Fuel 128 (2014) 21–29

Contents lists available at ScienceDirect

Fuel

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

A shock tube ignition delay study of conventional diesel fueland hydroprocessed renewable diesel fuel from algal oil

http://dx.doi.org/10.1016/j.fuel.2014.02.0640016-2361/� 2014 Elsevier Ltd. All rights reserved.

⇑ Corresponding author. Tel.: +1 518 276 8115.E-mail address: oehlsm@rpi.edu (M.A. Oehlschlaeger).

Sandeep Gowdagiri, Weijing Wang, Matthew A. Oehlschlaeger ⇑Mechanical, Aerospace, and Nuclear Engineering, Rensselaer Polytechnic Institute, Troy, NY, USA

h i g h l i g h t s

� Ignition delay measurements reported for conventional and alternative diesels.� Ignition delay indistinguishable for T > 1000 K, varies substantially for T < 1000 K.� Empirical model is presented for ignition variability.

a r t i c l e i n f o

Article history:Received 28 January 2014Received in revised form 25 February 2014Accepted 26 February 2014Available online 14 March 2014

Keywords:DieselHydroprocessed fuelsIgnition delay timeShock tube

a b s t r a c t

Ignition delay time measurements were made using the shock tube method for a conventional petro-leum-derived military diesel fuel, F-76, and an alternative hydroprocessed renewable diesel fuel derivedfrom hydroprocessing algal oils, HRD-76. Measurements were made in the reflected shock region usingpressure and chemiluminescence and carried out for temperatures ranging from 671 to 1266 K, at nom-inal pressures of 10 and 20 atm, and for fuel–air equivalence ratios of 0.5 and 1.0. The measurementsshow that the high-temperature (>1000 K) reactivity of the two fuels is indistinguishable and very similarto that for petroleum-derived and alternative jet fuels and biodiesels. While at low temperatures(<1000 K) the ignition delay time for HRD-76 is up to a factor of two shorter than F-76 due to the largerfraction of n-paraffins and lack of aromatics in HRD-76. The combination of the present diesel ignitionmeasurements with previous studies for jet fuel ignition illustrate common characteristics for the gas-phase ignition of these high molecular weight mostly-aliphatic fuels, notably common high-temperatureignition characteristics and low-temperature reactivity that correlates with derived cetane number(DCN). These characteristics allow for formulation of a simple three-Arrhenius model for ignition delaytime that incorporates DCN functionality to describe variability in fuel reactivity.

� 2014 Elsevier Ltd. All rights reserved.

1. Introduction

Worldwide consumption of petroleum distillate fuel oil, whichincludes diesel fuels and fuel oils, has been around 25 million bar-rels per day since 2007–2008 [1] where it accounts for about 10%of worldwide energy consumption. Diesel fuels are primarily usedin transportation and power-generation where they are most oftenburned in diesel engines. The prediction of autoignition, whichinitiates combustion in diesel and other compression-ignitionengines, is of critical importance to the computational design andoptimization of diesel engines, today largely based on computa-tional fluid dynamics calculations that include fuel chemistry.Additionally, many alternative diesel fuels have been developed

in recent years, including biodiesels from plant oils, animal fats,and waste oils; synthetic fuels from gas-to-liquid and coal-to-liquid processes; synthetic fuels from hydroprocessing bio-oils;and fuels from genetically engineered microbes. These fuels canhave different physical and chemical kinetic properties, in somecases unknown, which can influence their use in engines.

Most experimental work on the autoignition of diesel fuelsfound in the literature is based on the autoignition of fuel spraysin constant volume chambers [2], engines [3], or rapid compressionmachines [4]. While there have been many studies of the gas-phase autoignition of diesel fuel surrogate compounds [5–8], typi-cally smaller compounds (e.g., n-heptane), very few experimentshave been performed where the gas-phase autoignition of multi-component diesel fuels has been isolated from evaporation andmixing phenomena, due to the low volatility of diesel fuels. Thisvoid in the literature is problematic due to the importance oflow-volatility fuels in the transportation and energy sectors and

22 S. Gowdagiri et al. / Fuel 128 (2014) 21–29

the growing number of kinetic models (see reviews [9,10]) thatlack sufficient targets for validation.

