My Community

When running a surface simulation using MD capabilities, we obtain velocities around zero for x and y components, while in the z direction values are of the order they should be according to the expected Maxwell Boltzmann distribution.

The simulation is carried out with normal periodic boundary conditions in the x and y directions and in the z direction the box length is set to a large number. Temperature is controlled using a simple rescaling scheme.:
SMASS=-1,NBLOCK=1.

Does anybody know why velocities are so different in the different directions?

it may very well be possible (at least at low T) that the atoms have their highest mobility along the surface normal. Please check if you have set
1) either TEBEG,TEEND in INCAR (or some starting random velocities in CONTCAR).
2) switched off symmetry (otherwise, the forces are symmetrized according to the 2D symmetry of your surface)