A Molecular Density Functional Theory of aqueous electrolytic solution

Abstract: We propose a generalisation of molecular density functional theory to describe inhomogeneous solvent mixture, with the objective of modelling electrolytic solutions. Two electrolytic models are presented, both within the HNC approximation. The first one is a two-components mixture representing a primitive-like model of sodium chloride, where the solvent is described as a dielectric continuum. This popular model has the advantage of simplicity, as the ions densities solely depend on spatial coordinates. Additionally, we develop a realistic three-components electrolyte model, in which water solvent is described by a third density field that depends on both spatial and orientational coordinates. The proposed methodology and its tridimensional implementation (3 spatial coordinates and 3 Euler angles) are validated by comparing the solvation properties of a sodium cation with the predictions of integral equation theory solved in 1D (1 intermolecular distance and 5 Euler angles), showing near-perfect agreement. This methodology enables the study of solvation properties of solutes of arbitrary shapes in electrolytic solutions, as demonstrated with the prototypical N-methylacetamide molecule immersed in both electrolytic solution models.

• The paper presents a novel extension of molecular density functional theory to handle multi-component aqueous electrolyte mixtures with both spatial and orientational resolution.
• It employs advanced numerical schemes using FFTs and generalized spherical harmonics for efficient variational minimization, achieving excellent agreement with 1D integral equation theory.
• The explicit water model accurately captures ion solvation structure and hydrogen-bonding effects, paving the way for applications in biophysics, nanoconfinement, and electrochemical engineering.

Molecular Density Functional Theory for Aqueous Electrolytes: Formalism, Validation, and Implications

Introduction and Context

Electrolytic solutions, crucial in biological and electrochemical systems, present persistent challenges for theoretical modeling due to their inhomogeneity and the interplay between ionic correlations and solvent structure. Traditional explicit-solvent simulations such as MD or MC, while accurate, are prohibitively expensive for complex or large-scale systems. Field-theoretic approaches, especially continuum Poisson-Boltzmann (PB) frameworks, fail to capture finite-size effects, ion-specific interactions, and solvent structure.

Integral equation theory (IET) and reference interaction site models (RISM) offer an improved microscopic description, yet the complexity scales steeply with molecular detail and dimensionality, particularly for molecular solvents like water in the presence of three-dimensional solutes. While advances such as 3D-RISM and the inclusion of various closure relations improve practical utility, site-based decompositions can be physically uncontrolled.

Classical density functional theory (cDFT) and, more recently, molecular DFT (MDFT) provide a formally rigorous alternative, recasting the problem as the minimization of a free energy functional over spatial (and, for molecular solvents, orientational) densities. The presented work extends MDFT beyond single-component solvents to arbitrary mixtures of rigid molecular species, specifically targeting aqueous electrolyte solutions. This allows the simultaneous treatment of spatial inhomogeneity, realistic solvent structure, and ionic correlations within a variationally-controlled framework.

Theoretical Framework

The formalism generalizes MDFT to $N$-component mixtures of rigid molecular species, exploiting the grand canonical ensemble. Each species is characterized by a density $\rho_A(\mathbf{r},\Omega)$ depending on spatial coordinate $\mathbf{r}$ and orientation $\Omega$; monoatomic ions have no orientational degrees of freedom. The grand potential functional $F$ is universally decomposed into the sum of ideal, external, and excess parts:

$$F[\{\rho_A\}] = F_{\mathrm{id}}[\{\rho_A\}] + F_{\mathrm{ext}}[\{\rho_A\}] + F_{\mathrm{exc}}[\{\rho_A\}]$$

The ideal and external parts encode entropy and the solute-solvent potential, respectively; the excess term captures correlations and is expressed via a second-order truncation (hypernetted-chain, HNC, approximation) involving direct correlation functions (dcf) $c_{AB}$. Higher-order "bridge" terms are omitted here, expecting that they may need future development specifically for mixtures.

Key computational advances include:

• Efficient variational minimization schemes leveraging FFTs for spatial convolutions and generalized spherical harmonics (GSH) projections for orientational convolutions, following the machinery developed for molecular IET.
• Careful enforcement of electroneutrality in periodic simulation cells, crucial due to the divergence of Coulomb interactions in reciprocal space.
• Systematic validation against 1D-IET solutions exploiting spherical and orientational symmetry for for "test particle insertion" scenarios.

