Experimental and Modeling Study of Trends in the High-Temperature

Sep 8, 2011 - Experimental and Modeling Study of Trends in the High-Temperature Ignition of Methyl and Ethyl Esters. Benjamin ... Combined Ab Initio, ...
1 downloads 14 Views 1MB Size
ARTICLE pubs.acs.org/EF

Experimental and Modeling Study of Trends in the High-Temperature Ignition of Methyl and Ethyl Esters Benjamin Akih-Kumgeh* and Jeffrey M. Bergthorson Department of Mechanical Engineering, McGill University, Montreal, Quebec, Canada ABSTRACT: Methyl and ethyl esters of carboxylic acids can be obtained from the trans-esterification of fatty acid triglycerides using methanol and ethanol, respectively. In this work, the relative ignition of methyl and ethyl esters is reported, based on new hightemperature shock tube ignition data of ethyl acetate (EA) and ethyl propanoate (EP). These are compared with literature data for other methyl and ethyl esters, and isomer and alkyl group effects are also investigated. It is found that ethyl esters are generally characterized by shorter ignition delay times than those of methyl esters of the corresponding alkanoic acid. Increased reactivity of ethyl esters is also observed for isomeric ethyl and methyl esters. A combined high-temperature chemical kinetic model is proposed for methyl and ethyl acetates as well as for ethyl formate in order to shed light on the link between chemical structure and the observed reactivity trends. The proposed model reflects the experimental trends and generally predicts ignition delay times in close agreement with measured data. The reduced methyl acetate reactivity is attributed both to the absence of more reactive secondary CH sites and the ethyl group which facilitates complex unimolecular decomposition reactions in ethyl esters such as acid and ethylene elimination. The model and the new experimental data contribute toward improved understanding and modeling of the combustion properties of biodiesel surrogates.

’ INTRODUCTION A sustainable energy future will likely involve increased use of biofuels such as biodiesel. On the other hand, concerns about the associated costs and environmental neutrality prompt a closer look at biofuel production methods with the goal of improving the sustainability of the process. Although biodiesel is mainly produced by methanol-based trans-esterification of fatty acid triglycerides in acidic or alkaline media, the possibility of using other alcohols is being explored with increasing interest.1,2 Successful system-level tests of ethyl ester as biodiesel have been reported.2,3 The possibility of using mixtures of methanol and ethanol in the trans-esterification of soybean oil have been explored by Joshi et al.4 concluding that the resulting methyl and ethyl ester mix can offer improved physical properties. Further use of various alkyl esters beside methyl esters in combustion engines can be guided by understanding their relative combustion and physicochemical properties through systematic experimental studies. A review of the combustion chemistry of biodiesel surrogates by Lai et al.5 and biofuel combustion chemistry by Kohse-H€oinghaus et al.6 have been presented, highlighting the need for further studies. With respect to methyl and ethyl esters, previous studies by Metcalfe et al.7 explored the relative ignition of the ester isomers, ethyl propanoate (EP) and methyl butanoate (MB), concluding that EP is more readily ignited than MB under the same conditions. Their comparison is based on the same molar concentrations since these two esters are isomeric. Hakka et al.8 studied the ignition of methyl and ethyl butanoates in a shock tube. In their study, comparison is based on the fuel concentration. They concluded that the ethyl ester of butanoic acid is slightly more reactive than the methyl ester. As in the study of EP and MB by Metcalfe et al.,7 the authors link the increased reactivity of ethyl esters to the possibility of ethylene and alkanoic acid elimination by complex r 2011 American Chemical Society

unimolecular reaction. Such reactions typically proceed with barriers lower than those of direct unimolecular cleavage. Further, the products, ethylene and alkanoic acids, are more reactive than the esters. In a recent study, Bennadji et al.9 investigated the relative shock tube ignition of four methyl and ethyl esters of unsaturated esters—methyl and ethyl acrylates (CH2CHCOOCH3 and CH2CHCOOC2H5) as well as methyl and ethyl crotonates (CH3CHCHCOOCH3 and CH3CHCHCOOC2H5). Ignition delay times (ranging from 5 to 250 μs) at pressures of 7.09.6 atm were reported. The authors used the concentration of carbon atoms and the carbon/oxygen ratio as constraints to evaluate the relative behavior. Under these constraints, similar ignition behavior was observed. However, it was observed that the acrylates are more reactive than the crotonates. Because of the short chains of these esters, the methyl and ethyl effects are partly complicated by the low-activation barrier addition reactions of radicals to the carboncarbon double bonds, yielding reactive radicals. In a recent study of the ignition of selected C3 oxygenated hydrocarbons, the current authors showed that ethyl formate (EF) is more reactive than its ester isomer, methyl acetate (MA).10 The higher reactivity of EF is attributed to the presence of secondary CH bonds in EF, as well as the ability of EF to undergo complex unimolecular reactions to yield formic acid and ethylene, as mentioned previously. At the system level, Zhang et al.11 studied the relative ignition behavior of C9 fatty acid alkyl esters in a motored engine. They concluded that the lowtemperature heat release was higher for ethyl nonanoate than Received: July 4, 2011 Revised: September 4, 2011 Published: September 08, 2011 4345

dx.doi.org/10.1021/ef200977p | Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

ARTICLE

Figure 1. Methyl and ethyl esters considered in this study. In addition to these, reference is also made to methyl butanoate (MB).

observed for the methyl ester. It was also observed that the location of the double bond, CdC, influenced the overall reactivity. Reduced reactivity is observed for double bonds closer to the center of the molecule. As the short-chain esters are explored to improve understanding of ester combustion, there is a need to explore the trends in reactivity for a wide range of molecules with different molecular structures. Ethyl actetate (EA) ignition data have not been previously reported. Although ignition delay times of EP have been reported at pressures up to approximately 4 atm and under dilute conditions, further data at higher pressures are needed for further comparison with existing data of other short-chain esters.10,12 In this study, new experimental data for EP and EA are reported. Together with literature data for methyl formate (MF), EF, MA, and methyl propanoate (MP), their ignition behavior is explored to reveal the effects of methyl and ethyl ester groups on the same carboxylic acid. The structures of these methyl and ethyl esters are shown in Figure 1. Their relative ignition behavior is further put into perspective by considering the effect of ester isomers as exemplified by the three surrogates, MP, EP, and MB. These comparisons are done for mixtures of fuel, oxygen, and argon under conditions of constant equivalence ratios (ϕ), argon/oxygen ratios (D), and nominal pressures (p), over a range of postreflected shock temperatures (T). By adopting the constant argon/oxygen ratio (D), conclusions on the relative reactivity of these surrogates can be compared to those of fuel/air flame propagation studies, wherein the nitrogen/oxygen ratio in air is constant. To further examine this relative behavior, a chemical kinetic model is developed for MA, EA, and EF. The model is based on a C1C4 hydrocarbon combustion model by Wang et al.,13 excluding the chemistry of C4 species. Model predictions of ignition delay times and their trends are compared to experimental observations.

