Greetings,
If I were interested in writing post processing scripts for wave functions from a VASP calculation, how can I get a text file printout of the wave functions used for the system? WAVECAR is binary format.
Thanks
Jeff Mullen
ECE NCSU
Post processing WAVECAR file
Moderators: Global Moderator, Moderator
-
jmullen
- Newbie
- Posts: 8
- Joined: Wed Nov 04, 2015 10:45 pm
- License Nr.: 5-2236
-
jnokelai
- Newbie
- Posts: 5
- Joined: Thu Apr 14, 2016 10:58 am
- License Nr.: 5-726
Re: Post processing WAVECAR file
Hello,
There does exist a program for that purpose: http://www.andrew.cmu.edu/user/feenstra/wavetrans/
So WAVECAR contains the plane wave amplitudes, and the WaveTrans program writes them out in txt format. If I have understood it correctly, in order to be able to read the WAVECAR file, the number of k-points and the number of plane waves for each k-point has to be know beforehands to be able to read the WAVECAR file, so it calculates them based on present symmetries. However, in the case of rhombohedral systems, this program told me that the amount of estimated plane waves for the first k-point does not match with WAVECAR, so I have suspected that there might be a bug there. But for other systems this worked fine for me. By the way, I was thinking of using this program to plot plane wave coefficients as a function of energy in order to be able to see good values for ENCUT in more elegant way than doing energy convergence tests as a function of ENCUT (I thought that for energies higher than ENCUT the plane wave amplitudes should become neglible). But this is still in the TODO-list.
Moreover, there is also the WaveFcnPlot program, which calculates the wave function in real space for given x, y, k, band and spin as a function of z.
There does exist a program for that purpose: http://www.andrew.cmu.edu/user/feenstra/wavetrans/
So WAVECAR contains the plane wave amplitudes, and the WaveTrans program writes them out in txt format. If I have understood it correctly, in order to be able to read the WAVECAR file, the number of k-points and the number of plane waves for each k-point has to be know beforehands to be able to read the WAVECAR file, so it calculates them based on present symmetries. However, in the case of rhombohedral systems, this program told me that the amount of estimated plane waves for the first k-point does not match with WAVECAR, so I have suspected that there might be a bug there. But for other systems this worked fine for me. By the way, I was thinking of using this program to plot plane wave coefficients as a function of energy in order to be able to see good values for ENCUT in more elegant way than doing energy convergence tests as a function of ENCUT (I thought that for energies higher than ENCUT the plane wave amplitudes should become neglible). But this is still in the TODO-list.
Moreover, there is also the WaveFcnPlot program, which calculates the wave function in real space for given x, y, k, band and spin as a function of z.