How to estimate Curie temperature by VASP?

Message

spakinfo · #1 Post by **spakinfo** » Wed Nov 11, 2009 6:52 am

Hi,
Does anybody know a practical way to estimate Tc (Curie temperature) by VASP for a ferromagnetic system? Suppose I have a dilute magnetic semiconductor, is it correct to assume that Tc is proportional to energy difference between FM and AFM configurations? Is there a good mean-field approach (applicable in VASP) to this problem

I'd be grateful if somebody could help me out.

rgc · #2 Post by **rgc** » Wed Nov 11, 2009 1:20 pm

I think it is correct to assume that Tc is proportional to EFM-EAFM if you have only one possible AFM configuration (only one type of coupling, associated to a coupling constant J). In that case a mean field approach will give you Tc ~ J ~ EFM-EAFM.

I've done something similar for the Curie-Weiss temperature in an AFM system, not sure if that would help you, but this is the reference: J. Mater. Chem., 2003, 13, 2848â€“2850.

http://www.rsc.org/publishing/journals/ ... i=b309796j

spakinfo · #3 Post by **spakinfo** » Thu Nov 12, 2009 8:55 am

Thanks Recardo for your comment and paper.
It seems that your approach does give Tc values near experimental data. However, the estimated values tend to overestimate the experimental Tc. For instance, for BCC Fe I get Tc>1500 which is quite larger than the experimental value 1040 K. Any suggestion to improve the Tc results?