’ EXPERIMENTAL TECHNIQUE Ignition delay times were obtained using a 5-cm inner diameter shock tube described previously.14 Test mixtures were prepared in a 90-L mixing tank and allowed to mix for approximately 24 h. In the course of preparing the test mixtures, volumes of the injected liquid fuel were chosen such that the resulting partial pressure after complete evaporation was less than 50% of its vapor pressure at room temperature in order to avoid fuel condensation. The purities of EP and EA were 99+% and 99.5%, respectively. Those of the other fuels are reported in their earlier studies.10,12,14 The shock tube was helium-driven and the shock velocity was obtained from pressure signals of four fast-response pressure

Figure 2. End wall pressure and CH radical chemiluminescence measurements with corresponding ignition delay time, τ, for an EA/ O2/Ar mixture with ϕ = 2.0, D = 24, p = 10.3 atm, and T = 1162 K. The ignition delay time in this case is indicated by the star. transducers mounted 50 cm apart, close to the end wall. The shock velocity at the endwall was obtained by linear extrapolation of the measured attenuating discrete velocity values to the endwall. The postreflected shock temperature was calculated using the onedimensional shock equations in the GasEq software package 15 with initial compositions, thermodynamic conditions and the endwall shock velocity as input parameters. Uncertainty in the calculated temperature results from the uncertainties in the shock velocity and composition of the combustible gas. It is estimated that the average temperature uncertainty in this work is between 10 and 25 K. Ignition delay times were determined from CH emission profiles obtained using a photodiode with a filter centered at 430 ( 10 nm. Ignition delay time is defined as the time between the shock reflection from the endwall and the time corresponding to the intersection of the line of maximum slope in the CH profile to the initial photodiode signal prior to ignition onset. An example of the endwall profiles of pressure and CH emission as well as the ignition delay time definition is shown in Figure 2. In a previous study, 16 the pressure rise in the postreflected region was characterized for nonreactive oxygen/argon mixtures and found to be approximately 6%/ms. For ignition delay times less than 1 ms, the associated ignition delay reduction is within the limit of uncertainty and does not compromise the results of these studies. For much longer ignition delay times at temperatures below 1000 K, compressive heating of the reacting system leads to ignition delays significantly different from usual constant volume model predictions, as observed in a recent study by Heufer et al. 17 In this study, ignition delay times are obtained for EA and EP while the other data are obtained from previous work in our laboratory on alkyl ester ignition. 10,12,14,16 Representative ignition data for EA and EP are presented in the Appendix.

’ MODELING OF METHYL/ETHYL ACETATES AND ETHYL FORMATE To gain further insight into the oxidation of these methyl and ethyl esters, MA, EA, and EF are further considered. These esters are included in the chemical kinetic model for small alkyl esters proposed by Westbrook et al.18 However, a new chemical kinetic model is herein proposed for MA, EA, and EF to further demonstrate the role of structure on the reactivity trends. It is based on the chemical kinetic model for H2/CO/C1C4 4346

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

ARTICLE

hydrocarbons by Wang et al.,13 whereby reactions involving species with more than 3 carbon atoms have been excluded. The model is aimed at predicting high-temperature ignition, so that only the main reaction channels deemed important have been included. The fuel undergoes unimolecular reactions as well as hydrogen abstraction by radicals and molecular oxygen. The resulting radicals react mostly by unimolecular decomposition to lower carbon species, whose further oxidation chemistry blends with the detailed chemistry of C1C3 alkanes in the model by Wang et al.13 Whereas some of the elementary reactions involved in the process have been the subject of experimental and theoretical studies, others have not yet been addressed. The rate constants of the latter are assigned based on analogy with reactions primarily from the propane system, taken from the propane kinetic data recommendations by Tsang.19 Previous work by El Nahas et al.20 showed that the concerted unimolecular reaction of EP, yielding ethylene and propanoic acid, is the most favorable unimolecular reaction pathway. This channel is also expected to be important in the oxidation of ethyl formate and ethyl acetate, proceeding as shown in eqs 1 and 2. The rate for the EA reaction has been assigned based on work by Scheer et al.21 while that of the EF reaction is assigned based on the pyrolysis study of ethyl and isopropyl formates and acetates by Blades.22 EA H CH3 COOH þ C2 H4

ð1Þ

EF H HCOOH þ C2 H4

ð2Þ

The activation energies reported for reactions 121 and 222, are 46.580 and 44.172 kcal/mol, respectively. El Nahas et al.20 reported a barrier of 50.2 kcal/mol (210 kJ/mol) for the concerted elimination of ethylene and propanoic acid from ethyl propanoate. The products of these EA and EF concerted reactions are stable molecules, carboxylic acids and ethylene, whose further oxidation has been modeled by unimolecular decompositions and hydrogen abstraction by radicals and molecular oxygen. Ethylene chemistry is described in the base mechanism by Wang et al.23 Another concerted unimolecular reaction considered in fuel consumption reactions is the elimination of an alcohol and a ketene (or CO in the case of a formate). Methanol elimination has been addressed in modeling studies before.12,20,24,25 Barriers obtained for methyl esters have been used to assign rate parameters to ethanol elimination from EA and EF. EA H CH2 CO þ C2 H5 OH

ð3Þ

EF H CO þ C2 H5 OH

ð4Þ

Thermodynamic data for the species involved in reactions of the submechanisms have been estimated by Group Additivity methods using the THERM software package.26 The submechanisms are presented as in the Appendix. The proposed model is used hereafter to compare the ignition delay times of MA and EA as well as MA and EF. Simulations are carried out using the constant volume homogeneous reactor in CANTERA,27 whereby ignition delay times are obtained as the time from reaction onset to the time of maximum CH concentration. The present high-temperature kinetic model is developed as a basis for further reduced-order modeling of biodiesel surrogate fuels.

