The Leuven isoprene mechanism, proposed earlier to aid in rationalizing the unexpectedly high hydroxyl radical (OH) concentrations in isoprene-rich, low-nitric-oxide (NO) regions (Peeters; et al. Phys. Chem. Chem. Phys. 2009, 11, 5935), is presented in an upgraded and extended version, LIM1. The kinetics of the crucial reactions in the proposed isoprene-peroxy radical interconversion and isomerization pathways are re-evaluated theoretically, on the basis of energy barriers computed at the much higher CCSD(T)/aug-cc-pVTZ//QCISD/6-311G(d,p) level of theory, and using multiconformer partition functions obtained at the M06-2X/6-311++G(3df,2p) level that, different from the B3LYP level used in our earlier work, accounts for the crucial London dispersion effects in the H-bonded systems involved. The steady-state fraction of the specific Z-delta-OH-peroxy radical isomers/conformers that can isomerize by a 1,6-H shift is shown to be largely governed by hydrogen-bond strengths, whereas their isomerization itself is found to occur quasi-exclusively by hydrogen atom tunneling. The isomer-specific Z-delta-OH-peroxy 1,6-H-shift rate coefficients are predicted to be of the order of 1 s(-1) at 298 K, but the experimentally accessible bulk rate coefficients, which have to be clearly distinguished from the former, are 2 orders of magnitude lower due to the very low Z-delta-OH-peroxy steady-state fractions that are only around or below 0.01 at low to moderate NO and depend on the peroxy lifetime. Two pathways subsequent to the peroxy radical 1,6-H shift are identified, the earlier predicted route yielding the photolabile hydropercury-methylbutenals (HPALDs), and a second, about equally important path, to dihydroperoxy-carbonyl peroxy radicals (di-HPCARP). Taking this into account, our predicted bulk peroxy isomerization rate coefficients are about a factor 1.8 higher than the available experimental results for HPALD production (Crounse; et al. Phys. Chem. Chem. Phys. 2011, 13, 13607), which is within the respective uncertainty margins. We also show that the experimental temperature dependence of the HPALD production rates as well as the observed kinetic isotope effect for per-deuterated isoprene support quantitatively our theoretical peroxy interconversion rates. Global modeling implementing LIM1 indicates that on average about 28% of the isoprene peroxys react via the 1,6-H-shift isomerization route, representing 100-150 Tg carbon per year. The fast photolysis of HPALDs we proposed earlier as primary OH regeneration mechanism (Peeters and Muller. Phys. Chem. Chem. Phys. 2010, 12, 14227) found already experimental confirmation (Wolfe; et al. Phys. Chem. Chem. Phys. 2012, 14, 7276); based on further theoretical work in progress, reaction schemes are presented of the oxy coproduct radicals from HPALD photolysis and of the di-HPCARP radicals from the second pathway following peroxy isomerization that are both expected to initiate considerable additional OH recycling.
