Hi,
I am quiet new with ABINIT, and I am having some troubles using the anaddb code. I am wondering how/from where i can get or generate the list of nph1l wavevectors, at which the phonon frequencies will be interpolated. I have all the _DDBs for each qpt, that, as the tutorial does, I picked the same as the points of the kgrid I used for the GS run. Then, when i need the qph1 for anaddb, i am a bit lost,.. where are they from?
Thanks for your attention,
Valentina
qph1l phonon dispersion anaddb
Moderators: mverstra, joaocarloscabreu
-
- Posts: 271
- Joined: Sat Jan 16, 2010 12:05 pm
- Location: Dept. of Physics, University J. J. Strossmayer, Osijek, Croatia
- Contact:
Re: qph1l phonon dispersion anaddb
Dear Valentina,
The thing is that you have to adequately represent the dispersions (PD) throughout the BZ.
If you have some other PDs (from experiments or previous theoretical work) and you want to compare yours to theirs, it would be nice to follow their paths through the BZ. So, just take their high symmetry points and divide the line between each two points into 20 or 40 parts (depending how smooth curves you want). And then enter all these points into the qph1l list.
You can use the awk command to generate a list of points between a finite q-point and 0. Make a d.dat file which contains only one zero. Then execute
awk '{ for (i=1; i<=40; i++) {s=0.3333; k=0.0083325*i; s=s-k+$1; printf(" %g", s); printf("\n") } }' d.dat > d1.dat
In d1.dat you'll have all the points between 0.3333 and 0, with an interval of 0.0083325. There are probably better (and nicer) ways to do this, but this one's the one I use.
If you don't have any PDs to compare to, you'll have to think of your own path around and/or through the "octant" of your BZ.
Hope this helps.
Cheers!
Igor L.
The thing is that you have to adequately represent the dispersions (PD) throughout the BZ.
If you have some other PDs (from experiments or previous theoretical work) and you want to compare yours to theirs, it would be nice to follow their paths through the BZ. So, just take their high symmetry points and divide the line between each two points into 20 or 40 parts (depending how smooth curves you want). And then enter all these points into the qph1l list.
You can use the awk command to generate a list of points between a finite q-point and 0. Make a d.dat file which contains only one zero. Then execute
awk '{ for (i=1; i<=40; i++) {s=0.3333; k=0.0083325*i; s=s-k+$1; printf(" %g", s); printf("\n") } }' d.dat > d1.dat
In d1.dat you'll have all the points between 0.3333 and 0, with an interval of 0.0083325. There are probably better (and nicer) ways to do this, but this one's the one I use.
If you don't have any PDs to compare to, you'll have to think of your own path around and/or through the "octant" of your BZ.
Hope this helps.
Cheers!
Igor L.
Re: qph1l phonon dispersion anaddb
Thank you very much Igor, this is really helpful!
Regards,
Valentina
Regards,
Valentina