[Pw_forum] Issues with Phonons

[Hua Bao]

Dear Thierry Tsafack and QE users,

I actually have similar problems as Theirry. I am recently trying to
calculate phonon dispersion relation for two dimensional materials, but
always got negative frequencies.

First, I am pretty sure my structure is fully relaxed. It is also stable
based on the literature report (they use VASP for calculation).
I used norm-conserving type of pseudopotential.
The negative frequency appears near the zone-center. At zone center the
frequency is exactly zero.

I've tried to resolve the problem using the following attempts:
1. Change different ASR
2. Increase ecutrho to from 240 to 480 Ry (8 times ecut)
3. Increase ecutwfc from 45 to 90
4. Increase k-point mesh size from 16*16*1 to 32*32*1
5. Increase tr2_ph from 1e-18 to 1e-20
6. Increase conv_thr from 1e-8 to 1e-12

Since the calculations are pretty time consuming, it is difficult to try
even smaller convergence threshold.

I really hope someone can help with this issue.

Thanks,
Hua Bao






On Sat, Nov 2, 2013 at 4:03 AM, Thierry Tsafack <tr.thierry at rice.edu> wrote:

>
>
> Dear fellow QE users,
>
> I hope this email finds you well.
>
> Could you please help me fix the issues below?
>
> 1-) The executable ph.x requires the knowledge of nq1 nq2 nq3 to
> determine the density of the Monkhorst-Pack grid, i.e, number of
> points and their coordinates.  What is the linkage between nq1 nq2 nq3
> and the number of points in the grid?
>
> Everything else being equal, 3x3x1 yielded 5 points, 4x4x1 yielded 10
> points, 5x5x1 yielded 13 points, 6x6x1 yielded 20 points, 7x7x1
> yielded 25 points.
>
> For benchmark purposes, how do I know how many points nq1 x nq2 x 1
> will produce without doing the calculations?
>
>   2-) How to go about a situation in which ph.x produces w(q_i)>0 for
> all i in the above grid, but q2r.x and matdyn.x produce some w(q_i)<0,
> for all choices of the ASR or ZASR?  Assume the geometry relaxation
> was done as well as the scf calculation and you know that the
> structure is stable.
>
> 3-) Recover =.true. in QE-4.3 always fails saying:"from checkallsym :
> error".  How to go around it?
>
> Thanks in advance for your clues, if your answer necessitates further
> details, let me know.
>
>
> ~Thierry T.
> MEMS Rice University, Houston, TX
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20131103/ce8ee389/attachment.html>