Characterization of Biodiesel and Biodiesel Particulate Matter byTG, TG-MS, and FTIR

Yi-Chi Chien,*,† Mingming Lu,‡ Ming Chai,‡ and F. James Boreo§

Department of EnVironmental Engineering and Science, Fooyin UniVersity,Kaohsiung County 831, Taiwan, and Department of CiVil and EnVironmental Engineering and Department

of Materials Science and Engineering, UniVersity of Cincinnati, Cincinnati, Ohio 45221

ReceiVed May 23, 2008. ReVised Manuscript ReceiVed October 25, 2008

Biodiesel is a potential renewable and carbon-neutral alternative to fossil fuels, and it is environmentallyand economically attractive. This paper studies the decomposition kinetics of biodiesel using thermal gravimetricanalysis (TGA) in one-stage pyrolysis. Biodiesel can be decomposed at 119-237 °C. The kinetic parametersfor biodiesel pyrolysis were obtained from the TGA experiments. The global rate equation for biodiesel pyrolysiscan be expressed as dX/dt ) 2.6 × 107 exp(-16.2/8.314 × 10-3T)(1 - X)0.52 (X denotes the reaction conversion).Characteristics of diesel and biodiesel and the associated diesel particulate matter (DPM) emitted from a nonroaddiesel generator were also analyzed by Fourier transform infrared (FTIR) spectroscopic methods. The FTIRspectra of biodiesel showed a CdO stretching band of methyl ester at 1743 cm-1 and C-O bands at 1252,1200, and 1175 cm-1. Furthermore, the FTIR spectra of DPM were similar to those of the fuels, an indicationthat the chemical structures of DPM are closely related to the fuel and engine oil properties, consistent withour previous study. The temperature series of 11 fragments have been analyzed in nitrogen, which include m/z29, 31, 44, 74, 87, 105, 143, 263, 294, 296, and 298. The fragments at m/z 294, 296, and 298 represent themethyl ester components of biodiesel, and the fragment at m/z 44 is carbon dioxide, fragments at m/z 29 and105 represent aldehyde compounds, and fragments at m/z 87 and 143 are shorter chain methyl esters, all ofwhich can be byproducts from biodiesel combustion. The fragments at m/z 57, 67, 95, and 109 representhydrocarbon components, which may be fragmented from the long carbon chains of methyl esters, and thefragment at m/z 31 is a methoxy group, which may be fragmented from methyl esters. The information ofTG-MS as analyzed above can offer a better understanding of the byproduct formation mechanisms of biodieselcombustion.

1. Introduction

With the gradual depletion of fossil fuels, the developmentof renewable energy sources is of increasing importance.Biodiesel is defined as the monoalkyl ester derivatives of longchain fatty acids, produced from oil crops, waste cooking oil,or animal fat via a relative simple transesterification process.The transesterification process is the reaction of a lipid with ashort chain alcohol such as methanol to form esters andbyproducts.1-3 Biodiesel is a renewable and biodegrable fueland is becoming environmentally and economically attractive.1

The biodegradability of biodiesel is evidenced by the muchshorter half-life comparing with petroleum diesel. As anexample, the half-life of approximately 100 ppm B20 inrainwater is 6.8 days.4 Due to its high solubility and biodegrada-

tion, biodiesel has a potential to be used in bioremediation toremove oils or polycyclic aromatic hydrocarbons (PAHs) incontaminated soils.5,6

Biodiesel has an energy density that is slightly less than thatof petroleum diesel. It has been proven that it is a technicallysufficient alternative diesel fuel in the fuel market. In Europeancountries and the United States, biodiesel has been acceptedfor use in automobiles, ships, and heating systems.7 Generally,biodiesel can be used either as a direct substitute for fossil fuelsor blended with different amounts of petroleum diesel incompression-ignition engines without any modification withenvironmental and economic advantages.8 In a biodiesel/dieselblend, the greater the percentage of biodiesel present, the greaterthe reduction of air pollution emission observed, except NOx

and carbonyls.9,10 Kulkarni and Dalai also have reported thatthe biodiesel produced from waste cooking oil gives betterengine performance and emits less emission when tested on* To whom correspondence should be addressed. Tel 011-8867-783-

