Experimental Measurements of

Thermal Conductivity of Wood

Species in China: Effects of Density,
Temperature, and Moisture Content

Downloaded from http://meridian.allenpress.com/fpj/article-pdf/61/2/130/1699045/0015-7473-61_2_130.pdf by guest on 29 March 2024

Zi-Tao Yu Xu Xu Li-Wu Fan
Ya-Cai Hu Ke-Fa Cen

Abstract
Experimental measurements of thermal conductivity of wood were performed using the heat flow meter and transient
plane source technique. The specimens were prepared from five species of both softwoods and hardwoods widely available
and used in China, with a wide range of density and moisture content. The transverse thermal conductivity of ovendry
specimens is presented as a function of density and temperature up to 908C and is compared with that along the grain
direction for two select species. The influence of moisture content up to 23 percent, which is below the typical fiber
saturation point of wood, on the transverse thermal conductivity is presented as well. It is shown that the transverse thermal
conductivity of wood increases with density, temperature, and moisture content. Linear correlating equations are proposed in
terms of these factors.

W ood, a natural organic composite material that

consists of cellulosic fibers and lignin, has a long history of
measured thermal conductivity of both softwood and
hardwood species in Japan under a heating process up to
use both as a solid fuel and as a construction material. A 2708C. Soon thereafter, Suleiman et al. (1999) measured
need exists, however, to know thermal properties of wood to thermal conductivity of the hardwood birch (Betula
understand and model heat transfer processes in wood and alnoides) grown in Sweden at 218C and 1008C for both
wood-based materials. For example, the energy design and longitudinal and transverse directions. Samples obtained
evaluation of energy performance of wood-frame buildings from the softwoods white pine (Pinus strobus) and red pine
partially rely on the thermal properties of wood and wood (Pinus resinosa) grown in the United States were measured
products (TenWolde et al. 1988). The analysis of combus- by Rice and Shepard (2004) for two different moisture
tion and pyrolysis of wood exposed to fire also demands the content levels. Ngohe-Ekam et al. (2006) reported their
knowledge of thermal properties (Beall 1977, Ragland et al. experimental measurements on thermal conductivity of
1991, Thunman and Leckner 2002). Thermal conductivity
that represents quantitatively the ability of wood to conduct
heat is of great significance in heat transfer modeling (Olek The authors are, respectively, Associate Professor, Inst. of
et al. 2003). Thermal Sci. and Power Systems, Dept. of Energy Engineering,
Zhejiang Univ., Hangzhou, People’s Republic of China (yuzitao@
Measurement of thermal conductivity of wood dates back zju.edu.cn); Associate Professor, College of Metrological and
several decades (see, e.g., Wangaard 1940, MacLean 1941). Measurement Engineering, China Jiliang Univ., Hangzhou, People’s
The early experimental work was primarily performed using Republic of China ([email protected]); Research Assistant, Dept. of
the steady-state, guarded hot plate method. Thereafter, Mechanical Engineering, Auburn Univ., Auburn, Alabama
transient techniques, such as the laser flash method (Harada ([email protected]); Professor, Inst. of Thermal Sci.
et al. 1998), transient plane source technique (Suleiman et and Power Systems, Dept. of Energy Engineering, Zhejiang Univ.,
al. 1999), and transient hot wire method (Kol 2009), were Hangzhou, People’s Republic of China ([email protected]); and
Professor, State Key Lab. of Clean Energy Utilization, Zhejiang
developed. The main factors that significantly affect thermal
Univ., Hangzhou, People’s Republic of China ([email protected]).
conductivity of wood include density, temperature, and This paper was received for publication in December 2010. Article
moisture content. Steinhagen (1977) and TenWolde et al. no. 10-00075.
(1988) collected extensive data on thermal conductivity of ÓForest Products Society 2011.
wood from early experimental studies. Harada et al. (1998) Forest Prod. J. 61(2):130–135.

