Logo
ScienceStack
Table of Contents
Fetching ...
Sign In

Hydration Free Energies of Linear Alkanes: Evaluating and Correcting Classical Force Field Predictions with Different Water Models

Yalda Ramezani, Sumit Sharma

TL;DR

Problem: nonpolar solutes like linear alkanes exhibit hydration free energies ($ΔG_{hyd}$) that are overpredicted by common force fields, amplifying the hydrophobic effect. Approach: the authors systematically evaluate TraPPE-UA with four water models (SPC/E, OPC3, TIP4P/2005, OPC), compute $ΔG_{hyd}$ via free energy perturbation with a soft-core LJ and calibrate the alkane–water parameter $ε$ using cavity-formation free energies, while also assessing the impact of shifted LJ potentials. Findings: all water models overestimate $ΔG_{hyd}$; an approximate +5% increase in $ε$ relative to Lorentz–Berthelot brings predictions in line with experiments; HH-alkane with TIP4P/2005 and GAFF with TIP4P/2005 show good agreement, whereas LJ shifting worsens accuracy. Significance: calibrating mixture interaction parameters through cavity energetics improves hydration energetics predictions for nonpolar solutes and reveals limitations of standard mixing rules in hydrophobic contexts.

Abstract

Common force fields tend to overestimate the hydration free energies of hydrophobic solutes, leading to an exaggerated hydrophobic effect. We compute the hydration free energies of linear alkanes from methane to eicosane (C$_{20}$H$_{42}$) using free energy perturbation with various three-site (SPC/E, OPC3) and four-site (TIP4P/2005, OPC) water models and the TraPPE-UA force field for alkanes. All water models overestimate the hydration free energies, though the four-site models perform better than the three-site ones. By utilizing cavity free energies, we reparameterize the alkane--water well depth to bring simulation results in agreement with experimental and group-contribution estimates. We find that the General Amber Force Field (GAFF) combined with TIP4P/2005 water provides closer estimates of the hydration free energy. The HH alkane model (a reparameterized TraPPE-UA force field) with TIP4P/2005 reproduces experimental hydration free energies. We also show that applying a shifted Lennard--Jones potential leads to systematic deviations in the hydration free energy estimates.

Hydration Free Energies of Linear Alkanes: Evaluating and Correcting Classical Force Field Predictions with Different Water Models

TL;DR

Problem: nonpolar solutes like linear alkanes exhibit hydration free energies () that are overpredicted by common force fields, amplifying the hydrophobic effect. Approach: the authors systematically evaluate TraPPE-UA with four water models (SPC/E, OPC3, TIP4P/2005, OPC), compute via free energy perturbation with a soft-core LJ and calibrate the alkane–water parameter using cavity-formation free energies, while also assessing the impact of shifted LJ potentials. Findings: all water models overestimate ; an approximate +5% increase in relative to Lorentz–Berthelot brings predictions in line with experiments; HH-alkane with TIP4P/2005 and GAFF with TIP4P/2005 show good agreement, whereas LJ shifting worsens accuracy. Significance: calibrating mixture interaction parameters through cavity energetics improves hydration energetics predictions for nonpolar solutes and reveals limitations of standard mixing rules in hydrophobic contexts.

Abstract

Common force fields tend to overestimate the hydration free energies of hydrophobic solutes, leading to an exaggerated hydrophobic effect. We compute the hydration free energies of linear alkanes from methane to eicosane (CH) using free energy perturbation with various three-site (SPC/E, OPC3) and four-site (TIP4P/2005, OPC) water models and the TraPPE-UA force field for alkanes. All water models overestimate the hydration free energies, though the four-site models perform better than the three-site ones. By utilizing cavity free energies, we reparameterize the alkane--water well depth to bring simulation results in agreement with experimental and group-contribution estimates. We find that the General Amber Force Field (GAFF) combined with TIP4P/2005 water provides closer estimates of the hydration free energy. The HH alkane model (a reparameterized TraPPE-UA force field) with TIP4P/2005 reproduces experimental hydration free energies. We also show that applying a shifted Lennard--Jones potential leads to systematic deviations in the hydration free energy estimates.

Paper Structure

This paper contains 7 sections, 7 equations, 5 figures, 3 tables.

Table of Contents

  1. Introduction
  2. Simulation System and Methods
  3. Results and Discussion
  4. Conclusions
  5. Conflicts of Interest
  6. Data availability
  7. Supporting Information

Figures (5)

  • Figure 1: (a) Hydration free energies, $\Delta G_{hyd}$ of alkanes using the TraPPE-UA model and different water models. All models show positive deviation from the experiments/group contribution. The 3-point water models show larger deviation from the experimental/group contribution values compared to the 4-point water models. (b) $\Delta G_{hyd}$ of alkanes using the TraPPE-UA model but with the alkane-water well depth parameter updated via the method discussed in the text. An excellent match with the experimental/group-contribution values is found, except for some deviations observed with the OPC3 model for large alkanes.
  • Figure 2: Hydration free energy of alkanes calculated using the HH-alkane model is compared with the TraPPE-UA force field and the experimental/group-contribution values. TIP4P/2005 water model is employed. The HH-alkane model matches the experimental/group contribution values well up to $C_{20}$.
  • Figure 3: Hydration free energies of alkanes calculated using the General Amber Force Field (GAFF) all-atom model of alkanes and two different water models: SPC/E and TIP4P/2005. The results are compared with the TraPPE-UA model and experimental/group contribution values. Overall, results using GAFF are closer to the experimental/group contribution values compared to TraPPE. GAFF, SPC/E overestimates the $\Delta G_{hyd}$ values. GAFF, TIP4P/2005 overestimates the $\Delta G_{hyd}$ for small alkanes ($\leq C_5$) but underestimates the $\Delta G_{hyd}$ for large alkanes. Lines are guide to the eyes.
  • Figure 4: Hydration free energies, $\Delta G_{hyd}$ of alkanes for the unshifted HH-alkane (HHalkane) compared to those obtained for the shifted HH-alkane (HHalkane (shifted)); shifted HH-alkane with the potential shift corrected using equation \ref{['eq:energy_shift_approx']} and tail corrections added (equation \ref{['eq:long_range']}) (HHalkane (shift adj., tail)); and shifted potential with only tail corrections added (HHalkane (shifted, only tail)).
  • Figure 5: Carbon-oxygen radial distribution function for four different alkanes. The radial distribution functions of $C_{10}$, $C_{14}$, and $C_{20}$ are shifted vertically by 0.1 from the previous graph for ease of visualization.