The kinetic isotope effect of thereactions OH + CH3OCH3(DME) and OH + CD3OCD3(DME-d6) was experimentallyand theoretically studied.Experiments were carried out in a slow-flow reactor at pressures between5 and 21 bar (helium as bath gas) with production of OH by laser flashphotolysis of HNO3and time-resolved detection of OH bylaser-induced fluorescence. The temperature dependences of the ratecoefficients obtained can be described by the following modified Arrheniusexpressions: kOH+DME= (4.5 ± 1.3)× 10–16(T/K)1.48exp(66.6 K/T) cm3s–1(T= 292–650 K, P= 5.9–20.9bar) and kOH+DME-d6= (7.3 ± 2.2) × 10–23(T/K)3.57exp(759.8 K/T) cm3s–1(T= 387–554 K, P= 13.0–20.4 bar). A pressure dependence of therate coefficients was not observed. The agreement of our experimentalresults for kOH+DMEwith values from otherauthors is very good, and from a fit to all available literature data,we derived the following modified Arrhenius expression, which reproducesthe values obtained in the temperature range T=230–1500 K at pressures between 30 mbar and 21 bar to betterthan within ±20%: kOH+DME= 8.45× 10–18(T/K)2.07exp(262.2 K/T) cm3s–1. For kOH+DME-d6, to the best of our knowledge, this is the first experimental study.For the analysis of the reaction pathway and the kinetic isotope effect,potential energy diagrams were calculated by using three differentquantum chemical methods: (I) CCSD(T)/cc-pV(T,Q)Z//MP2/6-311G(d,p),(II) CCSD(T)/cc-pV(T,Q)Z//CCSD/cc-pVDZ, and (III) CBS-QB3. In all three cases, the reaction is predicted to proceedvia a prereaction OH–ether complex with subsequent intramolecularhydrogen abstraction and dissociation to give the methoxymethyl radicaland water. Overall rate coefficients were calculated by assuming athermal equilibrium between the reactants and the prereaction complexand by calculating the rate coefficients of the hydrogen abstractionstep from canonical transition state theory. The results based onthe molecular data from methods (I) and (II) showed a satisfactory agreement with the experimental values, whichindicates that the pre-equilibrium assumption is reasonable underour conditions. In the case of method (III), the isotopeeffect was significantly underpredicted. The reason for this discrepancywas identified in a fundamentally differing reaction coordinate. Obviously,the B3LYP functional applied in method (III) for geometryand frequency calculations is inadequate to describe such systems,which is in line with earlier findings of other authors. [ABSTRACT FROM AUTHOR]