get 1/3 P in each direction when setting P in x only and zero in y and z

The plots look normal. Pressures fluctuate a lot in MD, especially
in solids. Why do you think anything is wrong? And I don't see
why you are concerned with the pressure for one dimension. In a solid
the other dimensions will be coupled to it.


Dear Steve,

I am looking at crystal/liquid alloy interface system. You are right since in a liquid system if we initially apply different pressure in x/y/z, it will end up with uniform pressure due to Pascal's law.

Since LAMMPS is capable of applying different normal pressures in different directions, I just tried applying pressure in one direction and got surprised by the uniform pressure coming out.

Still I don't see why the pressure should be coupled in solid. Say, in uni-axial compression or tension, the normal pressure in the other two directions should be zero.


you are correct that a solid can sustain different pressures in different
dimensions, but only within limits. If the difference is too large,
or you try to change things too fast, there will be coupling.