Prediction of pyrolysis kinetic parameters from biomass constituents based on simplex-lattice mixture design☆

Panusit Sungsuk ,Sasiporn Chayaporn ,Sasithorn Sunphorka ,2,Prapan Kuchonthara ,3,Pornpote Piumsomboon ,3,Benjapon Chalermsinsuwan ,3,*

1 Fuels Research Center,Department of Chemical Technology,Faculty of Science,Chulalongkorn University,254 Phayathai Road,Patumwan,Bangkok 10330,Thailand

2 Faculty of Engineering and Architecture,Rajamangala University of Technology Tawan-ok,Uthenthawai Campus,225 Phayathai Road,Bangkok 10330,Thailand

3 Center of Excellence on Petrochemical and Material Technology,Chulalongkorn University,254 Phayathai Road,Patumwan,Bangkok 10330,Thailand

1.Introduction

Biomass,containing energy from the sun acquired through photosynthesis,is organic matter thatconsists ofmostly carbon and hydrogen atoms.It is considered as a source of renewable energy that can be used in short cycle,unlike coal and petroleum.Pyrolysis is one of the most important processes to convert biomass to energy by its decomposition under an absence of oxygen[1].Among the thermochemical processes for transforming biomass into energy and products,pyrolysis is the simplest to set up as compared to combustion and gasification[2,3].The products of biomass pyrolysis consist of char,tar and gas.

To understand the mechanism and to design a suitable process for biomass pyrolysis,acknowledgment of the kinetic parameters,including the activation energy,frequency factor and reaction order,is very important.Therefore,various methods,analytical,differential and integral,have been developed to calculate the kinetic parameters using a thermogravimetric analyzer[4-12].These methods provide slightly differentvalues based on theirassumptions.Huang et al.indicated thatthe analyticalmodelderived by an assumption of n th-orderkinetics was acquirable,representative and reliable for biomass pyrolysis[4].Besides the calculation of kinetic parameters from thermogravimetric analyzer data,the prediction of kinetic parameters from biomass constituents has been explored by many studies[11,13-16].Gani and Naruse observed the effectof cellulose and lignin contents on pyrolysis characteristics[16].The authors reported thathigher cellulose contentprovided a faster pyrolysis rate.Peters attempted to describe thermal degradation ofpistachio shells by a detailed reaction mechanism[14].The results indicated that the simple model derived from kinetic values for pistachio shells pyrolysis and a discrete particle method based on biomass composition provided a good agreement between predicted values and experimental values.However,the proposed models were still quite complicated.Therefore,the development of a simplified model that can potentially predictthe thermalbehaviorand kinetic valuesisstillattractive.Currently,mathematical and statistical methods have been developed to examine the relationship between input-outputparameters.Standard mixture design methods such as the simplex-lattice mixture design(SLD)and response surface methodology(RSM)were used to evaluate the effect of three correlated factors on the desired response while other experimentaldesigns are usually based on independentfactors[17-20].When using the simplex-lattice mixture design,the summation of three proportions must be equal to one.Therefore,it is suitable for applying it to a biomass system.There is literature that has reported that the RSM with the simplex design could be used to determine the effectofmain factors and to interpretthe interactions between the mixture components[18,21].Liu et al.used the simplexlattice mixture design to determine the interaction between biomass components during pyrolysis[18].The effect of components and their interactions on mass loss rate could be observed.However,the other kinetic values were not investigated in their work.In addition,this field of research still lacks studies in which this kind of model is applied.

The aim of this study is to develop a simplified model for the prediction of kinetic values for biomass pyrolysis.Mixture design methods were applied for this purpose.Pure biomass constituents,including cellulose,hemicellulose and lignin and their mixtures,were used as the biomass model.Pyrolysis of pure and mixed biomass components,at the desired ratios based on the mixture design,were investigated and then kinetic values were examined.An analytical method was used to calculate the kinetic values from TGA data since it is simpler and gives more accurate values compared to other methods such as the Kissinger-Akahira-Sunose(KAS)and Ozawa-Flynn-Wall(OFW)methods[22].To obtain the kinetic values,the analytical method(for reactor order equal to one and others)was fitted to thermogravimetric analysis data to calculate the kinetic values.With respect to these proposed regression models and RSM methods,the effects of pure biomass components and their interactions were identified.The calculated kinetic parameters and generated models are provided in this study and discussed in detail.This study provides useful information for simulation of biomass related processes.

