# Spin polarized calculation gives different solution in SIESTA and Quantum ESPRESSO

I was trying to optimize Vanadium doped 4x4 MoS2 using SIESTA. Plotting the band structure I was able to see the symmetry breaking at K and K'. But when I tried to do the same calculation in Quantum ESPRESSO, the first trouble I can across is that I needed to set a starting magnetization value which wasn't needed for SIESTA. Does SIESTA have a default value for it? Secondly, I tried using different values of starting magnetization for Vanadium but all of them converged to zero. SIESTA had output that Vanadium had an extra up spin.

Edit 1 : As suggested by Camps, I tried optimizing the structure using starting_magnetization value of 1 for all three atoms. Still, the system converged to a state without magnetization.

Edit 2 : I have finally got similar spin splitting i.e. up and down for Quantum espresso by adding Hubbard Potentials for the atoms. I believe this has got to do with difference in the basis sets used for SIESTA and QE. (I tried the following on smaller system where neither QE nor SIESTA showed any symmetry breaking at K and K')

• Your last edit might be better as an answer! Dec 31, 2021 at 7:29
• @NikeDattani I gave it as an edit because I thought a better answer would be one stating the actual difference between LCAO and PW basis that makes PW to require Hubbard potential for obtaining spin splitting. Jan 3 at 11:00

When running spin polarized calculations with SIESTA, the logical keyword DM.InitSpin.AF will define the initial magnetization. From the SIESTA manual (v4.1.5):

DM.InitSpin.AF false (logical)
It defines the initial spin density for a spin polarized calculation. The spin density is initially constructed with the maximum possible spin polarization for each atom in its atomic configuration. This variable defines the relative orientation of the atomic spins: If false the initial spin-configuration is a ferromagnetic order (all spins up). If true all odd atoms are initialized to spin-up, all even atoms are initialized to spin-down.

You can also use the block DM.InitSpin and define your own spin configuration for each atom. Example:

%block DM.InitSpin
5 -1. 90. 0. # Atom index, spin, theta, phi (deg)
3 +45. -90.
7 -
%endblock DM.InitSpin


This block has precedence over DM.InitSpin.AF.

• Is there any default value for this? I had not used this block in the input file for SIESTA. Dec 28, 2021 at 4:03
• The default value for DM.InitSpin.AF is false.
– Camps
Dec 28, 2021 at 16:15