HI all,
I need your kind help in phonon caculation.
I have taken the fully relaxed structure (1x1x1 surface slab) and the forces are relaxed up to 8.1e-05.
Then, I have taken 2x2x1 supercell of same structure and K POINTS are halved (i.e. K POINTS changed from 6 10 1 to 4 6 1). I have relaxed the 2x2x1 structure using ISIF= 2, IBRION =2 , NSW=1.but after one step the max force is of the order of 0.016105.
I am very doubtful whether I should go for phonon with this structure or I should relax the 2x2x1 structure fully (with max possible number of steps)?
Wait.
In general:
1) Optimize the single cell down to 1E-3 meV. (1/10 of VASP default)
2) With the FDM (or supercell method, which I guess is what you are doing) you do not need to reoptimize the supercell. The sueprcell is needed to decouple the displacement in the single cell. An optimization of the supercell is not negative for itself, but just moves you to a better approximation. Nonetheless if the supercell is not needed to describe any new effect, it is physically insignificant.
Great
Thanks for your reply.
Can I ask one more question related to IBRION =5 calc
I can tell you my procedure:
1) I optimize everything tightly with VASP
2) If I want to use FDM, I usually perform it with PhonoPy. I never tested FDM within VASP. I guess the problem was the absence of FDM away from Gamma ( if I remember) and the nice interface that phonopy offers.
3) I used IBRION 7 for the DFPT
Try to tight up the forces btw. And also check carefully Cutoff and Kpoint convergence.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Atomic, Molecular and Optical Physics