Re: How to generate a mixed pseudopotential ( No.1 ) |
- Date: 2015/06/08 14:04
- Name: T. Ozaki
- Hi,
One can generate a mixed pseudopotential located in between atoms A and B whose nuclear charges are Z and Z+1, respectively. Please see http://www.openmx-square.org/adpack_man2.2/node23.html
Mixing of pseudopotentials for arbitrary atoms is not supported.
Regards,
TO
|
Re: How to generate a mixed pseudopotential ( No.2 ) |
- Date: 2015/06/09 03:16
- Name: PR
- Dear Prof. Ozaki,
Thanks for your response, Is there any fundamental reason that it is not supported by OpenMX? Is it possible that I manually generate two Fully relativistic pseudopotentials and their related Pseudo-atomic orbitals and then mix them to get the desired mixing of pseudopotentials for my arbitrary atoms?
If I know what parts of Fully relativistic pseudopotentials and Pseudo-atomic orbitals are used in the OpenMX code after generation by ADPACK can be very helpful.
For example, I assume in Fully relativistic pseudopotentials, OpenMX use only Core electron densities for PCC <density.PCC density.PCC> and <Pseudo.Potentials Pseudo.Potentials>
And in Pseudo-atomic orbitals OpenMX use only Charge density of valence electrons <valence.charge.density valence.charge.density> and DATA for multiple pseudo atomic orbitals.
Is this true?
Is there any guidance you could probably give in case of mixing basis functions or getting the Pseudo-atomic orbitals for the mixed pseudopotential?
Thanks in advance,
Best regards,
PR
|
Re: How to generate a mixed pseudopotential ( No.3 ) |
- Date: 2015/06/09 17:32
- Name: T. Ozaki
- Hi,
OpenMX reads from *.vps the following information:
AtomSpecies total.electron valence.electro grid.num.output j.dependent.pseudo.potentials <project.energies ... project.energies> <Pseudo.Potentials ... Pseudo.Potentials> <density.PCC ... density.PCC>
and from *.pao the following information:
AtomSpecies grid.num.output <valence.charge.density ... valence.charge.density> radial.cutoff.pao PAO.Lmax PAO.Mul <pseudo.atomic.orbitals.L=* ... pseudo.atomic.orbitals.L=*>
It might be possible to generate a mixed pseudo-atom by properly mixing these information of arbitrary two atoms, being worth considering.
Regards,
TO
|