0542. Fax 011-8867-782-1221. E-mail PL036@mail.fy.edu.tw.† Fooyin University.‡ Department of Civil and Environmental Engineering, University of

Cincinnati.§ Department of Materials Science and Engineering, University of

Cincinnati.(1) Conceicao, M. M.; Fernandes, V. J.; Araujo, A. S.; Farias, M. F.;

Santos, I. M. G.; Souza, A. G. Energy Fuels 2007, 21, 1522–1527.(2) Hurley, M. D.; Ball, J. C.; Wallington, T. J.; Toft, A.; Nielsen, O. J.;

Bertman, S.; Perkovic, M. J. Phys. Chem. A 2007, 111, 2547–2554.(3) Encinar, J. M.; Gonzalez, J. F.; Rodrıguez-Reinares, A. Fuel Process.

Technol. 2007, 88, 513–522.(4) Prince, R. C.; Haitmanek, C.; Lee, C. C. Chemosphere 2008, 71,

1446–1451.

(5) Taylor, L. T.; Jones, D. M. Chemosphere 2001, 44, 1131–1136.(6) Pereira, G.; Mudge, S. Chemosphere 2004, 54, 297–304.(7) Cetinkaya, M.; Karaosmanoglu, F. Energy Fuels 2005, 19, 645–

652.(8) Conceicao, M. M.; Fernandes, V. J.; Bezerra, A. F.; Silva, M. C. D.;

Santos, I. M. G.; Silva, F. C.; Souza, A. G. J. Therm. Anal. Calorim. 2007,87, 865–869.

(9) Conceicao, M. M.; Candeia, R. A.; Silva, F. C.; Bezerra, A. F.;Fernandes, V. J.; Souza, A. G. Renew. Sust. Energy ReV. 2007, 11, 964–975.

(10) Chai, M.; Lu, M. The Proceedings of the 100th A&WMA AnnualConference and Exhibition; 2007; Paper No. 558.

Energy & Fuels 2009, 23, 202–206202

10.1021/ef800388m CCC: $40.75 2009 American Chemical SocietyPublished on Web 12/05/2008

commercial diesel engines.11 Furthermore, neat biodiesel is freefrom sulfur and yields a decrease of 48% CO, 55% hydrocarbon,and 53% particulate matter.12 Wang et al. has performed a studythat compared exhaust emissions from nine heavy trucks fueledby petroleum diesel and biodiesel blend without engine modi-fication. They have indicated that the heavy duty trucks fueledwith biodiesel blend emitted lower particulate matter, CO, andhydrocarbon than the same trucks fueled with No. 2 diesel.13

In addition, it has been reported that the engine performance isalmost unaffected by the use of biodiesel or biodiesel/dieselblend.9

Different vegetable oils, such as corn oil, castor oil, sunfloweroil, and soybean oil, have been used to produce biodiesel inBrazil. Some thermal behaviors of the corn biodiesel, babassubiodiesel, castor oil biodiesel, and the soybean biodiesel havebeen investigated. For example, the evaporation temperature ofbabassu biodiesel starts around 52 and 60 °C in air and nitrogen,respectively.14 Castor biodiesel presents stability up to 150 °C.Its thermal profile may be altered by the decomposition processdue to the formation of intermediate compounds.1 Dantas et al.indicates that the activation energy and pre-exponential constantof the corn biodiesel decomposition in synthetic air atmosphereare 73.94-87.65 kJ ·mol-1 and from 8.8 × 104 to 3.2 × 106

s-1, respectively.15 Because of the possibility of increasing useof biodiesel, it is important to quantify the kinetics parametersof biodiesel pyrolysis.1,14-16 However, the kinetics of thebiodiesel made from soybean, a main feedstock in the UnitedStates, is not well understood.

