the electrostatic force of terminal hydroxyl group on a nanotube

LAMMPS LAMMPS Mailing List Mirror
1

hi there.

I bulid a carbon nanotube with hydroxyl groups covalently linked to its port . I try to relax the model (e.g. under nvt 2K with a 0.01fs timestep )in lammps but it break up everytime because of the huge the electrostatic force between non-boned but neighboring O and H atom.

O 863 3 24.5895 10.2525 37.5525 0 -0 -0 9.45942e+006 -2.14143e+006 -1.55851e+007
H 889 4 24.5906 10.2522 37.5507 -0 0 0 -9.46005e+006 2.14138e+006 1.55852e+007

when I add a non-boned interaction between O & H atoms, the model functions well but according to oplsaa forcefileld there exists no interaction between O & H atomes in hydroxyl groups.

is there any other ways to solve the problem?

thanks in advance!

2

For the O and H atom you list, are you sure they are not bonded indirectly, e.g. a 1-3 or 1-4

interaction? If so, you should insure that your special_bonds settings are correct

for weighting that interaction, possibly the weight should be 0.0.

If they are not connected as close as 1-4, then I don’t see how there could be

no OPLSAA nonbond interaction defined for them.

Steve

3

it’s ture, the O and H atom are not bonded indirectly . （1-5 interaction）

in OPLSAA ， the lj parameters for H of “Phenol -OH” which i think is suitable for cnt-oh is 0 0 . maybe it is not the

way I think.

thanks a lot!

4

it's ture, the O and H atom are not bonded indirectly . （1-5 interaction）

in OPLSAA ， the lj parameters for H of "Phenol -OH" which i think is
suitable for cnt-oh is 0 0 . maybe it is not the

way I think.

it is very likely not the way you seem to think. it looks like you
haven't studied carefully enough what encompasses the OPLS/AA model.
LJ is not the only interaction. the OPLS potential is build from
non-bonded (via lennard-jones and coulomb) and bonded interactions
(bonds, angles, dihedrals) for which non-bonded interactions are
either excluded (1-2, 1-3) or scaled (0.5 for 1-4 pairs).with all
those interactions in place and using a correct bond topology, thing
should behave well.
from your description, it sounds as if your setup and your topology are a mess.

axel.

5

thanks for your timely reply!