Energy Education Science and Technology Part A: Energy Science and Research

2012 Volume (issues) 29(1): 419-426

Product yields and kinetics of

pyrolysis of sawdust and bagasse particles
L. Kapoor, A. S. Chaurasia*
Jaypee University of Engineering & Technology (JUET), Department of Chemical Engineering, Guna-473226
(Madhya Pradesh), India

Received: 19 June 2011; accepted: 04 August 2011

Abstract
Pyrolysis is one of the thermochemical biomass conversion processes where thermal destruction of
organics in the absence of oxygen happens. The destructed portion comprises of high energy content and
numerous organic content which leads to the possibility of extraction of energy and chemicals from
biomass. Pyrolysis generates three different products in different quantities: char, tar and gases, the
relative proportions of which depend on type of pyrolysis, the characteristics of biomass and the reaction
parameters. For engineering application knowledge of the pyrolysis kinetics is essential for predicting the
pyrolysis behavior of biomass materials as well as designing the suitable reactor. In the present study,
formation of gaseous volatiles released during primary pyrolysis of sawdust and bagasse samples has been
studied. The average particle size of sawdust and bagasse samples used in this study is 215 µm.
Experiments have been carried out in muffle furnace to estimate the kinetic parameters of gaseous
volatiles for both the samples. The temperature range used in this study is 500 0C - 800 0C with residence
time varying from 60 s – 600 s. The kinetic scheme of biomass decomposition by single reaction giving
gaseous volatiles and char is used. The proposed single reaction model is simulated and the best values of
the kinetic parameters i.e. activation energy and Arrhenius constant is found by using the two dimensional
surface fitting non linear regression algorithm. The kinetic parameters obtained in this study has been
compared with the earlier researchers and found to be matching well.
Keywords: Pyrolysis; Biomass; Kinetic modeling; Kinetic scheme
©Sila Science. All rights reserved.

1. Introduction
Biomass includes all water and land based vegetation and trees, all waste biomass such as
municipal solid waste (MSW), forestry and agricultural residues and certain types of industrial
wastes. Biomass provides a clean renewable energy source that could dramatically improve our
environment, economy, and energy security. The energetic and industrial usage of biomass is
becoming more and more technologically and economically attractive. Wood, woody and other
_______________
*
Corresponding author. Tel.: +91-754-426-7051; fax: +91-754-426-7011.
E-mail addresses: [email protected]; [email protected] (A. S. Chaurasia).
420 L. Kapoor, A. S. Chaurasia / EEST Part A: Energy Science and Research 29 (2012) 419-426

forms of biomass are some of the main renewable solid energy resources available and provide
the only source of liquid, gaseous and solid fuels [1-16]. The conversion processes of biomass
usually involve a reduction of the water content of the material, resulting in the simultaneous
increase in its thermal value and preservation potential, and in an improving of the handling
characteristics of the biomass, for example turning it into a fluid, which may be either gas or
liquid. Wood and other biomass can be treated in a number of different ways to provide such
fuels. In general such methods are divided into biological and thermal. Pyrolysis is one of the
thermochemical biomass conversion processes where thermal destruction of organics in the
absence of oxygen takes place resulting in the formation of three different products: char, tar, and
gases.
In the present study, the kinetic parameters has been estimated for gaseous volatiles by
performing pyrolysis of saw dust and bagasse samples for the temperature range of 500 °C – 800
°C and a residence time of 1 minute – 10 minute. Single reaction model is simulated to estimate
the kinetic parameters by using the two dimensional surface fitting non linear regression
algorithm.
2. Experiments
2. 1. Materials and experimental equipment
The materials selected for this study are saw dust and bagasse. The average particle size of the
sample is 215 μm. The sample is dried in an oven at 105 ºC for 8 hour before experiment. To
perform the pyrolysis of sawdust and bagasse samples, muffle furnace as shown in Fig. 1 is used.
It is a laboratory scale furnace that works at atmospheric pressure and temperature up to 1100 ºC.
Outer body is made of thick mild steel. Furnace is lagged with ceramic wool and hysil blocks for
insulation. Heaters are made of 80/20 Nichrome wire, uniformly wound on the muffle. Mild
steel control box is fitted on the top of the furnace. It consists of indicating lamps, on/off switch,
contactor and digital temperature controller along with Cr/Al thermocouple sensor.

