Some entries in CHO thermo library are severely wrong

[cgrambow]

trafficstars

I have encountered several entries in the CHO thermo library, which have incorrect heats of formation. Unfortunately, I didn't think of cataloging all the incorrect species as I encountered them so I just have two examples right now. I think there may also be incorrect values in the other CH* libraries. I discovered that quite a lot of values in literature are incorrect, presumably because of human error when copying over data. This is mostly the case for large collections of thermodynamic data.

The two examples are:

• 2-methoxyethanol (link): CHO value is -103.7 kcal/mol, whereas GA is -89.7. I also have a F12/DZ calculation that agrees exactly with GA. The actual experimental value is -90.1 +/- 1.9 (see here)
• 2,5-dihydrofuran (link): CHO: -26.0, GA: -15.7, F12/DZ: -15.8, expt: -16.7 (see here). Interestingly, the values in JetSurf1.0/2.0, CurranPentane, and USC-Mech-ii are also incorrect.

I expect GA to perform relatively well for the CHO library, so it may be worth comparing the entries to the GA predictions, flagging those with severe disagreement, and manually correcting them.

[mjohnson541]

GAV might have some difficulty with some of the more highly oxygenated species in CHO, but otherwise this sounds worthwhile, it may be best to just remove the entries from CHO and add any better values to a more trustworthy library since I don't believe we have individual sources for CHO entries.