Figure 3. MP and EP ignition trends for stoichiometric mixtures with argon/oxygen ratios, D, of 14.3 and 3.8 at 10 atm. Symbols: for D = 14.3: MP12 (9), EP (b); for D = 3.8: MP12 (0), EP (O); lines are exponential fits to data.

’ RESULTS AND DISCUSSION New experimental data on EA and EP are presented in this section in comparison to the methyl ester of the same organic acid root, methyl acetate and methyl propanoate, respectively. The MPEP comparison is first presented, followed by that of MAEA. An instance of MF versus EF is shown while the section concludes with a comparison of MB, EP, and MP to further highlight methyl, ethyl, and isomer effects in a manner similar to the MA versus EF isomer comparison presented in previous work.10 In the second section, model predictions are compared with data and experimental correlations. Experimental Results. Results of the relative high-temperature ignition behavior are presented as ignition delay times plotted on a logarithmic scale against inverse temperatures for test mixtures with the same equivalence ratio, ϕ, argon/oxygen ratio, D, and nominal pressures, p. Variations in the postreflected shock pressure are corrected by a power-law based on the known dependence of ignition delay times on pressure τ µ pγ

ð5Þ

where γ is determined from correlating ignition delay times at various pressures. Usually for oxygenated fuels and other hydrocarbons, γ lies between 0.5 and 1.0 and the ignition delay time is not very sensitive to small variations in pressure. Figure 3 shows the ignition delay times of MP and EP at an average pressure of 10 atm for stoichiometric mixtures with two different argon/oxygen ratios, D, of 14.3 and 3.8. The data for MP are obtained from ref 12 while the EP data are from this study. As expected, the less dilute mixtures with D = 3.8, are characterized by shorter ignition delay times than those for mixtures with D = 14.3, as a consequence of higher effective oxygen and fuel concentrations. For each level of dilution, ignition delay times of MP are longer than those of EP, indicating higher reactivity of the latter. The increased reactivity of EP is thought to be associated with the presence of the ethyl group bonded to the carboxylic acid part of the ester. The ethyl group, with secondary CH bonds offers more vulnerable sites for attack by H-abstracting radicals such as H, OH, HO2, and CH3. Further, as discussed in the modeling 4347

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

ARTICLE

Figure 4. MP and EP ignition trends for stoichiometric mixtures with argon/oxygen ratios, D, of 14.3 at 1 atm. Symbols: MP12 (9), EP (b); lines are exponential fits to data.

Figure 6. Methyl and ethyl propanoate ignition trends for rich mixtures (ϕ = 2.0) with argon/oxygen ratios, D, of 24 at 10 atm. Symbols: MP12 (9), EP (b); lines are exponential fits to data.

Figure 5. Methyl and ethyl propanoate ignition trends for lean mixtures (ϕ = 0.5) with argon/oxygen ratios, D, of 7.9 at 4 atm. Symbols: MP12 (9), EP (b); lines are exponential fits to data.

Figure 7. MA and EA ignition trends for stoichiometric mixtures with argon/oxygen ratios, D, of 14.3 and 3.8 at 10 atm. Symbols: for D = 14.3: MA12 (9), EA (b); for D = 3.8: MA12 (0), EA (O); lines are exponential fits to data.

section previously, the concerted elimination of ethylene and propanoic acid in EP is more favorable than other unimolecular reactions that these esters can undergo. This leads to increased reactivity especially at reaction onset when the concentration of the radical pool is not high. This concerted reaction is not possible for MP. For the stoichiometric mixtures with D = 14.3, the same trend in relative ignition of MP and EP is observed at an average pressure of 1 atm as shown in Figure 4. To ignite mixtures at lower pressures, higher reflected shock temperatures are required. The difference in the reactivity of the two esters becomes less pronounced as the temperature increases beyond approximately 1400 K. Differences in global reactivity such as ignition delays can be traced back to differences in activation energies of controlling elementary reactions. As the temperature increases the number of collisions with energy in excess of the activation energy also increases until at much higher temperature almost all collisions possess energy in excess of the required threshold for chemical reactions. This partly explains the diminishing difference in reactivity with increasing temperatures as reactions tend toward collision frequency controlled reaction rates.

Figure 8. MA and EA ignition trends for stoichiometric mixtures with argon/oxygen ratios, D, of 14.3 at 1 atm. Symbols: MA12 (9), EA (b); lines are exponential fits to data.

Further MP and EP comparison is carried out at lean and rich conditions. In Figure 5, results are shown for mixtures with an equivalence ratio, ϕ, of 0.5 and an argon/oxygen ratio, D, of 7.9 at 4348

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

Figure 9. Methyl and ethyl acetate ignition trends for lean mixtures (ϕ = 0.5) with argon/oxygen ratios, D, of 7.9 at 4 atm. Symbols: MA12 (9), EA (b); lines are exponential fits to data.

Figure 10. Methyl and ethyl acetate ignition trends for rich mixtures (ϕ = 2.0)with argon/oxygen ratios, D, of 24 at 10 atm. Symbols: MA12 (9), EA (b); lines are exponential fits to data.

average pressures of 4 atm. Under these conditions, EP is also observed to have shorter ignition delay times than MP. In the case of rich mixtures (ϕ = 2.0) at D = 24 and average pressures of 10 atm the same relative trend is seen (see Figure 6). These results lead to a general conclusion that EP is more reactive than MP at temperatures above 1000 K. In the case of alkyl esters of acetic acids, the relative ignition of MA and EA is shown in Figure 7 for stoichiometric mixtures with D = 14.3 and D = 3.8 at average pressures of 10 atm. Similar to the findings for methyl and ethyl esters of propanoic acid, EA is found to be more reactive than MA. The ignition delay times of mixtures with D = 3.8 are also correspondingly shorter than those obtained with more dilute mixtures with D = 14.3. Further, Figure 8 shows that at the lower average pressure of 1 atm, though EA ignition delay times are shorter than MA, the difference is not as pronounced as at higher pressure. This is partly related to the fact that higher temperatures are accessed at lower pressures. At the higher temperature end in these experiments, both methyl and ethyl esters tend to have comparable ignition delay times.

ARTICLE

Figure 11. Methyl and ethyl formate ignition trends for stoichiometric mixtures with argon/oxygen ratios, D, of 18.8 at 12 atm. The methyl formate correlation is based on a previous study.14 Legend: MF correlation14 (solid line), EF10 (O), dashed line is fit to EF data.

