My Community

For surface calculation, I found that LOCPOT has a lot of noise. If the surface is in the x-y plane, I pick an x and y on the surface and plot the potential along the z direction, I found that there are a lot of fluctuations when the potential is decaying into the vacuum. In addition, it seems like if I let VASP adding the exchange-correlation part into the local potential, the noise is even bigger. Is there any way to improve this situation? Thank you.

the XC part converges only slowly therefore it is omitted in LOCPOT if you run the default version of vasp.
concerning the fluctuations (without XC):
please check if
1) your calculations were really converged (basis set sizes, electronic and ionic ground state)
2) if the vacuum is thick enough to avoid spurious surface-surface interactions across the vacuum