Experimental and Theoretical Investigation of Heat and Mass Transfer Processes during Wood Pyrolysis Won Chan Park , Arvind Atreya1 Department of Mechanical Engineering University of Michigan & Howard R. Baum National Institute of Standards and Technology Building and Fire Research Laboratory Abstract Thermal decomposition of 25.4 mm diameter wood spheres is studied both experimentally and theoretically. Wood spheres were pyrolyzed in a vertical tube furnace at temperatures ranging from 638K to 879K. Mass loss and temperatures of the sample were measured during pyrolysis. Center temperature measurements showed two distinct thermal events consisting of sequential endothermic and exothermic reactions. A numerical investigation of these endo/exothermic reactions using various pyrolysis kinetics models was conducted to determine the pyrolysis mechanism and the heats of the pyrolysis reactions. A comparison of the experimental and numerical results showed that: (i) Contrary to the suggestions in the literature, the contributions of the secondary tar decomposition and lignin decomposition to the center temperature exothermic peak are small. (ii) Exothermic decomposition of the intermediate solid is responsible for the center temperature peak. (iii) The center temperature plateau is caused by the endothermic decomposition of cellulose. Based on the experimental and numerical results, a new wood pyrolysis model is proposed. The model consists of three endothermic parallel reactions producing tar, gas and intermediate solid and subsequent exothermic decomposition of the intermediate solid to char and exothermic decomposition of tar to char and gas. The proposed pyrolysis model shows good agreement with the experiments. Pressure calculations based on the new pyrolysis model revealed that high pressure is generated inside the biomass particle during pyrolysis. This is expected to split apart the biomass particle, as observed during the experiments. The splitting is due to both weakening of the structure and internal pressure generation during pyrolysis. At low heating rates, structural weakness is the primary factor, whereas at high heating rates, internal pressure is expected to be the determining factor, much like popcorn. It is expected that moisture, while not considered in this work will have a similar effect, but at lower temperatures. 1. Introduction Biomass is expected to be a major source of sustainable energy in the future as the World transitions from traditional carbon-based fuels such as coal, petroleum and natural gas to carbon-neutral fuels to combat global warming and fossil fuel depletion. As recognized in Ref [1], biomass-derived biofuels are the only current sustainable source of liquid hydrocarbons required for transportation. However, an economic way of making them has not yet been devised despite the fact that lignocellulosic biomass costs significantly less than crude oil (about $15 per barrel of oil energy equivalent). The U,S. forests, crops, and urban wood wastes can produce 1.3 billion dry tons of renewable biomass per year, without land use change [2]. Land use change is an important consideration because using good cropland for the production of biofuels exacerbates the very global warming problem they attempt to solve [3]. Further, if the excess forest biomass is not harvested, it only serves to produce devastating forest fires. If harvested, it can provide a significant amount of carbon-neutral liquid 1
Corresponding author
[email protected]
biofuels for transportation use. Harvesting the excess forest biomass also makes the forest healthy. Forestry data shows that well-managed healthy forests grow 55% faster. Thus, extraction of this forest biomass is better than carbon neutral. Barriers to usage of this biomass are: (i) It is voluminous – resulting in low energy density. (ii) It is distributed over large and remote areas. (iii) The cost of transportation from remote locations to a central processing facility is high. (iv) It is chemically diverse making biochemical fermentation techniques difficult, and (v) The method of waste disposal (minerals, ash, etc.) that is generated during processing at a central facility is important for sustainable forests and crops. The recent Tennessee toxic coal ash sludge spill points to proper disposal of the waste generated at any central processing facility. While it is toxic to people, it contain essential nutrients for sustainable plant growth. Currently available pyrolysis methods for conversion of this biomass to liquid transportation fuels require small particle size (~1mm) and several times the amount of
they attributed the center temperature peak to the exothermic secondary reaction between volatiles and char. Strezov et al. [9] measured the heat of pyrolysis ‘ ’ of cellulose, hemicellulose, lignin and four different sawdust biomass samples in an infrared furnace. They reported that all of the samples showed both endothermic and exothermic reactions during pyrolysis. They attributed the exothermic behavior to the exothermic tar cracking [10] and the decomposition of dehydrocellulose [8,11]. From this previous work, a few facts are clear: (i) Both endothermic and exothermic reactions have been observed previously by several investigators in addition to this work, however, the explanations for this behavior differ significantly. The reaction mechanism has not been clarified. (ii) The rates of the reactions involved and the heats of these reactions also have not been quantified. For thick wood particles, internal pressure generation is also an important factor influencing the pyrolysis process and the temperature. Pressure gradient drives the volatiles out of the particle and high internal pressure may also split a partially pyrolyzed wood particle. Pressure splitting of thick particles is desirable because it obviates the need to make small particles. This natural formation of small particles enhances the biomass conversion speed and increases the yield of liquid products. It also reduces the residence time of the volatiles in the pores reducing tar cracking that would otherwise be promoted by high pressure. There is a substantial amount of literature on thermal decomposition of lignocellulosic materials, however, only a few studies have considered pressure generation during pyrolysis. Fredlund [12] measured pressure variations at multiple locations in a pyrolyzing thick block of wood. Baum and Atreya [13] analytically solved the pressure distribution in a pyrolyzing charring solid undergoing flame spread. Park, et al. [14] developed a numerical model to solve for pressure generation during pyrolysis. Internal pressure measurements and/or models have not been reported for pyrolysis of small wood particles. The small sample size makes the pressure measurement very difficult and it is usually assumed that the gases that are generated simply leave the small particle. In this work, internal pressure generation in small particles is numerically investigated. The objective is to provide a better understanding of wood pyrolysis with emphasis on quantifying the heat and mass transfer that occurrs during pyrolysis. We employ both experimental and numerical methods to develop a wood pyrolysis model useful for producing liquid bio-fuels from chipper-size (~1 inch) biomass particles.
