Vibrational and rotational spectra can be calculated from the molecular dynamics trajectories, but this task is, many times, performed by a different software than that employed for the MD.
it basically depends on which type of spectrum you want to calculate... you just have to know your transition operator and then you could write your own short code.
A vibrational spectrum (vibrational density of states) for example can be calculated from the Fourier transform of the normalized mass-weighted atomic velocity autocorrelation function (v*v). You just need a trajectory long enough to give you proper resolution (20 ps correlation function will result in about 2 wavenumbers) and the trajectory has to have a sufficient data sampling rate (I usually go with something around 2-5 fs, which will give you a Nyquist limit of around 6000-4000 wavenumbers).
If you want to get an IR spectrum you would do the same with the dipole autocorrelation (q*r) function and maybe apply some frequency dependent quantum correction afterwards...
at the moment the only software I know to do the latter type of correlation is VMD via the spectral analysis function. You can find some more info on this on the doc page of VMD.
it basically depends on which type of spectrum you want to calculate... you just have to know your transition operator and then you could write your own short code.
A vibrational spectrum (vibrational density of states) for example can be calculated from the Fourier transform of the normalized mass-weighted atomic velocity autocorrelation function (v*v). You just need a trajectory long enough to give you proper resolution (20 ps correlation function will result in about 2 wavenumbers) and the trajectory has to have a sufficient data sampling rate (I usually go with something around 2-5 fs, which will give you a Nyquist limit of around 6000-4000 wavenumbers).
If you want to get an IR spectrum you would do the same with the dipole autocorrelation (q*r) function and maybe apply some frequency dependent quantum correction afterwards...
at the moment the only software I know to do the latter type of correlation is VMD via the spectral analysis function. You can find some more info on this on the doc page of VMD.
- Badges
- Science topic