Integrated Microring Spectrometer with In‐Hardware Compressed Sensing to Break the Resolution‐Bandwidth Limit for General Continuous Spectrum Analysis

Abstract

Microspectrometers have numerous applications in mobile optical sensing due to their dramatic advantages of compact size, light weight, and low power consumption. Reconstructive spectrometers, based on computational algorithms, have garnered considerable interest as they exhibit superior resolution or spectral bandwidth. However, existing reconstructive spectrometer designs face challenges in spectral applicability, algorithm robustness, and the resolution‐bandwidth limit. Here, a reconstructive spectrometer that utilizes a microring resonator (MRR) is proposed to achieve compressed sensing in hardware by decomposing an arbitrary continuous spectrum into a series of simple comb spectra. Owing to the presparse operation of the MRR, one needs to construct the comb spectra with a known line shape, only leaving the amplitude to be solved. Consequent random gratings measure the comb spectra, which are reconstructed with high robustness. Thanks to the independent engineering of spectral resolution and bandwidth, the approach breaks the traditional resolution‐bandwidth limit. A narrowband signal of a dual peak at 0.2 nm and a broadband spectrum with a large spectral bandwidth of 60 nm and 300 spectral channels using only eight physical channels is retrieved, achieving an ultrahigh reconstructive compression ratio of 37.5. This work opens up the possibility of on‐site spectral spectroscopy applications for the lab‐on‐a‐chip system.

Ultra-compact monolithically-integrated silicon single-microring spectrometer breaking the free-spectral-range limitation

“…More recently, the reconstructive spectrometer paradigm has emerged and received great attention 5,[19][20][21][22][23][24][25][26][27][28][29][30] , in which case diverse spectral signals can be retrieved computationally by pre-calibrating information encoded with orthogonal spectral responses. For reconstructive spectrometers, the performance depends on the diversity of calibration spectrum, and thus more accurate spectrum reconstruction is enabled when more spectrum channels are available 22,24 .…”