2.Experimental

2.1.Materials

Pure biomass components,including cellulose(α-cellulose,catalog number C8002),hemicellulose(xylan,catalog number X4252)and lignin(organosolv,catalog number 371,017),were purchased from Sigma-Aldrich and used to develop synthetic biomass.Leucaena leucocephala [mass ratio of cellulose:hemicellulose:lignin =43.06:30.77:26.17(dry,ash-free basis)]was used as a biomass model.Proximate analysis was performed by ASTM E870-82.All natural samples were dried at 110°C for 24 h.The dried L.leucocephala was then ground by a biomass grinder and filtered into particles at the size of 150-250 μm.

2.2.Thermogravimetric analysis(TGA)

Pyrolysis of synthetic biomass and natural biomass was carried out by using a thermogravimetric/differentialthermalanalyzer(Mettler Toledo TG Analyzer 851e model)under nitrogen atmosphere.Approximately 3.0 mg of sample was placed into an aluminum pan.The flow rate of nitrogen gas into the cell was 50 ml·min-1.The sample was heated from 30 to 1000°C at linear heating rates of 5,10,20 and 40 °C·min-1,respectively.

2.3.Derivation ofkinetic parametersfrom thermogravimetric analysis data

The TGA curves representthe overallweightloss rate of biomass pyrolysis.Therefore,they were used to determine the kinetic parameters of synthesized biomass pyrolysis in this study.For the non-isothermal system,the rate ofdecomposition(dα/d t)combined with the Arrhenius equation was expressed by Eq.(1).

where α is the fraction of solid materials decomposed at any time t which is defined using Eq.(2),n is the reaction order,k is the rate constantgiven by the Arrhenius equation,A is the frequency factor(s-1),Eais the activation energy (kJ·kmol-1),R is a gas constant(8.314 kJ·kmol-1·K-1)and T is the reaction temperature(K).

where W0is the original mass,W is the mass at any time t and Washis the ash content in the sample.

For linear heating rate,β=d T/d t,Eq.(1)becomes

An analytical method is the model for a single step decomposition reaction.It can be solved by integrating Eq.(3).Its integrated form of Eq.(3)for 1st order and n th order of reaction can be expressed as shown in Eqs.(4)and(5),respectively.

For 1st order of reaction

For n th order of reaction

For a case of n th order of reaction,n was calculated by the Kissinger index of shape equation.The shape index is defined as the absolute value ofthe ratio ofthe slope oftangents to the curve atin flection points of differential thermal analysis(DTA)curves[23,24].To calculate the other kinetic parameters,including Eaand A,the TGA curves were fitted with analytical models[Eqs.(4)and(5)]by means of maximizing the regression coefficient(R2).Then,the model which shows the better fit to the experimental data will be chosen to develop the correlation between biomass compositions and kinetic parameters.

2.4.Experimental design and statistical analysis

In this study,an analysis of variance(ANOVA)and response surface methodology(RSM)based on SLD was used to evaluate the effect of mass fraction of cellulose(X1),mass fraction of hemicellulose(X2)and mass fraction of lignin(X3)on each kinetic parameter.The SLD is suitable for three dependent factors when the fraction of three components is one.Three factors are main components of biomass,including cellulose,hemicellulose and lignin.Based on SLD,the samples were prepared by mixing the three components at differentratios with the summation of the proportions(X1+X2+X3)one.The sample codes and mass fraction of each experimental design point as 13 combinations are presented in Table 1.The three experimental points were pure component treatments;the six experimental points were two-component mixtures,and the four experimental points were three-component mixtures.For each mixture,the experiments were performed at four linear heating rates:5,10,20 and 40 °C·min-1with duplicates.

