Fuel 140 (2015) 626–632

Contents lists available at ScienceDirect

Fuel

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

Experimental and kinetic study on ignition delay times of dimethylcarbonate at high temperature

http://dx.doi.org/10.1016/j.fuel.2014.10.0130016-2361/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding authors. Tel.: +86 29 82665075; fax: +86 29 82668789.E-mail addresses: hujiang@mail.xjtu.edu.cn (E. Hu), zhhuang@mail.xjtu.edu.cn

(Z. Huang).

Erjiang Hu ⇑, Yizhen Chen, Zihang Zhang, Lun Pan, Qianqian Li, Yu Cheng, Zuohua Huang ⇑State Key Laboratory of Multiphase Flow in Power Engineering, Xi’an Jiaotong University, Xi’an 710049, People’s Republic of China

h i g h l i g h t s

� Ignition delay times of dimethyl carbonate were measured in a shock tube for the first time at different conditions.� A modified DMC kinetic model was proposed and can well predict ignition delay times and activation energies.� DMC is primarily consumed through the H-abstractions and not fuel unimolecular decompositions for high temperature ignition.

a r t i c l e i n f o

Article history:Received 17 July 2014Received in revised form 30 September2014Accepted 3 October 2014Available online 18 October 2014

Keywords:Dimethyl carbonateShock tubeIgnition delay timeChemical kinetic model

a b s t r a c t

Ignition delay times of dimethyl carbonate were measured in a shock tube for the first time at T = 1100–1600 K, p = 0.12–1.0 MPa, fuel concentration = 0.5–2.0%, and / = 0.5–2.0. A modified chemical kineticmodel was developed and can well predict ignition delay times and activation energies. Furthervalidation of the proposed kinetic model was made on the basis of the opposed flow diffusion flame data.Reasonable agreements were also achieved under the literature data. Reaction pathway analysis showsthat at high temperature the DMC molecule is primarily consumed through the H-abstractions but notfuel unimolecular decompositions. Sensitivity analysis provides some key fuel-species reactions forDMC high temperature ignition.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

It has been proven that oxygenated fuels are effective in reduc-ing soot emissions from diesel engines. Researches concerningengine performance and emissions have been conducted widelyon diesel engines fueled with various oxygenates, such as alcohols,ethers, esters, carbonates and acetates [1–4]. As a promising oxy-genated fuel, Dimethyl carbonate (DMC) possesses a very high per-centage of oxygen. Engine studies indicated that the DMC additionto the diesel could efficaciously decrease the smoke emissions[5–7]. Previous studies revealed that the addition of ethers wasmore efficient in reducing the smoke emissions compared to theaddition of alcohols [8], and that dimethyl carbonate was evenmore efficient than other ethers [7]. However, there are numeroussimultaneous processes in the engine which make it difficult todetermine the mechanism for such decrease.

Little work has been published regarding fundamental combus-tion researches of DMC in recent years. Sinha and Thomson [9]measured the species and temperatures in an opposed flow diffu-sion flames of iso-propanol, dimethoxy methane (DMM), anddimethyl carbonate (DMC). For DMM and DMC, the absence ofC–C bonds effectively decreases the formation of soot precursorssuch as ethylene, acetylene, and propylene. Compared to the pro-pane flame, the ethylene levels of iso-propanol, DMM, and DMCflames were reduced by 41%, 77%, and 93%, respectively. However,their work did not cover the development of chemical kineticmodel. Sinha and Thomson [9] studied the experimental speciesprofiles by experimenting in an opposed flow diffusion flame andit gave the exclusive validation of the only available DMC kineticmodel [10], which is developed by Glaude and can well simulatethe DMM and DMC diffusion flame results. This model was alsocompared with species profiles obtained in premixed low-pressure(30 Torr) flames of heptane with DMC addition, and an overallsatisfactory agreement was attained by Chen et al. [11]. Badinmeasured the laminar flame speed using the heat flux method,but Glaude DMC model significantly overpredicted the experimen-tal values [12].