Fig. 1. Muffle furnace.

2. 2. Experimental procedure
Experiments are carried out to determine the kinetic parameters for gaseous volatiles in a
muffle furnace. The operating conditions are:
• Initial temperature: Room temperature up to 303 K
• Final temperature range: 773 K – 1073 K
• Holding (residence) time range upon attaining final temperature: 1 minute – 10 minute.
The predefined weight of the sample is taken in the crucible and kept in the furnace upon
attaining the final temperature for different residence time. At the end of each experiment,
 L. Kapoor, A. S. Chaurasia / EEST Part A: Energy Science and Research 29 (2012) 419-426 421
crucible is weighed to determine the yield of gaseous volatiles released and char formed. Most of
the runs are replicated to verify the obtained values. The representative yield of gaseous volatiles
and char for different extents residence time and temperature for sawdust and bagasse (average
particle size: 215 μm) are given in Tables 1–4.
Table 1. Experimental data on volatiles plus gases yield at different residence time and
temperatures in terms of weight percentage of sawdust
Residence time (sec) Volatiles plus gases yield (Expressed as wt% of sawdust)
773 K 873 K 973 K 1073 K
60 0.6 1.4 3.4 9
120 2.2 4.2 18 39.8
180 4.2 13 36.8 59.8
240 6 17.8 65.6 71.8
300 19.6 58.4 69.8 73
360 33.2 67.4 71.6 73.4
420 37.8 67.2 72.6 74.8
480 40 69.2 72.6 74.8
540 61.4 69.2 72.6 74.8
600 65.2 69.2 72.6 74.8

Table 2. Experimental data on volatiles plus gases yield at different residence time and
temperatures in terms of weight percentage of bagasse.
Residence time (sec) Volatiles plus gases yield (Expressed as wt% of bagasse)
773 K 873 K 973 K 1073 K
60 2.2 4.8 14.6 30.8
120 5.0 16.2 53.6 78.8
180 5.8 45 76.0 80.4
240 8.6 58 78.0 81.0
300 15.2 74.2 78.8 81.2
360 46.8 76.8 79.4 81.2
420 62.2 77.4 80.4 81.2
480 65.6 77.6 80.4 81.2
540 73.0 77.6 80.4 81.2
600 73.2 77.6 80.4 81.2

Table 3. Experimental data on char plus unreacted biomass yield at different residence time
and temperatures in terms of weight percentage of sawdust.
Residence time (sec) Char plus unreacted biomass yield (Expressed as wt% of sawdust)
773 K 873 K 973 K 1073 K
60 99.4 98.6 96.6 91.0
120 97.8 95.8 82.0 60.2
180 95.8 87.0 49.8 40.2
240 94.0 82.2 34.4 28.2
300 80.4 41.6 30.2 27.0
360 66.8 32.8 28.4 26.6
420 62.2 32.6 27.4 26.6
480 60.0 30.8 27.4 26.6
540 38.6 30.8 27.4 26.6
600 34.8 30.8 27.4 26.6

3. Kinetic modeling and simulation

Biomass pyrolysis involves numerous extremely complex reactions and end up with large
number of intermediates and end products, devising an exact reaction mechanism and kinetic
modeling for biomass pyrolysis is extremely difficult, hence, pyrolysis models are modeled on
the basis of visible kinetics. Thus, pyrolysis kinetic modeling paved the way for researchers [1-
8] to propose different reaction schemes in different times. Hence, even today it is difficult to
develop a precise kinetic model taking into account all the parameters concerned.
422 L. Kapoor, A. S. Chaurasia / EEST Part A: Energy Science and Research 29 (2012) 419-426

