Insight into selected reactions in low-temperature dimethyl ether combustion from Born-Oppenheimer molecular dynamics

Dimethyl ether is under consideration as an alternative diesel fuel. Its combustion chemistry is as yet ill-characterized. Here we use Born-Oppenheimer molecular dynamics (BOMD) based on DFT-B3LYP forces to investigate the short-time dynamics of selected features of the low-temperature dimethyl ether (DME) oxidation potential energy surface. Along the chain propagation pathway, we run BOMD simulations from the transition state involving the decomposition of (*)CH(2)OCH(2)OOH to two CH(2)=O and an (*)OH radical. We predict that formaldehyde C-O stretch overtones are excited, consistent with laser photolysis experiments. We also predict that O-H overtones are excited for the (*)OH formed from (*)CH(2)OCH(2)OOH dissociation. We also investigate short-time dynamics involved in chain branching. First, we examine the isomerization transition state of (*)OOCH(2)OCH(2)OOH --> HOOCH(2)OCHOOH. The latter species is predicted to be a short-lived metastable radical that decomposes within 500 fs to hydroperoxymethyl formate (HPMF; HOOCH(2)OC(=O)H) and the first (*)OH of chain branching. The dissociation of HOOCH(2)OCHOOH exhibits non-RRKM behavior in its lifetime profile, which may be due to conformational constraints or slow intramolecular vibrational energy transfer (IVR) from the nascent H-O bond to the opposite end of the radical, where O-O scission occurs to form HPMF and (*)OH. In a few trajectories, we see HOOCH(2)OCHOOH recross back to (*)OOCH(2)OCH(2)OOH because the isomerization is endothermic, with only an 8 kcal/mol barrier to recrossing. Therefore, some inhibition of chain-branching may be due to recrossing. Second, trajectories run from the transition state leading to the direct decomposition of HPMF (an important source of the second (*)OH radical in chain branching) to HCO, (*)OH, and HC(=O)OH show that these products can recombine to form many other possible products. These products include CH(2)OO + HC(=O)OH, H(2)O + CO + HC(=O)OH, HC(=O)OH + HC(=O)OH, and HC(=O)C(=O)H + H(2)O, which (save CH(2)OO + HC(=O)OH) are all more thermodynamically stable than the original HCO + (*)OH + HC(=O)OH products. Moreover, the multitude of extra products suggest that standard statistical rate theories cannot completely describe the reaction kinetics of significantly oxygenated compounds such as HPMF. These secondary products consume the second (*)OH required for explosive combustion, suggesting an inhibition of DME fuel combustion is likely.