hybrid pair styles for CHARMM and TIP4P ew

Dear Lammps Users,

I am trying to run a simulation composed of glucose and water molecules. I am using CHARMM force field parameters for the glucose molecules and the TIP4P ew water model. I seem to be having some difficultly getting the simulation to proceed. The difficultly appears to centre upon the use of TIP4P ew with CHARMM and the choice of pair styles. Currently my input seems to work for a minimization, but immediately fails when trying to begin a simulation following from this minimization.

I am using the version of LAMMPS from 5 Jun 2019.

Could anyone point me in the right direction as to how I can correctly combine the pair styles? Or suggest some where else that the issues may be stemming from?

Many thanks in advance,

James

My current input is:

variable xyz_dump_rate index 1000

atom_style full
bond_style harmonic
angle_style charmm
dihedral_style charmmfsw
improper_style harmonic
special_bonds charmm

boundary p p p
pair_style hybrid lj/charmmfsw/coul/long 10 12 lj/cut/tip4p/long 1 2 1 1 0.125 12 12
pair_modify mix arithmetic
kspace_style pppm/tip4p 1e-6
read_data glucose100_tip4p5000.lmp

pair_coeff 1 1 lj/cut/tip4p/long 0.16275000 2.51569076
pair_coeff 1 2 lj/cut/tip4p/long 0.00000000 1.77886200
pair_coeff 2 2 lj/cut/tip4p/long 0.00000000 0.00000000
pair_coeff 1 3 lj/charmmfsw/coul/long 0.07216647 2.59382823
pair_coeff 1 4 lj/charmmfsw/coul/long 0.07216647 2.59382823
pair_coeff 1 5 lj/charmmfsw/coul/long 0.07216647 2.59382823
pair_coeff 1 6 lj/charmmfsw/coul/long 0.09546727 2.59726064
pair_coeff 1 7 lj/charmmfsw/coul/long 0.12757351 2.47069135
pair_coeff 1 8 lj/charmmfsw/coul/long 0.17681707 2.51181657
pair_coeff 1 9 lj/charmmfsw/coul/long 0.08652456 1.88795220
pair_coeff 1 10 lj/charmmfsw/coul/long 0.08557891 2.35625944
pair_coeff 1 11 lj/charmmfsw/coul/long 0.07547351 2.35625944
pair_coeff 2 3 lj/charmmfsw/coul/long 0.00000000 1.88774863
pair_coeff 2 4 lj/charmmfsw/coul/long 0.00000000 1.88774863
pair_coeff 2 5 lj/charmmfsw/coul/long 0.00000000 1.88774863
pair_coeff 2 6 lj/charmmfsw/coul/long 0.00000000 1.89246211
pair_coeff 2 7 lj/charmmfsw/coul/long 0.00000000 1.71463284
pair_coeff 2 8 lj/charmmfsw/coul/long 0.00000000 1.77337883
pair_coeff 2 9 lj/charmmfsw/coul/long 0.00000000 0.63246622
pair_coeff 2 10 lj/charmmfsw/coul/long 0.00000000 1.54518884
pair_coeff 2 11 lj/charmmfsw/coul/long 0.00000000 1.54518884
pair_coeff 3 3 lj/charmmfsw/coul/long 0.03200000 2.66967971
pair_coeff 3 4 lj/charmmfsw/coul/long 0.03200000 2.66967971
pair_coeff 3 5 lj/charmmfsw/coul/long 0.03200000 2.66967971
pair_coeff 3 6 lj/charmmfsw/coul/long 0.04233202 2.67301472
pair_coeff 3 7 lj/charmmfsw/coul/long 0.05656854 2.55020796
pair_coeff 3 8 lj/charmmfsw/coul/long 0.07840408 2.59007092
pair_coeff 3 9 lj/charmmfsw/coul/long 0.03836665 1.99088131
pair_coeff 3 10 lj/charmmfsw/coul/long 0.03794733 2.43950885
pair_coeff 3 11 lj/charmmfsw/coul/long 0.03346640 2.43950885
pair_coeff 4 4 lj/charmmfsw/coul/long 0.03200000 2.66967971
pair_coeff 4 5 lj/charmmfsw/coul/long 0.03200000 2.66967971
pair_coeff 4 6 lj/charmmfsw/coul/long 0.04233202 2.67301472
pair_coeff 4 7 lj/charmmfsw/coul/long 0.05656854 2.55020796
pair_coeff 4 8 lj/charmmfsw/coul/long 0.07840408 2.59007092
pair_coeff 4 9 lj/charmmfsw/coul/long 0.03836665 1.99088131
pair_coeff 4 10 lj/charmmfsw/coul/long 0.03794733 2.43950885
pair_coeff 4 11 lj/charmmfsw/coul/long 0.03346640 2.43950885
pair_coeff 5 5 lj/charmmfsw/coul/long 0.03200000 2.66967971
pair_coeff 5 6 lj/charmmfsw/coul/long 0.04233202 2.67301472
pair_coeff 5 7 lj/charmmfsw/coul/long 0.05656854 2.55020796
pair_coeff 5 8 lj/charmmfsw/coul/long 0.07840408 2.59007092
pair_coeff 5 9 lj/charmmfsw/coul/long 0.03836665 1.99088131
pair_coeff 5 10 lj/charmmfsw/coul/long 0.03794733 2.43950885
pair_coeff 5 11 lj/charmmfsw/coul/long 0.03346640 2.43950885
pair_coeff 6 6 lj/charmmfsw/coul/long 0.05600000 2.67634559
pair_coeff 6 7 lj/charmmfsw/coul/long 0.07483315 2.55369901
pair_coeff 6 8 lj/charmmfsw/coul/long 0.10371885 2.59350830
pair_coeff 6 9 lj/charmmfsw/coul/long 0.05075431 1.99535119
pair_coeff 6 10 lj/charmmfsw/coul/long 0.05019960 2.44315808
pair_coeff 6 11 lj/charmmfsw/coul/long 0.04427189 2.44315808
pair_coeff 7 7 lj/charmmfsw/coul/long 0.10000000 2.42485701
pair_coeff 7 8 lj/charmmfsw/coul/long 0.13860014 2.46674649
pair_coeff 7 9 lj/charmmfsw/coul/long 0.06782330 1.82756102
pair_coeff 7 10 lj/charmmfsw/coul/long 0.06708204 2.30815388
pair_coeff 7 11 lj/charmmfsw/coul/long 0.05916080 2.30815388
pair_coeff 8 8 lj/charmmfsw/coul/long 0.19210000 2.50793639
pair_coeff 8 9 lj/charmmfsw/coul/long 0.09400319 1.88278676
pair_coeff 8 10 lj/charmmfsw/coul/long 0.09297580 2.35212267
pair_coeff 8 11 lj/charmmfsw/coul/long 0.08199695 2.35212267
pair_coeff 9 9 lj/charmmfsw/coul/long 0.04600000 0.89444231
pair_coeff 9 10 lj/charmmfsw/coul/long 0.04549725 1.66961735
pair_coeff 9 11 lj/charmmfsw/coul/long 0.04012481 1.66961735
pair_coeff 10 10 lj/charmmfsw/coul/long 0.04500000 2.18522702
pair_coeff 10 11 lj/charmmfsw/coul/long 0.03968627 2.18522702
pair_coeff 11 11 lj/charmmfsw/coul/long 0.03500000 2.18522702

