NO gas missing R*ln(2) in entropy in PrimaryThermoLibrary

[kirkbadger18]

It appears that the species "NOJ" in PrimaryThermoLibrary.py is estimating entropy incorrectly. This was discovered when looking at the rate of NO oxidation to NO2 on Pt(111). We saw that oxidation was occurring beyond equilibrium at high temperature. This we discovered was due to an underestimation of entropy by about 1.4 cal/mol which is approximately R*ln(2), or the entropy from electronic degeneracy of NO. While PrimaryThermoLibrary.py gets it wrong, NOx2018.py gets it right, and stems from the ATcT.

This was discussed in the RMG-Cat development meeting today, and the recommendations from this discussion were to fix the instances where NO is assigned an incorrect NASA7 polynomial. It is recognized that there may be other species that are wrong too, and we should perhaps search for the root cause of this error and other species impacted. The other suggestion is to create a new thermo library using the ATcT NASA7 polynomials that we have. Here is a file with those values from C. Franklin Goldsmith, and before that, from Branko Ruscic I believe:

ATcT_polynomials.yaml

[rwest]

Continuing the discussion from the RMG-cat meeting, about NOj entropy in primaryThermoLibrary being off by R*ln(2), probably due to electronic degeneracy...

Apparently it comes from a calculation Beat ran in 2012, and the pull request (RMG-database #24) was merged in 2013.
Unfortunately the commit message is just "Update thermo library DFT_CCSDTF12_BAC"

The PR description includes:

a thermochemistry library based on CCSD(T)F12A/cc-pVXZ-F12//B3LYP/6-311++g(d,p) calculations corrected by Bond Additivity Correction (BAC) has been added (thermo_DFT_CCSDTF12_BAC.py) containing soon more than 430 molecules and radicals, mostly related to nitrogen

What's our best bet to find how systematic the mistake is?
Will all radicals (with multiplicity of 2) be off by Rln2 in S ?
Does someone have access to @bbuesser's scripts that turned his CCSD(T)F12A/cc-pVXZ-F12//B3LYP/6-311++g(d,p) calculations into the library, back in 2012? Or the calculations themselves?

Can we detect if this occurred in other libraries?

Alon's "N in RMG" paper https://doi.org/10.1002/kin.21154 (with Beat as co-author) says "The thermodynamic parameters of the CH3, NH2, C2H5, and CH2NH2 radicals in the current model are given at the CCSD(T)F12A/cc-pVTZ-F12//B3LYP/6-311++g(d,p) level of theory, taken from an RMG library." and the model in the SI refers to thermo_DFT_CCSDTF12_BAC for NO, among other species.