My Community

Dear all,

For nonmagnetic systems like Al and Si, we set ISPIN=1. For FM and AFM systems below their critical temperatures like Fe and NiO, we set ISPIN=2. But for Fe and NiO ABOVE their critical temperatures, how to set ISPIN? 1 or 2?

Thank you very much!

Hi

In my opinion, ISPIN doesn't have anything to do with being below or above the critical temperature. As soon as the system you are considering is magnetic, you need to set ISPIN=2.

Then, you need to set the correct magnetic moments (using the MAGMOM tag) in order to reproduce the magnetic ordering.

What you're trying to say, I presume, is that above the critical temperature, systems can be paramagnetic. For paramagnetism, I think the best choice is to set ISPIN=1, for DFT calculations. For DFT+U or hybrid calculations, where electrons are localized, then I think it is impossible to model paramagnetism. But I may be wrong.

Regards.

<span class='smallblacktext'>[ Edited Mon Sep 06 2010, 12:33PM ]</span>

There is always issues to model phases at high finite temperatures from T = 0 K calculations. If you truly want to model the paramagnetic phase I would say that neither ISPIN = 1 nor ISPIN = 2 is an approriate choice. To use the nonmagnetic solution as an approximation for the paramagnetic state should definitely be questioned before usage, since its validity and accuracy for the bulk properties of the paramagnetic phase will depend on the specific system. I would rather look through the literature and consider approaches such as the Disorder Local Moment model or alternatives.

Cheers,
/Dan

<span class='smallblacktext'>[ Edited Mon Sep 06 2010, 01:30PM ]</span>

I agree with Dan. ISPIN=1 remains definitely questionable for the modelling of paramagnetic phases.

Thank you, boris and Dan.