Figure 12. Comparison of MB, MP, and EP isomer and alkyl ester effects for stoichiometric mixtures with argon/oxygen ratios, D, of 14.3 at 10 atm. Legend: MB12 (Δ), MP12 (9), EP (0). Lines are fits to experimental data.

Exploring the relative behavior of acetates at lean and rich conditions leads to the same conclusions presented above. In Figure 9, results are shown for lean mixtures (ϕ = 0.5) with D = 7.9 at average pressures of 4 atm. Under these conditions, EA also has shorter ignition delay times than MA. The same trend is observed in Figure 10 for rich mixtures (ϕ = 2.0) with D = 24 at average pressures of 10 atm. Drawing from ignition data and ignition correlations reported in the literature, MF and EF can also be compared, yielding the same conclusion as above. In the example in Figure 11, EF ignition data from ref 10 are compared with predictions using the MF ignition correlation from ref 14. Whereas EF is found to be more reactive than MF, MF appears to be characterized by a weaker temperature sensitivity than EF. It is possible that the absence of CH bonds on the nonester side of the carbonyl group in formates modifies the methyl/ethyl ester trends in formates. It is also worth noting that the global activation energy of MF was found to be lower than that of other methyl esters in ref 12. 4349

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

Figure 13. Comparison of predictions of the EP model by Metcalfe et al.7 for stoichiometric mixtures with argon/oxygen ratios, D, of 14.3 at 1 and 10 atm. Symbols: EP at 1 atm (9), EP by Metcalfe et al. (0), EP 10 atm (b); lines: EP at 1 atm (solid line), EP at 10 atm (dashed line).

Figure 14. Comparison of model predictions of MA and EA for stoichiometric mixtures with argon/oxygen ratios, D, of 14.3 at 10 atm. Symbols: MA12 (9), EA (b); lines: MA sim. (solid line), EA sim. (dashed line).

The methyl/ethyl effect and isomer effects are illustrated in Figure 12, where ignition data for MB from ref 16 have been used to compare with MP and EP. Similar to observations by Metcalfe et al. 7 at lower pressures of 1 and 4 atm, EP ignites more readily than its ester isomer, MB, under the conditions investigated. With respect to MP, MB shows longer ignition delay times, as discussed by the current authors in ref 12. The reason for this behavior is not completely understood. In Figure 13, predictions of the EP model by Metcalfe et al.7 are compared with the present data for stoichiometric mixtures with D = 14.23 at 1 and 10 atm. Experimental data at 1 atm reported by Metcalfe et al.7 are also added. It should be noted that in their work, the authors defined their ignition as the time to the maximum of the product of the radicals, C2H  O, whereas in the current study the authors define ignition delay times as the time to the maximum CH concentration. It is observed that model predictions are in good agreement with both atmospheric data sets. At higher pressures, predicted

ARTICLE

Figure 15. Comparison of model predictions of MA and EA for stoichiometric mixtures with argon/oxygen ratios, D, of 3.8 at 10 atm. Symbols: MA12 (9), EA (b); lines: MA sim. (solid line), EA sim. (dashed line).

Figure 16. Comparison of model predictions of MA and EA for lean mixtures (ϕ = 0.5) with argon/oxygen ratios, D, of 7.9 at 4 atm. Symbols: MA12 (9), EA (b); lines: MA sim. (solid line), EA sim. (dashed line).

ignition delay times are slightly longer than those that were measured. Extension of this work to other combustion properties is required for better understanding of the role the alkyl group has in the oxidation of alkyl esters. The investigation of combustion properties can complement the evaluation of alternative alcohols for trans-esterification of fatty acid into biodiesel. However, the gas-phase combustion properties must also be considered along with the physical properties of the resulting esters. Properties such as fuel viscosity and vapor pressures will affect injection, atomization, and droplet evaporation. These issues need to be addressed as progress is made toward a more sustainable biodiesel energy economy. It seems possible that ethyl ester oxidation will be accompanied by higher concentrations of intermediate organic acid species which may present increased demands on material compatibility. Evaluation of Model Performance. The proposed model is evaluated with respect to its prediction of ignition delay trends. Ignition predictions by the current model are first compared for MA 4350

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

Figure 17. Comparison of model predictions of MA and EA for rich mixtures (ϕ = 2.0) with argon/oxygen ratios, D, of 24 at 10 atm. Symbols: MA12 (9), EA (b); lines: MA sim. (solid line), EA sim. (dashed line).

Figure 18. Comparison of model predictions of MA and EF for stoichiometric mixtures with argon/oxygen ratios, D, of 18.8 at 12 atm. Legend: MA correlation12 (dash-dotted lines), EF10 (b); MA sim. (solid line), EF sim. (dashed line).

ARTICLE

Figure 19. Comparison of model predictions of MA and EF for lean mixtures (ϕ = 0.5) with argon/oxygen ratios, D, of 8.88 at 13 atm. Legend: MA12 (9), EF10 (b); MA sim. (solid line), EF sim. (dashed line).

pressures of 4 atm. The model also predicts a reduction in the difference between MA and EA reactivity at high temperatures as observed in the data. In Figure 17, however, there is poor agreement between model and experiment for rich mixtures (ϕ = 2.0) at average pressures of 10 atm. At higher temperatures, predicted ignition delay times of MA are longer than measured ignition delay times. The agreement for EA is better but still not as good as that observed for the other cases. The EF submodel in the current model is tested by comparing model predictions for MA and EF against experimental data or correlations derived from experiments as presented in refs 10 and 12. In Figure 18, model predictions are compared against EF experimental data and a correlation for MA from a previous study.12 The higher reactivity of EF is captured and the model predictions of EF are in good agreement with the experiment. There is also agreement between MA predictions and ignition delay times calculated using the correlation from ref 12 τ ¼ ð4:5  105 Þϕ0:28 ( 0:13 D0:99 ( 0:09 p0:74 ( 0:07 expð38:2 ( 1:5=RTÞ

