[lammps-users] bond missing

Hi all,

i am simulating the interaction between for example CO2 and a nanotube. The script was :

dimension 3
boundary p p p
units real
lattice none
neighbor 0.3 bin
neigh_modify delay 0 every 1 check yes
atom_style full
bond_style harmonic
angle_style harmonic
pair_style lj/cut/coul/long 8 10
#pair_modify mix arithmetic
kspace_style pppm 1e-4
######Data file######################################
read_data MWNT-CO2.dat

2010/9/21 Jim Jim <drazentl@…8…>

Hi all,

i am simulating the interaction between for example CO2 and a nanotube. The script was :

dimension 3
boundary p p p
units real
lattice none
neighbor 0.3 bin
neigh_modify delay 0 every 1 check yes
atom_style full
bond_style harmonic
angle_style harmonic
pair_style lj/cut/coul/long 8 10
#pair_modify mix arithmetic
kspace_style pppm 1e-4
######Data file######################################
read_data MWNT-CO2.dat
#####################################################

Type 1 : C_CO2 atoms

Type 2 : O_CO2 atoms

Type 3 : C* atoms

group C type 1
group O type 2
group tubes type 3
group CO2 molecule <> 1 5000

group CO2 union C O

#####################################################
pair_coeff 1 1 0.0558980835347 2.757 #C-C epsilon-sigma
pair_coeff 2 2 0.159983896019 3.033 #O-O epsilon(kcal/mol) sigma(A)
pair_coeff 3 3 0.6 3.0 #C*-C*
pair_coeff 1 3 0.0558980835347 2.757 #C-C*
pair_coeff 1 2 0.0945663427646 2.895 # C-O
pair_coeff 2 3 0.2 2.6 #O-C*
######################################################
bond_coeff 1 1000 1.16
angle_coeff 1 1000 180.0

velocity tubes set 0.0 0.0 0.0 units box
velocity tubes zero angular
velocity tubes zero linear
fix 1a tubes setforce 0.0 0.0 0.0
fix 1b tubes momentum 1 linear 0 0 0 angular
fix 2a CO2 nve/limit 10.0
fix 2b CO2 temp/rescale 100 300.0 300.0 50.0 0.8
fix 3 tubes rigid single
#####################################################
velocity CO2 create 300.0 4928459 units box

Print Energies on log.lammps

thermo_modify lost warn
thermo_style multi
thermo 1000
#####################################################
#dump 1 all atom 500 nve_nvt.lammpstrj
timestep 0.005
run 50000

The error that i am receiving is : Bond atoms 6691 6693 missing on proc 1 at step 20
with the atoms that are missing to vary as i change some parameters in order to see what am i doing wrong. i read the manual which says that the problem is either a short cut off or the bond has blown apart. The first one is not happening so it may be the 2nd. However when i print the E_bond i see that it is always set to 0.00 even if i have put a bond coef. Why is this hapenning?

---------------- Step 0 ----- CPU = 0.0000 (sec) ----------------
TotEng = 15405821207.3231 KinEng = 13412.7510 Temp = 299.9600
PotEng = 15405807794.5722 E_bond = 0.0000 E_angle = 0.0000
E_dihed = 0.0000 E_impro = 0.0000 E_vdwl = 15405808494.6240
E_coul = 69582.6359 E_long = -70282.6878 Press = 12806437.2719
ERROR on proc 1: Bond atoms 6691 6693 missing on proc 1 at step 20

Jim:

The bond error is the symptom, not the cause of your problem! Look at the van der Waals energy of your system (> 1.5e10). What’s causing your system to blow up is that two nonbonded atoms are getting too close together. The resulting force pushes one of the atoms in the bond very far away, and thus you get the bond error.

(I don’t think the settings on your nve/limit command help to avoid the problem, either.)

–AEI

It's odd that the bond energy is 0.0, but only
if it stays that way for many timestep. It could be
0.0 on the step 0, b/c you built the system with
all bonds at equilibrium length. Try simulating CO2
w/out the tube and see if you can get normal dynamics
for them. Your timestep of 0.005 fmsec is very small - why
did you pick that?

Steve

2010/9/21 Jim Jim <[email protected]...>:

I put a small timestep in order to minimize my system , because when i have the molecules with the nanotube i cannot use the minimize. When i changed the equilibrium length bond energy changed immediately but again the there is the error about the bond missing.However when I had tried to simulate the CO2 w/out the tube and everything worked fine with bond energy different than 0.00 . Moreover when i tried to put a fix shake for the CO2 molecule i took nan for the pressure so in order to keep the molecules rigid i take the spring parameters with high values.

I don't think you can use fix shake with a linear molecule,
at least some people have reported problems. But you
can use fix rigid for them. If you create a bad model
initially (bad thermo output), then you cannot hope
to do dynamics on it.

Steve

2010/9/21 Jim Jim <[email protected]...>: