Mimicking efferent nerves using a graphdiyne-based artificial synapse with multiple ion diffusion dynamics

The human nervous system's efficiency and complexity inspire research in neuromorphic computing, bioinspired systems, and brain-machine interfaces. Mimicking synapses is crucial for building artificial sensorimotor systems and neuromorphic chips. Various structures, including metal/insulator/metal stacks, electrolyte/semiconductor heterojunctions, and multiterminal transistors, have been explored. However, challenges remain in achieving biomimetic functionalities, including thermal and environmental stability. Graphdiyne (GDY), a novel carbon allotrope, offers unique properties including efficient ion storage and transport due to its porous structure and sp-hybridized carbon atoms. This makes GDY a promising candidate for creating artificial synapses.

The existing literature highlights the use of various materials and architectures to mimic synaptic behavior in artificial systems. Metal/insulator/metal (MIM) structures, electrolyte/semiconductor heterojunctions, and multiterminal transistors have shown some success in emulating signal transmission. The focus on ion migration as a mechanism for emulating biological sensory/motor neurons through neuromorphic synapses is gaining traction in bioinspired ionotronic and biohybrid systems. However, limitations in thermal and environmental stability, along with a need for exploration of new materials and device responsivity, persist. The unique properties of graphdiyne (GDY), a new carbon allotrope, in ion storage and transport, makes it a promising candidate for creating artificial synapses. Previous research has demonstrated GDY's potential in other applications, such as batteries, catalysis, and solar cells, due to its remarkable optoelectronic properties and biocompatibility.

The researchers fabricated a junction-type GAS by coupling a GDY film with solid-state electrolytes (Li-GAS and Na-GAS). Two-dimensional GDY was synthesized using a catalytic coupling reaction of hexakis[(trimethylsilyl)ethynyl]benzene (HEB-TMS). The GDY film's morphology and crystallinity were characterized using AFM, SEM, TEM, and XPS. I-V curves were obtained to investigate ion migration dynamics. First-principles calculations using VASP were performed to study the diffusion behavior of Li and Na ions in GDY. Thermal and environmental stability were evaluated by testing the devices at elevated temperatures (up to 353 K) and humidity (up to 35%). The GAS's integration with artificial muscles was achieved to demonstrate an artificial efferent nerve. Electrical measurements were conducted using a Keithley 4200A semiconductor parameter analyzer in a nitrogen-filled glove box. The actuator was fabricated using poly(vinylidene fluoride-co-hexafluoropropylene) (PVDF-HFP) and 1-ethyl-3-methylimidazolium tetrafluoroborate (EMIBF4) solution casting, sandwiched between CNT electrodes. The synaptic device-amplifier circuit-polymer actuator system was constructed to control the actuator's operation.

The fabricated GAS demonstrated several key characteristics of biological synapses, including short-term plasticity (postsynaptic current, paired-pulse facilitation, dynamic filtering), ultralow power consumption (16.7 fW per synaptic event), and high sensitivity (millivolt-level response). The GAS exhibited good ion diffusion dynamics even at high temperatures and humidity. First-principles calculations supported the experimental findings by revealing low diffusion energy barriers for Li and Na ions in GDY. The integration of GAS with artificial muscles demonstrated real-time information integration and parallel processing capabilities in an artificial efferent nerve, enabling signal transduction and actuation. The GAS effectively integrates multiple inputs and modulates neural responses. The parallel processing capability was confirmed by applying impulses of different frequencies to the GAS, with the artificial muscle response reflecting the integrated information. The GAS could identify frequencies up to 5.6 Hz. The study demonstrated the construction of an artificial efferent nerve connecting the GAS with artificial muscles, resulting in muscle actuation proportional to the number and type of inputs.

The findings demonstrate a significant advancement in the creation of bio-inspired artificial synapses. The GAS's ability to mimic short-term plasticity, parallel processing, and integration of multiple inputs, combined with its low power consumption, high sensitivity, and stability, addresses a crucial gap in building more sophisticated neuromorphic systems and biohybrid interfaces. The success in creating an artificial efferent nerve highlights the potential for applications in diverse fields including neurorobotics, soft electronics, and brain-computer interfaces. The biocompatibility of GDY also opens avenues for integrating this artificial synapse with biological systems.

This research successfully demonstrated a graphdiyne-based artificial synapse with parallel processing and information integration capabilities. The GAS exhibits key synaptic characteristics, including short-term plasticity and ultralow power consumption, while maintaining stability under various conditions. Its integration into an artificial efferent nerve system demonstrates potential applications in bioinspired sensorimotor systems and biohybrid systems. Future research could focus on scaling up GDY production and exploring the integration of the GAS with more complex biological systems.

While the study demonstrated significant progress, limitations exist. The large-scale preparation of high-quality GDY films remains a challenge. The experiments were conducted with specific artificial muscles, and further research is needed to assess compatibility with a wider range of actuators. The frequency response of the GAS is limited (up to 5.6 Hz), and the study did not address long-term plasticity.