[Q-e-developers] wavefunction

[jqli14]

Dear Nicola,

  (1) I found that the elements of wavefunction arrays do not share a phase after several iterations unless convergence is achieved. is that right? 
(2)And even though convergence is achieved, the phases of every elements are not exactly the same. When conv_thr=1.d-14, the errors between phases are around 1.d-4~1.d-5. How to reduce the error?

Jiqiang
      

在 2015-10-28 19:17:55，"Nicola Marzari" <nicola.marzari at epfl.ch> 写道：

>
>
>You are correct - there is a phase at gamma that you can remove
>and make all the gamma wavefunctions real.
>
>Also, you do not to change the phases at other k-points - the phases at
>every k-point are arbitrary.
>
>			nicola
>
>On 28/10/2015 14:11, Jiqiang Li wrote:
>> Dear Paolo,
>>       Sorry to have expressed the wrong idea.
>> 1) I want to ask how to merely force the wavefunction at Gamma point to
>> real numbers in a calculation with many k points, I do not care the case
>> of other k points. "The former case is implemented", is it implemented?
>> However, I output the wavefunction at Gamma point, result shows that
>>   wavefunction at Gamma point is complex.
>>
>> 2)Furthermore, I found that the elements of wavefunction array belong to
>> identical k points share the same phase.  Then can I multiply
>>   wavefuntions at all k points by the oppsite phase of wavefunction at
>> Gamma point. Then the wavefunction at Gamma point becomes real numbers,
>> the relative phase between wavefunction at all k points keep unchange
>> and physical properties will remain. Correct me if I am wrong.
>>
>> Jiqiang Li
>>
>> ------------------------------------------------------------------------
>>
>>     *From:* Paolo Giannozzi <mailto:p.giannozzi at gmail.com>
>>     *Date:* 2015-10-28 17:31
>>     *To:* General discussion list for Quantum ESPRESSO developers
>>     <mailto:q-e-developers at qe-forge.org>
>>     *CC:* jqli14 <mailto:jqli14 at fudan.edu.cn>
>>     *Subject:* Re: [Q-e-developers] wavefunction
>>     You cannot, in general.  You can choose real wavefunctions only for
>>     Gamma (k=0), or if the system has inversion symmetry. The former
>>     case is implemented, the latter is not.
>>     It wouldn't be that difficult to implement it, but it must be done
>>     with care.
>>
>>     Paolo
>>
>>     On Wed, Oct 28, 2015 at 8:07 AM, jqli14 <jqli14 at fudan.edu.cn
>>     <mailto:jqli14 at fudan.edu.cn>> wrote:
>>
>>         Dear all,
>>               In a calculation with  many  k points, how to force the
>>         wavefunctions to real numbers but not complex numbers?
>>
>>         Jiqiang Li
>>
>>
>>
>>
>>
>>         _______________________________________________
>>         Q-e-developers mailing list
>>         Q-e-developers at qe-forge.org <mailto:Q-e-developers at qe-forge.org>
>>         http://qe-forge.org/mailman/listinfo/q-e-developers
>>
>>
>>
>>
>>     --
>>     Paolo Giannozzi, Dept. Chemistry&Physics&Environment,
>>     Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
>>     Phone +39-0432-558216 <tel:%2B39-0432-558216>, fax +39-0432-558222
>>     <tel:%2B39-0432-558222>
>>
>>
>>
>> _______________________________________________
>> Q-e-developers mailing list
>> Q-e-developers at qe-forge.org
>> http://qe-forge.org/mailman/listinfo/q-e-developers
>>
>
>-- 
>----------------------------------------------------------------------
>Prof Nicola Marzari, Chair of Theory and Simulation of Materials, EPFL
>Director, National Centre for Competence in Research NCCR MARVEL, EPFL
>http://theossrv1.epfl.ch/Main/Contact http://nccr-marvel.ch/en/project
>_______________________________________________
>Q-e-developers mailing list
>Q-e-developers at qe-forge.org
>http://qe-forge.org/mailman/listinfo/q-e-developers