topologically unique structures failed for methoxide about crest HOT 4 CLOSED

I'm not sure why the structure is sorted out and will check the routine again. This must not occur, you are totally right.

The strange structure is a result of its lower energy at GFN2/GBSA(H2O) level, things like this can happen because the workflow -in favor of computation time- simply removes each hydrogen atom and reoptimizes the structure. I.e., it is independent from reaction barriers for the dissociation of the proton. Sometimes this leads to strange artifacts.

from crest.

@pprcht Thank you for the reply. I wouldn't worry too much about the fact that GFN2/GBSA(H2O) level couldn't capture the true energy minimum. As it is an empirical method with a crude solvation method.

from crest.

Apparently there was an error in the charge handling for the deprotonation tool and this issue therefore not a result of the topology check. The error led to a neutral calculation instead of the correct anionic one. This seems to pose no practical issue for xtb, but the results were incorrect. Now the methanol deprotonation gives MeO^-, as one would expect. Thanks for finding this!

As far as I can see, this error has been present only in the latest version of the 2.11 development version. Other functionalities (conf. searches) were not affected. I updated the binaries here, please check if the program header has the correct date (today). This is hopefully the last time we update the development binary as we plan to do the actual release of 2.11 within the next week.

from crest.

@pprcht Thank you for fixing this.

from crest.