[Pw_forum] Bug in Raman calcs with no-symmetries
Paolo Giannozzi
p.giannozzi at gmail.com
Mon Apr 25 22:08:23 CEST 2016
Thank you. It will be fixed in the next release (coming really soon)
Paolo
On Mon, Apr 25, 2016 at 12:20 PM, Marc Blanchard <
marc.blanchard at impmc.upmc.fr> wrote:
> Dear Paolo Giannozzi,
>
> Your comment is correct.
>
> Recapitulating, one should change the subroutines symmatrix3 and
> symtensor3 (both written in
> the file PW/src/symme.f90)
> 1) The first instruction in both routines is
> IF (nsym == 1) RETURN
> This should be changed (in both routines) with
> IF (nsym > 1) THEN
> 2) then, one should add the line
> END IF
> before the comment "Bring to cartesian axis"
>
>
> To uncover the bug you might run the following example: Silicon bulk
> in the hexagonal cell with six atoms slightly displaced to kill the
> symmetries.
> The pseudopotential is taken from the quantum-espresso database.
> I report the input files I used for the pwscf and the phonon calculation:
>
> cat > input_pw <<EOF
> &control
> calculation='scf'
> restart_mode='from_scratch',
> pseudo_dir = '$PSEUDO_DIR/',
> outdir='$TMP_DIR/',
> prefix='si'
> /
> &system
> ibrav= 4, celldm(1) =7.2, celldm(3)=2.44948974278317810d0,
> nat= 6, ntyp= 1, ecutwfc =55
> /
> &electrons
> diagonalization='david', mixing_beta = 0.7, conv_thr=1.d-12,
> /
> /
> ATOMIC_SPECIES
> Si 28.085 Si.pw-mt_fhi.UPF
> ATOMIC_POSITIONS {crystal}
> Si 0.000100000000 0.000100000000 0.000100000000
> Si 0.000100000000 0.000000000000 0.250000000000
> Si 0.333333333333 0.666666666667 0.333333333333
> Si 0.333333333333 0.666666666667 0.583333333333
> Si 0.666666666667 0.333333333333 0.666666666667
> Si 0.666666666667 0.333333333333 0.916666666667
> K_POINTS {automatic}
> 2 2 1 0 0 1
> EOF
>
> cat > input_ph <<EOF
> phonons of si at Gamma
> &inputph
> prefix='si',
> epsil=.true.,
> trans=.false.,
> lraman=.true.,
> amass(1)=28.0855,
> outdir='$TMP_DIR/',
> fildyn='si.dynG',
> /
> 0.0 0.0 0.0
> EOF
>
>
> With the old version of the code (in which the bug is present) the Raman
> tensor
> associated to the first atom is:
> Raman tensor (au^-1) in cartesian axis
>
> atom 1
> ( 0.013699086 -3.920153995 5.788324065 )
> ( -3.920153995 3.923226229 -2.896625782 )
> ( 5.788324065 -2.896625782 -0.003485967 )
>
> ( -3.911816105 3.914537535 -2.896540201 )
> ( 3.914537535 0.014410663 5.790162015 )
> ( -2.896540201 5.790162015 0.023554646 )
>
> ( 8.842144308 -4.420575437 -0.000769476 )
> ( -4.420575437 8.847405130 0.011191688 )
> ( -0.000769476 0.011191688 -31.777282604 )
>
>
> After correcting the bug it becomes:
> Raman tensor (au^-1) in cartesian axis
>
> atom 1
> ( 0.013699086 -4.518694758 2.363073404 )
> ( -4.518694758 0.008662674 -0.001161423 )
> ( 2.363073404 -0.001161423 -0.000580994 )
>
> ( -4.509066992 0.002636341 -0.001121079 )
> ( 0.002636341 4.548354141 2.362756055 )
> ( -0.001121079 2.362756055 0.004197656 )
>
> ( 3.609790298 0.000234155 -0.000128246 )
> ( 0.000234155 3.612924305 0.002079798 )
> ( -0.000128246 0.002079798 -2.162170216 )
>
> In order to verify that the second run is correct you can do
> a third run in which the atomic positions are:
> Si 0.000000000000 0.000000000000 0.000000000000
> Si 0.000000000000 0.000000000000 0.250000000000
> Si 0.333333333333 0.666666666667 0.333333333333
> Si 0.333333333333 0.666666666667 0.583333333333
> Si 0.666666666667 0.333333333333 0.666666666667
> Si 0.666666666667 0.333333333333 0.916666666667
>
> By doing this, the system has now several symmetries and the bug does not
> appear.
