve almost completely in the same solvent at room temperature.

VII.

APPLICATION OF THE LIQUEFACTION AND/OR DISSOLUTION OF CHEMICALLY MODIFIED WOOD

There are many potential applications for the liquefaction and/or dissolution of chemically modified wood.Examples include the fractionation of modified wood components [34,36,37], the preparation of solvent-sensitive and/or reaction-sensitive wood-based adhesives 123 1,35331, the preparation of resinified wood-based moldings such as the foam type 131 1, and the preparation of wood-based fibers and their conversion to carbon fibers (391. To fractionate modified wood components, the dissolution-precipitation technique has been successfully used [34,36.37]. Attempts to prepare wood-based adhesive as well a s their curing based on the reaction of modified wood and the reactive solvents are reported in the literature. In these cases, phenols, bisphenols, and polyhydric alcohols have been used as reactive solvents [31,35,38,40,41 I. Combined use of these reactive solvents with reactive agents, crosslinking agents, and/or hardeners has given rise to phenol-formaldehyde resins (such as resol resin), polyurethane resins, epoxy resins, etc., all of which contain meaningful amounts of chemically modified wood. The chemically modified woods are not designed merely to dissolve and disperse in the final resins, but to react chemically and bond to thc resins. This can be achieved by liquefaction of the chemically modified wood into the reactive solvent using solvolysis techniques. In the case of epoxy resins. it can also be achieved by reacting various alcoholic hydroxyl groups remaining in the modified woods with epichlorohydrin, resulting i n introduction of glycidyl groups. Crosslinking within and between degraded wood components, especially between degraded polysaccharide components during the last stage of resinification, by reaction with crosslinking agcnts, can also be used. In order to prepare wood-based resins with meaningful amounts of the wood component, it is very important to liquefy or dissolve the chemically modified wood into reactive solvents in high concentrations (more than 50% is preferable). When hydrophilic chemically modified woods, such as carboxymethylated woods, hydroxyethylated woods, or ethylated woods are used in wood-based adhesives, aqueous resol resin adhesives that maintain their solution state during the preparation-that is, from the period of the completion of the liquefaction or dissolution in phenol to the final stage of prepolymer formation-are obtainable[35,38,40,41 I, Whcn a phenol solution with a concentration more than 50% is obtained, the chemically modified wood powder cannot bc completely immersed in phenol but can be only partly penetrated by the phenol during the first stage of dissolution (see Fig. 34). However, when the heterogeneous mixture is allowed to stand for about 30 min at 80°C (without stirring in the presence of appropriate amounts of hydrochloric acid as catalyst), a homogeneous paste can be obtained. Subsequent stirring of the paste for about I - 1 .S h enhances the liquefaction. In this liquefaction process, a certain degree of phenolysis ofwood components. especially that of lignin, takes place, which makes it easy to liquefy and dissolve them in phenol. After neutralizing the paste with aqueous sodium hydroxide, il definite amount of formalin and sodium hydroxide are added and resinified in accordance with the conventional procedure for preparing the resol resin adhesives. The appearance of thc resin obtained (Fig. 3 5 ) is similar to that of the corresponding commercial phenol resin adhesive.

Wood Plasticization

689

FIGURE 34 Early stage of liquefaction of hydroxyethylated wood meal into phenol. Photograph was taken after 10 min of liquefaction at 80°C.

The wood-based resol resin adhesives have superior gluability and workability. The adhesives can be used with fillers, thickeners, and fortifiers such as wheat flour, coconut shell, walnut flour, and polymeric MD1 (4,4’-diphenyl methane diisocyanate). It has been found that by using mildconditions for the first liquefaction and dissolution process, which does not completely dissolve all of the wood meal, the addition of fillers and thickeners into the adhesives become unnecessary. On the other hand, the addition of appropriate fortifiers, especially crosslinking agents such as polymeric MDI, into the wood-based adhesives enhances their dry-bond and waterproof gluabilities remarkably. In the cases of utilizing hydrophobic chemically modifiedwoodssuch as acetylated wood, butyrylated wood, etc., the liquefaction and dissolutionof these chemically modified woods into phenol is possible, but the products tend to become solids when they are subsequently resinified. In these cases, resinification should bedone within the kneader so that the products become solid powders (Fig. 36, center). The products can be dissolved in suitable solvents such as ethyl acetate, resulting in solutions as shown on the right side of Fig. 36, and can be used as reaction-sensitive liquid adhesives. For the preparation of wood-based polyurethane as well as epoxy resin adhesives, the above-mentioned hydrophilic chemically modified woods, prepared by conventional methods, are liquefied and dissolved in polyhydric alcohols or bisphenol A in a manner similar to the dissolution in phenol [31]. Concentration of the modified wood is usually more than 50%. Diluents such as ethanol or methanol are also often added to the dissolution system according to the requirements. After the liquefaction, the pastes are neutralized and the diluents are distilled off. When the pastes are used in combination with suitable polyisocyanatecompounds,theybecomewood-based polyurethane adhesives. When the pastes are further reacted with epichlorohydrin, glycidyl etherified resins are formed. These can be used with hardeners such as amines and acid anhydrides, and they

690

Shiraishi

FIGURE 35 Appearance of hydroxyethylated wood-phenol resin adhesive prepared.

become wood-based epoxy resin adhesives. Figure 37 shows epoxy resin adhesives prepared under various conditions after liquefying allylated wood into the same weight of bisphenolA. These are comparedwith a commercial epoxyresinadhesive. Figure 38 shows the fluidity of one of the epoxy resin adhesives prepared from allylated wood. Generally, the wood-based epoxy resins tend to become very viscose or solid, depending on the conditions of preparation, and require dilution or dissolution with solvents such as ethyl acetate, acetone, etc. These resins make satisfactory adhesives, which can be used in waterproof glues. Molding materials such as foams or shaped moldings can be obtained from chemically modified wood solutions of polyhydricalcohols, phenols, and bisphenolsas described above [31]. An example is shown in Fig.39. These can be prepared by adding anadequate amount of water as a foaming agent and a polyisocyanate compound (polymeric MDI) as a hardener to the 1,6-hexanediol solution of allylated wood, mixing well, and heating. When heated at 100°C, foaming and resinification of the resins are initiated within 2 min and are completed within several minutes. If promotors such as triethylamine are added, rapid reactions occur even at room temperature and foams can be obtained within several minutes. The foam shown in Fig. 39 has a very low apparent density (0.04 g/cm’). It also

Wood Plasticlzatlon

691

FIGURE 36 Preparation of acetylatedwood-phenolformaldehyde adhesive. Acetylated wood meal (left); phenol-resinified acetylated wood (center); ethyl acetate solution of phenol-resinified acetylated wood (right).

FIGURE 37 Appearance of epoxy resins prepared by liquefaction and dissolution of allylated wood into bisphenol A followed by glycidyl etherification, as well as a commercial epoxy resin (Araldite, CIBA-GAYGET, upper left).

692

Shlraishi

FIGURE 38 An example of the solution property shown by theappearance of allylated woodepoxy resin.

FIGURE 39 Appearance of polyurethane foams from allylated wood, prepared by liquefying and dissolving in 1,6-hexanediol, adding a foaming agent (water) and reacting agent (polymeric MDI), and resinifying.

Wood Plasticization

693

has substantial strength and a restoring force for compression deformation, as can be presumed from Fig. 40. The apparent density of the foams is dependent on the amount of foaming agent as well as the kind of reactive solvent (such as species of polyhydric alcohols, phenols, bisphenols, anddiluents) used. Foams preparedfrom the modified wood solution of bisphenol A using a similar procedure tend to have apparent densities around 0.1 g/cm3 and considerable strength. In order to elucidate the role of the liquefied chemically modified wood within the foams, comparative experiments of preparing the foams without the chemically modified wood have also been conducted. It was found that foaming actually occurs during the resinifying processbut, immediately after that, a contraction in volume of the foam occurs, resulting in resin moldings with apparent densities around 0.2 g/cm3 with little foamingcell structure remaining. This result is understandable because the open-mold, one-shot process used here is said to require rather high-molecular-weight polyhydric alcohols as one raw material of the foam. This result also reveals that the liquefied chemically modified wood plays a positive role in maintaining the shape of the foam during formation. One other application of the phenomenon of liquefying modified wood is in the formation of filaments or fibers produced by spinning methods. Tsujimoto [39] prepared wood-based fibers from acetylated wood. Acetylated wood is first liquefied and dissolved in phenol, and hexamethylenetetramine is added to the solution in various proportions. This is followed by heating to 150°C to promote addition-condensation, resulting in a resinified solution with high spinnability. From the spinning solution, filaments are spun and hardened in a heating oven at a definite heating rate (maximum temperature 250°C). An example of the filament is shown in Fig. 41. These filaments can be carbonized to give carbon filaments. Carbonization is carried out in an electrically heated furnace at a temperature of 900°C with a heating rate of 5.5"Umin. The strength of carbon filaments is measured according to JISR7601, and tensile strengths up to 120 kgf/mm' has been

FIGURE 40 Deformation and restoring property of foam prepared from allylated wood.

Shiraishi

694

FIGURE 41 Appearance of acetylated wood-phenol resinified filament. obtained, which is comparable to that of pitch carbon fibers of general-purposegrade. This strength value is expected to be increased by improvements in spinning and carbonization methods.

VIII.

LIQUEFACTION OF UNTREATED WOOD

The liquefaction and dissolution of chemically modified woods have been reviewed so far.Morerecently,untreatedwood has also beenfound to liquefyin several organic solvents [3,5-7,33,42-551. For example, after treating at around 250°C for 15-180 min, wood chips and wood meals were liquefied in phenols, bisphenols, alcohols (benzyl alcohol), polyhydric alcohols (1,6-hexanediol, l ,4-butanediol, and glycerin), and hydroxyethers (methyl cellosolve, ethyl cellosolve, diethylene glycol, triethylene glycol, and polyethylene glycol). The liquefaction of untreated wood can also be achieved at a lower temperature of 150°C and at atmospheric pressure inthe presence of an acid catalyst, phenolsulfonic acid, sulfuric acid, phosphoric acid, oxalic acid, and hydrochloric acid having been used [3,57,33,42-551. It is possible to obtain pastelike solutions with a high concentration of wood solute of up to 70%. After liquefaction, the wood components were found to have degraded and became reactive, which will be shown more detail in the section on the liquefaction mechanism for wood. The obtained wooden solute can be used to prepare adhesives and other moldings, opening a practical new field for utilization of wood materials, the details of which will also be explained later. During the liquefaction of wood, especially in the presence of acid catalyst, recondensation of degraded wood components occurs, as has been also observed inthe explosion and autohydrolysis process for wood. Because of the recondensation, it becomes very

Wood Plasticization

695

difficultto obtain a liquid with a large wood concentration, which is often considered undesirable from the viewpoint of biomass utilization. On the other hand, starch is very easy to liquefy even at a very small liquid ratio and catalyst concentration, when a catalyst is even necessary. Based on this background, a combined liquefaction process for wood and starch was proposed as a practical method for preparing large-biomass-content liquids [45]. That is, a stepwise liquefaction procedure in which the wood could be preliquefied alone at a relatively large liquid ratio, followed by the addition and liquefaction of the starch, was proposed. By this procedure, a large-biomass-content liquid was prepared with relatively small unliquefied residue [45]. In order to find an appropriate method for accurately determining the amounts of unliquefied residues, the soluble properties of liquefied wood and starch were investigated using a series of diluent solvents [47]. It was found that the solubility behavior of a liquefied biomass in a certain solvent was a kind of fractionation of the degraded and liquefied biomass components. In most cases, any single solvent could not dissolve all of the liquefied components completely. Several binary solvent mixtures composed of solvents considerably different in polarity were found to be good diluent solvents for liquefied biomasses. These phenomena can be illustrated by consulting previous works on physicochemical properties of binary solvent mixtures. Among several satisfactory binaries, the binary of dioxane and water has been studied in detail and found to be widely suitable for liquefied biomasses prepared in different liquefaction solvents. The range of dioxane/ water mixing ratio usable for the complete dilution of liquefied biomasses was wide enough for practical usage. Especially, a binary with a dioxanelwater composition of 812 was recommended as a universal diluent for liquefied biomass [47]. Phosphoric acid and even oxalic acid were found to be usable as catalysts for the liquefaction of wood [46,48-501. In the latter case, a small amount of hydrochloric acid was used simultaneously. These uses of catalyst were evaluated in connection with the flow properties and reactivities or curing properties of the liquefied wood. In this extension, phenolated woodlphenollformaldehyde co-condensed resins were proposed [53]. Wood was first liquefied in the presence of phenol by using an acid catalyst to produce a phenolated wood, and after the liquefaction, formalin was added to conduct a condensation reaction forconverting the remaining nonreacted phenol into resin components. It was found that this procedure can convert almost all the phenol remaining after liquefaction into resins, and therefore significantly upgrades the practical value of the liquefaction technique. Another advantage of this co-condensation isthat it can greatly improve the thermofluidity of the phenolated wood resins and the mechanical properties of their molded products. The flow temperatures and melt viscosities of the co-condensed resins were much lower than those of the phenolated wood resins. That is, these two properties were more or less similar to those of the conventional Novolac resin, resulting in excellent processability. The flexural properties of the molded products made from the co-condensed resins, although this point should be discussed in the next section, were much higher than those of the phenolated wood and also somewhat superior to those of the conventional Novolac resin [53].

A.

Comments on the Liquefaction of Wood

The term “liquefaction of lignocellulose” has hitherto referred chiefly to those procedures for producing oil from biomass under very severe conditions for conversion 156-581. For example, Appel et al. have converted cellulosics to oil by using homogeneous Na,CO, catalyst in water and a high-boiling-point mixture (anthracene oil, cresol, etc.)at a pressure

696

Shiraishi

of 140-240 atm, with synthesis gas CO/H2 1581. Treatment for 1 h at 300-350°C resulted in a 40-60% yield of benzene solubles(oil) and 95-99% conversion of the starting materials. This type of liquefaction can be more precisely called the oilification of lignocellulosics. The review presents recent progress on lignocellulosic liquefaction under milder treatment conditions, that is, at temperatures of 80-150°C with an acidic catalyst. One special group of chemically modified woods can be dissolved in cresols even at room temperature, as shown previously.

B. Application of the Liquefactionof Untreated Wood Almost the same products have been prepared from liquefaction solutions of untreated wood as from chemically modified wood. For example, resol-type phenolic resin adhesives prepared from five parts wood chips and two parts phenol did not require severe adhesion conditions and were comparable to the corresponding commercial adhesives in gluability. Acceptable waterproof adhesion was attained from the adhesives after gluing wood veneers at 120- 130°C with a hot-pressing time of 0.5 min to l-mm-thick plywood. This adhesion temperature of 120°C is at least 15°C lower than that ordinarily usedwith resol resin adhesives [43]. As a second example, foams can be prepared from untreated wood-polyethylene glycol solutions [59). Both soft and hard types of foams can be produced according to the preparation conditions. The prepared foams had a density of around 0.04 &m3, substantial strength, and strong restoring force against deformation. These results imply that the wood components were not merely blended within the foam bubbles, but also played an important role in maintaining the dimensional stability of the foams. Rigid polyurethane foams from combined liquefaction mixtures of wood and starch were proposed [ 5 I]. In this case. the large-biomass-content polyols were first prepared from a combined liquefaction of wood and starch, and the application of these polyols to the preparation of polyurethane foam was studied. The viscosity of a biomass polyol was influenced greatly by the composition of the biomass. At a constant total biomass content of 50%, an increase i n the wood content (i.e., decrease in starch content) drastically increased the viscosity of the polyol. Rigid polyurethane foams have been prepared successfully from the large-biomass-contcnt polyols. The foams had densities of about 0.03 g/cm', compressive strength of 80- 150 kPa, and elastic moduli of 3- I O MPa, being comparable to those of the conventional rigid polyurethane foams. The biomass composition in a biomass polyol had a significant influence on the properties of the resulting foams. The foams prepared from a biomass polyol containing only liquefied starch showed the greatest compressive strength and elastic modulus, but they were brittle and revealed poor restorability after deformation.Thefoams made from biomass polyols containing both wood and starch had somewhat smaller compressive strength and elastic moduli, but were much more resilient, revealing good balance in overall properties. Water-absorbing polyurethane foams were also prepared [54]. These were prepared from liquefied starch polyols and diphenylmethane diisocyanate (MDI) by using a cellopening foaming surfactant. The liquefied starch polyols were obtained by the liquefaction of starch in the presence of polyethylene glycol-dominant reaction reagents by using sulfuric acid :IS :l catalyst under either a refluxing condition or a reduced-pressure condition. The influences of the liquefaction conditions onthe properties of the liquefied starch polyols were investigated, taking into account the requirements for preparing appropriate polyurethane foam. Feasible formulation for the preparation of the water-absorbing foams

Wood Plasticization

were proposed and theproperties

697

of the foamsobtained

were systematically reported

[541. The third application example is Novolacresin-typemoldings prepared from untreated wood-phenol solutions [60]. After one part wood meal had been liquefied in two parts phenol, the untreated phenols were distilled under reduced pressure. The resulting liquefied and reacted wood-phenol powder could be used directly after wood meal filler and hexamethylene tetramine had been added and hot-pressed at 170-200°C. The flexural strength of the molding was comparable to those made from the commercialNovolac, when the curing temperatures for the former were settled 20°C or somewhat higher than that for the commercial Novolac. In connection with these moldings, it was described in the last part of the previous section that when the free phenol existing within the liquefied phenol solution was subsequently reacted with appropriate amounts of formaldehyde to give co-condensed resin, the thermal fluidity, the curing property of the liquefied wood, as well as the mechanical properties of the molding can be enhanced considerably [ 5 3 ] .Additionally, it was found that the flexural properties of the liquefied wood moldings were enhanced with an increase in the amount of combined phenol within the liquefied wood and became comparable to those of the commercial Novolac when the amounts of combined phenol were greater than 75%. Furthermore, with an increase in the content of wood fillers, the flexural properties of the liquefied wood moldings were enhanced more effectively than was the case for the commercial Novolac molding, showing that liquefied wood resins can gain a greater reinforcedeffect from compounding with wood fillers than did thecommercialNovolac resins. And the greater the amount of combined phenol, the higher was the reinforcing performance of the wood fillers. In addition, water-sorption measurements and SEM observations of the moldings indicated that the liquefied wood resins had much greater hydrophilicity than the Novolac and revealed greater compatibility with wood fillers 1521. Carbon fiber, already described in the section on the application of liquefied solutions from chemically modified wood, could also be prepared from an untreated wood solution, and atensilestrength ofup to 1.2 GPa has been obtained so far. Even better physical properties can be expected with more development 1391.

C.

Liquefaction Mechanism for Wood and Related Compounds

As described above, liquefaction of wood and its application have been developed during the last ten and more years. More recently, considerable studies have been carried out to elucidate the liquefaction mechanism for wood and its model compounds. First, cellobiose was used as the model compound for cellulose, and its liquefaction mechanism i n the presence of polyhydricalcoholor phenol and acatalyticamount of sulfuric acid was studied. As the conclusion, the following were shown: ( 1 ) Liquefaction of polysaccharides in the presence of alcohols or phenol with catalytic amount of sulfuric acid is accomplished via alcoholysis or phenolysis in the glucosidic linkage. (2) During this liquefaction reaction i n the presence of alcohols, the anomeric hydroxyl groups of the reducing end group or that of the free glucose are protonated and alcoholated, resulting in the same glucoside as is yielded by the above alcoholysis. (3) The rate of liquefaction depends on the accessibility of the liquefaction solvent to the polysaccharide. The liquefaction of an amorphouspolysaccharide, such asstarch, is very rapid, whereas that of crystalline cellulose proceeds at a much slower rate. which obeys pseudo-first-order kinetics. (4) The initial product of the liquefaction in the presence of an alcohol or phenol is the corresponding alcohol or phenol glucosides. (5)The reaction between polysaccharide

Shiraishi

698

and phenol is more complicated compared with that between polysaccharide and alcohols, because of the multifunctionality of phenol. As a result, liquefaction products prepared in the presence of phenol tends to convert to higher-molecular-weight substances with increase in the reaction time. On the other hand, the liquefaction mechanism for lignin in the presence of phenol was studied in relatively wider ranges, that is, without and with acidic catalysts [61-631. As the model compound for lignin, guaiacylglycerol-P-guaiacylether (GC) was used and the range of the liquefaction studied was as follows: ( I ) under elevated temperature (200250°C) without catalyst; (2) under an elevated temperature of 200-250°C in the presence of acetic acid (catalyst); (3) under a moderate temperature of 150°C in the presence of acetic acid (catalyst); (4) under a moderate temperature of 150°C in the presence of sulfuric acid (catalyst), which corresponds to the study on the liquefaction of cellobiose described above. Conclusions obtained from this study are as follows: ( 1 ) The liquefaction of GC in the presence of phenol under elevated temperature without catalyst proceeds very rapidly through homolysis, producing coniferyl alcohol radical and guaiacol radical through quinone methide as the initial main intermediate. However, various homolytic cleavages occur, which give various kinds of radical compounds. As the result, considerable compounds are produced through reaction among these radical species, with reactions among coniferyl alcohol radical, guaiacol radical, and phenoxy or phenyl radicals resulting in dominating reaction pathways. (2) Acetic acid can greatly promote the homolysis reaction of GG, but does not alter the reaction mechanism; that is, in the presence of acetic acid, homolytic cleavage and coupling can occur even at a mild temperature of 150"C, and the resulting reaction products resemble those obtained under elevated temperature without catalysts. (3) Under catalysis with sulfuric acid, GG is first transferred into mainly benzyl cation. Benzyl cation rapidly condenses with phenol to give four condensed products as the initial reaction intermediates, which are no sooner produced thanthey are further subjected to extensivecleavage in their P-0-4 linkages and C&, bondings. The resulting cleaved fragments further react with phenol to form various phenolated products. The characteristic of this liquefaction reaction is heterolytic, that is, ionic, giving relatively small numbers of products when compared with the homolytic reactions [61-63].

IX.

CONCLUSION

The present state of studies on wood plasticization has been briefly reviewed. The scope of the description has been somewhat limited, that is, somewhat focused on works of the author and his colleagues. This is because the study of wood plasticization isnewand immature and sometimes tends to confusion. However, it can be said that this is a new field for chemical processing of wood with high future potential, although more studies need to be undertaken.

REFERENCES

Wood Plasticization

699

N. Shiraishi, in N. Shiraishi, H. Kajita and M. Norimoto (eds.), ReceI?t Research on b%od C I l d Wood-Based Materials, Elsevier Applied Science, London, p. 155 (1993). 6. N. Shiraishi, in Y. Doi (ed.), Handbook on Biodegracluble Plastics, STS,Tokyo, p. 139 (1995). 7. N. Shiraishi and M. Yoshioka, in T.Haraguchi, I . Sakata, K. Shimizu, N. Shiraishi, M.Norimoto and G. Meshimha (eds.), Handbook on Wooden New Muterials, Gihoudou, Tokyo, pp. 45, 62, 72, 79, 81, 118, 152 (1996). 8. D. A. 1. Goring, Pulp Paper M a g . C m . , 64: T-517(1963). 9. S.-Z. Chow, Wood and Fiber; 3 : 166(1972). IO. S. H. Baldwinand D.A.I. Goring, Svensk Paperstid., 71: 641 (1968). 1 I . N. Shiraishi, T. Matsunaga, and T. Yokota, J . Appl. Polynwr Sci., 24: 2361 (1979). 12. H. Fnakoshi, N. Shiraishi, M. Norimoto, T. Aoki, S. Hayashi. and T. Yokota. Holiforsch.. 33: 159 ( 1979). 13. N. Shiraishi, T.Aoki, M. Norimoto,and M. Okumura, in D.N.-S.Hon(ed.), Grq? rnerizatiorl of Lignocellulosic Fibers, ACS Syrnp. Series, 187, AmericanChemical Society, Washington, DC, p. 32 I (1982). 14. N. Shiraishi, T. Aoki, M. Norimoto, and M. Okumura, Cltemteclz, p. 366 (June 1983). 15. M. Norimoto, T. Morooka, T. Aoki, N. Shiraishi, T. Yamada, and F. Tanaka. Wood Res. Tech. Notes. 1 7 181 (1983). 16. N. Shiraishi, H. Matsui, K. Tsubouchi, T. Yokota, and T.Aoki, A b s t ~Pcrpers . Presented u t 31st Nutl. Meeting, J q m n Wood Research Society. Tokyo, p. 262 ( I98 1 ). 17. N. Shiraishi, and Coda, Mokuzoi Kogyo, 39: 329(1984). 18. N. Shiraishi and M.Yoshioka, Sen-i Gakkoishi, 42(6): T-346 (1986). 19. M.Moritaand I . Sakata, J. Apj~l. Polyn~e,: Sci., 3 1 : 831 (1986). 20. M. Morita, K. Shigematsu, and 1. Sakata, Abstr. Papers Prc..sented ut 35th Nntl. Meeting, J q m r l Wood Re.search Society, Tokyo, pp. 215. 2 16 (1985). 21. M. Norimoto. KGK J.. 16(9): 18 (1982). 22. T. Morroka, M. Norimoto, T. Yamada, and N. Shiraishi, Wood Res. Tech. Notes, 17: 75 (1983). 23. N. Shiraishi and K. Shiratsuchi,JapanPatent PublicationUnexamined. 1989-40560. 24. H. Matsudaand M. Ueda, Mokuzcli Gnkktrislti, 31: 903 (1985). 25. H. Matsuda, M. Ueda, K. Murakami, and M. Hara, Mokuzni Gakknishi, 30: 735, 100.3 ( 1984); Mokuzcri Gnkkrrishi, 31: 103, 215, 267, 468 (1985). 26.N. Shiraishi, Y. Murakami, T. Yokota,M. Noritnoto, T. Aoki, 0. Fujioka,and K. Ito, Abstl: Pcrper.s Presrrtted (11 32rld Nrrtl. Meetinsq. J q x r r z Wood Rrserrrdl SocYety. Fukuoka, p. 326 ( 1982). 27. N. Shiraishi, H. Karakane, T. Yokota. M. Norimoto, 0. FLtjioka, and R. Kitamura, A h t , : P C I I J C ~ ~ S ~, p. 156 ( 1983). Preserlted (11 33rd N u t / . Meetir~g.J a p m Wood Resetrrch S o c i c ~ r Kyoto, 28. N. Shiraishi, in C1wrnisrr;y of' Wood Utili-trtiorl. Kyoritsu,Tokyo. p. 294 (1983). 29. N. Shiraishi. in Advmcwl Techticpe and FLrturc. A p p ~ ~ );)ld lWool/ C/lc>l?lic.tr/,y, CMC, Tokyo, p. 271 ( 1983). 30. N. Shiraishi. Sert-i t o K o g y o . 39: 95 (1983). 31. N. Shiraishi, S. Onodcra, M. Ohotani. and T. Masunloto, M o k u z ; Gctkkai,y/t;,3 1 : 418 (1985). 32. N. Shiraishi,Japan PatentPublication Examined. 1988-1992 (appl. June 6. 1986). 33. N. Shiraishi. N. Tsujimoto,and S. Pu. Japan PatentPublication UnexiUl1ined 1986-261358 (appl. May 14, 1985). 34. N. Shiraishi,Japan PatentPublication Unexamincd, l982-2301: lc)82-236(). 35. N. Shiraishi.and H. Kishi. J. App/. Po1ym.r: Sri., 32: 3189 (1986). I. Sakata, Cellulose Chcw~.7 i d t l l o l . . 21: 255 (1987). 36. M. Morita and 37. R. A. Yong, S. Achmadi. and D. Barkalow, Preprints of CELLUCON'84, Cartl.eIlc-Wrexhaln. Walcs, p. 64 ( 1984). 38. H. Kishi, and N. Shiraishi, Mnku:cri Gddkrishi, 32: 520 ( 1986). 3'9. N. Tsujimoto, i n N. Shilaishi. H. Kajita. and M. Noril1loto (e&.), R e c w l t Kc>,sc.trrc/ro l l woo^/ t r d Wr)od-Bcr.sd M ~ t ~ r i ~ r 1 Elsevier .s. Applied Science. London, p. I69 ( 1993). 40. N. Shiraishi, H. Ito, and S. V. Lonikar. J. Wood Chml. pdutol., 7 405 ( I 987). 5.

700 41. 42. 43. 44. 45. 46. 47. 48. 49. 50. 51. 52. 53. 54.

55. 56. 57. 58. 59.

60. 61. 62. 63.

Shiraishi M. Morita, Y. Yatnawaki, M. Shigematsu. and I . Sakata, Mokrrxri Gokkrrishi, 36: 659 (1990). S. Pu. and N. Shiraishi, M o k u x i Gcrkktrishi, 39: 446, 453 (1993). S. Pu. M. Yoshioka, Y. Tanihara, and N. Shiraishi. in C.-Y. Hse and B. Tomita (eds.),Adho.si~~e.s m d Ror~dedWood Proc/wts, Forest Products Society (USA). Madison, WI, p. 344 (1994). S. Pu, and N . Shiraishi, Mokuzcri G~rkktrishi.40: 824 (1994). Y. Yao. M. Yoshioka. and N. Shiraish, Mokrrz(ri Gtlkkniski. 39: 930 ( 1 993). L. Lin, M. Yoshioka, Y. Yao, and N. Shiraish, J. A/?/)/.P d y r ~ r t ~Sci., r 52: 1629 (1094). Y. Yao. M. Yoshioka. and N. Shiraish, Mokrc=.ni Gtrkkaiski, 40: 176 (1994). M. H. Alma. M. Yoshioka, Y. Yao, and N. Shiraishi, M o k ~ t z Gtrkktrishi, ~i 4 / : 1122 (1995). M. H. Alma. M. Yoshioka. Y. Yao, and N. Shiraishi. J. A p p / . Po/yrn~r:Sci.. 61: 675 (1996). M. H. Alma, M. Yoshioka, Y. Yao, and N. Shiraishi. Ho/$~r.sc~I~.. 50: 85 (1996). Y. Yao, M. Yoshioka, and N. Shiraish. Mokrzrri Gnkktrishi, 41: h59 (1995). L. Lin. M. Yoshioka, Y. Yao. and N. Shiraish. J . A/)/?/. Po/yr~rer:Sci.. 55: 1563 ( 1995). L. Lin, M. Yoshioka, Y. Yao, and N. Shiraish, J . A/>/>/.Po/yrrrtv: Sci.. 58: 1297 ( I 995). Y. Yao, M. Yoshioka. and N. Shiraish. J . AI?/>/.Po/yrrwc Sci., 60: 1939 (1996). M. H. Alma, M. Yoshioka, Y. Yao. and N. Shiraishi, Mokrrztri Gttkkoishi, 4 / : 741 (1995). C. Vanasse. E. Chornet, and R. P. Overend, C m . J . Cl~cwt.Ens., 66: I12 (1988). H. R. Appel. I. Wender. and R. D. Miller, U.S. Bur: Mirles. Ech. Prog. Kcp., 25: S (1969). H . R . Appel. Y. C. Fu, E. G. Illig, F. W. Stetfgen, and R. D. Miller, U.S. H r o : M i r ~ c xHI 801.3, p. 27 ( I 975). N. Shiraishi. in H. Inagaki and G. 0. Phillips (eds.). Ct4lctln.sic.s Ufiliztrfiorl; Htwrrrch t r r r d Rcwnrcls in Ce//rr/o.sic.s.Elsevier Applied Science, London, p. 97 ( 1989). N.Shtraishi, Japan Patent Publication Unexamined. 1989-179483. L. Lin. Y. Yao. M. Yoshioka. and N. Shiraishi, Ho/$)r.sch.. 51: 316 (1997). L. Lin, M. Yoshioka. Y. Yao. and N. Shiraishi. Ho/$orsch.. 51: 325 (1997). L. Lin. M. Yoshioka, Y. Yao, and N.Shimishi, Ho/$orsc./~..51: 333 (1997).

17 Wood-Polymer Composites Hiroshi Mizumachi The University of Tokyo, Tokyo, lapan

1.

INTRODUCTION

In composite systems consisting of polymers and other solid materials, polymer molecules are in close contact with surfaces of solids, and it is quite reasonable to think that there are some interactions between the two materials near the surface. Polymer segments in the vicinity of the surface will be less mobile than those of the bulk polymers beyond the range of influence of the surface. Physical properties of polymers filled with various solid particles have been extensively studied in the past, and aresummarized by Kraus [ l ] , Flory [ 2 ] , and Nielsen [3]. There is much literature showing that carbon black particles in vulcanized natural rubber modify the mobility of the polymer chains as a result of chemical linkages formed between the two materials. This modified polymer layer is estimated to be about SO W i n thickness. Therefore, the modulus and the ultimate strength of the filled rubber are greater than those of the unfilled one. Kraus and Gugone [4] studied the adsorption of elastomers onreinforcing fillers from solutions,analyzedtheadsorptionisothermsaccording to a statistical mechanical theory of polymer adsorption developed by Simha, Frisch, and Eirich [S-71, and obtained valuable information on the polymer-filler interaction responsible for reinforcement. Kwei [S] studied the sorption of water vapor by Ti02-filled and unfilled epoxy polymer, and performed a thermodynamic analysis that showed that polymer segments at a distance less than IS00 A from the surface of the filler are under the influence of the filler. Kambe and Kat0 191 studieddynamicmechanicalproperties of a series of amorphousepoxy resin films filled with homogeneous polyethyl methacrylate (PEMA) beads with different diameters, and showed that larger increases of T,of PEMA are found for smaller particles, and that the loss peak per unit volume of PEMA is lowered with increase in the particle size, thus indicating some interactions between beads and the matrix phase. I t is evident from these several examples that we need detailed knowledge of the interfacial interactions between the two phases if we want to study the physical properties of composite materials. We need to measure the physical or physicochemical properties of the composite matcrials a s a whole, as well as those of the individual constituents, in order to know the properties of the bound polymer near the surface, which reflect the interactions between the component phases. Physical properties such as dimensional stability, bending strength, abrasion, etc., of wood-plastic composites have been widely studied from a practical point of view, as 701

Mizumachi

702

reviewed by Kent et al. [IO], Siau et al. [ l l ] , and Burmester [12]. However. the intermolecular interactions between the two materials through the boundary surface has not been studied in detail from the physicochemical standpoint. In this chapter, some typical studies of the interactions between polymers and woods or wood components are reviewed.

II. DYNAMIC MECHANICAL PROPERTIES OF COMPOSITE SYSTEMS If a polymeric material is subjected to a sinusoidal strain E

= ell exp(iwr)

where is the amplitude of strain, W is the angular frequency of vibration, and r is time, then the stress U also will vary sinusoidally with the same angular frequency W, with the amplitude of v,,, and with a certain phase angle 6, namely, v=

v,, exp i(wr +

6)

The dynamic modulus of the material is expressed as a complex number,

where E' and E" are called storage modulus and loss modulus, respectively. Mechanical energy applied to a material is split into two parts; one is stored in the material as the mechanical energy, which is responsible for the elastic response of the material, and the other is converted to heat as a result of internal friction. The former is proportional to E' and the latter to E". Both E' and E" are functions of frequency and temperature. Typical curves of E' and E" at a constant frequency for an amorphous uncrosslinked polymer are shown schematically in Fig. 1 as a function of temperature. Generally, we find a few dispersions (stepwise drops of E ' ) accompanied by the same number of mechanical absorptions (peaks in E"). Each dispersion and absorption, which must occur simultaneously, correspond to a mode of molecular motion. For example, the primary absorption of a-absorption (or a-dispersion) that appears at the highest temperature is assigned to the initiation of micro-Brownian motion of polymer backbone segments. Around the temperature of this primary absorption, E' of the material changes drastically from IO"' to lo7 dyneskm'. Below the temperature of this absorption, backbone chains of the polymer are frozen. and the material is glassy. However, some kinds of local motions, such as rotation of side groups or local twisting of the chains, can happen at low temperature, and this iswhy some secondary absorptions-P-absorption, y-absorption, and so on. appear. We can get information on molecular motions of polymers from their dynamic mechanical properties measured over a wide temperature range. Therefore, if we measure the dynamic mechanical properties of polymers in the composite systems and compare them with those of the bulk polymers, useful information on the interactions between the component phases can be obtained.

703

Wood-Polymer Composites

log E '

l o g E"

TEMPERATURE FIGURE 1 Schematic representation of E' and E" at constant frequency as a function of temperature for a linear amorphous polymer.

A.

Dynamic Mechanical Propertiesof Wood Impregnated with Polymers

Figure 2 shows E" at 100 H z of Sugi (Cryptomeria Japonica D. Don, early wood, lon[ 131. There are two broad absorption gitudinal direction) as a function of temperature peaks: one around 0°C and the other below -50°C. Mechanisms of these absorptions are not known exactly. However, the change of dynamic mechanical properties of woods is generallyverysmallcomparedwiththoseofsyntheticpolymers.Wearenowmostly

10'

I

G

3.5 .:**tVC"..r-*"

lo8 '

.

-100

-50 0 TEMPERATURE

100

50 "C

FIGURE 2 Dynamic loss moduli of Sugi (earlywood, longitudinal direction) as a function of temperature at various frequencies. Different symbols refer to different specimens in each frequency of measurement.

Mizumachi

704

interested in the interactions between woods and polymers-in other words, in the extent to which molecular motions of polymers are influenced by solid materials. When polymers are impregnated into woods or filter papers by immersing the specimens in polymer solutions, a very thin polymer layer is formed on the very complicated surfaces of the wood or filter paper. These surfaces are porous and have such heterogeneous structures that the effective surface areas for polymer adsorption differ from specimen to specimen. Therefore, it is difficult to control, or even to measure, the thickness of the polymer layer or the surface of the material. Because an interaction between the solid surface and polymer molecules, if any, must be the consequence of a very shortranged molecular force (van der Waals force), it may be predicted that the thinner the polymer layer, the greater the influence of the solid surface. On the contrary, as the polymer layerbecomes thicker, the properties of the layer approach those of the bulk polymer which lies beyond the region of influence of the surface. Figure 3b shows variation of temperature of a-absorption, 7&!&x), with the amount of polymer. Obviously, T(E,',:c,x) of a-absorption of polymer in the composite systems is higher than that of the polymer itself, suggesting that polymer chains on the surfaces of cellulosic materials are somewhat im-

l

-40 0 40 TEMPERATURE

0

5 w t .'lo

80

"C

10 100 of polyrncr

(b)

FIGURE 3 Variation of dynamic loss moduli at 110 Hz with the amount of polymer impregnated: (a) log E" as a function of temperature; (h) T(E:i,,,x)as a function of the amount of polymer. Here a commercial NBR adhesive polymer is used.

Wood-Polymer Composites

705

mobilized, and are falling toward that of the bulk polymer as the amount of polymer increases. But when the polymer content is very small, i.e., below 10 wt% in the case of filter paper, T(E::,:,Jvaries very slightly. In other words, it is substantially constant within the experilnental error a s shown by the upper dashed line in Fig. 3b. When wood specimens are immersed into 5 % polymer solutions and evacuated to remove solvent, the amount of polymers impregnated in the wood generally falls within the region where T (E:;,~!,) is substantially constant. AT(E::,.d,or the difference between of polymer under the influence of the solid surface and that of the bulk polymer, which corresponds to the difference between the two dashed lines in Fig. 3b, will be one of the appropriate parameters to express the degree of interfacial interactions between the two components. AT(E::,,,)is approximately equal to ATv or the difference between T V of polymers in the colnposite and in the bulk phase. Typical data on the dynamic mechanical properties of wood-polymer composites are shown in Figs. 4-7. Data for the respective bulk polymer and for wood are shown for comparison. Generally, in composite systems the drop of E' is smaller and the E" peak shifts to higher temperature than in the respective bulk polymer. Dynamic mechanical properties of the composite materials can be calculated by means of the theory of polymer blends. Takayanagi et al. [ 141 proposed mechanical models like the ones shown in Fig. 8 to describe the viscoelastic behavior of polymer blends in terms of the known properties of the two component polymers. If phase cl is dispersed in phase M', there are two possible equivalent models for the system, model I and model 11. Complex moduli of these models are expressed as follows:

E2' (model I) = (+/[AE$ E* (model 11) = A[@E$

+ (1

+ (1

+ ( 1 - AYE:) ' A)/E,T]" + ( 1 - A)E:

-

A)E:l

where E;: and E: are complex moduli of phase N and phase W , respectively; the values of A and 4 correspond to the thickness fraction and the length fraction, respectively, of phase N as shown in Fig. 8, and the product A4 is equal to the volume fraction ( U , , ) of phase N. Practically, A and 4 are called the parameters of the mixing state, because the values of these parameters naturally vary with the change of mixing state for a system with the same value of U<,. For an actual composite system having separate phases of W and (I, the behavior can be described by either of the models with appropriate values of A and 4. It is experimentally verified that these equations can be applied to many polymer blends when the component polymers from two separate phases in the texture of the system without any interfacial interaction. Examples of numerical calculations for some woodpolymer composite systems according to the above equations are shown in Figs. 9 and 10, where W and N refer to wood and adhesive, respectively. As A approaches 1.0 (both of the models are reduced to the series connection of w and a ) , T(E:;,.,Jincreases somewhat, but at the same time the corresponding E' falls drastically. Because the experimentally observed E' curve shows very little dispersion in every case, a mechanical model with (b approaching 1.0 (the parallel connection of LV and a ) must be employed for these woodpolymer composite systems. Because the elevation of T(E::,L,x) cannot be accounted for, we have to take some interfacial interactions between the two phases into consideration in order to interpret the fact that the experimentally observed T(E::,J shifts to higher temperature than that of the bulk polymer without any sizable dispersion of E'. It is evident that polymer molecules adsorbed on the surface of cellulosic materials are immobilized by the intermolecular forces at the boundary surface and that the composite material consists of wood texture and a thin layer of the immobilized polymer on the surface, which is constructed almost in parallel combination.The E" peak of the composite materials

I

-.

E'

E'

-0

job

t

z

s'

! E"

10'

-

E'

c

r - -

i #

n?

E"

l

=*

.

<:

.

b .

E"

t

-50

0

TEMPERATURE

50

"C

-100

-50

TEMPERATURE

-100

0

"C

-50

TEMPERATURE

0

50 "C

I00

TEMPERATURE

150 "C

FIGURE 4 E' and E" versus temperature at 110 Hz for the composite system composed of wood (Sugi) and NBR co-polymer (acrylonitrile 50%). Data on the constituent bulk materials are also shown for comparison. 0,composite system; 0 , polymer; @, wood. FIGURE 5 E' and E" versus temperature at 110 Hz for the composite system composed of wood (Sugi) and SBR co-polymer (styrene 23.5%). Data on the constituent bulk materials are also shown for comparison. 0,composite system; 0 , polymer; @, wood. FIGURE 6 E' and E" versus temperature at 110 Hz for the composite system composed of wood (Sugi) and SBR co-polymer (styrene 46.0%). Data on the constituent bulk materials are also shown for comparison. 0,composite system; 0 , polymer; @, wood. FIGURE 7 E' and E" versus temperature at 110 Hz for the composite system composed of wood (Sugi) and polystyrene. Data on the constituent bulk materials are also shown for comparison. 0.composite system; 0 , polymer; @, wood.

707

Wood-Polymer Composites

/ t

I I

Model Il

Model I

FIGURE 8 Mechanical equivalent models for the composite system where phase a is dispersed in phase M' without interfacial interaction between the two phases. A and 4 are the parameters of mixing states, and A4 is equal to the volume fraction of phase a.

IO" .

(a)

(b)

E'

E'

.

J

'O'

-100

-50

TEMPERATURE

0900 0151 IOW C138

0

50 "C

-100 -50 0 TEMPERATURE

50 "C

FIGURE 9 E' and E" versus temperature at 110 Hz for the composite system composed of wood (Sugi) and SBR (styrene 23.5%). The experimentally obtained curves are compared with the calculated ones. Numerical calculations are carried out according to the equations based on (a) model I and (b) model 11, using the properties of the constituent bulk materials, which are also shown in the figure. Solid lines are the calculated curves with different parameters of mixing state.

Mizumachi

708

FIGURE 10 E’ and E” versus temperature at II O Hz for the composite system composed of wood (Sugi) and polystyrene. The experimentally obtained curves are compared with the calculated ones. Numerical calculationsare carried out accordingto the equations based on (a) model 1 and (b) model 11. using the properties of the constituent bulk materials which arc also shown in the figure. Solid lines are the calculated curves with different parameters of mixing state.

becomes broader than that of the bulk polymer, which suggests that there are various kinds of segmental environments in the adsorbed polymer layer, and this broadening cannot be interpreted by any of the above equations. Dependence of AT(E;,;,,)or ATg on the polymer composition is shown in Fig. 11 for co-polymers of styrenehutadiene (SBR) and acrylonitrilehutadiene (NBR). AT(Et,t,x)obviously varies systematically with polymer composition, but unfortunately there is no molecular theory to interpret these characteristics. It is well known that two materials with similar values of solubility parameter (S), which is a square root of the cohesive energy density (C.E.D.), are compatible with each other. The values of 8 for the materials are 8.4

20 15

d

‘0.0

PBD

0.5

1.0 PSt

i 4 0 PBD

65

0

N

(a>

FIGURE 11 Dependence of ATx on polymer composition in wt%: (a) SBR co-polymers; (b) NBR co-polymers. 0, Sugi; 0 , filter paper.

Wood-Polymer Composites

709

(callmol)”’for polybutadiene, 9.1 for polystyrene, 12.5 for polyacrylonitrile, and 15.65 for cellulose; there is a possibility that as the 6 of polymer approaches that of wood (or filter paper), the two materials become more compatible and polymer molecules will be adsorbed more tightly on the solid surfaces, and consequently AT(E::,.,) will become larger. Iyengar et al. 1151 studied the role of adhesive-substrate compatibility in adhesion and measured peel strength of Mylar/adhesive/Mylar systems using various kinds of adhesives. They found that maximum peel strength is obtained when the solubility parameter of the adhesive is close to that of the substrate; in other words, peel strength becomes greater as the two materials become more and more compatible. AT(E::,,,) of a Sugi-polymer system is slightly larger than that of the corresponding filter paper-polymer system, but the general tendency of AT(E::,ctx) variation with polymer composition is almost the samefor the two materials. It is possible to think that Sugi immobilizes the polymer slightly more than filter paper does, but this small difference will notbe significant because the amount of polymer impregnated is different for the two solids and there is a large scatter of experimental points in Fig. 1 1 . Studies of this kind have been advanced also by Tadokoro et al. [ 161, Okumura et al. [ 171, Handa et al. [ 181, and Motohashi and Tomita 1191, who made some interesting observations concerning the dependence of the rheological properties of various wood-polymer composite systems on such factors as polymer loading (the amount of polymer impregnated), the kind of solvent used for impregnation, the molecular weight of polymers, the method of preparation of the composite system (blend or graft), and so on. It is interesting to note that they have double peaks in E” curves of wood-polymercompositesystems when the amount of polymer impregnated is extremely large. Figure 12 illustrates some typical data of the dynamic mechanical properties of wood-PVAc emulsion composites plotted as a function of temperature. One peak which appears on the higher-temperature side is assigned to the initiation of micro-Brownian motion of the immobilized polymer segments in the vicinity of the solid surface. The other peak corresponds to that of the ordinary segments which lie in the bulk phase. It has been repeatedly ascertained in these reports that polymer molecules are immobilized near the surface of cellulosic materials as a result of molecular interactions.

B.

Dynamic Mechanical Propertiesof Polymers Filled with Powders of Wood or WoodComponents

The fact that polymer segments are immobilized by the surfaces of solids also can be ascertained by measuring dynamic mechanical properties of filled and unfilled polymers. Carbon black is well known as an active reinforcing filler for elastomers. When rubbers filled with carbon black are vulcanized with sulfur, the chemical reaction between rubber molecules and the surface of the filler can occur in addition to the crosslinking reaction in the rubber phase, and then some modified polymer layers are formed near the boundary surfaces [2]. Shimbo and Ochi [20] showed that the rubbery plateau that is always found in the case of crosslinked polymers appears in the modulus-versus-temperature curves of polyethylene filled with carbon black above the melting temperature of polyethylene crystals. This means that some three-dimensional network structures are formed in the polyethylenekarbon black system or, in other words, that carbon black is an active filler even for such polymers as polyethylene, which had been believed to be chemically inert. On the other hand, calcium carbonate is a typical inert filler, especially for SBR. There is no evidence of chemical reaction nor of physical interaction between polymer molecules and the surface of the filler during vulcanization. In fact, the modulus of SBR filled with

710

Mizumachi

t

20

o i

60

80

100

Temperature (‘C)

FIGURE 12 E” versus temperature at 110 Hz for the composite system composed of wood (Sugi) and PVAc emulsion with various polymer contents (A, 7.0; B. 15.1; C, 22.1; D, 37.0%).

calcium carbonate varies with the filler content in accordance with the theory of composite systems, where the interfacial interaction between the dispersed phase and the matrix phase is not taken into consideration [2]. Shimbo and Ochi [20]also showed that the modulus of polyethylene filled with calcium carbonate falls drastically above the melting temperature of polyethylene crystals without revealing any rubbery plateau region. We obtain information on the interaction between polymers and fillers by comparing the dynamic mechanical properties of filled polymers with those of the respective unfilled ones. In cases where polymer and fillers are mixed in solvent at room temperature, no chemical reaction is expected to occur at the boundary surface. Therefore, the rubbery plateau region is not observed in the modulus-versus-temperature curves, but the temperature of the absorption peak T(E:;,:,.Jis generally altered if the two components interact with each other. Figure 13 shows the loss modulus (E”) at 100 Hz versus temperature of filled (50 wt%) and unfilled polystyrene [21]. T(E::,;,,)of polystyrene filled with calcium carbonate is almost the same as that of the unfilled polymer, reflecting the inertness of the filler. Strictly speaking, T(E::,itx) of the filled polymer is slightly lower than that of the unfilled polymer in this case. The exact reason is not known, but it may be due to the incorporation of free volume during the process offilm forming. On the other hand, T(E::,rtx) of polystyrene filled with carbon black is the highest of all the composite systems.

711

Wood-Polymer Composites

CaCO, (unfilled) lignin cellulose xylan

I

carbon black

~~

30

50

70

90 110 temperature

130

150

‘C

FIGURE 13 Comparison of the temperature dependence of the loss moduli of polystyrene filled withcalciumcarbonate,carbonblack,cellulose,lignin,and xylan to 50% by weight. Data for unfilled polystyrene are also shown.

This means that carbon black alters the physical properties of polystyrene molecules in the matrix phase to the greatest extent. Similar data for tilled and untilled SBR and polybutadiene are shown in Figs. 14 and 15. Variations of T(E::,:,x) with tiller content for the three polymers are summarized in Figures 16- 18. In every case, calcium carbonate and carbon black play similar roles in the rheological behavior of the tilled polymers. Therefore it would be reasonable to think that the elevation of T(E::,ltJin the compositesystems, as compared with that of the respective bulk polymer, corresponds well to the extent of interfacial interaction between the component phases. Yim et al. [22], who referred to T(tan S,,,,) as “Tx” in their report, showed that the shift in “TK”is directly proportional to the extent of polymer-filler interaction energy estimated from the heats of adsorption of the model compounds of the polymers on the tiller surfaces, which is in agreement with the previously mentioned speculation. The extent of interfacial interaction in polymer-wood component composite systems can now be compared. As the content of wood component tilled in the polymer increases, 7‘(E::rdx) of the composite system increases in every case. This means that the three components are more active than calciumcarbonate, so it is clear thatwood components modify the physical properties of nearby polymer molecules. Fillers of wood components elevate T(.E::,z,x) of polystyrene to a greater extent than that of polybutadiene. This is consistent with the informations obtained from Fig. 11, where AT(,?::,,,) of a series of SBRimpregnated woods are plotted as a function of copolymer composition. Among the three wood components, xylan is slightly more active than the other two in every case.

Mizumachi

712

carbon black

-90

-70

-50

-30 -10

30

l0

"C

temperature

FIGURE 14 Comparison of the temperaturedependence of the loss moduli of SBR (styrene 46.0%) tilled with calcium carbonate, carbon black, cellulose, lignin. and xylan to SO% by weight. Data for unfilled SBR are also shown.

\

carbon black

I -150 -130 -110

-90

-70

temperature

-50 -30

'C

FIGURE 15 Comparison of the temperature dependence of the loss moduli of polybutadiene tilled with calciumcarbonate,carbonblack,cellulose. lignin, and xylan to SO76 by weight. Data for unfilled polybutadiene are also shown.

713

Wood-Polymer Composites

105

0

10

20 30 fillcr content

50

40 W 1. "1.

FIGURE 16 Dependence of T(.Ei:,ax) on filler content for polystyrene. Fillers: U, carbon black; 0 , cellulose; +, lignin: 0 , xylan.

111.

@.

calcium carbonate:

DIELECTRIC PROPERTIESOF COMPOSITE SYSTEMS

The dielectric constant of a material is defined as

"

0

-15

0

10

20

fillcr contont

30

40

50

wt.%

FIGURE 17 Dependence of T(E::,,,,)on filler content for SBR (styrene 46.0%). Fillers: carbonate: U, carbon black; 0 , cellulose; +, lignin; 0 , xylan.

0 , calcium

714

Mizumachi

- 85 V e

-90

:z W

0

Y

t"

-95

-100

10 20 30 filler contont

40

50

W 1. @l.

FIGURE 18 Dependence of T(E::,J on filler content for polybutadiene. Fillers: bonate; a, carbon black; cellulose; +, lignin; 0, xylan.

*,

Q,

calcium car-

where C is the capacitance of a pair of capacitors when the space between them is filled with the material, and C,, is the capacitance of the same arrangement of conductors in vacuo. When a material is placed in an electric field, the centers of gravity of the negative charges are displaced from those of the positive charges-in other words, dipolesare induced instantaneously, and the charge on the conductors becomes greater than that in the case of conductors in vacuo. In addition, if the material contains permanent dipoles, they tend to orient in the direction of the electric field, and the charge will appear on the conductors. However, the orientation of permanent dipoles is associated in general with some elementary processes of the passage over potential energy barriers as a consequence of Brownian motion, and therefore it isnot only time-dependent, but also temperaturedependent. Therefore, E of a material must be expressed as follows: E

=

E(t)

at T

The variation of E ( t ) with application of voltage is analogous to that of compliance J ( t ) with application of stress. If the voltage varies simultaneously with time, then the dielectric constant is expressed as a complex number: &* = &l - is" and the ratio E"/E' is usually called the dielectric loss tangent tan 6,. Electrical energy applied to a material is resolved into two components: one is stored in the material, and the other is dissipated as heat. The former is proportional to E' and the latter to E". Typical curves of E' and E" at a constant temperature are shown schematically as a function of angular frequency in Fig. 19. At infinite frequency, permanent dipoles cannot orient in accordance with the fast sinusoidal variation, and E' is equal to a certain constant value E,, which is related only to the electronic depolarization of the material. As frequency

715

Wood-Polymer Composites

I

c'I

FIGURE 19 Variation of

E'

and

E"

as a function of frequency.

approaches zero, on the other hand, all the permanent dipoles can orient in the direction of the electric field at any moment without delay, and E' approaches a larger limiting value E(,. A stepwise change or dispersion of E' occurs at the frequency L,,,, which is comparable to the time scale within which permanent dipoles move, and simultaneously a peak of E", or absorption, is observed at the same frequency. These characteristics are described mathematically as follows:

If E" is plotted as ordinate against E' as abscissa, which is called a Cole-Cole plot, a circular arc is obtained as shown in Fig. 20. The length of the cord of the circular arc is equal to ( E ( ) - E,) = A s , which is proportional to the concentration of the dipoles contributing to the orientation polarization; PT is equal to the angle between the radii of the arc drawn to the points E( ) and E, from the center of the circle, and the value of P (0 < P < 1 ) represents the degree of broadness of a dielectric absorption curve. The smaller P corresponds to the broader absorption peak. The relaxation time E' CO

- €CO cosec P,

E'

FIGURE 20

Cole-Cole plot.

Mizumachi

716

varies with temperature, according to either the Arrhenius equation or the WLF equation. When E' and E" of polymeric materials are measured over a wide range of temperature and frequency, there appear multiple absorption peaks (a, P, 7, . . .) in many cases, such as are shown in Fig. 2 I . Each absorption peak corresponds well to the dynamic mechanical absorption phenomena, as shown in the same tigure. Dielectric properties and dynamic mechanical properties give us substantially the same information on the molecular motions in the materials, the former being amplified by polar or polarizable groups.

A.

Dielectric Properties of Wood Impregnated with Polymers

Polymethyl methacrylate (PMMA) shows multiple absorption peaks, where a-absorption is attributed to the segmental rotation or translational micro-Brownian motions of backbone chains. The other absorptions are inferred to be related to rotation and/or some modes of motion of ester group, to rotation of methyl group attached to the backbone chain, etc. [23,24]. The a-absorption of PMMA is sometimes overlapped by increase of E'' due to the ionic condition at highertemperature, and hence the P-absorptionpeaksometimes becomes difficult to distinguish from the a-absorption peak. Dielectric properties of wood-PMMA composite systems were studied at room temperature (25°C) by Kitsuta [25] and at lower temperatures, between -70 and +20°C, by Handa et al. [26]. They were interested mainly in the influence of wood on subsidiary

TEMPERATURE

TEMPERATURE

FIGURE 21

Dielectricandmechanicalabsorptions

of polymers.

