Dear all,
What is the correct FFT window to be used to obtain phonon DOS from velocity auto-correlation function?
I have conducted MD simulations of bcc-Fe and get the VACF data.
Using a Rectangle window for FFT, I observe non-zero data for frequencies at 0 Hz and >10 THz, as if the whole DOS curve has been shifted up.
Some paper suggested to use the Welch window. With that FFT window, frequencies at 0 HZ and >10 THz ranges reach 0 as one expect normally, however, the shape of the DOS is distorted and different from the result by frozen phonon calculations.
So how should I properly choose FFT method for phonon DOS from VACF?
Is there any physical significance of the shift during the usage of Rectangle window?
Any suggestion related calculations on phonon DOS via MD is also appreciated.
Thanks in advance.
Yi Wang
PostDoc of Tsinghua university
Dear Yi Wang,
I don't know what I suggest is what you're looking for but to obtain a good FFT calculations you can use the FFT code in the PYXAID package for phonon spectral density in the following link (the do_fft function), I hope it helps:
https://github.com/Quantum-Dynamics-Hub/pyxaid-code/blob/master/PYXAID/spectrum.py
More information can be found in this paper:
Article The PYXAID Program for Non-Adiabatic Molecular Dynamics in C...
- Badges
- Science topic
- Similar topics
- Biological Science
- Bioecology
- Systems Ecology
- Landscape Ecology
- Connectivity