Search for: All records

Abstract A body that violates Kirchhoff’s law of thermal radiation exhibits an inequality in its spectral directional absorptivity and emissivity. Achieving such an inequality is of fundamental interest as well as a prerequisite for achieving thermodynamic limits in photonic energy conversion¹and radiative cooling². Thus far, inequalities in the spectral directional emissivity and absorptivity have been limited to narrow spectral resonances³, or wavelengths well beyond the infrared regime⁴. Bridging the gap from basic demonstrations to practical applications requires control over a broad spectral range of the unequal spectral directional absorptivity and emissivity. In this work, we demonstrate broadband nonreciprocal thermal emissivity and absorptivity by measuring the thermal emissivity and absorptivity of gradient epsilon-near-zero InAs layers of subwavelength thicknesses (50 nm and 150 nm) with an external magnetic field. The effect occurs in a spectral range (12.5–16 μm) that overlaps with the infrared transparency window and is observed at moderate (1 T) magnetic fields. 

From higher computational efficiency to enabling the discovery of novel and complex structures, deep learning has emerged as a powerful framework for the design and optimization of nanophotonic circuits and components. However, both data-driven and exploration-based machine learning strategies have limitations in their effectiveness for nanophotonic inverse design. Supervised machine learning approaches require large quantities of training data to produce high-performance models and have difficulty generalizing beyond training data given the complexity of the design space. Unsupervised and reinforcement learning-based approaches on the other hand can have very lengthy training or optimization times associated with them. Here we demonstrate a hybrid supervised learning and reinforcement learning approach to the inverse design of nanophotonic structures and show this approach can reduce training data dependence, improve the generalizability of model predictions, and significantly shorten exploratory training times. The presented strategy thus addresses several contemporary deep learning-based challenges, while opening the door for new design methodologies that leverage multiple classes of machine learning algorithms to produce more effective and practical solutions for photonic design. 

Abstract Controlling both the spectral bandwidth and directionality of emitted thermal radiation is a fundamental challenge in contemporary photonics. Recent work has shown that materials with a spatial gradient in the frequency range of their epsilon‐near‐zero (ENZ) response can support broad spectrum directionality in their emissivity, enabling high total radiance to specific angles of incidence. However, this capability is limited spectrally and directionally by the availability of materials with phonon‐polariton resonances over long‐wave infrared wavelengths. Here, an approach is designed and experimentally demonstrated using doped III–V semiconductors that can simultaneously tailor spectral peak, bandwidth, and directionality of infrared emissivity. InAs‐based gradient ENZ photonic structures that exhibit broadband directional emission with varying spectral bandwidths and directional ranges as a function of their doping concentration profile and thickness are epitaxially grown and characterized. Due to its easy‐to‐fabricate geometry, it is believed that this approach provides a versatile photonic platform to dynamically control broadband spectral and directional emissivity for a range of emerging applications in heat transfer and infrared sensing.