Skip to content

Development of a Reactive Force Field for Simulations on the Catalytic Conversion of C/H/O Molecules on Cu-Metal and Cu-Oxide Surfaces and Application to Cu/CuO-Based Chemical Looping

Evidence and attribution

Authority of statements

Prose below summarizes the J. Phys. Chem. C article identified by doi, title, and pdf_path. Quantitative convergence settings and extended tables appear in the published PDF and Supporting Information.

Summary

Zhu et al. develop an integrated Cu/C/H/O Reaxff for heterogeneous catalysis on Cu and Cu oxide via: (1) re-optimizing Cu on an expanded cluster training set; (2) merging with a C/H/O field; (3) fitting Cu–(C/H/O) cross-terms to DFT binding and elementary barriers. They introduce transition-state search and path tools on the Reaxff surface. LAMMPS reactive MD demonstrates H transfer and H₂ / CHO-class chemistry on Cu and chemical-looping Cu↔CuO cases with glucose and hydrocarbon oxidation. Adri C. T. van Duin is a senior author.

Methods

1 — MD application. LAMMPS reactive MD on Cu(100), (111), (211) and related cells (Section 3; SI for supercell sizes). NVT with Berendsen thermostat (0.1 ps damping as on [[2020wenbo-zhu-j-phys-chem-jp0c02573-2]]). Temperature setpoints include 600 K, 1000 K, 1400 K, 1600 K, and other values in 600–1600 K depending on case; 0.25 fs time step where stated; e.g. 400 ps gas-phase sampling for glucose + Cu at 1600 K. N/ANPT production stated in the short summary—confirm PDF; N/Ametadynamics; N/Aelectric field. PBC slab models as in Figures 8–14. Barostat N/A for NVT bracket as written on the companion slug.

2 — Force-field training. Parent Reaxff Cu subset + C/H/O library; reoptimize Cu; merge; fit cross-terms to DFT data on adsorption and barriers (Section 2). QM reference and training targets as in Section 3 below. Optimization by Reaxff least-squares-style updates in the Reaxff optimizer as described; TS-aware sampling using custom Reaxff path tools.

3 — Static QM (DFT reference). DMol3; GGA rPBE; ECP; unrestricted spin; 4×4×1 k-mesh (periodic); 0.006 Ha smearing; global 4.5 Å cutoff; SCF/geometry thresholds in Section 2.1. Binding energies via the adsorbate/surface partitioning; TS via LST/QST/CG as cited. N/Ahybrid functionals as mainline in the excerpt curated here.

4 — Review or non-simulation. N/A

Findings

Outcomes and mechanisms. The merged field reproduces DFT trends for adsorption and elementary steps in the validation set. MD shows H shuttling and H₂ / CHO-type events on Cu facets supporting network chemistry at interfaces. CLC case studies differentiate fuels by detailed reaction pathways on Cu/CuO as the abstract claims.

Comparisons and sensitivity. DFT vs Reaxff; T-dependent MD cases (600–1600 K range on illustrative systems).

Authored limitations and outlook. Real catalysts with promoters, alloys, and coking are not fully captured; see ## Limitations.

Corpus honesty. Error tables in JPCC/pdf_path.

Limitations

Industrial catalysts include promoters, alloys, and long-timescale coking not represented in idealized surface cells. ReaxFF accuracy remains tied to the training scope; extrapolation to new chemistries requires validation.

Relevance to group

Extends the group’s Reaxff portfolio into Cu catalysis and chemical looping, combining parameterization methodology with application trajectories.

Citations and evidence anchors

  • papers/Zhu_JPCC_CuCHO_2020.pdf; https://doi.org/10.1021/acs.jpcc.0c02573
  • Extract alignment: normalized/extracts/2020wenbo-zhu-j-phys-chem-jp0c02573_p1-2.txt