catalyst [4]. This is clearly not possible with the widely distributed 1.3 billion dry tons of renewable biomass supply available per year [2]. Thus, an economical way has to be found to process this biomass on-site without finely grinding it. The present work is a step toward this goal. Here we attempt to understand and model the behavior of a wood chipper-size (~1 inch) particle during pyrolysis. Wood pyrolysis is a complex process. It involves many physical and chemical processes such as heat transfer, moisture evaporation, decomposition kinetics, heat of pyrolysis, pressure build up in the solid, changes in material properties with the extent of pyrolysis and temperature, anisotropic property behavior, among others. To add to these complications, our experimental measurements show that during the pyrolysis of moisture free wood spheres, a distinct center temperature plateau appears between 610 and 640K owing to an endothermic reaction. Further, immediately after the endothermic reaction, the center temperature rises sharply and exceeds the surface temperature due to an exothermic reaction. Thus, from a thermal perspective, wood pyrolysis is a combination of successive endothermic and exothermic reactions. Similar observations have been reported by other researchers in the literature [5-9] but the current wood pyrolysis models do not account for this phenomenon. Nevertheless, it is important for determining the energy required to convert solid wood to liquid biofuels and from the point-of-view of pressure generation that may split the particle apart obviating the need for creating small particles. A single heat of pyrolysis, ‘ ’ of lignocellulosic materials quoted in the literature and it varies greatly depending on the type of material, experimental setup and conditions. Values found in the literature range from 0 to 1500 kJ/kg [7]. Bilbao et al. [5] measured of Pinus Pinaster using DSC (differential scanning calorimeter) in order to use it as input data for their wet wood pyrolysis model. They expressed their results as: for the first endothermic reaction stage (up to 60% conversion) and for the second exothermic reaction stage (remaining 40% conversion) . They presumed that the first endothermic reaction corresponds to cellulose and hemicellulose decomposition and lignin decomposition accounts for the second exothermic reaction. Di Blasi et al. [6] reported that the inner core of the wood cylinder decomposition showed endo-/exothermic reactions. They observed that the two reactions are more distinguishable at lower heating rates. At higher heating rates, the endothermic and exothermic reactions overlap producing a smaller temperature peak at the center. They also explained this thermal behavior by presuming endothermic decompositions of holocellulose and extractives and exothermic lignin decomposition. Koufopanos et al. [7] also measured the center temperature of dry wood cylinders during pyrolysis and obtained results similar to Di Blasi et al’s [6]. However,
2. Experiment setup Pyrolysis experiments were conducted on moisture free maple wood spheres 1” in diameter. Spherical geometry was used to maintain one dimensionality for
2
final char yield decreases from 31% to 17% of the initial sample mass, whereas mass loss rate increases. Wood sphere pyrolysis begins with low temperature hemicellulose decomposition showing a mild weight loss rate. As the temperature increases, cellulose starts to decompose and the solid mass fraction decreases rapidly. The majority of weight loss occurs during this period. After the completion of cellulose decomposition, the leftover lignin continues to decompose resulting in a gradual weight loss rate. Temperature measurements at the surface and at the center of the sample are shown in Fig. 3. The surface temperature rises quickly and attains thermal equilibrium with the furnace. Center temperature measurements show distinct plateaus in the temperature range of 610K to 640K, indicating an endothermic reaction. Immediately after completion of the endothermic reaction, the center temperatures rise sharply owing to the subsequent exothermic reaction. During the exothermic reaction, the center temperatures exceed the surface temperature by 10 ~ 70K except for 879K case where it still rises sharply but the surface temperature also rises sharply. While the center temperature peaks are more distinguishable for the low temperature cases (688K to 783K), the exothermicity of the second reaction enables the heating of the entire sample to occur with only a short lag from the surface. This is an important observation because large internal pressure is generated that splits apart the sample. Thus, the sample size is naturally reduced for producing liquid hydrocarbons via fast pyrolysis. For high furnace temperature cases, the center temperature peaks were small or not observed because of the overlap between the endo- and exothermic reactions. An interesting comparison of solid mass loss and temperature for T fn 688 K is shown in Fig. 4. It
ease of modeling. Weight and temperature changes of the sample during pyrolysis were measured. The apparatus used is schematically shown in Fig. 1. The spheres were heated in the vertical tube furnace at temperatures ranging from 638 to 879 K. The tube furnace (Carbolite® GVA 12/300) consists of a mullite tube (length: 558mm, inner diameter: 106mm), two electric heaters (total heating length: 300 mm), a temperature controller, insulators and a stainless steel outer casing. As shown in Fig. 1, weight loss of the sample was measured by a scale located above the furnace. Sample surface temperature was measured by a K-type thermocouple (bead diameter ~ 0.2mm). The thermocouple bead was positioned in a tiny slit on the surface of the sample and fixed by glue. Temperatures inside the sample were measured at two locations, center ( r 0) and middle (r r0 / 2) by two thin sheathed K-type thermocouple probes (0.25mm sheath diameter, 450mm length) inserted though holes drilled in the sample. Furnace temperatures were monitored by eight thermocouples on the tube inside surface and two thermocouples on the top and bottom end insulating caps. From temperature measurements at ten locations on the furnace, the effective furnace temperature T fn was calculated by the following equation.
4 Tfn FT i i i Where,
0.25
,
i 1 ...10
(1)
Fi is the view factor between the ith section area
of furnace inside surface and the sample surface [15]. Argon at ambient temperature was used to purge the furnace at a low flow rate of 0.21 g/s to displace oxygen and carry away the pyrolysis products. Gas temperature near the sample was also measured by a K-type thermocouple to determine the convective heat losses. The voltage signals from all the thermocouples were processed by two NI-SCXI-1112 (8 channels) boards installed in a PXI-1011 chassis. The temperature data was recorded at 1 Hz frequency. The wood spheres were dried at a temperature of 115℃ for at least 3 hours before the pyrolysis experiments to remove moisture. In order to avoid interference in the mass loss measurement due to the installation of temperature probes in the sample, mass measurements and temperature measurements were conducted separately for the same conditions.
indicates that the temperature plateau corresponds to the second half of the active mass loss period and gradual or negligible mass loss occurs during the center temperature peak. It is generally accepted that endothermicity during biomass pyrolysis is caused by the formation of tar or volatiles. Meanwhile, several mechanisms have been reported to be responsible for the exothermcity, such as exothermic lignin decomposition, secondary tar cracking reaction, dehydrocellulose decomposition to gas and char. These exothermic reaction mechanisms will be discussed in the next part. 4. Thermal mechanism of wood pyrolysis Various exothermic mechanisms were investigated numerically to determine their effect on thermal characteristics of wood pyrolysis and to determine the reason for the center temperature peak. Three different models and experimental measurements were used. 4.1 Model description 4.1.1 Kinetics model
3. Experimental results Wood sphere pyrolysis experiments were conducted at six different furnace temperatures ranging from 638K to 879K. Figure 2 shows the measured solid mass fraction (Y = sample weight / initial sample weight). As the furnace temperature increases from 638K to 879K, the
3
Hemicellulose decomposition is modeled by two-step reactions [18]. First, hemicellulose decomposes to volatiles and intermediate solid, and then the intermediate solid decomposes to volatiles and char in the subsequent reaction. The kinetic parameters of hemicellulose pyrolysis are listed in Table 2.
