LAMMPS WWW Site - LAMMPS Documentation - LAMMPS Mailing List Archives
Re: [lammps-users] Free energy calculation using TI method

Re: [lammps-users] Free energy calculation using TI method

 From: Stefan Paquay Date: Mon, 4 Dec 2017 09:12:03 -0500

Your lammps version is very old, consider updating if possible.

With respect to TI, if you do not get out what you expect, either what you are calculating is not what you think it is, or you are not sampling your systems properly. Your system seems very complex though, maybe you should try a simpler setup just  to get the hang of doing TI with LAMMPS?

On Dec 4, 2017 7:16 AM, "Manish Maurya" <manishmm@...33....71...> wrote:
Dear LAMMPS users,

I am trying to calculate free energy using thermodynamic integration algorithm implemented in LAMMPS. I have got the results for free energy but I am not getting correct trend in free energy with respect to coupling parameter in fact I am getting opposite results. I mean to say, at low value of lambda (0.01) magnitude of free energy should be less as compared to high value of lambda (1). So, I want to know where am I making mistakes? I am writing my input script and calculated free energy data.

#LAMMPS (11 JAN 2012)

dimension   3

units       real

boundary    p p p

atom_style   full

pair_style     lj/cut 15.0

set type 1 charge 0.0

set type 2 charge 0.0

pair_coeff      3 3  0.0556417148  3.4

pair_coeff      1 1  0.05366216    2.80

pair_coeff      2 2  0.15701151    3.05

pair_coeff      1 3  0.090333873   3.395

pair_coeff      2 3  0.093461816   3.225

bond_style       harmonic

bond_coeff     1  1000.00        1.16

angle_style      harmonic

angle_coeff    1  1000.00        180.0

group flakes      type 3

group so2         type 1 2

neighbor      3.0 bin

neigh_modify     delay 1 every 1 check yes

neigh_modify    exclude molecule flakes

variable        a loop  100

label         loop

variable       scale1  equal  ((0.0)+(\$a*0.01))

fix               5  flakes setforce 0.0 0.0 0.0

fix             6 so2 nvt temp 303.0 303.0 100.0

variable        lj equal  v_scale1

variable        dlj equal 1

fix             eps1 all adapt 100 pair lj/cut epsilon 1 3 v_scale1

fix             eps2 all adapt 100 pair lj/cut epsilon 2 3 v_scale1

compute   ThermoInt  all ti lj/cut 1*3 v_lj v_dlj

fix    12 all ave/time 100 50 100000 c_ThermoInt

timestep      1.0

thermo        100000

#thermo_style      custom step temp press vol pe ke etotal

thermo_style    custom step v_scale1  c_ThermoInt

thermo_modify   flush yes

restart          100000 wtr.restart

dump     10 all custom 10000 wtr.lammpstrj id mol type x y z

dump_modify 10 sort id

run        100000

unfix           5

unfix           6

unfix eps1

unfix eps2

uncompute   ThermoInt

undump          10

next           a

jump           in.dat loop

lambda               ThermoInt

 0.01 -259.563 0.109 -47.0543 0.208 -26.4778 0.307 -14.8638 0.406 -14.2288 0.505 -8.47696 0.604 -11.9544 0.703 -7.54119 0.802 -8.8976 0.901 -8.22874 1 -7.67377

Regards,

Manish Maurya
Research Scholar
Computational Nano-science Group
Department of Chemical Engineering
Indian Institute of Technology, Kanpur
Kanpur, UP, India -208016

------------------------------------------------------------------------------
Check out the vibrant tech community on one of the world's most