My Community

Dear VASP experts and users,

I am doing calculations of small metal clusters to obtain energy of the cluster and forces on each atom. I have several frustrating questions and and any suggestions are highly appreciated.

1) I used ISMEAR = 0 and SIGMA = 0.1 as suggested in the manual for atom and dimer calculations and I found the T*S term to be fairly large: in several tenth ev, much higher than the suggested value of a few meV. I wonder this ISMEAR setting is correct or not (for a metal cluster, I also tried ISMEAR = 1, which also gives large T*S) and how important is the relatively larger T*S and its influence on the final forces I want to obtain.

2) I did single point calculations for the clusters. Will the cluster geometry to the final energy convergience and forces?

Thank you for your help.

AuD

small clusters have a discrete eigenvalue spectrum, like molecules, therefore SIGMA has to be significantly smaller (0.001 or even less)
ISMEAR=0 and a Gamma-only k-point mesh are the correct choices for clusters.
the total energies and equilibrium geometries then should converge to within the chosen precision (EDIFF, EDIFFG).