Skip to content

ReaxFF molecular dynamics simulation of intermolecular structure formation in acetic acid-water mixtures at elevated temperatures and pressures

Evidence and attribution

Authority of statements

Prose below summarizes the JCP article identified by title and pdf_path.

Summary

ReaxFF reactive MD in ADF is used to study structure and hydrogen-bonding in acetic acid–water mixtures from ambient conditions to supercritical states (temperature–pressure ranges in the article, e.g. up to roughly 600–1200 K and 5–150 MPa as stated in the introduction). A combined ReaxFF (water potential plus biomolecular acetic acid subset, with O–C–O valence conjugation adjusted) is checked against metadynamics on acetic acid dissociation so that the acid dissociation free-energy difference yields a realistic \(K_a\). Canonical MD with 0.25 fs timesteps and Berendsen thermalization analyzes RDFs, hydrogen-bond accounting, and cluster statistics versus composition \(x_{\mathrm{HAc}}\).

Methods

  • Force-field merge and \(K_a\) calibration: Water ReaxFF from prior work; acetic acid subset from biomolecular O/H/C parameters; valence conjugation term adjusted (reported change from −19.93 to −5.0 for the optimized parameter) so metadynamics gives two minima in the \(S_{\mathrm{OH}}\) collective variable with \(\Delta G\) translating to \(K_a \approx 4.88\), near experiment (4.76). Metadynamics uses Gaussian hill deposition (height/width and deposition interval in the paper), NPT equilibration (1 ns, 300 K, 0.1 MPa) of one acetic acid in 60 waters in a 12.26 Å cube.
  • Bulk MD: 1000 molecules total, varying acetic acid mole fraction \(1.0 \ge x_{\mathrm{HAc}} \ge 0.2\); densities from 1.0 and 1.05 g/cm\(^3\) for water and acid components; canonical ensemble; 0.25 fs timestep; Berendsen thermostat (0.1 ps damping); heating in 100 K increments with 50 ps equilibration and extended runs (≥0.5 ns at 300 K, ≥1 ns at final \(T\)) until homogeneous density.
  • Analysis: RDFs for OO, carbonyl–hydroxyl oxygen, methyl–carbonyl carbon, etc.; hydrogen-bond criteria using donor/acceptor oxygen distances from RDF minima and angular cutoffs (relaxed in supercritical conditions); averages over the last 0.25 ns with 500 samples.

Bulk production MD (consolidated). ADF ReaxFF reactive MD on 1000 molecules in periodic cubic supercells with PBC; NVT Berendsen control (0.1 ps damping), 0.25 fs timestep, staged 100 K heating with 50 ps equilibration per step and ≥0.5 ns (300 K) / ≥1 ns (target T) sampling; supercritical grids per article. Metadynamics calibration uses NPT (1 ns at 300 K, 0.1 MPa) before bulk NVT sweeps. Barostat for bulk sweeps: N/A — see JCP Methods for each state point. Electric field: N/A — not used. Enhanced sampling: metadynamics for \(K_a\) only; N/A — umbrella / replica exchange otherwise.

Findings

At ambient \(T\), P = 0.1 MPa): for acid-rich compositions, cyclic dimers and chain-like assemblies appear (RDF peaks ~2.7 Å and longer-range peaks ~5–9 Å); increasing water disrupts dimers/chains, and released acid molecules show ~4 Å correlations consistent with dipole–dipole pairing. Water–water RDFs match experimental-like shell structure at high water content; distorted shells appear as acetic acid dominates. In the supercritical regime, long-range structural order (e.g. second/fourth neighbor features) weakens; water–water RDFs remain consistent with literature, while first-neighbor acid–water correlations persist at lower supercritical \(T\) but fade** at higher \(T\) within the studied range.

Comparisons, sensitivity, and limitations. The metadynamics calibration explicitly targets agreement with experimental \(K_a\) for acetic acid dissociation, anchoring the merged ReaxFF before composition/temperature/pressure sweeps. Sensitivity to \(x_{\mathrm{HAc}}\) shows systematic changes in hydrogen-bond motifs and RDF shell structure, with additional broadening of hydrogen-bond definitions in supercritical regimes as described in the article. Limitations include the usual ReaxFF approximations to quantum hydrogen-bonding effects. Corpus honesty: cite pdf_path for exact state points and any SI figures beyond this summary.

Limitations

Reactive MD with empirical parameters may not capture all quantum effects in hydrogen bonding; supercritical analysis relies on broader HB geometric cutoffs.

Reader notes (navigation)