USER-FEP package with LAMMPS


For tip4p pair styles you do need to give the site indices (I suppose these are your 2 3 4 0.125) because of the charge M on the O that does not sit on an atom but a short distance away (0.125 or 0.15 A according to the version of tip4p).
Besides, there seems to be no simple way at the moment to mix charmm style and tip4p water in LAMMPS.
I’d use a 3-site water model such as spce or tip3p. Here charges are located at the center of sites, so you do not need those indices nor the O-M distance. You can just treat the water sites as normal lj/charmm sites. This would be very similar to the example I sent you (I used spce).

One other thing: for charmm pair styles it is usually better to give inner and outer cutoffs which are different (say 8.0 and 10.0 A) otherwise you loose the interest of the charmm version.

So I’s say you would use, for lambda=1, alpha_LJ=0.5 and alpha_C=10.0:

pair_style hybrid &
lj/charmm/coul/long 8.0 10.0 &
lj/charmm/coul/long/soft 1.0 0.5 10.0 8.0 10.0

You can check the number and order of arguments to the pair_style command in the doc page for the soft-core lj pairs (in my repository).