There are a myriad of pair potentials available for simulating the interactions between water molecules, ranging from simpler models like the SPC/E to more sophisticated ones based on high-level quantum mechanical methods like coupled-cluster calculations.

I'm interested in identifying what is currently considered the "gold standard" ab initio pair potential for the interaction between two water molecules. My primary focus is on quantum mechanical accuracy, rather than the potential's ability to reproduce bulk behavior. Ideally, I'm seeking a highly accurate model that has been validated against quantum mechanical calculations, similar to the one presented in the paper An accurate analytic representation of the water pair potential by Cencek et Al

Additionally, if there is existing code that can reproduce the results of the potential, that would be highly beneficial. Your expert guidance is greatly appreciated, as I have limited experience in this specific area of computational chemistry.


1 Answer 1


I'm not an expert, but I believe Paesani's many-body MBPol force fields Acc. Chem. Res. 49, 1844 (2016) afford some of the most accurate descriptions of the interactions between water molecules. The MBPol models are fit to highly accurate coupled-cluster energies with singles, doubles and perturbative triples (CCSD(T)) at the complete basis set (CBS) limit. I have not followed the development closely, but the many-body decomposition is a good Ansatz for building interaction models.

I am a bit baffled, though, by your question which asks for a "pair potential". Pair potentials are typically based only on the distance, while the interaction of water molecules is definitely not of the pair form, as it is strongly affected also by the relative angle of the molecules. The molecules are polar and only form hydrogen bonds at certain magic angles, and this requires more complicated interactions than pair potentials. Also, from what I remember, the higher-than-two-body terms in MBPol are likewise important; the behavior of three water molecules is more complicated than what one would expect based on just a two-molecule model applied to three molecules.

  • $\begingroup$ Thank you for your answer. After reviewing the article you cited, I'll likely accept it. I believe "pair potential" correctly describes the interaction of two isolated molecules that only depend on their relative positions and orientations. I believe that for more than two molecules, "potential energy surface" (PES) is the correct term. In my case, I'm focused on estimating a third cross virial coefficient between two water molecules and a third particle. To get an exact expression, I'd need to find the non-additive parts of the PES, but I only want an estimate. $\endgroup$ Sep 14 at 7:42
  • $\begingroup$ Note: In your article the author has a reference to an excellent review of the different multibody potential energy functions. $\endgroup$ Sep 14 at 10:18

You must log in to answer this question.

Not the answer you're looking for? Browse other questions tagged .