Search for: All records

Gradient metasurfaces provide a well-established platform for the wavefront and polarization control of light propagating in free space. More recent studies have also focused on their integration with plasmonic and dielectric waveguides and optoelectronic devices such as light emitters and photodetectors. In this context, metasurfaces can be used to interface guided modes with free-space radiation, for example, to enforce the selective input coupling or detection of light with prescribed characteristics (in terms of direction of propagation, state of polarization, orbital angular momentum, etc.) or to manipulate the properties of the out-coupled or emitted light according to a desired system operation. These functionalities, which normally involve complex combinations of bulky optical elements, have important applications in multiple areas of high technological significance, from optical interconnects to microdisplays and multifunctional image sensors. Here, we provide a tutorial overview of this field of research, including a description of the basic building blocks (meta-atoms) used in integrated gradient metasurfaces, a review of the main capabilities reported to date, and an outlook on future directions of study and technological prospects. 

The non-Hermitian skin effect is a recently discovered phenomenon in periodic wave systems, whereby a macroscopic number of eigenmodes become tightly localized near the boundaries in the presence of losses or amplification. Its demonstration with semiconductor photonic crystals could create new opportunities to engineer their spectral, angular, and spatial response for multiple device applications. However, previous theoretical work has only investigated this effect in purely one- or two-dimensional structures with ad-hoc (and unusually large) values of the imaginary refractive index. Here, we show numerically that strongly confined skin modes can be established in realistic Si photonic crystal slabs of finite thickness, even at photon energies below the Si absorption edge, where the only source of non-Hermiticity is radiation losses. Furthermore, these modes exhibit a distinctive chiral behavior that enables their selective excitation on opposite sample boundaries with free-space radiation incident along different directions. 

We investigate a convolutional neural network with the first layer implemented by specially designed plasmonic metasurface photodetectors, showing an order of magnitude decrease in computational complexity for the accurate classification of transparent biological cells. 

We demonstrate the use of plasmonic gradient metasurfaces to tailor the angular response of generic planar photodetectors. The resulting devices are promising for a wide range of computational imaging applications with enhanced miniaturization and functionality. 

We use a generative deep learning method based on denoising diffusion probabilistic model to design plasmonic phase-imaging sensors for broadband operation. This flexible method enables optimized inverse design for a wide range of nanophotonic devices. 

Angle-sensitive photodetectors are a promising device technology for many advanced imaging functionalities, including lensless compound-eye vision, lightfield sensing, optical spatial filtering, and phase imaging. Here we demonstrate the use of plasmonic gradient metasurfaces to tailor the angular response of generic planar photodetectors. The resulting devices rely on the phase-matched coupling of light incident at select geometrically tunable angles into guided plasmonic modes, which are then scattered and absorbed in the underlying photodetector active layer. This approach naturally introduces sharp peaks in the angular response, with smaller footprint and reduced guided-mode radiative losses (and therefore improved spatial resolution and sensitivity) compared to analogous devices based on diffractive coupling. More broadly, these results highlight a promising new application space of flat optics, where gradient metasurfaces are integrated within image sensors to enable unconventional capabilities with enhanced system miniaturization and design flexibility. 

The development of photonic technologies for machine learning is a promising avenue toward reducing the computational cost of image classification tasks. Here we investigate a convolutional neural network (CNN) where the first layer is replaced by an image sensor array consisting of recently developed angle-sensitive metasurface photodetectors. This array can visualize transparent phase objects directly by recording multiple anisotropic edge-enhanced images, analogous to the feature maps computed by the first convolutional layer of a CNN. The resulting classification performance is evaluated for a realistic task (the identification of transparent cancer cells from seven different lines) through computational-imaging simulations based on the measured angular characteristics of prototype devices. Our results show that this hybrid optoelectronic network can provide accurate classification (>90%) similar to its fully digital baseline CNN but with an order-of-magnitude reduction in the number of calculations. 

We report a new technique for single-shot quantitative phase retrieval from transparent objects, based on plasmonic metasurface photodetectors featuring an asymmetric angular response around normal incidence combined with a particularly simple optical setup.