[lammps-users] compute fep pair style param not supported for buck/coul/long

Hi all,

I am running an thermodynamic integration calculation using ‘fix adapt/fep’ and ‘compute fep’
The calculation is to integrate atom type 1 into the system, the system has other two atom types.

My input related to the calculation:

pair_style hybrid buck/coul/long 15.0 15.0
variable lambda equal ramp(0.0,1.0)
variable q1 equal -1.13*v_lambda
variable a2 equal 96716.76488*v_lambda
variable a3 equal 1308861.891*v_lambda

fix ADAPT all adapt/fep 100000 &
  pair buck/coul/long a 1 2 v_a2 &
  pair buck/coul/long a 1 3 v_a3 &
  atom charge 1 v_q1 &
  after yes
variable dlambda equal 0.002
variable dq1 equal -1.13*v_dlambda
variable da2 equal 96716.76488*v_dlambda
variable da3 equal 1308861.891*v_dlambda

compute FEP all fep ${TK} &
  pair buck/coul/long a 1 2 v_da2 &
  pair buck/coul/long a 1 3 v_da3 &
atom charge 1 v_dq1

fix FEP all ave/time 20 4000 100000 c_FEP[1] c_FEP[2] file fdti01.lmp

dump TRAJ all custom 20000 dump.lammpstrj id mol type element x y z ix iy iz

run 2000000

The setting file:
    pair_coeff 1 1 buck/coul/long 24347.94944 0.45150245 744.4847164
    pair_coeff 1 2 buck/coul/long 96716.76488 0.361526225 0.0
    pair_coeff 1 3 buck/coul/long 1308861.891 0.291380225 0.0
    pair_coeff 2 2 buck/coul/long 384185.643 0.27155 0.0
    pair_coeff 2 3 buck/coul/long 5199160.121 0.201404 0.0
    pair_coeff 3 3 buck/coul/long 70359906.63 0.131258 0.0

Then I got the error message :
ERROR: compute fep pair style param not supported (src/USER-FEP/compute_fep.cpp:223)
Last command: run 2000000

I don’t understand why this is happening as ‘buck' style is supported according to the the documentation.
Please help me.
Thanks in advance.

Hi all,

[…]

I don’t understand why this is happening as ‘buck’ style is supported according to the the documentation.

pair style buck != pair style buck/coul/long

you could try with

pair_style hybrid/overlay buck 15.0 coul/long 15.0

and then replace in the pair_coeff statements buck/coul/long with buck and add pair_coeff * * coul/long

which does use pair style buck.

axel.