Model-1 In this model, wood decomposes into three major products: tar, char and gas, by three primary reactions. And then, a portion of tar decomposes to gas and char by successive secondary reactions [16]. A schematic is shown below. gas kg kt
wood
hemicellulose
kg2
?1 intermediate solid
kc2
i t , g , c, g 2, c 2
(1- ?2) volatiles
+
B) Cellulose Cellulose pyrolysis is modeled by two parallel reactions are expressed in Eq. (3) [19]. Initial mass fractions of cellulose for the two reactions are predetermined as: clA,0 f A cl ,0 , clB ,0 f B cl ,0 .
Reaction rate k is assumed to follow a 1st order Arrhenius type reaction with activation energy E and pre-exponential constant A : Ei RT
k2
?2 char
char
+
tar
kc
ki Ai e
(1- ?1 ) volatiles
k1
h ’s of the two reactions are assumed to be the same. The kinetic parameters of cellulose pyrolysis are listed in Table 3.
(2)
The kinetic parameters and heats of pyrolysis of secondary reactions of model-1 were obtained from the literature and they are listed in Table 1. To see if the model will explain the experimental results, the heats of pyrolysis for primary reactions were obtained by fitting with the experimental measurements.
A
celluloseA
B
celluloseB
(1- ?A) volatiles + ?A char (1- ?B) volatiles + ?B char
E i d i m Ai e RT ini 1 0.99i i dt
Model-2 Wood consists of three major components: hemicellulose (25~35%), cellulose (40~50%) and lignin (16~33%). Each component shows different characteristics during its thermal decomposition. Hemicellulose decomposition occurs at 200℃~260℃ and produces acetic acid. Cellulose decomposes to levoglucosan and dehydrocellulose at 240℃~350℃. Lignin decomposes exothermally over a broad temperature range of 280℃~500℃ and produces more char than the other two components [17]. Model-2 regards wood as the mixture of these three major components. This model is useful to analyze the effect of decomposition of individual wood component. In model2, the kinetic parameters of each component were chosen from literature values that showed good agreement with the experiment. Following were chosen: hemicellulose [18]; cellulose [19]; lignin [20]. Unlike model-1, wood in model-2 decomposes to two final products; char and volatiles. This was necessary because a three-product model is not available for hemicellulose and lignin. Thus, volatiles account for the sum of gas and tar. Also, the char yield for model-2 is predetermined. The mass fractions for hemicelluloses, cellulose, and lignin are: 30%, 43.6%, and 26.4% respectively [21]. The extractives portion is included in hemicelluloses. The individual models are: A) Hemicellulose
Where, A clA , B clB clB ,0 clB ,0
(3) .
C) Lignin Lignin decomposes to volatiles and char by a single Arrhenius type reaction [20]. The kinetic parameters of lignin pyrolysis are listed in Table 4.
lignin
kL
(1 - ?L) volatiles + ?L char
Model-3 Model-3 has the structure of the Kilzer-Broido cellulose pyrolysis model [22]. Originally, the KilzerBroido model was postulated for three reactions of cellulose decomposition; the slightly endothermic reaction from cellulose to dehydrocellulose, the strong endothermic tar (mainly levoglucosan) formation from cellulose and the exothermic decomposition of dehydrocellulose to char and gas. Milosavljevic and colleagues reported that the enthalpy of cellulose pyrolysis becomes more endothermic as char yield decreases, based on their experiment and the literature [8, 23]. Their experimental results support the exothermic char formation mechanism of the Kilzer-Broido model. Wood pyrolysis experiments in this work showed thermal behavior similar to the cellulose pyrolysis experiment by Milosavljevic et al. One reason may be that
4
approximately half of the wood by mass consists of cellulose. Hence, it is plausible to apply the Kilzer-Broido type model to wood pyrolysis. In model-3, wood is pyrolyzed through two paths. One is an endothermic tar producing reaction and the other is an intermediate solid producing reaction which is assumed to have a zero heat of pyrolysis. The intermediate solid, corresponding to dehydrocellulose in the Kilzer-Broido model, is decomposed into char and gas by an exothermic reaction. Each reaction rate is assumed to follow a 1st order Arrhenius reaction. The activation energy of reaction 1 is taken from wood to tar producing reaction of model-1. k1
tar
k2
intermediate solid
wood
k3
Intermediate solid:
c S c kc a k c 2 t t
Char:
hcl S hcl k1 hcl t is Intermediate solid: Sis 1k1 hcl k2 is t clA d A , Cellulose: SclA clA,0 dt t clB dB SclB clB ,0 dt t Where, cl clA clB l Lignin: Sl k L l t Char:
c
t Model-3
S c 2 k 2 is A S clA B S clB L S l
Virgin wood:
a t
S a k1 k 2 a
(12)
2
t
t
Gas:
s
r r
t
t
2
g
Where,
t
t
c2
a
g2
t
(14)
1 r
r V S r
2
2
g
g
k g a k g 2 t (15)
is porosity calculated by 1
is total solid mass per unit volume,
w
s w
1 . w
is initial wood
porosity, w 0.4 [24].
(4)
Model-2 Volatiles: v
(5)
t
Model-2
Hemicellulose:
S is k 2 a k3 is
(13) S c k3 is t Mass change per unit volume of each gas phase component can be expressed as the sum of mass flux through the control volume boundaries and mass generation in the volume due to pyrolysis reaction. The mass flux of each gaseous species consists of convective flux due to gas flow and diffusive flux due to diffusion among the gaseous species. However, as the effect of the diffusion is very small compared to the convection, thus only the convective mass flux is considered here. Equations are written in spherical coordinates. Model-1 1 Tar: r V S k k k
? char + (1 - ?) gas
a S a kt k g k c a t
t
c
Char:
Where, char yield of reaction 3 and gas is estimated as 0.65 . 4.1.2 Mass conservation Mass conservation of each component is governed by diffusive and convective mass flux in the gas phase and production or destruction owing to the decomposition reaction. It is assumed that volume shrinkage does not occur. Material properties and reaction parameters used in equations are listed in Table 5. The mass change per unit volume of each solid phase component depends on the pyrolysis reactions. Model-1 Virgin wood:
is
1 k 2
(6)
Model-3
(7)
Tar:
2
1 r r 2
r V S 2
v
v
1 1 k1 hcl
16
is 1 A k A clA 1 B k B clB 1 L k L l
t t g
1 r
2
r V S r
1
2
t
t
k1 a
(17)
r V g S g 1 k3 is (18) 2 t r r Pressure equation Since the wood and char structure is composed of numerous tiny pores, the internal gaseous components flow is dominated by the viscous force. Thus, the gaseous components flow velocity V is calculated by Darcy’s law. B P V (19) r Where, is the viscosity. The permeability B of partially pyrolyzed solid is linearly interpolated between char and virgin wood as.
