Why sign on dopant atoms ( for coupling study) is same in case of AFM state?

[Avaneesh]

Hi,

During Spin polarized coupling calculation of  doped nanowires, i found that local magnetic moment on both dopant atom is positive (up - down spin) for both AFM and FM state. While from literature survey local magnetic moment for AFM state is of opposite sign. Why is this so?

Is this is just a sign convention used for AFM state, when it is more stable than FM state or vice versa?

[Jess Wellendorff]

No, there should be no funny sign convention at play here.
Did you do a Mulliken Population Analysis to see the local magnetization?
Also, are you sure that your AFM calculation has actually converged to an anti-ferromagnetic electronic structure?

[Avaneesh]

Thanks for reply.

I have calculate the Mulliken Population to analyze the local magnetization on atoms and structure is also converged with the forces all less than the given tolerance limit.

Here AFM calculation  is done for unpassivated doped InP nanowire. In which dopant atoms (Mn) assign with opposite spin state, while other atoms (In and P) assign with +1 spin state. Is i am doing wrong?

Pls help me to find a solution

[Jess Wellendorff]

OK, so your initial spins are -1 and +1 for different Mn atoms, and +1 for In and P. But that is just your initial guess for the relative spin polarization on those atoms. What matters in the end is the spin structure that the calculation converges to during SCF. Perhaps the AFM state is "lost" because it is not energetically favorable. In that case the ground state you obtain after the SCF has converged may actually be a FM state, even though you started out from an AFM state.

So is the Mulliken Population Analysis for your AFM calculation consistent with an AFM state? Or does it look like an FM state?

[Avaneesh]

So, it means that the structure of nanowire which i choose here for coupling study  is not energetically favorable for AFM state.  While its  total energy diff. (AFM-FM < 0) and Mulliken population report shows different behavior.

Is this due to wrong calculation of DFT or parameter (mesh cut-off, k points etc.) used for calculation  (however initially, i had checked  its accuracy by calculating bulk material parameter like bandgap and lattice  parameter, which is consistent with previous reported works)?   

[Jess Wellendorff]

I have not seen your calculation, so I do not know if the AFM state is stable or not. I just mentioned it as a possibility. You say that AFM-FM energy differences and Mulliken report shows different behavior. What does this mean? I think I will need to see the python scripts and text outputs from your FM and AFM calculations before I can help you further.