neighbor 2 bin
neigh_modify delay 5 every 1

thermo_style custom etotal ke temp pe ebond eangle edihed eimp evdwl ecoul elong press vol etail density
thermo_modify line multi
thermo 10

dump 3 all xyz ${xyz_dump_rate} traj_sim_sug.xyz
write_data output.data

timestep 10.0
min_style fire
minimize 0.00000001 0.0000000001 15000 15000

reset_timestep 0
fix 1 all shake 1e-6 500 0 m 1.008 b 1 a 1 13
fix 2 all nvt temp 300.0 300.0 100.0
run 1000000

Unless stated otherwise above:
IBM United Kingdom Limited - Registered in England and Wales with number 741598.
Registered office: PO Box 41, North Harbour, Portsmouth, Hampshire PO6 3AU

sorry, but it won’t work correctly like this.
you must use a tip4p pair style for all coulomb interactions. otherwise, the coul/long interactions will not “see” the water oxygen charge at the M point location, but on the oxygen location.

axel.

Dear Axel,

Thanks you for your fast and helpful response.

To implement your suggestion, I can see that I can select tip4p/long as the pair style for the coulombic interactions, but I cannnot see a way to specify to use charmm LJ terms in conjuction. Am I missing something here or is this not a possible combination?

Many thanks in advance for any help.

James

Dear Axel,

Thanks you for your fast and helpful response.

To implement your suggestion, I can see that I can select tip4p/long as the pair style for the coulombic interactions, but I cannnot see a way to specify to use charmm LJ terms in conjuction. Am I missing something here or is this not a possible combination?

since CHARMM is parameterized for TIP3P (actually a modified TIP3P that was also known as TIPS2 at some time) and not TIP4P (or TIP4P-ew) it is not a big surprise, that there is no specific pair style that used the charmm switching function for lennard jones and coul/tip4p for long range electrostatics.
things get even more complicated when using charmm dihedrals due to the somewhat complex ways they are currently implemented in LAMMPS (for historical reasons).

so to enable using TIP4P style water in your setup, you would have to write a new pair style which merges the LJ and chamm parts of the coul/long style with the coulomb processing of lj/cut/coul/tip4p.

axel.

Hi Axel,

Thank you for your help once again and the pointer in how to solve my problem.

Many thanks,

James

Dr. James McDonagh
Research Staff Member
IBM Research UK
E-mail james.mcdonagh@…8604…
Mobile +44 7387 128078
Office +44 (0)1925 864242
IBM Profile
IBM Research@…8621…

Hi Axel,

Thank you for your help once again and the pointer in how to solve my problem.

if you are successful, please consider submitting the new pair style for inclusion into the LAMMPS distribution. this is best done via a pull request on https://github.com/lammps/lammps/pulls/

thanks,
axel.