Bioresource Technology 37 (1991) 161-168
Properties of Wood for Combustion Analysis
K. W. Ragland, D. J. Aerts
Department of Mechanical Engineering, University of Wisconsin-Madison, Madison, Wisconsin 53706, USA
&
A. J. Baker
USDA Forest Service, Forest Products Laboratory, Madison, Wisconsin 53705, USA*
(Received 4 June 1990; revised version received 1 October 1900; accepted 2 October 1900)
Abstract
A systematic compilation of 21 different property
values of wood and bark fue1 was made to facilitate
the engineering analysis and modeling of combustion
systems. Physical property values vary greatly
and properties such as density, porosity, and internal
surface area are related to wood species
whereas bulk density, particle size, and shape distribution
are related to fuel preparation methods.
Density of dry wood and bark varies from 300 to
550 kg m
-3
; bulk density of prepared wood fuel
varies from about 160 to 230 kg m
-3
. Thermal
property values such as specific heat, thermal conductivity,
and emissivity vary with moisture
content, temperature, and degree of thermal
degradation by one order of magnitude. The
carbon content of wood varies from about 47 to
53% due to varying lignin and extractives content.
Mineral content of wood is less than 1%, but it can
be over 10 times that value in bark. The composition
of mineral matter can vary between and within
each tree. Properties that need further investigation
are the temperature dependency of thermal conductivity
and the thermal emissivity of char-ash.
How mineral matter is transformed into various
sizes of particulates is not well understood.
A HC
AP
AV
c
D
EC
ECO
EHC
EP
Gg
hD
~
k
kr
M
M'
ρ
R
Re
RH C
RC
RCO
Rp
S
Sc
Tg
δ
ρc
ρg
ρs
µ
[ ]
Key words: Wood, combustion, fuel properties,
modeling.
NOMENCLATURE
AC
Pre-exponential factor for char
AC O Pre-exponential factor for CO
*The Forest Products Laboratory is maintained in cooperation
with the University of Wisconsin.
Pre-exponential factor for hydrocarbons
Pre-exponential factor for pyrolysis
(h
-1
)
Surface area per unit volume (cm
-1
)
Specific heat (J kg.
-1
K
-1
)
Diffusivity of oxygen (m
2
s
-1
)
Activation energy for char (cal mol
-1
)
Activation energy for CO (cal mol
-1
)
Activation energy for hydrocarbons (cal
mol
-1
)
Activation energy for pyrolysis (cal
mol
-1
)
Mass flux of gas (kg m
-2
h
-1
)
Mass transfer coefficient (kg s
-1
m
-2
)
Thermal conductivity (W m
-1
K
-1
)
Rate constant
Moisture mass fraction
Molecular weight (kg kgmol
-1
)
Pressure (atm)
Universal gas constant (cal mol
-1
K
-1
)
Reynolds number
Hydrocarbon reaction rate (kgmol m
-3
)
Char reaction rate (kg s
-1
)
CO reaction rate (kgmol m
-3
)
Pyrolysis rate (kg s
-1
)
Specific gravity
Schmidt number
Solid temperature (K)
Void fraction
Char density (kg m
-3
)
Density of gas phase (kg m
-3
)
Solid density (kg m
-3
)
Viscosity (N s m
-2
)
Species concentration (kgmol m
-3
)
161
Bioresource Technology 0960-8524/91/S03.50
Great Britain
1991 Elsevier Science Publishers Ltd, England. Printed in
162 K. W. Ragland, D. J. Aerts, A. J. Baker
INTRODUCTION
Analysis and modeling of combustion in stoves,
furnaces, boilers and industrial processes require
adequate knowledge of wood properties. Detailed
computer modeling of combustion processes
requires accurate property values, and these properties
are not generally available in one reference
source. The objectives of this paper are to review
existing property data on wood which are needed
for the analysis of combustion systems and to suggest
properties that need further quantification.
Fuel properties for combustion analysis of
wood can be conveniently grouped into physical,
thermal, chemical, and mineral properties. Bark
properties should be distinguished from wood
properties. Thermal degradation products of
wood consist of moisture, volatiles, char and ash.
Volatiles are further subdivided into gases and
tars. Some properties vary with species, location
within the tree, and growth conditions. Other properties
depend on the combustion environment.
