FET: Small: Hybrid Electrical, Ionic, and Biocompatible Artificial Synaptic Transistors

传统的计算系统具有限制数据存储和处理能力的根本性缺陷。与传统系统不同，人脑通过电化学过程运行，从而实现并行计算和高效数据存储。该项目旨在通过探索生物电子神经形态计算的潜力来克服传统计算系统的局限性。虽然在开发受大脑启发的神经形态硬件方面取得了进展，但这些人工架构与生物系统之间仍然存在差距。为了解决这一差距，该项目专注于创建混合生物电子神经形态计算系统，将人工突触与实时神经网络集成在一起。这项新技术在多个领域具有巨大的前景。在神经科学领域，它为研究突触对电信号和离子信号的反应提供了一个人工平台。在计算方面，它可以有效地处理非结构化数据，模拟大脑的能力。此外，在生物医学应用中，它促进了生物系统和计算机之间的无缝集成，为先进的生物电子混合动力车打开了大门。从长远来看，这些设备可以用作大脑植入物，允许复制大脑行为，并开发下一代假肢设备，用于治疗神经退行性疾病，如帕金森病和阿尔茨海默病。这项研究的影响将通过研究生培训进一步扩大；本科生全年参与研究，这样他们就可以为项目做出重大贡献；领导开发和领导一个项目，为研究生院的电气工程本科生做准备；以及传播教育视频，以提高对这一跨学科领域的认识和兴趣。研究人员提出了基于石墨烯晶体管的新型人工突触装置和阵列的设计，以满足无缝集成和响应生物信号的必要标准。基于最近的创新，该团队开发了人工突触晶体管，使用完全生物相容性材料，包括双层石墨烯和nafion基化合物。这些晶体管也表现出明显的低开关能量。该项目的主要目标是将石墨烯人造突触晶体管缩放到与生物学相关的尺寸，并研究单个器件的响应和阵列行为。该装置将被设计成对电信号和离子信号做出反应，特别是钾（K+）和钠（Na+）离子。器件的通道电导，与记忆状态相对应，将通过对国家门施加电脉冲来控制。这一过程促进了质子通过质子体的运动，导致电导的可控变化。结合数据驱动的多维建模和实验阵列测试的综合方法将用于验证设备及其阵列作为人工神经网络的功能和性能。通过建模和实验验证相结合，该项目旨在开发一种尖端的神经形态系统，该系统密切模仿哺乳动物神经元的特性。该奖项反映了美国国家科学基金会的法定使命，并通过使用基金会的知识价值和更广泛的影响审查标准进行评估，被认为值得支持。

Traditional computing systems have fundamental drawbacks that limit the capability of data storage and processing. Unlike conventional systems, the human brain operates through electrochemical processes, enabling parallel computing and efficient data storage. This project aims to overcome limitations in traditional computing systems by exploring the potential of bioelectronic neuromorphic computing. While progress has been made in developing brain-inspired neuromorphic hardware, there remains a gap between these artificial architectures and biological systems. To address the gap, the project focuses on creating hybrid bioelectronic neuromorphic computing systems that integrate artificial synapses with live neuronal networks. This novel technology holds immense promise across multiple domains. In neuroscience, it offers an artificial platform for studying synapse responses to electrical and ionic signals. In computing, it enables efficient processing of unstructured data, emulating the brain's capabilities. Moreover, in biomedical applications, it facilitates seamless integration between biological systems and computers, opening doors for advanced bioelectronic hybrids. In the long run, these devices could be used as brain implants, allowing the replication of brain behavior and the development of next-generation prosthetic devices for treating neurodegenerative diseases like Parkinson's and Alzheimer's. The impact of this research will be further broadened through graduate student training; year-round involvement of undergraduate students in the research, set up so they can significantly contribute to the project; leadership efforts in developing and leading a program to prepare undergraduates in electrical engineering for graduate school; and dissemination of educational videos to increase awareness and interest in this interdisciplinary area. The investigators propose the design of novel artificial synaptic devices and arrays based on graphene transistors to meet the necessary criteria for seamless integration and response to biological signals. Building upon recent innovations, the team has developed artificial synaptic transistors using fully biocompatible materials, including bilayer graphene and Nafion-based compounds. These transistors also exhibit significantly low switching energy. The project's key objectives are scaling the graphene artificial synaptic transistors to biologically-relevant sizes and investigating individual device response and array behavior. The devices will be engineered to respond to both electrical and ionic signals, specifically potassium (K+) and sodium (Na+) ions. The devices' channel conductance, which corresponds to memory states, will be manipulated by applying electrical pulses to the Nafion gate. This process facilitates the movement of protons through the Nafion body, leading to controlled changes in conductance. A comprehensive approach combining data-driven multidimensional modeling and experimental array testing will be used to validate the functionality and performance of the devices and their arrays as artificial neural networks. Through a combination of modeling and experimental validation, the project aims to develop a cutting-edge neuromorphic system that closely emulates the properties of mammalian neurons.This award reflects NSF's statutory mission and has been deemed worthy of support through evaluation using the Foundation's intellectual merit and broader impacts review criteria.