search

skelton-group / Phonopy-Spectroscopy

A collection of tools for simulating vibrational spectra, which interfaces with the Phonopy package.
MIT License
144 stars 55 forks source link

requesting more information #8

Closed iowachem closed 1 year ago

iowachem commented 2 years ago

I have installed all the needed packages and have setup the scrip properly.

My question is on generating the Raman and IR spectrum. I have setup a set of 2x2x2 supercells using phonopy and have done the calculation. I have generated the irreps.yaml, mesh.yaml, FORCE_SETS.

Do I need to do another set of calculations with phono3py? or can I proccess just with the phonopy calculation output files?

Can you please provide me with instruction on how to proceed from here.

demonic-daisy commented 2 years ago

I have almost the same problem. I am using dftb+ as driver (due to a very large unit cell) and am interested in a solid state IR spectrum. I got eigenvectors (binary format) and forces (xml or plain text format) from dftb+ calculations of my irreducible cell but don't see a path how to process from there (and how to convert my file formats) to obtain an IR spectrum using phonopy, phono3py and phonopy-ir. @iowachem Have you found a solution for your problem that you would share with me?

JMSkelton commented 1 year ago

@iowachem I suspect you have probably worked this out already, but: A Phono3py calculations is optional and is performed to calculate the mode linewidths from first principles. If you do not want to do this, you can specify a uniform linewidth using the --linewidth command-line argument (defaults to 0.5 THz ~ 16.5 cm^-1).

@demonic-daisy The code currently only interfaces with Phono(3)py. (There is a direct interface to the VASP code, but this has been temporarily removed pending further testing after the recent code update.) The recent code updates, and further updates to follow, should make it easier to write interfaces to other codes in the future. We are not familiar with dftb+, and writing an interface would therefore require a considerable amount of effort - presuming that dftb+ can compute the quantities required in addition to the phonon frequencies and eigenvectors (Born or other definition of atomic charges, dielectric tensors). If you still need this, we can work with you to implement it, but we cannot provide an indicative timescale at present.