Table 4. Experimental data on char plus unreacted biomass yield at different residence time
and temperatures in terms of weight percentage of bagasse.
Residence time (sec) Char plus unreacted biomass yield (Expressed as wt% of bagasse)
773 K 873 K 973 K 1073 K
60 97.8 95.2 85.4 69.2
120 95.0 83.8 46.4 21.2
180 92.2 55.0 24.0 19.6
240 91.4 42.0 22.0 19.0
300 84.8 25.8 21.2 18.8
360 53.2 22.6 20.6 18.8
420 37.8 22.4 19.6 18.8
480 34.4 22.4 19.6 18.8
540 28.0 22.4 19.6 18.8
600 26.8 22.4 19.6 18.8

Kinetic parameters usually include activation energy, pre exponential factor and order of
reaction, which describe the rate of mass loss. It is necessary to understand the kinetics of
pyrolysis in order to design a suitable pyrolysis reactor.
The one stage kinetic scheme for biomass pyrolysis given below is utilized in our study. Here
the biomass gives volatiles plus gases and char as the primary pyrolysis products.

Volatiles + Gases
(G2)
Biomass

Char

The experimental data are then fitted to the kinetic equation [9].

dVi
= k i (Vi * − Vi ) (1)
dt
Here Vi is the yield of material i at any time t , Vi* is the ultimate value of Vi at long
residence time and high temperature (determined as a part of fitting procedure by including
experimental data at high conversion) and k i is the rate constant.
−E
k i = Ai e RT
(2)
Here Ai is pre exponential factor, E i is the activation energy, R is the universal gas constant
and T is the temperature.
The rate of formation of individual product is modeled as first order in the difference between
the ultimate yield of that product and the amount of that product generated up to that time. The
kinetic parameters for volatiles plus gases and char are found by two dimensional surface fitting
non linear regression algorithm.
3. 1. Estimation of kinetic parameters
Kinetic parameters have been estimated by two dimensional surface fitting non linear
 L. Kapoor, A. S. Chaurasia / EEST Part A: Energy Science and Research 29 (2012) 419-426 423
regression algorithm as proposed by Xu and Kandiyoti [8]. By using the algorithm it is observed
that it searches close to the assumed starting value and are not able to reduce the error to larger
extent. We have modified last three steps of the original algorithm proposed by Xu and Kandiyoti
[8]. The modified algorithm is as follows:
(1) Initially, a pair of values of A and E are arbitrarily selected and the next set of
approximations for two parameters are calculated from
Ai = A0 + A0 (1 + cos i )
Ei = E0 + E0 (1 + sin  i )
where, A0 = A0 / N1 ; E0 = E0 / N 2 are the searching step lengths adjusted by changing the
values of N 1 and N 2 .
(2) Initially, N1 = 10.0 and N 2 = 10.0 .
 i = i(2 / N ) is the searching direction ( i = 1, 2, − −−, N )
For example, N = 8 represents the eight points surrounding the reference point.
(3) For each pair of kinetic parameters the weight loss (or conversion) data at any reaction
condition specifying holding temperature, residence time is determined by using Runge Kutta
fourth order method for single step reaction.
(4) The search for the best fitting kinetic parameters is performed by computing the standard
error for each pair of kinetic parameters
M

 (V
j =1
j − Vmj ) 2
i =
M
where,
V j = simulated weight loss calculated by using kinetic parameters ( Ai , Ei ).
Vmj = experimental data.
M = number of experimental data points.
 i = statistical least-squares standard error for Ai , Ei .
(5) Select the values of Ai and E i which give least  i from N points.
(6) The searching step lengths of A and E are then reduced by doubling N 1 and N 2 values
and the search loop is repeated until the difference between consecutive  i values are very
small.
(7) Take Ai and E i values from step (6) corresponding to least  i from N points. Use this
value Ai and E i in step (1).
(8) Repeat steps (1) – (7) until  i is reduced to maximum possible extent.