130 YU ET AL.
tropical wood as a function of basal density. Recently, Kol Furthermore, for each of the two select species (i.e., the
(2009) measured transverse thermal conductivity of five softwood masson pine [Pinus massoniana] and the hard-
hardwood species grown in Turkey with different moisture wood red lauan), three ovendry specimens along the grain
contents. direction were prepared so that both longitudinal and
A literature survey, however, reveals a lack of experi- transverse thermal conductivities could be measured. For
mental data on thermal conductivity of wood species grown two other select species (i.e., the softwood larch [Larix
in China and Southeast Asia. To extend the existing gmelinii] and the hardwood basswood [Tilia yunnanensis]),
knowledge, this study measured and documented thermal specimens with different moisture contents up to approxi-
conductivity of wood species that are widely available and mately 23 percent, which is below the typical fiber
used in China and investigated its variations with density, saturation point of wood, were prepared in such a way that
temperature, and moisture content. the samples were conditioned at different relative humidity
levels until an equilibrium moisture content had been
Experiments reached (Kol 2009). This wetting process usually took 3 to 7
days, depending on the microstructures of the wood species.
Preparation of specimens Both drying and wetting processes were monitored inter-
Five species of both softwoods and hardwoods were mittently by weighing the samples being treated. At each

Downloaded from http://meridian.allenpress.com/fpj/article-pdf/61/2/130/1699045/0015-7473-61_2_130.pdf by guest on 29 March 2024

selected for a total of 10 wood species studied, as listed in moisture level, only one specimen was prepared for both
Table 1. The origins of these wood species, all of which are species. To preserve any moisture from undesirable gain or
widely available and used in China as construction and loss, all of the treated specimens were stored in sealed glass
jars until measuring.
decoration materials, are also given. Except for the three
hardwood lauan species (red lauan [Shorea negrosensis], Heat flow meter technique
yellow lauan [Shorea kalunti], and white lauan [Pentacme
contorta]), all studied species were originally grown in In this study, thermal conductivity of ovendry specimens
China, covering an extremely broad area from the was measured by using the heat flow meter technique, which
uses two calibrated heat flux transducers to measure the heat
northernmost province (Heilongjiang) to southern provinces
flux passing through the specimen that is placed between
(Zhejiang, Fujian, and Yunan). Although lauan species have
them. A steady-state, unidirectional heat flux is maintained
been planted in southern China for years, the available logs by applying a given thermal gradient across the specimen.
in the Chinese market are mainly imported from Malaysia As given by Fourier’s law (Siau 1984), thermal conductivity
and the Philippines. of the specimen is evaluated as
Boards were sawn from portions of logs that were free of
evident cracks, checks, and knots. Specimens were then q 00 H
prepared by planing the board surfaces and sawing into flat k¼ ð1Þ
DT
slabs with nominal dimensions of 100 by 100 by 10 mm3.
where k is the thermal conductivity (W/mK), q 00 is the heat
For each wood species selected, three ovendry specimens flux measured (W/m2), H is the height of the specimen (m),
were prepared to measure the transverse thermal conduc- and DT is the temperature difference (K).
tivity. As described by Siau (1984), ovendry samples were An HFM 436/3 Lambda system (NETZSCH, Germany)
obtained using a drying temperature of 1058C for a based on the heat flow meter technique, which is designed
sufficiently long time until a constant mass was attained. for poorly conducting insulation materials that possess a low
After heating at such temperature, both the free and bound thermal conductivity up to 0.5 W/mK, was used. This
water in wood was released, and the specimens were instrument allows temperature-dependent measurements
completely dry. The actual dimensions and mass of the between 08C and 1008C, and the typical repeatability is
samples were then measured precisely by using digital less than 5 percent. In addition, it accepts rectangular
calipers (precision, 60.1 mm) and a digital scale (precision, specimens with a maximum size of 300 by 300 by 100 mm3.
60.1 g), respectively. Consequently, apparent ovendry Because of the use of two heat flux transducers, the
densities of the specimens were readily calculated, as given measurement times are typically within 20 minutes,
in Table 1. providing a rapid means of measuring thermal conductivity.

Table 1.—Wood samples measured.

Species Origin Ovendry density (kg/m3)
Softwoods (conifers)
China-fir (Cunninghamia lanceolata) Fujian, China 308.3
Korean pine (Pinus koraiensis) Heilongjiang, China 378.6
Masson pine (Pinus massoniana) Zhejiang, China 415.0
Scots pine (Pinus sylvestris var. sylvestriformis) Heilongjiang, China 416.4
Larch (Larix gmelinii) Heilongjiang, China 672.8
Hardwoods (angiosperms)
Red lauan (Shorea negrosensis) Imported from Malaysia 345.8
Yellow lauan (Shorea kalunti) Imported from Malaysia 401.7
White lauan (Pentacme contorta) Imported from Malaysia 436.9
Basswood (Tilia yunnanensis) Yunnan, China 456.0
Birch (Betula alnoides) Yunnan, China 613.8