Here we present measurements of ignition delay times for con-ventional and alternative diesel fuels carried out using the re-flected shock technique, which provides well defined reactionconditions of relevance to combustion engine applications. Priorshock tube studies of gas-phase multi-component petroleum die-sel fuel autoignition are limited, to our knowledge they only in-clude the work Haylett et al. [11] and Penyazkov et al. [12].Haylett et al. [11] measured ignition delay times for four dieselfuels at 3–7 atm, temperatures from 950 to 1400 K, and for mix-tures containing 4% and 21% O2 in argon with highly variable fuelloading. Haylett et al. used an aerosol shock tube technique wherethe fuel was introduced into the shock tube as an aerosol, the inci-dent shock was used to vaporize the fuel, and measurements weremade behind the reflected shock. Penyazkov et al. [12] used aheated shock tube method, similar to that utilized here, to measureignition delay times for a Diesel fuel No. 2 (DF-2) in air at 5–10 atm, 1065–1838 K, and for equivalence ratios of 0.5, 1, and 2.

There has been important recent progress towards developingsynthetic non-oxygenated alternative diesel and jet fuels fromrenewable biomass sources using hydroprocessing to convertbio-oils to paraffinic hydrocarbons. These fuels have been labeledas hydroprocessed renewable diesel or jet (HRD or HRJ) fuels andhave been demonstrated in application (e.g., the US Navy 2012Rim of the Pacific demonstration exercise in anticipation of the‘‘Great Green Fleet’’ [13]) and in some cases certified [14] for usein diesel and gas turbine engines. Due to their largely straight chainand branched paraffinic composition, HRD and HRJ fuels can havedifferent combustion properties than petroleum fuels containingappreciable quantities of aromatics and cyclo-alkanes with lowerreactivity, higher density, and other differing properties. Oneimportant observation has been the higher reactivity of HRJ fuelsrelative to petroleum jet fuels [15]. Here we present a comparisonof the ignition delay time, a measure of global reactivity, for a con-ventional military diesel fuel, F-76, and a HRD fuel synthesizedfrom algal oil, HRD-76.

2. Experimental method

Ignition delay time measurements were carried out in theRensselaer heated shock tube facility (5.7 cm inner diameter,2.59 m long driver, 4.11 m long driven) using the reflected shocktechnique. The facility and experimental methods for ignition de-lay time measurements have been previously described [16,17].Here only details pertinent to the present measurements are given.

Due to the high molecular weight of the diesel fuels studied, theshock tube and mixing vessel were heated to 200 �C to provide suf-ficient fuel volatility for gas-phase fuel/air mixture preparation (airis 99.995% pure O2 and N2 at a ratio of 1–3.76) and avoid fuel con-densation during shock tube experiments. The heated shock tubeexhibited axial temperature uniformity of ±4 K during the experi-ments. Gas chromatography/mass spectrometry measurements(Shimadzu QP5050) were made to ensure that the diesel fuels werenot partially decomposed at the 200 �C mixture preparationcondition.

Experiments were carried out for two diesel fuels: F-76, a mili-tary grade petroleum-derived diesel fuel, and HRD-76, an alterna-tive diesel fuel from hydroprocessed algal oil. Chromatograms andthe hydrocarbon class distribution for the two fuels are given inFigs. 1 and 2, respectively. Some fuel properties are given in Table 1,while additional detailed fuel data sheets are appended as supple-mentary material. The experimental conditions for which ignitiondelay time measurements were carried out are given in Table 2.Measurements at the lowest temperatures were limited to lean

conditions (/ = 0.5), requiring less fuel loading (lower temperatureshock tube experiments require higher initial total pressure), dueto the relatively low volatility of the fuels studied. It can be seenthat F-76 contains cyclo-alkanes, aromatics, and other compoundsnot found in HRD-76, which is wholly comprised of normal andlightly-branched alkanes. These compositional differences resultin a substantially higher derived cetane number (DCN measure-ments found in Table 1, carried out by Advanced Engine Technol-ogy Ltd. using ASTM D6890) for HRD-76 relative to F-76. F-76also contains a broader range of molecular weights, as shown bythe chromatograms. For the purposes of defining fuel/air mixturestoichiometry, the average molecular formula (CmHy in Table 1)was defined using a correlation between distillation curve andmolecular weight from Maxwell and Bonnell [18] and the hydro-gen-to-carbon ratio (H/C Table 1, hydrogen content determinedusing ASTM D7171).