ð6Þ and EA. Literature data and correlations for the ester isomers, MA and EF, are then used to compare with model predictions, to verify the trends discussed by the current authors in ref 10. In Figure 14, model predictions of MA and EA ignition delay times are compared with their respective measured delay times. This set of data was used in the development phase of the model; the model is subsequently tested against data under different conditions. It is observed that the agreement between the model and experiment in this case is good, with MA predicted to have longer ignition delay times than EA as observed in experiment. Figure 15 shows the comparison of the model predictions with measured ignition delay times for stoichiometric mixtures with D = 3.8 at an average pressure of 10 atm. It is also found that the model accurately predicts the observed ignition delay times of these two esters under the given conditions of high oxygen and fuel concentration as well as lower dilution. A similar level of agreement between model and experiment is observed in Figure 16, where ignition delay times are compared for lean mixtures (ϕ = 0.5) with D = 7.9 at average

where τ is the ignition delay time in μs, ϕ is the equivalence ratio, D is the ratio of the argon concentration to the oxygen concentration, p is the pressure in atm, 38.2 is the global activation energy in kcal/mol, and R = 1.986  103 kcal/(mol K) is the universal gas constant. Figure 19 shows the comparison between model and experiment for lean mixtures (ϕ = 0.5), with an argon/oxygen ratio, D, of 8.8 at pressure of 13 atm. Ignition delay times have been reported for MA and EF under these conditions, 10 thus these are directly compared with model predictions. It is observed that model predictions for both esters are in good agreement with experimental observations. However, at the lower pressure of 1 atm as shown in Figure 20, it is seen that the model predictions of MA ignition deviate from the measured data. Perhaps detailed treatment of pressure-dependent reactions in these submodels using RRKM or QRRK theories could improve this poor performance at low pressures such as 1 atm. Finally, for the rich mixtures (ϕ = 2.0) with D = 18.8, it is found that model predictions of EF are in good agreement (Figure 21). 4351

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

Figure 20. Comparison of model predictions of MA and EF for lean mixtures (ϕ = 0.5) with argon/oxygen ratios, D, of 8.88 at 1 atm. Legend: MA12 (9), EF10 (b); MA sim. (solid line), EF sim. (dashed line).

Figure 21. Comparison of model predictions of MA and EF for rich mixtures (ϕ = 2.0) with argon/oxygen ratios, D, of 18.8 at 12 atm. Legend: MA correlation12 (dash-dotted lines), EF10 (b); MA sim. (solid line), EF sim. (dashed line).

Because there are no MA ignition data under these conditions, the MA correlation from ref 12 is used. There is a fair level of agreement between the model and the correlation under these conditions. The observed differences in the reactivities of the esters as predicted by the model can be further put into perspective by considering stoichiometric fuel, oxygen, and argon systems with an argon/oxygen ratio of 14.3, undergoing ignition at a pressure of 10 atm and initial temperature of 1200 K. Three most important fuel consumption pathways are determined at the beginning of the simulations and a few microseconds just before ignition. In the case of MA, shown in Figure 22, the ignition delay time is 787 μs. At 10 μs after reaction onset, it is observed that the fuel is depleted mostly by H abstraction by CH3, resulting from direct cleavage reactions. Apart from the three most important reaction channels shown, other fuel consumption reactions at this time involve H abstraction by H atoms, O2, and OH radicals. Cleavage of the CO bond adjacent to the carbonyl has not been considered in the model on account of its higher bond energy, as shown in ref 12. At 760 μs, a sufficient radical pool is generated to attack the fuel.

ARTICLE

Figure 22. Fuel (MA) depletion pathways at two different time intervals prior to ignition for a stoichiometric MA/O2/Ar mixture at a pressure of 10 atm with an argon/oxygen ratio, D, of 14.3 at 1200 K. The ignition delay time is 787 μs, the species MAMJ is CH3COOCH2 and MA2J is CH2COOCH3.

Figure 23. Fuel (EA) depletion pathways at two different time intervals prior to ignition for a stoichiometric EA/O2/Ar mixture at a pressure of 10 atm with an argon/oxygen ratio, D, of 14.3 at 1200 K. The ignition delay time is 301 μs, the species EAMJ is CH3COOCH2CH2 and MA2J is CH3COOCHCH3, and EA3J in the model is CH2COOCH2CH3.

In this case, it is observed that abstraction by H atoms and O atoms dominates. In the case of the ethyl esters as exemplified by EA in Figure 23, the decomposition to ethylene and organic acid features as an important fuel depletion pathway both at reaction onset and just before ignition occurs, albeit with decreasing overall importance as more radicals are produced. A similar pattern is observed for EF ignition, though the ignition delay time in this case is 223 μs. By comparing methyl and ethyl esters, this study shows the general tendency of ethyl esters to ignite more readily than their methyl ester analogues. The trends observed in experiments can also be seen in modeling results. The model presented in this work shows an overall good performance. Further evaluation of its performance against other combustion experiments is needed. It should be noted that the performance of the small alkyl ester model of Westbrook 4352

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

ARTICLE

Table 1. Ethyl Propanoate and Ethyl Acetate Ignition Data Ester: EP

p [atm]

T [K]

τ [μs]

Ester: EA

p [atm]

T [K]

τ [μs]

ϕ = 1.0, D = 14.3

10.6

1143

571

ϕ = 1.0, D = 14.2

10.5

1143

1080

xEP = 1.0%

10.2

1177

373

xEA = 1.28%

10.4

1212

470

xO2 = 6.5%

10.0

1212

238

xO2 = 6.48%

10.1

1231

344

9.2

1231

169

10.2

1300

158

10.6

1301

82

9.6

1465

41

1.4

1242

558

1.1

1303

893

1.3

1304

368

1.0

1389

407

1.2 1.3

1431 1558

191 94

1.1 1.0

1421 1551

207 138

ϕ = 1.0, D = 3.8

1.0

1564

113

10.4

1060

522

ϕ = 1.0, D = 3.8

1.2

1572

118

10.7

1055

1154

xEP= 3.12%

10.0

1089

254

xEA= 4.04%

10.2

1107

657

xO2 = 20.3%

10.8

1120

129

xO2 = 20.14%

10.6

1128

471

10.0

1128

124

11.4

1134

258

10.3

1155

93

9.5

1225

89

10.1 4.2

1207 1157

43 660

ϕ = 0.5, D = 7.9

9.1 4.6

1235 1200

84 601 248

ϕ = 0.5, D = 7.9 xEP = 8.60%

4.3

1230

248

ϕ = 0.5, D = 7.9, xEA = 11.10%

4.0

1300

xO2 = 11.1%

3.8

1258

190

xO2 = 11.11%

3.8

1303

194

4.2

1307

114

4.0

1306

217

4.1

1335

78

4.0

1405

69

4.1

1384

43

3.7

1449

55

ϕ = 2.0, D = 24.0

11.0

1178

498

ϕ = 2.0, D = 24.0

10.3

1162

1134

xEP = 1.20% xO2 = 3.95%

10.3 9.7

1191 1251

414 172

xEA = 1.56% xO2 = 3.93%

10.8 10.2

1248 1298

445 260

10.0

1257

198

10.0

1300

359

9.1

1290

130

10.0

1388

91

9.8

1347

71

10.4

1432

65

Table 2. MA Submechanism (Units are kcal, cm3, mol, s, and K) reaction

A

n

MA=CH3COO+CH3 MA=OCOCH3+CH3 MA=CH2CO+CH3OH MA+H=MAMJ+H2 MA+OH=MAMJ+H2O MA+HO2=MAMJ+H2O2 MA+CH3=MAMJ+CH4 MA+CH3O=MAMJ+CH3OH MA+O=MAMJ+OH MA+O2=MAMJ+HO2

