📄 other

Interfacial Thermal Conductance Due To Electronic Heat Transport Across A Normal-metal to Superconductor Boundary Of Dissimilar Materials

This paper reanalyzes six decades of thermal transport data across normal-metal to superconductor interfaces to demonstrate that electronic heat transport, particularly when modeled using the Bardeen-Rickayzen-Teword (BRT) framework adapted with the Wiedemann-Franz law, is a dominant and necessary mechanism alongside phonons to explain experimental results across a wide temperature range, whereas other proposed electronic models fail to align with observations.

Original authors: Robert Young

Published 2026-07-24

📖 4 min read☕ Coffee break read

CC BY 4.0

Original authors: Robert Young

Original paper licensed under CC BY 4.0 (https://creativecommons.org/licenses/by/4.0/). ✨ This is an AI-generated explanation of the paper below. It is not written or endorsed by the authors. For technical accuracy, refer to the original paper. Read full disclaimer

Imagine heat as a bustling crowd of tiny messengers trying to cross a border between two different neighborhoods. In the world of physics, these messengers come in two main flavors: phonons and electrons. Phonons are like the vibrations of the ground itself—think of them as the "shaking" of the material's atoms, similar to how a ripple moves through a crowd of people holding hands. Electrons, on the other hand, are the actual charged particles zooming through the material, like individual runners sprinting through the crowd. Usually, in metals, these speedy electrons are the main carriers of heat, zooming past the slower, shuffling phonons.

Now, imagine one of those neighborhoods is a special place called a superconductor. When this neighborhood gets cold enough, something magical happens: the electrons pair up and form a synchronized dance called "Cooper pairs." These pairs are so coordinated they can move without any friction, but there's a catch: they stop carrying heat entirely. They become heat-silent dancers. The only electrons left to carry heat are the "unpaired" ones, which are few and far between when it's very cold. This creates a tricky puzzle for scientists: when heat tries to cross the border from a normal metal into this superconductor, how does it get across? For decades, researchers assumed the "ground vibrations" (phonons) were doing most of the work, but the data didn't add up. The heat was moving faster than the vibrations alone could explain, suggesting the few remaining unpaired electrons were still doing some heavy lifting, perhaps through secret tunnels or special shortcuts.

This paper by Robert Young tackles that mystery by looking at old experimental data from the last sixty years, specifically focusing on the "border crossing" between normal metals (like copper) and superconductors (like lead or tin). The author acts like a detective, testing four different theories about how those unpaired electrons might be sneaking across the boundary. The first theory, Andreev transport, suggests electrons bounce off the border and turn into pairs, but the math only works if the actual contact area is tiny—like a tiny pinprick in a giant wall. The second theory, Griffin and Maki, proposes that electrons tunnel through a thin oxide layer, but the data shows this model doesn't fit the real-world measurements very well.

The paper finds that the best explanations come from two other ideas, often working together. One is NIS tunneling, where electrons tunnel through a thin insulating layer (like a dirty spot on the contact), and the other is Bardeen-Rickayzen-Tewordt (BRT) scattering, where electrons bounce off defects or impurities at the interface. By applying a classic physics rule called the Wiedemann-Franz law (which links how well a material conducts electricity to how well it conducts heat) to these boundaries, the author shows that a combination of these electron mechanisms and the phonon vibrations can explain the data perfectly.

The study reveals that at very low temperatures (below about 20% of the superconductor's critical temperature), the "ground vibrations" (phonons) dominate the heat transfer. However, as it warms up slightly (between 30% and 50% of that critical temperature), the unpaired electrons take over, often carrying more heat than the vibrations. The author suggests that in many real-world samples, the contact isn't perfectly clean; it's likely "dirty" with oxide layers or defects. This means that while the "Andreev" theory (the clean bounce) might happen in a tiny fraction of the contact area (less than 1%), the rest of the heat is moving through the "dirty" paths of tunneling and scattering. The paper concludes that to accurately predict heat flow in these systems, we must account for both the vibrations and the electrons, and that the "dirty" nature of the contact is the key to unlocking why the heat moves the way it does.

Technical Summary: Interfacial Thermal Conductance Due to Electronic Heat Transport Across a Normal-Metal to Superconductor Boundary of Dissimilar Materials

Problem Statement
For over six decades, experimental measurements of heat transport across normal-metal to superconductor (NS) interfaces have presented a challenge: attributing this transport exclusively to lattice vibrations (phonons), which typically scale with the cube of temperature ( $T^3$ ), is inadequate. Historical data often showed temperature dependencies ranging from $T^5$ to $T^6$ , which researchers deemed inexplicable under pure phonon models. While it was hypothesized that electrons carry the "extra" heat, no single theory has fully described the thermal boundary resistance (TBR) from the normal state, through the superconducting transition temperature ( $T_c$ ), and into the deep superconducting regime. This paper addresses the lack of a unified theoretical framework to explain interfacial thermal conductance (ITC) in dissimilar metal NS contacts, specifically examining the role of electronic heat carriers alongside phonons.

