Water interactions with nanoporous silica: Comparison of ReaxFF and ab initio based molecular dynamics simulations

Summary¶

Rimsza et al. compare ReaxFF and ab initio molecular dynamics (AIMD) for water interacting with nanoporous silica, focusing on how well classical reactive MD reproduces QM-level structure, reaction barriers, and transport trends in confined Si/O/H environments. Two Si/O/H ReaxFF parametrizations—associated with Yeon et al. and Fogarty et al. in the article’s framing—are evaluated against AIMD references on high-defect, strained silica motifs where water dissociation, hydroxylation, and small-ring defect chemistry are sensitive tests of the classical model. Adri C. T. van Duin is a coauthor, linking the paper to the group’s long-running silica–water ReaxFF development lineage and to geochemical/materials applications where nanopores alter H-bond networks and reaction kinetics.

Methods¶

1 — MD application (ReaxFF and AIMD). The study compares ReaxFF and DFT-based ab initio molecular dynamics (AIMD) for water in nanoporous silica models that expose high-defect, strained Si/O/H environments. Reactive classical MD uses two literature Si/O/H parameterizations—the Yeon et al. set (J. Phys. Chem. C 2016) and the Fogarty et al. set (J. Chem. Phys. 2010)—on the same classes of structures simulated at the AIMD level. Observables include local Si–O–Si and Si–OH speciation, water dissociation propensity, pathways and energetics for two-membered (2-ring) siloxane defect removal, hydroxylation kinetics, and H versus O diffusion under nanoconfinement. Engine, system sizes (atom counts), PBC, ensemble, timestep, thermostat/barostat, temperature and pressure, run lengths, electrostatics, and QEq usage are N/A — not transcribed on this wiki page; read pdf_path and any SI. Electric fields and enhanced sampling beyond conventional MD/AIMD are N/A — not highlighted in the short abstract-level material used for this note.

2 — Force-field training. N/A — the paper benchmarks published ReaxFF parameterizations rather than publishing a full new fit as its main contribution.

3 — Static QM / DFT. AIMD on comparable nanoporous silica motifs supplies the QM reference for structures, barriers where reported, and diffusion trends; detailed DFT functional, basis, and k-sampling choices are N/A — not duplicated here—see the article.

Both ReaxFF and AIMD use PBC supercells with explicit Si/O/H stoichiometry as in the published models. Molecular dynamics protocol tables in pdf_path should list the engine (e.g. LAMMPS for ReaxFF), supercell atom counts, NVT/NPT staging, timestep (fs), equilibration/production run lengths (ps/ns), thermostat and barostat/pressure control, and target temperature (K). Those numerical fields are N/A — not transcribed on this page unless they appear above. Shear, electric field driving, and umbrella/metadynamics/replica-exchange sampling are N/A — not highlighted in the abstract-level excerpt used here unless the full text introduces them.

Findings¶

Outcomes and mechanisms. Competing reaction pathways and intermediate geometries affect 2-ring defect removal and hydroxylation rates, so defect-rich nanoporous silica is a strong discriminator between ReaxFF parametrizations. Nanoconfinement lowers water diffusion relative to bulk-like water. Hydrogen diffuses roughly 10–30% faster than oxygen in the nanoconfined regime discussed in the abstract, consistent with H-hopping contributions emphasized in the article.

Comparisons. For the high-defect Si/O/H benchmark suite in the paper, the Yeon et al. ReaxFF parametrization tracks AIMD more closely than the Fogarty et al. set on the observables highlighted in the abstract (mechanisms, hydroxylation, defect populations, activation energies where quoted).

Sensitivity and limitations. Results are parametrization-dependent; AIMD itself depends on DFT approximations. Quantitative values should be taken from J. Phys. Chem. C figures and tables rather than this summary.

Limitations¶

Parametrization dependence implies residual uncertainty for complex amorphous silica and high-pH chemistry outside the training scope. AIMD references themselves depend on DFT choices; treat disagreements as triangulation opportunities rather than automatic ReaxFF failure.

This benchmark is frequently cited next to geochemistry theme pages; keep task:validation prominent in frontmatter so application papers do not drown out method comparisons in retrieval.

Relevance to group¶

Direct ReaxFF validation work for silica–water interfaces with van Duin authorship.

Citations and evidence anchors¶

DOI: https://doi.org/10.1021/acs.jpcc.6b07939 (papers/Rimsza_JPC_2016.pdf).

reaxff-family