[QE-users] phonon dispersion

Thu Jun 5 17:38:25 CEST 2025

Hello,

to add a bit on Paolo's answer, as this is quite a common problem. If 
your materials is metallic or almost metallic at the DFT level that you 
are using, you have to treat it as such when doing a phonon calculation, 
or the dielectric response will not converge to a meaningful result, and 
give bad dielectric matrix and/or effective charges.

In particular, in your case, InN is a bad metal in PBE, i.e. it has a 
"negative gap" with a pocket of electrons. This pocket may give random 
results depending on the number of k-point, you definitely cannot ignore 
it. The fact that you compute a Dielectric tensor of >54'000 should have 
alerted you that something non-physical was going on (the code could 
also issue a warning for such large value, but then there are highly 
polarizable materials with very large epsilon).

If you really need it to be an insulator, you need to find a way to open 
the gap.

hth

p.s.  I'm not sure exactly why this breaks the phonon calculation, there 
is a sum whose cutoff that depends on the average value of epsilon, 
which may not be a great choice when it is so anisotropic.

On 05/06/2025 16:57, Paolo Giannozzi wrote:
> Hi
>
> I don't know the reasons for the weird behavior with v.7.3. If you 
> have split your calculation into partial one, that could be the reason 
> (sometimes the "splitting" machinery does not work as expected).
> I could reproduce your results with v.6.4.1, though. If you look at 
> the output (or dynamical matrix file) for q=0, you will notice very 
> large values for the dielectric tensor along xx and yy. This indicates 
> a quasi-metallic character of your system, presumably as a consequence 
> of DFT shortcomings, spin-orbit interactions absence, or both. If you 
> replace in the dynamical matrix files the dielectric tensor with the 
> experimental one (exx = eyy = 14.4, ezz = 13.1) and re-compute the 
> force constant matrices, the anomalous behaviour at q => 0 disappears.
>
> Thanks to Pietro Delugas and Lorenzo Paulatto for help on this subject.
>
> Paolo
>
> On 01/06/2025 01:07, Ivan Dario Arellano Ramirez wrote:
>>
>> Hi QE users
>>
>> Recently, I performed phonon calculations using the ph.x module 
>> implemented in QE. I initially used the version 6.4.1 and obtained 
>> the results shown in the attached figure (top). I noticed an 
>> unusual behavior near the gamma point, one of the acoustic branches 
>> exhibits a U-shaped anomaly  with some negative frequencies (small 
>> values). The acoustic modes should approach the Gamma point linearly.
>>
>> After some time, I decided to repeat the calculations using identical 
>> input file parameters, but on a different machine and with QE version 
>> 7.3. In this version, the band structures shows oscillations around 
>> both the gamma and A points.
>>
>> Could anyone help me understand the origin of these two issues:
>> 1) in the top figure, v.6.4.1 the anomalous U-shape of one of the 
>> acoustic branches near the gamma point.
>> 2) In v. 7.3, the oscillations around the gamma and A points.
>>
>> Any insights or suggestions would be appreciated
>>
>> Here are the ph.in <http://ph.in>, q2r.in <http://q2r.in>, matdyn.in 
>> <http://matdyn.in>
>>
>> Phonons of InN at G!=0
>> &inputph
>>   tr2_ph = 1.0d-14,
>>   ldisp = .true.,
>>   search_sym = .false.,
>>   prefix = 'inn-ph',
>>   recover = .true.
>>   fildyn = 'inn-ph.dyn',
>>   nq1 = 5,
>>   nq2 = 5,
>>   nq3 = 3,
>>   amass(1) = 114.818,
>>   amass(2) = 14.007,
>>   outdir =  './tmp'
>>   /
>>
>> &INPUT
>>   fildyn = 'inn-ph.dyn',
>>   zasr = 'crystal'
>>   flfrc =  'inn-ph.fc'
>>   /
>>
>> &input
>>   asr = 'crystal',
>>   flfrc = 'inn-ph.fc',
>>   flfrq = 'inn.band.freq'
>>   flvec = 'inn.band.modes'
>>   amass(1) = 114.818,
>>   amass(2) = 14.007,
>>   q_in_band_form = .true.
>>   /
>> 8
>>     0.0000000000     0.0000000000     0.0000000000  40  ! Gamma
>>     0.5000000000     0.0000000000     0.0000000000  25  ! M
>>     0.3333333333     0.3333333333     0.0000000000  40  ! K
>>     0.0000000000     0.0000000000     0.0000000000  25  ! Gamma
>>     0.0000000000     0.0000000000     0.5000000000  40  ! A
>>     0.5000000000     0.0000000000     0.5000000000  25  ! L
>>     0.3333333333     0.3333333333     0.5000000000  40  ! H
>>     0.0000000000     0.0000000000     0.5000000000  1   ! A
>>
>> Regards,
>>
>> -- 
>>
>> Iván Darío Arellano Ramírez
>>
>> Profesor Titular Programa de Ingeniería Física
>>
>> Director Grupo de Investigación en Modelado y Simulación Computacional
>>
>> Universidad Tecnológica de Pereira
>>
>>
>> El contenido de este mensaje y sus anexos son únicamente para el uso 
>> del destinatario y pueden contener información  clasificada o 
>> reservada. Si usted no es el destinatario intencional, absténgase de 
>> cualquier uso, difusión, distribución o copia de esta comunicación.
>>
>> _______________________________________________________________________________ 
>>
>> The Quantum ESPRESSO Foundation stands in solidarity with all 
>> civilians worldwide who are victims of terrorism, military 
>> aggression, and indiscriminate warfare.
>> -------------------------------------------------------------------------------- 
>>
>> Quantum ESPRESSO is supported by MaX (www.max-centre.eu)
>> users mailing list users at lists.quantum-espresso.org
>> https://lists.quantum-espresso.org/mailman/listinfo/users
>
-- 
Dr. Lorenzo Paulatto
IR IMPMC - CNRS UMR 7590 / Sorbonne Université / MNHN
phone: +33 (0)1 442 79822 / telegram: lpaulatto
http://www.impmc.upmc.fr/~paulatto/ - https://anharmonic.github.io/
23-24/423 B115, 4 place Jussieu 75252 Paris CX 05
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20250605/d8921a10/attachment.html>