Queries about input and output files, running specific calculations, etc.
Moderators: Global Moderator, Moderator
-
asiyeh_shokri2
- Newbie
- Posts: 34
- Joined: Wed Aug 03, 2022 10:42 am
#1
Post
by asiyeh_shokri2 » Mon Oct 14, 2024 3:02 pm
Hello
I have questions about calculation of Raman Intensities using VASP which I appreciate any answer and guide. I am using (https://github.com/raman-sc/VASP) guide and program for a unitcell of 20 atoms.
using DFTP or finit displacement and EPSILON = True in INCAR file I see at the end of OUTCAR file there are all the Raman and IR modes of this material. So my question is if I do not want to plot the raman intensities, would it be enough to just do a single calculation using EPSILON = True and IBRION=5-8? how can I distinguish between IR and raman modes? Do I have to use a supercell for this calculation for a pure structure? what about doped crystals? In this case how should I compare my result while I will have for example 480 modes while for pure structure number of all modes is 30?
Best regards,
Asiyeh
-
alex
- Hero Member
- Posts: 585
- Joined: Tue Nov 16, 2004 2:21 pm
- License Nr.: 5-67
- Location: Germany
#3
Post
by alex » Tue Oct 15, 2024 1:09 pm
Hello Asiyeh,
if you do _not_ need Raman _intensities_ a normal frequency run as VASP provides is fully sufficient. You may chose IBRION in a way that suits your purpose best. LEPSILON does provide additional information you may or may not wish to have. Again: your choice.
Since VASP is lacking most of the interesting parts of group theory you have to apply selection rules to identify IR and/or Raman active modes yourself or with the help of third party software.
Concerning supercells it's a bit dependend on what you are after: for simple assignment the small cell might be enough, but for higher accuracy wrt. energy comparisons I'd go for a supercell simulation for the pure system, too. On the plus side you might identify degeneracies if you go away from harmonic approximation, e.g. with MD.
Good luck,
alex
-
asiyeh_shokri2
- Newbie
- Posts: 34
- Joined: Wed Aug 03, 2022 10:42 am
#4
Post
by asiyeh_shokri2 » Wed Oct 16, 2024 10:14 am
Hello
First of all, thank you for your response. When I perform DFPT or finite difference calculations (using IBRION = 5, 6, 7, or 8 ) , the OUTCAR file provides the eigenvalues and eigenvectors of the dynamical matrix. It lists several frequencies, matching the expected number of IR and Raman modes, suggesting that both IR and Raman active modes are included.
My question is: if I only need the vibrational frequencies, are these calculations (IBRION = 5, 6, 7, or 8 ) sufficient? As far as I understand, extracting Raman tensors, intensities, activities, or polarization data would require additional tools like phonopy or raman_sc.
I hope my question is clear now.
Best regards,
Asiyeh
Last edited by
asiyeh_shokri2 on Wed Oct 16, 2024 10:15 am, edited 2 times in total.
-
asiyeh_shokri2
- Newbie
- Posts: 34
- Joined: Wed Aug 03, 2022 10:42 am
#5
Post
by asiyeh_shokri2 » Wed Oct 16, 2024 10:36 am
alex wrote: ↑Tue Oct 15, 2024 1:09 pm
Hello Asiyeh,
if you do _not_ need Raman _intensities_ a normal frequency run as VASP provides is fully sufficient. You may chose IBRION in a way that suits your purpose best. LEPSILON does provide additional information you may or may not wish to have. Again: your choice.
Since VASP is lacking most of the interesting parts of group theory you have to apply selection rules to identify IR and/or Raman active modes yourself or with the help of third party software.
Concerning supercells it's a bit dependend on what you are after: for simple assignment the small cell might be enough, but for higher accuracy wrt. energy comparisons I'd go for a supercell simulation for the pure system, too. On the plus side you might identify degeneracies if you go away from harmonic approximation, e.g. with MD.
Good luck,
alex
Hello Alex
Thank you so much for your respond. So if I correctly understand you the frequencies I see in the OUTCAR are the raman and IR modes I am looking for. Am I right?
Best regards,
Asiyeh
-
alex
- Hero Member
- Posts: 585
- Joined: Tue Nov 16, 2004 2:21 pm
- License Nr.: 5-67
- Location: Germany
#6
Post
by alex » Wed Oct 16, 2024 11:55 am
Yes Asiyeh,
all(!) vibrational modes will be computed by every code. No matter of their intensity.
The vibrational modes are eigenvalues of the Hesse matrix (or Hessian).
For systems with symmetry selection rules may apply, which (dis-)allow intensity at the mode under consideration.
However, computing intensities (IR or Raman) always is based on the vibrational modes you get from the Hessian.
No other positions/energies are possible.
Cheers,
alex
-
fabien_tran1
- Global Moderator
- Posts: 418
- Joined: Mon Sep 13, 2021 11:02 am
#7
Post
by fabien_tran1 » Wed Oct 16, 2024 1:40 pm
asiyeh_shokri2 wrote: ↑Wed Oct 16, 2024 10:14 am
My question is: if I only need the vibrational frequencies, are these calculations (IBRION = 5, 6, 7, or 8 ) sufficient? As far as I understand, extracting Raman tensors, intensities, activities, or polarization data would require additional tools like phonopy or raman_sc.
Yes you are right.