A comparative analysis of kinetic parameters from TGDTA of Jatropha curcas oil, biodiesel, petroleum diesel and B50 using different methods

This article appeared in a journal published by Elsevier. The attachedcopy is furnished to the author for internal non-commercial researchand education use, including for instruction at the authors institution

and sharing with colleagues.

Other uses, including reproduction and distribution, or selling orlicensing copies, or posting to personal, institutional or third party

websites are prohibited.

In most cases authors are permitted to post their version of thearticle (e.g. in Word or Tex form) to their personal website orinstitutional repository. Authors requiring further information

regarding Elsevier’s archiving and manuscript policies areencouraged to visit:

http://www.elsevier.com/authorsrights

Author's personal copy

A comparative analysis of kinetic parameters from TGDTA of Jatropha curcas oil,biodiesel, petroleum diesel and B50 using different methods

A.P. Singh Chouhan a, Neetu Singh b, A.K. Sarma a,⇑a Sardar Swaran Singh National Institute of Renewable Energy, 12KM Stone, Jalandhar Kapurthala Road, Punjab 144 601, India b Dr. BR Ambedkar NIT Jalandhar, Jalandhar, India

h i g h l i g h t s

" The TGA thermograms presented for different fuels viz. JCO, JCME (KOH), JCME (NaOH), petroleum diesel and B50 blends." Calculations of the different paramete rs applicable for kinetic study purposes using three different methods." The refined JCO has highest value of activa tion energy 49.97 kJ/mol while petroleum diesel has the lowest (9.11 kJ/mol).

a r t i c l e i n f o

Article history:Received 29 October 2012 Received in revised form 20 November 2012 Accepted 13 December 2012 Available online 10 January 2013

Keywords:Reined JCO TGDTA analysis FAMEKinetic methods Kinetic parameters

a b s t r a c t

Fatty Acid Methyl Ester (FAME) obtained from the transesterification reaction of refined Jatropha curcas oil (JCO) and methanol using homogeneous catalysts NaOH and KOH were investigated. Gas chromato- graphic analysis were performed to determine fatty acid profile of the JCO and percentage conversion of JCO to FAME. Thermogr avimetric experiments were conducted in nitrogen and air medium at constant heating rates 10 �C/min and constant flow of gases 20 ± 0.5 ml/min and calculated the weight losses with increasing temperature. The TGA thermograms were analyzed for fuels under test viz. JCO, FAME KOH,FAMENaOH, petroleum diesel and B50 blends. The calculations of the different parameters applicable for kinetic study purposes were carried out using three different methods available in literature viz. Redfern method, Friedman method and Differential method (Direct Arrhenius method), respectively. A compara- tive thermogravi metric and differential thermal analysis (TGDTA) of JCO, diesel, FAME and B50 can pre- dict the thermal behavior and combustion chara cteristics. It has been observed that the refined JCO has the highest value (49.97 kJ/mol, Differential method) and most stable thermally and during combustion while petroleum diesel has the lowest activation energy (9.11 kJ/mol, Friedman method) and easily com- bustible as comp ared to biodiesel or blends. On the basis of the kinetic analysis it can also be concluded that all the three methods are equ ally effective to predict the combustion beh avior of the fuel although there are variations in the quantitative value of activa tion energy.

� 2012 Elsevier Ltd. All rights reserved.

1. Introduction

The reserve to production ratio of petroleum crude oil is dimin- ishing day to day resulting crisis and price hikes of petroleum products. Moreover, detrimental environm ental concerns of using fossil fuels have been stimulati ng the growth of renewable liquid fuel. Jatropha curcas oil (JCO) has been emerging as the plausible substitute of petroleum derived diesel in major parts of the Asian

countries . Investigations are ongoing in several Southeast Asian countries like India, Thailand, and Malaysia, for mass scale produc- tion and application of JCO. It can be directly used in diesel engine blending with petroleum diesel or by preparing Fatty Acid Methyl Ester (FAME) commonl y used as biodiesel . Biodiesel production from vegetable oils and non-edible oils is costly as compare d toconventi onal diesel fuel but JCO provides promising solution [1,2]. The seed kernel of J. curcas L. contain about 30–35 wt.% ofexorable oil. The common by-products produced while processing for biodiesel are glycerol and seed cake. During the expeller pro- cessing about 65–70 wt.% seed cakes are produced as reported.Seed cake has been used as a feedstock for the biogas production purposes [2]. The use of JCO is beneficial due to the low cost ascompare d to other edible and non-edible oils. The oil or fat used

0016-2361/$ - see front matter � 2012 Elsevier Ltd. All rights reserved.http://dx.doi.org/10.1016/j.fuel.2012.12.059

Abbreviations: TGDTA, thermogravimetric and differential thermal analysis; JCO,Jatropha curcas oil; FAME, Fatty Acid Methyl Ester; CRM, Coats and Redfern method;FR, Friedman method; DM, Differential method (Direct Arrhenius method); FFA,free fatty acid; SD, Standard deviation; SE, Standard error.⇑ Corresponding author. Tel.: +91 1822255543.

