GFN-xTB organosilicon interactions

[fabbe...@gmail.com]

Hello CP2k community,

I am running some QS simulations using XTB on a sample containing Si, O, C, and H. Following some other threads here I have managed to get the system running reasonably well. However the H closest to Si gain way to much kinetic energy after the 1st time step.

Having found this paper (https://arxiv.org/pdf/2109.10416.pdf) I am wondering if anyone else here has had a problem getting XTB to work for organosilicon systems.

Best regards

Fabian

[Anna Hehn]

Dear Fabian

we did not investigate organosilicon systems in particular, but we found that for periodic framework materials in general, the default parameters of GFN1-xTB may in certain cases lead to instabilities. The keyword CHECK_ATOMIC_CHARGES in the xTB section is actually checking for unphysical Mulliken charges; if the keyword is set to true, CP2K stops if the self-consistently obtained Mulliken charges are significantly different from the expected, physically reasonable value.

Best regards

Anna

--
You received this message because you are subscribed to the Google Groups "cp2k" group.
To unsubscribe from this group and stop receiving emails from it, send an email to cp2k+uns...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/cp2k/dd901ca8-e715-4836-9f46-05d7eb6988d4n%40googlegroups.com.