ABINIT Discussion Forums

Dear all,

I'm trying to get information about pressure-induced phase-transitions (monoclinic/orthorhomb.) of lanthanum sulfides. For this task I would like to use Abinit and ELK as an comparison (as mentioned in one of the tutorials). I'm new to calculating and have several questions concerning convergence-criteria and in general.

I have chosen Abinit to geometry-optimize structures of two phases at given cell-volumes and then use the energies to get E/V-curves for equation-of-state fitting. I read here that a faster way is to give a target-pressure and take the resulting volume after the geometry-optimization. I would further like to use the output of those runs as an input for ELK-calculations (g.opt. runs really slow with it). I played around with Abinit the last days to get familiar with it, now I have the following questions.

1) As a „check“ of given PAW-Potentials, is it allowed to compare bandstructures of e.g. LaCl3 and Na2S calculated each with both ELK and Abinit? I would like to avoid long runs for that „potential-check“.
2) I skipped 1) yesterday (using PAW-files from abinit.org) and tried to get converged values of ecut, pawedgcut, and ngkpt 3*x with kptop 1. For monoclinic a_LaS2 i have reached values of ecut: 30Ha , pawedgcut 100 and ngkpt 3*3, what seems very high for me. Still I'm far away from dE=10E-6 Ha/run (that value is also set as conv.criterium for one run)

Is it senseless to begin any discussion before doing task 1) or is it possible that Lanthanum requires really large ecutoffs that i just have not reached yet?

Yours, Chr.