ISYM= -1 in non collinear calculations

[priya]

Hi

I found on the forum that I should introduce the flag ISYM=-1 in the INCAR file when I run a non-collinear calculation. Could someone explain why?

I ran a calculation with U for my system for the ferromagnetic case (q=0) using the non-collinear code and the vasp version compiled for collinear calculations. I got a lower energy in the non-collinear calculation. Examining my results I found that the non-collinear calculation was converging to an insulating state, while the collinear calculation which did not have this flag (ISYM) was converging to a metallic state. Introducing this flag in the INCAR for both calculations gave me the same result in the collinear and the non-collinear cases.

So, it would help to know why ISYM is needed.

Thanks,

Priya

[admin]

ISYM = -1 is needed because when including spin-orbit interactions
(as neded for non-collinear runs), time-reversal symmetry is broken.