Hey lammps users !
I try to simulate a lattice consisting of 336 molecules (7056 atoms) in an orthorombic geometry.
I would like to calculate the vibrational spectrum of the system under different conditions, and to do so I need to dump the velocities of, say, 20 arbitrary molecules, and then Discrete FFT the data. (Do I ?? would a single molecule be meaningless statistically?)
But, How can I tell lammps to output the velocity data for every atom that belongs to these 20 molecules (say in the mol-id range: 20-40).
The atom style ‘charge’ doesn’t include the mol-id attribute… so do I need to work with the ‘full’ atom style ?
If so, how should I specify each atom to its molecule in the data file ? (I have an xyz format of the crystal which has just the atom id’s and coordinates - no mol-id)
Thanks !
Hey lammps users !
I try to simulate a lattice consisting of 336 molecules (7056 atoms) in an
orthorombic geometry.
I would like to calculate the vibrational spectrum of the system under
different conditions, and to do so I need to dump the velocities of, say, 20
arbitrary molecules, and then Discrete FFT the data. (Do I ?? would a single
molecule be meaningless statistically?)
But, How can I tell lammps to output the velocity data for every atom that
belongs to these 20 molecules (say in the mol-id range: 20-40).
The atom style 'charge' doesn't include the mol-id attribute.... so do I
need to work with the 'full' atom style ?
you can define a group based on atom id. the order of these ids will not
change, so you can map them easily to the molecules, if needed.
with the group you can then generate a specific dump file that only dumps
out data for that group of atoms.
If so, how should I specify each atom to its molecule in the data file ? (I
have an xyz format of the crystal which has just the atom id's and
coordinates - no mol-id)
you may find using the topotools plugin in VMD useful. there is an
incomplete tutorial available here, too:
http://sites.google.com/site/akohlmey/software/topotools
cheers,
axel.
Thanks for your reply !
That surely seems like a simple way to do it.
Another question I have is regarding the reax/c and reax pair styles.
I’ve noticed that all of the new example files (AuO , Zn_O_H, Fe_O_H, V_O_H etc…) in the new versions of lammps use the reax/c pair style.
Is there any reason to prefer that pair style rather the reax pair style ? why did lammps moved to reaxc and left reax behind ?
If so, is there any code to analyze the bonds data (similar to molfrag that works with the reax pairstyle) that fits to reax/c pairstyle ?
I’d be very glad to know where i can get one from !
Best regards,
David
You can also define a group based on molecule IDs, or a
range of them - see the group command doc page for details.
Steve
The 2 versions were implemented by different people. The
C version is faster for most problems. It also has less featues.
E.g. it doesn't tally per-atom stress and energy (yet). So we keep
them both for now. They are a good check agaist each other.
We've found some bugs that way.
Re: analysis, the original ReaxFF has a fix bond/reax - I can't recall
if it also works with the C version. If not, someone could write
a modified version that does ... please volunteer!
Steve