Nobuo Shiraishi, Hiromu Kajita and Misato Norimoto (Eds.) - Recent Research On Wood and Wood-Based Materials. Current Japanese Materials Research (1993)

CURRENT JAPANESE MATERIALS RESEARCH
Volume 1
Current Research on Fatigue Cracks
T. TANAKA, M. JONO and K. KOMAI
Volume 2
Statistical Research on Fatigue and Fracture
T. TANAKA, S. NISHIJIMA and M. ICHIKAWA
Volume 3
High Temperature Creep-Fatigue
R. OHTANI, M. OHNAMI and T. INOUE
Volume 4
Localized Corrosion
F. HINE, K. KOMAI and K. YAMAKAWA
Volume 5
Recent Studies on Structural Safety
T. NAKAGAWA, H. ISHIKAWA and A. TSURUI
Volume 6
Fractography
R. KOTERAZAWA, R. EBARA and S. NISHIDA
Volume 7
Computational Plasticity
T. INOUE, H. KITAGAWA and S. SHIMA
Volume 8
Fracture Mechanics
H. OKAMURA and K. OGURA
Volume 9
Soil Improvement
T MISE, K. NISHIDA, M. KAMON and M. MASHIMA
Volume 10
X-Ray Diffraction Studies on the Deformation and Fracture of Solids
K. TANAKA, S. KODAMA and T. GOTO
RECENT RESEARCH ON WOOD
AND
WOOD-BASED MATERIALS
Edited by
N O B U O SHIRAISHI
Kyoto University, Japan
HIROMU KAJITA
Kyoto Prefectural University, Japan
MISATO NORIMOTO
Kyoto University, Japan
Current Japanese Materials ResearchVol 11
ELSEVIER APPLIED SCIENCE
L O N D O N and N E W YORK
ELSEVIER SCIENCE PUBLISHERS L T D
Crown House, Linton Road, Barking, Essex IG11 8JU, England
WITH 33 TABLES A N D 120 I L L U S T R A T I O N S
1993 ELSEVIER SCIENCE PUBLISHERS L T D and T H E SOCIETY O F

MATERIALS SCIENCE, J A P A N
British Library Cataloguing in Publication Data

Recent Research on Wood and Wood-based
Materials. (Current Japanese Materials
Research; Vol. 11)
I. Shiraishi, N. II. Series *
620.12
ISBN 1-85861-004-4
Library of Congress Cataloging-in-Publication Data

Recent research on wood and wood-based materials / edited by N.
Shiraishi, H. Kajita, M. Norimoto.
p. cm.(Current Japanese materials research ; vol. 11)
Includes bibliographical references and index.
ISBN 1-85861-004-4
1. Wood. I. Shiraishi, N o b u o , 1936-00. II. Kajita, H. (Hiromu)
III. Norimoto, M. (Misato) IV. Series.
TA419.R24 1993
620.1'2^dc20 92-42940
CIP
N o responsibility is assumed by the Publisher for any injury and/or damage to persons or
property as a matter of products liability, negligence or otherwise, or from any use or
operation of any methods, products, instructions or ideas contained in the material herein
Special regulations for readers in the USA

This publication has been registered with the Copyright Clearance Center Inc. (CCC),
Salem, Massachusetts. Information can be obtained from the CCC about conditions under
which photocopies of parts of this publication may be made in the USA. All other copyright
questions, including outside the USA, should be referred to the publisher.
All rights reserved. N o part of this publication may be reproduced, stored in a retrieval
system, or transmitted in any form or by any means, electronic, mechanical, photocopying,
recording, or otherwise, without the prior written permission of the publisher.
Photoset in Malta by Interprint Limited

Printed in Great Britain by the University Press, Cambridge
Foreword
The Current Japanese Materials Research (CJMR) series is edited by

the Society of Materials Science, Japan, and published by Elsevier Science
Publishers, UK, aiming at the overseas circulation of current Japanese
achievements in the field of materials science and technology. This
eleventh volume of the series is entitled Recent Research on Wood and
Wood-based Materials. All the papers have been selected by the authors
as representative of the current status and recent developments in wood
and wood-based materials.
Although many international meetings are held each year in a wide
range of specialized fields, it cannot be denied that most research results
in Japan are published only in Japanese and tend, therefore, to be
confined to the domestic audience. The publication of the CJMR series is
an attempt to offer these results to colleagues abroad and thereby
encourage the exchange of knowledge between us. I hope that our efforts
will interest engineers and scientists in different countries and may
contribute to the progress of materials science and technology through-
out the world.
HIROSHI JINNO
President, Society of Materials Science, Japan
V
Preface
Wood is composed of 50-55% cellulose, 15-25% hemicellulose and

20-30% lignin, with small quantities of ash and extractives. All the main
components are high polymers and form an interwoven network in the
cell wall. The minor components are mostly in cell lumina or special
tissues such as resin canals, and are directly or indirectly related to the
physiology of a living tree.
The main components are considered to be directly related to the
fundamental properties of wood, which result not from a simple summation
of the properties of the individual components, but from the integration of
these properties due to the mutual interaction of the components.
Because wood is a multicomponent, hygroscopic, anisotropic, fibrous,
porous, biodegradable and renewable material, it is unique compared to
such other materials as metal, plastic and glass.
In general, we can always find other materials to surpass wood in
individual properties. However, wood possesses moderate values for
almost all properties in good balance. This is the reason why wood is so
familiar to human beings and originates from the fact that it is a natural
material. On the other hand, wood can also be used as a raw material for
chemical processes, and its chemical modification and conversion have
recently been extensively investigated.
In the present volume, emphasis has been placed on a limited number of
recent developments in the field of wood science and technology in Japan.
The editors would like to express their utmost thanks to all the authors
for contributing the technical papers in this volume. We also acknowl-
edge the considerable help of Mr Yoshido Fujiwara, the director of
JSMS, in editing this volume.
N O B U O SHIRAISHI
HIROMU KAJITA
MISATO NORIMOTO
vii
List of Contributors
TAKANORI ARIMA
Department of Forest Products, Faculty of Agriculture, The Tokyo
University, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113, Japan
JUN-ICHI AZUMA
Department of Wood Science and Technology, Faculty of Agriculture,
Kyoto University, Kitashirakawa Oiwake-cho, Sakyo-ku, Kyoto 606-
01, Japan
JOSEPH GRIL
Laboratoire de Mecanique Generate des Milieux Continus, Universite
de Montpellier II, USTL, Place E. Bataillon, 34000 Montpellier Cdx
2, France
TOSHIMITSU HATA
Wood Reasearch Institute, Kyoto University, Uji, Kyoto 611, Japan
YUJI IMAMURA
Wood Research Institute, Kyoto University, Uji, Kyoto 611, Japan
RAZALI ABDUL KADER

Faculty of Forestry, University Pertanian Malaysia, Serdang, Selan-
gor, Malaysia
HIROMU KAJITA
Department of Forestry, Faculty of Agriculture, Kyoto Prefectural
University, Shimogamo, Sakyo-ku, Kyoto 606, Japan
xi
xii LIST OF CONTRIBUTORS
FUMIO KAMIYA
Forestry and Forest Products Research Institute, P.O. Box 16,
Tsukuba Norin Kenkyu Danchi-nai, Ibaraki 305, Japan
SHUICHI KAWAI
KOHEI KOMATSU
JUNSUKE MUKUDAI
Department of Forestry, Faculty of Agriculture, Kyoto Prefectural
University, Shimogamo, Sakyo-ku, Kyoto 606, Japan
TAKASHI NAKAI
MISATO NORIMOTO
KEIZO OKAMURA
01, Japan
SHIRO SAKA
01, Japan
HIKARU SASAKI
MIKIO SHIMADA
LIST OF CONTRIBUTORS xiii
NOBUO SHIRAISHI
Kyoto University, Kitashirakawa, Oiwake-cho, Sakyo-ku, Kyoto 606-
01, Japan
BAMBANG SUBIYANTO
Division of Material Science, R&D Center for Applied Physics, In-
donesian Institute of Sciences, Kompleks Puspiptek, Serpong, Tanger-
ang 13510, Indonesia
MASAHARU SUZUKI
Department of Applied Biological Science and Technology, Tokyo
Noko University, 3-5-8 Saiwai-cho, Fuchu, Tokyo 183, Japan
MUNEZOH TAKAHASHI
NAOHIKO TSUJIMOTO
Central Research Laboratory, Oji Paper Company Ltd, 1-10-6
Shinonome, Koto-ku, Tokyo 135, Japan
QIAN W A N G
M I N G ZHANG
01, Japan
Structure and Chemical Composition of
Wood as a Natural Composite Material
SHIRO SAKA
Department of Wood Science and Technology,

Faculty of Agriculture, Kyoto University,
Kitashirakawa Oiwake-cho, Sakyo-ku, Kyoto 606-01, Japan
ABSTRACT
The architecture of the molecular arrangement and ultrastructure of wood cell walls is
described in relation to the physical and mechanical properties of wood as a natural
composite material. The chemical composition of wood is also summarized to illustrate the
heterogeneity in distribution of the cell wall constituents.
Keywords: wood structure, chemical composition, cellulose, hemicellulose, lignin
INTRODUCTION
Wood is an easily renewable natural resource with adequate hus-

bandry and long-range planning. The principal sources of commercial
timber are softwood (gymnosperm) and hardwood (angiosperm) trees. A
tree of either the softwood or hardwood type consists of the stem,
branches and roots, the stem being the principal source of material for
manufacturing lumber and other wood products.
In its structure, wood is a natural composite material formed from a
chemical complex of cellulose, hemicellulose, lignin and extractives. These
polymeric substances are not uniformly distributed within the wood cell wall
and their concentration changes from one morphological region to another.
In order to use wood judiciously as a material, a basic knowledge of its
structure and chemical composition is essential. The aim of this paper is
therefore to present the current knowledge of the structure of wood and
its chemical composition.
1
2 SHIRO SAKA
WOOD CELLS
As illustrated in Fig. 1, wood has several different types of cell exist-

ing in both softwoods and hardwoods. The types of cell are different
in volumes, with their axes directed longitudinally and transversely. The
anatomy of softwoods is less complex than that of hardwoods.
As listed in Table 1, several types of cells are present in softwoods, with
tracheids and ray parenchyma cells always being found. Figure 2 shows a
scanning electron micrograph of the typical softwood cells. In hard-
V
FIG. 1. Schematic diagram of the major wood cells [ 1 ] . wf, wood fiber; tr,
tracheid; ve, vessel element; ap, axial parenchyma cell; rtr, ray tracheid; rp, ray
parenchyma cell.
TABLE 1
Cells in softwoods
Longitudinal Transverse
Tracheid Ray tracheid

Strand tracheid Ray parenchyma cell
Axial parenchyma cell Ray epithelial cell
Epithelial cell
STRUCTURE AND CHEMICAL COMPOSITION OF WOOD 3
FIG. 2. Scanning electron micrograph of a typical softwood (Pinus densiflora),

showing the radial (R), tangential (T) and transverse (X) surfaces. Note that
tracheids, rays and resin canals are present. (Courtesy of Prof. H. Saiki.)
woods, the principal types of cell are vessel, tracheid, fiber and paren-
chyma, which are more complex than those in softwoods, as shown in
Table 2. Since softwood tracheids and hardwood fibers constitute the
majority of wood cells, they contribute in a major way to the physical and
chemical properties of wood.
TABLE 2
Cells in hardwoods
Longitudinal Transverse
Vessel element Ray parenchyma cell

Tracheid Ray epithelial cell
Vascular tracheid
Vasicentric tracheid
Fiber
Fiber tracheid
Libriform fiber
Axial parenchyma cell
Epithelial cell
4 SHIRO SAKA
CELL WALL ORGANIZATION
The cell wall organization of typical softwood tracheids or hard-

wood fibers constituting the majority of wood cells is described in
Fig. 3, which has been modified by Harada [3] from the model by
Wardrop and Harada [ 4 ] . Basically, the cell wall consists of the
primary (P) and secondary (S) wall layers. The P layer is formed dur-
ing the surface growth of the cell wall, while the S layer is formed
during the thickening of the cell wall. As shown in Fig. 3, the texture of
cellulose microfibrils in the P layer is loosely aggregated, and is oriented
more or less axially to the cell axis on the outer surface ( P ) and G
transversely on the inner surface (P^. The S layer inside the P wall
is composed of three sublayers termed S S and S , based on dif-
1 ? 2 3
ferences in the microfibrillar orientation. The Si layer is a flat helix

and crossed fibrillar structure with alternating S- and Z-helices of
microfibrillar orientation, whereas the S layer is the thickest within the
2
S wall and comprises a steep Z-helix with a high degree of parallelism

in the microfibrils. The S layer is a thin layer with a flat S-helix that
3
is loosely textured [ 5 ] . Although the S layer exhibits steep helices in

2
the microfibrillar orientation, there exist transition lamellae on its inner

and outer surfaces. The microfibrils in these lamellae are gradually
changed in angle between S and S ( S ) and between S and S ( S )
x 2 1 2 2 3 23
FIG. 3. Schematic diagram of the microfibrillar orientation in the wood cell wall
from softwood tracheids and hardwood fibers: P and Pi are the external and
0
internal surfaces of the primary wall: S and S 3 are the intermediate layers
1 2 2
between S and S , and between S and S , respectively [3].

x 2 2 3
[6]. However, the change in microfibrillar angle is more abrupt in S 3 2
than in S [ 7 ] .
1 2
In addition to these wall layers, a layer called the middle lamella (ML)
or intercellular layer (I) is located between the P walls of adjacent cells.
Since it is difficult to distinguish ML from the two P walls on either side,
the term compound middle lamella (CML) is generally used to designate
the combination of ML with the two adjacent P walls (Fig. 4).
FIG. 4. Transmission electron micrograph of an ultra-thin transverse section

from white birch (Betula papyrifera Marsh.), showing different morphological
regions in the fiber (F) and vessel (V); middle lamella (ML) and different layers of
the secondary wall (Si, S and S ). Primary wall is not evident.
2 3
Such wood cells are composed of three major components, cellulose,

hemicellulose and lignin (Table 3). Cellulose is the structural framework
in the form of cellulose microfibrils, while hemicellulose is the matrix
substance present between the cellulose microfibrils. Lignin, on the other
hand, is the encrusting substance binding the wood cells together and
giving rigidity to the cell wall [ 7 ] .
In general, the S layer increases with increasing wall thickness,
2
whereas S and S remain fairly constant. Because of its greater thickness,

t 2
6 SHIRO SAKA
TABLE 3
Comparison between the chemical composition of hard-
woods and softwoods [2]
Softwoods Hardwoods
Cellulose 40-50 45-50

Hemicellulose
Xylan 6-10 15-20
Glucomannan 5-10 0
Lignin 27-30 20-25
Resin 2-5 2
Ash 0-2-0-5 0-5
the S layer is largely responsible for the physical and mechanical

2
properties of the cell walls. Therefore, it is of great interest to know the

relationship between the microfibrillar orientation in the S layer and its
2
cell wall properties.

As can be seen in Fig. 3, the cellulose microfibrils in the tracheid S 2
layer are oriented nearly parallel to the fiber axis [6, 7 ] . However, some
variations in the fibrillar orientation exist. The closest parallelism to the
cell axis has been reported for the gelatinous (G) layer in tension wood
[8], which is made up of the abnormal hardwood xylem tissues formed in
a leaning stem or branch. Saiki [ 9 ] has reported for several softwood
species that the microfibrillar angles in the latewood tracheid S layer are
2
generally lower than those in the earlywood. In contrast, compression

wood, which is made up of the abnormal softwood xylem tissue formed
on the lower or compression side of a leaning stem or branch has a larger
microfibrillar angle in the S layer [10, 11]. This evidence from the
2
literature suggests that the microfibrillar angles increase in the order:
Tension wood Latewood Earlywood Compression wood

G-layer tracheid (S ) 2 tracheid (S )
2 tracheid (S ) 2
Interestingly, this trend is parallel to the trend in lignin concentration; the

tension wood G-layer is known to have little or no lignin [4, 12], whereas
compression wood has an abnormally high lignin content in the S layer 2
compared to normal softwood [13, 14]. The tracheid S layer in latewood

2
has a lower lignin content than that in earlywood [13, 15-18], except for
juvenile wood [16]. Furthermore, Nanko et al. [19] have found in the
reaction phloem fiber that the unlignified layer had microfibrils oriented
parallel to the fiber axis, while the microfibrils in the lignified layer were
oriented at an angle to the fiber axis.
Figure 5 shows the relationship for softwoods between the lignin

content and microfibrillar angle (6) in the tracheid S layer determined2
from the degree of spreading in the equatorial diffraction arcs of X-ray

fiber diagrams. Since about 80% of the total lignin in softwoods is in the
secondary wall of the tracheids [20], and the S layer is the major portion
2
of the tracheid secondary wall, the whole lignin content of wood could be
expected to be closely correlated to the lignin concentration in the S 2
layer of the tracheid. Thus, Fig. 5 indicates that the lignin concentration
in the S layer increases with increasing microfibrillar angle of the
2
tracheid S layer [21].

2
content (a)
| 30
25
20
0 10 20 30 40 50
Microfibrillar angle (e)
FIG. 5. Relationship between the microfibrillar angle (9) in the tracheid S layer
2
and the lignin content of wood [21].
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. 3, the three major chemical constituents of wood
would mutually interact and strengthen each other to make up a natural
super-composite material. Figure 6 shows such an ultrastructural ar-
rangement of the cellulose microfibrils, hemicellulose and lignin in the
wood cell wall as proposed by Harada and Cote [22]; 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.
8 SHIRO SAKA
Mf
FIG. 6. Schematic diagram of the

ultrastructural arrangement of a cellu-
lose microfibril (Mf), hemicellulose (H)
and lignin (L) in the wood cell wall
[22].
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) [23, 24]. A recent
study by Shiraishi and co-workers [25, 26] has indicated that LCC must
be a compatibilizer-like substance localized at the interface between
macromolecules of lignin with a hydrophobic nature and hydrophilic
carbohydrates. The model in Fig. 6 suggests the presence of such LCC
molecules. In addition, it is worth while to note that they found a new
function of LCC to enhance the physical and mechanical properties of
wood by its presence in a small amount.
STRUCTURE OF THE CELLULOSE MICROFIBRILS
On cellulose microfibrils themselves, many investigations have been

carried out by electron microscopy, using the negative staining technique
[27, 28] or diffraction contrast in the bright-field mode [29]. Consequent-
ly, the size of the microfibrils has been found to be different in different
cellulose materials [28, 30, 31], being about 2-5 nm for wood hollocel-
lulose, about 3 nm for the gelatinous layer of hardwood tension wood
fibers, about 5 nm for whatman cellulose, and about 20 nm for valonia
[28, 31].
Of the studies on the crystalline ultrastructure of cellulose microfibrils,
Sugiyama et al [32-35] have done pioneering work by means of
high-resolution electron microscopy of ultra-thin sections. As shown in
Fig. 7, lattice lines of 0-60, 0-54 and 0-39 nm were successfully visualized in
Valonia macrophysa for the first time. The lattice spacing of these lines has
FIG. 7. Transmission electron micrographs of a cross-sectional view of the

cellulose microfibrils in Valonia macrophysa, showing lattice spacings of 0-60, 054
and 0-39 nm, and optical diffraction patterns of the corresponding micrographs.
(Courtesy of the late Prof. H. Harada.)
also been measured by electron and optical diffraction patterns. Since one
microfibril can be seen as a single crystal, the existence of 3-5 nm
elementary fibrils has been ruled out [33]. Furthermore, from obser-
vations of the lattice lines along the microfibrils, it has been concluded that
there is no periodicity in the longitudinal direction [32]. Unfortunately,
the lattice image of cellulose microfibrils in wood has not yet been taken
successfully due to the low crystallinity and small size of the crystallite.
DISTRIBUTION OF CHEMICAL CONSTITUENTS IN THE

CELL WALL
Three major cell wall components, cellulose, hemicellulose and lignin,

are not uniformly distributed in wood cells, and their concentration
changes from one morphological region to another. Therefore, knowledge
about the distribution of chemical constituents in the cell wall is of great
importance for understanding the properties of wood as a natural
composite material. In the next sections, the recent progress in inves-
tigations on the chemical composition of wood is presented.
10 SHIRO SAKA
Lignin
To investigate the distribution of lignin in wood, a number of reliable
methods can be used. As representative techniques, potassium permanga-
nate staining has been qualitatively used with electron microscopy [35],
while studies have also been reported by electron microscopy of the lignin
skeleton left after removing the constituent polysaccharides with hydro-
fluoric acid [36] or brown rot fungus [37].
In a quantitative assay of the lignin distribution, ultraviolet (UV) micro-
scopy by Goring and co-workers [39-41] has provided good results with
thin sections of wood. Saka et al [42-44] have developed a new method
based on specific bromination for lignin in a nonaqueous system, the
bromine concentration in each morphological region of wood then being
determined by transmission electron microscopy (TEM) or scanning
electron microscopy (SEM) coupled with energy-dispersive X-ray analysis
(EDXA). When the lignin reactivity toward bromination is known, the
distribution of lignin can be determined in various morphological regions
of wood. Figure 8 shows a direct comparison made between the two
1.00
Latewood Earlywood
i
Lignin concentration (g/g)
0.80
v -----
MIDDLE
LAMELLA (cc)
0.60 oo o
Q
O uv
0.40 EDXA
BOUNDARY
0.20
SECONDARY
WALL
I I
15 10 5 5 10
Cell number
FIG. 8. Variation of lignin concentration across the earlywood/latewood
boundary by UV microscopy ( O ) and TEM-EDXA ().
techniques of UV microscopy and EDXA with bromination [44] across the

earlywood/latewood boundary of black spruce (Picea mariana Mill.). It is
apparent that the results obtained by the two methods are in good
agreement with each other.
Table 4 shows the distribution of lignin in loblolly pine (Pinus taeda L.)
tracheids as determined by bromination with S E M - E D X A [45]. One of
the advantages of this technique compared with UV microscopy is the
ability to study the S S and S layers in the secondary wall as separate
1 ? 2 3
entities, such resolution often being difficult with U V microscopy. It is

interesting to note that the lignin concentration in the S layer is lower 2
than that in either the S or S layer. The line profile of the bromine
x 3
X-ray in Fig. 9 shows such differences clearly. Fukazawa and Imagawa
TABLE 4
Distribution of lignin in loblolly pine tracheids as determined by bromination
with SEM-EDXA
Wood Morphological Tissue Lignin Lignin

region volume (%) (% of total) concentration (g/g)
Early wood S
t 13 12 025
5
2 60 44 0-20
5
3 9 9 028
ML 12 21 0-49
a
ML cc 6 14 064
Latewood Sj 6 6 0-23
5
2 80 63 018
5
3 5 6 025
ML 6 14 051
a
ML cc 3 11 078
a
Middle lamella at the cell corner
[46] have also reported the similar finding of high U V absorbance near
the lumen/wall interface for juvenile wood tracheids of Japanese fir (Abies
sachalinensis Fr. Schm.).
Hardwood lignins consist mainly of guaiacyl and syringyl residues, and
their ratio seems to change from one morphological region to another.
Although Fergus and Goring [47, 48] have attempted to determine the
distribution of lignin in white birch (Betula papyrifera Marsh.) by U V
spectral analysis, the syringyl and guaiacyl residues have markedly
different UV absorptivity. Therefore, it is essential to know the exact ratio
before the lignin concentration in a particular morphological region can
be computed from the U V absorption.
12 SHIRO SAKA
FIG. 9. Scanning electron micrograph (a) of brominated latewood tracheids in

loblolly pine (05 fim 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 walls.
Saka et al [49, 50] have developed a new method for computing the ratio
of guaiacyl and syringyl residues in various morphological regions by
combining UV microscopy with bromination-EDXA (UV-EDXA), which
can be used to determine the lignin distribution in hardwoods. Shown in
Table 5 is the ratio of guaiacyl to syringyl residues in various morphological
regions of white birch wood as determined by the UV-EDXA technique
[49]. The results indicate that the fiber secondary wall (S ) contains pre-
2
dominantly syringyl residues, whereas the vessel secondary wall (S ) con-2
sists mostly of guaiacyl residues. For the ray parenchyma cell, about equal
proportions of guaiacyl and syringyl residues were contained in lignin. For
the cell corner middle lamella (ML ), 80 to 100% of the lignin was found to
CC
be guaiacyl residue, with the remaining 0 to 20% being syringyl residue. It

is apparent from Table 5 that, in hardwoods, the ratio of guaiacyl to
syringyl residues in lignin varies in different morphological regions.
The heterogeneity of the formation of lignin in hardwoods can also be
studied in the autoradiographic technique described by Terashima et al
[51]. These authors administered H-labeled guaiacyl and syringyl model
3
compounds to magnolia shoots and determined their location in the

growing cell wall by micrd-autoradiography. They found that the vessel
wall, cell corner and compound middle lamella were lignified by the
deposition of guaiacyl-type lignin, and that the fiber wall was composed
of syringyl-guaiacyl lignin. These results are in good agreement with the
findings by the UV-EDXA technique (Table 5).
TABLE 5
Ratio of guaiacyl and syringyl residues and the distribution of lignin in white birch
Element Morphological Tissue Guaiacyl-syringyl Lignin
region volume (%) ratio concentration (g/g)
Fiber Si 11-4 014
s 2
58-4 12:88 014
s 3
3-5 012
ML 5-2 0-36
a
ML (f/O cc
2-4 91:9 0-45
Vessel s, 1-6 0-26
s 2
4-3 88:12 0-26
S 3
2-3 0-27
ML 0-8 0-40
1
MLJf/v) -0 80:20 0-58
Ray S 80 49:51 012
parenchyma ML 20 0-38
ML (f/r)<cc -0 100:0 0-47
ML (r/r)"cc -0 88:12 0-41
STRUCTURE AND CHEMICAL COMPOSITION OF WOOD
Fiber/fiber.
b
Fiber/vessel.
c
Fiber/ray.
d
Ray/ray.
13
14 SHIRO SAKA
Also shown in Table 5 is the distribution of lignin in white birch

wood as determined by UV-EDXA [49]. For the fiber secondary wall,
the lignin concentration in the S layer is slightly lower than that in
3
either the S or S layer. However, its difference is so small that the

1 2
lignin may be considered to be distributed uniformly across the second-

ary wall. The vessel walls also reveal a uniform distribution of lignin,
although the concentration is about 1-9 times higher than that of the
ray parenchyma cells. The cell corner middle lamella (ML ) associated
CC
with fibers and vessels has the highest lignin concentration. In spite of
sufficient spatial resolution for analysis by the EDXA system, the
middle lamella between the cell corner areas (ML) was 10 to 30% lower
in concentration than the cell corner middle lamella (ML ). It is of
CC
interest to note that the lignin concentration in the middle lamella

regions of hardwoods is lower than that of softwoods, as can be seen in
Tables 4 and 5. Such a difference has been confirmed by a micro-Klason
lignin analysis of tissue fractions of ML from black spruce and white
birch [50, 51].
It is known that the ratio of syringyl to guaiacyl residues in hard-
wood lignin changes among wood species. However, at present, nothing
is known about the relationship between its distribution and the cell
wall organization of cellulose microfibrils illustrated in Fig. 5 for soft-
woods. It is, therefore, a future task to study the role of the heterogene-
ity of hardwood lignin in the mechanical and physical properties of
hardwoods.
Cellulose and Hemicellulose

The distribution of cellulose is probably the easiest to study in the
various morphological regions of wood. Although the orientation of the
cellulose microfibrils is quite different in the various cell wall layers,
cellulose is quite evenly distributed throughout the secondary wall. In the
primary wall, however, the microfibrils are rather loosely distributed (Fig.
3), so that the concentration of cellulose in the primary wall may be lower
than that in the secondary wall.
Unlike cellulose, a study of the distribution of hemicellulose is
difficult, because histochemical techniques are generally nonspecific
and frequently unreliable. The distribution of polysaccharides has, how-
ever, been examined in holocellulose skeletons after removing the
lignin with acid chlorite [53, 54]. Although the presence of some
residual lignin and partial removal of polysaccharides may obscure
the data, the overall results obtained by this method are in good agree-
ment with the holocellulose distribution inferred from the lignin

skeleton.
For a quantitative determination of the polysaccharide distribution,
the microdissection technique has often been used. Meier [55, 56] has
adopted this technique for woods at different stages of development that
were microscopically distinguished, isolated and subsequently subjected
to a microanalysis for sugar residues. From a knowledge of the chemical
composition of the different polysaccharides in wood, the content of
polysaccharides in various morphological regions could be calculated.
Although some doubt exists about the additional deposition of polysac-
charides during the later stages of secondary wall thickening, the tech-
nique developed by Meier still remains applicable. With this technique,
Takabe [57] has determined the distribution of polysaccharides across
100*
80
Sugar composition
60
40
20-
P ST S2 S3<W)
Gl ucuronoa rabi noxyl an
Galactoglucomannan
Cellulose
FIG. 10. Distribution of polysaccharides across the cell wall in tracheids from
Cryptomeria japonica [57].
16 SHIRO SAKA
the wood cell wall in tracheids from Cryptomeria japonica (Fig. 10).
Interestingly, cellulose is rich in the middle of the S layer, whereas the
2
hemicelluloses of glucuronoarabinoxylan and galactoglucomannan are

abundant in S and the outer part of the S and S layers. The warty
l9 2 3
layer (W) is composed mainly of galactoglucomannan.
Inorganic Constituents
As a general rule, elements other than carbon, hydrogen, oxygen and
nitrogen make up between 0 1 and 0-5% of the weight of wood [58, 59].
This proportion, although small, contains a wide variety of elements,
numbering over thirty [60]. Some of the inorganic elements present in
wood are essential for wood growth, whereas others are not necessarily
required.
Although a fair amount of information is available on the inorganic
constituents of wood [60-62], little has been published on their mor-
phological distribution in the cell wall. Saka and Goring [63] have
studied the distribution of inorganic constituents from the pith to the
outer ring of black spruce by the TEM-EDXA technique. This technique
provides a simple and powerful tool for determining any element above
neon in the periodic table.
Figure 11 shows the seven morphological regions in tracheids, ray
tracheids and ray parenchyma cells that have been investigated. Fifteen
different elements above neon in the periodic table were detected,
including Na, Mg, Al, S, CI, K, Ca, Cr, Fe, Ni, Cu, Zn and Pb.
Although the secondary walls usually contain detectable amounts of
only four elements (sulfur, chlorine, potassium and calcium), almost
all the elements are localized and concentrated in the torus and half-
bordered pit membrane regions (Fig. 12). This result may be related to a
high concentration of pectic materials in the pit membrane regions.
The total content of inorganic constituents decreases in the order of
torus (2%) > half-bordered pit membrane (1%)> middle lamella
(0-4%) > ray parenchyma cell wall (0*3%)> tracheid secondary wall (01
to 015%).
In addition, the total content of inorganic constituents is higher in
earlywood than latewood for all the morphological regions studied. This
finding supports the concept [64] that earlywood tracheids having large
lumens and abundant pits play an important role in water conduction,
while thick-walled latewood tracheids with far fewer pits may act as
physical or mechanical support for the wood.
FIG. 11. Transmission electron micrographs of a cross section of black spruce

showing the seven different morphological regions. The dark circle indicates the
location of the analysis, and its diameter corresponds to the spatial resolution of
analysis (400 nm). S , S , S , S layers in tracheid (T), ray tracheid (RT) and ray
T RT R 2
parenchyma (R); T , T , tori in T and RT; C C , cell corner middle lamella; M, a

T RT
half-bordered pit membrane between T and R.

18 SHIRO SAKA
FIG. 12. X-ray spectra from the tracheid secondary wall and tracheid torus.
CONCLUDING REMARKS
Recent results from investigations on the structure and chemical

composition of wood have been summarized to explain the nature of
wood as a natural composite material. However, since both fields have
been studied independently, these results are insufficient for a full under-
standing of the relationship between the cell wall organization and the
distribution of chemical constituents in wood. Extensive studies on the
heterogeneity of chemical constituents in relation to the ultrastructure of
wood will provide a better picture of wood as a natural composite
material.
REFERENCES
[1] K. Shimaji, Mokuzai-no Soshiki (1976) Morikita Shuppan, Tokyo.

[2] T. Koshijima, H. Sugihara, H. Hamada, M. Fukuyama and G. Fuse,
1
Kisomokuzai kougaku (1968) Futaba-shoten.

[3] H. Harada, Mokuzai Gakkaishi, 30, 513 (1984).
[4] A.B. Wardrop and H. Harada, J. Exp. BoL, 16, 356 (1965).
[5] R.D. Preston, The Physical Biology of Plant Cell Walls (1974) Chapman and
Hall, London.
[6] H. Harada, Y. Miyazaki and T. Wakashima, Bull Government Forest Exp.

Station, 104, 1 (1958).
[7] M. Fujita and H. Harada, Wood and Cellulosic Chemistry (1990) Marcel
Dekker, New York.
[8] A.B. Wardrop, The Formation of Wood in Forest Trees (1964) Academic
Press, New York.
[9] H. Saiki, Memoirs College Agri., Kyoto Univ., 96, 47 (1970).
[10] W.A. Cote, Jr., T.E. Timell and R.A. Zabel, Holz Roh-Werkt., 24, 432 (1966).
[11] M. Fujita, H. Saiki and H. Harada, Mokuzai Gakkaishi, 24, 158 (1978).
[12] Y. Musha, Ph.D. Thesis, McGill Univ., Montreal, Canada (1973).
[13] J.R. Wood and D.A.I. Goring, Pulp Paper Mag. Can., 72, T95 (1971).
[14] K. Fukazawa, Res. Bull College Exp. For., liokkaido Univ., Vol. 31, 87
(1974).
[15] B.J. Fergus, A.R. Procter, J.A.N. Scott and D.A.I. Goring, Wood Sci.
Technol, 3, 117 (1969).
[16] K. Fukazawa and H. Imagawa, Wood Sci. Technol, 15, 45 (1981).
[17] S. Saka and R.J. Thomas, Wood Sci. Technol, 16, 167 (1982).
[18] S. Saka, P. Whiting, K. Fukazawa and D.A.I. Goring, Wood Sci. Technol,
16 269 (1982).
[19] H. Nanko, H. Saiki and H. Harada, Mokuzai Gakkaishi, 28, 202 (1982).
[20] B.J. Fergus, A.R. Procter, J.A.N. Scott and D.A.I. Goring, Wood Sci.
Technol, 3, 117 (1969).
[21] S. Saka and M. Tsuji, Cellulose Chem. Technol, 21, 225 (1987).
[22] H. Harada and W.A. Cote, Jr., Biosynthesis and Biodegradation of Wood
Components (1985) Academic Press, New York.
[23] T. Koshijima, Wood Res. Tech. Notes, 19, 11 (1984).
[24] T. Watanabe, Ph.D. Thesis, Kyoto Univ. (1989).
[25] S. Takase, N. Shiraishi and M. Takahama, Wood Processing and Utilization
(1989) John Wiley & Sons, England.
[26] S. Shiraishi, Kobunshi Kako, 38, 338 (1989).
[27] T. Goto, H. Harada and H. Saiki, Mokuzai Gakkaishi, 19, 463 (1973).
[28] T. Goto, H. Harada and H. Saiki, Mokuzai Gakkaishi, 21, 537 (1975).
[29] J.F. Revol, Carbohydrate Polym., 2, 123 (1982).
[30] V. Balashov and R.D. Preston, Nature, 176, 64 (1955).
[31] H. Harada and T. Goto, Cellulose and Other Natural Polymer Systems
(1982) Plenum, New York.
[32] J. Sugiyama, H. Harada, Y. Fujiyoshi and N. Uyeda, Mokuzai Gakkaishi,
30, 98 (1984).
[33] J. Sugiyama, H. Harada, Y. Fujiyoshi and N. Uyeda, Planta, 166, 161 (1985).
[34] J. Sugiyama, H. Harada, Y. Fujiyoshi and N. Uyeda, Mokuzai Gakkaishi,
31, 61 (1985).
[35] J. Sugiyama, H. Harada, Y. Fujiyoshi and N. Uyeda, Denshikenbikyo, 20,
143 (1985).
[36] D.E. Bland, R.C. Foster and A.F. Logan, Holzforsch., 25, 137 (1971).
[37] A.L.K. Bentum, W.A. Cote, Jr., A.C. Day and T.E. Timell, Wood Sci.,
Technol, 3, 218 (1969).
[38] H. Meier, Holz Roh-Werkst., 13, 323 (1955).
20 SHIRO SAKA
[39] J.A.N. Scott, A.R. Procter, B.J. Fergus and D.A.I. Goring, Wood Sci.
TechnoU 3, 73 (1969).
[40] J.A.N. Scott, A.R. Procter, B.J. Fergus and D.A.I. Goring, Wood Sci.
TechnoU 4, 237 (1970).
[41] J.R. Wood and D.A.I. Goring, J. Microscopy, 100, 105 (1974).
[42] S. Saka, R.J. Thomas and J.S. Gratzl, Tappi, 6 1 , 73 (1978).
[43] S. Saka and R J . Thomas, Wood Sci. TechnoU 16, 1 (1982).
[44] S. Saka, P. Whiting, K. Fukazawa and D.A.I. Goring, Wood Sci. TechnoU
16, 269 (1982).
[45] S. Saka and R.J. Thomas, Wood Sci. TechnoU 16, 167 (1982).
[46] K. Fukazawa and H. Imagawa, Wood Sci. TechnoU 15, 45 (1981).
[47] BJ. Fergus and D.A.I. Goring, Holzforsch., 24, 118 (1970).
[48] B.J. Fergus and D.A.I. Goring, Holzforsch., 24, 113 (1970).
[49] S. Saka and D.A.I. Goring, Holzforsch., 42, 149 (1988).
[50] S. Saka, S. Hosoya, F.GT. St-Germain and D.A.I. Goring, Holzforsch., 42,
79 (1988).
[51] N. Terashima, K. Fukushima and K. Takabe, Holzforsch., 40, 101 (1986).
[52] P. Whiting, B.D. Favis, F.G.T. St-Germain and D.A.I. Goring, J. Wood
Chem. TechnoU 1, 29 (1981).
[53] W.A. Cote, Jr., N.P. Kutscha, B.W. Simson and T.E. Timell, Tappi, 5 1 , 33
(1968).
[54] T. Fujii, H. Harada and H. Saiki, Mokuzai Gakkaishi, 27, 149 (1981).
[55] H. Meier, J. Pol. Sci., 5 1 , 11 (1961).
[56] H. Meier and K.C.B. Wilkie, Holzforsch., 13, 177 (1959).
[57] K. Takabe, Ph.D. Thesis, Dept. Wood Sci. Technol., Kyoto Univ. (1984).
[58] A.J. Panshin and C. de Zeeuw, Textbook of Wood Technology, Vol. 1, 3rd
ed. (1970) McGraw-Hill, New York.
[59] B.L. Browning, The Chemistry of Wood (1963) John Wiley & Sons, New
York, Interscience Publishers.
[60] E.L. Ellis, Cellular Ultrastructure of Woody Plants (1965) ed. by W.A. Cote,
Jr., Syracuse Univ. Press, New York.
[61] E.L. Ellis, For. Prod. J., 12, 271 (1962).
[62] W.L. Galligan, H. Stern and P. Hohenschuh, For. Prod. J., 15, 185 (1965).
[63] S. Saka and D.A.I. Goring, Mokuzai Gakkaishi, 29, 648 (1983).
[64] H.P. Brownt, Wood Chemistry, Vol. 1, 2nd ed. (1952) ed. by L.E. Wise and
E.C. John, Reinhold Publishing Corp., New York.
Oriented Strand and Fiber Boards
HIROMU KAJITA and JUNSUKE MUKUDAI
Department of Forestry, Faculty of Agriculture,

Kyoto Prefectural University, Shimogamo, Sakyo-ku, Kyoto 606, Japan
ABSTRACT
Oriented mat-forming methods that employ mechanical and electrostatic orientation
were studied. The effects of some variables on the degree of particle alignment and on the
physical and mechanical properties of oriented boards were investigated. A production
system applying the electrostatic method, in which electrodes are located only on the reverse
side of the forming belt, is proposed.
Keywords: oriented strand board, oriented fiber board, mechanical orientation, electrostatic
orientation, alignment angle
INTRODUCTION
Wood is an anisotropic material having better mechanical properties

and dimensional stability in its longitudinal direction. In manufacturing
reconstituted wood products, the anisotropic nature of the raw material
can be enhanced by directionally aligning the elements. Oriented fiber
and/or particleboard which has been developed recently for structural use
is one of the rational applications of this characteristic of wood.
There are two methods for aligning the particles, one using electrical
and the other mechanical means.
1. Mechanical method
The particles were aligned by a laboratory-size mechanical orienter, in
which many reciprocating plates were set parallel to each other at regular
intervals [1, 2]. The stroke and the frequency of the reciprocating plates
were 13 cm and 80 times per minute, respectively. Wood particles were
21
22 HIROMU KAJITA AND JUNSUKE MUKUDAI
dropped onto the reciprocating plates and passed through the spaces
between them (10 mm), forming a mat on a caul. The degree of particle
alignment could be controlled by changing the free-fall distance between
the bottom of the reciprocating plates and the top of the particle mat.
This mechanical orienter could align some particles, especially those that
were long or in slender flakes [ 1 ] , but small particles and fines were more
difficult to align mechanically, the electrostatic method being the most
suitable for these. Figure 1 shows typical examples of the distribution of
alignment angle (the angle between the cardinal direction of alignment
and the longer axis of the particle). With increasing free-fall distance, the
distribution curve gradually flattens and approaches the random dis-
tribution curve. These results show the same trends as those reported by
Suzuki and Sekino [3] and by Kawei et al. [4]. Figure 2 shows the effect
of free-fall distance on the degree of alignment with flake- and strand-type
particles. The average alignment angle (AAA) of the strand-type particles
increased with increasing free-fall distance, and the value is larger than
that for the flake-type particles. It seems that the spacing distance
72.5 %
O 52.9 %
s
Free-fall
40 :
distance
2 cm
o 4 cm
6 cm
30
8 cm
* Random
(
Frequency
Plate spacing: 10 mm
20
10
30 60 90
Alignment angle (degrees)
FIG. 1. Distribution of alignment angles for flakes of 0-5 x 10 x 50 mm.
ORIENTED STRAND AND FIBER BOARDS 23
(899J6ep) a|6ue iu9uiu6;|e a6ejeAv 50
40
lauan, sliver^
Cl.8-2.3-19.2mnn,,
30
"^sugi , semi-strand
( 0 . 5 * 2 . 3 "26.4mm) C
20'
* ^ s u g i , long tlakc
( 0 . 5 - 1 0 5 0 mm )
10<
Plate spacing: 10mm
2 4 6 8 10
Free-fall distance (cm)
FIG. 2. Relationship between the average alignment angle and free-fall distance
for strand- and flake-type particles of sugi (Cryptomeria japonica D. Don) and
lauan (Shorea spp.).
between the plates was too wide for the strand-type particles. Therefore,
one way of improving the alignment of these particles is to adjust the
spacing distance to a value closer to the particle width [1, 5]. A free-fall
distance of less than 3 cm did not affect AAA of the flakes, the value being
about 10 degrees. On the other hand, with a free-fall distance of over
3 cm, AAA increased with increasing free-fall distance. This is because
particles which had been oriented while between the plates became
disordered during free falling between the bottom of the plates and the
top of the particle mat.
Figures 3 and 4 show the relationships between the bending properties
(modulus of rupture (MOR) and modulus of elasticity (MOE)) and board
density and particle alignment. These figures show parabolic curves for
the bending properties against board density, both MOR and M O E
increasing with increasing board density [ 2 ] .
By increasing the board density from 0-50 to 0-70 g/cm , MOR of the 3
single-layer board was approximately doubled in the direction parallel to

8
Single-layer board
Modulus of elasticity (GPa)
0
0.6 0.7 0.8
Board d e n s i t y (gem )
- 3
FIG. 3. Relationships between the modulus of elasticity, board density and

average alignment angle for a single-layer board of sugi (Cryptomeria japonica D.
Don).
the oriented axis, in contrast to an increase of 2-2 times in the random

boards [ 2 ] . On the other hand, MOR of the three-layer board was
increased by about 65% in the parallel direction, in contrast to an
increase of 80% in the random boards [ 2 ] . The relationships between
M O E and board density and between MOR and board density were
almost the same.
As expected, orienting the particles produced an increase of MOR in
the oriented direction of the boards. When AAA was reduced from 44 to
13 degrees, MOR in the oriented direction of the single-layer boards was
about twice that of the random boards, while the value in the orthogonal
direction decreased to half that of the random boards [ 2 ] . The bending
properties of the three-layer boards were greater than those of the
single-layer boards. These differences may be explained by a difference in
the slenderness ratio (length/thickness) of the particles and in the density
gradient existing across the thickness of the boards.
From Fig. 4 it can be seen that the MOR value (40MPa) of the
random board with a density of 0-70 g/cm corresponds to that in the 3
80
L Three-layer board
60
(MPa)
Modulus of rupture
40
3
O
= 20
J- J- J-
0.5 0.6 0.7
Board density (gem )
- 3
FIG. 4. Relationships between the modulus of rupture, board density and

average alignment angle for a three-layer board of sugi (Cryptomeria japonica D.
Don).
alignment direction of the aligned board with a density of 0-50g/cm and 3
with AAA of 18 degrees. This means that about 30% of the raw material
weight of manufactured boards can be saved.
Figure 5 shows how a decrease in AAA affects the ratio of parallelism to
orthogonality (||/) for bending properties. Board density had no signifi-
cant effect on the ratio, but a marked increase in the value of the
| | / l ratio for bending properties with decreasing AAA is clearly depicted.
Especially, the tendency for an increase in the ||/J_ ratio of the three-layer
boards is much greater than that of the single-layer boards. The pattern of
response for MOE of the single-layer boards is similar to that for MOR. On
the other hand, it seems that the ||/_L ratio for MOR of the three-layer
boards aligned to less than 20 degrees is greater than that for M O E ,
although there are not enough data to explain the results at this stage.
The internal bond-strength (IB) and thickness swelling (TS) measured
after 24 hours of soaking were not significantly different between the
random and oriented particleboard [ 2 ] , the effect of board density on IB
4
Single-layer board
2
!
i
MOE///MOE^
LE-L/LE//
0 10 20 30 40
Average alignment angle (degrees)
FIG. 5. Anisotropy of oriented particleboards.
being greater than that of AAA. TS is due to the release of residual

compressive stresses imparted to the board during hot-pressing, in
addition to normal swelling of the component wood. The compaction
ratio (board density divided by wood density) increases as board density
increases, the compaction ratio for a 0 5 and 0-8g/cm board density 3
being estimated as 1-3 and 2 1 , respectively. It seems that TS with water

immersion increases with increasing compaction ratio. As TS of particle-
boards is not markedly improved by particle alignment, it is necessary to
find a more effective method.
Linear stability is affected substantially by particle alignment. The
linear expansion (LE) values measured after a range of conditioning
procedures at 50 to 90% RH is small in the direction parallel to particle
alignment, and large in the orthogonal direction, as would be expected.
In particular, with AAA of 13 degrees, the value in the parallel direction
for a single-layer board decreased by up to 50% of the value for random
boards, while the value in the orthogonal direction was approximately
ORIENTED ^TRAND AND FIBER BOARDS 27
twice as large as the value for random boards [ 2 ] . LE in the two

directions approached the value for random boards with increasing AAA.
The effects of particle alignment on the ratio of _L/|| values for LE are
similar to the effects on bending properties (Fig. 5). The _L/|| ratio for LE
of single-layer boards is greater than that of three-layer boards. It is
thought that the randomly dispersed core of three-layer boards effectively
restrains LE in the direction orthogonal to particle alignment. It seems
that a cross-oriented core may be the most effective way to reduce this
dimensional change and to improve the low bending property values in
the direction orthogonal to particle alignment [6, 7].
2. Electrostatic method
With this method, the fibers or particles fall freely through an electro-
static field, become polarized and are aligned in the direction of the
electrical lines of force within the field. There are three types of electrode
arrangement in the high-voltage electrostatic orientation system (Fig. 6)
[8].
Particle Particle
^ 0 0
Particle
^ Teflon-coated 1 ^ I I
glass fiber mesh 0
( a ) Top electrodes (b) Top - bottom (c) Bottom

electrode electrodes
combination
FIG. 6. Possible electrode configurations with the electrostatic method of mat
formation. 0 and are the electrodes, and heavy, solid lines represent the
charges; thin lines represent the direction and shape of electrostatic lines of force.
Kawai et al. [4] have investigated the effects of species, particle shape,
particle moisture content and electrostatic field intensity on the degree of
alignment in aligned particle mats using the configuration shown in Fig.
6(a) (method A). The field intensity and particle moisture content are
significant factors for particle alignment, whereas none of the other
factors showed a statistically significant effect. With an increase in the
field intensity, the average alignment angle decreased and better-aligned

mats were obtained. The best value for AAA was obtained with a field
intensity of 4kV/cm and a moisture content of 10%. By this method,
particles are oriented by passing between the electrodes during the
mat-forming process, although there is significant unstable movement of
the particles as they fall. The tendency of particles to stick to the lower
edge of the electrodes and form 'bridges' between the electrodes and
between the mat and electrodes disturbs the orientation, and an even
thickness of the mat is also difficult to attain. Bridging can be eliminated
to some extent by putting another set of electrodes with correspondingly
similar (negative or positive) charges below the forming belt as shown in
Fig. 6(b) (method B). However, the instability of the particles still remains
[9].
An even mat and a better orientation in the bottom layer of the mat
can be achieved by eliminating the upper electrodes and using only the
electrodes on the underside of the forming belt (Fig. 6(c), method C) [10].
The unstable movement of the particles is reduced and the particles are
oriented along the electrostatic lines of force just before touching the belt
or the mat surface. This method is also much safer because the electrodes
can be covered with isolation plates so that actual contact with the
particles is eliminated. Another advantage is that these isolation plates
serve, at the same time, to intensify the aligning torque generated on the
particles when a special material is used for the plates [11]. Despite the
high voltage required with the electrostatic method, the electrical con-
sumption is relatively low, for instance in the range of 01-0-5 watt/cm
(length of electrode) at 4kV/cm when the electrode spacing is 15-20 cm
[8].
Sasaki et al. [8] have investigated the orientation of particles by using
the electrostatic methods (Fig. 7). With method C, the electrostatic field
intensity decreases as the distance from the electrode increases. AAA of
the particles, which is inversely proportional to the electrostatic field
intensity, increases as the distance from the electrodes increases. For
particles near the bottom of the mat, the orientation was better than
when lower and upper electrode pairs (method B) were used. It should be
noted that the upper portion of the mat was not as well oriented, which
indicates that this method is more useful for thinner board production
(less than 6 mm thick) because the oriented portion will be relatively
greater than in thicker boards. In Fig. 7, data for softwood fibers are also
plotted for comparison. The fibers appear well oriented by method C.
However, the crimp of the fibers and their intertwining decreases the
Orientation
direction
50
Average alignment angle (degrees)
40 Random ^ 1
30-
20 -
*
Electrode method B
A
4 L Electrode method C
O akamatsu , strand (m.c. 13 %, U M F )
A lauan, strand (m.c.13%, U M F )
softwood , fiber (m.c. 11 %)
J-
5 10 15 20
M a t height (mm)
FIG. 7. Relationship between the mat height and average alignment angle for
akamatsu (Pinus densiflora S. et Z.) and lauan (Shorea spp.) particleboards. (1)
Average orientation in a random-formed mat; (2) average orientation in an
oriented mat using the top-bottom electrode combination (method B); (3) average
orientation in an oriented mat using bottom electrodes only (method C).
effect of the actual orientation of individual fibers on the board proper-

ties. Generally, smaller particles such as semi-strand can be oriented more
easily than strand-type particles by the electrostatic method.
A disadvantage of using electrodes at the bottom of the forming belt is
that the particle mat has a graduated orientation, being better toward the
bottom and becoming less oriented as the distance from the electrodes
increases. This is a limiting factor in the thickness of particle mats that
can be made with this method. However, the less-oriented portion at the
top of the mat can be shaved off to make a more even mat, and two mats
can then be bonded together, resulting in a double mat with better
orientation toward the surfaces. Table 1 shows the properties of oriented
particleboards manufactured by laboratory-scale production apparatus
30
TABLE 1
3
Properties of semi-strand board adjusted to board density 0-7 g/cm
akamatsu
lauan (Shorea spp.) (Pinus densiflora S. et Z.)
Board property Oriented Random Ratio Oriented Random Ratio
I 1 R HA ll/R I 1 R HA ll/R
MOE, dry (GPa) 5-8 2-1 3-6 2-8 1-6 4-7 1-7 2-9 2-7 1-6
MOE, wet (GPa) 2-5 10 1-5 2-3 1-7 1-7 0-6 11 2-9 1-6
MOR, dry (MPa) 36-3 17-1 26-4 20 1-4 29-6 14-7 21-2 2-0 1-4
MOR, wet (MPa) 110 6-3 9-3 1-7 1-2 10-5 5-7 6-7 1-8 1-6
IB (MPa) 0-6 0-6 n.s. 0-5 0-5 n.s.
TS (%) 14-5 17-9 n.s. 37.6 36.2 n.s.
LE (%) 0-27 0-76 0-50 0-36 0-48 0-18 0-93 0-52 019 0-35
HIROMU KAJITA AND JUNSUKE MUKUDAI
Notes: Parallel to orientation; 1: perpendicular to orientation; wet: 70 C, 2 h then 20 C, 1 h immersion; IB: internal
bond strength; TS: thickness swelling after 24 h immersion at 20 C; LE: linear expansion measured on wet bending
specimens; n.s.: not significant.
using method C [8]. Oriented boards from lauan (Shorea spp.) and
akamatsu (Pinus densiflora Sieb. et Zucc.) are fairly orthotropic, with the
ratio for M O E in the range 2-3-2-9. M O E in the oriented direction
is 60 to 70% greater than that of random boards, while the ratios of
MOR are somewhat less than those of MOE. LE of oriented strand
board (OSB) in the oriented direction is 1/3 to 1/5 of that in the
perpendicular direction. Bending-creep deflection in the oriented direc-
tion is as low as that of plywood.
The proposed configuration for industrial application of this system is
shown in Fig. 8. One set each of electrostatic orienters for top and bottom
mats is used, the less-oriented top portion of each mat being removed by
shaving to make an evenly oriented mat. The upper mat is turned over
by a flexible reversing-belt and a rigid drum without disturbing the
orientation of the particles, the turned-over mat then being deposited on
top of the lower mat by using a sliding plate with a low coefficient of
friction. A continuous particle mat with well oriented surfaces results.
Coupled with a continuous press, this method is ideal for continuously
oriented thin-board production [8].
FIG. 8. Schematic diagram of equipment for producing oriented particle mats

with electrostatic fields charged from the underside of the belt conveyor. P and
P', particle inlet; R and R', particle return; 1 and 1', forming heads; 2, electrodes;
3, forming belt; 4, shave-off roller; 5, reversing belt; 6, particle mat.
REFERENCES
[1] H. Kajita and J. Mukudai, J. Soc. Mater. Sci. Jpn., 30, 736 (1981).
[2] H. Kajita, Mokuzai Gakkaishi, 33, 865 (1987).
[3] M. Suzuki and N. Sekino, Mokuzai Gakkaishi, 28, 91 (1982).

