微电子
神经形态工程学
材料科学
晶体管
计算机科学
能源消耗
光电子学
人工智能
人工神经网络
电气工程
电压
工程类
作者
Theodoros Serghiou,José Diego Fernandes,Vaithinathan Karthikeyan,Dani S. Assi,Douglas Henrique Vieira,Neri Alves,Jeff Kettle
标识
DOI:10.1002/adfm.202417355
摘要
Abstract The recent advances in optic neuromorphic devices have led to a subsequent rise in the development of energy‐efficient artificial‐vision systems. While the energy consumption of such devices is known to be much lower than conventional vision systems, it is known that manufacturing accounts for the largest share of the climate impact in microelectronics, dominating over the product use phase. Thus, there is a need to develop sustainable manufacturing processes and to adopt low‐impact materials for hardware solutions of the future. In this study, an Electrolyte‐Gated Organic Field‐effect Transistor (EGOFET) is experimentally demonstrated for the implementation of a high‐performing synaptic optical sensor using sustainable materials that degrade to benign products at the End of Life (EoL). The device shows remarkable light response with maximum Paired‐Pulse Facilitation (PPF) Index of up to 151% at a light power density of 38 µW cm −2 , which enables artificial synaptic applications with an average power consumption as low as 2.4 pJ for each training process, representing one of the best among the reported results. To demonstrate the tunability of the vision system, an ensemble decision tree is used to enable the EGOFET to distinguish and remember different primary colors at different power densities with 95.6% accuracy.
科研通智能强力驱动
Strongly Powered by AbleSci AI