Gas:
(8)
(9) (10)
(11)
5
2
B 1 Bw Bc
Model-2
(20)
a Cw c Cc Tt C pv m v Tx
Where, the degree of pyrolysis is 1 a for modelw is for model-2 and model-3. 1 and 1 a w
x n T Q T k a n x x x 1
Where,
Pressure of model-1 and model-3 is the sum of the partial pressures of tar and gas. (21) P Pt Pg
Q Shcl Sis hhcl SclA SclB hcl Sl hl
Model-3
C
Gaseous components are assumed to behave as ideal gas.
P
RT
(22)
M
1
t T
r r
2
r
2
BPg P
R
T r
Mg
Sg
a Cw is Cc c C pt t C pg g
T T 1 T 2 T C pt t C pg g V r t r r 2 r r
(28)
Specific heat capacity for solid components C and constant pressure specific heat capacity C p for the gaseous components are functions of temperature as shown in Table 5. Effective thermal conductivity is calculated as the sum of solid and volatiles conductivity and radiation heat transfer through the pore [25].
(23)
13.5 T 3 d (29) 1 w c v e Where, , e and d are the Stefan-Boltzmann constant,
Gas Partial Pressure:
Pg
w
Where, Q k1h1 k 2 h2 a k3 is h3
Where, M and R are molecular weight and the universal gas constant. Combining Eq.’s (14) – (18) and (22) gives pressure equations. Model-1 & -3 Tar Partial Pressure:
Pt 1 2 BPt P R St r t T r 2 r T r M t
(27)
emissivity and pore size. Since heat is conducted in all three directions (radial, tangential and grain) in the sample, the thermal conductivity used for the 1-D model is taken as the average value of three directions. (30) w (w,radial w,tangential w, grain ) / 3
(24)
B) Model-2 P 1 2 BP P R Pressure: S v (25) 2 r t T r r T r M v The pressure boundary condition at the surface is set to ambient pressure Ps P0 .
c (c , radial c ,tangential c , grain ) / 3
(31)
Thermal energy flux through the wood sphere boundary surface is determined by external convective and radiative heat transfer conditions.
4.1.3 Energy conservation The conservation of energy is governed by the change of energy stored in the control volume by, thermal conduction, convective flow of gaseous phase components and heat generation by pyrolysis reaction. In order to setup the energy equation, local thermal equilibrium between gas and the solid phase components is assumed. The energy equations for the three models are: Model-1 T Cw a Cc c C pg g C pt t t (26) T 1 T 2 T C pg g C pt t V 2 r Q r r r r Where, heat generation is calculated as:
T h Tg T es T f4 T 4 ; (r r0 ) r
(32)
Furnace and argon temperatures surrounding the sample ' Tg ' are listed in Table 6. Surface emissivity is calculated as [24]: Ts 450 K es ew Ts 450 ec ew 450 K Ts 550 K (33 es ew 550 450 550 K Ts es ec
The surface heat transfer coefficient is estimated as
Q kt ht k g hg k c hc a k g 2 hg 2 k c 2 hc 2 t
h 20 (W / m 2 K ) based on the temperatures of the sample surface, gas near the sample and in the furnace.
.
6
Q
model taken from the literature as described in the previous section. On the other hand, the heat of pyrolysis, h of hemicellulose, cellulose and lignin
4.2 Results 4.2.1 Exothermic reactions Secondary tar decomposition The exothermic effect of the secondary tar decomposition was investigated by the comparison between ‘without secondary tar decomposition’ (case A) and ‘with secondary tar decomposition’ (case B). This comparison was made based on model-1. The kinetics parameters and heats of pyrolysis are listed in Table 1.
kg wood
gas kt
char wood
tar kc2
kc char
By neglecting the secondary tar decomposition mechanism, Eq.’s (5), (14) and (15) are modified for case A as shown below: c (34) Sc kc a t t 1 2 r V t S t kt a (35) 2 t r r
g t
1 r
2
r V S r 2
g
g
kg a
found
as:
h1 44kJ / kg ,
hA hB 280kJ / kg
and
Fig. 6 shows solid mass fraction comparison between the experiment and model-2 for different temperature conditions. Since pre-determined fixed char yield ratios are used for pyrolysis models of all three components, the final char of model-2 does not account for the final char yield variation for different temperature conditions observed in the experiments. As a result, except for the 688K case, the final char of model-2 is considerably different from the experiment. Model-2 predicts faster solid mass loss rate in the beginning stage due to the hemicellulose pyrolysis model. Also, slow lignin decomposition results in a gradual slope in the final stage. In general, model-2 does not predict the mass loss accurately as shown in the figure. This discrepancy is believed to be due to the following reasons. Pure cellulose, hemicellulose and lignin used to determine the kinetics model in the literature do not have the same chemical structure as those existing in wood. The interaction among components would affect the individual pyrolysis of each component, which is not considered in model-2. However, the model-2 results depict the characteristics of experimental solid mass loss in the beginning and the final stage better than model-1 (not shown) owing to separate modeling of hemicellulose, cellulose and lignin. Fig. 7 shows center temperature comparison between model-2 and the experiment. The center ( r 0 ) temperature plateau of model-2 agrees well with the experiment for both temperature and time duration. Individual modeling of each component, especially cellulose decomposition which is strongly endothermic, enables good prediction of the center temperature plateau. On the other hand, the temperature rise after the plateau is gradual and the magnitude of the center temperature peak exceeding surface temperature is smaller compared with the experiment. This is despite using the best-fit values for the heats of pyrolysis. Decomposition rates of the three components at the center of the sphere are plotted in Fig. 8. Hemicellulose and cellulose decomposition appear as two distinct peaks in sequential order. The cellulose decomposition peak corresponds to the center temperature plateau in Fig. 7. The results indicate that the first and second parts of the main mass loss period correspond to hemicellulose and cellulose decomposition respectively. In the experiment, the center temperature plateau corresponds to the second part of mass loss period. Therefore, the center temperature plateau is caused by endothermic cellulose decomposition
kg2 kt
were
hL 350kJ / kg by fitting to the experimental result.
gas
kg
(B) Case B: primary and secondary decomposition.
h2 100kJ / kg ,
(A) Case A: primary decomposition
tar
kc
decompositions
(36)
Also Q of Eq. (26) is modified as:
Q kt ht k g hg kc hc a .
