Re: Include spin orbit coupling but turn off magnetism ( No.1 ) |
- Date: 2016/04/24 09:39
- Name: T. Ozaki
- Hi,
An easy way for that purpose is to perform a non-spin polarized calculation within the colliear DFT, and perform one shot diagonalization including SO coupling with the charge density obtained by the first SCF calculation. Though this is known to be a second variational scheme, OpenMX does NOT support the scheme. Instead of the scheme, in OpenMX one can control the direction and magnitude of spin moment using the constraint scheme 'on2' explained at http://www.openmx-square.org/openmx_man3.8/node106.html This is a more general scheme which allows us to include SO coupling in a self-consistent manner while controlling the spin polarization within the non-colliear DFT. With 'on2', if you start the calculation with a non-spin polarized charge density, the electronic structure will be constrained so that the non-spin polarized state can be kept as much as possible depending on the magnitude of 'scf.Constraint.NC.Spin.v'.
Regards,
TO
|
Re: Include spin orbit coupling but turn off magnetism ( No.2 ) |
- Date: 2016/04/25 02:41
- Name: Hung-Yu Yang <yanghw@bc.edu>
- Dear Prof. Ozaki,
Thank you for pointing out this direction; actually I thought about using constraint to achieve this, while when I checked the manual of 3.7 version, only the constraint on the spin direction was allowed and I didn't check the newest 3.8 version. Thank you again for the helpful information.
Yours sincerely,
Hung-Yu
|
Re: Include spin orbit coupling but turn off magnetism ( No.3 ) |
- Date: 2016/04/25 08:52
- Name: T. Ozaki
- Hi,
One more tips: In OpenMX Ver. 3.8, the 'RMM-DIISH' has been newly released as explained in http://www.openmx-square.org/openmx_man3.8/node42.html The scheme applies RMM-DIIS for the KS Hamiltonian matrix, and thereby may be suitable for the plus U method and the constraint schemes. In several cases, we found that RMM-DIISH outperforms RMM-DIISK. I wonder that the scheme can be an option for your case.
Regards,
TO
|
Re: Include spin orbit coupling but turn off magnetism ( No.4 ) |
- Date: 2016/04/27 21:51
- Name: Hung-Yu Yang <yanghw@bc.edu>
- Dear Prof. Ozaki,
What I am doing really involves both plus U and constraint (iridates), and this can possibly make a difference. Thank you very much for the update.
Yours sincerely,
Hung-Yu
|