For the past few months I have been having problems losing bonded atoms and other issues. Basically, I sometimes have issues with atoms not “seeing” eachother and sometimes exploding and causing problems. Sometimes I also have a problem where the molecule crumples up and somehow finds an unrealistic energy minimum which I assume is electrostatic (due to the atoms being closer) Basically, I think I am having Lennard-Jones problems somehow. Can anyone see anything immediately wrong with this in.file? I would really appreciate a second look.
Much thanks,
Josh Deetz
PhD Candidate Student
Chemical Engineering
University of California, Davis
§ 408-242-5523
Single molecule tri-ethoxysilane
units real
atom_style full
kspace_style ewald 0.0001
pair_style lj/class2/coul/long 15.0
dielectric 1.0
pair_modify mix sixthpower tail yes
bond_style class2
angle_style class2
dihedral_style class2
boundary p p p
read_data data.singleclass2modified
special_bonds lj/coul 0.0 0.0 1.0 dihedral yes
neighbor 2.0 bin
neigh_modify delay 5
timestep 0.5
thermo_style multi
thermo 50
fix 1 all nvt temp 275.0 275.0 100.0
group tes type <= 7
dump 1 tes atom 5 dump.singleclass2
minimize 1.0e-4 1.0e-6 1000 1000
run 30000
Joshua:
There’s not really enough to go on here. Does this happen all of the time, or just sometimes? Does the problem occur at the outset of the simulation, or is it something that occurs later on in the simulation? Is the problem reproducible if you rerun a problem under the same conditions (same restart file, same number of processors, etc.)?
Are you starting from a known “good” configuration? Many times problems occur because the starting configuration is energetically unstable, and a long time step will cause the system to blow up pretty quickly. But that’s just a hunch, without more information to work with.
—AEI
For the past few months I have been having problems losing bonded atoms and
other issues. Basically, I sometimes have issues with atoms not "seeing"
eachother and sometimes exploding and causing problems. Sometimes I also
have a problem where the molecule crumples up and somehow finds an
unrealistic energy minimum which I assume is electrostatic (due to the atoms
being closer) Basically, I think I am having Lennard-Jones problems somehow.
Can anyone see anything immediately wrong with this in.file? I would really
appreciate a second look.
what you describe sounds much more like you have
a bad model (i.e. bad or incorrectly assigned parameters)
and that would not show at all in this part of the input.
this can be as simple as bad unit conversion or missing
cross terms in the parameters, or something worse and
more subtle. it is impossible to debug this from just an
input and even for a complex system. i suggest you go
back to the basics and evaluate each part of your system
independently and try to identify which part is off and
then try to understand why.
axel.
Turns out that somehow the mixing rules were not being applied here, which is confusing as I thought the “pair_modify mix sixthpower” command would have done this automatically. After specifying the parameters for all homo- and hetero-atoms explicitly in the input file, I now have no issues. Thanks for giving it a second look.
From the class2 doc page:
For atom type pairs I,J and I != J, the epsilon and sigma coefficients
and cutoff distance for all of the lj/class2 pair styles can be mixed.
Epsilon and sigma are always mixed with the value {sixthpower}. The
cutoff distance is mixed by whatever option is set by the pair_modify
command (default = geometric). See the "pair_modify" command for
details.
You should not have to specify a mix command via pair_modify, as
it is on by default. If it's not working this way, let us know, as
it might be a bug.
Steve