My Community

Hi,

I am running calculations on the vibrational modes of molecular crystals using IBRION=5 feature. However, I found that the symmetry of the hessian matrix is not used and all the atoms are being desplaced, which dramatically increase the computation time. Since I am using gamma point for BZ integration anyway, I don't need to worry about the symmetry effects for the k point sampling. Could you let me know whether I can use the symmetry for displacing the atoms?

Thanks,
Maosheng

Hi,

It seems the current version does not use the symemtry to calculate the Hessian matrix. I think waht I can do is manually displace the necessary atoms and reconstruct the full Hessian matrix using the smaller symmetry reduced matrix. My question is then how to let VASP write out the calculated hessian.

Thanks,
Maosheng

Hi,

there is no way. You have to do 'by hand'. But once you got the Hessian for your symmetric system, it's just a little programming. Look at the source...

alex

the excited modes can be chosen by making use of the
Selective dynamics
option in POSCAR (please have a look in the online handbook,
section files, POSCAR how to use it).
Then, by choosing the excitations explicitely, you can select special vibration modes and keep any symmetry you want to.

Hi,

Thanks for everybody's reply. I am using the selective dynamics. However, my purpose is to obtain all the vibrational modes of the molecular crystal. I can select the minimum number of displacement according to the symmetry. But later on I wish to construct the full Hessian matrix and diagonalize it and obtain the full spectra of vibrations. As Alex said, I think VASP do not work that way. The selective dynamics is designed for a system like molecule on a solid surface. I am reading the code and may need your guys further help.

Thanks.
Maosheng