In an ANOVA,the p-value was used to test the statistical hypothesis.The chosen significance level was 5%which corresponded to p-value=0.05.The p-values below 0.05 indicated the statisticalsignificance of the factors.The general regression model is expressed in terms of polynomial equation[Eq.(6)].

where Y is an estimated response and a1,a2,a3,a12,a13and a23are the constant coefficients for linear and non-linear terms calculated from the experimental results using the ANOVA method[17,19].Both direction and magnitude of the constant coefficient indicate the effect andimpact of each term on the response.The ternary contour plots for estimated Ea,A and n were also generated from the proposed regression models.The plots were superimposed to observe the in fluence of biomass composition on each kinetic parameter.

Table 1 Mass fraction of biomass components based on SLD

To check the accuracy of the model,the pyrolysis of real biomass,L.leucocephala,was also performed.The kinetic parameters of pyrolysis of the real biomass would be predicted from the proposed regression models.Then,the conversion curve obtained from predicted kinetic values was compared to that obtained from the experiments.

3.Results and Discussion

3.1.Pyrolysis of pure component

Fig.1(a),(b)and(c)shows the TGA and DTA curves for pyrolysis of cellulose,hemicellulose and lignin,respectively,with heating rates of 5,10,20 and 40 °C·min-1.They illustrate the different trends of the decomposition profile for three components.As can be seen from the TGA curves,cellulose started to degrade at the highest temperature(～270 °C)when compared to the other components.This is due to the crystalline micro- fibril arrangement of cellulose which makes it resistant to thermal decomposition[25].Moreover,the cellulose is composed of only one simple repeating unit,cellubiose.Therefore,it dramatically decomposes within a short temperature range.This behavior provided only one DTA peak which exhibited the highest magnitude among the DTA curve of all components,resulting in the expected highest Ea.

Hemicellulose is composed of several monosaccharides,including glucose,xylose,mannose,galactose and arabinose.Therefore,it has more amorphous structure than cellulose[26].It started to decompose at a lower temperature than cellulose and then slowly decomposed under a broad temperature range.The DTA curve of hemicellulose pyrolysis had two main peaks because of its differentmain structures,glucose and xylan.

Lignin has a very complex structure and consists of three common phenylpropane structures:p-hydroxyphenyl,syringyl,and guaiacyl units.It is thus the most dificult one to decompose.Its decomposition happened slowly at a lower temperature and finished at the highest temperature.In a comparison ofDTAcurves ofhemicellulose and lignin,DTA obtained from lignin pyrolysis had the lowest magnitude with the widest range.Moreover,the amount of remaining solid from lignin pyrolysis was higher than those for hemicellulose and cellulose due to its largely complex structure that mostly contains aromatics.

Effect of heating rate is also demonstrated in Fig.1.It can be seen that the TGA curves and DTA peaks shifted towards higher temperature with an increase in the heating rate.This was because of the thermal lag at a higher heating rate atwhich time the difference of temperature between sample and furnace was greater than those at lower heating rates,resulting in slower decomposition[27].These results are correspondent with some of the published literature[28,29].

Fig.1.TG/DTA curves of cellulose(a),hemicellulose(b)and lignin(c)at different heating rates.

3.2.Pyrolysis of mixed components

The pyrolysis oftwo-componentmixtures and three-componentmixtures were explored.The TGA and DTA curves of pyrolysis of mixtures at 5 °C·min-1are shown in Fig.2.Fig.2(a)presents TGA and DTA curves of cellulose and hemicellulose mixtures.It can be seen that two main peaks ofDTAcurves appeared.The firstpeak mightbelong to the decomposition of the hemicellulose proportion since it appeared at the same position as can be observed in a case of pyrolysis of pure hemicellulose.The second peak belongs to the cellulose proportion for the same reason.Moreover,the quantity of the components is demonstrated by the means of peak magnitude.For example,the magnitude of the first peak decreased with the decrease of the hemicellulose fraction.

Fig.2.TG/DTA curves of synthesized biomass:cellulose:hemicellulose(a),cellulose:lignin(b),hemicellulose:lignin(c)and cellulose:hemicellulose:lignin(d)at different ratios.

Fig.2(b)presents the TGA and DTA curves of cellulose and lignin mixtures.The DTA curve had only one main peak and its magnitude agreed with the mass fraction of cellulose.It was probably because all cellulose proportion decomposed within a short temperature range as mentioned above.Therefore,the DTA curve of cellulose pyrolysis became dominant.The same phenomenon was found in a case ofhemicellulose and lignin mixtures[Fig.2(c)]which showed that the DTA curve had two main peaks at the same positions as the DTA curve obtained from pure hemicellulose pyrolysis.

