[CP2K-user] About TDDFT: version bugs and functionals availability

Lucas Lodeiro eluni... at gmail.com
Fri Feb 12 01:26:47 UTC 2021


Hi all,

I want to perform some TDDFT calculations for periodic systems. I study the
regtest examples and read all forum topics related to TDDFT... but some
doubts remain.

I understood from the forum that:
* TDDFT is only available for single points calculations with GPW for RKS
(not optimizations or gradients). There are some problems with 2° and 3°
derivatives in the HF module in 7.x versions when truncated coulomb
interaction is used. Also ADMM has problems with EXCH_CORRECTION_FUNC, only
PBEX or BECKE88 work (
https://groups.google.com/g/cp2k/c/SCHS1-VcFx4/m/Bc_tIIzHAQAJ and
https://groups.google.com/g/cp2k/c/VLWizWNCZ1Y/m/hF9x4-W7AQAJ and
https://groups.google.com/g/cp2k/c/_b4cRs7HxZE/m/1aJa0FjjAQAJ).
* Some people do some tests (without mention which version they used) and
saw that no hybrid functional works, full coulombic, hybrid metaGGA or
range-separated functionals... (
https://groups.google.com/g/cp2k/c/PcPUy6FouGE/m/CKOUiGxUDAAJ and
https://groups.google.com/g/cp2k/c/TnIuSucqsMg/m/s14nG6DrAgAJ)

Are all of these correct? And, which versions are affected? Just 7.x are
affected?

Other questions are:
* Double Hybrid functional are available to calculate TDDFT?
* Is it necessary to re-define the XC subsection into the TDDFT section?
Or, is it sufficient with the XC subsection in the DFT section?

I comment and ask all of these things, because I want to test a new
range-separate double hybrid functional in a crystalline periodic system...
which is a little bit tangled to test.

Regards
Lucas Lodeiro
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20210211/4a6cfe2a/attachment.htm>


More information about the CP2K-user mailing list