Hello!
My name is Anastasia Korovina. I am second year Master student at the Moscow Institute of Physics and Technology (MIPT) from Moscow.
Currently, I started to study ways for calculation of piezoelectric response of 2D materials using VASP and Quantum Espresso.
At this time, I carried out test calculations of piezoelectric coefficients for h-BN and MoS2. Unfortunately, obtained results are significantly different from reference. The value of the clamped-ion coefficients is the same as in reference data (3.7*10^-10 C/m), but the values for relaxed-ion are totally different than in reference value and equally the same comparing with clamped-ion approximation. Perhaps I misunderstand the origin of these approximations. I will be very grateful to you if you could please help me figure this out or could advise some person who can help?
For h-BN, we use rectangular unit cells of 4 atoms. Clamped-ion coefficients were achieved by setting of ISIF = 2, with no ionic relaxation (electronic only), so the ions cannot move to their position after deformation. Relaxed-ion coefficients were achieved by setting of ISIF = 2, but with the inclusion of electronic relaxation NSW = 900. However, this was not successful because I have got the same values for both configurations, although we clearly see in the literature that the values should be different. For the case of MoS2, we also used a rectangular unit cell (6 atoms). But in this case, I didn't obtain any correlation with the literature neither for clamped nor for relaxed-ion.
We used a several methods: we also replaced pseudopotentials for Mo atoms to more complete Mo_pv, set such flags as LCALCPOL, increased EDIFF, increased the number of k-points.
Now I am in a difficult situation, and I cannot move further in research. I would be very grateful if you would tell me where I am making a fundamental mistake.
Thanks in advance!
I need some more information to explain... such as total energy and other parameters... So directly message me.
Here are a few things to check.
Have you verified that the geometry is different for the two calculations you're doing, i.e. is the final geometry from NSW=900 (this is a bit high) actually different from the simple scf calculation in your first step? If so, are they significantly different? Also, is the reference data computational, and do they have the structural parameters available for you to compare with?
- Badges
- Science topic
- Similar topics
- Names