How to obtain correct ELF?

[Kazunori Shibata]

Dear all,

I'm calculating several atoms, such as H, He, O, and Ne. I encountered a strange behavior of the ELF. Please see the figure for the ELF and electron density of hydrogen. The horizontal line is the distance from a hydrogen nucleus (in au). The ELF(ELF1D.OUT) and electron density(RHO1D.OUT) are shown by the green and purple lines, respectively.

The ELF shows two strange behaviors.
1: An artificial-like spine at about 1.4 au. -> This may correspond to the "rms" of H.in.
2: ELF keeps about 1. -> It must close to zero where electron density closes to zero.

What parameters should I change?
The input file is also attached(H_atom.in). I've tried several values for gmaxvr,rgkmax,lradstp,lmaxapw,and lmaxvr. They haven't solved above problems.

I confirmed that the electron density accords with the exact value of the ground state of isolated hydrogen.
So, the electron density itself will be calculated correctly.

In addition, I haven't succeeded to reproduce the ELF of Neon atom, shown in J. Chem. Phys. 92, 5397-5403 (1990). If you have reproduced, please provide me the .in file.

I'd like to hear experts' comments.
Thank you for your help in advance!

Kazu-s

[Andrew Shyichuk]

Dear Kazu-s,
Long story short - you gotta optimize RMT, rgkmax, number of k-points, etc. RMT first, lmaxo/lmaxi/lmaxapw later. What I am doing now is trying to minimize error in calculated charge. Sad story of trial and effort.

[J. K. Dewhurst]

Hi All,

When in doubt, set highq to .true.

Here is my input file:

tasks
  0
 31
 51

highq
 .true.

ngridk
 1  1  1

avec
  8.0000  0.0000  0.0000
  0.0000  8.0000  0.0000
  0.0000  0.0000  8.0000

molecule
 .true.

atoms
 1
 'H.in'
 1
 0.0000 0.0000  0.000000 0.0 0.0 0.0

plot1d
 2
 2000
 0.0 0.0 0.0
 1.0 1.0 1.0

The ELF is much smoother. The parameters used by highq are written to stdout. These can be further increased to improve the smoothness at the muffin-tin boundary.

Regards,
Kay.