problem with introducing new bonds

Hello Lammps users:

I am simulating cross-linking among a bunch of LJ particles. The initial configurations were created by create_atoms command so there is no data file and there is no bond. After necessary equilibration of the LJ system, the script starts to form quartic style bonds between particles using the fix bond/create:

fix 2 all bond/create 10 1 1 1.3 1 prob 0.1 85784 iparam 4 1 jparam 4 1

According the Lammps manual, bond style and coefficients have to be defined. So I wrote in the script before the above line:

bond_style quartic
bond_coeff 1 1434.3 -0.7411 0.0 1.5 67.2234

Error message comes out at the line of bond coefficients saying “Incorrect args for bond coefficients”. If I define bond style and coefficients before fix bond/create, there would be problem of “Invalid bond type in fix bond/create command”. So can anyone kindly point out what mistake I made in doing? Thanks.

Shaorui

I assume your data file defined bond types even if it
had 0 bonds? And that you are using a atom_style
that allows for bonds? If this doesn't help, please
post a simple, small script and data file that gives an
error.

Steve

Hi Steve:

Thanks for responding. Problem solved. It is my first time using fix atom/create thus I did not know it has to start to with a data file which define bond types. What I first did is to create atoms using the create_atoms command (thus no data file), then equilibrate the LJ system and directly initiate cross-linking. I later created a data file using the restart file of the equilibration process and used it as the starting configuration for creating bonds (some modification was made such as define bond types). Now it works.

Another related questions: Since bond_style quartic requires special_bonds 1 1 1, and fix bond/create requires special_bonds 0 1 1, does it mean we cannot use create quartic style bonds using this fix? I am working around this conflict by first creating FENE bonds then re-defining bond type to be quartic. It seems fine.

Shaorui

Another related questions: Since bond_style quartic requires special_bonds 1 1 1, and fix bond/create requires special_bonds 0 >1 1, does it mean we cannot use create quartic style bonds using this fix? I am working around this conflict by first creating >FENE bonds then re-defining bond type to be quartic. It seems fine.

yes, the special bonds requirements for the 2 commands are incompatible.

Steve