<div dir="ltr"><br>PM6D3H+ was incorporated into GAMESS, see<br>
<br>
Kromann JC, Christensen AS, Steinmann C, Korth M, Jensen JH. (2014) A<br>
third-generation dispersion and third-generation hydrogen bonding<br>
corrected PM6 method: PM6-D3H+. PeerJ 2:e449<br>
<a href="https://www.google.com/url?q=https%3A%2F%2Fdoi.org%2F10.7717%2Fpeerj.449&sa=D&sntz=1&usg=AFQjCNG2BiLJy3PdPgDrZ43gGf4-lbK17A" rel="noreferrer" target="_blank">https://doi.org/10.7717/peerj.<wbr>449</a><br>
<br>
and <a href="http://www.google.com/url?q=http%3A%2F%2Fproteinsandwavefunctions.blogspot.com%2F2014%2F09%2Fpm6-dftb-gamess-enzyme-design-protein.html&sa=D&sntz=1&usg=AFQjCNFFI9oYsik4tCwFLTDl9EuxUNJX2w" rel="noreferrer" target="_blank">http://<wbr>proteinsandwavefunctions.<wbr>blogspot.com/2014/09/pm6-dftb-<wbr>gamess-enzyme-design-protein.<wbr>html</a><br>
...the linked paper tabulates accuracies on the S22 and S66 databases<br>
of about 0.6 kcal/mol MAD, 2.11 kcal/mol max. error...<br>
<br>but because GAMESS lacks general periodic and slab capability as well as some MD capabilities, it would be useful to use this in cp2k.<br><br>
The code is GPLv2 at <a href="https://www.google.com/url?q=https%3A%2F%2Fgithub.com%2Fjensengroup%2Fhydrogen-bond-correction-f3&sa=D&sntz=1&usg=AFQjCNEvS6PddSSjUb-PBW5U7YICwY4XoA" rel="noreferrer" target="_blank">https://github.com/<wbr>jensengroup/hydrogen-bond-<wbr>correction-f3</a><br>
(and gives energies and gradients in the standalone.)<br><br>
In addition, the accuracy of the related PM6-DH+ on the S22 and S66<br>
databases is listed in tables 2 & 3 of<br>
<a href="https://www.google.com/url?q=https%3A%2F%2Fwww.chem.purdue.edu%2Fslipchenko%2Fpubs%2Fpdf%2F2012_EFP_S22.pdf&sa=D&sntz=1&usg=AFQjCNHMm2MQNMGh4h700BiN5s2yZs-vzg" rel="noreferrer" target="_blank">https://www.chem.purdue.edu/<wbr>slipchenko/pubs/pdf/2012_EFP_<wbr>S22.pdf</a><br>
is compared to higher levels of theory and performs well.<br>
<br>
(Also, as I surmised, the DFT-D3 corrections under XC don't work with SE, so...)<br>
<br>
Is there any chance of including/interfacing this in cp2k?<br>
<br>
<br>
Similar corrections are included in PM7 and PM7-TS, so this would likely be a<br>
step to implementing these....<br>
<br>
PM7 and PM7-TS<br>
<a href="http://www.google.com/url?q=http%3A%2F%2Flink.springer.com%2Farticle%2F10.1007%2Fs00894-012-1667-x%2Ffulltext.html&sa=D&sntz=1&usg=AFQjCNHbCHbH9N77q6ZpTVYKHj4Rc7eo1A" rel="noreferrer" target="_blank">http://link.springer.com/<wbr>article/10.1007/s00894-012-<wbr>1667-x/fulltext.html</a><br>
(accuracies of PM6-DH+ and PM6-DH2 also tabulated here)<br>
Supplementary material has parameter listings for both.<br>
<br>The transition-state parameterization/capability would also be useful to have.<br><br><br></div>