4. Results and discussion

The best fit kinetic parameters are estimated for volatiles plus gases for both the samples.
Wide range of simulations is carried out with different sets of initial guess values of frequency
factor and activation energy to estimate the kinetic parameters. It is observed that the standard
error decreases as the simulation progresses, gives the minimum value for a particular set of
424 L. Kapoor, A. S. Chaurasia / EEST Part A: Energy Science and Research 29 (2012) 419-426

initial guess value and then increases as shown in Figs. 2 and 3 for sawdust and bagasse
respectively. Different sets are then compared to find the lowest possible standard error between
the experimental value and model prediction. The kinetic parameters i.e. frequency factor and
activation energy is then chosen from a set which has lowest value of standard error.

Fig. 2. Standard deviation error curve for sawdust.

Fig. 3. Standard deviation error curve for bagasse.

The experimental and predicted values of gaseous volatiles for sawdust and bagasse are shown
in Tables 5 and 6 respectively considering wide range of temperatures and residence time. It is
seen that predicted values are matching well with the entire range of experimental data. The
kinetic parameters obtained in this study are compared with earlier researchers as shown in Table
7. It is seen that kinetic parameters for both the samples obtained in this study are in good
agreement with earlier researchers.
Figs. 4 and 5 show the weight loss for sawdust and bagasse samples as a function of
temperature. It can be seen that the amount of char decreases with increase in temperature. At
higher temperature the larger molecules present in liquid and residual solid are broken down to
produce smaller molecules, which enrich gaseous fraction. Also, cellulose depolymerizes at
higher temperature. Hence, the production of volatiles plus gases increases with increase in
temperature. It can be seen that the amount of volatiles plus gases released is more for the
bagasse sample than for the saw dust sample. This is because each biomass showed quite
different devolatilization behavior.
5. Conclusions
 It can be seen that the char yield decreases with higher reactor temperatures probably due
to higher conversion rates.
 Production of volatiles plus gases increases both with temperature and residence time.
 L. Kapoor, A. S. Chaurasia / EEST Part A: Energy Science and Research 29 (2012) 419-426 425

Table 5. Initial guess values: A = 104 s-1, E = 86470 J/mol, New values: Anew = 104.12 s-1,
Enew = 111764.932551 J/mol, δ =21.603037 (Sawdust sample).
Time (s) 773 K 873 K 973 K 1073 K

(VG)E (VG)P (VG)E (VG)P (VG)E (VG)P (VG)E (VG)P

60 0.6 1.7976 1.4 11.8995 3.4 41.9978 9 70.8155
120 2.2 3.5519 4.2 21.9059 18 60.4152 39.8 74.5878
180 4.2 5.2641 13 30.3205 36.8 68.4918 59.8 74.7887
240 6 6.9352 17.8 37.3965 65.6 72.0336 71.8 74.7994
300 9.8 8.5661 58.4 43.3468 69.8 73.5869 73 74.8000
360 33.2 10.1578 67.4 48.3505 71.6 74.2680 73.4 74.8000
420 37.8 11.7113 67.2 52.5582 72.6 74.5667 74.8 74.8000
480 40 13.2274 69.2 56.0965 72.6 74.6977 74.8 74.8000
540 61.4 14.7071 69.2 59.0719 72.6 74.7551 74.8 74.8000
600 65.2 16.1512 69.2 61.5740 72.6 74.7803 74.8 74.8000

Table 6. Initial guess values: A = 105 s-1, E = 80000 J/mol, New values: Anew = 105.27 s-1,
Enew = 121250.612852 J/mol, δ = 19.80798 (Bagasse sample).
Time (s) 773 K 873 K 973 K 1073 K

(VG)E (VG)P (VG)E (VG)P (VG)E (VG)P (VG)E (VG)P

60 2.2 5.5779 4.8 36.8934 14.6 78.2655 30.8 81.1999
120 5.0 10.7726 16.2 57.0242 53.6 81.0939 78.8 81.2000
180 5.8 15.6105 45 68.0086 76.0 81.1962 80.0 81.2000
240 8.6 20.1161 58 74.0021 78.0 81.1999 81.0 81.2000
300 15.2 24.3121 74.2 77.2725 78.8 81.2000 81.2 81.2000
360 28.2199 79.0570 81.2000 81.2000
46.8 76.8 79.4 81.2
420 62.2 31.8593 77.4 80.0307 80.4 81.2000 81.2 81.2000
480 65.6 35.2487 77.6 80.5620 80.4 81.2000 81.2 81.2000
540 73.0 38.4052 77.6 80.8519 80.4 81.2000 81.2 81.2000
600 73.2 41.3450 77.6 81.1010 80.4 81.2000 81.2 81.2000