Where the properties are highly variable, the
likely range of the property is given. Hence, this
paper is a general review of properties rather than
a tabulation of specific data.
Combustion systems using wood fuel may be
generally grouped into fixed bed, suspension
burning, and fluidized bed systems. The systems
range from residential to commercial and industrial
to utility scale. The fuel property data
needed depend. of course. on the type of application
and the detail of the model. In general, combustion
models can be classified as macroscopic
or microscopic. Wood fuel properties for macroscopic
analysis, such as ultimate analysis, heating
value. moisture content, particle size, bulk density,
and ash fusion temperature, have recently been
reviewed (Bushnell et al., 1989). Properties for
microscopic analysis include thermal, chemical
kinetic, and mineral data, and these properties
have not been collected in one source for wood.
Although there is a risk of oversimplifying complex
processes, the authors hope that the information
provided will assist the combustion modeler
in the difficult job of assembling property data.
The paper is not a comprehensive review of all
available wood fuel property data, but rather gives
a data set sufficient for modeling purposes.
PHYSICAL PROPERTIES
Physical properties for analysis of combustion
systems include density, hulk density, particle size,
internal and external surface area per unit volume,
porosity, and color.
The density of commercially important wood
species in the United States ranges from 300 to
550 kg m
-3
on an ovendry basis (Forest Products
Laboratory. 1987). The range in density of bark is
similar (Koch, 1972). Tropical woods vary more
widely. with a dry density as high as 1040 kg m
-3
.
For a particular species, the variation in dry
density is not more than about 10%. The density
of the wood cell wall is 1450 to 1550 kg m
-3
(Weatherwax & Tarkow, 1968). With increasing
moisture content, both the density and volume
increase up to the fiber saturation point of about
30% moisture. Above this point, only the density
increases (Forest Products Laboratory, 1987).
Fresh. green wood has a moisture content of
35-60%, on a wet basis. Dried wood used for fuel
typically has a moisture content of 5-20%.
The particle size of wood fuels ranges from
whole trees to sander dust and includes processed
and pelletized fuels. Dry bulk density, which is
defined as mass of ovendry wood per unit of
green bulk volume. ranges from 157 kg m
-3
for
hardwood shavings to 227 kg m
-3
for hardwood
sawdust (Table 1) (Forest Products Laboratory,
1987). Approximately 80% of the sawdust is
smaller than 2·5 mm and 80% of the pulp chips
and hog fuel are less than 25 mm and 30 mm,
respectively, based on the particle width (Fig. 1)
(Seppanen, 1988). Of course, the size distributions
vary with the equipment used to make the
chips. Representative shape factors (length to
thickness ratio) for sawdust, pulp chips, and hog
fuel, are 3·5, 5·8 and 35, respectively (Seppanen,
1988).
The porosity and internal surface area of wood
char are needed for detailed modeling of pulverized
wood combustion. Upon completion of pyrolysis,
the porosity of char is 0·8-0·9 for wood
heated under combustion conditions. The initial
char density is 10-20% that of the dry wood.
Table 1. Representative dry bulk density of typical wood
fuels (Forest Products Laboratory, 1987)
Fuel type
Hardwood sawdust and bark
Mixed pine-hardwood sawdust
Clean hardwood pulp chips
Hardwood whole-tree chips
Pine whole-tree chips
Hogged dry trims (hardwood)
Shavings (hardwood)
Bulk density (kg m-3
)
227
219
210
211
181
221
157
Properties of wood for combustion analysis 163
Fig. 1. Cumulative size distribution for sawdust,
unscreened pulp chips and hog fuel.
depending on the heating rate. As the char reacts.
the density decreases further. The internal surface
area of wood char (Wark & Warner, 1981) is of
the order of 10
6
m
2
kg
-1
. The pore size distribution
of wood char (Baileys & Blankenhorn, 1982)
ranges from 0·01 to 30 µm. For packed bed combustion
of pulp chips, the void fraction and
external surface area per unit total volume are
used in some models. For pulp chips, the void
fraction is 0·5-0·6, and the surface area per unit
total volume (Ragland, K. W., Aerts, D. J. & Baker,
A. J., unpublished) is about 1·5 cm
-1
.