FOREST PRODUCTS JOURNAL Vol. 61, No. 2 131

Transient plane source technique Figure 1. Note that the data were measured at 308C, which is
Although the heat flow meter technique is somewhat higher than room temperature (;218C). In Figure 1, the data
faster than the conventional guided hot plate method, it is points for ovendry specimens at room temperature,
still a steady-state method, and the typical measurement presented by TenWolde et al. (1988), and two empirical
times of 20 minutes cannot be shortened. Therefore, this equations, proposed by TenWolde et al. (1988) and Harada
technique is not suitable for measuring moist wood et al. (1998), are given for comparison.
specimens. When a thermal gradient is applied to a moist It is shown that thermal conductivity of wood increases
sample, a redistribution of the moisture takes place, leading with density, which is consistent with the relationship found
to a transient heat flow that in turn confounds the by many other researchers. This is obvious in that for a
measurement (TenWolde et al. 1988). In view of this, the given volume, as the density of wood increases, more fibril
recently developed transient techniques, which offer rapid exists that is more conductive than air. Considering the
measurements within several minutes or even seconds, effective media theory of composite materials, the effective
appear desirable for measuring thermal conductivity of thermal conductivity is therefore increased. A generic linear
moist wood samples, because the moisture diffusion during increasing tendency is observed for both softwoods and
the short heating period would be negligible. hardwoods; hence, there is no need to distinguish a wood
In the present study, the measurements for moist species. For densities lower than 500 kg/m3, the measured

Downloaded from http://meridian.allenpress.com/fpj/article-pdf/61/2/130/1699045/0015-7473-61_2_130.pdf by guest on 29 March 2024

specimens were performed on a TPS 500 Thermal Constants data agree well with those reported in the literature, whereas
Analyzer (Hot Disk AB, Sweden) that is based on the for the two data points at higher densities, the measured data
transient plane source technique. The principle and data are smaller. However, the data presented by TenWolde et al.
reduction of this technique were summarized by Suleiman et (1988) are approximate and may vary from the actual values
al. (1999) and are not herein duplicated. This instrument by as much as 20 percent.
offers a simultaneous measurement of both thermal The correlating equation, which has a coefficient of
conductivity and diffusivity, although only the thermal determination of 0.853, for the measured data is
conductivity data are presented. The applicable thermal
k ¼ 0:04409 þ 0:0001278q
conductivity range is between 0.03 and 100 W/mK, and the
reproducibility is typically less than 2 percent. The for 300 kg=m3 , q , 700 kg=m3 ð2Þ
measurement time is between 2.5 and 640 seconds. In where q is the ovendry density (kg/m3). The three equations
addition, a controllable electric furnace was used to achieve presented in Figure 1 may be interchangeable with
measurement temperatures greater than room temperature. reasonable accuracy when the density of interest is
Once a desirable temperature had been attained, the moist relatively low, although the present fitting curve has a
samples were inserted into the furnace chamber, and the smaller slope than those of the others.
measurements were finished rapidly.
Temperature dependence of ovendry wood
Experimental details
In the literature, few studies have considered the
As mentioned, the ovendry specimens were measured temperature dependence of thermal conductivity of wood.
using the heat flow meter technique, and the moist For example, Steinhagen (1977) presented the variations of
specimens were measured using the transient plane source conductivity data for temperatures between 408C and
technique. The accuracy and reproducibility of the instru- 1008C in the form of piecewise linear curves, but no data
ments were assured by testing with stainless steel standard. points were given. Harada et al. (1998) measured thermal
For each ovendry specimen, five measurements were conductivity of some Japanese wood species up to 2708C,
performed at each of the temperatures of interest, and the but the variations below 1008C were not clearly revealed.
average thermal conductivity was obtained with a standard Furthermore, Suleiman et al. (1999) measured birch samples
deviation of less than 1 percent. The typical relative
deviation among the three specimens for each of the select
species was less than 3 percent. Furthermore, for the moist
specimens, three measurements were conducted with a
standard deviation of less than 5 percent. Although the
variation of moisture content was not monitored during the
measurements, this relatively small standard deviation
indicates that the influence of the moisture redistribution
was nearly negligible, because the three measurements were
completed in a short period of several minutes. In the next
section, average thermal conductivity at each of the data
points will be presented without noting the local uncertainty.
The ovendry density presented in Table 1 is also the average
value among the three specimens for each species. The
standard deviation for ovendry density is typically less than
3 percent.

