[lammps-users] NVT calculation for point dipoles


I wondered, if

# Dipole particles in a 3d box

units lj
atom_style dipole
dimension 3

lattice sc 0.095
region box block -15 15 -15 15 -15 15
create_box 1 box
create_atoms 1 box

mass 1 1.0
shape 1 1 1 1
dipole 1 1
set group all dipole/random 98934

velocity all create 0.1 87287 mom yes

pair_style dipole/cut 2.5
pair_coeff * * 1.0 1.0

neighbor 0.3 bin
neigh_modify delay 0

fix 1 all nvt/sphere temp 0.1 0.1 100

timestep 0.005

compute erot all erotate/sphere
thermo_style custom step temp epair c_erot etotal press
thermo 20

dump 1 all custom 1000 dump_T01.dipole id type x y z mux muy muz

run 500000

would treat the dipoles in the correct way, I mean nvt/sphere takes
rotational degrees of freedom into account, does it work for those point
dipoles? In the case of the nve ensembe also "update dipole" must be
included in the "fix" command, which does not work for nvt.

And do the commands:

velocity all create 0.1 87287 mom yes
# should the velocity temperature equal the desired one

fix 1 all nvt/sphere temp 0.1 0.1 100

#Nose/Hoover thermostat, should both temperature be equal (Tstart,Tend)

set the temperature to 0.1 in lj untis, or do I have to change s.th. else?

Gerald Rosenthal

I think an update dipole option needs to be
added to fix nvt/sphere to do what you want.

Re: setting temperature, the commands you
list should work (but again they do nothing
for the dipoles).