Optoelectronic synapses with chemical-electric behaviors in gallium nitride semiconductors for biorealistic neuromorphic functionality

Xin Liu, Danhao Wang, Wei Chen, Yang Kang, Shi Fang, Yuanmin Luo, Dongyang Luo, Huabin Yu, Haochen Zhang, Kun Liang, Lan Fu, Boon S. Ooi, Sheng Liu, Haiding SunNature Communications 2024 · DOI

TL;DR

Researchers have developed a "biorealistic" optoelectronic synapse using gallium nitride (GaN) nanowires that operates in a liquid electrolyte. Unlike traditional solid-state artificial synapses, this device mimics the chemical-electric nature of biological brains, allowing it to switch between short-term and long-term memory and respond to its chemical environment—even emulating complex phenomena like oxidative stress-induced cognitive decline.

What Problem Does This Paper Address?

Most current optoelectronic synapses—the building blocks for future artificial vision and light-based AI—are purely solid-state. They rely on physical processes like charge trapping in semiconductors to mimic synaptic plasticity.

However, biological synapses are far more complex; they are chemical-electric systems. They function within physiological electrolytes where neurotransmitters and ions dictate signal transmission. Because traditional artificial synapses lack this chemical dimension, they struggle to achieve "biorealistic" functionality, such as responding to their chemical surroundings or emulating how biological factors (like aging or toxins) affect learning and memory. This paper bridges that gap by introducing a photoelectrochemical (PEC) architecture.

Methods

The team fabricated vertically aligned p-AlGaN/n-GaN heterojunction nanowires on a silicon substrate using plasma-assisted molecular beam epitaxy (MBE).

The novel "secret sauce" is the device's operating environment:

  • Photoelectrochemical (PEC) Setup: The nanowires act as a photoelectrode immersed in an electrolyte (typically Phosphate Buffered Saline, or PBS).
  • Dual-Modal Modulation: The researchers found they could switch the device's behavior simply by changing the circuit state. A closed-circuit (current mode) produces short-term plasticity (STP), while an open-circuit (voltage mode) produces long-term plasticity (LTP).
  • Chemical Tuning: To mimic biological receptors, they decorated the nanowire surfaces with Platinum (Pt) nanoparticles. They also varied the electrolyte's pH, ion concentration, and introduced bioactive molecules like ascorbic acid and hydrogen peroxide ($H_2O_2$) to see how the "synapse" reacted.

Datasets

As this is a materials science and device physics paper, it does not use standard machine learning datasets. Instead, the authors generated original experimental data through:

  • Structural Characterization: SEM, TEM, and STEM imaging.
  • Electrochemical Measurements: Recording Post-Synaptic Current ($\Delta$PSC) and Post-Synaptic Voltage ($\Delta$PSV) under various light and chemical stimuli.
  • Simulated Visual Tasks: A "U" shaped image was used to demonstrate how chemical environments affect image retention and contrast.

Key Insights and Ideas

  • The Interface is the Key: The synaptic behavior isn't just happening inside the semiconductor; it’s driven by the interplay between internal charge transport and external redox reactions at the nanowire/liquid interface.
  • Circuit State as a Memory Switch: One of the most elegant findings is that memory duration can be toggled without changing the material. In current mode, holes dissipate quickly (STP). In voltage mode, the open circuit prevents easy charge dissipation, leading to much longer retention (LTP).
  • Artificial Receptors: The Pt nanoparticles act like transmembrane receptors in a cell. They facilitate faster electron transfer to the electrolyte, which significantly boosts the synaptic response.
  • Chemical Sensitivity: Because the device is electrochemical, its "learning" ability is directly affected by its environment. This allows it to sense and adapt to chemical changes, much like a biological brain responds to neurotransmitters or drugs.

Key Results

  • Dual Plasticity: In current mode, the synaptic response lasted about 60 seconds. In voltage mode, it lasted over 1500 seconds.
  • Paired-Pulse Facilitation (PPF): The device achieved a maximum PPF index of 217% (current mode) and 293% (voltage mode), comparable to biological benchmarks.
  • Biomimetic Emulation: By adding $H_2O_2$ to the electrolyte, the team successfully simulated oxidative stress. The "impaired" synapse showed a lower learning level and required 24 pulses to relearn a task that a "healthy" synapse could relearn in just 17 pulses.
  • Vision Loss Simulation: The researchers demonstrated that under oxidative stress, the device's ability to perceive and remember a visual "U" pattern was significantly degraded, mimicking age-related vision decline.

Limitations and Future Work

  • Scalability: The current study uses a standard electrochemical cell setup. For practical AI applications, these liquid-junction devices need to be integrated into high-density, microfabricated arrays.
  • Power Consumption: While the current is in the nanoampere (nA) range, further optimization is needed to compete with the extreme energy efficiency of the human brain.
  • System Integration: The authors suggest that future work should focus on CMOS-compatible fabrication and developing peripheral readout circuits that can handle both current and voltage signals simultaneously on a single chip.
  • Long-term Stability: While GaN is chemically robust, the long-term stability of the nanowire-electrolyte interface under constant operation requires further investigation.

Summary updated

Abstract

Optoelectronic synapses, leveraging the integration of classic photo-electric effect with synaptic plasticity, are emerging as building blocks for artificial vision and photonic neuromorphic computing. However, the fundamental working principles of most optoelectronic synapses mainly rely on physical behaviors while missing chemical-electric synaptic processes critical for mimicking biorealistic neuromorphic functionality. Herein, we report a photoelectrochemical synaptic device based on p-AlGaN/n-GaN semiconductor nanowires to incorporate chemical-electric synaptic behaviors into optoelectronic synapses, demonstrating unparalleled dual-modal plasticity and chemically-regulated neuromorphic functions through the interplay of internal photo-electric and external electrolyte-mediated chemical-electric processes. Electrical modulation by implementing closed or open-circuit enables switching of optoelectronic synaptic operation between short-term and long-term plasticity. Furthermore, inspired by transmembrane receptors that connect extracellular and intracellular events, synaptic responses can also be effectively amplified by applying chemical modifications to nanowire surfaces, which tune external and internal charge behaviors. Notably, under varied external electrolyte environments (ion/molecule species and concentrations), our device successfully mimics chemically-regulated synaptic activities and emulates intricate oxidative stress-induced biological phenomena. Essentially, we demonstrate that through the nanowire photoelectrochemical synapse configuration, optoelectronic synapses can be incorporated with chemical-electric behaviors to bridge the gap between classic optoelectronic synapses and biological synapses, providing a promising platform for multifunctional neuromorphic applications.
BibTeX
@inproceedings{liu2024optoelectronic,
  author = {Xin Liu and Danhao Wang and Wei Chen and Yang Kang and Shi Fang and Yuanmin Luo and Dongyang Luo and Huabin Yu and Haochen Zhang and Kun Liang and Lan Fu and Boon S. Ooi and Sheng Liu and Haiding Sun},
  title = {{Optoelectronic synapses with chemical-electric behaviors in gallium nitride semiconductors for biorealistic neuromorphic functionality}},
  booktitle = {Nature Communications},
  year = {2024},
  doi = {10.1038/s41467-024-51194-z},
}

Loading related content...

Figure