All-Inorganic CsPbBr3 Perovskite Planar-Type Memristors as Optoelectronic Synapses
Citations Over TimeTop 10% of 2024 papers
Abstract
Mimicking fundamental synaptic working principles with memristors contributes an essential step toward constructing brain-inspired, high-efficiency neuromorphic systems that surpass von Neumann system computers. Here, an electroforming-free planar-type memristor based on a CsPbBr3 single crystal is proposed and exhibits excellent resistive switching (RS) behaviors including stable endurance, ultralow power consumption, and fast switching speed. Furthermore, an optically tunable RS performance is demonstrated by manipulating irradiation intensity and wavelength. Optical analysis techniques such as steady-state photoluminescence and time-resolved photoluminescence are employed to investigate the distribution of Br ions and vacancies before and after quantitative polarization, describing migration dynamic processes to elucidate the RS mechanism. Importantly, a CsPbBr3 single crystal, as the optoelectronic synapse, shows unique potential to emulate photoenhanced synaptic functions such as excitatory postsynaptic current, paired-pulse facilitation, long-term potentiation/depression, spike-timing-dependent plasticity, spike-voltage-dependent plasticity, and learning-forgetting-relearning process with ultralow per synapse event energy consumption. A classical Pavlov's dog experiment is simulated with a combination of optical and electrical stimulation. Finally, pattern recognition with simulated artificial neural networks based on our synapse reached an accuracy of 93.11%. The special strategy and superior RS characteristics of optoelectronic synapses provide a pathway toward high-performance, energy-efficient neuromorphic electronics.
Related Papers
- → Generalised mathematical model of memristor(2016)31 cited
- → Area of Hysteresis Loop of a Generic Memristor Emulator(2019)4 cited
- Research progress of the memristor and its application foreground(2010)
- → Memristor devices and memristor-based circuits(2020)
- → Harnessing memristor circuits and device variability in emergent computing applications(2022)