Methodology
The study employs a comparative analysis of historical experimental data from eleven different samples (primarily Pb-Cu and Sn-Cu interfaces) against four distinct theoretical models for electronic heat transport. The analysis assumes that total thermal transport is the sum of parallel phonon and electron components. The four electronic models evaluated are:

Andreev Transport: Based on the theory of Andreev (1964) and extended by Bonnard et al., describing reflection at a clean NS interface.
Griffin-Maki Tunneling: Models quantum-mechanical tunneling of electron heat current across a thin oxide layer, with separate equations for low and high-temperature limits.
Giaever-Megerle NIS Tunneling: Describes tunneling across an insulating layer (Normal-Insulator-Superconductor) using an analytical series solution for electrical conductance, converted to thermal conductance.
Adapted Bardeen-Rickayzen-Tewordt (BRT): Originally a bulk thermal conductivity ratio, this model is adapted here for interfacial transport by applying the Wiedemann-Franz (WF) law to the interface.

A key methodological innovation is the application of the Wiedemann-Franz law to interfacial conductance. This allows the electronic thermal conductance to be expressed in terms of a single input parameter: the specific electrical contact resistance ( $R_cA$ ). This parameter can be derived either from direct electrical measurements or from the slope of the ITC versus temperature plot in the fully normal state.

Key Contributions

Unified Framework via Wiedemann-Franz Law: The paper demonstrates that by applying the WF law to the interface, the complex electronic transport models (specifically the adapted BRT and Giaever-Megerle NIS) can be reduced to a single fitting constant ( $R_cA$ ). This simplifies the prediction of ITC across the entire temperature range, from above $T_c$ (normal state) down to deep superconducting temperatures.
Quantification of "Dirty" Interfaces: The analysis reveals that for interfaces prepared in ambient conditions (the majority of historical data), the "clean" Andreev transport mechanism accounts for only a tiny fraction (1% or less) of the physical contact area. The majority of the interface behaves as if it contains insulating patches or defects.
Mechanism Identification: The study systematically evaluates which mechanisms align with experimental data. It finds that Griffin-Maki equations do not align well with the full temperature range. In contrast, the adapted BRT and NIS tunneling models, particularly when combined, successfully fit the data.
Phonon-Electron Transition: The paper maps the transition region where heat transport shifts from being phonon-dominated (below $\sim 0.2 T_c$ ) to electron-dominated (above $\sim 0.5 T_c$ ). It establishes that in the intermediate range ( $0.2 < T/T_c < 0.5$ ), both mechanisms must be accounted for in parallel to match experimental observations.

Results

Andreev Transport: To fit experimental data, the Andreev model required scaling down the effective contact area to approximately 1% or less of the physical area. This suggests that in "dirty" interfaces (prepared in air), the majority of the interface is insulating, preventing direct NS contact.
Griffin-Maki Models: These models failed to provide a consistent fit across the temperature range. The low-temperature and high-temperature limits required different contact resistance values that could not be reconciled with the full dataset.
Adapted BRT and NIS:
- The adapted BRT model, using a single $R_cA$ value derived from the normal-state slope, provided excellent fits for data spanning from above $T_c$ to well below it.
- The NIS tunneling model also provided good fits.
- Combined Mechanisms: The best overall fits were often achieved by assuming a parallel combination of mechanisms. For instance, a combination of scaled Andreev transport (representing the small clean patches) and NIS tunneling (representing the oxide barriers) could explain the data, predicting a peak in ITC just below $T_c$ that aligns with electrical resistance dips observed in other studies.
- Series vs. Parallel: In multi-layer samples (e.g., Challis and Cheeke's samples), the data suggested that bulk thermal resistance and interfacial resistance act in series, while phonon and electron interfacial resistances act in parallel.
Contact Resistance Values: The derived specific contact resistances ( $R_cA$ ) for the "dirty" ambient-prepared samples ranged from 0.018 to 0.72 $\Omega\cdot\mu m^2$ . This is one to two orders of magnitude higher than the values (0.001–0.002 $\Omega\cdot\mu m^2$ ) measured in clean, vacuum-deposited thin-film interfaces, confirming the "dirty" nature of the historical bulk samples.
Temperature Dependence: The analysis confirms that below $\sim 0.2 T_c$ , phonon transport dominates. Above $\sim 0.5 T_c$ , electronic transport dominates. In the transition zone, the sum of both is required.

Significance
The paper claims that the primary significance of this work is the provision of a practical, unified method to model interfacial thermal conductance in NS contacts using the Wiedemann-Franz law. By reducing the electronic component to a single parameter ( $R_cA$ ), the adapted BRT model offers a robust tool for predicting thermal performance across the full temperature spectrum without needing complex, multi-parameter fits for different regimes.

The study concludes that historical discrepancies in thermal data (e.g., $T^5$ or $T^6$ dependencies) are resolved by recognizing the simultaneous presence of phonon and electronic transport, where the electronic component is heavily influenced by interfacial defects or oxide layers (NIS tunneling) rather than just clean Andreev reflection. The paper modestly notes that while the adapted BRT model is the simplest approach for prediction, the simultaneous presence of multiple mechanisms (BRT and NIS) is likely in many real-world samples. It emphasizes that distinguishing the precise dominant mechanism in a specific sample would require simultaneous electrical and thermal measurements across a wide temperature range, a capability not present in the historical datasets analyzed.

More like this