Extracting phonon velocity

Running EPW
1

Dear EPW users,

I am using EPW-5.4.1. To extract phonon velocities of Si (I am actually interested in the +/- sign of the velocity not the correct magnitude), I have added the following to the ephwann_shuffle.f90:

CALL vmewan2blochp(xxq, nmodes, nrr_q, irvec_q, ndegen_q, uf, vmefp(:, :, :), wf(:, qind), rws, nrws)
DO imode = 1, nmodes
      vmeq_av(:) = zero
      DO jmode = 1, nmodes
          vmeq_av(:) = vmeq_av(:) + REAL(vmefp(:, jmode, jmode), KIND = DP)
      ENDDO
          vmeqf(:, imode) = vmeq_av(:) 
ENDDO
write(99,'(6E26.16E3)') vmeqf(1,1), vmeqf(1,2), vmeqf(1,3), vmeqf(1,4), vmeqf(1,5), vmeqf(1,6)

I ran the code on 100 q-points and for all the q-points, all the 6 x-components of velocity are identical. Different q-points have different velocities but each q-point has the same velocity for acoustic and optical modes. The routines in the latest EPW seems to be similar. Is this a bug in the code?

Thanks,
Vahid

Vahid Askarpour
Department of Physics and Atmospheric Science
Dalhousie University
Halifax, Canada

2

The velocities of the modes can be obtained by simply doing the following single loop:

CALL vmewan2blochp(xxq, nmodes, nrr_q, irvec_q, ndegen_q, uf, vmefp(:, :, :), wf(:, qind), rws, nrws)
DO imode=1,nmodes
vmeqf(:, imode)=REAL(vmefp(:, imode, imode), KIND = DP)
ENDDO

I was averaging the six modes before so they were all equal to average.

Vahid