Results and Discussion

Density dependence of ovendry wood
The transverse thermal conductivity of ovendry speci- Figure 1.—Transverse thermal conductivity of wood at 30 8C as
mens as a function of the ovendry density is presented in a function of ovendry density.

132 YU ET AL.
grown in Sweden at both 218C and 1008C, but only a single For the five hardwood species, the temperature dependence
temperature point other than room temperature was of transverse thermal conductivity is presented in Figure 3.
reported. Therefore, to investigate the temperature depen- Similar to the softwoods, the almost-linear increasing trend is
dence of thermal conductivity of wood below 1008C, observed for hardwoods between 308C and 808C. The linear
measurements were conducted in the present study at fitting equations, all of which possess a coefficient of
temperatures between 308C and 908C, in 108C increments. determination greater than 0.97, are expressed as
Note that this temperature range is out of the normal in-
k ¼ 0:08283 þ 0:0002437T for red lauan ð8Þ
service temperature of wood and may reflect an early stage
toward fire or pyrolysis. k ¼ 0:078 þ 0:0002532T for yellow lauan ð9Þ
In Figure 2, variations of transverse thermal conductivity
as a function of temperature are compared for the five k ¼ 0:1044 þ 0:0002859T for white lauan ð10Þ
softwood species. It is shown that the thermal conductivity
is proportional to temperature. For each wood species, the k ¼ 0:08635 þ 0:0003695T for basswood ð11Þ
increase between 308C and 808C is almost linear. However, k ¼ 0:1114 þ 0:00037T for birch ð12Þ
for Scots pine (Pinus sylvestris var. sylvestriformis), masson
pine, and Korean pine (Pinus koraiensis) specimens, the Except for birch, which has the greatest density among

Downloaded from http://meridian.allenpress.com/fpj/article-pdf/61/2/130/1699045/0015-7473-61_2_130.pdf by guest on 29 March 2024

thermal conductivity at 908C is nearly identical to the the hardwoods tested, the thermal conductivity becomes
corresponding value at 808C, although the linear increasing nearly unvaried when the temperature is further increased to
trend remains for the other two species (China-fir [Cunning- 908C. The greatest relative increase at 808C is 18.5 percent
hamia lanceolata] and larch). The most relative increase at for basswood, whereas the smallest relative increase is 11.7
808C is 19.8 percent for Korean pine, whereas the smallest percent for white lauan. In summary, for both softwoods and
increase is 10.1% for China-fir. hardwoods, the temperature dependence is nearly linear
As shown in Figure 2, the linear fitting curve for China-fir between 308C and 808C. Over this 508C temperature
is less steep than the other curves, which are nearly parallel difference, the relative increase of transverse thermal
to one another. The correlating equations, all of which conductivity is between 10 and 20 percent, which is greater
than the 10 percent increase for every 508C indicated by
exhibit a coefficient of determination greater than 0.97, are
Steinhagen (1977).
as follows:
The anisotropy of thermal conductivity of wood is
k ¼ 0:07538 þ 0:0001681T for China-fir ð3Þ presented for the two select species (the softwood masson
pine and the hardwood red lauan) in Figure 4. It is shown
k ¼ 0:08124 þ 0:0003695T for Korean pine ð4Þ that at constant temperatures, the longitudinal thermal
conductivities for both species are nearly identical, although
k ¼ 0:08653 þ 0:0003521T for masson pine ð5Þ more significant differences are present for the transverse
direction. In addition, the temperature dependence of
k ¼ 0:08805 þ 0:0003888T for Scots pine ð6Þ longitudinal thermal conductivity is much weaker than that
of transverse conductivity; that is, the longitudinal thermal
k ¼ 0:1185 þ 0:0003417T for larch ð7Þ conductivity is nearly independent of temperature, espe-
cially for red lauan (hardwood). Therefore, the ratio of
where T is temperature (8C). These equations are valid for longitudinal to transverse thermal conductivity becomes
temperature between 308C and 808C and may be extended to smaller with increasing temperature. The average ratios over
room temperature and up to 1008C with reasonable different temperatures are found to be 1.71 and 1.87 for
accuracy. masson pine (softwood) and red lauan (hardwood),

Figure 2.—Transverse thermal conductivity of softwood spe- Figure 3.—Transverse thermal conductivity of hardwood
cies as a function of temperature. species as a function of temperature.