[4] S. Kawai, L. Ling-Lee, H. Sasaki and S. Ishihara, Mokuzai Gakkaishi, 28,
295 (1982).
[5] R.L. Geimer, USDA Forest Serv. Res. Pap. FPL 275, 1 (1976).
[6] F. Saito, M. Ikeda and H. Tsuchiya, Mokuzai Gakkaishi, 33, 19 (1987).
[7] H. Kajita and J. Mukudai, J. Soc. Mater. Sci. Jpn., 39, 1218 (1990).
[8] H. Sasaki, S. Kawai, O. Pulido and N. Pengprecha, Mokuzai Gakkaishi, 35,
725 (1989).
[9] L. Ling-Lee, S. Kawai, H. Sasaki and S. Ishihara, Mokuzai Gakkaishi, 28,
301 (1982).
[10] H. Sasaki and T. Yamamoto, Jpn. Pat. 59-17946 (1984).
[11] Y. Yoshida, O. Pulido, S. Kawai and H. Sasaki, Mokuzai Gakkaishi, 34, 401
(1988).
Low-Density Particleboard
SHUICHI KAWAI and HIKARU SASAKI
ABSTRACT
Dimensionally stable low-density (0-4-0-6 g / c m ) particleboard is required as a substitute
3
for plywood panels. This paper discusses the limit for lowering the density of particleboards
by using isocyanate resins and the effects of various factors such as the species (density) and
configuration of the particles, the resin content and resin components on the properties of
low-density particleboard.
The mechanical properties of boards from different species had linear relationships with
the compaction ratio (the board density divided by raw material density), and the practical
lower limit of the compaction ratio was found to be 0-7-0-8 for isocyanate resin. The
thickness swelling of the boards was independent of the species of raw material, and
increased with increasing compaction ratio, i.e., low-density (low-compaction) particle-
boards were more dimensionally stable. Linear relationships between the mechanical
properties and shape factors based on the particle configuration and derived from fracture
mechanics theory were observed. Both the mechanical and dimensional properties of the
boards improved with increasing resin content, but were generally independent of the
formulation of isocyanate compound adhesives such as the free-isocyanate group content
and functionality of crude methylene diphenyl diisocyanate (MDI). Crude M D I alone,
compound resin, and unreacted mixed resin did not make much difference in the board
properties either.
Keywords: particleboard, low density, isocyanate resin, compaction ratio, dimensional

stability, mechanical properties
INTRODUCTION
In particleboard production, wood particles are plasticized and par-

tially compressed under an unsteady state of heat and moisture by
hot-pressing. This gives an increase in the contact area of the adhesive
and improves the bondability among the compacted particles. However,
a compressed particle tends to recover from its deformation, which is
33
34 SHUICHI KAWAI AND HIKARU SASAKI
known as the spring back of a drying set, when soaked in water.

Therefore, particleboard has the potential to swell, especially in the
thickness direction.
The low-density (low-compaction) particleboard developed in recent
years has reduced this potential swelling and thus improved not only the
dimensional stability, but also the thermal and sound insulation proper-
ties of the board [1, 2].
The important processes contributing to a low board density are the
preparation of the particles and the spraying of the resin. This paper
discusses the limit to which the density of particleboards can be reduced
for isocyanate resins and the effects of such factors as the species (density)
and configuration of particles, the resin content and resin components on
the physical and mechanical properties of low-density particleboard.
EFFECTS OF THE PARTICLE DENSITY AND COMPACTION

RATIO ON BOARD PROPERTIES
Particles of uniform dimensions prepared from various species in a

density range of 0-2-0-7 g/cm were pressed into particleboards in a
3
density range from 0-1 to 0-9 g/cm , using an isocyanate compound

3
adhesive at a 10% resin content.

Figures 1-3 show the effects of particle density and compaction ratio
(board density divided by particle density) on the mechanical strength of
boards [ 3 ] .
Both the moduli of rupture and elasticity of the boards in each species
have a linear relationship to the compaction ratio. The ratio of the
modulus of rupture of each board to that of solid wood in the fiber
direction is proportional to the compaction ratio, and this relationship
seems to be independent of the species. The modulus of rupture of each
board is about 30% of that of solid wood, showing that such recon-
stituted wood can attain a bending strength of 70-80% of that of the
solid wood when the particles are highly oriented.
The internal bond strength is a function of the bondability of glue, and
is affected by the resin content per unit area and by the contact area. The
former is proportional to the density of the particles when the resin is
sprayed on particles of uniform dimensions at the same level of resin
content. The latter increases with increasing compaction ratio, and this is
supposed to be accelerated when the particles are partially compressed.
Therefore, the internal bond strength is greater in particleboard from a
LOW-DENSITY PARTICLEBOARD 35
80
Modulus of rupture, MOR (MPa)
60 I
Light red meranti
PT = 0A0
Apitong p = 0.70/
T
I/
40 Balsa-apitong
,,-0.29.
20
0.5 1.0 1.5

Compaction ratio, CR
FIG. 1. Relationship between the modulus of rupture and compaction ratio.

The length, width and thickness of the particles were 30, 7 and 0-5 mm,
respectively; p is the density of the raw material
r
MOR (board) /MOR (wood)
O Balsa
100 Light red meranti
Apitong
50
0 0.5 1.0 1.5

FIG. 2. Relationship between the MOR ratio of the board to solid wood
(expressed as a percentage) and the compaction ratio.
higher density species at the same level of compaction ratio, and increases
exponentially with increasing compaction ratio. This explanation fits
Fig. 3 well. When the bonding strength of the adhesive is sufficient,
however, the internal bond strength will be limited by the tensile or
bending failure of the particles, which results in a saturation curve in the
O Balsa p = 0.21
r
I
2 Light red
<aw) 8/ *Mûaj;s puoq
meranti p = 0.4C
r
I Apitong p = 0.7C
r
4
I Balsa-apitong |~
P r = 0.29J
|BUJ9ju|
0 0.5 1.0 1.5 2.0

FIG. 3. Relationship between the internal bond strength and compaction ratio.
The configuration of particles is given in Fig. 1; p is the density of the raw
r
material.
higher range of compaction ratio, as shown in Fig. 3. The internal bond

strength of boards from the mixed species is principally determined by
the strength of the weaker (lighter) species, although it is reinforced
slightly by the stronger (heavier) species.
The practical lower limit of compaction ratio is about 0-7-0-8 under
these conditions for moduli of rupture and elasticity, and internal bond
strength.
Thickness swelling of the boards in each species after long-term water
immersion increased with increasing board density. This is supposed to
be directly related to the spring back of the compression set formed in
the particles during hot-pressing. Figure 4 shows that the thickness
swelling after 31 days of water immersion increased linearly with increas-
ing compaction ratio, and is independent of the species used, which
proves that low-density (low-compaction) particleboards have better
o Balsa P r = 0.21
40
Thickness swelling, TS (7.)

Light red
meranti /O = 0.40 R
! Apitong = 0.70 P r
& Balsa-apitong / M
20 p =0.29 X
r / ^
i 1 1 i i i
0 1.0 2.0
FIG. 4. Relationship between the thickness swelling and compaction ratio after
31 days of water immersion. M represents the master line for different boards of
each species, CW is the theoretical thickness swelling of a balsa-apitong board,
and p is the density of the raw material. The configuration of the particles is
r
given in Fig. 1.
dimensional stability. The thickness swelling of the boards with mixed

species follows the additive law.
EFFECT OF PARTICLE CONFIGURATION ON

BOARD PROPERTIES
Twenty-four types of lauan (Shorea spp. with a density of 0-40 g/cm ) 3
particles with different dimensions that were strictly controlled for length
(A), width (a>), and thickness (d) were each used to produce low-density
particleboards with a density of 0-40 g/cm , using an isocyanate com- 3
pound adhesive [ 4 ] .
The bending strength increased with increasing X and with decreasing
d. N o clear effect of co on the bending strength was observed, but the
bending strength seems to have decreased with increasing co. The inter-
nal bond strength increased with increasing d, and with decreasing X
and co.
Such an effect of particle configuration on the mechanical strength of a
board can be explained by fracture mechanics theory. In the layered
structure of a particle mat, voids formed along with the particles are
assumed to be cracks. The dominant crack which affects the bending
strength is a discontinuity formed at the cross section of a particle. When

the crack dimensions are considered to be proportional to the product
of the thickness S and average width 2co/n of a particle, i.e., the cross-
sectional area, the bending strength is inversely proportional to the
crack dimensions approximately to the power of 0 5 . On the other hand,
length X of a particle is a factor related to the distribution of these
dominant cracks. As X becomes smaller, interaction among the cracks
becomes larger, resulting in a decrease in strength. The theory states that
this change in bending strength is proportional to the square root of X.
Consolidating these relationships gives the bending strength as propor-
tional to (nX/25co) ' , which is called the shape factor in bending
0 5
strength.
In the same manner, the crack dimensions are considered to be
proportional to the product of X and co for tensile stress in the direction of
the board thickness, so the effect on internal bond strength is inversely
proportional to the square root of their product. The internal bond
strength is also proportional to the mass of adhesive per unit area, i.e.,
roughly to the thickness of the particles. Therefore, internal bond strength
is proportional to d(Xoj)~ , which is called the shape factor in internal
05
bond strength.
Figures 5 and 6 show the relationships between these shape factors,
and the bending strength and internal bond strength, respectively. A
linear relationship between the internal bond strength and the shape
factor can be seen in Fig. 6, showing that the effect of particle configur-
ation on the internal bond strength can be well explained by the shape
factor derived from fracture mechanics theory.
On the other hand, the data taken all together in Fig. 5 are not
linear. However, excluding the data from X = 1 0 mm and also for the
particles with the largest cross-sectional area (w = 7mm, <5 = 0-7 mm)
with X = 20 and 30 mm (those within the broken line), a linear rela-
tionship was also found between the bending strength and the shape
factor. This can be interpreted as follows: as the particles become
smaller, in this case at 10 mm in length, the number of cracks markedly
increases and thus the cracks along the board thickness are com-
pounded. This explains why fracture was initiated at a much lower
stress level than the theoretically predicted value based on simple crack
dimensions. In the case of an extremely large cross-sectional area, this
compounded effect on bending strength is extended to the boards with
X = 30 mm.
30 r-
25 h
20 h-
CL
3
Modulus
10 h-
5
2 4 6 8 10 12
M - (mm 05)
FIG. 5. Relationship between the modulus of rupture and (nX/2Sco) , the shape
05
factor in bending strength. The board density is 0-4 g/cm , and an isocyanate resin
3
adhesive was used at a 1 0 % resin content.
EFFECTS OF RESIN CONTENT AND RESIN COMPONENTS

ON THE BOARD PROPERTIES
Isocyanate compound resins were formulated from crude M D I

(isocyanate group (NCO) functional number, 2-3) and polyether polyol
(molecular weight, 3000; OH functional number, 3) with a free-NCO
group content of resin in the range of 5-30% at 5% intervals. Lauan
(Shorea spp. with a density of 0-4 g/cm ) particles prepared with a ring
3
flaker were pressed into particleboards with densities of 0-4 and 0-6 g/cm , 3
using these compound adhesives at an 8% resin content.

The results show that the N C O group content in the resin did not have
much effect on the mechanical and dimensional properties of particle-
boards, except at an extremely low value (5% N C O content).
In another series of experiments, particleboards were produced by
using crude MDIs (an N C O group content of about 30%) alone with
12
Internal bond strength (MPa)
1J0
3.8
0.6
D.4
t
FIG. 6. Relationship between the internal bond strength and S(X(o)~ ' , the 0 s
shape factor in internal bond strength. The board density and adhesive used are
given in Fig. 5.
N C O functional numbers from 2-3 to 3-2, and by using these resins (an
N C O group content of 18%) with polyether polyol and their unreacted
mixtures at a resin content level from 2 to 20%.
Figure 7 shows the effect of the isocyanate resin content on the internal
bond strength of the particleboards. An increase in the resin content
improved both the mechanical and dimensional properties of the particle-
boards. Crude MDI alone, the compound resin, and unreacted mixed
resin did not make any great difference in the board properties.
The effects of factors taken in this paper such as the species (density) and
configuraton of particles, the resin content, and the board density on the
mechanical and dimensional properties of particleboards were determined
by a multiple regression analysis [3, 4, 5]. Both the moduli of rupture and
elasticity could be predicted adequately with the board density and bulk
density of the particle mat as predictor variables (R = 0-90). The particle 2
mat bulk density correlated inversely with the shape factor in bending,
(nX/2d(o) divided by the density of the raw material (r= 0.85). This
05
9
Internal borid strength (MPa)

3
1 J
0 5 10 15 20
Resin content (%)
FIG. 7. Relationship between the internal bond strength and resin content for a
board density of 0 6 and 0 4 g/cm . 3
Board density
06 0 4 (g/cm ) 3
O Crude MDI alone (30% NCO content)

A A Compound resin (18% NCO content)
Unreacted mixed resin (18% NCO content)
suggests that the bulk density of a particle mat might be a useful index to
represent the factors of particle density and configuration in relation to the
bending strength of a board. The internal bond strength can be well
predicted by the board density to the second power and the particle density
(.R = 0-93). Resin content is definitely another important factor, although
2
factors related to the effective bonding area seem to be more significant in

the range used in this experiment. Thickness swelling can be estimated by
four predictor variables, these being the resin content, compaction ratio,
and particle length and thickness (K = 0-70). 2
REFERENCES
[1] S. Kawai, Wood Research and Technical Notes, No. 24, 31 (1988).
[2] S. Kawai, H. Sasaki, S. Ishihara, A. Takahashi and M. Nakaji, Mokuzai
Gakkaishi (J. Japan Wood Res. Soc), 3 4 , 973 (1988).
[3] S. Kawai, H. Suda and H. Sasaki, Mokuzai Gakkaishi, 32, 385 (1987).
[4] H. Suda, S. Kawai and H. Sasaki, Mokuzai Gakkaishi, 32, 376 (1987).
[5] S. Kawai, H. Suda, M Nakaji and H. Sasaki, Mokuzai Gakkaishi, 32, 876,
(1986).
Steam-Injection Pressing Technology
HIKARU SASAKI, SHUICHI KAWAI, TOSHIMITSU HATA

and
BAMBANG SUBIYANTO
Division of Material Science, R&D Center for Applied Physics,

Indonesian Institute of Sciences,
Kompleks Puspiptek, Serpong, Tangerang 13510, Indonesia
ABSTRACT
In order to establish the optimum steam-injection pressing technology, the characteristics
of curing adhesives under high steam pressure, the temperature behavior in the particle mat
core during steam-injection pressing, and the length of pressing time were investigated. A
semi-continuous steam-injection press of test-plant scale was then developed and applied to
the production of thick, low-density particleboard.
The duration and pressure of steam injection with urea formaldehyde (UF) and
urea-melamine formaldehyde ( U M F ) adhesives needed to be carefully determined for
particleboard production by high-temperature steam-injection pressing, too-high steam
temperatures resulting in a poor internal bond strength. On the other hand, the internal
bond strength of those boards bonded with other resins was not perceptibly influenced by
the steam temperature.
The temperature in the middle layer of a particle mat with a thickness of 20 or 40 mm
reached more than 100 C at the moment of steam injection, whereas it takes 4 and 11
minutes, respectively, in the case of hot pressing. The computer simulation predicts that
only a few seconds of steam injection is required for a mat thickness within 100 mm to reach
100C in the mat core.
Isocyanate bonded particleboards with a thickness of 20 mm could be produced with
a press time of around one minute. This corresponds to only about one-fifth of the
press time of a conventional hot-platen press, and this effect of steam-injection pressing
on shortening the press time would be greater for thicker board production.
The newly developed semi-continuous, steam-injection press worked satisfactorily, and
100 mm thick low-density particleboards bonded with an isocyanate resin could be
manufactured with a 90-second press cycle.
43
44 HIKARU SASAKI ET AL.
Keywords: pressing technology, steam-injection press, adhesive curing, press time, particle-
board, continuous press.
INTRODUCTION
A new pressing process for the production of thick particleboards with

very short pressing time has been developed. This new process involves
injecting high-pressure steam through perforated press platens toward
the mat surfaces. The mat acts as a seal to permit the pressure and
temperature to build up; consequently, the core temperature of the mat
rises rapidly, and the press time is much less than by conventional
hot-platen pressing.
This paper discusses the characteristics of curing adhesives under high
steam pressure or temperature, the temperature behavior of the particle
mat core during steam-injection pressing, and the reduced press time with
this process. The paper also reports the development of a semi-continu-
ous, steam-injection press of test-plant scale, and its application to the
production of 100 mm thick low-density particleboard bonded with an
isocyanate adhesive.
C U R I N G ADHESIVES WITH STEAM INJECTION PRESSING
Particleboard adhesives, i.e., urea formaldehyde (UF), urea-melamine

formaldehyde (UMF), phenol formaldehyde (PF), phenol-melamine for-
maldehyde (PMF), and isocyanate compounds (IC), were cured under
high steam pressure or temperature in a specially designed reaction cell
[1]. The heating temperature (steam pressure) was varied from 120 C
(0-2 MPa) to 200 C (1-5 MPa) with steam-injection times from 30 to 180
sec. After the steam had been injected for the scheduled time, the curing
reaction of the adhesives was immediately stopped by injecting liquid
nitrogen ( N ) into the reaction cell to freeze the adhesives. The hardness
2
of the resins cured by this process was then compared with the figure for
resins (control) cured in an oil bath under atmospheric pressure.
Figure 1 shows the hardness ratios of the U F and U M F adhesives in
relation to the steam injection conditions. At higher temperatures and
consequently greater steam pressures, the maximum value (peak) for the
hardness of the cured U F adhesive occurred with a short steam-injection
time. At a lower temperature, the peak occurred with a longer steam-
STEAM-INJECTION PRESSING TECHNOLOGY 45
100
Hardness ratio. R {%) 80
eo
40
20
0 30 00 90 120 150 180

Steam injection time (sec)
FIG. 1. Relationship between the hardness ratio and steam-injection time for
U F and U M F resins.
O 120C, A 140C, 160C

O 180C, T V 2 0 0 C.
UF, UMF
injection time. After this peak, the hardness decreased with increasing
steam-injection time at all temperature levels, and finally the resin
decomposed to a liquid. The time range in which the preferred hardness
of the resin for bonding exists was wider at lower temperatures than at
higher temperatures. Therefore, the duration and pressure for steam
injection with a U F adhesive needs to be determined carefully for particle-
board production by high-temperature steam-injection pressing.
The reason for this hardness change with the U F resin may have been
that the condensation and decomposition reactions occurred simulta-
neously in the resin during steam injection at high temperatures. The
former reaction is dominant until the hardness reaches its peak, whereas
the latter is dominant after the peak. In this case, decomposition may
have been caused by the presence of excessive water, which could not be
easily extracted from the reaction system under high pressure and would
have accelerated hydrolysis.
The trends in this relationship for the U M F and P M F resins were
similar to those for UF. However, the hardness peak of these two
resins occurred after a longer steam-injection time than that for the U F
resin, which suggests that these resins required more heat energy for
curing than did the U F resin. The hardness of the P F adhesive cured
with steam injection was quite low compared with that of the PF resin
cured under atmospheric pressure. The isocyanate resins could be
cured under all steam pressure conditions, and foamed after reacting
with water.
The foregoing observations indicate that curing adhesives under high
pressure (temperature) is different from that under atmospheric condi-
tions. The concept that both higher temperature and longer heating time
always results in better curing does not hold true under high-pressure
conditions. Polymerization and decomposition may occur one after the
other under such conditions.
Particleboards with a density of 0-6 g/cm were manufactured under
3
various conditions of steam-injection pressing, using the adhesives al-

ready described [ 2 ] . The maximum values of internal bond strength for
U F - and UMF-bonded boards were observed at a relatively low steam
temperature (120-140 C) and with long pressing time (120-150 sec).
Higher steam temperatures resulted in poor board properties. On the
other hand, the internal bond strengths of the PF-, PMF-, and IC-
bonded boards were not perceptibly influenced by the steam temperature,
and higher steam temperatures seemed to permit shorter pressing times.
All the boards produced by steam-injection pressing had better dimen-
sional stability than those produced with a conventional hot-platen press.
The trend in properties of the boards produced under high pressure
agrees well with that of the adhesives cured under the high-pressure
reaction cell conditions already discussed, but the optimum conditions
tended to favor a longer pressing time.
TEMPERATURE BEHAVIOR OF THE PARTICLE MAT

DURING STEAM-INJECTION PRESSING
The temperature behavior of the particle mat during steam-injection

pressing was investigated under various conditions, using both experi-
mental and analytical methods [3, 4 ] .
Figure 2 shows the temperature changes in the middle layer of the mat
during steam-injection pressing under various conditions. The temperature
in the middle layer of the mats with a thickness of 20 and 40 mm increased
beyond 100 C at the moment of steam injection, the vapor diffusing
through the mat, as soon as steam was injected. However, the maximum
temperature reached in the middle layer during steam injection tended to
decrease with increasing mat thickness. As the steam pressure for injection
and the mat density was increased, the temperature in the middle layer
increased accordingly. When steam injection was continued, the core mat
(C) 1501
100
Temperature
50
I Steam on
I I I I I
0 2 4 6 8 10
Pressing time (min)
FIG. 2. Temperature behavior in the middle layer of the mat during steam-
injection and conventional hot-platen pressing. A: Steam injection (steam pressure,
0 6 3 MPa; time, 3 0 sec; initial moisture content, 11%; board thickness, 2 0 mm); B:
steam injection (0-63 MPa, 30 sec, 0%, 20 mm); C: steam injection (0-63 MPa, 30 sec,
11%, 20 mm); D: steam injection (0-63 MPa, 30 sec, 11%, 4 0 mm); E: conventional
hot platen (initial moisture content, 11%; board thickness, 20 mm); F: conventional
hot platen (initial moisture content, 11%; board thickness, 4 0 mm). Board
parameters: particles, semi-flake; board density, 0-4 g/cm (only A, 0-6 g/cm );
3 3
adhesive, isocyanate; temperature of hot platens, 160 C.
temperature remained constant during injection. This suggests that with

longer steam-injection times, the inflow and outfow of steam in the middle
layer of the mat was balanced, and the internal vapor pressure or
temperature remained constant. After injection was stopped, the internal
vapor pressure or temperature of the middle layer fell due to the outflow of
steam from the edges of the mat. The temperature of the dry mats then con-
tinued to rise, whereas that of the mats with a moisture content of 11% de-
creased to about 100 C before gradually rising at a lower rate than that for
the dry mat. This is because the heat energy supplied from the hot platens
was partially consumed in evaporating the water present in the particles.
In the case of hot-platen pressing, it required about 4 and 11 minutes
to increase the temperature to 100 C in the middle layer of a mat with a
thickness of 20 and 40 mm, respectively. With increased moisture content,
the time necessary for the middle layer to reach 100 C tended to be less,
whereas the time to maintain a constant temperature (about 100 C) was
prolonged.
Such experimental data on the temperature behavior in a particle mat

during pressing suggests that the mechanism for heat transfer in the mat
during steam-injection pressing is different from that in conventional
hot-platen pressing. In the former case, steam flow at a high temperature
controls the heat transfer through the particle mat, while in the latter case
the heat energy supply depends on heat conduction from the hot platens
to the particle mat. The heat conduction law, with a driving force of
temperature gradient, and Fick's law of diffusion, with that of pressure
gradient, were applied as theoretical equations to the thermal flow in a
mat during hot pressing accompanied with steam injection. Since both
equations have similar formulas, the same type of finite element equations
could be derived for a numerical analysis. A computer program for
two-dimensional heat transfer was modified to allow an analysis of the
temperature distribution for both hot-platen pressing and steam-injection
pressing.
The calculated results agree comparatively well with the observed
results from both pressing methods. Figure 3 shows the calculated results
of the relationship between the mat thickness and steam-injection time
necessary for a mat core to reach 100 C within 10 sec after steam
injection. With increasing mat thickness, the injection time required to
reach 100 C increases. However, only a few seconds of steam injection
are required for a mat thickness within 100 mm to reach 100 C in the
mat core. Therefore, the normal range of mat thickness would not affect
the press time required for practical steam-injection pressing. The simula-
200
(sec)
Steam-injection pressure: 0.4 MPa

Mat density: 0.4 g/cm 3
150
Steam-injection time
100
50
100 200 300 400 500

Mat thickness (mm)
FIG. 3. Theoretical relationship between the mat thickness and steam-injection
time necessary for a mat core to reach 100 C within 10 sec after injection.
tion also predicts that about 180 sec of steam injection is required for the
core of 500 mm thick mat to reach 100 C, the temperature necessary for
an isocyanate resin to cure in about one minute [1]. Thus, such a thick
board could be produced with 4-5 min of press time by steam-injection
pressing.
OPTIMIZING PRESS TIME BY STEAM-INJECTION PRESSING
Particleboards with a thickness of 20 mm and using an isocyanate

compound adhesive were produced by steam-injection pressing in order
to determine the optimum conditions [5].
Enough energy to produce 20 mm thick boards could be supplied with
3 sec of steam injection. The bending properties tended to increase as the
initiation of the steam supply was delayed. However, the initiation of
steam injection for a compaction ratio of 1-0-1-3 would be rational to
optimize the total press time, because a longer pressing time and a greater
pressure are needed to compress the mats to a target density during the
production of high-density boards.
Figure 4 shows the relationship between the total pressing time and
internal bond strength for both steam-injection pressing and conven-
tional hot-platen pressing. Steam-injection pressing makes it possible to
produce 20 mm thick particleboards in a press time of 45 sec for boards of
0-4 g/cm density, and in 60 sec for boards of 0-6 g/cm density. This
3 3
corresponds to only 1/5-1/6 of the press time for conventional hot-platen

pressing. The results of the computer simulation already described
suggest that this effect of steam-injection pressing on press time would be
even greater for thicker board production.
The steam-injected particleboards had slightly inferior bending proper-
ties to those of conventional platen-pressed particleboards, because of
their more uniform density distribution through the board thickness.
They had, however, greater internal bond strength and screw-withdrawal
resistance at the end surfaces of the boards.
DEVELOPMENT OF A CONTINUOUS PRESS WITH

STEAM INJECTION
As the first step to develop a continuous press with steam injection, a

semi-continuous (intermittent) type was produced [ 6 ] . In this press,
(MPa)
2.5
2.0
Internal bond strength
1.5
1.0
0.5
0 1 2 3 4 5
Total pre8sing time (min)
FIG. 4. Relationship between the total pressing time and internal bond strength
of a particleboard.
Board density
0-4 0-6 (g/cm ) 3
Conventional hot-platen pressing O A

Steam-injection pressing # A
The ranges of mean value show 9 5 % confidence intervals. Injection time, 3 sec;
compaction ratio at the initiation of steam-injection, 1 0 (mat density, 0-4 g/cm ),
3
and 4 / 6 (mat density, 0-6 g/cm ); board thickness, 2 0 mm.

3
upper and lower flat hot platens with steam injection from each surface
were used. The press moves with the forming belt and mat during
pressing, and after moving a certain distance, opens and returns to the
initial position. This operation is repeated, and boards can be continu-
ously produced. A teflon-coated glass net was used for the forming belt to
allow steam to be injected from the hot platens into the mat through the
belt.
Figure 5 illustrates this press. Perforated steam-injection plates with a
thickness of 32 mm are attached to upper and lower hot platens with
dimensions of 1200 (L) x 850 (W) x 35 (T) mm. The upper platen can move
up and down with a maximum pressure of 1-5 MPa. At the surface of the
steam-injection plates, holes of 2-0 mm in diameter for injecting steam are
drilled in a 25 x 25 mm spacing pattern. These holes are grouped into four
sections with an area of 300 x 8 5 0 mm each. Steam injection in each
section can be opened or closed separately by control valves. The inlet to
the press is provided with a tapered cold press 200 mm in length to
Lifting jack Upper platen jack Unit: mm
^1
Oil t a n k |
Hot platens
200^"
Conveyor >
STEAM-INJECTION PRESSING TECHNOLOGY
I Steam-injection plates Jack for reciprocating motion
FIG. 5. Construction of a semi-continuous (intermittent) steam-injection press.

51
reduce any discontinuity in pressure and mat thickness. The angle of the
taper can be changed within the range 1/4-1/10 by changing the attach-
ment plate.
This press moves in the direction of the mat conveyor in a reciprocat-
ing motion. For this purpose, the press is supported by four wheels
running on two rails. The press is small size and operates at a low speed
because it was only designed for experimental purposes. However, the
equipment is suitable to study the fundamental problems of semi-
continuous steam-injection pressing.
Steam injected from the hot platens diffuses not only in the thick-
ness direction but also in the horizontal direction of the mat. There-
fore, the selection of position, time and duration for steam injection
are very important to prevent the adhesive from pre-curing in a
mat which is waiting its turn for hot pressing in the tapered cold
press.
Experiments were conducted to press 100 mm thick lauan low-density
particleboard with a density 0-4 g/cm with this semi-continuous steam-
3
injection press [ 7 ] . The change in temperature distribution in the mat

during pressing, and the internal bond strength, bending properties and
thickness swelling of the board produced were measured.
The optimum pressing conditions could be determined by adjusting
the steam injection duration and injection area; for instance, 10 sec of
injection over platen surface 30 cm in from the edges of the press
platen, and then 80 sec of hot-pressing was optimum for a 100 mm thick
board.
Repeated pressing with no steam injection was necessary for the
zone of the particle mat existing between consecutive steam-injection
pressing cycles in order to produce homogeneous properties through-
out the length of the boards. The optimum length of this addi-
tional pressing operation was around three times the thickness of the
board.
The deviation in internal bond strength of the steam-injection pressed
board was less than that of conventional hot-pressed board.
These results indicate that this type of semi-continuous (inter-
mittent) steam-injection press is applicable to the production of con-
tinuous thick particleboard. The size of the press would be much
smaller than that needed for a conventional continuous hot-platen
press due to the short pressing cycle with steam injection. Boards
produced by this press had an even thickness and homogeneous
properties.
REFERENCES
[1] B. Subiyanto, S. Kawai, M. Tanahashi and H. Sasaki, Mokuzai Gakkaishi (J.

Japan Wood Res. Soc), 35, 419 (1989).
[2] B. Subiyanto, S. Kawai and H. Sasaki, Mokuzai Gakkaishi, 35, 424 (1989).
[3] T. Hata, B. Subiyanto, S. Kawai and H. Sasaki, Wood Sci. and Technol., 2 3 ,
361 (1989).
[4] T. Hata, S. Kawai and H. Sasaki, Wood Sci. and Technol, 24, 65 (1990).
[5] T. Hata, B. Subiyanto, S. Kawai and H. Sasaki, Mokuzai Gakkaishi, 35, 1080,
1086 (1989).
[6] H. Sasaki, B. Subiyanto and S. Kawai, Mokuzai Kogyo (Wood Industry), 4 5 ,
409 (1990).
[7] B. Subiyanto, S. Takino, S. Kawai and H. Sasaki, Mokuzai Gakkaishi, 37, 24
(1991).
Laminated Veneer Lumber and Composite
Beams Produced from Tropical Hardwood
Thinnings
HIKARU SASAKI, QIAN WANG, SHUICHI KAWAI

and
RAZALI ABDUL KADER
Faculty of Forestry, University Pertanian Malaysia, Serdang,

Selangor, Malaysia
ABSTRACT
The mechancial properties of laminated veneer lumber (LVL) made from plantation
thinnings (9-year old) grown in Sabah, Malaysia were investigated. The species used were
Acacia mangium (AM), Gmelina arborea (GA), Albizia falcata (AF) and Eucalyptus deglupta
(ED). Laminated veneer lumber (LVL) was made up of 9-ply veneers of 2-5 mm thickness
from each species, using staggered scarf joints for the veneers. Composite beams were
constructed with LVL flanges and a low-density semi-strand board (LDB) web, and the
mechanical properties were tested.
The average standard deviation in modulus of elasticity (MOE) of LVL was \/y/n (where
n is the ply number) of that of solid sawn lumber of the same species.
The modulus of rupture (MOR) of LVL was almost the same as that of solid sawn
lumber, and the reduction in the variation of M O R was not as great as that found with
MOE.
M O E and M O R for the composite beam were lower than the figures for LVL, but much
higher than those for LDB. The variation in these properties was a little less than that of
LVL, so that the composite beam tested is potentially useful as a structural member.
With the short-span bending test specimens, shearing failure occurred near the glue layer
between the flange and web. The criteria derived for the onset of bending or shear failure
were useful for predicting the failure pattern.
Keywords: thinnings, laminated veneer lumber, composite beam, low-density particleboard.
55
INTRODUCTION
Fast-growing tropical species are the most promising wood resource

for the future. Some species have already been planted successfully and
are growing to produce attractive young forests.
The logs produced from thinning operations in these forests have not
so far been utilized, since their mechanical properties are generally lower
than those of the useful softwood species and the variation is wider.
This paper reports the technology for improving the properties by
processing the logs into laminated veneer lumber (LVL). An application
of the LVL material as the flanges of a composite beam with a
particleboard web is introduced.
EXPERIMENTAL P R O C E D U R E
Raw Materials
Nine-year old logs from thinning operations at a plantation in Sabah,
Malaysia were used. The logs were selected, cut into 2 m lengths and
shipped to Japan with the cooperation of the Sabah Foundation. The
species, volume and diameter of each type of log are shown in Table 1.
Each 2 m long log was cut in half, and one half was sawn into JAS
(Japanese Agricultural Standards) sized lumber including knotty and
clear specimens. The other half was peeled into veneers 2-5 mm thick by
using a lathe (Tanouchi type) with a powered back-up roll which was
developed by the authors' group.
LVL
Scarf joints slope) were used to join the ends of the veneers, this
being done with the Hashimoto automatic scarf joint method, automatic
assembly, and final semi-continuous (intermittent) hot-pressing.
Nine-ply LVL with well-staggered end scarf joints was made, using a
urea melamine formaldehyde resin adhesive, by hot-pressing at 105 C
and 0-8 MPa pressure for four minutes. The dimensions of each LVL
material were 4 m long, 920 mm wide and 21 mm thick.
Composite Beam
Composite beams were constructed from LVL flanges and a low-
density particleboard (LDB) web. This board, 60 mm thick with a specific
gravity of 0-4, was made from semi-strand particles of Shorea species, and
TABLE 1
Logs from the Sabah Foundation [1]
Symbol Species Number Average Volume Average

diameter air-dried density
3 3
mm m g/cm
AM Acacia mangium Will. 10 185 0-389 0-54

GA Gmelina arborea Linn. 15 213 1068 0-49
AF Albizia falcata Back. 19 228 2-528 0-29
ED Eucalyptus deglupta BI. 15 206 0-963 0-49
LAMINATED VENEER LUMBER AND COMPOSITE BEAMS
57
glued with an isocyanate compound adhesive at an 8% resin content,

based on the oven-dried weight of the particles. A semi-continuous
(intermittent) press with steam injection developed by the authors' group
was used to produce boards to 2700 x 750 x 60 mm in size. The pressing
time was 1-5 min at a platen temperature of 160 C, including 10 sec of
steam injection at a 6 atmosphere steam pressure.
The board was cut into 2-7 m long and 10 cm wide sections, which
corresponds to the height of the web for the composite beam. LVL strips,
2-7 m long and 6 cm wide, were bonded to both edges of the particle-
board section to make the composite beam (Fig. 1). Phenol resorcinol
resin adhesive was used at a 300 g / m glue application in the single glue
2
line. Cold pressing for 24 h at a pressure of 0-34 MPa was applied for
curing.
Before this composite process, block shear tests were carried out
on the glue bond made between the LVL material and the edge of the
low-density particleboards with varying application rates of the glue.
The results are shown in Fig. 2, which includes the results with the iso-
cyanate resin. The isocyanate resin had better bonding strength than
the phenol resorcinol resin under dry conditions. In other tests under
wet conditions (4h boiling, 2 h drying (105 C), cooling in water), both
resins showed equal bonding strength with high glue applications.
From this preliminary experiment, it was concluded that the optimum
amount of glue application in the liquid condition was 180 g / m 2
for the isocyanate resin and 300 g / m for the phenol resorcinol resin
2
adhesive.
[mm]
LVL
i!! 1111111 I!5

11111 O
o
i lllllMKl tt Xi
C ft)
ft) .
o
Milium I
m u
11111 f 1111 O (0
_a o.
1111111111
LVL
FIG. 1. Construction of the compo-
o
site beam.
LAMINATED VENEER LUMBER AND COMPOSITE BEAMS 59
Application of IC resin glue, liquid base (g/m ) 2
0 50 100 150 200 250 300 350 412
Application of PRF resin glue, liquid base (g/m ) 2
0 60 120 180 240 300 360 420 480 540
7 100
Wood failure (W.F ) in 1

6
( dW)
80
B
5
60
iia8uaaas Surpuog
3 40 -G
2
H 20
V
0
0 50 100 150 200 250 300 350
Glue application, solid base (g/m ) 2
FIG. 2. Bonding properties of LVL (flange) to the edge of low-density particle-

board (web) in the dry condition [4]. LVL was made from kamerere; bonding was
at 25 C and 0-5 MPa for 24 h; IC, isocyanate resin; PRF, phenol resorcinol resin.
RESULTS A N D DISCUSSION
Mechanical Properties of Solid Lumber and LVL [ 1 , 2 ]

A comparison between the mechanical properties of clear specimens of
solid wood and those of sized lumber, which included natural defects such
as knots, showed that the strength of the low-density species was not
sensitive to natural defects, while the high-density species were more
defect sensitive. The modulus of elasticity of albizia was increased to some
extent by the presence of knots.
Figures 3 and 4 show how the probability distribution of the mechan-
ical properties could be improved by the laminating process. In accord-
ance with the central limit theorem [ 3 ] , the standard deviation was
60
18r
16-
14-
12-
10-
8-
6-
Normalized frequency
HIKARU SASAKI ET AL.
4 -
2-
0 -
FIG. 3. Comparison of MOE distribution for dimension lumber and LVL (2-5 mm x 9 ply) from plantation thinnings from Sabah,
Malaysia [2]. Refer to Table 1 for definitions of AF, GA, ED and AM.
Sized lumber
T TTT
LAMINATED VENEER LUMBER AND COMPOSITE BEAMS
MOR MOR (MPa)
FIG. 4. Comparison of MOR distribution for dimension lumber and LVL (2-5 mm x 9 ply) from plantation thinnings from Sabah,
Malaysia [2]. Refer to Table 1 for definitions of AF, GA, ED and AM.
61
inversely proportional to the square root of the number of laminations.

