How to model annealing

[maxime_laniel]

Hi,

From previous threads in this forum (from many years ago) it seems that it was possible to model annealing. I was wondering if it is still the case. Also, how exactly would I go about it? From my understanding, correct me if I'm wrong, you would have to relax your structure, heat the structure in AIMD mode, cool it again in AIMD mode and finally relax the structure again (not in AIMD mode). From what I've gathered, you would be able to run a heating step using the TEBEG, TEEND and SMASS INCAR tags. Could you provide a clearer explanation as to what tags would have to be used throughout this process or refer me to a tutorial of some sort? Thank you for your help.

Best,

ML

[pedro_melo]

Dear ML,

I spoke with my colleagues and, while we do not have a tutorial for how to simulate annealing, I can tell you that it should be a matter of running a molecular dynamics where the TEBEG is set to some starting temperature and TEND has to be set to the final temperature to which the structure should be annealed. Since the annealing process should be done slowly you should set NSW to a high value.

Let me know if this helps. Kind regards,
Pedro

[maxime_laniel]

Dear Pedro,

Thank you for your response, it is helpful. Also, I just wanted to make sure with you that, to simulate the subsequent quenching, it would essentially be the same idea (molecular dynamics run where you set TEBEG and TEND), but the NSW value would be smaller. Thank you for your help.

Best regards,

ML