[QE-users] Absolute Magnetization diverging

[Nicola Marzari]

Dear Miles,


a few comments:

1) there are no (or not yet, to be precise) fully relativisitc 
pseudopotentials in the SSSP and the Materials Cloud - so nothing that 
has been tested and recommended.

2) you mention "increasing the mixing beta to 0.9"; but that's exactly 
the opposite of what you should do, i.e. use very small mixing betas, 
even 0.1 or 0.05, and play around with the different Thomas-Fermi 
approaches. If you are new to these kind of calculations, you might need 
to consider spending some time learning the basics (a very minimal intro 
is here: 
https://www.materialscloud.org/learn/sections/spVa3j/fireside-chats-for-lockdown-times)

3) maybe try collinear calculations first, to make sure you have a 
handle on running calculations, and double check your inputs against the 
QE input generator (https://www.materialscloud.org/work/tools/options) 
to make sure you are in the right ballpark.

HTH

nicola


On 14/02/2025 04:09, Johnson, Miles R. wrote:
> Hello,
> 
> I'm trying to use the PSlib pseudopotentials for an scf calculation on 
> an antiferromagnetic material, including spin-orbit, as the materials 
> cloud database and quantum espresso website seem to imply these are the 
> best. Unfortunately, the absolute magnetization has immediately diverged 
> in every computation I've tried, and I'm not sure how to stop it.
> 
> Using the pseudodojo pseudopotentials it converges to around 6 bohr/unit 
> cell. With the PSlib pseudopotentials even by the first or second 
> iteration it is at like 1000 bohr/unit cell, and never recovers from 
> there. I've tried setting startingwfc to 'atomic', and this makes the 
> first iteration have a reasonable absolute magnetization, but then in 
> the second iteration it already increases to like 100 bohr/unit cell and 
> diverges from there. Increasing the mixing beta to 0.9, it still 
> diverges very quickly.
> 
> The total energy also lowers by hundreds of Ry as the magnetization 
> diverges. This happens with both the fully reletavistic PBE and PBEsol 
> pseudopotentials. I've also tried modifying input_dft a bit, but the 
> functionals that I've tried either aren't compatible with noncolinear 
> spin, or haven't changed the divergence.
> 
> Does anyone have tips for why this might be diverging so badly, and how 
> to get it to converge to a reasonable ground state? I realize I can use 
> the pseudodojo potentials, but I'm thinking my bands might not be 
> correct so I want to compare how they look for different pseudopotentials.
> 
> Best,
> Miles Johnson
> PhD Candidate in Applied Physics
> California Institute of Technology
> 
> 
> 
> 
> 
> _______________________________________________________________________________
> The Quantum ESPRESSO Foundation stands in solidarity with all civilians worldwide who are victims of terrorism, military aggression, and indiscriminate warfare.
> --------------------------------------------------------------------------------
> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
> users mailing list users at lists.quantum-espresso.org
> https://lists.quantum-espresso.org/mailman/listinfo/users


-- 
----------------------------------------------------------------------
Prof Nicola Marzari, Chair of Theory and Simulation of Materials, EPFL
Director, National Centre for Competence in Research NCCR MARVEL, SNSF
Laboratory Head, PSI Center for Scientific Computing, Theory, and Data
Contact info and websites at http://theossrv1.epfl.ch/Main/Contact