Hi. I have a question asking for your help. I am trying to simulate the lattice constant and the corresponding cohesive energy of the Gallium Nitride (GaN) unit cell with in SW potential. I read the manual (http://lammps.sandia.gov/doc/pair_sw.html#Stillinger) and I was confused by the following statement.
"As an example, imagine a file SiC.sw has Stillinger-Weber values for Si and C. If your LAMMPS simulation has 4 atoms types and you want the 1st 3 to be Si, and the 4th to be C, you would use the following pair_coeff command:
pair_coeff * * SiC.sw Si Si Si C
"
How to define “4 atoms types”? I am thinking Si is one type and C is another. What kind of crystal structure is suitable for the above example?
For instance, I use " pair_coeff * * GaN.sw Ga N " for wurtzite, zinc blende and rocksalt GaN structures because there are only two types(Ga and N). Is that correct? Thanks a lot for your time.
Hi. I have a question asking for your help. I am trying to simulate the
lattice constant and the corresponding cohesive energy of the Gallium
Nitride (GaN) unit cell with in SW potential. I read the manual
(http://lammps.sandia.gov/doc/pair_sw.html#Stillinger) and I was confused by
the following statement.
"As an example, imagine a file SiC.sw has Stillinger-Weber values for Si and
C. If your LAMMPS simulation has 4 atoms types and you want the 1st 3 to be
Si, and the 4th to be C, you would use the following pair_coeff command:
pair_coeff * * SiC.sw Si Si Si C
"
How to define "4 atoms types"? I am thinking Si is one type and C is
another. What kind of crystal structure is suitable for the above example?
you are mixing up elements (or rather force field types) with atom
types. different elements (or force field types) *have* to be
different atom types, but you can assign multiple atom types to be of
the same element (or force field type). there are all kinds of reason
to do that. it is a convenient method to single out groups of atoms,
sometimes only for later analysis.
For instance, I use " pair_coeff * * GaN.sw Ga N " for wurtzite, zinc
blende and rocksalt GaN structures because there are only two types(Ga and
N). Is that correct? Thanks a lot for your time.
Here comes the following question. I believe I use the proper setting for create atom and potential (pair_coeff * * GaN.sw Ga N) syntax, then I want to calculate lattice parameter and the corresponding potential energy. I tried a (lattice parameter) from the range 3 to 4 with increment 0.01, I can get corresponding potential energy except a = 3.360. Then I tried the range from 3.350 to 3.370 with increment 0.001, I can get the corresponding potential energy except a = 3.360 (It shows NaN). It happens on both my GaN cell with wurtzite and zinc blende structures. The other thing is I can get reasonable potential energy value when I use a = 3.3599999 and a = 3.3600001, but can not when a is exactly equal to 3.3600. Any suggestions what might be wrong? Thanks in advance.
Here comes the following question. I believe I use the proper setting for
create atom and potential (pair_coeff * * GaN.sw Ga N) syntax, then I want
to calculate lattice parameter and the corresponding potential energy. I
tried a (lattice parameter) from the range 3 to 4 with increment 0.01, I can
get corresponding potential energy except a = 3.360. Then I tried the range
from 3.350 to 3.370 with increment 0.001, I can get the corresponding
potential energy except a = 3.360 (It shows NaN). It happens on both my GaN
cell with wurtzite and zinc blende structures. The other thing is I can get
reasonable potential energy value when I use a = 3.3599999 and a =
3.3600001, but can not when a is exactly equal to 3.3600. Any suggestions
what might be wrong? Thanks in advance.
You are so efficient. It might be the possible direction. I google a bit and find that for pair_style eim (embedded-ion method (EIM) potentials) saying it ignores atomic radius but nothing is mentioned for pair_style sw. Does it mean I have to consider atomic radius for SW potential? In my particular case, is that referring to when I use a = 3.36 two Gallium atoms overlap? I tried, but where can I find the default atomic radius list for Ga and N (or any other elements) in LAMMPS? Thanks.
You are so efficient. It might be the possible direction. I google a bit and
find that for pair_style eim (embedded-ion method (EIM) potentials) saying
it ignores atomic radius but nothing is mentioned for pair_style sw. Does it
mean I have to consider atomic radius for SW potential?
no. you have point particles.
In my particular case, is that referring to when I use a = 3.36 two Gallium atoms overlap? I
tried, but where can I find the default atomic radius list for Ga and N (or
any other elements) in LAMMPS? Thanks.
you are not making much sense here. what i am saying is that you are
likely to have created a bad initial configuration. your wildly
fluctuating energies for the other cases hint that your procedure is
not correct, or that what structure is created may not be what you
think it is.
the way to deal with this is to visualize it carefully and check. or
dump coordinates to files and inspect them.
I had tried that. The purple represents Ga atoms and blue represents N atoms. The attachments show the snapshot from different viewpoint. (Please see the attachment).
It seems generate the "reasonable" initial configuration even when I use a = 3.36, but it gives "nan" for energy. Any suggestions is deeply appreciated.
I had tried that. The purple represents Ga atoms and blue represents N atoms. The attachments show the snapshot from different viewpoint. (Please see the attachment).
It seems generate the "reasonable" initial configuration even when I use a = 3.36, but it gives "nan" for energy. Any suggestions is deeply appreciated.
if the atoms are on top of each other, you don't see anything.
remember, it is near impossible to discuss these issues without seeing
what you are doing.
crystal structure generation is something that is extremely well
tested in lammps, since people use it a lot. so chances are, that you
are doing something wrong.
you may consider moving your box a little bit, to not be bitten by
truncation errors due to floating point math.
