Unstable model of C16H34 using Reax potential

Dear LAMMSPS users,

I am using LAMMPS to simulate the dynamic of alkanes using Reax potential, which developed by Andri van Duin 2001 (ffield.reax.CH) and Mattsson 2010 (ffield.reax.mattsson).

Firstly, I started with a molecules of C16 in 3D model. The initial structure of C16 molecule was built in Material Studio software using COMPASS force field. The input script is described in in.C16 file (attached above), and the structural data is shown in C16_3D.lammps05 file.

The obtained componential energies (output_ReaxEnergyComponents_Temp300) show that the initial bonding, atomic, … energies are zeros (except Van der Waals, Coulomb, and qeq energies). The Coulombs and qeq energies are significantly larger than Van der Waal. For higher temperature, the model is unstable. The C-H bonds are broken and H atoms move away from C (C16_3D_Animation.avi). Furthermore, the number of species increased from 3 (H C(H3) C(H2)) at 1-10 K into 4 species at 100 K and 5 species at 300 K.

I has searched some discussions regarding this issue in lammps-user forum and tried some possible ways to fix this problem such as reducing time step (0.001-0.01 fs), gradually increasing the temperature (1-300 K), optimizing the structure before running dynamic. Unfortunately, I can’t fix it.

I write this email with an expectation that someone can tell me what is the problem with my model.

Many thanks for your help !

Best regards,

Thi Dinh TA

PhD Candidate

School of Mechanical, Materials & Mechatronic Engineering

University of Wollongong

NSW 2522, Australia

Phone : (+61) 405 234 812

Email: [email protected]…3974…

C16_3D.lammps05 (13.8 KB)

ffield.reax.CH (4.64 KB)

ffield.reax.mattsson (14.8 KB)

in.C16 (4.86 KB)

lmp_control (992 Bytes)

output_DumpAllCustom.lammpstrj (199 KB)

output_ReaxEnergyComponents_Temp300 (337 KB)

species.out (511 KB)

C16_3D_Animation.avi (1.26 MB)

Dear LAMMSPS users,

I am using LAMMPS to simulate the dynamic of alkanes using Reax potential,
which developed by Andri van Duin 2001 (ffield.reax.CH) and Mattsson 2010
(ffield.reax.mattsson).

Firstly, I started with a molecules of C16 in 3D model. The initial
structure of C16 molecule was built in Material Studio software using
COMPASS force field. The input script is described in in.C16 file (attached
above), and the structural data is shown in C16_3D.lammps05 file.

The obtained componential energies (output_ReaxEnergyComponents_Temp300)
show that the initial bonding, atomic, … energies are zeros (except Van der
Waals, Coulomb, and qeq energies). The Coulombs and qeq energies are
significantly larger than Van der Waal. For higher temperature, the model
is unstable. The C-H bonds are broken and H atoms move away from C
(C16_3D_Animation.avi). Furthermore, the number of species increased from 3
(H C(H3) C(H2)) at 1-10 K into 4 species at 100 K and 5 species at 300 K.

I has searched some discussions regarding this issue in lammps-user forum
and tried some possible ways to fix this problem such as reducing time step
(0.001-0.01 fs), gradually increasing the temperature (1-300 K), optimizing
the structure before running dynamic. Unfortunately, I can’t fix it.

I write this email with an expectation that someone can tell me what is
the problem with my model.

​there is one major problem with your input deck: you are defining explicit
bonds, angles and dihedrals in your ​data file. for reaxff, however, those
are defined implicitly by the pair style, so you must not list them
explicitly, as these explicit bonds cause the corresponding pairs to be
excluded from the nonbonded computation. as a workaround, you may try using:

special_bonds lj/coul 1.0 1.0 1.0

​but i strongly recommend to edit your data file to remove all bonds,
angles and dihedrals.

axel.​

1 Like

2016-03-31 12:11 GMT+02:00 Dinh Ta <[email protected]>:

Dear LAMMSPS users,

I am using LAMMPS to simulate the dynamic of alkanes using Reax potential,
which developed by Andri van Duin 2001 (ffield.reax.CH) and Mattsson 2010
(ffield.reax.mattsson).

Firstly, I started with a molecules of C16 in 3D model. The initial
structure of C16 molecule was built in Material Studio software using
COMPASS force field. The input script is described in in.C16 file (attached
above), and the structural data is shown in C16_3D.lammps05 file.

You should, at the very least, get rid of the bonds, angles, dihedrals, in
your data file (Reax doesn't need them).