When heated, wood changes color. At 250°C,
wood is dark brown; above 300°C, wood is black.
The brown color is associated with the onset of
pyrolysis. Above 300°C, shrinkage also begins
(Baileys & Blankenhorn, 1982).
THERMAL PROPERTIES
Important thermal properties in the analysis of
wood combustion include the specific heat of
wood and char, the thermal conductivity of wood
and char, and the emissivity of char. Specific heat
depends on temperature and moisture content but
not on density or species. The specific heat of dry
wood is given by (TenWolde et al., 1988)
c(dry) = 0·1031 + 0·003867T (kJ kg-1
K
-1
)
(1)
The specific heat of wet wood is greater than
would be expected from the simple law of mixture
as a result of the energy absorbed by the
wood-water bonds and can be represented by a
correction term (TenWolde et al., 1988)
c(wet)=[c(dry)+4·19 M]/(l+M)+A
(kJ kg
-1
K
-1
) (2)
where
A=(0·02355T - 1·32M - 6·191) M,
T = temperature in K, and
M = fractional moisture content. dry basis.
Above the fiber saturation point, A = 0. Specific
heat can vary by as much as 100% depending on
the temperature and moisture content (Table 2).
The specific heat of wood char can reasonably
be assumed to be the same as that of graphite. The
specific heat of graphite varies from 0·715 kJ kg
-1
K
-1
at 300 K to 2·04 kJ kg
-1
K
-1
at 2000 K. A
curve fit of the data from 700 to 2000 K, which
holds to within 5%, is given by (Stull, 1971)
c(char) = 1·39 + 0·00036 T (kJ kg
-1
K
-1
) (3)
The thermal conductivity of wood increases
with density, moisture content. and temperature.
Also, the thermal conductivity is approximately
1·8 times greater parallel to the grain than in
either the radial or tangential directions. For a
wide variety of wood species at room temperature.
the average thermal conductivity perpendicular
to the grain (TenWolde et al., 1988) is given
by
k = S(0·1941 + 0·4064M) + 0·01864
(W m
-1
K
-1
) (4)
where S is the specific gravity based on volume at
the current moisture content and weight when
ovendry. For example, if S = 0·47 and M = 0·10,
k = 0·129 W m
-1
K
-1
. The thermal conductivity
of wood increases with temperature at the rate of
approximately 0·2% per °C above room temperature.
Hence, thermal conductivity increases 10%
for every 50 K increase in temperature. The temTable
2. Specific heat of wood for selected temperatures
and moisture contents
Temperature
(K)
280
300
320
340
360
Specific heat (kJ kg-1
K-1
)
c(dry) c(5%) c(12%) c(20%)
1·2 1·3 1·5 1·7
1·3 1·4 1·7 1·9
1·3 1·5 1·8 2·0
1·4 1·6 1·9 2·2
1·5 1·7 2·0 2·3
164 K. W. Ragland, D. J. Aerts, A. J. Baker
perature correction is based on limited data, however
and should be used with caution.
The thermal conductivity of char is reported to
be 0·052 W m
-1
K
-1
at room temperature for
maple, beech, and birch char (Wheast, 1976).
Intuitively, the thermal conductivity of char
depends on density, which depends on the heating
rate. Evans and Emmons (1977) report that the
thermal conductivity is given by
k = 0·67S-0·071 (W m
-1
K
-1
) (5)
where S is the specific gravity as noted above.
Obviously. this relation fails at S < 0·106. The
variation of char conductivity with temperature is
not available.
The thermal radiation emissivity of carbon is
0·80 and is independent of temperature (Anon..
1972a). Actual wood char may have a lower emissivity
resulting from the ash layer on the surface of
the char. The character of the ash layer depends
on temperature, flow velocity, and oxygen concentration;
therefore, char emissivity is difficult to
assess. Refractory oxides typically have an emissivity
of about 0·4 at 1300 K, for example, and
wood ash emissivity may be similar. A char emissivity
of 0·75 has been used for modeling
purposes (Evans & Emmons, 1977).
CHEMICAL PROPERTIES
Important chemical properties for combustion are
the ultimate analysis, proximate analysis, analysis
of pyrolysis products, overall heating value. heat
of pyrolysis, heating value of the volatiles, and
heating value of the char.