Temperatures at the center, middle and surface for case A and B are compared in Fig. 5. The temperature difference between case A and case B is hardly noticeable, implying that the exothermicity of the secondary tar decomposition is not significant to cause the center temperature peak to exceed the surface temperature as observed in the experiments. Therefore, the center temperature peak in the experiment is caused by exothermic reactions other than secondary tar decomposition. Lignin decomposition The exothermic lignin decomposition has been presumed to be responsible for the exothermic thermal behavior in the final stage of wood pyrolysis in several cases found in the literature [5,6]. The effect of exothermic lignin decomposition has been studied using model-2 which consists of hemicellulose, cellulose and lignin pyrolysis models. The kinetics parameters of each
7
From the relationship between char yield and the furnace temperature, it is evident that the char producing process prefers lower temperature, whereas volatiles generation is promoted at higher temperature. Thus, the char producing reaction begins earlier than the volatiles generation. However, the exothermic reaction related to char occurs in the final stage of pyrolysis. Therefore, the exothermic reaction cannot be the primary char forming reaction which is competing with the volatiles producing reaction from virgin wood. Hence, it is concluded that the char forming process satisfying the experimental observations should be two step reaction whereby an intermediate solid is introduced. The first reaction is virgin solid to intermediate solid conversion which is competing with volatiles forming reaction and prefers lower temperature. This non-exothermic reaction determines the char yield. The second reaction is the exothermic conversion of the intermediate solid to char or both char and gas as in model-3. On the other hand, the center temperature plateau of model-3 is not as distinct as that of model-2, because wood is modeled as a single material as in model-1. However, satisfactory model prediction for both the thermal behavior and mass loss were achieved by model3. The results of model-3, showed that the exothermicity appearing after the endothermic reaction of wood pyrolysis is attributed to exothermic intermediate solid decomposition to char and gas.
and hemicellulose decomposition is less endothermic than cellulose. Lignin decomposition begins earlier and lasts longer than the other two components as it occurs over a wide temperature range. Since the lignin decomposition rate is less sensitive to temperature than other components, more lignin is left after the completion of endothermic cellulose decomposition for the 831K case than the 688K case. As a result, more exothermic effect appears for 831K than 688K in the calculations. This is not consistent with the experiment that shows a more distinct center temperature peak for 688K. Further, the center temperature rise after the temperature plateau is gradual, owing to steady exothermic lignin decomposition. Since the center temperature rise for model-2 is different from that of the experiments, it is concluded that the exothermic behavior observed in the experiment is not primarily caused by exothermic lignin decomposition. Intermediate solid decomposition In cellulose pyrolysis, exothermic reaction after the endothermic reaction has been reported, with the exothermic reaction being attributed to the exothermic decomposition of dehydrocellulose to char and gas by several researchers [8, 9]. Since cellulose comprises approximately half of the wood mass, it is plausible to assume wood pyrolysis process as similar to cellulose pyrolysis. The effect of the exothermic intermediate solid (corresponding to dehydrocellulose in cellulose pyrolysis) decomposition to char and gas was studied using model-3 to determine its role in the exothermic behavior of the wood pyrolysis. Kinetic parameters used in model-3 are listed in Table 7. Activation energy E1 is taken from the
5. Modeling of endo-/exothermic wood pyrolysis A new wood pyrolysis model is proposed based on the experimental and theoretical results of this work. This model should account for endo-/exothermic thermal behavior and solid mass loss measured in the experiment. In addition, the model should predict products yield satisfactorily for the application to liquid bio-fuel production. It has been known that gas yield increases at higher temperature, which is the opposite to char yield. In model-3, since both char and gas are produced from the intermediate solid at a fixed ratio between them, its gas yield is not realistic. Thus, it is more reasonable that the gas producing reaction is one of the primary competing reactions rather than the secondary decomposition of the intermediate solid. The proposed model has a modified form of model-1 which employs a two-step char producing reaction similar to model-3. However, unlike model-3, intermediate solid is converted into char only and gas is produced directly from virgin wood by a primary reaction. Kinetics parameters of primary and secondary tar decomposition
tar producing reaction of model-1. Other kinetic parameters and h are found by fitting with the experimental results. Solid mass fraction of model-3 shows good agreement with the experiments as shown in Fig. 9, except for the 638K case. The center temperature peak of model-3 agrees well with the experiment as shown in Fig. 10. In addition, model-3 shows a decrease of the temperature peak at higher furnace temperature as observed in the experiments. The concept of the exothermic secondary decomposition of the intermediate solid is plausible for the following reasons. (i) The center temperature peak appears after the active mass loss process and only small or negligible mass loss occurs during the exothermic reaction period. This indicates that the reactant of the exothermic reaction is in the solid phase and the product of exothermic reaction is also mainly in the solid phase, i.e. char. (ii) Exothermicity increases with the increase of char yield. (iii) The endothermic reaction and the exothermic reaction occur in consecutive order. (iv) It is possible that thermodynamically the solid rearranges to a more stable form (char) releasing energy.
reactions taken from model-1, except for
Ais which is
adjusted by experimental mass loss data from this work. Kinetic parameters of reaction c come from reaction 3 of model-3. Kinetic parameters of the model are listed in Table 8.
8
kt
wood kis
wood pyrolysis. On the other hand, low temperature hemicellulose decomposition in the beginning stage and high temperature lignin decomposition in the final stage are not taken into account well in proposed model. Large solid mass discrepancy between the model and the 638K experiment is attributed to the incapability of the proposed model to model low temperature pyrolysis, which may be acceptable for bio-oil production. Final product fractions of the proposed model are plotted in Fig. 13. As temperature increases, more volatiles and less char are produced. For furnace temperatures higher than 750K, secondary tar decomposition influences tar and gas yields. Tar yield for the model decreases for temperature higher than 800K due to secondary tar decomposition. Fig. 12 shows that the model predictions of the center temperature compare well with the experiments. The center temperature peaks are well predicted by employing the two-step char producing reactions. The center temperature plateaus of the proposed model are not as distinct as shown in the experiment. The approximation of single wood reactant leads to the gradual center temperature change during endothermic pyrolysis rather than distinct plateau shape observed during the experiments.
gas
kg
kg2 tar
kc2
intermediate solid
char kc
By introducing intermediate solid to char producing reaction, mass equations of virgin wood, intermediate solid, char are modified as shown below: a Virgin wood: (37) S a kt k g kis a t is (38) Intermediate solid : S is k is a k c is t Char:
c t
S c kc is kc 2 t
(39)
The tar and gas partial pressure equations are the same as those of model-1. Energy equation is the same as that of model-3. However, of Eq. (28) Q Error! Reference source not found. is modified as:
Q kt ht k g hg kis his a kc hc is k c 2 hc 2 k g 2 hg 2 t
6. Pressure generation Pressure variations at wood sphere center for furnace temperature range T fn 638 K ~ 1180 K were shown in
.
