I'm M. Jon Dadras.
I'm trying to run a simulation of tungsten (W) self-ion bombardment using the Tersoff/ZBL potential as developed by Juslin. I want to cumulatively bombard the surface with 1 keV W atoms...my method is the following I have a surface that is ~ (76 Angs)^3 about 26000 atoms; my insertion slab is about 12 Angs above the surface interface in Z. I strike the surface and allow the system to freely evolve for 10 ps, following this I thermostat the system with a Langevin thermostat for 4 ps, and finally I allow for another 1 ps of NVE evolution before the next impact. We loop over this eventually to get up to about 1000 impacts. The script works great up to this point...I'm running on the Kraken at NICS. When I try to delete atoms that are in a region above the insertion plane the program does delete the appropriate atoms and then immediately at the start of the next calculation whether it's still in the same loop or not. I ran some initial tests on my local desktop & everthing seemed fine...but, I cheated I used a much smaller surface and the inserted atoms were moving up, and hence the only ones being deleted. I include my input script & some basic results...the people at Kraken have so far been unable to help me.
LAMMPS input (the name of the file is called blimp.trz, which is why it's called at the end):
# Surface of W in bcc from to be bombarded
units metal
atom_style atomic
boundary p p m
pair_style tersoff/zbl
read_data 2bimpdat.tung
lattice bcc 1.0
region sput block 0.0 75.96 0.0 75.96 87.0 INF units box
region deps block 0.0 75.96 0.0 75.96 85.0 87.0 units box
region tran block 0.0 75.96 0.0 75.96 INF -10.0 units box
group butt id <= 1152
group surf id <> 1153 26511
group imps id >= 26512
#group spts region sput
group trns region tran
pair_coeff * * W-W1.tersoff.zbl W W1
neighbor 0.5 bin
neigh_modify every 10 check yes
label loopa
variable a loop 50
variable e equal $a+40
variable d equal round(random(1,20,$e))
variable b equal round(random(20,50,$e))
variable c equal round(random(50,99,$e))
fix ins imps deposit 1 2 51000 $d8$b$c1 region deps local 10 12 2 vz -323.77 -323.77
fix frz butt setforce 0.0 0.0 0.0
fix trmout all nve
thermo_modify lost ignore flush yes
compute 1$a all temp
compute_modify 1$a dynamic yes
timestep 0.0002
dump 2 all xyz 500 imp$a.xyz
dump 3 all custom 1000 imp$a.dat id type x y z vx vy vz
run 50000
undump 2
undump 3
group spts region sput
fix frz butt setforce 0.0 0.0 0.0
fix trmout all nve
thermo_modify lost ignore flush yes
compute 5$a all temp
compute_modify 5$a dynamic yes
timestep 0.0002
dump 11 spts xyz 10 spt$a.xyz
run 10
undump 11
delete_atoms group spts
group spts delete
variable T equal temp
fix trm surf langevin $T 1000.0 1 248158
fix frz butt setforce 0.0 0.0 0.0
fix trmout all nve
thermo_modify lost ignore flush yes
compute 2$a all temp
compute_modify 2$a dynamic yes
timestep 0.0002
dump 5 all custom 2000 trm$a.dat id type x y z vx vy vz
run 20000
undump 5
fix frz butt setforce 0.0 0.0 0.0
fix trmout all nve
thermo_modify lost ignore flush yes
compute 3$a all temp
compute_modify 3$a dynamic yes
timestep 0.0002
dump 7 all custom 1250 rlx$a.dat id type x y z vx vy vz
run 5000
undump 7
next a
jump blimp.trz loopa