6.29  10 7.00  1015 1.08  1013 6.99  1014 1.40  103 4.76  104 9.03  101 4.34  1011 1.93  105 2.04  1013

5.89 0 0 0 2.66 2.55 3.65 0 2.68 0

89577.8 81695.0 74210 9950 527 16484.7 7149.5 6454.4 3714.5 41325.2

28

MA+H=MA2J+H2 MA+OH=MA2J+H2O MA+HO2=MA2J+H2O2 MA+CH3=MA2J+CH4 MA+CH3O=MA2J+CH3OH MA+O=MA2J+OH MA+O2=MA2J+HO2

4.04  1014 1.40  103 4.76  104 4.76  104 4.34  1011 1.93  105 2.04  1013

0 2.66 2.55 2.55 0 2.68 0

9320 527 16484.7 16484.7 6454.4 3714.5 41325.2

12

MAMJ=CH2O+CH3CO MA2J=CH2CO+CH3O

4.45  1014 1.19  1013

0.22 0.38

27231.7 36786.6

31

OCOCH3=CO2+CH3

9.22  1010

0.87

30337.0

33

35

Ea

ref 29

for CH3OCH3=CH3O+CH3 for CH3CHO=CH3+HCO

12 12 19 19 19 19 19 30

for (CH3)2CHCH(CH3)2+O2= HO2+(CH3)3CCH(CH3)

4353

19 19 19 19 19 30

for (CH3)2CHCH(CH3)2+O2=

HO2+(CH3)3CCH(CH3) for CH2OCH3=CH2O+CH3 32 for CH2OOCCH2CH2CH3= CH2O+CH3CH2CH2CO for nC3H7=CH3+C2H4

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

ARTICLE

Table 3. EF Submechanism (Units are kcal, cm3, mol, s, and K) reaction

A

n

Ea

ref

EF=HCOOCH2+CH3

7.00  1015

0

81695.0

EF=HCOO+C2H5

2.96  1022

2.16

96505.2

34 for C2H5OH=C2H5+OH

EF=HCOOH+C2H4

2.13  1011

0

44191.7

22

EF=CO+C2H5OH

8.62  1013

0

697800

EF+H=EFMJ+H2

6.99  1014

0

9950

12

EF+OH=EFMJ+H2O

1.40  103

2.66

527

19

EF+HO2=EFMJ+H2O2

4.76  104

2.55

16484.7

19

EF+CH3=EFMJ+CH4 EF+CH3O=EFMJ+CH3OH

9.03  101 4.34  1011

3.65 0

7149.5 6454.4

19 19

EF+O=EFMJ+OH

1.93  105

2.68

3714.5

19

EF+O2=EFMJ+HO2

2.04  1013

0

41325.2

30

EF+H=EF2J+H2

4.66  1014

0

6700

12

EF+OH=EF2J+H2O

2.70  104

2.39

393

19

EF+HO2=EF2J+H2O2

9.62  103

2.6

13902.5

19

EF+CH3=EF2J+CH4

1.59  101

3.46

5482.2

19

EF+CH3O=EF2J+CH3OH EF+O=EF2J+OH

1.45  1011 4.77  104

0 2.71

4567.3 2104.5

19 19

EF+O2=EF2J+HO2

3.01  1013

0

39175.4

EF+H=EF3J+H2

4.86  1013

0

9310

12

EF+OH=EF3J+H2O

1.40  103

2.66

527

19

EF+HO2=EF3J+H2O2

4.76  104

2.55

16484.7

19

EF+CH3=EF3J+CH4

9.03  101

3.65

7149.5

19

EF+CH3O=EF3J+CH3OH

4.34  1011

0

6454.4

19

EF+O=EF3J+OH EF+O2=EF3J+HO2

1.93  105 2.04  1013

2.68 0

3714.5 41325.2

19 30

EFMJ=HCOO+C2H4

1.20  1011

0

24031.9

35 for C2H5OC2H4=C2H5O+C2H4

EF2J=CH3CHO+HCO

5.95  1012

0.38

36786.6

32 for CH2OOCCH2CH2CH3=

EF2J=2HCO+CH3

1.19  1013

0.38

36786.6

32 for CH2OOCCH2CH2CH3

EF3J=CO+C2H5O

8.06  1011

0.65

21111.5

32 for CH3OCO=CO+CH3O

EF3J=CO2+C2H5 HCOO=HCO+O

8.06  1011 2.96  1022

0.65 2.16

21111.4 96505.2

32 for CH3OCO=CO+CH3O 34 for C2H5OH=C2H5+OH

HCOO=H+CO2

1.58  1016

0

97460.7

36 for C3H8=nC3H7+H

HCOOCH2=HCO+CH2O

1.19  1013

0.38

36786.6

32 for CH2OOCCH2CH2CH3

HCOOH=CO+H2O

2.45  1012

0

60377.8

37

HCOOH=CO2+H2

4.46  1013

0

68202.4

38

HCOOH+H=HCOO+H2

4.04  1014

0

9320

12

HCOOH+OH=HCOO+H2O HCOOH+CH3=HCOO+CH4

1.40  103 9.03  101

2.66 3.65

527 7149.5

19 19

HCOOH+O2=HCOO+HO2

3.97  1013

0

50880.2

19

C2H5OH=CH3+CH2OH

2.00  1016

0

85000

39

C2H5OH=C2H4+H2O

2.11  108

1.36

65690

34

29 for CH3CHO=CH3+HCO

12 for MF=CO+CH3OH

29 for CH3CHO+O2=CH3CO+HO2

CH2O+CH3CH2CH2CO, A reduced by 1/2 =CH2O+CH3CH2CH2CO

=CH2O+CH3CH2CH2CO

et al.18 has been previously evaluated by the current authors for MF,12,14 MA,10,12 and EF.10 Good agreement in all cases is observed at high pressures of about 1013 atm while some discrepancies were also observed at lower pressures and less dilute mixtures. The MA, EF, and EA ignition predictions of the current model generally agree with those of Westbrook et al.18 at pressures of 1013 atm for argon/oxygen ratios, D, greater than 10. Longer ignition delay times are predicted by the model of Westbrook et al.,18 than observed for MA and

