Generating a pseudopotential for graphene: Is it done on an individual carbon atom, or on an entire unit cell of graphene?

Question

Let's take the example of a 2D graphene layer. The objective is to do an electronic calculation with only the valence electrons with a plane wave basis. Hence the need to pseudize the core electrons through an all-electron (AE) calculation.

Does the AE calculation contain only one carbon atom in my case? Or is it done on a primitive cell with the same basis functions as the valence electron (VE) only calculation?

Answer 1

Pseudopotentials tend to be provided for individual elements in the periodic table.

For example, if you go to the Pseudopotential library and click on the element C, you will get access to many files that contain pseudopotential data for the element C. For example, I see C.awfn in the list of available pseudopotentials and when I click on it, I get this which starts with the following:

Atomic C  wave function in real space
Atomic number
  6
Total number of orbitals
 2
From 1 configurations, orbitals 2S(2)2P(2) [PZ].
Number of up, downspin electrons
n/a
States
n/a
Radial grid (a.u.)
 957
    0.000000000000000E+00
    0.167862832095561E-05
    0.338140915368192E-05
    0.510869001038132E-05
    0.686082340334615E-05

I have not seen pseudopotentials that were built specifically for entire molecules or unit cells, although I don't see why it wouldn't be possible.

The Quantum ESPRESSO website also provides a pseudopotential library here and a list of other pseudopotential libraries here (which includes Pseudo Dojo which was created by my co-author Michiel van Setten). Most if not all of those resources will provide you with a periodic table from which you can select a periodic table element and get its pseudopotentials (and I haven't seen such a database that provides you with various materials like graphene, graphane, or graphyne, and allows you to download a pseudopotential that is built specifically for one of those materials).