717

Wood-Polymer Composites

absorptions of PMMA. Typical data by Handa et al. on the dielectric properties of a wood (Buna)-PMMA composite system are shown in Figs. 22 and 23. It is interesting to note that woods impregnated with polymers and those impregnated with monomers and then irradiated have different properties. Figure 24 shows the Cole-Cole plots for control wood and wood-PMMA composite systems at -60°C. The parameters A s and 0 are plotted against polymer content in Fig. 25. In composite systems where polymers are impregnated into woods, A s and 0 are constant and almost the same as the respective parameters of woods. However, in cases where monomers are impregnated into woods and then irradi-

.c,

2.4

-

2.3

-

2.2

-

2.1

2.0

-

-60

(a)

40

-20

0

20

Temperature ( C )

O-O~I

0.04

-

0.03

-

c,

0.02

-

t

-60

(b)

40

-20

Temperature ( C

0

20

1

FIGURE 22 (a) Dielectric constant (E') of PMMA-WPC (injection) system (polymercontent 21.2%) as ;L function of temperature at various frequencies. (b) Dielectric loss factor ( E " ) of PMMAWPC (injection) system (polymer content 21.2%) as a function of temperature at various frequencies. Frequency of measurements: a, 1 10 Hz; b, 300 Hz; c, 1 kHz; d, 3 kHz; e, I O kHz; f, 30 kHz; g , 100 kHz; h, 300 kHz; i, I MHz.

718

Mizumachi

1.7

60

003

-20

0

Temperature

-40

(C)

20

-

l

h

c, 9

002 - c f

bl

B

d

a

OOlL 60

-40

-20

0

20

Temperature ( C 1

FIGURE 23 (a)Dielectric constant ( E ‘ ) of PMMA-WPC(irradiation)system(polymercontent 19.2%) as a function of temperature at various frequencies. (b) Dielectric loss f x t o r ( E ” ) of PMMA19.2%) as a function of temperature at various freWPC (irradiation) system (polymer content quencies. Frequency of measurements: a, 110 Hz; b, 300 Hz: c, 1 kHz; d, 3 kHz: e, I O kHz; f, 30 kHz; g, 100 kHz; h, 300 kHz; i, 1 MHz.

ated, A& decreases as polymer content increases, which indicate that graft co-polymerization onto the methylol group occurs in wood, and as a consequence the concentration of dipoles contributing to the orientation polarization within the temperature and frequency range of this experiment decreases during irradiation. At the same time, the p-value also decreases, indicating the increase of nonuniformity in the environment around the side groups of the polymer.

719

Wood-Polymer Composites

0.031 0

,

,

L

1

0.03

0.031 0

A

7% . ,

0.03 0

I

0.03

0 04-

O 2.0

1.9

1.8

I

A

1.7

€* FIGURE 24 Cole-Cole plot forcontrolwoodandPMMA-WPCsystems at -60°C. X , control wood; 0, PMMA-WPC (injection) system (polymer content 21.2%); A , PMMA-WPC (injection) system (13.3%);0 , PMMA-WPC (irradiation) system (19.2%); A, PMMA-WPC (irradiation) system (9.2%).

0.4 I

0.2

l

20

15

10 0

5

1

l

25

Polymer Content ( % )

FIGURE 25 (a) A E versus polymer content. (b) p versus polymer content. PMMA-WPC (injection) system; 0 . PMMA-WPC (irradiation) system.

x, control wood;

0.

Mizumachi

720

B.

Dielectric Properties of Polymers Filled withWood or Wood Components

Interaction between woods or wood components and polymers can also be studied by measuring the dielectric properties of filled and unfilled polymers, just as in the case of dynamic mechanical studies mentioned earlier, because both properties give us substantially the same information about the molecular motions. Mizumachi and Kamidohzono [27] studied the dielectric properties of polyvinyl acetate (PVAc) tilled with powders of wood components and compared them with those of the unfilled PVAc within the region ofa-absorption of the polymer. Valuesof E" of untilled PVAc are plotted against the logarithm of ,f at several temperatures in Fig. 26. The peak shifts toward the higherfrequency side as the temperature of measurement is raised. It is well known that a single absorption curve is obtained from plots of the normalized losses E"/S::,;,~against the logarithm of reduced frequency ,$'f,,ynx at various temperatures, where .L,,ctx is the frequency corresponding to ateach temperature of measurement [28,29].The superposed normalized dielectric a-absorptioncurve thus obtained fromFig.26 for unfilled PVAc is shown in Fig. 27a. Normalized dielectric loss at frequencyf'is equal to the dielectric relaxation spectrum at T = im-f in a crudeapproximation, and therefore the curve for E ' ' / & ~ ~ versus ,~,~ log(f&,,) represents approximately the distribution of the relaxation times. Similar plots for cellulose-filled and lignin-filled PVAc are shown in Figs. 27b and 27c, respectively, and the superposed normalized dielectric absorption curves for the three systems are summarized in Fig. 28. There are scatters of points in the case of lignin-filled system, but it is obvious from the figure that absorption peaks or distributions of relaxation times of tilled PVAc, especially in the case of a cellulose-filled system, is somewhat broader than that of the unfilled ones. If a polymer film containing more solid particles can be prepared, the peak will become much broader. This phenomenon is in agreement with previous observations on the dynamic mechanical properties of polymer-impregnated woods that the E" peak of polymer in the composite system is broader than that of the respective bulk polymer. This broadening is indicative of the fact that there are various kinds of segmental environments in the boundary polymer layer.

Wood-Polymer Composites

721

i~j,,,~,

The average relaxation time 7~ = generally decreases as the temperature of measurements is raised, and the dependence of r on temperature can be approximately described by an Arrhenius-type equation:

or

where AH* is the apparent activation energy of the relaxation process, R is the gas constant, and T is the absolute temperature. Figure 29 shows the plots of logarithm of ,A,,.L, versus I/T for unfilled and filled PVAc. A straight line is obtained in every case, and AH* can be calculated from the slopes of the straight lines. The values of AH* can be regarded as the energy required for a mole of polymer segment to initiate micro-Brownian motion. A H N thus evaluated for untilled PVAc is 53 kcal/mol, which is in good agreement with the literature values 1301. It is evidence that AH* for filled PVAc is larger than that of the

- 3 - 2 - 1

0

1

2

3

4

log( f / f r n a x )

FIGURE 28 Comparison among the superposednormalizeddielectricabsorptions: PVAc; 0,PVAc lillecl with ccllulosc (40%;): X , PVAc tilled with lignin (40%).

0.

unlilled

722

Mizumachi

5.0

**

. .. ...... *\-.

\

4.0 log f max

3.0 2.0 73 K c a l h o l e 1.o

k 3.00 2.80

3.10

2.90

lo3 I T

FIGURE 29 Plots of logA,,asx versus 1/T for unfilled andfilledPVAc: filled with cellulose (40%); X, PVAc tilled with lignin (40%).

0,

unfilledPVAc;

0,

PVAc

unfilled one, and that the average relaxation time for the former system is longer than that for the latter. This means that the bound polymer that lies within the region of influence of the solid surfaces is less mobile than the polymers in the bulk phase.

IV. VAPOR SORPTION BY POLYMERS FILLEDWITH SOLID POWDERS Sorption of low-molecular-weightsubstances by an amorphouspolymer is irreversible below its glass transition temperature T,, whereas it becomes reversible and hysteresis is no longer observed above T,. Therefore,measurements of vapor sorption must be performed at elevated temperatures if we want to analyze the data according to the principles of thermodynamics. The amount of solvent sorbed by a filled polymer is generally lower than that of the corresponding unfilled polymer at low relative vapor pressure (p&:), but the difference lessens as pIIp: increases, and after a certain critical concentration w I C of the solvent is reached in the polymer phase, filled and unfilled polymers have the same sorptive properties. Kwei [8] noticed the difference in the vapor sorption of filled and unfilled polymers in the range 0 < W , ) v l c and proposed athermodynamiccycleasshown i n Fig. 30, by

Filled polymer

+ Solvent

A Fm

(Wlc )

Free polymer + Solvent (w1~

I Unfilled polymer + Solvent (Wlc I / FIGURE 30 Thermodynamiccycleproposcd by T. K. Kwci 181

Wood-Polymer Composites

723

which the free energy AF,,, of the transformation of filled polymer to the unfilled polymer can be evaluated. Kwei described his thermodynamic cycle and the corresponding equations on a mole basis, but here they are represented on a gram basis. The free energy changes AF and AF' in the thermodynamic cycle can be evaluated by analysis of the vapor sorptions of filled and unfilled polymer, respectively, using the Gibbs-Duhem relation:

AF = w , A F , w,dAF,

+ wzAt',

+ w,dAF2 = 0

where

AF, =

-1($

dAF,

(graphic integration)

and R is the gas constant, T is the absolute temperature, and M ,is the molecular weight of the solvent. From the values of AF and AF' thus obtained, the difference in the free energy of the filled and unfilled polymers can be evaluated as

AF,,, = AF'

-

AF

It is obvious from the thermodynamic cycle that the AF,,, value must be determined independently of the kind of solvent used in the experiment, and therefore we come to know the difference in the free energy of the filled and unfilled polymers through the value of AF,,,.Figure 31 shows the sorption of benzene at 50°C by PVAcIcellulose ( 5 % ) , and unfilled PVAc as a function of p,lpy. At very low p,lp';,there is a slight difference between the sorption of unfilled PVAc and that of the filled one, but the two isotherms coincide after a critical vapor pressure (p,Ip~ = 0.1) is reached. Data in Fig. 31 were applied to the Kwei's thermodynamic cycle and AF,,, is calculated to be 0.018 callg polymer at 50°C. This means that cellulose lowers the free energy ofPVAc molecules near

20

U

S

10

U U

cn E

0 0.00 FIGURE 31

Q05

0.10

Sorption of acetone by filled

0.15 ( 0 ) and

P,/P,O

unfilled

(X)

PVAc a t 50°C.

Mizumachi

724

the surface or, in other words, the latter is slightly stabilized as compared with the same polymer molecules in the bulk. The amount of benzene sorbed at 50°C and 60°C by NBRkellulose (50%) is compared with that of unfilled NBR in Figs. 32 and 33, respectively. In this case, there is no difference between the sorption isotherms of filled and unfilled polymer at both temperatures in the whole range of relative vapor pressure, in spite of the fact that the filler content is much increased as compared with the case of PVAc. This means that cellulose surfaces do not alter the thermodynamic properties of the surrounding NBR molecules in the filled system. Results of the thermodynamic analysis are summarized in Table 1. If the value of AF,,, is evaluated at several temperatures, the enthalpy (AH,,,)and the entropy (AS,,,)components in AF,,, can be separated by the following equation:

AF,,, = AH,,, - TA&, where

Kwei made this analysis on epoxy polymer tilled with TiO?, which has higher energy surfaces than cellulose has, and obtained AH,,, = 1.81 (at 80°C) to 2.47 (at 70°C) cal/g polymer, and AS,,, = 4.9 X 10.' (at 80°C) to 6.7 X (at 70°C) cal/deg/g polymer. He pointed out that these values are about 30 times smaller than the corresponding values for the melting of polyethylene crystal, but it is clear that the positive values of AH,,, and AS,,, indicate that there exist some structures with local ordering of polymer segments in the filled state. From the data in Figs. 32 and 33, AF,,, is zero at SO-6O0C, and therefore AH,,, = 0.0 cal/g polymer and AS,,, = 0.0 cal/deg/g polymer in this temperature range, meaning that there is no difference in enthalpy and entropy between filled and unfilled NBR within the experimental error. As there are no data of AF,,, at different temperatures in the case of the PVAckellulose system, AH,,, and AS,,, cannot be calculated, but it can safely be said that the thermodynamic interaction between cellulose and polymer is greater in the PVAd cellulose system than i n the NBWcellulose system, because AF,,, has a positive value in

V.-

00

02

0.4

Q6

0.8

P,/P,"

FIGURE 32 Sorption o f bcnzcne by lilled

( 0 ) and

untillcd ( X ) PVAc

at

50°C.

725

Wood-Polymer Composites

.

FIGURE 33 Sorption of benzene by tilled

( 0 ) and

untilled ( X ) NUR at 60°C.

the former case at 50°C, while it is zero in the latter case. This trend is consistent with the results obtained in rheological studies on the interaction between wood and polymers: the elevation of T(E;:,;,Jin the PVAc/filter paper system is 10°C, which means that the interfacial interaction between the component phases is greater in the former system.

V.

ADSORPTION OF POLYMERS ONTO SOLID SURFACES FROM SOLUTIONS

When solid powders are mixed with polymer solutions, polymer molecules will be adsorbed on the solid surfaces until adsorption equilibrium is attained. Adsorption behaviors are governed by balance of affinities between polymers and solid surfaces, between solvents and solid surfaces, and between polymers and solvents (i.e., solubility). So, if we can compare the adsorption isotherms with the other conditions being equal, it is possible to obtain information on the interactions between polymers and solid surfaces. It is sometimes pointed out that the adsorption isotherm for polymers in solutions is similar to the Langmuir isotherm or to the Freundlich isotherm within experimental error. For example, Luce et al. [3 1 1 studied the adsorption ofPVAc from benzene solution on filter pulp swollen i n various solvents, and obtained isotherms as shown in Fig. 34.

TABLE 1 ThermodynamicQuantities Polymer PVAc NUR Epoxy resin" "Kwci (1964).

Vapor

Ternperaturc ("C )

AF,,, (cal/g

Filler ( ~ 1 % ) Cellulose (S) Cellulose (SO)

Acetone Uenzenc

S0 S0

0.0 I8

Water

60 70 80

TiO, (14.31)

polymer) 0.00 0.00 0.08 0.1S

Mizumachi

80

FORMAMIDE

METHANOL L

BENZENE 0

1

2

3

4

EQUILIBRIUM CONCENTRATION

5

6

(dl)

FIGURE 34 Typical adsorption isotherms for PVAc from benzene solution onto filter pulps swollen in dimethylsulfoxide (DMSO), formamide, water, methanol, and benzene.

Surface area available for adsorption of polymers can be altered by swelling and solvent exchange prior to adsorption. It is found that the greater the degree of swelling in the pretreatment, the larger is the apparent saturation value of the adsorption. Although adsorption is initially rapid, the rate decreases gradually and it takes a long time to attain equilibrium adsorption, particularly when porous material is used as adsorbent. It has also been shown that the solvent plays an important role in the adsorption process. Poor solvent generally favors adsorption, as shown in Fig. 35, due to the facts that polymer molecules are tightly coiling in the poor solvent and therefore the more efficient packing of polymer molecules on the solid surfaces is realized, and also to the fact that polymer molecules are thermodynamically more stable in the adsorbed state than in the solution state when poor solvents are used. On the contrary, polymer molecules have extended conformations in good solvent, and they are more stable in the solution state than i n the adsorbed state, which iswhy the relative adsorption becomes smalleras the solubility parameter of a solvent comes closer to that of the solid as shown in Fig. 35. Adsorption isotherms in Fig. 34 are apparently conformable to both Langmuir and Freundlich isotherms, as illustrated in Figs. 36 and 37, respectively. Adsorption of polymers onto wood surfaces is usually analyzed according to the Langmuir equations 132-341. However, it is quite reasonable to imagine that polymer molecules are adsorbed on the solid surfaces at ;I number of segments along the molecular chains, with the remainder of the molecules extending into the solution phase. Luce and Robertson have pointed out that although there is some uncertainty about the surface area available for adsorption, the extent of adsorption is such as to suggest that the amount of polymer adsorbed is at least

727

Wood-Polymer Composites

L

2.0

0.5

9.5

9.0

10.0

SOLUBILITY PARAMETER

FIGURE 35 Therole of solvent in polymeradsorption.Solventsarecompared by plotting the relative adsorption from each (benzene= 1 ) as a function of their solubility parameters: 0 , adsorption of PVAc onto filter pulp fromcarbon tetrachloride,ethylacetate,benzene,methyl ethyl ketone, nitrobenzene. dioxane, ethylene chloride, and acetone;0 , adsorption of PVAc onto iron powder from carbon tetrachloride, benzene, chloroform. acetonitrile, and ethylene chloride.

0.4

/

I/

0

METHANOL

2

f-

4

EQUILIBRIUM CONCENTRATION

6

(g/l)

FIGURE 36 Langmuir isotherms for adsorption of PVAc from benzene solution onto filter pulp swollen in various solvents. The ordinate is the equilibrium polymer concentration divided by the corresponding equilibrium adsorption.

Mizumachi

728

WATER

cl

0.1

1

10

EQUILIBRILTM CONCENTRATION (g/ 1) FIGURE 37 Freundlich isotherms for adsorption of PVAc from benzene solution onto filter pulp swollen in various solvents. The plot is logarithmic.

one order greater than could be accommodated if all segments occupied adsorption areas. There must be a close relation between conformation of adsorbed polymer molecules and adsorption isotherm. Simha et al. [6] have developed theories of the adsorption of linear flexible polymers, and derived the following equation:

L/=I

d=.Z

FIGURE 38 Conformation of adsorbed polymer molecules.

LI=3

Wood-Polymer Composites

729

where 8 is the fraction of the surface covered by polymer segments; (v) is the average number of segments adsorbed per molecules as shown schematically in Fig. 38; K is a constant dependent on (v), the heat of adsorption, solvent interaction, molecular weight, and temperature; k , is a measure of the free energy of interaction between adsorbed segments in excess over the interaction in the solution phase; and C is the concentration of polymer in the solution phase that is in adsorption equilibrium. Kraus and Gugone 141 studied the adsorption of elastomers on various carbon blacks in detail, and tried to analyze the adsorption isotherms according to the aforementioned theory. They neglected the excess lateral interactions ( k , = O), and transformed the equation into a form suitable for plotting experimental data:

where LZ isthe adsorption at concentration C, and a, isits saturation value. To fit the experimental data. the left-hand side of the equation was plotted against log C for several assigned values of a,, and the best straight line through the experimental points chosen for the calculation of K and (U). They could fit all the isotherms to the equation, among which some typical examples are shown in Fig. 39, and the values of the parameters are summarized in Table 2. If (v) is equal to unity, the isotherm of Simha et al. is identical with that of Langmuir, but when (v) is larger than unity, amount of polymer adsorbed ( a ) appears to level outa very low concentration and then keeps on increasing gradually toward a, as the concentration increases. This means that the isotherm deviates from the Langmuir equation. Although the isotherms are not very sensitive to the exact choice of parameters, it is interesting to be able to get information on the state of adsorption of polymer molecules through this kind of analysis.

FIGURE 39 Typical plots of isotherms according t o Simha-Frisch-Eirich equation.

TABLE 2

Adsorption of Elastomers o n Carbon Black According to Simha-Frisch-Eirich Equation 4I0F, GR-S

Black

Fr S FR FEF-I FEF-I1 Acetylene EPC HAF Graphon SAF-I SAF-I1 S AF-I I1 (1%. saturation

(4

I .70 2.65 2.45 3.10 2.10 2.40 3.05 1.37 3.30 2.75 3.30

(2

\

0.036 0.067 0.119 0.108 0.169 0.160 0.155 0.140 0.245 0.328 0.354

GR-1-18

Hevea

K

(4

a,

K

(4

a,

K

3.5 1.2 0.8 3.1 1.2 0.7 2.8 0.85 4.2 2.4 4.4

2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2 2.2

0.014 0.0245 0.0435 0.044 0.057 0.064 0.062 0.0845 0.105 0.125 0.125

107 155 88 109

3.8 5.0 8.0 10.0 7.6 13.0 8.0 -

0.07 1 0.128

0.024 0.03 0.054 0.07 0.05 0.09 0.06 -

adsorption. grams per gram black. (v). average number of adsorbed segments per polymer molecule. K. constant. liters per gram.

100 100

199 40 323 250 393

0.144

0.172 0.192 0.095 0.150 -

Wood-Polymer Composites

731

REFERENCES G.Kraus, RubberChem. Techn., 38: 1070(1965). P. J. Flory, Principles of Polymer Chemistry, Cornell University Press, Ithaca, NY (1953). L. E. Nielsen, Mechanical Properties of Polymers and Composites, Marcel Dekker, New York, 1975. 4.G.Krausand J. Gugone, Ind. Eng. Chern., 47: 1809 (1955). 5. H. L. Frisch, R. Simha, and F. R.Eirich, J. Chem.Phys.. 21: 365 (1953). 6. R. Simha, H. L. Frisch, and F. R. Eirich, J. Phys. Chem., 5 7 584 (1954). 7. H. L. Frisch and F. R. Eirich, J. Phys. Chem., 58: 507 (1954). 8. T. K. Kwei, J. Polymer Sci. A, 3: 3229 (1965). 9. H. Kambeand T. Kato, Rep.Prog.PolymerPhys. Japan, 12: 317 (1969). IO. J. A. Kentand N. Nash, Woodworking Digest, 67: 31 (1965). 11. J. F. Siau, J. A. Meyer, and C. Skaar, Forest Prod. J., 15: 162 (1965). 12. A. Burmester, Holz-Zenrralbl., 93: 1191 (1967). 13. H. MizumachiandM.Fujino, Holiforsch., 26:164 (1972). 14. M. Takayanagi, H. Harima,and Y. Iwata, Mem. Fac. Eng. Kyushu Univ., 23: I (1963). 15. Y. lyengar and D. E. Erickson, J. Appl. Polymer Sci., 11: 231 1 (1967). 16. K. Tadokoro, T. Sadoh, and K. Nakato, Mokuzai Gakkaishi, 22: 309 (1976). 17. M. Okumura, N. Shiraishi, T. Sadoh,and T. Yokota, Proc.26th Ann. Meeting Japan Wood Research Society, p. 215 (1976). 18. T. Handa, S. Yoshizawa, M. Fukuoka, M. Suzuki, Y. Ikeda, and M. Saito, Proc. Rheol. Symp. Japan Wood Research Society, p. 3 (1976). 19. K. Motohashiand B. Tomita, Mokuzai Gakkaishi, 26: 87 (1980). 20. M. Shimbo and K. Ochi, Proc. 24th Ann. Meeting Chemical Society of Japan, 4 : 2044 (197 I). 21. H. MizumachiandM. Tsukiji, Holzjlorsch., 34: 122 (1980). 22. A. Yim, R. S. Chahal, and L. E. St. Pierre, J. Colloid Interjke Sci., 43: 583 (1973). 23. Y. Ishidaand K. Yamafuji, Kolloid-Z., 177 97 (1961). 24. T. Tanakaand Y. Ishida, J. Phys.Soc. Japrm, 15: 261 (1960). 25. K. Kitsuta, Mokuzai Kogyo (WoodIndustry), 25: 14 (1970). 26. T. Handa, S. Yoshizawa, andM.Fukuoka, KobunshiRonbunshu, 34: 617 (1977). 27. H.Mizumachiand M. Kamidohzono, Holdorsch., 29: 229 (1975). 28. Y. Ishida, M. Matsuo,and K. Yamafuji, Kolloid-Z., 180: 108 (1962). 29. S. Saito, Kolloid-Z., 189: I16 (1963). 30. H. A. Stuart, Die Physik derHochpolyn~eren, Springer-Verlag, Berlin/Gottingen/Heiderberg ( 1956) 31. J. E. Luce and A. A. Robertson. J. Polymer Sci., SI: 317 (1961). 32. K. Nakato, N. Shiraishi,and K. Yokoo, Zairyo, 16: 839 (1967). 33. N. Shiraishi, Y. Ishimaru, K.Nakato, and T. Yokota. Moku;ui Gakktrishi, 15: 20 (1969). 34. Y. Ishimaru. Mokuzcti Gnkknishi, 22: 22 ( I 976). 1.

2. 3.

This Page Intentionally Left Blank

18 Adhesion and Adhesives Hiroshi Mizumachi The University of Tokyo, Tokyo, lapan

1.

INTRODUCTION

Most of the materials currently used are composites, and they are manufactured by means of adhesives and adhesion techniques. Recently, increasingly diversified and advanced technical requirements have been imposed on the functions of those composite materials, and it is well known that there are many cases where adhesion between the constituents within the composite materials plays an important role. There are various types of performance evaluation tests forcomposite materials, among which tests for evaluating the bond performance may be regarded as the most interesting ones for those engaged in work related to adhesion. We can find numerous material characteristics concerning both adhesives and adherends, and also many factors concerning adhesion processing techniques that are expected to have some effects on adhesion. These factors must interact with each other in a complicated manner, resulting in development of various material functions, including a necessary level of bond performance. Engineering difficulties would be reduced if all the relations among these factors were clarified theoretically. However, as a matter of a fact, the greater part of them are hidden in a large black box. and a vast number of trials will have to be carried out to develop a new composite material. The present situation is far from one where we are ready to perform successful molecular design for optimum adhesives to produce a composite material with required physical product characteristics, and at the same time we can easily establish the optimum processing procedures. However, it is an important responsibility of adhesion scientists and engineers to continue efforts to accomplish basic or scientific approaches one by one until the final goal is achieved, even if it is a long way off. In this chapter, attention is concentrated on the relation between structure/properties of adhesives and adhesive strength, and adhesion mechanism is discussed rheologically.

II. ADHESIVESTRENGTH We adhere a pair of adherends by applying an adhesive between them, and after adhesion is completed, we have to ascertain to what extent the bonding is satisfactory. In what way do we perform the test? Usually, we measure the breaking strength of the composite system 733

Mizumachi

734

consisting of adherends and adhesives, and therefore it is fundamentally very difficult to define the degree of adhesion scientifically, because failure mode is very complicated and depends on experimental conditions. Failure occurs sometimes in the bulk phase of adhesive, sometimes in the bulk phases of adherends, and sometimes at the interface between the two materials. Sometimes failure occurs both at the interface and in the bulk phases of the two materials in a very complicated manner, as shown schematically in Fig. 1. However. it is the custom of adhesion scientists and engineers to call the breaking strength of thecompositesystem, which is measured according to some standardized methods, “adhesive strength” or “bond strength.” If failure is always of a completely interfacial nature, the adhesive strength is equal to the sum of the intermolecular forces between the two materials at the surface of contact, the value must be calculated according to the principles of physical chemistry or quantum mechanics, and the value must be independent of the method of testing. Actually, perfectly interfacial failure never occurs in practical cases. Even when interfacial failure is observed visually, evidence of traces of cohesive failure is sometimes clarified by means of electron microscopy, or surface analyses such as ESCA, FTIR, etc. If the failure mode is different, the physical meaning of adhesive strength must be different, which makes it quite difficult to analyze this complicated phenomenon on a scientific basis. Now, because the adhesive strength is practically defined as the breaking strength of the composite system of adherends and adhesive, the value is dependent not only on the physicochemical properties of the surface (critical surface energy, surface tension, surface irregularity, etc.), but also on the structure/properties of the two bulk materials (molecular structure, molecular-weight distribution, co-polymer composition, grafthlock, crosslinking, phase structure, morphology near the boundary surfaces, mechanical properties, thermal expansion coefficient, etc.). And it is true that the apparent adhesive strength is different if a different method of testing is employed, even for the same combination of

-

U

m

H-

L

I

m

FIGURE 1 Failure modes of adhesive joints.

735

Adhesion and Adhesives

t

4

00

-F

E 4

t

4

Pee 1

t

Shear

Tens i le

FIGURE 2 Typical types of adhesion tests.

adherend and adhesive. The typical types of adhesion tests most frequently employed are tensile test, shear test, and peel test, which are shown schematically in Fig. 2. Also, some types of tack tests are employed for pressure-sensitive adhesives. It is interesting to note that adhesive shear strength of a steel/epoxy residsteel system is usually 200-300 kg/cm’, while peel strength of the same combination of the materials is only 0-2 kg/25 mm. Although the two quantities have different dimensions and they cannot be compared in a simple manner, it is a kind of common sense that when adhesive shear strength is high for an adhesive, peel strength is low for the same adhesive, and vice versa. Adhesive tensile strength has the same dimension as shear strength, namely kg/cm’, but the two quantities are not necessarily the same. If the testing method is different, the situation is quite different, and therefore, it is absolutely imperative for development of the science and technology of adhesion to clarify what kinds of factors are dominant for each type of adhesive strength. So far, numerous data have already been accumulatedconcerning the relations between adhesive strength and various factors of woods, i.e., specific gravity, moisture content, morphological factors, physical properties, etc., and we can easily find them in handbooks [ 1,2], but few researchers have concentrated their attention on the effects of the structuresand physical properties of adhesives on adhesive strength of wood. This is also a very important problem, and in this chapter the dependence of adhesive tensile strength, shear strength, and peel strength, especially on rheological properties of adhesives, is reviewed. In addition, tack of pressure-sensitive adhesives is also discussed.

111.

DEPENDENCEOF ADHESIVE STRENGTHON STRUCTURES AND PROPERTIES OF ADHESIVES

A.

Adhesive Tensile Strength

In adhesive tensile tests, the external force is applied to the adhesively bonded joints in the direction perpendicular to the adhesive layer. When wood adhesion is concerned, the two pieces of adherends are combined so as to cross their fiber directions as shown in Fig. 3, which is called “cross-lap” joint. Motohashi et al. [3] studied the influence of the chemical properties and dynamic mechanical properties of polyvinyl acetate (PVAc) emulsion adhesives on tensile strength for cross lap wood joints. Many kinds of PVAc emulsions were prepared, using a redox system of tartaric acid-hydroperoxide as initiator and partially saponified polyvinyl alcohol (PVA) (degree of polymerization 1500, degree of saponification 86.5-89%) as emulsifier.

736

Mizumachi

FIGURE 3 Cross-lap wood joints.

The molecular weight of PVAc and the grafting efficiency or the amount of grafting of VAC monomer on PVA vary largely according to the differences of reaction conditions, some of which are listed in Table 1 . Figure 4 shows that there is no significant relation between adhesive tensile strength of cross-lap wood joints at room temperature and the molecular weight of PVAc. However, the adhesive tensile strength increases significantly asgrafting efficiency increases, as shown in Fig. 5. Adhesive films were prepared by casting the emulsions on Teflon sheets and drying at room temperature, and the ultimate strengths were measured. Obviously, the ultimate strength ofan emulsion film increases linearly with grafting efficiency, as illustrated in Fig. 6, so it is concluded that the adhesive tensile strength of a PVAc emulsion synthesized using PVA as emulsifier has a significant relation to grafting efficiency. This means that adhesive strength is dependent to a great extent on the mechanical properties of the adhesive. Polymeric materials are generally viscoelastic in nature, and physical or mechanical properties are time-dependent as well as temperature-dependent, especially around the glass transition region. Motohashi et al. [4] chose two typical PVAc emulsions, differing in grafting efficiency (see Table 2), measured adhesive tensile strength at a constant speed of crosshead separation as a function of temperature in the range from - 130 to + 140°C, and obtained the results shown in Fig. 7. At extremely low temperature, micro-Brownian motion of polymer segments is frozen and the adhesive is glassy. The tensile strength in this region is almost constant and the value is approximately 10 times as smallas the ultimate strength of the adhesive film. The adhesive tensile strength increases gradually as the temperature is raised, and reaches a maximum near the glass transition temperature of the polymer, where microBrownian motion of backbone chain segments is initiated. At higher temperatures, molecular cohesion of the polymer decreases greatly and accordingly adhesive tensile strength also becomes very low.

TABLE 1

Name of emulsion A- 1 2 3 4 5 6

7 DE- 1 2 3 C- I 2 3 4 5 6 T- 1 2 3 4 EB- I 2

Kinds of PVAc Emulsion and All Experimental Results

HPO amount initial and additional weight (g)

0.08-0.08 0.1-0. I 0.1-0.2 0.3-0.3 0.4-0.4 0.5-0.5 0.7-0.7 0.1-0.1 0.1-0. I 0.1-0.1

0.1-0.1 0.1-0.1

0.1-0.1 0.1-0.1

0.1-0.1 0.3-0.3 0.1-0.1 0.1-0.1 0.1-0.1

0.3-0.3 0.1-0.1 0.1-0.1

Chain transfer reagent" (ml)

-

DE 0.5 DE 1.0 DE 2.0 c 0.5 c 1.0

c 1.s c 2.0 c 3.0

C 0.5 T 1.0 T 2.0 T 3.0 T 1.0 EB 0.5 EB 1.0

Molecular weight of benzene-soluble part

Monomer conversion

Grafting efficiency

(%I

(c/o)

A,, x lo-'

98.3 98.7 97.9 99.1 92.9 98.4 98.1 97.8 96.6 97.0 97.4 97.6 95.5 95.5 95.5 97.4 96.9 97.3 98.8 99.3 98.5 95.0

48.4 54.4 36.6 27.0 22.8 17.1 10.3 55.8 50.7 57.6 33.6 21.5 20.7 15.9 11.9 15.6 40.8 32.7 23.8 15.0 30.6 20.3

6.13 5.12 4.90 3.30 3.4 1 2.42 2.38 5.29 5.91 5.38 3.53 8.15 7.28 6.96 8.01 3.66 5.58 5.50 5.81 3.97 6.15 6.88

"DE. dially ether: C. cumene; T. toluene: EB. ethylbenzene.

A,, X 4.0 I 2.08 2.23 1.51 1.54 1.13 0.8 I3 2.4 1 2.91 2.13 1.75 4.58 3.87 3.3 1 3.05 1.80 4.04 3.92 4.22 2.04 2.98 4.07

Ultimate film strength and standard deviation (MPa)

Bond strength of cross-lap joint and standard deviation (MPa)

37.7 (5.10) 44.3 (4.82) 39.0 (5.33) 35.6 (4.38) 31.3 (4.93) 28.7 (5.82) 2 I .7 (5.25)

4.00 (0.646) 4.10 (0.854) 3.80 (0.526) 3.44 (0.627) 3.46 (0.429) 2.67 (0.474) 2.72 (0.485) 3.91 (0.545) 3.84 (0.581) 3.64 (0.409) 3.53 (0.549) 3.14 (0.408) 2.88 (0.348) 2.73 (0.396) 2.79 (0.402) 2.79 (0.507) 3.27 (0.550) 3.33 (0.409) 3.09 (0.454) 2.77 (0.359) 3.31 (0.406) 2.90 (0.541)

-

30.0 (4.17) 28.7 (5.17) 24.9 (4.80) 30.0 (4.92) 22.0 (6.53) 24.8 (1.12) 33.3 (3.37) 33.2 (2.91) 31.3 (3.70) 27.1 (4.26) 36.0 (5.24) 26.6 (5.05)

738

Mizumachi

0

40,

0 0

A

o

0

n

d 15

e

I

A

Y

A A

&

L, m C

0

2 3.0,

0

L

cn

00

U

V

C

0

m

V

V V V

2.5.

2 .o,*

V

0

Gralling eflicirncy ( 48.4 57.6

A

30.6

0

- 40.8 20.3 -

V

10.3

)

27.0

- 17.1

1

0

J

1

3

2

FIGURE 4 Relation between grafting efficiency.

A?,,

4

S

and adhesive tensile strength of cross-lap joint at each level of

Hatano et al. [5] have investigated the relationship between the adhesive strength and the dynamic mechanical properties as well asthefailure characteristics of epoxy polymers at room temperature. Epikote 828 (epoxy equivalent 189 2 5) and Epikote 871 (epoxy equivalent 41 3) were blended at various ratios and cured with a stoichiometric amount of diethylenetriamine. Asthe weight fraction of Epikote 871 in the Epikote 828/ Epikote 871 blends increased, the temperature of viscoelastic absorption peak T(E:ax)at l10 Hz, which is about 15-20°C higher than the glass transition temperature of the cured polymers, was varied from 135 to 20°C. At the same time the state of the adhesive phase at room temperature changed from glassy (storage modulus E' = 1.5 X 10" dyneslcm2) to rubbery (E' = 10' dyneslcm'), and the ultimate tensile strength of the film decreased monotonically. Especially when Epikote 828 content was rich, the curing reaction did not proceed far enough, which was evident from the fact that there appeared somewhat anomalous behavior in the dynamic mechanical properties of the film at 55°C as shown in Fig. 8. The films were hard but brittle. However, the anomaly disappeared after the same sample was kept for a few hours at elevated temperature, and the film became tougher. In the practical case of wood adhesion, where epoxy resins are used as adhesives, they must be cured at room temperature, although they are usually cured at extremely high temperature in the case of adhesion of metals, etc. Figure 9 shows the datafor adhesive tensile strength of epoxy resins cured at room temperatureas a function of composition. Tensile strength was low either when the adhesive was very hard (Epikote

739

Adhesion and Adhesives

4.O

h

2 I

3.5

Y

f 0 C

2

c

3.0 0 C

0

m 2.5

1.75- 2.41 v 0.81 1.54

-

io

0

20

30

40

Grafting efficiency

50

60

(%l

FIGURE 5 Relation between grafting efficiency and adhesive tensile strength of cross-lap joint at each level of A?,$,.

87 1 -rich region) or when the adhesive was very soft (Epikote 87 1-rich region), and reached a maximum when the adhesive was moderately hard (theintermediate composition of Epikote 828Epikote 87 1 blends). Similar characteristics have also been clarified by Ishii and Yamaguchi [6], who adhered such adherends as steel,aluminum alloys, copper, polycarbonate, etc., with an epoxy resin, and measured adhesive tensile strength as a function of both temperature and deformation velocity. The master curve that they obtained shows a peak around U = lo410' mm/min, and tensile strength became lower either at higher or lower velocity. This is in agreement with the above-mentioned features of adhesive tensile strength.

B. AdhesiveShear Strength In adhesive shear tests, the external force is applied to the adhesion system in the direction parallel to the adhesive layer. Motohashi et al. [4] measured adhesive shear strength of woods joined with the same PVAc emulsions as they had used in the adhesive tensile tests, and obtained the results shown in Fig. 10. Here again, maximum adhesive strength is seen around the glass transition temperature of the adhesive, and it naturally decreases as temperature is raised further, which is a reflection of the decrease of molecular cohesion in the adhesive layer. However, adhesive strength in shear tests at lower temperatures is two or three times larger than in tensile tests in the same temperature range. Figure 11 shows similar data for epoxy

740

Mizumachi

50

LI

40.

5 m C

L

30.

L

m

-.-E

e

20.

nwx 10-5 2.91 m 1.75 v 0.81

- C.58 - 3.31 - 2.41 - 1.54

30

40

a 3.87 4

101

0

20

10

50

-

60

Grafting efficiency (%) FIGURE 6 films.

Relation between graftingefficiency and ultimate tensile strength of PVAc emulsion

resin adhesives, which were obtained by Hatano et al. [ S ] , and the same trends are observed. These characteristics are commonly seen in many types of adhesives. It isvery interesting to note that even in cases where such polymers as polystyrene and polyvinyl chloride, which are not expected to have affinity to solid surfaces, are artificially used as adhesives, adhesive shear strength for wood adherends is about 20-50 kgkm' at low temperature, although that for metal adherends is almost zero. An example is shown in Fig. 12. Of course, there are some cases where specific adhesion at the boundary surface is so strong as to retain a state of intimate contact even when the internal stress due to

TABLE 2

Chemical Properties of PVAc Emulsion Adhesives weight of benzene-soluble parth

Monomer Grafting Molecular conversion efficiency" Emulsion

("/.)

A B

96.2

M,, x 10" 54.4 12.2

4.06 5.12 2.18

M,,,x l 0 20.8 6.52

"Percentage of unextracted PVAc amount with benzene to total PVAc in drled film. hCalculatedfrom gel permeationchromatogram. h?,,, number-averagemolecularweight; molecular weight.

h?,$,

M%? lM,8 2.98 weight-average

741

Adhesion and Adhesives

7. Ernulslon A

6 8 :i,:

5.

I

a 7

FIGURE 7 Temperature dependence of adhesive tensile strength of cross-lap wood joints bonded with two emulsion adhesives.

-100

0 Temperature

200

100 (OC

1

.,

FIGURE 8 Temperature dependence of E' and E" at 1 10 Hz for Epikote 828 cured with dimethylenetriamine. A, A , E ' , E" of the resin cured at 20°C for 5 days; 0, E', ? !,' of the resin cured at 20°C for 5 days, and then at 90°C for 3 h.

742

Mizumachi

i 4

d

1. Epikote87l Epikole828 Weightfraction ~

FIGURE 9 Plot of adhesivetensile strength of cross-lap wood joints bonded with epoxy resin adhesives against weight fraction of Epikote 871 in Epikote 828/Epikote 871 blends. Wood failure is also shown.

the difference of thermal expansion coefficients of materials becomes great. For example, Koizumi and Matsunaga [7] measured adhesive shearstrength of steels bonded with a certain epoxy resin adhesive in the temperature range 20-80°C and in the velocity range 10-3-10" c d s . The master curve that they obtained by reducing the datato 20°C is sigmoid in shape and there is no region where adhesive shear strength decreases as velocity increases within their experimental conditions. Shimbo et al. 181 have studied a variety of epoxy resins and obtained similar results. They found that as the adhesive becomes harder, adhesive shear strength becomes greater, in spite of the fact that internal stress becomes larger. Mizumachi [9,10] has summarized the general features of wood adhesion in relation to the dynamic mechanical properties of adhesives as shown in Fig. 13. When the temperature of the adhesion test is lowered, when the rate of strain is elevated, or when polymers of high glass transition temperature are used as adhesives without plasticizers, adhesive strength for wood adherends has a certain positive value, even when specific affinity at the boundary surface is not expected. This seems to be due to the fact that adhesive anchors formed near the surfaces of woods can resist the external force. This anchor effect or mechanical interlocking effect plays an important role in wood adhesion when the adhesive isin glassy state. In adhesion shear tests, all the anchors formed are effective for adhesive strength; in tensile tests, however, some fraction of the anchors is withdrawn without any resistance, and that is the cause of the difference between adhesive shear strength and tensile strength. This region (E' > 10"' dynes/cm') is

743

Adhesion and Adhesives

I

2

121

.

Emulsion B

ul

0 . 0

0 . - . .1 -150

D

' 0 0, . . - . 1 . . . . ~ . " 1 . ~ - - - ~ l - @ - ~ l

-100

-50 0 50 Temperature ('C)

l00

IS0

FIGURE 10 Temperature dependence of adhesive shear strength of woods bonded with two emulsion adhesives.

called A-region (A refers to Anchors), and failure modes in this region are illustrated schematically in Fig. 14. When the test temperature is raised a little more or the rate of strain is lowered, or when glassy polymers are plasticized so as to make E' of the adhesive between 10" and 10"' (mostly when 5 X 10') dyneskm', adhesion through the boundary surface between adherend and adhesive becomes mostly effective, and in this region both adhesive tensile strength and shear strength for woods as well as for metal adherends become maximum. This region is called the B-region (B refers to Boundary surface), meaning that adhesion at the boundary surface predominates there. When the adhesive becomes extremely soft (E' < 10' dyneskm?), failureoccurs always in the adhesive phase, which means that adhesion at the boundary surface exceeds the cohesive strength of the adhesives. This region is called the C-region (C refers to Cohesion), because the experimentally measured adhesive strength is determined solely by the cohesive strength of the adhesives.

-

C.

Peel Strength

In peel tests, an adherend is largely bent and the external force is applied so as to pull it. Therefore, it is not possible to perform peel tests on wood adherends.

744

Mizurnachi

I

I

1

10 Epikote828 Epikate87I Weight fraction FIGURE 11 Plot of adhesive shear strength of woods bonded with epoxy resin adhesives against weight fraction of Epikote 871 in Epikote 828/Epikote 871 blends. Wood failure is also shown.

Birch/ PSt /Birch

FIGURE 12 Adhesive shear strength of polystyrene as a function of temperature, where Kaba and aluminum ( 0 ) are used as adherends.

(0)

745

Adhesion and Adhesives

5x109

dyne/crn2

B region

-#

E of adhesives

Temperature of adhesion tests

(

Degree of plasticization etc.

FIGURE 13 Dependence of adhesive strengths on storage modulus

*-

m

of adhesives.

n I

S hear

Tensile

FIGURE 14 Comparison of failure modes for shear and tensile tests in the case when the adhesive is in glassy state (A-region).

Mizumachi

746

Hofrichter and McLaren [ 1 l ] measured peel strength of cellophane, using terpolymers of vinyl acetatehinyl chloride/maleic acid as adhesives, and obtained the following empirical relation:

P = k[-COOH]" where k and n are constants ( n = 0.5-0.75). They believed that as the concentration of the "COOH group increases in the adhesive, the probability of interaction between the functional group and " O H group on the surface of the cellophane increases, and consequently peel strength also increases. However, if cohesive failure must be partly involved in adhesion tests, their discussion does not seem reasonable. Mizumachi et al. [ 121 studied the influence of chemical structures and dynamic mechanical properties of a series of methacrylic co-polymers on peel strength of cellophane. Monomers used in the study are listed in Fig. 15. One can easily prepare adhesive polymers with various viscoelastic properties and with various concentrations of functional groups such as hydroxyl groups and epoxy groups by combining these monomers. Monomer mixtures are polymerized in contact with adherend in the glass cell shown in Fig. 16, and after polymerization is completed, a peel test is performed. Some of the typical data are shown. Figure 17 shows the cases of co-polymers of 2-ethylhexyl methacrylate (EH) and 2-hydroxyethyl methacrylate (HO) or glycidyl methacrylate (G). It is true that

Methyl methacrylate (M)

CHt'

Ethyl methacrylate

(€1

F C.0

Butyl methacrvlate (B)

cnZ=

?43

c

$90

0 FH2

42M25 Lauryl methacrylate

'4'9'2'5

(L)

2 Ethyl hexyl methacrylate (EH) ~

P F"2 P 2 OH

2 - Hydroxy ethyl methacrylate WO)

FIGURE 15 Monomers.

Glycidyl methacrylate (G)

747

Adhesion and Adhesives

UV light

111 monomers cellophane

FIGURE 1 6 Glass cell used for polymerization and peel test.

1.0

EH

H0

G FIGURE 1 7 Peel strength of cellophane as a function of co-polymer composition for EH/HO and EH/G ( 0 ) systems.

(0)

748

Mizumachi

peel strength increases as concentration of hydroxyl group or epoxy group increases in some range. Hofrichter et al. made their experiments within a very narrow range of concentration (less than 5.7%). However, if the functional group in the adhesive is increased more and more, peel strength goes down. It does not keep on increasing, nor does it level off, but it decreases to almost zero. The left-hand side of the peak corresponds to cohesive failure, and the right-hand side corresponds to interfacial failure. These data show that chemical structure (or functional group) is not the only factor in adhesion. The dynamic mechanical properties of the copolymers were also measured. The temperature of the viscoelastic absorption peak at 100 Hz, which is close to the glass transition temperature, was plotted as a function of co-polymer composition as shown in Fig. 18. These curves will go down if frequency of measurement is lower. At temperatures below these curves, the backbone chains of the co-polymers are frozen in and the polymers are in the glassy state. At temperatures near the curves, micro-Brownian motion is initiated; and at temperatures above the curves, molecular motion will be great and the polymers will be in a rubbery state and then in a fluid state. It is evident that when the adhesive is rubbery or fluid at room temperature, the cohesive failure occurs at room temperature, regardless of the concentration of functional groups, and peel strength is low; and that when the adhesive is glassy, interfacial failure occurs and the peel strength is still low. Peel strength becomes maximum when the viscoelastic absorption peak of the polymer is near room temperature, where the peel test is performed.

120

I

FIGURE 18 Temperature of viscoelastic absorption peak at 100 Hz as a function of co-polymer composition for EH/HO ( 0 ) and EH/G ( 0 ) systems.

749

Adhesion and Adhesives

0

8 0.5 E

0 0

0

.

0

0

1.0

H0 C

FIGURE 19 Peel strength of cellophane as a function of co-polymer composition for E/HO and E/G ( 0 ) systems.

(0)

Figure 19 shows the cases of ethyl methacrylate and the two functional monomers ( H 0 and G ) . These co-polymers are glassy at room temperature over the whole range of co-polymer composition as shown in Fig. 20. Peel strength is almost zero, in spite of the fact that the concentration of functional groups increases to a great extent. Figure 2 1 shows the case of lauryl methacrylate (L) and methyl methacrylate (M), where no functional group is involved and only viscoelastic properties of the copolymers are varied, as shown in Fig. 22. Here again, maximum peel strength is obtained when the temperature of the viscoelastic absorption peak is near room temperature. Master curves of the rheological functions of some copolymers were obtained and compared with the data on peel strength. It was concluded that the viscoelastic properties of adhesives are the dominant factor in adhesion. Peel strength becomes maximum when the viscoelastic absorption peak of the adhesive at room temperature in the test appears at the frequency that corresponds to peel velocity. These characteristics are summarized as follows. The absolute value of the peel strength can be somewhat different if the chemical structure of the copolymerization system is different, but maximum peel strength is always obtained when the storage modulus E‘ of the adhesive is about 10’ dyneskm’ as shown in Fig. 13. If Hofrichter et al. had performed their experiments over a much wider range of concentration, they might have come to somewhat different conclusions.

D.

Tack of Pressure-Sensitive Adhesives

Tackis one of the most important properties of a pressure-sensitive adhesive, and it is measured mainly by two kinds of methods, which are summarized by Johnston [13]; one is the probe tack test and the other is the rolling-ball tack test. Because the former must

750

Mizumachi

120 100

-

0. 0

**.**-

n

u

0

-80

.m

*

-

0. 0"

be-" \o-o

60 h 8

:: W

c "10

20

.

O0oA E

a

A

0.5 '

a

'

a

1.0 H0 G

FIGURE 20 Temperature of viscoelastic absorption peak at 100 Hz as a function of co-polymer composition for E/HO ( 0 ) and E/G ( 0 ) systems.

L FIGURE 21

M Peel strength of cellophane as a function of co-polymer composition for L h 4 system.

751

Adhesion and Adhesives

120 100

’

n

u

-

0

-eo 60 n I 0

: E W

g40

-

/

0

OakL

0.5 ’ ’ ’

*

m

1.o M

FIGURE 22 Temperature of viscoelastic absorption peak at 100 Hz as a function of co-polymer composition for L/M system.

be classified as one of the adhesive tensile tests, only the latter is discussed here. It is believed that rolling motion of a ball on a pressure-sensitive adhesive reflects tackiness of the adhesive, and therefore rolling-ball tests have been employed in many countries for a very long time. In J. Dow’s method of measuring rolling-ball tack, balls are rolled on an inclined surface. A pressure-sensitive adhesive 10 cm in length is placed on a part of the surface. Angle of inclination is 30°, and leading distance is 10 cm. If a ball is too large, it may roll out across the pressure-sensitive adhesive zone and go down farther. Then a little smaller ball is rolled, and so on. If a ball of a certain diameter, namely, 11/32 in., stops within the pressure-sensitive adhesive zone, tack of the pressure-sensitive adhesive is expressed as “ball number 17.” In the PSTC-6 method, a ball of 14/32 in. diameter rolls down on an inclined path and onto a horizontal surface of a pressure-sensitive adhesive. Here, tack of the pressure-sensitive adhesive is expressed in terms of the rollout distance, because the reciprocal of it is considered to be proportional to the tack of the pressuresensitive adhesive. These ways of expressing tack are useful in some practical cases, but the physical meaning of the value is not necessarily clear. If the angle of inclination is different, or if the length of path on a rigid substrate is different, the ball number or the rollout distance might be different for the same pressure-sensitive adhesive. It is hoped that a method can be developed by which tack of pressure-sensitive adhesive can be expressed in terms of significant physical meaning. Mizumachi [ 14,151 pointed out that tack must reasonably be described in terms of the rolling friction coeffi-

Mizumachi

752

cient ( f ) of the pressure-sensitive adhesive, because rolling motion of a ball can be expressed by a set of equations of motion, where f is involved and ,f does not depend on such parameters as the angle of inclination of the surface or the leading distance, but depends only on the physical properties of the materials on which the ball rolls. He solved the equations of motion of a ball, assuming that

f = 40 + 4 l V and derived the following equations: cos a)/(5g4?cos’a)]

(x - x(,),, = [(7R(R sin a X

log[(R sin a

-

(+(,

+ ~ I v , ) c o sa)l(R sin a - 4(,cos a ) ]

+ 7Rv,,/(5g4f COS a ) where (x - x(,),, is the rollout distance of a ball on a pressure-sensitive adhesive, R is the radius of a ball, a is the angle of inclination of the surface, and v,, is the initial velocity of the ball in the pressure-sensitive adhesive zone, which is given as follows: v,, = ( X

-

X(,)”’( 1Og/7R)”’(R sin a - A) cos a)”’

where ( X - X(,) is the leading distance, and.f;, is the rolling friction coefficient of the rigid substrate, which is nearly equal to zero. Urushizaki et al. [ 161 made extensive experiments on rolling motion of a ball on pressure-sensitive adhesives, and showed that the data can be analyzed successfully according to the above equations. A typical example is given in Fig. 23. Values of the two parameters, 4” and 41,are determined so as to minimize the sum of the deviations of experimental data from the theoretical curves. Agreement between the theoretical curves and the experimental data is satisfactory. In this case, +(, = 0.67 cm and = -0.0043 S , regardless of the leading distance or the angle of inclination of the surface.

x-x0 FIGURE 23

(CM 1

Plot of rollout distance against leading distancc

o f a ball

753

Adhesion and Adhesives

“

FIGURE 24 Rollingcylindermethod

Mizumachi and Saito [ 171 also clarified that rolling processes of a ball can be analyzed by the theory over the whole range of velocity, and at the same time rollout distance can also be analyzed by the same theory. In any case, one must perform a rather complicated analysis in order to determine the values of ,f (or 4,)and because velocity of a ball changes at every moment and at the same time ,f varies as a function of velocity. However, Mizumachi [ 14,15,18] point out that it we adopt the pulling cylinder method as illustrated in Fig. 24, we can easily determine the value offwithout any elaborate analysis. Suppose that a cylinder of radius R, length 0,and weight M g is pulled by a force P on a horizontal plane of a pressure-sensitive adhesive at a constant velocity v. Then f of the material is given as

This type of experiment and also the analysis can be easily made, and we do not need any postulate concerning the dependence offon U . If we want to know f as a function of 71, we only have to pull a cylinder at several constant velocities. Naturally, .f‘depends on the physical properties of a pressure-sensitive adhesive, and if the value off is plotted against log 7~ for a viscoelastic material over a very wide range of velocity, it is expected that some curve will be obtained that increases from a relatively low value to a certain maximum and then decreases as the velocity becomes greater. A typical example is given in Fig. 25.

W. RHEOLOGICAL THEORYOF ADHESIVE STRENGTH If we measure stress and strain at break (c/,, c,,)of a viscoelastic material, and plot c, against E/,, a curve of common shape such as given in Fig. 26 is obtained. This is the “failure envelope” of Smith [ 191. Several attempts 120-241 have been made to interpret the failure envelope theoretically, among which Hata’s theory [23] is the simplest. He showed that the failure envelope of viscoelastic materials (adhesives) can be reproduced if we choose a simple mechanical model, a parallel connection of two Maxwell elements, and assume some failure criteria. Parameters concerning the elements in the model are as shown in Fig. 27. Failure starts in the weak point (the tirst Maxwell element), and then the residual part (the second Maxwell element) breaks down, carrying the whole load. The failure at the first Maxwell element may occur if either of the following two conditions is fulfilled: I.

Strain of the spring ( E ~ , reaches ) a certain critical value (cIIC). This corresponds to the region where strain rate ( r ) is high or temperature ( T ) is low.

754

Mizumachi

lr

FIGURE 25 adhesive.

Plot off measured by pulling-cylinder method against log v for a pressure-sensitive

C

-

Strain

FIGURE 26

Schematic representation of a failure envelope of a viscoelastic material.

FIGURE 27

Parallel connection of two Maxwellelements.

Adhesion

755

11. Strain of the dashpot (cl2) reaches a certain critical value to the region where r is low or T is high.

Thiscorresponds

Then the following equations are derived:

Results of the numerical calculations, substituting appropriate values for the parameters,are shown in Fig. 28, where characteristic features of the failureenvelopeare reproduced. It is possible to show according to this theory that the failure point (U!,, E!,) moves counterclockwise along the failure envelope as the temperature is lowered or the strain rate is increased. In the case of failure of an adhesively bonded joint, the same treatment must be possible in the region where cohesive failure occurs. However, it is evident that an additional failure criterion is needed in order to interpret the interfacial failure. Hata [24] postulated the third criterion: 111. Interfacial failure occurs when energy stored in the springs of the model (W) reaches a certain critical value (Wc-):

Y. Hatano and H. Mizumachi [25,26] calculated adhesive strength of various kinds according to the theory, and some of the results are shown. A.

Adhesive Tensile Strength and Adhesive Shear Strength

Theoretical expressions for both adhesive tensile strength and shear strength have already been given above. However, the parameters must be properly chosen, as illustrated in Fig. 29. An example of the numerical calculation of adhesive shear strength is given in Fig. 30. The curves vary naturally if we choose different values for the parameters. For example, if and W,. is larger, curves I and 111 will go up in almost parallel, and if elZC becomes larger, the slope of curve I1 will be greater. And if only the relaxation time 7,= q / E , varies, all the curves will shift along the log U axis.

d

“Q c-

-

””“”

& FIGURE 28

Explanatory diagrams of the model theory for the failure envelope.