EA for mixtures with D of 3.76 at 10 atm, whereas under these conditions, predictions of both models are similar for EF and in agreement with experimental data. For all three fuels, discrepancies are observed at pressures of 4 atm and lower, with the model of Westbrook et al.18 predicting shorter ignition delay times than measured or predicted by the current model. The EA data presented in this work could also be useful for the further improvement of the model of Westbrook et al. 18 4354

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels

ARTICLE

Table 4. EA Submechanism (Units are kcal, cm3, mol, s, and K) reaction

A

n

Ea

ref

EA=CH3+EF3J

7.00  1015

0

81694.97

29 for CH3CHO=CH3+HCO

EA=MAMJ+CH3

7.89  1022

1.79

88622.32

40

EA=OCOCH3+C2H5

2.96  1022

2.16

96505.17

34 for C2H5OH=C2H5+OH

EA=CH3COOH+C2H4

1.66  109

1

46580.47

21

EA=CH2CO+C2H5OH

1.08  1013

0

74210

12 from MA

EA+HEAMJ+H2

6.99  1014

0

9950

12

EA+OH=EAMJ+H2O

1.40  103

2.66

527

19

EA+HO2=EAMJ+H2O2 EA+CH3=EAMJ+CH4

4.76  104 9.03  101

2.55 3.65

16484.71 7149.50

19 19

EA+CH3O=EAMJ+CH3OH

4.34  1011

0

6454.38

19

EA+O=EAMJ+OH

1.93  105

2.68

3714.49

19

EA+O2=EAMJ+HO2

3.97  1013

0

50880.20

19

EA+HdEA2J+H2

4.66  1014

0

6700

12

EA+OH==EA2J+H2O

2.70  104

2.39

393

19

EA+HO2=EA2J+H2O2

9.62  103

2.6

13902.48

19

EA+CH3=EA2J+CH4 EA+CH3O=EA2J+CH3OH

1.59  101 1.45  1011

3.46 0

5482.16 4567.27

19 19

EA+O=EA2J+OH

4.77  104

2.71

EA+O2=EA2J+HO2

3.01  1013

0

EA+H=EA3J+H2

4.04  1014

0

EA+OH=EA3J+H2O

1.40  103

EA+HO2=EA3J+H2O2

4.76  104

EA+CH3=EA3J+CH4

2104.48 39175.37

19 29

for CH3CHO+O2=CH3CO+HO2

9320

12

2.66

527

19

2.55

16484.71

19

9.03  101

3.65

7149.50

19

EA+CH3O=EA3J+CH3OH EA+O=EA3J+OH

4.34  1011 1.93  105

0 2.68

6454.38 3714.49

19 19

EA+O2=EA3J+HO2

3.97  1013

0

50880.20

19

EAMJ=OCOCH3+C2H4

2.94  1013

0.27

34636.76

32

EA2J=CH3CHO+CH3CO

1.19  1013

0.38

36786.62

32 for CH2OOCCH2CH2CH3=CH2O+CH3CH2CH2CO

EA2J=CH3CO+HCO+CH3

1.19  1013

0.38

36786.62

32 for CH2OOCCH2CH2CH3=CH2O+CH3CH2CH2CO

EA3J=CH2CO+C2H5O

1.27  1012

0.66

49208.08

CH3COOH+H=CH2COOH+H2

4.04  1014

0

9320

12

CH3COOH+OH=CH2COOH+H2O CH3COOH+CH3=CH2COOH+CH4

1.40  103 9.03  101

2.66 3.65

527 7149.50

19 19

CH3COOH+O2=CH2COOH+HO2

3.97  1013

0

50880.20

19

CH3COOH=CH2CO+H2O

2.82  1012

0

64945.83

41

OCOCH3=CH2COOH

9.95  1011

6.55

21739.94

32

CH2COOH=CH2CO+OH

2.96  1022

2.16

96505.17

34 for C2H5OH=C2H5+OH

CH2COOH=HCCO+H2O

1.27  1012

0.66

49208.08

32 for EA3J=CH2CO+C2H5O

’ CONCLUSION The relative ignition behavior of a series of short-chain methyl and ethyl esters is presented based on new shock tube ignition data and published ignition delay times by the current authors. Comparison is based on the shock ignition of homogeneous gasphase mixtures of esters, oxygen, and argon under the constraints of constant equivalence ratio, argon/oxygen ratio, and nominal pressure, over a range of postreflected shock temperatures. It is observed that ethyl esters generally have shorter ignition delay times than the methyl esters of the same carboxylic acid. This increased reactivity of the ethyl esters is thought to be linked to the presence of the ethyl group which provides a weaker CH site for H abstraction reactions as well as further reactivity enhancement through the more favorable elimination of ethene by concerted unimolecular reaction to yield the corresponding carboxylic acid.

32

In the case of methyl and ethyl ester isomers, ethyl esters also have shorter ignition delay times for similar reasons. A chemical kinetic model is proposed for MA, EA, and EF. The proposed model generally performs well in capturing the reactivity trends, and gives a reasonable quantitative prediction of the measured ignition delay times. Some discrepancies are observed especially at lower pressures and rich MA mixtures. This work contributes toward further understanding of the implications of the trans-esterification alcohol on the combustion chemistry of the resulting biodiesel.

’ APPENDIX Representative Ignition Delay Times of Ethyl Acetate and Ethyl Propanoate. Ignition data for ethyl propanoate and ethyl

acetate are provided in Table 1. 4355

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356

Energy & Fuels Ester Submechanisms. Submechanism data for MA, EF, and EA are provided in Tables 2, 3, and 4, respectively.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected].

