Researchers at the International Iberian Nanotechnology Laboratory (INL) have developed a neuromorphic photonic neuron that processes optical information through self-sustaining oscillations. The technology is based on controlling negative differential resistance (NDR) in a micro-pillar quantum resonant tunneling diode (RTD) using light. This breakthrough could lead to more efficient and compact neuromorphic computing systems operating with photonic signals.
Biological neurons utilize rhythmic activity for information processing, and traditional neuromorphic systems replicate these processes using electrical, mechanical, or thermal signals. However, photonic technologies offer higher speed, energy efficiency, and compactness. Unlike previous designs that required additional circuitry and increased power consumption, the new device integrates sensory perception and oscillatory behavior into a single semiconductor component based on III-V materials.
The developed RTD photodetectors, published in Scientific Reports, function as artificial oscillatory neurons activated by near-infrared light. These devices, made from n-type gallium arsenide, have diameters of 6–10 micrometers and incorporate double-barrier quantum wells that enable resonant tunneling.
Experiments demonstrated that these devices exhibit characteristic current-voltage behavior with an NDR region under illumination and can respond to pulse-modulated light. This advancement paves the way for the development of more powerful and energy-efficient neuromorphic computing systems based on photonic technologies.
#space #science #educational #technology
