Development of a ReaxFF reactive force field for the Pt–Ni alloy catalyst
Summary¶
This paper develops a ReaxFF description of Pt/Ni/C/H/O aimed at heterogeneous catalysis, fitting to DFT data on Pt–Ni alloy equations of state, surface energies of Pt\(_x\)Ni\(_{1-x}\) facets, and adsorption of O, H, C, hydrocarbon fragments, CO, OH, and H\(_2\)O. GCMC and MD on slabs and nanoparticles probe Pt vs Ni segregation under vacuum, H\(_2\), and O\(_2\), matching trends for Pt-rich compositions while noting weaker behavior for Ni-rich alloys. ExxonMobil coauthorship reflects industrial catalysis context.
Methods¶
1 — MD application (GCMC and classical MD). After parametrization, grand-canonical Monte Carlo (GCMC) and molecular dynamics (MD) probe surface segregation for a Pt\(_3\)Ni slab and a truncated cuboctahedral nanoparticle exposing (111) and (100) facets under vacuum, H\(_2\), and O\(_2\), with PBC as in the published models. Timestep, thermostat/barostat choices, temperature and pressure setpoints, run lengths, and electrostatic details are N/A — not restated in the indexed excerpt; use pdf_path. Electric fields and enhanced sampling beyond GCMC/MD are N/A — not indicated in that excerpt.
2 — Force-field training. The authors develop Pt/Ni/C/H/O ReaxFF parameters for heterogeneous catalysis on Pt–Ni alloys using DFT reference data for equations of state of Pt\(_3\)Ni, PtNi, and PtNi\(_3\), Pt\(_x\)Ni\(_{1-x}\)(111) surface energies with \(x = 0.67\)–\(0.83\), and adsorption energies of O, H, C, CH, CH\(_2\), CH\(_3\), CO, OH, and H\(_2\)O on Pt\(_8\)Ni\(_4\), Pt\(_9\)Ni\(_3\), and Pt\(_{10}\)Ni\(_2\) reference cells, emphasizing reproduction of composition-dependent binding-strength orderings tabulated in the paper. DFT program, functional, basis, and k-mesh settings for the training set are N/A — not duplicated here.
3 — Static QM / DFT. DFT supplies equations of state, surface energies, and adsorption energies that define the ReaxFF training targets. Full DFT software, functional, basis, and k-mesh tables are in pdf_path rather than duplicated here.
GCMC/MD segregation studies use PBC slabs and a truncated cuboctahedral nanoparticle with explicit H\(_2\) or O\(_2\) as in J. Phys. Chem. A. Applied electric fields and umbrella, metadynamics, or replica-exchange sampling are N/A — not indicated in the abstract-level excerpt. Timestep, equilibration and production lengths, and thermostat or barostat parameters are N/A — not transcribed here.
Findings¶
Segregation trends (vacuum). The fitted field reports Pt surface segregation for (111) at \(x \geq 0.67\) and for (100) at \(x \geq 0.62\) in vacuum in the authors’ analysis.
Adsorbate-driven segregation. H and CH\(_3\) are associated with driving Pt to the surface, whereas oxidation intermediates/products such as C, O, OH, H\(_2\)O, CO, CH, and CH\(_2\) tend to draw Ni to the surface. Under O\(_2\), Ni segregates to the surface of a model PtNi nanoparticle, whereas under H\(_2\) or vacuum, Pt segregation is favored—consistent with the differential affinities discussed in the article.
Transferability limits. The authors state the parametrization performs reasonably for Pt:Ni ratios greater than ~0.6 but has difficulty with Ni-rich compositions (\(x < 0.6\)), motivating additional training for Ni-rich alloys.
Limitations¶
- Bimetallic reactive spaces are large; remaining errors map to specific composition/adsorbate corners.
Relevance to group¶
Core ReaxFF parametrization publication from van Duin’s group on alloy catalyst modeling.
Citations and evidence anchors¶
- DOI:
https://doi.org/10.1021/acs.jpca.6b06770(papers/Shin_PtNiCHO_JPCA_2016.pdf).