When I relax a structure in DFT calculation using vasp, I get a new structure in the contcar file , When I use the coordinates in the contcar file to see the structure in vesta , it looks a lot different from the initial structure. The bonds are broken and I am not able to see a stand of the PILs structure I used. If anybody knows why this happened or this is normal during structure relaxation. Anybody else has same problem or I am doing something wrong. Any help will be much appreciated.
Thank you
DFT, like all ab initio methods, doesn't "know" bonds, it only "knows" orbitals. Therefore, if there are no significant interactions between the atoms between you drew a bond, DFT doesn't care that you drew something there or not, it will just optimize to a state in which the interactions are more favorable - and if that involves major structural changes, it will do that.
There are some things that can cause an optimization to go wrong, here are a few:
1) if the potential surfaces of two structures are close together, at some optimization step there may be a leap from one to another. If you have performed an optimization with checkpoints (these may have a different name in VASP, I've only worked with other tools), you can check if that was the case at any of your points and restart the optimization from there with different parameters.
2) if strange things happen during the optimization, parameters to be varied would be the functional, basis set (going larger) or also the SCF criterium.
One thing that you should try is to load your CONTCAR file in vesta and in the boundary option (object -> boundary), set each min in -1 and each max in 2. It duplicates the cell in each direction and if in that visualization you still finding something wrong, definitely the optimization is in the wrong direction. You should check your cut-off energy, the incorporation of vdW corrections, and/or the energy criteria (EDIFF) for the SCF.
- Badges
- Science topic
- Similar topics
- Filing