Near-infrared lensless holographic microscopy on a standard sensor for label-free mesoscale imaging in multiple scattering

Lensless in-line digital holographic microscopy (LDHM) uniquely offers mesoscale, label-free recovery of amplitude and phase in weakly scattering samples, yet it quickly fails in thick and turbid specimens due to multiple scattering and attenuation. We propose near-infrared (NIR) LDHM with operating wavelength range considerably extended up to the silicon-cutoff (~1100 nm) on a standard board-level CMOS sensor. Using tissue-mimicking phantoms and resolution targets, we quantitatively investigate reconstruction performance and show that switching from visible to NIR illumination substantially improves the robustness of conventional LDHM to multiple scattering, with amplitude features remaining resolvable through phantom layers up to ~1.4 mm at 1100 nm compared to only ~350 µm in the visible. Despite a low detector quantum efficiency (0.19% at 1100 nm), robust reconstructions are obtained under low-photon-budget acquisition. We further report that increasing the sample-sensor distance can improve lateral resolution by up to twofold under multiple scattering, consistent with reduced influence of strongly scattered light on hologram formation. Finally, we demonstrate wide-field, label-free, amplitude-phase imaging of thick uncleared mouse tissues, resolving internal structure in brain and liver slices up to ~250 µm and ~60 µm, respectively. These results establish NIR-LDHM as a hardware-minimal detection approach for mesoscale label-free imaging in-and-through complex media, with clear relevance to biophotonics and computational microscopy.