I have a pdb file with four atoms, say, A-B-C-D. I used topotools to generate a data file with a list of bond, angles, and dihedrals using this pdb file. In the LAMMPS input file I am using “boundary p p p”, ie. periodic boundary condition.
Will LAMMPS treat the chain as disconnected chains in its periodic images or LAMMPS will use nearest neighbor list to generate bonds between atom in the original box with its periodic images ? More explicitly, do I need to add a bond between end atoms (A and D) in topotools script to generate a data file where the original chain is connected to its periodic images combined with periodic boundary condition for LAMMPS to treat the systems as connected infinite chain.
Lammps will not create any bonds unless you ask it to. Since there is no bond between A and D, you have to add it. How should lammps know that you want that bond to be there?
If I change the above example to A-B-A-B, do I still need to add a bond between the end atoms (first A and the the last B) to create an infinite chain ? or the periodic boundary condition in LAMMPS will take care of connecting chains based on nearest neighbor, since now I have a bond defined between A and B atoms within a unit cell ?
is this a joke?
Are you trying to provoke me into verbally abusing you?
Changing the type of an atom does not change the topology of the molecule.
So: same question, same answer.
What Axel is telling you is that bonding is between atom ids not atom types. Simply turn your example into: A-B-A-B-A-B and get a idea for the mess that quickly develops if bonding was to be based on atom types.
Vivek, you “need” to add the bond. think it in this way–> PBC means whatever is in the unit cell, is replicated in the images. "Whatever means whatever, i.e. if bond is there, then its replicated and vice-versa. Now you say that you have bond “within” the unit cell, ok.! But what about the boundaries?? At the farthermost boundary, if there is no bond, how can you expect your chain to be infinitely long then. If u still didnt get the point, then just test the concept with a small test run and see the difference tin the results, you will have the insight…