In this case, the number of laminations was nine, because 9-ply LVL was
used. Theoretically, therefore, the standard deviation of the mechanical
properties of LVL could be one-third of that of solid sawn lumber.
In the experiment, the standard deviation of M O E for LVL was 1/2-5
(AM), 1/1-91 (GA), 1/5-72 (AF) and 1/3-67 (ED) of that for solid sawn
lumber, the 1/3-45 average of these values showing good applicability of
the theorem. M O E for LVL had a higher average value than that of solid
sawn lumber by 12-5% (AM), 9-1% (GA), 18-5% (AF) and 15-3% (ED).
The average value of M O E was increased by lamination, the main
reason for this being the increase in specific gravity of LVL after
hot-pressing. On the other hand, the average values of MOR were almost
the same for LVL and for the solid sawn lumber. This means that the
strength was related more to the size of the defects than to the average
density of the material. The reduction in the variation of MOR was not as
great as that found in the M O E value.
Mechanical Properties of the Composite Beams [ 4 ]

Figures 5 to 9 show the results of tests on the mechanical properties
of the composite beams. Data for solid sawn lumber, LVL and low-
A: LVL Y=161X+18.7, r 0.98

B: Composite beam Y=229X-35.6, r = 0 . 9 8
C: PB (web) Y=72.7X-18.4, r l = 0 . 7 0
12
O ED-LVL f l a n g e d c o m p o s i t e beam
AF-LVL f l a n g e d c o m p o s i t e beam A
10
(GPa)
8
moe
4 h A ED-LVL O PB ( Edgewise b e n d i n g )
AF-LVL PB ( H o r i z o n t a l b e n d i n g )
C
2 h
0
0.3 0.35 0.4 0.45 0.5 0.55
Specific gravity
FIG. 5. MOE as a function of specific gravity in a composite beam, LVL and
low-density particleboard (web) [4]. Refer to Table 1 for definitions of ED and A F .
A: LVL Y-1193X-45.9, r -0.88

a
B: Composite beam Y-1605X-401, rf-0.98

80
ED-LVL O ED-LVL flanged composite beam
AJ-LVL 0 AF-LVL flanged composite beam
PB(II) A A
OMR)
60
*oh
40
20
0.3 0.35 0.4 0.45 0.5 0.55

Specific gravity
FIG. 6. Bending strength as a function of specific gravity in a composite beam,

L V L , and low density particleboard (web) [4]. Refer to Table 1 for definitions of
ED and A F .
A: Sized lumber Y-6.81X-25.3,

B: LVL Y-7.76X-215,
C: Composite beam Y-6.23X-105,
O Sized lumber of ED
# Sized lumber of AF
A ED-LVL
A AF-LVL
CD
40
20 ED-LVL flanged composite beam

AF-LVL flanged composite beam
JL _L
20 40 60 80 100 120 140
MOE (GPa)
FIG. 7. Relationship between MOE and MOR for sized lumber, L V L , and a
composite beam [4]. Refer to Table 1 for definitions of ED and A F .
density particleboard are included for an easy comparison with the

properties of the composite beam.
In Figs. 5 to 7, MOE and MOR for the composite beams are a little lower
than those for the solid sawn lumber of LVL. The variation of each
property at each density level, however, is much smaller than those of the
solid sawn lumber and LVL. This means that the beam is a reliable material
for structural purposes, despite its inferior average mechanical properties.
Figures 8 and 9 show the shear strength of the composite beam, which
is important in short-span uses. Shearing failure occurred along the glue
line of the flange/web interface, while bending failure was observed when
the span of the beam was increased. The criteria for these modes of failure
are as follow:
Shearing failure occurred when
2LR//*[l-f (l-e)]<l
2
(1)
Bending failure occurred when
2LRIh[\-t {\-e)\>\
2
(2)
A: LVL Y-16.6X+35.1, r -0.45

B: Composite beam Y54.8X+3.33, r0.30
C: PB (web) Y=253X-68.2, r-0.95
A ED-LVL O ED-LVL flanged composite beam
6
A AF-LVL 0 AF-LVL flanged composite beam
Si 5
i 4
u
" 3
I2
U
A : Horizontal shear in bending
B : Horizontal shear along the
glue line in bending
1 C : ASTM edgewise shear
_L -L
0.3 0.35 0.4 0.45 0.5 0.55
Specific gravity
FIG. 8. Shearing strength as a function of specific gravity in a composite beam,
LVL, and low-density particleboard (web) [ 4 ] . Refer to Table 1 for definitions of
ED and A F .
A: ED Y= -0.02X+33.6, r =-0.76
B: AF Y= -0.02X+31.5, r*=-0.91
5 r- ft Compression failure of ED-LVL flanged composite beam

it Compression failure of AF-LVL flanged composite beam
O Shear failure of ED-LVL flanged composite beam
Shearing strength (MPa)
0 Bending failure of ED-LVL flanged composite beam

4 A Shear failure of AF-LVL flanged composite beam
Bending failure of AF-LVL flanged composite beam
2 A
Critical span
^ C r i t i c a l span
length for AFv
length for E D
(from eq. 1 )
(from eq. 1 )
Li L_
200 400 600 800 1000 1200
Span length (mm)
FIG. 9. Shearing strength of a composite beam as a function of span length [4].
Refer to Table 1 for definitions of E D and AF.
where L = span, /z = beam height, t = h /h (h = height of web), R = t / g

c c 0 0
( T = shear strength of flange/web interface, a = tensile strength of flange),

0 0
e = EJEf (E = Young's modulus of web, = Young's modulus of flange).

c f
The predicted values from these equations agree qualitatively with the
experimental results shown in Fig. 9.
Although the test results from the compressive indentation strength
perpendicular to the lamination of the beam are abridged, the compres-
sive strength of the composite beams was found to be similar to that of
the solid lumber and of LVL, and the deformation was very much
influenced by the species of the LVL flanges.
CONCLUSION
The mechanical properties of Sabah plantation thinnings are not

consistent enough when the material is used as solid sawn lumber. When
it is processed into LVL, however, the consistency could be increased in
proportion to the square root of the ply number. The application of LVL
to the flanges of a composite beam provided high stiffness with little
variation. However, the strength level is still not high enough, and further
technological development is necessary to increase the strength.
REFERENCES
[1] Q. Wang, H. Sasaki and R. A. Kader, Mokuzai Kenkyu Shiryo (Wood Res.
and Tech. Notes), No. 25, 45 (1989).
[2] Q. Wang, T. Hayashi, H. Sasaki and Y. Nagaya, Mokuzai Gakkaishi, (J.
Japan Wood Res. Soc), 36, 624 (1990).
[3] M. R. Spiegel, Theory and Problems of Statistics, McGraw-Hill, New York
(1972).
[4] H. Sasaki, Q. Wang, S. Kawai and R. A. Kader, Proc. 1990 Joint Int'l. Conf.
on Processing and Utilization of Low-Grade Hardwoods, Taipei, 173 (1990).
Chemically Modified Particleboards
HIROMU KAJITA
Department of Forestry, Faculty of Agriculture,

Kyoto Prefectural University, Shimogamo, Sakyo-ku, Kyoto 606, Japan
and
YUJI IMAMURA
ABSTRACT
Studies were carried out on chemically modified particle- and/or fiber-boards to improve
their mechanical properties, dimensional stability, and biological properties. The technology
developed from chemical modification of solid wood was applied to these reconstituted
wood products. This report describes the production and properties of the chemically
modified boards.
Keywords: dimensional stability, chemical modification, phenol resin impregnation, acetyla-
tion, formalization
INTRODUCTION
Dimensional instability is a greater problem in particleboards than

in solid wood products. Water absorption causes both irreversible
swelling due to a release of the residual compressive stresses imparted to
the board during the pressing process, and reversible swelling due
to water absorption by wood cell walls. There have been many efforts
to improve the mechanical properties, dimensional stability, and bio-
logical properties, using phenol-formaldehyde resin impregnation,
acetylation, maleic acid and glycerol mixture treatment, formaldehyde
treatment, etc.
67
68 HIROMU KAJITA AND YUJI IMAMURA
Phenol-formaldehyde Resin Impregnation

When wood is impregnated with a water-soluble phenolic resin that
penetrates into the cell-walls and is then dried and heat-cured, a bulked
product with insoluble polymers is formed which will not leach out in
water. As a result, hygroscopicity, swelling, and susceptibility to bio-
degradative organisms are all reduced. This treatment of incorporating
phenolic resin into wood was mainly developed for veneers to produce
plywood with better strength properties, and the same treatment for
particleboards was tried in order to improve their physical and biological
properties [1, 2 ] .
Strand particles prepared from sugi (Cryptomeria japonica D. Don)
thinnings were treated with a low-molecular weight (MW, 389) phenolic
impregnating resin (IPR, 0-30 resin solids) by dipping or spraying. The
particles were air-dried for several days and then dried at 60 C, before
being sprayed with 8% bonding phenolic resin (BPR, MW = 962) and
produced into single layer, 10 mm thick particleboards. A further set of
boards was produced from particles that had been sprayed with a mixture
of IPR and BPR in a rotary drum mixer. Boards with 10% IPR had
modulus of rupture (MOR) values that were 40-50% greater after two
hours of boiling than boards without IPR; the MOR values were not
significantly different between the methods for treating the particles with
IPR. Boards prepared by using only 15% BPR or only 15% IPR had
similar MOR values. Internal bond (IB) strength increased and thickness
swelling (TS) decreased with increasing IPR resin solids. TS after six
cycles of the ASTM accelerated aging test was less in the boards bonded
with only 15% IPR than in those bonded with only 15% BPR (Fig. 1)
[1]. The results obtained from accelerated laboratory tests on bio-
degradation suggest that the incorporated resin solids worked well to
enhance the decay- and termite-resistance of the particleboards. Decay
was assessed as nil or very low in those boards with 15% IPR. Brown-rot
fungus Tyromyces palustris was less affected by the phenolic resin
treatment, causing weight losses even for those boards with 15% IPR,
whereas this loading almost suppressed the decay by white-rot fungus
Coriolus versicolor. The resistance to attack by Comptotermes formosanus
termites was greatest in those boards with 15% IPR, which had the least
weight loss, termite mortality increasing with increasing IPR resin solid.
The mortality curves for termites fed on phenolic resin-treated particle-
boards gradually increased as the test period continued, which demon-
strates that the effect was not due to toxic chemicals but to an inability to
digest the material [ 2 ] .
CHEMICALLY MODIFIED PARTICLEBOARDS 69
IPR co
0 * 10
I 1 1 I I I I
0 1 2 3 4 5 6
Number of c y c l e s
FIG. 1. Change in thickness swelling (TS) of phenolic resin-treated particle-

boards during the A S T M 6-cycle accelerated aging test.
Because of the molecular size of phenolic resin, the resin solids cannot
occupy all the sites of potentially accessible regions. A low-molecular
weight resin in an aqueous solution can easily penetrate into the cell wall,
whereas a resin of larger molecular size would be trapped at the lumen
surfaces of the cell walls [1]. As a result, dimensional stability and
biological resistance can be more easily achieved with low-molecular
weight resin.
Acetylation
A number of studies on wood acetylation have been reported in the
past, and acetylated wood is said to have excellent physical and biological
properties related to dimensional stability, and resistance to decay and
termite attack. Acetylation occurs in the reaction with acetic anhydride,
in which acetic acid is split out as a by-product. The hydroxyl groups on
cellulose, hemicellulose, and lignin are converted into hydrophobic acetyl
groups.
Kiguchi and Suzuki [3] have made particleboards from acetylated sugi
particles without a binder, the particles being acetylated by acetic
anhydride with a catalyst of sulfuric acid. The particle surfaces with
acetylated wood components were melted by applied heat, and the
particles were bonded. However, under the optimum conditions for
acetylation, the particle surfaces were darkened, and the wood compo-
nent deteriorated by the hydrolytic reaction of the sulfuric acid. TS of the
acetylated particleboards was reduced considerably, but the IB strength
was one-fifth that of phenol-formaldehyde resin-bonded particleboard.
The mechanical properties and dimensional stability of boards made

from albizia (Albizia falcata Backer) particles, which had been acetylated
by acetic anhydride without a catalyst, were affected significantly by the
acetyl weight-percent-gain (WPG) depending on the type of adhesive [ 4 ] .
The bending properties of isocyanate (IC)-bonded boards decreased with
increasing WPG, whereas no serious effects were observed with phenol
formaldehyde resin (PF)-bonded boards [4]. The creep compliance of the
boards decreased with increasing WPG, the effects of acetylation on creep
compliance occurring with decreasing WPG in the PF boards compared
with the IC boards [5], The IC boards suffered from a considerable
decrease in IB strength with increasing WPG up to 12%, although no
significant effect of increasing WPG on the IB strength of PF boards was
noted. Dimensional stability was improved with increasing WPG for
both the IC and PF boards. This may have been due to the fact that the
less-hygroscopic groups in wood cause less water to be absorbed, and
that most of the irreversible stresses do not occur in boards made from
acetylated particles [4]. The acetylated boards with 20% WPG were
highly resistant to decay by brown-rot fungus Tyromyces palustris,
white-rot fungus Coriolus versicolor and soft-rot fungus Chaetomium
globosum. Acetylation significantly enhanced their termite resistance
based on laboratory tests and field exposure in wet tropical areas [6].
On the other hand, Nishimoto and Imamura [7] made particleboards
from mixtures of control and acetylated spruce (Picea jezoencis Carr.)
particles. The mixture of acetylated particles (the average acetyl content
was about 17%) reduced the IB strength of the boards, and the IB values
were markedly decreased when bonded with urea-melamine and phenol-
melamine formaldehyde resins. However, the IC resin retained relatively
high gluability, even though the ratio of acetylated particles amounted to
100%. There was a trend of decreasing value of TS and water absorption
(WA) after soaking in water for 24 hours as the amount of acetylated
particles increased, and the fully acetylated boards exhibited 2 to 3 % of
TS and values below 20% in WA. The equilibrium moisture content of
the acetylated boards in the air-dried condition was reduced to the range
of 3 to 5%, and under moisture-saturated conditions, was below 10%.
Soil block tests with brown-rot fungus Tyromyces palustris and white-
rot fungus Coriolus versicolor showed very slow decay in those boards con-
taining up to 50% of acetylated particles, and no decay in the boards made
only from acetylated particles [4, 6]. The boards made with 75% of acety-
lated particles were resistant to attack by Reticulitermes speratus termites,
although less resistant to attack by Coptotermes formosanus termites [ 5 ] .
Yoshida et al [8] have made low-density particleboards from acetylated

seraya (Shorea spp.) particles, which had been treated with acetic anhydride
to produce a 16% weight gain in the acetic group. Acetylated, non-
acetylated, and mixed (50:50 weight ratio) particles were pressed into
low-density particleboards (0-4 and 0-5 g/cm ) using an isocyanate resin
3
adhesive. The MOR was decreased slightly by the acetylation treatment,

because the dimensions of the particles were changed during the acetylation
process and the compaction ratio of the treated boards was consequently
reduced. However, there was no difference in IB between the acetylated and
non-acetylated boards at approximately the same density level.
The acetylated particleboards had high dimensional stability, and
hysteresis in TS rarely occurred after and during immersion in water for a
24-hour period. This means that the acetylated low-density particle-
boards performed like solid wood. TS after 6 cycles of the ASTM
accelerated aging test was less on the acetylated boards than in the
non-acetylated boards (Fig. 2). Twelve-week soil block tests showed very
slow decay in those boards containing 50% of acetylated particles, and
no decay in the boards made only from acetylated particles. When
exposed to fungal attack during bending-creep tests, the low-density
particleboards (0-5 g/cm ) containing up to 50% of acetylated particles
3
maintained their strength for longer than the control boards. The boards
made only from acetylated particles showed little strength reduction after
100 days, and any additional deflection after longer exposure was
minimal. Although the weight loss from termite attack was not as large as
Acetylated particle
c o n t e n t (%]
40
5
30
c/>20
10
I I I i I I I
0 1 2 3 4 5 6
N u m b e r of c y c l e s
FIG. 2. Change of thickness swelling (TS) of acetylated particleboards during

the A S T M 6-cycle accelerated aging test.
that in the controls, the acetylated boards were attacked by Coptotermes

formosanus termites, although they were hardly affected by Reticulitermes
speratus termites [ 9 ] .
Wood modification techniques are currently creating interest as a way
to prevent biodeterioration. The toxicity of conventional wood preserv-
atives is causing growing environmental concern, and efforts to develop
non-toxic methods for preserving wood have intensified all over the
world. Chemical modification of wood is a potentially good alternative to
conventional wood preservation [10], and modification by acetylation
[11] has received considerable attention.
Other Chemical Treatments

Fujimoto et al. [12, 13] have made particleboards from larch (Larix
leptolepis G.) strands treated with a maleic acid and glycerol mixture
(MG). This method combines three elements: (1) the use of a three-
functional group compound (glycerol) and a two-functional group com-
pound (maleic acid), which form a three-dimensional polymer when
reacted; (2) the use of 25% excess dicarboxylic acid, which reacts with the
hydroxyl groups in the cellulose; (3) the use of a high temperature (over
200 C) to free the carboxylic acid to react with the hydroxyl groups
without a catalyst. TS and WA after soaking in water for 96 hours were
greatly reduced with increasing M G weight gain. The strength properties
of these particleboards, especially their modulus of elasticity and IB
strength, were superior to those of non-treated boards [13]. However, the
decay resistance of MG-modified giant ipil-ipil (Leucaena leucocephala
Lam. de Wit) particleboards was lower than that of propylene oxide- or
acetic anhydride-modified boards [14].
Self-bonding by hot-melted surfaces was apparent in the etherification
processes of benzylation, allylation and cyanoethylation [15]. Especially,
benzylation provided the best bonding shear strength in tests of the three.
Hot-melt bonded board manufactured by benzylated akamatsu (Pinus
densiflora Sieb. et Zucc.) particles had as good bending strength as
commercial particleboard, with superior water-resistance and dimen-
sional stability [15].
Yamaguchi et al. [16] have investigated the effect of wood-surface
activation on the IB strength of particleboards. Sugi particles were
activated by such oxidizing agents as H N 0 , C H C O O O H , H 0 , and
3 3 2 2
a ( N H ) 2 S 0 - C u S 0 4 mixture. Boards were made up from the surface-

4 2 8
activated particles with Mimosa-tannin adhesives. They found that am-

monium persulfate was the best agent for activating the wood surfaces, an
adequate amount of oxidizing agents for wood-surface activation being

1-5-3% of the wood. The dry IB strength values of the boards made from
activated wood were 1*3 18 times those of the control boards.
The effectiveness of trioxane and tetraoxane as a reagent source for the
formalization of solid wood has been investigated in a sulfur dioxide-
catalyzed reaction system [17, 18]. The antiswelling efficiency of the
specimens treated with tetraoxane was increased to nearly 70%, this
value being greater than that obtained by using a formaldehyde mono-
mer of trioxane. S0 -catalyzed formalization with tetraoxane was the
2
most effective among the various formalization methods for improving

the dimensional stability and retention of strength. This treatment
method has been applied to medium-density fiberboard (MDF) [19]. TS
of tetraoxane-treated M D F after immersing in water for 5 days was half
that of the control boards, and the bending strength after immersing in
water was dramatically improved.
REFERENCES
[1] H. Kajita, Mokuzai Gakkaishi (J. Japan Wood Res. Soc), 35, 406 (1989).
[2] H. Kajita and Y. Imamura, Wood Sci. Techn., 26, 63 (1992).
[3] M. Kiguchi and M. Suzuki, Mokuzai Gakkaishi, 3 1 , 200 (1985).
[4] B. Subiyanto, S. Yusuf, S. Kawai and Y. Imamura, Mokuzai Gakkaishi, 35,
412 (1989).
[5] S. Takino, M. Norimoto, S. Kawai and H. Sasaki, Mokuzai Gakkaishi, 35,
625 (1989).
[6] S. Yusuf, M. Takahashi, and Y. Imamura, Mokuzai Gakkaishi, 35, 633
(1989).
[7] K. Nishimoto and Y. Imamura, Mokuzai Kogyo (J. Wood Technol. Assoc.,
Japan), 40, 414 (1985).
[8] Y. Yoshida, S. Kawai, Y. Imamura, K. Nishimoto, T. Sato and M. Nakaji,
Mokuzai Gakkaishi, 32, 965 (1986).
[9] Y. Imamura, K. Nishimoto, Y. Yoshida, S. Kawai, T. Sato and M. Nakaji,
Wood Res., No. 73, 35 (1986).
[10] R.M. Rowell, For. Prod. Abstr., 6, 363 (1983).
[11] Y. Imamura and K. Nishimoto, Wood Res., No. 72, 37 (1986).
[12] H. Fujimoto, T. Anazawa and K. Yamagishi, Mokuzai Gakkaishi, 33, 610
(1987).
[13] H. Fujimoto, T. Anazawa and K. Yamagishi, Proceedings of 23rd Interna-
tional Particleboards/Composite Materials Symposium, ed. by T.M. Mal-
oney, Washington State University, Pullman WS, 95 (1989).
[14] V.C. Mallari, Jr., K. Fukuda, N. Morohoshi and T. Haraguchi, Mokuzai
Gakkaishi, 36, 139 (1990).
[15] M. Kiguchi, Mokuzai Gakkaishi, 36, 867 (1990).
[16] H. Yamaguchi, M. Koga, M. Higuchi and I. Sakata, Mokuzai Gakkaishi, 36,

725 (1990).
[17] K. Minato, R. Yasuda and H. Yano, Mokuzai Gakkaishi, 36, 860 (1990).
[19] K. Minato, M. Norimoto, H. Sasaki, A. Yamamoto and N. Sawada,
Proceedings of 40th Annual Meeting of Japan Wood Research Society,
Tsukuba, 338 (1990).
Estimation of the Fungal Resistance of
Wood Composites for Structural Use
YUJI IMAMURA
ABSTRACT
Fungal decay considerably reduces the strength of wood composite materials such as
particleboards and flakeboards with only slight weight loss, so that it is not proper to
estimate their fungal resistance by only using the value of weight loss as an index. A
testing method that combined bending deformation and decay hazard was applied to
wood-based boards to evaluate their mechanical performance under fungal attack. The
rapid reduction of mechanical strength during the early stages of decay is assumed to have
been due to active fungal invasion of the surface of wood elements and consequent glue
failure. The structural performance of boards under biological attack is assessed in
relation to the manufacturing conditions and type of treatment used as well as to the
species of fungi applied. Scanning electron-microscopic observations of the fractured
surfaces were conducted to present visible evidence of glue-line failures due to fungal
invasion and their prevention by introducing a more adequate bonding or glue-additive
treatment.
Keywords: Particleboard, fungal attack, glue-line, bending-creep, fracture, SEM
INTRODUCTION
The increased utilization of wood composite boards for structural use

in such applications as the underlay of flooring materials and as wall
panels dictates the need for a better understanding of the detrimental
effects of wood-destroying fungi, and for the development of preventive
measures in controlling the problem. Although wood composite boards
such as particleboards are generally considered less susceptible to bio-
deterioration than solid wood, their resistance can be affected by the
species of wood in the board, the type of adhesive used as a binder, the
75
76 YUJI IMAMURA
size and shape of the particles, and the composition and physical
properties of the board [ 1 ] .
Much of the laboratory test data published on the decay of wood uses
a method based on the weight loss after a certain period of time in the
presence of single or mixed organisms. It is known, however, that the
brown-rot fungal attack on solid wood results in a large strength loss
with very low wood weight loss [ 2 ] . Consequently, the strength reduction
in a particleboard due to decay cannot be assumed to be always
associated with weight loss in the specimen.
A test method has been developed to determine the strength loss in
wood while under attack by brown- and white-rot fungi. This test method
determines the deflection under load by bending creep apparatus when
coupled with fungal attack [3, 4 ] .
To evaluate the mechanical performance of particleboards when ex-
posed to wood-decaying fungi, commercial particleboards and two kinds
of treated ones, which were made from acetylated chips and treated with
a fungicide incorporated in the glue, were subjected to bending-creep tests
under progressive fungal attack.
A scanning electron microscope (SEM) was employed to observe the
fractured surfaces, and morphological characteristics of the decayed
particleboards were presented in relation to the board quality and the
species of fungi inoculated into the specimens.
BENDING-CREEP TEST U N D E R F U N G A L ATTACK
The test boards were subjected to bending-creep tests under fungal

attack in a newly designed decay chamber (Fig. 1). This chamber consisted
of a metal-wire frame covered with a polyethylene bag and sealed with a
porous plug of silicon rubber. A brown-rot fungus Tyromyces palustris
(Berk, et Curt) Murr and a white-rot fungus Coriolus versicolor (L. ex Fr.)
Quel, which are standard strains for Japanese Industrial Standards (JIS),
were used for decay attack. Mycelial fragments prepared from a shake
culture were spread on the bottom tension surfaces of sterilized board
materials. A tray filled with sterilized water was placed in the bottom of the
chamber to keep the specimens in a moist condition.
A load was applied at the center of each specimen from outside the
decay chamber, this applied load being sufficient to cause an initial
deflection equivalent to 1/300 of the span length. The value of the load
was approximtely equivalent to 10% of the wet bending strength of the
FUNGAL RESISTANCE OF WOOD COMPOSITES FOR STRUCTURAL USE 77
FIG. 1. A testing device of bending-creep measurement coupled with fungal

attack. Weight is hung from outside the decay chamber consisting of a wire frame
and a polyethylene bag.
particleboards. The subsequent deflection of each board at the center of

the span (300 mm) was measured regularly with an electric dial-gauge as a
criterion to determine the performance. The bending-creep test was
carried out until the specimens broke under the load or for a longer time
of more than three or six months. The test apparatus was maintained in a
conditioned room at 26 C, which was suitable for the incubation of the
fungus. After failure or a fixed time, the oven-dry weight loss was
determined for each composite board.
78 YUJI IMAMURA
Since creep in wood is dependent on the moisture content, it is

important to determine the creep due to the moisture present in the wood
separately from the creep due to a strength loss associated with decay.
One set of specimens had liquid inoculum applied to bottom tension
surface, while a second set had an equal amount of sterile water applied
to the bottom side. The wet sterile controls showed a deflection of less
than a few millimeters during the 200-day test, while the specimens with
fungal decay failed within the 200-day test showing a large deflection.
This shows that only a small part of the observed deflection was
associated with creep due to moisture.
The deflection-time curve for particleboards exposed to fungal decay
shows a typical pattern comprising an initial increase of deflection, a
stable zone, and a final rapid increase in the slope leading to fracture. N o
difference was apparent in initial deflection between the inoculated
boards and sterile wet controls; however, the rapid increase in deflections
was only observed with the former boards. This difference is assumed to
have been due to the strength reduction caused by the fungal action. The
board resistance detected in the creep tests under fungal attack varied
depending on the manufacturing conditions, which included preservative
treatments, as well as on the fungal species.
PERFORMANCE OF COMMERCIAL PARTICLEBOARDS

UNDER FUNGAL ATTACK
Figure 2 shows the deflection-time curves of commercial particle-

boards bonded with urea-melamine-formaldehyde (UMF) resin (ab-
breviated as M-type board) and phenol-formaldehyde (PF) resin (P-type
board). The two types of M board show a different behavior in their
deflection curves according to the board density and particle size; the M l
board had a lower density and larger particle size than the M2 board. A
rapid increase in the deflection of the M l board occurred relatively early,
and the board failed within the shortest period of 35 days with a slight
weight loss. Large particles in a low-density board seem to have promo-
ted the fungal growth, which caused the rapid increase in deflection
during the early period of incubation.
Figure 3 shows the fractured surfaces of the M l board, in which active
colonization of the fungus was detected prior to wood decay. The SEM
analysis showed that mycelial invasion was initiated within a short
period, and the rapid development of colonization occurred. The initial
20r~
6.7
E
E M2 t
88
3.3
c
M . !
g
o
a>iO|
*-
Q
Sterile wet controls
20 40 60 80 100
Loading time (days)
FIG. 2. Deflection-time curves for commercial particleboards bonded with
UMF resin (Ml and M2) and PF resin in a creep test under fungal attack by T.
palustris. Arrows indicate fracture of the test boards, and the number above each
arrow shows the value of the weight loss at the time of fracture.
colonization of the fungal mycelium was established mostly by invasion

of the surface of the wood particles, which served as the main route of
fungal growth. Mycelia and strands may have developed on the particle
surfaces and penetrated deeply into the inner parts of boards to find a
new source of nutrition or a congenial environment. Active fungal
invasion of the surface of the wood particles is assumed to have caused
the strength reduction with little weight loss consequent upon glue
failure.
The P board carried the load for a longer period than the M l and
M2 boards, and failed within 95 days (Fig. 2). Some observations of the
P board in the early stages of fungal exposure showed that the extent of
mycelial colonization was significantly smaller than that in the M
board. After exposure for 70 days, the P board was estimated to have a
weight loss similar to that of the M2 board, but was unique in that it
continuously carried the load, whereas the M 2 failed. The P board
ultimately failed after 25 more days, at which time the weight loss
80 YUJI IMAMURA
FIG. 3. Scanning electron micrograph of the fractured surface of the Ml

particleboard after exposure to T. palustris.
reached more than 10% and the wood particles were substantially
destroyed.
The adhesives used in manufacture are of greater importance for the
fungal growth, as urea-melamine and isocyanate (IS) resins can have a
promoting, and phenolic resins an impeding effect. It has been reported
that unleached phenol-formaldehyde boards showed some toxicity to a
brown-rot fungus in sand-block tests. The presence of some soluble toxic
materials in PF resins may inhibit or slow the attack by T. palustris. This
may account for the longer time to failure of the PF boards compared
with the U M F and IS boards. However, these results do not necessarily
indicate superior performance from untreated PF boards in service, since

the free phenols are easily leached away in water.
Although the primary mode of fungal attack varied according to the
board quality, fractured specimens of the undecayed controls demon-
strated a bristling surface of the wood particles which had been torn off
[5]. On the contrary, the severely decayed boards infested with the
brown-rot fungus had particles which were so fragile as to permit clear
breaks in the transverse as well as in the longitudinal direction.
The white-rot fungus resulted in a longer time for creep fracture and
caused a higher weight loss in the sample boards before inducing a
substantial strength reduction as compared with the brown-rot fungus.
SEM observations of the fractured surfaces of the board decayed by C.
versicolor showed degradation of the cell walls of the wood particles in
the form of voids around pit openings and thinning of the cell walls
accompanied by fungal colonization. These observations were confirmed
by the bending behavior, which showed a slow increase in deflection and
a large weight loss when fractured. It is concluded that the white-
rot fungus had less ability to break down the glue bonding and depended
on decomposition of the wood substances for fungal growth.
PERFORMANCE OF CHEMICALLY MODIFIED

PARTICLEBOARDS UNDER FUNGAL ATTACK
Small particles of spruce (Picea jezoencis Can.) were used for

acetylated particleboard production. The particles were acetylated with
acetic anhydride using sodium acetate as a catalyst to an average acetyl
weight gain of 17%, based on the original oven-dry weight. Control and
acetylated particles were pressed into particleboards, using 8% U M F , PF
or isocyanate (IS) resins.
The acetylated particleboards showed very little weight loss in both a
standard sand-block test with brown- and white-rot fungi, and unsterile-
soil burial tests with brown-, white-, and soft-rot fungi, and with wood-
attacking bacteria [6, 7, 8].
When exposed to fungal attack in bending-creep tests, all the in-
oculated pine control boards bonded with U M F resin failed within 30
days from brown-rot fungus T. palustris, and within 75 days from
white-rot fungus C. versicolor, while the acetylated pine boards did not
fail within 200 days from the white-rot fungus, but did fail within 50 days
from the brown-rot fungus (Fig. 4).
82 YUJI IMAMURA
6.6
A 17.8
8.8
I !!
E
20
Deflection
0)
Q 10

J I I I I 111 J I I I I I 111
1 10 100
Loading time (days)
FIG. 4. Deflection-time curves for U M F ( A , A)-, P F ( , ) - , and IS ( O ,
)-bonded particleboards made from untreated ( A , , O ) and acetylated (A, ,
) particles in bending-creep tests under attack by T. palustris.
Electron micrographs of all the failed surfaces from the brown-rot

fungal tests showed that the U M F resin was attacked by the fungus,
which caused the boards to fail. In the control boards, the wood was also
attacked, while in the acetylated boards, no attack on the acetylated
particles was apparent. Even though the acetylated particleboards lost
strength during brown-rot fungal attack, these boards had lost less than
1 % in weight at the time of board failure.
The time to failure for the IS control boards ranged from 16 to 28 days,
and from 48 to 55 days for the PF control boards under brown-rot fungal
attack. Even though these ranges are quite large, the deflection-time
curves for a given set of experimental conditions are similar. On the
contrary, the IS and P F boards made from acetylated particles show a
very small deflection after more than 200 days.
Mycelia of T. palustris fully covered the surfaces of all boards within one
week, using medium-nutrition contained in the mycelial fragments in-
oculated. The mycelium could invade the inner part of the control boards,
and UMF-bonded acetylated boards treated at a sufficient level to provide
biological resistance. Rapid mycelial invasion was observed in those

boards, however, this being slower for colonization in the acetylated boards
bonded with the PF or IS resins. The colonization by decay fungi was also
detected by SEM in the void spaces of the acetylated boards bonded with
the U M F resin, and it was noted that fungal mycelia developed only on the
glue lines, which would cause a strength reduction. The proportion of
glue-line failures increased after exposure to decay attack in the control
boards, while acetylation reduced the attack on the glue-lines as well as
reducing decay of the wood particles when PF or IS adhesives were used.
Although the mixture of acetylated particles maintained their strength
longer than the control boards did, blended boards containing acetylated
particles at 50% did fail. Particleboards made from 75% acetylated
particles and 25% untreated ones could carry the load for more than 200
days with a small deflection when bonded with IS or PF resin [ 6 ] .
PERFORMANCE OF GLUE-ADDITIVE-TREATED
PARTICLEBOARDS UNDER FUNGAL ATTACK
When particleboards are to be used under humid conditions, it is

generally agreed that they must be protected from biodeterioration by the
addition of a suitable preservative. Adding the preservative to the
adhesive has usually been the method used [5, 9, 10], and can be expected
to be a practical means for increasing the decay resistance.
Treated particleboards were prepared from undurable spruce chips and
U M F resin or IS resin containing 0-6, 0-8 or 1 0 k g / m of a fungicide (a
3
triazole compound) [ 5 ] . The untreated control board failed within 10-20

days under fungal attack by T. palustris. In contrast, the treated boards
containing 0-6 k g / m of fungicides tolerated the attack for longer than the
3
untreated controls and failed by 60 days from loading. The particleboards

treated with the fungicide incorporated in the glue at a retention level of
1 0 k g / m showed only a very small deflection (almost the same as sterile
3
wet boards) even after 200 days (Fig. 5).

Mould was able to grow on the particleboards and deteriorate the glue
bonding even when no decay was found in the wood particles. It was
suggested that the glue-additive treatment with preservatives would be
effective against mould as well as fungal decay [10].
The incorporation of fungicides into the glue is considered to be more
effective against fungal growth on the surfaces of the particles, possibly
because of the more even distribution of toxicant chemicals.
84 YUJI IMAMURA
20r 8.4
E 3.5
E
I
C
o>
0)
Q
0 I I I l I M i l l I I I I I I I I I I
1 5 10 50 100
Loading time (days)
FIG. 5. Deflection-time curves of untreated low-density particleboards () and
those treated by incorporating azaconazole in IS resin at retention levels of 0 6
(A) and l-0kg/m (). 3
REFERENCES
[1] Y. Imamura, Mokuzai Kogyo (J. Wood Technol Assoc., Japan), 42, 306
(1987).
[2] Y. Imamura and K. Nishimoto, Mokuzai Gakkaishi (J. Japan Wood Res.
Soc), 30, 1027 (1984).
[3] M. Takahashi, Int. Res. Group on Wood Preserv., Document No. IRG/WP/
2163 (1981).
[4] Y. Imamura and K. Nishimoto, J. Soc. Mater. Sci., Japan, 34, 985 (1985).
[5] Y. Imamura and K. Nishimoto, Proceedings of Pacific Regional Wood
Anatomy Conference, 37 (1984).
[6] Y. Imamura and K. Nishimoto, Mokuzai Gakkaishi, 33, 25 (1987).
[7] R.M. Rowell, J.A. Youngquist and Y. Imamura, Wood and Fiber Sci., 20,
266 (1988).
[8] Y. Imamura et al, Paperi ja Puu, 7, 816 (1989).
[9] H. Motoki, M. Sagioka and M. Sakai, Mokuzai Gakkaishi, 31, 193 (1985).
[10] K. Fukuda, Y. Tominaga and T. Haraguchi, Mokuzai Hozon (Wood
Preservation), 15, 10 (1989).
Full-Scale Tests on Wooden Houses During
Progressive Stages of Construction
TAKANORI ARIMA
Department of Forest Products, Faculty of Agriculture,

The Tokyo University, 1-1-1 Yayoi, Bunkyo-ku, Tokyo 113, Japan
ABSTRACT
Full-scale tests on houses during progressive stages of construction were conducted with
a vibration generator to determine the influence of components, and the relationships
between the natural frequency and the racking resistance of a house is discussed. The
changes in natural frequency during construction are significant and related to the stiffness
of the load-bearing walls and the damping ratio of the structure. Since the relative stiffness
depends on the amount of lining and cladding materials, as well as on the load-bearing
walls, the non-load-bearing materials play an important role in providing additional
resistive elements. This result was confirmed by the fact that gypsum cracking at the corner
of a wall was initially observed in the full-scale tests and from actual earthquake damage.
The effects of the number of stories, unbalanced walls and floor openings on the
frequency-response curve and the natural frequency are also discussed.
Keywords: Full-scale tests, wooden house, natural frequency, racking resistance, shear wall,
non-destructive testing
INTRODUCTION
Testing full-scale wooden structures is similar in many respects to

testing structures made with other materials, although the wooden
structure for residential houses has many unique aspects that must be
considered. The variety of designs and components that constitute the
materials and joints make actual testing of many types of residential
houses of great importance to understand how a completed structure
performs in use. Many full-scale tests by applying a horizontal load on
various types of wooden houses have recently been conducted in Japan
85
86 TAKANORI ARIMA
[10, 11]. As static tests of full-scale structures are difficult and expensive
to conduct, it is necessary to develop a non-destructive testing method
that is easy to apply on site, and to measure the dynamic behavior and
stiffness of full-scale houses during progressive stages of construction, and
during the stages of demolition.
The primary purpose of this report is to investigate the relationship
between the stiffness of a full-scale structure and its dynamic natural
frequency, and to evaluate the effects of the construction building stages
and components in providing racking resistance to lateral loads. Based
on the results observed with actual houses, the effects of the number of
stories, unbalanced walls and floor openings on the frequency-response
curve and the natural frequency will also be discussed.
FREQUENCY-RESPONSE CURVE F R O M A PORTABLE

VIBRATION GENERATOR
In order to produce the frequency-response curve for a full-scale house,

a portable vibration generator of 40 kg in weight was set close to the
center of the structure, and accelerometers were located in various
positions.
Figure 1 shows various response curves [8]. The frequency response
and natural frequency depend on the number of stories, height and
components of the house. The changes in natural frequency during the
.25
Conventional 2-story structure
.2
I
.15
-P
C
.i I
U
(0
a
w .05
a
"0 5 10 15
FREQUENCY (HZ)
FIG. 1. Frequency-response curves and natural frequency during progressive
stages of construction. (I: diagonal bracings; II and III: + gypsum board cladding;
IV: + siding).
TESTS ON WOODEN HOUSES DURING STAGES OF CONSTRUCTION 87
progressive stages of construction were significant and related to the

stiffness of the bearing walls.
It was also found that the natural frequency of free vibration when
giving a sharp jolt to the structure was approximately equal to that
obtained by the vibration generator method.
SHEAR WALLS A N D N O N - L O A D - B E A R I N G WALLS
The main bearing walls for the conventional Japanese structure are
diagonal-braced frames. Increasing the diagonal bracing provided in-
creased racking resistance and natural frequency in the test house.
Gypsum board shear walls of a light-frame construction as internal
bearing walls also provided a significant bracing effect proportional to
the number of walls as shown in Fig. 2. These results show that increasing
the number of bearing walls provided increased natural frequency in the
full-scale house, as was reported for wood structure models in previous
papers.
8
Natural frequency (Hz)
~0 30 60 90 120
Length of wall 1 P (P=910mm)
FIG. 2. Natural frequency from increasing-length gypsum board shear walls of

light-frame construction. The length of wall is typical of a two-storied house with
a total floor area of 132 m 2
The natural frequency of a house after installing lining or cladding

materials as non-load-bearing walls might be expected to reduce due to
the increasing weight of the structure. But as shown in Fig. 3 [11], the
change in natural frequency of a one-story house during demolition by
reducing the number of gypsum-plaster cladding walls indicates that the
88 TAKANORI ARIMA
10
(Hz)
9
Natural frequency
8
.
0 1.0 1.2 1.4 1.6
Racking load (x 1 0 3
kgf)
FIG. 3. Change in natural frequency of a one-story house during demolition by
reducing the length of gypsum-plaster cladding walls.
lining and cladding materials could significantly contribute to increasing

the natural frequency.
The effect of lining and cladding materials on the natural frequency
during progressive stages of construction is shown in Fig. 4 [11].
Although the relative stiffness naturally depends on the amount of
lining and cladding materials and on the specification [5, 11], it can be
seen that these non-load-bearing materials provide additional structural
stiffness [2, 7, 8, 12]. In addition, the damping ratio of a wooden structure
generally tends to be increased with increasing lining materials.
RELATIONSHIP BETWEEN THE N A T U R A L F R E Q U E N C Y

A N D STATIC STIFFNESS
Apparent spring constant X, which is related to the stiffness of a

structure against a horizontal load, can be calculated by the equation of
a model having one degree of freedom:
K= (2n-f) -W/g
2
where / is the frequency, W is the weight, and g is 980 cm/sec . 2
In an earlier report on a model of a wooden structure, the relationship

between static stiffness K , which is expressed as the ratio of the load to
s
deformation, and apparent stiffness K calculated from the natural fre-

quency was shown to be linear. However, stiffness K depends on the s
Light-frame structure
8 ,, EW-direction
SN-direction
6 a-
'ft.
4
i
r
1 1
N
o
2
0
i
I 1p
V/,
f
c I II III
0)
D
CD 3-story structure
I
8 EW-direction
CO* SN-direction
6
-P
CO
2 4
2
0
i ii III
Stage of construction
FIG. 4. Effect on the natural frequency of lining and cladding materials during
progressive stages of construction. Light-frame structure: I: plywood shear wall;
II: + gypsum board shear wall; III: + siding. 3-stories conventional structure: I:
diagonal bracing: II: + gypsum board cladding; III: + siding).
amount of deformation, because the load-deformation curve is not linear.

Figure 5 shows the relationship between K obtained from vibration tests
and K from the resistance against a horizontal load for a full-scale house.
s
Since the deformation resulting from vibration is minute, the apparent

stiffness K tends to be less than static stiffness K as described in the
s
earlier report. When actual values for K are estimated from X, the
s
coefficient for estimation must correspond to the amount of deformation.

The relationship between K obtained from vibration tests on full-scale
houses and K estimated from racking-resistance test on the constituent
s
walls is shown in Fig. 6. In previous reports [3, 6, 9 ] , it was stated that

the resistance against a horizontal load of a full-scale house was 1-21*7
times as much as the resistance estimated from racking-resistance tests of
a wall.
90 TAKANORI ARIMA
Spring constant, K (x 10 kgf / cm) 12

/ 8/ z
3
10
X J 3/
8
4,
0 2 4 6 8 10 12 14 16
Racking load (x 10* kgf)

CO
FIG. 5. Relationship between K obtained from vibration tests and the racking
resistance to a horizontal load of a full-scale house.
kgf / cm)
15 r
3
Spring constant, K (x 10
10 -
O
FIG. 6. Relationship between K ob-
tained from vibration tests and the
0 10 20 30 ultimate strength estimated from
Ultimate strength (x 1 0 3
kgf) racking resistance test on a wall.
LOCATION OF LOAD-BEARING WALLS, FLOOR O P E N I N G S

A N D N U M B E R OF STORIES
The mode of vibration is influenced by the location of the load-

bearing walls and an unbalanced walling in the structure could cause
torsional deformation. In particular, large floor openings have been
identified as the cause of severe torsion at the second resonant fre-

quency [11].
The number of stories of a structure has an effect on the natural
frequency and mode of vibration, because of the increasing weight and
distribution of stiffness. The lowest frequency, called the first mode, can
be calculated by using a system having multi-degrees of freedom as the
model for a multi-storied wooden structure. When structures have a
similar plan for each story, the natural frequency of a three-storied
structure is lower than that of two or one-storied structure as shown in
Fig. 7 [11, 12].
10
Natural frequency (Hz)
4 Stage II
2
8 1
^ 6 Stage I
1 2 '3 4
Number of stories
FIG. 7. Effect of the number of stories on natural frequency.
D A M A G E OBSERVED F R O M FULL-SCALE TESTS A N D

F R O M A N ACTUAL EARTHQUAKE
The damage from full-scale tests has been recognized to be similar to

observed earthquake damage. A newly built house which complied with
building regulations was hardly affected in a recent earthquake, although
cracking of gypsum at the corners of walls and cracking mortar were
found as typical damage in conventional wooden houses. It is to be noted
that degree of damage depended on the total load-bearing wall strength,
which is represented by the product of the wall length and the resistance
factor in Japanese building regulations as shown in Fig. 8 [10]. As
described previously, current house construction can provide various
methods for load sharing that will result in the house having a greater
92 TAKANORI ARIMA
100
COLLAPSE OF EXTERNAL WALLS
50 L\ \ \ N CRACKING OF EXTERNAL WALLS
CRACKING OF FOUNDATION
I n IV
<1.2 <1.5 <2.0 2.0<
Ratio of damage {%)
100
CRACKING OF INTERNAL WALLS
50 ILL-FITTING FRAMES
DELAMINATED TILES
i n id iv
100
50 DAMAGED SHOJI (SLIDING PAPER DOORS)
1\ \ \ \| GAPS BETWEEN POSTS AND WALLS
I n ID IV
R e l a t i v e amount of l o a d - b e a r i n g walls
FIG. 8. Damage to houses in the 1978 earthquake expressed as a function of the
relative amount of load-bearing walls (actual amount of bearing walls/required
amount of bearing walls). The ratio of damage is denned as actual number of
damaged houses to the total number of houses investigated. The extent of damage
was slight, needing only minor repairs.
resistance to lateral loading that can be determined from a normal

structural analysis or from testing the load-bearing components in
isolation.
CONCLUSIONS
Non-destructive testing by using a vibration generator was effective for

measuring the frequency-response curve for a full-scale house and for
evaluating the relationship between the natural frequency and the racking
resistance of the house.
Full-scale testing during the progressive stages of construction could be
used to evaluate the contribution of the building elements in providing
racking resistance to lateral loads.
The frequency-response curve and natural frequency depended on the
number of stories, height and components of a house.
Increasing the number of diagonal braced frames and shear walls
provided increased racking resistance and natural frequency in the test
house. Increasing the lining and cladding materials could contribute to

increasing the natural frequency as additional resistance elements.
Apparent stiffness K calculated from the natural frequency was linear
to static stiffness K , but tended to be greater than K . When actual values
s s
of K are estimated from K, the coefficient for estimation must corre-

s
spond to the amount of deformation.

