Lattice command's bonding topology

Hello all,

Can anyone tell me how the bonding topology is created when the lattice command is used? I have tried viewing the restart file of Lammps’ diamond silicon structure with the Tersoff pair_style (using the restart2data command), but all I am able to see are the coordinates and velocities with no bonding information. My thought is that the Tersoff potential determines how the atoms are bonded, but I’d like to be able to see this explicitly so that I can make sure I’ve built my own structure correctly using the “lattice custom” command or an input data file.

I am using lammps-8Feb12. I will include the small input file and restart (data) file that I used.

Many thanks for your help,

~Elyse McEntee

d-Si_lattice_topology.in (1.25 KB)

restart_top.dat (1.64 KB)

Hello all,

  Can anyone tell me how the bonding topology is created when the lattice
command is used? I have tried viewing the restart file of Lammps’
diamond silicon structure with the Tersoff pair_style (using the
restart2data command), but all I am able to see are the coordinates and
velocities with no bonding information. My thought is that the Tersoff
potential determines how the atoms are bonded, but I’d like to be able to
see this explicitly so that I can make sure I've built my own structure
correctly using the “lattice custom” command or an input data file.

the lattice command has nothing to do with bonds.
all it does is defining a grid of positions in space for
use by other lammps commands.

the bonds in tersoff potentials are implicit
and determined during the evaluation of the
non-bonded calculations. no explicit bonds
need to be given. in fact, that would be an error.

axel.

Hi Axel,

Thank you so much for your reply. Is there no way to view a data file of the bonds implicitly determined by the Tersoff potential?

I would like to make sure that I view the bonds in my system correctly in VMD when using the “DynamicBonds” setting in the Graphical Representations menu. Should I assume that any atoms within the cutoff distance of each other in the Tersoff potential are implicitly bonded?

Thank you again for your help,

~Elyse

Hi Axel,

  Thank you so much for your reply. Is there no way to view a data file
of the bonds implicitly determined by the Tersoff potential?

there is no bond on/off switch in the tersoff potential.
please have a look at the functional form of the potential
in the documentation and the relevant publications.

   I would like to make sure that I view the bonds in my system correctly
in VMD when using the "DynamicBonds" setting in the Graphical
Representations menu. Should I assume that any atoms within the cutoff
distance of each other in the Tersoff potential are implicitly bonded?

VMD has no clue of the tersoff potential (or chemistry
for that matter). it shows what you tell it to show based
on the information you provide. there is no simple way
to correctly visualize the bonding based on a manybody
potential in a tool that assumes a more conventional
force field with explicit bonds. in fact, the entire concept
of drawing explicit bonds is only fully correct in that
scenario. for everything else, it *always* has to be an
approximation, and it is up to you, to pick the best
settings to minimize the error and to not overinterpret
what you are seeing.

far too many people make the mistake to assume
that there is a little chemist hidden in every computer
and that VMD awakens him and lets him decide what
is a bond or not, or what element an atom is and how
many bonds it can have, even if the data they feed
VMD contains none of that information.

cheers,
     axel.

Hello again,

I’m sorry, I wasn’t very clear in the last email. I completely agree that we don’t have little chemists in our computers. It’s a good thing too, otherwise there would be no use for us graduate students.

I shouldn’t have brought VMD into this thread. I’m very aware that when I tell VMD to draw bonds between any atoms that are within 3.2 Angstroms of each other, it will draw those bonds whether or not they actually exist. This is not my concern. My concern was that I was incorrect to tell VMD to visualize my system in this way. I realize now that obtaining a bonding topology file for the initial step of my simulation will not be feasible or very meaningful for my system.

Thank you for all of your help, time and patience.

Take care,

~Elyse

hi!

Hello again,

  I’m sorry, I wasn’t very clear in the last email. I completely agree
that we don’t have little chemists in our computers. It’s a good thing
too, otherwise there would be no use for us graduate students.

i think you are not giving yourself enough credit.
graduate students have a lot of value beyond that. :wink:

  I shouldn’t have brought VMD into this thread. I’m very aware that
when I tell VMD to draw bonds between any atoms that are within 3.2
Angstroms of each other, it will draw those bonds whether or not they
actually exist. This is not my concern. My concern was that I was
incorrect to tell VMD to visualize my system in this way. I realize now
that obtaining a bonding topology file for the initial step of my
simulation will not be feasible or very meaningful for my system.

you have discovered the big secret of successful research:
it is not about knowing the answers, it is all about asking
the right questions. you should always keep this in mind
when you are looking for answers.

Thank you for all of your help, time and patience.

no problem. it is always rewarding to see when
discussions here on the mailing list not only lead
to solving the immediate problem, but also help
to obtain a better understanding of the methodology

good luck with your simulations,
      axel.

I completely agree that we don’t have little chemists in our computers.

I think John Stone (VMD) is probably working on that for a future release.
He's part of a bio group after all.

Steve