Single-detector MIR microspectrometer based on a MEMS microshutter array.

Area of Science:

• Optics and Photonics
• Microsystems Engineering
• Spectroscopy

Background:

• Mid-infrared (MIR) spectroscopy is crucial for material analysis but often requires expensive detector arrays.
• Existing MIR spectrometers face limitations in cost and complexity.

Purpose of the Study:

• To develop a cost-effective MIR spectrum reconstruction method.
• To demonstrate a novel spectrometer design utilizing a MEMS microshutter array (MSA) and a single photodetector (PD).

Main Methods:

• A pre-patterned MIR filter's transmittance was modulated using a MEMS microshutter array (MSA) with 4x8 pixels.
• A single-pixel photodetector (PD) recorded optical power for various MSA configurations.
• A deep neural network (DNN) reconstructed the full spectrum from modulated optical powers.

Main Results:

• Successfully reconstructed the full spectrum of a light source from 3 to 5 μm using 10 MSA configurations.
• Achieved a mean squared error of less than 1% between the reconstructed and ground truth spectra.
• Demonstrated a low-cost alternative to expensive MIR photodetector array systems.

Conclusions:

• The proposed method offers a viable, low-cost solution for MIR spectroscopy.
• This approach significantly reduces the cost of MIR spectrometers.
• The DNN-based reconstruction enables high-accuracy spectral analysis with simplified hardware.