Mizumachi

756

l Tens i l e

Shear FIGURE 29

Comparison of shear and tensile deformation: Shear

Strain Strain rate Modulus Viscosity

E

=d h = 7dh G 77

d&/ldt

Tcnsile E = .r/h d&/lrlt = lllh E (= 3G)

77,

Because we have to select the failure mechanism with the lowest m,, at every velocity, the solid line in the figure expresses the overall adhesive shear strength as a function of velocity. According to the rheological principle, larger velocity is equivalent to lower temperature, or vice versa, and therefore we can imagine how adhesive shear strength depends on temperature at a constant velocity. Curves I, 11, and 111 in Fig. 30 correspond to the C-region, B-region, and A-region in Fig. 13, respectively. If we want to take into

Mizumachi

758 10-

il \‘

0-

\’4

g

6-

. U? P

m

Y

4-

h

(L

2-

104

10.1

10.2

100

101

Velocity (crn/s)

FIGURE 32 An example of the results ofthe numerical calculations of peel strength, which is plotted against log 7). Values of the parameters are as follows: b = 1.5 cm, h = 0.025 cm, W,, = 5 X 10’ dyneskm’, E , = 10“’ dyneslcm’, 77, = 10’ poise, E’ = 10“’ dyneslcm’, 77’ = 10’ poise, q I C = 0.02, = 0.005, W , = 10’ ergslcm’. It is assumed that a suddenjumpfromcurve I tocurve 111 occurs at the same strain rate as in the case of shear test.

interesting to notice that the absolute value of P is very low (order of magnitude of 10” kg/1.5 cm), in spite of the fact that the adhesive shear strength for the same model is about 10’ kg/cm?. A second interesting point is that the velocity corresponding to maximum peel strength is much lower than that corresponding to maximum shear strength, which is in good agreement with the previously described fact that adhesive shear strength becomes maximum when E‘ is about 5 X 10’ dyneslcm’, while peel strength becomes maximum when the adhesive is much softer (E’ = lo8 dyneslcm’).

C.Tack

(Rolling Friction Coefficient)

Mizumachi [ 181 developed a rheological theory on rolling friction of pressure-sensitive adhesives in the case of a cylinder of radius R , length b, and weight M g pulled by a force P at a constant velocity 71, using the same model. Parameters of a cylinder are shown in Fig. 33. The corresponding equations are as follows:

FIGURE 33 Cylinder tacktest.

759

Adhesion and Adhesives

v (cmlsl

FIGURE 34 An example of the results of the numerical calculations of,f. which is plotted agianst log V. Values of the parameters are as follows: 17 = 2.0 cm, R = 1 . 0 cm, IT = 0.001 cm, M g = 6.0 X 10' dynes. E , = IO7 dyneskm', r], = IO7 poise. E, = IO7 dyneskm', = 10' poise. E,,' = 3.0, cl'(.= 7.0, W , = 7.0 X IO7 ergs/cm'.

r].

V.

FRACTURE MECHANICS OF ADHESIVE JOINTS

I t has been clarified in the previous section that the values of adhesiveshearstrength, adhesive tensile strength, and ped strength can be calculated theoretically as a function of rate of deformation, according to the mechanical model proposed by Hata 123,241. If we postulate an additional criterion concerning the abrupt transition from cohesive failure to interfacial failure, we can construct a curve having a peak at some rate for each adhesive strength. The rate at which peel strength becomes maximum is lower than that at which both adhesive shear strength and adhesive tensile strength do. The peak in each adhesive

Mizumachi

760

strength shifts toward the higher-rate side as the relaxation time of the viscoelastic material (adhesive) decreases. or vice versa. In other words, for a particular adhesive, the peak shifts toward the higher-rate side when the temperature of measurements is raised, and when we compare the curves for various adhesives at some fixed temperature, an adhesive of lower T, will have its maximum adhesive strength in a higher-rate region. These features are in agreement with most of our experimental observations so far. However, there is a large discrepancy, that is, adhesive tensile strength is calculated to be much greater than adhesive shear strength, whereas the experimentally measured values are generally in the reverse order. In spite of this discrepancy, this simplified theory may be useful in describing qualitatively the complicated phenomena of adhesion, but we have to think seriously of the case of the discrepancy. Of course, the biggest problem is stress concentration. It is postulated in the abovementioned theory that stress and strain are uniform within the specimen, but this is not true especially in case where the modulus of the adhesive is high. When an external force is applied to the adhesive joint, stress is extremely concentrated at the edge or corner of the adhesive layer, and failure of the joint initiates there and propagates along the bond line as the force reaches a critical value. Stress concentration within a material and its fracture are the major objects of the fracture mechanics which has been developed in the fields of solid materials such as glass, ceramics, hard plastics, and others, and has also been applied to the fracture of adhesive joints [27-391. In fracture mechanics, the critical strain energy release rate G,. is an important parameter, a measure of the toughness of the material, and is defined as follows:

G,. =

(2)(g)

where. P,. and h refer to load at failure and width of the specimen, respectively, and ( K / ilA) is a partial differentiation of compliance (C) with respect to crack length ( A ) . G , is the energy required to increase the unit area of the crack surface in the specimen, and is different if the deformation mode is different. There are three typical deformation modes: mode I or tensile-opening mode, mode I1 or in-plane shear mode, and mode I11 or antiplane shear mode, which are illustrated schematically in Fig. 35. Lim et al. [38,393 measured G, for three deformation modes: G,,. for mode I, GIICfor mode 11, and GI,,,. for mode 111, using double-cantilever beam specimens of wood joints by the compliance method. Characteristics of both adhesives and adherends are listed i n Tables 3 and 4, respectively.

Adhesion and Adhesives

AVUT EPOO 1

EP007 EC34569 EsetR PM200 KU224 KU66 1/2 CH18 Y 400 SGA

A-ff 3000DXH

761

Water-based vinyl polymer isocyanates H-3" Epoxy polyamines" Epoxy polyamines" Epoxy polyamines" Epoxy polyamines" Epoxy silicon" Polyurethanes Polyester (polyol) polyisocyanates" Polyacetates Polyacrylates polyacrylates" Polyacrylates polyamines" PoIy(cY-cyano acrylates)" PoIy(a-cyano acrylates)"

7.5 x I O " -54.7 64.6 84 55.5 -51.3 21 46.3 27.3 72.2 -11.3

1.54

X IO" 1.0 x IO"' 1.12 x 10"' 1.16 x IO"'

4.65 X IOx 1.04 x IO"' 1.01 x 10"' 1.08 x IO"' I . 18 x IO"' 6.06

X

10"

4.34

X

10'

1.00 X 1 0 ' 3.42 X IO"

3.03 x I O " 5.81 x IO' 6.23 X IO" 2.99 x 10" 2.64 X I O " 1.13 x IO'' 7.58 x I O "

0.075 0.26 0.0382 0.0305 0.026 0.125 0.06

0.027 0.245 0.096 0. I25

The experimentally obtained GC values for the joints. where the commercially available adhesives are used, are summarized in Table S. It is shown that G,,. < G,,,,. < G,,,., and this tendency is in agreement with that found in the literature 134,361. An interesting G, value and the corresponding conquestion is what kind of correlation exists between in Fig. 36. whereadhesivctensile ventionaladhesivestrength.Ancxampleisshown strength of a series of the joints is plotted against the square root of G,,, because they are obtained by the tests of similar deformation mode, and G,, is proportional to the square of the load at failure. The correlation coefficientin this case is not necessarily high enough, but if well-characterized adhesives are employed, a moresignificantcorrelation will be o n the state of molecular found. It is quite natural to think that the correlation depends motion of the adhesive (i.e., glassy state, transient state, rubbery state, or fluid state), and systematic data are being accumulated 1401. A rheological approach as well as a fracture mechanical approach will be necessary i n order to clarify the mechanism of adhcsion.

TABLE 4

Characteristics o f Adherends Specilic gravity

Moisture content

Young's modulus ( 1 Or kgl/cm')

Adherends

Air

Dry

(%)

E , . E,,

Ell,

Kaba 1 Kahn2

0.6X

0.64 0.78

14.8-16.5 14.9

I .38 1.16

1.34 1.12

0.88

Mizumachi

762 TABLE 5

Critical Strain Energy Release Rates for Modes

I, II, and I11

GllC

Adhesives

G,,

(kgf cm/cm')

G,,,,.

AV UT EP007 EPOO I PM200

0.34 0.28 1.18 0.34 0.19 0.39 0.10 0.20 0.24 0.07 0.24 0.10 0.1 1

2.05 2.77 2.15 2.80 2.57 5.55 I .72 3.7 I 2.93 2.1 1 2.54 9.47 4.06

0.93 0.72 0.52 0.5 I 0.69 2.53 0.47 0.80 0.82 0.76 0.50 0.83 1.36

EsetR EC3569 KU224 KU66 1/2 CH18 Y400 SGA A- a

3000DHX

VI.

GIICGC

6.0 9.9 11.9 8.1 13.3 14.2 16.6 19.0 12.3 30.4 10.8 54.8 38.3

GlllC/Gl,

2.7 2.6 2.9 1.S 3.6 6.5 4.5 4.1 3.4 10.9 2. I 8.6 12.9

CONCLUDING REMARKS

Adhesion involves a variety of factors, among which structures and propertiesof adhesives and/or adherends are most important, because failure of the materials is always involved in adhesion tests. Takayanagi [41] has pointed out that physical properties of a material ( Z ) are generally expressed in terms of two variables, namely, the state of aggregation ( X ) of molecules and the state of molecular motion (Y). The variable X represents such aspects of material ascrystalline/amorphous,orientedhonoriented, and homogeneous/heterogeneous, as well as superstructures. Information related to X can be obtained through studies utilizing electron microscopy, light-scattering observation, and X-ray diffraction measurements. The variable X also includes thermodynamic measurements with respect to compatibility among the components of the material, and phase separation. 1o+"----

.

FIGURE 36 Relationbetweenadhesivetensilestrength various adhesives.

and strainenergyreleaserate.

GI,., for

Adhesion

763

On the other hand, the variable Y represents such aspects as the micro-Brownian motion of molecular segments and some local modes of molecular motion. Information concerning these aspectscan be obtained through such observationsas viscoelasticity, dielectric characterization, and NMR. Let us assume, for example, that we have two materials, both of which are completely amorphous and located at the same point on the X axis. If the micro-Brownian motion of the backbone chains is restricted in one material and nonrestricted in the other-in other words, if they are located at different points on the Y axis-then the two materials will show completely different physical properties. The former will be in a glassy state with modulus E' of the order of IO"' dynes/cm', while the latter will be a rubbery material with E' of about lo7 dynedcm'. If we take two materials, each having a two-phase structure, the overall material characteristics of a system consisting of a continuous phase in a glassy state and a dispersed phase in a rubbery state would differ extremely from those of the other in which the state of each phase is reversed. This indicates that a conclusion with generality cannot always be achieved if relations of the adhesion performance with either of the variables X and Y are investigated separately. And if some data on the practical performance of adhesion can be connected to Z as a function of X and Y , it will enhance the understanding of adhesion phenomena from the viewpoint of polymer science. The so-called adhesion theories and adhesion principles presented in the past may be said to have clarified some theoretical or empirical laws to connect practical adhesion performance to physical properties of materials. Even in the case where a law is only an empirical one, it will not only be helpful for the efficient development of new materials, it will also serve to stimulate theoretical advancement in the future if the law is expressed in terms of Z =,fix,Y ) . In this chapter, adhesive strengths of various kinds are connected to dynamic mechanical properties of an adhesive, and it became evident that there exists some common tendency, which was interpreted by a simplified rheological theory. We have come to know that although adhesion is a very complicated phenomenon, lots of common elementary processes are involved in various aspects of adhesion. If we continue to accumulate data on adhesion systematically, using well-characterized adhesives (molecular-characterized as well as material-characterized adhesives), the mechanism of adhesion will be clarified scientifically in more detail.

REFERENCES 1

2. 3. 4. 5. 6. 7. 8. 9. 10. II. 12. 13. 14.

IS.

E. A. Davies, Adhesion t r n d Arlhesives, Furldnrrlentnls rrnd Prcrctice, Society of the Chemical Industry (1954). Seccyczk~c-Riron to Oqo, Kobunshi Gakkai ( 1 959). K. Motohashi, B. Tomita, and H. Mizumachi, Holdi~rsch..36: 183 (1982). K. Motohashi, B. Tomita, H. Mizumachi, and H. Sakaguchi, Wood Fiber Sci., 16: 72 (1984). Y. Hatano, B. Tomita, and H. Mizumachi, Moklczcri Gakkaishi, 29: 578 (1983). H. Ishii and Y. Yamaguchi, J . Adlwsiorl Soc. Japan, l / : 59 (1975). S. Koizumi and T. Matsunaga, J. Adkesior~Soc. Jtrparl. 6 : 437 (1970). Seccynkrc Hcrrldbook, Nikkan Kogyo ( 1980). H. Mizumachi, Mokuzcri K o g y , 36: 3 (1981). H. Mizumachi, Mokuzai Kogyo, 36: 57 (1981 ). C. F. Hifrichter and A. D. McLaren, / m f . Eng. Chern., 40: 329 (1948). H. Mizurnachi, M. Tsukiji, Y. Konishi, and A. Tsujita, J. Adhesior? Sor. J t p t r r l , 12: 378 (1976). J. Johnston, A d h e s i ~Age, ~ ~ 26: 34 (1983). H . Mizumachi, Zcriry Gijutsu, 2: 72 (1984). H. Mizumachi, J . Adhesiorl Soc. h p m t , 20: 522 (1984).

764

Mizumachi

16. F. Urushizaki, H. Yamaguchi. and H. Mizumachi, J . Atlhe.rior~Soc. J q x r n , 20: 295 (1984). 17. H. Mizurnachi and T. Saito, J. Arlhrsiorl. 20: 83 (1986). 18. H. Mizumachi, J. App/. P o / w w r Sci., 30: 2675 (1985). 19. T. L. Smith, J. P o / y l r r Sci., 32: 99 (1958). 20. R.Sato. K o h r m s / z i . 15: 665 (1966). 21. R. Sato. K o h l r n s h i . 15: 768 (1966). 22. T. Saito. K o b u r ~ s / R ~ iO I I ~ L ~ I I41: S / Z19 L I( ,1984). 23. T. Hata, Z ~ r i t y ,1 3 : 322 ( 1964). 24. T. Hnta, ./. Adlwsiorl Soc. J q x r t l . h': 64 ( 1972). 25. Y. Hatano and H. Mizumachi, Mokuxri Gnkkrtishi. 3.5: 243 ( 1989). 26. H. Mizumachi, Semi t o K o g x o . 42: 33 (1986). 27. S. Mostovoy and E. J. Ripling, J . Appl. P o / y r w r Sci., 15: 641 (1971). 28. S. Mostovoy and E. 3 . Ripling. J. AI?/?/. P o / w v r Sci., 1 5 : 661 ( I 97 I ) . 29. W. D. Bascom, R. L. Cottington. R. L. Jones, and P. Peyser. J . Appl. Polyr~lrrSci.. / Y : 2545 (1975). 30. R. Ebewele, B. River, and J. Koutsky. Wood Fiher; l / : 197 (1979). 31. J. L. Binter. J. L. Rushford, W. S. Rose. D. L. Hunston, and C. K. Riew, J. Ad/w.sior~.13: 3 32. 33. 34. 35. 36. 37. 38. 39. 40. 41.

(1981). M. Takatani, R. Hamada, and H. Sasaki. M o k l r x i Gtrkkaishi, 3 0 : 130 ( 1984). H. Chai, ASTM STP 893. 209 ( 1986). H. Chai. / t i t . ./. Frrrcmrc., 3 7 137 ( 1988). M. B. Ouezdou and A. Chudnovsky. J . A d 1 ~ 4 0 r l25: . I69 ( 1988). K. M. Liechti and T. Freda. J . At/hr.siotl. 28: I45 ( I 989). T. Kobayashi, Y. Hatano, and H. Mizumachi, Mok~r:cri Gtrkkrri.shi, 37: 331 (1991). W. W. Lirn, Y. Hatano. and H. Mizumachi. J. App/. P o / w w r Sci., 52: 967 (1994). W. W. Lim and H. Mizumachi, J. AppL Po!\rtrc>r Sci.. 57: 55 (1995). W. W. Lim. Ph.D. thesis. The University of Tokyo. 1995. M. Takayanagi. Kohtcrd1i, 3 1 : 142 (1982).

Pressure-Sensitive Adhesives and Forest Products Hiroshi Mizumachi The University of Tokyo, Tokyo, lapan

1.

INTRODUCTION

Because pressure-sensitive adhesives (PSA) are used to bond one material to another, they can be regarded as a kind of adhesive, but in industry they are treated substantially as if they were materials which are different from the so-called adhesives other than PSA. The industrial societies are organized separately, and various statistical data are collected separately. If we look at the situation in detail, we come to know that there surely are some differences between PSA and other adhesives. First, other adhesives are converted from liquid state to solid state after they are applied on the surface of the adherends, but PSA is never hardened. Second, other adhesives are usually supplied to the consumer as liquids or solids (hot-melt adhesives), but PSA is supplied as a PSA product in which PSA is coated on some film material. Usually, PSA itself, eitherasa solid or a liquid, is not available commercially. The most abundant PSA products are PSA tapes, which are used in the fields of packaging, office uses, electronics, vehicles, buildings, medicine, etc. The next abundant PSA products are PSA labels or decals, for which no explanation will be needed about the labels. Decals are sometimes called “sticking paints,” meaning that the necessary information is printed on a film and PSA is coated on the back side of the print so that one can stick them anywhere one likes. Decals can be conveniently used instead of coatings. (For example, in many cases decals are applied on the outsides of cars and trains or on the walls and windows of buildings.) Many kinds of PSA products for medical uses have been developed recently. The industrial scale of PSA-related fields in Japan in 1990 is shown in Fig. 1 [ 1,2]. The total scale of the PSA industry was 452,000 million Y (about $3,480 million U S . ) . On the other hand, the total industrial scale of adhesives other than PSA in the same year was 246,000 million Y (about $1,900 million U.S.), according to the statistics of the Industrial Society of Adhesives in Japan. Scale ofthe former approximately doubles that of the latter. The amount of polymers used as PSA is about 10%ofthatused as adhesives other than PSA, but nevertheless PSA has such a large industrial scale not simply because it includes the cost of various substrates such as papers, fabrics, plastic films, metal foils, etc., but because PSA has a tremendous function of combining chemical components of PSA and substrates. Values of polymers in PSA are added greatly. 765

I

Mizumachi

766 60.0%

l

4.0%

-

36.0%

42.0%

INDUSTRIAL PSA TAPES

0 LABEL & DECALS MEDICAL PRODUCTS

FIGURE 1 Market share of products in Japan in 1990. (a) Area of PSA products (%). Total area is about 2,268 million m’. (b) Amount of money (%). Total money is $3,480 million U.S.

There are several kinds of PSA, such as rubber-based PSA, hot-melt PSA, acrylic PSA, silicone-based PSA, and others. Rubber-based PSA includes natural rubber and synthetic rubber, but the former is much more used than the latter. Because rubber alone is not sticky enough as PSA, we have to blend some tackifier resins with the rubber. It must be pointed out that solvents must be used in the manufacturing process of the rubberbased PSA. Polymers mostly used in hot-melt PSA are SIS, SBS (block co-polymers of styrene-isoprene-styrene or styrene-butadiene-styrene) and SEBS (hydrogenated SBS). Tackifier resins must be blended with these block co-polymers, too. No solvent is needed for these co-polymers either in the blending process or in the coating process. This is why these systems are called hot-melt PSA, and this is a great advantage which this type of PSA has. There are many kinds of acrylic PSA because they are produced by combining various acrylic co-monomers (acrylates and/or methacrylates), some of which are made by solution polymerization and others by emulsion polymerization. Everybody thinks that any industrial products must be environmentally friendly, and therefore in the field of PSA, people have tried to reduce solvent-based materials as much as possible, which is the reason why the production of water-based (emulsion) PSA is gradually increasing. Tackifier resins used not to be blended with acrylic co-polymers because it is quite easy to make polymers with any T, and modulus by controlling the kind and composition of co-monomers. Recently, however,there have been many cases where tackifier resins have been introduced in this type of PSA for the purpose of modifying the adhesion properties against polyolefins. Silicone-based PSA is a blend of silicone rubbers and silicone resins. Because surface tension of this series of polymers is low,they easily stick to Teflon, polyimides, silicone rubbers, etc., which are known to be inert to most adhesives. At the same time, this type of PSA is resistant to heat, chemicals, weathering, etc. All these blends are coated on papers, cellophane, fabrics, plastic film, or metal foil to produce PSAtapes, labels, or decals, and theyare abundantly used in industry, offices, in and in homes.

767

Pressure-Sensitive Adhesives

II. PSAAND FOREST PRODUCTS Figure 2 [1,2] shows the market shares of both PSA products (substrate plus PSA) and PSA itself in 1990 in Japan. Natural rubber, which is a forest product, has the biggest share: 43.3% by area of PSA products (tapes and labels) and 53.3% by weight of PSA itself. SBR (random co-polymers of styrene and butadiene), which is the most popular synthetic rubber, is used at about 10% of natural rubber as PSA. Block co-polymers such as SIS, SBS, and SEBS, which are used as hot-melt PSA, are sometimes called thermoplastic rubbers, and there are cases where they are classified as one of the rubber-based PSAs. Anyway, the amount of these block co-polymers used as PSA is much lower than that of natural rubber. Recently, the legal regulations on environmental problems have become more andmore severe, and manyresearchers have been seriously trying to develop some production systems where solvents are not needed. However, according to the anticipation by PSA specialists [3], natural rubber will continue to be used for some period of time, in spite of the fact that we cannot avoid the solvent problem in this system. This may be due not only to the low cost, but to the tremendous accumulation oftechnological data, and also to the fact that this type of PSA has delicate performance which the other PSA do not have. It must be pointed out that there is a great difference between the solvent problem in PSA and that in other adhesives or coatings. In the case of adhesives (other than PSA) or coatings, solvent will vaporize into the air anywhere they are used. On the other hand, in the case of PSA, solvent is used within the plant, andwhen the PSA products are transferred to consumers, the solvent problem does not occur at all. If the solvent is

2500

100

2000

80

E 1500

60

(Y

C 0

.-

c 0

c,

1000

40

500

20

0

Y

0

PSA PRODUCTS

PS ADHESIVES

FIGURE 2 Polymers usedas PSA in Japan in 1990 NR, natural rubber; SBR,random co-polymers of styrene and butadiene; SIS, styrene-isoprene-styrene block co-polymer; Sol.Acryl., solution of acrylic copolymers; Em.Acryl., emulsion of acrylic copolymers.

I

Mizurnachi

768

recovered effectively within the plant as is the case in most of the PSA industry, there is almost no problem. Nevertheless, much effort has been concentrated to developsome manufacturing processes without any solvents, such as emulsion techniques, hot-melt techniques, radiation-cure techniques, and others. As mentioned earlier, the modulus and viscosity of rubber itself are usually too high for PSA, and we have to add some tackifier resins to the system. The most popular tackifier resins which have been used in PSA are rosins and terpene resins. Rosin is harvested from pine trees, and its main component is abietic acid. Terpene resins are polymers of CYpinene, P-pinene, dipentene, etc. Molecular weight of the resins is around 1000. There are many kinds of tackifier resins because both rosin and polyterpenes are chemically modified in various ways. They are hydrogenated, dehydrogenated, dimerized, or polymerized further, esterified with glycol, glycerol, or pentaerythritol. In some cases, they are co-polymerized with phenolic compounds. Most of the resins are solid, although there are some liquid tackifiers. The solids are very brittle, and if mechanical shock is given to them by a hammer, they are tinely divided into very small fragments or powders, but if we blend the resins with natural rubber, the viscosity of the system becomes extremely low, as shown in Fig. 3 [4]. Thisseems to be an anomalousphenomenon. If molecules of natural rubber and tackifier resins share the free volumes within the material, then the viscosity of the blend must vary monotonically with the blend ratio [ 5 ] .How to analyze this phenomenon on a scientific basis is a great problem which has not been

l O’O

I

Q

In

lo9 ‘g v

>. c, .v)

n

0

io8 g

: €

1

5

Y

c,

v

C

Q

Jz

G c

107

3-; Q)

Q)

g

2

2

a

L

0.5

1 o6 -Et Polyethylene

0

0 10 20 30 40 50 60 70 80 90 100 Resin Content (%> FIGURE 3 Viscosity and peel strength of natural rubber-based PSA (a blend of naturalrubber and polyterpene resin Piccolyte S 125).

Pressure-Sensitive

769

TABLE 1 Tackifier Resins I. Resin made from forest products A. Polar compounds 1. Rosin or rosin derivatives (a) Rosin: gum rosin, tall oil resin, wood rosin (b) Modified rosin: hydrogenated rosin, disproportionated rosin, polymerized rosin (c) Rosin ester: rosins or modified rosins esterified with glycol, glycerol, or pentaerythritol 2. Terpene-phenol resin B. Nonpolar compounds 1. Terpene resin: a-pinene resin, ppinene resin, dipentene resin 2. Terpene resin modified by hydrocarbon II. Resin made from petroleum, etc. A. Polymerization resin 1. Petroleum resin: aliphatic resin, cycloaliphatic resin, aromatic resin 2. Cumarone-indene resin 3. Styrenic resin: styrene resin, substituted styrene resin B. Condensation resin 1. Phenolic resin: alkyl phenolic resin, rosin-modified resin 2. Xylene resin

solved yet. Practically,PSA specialists are interested in the blend ratio where the viscosity and modulus are extremely low. There are tackifier resins -which are synthesized from petroleum, besides those which come from forest resources, and all of themare competing in the industrial markets. Table 1 [6] shows the classification of the tackifier resins. Thus, the fact that not only natural rubber, but also such tackifier resins as rosin derivatives and polyterpene derivatives, which originate from forest products, are used abundantly in industry will be attracting the interest of many forest products researchers. In addition, it must be pointed out that a large amount of kraft paper is used as backing material for PSA, as shown in Fig. 4. For example, OPP (oriented polypropylene) film is mostly used inthe United States as the backing material for packaging tapes, but inJapan, kraft paper is used in the same area as the majority of all the film materials. ,Others

(4

,Others

(W

FIGURE4 Materials used as backings of the PSA products. (a) Fraction by area of PSA products. (b) Fraction by money of PSA products.

Mizumachi

770

Of course, such plastic films as polyethylene terephthalate (PET), polyvinyl chloride, etc., are used to a great extent in areas other than packaging.

111.

PRACTICAL PERFORMANCE OF PSA-THREE FUNDAMENTAL PSA PERFORMANCE CHARACTERISTICS

PSA tapes stick easily to any material upon light touch, which is the most important property of PSA. They are not usually expected to show strong adhesive strength, contrary to the case of structural adhesives. In some applications low adhesive strength is a great merit of the products. On the other hand, there are some PSA specialists who are trying to develop PSA products which stick to various adherends like ordinary PSA does, and at the same time can be used in some structural applications. So, there are tremendous numbers of PSA products which are commercially available now. The practically important properties of PSA are sometimes called “the three fundamental PSA performances.” Testing methods are standardized by the JIS (Japan Industrial Standards), ASTM (American Society for Testing Materials), PSTC (Pressure Sensitive Tape Council), and others. Some of the typical testing methods are shown in Figs. 5-7 [ 6 ] .

A.

Adhesion

“Adhesion” expresses a degree of adhesion of PSA under a normal condition, and it is measured by 180” peel test after PSA tape is thoroughly adhered on an adherend. Theoretical approaches have been tried by many PSA researchers [ 7 ] .

B. Tack “Tack” means the instantaneous adhesion of PSA. There are three types of tack test. The first one is called the rolling-ball method, where steel balls are rolled on PSA, and a rollout distance or some related quantities are regarded as a measure of tack of PSA. Mizumachi et al. [8-191 analyzed the rolling motion of a ball on PSA according to the equation of motion of a solid ball, and proposed that the rolling friction coefficient f of PSA must be

1

PSA tape

FIGURE 5

180” peel test.

Pressure-Sensitive Adhesives

771

“ 0 0 Steel balls, diameter increasing in 1 /32” steps

FIGURE 6

J.Dow ball tack test.

taken as a measure of tack. The second method is called the “quick stick” or “loop tack” method, where a looped tape with the sticky surface facing outside is hung in a tube, and upon light touch on an adherend it is pulled away. The measured resistance is a measure of tack. The last test is the “probe tack” test. This is a simulation of the finger test. An end surface of a cylinder touches the PSA surface and moves away. The resistance is regarded as the probe tack of the PSA tape.

C . Cohesion or Holding Power The upper part of a strip of PSA tape is adhered on an upright adherend surface, and a specified load is applied on the lower unbonded part of the tape. Then the PSA tape bears a constant shear stress U(,, and the tape finally falls down in time r),. This is called the holding power. Of course, r , is dependent on such factors as the dimensions of the bonded part and load, which is why they are specified in the standards mentioned above. For example, when we close a corrugated paper box using PSA tape for packaging, the tendency of the box to open is suppressed by the tape, which means that the PSA layer of the tape is continuously bearing the shear stress. It is a very serious problem to know to what extent of stress and how long the PSA tape can resist the shear. This is why this type of testing has been standardized. The three fundamental PSA performances thus obtained depend on the temperature and time scale of the measurements (i.e., contact time, rate of separation, etc.), and therefore the conditions of measurements are specified in detail by some standards, as mentioned above. Because PSAs are viscoelastic materials, the PSA performances are a function of the physical properties of PSA, such as phase structure, viscoelastic properties, T,, viscosity, molecular weight, and so on. Figure 8 [20] shows the dependence of the three fundamental PSA performances on T, of the PSA. Similar empirical relations have been found for other factors as well. PSA specialists are trying to develop a variety of PSA products by taking

Mizumachi

FIGURE 7 Polykenprobetack test: B. backing: A, pressure sensitiveadhesive: W. weight; P, probe: C. carrier; I, insulation: E, electric contacts; CC, collect chuck (probe holder);G, force gauge; D, dashpot; L. lead screw; CL, clutches; T, transmission (multispeed); H, motor; TC. timer and controls.

into account the balance of the three performances which willbe appropriate for the specific applications. For example, in the case of memorandum notes which stick on anything, the necessary conditions are that some degree of tack is needed and at the same time the adhesive strength (peel strength) must not be high so as not to withdraw any fibers from the paper to which it sticks. Similarly, in the case of bandages where human skins are expected to be the adherends, the peel strength must be controlled at a relatively low level in order to avoid damaging the skin. On the other hand, in the case of packaging tape or industrial tapes for semistructural uses (e.g., VHB), both peel strength and holding power must be especially great. If the practical requirements for a variety of applications are clarified quantitatively, we can control the appropriate levels of the three PSA performances by

773

Pressure-Sensitive Adhesives

0 Sheaf e Peel

* Quick lack

-

-3

-

E

.-“Ine

m

8

all

U

-2

g .VI

L L

4

-1

I

-40-30-20-10

FIGURE 8

$

I

0 +l0 +20

Relation between thethree fundamental PSA Performances and 7‘, of PSA

choosing chemical structures, composition, molecular weight, and tackifier resins and their concentrations. There are many empirical guidelines which have been proposed by some well-known specialists [2 l]. However, if we want to understand the phenomena related to PSA on a scientific basis, we are usually confronted with some difficulties because the empirical rules of PSA are expressed in terms of the PSA performances measured within a very narrow range of experimental conditions (temperature, time scale, stress level, etc.), and we cannot analyze them according to the principles of physics, physicochemistry, surface chemistry, and/or rheology. Experimentally evaluated quantities of PSA performances vary systematically in accordance with the change of the experimental condition, and it is recommended that the PSA performances be measured over a wide range of experimental conditions, and the results compared with the physical properties of PSA such as viscoelastic properties, or other structure/property of the materials. Accumulation of such data will result in clarification of the mechanisms of PSA phenomena.

IV.

COMPARISON BETWEEN ACRYLIC PSA AND NATURAL RUBBER-BASED PSA

The PSA performances depend greatly on the viscoelastic properties of the PSA, which may suggest in turn that PSA of the same viscoelastic properties will have the same PSA performances, but actually the PSA performances are delicately different from one PSA to another. Here, the rheological aspects of the PSA performances of acrylic PSA are compared with those of natural rubber-based PSA.

A.

Viscoelastic Properties [22]

Figure 9 shows the temperature dependence of the Young’s modulus of a series of acrylic polymers. The shape of the curve is almost the same for any acrylic polymer, and if the

774

Mizumachi

FIGURE 9 Young’s modulus of acrylic polymers: 4, polymethylacrylate; 0, polyethylacrylate; H, polyisobutylacrylate; 0 , poly-n-butylacrylate; 0 , polyethylhexylacrylate.

curves are shifted along the temperature axis by the difference of T,, all the curves can be superimposed. Similar behavior is seen in a series of co-polymers. This means that the modulus at the rubbery plateau region is higher for a polymer of higher T,q. On the other hand, Figure I O shows similar plots for a natural rubber-based PSA which is a blend of natural rubber and glycerol ester of hydrogenated rosin. Crossing over

FIGURE 10 Young’s modulus of thenatural rubber-based PSA (blends of natural rubberand hydrogenated rosin esterified with glycerol). Content of tackifier resin: f , 0%; 0 , 20%; 0, 40%; 0 , 60%; A, 80%.

Pressure-Sensitive

775

of the curves is seen in the figure, i.e., the modulus at the plateau region is lower for higher T,. These trends are sometimes seen in natural rubber-resin systems.

B. Adhesion or Peel Strength [23,24] Figure I 1 shows peel strength P of an acrylic PSA (a co-polymer of butyl acrylate and acrylic acid, 90/10) as a function of rate of separation v. In the region of very low rate, P increases with increase of U , and cohesive failure is observed. At around a certain v, failure mode abruptly changes from cohesive to interfacial, and P decreases drastically. In the region of very high v, P becomes nearly constant (or increases very slightly), and the value of P there depends on the critical surface tension yc of the adherends. Curves in the figure shift along the log v axis toward the higher-v side as the relaxation time of the PSA becomes shorter (or as the temperature is elevated, or as T, of the PSA is lowered). Figure 12 shows similar plots for natural rubber-based PSA. It is evident that the shape of the curves is quite different from that in Fig. 11. However, there are some similarities: P at very high v depends on yc.of the adherends, and the curve shifts along the rate axis according to the change of relaxation time of the PSA.

C. Tack

or Rolling Friction Coefficient [23,24]

Mizumachi et al. proposed that the rolling friction coefficient f of PSA must be regarded as a measure of tack of the PSA and can be evaluated by a simple method of pulling a cylinder on PSA. Figure 13 shows the plots off of an acrylic PSA (a co-polymer of butyl acrylate and acrylic acid, 90/10) against rate of pulling v. Curves of the plots of P against log U are similar in shape to those o f f against log v. In the region of very low rate, f is an increasing function of U , and cohesive failure is observed there. After v reaches a certain level, failure mode changes from cohesive to interfacial, and f decreases drastically. The curves shift toward the rate axis according to the change of the relaxation time of the PSA.

l

2000

E

$1500 L

p 1000 I

% 500 0

? *

e

r r r

IO-^

10-~

IO-'

v (mfsec) FIGURE 11 Dependence of peel strength P of an acrylic PSA (a co-polymer of butyl acrylate and acrylicacid) on rate of separation. Adherends are: X, polyvinylchloride; 0, nylon 6,6; polyoxymethylene; 0 , polypropylene; 0 , polyethylene; v, polytetrafluoroethylene.

+,

776

Mizumachi

There is a remarkable difference between f and P in the region of extremely high U . It is evident that j ' decreases to almost zero at high U , whereas P becomes constant or slightly increases in the same region. Mizumachi et al. tried to analyze these phenomena by assuming that f is related to both the bonding process and the debonding process at the same time, and that P is related to the debonding process only. Figure 14 shows a plot off against log 71 for natural rubber-based PSA. Here some differences are found in the curves of the two PSA series, and specific dependence on yc is seen for the two systems. It is true that practical performances of any PSA system are dependent upon structure and

10-~

10-3 IO-^ v (mlsec)

IO-'

FIGURE 13 Dependence of rolling friction coefficient f of acrylic PSA (a co-polymer of butyl acrylate and acrylic acid) on rate of pulling the cylinder. Cylinder is madeof: 0 , brass; X, polyvinyl chloride; 0, nylon 6.6; polyoxymethylene; 0,polypropylene; 0 , polyethylene; A, polytetrafluoroethylene.

+,

777

Pressure-Sensitive Adhesives

10 -

2 2 %5-

0'

-10-5 "0-4 "10-3' '

'

'

'10-2

'

-10-1'

v (mlsec)

Dependence of rolling friction coefficient f of natural rubber-based PSA (a blend of naturalrubberandrosin)onrateof pulling the cylinder. Cylinder is made of A, polypropylene; , polyethylene; U, polyoxymethylene;0, nylon 6,6; X , polytetrafluoroethylene. FIGURE 14

+

properties of PSA and substrate materials, but detailed study is needed to clarify some delicate differences between PSA systems based on natural resources and those based on synthetic polymers.

D.

Holding Power or Shear Creep Resistance [23,24]

A strip of PSA tape is adhered on a vertical adherend surface, and a constant load is applied on the end of the tape. Then a bonded part of the tape resists a constant shear stress CT[, due to the load, and at last the tape falls off after a time tl, has passed. Of course, the larger the stress U[), the smaller the time to break l!,,or vice versa. A plot of LT,, against log t,, gives a descending curve. If these curves are obtained at different temperatures, we can construct a master curve according to the time-temperature superposition principle. Figure 15 shows some examples of the master curve for acrylic PSA. It is evident that the higher the T q of acrylic co-polymer, the more the curve shifts toward longer time scale. Figure 16 shows similar curves for natural rubber-based PSA, where the opposite trend is seen: the higher the T,
778

Mizumachi

FIGURE 15 Dependence of shearcreepresistance acrylate and acrylic acid) on stress a,,.

V.

(f,,)

of acrylic PSA (a co-polymer of butyl

CONCLUDING REMARKS

It has so far been mentioned that the industrial scale of PSA is approximately double that of the adhesives other than PSA, that the fractions of forest products such asnatural rubber, rosins, polyterpene, etc., used as PSA components are very large, and that a tremendous amount of paper is also used as backing material for tapes and labels. However, very few researchers in the field of forest products science are interested in PSA. The reason might be that usually PSA products are not used abundantly in the wood industry. But it must be pointed out that the study of PSA is indeed an important part of forest products science, for which the industrial utilization of forest products is the main aim of research.

10’

io3 t, (sec)

lo2

io4

FIGURE 16 Dependence of shear creep resistance natural rubber and rosin) on stress cq).

(t,,)

of naturalrubber-based PSA (a blend of

Pressure-Sensitive Adhesives

779

Especially in the case of natural rubber-based PSA, we have to study the blends of natural rubber and tackifier resins, and it is quite natural to think that the PSA performances depend greatly on the physical or physicochemical properties of the blends, such as miscibility between the components, viscoelastic properties of the blends, surface chemical properties of the materials, and so on. Variation of the chemical structure and composition of gum and resin may have very complicated influences on various aspects of PSA phenomena, but it is ratherdoubtfulwhether the structure/properties of the blends or the components, which originallycamefrom woods, have been fully characterized on the basis of forest products science. Some PSA specialists used to speculate that the reason natural rubber is compatible with most of the tackifiers made from forest products is that they initially existed in the woods. Of course, this explanation is quite emotional rather than scientific, which indicates that there is a lot of room left in the field of PSA for researchers of forest products science. They are expected to approachPSA science from the standpoint of forest products science.

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

12. 13. 14. 15. 16. 17. 18. 19. 20. 21. 22. 23. 24.

K. Fukuzawa, AFERA Conf., Amsterdam (1991). K. Fukuzawa, European Adhesives and Sealants, p. 19 (March 1992). For example, D. Satas, AFERA Conf., Amsterdam (1991). K. Fukuzawa, J . Adhesion Soc. Japan, 6 : 441 (1970). F. Bueche, Physical Properties of Polymers, Wiley, New York (1970). Japan Adhesive Tapes Makers Association, Handbook of PSA (1995). T. Saito, J. Adhesion Soc. Japan, 2 f : 220 (1985). F. Urushizaki, H. Yamaguchi, and H. Mizumachi, J. Adhesion Soc. Jupun, 20: 295 (1984). H. Mizumachi, Matex Technol., 2: 72 (1984). H. Mizumachi, J. Adhesion Soc. Japan, 20: 522 (1984). H. Mizumachi, J. Appl. Polymer Sci., 30: 2675 (1985). F. Urushizaki, H. Yamaguchi, and H. Mizumachi, Yakugaku Zasshi, 106: 491 (1986). H. MizumachiandT. Saito, J . Adhesion, 20: 83 (1986). H. Mizumachi and Y. Hatano, J . Adhesion, 21: 251 (1987). H. Mizumachi and Y. Hatano, J. Appl. Polymer Sci., 3 7 3097 (1989). T. Tsukatani, Y. Hatano, and H. Mizumachi, J. Adhesion. 31: 59 ( I 989). F.Urushizakiand H. Mizumachi, Chem. Pharm. Bull., 3 9 159 (1991). T. Tsukatani, Y. Hatano, and H. Mizumachi, J. Adhesion Soc. Japan, 27: 217 (1991). T. Tsukatani, Y. Hatano, and H. Mizumachi, J. Adhesion Soc. Japan, 2 7 217 (1991). J. W. Hagan, C. B. Mallon, andM. R. Rifi, Adhesives Age, 29: (March 1979). H. Mizumachi, in Adhesion Society of Japan (ed.), Handbook of Adhesion, p. 257 (1996). A. Zosel, in Advances in PSA Technology-2, p. 92 (1992). T. Tsukatani, T. Hata, Y. Hatano, H. Mizumachi, and R. Ramharack, Proc. Eurocoat Congress, Nice, France, p. 319 (1991). T. Hata, T. Tsukatani, and H. Mizumachi, J. Adhesion Soc. Japan, 30: 307 (1994).

This Page Intentionally Left Blank

20 Wood-Inorganic Composites as Prepared by the Sol-Gel Process Shiro Saka Kyoto University, Kyoto, japan

1.

INTRODUCTION

Wood has been used by mankind since ancient times. However, it has some defects as a natural material. A recent trend in wood research is to remove such defects and add the new value to woody materials. A study of wood-inorganic composites is also one of its trials and has been actively carried out to develop new functional woody materials. In a study of wood-inorganic composites, the double-diffusion process of barium phosphate must be mentioned. This has been developed at the Wood Research Institute, Kyoto University, in collaboration with Central Research Laboratory, Matsushita Electric Works Ltd. From a hint that old electrical wood poles buried in the ground cannot be burned out, development of fire-resistant wood has been started, trying to improve dimensional stability and biodeterioration together with fire-resistance [ 1,2]. Different from this process, we have developed a method using a metal alkoxide to prepare wood-inorganic composites by sol-gel process. For this preparation, efforts have been made to add new value to the wood without losing its characteristic properties such as porous structure [3- 131.

II. SOL-GEL PROCESS A basic principle of the sol-gel process involves hydrolysis of the metal alkoxide M(OR),, (M = Si, Ti, Ba, Zr, etc.; R = alkyl group; n = oxidation number) with water and subsequent polycondensation by dehydration or dealcoholation reaction to produce metaloxane sols as described below. M(OR),, + xH,O

+ M(OH),(OR),,_,+ xROH

Dehydration reaction: -M-OH

+ HO-M-

+ -M-O-M-

+ HZ0

Dealcoholation reaction: -M-OH

+ R-0-M-

+

"-0"-

+ ROH 781

Saka

782

The sols formed are then solidified as wet metaloxane gels in a temperature range between 25 and 80°C and then heated to 120°C and dried [14]. In an ordinary sol-gel process, the gelled masses are further heated to 800°C to prepare glasses. However, the gels in this study must be treated under the temperature at which wood is not thermally degraded. Therefore, in a reaction medium of the metal alkoxide/alcohol/acetic acid (catalyst), the moisture-conditioned wood or water-saturated wood was soaked at ambient temperature under reduced pressure or atmospheric pressure. The water present within the wood cells initiates the reaction of the hydrolysis and polycondensationof metal alkoxide.The soaked wood was subsequently treated at a temperature between 50 and 60°C for 24 h, and at 105°C for another 24 h to prepare wood-inorganic composites [3,4]. 111.

MONOCOMPONENTWOOD-INORGANICCOMPOSITES

A.

Distribution of Inorganic Substances inWood Cells

Recent studies have indicated that the distribution of inorganic substances in wood cell walls is dependent on a combination ofthe metal alkoxide species and the water-retaining conditions of the wood specimens.That is, the distribution of inorganic substances formed in the moisture-conditioned specimens is completelydifferent from thatin the watersaturated specimens. Figure 1 shows the types of distribution of inorganic substances. Table 1 summarizes the results of its distribution for different metal alkoxide/alcohol reaction systems. The moisture-conditioned specimens refer to woodwithonlyboundwaterbelow the fiber saturation point; therefore, water is distributed only within the cell walls. Water-saturated specimens have free water in the cell cavities in addition to the bound water within the cell walls. In this way,the use of these specimens with differently distributed water in the cells can make it possible to prepare wood composites with inorganic substances distributed differently [3,4,6]. In type I, the metal alkoxides that form metaloxane gels specifically within the cell walls are silicon alkoxide and boron alkoxide with the moisture-conditioned specimens (Fig. 2a). The use of water-saturated specimens makes the inorganic substances distributed

0

I

II

m

Iv

v

FIGURE 1 Distribution of inorganic substances

in wood-inorganic composites. (From Ref.

6.)

Wood-Inorganic Composites

783

TABLE 1 Inorganic Gels Formed and Their Distribution in Wood Cells as Prepared in Various Metal Alkoxide/Alcohol Reaction Media Types of gels distributed in Fig. 1 Metal alkoxides

Alcohols

Gels

s i (OC2Hd4 Si (OCH3)4 Si (OC3Hd4 Ti (OC3Hd4 Al(OCH(CHA), A1 (OC~HS), Zr (OC2Hd4 B (WHA

EtOH MeOH i-MH i-PrOH i-PrOH EtOH EtOH MeOH

Si02 SiO, SiO, TiO, A1203 A1203

ZrO, B203

Moisture-conditioned specimen

I I I 11, I n In HI I11 I

Water-saturated specimen

Iv

Iv Iv 0 0 0 0 W. v

as type IV, due to the presence of water in the cell cavities in addition to the cell walls (Fig. 2b). In these reaction systems,the distribution of the inorganic substances is really parallel to that of water in the wood specimens. However, metal alkoxides with Ti, Al, andZr are gelled as type 11 or 111, filled or surrounded with gels only in the cell cavities, by use of moisture-conditioned specimens (Fig. 2c) but not as in type I. Additionally, the use of water-saturated wood did not allow inorganic gels in either cell walls or cell cavities, but only allowed them to cover the outer surface of the specimens with inorganic substances.

FIGURE 2 SEM micrographs (upper) and Si-K, and Ti-& X-ray mapping (lower) over the corresponding area of the SEM micrographs: (a) composites prepared from moisture-conditioned specimens (9.5 WPG) with SiO, gels in the cell walls; (b) composites prepared from water-saturated specimens (122 WPG) with SiO, gels in the cell cavities; (c) composites prepared from moistureconditioned specimens (45 WPG) with TiO, gels in the cell cavities. (From Refs. 3-5.)

Saka

784

These distribution results are due to the hydrolysis rate of metal alkoxides and the subsequent sol rate by the polycondensation reaction, which increases in the order

B < AI,

Si,

Zr < Ti

A difference in the hydrolysis/polycondensation rate seems to result in the difference in the distribution of inorganic substances within the wood cells [6]. Evidence for this concept can be found in a study of TiO, wood-inorganic composites prepared with titanium alkoxides and titanium chelates [IO]. These agents have different rates of hydrolysis and polycondensation, and accordingly result in different distribution of TiO, gels in wood cells.

B. Topochemistry in Enhancingthe Properties of Wood Wood has characteristic properties of dimensional instability, flammability, and biodeterioration. Considering these properties as defects of wood, effective treatments have been investigated to improve wood properties, without losing any favorable properties. The following discussion therefore deals with the effective property enhancement of wood by inorganic treatments and its topochemistry.

1. DimensionalStability Figure 3 shows relationship between antiswelling efficiency (ASE) and weight percent gain (WPG) in various wood-inorganic composites. Here, the ASE is a measure to evaluate the dimensional stability, i.e., an ASE of 0% refers to no control of dimensional stabilization, whereas 100% ASE refers to its complete control. In wood-inorganic composites with SiOz or BzO, gels (type I), inorganic substances are formed selectively within the cell walls and their dimensional stability is effectively enhanced with a small increase in WPG [4,6]. It is further demonstrated that even higher 50

l

‘‘

-30 0

I

l

I

I

I

10

20

30

40

50

60

W P G (%> FIGURE 3 Relationship between antiswelling efficiency (ASE) and weightpercent gain (WPG) in various wood-inorganic composites. (From Ref. 6.)

Wood-Inorganic Composites

785

dimensional stabilization can be achieved if the SiO, gels are bound directly with cell wall components through isocyanate- or epoxy-type silane coupling agents [7]. On the other hand, for wood-inorganic composites with TiO,, AI2O3,and ZrO, gels (type 11 or 111), in which inorganic substances are present in the cell cavities, dimensional stabilization cannot be achieved. Interestingly, however, TiO, wood-inorganic composites from titanium chelates (type I) as mentioned above revealed an improvement of dimensional stabilization [lo].

2. Fire-RetardantProperties Figure 4 shows the combustibility test after 45-S ignition of model houses made from 2mm-thick veneers of hinoki (Charnaecyparis obfusa Endl.). It is apparent that the SiO, wood-inorganic composites (right) show much higher resistance to burning, compared with the untreated wood (left). Figures 5a-5c show micrographs obtained by scanning electron microscope (SEM) of the transverse surfaces after the combustibility test [4].It should be noted that the untreated wood (a) has thin cell walls, whereas thecomposites prepared fromthe moistureconditioned specimens with 9.5 WPG (b) retain rather thick cell walls by the deposition of SiO, gels and carbonizationof the cell walls. This apparently shows that a small amount of SiO, gel formed within the cell walls can enhance fire resistance. On the other hand, the composites with 40 WPG prepared with water-saturated specimens (c) cannot protect the cell walls and the cell walls are as thin as in the untreated wood (a).

3. TermiteDeterioration Figure 6 shows the results of termite tests with Reticulitermes sperafus Kolbe. On the relationship between the number of dead termites and the test periods [4,6]. Compared with untreated specimens, the composites from water-saturated specimens with 70 WPG (type IV) show significant resistance to termite attack, due perhaps to the blocking effects of the gels. However, the composites prepared from moisture-conditioned specimens (type I) also reveal termite resistance with only 10 WPG. It is therefore noteworthy that a small amount of SiO, gel formed within the cell walls is as effective against termite attack as gels formed in the cell cavities.

FIGURE 4 Comparison of combustibility tests of model houses made from untreated (left) and SiO, composite wood veneer (right) after 45-S of ignition.

Saka

FIGURE 5 SEM micrographs of wood-inorganic composites after combustibility tests: (a) untreated wood; (b) composites with Si02 gels in the cell walls (9.5 WPG); (c) composites with Si02 gels in the cell cavities (40WPG); (d) composites with Si0,-P,O, gels (16.9 WC);(e) composites with Si02-B20, gels (38.7WPG); (f) composites with SiO2-P2O5-B2O3 gels (34.3WPG). (From Refs. 4.8.)

100 W

a,

c, .l4

E 8

rd

8

50

a %l

0

35

= o 0

10

20

30

40

Test Period (days) FIGURE 6 Relationship between the number of dead termites and the test periods for various wood-inorganic composites: 0 SiOz composites (type W ) ;0 SiO, composites (type I); 0 TiO, composites (type 11); ZrO, composites (type 111); A A1,0, composites (type 111); A untreated wood. (From Refs. 4,6.)

787

Wood-Inorganic Composites

WaterRepellency Figure 7 shows the change of water absorption ratio (WAR) in Si02 wood-inorganic composites prepared with and without a water-repellent agent as a property enhancer. As SiO, wood-inorganic composites were prepared in the reaction medium of tetraethoxysilane (TEOS)/ethanol(EtOH)/acetic acid (catalyst) with a molar ratio of 1 :1:0.01, the water-repellent agent silicon alkoxide with a long-chain alkyl residue or perfluoroalkyl residue was added to the reaction medium in a molar ratio of 0.0 I . These water-repellent agents are, for example, DTMOS [decyltrimethoxysilane, CH,(CH,),)Si(OCH,),] and HFOETMOS 12-heptadecafluorooctylethyltrimethoxysilane;CF,(CF,),CH2CH2Si(OCH3),],and have a large molecular weight so that they are mainly distributed over the surfaces of the cell cavities, as observed by SEM-EDXA study. However, due to the trimethoxysilyl residue present in the water-repellent agent, it is bound with SiO, gels formed within the cell walls as described below and thus fixed in the wood cells with low-surface-energy residues (R') exposed over the surface of the cell cavities.

4.

l SiO,

\

+

OH

\

H-0-Si-R'

\

Y

- ( CH,) ,CH,

R ;

or

-CH,CH, (CF, ) ,CF,

Compared with SiO, composites and untreated wood, composites with a water-repellent agent revealed lower WARin Fig. 7, indicating a water-repellent property added to the composites [ S ] .

5. Antibacterial Properties Creosote and chromated copper arsenate (CCA) are widely used as preservatives for antibacterial treatment of wood. In spite of their excellence in this property, they have some

Untrealed

TFPTMOS-SiO,

HFOETMOS-SiO,

1

0

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

1

5 10 15 Test period ( days )

FIGURE 7 Change of waterabsorptionratio in Si02 wood-inorganic composites prepared with and without a water-rcpellentagent:TFPTMOS, 3,3,3-trifluoropropyItrimethoxysilane;DTMOS, decyltrimethoxysilane: HFOETMOS. 2-heptadecaHuorooctylcthyltrin~ethoxysilane. (From Rcf. 9.)

Saka

788

drawbacks in terms of toxicity. Therefore, less toxic and environmentally acceptable chemicals are expected to be used. Quarternary alkylammonium salts are among the candidates for the antimicrobial treatment ofwood.Inthisstudyofwood-inorganic composites, trimethoxysilylpropyldimethyloctadecylammonium chloride (TMSAC), shown below, was used as a property enhancer to add an antibacterial property to wood [ll].

L As in water-repellent agents, TMSAC can be expected to be bound with SiO, gels and fixed in the wood cells, with quarternary alkylammonium salt residue exposed over the surface of the cell cavities. Therefore, TMSAC with a molar ratio of 0.005 was added to the reaction medium of TEOSEtOWacetic acid (molar ratio 1: 1:O.Ol).It is quite apparent in Fig. 8 that TMSAC-added SiO, composites (TMSAC-Si02) reveal a significant resistance against attack by white-rot fungi [Coriolus versicolor (L.ex Fr.) Quell. It should be noted further that TMSAC-Si0, composites are more resistant than TMSAC wood. This result canbeexplained by Fig. 9, inwhichTMSAC-SiO, composites are more water repellent compared with TMSAC wood, due perhaps to more uniformly distributed TMSAC with its long alkyl residue, chemically bound with Si02 gels in wood [l l]. In summary, if the composites have inorganic gels distributed selectively within the cell walls, as in type I, effectivelyenhanced properties can be achieved.However,in composites of type 11 or 111, improvement of the properties cannot be expected. In types IV and V, property improvement can be expected to some extent, but the porous structure characteristic of wood would be diminished. The results obtained therefore indicate that it is more effective for enhancement of the wood properties to incorporate inorganic substances into wood in neighbor wood cell wall components rather than to deposite them in the cell cavities, far from the cell wall components. For the property enhancers of silicon alkoxide with hydrophobic residue, such as a long-chain alkyl or perfluoroalkyl residue, a water-repellent property can be added to woodby covering the surfaces of the cell cavities. Similarly, property enhancers with a quarternary alkylammonium salt residue can add an antibacterial property to wood. Therefore, it may be concluded that the topochem-

FIGURE8 Comparisons of fungal attack by whlte-rot fungi:SiOz composites (6.7 WPG); TMSAC wood (0.7 WPG); TMSAC-SiO, composites (4.1 WPG). (From Ref. 11.)

789

Wood-Inorganic Composites

1.o

l

~

I

~~

Untreated *

x

v

,a

E 0

Y

.m

e 0.5

c)

2

0

I

I

5

10

15

Test periods ( days ) FIGURE 9 Changes of water absorption ratio for various composites: SiO, composites (6.7 WPG); TMSAC wood (0.7 WPG); TMSAC-SiO, composites (4.1 WPG). (From Ref. 1 I .) ical effects of inorganic substances and property-enhancers exist for wood property enhancement in wood-inorganic composites.

IV.

MULTICOMPONENTWOOD-INORGANICCOMPOSITES

Through a study of monocomponent wood-inorganic composites, S i 0 2 composites have been found to be harmless and environmentally friendly. Additionally, composites prepared from moisture-conditioned wood specimens have porous structures characteristic of untreated wood. To improve the properties of these SiO, composites further, a study of multicomponent wood-inorganic composites has been made with two or more kinds of metal alkoxides.

A.

Si0,-P,O,-B,O,

Wood-InorganicComposites

The preparation of Si02-P,0s, Si02-B203, and SiOZ-Pz0,-B2O3 wood-inorganic composites is based on the reaction medium of TEOS/EtOH/acetic acid (molar ratio 1: 1:0.01) with trimethylphosphite (TMP) and/or trimethylborate (TMB) with a molar ratio of 0.05 181. Figure I O shows the results of thermogravirnetric (TG) analyses of the composites obtained. In the TG curve of the untreated wood (a), an abrupt decrease in its weight can be observed in a temperature range between 300 and 350°C due to the flaming. However, in the SiO, composites formed within the cell walls (8.4 WPG) (b),the flaming temperature was shifted higher with the higher residual weight, compared with untreated wood. The difference between (a) and (b) in the TG curves would have reflected upon the differences observed in Figs. 4 and 5. On the other hand, binary and ternary composites (c, d, and e)

790

Sa ka

-

Flaming.

100-

.

