[Wien] Fwd: Re: Spin-orbit coupling SCF not converging

[Peter Blaha]

It is certainly TETRA.

Your BZ is a 2D BZ. You have only ONE kz value, which means all your 
"tetrahedra" are actually triangles and this is a highly degenerate 
situation. Even worse is SO coupling, because in a non-spinpolarized 
calculation all eigenvalues are "doubled", i.e. even more degenerate.

I'm pretty sure you get even errors in the number of electrons. Try

grep ':CHA ' case.scf    (mind the blank), or
grep :NEC01 case.scf

In more recent versions of WIEN2k the init_lapw script even warns you, 
because it detects a 2D BZ and suggests to switch to TEMP.

Maybe I'll make this automatic in future.

PS: I would expect you have to do a spin-polarized calculation, 
otherwise the simulation is unrealistic.

-------- Weitergeleitete Nachricht --------
Betreff: 	Re: [Wien] Spin-orbit coupling SCF not converging
Datum: 	Tue, 26 Nov 2019 17:21:59 +0000
Von: 	Luigi Maduro - TNW <L.A.Maduro at tudelft.nl>
An: 	'pblaha at theochem.tuwien.ac.at' <pblaha at theochem.tuwien.ac.at>



Hello Peter,

According to other DFT studies a monolayer of MoS2 with zig-zag edges is 
metallic and ferromagnetic. However, I do not know if the current system 
under study is ferromagnetic. Without SOC the system under study is 
metallic: this was done with efmod set to TETRA. I am interested in how 
the SOC affects valence band splitting in this system. I will try using 
TEMP(S) with the suggested smearing. I do not understand how changing to 
TEMP(S) affects the SCF convergence though.

Thanks for the advice!

Cheers,

Luigi

PhD candidate
Kavli Institute of Nanoscience

Department of Quantum Nanoscience

Faculty of Applied Sciences

Delft University of Technology