E-mail address: [email protected] (A.K. Sarma).

Fuel 109 (2013) 217–224

Contents lists available at SciVerse ScienceDi rect

Fuel

journal homepage: www.elsevier .com/locate / fuel

Author's personal copy

in alkaline transeste rification reaction should not contain more than 0.1 wt.% free fatty acid (FFA), which is equivalent to 0.2 mgKOH/g triglyceride. If the FFA level exceeds this threshold, saponi- fication hinders separation of the ester from glycerol and reduces the formation rate thereby reducing the yield of FAME [3].

A large volume of research articles and patents are available inthe area of biodiesel production, fuel property analysis, effect ofadditives during storage and combustion using JCO as feed stock [1,4]. Thus, the perceptive of thermochemical conversion processes of the biofuel in comparison to petroleum derived fuel are also found necessar y [5]. Thermal analysis techniques used for edible oil and fat characteri zed by measuring several properties such asthermo oxidative behavior and stability, specific heat, thermal decompositi on, activation energy, temperat ure and enthalpy ofcrystallization, action of antioxidant s in oil thermal stability, high pressure oxidation, induction time measureme nt, etc., had been re- ported [6,7]. The combusti on mechanism of fuel, fuel property,ignition behavior and ignition delay, etc., plays vital role during the developmen t of combustion engines.

Very recently, thermal oxidation stability measure ment of bio- diesel using different antioxidants was reported by Jain and Shar- ma [8]. The kinetic parameters were calculated from Friedman kinetic equation using thermogravi metric analysis (TGA). Crnkovic et al. [9] calculated and presented the activation energy and com- bustion behavior of beef tallow and crude glycerine using direct Arrhenius equation and model free kinetic methods.

A detail investigation and comparison of different methods used for calculations of different kinetic parameters that can be applied for biofuel combustion and thermal degradation are not available in literature and found necessary. The calculations of the different parameters applicable for the kinetic study purposes i.e. R2, activa- tion energy, frequency factor and rate of reaction are found neces- sary for three different methods available in literature viz. Coats and Redfern method(CRM), Friedman method(FM) and Differential method (DM) (Direct Arrhenius method), respectively .

In this article, the authors investigated and focused the compar- ative study of JCO, FAME NaOH, FAME KOH, petroleum diesel, B50 blends using thermogravi metric profile and kinetic parameters.This study has been conducted to evaluate and co-relate the ther- mal decompo sition and the weight losses of the fuels at different temperature s. TGA thermograms have also been used for the calcu- lation of activation energy, kinetic rate constant and frequenc y fac- tor of the fuels under study that leads to establish the stability factor of tested fuels.

2. Materials and methods

2.1. Oil collection and refining

J. Curcas oil (JCO) was purchased from an oil expeller Industry,Udaipur, Rajasthan, India. The JCO had a density 0.918 g/cm 3 at20 �C with an initial acid value 8.9 mgKOH/g. The JCO was later treated with 1% KOH solution prepared using Millipore distilled water and washed several times. The washed oil was dried in abeaker using Na2SO4 and then in an oven at 50 �C for overnight.

The same was later tested for free fatty acids and water content and was named as refined JCO. Methanol was purchased from Loba chemical s with a purity of 99.5% and used as such. The petroleum diesel required for the experiment was purchased from the nearest petrol pump of Hindusta n Oil Corporation located in Wadala Kalan,Punjab, India. This was used for TGA experiments and for preparing blends without further purification.

2.2. Biodiesel preparati on

NaOH and KOH were weighed separately 1 wt.% of the total amount of oil using standard weighing balance and dissolved in re- quired quantity of methanol at 50 �C. The biodiesel preparation was carried out separately , in 10 ml capacity vials using 1:9 M ratio of oil to alcohol using methanol(NaOH) and methanol(KOH) and refined JCO. These vials were sealed and shaked vigorously for 1 h and kept in an oven at 65 �C for 24 h. Then the vials were opened and the reaction mixtures were washed 5–6 times to re- move the catalyst and methano l. These were later centrifuged toseparate the glycerol (Model No. 5430R, manufactured by Eppen- dorf Company, UK) for about 15 min at 10,000 rpm at 15 �C con- stant temperature. The left over glycerol was separated from the esters i.e. FAME NaOH and FAME KOH by decantation. This process was repeated to produce substantial amount of FAME NaOH andFAMEKOH for further characteri zation. A portion of these biodiesel was mixed with equal weight of petroleum diesel to prepare their B50 blends.

2.3. Characteri zation of biodiesel

The FAMEs were characterized for density by digital density meter (model No. DMA 5000 Series, manufactur ed by Anton Paar,Germany ), viscosity by Ostwald viscometer (model No. 86-18D Lawler viscometer bath, manufactur ed is USA), flash point by auto- matic flashpoint tester (model No. ACO-7, made by Tanaka Com- pany, Japan), calorific value using automatic bomb calorimeter (Toshniwal, India) and fatty acid methyl ester conversion identifiedand quantified using the Gas Chromatograp h (AgilentTechnolo gies).

