Reactive Force Field-Based Molecular Dynamics Simulations on the Thermal Stability of Trimesic Acid on Graphene: Implications for the Design of Supramolecular Networks
Summary¶
Jacquelin et al. use ReaxFF to study how trimesic acid (TMA) self-assembles into several ordered 2D hydrogen-bonded motifs on graphene—honeycomb, filled honeycomb, flower, zigzag, and close-packed—and how those networks evolve with temperature. Simulations combine constant-NVT molecular dynamics and force-biased Monte Carlo (fbMC) up to 650 K, tracking hydrogen-bond counts, OH radial distribution functions, and carboxyl-group torsion as order–disorder indicators. The work argues that honeycomb networks are more thermally robust than the dense zigzag and close-packed arrangements, that extra TMA “guests” in honeycomb pores further stabilize the lattice through host–guest interactions, and that kinetics (not only static energetics) governs when high-coverage close-packed regions reorder toward honeycomb-like packing.
Methods¶
MD application (ReaxFF on graphene)¶
- Engine / code: ReaxFF-based molecular dynamics and force-biased Monte Carlo as described in ACS Appl. Nano Mater.; the article does not restate a specific MD engine name in the indexed front matter—N/A — package string (e.g. LAMMPS) if not explicit in the PDF/SI, confirm
pdf_path. - System size & composition: TMA monolayer polymorphs on a graphene support in the named motifs; exact supercell sizes, TMA coverages, and atom counts are given in the article/SI—N/A — not duplicated in this note.
- Boundaries / periodicity: In-plane periodic 2D supercells for the supported monolayer models (standard for graphene/TMA slabs); normal-direction padding as in the paper.
- Ensemble: NVT MD for the thermal-stability sweeps; fbMC used for the compact close-packed case (per the article’s protocol for that segment).
- Timestep & duration: Integration settings and run lengths for NVT MD and fbMC stages are in ACS Appl. Nano Mater.—N/A — not restated here; use
pdf_path. - Thermostat / barostat: NVT thermostating for MD as implemented in the published workflow; N/A — explicit thermostat family string if not on the first pages of the PDF. N/A — isotropic NPT barostat for the monolayer studies unless the paper lists bulk NPT equilibration segments.
- Temperature: Simulations up to 650 K (article abstract) to probe melting/reordering; intermediate setpoints and ramps are in the Methods.
- Pressure, electric field, shear/shock, enhanced sampling: N/A — no static electric field, shock loading, or umbrella/metadynamics in the reported workflow. Long-range and ReaxFF charge equilibration settings—N/A — in this short note; see
pdf_path.
Force-field training¶
N/A — this is an application study: the authors employ a published ReaxFF description for TMA, TMA–TMA, and TMA–graphene interactions and validate it against their QM reference comparators in the article (not a de novo open-parameter fit reported as the paper’s main contribution). Copy parameter lineage and any reweighting from pdf_path/SI for reproduction.
Static QM¶
The paper uses DFT (and related) reference data to benchmark ReaxFF against TMA–TMA and TMA–graphene interaction strengths for selected configurations; functional, basis, and k-mesh details are tabulated in the article—N/A — not copied into this summary.
Findings¶
- Stability ordering: By the melting temperatures and disorder metrics reported (H-bond counts, OH RDFs, carboxyl twisting), the honeycomb motif is more thermally stable than high-coverage zigzag and close-packed networks.
- Host–guest: Additional TMA molecules placed in honeycomb pores increase stability, consistent with host–guest stabilization beyond the empty honeycomb case.
- Kinetics vs static scores: The authors show that energetics alone do not explain all structural outcomes: close-packed regions can be kinetically destabilized even when static energy rankings might suggest otherwise, and MD captures rapid loss of order in some dense mesophases.
- Reordering: For close-packed islands, fbMC samples a transition toward a quasi-honeycomb arrangement, which the paper ties to stronger dimeric –COOH hydrogen bonding in honeycomb-like packing versus trimeric motifs in the compact structure.
- Disorder mechanism: Twisting/rotation of carboxyl groups with temperature disrupts hydrogen bonds and drives network “melting”; partial desorption at the onset of disorder is discussed in terms of intermolecular vibrational energy transfer.
- Corpus honesty: Melting temperatures, RDFs, and H-bond statistics should be quoted from the version-of-record PDF/SI; this page does not tabulate those numbers.
Limitations¶
ReaxFF remains a semiempirical bond-order model; long-timescale rare events and subtle electronic effects in π-conjugated organics may require higher-level validation. Guest loading and coverage choices in simulation cells may not span all experimental STM solution conditions.