Machine-learning designed smart coating: temperature-dependent self-adaptation between a solar absorber and a radiative cooler

Abstract

We designed a multilayered self-adaptive absorber/emitter metamaterial, which can smartly switch between a solar absorber and a radiative cooler based on temperature change. The switching capability is facilitated by the phase change material and the structure is optimized by machine learning. Our design not only advances the machine-learning-based development of metamaterials but also has the potential to significantly reduce carbon emissions and contribute to the goal of achieving carbon neutrality.