2.4. Gas chromatograp hy for biodiesel analysis

The refined JCO was further characterized for fatty acid profileusing GC–FID detector as per AOCS Official method 1998, Ce 1-62 and Ce 2-26.

The FAMEs were analyzed using GC in a column of dimethyl si- loxane with dimension 30 m � 250 lm and for the determination of glycerides viz. glycerol, monoglyceride, diglyceride and triglyc- eride another column was used. The oven temperature was ini- tially maintain ed at 210 �C for overall process in FAME analysis and for glycerides it was kept at 50 �C. The GC worked with two FID detectors using helium, nitrogen, air and hydrogen gases. The injector, transfer and source temperat ures were 220 �C, 200 �Cand 150 �C, respectively . The carrier gases used was helium while make up gas was nitrogen with a total scan time 35 min. The FAME

Nomenc lature

Ea activation energy R2 regression coefficientR gas constant = 8.31 J mol �1 K�1

T temperatur e (�C)b heating rate

A frequency factor x fractional weight loss n number of observat ion/method sX1,X2, . . . ,Xn Input value 1, Input value 2 . . . Input value n.Xm mean value of the total observatio n.

218 A.P. Singh Chouhan et al. / Fuel 109 (2013) 217–224

Author's personal copy

conversion, expressed as wt.% has been calculated by the following formula as per procedure EN 14103.

%FAME Conversion ¼ ½XðA� AEIÞ=AEI� ðWEI=WÞ � 100�

XA ¼ Total peak area from C6 : 0 to C24 : 1

XAEI ¼ Peak area corresponding to nonadecanoic acid methyl

ester

WEI ¼WeightðmgÞof nonadecanoic acid methyl ester being used as internal standard

W ¼Weight ðmgÞ of the sample

2.5. Thermograv imetric (TGA) analyzer

The TGDTA instrument manufactur ed by Perkin Elmer, model STA 6000 was used to study the thermochemi cal behavior of the samples within 30–1000 �C, heating rate 10 �C/min and constant flow of nitrogen and air at a rate (20 ± 0.5) ml/min.

3. Theory/calculati ons

3.1. Kinetic study

The kinetic study was carried out using three established meth- ods viz. Coats and Redfern method, Friedman method and Direct Arrhenius method (Differential method). These methods were used for the analysis of various kinetics parameters like activation en- ergy (E), kinetic rate constant (k) and Frequency factor (A) based on the TGA thermograms data.

3.1.1. Coats and Redfern method Coats and Redfern method is an integral method used for non-

isothermal kinetic analysis. This method eliminates the rate con- stant and gives activation energy and frequenc y factor directly [10–18]. It can be written in the equation form as:

ln1� ð1� xÞð1�nÞ

T2ð1� nÞ

" #¼ ln

ARbEa

� �1� 2RT

Ea

� �� Ea

RTðfor n–1Þ ð1Þ

for n = 1, the above equation modifies as

lnlnð1� xÞ

T2

� �¼ ln

ARbEa

� �1� 2RT

Ea

� �� Ea

RTðfor n ¼ 1Þ ð2Þ

where x is the fractional weight loss. In Eq. (2) except fractiona lweight loss (x) and TGA temperatu re T, other paramete rs are con- stant. Thus plotting a graph betwee n ln lnð1�xÞ

T2

h iand 1

T the value of�Ea/R can be obtained, from which activati on energy (Ea) can becalculate d.

3.1.2. Friedman method The Friedman equation can be expressed as

lndxdt

� �¼ ln f ðxÞ þ ln A� Ea

RTð3Þ

Eq. (3) can be used to calculate the kinetic paramete rs without the call for a mathemat ical model. By plotting the ln(dx/dt) and 1/Tcurves, one can get a straight line which gives the activati on energy (Ea), frequen cy factor (A) and kinetic rate constant (K). This method was used for non-isoth ermal kinetic analysis of the devolat ilization of corn cobs and sugar cane bagasse in inert atmospher e [19,20].

3.1.3. Direct Arrhenius method Thermogr avemetric data can be used to characteri ze the mate-

rials (solid or liquid samples) to determine the thermodyna mics and kinetics of the reactions as depicted in Arrhenius equation [21–23].

The Differential method or Direct Arrhenius method can be ex- pressed mathematicall y as

ln1

1� x

� �¼ ln

Ab

� �� Ea

RT

� �ð4Þ

The plot of ln 11�x

dxdT

� �versus 1

T should give a straight line with slope (�Ea/RT) from which the activation energy, Ea can be calculate d.

3.2. Error calculation

Standard error (SE) for different methods applied can be calcu- lated using the following equation:

Standard error ðSEÞ ¼ Standard deviation ðSDÞ=ffiffiffinp

ð5Þ

where

Standard deviation ðSDÞ¼

ffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi1

ðn�1Þ�ðX1�XmÞ�ðX1�XmÞþðX2�XmÞ�ðX2�XmÞþ���ðXn�XmÞ�ðXn�XmÞ

s

where Xm is the mean value of total inputs, n is the the number ofmethods/ observati on.

