I calculate DOS from VASP code with default NEDOS as well as increasing NEDOS with 3000. When I plot DOS from DOSCAR file it does not look good as we see in literature. Sometime one peak dos very large compared to other sometime all DOS peaks look zigzag.
I plot DOS in origin software. Sometime I use GNUPLOT also. How can I improve plot quality ?
1: By any input parameter in VASP
2: By plotting software
3: Any other
I have attached here my simulation output files files as well as DOS plot and origin file.
Dear SK Sharma,
1: By any input parameter in VASP
Tetrahedron method and Fine k meshes are few important parameters to get a good DOS.
http://cms.mpi.univie.ac.at/vasp/vasp/Accurate_DOS_Band_structure_calculations.html
2: By plotting software
Both origin and Gnuplot will give the good quality pictures.
Dear SK Sharma,
1: By any input parameter in VASP
Tetrahedron method and Fine k meshes are few important parameters to get a good DOS.
http://cms.mpi.univie.ac.at/vasp/vasp/Accurate_DOS_Band_structure_calculations.html
2: By plotting software
Both origin and Gnuplot will give the good quality pictures.
You are running a small cell with few k-points. As Manoharan said, increase the k-points if you want a more accurate DOS. NEDOS at 3000 is more than enough. The plotting software just plots the data it gets. No problem there.
I have increased KPOINTS upto 24*24*24 still I am getting almost same DOS plot as in my previous post.
I have attached here my DOS plot and INCAR file.
Any suggestion !!!
Shilendra
Increasing SIGMA (perhaps to 0.1) in INCAR will make the DOS features broader and smoother, but check that this does not significantly affect the total energy of your simulation. See the linked page in the VASP manual for some guidelines.
http://cms.mpi.univie.ac.at/vasp/guide/node124.html
Dear Prof. Martin I applied your suggestion SIGMA=0.1. Now DOS is looking somewhat better than previous DOS in above post. I have attached here new DOS.
Please look my DOS and any more suggestion and comment ?
Thank You for suggestions !
Shilendra
Dear Federico,
I use png format generally . Any other suggestion for figure format ?
Thank You
Shilendra
Make sure that you use tetrahedron method (ISMEAR=-5) as Kamajai suggested. We do not know what is the size of your system, so please check k-point density about 0.03/A. If you did, then you do not need to do anything else with VASP calculations.
For plotting, people prefer to shift the Fevel level to 0 (eV), so I recommend you do the same thing. Also, keep the range of energy to your region of interest, you should not draw the wide range of energy as the figures. Usually, DOS about the Fermi level is more important.
I prefer plotting with GNU plot or Xmgrace, print as eps and then convert eps to PDF. You will get high quality images.
Shilendra Kumar Sharma
You can use eps format for images. They are non-pixelated (vector graphics). In simple words, you can keep zooming in the image without decreasing quality.
Use smooth unique/csplines in gnuplot or method of running average of 2 or 3 points ( I don't have a script for that).
Also plot DOS per atom., select only meaningful part of the valance and conduction band.
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Materials Science