> The Raman tensor in this case is:
> Raman tensor (au^-1) in cartesian axis
>
> atom 1
> ( 0.000000000 -4.524263555 2.358129910 )
> ( -4.524263555 0.000000000 -0.000000000 )
> ( 2.358129910 -0.000000000 0.000000000 )
>
> ( -4.524263555 0.000000000 0.000000000 )
> ( 0.000000000 4.524263555 2.358129910 )
> ( 0.000000000 2.358129910 -0.000000000 )
>
> ( 3.592025004 0.000000000 0.000000000 )
> ( 0.000000000 3.592025004 0.000000000 )
> ( 0.000000000 0.000000000 -2.186618341 )
>
>
> Sincerely
>
> Marc Blanchard & Michele Lazzeri
>
> ___________________________
> Dr Marc Blanchard
> Equipe de Minéralogie Environnementale
> Institut de Minéralogie, de Physique des Matériaux et de Cosmochimie
> (IMPMC)
> UPMC - Sorbonne Universités
> Tour 23 - Couloir 23/24 - 4e étage - Bureau 23
> Case courrier 115
> 4 place Jussieu
> 75252 Paris Cedex 05
> Tel: +33 (0)1 44 27 98 22
> Fax: +33 (0)1 44 27 37 85
> http://www.impmc.upmc.fr/~blanchard/
>
> Le 22 Apr 2016 à 11:45, Paolo Giannozzi a écrit :
>
>
>
> On Fri, Apr 22, 2016 at 11:02 AM, Marc Blanchard <
> marc.blanchard at impmc.upmc.fr> wrote:
>
>>
>> To correct the bug:
>> 1) The first instruction in both routines is
>> IF (nsym == 1) RETURN
>> This should be changed (in both routines) with
>> IF (nsym >= 1) THEN
>>
>
> "If (nsym > 1 THEN", I guess, or else the IF is completely useless. Could
> you please point to a (simple!) example showing the bug? thank you
>
> Paolo
>
> 2) then, you should add the line
> END IF
> before the comment "Bring to cartesian axis"
>
>
> Best regards
> Marc Blanchard & Michele Lazzeri
>
> ___________________________
> Dr Marc Blanchard
> Equipe de Minéralogie Environnementale
> Institut de Minéralogie, de Physique des Matériaux et de Cosmochimie
> (IMPMC)
> UPMC - Sorbonne Universités
> Tour 23 - Couloir 23/24 - 4e étage - Bureau 23
> Case courrier 115
> 4 place Jussieu
> 75252 Paris Cedex 05
> Tel: +33 (0)1 44 27 98 22
> Fax: +33 (0)1 44 27 37 85
> http://www.impmc.upmc.fr/~blanchard/
>
>
> _______________________________________________
> Pw_forum mailing list
> Pw_forum at pwscf.org
> http://pwscf.org/mailman/listinfo/pw_forum
>
>
>
> --
> Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
> Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
> Phone +39-0432-558216, fax +39-0432-558222
>
>
>
--
Paolo Giannozzi, Dip. Scienze Matematiche Informatiche e Fisiche,
Univ. Udine, via delle Scienze 208, 33100 Udine, Italy
Phone +39-0432-558216, fax +39-0432-558222
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.quantum-espresso.org/pipermail/users/attachments/20160425/1717f313/attachment.html>
More information about the users
mailing list