4. Results and discussion

4.1. Transesteri fication reaction and fatty ester profile of the refinedJCO

The transesterifcat ion reaction carried out for refined JCO invials with an initial acid value 0.00 mg KOH/g (Table 1) and fatty ester profile was determined by GC analysis. Refined JCO consists of 16.95 wt.% of methyl palmitate, 33.31 wt.% of methyl linoleate,38.4 wt.% of methyl oleate, 8.57 wt.% of methyl stearate and 0.2 wt.% of methyl archidate . The oleic acid is the major fatty acid followed by linoleic acid, palmitic acid and stearic acid in JCO com- prising of about 25.8 wt.% unsaturated and 70.4 wt.% saturated fatty acids. The individual peaks of the gas chromatogram of the transeste rified product FAME NaOH have been presented in Fig. 1.

The percentage conversio n of oil to FAME during batch reaction,after separating the glycerol was found to be 96.5% for FAME with NaOH while the same was 96.9% with KOH.

4.2. Refined JCO, FAMEs and petroleum diesel characterization

The FAMEs and blends prepared were characteri zed for the minimum basic fuel properties as per ASTM and DIN EN14214 methods and presente d in Table 1. Because the wt.% conversion of FAMEs were as per the ASTM standards, the primary fuel prop- erties applicabl e for biodiesel were also found within the pre- scribed limit. The fuel properties were determined using modern equipme nts available in our laboratory and IOCL Jalandhar asapplicabl e to petroleum based fuels. The quantitative value re- corded is the mean value of at least the three consecut ive measure- ments. Obviously the density, viscosity , flash point of refined JCO are much higher as compared to its correspondi ng esters, B50 blends and petroleum diesel.

4.3. Thermogr ams of the tested samples

The TGA data were extracted from the instrument and plotted the same for the refined JCO, petroleum diesel, FAME NaOH, FAME KOH

and their B50 blends as shown in Figs. 2–7. Each of the plots has

A.P. Singh Chouhan et al. / Fuel 109 (2013) 217–224 219

Author's personal copy

Table 1Fuel propert ies of refined JCO, FAMEs and petroleum diesel.

S.No. Property Refined Jatropha oil FAME NaOH/FAMEKoH Petroleum diesel Biodiesel ASTD 6751-02 Standards DIN EN 14214

1 Density (15 �C, g/cm 3) 0.919 0.881/0.880 0.838 – 860–9002 Viscosity cSt 26.5 5.7/5.5 4.4 1.9–6.0 3.5–5J Flash point (�C) 225 110/110 83 >130 >120 4 Pour point (�C) 4 2/2 �20 – –5 Water content (%) 1.4 0.025/0.025 0.02 <0.03 <0.02 6 Ash content (%) 0.8 0.012/0.010 0.01 <0.02 <0.02 7 Carbon residue (%) 1 0.30/0.3 0.17 – <0.30 8 Acid value (mgKOH/g) 0.00 0.00/0.00 – <0.80 <0.50 9 Calorific value (kJ/g) 38.65 39.23/39.4 45.01 – –10 Copper strip corrosion – la/la lb – –11 Monoglyceride % – 0.180/0.176 – –12 Diglyceride % – 0.000/0.001% – – EN 14106 13 Triglyceride % – 0.020/0.021 % – –14 Glycerol % – 0.107/0.100% – –

Fig. 1. GC profile of FAME (NaOH).

Fig. 2. Thermograms of Refined JCO in air and nitrogen medium. Fig. 3. Thermograms of petroleum diesel in air and nitrogen medium.

220 A.P. Singh Chouhan et al. / Fuel 109 (2013) 217–224

Author's personal copy

two curves overlapping each other in nitrogen and air medium,respectively . The fractional weight losses were plotted against con- stant temperature interval 50 �C.

Fig. 2 shows the thermogram s of refined JCO with constant heating rate and constant flow of nitrogen and air. There is no sig- nificant difference observed between the two thermograms although the mediums are different. The initial evaporati ve losses observed at 100 �C which could be due to the presence of moisture with the refined oil. The thermal and evaporative losses within 100–300 �C is only about 2 wt.%. This result is attributed to the presence of mono and diglyceride in the oil structure. The signifi-cant weight losses are observed within 300–450 �C which is about 94 wt.% to the original weight of the sample whereas about 4 wt.%degradation is observed within 450–500 �C. During this thermal treatment process primarily the triglyceride molecules are vapor- ized while a fraction may undergo cracking and polymerization and subsequent vaporiza tion in the absence of air (N2 medium).In the presence of air, however, oxidative polymerizat ion and evaporation are feasible within the temperature range 300–500 �C, in addition to cracking and combustion. The remaining part of the thermogram is parallel to X-axis at zero showing no residual inorganic. Similar types of thermograms were reported for fodder radish crude oil in a recent article [24]. The thermogram obtained for refined JCO is very similar to the recently reported thermo- grams at a constant heating rate 10 �C by Mazumdar et al. [25]for JCO of North East India viz. IITJ 8, IITJ 9, IITJ 18 and IITJ 20 under air and nitrogen environment.