In the case of three-component mixtures,Fig.2(d)shows the TGA and DTA curves of the pyrolysis of mixtures.It can be seen that the DTA curves have two or three main peaks depending on the mass fraction ofcellulose and hemicellulose.For the results ofotherheating rates,the TGA and DTA curves showed the same trend as Fig.2,butthe curves shifted forward with an increase in heat rates(the results not shown here).

3.3.The selection of analytical method

Before all kinetic parameters were calculated,a suitable analytical method had to be selected from two different cases:for n=1 and n≠1.Fig.3 illustrates the relationship between conversion and temperature of the three-component mixture from TGA and two analytical methods at 5 °C·min-1.It shows that the curves from the analytical method for n≠1 was closed to the curve obtained from TGA(R2=0.96)when compared to the curves obtained from the analytical method for n=1.Therefore,this method(n≠1)is accurate enough and was thus chosen for the further analysis.

From the TGA and DTA data,the activation energy,frequency factor and reaction order of each composition were calculated by the selected method as described in Section 2.3.Then,the effect of component proportion on each kinetic parameter and mathematical relation of biomass constituents and kinetic parameters were identified as described below.

3.4.Calculated kinetic parameters

The TGA curves were fitted according to the analytical methods for n=1[Eq.(4)]and n≠1[Eq.(5)].The kinetic parameters for both cases were then calculated as presented in Table 2.Among pure components(sample nos.6,11 and 12),pyrolysis of cellulose showed the highest Eaand A as described in Section 3.1.However,the trend ofkinetic parameters obtained from other synthesized biomass could not be clearly observed.This might be due to the interaction between biomass components.

Moreover,the trend of kinetic parameters obtained from both analytical methods was different because the changes in reaction order influenced the changes of other parameters during the calculation.From the results obtained from Section 3.3,an analytical model for n≠1 showed a better fit to the experimental data.Therefore,the kinetic parameters calculated from this model were next used in the statistical analysis.

Fig.3.The curve obtained from experimental data(cellulose:hemicellulose:lignin=0.67:0.17:0.17),the fitted curve obtained from analytical method for n=1 and the fitted curve obtained from analytical method for n≠1.

Table 2 Calculated kinetic parameters of biomass pyrolysis from analytical methods(for reaction order=1 and reaction order≠1)

3.5.Statistical analysis and modeling

To identify the effectof each biomass components on kinetic parameters and generate a statistical model for prediction of kinetic parameters,the RSM was employed.An analysis of variance(ANOVA)for Ea,A and n are shown in Tables 3(a),3(b)and 3(c),respectively.The pvalue shows the probability of kinetic parameters affected by the main composition during biomass pyrolysis.The terms which have p-value less than 0.05 have an important effect on the biomass pyrolysis.The ANOVA indicated that all independent biomass components:cellulose(X1),hemicellulose(X2)and lignin(X3),were the most significant factors affecting Eaand A.In case of n,independent components and twolevel interaction of components,including X1X2,X1X3and X2X3,were significant factors.To verify the ANOVA,the plots of normal probability of residuals,residuals versus predicted value and residuals versus effect for Ea,A and n were generated and are shown in Figs.A1,A2 and A3,respectively.The normal probability plots for all parameters show a straight line,indicating normal distribution of errors.The plots of residuals versus predicted value and residuals versus effectshow that the residuals have no pattern.The validation revealed that there were no problems in the analysis.The final mathematical models associated to the responses in terms of actual factors[Eq.(6)]are shown below.

Table 3(a)The analysis of variance(ANOVA)of activation energy

Table 3(b)The analysis of variance(ANOVA)of frequency factor

Table 3(c)The analysis of variance(ANOVA)of reaction order

From these regression models,the direction and magnitude of the constant coefficient of each term demonstrated that cellulose proportion had the most effect on Eaand A.An increase in the mass fraction of cellulose increased Eaand A,as a simple repeating unit of cellulose can be dramatically decomposed within a short time and temperature range.Therefore,an increase in this fraction leads to the simple decomposition of biomass at appropriate temperature,resulting in higher Eaand A.For n,interaction between the mass fractions of cellulose(X1)and lignin(X3)had the most effect.

