Spectral mapping of thermal transport across SiC-water interfaces
Summary¶
Nonequilibrium molecular dynamics (NEMD) with spectral heat-flux mapping analyzes thermal boundary conductance (TBC) between 3C–SiC and water for C- vs Si-terminated (100) and (111) surfaces under hydrophobic vs hydrophilic LJ tuning. The International Journal of Heat and Mass Transfer article (DOI 10.1016/j.ijheatmasstransfer.2018.11.101) decomposes interfacial heat flux into frequency-resolved contributions q(ω) and relates mode overlap integrals to temperature-jump thermal boundary conductance, letting the authors compare wetting trends, phonon density of states, and spectral transport anisotropy across four termination cases. Low-frequency modes dominate transport for several terminations, while Si-terminated SiC(111) shows notable high-frequency contributions. Out-of-plane vibrational components carry most of the interfacial heat flux; in-plane fractions are smaller—especially on (111)—and correlate with interfacial liquid structure and DOS overlap more cleanly than with interfacial bonding strength alone.
Methods¶
Classical water and SiC bulk models (A/B)¶
Water: SPC/E, rigid (SHAKE), PPPM electrostatics (accuracy 10⁻⁶). SiC: MEAM for Si–Si, C–C, Si–C; solid–liquid coupling via 12–6 Lennard-Jones on Si–O and C–O only (C–O r and ε fixed; Si–O ε scanned 0.005–0.025 eV at fixed Si–O r = 2.63 Å to span wetting).
NEMD geometry and protocol (B)¶
Cells: two SiC slabs sandwiching ~6 nm water; slab ~10 nm long (z); cross section 2.62×2.62 nm² (100) or 2.78×2.67 nm² (111); fixed outer layers; 1.5 nm hot/cold NEMD regions.
Workflow (LAMMPS + VMD): minimize → NVT 300 K, 1 ns (Nosé–Hoover, 0.1 ps τ) → NVE 1 ns sanity check → 5 ns NEMD relax → 5 ns production (sample 25 ps).
Spectral binning: q(ω) is accumulated in frequency bins after 5 ns NEMD relaxation and 5 ns production, with 25 ps chunks used for time averaging of interfacial heat flux spectra as stated in the Methods section of the article.
Analysis¶
Spectral heat flux q(ω) from force–velocity correlations; DOS from VACF; TBC from temperature-jump fits.
ReaxFF / reactive training (A)¶
Not used—non-reactive MEAM + LJ interface study.
Integration parameters (NEMD). LAMMPS molecular dynamics on SiC/water slabs uses periodic in-plane PBC with fixed outer layers, sandwiching ~6 nm water between ~10 nm slabs (cross sections stated above). Workflow: NVT at 300 K for 1 ns with Nose–Hoover thermostat (0.1 ps time constant) → NVE sanity 1 ns → 5 ns NEMD relaxation → 5 ns production with 25 ps chunks for spectral averaging. Timestep: the article’s Methods tabulate the stable timestep in fs for MEAM/SPC/E stability (see pdf_path if not repeated here). Barostat / hydrostatic pressure: N/A — fixed lateral dimensions without GPa pressure servo in the summarized NEMD path. External electric field: N/A. Enhanced sampling: N/A.
Findings¶
Comparisons and sensitivity¶
Literature context: the article contrasts TBC trends with prior wettability scaling arguments (e.g., contact-angle–based correlations in the literature) and shows cases where DOS overlap aligns better with thermal boundary conductance than bond strength alone. Sensitivity: Si–O Lennard-Jones ε scans (0.005–0.025 eV) modulate wetting and, through phonon overlap, shift q(ω) and TBC between hydrophilic and hydrophobic (100)/(111) terminations. Temperature: 300 K liquid preparation before NEMD establishes the thermal boundary conditions quoted above.
Limitations and corpus honesty¶
Rigid SPC/E water and empirical SiC–water LJ coupling omit polarization and quantum effects that could matter for some interfaces—limitations stated in the article discussion. PDF-grounded detail: numeric timestep and any auxiliary pressure/stress controls beyond this summary appear only in the Methods of pdf_path; this page does not invent missing numbers.
Mechanisms¶
TBC correlates with phonon DOS overlap S more cleanly than wettability alone; Si-terminated (100) hydrophilic cases show among the strongest S/highest TBC vs hydrophobic C-terminated (100). Out-of-plane modes dominate q(ω) on (100); in-plane fraction drops on (111). Si-terminated (111) can contribute high-ω DOS beyond the bulk gap; C-terminated surfaces concentrate q(ω) at lower THz cutoffs than Si-terminated on the same plane. Depletion length reconciles TBC scatter vs contact-angle expectations.
Limitations¶
Empirical SiC–water LJ coupling is minimal; rigid SPC/E omits quantum/polarization effects for some interfaces.
Relevance to group¶
Penn State MNE collaboration on nanoscale thermal transport at solid–liquid interfaces.
Citations and evidence anchors¶
papers/Others/Gonzalez_Valle_water_SiC_IntJHeatTransfer_2019.pdf