forceTheorem makes wrong estimations of spin spiral energy
The force theorem calculation of spin spiral dispersion seems to not work properly.
Let us consider a fcc bulk Fe structure.
The problem is that when I run a forceTheorem calculation with beta=0.000 I expect to see no energy difference between different q-vectors since there is no actual spin rotation (SQA || z). However, I obtain such results:
| q-vector | Energy, eV |
|---|---|
| 0.000 0.000 0.000 | 1.87833122735693 |
| 0.125 0.125 0.000 | 0.496156085857061 |
| 0.250 0.250 0.000 | 0.0921295896478881 |
| 0.375 0.375 0.000 | 0.0 |
| 0.500 0.500 0.000 | 0.0487845728566691 |
| 0.000 0.000 0.000 | 1.87833122735693 |
| 0.125 0.125 0.000 | 0.496156085857061 |
e.g. it is reproducible in a single forceTheorem loop. ForceTheorem input: inp.xml
The same calculation using the OLD FLEUR gives correct difference in energy between q-vectors:
| q-vector | sum of eigenvalues, Htr |
|---|---|
| 0.000 0.000 0.000 | -735.499128072800 |
| 0.125 0.125 0.000 | -735.499128072800 |
| 0.250 0.250 0.000 | -735.499128072800 |
| 0.375 0.375 0.000 | -735.499128072800 |
| 0.500 0.500 0.000 | -735.499128072800 |
ForceTheorem input for q=0.125 0.125 0.000: inp_old nocoinp_old