intra_H_migration hardcoded to force ROO -> QOOH as forward direction #259
Description
Currently the RMG database is hardcoded such that ROO -> QOOH reactions always have their kinetics estimated in the direction given. This is unusual because many of these reactions are actually endothermic, which is inconsistent with our general approach of preferring rate estimates in the exothermic direction. However, it seems that "the rate estimates for migration from O-H to C. are worse than from C-H to O. (i.e. entirely missing and estimated by inappopriate averaging)" (c/o comment from @rwest at L474 of ReactionTemplate.java). It would be better to put improved estimates in the database rather than hardcode this specific change.
This is not new, but I'm adding the issue as a reminder. This is not currently implemented in RMG-Py, which is one big reason why models involving low-temperature peroxy chemistry are different between the two.