← 返回
储能系统技术 储能系统 SiC器件 GaN器件 ★ 5.0

有机光电子突触晶体管阵列的研究进展:系统集成的制备策略与创新应用

Recent progress in organic optoelectronic synaptic transistor arrays: fabrication strategies and innovative applications of system integration

作者 Pu GuoJunyao ZhangJia Huang
期刊 半导体学报
出版日期 2025年1月
卷/期 第 46 卷 第 2 期
技术分类 储能系统技术
技术标签 储能系统 SiC器件 GaN器件
相关度评分 ★★★★★ 5.0 / 5.0
关键词 Pu Guo Junyao Zhang Jia Huang 半导体学报(英文版) Journal of Semiconductors
语言:

中文摘要

人工智能的快速发展加剧了传统计算架构在能耗和数据延迟方面的瓶颈。数据为中心的存算一体架构有望解决上述问题。有机光电子突触晶体管因其柔性、低成本和大面积制备优势,成为实现该架构的候选器件之一。然而,单个器件难以完成矩阵向量乘法等复杂功能,因此亟需构建有机光电子突触晶体管阵列(OOSTAs)。本文综述了OOSTAs的最新进展,涵盖涂布、物理气相沉积、印刷及光刻等多种制备策略,并探讨其在神经形态视觉与计算系统中的集成应用,最后分析了其在实际应用中面临的挑战与未来发展方向。

English Abstract

The rapid growth of artificial intelligence has accelerated data generation,which increasingly exposes the limitations faced by traditional computational architectures,particularly in terms of energy consumption and data latency.In contrast,data-centric computing that integrates processing and storage has the potential of reducing latency and energy usage.Organic optoelectronic synaptic transistors have emerged as one type of promising devices to implement the data-centric com-puting paradigm owing to their superiority of flexibility,low cost,and large-area fabrication.However,sophisticated functions including vector-matrix multiplication that a single device can achieve are limited.Thus,the fabrication and utilization of organic optoelectronic synaptic transistor arrays(OOSTAs)are imperative.Here,we summarize the recent advances in OOSTAs.Various strategies for manufacturing OOSTAs are introduced,including coating and casting,physical vapor deposition,printing,and photolithography.Furthermore,innovative applications of the OOSTA system integration are discussed,including neuromor-phic visual systems and neuromorphic computing systems.At last,challenges and future perspectives of utilizing OOSTAs in real-world applications are discussed.
S

SunView 深度解读

该有机光电子突触晶体管阵列技术对阳光电源智能控制系统具有前瞻性启发价值。其存算一体架构可应用于iSolarCloud云平台的边缘计算节点,实现光伏电站和储能系统的低功耗实时数据处理与智能诊断。神经形态计算特性可优化ST系列储能变流器的预测性维护算法,通过模式识别提升故障预警准确率。柔性大面积制备优势为分布式传感器网络提供低成本方案,可集成于PowerTitan储能系统的温度、电流监测模块。虽然当前技术成熟度有限,但其低延迟、低能耗特性契合新能源装备智能化趋势,为下一代边缘AI芯片开发提供技术储备,支撑构网型控制等复杂算法的本地化部署。