Combustion and Flame 156 (2009) 1852–1855

Contents lists available at ScienceDirect

Combustion and Flame

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

Brief Communications

Autoignition of methylcyclohexane at elevated pressures

Gaurav Mittal 1, Chih-Jen Sung *

Department of Mechanical and Aerospace Engineering, Case Western Reserve University, Cleveland, OH 44106, USA

a r t i c l e i n f o a b s t r a c t

Article history:Received 15 February 2009Received in revised form 20 May 2009Accepted 25 May 2009Available online 24 June 2009

0010-2180/$ - see front matter � 2009 The Combustdoi:10.1016/j.combustflame.2009.05.009

* Corresponding author. Present address: DepartmeUniversity of Connecticut, Storrs, CT 06269, USA. Fax:

E-mail address: cjsung@engr.uconn.edu (C.-J. Sung1 Present address: Department of Mechanical Engine

Akron, OH 44325, USA.

Autoignition of methylcyclohexane (MCH) is investigated in a rapid compression machine (RCM). Exper-iments are conducted for homogeneous MCH/O2/N2/Ar mixtures with equivalence ratios varying from 0.5to 1.5, at compressed charge pressures of 15.1 and 25.5 bar, and over a compressed gas temperaturerange of 680–905 K. In the RCM experiments, the mole percentage of MCH is kept constant at 1.047%,while the equivalence ratio is varied through variation in oxygen mole percentage. Ignition delay mea-surements demonstrate the features of two-stage ignition characteristics and strong negative tempera-ture coefficient (NTC) behavior. The NTC behavior is also noted to shift to higher temperatures athigher pressures and for oxygen-rich mixtures. In addition, both the first-stage and total ignition delaysshow significant dependence on oxygen mole fraction. Furthermore, the current ignition delay measure-ments provide additional benchmark data for the validation of the oxidation kinetics of MCH. Comparisonof the experimental ignition delays with the simulated results using an available detailed kinetic mech-anism is conducted and discussed.

� 2009 The Combustion Institute. Published by Elsevier Inc. All rights reserved.

1. Introduction

Although cycloalkanes constitute a significant fraction in dieseland jet fuels, kinetic studies on cyclic alkanes are meager. Methyl-cyclohexane (MCH) is recognized as a candidate cycloalkane forsurrogate formulation, and is also considered as the starting pointfor the development of kinetic mechanisms of other cycloalkanes.Previous investigations on MCH combustion include low pressurepyrolysis [1], pyrolysis and oxidation of MCH and the MCH/tolueneblends in a turbulent plug flow reactor [2], laminar diffusion flameexperiments [3], high temperature shock tube experiments andmodeling [4], experiments and modeling in a rapid compressionmachine (RCM) [5], and shock tube ignition delay measurementsat elevated pressures [6,7]. In an attempt to develop a comprehen-sive mechanism for MCH, Pitz et al. [5] developed a low tempera-ture mechanism for MCH and merged it with a recent hightemperature mechanism from [4] to simulate their experimentsin an RCM.

It is recognized that MCH displays the features of low tempera-ture heat release and two-stage ignition behavior; resulting typi-cally from the reactions of peroxy radical. Because the influenceof low temperature heat release can become important in practicalcombustors and engines, a comprehensive detailed reaction mech-anism must take the reactions in the low temperature kinetic

ion Institute. Published by Elsevier

nt of Mechanical Engineering,+1 860 486 5088.

).ering, The University of Akron,

regime into account, and should correctly predict the features ofautoignition, such as the first- and second-stage ignition delaysand the pre-hot ignition heat release.

Experiments in an RCM allow access to conditions of low tem-peratures where two-stage ignition phenomenon is present. Apartfrom the recent investigation by Pitz et al. [5] at elevated pres-sures and low temperatures, there is no other RCM experimentin this operating regime. In this work, in order to provide bench-mark experimental data for validating and developing a compre-hensive chemical kinetic mechanism of MCH, experiments areconducted for autoignition of MCH in an RCM over a range ofcompressed charge pressures, compressed gas temperatures, andequivalence ratios. This experimental dataset is then used to as-sess the performance of an available reaction mechanism takenfrom [5].

In the following, the experimental and computational specifica-tions are first described, followed by results and discussions ofMCH autoignition experiments and model performance.

