paw+gw with spin-orbit  [SOLVED]

GW, Bethe-Salpeter …

Moderators: maryam.azizi, bruneval

Locked
raul_l
Posts: 74
Joined: Sun Jan 08, 2012 7:45 pm

paw+gw with spin-orbit  [SOLVED]

Post by raul_l » Fri May 31, 2013 7:18 pm

Is it possible to include spin-orbit coupling in GW calculations? I tried pawspnorb 1 but the screening calculation stopped with

Code: Select all

 chkint_prt: ERROR -
  Context : the value of the variable ecut is  16.
  The value of the input variable nspinor is    2, while it must be
  equal to    1
  Action : you should change the input variables nspinor or ecut.

 chkint_prt: ERROR -
  Context : the value of the variable ecut is  16.
  The value of the input variable nspinor is    2, while it must be
  equal to    1
  Action : you should change the input variables nspinor or ecut.

 Subroutine Unknown:0:ERROR
 
 chkinp: ERROR -
 Checking consistency of input data against itself gave    1 inconsistencies.

 Subroutine Unknown:0:ERROR

which I sort of expected because as far I know spin-dependent self-energy isn't implemented yet. However, is it possible to "use the vector part of the pseudopotential to treat spin–orbit as a perturbation on top of the GW results" like they do in Malone et al., J. Phys.: Condens. Matter 25, 105503 (2013), except with PAW datasets? Or is there some other way to include SO coupling?
Raul Laasner
Netherlands Institute for Space Research

rangel
Posts: 45
Joined: Tue Aug 18, 2009 9:50 pm

Re: paw+gw with spin-orbit

Post by rangel » Wed Jun 05, 2013 12:29 pm

Not in ABINIT.
You could add the SO term to the hamiltonian with an external code, but this requires some coding.
Tonatiuh Rangel

raul_l
Posts: 74
Joined: Sun Jan 08, 2012 7:45 pm

Re: paw+gw with spin-orbit

Post by raul_l » Wed Jun 05, 2013 8:02 pm

OK. Then I will include the spin-orbit induced shift only in the band gap value, assuming it is the same at DFT and QP level.
Raul Laasner
Netherlands Institute for Space Research

hammouri
Posts: 57
Joined: Fri Feb 13, 2015 6:44 am

Re: paw+gw with spin-orbit

Post by hammouri » Sat Feb 17, 2018 6:04 pm

raul,
how would you add the soc effect to the gw calculations?
And the way, does the new version of abinit supports GW+SOC?

Locked