Large-Aperture Inverse-Designed Sparse Meta-Optics in the Mid-Wave Infrared

Abstract Long-range midwave infrared (MWIR) imaging demands large apertures for high angular resolution, yet scaling conventional flat optics to the centimeter scale severely throttles multiwavelength design tractability and large-scale physical realization. While flat sparse meta-optics systems have recently emerged to mitigate the footprint of monochromatic large-scale optics, their functional scaling remains fundamentally bounded by single-wavelength operational limits and staggering computational costs under full-scale global optimization. Here, we shatter these long-standing boundaries by demonstrating an inverse-designed sparse meta-optics paradigm that transcends traditional monochromatic constraints to achieve multiband and broad-spectrum MWIR imaging. By implementing a reduced-dimensional optimization framework, we compress multiwavelength phase design complexity by orders of magnitude, mapping optimal radial phase distributions directly onto a prescribed sparse-aperture geometry. The sparse-aperture configuration is selected for its straightforward layout, mechanical robustness, and inherently compact point spread function (PSF), thereby isolating the phase optimization from the macro-scale aperture geometry. The resulting algorithmic tractability, combined with maskless, large-area laser direct writing (LDW) lithography, enables the physical realization of complex wavefront control over large areas while easing the fabrication burden. Using the integrated design-to-fabrication framework, two 89 mm-diameter all-silicon sparse metalenses are fabricated and experimentally validated, demonstrating enhanced performance across multiple wavelengths, varying object distances, and diverse sensing scenarios. Combined with a complementary backend reconstruction algorithm to mitigate inherent PSF artifacts, this work establishes a comprehensive computational design and scalable laser manufacturing platform for next-generation, large-aperture sparse meta-optics in thermal imaging systems.