[Pw_forum] Negative phonon frequencies at gamma point

[Marton]

Hi Pablo,

Upon a quick glance your inputs look reasonable to me (although your pw.x
convergence thresholds are really really unreasonably too tight) so this is
just a wild guess but it could happen that the relaxation of the cluster
converged to a saddle point of the potential energy surface instead of a
local minimum (maybe due to the high initial symmetry of your structure).

If this is indeed the reason you could try to displace the cluster along
the mode with imaginary frequency to gently push it down the hill. Another
approach would be to do a short MD or an even simpler approach would be to
apply random and small displacements to the initial atomic positions just
to remove the high symmetry. I'm not sure if it is related but there are
hydrogens on the surface of your cluster that repel each other: it was
suggested in some early papers that surface reconstruction (=removal of
those hydrogens and formation of long Si-Si bonds) could be favorable on
these tiny clusters, too (similar to the bulk case).

Again, I'm not sure if this is indeed the reason but maybe it is worth
trying!

Cheers,
Marton Voros

--
Materials Science Division
Argonne National Laboratory

On Fri, Apr 14, 2017 at 8:47 AM, Pablo García Risueño <
garcia.risueno at gmail.com> wrote:

> Dear Espresso community
>
> I come to you because I find a problem that I cannot solve. I am running
> phonon calculations in different Si-H clusters. Some of them (the smallest
> ones) work well, but for the rest I always find negative frequencies beyond
> the 6th frequency (at q=0, I deal with isolated systems).
>
> I obtain very nice relaxations (e.g. <<  Total force =     0.000093 >>)
> but ph.x gives some negative frequencies beyond the 6th one:
>
>      freq (    6) =      -2.042467 [THz] =     -68.129354 [cm-1]
>      freq (    7) =      -2.040393 [THz] =     -68.060182 [cm-1]
>      freq (    8) =      -2.040381 [THz] =     -68.059778 [cm-1]
>      freq (    9) =       0.396144 [THz] =      13.213941 [cm-1]
>
> When I run dynmat.x, the results get worse instead of improving:
>
>   10   -134.72   -4.0388    0.0499
>    11   -134.72   -4.0388    0.0499
>    12    -77.44   -2.3217    0.0000
>    13    -77.44   -2.3217    0.0000
>    14    -76.18   -2.2838    0.0000
>    15    -76.18   -2.2838    0.0000
>    16    -76.18   -2.2838    0.0000
>    17    -71.69   -2.1492    0.0201
>    18    -71.69   -2.1492    0.0201
>    19    -71.69   -2.1492    0.0201
>    20    -58.72   -1.7604    0.0000
>    21    -52.87   -1.5849    0.0000
>    22    -45.50   -1.3642    0.0000
>    23    -45.50   -1.3642    0.0000
>    24    -45.50   -1.3642    0.0000
>    25    -38.77   -1.1622    0.0000
>    26    -38.77   -1.1622    0.0000
>    27      2.10    0.0628    0.0103
>    28      2.10    0.0628    0.0103
>    29      2.10    0.0628    0.0103
>    30     31.92    0.9568    0.0000
>
>
> I have tried many recommendations that I found at the forum (increase the
> box size, reduce the conv_thr, try other pseudopotentials...) but
> everything failed. Could you give me a clue on how this problem could be
> solved? I am attaching a document with the input files and part of the
> output informations.
>
> Thank you very much. Best regards.
>
> --
> --
>
> Dr. Pablo García Risueño
>
> Institut für Physikalische Chemie, Universität Hamburg, Grindelallee 117,
> 20146 Hamburg
>
> Tel. +49 040 42 83 84 82 7
>
> _______________________________________________
> 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/20170416/a25e70ee/attachment.html>