Skip to content

Thermal transport across SiC–water interfaces

Summary

Nonequilibrium classical molecular dynamics computes thermal boundary conductance (TBC) across 3C–SiC | water interfaces, scanning crystallographic orientation, Si vs C termination, and hydrophilic vs hydrophobic solid–liquid coupling strengths, then relates interfacial liquid structure—including density depletion lengths—to TBC, showing wetting metrics alone are not universal predictors of heat flow. The work targets thermal management at wide-bandgap semiconductor cooling interfaces where Kapitza resistance can dominate junction temperatures; by sweeping LJ solid–water ε across hydrophilic/hydrophobic regimes, the paper separates contact-angle-like metrics from structure-resolved heat-transfer descriptors.

Methods

From ACS Appl. Mater. Interfaces PDF (pdf_path).

  • Interactions: SPC/E water; LAMMPS + VMD; SiC slabs with fixed bottom layers and frozen constraint region; 6 nm gap between opposing slabs filled with 1000-1100 H2O molecules (adjusted with wettability to keep bulk pressure consistent per article). LJ solid-liquid parameters scanned to vary hydrophilic / hydrophobic coupling (epsilon_SL ranges tied to prior wetting work).
  • Equilibration: Minimization; NVT at 300 K (1 ns) with Nose-Hoover (time constant 100 fs); NVE 1 ns stability check.
  • Nonequilibrium thermal: Energy add/remove in hot/cold slab regions (5 ns ramp); production 7.5 ns sampling KE and coordinates every 10 ps. Timestep 1 fs. Thermal boundary conductance G from G = J / Delta T_int with heat currents 5-15 nW tested for linear response.

Findings

  • TBC varies with plane and termination beyond what a single wetting descriptor captures; interfacial water structure must be included.
  • Hydrophilic coupling tends to increase TBC vs hydrophobic cases in the trends summarized, but exceptions known from other systems motivate the structure-aware analysis used here.
  • Using density depletion length as a descriptor reconciles disparate TBC estimates that would otherwise appear inconsistent when only contact-angle–like metrics are used.
  • Linear-response checks (5–15 nW heat currents) support extracting G from steady ΔT across the interface without spurious nonlinear heating artifacts in the reported protocol window.

Limitations

  • Classical potentials for SiC–water may miss quantum corrections and polarization effects important for some quantitative comparisons.
  • LJ mixing rules used to dial hydrophilic/hydrophobic coupling are phenomenological; ab initio interfacial spectroscopy would be needed to tie ε scans to specific surface terminations in experiment.
  • Surface oxidation or hydrocarbon contamination on SiC in device fabrication can alter wetting and TBC relative to the idealized cleaved slabs simulated.

Reader notes (navigation)

Cross-link thermal boundary conductance surveys when comparing SiC power electronics cooling concepts to diamond or oxide substrates in other corpus entries.