Fig. 3 shows the thermograms of petroleum diesel derived from the TGA data with the same heating rate in air and nitrogen envi- ronment. There is very little variation between the two curves un- der nitrogen and air medium. It is expected that the weight loss with time in nitrogen medium is due to the phase change and isendothermi c in nature while in air medium the change is attrib- uted to both phase change and subsequent combustion of the die- sel fuel after attaining the ignition temperature in an exothermic process. About 98 wt.% petroleum diesel losses are observed within 100–370 �C which is the near normal (140–370 �C) boiling range ofdiesel [26]. The weight losses are mostly evaporative up to 210 �C(ignition temperature ) in either environm ent while after 210 �Cthere is the possibility of auto ignition in diesel fuel in air medium [25]. However , little significant differenc es were observed as be- cause the homogen eous mixture of diesel fuel evaporated smoothly with increasing temperature .

Fig. 4. Thermograms of biodiesel with NaOH catalyst in air and nitrogen medium.

Fig. 5. Thermograms of biodiesel with KOH catalyst in air and nitrogen medium.

Fig. 6. Thermograms of B50 (NaOH) with petroleum diesel air and nitrogen medium.

Fig. 7. Thermograms of B50 (KOH) with petroleum diesel in air and nitrogen medium.

A.P. Singh Chouhan et al. / Fuel 109 (2013) 217–224 221

Author's personal copy

Figs. 4 and 5 show the thermogr ams of FAME NaOH andFAMEKOH respectively , from which it has been observed that both the fuels show major weight loss within 280–400 �C which is about 96 wt.% to the original sample. This boiling range ismuch inferior as compared to the complex triglyceride molecules of refined JCO. The major components of both the biodiesel samples are methyl oleate and methyl linolieate and hence weight loss observed within 350–390 �C is found most promi- nent as expected from the compositi on. However, the earlier re- ported results for IITJ 8, IITJ 9, IIJ 19 and IITJ20 showed major weight losses (90%) at 280 �C [25]. There is no appreciable distinction of the TGA profiles of these two fuels as the source of the fuel is same and percent conversio ns to esters are also equivalent. However, in oxidative mood of operation the thero- mograms showed little different behavior as expected due torapid combustion along with evaporation [24].

Figs. 6 and 7 show the thermograms of B50 i.e. the blends ofbiodiesel with petroleum diesel. Because of the fact that the blends are homogeneous mixtures of biodiesel with petroleum diesel,these thermogram s are much wider as compared to biodiesel,while these are relatively steeper as compared to petroleum diesel.In these thermograms about 80 wt.% losses are observed within the temperature range 150–370 �C making it a very perfect fuel interms of boiling point. In air environment the process is highly exo- thermic whereas in nitrogen environment it is endothermic alto- gether. In nitrogen environment the process involves mainly phase change although at high temperature (>300 �C) there is bond cleavage and cracking is a side reaction which is exothermic innature.

4.4. Activation energy

The activation energy (kJ/mol) i.e. the minimum energy re- quired to cross the barriers for initiating a chemical reaction has been calculated using Eqs. (2)–(4) while SEs were calculated using Eq. (5) for the fuels under tests and presented in Tables 2 and 3,respectively .

It has been clearly observed that the activation energies ob- tained for the refined JCO under nitrogen flow are 42.29 kJ/mol,37.39 kJ/mol 39.19 kJ/mol from CRM, FM and DM methods respec- tively with calculated SE equal to 1.53 are the highest while the same are lowest for petroleum diesel viz. 15.74 kJ/mol, 11.02 kJ/ mol and 12.23 kJ/mol respectively with calculated SE equal to1.41. The activation energies calculated under air flow are 49.76 kJ/mol, 43.18 kJ/mol and 49.97 kJ/mol obtained from CRM,FM and DM methods , respectivel y with SE equal to 2.64. For petro- leum diesel the activation energy values under air flow are 16.47 kJ/mol, 9.11 kJ/mol and 10.88 kJ/mol, respectively , with SEvalue 2.22. It was earlier reported that the activation energy calcu- lated for JCO oil is 39 kJ/mol under air flow as calculated from dif- ferential method [8]. The difference in quantitat ive values may beattributed to the composition difference between the JCO oil. In re- fined JCO the lighter molecular weight free fatty acids has been re- moved by washing with dilute alkali solutions which are responsib le for widening the boiling range. Aligrot et al. [27] deter-mined activation energy for six different diesel fuels and values ranged from 65 kJ/mol and 85 kJ/mol. However, Crnkovic et al.[9] used model free kinetics for determining the activation ener- gies of beef tallow and crude glycerin and ascertained that the acti- vation energy parameter did not remain constant during the reaction. In addition, the activation energy range determined for both crude glycerine (90–42 kJ/mol) and beef tallow (50 kJ/mol and 113 kJ/mol) were an indicative that the activation energy should be considered as several values instead of an average value,and this information is expected to play an important role mainly when new fuels have been used in combustion processes . In aver- age, values of activation energies for both crude glycerin and beef tallow were close to the range for diesel. Furthermore, in this investiga tion it is found invariably important to determine the activation energies using three available methods so as to ascertain the variation s.