E. Hu et al. / Fuel 140 (2015) 626–632 627

In fact, many reaction rate constants in this DMC model weredetermined by estimation. Further development and validation ofthe DMC kinetic model are lagging due to the lack of more accuratefundamental experimental data. For instance, there are notreported ignition delay time data yet.

Therefore, the objectives of the present work include the mea-surement the ignition delay time of DMC under different condi-tions, the validation of the DMC chemical kinetic model, and theinterpretation of the ignition chemistry through reaction pathwayand sensitivity analysis.

2. Experimental and calculated approach

Ignition delay times were measured in a shock tube and thedetailed description of the experimental apparatus can be referredto the previous publications [13,14]. The time interval between thearrival of incident shock wave at the endwall and the intercept ofthe maximum slope of the CH⁄ trajectory with the zero line isdefined as the measured ignition delay time (s) in the currentstudy. Ignition temperatures (T) are calculated with Gaseq [15].The tested ranges of the equivalence ratio and the pressure ofDMC/oxygen/argon mixtures were 0.5–2.0 and 0.12–1.0 MPa,respectively. The respective purities of DMC, oxygen and argonare 99.9%, 99.999% and 99.999%. Table 1 shows the compositionsof test mixtures in detail.

Calculations were carried out using Chemkin [16] and Senkin[17] codes and applying the constant volume adiabatic model.The maximum increase rate of temperature profile (max dT/dt) isused to define the calculated ignition delay time in this study.

3. Chemical kinetic model

The unique DMC chemical model available at present wasdeveloped by Glaude in 2005 [10]. It consists of the C1–C3 modeldeveloped by Curran and the DMC sub-model developed by Gla-ude. Although this model was validated with DMC opposed-flowdiffusion flames [10], it cannot give a satisfactory prediction tothe experimentally measured ignition delay times of this study.

The modified chemical kinetic model in this study wasenhanced by adding the DMC sub-model to Aramco Mech 1.3model [18] and it consists of 275 species and 1586 reactions as awhole. DMC sub-model consists of 24 elementary reactionsabsorbing the Glaude DMC sub-model [10] and Dooley MB sub-model [19]. Detailed mechanism including thermochemical datais given in Supporting Information. It can be observed that themodification was mainly conducted on the DMC sub-model, whilethe original C4 chemistry remained the same. Table 2 [10,19–23]provides further details of the modified DMC sub-model.

3.1. Unimolecular decomposition

Fig. 1 gives the bond dissociation energies of dimethyl carbon-ate, methyl butanoate and dimethyl ether [24]. The bond energiesof CH3OC⁄OO-CH3 (374.5 kJ/mol) and CH3OC⁄O-OCH3 (427.2 kJ/mol) in DMC are different and they are also unequal to that of

Table 1Composition of DMC–O2–Ar mixtures.

Mixtures / XDMC (%) XO2 (%) XAr (%) p (MPa)

6 0.5 1.0 6.0 93.0 0.12, 0.5, 1.07 1.0 1.0 3.0 96.0 0.12, 0.5, 1.08 2.0 1.0 1.5 97.5 0.12, 0.5, 1.09 1.0 0.5 1.5 98.0 0.5

10 1.0 1.5 4.5 94.0 0.511 1.0 2.0 6.0 92.0 0.5

CH3O-CH3 (351.5 kJ/mol). Therefore, in Glaude DMC model [10],it is improper to adopt the rate constant of CH3O-CH3 decomposi-tion reaction as that of R1565 and R1566.

Bond energies of CH3OC⁄OO-CH3 (374.5 kJ/mol) and CH3OC⁄O-OCH3 (427.2 kJ/mol) in DMC are almost identical to those ofC3H7-C⁄OO-CH3 (364.1 kJ/mol) and C3H7C⁄O-OCH3 (423.9 kJ/mol)in methyl butanoate. On the basis of analogical method, thereaction rates of these two bond fission should have the samevalues at the same temperature. As a result, the rate constants ofCH3OC⁄OO-CH3 (R1565) and CH3OC⁄O-OCH3 (R1566) decomposi-tion, recommended by Dooley et al. [19] for the decompositionsof methyl butanoate, were adopted in the present model.