2. Experimental and computational specifications

Experiments are conducted in an RCM. The details of thepresent RCM and experimental procedure can be found in[8,9]. Autoignition investigations for three MCH/O2/N2/Ar mix-tures (denoted as Mix #1–3) are conducted over the compressedgas temperature range of TC = 680–905 K and at two compressedcharge pressure conditions of PC = 15.1 ± 0.3 and 25.5 ± 0.5 bar.Table 1 lists the compositions of the three gas mixtures testedherein.

Inc. All rights reserved.

0

5

10

15

20

25

30

-24 -16 -8 0 8 16 24 32 40

Pres

sure

(ba

r)

Time (ms)

P0 = 824 Torr

T0 = 296.4 K

TC = 705.7 K

PC = 15 bar

τ1

τ

Fig. 1. Pressure trace illustrating the definitions of the first-stage ignition delay (s1)and the total ignition delay (s). Conditions: Mix #1 (/ = 1.0), P0 = 824 Torr,T0 = 296.4 K, TC = 705.7 K, and PC = 15 bar.

Table 1List of compositions of gas mixtures investigated (in molar proportion).

Mix # / MCH O2 N2 Ar

1 1.0 1 10.5 12.25 71.752 0.5 1 21 0.00 73.53 1.5 1 7 16.35 71.15

G. Mittal, C.-J. Sung / Combustion and Flame 156 (2009) 1852–1855 1853

Moreover, the equivalence ratios corresponding to Mix #1, #2,and #3 are, respectively, / = 1.0, 0.5, and 1.5, with varying propor-tions of diluting N2 and Ar. MCH employed is 99.5% pure (suppliedby Fisher Scientific).

In the experiments, the mole percentage of MCH in the reac-tive mixture is kept constant at 1.047% and the equivalence ra-tio is varied through variation in the mole percentage of O2, asshown in Table 1. Additionally, the same specific heat ratio ismaintained for all of the mixtures investigated by adjustingthe relative proportion of N2 and Ar in the diluent. The ap-proach of maintaining the same specific heat ratio leads toidentical gas temperature and pressure at the end of compres-sion for all mixtures with the same compression ratio and ini-tial conditions of pressure and temperature. For a given mixturecomposition with initial temperature of room temperature, thecompressed gas temperature at the end of compression (topdead center, TDC), TC, is varied by altering the compression ra-tio; whereas the desired pressure at TDC, PC, is obtained byvarying the initial pressure of the reacting mixture for a givencompression ratio. The temperature at TDC is determined bythe adiabatic core hypothesis according to the relation ofR TC

T0

cc�1

dTT ¼ ln½PC=P0�, where P0 is the initial pressure, T0 the ini-

tial temperature, and c is the specific heat ratio that is a func-tion of temperature. The value of TC is taken as referencetemperature for reporting the data.

Numerical modeling of RCM experiments is performed usingthe SENKIN code [10] in conjunction with the CHEMKIN package[11]. The modeling begins from the start of compression stroke,and includes the effect of heat loss during the compression strokeand after the end of compression. In particular, an approach basedon ‘effective volume’ is applied to model the effect of heat loss.Description and validation of this heat transfer model are detailedin [8,9,12]. Simulations are conducted using the reaction mecha-nism of [5].

The experimental pressure traces and the associated empiricaleffective volume parameters are available in [13].

3. Results and discussion

Fig. 1 shows a typical pressure trace for the ignition of Mix#1 (/ = 1.0) with T0 = 296.4 K and P0 = 824 Torr, along with thedefinitions of the first-stage ignition delay (s1) and the total igni-tion delay (s). Specifically, the ignition delays (s1,s) are definedas the time from the end of the compression stroke, where thepressure peaks at t = 0, to the corresponding instant of inflectionpoint in the pressure history. For this case, the measured pres-sure and the deduced temperature at the end of compression(t = 0) are PC = 15 bar and TC = 705.7 K, respectively. There is anuncertainty in the reported compressed gas temperature due tothe errors associated with the partial pressure measurementsfor mixture preparation, the dynamic pressure measurements re-lated to the accuracy of piezoelectric transducer, and the initialtemperature measurements using a thermocouple. Followingthe uncertainty propagation approach of [14], the maximumuncertainty in TC for all experiments is estimated to be less than3.5 K and is mainly due to the uncertainty in the initial temper-ature measurement.

