I want to study the thermal properties of a mixed system which is constructed by virtual crystal approximation in VASP. When I try to run the ab initio Molecular Dynamics of this system in VASP, I found that there would be either nearly infinite Temperature (kinetic energy as well) or large ERROR output. So I am wondering if the system with VCA could be used to run the AIMD via vasp or any other packages?
Look forward your advice and thanks a lot!
Don't use VCA method in any systems ! The result of VAC is unbelievable and this method has been discarded by the scientific community. Use SQS or other randomly method for mixture systems.
Dr. Wang
Beyond what Wang Jinkai said, I'm pretty sure the VCA implementation in VASP is still being worked on (for example, as far as I can remember, you can't do phonon calculations yet), so there may be bugs.
Unless you've got a very complex mixed system, you might want to use any of the supercell based methods for AIMD, although I realize that the computational costs might be very high.
If you are doing MD simulations, you are already probably using a supercell large enough to allow you to distribute atoms randomly or quasi-randomly over the shared sites, getting a better representation of the disordered system. VCA only makes sense, if it does at all, to force a small-cell calculation for a disordered system. For MD simulations, use an SQS (special quasi-random structure, where the atoms are place over the sites in a way that mimicks the distribution functions of a truly disordered system). Or, if you supercell is very large, just assign random occupations.
- Badges
- Science topic
- Similar topics
- Filing