[QE-users] question about wave functions phonon code

[Jacopo Simoni]

Hi thanks for the reply. I need to compute a sort of finite temperature
electron-phonon coupling, so I needed to go into the code and I wrote my
own routine to do that, I am debugging the code and I noticed that the
result at different but equivalent q points, let's say X=(1,0,0) and
X=(1,1,0) in FCC are not the same as I should expect.
Then I noticed that the wave function (evq) in reciprocal space differs
(that is expected) but I do not understand  the relation between the two,
that was why I was asking the previous question. In the meantime I found a
mistake in the way I was computing the potential derivative, that is
probably the main source of my problem. I was wondering also what does the
routine symdyn_munu_new exactly do? Why does the dynamical matrix have to
be symmetrized in the basis of the modes ?

Thanks in advance,
Jacopo Simoni, Lawrence Berkeley National Lab.

On Mon, 27 Sept 2021 at 00:20, Lorenzo Paulatto <paulatz at gmail.com> wrote:

> Hello Jacopo,
> instead of answering your question, I may ask you what you actually want
> to do, because chances are that your problem has already been met by others.
>
> E.g. there are ways to eliminate this phase, or to neutralize it to
> compute the derivative w.r.t. the wave vector. But more often, when you
> have an observable quantity that comes from a sum over the k-points, the
> phases will cancel out in the total, if done correctly. It is actually a
> good way to check that your formulas are correct.
>
> Hth
>
> --
> Lorenzo Paulatto
>
> On Sat, Sep 25, 2021, 21:28 Jacopo Simoni via users <
> users at lists.quantum-espresso.org> wrote:
>
>> So this means the phase factor is of the form e^{iG(k)r} where G(k) is
>> the translation vector in rec. space such that
>> G(k)+q+k=q'+k' (inside the first Brillouin zone)
>> the phase factor is therefore dependent on the k vector and in reciprocal
>> space the wave functions of the two equivalent q points are shifted by the
>> vector G(k) ?
>> Is there a variable in the code corresponding to this vector G or to the
>> phase factor itself ?
>>
>>
>> On Sat, 25 Sept 2021 at 03:42, Paolo Giannozzi <p.giannozzi at gmail.com>
>> wrote:
>>
>>> On Sat, Sep 25, 2021 at 2:00 AM Jacopo Simoni via users <
>>> users at lists.quantum-espresso.org> wrote:
>>>
>>> This wave function appears different from the wave function at an
>>>> equivalent q point, for instance if I look at evq at q=(0,0,1), this is
>>>> different from evq at (1,1,0) that are equivalent by translation of a G
>>>> vector (I am thinking here at a FCC periodic lattice). The two functions
>>>> just differ by a phase factor or I am missing something ?
>>>>
>>>
>>> They differ by a phase factor; moreover, in the presence of degenerate
>>> eigenvalues, you have no guarantee that the eigenvectors in the degenerate
>>> subspace are the same. Finally, the ordering of k+G components is not
>>> necessarily the same in the two cases
>>>
>>> Paolo
>>>
>>> Thanks in advance,
>>>> Jacopo Simoni, Lawrence Berkeley National Lab.
>>>> _______________________________________________
>>>> 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
>>>
>>>
>>>
>>> --
>>> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
>>> Univ. Udine, via delle Scienze 206, 33100 Udine, Italy
>>> Phone +39-0432-558216, fax +39-0432-558222
>>>
>>> _______________________________________________
>> 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
>
>
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20210927/7058b9e4/attachment.html>