Lennard-Jones 2D pressure changes with potential cut off

Dear All,

Version of LAMMPS: 5 May 2012

I ran a 2D simulation of Lennard-Jones particles at various temperatures
and densities using an NVT ensemble and noted that the pressure (after
including long-range correction) increases with increasing cut off length
used for the potential.

pair potential used: lj/charmm/coul/charmm
Charge on particles: 0

The system contains 2500 particles and was run with periodic boundary
conditions. All data stated below are in Lennard-Jones units. I used the
expression given below for long-range correction.

P_correction=12/5(pi)*(rho)^2*[(1/r_i)^10-5/4*(1/r_i)^4]

rho=density, T=temperature, P=pressure, delta P=uncertainty in pressure,
pi=3.14, r_i=inner cut-off radius used for potential, r_o=outer cut-off
radius used for potential.

T rho r_i r_o P delta P
0.5 0.75 3.0 3.1 0.2936 0.0014
    3.5 3.6 0.3065 0.0016
    4.0 4.1 0.3216 0.0007
    4.5 4.6 0.3230 0.0024

0.7 0.75 3.0 3.1 1.1724 0.0026
    3.5 3.6 1.1874 0.0028
    4.0 4.1 1.2020 0.0031
    4.5 4.6 1.2037 0.0023

1.0 0.75 3.0 3.1 2.3817 0.0037
    3.5 3.6 2.3951 0.0019
    4.0 4.1 2.4063 0.0037
    4.5 4.6 2.4116 0.0030

1.0 0.50 3.0 3.1 0.5510 0.0010
    3.5 3.6 0.5589 0.0008
    4.0 4.1 0.5648 0.0008
    4.5 4.6 0.5658 0.0004

1.4 0.10 3.0 3.1 0.1386 0.0001
    3.5 3.6 0.1390 0.0001
    4.0 4.1 0.1391 0.0001
    4.5 4.6 0.1392 0.0001

Regards,
Sindhana Selvi P.S.
Masters student (Physics)
Indian Institute of Science

Dear All,

Version of LAMMPS: 5 May 2012

Update.

Dear All,

Version of LAMMPS: 5 May 2012

I ran a 2D simulation of Lennard-Jones particles at various temperatures
and densities using an NVT ensemble and noted that the pressure (after
including long-range correction) increases with increasing cut off length
used for the potential.

pair potential used: lj/charmm/coul/charmm
Charge on particles: 0

The system contains 2500 particles and was run with periodic boundary
conditions. All data stated below are in Lennard-Jones units. I used the
expression given below for long-range correction.

P_correction=12/5(pi)*(rho)^2*[(1/r_i)^10-5/4*(1/r_i)^4]

​is this correction specifically for the CHARMM variant of the
lennard-jones interaction which uses a smooth truncation of the potential?
or for plain cutoff?
please also note, that if you want to *exactly* reproduce CHARMM, you need
to use a slightly different switching function, i.e. the one implemented in
the lj/charmmfsw/coul/*
the one in lj/charmm/coul/* corresponds to an extremely old variant of
CHARMM.

axel.

Dear Prof. Axel Kohlmeyer,

The pressure correction used is for the plain cutoff and does not take
into account the smooth truncation of the potential.

Regards,
Sindhana Selvi P.S.
Masters student (Physics)
Indian Institute of Science