’ ACKNOWLEDGMENT We acknowledge support from the Natural Sciences and Engineering Research Council (NSERC) of Canada and the “Fonds Quebecois de la Recherche sur la Nature et les Technologies” (FQRNT). ’ REFERENCES (1) Demirbas, A. Energy Convers. Manage. 2003, 44, 2093–2109. (2) Baiju, B.; Naik, M.; Das, L. Renewable Energy 2009, 34, 1616– 1621. (3) Peterson, C.; Reece, D.; Thompson, J.; Beck, S.; Chase, C. Biomass Bioenergy 1996, 10, 331–336. (4) Joshi, H.; Moser, B.; Toler, J.; Walker, T. Biomass Bioenergy 2010, 34, 14–20. (5) Lai, J. Y.; Lin, K. C.; Violi, A. Prog. Energy Combust. Sci. 2011, 37, 1–14. (6) Kohse-H€oinghaus, K.; Oßwald, P.; Cool, T. A.; Kasper, T.; Hansen, N.; Qi, F.; Westbrook, C. K.; Westmoreland, P. R. Angew. Chem., Int. Ed. 2010, 49, 3572–3597. (7) Metcalfe, W.; Dooley, S.; Curran, H.; Simmie, J.; El-Nahas, A.; Navarro, M. J. Phys. Chem. A 2007, 111, 4001–4014. (8) Hakka, M. H.; Bennadji, H.; Biet, J.; Yahyaoui, M.; Sirjean, B.; Warth, V.; Coniglio, L.; Herbinet, O.; Glaude, P. A.; Billaud, F.; BattinLeClerc, F. Int. J. Chem. Kinet. 2010, 42, 226–252. (9) Bennadji, H.; Coniglio, L.; Billaud, F.; Bounaceur, R.; Warth, V.; Glaude, P.-A.; Battin-Leclerc, F. Int. J. Chem. Kinet. 2011, 43, 204–218. (10) Akih-Kumgeh, B.; Bergthorson, J. M. Combust. Flame 2011, 158, 1877–1889. (11) Zhang, Y.; Yang, Y.; Boehman, A. L. Combust. Flame 2009, 156, 1202–1213. (12) Akih-Kumgeh, B.; Bergthorson, J. M. Combust. Flame 2011, 158, 1037–1048. (13) Wang, H.; You, X.; Joshi, A. V.; Davis, S. G.; Laskin, A.; Egolfopoulos, F.; Law, C. K. USC Mech Version II. High-Temperature Combustion Reaction Model of H2/CO/C1-C4 Compounds. http://ignis. usc.edu/USC_Mech_II.htm, (accessed January 4, 2011). (14) Akih-Kumgeh, B.; Bergthorson, J. M. Energy Fuels 2010, 24, 396–403. (15) Morley, C. A Chemical Equilibrium Program for Windows. http:// www.arcl02.dsl.pipex.com (accessed June 22, 2009). (16) Akih-Kumgeh, B.; Bergthorson, J. M. Energy Fuels 2010, 24, 2439–2448. (17) Heufer, K.; Fernandes, R.; Olivier, H.; Beeckmann, J.; R€ohl, O.; Peters, N. Proc. Combust. Inst. 2011, 33, 359–366. (18) Westbrook, C.; Pitz, W.; Westmoreland, P.; Dryer, F.; Chaos, M.; Oßwald, P.; Kohse-H€oinghaus, K.; Cool, T.; Wang, J.; Yang, B.; Hansen, N.; Kasper, T. Proc. Combust. Inst. 2009, 32, 221–228. (19) Tsang, W. J. Phys. Chem. Ref. Data 1988, 17, 887–951. (20) El-Nahas, A.; Navarro, M.; Simmie, J.; Bozzelli, J.; Curran, H.; Dooley, S.; Metcalfe, W. J. Phys. Chem. A 2007, 111, 3727–3739. (21) Scheer, J.; Kooyman, E.; Sixma, F. Recueil 1963, 82, 1123–1154. (22) Blades, A. Can. J. Chem. 1954, 32, 366–372. (23) Wang, Y. L.; Feng, Q.; Egolfopoulos, F. N.; Tsotsis, T. T. Combust. Flame 2011, 158, 1507–1519. (24) Francisco, J. J. Am. Chem. Soc. 2003, 125, 10475–10480. (25) Dooley, S.; Burke, M. P.; Chaos, M.; Stein, Y.; Dryer, F. L.; Zhukov, V. P.; Finch, O.; Simmie, J. M.; Curran, H. J. Int. J. Chem. Kinet. 2010, 42, 527–549.

ARTICLE

(26) Ritter, E.; Bozzelli, J. Int. J. Chem. Kinet. 2004, 23, 767–778. (27) Goodwin, D. Proc. CVD XVI and EuroCVD Fourteen, Electrochem. Soc. 2003, 14, 155–162. (28) Zhao, Z.; Chaos, M.; Kazakov, A.; Dryer, F. Int. J. Chem. Kinet. 2008, 40, 1–18. (29) Baulch, D.; Cobos, C.; Cox, R.; Esser, C.; Frank, P.; Just, T.; Kerr, J.; Pilling, M.; Troe, J.; Walker, R.; Warnatz, J. J. Phys. Chem. Ref. Data 1992, 21, 411–429. (30) Baldwin, R.; Drewery, G.; Walker, R. J. Chem. Soc. Faraday Trans. 1 1984, 80, 3195–3207. (31) Li, Q.; Zhang, Y.; Zhang, S. J. Phys. Chem. A 2004, 108, 2014–2019. (32) Huynh, L.; Violi, A. J. Org. Chem. 2008, 73, 94–101. (33) Curran, H. Int. J. Chem. Kinet. 2006, 38, 250–275. (34) Tsang, W. Int. J. Chem. Kinet. 2004, 36, 456–465. (35) Foucaut, J.-F.; Martin, R. J. Chim. Phys. 1978, 132. (36) Dean, A. J. Phys. Chem. 1985, 89, 4600–4608. (37) Blake, P.; Davies, H.; Jackson, G. J. Chem. Soc. B 1971, 1923– 1925. (38) Chang, J.; Chen, H.; Xu, S.; Lin, M. J. Phys. Chem. A 2007, 111, 6789–6797. (39) Frassoldati, A.; Cuoci, A.; Faravelli, T.; Niemann, U.; Ranzi, E.; Seiser, R.; Seshadri, K. Combust. Flame 2010, 157, 2–16. (40) Tsang, W. Combust. Flame 1989, 78, 71–86. (41) Blake, P.; Jackson, G. J. Chem. Soc. B 1969, 94–96.

4356

dx.doi.org/10.1021/ef200977p |Energy Fuels 2011, 25, 4345–4356