Selective cell relaxation

[sergey_lisenkov1]

Dear VASP developers,

I was wondering if there are plans to implement a selective cell relaxation (like only "a", "b", "c", or in-plane only for 2D material) - similar to the following method:
https://github.com/Chengcheng-Xiao/VASP_OPT_AXIS

It is just very annoying to patch every new version of VASP with this and recompile every time. And since the format/content of the VASP's files changes from one release to another, this patch may not work in the future.

Thank you.

[jonathan_lahnsteiner2]

Dear sergey_lisenkov1,

I am not sure if I understand your question correctly. But what you want
to do is to constrain certain lattice vectors during cell shape relaxation with vasp.
As far is I know and understand your question this feature is available in vasp and is documented
here
wiki/index.php/ICONST

under the header Restrictions on the volume and/or shape of the simulation cell.
I hope this helps and that this is the feature your are looking for.


All the best

Jonathan

[sergey_lisenkov1]

Hi Jonathan,

Thanks for your reply. I'm sorry if my original message wasn't clear.

What I meant is "constrained cell relaxation", something like explained in the "constr_cell_relax.F" file, but for more general use, without recompiling every time.

I will check ICONST settings and how they applied. If you are familiar with it, can it be used to do full relaxation (cell+ionic) of 2D material (no cell relaxation along third lattice vector)?

Thanks,
Sergey

[jonathan_lahnsteiner2]

Dear sergey_lisenkov1,

Yes this should be possible. To constrain the length of the third lattice vector you just have to put

Code: Select all

LR 3 0

into your ICONST file. To additionally constrain the angle of the third lattice vector with respect to the second and first you have to add

Code: Select all

LA 1 3 0
LA 2 3 0

But this works only in combination with IBRION=0.
Another feature which might be interesting for you and which is working for IBRION>0 is the INCAR tag LATTICE_CONSTRAINTS:
wiki/index.php/LATTICE_CONSTRAINTS

And last I would like to recommend this post to you, which might also be helpful:
forum/viewtopic.php?f=4&t=18723&p=22915#p22915

All the best Jonathan

[sergey_lisenkov1]

Hello,

it seems that all these options are only used for MD simulations. But i'm mostly interested in geometry relaxation (IBRION = 2).

[jonathan_lahnsteiner2]

Dear sergey_lisenkov1,

I think this is a miss-understanding. Both links I sent to you are meant for structure relaxation. So they are usable with IBRION=2. Especially check the link https://www.vasp.at/forum/viewtopic.php ... 915#p22915 there was some update until our last
conversation. The more sophisticated feature for relaxation mentioned in this post will become available in future vasp version.

Best Jonathan