Infrared image sensors are commonly assembled by manufacturing a detector array separately and bonding it to a readout integrated circuit. Researchers have now demonstrated a different route: growing the light-sensitive semiconductor directly on the readout chip using a room-temperature water-based process.
The study, published in Nature, reports monolithic in situ integration of lead sulfide, or PbS, photodetectors on readout integrated circuits. The team produced quasi-single-crystalline films with wafer-scale pixel uniformity, fabricated a 640 × 512-pixel focal-plane array and captured clear infrared images.
Growing a detector where it will operate
Direct growth could remove several alignment and bonding steps from conventional hybrid sensor manufacturing. The challenge is that a finished readout circuit cannot tolerate many of the high temperatures or aggressive processing conditions used to make electronic-grade semiconductor films.
The researchers drew on biomineralization principles and controlled lead–hydroxide complexation in an aqueous chemical bath. By regulating the reactive species and suppressing environmental disturbances, they encouraged PbS to crystallize preferentially at the chip surface rather than forming particles throughout the solution.
The resulting films had a reported carrier density of approximately 2 × 1015 per cubic centimetre and carrier mobility of 124 square centimetres per volt-second. Measurements across the wafer showed broadly uniform thickness and crystallographic properties at the tested positions.
Photodiodes and a complete focal-plane array
Antimony doping enabled the researchers to tune the electrical properties of the PbS films. The resulting photodiodes reached a reported specific detectivity of 4.4 × 1011 Jones at a wavelength of 2.5 micrometres. In the study’s comparison, that value exceeded the performance of commercial indium gallium arsenide detectors at that wavelength.
The team then integrated the detectors into a 640 × 512-pixel focal-plane array. The array generated recognizable infrared images, showing that the process could support a complete imaging system rather than only individual laboratory devices.
Devices were also grown directly on flexible readout circuits. This demonstration suggests a possible route toward curved or wearable infrared sensors, although the study does not establish the durability or field performance of such systems.
Potential manufacturing advantages
A solution-based process operating at room temperature could simplify fabrication and reduce the thermal budget for integrating infrared-sensitive materials with silicon electronics. It may also support larger areas or unconventional substrates that are difficult to address through conventional bonding.
However, scaling from research devices to manufacturing will require close control of bath chemistry, vibration, temperature, contamination and pixel-to-pixel reproducibility. The use of lead also requires environmental, occupational and end-of-life safeguards. Performance must ultimately be evaluated under application-specific conditions, including noise, temperature cycling, long-term stability and yield.
Extending monolithic sensor integration
The researchers also reported a proof-of-concept extension of the growth strategy from PbS to lead selenide. That result suggests the chemistry may be adaptable to other infrared materials rather than being limited to one compound.
The central advance is therefore both a device and a manufacturing concept: instead of attaching a finished detector layer to a circuit, the detector can be crystallized directly on the electronics under conditions the circuit can tolerate. If reliability and scale can be demonstrated, the approach could influence infrared cameras used in machine vision, industrial inspection, scientific instrumentation and flexible sensing.