The ternary plots were generated by the regression models.Fig.4 shows the ternary contour plots indicating the effects of the interaction between main components and Ea[Fig.4(a)],A[Fig.4(b)]and n[Fig.4(c)].The effects of the components and interactions on the kinetic parameters can be observed by these plots.These results provide useful information to predict the kinetic parameters of biomass pyrolysis at different compositions.

3.6.Application to real biomass

The biomass sample,L.leucocephala,with known composition was used to test the accuracy of models.Its composition was substituted into Eqs.(7)-(9)in order to calculate the kinetic parameters.Then,the calculated parameters were substituted into Eq.(5).The relationship between temperature and conversion was plotted and compared to the TGA curve as shown in Fig.5.The R2value was close to unity(0.98)which demonstrates a good agreement of the generated model and experimental data.In addition,these models were applied to calculate the kinetic values of other biomass pyrolysis.The results and R2values are presented in Table 4.As high R2values were obtained,the analytical method for n≠1 and statistical analysis have the potential to identify the correlation between biomass components and kinetic values for biomass pyrolysis.

However,while the error was acceptable with respect to R2value,the difference between experimental and predicted data was detected at moderate temperature(Fig.5).This is due to the assumption of the kinetic model.In addition,xylan and organosolv lignin were used for the hemicellulose and lignin model,respectively,since these forms of biomass components are largely found in nature.Still,while their behaviors are different from real hemicellulose and lignin,the obtained model can be used as a simple approximation for predicting the overall trend of chemical reaction kinetics.

Fig.4.Ternary contour plots of predicted activation energy(a),logarithm of frequency factor(b)and reaction order(c).

Fig.5.The conversion curves obtained from experimental data and predicted data of pyrolysis of Leucaena leucocephala at 20 °C·min-1;R2=0.98.

Table 4 Predicted kinetic values of different biomass

To generate a more precise conversion curve,othersolid-state kinetic reactions,including nucleation,geometricalcontraction and diffusion,should be applied into the kinetic models.In real biomass,the interaction between all components,particles-volatiles and the effect of alkali and alkaline earth metal species can be very strong[15,38].However,the effect of ash on kinetic parameters is not important comparing to the used reactive species.Thus,besides applying other solid-state kinetic reactions,the terms for those involved the interaction should be included for more accuracy.

4.Conclusions

The relationship between biomass constituent and kinetic parameters for biomass pyrolysis was determined.Response surface methodology and simplex-lattice mixture design were used for this purpose.The results show that a model obtained from the analytical method(for n≠1)can be appropriately used to calculate the activation energy,frequency factor reaction order from thermogravimetric analysis data that provided a high regression coefficient.From statistical analysis,cellulose has the most significant effect on the activation energy and the frequency factor while interaction between cellulose and lignin has the most effect on the reaction order.The generated regression models for each kinetic parameter were used to produce ternary contour plots and determine the kinetic parameters.Then,L.leucocephala was used as an example for the accuracy test of these models.The results show that the calculated kinetic values provide a very acceptable thermogravimetric analysis curve compared to the curve obtained from the experiment.It is expected that this work provides useful information for estimating the kinetic parameters for other biomass pyrolysis and designing suitable process operating conditions.

Acknowledgements

The authors would like to express their thanks to the Graduate School,Chulalongkorn University for partial financial support.The authors also thank the Publication Counseling Unit,Chulalongkorn University for English language editing.

Appendix

Fig.A1.Normal probability plot of residuals(a),residual vs predicted values(b)and residual vs factor(c)for the prediction of activation energy.

Fig.A2.Normal probability plot of residuals(a),residual vs predicted values(b)and residual vs factor(c)for the prediction of logarithm of frequency factor.

Fig.A3.Normal probability plot of residuals(a),residual vs.predicted values(b)and residual vs factor(c)for the prediction of reaction order.

References

[1]J.Jones,Converting solid wastes and residues to fuel,Chem.Eng.85(1978)87-94.