Unbalanced walling and floor openings in the structure can cause
torsional deformation.
ACKNOWLEDGEMENT
Thanks are given to the many researchers and scientists whose work
contributed to this report.
REFERENCES
[1] T. Arima, M. Okabe and N. Maruyama, Mokuzai Gakkaishi (J. Japan Wood
Res. Soc), 34, 813 (1988).
[2] T. Murota, T. Arima, H. Okada, M. Sato, K. Saito and K. Mashita,
Kenchiku Kenkyu Shiryo No. 28 (1981).
[3] Ministry of Construction, Development of a New Wood Construction System
for Small Houses, Technology Center for National Land Development of
Japan (1975).
[4] S. Nakajima and T. Arima, Mokuzai Gakkaishi, 36, 291 (1990).
[5] H. Sugiyama, T. Uchisako, N. Ando, H. Hirano and N. Nakamura, ICTE
Seattle Proceedings, 55 (1988).
[6] H. Sugiyama, T. Uchisako, N. Ando, T. Arima, H. Hirano and N.
Nakamura, ICTE Seattle Proceedings, 804 (1988).
[7] G.F. Reardon, PTEC. Auckland Proceedings, 276 (1989).
[8] Japan Housing and Wood Technology Center, Full-scale Test of Conven-
tional Timber Structure House (1986, 1987).
[9] Japan Wood Frame Housing Association, Full-scale Test of Three-storey
2x4 Construction (1979),
[10] T. Arima, Full-scale Test of Timber Construction Using a Vibration Gener-
tor, The Building Center of Japan (1989).
[11] T. Arima, M. Okabe, N. Maruyama and S. Hayamura, IT EC. Tokyo
Proceedings, 778 (1990).
[12] S. Nakajima, T. Arima, N. Nakamura and J. Ichikawa, Mokuzai Gakkaishi,
37, 128 (1991).
Structural Properties of Full-Size
Sawn Timber
TAKASHI NAKAI

ABSTRACT
For developing more reliable stress-grading methods, full-size bending tests were con-
ducted on square sawn timber. It was found that the concept of 'strength ratio', which is
used in ASTM D 245, could not always be applied to all species, but that the modulus of
elasticity could be used as the most suitable practical parameter for grading regardless of
species.
Various methods were studied for measuring the modulus of elasticity, both static and
dynamic, including the methods of stress-wave propagation time measurement and of
fundamental vibration frequency measurement.
A series of destructive tests was conducted in bending, compression and tension of
full-size sawn timber after measuring the static and dynamic moduli of elasticity. The
moduli of elasticity obtained from the deflection at mid-span in the third point loading
condition and from the fundamental vibration frequency and density are regarded as the
most suitable parameters for grading for the Japanese saw-mill industry. A new grading
system should be developed that applies these parameters for large cross-sections such as
flat square sawn timber.
Keywords: structural timber, stress grading, modulus of elasticity, non-destructive test,

fundamental vibration frequency
INTRODUCTION
Historically, a great deal of effort has been made to obtain basic

strength [1] with the results from small clear specimens, which has been
regarded as a starting point for deriving the allowable unit stress on
graded sawn timber. The grading systems of almost all countries have
95
96 TAKASHI NAKAI
been based upon two important hypotheses:
(a) The strength of sawn timber is proportional to the strength of small

clear specimens of the species used.
(b) The effect of natural growth defects such as knots on the strength of
sawn timber is independent of the species. This hypothesis is called
the 'strength ratio' concept.
Many full-size strength test results have revealed that these two
hypotheses are not always valid. North America and some countries in
Europe have finished what they called an 'in-grade test' project for
dimension lumber, which is from 38 mm to 89 mm in thickness and
89 mm or more in width. There is a growing trend to change from the
allowable unit stress design method to the limit state design method, for
which it is essential to determine the characteristic values for graded sawn
timber.
Unlike other countries, Japanese structural sawn timbers that are used
for construction have a relatively large cross section, and are known as
squares (e.g., 10-5 cm by 10-5 cm in cross section) and flat squares (e.g.,
1 2 0 c m by 2 7 0 c m ) , which resemble North America's 'post and timbers'
and 'beam and stringers' respectively. Very little research work has been
carried out on the structural performance of these sections.
Forestry and Forest Products Research Institute has conducted a
series of full-size bending tests on timber cross sections, in the square and
flat square series. As there was a high correlation coefficient regardless of
species between the modulus of elasticity and various strengths such as
modulus of rupture, compressive and tensile strength, several non-
destructive tests to measure the modulus of elasticity were conducted to
seek a practical application of this parameter at saw-mills or lumber
yards for large cross-section timber, which cannot be evaluated by
commercial grading machines which have been developed mainly for
dimension lumber.
MATERIALS A N D TEST M E T H O D S
(1) Squares bending test. This research was carried out with the
co-operation of nine prefectural research organizations under the supervi-
sion of FFPRI [ 2 ] . The species tested in this project were akamatsu
(Pinus densiflora), Douglas fir (Pseudotsuga menziesii), karamatsu (Larix
STRUCTURAL PROPERTIES OF FULL-SIZE SAWN TIMBER 97
leptolepis), Dahurian larch (Larix gmelinii), hinoki (Chamaecyparis ob-

tsusa), Hem fir (Tsuga heterophylla and Abies spp.), sugi (Cryptomeria
japonica) and touhi (Picea sp.). The total number of specimens tested was
1471 with a size of 10-5 cm x 10-5 cm x 300 cm. The third point loading
system was used with a 270 cm total span and all tests were conducted in
air dry conditions.
(2) Comparison of various modulus of elasticity measurements. By
mainly using sugi and hinoki squares, four methods for measuring the
modulus of elasticity (MOE) were compared [3, 4]: measuring the
deflection at mid-span by applying dead weights, measuring the stress
wave and ultrasonic wave propagating time and density, and measuring
the fundamental vibration frequency and density. The equation for M O E
Ef is as follows:
f= (V2L)(Efg/py'2
where / is the fundamental vibration frequency (Hz), L is the length of the

specimen (cm), g is the acceleration due to gravity (980cm/sec ), p is the
2
density (g/cm ), and Ef is the dynamic longitudinal modulus of elasticity

3
(GPa).
(3) Squares compression and tension tests. This research was conduc-
ted at FFPRI [5, 6 ] . Species tested were akamatsu., Douglas fir, Hem
fir, ezomatsu (Picea jezoensis) and sugi. The size of each specimen was
5 cm x 10 cm x 350 cm, except for sugi, which was 10 cm x 10 cm x
350 cm, and each sample size was about 100, except 200 sugi samples.
A 980 000 N-capacity testing machine was used with 50 cm long tapered
wedge grips, the elongation being measured by digital dial gauges
with 1/100 mm graduations for a 1000 mm gauge length at the center.
Compressive strength was measured for a short column specimen with a
slenderness ratio of 10, which was end-matched to the tensile speci-
men.
(4) Flat squares bending tests. Seven prefectural research organiza-
tions took part in this project under the supervision of FFPRI [ 7 ] . The
species tested were akamatsu, Douglas fir, karamatsu, ezomatsu,
todomatsu (Abies sachalinensis) and sugi. The total number of specimens
tested was 1057 with a typical size of 1 2 0 c m in width and from 2 1 0 c m
to 3 0 0 c m in height. After recording the knot size and location and
measuring Ef the third point loading was applied edgewise with a total
span from 360 cm to 390 cm depending on the length and height of the
specimen.
98 TAKASHI NAKAI
1. The most important result obtained in this series of tests was that
the value for correlation coefficient r between the knot size and modulus
of rupture (MOR) was high enough for grading, being in the order of 0 6 5
for akamatsu, Dahurian larch and Hem fir. In the case of hinoki and sugi,
the values for r were in the oder of 0 1 5 to 0-25, and although statistically
significant, this is not considered to be a practical value for grading.
From these results, it is obvious that the 'strength ratio' concept cannot
be applied to all species. Only MOE showed a high correlation coefficient
in the order of 0-6 to 0-8 regardless of species.
(2) According to the results obtained in this study, all methods for
measuring the modulus of elasticity gave similar values for the correlation
coefficient between the modulus of elasticity and modulus of rupture. For
practical applications, however, a non-contact measuring method, i.e., Ef
obtained by measuring the fundamental vibration frequency and density,
is regarded as the most suitable. It is also pointed out that EF can be
used quite effectively for grading or sorting logs before sawing. Contact
methods such as measuring the stress wave and ultrasonic wave propaga-
tion time required close contact between the sensor and surface of the
timber cross section, which is not easy to maintain. When using ultra-
sonic waves, a large volume of timber may cause too much decay of wave
for its arrival to be detected.
3. When comparing the fifth percentile value for bending, compressive
and tensile strengths, the ratio obtained for sugi was 1 00:0-84:0-64,
respectively. The results of this test indicate that the allowable tensile unit
stress should be reduced to 60% of the allowable bending unit stress.
4. The relationship between Ef and MOR for flat square sawn timber
was established as shown in Fig. 1, from which it appears that sugi and
the other species should be graded separately. These relationships could
be used for allocating new allowable unit stresses in bending when Ef is
measured for flat square sawn timber.
Other valuable research work [8, 9] for simultaneously measuring
modulus of elasticity E and shear modulus G for full-size timber should
be mentioned for the case of long span beams with a very high ratio of
height to width for large-scale timber construction, because estimating G
for full-size timber is very important to prevent lateral buckling of the
beam. Although the ratio of E/G is assumed to be 15 or 16 worldwide, the
cutting pattern and quality of timber may affect this relationship. Further
research needs to be done to clarify this point.
STRUCTURAL PROPERTIES OF FULL-SIZE SAWN TIMBER 99
100
80
MOR (MPa)
60
40
20
0 5 10 15 20
Ef(GPa)
FIG. 1. Relationship between Ef and MOR for flat square sawn timber.
REFERENCES
[1] T. Nakai and R. Yamai, Bulletin of the Forestry and Forest Products R.I., No.
319, 13 (1982).
[2] Forestry Agency, Rep. No. 25, 196 (1985).
[3] T. Nakai and T. Tanaka, The Journal ofCIB, 27(1), 23 (1989).
[4] T. Nakai, T. Tanaka and H. Nagao, Proceedings of 1990 International
Timber Engineering Conference, Vol. II, 593 (1990).
[5] T. Nakai, T. Ebihara and T. Tanaka, Wood Utilization Division Note, 62-1
(1987).
[6] Strength of Timber and Wood-based Structures Group, Structural Timber: A
Collection and Analysis of Strength Data, Japan Wood Research Society,
Tokyo (1988).
[7] T. Nakai, H. Nagao and T. Tanaka, Summaries of technical papers of annual
meeting, Architectural Institute of Japan, Structures II, 5 (1991).
[8] N. Sobue, Mokuzai Gakkaishi (J. Japan Wood Res. Soc), 32, 274 (1986).
[9] N. Sobue, Mokuzai Gakkaishi, 34, 652 (1988).
Structural Analysis of Subassemblies Using
Plywood Panels
FUMIO KAMIYA

ABSTRACT
Plywood panels are used for the construction of all types of wooden residential buildings
in Japan. It is recognized that such buildings are safe from seismic loads, due to the
contribution of plywood for shear strength. The analytical work carried out in Japan for
predicting the behavior of plywood-sheathed walls and floors is summarized.
Keywords: plywood, walling, flooring, non-linear analysis, dynamic analysis
INTRODUCTION
The first production runs of structural-use plywood in Japan took

place in the early 1960s when prefabricated-panel wooden houses were
introduced. Production has increased since the light-frame building
system was imported from North America. Today, structural-plywood is
also used for constructing the post-and-beam type of conventional
houses, the contribution of plywood being to stiffen and strengthen the
structure against lateral forces such as wind and seismic loading.
Besides the prefabricated-panel system where glue is used, plywood
panels are also fastened to the framing by nails. One of the difficulties for
analysis is that the structure shows a non-linear relationship between the
load and the deformation. The behavior of each has been studied by tests
in earlier works and subsequently by analytical methods.
It is now possible to predict analytically the stiffness and the strength
of plywood-sheathed walls and floors from information on the material
and the nailed joint which fastens plywood panels to the framing.
101
102 FUMIO KAMIYA
PLYWOOD-SHEATHED WALLS
When post-and-beam or light weight framing without braces is

sheathed with plywood panels, it provides good resistance to lateral
forces by the contribution of the plywood in shear. Plywood-sheathed
walls are the main shear walls that are acceptable by Japanese building
codes.
The allowable shear is basically evaluated by the racking test. This
test, however, is costly and time consuming, and pioneering work for
predicting analytically the racking behavior of a wall has been carried
out by Tuomi and McCucheon [ 1 ] . Their analysis assumes the principle
of statics and the direction of the nail deformation. Analytical ap-
proaches followed this work in the early 1980s in Japan [ 2 - 5 ] .
With Kamiya's approach [ 3 ] , the relationship between the slip and
force at each nail was formulated under the assumption that the bend-
ing deflection of the framing was negligible. The non-linear load-
deflection curve obtained by repetitive numerical calculation agrees well
with the experimental curve (see Fig. 1). Kamiya also developed an
approximate solution that can predict the non-linear behavior without a
computer.
Pioneering work on a finite element model was done by Noguchi [6],
and mathematical models are now commonly used for the analysis
[7-9].
3500
3000
2500
2000
CO
5 1500
1000
500
0 I 1 1 1 1 1
0 10 20 30 40 50 G0
Shear strain (1CT rad)
3
FIG. 1. Validity of the analytical method for predicting the collapse-resistance

behavior of plywood-sheathed walls.
STRUCTURAL ANALYSIS OF SUBASSEMBLIES USING PLYWOOD PANELS 103
HORIZONTAL PLYWOOD DIAPHRAGMS
Plywood-sheathed floors are important subassemblies which transfer

lateral loads to the seismic-resistant elements of a building such as the
shear walls. They are called horizontal plywood diaphragms.
Plywood diaphragms are basically beams in which lumber on edge
carries the bending component and plywood carries the shear. The
mechanism is almost the same as that of a plywood-sheathed wall if the
ratio of the floor length to the depth is small and the bending deforma-
tion is therefore negligible.
When diaphragms have openings, however, they cannot transfer the
design loads unless suitable allowance is made for the parameters of the
distribution of shear around the opening, and for the stress in the
connection at the corners of the opening.
A finite element analysis that handles the non-linear behavior of a nailed
joint is available to determine these stresses [10]. This analysis, however, is
too complicated for use by designers and does not always suit design
practices. To the author's knowledge, the 'Guidelines for the Design of
Horizontal Wood Diaphragms' [11] published by the Applied Technology
Council (ATC) of the USA is the only guideline which addresses the design
methodology of diaphragms with openings. The ATC analysis assumes that
a diaphragm with openings behaves similarly to a Virendeel Truss. This
analysis, however, still seems to be too complicated for use by designers.
Kamiya [12] developed a method for calculating the shear around
openings with a commonly-used FEM computer program. His analysis is
the membrane-to-brace replacement method shown in Fig. 2. With this
FIG. 2. Brace analysis of a plywood-sheathed diaphragm.

104 FUMIO KAMIYA
analysis, the stiffness of the framing without the braces is assumed to be

so high that deformation can be neglected and the stiffness of the braces
can be determined as equivalent to that of the diaphragm element. The
calculation procedure calls for an algorithm:
Step 1. Calculate the shear under the assumption that the stiffness of
each brace is the same. The equivalent stiffness of braces can then
be determined.
Step 2. Determine the number of nails in each diaphragm element
according to the magnitude of the calculated shear.
Step 3. Recalculate the shear for the new stiffness value.
Step 4. Check if the nail stress is less than the allowable figure. Go back
to Step 2 if the answer is 'No', or end if 'Yes'.
The stress in the connection at the corner of an opening can be

calculated from a summation of the shear.
A static loading test of diaphragms with and without openings was
conducted to verify this analysis. The test results show that the dia-
phragms had a similar ultimate strength regardless of the presence of
openings (see Fig. 3), the load factor (ratio of the ultimate strength to the
design load) being about 2-8 on average. The observed shear around the
openings was almost equal to the calculated values (see Fig. 4). From
these results, it was proved that the design methodology was valid.
SEISMIC RESPONSE ANALYSIS
Several vibration tests on full-scale wooden residential buildings were

conducted by using a vibration generator. The conclusion from such tests
is that the natural frequency of a building is generally almost equal to the
value predicted from the stiffness data of a static test. However, because
the amplitude of such a vibration test is very small, it is impossible to
obtain the seismic response of the structure to the real scale of ground
motion induced by an earthquake from this type of vibration test.
Although it is possible to determine the stiffness and strength of the
structure, the values are not applicable for determining its dynamic
behavior. One analytical method that could be used to obtain the seismic
behavior of a wooden structure is the dynamic response analysis which is
commonly used for steel or reinforced concrete structures. The vibration
equation for a single degree of freedom system is
MX + CX + P=-MX C
14
Total load (tonf)
20 40 60 80 100 120 140
Deflection at midspan (mm)

FIG. 3. Experimental load-deflection curves for plywood-sheathed diaphragms
with and without an opening.
Right: E x p e r i m e n t a l shear 1000kg/m

Left: C a l c u l a t e d shear
FIG. 4. Comparison of the calculated shear around an opening with the

experimental value.
where M is the mass, C is the damping coefficient, P is the resistance of

the structure, X is the displacement, and X is the earthquake ground c
motion.
106 FUMIO KAMIYA
In order to solve this equation, a model of the load-deflection

relationship is essential. Repetitive loading tests on plywood-sheathed
walls, however, show that the relationship between the load and the
deflection is significantly hysteretic and easily affected by the history
of the load and deflection. Therefore, it is questionable whether the
load-deflection model developed for steel or concrete structures can be
used.
Kamiya [13] conducted a pseudo-dynamic test on plywood-sheathed
walls, which is a hybrid method consisting of a dynamic analysis and
loading test conducted simultaneously. The objective of the loading test is
to provide the real resistance of the structure to compute the foregoing
equation.
The test requires an actuator for loading and a computer for controll-
ing the actuator and analyzing the response. Figure 5 shows the system.
Typical results indicate that earthquake waves with a maximum acceler-
ation of 341-7 gal induce a maximum load response of 2-0 to 2-7 times the
allowable design shear, and that the maximum deflection response was
1-07 cm to 1-45 cm.
If a hysteretic load-deflection model that incorporates the change due
to fatigue was developed, the seismic response could be calculated by a
normal dynamic response analysis. Kamiya developed such a model after
careful observation of the load-deflection characteristics obtained from
D/A Actuator
MX+ C X + P = - MXe converter controller
HLoad datah A/D Data ac-

converter quisition
Other d a t a M system
Assumed values|
of M a n d C LVDT
z 1
Earthquake data|
Xe
Test wall
/ Actuator
M a s s s t o r a g e unit ^ - L o a d cell
FIG. 5. Pseudo-dynamic testing system.

the test. This model consists of multiple straight lines passing through
several points such as the residual deflections and loads. The computer
program for this analysis is written in the BASIC language, and about
500 lines are required for the program. The calculated and observed
responses are shown in Fig. 6, indicating that the maximum load and
deflection responses could be predicted with an error of less than 5%,
indicating that this analytical method would be suitable for improving a
design.
Def1.(mm)
20
Pseudo dynamic test

Usual response analysts using the model developed
Load (ton-f)
-2
5 10 15
Time (sec)
FIG. 6. Seismic response of plywood-sheathed walls obtained by the pseudo-
dynamic test and by the usual response analysis using the load-deflection model
developed. Earthquake data is from El Centro, NS, 1940.
FURTHER STUDIES
The hysteresis of wooden diaphragms is mainly due to the hysteretic

characteristics of the nailed joint. Therefore, if information on the
characteristics of nailed joints is obtained, and if a theoretical approach is
developed for predicting the stiffness of a wall which can handle the
non-linear hysteresis loop of the nailed joint, the dynamic response of
wooden diaphragms could be analytically predicted without a full-scale
loading test. Such studies are now under way.
108 FUMIO KAMIYA
REFERENCES
[1] R.L. Tuomi and W.J. McCucheon, J. ofStruc. Div., ASCE, 104 (ST7), 1131
(1978).
[2] I. Sakamoto, Abstracts of Annual Meeting Papers, Arch. Tech. Inst, of
Japan, 2003 (1980).
[3] F. Kamiya, Trans, of Arc. Inst, of Japan, 309, 86 (1981).
[4] Y. Hirashima, Mokuzai Gakkaishi (J. Japan Wood Res. Soc), 27(2), 141
(1981).
[5] M. Yasumura and H. Sugiyama, Abstracts of Annual Meeting Papers, Arch.
Tech. Inst, of Japan, 2253 (1981).
[6] H. Noguchi, Abstracts of Annual Meeting Papers, Arch. Tech. Inst, of Japan,
1793 (1979).
[7] K. Miyazawa and S. Hirano, Abstracts of Annual Meeting Papers, Arch.
[8] M. Hata, S. Takino and H. Sasaki, Mokuzai Gakkaishi, 34(9), 718 (1988).
[9] S. Nishimori and S. Kuromasa, Abstracts of Annual Meeting Papers, Arch.
[10] H. Nasu and H. Noguchi, Abstracts of Annual Meeting Papers, Arch. Tech.
Inst, of Japan, 43 (1989).
[11] Applied Technology Council, Guidelines for the Design of Horizontal Wood
Diaphragms, USA (1981).
[12] F. Kamiya, Proc. 1990 Intl. Timber Eng. Conf, 502 (1990).
[13] F. Kamiya, Proc. 1989 Intl. Conf. on Timber Eng., 838 (1988).
Mechanical Timber Joints and their
Application to Glulam Portal Frames
KOHEI KOMATSU

ABSTRACT
The nail, drift-pin and bolt are representative fasteners to produce moment-resisting
timber joints, and the characteristics of these fasteners are briefly outlined. A method for
non-linear analysis for glulam frame structures is then introduced. Finally, the analytical
approach is verified by full-size experiments on three kinds of two-storey glulam portal
frames utilizing nailed joints, drift-pinned joints and bolted joints, respectively.
Keywords: nail, drift-pin, bolt, moment-resisting joint, glulam portal frame.
INTRODUCTION
A new form of timber construction employing portal frames of straight

glued laminated timber (called 'glulam') with moment-resisting joints has
recently become popular in Japan. Various fasteners and joining tech-
niques have been proposed to achieve the moment-resisting joints. In this
report, the characteristics of representative mechanical fasteners being
used in the moment-resisting joints are introduced and their application
to glulam portal frames is outlined.
MECHANICAL FASTENERS FOR PORTAL FRAMES
Nails
In New Zealand, where the glulam portal frame was first developed
[1], nailed timber joints with steel gusset plates or plywood gusset plates
109
110 KOHEI KOMATSU
are commonly used for the moment-resisting joints. Generally speaking,

a nailed joint with steel side plates gives more stiffness and strength
than with plywood gussets. In Japan, however, the nailed timber joint
has not been considered as a viable method for large-scale timber
construction until recently. Komatsu [2, 3] did a series of experiments
on the performance of single-shear nailed timber joints with steel side
plates to provide fundamental data for designing the glulam moment-
resisting joint with nailed steel side plates. The results obtained in these
experiments [2, 3] can be summarized as follows:
1. There is no linear portion in the load-slip relationship of nailed

timber joints with steel side plates as shown in Fig. 1.
2. For practical design purposes, however, initial slip modulus K s
could be defined as K = (P . ô-2s)/0'25, where P . is the load

S 0 5 0 5
for a slip of 0-5 mm and JP0.25 is the load for a slip of 0-25 mm.
3. For both initial slip modulus K and ultimate strength P of the
s u
nailed timber joint with steel side plates, no statistical significance

was found between the cases of loading parallel to the grain and
perpendicular to the grain.
4. Permissible nail forces evaluated by using the test results are about
30% higher for low-density timber and about 60% higher for
high-density timber than those currently assigned in the timber
design code [ 4 ] .
Drift-pins
In Japan, drift-pins are used for the moment-resisting joints in glulam
portal frames in preference to nails because of the good joint appear-
ance and its excellent fire-resisting performance. The performance of
individual drift-pin joints with inserted steel plate(s) has been in-
vestigated first by Hata et al. [ 5 ] , then by Maeda [6] and recently by
Kawamoto et al. [7, 8, 9 ] . Unlike the case of nailed joints, initial slip
modulus K and ultimate strength P of drift-pin joint are strongly
s u
dependent on the loading angle to the grain direction. According to the

test results, the characteristics of drift-pin joint can be summarized as
follows:
1. Initial slip modulus K is strongly affected by the clearance dia-

s
metral between the pin and the pin-hole in the steel plate.
2. In the case of loading parallel to the grain, the yield strength of a
drift-pin joint can be well predicted by the so-called 'European yield
theory' [10]. The ultimate strength of a drift-pin joint is slightly
(kgf)
MECHANICAL TIMBER JOINTS APPLIED TO GLULAM PORTAL FRAMES Ill
800 800
(kgf)
Timber: Sugi Timber: Sugi
P
Nai1: CN-50 (d=0. 28cm) 600 r Nail: CN-90 (d=0.41cm)
P
Load per n a i 1 ,
Load per n a i 1 ,
400 400
o 200 200
2 4 6 8 10 0 2 4 6
Slip deflection, S (mm) Slip deflection, S (mm)
800
800
(kgf)
(kgf)
Timber: Douglas-fir
600
600 Nail: CN-90 (d=0.41cm)
P
P
Load per n a i 1 ,
Load per n a i l ,
400
400
200
200
Timber: Douglas-fir
Nail: CN-90 (d=0.41cm)
2 4 6 8 10 0 2 4 6 8
Slip deflection, S (mm) Slip deflection, S (mm)
///////AMM///
Load cell
lOOOkgf
or 2000kgf
Nai I Nai I
\
Mr
Del I e c t I on
measurIng
device Output
Output^ 1= Output
I
Steel plate
Timber C l e a r ance

0. 20. 3mm
FIG. 1. Examples of the load (P)-slip (S) relationship for nailed timber joints
with steel side plates.
112 KOHEI KOMATSU
higher than the yield strength, so that the European yield theory
can also be used for estimating ultimate strength P . u
3. In the case of loading perpendicular to the grain, it is difficult to

define yield strength P , and the ultimate strength depends on the
y
testing method and joint configuration. In consequence, the Euro-

pean yield theory cannot well predict either the yield strength or the
ultimate strength.
4. Load (P)-slip (S) characteristics of drift-pin joints loaded perpen-
dicular to the grain depend on the end distance, edge distance, and
the ratio of length (/)/diameter (d) of the drift-pin. Figure 2 shows
an example of the dependence of load (P)-slip (S) curves on the joint
geometry.
Bolts
According to the experiments by Kanaya et al. [11], the behavior of a
single-bolted timber joint is similar to that of a drift-pin joint up to yield
strength P as shown in Fig. 3. After exceeding the yield point, however,
y
the bolted joint can deform much more and is stronger than the drift-pin
joint, because the presence of washers can generate additional resistance
to deformation as the bending deformation of the bolt shank increases.
By this effect, a moment-resisting joint composed of bolts with steel side
plates is much more ductile and stronger than a drift-pin joint [12].
Generally speaking, however, bolts are not used widely for moment-
resisting joints, because the presence of bolt heads makes the appearance
and fire resistance worse.
G L U L A M PORTAL FRAME
Nonlinear analysis
The deformation of a glulam portal frame strongly depends on the
characteristics of the fastener used in the moment-resisting joints. To
analyze this effect precisely, a non-linear finite element method (FEM)
was developed by Komatsu [13]. As the first step in the analysis, the
load (P)-slip (S) relationship of individual fastener was theoretically
determined by extending the concept of a beam on an elastic foundation
into a non-linear problem in accordance with the previous study by
Tsujino and Hirai [14]. Then, the stiffness matrices of the joint element
were ostensively derived by estimating the total strain energy of the
joint element as the summation of the strain energy from individual
p
Steel plate
ate 1(9BB)
I (91) II G lu lam
rji 3 0 0 0
e
e : Ehd distanoe : Edge distance
Deflection measuring
device
Slip (mm)
v 0 40 70 100
Edge distance ( mm )
MECHANICAL TIMBER JOINTS APPLIED TO GLULAM PORTAL FRAMES
Fig. 2. Dependence of the load (P)-slip (5) characteristics of drift-pin joints on the edge distance and the end distance when
loaded perpendicular to the grain. Main members, sugi (Japanese cedar) glulam of 150 mm width (/); gusset plates: 9 mm
SS41 insert-type steel plate; drift-pin, 16 mm diameter (d) SS41 steel pin.
113
114
6000
5000
4000
(kgf)
3000
2000
Load, P
Drift-pip
1000
KOHEI KOMATSU
a) Drift pin j o i n t b) Bolted j o i n t

Bolt
0' . . . .
0 5 10 15 20 25
sup , s ( mm)
FIG. 3. Comparison of load (P)-slip (S) characteristics between drift-pin and bolted joints with the same joint geometry.
Main members, sugi (Japanese cedar) glulam of 150 mm width (/); gusset plate, 9 mm thick SS41 insert-type steel plate;
drift-pin, 16mm diameter (d) SS41 steel pin. Each plot corresponds to the individual test results.
MECHANICAL TIMBER JOINTS APPLIED TO GLULAM PORTAL FRAMES 115
fasteners in the joint region. Once the stiffness matrices of the joint
element have been determined, any two-dimensional glulam frame
structure involving various moment-resisting joints can be solved
nearly up to the ultimate stage by employing an FEM program in-
corporating the Newton-Raphson method as a non-linear solu-
tion.
Verification by full-scale experiment

In order to verify the analysis just mentioned, push-pull cyclic hori-
zontal loading experiments were done on full-size two-story glulam
portal frame specimens having an 8 m span and 6-8 m height as
shown in Fig. 4 [12]. Figure 5(a) shows the test results from a portal frame
composed of nailed joints with steel side plates. In Figs 5(a),
(b) and (c), the continuous curve indicates the load-deformation relation-
ship for a 'roof-beam', while each plot indicates the results calculated
by non-linear FEM. Good agreement between the observed and analy-
tical result was obtained in the case of the nailed joint. Figure 5(b)
shows the case of a portal frame composed of drift-pin joints with double-
insert steel plates. In this case, a relatively large deformation was
observed compared with the result by non-linear analysis. This dis-
crepancy was caused by the fact that there was a relatively large
diametral clearance between the drift-pin and pin-hole in the steel
plates (about 1 mm on average). Figure 5(c) shows the case of a portal
frame composed of bolted joints with steel side plates. In this case,
the observed deformation was smaller than the result by non-linear
analysis due to the effect of friction between the steel plates and glulam
members.
CONCLUSION
The behavior of glulam portal frames strongly depends on the charac-

teristics of the fasteners used in the moment-resisting joints. A nailed joint
can give the most feasible performance, judging from purely structural
aspects, but its appearance and fire-resisting performance make its use in
Japan difficult. In consequence, moment-resisting joints using drift-pins
with steel insert plates are the most widely used in Japan as the
second-best choice.
116
o
, o
0
KOHEI KOMATSU
2000 1 2000 1 2000 1 2000

8000
FIG. 4. Schematic diagram of the two-story glulam portal frame specimen and loading system. Glulam members; Ezo-Todo
(Picea-Abies) 150 mm x 600 mm.
MECHANICAL TIMBER JOINTS APPLIED TO GLULAM PORTAL FRAMES 117
c 25000
~ 20000
] 15000
10000
5000
-200-150-100-50
50 100 150 200250300350400450500
-5000 Displacement, S (u)
10000
25000
20000 b) Drift-pinned
| 15000
10000
5000
-200-150-100-50
50 100 150 200250300350400450500
Displacement, 8 ()
25000]
(kgf)
20000
Load. P
150001 c) Bolted
10000
50001
-200-150-100-50
50 100 150 200 250 300 350 400 450 500
Displacement, 8 (MM)
Observed
Computed
FIG. 5. Load (P)-displacement (S) relationship at roof beam of test specimens,

(a) Nailed (d = 033cm), (b) Drift-pinned (d=18mm), (c) Bolted (d=\2mm). P is
the total applied load. S is the horizontal deflection of roof-beam. The dot and
dash line in each figure indicates the PS relationship obtained by assuming that
all the beam-column joints are perfectly rigid.
118 KOHEI KOMATSU
REFERENCES
[1] G.B. Walford, Timber Use Manual, Section B2, NZTIF (undated).
[2] K. Komatsu, Mokuzai Gakkaishi (J. Japan Wood Res. Soc), 36(12), 1035-
1041 (1990).
[3] K. Komatsu, Mokuzai Gakkaishi, 36(12), 1042-1049 (1990).
[4] Architectural Institute of Japan: Standard for Structural Calculation of
Timber Structures, 155, Gihoudou (1988).
[5] S. Hata, K. Mori and M. Satou, Summaries of Technical Papers at the
Annual Meeting, Architectural Institute of Japan, 1253-1254 (1986).
[6] M. Maeda, Journal of The Hokkaido Forest Products Research Institute,
4(2), 11-18 (1990).
[7] N. Kawamoto and N. Kanaya, Summaries of Technical Papers at the
Annual Meeting, Architectural Institute of Japan, 99-100 (1988).
[8] N. Kawamoto, N. Kanaya and K. Komatsu, Summaries of Technical
Papers at the Annual Meeting, Architectural Institute of Japan, 101-102
(1989).
[9] N. Kawamoto, K. Komatsu and N. Kanaya, Summaries of Technical
(1990).
[10] H.J. Larsen, Proceedings of the IUFRO-5 Conference, Pretoria, 646-654
(1973).
[11] N. Kanaya, N. Kawamoto and K. Komatsu, Summaries of Technical
(1989).
[12] K. Komatsu, F. Kamiya and Y. Hirashima, Proceedings of The 31st Japan
Congress on Materials Research, 185-191 (1988).
[13] K. Komatsu, Mokuzai Gakkaishi, 34(7), 581-589 (1988).
[14] T. Tsujino and T. Hirai, Mokuzai Gakkaishi, 29(12), 833-838 (1983).
Wood as a Livable Interior Material
MASAHARU SUZUKI
Department of Applied Biological Science and Technology,

Tokyo Noko University, 3-5-8 Saiwai-cho, Fuchu, Tokyo 183, Japan
ABSTRACT
It is desirable for new dwelling houses to use renewable plant materials, wood having
many advantages in this respect. It was proved from measurements of the heat flow from
human fingers and by rubbing the finger tips that some wood surfaces were warm and
smooth to the touch in comparison with other materials. The abrasive wear properties of
wood and other interior materials were also assessed, the surface of wood-plastic compos-
ites showing a high level of abrasion resistance.
The hygroscopic behavior of wood and its related materials was determined, and it is
suggested that controlling indoor air humidity is influenced by the water vapor affinity of the
surrounding wall materials. The moisture permeability of face-finished materials and hollow
core panels was obtained, and diagrams to predict moisture condensation were constructed.
The sound transmission loss of wood-based materials was also measured and compared
with the calculated value by using London's modified formula.
Keywords: Surface characteristics, abrasion resistance, humidity control, moisture con-

densation, sound transmission loss.
INTRODUCTION
Although wood is characterized by a high degree of diversity in its

properties, the surface properties as related to the amenities of a wooden
house are really remarkable. For instance, when we touch some objects
(wooden furniture or panels) with the fingers, a feeling of warmth and
smoothness is imparted. The physical and psychological mechanisms for
these effects are not yet clear [ 1 ] .
In connection with these effects, the interior materials in a room are
often rubbed, and their service life is dependent on the abrasion resistance
of the finished layer.
119
120 MASAHARU SUZUKI
Japan has a humid climate, and it is consequently appropriate to use

hygroscopic materials in a room, because materials like wood take up or
give off moisture depending on the surrounding air conditions to main-
tain the relative humidity in a room at the optimum level. Many aspects
of such a humidity control process, however, remain unresolved [2, 3].
It is important to limit the sound transmission from adjoining rooms.
While a number of papers have been published [ 4 ] , the prediction of
sound transmission loss in an intervening panel is inaccurate because of
its dependence on such factors as mass, frequency and configuration.
In this review, explanations and some suggestions on these aspects are
given.
SURFACE CHARACTERISTICS OF W O O D RELATED TO

LIVABILITY
1. Warmth to the touch

Many different interior materials were touched with the forefinger, and
the skin surface temperature was measured by a thermocouple. The
variation of temperature with time was recorded, and thermal trans-
mission from the forefinger to each sample was calculated [5].
Figure 1 shows the relationship between the thermal transmission and
heat capacity (specific heat x density) of interior materials. When the
degree of thermal transmission is small, a feeling of warmth and comfort
is aroused, this being associated with low thermal conductivity of the
materials being touched. From Fig. 1, we can identify a group of
materials which possess livable characteristics, i.e., group A. The interior
materials of group B have higher heat capacity and a greater ability to
receive or hold heat energy.
Apparatus which could duplicate the thermal sensation of human
touch was constructed and used to evaluate sensory warmth quantitative-
ly. For example, the rate of temperature change at the interface was
0-037C/s for Paulownia tomentosa and 0061 C/s for Quercus myr.
The corresponding values for sensory warmth are 5 0 and 3 1 , respective-
ly. In addition to these figures, the rate of temperature change was closely
related to the thermal conductivity of the material [6].
2. Smoothness to the touch

When we touch an object with the fingers to sense its smoothness, we
physically slide the fingers on the object's surface. In order to find the
WOOD AS A LIVABLE INTERIOR MATERIAL 121
X10 6
r
3
( /m K)
3
J
2 B
Specific heat
x dens i ty
2
1 2 3 4 x10
Thermal t r a n s m i s s i o n ( / s i . 2 x 10 m )
J 4 2
FIG. 1. Relationship between the thermal transmission from the forefinger and
the heat capacity (specific heat x density) of the material touched. 1, wool; 2, soft
vinyl leather; 3, clothes; 4, foamed polystyrene; 5, insulation board; 6, tatami mat;
7, Chamaecyparis obtusa Endl.; 8, painted wood; 9, plaster board; 10, particle-
board; 11, polyester resin-overlaid plywood; 12, hardboard; 13, phenolic resin; 14,
glass; 15, ceramics; 16, Portland cement; 17, aluminium (1-2 x 1 0 " m ): Contact
4 2
area of forefinger tip to material.
relationship between the frictional phenomenon and the sense of

smoothness, models of the finger tip were made by using paraffin wax
to reproduce the finger print, foamed polyurethane, and artificial
leather.
The frictional coefficients between these models and interior materials
were then measured for wood, paper, synthetic resin, glass, ceramics and
steel [ 7 ] . A sensory test on each material sample was also conducted,
and the results compared.
In the foregoing experiments, an increase in sensory value indicates
an increase in smoothness, and it was found that there was a good
relationship between the static frictional coefficient of sliding models
and the sensory value as shown in Fig. 2.
To evaluate the surface characteristics of softwood, the frictional
coefficient was measured for various models of human skin at differ-
ent growth stages from early wood to late wood. A decrease in fric-
tional coefficient from early wood to late wood was observed as shown
in Fig. 3.
122 MASAHARU SUZUKI
Frictional coefficient
" 0.4 J 2
o 1 0
13 14
15
17
16
2 0.3
2 -1 1 2
Slightly Smooth
rough
FIG. 2. Relationship between the sensory value of the smoothness of a ma-
terial surface and the frictional coefficient in a sliding human skin model. The
sensory value indicates the ease of smooth sliding. 1. Urea resin sheet; 2, paper
for a name card; 3, glass; 4, Paulownia tomentosa Steud.; 5, Fagus crenata Blume;
6, amino alkyd-resin-painted teak; 7, stainless steel; 8, ceramics; 9, melamine-
resin-overlaid sheet; 10, colored amino alkyd-resin-painted plywood; 11, poly-
vinyl chloride sheet; 12, Cercidiphyllum japonicum Sieb.; 13, Betula maximowic-
ziana Regel; 14, Zelkowa serrata Makino; 15, Chamaecyparis obtusa Endl.; 16,
Acer mono Maxim.; 17, Quercus mongolica Fisch. var.
3 . Abrasion resistance of interior materials

The finished interior boards in a house are subjected to rubbing or
abrasion. The floor suffers from the most serious abrasion to the
detriment of its service life. In order to evaluate the abrasion resistance
of interior materials, the Taber instrument was used. This laboratory
test method for abrasion produces an abraded area that is annular in
shape, and is considered to be a realistic simulation of actual wear
phenomena.
The volume removed by abrasion from the surface of a sample
increases with increasing grain size and vertical load on the testing wheel.
This abrasive wear is caused by shear stress between the wheel and the
surface of material, and can be expressed by the following relationship:
(1)
where d is the loss in thickness, a is the abrasive-wear coefficient, T is the

shear stress, AR is half of the width of the abraded annular area, n is the
Frictional coefficient
0.7 o o
o o n o
o o
o o
o o
0.5 o
I I I I I I I 1 I I I I I I I I I 1 I I I 1 I I
A 6 8 10 12 14 16 18 20 22 24 26
kEa r l y wood *Late
wood^
Sample number
FIG. 3. Frictional coefficient by sliding a paraffin wax replica of the skin of the
forefinger on the surface of wood. This is plotted along the annual ring growth of
Cryptomeria japonica D. Don for a vertical load of 1*55 g.
number of cycles, \i is the frictional coefficient of the abrasive paper

wound around the wheel, and P is the vertical load on the wheel. From
Eqn. 1, we can get
(2)
By obtaining the abrasion in a steady state with a linear relationship

between d and n, the value of d can be calculated. Table 1 indicates the
value of d for various kinds of interior materials, in which d is the
measured loss in thickness for n= 100 cycles. Compared with solid wood,
the value of d for synthetic resin and wood-methyl methacrylate (MMA)
composite is less. There are differences in the wear of paint-impregnated
or adhering materials, and microscopic observation shows many scratch
lines produced by the abrasive grains.
In a practical wear situation, there are many kinds of appearance
changes and variations in the abraded depth. When testing wood, it is
necessary for the size of the test specimen to be small so that its abrasion
property is that of the local wood tissue.
124 MASAHARU SUZUKI
TABLE 1
Abrasive-wear property of interior materials by the Taber abrader expressed as
the loss in thickness per 100 cycles under load P = 500g
Material Loss Material Loss

in \im in \im
Pinus densiflora S. et Z. 34-45 Melamine formaldehyde resin 16

Fagus crenata Blume 30 Polyurethane resin 22
Quercus crispula Blume 27 Epoxy resin 23
Hardboard 35 Amino alkyd resin 24
Maple wood-Poly 22 Pure PVC sheet 13
(ST-MMA)* composite
Diallyl phthalate resin- 23 Solid PVC 32
impregnated paper
* ST: styrene, MMA: methyl methacrylate
c x1 0 '
6
g\ 10 i
3
Abrasion resistance u (kg
A* 5
8 2A
4
A
6h
A

8ooo 1
A A
D Q
O D m
D
A A a *
q
2hD fiO D
A
AA
Â D6 ^ A
0 0.4 0.8
Annual ring
on proportion ofgrowth process
FIG. 4. Change of abrasion resistance according to annual ring growth. 1,

Chamaecyparis obtusa Endl.; 2, Cryptomeria japonica D. Don.; 3, Pinus densiflora
S. et Z.; 4, Fa#s crenata Blume; 5, Quercus mongolica Fisch. var.; 6, Betula
maximowicziana Regel.
A theory to represent abrasion phenomena has been derived [ 8 ] , the

volume V removed by abrasion being given by
(3)
where P is the internal pressure induced by contact with the abrasive, <5 is
the frictional force (shear force) and L is the length of friction area.
Abrasion resistance, U, is defined as the amount of the work needed to
remove a unit volume of wood.
In an experiment, a small strip of a specimen would around an
attachment was slid on abrasive paper (no. 180) with 80 jam grit size and
104 degree edge angle.
Figure 4 shows the abrasion resistance of various wood samples
according to annual ring growth. The abrasion resistance of late wood
was higher than that of early wood.
HYGROSCOPIC A N D THERMAL PROPERTIES
1. Moisture adsorption and humidity control behavior

The adsorption of moisture at about 75% humidity was assessed for
each interior material and is shown in Table 2. These materials are
divided into three groups according to the quantity of adsorbed water in
moisture content range of 0-6%, 7-14% and 15-20%.
Indoor air humidity is influenced by the surrounding wall materials
and comfort dictates that it should not be too high or too low.
TABLE 2
Hygroscopic property of interior material. Moisture content under 75% relative
humidity at 20 C
Material MC(%) Material MC(%)
Chamaecyparis obtusa Endl. 11 Paper (LBKP) 9
Hydrogen chloride treated 7 Nitrocellulose film 2
Formaldehyde treated 5 Polyurethane resin film 1-2
Acetylated 5 Porous calcium silicate 5
(acetyl content: 16%)
Fagus crenata Blume 13 Polyvinyl chloride film 0-2
170 C heat treated 55 Polyvinyl chloride powder 0-7
126 MASAHARU SUZUKI
Consequently, the hygroscopic properties of interior materials are very

important.
The relative humidity in a room whose walls are sheathed with non-
hygroscopic materials and with no ventilation is affected by temperature.
It increases with decreasing temperature when a small amount of wood
or a large amount of polyvinyl chloride is used as the wall materials of the
room. By using a certain amount of wood on the surface of the walls, the
relative humidity will not change due to sufficient adsorption or desorp-
tion with temperature.
This controlling effect can be expressed by the humidity control
efficiency (HCE) as
(4)
where h is the initial humidity, h is the humidity in relation to the

{ s
temperature change and amount of hygroscopic material, and h is the c
extreme humidity in free material with the same temperature change.

HCE obtained for a space of 176 c m surrounded by wooded walls is
3
63%, and that by Portland cement is 48%. The absolute value of HCE
can be used and the values for some interior materials are tabulated in
Table 3 [9]. If walling is made from several kinds of finished materials
and their area percentages are a a , a ,..., then
l 5 2 3
HCE = c a +c a
l l 2 2 + c a -\
3 3 (5)
TABLE 3
Humidity control efficiency in a space of 176 cm 3
Material Coefficient Material Coefficient
Shorea negrosensis Foxw. 0-63 Polyvinyl chloride 0-22

film
Nitrocellulose paint coating 0-39 Tatami mat (straw) 054
Kraft pulp sheet (LB) 0-58 Portland cement 0-48
where c , c , c , . . . are the absolute value of HCE for each material. It

1 2 3
seems optimum for a dwelling that HCE should be about 55% for a
comfortable range of indoor humidity.
The capacity for absorbing or giving out moisture has been inves-
tigated from two points of view. One is the hygrothermic capacity that is
related to the relative humidity change which follows a temperature
change. The other is the moisture-absorption capacity (or moisture-

desorption capacity) when exposed to an ambient atmosphere of high
relative humidity (or low relative humidity).
From Maki et al. [10, 11], the logarithm of relative humidity (log h) is a
linear function of temperature T under conditions free from hygroscopic
materials with a slope of - 0 0 2 4 5 (%/C). If slope value | B | (the absolute
value) is smaller than 00245, the relative humidity is influenced by
hygroscopic materials in the space without ventilation. When |B| ap-
proaches zero, the relative humidity in the space is unchanged with
temperature. Table 4 shows the value for | B | with some interior materials
in comparison with their moisture absorption capacity [10].
TABLE 4
|B| value in a house lined with various internal wall materials and relative
humidity increment Ah in a test chamber lined with these same materials (Maki et
al. [10]). Ah was measured by injecting water vapor into the test chamber in a
vacuum at 20 C. The water vapor was injected for 15 min, and Ah was measured
after 60 min.
Interior wall materials ||x/0 4

Ah
Calcium silicate insulation board 5 26-5

Wooden fiber insulation board 15 19-5
Plywood 20 24-28
Phenol-melamine high-pressure laminate 66 44
Rock wool ceiling 83 (ref. 74) 41
Melamine coated plywood 139 36
Vinyl film 265 72
From Okano's paper, the humidity control behavior of wood is related

to the change of its moisture content, and HCE is affected by the quantity
and shape of the wood [12]. If the relative humidity surrounding wood is
changed, the moisture content in the core of the wood will be changed
much later than that in the surface layer.
Figure 5 shows the change in moisture content of the surface layer and
core in a day caused by changes in the temperature and relative humidity
of the surrounding atmosphere.
2. Moisture permeability, thermal gradient and moisture condensation

for a wooden panel
A double wall with a hollow core is widely used in dwelling-house
construction. If the water vapor pressure outside the wall is higher than
128 MASAHARU SUZUKI
20r
Moisture content (%)
16 L
8
I
e
12
8 FIG. 5. Moisture content change in

the surface layer and core of a wood
sample during 24 h in relation to the
equilibrium moisture content es-
timated from the temperature and
(hr) humidity of the surrounding atmos-
phere (Okano [12]).
that in the core, the moisture transmitted through the surface increases
the relative humidity in the inner core. The amount of moisture transmit-
ted may depend on the moisture resistance of the intervening materials
and the difference in water vapor pressure on both sides.
From the increment of relative humidity dh in time dt, the moisture
transmission coefficient, J , is given by the following equation:
t
(6)
where h (t) is the relative humidity at time t in a hollow core (air space),
c
h is the relative humidity outside the panel, E is the thickness of the

Q
hollow core, a is the amount of water vapor per unit volume for 100%
relative humidity, and p is the saturated water vapor pressure. Figure 6
s
shows the moisture transmission coefficients of wood-based materials

against time [13].
In order to measure the moisture and heat transfer through a wall, an
air-conditioner was designed to regulate the atmospheric conditions
which are kept constant or changed.
A double-wall panel with a hollow core was constructed on a wooden
frame and sheathed with plywood.
Figure 7 shows a side view of the test panel, and the measuring points
of relative humidity and temperature. The electrode of an insulation
M o i s t u r e transmission J ( /m hmmHg) WOOD AS A LIVABLE INTERIOR MATERIAL 129
x1
20
2
9
10
10 20 30
T i m e ( h r)
FIG. 6. Time-dependent moisture transmission coefficient. 1. Chamaecyparis

obtusa Endl. of 31mm thickness; 2, hardboard of 3-6 mm; 3, dap resin sheet-
overlaid plywood of 30mm.
resistance meter was attached to the board, the meter indicating the
changing resistance with increasing moisture content.
Figure 8 shows time plots of log p (electrical resistance) u (moisture
content) and h (relative humidity). When the B-chamber was cooled (see
'cool' in [A]), there was a marked increase in u and finally surface x
moisture condensed. From Fig. 8 [ B ] , as h in the internal air space 2
increased and moisture content M at the lower corner of S reached more

3 5
than 20%, then surface moisture condensed. As relative humidity h in x
Fig. 8 [ D ] is reduced, the value of u is less than 17%, showing no 3
moisture condensation.
In order to predict the internal moisture condensation, the following
equations can be derived:
(7)
130 MASAHARU SUZUKI
S7 6 5 4 3 2 S1
O O D O 0 fc> O
U
A -h
P P
B A FIG. 7. Cross-section of a plywood
PI Ar PI Bn
sheathing panel. S, lateral position; T,
temperature; p, water vapor pressure;
h, relative humidity; w, moisture con-
tent; p, electrical resistances; Ar, air
H6h-60-H6h- n
space; PI, plywood; Bn, boundary
layer.
(8)
where T and P are the temperature and water vapor pressure at

x x
distance x (x = 0 is the terminal of the boundary layer on S J , R (x) t
and R (x) are the resistance of thermal conduction and that of mois-
m
ture permeance up to the x point, and R and R are the total resist-
t m
ance to thermal transmission and that to moisture transmission, respect-

ively.
We can find the value for saturated water vapor pressure P at every s
point of the panel section and compare them with P . If P >P this x X S9
produces moisture condensation, while P <P produces no moisture

X S
condensation. Figure 9 indicates the relationship between P and P for x s
each of Fig. 8 [ A ] , [ B ] , [ C ] and [ D ] . The prediction from Eqn. 8 is

validated by the experimental results shown in Fig. 8.
Although temperature at each point ( S S , . . . , S ) fell reason-
l 9 2 7
ably well onto a straight line, that at S (the lower portion of the inside
5
surface of the panel) was a little lower than predicted. Heat convec-
tion in the hollow core may have influenced the temperature distribu-
tion. In the case of a multilayer wall with an insulation layer and
underlayer, the temperature distribution could be well predicted by
using Eqn. 7.
[A]
0 ^ 2 0
- C 6 0
0 ^ 2 0
^ 10
^ 7 0
^ 60
o^ 2C
^110
1
2
60
- 1
4< - cool t
i i i i
8 1 6
xl(
0
Time (min)
FIG. 8. Time plots of p, u and h in the panel shown in Fig. 7. The initial states of
T and h were 25 C and 65%, respectively. Following the change of temperature,
T was reduced to 0C. Dashed lines show the moisture content at the internal
7
surface of the corner (S ). 5
LIMITING S O U N D TRANSMISSION
The transmission loss of reverberating sound through a single wood-

based panel having been determined, attention needs to be drawn to the
132 MASAHARU SUZUKI
(kPa)
o
Water vapor pressure

Temperature (C)
ii 11
BnPI Ar PIBr? BnPI
i i
Ar

1i
PI Bn
Thermal resistance
(ratio)
FIG. 9. Prediction diagram for moisture condensation. Refer to Fig. 7 for the
definition of each symbol.
difference in measured value from that computed by the random inci-

dence mass law. This difference was 20-30% high in the low-to-middle
frequencies, and may have been caused by a resistive component in the
mechanical impedance of the wall.
This resistive component is expressed as resistive term R in London's
formula. The value of R for a test panel can be evaluated to give the best fit
to the experimental data. It was found that R/M could be expressed as a
function of the thickness with a high correlation coefficient, where M is the
mass per unit area of the panel. The relationship between panel thickness J
and R/M is shown in Fig. 10 for plywood and particleboard. If R is
unknown, it can be calculated from the following experimental formulae:
plywood R / M = 1 1 8 J - c.c.= - 0 - 8 8 6
9 5 3
particleboard R/M = 0'l\S-' c.c.= - 0 - 8 7 9

120
hardboard R/M = 0-72J" c.c. = - 0 - 9 2 9

0 8 7 9
solid wood R/M = - 0-254J-h 117 c.c. = - 0-708

where the units of R/M are m / k g and of J are cm, and c.c. indicates the
2
correlation coefficient.
3f (A)
(B)
0.8
o <& o
0.4 8 ^
1 2 3 4 0 2 ~4
Thickness (cm)
FIG. 10. Relationship between the thickness and R/M ratio for plywood (A) and
particleboard (B). R is the resistive term, and M is the area density.
By substituting R into the mass law equation, we can compute the

sound transmission loss (STL) in the low-to-middle frequencies
STL = 10 log a\ -10 log lnj^l +
(9)
coM
where co is the angular frequency, M is the area density, p' is the density
of air, and C is the velocity of sound in air.
It has been found from the vibration properties of a flat plate that the
resistive term had a close relationship with the two dimensional loss
moduli and its vibration amplitude [14].
The sound transmission loss of a double-panel wall is influenced by
the thickness of the wall and the material used for the facing board. A
double-panel wall incorporating a hollow core resonates in the low
frequency range and transmits sound. When the hollow core is filled
with a glass wool mat, the sound transmission loss is increased by
2-3 dB.
Another problem is foot-step noise generated on the second floor being
transmitted to the first floor. There has been some research on the
sound-proofing of wooden structures to achieve lower sound transmit-
tance [15]. Experiments have also been conducted on a floating floor and
on the application of sound-insulating material to reduce floor-impact
noise [16].
134 MASAHARU SUZUKI
CONCLUSION
Wood material has the advantages of high specific strength and good
workability compared with other building materials. In particular, a
wood surface is characterized by warmth to the touch and by some
degree of smoothness. This was clearly proved by measurements on heat
flow and rubbing with a finger tip. The abrasion properties of wood and
other interior materials were measured, and their performance in use was
estimated. The effect of the inhomogeneity in wood tissue on the abrasion
property was shown to be due to its major structural factors.
It is apparent that wood, a hygroscopic material, has a high level of
humidity control efficiency which is important for a dwelling-house. The
transmitted moisture through the face material of a double-panel wall
increased the relative humidity in the intervening air space. These
experimental data confirmed the prediction of water vapour pressure and
the formation of moisture condensation outside or inside the wall.
The sound transmission loss of wood-based materials could be cal-
culated precisely by using London's modified formula. The resistive term
has a close relationship to the two dimensional loss moduli and vibration
amplitude of each wood-based material.
REFERENCES
[1] T. Okajima, S. Wakayama, K. Noda and H. Konishi, J. Arch. Plan. Env.

