Fluorescence Spectral Imaging Based on Computational Spectral Sensing

Abstract

As a powerful optical tool, spectrometers are widely used in materials analysis and light-source characterization. Although commercial spectrometers based on dispersive diffraction or Fourier transform offer high resolution and broad dynamic range, spatial information of the sample is lost, limiting its application in snapshot hyperspectral imaging. Herein, we propose a computational spectrometer based on the tailored point spread function of the imaging system that uses disordered microstructures fabricated by low-cost photolithography. The spectrometer achieves a high quantum efficiency (${Q}_{\mathrm{avg}}\phantom{\rule{0.2em}{0ex}}\ensuremath{\sim}\phantom{\rule{0.2em}{0ex}}60\mathrm{%}$) and ultracompact footprint as low as $24.5\ifmmode\times\else\texttimes\fi{}24.5\phantom{\rule{0.2em}{0ex}}\ensuremath{\mu}{\mathrm{m}}^{2}$, exhibits outstanding performance in weak-light hyperspectral imaging (e.g., Raman or fluorescence spectroscopy) in the microscopic system. Apart from the advancements mentioned above, our spectrometer offers good compatibility with commercial microscopes, implying it is a promising candidate for spectral imaging in chemistry, biology, and others.