[2]E.Mura,O.Debono,A.Villot,F.Paviet,Pyrolysis of biomass in a semi-industrial scale reactor:Study of the fuel-nitrogen oxidation during combustion of volatiles,Biomass Bioenergy 59(2013)187-194.

[3]J.M.Encinar,J.F.Gonzalez,J.Gonzalez,Fixed-bed pyrolysis of Cynara cardunculus L.Product yields and compositions,Fuel Process.Technol.68(3)(2000)209-222.

[4]Y.F.Huang,W.H.Kuan,P.T.Chiueh,S.L.Lo,A sequential method to analyze the kinetics of biomass pyrolysis,Bioresour.Technol.102(19)(2011)9241-9246.

[5]K.L.Lam,A.O.Oyedun,C.W.Hui,Experimental and modelling studies of biomass pyrolysis,Chin.J.Chem.Eng.20(3)(2012)543-550.

[6]Y.F.Huang,P.T.Chiueh,W.H.Kuan,S.L.Lo,Pyrolysis kinetics of biomass from product information,Appl.Energy 110(2013)1-8.

[7]S.Hu,A.Jess,M.Xu,Kinetic study of Chinese biomass slow pyrolysis:Comparison of different kinetic models,Fuel 86(17-18)(2007)2778-2788.

[8]A.Meng,H.Zhou,L.Qin,Y.Zhang,Q.Li,Quantitative and kinetic TG-FTIR investigation on three kinds of biomass pyrolysis,J.Anal.Appl.Pyrolysis 104(2013)28-37.

[9] ?.?epelio?ullar,A.E.Pütün,Thermal and kinetic behaviors of biomass and plastic wastes in co-pyrolysis,Energy Convers.Manag.75(2013)263-270.

[10]J.Chattopadhyay,C.Kim,R.Kim,D.Pak,Thermogravimetric characteristics and kinetic study of biomass co-pyrolysis with plastics,Korean J.Chem.Eng.25(5)(2008)1047-1053.

[11]A.Garcia-Maraver,D.Salvachua,M.J.Martinez,L.F.Diaz,M.Zamorano,Analysis of the relation between the cellulose,hemicellulose and lignin content and the thermal behavior of residual biomass from olive trees,Waste Manag.(11)(2013)2245-2249.

[12]G.Wang,A.Li,Thermal decomposition and kinetics of mixtures of polylactic acid and biomass during copyrolysis,Chin.J.Chem.Eng.16(6)(2008)929-933.

[13]L.Burhenne,J.Messmer,T.Aicher,M.-P.Laborie,The effect of the biomass components lignin,cellulose and hemicellulose on TGA and fixed bed pyrolysis,J.Anal.Appl.Pyrolysis 101(2013)177-184.

[14]B.Peters,Prediction of pyrolysis of pistachio shells based on its components hemicellulose,cellulose and lignin,Fuel Process.Technol.92(10)(2011)1993-1998.

[15]C.Couhert,J.-M.Commandre,S.Salvador,Is it possible to predict gas yields of any biomass after rapid pyrolysis at high temperature from its composition in cellulose,hemicellulose and lignin?Fuel 88(3)(2009)408-417.

[16]A.Gani,I.Naruse,Effect of cellulose and lignin content on pyrolysis and combustion characteristics for several types of biomass,Renew.Energy 32(4)(2007)649-661.

[17]S.Karaman,M.T.Yilmaz,A.Kayacier,Simplex lattice mixture design approach on the rheological behavior of glucomannan based salep-honey drink mixtures:An optimization study based on the sensory properties,Food Hydrocoll.25(5)(2011)1319-1326.

[18]Q.Liu,Z.Zhong,S.Wang,Z.Luo,Interactions of biomass components during pyrolysis:A TG-FTIR study,J.Anal.Appl.Pyrolysis 90(2)(2011)213-218.

[19]P.V.Rao,S.S.Baral,Experimental design of mixture for the anaerobic co-digestion of sewage sludge,Chem.Eng.J.172(2-3)(2011)977-986.

[20]S.Azevedo,L.M.Cunha,P.V.Mahajan,S.C.Fonseca,Application of simplex lattice design for development of moisture absorber for oyster mushrooms,Procedia Food Sci.1(2011)184-189.