It is reported by Glaude [10] that R1567 is a new molecularelimination path for the DMC and the activation energy of thiselementary reaction was determined by CBS-Q plus correctionswith isodesmic reactions. The present model employed the rateconstant of R1567 of Glaude’s calculation. R1568 is another DMCunimolecular decomposition reaction, and its rate constant wasdetermined by Glaude’s estimation [10,20].

3.2. Hydrogen abstraction

It was proved that H-abstraction is the main pathway for the con-sumption of typical fuels at high temperature ignition [14,25,26]. InGlaude DMC model, the rate constants of H-abstraction by radicals(R1573, R1575 and R1577) are assumed to that of the secondaryC–H of iso-octane. Analogism can be made between the otherH-abstraction reactions and the primary and secondary C–H ofn-heptane. Fig. 2 gives the bond dissociation energies in GlaudeDMC model of n-heptane and iso-octane at 298 K [24]. The bonddissociation energy of C–H in DMC molecule (422.6 kJ/mol) isdifferent to that of n-heptane (367.8 kJ/mol and 409.6 kJ/mol)and iso-octane (399.6 kJ/mol), but it is close to that of primaryC–H (414.2 kJ/mol) in methyl butanoate. Therefore, the presentmodel assumed the rate constants of H-abstraction reactions ofDMC to be identical to that of methyl butanoate.

3.3. Ether-acid conversion

R1579 is another consumption pathway of DMC moleculeproducing the COC⁄OOH. Since this reaction has small branchingratio and consequently has little effect on the ignition delay time,the rate constant of this elementary reaction was also adoptedby Glaude’s estimation [10].

3.4. Hydrogen abstraction of COC⁄OOH

COC⁄OOH is mainly consumed by hydrogen abstraction withsmall radicals (R1580–R1583). These reactions are also taken fromthe Glaude’s estimation [10].

3.5. Radical decomposition

CJOC⁄OOH radical is the product of hydrogen abstraction ofCOC⁄OOH, and it decomposes to CH2O, CO and OH throughR1584. COC⁄OOJ radical, the product of R1565, produces CH3Oand CO2. COC⁄COOCJ radical is the main product of DMC moleculeconsumption, decomposing CH3OCO and CO2. These three rateconstants (R1584–R1586) were assumed by Glaude [10].

Decomposition of the CH3OCO radical has been the subject ofrecent attention [10,27]. In this model, we employed Glaude’sexpression of CH3OCO decomposition calculated by CBS-Q method,which has been validated with speciation data of an opposed flowdiffusion flame in his modeling study. The model prediction indi-cates the importance of the decomposition of CH3OCO, becauseits productions – CO2 and CH3 radicals – dominate free radical

Table 2DMC sub-model.

Num Reaction Rate constant References

Unimolecular decompositionR1565 COC⁄OOC(+M), COC⁄OOJ + CH3(+M) 2.552E+23–1.99 8.810E+04 [19]

LOW/1.744E+73–1.596E+01 8.532E+04/TROE/2.18E-01 1.00E+00 6.37E+03 8.21E+09/

R1566 COC⁄OOC(+M), CH3OCO + CH3O(+M) 7.426E+21–1.38 9.890E+04 [19]LOW/1.807E+58–1.170E+01 9.127E+04/TROE/2.51E-01 3.72E+02 9.48E+09 5.00E+09 /

R1567 COC⁄OOC, CH3OCH3 + CO2 5.0E+11 0.19 69800.0 [10]R1568 COC⁄OOCJ + H, COC⁄OOC 5.0E+13 0.0 0.0 [10,20]