In recent years, thermal analysis tools including thermo-gravimetry (TG) and differential scanning calorimetry (DSC)are becoming important in providing useful information in termsof kinetic parameters, thermal stability, etc.17 Continuous real-time information regarding the weight loss and gaseous emissioncan be obtained by coupling mass spectrometry to the thermo-gravimetry (TG-MS).18 It also gives the advantage of avoidingthe multistage sample preparation such as the digestion ofsample and can provide more detailed information on thedecomposition of biodiesel compounds by identifying fragmentsions from the original compounds with temperature changing.19

In addition, characteristics of fresh biodiesel and the dieselparticulate matter (DPM) emitted from the nonroad engine werealso analyzed using Fourier transform infrared (FTIR) spec-troscopy. Thus, the objective of this work is to obtain thepyrolysis kinetics of biodiesel using TG analysis and understand

the fragmentation patterns of biodiesel, which can be used tohelp interpret byproduct formation in biodiesel combustionstudies.

2. Experimental Methods

The sample of soybean biodiesel (BD-100, Nexsol biodiesel) usedin this study was purchased from Peter Cremer Co. The samplewas used as received for TG and was blended with diesel fuel foruse in the nonroad diesel engine. Ultimate analysis of the biodieselsample was determined by OKI Analytical, a commercial laboratoryin Cincinnati, OH.

Preliminary pyrolysis kinetics of biodiesel was determined usingTG (TA Instrument 5100/Dynamic TGA 2950, with the capabilityof determining weight loss and temperature difference simulta-neously). Approximately 4-6 mg of the biodiesel sample washeated from room temperature to 400 °C at heating rates of 3, 5, 8,and 10 °C ·min-1, respectively, in 100 mL min-1 high puritynitrogen. When the experiment was finished, the furnace powerwas turned off but the carrier gas was kept flowing until the furnacewas cooled down to room temperature. Aluminum oxide was usedas a reference in all TG experiments.

The TG-MS experiments were performed simultaneously usinga thermogravimeter (STA 409 CD, Netzsch Instruments, Inc.) anda quadrupole mass spectrometer (QMA 400, Balzers Instruments,Inc.). A Skimmer coupling system (Netzsch Instruments, Inc.) isequipped to combine these two instruments together. About 2-8mg biodiesel was decomposed with TG and the gas products wereintroduced to the mass spectrometry for obtaining evolution curves.The sample was heated up to 400 °C at a heating rate of 5 °Cmin-1 in 100 mL min-1 of high purity nitrogen.

In addition, the functional groups of biodiesel and DPM wereinvestigated by using FTIR spectroscopy (Nicolet Nexus 870 FT-IR, Thermo Electron Corp.). The DPM was collected from anonroad Generac diesel generator (1992, Model SD080, model No.92A-03040-S), which operated at 0, 50, and 75 kW for idle, low-load, and high-load modes, respectively. Detailed procedures ofDPM collection have been published elsewhere and are brieflydescribed here.20 DPM samples were collected by a high volumesampler at a flow rate of approximately 300 L min-1 with asampling time of approximately 30 min. Approximately 15 mg ofDPM was dissolved in dichloromethane (DCM) solution with sonicvibration for 30 min. Then the solution was concentrated to about1 mL from which three drops were placed onto a potassium bromide(KBr) pellet for the FTIR analysis. In addition, a droplet of thediesel or biodiesel sample was placed onto a KBr pellet andmeasured by FTIR spectroscopy. For all spectra reported, a 64 scandata accumulation was conducted at a resolution of 4 cm-1.

3. Results and Discussion

3.1. Ultimate Analysis. Ultimate analysis of the commercialbiodiesel sample is shown in Table 1. Note that the biodieselcontains approximately 8% oxygen, which is not present inpetroleum diesel. In addition, there was no sulfur detected. Thehigh heating value of the biodiesel is 39 594 kJ kg-1 (17 038Btu lb-1).

