The Optical-Electronic Integrated Spiking Neurons Based on Antiferroelectric Thin-Film Transistors
Sensory neurons are an important componentin achieving flexible and efficient intelligent sensing systems. Multimodal sensory neurons provide more comprehensive and reliable sensing than unimodal neurons, enabling them to better perform complex recognition or decision-making tasks. This article prop...
Gespeichert in:
Veröffentlicht in: | IEEE transactions on electron devices 2024-10, Vol.71 (10), p.6442-6447 |
---|---|
Hauptverfasser: | , , , , , , , , , |
Format: | Artikel |
Sprache: | eng |
Schlagworte: | |
Online-Zugang: | Volltext bestellen |
Tags: |
Tag hinzufügen
Keine Tags, Fügen Sie den ersten Tag hinzu!
|
Zusammenfassung: | Sensory neurons are an important componentin achieving flexible and efficient intelligent sensing systems. Multimodal sensory neurons provide more comprehensive and reliable sensing than unimodal neurons, enabling them to better perform complex recognition or decision-making tasks. This article proposes a multimodal sensory neuron based on antiferroelectric thin-film transistors (AFeTFTs) for the first time. It can realize leaky integrate-and-fire (LIF) neuron behavior in both optical and electrical modes. The hafnium-based antiferroelectric capacitor is integrated on the gate of amorphous indium gallium zinc oxide thin-film transistors (a-IGZO, TFTs). By utilizing the polarization and depolarization properties of the antiferroelectric thin film Hf 0.14 Zr 0.86 O 2 , achieving the integration and firing behavior of LIF neurons in electrical mode. Additionally, based on the photosensitive properties of IGZO films, the device also realizes LIF neuron behavior in optical mode. The device exhibits reliable performance and high reproducibility. Since the IGZO film serves as both the channel layer of the transistor and the photosensitive unit of the neuron, the device has a compact structure and is convenient for integration. This work provides an effective approach for multimodal sensing fusion and functional integration, offering a solution for the realization of more efficient and flexible intelligent sensing systems. |
---|---|
ISSN: | 0018-9383 1557-9646 |
DOI: | 10.1109/TED.2024.3450440 |