ABSTRACT
The influence of methanol and butanol on soot formation during the pyrolysis of a toluene primary reference fuel mixture with a research octane number (RON) of 91 (TPRF91) was investigated by conducting shock-tube experiments. The TPRF91 mixture contained 17 mol % n-heptane, 29 mol % iso-octane, and 54 mol % toluene. To assess the contribution of individual fuel compounds on soot formation during TPRF91 pyrolysis, the pyrolysis of argon diluted (1) toluene, (2) iso-octane, and (3) n-heptane mixtures were also studied. To enable the interpretation of the TPRF91 + methanol and TPRF91 + butanol experiments, the influence of both alcohols on soot formation during the thermal decomposition of toluene and iso-octane was also investigated in a separate series of measurements. Pyrolysis was monitored behind reflected shock waves at pressures between 2.1 and 4.2 bar and in the temperature range of 2060-2815 K. Laser extinction at 633 nm was used to determine the soot yield as a function of reaction time. For selected experiments, the temporal variation in temperature was also measured via time-resolved two-color CO absorption using two quantum-cascade lasers at 4.73 and 4.56 µm. It was found that soot formed during TPRF91 pyrolysis is primarily caused by the thermal decomposition of toluene. Adding methanol to TPRF91 results in a slight reduction of soot formation, whereas admixing butanol results in shifting soot formation to higher temperatures, but in that case, no overall soot reduction was observed during TPRF91 pyrolysis. Measured soot yields were compared to simulations based on a previous and an updated version of a detailed reaction mechanism from the CRECK modeling group [Nobili, A.; Cuoci, A.; Pejpichestakul, W.; Pelucchi, M.; Cavallotti, C.; Faravelli, T. Combust. Flame 2022; 10.1016/j.combustflame.2022.112073]. Rate-of-production analyses for reactions involving BINS at different experimental conditions were carried out. Although in the case of TPRF91 and toluene pyrolysis, no quantitative agreement was obtained between the experiment and simulation, the comparison nevertheless shows that the new version of the CRECK mechanism is a significant improvement over the previous one. In the case of n-heptane decomposition and iso-octane pyrolysis with and without alcohols, the updated reaction mechanism shows excellent agreement between simulation and measured soot yields.
ABSTRACT
Shock-tube experiments have been performed to investigate the thermal decomposition of the oxygenated hydrocarbon dimethoxymethane (DMM; CH3OCH2OCH3). The primary initial reaction channels of DMM decomposition are considered to be the two bond fissions: CH3OCH2OCH3 â CH3O + CH2OCH3 (1) and CH3OCH2OCH3 â CH3 + OCH2OCH3 (2). In the present work, two shock-tube facilities and three different detection techniques have been combined: Behind reflected shock waves, we have carried out time-resolved measurements of (i) the formation of H atoms using the highly sensitive H-ARAS (Atomic Resonance Absorption Spectrometry) technique and (ii) the depletion of the DMM reactant by high-repetition-rate time-of-flight mass spectrometry (HRR-TOF-MS). In addition, (iii) the temperature-dependent composition of stable reaction products was measured in single-pulse shock-tube experiments via gas chromatography (GC/MS). The experiments span a temperature range of 1100-1430 K, a pressure range of 1.2-2.5 bar, and initial reactant mole fractions from 0.5 ppm (for H-ARAS experiments) up to 10â¯000 ppm (for HRR-TOF-MS experiments). Experimental rate constants ktotal, ktotal = k1 + k2, obtained from these three completely different methods were in excellent agreement among each other, i.e., deviations are within ±30-40%, and they can be well represented by the Arrhenius expression ktotal( T) = 1013.28±0.27 exp(-247.90 ± 6.36 kJ mol-1/ RT) s-1 (valid over the 1100-1400 K temperature and the 1.2-2.5 bar pressure range). By replacing the respective ktotal values used in a recently published DMM chemical kinetics combustion mechanism (Vermeire et al. Combust. Flame 2018, 190, 270-283), it was also possible to successfully reproduce measured product distributions.