The ultimate analysis (Table 3) provides weight
percentage of C, H, O, N, and S. The C content of
softwood species is 50-53%, and that of hardwood
species 47-50% due to the varying lignin
and extractives content. All wood species contain
about 6% H (Petura, 1979). Oxygen content
Table 3. Ultimate analysis of wood (dry, ash-free weight per-
cent)
Element Average of 11 Average of 9 Oak Pine
hardwoodsa
softwoodsa
barkb
barkb
C 50·2 52·7 52·6 54·9
H
O
6·2 6·3 5·7 5·8
43·5 40·8 41.5 39·0
N 0·1 0·2 0·1 0·2
S — 0·0 0·1 0·1
a
Tillman et al. (1981).b
Anon. (1972b).
ranges from 40% to 44%, S is less than 0·1% and
N ranges from 0·1 -0·2%.
Proximate analysis gives the weight fraction of
moisture, and volatiles including tar. char. and
ash. according to ASTM standard test method
E870-82. The ground wood sample is heated in
air to 103°C to constant weight to obtain dry
wood, to 950°C in a capped crucible to obtain
char. and to 600°C in air to obtain ash. The
volatile yield can be increased by increasing the
heating rate (i.e. increasing the heat transfer rate
or increasing the exposed surface area). Conversely,
the char yield is increased by slow heating
of large particles. The standard procedure for
proximate analysis is more representative of combustion
on a grate than in suspension. The ASTM
volatile yield for dry ash-free wood ranges from
70% to 90%, depending on the species. The
ASTM volatile yield for bark is typically closer to
70%.
The volatile yield of dry wood under combustion
conditions may be further subdivided into
light hydrocarbons. tar. carbon monoxide, carbon
dioxide, hydrogen. and moisture yield (Table 4).
The yields depend on the temperature and heating
rate of pyrolysis. The tar has an average composition
(Adams. 1980) of C6H6.2O0.2, and the
light hydrocarbons are primarily methane.
The higher heating value of different wood
species on a moisture-free basis varies less than
15%. The higher heating value of softwoods
(Baker. 1989) is 20-22 MJ kg
-1
and of hardwoods,
19-21 MJ kg
-1
. The average higher heating
value of seven Populus hybrids (Bowersox et
al., 1979) is 19·49 MJ kg
-1
. Typically, softwoods
have more extractives and more lignin than do
hardwoods, which accounts for the slightly higher
heating value.
The heating value of the volatiles may be
estimated by the following method. Pyrolysis is
Table 4. Distribution of pyrolysis products
for dry wood under combustion conditionsa
product
H2O
CO2
CO
H 2
Light hydrocarbons
Tar
Char
Mass fraction
0·25
0·183
0·115
0·005
0·047
0·20
0·20
a
Bowersox et al. (1979) and Forests Products
Laboratory (1987).
Properties of wood for combustion analysis 165
initially slightly endothermic and then exothermic,
so that overall pyrolysis is considered essentially
thermally neutral (Roberts, 1971). Using the data
in Table 4 and assuming that char is all carbon
with a higher heating value of 32·8 MJ kg
-1
, an
average higher heating value of the volatiles is
calculated to be 16·2 MJ kg
-1
. Subtracting the
heating value of CO and H2, the light hydrocarbons
and tars have a higher heating value of
41·0 MJ kg
-1
.
Above 700°C char contains little hydrogen.
Stimely and Blankenhorn (1985) reported the
hydrogen content of char as a function of temperature
for four woods. For example, for hybrid
poplar, the percentage of hydrogen by weight is
H2 = 7·39-0.01 T (6)
where T is the final carbonization temperature
(°C). Hence for temperatures above 739°C, which
is typically the case for most combustion applications,
char does not contain hydrogen.
