hello all,
i am trying to do two phase simulation…
first i equilibrate atoms at certain temp with periodic boundary condition and npt ensemble…i used simulation box size 0 20 0 20 -21 -1
then i equilibrate atoms at another temp with periodic boundary condition and npt ensemble… i used simulation box size 0 20 0 20 0 20
then i read this dump file to another simulation box and equilibrate it with certain temp with periodic boundary conditions . simulation box size is 0 20 0 20 -21 20 . i used npt ensemble .
problem is dump file it creates is not shows two different region. they mixed with each other
boundary p p p
lattice fcc 4.046
region whole block 0 20 0 20 -21 20 units box
create_box 2 whole
region solid block 0 20 0 20 -21 -1 units box
create_atoms 1 region solid
mass 1 26.98
region liquid block 0 20 0 20 0 20 units box
create_atoms 2 region liquid
mass 2 26.98
read_dump dump_solid 80000 type vx vy vz box yes replace yes
read_dump dump_liquid 79990 type vx vy vz box yes replace yes
echo screen
group solid type 1
group liquid type 2
dump 5 all custom 10 dump_liquidrelax id type x y z vx vy vz
thanks in advance
hello all,
i am trying to do two phase simulation..
first i equilibrate atoms at certain temp with periodic boundary condition
and npt ensemble..i used simulation box size 0 20 0 20 -21 -1
then i equilibrate atoms at another temp with periodic boundary condition
and npt ensemble.. i used simulation box size 0 20 0 20 0 20
then i read this dump file to another simulation box and equilibrate it with
certain temp with periodic boundary conditions . simulation box size is 0 20
0 20 -21 20 . i used npt ensemble .
this is a bad protocol, especially since you have periodic boundary
conditions. image flags and thus positions will be invalid after
merging the systems. even, if you correct for that (or use fixed
boundaries with walls in z-direction), you are creating a high
potential energy area near the interfaces.
a) compute your target density for the entire system and determine a
suitable simulation cell size
b) build a system that has both phases already included. define each
phase into a separate group
c) run a minimization to remove close contacts and thus high potential
energy collisions.
d) set up a simulation with fix nve and fix langevin or fix temp/csld
that operates at only one of the two subsystems, the other remains
immobile
e) switch fix nve and the thermostat fix to the other subsystem and
equilibrate that for a bit, while keeping the other part immobile.
f) repeat steps d) and e) a couple of times while starting with a
rather short Tdamp and increasing it for each iteration.
g) switch to fix nvt for the whole system for a bit and only then
switch to fix npt.
this is simpler, faster, and safer.
axel.