Homogeneous gas phase kinetic experiments at high pressures and intermediate temperatures are critical for furthering an understanding of the autoignition chemistry of spark ignition engine fuels. In order to perform such experiments, a new variable pressure flow reactor has been developed. This device can be used to study homogeneous gas phase chemical kinetics over the pressure range from 1 to 28 atmospheres and the temperature range from 700 to 1300 K. Initial validation experiments were performed at pressures from 1 to 9 atmospheres using stoichiometric hydrogen and oxygen mixtures. Post-induction chemical kinetic results compare favorably with earlier atmospheric pressure flow reactor data and with numerical results predicted using a recently published comprehensive kinetic CO/H2/O2 mechanism.
Near stoichiometric n-butane and oxygen flow reactor experiments were performed at pressures of 3 atm to 8 atm and with initial temperatures from 708 K to 945 K. Results were compared with predictions generated from a comprehensive kinetic mechanism which has been used extensively for predicting and interpreting n-butane autoignition phenomena. Relative to the experiments, the model predicts autoignition time scales that are too short at lower temperatures/higher pressures and too long at higher temperatures/lower pressures. The present model is therefore likely to over-emphasize the importance of lower temperature kinetic processes on autoignition phenomena. The mechanism fails to reproduce some of the high temperature atmospheric pressure flow reactor data originally utilized in the initial model development, the hydrogen/oxygen data discussed above, and other flow reactor results on the oxidation of methane at pressures of 3, 6 and 9 atmospheres.
Sources of disagreement are multifaceted and cannot be simply repaired. Comprehensive, hierarchical redevelopment of a mechanism for n-butane oxidation, validated over an appropriate range of pressures as well as temperatures, is needed. Additional experimental data, similar to that presented here, will be required in order to generate and validate a more realistic, accurate comprehensive mechanism for the pressure and temperature ranges relevant to engine autoignition chemistry. While present models may reasonably predict autoignition times, model improvements will be required to quantitatively assess the relative affects of various chemical kinetic phenomena on autoignition behavior.