For the determination of ignition delay times behind reflectedshock waves, pressure and electronically-excited OH chemilumi-nescence were measured. See Fig. 3 for an example ignition delaytime measurement. Ignition delay time was defined as the timeinterval between shock reflection at the driven section end wall,defined using the measured pressure profile at a side wall location(2 cm from end wall) and the measured incident shock velocity,and the onset of ignition at the end wall, defined by extrapolatingthe maximum slope in the OH signal to the baseline (acquiredthrough the shock tube end wall). The pressure gradient behindthe reflected shock wave, due to viscous gasdynamics, for theseexperiments was (dP/dt)(1/P0) = 2–3% ms�1. This pressure gradientshould be considered in kinetic modeling of these experiments,particularly for ignition delay times longer than 2 ms which couldbe reduced by tens to near one hundred percent compared to igni-tion delay times predicted for constant volume conditions, depend-ing on the conditions and model.

Reflected shock conditions were determined using the normalshock equations with the measured shock velocity and reactantmixture thermodynamic properties as inputs. Thermodynamic sur-rogates were used to represent F-76 and HRD-76 for the definingreactant thermodynamic properties and were formulated to emu-late the fuel H/C, average molecular weight, and organic distribu-tion given in Fig. 2, as in [16]. The thermodynamic surrogate forthe fuel has minimal influence on the calculated reflected shockcondition because the reactant thermodynamic properties aredominated by the �99% air contained in the fuel/air mixtures.For example, the difference in reflected shock temperature for acalculation carried out using a thermodynamic surrogate, as de-scribed above, and a calculation carried out assuming a pure n-al-kane fuel, with both calculations containing the same carboncontent (that of the experiment), is approximately 1 K.

All measured ignition delay times are reported in the appendedsupplementary material. Ignition delay times have uncertaintiesestimated at ±25% based on uncertainties in determination of igni-tion delay time from measured signals and the uncertainties inreactant mixture composition and reflected shock conditions. Theuncertainty in reflected shock temperature (±1–1.5%), based onuncertainty in incident shock velocity, has the largest contributionto ignition delay time uncertainty.

3. Results and discussion

Measurements of ignition delay times are shown in Fig. 4 for F-76 and HRD-76. In both cases the high-temperature results(>1000 K) exhibit an expected decrease in ignition delay time withincreasing equivalence ratio and pressure. The / = 0.5 fuel/air re-sults at 20 atm demonstrate the three regions of reactivity charac-teristic of aliphatic fuels: the high-temperature region, exhibiting

Fig. 1. Chromatograms for F-76 and HRD-76.

Fig. 2. Distribution among hydrocarbon classes for F-76 and HRD-76 [19].

Table 1Fuel properties.

Fuel MWavg (g/mol) H/C Empirical formula DCN

F-76 193 1.80 C14.0H25.1 48.8HRD-76 218 2.12 C15.4H32.7 78.5

S. Gowdagiri et al. / Fuel 128 (2014) 21–29 23

Arrhenius-like temperature dependence; the negative-tempera-ture-coefficient (NTC) region, exhibiting a reduction in ignition de-lay time with decreasing temperature (in this case very moderateNTC behavior occurs); and the low-temperature region, wherethe ignition delay time returns to positive apparent activationenergy. A comparison of F-76 and HRD-76 at common conditions(/ = 0.5, 20 atm) in Fig. 5 shows that the high-temperature ignition

delay times for these fuels are indistinguishable. This is also thecase at / = 1 and 10 and 20 atm. However, at temperatures lowerthan 1000 K HRD-76 has significantly shorter ignition delay timethan F-76, up to a factor of two at 750–800 K.