Moreove r, the activation energy of biodiesels i.e. FAME KOH

(Nitrogen: 20.71 kJ/mol, 15.59 kJ/mol and 20.7 kJ/mol, SE 1.70;Air: 30.39 kJ/mol, 14.53 kJ/mol and 24.71 kJ/mol, SE 4.46) and

Table 2Kinetic parameters calculated from TGA thermograms for fuels under test in nitrogen medium .

Fuels under test Kinetic methods Kinetic parameter R2

Activation energy, E (kJ/mol) Frequency factor (A) s�1 Rate constant (K)

Refined JCO CRM 42.29 177.84 0.51 0.91 FM 37.39 136.25 0.46 0.93 DM 39.19 175.67 0.76 0.95 SE 1.53 13.53 0.09

Petroleum diesel CRM 15.74 5.97 0.27 0.98 FM 11.02 1.39 0.11 0.99 DM 12.23 1.82 0.12 0.99 SE 1.41 1.46 0.06

Biodiesel (NaOH) CRM 23.56 32.21 0.30 0.91 FM 17.15 24.75 0.84 0.98 DM 17.81 6.55 0.15 0.98 SE 2.03 7.63 0.20

Biodiesel (KOH) CRM 20.71 12.11 0.15 0.91 FM 15.59 18.85 0.82 0.92 DM 20.7 1.27 0.13 0.95 SE 1.7 5.12 0.22

B50 (NaOH) with petroleum diesel CRM 15.72 4.09 0.19 0.94 FM 10.86 0.84 0.07 0.97 DM 17.02 0.59 0.17 0.91 SE 1.87 1.12 0.03

B50 (KOH)with petroleum diesel CRM 15.05 2.41 0.13 0.93 FM 10.03 0.74 0.08 0.93 DM 13.78 1.84 0.1 0.94 SE 0.87 0.49 0.04

222 A.P. Singh Chouhan et al. / Fuel 109 (2013) 217–224

Author's personal copy

FAMENaOH (Nitrogen: 23.56 kJ/mol, 17.15 kJ/mol and 17.81 kJ/mol,SE 2.03; Air: 30.05 kJ/mol, 18.33 kJ/mol and 22.69 kJ/mol, SE3.42) are higher than their blends with petroleum diesel but lower than the refined JCO irrespective of the medium of flow as shown in Tables 2 and 3, respectively. These values are however, lower than the reported values by Jain and Sharma [8]. This may bedue to complete conversion of the refined JCO to corresponding es- ters and absence of free fatty acids in their compositions .

This is a well known fact that the activation energy is inversely proportional to the thermal stability of a fuel and the thermal stability has been found in the order: refined JCO > FAME KOH

� FAMENaOH > B50 FAME(KOH) � B50FAME(NaOH) > petroleum diesel.The calculated values are dependent upon the regression coeffi-cient R2. The values appears most accurate when the R2 approachesto one. However, oxidation stability and biodegra dability may not follow the same order because JCO and FAME being plant origin are easily oxidizable by microorganism . To the contrary, petroleum diesel being matured undergro und hydrocarbo ns for millions ofyears are less prone to oxidation by biological means. This requires further investiga tion to co-relate thermal stability with biological stability using activation energy concept.

It has also been ascertained from Tables 2 and 3 that the activa- tion energy calculated from the three methods are not exactly identical for any of the fuel under test in either air or nitrogen envi- ronment. Thus, it is very difficult to ascertain the exact quantitative value of activation energy through thermogr avimetric analysis.Therefore, the SE calculatio ns for the three methods are found more appropriate to be introduce d as presente d in Tables 2 and 3.

4.5. Frequency factor (A)

The frequency factor of a reaction expressed by ‘A’ which nor- mally signifies the degree of collisions of a reaction per second are presented in Tables 2 and 3. It has been observed that the fre- quency factor for JCO is highest under all the three methods ofstudy in both modes of operation although there are variations intheir quantitative value. It may be attributed to the bulky

molecula r structure of the triglycerides as compared to the other fuels under study that facilitate s increased frequency with increas- ing temperature. Topa [28] investigated in detail about the activa- tion energy, frequency factor, order of reaction and regression coefficient of canola oil, its biodiesel and diesel using several meth- ods. It was reported that the frequency factor varied from method to method and found 1.74 � 101–2.96 � 103 s�1, while, the activa- tion energy varies from 51.67 to 97.67 kJ/mol with R2 equal to0.93–0.99 for canola oil. The frequency factor for refined JCO, how- ever, varies from 5.726 � 101 s�1 to 8.01 � 102 s�1 in air medium with R2 varies from 0.91 to 0.95, while, 1.36 � 101 s�1 to1.77 � 101 s�1 in nitrogen medium with R2 varies between 0.91 and 0.97. This clearly unveil that the frequency factor calculated for different fuels may vary within certain undefined limit.

