Stress-engineered ultra-broadband spectrometers

Abstract

The evolution of miniaturized spectroscopic tools is pivotal for expanding the application of spectral data across scientific, industrial, and consumer domains. Recent advancements in computationally augmented systems have dramatically reduced device form factors toward those compatible with consumer tech integration. However, for a commercial reality, most applications demand operation across visible to short-wave infrared (SWIR) range. In this regard, existing miniaturized devices are either constrained by physical properties; use complex, costly, or unscalable fabrication techniques; or require multiple components to address separate parts of the spectrum. Here, we report on a low-cost, visible to SWIR, miniaturized spectrometer design enabled by a mass-producible, nonlithographic method of engineering planar dispersive elements from widely available plastics. By deforming shape memory epoxies, we encode spectral information, which is processed by a complementary metal oxide semiconductor sensor array and reconstructed via algorithms. This design offers broadband capability from 400 to 1600 nanometers and enables line-scanning spectral imaging, paving the way for affordable spectrometers.

Fig. 1.. Application and characterization of plastics.

(A and B) Illustration of plastic replacing glass in small camera and spectrometer modules, respectively. (C) Birefringence of plastic products in the laboratory (scale bar, 1 cm). (D) Illustration of birefringence effect. (E) The introduction of stress into the network. (F) Glass transition curves of epoxy film on the DMA. (G) DMA curves of strain, stress, and temperature during the internal stress programming. (H) The stress retention of epoxy compared to the stress relaxation of PMMA. a.u., arbitrary units.

Fig. 2.. Design of stress and dispersion engineering in epoxy films.

(A) Simulation of rectangular films stretching. (B) Colors of rectangular films at different strains (scale bar, 2 mm). (C) Transmission spectra corresponding to different strains in rectangular films. (D) Simulation of triangle film stretching [same color bar as (A)]. (E) Stress-engineered color obtained by stretching a triangular sample (scale bar, 5 mm). (F) Transmission spectra along the stretching direction [white arrow in (E)] in triangle films. (G) Simulation of large-scale film stretching [same color bar as (A)]. (H) Picture of the large-scale film under polarized light (scale bar, 2 cm). (I) The variation curves of the main-peak wavelengths in the transmission spectra in the stretching direction in different locations and samples [corresponding to L1–L5 in (H)]. (J) Schematic diagram of the designed system for producing stress-engineered filters.

Fig. 3.. Schematic diagram, reconstruction, and spectral imaging results of the spectrometers.

(A) Picture of miniaturized spectrometer fabricated on the basis of CMOS chip. (B) Schematic and operation of the spectrometer. The CMOS sensor is Sony IMX178 with 3072 by 2048 pixels. (C) The response functions obtained from calibration. (D) Reconstruction of 19 monochromatic spectra from 400 to 800 nm compared with measured spectrum. (E) Reconstruction of a double-peak spectrum separated by 10 nm compared with the measured spectrum. (F) Reconstruction of a halogen lamp spectrum compared with the measured spectrum. (G) Spectral imaging results range from 460 to 620 nm (artificially colored). (H) Images captured by RGB camera and synthetic RGB image reconstructed from spectral data (scale bar, 500 μm). (I) Reconstruction of the red, green, and blue pixels compared with measured spectra.

Fig. 4.. Reconstruction performance of the ultra-broadband spectrometers.

(A) Reconstruction of monochromatic spectra with different wavelengths varying from 400 to 1600 nm compared with measured spectra using commercial spectrometers. (B) Reconstruction of double-peak spectra at (488 and 498 nm), (979 and 994 nm), (1310 and 1330 nm), and a broadband tungsten light source spectrum compared with measured spectra, respectively.