Hi,

i'm trying to define embed a prism region in a simulation but i'm afraid there is something i don't understand well.

Briefly, i have a drop of water and i bound it with a wall below and above in the z-direction. In the wall below, i embed particular regions (in the example below the embedded region is a strip infinite in y and finite in x and z)

This example works:

region 1 block INF -5 INF INF INF -6 units box side out

region 2 block -5 -5 INF INF -6 -2 units box side out

region 3 block -5 5 INF INF INF -2 units box side out

region 4 block 5 5 INF INF -6 -2 units box side out

region 5 block 5 INF INF INF INF -6 units box side out

fix zero all wall/lj126 zhi 200.0 0.03 3.0 12.0 units box pbc yes

fix first all wall/region 1 lj126 0.03 3.0 12.0

fix second all wall/region 2 lj126 0.03 3.0 12.0

fix third all wall/region 3 lj126 0.03 3.0 12.0

fix fourth all wall/region 4 lj126 0.03 3.0 12.0

fix fifth all wall/region 5 lj126 0.03 3.0 12.0

This other, however, which now includes a prism, reports lost atoms

region 1 block INF -5 INF INF INF -6 units box side out

region 2 prism -5 5 INF INF -6 -2 0 0 0 units box side out

region 5 block 5 INF INF INF INF -10 units box side out

fix zero all wall/lj126 zhi 200.0 0.03 3.0 12.0 units box pbc yes

fix first all wall/region 1 lj126 0.03 3.0 12.0

fix second all wall/region 2 lj126 0.03 3.0 12.0

fix fifth all wall/region 5 lj126 0.03 3.0 12.0

I thought both are equivalent but clearly they are not. Not sure why though.

I'm using LAMMPS 11 Sep 2014.

Thanks

Miguel