[CPMD-list] Gold pseudopotentials and cell parameters

[Ari P Seitsonen]

Dear Stefano,

  Not much about pseudo (well, you CERTAINLY shouldn't use LOC=D, you get 
ghosts!), your pseudo is probably okay (otherwise I can give you one). 
However bulk gold is METALLIC, you'll certainly have problems with ODIIS 
and PCG; there are too many states around the Fermi energy. You have to 
use broadening (FREE ENERGY FUNCTIONAL), or if you only want to dynamics, 
you can try using a thermostat for the electrons.

  Or maybe I misunderstood your problem.

    Greetings from Zurich,

       apsi

-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-=*=-
 Ari Paavo Seitsonen / Ari.P.Seitsonen at iki.fi / http://www.iki.fi/~apsi/
 Tel +41 1 635 44 97 / Fax +41 1 635 68 38 / GSM +41 79 719 09 35      
 Anschrift: Physikalisch Chemisches Institut (PCI), Universität Zürich (UniZh)
 Indirizzo: Winterthurerstrasse 190, CH-8057 Zürich
 Address:   Schweiz / Svizzera / Suisse / Svizra / Switzerland

On Fri, 10 Dec 2004, Stefano Piana wrote:

> Dear CPMD users,
> I am still generating MT PBE gold pseudopotentials with not a lot of success 
> so far.
> At the moment I am able to generate pseudopotentials that give a reasonable 
> geometry for the gold dimer but that have serious convergence problems when I 
> try to simulate the bulk gold. I tried to generate the pseudos from the Au and 
> Au+ atom, but they do not differ much. I tried with and without NLCC but, 
> again, not much difference there.
> I tried both LOC=S;P and D. LOC=S seems to give the best results. I also 
> played a bit with the cutoff (from 50 to 80) and with the GC-CUTOFF (from 1e-6 
> to 5e-5) but these do not seem to change things much, a cutoff of 50 and a 
> GC-CUTOFF of 5e-6 seem to be reasonable.
> 
> Somebody suggested me that I should use a _LARGE_ number of K-Points for the 
> bulk gold. Although I tried to run with up to 40 K-Points, I still have 
> convergence problems and moreover it is not possible to optimize the cell with 
> K-POints, therefore I cannot calculate the bulk properties of gold with my 
> pseudo. An alternative is to use a bigger cell. I tried to run with 2x2x2 
> supercell containing 32 gold atoms, but still convergence problems arise (i.e. 
> the wavefunction never converges). I tried also INITIALIZE WAVEFUNCTION RANDOM 
> and to run a few steps of PCG MINIMIZE before switching to ODIIS. The second 
> method sometimes works, but convergence seems to be a bit erratic.
> 
> At the moment I am including only the 6S and 5D in the valence. IF I include 
> the 5P as well I get very good results for the gold dimer (as expected from 
> the literature), but the calculation is more expensive and it looks like the 
> convergence problems for the solid state worsen.
> I would like to have some suggestions. If anybody has already generated a MT 
> pseudopotential for gold could he send it to me so that I can compare with my 
> pseudos?
> 
> Stefano
> -- 
> Nanochemistry Research Institute
> Curtin University of Technology
> GPO Box U1987
> Perth 6845 - Western Australia
> tel. +61 8 9266 2687
> piana at power.curtin.edu.au
> 
> _______________________________________________
> CPMD-list mailing list
> CPMD-list at cpmd.org
> http://cpmd.org/mailman/listinfo/cpmd-list
>