Dear Pedram, I think it could be good if you carefully think about what actually you want to achieve, and why not get something to read about molecular dynamics.
You have a periodic system with 4 atoms per cell. You say you relaxed your system in CG scheme; fine. Now, molecular dynamics needs some kind of ensemble. It all is about how the atoms wildly move in all different directions. If you'd have a molecule, these 4 atoms would be all you have, so running a MD simulation on them would be legitimate. However, you impose periodic boundary conditions, so the translated atoms in all cells are forced to move the same way. This is not the way the Nature does it. In other words, you want to explore the phase space (of all particles in a crystal) yet constrain yourself to an infinitesimal specially selected part of the phase space. I'd not expect anything reasonable from such simulation even if done without technical errors. If you check the literature you'll see that MD simulations are done either on molecules, or - in dense systems - on large enough supercells. So if you want to simulate how ZnO vibrates at some temperature, you have to go to a much much larger supercell, no way around. Good luck Andrei Postnikov > Hi, > > I really hope someone would kindly help me solve this problem. > > Enclosed, there's a .fdf file for bulk zno running for optimization under > NPT ensemble. > The geometry has been first relaxed up to 0.001 eV/Ang in CG scheme and > then I started running it in MD. > > The system is not showing any convergence in step 1800 and the trends > seems > to be really odd. > > I would be grateful if you kindly take a look at the files to find the > source of the problem. > > Best Regards, > Pedram >
