Incubation times, fall-off and branching ratios in the thermal decomposition of toluene: Experiments and theory

We have investigated the thermal unimolecular decomposition of toluene in a low-pressure shock tube (0.1 bar < p < 2 bar) in the temperature range 1350 K to 1900 K. The decomposition of toluene proceeds in two competing parallel reactions C(6)H(5)CH(3)+M-->C(6)H(5)CH(2)+H+M (R1) and C(6)H(5)CH(3)+M-->C(6)H(5) + CH(3) + M (R2). Reaction (R1) generally dominates under all conditions. The rate constant k(1) has been determined using H-atom detection via calibrated atomic resonance absorption spectroscopy (ARAS) at 121.6 nm. At the highest temperatures a slight pressure dependence of k(1) was noticed while at lower temperatures k(1) was found to be almost pressure independent. At the lowest pressures and highest temperatures we additionally identified significant induction times. We analyzed induction times, fall-off, and rate constants with a two-channel master equation model. Collisional energy transfer probabilities for toluene/Ar were taken as determined by Luther and co-workers (ref. 1, U. Hold, T. Lenzer, K. Luther, K. Reihs and A. C. Symonds, Ber. Bunsen-Ges. Phys. Chem., 1997, 101, 552; T. Lenzer, K. Luther and K. Reihs, J. Chem. Phys., 2000, 112, 4090) and specific rate constants from simplified SACM calculations, which have been confirmed by experiments (ref. 2, U. Brand, H. Hippler, L. Lindemann and J. Troe, J. Phys. Chem., 1990, 94, 6305; ref. 3, H. Hippler, Ch. Riehn, J. Troe and K. - M. Weitzel, J. Phys. Chem., 1990, 94, 6321). The numerical solution of the master equation was obtained considering a maximum energy of 80 000 cm(-1) and a matrix dimension of 1000. The lower 500 energy levels have been considered as discrete levels taking into account the molecular specific structure in the density of states. The unimolecular pressure dependent rate constants k(1) (M) and k(2) (M) were determined from the eigenvalues of the system. The induction time was identified as the delay obtained from back extrapolation of the stationary reaction rate. The agreement between experimental and modeled pressure and temperature dependent rate constants k(1) and k(2) was excellent. The experimental incubation times were predicted within a factor of three indicating three times slower energy transfer rates at the bottleneck than used in the model.