My Community

Hi all,
I'm trying to compute elastic constants of MgCO3 via the stress-strain method.
I initially relax the structure (using ISIF = 3) which works fine and the stress tensor is approx. zero across all elements.
However, when I apply a strain of, say, 0.5% (either compressive or tension) and relax only the atoms (using ISIF = 2), I get very high stress tensor values. An x-direction loading of 0.5% gives me an XX stress tensor value of something around 400 kbar.
Applying 1%, 1.5% and 2% strain from there on gives regular jumps in stress tensor value, but it's the initial 0 kbar to 400 kbar jump that's causing me a lot of problems.

Can anybody help me with why this may be?

Relaxed POSCAR of MgCO3 lattice parameters:

2.3179487194980619 -4.0148049515098636 0.0000000000000000
2.3179487194980619 4.0148049515098636 0.0000000000000000
0.0000000000000000 0.0000000000000000 14.9269277154155713

My INCAR looks like this when computing strained structures:

ISTART = 0
ICHARG = 2
ENCUT = 520
PREC = Accurate
ISMEAR = 0
SIGMA = 0.1
ISIF = 2
IBRION = 2
NSW = 100
NELM = 100
EDIFF = 1E-7
EDIFFG = -0.001
LREAL = .FALSE.
ISYM = 0

Thank you

Hi Alina,

Thank you for your question. Could you upload the POSCAR, POTCAR, KPOINTS, INCAR, and OUTCAR files? We can then take a look at your calculation and try to reproduce it.

Best wishes,

Chris

Hi Christopher,

Thank you so much for your reply.
Thankfully I identified the issue - my ENCUT was too low. Increasing ENCUT has now yielded reasonable results
I appreciate the response regardless.

Alina

Hi Alina,

I'm glad that you could solve it yourself. If the ENCUT was too low, then I imagine that it was an issue with Pulay stress?

I'll lock the thread at the end of the week, unless you have any further questions?

Best wishes,

Chris