Hyperspectral quantum-dot image sensors via in-pixel reconfigurable band-alignment

Abstract

High-resolution short-wave infrared hyperspectral imaging enables non-destructive material identification and imaging through scattering media, paving the way for transformative applications in portable diagnostics, precision agriculture, environmental monitoring and space exploration. However, conventional hyperspectral imagers face a compromise between spatial resolution, spectral resolution and device footprint. Here we report a miniaturized hyperspectral image sensor that mitigates this trade-off by leveraging monolithically integrated, bias-reconfigurable stacked colloidal quantum dot junctions and a bias-programmable spectral reconstruction algorithm. By applying a defined sequence of single-polarity increasing bias voltages, the interfacial band alignment can be tuned, thus mediating the collection of photon-generated carriers in colloidal quantum dot layers with different energy gap. Our imager achieves spatial resolution of 1,280 × 1,024, spectral resolution of 1 nm, reconstruction accuracy of 0.055 nm, peak detectivity above 10¹³ jones and broadband coverage (400–1,700 nm), all within a compact pixel footprint of 15 × 15 µm². The high signal-to-noise ratio and spatial resolution result in accurate reconstruction of hyperspectral image information, enabling food quality monitoring, chemical solvents discrimination and materials identification. Bias-voltage-tuneable colloidal quantum dots junctions enable short-wave infrared hyperspectral imagers with megapixel spatial resolution and nanometre-scale spectral resolution for food quality monitoring and materials discrimination.