We present the next generation covert plasmonic security labels based on Ag nanowire structures and their polarization dependent surface-enhanced Raman scattering (SERS) imaging. The consist of nanowires fabricated by two-photon lithography thermal evaporation, where molecular probes choice are deposited. Simulation experimental results show that SERS signals from embedded molecules depend significantly incident field. information cannot be revealed directly physical features, but can only...

The single mode hybrid dielectric-loaded plasmonic waveguide is presented at the wavelength of 1.55 μm. We show that this waveguiding structure, consisting a low-index SiO2-stripe sandwiched between high-index Si-nanowire and silver film, achieves both long propagation length strong field confinement with high power intensity. Components such as 90°-circular S-shaped bends, based on proposed an intensity area 50×200 nm2, can obtain total transmission efficiency exceeding 85% for various bend...

Printing technology based on plasmonic structures has many advantages over pigment color printing such as high resolution, ultra-compact size and low power consumption. However, due to losses broad resonance behavior of metals in the visible spectrum, it becomes challenging produce well-defined colors. Here, we investigate cross-shaped dielectric nanoresonators which enable quality spectral regime and, hence, We numerically predict experimentally demonstrate that proposed all-dielectric...

Deep Neural Networks (DNN) have shown early promise for inverse design with their ability to arrive at working designs much faster than conventional optimization techniques. Current approaches, however, require complicated workflows involving training more one DNN address the problem of non-uniqueness in inversion and emphasis on speed has overshadowed far important consideration solution optimality. We propose demonstrate a simplified workflow that pairs forward-model evolutionary...

Transfer learning through the use of synthetic images and pretrained convolutional neural networks offers a promising approach to improve object detection performance deep networks. In this paper, we explore different strategies generate datasets subsequently them achieve better accuracy (mAP) when trained with state-of-the-art networks, focusing on packed food products in refrigerator. We develop novel techniques like dynamic stacking, pseudo random placement, variable pose, distractor...

We analyze and optimize the performance of coupled plasmonic nanoantennas for refractive index sensing. The investigated structure supports a sub- super-radiant mode that originates from weak coupling dipolar quadrupolar mode, resulting in Fano-type spectral line shape. In our study, we vary near-field two modes particularly examine influence detuning between them on sensing performance. Surprisingly, case matched resonance frequencies does not provide best sensor. Instead, find right amount...

Mid-infrared supercontinuum (SC) extending to ~4.0 mum is generated with 1.3 W time-averaged power, the highest power our knowledge, in ZBLAN (ZrF(4)-BaF(2)-LaF(3)-AlF(3)-NaF...) fluoride fiber by using cladding-pumped amplifiers and modulated laser diode pulses. We demonstrate scalability of SC average varying pump pulse repetition rate while maintaining similar peak power. Simulation results obtained solving generalized nonlinear Schrödinger equation show that long wavelength edge...

The surface-enhanced Raman scattering (SERS) spectrum exhibits huge potential as an alternative data storage element. Using plasmonic nanostructures the physical building blocks where probe molecules are adsorbed, their corresponding structural and SERS information stored within a finite volume of nanostructures. However, current development is hampered by difficulty in fabricating quantitative homogeneous platforms. Here, we introduce concept "plasmonic molecular storage" using intensity...

A major challenge in plasmonic hot spot fabrication is to efficiently increase the volumes on single metal nanoparticles generate stronger signals plasmon‐enhanced applications. Here, synthesis of designer nanoparticles, where plasmonic‐active Au nanodots are selectively deposited onto edge/tip regions Ag demonstrated using a two‐step seed‐mediated precision approach. Such “hot spots over spots” strategy leads an efficient enhancement nanoparticles. Through cathodoluminescence hyperspectral...

The need for large annotated image datasets training Convolutional Neural Networks (CNNs) has been a significant impediment their adoption in computer vision applications. We show that with transfer learning an effective object detector can be trained almost entirely on synthetically rendered datasets. apply this strategy detecting pack- aged food products clustered refrigerator scenes. Our CNN only 4000 synthetic images achieves mean average precision (mAP) of 24 test set 55 distinct as...

We show that design optimizations, an integral but time-consuming component of optical engineering, can be significantly sped-up when paired with deep neural networks (DNNs). By using the DNN indirectly for choosing initializations and candidate preselection, our approach obviates need large networks, big datasets, long training epochs, excessive hyperparameter optimization. For a 16-layered thin-film problem, surrogate-assisted differential evolution (DE) algorithm is able to achieve...

Constructing an artificial assembly for efficient photocatalytic water splitting is key in the pursuit a solar-driven renewable energy economy. Here, we have fabricated covalently linked gold nanoprism–cobalt molecular catalyst construct that stimulated photoelectrocatalytic oxidation reaction. This generated significant photocurrent (∼50 μA/cm2) neutral aqueous conditions with minimal onset overpotential (∼250 mV). dyad imitates light-harvesting properties of natural photosystem-II by...

Periodic structures consisting of alternating layers positive-index and negative-index materials have a novel bandgap at the frequency which average refractive index is zero. We show that, in presence Kerr nonlinearity, this zero-n gap can switch from low transmission to perfectly transmitting state, forming nonlinear resonance or soliton process. This omnidirectional, contrast usual Bragg periodic structures.

Harvesting the full bandwidth of solar spectrum, especially near-infrared portion, remains a challenge for solar-to-fuel conversion technology. Plasmonic nanostructures have recently attracted attention in this connection due to their enhanced yet tunable broadband absorption and photochemical stability. Here we report nanoplasmonic photocatalytic construct by decorating plasmonic Au nanorods with CoO harvesting visible NIR light via photoelectrochemical water oxidation reaction (WOR). In...

The strong dispersion, ultra-thin form-factor and robustness to degradation make metasurfaces attractive for color filter applications. In particular, transmission-mode filters using silicon could potentially replace conventional arrays in backside-illuminated CMOS image sensors enable novel multispectral sensors. We report a robust inverse-design methodology polygon-shaped, particle void, meta-atoms. predict that metasurface primary (RGB) designed with this approach exhibit enhanced gamut,...

Silicon nitride (Si 3 N 4 ) subwavelength medium contrast gratings (MCGs) directly integrated with CMOS photodetectors are a promising option for on-chip label-free biosensing. The narrow spectral features required sensing often realized in Si nanostructures by weakly corrugated which limit design flexibility. We numerically investigate the optical properties of asymmetry-engineered MCG and predict formation ultra-sharp via excitation quasi-bound states continuum (QBIC) resonances....