Strange exchange interaction: Example fcc-Ni

[Dongzhe Li]

Dear Vampire experts,

I am a new user of Vampire code.

For the example of fcc-Ni (which is from the previous workshop), the exchange interaction is given as follows:

-----------------------------------------------------------------

material[1]:exchange-matrix[1]=2.757e-21

------------------------------------------------------------------

However, if I check the literature (see the PRB below), it is 0.206 mRy, if I transform it to Vampire input unit, it becomes 8.9799e-22 which is much smaller than 2.757e-21. Here, I took into account the double summation. Therefore, I get a much smaller Curie temperature than expected.

https://journals.aps.org/prb/abstract/10.1103/PhysRevB.64.174402

My DFT calculation also reproduces a similar value as reported in the PRB above. 

Can anyone point out where is the problem? Thanks in advance!

Best regards,

Dongzhe

[milton...@gmail.com]

I too am facing this problem and have never been able to use exchange values from DFT articles in Vampire, always getting a much lower Curie temperature.

Looking in the article for the design and verification of Vampire (https://doi.org/10.1088/0953-8984/26/10/103202) the key expression seems to be "multiscale modelling", and there are some articles about it like Towards multiscale modeling of magnetic materials: Simulations of FePt where they have some formula for an effective exchange constant. This is referred to in an article which Richard Evans was on (Multiscale model approaches to the design of advanced permanent magnets) where the exchange constants are instead fitted from experimental data.

Thank you for making this thread, Dongzhe. Let's post our progress on the issue here, while we wait for someone who knows.

[LeBert Sam]

The equation (12) in the article '(Multiscale model approaches to the design of advanced permanent magnets') is cited to reference [30]. Is that cited article published? If so can you provide the link.

Regards,

LeBert Sam Billgates