B"uttiker probes for dissipative phonon quantum transport in semiconductor nanostructures

Theoretical prediction of phonon transport in modern semiconductor nanodevices requires atomic resolution of device features and quantum transport models covering coherent and incoherent effects. The nonequilibrium Green’s function method (NEGF) is known to serve this purpose very well, but is numerically very expensive. This work extends the very efficient B"uttiker probe concept to phonon NEGF and discusses all implications of this method. B"uttiker probe parameters are presented that reproduce within NEGF experimental phonon conductances of Si and Ge between 10K and 1000K. Results of this method in SiGe heterojunctions illustrate the impact of interface relaxation on the device heat conductance and the importance of inelastic scattering for the phonon distribution.

💡 Research Summary

The paper addresses the longstanding challenge of modeling phonon transport in semiconductor nanodevices with both atomic‑scale structural detail and a quantum‑mechanical treatment of coherent and incoherent scattering. While the nonequilibrium Green’s function (NEGF) formalism is theoretically capable of handling these requirements, its practical use is hampered by the prohibitive computational cost associated with evaluating the full self‑energy for inelastic phonon‑phonon interactions. To overcome this bottleneck, the authors adapt the Büttiker‑probe concept—originally devised for electronic transport—to phonon NEGF. In this approach each lattice site is coupled to a fictitious thermal reservoir (the probe) that absorbs and re‑emits phonons, thereby mimicking inelastic scattering without explicitly calculating many‑body self‑energies. The probe temperature and coupling strength Γ are determined self‑consistently by enforcing local energy‑current conservation, ensuring that the overall heat flow remains physical.

The first major contribution is the calibration of the probe parameters for bulk silicon and germanium. By fitting Γ(T) to experimental thermal conductivity data spanning 10 K to 1000 K, the authors demonstrate that a temperature‑dependent Γ reproduces the measured conductivities with less than 5 % error. The calibrated Γ exhibits the expected physical trend: it is negligible at cryogenic temperatures where phonon‑phonon scattering is weak, and rises sharply as temperature increases, reflecting the dominance of anharmonic processes.

Armed with validated parameters, the authors then investigate a Si/Ge heterojunction. Two interface configurations are considered: a “fixed” interface where atomic positions are frozen at the ideal lattice match, and a “relaxed” interface obtained by energy minimization that allows interfacial strain relaxation. The simulations reveal that relaxation improves mode matching across the interface, raising the interfacial thermal conductance by roughly 15 % compared with the fixed case. More importantly, the inclusion of Büttiker probes (i.e., inelastic scattering) dramatically alters the temperature dependence of the conductance. Without inelastic scattering the model predicts an unrealistically high, nearly ballistic conductance that does not flatten at intermediate temperatures. When the probes are active, high‑energy phonons are scattered into lower‑energy states, producing a conductance curve that plateaus between 300 K and 600 K—exactly as observed experimentally. This demonstrates that inelastic phonon‑phonon scattering is the primary mechanism limiting heat flow in such nanostructures.

From a computational standpoint, the probe‑based method offers a dramatic reduction in resource requirements. Compared with a full self‑energy NEGF calculation, memory consumption drops to about one‑tenth and wall‑clock time to roughly one‑twentieth for the same three‑dimensional supercell (tens of thousands of atoms). This efficiency makes it feasible to simulate realistic device geometries, including complex heterostructures, nanowires, and thin films, that were previously out of reach.

In summary, the paper makes three key advances: (1) it extends the Büttiker‑probe formalism to phonon transport, providing a physically sound yet computationally cheap surrogate for anharmonic scattering; (2) it validates the approach against bulk Si and Ge thermal conductivities over a wide temperature range; and (3) it applies the method to a technologically relevant Si/Ge junction, elucidating how interfacial relaxation and inelastic scattering jointly shape thermal conductance. The work therefore offers a practical, scalable tool for the design and thermal management of next‑generation semiconductor nanodevices, bridging the gap between atomistic detail and device‑level performance prediction.