Making short-wave infrared (SWIR) cameras practical at scale requires more than high sensitivity—it also requires uniformity. In an imaging chip, thousands or even millions of pixels must respond consistently. Variations in the light-sensitive film can cause some pixels to appear brighter, become noisier, or fail altogether, degrading image quality even when individual devices perform well.
Among the materials being explored for low-cost SWIR imaging, PbS QDs are particularly attractive because they can be processed from solution, their absorption wavelength can be tuned through size and surface chemistry, and they are compatible with scalable coating and printing techniques. High-performance PbS QD photodetectors typically rely on p–n junctions to separate photogenerated charges. While n-type QD inks are now relatively mature, producing uniform p-type QD films over large areas remains a major challenge.
Conventional fabrication of p-type PbS films relies on solid-state ligand exchange, in which the surface molecules are replaced after the QD film has been deposited. This post-deposition process is often incomplete and spatially nonuniform, while the accompanying volume contraction can induce cracks and electrical inhomogeneity across large-area films, limiting imaging performance.
A research team has now developed a directly synthesized p-type PbS QD ink to overcome this bottleneck. By tuning the functional groups on the thiol ligands, they controlled both the ligand-surface interaction and the p-type electronic properties of the QDs. Combined with a seeded-growth strategy, this approach also enabled broad spectral tunability, extending the first excitonic absorption peak from approximately 1100 nm to beyond 2000 nm.
The most significant improvement was achieved in large-area film uniformity. The directly synthesized ink formed dense, crack-free p-type films with substantially reduced thickness variation across 6-inch silicon wafers. Whereas films prepared by conventional solid-state ligand exchange exhibited a thickness variation of approximately ±7.82 nm, the directly synthesized films varied by only ±0.31 nm. Surface-potential mapping and sheet-resistance measurements likewise revealed much more homogeneous electrical properties, indicating that the direct-ink approach provides a highly uniform material platform for wafer-level SWIR devices.
The researchers then evaluated the material in SWIR photodetectors and prototype imaging modules. At the single-device level, the ink-based photodetectors performed comparably to conventional devices while exhibiting lower noise in some measurements. More importantly, the improved film uniformity translated directly into better imaging performance. Photoresponse non-uniformity decreased from 10.57% to 5.79%, and the fraction of dead pixels was reduced from 0.21% to 0.03%.
To demonstrate the practical advantages of the technology, the team combined the prototype SWIR imager with an object-recognition algorithm for vehicle detection. Under clear conditions, both visible-light and SWIR images enabled reliable recognition. Under foggy conditions, however, visible-light images lost contrast and missed some vehicles, whereas the SWIR images preserved clearer scene information and achieved higher recall and F1 scores.
These results demonstrate that directly synthesized p-type PbS QD inks provide a scalable route toward highly uniform wafer-level SWIR imaging. The approach could facilitate the development of lower-cost infrared cameras for traffic monitoring, autonomous driving, machine vision, environmental sensing, and other applications that require large-area imaging with high pixel-to-pixel consistency.
National Science Review
Experimental study