Eng. Trans. AIJ. No. 369, 9 (1986).
[2] T. Okano, Y. Miwa, A. Mishiro and I. Asano, Mokuzai Gakkaishi (J. Japan
Wood Res. Soc), 25, 379 (1979).
[3] M. Norimoto, T. Ohgama and T. Yamada, Mokuzai Gakkaishi, 36, 341 (1990).
[4] M. Suzuki, H. Saito and T. Tanaka, Bull. For. Pro. Res. Inst., No. 306, 95 (1979).
[5] M. Suzuki, Bull. Gov. For. Exp. St., No. 231, 27 (1970).
[6] Y. Harada, K. Nakato and T. Sadoh, Mokuzai Gakkaishi, 29, 205 (1983).
[9] M. Suzuki, Bull. Exp. For. Tokyo Noko Univ., No. 24, 89 (1987).
[10] F. Maki, M. Norimoto and T. Yamada, Mokuzai Gakkaishi, 24, 797 (1978).
[11] F. Maki, M. Norimoto and T. Yamada, Mokuzai Gakkaishi, 27, 828 (1981).
[12] T. Okano, Mokuzaikankyo no Kagaku (edited by T. Yamada), Kaiseisha,
308 (1987).
[13] M. Suzuki, Bull Gov. For. Exp. St., No. 257, 1 (1973).
[14] M. Suzuki, Y. Kozima and M. Koyasu, Mokuzai Gakkaishi, 32, 155 (1986).
[15] T. Nakao, A. Takahashi and C. Tanaka, Mokuzai Gakkaishi, 34, 14 (1988).
[16] A. Takahashi, M. Nanba, T. Nakao and C. Tanaka, Mokuzai Gakkaishi, 33,
950 (1987).
Structure and Properties of Chemically
Treated Woods
MISATO NORIMOTO
and
JOSEPH GRIL
Laboratoire de Mecanique Generate des Milieux Continus,

Universite de Montpellier II, USTL, Place E. Bataillon,
34000 Montpellier Cdx 2, France
ABSTRACT
Chemically treated woods are classified by using a model of the structural change in the
cell wall and the lumen of a cell resulting from the treatment. The dimensional stabilization
of wood by chemical treatment is discussed, based on a generalized concept of the
dimensional stability that allows for the deformation due to the simultaneous action of
moisture changes and external forces applied to the material. Typical cases of chemical
treatment are used to outline the relationship between the structural changes involved and
the resulting properties.
Keywords: chemically treated wood, dimensional stability, mechanosorptive creep, acoustic

properties, static mechanical properties
INTRODUCTION
Chemical treatment is one of the most effective approaches to enhance

the properties of wood. However, the mechanical and surface properties
so much appreciated in wood result from the existence of many compos-
ite structures ranging from the macroscopic to microscopic and molecu-
lar level. Some treatments of wood reach the core of cellulosic microfib-
rils, destroying the crystalline structure and eliminating most of the
135
136 MISATO NORIMOTO AND JOSEPH GRIL
composite structures of 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. A
chemical treatment of wood as defined in this paper excludes such radical
modifications. A chemical treatment may reduce some defects relative to
wood utilization, enhance its properties, and create new performance or
functions, while keeping the bulk of the superior mechanical properties
of wood. Chemical treatments include the use of heat and moisture, resins
or other chemicals. In most cases, they affect the amorphous components
of the cell wall material, the cell lumens sometimes being filled by
chemicals.
In many situations, the difficulties encountered in wood utilization are
related to its dimensional instability to moisture, biodegradability and
flammability. In the case of soft coniferous woods, especially when used
for flooring boards, hardness and abrasion resistance may also become a
problem. Among the difficulties, changes of wood properties caused by
the variation of ambient conditions are of great importance: usually they
are undesirable effects, and one of the main reasons for treating wood is
to reduce them.
Hitherto, dimensional stability has been evaluated by measuring the
dimensional change to load-free wood specimens. However, in actual
practice, wood components are often subjected to mechanical force, so
that the action of external stresses should be taken into account when
measuring the dimensional stability. Moreover, the simultaneous action
of mechanical stresses and moisture changes can cause an exceptionally
high level of creep strain. This phenomenon is called 'mechanosorptive
creep' [1]. The viscoelastic properties such as the dynamic Young's
modulus and internal friction may also depend significantly on humid-
ity. This is one of the main factors causing tonal instability in wooden
musical instruments. Generally speaking, any type of dimensional
change caused by humidity variations can be regarded as dimensional
instability and originates in high hygroscopicity. In order to study how
chemical treatments can effectively reduce this dimensional instability
of wood to moisture, it is important to have a clear understanding of
the relationship between the structural modifications and changes in
properties.
In this paper, a classification of chemically treated woods based on the
modification of the microstructure will be proposed. The properties of a
number of treated woods will then be presented and analyzed in relation
to the type of chemical treatment involved.
STRUCTURE AND PROPERTIES OF CHEMICALLY TREATED WOODS 137
M O D E L FOR W O O D WATER INTERACTION
B-l in Fig. 1 shows a molecular model of the lignocellulosic materials

that form the cell wall of wood. The 'chain' indicated by a refers to the
crystalline core of a cellulosic microfibril, but it must be understood in a
wider definition. The chain may be any part of the lignocellulosic
materials of the cell wall that is not affected by water movement. It has a
water-reactive zone at its boundary, illustrated by 'hydroxyl group' b.
Neighboring chains are linked to each other by a set of water-reactive
zones, shown here by the 'hydrogen bond' c. Water sorption on c has a
double effect. First, the change of matrix volume forces lateral displace-
ment of the two chains a in the direction indicated by d. Second, it
P:::3
ttt
1 2 3
B 4 5 6
H
d
rrl rr
FIG. 1. Model for the chemical treatment of wood. A, Cellular level: 1-3,
untreated cell wall; 4-6, treated cell wall; 1, untreated; 4, no chemical deposits in
lumen, 2 and 5, deposits on cell wall surface; 3 and 6, filling of lumen. B,
Modification of ligno-cellulosic material at the molecular level. O , hydroxyl
group available for hydrogen bonding; , substitution of hydroxyl group;
# , bulking agent.
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 moisture expansion (swelling)
or shrinkage, and the second in mechanosorptive creep or relaxation. In
the model, the two chains are represented as two independent members
for the sake of clarity, but in reality, they belong to a fully interconnec-
ted framework including not only the crystalline core of the microfibrils,
but also some of the surrounding molecules that do not react to water.
This framework behaves elastically while molecular movement in the
matrix is possible, and is enhanced by water movement (mechano-
sorptive effect). The existence of this framework prevents moisture-
induced expansion d as well as stress-induced creep deformation e from
exceeding a limit at which the tendency for the matrix to expand or to
modify its shape is counterbalanced by the pressure and the shear
stresses exerted by the framework. This fact also accounts for the
mechanosorptive recovery: whatever molecular movements occur in the
water-reactive zones of the matrix, the surrounding framework tends to
recover its original shape in a stress-free state. The simplistic image used
here suggests that the basic phenomena associated with expansion d and
creep e are not connected with each other because they occur in
different 'directions' at the local level. Moreover, the deformation is
pictured as a two-dimensional process. Although the actual situation is
more complex and involves a three-dimensional interaction between
sorption and local stresses, the model has, however, some basis regard-
ing the cell wall structure and the anisotropic nature of the microfibril-
lar framework. At the level of the cell wall, moisture-induced expansion
is more or less allowed in the thickness direction, but is partially
prevented in the circumferential direction due to mutual restrictions of
the cell wall layers. On the other hand, an external stress applied to
wood results at the cellular level in local stresses directed mainly in
the axial-circumferential plane of the cell wall. Therefore, we have
locally a partial disconnection between moisture-induced expansion and
stress-induced strain. Although this analysis only holds, strictly speak-
ing, for thin cell walls, it underlies the two separate directions shown by
d and e. The cellulosic framework behaves elastically, while, in the
matrix, molecular rearrangements are possible and result in a delayed
mechanical response in the form of viscous strain, stress relaxation or
vibrational damping. The mechanical behavior of the matrix (apart from
moisture expansion) is influenced by water in two ways. First, when the
matrix contains a large amount of water, it has a much higher viscosity
than in the dry state. Second, the sorption process itself (either adsorption
or desorption) contributes further to the viscosity of the matrix, because
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 that rearrangement (longer than the time needed for
sorption because it involves the movement of longer molecular segments),
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 (the 'mechanosorptive effect'), or a
sorption-induced viscosity that cannot be predicted by the direct effect of
moisture or viscosity. The amount of viscosity (compared to elasticity) in
the mechanical response of wood to external forces depends on the level
of the local stresses supported by the matrix at the molecular level. In the
case of longitudinal loading, for example, the proportion of matrix
involvement in the deformation process is determined by the average
inclination of the microfibrillar framework relative to the fiber axis in the
axial-circumferential plane. The greater this inclination, the more the
matrix is involved, and the more the material is subjected to time-
dependent phenomena such as hygrothermal softening, hygromechanical
coupling and mechanical damping. The molecular model of B-l in Fig. 1
does not contain this information. However, it is implicit by the stress-
induced strain being shown as a shear deformation in direction e, which
thus stands for the direction of local stresses supported by the matrix.
The direction of the 'chains' may not be, strictly speaking, that of the
microfibrillar framework. Moreover, the water-reactive zones of the
matrix responsible for the activation of viscosity c acting on e may not be
exactly the same as those involved in moisture-induced expansion c
acting on d. These aspects do not detract from the validity of the
model as a synthetic representation of an extremely complex actual
system.
CLASSIFICATION O F CHEMICALLY TREATED W O O D S
We classify chemically treated wood according to a double criterion

[2]: (A) modification of the lumens (cellular level), and (B) modificaton of
the cell wall material (molecular level). Figure 1A shows the cross section
of a single cell. A-l is the untreated wood. In A-2 and -3, the cell wall
remains untreated, while a resin or any other product is either deposited
on the internal face of the lumen (A-2) or fills (partially or totally) the
lumen (A-3). In A-4, -5 and -6, the cell wall is modified. Again three cases
are distinguished: no deposit in the lumen (A-4), deposit on the internal
face of the lumen (A-5) and partial or total filling of the lumen (A-6).
Figure IB analyzes several cases of cell wall modification according to
the molecular model just detailed. The hydroxyl group represented by
the open circles is available for hydrogen bonding with water; the small
closed circles indicate that the hydroxyl group has been substituted with
chemical bonding and is not available any more. The large closed circles
indicate the bulking effect caused by the introduction of large molecules
between the constituents, which replace water molecules. Weak connec-
tions such as hydrogen bonding are represented by two hydroxyl
groups (open circles) facing each other; strong molecular bonding is
shown by a line. The crosslinking effect is thus pictured by an unbroken
sequence of lines and closed circles, either small or large, linking the two
chains. The initial model (B-l) can be modified according to the eight
patterns (2 to 9) of Fig. IB. In patterns 2 and 3, crosslinking occurs
without the bulking effect by using molecules of low molecular weight
that link the two reaction sites with hydroxyl groups. In 2, the reaction
is done in a dry state with a short linkage, preventing both lateral
expansion d and molecular movement e. In 3, the reaction is done in a
swollen state, so that after drying, the structure develops considerable
looseness. Both a crosslinking and a bulking effect occur in patterns 4
and 5. In 4, the reactant is hydrophobic, and in 5 it is hydrophilic
(as shown by the small open circles attached to the big closed circle).
In the remaining four patterns there is a bulking effect but no crosslink-
ing. In 6 and 7, the reactant establishes a stable bond on one side.
Again, it is hydrophobic in 6 and hydrophilic in 7; in pattern 7, the
reactant suppresses hydrogen bonding on one side but simultaneously
creates a new sorption site at the other. In the last two patterns, no
stable bond is established between the reactant and the constituents. In
pattern 8, the bulking agent is extremely hydrophilic and establishes
extensive hydrogen bondings with the constituents, and in pattern 9, it
is hydrophobic and does not interact with the constituents nor with
water.
Any type of chemical modification can be characterized by a combina-
tion of the two criteria A and B of Fig. 1. Several combinations of these
criteria may occur at the same time. In some cases, it is not possible to
determine which combinations are involved in the treatment because of a
lack of knowledge on the mechanism for action that underlies the
treatment.
CHARACTERIZATION O F W O O D STABILITY
Conventional Dimensional Stability

Moisture-induced expansion, represented by d in Fig. 1B-1, may be
reduced by the crosslinking in patterns B-2, -4 and -5 and by the bulking
effect in patterns B-4-9. Treatment of types A-2 and -3 with no modifica-
tion of the cell wall material (obtained, for example, by resin impregna-
tion) can delay, but not reduce the adsorption of water and consequential
dimensional changes. On the other hand, types B-7 and -8 may be used in
the case of dimensional stability without any reduction of hygroscopicity
being desired.
The dimensional stability provided by treatment is evaluated by use of
an 'anti-swelling efficiency' (ASE) factor given by the following express-
ion:
where S is the relative volume increase (volumetric swelling) of a

load-free, chemically treated specimen between a dry state and a wet
state, and S is the value measured for untreated wood under the same
u
experimental conditions.
where V is the dry volume (low humidity) and V is the wet volume (high
x h
humidity). An ASE value of 100% would mean that the specimen is

perfectly stabilized; ASE of 0% means that the treatment has had no
effect at all on the dimensional stability; and negative ASE that it has the
opposite effect on stabilization. Wet and dry states were obtained here by
subjecting an initially dried specimen to 20 C and 89% RH for one week
(wet), and then for one week at 20 C and 29% RH (dry). The experiment
to evaluate dimensional stability by ASE involves measuring the dimen-
sional change of load-free specimens. Therefore, ASE alone does not give
any indication of the modification of mechanical properties that involve
the action of external forces on the material. Appropriate methods to
obtain this additional information depends on the hygromechanical
loading to which the material is subjected in practice. Wood is a highly
anisotropic material that resists much more along the grain, so that for
structural uses, its longitudinal behavior is of major importance. The
modification of wood by chemical treatment will be evaluated on three
criteria: (1) the mechanosorptive effect, (2) the acoustic properties, and (3)
the static mechanical properties.
Mechanosorptive Creep
Dimensional stability with respect to such stress-induced deformation
as creep or vibrational damping requires a reduction of matrix viscosity
due to moisture sorption. Crosslinking of type B-2, -4 and -5 may
efficiently prevent viscous deformation in the matrix. Bulking along,
however, does not necessarily provide this effect. Only 'hydrophobic
bulking' types B-5 and -9 have a good chance of reducing hygroscopicity.
'Hygrophilic bulking' in types B-7 and -8, however, favors matrix hygro-
scopicity and may thus enhance the plasticizing action of water on the
matrix, resulting in instability.
Mechanosorptive creep was measured in three-point bending tests at
30 C under humidity cycling between 29 and 86% RH. The initial load
was adjusted to obtain a stress level of 10 MPa in the dry state. The creep
test lasted two days for each specimen, which was loaded first at 29%
RH, kept for lOmin., moved to 86% RH for 24 h, and then returned to
29% RH for the remaining 24 h. Specimens of sitka spruce (Picea
sitchensis; 100 mm [ L ] x 15 mm [R] x 2 mm [T]) were modified by vari-
ous treatments and tested according to this procedure.
To characterize the ability of a treatment to reduce mechanosorptive
creep, an 'anti-creep efficiency' (ACE) factor is defined by an expression
similar in format to that for ASE:
where SJ is the increase of compliance (the strain level divided by stress

level) induced by a complete cycle of humidity, and SJ is the value
U
obtained from an untreated specimen. Like ASE, a specimen with 100%

ACE would have no creep at all; negative ACE indicates mechano-
sorptive instability. The expression used to calculate 5J is
where w is the width of the specimen and t is its thickness (in the dry
state), / is the span (80 mm), F is the applied load, / is the deflection after
x
loading, f is the deflection at the end of the experiment, d J describes the

2
potential for stress-induced strain along the grain under changing humid-
ity conditions. It is the formal equivalent of swelling ratio S, which
represents the potential for the moisture-induced deformation of an

unloaded specimen. Figure 2 shows the relationship between ASE and
ACE after several treatments [2].
100
FIG. 2. Anti-creep efficiency (ACE)

compared to anti-swelling efficiency
ACE(
(ASE). F, formalization; H, heat

treatment; M, treatment with maleic
100
acid; MG, treatment with maleic
acid and glycerol; A, acetylation;
WPC, formation of a wood-polymer
composite; PO and BO, etherifica-
tion with propylene oxide and buty-
200 lene oxide, respectively; PEG, im-
pregnation with polyethylene glycol;
50 P, impregnation with phenol-formal-
ASE(M dehyde resin.
Acoustic Properties
Another way to evaluate wood viscoelasticity is to measure the
specific Young's modulus (E/y) and internal friction (tan <5). E/y is related
to the sound velocity, and tan 5 to sound absorption, and both are thus
important parameters of acoustic properties. Pattern A-4 + B-2 should
be effective for increasing E/y, because the small molecular bridges made
while chains a were closed to each other improve the molecular cohe-
sion with a very small weight increase. In all other types (except for
A-4 + B-3), a decrease of E/y would be expected, because bulking either
reduces the molecular cohesion, or improves it, not enough to compen-
sate for the increase in density. The dependency of tan d on the type of
treatment, however, should be close to that of mechanosorptive creep (a
decrease with B-2, -4, -5, -6 and -9 and an increase with B-7 and -8)
because both measurements are related to viscous movement in the
matrix.
The following quantities can be used to evaluate the effect of chemical

treatment on acoustic properties:
Figure 3 shows the relationship between A(E/y) and A(tan<5) , and T T
Fig. 4 the relationship between l o g tan 3 and l o g E/y for a number of

1 0 1 0
treatments performed on akaezomatsu (Picea glehnii) at 20 C and 60%

RH, using the free-free beam method [ 3 ] . The dotted line on Fig. 4 is the
regression line obtained for untreated wood of the same species. Some
comments should be made concerning this regression line. In the longi-
tudinal direction, E/y (Young's modulus divided by density) characterizes
the average rigidity of the cell wall [ 3 ] , while tan d represents the relative
amount of viscous strain to elastic strain and the participation of the
matrix in the deformation process. Both quantities depend mainly on the
internal anisotropy of the cell wall (mean microfibrillar angle) and very
little on the wood density. As the variations of both E/y and tan<5
originate from the same ultrastructural factor, they should not be
6_
PEG 3
o-
-150
^ -100
PEGAO^T
WPC2O g 1
-50
< -F 5
PO3-O
B03-O . FIG. 3. Relationship between the
wiC2-W>
- 5 0 ^ >0
A 3
50 increase of specific Young's modulus
(A(/y) ) and increase of internal fric-
P T A(^) C/.) T
3
6 t
tion (Atandj) after chemical treat-
-50 ment. Refer to Fig. 2 for definition of
symbols.
-1.7
-1-8
o -1-9
c
o
"o -2.0
cn
o
-2.1
-2.2
-2.311.0 11.1 11.2 11.3 1U
FIG. 4. Relationship between the logarithm of specific Young's modulus

(log E/y) and the logarithm of internal friction (log! tan S). U , untreated. The
10 0
dotted line is the regression line for untreated wood. Refer to Fig. 2 for definitions
of the other symbols.
independent of each other. In practice, it has been observed for a number

of untreated specimens from different wood species that the higher E/y
(corresponding to a smaller microfibrillar angle), the smaller tan d [ 4 ] .
Moreover, E/y and tan S are related to each other by a power law, i.e., the
line on a log-log graph such as the dotted line in Fig. 4. When different
viscous behavior is apparent between a treated specimen and the un-
treated reference, we do not know a priori whether the difference is due to
the treatment or originates from bad matching of the two specimens; for
example, a considerable difference in their mean microfibrillar angle.
However, we did confirm that it is not the latter case (at least for
treatments of type A-4 in which only the cell wall was modified) by
observing that the point representing the modified wood data in Fig. 4
not only differs from that of untreated wood ( U ) but also lies far from
2
the dotted line.
Static Mechanical Properties

Treatments of type A-2 and -3 and A-5 and -6 + B-4, -5, -6 and -9
should improve such static properties as the rigidity (elastic moduli) and
strength. With a combination A-4 + B-4, -5, -6 and -9, this may or may
not be the case, depending on the amount of degradation provoked by
the treatment.
The following factors serve to evaluate the improvement of elastic
modulus E and strength a resulting from the treatment:
where E and a relate to untreated wood. Figure 5 shows in the AEj-Aaj

u u
plane the values obtained from static bending tests in the longitudinal
direction on sitka spruce [2] at 20 C and at either 65% RH (solid circles) or
30% RH (open circles). In practice, it is important to know by how much
the rigidity or strength are reduced in high humidity conditions.
The following factors have been used to evaluate the change in
sensitivity to humidity conditions resulting from the treatment:
9
-50 i 2
i
v S"
-so jsfca
AETCM
V POi
WIC1
P -50
PEGt
FIG. 5. Relationship between the increase of MOR (Aa ) and increase of MOE
T
(AEj) by chemical treatment. Refer to Fig. 2 for definition of the symbols.

where SE and da
RH are the relative increase of E and cr, respectively,
RH
between the low humidity condition (30% RH) and high humidity
condition (98% RH), and 5E and <5<x are the corresponding values
RHxi RHu
for untreated wood.
where 1 and h refer to the low and high humidity conditions, respective-
ly. Figure 6 shows in the RH A-A<7
H plane the results obtained from
R
longitudinal bending of sitka spruce. We have called A the modulus R H
of elasticity (MOE) stability A<r the modulus of rupture (MOR)

RH
stability.
0-F4
-60 6
-AO Q
D A 2
<
-20
1 0 , | ,
-80 -60 -AO 1
Q 20 40 60
-20
Q
PO2 B02
6
Fig. 6. MOR stability (A<x ) compared to MOE stability (A ). Refer to

RH RH
Fig. 2 for definition of the symbols.
In order to improve the hardness or abrasion resistance, it is necess-

ary to have at least partial filling of the lumens as in A-2, -3, -5 and -6.
An improvement of the resistance to decay or to termite attack is
usually the case in any of the B-2-9 patterns, although there may be
exceptions.
EFFICIENCY O F CHEMICALLY TREATED W O O D S
Heat Treatment
Heating wood in a confined atmosphere induces the formation of
cohesive structures between the cell wall components, but at the same
time causes chemical changes as well as degradation. Wood treated with
heat belongs to type A-4 + B-2 of Fig. 1. Degradation of water-reactive
polysaccharides accompanies an increase of dimensional stability, so that
high levels of stability are obtained at the expense of strength properties,
hardness and abrasion resistance are reduced, permeability to water
vapor is reduced, liquid water repellency is increased, wood color
becomes darker, but density is almost unchanged [5]. However, if the
effect of degradation remains small and the formation of cohesive
structures dominates, the strength and specific elastic modulus will
increase, and the internal friction decrease [6]. Figure 5 shows two
examples of heat treatment yielding a reduction in the strength proper-
ties, H (3 hours at 180C and atmospheric pressure, weight percent gain
x
( W P G ) = - 2 - 4 % ) and H (idem for 10 hours, W P G = - 3 - 5 % ) , and one

2
example in which they increase slightly, H (idem for 10 hours). Figure 2

3
shows that ASE and ACE increase, but only very slightly. Figure 6 shows,
in the case of H , no change in M O E stability and a slight increase in
3
MOR stability.
Formalization
Formalization is a reaction involving the formation of oxymethylene
bridges between the hydroxyl groups of the constituents by formalde-
hyde. Reaction in the vapor phase yields pattern A-4 + B-2, and in the
liquid phase, pattern A-4 + B-3. In Fig. 2, F (vapor phase, HC1 as a
2
catalyst) and F (vapor phase, S 0 as a catalyst) have high ASE, but F

3 2 x
(liquid phase) has almost no dimensional stability. Vapor-phase treatment

improves the dimensional stability for a very small weight increase, and
ASE values up to 80-90% have been obtained. Water repellency is also
increased [5]. The creep from changing humidity is much reduced by
formalization, which is evident from the high ACE value for both F and 2
F in Fig. 2. Figure 5, on the other hand, shows that MOE remains

3
unchanged in the case of liquid-phase treatment ( F WPG = 0-5%) or

l 9
vapor-phase treatment with HC1 as a catalyst ( F , WPG = 0*7%, and F ,

2 4
W P G < 1 % ) , but is improved by vapor-phase treatment with S 0 as a 2
catalyst ( F , W P G = 21%). MOR is reduced in all cases, but especially

3
by vapor-phase treatment with HC1 as a catalyst. Figure 6 shows that, for

F , M O E stability is slightly increased while MOR stability is markedly

4
increased. Compressive strength and hardness are slightly decreased, and

tensile strength and abrasion resistance likewise. Treatment with a-
trioxymethylene or tetraoxymethylene and S 0 as a catalyst results in
2
less strength reduction than that with HC1 as a catalyst [ 7 ] . Figure 3

shows, for F (WPG = 2-9%, HC1 as a catalyst) and F (WPG = 3-9%,
6 7
S 0 as a catalyst), a large reduction in tan 3 but almost no change in E/y.

2
In the case of liquid-phase treatment, F (WPG =2-8%), tan<5 has

5
increased while E/y has decreased. In Fig. 4, the vapor phase data ( F , 6
F ) lie under the standard line, and those for the liquid phase ( F ) above.
7 5
Compared to untreated wood, the treatment with S 0 as a catalyst may

2
provide increases of up to 20% for E/y in the radial direction and

reduction for tan 3 of up to 30% and 45% in the fiber and radial
directions, respectively [8]. An overall improvement to the sound quality
of violins in which the soundboard or bridge has been formalized was
reported [ 9 ] . Vapor-phase treatment of medium-density fiberboards
markedly improves their dimensional stability, while formalization also
improves the resistance to decay.
Esterification
Esterification examples can be mentioned of acetylated wood,
phthalyated wood, wood oligoesterified with maleic anhydride, or
phthalic anhydride, and epichlorohydrin, wood esterified with maleic acid
and glycerol, wood treated with isocyanates, and wood esterified with
alkylketendimer. Among them, acetylation has been most thoroughly
studied [10]. Acetylation of wood is performed with acetic anhydride in
the liquid or vapor phase, and sometimes with ketene gas. Acetylation
leads to the formation of an A-4 + B-6 type of structure, hydrophilic
hydroxyl groups being substituted with hydrophobic and bulky acetyl
groups. Consequently, hygroscopicity decreases and water repellency
increases, density increases depending on the W P G value, a slight color
change occurs and stability to sunlight is improved. ASE increases with
increasing W P G up to 20%, but becomes sluggish for higher WPG; ASE
can reach 65-75%. The creep during humidity changes is much reduced
by acetylation; both ASE and ACE are improved as shown by A x
(WPG = 22-8%) in Fig. 2. E/y decreases proportionately to an increase in

WPG; the decrease of tan 3 with WPG, however, is preceded by a slight
increase. As both E/y and tan 3 have decreased in the case of A 3
(WPG = 20-4%) in Fig. 3, it lies under the standard line in Fig. 4.

Although the change in M O E or MOR remains small with increasing
WPG, Aj and A ( W P G = 19%) in Fig. 5 show that, for WPG approach-

2
ing 20%, M O E is slightly reduced while MOR has increased. Figure 6

shows that both M O E and MOR stability is improved by the treatment.
Hardness and abrasion resistance are slightly reduced. As it provides
dimensional stability without toxicity, acetylation is used in bathtubs and
on the internal walls of bathrooms, as well as for a flooring material in
combination with WPC treatment. The improvement in acoustic proper-
ties provided by acetylation, especially the tan S reduction together with
better stability in changing humidity conditions, led to the development
of a stereo speaker made with acetylated sounding boards [10]. Such a
product creates a natural sound and, at the same time, emphasizes the
beauty of the material. Acetylation of piano pin blocks [11] is an example
of an application that takes advantage of the reduction in mechano-
sorptive creep. Acetylation also improves the resistance to decay and
termite attack.
Esteriflcation with maleic acid (M) belongs to pattern A-4 + B-4 and -7
and, together with glycerol (MG), to pattern A-5 + B-5 and -7. These
treatments induce a density increase and a drastic color change. In Fig. 2,
both ASE and ACE have increased for M (WPG=15-8%) and
M G i ( W P G = 20-5%). Figure 5 shows that MOE is slightly reduced and
MOR much reduced by the M treatment; for the MG treatment, MOE is
almost unchanged and MOR either decreases slightly ( M G ^ or increases
slightly ( M G , W P G =18-23%). Figure 6 shows that both M O E and
2
MOR stability have been improved for M G . Figure 3 shows for M G

2 3
(WPG = 22-9%) a slight reduction in E/y and a marked reduction in

tan S, so that, in Fig. 4 it lies well beneath the standard line. MG
treatment improves hardness and abrasion resistance. It is used to obtain
a permanent set after a wood-bending operation and to improve the
dimensional stability of particleboards [12].
Etherification
Etherification of wood has been performed with epoxides, alkyl halide
or acrylonitrile. The addition of epoxides provides wood with the
structure of type A-4 + B-7. Under high humidity conditions, the treated
material becomes more hygroscopic than untreated wood, and permea-
bility to water vapor is somewhat increased. Propylene oxide ( P O , x
W P G = 25-7%) treatment yields much more mechanosorptive creep than

with untreated wood; in Fig. 2, P O ACE has a large negative value. As
x
was previously explained, the B-7 pattern may provide wood with high
ASE but, at the same time, it reduces ACE considerably, thus contrasting
with acetylation which belongs to the B-6 type. Figure 3 shows for both
P 0 (WPG = 21-9%) and B 0 (WPG = 18-7%) some decrease in E/y and
3 3
a large increase in tan <5, so that, in Fig. 4, both lie above the standard
line. Figure 5 shows for P O , P 0 (WPG = 25-28%), BO and B 0
t 2 i 2
(WPG = 23-27%) a diminution of MOE, and either a small decrease or a

slight increase of MOR. Compressive strength in the fiber direction is
reduced. Figure 6 shows a large reduction of M O E and MOR stability.
Hardness is almost unmodified by the PO treatment (WPG = 20-22%),
color change is small, and resistance to decay and fungal degradation is
improved [13]. The treatment of wood with epoxides improves their
dimensional stability (high ASE) and, at the same time, yields high
hygroscopicity; it is recommended, therefore, for the treatment of wood
used for internal wall materials. The increased plasticity also suggests
applications in such forming operations as wood bending.
Treatment with Phenol Formaldehyde Resin

An aqueous solution of phenol formaldehyde resin with a low molecu-
lar weight of around 300 penetrates the cell wall, and patterns A-5 + B-9,
-6 and -3 may be expected. Density is increased; generally the color turns
to reddish brown, but with saligenin or a low molecular weight resin,
there is almost no color change. Figure 2 shows that by treating with low
molecular weight resin ( P WPG =12-9%), both ASE and ACE are
l 5
improved. ASE of up to 60-70% can be attained [ 5 ] . Permeability to

water vapor decreases and liquid water repellency decreases. Figure 3
shows that, in the case of a low molecular weight resin ( P , 3
WPG = 44-6%), E/y is slightly reduced and tan 5 somewhat reduced,

while, for a resin with high molecular weight that does not penetrate into
the cell wall ( P , W P G = 22-7%), E/y is somewhat reduced and tan 8 is
4
increased. These results illustrate the necessity of resin penetration into

the cell wall to improve the acoustic properties of wood used for a
soundboard, for example. In the case of a saligenin treatment with
WPG = 5-8%, E/y is almost unchanged in the fiber direction, but
increased by 15% in the radial direction; tan d is reduced by 17% in the
fiber direction and by 37% in the radial direction [14]. Figure 5 shows
for both P i (WPG =12-9%) and P (WPG = 40-47%) an increase of
2
MOE and MOR, and especially for the higher WPG. Figure 6 shows that
both M O E and MOR stability are increased. Compressive strength is
improved in the fiber direction as well as transversely to the fiber.
Regarding the tensile properties, the longitudinal elastic modulus in-
creases, but the longitudinal strength decreases while the transverse
strength increases. Impact bending strength is reduced, hardness is

increased, but abrasion resistance is reduced. Electrical insulation and
heat resistance are improved. The treatment provides better resistance to
decay and termite attacks, as well as high resistance to acids [5].
Impregnation of the resin with low molecular weight compounds has
been tried for the surface densification of low-density conifers [15], as
well as for the dimensional stabilization of particleboard.
Wood-Inorganic Material Composite (WIC)

This product is obtained by immersing wood in an aqueous solution of
BaCl , and then by immersing in an aqueous solution of ( N H ) H P 0
2 4 2 4
[16]. This treatment yields the pattern A-5 + B-8, and provides a material
with fireproof qualities that can be used for fire prevention. It increases
the density but causes almost no color change. The penetration of salts
( B a P 0 and B a H P 0 ) into the cells may result in some ASE increase,
4 4
but ACE should be reduced. Figure 2 shows for WIC (WPG = 42-5%)
X
no dimensional stability with a negligible ASE value, but a negative ACE

(i.e. more creep than untreated wood). In Fig. 5, no change in M O E and
MOR is apparent. However, for WPG approaching 100%, an MOR
increase of 30-40% together with a decrease of impact strength can be
seen. Figure 3 shows for W I C (WPG = 49-1%) a drastic diminution of
2
E/y and an increase of tan S. Protection from decay and fungal attack are
improved.
Wood-Plastic Composites (WPC)

These are obtained by the introducing vinyl monomers to wood, which
are then polymerized. Structures of type A-5 and -6 + B-6 and -9 can be
obtained under experimental conditions, but pattern A-2 and -3 + B-l
can be expected in industrial practice. The density can increase by
30-80%, and color change remains slight. Hygroscopicity and permea-
bility to water vapor decrease and liquid water repellency increases. In
Fig. 5, W P C (WPG =105%) combined with methylmethacrylate
X
(MMA) yields an increase of 50% for M O E and 60% for MOR. Tensile
strength, compressive strength, bending strength, impact bending
strength and hardness all are increased by 100-200% [ 5 ] . Pattern A-5
and -6 may improve ASE, but not pattern A-2 and -3, as shown by
W P C ! in Fig. 2, where both the ASE and ACE factors are close to 0%.
MMA induces a large increase in elastic modulus but, at the same time,
a higher relative increase of density, so that for W P C (WPG =121%)
2
in Fig. 3, the net result is a decrease of E/y. The large increase in tan d
can be attributed to the influence of M M A viscosity contained in the

lumen on the internal friction. As a result, in Fig. 4, W P C lies above
2
the regression line and on the left side. In addition to its effect on
mechanical properties, WPC improves the surface quality. It is widely
used in architectural features, furniture, sporting goods, crafted objects,
etc.
Wood impregnated with expoxy resins is another case of pattern A-2
and -3 + B-l [5]. Epoxy resin that is partially polymerized is intro-
duced into the cell lumens and cured afterwards. There is almost no
ASE as there is no penetration into the cell wall. Permeability to water
vapor is much reduced and water repellency is increased. Strength,
hardness and abrasion resistance increase drastically. It is used for
wooden boat hulls or the outer ply of plywood, and also for strengthen-
ing decayed wood.
Impregnation with Polyethylene Glycol (PEG)

This composite is prepared by introducing a PEG solution into the
material and then drying. The molecular weight should be lower than
1000 to allow penetration into the cell wall and result in pattern A-
5 + B-8; for a molecular weight of about 10000, there is no penetration at
all and pattern A-2 and -3-fB-l results. Figure 2 shows the result
PEG-1000 ( P E G i , WPG =110%) where ASE is extremely high (81%),
but ACE extremely negative (around 200%, i.e. three times more
mechanosorptive creep than in untreated wood). PEG molecules intro-
duced into the cell wall participate in the sorption process, and the
plasticizing effect of the adsorbed water triggers molecular mobility.
Consequently, E/y is reduced and tan 3 is increased. Figure 3 shows for
P E G (WPG = 48-8%, PEG-600) and P E G (WPG = 38-8%, PEG-1000)
3 4
a large decrease in E/y and a drastic increase in tan 8, so that, in Fig. 4,

the corresponding marks lie well above the regression line. Comparative-
ly, for P E G (WPG = 3 6 1 % , PEG-10 000), the changes remain very
5
small. Figure 5 shows for PEG and P E G (WPG = 36-42, PEG-1000), a

X 2
drastic reduction of M O E and MOR. However, in Fig. 6, their stability

factor is almost the same as that of untreated wood. The extremely good
ASE value obtained allows the use of PEG treatment to reduce checking
during drying.
Paraffin impregnation is similar to PEG-10 000 as giving an A-2 + B-1
type. The impregnation of wood with paraffin wax mixed with preserv-
atives is used to prevent checking and to provide decay resistance and
water repellency.
154 MISATO NORIMOTO AND JOSEPH GRII
REFERENCES
[1] P.U.A. Grossman, Wood Sci. Techn., 10, 165 (1976).

[2] M. Norimoto, J. Gril and R.M. Rowell, Wood and Fiber Sci., 24, 25 (1992).
[3] H. Akitsu, M. Norimoto and T. Morooka, Mokuzai Gakkaishi (J. Japan
Wood Res. Soc), 37, 590 (1991).
[4] M. Norimoto, F. Tanaka, T. Ohogama and R. Ikimune, Wood Res. Techn.
Notes, No. 22, 53 (1986).
[5] R.M. Rowell and P. Konkol, Gen. Tech. Rep. FPL-GTR-55, 1 (1987).
[6] S. Kadita, T. Yamada, M. Suzuki and K. Komatsu, Mokuzai Gakkaishi, 1,
34 (1961).
[8] K. Minato and H. Yano, Mokuzai Gakkaishi, 36, 362 (1990).
[9] H. Yano and K. Minato, Proc. 13th Internat. Cong, on Acoustics, Yugo-
slavia, 3 , 45 (1989).
[10] M. Norimoto, Wood Res. and Techn. Notes, No. 24, 13 (1988).
[11] H. Fujimoto, T. Anazawa and K. Yamagishi, Proc. 23rd Internat. Particle-
board/Composite Materials Symp., Wash. State Univ., 95 (1989).
[12] H. Yano, J. Mukudai and M. Norimoto, Mokuzai Gakkaishi, 34, 94 (1988).
[13] R.M. Rowell and W.D. Ellis, FPL Res. Paper, FPL-451, 1 (1984).
[14] H. Yano, N. Kanou and J. Mukudai, Mokuzai Gakkaishi, 36, 923 (1990).
[15] H. Inoue, M. Norimoto, Y. Otsuka and T. Yamada, Mokuzai Gakkaishi, 37,
227 (1991).
[16] S. Hirao, Wood Industry, 44, 628 (1989).
Plasticization of Wood and its Application
NOBUO SHIRAISHI

Kyoto University, Kitashirakawa Oiwake-cho, Sakyo-ku, Kyoto 606-01,
Japan
ABSTRACT
W o o d has been found to be convertible to a thermally flowable material by chemical
modifications, in which thermoplastic properties can be conferred by internal plasticization.
In some cases, the thermoplastic properties must be supplemented by externally blending
the modified wood with plasticizers. Chemically modified wood and untreated wood have
both been found to dissolve or liquefy in various solvents and solutions. The applications
of liquefied wood solutions and plasticized wood are described for preparing films, sheets
and other moldings from the chemically modified wood, for three-dimensionally cured
plastic-like wooden boards, for deep-drawable hard boards, and for hot-melt adhesives.
Also described is the surface-layer plasticization of wood for surface-densification and
embossing, and the preparation of wood-based reactive adhesives.
Keywords: plasticization of wood, chemically modified wood, thermoplastic properties,

plasticizers, wood utilization
INTRODUCTION
The methods for processing wood are very limited. While metals,
plastics and glass can be processed in the liquid phase at high tempera-
tures, 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
155
156 NOBUO SHIRAISHI
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 - 5 ] .
THERMOPLASTICIZATION OF W O O D
Wood differs from the plastics in the basic property of thermoplasticity.

A search of previous literature on the thermoplasticity of wood revealed
that only phenomena up to the second order transition and/or thermal
softening had been investigated, and not the phenomenon of thermal flow
[6-8].
The author recently found that wood could be converted into a
thermally flowable material by chemical modification such as that by
esterification and etherification [ 9 ] . Chemical modification does not
necessarily require special techniques and therefore conventional and
simple methods can be used for this purpose [ 1 - 5 ] .
The phenomenon of thermal flowability can be explained in terms of the
internal plasticization of wood, which can change the fundamental
properties, including thermoplasticity. The introduction of large non-polar
substituent groups into wood can result in a chemically modified material
with high thermoplasticity. For example, esterified woods with large acyl
groups resulting from caproylation and lauroylation exhibit thermal flow
behavior. Photos (a) and (b) in Fig. 1 reveal that lauroylated wood meal can
be molded into a transparent sheet by hot-pressing at 160 C.
However, when a small substituent group and/or polar group was
introduced, such thermal fluidity could not be achieved, so that this
modification alone could not produce plastic properties. This lack of
plasticity could be influenced by additional external plasticization. For
example, allylated wood, which did not show complete thermal fluidity,
was changed to a flowable and moldable material after blending with
certain amounts of polymethylmethacrylate (PMMA) and dimethyl
phthalate (Fig. 2). From these results, it can be concluded in principle
that, irrespective of the molecular size of substituent groups, thermoplas-
tic properties can be imparted by internal plasticization through chemical
modification. If the effect is insufficient, it can be supplemented by
external plasticization.
In addition to this fundamental technique offering enough thermoplas-
ticity to chemically modified wood with a small or polar substituent,
several other attempts have been made by acetylated wood.
PLASTICIZATION OF WOOD AND ITS APPLICATION 157
,<v . -1,1*14 V" <
FIG. 1. Lauroylated wood meal (a), a film prepared from the lauroylated wood
meal by compression molding (b), a piece of acetyl-propionylated wood (c), and
compression moldings from acetyl-propionylated wood (d and e).
The thermoplasticity of acetylated wood was found to vary with the

acetylation method adopted. The plasticity of acetylated wood prepared
by reacting with a trifluoroacetic anhydride-acetic acid mixture (the
TFAA method) is high, and the products clearly melted at about
300-320 C under a low pressure of 0-3 MPa. However, this was an
exceptional case, because acetylated woods prepared by other conven-
tional procedures have not shown thermal flow. Thus, techniques for
converting these thermally non-meltable acetylated woods into plastic
materials have been investigated.
As a more effective method than that of mere blending, a method of
grafting a vinyl monomer after chemical modification has been develop-
ed [10, 11]. In this case, the resultant vinyl polymers act as a plasticizer
within the wood. This grafting method is similar in principle to the
already mentioned method of external plasticization, in which synthetic
polymers are used as the plasticizer in some cases. However, in the latter
case, it has not been easy to find the most appropriate synthetic
158 NOBUO SHIRAISHI
FIG. 2. Plasticization of allylated wood (left) by blending with dimethyl

phthalate (center) and polymethylmethacrylate-dimethyl phthalate (right).
polymers to act as suitable plasticizers, because compatibility between

polymers rarely occurs. To overcome this difficulty, grafting techniques
can be used. Even when a combination of acetylated wood and synthetic
polymers could not produce a homogeneous transparent film by a simple
blending technique, grafting can result in a molded transparent film.
Investigations have also been advancing to further enhance the
thermofluidity of chemically modified wood. Among them, the method of
using trifluoroacetic acid (TFA) as a reagent for pretreatment to selective-
ly split bonds combining lignin units [12], as well as methods for
chlorinating the aromatic ring of lignin in chemically modified wood [13]
are very effective and specific. Acetylated wood obtained by the former
method shows an apparent melting temperature of about 210 C, this
temperature being almost 90 C lower than the apparent melting tem-
perature of cellulose triacetate. The compatibility and/or mutual plastic-
ization among the acetylated wood components can be considered to
account for difference in flow temperature. The plasticization of cellulose
acetate by interacting with acetylated lignin has been postulated and
proved experimentally [12].
As to the second method for enhancing the thermofluidity of chemical-
ly modified wood, Morita and Sakata have investigated the effect of
post-chlorinating cyanoethylated wood [13]. It was shown that the flow
temperature of the cyanoethylated wood could be reduced by 100-
120 C with this method. This lowering of the flow temperature is thought
to be attributable to a loosening of the lignin structure by chlorine
substitution, the action of the chlorinated lignin as an external plasticizer
on the cyanoethylated cellulose within the modified wood, and the

decrystallization effect from chlorination [13].
Mixed esterification has also enhanced the thermoplasticity of acetylated
wood [4, 14, 15]. As shown in (c), (d) and (e) of Fig. 1, thermofluidity was
conferred by mixed acetyl-propionylation. It is known that this kind of
mixed esterification often gives amorphous cellulose derivatives.
APPLICATION OF THERMOPLASTICIZED W O O D
Based on the findings just described, several applications have been

attempted. Examples are the preparation of films, sheets and moldings
from chemically modified wood [10, 14-19], the preparation of three-
dimensionally cured plastic-like wooden boards [20, 21], the preparation
of deep-drawable hard boards [1, 5], application to hot-melt adhesives
[1], and plasticization of the surface-layer of wood for surface-densifica-
tion and embossing [22].
There have been many trials for preparing films, sheets and moldings.
Moldings with various physical properties could be obtained from
chemically modified woods and their blends with synthetic polymers. For
example, the mechanical properties of films from benzylated wood are
comparable with those of the common synthetic polymers. The mechan-
ical properties of a benzylated wood film can be further enhanced by
blending with polystyrene (PS), in which styrene-maleic anhydride
copolymer (SMA) is added in a small amount. The presence of small
amounts of SMA also improves the tensile properties of blended films
from benzylated wood and modified polyphenylene oxide. In these cases,
SMA apparently acts to create compatibility.
Three-dimensionally cured plastic-like wooden boards have been pre-
pared from chemically modified wood [20, 21]. This is a case of
thermosetting molding, and the important point is that of making
use of chemicals which can act as plasticizers for the modified wood,
and at the same time can react with the components of the modified
wood during hot-press molding. For instance, phthaloylated wood as
an example of carboxyl-group-bearing esterified wood was used as
the chemically modified wood with bisphenol A diglycidyl ether as the
plasticizer and cross-linking agent. In the presence of the latter, a
thermally flowable molding results with the cross-linking reaction
during hot-pressing. The prepared board containing 60% of wood has
a compressive strength as high as 200 MPa, and other properties
160 NOBUO SHIRAISHI
such as the hardness, thermo-moldability and water resistance are also

satisfactory.
The production of three-dimensionally moldable fiber boards, or
deep-drawable fiber boards has also been advancing. In this case, wood
fibers are chemically modified to a level that makes the product thermo-
plastic but not thermally flowable, and fiber boards that are highly
moldable have been obtained (Fig. 3).
FIG. 3. Hot-press drawability of hard boards at the density of approximately

10. An ordinary hard board (right) and a hard board prepared from acetyl-
propionylated wood pulp with low substitution (left) are shown.
C O M M E N T S O N THE LIQUEFACTION O F W O O D
The term 'liquefaction of lignocellulosics' has hitherto chiefly referred

to those procedures for producing oil from biomass under very severe
conditions for conversion 23-25]. For example, Appel et al have
converted cellulosics to oil by using homogeneous N a C 0 catalyst in
2 3
water and a high-boiling-point solvent mixture (anthracene oil, cresol,

etc.) at a pressure of 140-240 atm. with synthesis gas, C O / H [25].
2
Treatment for one hour 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.
This review presents recent progress on lignocellulosic liquefaction

under milder treatment conditions; that is, at a temperature of 2 4 0 -
260 C without a catalyst, or at a temperature of 80-150 C with an acidic
catalyst. One special group of chemically modified woods can be dis-
solved in cresols even at room temperature, as will be shown later. The
liquefaction or dissolution of chemically modified wood has been devel-
oped [17, 19, 26, 27], and the liquefaction of untreated wood has also
been found to be possible [2, 28].
LIQUEFACTION O F CHEMICALLY M O D I F I E D W O O D
Chemically modified woods have been found to liquefy and dissolve in

neutral aqueous solvents, organic solvents or solutions, depending on the
characteristics of the modified wood [1, 14, 15, 18]. So far, three methods
have been found for wood liquefaction.
The first trial on the liquefaction of wood was accomplished by using
very severe dissolution conditions [29]. One example used wood samples
esterified with a series of aliphatic acids which could be liquefied in
benzyl ether, styrene oxide, phenol, resorcinol, benzaldehyde, aqueous
phenols, and a chloroform-dioxane mixture after treating at 200-260C
for 20-150 min.
Another method for liquefaction makes use of solvolysis during
the process [28, 30]. By using conditions which allow phenolysis of
part of the lignin, especially in the presence of an appropriate catalyst,
the liquefaction of chemically modified wood into phenols could be
accomplished under milder conditions (at 80 C for 30 to 150 min).
Allylated wood, methylated wood, ethylated wood, hydroxyethylated
wood, acetylated wood and others have been found to dissolve in
polyhydric alcohols such as 1,6-hexanediol, 1,4-butanediol, 1,2-
ethanediol, 1,2,3-propane triol and bisphenol A, using the lique-
faction conditions just described. The liquefaction process can produce
a paste-like solution with a quite high concentration of wood solute
(70%).
The third method of liquefaction or dissolution involves post-chlorina-
tion [31]. When chemically modified woods are chlorinated, their solubil-
ity in solvents is tremendously enhanced. For example, at room tempera-
ture, cyanoethylated wood can dissolve in o-cresol by only 9-25%.
However, once chlorinated, it can dissolve almost completely in the same
solvent at room temperature.
162 NOBUO SHIRAISHI
APPLICATIONS O F THE LIQUEFACTION OR OF

CHEMICALLY M O D I F I E D W O O D
There are many potential applications for the liquefaction and dissol-
ution of chemically modified wood. Examples include the fractionation of
modified wood components [31, 32], the preparation of solvent-sensitive
and/or reaction-sensitive wood-based adhesives [1, 3, 27, 30, 33], the
preparation of resinified wood-based moldings such as the foam type [27],
and the preparation of wood-based fibers and their conversion to carbon
fibers [34].
To fractionate modified wood components, the dissolution-precipita-
tion technique has been successfully used [31, 32].
In the preparation of adhesives from chemically modified wood, phenols,
bisphenols and polyhydric alcohols have been used as solvents, and the
resins contained a significant amount of modified wood [1, 3, 27, 30, 33].
The combined use of these reactive solvents with reactive agents, e.g., a
cross-linking agent and/or hardener, has given rise to phenol-formalde-
hyde resins (such as resol resin), polyurethane resins, epoxy resins, etc., all
of which contained a significant amount of chemically modified wood
and possessed outstanding gluability [1, 3, 30, 33].
Molded materials such as foams or pressed shapes can also be obtained
from chemically modified wood solutions. Foams, for example, can be
prepared by adding an adequate amount of water as a foaming agent and a
polyisocyanate compound as a hardener to a 1,6-hexanediol solution of
allylated wood, which is then mixed well and heated. When heated at
100 C, foaming and resinification are initiated within two minutes and
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 has a low apparent density of
0 0 4 g / c m , substantial strength, and significant restoring force to compres-
3
sive deformation [27] (Fig. 4).

