Hello,
I want to simulate laser ablation on silicon , I generated the input file and the results show that the temperature raise as it is expected but the atoms don’t move at all. Any idea about the source of this problem is appreciated. My input file is as below:
1 Like
dear leyla,
Hello,
I want to simulate laser ablation on silicon , I generated the input file
and the results show that the temperature raise as it is expected but the
atoms don't move at all. Any idea about the source of this problem
there are multiple problems with your input.
is appreciated. My input file is as below:
______________________________________
units metal
atom_style atomic
boundary p p f
pair_style sw
read_data DATAFILE.LAMMPS
group silicon type 1
group base type 2
pair_coeff * * si.sw Si Si #All
neighbor 0.30 bin
why change this from the default?
neigh_modify delay 0 every 1 check yes
run_style verlet
thermo_style custom pxx pyy pzz pxy pxz pyz temp
thermo 10
velocity silicon create 6.4 4928459 dist gaussian
you are looking at a system at 6.4 kelvin? seriously?
fix 1 silicon heat 10 1
fix 2 silicon nve
fix 3 base nvt temp 6.4 6.4 2
compute Mytemp silicon temp
restart 10000 ./restart/restart.file.*
dump 1 all atom 10 ./results/dump.SI.*
timestep 0.0000001
you have a time step of a 1/10th of an attosecond
are you serious about this?
remember, that you have classical particles here
not electrons or even a plasma.
your timestep is off by 3-4 orders of magnitude.
run 5000
5000 steps is "nothing".
axel.
Dear Axel,
Thanks for your reply, the temperature is 300K which is 6.4 in reduced units. I know the time steps is too small but I just wanted to check to see how it goes. Even in that time steps, the temperature rises dramatically but the atom don't move. I still don't understand why the atoms are not moving.
Regards,
Leyla
Leyla,
You are using "units metal" so there is no need to convert to reduced
units. (The temperature defined for units metal is in degrees K
already). See http://lammps.sandia.gov/doc/units.html
Jan-Michael
Dear Axel,
Thanks for your reply, the temperature is 300K which is 6.4 in reduced units. I know the time steps is too small but I just wanted to check to see how it goes. Even in that time steps, the temperature rises dramatically but the atom don't move. I still don't understand why the atoms are not moving.
your reasoning doesn't make sense. the atoms are not
moving *because* of the small time step.
before using fix heat, you should first equilibrate the system
and verify that you can run a regular MD correctly.
axel.
Hi,
Maybe the time step is soo small that there is not enough time for the
atoms to relact....Usually A relaxation time must be incorporated in
the simulations, if the time step is smaller than the max frequency
of vibration (density of states), then the simulation makes no
sence...
PS : My answers are based on pure intuition and some luck
Oscar G.
Why the temperature rises dramatically. I used these commands to fix the temperature at 300k. But the temperature becomes 10 times larger every time steps!
fix 1 silicon nvt temp 300.0 300.0 100.0
fix 2 silicon temp/rescale 100 300 300 0.01 1.0
Why the temperature rises dramatically. I used these commands to fix the temperature at 300k. But the temperature becomes 10 times larger every time steps!
a bad initial configuration??
does the same happen without a thermostat?
please keep in mind that MD is strictly "garbage-in, garbage-out",
i.e., if you don't feed the program the right data, you won't get the
right answer. software is extremely unsophisticated. it doesn't
know what you want, it only does exactly what you tell it to do,
regardless how meaningless or how "wrong".
axel.
Hi,
Perhaps the integrator is having a hard time in trying to mantain the
temperature at a constant, this probably caused by and instability (a
sudden phase change) or a bad input script...
Oscar G.