[QE-users] About the ONCVPSP pseudopotential suitable for Quantum ESPRESSO.

[Hongyi Zhao]

Hi,

I noted the following github repo: https://github.com/pipidog/ONCVPSP
which told the following things:

------------------------------------------
Optimized Norm-Conserving Vanderbilt Pseudopotential (ONCVPSP) is an
accurate and inexpansive NCPP:
https://journals.aps.org/prb/abstract/10.1103/PhysRevB.88.085117
Currnetly there are two major ONCVPSP database, one is provided by
ABINIT's official website:
http://www.abinit.org/downloads/pseudodojo/pseudodojo
The other one is provided by sg15 database:
http://www.quantum-simulation.org/potentials/sg15_oncv/

The former is in psp8 format which is not compatible with Quantum
Espresso (QE). The later is already in UPF format but these ONCVPSPs
do not contain the PSWFC section (wave functions in PSP) which is
required by projwfc.x, the code in QE to calculate PDOS and fatband,
to project. As a result, neither of them is fully compatible with QE
and, more importantly, neither one provide fully-relativisitc version
(needed for spin-orbit coupling).

In short, you will always need to regenerate your own ONCVPSP in most
QE calculatons especially when it comes to spin-orbit coupling.
------------------------------------------------

Based on the above description, the sg15_oncv supplied here
http://www.quantum-simulation.org/potentials/sg15_oncv/, has not the
out-of-the-box feature for Quantum ESPRESSO.

On the other hand, the https://github.com/pipidog/ONCVPSP itself was
recently  updated 2017, so it maybe outdated.

As a result, I still confused on the selection of ONCVPSP
pseudopotential suitable for Quantum ESPRESSO.  Could you please give
some hints on this issue?

Thanks a lot in advance.

Regards
-- 
Hongyi Zhao <hongyi.zhao at gmail.com>