Deep learning for the design of photonic structures

Ma, Wei; Liu, Zhaocheng; Kudyshev, Zhaxylyk A.; Boltasseva, Alexandra; Cai, Wenshan; Liu, Yongmin

doi:10.1038/s41566-020-0685-y

Cited by 656 publications

(352 citation statements)

References 95 publications

Supporting

Mentioning

314

Contrasting

Order By: Relevance

“…As a result, since the complex refractive index of a‐Si:H can be fine‐tuned by controlling the structural disorder of silicon‐hydrogen bonding through the deposition parameters of PECVD, in addition to the ease of fabrication and increased substrate selectivity, low‐loss a‐Si:H is a promising material to significantly improve the efficiency of silicon‐based photonic devices that could play a key role in future photonic applications. [ 51,52 ] In combination with promising design methods such as semi‐continuous metasurfaces, [ 53–55 ] deep‐learning processes, [ 56,57 ] and topological optimization, [ 58 ] low‐loss a‐Si:H is a candidate for the use in all‐dielectric metasurfaces at visible frequencies.…”

Section: Discussionmentioning

confidence: 99%

Revealing Structural Disorder in Hydrogenated Amorphous Silicon for a Low‐Loss Photonic Platform at Visible Frequencies

et al. 2021

View full text Add to dashboard Cite

The high refractive index of hydrogenated amorphous silicon (a‐Si:H) at optical frequencies is an essential property for the efficient modulation of the phase and amplitude of light. However, substantial optical loss represented by its high extinction coefficient prevents it from being utilized widely. Here, the bonding configurations of a‐Si:H are investigated, in order to manipulate the extinction coefficient and produce a material that is competitive with conventional transparent materials, such as titanium dioxide and gallium nitride. This is achieved by controlling the hydrogenation and silicon disorder by adjusting the chemical deposition conditions. The extinction coefficient of the low‐loss a‐Si:H reaches a minimum of 0.082 at the wavelength of 450 nm, which is lower than that of crystalline silicon (0.13). Beam‐steering metasurfaces are demonstrated to validate the low‐loss optical properties, reaching measured efficiencies of 42%, 62%, and 75% at the wavelengths of 450, 532, and 635 nm, respectively. Considering its compatibility with mature complementary metal–oxide–semiconductor processes, the low‐loss a‐Si:H will provide a platform for efficient photonic operating in the full visible regime.

show abstract

Section: Discussionmentioning

confidence: 99%

Revealing Structural Disorder in Hydrogenated Amorphous Silicon for a Low‐Loss Photonic Platform at Visible Frequencies

et al. 2021

View full text Add to dashboard Cite

show abstract

“…Simulation-based machine learning approaches have been applied to GaN-LED design optimization [ 126 , 127 ] but produced unreliable results [ 128 ]. The great popularity of such methods in materials science [ 129 ] and in photonics [ 130 ] seems hard to transfer to optoelectronic devices considering their complex internal physics and their material parameter uncertainties. In fact, the strength of machine learning lies in the analysis of large amounts of experimental data which are often routinely collected in the industrial LED production.…”

Section: Key Modeling and Simulation Challengesmentioning

confidence: 99%

Efficiency Models for GaN-Based Light-Emitting Diodes: Status and Challenges

Piprek¹

2020

Materials

View full text Add to dashboard Cite

Light-emitting diodes (LEDs) based on Gallium Nitride (GaN) have been revolutionizing various applications in lighting, displays, biotechnology, and other fields. However, their energy efficiency is still below expectations in many cases. An unprecedented diversity of theoretical models has been developed for efficiency analysis and GaN-LED design optimization, including carrier transport models, quantum well recombination models, and light extraction models. This invited review paper provides an overview of the modeling landscape and pays special attention to the influence of III-nitride material properties. It thereby identifies some key challenges and directions for future improvements.

show abstract

“…Although conventional structural design methods, including physics‐based approaches and numerical simulations, offer important guidelines, it is not trivial to find the right structures with ideal selective emission spectra. We envisage simple photonic materials and structures designed and optimized by advanced methods such as the inverse design methods, [ 64–75 ] which enable nonintuitive and irregularly shaped structures, outperforming physically or empirically designed structures. Particularly, the artificial neural networks, as one of the most powerful machine learning methods, have shown orders of magnitude faster and much accuracy in optimizing structures in high dimensional space.…”

Section: Discussionmentioning

confidence: 99%

Photonics Empowered Passive Radiative Cooling

Zhang

Chu

Cai

et al. 2021

Advanced Photonics Research

View full text Add to dashboard Cite

Passive radiative cooling has recently received renewed interest because of its unprecedented capabilities in cooling terrestrial objects below ambient air temperature without external energy consumption and greenhouse gas emission. This technology has been demonstrated as promising as replacements/complements of conventional compressed air‐based active cooling systems, which can significantly impact the global energy landscape by providing a green and efficient cooling way. The key to this success is judiciously designed photonic micro/nanostructures, which simultaneously reflect solar irradiation and emit thermal infrared emission across the atmospheric transparency window 8–13 μm. Herein, an introduction of the fundamental principles of passive radiative cooling is given, discussing the critical factors associated with the net cooling power of radiative cooling. Following this, the recently emerged photonic materials and structures (e.g., multilayer thin films, micro/nanoparticles, photonic crystals, metamaterials, metasurfaces, etc.) that facilitate radiative cooling are reviewed and fruitfully analyzed and discussed. Some possible scale‐up manufacturing ways toward the practical deployment of this energy‐efficient technology in real‐world applications are then discussed. The potential applications are also summarized and envisioned. Finally, perspectives on the future development in conjunction with artificial intelligent design of photonic structures and materials are presented and discussed.

show abstract

Deep learning for the design of photonic structures

Cited by 656 publications

References 95 publications

Revealing Structural Disorder in Hydrogenated Amorphous Silicon for a Low‐Loss Photonic Platform at Visible Frequencies

Revealing Structural Disorder in Hydrogenated Amorphous Silicon for a Low‐Loss Photonic Platform at Visible Frequencies

Efficiency Models for GaN-Based Light-Emitting Diodes: Status and Challenges

Photonics Empowered Passive Radiative Cooling

Contact Info

Product

Resources

About