Nano Energy｜A Bionic Flexible Afferent Nervous System Based on Artificial Synapses

已传文件：photo/1655951975.png North Konami can provide MXene materials (customizable)







Research abstract






Introduction





Graphical guide



Figure 1. Schematic diagram of biological afferent nervous system and artificial afferent nervous system.





Figure 2. Basic electrical properties of two MXene artificial synapses and ANS devices. (a) Schematic diagram of the structure of the MXene artificial synapse. (b)-(d) HRTEM images of MXene nanosheets. (e) I-V characteristics of MXene@ITO, (f) turn-on voltage and current statistics, (g) statistical distribution of high and low resistance, (h) temperature dependence test. (i) I-V characteristics of the MXene@PET artificial synapse, (j) turn-on voltage and current statistics, (k) bending durability test, (l) retention characteristics after bending. (m) The current stability test of the ICE layer after the artificial synapse was suspended. (n) The current stability test of the ICE layer after the artificial synapse was biased with 0.5 V. (o) Schematic diagram of the ion migration of ICE and MXene layers after the artificial synapse is suspended. (p) Schematic diagram of ion migration of ICE and MXene layers after artificial synapse is biased at 0.5 V.







Figure 3. Tactile response testing of ANS devices and the corresponding statistics of plasticity synaptic current rate of change. (a), (b) Click test. (c), (d) Sliding test. (e), (f) Tactile test. (g), (h) Grip test. (i), (j) Grip test. (k), (l) Bend test. (m), (n) Grip test of the array. (o), (p) Radial artery pulse test.







Figure 4. Exploration of the physical mechanism of the ANS device. (a)-(e) Optical and SEM images of the surface and sides of the ICE layer. (f)-(i) The cross-sectional force analysis simulation of the porous ICE layer under external forces in different directions. (j)-(m) Simulation of cross-sectional force analysis of a single hole in the ICE layer under external forces in different directions. (n), (o) Schematic illustration of the ion migration of EMIM+ and BF4- inside the ICE layer with and without pressure. (p) Equivalent circuit diagram of the ICE layer. (q)-(t) Current response curves of ANS devices at different temperatures. (u) The rate of change of plastic synaptic current in ANS devices at different temperatures.





Figure 5. (a) A schematic diagram of operant conditioning based on “trial and error learning”. (b), (c) Postsynaptic current response test of ANS devices under combined pressure and temperature stimulation, realizing the process from short-term memory to long-term memory.



Summarize

This work proposes a low-power, miniaturized flexible ANS device and establishes a link between artificial synapses and real external stimuli. The device successfully simulates the neural response behavior triggered by real external stimuli, and the excellent performance of simulating operant conditioning makes it a strong candidate for tactile neuromorphic chips and artificial intelligence robots, for neuromorphic chips in tactile and sensory learning provides more possibilities.



Literature link

https://www.sciencedirect.com/science/article/pii/S221128552200564X

For the original text, please click the lower left corner of the tweet to read the original text




Special thanks to the author of this article for their strong support