My Community

Hi,

I am reading a paper ( PHYSICAL REVIEW B 83, 045409, 2011) which use the a combination of revPBE+LDA+dispersion) to calculate the exchange-correlation energy via QUANTUM-ESPRESSO. I am wondering if it is possible to such calculation using VASP? Any hint or guidance are appreciated!

Thank you,

In order to calculate adsorption energies and
properly take into account dispersive interactions, we use
the vdW-DF approach proposed by Dion et al.23,24 showing
transferability across a broad spectrum of interactions.25,26
In this method, the exchange-correlation energy functional
consists of three parts: (i) the exchange part of the revised
Pedrew-Burke-Ernzerhof (revPBE) functional,27 (ii) the local
correlation part of the standard local density approximation
(LDA) functional, and (iii) the nonlocal correlation part,
incorporating effective many-body density response and allowing treatment of dispersive interactions without any fitting
parameters.

There is a limited amount of these combinations that are available in VASP. I cannot comment on that particular work, though mixing LDA correlation with revPBE exchange seems strange and is not possible at the moment. What is implemented in VASP is a link to LIBXC for GGA functionals only, so you could check if there is a GGA functional that behaves close to this particular combination.

As for the dispersion correction: Here is a link to all the van der Waals functionals implemented in VASP.

Hi,

On this page vdW-DF it is explained how to use vdW-DF. "GGA = RE" corresponds to revPBE and with "AGGAC = 0.0000" you can turn off the GGA component of correlation (i.e., use LDA for correlation).

[MS: fixed link]

Thank you, fabien. Yes, we figured it out as what you point out.