4.6. Kinetic rate constant(K)

The reaction rate constant normally denotes how fast or slow the reaction is moving and the rate constant calculated for the fuels under tests are presented in Tables 2 and 3. The rate constants ofthe fuels under study in air and nitrogen medium showed that the values are less than one in most cases. This implies that there is an overall drive towards product formatio n during thermal applicati on either in nitrogen or air medium of TGA operation .Due to scarce literature available for rate constant calculation,these values may require further critical study with different other fuels.

5. Conclusion

The kinetics of thermal decompositi on reactions of carbona- ceous liquid materials (i.e. hydrocarbon or biofuels) are complex as the process involves a large number of reactions such as com- bustion, thermal degradat ion and phase change. Although TGA provides general informat ion on the overall reaction kinetics,rather than individual reactions, it can be used as a tool for com- parison of kinetic data of various fuels. Different methods available

Table 3Kinetic parameters calculated from TGA thermograms for fuels under test in air medium.

Fuels under test Kinetic methods Kinetic parameter R2

Activation energy, E (kJ/mol) Frequency factor (A) s�1 Rate constant (K)

Refined JCO CRM 49.76 740.45 0.11 0.91 FM 43.18 801.59 0.36 0.97 DM 49.97 57.26 0.12 0.93 SE 2.64 238.85 0.08

Petroleum diesel CRM 16.47 3.91 0.07 0.92 FM 9.11 1.32 0.19 0.95 DM 10.88 1.33 0.11 0.97 SE 2.22 0.86 0.03

Biodiesel (NaOH) CRM 30.05 38.06 0.11 0.94 FM 18.33 20.62 0.42 0.96 DM 22.69 12.2 0.11 0.99 SE 3.42 5.07 0.1

Biodiesel (KOH) CRM 30.39 64.62 0.12 0.96 FM 14.53 0.54 0.02 0.93 DM 24.71 10.44 0.05 0.89 SE 4.64 19.93 0.05

B50 (NaOH) with petroleum diesel CRM 17.66 6.72 0.2 0.99 FM 14.72 2.84 0.11 0.93 DM 16.43 4.69 0.12 0.97 SE 0.85 1.12 0.03

B50 (KOH)with petroleum diesel CRM 19.54 8.79 0.27 0.92 FM 10.18 0.5 0.06 0.99 DM 13.51 1.79 0.12 0.97 SE 2.74 2.57 0.06

A.P. Singh Chouhan et al. / Fuel 109 (2013) 217–224 223

Author's personal copy

for reaction kinetics should be used such that they can give appro- priate information regarding the combusti on behavior o the fuel.

TGA thermograms used for the analysis of activation energy, ki- netic rate constant and frequency factor for the different fuels indi- cate the stability of the JCO and biodiesel samples at the different temperature ranges are beneficial for the determinati on of the thermal behavior of the tested fuels. This study can be advanta- geous for the combustion chamber design, modification and blend- ing purposes. Operating temperature is very effective for the decompositi on and combustion of different chemical compositions of the biodiesel and blends. Although the kinetic paramete rs such as activation energy are useful in determini ng the thermal and oxi- dative stability of a fuel, but this is not possible to ascertain the ex- act quantitative value and hence the range may be considered.However, it must also be kept in mind that combustion in a CI en- gine and TGA processes are not similar and, therefore, the two pro- cesses cannot be identical.

Acknowled gments

The authors acknowled ge the Ministry of New and Renewable Energy, Govt of India, New Delhi for financial assistance in the form of a project grant (Sanction F No. 7/144/2009 -NT dt. 01.10.2010).The authors are also thankful to the Indian oil Corporat ion Limited (IOCL) Jalandhar , Punjab India for providing the facility for some ofthe quality test of biodiesel in their laboratory.

References

[1] Changbo L, Wenli S, Weigang L. Kinetics of biomass catalytic pyrolysis.Biotechnol Adv 2009;27:583–7.

[2] Singh RK, Padhi SK. Characterization of jatropha oil for the preparation ofbiodiesel. Nat Prod Radian 2009;8(2):127–32.

[3] Berchmans, Hanny J, Hirata S. Biodiesel production from crude Jatropha curcas L. seed oil with a high content of free fatty acids. Bioresour Technol 2008;99(6):1716–21.

[4] Sarma AK, Sarmah JK, Barbora L, Kalita p, Chatterjee C, Mahanta, et al. Recent inventions in biodiesel production and processing – a review. Recent patent onEngin, vol. 2; 2008. p. 47–58.

