Question regarding bond/create

Hello everyone,

I’m kind of new to LAMMPS and I’m stuck with a problem with the bond/create function. I’m getting the following error:

Fix bond/create cutoff is longer than pairwise cutoff

but I haven’t defined any pairwise cutoff nor know where they are!

I tried to use lower distances in my bond/create but that didn’t work either.

Any suggestion?

This is my input file:

units real
atom_style full
log sim.log

For a single processor calculation

variable T equal 273 # Simulation temperature
variable salt equal 100.0 # Salt concentration [mM]

variable random equal 12345

bond_style morse
angle_style none
dihedral_style none
pair_style none

boundary p p p
newton on #yes
read_data bdna_conf.in

neighbor 4.0 multi
neigh_modify check no
atom_modify sort 0 0.0

A timestep of 0.01 ps

timestep 10

group mobileatom id 1:11
group fixatom type 5

Initialize velocities from a Gaussian distribution

velocity mobileatom create {T} {random} rot yes mom yes dist gaussian

Specifying a Langevin integrator to perform a simulation in the NVT ensemble

fix 1 mobileatom langevin {T} {T} 5000 ${random} gjf yes
fix 2 mobileatom nve
fix 3 all bond/break2 1 10 0.5
fix 4 all bond/break2 1 9 0.5
fix 5 all bond/create 1 1 5 0.5 9
fix 6 all bond/create 1 2 5 0.5 9
fix 7 all bond/create 1 3 5 0.5 10
fix 8 all bond/create 1 4 5 0.5 10

Calculating the different components of the non-bonded energy

compute 1 all bond/local dist eng

Specifying the frequency of thermodynamic output

thermo 500
thermo_style custom step ebond etotal

Write configuration to file

dump 1 mobileatom custom 500 traj.xyz x y z

Run X number of steps

run 25000

Thanks a lot!

A pair_style command will define the pair cutoff.

Do you not need a pair style defined to prevent

atoms that are not part of the same bond/angle/etc

from overlapping, in your model?

Steve

So far I’m only working with bonds, no pair interactions

Had to refresh myself on how fix bond/create works. You must define

a pair style, b/c the way each atom finds possible new atoms to

bond with is by checking its neighbor list. If you really want a model

where there are no pair interactions (between atoms that are not bonded),

then you can just define pair lj/cut and set all the epsilon values (pair_coeff)

to 0.0. The cutoff(s) you assign with the pair style/coeff commands will allow

fix bond/create to work.

Steve