A Hybrid Density Functional Theory Study of the Low-Temperature Dimethyl Ether Combustion Pathways. I: Chain-Propagation

Dimethyl ether (DME) has been proposed to be a promising alternative to conventional diesel fuel because of its favorable compression ignition property (high cetane number) and its soot-free combustion. A radical chain mechanism for hydrocarbon autoignition has been proposed for DME at low temperatures. In this mechanism, the chain initiation step consists of DME undergoing hydrogen abstraction by a highly reactive species (typically ⋅OH). The CH 3 O ⋅ CH 2 created in the initiation step then combines with O 2 ; the subsequent CH 3 OCH 2 OO⋅ radical is involved in a Lindemann-type mechanism, which can lead to the production of formaldehyde (CH 2 = O) and ⋅OH. This concludes the chain-propagating step: the one ⋅OH produced then sustains the chain-reaction by creating another CH 3 O ⋅ CH 2 . A relatively stable intermediate (⋅CH 2 OCH 2 OOH), formed via isomerization of CH 3 OCH 2 OO⋅ in the chain-propagation step, can combine with a second O 2 to produce a radical (⋅OOCH 2 OCH 2 OOH) that can potentially decompose into two ⋅OH radical (and other products). This path leads to chain-branching and an exponential increase in the rate of DME oxidation. We have used spin-polarized density functional theory with the Becke-3-parameter Lee–Parr–Yang exchange-correlation functional to calculate the structures and energies of key reactants, intermediates, and products involved in (and competing with) the chain-propagating and chain-branching steps of low-temperature DME oxidation. In this article, Part I, we consider only the chain-propagation mechanism and its competing mechanisms for DME combustion. Here, we show that only certain conformers can undergo the isomerization to ⋅CH 2 OCH 2 OOH. A new transition state has been discovered for the disproportionation reaction ⋅CH 2 OCH 2 OOH → 2CH 2 O + ⋅OH in the chain-propagating step of DME autoignition that is much lower than previous barriers. The key to making this decomposition pathway facile is initial cleavage of the O–O rather than the C–O bond. This renders all transition states along the chain-propagation potential energy surface below the CH 3 O ⋅ CH 2 + O 2 reactants. In contrast with the more well-studied CH 3 ⋅ CH 2 (ethyl radical) + O 2 system, the H-transfer isomerization of CH 3 OCH 2 OO⋅ to ⋅CH 2 OCH 2 OOH in low-temperature DME oxidation has a much lower activation energy. This is most likely due to the larger ring strain of the analogous transition state in ethane oxidation, which is a five-membered ring opposed to a six-membered ring in dimethyl ether oxidation. Thus low-temperature ethane oxidation is much less likely to form the ⋅ROOH (where R is a generic group) radicals necessary for chain-branching, which leads to autoignition. Three competing reactions are considered: CH 3 O ⋅ CH 2 → CH 2 O + ⋅CH 3 ; ⋅CH 2 OCH 2 OOH → 1,3dioxetane + ⋅OH; and ⋅CH 2 OCH 2 OOH → ethylene oxide + HOO⋅. The reaction barriers of all these competing paths are much higher in energy (7–10 kcal/mol) than the reactants CH 3 O ⋅ CH 2 + O 2 and, therefore, are unlikely low-temperature paths. Interestingly, an analysis of the highest occupied molecular orbital along the CH 3 O ⋅ CH 2 decomposition path shows that electronically excited (A 2 or A 2 ) CH 2 O can form; this can also be shown for ⋅CH 2 OCH 2 OOH, which forms two formaldehyde molecules. This may explain the luminosity of DME’s low-temperature flames.