The high-temperature (>1000 K) ignition delay time results forboth fuels can be correlated in combination using power-law scal-ing for ignition delay time dependence on pressure and equiva-lence ratio and Arrhenius temperature dependence [11,16]. Aleast squares regression provides the following correlations forT > 1000 K:

s ¼ 1:7ð�0:4Þ � 10�8/�0:36ð�0:05ÞP�0:71ð�0:05Þ

expð13000� 200=T½K�Þ seconds

The efficacy of the correlation is demonstrated in Fig. 6, wherethe high-temperature data are correlated to common conditionswhere they exhibit extremely low scatter (1r scatter is ± 8%) andagain indistinguishable ignition delay time for the two fuels justi-fying the common high-temperature correlation.

The high-temperature results are compared to the previousstudies of diesel fuel ignition by Haylett et al. [11] and Penyazkovet al. [12] in Fig. 7. Because these studies were carried out at differ-ent conditions, the above correlation was used to scale the studiesto a common condition (/ = 1 and 20 atm) for comparison. TheHaylett et al. measurements for United States and Europeansourced DF-2 were carried out with DF-2/21% O2/argon mixturesfor / = 0.32–1.39 and P = 2.3–8 atm (US DF-2) and / = 0.17–0.38and P = 3.5–7.4 atm (Europe DF-2). The Penyazkov et al. studywas for DF-2/air mixtures at / = 0.5, 1, and 2 and P = 4.7–10.4 atm. Hence the three studies represent a range in equivalenceratio from �0.2 to 2 and pressure from �2 to 20 atm but all threestudies were carried out for mixtures containing 20–21% O2. TheFig. 7 comparisons illustrate that all three studies agree for ignitiondelay time, within the scatter of the data.

Table 2Experimental conditions.

Mixture Equivalenceratio, /

Pressure(atm)

Temperature(K)

Ignition delay time, s(ls)

F-76/air 0.5 20 705–1241 81–4716F-76/air 1 10 951–1266 86–1988F-76/air 1 20 922–1219 82–1428HRD-76/air 0.5 20 671–1224 97–5633HRD-76/air 1 10 947–1212 141–1901HRD-76/air 1 20 883–1207 90–1654

Fig. 3. Example ignition delay time measurement. Pressure measurement made atside wall location 2 cm from end wall and OH� measurement made through endwall.

24 S. Gowdagiri et al. / Fuel 128 (2014) 21–29

Further comparisons of the diesel high-temperature ignitiontimes with literature data for petroleum-derived and alternativejet fuels are shown in Fig. 8 and comparisons with biodiesels areshown in Fig. 9. Again all data exhibit similar ignition delay time,irrespective of fuel composition, indicating that common oxidationchemistry, in this case C0–C3 (fuel fragments) chemistry, governsthe high-temperature ignition of fuels containing primarily paraf-finic functionalities and that kinetic modeling of this region canbe severely reduced in complexity if ignition is the sole desiredmodel target.

At lower temperatures (<1000 K) combustion chemistry is sig-nificantly more complicated [24] and intricacies of the molecularstructure or fuel composition play an important role in governing

Fig. 4. Ignition delay time measurements for F-7

overall fuel reactivity. However, much of the details of low-tem-perature are captured by the derived cetane number (DCN), mea-sured in the ignition quality tester (IQT) constant volumecombustion chamber, or the cetane number, measured using aCooperative Fuel Research (CFR) engine. To demonstrate, Fig. 10shows a comparison of ignition delay times for F-76 and HRD-76with measurements for a Jet A sample also made in our laboratoryusing similar methods [16]. The ignition delay time temperaturedependence for all the three fuels is similar and at low tempera-tures (700–900 K) there is a clear ordering in ignition delay timefrom longest to shortest with DCN from smallest to largest, withJet A (DCN = 47.1) only very slightly longer than F-76(DCN = 48.8) which is much longer than HRD-76 (DCN = 78.5).

