Event-Trigger Dual-Signal-Mode Flexible Piezotronic Bipolar Junction Transistor With Machine Learning for AI-Sound Recognition

Emad Iranmanesh,Jucai Zhai,Congwei Liao,Zihao Liang,Yong Zhao,Hang Zhou,Shengdong Zhang,Charalampos Doumanidis,Anand Prakash Dwivedi,Kai Wang
DOI: https://doi.org/10.1109/led.2023.3238065
IF: 4.8157
2023-03-01
IEEE Electron Device Letters
Abstract:This work reports on a piezotronic n-p-n bipolar heterojunction transistor (PBJT) intended for AI-sound recognition with low power consumption footprint. An organic polymer (Poly (3-hexylthiophene)) as Base is sandwiched between two inorganic piezoelectric semiconductor ZnO layers as Emitter and Collector, respectively to form a vertically-stacked heterogeneous n-p-n BJT, which differs from the traditional BJT in that it can provide either current or voltage as an output signal to favor application's preference. It can be regarded as two back-to-back auto-biased piezotronic heterojunction diodes where Base-Collector diode is reversely biased and Base-Emitter one is in a forward bias (or vice versa) upon sound wave stimuli. In addition, polarized charges generated at Emitter and Collector sides redistribute and as a result, give a significant enhancement in the output voltage. With machine learning AI, low frequency (LF) sound detection, mapping, and recognition have been demonstrated by this device.
engineering, electrical & electronic
What problem does this paper attempt to address?