Nanophotonic color splitters for high-efficiency imaging

Johlin, Eric

doi:10.1016/j.isci.2021.102268

Cited by 29 publications

(22 citation statements)

References 32 publications

Supporting

Mentioning

Contrasting

Order By: Relevance

“…Moreover, these numerical simulations are not assured to find globally optimal solutions, and often, a large number of initial conditions must be tested to map out the configuration space, even when single designs are desired. 5 …”

Section: Introductionmentioning

confidence: 99%

Machine Learning in Interpolation and Extrapolation for Nanophotonic Inverse Design

Acharige

Johlin

2022

ACS Omega

Self Cite

View full text Add to dashboard Cite

The algorithmic design of nanophotonic structures promises to significantly improve the efficiency of nanophotonic components due to the strong dependence of electromagnetic function on geometry and the unintuitive connection between structure and response. Such approaches, however, can be highly computationally intensive and do not ensure a globally optimal solution. Recent theoretical results suggest that machine learning techniques could address these issues as they are capable of modeling the response of nanophotonic structures at orders of magnitude lower time per result. In this work, we explore the utilization of artificial neural network (ANN) techniques to improve the algorithmic design of simple absorbing nanophotonic structures. We show that different approaches show various aptitudes in interpolation versus extrapolation, as well as peak performances versus consistency. Combining ANNs with classical machine learning techniques can outperform some standard ANN techniques for forward design, both in terms of training speed and accuracy in interpolation, but extrapolative performance can suffer. Networks pretrained on general image classification perform well in predicting optical responses of both interpolative and extrapolative structures, with very little additional training time required. Furthermore, we show that traditional deep neural networks are able to perform significantly better in extrapolation than more complicated architectures using convolutional or autoencoder layers. Finally, we show that such networks are able to perform extrapolation tasks in structure generation to produce structures with spectral responses significantly outside those of the structures on which they are trained.

show abstract

Section: Introductionmentioning

confidence: 99%

Machine Learning in Interpolation and Extrapolation for Nanophotonic Inverse Design

Acharige

Johlin

2022

ACS Omega

Self Cite

View full text Add to dashboard Cite

show abstract

“…Filtering schemes like this come at a cost of transmission efficiency because they absorb all light outside of their passband leading to average transmission values of approximately 33% under uniform spectral illumination. Solutions to this problem have converged on the concept of color routing where scattering structures accept light incident on a group of pixels and redirect each wavelength band to a different pixel [2,9,[15][16][17]. In this manuscript, we demonstrate an efficient, multilayer inverse-designed device in the mid-infrared for accomplishing this task and further augment it to sense linear polarization.…”

Section: Mainmentioning

confidence: 99%

3D-Patterned Inverse-Designed Mid-Infrared Metaoptics

Roberts¹,

Ballew²,

Zheng³

et al. 2022

Preprint

View full text Add to dashboard Cite

“…Additionally, the CRs reported can fully utilise the light incident on the imaging sensors and route different colours to the corresponding pixels. Three-dimensional (3D) CRs have been demonstrated with the aim of improving the energy utilisation efficiency and colour collection efficiencies of RGB light by using multilayer stacked structures or 3D structures 14 , 15 designed using the inverse-design method 20 – 23 . Despite the theoretically great performance, the difficulties in fabricating 3D nanostructures in multiple lithographies and precise alignment hinder their realisation, especially for devices working in the visible region.…”

Section: Introductionmentioning

confidence: 99%

Pixel-level Bayer-type colour router based on metasurfaces

et al. 2022

View full text Add to dashboard Cite

The three primary colour model, i.e., red, green, and blue model, based on the colour perception of the human eye, has been widely used in colour imaging. The most common approach for obtaining colour information is to use a Bayer colour filter, which filters colour light with four pixels of an imaging sensor to form an effective colour pixel. However, its energy utilization efficiency and colour collection efficiency are limited to a low level due to the three-channel filtering nature. Here, by employing an inverse-design method, we demonstrate a pixel-level metasurface-based Bayer-type colour router that presents peak colour collection efficiencies of 58%, 59%, and 49% for red, green and blue light, and an average energy utilization efficiency as high as 84% over the visible region (400 nm–700 nm), which is twice as high as that of a commercial Bayer colour filter. Furthermore, by using a 200 µm × 200 µm metasurface-based colour router sample working with a monochromatic imaging sensor, colour imaging is further realized, obtaining an image intensity twice that achieved by a commercial Bayer colour filter. Our work innovates the mechanism of high-efficiency spectrum information acquisition, which is expected to have promising applications in the development of next-generation imaging systems.

show abstract

Nanophotonic color splitters for high-efficiency imaging

Cited by 29 publications

References 32 publications

Machine Learning in Interpolation and Extrapolation for Nanophotonic Inverse Design

Machine Learning in Interpolation and Extrapolation for Nanophotonic Inverse Design

3D-Patterned Inverse-Designed Mid-Infrared Metaoptics

Pixel-level Bayer-type colour router based on metasurfaces

Contact Info

Product

Resources

About