Skip to content

Thermal Stability of Organic Monolayers Grafted to Si(111): Insights from ReaxFF Reactive Molecular Dynamics Simulations

Abstract

ReaxFF MD (ADF + LAMMPS, 0.1 fs, 1500–2000 K) on Si(111) slabs with alkyl (50% coverage) and methyl/propynyl/propenyl full-coverage monolayers reveals two-silicon dehydrogenation pathways and coverage-dependent alkene desorption.

Summary

ReaxFF parametrization for Si/C/H optimized to DFT training sets for alkyl-on-Si(111) chemistry (van Duin parameter lineage) targets pentacoordinate Si intermediates and homolytic Si–C events that dominate hot organosilicon decomposition. MD uses velocity Verlet, 0.1 fs timestep, temperature damping 100 fs, 800 ps runs from 1500–2000 K after 300 K, 1 ps thermalization (ADF2016 and LAMMPS). Si(111) slabs use six bilayers; 6×8 and 8√3×8 cells (48 or 64 top Si atoms) with 200 Å vacuum; both surfaces functionalized in many cases. Systems include 50% alkyl + 50% hydrogen at top sites (ethyl, propyl, pentyl, decyl) and 100% Si–CH₃, Si–CCCH₃, Si–CHCHCH₃. High temperatures accelerate decomposition within accessible MD timescales (authors justify 1500–2000 K window after ethyl scan 1200–2200 K), trading laboratory temperatures for activated chemistry within nanosecond trajectories.

Methods

1 — MD application (atomistic dynamics)

ReaxFF reactive molecular dynamics probes thermal decomposition of organic monolayers on Si(111) at elevated temperatures so that bond-breaking events occur within accessible nanosecond trajectories. Simulations use velocity Verlet integration with Δt = 0.1 fs, Berendsen thermostat coupling (100 fs damping constant in the article), 800 ps production segments after 1 ps thermalization from 300 K into the 1500–2000 K window used for the grafted layers (authors motivate the hot window with a prior ethyl-layer scan up to 2200 K). Calculations are reported from ADF2016 and cross-checked in LAMMPS as stated in the article.

  • Engine / code: ADF2016 and LAMMPS with ReaxFF (same DOI as 2017federico-a-soria-acs-thermal-stability; this corpus path may be a proof PDF).
  • System size & composition: Si(111) slabs with six bilayers, 6×8 and 8√3×8 surface unit cells (48 or 64 top-layer Si sites), ~200 Å vacuum gap, and both surfaces functionalized in the geometries discussed in the article. Monolayers include 50% alkyl + 50% H (ethyl, propyl, pentyl, decyl) and full-coverage Si–CH₃, Si–CCCH₃, and Si–CHCHCH₃ terminations.
  • Boundaries / periodicity: Three-dimensional periodic slab supercells with vacuum along the surface normal (geometry conventions in the article figures).
  • Ensemble: NVT hot-MD segments after initial thermalization (canonical sampling with Berendsen coupling as reported).
  • Timestep: 0.1 fs (required for stiff Si–C/H motion at 1500–2000 K in the authors’ protocol).
  • Duration / stages: 1 ps ramp/thermalization from 300 K, then 800 ps reactive sampling in the reported high-temperature window(s).
  • Thermostat: Berendsen temperature control with 100 fs damping constant (article text).
  • Barostat: N/A — constant-volume NVT hot runs without hydrostatic NPT control in the protocol summarized here.
  • Temperature: 300 K initialization, then 1500–2000 K production temperatures depending on layer chemistry (per article).
  • Pressure: N/A — no stress-control barostat in the summarized NVT decomposition protocol.
  • Electric field: N/A — not used.
  • Replica / enhanced sampling: N/A — direct hot MD rather than umbrella / metadynamics.

2 — Force-field training

Si/C/H ReaxFF parameters optimized against DFT training data for alkyl-grafted Si(111) chemistry (including pentacoordinate Si motifs and homolytic Si–C pathways) along the van Duin reactive silicon–hydrocarbon lineage cited in the article.

3 — Static QM / DFT-only

DFT reference energies/structures enter only as training data for ReaxFF; ab initio MD is not the production tool for the reported thermal decomposition trajectories.

Findings

  • Silyl radicals from Si–C homolysis drive dehydrogenation; main pathways require two lattice Si centers, not only one, so pairwise Si reactivity matters beyond local substituent chemistry.
  • Flexible n-alkyl chains dehydrogenate readily (β-H routes → 1-alkene desorption); lower coverage allows deeper methylene dehydrogenation including terminal methyl on long chains.
  • Rigid alkynyl/alkenyl substituents hinder terminal methyl abstraction → higher thermal stability versus saturated analogs at comparable coverages.
  • Surfaces trend toward hydrogen termination as Si–C bonds break and Si–H forms during decomposition, altering subsequent radical recycling.

Limitations

Proof PDF in corpus; elevated temperatures accelerate kinetics—extrapolate to lower T with care.

Relevance to group

Adri C. T. van Duin co-authorship; demonstrates ReaxFF for semiconductor organics degradation. The ACS Appl. Mater. Interfaces article is a useful reference for Si(111) graft thermolysis benchmarks when calibrating Si/C/H reactive models against DFT barriers and desorption product channels.

Citations and evidence anchors

  • DOI: 10.1021/acsami.7b05444