[Paper Review] Pressure Correction in Classical Density Functional Theory: Hyper Netted Chain and Hard Sphere Bridge Functionals

Volodymyr P. Sergiievskyi, Guillaume Jeanmairet+2 more|arXiv (Cornell University)|Sep 4, 2015

Phase Equilibria and Thermodynamics27 references3 citations

TL;DR

This paper proposes an a posteriori pressure correction for classical density functional theory (DFT) to improve solvation free energy (SFE) predictions, showing that errors in SFE calculations stem from incorrect system pressure in HNC and hard-sphere bridge functionals. The correction—based on the compressibility route pressure and an additional ρkTΔV term—significantly enhances accuracy, especially when combined with HNC/PC or HS-Bridge functionals, achieving SFE errors comparable to molecular dynamics simulations for 122 organic molecules.

ABSTRACT

Low accuracy of the Solvation Free Energy (SFE) calculation is a known problem of the numerical methods of the Integral Equation Theory of Liquids and the Classical Density Functional Theory (Classical DFT). Although functionals with empirical corrections can essentially improve the predictability of the methods, their universality is still a question. In our recent paper we connected the SFE calculation errors with the incorrect pressure in the Classical DFT and proposed the a posteriory correction to improve the results (J. Phys. Chem. Lett., 5, 1925-1942 ). This paper raised a discussion in the community. In particular, recently appeared a critical reply where pointed some thermodynamical inconsistencies of the derivations in our paper (J. Chem. Theory Comput., 11, 378-380). In the present work we re-derive the pressure correction in a more simple way and show that despite the inaccuracies during the derivation, the final form of the previously derived correction is correct. We also test the applicability of the proposed correction to the functionals which include a three- and many- body terms from the fundamental measure theory (FMT) for hard sphere fluid. We test all the functionals on a set of model systems and discuss the obtained results.

Motivation & Objective

• Address the long-standing issue of inaccurate solvation free energy (SFE) predictions in classical DFT, particularly in HNC and hard-sphere bridge functionals.
• Investigate the root cause of SFE errors, identifying incorrect system pressure as a primary driver, especially in HNC where pressure reaches 11.5 kBar.
• Develop and validate a thermodynamically consistent a posteriori pressure correction to improve SFE predictions without altering the underlying functional form.
• Assess the trade-off between thermodynamic consistency (correct pressure) and accuracy in SFE for small solutes when parameterizing hard-sphere diameter.
• Propose a universal ρkTΔV correction term to further reduce SFE errors, especially for large solutes with significant partial molar volume.

Proposed method

• Re-derive the pressure correction in the HNC approximation using a simpler, more transparent approach to validate the correctness of the prior pressure correction formula.
• Apply the pressure correction to HNC and hard-sphere bridge (HS-Bridge) functionals, using the compressibility route to compute system pressure.
• Parameterize the HS-Bridge functional to either reproduce correct pressure or correct SFE for methane, exploring the trade-off between thermodynamic consistency and predictive accuracy.
• Introduce an additional ρkTΔV correction term based on the dependence of SFE errors on partial molar volume, derived from systematic analysis of errors.
• Test the corrected functionals (HNC/PC+, HS/PRES+) on a benchmark set of 122 organic molecules to evaluate predictive accuracy against MD simulations.
• Use radial distribution functions and solvent structure analysis to assess the quality of solvation shell predictions beyond SFE.

Experimental results

Research questions

• RQ1 Why do HNC-based functionals produce highly inaccurate solvation free energies despite qualitatively correct solvent structure predictions?
• RQ2 To what extent does the incorrect system pressure in HNC and HS-Bridge functionals contribute to SFE errors, particularly for solutes with large partial molar volumes?
• RQ3 Can a universal, a posteriori pressure correction be derived that improves SFE accuracy without compromising thermodynamic consistency?
• RQ4 How does the choice of hard-sphere diameter affect both the system pressure and SFE predictions, and can a single parameterization achieve both correct pressure and accurate SFE for small solutes?
• RQ5 Can the systematic error in SFE predictions be corrected by adding a ρkTΔV term, and does this correction generalize across diverse solute types?

Key findings

• The HNC approximation produces a system pressure of 11.5 kBar, which is the primary cause of the overestimated solvation free energy, as the −PΔV term dominates the error.
• The a posteriori pressure correction (HNC/PC) successfully restores the correct qualitative dependence of SFE on solute size, particularly for large hard spheres, where HNC incorrectly predicts a volume-dependent trend.
• Parameterizing the HS-Bridge functional to reproduce correct pressure leads to accurate SFE and solvent structure for large solutes, but results in significant underestimation of SFE for small solutes like methane.
• A hybrid functional combining the HS-Bridge and pressure correction (HS/PRES+) fails to improve SFE for small solutes, indicating a fundamental incompatibility between pressure consistency and small-solute accuracy in current formulations.
• The additional ρkTΔV correction term effectively reduces SFE errors across a wide range of solutes, especially those with large partial molar volumes, and when applied to HNC/PC+, reduces errors to levels comparable to molecular dynamics simulations.
• For 122 organic molecules, the HNC/PC+ functional with the ρkTΔV correction achieves the best overall SFE accuracy, suggesting this correction is a robust and generalizable improvement for classical DFT solvation free energy calculations.