REACTION RATES
To analyze combustion systems, it is often useful
to model the rate at which drying, pyrolysis, and
char bum take place. Drying rates are outside the
scope of this paper. During pyrolysis the components
of wood. such as cellulose, hemicellulose,
and lignin, pyrolyze at different rates; however,
for combustion analysis, a single global rate
expression of the Arrhenius form often gives
acceptable results (Table 5). Roberts (1970)
reviewed the literature on the pyrolysis of wood
and suggested that an activation energy of 31 kcal
mol
-1
is appropriate for both small and large
particles (Table 6); a conclusion that was supported
by Belleville (1984). The suggested preexponential
constant (Roberts, 1970; Belleville,
1981) in the Arrhenius equation is 7 × 10
7
s
-1
. As
noted previously, the pyrolysis gases contain
hydrogen, carbon monoxide, light hydrocarbons
and tars. The hydrogen can be considered to react
instantaneously at combustion temperatures. The
global oxidation rate of CO to CO, is well known
and depends on the concentration of water vapor
(Table 5). Global kinetic reaction rates for hydrocarbon
volatiles from wood combustion are not
available. Smoot and Smith (1985) have reviewed
the kinetics for coal volatiles, and their results
indicated that long-chain and cyclic hydrocarbons
react to form hydrogen and carbon monoxide
according to the following global reaction:
CnHm + 0·5nO2 = 0·5mH2 + nCO
Light hydrocarbons and tars are lumped together
in this approach.
Table 5. Reaction rate equations
Reference
Roberts (1970)
Smoot and Smith (1985)
Smoot and Smith (1985)
Smoot and Smith (1985)
hD
~
= 1·46 ρN2
M'gGgSc
-2/3
Re-0·41
( 1 - δ )0·2
Re = Gg/(Av(1-δ)µ); Sc=µ/ρg
Rate equation
Rp = Ap (ρs - ρc) exp( - Ep /RTs )
Rc = (Avρo 2
n
(M'g /M's) kr + 2hD
~
)/((M'g/M's) kr+ 2hD
~
)
RHC = AHC Tg ρ 0·3
[HC]0·5
[O2]exp(–EHC/RTg)
RCO = ACO[CO] [O2]0·25
[H2O]0·5
exp(–ECO/RTg)
where
kr= Acexp(– Ec /RTs )
Gamson (1951)
Table 6. Reaction rate constant values for equations in Table 5
Parameter Value Reference
Ap
2·52 × 1011
h-1
Roberts (1970)
Ep
3·1 × 104
cal mol-1
Roberts (1970)
ACO
4·0 × 105
(see reaction rate) Smoot and Smith (1985)
ECO
4·10 × 104
(cal mol-1
) Smoot and Smith (1985)
AH C
2·07 × 104
(see reaction rate) Smoot and Smith (1985)
EHC
1·917 × 107
cal mol-1
Smoot and Smith (1985)
AC
1·71 × 107
kg m-2
h-1
EC
23·8 × 104
cal mol-1
Bhagat (1980)
Bhagat (1980)
166 K. W. Ragland D. J. Aerts, A. J. Baker
Char reacts with oxygen, carbon dioxide, water
and hydrogen. Char oxidation is the dominant
reaction in a combustion environment. Since this
is a surface reaction. the reaction rate depends on
kinetics and diffusion of oxygen to the char
surface (Tables 5 and 6). Minimal wood-char
kinetic data are available compared to coal-char
data, although this is important because char
burnout is much longer than volatile burnout. The
diffusion parameters depend on the gas conditions
and are beyond the scope of this paper. as is
the use of intrinsic kinetic data instead of global
kinetics.
Table 8. Mineral analysis of whole tree and bark
Element Mineral content (ppm)
Aspen Pine barkb
Oak barkc
(whole tree)a
MINERAL PROPERTIES
The mineral content of clean wood of temperate
tree species is 0.1% to 06% and that of bark 3%
to 5%. Mineral matter in wood consists mostly of
salts of calcium, potassium, and magnesium. but
salts of many other elements are also present in
lesser amounts (Table 7). Some salts are formed
with the organic acid groups of the cell wall components.
whereas others occur as carbonates.
phosphates, sulfates, silicates and oxalates. The
mineral content of wood and bark is highly
variable between and within species and can vary
with soil and growth rate. Whole-tree analyses of
minerals and trace elements in aspen. pine bark.
and oak bark (Table 8) are different from that of
pure wood. Silica is rarely present in more than
trace amounts in the wood of temperate tree
species (Pettersen, 1984). The silica content of
clean bark is higher than that of wood. Analysis of
bark from 18 softwood species and 24 hardwood
species indicated an average silica content of
Total ash: a
= 1·45%: b
= 2·9%: c
= 5·3%.
