Design of All-Optical Bidirectional Self-Powered Synaptic Devices for Neuromorphic Computing Applications

Area of Science:

• Optoelectronics
• Materials Science
• Neuroscience

Background:

• Self-powered optoelectronic synaptic devices offer low energy consumption but lack clear mechanisms for inhibitory behaviors.
• Mimicking optically inhibitory behaviors in these devices is a significant challenge for neuromorphic computing.

Purpose of the Study:

• To fabricate perovskite-based synaptic devices with self-powered optical potentiation and electrical inhibition.
• To elucidate the mechanisms behind inhibitory behaviors in these devices.
• To engineer devices exhibiting optical inhibition for neuromorphic applications.

Main Methods:

• Fabrication of formamidinium lead iodide perovskite synaptic devices.
• Investigation of defect traps and iodine ion migration for inhibitory mechanisms.
• Incorporation of ethanediamine dihydroiodide (EDADI) to modulate synaptic behavior.
• First-principles calculations to understand the role of iodide vacancy formation energies.

Main Results:

• Devices demonstrated self-powered optical potentiation and electrical inhibition.
• Mechanisms identified as defect traps (room temp) and iodine ion migration (low temp).
• EDADI addition successfully shifted behavior from optical potentiation to optical inhibition.
• First-principles calculations supported the role of iodide vacancies in this transformation.
• Achieved high recognition rates (97.95% for MNIST, 77.36% for CIFAR-10) using the engineered devices.

Conclusions:

• The study clarifies the mechanisms of self-powered optically inhibitory synaptic behaviors.
• EDADI incorporation is a viable strategy to achieve optical inhibition in perovskite synapses.
• This research advances the development of all-optical, bidirectional, self-powered neuromorphic computing systems.