(11) Kulkarni, M. G.; Dalai, A. K. Ind. Eng. Chem. Res. 2006, 45, 2901–2913.

(12) Haas, M. J.; Scott, K. M.; Alleman, T. L.; McCormick, R. L. EnergyFuels 2001, 15, 1207–1212.

(13) Wang, W. G.; Lyons, D. W.; Clark, N. N.; Gautam, M.; Norton,P. M. EnViron. Sci. Technol. 2000, 34, 933–939.

(14) Santos, N. A.; Tavares, M. L. A.; Rosenhaim, R.; Silva, F. C.;Fernandes, V. J.; Santos, I. M. G.; Souza, A. G. J. Therm. Anal. Calorim.2007, 87, 649–652.

(15) Dantas, M. B.; Conceicao, M. M.; Fernandes, V. J.; Santos, N. A.;Rosenhaim, R.; Marques, A. L. B.; Santos, I. M. G.; Souza, A. G. J. Therm.Anal. Calorim. 2007, 87, 835–839.

(16) Candeia, R. A.; Freitas, J. C. O.; Souza, M. A. F.; Conceicao, M. M.;Santos, I. M. G.; Soledade, L. E. B.; Souza, A. G. J. Therm. Anal. Calorim.2007, 87, 653–656.

(17) Chien, Y. C.; Shih, P. H.; Hsien, I. H. EnViron. Eng. Sci. 2005,22, 601–607.

(18) Otero, M.; Dıez, C.; Calvo, L. F.; Garcıa, A. I.; Moran, A. BiomassBioenerg. 2002, 22, 319–329.

(19) Ehen, Z.; Novak, C.; Sztatisz, J.; Bene, O. J. Therm. Anal. Calorim.2004, 78, 427–440.

(20) Liang, F.; Lu, M.; Keener, T. C.; Liu, Z.; Khang, S. J. J. EnViron.Monit. 2005, 7, 983–988.

Table 1. Typical Biodiesel Composition and Energy Content

element analysis (wt %)

carbon 79.01hydrogen 12.90oxygen 8.04 (balanced)nitrogen 0.02sulfur not detectedchlorine 0.03

high heating value 39 594 kJ ·kg-1 (17 038 Btu · lb-1)

Characterization of Biodiesel and DPM Energy & Fuels, Vol. 23, 2009 203

3.2. Infrared Analysis of Fuels. The FTIR spectra of B50biodiesel (50% biodiesel and 50% diesel blend), diesel, andB100 biodiesel (100% biodiesel) are shown in spectra a, b, andc of Figure 1, respectively. The main components of diesel arealiphatic hydrocarcons, whose chemical structures are similarto long carbon sides chain of the main components of biodiesel.The aliphatic hydrogen at 2928 and 2856 cm-1 are indicated inFigure 1. The observation of an absorption peak at 727 cm-1

suggested the CH2 out-of-plane bending. In addition, since thebiodiesel is mainly monoalkyl ester, the intense CdO stretchingband of methyl ester appears at 1743 cm-1. The medium C-Obands at 1252, 1200, and 1175 cm-1 are also expected in Figure

1a,c. The absorbance at 3010 cm-1 indicated the HCdCH bondassociation, while absorbance at 1376 cm-1 indicated the -CH3