Fig. 2 presents the raw experimental pressure traces that areused to obtain the ignition delay results. Each plot in Fig. 2shows pressure traces for given mixture composition and com-pressed charge pressure, while varying compressed gas temper-ature. The pressure traces in Fig. 2 highlight the detailedfeatures observed in the ignition behavior, including the first-stage pressure rise and the hot ignition event. It is seen fromFig. 2 that two-stage ignition response is apparent at low com-pressed gas temperatures, while single-stage ignition is observeat high TC. Furthermore, as TC is increased, the pressure rise dueto the first-stage activity decreases and eventually it ceases toexist. This decrease in the first-stage pressure rise with increas-ing TC suggests the existence of a cut-off temperature for thefirst-stage activity. As a result, at low TC the first-stage activityyields higher pressure rise to attain the cut-off temperature atthe end of the first-stage event. It is noted that Griffiths et al.[15] used pressure traces from an RCM to estimate the averagegas temperature in the transition from the first stage to thesecond stage for various alkanes, and reported that the averagetemperature after the first-stage activity remained approxi-mately constant.

Fig. 2 also shows that for given PC and TC, the first-stage pres-sure rise for an oxygen-rich mixture is higher than oxygen-leanmixture. This behavior is due to the enhancement of peroxida-tion reactions when O2 concentration is increased. An increasedO2 concentration also favors low temperature heat release andincreases the cut-off temperature for the first-stage activity. Itis further seen from Fig. 2 that the limiting value of TC beyondwhich the first-stage activity ceases is higher for oxygen-richmixtures.

Fig. 3 shows the measured s and s1 with varying TC for differentequivalence ratios at PC � 15.1 and 25.5 bar. Since / is variedthrough the change in the mole fraction of O2 while keeping themole fraction of MCH constant, the effect of / on ignition delaysobserved here represents the influence of O2 concentration. It isseen from Fig. 3 that both s and s1 strongly depend on / and igni-tion delays increase with increasing / (decreasing O2 mole frac-tion). At lower TC, ignition delays tend to converge and becomeindependent of /. The observed behavior of strong dependenceof s1 on / (O2 mole fraction) for MCH is a feature different fromthe behaviors of other fuels such as dimethylether [16] and n-dec-ane [17] that show negligible dependence of s1 on /. It is alsonoted that the onset of negative temperature coefficient (NTC) re-gion is shifted to lower temperatures as / (O2 mole fraction) is in-

0

5

10

15

20

25

30

35

-20 0 20 40 60 80 100 120

Pres

sure

(bar

)

Time (ms)

TC=705 K680692

737

754772

837791

(a)

Mix #1, φ=1.0PC~15.1 bar

0

10

20

30

40

-16 -8 0 8 16 24 32 40

Pres

sure

(bar

)

Time (ms)

682694

708TC=741 K

758 778822

(b)

Mix #1, φ=1.0 PC~25.5 bar

0

5

10

15

20

25

30

35

-20 -10 0 10 20 30 40 50

Pres

sure

(bar

)

Time (ms)

680693

706TC=736 K754

771773

(c)

Mix #2, φ=0.5PC~15.1 bar

0

10

20

30

40

-20 -10 0 10 20 30 40

Pres

sure

(bar

)

Time (ms)

681695TC=758 K

797777

821(d)

Mix #2, φ=0.5PC~25.5 bar

0

5

10

15

20

25

30

35

0 40 80 120 160 200 240 280

Pres

sure

(bar

)

Time (ms)

TC=706 K692

721 736753 772

793817

(e)

Mix #3, φ=1.5PC~15.1 bar

0

10

20

30

40

-20 0 20 40 60

Pres

sure

(bar

)

Time (ms)

799

696TC=710 K

725

758

824

741

778

(f)

Mix #3, φ=1.5PC~25.5 bar

Fig. 2. Experimental pressure history for various mixtures and compressed conditions (a) Mix #1 (/ = 1.0), PC � 15.1 bar, (b) Mix #1 (/ = 1.0), PC � 25.5 bar, (c) Mix #2(/ = 0.5), PC � 15.1 bar, (d) Mix #2 (/ = 0.5), PC � 25.5 bar, (e) Mix #3 (/ = 1.5), PC � 15.1 bar, and (f) Mix #3 (/ = 1.5), PC � 25.5 bar.

1854 G. Mittal, C.-J. Sung / Combustion and Flame 156 (2009) 1852–1855

creased (decreased). Furthermore, at higher PC, the NTC region isshifted to higher TC.

