negative phonon frequencies near q=0

[jyxi]

Dear Dr. Samuel Ponc
Thanks for your reply. According to your suggestion, we recalculated the EPW-interpolated phonon dispersion by QE-6.2.1 and EPW-4.3.1 with asr_typ = crystal, however, there is no difference between crystal and simple , and the phonon frequencies are still negative as well as no degenerate between ZA and TA mode near q=0. Is there another parameter for affecting phonon frequencies ? Thank you.
Best,

Jinyang Xi

.....
......
0.022919 0.022919 0.022919
-0.2237 -0.2237 2.2062 21.4907 21.4907 21.8887 [ in unit of mev]
0.018335 0.018335 0.018335
-0.2877 -0.2877 1.7607 21.4788 21.4788 21.7392
0.013751 0.013751 0.013751
-0.2646 -0.2646 1.3178 21.4680 21.4680 21.6171
0.009167 0.009167 0.009167
-0.1972 -0.1972 0.8771 21.4595 21.4595 21.5266
0.004584 0.004584 0.004584
-0.1045 -0.1045 0.4381 21.4541 21.4541 21.4710
0.000000 0.000000 0.000000
0.0000 0.0000 0.0000 21.4522 21.4522 21.4522
0.000000 0.009167 0.000000
0.1514 0.1989 0.5383 21.4523 21.4539 21.4776
0.000000 0.018335 0.000000
0.4030 0.5350 0.9813 21.4532 21.4587 21.5517
0.000000 0.027502 0.000000
0.6170 0.8684 1.4324 21.4545 21.4660 21.6717
0.000000 0.036670 0.000000
0.8442 1.1878 1.8878 21.4556 21.4748 21.8328
......
......

[sponce]

Dear Jinyang Xi,

First, are you sure you correctly used the crystal sum rule? It does work differently from the simple one.

You need use q2r.x to generate the IFC matrix from the dyn files produced by QE.
Then you need to place it inside the save folder.
After that you can run an EPW calculation and specify

asr_typ = 'crystal'
lifc = 'true'

If forgot to tell you about the lifc in the email.

If this does not help and you still get the same phonon frequencies, then it might be a problem of Wannierization.
In that case can you plot the various decay files and make sure that they decay properly.

Best,
Samuel