Re: [Pw_forum] Supercell size in phonon calculations

[Charlie Ruffman]

Hi Stefano and others,


Thank you for your helpful response!


We are fairly new to both QE and phonon calculations, and have been trying to 
simulate the Raman spectrum of an extended 2D sheet using the primitive cell 
and also a 3x3 supercell. A phonon calculation at q=0 of the primitive cell 
produces results that are very close to experiment (at 16x16x1 Monkhorst pack 
k-points, 1e-8 Ry for electronic convergence, and a kinetic energy cutoff of 50 
Ry). However, calculation of the 3x3 supercell at q=0 (using the same 
convergence criteria and 8x8x1 k-points) produces additional Raman active 
modes, as well as shifting some of the modes present in the primitive cell to 
different energies. If the additional Raman active modes in the supercell are 
from  q/=0 in the primitive cell, why might these be Raman active? Also, is it 
possible to reliably simulate the Raman spectrum of a 3x3 or larger supercell?


Thank you again for your help with this.


--
Charlie Ruffman | BSc
BSc(Hons) Student

Chemistry Department
University of Otago
Tel 64 3 479 7930
PO Box 56, Dunedin 9054
New Zealand

Website http://neon.otago.ac.nz/research/alg/
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

Re: [Pw_forum] Supercell size in phonon calculations

[stefano de gironcoli]

On 06/09/2017 11:22, Charlie Ruffman wrote:


Hi all,


I have recently become interested in how the size of the unit cell in 
phonon calculations effects the results produced by the calculation 
(Raman spectra in particular). It seems that increasing the number of 
atoms explicitly simulated should also increase the number of normal 
modes calcuated?



It does. these are hovewer modes that would correspond to a q/=0 in the 
original primitive cell... as such they should be raman-inactive.
The calculation in the supercell, if properly done (same cutoff, 
equivalent k-points, sufficient scf convergence) should be exactly 
comparable to the superposition of the results in the primitive cell for 
the q-vectors that map into the Gamma point of the supercell.


stefano

Does anyone know if using the primitive cell will produce more or less 
accurate results than a larger supercell?



Thanks


--
Charlie Ruffman | BSc
BSc(Hons) Student

Chemistry Department
University of Otago
Tel 64 3 479 7930
PO Box 56, Dunedin 9054
New Zealand

Website _http://neon.otago.ac.nz/research/alg/ 
_



___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum



___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum

[Pw_forum] Supercell size in phonon calculations

[Charlie Ruffman]

Hi all,


I have recently become interested in how the size of the unit cell in phonon 
calculations effects the results produced by the calculation (Raman spectra in 
particular). It seems that increasing the number of atoms explicitly simulated 
should also increase the number of normal modes calcuated?


Does anyone know if using the primitive cell will produce more or less accurate 
results than a larger supercell?


Thanks


--
Charlie Ruffman | BSc
BSc(Hons) Student

Chemistry Department
University of Otago
Tel 64 3 479 7930
PO Box 56, Dunedin 9054
New Zealand

Website http://neon.otago.ac.nz/research/alg/
___
Pw_forum mailing list
Pw_forum@pwscf.org
http://pwscf.org/mailman/listinfo/pw_forum