[21]S.M.Teimouri Sendesi,J.Tow fighi,K.Keyvanloo,Effect of iron,phosphorous,and Si/Al on HZSM-5 catalytic performance and stability by response surface methodology,J.Anal.Appl.Pyrolysis 104(2013)695-702.

[22]S.Chayaporn,P.Sungsuk,S.Sunphorka,P.Kuchonthara,P.Piumsomboon,B.Chalermsinsuwan,Evaluation of biomass component effect on kinetic values for biomass pyrolysis using simplex lattice design,Korean J.Chem.Eng.32(6)(2015)1081-1093.

[23]C.-P.Lin,Y.-M.Chang,J.P.Gupta,C.-M.Shu,Comparisons of TGA and DSC approaches to evaluate nitrocellulose thermal degradation energy and stabilizer ef ficiencies,Process Saf.Environ.88(6)(2010)413-419.

[24]H.E.Kissinger,Reaction kinetics in differential thermal analysis,Anal.Chem.29(11)(1957)1702-1706.

[25]J.Lehmann,S.Joseph,Biochar for environmentalmanagement:Science and technology,Earthscan,London,UK,2009.

[26]S.Irmak,M.Kurtulus,A.Hasanoglu,O.Erbatur,Gasification efficiencies of cellulose,hemicellulose and lignin fractions of biomass in aqueous media by using Pt on activated carbon catalyst,Biomass Bioenergy 49(2013)102-108.

[27]R.Comesana,M.A.Gomez,M.A.Alvarez,P.Eguia,Thermal lag analysis on a simulated TGA-DSC device,Thermochim.Acta 547(2012)13-21.

[28]H.Yang,R.Yan,H.Chen,C.Zheng,D.H.Lee,D.T.Liang,In-depth investigation of biomass pyrolysis based on three major components:Hemicellulose,cellulose and lignin,Energy Fuel 20(1)(2005)388-393.

[29]G.Wang,W.Li,B.Li,H.Chen,TG study on pyrolysis of biomass and its three components under syngas,Fuel 87(4-5)(2008)552-558.

[30]M.Brebu,S.Ucar,C.Vasile,J.Yanik,Co-pyrolysis of pine cone with synthetic polymers,Fuel 89(2010)1911-1918.

[31]N.Worasuwannarak,T.Sonobe,W.Tanthapanichakoon,Pyrolysis behaviors of rice straw,rice husk,and corncob by TG-MS technique,J.Anal.Appl.Pyrolysis 78(2007)265-271.

[32]T.E.Odetoye,K.R.Onifade,M.S.AbuBakar,J.O.Titiloye,Thermochemical characterisation of Parinari polyandra Benth fruit shell,Ind.Crop.Prod.44(2013)62-66.

[33]M.Carrier,A.Loppinet-Serani,D.Denux,J.-M.Lasnier,F.Ham-Pichavant,F.Cansell,et al.,Thermogravimetric analysis as a new method to determine the lignocellulosic composition of biomass,Biomass Bioenergy 35(2011)298-307.

[34]D.Li,L.Chen,X.Zhang,N.Ye,F.Xing,Pyrolytic characteristics and kinetic studies of three kinds of red algae,Biomass Bioenergy 35(2011)1765-1772.

[35]W.-H.Chen,H.-C.Hsu,K.-M.Lu,W.-J.Lee,T.-C.Lin,Thermal pretreatment of wood(Lauan)block by torrefaction and its in fluence on the properties of the biomass,Energy 36(2011)3012-3021.

[36]J.Zeng,D.Singh,S.Chen,Thermal decomposition kinetics of wheat straw treated by Phanerochaete chrysosporium,Int.Biodeterior.Biodegrad.65(2011)410-414.

[37]J.O.Titiloye,M.S.Abu Bakar,T.E.Odetoye,Thermochemical characterisation of agricultural wastes from West Africa,Ind.Crop.Prod.47(2013)199-203.

[38]N.Sonoyama,J.-I.Hayashi,Characterisation of coal and biomass based on kinetic parameter distributions for pyrolysis,Fuel 114(2013)206-215.