Hydrogen abstraction of DMCR1569 COC⁄OOC + C2H3, C2H4 + COC⁄OOCJ 5.010E+11 0.00 1.800E+04 [19]R1570 COC⁄OOC + C2H5, C2H6 + COC⁄OOCJ 5.010E+10 0.00 1.340E+04 [19]R1571 COC⁄OOC + CH3, CH4 + COC⁄OOCJ 2.265E+00 3.46 5.481E+03 [19]R1572 COC⁄OOC + CH3O, CH3OH + COC⁄OOCJ 2.175E+11 0.00 4.571E+03 [19]R1573 COC⁄OOC + CH3O2, CH3O2H + COC⁄OOCJ 1.229E+04 2.60 1.391E+04 [19]R1574 COC⁄OOC + H, H2 + COC⁄OOCJ 0.975E+06 2.40 4.471E+03 [19]R1575 COC⁄OOC + HO2, H2O2 + COC⁄OOCJ 1.229E+04 2.60 1.391E+04 [19]R1576 COC⁄OOC + O, OH + COC⁄OOCJ 8.280E+05 2.45 2.830E+03 [19]R1577 COC⁄OOC + O2, HO2 + COC⁄OOCJ 3.000E+13 0.00 4.964E+04 [19]R1578 COC⁄OOC + OH, H2O + COC⁄OOCJ 7.020E+07 1.61–3.500E+01 [19]

Ether-acid conversionR1579 COC⁄OOC + H) COC⁄OOH + CH3 3.79E+16–1.39 5402.0 [10,21]

Hydrogen abstraction of COC⁄OOHR1580 COC⁄OOH + OH, CJOC⁄OOH + H2O 5.25E+09 0.97 1590.0 [10,22]R1581 COC⁄OOH + H, CJOC⁄OOH + H2 9.40E+04 2.75 6280.0 [10,22]R1582 COC⁄OOH + CH3, CJOC⁄OOH + CH4 4.52E-01 3.65 7154.0 [10,22]R1583 COC⁄OOH + O, CJOC⁄OOH + OH 9.65E+04 2.68 3716.0 [10,22]

Radical decompositionR1584 CJOC⁄OOH) CH2O + CO + OH 6.10E+21–2.40 3.252E4 [10,22]R1585 CH3O + CO2, COC⁄OOJ 1.000E+11 0.00 9.20E+03 [10,23]R1586 CH3OCO + CH2O, COC⁄OOCJ 1.06E+11 0.0 7350.0 [10,20]R488 CH3 + CO2, CH3OCO 4.760E+007 1.540 34700.0 [10]R489 CH3O + CO, CH3OCO 1.550E+006 2.016 5730.0 [10]

C

H

H

H 427.2 O C

H

H

H422.6C

O

O 374.5

C

H

H

C 423.9 O C

H

H

H414.2C

O

C 364.1

H

H

H

H

H

C

H

H

H O C

H

H

H401.7351.5

Fig. 1. Bond dissociation energies (BDE298, kJ/mol) of dimethyl carbonate, methylbutanoate and dimethyl ether [24].

C

H

H

H C

CH3

CH3

399.6

C

H

H

C

H

CH3

C

H

H

H

C

H

H

H C

H

H

367.8

C

H

H

C

H

H

C

H

H

C

409.6

C

H

H

H

H

H

Fig. 2. Bond dissociation energies (BDE298, kJ/mol) of n-heptane and i-octane [24].

628 E. Hu et al. / Fuel 140 (2015) 626–632

chain branching and termination in most of sensitive reactions inthe system. Moreover, the rate constants of R488 and R489 in Ara-mco Mech 1.3 model developed by Metcalfe et al. [18], the methylbutanoate model developed by Dooley et al. [19] and the methyl

esters model developed by Pascal [28], were also taken from theGlaude’s calculation.