The DCN-ignition delay time relationship is expanded upon inFig. 11, where ignition delay time is plotted against DCN for threecases: (1) the IQT experiment, where DCN is determined from theignition of a fuel spray into heated air at 829 K and 20 atm; (2)shock tube gas-phase fuel/air ignition at / = 0.5, 20 atm, and830 K; and (3) shock tube gas-phase fuel/air ignition at / = 1.0,20 atm, and 830 K. Data in Fig. 11 comes from the present studyand the jet fuels studies of Wang and Oehlschlaeger [16] and Valcoet al. [25]; ignition delay time data for jet fuels is available in theliterature to a larger degree than for diesel fuel. Fig. 11 illustratesthat the ignition delay time correlation with DCN in the NTC regionaround 830 K, chosen because it is the temperature at which theIQT experiment is used to determine DCN. Shock tube ignition de-lay time for fuel/air mixtures at 830 K to first order obeys a powerlaw scaling with DCN (in the range s / DCN�1 to s / DCN�1:5), withno discernible dependence on the molecular weight of the fuels(diesel vs. heavy jet fuels vs. light jet fuels) or the specifics of thehydrocarbon distribution of the fuels. The DCN-ignition time rela-tionship is useful because it indicates that many of the complexi-ties of low-temperature and NTC combustion kinetics can to firstorder be represented with a small number of parameters or a sin-gle parameter such as the DCN, effectively a low-temperaturechemistry transfer function, as proposed and utilized to formulatesimple surrogate mixtures to emulate the reactivity of composi-tionally-complex fuels by Dooley et al. [26,27]. The simple depen-dence of global reactivity (ignition delay time) on DCN can allowfor simple representations of low-temperature chemistry in com-putationally-intensive CFD calculations used to model low-tem-perature ignition phenomena, for example in diesel engines. Ofcourse the use of the DCN as a single parameter to representlow-temperature chemistry is a first-order approximation andhas limitations especially for compositionally unique fuels [28].

6/air (left) and HRD-76/air (right) mixtures.

Fig. 5. Comparison of ignition delay times for F-76 and HRD-76 at / = 0.5 and20 atm.

Fig. 6. Ignition delay times for F-76 and HRD-76 scaled to common conditions (/= 1 and 20 atm) using the high-temperature correlation.

Fig. 7. Comparison of ignition delay times from the present study with dieselmeasurements from Haylett et al. [11] and Penyazkov et al. [12]. Ignition timesscaled to common conditions (/ = 1 and 20 atm) using the high-temperaturecorrelation.

Fig. 8. Comparison of ignition delay times from the present study with measure-ments for jet fuels from Wang and Oehlschlaeger [16] and Vasu et al. [20]. Ignitiontimes scaled to common conditions (/ = 1 and 20 atm) using the high-temperaturecorrelation.

S. Gowdagiri et al. / Fuel 128 (2014) 21–29 25

Fundamentally, low-temperature reactivity is governed by thecomplex multi-pathway low-temperature oxidation reactionkinetics [24] that must be resolved should a first principles rela-tionship between fuel structure and reactivity be desired or othernon-ignition fuel chemistry characterizes be of interest.

The similarity of high-temperature ignition delay time shown inFigs. 5–8 for diverse jet and diesel fuels and the simple dependenceof ignition delay time on DCN shown in Fig. 11 motivates a prag-matic empirical model for ignition delay time based on a slightmodification to the three-Arrhenius model of Weisser [29] and alsoutilized by Vandersickel et al. [30]. The Weisser model definesthree time constants for low-, moderate-, and high-temperatureoxidation. The overall ignition time at any temperature is ex-pressed as a combination of those three time constants:

1s¼ 1

sL þ sMþ 1

sH

where s is the ignition time and sL, sM, and sH the time constants forlow-, moderate-, and high-temperature chemistry, respectively. SeeFig. 12 for a graphical description of the model.

The three-Arrhenius model time constants can be expressedwith Arrhenius temperature dependence and power-law functionsfor pressure, equivalence ratio, and DCN dependence, where incor-porating DCN functionality to describe ignition delay time depen-dence on fuel variability is the unique contribution of the presentwork:

si ¼ AiP

Pref

� �ai

/biDCN

DCNref

� �ci

expTact;i

T

� �for i ¼ L;M;H

Pref and DCNref are a reference pressure and reference DCN, and Ai, ai,bi, ci, and Tact;i are chosen for each temperature range to fit experi-mental data. Here this model was adjusted to best fit a large rangeof target data from measurements carried out in our laboratory fordiesel fuels (present study) and jet fuels [16,25]. The data repre-sents a range of experimental conditions relevant to engines:DCN = 31–78.5, / = 0.25–1.5, P = 8–39 atm, and T = 650–1300 K.The best-fit model parameters are given in Table 3.

