Hi All,
I am working on Carbon Nanotubes and want to generate the Vibration Spectrum for the carbon atoms of the nanotube. I have read in some papers that “Fourier transform of
the velocity autocorrelation function” can be used to generate it, but no one provided a reference book/article describing how to do that. If anyone know how to do it, please describe how I can do it. Also if you know of any good book/paper containing the details, please let me know.
Thank you all for the help,
Navdeep
Navdeep,
This paper might help you:
Superlattices and Microstructures. Volume 35, Issues 3-6, March-June 2004,
Pages 227-237.
Ali Rajabpour
Navdeep,
This paper might help you:
Superlattices and Microstructures. Volume 35, Issues 3-6, March-June 2004,
Pages 227-237.
Ali Rajabpour
Hi All,
I am working on Carbon Nanotubes and want to generate the Vibration Spectrum
for the carbon atoms of the nanotube. I have read in some papers that
"Fourier transform of
the velocity autocorrelation function" can be used to generate it, but no
one provided a reference book/article describing how to do that. If anyone
know how to do it, please describe how I can do it. Also if you know of any
good book/paper containing the details, please let me know.
most text books on molecular dynamics contain this information.
if you want more insight into the technical aspects of implementation,
have a look at the classic book on MD by allen and tildesley.
if you want to know more about the theory, check out a text book
on statistical mechanics, e.g. the one by mcquarrie.
if you want to be lazy, just dump the velocities into a suitable
lammps custom trajectory file, read it into VMD and use the
specden plugin to compute the spectral densities directly.
cheers,
axel.
Axel and Ali, Tanks a lot for the help…
Navdeep
Hi Navdeep,
If you treat the velocity of the atom as a time series data/signal,
the vibrational spectrum is nothing more than power spectral density
of that time series data.
If you want the vibrational spectrum, then just collect the atom
velocity at fixed time intervals over the duration of your simulation,
load the data into Matlab (or any of your favourite numerics
software), take its discrete FFT and then absolute-square it.
According to the Wiener–Khinchin theorem, it's equivalent to taking
the Fourier transform of the velocity auto-correlation function.
Good luck,
Zhun-Yong