Computing Phonon dispersion from phelel_params.hdf5 file (VASP+phelel workflow)

Queries about input and output files, running specific calculations, etc.


Moderators: Moderator, Global Moderator

Post Reply
Message
Author
saranya_velliyarat
Newbie
Newbie
Posts: 7
Joined: Thu Sep 12, 2024 6:07 pm

Computing Phonon dispersion from phelel_params.hdf5 file (VASP+phelel workflow)

#1 Post by saranya_velliyarat » Tue Sep 29, 2026 11:10 pm

Hello,

I am working on calculating the phonon dispersion for my system. I am using a 2×2×2 supercell containing 120 atoms, which requires 90 displaced supercells. The finite-difference calculation on VASP using IBRION = 6 tag is time-consuming for a supercell of this size, so I used phelel to split the calculation into individual jobs.

I generated the displaced supercells with phelel and ran each one as a separate VASP calculation using the ELPH_PREPARE = .TRUE. tag and combined the results using phelel to obtain the "phelel_params.hdf5" file.

I understand that in the VASP-only workflow, the force constants can be written to vaspout.h5 and then reused with LPHON_READ_FORCE_CONSTANTS to calculate the phonon dispersion. Since the "phelel_params.hdf5" file generated from the same finite-displacement calculations also contains the information needed for the interatomic force constants, my main question is whether I can use the existing "phelel_params.hdf5" directly to calculate the phonon dispersion, instead of using "vaspout.h5".

If this is possible, could you please let me know the appropriate workflow or command for doing this?

Thank you very much for your help!

Best regards,
Saranya


henrique_miranda
Global Moderator
Global Moderator
Posts: 598
Joined: Mon Nov 04, 2019 12:41 pm
Contact:

Re: Computing Phonon dispersion from phelel_params.hdf5 file (VASP+phelel workflow)

#2 Post by henrique_miranda » Thu Oct 08, 2026 12:03 pm

Yes, you can use phelel_params.hdf5 directly. No need to go through vaspout.h5 / LPHON_READ_FORCE_CONSTANTS.

Put these in a folder:

  • phelel_params.hdf5
  • the primitive-cell POSCAR and POTCAR
  • a QPOINTS file with your q-path (same format as KPOINTS, e.g. line mode, reciprocal coordinates)
  • a small KPOINTS and a KPOINTS_ELPH (e.g. a Gamma 4x4x4 mesh in both)

and run with this INCAR:

PREC = Accurate
ISMEAR = 0; SIGMA = 0.01
LWAVE = .FALSE.
LCHARG = .FALSE.
NBANDS = 24
ELPH_RUN = .TRUE.
ELPH_DRIVER = MELS
ELPH_DECOMPOSE = "P"
KPOINTS_OPT_MODE = 2
IFC_ASR = -2

If a QPOINTS file is present, VASP builds the force constants from phelel_params.hdf5 and writes the phonon dispersion along that path to vaspout.h5 (results/phonons: eigenvalues, eigenvectors, kpoint_coords, k_path_length).


Post Reply