bond.3.3. Thermogravimetric Analysis. In order to have a better

understanding of the biodiesel decomposition, the thermaldegradation behavior of the biodiesel molecules was investi-gated. Typical TG curves of the biodiesel decompositionobtained in this study are shown in the top of Figure 2.Experimentally, the thermal decomposition of biodiesel atheating rates of 3, 5, 8, and 10 °C min-1, respectively, are in asingle step that describe the evaporation and decomposition ofbiodiesel. It is consistent with that of corn-oil biodiesel(produced by the ethanol routes) pyrolysis. However, thedecomposition begins at 119-125 °C and is completed at212-237 °C with no residue left that is lower than that of corn-oil biodiesel pyrolysis.15 The heating rate is a crucial factoraffecting the decomposition results. The higher heating rate maydecrease the distribution of the heat in the biodiesel moleculesand make the decomposition start at higher temperature. Thus,the shape of the TG curves, the initial decomposition temper-ature, and the temperature for a given weight loss are increasingwith the increasing heating rates (shown in Figure 2, top). Inthe bottom of Figure 2, DTG curves of biodiesel decompositionpresented one transition in the peak temperature of 183-219°C. The peak temperature was also increased with the higherheating rate.

3.4. Kinetics Calculation. The kinetic parameters for thepyrolysis of biodiesel molecules were calculated on the basisof weight loss data at different heating rates (from roomtemperature to 400 °C at 3, 5, 8, and 10 °C min-1). Thecalculation was based on the classical law of kinetics and theoverall rate equation is expressed in the Arrhenius equation

dXdt

)A exp(- Ea

RT)(1-X)n

X)w0 -w

w0 -wf

k)A exp(- Ea

RT)where t ) time (min), A ) pre-exponential factor (min-1), Ea

) activation energy (kcal mol-1), T ) reaction temperature (K),R ) gas constant, w ) mass of the biodiesel sample at time t,w0 ) initial mass of the samples, wf ) final mass of the samples,X ) conversion, n ) reaction order for the unreacted sample,and k ) rate constant (min-1).

The kinetic analysis indicates that the activation energy andpre-exponential constant of the biodiesel sample are 68.75 kJmol-1 (16.2 kcal mol-1) and 2.6 × 107 min-1, respectively.The activation energy is slightly lower than that of the methanol-based biodiesel (71.51-88.66 kJ mol-1) and ethanol-basedbiodiesel (80.71-92.84 kJ ·mol-1), as reported by an oxidativeTG study in Brazil.8 The lower activation energy may beassociated with better combustion properties and also as a resultof a different feed stock. In addition, the reaction order of thedecomposition of biodiesel sample is 0.52. The decompositionof biodiesel molecules can be satisfactorily described by thefollowing rate equation:

dXdt

) 2.6 × 107 × exp(- 16.2

8.314 × 10-3T)(1-X)0.52

The conversion and reaction time are the important informa-tion related to optimize the engine operation conditions. On theother hand, the decomposition temperature is the key factor thataffects the exhaust gas distribution.

Figure 1. Typical FTIR spectra of the (a) B50, (b) diesel, and (c)biodiesel sample.

Figure 2. TG curves (top) and DTG curves (bottom) of biodieseldecomposition at heating rates of (a) 3 °C min-1, (b) 5 °C min-1, (c)8 °C min-1, and (d) 10 °C min-1 in nitrogen.

204 Energy & Fuels, Vol. 23, 2009 Chien et al.

3.5. Infrared Analysis of DPM. The chemical structure ofDPM emitted from the petroleum diesel and biodiesel used ina nonroad diesel engine operated in different loads was alsoinvestigated by FTIR spectroscopy (shown in Figure 3, top). Ina previous study, we indicated that the chemical compositionof DPM emitted from the same diesel engine may be acombination of fuel evaporation and engine oil generated bycombustion and is mostly alkanes, PAHs, and carboxylic acids.20

Therefore, the FTIR spectra of DPM bear some resemblance tothat of diesel fuel (shown in Figure 3, top and Figure 1), withthe exception of the absorbance at 1747 cm-1, which indicateda carboxylic functional group found only in the DPM FTIRspectra. This is consistent with our previous studies that theengine oil contains carboxylic and benzoic acids not present indiesel.20 Furthermore, there was a small peak at 1577 cm-1 thatis due to a functional group containing a nitrogen atom. Overall,operation at different loads has not significantly altered the DPMcomposition, as the FTIR spectra of DPM at different loadsappear similar. However, the distributions of individual com-position at each load required further study, as FTIR is unableto provide this information.

