[CP2K-user] Transition State Optimisation using PM6

[ma...@gmail.com]

Just to reply to myself (and potential readers from the future), the 
keyword LEVEL_SHIFT 
(https://manual.cp2k.org/trunk/CP2K_INPUT/FORCE_EVAL/DFT/SCF.html#LEVEL_SHIFT) 
helped me. I increased it to 30, the structure converged and then I used 
the converged structure in the next calculation without the LEVEL_SHIFT 
keyword and it worked :) .

On Friday, 21 August 2020 at 16:07:54 UTC+1 ma... at gmail.com wrote:

>
> Dear CP2K community,
>
> I encountered a peculiar problem with optimising transition states using 
> the semi-empirical method PM6. During the optimisation, the forces acted 
> weirdly and the whole structure started flying apart. I should say that 
> similar transition states for other molecules have been successful.
>
> I have seen such behaviour before, but in the previous case, the SCF cycle 
> failed to converge, which lead to unreasonable forces. In the current case, 
> there are no problems with the SCF cycles and as I said, similar transition 
> states converge quite quickly.
>
> Furthermore, I should say that I did check the vibrational frequencies and 
> the structure is a very good initial guess for the transition state. I 
> tried to attach the necessary files (including the geometry optimisation 
> frames and the vibrational frequency output), but for some reason, I 
> couldn't. If you need the inputs, I can copy them here. The program version 
> I am using is CP2K 6.1. If you have any further questions, let me know.
>
> Thank you in advance,
> Mario

-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20201203/bff02304/attachment.htm>