Glowing

A

x

v

c,

-c

M

.F-

a

M

2

50-

2

-...---

v1

a,

_._-__ e

d "." C _______.__ -. b

CT

. " 4 2 " " -

-0-1

0

I

l

I

100

200

300

I

1

500 Temperature("C) 400

""""

a

I

l

I

600

700

800

FIGURE 10 Thermogravimetricanalyses of variouswood-inorganiccomposites: (a) untreated wood: (b) SiOzcomposites (8.4 WPG); (c) Si02-P20, composites (16.9 WPG); (d) SiO,-BzO2 composites (38.7 WPG); (e) Si02-Pz0,-Bz0, composites (34.3 WPG). (From Ref. 8.)

revealed fairly high residual weight for flaming, showing very high tire resistance. Furthermore, for glowing at a higher temperature over 350°C in (d), or over 300°C in (c) and (e), both binary and ternary composites show fairly high tire resistance. These composites have the highest WPG at 38.7, keeping the cell cavities nearly empty. Figure 11 shows comparisons of the combustibility tests among five plywood veneers made from untreated wood (left) and inorganic composites (right) [S]. After 30-S of ignition with a gas burner, the burner was removed and combustibility with the remaining flame was observed. Compared with SiOLcomposites, binary and ternary composites reveal stronger fire resistance. Furthermore, as shown in Figs. Sd-Sf, SEM micrographs of these carbonized composites after this combustibility test apparently show that the thinning of the cell walls is much less in these binary and ternary composites, compared with untreated wood (Fig. Sa). Figure 12 shows the results of differential thermal analyses (DTA) of those binary and ternary composites [S]. In the untreated wood (a), significant endothermic peaks corresponding to flaming and glowing can be observed. However, these peaks weaken i n the SiO, composites (b). and in binary and ternary composites they disappear and broaden to the higher temperature in SiO,-B,O, composites (d). In Si0,-P,O, (c) and Si02-P,0,B,O, composites (e), the endothermic peaks broaden to both higher and lower temperatures, revealing high resistance to combustion. For the tire retardance observed above, the mechanism of B,O, gels is different from that of P,O, gels. In Fig. 10, the flaming temperature is shifted lower for composites with P,O, gels, whereas for the composites with B,O, gels or SiO? composites, it is shifted higher. Thegels in the latter composites are believed to be melted during the flaming process and form a glassy layer to cover the cell wall components. As a result, fireretardant properties are added through the physical barrier against heat and oxygen. On the other hand, for composites with P,O, gels, chemical effect by dehydration with phosphorus has promoted the carbonization of the composites which have revealed fire-retardant properties [ IS]. Therefore, as seen in Fig. IO, ternary composites with Si0,-P,O,-

791

Wood-Inorganic Composites -

~-

~.

~

~.~

FIGURE 11 Comparisons of various wood-inorganic composites after combustibility tests: (left) untreated wood veneer; (right) wood-inorganiccomposites corresponding to those in Fig. 10. (From Ref. 8.)

I

0

I

100

I

I

I

I

1

200

300

400

500

600

l

700

l

800

Temperature ("C1 FIGURE 12 Differential thermal analyses of variouswood-inorganic composites: (a) untreated wood; (b) SiO, composites (8.4 WPG); (c) Si0,-P,O, composites (16.9 W C ) ; (d) Si0,-B,O, composites (38.7 WPG);(e) Si0,-PZ05-B,0, composites (34.3 WPG). (From Ref. 8.)

792

Saka

B20, gels would have high fire-retardant properties through effects of both a physical barrier and chemical reaction as mentioned above.

B. Leachability of Inorganic Substances For those wood-inorganic composites with fire-retardant properties, a leaching test for inorganic substances was made under the severe conditions of water stirred 160 times per minute in a 250-mL beaker [9]. The results obtained clearly indicated that the P20, and B,O, gels are leached out readily, but SiO, gels are stable in composites. To overcome this property, we tried to use the water-repellent agents studied in Fig. 7 with a molar ratio of 0.01 on the reaction system of TEOS/EtOH/acetic acid with TMP or/and TMB(molar ratio 1: 1:0.01 :O.OS:O.OS). The water-repellent agents used were DTMOS (decyltrimethoxysilane) and HFOETMOS (2-heptadecafluorooctylethyltrimethoxy silane), as already mentioned. Figure 13 shows one example of the results obtained, which clearly indicate that the leaching of P,O, gels is fairly prevented with an addition of such property enhancers, particularly HFOETMOS, and of course, fire-retardant property was maintained. Similar results were obtained with B 2 0 , gels 191. However, leachability was slightly higher than for P,O, gels, so some improvement would be necessary for permanent achievement of antileachability.

V.

WOOD-INORGANIC COMPOSITES WITH MULTICOMPONENT OLIGOMERS

In a practical sense, TMP needs to be handled cautiously because of its odor and toxicity, while BzO, gels from TMB are not stable and are readily leachable. Additionally, consid-

Treatment time (h) FIGURE 13 Leachability of P,O, gels in Si0,-P,O, compositcs and effects of the water-repellent agent added to composites on the prevention of leaching. (From Ref. 9.)

Wood-Inorganic Composites

793

ering operational and processing environments, methyltrimcthoxysilane (MTMOS) is more appropriate than TEOS because it is a safer agent, and its potential for displacement of TEOS has already been demonstrated to prepare SiO, wood-inorganic composites [ 121. Therefore, to pursue the development of “superwood” with topochemical effects, some silicon alkoxide oligomers with ethylphosphite and boric hydroxide residues were prepared from MTMOS, TMP and boric acid, and applied for the preparation of SiO2-P2O,-Bz0, wood-inorganic composites [ 131. The oligomers prepared are described below.

Me

I

OMe

I

Me

I

MeO-Si-0-Si-0-Si-0-Si-OMe I l I 0 Me OMe

0 I

I

Me

R = CH2CH3 or H

Through their evaluation, the composites had a fire resistance as high as the composites prepared from the reaction system TEOS/EtOH/acetic acid with TMP and TMB. Furthermore, leaching of the gels was prevented to ;I greater extent, due perhaps to chemical bonding of multicomponents in oligomer levels which could have stabilized inorganic substances in wood cell walls. Additionally, the oligomers prepared are nontoxic, so environmental safety in their preparation was achieved. By adding HFOETMOS in a small quantity as a property enhancer to the oligomer reaction system, the composites obtained could improve further the antileachability of PzO, and B,O, gels.

VI.

CONCLUDING REMARKS

To develop wood with high function and remarkable properties, we have tried to prepare inorganic composites of wood without losing characteristic properties of the wood as seen in its porous structure. As mentioned already, in spite of the same inorganic substances used, the observed properties are different if inorganic substances are distributed differently in the wood cells.This is because topochemical effects exist in wood for property enhancement.Our goal fordeveloping ideal “superwood” is to achievewood-inorganic composites from environmentally friendly materials with their minimal use and maximal effect on property enhancement. More extensive study of topochemistry in wood property enhancement will provide a clue to a development of “superwood.”

794 7. K. Ogisoand S. Saka. Mokuzcci Gakkaishi,40:1100 (1994). 8. H. Miyafuji and S. Saka, M o k l c x i Gakkaishi, 42:74 (1996). 9. S. Saka and F. Tanno. Mokuzni Gdknishi, 42:81 (1996). IO. H. Miyafuji and S. Saka, Wood Sci. 7 2 ~ . h r 1 o l .3/:449 , (1997). 1 1. F. Tanno, S. Saka, and K. Takabe, Mater: Sci. Res. h f . , 3: 137 (1997). 12. S. Saka and T. Ueno. Wood Sci. Techr~ol.. 31:4S7 (1997). 13. H. Miyafuji, S. Saka. and A. YaInamoto, Hol;for.sch. 52:410 (1998). 14. S. Sakka, Scicwcc. by Sol-Gd Process, Agune-shofusha, Tokyo, p. 8 (1988). IS. F. L. Browne, U.S. FPL Rep. 2136, U.S. Forest Service, Forest Products Lab. (1958).

Sa ka

21 Preservation of Wood Darrel D. Nicholas Mississippi State University, Mississippi State, Mississippi

1.

INTRODUCTION

A number of books and summaries that deal with wood preservation are currently available [ 1-51. Consequently, there is little need for another general review of this subject. However, in the past few years there have been significant changes in the wood preserving industry, so a review of these trends seems appropriate. Accordingly, in this chapter emphasis will be placed on new developments and trends in the wood preserving industry. New developments in wood preservation have been mainly in the area of new preservatives, so this chapter will focus heavily on these advances.

II. TREATMENT PROCESSES AND TECHNOLOGY The selection of wood preservatives, formulations, and treatment methods is dependent on the product and type of protection required. For example, control of sapstain and mold in green lumber is accomplished by dip- or spray-treating the wood with aqueous formulations. Since only short-term protection is required, this type of treatment is adequate. Millwork is also treated by the dip method, but somewhat higher preservative loadings are attained in dry, highly permeable wood species. Pressure processes are used for products which are used in adverse environments. The use of pressure/vacuum systems makes it possible to achieve good penetration and retention of the biocidcs, which are required for these products.

A.

PressureProcesses

The majority of wood products are treated by conventional pressure methods using either the full-cell or empty-cell process. The full-cell process uses an initial vacuum to evacuate air from the wood, followed by filling the cylinder with preservative solution under vacuum prior to the application of pressure. This process is generally used for water-borne preservatives, where maximum treating solution retention is desired. In recent years the lnoditied full-cell process has become increasingly popular. The same basic cycle is used with the exception that the initial vacuum is reduced by about 40-50% and a final vacuum is applied after the pressure period. Use of the modified full-cell treatmentprovidesa 795

Nicholas

796

means of reducing the final solution retention. which minimizes the dripping of preservative solution and subsequent holding time after treatment. Empty-cell processes (Lowry and Rueping) do not employ an initial vacuum and as a consequence result in much lower net preservative solution retentions. By applying some degree of initial air pressure and filling the cylinder at this pressure (Rueping process), the net solution retention can be reduced even further. These treating processes are generally used with oil-borne preservatives where it is desirable to minimize the amount of carrier oil used.

B. Vapor-PhaseProcess A major concern i n the treatment of many wood species is achieving adequate penetration of the preservative. One possible approach to this problem is to use vapor-phase treatments. The validity of this concept has been demonstrated in New Zealand, where they successfully treated various wood products with borates Barnes and Murphy [6]. In this process, trimethyl borate is vaporized by heating and then introduced into an evacuated cylinder containing wood at a low moisture content. The trimethyl borate rapidly diffuses into the wood and reacts with residual water to form boric acid in situ.

C. Supercritical-Fluid Process Another approach to the problem of poor preservative penetration is to use supercritical fluids as carrier solvents in the treating process. Supercritical fluids are capable of penetrating the small openings i n wood because they effectively eliminate the interfacial problems associated with conventional liquids. The feasibility of this process has been demonstrated in laboratory studies 171. However, a substantial amount of research willbe required before it can be determined whether this process has any commercial applications.

111.

PRESERVATIVES

Biodeterioration of wood products by microorganisms and insects is a major problem and results in millions of dollars lost annually. A substantial portion of this money could be saved if appropriate control methods were used. Some wood species are naturally resistant to biodegradation because they contain toxic heartwood extractive. However, as a result of considerable variation in durability among trees and a shortage of material, nondurable species are generally used for these applications. In order to attain a reasonable service life from nondurable woods used in exterior applications, they must be treated with wood preservatives. Currently, all commercial wood preservative formulations contain chemicals that are toxic to microorganisms and insects. These chemicals protect wood by preventing the attack of wood-decay fungi and in some cases insects. In order t o be commercially viable wood preservatives, biocide formulations must have the following characteristics: Cost effective Good permanence in the wood under use conditions No significant cff'cct on the strength properties of wood Low corrosivity to metal F'Istcners Good penetration properties Safe to handle and use

Preservation of Wood

797

Low mammalian toxicity N o detrimental effects on the environment

A.

Types of Preservatives

Wood preservatives are generally classified into two basic types-oil-borne and waterborne-which are distinguished by the type of carrier used to solubilize the biocides.

1. Oil-borne Preservatives Creosote and pentachlorophenol (penta) are the major wood preservatives currently being used. Another oil-borne biocide that has been used on a limited basis is tributylin oxide (TBTO). Abrief discussion of the general characteristics of these preservatives is presented below. a. Creosote. The use of creosote as a wood preservative was patented in 1838 by Bethell, and since that time creosote has remained an effective, widely used chemical for treating wood [S]. Creosote is a complex mixture containing at least 200 identifiable compounds. However, it is generally agreed that several thousand different compounds are present in very small amounts [91. The greater part of the composition of creosote consists of neutral fractions (Table l ) . Tar acids, such as phenol and the creosols,as well as such tar bases as pyridenes, quinolines, and acridines, constitute a rather small percentage of the total weight of creosote. Unlike the neutral fractions, the tar acids and bases are usually soluble in water and hence contribute very little to the efficacy of creosote as a wood preservative. It follows

TABLE l ChemicalComposition of a Typical Creosote Produced in the United States Compound or Component Naphthalene Methyl naphthalene Diphenyl dimethylnaphthalene Biphenyl Acenaphthene Dimethylnnphthalcne Diphcnyloxide Dibenzofuran Fluorene-related compounds Methyl Huorenes Phenanthrene Anthracene Carbazole Methylphcnanthrene Methyl anthracenes Fluoranthene Pyrene Benzofluorene Chrysene Other components not identified

U S . Creosote [34] 3.0 2.1

0.8 9.0 2.0 -

S.o 10.0 3.o 21.0 2.0 2.0 3.0 4.0 10.0 8.5 2.0 3.0

Nicholas

798

from the foregoing statements that the chemistry of creosote and that of the coal-tar neutral fractions are quite similar. So, for that matter, is the chemistry of the parent materialcoal tar. Compositional data for coke-oven coal tar produced in the United States is given in Table 2. The majority of the compounds in creosote are aromatic hydrocarbons with condensed ring systems. In addition, tar acids, which are heterocyclic compounds containing nitrogen plus some neutral oxygenated compounds, are present [ 101. The tar acids and bases contain a wide range of fungicidal constituents. The fungicidal activity of creosote fractions obtained by distillation varies widely. In this regard, Schulze and Becker [ 1 l ] investigated the fungicidal activity of I5 distillate fractions boiling between 120 and 360°C against three wood-decay fungi. They found that the toxicity of the different fractions varied widely, with those boiling between 180 and

TABLE 2 Chemical Composition of U.S. Coke-Oven Tars Component Water, '70 Carbon, o/o (on dry tar) Hydrogen, 96 (on dry tar) Sulfur, 95 (on dry tar) Nitrogen, C7c. (on dry tar) Ash, '3- (on dry tar) Toluene insolubles, c/c (on dry tar) Components wt, Oh (on dry tar) Benzene Toluene o-Xylene rtl-Xylene />-Xylene Ethylbenzene Styrene Phenol 0-Cresol m-Cresol p-Cresol Xylenols Higher-boiling tar acids Naphtha fraction (bp 150-200°C) l-Methyl naphthalene 2-Methyl naphthalene Acenaphthene Fluorenc Diphenylenr oxide Anthracene Phenanthrene Carbazole Tar bases Medium-soft pitch (70°C. R, and B softening pt.)

Wt (95) 2.2 91.3 5. I I .2 0.67 0.03 9.1 0.12 0.25 0.04 0.07 0.03 0.02 0.02 0.6 1 0.25 0.45 0.27 0.36 0.83 0.97 8.80 0.65 1.23 0.84 0.75 1.66 0.60 2.08 63.5

Preservation of Wood

799

240°C being the most active. Within these fractions, thionaphthene, 2-naphtho1, 2-methylnaphthalene, and isoquinoline were the most active compounds. However, it is likely that some of the components in creosote exhibit synergism, so the toxicity of individual components is probably not of major significance. In recent years attempts have been made to improve the surface characteristics of creosote-treated wood. Thisconcept was termed “clean”creosote andis achieved by reducing the xylene insolubles (XI) present in creosote solutions [ 121. By reducing the XI content to 0.1% or lower, a much cleaner product is produced. b. Pentachlor-ophenol (Petzta). Penta was first used asa wood preservative in the 1930s and rapidly became established as the most widely used single oil-soluble biocide in wood preservation. The performance of penta-treated wood for ground-contact applications is highly dependent on the carrier system used [ 13- 171. The effect of the carrier oil on performance has been attributed to the following factors: ( 1 ) its effect on penta depletion; (2) its effect on distribution of penta in the wood structure; and (3) its intrinsic biological activity [ 171. The fungicidal activity of the carriers varies considerably among the petroleum oils used for this purpose, with some oils demonstrating reasonably good performance in the soil block test [ 171. With regard to penta depletion from treated wood, there is considerable variation among the oils. Furthermore, there appears to be an inverse relationship between the penta depletion rate and performance of field stakes [ 171. Over the years, a number of different carriers other than petroleum oils have been used for penta treatments. These include transient light solvent systems-liquefied petroleum gas(LPG), methylene chloride, and mineral spirits-and water-based emulsions. The LPG (Cellon) and methylene chloride (Dow process) systems were used extensively for a number of years to treat utility poles. However, poor performance of the treated wood due to erratic penta distribution and environmental concerns resulted in abandonment of these processes. The water-borne emulsion penta system also failed due to poor performance of the treated wood products. Inadequate performance was probably due to excessive leaching of penta from treated products subjected to exterior exposure.Another water-borne system, the water-soluble sodium salt of penta, was used extensively for dip or spray treatment of green wood tocontrol stain and mold fungi. Because of environmental concerns this compound is no longer used for this application. c. Tributylin Oxide. Organotin compounds were shown to have fungicidal properties in the 1950s. Subsequently, laboratory and field tests demonstrated that tributylin oxide (TBTO) was the best wood preservative on the basis of cost, permanence, and mammalian toxicity [ 181. TBTO has been used extensively in Europe as a wood preservative for aboveground applications. It has also been used to a limited extent in the United States as a replacement for penta in millwork applications. However, recent studies have shown that this compound is not stable in contact with wood and gradually decomposes over a period of time [19]. As a consequence, TBTO is no longer used in the United States. In Europe, tributylin naphthenate has replaced TBTO and apparently is equally as effective and more stable.

2. Water-Borne Preservatives Chromated copper arsenate (CCA) and ammoniacal copper zinc arsenate (ACZA) are the major water-borne wood preservatives currently being used commercially in the United States. A brief description of the general characteristics of these preservatives is presented below.

Nicholas

800

(1. Chromlted Copper Arsenate. CCA is unquestionably the most important wood preservative in the United States, representing 78% of all preservatives used in 1993 [6]. Although CCA is water-soluble, it undergoes a series of complex fixation reactions in wood. These reactions involve both lignin and carbohydratecomplexesas well as inorganic precipitates (Fig. l ) . As a consequence of these reactions, CCA is highly fixed in wood and resists leaching even under severe exposure conditions. This interaction with wood results in a decrease in strength properties, with toughness being particularly sensitive to this treatment. However, these strength losses can be minimized to 10% or less by drying the treated wood at temperatures of 71°C or less 120-221. CCA is a very effective wood preservative andis used for numerous applications suchas lumber fordecks, utility poles, marine piling, etc. When the wood is properly treated. an extremely long service life can be obtained with these products. h. Anmoniacal Copper Zinc Arsenrrre. The useof ACZA is limited to the West Coast area, where it is used to treat Douglas fir and other local wood species. The alkaline solution provides better penetration of these relatively refractory wood species. ACZA is not as highly fixed as CCA, and the chemical reactions responsible for this insolubilization are not clearly understood. The main mechanism of fixation of copper and zinc is postulated to be the formation of insolublecopper arsenate and zinc arsenate. However, the overall mechanism is undoubtedly more complex because cuprammonium ions react by ion exchange with functional groups in wood 1231. In addition, both copper and zinc complexes can be formed with the wood substrate, but copper and zinc complex formation do not appear to be related.

3. New Wood PreservativeSystems In recent years considerable research has been directed at the development of new wood preservatives. This activity was stimulated by the actions of the U.S. Environmental Protection Agency, which questioned the environmental impact of the major wood preservatives used in the United States. As a result of this work, a number of biocides have been identified as potential new wood preservatives (Table 3 ) . The majority of these chemicals have been used for other applications in agriculture, paints, etc. Others, such as copper

CCA (Cu”, Crs+, As5’)

Cu2+,Cr3+,As5+

W + , CuCrO,,

l

l

Carbohydrate

Inorganic Precipitates

cu2+

CrAsO,

FIGURE 1 CCA reactions with wood.

Cr(OH1,. CrAsO,, Cr,(OH),CrO,, Cu(OH)CuAsO,

TABLE 3 Biocides withPotential a s Wood Preservatives Trade name

Chemical name 3-Iodo-2-propynyl butyl carbamate

Ti mbor2'In

Disodium octaborate tetrahydratc

Copper naphthcnate

Copper( 11) naphthenatc

Oxine copper/ copper-8

Copper-8-quinolinolate

structureChemical

H 9 l I -c=c-~-o--~-N - C * H ~

A

Na2B,0,,.4H20

4,S-Dichloro-2-,1-octyl-4isothiazolin-3-one

Busan 3 ( P / TCMTB

2-(Thiocyanomcthylthio) benzothiaxole

WocosenW propiconazole

(2RS, 4RS)-2-(2,4dichlorophenyl)-2-[ 1- 1H ( I .2,4-triazoIe)methyI1-4propyl- 1.3-dioxolanc

Tebuconazole

(3RS)-5-(4-~hlorophenyl)2. 2-dimethylethyl-3-1H [ I .2.4-triazole)methyl)-3pentanol

Amical 4 8 O

Diiodomethy-/,-tolysulfone

Chlorothalonill tuffgard@>B

2,4,5.6-Tetrachloroisophthalonitrile

Cl

I

OH C I O C H ~ - C H Z - C - C (I C H ~ ) ~ ?Hz

C=N

C=N

802

Nicholas

naphthenate and Cu-8, have been used to a limited extent as wood preservatives in the past but are now gaining more popularity because of their relatively low mammalian toxicity. A brief discussion of each of these biocides is presented below. N. Polyphusea (ZPBC). IPBC is an organic biocide that exhibits low mammalian toxicity and has broad-range activity against common wood decay, mold, and stain fungi, but is not effective against wood-destroying insects. It is currently being used to treat wood for above-ground applications for millwork and similar products. A combination of IPBC and DDAC is effective against mold and sapstain fungi and is used extensively for controlling these microorganisms in freshly sawn lumber. b . TinzborQ ( B o r d B o r i c Acid). T i m b o e is an inorganic biocide with boron being the active component. It hasa very low mammalian toxicity and exhibits broad-range activity against both wood-decay fungi and insects. It is highly soluble in water and readily diffuses in and out of wet wood. Consequently, its use as a wood preservative is limited to above-ground applications which are protected from the weather. It is currently being used to a limited extent in the United States for commercial products. c. Copper- Nuphthenate. Copper naphthenate is an organometallic compound that is normally prepared by the direct reaction of copper hydroxide with naphthenic acid at elevated temperatures in a hydrocarbon solvent. It exhibits low mammalian toxicity and has broad-range activity against wood-decay fungi and insects. Copper naphthenate is not a new wood preservative and has been used to a limited extent as a preservative for a variety of wood and textile products over the years. As a result of recent environmental concerns with the major wood preservatives, interest in copper naphthenate has increased. This led to its use by several companies as a preservative for utility poles. However, some utilities have experienced early failure of some of these poles, and this has resulted in curtailment in the use of copper naphthenate. The exact cause of these early failures is not known, but it appears that it may be due at least in part to inactivation of copper naphthenate when it is used to treat green poles that are steam-conditioned. In any event, this experience will undoubtedly result in limited use of this preservative in the future. d. Oxine Copper (Cu-S). Cu-8 is an organometallic compound formed by the reaction of copper with 8-quinolinol. It exhibits low mammalian toxicity and has broadrange activity against wood-decay fungi and insects. Cu-8 is very insoluble in water and most organic solvents, butan oil-soluble form can be made by reaction with 2-ethyl hexoate. A water-soluble form can be made with dodecylbenzene sulfonic acid, but this formulation is highly corrosive to metals. Cu-8 is not a new preservative and has been used to a limited extent as apreservative for a variety of wood and textile products over the years. It is currently the only preservative that is approved for treating wood that is in contact with foodstuffs. It isused primarily for treating specialty items such as food pallets and picnic tables. e. Burduc 2 2 a (DDAC). DDAC is an organic biocide that exhibits low mammalian toxicity and has broad-range activity against wood-decay fungi and insects. It is a watersoluble compound, but undergoes an ion-exchange reaction with wood which greatly reduces its leachability from wood exposed to water or wet soil. In order to improve the efficacy of DDAC as a wood preservative, it is generally combined with other biocides. In this regard, DDAC is currently used in two commercial wood preservative formulations.As mentioned previously, a formulation composed of DDAC and IPBC is widely used to control mold/sapstain in green wood. More recently, a combination of DDAC and ammoniacal copper (ACQ) has emerged as a commercially viable alternative to CCA for many applications.

Preservation of Wood

803

J: Kathon 9 3 0 a ( R H 287). RH 287 is an organic biocide which exhibits low mammalian toxicity and has broad-range activity against wood-decay fungi and insects. It is readily soluble in hydrocarbon solvents and is practically insoluble in water. This compound is not currently being used commercially as a wood preservative, but it has considerable future potential. g. Busan 30Q (TCMTB). TCMTB is an organic biocide which exhibits low mammalian toxicity and has broad-range activity against wood-decay fungi and insects. It is readily soluble in hydrocarbon solvents and is practically insoluble in water. A TCMTB formulation which contains methylene bis thiocyanate is currently being used commercially for sapstain/mold control in freshly sawn lumber. 11. Propiconcczole ( WocnsenQ). Propiconazole is an organic triazole biocide which has low mammalian toxicity and exhibits broad-range activity against wood-decay fungi, sapstain/mold fungi, and insects. It is readily soluble in organic solvents and exhibits very low solubility in water. Propiconazole is currently being used commercially for above-ground treatments and sapstain/mold control applications in Europe and Canada. i. Tebuconazole. Tebconazole is similar in structure to propiconizole and exhibits many of the same properties. It has a somewhat lower mammalian toxicity and exhibits good activity against wood-decay fungi. It has no commercial applications as a wood preservative at the present time. j . Arnica1 4 8 a . Amical 4 8 0 is an organic biocide that has extremely low mammalian toxicity and exhibits broad-range activity against wood-decay fungi and insects. It is readily soluble in a number of organic solvents and exhibits low solubility in water. Currently, it does not have any commercial applications as a wood preservative. k. Chlorothalonil (Nopcocidea, Tuffgarda). Chlorothalonil is an organic biocide which exhibits extremely low mammalian toxicity and has broad-range activity against wood-decay fungi and insects. It has limited solubility in organic solvents and very low solubility in water. Chlorothalonil is currently being used to a limited degree as an additive for mold control in CCA-treated wood. Because of its relatively low cost and good efficacy, it has considerable potential for both above-ground and ground-contact applications. Itis also used as a sapstain chemical as NexGenQ.

4. TrendsinWoodPreservativeDevelopment The current trend in the development of wood preservatives is to use biocide combinations. These combinations include both inorganic-organic and organic-organic binary mixtures. Examples of these are ammoniacal copper quat (ACQ), copper dimethyldithiocarbamate (CDDC), ammoniacal copper citrate (ACC), ammoniacal copper azole, DDAC-IPBC (NPl), chlorothalonil-chlorpyrfos,and DDAC-Na omadine. ACQ has two different formulations-ACQ B and ACQ D. The only difference in these formulations is the copper-complexing agent. Type B contains ammonia and type D contains ethanolamine. In both formulations, the active ingredients are copper and DDAC in a ratio of 2: 1 (Cu0:DDAC). This combination provides broad-range efficacy against wood-decay fungi and insects in both above-ground and ground-contact applications. CDDC is formulated with copper ethanolamine and sodium dimethyldithiocarbamate (SDDC). Since copper reacts rapidly with SDDC to form an insoluble complex, a twostep treating process is required with this preservative system. Accordingly, the wood is first treated with a copper ethanolamine solution in one treating cylinder and then moved

Nicholas

804

to a second treating cylinder for treatment with SDDC.The reaction product is a 1.2 c0pper:dimethyldithiocarbamate chelate which is highly insoluble in water. This chelate appears to have reasonable broad-range efficacy against wood-decay fungi and insects in both above-ground and ground-contact applications. ACC is formulated with ammoniacal copper carbonate and citric acid, using a ratio of 1.6S:l (Cu0:citric acid). Although the AWPA has developed a Standard, the value of this wood preservative system is questionable because both soil block and field stake test data indicate that the treated wood is not resistant to copper-tolerant fungi. This weakness is not surprising, since it does not contain a co-biocide. NP- I Q is formulated with DDAC and IPBC at an 8.5: 1 ratio of DDAC:IPBC. NP-I is used mainly for sapstain and mold control in freshly sawn wood but may have potential for millwork and similar applications. Copper azole is formulated with a combination of copper (49%), boric acid (49%) and tebuconazole (2%), using ethanolamine as the complexing agent for copper. This system is still undergoing evaluation, but shows promise as a viable wood preservative for some applications. The trend toward the use of binary and tertiary biocide combinations in wood preservative formulations is expected to continue.This formulation strategy is particularly significant when the biocides exhibit synergism. Indeed, the development ofwood preservatives on the basis of synergistic mixtures is currently being explored and appears to have considerable potential [24].

5. Biocontrol Another approach to wood preservation is to use antagonistic microorganisms rather than toxic chemicals. The first report concerning the potential to this approach was by Richard and Bollen [2S], using Scytalidiun~sp. to inhibit Antrodicr cclrhonica in Douglas fir poles. The results of this study stimulated additional research on the validity of this method for wood preservation, and this led to the development of Binab AB@. This bioprotectant has been marketed in Europe a s a preservative for Scots pine [26,27]. Although some trials were very positive, there is concern about the long-term effectiveness of this formulation 127-291. Binab AB was also evaluated by Morrell and Sexton [30] as a possible preservative for Douglas fir and southern pine. Poor results were obtained in this latter study, and failure of the system was attributed to the fact that it did not control all the decay fungi that colonize these wood species. Consequently, it was concluded that biocontrol is not a currently viable method for preserving wood products against decay fungi. Although the possibility of using microorganisms for controlling wood-decay fungi is discouraging, this approach may have potential for controlling sapstain/mold fungi in freshly sawn wood products [2]. This particular application requires only short-term protection, so concern about long-term survival of the microorganisms is eliminated. Both bacteria [31] and fungi 1321 show promise for this application.

IV.

FUTUREDEVELOPMENTS

Environmental issues have been the major driving forces behind the development of new wood preservatives in recent years. This trend will undoubtedly continue into the foreseeable future because there currently is considerable concern about disposal of treated wood after it has ended its useful life. This is particularly true for CCA-treated wood, which is

Preservation of Wood

805

not amenable to disposal by incineration. This dilemma will encourage research to develop practical methods for removing CCA from treated wood so that it can be recycled. At the same time, efforts will continue toward the development of alternative wood preservatives that do not pose disposal problems for treated wood. The development of ACQ, which eliminated chromium and arsenic, was a step in the right direction, but the presence of copper complicates disposal problems. The major thrust in wood preservative development in the future will probably be based on total organic biocide systems. If the mammalian toxicity of these systems is low, then the environmental concerns will be effectively eliminated. Other, more specific, and possibly nonbiocidal,approaches to wood preservation may develop as we learn more about the complex reactions involved in the overall decay mechanisms utilized by fungi. The problem of poor weathering characteristics of treated wood, especially CCAtreated wood, will probably alsoreceiveconsiderable attention in the future. If a high degree of water repellency canbeimpartedtowood, it will minimize the weathering problem and also minimize the biocide levels required to inhibit biodeterioration. Significant advancements in this area will help pave the way for development of cost-effective organic wood preservative systems. With regard to treating processes, significant improvements have been minimal in the past, and this trend probably will continue. One of the main problems that needs to be addressed is providing adequate treatment of refractory heartwood. More research is needed in this area, and progress in solving this problem could be very rewarding. Continued progress with the supercritical-fluid treatment process could make a major contribution in this area [ 3 3 ] .Other developments, such as the shock-wave treating process and a compression/vibration pretreatment of the wood, may also prove to be effective methods for improving the treatability of refractory woods.

REFERENCES I.

2. 3. 4. 5. 6. 7.

X. 9.

IO.

11.

12. 13.

R. A. Eaton and M. D. C. Hale, Wood D e c q Pests trr~dProtectiorl. Chapman & Hall, London ( 199.3). R. A. Zabel and J. J. Morrell, Wood Microbiology: 1)ccay m l d I t s Prclvrltiorl, Academic Press, New York ( 1992). J. G. Wilkinson, Irldrtstrid Tirnher Preservntiorz, Associated Business Press, London ( 1979). D. D. Nicholas(ed.), Wood Deteriorntion clnd I t s P revention by Presenwtivo Trcyttnlerlts, Syracuse University Press, Syracuse, NY (1973). G.M. Hunt and G. A. Garrett, Wood Preservntiorz. 3rd ed., McGraw-Hill, New York (1967). H. M. Barnes and R. J. Murphy, Forest Prod. J . , 4.5(9):16 (1995). J. J. Morrell, K. L. Levien, E. S. Demessie, S. Kutnar, S. Smith, and H. M.Barnes, Proc. Curl. Wood Preservntiorl Assrl., 14:6 ( I 993). M. P. Levi, in Wood Drteriortrtiorl crntl I t s Prevelltion by Presenrrtive Trmtnlents, Vol. l , D e g r d r t i o n u ~ l dProtection of Wood (D. D. Nicholas, ed.), Syracuse University Press. Syrocuse, NY. p. I83 ( 1973). Anonymous, The Biologic and Economic Assessment of Pcntachlorophenol, Inorganic Arsenicals and Creosote, Vol. I. Wood Preservatives. USDA Tech. Bull. 16.58-1:28 (1981 ). W. H. Hartford, in Wood Deteriorrrtion and I t s Preverltior? by Pre.ser\*nti\v Tremmv1t.s. Vol. 11, Presenutivcv trrltl Prescwcrtive Systerus, Syracuse University Press,Syracuse, NY, p. I O ( 1973). B. Schulze and G. Becker, Holz$~r.scl~.,2 3 7 (1948). H. M. Barnes and L. L. Ingram, Jr.. Pmc. Am. Wood Pre.srl?.c.r.s’A.s.srl.,Y1:108 (1995). R. H. Baechler and H. G. Roth, Forest Prod. J., 12: 187 (1962).

806

Nicholas

14. I. Hatfield and S. S. Sakornbut. Forest Prod J., 5:361 (1955). 15. F. J. Meyer and R. M. Gooch. Forest Prod. J., 6:117 (1956). 16. W. C. Kelso. E. A. Behr, and R. E. Hill, Forest Prod. J., 5369 (1955). 17. D. D. Nicholas, L. Sites, H. M. Barnes, and H. Ng, Proc. Am. Wood Pre.sen~ers'As.srr.,Y0:44 (1994). 18. T. Hof. J . Inst. Woocl Sci., 4(5)23:19 ( 1 969). 19. R. Meder and K. J. Archer, Holgorsch.,45(2):103 (1991). 20. J. E. Winandy, B. A. Bendsten. and R. S. Boone, Forest Prod. J.. 33(6):53 (1983). 2 1. H. M. Barnesand J. E. Winandy, The InterrrcttionnlResecrrch Group on Wood Preserwtion IRG/WP/3543 ( 1 989). 22. H. M. Barnes, J. E. Winandy, and P. H. Mitchell, Inst. Wood Sci., 11(6):222 (1990). 23. M. A. Hulme, Record Anrrucrl Convention British Wood Preservers' Assn., pp. 38 (1979). 24. T. P. Schultzand D. D. Nicholas, Proc., Wood Presenntiorr i n the '90s nnd Beyond, Forest Products Society, Madison, WI, p. I87 ( 1995). 25. J. L. Richard and W. B.Bollen. Cm. J. Botany, 46:643 (1967). 26. J. L. Richard, J . I n s t . Wood Sci., 7(4):6 (1976). 27. P. 1. Morris, D. J. Dickinson,and J. F. Levy, R e c ~ r dAnrr~ralConvention British Wood Preservers' Assn.. p. 42 (1984). 28. P. I . Morris and D. J. Dickinson, Tlw Intenrntiorrcrl Re.seorch Group o n Wood Pmservotion IKG/WP/IISO, Stockholm, Sweden (1981). 29. A. Bruce and B. King, Materictl tcncl Orgctnisrrren. 18(3):171 (1983). 30. J. J. Morrell and C. M. Sexton, Wood Fiber Sci.. 22:10 (1990). 31. R. K. Velicheti and J . J. Morrell, in Wood Presrrvcrtion i n the '90s crnd Beyond, Forest Products Society Proc. 7308, Madison, WI, p. 245 (1995). 32. S. C.Croan, Tlw Internntiorrcrl Resecrrch Group o n Wood Prrservution IRG/wP/96-10158 (1996). 33. J. J. Morrell and K. L. Levien, in Wood Preservcctiorr i n the '90s crnd Beyond, Forest Products Society Proc. 7308, Madison, W1 (1994). 34. L. F. Lorenzand L. R. Gjovik, Proc. A m Wood Presc.rvers'As.srr.. 6832 (1972).

Preservation of Waterlogged Wood David N.-S. Hon Clernson University, Clenlson, South Carolina

Hay surlk, a shattered visage lies . . . Nothing beside remnirls. Round the clecccy Of that colossal wreck, hourldless c m 1 hare, The lone c z r d level .sc~r~d.s stretch fiw clwcly.

“Ozymandias” ( 1817) Percy Bysshe Shelley

1.

INTRODUCTION

Wood is a versatile material which has been used since the dawn of civilization. At the basic level wood satisfied human beings’ needs or wants in shelter, defense, transport, and leisure. Archeologists have discovered, from time to time, weapons, domestic utensils, tools, building materials, and boats made of wood. It is a durable material under a benign environment.Since woodis acomposite biological material, it inevitably continues to deteriorate asa result of physical, chemical, mechanical, and biological processes. To prolong the service life of wood, many chemical treatments have been developed to protect it from attack by microorganisms such as bacteria and fungi, and insects such as termites (see Chapters 12 and 21). Surface treatments are used to protect wood against moisture and weathering (see Chapters 9 and 11). Since wood is one of the oldest materials used, many artifacts, from cradlesto coffins, which were used in the past have been discovered by archeologists. Thus, specific kinds of treatments have been developed to preserve waterlogged wood which has been stored under wet conditions such as burial in soil below the permanent water table, at the bottom of rivers or lakes, or i n a marine environment. More than 40% of all shipwreck losses in the Western Hemisphere have occurred due to ships wrecking in shallow water. Wood is a biological material; when it is submerged in the marine environment, it comes under immediate attack from the teredos, fungi, and different bacteria. The more wood items are exposed to salt water, the more they suffer and the more quickly some of them vanish. Oftentimes, wood-based materials, when buried deep under sediment, will suffer less, or not at all, and on occasion can be discovered in an excellent state of preservation. Sometimes, wood is completely preserved by being saturated by iron oxide while lying 807

808

Hon

close to iron objects under water, as they tend to become mineralized and hard in texture. Wood will also be preserved if it is saturated in fresh water. In 1997, Titanic was a popular movie which broke the box-office sale record for the century. The story was based on the S.S. Titanic, which went down in the Atlantic Ocean in April 19 12, eighty-eight years ago. On July 14, 1986, the sunken Tirctrlic was discovered, and more than 3000 artifacts have been lifted from the debris field. Contrary to many expectations, the deep ocean did not preserve the Titanic from decay. Organic materials such as wood, paper, and cloth all perished. The wood decking and furniture were nearly all gone. Media attention, in print and on television, was usually intense and at times frenzied, as anyone who remembers the discovery of the wreck of the Titanic in 1985 can attest. When the Titunic- broke the surface of the water after 87 years of marine burial, we can only imagine the depth of emotions experienced by those who were there. Hearts must have been bursting and tears flowing. The sight of the ship must have conjured up many different thoughts in the people watching: thoughts of its preciousness as an object of antiquity and the incredible tie it forms between the people of today and those 87 years ago; thoughts about the ship and its repository of information about past ship-building technology. To study shipwrecks is to study human history. To study wood in the past is also to study human activity. Such studies have brought to light cultures that existed long before written records, and have transformed featureless prehistory into a fascinating landscape of evolution, cultural change, and technological advance. They have delineated sequences of events in the past and have discovered and illustrated the course of human civilization. Because wood-based artifacts provide a rich and varied record of our early activities and technology, there is a great need to preserve them in the many forms that remain from ancient to modern times. It is not only because it will be interesting to future generations, but to use it to study the cultures, human behavior, and stratagems of intelligent human minds throughout history and to study the wood aging process itself. Cultural materials used by society exist in systemic context. Scientists and technologists study waterlogged woods, including shipwrecks, in archeological context in order to understand the systemic contexts of past societies. Preservation can be viewed as actions by a contemporary society to slow the rate of deterioration and destruction of cultural materials. The majority of waterlogged objects need to be preserved with special chemicals. If this is not done quickly and properly, the artifacts will rapidly deteriorate. The conservation and preservation of waterlogged wood tend to be expensive, require specialized knowledge and facilities, and be complex and time consuming. Preservation of waterlogged wood depends greatly on the condition of the wood and the conditions of the environment where the wood will be stored after preservation. Present techniques range from art to science. The history of preservation or conservation treatments is a story of success and failures [ l ] . Until recently, many techniques have been based on empirical approaches rather than hard scientific data. Because of the understanding of degradation mechanisms of wood and availability of new instrumentations [2], scientists are able to tackle preservation problems with more scientific approaches and thus acquire better results. For most waterlogged woods, preservation consists of both careful removal of water from the wood to minimize shrinkage, and introduction of a substance into the wood to improve its strength. Generally, the processes developed to preserve waterlogged woods are classified into two groups. One consists of dehydrating the wood first, before treating it with a consolidant. The process is commonly employed only on small objects, because careful removal of water from wood involves solvent exchange, a process which takes longer as the size of the object increases. The other group, which is employed with

Preservation of Waterlogged Wood

809

larger artifacts such as entire ships, uses injection of a consolidant first and dehydrating the artifact after treatment. These processes use water-soluble consolidants and introduce the consolidants into the wood through the time-consuming process of diffusion. In this chapter. many methods that have been developed to preserve waterlogged wood are reviewed. Additional information can be obtained from several excellent monographs [3S]. A case study of preservation of the historic gunboat U.S.S. Cairo is included.

II.

PROPERTIES OF WATERLOGGEDWOOD

As mentioned earlier, wood is a biological material that deteriorates in almost any environmental condition. If wood is kept in a very moist or wet environment, it will absorb water and eventually become waterlogged. Wood normally decays under combined biological and chemical attack when buried in the ground or submerged in water. Buried wood will generally not survive unless it becomes waterlogged to create the anaerobic conditions necessary for protection from decay fungi and inserts. Under this condition, wood is still vulnerable to chemical degradation and attack by anaerobic bacteria. In time, changes will occur in the wood structure which will affect its physical and chemical properties. These changes adversely affect the integrity of the wood. At the outset, the readily water-soluble extractives and mineral salts in the wood will diffuse into the surrounding medium, followed by readily hydrolyzed compounds such as the pectins and pentosans. Then a microbiological degradation of the more stable hemicelluloses and cellulose will follow. Finally, mainly lignin remains, which can also be decomposed slowly by microorganisms in the anaerobicenvironment. Both fungi and bacteria cause the microbiological breakdown of wood and both use extracellular enzymes to hydrolyze cellulose, hemicelluloses, and lignin. In the case of afungus breakdown, degradation spreads throughout the capillary system of the wood, at least as far as the extracellular enzymes can penetrate into this system. Because of the loss of polymeric components in the cell structure, voids increase throughout the cell and the wood becomes more porous and permeable to water. Hence, for highly waterlogged wood, it is not unusual to have a moisture content of over 800% (based on oven-dried weight). Thehygroscopicity of waterlogged wood tends to increase, and equilibrium moisture content values as much as twice those of recent wood have been found [6]. Because of increased hygroscopicity, deteriorated waterlogged wood also exhibits significant increase in shrinkage, reaching values of the order of 70% for volumetric shrinkage. The shrinkage in volume is found to be related linearly to maximum moisture content of wood 171. As long as the waterlogged wood is kept wet, it will retain its shape. If the wood is exposed to the air, it cannot redistribute internal moisture properly during drying due to various factors. The most important one is due to capillary tension collapse above the fiber saturation point in the deteriorated wood. Cell wall shrinkage occurs below the fiber saturation point because of desorption and resultant dimensional change [ 7 ] . Typically, moisture attacks the secondary cell wall, reducing the resistant strength and flexibility of the wood [g]. Hence, when the excess water evaporates, the resulting surface tension forces of the evaporating water cause the weakened cell walls to collapse. Such an action subsequently leads to significant warping, shrinkage, and cracking. Drying of highly deteriorated waterlogged wood therefore results in severe damage and distortion of artifacts. Shrinkage values of more than 30% in the tangential and more than 10% in the longitudinal directions have been recorded [6]. Moreover, because of the loss of major polymeric structural components in the cell, deteriorated waterlogged wood also shows

Hon

81 0

major losses in strength. To prevent shrinkage, stabilization treatments should be performed to control dimensional change as well as the strength and stiffness of the wood. The degradation of the wood also results in a decreased swelling capacity of the old waterlogged wood when compared to new wood. The amount of shrinkage upon drying occurring in old waterlogged wood appears to be correlated with the amount of degradation of the wood as reflected in its chemical composition. It should be borne in mind that wood can be preserved for long periods of time under anoxic, waterlogged conditions. Thus, ideally, no submerged sites should be excavated unless archeologists can guarantee a proper preservation method for the recovered artifacts. Or, after excavating and surveying, the waterlogged wood should be left or reburied and the site preserved in situ, rather than “collecting” items that often deteriorate in air. These simple approaches would allow us to preserve cultural heritage, without all of the expense of conservation risk to the archeological or waterlogged materials. To do this effectively we must be able to measure how stable the archeological material is within a site.

Ill. PRESERVATIONTREATMENTS Conservation treatments for waterlogged wood have been designed to prevent dramatic dimensional changes caused by cell collapse and cell wall shrinkage during drying. Wood artifacts to be conserved are usually physically weak and chemically very complex. The most active phase of a conservation treatment is in controlling the processes as the waterlogged wood is transferred from its deposition to its new “home.” Nonnally, one of the first steps that must be taken is to remove contaminants such as sea salt and iron sulfide which will be unstable in the new environment and damage the wood’s structure. Hence, the specimens must be first cleaned and kept wet after removal from the discovery site. If they are allowed to dry, they will shrink and split. Normally it is recommended to let wood soak for 1-12 months in fresh water, depending on the size of the artifact. Hardwoods can handle a mild solution ( 5 % ) of muriatic acid. In this case, no steel or iron should be involved, as muriatic acid will destroy these items. After muriatic acid is used, the wood must be soaked again in fresh water. This will help remove the smell of acid. Preservation of waterlogged archeological wood involves stabilization a s tosize, shape, and durability. The state of the wood surface as seen from an esthetic viewpoint is also taken into consideration in preservation, whereas questions of restoration, completion, and related aspects are considered outside of preservation. Most if not all processes developed to preserve waterlogged wood have as their objectives to achieve dimensional stabilization and improvement of wood strength. Wood dimensions can be stabilized by three different means or combination of these means. These are (1) reducing the hygroscopicity of the wood so that less water can be taken up, ( 2 ) forming crosslinks between cellulose chains so as to minimize separation of these units, and (3) depositing a bulking agent within the swollen wood so as to reduce shrinkage 191.

A.

On-Site Preservation

On-site preservation consists of preservation measures that are employed at the excavation site. Immediate treatments are usually aimed at preventing shrinkage and further deterioration of the wood before it can be treated at a conservation facility. Techniques include spraying with a 5% solution of sodium borate, boric acid, or any other wood-preservative

Preservation of Waterlogged Wood

811

to arrest and prevent further attack of deteriorating organisms, immersing in fresh water, or covering with plastic foam sheets saturated with water to prevent drying and shrinkage.

B. Classic Impregnation Treatments One of the earliest methods for treating waterlogged wood was to impregnate it with a mixture of petroleum and a drying oil, such as linseed oil, while the wood was allowed to dry slowly. Most of the drying oils do not penetrate well, and they tend to deteriorate in time. The treated object becomes sticky and dark. Glycerin had been used in the early 1990s. It was used to replace water in the wood. It evaporates extremely slowly at ambient conditions. Since air does not enter the wood, the shrinkage which is caused by the surface tension at the interface between air and water does not occur. Glycerin is a highly hygroscopic material; it absorbs and releases moisture as the atmospheric humidity changes. It gives the treated object a wet and sticky appearance. It also does not strengthen the wood. Aluminum potassium sulfate or alum has also been used. When waterlogged wood is treated with alum in water solution, the potassium, aluminum, and sulfate ions will diffuse into wood. When the wood is cooled, alum recrystallizes to function as a bulking agent to prevent wood from shrinking. The treated object may be brushed with warm linseed oil and a thin coat of shellac to prevent reabsorption of moisture from the air. Unfortunately, alum treatment only reinforces a thin surface layer of wood.

C.

Consolidation Using Water-Borne Chemicals

Consolidation using water-borne chemicals relies on strengthening the wood structure through the introduction of a consolidant into the cells. The method utilizes chemicals that are soluble in water and are introduced directly into the wood without drying the wood. Consequently, preserving an object using water-borne chemicals does not take as long as using resins or other consolidants that are not water-soluble. Nevertheless, the treatment time is governed by the rate of diffusion of consolidant through the wood to where it is required. These water-borne chemicals include sugars,salts,alum, polyethylene glycol (PEG), tetraethyl ortho silicate (TEOS), and phenol. Of these, PEG is the only one which has gained wide acceptance among conservators. PEG has been considered since the 1950s to be the most reliable method for treating waterlogged wood. PEG is a synthetic polymer of ethylene oxide with the general formula HOCH2-(CH20CH2),,-CH20H, where n represents the average degree of polymerization. PEGs are available in molecular weights (MW) ranging from 200 to 6000. The fractions with MWs ranging from 200 to 600 are liquid at 20°C and are soluble in water inall proportions. Those with MWs above 600 are white and waxy, with the higher MWs being more viscous than the lower MWs. PEGs with higher MWs are soluble freely in alcohols, such as ethanol and methanol, as well as water. For treating waterlogged wood, the low-MW PEGs are dissolved in water or alcohol and diffuse into the wood slowly at 60°C for a period of several days to weeks to replace water in the wood. Then the next range of higher-MWs PEGs is successively and gradually introduced into the solution and allowed to diffuse into the wood to replace the low-MW PEGs. Apparently the low-MW PEGs have higher rates of diffusion and penetration into the cell wall. Eventually, higher-MW PEGs (>3000) are used. The size of the PEG increments is dependent on the condition, size, and species of the wood. While low-MW PEGs are most effective for less degraded wood, higher-MW PEGs tend to be more suitable for highly degraded wood. In any cases, it is a very time-consuming process. Fungicides may

812

Hon

be used simultaneously with PEG. After PEG treatment, the excess wax is wiped off and the wool is allowed to cool. When cooled, any excess that was on the surface is removed with a hot-air gun or with hot water. Alcohols may finally be used to clean the surface to remove dark color and to regenerate woodlike color. When the Swedish warship Wasrl was resurrected in 1961 after 333 years at the bottom of Stockholm Harbor, PEG was selected to treat the entire vessel [IO]. Even though PEG treatment is a popular method, it has several significant drawbacks. It is a relatively expensive and time-consuming process. Perhaps most significantly, PEGtreated wood is sensitive to heat and high humidity. These unfavorable characteristics require that treated wood be curated in climate-controlled facilities.

D. SucroseTreatment To overcome the PEG-related limitations, sucrose has been developed as an alternative method [ 11. In practice, the procedure is identical to that described for PEG. where sucrose is applied by aqueous diffusion as a bulking agent. Itis less expensive and penetrates wood quite well. For better results, refined white sugar (white sucrose) is recommended. Unbleached, brown-colored sugar should be avoided because of its high hygroscopic property. For less degraded waterlogged wood, a sucrose solution with a sufficiently low concentration ( 1 - 5 % ) may be used to avoid dehydration of sound wood. For highly degraded wood, a higher concentration of sucrose may be used. It is recommended to start with low percentage increases, for example, 1-596, until a concentration of 50% is reached. Then the solution can be increased in 10% increments. The treatment continues until sucrose concentration reaches 70% and the wood has equalized at this concentration. During the treatment, an antimicrobial agent may be added to the first batch of solution. When the wood has reached equilibrium with the highest solution concentration, the treated wood is air-dried slowly under conditions of controlled high humidity. The properly dried wood should be, if possible, stored under conditions of less than 70% humidity. Sucrose-treated wood has the appearance, density, and much of the strength of nondegraded wood.

E. Acetone-RosinTreatment Acetone-rosin treatment involves the exchange of free water in the waterlogged wood with a natural rosin, such as pine rosin. In this method, the wood is dehydrated completely with acetone or a mixture of acetone and ether [ I 1 , l 21. It is important to remove all the water because it is not compatible with rosin. The well-dehydrated wood is placed in a saturated solution (67%) of rosin dissolved in acetone at about 50-55°C. After treatment, treated wood is removed from the solution and the excess rosin is wiped off, This method can give excellent results, but it involves heating volatile flammable solvents and is difficult to control. Acetone is also an expensive solvent. It is recommended only for treating small artifacts.

F. Alcohol-EtherTreatment In alcohol-ether treatment, waterlogged wood isfirst immersed in successive baths of alcohol until all the free water is replaced by either isopropyl alcohol or ethanol. This is followed by successive baths of acetone. The treated wood is then dried under vacuum to remove ether. Since ether has a very low surface tension (0.7 dynekm), when it evaporates,

Preservation of Waterlogged Wood

813

the surface tension forces are so low that there is no appreciable collapse of the weakened cell wall. Like acetone, alcohols and ether are highly flammable solvents, and they must be handled very carefully. After the water is replaced by alcohols, camphor can be used instead of ether. After the camphor fills the cavities and cell walls of the waterlogged wood, the camphor then slowly sublimates without exerting any surface tension on the cell walls. Thus the wood does not collapse, shrink, or distort.

G.

In-SituPolymerization

PEG is a polymer. It takes time for the polymer to diffuse or penetrate into wood cells. To overcome this problem, considerable work has been done in an attempt to develop monomers which can be polymerized in situ after infusion [ 13,141. The polymerization process maybe initiated using either heat or high-energy radiation to convert monomer into polymer. The polymer functions as a consolidant, providing strength and dimensional stability for the degraded wood cells. Many monomers can be used for this purpose. Styrene, vinyl acetate, acrylonitrile, acrylates, and methacrylates are among the most commonly used monomers. Melamine formaldehyde resin has been used quite successfully for this purpose. After infusion, copolymerization of unsaturated polyester oligomeric resins with styrene by high-energy radiation has been employed successfully. The treated products are mechanically strong. durable, and stable to a wide range of environmental conditions [ 131.

H.Freeze-DryingTreatment Freeze-drying is not actually a conservation process but rather only a dehydration process. It is a physical process which sublimes ice. Water in the wood is frozen and leaves in the form of gas without actually passing through the liquid state. There are several variations of this process, including directly freeze-drying the wood, freeze-drying after impregnation with a consolidant, freeze-drying after exchanging water in the wood with another solvent, and consolidating the wood before freeze-drying and then freeze-drying in a natural environment [ 15,161. The technique is not as straightforward as it might seem to be, as there is a considerable reduction in density when water freezes, and hence a physical expansion of the material within the specimen. Accordingly, for freeze-drying to be effective in treating waterlogged wood, a cryoprotectant must be added. The purpose of a cryoprotectant is to reduce the volume change by continuing to increase in density as the mixture cools through the water’s freezing point. The most commonly used cryoprotectant is PEG with a MWof 400. I t is usual for waterlogged wood to be i n equilibrium with a 20% solution of PEG prior to commencing the freeze-drying process. Vacuum freeze-drying to remove water from waterlogged wood has been done at several conservation laboratories. Of the many methods tested, pretreatment with PEG 200 and 400 worked well with undeteriorated wood. Highly deteriorated woods responded best to pretreatment with a combination of higher-molecular-weight PEG or PEG dissolved in solvents such as f-butyl alcohol [ 171. Directly freeze-drying wood was the process chosen in the conservation of some parts of the Mary Rose [ 5 ] ,which was a ship built as part of King Henry VIII’s naval expansion program.

1.

SupercriticalDrying

The supercritical drying method was invented in the 1950s by Kistler 1181. This method uses a high-density or supercritical fluid to replace the water in the wood. The supercritical

Hon

814

fluid is then removed from the wood by decompression, without forming a liquid phase. As shrinkage is due to surface tension forces at a liquid surface, supercritical drying does not damage the artifact. Carbon dioxide is a suitable supercritical fluid that can be used. For treatment, water in the wood is first replaced with an organic solvent or alcohol. At high pressure the carbon dioxide’s density is similar to thatof common liquids, and it readily dissolves the alcohols.

IV. PRESERVATION OF HISTORIC GUNBOAT A CASE STUDY1221 A.

U.S.S. CAIRO.

