Phonon Transmission Function

• Phonon Transmission Function is a dimensionless measure that quantifies the probability of a phonon at frequency ω transmitting across an interface, capturing microscopic details.
• It is computed using methods like atomistic Green’s function and scattering approaches, which incorporate effects of lattice mismatch, polarization, and structural disorder.
• The function is pivotal for thermal management and nano-engineering, informing designs of phonon filters, superlattices, and thermoelectric devices.

The phonon transmission function, typically denoted $T(\omega)$, quantifies the probability that a vibrational excitation (phonon) of angular frequency $\omega$ incident on an interface or nanoscale region is transmitted from one lead or material to another. It is fundamentally important for modeling thermal boundary resistance, nanoscale heat transport, and the spectral content of energy flow at solid interfaces. The frequency dependence, polarization selectivity, and structural sensitivity of $T(\omega)$ encode microscopic details essential for understanding and engineering interfacial thermal conductance.

1. Physical Definition and Theoretical Foundations

The phonon transmission function emerges from the theoretical framework describing energy flow between coupled elastic media, most rigorously in the harmonic (elastic, coherent) regime. For a pair of materials or device-lead system, the transmitted heat flux per unit area, under steady-state conditions and small temperature difference $\Delta T$, is given by the Landauer formula: $$J = \frac{1}{2\pi} \int_0^\infty \hbar \omega\, T(\omega)\, [n_L(\omega) - n_R(\omega)]\, d\omega$$ where $n_{L/R}(\omega)$ are the Bose–Einstein distributions in the left and right leads, and $T(\omega)$ is the dimensionless transmission function, bounded between 0 and the total number of propagating modes per frequency. $T(\omega)$ subsumes all effects of interface structure and atomic-scale disorder, representing the net probability of successful phonon transmission across the interface per unit frequency (Dai et al., 30 Jan 2026, Hua et al., 2016, Das et al., 2012).

In mode-resolved terms, $T(\omega)$ sums the probabilities from all incoming states in the injecting medium to all outgoing states in the receiving medium: $$T(\omega) = \sum_{n\in L}\sum_{m \in R} |t_{mn}(\omega)|^2$$ where $\omega$0 is the transmission amplitude from mode $\omega$1 (left) to $\omega$2 (right) (Ong, 2018, Ong, 2018).

2. Atomistic Green's Function and Scattering Approaches

The most general and rigorous formalism for computing $\omega$3 is the atomistic Green's function (AGF) method. Under the harmonic approximation, the device or interface region is described by a real-space dynamical matrix $\omega$4, and the coupling to (semi-infinite) leads is treated via self-energy matrices $\omega$5: $\omega$6 The "broadening" or coupling matrices are $\omega$7, encoding the escape rate at each frequency.

The Landauer-Caroli formula for the total transmission function is then: $\omega$8 This expression remains valid for any interface structure—ideal, disordered, or defective—and does not require the central region to possess well-defined phonon momentum or periodicity (Dai et al., 30 Jan 2026, Ong, 2018, Bachmann et al., 2011, Gu et al., 2015).

Alternative approaches make explicit connection to wave-scattering. For 1D or layered systems, the transmission amplitude $\omega$9 is related to the energy flux, and

$T(\omega)$0

where $T(\omega)$1 are the group velocities in the leads at frequency $T(\omega)$2. In multi-mode systems, summation over branches and polarizations is implied (Das et al., 2012, Zhang et al., 2011).

3. Dependence on Interface Structure, Lattice Mismatch, and Polarization

The microscopic form of the phonon transmission function $T(\omega)$3 is highly sensitive to atomic-scale interface properties, mass mismatch, bonding strength, and symmetry breaking.

• Crystalline-Crystal Interfaces: Lattice mismatch leads to frequency-dependent reflection and transmission even in the absence of defects. Longitudinal phonons transmit much more efficiently than transverse phonons except at special commensurability resonances. For incommensurate lattices, the transverse transmission often vanishes (Meilakhs, 2015).
• Surface Roughness & Disorder: Atomic roughness and intermixing relax momentum conservation, enabling off-diagonal mode conversion and increasing phase space for transmission at the expense of coherence (Sadasivam et al., 2017).
• Density of States Overlap: Interfaces between materials with mismatched phonon spectra exhibit pronounced frequency cutoffs, with transmission suppressed when the density of states of one side vanishes (Gu et al., 2015).
• Angle and Polarization: Transmission is strongly angle-dependent; nearly all spectral conductance at a Si/Ge interface is carried by modes within a critical angle θ_c ≈ 50°, independent of interface details above ~30 K (Alkurdi et al., 2017). Mode-resolved evaluation reveals that low-frequency, long-wavelength longitudinal acoustic (LA) modes typically dominate the transmission spectrum (Hua et al., 2016, Hua et al., 2015).