We further note that the reported s1 in Fig. 3 first decreaseswith increasing temperature, levels off, and then increases slightly.This behavior is due to the combined influence of the heat loss ef-fect and the definition of s1 on the basis of inflection point duringthe first-stage activity. As TC is increased, the induction time for thefirst-stage activity decreases. However, at the same time, the rateof pressure rise associated with the first-stage activity also de-creases. Since the first-stage activity has to overcome the pressuredrop due to heat loss for a net pressure rise and an apparent inflex-ion point, the deduced first-stage ignition delay may show an in-crease with increasing TC when the rate of first-stage heat releaseis very small.

A comparison of experimental and simulated ignition delaysfor / = 1.0/PC = 25.5 bar and / = 0.5/PC = 15.1 bar is shown inFig. 4. Although the reaction mechanism of [5] does show the fea-tures of NTC, substantial quantitative discrepancy is noted. Themechanism predicts higher s and s1. Pitz et al. [5] had noted thatthe predictions from kinetic mechanism were particularly sensi-tive to isomerization rates of methylcyclohexyl radicals. Specifi-cally, the importance of relative rates of isomerization through5-, 6-, and 7-membered transition state ring structures and theirimpacts on chain branching of the system were highlighted. The-oretical calculations and experimental measurements of speciesevolution profiles under conditions of low-to-intermediate tem-peratures are therefore required for further kinetic mechanismdevelopment.

Fig. 3. Experimental data of total (filled symbols) and first-stage (open symbols) ignition delays for different equivalence ratios (Mix #1–3): (a) PC � 15.1 bar and (b)PC � 25.5 bar.

Fig. 4. Comparison of experimental (symbols) and computed (lines) ignition delaysfor (a) PC � 25.5 bar, Mix #1 (/ = 1.0) and (b) PC � 15.1 bar, Mix #2 (/ = 0.5).

G. Mittal, C.-J. Sung / Combustion and Flame 156 (2009) 1852–1855 1855

Acknowledgments

This work has been supported by the Air Force Office of Scien-tific Research under Grant No. FA9550-07-1-0515 and by theNational Aeronautics and Space Administration under Grant No.NNX07AB36Z.

References

[1] T.C. Brown, K.D. King, Int. J. Chem. Kinet. 21 (1989) 251–266.[2] S. Zeppieri, K. Brezinsky, I. Glassman, Combust. Flame 108 (1997) 266–286.[3] C.S. McEnally, L.D. Pfefferle, Proc. Combust. Inst. 30 (2005) 1425–1432.[4] J.P. Orme, H.J. Curran, J.M. Simmie, J. Phys. Chem. A 110 (2006) 114–131.[5] W.J. Pitz, C.V. Naik, T. Ní Mhaoldúin, C.K. Westbrook, H.J. Curran, J.P. Orme, J.M.

Simmie, Proc. Combust. Inst. 31 (2007) 267–275.[6] S.S. Vasu, D.F. Davidson, R.K. Hanson, Combust. Flame 156 (2009) 736–749.[7] J. Vanderover, M.A. Oehlschlaeger, Int. J. Chem. Kinet. 41 (2009) 82–91.[8] G. Mittal, C.J. Sung, Combust. Sci. Technol. 179 (3) (2007) 497–530.[9] G. Mittal, C.J. Sung, Combust. Flame 150 (2007) 355–368.

[10] A.E. Lutz, R.J. Kee, J.A. Miller, Sandia Report SAND 87-8248, Sandia NationalLaboratories, 1987.

[11] R.J. Kee, F.M. Rupley, J.A. Miller, Sandia Report SAND 89-8009B, SandiaNational Laboratories, 1989.

[12] G. Mittal, M.P. Raju, C.J. Sung, Combust. Flame 155 (2008) 417–428.[13] Available from: <http://www.mae.case.edu/facilities/cdl/projects/rapidcomp/

rapiddatabase/>.[14] H.W. Coleman, W.G. Steele, Experimentation and Uncertainty Analysis for

Engineers, Wiley, New York, 1989 (Chapter 3).[15] J.F. Griffiths, P.A. Halford-Maw, D.J. Rose, Combust. Flame 95 (1993) 291–306.[16] G. Mittal, M. Chaos, C.J. Sung, F.L. Dryer, Fuel Process. Technol. 89 (2008) 1244–

1254.[17] K. Kumar, G. Mittal, C.J. Sung, Combust. Flame 156 (2009) 1278–1288.