A Reconfigurable and Nonvolatile Optoelectronic Memtransistor with Ultralong Retention Time Based on In₂Se₃/Te van...

Area of Science:

• Materials Science
• Nanotechnology
• Neuroscience

Background:

• Memtransistors offer reconfigurable, nonvolatile properties crucial for simulating neural synapses.
• Conventional von Neumann architecture faces challenges that neuromorphic computing aims to overcome.

Purpose of the Study:

• To fabricate and characterize an indium selenide/tellurene van der Waals p-n junction memtransistor.
• To demonstrate its artificial synaptic properties modulated by dual electric and optoelectronic stimuli.
• To explore its potential in neuromorphic computing and artificial intelligence.

Main Methods:

• Fabrication of an indium selenide (In2Se3)/tellurene (Te) van der Waals p-n junction memtransistor.
• Modulation of synaptic weights using electric and optoelectronic dual stimuli.
• Implementation of logic operations and a Pavlov's dog experiment.
• Performance evaluation using Modified National Institute of Standards and Technology image recognition with an artificial neural network.

Main Results:

• The memtransistor exhibited significant artificial synaptic properties with nonvolatile retention exceeding 10^5 s.
• Dual modulation modes enabled basic logic operations and a Pavlov's dog experiment.
• Achieved 90.15% accuracy in image recognition, maintaining 70.38% under multilevel noise.

Conclusions:

• The In2Se3/Te memtransistor demonstrates excellent artificial synaptic behavior and dual-stimuli response.
• The device shows high potential for developing brain-inspired computing systems and neuromorphic applications.
• These findings pave the way for advanced AI hardware.