Background and History of Chemical Treatments [ 19-21]

The U.S.S. Cairo was constructed at Mound City, Illinois, in the fall and early winter of 1861. It was 175 ft in length, 5 1 ft in breadth, and 15 ft in height. The vessel was a flatbottomed, three-keeled ship, designed for use solely on Western rivers. Theship was constructed of heavy timber, mostly white oak, with additional protection provided by 2.5in.-thick armor plates to protect the front casemate, sides abreast machinery, and the pilothouse. Approximate tonnage of the vessel was 512 tons, and it was manned by a crew of 174 officers and seamen. The U.S.S. Cairo served the Union forces for approximately 1 1 months, from January 1862 to December 1862, when she was sunk by a submerged Confederate“torpedo” (mine) in the Yazoo River near Vicksburg, Mississippi, on December 12, 1862. The U.S.S. Cairo had retained virtually all its structural integrity beneath the muddy left bank of the Yazoo riverbed when it was discovered in 1956. Following a three-year hiatus of exploration, the exact site of the Cairo was determined conclusively and the gunboat was found to be in an apparently excellent state of preservation. In 1960, a gun port, winched from its mounting on an inclined casemate, the pilothouse, and a naval cannon were the first items to be raised. However, the entire vessel was not raised until 1964 by the State of Mississippi. During this process, unfortunately, the vessel was heavily damaged and ultimately was cut into three pieces to get it ashore, i.e., to the barge. The immediate need for preservation of wooden objects was recognized. A huge polyethylene tank (45 ft X 7 ft X 4 ft) was made, in which smaller wooden artifacts were preserved with 25% polyethylene glycol (PEG-1000) solution. Sodium salt of pentachlorophenate was also added to the solution. No preservation or protection of large wooden objects was done. In 1965, the Cairo was barged to Pascagoula, Mississippi, for mock-up and restoration. The bulk of the wooden remains of the Cairo was comprised of approximately 17 large separate wood sections after the gunboat was moved to the Ingalls’ shipyard at Pascagoula. After the armor was separated, the wooden parts (ranging in size from whole sections of the casemate to single planks) were stored and stacked. Due to the high cost of polyethylene glycol, the wooden fabric was placed under sprinklers of tap water for occasional “sprinkling treatment.” This practice was terminated in the early- to mid- 1970s, but the Cairo stayed outdoors without any protection at Pascagoula until 1977. In that year, the Cairo was removed to the National Military Park at Vicksburg, Mississippi. In the course of preparing the Cairo for transport, workers discarded unidentifiable, highly deteriorated lumber that had separated from the gunboat sections reassembled in 1965. In order to facilitate truck transit to the park, many of the 17wood sections were further divided by chain saws into smaller sections. By the time the Cairo arrived at Vicksburg National Military Park, there were approximately 27 separate significant wood sections. All of the wood sections that were transported to Vicksburg were sprayed with a hydrozol

Preservation Wood of Waterlogged

815

5% pentachlorophenol solution before shipment.No further chemical stabilization of wood fabric was done until 1979, when the National Park Service decided on spray treatment of the Cairo remains with polyethylene glycol (PEG 4000) and copper-o-quinolinolate (PQ-57) solution six times between 1979 and 1982. Since then, the Cairo has been occasionally treated with insecticide. Meanwhile, a shelter with a covering structure for the gunboat was completed in 1980. B. Specimens from the Gunboat Specimens from various location of the gunboat, as shown in Fig. 1, are listedbelow. They were collected for chemical analysis. 1. Port (Right, Upper,Forward) 2. Port (Right, Lower,Forward) 3. Hurricane Deck (Left, Under,Forward) 4. Port (Left,Upper, Forward) 5. Port (Right, Upper,Middle) 6. Gundeck (Right, Upper,Middle)

FIGURE 1 Specific locations of wood specimens collected from the U.S.S. Cairo, Vicksburg Military Park, Vicksburg, MS.

816

Hon

7. Gundeck (Right, Upper, Middle) 8. Hurricane Deck (Left, Under, Middle) 9. Gundeck (Center, Upper, Middle) 10. Hurricane Deck (Left, Under, Middle) 11. Hurricane Deck (Left, Under, Middle) 12. Hurricane Frame-Cross Section (Left, Under, Middle) 13. Gundeck (Right, Upper, Forward, near Torpedo-damaged bow) 14. Gundeck (Right, Upper, Forward, near Torpedo-damaged bow) 1 S . Gundeck. Side Plank (Right, Upper, Forward. near Torpedo-hit site) 16. Gundeck (Right, Upper, Middle) 17. Gundeck (Center, Upper, Middle, in front of Paddlerwheel) 18. Hurricane Deck (Right, Under, Middle) 19. Hurricane Deck (Center, Under Paddlewheel, Middle) 20. Gundeck, Side Plank (Left, Under, Aft) Most of the wood elements of the Cairo are deteriorated-some to a critical state. For example, the wood fabric located in the port areas is light in weight, very brittle, and fragile. Planks located at the gundeck and hurricane deck are significantly discolored and weathered, but have retained their strength. Portions of deteriorated wood regularly fall from the structure and canbefound on the floor under the boat. In general, the wood located on the upper part of the gunboat is suffering more degradation than that at the bottom. This implies that the degradation process continues to proceed due to weathering and other environmental factors. The bulk of the remaining Cairo fabric is white oak that is essentially sound. However, there is considerable surface decay caused largely by biological agents. There is also some chemical decay and significant weathering. The deep cracks in most members are usually the result of repeated wet/dry cycles. These checks can present serious preservation problemswhere they have allowedmoisture to penetrateto the timbercenter, thereby causing internal pockets of decay. The chemicalcomposition of the Cairo fabric was analyzed for the specimens collected from the Cairo. Scanning electron microscopy was also used to examine the ultrastructures of the wood specimens. Details of these studies will be discussed in a subsequent section. Chemicalcomposition of wood specimens from the Cairo wasanalyzed using a variety of laboratory techniques.Holocellulose,alpha-cellulose, Klasson lignin,extractives, and ash were determined based on wood chemistry standard methods described in the literature. A scanning electron microscope (SEM; JEOL JSM-IC848) was used to study the ultrastructure of the wood specimens.

C. Wood Identification The wood used in the construction of the U.S.S. Cairo was identified as a species of the white oak group (QuCrcu.7 sp.). Some of the gundecks of the ship were made of southern yellow pine (Pirzus sp.). It is not possible to identify the specific species of these group of woods on the basis of their wood anatomyalone; fruit and flowers are needed for positive species identification.

1. Chemical Analysis Chemical analysis of specimens from the gunboat revealed that the wood fabric is severely degraded, particularly at the surface layers. Chemical composition of the wood is tabulated

Preservation of Waterlogged Wood TABLE 1 Sample no.''

817

Chemical Composition of Deteriorated Wood Fabric of U.S.S. Ctriro Alpha-

Holocellulose cellulose

82.27 55.23 56.84 61.99 47.96 66. I0 47.07 62.04 45.48 6 1.92 49.65 67.67 48.43 66.9 I 5 I .68

47.68 23.09 18.29 27.90 -

40.46 -

36.17 -

36.97 -

35.93 -

33.10

Klason lignin

Ethohenzene extractives extractives

1% NaOH

20.76 23.39 29.20 22.55 25.98 25.97 29.68 27.06 33.2 I 24.69 25.28 22.47 28. I 1 23. I 1 27.19

3.67 10.06 5.25 5.03 14.95 5.04 12.77 4.59 11.51 8.53 15.25 5.41 12.36 6.88 10.06

18.76 52.12 55.44 47.26

Ash

-

28.15 66.18 34.78 59.30 37.9 1 54.16 33.26 56.94 34.47 56.13

0 . I3

2.0 I 1.25 1.42 1.85 0.60 0.93 1.10

2.67 0.9 I 1.30 1.43 1.77 1.95 2.69

in Table I . It is apparent that cellulose and hemicelluloses content have decreased significantly. Lignin appeared to suffer less degradation than cellulose. (The increase in lignin content was due to the loss of cellulose and hemicelluloses.) Reduction of degree of polymerization of cellulose is shown in Table 2. It also revealed that low-molecular-weight fragments of detcriorated products can be extracted with co-solvents of ether and benzene and with 1 % sodium hydroxide solution.

2. Scanning ElectronMicroscopy Examination The wood of white oak is made up of six types of cells: springwood vessels. summerwood vessels, vasicentric tracheids, longitudinal parenchyma, ray parenchyma, and libriform fibers. SEM micrographs of the cross section of a normal piece of white oak show extremely thick-walled fibers which make up the highest proportion of the cell types in white oak. The vessel elements, vasicentric tracheids, and parenchyma arc generally thin-walled cell. SEM micrographs of the wood from the U.S.S. Cairo show considerable deterioration and reduction in the secondary walls of the thick-walled fibers as well a s complete deterioration of some of the thin-walled ccllular elements (i.e., latewood vessels and vascicentric tracheids). The deterioration which has occurred to the anatomical structure is due in part to fungi, evidenced by the presence ofan abundance of hyphae (Figs. 2 and 3) and bore holes in the cell walls, as well as possibly some bacteria (Fig. 2).

D. Consolidation of Historic U.S.S. Cairo Gunboat

In order to restore or improve the strength of the critically deteriorated, once-waterlogged gunboat, conventional polyethylenc glycol trcatments are undesirable because the wood structure has already shrunk and collapsed. I t is very difficult for polyethylene glycol to penetrate into such wood. Likewise, the use of borate has the same problem for oncewaterlogged wood. In addition, borate is too casily dissolved in water. It will leach out

3

Hon

818

TABLE 2 Changes in Degree of Polymerization of Cellulose from the U.S.S. Cairo Sample no.'

Degree of polymerization ~~

Control 1

4 5(c) 6(c) 9(c) 9(s) 17(c) 17(s) 19(c) 19(s) 20(c) 20(s)

1471 467 934 1634 770

957 1447. 1027

'See Fig. 1 for location of speclmens; c, core; S, surface.

FIGURE 2 Springwoodzoneshowingspringwoodvesselelements in lowerrightcornerwith tyloses and smallroundocclusionswhicharepossiblybacteria.Secondarywall of allcellsis practically nonexistent. Magnification 15OX.

Preservation of Waterlogged Wood

819

FIGURE 3 Cross section of libriform fiber tissue showing extreme deterioration of the secondary walls of the fiber. Normally these cells would have extremely thick walls with a very small lumen. Magnification 300X.

readily from wood in a high-humidity climate. The addition of polyethylene glycol and borate would only increase the risk of further damage of the ruined fabric. One hundred and two years of submersion in the Yazoo River as well as a decade-long weathering in Pascagoula and Vicksburg have brought the gunboat to a state of extreme fragility. As discussed earlier, the gunboat has lost most of its polysaccharides(i.e., cellulose and hemicellulose) and some parts of its lignin. Loss of these chemical components and physical stress have resulted in the formation of cracks, checks, crevices, and cavities in the wood structure at the macro- and microscopic levels. Given these conditions, the buildup of polymer consolidants within the cell walls appears to be the most acceptable approach to restoring fabric solidity. Six different types of polymer resins were selected to consolidate the deteriorated wood samples. The resins were phenol formaldehyde, resorcinol-phenol-formaldehyde, epoxy, polyacrylic, polyurethane, and wood rosin. With the exception of wood rosin, all the polymer consolidants favorably restored the strength of the wood fabric. Of these polymers, only resorcinol-phenol-formaldehyde resin exhibited acceptable appearance, i.e., color and texture of wood after impregnation; wood treated withother resins displayed a glossy surface which is not acceptable for restoration and conservation purposes. The resorcinol-phenol-formaldehyde-treated wood specimens exhibited significant improvement of hardness, reduced water pick-up, and good dimensional stability. 1. Resorcinol-Phenol-Formaldehyde Treatments Various concentrations of resorcinol-phenol-formaldehyde resins can be prepared by using water and alcohol as the diluents, as shown in Tables 3 and 4. Several solutions were

Hon

820

TABLE 3 Alternative Formulations of Resorcinol-Phenol-Formaldehyde Resin (R301)" for Application to Wood R30

Sample

I

Hardener Viscosity Alcohol Water

100 100

I 00 100 I00

20

16 30 30 30 20

"K301 is a rcsorclnol-phenol-formaldehyde "cp = centipoise.

0 0 0 0 I20

80 80 120 I60

120

236 76 36 52

rcsin manufactured hy Koppcrs Company. Inc.

used in tests in which wood fabric was treated-each with a given solution and cured at room temperature. Allof the formulations worked well with the wood fabric to achieve good consolidation. Strength was improved. Original color and texture were retained. It was found that the more alcohol used, the darker the color of wood surf'ace resulted. The use of filler (peanut flour) could fill the large gaps and cavities properly. Scanning electron micrographs showed that the resorcinol-phenol-formaldehyde resin deposited on the surface of cell lumens properly to achieve good consolidation. N. Chamcteristics qf Resorcinolic Resins. Resorcinolic resins are condensation products of resorcinol with formaldehyde, or with various phenol-formaldehyde resoles. The latter were used for this study due to their much lower cost. Although the resorcinolic adhesives were initially found to be capable of being efficiently cured under neutral conditions, it has generally been found to be advantageous for commercial purposes to cure them under mildly alkaline conditions. Basic catalysts may provide bonds of strength at low curing temperature, such as that used in this work. Resins are prepared from resorcinol by reaction with efficient amounts of formaldehyde. The theoretical amounts necessary to produce a cured resin are slightly in excess of 1 molof formaldehyde per mole of resorcinol. However, even 80% of this amount would be sufficient to give an unstable or gelatinous product, since the molecules react disproportionately in bulk situations, so that

TABLE 4 Altcrnative Formulations of Resorcinol-Phenol-Formaldehyde Resin (G 1260-A)" for Application to Wood Samplc

1 2 3 4 S

6 7 8 9 10

G 1260-A

Hardcner

Water

Alcohol

100 100 100 100 I 00 1 00

20 20 20 20 20 20 20 20 20 20

0 0 S0

30 60 0

IO0 100 100 100

IS

18

25 35

2s 35 40 20

40 h0

80 80

IS 0

(cP)" Viscosity 280 70 200 280 120 72 64 S2 64 -

Preservationof Waterlogged Wood

821

OH

OH

7

OH I

CH2OH

FIGURE 4

Substitutionreaction in a resorcinol structure.

some of the resorcinol molecules are left unreacted, while some of the first-formed oligomers acquire a superior share of the formaldehyde to form polymers of higher molecular weights. Resorcinolic resins of between 0.5 and 0.7 mole of formaldehyde per mole of resorcinol are made having infinite stability. At the point of eventual use, some additional formaldehyde is provided and the resin is converted, within a short period of time, to a very highly cross-linked resin. This product is characteristically insoluble, infusible, and physically strong when properly aged. b. Chemistry of' Resorcinol- Fornlaldehyde Resin Formation. Resorcinol readily combines with formaldehyde to form methylol derivatives, with the methylol groups occupying either the positions ortho to both hydroxyl groups, or ortho to one and para to the other. The meta position is not ordinarily reacted (see Fig. 4). The reactivity of these methylol derivativesis so high that they cannot easily be isolated in pure stable form. They continue to react under ambient, uncatalyzed conditions with formaldehyde,resorcinol,phenol, or othermethylol-containingmolecules to form polymer chains of higher molecular weight, with branched, as well as linear, configurations of great complexity. These reactions continue until spatial considerations prevent further interaction. In these polymers the resorcinol nuclei are joined together through methylene linkages to give complex molecules as shown schematically in Fig. 5 . c. Resorcinol-Pherzol- Fr~rnluldehyne Resin. In orderto produce resorcinol-phenol-co-polymer resin, the phenol is combined with the formaldehyde before the resorcinol is introduced. If the resorcinol were added to the initial charge, it would preempt most of

&

CH,---

CH,2H:+

OH

FIGURE 5

&CH2

OH

9 CH,---

OH

CH,---

CH2---

A typical polymcric structurc of resorcinol-formaldehyde rcsin

822

Hon

the formaldehyde and form a gel, because it is many times as reactive as phenol. As a result, most of the phenol would remain unreacted. In order to obtain a block co-polymer, phenol would be combined withformaldehyde in one reactor to form a resole, and resorcinol would be combined with formaldehyde in a second reactor. The two resins would then be mixed, and the methylol groups on the resole would combine with available nuclear positions of the resorcinol to form mixed polymers. The other way to do it is to let phenol react with an excess of formaldehyde to form a low-condensed resol. Resorcinol would then be added in sufficient proportion to combine with all of the methylol groups, thus forming a resinous co-polymer.

2. ScanningElectronMicroscopyExamination Since the resorcinol-phenol-formaldehyde resin consolidates wood fabric successfully, it is assumed that the penetration and solidification of the polymer in the cell walls is efficient. SEM was used to studythe characteristics of resin treated wood fabric. Typical SEM pictures are shown in Figs. 6-9. Figures 6 and 7 showed that the cross section of the red oak was coated with resorcinol resin. Some of the small cells still can be seen. Figure 6 shows that most of the cells in the cross section were filled with resins. A large vessel on the left was also filled. Vessels on the right were still left empty. Figure 8 shows the tangential surface aspect. Some of the procumbent cells in the body of the ray still can be seen, but are heavily coated with the resin. The surface of the radial section is also heavily coated with resorcinol resin, in which some of the pit cavity was filled with resin too (Fig. 9).

FIGURE 6 SEM micrograph of the cross section of red oak fabric of the gunboat, treated with resorcinol-phenol-formaldehyde resin. The vessel at the left was filled with resin. The surface of the lumen was also coated with the resin. Magnification 250X.

Preservation of Waterlogged Wood

823

FIGURE 7 SEM micrograph of the cross section of red oak fabric of the gunboat, treated with resorcinol-formaldehyde resin. The surface was heavily coated with the resin. Many small cells still can be seen. Magnlfication 250X.

FIGURE 8 SEM picture of the tangential section of red oak fabric of the gunboat, treated with resorcinol-formaldehyde resin. One of the ray cells at the left was filled with resin. Magnification 250X.

824

Hon

FIGURE 9 SEM micrograph of radial section of red oak of the gunboat, treated with resorcinolformaldehyde resin. The surface was heavily coated with the resin. Some of the pits were entirely coated. Magnification 250X.

It is obvious that resorcinol resin can penetrate, wet, and consolidate fabric favorably to increase the strength of the deteriorated fabric. It may serve as a useful resin to preserve the U.S.S. Cairo gunboat.

3. Large-scale Treatment of U.S.S. Cairo Gunboat Fabrics In order to evaluate alternative treatments ofwood fabric, initial tests were conducted using small samples. Since resorcinol-phenol-formaldehyde resin exhibited positive effect on consolidation, a large-scale consolidation with resorcinol-phenol-formaldehyde wasattempted. Wood specimens with dimensions of 12 cm X 12 cm X 30 cmwere treated. Similar proportions of resin were used as for the small-scale study. Results are very impressive. The penetration of resins into the wood fabric was good, and the appearance, hardness, dimensional stability, and water pick-up properties were excellent. It is obvious that by proper treatments, mechanical and physical propertiesof Cairo gunboat fabric can be successfully consolidated. The improvement of 62.6%and 78.7% of hardness for severely decayed wood specimens and moderately decayed wood specimens, respectively, was a good demonstration.

REFERENCES 1.

2.

D. W. Grattan and R. W. Clarke, in Conservation of Marine Archaeological Objects (C. Pearson, ed.). Butterworths, London, p. 164 (1987). J. I. Hedges, in Archaeological Wood (R. M. Rowel1 and R. J. Barbour, eds.), American Chemical Society, Washington, DC, p. 111 (1990).

Preservation of Waterlogged Wood 3. 4.

5. 6. 7. 8. 9.

IO. 1I.

12.

13.

14. 15. 16.

17. 18. 19.

20. 21.

22.

825

R.M.Rowell and R. J. Barbour (eds.), Arc~htreologic~rrl Wood, American Chemical Society. Washington. DC ( 1990). D. L. Hamilton. Rnsic Metl~odsof’ Cor~.servir~~g Urlder-rvnterArr./~rreolo~~icrr/ Matcricrl Crrlt~rre. U.S. Department of Defense, Legacy Resource Management Program,Washington, DC ( 1 996). B. B. Christensen, The Conser~~crtior~ of Wcrterlogg~rl Woodi r l the Ntrtiorltrl M I I S P I I I~fI I DCIIm~rrk,The National Museum of Denmark. Copenhagen ( 1970). R. J. Barbour and L. Leney, in Proc. ICOM Wtrterloggecl Wood Workiug Grorcp Conf: (D. W. Grattan and J. C. McCawley, eds.), International Council of Museums. Ottawa (1982). P. Hoffmann, Strrd. Corlser-l~., 3 1 : 103 (1986). L. F. Hawley, Wood Liquid R&tiom, USDA Tech. Bull. 248. Washington, DC ( 1937). A. J. Stamm, Dimensional Stabilization of Wood with Carbowaxes, Forest Prod. J . , 6(5):201 (1956). B. Hafors. in Archaeologiccrl Wood (R. M. Rowell and R. J . Barbour, eds.), American Chemical Society, Washington, DC. p. 195 ( 1990). H. McKerrel. E. Roger, and A. Varsanyi. Strrtl. Cor~ser-~~.. 17:lIl (1972). T. Bryce, H. McKerrel,and A. Varsanyi, in Prob1ml.s i n the Cor~.s~rl.ntior~ of’ Wrrterloggc~d Wood, Marine Monographs and Reports No. 16 (W. A. Oddy). National Maritime Museum, London ( 1975). Q.-K. Tran, R. Ranliere, and A. Ginier-Gillet, in Archtreologicd Wood (R. M. Rowell and R. J. Barbour. eds.). American Chemical Society, Washington, DC, p. 217 (1990). R. W. Clarke and J. P. Squirrel, in Proc. ICOM Wtrterlo~ggerl Wood Workir~~g Gro~q’Con$ (D. W. Grattan and J. C. McCawley, eds.),International Council of Museums, Ottawa, p. 9 ( 1982). W. R. Ambrose, in IIC Ne,$. York Corlfc~rel1c.e0 1 1 ComcJr-wrtiorlof‘ Stone a r l d Woodcw Ol?jrc.t.s, 2nd ed.. 25.3 (1970). J. Watson, in Proc. ICOM Wrrterlogged Wood Working Gmtrp Cor!/: (D. W. Grattan and J. C. McCawley, cds.), International Council of Museums, Ottawa, p. 19 (1982). R. J. Barbour, in Arr~hrrrolo,gir~rr/ Wood (R.M.RowellandR. J. Barbour, eds.), American Chemical Society. Washington, DC, p. 176 (1990). S. Kistler. Nlrtrrre. 127:741 (1931). E. C. BCXSS, Htrrd1rrc.k 11-orlclncl: Tllc. Sir~kirzgt r r d Srrll1age of’ the Ctriro, LouisianaState University Press, Baton Rouge and London ( 1980). V. C. Jones and H. L.Peterson. The Sro~;v of’ ( I C i ~ ~War i l Grrnhoot: U.S.S. Ct1;r.o. U.S. Department of Defense, Washington. DC ( l97 l ). T. McGrath and D. Ashley, Historic Structure Report: U.S.S. Cairo, Vicksburg National Military Park, Vicksburg. Mississippi. U.S. Department ofInterior, Denver Service Centcr.Denver, C O (1981). D. N.-S. Hon and M. A. Taras. A.ssc~.s.srnc~r~t of‘ USS Clriro t r r r d R c c , o ~ l l ~ ~ ~ e r l t l f rft bj [r’ ~Preser~.s lwfintl ~ ~ ~ ~ I I I I ~ ~Vicksburg I I / , s , National Military Park, Southeast Region, National Park Service ( 1988).

This Page Intentionally Left Blank

Biodegradable Plastics from Lignocellulosics Mariko Yoshioka and Nobuo Shiraishi Kyoto University, Kyoto, Japan

1.

INTRODUCTION

During the past 50 years, synthetic polymers utilizing petroleum as their raw material have been advanced. A wide variety of plastic materials is now being used commercially that supportour daily life with considerablecomfort. Even plastics which have properties similar to those of metals have appeared as engineering plastics. Before the start of the synthetic polymer industry, there were a number of attempts to obtain moldable materials from natural polymers, mainly from cellulose. Trials of developing cellulose derivatives into industrially acceptable materials [ I ] , as well as efforts to identify excellent plasticizers for cellulose acetate (CA) [2,3], provide good examples. However, these effortsbecame unpopular with thestart-up of theindustrialization of petrochemistry. Actually, many synthetic polymers exhibit properties that polymers of natural origin do not possess, especially in relation to melt processability. Many opinions expressed i n textbooks claim that cellulose has such a rigid backbone that it cannot be converted to plastic materials. Because of these circumstances, it becomes understandable that attempts to develop plastics of natural origin have not been emphasized during the past half-century. Recently, however, several changes have occurred and provided motivation for the authors to start and continue studies on the conversion of biomass into plastics. The first change concerns gradually growing demands for circulating materials which are desirably originated from biomass. One of the works of the authors’ group must be includable, in which wood could be converted into thermally flowable material by chemical modification, such as esterification and etherification [4-71. Thatis, plastics couldbeobtainedfrom suchlow-cost materials as wood wastes. At the present, there are actually almost no sophisticated methods or technologies generally available that can make use of biomass wastes for the purpose of adding satisfactory value.Thus, wood plasticization canbe considered as one attempt to pursue recycling technology. The second motivation resulted from a recently occurring requirement for developing biodegradable plastics. To meet this need,development of biodegradable plastics from natural polymers becomes attractive, together with that of bacterial polyesters and synthetic 827

828

Yoshioka and Shiraishi

polymers (aliphatic polyesters and water-soluble polymers). Many reviews concerning the biodegradable plastics have appeared recently [S- I O ] . Among these biodegradable plastics, investigations in the field of bacterially produced polymers and synthetic polymers are more actively and extensively pursued than polymers from natural origin, especially those from lignocellulosic and cellulosic materials. Actually, even in recently published books on biodegradable polymers, it is difficult to find descriptions of lignocellulosic biomass wastes being converted to meaningful biodegradable plastics [8-10]. These biodegradable polymers must not only be cost-effective; they must also have performance characteristics that are comparable to common synthetic polymers and they must be degradable in the cnvironment. These requirements, however, are often mutually exclusive, and practical biodegradable polymers have not yet been realized. It has been pointed out that it will take S - 10 more years before the development of biodegradable polymers reaches a practical level [ l l]. It is known that the biodegradable chemical intermonomer bonds include glycosides, peptides, and aliphatic esters [ 121. Thus, some of the most attractive materials with greatest potential i n terms of cost, material applications, and environmental compatibility include cellulose derivatives, especially cellulose esters. Among the cellulose esters, cellulose acetate (CA) has been produced industrially in the largest amount. Thus, the big interest has recently focused on the potential biodegradability of CAS. Until the end of the 1960s, it was accepted as axiomatic that cellulose acetates having DS 2 1.0 are resistant to hydrolysis by enzymes [ 131. In 1969, however, Cantor and Mechalas [ 141 found that even cellulose diacetates (CDAs) having a degree of substitution up to 2.5 could be degraded by microbial attack. Their investigation was carried out with the objective of relating losses i n semipermeability of cellulose acetate (DS 2.5) reverseosmosis membranes to microbiological degradation. They were discussing the durability of cellulose diacetate used for the reverse-osmosis membranes; that is, they did not have any interest in biodegradable plastics. In that sense it can be said that the first people to find the microbiological degradation of CDA in relation to biodegradable polymers were the research groups of Eastman Chemical Company and the University of Massachusetts. Buchanan et al. [IS] and Komarek et al. (161 demonstrated that CA with a degree of substitution (DS) up to 2.5 can be degraded microbially. In the former study [IS], they used two separate assay systems to evaluate the biodegradability of CA: an in-vitro cnrichment cultivation technique (closed batch system), and a system in which CDA films were suspended in a water treatment system (open continuous-feed system). The in-vitro assay employed a stable enrichment culture. which was initiated by activated sludge into a basal salts medium containing CA with 5 % (v/v). Extensive degradation of CDA (DS = 2.5) fibers was found after 2-3 weeks of incubation. In-vitro enrichments with CMA (DS = 1.7) films were able to degrade 80% of the films in 4-5 days. Films prepared from cellulose triacetate remained essentially unchanged after 28 days in the in-vitro assay. The wastewater treatment assay was less active than the in-vitro enrichment system. For example, approximately 27 days were required for 70% degradation of CMA (DS = 1.7) fill11s to occur, while CDA (DS = 2.5) films required approximately 10 weeks before significant degradation was obtained.Evidence for the biodegradation ofCA was also obtained through the conversion of cellulose~l-“C]-acetateto “CO, in the in-VitI-0 assay [ 151. The last viewpoint was conclusively established in their successive study by use of naturally derived mixed microbial culture derived from activated sludge and “C-labeled CA and “C-labeled cellulose propionate [ 161. Biodegradation was measured in an in-VitrO aerobic culture system that was designed to capture “CO, produced by the aerobic mi-

Biodegradable Plastics from Lignocellulosics

829

crobial metabolism of cellulose esters. More than 80% and 60% of the original lJc-polymeric carbon was biodegraded to ]'CO; for CA substrates with a DS of 1.85 and those of 2.07 and 2.57, respectively, over periods of 14-31 days. The amount of biodegradation that was observed with cellulose[ I-'"C] propionate with DS of 2.1 I , 2.44, and 2.64 was lower than that of the corresponding acetyl ester and ranged from 0.09% to 1.1 %. However, cellulose[I-"C] propionate with a DS of 1.77 and 1.84 underwent very rapid degradation in the mixed culture system, with from 70% to over 80% conversion of labeled polymeric carbon metabolized to "CO2 in 29 days. The high level of microbial utilization of carbon from both cellulose esters and its conversion to CO, confirms the biodegradability of these polymers and the potential they have for total mineralization in natural microbiologically active environments [ 16). Gu et al. [ 171 studied the cellulose acetate biodegradation upon exposure to simulated aerobic composting and anaerobic bioreactor environments. CA films with DS of 1.7 and 2.5 were exposed to biologically active in-laboratory aerobic composting test vessels maintained at approximately 53°C. The CA 1.7- and 2.5-DS films (thickness values of 1.3 X 10"-2.5 X 10 ' and 5.1 X 10" mm,respectively) had completelydisappeared by the end of 7- and 18-day exposure time periods in the biologically active bioreactors, respectively. On the other hand, almost no change in CA film weight losses was observed when the samples were exposed in the poisoned control vessels(aqueous KCN was added), showing the conclusion that CA film erosion during the composting exposures resulted from, at least in part, biologically mediated processes.Treatments of CA 1.7-DS film samples (1.3 X 10 '"2.5 X IO-' and 5. l X 10 mm thickness) in anaerobic serum bottles with or without KCN poison gave the same natures of results mentioned above. Therefore, it was concluded that degradation of the CA 1.7-DS films upon exposure to the anaerobic bioreactors was due, also at least in part, to biologically mediated processes. Gu et al. [ 181 also reported the degradation and mineralization of CA in simulated thermophilic compost environments. They studied the aerobic degradation of CA (DS 1.7 and 2.5) films exposed for up to 7 and 18 days, respectively. The number- and weight-average molecular weight (Mn and MW) values for both 1.7- and 2.5-DS CAS decreased significantly for extended composting exposure times. For example, Mn of residual polymers (CA 1.7 and 2.5 DS) decreased by 30.4% by day 5 and 20.3% by day 16, respectively. Furthermore, a decrease i n the DS from 1.69 to 1.27 (4-day exposure) and from 2.5 1 to 2.18 (12-day exposure) was observed for the respective CA samples. In contrast, CA films (DS 1.7 and 2.5) which were exposed to poisoned control vessels for identical time periods showed no significant changes in Mn and DS. Scanning electron microscopy (SEM) photographs of CA (DS 1.7 and 2.5) film surfaces after compost exposures revealed severe erosion and corresponding microbial colonization. Similar exposure times for CA films in poisoned control vessels resulted in only minor changes in surface characterization by SEM observations. Theconversion of CAS (DS 1.7 and 2.5) to CO, was monitored by respirometry. A lag phase of IO- and 25-day duration for CA DS 1.7 and DS 2.5, respectively, was observed, after which the rate of degradation increased rapidly. Mineralization of 1.7- and 2.5-DS CA powders, reported as the percentage of theoretical CO, recovered, reached 72.4% and 77.6% in 24 and 60 days, respectively. The results of this study demonstrated that microbial degradation of CA films exposed to aerobic thermophilic compost reactors not only results in film weight loss but also causes severe film pitting and a corresponding decrease in chain Mn and DS for the residual material. Furthermore, a high degree of CA mineralization was observed by the significant attainments of CO, conversion. Buchanan et al. [ 191 studied the influence of DS on blend miscibility and biodegradation of cellulose acetate blends. They reported their findings on blends of CA having a

830

Yoshioka and Shiraishi

DS of 2.49 with those having a DS of 2.06. This blend system was examined over the composition range of 0- 100% 2.06-DS CA employing both solvent casting of films (no plasticizer) and thern-ral processing (melt-compressed films and injection molding) using poly(ethy1ene glycol) as a common plasticizer. Thermal analysis and measurement of physical properties indicate that blends in the middle composition range are partially miscible. while those at the end of the composition range are fully miscible. The miscibility of these cellulose acetate blends is suggested as being influenced primarily by the monomer composition of the co-polymers. Bench-scale simulated municipal composting confirmed the biodestructability of these blends and indicated that incorporation of a plasticizer accelerated the composting rate of the blends. In-vitro aerobic biodegradation testing involving radiochemical labeling demonstrated conclusively that both the lower DS CA (DS 2.06) and plasticizer significantly enhanced the biodegradation of the more highly substituted CA (DS 2.49). This last point, that the presence of the lower DSCA significantly enriches the biodegradability of the more highly substituted CA, was already reported by Itoh et al. [20,21]. They reported that, when the former, more degradable CA, exists in an amount of more than 10%. the more highly substituted, thus less degradable CA can be significantly enhanced in its biodegradability. It was estimated that because of the presence of the lower-DS CA, microorganisms. usually not workable for destroying the higher-DS CA, can attain the degradation ability through so-called acclimatization. On the other hand, Sakai et al. [22] have searched for fungi that decompose CAS. They demonstrated that Neisseritr SI'CCN can degrade CA with a DS at least up to 2.3. The isolated strains, identified as Neisseria s i c m , degraded CA membrane filter (DS mixture of 2.8 and 2.0) and textiles (DS 2.34) in a cultivating medium. Biodegradation of 1.81and 2.34-DS CAS on the basis of biochemical oxygen demand reached S 1-60 and 4045%, respectively, in the culture of N . sicccr within 20 days. It also has been suggested that CA would undergo an enzymatic splitting by acetyl esterase in a first stage, down to a DS of 1 .O, before the degradation would continue by the action of cellulase enzymes [ l6.17,22]. Since cellulose diacetate (CDA) became thus recognized as a biodegradable polymer, various trials have been undertaken to impart sufficient thermoplasticity to CAS in order to render them melt-processable. This is because CDA, which has the greatest thermoplasticity among all kinds of CAS, fails to show adequate melting behavior without decomposition or discoloring. Thus, lowering the flow temperature of CAS is necessary, and it requires the addition of plasticizers and/or flow promoters. Traditional plasticization of CAS has been accomplished by using conventional p k ticizers with low molecular weights, such a s phthalates. glycerol derivatives, phosphates. etc. At present. phthalates and phosphates are used industrially i n procedures that are often very time-c(~nsuming (i.e.,4-5 h per batch). These plasticizers are usLlally not suitable for the prep:1ration of biodegradable polymers because of the harmful natures of their decomposition products. In this connection, there have been several attempts to utilize aliphatic polyesters of bacterial origins a s well as synthetic ones as plasticizers for CAS 123,341. Of these experimental studies, the following are includable as typical examples. Scandola et a l . reported miscibility of bacterial poly(3-hydroxybutyrate) (P(3HB)) with cellL1lose esters i n 1992 1231. They prepared blends of P(3HB) with cellulose acetate butyrate (CAB) and cellulose acetate propionate (CAP) by melt compounding. It is known that P(3HB)/CAB blendscontaining S-SOQI P(3HB) and P(.?HB)/CAP blends with 5h()(%p(3HB) arc transparent. stable homogeneous amorphous glasses, while blends with highel- p(3HB) content are partially crystalline. When i n the amorphous state. both P(3HB)/

Biodegradable Plastics from LignOCellUlOSiCS

83 1

CAB and P(3HB)/CAP blends show aglass transition which decreases regularly with increasing P(3HB) content, in excellent agreement with the behavior predicted for totally miscible blends. Both dynamic mechanical thermal analysis (DMTA) and differential scanning calorimetric (DSc) show that P(3HB) and CAB can crystallize from the blends only at temperatures higher than the composition-dependent T,. When crystallization is induced by thermal treatments. the melting temperature of the crystalline phase obtained depends on composition, a s expected for miscible blends of crystallizable polymers. Besides the strongly composition-dependent glass transition, another relaxation is observed, located in proximity to the T , of P(3HB) and slightly shifting to higher temperature with increasing CAB or CAP content. That is, another relaxation associated with mobilization of the IOWT, component is observed at alowertemperature. It was suggested that the two glass transitions are the manifestation of two mobilization processes coexisting in blends which appear in all respects to be single-phase, homogenous mixtures. After getting this information, Scandola’s group studied the effect of a low-molecular-weight plasticizer on the thermal and viscoelastic properties of the miscible blends of P(3HB) with CAB [24]. The low-molecular-weight plasticizer selected was di-n-butyl phthalate (DBP). The plasicizer DBP is miscible i n all proportions with both CAB and PHB. It is known that, analogous to the polymeric CAB/P(3HB) blends, the two polymer/diluent systems [CAB/DBP and P(3HB)/DBP] show a dual dependence on T, in composition. Thus, it can be said that i n binary mixtures such behavior appears to be independent on the macromolecular or lowmolecular-weight nature of the Iow-T, component. On the other hand, addition of a fixed amount of DBP plasticizer to CAB/P(3HB)blends with varyingcomposition [P(3HB) content from 0 to 100%] causes a significant decrease of T, of the binary polymer blends; the higher the amount of DBP in the ternary blend, the greater the T, depression. Concomitant with the expected plasticizing effect on T,, the presence of DBP also induces a decrease in the characteristic temperature of the additional low-temperature transition observed in CAB/P(3HB) blends. In the ternary blends, the temperature of such a transition is a function of DBP content only, being independent of the relative amount of the two polymers [CAB and P(3HB)I. Buchanan et al. (251 reported on their work on CAB and bacterial poly(hydroxybutyrateco-valerate)(PHBV)co-polymerblends.They prepared blends in the composition range 20-80 wt% of CAB and aco-polymer of PHBV by thermal compounding. Measurement by I3C-NMR and gel permeation chromatography (GPC) showed that no transesterification occurred during thermal mixing and that little change in molecular weight occurred.Blendscontaining20-50% PHBV were found to be amorphous, optically clear, miscible blends, while the blends containing 60-80% PHBV were semicrystalline, partially miscible blends. Both thermal and DMTA revealed the presence of a high-temperature transition that was sensitive to blend composition and a low-temperature transition whose position was uninfluenced by the blend composition. The high-temperature transitions of the 20-50% PHBV blendsclosely match calculated T V ’ s fora fully miscible blend. It was proposed that the dual transitions in the blends containing 20-50% PHBV arise from dynamic heterogeneity and not from a classical miscibility gap. Blend morphology was found to strongly influence physical properties such as tensile strength and tangent modulus. Blends containing 70% and 80% PHBV were found to exhibit tear strengths that were superior to either of the blend components. Buchanan et al. [26] also studied CAP and synthetic poly(tetramethy1ene glutarate) (PTC) blends. They prepared blends of synthetic polyester, PTC, and CAP, in the range of 50-90 wt% of the latter, by thermal compounding. During the compounding, no transesterification and little loss in molecular weight occurred. PTC was found to be a low-

832

Yoshioka and Shiraishi

melting (39"C), low-T, (-%"C), semicrystalline polymer. CAP is known as an amorphous, high-T, (136°C) polymer. It is known that the obtained CAP/PTG blends are optically clear and, when quenched from the melt, amorphous. Some blend compositions did exhibit small crystallization exotherms and melting endotherms in DSC experiments. The temperature of these melting endotherms lowered linearly from ca.168 to 148°C with decreasing CAP content over the range 85-60% CAP, while the AH, reached a maximum at 75% CAP in the blend. The T, of the blends containing more than 50% PTG lie near or below room temperature, and they are tacky and difficult to handle. No attempt was made to prepare and analyze these blends containing more than 50% PTG. Like the parent CAP,the blends containing 50-90% CAP exhibit a single relaxation process in their DMTA data, with no evidence of a low-temperature transition that could be associated with the polyester. The T, of these same blends agreed well with predicted values from Wood's equation. This fact offers good evidence for the miscibility of these blends. Buchanan et al. 1271 studied the influence of diol length of aliphatic polyesters on blend miscibility with CAP. A series of aliphatic polyesters containing a C5 dicarboxylic acid (glutaric acid) and C2-C8 straight-chain diols were blended with CAP at different composition levels. Characterization by DMTA revealed that, when blended with CAP, the polyesters prepared from C2-C6diols formed transparent, stable, amorphous glasses which exhibited a single composition-dependent T q . Upon reaching a C8 diol, the blend became partially miscible. It is known that within the miscible blends, analysis of their DMTA spectra indicates that the polyester prepared from the C4 diol had the highest level of miscibility with CAP, while the polyesters prepared from C5 and -CH,-CH,-0 -CH2-CH2diols gave the lowest degree of miscibility. Sub-T, mobilization processes, centered in the range -60 to -5O"C, were observed for the blends prepared from polyesters which contained C2, -CH2CH,0CH2CH2and C6 diols. In this connection, the activation energy for the sub-T, relaxation process for 40% poly(diethy1ene glutarakCAP blend was measured (210 kJ/mol). This result suggests cooperative, localized motion of a CAP-polyester complex. However, no relation was found between low-temperature relaxation processes and blend miscibility. Buchanan's group also reported a study on mechanical properties of CAP/synthetic aliphatic polyester blends 1281. This study was done because useful blends of cellulose ester with other high-molecular-weight polymers are generally unknown. Two aliphatic polyesters, PTC and poly(tetramethy1ene succinate)(PTS), have been thermally compounded with CAP in the range of 10-40% polyester. These blends have been injectionmolded, and the rnechanical properties of the molded bars were compared to bars molded from CAP plasticized with a low-molecular-weight diester, dioctyl adipate (DOA). The CAP/aliphatic polyester blends have significantly higher tensile strengths, flexural moduli, heat deflection temperatures, and greater hardness values than the corresponding CAP/ DOA blends. Buchanan et al. [29] studied the biodegradation of cellulose esters and cellulose ester/ diluent mixtures by composting.They evaluated a number of biodegradable polymers including CA with different DS and celluloseesteddiluent mixtures in a static, benchscale simulated municipal compost environment. Of the polymers evaluated,cellulose acetate (DS < 2.2), PHBV, and polycaprolactone (PCL) exhibited the fastest composting rate, disappearing completely after 14 days.Compression-molded films and injectionlnolded bars of CA (DS 2.06)hriethyl citrate (TEC) and of a series of miscible blends consisting of CAP andpoly(ethy1ene glutarate) (PEG) or PTC were evaluated in c m posting. Samples were removed from the compost at different intervals and evaluated by gravimetric analysis, GPC, and 'H-NMR. As expected, the CA/TEC film disappeared rap-

Biodegradable Plastics from Lignocellulosics

833

idly upon composting, while the injection-molded bars exhibited weight loss of 10- 12%. For the CAP/polyester blends. the type of polyester (PEG versus PTC) in the blends made no difference in cornposting rate. In general, a s the DS of the CAP decreased and the amount of polyester in the blend increased, the rate of composting and the weight loss due to colnposting increased. When the CAP was highly substituted, almost all the weight loss was ascribed to loss to polyester. When the DS of CAP was below approximately 2.0, both components degraded. Buchanan et al. also reported their work on composting of miscible CAPhliphatic polyester blends in another journal 1301. In the article, they evaluated a series of miscible blends consisting of CAP and PEG or PTC in a static bench-scale simulated nlunicipal compost environment. Samples were removed from the compost at different intervals, and the weight loss was determined before evaluation by GPC, SEM, and 'H-NMR. The type of polyester (PEG versus PTC) in the blend made no difference in composting rate. At fixed CAP DS. when thecontent of polyester in the blend was increased, the rate of composting and the weight loss due to composting increased. When the CAP was highly substituted, little degradation was observed within 30 days and almost all of the weight loss wasascribedto loss of polyesters. Although the polyesters were still observed to degrade faster, when the CAP DS was below approximately 2.0, both components were observed to degrade. The data obtainedin this experiment suggested that initial degradation of the polyester is by chemical hydrolysis and the rate of this hydrolysis is very dependent on the temperature profile of the compost and on the DS of CAP. Buchanan's group also developed and reported a bench-scale compost methodology that emulates a high-efficiency municipal windrow composting operation [ 31 1. A series of CA films, differing in DS,wasevaluated in this bench-scalesystem. In addition,commercially availablebiodegradablepolymers such as PHBV and PCLwere included as points of reference. Based on film disintegration and on film weight loss, CA having DS less than approximately 2.2 composts at rates comparable to that ofPHBV. NMR and GPC analyses of composted films indicate that low-molecular-weightfractionsare removed preferentially from the more highly substituted and more slowly degrading CA. Buchanan et al. submitted a patent application consisting of various data included in their above-mentioned publications 1321. Vazquez-Torres and Cruz-Ramos [33] studied binary blends of PCL with cellulose esters(CDA,CAB. and CTA) by using D S c , DMTA, and wide-angle X-ray scattering (WAXS)techniques.Aqualitativecomparisonwasmade with the results obtained by polarizing optical microscopy. PCL having MW of 35,000-45,000 was used. The PCL/ CAB system was proved to be partially miscible, whereas PCL/CDA and PCL/CTA appeared to be immiscible. A double-melting behavior was showed for PCL/CAB and PCL/ CTA blends. As these peaks did not shift by varying the heating rate of D S c run, this behavior can be due to melting of two populations of crystals of PCL, which maybe different in size. On the other hand, blends of PCL containing a small amount of CAB or CDA seem to develop more crystallinity for the PCL than this polymer alone. The solvent seems to have a certain influence on the thermal and morphological behaviors of the ascast blends of these three systems, affecting the extent of crystallinity of PCL, as well as its T,,, and AH,. This finding is discussed in the light ofWAXS and polarizing optical microscopy results. Zhang et al. [34] studied the melting and crystallization behavior and phase morphology of bacterial P(3HB) and hydroxyethyl cellulose acetate (HECA) blends prepared by casting films by D S c , Fourier-transform infrared spectroscopy (FT-R), SEM, and POlarizing optical microscopy. The melting temperatures of P(3HB) in the blends were in-

Shiraishi 834

and

Yoshioka

dependent of the blend composition with P(3HB) contents above 20%. The melting enthalpy of the blends decreased with increase in the HECA component and was close to the additive value of the enthalpy of the two components.The glass transition temperatures of P(3HB) in the blend were constant at about 8°C. No specific interaction between the two components was found by FT-IR. The crystallization of P(3HB) in the blend was affected by the HECA component, especially in the P(3HB)/HECA (20/80) blend. During the DSC cooling run at a lower cooling rate, two separate transitions werefoundfor P(3HB)IHECA (80/20), (60/40), and (40/60) blends, which corresponded to the crystallization of P(3HB) and the phase transformation of HECA from an isotropic phase to mesophase in the blends, respectively. The phase transformation of HECA from an isotropic phase to a mesophase was almost independent of the P(3HB) component. As has been shown in the above literatures [23-341, thermoplastic miscibilities could be found between bacterial or synthetic polyesters and cellulose esters, and amorphous optically clear miscible blends could be formed in certain ranges of combinations. It can be pointed out that novel thermoplastic materials moldable to give transparent homogeneous sheets could be prepared by these blendings. The problems with these are the main uses of cellulose esterssuch as CAB, the biodegradabilities of which have not been studied to a sufficient level, and higher cost performance of CAB compared with CA. At any rate, these studies have been continued with the intent to enhance the biodegradabilities of the materials obtained and furthermore to increase their thermoplasticities by blending the polymeric materials with high biodegradabilities. On the other hand, the addition of low-molecular-weight plasticizers has been taken as the second type of plasticization of cellulose acetates. Buchanan et al. [29] reported that CA can be effectively plasticized by thermal compounding with TEC.Thecompounded resins were converted to compression-molded film and injection-molded bars. Their biodegradabilities were evaluated and confirmed by composting. Concerning this plasticization of CA with low-molecular-weight plasticizers, there appeared in newspapers several announcements of commercialization of biodegradable plasticized CA. One was an announcement from Planet Polymer Technologies. Inc. (California, USA) that CA plasticized with triacetine was entering the Japanese market with the trade name of Lunare. Theother was from Daicel Chemical Industries Ltd., using polycaprolactone oligomer having a molecular weight of 500 as plasticizer for CA. As for methods for plasticization of CA, in addition to the two kinds of external plasticizations mentioned above, which are usually adopted, it is possible to make use of internal plasticization-that is, chemical modification or grafting methods. In view of this situation, we have been attempting in the past 5 years to find novel plasticizers and plasticizing procedures by which biodegradable thermoplastic polymers can be obtained from CAS. In the early stage, attempts have been made to introduce oligoester side chains into CA molecules by reaction of CA with dicarboxylic acid anhydrides such as maleic anhydride (MA) and succinic anhydride (SA) together with monoepoxides such as phenyl glycidyl ether (PGE), styrene oxide (SO), and allyl glycidyl ether (AGE). Theoligoesterifications of CAS have been carried out by the use of a compounding machine at high temperatures with constant kneading speed. The results of these attempts are shown later, With the advancement of this study, it became clear that CA must be sufficiently graft co-polymerized to prevent the bleeding of co-produced homo-oligomers or homopolymers. From these findings came the understanding that the more effective the grafting attained, the more ideal plasticization of CA can be effected. With this idea in mind, the

Biodegradable Plasticsfrom Lignocellulosics

835

authors have come to grafting work, in which CAS are plasticized by cyclic ester grafting using tin(II)2-ethylhexanoate (SnEht,) as catalyst.

II. PLASTICS FROM CELLULOSE ACETATE A.

Cellulose Acetate Plasticized by Reaction with Dibasic Acid Anhydrides and Monoepoxides During Melt Processing, and Their Biodegradabilities

The authors tried to develop a methodology for the plasticization of CAS by reaction with dibasic acid anhydrides and monoepoxide during melt processing [35].In this case, oligoesterified CA, prepared by the reaction of CA with SA and PGE, for example, at a temperature between 70 and 1 80"C, has the hypothetical structure shown in Fig. 1. This kind of introduction of the oligoester side chain into the CA molecule would enhance the thermoplasticity of CA (internal plasticization). At the same time, homooligomers of the oligoester would be produced and these would be able to act as external plasticizers. The effect of the plasticization treatment can be seen in Fig. 2. From the figure, it is known that CDA, though it has the greatest thermoplasticity among cellulose acetates, still lacks melt processability, whereas when it was reacted with SA and PGE at 120°C for 20 min under kneading conditions, it was converted easily into a thermoplastic material. Thus, the above supposition, that the formation of oligoester chains as grafted branches of CAS together with homo-oligomers enhances the thermoplasticity of the products, was confirmed. Since these results are very promising, the plasticization of CAS was examined under different conditions, including amounts of reactive plasticizers added, kneading reaction temperature, molding (hot pressing) time, and so forth. Thus, moldable plasticized products having various mechanical and thermal properties could be obtained. Representative stress-strain curves of CDA and cellulose monoacetate(CMA) oligoesterified are shown comparatively in Fig. 3, along with those for polystyrene (PS) and polyethyrene (PE). From this figure, it appears that both SP-10 and SP-40 have attractive mechanical properties. They are tougher than PS and stronger than PE. Thermal softening curves of L-40 and its plasticized materials are shown in Fig. 4. Among the latter, SS-40 (oligoesterified L-40 reacted with SA and SO) is included as well. From the figure, it can be seen that when CDA was oligoesterified by the reaction with SA and PGE, for example, i n an amount of 30.9 wt% at 120°C for 20 min in the kneader, the flow temperature dropped to 170°C from 250°C. Considerabledecrease in the flow temperature of CDA can be attained by this oligoesterification grafting.

0

0

II

II

OH

FIGURE 1 Schematic chemical structure of CA oligoesterified with SA and PGE.

036

Yoshioka and Shiralshi

FIGURE 2 Effects of oligoesterification on L-40: upper left, untreated L-40; upper right, L-40 hot-pressed at 190°C under 15 MPa for 30 S; lower left, L-40/SA/PGE (100;11.0/33.8) just after being kneaded at 120°C for 20 min without pretreatment; lower right, a sheet from the kneaded sample (lower left) prepared by hot-pressing at 190°C under 15 MPa for 30 S.

Although these results concerning thermoplasticization of CDA and the mechanical properties of the products were extremely fascinating, there often occurred a plasticizer migration (bleeding) problem. This was caused by fugitive plasticizers which are mostly homo-oligomers. The occurrence of plasticizer bleeding was found to be much more pronounced for plasticized CDA than for plasticized CMA (Fig. 5). In the latter, almost no bleeding was found. This difference is thought to be caused by a lack of miscibility, that is, a lack of affinity between the CAS and the oligoesters. To make immiscible polymers (or polymer A and oligomer B) miscible or compatible, a compatibilizer is often added. As one type of compatibilizer, A-B blockor graft co-polymers have been found to beeffective. There-

FIGURE 3 Stress-strain curves of representative CAS plasticized with SA and PGE. SP-l0 (40): LL-l0 (L-40)/SA/PGE = 100/11.0/33.8; prepared without pretreatment; kneading condition 120"C, 30 rpm (5 min)-90 rpm (20 min); hot-pressing condition 190"C,15 MPa, 30 S ; tensile test, span 40 mm, crosshead speed 5 mmlmin.

rom

Plastics Biodegradable

Llgnoce~~u~osics

837

I

50

100 150 200 250 300 Temperature ('C)

FIGURE 4 Thermal softening curves for L-40 and its plasticized materials. L-40: untreated CDA, SP-40: L-40ISNPGE = 100/11.0/33.8, MP-40: L40IMAPGE = 100/10.8/33.8, SS-40: U O I S A I S O = 100/11.0/27.0. Prepared without pretreatment; kneading condition 120°C, 30 rpm (5 min)-90 rpm (20 min); flow test condition, load 10 kgf; heating rate lO"C/min; T,, flow temperature.

FIGURE 5 Bleeding of plasticizer from sheets of plasticized L-40 and LL-10. SP-10: LL-IOISAI PGE = 100/11.0/33.8; prepared on April 21, 1993; SP-40: L40/SA/PGE = 100/11.0/33.8; prepared onApril19,1993: kneading condition 120°C, 30 rpm (5 min)-90 rpm (20 min); hot-pressing condition 190°C, 15 MPa, 30 S; photographed on June 26, 1995.

838

Yoshioka and Shiraishi

A

B

C

D

E

FIGURE 6 Effect of catalyst (Na2COS) on weight gain of cellulose acetate. Kneading, 12OoC, 90 rpm, 15 min; flow test, die diameter 1 mm, length 2 mm; plunger, 1 cm2;load 5 MPa; heating rate 1O0C/min. A, L-4O/MA/PGE = 100/16.9/25.9 (30%); B, L-40/SAPGE = 100/17.3/25.9 (30%); C, L-40OISAPGE = 100/21.7/32.6 (35%); D, LL-lO/SA/PGE = 100/21.7/32.6 (35%); E, L-40/SA/GMA = 100/17.7/25.1 (30%). Each value in parentheses shows reactive plasticizers content (wt%) in the startingmaterial. 0,withoutcatalyst(Na,CO,);withcatalyst(Na,CO,).

fore, a large amount of oligoester side chain attached to the CA molecule can beexpected to enhance the affinity betweenthe modified CAS andthe homo-oligomer. Thatis, grafting can effectively suppress or prevent the bleeding of nongrafting homo-oligomers. CMA can be proceeded by grafting to a higher level compared with the case for CDA. Based on these considerations, methods for enhancing the amount of grafting were pursued by varying the combination of dibasic acid anhydride and monoepoxide, by extending the kneading reaction period, by using grafting catalyst (Fig. 6), and so forth [36]. By these trials, the grafting could be enhanced, which actually resulted in the suppression of the plasticizer bleeding. The biodegradabilities of the representative samples obtained were examined by a soil burial test in a controlled environment (30°C, 80% RH) and by the measurement of oxygen consumption in a closed system where test samples were exposed to standard activated sludge [36]. Concerning the former, it was found in a few examples, as in Fig. 7, that the plasticized CA samples were degraded and disappeared within relatively short periods (i.e., within 3-12 months). Results of the biodegradation test measured by means of the oxygen consumption within a closed activated sludge suspension are shown in Fig. 8. It is apparent that all samples are subjected to significant biodegradation. The CMA control sample, of which biodegradation had been literarily confirmed, was degraded more slowly than any of the oligoesterified samples.

B. Cellulose Acetate Plasticized by Grafting with Cyclic Esters [36-391 In the previous section, it was shown that CA must be sufficiently graft polymerized to achieve effective plasticization and to prevent the bleeding of homo-oligomers. Thus, the amounts of grafting and the graft efficiency were intended to increase. As an extensionof the efforts, the authors have come to grafting work in which cyclic esters are reacted with CA using SnEhtz as a catalyst. This is based on the following information found in relatively recent publications. It is often said that little is known about the polymerization mechanism of cyclic esters in the presence of SnEht, [40,41]. Ikada described the same in his review article [40]on “polylactic acid,” butintroduced a mechanismbywhichhomo-polymers are produced predominantly, even though the ring-opening polymerization of cyclic esters is

Samples disapped after 3 months

Samples disappeared

l

*er l2 months

FIGURE 7 Changes SA-l0 and SP-IO specimens during the soil burial test in the incubator.