One other application of modified wood solutions is the formation of
filaments or fibers [34]. After making a phenol solution of acetylated wood,
hexamethylene tetramine is added, and the mixture heated at 120 C to
promote the addition-condensation reaction for a resinified solution with
high spinnability. From this solution, filaments are spun and hardened in
an oven at a certain heating rate. The maximum temperature for hardening
is 250 C, and endless filaments can easily be obtained by this method.
These filaments can then be carbonized to produce carbon filaments.
The strength of these carbon filaments has been measured according to
FIG. 4. Appearance of polyurethane foam from allylated wood.
Japan Industrial Standards, and a tensile strength of up to 1 GPa has so

far been obtained. This strength is comparable to that of pitch carbon
fibers of general-purpose grade. Greater strength can be expected by
improving the methods for spinning and carbonization.
Morita et al. [35] have prepared latex from chemically modified wood
(cyano ethylated (CE) wood). From a solution of chlorinated CE wood
(Cl-CE wood)/acetone/water, the acetone was gradually evaporated with
stirring, resulting in latex. The viscosity of the latex was as low as
one-tenth of that of a corresponding dimethylformamide solution of
Cl-CE wood, and the mechanical strength of the film prepared from the
latex was almost the same as that of film cast from an organic solution.
LIQUEFACTION O F U N T R E A T E D W O O D
The liquefaction and dissolution of chemically modified woods has so

far been reviewed. More recently, untreated wood has also been found
164 NOBUO SHIRAISHI
to liquefy in several organic solvents [2, 36]. For example, after treating
at around 250 C for 15-180 min, wood chips and wood meal were
liquefied in phenols, bisphenols, alcohols (benzyl alcohol), polyhydric
alcohols (1,6-hexanediol and 1,4-butanediol), oxyethers (methyl cello-
solve, ethyl cellosolve, diethylene glycol, triethylene glycol, and poly-
ethylene glycol), 1,4-dioxane, cyclohexanone, diethyl ketone, and ethyl
n-propyl ketone.
The liquefaction of untreated wood can also be achieved at a lower
temperature of 150C and at atmospheric pressure in the presence of an
acid catalyst, phenolsulfonic acid and sulfuric acid having been used [37].
It is possible to obtain paste-like solutions with a high concentration of
wood solute of up to 70% (Fig. 5). After liquefaction, the wood compo-
nents were found to have been degraded and became reactive. The
obtained wooden solute can be used to prepare adhesives and other
moldings, opening a practical new field for utilizing wood materials.
FIG. 5. Wood meal and a phenol solution of wood.
APPLICATION OF THE LIQUEFACTION OF U N T R E A T E D

WOOD
From solutions of untreated wood, almost the same products have

been prepared as those from chemically modified wood [ 3 ] . For
example, resol-type phenol resin adhesives prepared from five parts of
wood chips and two parts of phenol did not require severe adhesion
conditions and were comparable to the corresponding commercial
adhesives in their gluability. Acceptable waterproof adhesion was at-
tained from the adhesives after glueing wood veneers at 120-130 C with
a hot-pressing time of 0-5 min to 1 mm thick plywood. This adhesion
temperature of 120 C is at least 15 C lower than that ordinarily used
for resol resin adhesives.
As a second example, foams can be prepared from untreated w o o d -
polyethylene glycol solutions [5]. Both soft and hard types of foams can be
produced according to the preparation conditions. The prepared foams
had a density of around 0 0 4 g/cm , substantial strength, and strong restor-
3
ing force against deformation. These results imply that the wood compo-
nents were not merely blended within the foam bubbles, but also played an
important role in maintaining the dimensional stability of the foams.
The third application example is Novolak-resin type of moldings
prepared from untreated wood-phenol solutions [38]. After one part of
wood meal had been liquefied in two parts of phenol, the unreacted phenols
were distilled under reduced pressure. The resulting liquefied and reacted
wood-phenol powder can be cured directly after wood meal filler and
hexamethylene tetramine have been added and hot-pressed at 170-200 C.
The flexural strength of the moldings was comparable to those made from
commercial Novolak, and an example is shown in Fig. 6.
FIG. 6. Appearance of Novolak-like moldings from a phenol solution of wood.

166 NOBUO SHIRAISHI
The carbon fiber already described could also be prepared from an

untreated wood solution, and a tensile strength of up to 1-2 GPa has been
obtained so far. Even better physical properties can be expected with
more development.
The present stage of studies on wood plasticization has been

briefly reviewed. This is a new field for the chemical processing of
wood with high future potential, although more studies need to be
undertaken.
REFERENCES
[1] N. Shiraishi, Mokuzai Gakkaishi (J. Japan Wood Res. Soc), 32, 755 (1986).
[2] N. Shiraishi, Mokuzai Kogyo (J. Wood Technol. Assoc., Japan), 42, 42 (1986).
[3] N. Shiraishi, Y. Tamura and N. Tsujimoto, Mokuzai Kogyo, 42, 492 (1987;
43, 2 (1988).
[4] N. Shiraishi, Wood and Cellulose Chemistry, Marcel Dekker Inc., New York,
861 (1990).
[5] N. Shiraishi, Cellusosics Utilization; Research and Rewards in Cellulosics,
Elsevier Applied Science, London, 97 (1989).
[6] D.A.I. Goring, Pulp and Paper Mag. Can., 64, T-517 (1963).
[7] S.-Z. Chow, Wood and Fiber, 3, 166 (1972).
[8] S.H. Baldwin and D.A.I. Goring, Svensk Paperstid., 71, 641 (1968).
[9] N. Shiraishi, T. Matsunaga and T. Yokota, J. Appl. Polymer Sci., 24, 2361
(1979).
[10] N. Shiraishi, T. Aoki, M. Norimoto and M. Okumura, Graft Copolymeriz-
ation of Lignocellulosic Fibers, A. C. S. Symp. Series, 187, Amer. Chem. Soc,
Washington D.C., 321 (1982).
[11] N. Shiraishi, T. Aoki, M. Norimoto and M. Okumura, Chemtech, (June)
366, 1983.
[12] N. Shiraishi and M. Yoshioka, Sen-i Gakkaishi, 42(6), T-346 (1986).
[13] M. Morita and I. Sakata, J. Appl. Polymer Sci., 31, 831 (1986).
[14] N. Shiraishi, Chemistry of Wood Utilization, Kyoritsu Publ. Co., Tokyo, 294
(1983).
[15] N. Shiraishi, Advanced Technique and Future Approach in Wood Chemicals,
CMC Inc., Tokyo, 271 (1983).
[16] N. Shiraishi, Mokuzai Kogyo, 37, 194 (1982).
[17] N. Shiraishi, Kobunshi Kako, 31, 500 (1982).
[18] N. Shiraishi, Sen-i to Kogyo, 39, 95 (1983).
[19] N. Shiraishi and K. Goda, Mokuzai Kogyo, 39, 329 (1984).
[20] H. Matsuda and M. Ueda, Mokuzai Gakkaishi, 31, 903 (1985).
[21] H. Matsuda, Mokuzai Kogyo, 44, 637 (1989).
[22] T. Aoki, Proceedings of 11th Symposium on Chemical Processing of Wood,
Tsukuba, 7 (1981).
[23] C. Vanasse, E. Chornet and R.P. Overend, Can. J. Chem. Eng., 6 6 , 112
(1988).
[24] H.R. Appel, I. Wender and R.D. Miller, U.S. Bureau of Mines, Technical
Progress Report 25, 1969, 5.
[25] H.R. Appel, Y.C. Fu, E.G. Illig, F.W. Steffgen and R.D. Miller, U.S. Bureau
of Mines, RI 8013, 1975, 27.
[26] N. Shiraishi, Japan Patent Publication Examined, 1988-1992 (appl. June 6,
1986).
[27] N. Shiraishi, S. Onodera, M. Ohtani and T. Masumoto, Mokuzai Gakkaishi,
31, 418 (1985).
[28] N. Shiraishi, N. Tsujimoto and S. Pu, Japan Patent Publication Unex-
amined, 1986-261358 (appl. May 14, 1985).
[29] N. Shiraishi, Japan Patent Publication Unexamined, 1982-2301; 1982-2360.
[30] N. Shiraishi and H. Kishi, J. Appl. Polymer Sci., 32, 3189 (1986).
[31] M. Morita and I. Sakata, Cellulose Chem. Technol, 2 1 , 255 (1987).
[32] R.A. Young, S. Achmadi and D. Barkalow, Preprints of CELLUCON' 84,
Cartrefle-Wrexham, Wales, 64 (1984).
[33] H. Kishi and N. Shiraishi, Mokuzai Gakkaishi, 32, 520 (1986).
[34] N. Tsujimoto, Proceedings of Nth Symposium on Chemical Processing of
Wood, Kyoto, 17 (1984).
[35] M. Morita, Y. Yamawaki, M. Shigematsu and I. Sakata, Mokuzai Gak-
kaishi, 36, 659 (1990).
[36] N. Shiraishi, N. Tsujimoto and S. Pu, Japan Patent Publication Unex-
amined, 1986-261358 (appl. May 14, 1985).
[37] Y. Tanihara, K. Kato, S. Pu, S. Saka and N. Shiraishi, Abstracts of Papers
Presented at 39th National Meeting, Japan Wood Res. Soc, Okinawa, 321
(1989).
[38] K. Kato, M. Yoshioka and N. Shiraishi, Abstracts of Papers Presented at
40th National Meeting, Japan Wood Res. Soc, Tsukuba, 26 (1990).
Wood-Phenol Resin Fiber
NAOHIKO TSUJIMOTO
Central Research Laboratory, Oji Paper Company Ltd, 1-10-6 Shinonome,

Koto-ku, Tokyo 135, Japan
ABSTRACT
Completely new fields for wood utilization have been opened by the finding that
chemically modified and non-modified wood can both be dissolved or liquefied in several
organic solvents. As an application of this phenomenon, wood-based filaments were formed
by spinning from a wood-phenol solution with hexamethylenetetramine. Successful carbon-
ization of the filaments was carried out at 900 C, and the tensile strength of the resulting
carbon fiber was 11 GPa.
Keywords: wood-phenol resin fiber, plasticization of wood, fiber-spinning, carbon fiber
INTRODUCTION
New fields for wood utilization of high added value have been
developed by the plasticization of wood through relatively simple chemi-
cal modifications such as esterification and etherification [1]. There are
two application categories for wood plasticization: the first is related to
products involving thermal fluidity such as hardboard, surface-plasticized
wood and wood-based films, and the second is related to products of
solubilization in solvents, like adhesives [1]. In addition to the dissol-
ution or liquefaction of plasticized wood, new techniques for achieving
the same with non-plasticized wood have been developed [2].
In this paper, three fiber-forming techniques involving different starting
materials and spinning methods are studied: (1) swelling or dissolving
plasticized wood in an organic solvent, before spinning and evaporating
the solvent; (2) similar to the first method, but including a solvent as one
of the fiber-forming components; (3) using a phenol resin solution of
non-modified wood to spin from.
169
170 NAOHIKO TSUJIMOTO
The object of spinning wood fiber is to induce the valuable character-

istics of a filament such as its formability into various shapes, and
chemical and physical availability of its relatively large surface area.
These are the first fiber-forming techniques to utilize almost all the
components of wood.
FILAMENTS F R O M PLASTICIZED W O O D
The fiber-forming procedure that was used is similar to the dry

spinning method for acetate filaments. Two spinning solutions were
chosen, one of acetylated wood swollen with an organic solvent, and the
other of acetylated wood half dissolved in an organic solvent.
Experimental Procedure
The wood samples were radiata pine wood meal (48 mesh pass-60
mesh on, 170 mesh pass-300 mesh on, and 300 mesh pass) and refiner
mechanical pulp (RMP) was made from radiata pine. The chemical
composition of RMP was the same as that of the original wood, except
for the hot-water extractives. Chemical modification was carried out by
acetylating with trifluoro acetic anhydride [ 3 ] . The yield of acetylated
wood in both the meal and RMP was 145% based on weight, and its
converted substitution degree of acetylation was more than 2-5. The
acetylated wood was swollen or half-dissolved in chloroform at 25 C,
130C or 150C. These swollen or half-dissolved materials were put into
a syringe-shaped spinning device, and then spun through a nozzle of
0-5 mm or 1 0 mm in diameter. The filaments were fixed in shape by
evaporating the chloroform with heated air. These fixed filaments were
elongated in a range from 20% to 60% at 160-200 C, and the tensile
strength of the elongated filaments was measured according to the
procedure of Japan Industrial Standard (JIS) L 1069.
Filament Strength from the Swollen Wood

The acetylated wood meal and RMP were suspended, swollen in
chloroform and then spun at 25 C, before being elongated and measured
for filament strength.
The influence of particle size in the wool meal on the tensile strength of
the elongated filaments is shown in Fig. 1. The tensile strength of the
filaments composed of small particles was higher than that of larger
particles, due to the higher density from the smaller particle size.
WOOD-PHENOL RESIN FIBER 171
1001 1 1 1 1 1 1 1 r
90 h
(MPa)
80-
70 "
Tensile strength
60-
50-
40-
30-
201 i i I i I l i I
0.5 1.0 1-5
Density (g/cm ) 3
FIG. 1. Relationship between tensile strength and fiber density for differing
diameter of spinning nozzle.
1 mm and [] 0-5 mm for mechanical pulp at high temperature

O 1 rnm and d 0-5 mm for mechanical pulp at room temperature
A 1 mm and A 0-5 mm for wood meal at room temperature
48P: 48 mesh pass-60 mesh on; 170P: 170 mesh pass-300 mesh on; 300P: 300 mesh
pass.
The filaments composed of acetylated R M P were stronger than those

from the wood meal, because the former filaments had a higher density,
and their fibrous structure also seems to have contributed to the strength
by entanglement. The wood meal particles had lost their fiber shape at
the stage of wood milling.
The tensile strength also depended on the size of the spinning nozzle,
the fibers spun through the narrower nozzle showing higher tensile
strength. This is because one of the factors for determining the tensile
strength of a filament is the frequency of defects occurring in the fila-
ment.
Filaments from Half-Dissolved Wood

To obtain higher density filaments from swollen acetylated wood, the
acetylated wood was treated in chloroform at temperatures of 130 C and
150C, because room temperature was not enough to achieve complete
swelling. As shown in Fig. 1, this higher temperature treatment resulted in
a higher tensile strength from an increase in the density. At 130C and
150 C when compared with room temperature, the acetylated wood
swelled more or half-dissolved, while still keeping its own shape. In other
words, the microstructure of the spun filament became more loose in the
solvent and, consequently, was fixed more tightly at both stages of
solvent evaporation and of elongation, which led to the increase in
density.
Elongation Treatment
The relationship between the elongation rate and the tensile strength of
filaments was investigated.
As shown in Fig. 2, the filaments made from acetylated RMP swollen
in chloroform at 130C and 150C had a tensile strength increasing in
proportion to increasing elongation rate. The determining factor for
tensile strength seems to have been not only the density already de-
scribed, but also the orientation of fibrils in a filament after the elonga-
tion treatment. In comparison with an acetate fiber (the tensile strength is
around 150 MPa), a dry spun filament obtained from the half-dissolved
solution of plasticized wood had about half the tensile strength.
100 r
90 -
. 1 . , , , , 1
Tensile strength (MPa)
80
70-
60 -
50 -
40
30 >
C
20 -
10 20 30 40 50 60 70 80
Elongation (%)
FIG. 2. Relationship between the tensile strength and elongation rate for
acetylated RMP filaments.
130 C O 150 C
To obtain a stronger filament, the density of the filament should be

increased, and the filament should be more highly oriented. The former
method is limited, because the true specific gravity of both wood
components and their derivatives is 1-40-1-55, these figures being close to
those of the filaments already. Filament orientation also has its limitation
in terms of obtaining stronger filaments, because the maximum elonga-
tion rate was around 65% as shown in Fig. 2.
W O O D - P H E N O L RESIN FIBER
Fibers which are only composed of esterified wood components have

just been described with respect to the forming conditions and tensile
strength.
In addition, plasticized wood can be dissolved into organic solvents, so
that a new fiber-forming process utilizing fibers composed of both wood
components and an organic solvent such as phenol is available. This
method is called the 'wood-phenol resin fiber-forming process' and utilizes
plasticized wood that has been obtained by acetylation. This acetylated
wood is dissolved in phenol, and hexamethylenetetramine (which is
abbreviated to hexamine) is then added to the solution as a curing reagent
to promote initial condensation between the wood components and
phenol. This condensation imparts spinnability to the solution, from which
a filament can be spun, cured by heat and then carbonized.
In the final section of this paper, it will be shown that non-modified
wood is also applicable as the starting material instead of plasticized
wood.
Experimental Procedure
Radiata pine RMP was used as the starting material. Chemical
modification of the pulp was carried out by acetylating with trifluoro acetic
anhydride (TFAA method), and also by using perchloric acid as a catalyst
(PCA method) [3]. The dissolution (liquefaction) of the acetylated wood in
the presence of phenol was performed within a stainless steel pressure-
resisting tube at 260 C for a certain period of time without stirring.
Hexamine was added to the solution, and an initial condensation reaction
proceeded with heating until spinning became possible.
Two spinning methods were applied. With the first, a filament was
spun out or picked up from the surface of the spinning solution contained
in a crucible located in an oil bath and was wound by a machine
positioned above the crucible. The second method used an extruding type
of spinning machine equipped with a monohole spinning nozzle, from
which a filament was spun under nitrogen pressure, and wound by a
machine positioned below the nozzle. The would filament was heated to
250 C at various heating rates until the curing reaction was complete.
The cured filament was then carbonized under increasing temperature up
to 900 C in a nitrogen-filled electric furnace. The cured filament strength
was measured according to JIS L 1069, and the carbon fiber strength
according to JIS R 7601.
Acetylation Yield
The yields of acetylated wood prepared by the TFAA and PCA
methods were 145*3 and 143-4%, respectively, based on the weight of the
starting material. These figures indicate a similar level of acetylation with
each method.
Solubility in Phenol
The solubility of the acetylated wood in phenol was then investigated.
The time to completely liquefy was measured for acetylated wood-phenol
solutions of different concentration, the results being shown in Table 1.
The solubility of the acetylated wood prepared by the TFAA method was
superior to that prepared by the PCA method. This solubility difference
can be explained by the different decomposition of lignin: during acetyla-
tion by the TFAA method, selective dissociation of the a-0-4 bond in
lignin by trifluoroacetic acid occurs, which promotes the solubility of
lignin; this is not the case with acetylation by the PCA method.
Preparing the Spinning Solution

When hexamine was added to the acetylated wood-phenol solution
and the mixture reacted at 120 C for about 5 min in the typical case,
initial condensation occurred, and the viscosity of the solution increased
TABLE 1
Dissolution times by the TFAA and PCA methods
Acetylated wood content 25 35 40 50
(%)
Dissolution TFAA method 30 30 30
time (min.) 120 120 150
PCA method
to a level at which suitable spinning became possible. At this stage, the

reaction was stopped by cooling the solution. In this study, two major
factors affecting the rate of the reaction were investigated, one being the
added amount of hexamine, and the other the dissolving (liquefaction)
time.
As the progress of condensation can be detected by the increase in
viscosity of the solution, 14*3 and 12*7% of hexamine were added, based
on the weight of the solution, the influence of each quantity on the
viscosity increase being compared in Fig. 3. A difference in reaction time
of 3 h existed in order to reach the same value of viscosity at 130 C.
From this large difference in the required reaction time caused by a small
difference in the added amount of hexamine (only 2-4 wt%), it can be
deduced that the amount of hexamine is critical to stabilize the viscosity.
x10
VISCOSITY (Pa-s)
50
40
30
20
10
1 2 3 4 5 6 7
LIQUEFACTION TIME ( hr. )
FIG. 3. Relationship between the viscosity of the spinning solution and the
added amount of hexamine.
A 14-3% and 012-7% of added hexamine
The period for liquefying acetylated wood also influenced the conden-
sation rate, which is shown in Fig. 4. During the liquefaction of acetylated
wood by phenol, some of the acetic acid was eliminated, the amount of
this elimination increasing with increasing liquefaction period. This
observation and the results will be discussed in more detail in the last
section of this paper. As shown in Fig. 4, the condensation reaction was
influenced by the amount of eliminated acetic acid, a small eliminated
amount contributing to a rapid viscosity increase. In order to stabilize the
10
50
VISCOSITY (Pa-s)
40
30
20
10
o i - I 1 1 1 1 1 1
0 1 2 3 4 5 6 7
LIQUEFACTION T I M E ( n r . )
FIG. 4. Relationship between the viscosity of the spinning solution and the
amount of eliminated acetic acid. No. 1: 60 wt%; No. 2: 6-4 wt%; No. 3: 151 wt%
of eliminated acetic acid.
spinning solution, and provide consistent viscosity at a certain spinning

temperature, the presence of a specific amount of acetic acid, more than
10% based on the weight of the solution, proved to be necessary.
Spinning
The most suitable liquefaction period was assessed, based on the
spinnability of the fiber and the facility to pick up a filament from the
surface of the spinning solution.
This method for fiber forming is illustrated in Fig. 5. The spinnability is
governed by the ease of elongating a filament at point A and that of
supplying the solution at point B. The characteristics required are
FIG. 5. Pick-up spinning method.

suitable tensile strength and elongation of a filament at point A, and

suitable viscosity for supplying the solution at point B.
After optimizing all the factors just described, the winding speed could
be selected as the determining factor for spinnability, and it was found to
be necessary to take more than 3 h for liquefaction [ 4 ] .
Curing
One of the advantages of wood-phenol resin fiber is that a heating
process is enough to completely cure the filaments.
In a previous report, it was shown that the existence of acetylated wood
in the two-component system of phenol and hexamine, which allows even
chemical bonding to develop, made it possible to cure a filament while
maintaining its shape. On the other hand, a filament spun from a solution
composed only of phenol and hexamine could not maintain its shape
when exposed to an elevated temperature. It was shown that a filament
containing a larger quantity of acetylated wood exhibited greater ability
to maintain its shape during the curing process. Both atmospheres of air
and nitrogen had the same effect on the cured strength of a filament, and
a synergistic effect between acetylated wood and phenol on the filament
strength was also found [ 4 ] .
The reaction mechanism between phenol and hexamine has been
reported [ 5 ] , benzylamine derivatives being formed through amino-
methylation of hexamine, before dimethyleneamine bonds and methylene
bonds are formed with the elimination of ammonia.
In order to clarify that a chemical reaction occurs between wood
substances and hexamine, the interaction between acetylated wood and
hexamine was thermally analyzed. Figure 6 shows the results of a
differential scanning calorimetry (DSC) analysis of the related compounds.
At temperatures above 195 C, a reaction between hexamine and
acetylated wood occurred and resulted in an exothermal peak at 225 C.
On the other hand, acetylated wood alone did not show any peak, and
hexamine by itself showed the exothermal peaks beyond 250 C. The
exothermal peak which started at 195 C, therefore, seems to have been
caused by the aminomethylation between hexamine and wood com-
ponents such as lignin, carbohydrates and their acetyl derivatives.
The reaction mechanism between hexamine, phenol and the solutes
was investigated by identifying and quantifying the gases generated from
a spinning solution at various spinning temperatures.
Gas chromatograms of the generated gases under conditions of dif-
ferent heating temperature and heating period are shown in Fig. 7.
Acetylated wood
Hexamine .
EXO-
Hexar^f
A c e t y l * ^
FIG. 6. Differential scanning calori-
D 100 200 300 metric thermograms of acetylated
Temperature (C ) wood and hexamine.
H20
NH3
A
V j v
\ 200C,1hr
^^NH3 _J^^ 150C.2hr"
130C,2hr
I 1 11 1
1 1u
0 5 10 15 20 FIG. 7. Generated gases during the

RETENTION TIME (min) condensation reaction.
WOOD PHENOL RESIN FIBER 179
Ammonia and water were identified as two generated gases by means of

mass spectrometry. The spinning temperature of 130C did not produce
any gas, except for a trace water peak, while at a curing temperature of
200 C, the generation of both ammonia and water was quite obvious.
The ammonia is considered to have been generated by the condensation
of aminomethyl derivatives, and the water would have been mainly
formed from condensation between the aminomethyl derivatives and the
hydrogen oxide groups of carbohydrates.
From these results, it can be concluded that each wood component
performs as a constituent of 'wood-phenol resin fiber' through chemical
bonding between the wood components and the phenol.
Carbonization and Mechanical Properties of the Carbon Fiber

The cured filaments obtained from the foregoing processes were
carbonized at temperatures up to 900 C in nitrogen.
The tensile strength of the filament formed by pick-up spinning is
shown in Fig. 8, both before and after carbonization. Both samples
prepared by the PCA and TFAA methods were found to have a
carbonizing effect, exhibiting a maximum tensile strength of about
0-5 GPa.
0.6
A
0.5 [
A
TENSILE STRENGTH (Gpa)
0.4
FIG. 8. Tensile strength of fibers

from pick-up spinning before and
0.3
after carbonization (open symbols:
after carbonization, solid symbols:
before carbonization).
0.2 k o
\ O , 50% of acetylated wood con-
tent by the PCA method
A , A 60% of acetylated wood con-
0.1 tent by the TFAA method
0 50 100 , 35% of acetylated wood con-
FIBER W I D T H (urn) tent by the TFAA method
Because the tensile strength of the carbon fiber increased with decreasing
diameter, extrusion spinning was used to obtain finer filaments. Figure 9
shows the strength of carbon fibers from extrusion spinning, by which a
tensile strength of more than IT GPa could be obtained. This strength is
higher than that of the pitch carbon fiber of general-purpose grade.
1.3i . . 1 1 1
1.2-
1.1 -
(GPa)
1.0-
TENSILE STRENGTH
0.9-
0.8-
0.71 . . . i i 1 1
3 4 5 6 7 8 9 10
FIBER WIDTH (|jm)
FIG. 9. Tensile strength of carbon fibers from extrusion spinning of a solution

with 40% of acetylated wood content by the PCA method.
W O O D - P H E N O L RESIN FIBER F R O M U N M O D I F I E D W O O D
Liquefaction of Unmodified Wood into Phenol

Attempts were made to clarify the phenomena occurring during the
liquefaction of acetylated wood in phenol by using cellulose acetate,
which is simpler to analyze than acetylated wood.
The elimination of acetyl groups in the form of acetic acid was confirmed
by its identification and quantification with gas chromatography and
mass spectrometry. The same analytical methods were applied to

acetylated woods, and the amount of eliminated acetic acid with liquefac-
tion time and temperature was measured. Figure 10 shows clearly that
acetylated wood also generated acetic acid as the liquefaction time
progressed. From this observation, it is speculated that the generated acetic
acid might have the ability to liquefy the wood material. On this basis, a new
dissolution system comprising unmodified wood, phenol and acetic acid
was studied.
Figure 11 shows the weight fractions of undissolved wood after 5 h of
liquefaction at 260 C. When the unliquefied wood percentages are
20
40% acetylated wood
G E N E R A T E D CHjCOOH (9
P 260*C .
10
0
0 5 10
LIQUEFACTION TIME ( nr. )
FIG. 10. Regenerated acetic acid fraction and dissolution time.
compared at a constant wood content of 2 5 w t % in the solution, the

liquefaction effect of acetic acid is apparent, while the decreasing liquefac-
tion effect of an increasing amount of phenol can also be seen.
Spinning an Unmodified Wood Solution

Spinning, curing and carbonization of an unmodified wood solution
could be carried out in the same way as that described for acetylated
wood. The performance and properties of carbon fibers from unmodified
wood were the same as those from acetylated wood. When the relation-
ship between the viscosity of the spinning solution and the extrusion
pressure were studied for spinning the same fiber diameter (9 fim) under
the same winding rate (550m/min), the relationship was found to obey
Phenol CH3COOH
WOOD
FIG. 11. Undissolved wood fraction for unmodified wood.
Hagen-Poiseulle's law. The tensile strength of the carbon fiber thus

obtained was found to be about IT GPa, which is the same as that of
carbon fiber from acetylated wood.
CONCLUSION
Fiber forming methods have been proposed as a wood-utilization

technology with high added value. Three fiber-forming methods were
demonstrated. The first involved spinning wood plasticized both by
chemical modification and additional dissolution in an organic solvent;
the second was concerned with phenol resinified and chemically modified
wood; the third involved phenol resinified untreated wood. The first
method produced filaments with half the tensile strength of cellulose
acetate fibers. The other two methods gave carbon fibers with a tensile
strength of more than 1 0 GPa.
ACKNOWLEDGEMENT
The author is indebted to Professor Nobuo Shiraishi of Kyoto Univer-

sity for his valuable discussions and help in the preparation of this paper,
and also to Mr Masaru Yamakoshi and Miss Yukiko Horiuchi for their
experimental assistance.
REFERENCES
[1] N. Shiraishi, Mokuzai Gakkaishi (J. Japan Wood Res. Soc), 32, 755 (1986).
[2] N. Shiraishi, H. Tamura and N. Tsujimoto, Mokuzai Kogyo (J. Wood
Technol. Assoc., Japan), 42, 492 (1987); 43, 2 (1988).
[3] N. Shiraishi, K. Tsubouchi, T. Matsunaga and T. Yokota, Proceedings of
30th Annual Meeting of the Japan Wood Research Society, 34 (1980).
[4] N. Tsujimoto, Proceedings of the Nth Symvosium on Chemical Processing of
Wood, Jpn. Wood Res. Soc., 17 (1984).
[5] E.W. Orrell and R. Burns, Plastics and Polymers, 469 (1968).
Carbohydrate Conversion by Irradiation
with a Carbon Dioxide Laser
JUN-ICHI AZUMA, MING ZHANG and KEIZO OKAMURA
Kyoto University, Kitashirakawa Oiwake-cho, Sakyo-ku,
Kyoto 606-01, Japan
ABSTRACT
Four polysaccharides, cellulose, glucomannan, /M,3-glucan and starch, were irradiated
with a C O laser beam under a stream of nitrogen, and the chemical properties of the
2
degraded products were characterized. In all cases, melted materials were formed by
blasting with vaporized materials. The amounts of melted materials and vaporized materials
varied greatly depending upon the energy density and duration of irradiation. Scanning
electron microscopic observations indicated that the pyrolytic action of C 0 laser irradi-
2
ation was strictly limited to the surface of the melted layer. The melted materials were
mainly composed of a series of anhydro-oligosaccharides corresponding to the original
polysaccharides, together with monomeric 1,6-anhydrosugars. The vaporized materials
from cellulose were composed of volatile gases and an oily material rich in levoglucosan,
and also contained anhydro-cello-oligosaccharides identical to those observed in the melted
materials. The results of this study demonstrate that C 0 laser irradiation can be used as
2
a practical carbohydrate conversion method, and opens a new field of carbohydrate

technology in which anhydrosugars can be utilized as the raw material.
Keywords: C 0 laser, carbohydrate technology, anhydro-oligosaccharide, polysaccharide

2
conversion, C 0 laser irradiation.

2
INTRODUCTION
Laser is an abbreviation of the words 'light amplification by stimulated

emission of radiation', and produces a new-born high-power light that is
different from ordinary light. Its outstanding characteristics are the
intensity, directionality, monochromaticity, and coherence of its waves.
Among the laser beams, only a C 0 laser can emit high-power light in
2
185
186 JUN-ICHI AZUMA ET AL.
the medium infrared region at 106 /mi, which is suitable for cutting and
drilling a wide variety of non-metallic materials including textiles, leather,
confectionery, plastics, rubber, wood and paper [ 1 - 5 ] . However, prob-
lems concerning the chemical properties of the melted and vaporized
products, and characterization of the reactions caused by laser irradiation
remain to be solved. In order to elucidate these aspects, we have
previously cut filter papers by irradiating with a C 0 laser and partially
2
characterized the structure of the melted degradation products [6].

In this study, we investigate the chemical properties of the degradation
materials produced by C 0 laser irradiation of cellulose, and extend this
2
carbohydrate conversion method to three other polysaccharides,

glucomannan, /?-l,3-glucan and starch.
EXPERIMENTAL P R O C E D U R E
Laser irradiation was carried out by using Japan Chemical Engineering

& Machinery Model NRT-1 laser-heating apparatus equipped with a
Shin Nihon Kagaku Seisakusho Model SL-116C multiple-mode continu-
ous C 0 laser and a zinc-selenium lens with a focal length of 12-7 cm
2
The output power of the laser beam emitted at a wavelength of 10-6 jjm
was determined by a Nihon Kagaku Engineering Model PM-315 power
meter, and the spot size of the laser beam was determined by a Toshiba
beam profile meter. A schematic illustration of the laser irradiation
equipment is shown in Fig. 1. The sample holder for irradiation was made
up with two methylmethacrylate ring plates, 1-5 cm in height and 7-7 cm
and 10-2 cm in respective internal diameter, the bottoms of which were
bonded to the surface of a ceramic plate (95% aluminum oxide) with an
adhesive to make an annulus between the two methylmethacrylate plates.
The diameter of the annulus at the center was 8-9 cm, corresponding to
2 8 0 c m in length per revolution. Polysaccharide powder was inserted
into the annulus and compressed by hand with a methylmethacrylate ring
that fitted in the annulus. The sample holder was then laid on a rotating
plate, and the C 0 laser irradiated for one revolution at various combi-
2
nations of linear speed of 0-92-4-4 m/min, output power of 35-51 W, and

defocusing distance of 0-20 cm, using nitrogen as an assist gas (about
11/min). The energy density and irradiation time were calculated from
these values, and the amount of vaporized material was determined by
measuring the weight of the sample before and after irradiating.
The irradiated portion adhering with melted material was picked up by
CARBOHYDRATE CONVERSION BY CARBON DIOXIDE LASER IRRADIATION 187
Mirror
Oscillator
Lens
- N i t r o g e n gas
Focus
Defocus
Sample holder
Ceramic p l a t e
Melted material
Rotating plate
M e t h y l m e t h a c r y l a t e ring p l a t e
FIG. 1. Schematic diagram of the C 0 laser irradiation equipment.

2
a pair of tweezers and extracted thoroughly with distilled water or 80%

aqueous ethanol. A new sample was then placed in the annulus, and
irradiation was started again. The molecular weight distribution of the
melted materials was analyzed by size-exclusion chromatography (SEC)
on Asahipak GS-220 [ 7 ] . Fractionation of the melted materials was
carried out by repeated preparative SEC on Toyopearl HW-40S [7, 8].
The fractionated materials were purified by preparative high-performance
liquid chromatography (HPLC) on Asahipak NH2P-90 or by preparative
silica gel thin-layer chromatography (TLC). The structure of each isolated
material was characterized by a combination of chemical composition
and methylation analyses and spectroscopic analysis [ 9 ] . The morphol-
ogy of the irradiated surface was observed with a Jeol JSM-T330A
scanning electron microscope.
Degradation of Cellulose by C 0 Laser Irradiation

2
When Whatman CF-11 cellulose powder was irradiated by the C 0 2
laser beam, light yellow to brown-colored melted materials were formed

by blasting from the white cloud of vaporized materials. Figures 2 and 3

show the energy density dependence of the amounts of water-soluble
materials and vaporized materials. The results indicate that a high energy
density with low irradiation speed (longer duration of irradiation time)
produced a deep, narrow V-shaped valley in the cellulose layer, giving
rise to the production of a large amount of vaporized materials, whereas
a low energy density with high irradiation speed (shorter duration of
irradiation time) produced only a small amount of vaporized materials.
With an irradiation time of 150 msec, the amount of vaporized materials
was a clear maximum at about 200 J/cm . The amount of water-soluble 2
materials formed by C 0 laser irradiation shows a sigmoidal energy 2
density dependence in Fig. 3. The values of energy density which gave the
maximal amount of water-soluble materials below 100 J/cm diminished 2
from 65 J/cm to 16 J/cm with reduction of the duration of irradiation

2 2
time. A similar phenomenon was also observed in the minimum points

between 60 J/cm and 200 J/cm . In the energy density region higher than
2 2
1 .<4 i i i l i 1 i i 1 I 11 ! | 1 1 1 1 111 1
cz 1.2
\
cn
Speed
in 1.0 - /a
(m/min) -
DIJ9.
/ 0.92
D
E 0.8 -
A m o u n t o"f v a p o rized
0.6
* ' * 1.62
0.4 a 2.49
3.51
JS O o
0.2 ô~-^ 4.40
n i i iJ I i i i i i i i i I i i l l i iiil i i i ii i I I I
10 10 2
10 3
10 4
Energy density (J/cm ) 2
FIG. 2. Energy density dependence of vaporized materials produced by C 0 2
laser irradiation of cellulose.

350 :
Amount of water-soluble materials (mg/m)
Speed
300 - (m/min)
0.92
250 -
200 -
150 -
100 -
50 -
10 10 2
10 3
10 4
Energy density (J/cm )

2
FIG. 3. Energy density dependence of water-soluble materials produced by C 0 2
laser irradiation of cellulose.
the minimum points, the amount of water-soluble materials increased

with increasing energy density.
Figure 4 shows again the sigmoidal energy density dependence of the
proportion of water-soluble materials composed with the sum of the
water-soluble materials and the vaporized materials. The profiles indicate
that the values of energy density which gave the maximal proportion of
water-soluble materials diminished with reducing irradiation time. The
maximal proportion of water-soluble materials increased with decreasing
irradiation time, and attained 39-1% (w/w) after irradiating for 87 msec at
13-7 J/cm . 2
The water-soluble materials were composed of various components

having average molecular weights (Mw) lower than 3000. For clear
separation of the lower molecular weight components, the water-soluble
materials were extracted with 80% aqueous ethanol to remove the higher
molecular weight components. By combining preparative SEC, HPLC
and TLC, two reducing sugars and nine anhydrosugars could be isolated.
The major components were 1,6-anhydro-jS-D-glucopyranose (levo-
glucosan) and a homologous series of seven anhydro-cello-oligosac-
40 -
Proportion of water-soluble materials {%)
35 -
30 -
25 -
20 -
1 5-
10 -
5-
10 10 2
10
5
104
Energy density (J/cm )

2
FIG. 4. Energy density dependence of the proportion of water-soluble materials

produced by C 0 laser irradiation of cellulose.
2
charides up to cello-octaosan. The chemical structures of these anhydro-

sugars were characterized by H - and C-nuclear magnetic resonance 1 13
(NMR) spectroscopic and methylation analyses, and the results are

shown in Fig. 5 and in Tables 1, 2 and 3. The distribution of these
anhydrosugars was estimated to be 12-5% (levoglucosan), 15-0% (cel-
lobiosan), 15-9% (cellotriosan), 12-8% (cellotetraosan), 12-5% (cellopen-
taosan), 8-1% (cellohexaosan), 6-7% (celloheptaosan) and 5-3% (cello-
octaosan) of the dry weight of the 80% ethanol-soluble materials pro-
duced by irradiating at 2-9 x 10 J/cm . l,6-Anhydro-/?-D-glucofuranose 3 2
was also present, but its content was as low as 0 1 % of the original
water-soluble materials. Regarding the preparation of the anhydro-
cello-oligosaccharides, only cellobiosan [10] and cellotriosan [11] have
been synthesized, but the anhydro-cello-oligosaccharides having D . P . ^ 4
are newly isolated components which could not be easily separated by
conventional synthetic methods. The reducing sugars were composed of
glucose, cellobiose and cellotriose. All their contents increased with
TABLE 1
!
H-NMR data for the representative anhydrosugars produced by C 0 2 laser
irradiation of cellulose (S in p.p.m.)
Anhydrosugars DP* Anomeric protons
Levoglucosan 1 5-44(s)
461(d) 5-45(s)
Cellobiosan 2 (7-8)
[10] [10]
452(d) 4-62(d) 5-45(s)
Cellotriosan 3 (7-8) (7-8)
[1-2] [1-2] [10]
452(d) 454(d) 463(d) 5-46(s)
Cellotetraosan 4 (7-8) (7-8) (7-8)
[1-2] [1-3] [10] [10]
4-52(d) 4-55(d) 463(d) 5-45(s)
Cello-octaosan 8 (7-7) (7-8) (7-8)
[1-4] [5-3] [1-3] [10]
Chemical shifts for anomeric protons in D 0 are given with sodium 2,2,3,3-
2
tetradeuterio-3-(trimethylsilyl) propionate (TSP) as an internal standard.

H l : non-reducing end residue
n
H lj! intermediate residue(s)

H l : intermediate residue adjacent to an anhydroglucose residue
ia
H l : anhydroglucose residue
a
s, d: multiplicity of signals
( ) coupling constants (J in Hz)
l 2
[ ] molar ratio
* degree of polymerization
increasing moisture content of the original cellulose. The presence of

water contributed to accelerating the hydrolysis of glucosidic bonds. The
contents of glucose and cellobiose would be reduced to 0-8% and 0% by
using anhydrous cellulose powder.
Figure 6 shows a typical scanning electron micrograph of the irra-
diated surface. Part of the irradiated fibers melted and attached to each
other to form three layers, the irradiated surface (a), melted layer (b) and
fibrous layer (c). The thickness of the melted layer ranged from 150 fim to
900 fim depending on the irradiation conditions. The melted layer disap-
peared after washing with water, and a porous structure with a thickness
of 1 fim to 5 um was left as an insoluble residue. This porous structure
TABLE 2
13 C-NMR data for the representative anhydrosugars produced by C 0 2 laser irradiation of cellulose (8 in 192
p.p.m.)
Levoglucosan Cellobiosan Cellotriosan Cellotetraosan Cello-octaosan
C-l n 102-48 103-48 103-47 103-38

C-2 n 73-92 74-12 74-10 7401
Non-reducing C-3 n 76-61 76-57 76-54 76-38
end residue C-4 n 70-66 70-27 70-48 70-34
C-5 n 76-90 76-89 76-88 76-83
C-6 n 61-79 61-63 61-62 61-48
C-li 103-35 103-28(s)
C-2i 73-88 73-80(s)
Intermediate C-3i 7506 74-91(s)
residue(s) C-4i 79-53 79-23(s)
C-5j 75-72 75-66(s)
C-6i 6102 60-88(s)
Intermediate c-i i a 10219 10216 10214(w)
JUN-ICHI AZUMA ET AL.
residue C-2 i a 73-71 73-71 73-61

adjacent to C-3 i a 7506 75-06 74-91
anhydroglucose C-4 i a 79-49 79-67 79-49
residue C-5 i a 75-75 75-72 75-66
C-6 i a 6119 6113 60-81
C-l a 102-28 102-33 102-38 10219 10217
C-2 a 71-02 71-34 71-35 71-33 70-84
Anhydroglucose C-3 a 73-35 72-53 72-50 72-49 72-25
residue C-4 a 71-65 7919 79-25 79-24 78-72
C-5 a 77-10 75-05 7506 7506 74-91
C-6 a 6603 6608 6608 6606 65-92
2
Chemical shifts in D O are given with dioxane (67-40 p.p.m.) as an internal standard, s: strong, w: weak.
TABLE 3
Methylation analysis of the representative anhydrosugars produced by C 0 laser2
irradiation of cellulose
Anhydrosugar DP* Methylated sugar Molar ratio
Levoglucosan 1 2,3,4-Glc**
2,3,4,6-Glc 10
Cellobiosan 2 2,3,6-Glc
2,3-Glc 10
2,3,4,6-Glc 10
Cellotriosan 3 2,3,6-Glc 1-2
2,3-Glc 1-2
2,3,4,6-Glc 10
Cellotetraosan 4 2,3,6-Glc 1-8
2,3-Glc 10
2,3,4,6-Glc 10
Cello-octaosan 8 2,3,6-Glc 5-7
2,3-Glc 10
** 2,3,4-Glc = 2,3,4-tri-0-methyl-D-glucopyranose, etc.
may be the transition area from the melted layer to the native cellulose.
The crystallinity indices and molecular weights of the irradiated celluloses
decreased with increasing amounts of water-soluble materials, but re-
covered to the initial level after washing with water. The present results
show that the structural change produced by heating with the C 0 laser 2
beam was limited on the surface without producing any carbonized

materials when the heating time was too short to vaporize the cellulose.
Degradation of Glucomannan by C 0 Laser Irradiation

2
Wood glucomannan was isolated from a mercerated pulp liquor

(Unitika Rayon Co. Ltd.) in a yield of 0-56% (w/v) according to the
method of Meier [12]. This glucomannan, whose molar ratio of glucose,
galactose and mannose was 10:0-2:2-8, was irradiated by the C 0 laser 2
under the same conditions as those applied to the cellulose, and the
melted materials were extracted with 80% aqueous ethanol instead of
water, because glucomannan itself is partially soluble in water. The
energy density dependence of the amounts of vaporized materials and
80% aqueous ethanol-soluble materials were similar to those for the case
of cellulose. Because glucomannan has a linear main chain composed of
Levoglucosan
CH OH
2 CH OH
2
OH OH OH
n = 0 - 6
Anhydro-celtooligosaccharides
CH OH
2 CH OH
2
OH OH OH
n = 0 - 8
Anhydro-lamlnarioligosaccharides
FIG. 5. Structures of the anhydro-gluco-oligosaccharides and levoglucosan
formed by C 0 laser irradiation of cellulose, /M,3-glucan and starch.
2
FIG. 6. Morphological surface structure of cellulose irradiated by the C 0 laser.