[5] Chen YH, Tang TC, Chiang TH, Huang Bo-Yu, Chang Ching-Yuan, Chiang Pen- Chi, et al. A complementary biodiesel blend from soap nut oil and free fatty acids. Energies 2012;5(8):3137–48.

[6] Tiwari P. Isothermal and nonisothermal kinetic analyses of mahogamy oil shale with TGA. Biotechnol Adv 2009;27:583–7.

[7] Silva SA, Conceio MM, Souza AG, Prasad S, Silva Marta CD, Fernandes VJ, et al.Thermal analysis of the powder and the bran of algaroba. J Therm Anal Calorim 2004;75(2):411.

[8] Jain S, Sharma MP. Oxidation and thermal behavior of Jatropha curcas biodieselinfluenced by antioxidants and metal contaminants. Int J Eng Sci Technol 2011;3(4):65–75.

[9] Crnkovic PM, Koch C, Ávila I, Mortari DA, Cordoba AM, Moreira dos SA.Determination of the activation energies of beef tallow and crude glycerin combustion using thermogravimetry. Biomass Bioenergy 2012;44:8–16.

[10] Andrade Albuquerque RD, Pozzebom E, Faria EA, Filho FD, Ziani Suarez AZ,Soares do Prado AG. Thermal behavior of diesel/biodiesel blends of biodiesel obtained from buriti oil. Acta Scientiarum Technol Maring v Maring 2012;349(2):243–8.

[11] Iyer PVR. Biomass thermo-chemical characterization. 3rd ed. New Delhi: IIT Delhi; 2002.

[12] Ke-Ke Yang, Xiu-Li Wang, Yu-Zhong Wang, Bo Wu, Yong-Dong Jin, Bing Yang.Kinetics of thermal degradation and thermal oxidative degradation of poly (p-dioxanone). Eur Polym J 2003;39:1567–74.

[13] Yalcm Tonbul. Thermogravemetric investigation of the dehydration kinetics ofKSF, K10 and Turkish Bentonite. Turk J Chem 2001;25:333–9.

[14] Holland BJ, Hay JN. The value and limitations of non-isothermal kinetics in the study of polymer degradation. Thermochim Acta 2002;388:253–73.

[15] Singh BK, Kumari P, Prakash A, Adhikari D. Thermal decomposition kinetics ofpeanut shell. Nat Sci 2009;7:73–8.

[16] Barkia H, Belkbir L, Jayaweera AA. Non-isothermal kinetics of gasification byCO2 of residual carbon from Timahdit and Tarfaya oil shale kerogens. J Therm Anal Calorim 2004;76(2):623–32.

[17] Georgiera V, Zvezdova D, Vlaev L. Non-isothermal kinetics of thermal degradation of chitosan. Chem Cent J 2012;6:81.

[18] Galway AK. Perennial problems and promising prospects in the kinetics analysis of non-isothermal rate data. Thermochim Acta 2003;407:93–103.

[19] Krishnan D, Dass DM. A kinetic study of biodiesel in waste cooking oil. Afr JBiotechnol 2012;11(41):9797–804.

[20] Aboulkas A, Elharfi K. Study of the kinetics and mechanism of thermal decomposition of moroccan tarfaya oil shale and its kerogen. J Oil Shale 2008;25(4):426–43.

[21] Shukla R, Agrawal P, Kumar A. Crystallization kinetics in glassy Se100-x Inx system using iso-conversional methods. Chalcogen Lett 2010;7(4):249–55.

[22] Jain S, Sharma MP. Application of thermogravemetric analysis for thermal stability of Jatropha curcas biodiesel. Fuel 2012;93:252–7.

[23] Berrios M, Siles J, Martin MA, Martin A. A kinetic study of the esterification offree fatty acids (FFA) in sunflower oil. Fuel 2009;86:2383–8.

[24] Avila RN, Sodre JR. Physical-chemical properties and thermal behavior offodder radish crude oil and biodiesel. Ind Crop Prod 2012;38:54–7.

[25] Mazumdar P, Borugadda VB, Goud VV, Sahoo L. Physical–chemicalcharacteristics of Jatropha curcas L. of North East India for exploration ofbiodiesel. Biomass Bioenergy 2012;46:546–54.

[26] Sarma AK, Konwer D, Bordoli PK. A comprehensive analysis of fuel properties from koroch seed oil. Energy Fuels 2005;19:656–7.

[27] Aligrot C, Champoussin JC, Guerrassi N, Claus G. Prediction of self-ignition delay of different liquid diesel fuels. In: Int Symp COMODIA, vol. 94; 1994. p.331e6.

[28] Topa EH. Thermal characterization and kinetics of diesel, methanol route biodiesel, canola oil and diesel–biodiesel blends at different blending rates byTGA and DSC: a thesis. The graduate school of natural and applied sciences ofmiddle East Technical University, Turkey. <https://etd.lib.metu.edu.tr/upload/ 12612460/index.pdf> (19.11.12).

224 A.P. Singh Chouhan et al. / Fuel 109 (2013) 217–224