FOREST PRODUCTS JOURNAL Vol. 61, No. 2 133

 Because of the presence of more water, which is more
conductive than fibril and air, the thermal conductivity
increases proportionally as the moisture level is increased.
At constant moisture content levels, thermal conductivity
increases nonlinearly with temperature. Although the
nonlinearity is insignificant at relatively lower moisture
content levels, the fitting curves are apparently in the form
of a quadratic equation when the moisture content is high.
On the other hand, five different moisture content levels
(8.7% [air dry], 12.6%, 15.8%, 19.7%, and 22.2%) were
studied for basswood specimens. As presented in Figure 6,
the increase of thermal conductivity with temperature
becomes generally linear for this hardwood. It is noted that
for two of the intermediate moisture content levels (12.6%
and 15.8%), the fitting curves are slightly nonlinear and are
better expressed as a quadratic equation with a negative

Downloaded from http://meridian.allenpress.com/fpj/article-pdf/61/2/130/1699045/0015-7473-61_2_130.pdf by guest on 29 March 2024

quadratic coefficient. At similar moisture content levels, the
difference of the increasing trends between softwood and
Figure 4.—Comparison of thermal conductivity of wood hardwood species might be attributed to microstructures/
between transverse and longitudinal directions. anatomy of wood, which in turn leads to different
distributions of water in wood.
In Figure 7, the data presented in Figures 5 and 6 are
respectively. These values are consistent with the range of collected in such a way that the variations of thermal
1.5 to 2.8 reported in the literature (TenWolde et al. 1988). conductivity with moisture content are shown explicitly. At
constant temperatures, thermal conductivity increases line-
Combined effects of moisture content arly with moisture content. Two empirical equations for
and temperature moist wood at room temperature are shown in Figure 7. The
As an example for softwood species, transverse thermal equation proposed by MacLean (1941) is
conductivities of larch specimens at different moisture
k ¼ 0:02376 þ ð0:0002001 þ 0:000004031MCÞq
content levels (6.7% [air dry], 9.7%, 12.6%, 19.1%, and
22.9%) are presented in Figure 5. All of these moisture for MC , 40% ð13Þ
content levels are below the nominal fiber saturation point where MC is moisture content, and the equation proposed
of wood (;30%), which is defined as the moisture content by TenWolde et al. (1988) is
that corresponds with pronounced changes in the mechan-
ical and physical properties of wood (Siau 1984). In other k ¼ 0:01864 þ ð0:0001941 þ 0:000004064MCÞq
words, the thermal properties of wood begin to change as a
for MC , 25% ð14Þ
function of moisture content when it is below the fiber
saturation point (TenWolde et al. 1988). Also note that the Note that Equations 13 and 14 are in the form of a binary
moisture condition of wood in service is normally below the equation that relates thermal conductivity of wood to its
fiber saturation point. Therefore, thermal conductivity of ovendry density and moisture content. For the wood
wood above the fiber saturation point is of little practical specimens studied, the original binary equations have been
interest. reduced by substituting the corresponding known ovendry

Figure 5.—Transverse thermal conductivity of larch at different Figure 6.—Transverse thermal conductivity of basswood at
moisture content levels. different moisture content levels.

134 YU ET AL.
 softwood and hardwood species between 308C and 808C.
For the two select species (masson pine and red lauan), the
ratios of longitudinal to transverse thermal conductivity
were consistent with the range given in the literature.
Finally, transverse thermal conductivity of two other select
species (larch and basswood) at different moisture content
levels up to approximately 23 percent was presented in
different ways. At each moisture content level studied, a
quadratic increase of thermal conductivity with temperature
was observed for the softwood species (larch), whereas the
trend for the hardwood species (basswood) was generally
linear. Furthermore, at constant temperatures, thermal
conductivity was shown to increase linearly with moisture
content, and a correlating equation was proposed for each of
the species tested.
Acknowledgments

Downloaded from http://meridian.allenpress.com/fpj/article-pdf/61/2/130/1699045/0015-7473-61_2_130.pdf by guest on 29 March 2024