4. Results and discussion

4.1. Ignition delay time measurement

Fig. 3 gives the measured DMC ignition data. The logarithmicignition delay times under all conditions exhibit good lineardependence upon 1000/T. Therefore, through regression, a fittingcorrelation of DMC ignition delay times as a function of p, XDMC,/ and T for the DMC–O2–Ar mixtures is correlated as follows:

Fig. 3. Measured and fitted ignition delay times of DMC. (Symbols: measured values; lines: fitted values using Eq. (1)).

E. Hu et al. / Fuel 140 (2015) 626–632 629

s ¼ 1:11� 10�4p�0:38X�0:71DMC /0:65 exp

162:59� 1:68 kJ mol�1

RT

!;

R2 ¼ 0:987 ð1Þ

where s, p and / are ignition delay time in ls, pressure in MPa andequivalence ratio, respectively. R is the universal gas constant andits value is 8.314 J/mol K, and T is the temperature in K. R2 = 0.987indicate that Eq. (1) has good regressions to the ignition data. Thiscorrelation is only applicable to the temperature range of 1100–1550 K, pressure range of 0.12–1.0 MPa, equivalence ratio range of0.5–2.0 and fuel mole fraction range of 0.5–2.0% because of the lim-ited range of the study.

4.2. Validation of the kinetic model

4.2.1. Ignition delay timeIn this section, we used the Glaude DMC model and modified

model to calculate the ignition delay times in order to compareand validate their respective performances under different fuelmole fractions, pressures, and equivalence ratios.

Fig. 4 shows the measured ignition delay times and the simu-lated ones for DMC with modified model and Glaude model. Thekinetic models were validated at different fuel mole fractions. Asa matter of fact, the fuel mole fraction in our study has an equiva-lent effect as the dilution ratio used by other researchers [29].Fig. 4a gives the measured and calculated ignition delay times atdifferent DMC mole fractions. The ignition delay times decreasewith the increase of fuel fraction. The reason is that high fuel andoxygen concentrations enhance the overall reaction rate. Accord-ing to the figure, both Glaude model and modified model are capa-ble of capturing the trend of the curves and predicting the

activation energy. However, the overprediction of Glaude modelin ignition delay times of DMC makes it incompetent for quantita-tive simulation, while the modified model can well predict theignition delay times and activation energies of DMC across thewhole temperature range.

As shown in Fig. 4b, the ignition delay times are dependent onpressure as well besides the fuel mole fraction. Ignition delay timesdecrease with the increase of pressure. The reactant concentrationsincrease significantly at high pressures, which promotes the igni-tion process. The negative pressure exponent in the ignition delaycorrelation shown in Eq. (1) also suggests this phenomenon.Similarly, the modified model can well predict the pressure depen-dence and gives good quantitative agreement with the experimen-tal data at different pressures, while the Glaude model again givesover-predictions on the ignition delay times of DMC.

In order to extend the application of the modified model, theeffect of equivalence ratio on ignition delay times at low and highpressures is given in Fig. 4c and d. The same trend is presented at0.12 MPa and 1.0 MPa, showing that the ignition delay timesdecrease with decreasing equivalence ratio. The reason for this isthat, at high temperatures, the chain branching reaction (H + O2 =OH + O) plays a dominant role in the ignition process. Higher oxy-gen concentration (lower equivalence ratio) favors to the increaseof chain branching, resulting increased reactivity and decreasedignition delay times. Compared to the Glaude DMC model, themodified model can capture the dependence of equivalence ratioand gives good quantitative prediction to the experimental values.

4.2.2. Opposed-flow diffusion flame dataSince Glaude model was developed using mole fractions in an

opposed flow diffusion flame of DMC, a comparison between

Fig. 4. Measured and calculated ignition delay times with modified model and Glaude model for DMC. (Symbols: measured values; lines: calculated values.)

