Numerical Evaluation of Angle-Dependent IR-Transparent Radiative Cooling Performance for Asymmetric Periodic Structures

TL;DR

The paper tackles the challenge that single-angle asymmetric IR transmission is not a reliable predictor of non-contact radiative cooling performance. It develops a two-dimensional angle-resolved EM framework based on discrete exterior calculus with Bloch periodic boundary conditions and Floquet mode decomposition to obtain wavelength- and angle-dependent reflection and transmission for periodic IR-transparent PRC structures, and couples this with an energy-balance thermal model to predict transient and steady-state temperatures. The key finding is that practical cooling relies on angularly distributed asymmetry across the hemisphere; relying on normal-incidence data can overestimate cooling and even predict cooling where none occurs when full angular distribution is considered. This work establishes angularly distributed asymmetric transparency as a fundamental EM design principle for wide-angle metasurface–based radiative cooling and provides a robust methodology for accurate performance prediction in realistic environments.

Abstract

Infrared (IR)-transparent passive radiative cooling (PRC) enables non-contact thermal management by regulating radiative heat exchange without direct attachment to the cooling object. While asymmetric IR transmission at a specific incidence angle -- typically normal incidence -- is often emphasized, we show that such single-angle asymmetry is neither sufficient nor predictive of practical cooling performance. In this work, we demonstrate that effective non-contact PRC requires angularly distributed asymmetric IR transparency evaluated through hemispherical integration over emission directions, rather than asymmetry at a single incidence angle. To quantify this effect, an angle-resolved full-wave electromagnetic (EM) model with Bloch periodic boundary conditions and Floquet mode decomposition is employed to compute wavelength- and angle-dependent bidirectional reflection and transmission of periodic PRC structures. The resulting EM response is coupled to an energy-balance-based thermal model to predict the transient temperature evolution of the cooling object. By comparing models that account for the full angular distribution with normal-incidence-only approximations, we show that pronounced asymmetric transmission at normal incidence is generally not preserved at oblique angles. As a result, angular integration yields only marginal cooling or may even result in net heating, whereas normal-incidence-based models can substantially overestimate cooling performance. These results establish angularly distributed asymmetric transparency as a key EM design principle for IR-transparent PRC and wide-angle asymmetric metasurfaces.

Figures (6)

• Figure 1: Schematic of the AEMT device equipped with the ALT function.
• Figure 2: (a) Corresponding test structure, and (b) the 2D periodic unit cell employed for numerical analysis with Bloch PBC.
• Figure 3: Comparison of reflectance (R) and transmittance (T) spectra of the AEMT structure obtained using the FDTD and the in-house DEC solvers under normal-incidence conditions. The agreement between the two results verifies the accuracy of the developed DEC solver.
• Figure 4: Angle-wavelength maps of the reflection and transmission coefficients for sky-to-object and object-to-sky incidence: (a) $R_{\mathrm{STO}}(\lambda,\theta_i)$, (b) $T_{\mathrm{STO}}(\lambda,\theta_i)$, (c) $R_{\mathrm{OTS}}(\lambda,\theta_i)$, and (d) $T_{\mathrm{OTS}}(\lambda,\theta_i)$. The pronounced normal-incidence asymmetry progressively degrades at oblique angles, with increased STO transmission and reduced OTS transmission over a broad spectral range.
• Figure 5: Spatial distributions of the electric-field magnitude $|\mathbf{E}|$ at a wavelength of $10~\mu$m: (a) normal incidence for STO and OTS, and (b) oblique incidence ($50^\circ$) for STO and OTS. The results highlight the distinct transmission behaviors of STO- and OTS-based structures at grazing incidence.