The primary reactions t, g and is have the same endothermic heat of pyrolysis, ht hg his 80 kJ / kg and the second step
Fig. 14. The magnitude of pressure peak increases with the furnace temperature. It is noticed that pressure profiles have different shapes for different furnace temperatures. At high furnace temperature, center pressure rises steadily until pressure peak and then it drops rapidly to ambient pressure. At low furnace temperatures, the center pressure rises to the pressure peak rapidly but drops to ambient pressure gradually. This observation indicates that pressure peak occurs at different stage of pyrolysis depending on the furnace temperature. Center and middle pressures are shown with solid mass fraction at T fn 638 K and T fn 879 K in Fig. 15 and Fig. 16. At
exothermic reaction c has hc of 300 kJ / kg . These values were determined by matching with the experiments. The kinetic parameters and h of tar cracking reactions are the same as those used in model-1. Solid mass fraction for the proposed model is compared with the experiments in Fig. 11. In general, a good agreement of solid mass loss was achieved between the proposed model and the experiments except for the 638K case. 736K and higher temperature cases show better agreement with the experiment than lower temperature cases. However, the solid mass loss characteristics of the beginning and the final stages shown in the experiment are not predicted well by the proposed model. Solid mass of the proposed model remains constant until the initial mass loss begins, whereas slight solid mass loss was observed from the beginning in the experiments. In the final stage of wood pyrolysis, the proposed model shows a distinct end of solid mass loss, whereas a gradual solid mass loss was observed in the experiments. This discrepancy is attributed to the limitations of the 1st order Arrhenius type kinetics scheme focusing on the primary pyrolysis process [26]. Consequently, the kinetics scheme better represents the cellulose decomposition which is the main part of the
T fn 638 K , pressure peak appeared at 11.4% . In
contrast, at T fn 879 K , pressure peak occurred at
79.9% . Although the center pressure is always higher than the middle pressure, the reason for the pressure peaks occurring at different extents of pyrolysis may be explained as follows: At lower temperatures, the sample is heated slowly and the temperature gradient along radial direction is small. This results in a wide pyrolysis zone in the radial direction. Since the virgin solid permeability is significantly smaller than char and the permeability is slowly increasing everywhere due to partial pyrolysis, pressure peak occurs at lower extent of pyrolysis despite the fact that relatively small amount of volatiles are generated. Thus, the pressure peak is caused 9
by small permeability and not by large volatile generation. At higher temperatures, the pyrolysis zone is relatively thin and the surface has much higher permeability than the center. Thus the resistance offered by the partially pyrolyzed solid is small. Pressure build up is thus the result of large volatiles flow in thick char and thin partially pyrolyzed wood. As a result, the pressure peak appears near the end of pyrolysis when volatiles have to travel the longest path through the char. In the experiment, sample split was observed for some cases especially at high temperatures. Fig. 14 shows these events along with the center pressure. The event of sample split was identified by either a sudden mass loss of the sample or a sudden change of temperature measurement when the center thermocouple was exposed to furnace atmosphere. This split is caused by internal pressure, cracks due to non-uniform shrinkage, and structural weakness due to charring. Fig. 14 shows the relation between internal pressure and sample split. The split did not occur for low temperature cases T fn 638 K , 688 K . At low temperature, the
confirms that exothermic intermediate solid decomposition is responsible for the center temperature peak. The contributions of the secondary tar decomposition and lignin decomposition to the center temperature peak are small. Further, the results of the mixture model (model-2) indicate that endothermic cellulose decomposition is responsible for the center temperature plateau. A new wood pyrolysis model is proposed based based on the experimental and theoretical study of this work. The model consists of three endothermic parallel reactions for tar, gas and intermediate solid, a subsequent exothermic reaction of the intermediate solid to char conversion and the secondary tar decomposition mechanism. Comparison between the model and experiments show good agreement with both the temperature and the solid mass loss measurements. Pressure calculations with the new pyrolysis model revealed that the internal pressure build up mechanism is different for different furnace temperatures. Further, at high temperatures, a thick wood particle might split by combination of high internal pressure and weakened structure. On the other hand, a thick wood particle does not split during low temperature pyrolysis. Therefore, pressure becomes more important for the high temperature fast pyrolysis process.
pressure peak is small and appears in early stage of pyrolysis when the sample structure has not been weakened. In the mid temperature region, the sample split occurs at the end of the pyrolysis when the most of sample was nearly all converted into char. Thus in the mid temperature region, the sample split is dominated by structural weakness rather than internal pressure. At high temperature T fn 879 K , the sample split event was
Acknowledgments The research at the University of Michigan was supported by NIST grant # 3D1086. References
coincident with the pressure peak when its structure has weakened and could not endure the internal pressure. Sample split is not shown for temperatures higher than T fn 879 K because experiment was not performed in
1. Huber, G. W. Ed., National Science Foundation Report (2008), “Breaking the Chemical and Engineering Barriers to Lignocellulosic Biofuels: Next Generation Hydrocarbon Biorefineries,” Chemical, Bioengineering, Environmental, and Transport Systems Division. Washington D.C. 180 p. 2. R. D. Perlack, L. L. Wright, A. F. Turhollow, R. L. Graham, B. J. Stokes, and D. C. Erbach, “Biomass as feedstock for a bioenergy and bioproducts industry: the technical feasibility of a billion-ton annual supply. A joint study sponsored by U.S. Department of Energy and U.S. Department of Agriculture.” April, 2005. 3. J. Fargione, J. Hill, D. Tilman, S. Polasky, and P. Hawthorne, “Land Clearing and the Biofuel Carbon Debt,” Published online February 7, 2008; 10.1126/science.1152747. 4. Huber, G. W., Iborra, S and Corma, A. (2006) “Synthesis of Transportation Fuels from Biomass: Chemistry, Catalysts, and Engineering,” American Chemical Society, Published on Web 06/27/2006. 5. Bilbao, R., et al., Modelling of the pyrolysis of wet wood. Journal of Analytical and Applied Pyrolysis, 1996. 36(1): p. 81-97.
