I want to simulate ion-pairing between cations and anions (not single atom ions like Li+ and Cl-, but multiple atom ions like NH4+ and BF4-) in different nonaqueous solvents. Most of the huge molecular dynamics packages available there boast of forcefield parameters for proteins and lipids of hundreds of atoms but simple multi-atom ions are routinely missing from these software/force field packages. Is it possible to develop force field parameters systematically for any one of these open source packages and implement and test them over parallel machines ?
I am baffled by the number of software choices available and the number of force field packages, but at the same time, they have the data for the exact same molecules/atoms/species and seem equally useless for my case. Asking for the parameter files directly has proven to be futile. Also, I notice persistent discouragement of any effort to build these by self and a routine answer ends like "Its going to be extremely difficult". I would like to understand why is is this so difficult to do as a procedure. I totally understand that validation of these is another game altogether, but the development process should be fairly standard or ?
On another note, if i try to develop or use non-polarizable FF parameters like OPLS-AA for my problem, will they be any good to calculate properties like diffusivity/mobility/solvent-shell configuration/etc? Or would these things be better revealed if I use CPMD like techniques? Has the CPMD method evolved enough to give me these answers ?
Any links, pages, methods, procedural guidance is gratefully appreciated.
It's true that most FF focus on biological systems. The reason is very simple: funding $$$ for biolgical sciences (particularly for disease-oriented research) >>>>>>> funding $$$ for chemistry.
As to the main question: it can be definitely done and it will be extremely arduous, particularly the validation part (though perhaps not particularly difficult). But a quick scan of web reveals that there are numerous works on ionic liquids MD (doi: 10.1021/jp047619y, 10.1016/j.fluid.2010.02.034, 10.1039/C3CP53492H and many more) - perhaps these would be a good starting point.
It is indeed possible and some attempts have been already presented in the literature. In this respect, I would suggest to use Symmetry-Adapted Perturbation Theory as reference for parameterization:
http://pubs.acs.org/doi/abs/10.1021/ct5010593
The AMOEBA polarizable force field can be used (as implemented in the TINKER program package):
http://pubs.acs.org/doi/abs/10.1021/jp910674d
OpenMM:
https://simtk.org/home/openmm-amoeba
Some examples:
http://pubs.acs.org/doi/abs/10.1021/jp403578r
http://pubs.acs.org/doi/abs/10.1021/jp503347f
http://pubs.acs.org/doi/abs/10.1021/jp302548u
We just published a review article on our own Drude force field development, including a how-to for our parametrization process. There are also some references to Drude models of ionic liquids in it that may be of interest; polarization response is very important in such systems.
http://pubs.acs.org/doi/abs/10.1021/acs.chemrev.5b00505
The Drude model can be used with CHARMM, NAMD, OpenMM, and GROMACS. We do our development within CHARMM because it is very simple to interface the QM target data with the facilities in CHARMM for charge and polarizability fitting, vibrational analysis, etc.
There is no need to 'develop': look at
Molecular Simulation, Vol. 32, No. 6, 15 May 2006, 471–484
Thank you everyone, for the answers, I am already on the leads but it will take some time to go before I can report if i made progress or not.
Best
Abhishek
- Badges
- Science topic
- Similar topics
- Physical Sciences
- Atomic, Molecular and Optical Physics
- Atoms