630 E. Hu et al. / Fuel 140 (2015) 626–632

calculated and measured diffusion flame data of DMC was made tofurther validate the present kinetic model. The DMC diffusionflame experiments were conducted at p = 0.1 MPa [9]. The oxidizerstream containing 39% O2 and 61% N2 was sent through the topburner port; 8% fuel (DMC) and 92% N2 was sent through the bot-tom. Fuel flow line and burner port were heated to 45 �C to avoidany condensation. Validations against experimental diffusionflame data for the Glaude model and present model are given inFig. 5. The present model almost gives the same calculated valueson DMC diffusion flame data as the Glaude model, which indicatesthat the present model can well predict not only the ignition delaytimes but also the diffusion flame data of DMC.

Fig. 5. Measured and calculated mole fraction profiles of major species and temperaturcalculated values.)

4.3. Reaction pathway and sensitivity analysis

4.3.1. Reaction pathway analysisReaction pathway analysis is an effective method in determin-

ing main reaction pathways for the concerned species. Fig. 6 showsthe reaction pathways using the present model at T = 1350 K,p = 0.12 MPa and / = 1.0 for DMC ignition. We chose the timingof 20% fuel consumption for the analysis like other works[30,31]. It can be observed that DMC molecule is primarily con-sumed through the H-abstractions (their branching ratios are79.2%), while the unimolecular decomposition gives smallcontribution (20.8%). Obviously, the unimolecular decomposition

e in the DMC opposed-flow diffusion flame. (Symbols: measured values [9]; lines:

COC*OOC

COC*OOHCOC*OOJ

R1565 R1579

R1585 R1580-1583

20.8% 1.7%

100%100%

CH3O CJOC*OOHR1584 100%

CH2O

R1571R1574R1576R1578

77.5%

COC*OOCJR1586100%

CH3OCOR488

59.4%R489

40.6%

CH3OCH3

Fig. 6. Reaction pathway of present model for DMC in shock tube at p = 0.12 MPa, /= 1.0, T = 1300 K, 20% fuel consumption.

E. Hu et al. / Fuel 140 (2015) 626–632 631

reaction R1565 dominates the fuel breakdown, while the branch-ing ratio of R1566 is negligible. This is because the bond energiesof CH3OC⁄OO-CH3 (374.5 kJ/mol) is much smaller than that of CH3-

OC⁄O-OCH3 (427.2 kJ/mol) as shown in Fig. 1.

4.3.2. Sensitivity analysisTo identify the above variation and clarify the elementary reac-

tions that dominate the ignition chemistry, sensitivity analysis

(Sensitivity coefficient, S ¼ sð2:0kiÞ�sð0:5kiÞ1:5sðkiÞ

, here s is ignition delay

time and ki is pre-exponential factor of ith reaction). was con-ducted in this study. Positive sensitivity coefficients indicate aninhibiting influence on the ignition process, and vice versa. Fig. 7gives the key sensitive reactions for the DMC at T = 1350 K,p = 0.12 MPa and / = 1.0.

For DMC as shown in Fig. 7, chain branching reaction (R1) domi-nants the ignition chemistry. Small radical reactions have aremarkable influence on ignition delay times, and only two ofthe twelve most important reactions are fuel-related reactions.Fuel-related reaction R1574, as an H-abstraction reaction by H rad-ical, has the largest positive value of S at high temperature. Thissuggests that the ignition process can be accelerated by theincrease of the reaction rate. The fuel-related reaction R1565, asan important unimolecular decomposition reaction, gives the sec-ond largest negative sensitivity coefficient because of its low bonddissociation energy. Moreover, R488 and R489 are also found to beimportant for the DMC ignition.

Fig. 7. Sensitivity analysis for DMC in shock tube at T = 1350 K, p = 0.12 MPaand / = 1.0.

5. Conclusions

Experimental and kinetics study on ignition delay times of DMCwere conducted in a shock tube at different equivalence ratios(0.5–2.0), pressures (0.12–1.0 MPa) and fuel mole fractions (0.5–2.0%) at high temperatures. Main conclusions are summarized asfollows:

(1) Ignition delay times of DMC were measured for the first timein a shock tube. Similar to typical hydrocarbon fuels, ignitiondelay times are decreased with the increase of pressure, thedecrease of equivalence ratio and the increase of fuel con-centration. Correlation of ignition delay time is given bymultiple linear regression on experimental data.

