Hydration Free Energies of Linear Alkanes: Systematic Deviations in Common Water Models and Their Correction

Referee: [Reparameterization procedure (description of cavity free energies and well-depth adjustment)] The reparameterization procedure computes cavity free energies using the original TraPPE-UA/water Lennard-Jones parameters and then solves for a new well depth ε that reproduces experimental ΔG_hyd at 300 K. This treats the cavity term as fixed and independent of ε. Because the cavity contribution arises from growing a repulsive core whose effective size and packing depend on the full LJ interaction (including the attractive well), altering ε changes the equilibrium density and the free-energy cost of cavity creation. The original cavity value is therefore no longer the correct baseline once ε is updated. This assumption is load-bearing for the proposed correction and requires either an iterative recalculation or a quantitative estimate of the induced change in the cavity term.

Authors: We agree that a fully self-consistent treatment would require iterating the cavity calculation after updating ε, since the attractive well modestly influences solvent packing around the solute. In the original procedure the cavity term was evaluated with attractions disabled, which is a standard decomposition, but the referee correctly notes that the new ε alters the reference state. To address this, we will add an iterative recalculation in the revised manuscript: after determining the initial ε from the original cavity value, we recompute the cavity free energy with the updated parameters and refine ε until convergence. Preliminary checks indicate that the adjustment to ε is small (∼5–8 %) and that the final hydration free energies change by less than 0.3 kJ mol⁻¹, preserving the reported agreement with experiment. We will document the iteration count, the magnitude of the correction, and the resulting ε values. revision: yes

Referee: [Results section on temperature dependence] The improvement across 290–350 K is presented as a key result, yet the manuscript does not report root-mean-square deviations or mean absolute errors for the original versus reparameterized models at each temperature. Without these quantitative metrics, it is difficult to assess whether the temperature dependence is genuinely captured or whether the improvement is largely confined to the 300 K fitting point.

Authors: We concur that explicit error metrics at each temperature would strengthen the claim of transferability. In the revised manuscript we will add a table (or supplementary table) that reports both the root-mean-square deviation and the mean absolute error between simulated and experimental hydration free energies for the original and reparameterized models at every temperature from 290 K to 350 K. These statistics will be computed separately for the three-site and four-site water models and will include the full set of alkanes studied. revision: yes