[Pw_forum] external potential and partial wavefunctions

[Guido Fratesi]

> The external potential should contain both the local part (which
> should be the local ionic potential) and the non-local part which is
> V_NL. In the USPP,
> V_NL=sum_ij (D_ij *|beta_i> <beta_j|
> My question is: in PWscf, how to get out the non-local potential V_NL.
> Where are the D_ij and |beta_i> stored?  In the paper Vanderbilt, PRB
> 41, 7892 (1990), the author suggested that "the pseudopotential itself
> must be updated as part of the self-consistent screening process". I
> wonder whether PWscf does this in the self-consistent iterations. If
> yes, where are the new pseudopotentials stored? Regarding to the
> non-local psedupotential, I guess, in the pp.x, the total potential
> V_bare+V_H + V_xc doesn't contain the non-local part V_NL, right?  Do
> the V_bare and local ionic potential refer to the same thing? Or what
> does the V_bare mean here?

pp.x plots the LOCAL part only, possibly adding Hartree and XC 
contributions. One cannot plot as function of r a nonlocal potential.

The NL part is defined by matrix D and projectors beta. IT IS updated in 
the scf cycle by PW/newd.f90 (variable deeq):
* projectors    are read from the pseudopotential file
* D^0    (unscreened D) as well
* D = D^0 + \int V Q dr   is computed by PW/newd.f90
Look at PRB 47, 10142 (1993), Laasonen, CPMD with Vanderbilt USPP in 
particular Eq. (13) and following note discussing the screening of D.

Guido Fratesi

Dipartimento di Scienza dei Materiali
Universita` degli studi di Milano-Bicocca
via Cozzi 53, 20125 Milano, Italy

Phone: +39 02 6448 5214
email: fratesi at mater.unimib.it