XAS Calculation

[Gerard pareras niell]

Hello everyone,

I am trying to run a XAS calculation, however the convergence for the XAS section is very bad and there is no way the SCF is converging. I have reduced a lot the threshold just for testing but still not working. The system I am using is a bit tricky, 13Fe in a cluster on a SiO2 surface. I was wondering, is there any way to apply the CG minimizer for the SCF section on XAS? If not, how we can improve the convergence for that section?

Please find attached here the input and output for the calculation.

Any suggestion for improving the calc is more than welcome!
Thank you very much in advance.

Gerard

[Marcella Iannuzzi]

Hello Gerard, 

There are several problems in the simulation you are running.

First of all, for Fe an all electron potential is indicated, but a basis set for the valence electron only is used. This is not going to work.

The ground state electronic structure calculation is, not surprisingly, converging very badly.

Not knowing the properties of your system, it is difficult to judge, what the optimal settings would be, in terms of PP, basis set, cutoff, spin polarization, etc

I suggest that for the PP description you use POTENTIAL_UZH and BASIS_MOLOPT_UZH libraries. For the ALL electron description check you can take from the EMSL_BASIS_SET or check in the literature what is suggested for K-edge spectra of transition metals. 

It is possible that also for the ground state one should use diagonalization and smearing to properly converge the SCF. 

For the TP method to obtain XAS it is not possible to use OT because the occupation number of the excited core state is modified, meaning that the diagonalization approach has to be used. Sometime G-space mixing and smearing help the convergence.

Consider also the possibility to compute XAS with XAS_TDP.

Regards

Marcella