Multi-Layered Triboelectric Nanogenerator for Human-Machine Interface Using an Artificial Synaptic Device

Artificial synaptic device provides functions of sensing, short-term and long-term memory, which can find applications in various human-machine interfaces. Memory and sensing functions with low energy consumption are considered as a key of the neuromorphic system that is superior to high energy consuming von Neumann computing architecture. Triboelectric nanogenerator (TENG) provides spike-like impulses similar to the action potential propagation in the biological system, which can be utilized for the operation of artificial synaptic devices. Herein, we propose multi-layered TENG (M-TENG) capable of generating multiple spike impulses by a single mechanical stimulus based on tribo-positive microdome-patterned polydimethylsiloxane (PDMS) and tribo-negative 1H,1H,2H,2H-perfluorooctyltrichlorosilan (FOTS)-coated BaTiO₃/PDMS (FBP) composite films. While conventional triboelectric devices generate two spike peaks when the top and bottom triboelectric layers contact and separate with each other <i>via</i> contact electrification and electrostatic induction, our M-TENG can generate distinct multiple spikes corresponding to the number of layers under a single mechanical stimulus of the same applied force and frequency. Moreover, our M-TENG shows 1.33 times higher output charge compared to conventional single-layered TENG and the power density of 7.56 μW/cm² at 100 MΩ under the applied force of 19.6 kPa. For the application in artificial synaptic devices, M-TENG is coupled with organic electrochemical transistor (OECT), where multiple spikes generated from M-TENG are served as the gate voltage for the self-powered system. Our synaptic device exhibits continuous increase and gradual decay of post-synaptic current under several stimulations, leading to long-term plasticity (LTP). LTP is formed by paired-pulse facilitation (PPF) of excitatory post-synaptic current (EPSC). The efficiency of the triggered EPSC can be expressed by PPF index with different time intervals of stimulations. To further investigate human-machine interface using the artificial synaptic devices, a robotic hand is operated by training process mimicking memory consolidation.