make adsorptionPt111 more flexible with lone pairs px

[bjkreitz]

trafficstars

Description

The nodes in the adsorption correction file adsorptionPt111.py were rather restrictive. In most nodes it was specified that the generalR cannot have lone pairs p0. I didn't pay close attention to this when I restructured the adsorption correction database because the R was mostly C in the training data. However, this is not correct. In most cases, the R can also be replaced with O/N. Actually, in some cases the training data that I used contained O, but the node would not have been used to estimate the thermo for this species. @kirkbadger18 caught this detail.

I revised now the adjacency lists and made the nodes more permissive, by changing the p0 to px whenever it is reasonable. I did not touch the nodes containing nitrogen, because @kirkbadger18 is doing a complete overhaul of the nitrogen adsorption corrections.

Testing

I generated a mechanism for the oxidation of ethane on Pt(111) without (left log file) and with (right log file) these changes. The mechanisms are indeed slightly different.

With the more permissible adsorption corrections, RMG deems a bidentate carboxyl to be important

Previously, this species was discovered, but remained in the edge.

You see that the thermochemistry of the species has changed because RMG used the top level node C*O* to estimate the thermo and now it uses the more specific C=*RO-*. Whether this species should exist is another story (because there is also a monodentate carboxyl). However, I think that these updates to the adsorption corrections are necessary since they can affect our thermo estimate for some species and make them (supposedly) more accurate.

Review Carefully review that I haven't missed a node or an R. @kirkbadger18 will make another PR soon that changes the N-containing nodes. Generate a mechanism for system of your choice and check if everything works accordingly.