Fig. 9. Comparison of ignition delay times from the present study with biodieseland biodiesel surrogate measurements from Wang et al. [21–22] (methyl decanoateat / = 1 and 16 atm; soybean and animal fat biodiesel at / = 0.25 and 10 atm) andCampbell et al. [23] (methyl decanoate at / = 0.3–0.6 and 7 atm). Ignition timesscaled to common conditions (/ = 1 and 20 atm) using the high-temperaturecorrelation.

Fig. 10. Comparison of ignition delay times from the present study with measure-ments for Jet A from Wang and Oehlschlaeger [16].

Fig. 11. Variation in ignition delay times with DCN for jet and diesel fuels atcommon experimental conditions. The jet fuel data is from Wang and Oehlschlaeger[16] and Valco et al. [25].

Fig. 12. Three-Arrhenius model for ignition delay time, adapted from [29,30].

Table 3Three-Arrhenius model parameters.

L, low T M, moderate T H, high T

Ai (ls) 1.4 � 10�9 2.0 � 101 2.2 � 10�8

ai �1 �1 �0.8bi �0.7 �0.7 �0.4ci �1.5 �1.5 0Tact,i (K) 12,000 �5000 13,000Pref (atm) 1 1 1DCNref 50 50 50

26 S. Gowdagiri et al. / Fuel 128 (2014) 21–29

Comparisons between the three-Arrhenius model and the targetdiesel and jet fuel experimental ignition delay time data are givenin Figs. 13–15, exhibiting relatively good agreement that is charac-terized for the entire target data set in Fig. 16. The standard devi-ation (1r) for the dispersion between experimental data and thethree-Arrhenius model (Fig. 16) is ±20% of ignition delay time.The ignition delay time measurements show shot-to-shot scatter(reproducibility) of around ± 10% (1r) and we report estimateduncertainty of ±25% (2r confidence interval) for these data basedon an analysis of the potential systematic uncertainties. Hence,the present three-Arrhenius model is predictive to within abouttwice the experimental uncertainty limits for a 2r confidenceinterval; i.e., the 2r (95%) confidence interval for experimentaluncertainty is ±25% and 2r model-experiment correlation is±40%. As this suggests, the three-Arrhenius model provides afirst-order description of the dependence of ignition delay timeon conditions and DCN, chosen to represent fuel variability, butdoes not capture all the complexities of fuel chemistry. However,

it is predictive to within the limits of most a priori kinetic modelingstudies found in the literature and provides a simple way of pre-dicting fuel global reactivity for different fuels with known DCNsand for differing conditions.

The most obvious experimental observation not captured by thethree-Arrhenius model is exhibited in the lowest temperature dataaround 700 K, where the model overpredicts in some cases andunderpredicts in other cases measured ignition delay time. Theselowest-temperature deviations do not correlate with DCN. Thiscould be due to fuel compositional influences on ignition at700 K that are not captured by DCN, measured at 830 K. Onehypothesis is that these effects are caused by low reactivity

Fig. 13. Comparison of diesel fuel ignition delay time measurements with the present three-Arrhenius model.

Fig. 14. Comparison of jet fuel ignition delay time measurements [16,25] with the present three-Arrhenius model.

Fig. 15. Comparison of Jet A ignition delay time measurements [16] at variable equivalence ratio (left) and pressure (right) with the present three-Arrhenius model.

S. Gowdagiri et al. / Fuel 128 (2014) 21–29 27

aromatics and cyclo-alkanes. Perhaps, these compounds have amore inhibiting influence on reactivity at 700 K than they do at830 K and hence DCN does not fully capture their influence atthe lowest temperatures considered.

4. Summary

Ignition delay time measurements have been carried out for twodiesel fuels, conventional military-grade F-76 and an alternative

Fig. 16. Correlation between all target diesel and jet fuel ignition delay time data and the present three-Arrhenius model (left); distribution of experiment-model deviation(right).

