[Pw_forum] a few frequencies in PHONON calculation

Andrea Dal Corso dalcorso at sissa.it
Sat Dec 26 16:12:42 CET 2015


Actually the dynamical matrix is block diagonal in the basis of the  
modes chosen to compute the perturbation and the size of the blocks is  
the number of modes that transform according to the same irreducible  
representation of the small point group of the q vector.

If you look at the symmetry of the modes chosen by the ph.x code to  
compute the perturbation you will notice that they are grouped in such  
a way that there are first all the modes that transform according to  
representation 1,
then all the modes that transform according to  the representation 2,
etc. If you choose start_irr and last_irr so that all the modes of  
representation 1 are calculated and then force the diagonalization of  
the dynamical matrix, the frequencies of the modes that belong to  
representation 1 will be correct.
If your start_irr, last_irr include also some modes of representation  
2 but not all of them, the frequencies of the modes of representation  
2 will be wrong and change when you change last_irr.

Unfortunately if your system has no symmetry then all modes will  
transform according to the representation 1 and in order to have  
correct frequencies you will have to diagonalize the complete  
dynamical matrix, but if you have some symmetry it will be possible to  
select only some modes.

Moreover this trick cannot be used for q points at Brillouin zone  
border for nonsymmorphic space groups, because the code is not able to  
find the representations of these modes.

HTH,

Andrea

Quoting Lorenzo Paulatto <lorenzo.paulatto at impmc.upmc.fr>:

> Dear Eduardo,
> The frequency are the square roots of the eigenvalues of the dynamical
> matrix, it is not possible to compute just a few eigenvalues without having
> the whole matrix.
>
> Kind regards
>
> ▼ Hide quoted text
> On 22 Dec 2015 6:55 pm, "Eduardo Menendez" <eariel99 at gmail.com> wrote:
>
>> Hi,
>> I have one doubt about the PHONON code. Is it possible to compute just a
>> few phonon frequencies using the PHONON keywords last_irr, ldiag ? For
>> example, using
>>
>> last_irr = 10,
>>   ldiag=.true.,,
>>
>> may one find the frequencies of just the first irreducible representations
>> ?
>>
>> Thank you,
>>
>> Eduardo Menendez Proupin
>> Departamento de Fisica, Facultad de Ciencias, Universidad de Chile
>> URL: http://www.gnm.cl/emenendez
>>
>>
>>
>> _______________________________________________
>> Pw_forum mailing list
>> Pw_forum at pwscf.org
>> http://pwscf.org/mailman/listinfo/pw_forum
>>






More information about the users mailing list