bond atoms missing on proc error

Dear fellow users,

My simulations ended with the error: both bond atoms missing on proc. I know it means that “typically this is because the pairwise cutoff is set too short or the bond has blown apart and an atom is too far away”.
I did some research, but I really have no clue what went wrong. My system has 4096 atoms. Two of them are methane, and the rest are water. The two methanes are bonded. And
a special_bond command was used to calculate pairwise interaction between the two methanes. My major concern is the bond length, which ranged from 8.4-16.4 angstrom. Do you think the bond length is the reason
that simulations keep failing?

My input file is as below:

variable k equal 1
variable Ro equal 13.4

variable dx equal x[1]-x[2]
variable dy equal y[1]-y[2]
variable dz equal z[1]-z[2]

variable bx equal xhi-xlo
variable by equal yhi-ylo
variable bz equal zhi-zlo

units real
atom_style full
boundary p p p

special_bonds lj/coul 1. 0. 0.

pair_style hybrid sw0 sw
bond_style harmonic

read_data data.4094w_2methanes5A

newton on
pair_coeff * * sw0 mW-methane.sw Si Ge
pair_coeff * * sw mW-methane.sw Si NULL

neighbor 3.0 bin
neigh_modify every 2 delay 4 check yes

timestep 5

thermo 2000
thermo_modify flush yes

velocity all create 300 4928459 rot yes dist gaussian
fix 5 all nvt temp 300 300 1000
#dump 1 all custom 10000 nvt-300K.lammpstrj id type xs ys zs

fix scales methane print 100 ‘{dx} {dy} {dz} {bx} {by} {bz}’ file scales.${Ro}-$k screen no
restart 500000 restart_300K
run 5000000
write_restart restart.end_300K

Thank you.

Bin

Dear fellow users,

My simulations ended with the error: both bond atoms missing on proc. I know
it means that "typically this is because the pairwise cutoff is set too
short or the bond has blown apart and an atom is too far away".
I did some research, but I really have no clue what went wrong. My system
has 4096 atoms. Two of them are methane, and the rest are water. The two
methanes are bonded. And
a special_bond command was used to calculate pairwise interaction between
the two methanes. My major concern is the bond length, which ranged from
8.4-16.4 angstrom. Do you think the bond length is the reason
that simulations keep failing?

you cannot have a bond that is longer than the largest non-bonded
cutoff plus "neighbor list skin" distance.

if you need a harmonic constraint between those two molecules,
you can use fix spring in couple mode with 0.0 0.0 0.0 as vector and
r0 as the equilibrium distance and k as the force constant.

axel.