# How to do spin polarization calculations using Quantum ESPRESSO?

I am currently new to DFT calculations in Quantum ESPRESSO. Though, I have read some research papers for determining spin polarization and Half-metallicity in certain Heusler alloys. These research papers shows some theoretical calculations about Density of States at fermi level and showing band structures in first brillouin zone. But if I want to predict by myself such new materials, how can I do that? For eg:Mn2CoAl.
I have basic knowledge of DFT calculation, SCF loops, Pseudopotentials and running simple calculations in Quantum ESPRESSO.
What are the steps do I have to follow for spin polarization and Half metallicity prediction in new materials?

Thanking you!

• Could you cite a paper?
– Jack
Sep 23, 2020 at 2:04
• @Jack Thanks for your concern! Here I have cited the paper. DOI: 10.1039/C9RA05212G (Paper) RSC Adv., 2019, 9, 30462-30478
– UJM
Sep 23, 2020 at 5:01
• I think it might be better if you split up the question. You can as well ask how to identify/find a half metal using DFT
– Thomas
Sep 24, 2020 at 17:26

First, I would recommend reading through the pw.scf input file description, provided here.

The relevant parameters are in the &SYSTEM namelist of the input file. To do a basic, linear spin-polarized calculation, you would need to set at least two additional parameters. If I have two types of atom, say, Fe and O, then if Fe is listed first under atomic types (it is type '1'), then the additional parameters you need to set are:

nspin = 2
starting_magnetization(1) = 0.5


If you want to change the magnetic ordering, such as the various antiferromagnetic possibilities, or in general study different magnetizations on different atoms of the same type, you need to define different atomic types for the same element, such as Fe1, Fe2, etc. So you could have

nspin = 2
starting_magnetization(1) = 0.5
starting_magnetization(2) = -0.5


Where atomic type 1 is "Fe1", and type 2 is "Fe2". Keep in mind this isn't a constraint, it's just a starting point that hopefully leads you to a certain local minimum.

There are many parameters relevant for magnetic systems. It's worth going through the documentation carefully to understand what is available (non-collinear magnetism, spin-orbit coupling, starting charge, occupation matrices with DFT+U, etc.)

I can't comment on half-metallicity prediction since I'm not familiar with the topic, but hopefully this can get you going in the right direction with starting your calculations.

• Also I want to add that starting_magnetization is strictly an initial guess. It does not have to be say the exact magnetic moment that you find in studies. But this initial guess needs to have the right sign as Kevin has shown - If you want to simulate say an AFM material, the two magnetic atoms need to have opposite signs in the input (like -0.5,0.5) but the actual magnitude does not matter as the system calculates it by itself. Sep 26, 2020 at 1:12
• True. Sometimes the magnitude of the magnetization can steer the system into different local minima though. Sep 26, 2020 at 1:46