Molecules to study

[justinGilmer]

A table of molecules we plan to include in this study, as well as conditions. Edit as needed.

NOTE: This does not currently include any challenge molecules we discussed before.

Molecules	Forcefield	Conditions	Ensembles	Reasoning
United-Atom Methane	TraPPE-UA	3 (T, rho), 3 (T,P)	NVT, NpT, GEMC-NVT	LJ only
United-Atom Ethane	TraPPE-UA	1 (T, rho), 1 (T,P)	NVT, NpT	Adds bonds
United-Atom Propane	TraPPE-UA	1 (T, rho), 1 (T,P)	NVT, NpT	Adds angles
United-Atom Butane	TraPPE-UA	1 (T, rho), 1 (T,P)	NVT, NpT GEMC-NVT	Adds dihedrals
United-Atom Hexane	TraPPE-UA	3 (T, rho), 3 (T,P)	NVT, NpT	Adds dihedrals and angles
SPC/E Water	SPC/E	3 (T, rho), 3(T, P)	NVT, NpT	Electrostatics
Atomistic Ethanol	OPLS-AA	3 (T, rho), 3 (T,P)	NVT, NpT	Adds dihedrals

[bc118]

In GOMC, the bonds need to be constrained for the Trappe-UA FF (per Trappe-UA FF requirements). Water bonds and angles need to be constrained in GOMC (per water FF requirements). Note: GOMC is not capable of sampling bonds.

In order to minimize the simulations, getting the most benefit for each simulation being run, and reduce the overall project time:

• We should consider first running Methane and Hexane, and if we get matching results then the rest of the simulations (Ethane, Propane, Butane) should all be OK since hexane has all the bonds, angles, and dihedrals. I'm not sure the Ethane, Propane, Butane simulations add a lot of value, unless we are having differences in the simulations or errors, and they need to be run specifically for troubleshooting. If troubleshooting, all the groups don't need to run the C2-C4 molecules if like only 1 group is showing differences.

• We may want to run only NPT simulations, instead of running NVT (except if we are doing diffusion for MD, etc.). I would think that if NPT simulations work then the NVT simulation should follow, provided we are analyzing the same values. This would reduce the simulation by half.

[ramanishsingh]

Regarding Brad's first point:

If C1 and C6 results match, we can skip C2-C4 and instead simulate a couple of more complicated molecules. If C1 and C6 results do not match, then we can design the troubleshooting process, like one he has mentioned in his comment.

[tcmoore3]

Engines listed here: https://docs.google.com/spreadsheets/d/13Hx9UnN9bzmnUhwWLLfUcGq68s1I8VH1tprwh0r98NY/edit#gid=2017555854

[jennyfothergill]

This information seems outdated. I'd like to move the molecules to a reference card in the Project board and close this issue. @bc118 @ramanishsingh if there is still an issue with the topic you mention, would you mind opening a new issue?

[bc118]

I dont think this is an issue, as we made a decision. No need to reopen. I will close