28 S. Gowdagiri et al. / Fuel 128 (2014) 21–29

hydroprocessed renewable diesel fuel derived from algal oil HRD-76. Experiments were performed in reflected-shock-heated gasesfor 671–1266 K, 10 and 20 atm, and fuel–air equivalence ratios of0.5 and 1.0. At temperatures in excess of 1000 K, F-76 and HRD-76show indistinguishable ignition delay time and also similar ignitiondelay time to jet fuels [16,20] and biodiesels [21–23]. Comparisonsof the present measurements with high-temperature ignition delaytime measurements for diesel fuels (DF-2) carried out in other lab-oratories [11,12] also show good agreement. At lower temperatures,F-76 has an ignition delay time that is significantly longer than forHRD-76, around a factor of two longer at 750–800 K, owing to thelarger n-paraffin content and lack of aromatics in HRD-76.

The present study, focused on diesel fuels, and prior studies onjet fuel ignition [20,25] show common high-temperature ignitionbehavior amongst compositionally-variable fuels and low-temper-ature ignition that varies appreciably with fuel composition butthat correlates well with the derived cetane number (DCN). Theseignition delay time behaviors allow for the formulation of a simplethree-Arrhenius model [29,30] to describe the ignition of highmolecular weight and mostly aliphatic diesel and jet fuels withinthe bounds of the experimental data considered. The simple modelutilizes three time constants to describe time scales for chemistryin three temperature regimes with temperature, pressure, equiva-lence ratio, and DCN dependence incorporated into each timeconstant. The inclusion of DCN functionality in the presentthree-Arrhenius model provides a simple way for predicting igni-tion delay time for reactivity-variable diesel and jet fuels across arange of DCN (31–78.5) relevant to their utilization in compressionignition engines.

Acknowledgement

This work was supported by the Office of Naval Research undergrant #N00014-11-1-0839 with Dr. Sharon Beermann-Curtin astechnical monitor.

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.2014.02.064.

References

[1] U.S. Energy Information Administration. Petroleum Supply Annual, volume 1;2013. <http://www.eia.gov/petroleum/supply/annual/volume1/>.

[2] Knothe G, Matheaus AC, Ryan III TW. Cetane numbers of branched andstraight-chain fatty esters determined in an ignition quality tester. Fuel2003;82:971–5.

[3] Rothamer DA, Murphy L. Systematic study of ignition delay for jet fuels anddiesel fuel in a heavy-duty diesel engine. Proc Combust Inst 2013;34:3021–9.

[4] Kobori S, Kamimoto T, Aradi AA. A stude of ignition delay of diesel fuel sprarys.Int J Engine Res 2000;1:29–39.

[5] Pfahl U, Fieweger K, Adometi G. Self-ignition of diesel-relevant hydrocarbon-air mixtures under engine conditions. Proc Combust Inst 1996;26:781–9.

[6] Vasu SS, Davidson DF, Hong Z, Vasudevan V, Hanson RK. N-Dodecane oxidationat high-pressures: measurements of ignition delay times and OHconcentration time-histories. Proc Combust Inst 2009;32:173–80.

[7] Oehlschlaeger MA, Steinberg J, Westbrook CK, Pitz WJ. The autoignition of iso-cetane at high to moderate temperatures and elevated pressures: shock tubeexperiments and kinetic modeling. Combust Flame 2009;156:2165–72.

[8] Wang H, Warner SJ, Oehlschlaeger MA. An experimental and kinetic modelingstudy of the autoignition of a-methylnaphthalene/air and a-methylnaphthalene/n-decane/air mixtures at elevated pressures. CombustFlame 2010;157:1976–88.

[9] Pitz WJ, Mueller CJ. Recent progress in the development of diesel surrogatefuels. Prog Energy Combust Sci 2011;37:330–50.

[10] Battin-Leclerc F, Blurock E, Bounaceur R, Fournet R, Glaude PA, Herbinet O,et al. Towards cleaner combustion engines through groundbreaking detailedchemical kinetic models. Chem Soc Rev 2011;9:4762–82.

[11] Haylett DR, Davidson DF, Hanson RK. Ignition delay times of low-vapor-pressure fuels measured using an aerosol shock tube. Combust Flame2012;159:552–61.

