High-resolution on-chip spatial heterodyne Fourier transform spectrometer based on artificial neural network and...

Area of Science:

• Photonics
• Spectroscopy
• Integrated Optics

Background:

• Fourier Transform (FT) spectroscopy is crucial for spectral analysis.
• On-chip spectrometers offer miniaturization and portability advantages.
• Silicon-on-insulator (SOI) platforms enable advanced photonic integrated circuits.

Purpose of the Study:

• To propose and demonstrate a novel compact on-chip Fourier transform spectrometer.
• To achieve wide operating bandwidth and high spectral resolution.
• To develop and validate spectral retrieval algorithms for accurate signal reconstruction.

Main Methods:

• Design of a spectrometer utilizing a 16-channel power splitter and a Mach-Zehnder interferometer (MZI) array.
• Incorporation of MZIs with linearly increasing optical path length (OPL) differences.
• Development of a spectral retrieval algorithm combining pattern-coupled sparse Bayesian learning (PCSBL) and artificial neural networks (ANN).

Main Results:

• Demonstration of a flat transmission characteristic over a 100 nm bandwidth (1500-1600 nm).
• Achieved high resolution, capable of reconstructing narrowband signals with a 0.5 nm full width at half maximum (FWHM).
• Successfully reconstructed a triple-peaked signal with 3 nm separation using the developed algorithm.

Conclusions:

• The proposed on-chip FT spectrometer on SOI platform achieves wide bandwidth and high resolution.
• The integrated PCSBL and ANN spectral retrieval algorithm enhances signal reconstruction accuracy.
• This compact device shows significant potential for various spectroscopic applications.