# Silicon Photonic On-Chip Spectrometer Based on Cascaded Mach–Zehnder Interferometer

## TL;DR

This paper introduces a compact, high-precision on-chip spectrometer using cascaded Mach–Zehnder interferometers for efficient spectral analysis.

## Contribution

A reconfigurable silicon photonic on-chip spectrometer with cascaded MZIs and a novel reconstruction algorithm is proposed and implemented.

## Key findings

- A five-stage MZI design achieves a spectral reconstruction resolution of 5 pm over 1500–1600 nm.
- The spectrometer maintains low reconstruction errors for various spectral profiles.
- The device offers a compact, low-power solution for high-precision on-chip spectral analysis.

## Abstract

Spectrometers are essential tools for revealing the interaction between light and matter and analyzing the composition and state of materials, widely employed in scientific research, industrial inspection, and biomedicine applications. With the continuous expansion of application scenarios, higher demands are placed on the miniaturization, integration, and portability of spectrometers. This paper proposes and implements a reconfigurable silicon photonic on-chip spectrometer based on cascaded multi-stage Mach–Zehnder interferometers (MZIs). This structure achieves efficient sampling of the input spectrum by applying adjustable phase shifts to each MZI stage to construct different spectral responses. Combined with a convex optimization algorithm incorporating differential operators, the unknown input signals are decomposed into sparse and smooth components, achieving high-accuracy reconstruction. Experimental results show that the proposed five-stage MZI design with a total of 216 sampling channels achieves a spectral reconstruction resolution of 5 pm over the wavelength range from 1500 nm to 1600 nm. Moreover, the spectrometer exhibits consistently low reconstruction errors for broadband spectra, sparse spectra, and their hybrid spectral profiles. This research demonstrates excellent comprehensive performances in device structure design, phase modulation strategy, and reconstruction algorithm, providing an effective solution for realizing low-power, small-footprint, and high-precision on-chip spectral analysis.