2
1,4-linked j5-D-glucopyranose and /?-D-mannopyranose residues, various

kinds of anhydro-hetero-oligosaccharides together with levoglucosan and
l,6-anhydro-j8-D-mannopyranose (AM) were expected to be present in the
melted materials. The chemical structures of the four isolated anhydro-
sugars characterized by N M R spectroscopic and methylation analyses
(Tables 4 and 5) are shown in Fig. 7. The yields of these anhydrosugars
were 1-2% (AM), 4 0 % (4-j3-glucosyl-l,6-anhydromannose, G-AM), 5-7%
(4-/?-glucosyl-4-/?-mannosyl-l,6-anhydromannose, (G-M-AM) and 5-5%
(4-jS-cellobiosyl-4-j?-mannosyl-l,6-anhydromannose, G-G-M-AM) of the
80% ethanol-soluble materials obtained by irradiating at 2-9 x 10 J/cm . 3 2
These anhydrosugars are characteristic of the presence of 1,6-anhydro-

mannopyranose as the reducing ends. Anhydrosugars containing 1,6-
anhydroglucopyranose were also present, and their structural character-
ization is now in progress.
Degradation of the Other Water-Insoluble Polysaccharides by C 0 2
Laser Irradiation
Based on the results just presented, the water-insoluble homologous
polysaccharides were concluded to be easier to handle than the water-
soluble ones with respect to an exact estimation of the amount of
melted materials and the characterization of their structures. We there-
fore applied the present laser-induced conversion method to the two
TABLE 4
X
H-NMR data for the representative anhydrosugars produced by C 0 2 laser
irradiation of glucomannan (S in p.p.m.)
H-ln H-U tf-lia H-K
1,6-Anhydromannose (AM) 1 5-44(s)

4--Glucosyl-l,6- 4-53(d) 5-46(s)
anhydromannose (G-AM) 2 (7-7)
[10] [10]
4-0-Glucosyl-4-j3- 4-54(d) 491(d) 5-46(s)
mannosyl-1,6-anhydro- 3 (7-4)
mannose (G-M-AM) [0-8] [10] [10]
4--Cellobiosyl-4-- 4-53(d) 455(d) 4-88(s) 5-44(s)
mannosyl-1,6-anhydro- 4 (79) (7-7)
mannose (G-G-M-AM) [0-9] [0-9] [09] [10]
2
tetradeuterio-3-(trimethylsilyl)propionate (TSP) as an internal standard.

n
H lj: intermediate residue(s)

H l : intermediate residue adjacent to an anhydromannose residue
ia
H l : anhydromannose residue
a
( ) coupling constants ( J in Hz)
1>2
[ ] molar ratio
representative polysaccharides, /M,3-glucan (curdlan, Wako Pure

Chemical Ind. Ltd.) and starch (corn starch, Sanwa Denpun Kogyo Co.
Ltd.).
In the case of starch, both the energy density dependence of the
water-soluble and vaporized materials was quite similar to that obtained
with cellulose. The water-soluble materials contained at least 10 anhydro-
oligosaccharides, six of which could be isolated from the 80% ethanol-
soluble materials, and their structures are characterized in Fig. 5. The
representative results of N M R spectroscopic and methylation analyses
are shown in Tables 6, 7 and 8. The distribution of these anhydrosugars
and levoglucosan was estimated to be 18*8% (levoglucosan), 2 3 1 %
(maltosan), 1 6 1 % (maltotriosan), 11-4% (maltotetraosan), 9-5% (malto-
pentaosan), 8 0 % (maltohexaosan) and 3 0 % (maltoheptaosan) of the
80% ethanol-soluble materials produced by irradiating at 2-9 x 10 J/cm . 3 2
TABLE 5
Methylation analysis of the representative anhydrosugars produced by C 0 laser
2
irradiation of glucomannan
1,6-Anhydromannose (AM) 1 2,3,4-Man**
4-0-Glucosyl-l,6- 2,3,4,6-Glc 10
anhydromannose (G-AM) 2 2,3-Man 10
4-0-Glucosyl-4-j8- 2,3,4,6-Glc 11
mannosyl-1,6-anhydro- 3 2,3,6-Man 0-9
mannose (G-M-AM) 2,3-Man 1-3
4-0-Cellobiosyl- 2,3,4,6-Glc 10
mannosyl-1,6-anhydro- 4 2,3,6-Glc 11
mannose (G-G-M-AM) 2,3,6-Man 1-3
2,3-Man 10
** 2,3,4-Man = 2,3,4-tri-O-methyl-D-mannopyranose, etc.
The isolated oligosaccharides were exclusively a-l,4-linked malto-

oligosaccharides having their reducing ends blocked by 1,6-anhydro
rings, and did not contain 1,6-linked branch points. The effects of C 0 2
laser irradiation of the a-l,6-glucosidic linkages are now under investiga-

tion, using an a-1,6-linked glucan, dextran.
In the case of curdlan, the energy dependence of the vaporized
materials was similar to that of cellulose, but the energy dependence of
the water-soluble materials was somewhat different from that of cellulose
as shown in Fig. 8. Under conditions of low energy density ( < 2 0 J/cm ) 2
and high irradiation speed (5*3-51 m/min), a clear maximum was detec-
ted. However, under the other conditions, the amount of the water-
soluble materials monotonously decreased with increasing energy den-
sity. Figure 5 shows the structures of the homologous series of 9
anhydro-laminarioligosaccharides isolated from the water-soluble ma-
terials. The representative results of N M R spectroscopic and methyla-
tion analyses are shown in Tables 9, 10 and 11. The distribution of
these anhydrosugars and levoglucosan was 4-3% (levoglucosan),
6-6% (laminaribiosan), 9-4% (laminaritriosan), 11-4% (laminaritetraosan),
12-9% (laminaripentaosan), 11-7% (laminarihexaosan), 6*5% (lamin-
ariheptaosan), 2-7% (laminarioctaosan), 2-2% (laminarinonaosan) and
2-2% (laminaridecaosan) of the 80% ethanol-soluble materials produced
by irradiating at 2-9 x 1 0 J/cm .
3 2
l 6-Anhydro-0-
f (G-AM)
D-mannopyranose
Ori OH
(G-G-M-AM)
Anhydro-glucomannooligosaccharides
FIG. 7. Structures of the anhydro-hetero-oligosaccharides and 1,6-anhydro-/?-

D-mannopyranose produced by C 0 laser irradiation of glucomannan.
2
Chemical Properties of the Vaporized Materials

The vaporized materials formed from cellulose were collected in an
impinger type of glass trap, which was cooled with liquid nitrogen by
sucking with an electric cleaner. The vaporized materials were separated
into three parts, gases, transparent condensable liquid, and brown
colored oily material. The condensable liquid was immediately vaporized
by warming, whereas the brown colored material still remained after
removing the vaporized materials. The condensable liquid materials may
TABLE 6
^ - N M R data for the representative anhydro-oligosaccharides produced by C 0 2
laser irradiation of corn starch (S in p.p.m.)
H-\ n H-l, H-l,

5-16(d) 548(d)
Maltosan 2 (39)
[09] [10]
516(d) 5-36(d) 5-48(s)
Maltotriosan 3 (39) (3-7)
[0-9] [0-9] [10]
5-16(d) 5-38(d) 5-48(s)
Maltotetraosan 4 (39) (3-6)
[09] [1-8] [10]
5-16(d) 536(d) 5-48(s)
Maltoheptaosan 7 (39) (3-6)
[0-9] [4-8] [10]
2
tetradeuterio-3-(trimethylsilyl) propionate (TSP) as an internal standard.

n

a
( ) coupling constants (J l 2H) m z
[ ] molar ratio
have been composed of volatile hydrocarbon gases, as was indicated by

Madison and Keehn [13]. The colored materials obtained by irradiating
at 72 J/cm for 0-92 m/min were composed of at least six sugars: glucose
2
(1-3%), levoglucosan (40-5%), cellobiose (4-4%), cellobiosan (32-3%),

cellotriosan (13*7%) and cellotetraosan (7*8%). The marked accumula-
tion of levoglucosan is quite understandable when considering its high
distillability, while the presence of nonvolatile sugars in the vaporized
materials is indicative of the violent blasting induced by irradiating with
the CO laser. The present results indicate that the vaporized materials
2
could also be useful as starting material for the preparation of anhydro-

sugars.
TABLE 7
13
2
C-NMR data for the representative anhydro-oligosaccharides produced by C 0 laser irradi- 200
ation of corn starch (5 in p.p.m.)
Maltosan Maltotriosan Maltotetraosan Maltoheptaosan
C-l n
98-82 98-71 98-75 98-70
C-2 n 72-34 72-72 72-73 72-71
Non-reducing C-3 n 73-87 73-89 73-89 73-88
end residue C-4 n 71-08 70-44 70-43 70-42
C-5 n 73-33 73-77 73-65 73-62
C-6 n 61-67 61-59 61-59 61-57
C-lj 100-50 100-54(s)
C-2; 72-52 72-49(s)
Intermediate C-3i 74-15 74-1 l(s)
residue(s) C-4j 78-28 78-23(s)
C-5i 72-25 72-23(s)
C-6i 61-59 61-57(s)
Intermediate c-i i a
100-64 100-75 100-72
JUN-ICHI AZUMA ET AL.
residue C-2 i a
72-14 72-14 72-14
adjacent to C-3 i a
74-14 74-15 74-11
residue C-5 i a 71-91 71-89 71-90
C-6 i a
61-59 61-59 61-57
C-l. 102-18 10218 102-20 10218
C-2 a 70-96 7101 71-02 7103
Anhydroglucose C-3 a 70-69 70-66 70-64 70-59
residue C-4 a 77-34 77-53 77-57 77-54
C-5 a 76-21 76-15 7616 76-13
C-6 a 6613 6610 6610 6609
2
Chemical shifts in D 0 are given with dioxane (67-40 p.p.m.) as an internal standard, s: strong
TABLE 8
Methylation analysis of the representative anhydro-oligosaccharides produced by
CO laser irradiation of corn starch2
2,3,4,6-Glc** 10
Maltosan 2 2,3,6-Glc
2,3-Glc 11
2,3,4,6-Glc 10
Maltotriosan 3 2,3,6-Glc 0-9
2,3-Glc 11
2,3,4,6-Glc 10
Maltotetraosan 4 2,3,6-Glc 1-9
2,3-Glc 1-2
2,3,4,6-Glc 10
Maltoheptaosan 7 2,3,6-Glc 4-7
2,3-Glc 0-9
*2,3,4,6-Glc = 2,3,4,6-tetra-O-methyl-D-glucopyranose, etc.
(mg/m)
300
250
Amount of water-soluble materials
200
150
100
50
n
10 10 2
10 3
10 4
Energy density (J/cm ) 2
FIG. 8. Energy density dependence of water-soluble materials produced by C 0 2
laser irradiation of /M,3-glucan (curdlan).

TABLE 9
^ - N M R data for the representative anhydro-oligosaccharides produced by C 0 2
laser irradiation of curdlan (S in p.p.m.)

Anhydrosugar DP* Anomeric protons
H-K H-l a
456(d) 5-48(s)
Laminaribiose 2 (7-8)
[0-8] [10]
460(d) 4-74(d) 5-48(s)
Laminaritriose 3 (7-9) (7-8)
[0-8] [0-9] [10]
459(d) 479(d) 474(d) 5-48(s)
Laminaritetraose 4 (7-9) (7-9) (7-8)
[09] [11] [0-8] [10]
456(d) 4-79(d) 474(d) 5-48(s)
Laminaridecaose 10 (7-8) (7-8) (7-9)
[1-2] [6-7] [08] [10]
2
tetradeuterio-3-(trimethylsilyl)propionate (TSP) as an internal standard.

n

H l : intermediate residue adjacent to an anhydroglucose residue
ia
a
( ) coupling constants ( J in Hz)
1>2
[ ] molar ratio
*degree of polymerization
CONCLUSION
The effects of C 0 laser irradiation of polysaccharides were inves-

2
tigated to characterize the melted water-soluble materials. The results

show the possibility for chemical conversion of polysaccharides to a series
of anhydro-oligosaccharides which cannot be prepared easily by the
conventional synthetic methods. To prepare an homologous series of
anhydro-oligosaccharides, it is necessary to use linear homopolysacchar-
ides under anhydrous conditions. Separation of the anhydro-oligosac-
charides was achieved easily by a combination of preparative SEC and
HPLC. The anhydro-oligosaccharides may be useful as substrates for
TABLE 10
13
2
C-NMR data for the representative anhydro-oligosaccharides produced by C 0 laser irradiation of
curdlan (3 in p.p.m.)
Laminaribiosan Laminaritriosan Laminaritetraose Laminaridecaose
C-l 103-20 102-70 102-60 102-67

C-2 n 72-90 73-28 73-27 73-28
Non-reducing C-3 n 75-27 75-40 75-40 75-44
end residue C-4 n 69-51 69-43 69-40 69-40
C-5 n 75-92 75-84 75-82 75-82
C-6 n 60-53 60-49 60-51 60-51
C-li 102-29 102-35(s)
C-2i 73-03 73-13(s)
Intermediate C-3i 83-40 83-50(s)
residue(s) C-4, 67-96 67-93(s)
C-5i 75-49 75-50(s)
C-6j 60-51 60-5 l(s)
Intermediate c-i i a
102-81 102-69 102-64
residue C-2 i a
72-77 72-79 72-79
adjacent to C-3 i a
83-59 8404 8406
residue C-5 i a 75-52 75-49 75-44
C-6 i a 60-49 60-51 60-51
c-i a
100-93 100-91 100-83 100-83
C-2 a 67-97 67-70 67-96 67-93
CARBOHYDRATE CONVERSION BY CARBON DIOXIDE LASER IRRADIATION
Anhydroglucose C-3 a 79-93 79-98 79-94 79-95

residue C-4 a 6900 68-99 6900 6900
C-5 a 75-73 75-71 75-67 75-67
C-6 a 64-66 64-63 64-64 74-65
203
2
Chemical shifts in D O are given with dioxane (67-40 p.p.m.) as an internal standard, s: strong
TABLE 11
Methylation analysis of the representative anhydro-oligosaccharides produced by
C 0 laser irradiation of curdlan
2
2,3,4,6-Glc** 10
Laminaribiosan 2 2,4,6-Glc
2,4-Glc 1-2
2,3,4,6-Glc 10
Laminaritriosan 3 2,4,6-Glc 10
2,4-Glc 0-9
2,3,4,6-Glc 10
Laminaritetraosan 4 2,4,6-Glc 1-7
2,4-Glc 1-2
2,3,4,6-Glc 10
Laminaridecaosan 10 2,4,6-Glc 7-8
2,4-Glc 1-3
*2,3,4,6-Glc = 2,3,4,6-tetra-O-methyl-D-glucopyranose, etc.
glycosidases, because of the lack of reducing ability in situ, as well as

chemical tools for investigating the configurational requirements of the
reducing end for an enzyme reaction. The anhydro-oligosaccharides are
the useful starting materials for the synthesis of regioselectively branched
or modified polysaccharides via ring-opening polymerization [14, 15].
These polysaccharides will help to elucidate the relationships between
structure and function.
In conclusion, the present investigation presents a new conversion
method for polysaccharides by using a C 0 laser, and opens a new field
2
of carbohydrate technology in which anhydro-sugars can be utilized as

raw materials.
REFERENCES
[1] W.S. Duley, C0 LasersEffects and Applications, Academic Press, New

2
York (1976).
[2] C.H. Miller, T.A. Osial and A. Van der Jagt, Pulp Paper Mag. Can., 71, 63
(1973).
[3] C.W. McMillin and J.E. Harry, Forest Prod. J., 21(10), 34 (1971).
[4] S.B. Nordin, J.O. Nyren and E.L. Back, Textile Res. J., 44, 152 (1974).
[5] C.C. Peters and C M . Banas, Forest Prod. J., 27(11), 41 (1977).
[6] J. Suzuki, J. Azuma, T. Koshijima, K. Okamura and H. Okamoto, Chem.

Lett, 481 (1983).
[7] J. Azuma, M. Sakanaka, M. Zhang and K. Okamura, Bull. Kyoto Univ.
Forest, 60, 319 (1988).
[8] M. Murayama, B. Chun, J. Azuma and K. Okamura, Bull. Kyoto Univ.
Forest, 59, 310 (1987).
[9] J. Azuma, Modern Methods of Plant Analysis, New Series, Vol. 10, Plant
Fibers, 100-126, Springer-Verlag, Berlin (1989).
[10] Y. Okamori, M. Haga and S. Tejima, Chem. Pharm. Bull, 21, 2538 (1974).
[11] K. Takeo, T. Yasato and T. Kuge, Carbohydr. Res., 93, 148 (1981).
[12] H. Meier, Acta Chem. Scand., 12, 144 (1958).
[13] S.A. Madison and P.M. Keehn, J. Anal. Appl. Pyr., 9, 237 (1986).
[14] C. Schuerch, Adv. Carbohydr. Chem. Biochem., 39, 157 (1981).
[15] H. Sumitomo and M. Okada, Current Topics in Polymer Science, Vol. 1,
15-29, Hanser, Munich (1987).
Biochemical Mechanisms for the
Biodegradation of Wood
MIKIO SHIMADA
ABSTRACT
The biochemical mechanisms for the biodegradation of wood by white-rot and brown-rot
fungi are discussed on the basis of recently reported findings. The white-rot wood decay
process comprises the enzymatic hydrolysis of cellulose and oxidative breakdown of lignin;
the one-electron oxidation mechanism involved in the initial oxidation of lignin is described.
In the enzymatic breakdown of lignin, a cation radical intermediate plays a key role. In the
brown-rot wood decay process, cellulose degradation is more critical than that of lignin,
and the possible role of Fenton's system, which has been receiving keen attention recently,
is discussed in relation to the oxidative breakdown of cellulose. As an alternative hypothesis,
however, the possible role of oxalic acid in the hydrolytic cleavage of the cellulose chain is
proposed, because oxalic acid is one of the strongest physiological organic acids, and is
commonly known as a peculiar secondary metabolite produced by brown-rot fungi.
Keywords: white rot, brown rot, wood decay, biochemical mechanism, one-electron oxida-
tion, ligninase, cellulase, oxalic acid
INTRODUCTION
Research and development into the biodeterioration of materials is of

great importance from both economic and ecological viewpoints. Wood
is a unique material in many respects, and sharply contrasts synthetic
inorganic and organic materials. Wood is a natural product synthesized
by plants, and is also a substrate that can be metabolized by living matter
such as wood-destroying fungi. Wood is the food for these fungi, like
bread and butter for humans; it consists of hydrophilic carbohydrates
and hydrophobic lignin that provide an energy source for the fungi.
Biodeterioration of wood obviously occurs as a result of the action of
207
208 MIKIO SHIMADA
living matter. However, its biodegradability has recently been receiving

much attention as it has been accepted as a possible beneficial trait,
because woody materials are a non-polluting, renewable and recyclable
resource from an environmental viewpoint. Therefore, it is of great
importance to elucidate the mechanisms for the biodegradation of wood
and wood-based materials at the level of the enzymatic reactions in-
volved.
The proceedings of the first International Biodeterioration Symposium
[1], which was held in England in 1968, cover a wide variety of research
fields: (a) fundamentals of biodegradation, (b) mechanisms for bio-
degradation, (c) ecological aspects of biodeterioration, (d) testing for bio-
deterioration resistance, (e) control of biodeterioration, (f) deterioration
and protection of materials, (g) hydrocarbon microbiology and metallic
corrosion, (h) ecological aspects of timber decay, (i) biocides, (j) enzymes
and physiology of wood-destroying organisms, (k) marine fouling, and (1)
stored products microbiology.
Sixteen years later, the 6th International Symposium on Biodeteriora-
tion of Materials [ 2 ] , held in the USA in 1984, covered an even wider
range of research fields, presenting recent biotechnology-oriented papers
as a new trend. The increase in the number of papers presented clearly
shows that research on the biodeterioration of materials is consistently
attracting wide interest in many countries. However, enzymatic and
mechanistic investigations are still minor compared with the major
phenomenological ones, which are to be ultimately elucidated in chemical
language.
This paper reviews the current thinking on the biochemical mechan-
isms for the biodegradation of wood caused by wood-destroying white-
rot and brown-rot fungi, since a fuller understanding of the mechanisms
for wood decay provides important insight into the development of more
rational wood preservation methods as well as into the synthesis methods
for 'wood-mimetic' polymers or wood-based plastics which are in good
harmony with the natural environment.
DECAY OF W O O D C O M P O N E N T S
In order to develop rational wood preservation technology, it is

important to investigate wood biodegradation processes at the molecular
level. At present, in spite of some work that has been reported, the
biochemical interaction between wood-destroying fungi and timber sub-
BIOCHEMICAL MECHANISMS FOR THE BIODEGRADATION OF WOOD 209
strates is so complex that virtually nothing is known, in terms of

mechanistic studies with enzymes, of the whole biodegradation processes
in woods due to a wide variety of fungi. There are three types of wood
decaywhite rot, brown rot, and soft rotwhich are respectively caused
by white-rot, brown-rot, and soft-rot fungi. The most prominent lignin
decomposers are white-rot fungi which belong to a group of
basidiomycetes or so-called mushrooms. On the other hand, brown-rot
basidiomycetes preferentially attack cellulose and hemicellulose, and
hardly decompose lignin. Soft-rot fungi belonging to a group of molds
can moderately degrade lignin, but prefer to degrade wood polysacchar-
ides. The biodegradation of lignin by soil bacteria is considered to be
minor when compared with that of these fungi. However, since the recent
discovery of the lignin-degrading enzyme, 'ligninase', in 1983, enzymatic
studies on lignin degradation have become popular [ 3 ] and extensively
conducted with the white-rot fungus, Phanerochaete chrysosporium. The
biochemical mechanisms for lignin degradation by white-rot fungi and
for cellulose degradation by brown-rot fungi are described next.
L I G N I N D E G R A D A T I O N BY WHITE-ROT F U N G I
In order to understand the enzymatic mechanism for lignin degrada-

tion, knowledge about the chemical properties of the lignin substrate is
required. The chemical structure of a guaiacyl lignin model was first
proposed by Freudenberg as shown in Fig. 1, which indicates that the
lignin polymer consists of alkyl-aryl ether with CC bonding [ 4 ] . Lignin,
as a unique heterogeneous aromatic biopolymer sharply contrasting with
the other important biopolymers (nucleic acid, protein, starch, cellulose,
etc.), has been frequently compared to natural 'plastics', 'cementing
agents' and 'wood preservatives'.
Because specificity has been the major biological principle for explain-
ing the natural occurrence of varieties of optically active compounds that
have been synthesized by the specific action of enzyme catalysts, most
biochemists are likely to suggest that each structure is affected by a
specific enzyme of white-rot fungi. However, the author first proposed
that the non-(stereo)specificity of the biochemical devices in white-rot
fungi is important in initial lignin biodegradation, speculating the poss-
ible participation of powerful non-stereospecific oxidative agents, includ-
ing active oxygens and radical species, or laccase and peroxidase, in the
initial fungal attack on lignin [ 5 ] .
2
H COH 2
H COH
210
MIKIO SHIMADA
FIG. 1. Chemical model of the spruce lignin structure proposed by Freudenberg [4].
Although the name 'ligninase' has not been rigorously defined, it may
be used as a general term for enzymes that catalyze the breakdown of the
intermonomeric linkages in lignin. The ligninase' isolated from
Phanerochaete chrysosporium was shown to catalyze predominantly CC
bond cleavage, yielding aromatic aldehydes as the major products as
depicted in Fig. 2 [6]. This 'ligninase' is now called lignin peroxidase
(LiP), due to the enzymic requirement of hydrogen peroxide as an
oxidant like other classical peroxidases. LiP contains iron-protopor-
phyrin IX complex at the active site (Fig. 3). The synthetic iron-
tetraphenylporphyrin complex has been demonstrated to exhibit the
same reaction as that of ligninase, and has been a biomimetic model
catalyst for LiP due to its enzyme-mimicking.
OMe
CH OH
2
FIG. 2. Schematic presentation of C C bond cleavage in a lignin polymer

A p
catalyzed by ligninase as proposed by Tien and Kirk [6].
On the basis of our first report on the C Cp bond cleavage reac-

a
tion by using the 'ligninase-mimetic' iron porphyrin complex [ 7 ] ,

Schoemaker et al. [8] has proposed a one-electron oxidation mechan-
ism for the enzymatic lignin breakdown, and in the same year, Kirk et
al. [9] and our laboratory [10] independently reported almost the same
212 MIKIO SHIMADA
Hemin (Protoporphyrin IX) Tetrapheny1porphyrin

(TPP)
FIG. 3. Chemical structures of natural and synethetic iron-porphyrin complex.
one-electron oxidation mechanism which was supported by exper-

imental evidence.
Figure 4 shows the one-electron oxidation mechanism for the C C^ a
bond cleavage of a /? 1 lignin model substrate catalyzed by LiP in the

presence of hydrogen peroxide. In this reaction, hydrogen peroxide is
first reduced to water by the iron-porphyrin of the enzyme. The driving
force for CC bond cleavage is the reactive species called the 'oxenoid'
intermediate in the form of a two-electron deficient oxo-iron pophyrin
complex (X), which initially abstracts one electron from the aromatic
ring of the lignin model substrate, yielding a cation radical species (Y).
This intermediate (Y) then undergoes C C bond cleavage, yielding a
a p
benzaldehyde product and another radical intermediate (Z). The car-

bon-centered radical (Z) is instantly attacked by the dioxygen radical.
Alternatively, radical intermediate Z undergoes a further one-electron
transfer to form a benzyl cation. The phenyglycol product, which is a
counterpart of the benzaldehyde product, is produced either by oxygen-
ation with dioxygen or by hydroxylation with water. An important
finding in this reaction is the evidence that dioxygen was incorporated
into the newly formed benzylic hydroxyl group of the phenylglycol
product after CC bond cleavage.
Like the CC bond cleavage of the side chain in the dimeric lignin
model substrate, aromatic ring cleavage proceeds after the initial trans-
fer of one electron from the benzene ring of the veratryl alcohol
substrate (I), yielding the corresponding cation radical intermediate (II),
which is then oxidized to veratraldehyde (II) as the major product as
shown in Fig. 5 [11]. However, when the cation radical species is first
BIOCHEMICAL MECHANISMS FOR THE BIODEGRADATION OF WOOD
213
FIG. 4. A p
One-electron oxidation mechanism for the enzymatic C C bond cleavage of a / M lignin model substrate [10].
214 MIKIO SHIMADA
attacked by water and the radical intermediate is then oxygenated by

dioxygen, this yields five- or six-membered lactones (IV, V and VI) as
the ring cleavage products [11]. In this reaction, regiospecific oxygena-
tion with water and molecular oxygen (route a in Fig. 5) have been
demonstrated. Similarly, the /J-0-4 lignin substructure model and syn-
thetic lignin substrate undergoes aromatic ring cleavage [12]. The LiP
system is now known to oxidize a wide variety of compounds in the
presence of hydrogen peroxide as described previously [3].
In addition to LiP, laccase [5] and Mn-dependent peroxidase [Mn-
P ] [4] have also been receiving considerable attention for lignin oxidation.
However, these enzymes oxidized only the phenolic moieties of lignin and
then catalyzed the further polymerization of lignin with enzymatic
reactions that were carried out in an in vitro (test tube system). Despite this,
the enzymatic oxidation of lignin polymers by ligninase as well as laccase
did not achieve complete depolymerization, and many aspects of lignin
decomposition by white-rot fungi remain to be clarified.
CELLULOSE D E G R A D A T I O N BY BROWN-ROT F U N G I
Cellulose biodegradation principally involves hydrolytic depolymeriz-

ation of cellulosic materials to lower-molecular-weight compounds, final-
ly yielding monomeric glucose units by catalysis of the cellulase enzyme
system in living organisms. Major deterioration of wood-based lignocel-
lulosic materials is caused by microorganisms [14].
In general, wood-destroying brown-rot fungi seem to have host prefer-
ence for gymnospermous (soft) woods in nature [15], and since softwood
timber is used principally for wooden building [16], major biological
damage to wooden building materials is caused by brown-rot fungi [17].
The brown-rot fungus, Serpula lacrymans, has been reported to cause
serious damage to ezomatsu softwood (Picea jezoensis) that is used as
housing lumber in northern parts of Japan [18]. It is interesting to note
the sharp contrast in the degradation patterns of cellulose between
brown-rot and white-rot wood decay, as shown in Fig. 6 [19]. This
demonstrates that brown-rot fungus depolymerizes cellulose much more
rapidly in the early stages of wood decay, whereas white-rot fungus
depolymerizes it more slowly through the whole wood decay process. The
mechanism for the depolymerization of cellulose by white-rot fungus can
be explained in terms of the cooperative action of the exo- and endo-
cellulase system shown in Fig. 7, although the C x - C l hypothesis has been
BIOCHEMICAL MECHANISMS FOR THE BIODEGRADATION OF WOOD
FIG. 5. One-electron oxidation mechanism for the aromatic ring cleavage of a monometric lignin model substrate [11].
215
216 MIKIO SHIMADA
1600
o
H
4->
cd
M
H 1200
6
it
o
O 800
<D
<D
W>
<D
'd
<D 400
W>
CD
>
<
0 20 40 60
Average weight loss (%)
FIG. 6. Difference of the D.P. decrease in cellulose during progressive decay by

white-rot and brown-rot fungi [19].
proposed to explain the cellulose degradation mechanism. The cellulase

system consists of three enzymes that are considered to act sequentially.
First, endo-cellulase (Cx) randomly attacks amorphous regions of the
cellulose substrate to prepare the reaction sites for exo-cellulase (step A).
Then, exo-cellulase (the CI-enzyme according to Wood's definition) splits
the exo-wise cellobiose units (step B). These sequential hydrolysis reac-
tions are repeated, and the cellulose bundle is chopped down (step C).
The cellobiose and cello-oligo saccharides thus formed are finally hy-
drolyzed to glucose by the third enzyme jS-glucosidase (step D).
The mechanism for cellulose biodegradation by brown-rot fungi is not
as well characterized as that by white-rot fungi. In order to explain the
peculiarly drastic depolymerization (Fig. 6), the role of endo-glucanase
(Cx) has been receiving much attention, because brown-rot fungi such as
Poria placenta [20] and Tyromyces palustris [21] do not produce
exo-cellulase. An oxidative system has also been receiving keen attention
to explain the drastic depolymerization of cellulose, since Koenig [22]
proposed the hypothesis in his earlier work that brown-rot fungi utilize a
nonenzymatic oxidative reaction with the OH radical produced by a
Fenton-type reaction to depolymerize cellulose. He has demonstrated
Crystalline Amorphous
regions regions
Cellulose substrate
A | Endo-cellulase (Cx)
C Endo- and Exo =
cellulases
D
p-biucosidase
T\
Glucose
FIG. 7. Mechanism for the sequential hydrolysis of cellulose in the cellulase

enzyme system proposed by Wood and McCrae in Ref. 13.
that the celluloses in cotton and in agiosperm and gymnosperm woods

were rapidly depolymerized with Fenton's system, as shown in Fig. 8.
Recently, Enoki et al. [23] have suggested that a Fenton-type oxidation
occurred in a culture of brown-rot fungi, since ethylene was produced
from a-keto-methylthiobutyric acid, which is an indicator for a non-
stereospecific oxidation system. Highley et al. [24] provided more
positive evidence to support Koenig's hypothesis, reporting that cellulose
218 MIKIO SHIMADA
80
3soinTiao jo
70
60
Cotton
50
Angiosperm
'd'G.
40
SATIETY
30
20
0 10 20 30 40 50
Weight loss (%)
FIG. 8. Relationship between the D.P. decrease in cellulose in cotton and wood
samples and the weight loss of cellulosic samples by oxidation with Fenton's
system [22].
degradation occurred in the paracrystalline region of cellulose. They

isolated such oxidatively broken-down products as glyceric, erythronic,
gluconic and arabonic acids, which were also isolated from the degrada-
tion products of cotton cellulose treated with Fenton's system. They
proposed that CC bond cleavage might result in shortening of the
cellulose chain as shown in Fig. 9. However, the identification of these
aliphatic acids does not always provide proof of the involvement of
Fenton's system in physiological wood decay processes, since enzymatic
evidence for the production of these acids has not yet been offered.
Furthermore, Fenton's system yields such a destructive OH radical to
living matter as to degrade their own cell wall polymers. Schmidt et al.
[25] have demonstrated that Fenton's system is drastic enough to
decompose cellulose to carbon dioxide under the appropriate conditions
as long as hydrogen peroxide is continuously supplied. Because the OH
radical species is a very destructive oxidant, it is unlikely that such a
powerful oxidant is generated to attack cellulose selectively; lignin, which
is resistant to hydrolysis, might be decomposed as well by such drastic
oxidation. In reality, however, brown-rot wood decay does not decom-
cellulose
I Brown-roc
cellulose
cdlulos'
ceDulosc-cryihronic acid aUuiosc-eryTJironic and glycolic acids
ceilulosi .cellulose
ceMose-arabonic acid cellulose-glyceric acid

FIG. 9. Schematic model for oxidative cleavage of the cellulose chain proposed
by Kirk et al [24].
pose lignin. Interestingly, Schmidt et al [25] have also reported that

cellulose decomposition by Fenton's system was potently inhibited by
oxalic acid, which is an important physiological metabolite of brown-rot
fungi. Therefore, it is unlikely that Fenton's system functions physiologi-
cally for the cellulose breakdown during brown-rot wood decay.
At this point, on the basis of our recent experimental evidence [26], we
want to present an alternative hypothesis that, under physiological
conditions, cellulose is depolymerized by oxalic acid-catalyzed hydrolysis,
since 0 1 M or 0 9 % oxalic acid solution gives a pH of 1-3, which is strong
enough to hydrolyze cellulose randomly. However, as far as the nonenzy-
matic hydrolysis of cellulose by organic acids at physiological concentra-
tions is concerned, cellulose degradation by oxalic acid has not yet been
reported, so we tested the effect of oxalic acid on the viscosity of cellulose.
220 MIKIO SHIMADA
Table 1 shows that the incubation of pine kraft pulp with 1% oxalic acid at
35 C for four weeks decreased the viscosity of cellulose to 60% of the
original value. An increase in the concentration of oxalic acid to 5% caused
a more drastic decrease in viscosity to about 30% of the original value.
Although the physiological concentration of the acid in rotten wood has
not been measured, in view of the optimum moisture content for wood
decay, the oxalic acid concentration might be considerably higher than is
thought; cellulose as well as hemicellulose is easily depolymerized by oxalic
acid during the considerably longer period involved in the natural wood
decay process. Recently, Bech-Anderson [27] has reported that hemicel-
lulose (xylan) was hydrolyzed by treating wood meal with a 10% oxalic acid
solution. Therefore, it is proposed that oxalic acid plays a key role in the
degradation of carbohydrates during the brown-rot wood decay process.
TABLE 1
Effects of oxalic acid on the viscosity of cellulose [26]
Treatment method Viscosity (cP) Relative value
Before autoclaving 23-4 100

Incubation with water 23-0 98
Incubation with 1% oxalic acid 14-2 61
Incubation with 5% oxalic acid 7-6 32
The fact that the viscosity of cellulose is reduced by treating with

physiological concentrations of oxalic acid makes it of great import-
ance to elucidate the biochemical mechanism by which oxalic acid is
produced by brown-rot fungi in order to prevent the brown-rot wood
decay caused by oxalic acid. Quite recently, we successfully isolated from
the brown-rot fungus, Tyromyces palustris, the enzyme which catalyzes
the hydrolysis of oxaloacetic acid to yield oxalic acid and acetic acid, as
shown in Scheme 1 [26]. This is the first evidence for the enzymatic
COOH
COOH Oxaloacetase ^ COOH
6=0
6H 2
COOH H
* X
CH 3
COOH
SCHEME. 1. Hydrolytic formation of oxalic acid from oxaloacetic acid catalyzed

by oxaloacetase.
formation of oxalic acid from cell-free extracts of brown-rot fungus,

although the same enzyme has been isolated from the mold Aspergillus
niger [28], which was not being investigated with relevance to cellulose
biodegradation.
The reason for the accumulation of oxalic acid can be explained in
terms of both the presence of oxaloacetase and the absence of an
oxalic-acid-decomposing system [29, 30] in brown-rot fungi. In this
context, further research on the biochemical relationship between oxalate
formation and cellulose biodegradation is under way in our laboratory.
Currently, the biodegradation of lignocellulosic materials is receiving
widespread attention, and readers are recommended to consult the recent
relevant publications [31, 32].
CONCLUSION
Both the enzymatic hydrolysis of cellulose and hemicellulose, and the

oxidative decomposition of lignin are important in the white-rot wood
decay process. A lignin-degrading enzyme called lignin peroxidase (LiP)
seems to play an important role in lignin decomposition, involving the
one-electron oxidation mechanism at the incipient stage of white-rot wood
decay. On the other hand, brown-rot fungi utilize oxalic acid as an alter-
native diffusible hydrolytic agent which attacks the acid-labile hemicel-
lulose and cellulose chain during the brown-rot wood decay process.
REFERENCES
[1] A.H. Walters and J.J. Elphick, in Biodeterioration of Materials, Elsevier

Publishing Co., New York (1968).
[2] S. Barry, D.R. Houghton, G.C. Llewellyn and C.E. O'Rear, Biodeterioration
6, Proc. of the 6th International Biodeterioration Symposium (1984).
[3] M. Shimada and T. Higuchi, in Wood and Cellulosic Chemistry (ed. by
D.N.S. Hon and N. Shiraishi), Marcel Dekker, 557(1991).
[4] K. Freudenberg and A.C. Neish, in Molecular Biology, and Biochemistry and
Biophysics, Vol. 2 (ed. A. Kleinzeller, G.G. Springer and H . G Whittmann),
Springer Verlag, Berlin (1968).
[5] M. Shimada, in Lignin Biodegradation, Vol. I (ed. T.K. Kirk, H.-M. Chang
and T. Higuchi), CRC Press, Boca Raton, Florida, 195 (1980).
[6] M. Tien and T.K. Kirk, Science, 221, 661 (1983).
[7] M. Shimada, T. Habe, T. Umezawa, T. Higuchi and T. Okamoto, Biochem.
Biophys. Res. Commun., 122, 1247 (1984).
222 MIKIO SHIMADA
[8] H.E. Schoemaker, P J . Harvey, R.M. Bowen and J.M. Palmer, FEBS Lett,
183, 7 (1985).
[9] P.J. Kersten, M. Tien, B. Kalyanaraman and T.K. Kirk, J. Biol. Chem., 260,
2609 (1985).
[10] M. Shimada, T. Habe, T. Higuchi, T. Okamoto and B. Panijpan, Holzfor-
schung, 41, 277 (1987).
[11] M. Shimada, T. Hattori, T. Umezawa and T. Higuchi, FEBS Lett, 221, 327
(1987).
[12] T. Umezawa and T. Higuchi, FEBS Lett., 218, 255 (1987).
[13] M. Shimada and M. Takahashi, in Wood and Cellulosic Chemistry (ed.
D.N.S. Hon and N. Shiraishi), Marcel Dekker, 621 (1991).
[14] H. Wariishi, K. Valli, V. Renganathan and M.H. Gold, J. Biol. Chem., 264,
1418 (1989).
[15] D.A. Kaarik, in Biology of Plant Litter Decomposition, Vol. 1 (ed. by C.H.
Dikkerson and G.J.F. Pugh), Academic Press, New York, 129 (1974).
[16] J.G. Savory, in Biodeterioration of Materials, Elsevier Publishing Co., New
York (1968).
[17] T.C. Scheffer, in Wood Preservation and Its Prevention by Preservative
Treatments (ed. N.S. Cote), Syracuse University Press, Syracuse, New York,
31 (1973).
[18] S. Doi, Mokuzai Gakkaishi (J. Japan Wood Res. Soc), 31, 843 (1985).
[19] E.B. Cowling and W. Brown, Adv. Chem. Ser., 95, 152 (1969).
[20] T.L. Highley, Wood and Fiber, 5, 50 (1973).
[21] M. Ishihara and K. Shimizu, Mokuzai Gakkaishi, 30, 79 (1984).
[22] J.W. Koenig, Wood and Fiber, 6, 66 (1974).
[23] A. Enoki, H. Tanaka and G. Fuse, Wood Sci. Technol, 23, 1 (1989).
[24] T.K. Kirk, T.L. Highley, R. Ibach and M.D. Mozuch, Intern. Res. Group on
Wood Preserv., Document No. 1389 (1989).
[25] C.J. Schmidt, B.K. Whitten and D.D. Nicholas, Proc. Am. Wood Preserv.
Assoc, 11, 157 (1981).
[26] M. Shimada, Y. Akamatsu and M. Takahashi, Proc. Intern. Symposium of
Wood and Pulping Chemistry, Melbourne, Australia, April 29-May 3 (1991).
[27] J. Bech-Anderson, Intern. Res. Group on Wood Preserv., Document No.
1130 (1987).
[28] H. Simazono and O. Hayaishi, J. Am. Chem. Soc, 18, 5126 (1956).
[29] H. Shimazono and O. Hayaishi, J. Biol Chem., 227, 151 (1957).
[30] Y. Akamatsu, D.B. Ma, T. Higuchi and M. Shimada, FEBS Lett., 269, 261
(1990).
[31] N.G. Lewis and M.G. Paice, Plant Cell Wall Polymer: Biogenesis and
Biodegradation, Am. Chem. Soc. Symp. Ser., 399 (1989).
[32] T.K. Kirk and H.-M. Chang, Biotechnology in Pulp and Paper Manufacture,
Butterworth-Heinemann, Boston (1990).
Wood-Preserving Techniques to Prevent
Biodeterioration
MUNEZOH TAKAHASHI
ABSTRACT
Current wood-preserving techniques for preventing biodeterioration are described. Al-
though both toxic and non-toxie chemical treatments for protecting wood have been
extensively studied, the former is still the only practical method for preserving wood.
However, toxic preservatives in wide commercial use have been subjected to public
criticism, so that there is a demand for the development of low-toxicity alternatives and
safer treatment techniques. Several promising examples are described. Among the various
non-toxic treatments, acetylation, phenolic resin impregnation and wood-inorganic material
composite are worthy of consideration for their potential practical use. These treatments
are more applicable to reconstituted wood products than to thick solid timber since the
chemicals penetrate easily into the wood cell wall and voids in small-size elements.
However, a more economical processing system needs to be developed to improve the
cost-effectiveness. Safe disposal of wood products that have been treated with toxic or
non-toxic chemicals should be further considered to avoid an excessive accumulation of
non-degradable wood waste in the environment.
Keywords: wood preservative, low toxicity, non-toxic treatment, chemical modification,

resconstituted wood products
INTRODUCTION
Wood is composed essentially of only cell walls and voids. The cell wall
is a composite material constructed from a variety of organic polymers,
and the voids are filled alternately with air and water. While this basic
feature provides various advantages for their utilization, it quite often
causes undesirable biodeterioration such as microbial and insect attack
on wood in service. The organic polymers of wood cell walls, which are
not necessarily easily degradable, are abundant nutrient sources for the
223
224 MUNEZOH TAKAHASHI
specific wood-deteriorating organisms like decay fungi and termites.

Oxygen and liquid water occupying the voids in wood are essential for
the growth of these organisms.
As a general rule, however, the biodeterioration of wood, and particu-
larly microbial degradation, can only occur if there is an adequate supply
of water, a favorable temperature and oxygen present. Therefore, the
most effective means for protecting wood is to put it into service with
appropriate precautions to keep it dry. Where a safe level of these
requirements cannot be reasonably assured, protection by chemical
treatment is the alternative.
Preservatives make wood toxic to wood-deteriorating organisms, while
other chemicals modify the characteristics of wood and make it resistant
to biological attack.
This paper reports the general aspects of current treatments, focusing
on the control of decay and termite damage.
D E V E L O P M E N T OF ALTERNATIVES TO C O N V E N T I O N A L
W O O D PRESERVATIVES
Background
Until the 1960s, creosote, copper-chrome-arsenate (CCA), chlorinated
hydrocarbons {insecticidal D D T , y-BHC, drin chemicals, heptachlor and
chlordane and fungicidal pentachlorophenol (PCP)}, and organotin com-
pounds were the predominant wood preservatives.
Among these, most insecticidal chlorinated carbons were prohibited for
use in many countries, because of the detrimental side effects which were
demonstrated by environmental surveys and toxicological studies. How-
ever, chlordane was allowed for use in Japan until 1986. PCP and
water-soluble Na-PCP have been used world-wide, but they were banned
in Japan for the same reason about ten years ago.
Public criticism about the toxicity and detrimental environmental
effect is also strong against the other major historical preservatives. Bis
(tri-n-butyltin) oxide (TnBTO) also disappeared recently from the market,
this being triggered by the detection of TnBTO in sea-water bred fish
which were enclosed by a preservative-treated breeding net.
The preserving performance of creosoted wood products has been well
evaluated, but a few disadvantages have also been pointed out. These
include the poor paintability of treated products, unfavorable odor,
dermato-inflammation and the carcinogenetic substances in creosote.
WOOD-PRESERVING TECHNIQUES TO PREVENT BIODETERIORATION 225
Although some improvements to these undesirable properties have been

achieved in Australia [1, 2 ] , the consumption of creosote has been
dropping in Japan [ 3 ] , following the increasing world-wide trend to
replace wooden cross-ties and poles by concrete and steel products [4, 5].
As is well recognized, CCA treatment can provide good protection
against decay fungi, termites and other wood-boring insects. However,
CCA has been exposed to serious criticism against its future utilization
due to the yield of toxic hexavalent chromium ash and gaseous arsenious
acid when treated waste is burned [ 6 ] .
The search for acceptable alternatives to these preservatives is an
important and urgent research project. Risky preservatives should be
banned, and alternatives must be environmentally and toxicologically
acceptable on the basis of current scientific criteria. The cost effectiveness
of alternatives is also important for practical commercialization.
Oil-borne Preservatives
Several oil-borne chemicals were recently evaluated as wood preserv-
atives and commercialized.
Two organoiodine compounds, 4-chlorophenyl-3-iodopropargyl for-
mal (IF-1000) [ 7 - 1 2 ] and 3-bromo-2,3-diiodo-2-propenyl carbonate
(EBIP) [12-14], were developed in Japan to protect wood from fungal
attack. Two other organoiodine chemicals, 3-iodo-2-propynyl butyl car-
bonate (IPBC) [15-24] and 2,3,3-triiodoallyl alcohol (TIAA) [25], also
became available for the same purpose. Comprehensive studies have been
conducted by Lee et al. to evaluate these compounds as wood preserv-
atives [14, 24-28].
The order of their anti-decay effect after a single-brushing treatment
was TIAA ( 0 5 % ) > I P B C ( 0 - 7 5 % ) > I F - 1000 ( 1 0 % ) > E B I P (10%). The
effect of TnBTO was equal to IF1000. Mixing each compound with
termiticidal chlorpyrifos and a surface active agent gave a synergistic
fungicidal effect against white-rot fungus, Coriolus versicolor. However,
an effect was not always found against two brown-rot fungi, Tyromyces
palustris and Serpula lacrymans. Only IPBC gave a desirable synergistic
effect against all the fungi tested by mixing with a surface active agent.
IF1000 and T N B T O caused an undesirable antagonistic effect against
T palustris when mixed with chlorpyrifos. EBIP showed a synergistic
and/or additive effect against T. palustris and S. lacrymans only when
they were mixed with commercial synergistic and stabilizing agents.
The crossed-paper technique [29, 30] has been extensively used in
biochemical and medical research to determine the effectiveness of
various chemical formulations. This technique was first applied in wood

preservation studies to evaluate the mixing effects of different chemicals
[26] (Tables 1 and 2).
TABLE 1
Effects of mixing organoiodine fungicides with chlorpyrifos (CP) and/or a surface
active agent (S, polyoxyethylene sorbitan fatty acid ester) as evaluated by
crossed-paper tests [26]
Chemical TYP SEL cov
CP S CP +S CP S CP +S CP s CP +S
IF-1000 _ + ++ + +++ + + + + +
IPBC ++ ++ +++ ++ ++ +++ ++ ++ + + +
EBIP + + + + + + + + + +
TIAA + ++ + + ++ ++ - + + +
TnBTO - ++ + +++ +++ + + +
TYP, Tyromyces palustris; SEL, Serpula lacrymans; COV, Coriolus versicolor.
+ independent action, + + additive action, + + + synergistic action,
antagonistic action.
TABLE 2
Effects of mixing EBIP with chlorpyrifos (CP), a synergistic agent (S, octa-
chlorodipropylether), and a stabilizer (N, sodium-dioctylsulphosuccinate) as
evaluated by crossed-paper tests [26]
Fungus CP S N CP + S CP + N CP + S + N
TYP ++ + + ++ + +++
SEL ++ ++ ++
COV ++ + ++ ++ +++ +++
Symbols are the same as those in Table 1.
A triazole compound, 1 -{2-(2,4-dichlorophenyl)-1,3-dioxolon-2-yl}

methyl-lH-l,2,4-triazole (Azaconazole) was recently introduced by the
wood preservation industry after the chemical performed well against
decay fungi and sap-stain fungi [31, 32]. However, the preserving
performance was not satisfactory against a Japanese strain of S. la-
crymans [33]. A benzothiazole compound, 2-(thiocyanomethylthio) ben-
zothiazole (TCMTB), performed well as an active ingredient of anti-
sap-strain formulations [34-38].
Several organophosphate and carbamate compounds have been mar-
keted in Japan as post-chlordane termiticide: 0,0-diethyl-0-(3,5,6-tri-
chloro-2-pyridil) phosphorothioate (chlorpyrifos), O,0-diethyl-0-(a-cyano-

benziridenamino) thiophosphate (phoxim), 0,0-diethyl-0-(3-oxo-2-phenyl-
2H-pyridazine-6-yl) phosphorothioate (pyridaphenthion), 0,0-dimethyl-0-
(3-methyl-4-nitrophenyl) thiophosphate (fenitrothion), 2-chloro-l-(2,4,5-
trichlorophenyl) vinyldimethylphosphate (tetrachlorovinphos), 2-iso-
propoxyphenylmethylcarbamate (propoxur), 2-sec-butylphenyl-N-
methylcarbamate (BPMC), and 1-naphthyl-N-methyl-carbamate (car-
baryl).
These termiticidal chemicals are used mainly for the in situ treatment
of housing timber members. Their emulsified formulations are applied to
treat the soil under the floor. Although their performance is generally
satisfactory, some disadvantages also need to be pointed out. The
comparatively low persistency as a soil termiticide is due to their weak
resistance to microbes and alkalinity [39]. With the increasing human
health awareness, occasional poisoning caused by the inhibition of
acetylcholine-esterase activity might make workers reluctant to treat soil
with the above-mentioned termiticides. Several new techniques have been
developed to improve the application environment such as ventilated
working spaces, automatic foam spraying, filmed, microcapsuled, pow-
dered and film-forming types of soil termiticide, and termiticide-treated
vinyl ground sheets [6, 40].
Insecticidal synthetic pyrethroids are advantageous for safety because
of their rapid detoxification in warm-blooded animals. Bifenthrin, cyflu-
thrin, cyhalothrin, cypermthrin, deltamethrin, esfenvalerate, fenpropath-
rin, fenvalerate, fluvarinate, permethrin, tralomethrin, etc. are applicable
as wood and soil termiticides [6, 23, 41, 42]. Pyrethroids might take the
place of organophosphate and carbamate termiticides in the near future.
l,3,5-Tri-n-propyl-l,3,5,-triazine-2,4,6-(lH, 3H, 5H)-trion (TPIC) is
characterized by its slow-acting but highly repellent performance, and is
commercially available [43-51].
Water-borne Preservatives
Alkylammonium compounds (AAC), which were developed as in-
dustrial biocides, produced promising results in the laboratory as an
alternative to CCA [52-58]. However, fungus-cellar and field trials
did not allow their use in soil-contract situations [59, 60]. The low
toxicity and high anti-termite performance of AAC were well appreciated,
and combination with copper and other agents enhanced their biological
resistance [58, 59, 61, 62]. Although no kind of AAC is currently
used in above-ground situations in New Zealand where it was originally
marketed, didecyldimethyl ammonium chloride (DDAC) was recently

approved in Japan as an active ingredient of anti-fungal and anti-termite
formulations.
The substitution of phosphorus, boron, fluorine, or zinc for arsenic or
chromium is another important research project to develop alternatives
to CCA. Several formulations (Cu-Cr-P, Cu-Cr-B, Cu-Cr-F and C u -
Zn-F) have recently been re-evaluated in European countries as wood
preservatives for unweathered housing members. More recently in Japan,
micro-emulsified metal naphthenates have become available as wood
preservatives for pressure treatment, instead of the original oil-borne
compounds [63, 64].
The next generation of wood preservatives must not only be of low
toxicity but also be odorless and colorless formulations with good
paintability and low metal corrosiveness, since the demand for preservat-
ive-treated wood is diversifying into a wide range of commodities with
different biological hazard levels.
I M P R O V E M E N T OF PRESERVATION TREATMENT
Pressure Treatment
The Bethell process is still the principal technique in Japan for pressure
impregnating conventional water-borne preservatives. The Ruping and
Lowry processes are employed for treating with oily and oil-borne
preservatives. These treatments were introduced into Japan early in this
century, and several undesirable features have long been recognized.
Seasoning or kiln drying before and after treatment, pre-incising for
treating refractory wood species, low recovery of expensive organic
solvent, and low quality control of treated timber as an industrial
material might inhibit the Japanese wood preserving industry from
expanding the potential market even after the successful development of
treated wood foundation sills during the past two decades, instead of the
decline of wooden poles and sleepers [64].
Improvements to the pressure treatment process have been made by
several workers and have been highly evaluated. O P M (the oscillating
pressure method) and APM (the alternative pressure method) were
developed to treat green timber and to obtain even and deep penetration
of the wood preservatives [66, 67]. Treatment with light organic solvent
preservative (LOSP) is aimed primarily at protecting finished joinery
against accidental wetting during the interval between leaving the factory
and installation in a building [68]. The technique is worthy of consider-

ation for treating pre-cut housing timber and other dried materials
without undesirable dimensional changes [65], while the recovery of
solvent is essential if the technique is to be economically successful. Two
solvent recovery processes, using petroleum solvent, were developed in
USA in the late 1960s, under the trade names of'Celon' and 'Drilon'. The
succeeding 'Dow' and 'RentokiF processes are different from them in the
use of the non-flammable solvent, methylene chloride [66]. These
methods were devised for the special treatment with PCP or TnBTO.
More recently, a new solvent recovery process has been developed in
Japan [69]. This process is characterized by the high availability to
various combinations of preservative and solvent, and is thought to be
particularly applicable for treating wood and wood-based materials in a
factory manufacturing process such as the production of plywood,
particleboard and joinery.
As yet, however, no commercial plants offering the above-mentioned
processes are in operation in Japan. This is principally due to the high
installation and operating costs, and the otherwise passive attitude of the
industry for developing a potentially good market.
Superficial Treatment (Brushing, Spraying and Dipping)

