Skip to content

Improvement of the ReaxFF Description for Functionalized Hydrocarbon/Water Weak Interactions in the Condensed Phase

Summary

The work introduces CHON-2017_weak, a ReaxFF reparameterization built on the protein-2013 parameter set to fix condensed-phase densities and weak C/H/O/N interactions for functionalized hydrocarbons with water. Validation spans pure liquids and mixtures (including hexane–water demixing, ethanol and tetramethylammonium (TMA) in water), plus AEM-relevant simulations of TMA structure and base-induced degradation in alkaline solution, where an alternate degradation pathway is proposed relative to prior QM pictures. The J. Phys. Chem. B article is positioned as a practical fix for aqueous organic density and demixing errors that otherwise propagate into membrane degradation studies.

Methods

1 — MD application. Engine / code: ReaxFF molecular dynamics is carried out with LAMMPS (cited in §2.2 together with the standalone ReaxFF code for development tasks). System size & composition: validation cells include, e.g., 500-molecule liquid water boxes, 200-molecule n-alkane periodic samples, hexane–water mixtures, ethanol and TMA in water, and crystal amino acid cells for selected properties; atom counts and box preparation (low-density start, compression, NPT density equilibration) follow §2.2. Boundaries / periodicity: 3D PBC throughout. Ensemble: NPT MD at room temperature (plus 100 K NPT for amino acid crystals to limit thermal disorder) to equilibrate liquid density—500 ps total with density averaged over the last 400 ps after box compression from 0.1 g/cm³; separate 500 ps NVT MD trajectories provide RDF/MSD for liquid water and TMA/OH⁻/water mixtures. Timestep: 0.25 fs for all ReaxFF MD segments reported in §2.2. Thermostat / barostat: Berendsen thermostat with temperature damping 100 fs for NVT runs and Berendsen barostat with pressure damping 2500 fs for NPT runs. Duration / stages: 500 ps per production block for the liquid/mixture protocols above after NPT equilibration. Temperature: ~298 K “room temperature” liquid runs; 100 K NPT for crystal examples; small-molecule reaction mechanism comparisons at 0.1 K in the ReaxFF driver to match 0 K QM references. Pressure: NPT hydrostatic control only where NPT is used; NVT blocks are constant volume. Electric field: N/A — not used in the validation suite summarized in §2.2. Enhanced sampling: N/A — not reported (standard NVT/NPT MD). Shear / shock / AIMD: N/A — not part of this parameterization study.

2 — Force-field training. Parent FF / elements: protein-2013 (aqueous branch) with C/H/O/N van der Waals and hydrogen-bond-related terms reoptimized to form CHON-2017_weak (§2.1). QM reference: DFT (and related QM) data enter the training and reaction-scan comparisons as cited in the article; full levels and data sets are tabulated in the main text and SI. Training set / targets: liquid densities, RDFs, hexane–water demixing, TMA structural and degradation chemistry, and small-molecule reaction benchmarks. Optimization / software: ReaxFF parameter optimization (successive / least-squares style updates) as described in §2.1, implemented with the standalone ReaxFF code for development and LAMMPS for large MD validation. Reference data / validation: experimental densities and RDF/MSD benchmarks for water and organics, plus QM pathways for the TMA discussion.

Findings

  • CHON-2017_weak reproduces experimental density trends for the tested pure and mixed CHON systems substantially better than protein-2013 for the weak-interaction issues highlighted in the abstract.
  • The force field captures macrophase separation for hexane + water and sensible solvation behavior for ethanol and TMA in water (abstract).
  • For TMA degradation in base, simulations motivate an additional pathway that is more favorable energetically than the dominant QM-predicted route in the prior literature (abstract—see article for structures and energetics).

The discussion stresses that weak-interaction fixes should be regression-tested on reactive benchmarks relevant to each new application because van der Waals reoptimization can perturb barriers outside the training window.

Limitations

Weak-interaction fits can trade accuracy against reactive channels not included in the training set; confirm any new chemistry against QM or experiment before high-stakes predictions. Parameter files and full validation tables are detailed in the article and Supporting Information.

Wiki prose here is a navigation aid. Definitive numbers, protocol details, and figure-level claims should be taken from the peer-reviewed article at pdf_path (and any Supporting Information cited there), not from this page alone.

Relevance to group

Defines CHON-2017_weak, a widely reused aqueous-organic ReaxFF branch for membrane and solution chemistry in the group.

Citations and evidence anchors

  • DOI: 10.1021/acs.jpcb.8b01127