Experimental measurements of n-heptane flame speeds behind reflected shock waves with variable extents of pre-flame auto-ignition chemistry

The flame speeds of premixed stoichiometric n-heptane/21% O₂-79% Ar (so-called “airgon”) flames with different extents of pre-flame auto-ignition chemistry were experimentally investigated using an extended test-time, side-wall-imaging shock tube. n-Heptane/airgon mixtures were impulsively heated by reflected shock waves to a temperature of 703 ± 8 K and pressure of 1.55 ± 0.05 atm, exhibiting first-stage auto-ignition at 20.1 ± 0.6 ms. Flames were spark-ignited using a laser from 0.45 ms to 39 ms after reflected-shock heating, thus probing the augmentation of flame speeds with variable extents of pre-flame auto-ignition chemistry. The fixed-initial-temperature experiments displayed multiple distinctive flame speed regimes across first-stage auto-ignition. A burned-gas flame speed ($S_b^0$) increase of 6% was first observed with spark-ignition at $\sim$7 ms after reflected-shock heating. At spark-ignition timing very close to the first-stage auto-ignition, $S_b^0$ displayed a sharp 24% rise, which then gradually declined to $\sim$10% above the reference value where pre-flame chemistry is negligible. Experiments were additionally performed at 1.55 ± 0.05 atm using two fixed spark-ignition timings (0.45 ms and 25 ms after reflected-shock heating) for initial temperatures between 641 K and 771 K. In the experiments with variable initial temperatures, a non-monotonic $S_b^0$ dependence on initial temperature was observed for the 25-ms experiments, which showed a maximum $S_b^0$ increase of 14% relative to the 0.45-ms experiments. To provide modeling comparisons, the thermochemical time history of the reacting gas was first simulated; the species profiles and the temperature at a given residence time were then used to obtain the flame speed from 1D steady-state simulations. The multi-regime flame speed behavior was not observed in fixed-initial-temperature simulations, which predicted a single rise in flame speed only near the first-stage auto-ignition time. The simulations with variable initial temperatures qualitatively recovered the non-monotonic flame speed trend, but generally showed underprediction of flame speeds relative to experimental results. These new experiments provide insight into the effect of pre-flame chemistry on flame propagation and offer targets for improving flame modeling, potentially aiding the development of next-generation engine concepts.