Defect-engineered 2D Bi₂Se₃-based broadband optoelectronic synapses with ultralow energy consumption for neuromorphic...

Area of Science:

• Materials Science
• Condensed Matter Physics
• Neuroscience

Background:

• Optoelectronic synapses (OES) are crucial for overcoming traditional computing limitations.
• Bismuth selenide (Bi2Se3) is a promising material for OES due to tunable selenium vacancies.
• Selenium vacancies act as charge traps, enabling persistent photoconductivity for synaptic function.

Purpose of the Study:

• To develop and characterize a Bi2Se3-based optoelectronic synapse.
• To investigate the role of selenium vacancies in synaptic device operation.
• To evaluate the device's performance in artificial neural network simulations.

Main Methods:

• Fabrication of an ultrathin Bi2Se3 layer using chemical vapor deposition.
• Experimental characterization of synaptic behaviors (PPF, memory, learning).
• Density functional theory (DFT) simulations to understand defect mechanisms.

Main Results:

• Demonstrated key synaptic characteristics: paired-pulse facilitation (PPF), short-term and long-term memory, learning-relearning.
• Achieved a high PPF index (201.7%), long memory retention (523.1 s), and ultralow energy consumption (9.2 fJ/spike).
• Simulations confirmed selenium vacancies as critical trap centers.

Conclusions:

• Bi2Se3 is a viable material for high-performance optoelectronic synapses.
• The Se-vacancy-induced mechanism enables efficient synaptic function.
• This research advances neuromorphic computing and artificial visual perception.