I have a simulation box with about 120*40*40 (x*y*z) angstrom^3 and set npt ensemble via:
fix ffix all npt temp 1300 1300 0.1 iso 0.0 0.0 1.0 # timestep 1 fs
I checked the pressure tensor via command:
fix fpressure all ave/time 100 10 1000 c_thermo_press file pressure.lmp mode vector ave one
and found xx, yy, and zz pressure (when system is in equilibrium) are about +4000, -2500 and -1500 (unit: metal). I was wondering why the pressure is not isotropic, i.e., the same in xx, yy and zz directions (and around zero). Thanks.
The system is 120x40x40 angstrom along x-y-z direction. Does iso
pressure only applies for cubic system?
no, but if it _was_ a cubic system,
then the pressure _should_ be isotropic.
the box dimensions themselves have rather little relevance
as to whether the pressure should be isotropic or not, but
what you have _in_ your system matters a lot.
since you don't provide any information, there is not much
else to do than to speculate.
The simulation cell contains CuZr alloy, with a fat slab of CuZr crystal in the middle and CuZr liquid filled in the cell. The solid/liquid interfaces are along y-z plane and all atoms are free to move. Thanks.
I guess the non-equivalence of x,y,and z pressure components upon iso setting has to do with the initial dimensions of the simulation cell. If, for example, the initial sizes of the cell are such that pressure_xx is positive and pressure_yy and pressure_zz are negative, then the 'dimension coupling' of x,y,z direction due to iso setting makes it difficult for the pressure to be uniform, especially for non-liquid system. In this case aniso setting may help.