UT theoretical chemistry code forum

Dear all,

I have a general NEB question that I would like to pick your brains on.

I'm trying to find the diffusion barrier for an atom adsorbed on a nanowire. On nanowires that others have determined to be structurally rigid, my NEB calculations converge properly. On nanowires that are flexible, my calculations have difficulties converging. Is it a lost cause to try to find a barrier for diffusion on a flexible substrate?

This is a very good point. The NEB as implemented in VASP can have problems with soft modes, including rotation and translation. There are two issues; one is that the optimizers are in Cartesian coordinates -- internal coordinates are preferred for molecular systems with soft modes. The second issue is that the band can lengthen along soft modes giving poor resolution along the path. One strategy that you might try, if you are having problems, is to relax the band on the rigid substrate (or with whatever constraints make sense) and then re-relax the band in the more flexible system.