Hi all,
I have a question about the effect of nve ensemble and fix langevin command
Q1:In order to simulate ion-beam processes at nonzero temperatures,
the simulation cell must be kept at constant temperature algorithm.
I employ the Langevin equation of motion, but apply it only to the atoms at and near the cell boundaries, as described below.
-----------------lammps code----------------------
units metal
timestep 0.0001
velocity group_si create 300.0 42169 dist gaussian units box
velocity helium set 0.0 0.0 $v units box
fix 1 group_si_boundary langevin 300.0 300.0 10000 234103
fix 2 all nve
thermo_style custom step etotal temp press lx ly lz vol
Hi all,
I have a question about the effect of nve ensemble and fix langevin command
Q1:In order to simulate ion-beam processes at nonzero temperatures,
the simulation cell must be kept at constant temperature algorithm.
I employ the Langevin equation of motion, but apply it only to the atoms
at and near the cell boundaries, as described below.
-----------------lammps code----------------------
units metal
timestep 0.0001
velocity group_si create 300.0 42169 dist gaussian units box
velocity helium set 0.0 0.0 $v units box
fix 1 group_si_boundary langevin 300.0 300.0 10000 234103
fix 2 all nve
thermo_style custom step etotal temp press lx ly lz vol
--------------------------------------------------
"v" is initial velocity of the helium ion.
When I run this lammps code, value of "Temp" comes down (under 300K),
and value of Press rises, as below.
Why ?
why not?
you have only a few lines of input there,
so it is very difficult to make any comments
without speculating about what is not there.
the "temp" you measure is the temperature of
the total system. you are thermalizing only
a subset and you make no mention of equilibration.
if you have your initial atom positions on ideal
lattice positions, assign kinetic energy and
then integrate with nve, about half of that kinetic
energy will convert into potential energy as
part of the equilibration process.
Should I use "fix temp/rescale" command ?
NO!! temp/rescale is evil. never use it.
the only good thing about it is that it
is extremely easy to implement.
axel.
Dear Axel,
thanks for your advices.
> why not?
>
> you have only a few lines of input there,
> so it is very difficult to make any comments
> without speculating about what is not there.
I'm sorry.
Initial Si atoms setting commands are described below
#--Initial Si atoms setting commmad--
lattice diamond 5.43
region simbox block -10 10 -10 10 -570 5
region target block -10 10 -10 10 -550 -0
region target_in block -8 8 -8 8 -548 -2
region wall_x1 block -10 10 -10 -8 -548 -2
region wall_x2 block -10 10 8 10 -548 -2
region wall_y1 block -10 -8 -8 8 -548 -2
region wall_y2 block 8 10 -8 8 -548 -2
region wall_z1 block -10 10 -10 10 -2 0
region wall_z2 block -10 10 -10 10 -550 -548
create_box 3 simbox
create_atoms 1 region target
mass 1 28.0
group group_all region simbox
group group_si region target
group group_si_in region target_in
group group_si_wall region wall_x1
group group_si_wall region wall_x2
group group_si_wall region wall_y1
group group_si_wall region wall_y2
group group_si_wall region wall_z1
group group_si_wall region wall_z2
if you have your initial atom positions on ideal
lattice positions, assign kinetic energy and
then integrate with nve, about half of that kinetic
energy will convert into potential energy as
part of the equilibration process.
It's surprising!!
this is not surprising at all.
there is a very simple way to rationalize this.
consider an ensemble of harmonic oscillators
that are weakly coupled, e.g. a chunk of atoms
attached to its lattice positions by a spring.
now, for each single oscillator, you start with
the atom at rest in the location where the potential
energy is minimal and the kinetic energy would be
maximal and you assign 300K worth of kinetic
energy to it. next you do the integration of the equation
of motion and for each oscillator the kinetic energy
gets completely(!) converted into potential energy
at the turning point and converted back into potential
energy and back and forth. if you finally consider
that each individual oscillator would "started" at a
random time, then you would have your entire system
somewhere *between* the two extrema of potential
and kinetic energy and you would expect that about
half the kinetic energy is converted into potential energy.
I would have misunderstood about equilibration process.
To maintain the silicon's temperature, Should I set like below?
velocity group_si create 600.0 4223169 dist gaussian units box
that would help as you would insert more kinetic energy.
Before ion implantation, do I have to wait a silicon substrate becomes
equilibrating to about 300[K] ?
exactly. and for that you can also assign fix langevin
to those atoms to speed up the process of equipartitioning
the kinetic energy. that won't guarantee equilibration,
but is a prerequisite.
it may help to look up these issues in a MD and stat mech
text book where they are presented in much greater detail.
axel.