Neural Network-Based On-Chip Spectroscopy Using a Scalable Plasmonic Encoder

Abstract

Conventional spectrometers are limited by trade-offs set by size, cost, signal-to-noise ratio (SNR), and spectral resolution. Here, we demonstrate a deep learning-based spectral reconstruction framework, using a compact and low-cost on-chip sensing scheme that is not constrained by the design trade-offs inherent to grating-based spectroscopy. The system employs a plasmonic spectral encoder chip containing 252 different tiles of nanohole arrays fabricated using a scalable and low-cost imprint lithography method, where each tile has a unique geometry and, thus, a unique optical transmission spectrum. The illumination spectrum of interest directly impinges upon the plasmonic encoder, and a CMOS image sensor captures the transmitted light, without any lenses, gratings, or other optical components in between, making the entire hardware highly compact, light-weight and field-portable. A trained neural network then reconstructs the unknown spectrum using the transmitted intensity information from the spectral encoder in a feed-forward and non-iterative manner. Benefiting from the parallelization of neural networks, the average inference time per spectrum is ~28 µs, which is orders of magnitude faster compared to other computational spectroscopy approaches. When blindly tested on unseen new spectra (N = 14,648) with varying complexity, our deep-learning based system identified 96.86% of the spectral peaks with an average peak localization error, bandwidth error, and height error of 0.19 nm, 0.18 nm, and 7.60%, respectively. This system is also highly tolerant to fabrication defects that may arise during the imprint lithography process, which further makes it ideal for applications that demand cost-effective, field-portable and sensitive high-resolution spectroscopy tools.

Single-Detector Spectrometer Using a Superconducting Nanowire

“…These errors resulted primarily from the artifacts of the reconstruction method, since the noise is typically amplified in ill-posed inverse problems. Future improvement of the accuracy can be made by using more robust spectrum-specific recovery algorithms, such as applying neural networks − that have been demonstrated as a better performance than traditional regularized inversions.…”

Mie Sensing with Neural Networks: Recognition of Nano‐Object Parameters, the Invisibility Point, and Restricted Models

“…In particular, the ML ideology was successfully applied in several experimental reports on photonic metamaterials: refs. [19,22,24,31,61,62]. One of the goals of this paper is to motivate experimental work in electromagnetic Mie sensing.…”

Towards ‘Fourth Paradigm’ Spectral Sensing

“…Looking at Figure 2 , one may also question why we would bother forming the reconstruction problem as Equation ( 4 ) in lieu of training a neural network on K , thereby learning a map between the measured and target dimensions. We refrain from taking such an approach for two reasons: first, it is well known that “black box” models do not generalize well and this has, in fact, been demonstrated to be the case in the application of recovering visible spectral distributions from encoder-array spectrometers [ 36 , 37 ]; second, we lose any interpretability or contextual reference of the scene. While Equation ( 4 ) may seem trivially simple, enforcing sparsity via the -norm has demonstrated incredible success in not only reconstructing low-dimensional measurements, but also in reconstructing them in a way that mimics the underlying physical system [ 38 ].…”