I am trying to simulation rigid sphere (C180 Buckyball) impact on the
graphene sheet using NVE ensemble with AIREBO potential.
I have relaxed the graphene and the sphere for 10 ps before impact.

According to the impact mechanics, during impact kinetic energy (KE) of the
sphere will transfer into potential energy (PE) and kinetic energy
of the graphene and total energy (PE + KE) of the sphere and graphene should
be conserved. In the simulation,
I am observing before and after impact system total energy is conserved.
However, during impact (around 1.6 ps) total energy is not conserved.
Total energy is changing from -113000 eV to - 112564 eV during the impact.
I have also attached the Fig as ppt. I am sure whether you
can see it.

I have also tried initial relaxation with NVT ensemble. No significant

In another simulation, I have checked the only the rigid body motion of the
C180 Buckyball at 5 km/s with "rigid/nve single" fix command.
This simulation result matches with the theoretical prediction of KE,
displacement etc.

I am using 2012 version of lammps. I have attached input script, and data

Would you please make comments regarding why energy is not conserved during
impact? How can energy be conserved during impact?


Kind regards,
University of Delaware

# input script for Sphere Impact on Graphene

# Variables
variable input index input_1-20-20-gs-1-C180.txt

units metal
boundary p p p
atom_style atomic

processors 6 6 1


read_data ${input}

# Define regions

region left block INF -90.0 INF INF INF INF units box
region right block 90.0 INF INF INF INF INF units box
region front block INF INF INF -90.5 INF INF units box
region rear block INF INF 90.5 INF INF INF units box
region section block INF INF 0.00 INF INF INF units box

group left region left
group right region right
group front region front
group rear region rear
group section region section

group gs type 1
group bball type 2

group center subtract gs left right front rear


pair_style airebo 3.0 1 1
pair_coeff * * CH.airebo C C


neighbor 1.0 bin
neigh_modify delay 5

# ------------------------------ RELAXATION

GS-Impact-Results.pptx (182 KB)


Have you tried with smaller timesteps? The default timestep of 1 fs may be too large for a such high impact velocity.



Any macroscopic collision between object will convert some of the energy
kinetic energy into internal energy of other forms of energy … I dont belive KE = PE (exactly) … I bet the temperature rises, which mean HEAT must also be considered in the impact …

Since you already made an script that presumptuously conserve the energy during the impact , then I would suggest to compare the 2 scripts: What difference is in your model ? what initial condition are different ? , etc …

PS : I’m not sure of what i’m saying its just an humble opinion …

A Salute
Oscar G.