My Community

I have some questions related with drift force. Firstly, i cannot find the exact definition of it. Secondly, i would like to ask how it is calculated. In some of my calculations the drift force is much larger than the forces on each atom. What is the convergence criteria for it? Can i still trust my results or geometry optimization?. I would appreciate any information related with this subject. thank in advance...

the total drift of the forces is a drift of the center of gravity of the system. It should be smaller than the forces on the individual atoms.
usually a considerable total drift in the forces may be caused by either
1) insufficient basis set sizes (--> please increase PREC and ENCUT)
or
2) if you do a molecule in a box or slabs, the whole molecule or slab might start drifting. In this case it is sufficient to fix the co-ordinates of at least one atom (of the molecule) or one layer of the slab (see: POSCAR/Selective Dynamics)

I have one question about the total drift forces. You said the total drift should be smaller than the forces on the individual atoms. In my calculations (slab, large system) I never got the total drift force is smaller than the force on individual atoms. I often use at least 80 atoms per cell. I fix at least two atomic layer and use high cut off and PREC=accurate. But I have not seen that the total drift is less than the force on each atoms. However I get a total drift of less than 0.02eV in most of the cases.
Now I am doubt about my calculations and your statement. Could you please explain us more?
Thank you

<span class='smallblacktext'>[ Edited Fri Mar 10 2006, 07:40AM ]</span>