Page 1 of 1

elastic constants and ultrasoft pseudopotentials

Posted: Tue Aug 24, 2010 10:36 pm
by alejandro.bautista
Dear friends:
can be calculated elastic constants using ultrasoft pseudopotentials?
regards

alejandro

Re: elastic constants and ultrasoft pseudopotentials

Posted: Thu Aug 26, 2010 2:23 pm
by jzwanzig
Hi,
first of all, currently abinit computes elastic constants using density functional perturbation theory only from norm-conserving pseudopotentials. And in any case, Abinit never uses ultrasoft pseudopotentials, your two choices here are norm-conserving pseudopotentials or the Projector Augmented Wave (PAW) formalism. In PAW the cut-off energies are as low as with ultrasoft but the results are in principle even more accurate. At this moment abinit does not implement strain perturbations using PAW, however you can easily apply the strains yourself (by hand, by modifying rprim) and computing the stresses with PAW; you then obtain the elastic constants by finite difference. This approach works really well, my group uses it all the time.

Joe Zwanziger