Phonon calculations after scf calculation using pw.x + environ

[Johanna W.]

Hello all,

I am looking at isolated molecules and slab systems in implicit water environment. In vacuum, I obtained realistic vibrational frequencies from phonon calculations using ph.x in quantum espresso (and q2r.x and matdyn.x as post-processing tools) - also if I include environ to create the vacuum environment (environ_type='vacuum'). For a diatomic oxygen molecule, for example, this means that I get five frequencies which are almost zero (rotational and translational modes) and one non-zero vibrational mode around 46 THz.

 However, if I move to water environment (environ_type='water'), at least for the isolated molecules, the phonon calculation gives non-zero negative frequencies for the rotational and translational modes in the order of -10 THz. Using q2r.x and matdyn.x, I can "force" them back to zero but still the negativity in the phonon (ph.x) output worries me. Has anyone have similar experiences? Are there important remarks to take into account when combining environ and phonon calculations?

For reference, I attach my input and output files for the O2 calculation.

Best wishes and thank you very much

Johanna

[andr...@gmail.com]

Hi Johanna,

Great that you are using Environ, your question is very relevant and it is possible that the answer is not as clear as it should on the Environ documentation. Environ is not coupled with the PHonons code. ph.x uses linear response to compute vibrational modes. This is great if you have a small cell and you use reciprocal space points. However with Environ it is often the case that your simulation uses a large cell and you do not need q-points. This is to say, the effort of coupling Environ in a linear response code is significantly larger than just using finite difference schemes, say through ASE vibrations or ASE phonons, or Phonopy. In these approaches, you are basically displacing each atom along each Cartesian direction and then running a pw.x calculation, whose output allow to build the dynamic matrix. If you run each pw.x calculation with Environ you automatically have vibrations in the continuum solvent (where the main approximation is that you are still using the static permittivity of the solvent, but there is not easy way around this). 

I am not sure if your simulations strictly required you to use ph.x, I hope you can translate to the frozen phonons schemes. Let me know if there is anything that is not clear.

Best,

Oliviero

Oliviero Andreussi (he/him)
--
Associate Professor
Department of Chemistry and Biochemistry
Boise State University
Office: SCNC 316
Email: oliviero...@boisestate.edu
Web: http://www.materialab.org

[Johanna W.]

Hi Oliviero, 

thank you very much for your reply!

Since I use phonon calculation as base for thermodynamic corrections, I for now decided to perform them in vacuum environment and then add the corrections to the energy I get when regarding the slab in implicit electrolyte environment using environ. It's an approximation for sure but for now it seems valid - I will also try the frozen phonon schemes though, thank you for the hint!

Best wishes

Johanna