(2) A modified chemical kinetic model for DMC is proposed. Themodel yields fairly good agreement with the measured igni-tion delay times. The present model was validated againstthe diffusion flame data.

(3) Reaction pathway analysis shows that DMC molecule is pri-marily consumed through the H-abstractions. Sensitivityanalysis reveals the importance of small radical reactions,especially the main chain branching reaction H + O2 =O + OH at the tested temperature range. Some fuel-specificreactions are also found to have relatively large sensitivitycoefficients.

Acknowledgments

This study is supported by the National Natural Science Founda-tion of China (51306144), the National Basic Research Program(2013CB228406) and the State Key Laboratory of Engines at TianjinUniversity (SKLE201302). The support from the FundamentalResearch Funds for the Central Universities is also appreciated.

Appendix A. Supplementary material

Supplementary data associated with this article can be found, inthe online version, at http://dx.doi.org/10.1016/j.fuel.2014.10.013.

References

[1] Ren Y, Huang ZH, Miao HY, Di YG, Jiang DM, Zeng K, et al. Combustion andemissions of a DI diesel engine fuelled with diesel-oxygenate blends. Fuel2008;87(12):2691–7.

[2] Wang XG, Cheung CS, Di YG, Huang ZH. Diesel engine gaseous and particleemissions fueled with diesel-oxygenate blends. Fuel 2012;94:317–23.

[3] Wen LB, Xin CY, Yang SC. The effect of adding dimethyl carbonate (DMC) andethanol to unleaded gasoline on exhaust emission. Appl Energy2010;87(1):115–21.

[4] Zhu RJ, Miao HY, Wang XB, Huang ZH. Effects of fuel constituents and injectiontiming on combustion and emission characteristics of a compression-ignitionengine fueled with diesel-DMM blends. Proc Combust Inst2013;34(2):3013–20.

[5] Maricq MM, Chase RE, Podsiadlik DH, Seigl WO, Kaiser EW. The effect ofdimethoxy methane additive on diesel vehicle particulate emissions. SAEtechnical paper 982572; 1998.

[6] Murayama T, Zeng M, Chikahisa T, Arima T. Simultaneous reductions of smokeand NOx from a DI diesel engine with EGR and dimethyl carbonate. SAEtechnical paper 952518; 1995.

[7] Miyamoto N, Ogawa H, Arima T. Improvement of diesel combustion andemissions with addition of various oxygenated agents to diesel fuels. SAEtechnical paper 962115; 1996.

[8] Liotta FJ, Montalvo DM. The effect of oxygenated fuels on emissions from amodern heavy-duty diesel engine. SAE technical paper 932734; 1993.

[9] Sinha A, Thomson MJ. The chemical structures of opposed flow diffusionflames of C3 oxygenated hydrocarbons (isopropanol, dimethoxy methane, anddimethyl carbonate) and their mixtures. Combust Flame 2004;136(4):548–56.

[10] Glaude PA, Pitz WJ, Thomson MJ. Chemical kinetic modeling of dimethylcarbonate in an opposed-flow diffusion flame. Proc Combust Inst2005;30(1):1111–8.

632 E. Hu et al. / Fuel 140 (2015) 626–632

[11] Chen G, Yu W, Fu J, Mo J, Huang ZH, Yang JC, et al. Experimental and modelingstudy of the effects of adding oxygenated fuels to premixed n-heptane flames.Combust Flame 2012;159:2324–35.

[12] Bardin ME, Ivanov EV, Nilsson EJK, Vinokurov VA, Konnov AA. Laminar burningvelocities of dimethyl carbonate with air. Energy Fuels 2013;27:5513–7.