Interestingly, the spectra of DPM emitted from the B50 usedin the nonroad engine also bore close resemblance to that ofthe B50 fuel (shown in Figure 3 bottom). The spectra alsoshowed the same trend in the different operation loads. It isnoted that the absorbances at 1648 and 1558 cm-1 were notfound in the spectra of B50. It may be due to the monosubsti-tuted amides in the solid state, which may be due to the minornitrogen composition in the biodiesel fuel.

3.6. Thermogravimetry-Mass Spectrometry (TG-MS)Analysis. The following m/z ratio (mass-to-charge ratio)analyses were obtained from TG-MS to better understand thedecomposition mechanisms of biodiesel: 29, 31, 44, 74, 87, 105,143, 263, 294, 296, and 298. The temperature series of selectedMS fragments are shown in Figure 4. It suggests that thedecomposition of biodiesel occurs at approximately 120 °C,which is consistent with the previous TG result. The m/z 294,296, and 298 fragments are considered molecular weight ionsof biodiesel components C18:0, C18:1, and C18:2, respectively.The intensities of these three molecular peaks are lower thanthose of other fragments peaks, which is consistent with MSspectra reference.21

Between 140 and 250 °C, m/z 74 has the highest ion current,followed by m/z 87. The other peaks follow the same trend:increasing under low temperature, showing a peak between 150and 240 °C, and slightly decreasing under high temperature.The m/z ratios not plotted in Figure 4 include m/z 57 (C4H9

+),which may be fragment of alkanes; m/z 41 (C3H5

+) and 55(C4H7

+), which may be fragments of alkenes; and m/z 67(C5H7

+), 95 (C7H11+), and 109 (C8H13

+), which may befragments of dienes. Those fragments are from C-C bondcleavage of the long carbon chain of methyl esters, dependingon the numbers and locations of the CdC double bonds of theoriginal esters.22

Peaks at m/z 87 (C2H4COOCH3+) and 143 (C6H12COOCH3

+)are shorter chain methyl esters, which are also from C-C bondcleavage of the carbon chain of methyl esters. These compoundsmay be expected from the byproducts of biodiesel decomposi-tion, as the shorter chain esters have been reported.23

A well-known McLafferty rearrangement could occur duringthe MS analysis due to a six-member ring structure of anintermediate, which will form the ion with m/z 74.24 Themechanism of this rearrangement is shown in Figure 5. The R1

group in the figure could be C12H25, C14H29, C14H27, or C14H25,representing the major fractions of biodiesel composition, suchas C16:0, C18:0, C18:1, and C18:2.

(21) Mass Spectrometry Data Centre. Eight Peak Index of Mass Spectra;The Royal Society of Chemistry. Nottingham, UK, 1983.

(22) Silverstein, R. M.; Webster, F. X.; Kiemle, D. SpectrometricIdentification of Organic Compounds; John Wiley & Sons: Hoboken, NJ,2005.

(23) Archambault, D.; Billaud, F. J. Chim. Phys. Phys.-Chim. Biol. 1999,96, 778–796.

(24) McLafferty, F. W. Interpretation of Mass Spectra; UniversityScience Books: Mill Valley, CA, 1980.

Figure 3. FTIR spectra of DPM emitted from the (top) diesel and(bottom) biodiesel used in the nonroad diesel engine in the load of (a)50 kW, (b) 0 kW, and (c) 75 kW.

Figure 4. TG-MS plots of selected ions of biodiesel and TG curve.

