[lammps-users] 3D-IPS/DFFT

Hello,

I am using the 12 Dec 2018 version of lammps.

Is there 3D-IPS/DFFT command or a similar method implemented in lammps that can be used with the pairstyle charmmfsw. I need long range Lennard Jones and Coulombic interactions for a DPPC monolayer I am trying to model with CHARMM36.

3D-IPS/DFFT is explained in the attached link below.

https://aip.scitation.org/doi/full/10.1063/1.2992601

Philip Skeps

Undergraduate

2nd Year

(612)-346-3663

Hello,

I am using the 12 Dec 2018 version of lammps.

Is there 3D-IPS/DFFT command or a similar method implemented in lammps that can be used with the pairstyle charmmfsw. I need long range Lennard Jones and Coulombic interactions for a DPPC monolayer I am trying to model with CHARMM36.

Please note that using the CHARMM switching function for Lennard-Jones in combination with a long-range solver for Lennard-Jones makes no sense. After all you are trying to compute the r^6 term exactly and a switching function would be switching off the long-range contributions.

What is available in LAMMPS is listed in the manual: https://docs.lammps.org/Commands_all.html
Anything else would need to be implemented (as c++ code).

Axel.

Please also note that if you plan to use the long-range Lennard-Jones support that exists in LAMMPS, you should be updating your LAMMPS version. There were several issues detected and fixed with the corresponding pair and kspace styles when we added unit testing last summer.

Axel.

Hey Axel,

Sorry about not responding to the mailing list, is Ccing the mailing list considered responding to the mailing list;

Thank You for the quick response. I assume I should use just the standard charm pairstyle then with long range LJ solvers, or is this still wrong. I am fairly new to Molecular Dynamic simulations and any help is appreciated as I would love to get better at this sorta stuff.

Philip Skeps

Undergraduate

2nd Year

(612)-346-3663

Hey Axel,

Sorry about not responding to the mailing list, is Ccing the mailing list considered responding to the mailing list;

yes. the point it that the conclusion of any discussion remains visible and will be stored in the mailing list archives and thus can be searched for by people running into similar issues and thus reduces the need to answer the same questions over and over again (although some people don’t take as much advantage of that as they should).

Thank You for the quick response. I assume I should use just the standard charm pairstyle then with long range LJ solvers, or is this still wrong. I am fairly new to Molecular Dynamic simulations and any help is appreciated as I would love to get better at this sorta stuff.

the charmm pair styles all have a switching function for lj (as that is what CHARMM does). the difference between lj/charmm and lj/charmmfsw is the formulation of the switching function. the former implements what was used in CHARMM19(?) and the latter what is used currently. It should be explained in the manual of those pair styles.

If you employ a long-range solver with dispersion support you may only use a compatible pair style.
What is the best choice is impossible to say from remote and not trivial to answer anyways. CHARMM parameters are created under the assumption of using switching functions. How this affects accuracy when changing to a long-range treatment is difficult to say in general. Overall, this is a discussion about the research and independent from LAMMPS, so you need to have this discussion with your adviser/supervisor or whoever recommended using long-range dispersion solver.

Axel.