those temperature ranges. However, it is expected that the sample would be split at pressure peak or before reaching the pressure peak. In that case, internal pressure might not reach the predicted pressure peak. 7. Conclusions Pyrolysis of a wood spheres has been studied both experimentally and theoretically. Temperature measurements show two distinct thermal regimes during wood pyrolysis. First, endothermic reaction causes a center temperature plateau. During this period, a large amount of solid mass loss occurs. Second, a steep temperature rise and the center temperature peak exceeding surface temperature occurs after the center temperature plateau. A small or negligible amount of solid mass loss occurs during this exothermic reaction period. Therefore, the endothermic period corresponds to primary reactions of active volatiles generation, especially cellulose decomposition, and the exothermic period corresponds to intermediate solid conversion to final char. Numerical study of various exothermic mechanisms
10
Journal of Applied Polymer Science, 1981. 26(5): p. 1533-1553. 21. Sjostrom, E., Wood Chemistry Fundamentals and Applications. 1981, New York, NY: Academic Press. 22. Kilzer, F.J. and A. Broido, Speculation on the Nature of Cellulose Pyrolysis. Pyrodynamics, 1965. 2: p. 151. 23. Mok, W.S.L. and M.J. Antal, Effects of pressure on biomass pyrolysis. II. Heats of reaction of cellulose pyrolysis. Thermochimica Acta, 1983. 68(2-3): p. 165186. 24. Galgano, A. and C.D. Blasi, Modeling Wood Degradation by the Unreacted-Core-Shrinking Approximation. Ind. Eng. Chem. Res., 2003. 42(10): p. 2101-2111. 25. Di Blasi, C., Heat, momentum and mass transport through a shrinking biomass particle exposed to thermal radiation. Chemical Engineering Science, 1996. 51(7): p. 1121-1132. 26. Di Blasi, C. and C. Branca, Kinetics of Primary Product Formation from Wood Pyrolysis. Ind. Eng. Chem. Res., 2001. 40(23): p. 5547-5556. 27. Di Blasi, C., Analysis of convection and secondary reaction effects within porous solid fuels undergoing pyrolysis. Combustion Science and Technology, 1993. 30: p. 315-340. 28. Liden, A.G., F. Berruti, and D.S. Scott, A kinetic model for the production of liquids from the flash pyrolysis of biomass. Chemical engineering communications 1988. 65: p. 207-221. 29. Park, W.C., A Study of Pyrolysis of Charring Materials and Its Application to Fire Safety and Biomass Utilization, in Mechanical Engineering. 2008, The University of Michigan: Ann Arbor. 30. Chan, W.-C.R., M. Kelbon, and B.B. Krieger, Modelling and experimental verification of physical and chemical processes during pyrolysis of a large biomass particle. Fuel, 1985. 64(11): p. 1505-1513. 31. Lee, C.K., R.F. Chaiken, and J.M. Singer, Charring Pyrolysis of Wood in Fires by Laser Simulation. Proceedings of the Combustion Institute, 1976: p. 1459-1470. 32. Di Blasi, C., Modelling the fast pyrolysis of cellulosic particles in fluid-bed reactors. Chemical Engineering Science, 2000. 55(24): p. 5999-6013. 33. Kansa, E.J., H.E. Perlee, and R.F. Chaiken, Mathematical model of wood pyrolysis including internal forced convection. Combustion and Flame, 1977. 29: p. 311-324.
6. Di Blasi, C., et al., Pyrolytic behavior and products of some wood varieties. Combustion and Flame, 2001. 124(1-2): p. 165-177. 7. Koufopanos, C.A., et al., Modelling of the pyrolysis of biomass particles. Studies on kinetics, thermal and heat transfer effects. Canadian Journal of Chemical Engineering, 1991. 69(4): p. 907-915. 8. Milosavljevic, I., V. Oja, and E.M. Suuberg, Thermal effects in cellulose pyrolysis: relationship to char formation processes. Industrial & Engineering Chemistry Research, 1996. 35(3): p. 653. 9. Strezov, V., B. Moghtaderi, and J. Lucas, Thermal study of decomposition of selected biomass samples. Journal of Thermal Analysis and Calorimetry, 2003. 72(3): p. 1041-1048. 10. Gronli, M.G. and M.C. Melaaen, Mathematical model for wood pyrolysis - comparison of experimental measurements with model predictions. Energy and Fuels, 2000. 14(4): p. 791-800. 11. Milosavljevic, I. and E.M. Suuberg, Cellulose thermal decomposition kinetics: global mass loss kinetics. Industrial & Engineering Chemistry Research, 1995. 34(4): p. 1081-1091. 12. Fredlund, B., A Model for Heat and Mass Transfer in Timber Structures During Fire, A Theoretical, Numerical and Experimental Study, in Report LUTVDG/(TVBB-1003). 1988, Lund University: Sweden. 13. Baum, H.R. & Atreya, A., A model of transport of fuel gases in a charring solid and its application to opposed-flow flame spread, Proceedings of the Combustion Institute, 2007. 31(2): p. 2633-2641. 14. Park ,W.C., Atreya, A. and Baum, H.R., Numerical study of thermal decomposition and pressure generation in charring solids undergoing opposedflow flame spread, Proceedings of the Combustion Institute, 2007. 31(2): p. 2643-2652. 15. Siegel, R. and J.R. Howell, Thermal Radiation Heat Transfer. Fourth ed. 2002, New York, NY, U.S.A: Taylors & Francis. 16. Di Blasi, C., Modeling Intra- and Extra-Particle Processes of Wood Fast Pyrolysis, AICheE Journal 2002. 48(10): p. 2386-2397. 17. Mohan, D., C.U. Pittman Jr, and P.H. Steele, Pyrolysis of wood/biomass for bio-oil: A critical review. Energy and Fuels, 2006. 20(3): p. 848-889. 18. Varhegyi, G., et al., Kinetics of the thermal decomposition of cellulose, hemicellulose, and sugar cane bagasse. Energy & Fuels, 1989. 3(3): p. 329-335. 19. Capart, R., L. Khezami, and A.K. Burnham, Assessment of various kinetic models for the pyrolysis of a microgranular cellulose. Thermochimica Acta, 2004. 417(1): p. 79-89. 20. Chan, R.W.C., Krieger, B.B., Kinetics of dielectricloss microwave degradation of polymers: Lignin.
11
800
1 0.9
exhaust computer
surface temperature 700
0.8
solid mass fraction (Y)
insulatingtop cap
insulation heating elements
sample holder temperature measurement
heater controller
biomass sample
0.7 600
0.6
center temperature 0.5 500
0.4
solid mass fraction (Y) 0.3
middle temperature
0.2
thermocouples
temperature (K)
scale
400
0.1 0 insulating bottom cap
0
argon tank
sonic orifice
200
400
600
800
1000
300 1200
time (s)
Fig. 4: Solid mass fraction and temperatures of wood sphere pyrolysis at 688K.
pressure valve gauge
Fig. 1 Schematic of the experimental apparatus
900
1
surface
800
0.9
temperature (K)
solid mass fraction (Y)
0.8 0.7 0.6
638K
0.5
688K
0.4
700
case A
middle
case B center
600
500
783K 736K
0.3
400
831K
0.2
879K 0.1 0
300 0
200
400
600
800
0
100
200
300
400
500
600
time (s)
1000
Fig. 5: Temperature comparison between case A and case B of model-1 at center, middle and surface; furnace temperature: 831K
time (s)
Fig. 2 Solid mass fraction measurements, the average furnace surface temperature is indicated.
1 900
surface
879K
0.9
831K
0.8
800
solid mass fraction (Y)
783K
temperature (K)
736K 700
688K 638K 600
500
0.7
0.5
688K
0.4
783K
400
736K
879K
0.3
831K
0.2
center
0.1 0
300
638K
0.6
0
200
400
600
800
1000
1200
0
200
400
600
800
time (s)
time (s)
Fig. 6: Solid mass fraction comparison between model-2 and experiment (solid line: model-2, dotted line: experiment).