Characterization of Biodiesel and DPM Energy & Fuels, Vol. 23, 2009 205

Peaks at m/z 31 (methoxy, CH3O+) and 263 (C17H31CO+)may be fragments of C-O bond cleavage of C18:2 methyl ester(shown in Figure 6 with other selected MS ions). The pyrolysisof an oxygen-containing compound can result in formaldehydeand similar fragments.25 The m/z 29 fragment representsformaldehyde ion and m/z )105 represents benzalaldehydeion.26 This is an indication that aldehydes can be formed inbiodiesel decomposition and is consistent with observations fromengine tests.10,27

In addition to hydrocarbon fragments, the temperature seriesof CO2 (m/z 44) is also shown in Figure 6 (under nitrogenenvironment, however, we were not able to track the m/z ratioof CO). CO2 has been reported from pyrolysis of vegetable oil(triglycerides), which have similar structures as methyl esters(biodiesel).28,29 CO2 can be a unique byproduct in the pyrolysis

of oxygenated fuels. Therefore, the TG-MS information maybe helpful in understanding the decomposition mechanisms ofbiodiesel compounds and the potential byproducts formed.

4. Conclusions

The chemical functional groups of biodiesel and associatedparticulate emissions from a nonroad diesel engine have beendetermined by FTIR spectroscopy. The FTIR spectra of biodieselshowed an CdO stretching band of methyl ester at 1743 cm-1

and C-O bands at 1252, 1200, and 1175 cm-1, respectively.The FTIR spectra of its particulate emissions were largelysimilar to that of the fuel used. However, compared with dieselfuel, the FTIR spectra of the particulate emissions show theOdC-O functional group at 1747 cm-1 that may be due tothe n-alkanoic acids. The absorbances at 1648 and 1558 cm-1

were not found in the spectra of B50 that may be due to themonosubstituted amides formed in solid state in the thermaldecomposition process. The thermal behavior of biodieseldecomposition was also investigated in this paper with TG andTG-MS. The activation energy, pre-exponential constant, andreaction order of the pyrolysis of biodiesel sample are 67.75 kJmol-1 (16.2 kcal mol-1), 2.6 × 107 min-1, and 0.52, respec-tively. The global rate equation for pyrolysis of the biodieselcan be expressed as dX/dt ) 2.6 × 107 × exp(-16.2/8.314 ×10-3T)(1 - X)0.52 (X denotes the reaction conversion). Thetemperature series of 11 m/z fragments have been analyzed bymeans of TG-MS. On the basis of the analysis above, it isexpected that smaller esters, CO2, and aldehydes may be foundas byproducts of biodiesel pyrolysis. The methoxy group maybe fragmented from the methyl ester of biodiesel components,and many hydrocarbon fragments may be expected from thecarbon chain of biodiesel components. TG-MS can offer moreinformation on the expected byproducts from biodiesel decom-position in a quicker and simpler way than the actual pyrolyticexperiments and may be helpful in understanding the reactionmechanisms of byproduct formation.

Acknowledgment. The financial support of the Taiwan NationalScience Council is gratefully acknowledged. We are also gratefulfor the help provided by Mr. Michael Starr, University of Cincinnati,with the FTIR analysis and Prof. Soofin Cheng and Mr. Chung-Shen Kao, Taiwan University, with the TG-MS analysis.

EF800388M

(25) Luftl, S.; Archodoulaki, V. M.; Seidler, S. Polym. Degrad. Stab.2006, 91, 464–471.

(26) McLafferty, F. W.; Stauffer, D. B. The Wiley/NBS Registry of MassSpectral Data; John Wiley & Sons: New York, 1989.

(27) Krahl, J.; Baum, K.; Hackbarth, U.; Jeberien, H. E.; Munack, A.;Schutt, C.; Schroder, O.; Walter, N.; Bunger, J.; Muller, M. M.; Weigel,A. ASAE Meeting Paper No. 996136, 1999.

(28) Idem, R. O.; Katikaneni, S. P. R.; Bakhshi, N. N. Energy Fuels1996, 10, 1150–1162.

(29) Adebanjo, A. O.; Dalai, A. K.; Bakhshi, N. N. Energy Fuels 2005,19, 1735–1741.

Figure 5. Mechanism of McLafferty rearrangement.

Figure 6. Selected TG-MS ions and TG curve.

206 Energy & Fuels, Vol. 23, 2009 Chien et al.