EIGENVALUE file: What are these numbers?

[lahaye]

Hello,

Step by step I try to understand the in/output files of VASP.

The guidebook says about the EIGENVALUE file:

Code: Select all

The file EIGENVALUE contains the Kohn-Sham-eigenvalues
for all k-points, at the end of the simulation. For dynamic simulation
(IBRION=0) the eigenvalues on the file are usually that one predicted
for the next step: i.e. the file is compatible with CONTCAR.
For static calculations and relaxations (IBRION=-1,1,2) the eigenvalues
are the solution of KS-equations for the last step.

As an exercise, I have simulated the CO molecule from the on-line
workshop/notes, which produces this EIGENVALUE file:

Code: Select all

    2    2    5    1
  0.2560000E+03  0.8000000E-09  0.8000000E-09  0.8000000E-09  0.5000000E-15
  1.000000000000000E-004
  CAR 
 CO in a box                       
   10    1   10
 
  0.0000000E+00  0.0000000E+00  0.0000000E+00  0.1000000E+01
   1      -28.9436
   2      -13.7661
   3      -11.4387
   4      -11.4387
   5       -8.7734
   6       -1.8803
   7       -1.8803
   8       -0.3114
   9        1.6888
  10        1.8239

Now, I think the two columns of the last 10 lines are the 10
eigenvalues of the system (I suppose I understand that
part correctly). But all the other numbers above do not make sense
at all to me. Can someone help me understand this data?

Thank you.
Rob.

[admin]

the numbers are
1) # of ions , # of ions, the number of loops after which the averaged pair correlation functions and the DOS are written, ISPIN
2) the volume of the cell (in Ã…^3) and the lattice parameters of the box (in m)
3) T
4) the string 'CAR'
5) the header
6) # of electrons, # of k-points, #of bands
7) the k-point and its weight

[lahaye]

Thank you. That helps a lot.
This important information is not in the guidebook. Shouldn't it be added
in the chapter of the EIGENVALUE file?

Code: Select all

the numbers are
[...]
2) the volume of the cell (in Ã…^3) and the lattice parameters of the box (in m) 

In general, I find the VASP input/output not clear (and consistent) in
its choice of units. At least there's no mentioning of it in the guidebook.
I only can guess the default units to be Angstrom and eV at some
places, whereas at other places they appear to be meter and what not.

Wouldn't it be better to have a more consistent policy towards the use
of units, and whereever there's a deviation, print the units explicitly?

Regards,
Rob.

[admin]

VASP uses eV and Ã… as default units.

[jlbettis]

With respect to Band Structure plots obtained from recomposing the EIGENVALUE file, is the Fermi-level at 0 eV?

[jlbettis]

I believe I should use the Fermi-level obtained from a previously converged SCF run.