Figure 7.—Transverse thermal conductivity of wood as a The authors thank Ms. Li-Ping Shi and Mr. Liang Xu at
function of moisture content at different temperatures. NETZSCH Scientific Instruments Trading (Shanghai) Co.
Ltd. for their assistance in conducting the experiments.

densities. The linear fitting curves for the present data at Literature Cited
308C, which is close to room temperature, are depicted as Beall, F. C. 1977. Properties of wood during carbonization under fire
well for comparison. The equation, which is free of density, conditions. In: Wood Technology: Chemical Aspects. I. S. Goldstein
(Ed.). American Chemical Society, Washington, D.C. ACS Sympo-
for larch is sium Series. Vol. 43. Chap. 7, pp. 107–114.
Harada, T., T. Hata, and S. Ishihara. 1998. Thermal constants of wood
k ¼ 0:1425 þ 0:094MC for MC , 25% ð15Þ during the heating process measured with the laser flash method. J.
with a coefficient of determination of 0.972. In this case, the Wood Sci. 44:425–431.
Kol, H. Sx . 2009. The transverse thermal conductivity coefficients of some
slopes of the two empirical equations proposed in the
hardwood species grown in Turkey. Forest Prod. J. 59:58–63.
literature are greater than the slope in Equation (16). MacLean, J. D. 1941. Thermal conductivity of wood. Heat. Piping Air
Concurrently, the linear regression equation for basswood is Cond. 13:380–391.
Ngohe-Ekam, P. S., P. Meukam, G. Menguy, and P. Girard. 2006.
k ¼ 0:0731 þ 0:2497MC for MC , 25% ð16Þ Thermophysical characterization of tropical wood used as building
materials: With respect to the basal density. Constr. Build. Mater. 20:
with a coefficient of determination of 0.996. The slope in 929–938.
this equation becomes slightly greater than those in Olek, W., J. Weres, and R. Guzenda. 2003. Effects of thermal
Equations 14 and 15. In general, the two empirical equations conductivity data on accuracy of modeling heat transfer in wood.
predict thermal conductivity with acceptable accuracy for a Holzforschung 57:317–325.
wide range of density and moisture content. However, for Ragland, K. W., D. J. Aerts, and A. J. Baker. 1991. Properties of wood
the specific wood species studied, the correlating equations for combustion analysis. Bioresour. Technol. 37:161–168.
proposed by the present study would lead to more accurate Rice, R. W. and R. Shepard. 2004. The thermal conductivity of plantation
grown white pine (Pinus strobus) and red pine (Pinus resinosa) at two
predictions. moisture content levels. Forest Prod. J. 54:92–94.
Siau, J. F. 1984. Transport Processes in Wood. Springer-Verlag, Berlin.
Conclusions 245 pp.
Thermal conductivity of wood was measured experimen- Steinhagen, H. P. 1977. Thermal conductive properties of wood, green or
tally by using the heat flow meter and transient plane source dry, from 408 to þ1008C: A literature review. FPL-9. USDA Forest
Service, Forest Products Laboratory, Madison, Wisconsin. 10 pp.
technique for ovendry and moist specimens, respectively. Suleiman, B. M., J. Larfeldt, B. Leckner, and M. Gustavsson. 1999.
The specimens included five species for both softwoods Thermal conductivity and diffusivity of wood. Wood Sci. Technol. 33:
(China-fir, Korean pine, masson pine, Scots pine, and larch) 465–473.
and hardwoods (basswood, birch, and red, yellow, and white TenWolde, A., J. D. McNatt, and L. Krahn. 1988. Thermal properties of
lauan), which are widely available and used in China. wood and wood panel products for use in buildings. DOE/USDA-
First, transverse thermal conductivity was presented as a 21697/1. U.S. Department of Energy, Oak Ridge National Laboratory,
function of ovendry density for ovendry specimens, and a Oak Ridge, Tennessee. 40 pp.
Thunman, H. and B. Leckner. 2002. Thermal conductivity of wood—
linear correlating equation was obtained. Second, temper- Models for different stages of combustion. Biomass Bioenergy 23:
ature dependence of transverse thermal conductivity of the 47–54.
ovendry specimens was presented up to 908C. A nearly Wangaard, F. F. 1940. Transverse thermal conductivity of wood. Heat.
linear increasing trend was observed for each of the Piping Air Cond. 12:459–464.

FOREST PRODUCTS JOURNAL Vol. 61, No. 2 135