839

Shiraishi

840

Yoshioka and

Exposure time (week) FIGURE 8 Results of' the exposure test on the closed activated sludge system. Degree of degradation was calculated using oxygen consumption and theoretical initial oxygen demand. 0.LL-IO: 0, SP-40: L40/SA/PGE = 100/11.0/25.9; kneading, 120°C. 20 min. A, MP-IO: LL- IO/MA/F'GE = 100111.0/33.1; kneading,120°C.20 min, V,SA-IO:LL-IO/SA/AGE = 100/11.0/2S.9; kneading. 80°C, 15 min.

conducted in the presence of CDA. On the contrary, Kricheldorf and his co-workers studied the polymerization of L-lactide catalyzed with SnEht, i n the presence or absence of benzyl alcohol [42]. When SnEhtz and benzyl alcohol are used as a catalyst and a co-initiator, respectively, NMR spectroscopicexamination of all polylactidesobtained by the ringopening polymerization revealed the presence of benzyl ester end groups but the absence of 2-ethylhexanoate end groups [30]. This result means that the hydroxyl group of alcohols plays an essential and direct role in initiating the ring-opening polymerization of cyclic esters. In this sense, graft polymerization can be expected to occur selectively when the cyclic esters are polymerized in the presence of CA and SnEht?. Thus, the authors started a co-grafting study of s-caprolactone (CL) and I,-lactide (LACD) onto CDA using SnEhtz as a catalystin order to realize grafting with considerably

2011

O

O

,

, 20

,

, 40

,

, 60

Reaction time (min)

FIGURE 9 Effects of thereactiontime on theyield.Reactiontemperature140°C;L-40/(LACD + CL)/catalyst, 100/600/15 (by weight); LACD/CL, I.O/I.O (by mole).

Biodegradable Plastics from Lignocellulosics

0

20 40 Reaction time (min)

841

60

FIGURE 10 Effects o f thereactiontimeonthe flow temperature. Reaction temperature 140°C; L-40/(LACD CL)/catalyst. 100/600/15 (by weight): LACDICL, 1.0/1.0 (by mole).

+

high graft efficiency. In these cases, the reaction temperature was kept constant at 140°C and the reaction time was changed from 0 to 60 min. The other reaction conditions are shown in the footnotes of the relevant figures (Figs. 9-12). One of the characteristics of this grafting is that the grafting reaction can proceed rapidly and can be completed within 10-30 min (Fig. 9), and in correspondence with this, the flow temperature decreases to about halfof that of CDA(Fig. 10). The products obtained are moldable to films or sheets without using any plasticizer. Furthermore, it is known from Fig. 1 1 that LACD is grafted more rapidly than CL, producing relatively rigid and brittle products in the earlier stages and elastomer-like ones in the latter stages of the grafting. From Fig. I I , it can also be said that the total molar substitution reached its theoretical maximum value after 10 min of grafting. This result confirms that the graft reaction proceeds at a high rate and shows that selective grafting is achieved completely without significant production of homo-polymers or homo-oligomers. This supports the reaction mechanism proposed by Kricheldorf et al. (421. Transparent sheets were obtained depending on the reaction conditions, showing their amorphous nature. In this regard, the triad structure of the grafted side chain was analyzed by means of high-resolution NMR spectroscopy. An example of the NMR data is shown in Fig. 12, in which the E-oxycaproyl unit is denoted by C and the lactidyl unit by LL. Each of the splitting spectral lines for a, /3, y, &methylene carbons of c-oxycaproyl units and the methyl carbon of lactidyl unit

LL -1""

20 40 Rcaction timc (min)

GO

FIGURE 11 Effects of thereaction time on the molar substitution. Reaction temperature 140°C; L40/(LACD + CL)/catalyst, 100/600/15 (by weight); LACD/CL, I.O/I.O(by mole). 0 , LACD; A, CL: 0 , LACD + CL; , theoretical maximum value of (LACD + CL). ~

842

Yoshioka and Shiraishi

Y

38

3‘

31

0

1I

26

PP*

14

B

10

I,

l8

11

I1

FIGURE 12 ”C-NMR spectrum of (CL-CO-LACD) grafted CDA. H- 1. Region of a-,p-, y -, Smethylenecarbonatoms of coxycaproyl units, methylcarbonatom of lactidylunit, and acetyl methyl carbon atoms of CDA.

in a grafted product was assigned by reference to Kasperczyk and Bero’s work 1431. The spectrum of Fig. 12 means that even for a grafted CDA (H-l) prepared by using a liquid ratio of 2, LACD/CL = 2/5 (by mole), reaction time of 30 min, and reaction temperature of 140°C, the introduced graft side chains are long enough to reveal triad sensitivity. Appearance of LLCC, CCLL, and LLCLL sequences, total signal strength of which are comparable or larger than that due to the CCC sequence, reveals meaningful occurrences of random polymerization of CL and LACD within the graft side chains. This fact confers irregularity in graft chains and is considered to be related to the appearance of the high thermoplasticity and amorphous nature found and discussed above. In other words, it can be said that the analysis of the structure of the grafting side chain by means of NMR spectroscopy showed that, although the side chains are composed of large amounts of c-oxycaproyl or lactidyl block polymer portions, considerable amounts of randomly polymerized parts coexist in the grafting chains, which confers high thermoplasticity and amorphous nature on the grafted product obtained.

111.

PLASTICS FROM LIGNOCELLULOSE AND APPLICATION TO BIODEGRADABLE POLYMERS

A.

Plasticization of Wood by Benzylation and Blending with Polycaprolactone

Since lignocellulosics, including wood, are not thermally flowable materials, the methods for processing them are limited. If appropriate thermoplastic properties could be imparted to the lignocellulosics, they would become more useful materials. More than 15 years ago, it was found that wood could be converted into a plastic material by chemical modification, such as by esterification and etherification [4,5]. Chem-

Biodegradable Lignocellulosics Plasticsfrom

843

ical modification does not necessarily require special techniques; conventional and simple methods work satisfactorily for this purpose [6,7]. The phenomenon of thermal flow can be explained in terms of internal plasticization of wood. The introduction of large nonpolar substituent groups into wood can result in a chemically modified material with high thermoplasticity. When small substituent groups and/or polar groups are introduced, such thermal fluidity cannot be achieved; therefore, the latter modification alone cannot produce plastic properties. However, this lack of plasticity can be solved by external plasticization. All of the above reveals that chemically modified wood derivatives can be considered as novel biobased plastic materials, and experimental studies have targeted the development of composites with enhanced physical properties. There are various types of thermoplasticized wood derivatives. Among them, BzW is known as a moldable material giving excellent mechanical properties. Indeed, a molded sample of BzW showed a tensile strength of 42.7 MPa, which is fairly high when compared, for example, with that of polystyrene (PS) (Styron 666, a product of Asahi Chemical Industry Co. Ltd.). The tensile strength of Styron is 29.4 MPa under the same conditions. Differential scanning calorimetric measurements have revealed that both polymers are amorphous. Thus, the thermal softening behavior is similar, qualitatively, when measured with a thermomechanical analyzer. Quantitatively, there aresomedifferences. Experimental observations with a flow tester revealed that PS undergoes flow at 153"C, while BzW starts to flow at 175°C. Almost identical results are found when BzW is compared with polypropylene (PP) (J700G; MI = 7, a product of Idemitsu Petrochemical Co. Ltd.), which is also widely used as a thermoplastic polymer. However, even though the flow temperature of BzW is higher than that of PS or PP, the temperature difference can be reduced to almost zero by blending polycaprolactone (PCL) (PLACCEL H-4; Mn = 40,000, a product of Daicel Chemical Industries, Ltd.) in an amount of 20% with BzW. Figure 13 shows the dependence of melt viscosity (p)and shear stress (TU) on the shear rate ( 7 ) of BzW and PP, measured with a Capirograph. Itis seen that both the

7 n

W

.-0

v)

c

W

5t

FIGURE 13 Melt viscosity and shearstress versus shear rate for BzW and PP.

Yoshioka and Shiraishi

844

viscosity and the shear stress of BzW reveal similar shear rate dependences as those for PP. However, the values for BzW are one order larger than those for PP. This means that there exists a greater resistance to thermal processing for BzW. In order to reduce this resistance, blends of BzW with PCL were studied. since the latter has considerably greater thermoplasticity. Since PCL itself has a low melting point (ca. 60°C) and low tensile strength (19 MPa) at room temperature, its compounding with BzW would be of great significance. Results obtained are shown in Fig. 14. As BzW content increases. both the viscosity and shear stress are seen to decrease. At a PCL content of 30%, the characteristic melt flow values converge with those for PP, which are shown as dotted lines. Thus, the blended composite consisting of 70% BzW and 30% PCL has practically the same fluidity as PP. However, the mechanical properties of the blends were reduced by this blending as shown in Fig. 15. This is a widely observed phenomenon in polymer blends. The mechanical properties of the latter. can, however, be improved by the useof appropriate compatibilizers, which enhance the adhesion of interfaces between or anlong the domains or phases of blended polymers. As one of such attempts, the styrene-maleic acid anhydride co-polymer (SMA) was used and its effect as a compatibilizer was investigated. The results obtained are shown in Fig. 16. With an increase in the amount of added SMA, the tensile strength of the BzW/ PCL sheet recovered; and when the SMA content became 596, the tensile strength rose to 200 kgflcm' ( 19.6 MPa). The composite also showed practical moldability as shown in Fig. 17. The photograph displays trays that were vacuum-formed from sheets of the blends. This kind of tray is commonly used in grocery stores. These trays possess not only reasonable strength but they also have functional properties as biodegradability and photodegradability, as shown later.

B.

Bio- and Photodegradabilities of BzW and BzW/PCL Composites

Novel thermoplastic materials, which have been developed by the chemical modification of wood, have been described in Section 1II.A. Although they are of biological origin, an

n

W

.-m0

c

U

a M 0 -

M

0 -

845

Biodegradable Plastics from Lignocellulosics

GOO

-

-600 - 6

-

0 : Tensile strength

0:Break elongation A

'

Tensile modulus

0

o

400 -A

- 4 0 0 .c2 . 4

F

-5M

-U

C

z .-- 200 -

Y m U

Y

U

U

0'

YI

200 G-2

c

0.8 o'o', 10/0

Q

,:,

I

5/5

Blcntl ratio

FIGURE 15

-U .U

c

c

0

C

? I

I

A c

I .

0 -0

0/10

(BzW/PCL)

Relationship between mechanical properties of the BzW sheets and PCL content

evaluation of the biodegradability and photodegradability characteristics of these thermoplasticized wood derivatives is indispensable. Thus, the degradability of BzW was compared with that of synthetic polymers including PP, PS, polycarbonate, and others.The results reveal that while the sheets of synthetic polymers did not lose any strength at all even after 80 days of immersion in aerobic and anaerobic activated sludge, sheets of BzW exhibited continuous deterioration and strength reduction. The results for BzW are shown in Table 1 . The results were most impressive since, although PP is said to be photodegradable to a certain extent, it did not reveal any change in its external appearance or its strength; whereas BzW became very brittle and was destroyed after 80 days of exposure to sunlight.

SMA (part)

FIGURE 16 Relationship betweentensilestrength content.

of BzW/PCL 1713 (w/w)lsheetsand

SMA

Shlraishi 846

and

Yoshloka

FIGURE 17 Trays from BzWPCL blended composites.

Furthermore, the BzWrPCL sheets were found to undergo faster biodegradation than those of each individual sheet component, as revealed by tests carried out under the same conditions (not shown). This fact is remarkable inasmuch as PCL is well known for its high biodegradability.

W.

CONCLUSION

In this review, recent studies on biodegradable plastics from cellulose and lignocellulose were elucidated.It is known that thesestudies have just been developing. Even elastomeric biodegradable polymers could be introduced from cellulose, which had been considered a rigid polymer. This new development anticipates a wide and prospective future for that kind of development.

TABLE 1

Biodegradability andPhotodegradability of BzW

Control (BzW) Dipping in aerobic activated sludge for 80 days Dipping in anaerobic activated4.8 sludge for 80 days Dipping in sea water for 80 days Exposuring to sunlight

Tensile strength (MW

Tensile breaking elongation (%)

42.7 37.7 37.9 39.8

9.8 4.1

-

6.1

-

Biodegradable Plastics from Lignocellulosics

047

REFERENCES 1.

2.

3. 4. S.

6. 7.

8. 9.

IO. 11.

12. 13. 14.

IS. 16. 17. 18. 19.

20. 21. 22.

23. 24. 25.

H. Marusawa.and K. Uda, Plrrstic Mrttc~riczlsSeries 1/71. Ct~llulosic~ Re.sirt.s, Nikkan Kogyo Shinbun. Tokyo ( 1970). M. Kita. Htctrdlm)k of N m d Wootklw Mtrteritrls, (T. Haragtlchi et a l . . eds.). Gihodo, Tokyo. p. S94 (1996). C. R. Fordyce, and L. W. A. Meyer, Irrtl. Errs. Cherrr., 32: 1053 ( 1940). N. Shiraishi, T. Matsunaga, and T. Yokota. J. A p p / . Polyrm‘r Sri.. 24:2361 (1979). N. Shiraishi, in Cllrrnistry of Wootl U t i l i x t i o n (H. Imamura, H. Okamoto, T. Goto. Y. Ynsue, T. Yokota. and T. Yoshinloto, eds.). Kyoritsu. Tokyo, p. 294 (1983). N. Shiraishi, in Ct~ll~tlosc Utj/j:cctiorr, Kt~stwrchrrrrd Kmwrds in Ce//lrlosic~.s (H. lnagaki and G. 0 . Phillips, eds.). Elsevier Applied Science, London & New York. p. 97 (1989). N. Shiraishi, in Wood t r r d Cc4/rrlo.sic. Chmrisrry (D. N. S. Hon and N. Shiraishi. eds.). Marcel Dekker. New York & Basel. p. 861 ( 1991). G. J. L. Griffin (ed.), C h ~ ~ r ~ ~nrtd i s t rTyw h r l o l o g y t~f’Biotlc~~~rcrtla/~le P o / y r w r s , Blackie Academic &L Professional, London ( 1994). G. Scott and D. Gilead (eds.). Drgrurlnblt~PoIyrrrrr.s-Prit~ci~~/e~s t r i d Applic.trtiorl. Chapman & Hall, London (1995). A. J. Domb. J. Lost, and D. M. Wiseman (eds.). H m t l l x ) o k of Bioc/qrcrtltrD/r Polyrrrrrs, Harwood Academic, Vienna (1997). Biodegradable Plastic Society. Japan. Opening of New Plastic Era-Report of the Committee for Developing Biodegradable Polymers t o Practical Usage. p. I (March 1995). H. Sawada. in Prtrcriccrl Biodt~grrctitrhleP1rr.sfic.s. CMC. Tokyo. p. 38 (1992). E. T. Reese. b z d . Ertg. Cherrr., 4% 1 ):89 ( 1957). P. A. Cantor. and B. J. Mechalas, J. Po/yrrwr Sei.. 213:225 (1969). C. M. Buchanan, R. M. Gardner. and R. J. Komarek, J . A/>/>/.Po/yri?c,r Sci.. 47:1709 (1993). R. J. Komarek. R. M. Gardner. C. M. Buchanan, and S. C. Gedon, J . Appl. Polyrrtcr Sei.. SO: 1739 ( 1993). Ji-Dong Gu, D. T. Eheriel. S. P. McCarthy. and R. A. Gross, J . Erniron. Polyrwr Degrnrltrtior?, /(2): 143 (199.3). Po/yrr~c,rDrgrrrtkttiort, Ji-Dong Gu. D. T. Eberiel. S. P. McCarthy, and R. A. Gross. J . Err~~irotr. 1:2X1 (1993). C. M. Buchanan. D. Dorschel. R. M. Gardner, R. J. Komarek. A. J. Matosky, A. W. White, and M. D. Wood. J . Ettvirort. Polyrrwr Ikgrrrcltrriorl, 4: 179 ( 1996). M. Itoh. A. Kiyose. and Hirao, Japanese Patent, Hei 7-76632 (1997). M. Itoh. Ahstt: of Erttirorr. Cor![ E r h i h i t i o r ~o r 1 the Role of Bioc/c~grtrduhleMtrtc~ritr1.sirr Wtrstr MarttrgcJrrtcJrrt.Tokyo. p. 45 ( 1995). K. Sakai. T. Yamallchi. F. Nakatsu. and T. Ohe. Ahstr: of the 1’9’94 Ntrtiorttrl M e e t i r ~ gofAgric,. Cltc,t,r. of Jtrptrrr. Tokyo. p. S I I . 3FP5 ( 1994). M. Scandola. G. Ceccorulli. and M. Pizzoli, Mrrt~rorrrolc~c~rtles. 25:6441 (1992). G. Ceccorulli. M. Pizzoli. and M. Scandolo. Mnc.,r,rrrolr,c.rrl~,,s, 265722 ( 1993). C. M. Buch:unan. S. C.Gedon, A. W. White. and M. D. Wood. Mat,rorrrolc(,rtlcs, 25:7373 ( 1992).

36.

C.M.

Buchrunnn. S. C.Gedon. A. W. White, and M . D. Wood. Mtcc,rorrro/ec,rrl~,,s,26:2963

(1993).

27.

C. M. Buchanan. S. C. Gedon. B. G. Pearcy. A. W. White, and M. D. Wood. M t r c ~ , n n r o l l , . c ~ ~ r / ~ , . ~ . 26:5704 (1993). ?X. A. J. White. C. M. Buchanan. B. G. Percy. and M. D. Wood. ./. Appl. Po/yrrtrr Sci.. 52:52S ( 1994). 29. C. M. Buchnnan. D. D. Dorschel. R. M. Gnrdoner. R. J. Komarek. and A. W. White. J. M. S. “Prrrc, App/, Chtwt., A32(4):683 (1995). 30. C. M. Buchanan. C. N. Boggs. D. D. Dorschel, R. M. Gardner, R. J. Komarek, T. L. Wotlerson and A. W. White. ./. Errllirorz. Polyrrrrr Ilc,~yrtrt/rctior,.3 : 1 ( 1995).

a48 31. 32.

33. 34. 35. 36. 37. 38. 39. 40. 41.

42. 43.

and

Yoshioka

Shiraishi

R. M. Gardner, C. M. Buchanan, R. Komarek. D. Dorschel. C. Boggs, and A. W. White, J . Appl. Po/yr?~er Sci.. 52: 1477 ( 1 994). C. M. Buchanan, R. M. Gardner, M. D. Wood, A. W. White, S. C. Gedon. and F. D. Barlow. U.S. Patent 5292783 (1994). H. Viizquez-Torres and C. A. Cruz-Ramos, J. Appl. Polynrer Sci.. 54: 1141 (1994). L. Zhang, X. Deng. S. Zhao, and Z. Huang. Polyrr~er;38(24):6001 (1997). M. Yoshioka, T. Miyazaki, and N. Shiraishi, Mokuzcti Gtrkknishi, 42(4):406 (1996). M. Yoshioka, K. Okajima, T. Miyazaki, and N. Shiraishi. J. Wood Sci., 46( 1):22 (2000). M. Yoshioka, Preprints of YN-2 R e g d m Meeting o f t h e Society j b r rhe Stcrcly c f Eco-rrlutc,ricrl. Society o f Polymer Scientists, Japan. p. I (1998). M. Yoshioka, H. Mizumoto, N. Hagiwara, and N. Shiraishi, Pwprirlr oj' '98 Cellrrlo.se K & D , 5th Annual Meeting of the Cellulose Society. Japan, p. 26 (1998). M. Yoshioka, N. Hagiwara, and N. Shiraishi. Cellulose, 6 ( 3 ) :193 (1999). Y. Ikada, in Hurdhook of Hiodcgrurlcrble Plastics (Y. Doi et al.. eds.), N. T. S. Ltd.. Tokyo. p. 279 (1995). A. Kowalski, A. Duda, and S. Penczek, Mncrorrrol. Kcrpid Cornmrr~..19567 (1998). H. R. Kricheldorf, 1. Kreiser-Saunders, and C. Boetlcher, Po/ym>r:36(6):1253 (1995). J. Kasperczyk and M. Bero, Mrlcrortzol. C / I ~ I I192: I . , 1777 (199 I ).

Recycling of Wood and Fiber Products Takanori Arima The University of Tokyo, Tokyo, Japan

Wood has been the most important construction and decorative material since ancient times. The importance of wood will certainly increase in future years, since among all the construction materials, it is the only major renewable resource. Further, in view of growing energy consciousness and the resulting emissions of carbon dioxide to the earth's atmosphere, greater wood utilization is to be encouraged, since wood processing requires relatively smallamounts of energy. Since wood is a typical organic compound which is constituted by absorbing carbon dioxide from the atmosphere, biodegradation and burning leads to the release of carbon dioxide to the atmosphere. The effective use of wood waste produced from industries that are processing or consuming wood and of timber recovered from demolished houses has become a serious issue. Solving this problem could be an effective means for saving wood resources and reducing the emission of carbon dioxide, even though the processing requires a reasonable amount of energy. For wood-based resources, the breakdown ofraw material basically follows a sequence of steps from logs to lumber, to chips, to fibers, to charcoal, and tinally to fuel. After wood-based products such as building products, furniture. paper, etc., leave the factory, they are utilized and eventually disposed of. With disposal, they again become available as raw materials to various wood-utilizing industries if they are collected, shipped, and reprocessed. In such a case, it can be said that there is relatively little negative impact associated with this type of recycling. For example, there are many cases in Japan where, after removing foreign matter, wood recovered from demolished timber structures is reused in chip form for particleboard production or as a fuel source. If the material is prepared in this way, it presents no technical difficulties and can therefore be recycled. However, when the recycling conditions are not met, the wood material will simply be thrown away or incinerated along with other garbage. The essential point is to address adequately the problems of (1) usable life spans of wooden materials such as furniture and building materials; ( 2 ) their subsequent quality considerations asa recyclable raw material, including the removal of foreign debris; and ( 3 ) the collection ofthewood material. It is vital that this should be coordinated in order to address effectively the issues of environmental protection and energy concerns. In this chapter, an overview is given of the current conditions and issues related to the recycling of wood materials both from the wood industries and from the demolition of wooden structures in Japan. 849

Arirna

850

1. A.

EXISTING DISPOSAL AND UTILIZATIONIN WOOD INDUSTRY General

An outline of the use of wood wastes from wood industries is shown in Fig. 1. The breakdown of raw materials basically follows a sequence of steps in change of form from the unprocessed or semiprocessed state (such as logs or large timbers to be further processed) to lumber to chips and finally to fibers. One can imagine this process as a kind of “cascade” as the material is sequentially broken down into smaller units. Accordingly, provided that foreign objects or otherimpurities do not enter into thisprocess, it is possible to capture and utilize this “cascade”for the production of wood-based materials. For example, particleboard producers utilize the waste of lumber and plywood production which are farther upstream in this “cascade,” usually in chipform,as do many pulp producers as well. The waste trimmings, sawdust, bark, etc., of the wood-utilizing industries such as plywood, particleboard, fiberboard, and so on, are often burned to supplement the energy needs of the factories where the wood waste is generated. Although there has been a considerable amount of technical research into the uses of wood waste, whether it is economical for industry depends very much on its location and on the method of transport and the cost. Most of the residues from lumber mills are sent to the paper, particleboard, fiberboard, and wood-particle cement board industries as chips or flakes. Existing disposal and utilization of sawdust produced from lumber mills consists of fertilizer, briquettes, charcoal, and fuel. Approximately 70% of sawdust seems to be used effectively. Other wood waste is burned as fuel in the wood waste-generating industries’ own boilers or is incinerated or dumped. On the other hand, most of the wood waste from the building, furniture, and pallet industries seems to be disposed of by dumping or burning. Effective utilization of the wood waste generated on construction sites, such as used concrete-form plywood and the timber recovered from demolished houses, remains as one of the most important problems to be solved technically and economically, as described in Section 11.

B. Wood Industry Integration Plan The wood industry integration plan calls for the integration of the various industries that produce or use wood into segregated areas removed from residential districts. Such integration aims to improve the circulation of raw materials and products and to solve problems of pollution. In 1997, about 100 such areas were completed or under construction, and this seems to be a reasonable achievement. However, they still present some problems, especially cooperation between various manufacturers and the difficulty of incorporating the building and furniture industries. These areas have helped to solve the disposal problem and have enabled better utilization of wood waste for various purposes, since the production of waste from many industries is concentrated in a particular area.

C.Wood

Chip Industry

Residues produced from forest, lumber, and plywood mills, consisting of branches, timber, and plywood off-cuts, are generally used to produce wood chips for pulp, particleboard, and fiberboard. The volume and type of chips produced from the various sources are shown in Table 1. The softwood chips are produced from wastes from lumber sawmills more than from logs of small diameter. Use of the waste from demolition of wooden buildings and pallets, which are composed mainly of softwood, tends lo be gradually increasing

851

Recycling of Wood and Fiber Products

Wood waste from forest l l

l-lCharcoal Ir

Wood particle cement board Fiberboard Pulp cement board

FIGURE 1

Brief outline of distribution of woodwaste.

around the urban areas. Most of the hardwood chips are produced from logs of miscellaneous species. As most of the lumber mills in Japan are minor enterprises and are widely dispersed, transportation costs and the low prices of imported chips often restrict the effective use of wood waste. The wood integration plan as described could make a definite contribution to better utilization of wood waste.

D.

ParticleboardandFiberboard

The particleboard and fiberboard industries have made a considerable contribution to the utilization ofwood waste produced from lumber and plywood mills, so-called ordinary chips. Figure 2 shows the annual particleboard production in 11 Japanese industries, which produce about 80% of the total amount of Japanese production. The shaded portion indicates the ratio of recycled wood chips. The sum of the shaded portion reached about 20% of the total production. According to a survey 10 years ago, only a few mills used

TABLE 1 Origin of Wood Chips, 1997 ( X 1000 m') [3] Wastes from Grouping

Total

~ _ _ __

~~

Total Softwood Species Type of mill

Log

Own mill

Wastes . . from from Other forest

Wastes demolition

~~

11,165 5,812 4,708

Hardwood Lumber and chips Plywood, Ilooring. and chips Chips only

Total

mill

7,009 1,234 4,156 3,474 7,083 4.533 4.952 2,063 76

4,006

4,623 5,290 4,244 522143 379

141.189 1,046 419

76 -

76

2,645

784

-

14

770

7 7 7

63 I 47 8 IS3 61

-

-

7

570

852

Arlma

0

4.’

0 0 0 0

Recycledwoodchip

10

A

B

C

D

E

F

G

H

I

J

K

PB Maker

FIGURE 2 Particleboard production by 1 1 makers in 1992, and the proportion of recycled wood used.

recycled chips. Changes in conditions for procuring raw material could have elevated the importance of recycled wood in particleboard production over the last decade, as described in Section 111.

E.Wood-ParticleCementBoard Wood-particlecementboard of Japanese Industrial Standard is used as sheathing and fireproof material for buildings. The wood chips are produced from lumber, plywood, and from the demolition lumber of wooden structures. InJapan, large quantities of plywood are used as formworkpanels for concrete construction in buildings and civil engineering works. The waste panels from this work are contaminated withcementand sand, and are usuallydumped or burned.Ways of utilizing this material by crushing the waste panels and producinga type of wood-particle cementboardhavebeen investigated, butwhetherthis is technically or economically feasible depends on collecting wood waste constantly and expanding the market for this type of board. II. WOOD RESOURCES FROM DEMOLISHED BUILDINGS According to the Ministry of Health and Welfare’s “Change in Waste by Industry” report, in 1994 approximately 8 million tons of wood debris were transported from demolition sites. When this is converted to a volume basis, the figure exceeds 11 million cubic meters of wood waste. The ratio of reuse is approximately 36%. Additional waste such as furniture, pallets, leftover waste materials from new construction sites, etc., are thrown together, making up the difference. The actual volume of wood which can be reused varies considerably according to the demolition method employed. Below, the desired situation and problems of recycling or cascade use of wood demolition material are stated. (1) The nature of wood waste has changed, the wood particles have become smaller, more mixed with foreign matter, and are of many varieties. The use of a magnet can easily remove the steel objects such as nails, but nonferrous objects such as aluminum, plastic, cement, gypsum board, etc., present difficulttechnical problems inthe sorting and selection of chips, which need to be solved.

Recycling of Wood and Fiber Products

853

( 2 ) The condition and amount of the chips are determined by the manner in which the building is demolished-primarily, whether it was done by machine or by hand. AS shown in Table 2, this determines whether the chips are of suitable quality for pulping, board production, fuel chips, or just plain garbage, and has a tremendous effect on the price and volume of chipsobtained.Therefore, it is necessary to establish a means of sorting the chips according to quality so that they can be properly utilized to their maximum potential. Already, there is a slowly growing trend at the chip collection sites for a minimum acceptable level of chip quality. On a regional basis, there are instances where one can see thought has been put into the recycling of demolition material as well as cases where there appears to be no consideration at all as to the reuse of the recovered wood waste. The important deciding factor seems to be whether there is a facility to handle the wood waste in the locality, or sufficient volume to justify such a facility. (3) Ideally, the demolition and site preparation would be done carefully, but in many instances these steps are rushed, exacerbated by a shortage of labor that does not allow for hand demolition andwood waste segregation. As a result, the reliance on machine demolition damages muchof the material and introduces foreign matter into thewood waste. This also adversely affects the efficiency of the waste transportation, as it becomes impossible to load the machine-demolished and -loaded material in a way which maximizes the use of the truck. This adds further cost to a low-value waste. By wayof comparison, let us lookat an example of a machine-demolished but material-segregated case and a totally machine-demolished case where no effort is made to recover any of the material. In the sorted case, where the segregated demolition material can be more neatly loaded onto a truck, the load will weigh about 0.72 ton per cubic meter. compared to 0.38 ton per cubic meter for the unsorted case. Thus, a truck in the sorted case will carry almost twice the load of one in the unsorted case. Also, among the mixed waste to dump, the volume of wood in the unsorted case is about five times that of the sorted case. If the demolition material were to be separated, it would be possible to recycle a greater portion of it; however, when this is not done, the material will be disposed of as mixed waste and will not be recycled. The reliance on mechanical demolition arises from a limited amount of time available to devote to the job, the ease with which demolition may be conducted by machine compared to hand demolition,and a general shortage of labor. As a result, a larger volume of garbage is generated which is difficult to reuse or recycle.

TABLE 2

Wood Waste from Timber Structures Demolished by Machine or by Hand 171 VoIume/floor area (m’/rn2)

Demolished by machine Demolished by machine and hand Dcmolished by hand Raw material for

Mixed wood waste with debris

Wood waste without debris

0.038 0.033 0.07 1

0.070 0 . I00 0.120

Fuel Particleboard

Paper Fiberboard Particleboard

Total IO8 0. I33

0 .

0.19 1

Arima

854

(4) The generation and collection of waste material varies widely across regions and also by seasons, making it very difficult to obtain a stable supply and quality. Therefore, individual companies that must consider profit implications of using a new material will be reluctant to do so unless the supply is stable and the costs are low. Presently, the price of chips from demolition material is determined by their quality and suitability for fuel, board, or pulp production. At the same time, chips of pulping quality must compete with chips of a virgin nature, while fuel chips must compete with fossil fuels. Foreign currencies have a huge impact on the competition of these materials. (S) With greater urbanization, it has become very difficult to build recycling facilities, due to site restrictions. operating hours, and noise-prevention problems. For more distantly located sites, working hours and transportation considerations create a large burden, making it unprofitable and unfeasible to collect and process waste wood material. Because efforts on behalf of the government and individual companies are being conducted piecemeal and according to individual agendas, there is little active effort at coordinating the policies of recycling and garbage disposal. (6) The wood products industry has accepted the inevitability of using demolition from only a cost-and-supply viewpoint, but the usage will be determined by a balance of technological risk associated with manufacturing and the cost of raw materials. Accordingly. if a new raw material were to have stable price and supply, the wood products industry would naturally tend to use the new material. The recovered wood waste would tend to become a reduction i n trash. In this scenario, competition between new raw materials and the demolition wood waste on these simple economic terms would potcntially negate environmental conservation efforts in regard to the utilization of wood waste from demolition sites.

111.

THE PRESENT SITUATION FOR DEMOLITION WOOD CHIPS

From a technological standpoint it has become possible to utilize wood resources, including recovered demolition material. in a systematic cascade-type approach described earlier, where finally the wood is converted to carbon dioxide (CO,) by incineration. However, the great volumc of material generated creates difficulties in recycling and incineration with respect to social and economic efficiencies. According to the different methods of demolition, primarily mechanical and hand. the quality of the wood waste is changed where the material may be suitable raw material for manufactured wood products, chips, or unusable trash. I n order to prevent the recovered wood from becoming trash to be thrown away, it has become clear that an active and aggressive effort is needed. In spite of the fact that wood waste recycling efforts have shown positive results compared to the past, the recycling of demolition wood waste (demolition chips) still finds itself i n a very difficult situation. Wood chips arc generally segregated for utilization i n pulp, wood-based composite boards, or fuel production. In the regions around Tokyo, Chubu, and Osaka which are close to the typical urban areas, the composition of the chip material in the region’s 35 chip facilities is shown i n Table 3 . Approximately half of the chip materials are collected from the waste of house demolition. The waste from concrete-form. pallet, and packaging demolition also is considerably used because it is easy to collect and also to remove foreign debris. I n the casc of the board industrics, the large proportion of chips originating from the demolition of houses can easily be seen, a s shown i n Table 4.

Recycling of Wood and Fiber Products

855

TABLE 3 Composition Ratio of Chip Materials in Chip Recycling Facilities

[lo]

Waste from m i l l

Grouping

X 1000

area of Total Kanto

Chubu Kansai

(tons) 654 21 I 272 3.9171

Waste from demolition (%)

(%o)

Total

Wood product Housing House 5.3 1.9 9.2 3.2

Other Pallet

Concrete

Packaging

5 .4 12.0 I .2

54.4 48.4 54.1 62.3

5.9 6.7 4.8 6.5

8.5 9.5 7.0 9.7

forms

(c/)

13.2 19.2 IO. I 10.8

7.3 2.3 13.6 3.5

Table 5 shows the breakdown of industries which utilize wood waste chips, where the users are forest products-related industries.Particleboard and wood-particlecement board industries use recycled chip materials mostly as raw material. In fiberboard and pulp industries, the ratio of fuel use tends to increase because the quality level of raw material is required to be high. In the plywood, gypsum board, and dye industries, wood chips are used as fuel for drying. Table 6 shows, in order of relative importance, the primary considerations in chip quality as a raw material for production of pulp and boards, and for use asfuel.Pulp producers presently accept very little demolition wood chips for pulp production and are not actively pursuing furtherexpansion i n demolition wood chipacquisition.Thepulp sectors which do not accept demolition chips are involved primarily in the production of paperboard products such as cardboard. Demolition chips have great difficulty competing with virgin fiber sources due to the presence of foreign particles in the chips, cost, and other factors which make demolition chips unattractive to pulp producers. At present the supply and price of imported chips and waste from other wood industry sectors is both stable and economical, which is probably the main reason for the lack of interest and effort regarding greater demolition wood chip utilization in the pulp industry. In Japan, the recycling of old paper such as newspaper is very high; nevertheless, the industry is in a difficult situation and the incineration of large volumes of paper presents a considerable burden when considering society's economic and social efficiencies and priorities. If the chips from demolition wood waste are sorted, the negative aspects of these chips become negligible or nonexistent and the wood-based boards produced from such material can receive recognition by means of the Ecomark seal. Using demolition chips alsosaves on dryingexpenses,as they are usually drier than material from alternative sources. However, in recent years the percentage of nonferrous waste material mixed with

TABLE 4

Composition Ratio of Recycling Chip Materials in theBoard Industries [ I O ] Waste from demolition ( 7 6 )

Grouping of

Total xIO00

products

(tons)

House

Packaging

Pallet

form

I37 95 35

67.7 67.7 98.0

13.8 8.6

18.7 18.7

12.9 I .4 0.9

Particleboard Fiberboard Wood-particle ccment board

Concretc

1.1

~

Other (c/)

3.6 3.6 -

Arirna

856

TABLE 5 UsageRatio of Recycling Chip Materials in Industries [ I O ]

Grouping of

material Fuel, Raw

industry

tons

tons (%)

Particleboard Fiberboard Pulp Wood-particle cement board Plywood Gypsum board Dye Other Total

(%)

Other, tons

Total, tons

(%l

(%Io)

29,956 (2 1.6) 3,8 18 (2.7) 85,264 (6 I .4) 9,920 (7.1) 960 (0.7) 4,000 (2.9) 0 (0) 4,968 (3.6)

38,316 (7.5) 170,993 (33.5) 125,300 (24.5) 63,125 ( 12.4)

4,120 ( 100)

34,168 (5.2) 20,575 (3.1) 177,449 (27.1) 9,920 ( 1 .S) 39,276 (6.0) 174,993 (26.8) 125,300 ( 19.2) 8,093 ( I 1 .O)

138,886 (100) (2 I .2)

510,888 (100) (78.1)

4,120 ( 100) (0.6)

653.894 ( 100) (100)

4,2 I2 (0.8) 16,757 (3.3) 92,185 (18.0) 0 (0)

the wood chips has created difficulties, as the simple use of a magnetic sorting system is no longer sufficient. Accordingly, it becomesquiteimportant to what degreesortingis carried out at the demolition site. Moreover, the nonselection of materials, especially ones which will create waste disposal problems when demolition ultimately occurs, becomes vital from theconception of the building process. Along theselines, material selection based simply on functionality should be discarded in favor of greater consideration of the complete life cycle of the materials up to the ultimate conversion to CO, with incineration. Because the scale of wood-based material factories, such a particleboard, is limited due to site restrictions among other factors, geographic factors become important when considering the use of demolition chips for board production. In the wood industries, the waste raw material which does not find its way into pulp or board production is often burned to supplement energy requirements of the mills or sent for incineration elsewhere. Demolition wood chips used for fuel have comeinto greater use as an economic alternative to fossil fuels, and are often used as fuel for public baths and burned in boilers of various industries. However, in recent years the use of fuel chips has come into difficult times due to sagging prices and volumes. The economic costs of conveyor maintenance, ash disposal, and personnel for wood waste boilers have become increasingly burdensome, and many facilities are increasingly switching to oil and, especially, gas.

TABLE 6 Problems in Utilizing Demolition Chips as a RawMaterial Input (ranked in order of importance) [ 101 Rank

1 2 3 4 5 6

Pulp

Foreign matter Collection Chip quality Cost Finished goods quality Demand

Fuel Board production Collection Foreign matter Chip quality Demand Finished goods quality Lack of' awareness

Collection Demand Pricing Chip quality

Foreign nlatter Cost

Recycling Products Fiber of Wood and

857

The oil shocks of the past sounded the alarm of too much reliance on limited oil resources across all industries and sectors, including those related to building materials, where too much energy was consumed in the production, assembly, transportation, and demolition relatcd to those materials, at the same time came increased awareness of wasteful resource consumption. The use of wood and wood materials as a supplemental and alternative energy source was examined; however, burning wood has a lower heat efficiency, and in ordertoobtain the sameamount of energyas by burningoil, a greater volume of CO2 would be generated and released to the atmosphere, contributing a greater amount to the global warmingphenomenon. In otherwords, the move toward burning wood as an alternative to oil, though possible, provides an inferior choice concerning the environmental problems associated with excessive COz generation. When viewed from the point of efficiency and economics, the only likely trend is a move to oil and gas. However, at the end of its life cycle, wood will eventually be burned, and wood waste destined for incineration (versus fuel generation) combined with the burning of oil or gas used to fuel the incinerators themselves results in essentially a double release of carbon to the atmosphere. As a consequence, the burning of wood waste should be the last step i n a series of steps which utilizes the wood through a cascade-type recycling process. In this vein, it is necessary to give due consideration to effective use of waste materials, including the safety of the waste and any by-products from utilizing that waste. In other words, simply incinerating wood waste materials should be avoided if possible, and consideration should be given to systems which can capture and utilize the energy contained in the wood. The city of Sapporo,Japan, has wastedisposalresourcefacilities which produce solid-waste fuel, fuel chips, and chips for particleboard production. The uses of this waste include heat generation for greenhouses, electricity generation, and supplemental energy for industrial facilities. In order to carry out the proper sorting and material selection and to adjust to the seasonal variations in the waste types and quality, the use of silos becomes a vital factor in enabling operations to proceed smoothly.

IV. CONVERTING WOOD WASTE TO CHARCOAL AS A MEANS OF UTILIZATION AND DISPOSAL For Japan’s construction industry, and the housing industry in particular, the era of emphasis on volume of construction has yielded to a new era where emphasis is placed on quality. From this has arisen the “scrap and build” phenomenon, where large numbers of structures built i n previous years when volume was emphasized are demolished and new, high-quality buildings take their place. Because of the large numbers of buildings being scrapped after a relatively short life span, the problem of waste disposal and recycling has become an unavoidable issue. It is possible to reuse and recycle the waste from residential demolition sites as well asother public works and construction/demolitionsitesforvarious uses such as pulp, board, and fuel chips in the cascade-type manner describedearlicr. However, the huge volumes of waste being generated and the economic efficiency problems associated with utilizing waste chips have created a very difficult situation regarding the full utilization of the demolition chips. To meet the demand for reducing the volume of waste, i n the worst case, it could simply be incinerated or decomposed in a manner with comparatively few pollution problems, but disposing of the huge volumes of waste i n such a manner would emit large quantities of CO, into the atmosphere, which should be avoided to the degree possible.

Arima

858

From this background, the goal is to develop further the reuse and recycling alternatives while simultaneously reducing the volume of wood waste and keeping the CO, in a stored form (i.e., not released into the atmosphere until as late a time as possible) by converting waste wood resources into charcoal products with sequential carbonizing heat treatment steps. In the past the production of the charcoals, from ordinary general-purpose charcoal to special types such as for filters, has relied on relatively clean wood resources such as virgin material, and not material such as demolition waste. Charcoal is a very stable material compared to other biodegradable materials. Because of this, it is possible to change the method of disposal of charcoal, since it may be added to just about any soil anywhere without any adverse effects, thus alleviating some pressure on landfills. Because the physical properties differ according to the different levels of heat treatment, proper heat treatment is aimed at producing a charcoal suitable for prescribed uses. Charcoal is used for soil improvement, absorption of odors, moisture, and for preservatives. Charcoal was mixed with soil in ordinary house planters and plant growth and soil conditions were shown to be improved.

REFERENCES I. 2. 3. 4. S.

6. 7. 8. 9.

IO. 11.

12.

13. 14. 15.

16. 17. 1 8.

T. Arima, Rev. Forest Culture, 13:IO9 ( 1992). Research Group on Forest Products Administration, Wood Supply and Wood Industry 1997, Wood Industry Integration Plan Areas 27 1-283 ( 1997). Research Group on Forest Products Administration, Wood Supply and Wood Industry 1997, Wood Chip Industry, 189-196 (1997). S. Suzuki, Proc. Int. RlLEM Workxllop, Tokyo, p. 223 (1995). T. Arima, Proc. Preserlt arrd Future of‘ Wood Wctstc~,Japan Wood Research Society Hokkaido Branch. p. I14 (1993). T. Arima, J. J p 7 . Agric. Sys. Soc.. 8:69 ( 1992). Osaka Industrial Waste Corporation. Report on Generating and Recycling of Wood Waste i n the Construction Site ( 1988). Japan Housing Wood Technology Center, Report on Recycling of Wood Waste, p. 36 ( 1992). Association Housing Demolition, Composition Analysis on Waste from Wooden House Demolition ( 1993). Japan Housing Wood Technology Center, Report on Recycling of Wood Waste (I). p. 6 (1994). Japan Wood Research Society, Wood Scirncc~m t d T ~ ~ . k r ~ o /IV o g l~~t/rt.sfrirrl y r r r d 1Ir~il.vW ~ t c . . Wood Chip, p. 7 ( 1996). Japan Wood Research Society. Wood Scierrcc. ( I d Techrrology IV Irltlustrirrl rrrrd Drti1.v W~ISIL’, Roctrd, p. 104 (1996). T. Arirna. Rcsectrck Ptcp”:s o r 1 H o u s i n g trntl L ~ I I I ~211255 , ( 1997). K. Fujike, Pro(.. Prt~serrrt r r r d Future qf’ Wood Wrl.stC.Japan Wood Research Society Hokkaido Branch, p.22 (1993). T. Arima, Reports on Grant-in-Aid for Developmental Scientific Research Development of Fundamental Technics for Recycling Wood Resources, p. 42 ( 1994). T. Arima. S. Kobayashi, F. Socda, and M. Tokuda, RC): Forest Culfttrr. IS: I78 ( 1 997). Japan Wood Research Society. Wood Scicwcc t u r d T d ~ ~ r o l o II! g y Irtthrstritrl rrrlrl L)tti!\. Wtr.stt>. Chcrrt.otr/. p. 78 (1996). Japan Housing Wood Technology Center. Report on Recycling 01’ Wood Waste, Charcoal. p. 66 ( 1997).

25 Pulping Chemistry Goran Gellerstedt Royal Institute of Technology, Stockholm, Sweden

1.

INTRODUCTION

The pulping of wood for the production of paper, board, and other cellulosic products has a large economic impact in several countries in the world. North America, Scandinavia and, more recently, SouthAmerica are importantareasforpulpproduction,sincethe availability of suitable wood species is high. Japan has a large pulp industry which, to a large extent, is based on imported wood, whereas the pulp industry in Western Europe is of amoderatesize. In otherareas of theworld,suchasChina, annual plants play an important role as fiber supply, but here, as well as in other developing areas, new pulp mills utilizing wood are rapidly increasing in number. Secondary fibers constitute an important fiber resource that has been used for paper production in countries with no or little domestic wood supply. In the future, the importance of secondary fibers will increase considerably, and legislation in several countries has been used to force the paper industry to further increase the content of secondary fibers in a variety of paper products. The chemistry of pulping has developed in parallel with the development in pulping technologies. In several pulp-producingcountries, central researchinstitutes have been created to support the industry with fundamental research on selected problem areas of technical significance. Today, the chemicalknowledgeaboutpulping and bleaching of wood is comprehensive, and numerous articles, reviews, and book chapters are available on the subject. Nevertheless, due to the complexity of wood as a composite of polymers, our knowledge about the chemistry and chemical reactions involved when wood is converted to pulp of different types is still far from being complete. In this chapter, the chemistry involved in the production of unbleached and bleached mechanical and chemical pulps is discussed. with an emphasis on reactions which are thought to play a significant role in technical pulping systems.

II.

(CHEM1)MECHANICALPULPING

A.

TechnicalOverview

In mechanical pulping, wood is disintegrated into fibers by grinding or refining, using a rotating grindstonc or a disk refiner, respcctively. In the former case, wood logs are pressed against the rough stone surface with simultaneous water spraying. In the latter case, refin859

Gellerstedt

860

ing is done by forcing wood chipsto pass from the center and outward between two circular disks, with at least one rotating at a high speed. The inner surfaces of the disks are covered with bars and grooves of successively smaller dimensions. The disintegration of wood to mechanical pulp results in a broad distribution of fiber sizes, including course fibers and fiber bundles, whole fibers, fiber fragments, and small particles. To some extent the relative distribution of the various fractions is determined by the process and the conditions needed to meet various end uses. The resulting pulps are termed stone groundwood (SGW) and thermomechanical pulp (TMP), respectively. Refining is also used to make refiner mechanical pulp (RMP), which differs from TMP by the way the chips are pretreated in the process. Fresh spruce is the preferred wood species for making mechanical pulp, although other species can be used as well. The majority of the mechanical pulp produced is used for making newsprint and other “wood-containing” printing papers in large integrated mills. The pulp yield in mechanical pulping is usually of the order of 95-989’0; i.e., a l l the wood constituents are retained i n the finished product. Brightness isan important quality parameter for spruce mechanical pulps. Values around 60-63% IS0 are generally encountered. For several paper products this is too low, and bleaching is frequently used to adjust the brightness to different levels in the range of around 80% ISO, which, at present, constitutes the upper technical limit. The bleaching of mechanical pulps is done in such a way that a l l wood constituents are retained, using either dithionite under neutral conditions or alkaline hydrogen peroxide. This“ligninretaining bleaching” results in bright pulps which, however, suffer from brightness reversion caused by either heat or light. For this reason, bleached mechanical pulps cannot replace bleached chemical pulps in paper products where brightness permanence is important. Mechanical pulping can also be performed on chips which have been pretreated with small amounts of chemicals, usually sodium sulfite, either alone (softwoods) or in combination with sodium hydroxide (hardwoods). A somewhat lower pulp yield is obtained, but strength and brightness properties are improved. These pulps are referred to as chemithermomechanical (CTMP) or chemimechanical pulps (CMP), and they are often further bleached with alkaline hydrogen peroxide.

B.

Reactive Structures in Wood

The native color of wood can vary largely between species, but in bright softwoods, such as spruce, the predominant contribution comes from the lignin. Although the majority of lignin structures are colorless, the yellowish hue of wood can be attributed to the presence of small amounts of colored lignin structures. The quantitatively most important of these are end groups of the cinnamaldehyde type, which are present in softwoods with a frequency of approximately 3-4 units per 100 phenylpropane units [ 11. Among other chromophoric structures in native lignin, trace amounts of ortho- and pum-quinones are thought to be major contributors [2,3], although their presence is difficult to demonstrate directly. Phenolic lignin structures containing a-carbonyl groups may also contribute to the color of wood, but also in this case, no unequivocal proof of their existence is available. The structures and wavelengths of maximum absorbance for some wood chromophores are given in Fig. I . Native lignin also contains a variety of colorlcss structures which under rather mild reaction conditions may be converted into chromophores. The presence of small amounts of hydroquinoncs and catechols has been demonstrated 1431 and are probably the result

861

Pulping Chemistry

0

OCH,

0

+o 0

0

a,

-

amax

-

-

a,,

390 410 nm (420 440 nm)

a,,

-

-

-

335 350 nm (340 350 nm)

-

360 nm (410 - 420 nm)

-

310 nm

FIGURE 1 Suggested chromophores in wood and their approximate light absorption maxima (values in parentheses are for the solid state).

of redox reactions and incomplete methylation, respectively, in the growing tree. Under oxidative conditions, as in the presence of air, such structures are easily oxidized, with formation of the corresponding quinone structure [6]. End groups of the cinnamyl alcohol type do not contribute to the wood color directly. They constitute, however, structures of high reactivity in, e.g., oxidative processes. Since the frequency of these structures is rather high in the native lignin (approximately 3-4 per 100 phenylpropaneunits) 171, evena reaction proceeding to aminuteextent may contribute to color changes. Among the major structural units in lignin, the p-1 structures are of particular interest. In mild acidic or alkaline conditions such structures are easily converted into the corresponding stilbene structure according to the reaction shown in Fig. 2 [ & S ] . The stilbene, in turn, is easily oxidized to form deeply colored quinones of various types.

C.

Reactions with Sodium Sulfite

The pretreatment of chips at temperatures around 120- 130°C with 1-4% sodium sulfite solution, followed by refining, results in the production of CTMP. The weakly alkaline sulfitesolution reacts with the lignin to someextent,asshown in Fig. 3 [ 101. These reactions are not, however, uniform throughout the lignin matrix but concentrated to the outer part of the fiber [ I I ] . The preferential hydrophilization of the primary wall gives a

Gellerstedt

862

H'or

HO-

P

OH

OH

FIGURE 2 Formation of a stilbene from p- 1 structures in wood.

swelling effect which directs the fiber-fiber fracture to this cell wall layer. This results in a pulp having more intact and homogeneous fibers as compared to a mechanical pulp. The reactions between sulfite and wood are thought to involve sulfonation of reactive lignin structures as the predominant reaction mode. Thus, coniferyl alcohol structures, which are abundant in the primary wall 1121, react readily with the strongly nucleophilic sulfite ions to form a series of sulfonated coniferyl alcohol structures [ 13,141. The related structure, coniferaldehyde, is even more reactive and easily converted into sulfonatesunder CTMP conditions [ 151. Other aldehydes, such as vanillin structures, can be assumed to form a-hydroxy-sulfonates in the well-known aldehyde-bisulfite equilibrium reaction. Conjugated carbonyl structures, such as orrho- and para-quinones, will also react with the sulfite ion to form the corresponding aromatic sulfonates [ 161. The various reactions taking place between sulfite and wood chromophores result in a certain bleaching effect, leading to CTMP being somewhat brighter as compared to TMP.

D. Chemical Changes in Refining

1. Stone Groundwood and TMP The defibration of wood to mechanical pulp involves exposure of the wood constituents to high temperatures (150- 170°C) in the presence of water. These conditions result in a

Sulfur content, mmoVkg wood

5

10

15 Time, min

FIGURE 3 Sulfur uptake in wood under CTMP conditions [ I O ] .

Pulping Chemistry

863

color formation which, together with the native wood chromophores. determinesthe brightness of the final pulp. Several factors can influence the degree of color formation in the fibers, including wood quality, content ofbark in the wood, content of transition metal ions in the wood, pH during fiber liberation, presence of sulfite, and presence of oxygen. The importance of these parameters indicates that autoxidation reactions play a role in the discoloration reactions. This is further supported by experiments both with wood [ 17,l S] and with various model compounds [6]. The major contribution is likely to come from the oxidation of reactive phenols in ligninlike catechols and hydroquinones. Such reactions occur rapidly even at room temperature and result in the formation of the corresponding quinone structure. In the presence of metal ions, such as copper or manganese, the oxidation rates are further enhanced. Catechol structures in lignin are also able to form strong chelates with ferrous/ferric ions; these are colored and, once formed,difficult or impossible to remove [19]. Thus, the presence of high amounts of iron in the wood results in very dark-colored mechanical pulp [20]. The presence of bark in defibration is detrimental, since reactive phenolic compounds are easily leached out in the surrounding aqueous solution. Subsequently, these may participate in lignin autoxidation reactions, and new, and more intensely colored addition products linked to the fiber, may be formed [6]. The mechanical action on wood in grinding or refining can alsoalter the lignin structure in such a way that more reactive structures are created. In a series of investigations, several different lignin model compounds have been subjected to either ball milling at room temperature or added to coarse mechanical pulp which subsequently was subjected to further refining [21]. The products obtained from the model compounds reveal that mechanical action on lignin structures may result in both homolytic and heterolytic reactions. Important products include structures with an a-carbonyl group in the side chain, as well as stilbenes; the later originate from p-5 and p-l structures. One example is shown in Fig. 4.

2.

CTMP

The addition of sulfite to wood prior to defibration results in a certain brightening effect, since some wood chromophores are eliminated due to sulfonation. The effect is limited,

CH0

Hi:

II

CH

I

I

HC

It

CH

uo FIGURE 4

Chemicalchange in a p-5 structure as the result of mechanical treatment of wood.

Gellerstedt

864

however, since new chromophores are created in the defibration process. Therefore, the net result on brightness of the simultaneous creation and elimination of chromophores can vary widely due to the exact reaction conditions before and during refining. In particular, the presence of transition metal ions in the wood and/or the process water will affect the extent of autoxidation and other homolytic processes. The simultaneous presence of sulfite and a strong chelating agent such as EDTA or DTPAin the defibration process has been shown to give a synergistic bleaching effect, since the color-forming reactions are minimized while the bleaching effect of sulfite is still operating. Thus, a brightness gain of the order of 8-10% I S 0 is attainable using this mixture [22].

E.

Bleaching Chemistry

1.

Dithionite

Sodium dithionite (sodium hydrosulfite) is a mild reductive bleaching agent which is used to obtain small or moderate brightness gains in mechanical pulps with a maximum increase of around 8 I S 0 units. Although the reducing power in alkaline solution is much higher (eo = 1.12 V versus e,, = 0.66 V), such conditions result in discoloration of the pulp due to rapid alkali-induced discoloration reactions. Consequently, dithionite isusedat a pH around 5-6 and at temperatures around 60°C. Under aerobic conditions, dithionite easily decomposes into sulfite and sulfate according to reaction ( l ) , resulting in a loss of bleaching power, even though the sulfite formed may participate in reactions of the type discussed above. A slower decomposition is encountered under anaerobic conditions, resulting in the formation of sulfite and thiosulfate [reaction (2)]. Thus, the bleaching solution mustbe freshly prepared, which is done either by simple dissolution of the sodium salt in water or by reduction of sulfur dioxide by sodium borohydride. S,Oi-

2S,O:-

+ H,O + 0, = HSO, + HSO, + H,O = 2HSO; + S @ -

The bleaching chemistry of dithionite is a simple reduction of easily reduced chromophores, such as quinones present in lignin [23]. Accordingly. the maximum color reduction is found in the wavelength region of 440-480 nm.

Hydrogen Peroxide The bleaching of mechanical pulp with hydrogen peroxide in alkaline media requires a prior elimination of transition metal ions from the pulp since, otherwise, a rapid decomposition of hydrogen peroxide to oxygen and water takes place according to reaction (3). The metal ion concentration is reduced by treatment of the pulp with DTPA or EDTA at a neutral or slightly alkaline pH, followed by dewatering. The bleaching itself is carried out in the presence of silicate, which acts as both a buffer agent and a stabilizer for the peroxide. The brightness gain in peroxide bleaching is dependent on the charge of peroxide and alkali; however, the bleaching liquor cannot be allowed to contain a large excess of alkali, sincesuchconditions will result in darkening reactions in competition with the bleaching reactions [24]. 2.

H,O,

+ H 0 2 = O2 + H,O + HO-

(3)

The chemistry of peroxide bleaching has been thoroughly studied both with lignin model compounds and with pulps. Analysis of pulp reflectance spectra before and after a

865

Pulping Chemistry

1.6 1.4 1.2

t

A

400

1 I

l

0.8

600

500

wavelength, nm

I

457

FIGURE 5

Relative reflectance spectra of mechanical pulp beforehfter a peroxide stage [25].

peroxide bleaching reveals that two major areas of the spectrum are subject to changes, viz., 320-400 nm and 400-600 nm, as shown in Fig. 5 [25]. The former of these is attributed mainly to coniferaldehyde structures, which are known to react rapidly with the peroxy anions to form the corresponding aromatic aldehyde(Fig. 6) [26]. In phenolic units, the aldehyde may react further in a Dakin reaction, with formation of a hydroquinone structure [27,28].

