special_bonds for Tip4p water model

Dear LAMMPS

It might be a very naive question but I wonder whether the special bonds command is required to model liquid water Tip4p. I have this question because the benchmark code for SPC/E water models from LAMMPS has include this in the input file.

This is my input file:

imension 3
units real
atom_style full

read_data config-water.dat
#read_restart old_config_nvt.dat
replicate 1 1 1

group all type 1 2
group O type 1
group H type 2

neighbor 2.0 bin
neigh_modify delay 0 every 1 check yes one 4000
#neigh_modify every 1 delay 10 check yes

pair_style lj/cut/tip4p/long 1 2 1 1 0.15 8.5 8.5
kspace_style pppm/tip4p 1.0e-5

pair_modify shift yes
pair_modify tail no

pair_coeff 1 1 0.155 3.1536
pair_coeff 1 2 0.0000 0.0000
pair_coeff 2 2 0.0000 0.0000

bond_style harmonic
angle_style harmonic
dihedral_style none
improper_style none

bond_coeff 1 10000.00 0.9572
angle_coeff 1 10000.0 104.52

#special_bonds lj/coul 0.0 0.0 0.5

velocity all create 298 427 dist gaussian mom yes
restart 1000 old_config_1.dat old_config_2.dat

fix 1 all shake 0.00001 50 0 b 1 a 1

choose ensemble

fix 5 all npt temp 298.0 298.0 100 iso 1 1 50 drag 0.0
#fix 5 all nvt temp 298.0 298.0 100.0
#fix 5 all nve
#fix 7 all langevin 298.0 298.0 50 100 zero yes tally yes
#fix_modify 7 energy yes

compute 8 O pe/atom
compute 9 O reduce sum c_8
compute 10 H pe/atom
compute 11 H reduce sum c_10
compute 12 all pair lj/cut/tip4p/long
compute 13 all msd/molecule

compute 15 all rdf 100 1 1
fix 16 all ave/time 100 1 100 c_15 file O_O.rdf mode vector ave running overwrite
compute 17 all rdf 100 1 2
fix 18 all ave/time 100 1 100 c_17 file O_H.rdf mode vector ave running overwrite
compute 19 all rdf 100 2 2
fix 20 all ave/time 100 1 100 c_19 file H_H.rdf mode vector ave running overwrite

thermo 100
thermo_style custom step temp pe evdwl ecoul epair ebond eangle edihed eimp emol elong c_9 c_11 c_12 c_13[1][4] c_13[1][4] press vol etail xlo xhi
dump id all xyz 100 dump.xyz
dump_modify id element O H

timestep 2.0
run 7000000

Dear LAMMPS

It might be a very naive question but I wonder whether the special bonds
command is required to model liquid water Tip4p. I have this question

yes and no. you can easily answer your question yourself, by making a
few simple tests.

because the benchmark code for SPC/E water models from LAMMPS has include
this in the input file.

please note that the default setting in LAMMPS is:

special_bonds lj/coul 0.0 0.0 1.0

thus even if you don't explicitly issue a special_bonds command, it
will be used.

axel.

Dear LAMMPS

I thought the default setting of special bond as shown from the manual is:

Default:

All 3 Lennard-Jones and 3 Coulobmic weighting coefficients = 0.0, angle = no, dihedral = no, and extra = 0.

But as you suggested I have set the special_bonds to lj/coul 0.0 0.0 1.0 and rerun simulations and the log.lammps is exactly the same as the one that I have no special_bonds command. Hence should I always believe that 0.0 0.0 1.0 is the default setting?

Best Wishes
Lunna

Dear LAMMPS

I thought the default setting of special bond as shown from the manual is:

Default:

All 3 Lennard-Jones and 3 Coulobmic weighting coefficients = 0.0, angle =
no, dihedral = no, and extra = 0.

But as you suggested I have set the special_bonds to lj/coul 0.0 0.0 1.0 and
rerun simulations and the log.lammps is exactly the same as the one that I
have no special_bonds command. Hence should I always believe that 0.0 0.0
1.0 is the default setting?

what is written in the documentation takes precedence over what
somebody (e.g. me) claims in an e-mail without checking the docs and
what is written in the source code takes precedence over both (cf.
force.cpp).

as i already pointed out to you *twice* and as you have noticed
yourself. the setting of the 1-4 (aka dihedral) exclusions is of *no
consequence* for water models, since they DO NOT HAVE DIHEDRALS. full
stop.

your test is thus not proving anything about dihedral exclusions. you
need to have a dihedral for that.

axel.