[CP2K-user] energy fluctuation in GEO-OPT

[Marcella Iannuzzi]

Either the structure or the DFT setting must have serious problems, since 
the SCF is never converging. 
For example in SUBSYS you use "periodic none" with a box of 15 Å, but your 
system seems to be larger than that.
On the other hand, POISSON is still using PBC.  What do you want to model, 
an isolated subsystem of a crystal?
With such a wrong electronic structure you cannot expect any meaningful 
evaluation of the forces.
Before starting the GEO_OPT you need to be able to get a reasonable 
electronic structure of the initial state.

Regards
Marcella

On Wednesday, February 20, 2019 at 6:16:39 AM UTC+1, M. Memol wrote:
>
> I am a very recent user of cp2k and I am really sorry if my question is a 
> naive one. I am getting familiar with calculation by cp2k. I am doing an 
> geometry optimization run on a MOF building block. But comparing to the 
> examples, the energy fluctuates. input and output are given below. any 
> comment?
> thanks in advance 
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <https://lists.cp2k.org/archives/cp2k-user/attachments/20190220/29fdf820/attachment.htm>