[lammps-users] Temperature selection and control for ellipse simulation

Dear all,

In ellipse bulks with GB potential two temperatures were gathered. One is from compute temp/asphere command for all ellipses. The other is from thermo_style custom temp for whole system. As output shows the former is smaller than the latter about 0.1 reduced temperature. From the manual I found that the “compute temp/asphere” includes more freedom degree.What I want to get is the phase graph of GB bulks. Which temperature is the temperature for phase graph?

Also with the running goes on, the system temperature will increase continuously sometime break the system. It seems that the temperature control is out of works.

Some parts of my input file for lammpse is bellow:

GayBerne ellipsoids bulks

units lj
atom_style ellipsoid
dimension 3
boundary p p p

lattice sc 0.01

region AllCell block -5 5 -5 5 -2 2

create_box 1 AllCell
create_atoms 1 box
group AllCell type 1

set group AllCell quat 0 0 1 0 # type, name, arg, a b c

set group AllCell quat/random 18238

mass 1 1.0
velocity AllCell create 1.9 6668 loop local # group action temperature seed node

GB(sigma_ee/sigma_ss,epsilon_ss/epsilon_ee,upsilon,mu) (4.4,20,1,1) epsilon0=2 cutoff 5.5

shape 1 1 1 4.4
pair_style gayberne 1.0 1.0 1.0 5.5
pair_coeff 1 1 2 1 1 1 0.05 1 1 0.05 5.5

neighbor 0.3 bin

compute a_temp AllCell temp/asphere

thermo_style custom step temp press pe ke vol lx ly lz pxx pyy pzz pxy pxz pyz c_a_temp

timestep 0.002

restart 50000 bulk.restart

thermo 300 # output step
dump 1 AllCell custom 300 dump.GB_bulk tag type x y z quatw quati quatj quatk

fix npt AllCell npt/asphere 1.9 1.9 100 xyz 2 2 1000 #ID group.ID npt/asphere Tstart Tstop Tdamp p.style args keyword value
run 100000

Both temperatures should be similar in value, but you probably want to use the asphere temperature for your case because that includes the rotational kinetic energy and that is what is used by npt/asphere. Try decreasing your damp parameters for your temperature problem. Remember that these are time and not number of timesteps.

  • Mike