EPW Forum

Dear developers and users,
Hi, I'm using EPW for a while and I think it's really a good code. Thanks for all the contribution.
While I'm using it, I feel a little confused by some output. I hope I can get some advice from you guys.
Here's the problem. I want to know the exact number of the electron phonon coupling matrix element (4*4*4 coarse mesh). I put "prtgkk = true" in epw.in file and nkf=4 nqf=4 (just the same with coarse mesh). And I can get the electron phonon coupling matrix element |g| (meV) in epw.out.

But the results confused me: why there are so many zero at some k point in epw.out? (shown bellow)

at this point, the results look ok:
iq = 2 coord.: 0.0000000 0.0000000 0.2500000
ik = 1 coord.: 0.0000000 0.0000000 0.0000000
ibnd jbnd imode enk[eV] enk+q[eV] omega(q)[meV] |g|[meV]
------------------------------------------------------------------------------
6 6 1 6.3656 2.9061 6.8310 0.1515486919E+02
6 6 2 6.3656 2.9061 6.8310 0.1515486919E+02
6 6 3 6.3656 2.9061 18.7931 0.2598448010E+02
6 6 4 6.3656 2.9061 45.7747 0.3717149108E+02
6 6 5 6.3656 2.9061 45.7747 0.3717149108E+02
6 6 6 6.3656 2.9061 49.8697 0.9199567240E+02

but at this point, the results look not that ok: (the |g| of 1,2,4,5 modes are zero)
iq = 2 coord.: 0.0000000 0.0000000 0.2500000
ik = 2 coord.: 0.0000000 0.0000000 0.2500000
ibnd jbnd imode enk[eV] enk+q[eV] omega(q)[meV] |g|[meV]
------------------------------------------------------------------------------
6 6 1 2.9061 0.6240 6.8310 0.4728393878E-05
6 6 2 2.9061 0.6240 6.8310 0.4728393878E-05
6 6 3 2.9061 0.6240 18.7931 0.5845905745E+02
6 6 4 2.9061 0.6240 45.7747 0.2930626511E-05
6 6 5 2.9061 0.6240 45.7747 0.2930626511E-05
6 6 6 2.9061 0.6240 49.8697 0.1640789419E+03

even worse, at this point, all the output is zero:
iq = 6 coord.: 0.0000000 0.2500000 0.2500000
ik = 1 coord.: 0.0000000 0.0000000 0.0000000
ibnd jbnd imode enk[eV] enk+q[eV] omega(q)[meV] |g|[meV]
------------------------------------------------------------------------------
6 6 1 6.3656 3.1230 9.8251 0.0000000000E+00
6 6 2 6.3656 3.1230 9.8251 0.0000000000E+00
6 6 3 6.3656 3.1230 19.1482 0.0000000000E+00
6 6 4 6.3656 3.1230 44.5027 0.0000000000E+00
6 6 5 6.3656 3.1230 44.5027 0.0000000000E+00
6 6 6 6.3656 3.1230 51.0110 0.0000000000E+00

This is a AlAs system. But I found similar situation in other systems (like MgB2 or diamond etc)

As far as I'm concerned, I feel confused by the zero output. Is there any rule to determine this that I didn't know? Or is there anything I didn't understand correctly? I'm looking forward to your reply. Thanks a lot.


Yaokun Ye
A Ph.D student.

Dear Yaokun Ye:

First, it might not be directly related to your issue, but there was a bug in the subroutine for prtgkk of the old versions of EPW which I fixed it. So unless you did, I would suggest you to use the recent versions of EPW such as v5.4.

Returning to your main question:
It can be a bug, but it might not be a bug; something like symmetry, etc. can enforce it.
So the best way to check whether it is a bug is to calculate the e-ph vertices on the same pairs of k and q points using DFPT (ph.x code) and compare them with those from EPW.

Sincerely,

H. Lee

Dear Lee,
Thanks for your kindly reply. I'll try your suggest.
Best wishes.

Yaokun

hlee wrote: ↑Wed Aug 11, 2021 3:51 pm Dear Yaokun Ye:

First, it might not be directly related to your issue, but there was a bug in the subroutine for prtgkk of the old versions of EPW which I fixed it. So unless you did, I would suggest you to use the recent versions of EPW such as v5.4.

Returning to your main question:
It can be a bug, but it might not be a bug; something like symmetry, etc. can enforce it.
So the best way to check whether it is a bug is to calculate the e-ph vertices on the same pairs of k and q points using DFPT (ph.x code) and compare them with those from EPW.

Sincerely,

H. Lee

Dear Lee,
Sorry to bother you again. I tried your suggest and I found EPW got almost the same results as QE ph.x.
But I'm still kind of confused. Like you said, something like symmetry can enforce it to be zero. I can't understand this point.
Using the formula of the paper by Prof. Giustino (2017, Review of Modern Physics), I can't see how can the ele-phon matrix be zero, even by adding symmetry or something. I really appreciate it if you can help me with more details to understand this (I'm sorry maybe this question sounds kind of "stupid" to you)
Looking forward to your reply. Thanks a lot.

Yaokun

I've figured out this problem. Thanks for all the help.