Researchers from Korea University and the Korea Institute of Science and Technology have developed a mixed-dimensional perovskite quantum dot–based synaptic array that utilizes the unique optoelectronic properties of halide perovskites for multitasking (MT) learning. The team fabricated dual-output electroluminescent synaptic devices using Cs₁₋ₓFAₓPbBr₃ (0.00 ≤ x ≤ 0.15) quantum dots, enabling the concurrent processing of postsynaptic current (PSC) and postsynaptic electroluminescence (PSEL) signals within a single device architecture.
Cognitive-inspired MT processing with PSC and PSEL outputs. Image from: Science Advances
Halide perovskites combine high electronic conductivity with strong light emission and compositional tunability, making them suitable for applications requiring coupled electrical–optical responses. In this system, adjusting the FA⁺ concentration and input pulse parameters modulated both the electrical and optical signal ranges, allowing the devices to achieve up to 1,000 accessible synaptic states. The array exhibited stable long-term potentiation/depression, paired-pulse facilitation, and spike-rate-dependent plasticity.
By integrating updates from both PSC and PSEL channels, the MT learning framework could perform classification–regression and classification–image reconstruction tasks simultaneously. Compared to a combined single-tasking architecture, computational speed improved by 47.09 % and 29.17 %, while energy consumption decreased by 8.2× and 32.4× relative to GPU-based accelerators.
The dual-output mechanism leverages the perovskite layer’s ability to support parallel charge transport and electroluminescence, providing a physical basis for efficient co-processing of electrical and optical information. The results indicate that mixed-dimensional Cs–FA–Pb–Br perovskite systems can serve as versatile materials for neuromorphic computing platforms that emphasize low power consumption and multitasking performance.