A large amount of timber is treated by these simple methods, although,
as is generally known, the penetration and retention of preservatives are
much lower than is possible with pressure methods.
Brushing and spraying are the only methods for treating wood once
erected, but periodic re-treatment is necessary because of the shallow
penetration. The usual preservatives are creosote and organic solvent
types. The major application is for the preventive and remedial treatment
of wood in buildings, but both are employed for outdoor timber. The
resultant risk of incomplete coverage and of insufficient penetration often
causes unreliable performance from such treatment. The latest amend-
ment to the standardized testing method for evaluating the effectiveness
of wood preservatives by superficial treatment [70] will require them to
offer higher resistance to weathering. Therefore, higher treating strength
will be necessary for most preservatives to meet the required performance
of the amended standard. This will stimulate the partial replacement of
brushing and spraying by LOSP pressure treatment, because of its
advantages of more reliable performance, productivity, quality control
and resultant cost-effectiveness in comparison with superficial coating
[65].
Dipping is better than brushing and spraying because all surfaces can
absorb liquid freely, and the time for dipping can be suited to the
standard of treatment required. In Japan, dipping is applied exclusively
to control sap-staining fungi and surface mold in fresh-sawn timber.
Timber is loaded on pallets and submerged in a preservative-containing
tank for a short time. Although the treatment is targeted for short-term
protection until leaving the factory or until the next process, undesirable
failures have often occurred during the interval. In most cases, this has
been due to the degeneration of the working solution in a dipping tank.
The concentration of a freshly-prepared solution is checked as a matter
of routine, but periodic re-checking of the solution is poorly conducted
in many treatment plants. Preservative suppliers have to offer a frequent
analysis service to maintain the preservative performance.
E N H A N C E M E N T O F BIOLOGICAL RESISTANCE BY
CHEMICAL TREATMENT
Chemical Modification
Chemical modification of wood generally involves any chemical reac-
tion between the hydroxyl groups of principal wood components and a
single chemical reagent. Hydroxyl groups are the most abundant reactive
sites in the wood cell wall polymers and, therefore, provide the biological
enzymatic reaction sites. Chemical modification has been studied to
enhance such end-product properties as dimensional stability, fire retar-
dancy and biological resistance. There are often some problems with the
chemical treatment of wood that may restrict commercialization. How-
ever, the improvement of biological resistance will be more important for
wood preservation in the near future because it yields no risk of hazard
to health and the environment in the use of end-products.
The reagents used for chemical modification can be divided into the
three classes by the type of bond formed between them and the hydroxyl
group in the wood cell wall: esters, acetals and ethers [71]. An increase in
the resistance to biological attack has been recognized in all classes of
chemical modification with esters: acetic anhydride [72-81], 2,4-tolylene
diisocyanate [82, 83], methyl isocyanate [83]; with acetals; formaldehyde
[84, 85]; and with ethers: jS-propiolactone [86], acrylonitrile [86, 87],
epichlorohydrin or dichlorohydrin [88], propylene oxide and butylene
oxide [79, 89-91]. Some examples of the results are shown in Tables 3-7.
Wood-rotting fungi and termites each have a very specific enzyme
TABLE 3
Percentage weight loss in chemically-modified southern
yellow pine blocks after a 12-week exposure to the
brown-rot fungus, Gloeophyllum trabeum [71]
Chemical Percentage Percentage

weight weight
gain loss
Control 0 62-9
Propylene 200 38-6
oxide 26-7 32-8
Butylene 7-0 18-8
oxide 230 20
Methyl 11-6 12-8
isocyanate 24-3 2-3
Acetic 15-6 2-3
anhydride 24-6 0-8
TABLE 4
Percentage weight loss in chemically-modified southern
yellow pine blocks after a 12-week exposure to the
brown-rot fungus, Lentinus lepideus [71]
Chemical Percentage Percentage

weight weight
gain loss
Control 0 44-2
Propylene 51 17-5
oxide 24-0 4-8
Butylene 26-3 1-8
oxide
Methyl 21-6 10
isocyanate
Acetic 15-4 3-9
anhydride 18-6 0-7
system capable of degrading wood polymers into digestible units. If the

substrate for these systems is chemically changed, this enzymatic action
cannot take place. Therefore, the chemicals used for modification need
not be toxic to the wood-deteriorating organisms, because their action
renders the substrate unrecognizable or indigestible as a food source to
support biological growth [76].
TABLE 5
Percentage weight loss in chemically-modified southern yellow pine
blocks after a 2-week exposure to the subterranean termite. Re-
ticulitermes flavipes [71]
Chemical Percentage Block type Percentage

weight weight
gain loss
Control 0 31
Propylene 9 Interior 24
oxide Surface 19
17 Interior 15
Surface 13
34 Interior 5
Surface 8
Butylene 27 Interior 4
oxide Surface 4
34 Interior 3
Surface 3
TABLE 6
Percentage weight loss in acetylated southern pine and aspen flake-
boards after a 12-week exposure to the brown-rot fungus, Gloeophyl-
lum traheum (6% phenolic adhesive) [71]
Wood Percentage Percentage weight loss

weight
gain Un-leached Leached*
Southern 0 151 34-3

pine 60 4-3 29-3
14-8 0 0-8
18-4 0 0
Aspen 0 28-5 44-1
7-3 30 22-4
13 6 0-4 2-6
17-9 0-9 01
* After 2 weeks of water leaching.
Chemical modification has been studied on limited species of wood and

decay fungi, so that generalizations need to be presented very carefully.
However, the results with acetylated wood by the present author suggest
that a different effect exists between brown- and white-rot fungi [80].
Brown-rot fungus Tyromyces palustris was more resistant against acetyla-
TABLE 7
Percentage weight loss in acetylated albizzia particleboards after
an 8-week exposure to brown-, white-, and soft-rot fungi [81].
Percentage Adhesive Percentage weight loss

weight
gain TYP COV CHG
0 PF 21-7 100
5 PF 18-4 7-7
12 PF 11 0-7
20 PF 0 0
0 IC 26-8 24-3 150
5 IC 24-1 21-6 0-7
12 IC 4-3 0 0
20 IC 0 0 0
PF, phenol-formaldehyde resin; IC, isocyanate resin. TYP,
Tyromyces palustris (brown-rot); COV, Coriolus versicolor (white-
rot); CHG, Chaetomium globosum (soft-rot).
tion than white-rot Coriolus versicolor. A 20% weight gain (WPG) was
required to suppress decay by T. palustris in any wood species. On the
other hand, C. versicolor failed to attack acetylated Japanese cedar even
at 6 WPG, although 20 W P G was necessary to suppress its decay of
beech and albizzia (Indonesian fast-growing hardwood).
In spite of the numerous works on the chemical modification of wood,
commercial applications have not yet been attempted in any country
except for acetylation in Japan [92]. Wood modification is effective for
thin or small wood sizes because of its good chemical penetration. A
Japanese commercialized acetylated product is a laminated veneer lum-
ber (LVL) in which 3 mm-thick veneer is treated with an acetic anhydride,
using sodium acetate as a catalyst, prior to hot-pressing. Acetylated LVL
is applied to end-products that are exposed to continued dampness or
outdoor weathering conditions such as bathroom elements, external sign
boards, and roof shingles on traditional wooden shrine buildings.
Property enhancement of reconstituted wood products by acetylation
is also considered promising since the required chemicals react easily with
wood constituents in small-size elements. The biological resistance and
dimensional stability of flakeboard and particleboard were greatly im-
proved by acetylating the wood elements before board production [71,
78-81, 93-95]. Reconstituted wood products are becoming more import-
ant throughout the world as an effective solution to decreasing forest
resources. However, a more economical processing system needs to be
developed to improve the cost-effectiveness before chemically modified

reconstituted products can be commercially produced.
Treatment Other than by Chemical Modification

Many chemical systems that affect the cell wall and fill the void spaces
in wood have been studied to increase the utility of solid wood. W o o d -
polymer composites (WPC) include various vinyl composites as well as
several earlier products such as heat-stabilized wood (Staybwood), heat-
stabilized compressed wood (Staypak), phenol-formaldehyde-treated wood
(Impreg and Compreg), polyethylene glycol(PEG)-bulked wood, etc.
Some of these treatments have been applied for dimensional stabilization
and improvement of the physical and mechanical characteristics.
Recently, an earlier treatment with phenol-formaldehyde (PF)-resin has
been re-examined for the production of biologically resistant particle-
board. The decay of Japanese cedar and western hemlock blocks was
suppressed completely with less than 10% impregnation by PF-resin of
170 molecular weight, while a weaker effect existed when using another
PF-resin of 300 molecular weight [96] (Fig. 1). This suggests the possible
better penetration of small-molecule resin into the wood cell walls. The
100
$ 60
o
80
1 1 1LI11 >1 1
40
0o) 20
i SX' i1 1 1 Dl
* 0 1 1
I

20
\ *s
V1
\
5 40 \
\
\
tt
60
I
\
80 \-
\
\
20 16 12 8 4 0 4 8 12 16 20
10 0.
Percentage resin loading
\
FIG. 1. Effect of a phenol-formaldehyde resin treatment of western hemlock on
i> i i
the resistance to fungal attack for 12 weeks [96]. Relative weight
loss = W2/Wl x 100. Wl is the % weight loss of untreated wood, W2 is the %
weight loss of phenol-formaldehyde resin-treated wood. 170 molecular weight,
300 molecular weight. Solid line: Tyromyces palustris (brown-rot); broken line:
Coriolus versicolor (white-rot).
PF-resin treatment also affected the virulent termite, Coptotermes for-

mosanus, causing severe depletion of feeding activity and higher mortality
with a 5-15% resin impregnation. Thus, a low-molecular-weight PF-resin
treatment was evaluated for yielding better biological resistance than that
by acetylation. Enhanced biological resistance was again found in particle-
boards from PF-resin-impregnated particles [97]. These boards were
dimensionally stable and performed well mechanically, and are worthy of
consideration for commercial applications.
A new complex material, WIC (wood-inorganic material composite)
has recently been developed in Japan [98]. A double diffusion treatment
of veneers by two aqueous solutions of inorganic chemicals leads to the
formation of water-insoluble deposits within the cell walls and voids.
Water-impregnated veneers were introduced into solutions (I and II)
successively at 50 C for a desired period of time. Saturated aqueous
solutions were prepared from (I) barium chloride plus a small amount of
boric acid, and (II) ammonium phosphate plus a small amount of boric
acid. After treating, the veneers were exposed to running water to wash
out the excess chemicals, dried under ambient conditions for a suitable
period and then oven-dried. The percent weight gain (WPG) calculated
on an oven-dried weight basis indicates the level of treatment. Treated
veneers were glued and made into laminated veneer lumber (LVL) or
plywood. The main research target of this treatment has been the
improvement to the fire resistance of wood. Plywood made from a treated
veneer of Douglas fir was tested for its incombustibility, and the results
indicated that it met the performance requirements for a 'very slow-
burning material' designated in Notification No. 1231 of the Ministry of
Construction of Japan. While a higher WPG (approximately 100) is
needed to improve fire resistance, the inhibition of biological attack was
achieved by a relatively small W P G (Tables 8 and 9). These facts
TABLE 8
Percentage weight loss in wood-inorganic materials (WIC; 9 mm thick
LVLs, 25% weight gain) after an 8-week exposure to Tyromyces palustris
(brown-rot) and Coriolus versicolor (white-rot) [98]
Wood T. palustris C. versicolor
WIC Untreated WIC Untreated
Douglas fir 0-8 9-8 1-4 171

Lauan 11 250 11 19-3
Beech 1-3 300 1-3 30-5
TABLE 9
Percentage weight loss in wood-inorganic materials (WIC: 9 mm thick
LVLs, 25% weight gain) caused by subterranean termite Coptotermes
formosanus, and the mortality of the termites after a 3-week test [98]
Wood Weight loss (%) Mortality (%)
WIC U ntreated WIC U ntreated
Douglas fir 2-3 190 100 11

Lauan 3-5 291 100 13
Beech 2-4 29-7 100 2
encourage the extension of research into examining other combinations

of inorganic chemicals [99] and the development of more economical
treating systems. More recently, a small scale commercialization of WIC
resulted in a fire-retardant wooden door [100].
F U T U R E CONSIDERATIONS
As already pointed out throughout the world, some widely used wood
preservatives have been publicly criticized because of their relatively high
mammalian toxicity and environmental impact. This fact supports the
strong social demand for low-toxicity alternatives to these unsafe chemi-
cals that are on the list of commercial wood preservatives. Investigations
need to be conducted to develop new wood preservatives based on
screening tests of various chemicals in the laboratory and field. In
addition, the practical application of potentially useful chemicals for
treating and improving wood should be tested, and the quality of the
treated products should be fully evaluated. Various chemical treatments
of wood have been categorized as a non-toxic preservation process, and
some have been evaluated as suitable for wider utilization with recon-
stituted wood products. However, such non-toxic treatments cannot
replace preservative treatments, even if the current low cost-effectiveness
could be improved. Non-toxic chemical treatments are principally for
prevention, and not for remedial and extermination treatments. In
addition, chemical treatments are far more applicable to the production
of reconstituted wood products than to thicker solid timber.
The ecological and physiological approaches to prevent damage by
wood-attacking termites need to be investigated so that the use of
conventional termiticides can be reduced. It would be worth searching for
chemicals which disturb the close relationship between termites and

symbolic protozoa, and for slow-acting or repellent chemicals related to
the habits and behavior of termites.
Biological susceptibility and flammability usually decrease the value of
wood in use, although these features are advantageous for waste treat-
ment. Wood products treated with toxic or non-toxic chemicals eventual-
ly become waste at the end of their service life, and an excessive
accumulation of non-degradable and non-flammable wood waste would
cause a difficult problem in the future. We must consider the proper use
of treated wood products which can be safely disposed of and easily
degraded after use.
REFERENCES
[1] H. Greaves, C.W. Chin and J.B. Watkins, Intern. Res. Group on Wood
Preserv., Document No.: IRG/WP/3322 (1985).
[2] L.T. Cookson and H. Greaves, Holzforschung, 40, 59 (1986).
[3] Japan Wood Preserving Industry Association, Statistical Data1989
(1990).
[4] J.T. Micklewright, Proc. Am. Wood-Preserv. Assoc., 84, 343 (1986).
[5] New Zealand Ministry of Forestry, Statistics of the Forests and Forest
Industries of New Zealand to 1987, 100 pp (1988).
[6] G. Fuse, Mokuzai Gakkaishi (J. Japan Wood Res. Soc), 34, 709 (1988).
[7] M. Kajita, Text of a lecture on wood preservatives (July 18,1980, Osaka), 57
(1980).
[8] G. Fuse, A. Enoki and H. Tanaka, J. Antibact. Antifung. Agents, 9,273 (1981).
[9] G. Fuse, A. Enoki and H. Tanaka, Kaoku Gaichu, 1985, 25 (1985).
[10] M. Takahashi, K. Tsunoda, Y. Imamura, A. Adachi and K. Nishimoto,
Mokuzai Hozon (Wood Preservation), 11, 84 (1986).
[11] G. Fuse, A. Enoki and H. Tanaka, J. Antibact. Antifung. Agents, 14, 215
(1986).
[12] K. Tsunoda and K. Nishimoto, Intern. Biodeterioration Bull, 22, 27(1986).
[13] Y. Morisawa, M. Kataoka, H. Nagahori, T. Sakamoto, Y. Nakata, K.
Konishi, T. Aiba, T. Mizuno, S. Ono, H. Matsunaga and K. Nishimoto,
Holzforschung, 38, 225 (1984).
[14] D.H. Lee, M. Takahashi and K. Nishimoto, J. Antibact. Antifung. Agents,
16, 173 (1988).
[15] C.R. Coggins, Ann. Conv. Brit. Wood-Preserv. Assoc., 1982, 65 (1982).
[16] A.J. Cserjesi, E.L. Johnson and A. Bryne, Forintek Canada Corp. Con-
tract, No. 65-68-564 (1983).
[17] A.J. Cserjesi, E.L. Johnson, Intern. Res. Group of Wood Preserv., Docu-
ment No: IRG/WP/3281 (1984).
[18] J. Hansen, Intern. Res. Group on Wood Preserv., Document No.:
IRG/WP/3295 (1984).
[19] H.M. Barnes, B. Buchman and T.L. Amburgey, Proc. Am. Wood-Preserv.
Assoc., 81, 1 (1985).
[20] R.S. Smith, A. Bryne and A.J. Cserjesi, Forintek Canada Corp., Special
Publication, No. sp-25 (1985).
[21] R.C. DeGroot and M. Stroukoff, USDA-Forest Service, FRL-RP-481
(1985).
[22] T.C. Scheffer, J.J. Morrell and M.A. Newbill, Forest Prod. J., 37(7/8), 51
(1987).
[23] T. Yamashita, Y. Minamite, Y. Katsuda and K. Nishimoto, Mokuzai
Hozon, 14, 268 (1988).
[24] D.H. Lee, K. Tsunoda and M. Takahashi, J. Antibact. Antifung. Agents, 18,
365 (1990).
[25] D.H. Lee, K. Tsunoda and M. Takahashi, Mat. u. Org., 25, 145 (1990).
[26] D.H. Lee, K. Tsunoda and M. Takahashi, Mokuzai Gakkaishi, 36, 1095
(1990).
(1991).
(1991).
[29] Y. Uesugi, M. Katagiri and O. Noda, Agr. Biol. Chem., 38, 907 (1974).
[30] M. Katagiri and Y. Uesugi, Phytopathology, 67, 1415 (1977).
[31] A. Leclercq, Mat. u. Org., 18, 65 (1983).
[32] A.R. Valcke and W.R. Goodwine, Proc. Am. Wood-Preserv. Assoc., 81, 196
(1985).
[33] D.H. Lee, M. Takahashi and K. Tsunoda, J. Antibact. Antifung. Agents, 18,
169 (1990).
[34] L.E. Leightley, Intern. Res. Group on Wood Preserv., Document No.:
IRG/WP/3331 (1985).
[35] L.E. Leightley, Intern. Res. Group on Wood Preserv., Document No.:
IRG/WP/3374 (1986).
[36] J.A. Drysdale, Intern. Res. Group on Wood Preserv., Document No.:
IRG/WP/3416 (1987).
[37] K. Tsunoda, Mokuzai Hozon, 14, 215 (1988).
[38] K. Tsunoda, Mokuzai Hozon, 16, 130 (1990).
[39] M. Tanaka, T. Iijima and M. Higaki, Mokuzai Hozon, 15, 105 (1989).
[40] T. Takeuchi, T. Kobayashi, T. Nishikawa, K. Hirooka, T. Yamaguchi and
G. Fuse, Mokuzai Hozon, 13, 99 (1987).
[41] B.M. Kard, J.K. Mauldin and S.C. Jones, Pest Control, 57(10), 58 (1989).
[42] M. Takahashi, Mokuzai Kogyo {Wood Industry), 45, 150 (1990).
[43] K. Nishimoto, Kankyo Kami Gijutsu, 3, 421 (1985).
[44] S. Yaga, Kankyo Kami Gijutsu, 3, 478 (1985).
[45] K. Samejima, Kankyo Kami Gijutsu, 3, 541 (1985).
[46] Y. Nakajima, Kankyo Kami Gijutsu, 3, 600 (1985).
[47] K. Kanai, Y. Takebe and T. Iwamoto, Kankyo Kami Gijutsu, 3, 665 (1985).
[48] K. Kanai, Y. Takebe and T. Iwamoto, Kankyo Kami Gijutsu, 3, 747 (1985).
[49] K. Kido, H. Takahashi and M. Yoshimoto, Mokuzai Hozon, 12, 190 (1986).
[50] H. Takahashi, Y. Murakami, M. Yoshimoto, K. Kido, Y. Iwane, N.
Tanaka and K. Nishimoto, Mokuzai Gakkaishi, 31, 504 (1985).
[51] K. Kanai, H. Takahashi and S. Yaga, Appl. Ent. ZooL, 23, 15 (1988).
[52] J.A. Butcher, A.F. Preston and J.A. Drysdale, Forest Prod. J., 27(7), 19
(1977).
[53] D.J. Cross, Mat. u. Org., 14, 105 (1979).
[54] A.F. Preston and D.D. Nicholas, Wood and Fiber, 14, 37 (1982).
[55] C D . Howick, J.W. Crefield, J.A. Butcher and H. Greaves, Mat. u. Org., 17,
81 (1982).
[56] K. Tsunoda and K. Nishimoto, Intern. Res. Group on Wood Preserv.,
Document No.: IRG/WP/3232 (1983).
[57] C E . Sundman, Intern. Res. Group on Wood Preserv., Document No.:
IRG/WP/3299 (1984).
[58] L.R. Wallace, Proc. Am. Wood-Preserv. Assoc., 82, 159 (1986).
[59] J.A. Drysdale, New Zealand J. For. Sci., 13, 354 (1983).
[60] J.N.R. Ruddick, Intern. Res. Group on Wood Preserv., Document No.:
IRG/WP/3248 (1983).
[61] M. Hedley, K. Tsunoda and K. Nishimoto, Wood Research, No. 68, 37
(1982).
[62] M. Takahashi, Y. Imamura, K. Tsunoda, A. Adachi and K. Nishimoto,
Intern. Res. Group on Wood Preserv., Document No.: IRG/WP/3633 (1990).
[63] T. Takeuchi, H. Matsusaka, T. Kobayashi, S. Akao and M. Higaki,
Mokuzai Hozon, 15, 15 (1989).
[64] T. Takeuchi, H. Matsusaka, T. Kobayashi, S. Akao and M. Higaki,
Mokuzai Hozon, 15, 27 (1989).
[65] K. Tsunoda, Intern. Res. Group on Wood Preserv., Document No.:
IRG/WP/3596 (1990).
[66] J.G. Wilkinson, Industrial Timber Preservation, The Rentokil Library,
Associated Business Press, London, 532 pp (1979).
[67] A.J. Bergevoet, Intern. Res. Group on Wood Preserv., Document No.:
IRG/WP/3187 (1982).
[68] H. Greaves, Styrelsen f. Tekniskutveckling, STU information No. 438-1984
(1982).
[69] K. Iwasaki, Mokuzai Hozon, No. 26, 40 (1985).
[70] Japan Wood Preserving Association, JWPA Standard-I (1989).
[71] R.M. Rowell, in Wood and Cellulosic Chemistry, ed. D.N.S. Hon and N.
Shiraishi, Marcel Dekker, New York, 703 (1991).
[72] I.S. Goldstein, E.B. Jeroski, A.E. Lund, J.F. Nielson and M. Weater, For.
Prod. J., 11, 363 (1961).
[73] H. Tarkow, A.J. Stamm and E.C.O. Erickson, USDA-Forest Service, For.
Pest Lean. Rep. No. 1953 (1961).
[74] Koopers' Acetylated Wood, New Materials Technical Information (RDW-
400) E-106 (1961).
[75] M.D. Preston and R.J. Thomas, Wood and Fiber, 10(3), 149 (1978).
[76] R.M. Rowell, in The Chemistry of Solid Wood, ed. R.M. Rowell, ACS
Symposium Series 207, ACS, Washington D.C., 175 (1978).
[77] L.R. Gjovik and H.L. Davidson, USDA-Forest Service, FRL-RP-02
(1979).
[78] Y. Imamura, K. Nishimoto, Y. Yoshida, S. Kawai, T. Sato and M. Nakaji,
Wood Research, No. 73, 35 (1986).
[79] Y. Imamura and K. Nishimoto, Mokuzai Gakkaishi, 33, 25 (1987).

[80] M. Takahashi, Y. Imamura and M. Tanahashi, Intern. Res. Group on
Wood Preserv., Document No.: IRG/WP/3540 (1989).
[81] S. Yusuf, M. Takahashi and Y. Imamura, Mokuzai Gakkaishi, 35, 633
(1989).
[82] K. Lutomski, Mat. u. Org., 10, 255 (1962).
[83] R.M. Rowell and W.D. Ellis, in Urethane Chemistry and Applications, ed.
K.N. Edwards, ACS Symposium Series No. 172, ACS, Washington D.C.,
263 (1981).
[84] A.J. Stamm and R.H. Baechler, For. Prod. J., 10, 22 (1960).
[85] M. Stevens, J. Schalck and J.V. Raemdonck, Int. J. Wood Preserv., 1979, 57
(1979).
[86] I.S. Goldstein, W.A. Dresher and E.B. Jeroski, Ind. Eng. Chem., 51, 1313
(1959).
[87] G. Fuse and K. Nishimoto, Mokuzai Gakkaishi, 1, 157 (1961).
[88] L.O. Pihl and I. Olsson, British Patent 1 107 777 (1968).
[89] R.M. Rowell, Wood Sci, 1, 240 (1975).
[90] R.M. Rowell, Swedish For. Prod. Res. Lab., STF1 Series A, No. 772, 32
(1982).
[91] R.M. Rowell, S.V. Hart and GR. Esenther, Wood Sci., 11, 271 (1974).
[92] M. Sadachi, Mokuzai Hozon, No. 26. 47 (1985).
[93] R.M. Rowell, in Wood Modification, ed. M. Lawniczac, Polish Academy of
Science, Posnan, 358 (1985).
[94] R.M. Rowell, A.-M. Tilman and R. Simonson, J. Wood Chem. Technol, 6,
427 (1986).
[95] A.-M. Tilman, R. Simonson and R.M. Rowell, in Wood Modification, ed.
M. Lawniczak, Polish Academy of Science, Posnan, 436 (1985).
[96] M. Takahashi and Y. Imamura, Intern. Res. Group on Wood Preserv.,
Document No.: IRG/WP/3602 (1990).
[97] Y. Imamura and H. Kajita, Intern. Res. Group on Wood Preserv., Docu-
ment No.: IRG/WP/3622 (1990).
[98] K. Nishimoto, K. Tsunoda and Y. Imamura, Proceedings of IUFRO Div.
5 Conference, Brazil (1988).
[99] K. Tsunoda, Y. Imamura, M. Takahashi, S. Hirao, H. Usui and K.
Nomura, Proceedings of 40th Annual Meeting of Japan Wood Research
Society, Tsukuba, April 1990, 357 (1990).
[100] S. Hiaro, Mokuzai Kogyo, 44, 626 (1989).
Index
Abrasion resistance, 122-5, 147 methylation analysis of, 193, 197

and annual ring growth, 125 Anti-creep efficiency (ACE) factor,
Abrasive-wear coefficient, 122 142-3
Acacia mangium, 55-66 Anti-swelling efficiency (ASE) factor,
Acer mono Maxim., 122 141, 143
Acetic acid, 69 Aspergillus niger, 221
Acetic anhydride, 69, 72 Average alignment angle (AAA), 22-8
Acetylated particleboards, 81-3 Azaconazole, 84, 226
Acetylated RMP, 171, 172
Acetylated wood, 158, 174, 175
Acetylated wood meal, 170 Bending-creep test under fungal
Acetylation, 69-72, 149, 151, 157 attack, 76-8
Acetylcholine-esterase, 227 Benzothiazole compounds, 226
Acoustic properties, 143 Benzylated wood, 159
effect of chemical treatment, 144 Betula maximomcziana Regel., 122,
Adhesives, 162, 169 124
curing, 44-6 Biodegradation of wood, 136, 207-22,
Akaezomatsu (Picea glehnii), 144 224
Akamatsu (Pinus densiflora), 29, 31, Black spruce (Picea mariana Mill.),
72, 96, 98, 124 11, 17
Albizia falcata, 55-66 Bolted joints for glulam portal
Alkylammonium compounds, 227 frames, 112
Allylated wood, 156 BPMC, 227
Anhydro-cello-oligosaccharides, Brown-rot fungus
189-90 (Gloeophyllum trabeum), 231, 232
Anhydro-gluco-oligosaccharides, 194 (Lentinus lepideurs), 231
Anhydro-hetero-oligosaccharides, 198 (Poria placenta), 216
Anhydro-oligosaccharides (Serpula lacrymans), 214, 225
13
C-NMR data for, 200, 203 (Tyromyces palustris), 68, 76, 80-3,
'H-NMR data for, 199, 202 216, 220, 225, 232-3
methylation analysis of, 201, 204 (Tyromyces palustris (Berk, et Curt)
1,6-Anhydro-jS-D-glucopyranose, 189, Murr.), 76
190
l,6-Anhydro-/?-D-mannopyranose,
195, 198 Carbaryl, 227
Anhydrosugars Carbohydrate, 8
13
C-NMR data for, 192 conversion, 185-205
'H-NMR tlata for, 191, 196 Carbon fibers, 166, 179-80
241
242 INDEX
Carbon filaments, 162 energy density dependence of

Cellobiosan, 190 vaporized materials produced by,
4-/?-Cellobiosyl-4-/?-mannosyl-1,6- 188-9
anydromannose (G-G-M-AM), experimental procedure, 186
195 results and discussion, 187-99
Cellotriosan, 190 water-soluble materials formed by,
Cellulose, 69, 194 188-9
C 0 laser irradiation, 185-205
2
Composite beams, 55-66
degradation by brown-rot fungi, mechanical properties, 62-5
214-21 Compound middle lamella, 5
degradation by C 0 laser
2
Copper-chrome-arsenate (CCA),
irradiation, 187-93 224-5
distribution, 14 Corn starch, 199-201
effects of oxalic acid on viscosity, Creep strain, 136
220 Curdlan, 197, 201-3
microfibrils, 5-9, 14, 135 Cyanoethylated wood, 161, 163
morphological surface structure,
195
vaporized materials formed from, Dahurian larch (Larix gmelinii), 97,
198-9 98
Cercidiphyllum japonicum Sieb., 122 DDAC, 228
Chamaecyparis obtusa Endl., 121, 122, Dimensional stability, 136
124, 125, 129 conventional, 141
Chemically modified particleboards, Dimethyl phthalate, 156
67-74 Douglas fir (Pseudotsuga menziesii\
fungal attack, 81-3 96, 97
modulus of rupture (MOR), 68 Drift-pin joints for glulam portal
thickness swelling, 69 frames, 110-12
Chemically modified wood
enhancement of biological
resistance by, 230-4 EBIP, 225, 226
liquefaction of, 161-3 Electrostatic orientation, 27-31
Chemically treated woods, 135-54 Epoxy resins, 162
classification, 139^0 Esterification, 149-50, 156, 159
efficiency, 148-53 Etherification, 150-1, 156
mechanical properties, 145-7 Eucalyptus deglupta, 55-66
modulus of elasticity (MOE), Extrusion spinning, 180
146-53 Ezomatsu softwood (Picea jezoensis),
modulus of rupture (MOR), 147-53 97, 214
see also under specific properties
and types
Chlorination, 159, 163 Fagus crenata Blume., 122, 124, 125
Chlorpyrifos (CP), 226, 227 Fasteners, load (P)-slip (S)
C 0 laser irradiation of cellulose,
2
relationship, 112
185-205 Fenitrothion, 227
energy density dependence of the Fiber boards
proportion of water-soluble deep-drawable, 160
materials produced by, 190 moldable, 160
INDEX 243
Fiber-foring techniques, 169-83 Gmelina arborea, 55-66

Fibers, 162 Guaiacyl residues, 12, 14
Filaments, 162 Gypsum, board shear, 87
elongation rate and tensile strength
relationship, 172
from half-dissolved wood, 172 Hagen-Poiseulle's law, 182
from plasticized wood, 170-3 Hardwoods
Flammability, 136 cell types, 3
Flat squares bending tests, 97 chemical composition, 6
Floor-impact noise, 133 trimmings, 55
Formalization reaction, 148-9 Heat treatment, 148
Fractionation of modified wood Hem fir (Tsuga heteroplylla and
components, 162 Abies spp.), 97, 98
Frequency-response curve for wooden Hemicellulose, 5, 7, 14, 69
houses, 86-7 Hemin, 212
Frictional coefficients, 121 Hexamethylenetetramine. See
Fungal attack Hexamine
bending-creep test under, 76-8 Hexamine, 162, 173, 175, 177
chemically modified particleboards, Hinoki (Chamaecyparis obtsusa), 97
81-3 Holocellulose distribution, 15
glue-additive-treated particleboards, Houses. See Wooden houses
83 Humidity control, 125-7
particleboards, 78-81 Hydrogen bonding, 140
Fungal resistance of wood Hydrophilic effects, 140, 142
composites, 75-84 Hydrophobic acetyl groups, 69
Fungicides, 83 Hydrophobic effects, 140, 142, 149
Hygroscopicity, 120, 125-31, 136, 149
Galactoglucomannan, 16
Giant ipil-ipil (Leucaena leucocephala IF-1000, 225
Lam, de Wit), 72 IMF, 46
j9-l,3-Glucan, 194, 196, 201 Inorganic constituents of wood, 16
Glucomannan, 196-8 Intercellular layer, 5
degradation by C 0 laser
2 Internal friction, 143-5
irradiation, 193-5 IPBC, 225
4-/?-Glucosyl-1,6-anhydromannose Iron porphyrin complex, 211
(G-AM), 195 Isocyanates, 39, 44, 46, 80-3
4-/?-Glucosyl-4-/?-mannosyl-1,6-
anhydromannose (G-M-AM),
195 Japanese fir (Abies sachalinensis Fr.
Glucuronoarabinoxylan, 16 Schm.), 11
Glue-additive-treated particleboards,
83
Glue-line failures, 83 Karamatsu (Larix \epto\epis\ 96-97
Glulam portal frames, 109-18
deformation, 112
full-scale experiments, 115 Laminaribiosan, 197
nonlinear analysis, 112-15 Laminaridecaosan, 197
244 INDEX
Laminariheptaosan, 197 effects of particle configuration on

Laminarihexaosan, 197 properties, 37-8
Laminarinonaosan, 197 effects of particle density and
Laminarioctaosan, 197 compaction ratio on properties,
Laminaripentaosan, 197 34-6
Laminaritetraosan, 197 effects of resin content and resin
Laminaritriosan, 197 components on properties, 39-41
Laminated veneer lumber (LVL), internal bond strength, 34, 38, 40,
55-66, 233, 235 41, 52
bonding properties, 59 mechanical strength, 37
experimental procedure, 56-8 modulus of elasticity (MOE), 34
mechanical properties, 59 modulus of rupture (MOR), 34, 39
modulus of elasticity (MOE), 62-4 resin components in, 39
modulus of rigidity (MOR), 6 2 ^ semi-continuous steam-injection
Larch (Larix leptolepis G.), 72 pressing, 52
Laser beams, 185 shape factor, 38, 40
Latex, 163 thickness swelling, 36, 41
Lauan (Shorea spp), 29, 31, 37
Lauroylated wood meal, 156
Levoglucosan, 189, 194, 195, 197 Maleic acid and glycerol mixture, 72
Lignin, 5, 7, 8, 69 Mean microfibrillar angle, 144
concentration, 10 Mechanical orientation, 21-7
degradation, 209-14 Mechanical properties, chemically
distribution, 10, 14 treated woods, 145-7
formation, 12 Mechanical stress effects, 136
ratio of guaiacyl to syringyl Mechanosorptive creep, 136, 142, 150
residues, 12, 14 Medium-sensity fibreboard (MDF),
Lignin-carbohydrate complex, 8 73
Ligninase, 211 Methylene diphenyl diisocyanate
Lignocelulosic liquefaction, 161 (MDI), 39-40
Lignocellulosic materials, 137 Microfibrillar angle, 7
Linear expansion, 26 Middle lamella, 5, 14
Linear stability, 26 Mimosa-tannin adhesives, 72
Liquefaction Modulus of elasticity (MOE)
of acetylated wood, 175 chemically modified woods, 146-53
of chemically modified wood, chemically treated woods, 146-53
161-3 laminated veneer lumber (LVL),
of unmodified wood, 180-1 62-4
of untreated wood, 163-6 oriented strand and fiber boards,
of wood, 160-1 23-5, 31
Load-deflection curves Modulus of rigidity (MOR),
plywood-sheathed diaphragms, 105 laminated veneer lumber (LVL),
plywood-sheathed walls, 106 62-4
Loblolly pine (Pinus taeda L.), 11 Modulus of rupture (MOR)
Low-density particleboard, 33-41 chemically modified woods, 147-53
bending strength, 37-8 chemically treated woods, 147-53
crack dimensions, 38 oriented strand and fiber boards,
deflection-time curves, 84 23-5, 31
INDEX 245
Moisture Paulownia tomentosa, 120

condensation, 127-30 Paulownia tomentosa Steud., 122
expansion induced by, 141 PCA, 174, 179
humidity effects, 136 Phanerochaete chrysosporium, 211
permeability, 127 Phenol, liquefaction of unmodified
sorption, 142 wood into, 180-1
transmission coefficient, 129 Phenol-formaldehyde resins, 44, 46,
see also Hygroscopicity 68-9, 78-83, 151-2, 162, 234-5
Moldings, 165 Phenol-melamine formaldehyde
Moment-resisting joints, 109-18 (PMF), 44-6
Phenol resins
adhesives, 164
Nailed joints for glulam portal frame, see also Wood-phenol resin fiber
109-10 Phoxim, 227
Natural frequency of wooden houses, Plasticization of wood, 155-67
87, 88 application categories, 169
Novolak-resin type moldings, 165 filaments from, 170-3
Plywood panels, structural analysis,
101-8
Oligosaccharides, 197 Plywood-sheated diaphragms, 103
Organoiodine compounds, 225 load-deflection curves, 105
Oriented fiber board, 21-32 Plywood-sheated floors, 103
Oriented strand board, 21-32 Plywood-sheated walls, 102
Oxalic acid, effects on viscosity of load-deflection relationship, 106
cellulose, 220 Polyethylene glycol (PEG), 153, 165
Oxaloacetase, 220 Polymethylmethacrylate (PMMA),
Oxaloacetic acid, 220 156
Polyphenylene oxide, 159
Polysaccharides, 15, 148
Paraffin impregnation, 153 degradation by C 0 laser
2
Particle mat, temperature behavior irradiation, 195-7

during steam-injection pressing, Polystyrene, 159
46-9 Polyurethane resins, 162
Particleboards Portal frames, 109-18
acetylated, 81-3 mechanical fasteners, 109-12
acetylated sugi, 69 Propoxur, 227
chemically modified. See Propylene oxide, 72, 150
Chemically modified Pseudo-dynamic testing system, 106
particleboards Pyrethroids, 227
deflection-time curves, 78-9, 82 Pyridaphenthion, 227
fungal attack, 78-81
glue-additive-treated, 83
mechanical performance, 76 Quercus mongolica Fisch. var., 122,
production, 33 124
strength reduction due to decay, Quercus Myr., 120
76
see also Low-density particleboard;
Steam-injection pressing Radiatea pine, 170, 173
246 INDEX
Refiner mechanical pulp (RMP), 170, Sugi (Cryptomeria japonica), 15, 16,
173 97
Relative humidity, 126, 128 Sugi (Cryptomeria japonica D. Don),
Resin components in low-density 68, 123, 124
particleboard, 39 Surface characteristics
abrasion resistance, 122-5
smoothness to the touch, 120-1
Sawn timber warmth to the touch, 120
correlation coefficient between knot Syringyl residues, 12, 14
size and modulus of rupture
(MOR), 98
modulus of elasticity (MOE), 97 TCMTB, 226
structural properties, 95-9 Termites
Seismic response analysis, 104-7 (Coptotermes formosanus), 68, 72,
Semi-strand board, 30 235, 236
Shorea negrosensis Foxw., 126 (Reticulitermes flavipes), 232
Softwoods (Reticulitermes speratus), 12
cell types, 2 Termiticides, 226-7
cell wall organization, 4-8 Tetrachlorovinphos, 227
chemical composition, 6 Tetraoxane, 73
surface characteristics, 121 Tetraphenyl porphyrin (TPP), 212
Sound absorption, 143 Thermal gradient, 127-30
Sound-insulating material, 133 Thermal properties, 125
Sound transmission, 131-3 Thermoplasticity, 136, 156-60
in wooden houses, 120 TIAA, 225
Sound tansmission loss (STL), 133 Timber joints, 109
Sound velocity, 143 TnBTO, 224
Spinning methods, 173, 176-7 Todomatsu (Abies sachalinensis), 97
unmodified wood solution, 181-2 Touhi (Picea sp.), 97
Spinning nozzle, 171, 174 TPIC, 227
Spinning solutions, 170, 173-6, 181 Triazole compounds, 226
Spinning temperature, 179 Trifluoroacetic acid (TFA), 158
Spruce (Picea jezoencis Carr.), 81 Trifluoroacetic anhydride (TFAA),
Squares bending test, 96 170, 173, 174, 179
Squares compression test, 97 Trioxane, 73
Squares tension test, 97
Starch, 194, 196 Urea formaldehyde (UF), 44-6
Steam-injection pressing, 43-53 Urea-melamine-formaldehyde (UMF)
continuous press development, resin, 44, 45, 78-83
49-52
optimizing press time, 49
temperature behavior of particle Valonia macrophysa, 8-9
mat, 46-9 Vapor-phase treatment, 149
Strength ratio concept, 96 Vibration equation, 104
Stress-induced deformation, 142 Vibration generator, 104
Structural timber. See Sawn timber frequency-response curves from, 86
Styrene-maleic anhydride copolymer Vibration tests on wooden buildings,
(SMA), 159 104
INDEX 247
Vinyl composites, 234 Wood-plastic composites (WPC),

Vinyl polymers, 157 152-3, 234
Viscoelasticity evaluation, 143 Wood preservatives, 223-40
Volumetric swelling, 141 chemical modification, 230-4
development of alternatives, 224-8
oil-borne, 225-7
Water sorption, 137 pressure treatment, 228-9
Water vapor pressure, 127-8, 130 superficial treatment, 229-30
White birch (Betula papyrifera water-borne, 227-8
Marsh.), 5, 11-14 Wood stability, characterization of,
White-rot fungus 141-7
(Coriolus versicolor), 68, 81, 225, Wood structure, 1
233 Wood-water interaction, 137-9
(Coriolus versicolor (L. ex Fr.) Wooden houses
Quel), 76 damage from full-scale tests, 91-2
(Phanerochaete chrysosporium), earthquake damage, 91-2
209-14 floor openings, 90
WIC (wood-inorganic material frequency-response curve, 86-7
composite), 152, 235-6 full-scale tests on, 85-93
Wood interior material, 119-34
chemical composition, 1-20 location of load-bearing walls, 90
inorganic constituents, 16 natural frequency, 87, 88
Wood cells, 2-3 non-load-bearing walls, 87-8
distribution of chemical number of stories, 91
constituents in cell wall, 9-16 shear walls, 87-8
major components, 5 sound transmission in, 120
model of, 137 static stiffness, 88
wall organization, 4-8, 223 surface properties, 119
Wood composites, fungal resistance vibration tests, 104
of, 75-84 Wooden-sheated diaphragms,
Wood decay, 208 hysteresis of, 107
Wood-inorganic material composite
(WIC), 152, 235-6
Wood-methyl methacrylate (MMA) Young's modulus, 143-5
composite, 123
Wood-phenol resin fiber, 173-82
from unmodified wood, 180-2 Zelkowa serrata Makino, 122

Nobuo Shiraishi, Hiromu Kajita and Misato Norimoto (Eds.) - Recent Research On Wood and Wood-Based Materials. Current Japanese Materials Research (1993)

Uploaded by

Document Information

Original Title

Copyright

Available Formats

Share this document

Share or Embed Document

Sharing Options

Did you find this document useful?

Is this content inappropriate?

Copyright:

Available Formats

Nobuo Shiraishi, Hiromu Kajita and Misato Norimoto (Eds.) - Recent Research On Wood and Wood-Based Materials. Current Japanese Materials Research (1993)

Uploaded by

Copyright:

Available Formats

CURRENT JAPANESE MATERIALS RESEARCH

R. KOTERAZAWA, R. EBARA and S. NISHIDA

T MISE, K. NISHIDA, M. KAMON and M. MASHIMA

Current Japanese Materials ResearchVol 11

ELSEVIER APPLIED SCIENCE

WITH 33 TABLES A N D 120 I L L U S T R A T I O N S

1993 ELSEVIER SCIENCE PUBLISHERS L T D and T H E SOCIETY O F

British Library Cataloguing in Publication Data

Library of Congress Cataloging-in-Publication Data

Special regulations for readers in the USA

Photoset in Malta by Interprint Limited

The Current Japanese Materials Research (CJMR) series is edited by

Wood is composed of 50-55% cellulose, 15-25% hemicellulose and

RAZALI ABDUL KADER

Wood Research Institute, Kyoto University, Uji, Kyoto 611, Japan

Wood Research Institute, Kyoto University, Uji, Kyoto 611, Japan

Department of Wood Science and Technology,

Keywords: wood structure, chemical composition, cellulose, hemicellulose, lignin

Wood is an easily renewable natural resource with adequate hus-

As illustrated in Fig. 1, wood has several different types of cell exist-

Tracheid Ray tracheid

FIG. 2. Scanning electron micrograph of a typical softwood (Pinus densiflora),

Vessel element Ray parenchyma cell

CELL WALL ORGANIZATION

The cell wall organization of typical softwood tracheids or hard-

ferences in the microfibrillar orientation. The Si layer is a flat helix

S wall and comprises a steep Z-helix with a high degree of parallelism

is loosely textured [ 5 ] . Although the S layer exhibits steep helices in

the microfibrillar orientation, there exist transition lamellae on its inner

between S and S , and between S and S , respectively [3].

[6]. However, the change in microfibrillar angle is more abrupt in S 3 2

FIG. 4. Transmission electron micrograph of an ultra-thin transverse section

Such wood cells are composed of three major components, cellulose,

whereas S and S remain fairly constant. Because of its greater thickness,

Cellulose 40-50 45-50

the S layer is largely responsible for the physical and mechanical

properties of the cell walls. Therefore, it is of great interest to know the

cell wall properties.

generally lower than those in the earlywood. In contrast, compression

literature suggests that the microfibrillar angles increase in the order:

Tension wood Latewood Earlywood Compression wood

Interestingly, this trend is parallel to the trend in lignin concentration; the

compared to normal softwood [13, 14]. The tracheid S layer in latewood

Figure 5 shows the relationship for softwoods between the lignin

from the degree of spreading in the equatorial diffraction arcs of X-ray

tracheid S layer [21].

and the lignin content of wood [21].

The biosynthetic origin of this relationship is not known. However, it

FIG. 6. Schematic diagram of the

At the molecular level of arrangement of the chemical composition, the

STRUCTURE OF THE CELLULOSE MICROFIBRILS

On cellulose microfibrils themselves, many investigations have been

FIG. 7. Transmission electron micrographs of a cross-sectional view of the

DISTRIBUTION OF CHEMICAL CONSTITUENTS IN THE

Three major cell wall components, cellulose, hemicellulose and lignin,

techniques of UV microscopy and EDXA with bromination [44] across the

entities, such resolution often being difficult with U V microscopy. It is

X-ray in Fig. 9 shows such differences clearly. Fukazawa and Imagawa

Wood Morphological Tissue Lignin Lignin

FIG. 9. Scanning electron micrograph (a) of brominated latewood tracheids in

dominantly syringyl residues, whereas the vessel secondary wall (S ) con-2

be guaiacyl residue, with the remaining 0 to 20% being syringyl residue. It

compounds to magnolia shoots and determined their location in the