4. Experimental Metrology and Inverse Reconstruction

Experimental access to spectral and mode-resolved phonon transmission has been achieved via time-domain thermoreflectance (TDTR) in combination with ab-initio phonon modeling and kinetic inversion:

• TDTR/BTE Inversion: By measuring the frequency-dependent phase and amplitude of thermal responses and fitting them to solutions of the Boltzmann transport equation (BTE) parametrized by a trial transmission function τ(ω), the spectral dependence of τ(ω) can be extracted. This reveals that atomically clean interfaces act as low-pass "phonon filters"—transmitting with near-unity probability at low ω, but rapidly rolling off in the high-ω regime (Hua et al., 2016, Hua et al., 2015).
• PDE-Constrained Optimization: Recent method development recasts the reconstruction of the frequency-dependent reflection coefficient η(ω) (with T(ω) = 1 - η(ω)) from surface thermal response data as a PDE-constrained optimization problem. The method uses forward and adjoint solutions of the linearized phonon BTE and computes gradients via Fréchet derivatives, enabling nonparametric recovery of T(ω) from experimental observables (Gamba et al., 2020).
• Stochastic Gradient Descent (SGD): Optimization is performed via SGD, whose convergence is justified by the maximum principle and Lipschitz continuity of the forward map gradient. The method is robust to moderate experimental noise and can recover both parametric and nonparametric representations of η(ω) over relevant frequency grids.

5. Influence of Inelastic (Anharmonic) Processes

While most theoretical and experimental approaches focus on the harmonic (elastic) regime, real interfaces may exhibit substantial anharmonic contributions to T(ω), especially in weakly interacting or highly mismatched cases:

• Three-Phonon Scattering: Inelastic (three-phonon) processes contribute an explicitly temperature-dependent term to T(ω), which can become dominant above ~90 K in van der Waals-coupled interfaces. In the weak-coupling limit, the anharmonic transmission function can grow linearly with temperature and even exceed the elastic channel, but this growth is ultimately bounded by higher-order processes and quasiparticle breakdown (Zhou et al., 2022).
• Physical Implication: These findings suggest that at room temperature, inelastic channels must be included for quantitative predictions of interfacial conductance in heterogeneous or soft-bonded material systems.

6. Engineering and Applications

Phonon transmission engineering underpins strategies for thermal management and thermoelectric enhancement:

• Phonon Filters and Superlattices: Specific geometries (isotope/mass barriers, antidot arrays, edge clamping) in low-dimensional systems like graphene nanoribbons can introduce sharp dips or windows in T(ω), enabling selected mode filtering for thermal diode/rectifier and thermoelectric applications (Scuracchio et al., 2014).
• Material Selection and Nano-structuring: Transmission function calculations allow identification of phonon blockers (e.g., ZnO/ZnS interfaces) that suppress conductance and increase the thermoelectric figure of merit ZT by restricting the spectral overlap of bulk modes (Bachmann et al., 2011).

7. Summary Table: Key Properties and Methods

Aspect	Physical/Mathematical Role	Typical Reference Equations/Papers
Definition	Probability of frequency-resolved phonon transfer	$T(\omega)$4 (Dai et al., 30 Jan 2026)
Computation	AGF/NEGF using dynamical matrices & self-energies	(Ong, 2018, Ong, 2018, Sadasivam et al., 2017)
Experimental access	TDTR + spectral BTE inversion	(Hua et al., 2016, Hua et al., 2015, Gamba et al., 2020)
Harmonic regime	Elastic, temperature-independent T(ω)	(Das et al., 2012, Zhang et al., 2011)
Anharmonic regime	Inelastic, T(ω,T) grows with T	(Zhou et al., 2022)
Angle/polarization	Selective, e.g., θ_c ≈ 50°, LA vs. TA	(Alkurdi et al., 2017, Meilakhs, 2015, Hua et al., 2016)
Engineering	Filters via superlattices, nanostructuring	(Scuracchio et al., 2014, Bachmann et al., 2011)

The phonon transmission function remains the cornerstone of modern understanding and engineering of interfacial and nanoscale thermal transport, encompassing both fundamental atomic-scale physics and the applied manipulation of heat flow in advanced materials systems.