[Pw_forum] problem with pseudopotential

[Paolo Giannozzi]

On Oct 13, 2006, at 16:40 , luch001 wrote:

> I want to determine the stable structure by searching the global  
> minimum of the
> ab initio free energy. Therefore, I should get reliable free energy  
> with acceptable parameters (e.g., k-mesh and Ecut). In other words,  
> the convergence of the PP
> should be good enough, namely a bit softer PP. But simultaneously I  
> should
> guarantee the pseudo cores do not overlap too much, that is to say  
> the rc should
> be small enough. This is paradox.

it is not a paradox, it is life. You have to strike a compromise  
between two
conflicting requirements. Anyway: I don't think that PP cores  
overlap, even
at very high pressure, except for some large-core elements. If this  
happens,
use smaller cores, or move semicore states into valence

Paolo