Table 7. Mineral analysis of wood
Element Mineral content (ppm)
Yellow Poplara
White Oaka
Loblolly Pineb
Ca 1182 1976 1141
K 517 1128 116
Mg 265 294 326
Na 258 216 230
P 108 75 40
Mn 48 59 121
Fe 35 70 72
Al 14 32 —
Cu 3 2 —
Zn 8 62 —
a
Koch (1985).b
Koch (1972).
Ca 4160 5282 24418
K 1424 1333 88
Mg 493 1137 3816
P — —323
S 107 — —
Fe 30 609 1224
Zn 36 — —
Mn 17 — —
Al 10 2149 29
B 6 — —
Cu 13 — —
Na 17 280 3500
Pb 1·4 — —
Cd 0·30 — —
Si — 5278 2745
0·05% and 0·11%, respectively (Harder &
Einspahr, 1980). The silca content of bark. as
used for fuel. however. is higher because of soil
contamination by wind and harvesting methods.
Ash is formed from mineral matter during
combustion and gasification. The ash yield of
wood grown in the temperate zones is 0·1-1·0%,
whereas wood grown in the tropics contains up to
5% ash (Fengel & Wegener, 1984). Bark contains
3-8%, ash. Wood ash typically includes 40-70%
calcium oxide and 10-30% potassium oxide.
During combustion. the mineral ions oxidize
and volatilize or form particulates. The char
surface is hotter than the gas or the interior of the
particle. and ash particles tend to form on the char
surface. As the char burns away, the ash particles
are released. Knowledge of mineral matter
behavior is needed for analysis of ash deposition,
erosion and corrosion in combustion systems, as
well as for understanding stack gas opacity.
The ash-fusion temperature, which is measured
in the laboratory by ASTM method D 1857, is
used as an indicator of deposition tendency in
actual combustion systems (Table 9). The ashfusion
temperature is determined by mixing the
ash sample from proximate analysis with a solution
of dextrin to form a stiff paste that is pressed
into small cone-shaped molds. The oven temperature
at which the cone first deforms is reported as
the onset of ash fusion. In addition to the initial
deformation temperature, the softening temperature,
hemispherical temperature, and fluid temperature
are reported. Both oxidizing (air) and
Properties of wood for combustion analysis 167
Table 9. Ash-fusion temperatures for wood and bark
Source of ash Reducing temperature
(°C)
Oxidizing temperature
(°C)
Wood a
Initial deformation
Fluid
Oak bark b
Initial deformation
Softening
Fluid
Pine bark b
Initial deformation
Softening
Fluid
1450-1515 —
1500-1550 —
1477 1471
1493 1499
1504 1510
1193 1210
1226 1249
1266 1288
a
Baumeister et al. (1978).
b
Anon. (1972b).
reducing (60% CO, 40% CO2) atmospheres are
used. Wood has a higher ash-fusion temperature
than do many coals (Anon., 1972b) and hence can
sometimes be less prone to ash deposition than
coal. However, detailed testing is required to
investigate fouling. slagging and agglomeration.
CONCLUDING REMARKS
Property values for the analysis of combustion of
wood fuels have been collected and are reviewed
in this paper. Selected properties arc given for
unheated wood and bark and for the thermal
degradation products of volatiles, char and ash. In
some geographical locations, bark is the main
wood fuel; generally, less basic data on bark are
available when compared to available data on
wood.
Physical properties of density, size. and surface
area per unit volume are presented for various
types of wood fuels. The thermal properties of
specific heat and thermal conductivity of wood
and char are well known; however. the temperature
dependency of the char thermal conductivity
is not known. The thermal emissivity of char-ash
is not well known. Overall heating values and elemental
analyses are readily available. Pyrolysis
product yields are highly dependent on heating
conditions. Depending on the actual conditions,
the standard proximate analysis may or may not
be applicable. Suggestions arc made above for
approximating the volatile reaction rates, but
further investigation in this area is needed. Global
char reaction rates also need further observation.
Mineral matter in wood is well characterized.
Little is known. however. on how the mineral
matter is transformed into particulate ash of
various sizes.
168 K. W. Ragland, D. J. Aerts, A. J. Baker
Printed on recycled paper