[13] Zhang JX, Niu SD, Zhang YJ, Tang CL, Jiang X, Hu EJ, et al. Experimental andmodeling study of the auto-ignition of n-heptane/n-butanol mixtures.Combust Flame 2013;160(1):31–9.

[14] Hu EJ, Jiang X, Huang ZH, Zhang JX, Zhang ZH, Man XJ. Experimental andkinetic studies on ignition delay times of dimethyl ether/n-butane/O2/armixtures. Energy Fuels 2012;27(1):530–6.

[15] Morley C. Gaseq v0.76. <http://www.gaseq.co.uk>.[16] Kee RJ, Rupley FM, Miller JA. Chemkin-II: A Fortran chemical kinetics package

for the analysis of gas-phase chemical kinetics. Report no. SAND89-8009.Sandia National Laboratories; 1989.

[17] Lutz AE, Kee RJ, Miller JA. Senkin: A Fortran program for predictinghomogeneous gas phase chemical kinetics with sensitivity analysis. Reportno. SAND87-8248. Sandia National Laboratories; 1988.

[18] Metcalfe WK, Burke SM, Ahmed SS, Curran HJ. A hierarchical and comparativekinetic modeling study of C1–C2 hydrocarbon and oxygenated fuels. Int JChem Kinet 2013;45(10):638–75.

[19] Dooley S, Curran HJ, Simmie JM. Autoignition measurements and a validatedkinetic model for the biodiesel surrogate, methyl butanoate. Combust Flame2008;153(1–2):2–32.

[20] Curran HJ, Gaffuri P, Pitz WJ, Westbrook CK. A comprehensive modeling studyof iso-octane oxidation. Combust Flame 2002;129(3):253–80.

[21] Held TJ, Dryer FL. A comprehensive mechanism for methanol oxidation. Int JChem Kinet 1998;30(11):805–30.

[22] Curran HJ, Gaffuri P, Pitz WJ, Westbrook CK. A comprehensive modeling studyof n-heptane oxidation. Combust Flame 1998;114(1–2):149–77.

[23] Fisher EM, Pitz WJ, Curran HJ, Westbrook CK. Detailed chemical kineticmechanisms for combustion of oxygenated fuels. Proc Combust Inst2000;28(2):1579–86.

[24] Luo YR. Comprehensive handbook of chemical bond energies. CRC Press;2007.

[25] Zhang JX, Wei LJ, Man XJ, Jiang X, Zhang YJ, Hu EJ, et al. Experimental andmodeling study of n-butanol oxidation at high temperature. Energy Fuels2012;26(6):3368–80.

[26] Man XJ, Tang CL, Zhang JX, Zhang YJ, Pan L, Huang ZH, et al. An experimentaland kinetic modeling study of n-propanol and i-propanol ignition at hightemperatures. Combust Flame 2014;161(3):644–56.

[27] McCunn LR, Lau KC, Krisch MJ, Butler LJ, Tsung JW, Lin JJ. Unimoleculardissociation of the CH3OCO radical: an intermediate in the CH3O + COreaction. J Phys Chem A 2005;110(4):1625–34.

[28] Diévart P, Won SH, Gong J, Dooley S, Ju YG. A comparative study of thechemical kinetic characteristics of small methyl esters in diffusion flameextinction. Proc Combust Inst 2013;34(1):821–9.

[29] Akih-Kumgeh B, Bergthorson JM. Experimental and modeling study of trendsin the high-temperature ignition of methyl and ethyl esters. Energy Fuels2011;25(10):4345–56.

[30] Black G, Curran HJ, Pichon S, Simmie JM, Zhukov V. Bio-butanol: combustionproperties and detailed chemical kinetic model. Combust Flame2010;157(2):363–73.

[31] Yasunaga K, Mikajiri T, Sarathy SM, Koike T, Gillespie F, Nagy T, et al. A shocktube and chemical kinetic modeling study of the pyrolysis and oxidation ofbutanols. Combust Flame 2012;159(6):2009–27.