Model Systems

Two hierarchically constructed electrolyte models are introduced and contrasted:

1. Primitive-Like Model: Na$^+$ and Cl$^-$ ions as Lennard-Jones spheres in a dielectric continuum (water as an implicit, uniform dielectric). Computationally efficient, spatial densities only, but neglects solvent-specific structuring and explicit hydrogen-bond networks.
2. Explicit-Water Three-Component Model: Na$^+$, Cl$^-$, and water (SPC/E model) treated as three components, with water described by a full position- and orientation-dependent density. Six dcf projections between components (including angularly-dependent water-water and water-ion terms) are required. This model accounts for solvent reorganization, hydrogen bonding, and local dielectric inhomogeneity.

Numerical Results and Validation

Primitive Model

Test-particle (ion in electrolyte) solvation structure and free energy computed via both 3D-MDFT and 1D-IET display excellent agreement: solvation free energies for Na$^+$ and Cl$^-$ differ by less than $0.01$ kJ/mol. For 3D solutes, such as N-methylacetamide (NMA), spatial ion exclusion zones and the first solvation shell are resolved. Inclusion of the NMA solute with neutralized charges confirms that finite-size (Lennard-Jones) effects dominate structure, with electrostatic effects providing secondary modulation. The model correctly predicts cavity sizes and local charge distributions, though fails to reproduce the expected water-induced structuring and underestimates the role of specific ion-solute interactions due to mean-field screening.

Explicit Water Model

With water treated explicitly, the theory achieves near-perfect agreement with 1D-IET for Na$^+$ solvation free energies and radial distributions. Now, multi-shell solvation structure manifests for both ions and water, with strong water structuring controlling ionic arrangement. In the NMA case, orientational order of water and oscillatory ion distributions are observed, consistent with hydrogen-bonding and solvent polarization. The charge density profile and water polarization (as computed from the orientational distribution) show non-trivial, solvent-mediated solute-ion coupling and alternating charge layering not accessible to the primitive model.

The results demonstrate the importance of explicit solvent degrees of freedom for accurately capturing the interplay between steric, electrostatic, and hydrogen-bonding interactions, and validate the implementation's ability to treat arbitrarily-shaped solutes in complex electrolytes.

Implications and Future Perspectives

The extension of MDFT to multi-component, fully molecular mixtures opens the way for efficient, variationally-controlled studies of ion–solvent–solute interplay in biology (e.g., ion-mediated protein stability, macromolecular recognition), nanoconfined systems (electric double layers in capacitors), and heterogeneous catalysis. The explicit inclusion of molecular solvent structure overcomes limitations of dielectric continuum or primitive models, and enables full coupling between molecular shape, solvent orientation, and local electrostatics.

From a theoretical perspective:

• The formalism is extensible to mixtures with any number of components, as well as to other molecular solvents and non-aqueous systems.
• The HNC-level description is variationally well-defined, but further systematic inclusion of three-body/bridge terms (currently underdeveloped for mixtures) will improve prediction of thermodynamics and correlation-driven phenomena near criticality or in strongly-coupled regimes.
• The methodology provides a bridge between site-based IET/RISM (with uncontrolled approximations on the dcf decomposition) and fully molecular field-theoretic DFT.

From a computational standpoint:

• The framework provides a tractable route to rapidly screen solvation and ion-specific effects around large, flexible, or low-symmetry solutes, with much lower cost than explicit MD.
• Direct incorporation of high-level quantum mechanical or machine-learned force fields is possible via the external field term, suggesting future application to ab initio embedding and reactive systems.

Potential future developments are anticipated in:

• Incorporation of advanced bridge functionals for multi-component electrolytes, building on existing developments for pure solvents;
• Coupling with quantum mechanical descriptions for polarizable or electronically active solutes;
• Design and interpretation of experiments requiring spatially-resolved ion and solvent structure, including time-resolved spectroscopies and force measurements.

Conclusion

This work provides a formally exact and computationally validated extension of molecular density functional theory to inhomogeneous aqueous electrolytes, overcoming the limitations of both site-based IET and continuum models. The variational, orientation-resolved molecular framework enables accurate computation of solvation structure and energetics for arbitrary 3D solutes in multicomponent electrolytes. The results establish a platform for predictive, physically transparent modeling of complex systems where ions and molecular solvents mediate structure and function, with broad implications for physical chemistry, biophysics, and electrochemical engineering (2511.09346).