Fig. 3: Temperature measurements at the center and the surface of the sample.
12
1000
900
879K 831K
900
879K
800
783K
831K 800
736K
temperature (K)
783K 736K
temperature (K)
700
688K 638K 600
500
700
688K 638K 600
500
400
400
300
0
200
400
600
800
1000
1200
time (s) 300
0
200
400
600
800
1000
1200
Fig. 10: Center temperature comparison between model3 and experiment (solid line: model, dotted line: experiment)
time (s)
Fig. 7: Center temperature comparison between model-2 and experiment (solid line: model-2, dotted line: experiment).
1 0.9
9
7
3
1.5
1
0.5
solid mass fraction (Y)
2
0.8
hemicellulose cellulose lignin
8
decomposition rate (kg/m s)
hemicellulose cellulose lignin
3
decomposition rate (kg/m s)
2.5
6 5 4 3 2 1
0
0
100
200
300
400
500
600
700
800
0
900 1000
0
100
200
time (s)
300
400
0.6
638K
0.5
688K
0.4
783K
0.3
736K
879K
500
831K
0.2
time (s)
Figure 8 Decomposition rate of three components at the center of the wood sphere; Furnace temperature (a) 688K, (b) 832K, Model-2.
0.1 0
0
200
400
600
800
1000
time (s)
1
Fig. 11: Solid mass fraction comparisons between proposed model and experiment (solid line: proposed model, dotted line: experiment)
0.9 0.8
900
0.7
879K
638K
831K
0.6 800
0.5
783K
688K 736K
0.4
temperature (K)
solid mass fraction (Y)
0.7
783K
0.3
879K
736K
831K
0.2 0.1 0
0
200
400
600
800
700
688K 638K 600
500
1000
time (s)
400
Figure 9 Solid mass fraction comparison between model-3 and experiment (solid line: model, dotted line: experiment)
300
0
200
400
600
800
1000
1200
time (s)
Figure 12 Center temperature comparisons between proposed model and experiment (solid line: proposed model, dotted line: experiment)
13
1
1.45
Y
0.6
1.40
Pc / P0
Tar
0.8
solid mass fraction (Y)
0.5
n o tic ra 0.4 F s tc u 0.3 d ro P l a 0.2 in F
Char
Gas
0.1
1.35 1.30
0.6
Pm / P0
1.25 1.20
0.4 1.15 1.10
0.2
0 600
650
700
750
800
850
1.05
900
Furnace Temperature (K)
0 0
Figure 13 Final product fractions of proposed model 1.6
pressure
0.7
70
140
210
1.00 350
280
time (s)
Figure 16 Pressures at center and middle locations and solid mass fraction (Y) at furnace temperature 879 K
1180K 1080K
1.5
Table 1 Model-1 kinetics parameters and heats of pyrolysis
980K 879K 831K 783K 736K
P/P0
1.4
1.3
Reaction
c
Ai s1
688K
Ei
J / mol
64
b
152,700
d
64
g2 5b
1.00 x 10
a
148,000
d
c2 9a
4.38 x 10
a
111,700 64
g 10a
1.08 x 10
a
hi kJ / kg
638K
t 6a
3.27 x 10
108,000
d
6c
4.28 x 10
c
108,000
e
-42
-42
1.2
a:[26], b:[27], c:[28], d:[29], e: [30] 1.1
1.0
Table 2 Model-2 hemicellulose pyrolysis model parameters [18] 0
250
500
750
Reaction
1000
time (s)
Ei
Figure 14 Center pressure of proposed model (solid lines: pressure P/P0, symbols: sample crack during experiments). 1
1.26 x 10 96,000
0.56
0.45
Tbl 3 Mod-2 cellulose pyrolysis model parameters [19] Ai s
1.20
Ei 1.16
Y 1.12
0.4
pressure
solid mass fraction (Y)
J / mol
7
7.94 x 10 195,000
Reaction
0.8
2 16
1.24
Pc / P0
0.6
1
z
1
A
B
7.94 x 10
J / mol
16
7
1.26 x 10
202,650
255,000
f
0.821
0.179
0.087
0.087
n
1
22
m
0.481
1
1.08
Table 4 Model-2 lignin pyrolysis model parameters [20]
Pm / P0
0.2
1.04
Reaction Ai s
0 0
300
600
900
1200
1.00 1500
Ei
time (s)
14
J / mol
Figure 15 Pressures at center and middle locations and solid mass fraction (Y) at furnace temperature 638K
1
L 5
5.09 x 10 95,000 0.335
e
Table 5 Material properties and kinetics parameters property
value
w
source measure
630 (kg / m3 )
Cw
1500 1.0T ( J / kgK )
Cc
420 2.09T 6.85 10 -4 T 2
C pt C pg
100 4.4T 1.57 10 3 T 2
C pv
0.85cT 0.15cG 5 105 1 1 104 m
[10]
770 0.629T 1.91 104 T 2
d e
( J / kgK ) J / kgK J / kgK
[10] [10] [10] estimate [10] [30]
1
5.67 108
(W / m2 K4 )
W / mK
w , r ad ial
0.1046
w , g r ain
0.255
w , ta ng en tial
0.255
c , ra dia l
0.071
c , gr ain
0.105
c , ta ng entia l
0.105
c , ta ng entia l
0.105 0.0258 (W / mK )
Estimat [27]
Bw
5 10 16 (m 2 )
Bc
1 10 13 ( m 2 )
ew ec
0.7
[32] [32] [10]
0.92
[10]
20 (W / m2 K ) 3.0105 (kg / m s )
estimate [33]
Mg
0.038 kg / mol
[10]
Mt
0.11 kg / mol
[10]
Mv
0.076 kg / mol 8.314 J / mol K
[12]
h
R
[31]
W / mK W / mK
[31] estimate
W / mK W / mK W / mK W / mK
[31] [31] Estimat
Table 6 Furnace and gas temperatures T fn
638
688
736
783
831
879
Tg
585
635
670
720
770
820
Table 7 Model-3 kinetic parameters and Δh Reaction Ai s 1 Ei
1
2 10
J / mol
h kJ / kg
3 7
10
2.00 x 10
2.51 x 10
1.38 x 10
148,000
117,000
161,000
110
0
-210
Table 8 Proposed model kinetic parameters and Δh Reaction Ai s
1
J / mol hi kJ / kg Ei
t
g 10
1.08 x 10
4.38 x 10
148,000 80
is 9
c 6
10
3.75 x 10
1.38 x 10
152,700
111,700
161,000
80
80
-300
15
