Couldn't find the rate constant for reaction __ with H_Abstraction

[rwest]

There are a lot of lines like:

ERROR: Couldn't find the rate constant for reaction: SPC(54898)+OH(32)=SPC(54265)+O(34) with H_Abstraction
ERROR: Couldn't find the rate constant for reaction: SPC(54904)+OH(32)=SPC(54265)+O(34) with H_Abstraction

in Rajesh's log file in /home/rajesh/Rajesh/New_jobs/MultiT_pdep_Prun_15000_single_concentration They all (?) have CH3 or OH as a reactant and CH2 or O as a product.

[mrharper]

If @rwest is correct about each of the reactions involving CH3 or OH, the problem could be:

For the H_Abstraction rxn family, RMG checks if any of the species is a biradical. If so, RMG will force that side of the reaction to be the "reactants". So, it appears RMG may be having problems finding a rate coefficient for abstraction from X (anything) by a radical with a birad center (CH2, O). I've assigned myself to this issue.

[mrharper]

I misspoke before. RMG no longer has the rule of "if this is an H_abstraction reaction and one of the species is a biradical, make it the reactants" RMG will always grab kinetics for H_abstraction reactions in the exothermic direction.

Anyway, based on some preliminary debugging, it appears this "ERROR: Couldn't find the rate constant ..." always occurs for radical+radical H_abstraction reactions. In particular, if the abstraction is from a single heavy-atom radical (e.g. OH, CH3, as @rwest mentioned before).

I will add new rate rules to the library to see if this resolves the issue.

[rwest]

Presumably we're not even matching the top level node, otherwise we'd at least match that (even if the kinetics are hopeless)? ...and yet we generate the reaction. Do we generate it in one direction (that's DOES match the top nodes) then look for the kinetics in the reverse (because the direction in which it was generated is endothermic)? ... in which case would "if you can't find the forward kinetics, look for the reverse" be a smart move or a dumb one?

[rwest]

I wil re-run this case

[ajalan]

Quick comment: Populate reactions does not report H_abstraction between O_atom_triplet(O_2T) and methane even though we have a node in the rate library corresponding to X_H and O_atom_triplet. However, we do report an averaged rate constant for (O_2T) and ethane.

C2H6(1) + OJ(molecule/cm3)(2) --> C2H5J(4) + HOJ(6) 5700.000000000001 3.05 3.123 H_Abstraction estimate: (Average:) [ InChI=1/C2H6/c1-2/h1-2H3 , O_atom_triplet ] deltaHrxn(T=298K) = -1.1479859289917655 kcal/mol

So it does seem that we are not matching the top level node , in this case [X_H, O_atom_triplet] as @rwest alluded to earlier. Quite obviously, RMG-Py reports a rate for this reaction which seems to be in fair agreement with other estimates (though most values seem to be coming from 3-4 common sources)

http://rmg.mit.edu/database/kinetics/reaction/reactant1=1%20O%202T;__reactant2=1%20C%200;__product1=1%20O%201;__product2=1%20C%201;