[DFTB-Plus-User] Correction to dispUFF.f90

[Sharma SRK Chaitanya Yamijala]

Dear Developers,

Could you please help us to understand whether the geometry optimization
will be affected by the following changes that you made to the dispuff.f90
in the 18.2 version (On the GitHub issues page, you mentioned that this
problem was there from v1.3)?

-call matinv(invRecVecs)
+call matinv(recVecs)

Specifically, since the changes are only in the subroutine named
UpdateLatVecs, I want to know whether the optimization will be affected
even when the cell parameters are fixed or not?

*Background:*
We are currently running a periodic Mg-surface interfaced with different
solvents (containing C, H, N, O, F species) using DFTB-17.1 along with the
3ob parameter set and UFF dispersion correction. We find that the
optimization results from the DFTB are completely different from that of
the DFT-PBE results, where we observed the dissociation of certain solvents
with DFTB. While trying to understand the reason for these differences,
yesterday, we noticed this fix in the dispUFF.f90 routine on the Github
issues page. Although we are re-running our systems with v19.1 to know the
effect of this fix, it would be really helpful for us if you can tell us
your experience with this fix.

Thanks,
Sharma.

--------------------------------------------------------------------
SRK *Chaitanya* Sharma, Yamijala.
https://scholar.google.com/citations?hl=en&user=IglNkREAAAAJ
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://mailman.zfn.uni-bremen.de/pipermail/dftb-plus-user/attachments/20200520/c882fc49/attachment.htm>