Oxidation Kinetics of Pure and Blended Methyl Octanoate/n-Nonane/Methylcyclohexane: Measurements and Modeling of OH*/CH* Chemiluminescence, Ignition Delay Times and Laminar Flame Speeds

The focus of the present work is on the empirical characterization and modeling of ignition trends of ternary blends of three distinct hydrocarbon classes, namely a methyl ester (C9H18O2), a linear alkane (n-C9H20), and a cycloalkane (MCH). Numerous surrogate biofuel formulations have been proposed in the literature, yet specific blending of these species has not been studied. Moreover, the effects of blending biofuel compounds with conventional hydrocarbons are not widely studied and a further point is the lack of studies paying specific attention to the effects of fuel variation within a given blended biofuel. To this end, a statistical Design of Experiments L9 array, comprised of 4 parameters (%MO, %MCH, pressure, and equivalence ratio) with 3 levels of variation, constructed in order to systematically study the effects of relative fuel concentrations within the ternary blend enabled variations in fuel concentration for methyl octanoate and MCH of 10% - 30% and 20% - 40%, respectively. Variation in pressure of 1 atm, 5 atm, and 10 atm and in equivalence ratio of 0.5, 1.0, and 2.0 were used, respectively. The fuel-volume percentage of n-nonane varied from 30% - 70%. In total, 10 ternary blends were studied.

Ignition delay times for the ternary blends and for the three constituents were obtained by monitoring excited-state OH or CH transitions, A2Epsilon+ -> X2Pi or A2Delta -> X2Pi, respectively, behind reflected shock waves using a heated shock tube facility. Dilute conditions of 99% Ar (vol.) were maintained in all shock tube experiments with the exception of a separate series of n-nonane and MCH experiments under stoichiometric conditions which used 4% oxygen (corresponding to ~ 95% Ar dilution). Temperatures behind reflected shock waves were varied over the range 1243 < T (K) < 1672. From over 450 shock tube experiments, empirical ignition delay time correlations were constructed for all three pure fuels and a master correlation equation for the blended fuels. Ignition experiments conducted on the pure fuels at 1.5 atm indicated the following ignition delay time order, from shortest to longest: methyl octanoate < n-nonane < MCH. With increased pressure to 10 atm (nominal) the order remained, in general, consistent. Under fuel-lean conditions, ignition trends between methyl octanoate and n-nonane exhibited overlap at temperatures below 1350 K, below which the trends diverged with methyl octanoate having shorter ignition delay times. Similar behavior was observed under fuel-rich conditions, yet with the overlap occurring above 1450 K. Stoichiometric ignition trends did not display overlapping behavior under either 1.5 atm or 10 atm pressure. Laminar flame speed measurements were performed at 1 atm and an initial temperature of 443 K on the pure fuel constituents. Additional flame speed measurements of MCH were conducted at 403 K to compare with literature values and were shown to agree strongly with experiments conducted in a constant-volume apparatus. The experiments conducted herein, for the first time, measure laminar flame speeds methyl octanoate.

A detailed chemical kinetics mechanism was compiled from three independent, well-validated models for the constituent fuels, where the sub-mechanisms for methyl octanoate and MCH were extracted for integration into a base n-nonane model. The compiled mechanism in the present study (4785 reactions and 1082 species) enables modeling of oxidation processes of the ternary fuel blends of interest. Calculations were performed using the compiled model relative to the base models to assess the impact of utilizing different base chemistry sets. In general, results were reproduced well relative to base models for both n-nonane and MCH, however results for methyl octanoate from both the compiled model and the base model are in disagreement with the results measured herein. Ignition delay times of the fuel blends are well-predicted for several conditions, specifically for blends at lean/high-pressure and stoichiometric/high-pressure conditions, however are not accurately modeled at fuel-rich, high-pressure conditions.