CH0

I

CH

I

-

CH0

H202 I HO-

OCH,

/

2 HCOOH

+

HCOOH

OCH,

/

OR

+

OR

PH OCH,

0

OH

FIGURE 6

Alkaline peroxide oxidation of a coniferaldehyde structure in wood.

Gellerstedt

866

4

H*~dHO”

OCHj

+

COOH

others

COOH

0

FIGURE 7

Alkaline peroxide oxidation of a pnm-quinone structure in wood.

The elimination of coniferaldehyde structures with peroxide bleaching has been supported both by analysis of wood tissue using UV microscopy [29] and by chemical analysis of pulp samples [30].However, this elimination does not seem to be quantitativeand, even in highly bleached mechanical pulp, coniferaldehyde structures can still be found, thus indicating that the bleaching process is not optimal. Ortho- and yaru-quinones are also attacked by the strongly nucleophilic peroxide anion and degraded to a mixture of carboxylic acids (Fig. 7) [31]. In a second reaction pathway, quinones may also react with peroxide to form the corresponding hydroxy-quinone. Although occurring to only a minute extent, this reaction creates a new chromophore which, despite being a quinone, has a low reactivity toward alkaline peroxide, due to the acidic hydroxyl group. The stoichiometry in peroxide bleaching has been found to vary depending on the charge of peroxide. Thus, at high charges of peroxide, more peroxide is consumed to obtain a certain brightening effect (Fig. 8) [32]. A high charge of peroxide is, however, necessary to reach the highest brightness levels and, consequently, such bleaching systems are quite expensive. The reason for the increase in peroxide consumption seems to be a formation of radicals through a homolytical decomposition reaction of hydrogen peroxide according to reaction (4) [33-351. This reaction is catalyzed by certain transition metal ions, such as manganese, and results in the formation of hydroxyl radicals and superoxide ions. In the absence of other substrates, these radicals will combine to give oxygen and water [reaction (5)]; but the presence of pulp may alter the reaction pathways and a large

Relative bleaching effect

Relative bleaching effect

5-

5-

L-

10.7 11.1 11.5 pH;

L-

P

3-

3-

2-

2-

‘1

1-

I

0.2

0.6

1.0

Consumption of

0.2

b o 2, mol/kg

0.6

1.0

pulp

FIGURE 8 Stoichiometry in the alkaline peroxide bleaching of mechanical pulp [ 3 2 ] .PH, denotes the initial pH in the pure bleaching liquor.

Pulping Chemistry

867

number of both lignin and carbohydrate reactions are possible. The effect of these radicals on wood chromophores is unknown; however, it has been shown that bleaching mechanical pulp in the presence of a radical scavenger results i n asomewhat reduced bleaching response under otherwise identical conditions [36].

+ HO, = 0,- + HO' + H,O + HO' = 0, + HO-

H20,

(4)

0;

(5)

A complete conversion of coniferaldehyde and quinone structures to the reaction products outlined in Figs. 6 and 7 would require a consumption of hydrogen peroxide of the order of 0.20-0.24 mol/kg of pulp. Such values are found when pulp is bleached with a low charge of peroxide, as shown in Fig. 8. Therefore, the presence of unknown types of chromophores in wood, which have a more unfavorable stoichiometry or, alternatively, require the presence of oxygen radicals to be eliminated, cannot be ruled out. An alternative explanation, involving the formation of new chromophores, e.g., due to alkali- and/ or peroxide-induced reactions, is also possible.

F. Chemistry of Yellowing Both unbleached and bleached mechanical pulps undergo brightness reversion when exposed to daylight or heat treatment. For a given set of aging conditions, the extent of color formation is highest when the brightness of the original pulp is high [37]. For bleached mechanical pulps, these yellowing reactions prohibit a wide use of the pulp in high-quality paper products and thus constitute a severe technical problem. The reflectance spectra of bleached mechanical pulp before and after accelerated light-induced yellowing reveal that two wavelength areas change as the result of irradiation, viz., 310-350 nrn and 380-470 nm (Fig. 9) [38]. The former of these can be attributed to the formation of carbonyl groups in conjugation with aromatic rings; the latter is from quinone structures. The possibility of reducing or eliminating the yellowing of mechanical pulps is small, due to the fact that several functional groups present in the lignin absorb daylight and therefore act as sensitizers for photochemical oxidation reactions. The addition of ultra-

AAbs

0.3

t

~ 0 5

0.2

0.1

3

0

250

350

L50

550 Xnm

FIGURE 9 Absorption spectra beforehfter light-induced yellowing of mechanical pulp [ X ] .

868

Gellerstedt

0

2

1

3

Accelerated (h)

Irradiation time & 1

2

3

Daylight (months)

FIGURE 10 Light-induced yellowing of mechanical pulp as a function of irradiation time [40].

violet absorbers, such as hydroxylated benzophenones [ 3 7 ] , or radical scavengers, such as the combination of sodium sulfite and ascorbic acid [391, to paper containing mechanical pulp has been suggested, but the cost is high. The chemistry of’ light-induced yellowing involves at least two different types of reactions. One initial fast reaction takes place within the first hour of accelerated laboratory irradiation and is followed by a second reaction proceeding at a lower rate (Fig. IO) [40]. A suggested mechanism for the initial yellowing reaction is based on the photo-induced oxidation of reactive phenols in lignin, such as catechols and hydroquinones (Fig. 11) [41]. Such structures are present in the native lignin, and additional amounts can be formed on

OH

0

? FIGURE 11 Suggested mechanism for the light-induced yellowing of mechanical pulps.

Pulping Chemistry

869

dithionite or peroxide bleaching, as discussed above. The first step in the reaction, the absorption of light and formation of a phenoxy radical, can be accomplished by the phenol itself [42,43] or by a reaction involving an excited conjugated carbonyl structure [44-461. The latter, in turn, can be either native or formed in the peroxide bleaching stage. At a later stage of yellowing, unspecific monohydric phenols in lignin can be converted to phenoxy radicals and further oxidized to quinones by the action of light and oxygen. Again, conjugated carbonyl structures can act as photosensitizers. In addition to available carbonyl structures, lignin end groups containing conjugated double bonds are themselves photo-active. Forexample, coniferyl alcohol structures can be oxidatively cleaved to yield further aromatic aldehydes [47]. Photo-induced oxidation of benzyl alcohol groups to the corresponding ketyl radical may also constitute an important pathway for discoloration, since such structures, when present in P-aryl ether structures, induce a cleavage of the P-aryl ether linkage and direct formation of a phenoxy radical (Fig. 12) [46, cf. 451. The heat-induced (thermal) aging of mechanical pulps can take place i n the manufacturing process itself, during pulp storage, and in the finished paper product. The color formation seems to be due to autoxidation processes of catechols and hydroquinones, resulting in the formation of the correspondingquinones in reactions similar to those discussed above [6]. In the absence of light, these reactions proceed through phenolate anion intermediates, making them dependent on pH with a minimum around pH 5 . The presence of certain transition metal ions, such as copper or manganese, accelerates the reactions. A simple way of reducing the autooxidation of mechanical pulps is by addition of sulfite at a pH of around 5-6. The simultaneous presence of a chelating agent such as DTPA or EDTA prevents autooxidation of the sulfite and leads to long-term stabilization [37].

-

CHzOH I

Hi:

I: a

+

hv

OCH,

CH30 OCH,

/O

further reactions

FIGURE 12 Direct photo-induced cleavage of an oxidized p-0-4 structure in lignin with formation of a phenoxy radical.

Gellerstedt

870

111.

CHEMICAL PULPING

A.

TechnicalOverview

In chemical pulping, the cellulosic fibers are separated from each other by dissolution of lignin (and part of the carbohydrates) under acidic or alkaline hydrolytic conditions. All chemical pulping processes give fibers which still contain some residual lignin. This is usually measured in an indirect way by determining the consumption of a strong oxidant, potassium permanganate, in a given amount of pulp. The resulting value is termed the kappa number and, although not measuring only lignin 1481, is frequently used to classify the process and the quality of the pulp. The predominant process, kraft pulping, is done by reacting wood chips with a mixture of sodium hydroxide and sodium sulfide for 1-2 h at a temperature in the range of 140-170°C depending on wood species. The process gives brownish pulps with high fiber strength which can be used in a broad range of paper products. Unbleached kraft pulps in a yield range of SO-60% still contain some 8- 15% lignin and are used as such, whereas pulps in lower yield (4S-S0%) with 3-5% residual lignin usually are bleached to full brightness with oxidative bleaching agents. Major advantages of the kraft process are the low sensitivity to wood species and the possibility of producing pulp from wood of inferior quality. However, the malodorous compounds produced in the process and the high capital costs involved in new installations constitute important drawbacks. Nevertheless, the kraft process is the totally dominating chemical pulping process in the world. An alternative to kraft pulping, sulfite pulping, is somewhat olderasa technical process and gives brighter and more easily bleached pulps. Pulping is carried out in an acidic solution containing sulfur dioxide and either calcium, sodium, or magnesium ions at temperatures in the range of 120- 150°C for 6-9 h. The sulfite process is more sensitive to wood species and the resulting fibers are somewhat weaker than those from the kraft process. For special purposes, sulfite pulping of hardwood at a neutral or slightly alkaline pH is carried out. That process results in a high-yield sulfite pulp (NSSC) used for making corrugated medium in board. Several alternative pulping processes have been suggested but, with a few exceptions, these have not reached commercial scale. Various modifications of the kraft process. which are the most important of these, involve addition of either polysulfide, anthraquinone (AQ), or both to the cooking liquor. The resulting pulps are obtained at a somewhat higher yield for a given amount of residual lignin. Sulfur-free pulping with only sodium hydroxide and AQ is a further interesting alternative, resulting in “kraftlike” pulps. Certain hardwood species such as aspen are easily delignified, and solvent pulping can be used to obtain chemical pulp. In this process ethanol is used and the wood chips are pulped at 180- 190°C for 4-6 h.

B.

Chemistry of Pulping in Acidic Media

1. Sulfite Pulping The pulping ofspruce wood using an aqueous solution of acid bisulfite started as an industrial process in Sweden in 1874. The early industry was based on calcium as a counter ion, resulting in severe pollution problems since the used liquor could not be burned to recover the pulping chemicals. Calcium based mills are still in operation in the world, but the more modern mills utilize either sodium or magnesium, thus permitting chemical recovery and energy production through burning of the partially evaporated spent liquor.

Pulping Chemistry

871

The sulfite system contains two equilibrium reactions [reactions (6) and (7)], and the pulping liquor is usually prepared by dissolution of sulfur dioxide in an aqueous solution or a slurry of the appropriate hydroxide. The proportions are chosen such that a certain excess of sulfur dioxide is obtained, resulting in a pH in the range of 1.5-4.0, depending on process and product requirements. In NSSC (neutral sulfite semi-chemical) pulping, a solution of sodium sulfite is used at a slightly alkaline pH.

+ H,O = HSO; + H30' + H,O = SO:- + H30'

SOz H,O

pK,, = 1.9

(6)

HSO,

pK,, = 7.0

(7)

The presence of a base is essential in sulfite pulping, and pulping with sulfur dioxide alone cannot be done since such conditions would result in a dark wood residue with very little lignin dissolution.This is due to the fact that the strongly acidic aqueoussulfur dioxide promotes condensation reactions in the lignin, e.g., between aromatic and benzyl alcoholic carbon atoms, and does not lead to any appreciable amount of sulfonation reaction. The relationship between the total amount of sulfurdioxide and the required amount of base in the cooking liquor has been determined experimentally. Thus, if a certain relationship between the combined and total sulfur dioxide is maintained in the system, the proportions between sulfonation and condensation are such that sulfonation is favored

[W. The desired chemical reaction types in sulfite pulping are sulfonation and acid hydrolysis. Acid catalysed condensation is a nondesirable side reaction. In these reactions, the lignin is solubilized through the introduction of a large number of sulfonate groups in the lignin side chains. This solubilization is further facilitated by the acid-catalyzed hydrolysis of alkyl aryl ether linkages in lignin and of benzyl alkyl ether linkages between lignin and carbohydrates. Further acid hydrolysis takes place in the polysaccharides, resulting in a certain loss of pulp yield and in the formation of low-molecular-weight neutral sugars suitable for fermentation to ethanol. ( I . Lignin Chemistry. The rates for sulfonation and dissolution of lignin in sulfite pulping have been shown to be dependent on the pH of the liquor. Thus, sulfonation is always the fastest reaction type, and at a low pH a complete sulfonation of all lignin units takes place within a few hours of pulping (Fig. 13) [50].The hydrolytic reactions resulting in cleavage of ether linkages between lignin and polysaccharides, as well as between lignin units, seem to be somewhat slower. Together with possible restrictions in the mobility of dissolved lignin fragments through the fiber wall, the resulting delignification thus shows an apparent retardation as compared to sulfonation. Around neutral pH values, the sulfonation of lignin is much more selective and only around 20% of the phenylpropane units react, although at a comparably fast rate. Therefore, the dissolution of lignin is limited. The initial sulfonation reactions in the pH region are the same as those described for CTMP. Under more strongly alkaline conditions, sulfite pulping gives chemical pulps with characteristics similar to kraft pulps [51,52]. The reactions of lignin in acid sulfite pulping have been studied using a variety of model compounds representing the different structural units [53,54].The major reaction mode is a sulfonation of benzylic carbon atoms through the intermediate formationof carbonium ions, as outlined in Fig. 14. In principle, both phenolic and etherified lignin units can react, although the rates for phenolic structures seem to be somewhat higher [ S ] . A stronginfluence on the rate of sulfonation is also exerted by the pH under otherwise identical conditions. At higher pH values, i.e., around neutral and above, the sulfonation of lignin becomes more selective and only phenolic structures react. The reaction involves addition of sulfite

872

Gellerstedt

FIGURE 13 Rate of sulfonation (-pulping [ 5 0 ] .

-)

and dissolution (-

) of lignin

in acidic sulfite

CHZOH

I

CHzOH

I

CHzOH

I

HC-

HC-

I

HC-

HC-0-R

I

HSO,0

SOz/ H @

o +

-e

' @

OCH,

' -e

/O

OCH,

'

R-OH

+ H

OCH,

/O

FIGURE 14 Mechanism of sulfonation of lignin in acidic sulfitepulping.

CHzOH

CHzOH

I HCI

I

HC-

I

HC-OR

CHzOH

I

HC-

I

H0 ~

OCH3 OH

OCH,

0

OCH3 OH

FIGURE 15 Sulfonation of a phenolic lignin structure in neutral sulfite pulping.

further reactions

0

Pulping Chemistry

873

ions to intermediate quinone methide structures, as demonstrated in Fig. 15 1561. Under these conditions,further sulfonation may occurand, in P-aryl ether structures, the P-substituent can be eliminated, resulting in a partial degradation of the lignin macromolecule. Alkaline sulfite pulping conditions will result in more fragmentation of the lignin, since the cleavage of @-aryl ether structures is enhanced due to the presence of hydroxyl ions in addition to the sulfite 1541. The degree of sulfonation of the lignin will decrease, however, due to secondary elimination reactions of sulfonate groups 1571. The facile formation of new carbon-carbon bonds in lignin during acid conditions is due to the presence of free aromatic carbon atoms in the para position to a methoxyl group (C-6 position). This carbon atom will easily combine with an adjacent carbonium ion formed through elimination of the oxygen function from a benzyl alcohol or ether. In model experiments, this type of condensation has been shown to be favored over sulfonation in structures where both types of carbon atoms are present in a sterically suitable arrangement [58,591. It is also well known that acid sulfite pulping of wood species such as pine is less favorable than of spruce, since the heartwood in pine contains extractives of the pinosylvin type [60]. Under acidicconditions, the C-6 carbon atom in the 3 5 dihydroxy (or 3-hydroxy-5-methoxy)-substitutedaromatic ring in pinosylvin will compete with the bisulfite ion for the intermediate carbonium ions in lignin. As a result, several positions in the lignin will never become sulfonated and, consequently, such lignin fragments will be less soluble in the pulping liquor. b. Carbohydrates.The major reaction of carbohydrates in acid sulfite pulping is the hydrolysis of glucosidic linkages, resulting in a loss of polysaccharides and thus of yield. In particular, the low-molecular-weight hemicelluloses are lost in the process; a spruce sulfite pulp may contain less than 30% of the original glucomannan and around 50% of the xylan 1611. Most of the degraded polysaccharides are present in the pulping liquor as low-molecular-weight neutral sugars. To some extent these may be further converted in acid-catalyzed reactions to furfural (from pentoses) and hydroxymethylfurfural by repeated water elimination, followed by cyclization.

2. Organosolv Pulping Pulping with ethanol at a high temperature can be done with wood species such as aspen and poplar, as well as with annual plants [62,63]. Due to the facile liberation of lowmolecular-weight organic acids, such as acetic acid, the pulping process is carried out in a weakly acidic environment. Hydrolytic reactions, resulting in a liberation of lignin from the lignin-carbohydrate matrix, is assumed to play an important role [64]. In poplar species this is particularly facilitated since the lignin contains para-hydroxybenzoic acid end groups that are linked to carbohydrates through ester linkages 165,661. In grass species, ferulic acid plays a similar role as “spacer” 1671. Homolytic cleavage of linkages in lignin may, however, also occur and contribute to a lignin fragmentation and subsequent dissolution in the solvent 168,691. The carbohydrates in organosolv pulping seem to undergo reactions similar to those occurring in sulfite pulping; neutral sugars, as well as furfural, can be found in the spent liquor 1621.

C. Chemistry of Alkaline Pulping 1. Introduction Approximately 10 years after the first sulfite mill was started in the world, kraft pulping was introduced on a commercial scale. The process gained considerable value with the

Gellerstedt

874

finding that the addition of sulfide to a soda process gave a pulp of superior strength as compared to soda alone. In addition, the pulping time could be considerably shortened. Although the pulp was dark in color, the possibility of obtaining strong pulps from a variety ofraw materials made the process attractive. The poor bleachability prevented major expansion, however, until chlorine dioxide was developed as a bleaching agent in the 1940s, thus permitting the production of fully bleached kraft pulps. The kraft pulping liquor is a mixture of sodium hydroxide and sodium sulfide; the charge of each chemical is usually expressed on a wood basis as effective alkali and sulfidity, respectively. The sulfide system contains two equilibria, as shown in reactions (8) and (9). The equilibrium constants are such that, under all conditions prevailing in the digester, the sulfur is present as hydrosulfide ions [70]. These ions accelerate the rate of delignification, the rate increase being proportional to the charge of sodium sulfide. No similar influence on the rate of carbohydrate dissolution is found according to Fig. 16 [71], but from the figure it is obvious that the carbohydrate losses in kraft pulping are substantial.

+ H,O = HS- + H,O+ HS- + HO- = S” + H 2 0 H,S

pK,, = 7.1

(8)

pK,, = -0.7

(9)

gOl 80

J

20

\mc. 0

135OC. 165OC.

40

9

80

-

I I I 160 200 240 I20 COOKING TIME (TOTAL) MIN

FIGURE 16 Dissolved amounts of lignin and carbohydrates in kraft and soda pulping [71].

Pulping Chemistry

a75

If the dissolutions of lignin and carbohydrates in kraft pulping are plotted against each other, the resulting curve can be divided into three distinct phases. These are termed initial, bulk, and residual delignification, respectively; each has a different selectivity with respect to carbohydrate losses. The kinetics and activation energies for the initial and bulk phases have been determined from laboratory kraft cooks and are given in reactions ( 1 0) and ( 1 1) [72]. Obviously, there is no apparent influence of either alkali or sulfide on the delignification in the initial phase. The activation energy here is low. The bulk phase, on the other hand, shows a strong dependency on alkali and has an activation energy typical of chemical reactions. A certain influence of sulfide is also noticed.

dL = k . L - [HO-I0[HS-]” dr

--

E ,= 40 kJ/mol

The apparent slight influence of sulfide on the delignification kinetics has, however, a great impact on the resulting pulp characteristics. Both the transition point from bulk to residual delignification and the viscosity of the pulp at a given kappa number will be greatly changed if the sulfide concentration is so low that a sulfide deficiency occurs toward the end of the initial phase [73,74]. Thus, a low availability of sulfide in this part of the cook will result in a high amount of lignin when the cook reaches the residual (slow) delignification phase. Continued cooking results in a pulp of inferior quality with a low viscosity. In the traditional kraft cook, the whole charge of chemicals present in the “white liquor” is mixed directly with the wood, resulting in a very high concentration of alkali at the beginning of pulping. Simultaneously, the charge of sulfide is such that after a short period of pulping, the sulfide concentration may decrease to very low levels before it again increases [70]. These concentration profiles are highly detrimental for pulping selectivity and, in modern pulping, split charge of white liquor and recirculation of the used pulping liquor (black liquor) have been introduced. These changes result in a more even alkali profile during the cook and in a high initial concentration of sulfide, thus permitting a more selective delignification [75].

2. Lignin Reactions a. @Aryl EtherStructures. The dissolution of lignin in kraft pulping is mainly a consequence of the cleavage of p-aryl ether linkagesin p-0-4 structures, resulting in lignin fragmentation, together with the liberation of free phenolic hydroxyl groups. The latter, together with the originally present phenolic lignin end groups, are ionized in the alkaline solution. Phenolate salts are more soluble in water than un-ionized phenols, and this results in dissolution of lignin fragments.The reactions of p-aryletherstructures have been studied extensively, both in experiments with low-molecular-weight model compounds and through analyses of black liquor and pulp lignin samples [76-811. In phenolic structures of the guaiacyl type, the rate of cleavage has been found to follow the expression given in reaction (12), which, in accordance with the relationship found for wood, shows no influence of alkali or sulfide. In nonphenolic units [reaction (13)1, the corresponding relationship shows a first-order dependency on alkali [82,83].

876

Gellerstedt

The published pseudo-first-order rate constants for these reactions [84] suggest that the kinetic half-life for a phenolic P-aryl ether structure under realistic pulping conditions is around 1.3 min at 170°C, whereas the corresponding value for a nonphenolic structure is 45 min. These values are based on model compound studies and can be regarded as ideal, since complete accessibility between model and pulping liquor exists. The reaction mechanism for cleavage of a phenolic P-0-4 structure is shown in Fig. 17. The first reaction step, the formation of a quinone methide, is reversible, but once it is formed this intermediate may react further with nucleophiles present in the pulping liquor. The most reactive nucleophiles are hydrosulfide and/or polysulfide ions. Under conditions where sulfide is absent or present in very low concentration, carbanions from

Reduction

Condensation

Further reactions

FIGURE 17 Reactions of phenolic p-0-4structuresinkraft pulping.

Pulping Chemistry

077

lignin or carbohydrates may compete successfully [85-881. The latter reactions would result in condensation products, the presence of which is discussed further below. By addition of hydrosulfide or polysulfide to the quinone methide, a thiol (polythio) structure is formed (Fig.17) [76,89-911. In the alkaline solution this may undergo an intramolecular attack to form a cyclic intermediate with simultaneous release of the psubstituent. In further reaction steps, the cyclic intermediate loses elemental sulfur, resulting in a temporary loss of sulfide ions in the pulping liquor. In the alkaline solution, elemental sulfur reacts with sulfide to form polysulfide, which finally disproportionates into hydrosulfide and thiosulfate (Fig. 18). Although this reaction sequence is not fully supported by experimental data under conditions prevailing in a digester [92], it offers an explanation of the formation of thiosulfate and the re-formation of hydrosulfide in kraft pulping. In competition with the addition of a hydrosulfide ion or a carbanion to the intermediate quinone methide, two further reaction routes are possible on this intermediate. The first of these is elimination of either the P-hydrogen or the y-hydroxymethyl group, resulting in the formation of an enol ether structure 1761. Once formed, such a structure is relatively stable under alkaline conditions and may thus survive the kraft cook [79]. As a consequence, the lignin fragmentation will not proceed to an optimum extent. Thesecond reaction type is less established but involves a reduction of quinone methides to the corresponding aromatic a-methylene structure, a reaction which has been indicated by analysis of lignin samples with "C-NMR [93]. Further support for this reaction type has been obtained by alkaline treatment of quinone methide model compounds with alkali in the presence of either glucose, anthrahydroquinone, or both [94]. Again, such reactions would constitute a limitation of the lignin fragmentation, since p-0-4 structures are withdrawn from the desirable cleavage reaction. The cleavage of nonphenolic p-0-4 structures is dependent on the concentration of hydroxy ions and, as shown above, the reaction is comparably slow. The mechanism involves a formation ofan epoxide between the a - and the p-carbon atoms as shown in Fig. 19 1951. Subsequent hydrolysis would give a glycerol structure, together with the released @-substituent. Again, a competing reaction iwolving anions from lignin or carbohydrates may lead to condensation products [96]. The chemistry of p-0-4 structures in kraft pulping has been supported by analyses of the remaining fiber lignin, as well as the corresponding dissolved lignin, as a function of delignification. Analysis by acidolysis or thioacidolysis provides semiquantitative data of the amount of remaining p-0-4 structures in lignin. Figure 20shows the expected decrease in p-0-4 structures as pulping progresses but also that there is an amount remaining at the end of the cook (in softwood lignins). Furthermore, the dissolved lignin contains an appreciable amount of this structure throughout the cook [78,81]. Obviously, the rapid cleavage reaction indicated by the kinetic data above is not fully supported by the analytical data. The reason for this is not known, but factors such as accessibility of hydrosulfide, both in the fiber wall and in the lignin after dissolution, must be assumed

4nSo + 4HS- + 4HO- =

&,S2-

43,s'-

+ 4H20

+ 4(n-l)HO = nS2O3'- + 2(n+2)HS + (n4)H20 4S0 + 4HO- = S203'- + 2HS'

+ H20

FIGURE 18 Inorganic reactions of elemental sulfur in kraft pulping.

878

Gellerstedt CH2OH

I

HC-OR

I

HC-OH

CHIOH H

I

L

CH2OH -

O

-

I

e

HC-OH

+

"Q/ \ CH30

OCH,

/o

o /

CHZOH

I

HC-OH

I

HC-OH

I

FIGURE 19 The cleavage reaction of a nonphenolic p-0-4 structure in alkaline pulping.

879

Pulping Chemistry P-aryl ether structures l m o l l g of lignin

500

-

300 -

100

I

1

0

1

I

I

1

20 40 60 80 100 Degree of delignificotion. D/'' on wood

FIGURE 20 The presence of noncondensed phenolic p-0-4structures in pulp and dissolved lignin as a function of delignification (analysis by acidolysis).

to be important. The consumption of hydrosulfide and formation of sulfur in the fiber wall may, for example, result in a certain deficit of sulfide at some of the reactive sites during pulping. Even if the charge of sulfide is high and added to the wood through an efficient impregnation stage, this type of deficit seems to be present during pulping. The presence of p - 0 - 4 structures in the dissolved lignin can, for similar reasons, be assumed to be due to the formation of molecules or molecular aggregates having negatively charged shells with an interior that is inaccessible to the hydrosulfide ions. b. Other Lignin Substructure.s. Phenolicphenylcoumaran (p-S) and 1,2-diarylpropane-l ,3-diol (p-1) structures in lignin are reactive under alkaline conditions, since they contain an a-hydroxy or a-ether function that can be lost in the reversible formation of a quinone methide. In analogy with the reactions of p-aryl ether structures discussed above, addition of a nucleophile or elimination of a hydrogen or formaldehyde may take place. Thepresence of acarbon-carbonlinkage between the phenylpropane units, however, prevents any fragmentation of the lignin macromolecule around this linkage. Model studies indicate that phenolic p-5 or p-l structures preferentially undergo elimination reactions, resulting in the formation of stilbene structures (cf. Fig. 2) [97]. The nonphenolic counterparts, on the other hand, are essentially stable under alkaline pulping conditions. Other major lignin linkages, such as those in biphenyl (S-S) and biphenyl ether (40 - S ) structures, are completely stable in alkaline pulping, whether they are phenolic or not. In phenolic units, the side chains in such structures can, however, react via formation of a quinone methide according to the reactions outlined above. Recently, the behavior of a phenolic dibenzodioxocin (S-S-0-4) structure has been studied by the use of a model compound 1981. Under kraft cookingconditions, the aryl etherlinkageiscleaved to a large extent, leaving the S-S structure as a major reaction product. c. Aronzrrtic Metl1o.ry Groups. In softwood species, the lignin contains one aromatic methoxyl group per phenylpropane unit, with the exception of compression wood lignin, where some pcrrrr-hydroxy-phenylpropane units are present. In hardwoods, a mixture of

880

Gellerstedt

guaiacyl and syringyl structures are found, resulting in a methoxyl value per C-9 unit of between 1 and 2, depending on species. In kraft pulping, these methoxyl groups are attacked by the nucleophilic hydrosulfide ions, resulting in a partial demethylation and formation of methyl mercaptan. The formation of a new phenolic hydroxyl group in lignin adds to the hydrophilicity, whereas the other product, methyl mercaptan, is able to react further as a nucleophile to form dimethyl sulfide by further attack on aromatic methoxyl groups (Fig. 21) [99]. The two low-molecular-weight sulfur compounds survive the kraft cook and constitute major components in the mixture of malodorous gases leaving the digester. From both softwoods and hardwoods, the amount of these products corresponds to a degree of demethylation in the lignin of around 1-2%. In modern kraft mills, methyl mercaptan, dimethyl sulfide, and small amounts of dimethyl disulfide, formed through air (oxygen) oxidation of methyl mercaptan, and some hydrogen sulfide, are usually collected and burned to eliminate almost all of the odor. d. Condensation Reuctions. The presence of condensation reactions in lignin has been used frequently to explain the slow rate of delignification toward the end of kraft pulping, as well as the high demand of chemicals in a subsequent bleaching operation. This assumption is further supported by a variety of lignin model studies demonstrating that condensation reactions may easily occur, e.g., between quinone methides and different types of carbanions or between carbanions and liberated formaldehyde (Fig. 22) [85].The possibility of establishingequilibria between quinone methides and (ionized) hydroxyl groups, as well as reactions between quinone methides and enol structures, have also been suggested as modes of formation of new linkages between lignin and carbohydrates [87,88]. However, conclusive evidence for the formation of condensed structures in the wood residue during pulping, as well as their presence in the unbleached pulp, is not available. Although the existence of comprehensive condensation reactions giving rise to diarylmethane structures during pulping has been suggested [ 100,101], the analyses used in that work cannot be used unequivocally for making such an interpretation [ 1021. The presence of lignin-carbohydrate linkages in unbleached pulps. however, is strongly supported by indirect means; but, as discussed further below, it is not known whether these are native or not. The presence of condensed structures and products in dissolved lignin have been identified, although the extent of such reactions seems to be small [1031. Recently, it has also been suggested that cellulose and glucomannan. when subjected to a kraft cook in the presence of coniferyl alcohol, will result in a “grafting” reaction [ 1041. In summary,

(R = H or CH,) CHS-S-S-CH,

FIGURE 21 Formation of the malodorouscompoundsmethylmercaptan,dimethylsulfide.and dimethyldisulfide in kraft pulping.

Pulping Chemistry

W

(v (v

881

882

Gellerstedt

however, the presence of condensation reactions in draft pulping is still a matter of uncertainty which requires further attention.

3. Carbohydrate Reactions In alkaline pulping, the initial consumption of alkali is caused mainly by the rapid hydrolysis of acetyl groups from hemicelluloses: galactoglucomannans in softwood and glucuronoxylan in hardwoods. Essentially all acetyl groupsare eliminated as acetic acid, requiring an equivalent molar amount of sodium hydroxide. Thus, in softwood pulping, this reaction consumes approximately 0.35 mol of alkali per kilogram of wood, whereas the corresponding figure for hardwoods is in the range of 0.9 mol. The yield loss in kraft pulping due to degradation and dissolution of polysaccharides is substantial and constitutes a serious drawback of the process. Typical values for the amount of wood (pulp) components after kraft pulping of pine and birch are shown in Fig. 23. Thus, in addition to the desirable dissolution of lignin, all types of polysaccharides are degraded to some extent. In softwoods, substantial amounts of the xylan and around 75% of the dominating hemicellulose, glucomannan, are lost. In hardwoods, the xylan loss is almost 50%. The predominant reaction responsible for the degradation and dissolution of polysaccharides is an alkali-induced stepwise elimination of monomeric sugar units, starting from the reducing end; this is the so-called peeling reaction. Even at low temperatures, around 100°C, the rate of reaction is substantial and a decrease in the degree of polymerization (DP) of the polysaccharide is encountered [ 1051. For hemicelluloses having a short chain length, such as in glucomannans, the change in DP facilitates nearly acomplete dissolution already in the early part of the kraft cook. Xylans, being somewhat larger polymers and having a less efficient peeling reaction, survive the cook to a larger extent, but the remaining xylan chains can be assumed to have suffered a substantial loss i n DP. The mechanism for the peeling reaction is shown in Fig. 24. The major polysaccharides in wood are linked through 1 + 4 glucosidic linkages; the overall reaction results in a loss of the terminal sugar moiety as an isosaccarinic acid and the formation of a new reducing end group capable of undergoing the same reaction. Intermediate reaction steps involve a reversible opening of the hemiacetal ring, reversible isomerization to a fructose unit, a p-elimination of the 4-substituent, a keto-enol equilibrium, and finally a benzilic acid rearrangement. As a side reaction, the p-elimination reaction may occur already in the first-formed intermediate, the open aldehyde structure, with formation of a mdcr-~ac-

Pine Cellulose Glycomannan Xylan Other carbohydrates and various components Sum of carbohydrates Lignin Pitch Sum of components (yield) a

35

-

Birch

34

4 5

1 16

44 3 0.5 47

51

2

0.5 53

Figures in parentheses refer to original wood composition,

FIGURE 23

Yield loss in kraft pulping for pine and birch. (All figures based on wood.)

883

Pulping Chemistry

(OH

cell-0 H QH 0

H

OH

&

H0

-

(OH

/y-"oz,

cell-oH 0

OH

Metasaccharinic acid ("stopping" reaction)

-

p-elirn

cell-0

CHZOH 0

(Fructose unit)

I

OH

H

HoYCooH CHIOH

lsosaccharinic acid

FIGURE 24

The mechanism for thepeelingreaction

in alkaline pulping.

carinic acid. For cellulose, the latter reaction, termed the stopping reaction, occurs when an average of 65-70 sugar units have been peeled off the chain. In softwood xylans, the presence of arabinose substituents in the 3-position favors a direct @-elimination from the corresponding open aldehyde form of the end group. This results in a somewhat higher stability of this xylan as compared to a hardwood xylan, since the stopping reaction is facilitated. Both softwood and hardwood xylans also contain 4-0-methyl-glucuronic acid as substituent in the 2-positions, thereby making the isomerization to a terminal keto sugar more difficult and decreasing the extent of the peeling reaction. In alkaline pulping, the 4-0-methyl-glucuronic acid substituents in xylans are to a large extent converted to the corresponding unsaturated structure by loss of methanol (Fig. 25) [ 1061. The formed "hexenuronic acid" groups, which are still attached to the xylan backbone, are rclativcly stable under alkaline conditions and contributc a substantial proportion of the acidic groups in kraft pulps (Fig. 26) [ 1071.

Gellerstedt

884

HO'

cH30*1

~

- CHaOH

H0

H0

H0

"Xyl-xyl-xyl-

"Xyl-xyl-xyl-

FIGURE 25 Alkali-inducedelimination ofmethanolfrom4-OMe-glucuronicacid wood and hardwood xylans.

H0

units in soft-

The extent of peeling reactions is high in the beginning of a kraft cook, resulting in a rapid loss of hemicelluloses from the wood. In later parts of the cook, a further loss of carbohydrates can be initiated through alkaline hydrolysis of polysaccharide chains. This reaction occurs randomly along the chain and results in the formation of shorter chains or alkali-soluble chain fragments(Fig. 27). Since each cleavage reaction alsoleads to the formation of a new reducing end group, this in turn may rapidly undergo further degradation by the peeling reaction ("secondary peeling").

4. Additives in Alkaline Pulping The large yield loss in kraft pulping due to the degradation of polysaccharidescanbe reduced if the carbohydrate end groups are either reduced or oxidized, thus reducing the extent of the peeling reaction [ 108- 1 IO]. Both reaction modes have been tried successfully on a laboratory scale with polysulfide oranthraquinone (AQ) asoxidizingagents and sodiumborohydrideor hydrogen sulfide as reducingagents. Today, however, only the oxidants are used commercially.

20

I

18

-

16 W

4 3;

1412-

J

0 10-

S

0

o

i0 E

a6'

"

175 l70

HexA

-8-

165

"

MeGlcA

160

"

Ara

155

-El-

Y

g 3 5a W

150 t l

remperature

4

145

5F

4 '

140

2' n

"0

50

150

100

200

- 135 250

TIME, min FIGURE 26 The presence of hexenuronic acid groups i n pulp xylan as a function of cooking time in a kraft cook [1071.

Pulping Chemistry

+

'0 Xt

9+

+

0

X

C .*

885

Gellerstedt

886

The reaction mechanism in the oxidation of a carbohydrate end group can be envisaged as involving the enediol intermediate which is oxidized to the corresponding diketo structure (Fig. 28). A benzilic acid rearrangement then results in the formation of aldonic acid structures [cf. 11 I]. The reductive reactions give rise to the corresponding sugar alcohol or thioalcohol end groups with borohydride or hydrogen sulfide, respectively. The oxidation of carbohydrates with either polysulfide or AQ results in the formation of hydrosulfide ions or anthrahydroquinone (AHQ). In the alkaline pulping liquor, the latter is ionized and may add to quinone methides in lignin analogously to the hydrosulfide ion. The adduct formed, when present in a P-0-4 structure, has been shown to decompose with cleavage of the P-aryl ether linkage and re-formation of AQ, thus resulting in an efficient AQ redox cycle (Fig. 29) [ 112,113]. An alternative mechanism, which has been supported by several experimental studies, indicates that the reaction of AHQ with lignin involves single-electron-transfer reactions, leading to efficient P-aryl ether cleavage [ 1 141. With either mechanism, the cleavage reactions result in the formation of a cinnamyl alcoholstructure. In analogy to kraft pulping, a mixture of sodium hydroxide and AQ therefore acts as a powerful pulping system able to give “kraftlike” chemical pulps. In practice, however, AQ is used mostly as an additive in kraft pulping and, like polysulfide addition, this will result in a yield increase. The oxidizing power of polysulfide is not restricted to the carbohydrates; certain lignin structures can be oxidized as well. Some examples are shown in Fig. 30 [89,115]. Since oxidized lignin structures generally are more easily degraded in alkaline media, it can be assumed that the presence of polysulfide ions in kraft pulping will facilitate the delignification. Further support for such reactions was recently obtained by impregnating wood with polysulfide prior to kraft pulping and observing a better selectivity in the cook [ 1161.

5. The Chemical Structure of Unbleached Pulp Lignin. The comprehensive degradation and dissolution of lignin in kraft pulping is the result of cleavage of P-0-4 structures. However, the cleavage reaction is not complete, and several P-aryl ether linkages survive the kraft cook. Analytical data, although a.

GOH

CHIOH

= HOQ

-0

OH

C C H O -0

7

-0

OH

CHIOH

I

OCH3

0

0

0

OH

FIGURE 29

Mechanism for cleavage of a phenolic p-0-4structure in alkaline anthraquinone pulping. 03 03 -l

888

Gellerstedt

HC-0

PS

0. CH,

I

c=o

+

&H3

OCH,

OCH,

OH

OH

CH0

CHZOH

I

I

CH

CH

II

CH3

I

It

PS OCH3 OH

FIGURE 30 Alkaline polysulfide (PS) oxidation of an enol ether and a coniferyl alcohol structure.

at best semiquantitative, indicate a remaining amount, around 15% in spruce kraft pulp lignin [78]; this figure probably constitutes a lower limit. In birch pulp, on the other hand, the figure seems to be much lower [SO]. In addition to intact p-0-4 structures in the remaining kraft pulp lignin, some enol ether structures can also be found [79]. Obviously, the availability of hydrogen sulfide ions at the reacting sites in the fiber wall is not sufficient to suppress the elimination reactions completely throughout the cook, as discussed above. The presence of phenolic hydroxyl groups in lignin determines the solubility in alkaline solution. To some extent, such groups are present in the native lignin, but the liberation ofnew such groupsduring pulping is essential for the lignin dissolution to proceed efficiently. In dissolved lignins from a spruce kraft cook, it has been demonstrated that the content of phenols should be in the range of 3-4 mmol/g of lignin for attaining solubility in the pulping liquor; higher values are needed for the higher-molecular-weight molecules. In addition,a kraft lignin contains around 0.7 mmol/g of carboxyl groups 1I17,l IS]. In the residual fiber lignin, the values are much lower and the content of phenolic hydroxyl groups changes from approximately 0.7 in wood to 1.5 mmol/g of lignin in the unbleached pulp with a kappa number around 30 [ 1 17,1191. The desirable fragmentation of lignin in kraft pulping, brought about by the cleavage of p-0-4 structures, can be restricted due to a low availability of hydrosulfide and by nondesirable side reactions. The reduction of quinone methides mentioned above is one such reaction, but condensation reactions between lignin and polysaccharides would also

Pulping Chemistry

889

act in a restrictive way, since the resulting product would be of much higher molecular weight and have the possibility of being firmly attached to the polysaccharides. The formation of such linkages during pulping has been suggested [ 1201, but lignin-carbohydrate linkages may also be native. It has been found that the residual delignification phase is accompanied by a slow dissolution of lignin, which successively contains higher amounts of polysaccharides, most notably xylan [ 1181. Extraction of unbleached kraft pulp with strong alkali at an elevated temperature will also result in the dissolution of a material which contains lignin linked to xylan as the predominant polysaccharide [ 12 1 1. It is reasonable to believe that the lignin reactions in kraft pulping proceed toa different extent in different parts of the fiber wall and in different types of fibers. Therefore, all analytical data on residual lignin must be regarded as averages, and the true values for phenolic hydroxyl groups, etc., must, in fact, include a range of values. The lignin macromolecules also include a range of structural features from rather intact “native” molecules to those which have been severely degraded and chemically changed as a result of the pulping conditions. In line with this hypothesis, it has been shown that a substantial portion of the residual lignin can be leached out of the fibers in a process which is governed by a high temperature and the presence of alkali [ 122,1231. b. C~rr1~okyclmte.s.The peeling reactions occurring in the polysaccharides during pulping result in the formation of acidic end groups. In addition, some side groups in the hemicelluloses are either eliminated, such as galactose from galactoglucomannan, or chemically modified, like 4-0-methyl-glucuronic acid from xylan. The latter, together with the formed unsaturated hexenuronic acid groups, add a substantial portion of the acid groups i n unbleached kraft pulps, the total of which can be around 60-80 mmolkg of (softwood) pulp [ 1241. The presence of hexenuronic acid groups also add to the kappa number of the pulp (Fig. 3 1 ) [ l 25,1261. In recent work, it has been found that this structure alone corresponds to around 2 kappa units inan unbleached softwood kraft pulp (kappa no. 18) whereas the corresponding figure for a hardwood pulp (kappa no. 1.5) is around 5 units 11271.

Pulp

Kappa Hexenuronic number

acid, pMollg pulp

Kappa number equivalence

Birch, unbleached 11.3 14.0 14.5 15.9 18.8

64 85

3.6 5.5 4.9 5.6 5.8

Birch, bleached 6.0 4.5

53

39

4.6 3.4

22

1.2 1.9

Pine, unbleached 18.2 18.6 18.0

FIGURE 31 kraft pulps.

Contribution to kappa nulnbcr of hcxcnuronic acid in solnc unbleached and bleached

890

Gellerstedt

W. PULP BLEACHING A.

TechnicalOverview

The poor selectivity and the slow delignification encountered in the final delignification phase in kraft (as well as in sulfite) pulping cannot be tolerated from a pulp quality point of view. Consequently, the cook is interrupted and, when required, the pulp is further delignified and bleached using oxidation as the principal reaction mode. The residual lignin in chemical pulps, notably kraft pulps, is difficult to remove, and several oxidative steps are required with intermittent extraction of the lignin which was oxidized in the previous acidic stage. Contrary to mechanical pulp bleaching, the complete bleaching of chemical pulps to full brightness must be accompanied by the almost total removal of all residual lignin from the fibers. Obviously, the chromophoric structures introduced in the pulping process are such that no bleaching agent can be used for the selective oxidation of the colored groups without a simultaneous dissolution of the lignin. The most efficient bleaching agent for chemical pulps is elemental chlorine, which has been used for several decades (together with hypochlorite) as the predominant oxidant in chemical pulp bleaching. With the development of chlorine dioxide bleaching, the possibility of making fully bleached kraft pulp in sequences such as C E D E D* was realized and introduced in the 1940s. The growing awareness that elemental chlorine might produce chlorinated products with a negative impact on the environment and the large leakage of organic matter occurring from bleaching plants paved the way for further changes in the bleaching technology. Today, these changes are aimed at strongly decreased levels of chlorinated organic products. As a consequence, new possibilities arise for reducing the use of fresh water and closing up the pulp mills. The discovery that magnesium compounds excerted a positive influence on the selectivity in oxygen (0)bleaching opened up the possibility of starting the bleaching sequence with an 0 stage. Here, around 50-6570 of the residual lignin in a softwood kraft pulp can be eliminated without detrimentally affecting the pulp strength characteristics. The dissolved material can be introduced into the chemical recovery system, thus considerably reducing the content of organic matter in the bleaching effluent. In modern bleaching technology, further changes have been introduced together with new bleaching agents. The use of elemental chlorinehas been reduced or completely eliminated, while the use of chlorinedioxide has increased. In elemental chlorine-free (ECF) bleaching, only chlorinedioxide is used. Alternatives, not involving the useof chlorine-containing bleaching agents, have been developed (totally chlorine-free, TCF, bleaching) and include oxygen, hydrogen peroxide (P), and ozone ( Z ) as major oxidants. This development has resulted in large improvements in the effluent quality and in increasing possibilities for reducing the discharge of both water and organic material from pulp mills. Bleaching reactions with oxygen, hydrogen peroxide, or ozone involves, in addition to ionic reactions, hydroxyl radical reactions. Oxygen is itself a radical, and therefore the initial reaction with lignin is a one-electron reaction giving rise to the formation of SUperoxide ions, hydrogen peroxide, and organic peroxides, which in turn can give hydroxyl radicals in secondary reactions. Hydroxyl radicals can also be formed in hydrogen peroxide and in ozone bleaching stages. The formation of the highly reactive and unselective hy-

*:C = chlorine, E = alkaline extraction, D = chlorine dioxide.

Pulping Chemistry

891

droxyl radical in these bleaching systems will result in a certain oxidation and degradation of the cellulose unless special precautions are taken. Even so, certain cellulose degradation cannot be completely avoided.

B. Oxygen 1. Lignin Reactions In oxygen &lignification, carried out in an alkaline medium, a partial oxidation of phenolic structures in lignin takes place. In structures with conjugated double bonds, such as in enol ethers and stilbenes, the oxidation may also take place in the side chain, resulting in a fragmentation of the lignin. The reaction starts with the formation of a phenoxy radical by electron transfer to the oxygen. In subsequent reaction steps, the phenoxy radical is converted into a hydroperoxide, which in turn reacts further, with formation of oxidized products. The reaction sequence is outlined in Fig. 32 [ 128,1291. The kinetics for the oxidation of a series of ligninlike phenols has been studied and found to be rather slow [ 1301. Thus, the kinetic half-life for phenolic structures having a guaiacyl structure is of the order of 14 min under technically relevant conditions. For structures like stilbenes and diphenols, such as catechols, much higher reaction rates are encountered. The importance of these structures for the lignin degradation is, however, limited due to their low abundance. Since the solubility of oxygen in water is low, it must be assumed that the extent of oxygen oxidation of lignin is highly dependent on the efficiency of mixing the pulp and the oxygen gas. Once the concentration of oxygen in the aqueous phase has ceased, only a very limited further oxidation of phenols can take place and secondary reactions, such as alkaline hydrolysis and lignin extraction from the fibers, become important. These reactions are accompanied by a certain brightening action caused by the liberated hydrogen peroxide in the system. The latter, together with organic peroxides, may also give rise to hydroxyl radicals by metal ion-induced decomposition and, thus, degradation of the cellulose. The consumption of phenolic hydroxyl groups in oxygen delignification has been found to be rather limited [ 131,l 321. In accordance with the kinetics discussed above, the residual lignin after an oxygen stage contains a considerable amount of remaining phenolic groups, whereas the solubilized lignin has a higher amount of phenolic groups as compared to the average for the residual lignin in the unbleached pulp. Thus, it seems likely that the oxygen oxidation occurs predominantly with those parts of the residual lignin which are the most phenolic and the most accessible, i.e., a lignin which already is rather soluble but trapped in the fiber wall for reasons of molecular size or a few remaining linkages to other pulp constituents. In addition to phenolic groups,thedissolved lignin from the oxygen stage contains an increased number of carboxyl groups. The total number of acidic groups is of the order of 5 mmol/g of lignin. After an oxygen stage, the remaining lignin contains a higher amount of carboxyl and a lower amount of phenolic groups as compared to the unbleached fiber lignin. Other differences are small, but the residual lignin has a somewhat higher amount of condensed structures [ 1321 and the content of p-0-4 structures has decreased in comparison with the residual lignin after the kraft cook. A second oxygen stage results in further &lignification of the pulp, but again the structural differences in the residual lignin are minor [ 1331. 2. Carbohydrate Reactions The carbohydrate reactions in oxygen delignification result in a loss of viscosity due to attack by hydroxyl radicals along the cellulose chains. The reaction, which is shown in

892

9

\ \

0

c

Gellerstedt

Pulping Chemistry

a93

Fig. 33, involves the formation of a carbonyl group in the sugar moiety, followed by a pelimination reaction that results in a chain scission [134]. Unless magnesium ions are present in the bleaching liquor, the viscosity loss isof such a magnitude that technical utilization of oxygen delignification is prohibited 11351. Addition of magnesium ions, however, will stabilize the formed hydrogen peroxide to some extent [136], thereby reducing the formation of hydroxyl radicals. It has also been shown that the presence of magnesium ions gives a certain reduction in the rate of reaction between hydroxyl radicals and carbohydrate structures [ 1371. The yield loss in oxygen delignification is limited, since the carbohydrate peeling reactions (see above) are suppressed in favor of an oxidative stabilization reaction. Thus, the reducing end groups in the polysaccharide chains are converted to aldonic acid groups through oxygen oxidation of the open enediol structure as shown in Fig. 28 [ 1381. After a kraft cook, the pulp contains a xylan in which a major fraction of the 4-OMe-glucuronic acid groups has been converted into the corresponding unsaturated structure, "hexenuronic acid," by loss of methanol, as described in Fig. 25. A subsequent oxygen delignification step is not able to degrade this structure, and consequently the relative contribution to the resulting pulp kappa number from the hexenuronic acid groups will increase after the 0 stage.

CHpOH H0

o

H0

'

o

m0 H

o

'0

aOQow HH 00

CH20H

0

- Hop H0

CHpOH

t "

0

0

H0

1

'0

- @-OH

CHZOH

-

H CHpOH

rearrangement

H0

0

HeLow coo0

FIGURE 33 Mechanism for the oxidative degradation of cellulose by hydroxyl radicals.

894

C.

Gellerstedt

HydrogenPeroxide

Alkaline hydrogen peroxide in the presence of sodium silicate is the traditional bleaching system for mechanical pulps, as discussed earlier. An efficient bleaching action requires that all transition metal ions be removed as completely as possible prior to bleaching, an operation usually done by pretreatment of the pulp with a strong chelant such as EDTA or DTPA. The same criteria must be fulfilled in the peroxide bleaching of chemical pulps. Unlike mechanical pulps, however, chemical pulps cannot be bleached to high brightness values with hydrogen peroxide unless the residual lignin is removed, since the chromophoric systems are much less reactive [ 1391. Efficient delignification and bleaching of chemical (kraft) pulps can be obtained if the bleaching stage is preceded by a chelating stage with EDTA, carried out under controlled conditions [140,141] in order to eliminate manganese, iron, etc., as completely as possible while keeping the concentration of ions such as magnesium and calcium high. The bleaching stage itself can be done with alkaline hydrogen peroxide (without silicate) at temperatures around or slightly above 100°C for 2-4 h.In the latter case, an oxygen pressure can be applied (POstage). Addition of magnesium ions to the bleaching liquor is beneficial. The ability of alkaline hydrogen peroxide to attack and degrade chromophoric structures, such as conjugated carbonyl structures, is well known. In the bleaching of chemical pulp, however, this mode of reaction must be accompanied by reactions resulting in a more comprehensive lignin oxidation and dissolution. One possibility for achieving such a reaction would be a radical-induced oxidation of phenolic structures in the residual lignin. Thus, the (controlled) decomposition of hydrogen peroxide leading to hydroxyl radicals and superoxide ions should result in the degradation of aromatic rings through reactions similar to those occurring in oxygen delignification. Analyses of the structure of the dissolved and residual lignin from a peroxide bleaching stage do not, however, unequivocally support such a mechanism since, i n contrast to the corresponding lignins from an oxygen stage, the content of aromatic rings in these lignins does not change to any large extent [ 1321. On the other hand, the content of carboxyl groups i n lignins from a peroxide stage is high, demonstrating the capability of hydrogen peroxide as an oxidant for lignin. Based on I3C-NMR data, the majority of the formed carboxyl groups seem to be aliphatic rather than aromatic, thus indicating that side-chain oxidation in lignin may be a predominant reaction mode in peroxide bleaching [142]. Recently, this view was further supported by a model compound study in which a phenolic lignin structure was shown to be easily degraded by the action of alkaline hydrogen peroxide at an elevated temperature (Fig. 34) [ 1431. Analysis by "C-NMR of isolated residual lignin after a peroxide stage (in the sequence OQP) demonstrates that this lignin, although coming from a pulp with less than 1 % lignin, still has most of the features of a lignin structure; the major differences are a low amount of p-0-4 structures, the presence of carboxyl groups, and the presence of aliphatic methylene and methine groups [142]. Therefore, the successive delignification that takes place in the kraft cook and in the subsequent oxygen delignification and hydrogen peroxide bleaching stage does not seem to alter (degrade) the remaining lignin structure in a profound way. The reactions of carbohydrates in a peroxide bleaching stage seem to be very similar to an oxygen stage. Thus, any decomposition of hydrogen peroxide, resulting in the formation of hydroxyl radicals, is probably inducing an oxidation of hydroxyl groups to carbonyl groups along the polysaccharide chains followed by a chain scission. In analogy with oxygen delignification, the hexenuronic acid groups present in the xylan will survive an alkaline peroxide stage, thus resulting in a comprehensive contribution to the kappa nlltnher :d'ter the stape

(Fig. 31).

Pulping Chemistry

895

I

I

HC-0-OH

-

OCH,

/

0

0

OH

CHZOH

I

HC-O*

I

HC-O* coo@

+

I

HC

-

OCH3

OCH,

+

ox.

f"

-

I

HO'

OH

I

OCH, 0

FIGURE 34

OCH,

OH

Oxidation of a phenolic p-0-4 structure with alkaline hydrogen peroxide.

