JMSkelton
commented
4 years ago Dear Chris,
From what you said, it sounds like there is a problem with the Phonopy QE interface. If after further investigation you think this is the case, it may be worth putting it to the Phonopy developer.
It is particularly strange that you only see the frequency offset in the "molecule in a box" calculations. The way I run these in VASP they are treated exactly the same way as extended systems. Have you have used any of the settings tags in QE for aperiodic calculations?
phonopy-ir reads the frequencies from the mesh.yaml (or mesh.hdf5) file produced by Phonopy. I have not used the Phonopy QE interface before, but some things to check would be:
- The frequencies: if the frequencies calculated by Phonopy do not match the QE output, then as I said the problem may be the Phonopy QE interface.
- The physical units: phonopy-ir assumes the frequencies are in THz, but if the Phonopy QE interface outputs them in different units by default, this may be the source of the problem.
Best wishes,
Jonathan
On Mon, 13 Jan 2020 at 04:46, chrisewolf notifications@github.com wrote:
Dear all,
I am experiencing some weird red-shift when using PWSCF as calculator. I know that this was developed with VASP in mind but I was hoping someone here could help me out.
I have compared the DFPT results of PWSCF with the VASP results of the benzene example (using the same unit cell, but relaxed geometry) and I find them in excellent agreement but when using phonopy and phonopy-spectroscopy the modes are red-shifted by a factor of sqrt(2) (at least it seems that way). This does not happen for any extended system but for all the molecules-in-vacuum-box I tested.
Has anyone experienced this? Any command line input I might have missed...?
As an example I attach the highest modes (first the DFPT results and then phonopy) mode cm-1 THz IR | Phonopy
31 3073.41 92.1384 0.0000 127.3836904483 32 3082.09 92.3987 0.0000 127.7387235024 33 3082.23 92.4028 0.0000 127.8205031638 34 3095.56 92.8026 0.7386 128.3666018221 35 3095.73 92.8077 0.7432 128.4551561033 36 3104.16 93.0603 0.0000 128.8032547943
to get to these numbers I used exactly the procedure outlined in the bezene example
phonopy --qe -d --dim="1 1 1" -c scf.in -v phonopy --qe -f supercell-{1..048}.out
Dielectric constant and Born charges are basically identical with the ones given in the example.
phonopy --dim="1 1 1" --fc-symmetry --mesh="1 1 1" --eigenvectors --qe -c scf.in
`Python version 3.7.3 Spglib version 1.13.0 Calculator interface: qe Crystal structure was read from "scf.in". Mesh sampling mode Settings: Force constants symmetrization: on Sampling mesh: [1 1 1] Supercell: [1 1 1] Spacegroup: Pm (6) Use -v option to watch primitive cell, unit cell, and supercell structures.
Forces and displacements are read from "FORCE_SETS". Computing force constants... Max drift of force constants: -0.000244 (xx) 0.000000 (zz) Max drift after symmetrization: -0.000000 (xx) -0.000000 (xx) Mesh numbers: [1 1 1] Number of irreducible q-points on sampling mesh: 1/1 Calculating phonons on sampling mesh... `
phonopy-ir --linewidth=16.5 --spectrum_range="000 6000"
[image: IR-Spectrum] https://user-images.githubusercontent.com/27327867/72233493-0387e800-360b-11ea-9568-39a2d3746cad.png
I understand that this is not really an issue of Phonopy-Spectroscopy so please feel free to close it at any time. Any help would be much appreciated!
Best, Chris
— You are receiving this because you are subscribed to this thread. Reply to this email directly, view it on GitHub https://github.com/JMSkelton/Phonopy-Spectroscopy/issues/4?email_source=notifications&email_token=ACCK4FCQHYV5DGPENYHXNJDQ5PW27A5CNFSM4KF4K4NKYY3PNVWWK3TUL52HS4DFUVEXG43VMWVGG33NNVSW45C7NFSM4IFUQX7Q, or unsubscribe https://github.com/notifications/unsubscribe-auth/ACCK4FECFONMPFVNWJ6DYXLQ5PW27ANCNFSM4KF4K4NA .
chrisewolf
Dear all,
I am experiencing some weird red-shift when using PWSCF as calculator. I know that this was developed with VASP in mind but I was hoping someone here could help me out.
I have compared the DFPT results of PWSCF with the VASP results of the benzene example (using the same unit cell, but relaxed geometry) and I find them in excellent agreement but when using phonopy and phonopy-spectroscopy the modes are red-shifted by a factor of sqrt(2) (at least it seems that way). This does not happen for any extended system but for all the molecules-in-vacuum-box I tested.
Has anyone experienced this? Any command line input I might have missed...?
As an example I attach the highest modes (first the DFPT results and then phonopy)
mode cm-1 THz IR Phonopy 31 3073.41 92.1384 0.0000 127.3836904483 32 3082.09 92.3987 0.0000 127.7387235024 33 3082.23 92.4028 0.0000 127.8205031638 34 3095.56 92.8026 0.7386 128.3666018221 35 3095.73 92.8077 0.7432 128.4551561033 36 3104.16 93.0603 0.0000 128.8032547943
to get to these numbers I used exactly the procedure outlined in the bezene example
phonopy --qe -d --dim="1 1 1" -c scf.in -v phonopy --qe -f supercell-{1..048}.out
Dielectric constant and Born charges are basically identical with the ones given in the example.
phonopy --dim="1 1 1" --fc-symmetry --mesh="1 1 1" --eigenvectors --qe -c scf.in
_Python version 3.7.3 Spglib version 1.13.0 Calculator interface: qe Crystal structure was read from "scf.in". Mesh sampling mode Settings: Force constants symmetrization: on Sampling mesh: [1 1 1] Supercell: [1 1 1] Spacegroup: Pm (6) Use -v option to watch primitive cell, unit cell, and supercell structures.
Forces and displacements are read from "FORCESETS". Computing force constants... Max drift of force constants: -0.000244 (xx) 0.000000 (zz) Max drift after symmetrization: -0.000000 (xx) -0.000000 (xx) Mesh numbers: [1 1 1] Number of irreducible q-points on sampling mesh: 1/1 Calculating phonons on sampling mesh...
phonopy-ir --linewidth=16.5 --spectrum_range="000 6000"
I understand that this is not really an issue of Phonopy-Spectroscopy so please feel free to close it at any time. Any help would be much appreciated!
Best, Chris