Table 7. Comparison of kinetic parameters for gaseous volatiles obtained in this study with
earlier researchers.
Kinetic Parameters Boroson et al. [9] Diebold [10] Liden et al. [11] This study
Sawdust Bagasse
Temperature range (K) 773-1073 773-1073 773-1073 773-1073 773-1073
Pre-exponential factor (s-1) 1  104.98 1  105.19 1  106.49 1  104.12 1  105.27
Energy of activation (kJ mol-1) 93.3 87.5 108.0 111.76 121.251

Fig. 4. Weight loss vs. temperature curve for the sawdust sample.

 Kinetic parameters have been estimated by using two dimensional surface fitting non
linear regression algorithms.
 The parameters estimated are in good agreement with other researchers.
426 L. Kapoor, A. S. Chaurasia / EEST Part A: Energy Science and Research 29 (2012) 419-426

Fig. 5 Weight loss vs. temperature curve for bagasse sample.

References
[1] Prakash N, Karunanithi T. Advances in modeling and simulation of biomass pyrolysis. Asian J Sci
Res 2009;2:1-27.
[2] Sheth PN, Babu BV. Kinetic modeling for pyrolysis of hazelnut shell: optimal parameter estimation
using differential evolution (DE). Proceedings of International Symposium and 60 th Annual Session
of IIChE in association with International Partners (CHEMCON-2007), The Diamond Jubilee
Conference of IIChE, Heritage Institute of Technology, Kolkata, 2007;33.
[3] Sheth PN, Babu BV. Kinetic modeling of the pyrolysis of biomass. Proceedings of National
Conference on Environmental Conservation, BITS-Pilani, India, p 453, 2006.
[4] Ertas M, Alma MH. Slow pyrolysis of chinaberry (Melia azedarach L.) seeds: Part I. The influence of
pyrolysis parameters on the product yields. Energ Educ Sci Tech-A 2011;26:143–154.
[5] Olivella MA, De Las Deras FXC. Evaluation of linear kinetic methods from pyrolysis data of Spanish
oil shales and coals. Oil Shale 2008;25:227-245.
[6] Sonobe T, Worasuwannarak N. Kinetic analyses of biomass pyrolysis using the distributed activation
energy model. Fuel 2008;87:414-421.
[7] Babu BV, Chaurasia AS. Modeling, simulation and estimation of optimum parameters in pyrolysis of
biomass. Energy Convers Manage 2002;44:2135-2158.
[8] Xu B, Kandiyoti R. Two-stage kinetic model of primary coal liquefaction. Energy Fuels
1996;10:1115-1122.
[9] Boroson ML, Howard JB, Longwell JP, Peters AW. Products yields and kinetics from the vapor
phase cracking of wood pyrolysis tars. AIChE 1989;35:120-128.
[10] Diebold JP. The cracking kinetics of depolymerized biomass vapours in a continuous reactor. Thesis
T-3007, Colorado School of Mines, Golden, CO, 1985.
[11] Liden AG, Berruti F, Scott DS. A kinetic model for production of liquids from the fast pyrolysis of
biomass. Chem Eng Commun 1988;65:207-221.
[12] Demirbas A. Energy issues in energy education. Energ Educ Sci Tech-A 2011;27:209–220.
[13] Demirbas B. Sustainable biomass-to-energy business. Energ Educ Sci Tech-A 2011;28:453-458.
[14] Konur O. The evaluation of the research on the biofuels: a scientometric approach. Energ Educ Sci
Tech-A 2011;28: 903-916.
[15] Ak A. Waste management: Some energy recovery processes. Ener Educ Sci Technol Part-A
2012;28:1175-1182.
[16] Yumurtaci M, Kecebas A. Renewable energy and its university level education in Turkey. Energ
Educ Sci Tech-B 2011;3:143–152.