[12] Penyazkov OG, Sevrouk KL, Tangirala V, Joshi N. Autoignitions of diesel fuel/airmixtures behind reflected shock waves. in: Proceedings of the 4th Europeancombustion meeting, Vienna, Austria; 2009.

[13] U.S. Navy. Great Green Fleet <http://greenfleet.dodlive.mil/energy/great-green-fleet/>.

[14] Corporan E, Edwards T, Shafer L, DeWitt MJ, Klingshirn C, Zabarnick S, et al.Chemical, thermal stability, seal swell, and emission studies of alternative jetfuels. Energy Fuels 2011;25:955–66.

[15] Allen C, Valco D, Toulson E, Yoo JH, Lee T. JP-5 and HRJ-5 autoignitioncharacteristics and surrogate modeling. Energy Fuels 2013;27:7790–9.

[16] Wang H, Oehlschlaeger MA. Autoignition studies of conventional and Fischer-Tropsch jet fuels. Fuel 2012;98:249–58.

[17] Shen HPS, Oehlschlaeger MA. The autoignition of C8H10 aromatics at moderatetemperatures and elevated pressures. Combust Flame 2009;156:1053–62.

[18] Maxwell JB, Bonnell LS. Derivation of precision of a new vapor pressurecorrelation for petroleum hydrocarbons. Ind Eng Chem 1957;49:1187–96.

[19] Morris RE, Begue NJ. Compositional comparison of algae derived diesel fuels totheir petroleum derived counterparts. Naval Research Laboratory letter 6180/0128; 2011.

[20] Vasu SS, Davidson DF, Hanson RK. Jet fuel ignition delay times: shock tubeexperiments over wide conditions and surrogate model predictions. CombustFlame 2008;152:125–43.

[21] Wang W, Oehlschlaeger MA. A shock tube study of methyl decanoateautoignition at elevated pressures. Combust Flame 2012;159:476–81.

[22] Wang W, Oehlschlaeger MA. The high-temperature autoignition of biodieselsand biodiesel components. Combust Flame 2014, in press.

[23] Campbell MF, Davidson DF, Hanson RK. Ignition delay times of very-low-vapor-pressure biodiesel surrogates behind reflected shock waves. Fuel 2014;126: 271–281.

[24] Westbrook CK, Pitz WJ, Herbinet O, Curran HJ, Silke EJ. A comprehensivedetailed chemical kinetic reaction mechanism for combustion of n-alkanehydrocarbons from n-octane to n-hexadecane. Combust Flame2009;156:181–99.

S. Gowdagiri et al. / Fuel 128 (2014) 21–29 29

[25] Valco D, Gentz G, Allen C, Colket M, Edwards T, Gowdagiri S, OehlschlaegerMA, Toulson E, Lee T. Autoignition behavior of synthetic alternative jet fuels:an examination of chemical composition effects on ignition delays at low tointermediate temperatures. Proc Combust Inst, submitted for publication.

[26] Dooley S, Won SH, Chaos M, Heyne J, Ju Y, Dryer FL, et al. A jet fuel surrogateformulated by real fuel properties. Combust Flame 2010;157:2333–9.

[27] Dooley S, Won SH, Heyne J, Farouk TI, Ju Y, Dryer FL, et al. The experimentalevaluation of a methodology for surrogate fuel formulation to emulate gasphase combustion kinetic phenomena. Combust Flame 2012;159:1444–66.

[28] Won SH, Dooley S, Veloo PS, Wang H, Oehlschlaeger MA, Dryer FL, et al. Thecombustion properties of 2,6,10-trimethyl dodecane and a chemical functionalgroup analysis. Combust Flame 2014;161:826–34.

[29] Weisser GA. Modeling of combustion and nitric oxide formation for medium-speed DI diesel engines: a comparative evaluation of zero- and there-dimensional approaches. In: Diss. ETH no. 14465, LAV, ETH Zurch; 2001.

[30] Vandersickel A, Hartmann M, Vogel K, Wright YM, Fikri M, Starke R, et al. Theautoignition of practical fuels at HCCI conditions: high-pressure shock tubeexperiments and phenomenological modeling. Fuel 2012;93:492–501.