Vibrational relaxation, incubation times, and unimolecular dissociation of C4H4O have been investigated over the extended temperature range 500−3000 K in 2−5% furan−krypton mixtures, 2% furan−neon mixtures, and in pure furan. The experiments were performed in shock waves using laser-schlieren (LS) densitometry and time-of-flight (TOF) mass spectrometry. At low temperatures and low pressures, only vibrational relaxation was observed using the LS technique. This relaxation is unexpectedly slow and shows a strong nonexponential time dependence. Unimolecular dissociation is observed in TOF experiments between 1300 and 1700 K in a pressure range of 175−250 Torr as well as LS experiments between 1700 and 3000 K for pressures between 100 and 600 Torr. The TOF experiments show that under the given conditions two molecular dissociation channels leading to C2H2 + CH2CO or to C3H4 + CO are dominant. The branching ratio between these channels has been determined between 1300 and 1700 K. At low temperatures, the molecular channel leading to C3H4 and CO is preferred, but a channel switching was observed around 1700 K. The domination of these molecular channels is consistent with the shape of the LS profiles, and these have been successfully modeled with just these two reactions. The overall unimolecular rate constant is in the falloff regime close to the low-pressure limit. By use of statistical reaction rate theory, the total unimolecular rate constant could be modeled over an extended temperature and pressure range using a value of 〈ΔE〉all = 50 cm-1 for the furan dissociation. In a small range of conditions at low pressures and high temperatures, both the vibrational relaxation and dissociation were resolved and incubation times estimated.

We report the observation of highly nonlinear vibrational relaxation for a number of large molecules in shock waves, together with an attempt at a master-equation modeling of this phenomenon. In all these molecules laser-schlieren measurements show a clear and often well-resolved local maximum in the density gradient, indicating a similar maximum in the rate of energy transfer. This unexpected phenomenon is seen in the relaxation of benzene (C6H6), cubane (C8H8), cyclopropane (C3H6), furan (C4H4O), norbornadiene (C7H8), oxirane (C2H4O), and quadricyclane (C7H8). It has also been detected in cyclopentadiene (C5H6) and pyrazine (C4N2H4), as well as CF3Br and CF3Cl but in these was not well resolved. The phenomenon thus seems nearly ubiquitous; of the “large” molecules where relaxation could be resolved, only norbornene (at C7H10 the largest such molecule) exhibits a fully linear relaxation. The gradients are clearly and solely from vibrational relaxation; integrated gradients are in good agreement with thermodynamic calculations of total density change, and near-equilibrium relaxation times in pure cyclopropane and oxirane are fully consistent with overlapping ultrasonic results. It appears we are seeing a delay in the development of series coupling in these experiments. An attempt is made to model the process using a linear master equation with exponential gap probabilities having an α(∼〈ΔE〉down) linear in energy. Although this does introduce sufficient nonlinearity through the rate coefficients to produce maxima, these and the overall gradients are consistently too small. It is suggested that inclusion of a true nonlinearity through VV transfer will be needed to explain the observations, and a possible mechanism for this is proposed.

No abstract