I'd like to investigate the structural relaxations of the excited state for periodic systems. BSE and TDDFT are two promising methods for this purpose. While there is a discussion on forces with BSE here: Is there a software that has implemented forces of BSE calculations for solids?, I'm wondering if there is a software for forces with TDDFT.

  • $\begingroup$ I believe the term you're looking for is "Ehrenfest dynamics". A quick Google search shows many codes that implement it (CP2K, CPMD, GPAW, Octopus, Siesta, ...). Haven't used this, though, so I can't recommend a particular code. $\endgroup$ – leopold.talirz Sep 17 '20 at 17:05
  • $\begingroup$ Thanks. Adiabatic dynamics is OK for my purpose, although the Ehrenfest scheme certainly does the job. I'm wondering how those codes deal with the self-interaction error for solids? $\endgroup$ – Xiaoming Wang Sep 18 '20 at 3:45
  • 1
    $\begingroup$ Thanks @leopold.talirz and NikeDattani. Perhaps, my question is not appropriate, what I'm looking for is a workable code that could do excited-state relaxations of solids. Ehrenfest dynamics is not necessary, since I'm only concerning the stable configuration of the lowest excited state. Most TDDFT relaxations work with finite systems, for solids with significant self-interaction error I think TDDFT+force+hybrid functional is one practical solution for this purpose. It seems CP2K has TDDFT+hybrid but no force, while Siesta has TDDFT+force but no hybrid. Not sure the others. $\endgroup$ – Xiaoming Wang Sep 27 '20 at 1:42
  • 2
    $\begingroup$ The original question was if there's a TDDFT code that can relax the first excited state in solids, and has forces implemented. @leopold.talirz mentioned Siesta, which you say can do TDDFT+force, so I think that's a perfectly appropriate answer. Hopefully someone else can suggest a code that can do TDDFT+force+hybrid, but that's not what the question originally asked for. We need to get the ball rolling in the one-topic-per-answer framework and then I'm sure everything will fall into play. $\endgroup$ – Nike Dattani Sep 27 '20 at 4:34
  • 1
    $\begingroup$ Thanks @NikeDattani. I agree. $\endgroup$ – Xiaoming Wang Sep 27 '20 at 5:40

The SIESTA code has a branch (rel-Max-2) developed by researchers from Max Plank institute that include the calculations of forces and real-time TDDFT.

The TDDFT is merged into the main development branch and will be released in versions newer than 4.1 (i.e. 4.2 or 5.0).

To download it, go to the Gitlab page: https://gitlab.com/siesta-project/siesta/-/tree/rel-MaX-2

  • $\begingroup$ Thanks. Just want to add that the TDDFT implementation in Siesta can right now only consider ALDA xc kernel which is good for isolated systems but fails for solids. $\endgroup$ – Xiaoming Wang Apr 26 at 13:38

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.