Hydration free energies of monovalent ions in transferable intermolecular potential four point fluctuating charge water: An assessment of simulation methodology and force field performance and transferability

Abstract

Hydration free energies of nonpolarizable monovalent atomic ions in transferable intermolecular potential four point fluctuating charge (TIP4P-FQ) are computed using several commonly employed ion-water force fields including two complete model sets recently developed for use with the simple water model with four sites and Drude polarizability and TIP4P water models. A simulation methodology is presented which incorporates a number of finite-system free energy corrections within the context of constant pressure molecular dynamics simulations employing the Ewald method and periodic boundary conditions. The agreement of the computed free energies and solvation structures with previously reported results for these models in finite droplet systems indicates good transferability of ion force fields from these water models to TIP4Q-FQ even when ion polarizability is neglected. To assess the performance of the ion models in TIP4P-FQ, we compare with consensus values for single-ion hydration free energies arising from recently improved cluster-pair estimates and a reevaluation of commonly cited, experimentally derived single-ion hydration free energies; we couple the observed consistency of these energies with a justification of the cluster-pair approximation in assigning single-ion hydration free energies to advocate the use of these consensus energies as a benchmark set in the parametrization of future ion force fields.

Pairwise-additive force fields for selected aqueous monovalent ions from adaptive force matching

“…where q I is the charge of the ion and surface potential φ surf was chosen to be −0.527 V. 59 Although this term depends on the choice of φ surf , it only influences the single ion solvation free energies and cancels for the salt.…”

Single-ion hydration thermodynamics from clusters to bulk solutions: Recent insights from molecular modeling

“…As a result, the values for monovalent anions are more positive by about 5-6 kJ/ mol and values for divalent cations are more negative by about the same amount. This deviation was noticed in earlier publications [26,51,90], but the tabulated quantities are frequently used uncorrected. We consider these deviations a result of numerical errors (transferred from earlier publications) rather than different conventions, as sometimes suggested [26,51].…”

“…Sometimes the term 'absolute' is used to characterize hydration properties, typically invoked as a synonymous for 'intrinsic' [20,21], but it is also used to describe the results of cluster experiments [7,22,23], whose outcome has been extensively considered to measure 'real' properties [21,[24][25][26], which naturally include the cluster surface potential [20]. To avoid confusion, here we will avoid the term 'absolute' in favor of 'real' and 'intrinsic'.…”

“…As these terms suggest, the intrinsic properties are not real in the sense that they cannot be realized in any physical process because ions have to cross the interface when transferred from gas to the aqueous phase. However, a bulk solution without an interface can be achieved in computer simulations in periodic boundary conditions involving the Ewald summation (or for that matter, any other lattice summation) treatment of long-range interactions [26][27][28]. It follows that to determine real thermodynamics of ion hydration from computer simulations, it is critical to understand the planar vapor-liquid interface of pure water.…”

“…The original and most cited method relies on the socalled cluster-pair-based approximation (CPA) [7], which has also been considered to provide the most accurate proton hydration free energies and enthalpies [12,24,26,[93][94][95]. According to the CPA, the estimate for increasingly larger and infinite clusters is achieved by scaling the first term in Eq.…”