Doping-modulated lateral asymmetric Schottky diode as a high-performance self-powered synaptic device.

Area of Science:

• Materials Science
• Computer Engineering
• Neuroscience

Background:

• Conventional computing faces limitations due to increasing data volumes and saturated computational power.
• Neuromorphic computing offers a promising alternative by mimicking the brain's parallel processing for enhanced efficiency.
• A bottleneck exists in energy consumption and data transmission for advanced computing architectures.

Purpose of the Study:

• To demonstrate a compact, self-powered synaptic device technology for neuromorphic computing.
• To achieve zero-energy consumption in synaptic devices through energy harvesting.
• To tune device behavior for optimized information processing and memory functions.

Main Methods:

• Device fabrication utilizing barrier asymmetry.
• Chemical doping to tailor asymmetry and enable energy harvesting.
• Characterization of photovoltaic properties, including open-circuit voltage (V_OC) and power-conversion efficiency (PCE).

Main Results:

• Demonstrated compact, self-powered synaptic devices with zero-energy consumption.
• Achieved tunable V_OC up to 0.77 V and PCE up to 6% through chemical doping.
• Optimized photovoltaic features enabled outstanding programming weight states for training and memory consolidation.

Conclusions:

• The developed synaptic devices offer a pathway to overcome computational bottlenecks in the post-Moore era.
• Self-powered neuromorphic devices with efficient energy harvesting are feasible.
• This technology facilitates advanced functionalities like stimulus reinforcement and consolidation with zero energy cost.