[Pw_forum] improper rotation in the non-collinear case in EXX routine

[H. Lee]

Hi, everyone:

I have one question on the exxinit subroutine in exx.f90:

It seems that in this subroutine and the exx_set_symm subroutine, all
symmetries including inversion (improper rotation) are not explicitly
considered in the non-collinear case.
That is, the command, abs(index_sym(ikq)) is used in both subroutines,
thus, only corresponding symmetries without inversion are considered.
(In these subroutines, the symmetry including inversion is denoted by the
minus index.)

In collinear case, the following line is added to consider the effect of
inversion:
IF (index_sym(ikq) < 0 ) psic(1:nrxxs) = conjg(psic(1:nrxxs))

However, in the non-collinear case, I can't find such lines needed to
consider inversion symmetry.

I think that since inversion symmetry has no effect in 2x2 spin-rotation
matrix, it is necessary to consider inversion symmetry in 3x3 rotation
matrix.

Do I miss something?

Sincerely,

Hyungjun Lee
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20161007/2b26464c/attachment.html>