A preprint proposes a revised way to estimate how heat bends through layered thermal metamaterials after the conventional effective medium approximation, or EMA, tended to depart from numerical heat-flow calculations in some high-contrast settings. EMA is the analytical estimate the paper uses for the bending angle, while the new version adds two corrections: f(r), which represents reflection-like behavior where adjacent layers have sharply different thermal conductivities, and g(β), which represents resistance from the surrounding background medium. The authors report good agreement between this generalized EMA and numerical simulations when both high conductivity contrast and background-medium embedding are present.
The gap appears at high contrast
The work is a modeling study built around simulated configurations. In a representative case, the modeled system used two metallic layers in a seven-layer configuration aligned at 45 degrees. Heat-flow fields were generated with finite-element simulations in COMSOL Multiphysics, while conventional and generalized EMA calculations supplied the analytical comparisons.
To compare the approaches, the study tracked the heat-flow bending angle, φ. Its main gap measure, Δφ, was the difference between the conventional EMA angle and the angle from the finite-element model. The conventional EMA tended to depart from the simulations as the conductivity contrast r rose, the layer count n fell, and the background conductivity β fell. The numerical setup fixed the heat source at 373 K and the heat sink at 273 K, and applied perfectly matched layers to suppress boundary radiation and convection effects.
The two corrections
The paper then examines the thermal-energy ratio A2/A1 at the interface. The ratio increased monotonically for r below 5 and then slowly saturated. An empirical fit gave α1 = 0.56 and α2 = 0.71, following a dependence of the form (r-1)/(r+1), mathematically identical to the Fresnel reflection form used in optics. In the authors' interpretation, this is an optics-like analogy for reflection-like heat diffusion. The supplied analysis reports no goodness-of-fit measure or uncertainty for the fitted parameters.
That analogy becomes the first correction, f(r), which the model uses to account for reflection-like heat diffusion crossing adjacent layers with high conductivity contrast. The second, g(β), represents the extra thermal resistance of the surrounding medium, using an analogy to series-connected circuits. The generalized EMA revises the heat-flux coefficients by multiplying f(r) and g(β). In the authors' reading, f(r) is the main source of the large-r departure from conventional EMA, while g(β) accounts mainly for the overall reduction in the predicted bending angle φ.
In one comparison, at layer angles of 30, 45 and 60 degrees, with n = 14 and β = 30, generalized EMA predicted a slightly smaller φ than conventional EMA across r; the discrepancy increased as r increased. In a separate condition, with n = 32 and r = 30, conventional EMA was reported to describe the heat flow accurately when β exceeded 60 W m-1K-1. That example is not accompanied by a numerical error measure or uncertainty estimate in the supplied analysis.
A result still confined to simulation
The evidence remains a model-level result. The authors report good agreement between generalized EMA and numerical simulations when high conductivity contrast and background-medium embedding are present, but the supplied analysis does not provide comprehensive numerical error values or experimental confirmation. The simulations also used fixed thermal boundaries, and a mesh and domain check found stable results when sizes were below approximately 0.37 mm. Performance beyond the reported simulated conditions is therefore not established by this evidence.
The preprint reports support from the Regional Innovation System & Education (RISE) Glocal University 30 Program through the Gwangju RISE Center, funded by the Ministry of Education and Gwangju Metropolitan City in the Republic of Korea under identifier 2026-RISE(Glocal University 30)-05-011. The authors declare no known competing financial interests or personal relationships that could have appeared to influence the work. They state that data will be made available on request.
Paper data and sources
Original title: Generalizing Thermal Transport in High-Contrast Metamaterials through Interfacial Fresnel Reflection
Authors: Seung Hyeon Ham, Yu Min Kim, In Hyeok Choi, Jeong Woo Han
Journal/Repository: arXiv
Status: Preprint, not yet peer-reviewed
First online: 2026-08-26
DOI: Not available
Original paper · Full text