non-zero pressure at minimum energy structure from equation of states

[IBRAHIM]

Dear VASP admin,

I found the same issue described here "https://www.vasp.at/forum/viewtopic.php ... ure#p20492" but this time with PBE functional!

I follow the procedures recommended by VASP for lattice volume optimization "https://www.vasp.at/wiki/index.php/Ener ... lay_stress"

When I run a volume optimization (step2 in the above link) with different volumes (and ISIF=2,4), then fit to Birch-Murnaghan equation of state to find the equilibrium cell volume corresponding to
minimal energy and then relaxed again this cell, I got a non-zero external pressure (-2 to -7 KB).

However, if I did direct optimization (step 1 in the above link) with ISIF=3, I get an external pressure close to zero for a cell volume that is smaller than that corresponding to the minimum energy.

I run both calculations at the same increased ENCUT=1.3*Enmax and PREC=Accurate. I did it for simple metal and for my bulk systems with the same issue.

Is it a bug in the calculated stress tensor elements same as the first link? or is it normal?
because after studying the electronic properties of my systems optimized accurately using different volumes and fit to an equation of state,
I will also study the mechanical properties using the elastic tensor from VASP for those optimized structures.
The elastic tensor components will be adjusted MechElastic code for any residual pressure indicated by VASP

A similar issue was also reported in the forum "https://www.vasp.at/forum/viewtopic.php ... ure#p17494"

Many thanks in advance

[pedro_melo]

Dear IBRAHIM,

Your calculations involve different types of geometry optimisation:

1) ISIF = 2 => cell shape change: no; cell volume change: no
2) ISIF = 4 => cell shape change: yes; cell volume change: no
3) ISIF = 3 => cell shape change: yes; cell volume change: yes

If you let the cell shape change you are likely to find another configuration that has a lower pressure, but where the symmetry group will no longer be that with which you started. I would suggest that you go through the options for ISIF given here https://www.vasp.at/wiki/index.php/ISIF and select the most appropriate for your calculation. If you do not want the symmetries to change, it is probably better to pick a choice where the cell shape is unaltered.

If you want a better fit to the equation of state, you might need have smaller the steps in change of volume.

Kind regards,
Pedro Melo