Dear Kirill,

I must confess I am having some problems understanding the details of your issue. I agree that one would expect a uniform splitting of the spin up/spin down states as the first effect of the B-field. However, what exactly are you showing in your plot?
- Is this E_up-E_down for four bands as a function of the B-field?
- Since you have for x=0 (B-field=0?) a splitting you have a magnetic system, what is this in detail?
- Is this the splitting induced by switching on the B-Field and then doing a single iteration or is this the splitting after full SCF for each field situation?
- Is the magnetic moment increased or decreased by the field?

Best regards
Daniel

Dear Kirill,

thanks for the clarification. After looking into the code I concluded that there might indeed be a bug regarding the treatment of the interstitial. I just committed a fix to the 'develop' branch which hopefully :-) fixes the issue.
I would like to point out that however, that I do not expect the effect to be so straight forward in the fully SCF case, instead the response of the density to the field will add an additional splitting that no longer needs to be the same for all k-points, so I would recommend to check the effect of the fix for a single shot calculation first. You might be right that the difference is small :-)

Hope this helps.

Daniel

Dear Kirill,

Next try :-) I looked into this again and realized we operated on the wrong variable with the b-field. Hope my second fix is now correct.

Please also be aware that the contribution of the B-field to the total energy is not included correctly as the density-potential integrals that enter into the
total energy are calculated before the modification of the potential. So probably some term M\timesB is missing.

Daniel

Dear Kirill,

this starts to be slightly embarrassing :-). Gregor pointed out that there might be another problem, this time in the MT-spheres. I committed a fix that I hope will now solve this problem.

Hope this now really fixes the issue.

Daniel

Just for clarification, these problems are probably due to a refactoring that moved the call to the subroutine adding the B field to a different location. At this other location we have different conventions of how certain quantities are stored. Unfortunately we seem to have a test gap for the addition of external B fields and thus this problem was not detected by us. The addition of external B fields is also used very seldomly. We have to add a test case for this.

Thank you a lot for pointing out these problems!

Yes, please attach your input. Up to now these were quick tries to fix the problem without any test. I will investigate this in more detail.

I committed a fix to the develop branch. I think it will work now. At least the tests looked promising.

Dear Dongwook,

from what you describe (or I understand) your results should be fine. What is "missing" in our treatment of a Zeeman field is a contribution to the total energy. The total energy has basically two contributions. a) the sum of the eigenvalues and b) direct integrals over products of the charge densities and the potentials. As we include the Zeeman field in the potential only in a "post processing" step these (b) terms are not evaluated. But this is relevant for the total energy only, for example the magnetization - you are interested in- is evaluated from the eigenvectors which are calculated taking the Zeeman field into account. The missing term in the total energy does not change anything in the self-consistency. Actually, as long as the Zeeman field you add is constant in space also the missing M.B term can be evaluated "by hand" from the output ;-)

Hope this helps
Daniel