D. Chlorine and Chlorine Dioxide In the bleaching of pulp with chlorine, both chlorination and oxidation reactions take place in the lignin. The dissolution of lignin is limited, however, unless a subsequent alkaline extraction is carried out in which the formed carboxyl groups are ionized. Other acidic bleaching stages behave similarly. Experiments with pulp have demonstrated that despite the great efficiency of chlorine bleaching (in combination with alkaline extraction), one DE sequence is not sufficient to remove the residual lignin completely; in fact, several DE sequences applied on the same pulp do not result in a complete lignin removal, as measured by the kappa number (Fig. 35) [ 1441. The reason for this behavior is thought to depend on the formation of specific groups in lignin which prevent any further penetration of aqueous chlorine into the lignin matrix. Alkaline extraction removes that portion of the lignin containing these groups, and new unoxidized lignin structures become exposed for an oxidationlchlorination sequence. Indirect support for this view is found in chlorine analysis of kraft fibers before and after an E stage in a CE sequence. That part of the lignin which had a high content of chlorine (and presumably of oxidized groups) could be removed by alkali, leaving a residual lignin with a uniform and low residual amount of chlorine [145]. Estimation of the chlorine mass balance from the bleaching sequence (Css + D,J E D E D (Fig. 36) [l461 reveals that the majority of added chlorine ends up in the effluent

Gellerstedt

30

6

-i

0 CE 0 CRECE A C~ECRECE cR: 2 min,

l§

5.74% cl2 on 0.d. pulp C : t mln, 5.74% Cl2 on 0.d. pulp

10

51k:n*

,

om*

,

0

0

10

20

30

40

o%

, 50

Time (t), min

FIGURE 35 Successive dissolution of lignin in a kraft pulp by repeated treatment with followed by an alkaline extraction stage [144].

chlorine

as chloride ions, most of which are formed in the E stage. In addition, chlorate is formed in the D stages.The remaining chlorine is found as organically bound chlorine in the effluent and, to a small extent, in the bleached pulp. The reactions between lignin and aqueous chlorine have been thoroughly studied, and four reaction principles have been identified I 1341. Chlorination of aromatic rings may

Cl (<0,1 %) to air

Cl added 100%

Bleaching

Cl in pulp (c 0,5 Yo)

Pulping Chemistry

897

proceed either by substitution or by electrophilic displacement, the latter reaction resulting in ;L lignin fragmentation. Direct addition of chlorine to aliphatic double bonds can also take place both to lignin and to traces of remaining extractives. The desirable oxidation reaction takes place in the aromatic rings, whether chlorinated or not, resulting in a large number of carboxyl groups via the elimination of methoxyl groups and the formation of new phenolic hydroxyl groups. The "C-NMR data of dissolved lignin from a CE sequence reveals that the aromatic structure is almost completely lost and that the remaining amount of methoxyl groups must be very low [ 1471. The degree of chlorination in the polymeric lignin is high when elemental chlorine is used in bleaching. The ratio of C/Cl has been found to be of the order of 13- 15: I , i.e., an average of almost one chlorine per 1-2 monomeric lignin units is formed [ 1471. In addition, a large number of low-molecular-weight chlorinated compounds, both aromatic and aliphatic, are formed and found in the bleaching effluent. Contrary to elemental chlorine, chlorine dioxide itself is not able to chlorinate lignin; nevertheless, some chlorination does occur in chlorine dioxide prebleaching. The initial oxidation of lignin results in the formation of chlorite and hypochlorous acid. with the latter acting both as an oxidant and, in equilibrium with chlorine, as a chlorination agent. Furthermore, reaction between chlorite and hypochlorous acid results in the formation of chlorate as an undesirable by-product 11481. The gradual formation of chlorine in a D stage, as opposed to a C stage in which all the chlorine is added in one charge, however, results in a large difference in the amount of organically bound chlorine [149]. The latter can be further reduced by adjusting the bleaching pH in such a way that the formation of elemental chlorine is suppressed, i.e., by allowing a final pH = 3-4 in the bleaching liquor [ 1501. In technical D bleaching the remaining fiber lignin after the first chlorine dioxide stage has a C/CI ratio of around 100- 140:I , but in later D stages, a major portion of the introduced chlorine is again eliminated, resulting in very low residual amounts of organically bound chlorine in the fully bleached pulp ( 1 00-500 ppm of chlorine, depending on pulp type and wood species) [ 15 1,1521. By analogy with the kinetics of chlorine bleaching, chlorine dioxide reacts very fast with lignin, as shown in Fig. 37 [ 1531. Again, the rapid phase is followed by a much slower secondary lignin removal. Methylation ofall phenolic hydroxyl groups in lignin prior to the bleaching retards the degree of delignification to a considerable extent [154]. Therefore, the presence of free phenolic groups in lignin is essential for the reactions with chlorine dioxide to proceed efficiently. It has been well established that the first step in the chlorine dioxide reaction is an electron transfer resulting in the formation of chlorite and a phenoxy radical from phenolic or a radical cation from nonphenolic structures (Fig. 38) [155,156]. Both types of aromatic structures react at about equal rates with chlorine dioxide, but in the ionized form the rate of the phenol is approximately 108 times faster than that of the phenol ether. This results in a difference in reactivity between the phenolic and the nonphenolic lignin structure which depends on the pH, with a rate difference of about 100 at pH = 4. The next step in the oxidation chain involves an addition of a second molecule of chlorine dioxide to the radical intermediate, with formation of a chlorite ester. This, in turn, decomposes along several different routes depending on structure and, as shown in Fig. 38, both quinone and muconic acid structures are formed, together with hypochlorous acid and chlorite. The structural changes introduced in lignin during the oxidation with chlorine dioxide include acomprehensive elimination of aromatic rings and the formation of carboxyl groups. Chlorine substitution takes place to a small extent. After the oxidation and sub-

898

Gellerstedt

IC1021 ICl,I g act. CllL

0.50

0.25-\

CIO,

" "

,'C-

0

I

I

20

10

0

phase 1

phase 3

I

30 Time, min

sequent extraction, the remaining fiber lignin still has the general feature of a lignin polymer [ 1571, indicating that oxidative reactions in the fiber wall cease before a complete lignin oxidation has been achieved. After the extraction stage, the remaining fiber lignin can be subjected to further oxidation, again involving more or less the same structures as before.

E. Ozone In recent years, ozone has gained considerable interest as a bleaching agent for chemical pulps, and during the 1990s, several mill installations have occurred. The one important objective being met was the reduction or elimination of chlorine-containing bleaching effluents. Bleaching with ozone is always carried out in acidic media in order to minimize the self-decomposition of ozone in water according to reaction (14) [ 1581. 20,

+ HO-

= 0;-

+ HO' + 20,

(14)

Despite this, a certain formation of radicals seems to take place during technical bleaching conditions, i.e., 40-50°C and a reaction time of a few seconds. Most of these radicals are probably formed in a direct reaction between lignin and ozone rather than through ozone decomposition, which is a comparatively slow reaction. The ionic reactions between ozone and structures containing aromatic rings or olefinic double bonds are very fast. Thus, the kinetic half-life for phenol is of the order of 5 X S at room temperature [ 1591. This so-called ozonolysis reaction results in an oxidative cleavage of the double bond(s) and formation of carbonyl and carboxyl groups (Fig. 39) [ 1601. Analytical data on residual lignin isolated from kraft pulp after a Z stage indicates that this oxidation is comprehensive, giving rise to large amounts of carboxyl groups [161]. In the ozonation of pulp, carbonyl groups, however, are also introduced in the polysaccharides [ 162,1631. This undesirable oxidation can be induced by hydroxyl radicals formedeither by ozone decomposition [reaction (14)]or by a direct reaction between phenolic lignin structures and ozone, as discussed below. A direct ozoneoxidation

FIGURE 38 Reactions between chlorine dioxide and a phenolic lignin structure.

Q)

W W

900

Gellerstedt

0

1

c ”

0-0

0? J O C H 3

0

I

0’

FIGURE 39 Mechanism for the ozonolysis reaction of a double bond and formation of carboxyl groups in the presence of water.

of secondary alcohols in the polysaccharides to form carbonyl groups is also possible, since this type of reaction has a kinetic half-life of approximately 1-2 S [ 1641. The introduction of carbonyl groups along the carbohydrate chains is detrimental, since any subsequent alkaline treatment of the pulp will result in chain cleavage through p-elimination reactions and, thus, a viscosity loss, as discussed before. The direct formation of hydroxyl radicals from reactions between ozone and lignin structures competes with the ozonolysis reaction described above. The former reaction has been suggested to proceed through an electron transfer from the aromatic ring to ozone, followed by a rapid decomposition of the formed hydrotrioxide radical into oxygen and a hydroxyl radical (Fig. 40) [ 1651. For aromatic nonphenolic structures with a low redox potential, the initial radical yield is of the order of 5%. The presence of a phenolic hydroxyl

FIGURE 40

Formation of a hydroxyl radical

from the reaction between

ozone and a phenol.

901

Pulping Chemistry

0

H0 /

0

HOC+ 0

0

0

FIGURE 41 Suggested elimination of chloride ions from chlorinated aromatic structures in lignin upon treatment with alkali.

group will increase this yield for such structures. The degree of ionization of the phenol (i.e., the pH value) will determine the extent of radical formation.

F. Alkaline Extraction The role of alkaline extraction (E) following an acidic bleaching stage is to remove the oxidized lignin from the cellulosic fiber by neutralization of the formed carboxyl and other acidicgroups. In modern bleaching,oxygen, hydrogen peroxide,or both areadded in small amounts to furtherenhance the degree of oxidation and,thus, the dissolution of lignin. In addition to being dissolved, chlorinated lignin from a previous C or D stage will also undergo a substantial dechlorination by nucleophilic attack of hydroxyl and hydroperoxy ions [ 1661. This type of substitution may proceed through reactions between chlorinatedquinonestructures in lignin and a nucleophile,asoutlinedschematically in Fig. 4 I . Under acidic (as well as alkaline) bleaching conditions, carbohydrates may become oxidized along the polysaccharide chain, resulting in the formation of carbonyl groups in the sugar units. This type of reaction takes place particularly in ozonebleaching. In a subsequent extraction or alkaline bleaching stage, the introduced carbonyl group will induce a p-elimination reaction that results in a cleavage of the polysaccharide chain and, thus, a loss of viscosity.

REFERENCES 1. E. Adler and J. Marton, ActLI Chern. Scnrzd., 13:75 (1959). 2. J. C. Pew and W. J. Connors, Z ~ p p i 54r245 , (1971). 3. W. J. Connors, J. S. Ayers. K. V. Sarkanen, and J. S. Gratzl, Tappi, 54:1284 ( l 9 7 I). 4. S. Larsson and G. E. Miksche, Acto C/lem. Scrrnd., 26:2031 (1972). 5. G.Gellerstedtand E-L. Lindfors, Hol$o,sclz., 38:151 (1984).

902

Gellerstedt

6. G. Gellerstedt and B. Pettersson, Svensk Pupperstidn., 83:314 (1980). 7. J . Marton and E. Adler, Actu Chern. Sctrnd., 15:370 (1961). 8. K. Lundquist, Actu Chem. Scund., 18:1316 (1964). 9. H. Nimz, Chem. Ber:, 99469 (1966). IO. P. Engstrand, L-A. Hammar, and M. Htun, Proc. Int. Symp. Wood Pulping Clmnisty, Vancouver, B.C., Canada, Vol. l , p. 275 (1985). 11. U. Westermark, B. Samuelsson, R. Simonson, and R. Phil, Nordic Pulp Paper Res. J., 2:146 ( 1 987). 12. U. Westermark and B. Samuelsson, Proc. Int. Synlp. Wood P~tlpingCl1erni.stry, Melbourne, Australia, Vol. I , p. 595 (1991). 13. A. M. Bialski, C. E. Luthe, J. L. Fong, and N. G. Lewis, Can. J. Chern., 64:1336 (1986). 14. G. Gellerstedt and L. Zhang, Nordic Pulp Puper Res. J., 7:75 (1992). 15. I. D. Suckling, Proc. 1111. Symp. WoodPulping Chernistty, Melbourne, Australia, Vol. I , p. 587 ( 1 99 1). 16. J. E. LuValle, J. Am. Chenl. Soc., 742970 (1952). 17. H. I. Bolker, B. I. Fleming, and C. Heitner, Truns. Techn. Sect. CPPA, 4(1):30 (1978). 18. G. Leary and D. Giampaolo, Proc. Int. Svtnp. WoodPulping Clzemistn, Montreal, Que., Canada, Vol. I , p. D4 (1997). 19. F. Imsgard, I. Falkehag, and K. P. Kringstad, Tuppi, 54:1680 (1971). 20. G. E. Styan, Pulp Puper Cl7n., 76(12):88 (1975). 21. D. Y. Lee and M. Sumimoto, Hol$orsch., 45(Suppl.): 15 (1991). 22. G. Gellerstedt, I. Pettersson, and S. Sundin, Proc. Inr. Sytnp. Wood Pulping Chemistry. Tsukuba Science City, Japan, Vol. 2, p. 90 ( I 983). 23. J . Polcin, Paper Celul., 31(3):V9 (1976). 24. S. Moldenius and B. Sjogren, J. Wood Chern. Technol., 2:447 (1982). 25. B. vdsudevan, B. Panchapakesan, J. S. Gratzl. and B. Holmbom, Proc. 1987 Tuppi Pulping Conf, Washington, DC, Vol. 3, p. 517 (1987). 26. G. Gellerstedt and R. Agnemo, Acta Clzem. Scand., B34:275 (1980). 27. R. H. Reeves and 1. A. Pearl, Tappi, 48:121 (1965). 28. M. B. Hocking, M. KO, and T. A. Smyth, Can. J . Chem., 56:2646 (1978). 29. M. Douek and D. A. I. Goring, Wood Sci. Technol.. 10:29 (1976). 30. G. Gellerstedt and L. Zhang, J. Wood Chern. Technol.. 12387 (1992). 31. G. Gellerstedt, H-L. Hardell, and E-L. Lindfors, Actu Chenz. Scand.. B34:669 (1980). 32. S. Moldenius, Svensk Pnpperstidn., 85( 15):R116 (1982). 33. J. L. Roberts, Jr., M. M. Morrison, and D. T. Sawyer, J. Am. Chem. Soc., 100:329 (1978). 34. R. Agnemo, G. Gellerstedt, and E-L. Lindfors, Acta Chern. Scand., B33:154 (1979). 35. R. Agnemo and G. Gellerstedt, Actu Chenz. Scund., B33:337 (1979). 36. S. Moldenius, Some Aspects of Hydrogen Peroxide Bleaching of Mechanical Pulps, Ph.D. thesis, KTH, Stockholm, Sweden (1983). 37. G. Gellerstedt, I. Pettersson, and S. Sundin, Svensk Papperstidn., 86(15):R157 (1983). 38. A. Castellan, N. Colombo, P. Fornier de Violet, A. Nourmamode, and H. Bouas-Laurent, Proc. Int. Symnp. Wood Pulping Chetnisry, Raleigh, NC, Vol. I , p. 421 (1989). 39. R. Agnemo, Paper and a Method of Paper Manufacture, SW. Patent 468054 ( I 992). 40. M. Ek, Some Aspects on the Mechanism of Photoyellowing of High-Yield Pulps, Ph.D. thesis, KTH, Stockholm, Sweden, 1992. 41. M. Ek, H. Lennholm, G. Lindblad, and T. Iversen, Nordic Pulp Paper Res. J., 7:108 (1992). 42. H. Lennholm, M. Rosenqvist, M. Ek, and T. Iversen, Nordic Pulp Paper Res. J., 9: 10 (1994). 43. L. Zhang and G. Gellerstedt, Acra Chern. Scand., 48:490 (1994). 44 K. P. Kringstad and S. Y. Lin, Tappi, 53:2296 (1970). 45, J. Gierer and S. Y. Lin, Svensk Pupperstidn., 75:233 (1972). 46. J. A. Schmidt and C. Heitner, J. Wood Cl7etn. Technol., 13:309 (1993). 47 G. Gellerstedt and E-L. Pettersson, Acta Chem. Scand., B29:1005 (1975). 48. J. Li and G. Gellerstedt, Nordic Pulp Puper Res. J., 13:153 (1998).

Pulping Chemistry

903

49. S. A. Rydholm, Pulping Processes, Interscience, London, p. 467 (1965). 50. S. Haggroth, B. 0. Lindgren, and U. Saeden, Svensk Papperstidn., 56:660 (1953). 5 1. 0. Ingruber. Papiec 2 4 7 11 (1970). 52. H.-L. Schubert, K. Fuchs, R. Patt, 0. Kordsachia, and M. Bobik, Papier; 40(10A):V6(1986). 53. D. V. Glennie, in Lignins: Occurrence, Formation, Structure and Reactions (K. V. Sarkanen and C. H. Ludwig, eds.), Wiley-Interscience, New York, p. 597 (1971). 54. G. Gellerstedt, Svensk Papperstidn., 79537 (1976). 55. B. Lindgren, ActaChem. Scund.,3:1011 (1949). 56. G. Gellerstedt and J. Gierer. Acta Chem. Scand., 24: 1645 (1970). 57. K.Kratzl, E. Risnyovszky-Schafer, P. Claus, and E. Wittman, Holiforsch., 20:21 (1966). 58. 0. Goldschmidand H. L.Hergert, Tappi, 44:858 (1961). 59. G. Gellerstedt and J . Gierer, Svensk Papperstidn., 74:117 (1971). 60. H. Erdtman, Tappi. 32:303 (1949). 6 1 . E. Sjostrom, Wood Chemistry: Fundamentals and Applications, Academic Press, San Diego, p. 133 (1993). 62. S. Aziz and K. Sarkanen, Tappi J.. 72(3):l69 (1989). 63. D. K. Gallagher, H. L.Hergert, M. Cronlund,and L. L.Landucci, Proc. lnt. Symp. Wood Pulping Chemistry, Raleigh, NC, Vol. I , p. 709 (1989). 64. K. V. Sarkanen, Tappi J., 7.?(10):215 (1990). 65. C.J. Venverloo, Holiforsch.. 25: 18 ( 1 97 1). 66. C.Lapierre,Heterogenitedesligninesde peuplier: Miseenevidencesystematique, Ph.D. thesis, 1’Universite de Paris-sud, Centre d’Orsay, Paris (1986). 67. A. Scalbert, Caracterisation des lignines de paille de ble: Fractionnements, associations avec les oses et les acides phenoliques. Ph.D. thesis, I’Institut National Agronomique, Paris-Grignon ( 1984). 68. H. Nimz, Angew. Chern., 78:821 (1966). 69. U. Westermark, B. Samuelsson, and K. Lundquist, Proc. Int. Symp. Wood Pulping Chemistry, Beijing, China, Vol. 1, p. 93 (1993). 70. D. TormundandA. Teder, Proc. Int. Symp. Wood Pulping Chemistry, Raleigh,NC, p. 247 (1989). 71. J . E. Stone and D. W. Clayton, Pulp Paper Mag. Can., 61(6):T307 (1960). 72. S. Norden and A. Teder, Tappi, 62(7):49 (1979). 73. A. Teder and L. Olm, Pap. Puu, 63:315 (1981). 74. K. Sjoblom, J. Mjoberg,and N.Hartler, Pap. Puu, 65227 (1983). 75. B. Johansson, J. Mjoberg, P. Sandstrom, and A. Teder, Svensk Papperstidn., 87(10):30 (1984). 76. J.Gierer, B. Lenz, and N-H. Wallin, Acta Chern. Scand., 18:1469 (1964). 77. J. Gierer, Holiforsch., 36:43 (1982). B. Monties, Svensk Papperstidn., 87:R61 78. G.Gellerstedt, E-L. Lindfors,C.Lapierre,and ( 1984). 79. G. Gellerstedt and E-L. Lindfors, Nordic Pulp Paper Res. J., 2(2):71 (1987). 80. G. Gellerstedt, E-L. Lindfors, C. Lapierre, and D. Robert, Proc. 1st EuropeanWorkshop on Lignocellulosics rind Pulp, Hamburg, p. 224 (1991). 81. C.Gustavsson, K. Sjostrom,and W. WafaAI-Dajani, Nordic Pulp Paper Res. J., 14(1):71 ( 1 999). 82. J. Giererand S. Ljunggren. Svensk Papperstidn., 82(3):71 (1979). 83. J. Gierer and S. Ljunggren, Svensk Papperstidn., 82(17):503 (1979). 84. S. Ljunggren, Svensk Papperstidn., 83(13):363 (1980). 85. J. Gierer, F. Imsgard, and I. Pettersson, Appl. Pol.yrner Symp., 28: I195 ( 1 976). 86. J. Giererand S. Wannstrom, Hol7j%rsch., 38:181 (1984). 87. S. Ohara, S. Hosoya, and J. Nakano, Mokkuzai Gakkaishi, 26:408 (1980). 88. T. Fullerton and A. L. Wilkins, J. Wood Chem.Technol., 5(2):189 (1985). 89. G . Brunow and G. E. Miksche, Appl. Polymer Symp., 28:l 155(1976). 90. F. Berthold, E-L. Lindfors, and G. Gellerstedt, Holiforsch., 52:398 (1998).

904

Gellerstedt

91. F. Berthold, E-L. Lindfors,andG.Gellerstedt, Hol$orsch., 52:481 (1998). 92. L.GustafssonandA. Teder, Svensk Papperstidn., 72(8):249 (1969). 93. G. Gellerstedt, K. Gustafsson, E-L. Lindfors, and D. Robert, Proc. I n t . Symp. on Woodand Pulping Chemistry, Raleigh, NC, p. 21 1 (1989). 94. D. R. Dimmel, L. F. Bovee, and B. N. Brogdon, J. Wood Chern. Technol., 14(1):I (1994). 95. J. Gierer and I. Noren, Acta Chem. Scand., 16:1976 (1962). 96. J.Giererand S. Wannstrom, Hol$orsch., 40:347 (1986). 97. J. Gierer, Wood Sri. Technol., 19:289 (1985). 98. P. Karhunen, J. Mikkola, A. Pajunen, and G. Brunow, Nordic Pulp Puper Res. J., 14(2):123 ( 1 999). 99. J. Turunen, Pup. Puu, 43:663 (1961). Hol7forsch.. 42:385 (1988). 100. V. L. ChiangandM.Funaoka, 101. V. L. Chiang and M. Funaoka, Hol$orsch., 44: l47 (1990). 102. F. D. Chan, K. L. Nguyen, and A. F. A. Wallis, J . Wood Chern. Technol., 15(3):329 (1995). 103. G.Gellerstedtand D. Robert, Acta Clzem. Scund., B41:541 (1987). 104. 0. Karlsson and U. Westermark, NordicPulpPuper Res. J., 12(3):203 (1997). 105. R.KondoandK. V. Sarkanen, Hol;forsch., 3831 (1984). 106. M. H. Johanssonand 0. Samuelsson, Curbohydr. Res., 54295 (1977). 107. J. Buchert, A.Teleman, V. Harjunpaa,M.Tenkanen, L. Viikari, and T. Vuorinen,Tuppi J., 78(11):125 (1995). 108. R. Aurell and N.Hartler, Rippi, 46:209 (1963). 109. P. J.Kleppeand K. Kringstad, Norsk Skogirzd., 17(11):428 (1963). 1 10. H. H. Holton, 63rd Annual Meeting Technical Section CPPA 1977, Preprints, p. A107. I 1 1. L. Lowendahl and 0. Samuelsson, Svensk Papperstidn., 80(17):549 (1977). H. I. Bolker, Wood Sei. Technol., 14:207 112. G. J. Kubes, B. I. Fleming,J.M.Macleod,and (1980). 113. J. Gierer, M. Kjellman,and I. Noren, Hol~ot-.sch.,3717 (1983). 114. D. R. Dimmel and L. F. Schuller, J . Wood Chern. Technol., 6(3):345 (1986). 1 IS. F. Berthold and G. Gellerstedt, Hol&rsch., 52:490 (1998). 116. M. Lindstrom and A. Teder, Nordic Pulp Paper Res. J., 10(1):8 (1995). 117. G. Gellerstedt and E-L. Lindfors, Svensk Puppersfidn., 87(15):R115 (1984). J . Wood Chetn. Technol., 4:239 118. D.Robert, M. Bardet,G.Gellerstedt,andE-L.Lindfors, (1984). 119. R. C. Francis, Y-Z. Lai, C. W. Dence, and T. C. Alexander, Tappi J . . 74(Y):219 (1991). 120. T. Iversen and S. Wannstrom, Hol$or.sch., 40: l9 ( I 986). 121 . G. Gellerstedt, W. Wafa Ai-Dajani, and L. Zhang, Proc. 10th lnt. Swnp. on Wood and Pulping Chernisrn, Yokohama, Japan ( I 999). 122. B. D. Favis, J. M. Willis, and D. A. I. Goring, J . Wood Chem. Technol., 3(1):1 ( 1983). 123. G. Gellerstedt and A. Torngren, Proc. 8th Int. S>vnp. on Wood and Pulping Chemistry, Helsinki, Finland, Vol. 2, p. 285 (1995). 124. B. Sjogren, D. EhrengBrd, P. Engstrand,and M. Htun, in Properties of Ionic Polymers. Nuturul and Synthetic (L.SalmenandM.Htun,eds.)SwedishPulpandPaperResearch Institute, STFI-Meddelande A-989, p. 91 (1991). 125. G. GellerstedtandJ.Li, Curbohydr. Res., 29441 (1996). 126. J. Li andG.Gellerstedt, Carbohydr. Res., 302:213 (1997). 127. J. Li and G. Gellerstedt, Nordic Pulp Puper Res. J., 13(2):153 (1998). 128. K. Kratzl, P. K. Claus,A.Hruschka, and F. W. Vierhapper, Cell. Chern. Tecl~nol.,12:445 (1978). 129. S. LjunggrenandE.Johansson, Hol$orsch., 44:291 (1990). 130. E. Johanssonand S. Ljunggren. J. Wood Chem. Technol., 14:507 (1994). 13 I . G. Gellerstedt and E-L. Lindfors, Tappi J.. 70(6):1 19 (1987). 132. G. Gellerstedt and L. Heuts, J. Pulp Puper Sei., 23:J335 (1997). 133. E. C. Johansson, The effect of oxygen on the degradation of ligninmodel compounds and residual lignin, Ph.D. thesis KTH, Stockholm, Sweden, 1997.

Pulping Chemistry

905

134. C. W. Dence, in Pulp Bleaching.Principles and Pructice (C. W. Denceand D. W. Reeve, eds.), TAPPI Press, p. 125 (1996). 135. A. Robert, P. A. Rerolle. and C. Martin-Borret, Revue ATIR 18:15 I (1964). 136. J. D. Sinkeyand N. S. Thompson. Pup. Puu. 563473 (1974). 137. M. Ek, J. Gierer,and T. Reitberger, Proc. 3rd Int. Symp. on Wood and Pulping Chemistry. Vancouver, B.C., Canada, p. 209 (1985). 138. R. Malinen, E. Sjostrom,and J. Ylijoki, Pap. Puu. 55:5 (1973). 139. A. GBrtner and G. Gellerstedt. J. Pulp Paper Sci., in press. 140. G. Gellerstedt and 1. Pettersson, J. Wood Chern. Technol., 2:23 I (1982). 141. J. Basta, L.Holtinger, and J. Hook, Proc. 6th In?. Swnp. on Wood and PulpingChernistrv, Melbourne, Australia, Vol. 1, p. 237 (1991). 142. G. Gellerstedt, L. Heuts, and D. Robert, J . Pulp Puper Sci., 25: I 1 I (1999). 143. L. Heuts and G. Gellerstedt, Nordic Pulp Puper Res. J., 13:107 (1998). 144. R. M.Berry and B. I. Fleming, Hol
This Page Intentionally Left Blank

Index

1,4-Anhydroglucopyranose,642 6-0-Alkylcellulose, 6 13 Absorption, 98, 111, 264, 267, 349, 357, 364, 476, 577, 582, 621, 701, 704, 711, 725 Acetal linkage, 620 Acetic acid, 184, 267, 280, 293, 302, 303, 3 IO, 312, 313, 315, 455, 466, 490, 600, 604, 698, 873, 882 Acetic anhydride, 313, 469, 662 Acetone-rosin, 8 12 Acetylation, 119, 128, 184, 207, 300, 3 13, 350, 405, 406. 456, 476, 575, 579, 780, 582, 591, 594, 600, 604, 615, 616, 631, 639, 661-663, 667, 670, 684 Acid hydrolysis, 105, 153,183,186,190, 263, 282, 302, 305, 307, 31 I , 444, 450, 460, 467, 474, 530 Acid rain, 514, 528, 530, 544 Acid stain, 425, 426, 428, 433 Acidolysis, 110, 112,116,129,130,131,132, 150, 153,158,159,169,317,319,342, 459, 464, 877 Acrylonitrile, 708, 8 13 Activated sludge, 828, 838, 840, 845 Active oxygen species, 551, 559, 560 Acylation, 617, 657, 658, 659 ACZA (see Ammoniacal copper zinc arsenate) Adhesion, 213, 234, 265, 412, 428, 439, 684, 696, 709, 733, 749, 757, 763, 766, 770, 775, 844 Adhesive shear strength, 735, 739, 740, 742, 755, 756, 758, 760 Adhesive strength, 409, 684, 733-735, 738, 739, 742, 753, 757, 760, 763, 770

Adhesive tensile strength, 734-736, 739, 743, 755, 759, 760 Adhesives, 232, 264, 412, 424, 426, 439, 673, 684, 688, 694, 705, 709, 733, 765 Adsorption isotherm, 701, 725, 726, 729 Alcohol-ether, 8 12 Alcoholysis,131,166, 687, 697 Aliphatic hydroxy (see Hydroxyl group) Alkaline hydrolysis, 206, 448, 450, 475, 476, 485, 486, 884, 891 Alkaline pulping, 448, 469, 475, 476, 873, 879, 883, 886 Alkaline stain, 427 Alkaloid, 226, 227, 232, 234 Alkenylsuccinic anhydride, 601 Alkylketene dimer, 6 I O a-carbonyl group (see Carbonyl group) a-pinene, 222, 357, 768 Amidation, 600, 621 Ammoniacal copper azole, 803 Ammoniacal copper citrate, 803 Ammoniacal copper quat, 803 Ammoniacal copper zinc arsenate, 803 Amorphous cellulose, 298, 444, 497, 601 Amphoteric cellulose, 615 Anatomy, 5 1, 242, 244, 8 16 Anchor effect, 742 Angiosperm, 1, 175, 259, 260, 3 19, 43 1 Anthraquinone, 226, 467, 473, 481, 493, 621, 870, 884 Antioxidant, 214, 218, 229, 263, 288, 436 Antiswelling efficiency (ASE), 576 Arabinan, 61-63, 185,189,191, 277 Arabinofuranan, 65 1 907

908 Arabinose,46,47. 61, 176,186,195,199, 205, 232, 281. 303, 307, 450. 474, 477, 541, 645. 883 Arabinoxylan. 52. 57. 62, 184. 201, 203, 205, 303. 476 Archeologists. 801, 8 10 Aromatic ring opening, 464, 478, 482. 491, 562, 886 (see also Ring opening) Aryl ether cleavage, 150, 464,478, 482. 49 I , 886 Ash, 30, 52, 56, 72, 367, 368, 816, 856 (see o l s o Inorganic constituent) Autohydrolysis, 350, 466, 467, 694 Autoxidation, 264, 559, 863, 864. 869 Auxochrome, 386 Backing material, 769, 778 Bacterial cellulose, 85, 91, 92, 96, 98, 100. 295, 300. 601, 609 Balltack test. 749 Bark, 31, 71, 218, 227, 236, 242, 365, 368, 431. 8.50, 863 inner, 244, 247, 251. 253, 255, 263. 266, 267 outer, 242, 244, 25 l , 263, 267 Bast fiber, 266. 267 Benzyl alcohols, 116, 129-13 I . 156, 447, 498, 550, 552, 565, 694, 840. 869. 871, 873 Benzyl ether, 1 19, 129, 130, 448, 450, 46 I , 488. 492, 554, 686, 687 Benzylation, 660, 842 P-caprolactone, 840 (see N ~ S OPolycaprolactone) P-ether bond cleavage. 485. 552 Betulin. 251, 253 Biobleaching, 547, 557 Biocide combinations, 803, 804 Biocides, 264. 795. 797, 800, 802, 804 Biodegradability, 573, 828. 829, 830, 844, 845 Biodegradable plastics, 827, 828, 83 I , 846 Biological activity, 189, 221, 234, 251, 261, 799 Biomass, 169, 450, 466. 695, 696, 827, 828 Biosynthesis. 21 3, 220, 228. 249, 261, 627 flavonoid, 214. 217 lignan, 214, 217 lignin.108,146, 559 stilbene, 220 tannins, 256 taxol, 236 Bleaching, 304, 315. 325. 330, 364, 400, 403. 406, 426. 429, 431, 436, 493, 496-500, 510, 520, 547, 548, 559, 562, 565, 567, 618, 621. 859, 860, 862, 864, 866, 868, 869, 874, 880, 889, 890, 893, 859, 897, 90 1

Index Bleeding, 834. 836, 838 Boric acid. 284, 438, 793. 796. 804, 810 Briquette, 850 Bromination, 65, 66. 69, 617 Brown-rot fungi, 263. 429 Bulking effect, 575 Cadoxen, 288, 3 I 5 Carbonization. 659. 693, 694. 785. 790 Carbonyl group, I 14. I 16,124,125,146,153, 158, 405, 407. 447, 515. 516, 522. 529. 867, 893, 894. 898, 900.901 a-carbonyl, 119.135,142, 329, 407, 520, 863 Carboxyl group, 7 I , 186, 309, 3 I 1-3 13. 437. 520, 528. 544. 588, 599, 611, 615, 620, 621. 683. 891. 893. 894, 897. 898, 900 Carboxynlethylcellulose (see Cellulose derivatives)

Cascade, 850, 852, 854, 857 Catcchol, 322, 346, 425, 432, 860, 863, 868. 89 I CCA (sec Chromated copper arsenate) Cello-octaose, 629, 63 I . 636 Cello-oligosaccharides,628. 629, 63 I , 634, 636, 639, 642 Cello-uronic acid. 620 Cellulase, 105. 305.627 Cellulose, 35-39 Cellulose acetate (see Cellulose derivatives) Cellulose carbamate (see Cellulose derivatives) Cellulose derivatives, 286, 288, 293, 298, 495, 599, 601, 603, 604, 607, 609, 612, 615, 617, 62 I , 622, 627, 628, 639, 649-65 I , 827. 828 carboxymethylcellulose,98, 61 2, 613 CMC. 98. 103, 603, 606. 607, 612, 615 cellulose acetate, 600, 601, 604-606, 609, 613, 615, 617, 657, 662, 664, 675. 827, 828-830, 832-835, 838 (.see SO Cellulose diacetate and cellulose triacetate) cellulose carbamate, 599 cellulose diacetate. 600, 601, 603, 604. 830 (see c ~ l s oCellulose acetate and cellulose triacetate) cellulose nitrate. 290. 292, 293, 600, 657 cellulose triacetate, 85, 87, 600, 604, 639, 645, 647, 662, 663 cellulose xanthate, 496 hydroxyethylcellulose, 6 13, 6 16 hydroxypropylcellulose. 607, 6 13, 6 1 5 methylated cellulose, 652 methylcellulose, 613. 615, 616

Index Cellulose diacetate (see Cellulose derivatives) Cellulose ester, 292, 600, 603, 605-607, 609, 627, 658, 828-830. 832, 834 (see ulso Cellulose derivatives) Cellulose nitrate (see Cellulose derivatives) Cellulose solution, 288, 289, 444, 601, 606, 607, 609, 6 I3 Cellulose solvents, 305, 603, 605-607, 612, 617, 622, 657, 658 Cellulose triacetate (see Cellulose derivatives) Cellulose xanthate (see Cellulose derivatives) Charcoal, 357, 849, 850. 857, 858 Chemical composition, 51, 76, 78, 242, 244, 274, 282, 330, 810, 816 Chemical modification, 1 10, 114, 3 13, 473, 538, 573, 599, 657. 661, 670, 827, 834. 842. 844 Chlorine, 72, 282. 320, 321, 325, 364, 368, 497, 498, 565, 567. 672, 673, 874. 889, 890, 895, 897 Chlorine dioxide, 282, 364, 497, 874, 889, 890, 895, 897 Chlorothalonil, 801, 803 Chromaticness index. 394 Chromophores, 124,152,328, 386, 387, 515. 860, 862-864, 867 CMC (see Carboxymethylcellulose) Cohesive failure. 734, 746. 748, 755, 759, 775 Cole-Cole plot, 7 15, 7 17 Color difference, 394. 588 Color, 2, 39, 70, 123, 124, 130, 213, 234, 267, 286, 309. 332. 353. 38.5, 514-516, 518. 520, 528, 530, 536, 573. 583. 588, 812, 819. 820, 860, 861, 863, 869, 874, 889 Combustibility, 785, 786, 790 Compatibilizer, 76, 672, 679, 836, 844 Composite crystal model, 83, 85, 96, IO5 Compression wood (see Wood) Concrete-form plywood, 850 Condensation, 1 15 , 1 16, 144. 169. 22 1, 222, 229, 266, 278, 319. 344, 346. 352, 357, 450, 452, 458, 459, 46 1-464, 467, 470, 478, 486, 488. 491, 558, 601, 627, 667, 693, 695, 781, 782, 784, 820, 871, 873, 877, 880, 882, 888 (see also Lignm condensation) Coniferyl alcohol, 108, 119,123.134,146.148. 150, 156,161.165.169,214,233, 315, 349, 447, 478, 486, 558, 698, 862, 869, 880 Consolidant, 808, 809, 81 I , 81 3, 819 Convergent synthesis, 629, 631, 633, 636 Copper dimethyldithiocarbamate,803

909 Copper naphthenate, 802 Cork cell, 245 CP/MAS ''C NMR (see Spectroscopy) Creosote, 787, 797, 798, 799 Cross-linking, 200, 201, 203, 235. 246, 462, 492, 575, 580, 586, 609 (see also Chemical modification) Crystal structure, 83, 95, 96, 295, 604, 607, 627, 639, 647 Crystal transformation, 85, 87, 95, 96 Crystallization process, 96, 105 Cu-8 (see Oxide copper) Cuen, 288. 314 Cyclic ester, 838, 840 DABCO, 528 (see also 1,4-Diazobicyclo (2,2,2]-octane) 1,4-Diazobicyclo[2,2,2]-octane. 528 Deacetylation, 405. 456, 601, 604, 613, 636, 647 Degree of substitution, 293. 298, 600, 657, 662, 828 Delignification, 21, 56, 57. 61, 150, 267, 282. 284. 300, 303, 309, 337. 352, 443, 445, 464, 466, 467, 469, 489, 490, 491, 493, 497. 499, 500, 547, 558, 564, 565, 660, 871, 875, 877, 880, 886, 889, 891, 894, 897 Demolished house, 849, 850 Deoxyhalogenation. 599, 61 7 Dialdehyde cellulose, 619, 620 Diarylheptanoid, 221, 222, 432 Dibasic acid anhydride, 835, 838 Dielectric properties, 713, 716, 720 Dimensional stability, 279, 573. 576-578, 580, 587, 696, 701, 781. 784, 813, 819. 824 Dipentene, 768 Discoloration, 385, 393. 394, 397, 402, 404, 409, 419, 425, 429, 436, 443, 513, 5 16, 518, 520, 532, 538, 863, 869 by acid, 425 by alkali, 427, 864 by enzyme, 432. 434 by exudation of resin, 436 by iron, 410 by light, 394 by microorganism, 429 by oxidation, 434 photoinduced, 394, 397, 398, 400. 402-404, 407. 420, 5 I6 Dithionite, 860, 864, 868 Durability, 2, 213, 234, 391, 514, 796, 810, 828

910

Dynamic mechanical properties. 581, 590, 701703, 705. 709, 710, 716, 720, 735, 738, 742, 746, 748, 763 loss modulus, 702, 710 loss tangent, 235, 580, 583, 591-593, 671, 672. 674, 676, 714 storage modulus, 702, 738, 749 Earlywood, I , 4, 13, 28, 57, 67, 72, 74, 435, 526. 534, 536, 584 (see also Spring wood) Eicosaose, 636, 639, 642 Electrondiffractometry, 101, 105 Endwise depolymerization, 469, 470, 478 Epoxy resin, 9, 688-690, 701, 738, 742 Esterification, 207, 310, 599, 601, 603, 605, 607, 610, 612. 622, 657, 662, 684, 827, 831. 834, 835, 842 (see also Chemical modification) Etherification, 599, 601, 612, 621, 657, 827, 842 (see also Chemical modification) Etherified units, 129, 137, 448, 458, 462, 480 External plasticizer (see Plasticizer) Extraction, 124,169,175,180,188,235, 244, 245, 246, 258, 265-267, 218, 282, 284, 286, 300, 302, 315, 353, 355, 357, 360, 362, 366, 409, 429, 432, 467, 557, 610, 664, 889, 891, 895, 898, 901 Extractives, 5, 54, 124,163, 213, 218, 232, 234, 244, 246, 247, 248, 251, 265, 274, 276, 279, 332, 353, 355, 357, 359. 364, 394, 425, 429, 435, 5 12, 514, 655, 809, 816, 973, 817 Fertilizer, 367, 850 Feruloyl group, 199, 201 Feruloyl oligosaccharides, 188, 201 Fiberboard, 280, 586, 683, 684, 850, 851, 855 Filled polymer, 710, 71 l , 722-724 Fire-retardant property, 785, 790, 792 Flavonoid, 218, 232, 251, 255, 258, 263, 366, 388 Formalization, 575-577, 582, 586, 592 Fracture mechanics, 759, 760 Free radicals, 108,135,142,144,148,151, 319, 404, 407, 409, 459, 467, 497, 516, 517, 524, 526, 557, 562, 898 cellulase, 5 18 coniferyl, 137,156 hemicelluloses, 5 I8 hydrazyl, 152 hydroxyperoxide, 520, 522, 524 lignin, 520

Index [Free radicals] oxygen, 559, 560 peroxy, 408, 522, 527 phenoxy,132,148, 151, 152,405,407, 432, 524, 544, 868, 891, 897 sulfonyl, 524 Freeze-drying, 8 13 FTlR (see Spectroscopy) Fuel, 267, 353, 467, 849, 850, 854, 857 Galactan, 55, 57, 61, 63, 185, 189, 192, 232, 300, 314, 456 Galactofuranan, 65 1 Galactoglucomannan, 46, 52, 55, 56, 63, 284, 300, 303, 314, 450, 456, 476, 478, 882, 889 Galloyl, 228, 256 Gas pollutants, 524 Glacial acetic acid, 309, 329 Glass transition temperature, 676, 677, 722, 736, 738, 739, 742 Glucomannan, 46, 52, 56, 57. 61, 63, 98, 203, 207, 284, 300, 302, 3 12. 3 14, 444, 450, 456, 471, 478, 490, 873, 880, 882 Glucuronoarabinoxylan,52, 57, 62, 63 Glyceride, 232, 353, 360, 362 Glycerine, 694, 8 1 1 Glycosidic cleavage, 475, 476, 495, 497 Glycosylation, 26 l , 628, 629, 63 I , 634, 636, 642, 644, 649, 651 Grafting, 599, 617, 670, 671, 683, 736, 738, 740, 834, 835, 838, 840, 880 Guaiacylunits, 52, 70, 108,117,126,128,135, 154, 322, 344. 462 Gymnosperms, l , 2, 175. 192, 203, 3 19, 43 1 Hardwood (see Wood) Heartwood (see Wood) Hemicellulose, 45-47 Holocellulose, 7, 58, 59, 244, 267, 274, 276, 282, 300, 309, 3 14, 514, 530, 540, 816 Homogalacturonan, 185, 186,190, 207 Hot-melt adhesives, 673, 684, 765 Hue, 394, 400, 402, 860 Hydrogen bonding, 84, 91, 96, 98, 105, 207, 235, 286, 293, 294, 444, 448, 574, 575, 599, 657 Hydrogen peroxide, 137, 264, 267, 320, 346, 368, 403, 406, 426, 431, 437, 499, 549, 552, 554, 557, 559, 564, 860, 864, 866, 890, 893, 894, 901 Hydroperoxidation, 520, 523, 544 Hydroxyethylcellulose (see Cellulose derivatives)

Index Hydroxyl group aliphatic, 116,125,128,447 phenolic, 110, 111, 113,117,126.127-130, 135,137.148,152, 235, 267,320, 328, 330, 334, 405, 421, 445, 447, 448, 450. 458, 459, 462, 468, 491, 5 15, 520, 526, 540. 875, 880,888, 889, 891. 897, 900 Hydroxypropylcellulose (see Cellulose derivatives) Hygroscopicity, 213, 573, 576-578, 593, 595, 809. 810 Imidate method, 628, 631 Inorganic constituent, 734, 748, 755, 759 (see also Ash) Interfacial failure, 734, 748, 755, 759 IR (see Spectroscopy) IR spectrum, 114, 298, 647 Iron oxide. 412, 807 Iron stain, 410-414, 420, 427, 435 Isoflavonoid, 2 18 Isoprenoid, 222, 225, 226 Kathon 930, 803 Klasson lignin (see Lignin) Kraft pulping, 251, 331, 333, 360, 364, 466, 476, 490, 491, 493, 499, 621, 869, 873875. 877, 880, 882, 884, 888 Laccase. 547, 552, 558, 566 Langmuir isotherm, 725 Latewood, 1, 4, 57, 72, 534, 536, 537, 584, 817 (see also Summer wood) Lightness, 387-389, 39 I , 397, 398, 405, 409, 413, 437, 438, 513. 538 Lignan, 57, 134,169,214.216. 236, 251, 254, 366, 388, 402, 435 Lignin, 39-44 hardwood, 52, 70, 108,110,119.134,166, 3 15. 3 19, 328. 447 Klasson lignin, 816 mill wood,108, 114, 116,117,119,123,124, 126,128,129,130,131,134,146,153, 156,158,161,169,316, 322, 325, 329. 346, 350, 405, 445. 450, 520 softwood, 52.108, 1 10,12,134, 150, 165, 315, 318, 322, 328, 447, 877 Lignin-carbohydrate complex, 47. 76, 168. 443. 448, 461, 492, 656 Lignin biodegradation. 547-549. 558 Lignin condensation, 278, 461 -464, 467. 469. 478. 49 1 ( w e n l s o Condensation)

91 1

Lignin oxidation, 135 nitrobenzene, 135, 246, 319-322, 325, 464 permanganate, 133- 137, 320, 32l , 346, 49 1 Lignin peroxidase, 547, 550, 557, 558 Lignocellulosics, 696, 827, 835, 842 Loss modulus (see Dynamic mechanical properties) Loss tangent (see Dynamic mechanical properties) Manganese peroxidase, 547, 557, 566 Marine environment, 807 Mechanical pulping, 548, 859, 860 Metal alkoxide, 781, 782, 784, 789 Metalloporphyrins (see Cellulose derivatives) Methylated cellulose (see Cellulose derivatives) Methylcellulose, 613, 615, 616 Micro-Brownian motion, 592, 593, 595, 596, 675, 702, 709, 716, 721, 763 Microfibril, 5-12, 15, 25, 30, 35, 38, 40, 44, 46, 58, 75, 78, 87, 92, 96, 98, 100, 103, 207, 294, 476, 573, 578, 581, 585, 590, 602 Middle lamella, 12, 51, 59, 61-63, 70, 166, 168, 185, 300, 353, 369, 500, 534, 536, 538, 655 Mill wood lignin (see Lignin) Moisture-conditioned wood, 782, 789 Molecular dynamic simulation, 95 Monoclinic unit cell, 92, 95, 295 Monoepoxide, 834, 835, 838 Monosaccharides, 176,183,184,186,284,304, 305, 309, 312, 452, 466, 639 Morphological region, 51, 57, 5 8 , 60. 65, 68, 69, 72, 166, 274. 369, 445, 500 Munsell color specification, 39 I MWL (see Lignin, mill wood lignin) Native cellulose, 83, 85, 91, 96, 105, 282, 294, 295, 476, 601, 620 Neoflavonoid, 2 I8 Neolignan, 214, 254, 402 Nitration, 284, 293, 600, 609 Nitric oxide, 518, 524, 525, 544 Nitroso group, 525 Norlignan, 2 16-2 18, 435 Normal wood (see Wood) o-quinone (see Quinones) Oil-borne preservatives, 797 Oligoesterification, 834, 835 Oligosaccharides, 188. 190.199, 206, 306, 308, 313, 452, 628

912 One-electron oxidation mechanism, 547, 549, 550, 559. 562 Oxalic acid, 413, 415, 418, 420, 421, 425, 435, 456, 694, 695 Oxidative degradation, 161, 288, 320, 321, 344, 346,445, 456, 493, 5 13, 530, 566, 894 p-quinone (see Quinones) Pallet, 802, 850 Particleboard, 264, 266, 280, 405, 577, 580, 849, 850, 852, 855, 857 Pectin, 26, 52, 56, 59, 175,180,185,189,190, 203, 205, 245, 266, 353, 809 Peel strength, 709. 735, 743, 748, 749, 757, 772, 77.5 Peeling reaction, 470, 471, 473, 475, 476, 478, 493, 496, 882, 884, 889, 893 Pentachlorophenol, 263, 797, 799, 815 Pentosan, 55, 59, 302, 310, 452, 809 Peracetic acid, 282, 498, 499, 565, 664, 666 Periderm, 242, 244 Periodate, 124-126, 137, 3 10. 31 I , 320, 493, 496, 619, 620 Permanent fixation, 586 Permethylation, 184, 286, 613, 616 Peroxyacetic acid, 284 Phanerochaete chrysosporium, 547-550, 554. 557 Phenol-formaldehyde resin, 264, 265, 439, 688. 819-821 Phenolic acid, 206, 228, 246, 258, 557, 561 Phenolic hydroxyl (see Hydroxyl group) Phenolic units, 128, 137, 229, 344, 478, 483, 486, 865, 875. 879 Phenolics, 175. 188. 232, 246, 25 l , 3 15 Phenylcarbanilation,609 Phenylpropanoid. 214, 226, 232, 251 Phloem, 4. 26, 28, 242, 244. 252 Photo-induced discoloration (see Discoloration) Photodegradability, 525. 845 Plasticizer, 580, 582, 583, 593, 595. 656, 667. 668. 670, 68 I. 682, 830, 834. 841 external, 667. 668, 673. 681 Polycaprolactone. 832, 834, 842. 843 Polyethylene glycol. 406. 575, 579, 694, 696, 811. 814. 815, 817, 819 Polyflavanoid, 258, 264-266 Polysulfide, 473. 870, 876, 884, 886, 888 Polyterpene, 222. 768, 769, 778 Polyurethane, 263, 267. 405, 409, 420, 432, 437, 688. 689, 696. 819 Primarycellwalls,175.176.178. 192. 201. 207, 2 10

Index Primary wall, 12, 13, 15, 18-20,25,28.30, 58, 60, 62, 63, 176, 185, 861. 862 Proanthocyanidin, 218, 229, 246, 255, 256, 258, 260, 26 1, 263, 265 Probe tack test, 749 Property enhancement of wood, 784, 789, 793 Property enhancer, 787, 788, 792, 793 Propiconazole, 803 Propylene oxide treatment, 575, 577. 580, 582, 59 I , 592 Quinone methide, 698, 871, 876-879, 880, 886, 888 Quinones, 123,125,128,130, 226, 388. 539, 861, 863, 864, 866, 867, 869, 897, 901 o-quinone, 344, 499 p-quinone, 860, 862, 866 Radial compression, 584 Radicals (see Free radicals) Ray parenchyma, 2, 4. 19, 20, 21, 66, 67. 70, 72, 74, 369, 817 Reaction kinetics, 482 Reaction wood (see Wood) Regenerated cellulose, 85, 444, 601, 606, 613, 647 Regiospecific control, 628, 629 Relaxation process, 592. 593, 595, 675, 721, 832 Resorcinol-phenol-formaldehyde,8 19-822, 824 Rhamnogalacturonan, 185, 189,190.203 Rhytidome, 244 Ribbonassembly,96-98.100-104 Ring opening, 156. 498, 552, 562, 564 Ring-opening polymerization, 627, 628, 639, 641, 644. 645, 648, 650, 65 I , 838, 840 Rolling friction coefficient. 752, 758, 770, 775 Rosin, 222, 364, 768, 769, 778, 812, 819 Saccharitication, 267, 444, 466, 469 Sapwood (see Wood) Sawdust. 278. 358, 369, 432, 850 Secondary wall, 4. 5 , 12, 13, 16, 18. 24, 28. 30, 35. 37. 39. 40, 57, 58, 61. 66. 70. 72. 74, 75. 201, 203, 346. 369. 445. 500, 534. 536, 574, 817 Semicarbazide. 405, 436 Shipwreck, 807, 808 Shrinkage, 575, 808, 810, 814 Sizing. 97. 222. 610, 61 I Sodium borate. 302. 433. 8 I0 Sodium dihydrogen phosphate. 4 I9 Softwood lignin ( s c v Lignin)

913

index Soil burial test, 838 Sol-gel proccss, 781 Solubility. 119,201,235,245.282,292, 302, 304. 353, 366, 492, 539, 540, 599, 604, 612. 655. 687, 695, 725. 803, 888. 891 Solubility parameter, 708, 709, 726 Solvent exchange. 289, 294, 726, 808 Spectroscopy AES. 367-369 capillary GC-MS, 180 CP/MAS "C NMR, 83, 85-87 EDXA. 65, 66. 69.70, 72. 77, 315, 369, 787 ESCA, 538, 544, 734 ESI-MS, 1 82 ESR, 5 16-518. 525, 544. 550, 554 FAB-MS, I82 FTIR. 89, 91, 295. 330, 332, 529, 734 IR, 112-1 14. 330 LC-MS. 180 MALDI-TOF-MS, 182 NMR, 7, 110, 115, 116.119,12.5,131.156, 182,199.258.281,294,334.763,X40 "C-NMR, 83.85, 91, 92, 96, 1 17, 256. 294. 298, 334. 340. 639. 645, 831 SEM, 65. 66, 74, 369. 535. 584, 657, 697, 785, 790, 816,829,833 TEM, 17, 72 UV, 65.66, 70. 71, 110. 125.132,146,166, 328, 866 Splaycd microfibrils. 98,103 Spring wood, I ( s e e d s o Earlywood) 695, 696, 697 Starch. 2-4, 32. 52, 294, 620. Stereospecitic control, 628, 629 Steroid, 222. 224 Stilbene, 156.218,220,253,366,388.486, 49X.861,863,X79,X91 Stopping reactions. 472. 473, 476, 477, 883 Storage modulus ( w e Dynamic mechanical properties) Strain energy releasc rate, 760 Stress-freecrystallization, 104, 105 Styrcne-maleic acid anhydride copolymcr. 844 Suberin, 244-246 Substituent effects, 631. 642-645, 648, 649. 65 I Sucrose. 232, 812 Sultite pulping. 456, 464. 466. 869, 870. 87 1. 873 Sultite radical, 524, 525, 544 Sulfonation. 131. 265, 464. 466, 606, 609, 862. X7 l , 873 Sulfonyl radical. 524 Sulfur dioxide. 280, 518. 524, 525, 528. 530, 544. 864, 869-871

Summer wood, I (.see d s o Latewood) Superimposed normalized dielectric absorption, 720 Surface treatment, 403, 62 I , 807 Swelling, 288, 294, 455, 476, 575-577. 582. 591. 726, 810, 862 Syringyl moiety, 458 Tack, 749, 75 I , 758, 770, 771, 775 Tackitier, 766, 768, 769, 773, 779 Tannin, 218, 227, 232, 235, 244. 246. 253, 256. 258. 263, 264, 267, 353, 365, 388. 413. 425, 429. 43 I . 439 Taxol. 236,249,261 TBTO (sec Tributylin oxide) Tebuconazole. 801, 803, 804 TEMPO, 620 Tension wood (see Wood) Termite deterioration, 785 Terpenoid, 222, 247, 355, 358, 437 Thermoplasticity. 573, 655. 656, 657, 660, 672, 676.830,835, 842-844 Thioaceticacid, 133,135,142,159,166. 469 Titanic, 809 Topochemical effect, 344. 500, 793 Topochemistry. 51, 344, 500, 525. 7x4. 793 Tracheid, I , 13, 25, 42,46, 51, 438,536.537, 817 Treating processes empty cell, 795 full cell, 795 supercritical fluid, 796 vapor phase, 796 Triclinic unit cell. 89, 95 Trifluoroacetate. 607 Trifluoroacetic actd, 183,304,305.334.607. 662, 666 Tropolone, 222, 224, 226, 248. 366 U.S.S. Cairo gunboat, 817 background. 814. 8 15 consolidation. 8 17 history, 814,815 UV microscopy (see Spectroscopy) Vapor sorption. 722, 723 Veratryl alcohol, 550.554. 556 Vessel, 1. 4,18. 3 l , 3.5,70. 74. 368. 4 12. 420, 424, 575. 812, 814, 817, 822. 829 Watcr-borne preservatives, 799 Water-saturated wood, 782, 783 Water absorption ratio (WAR). 787

914

Water repellency, 610, 805 Waterlogged wood, 807 preservation, 8 I O properties, 809, 810 Wattle, 232, 263, 264, 266. 267 Wax, 232, 246-248, 353, 362,409,526, 607, 81 1 Weathering, 512, 528, 535, 541, 766, 805, 807, 816, 819 Weight percent gain (WPG), 577, 784 Whiteness, 429, 547-549, 557, 566, 788 Wood compression, 2, 22, 32, 36, 37, 39, 41, 42, 44, 45, 54, 55, 57, 58, 63, 68, 116, 134, 142,168, 879 hardwood, 1, 7, 25, 31, SI, 69, 92, 108, 112, 119. 123, 1 4 4 , 150, 175, 214, 246, 251, 265, 282, 300, 350, 434, 587, 604, 490, 810, 870, 879, 880 heartwood, 2, 4, 20, 54, 114, 213, 217, 218, 221, 232, 235, 264, 281, 366, 385, 388, 391-393, 400, 402, 410-413, 435, 438, 796, 805, 873 normal, 2, 12, 22, 40, 43, 5 1, 55, 61, 116, 142, 168, 658 reaction, 2, 22, 36, 51, 52, 63 sapwood, 2, 4, 20, 32, 54, 114, 213, 232, 244, 28 1, 341, 385, 388, 391, 393, 400, 402, 410, 430-432 softwood, 1-4, 17-20, 26, 28, 3 I , 5 I , 56, 59, 63, 65, 68, 71, 75, 92, 108,116, 144, 156,165,175, 214, 246, 252, 274, 284, 300, 319. 337, 358. 394, 425, 467, 490, 532, 573, 601, 850, 860, 880 tension, 2 , 9, 12, 22-24, 36, 40, 44, 51, 54, 55, 60, 445

Index Wood chips, 251, 325, 362, 432. 548, 694, 696, 850, 855, 860, 869, 870 Wood chromophores, 860, 862, 863, 867 (see also Chromophores) Wood industry integration, 850, 858 Wood preservation, 795, 799, 804 pressure processes, 795 supercritical-fluid process, 796 vapor-phase process, 796 Wood preservatives, 795 copper naphthenate, 802. oil-borne, 797 creosote, 797 pentachlorophenol, 799 tributylin oxide, 799 oxide copper, 802 Timbor, 802 water-borne, 799 ammoniacal copper zinc arsenate, 800 chromated copper arsenate, 800 Wood-inorganic composites, 782, 784, 788, 789, 792 Wood-methyl methacrylate composite, 575 Xylan, 56-58, 201, 207, 293, 300, 303, 305, 314, 325, 447, 450, 456. 466, 476, 477, 490, 497, 71 I , 873, 882, 883, 889. 894 Xylem, 1-4. 20, 26, 28, 3 1, 46, 66, 72, 175, 176, 242, 244 Xyloglucan, 98, 192, 195, 199, 201, 205, 207, 314 Yellowing, 332, 402, 540, 589, 867-869