Traditional photodetectors rely on fixed designs: responsivity, dynamic range and spectral selectivity are set during fabrication. Silicon, the dominant material, has a relatively fixed bandgap and limited carrier mobility, making broadband, high-speed and multifunctional detection difficult. Solution-processed metal halide perovskites offer a compelling alternative because their composition and dimensionality can be tuned across ultraviolet (UV), visible and near-infrared (NIR) wavelengths, while low-temperature printing enables flexible and large-area arrays. However, long-term stability, lead toxicity, large-area uniformity and the seamless integration of memory, logic and learning remain unresolved. Based on these challenges, or because of these issues, in-depth research is needed on stable, scalable and intelligent perovskite photodetectors that combine sensing, memory and computing.
A review led by Jilin University, with collaborators from Universiti Teknologi Malaysia, CCS University and other international institutions, was accepted on 1 April 2026 and published (DOI: 10.1002/cey2.70273) in Carbon Energy. The article examines perovskite photodetectors at the intelligence frontier, from tunable materials to adaptive optoelectronic systems. It covers material stability, device architectures, multifunctional arrays, neuromorphic imaging, optical communication and embedded intelligence, and outlines a practical path toward scalable, manufacturable intelligent sensors.
The review details how perovskite photodetectors now achieve ultralow dark currents below 10⁻¹¹ A, responsivities above 5 A W⁻¹, detectivities up to 10¹⁴ Jones and on/off ratios near 10⁴. Solution-based monolithic integration allows wafer-scale or freeform arrays on flexible substrates, enabling multispectral detection, polarization sensitivity and memory in one platform. The authors highlight narrowband, dual-band and bipolar responses for wavelength-selective encryption and anti-interference communication; flexible and hemispherical arrays for wide-angle, lens-free color and motion-tracking imaging; and perovskite thin-film transistors (TFTs) and optoelectronic synapses for neuromorphic computing, including excitatory postsynaptic current (EPSC), paired-pulse facilitation (PPF), image preprocessing and all-optical write/read. A central proposal is “spatial decoupling”: keeping a high-quality absorber for low-noise sensing while confining reversible ion migration and charge trapping to engineered interface layers for reliable memory and learning. The review also surveys high-speed, underwater and encrypted optical communication, bifunctional light-emitting/detecting units, and barriers such as lead toxicity, large-area uniformity, long-term reliability and artificial intelligence (AI) co-design.
The authors said the field is shifting from simply boosting responsivity to embedding useful functions at the sensor itself. They said stable materials, scalable printing and precise interface engineering will decide whether perovskite photodetectors can move from laboratory demonstrations to commercial systems. They said the biggest opportunity is co-designing materials, devices and algorithms so that sensing, memory and simple decisions happen together at the edge, reducing latency and energy use.
If realized, these devices could support intelligent imaging for autonomous vehicles, drones, robots and security cameras; compact spectrometers and medical endoscopy; wearable health monitors; underwater and free-space optical links; and low-power neuromorphic edge computing. Perovskite arrays might also enable lens-free color cameras, wide-angle artificial retinas and in-sensor preprocessing for the Internet of Things (IoT). Real-world adoption will require robust encapsulation, lead-free or low-toxicity formulations, standardized manufacturing, data security and lifecycle assessment. The review argues that combining tunable perovskites with embedded intelligence could turn photodetectors from passive components into adaptive nodes in future optoelectronic infrastructures.
###
References
DOI
10.1002/cey2.70273
Original Source URL
https://doi.org/10.1002/cey2.70273
Funding information
National Natural Science Foundation of China (52173166, 52573274), Fundamental Research Funds for Central Universities, Council of Scientific and Industrial Research–Senior Research Fellowship (09/0113(13402)/2022-EMR-I), Universiti Teknologi Malaysia (AJ090000.6700.09453), JTNCPI and CST/D-1524.
About Carbon Energy
Carbon Energy is an open access energy technology journal publishing innovative interdisciplinary clean energy research from around the world. The journal welcomes contributions detailing cutting-edge energy technology involving carbon utilization and carbon emission control, such as energy storage, photocatalysis, electrocatalysis, photoelectrocatalysis and thermocatalysis.