Exchange matrix decomposition.

[adrybakov]

Hello!

Last year I've asked the question about the matrix decomposition into isotropic, anisotropic and dmi parts on the google forum. Lately I noticed that TB2J produces correctly decomposed matrices, however, there is nothing written about it in the release notes.

Is the problem resolved or am I just lucky with a particular system?

Best,
Andrey

[mailhexu]

I think it is just for the particular system. I am now working on an alternative way to compute/merge the DMI and anisotropic exchanges. Not sure if the anisotropic J will be traceless though. More tests needed.
Best wishes,
HeXu