CAREER: Mixed-Signal Photonic Integrated Circuits for Energy-Efficient High-Speed Data Interfaces

职业：用于节能高速数据接口的混合信号光子集成电路

• 批准号: 2014109
• 负责人: Vishal Saxena
• 金额: $ 30.01万
• 依托单位: University of Delaware
• 依托单位国家: 美国
• 项目类别: Standard Grant
• 财政年份: 2019
• 资助国家: 美国
• 起止时间: 2019-08-23 至 2022-02-28
• 项目状态: 已结题
• 来源: https://www.nsf.gov/awardsearch/showAward?AWD_ID=2014109&HistoricalAwards=false
• 关键词: CAREER; Mixed; Signal; Photonic; Integrated

Proposal No.: 1454411 CAREER: Mixed Signal Photonics Integrated Circuits for Energy-Efficient High-Speed Data InterfacesVishal Saxena (Boise State University)Abstract: This research harnesses photonics to satisfy ever-growing government, industry, and consumer needs for data bandwidth, while significantly reducing the increasingly voracious energy footprint that accompanies Internet cloud use. Photonics uses light instead of electronics to perform a variety of functions such as information processing and transfer. While most of us still work in offices and meet friends in person, mobile and networking device capability has enabled massive online business and social transaction growth. The Internet and its cloud-based services are used for production systems, banking, entertainment, social interaction, information distribution and research. Resulting information accumulation is fueling the rise of "big data," as well as the powerful correlation tools that help analysts spot financial trends, prevent diseases, combat crime and improve quality of research. More and more, we put content in the cloud for easy access from anywhere and at any time, with data growing exponentially. All this data transfer uses a surprising amount of energy. To reduce data center energy consumption while increasing data capacity by over ten fold, this research will investigate novel hybrid data communication interfaces, using light rather than electrons to process and transfer data at higher speeds. The potentially explosive increase in data rates enabled by these hybrid photonic interconnects could lead to several transformative applications, such as future exascale data centers, terabit speed local area networks, and massively-parallel computing for big data applications. These hybrid photonic interconnects will not only have a broad impact on the semiconductor industry, but also US energy sustainability and security, as more energy-efficient computing systems would reduce carbon footprint of the Internet cloud. Further, to prepare students for the workforce with the necessary skills to drive future technology, the project includes a strong educational component. Interactive learning methods will be employed to teach electronic circuits and to bring photonics to integrated circuit design. The project also incorporates a high school outreach program, with an annual Smart Environments for Sustainability-themed one-week summer camp for high school students that commits to fostering women and minority group representation in integrated circuit design.Technology development leveraging integrated photonic circuits and optical interconnects has thus far largely focused on binary communication using silicon photonic modulators. To enable future optical interconnects for higher data rates and energy-efficiency, researchers must reconsider the hybrid integrated circuit paradigm. An important technology enabler is the high-speed signal processing capability of integrated photonic devices. The research approach will first be to develop a photonic design kit with standard cell libraries and compact models to enable large scale integration of photonic devices into hybrid integrated circuits. Researchers will employ photonic device high-speed optical domain signal processing into novel circuit configurations, exploit synergistic interaction between electronic and photonic components, and form a mixed-signal photonic architecture. Next, to exploit photonics beyond binary interconnects; researchers will develop novel mixed-signal photonic data converters, which they will use to demonstrate an advanced modulation transceiver architecture that is scalable to terabits per second data rates with order-of-magnitude lower energy consumption. Research outcomes will empower integrated circuit researchers by equipping them with a new photonics expertise to tackle nano-scaled technology design challenges, where data transfer bottlenecks constrain system performance. The photonic design kit will lower industry barriers to help facilitate photonics adoption into integrated circuits; resulting mixed-signal photonic data converter architectures will set a new paradigm by achieving greater than 10 GHz sampling rates with significantly reduced energy consumption over existing complementary metal?oxide?semiconductor (CMOS)-only architectures. Researchers will broadly disseminate project results by developing online educational material for a new CMOS photonics integrated circuit design graduate course, and through international journals and conferences.

方案编号：1454411职业生涯：混合信号光子集成电路的节能高速数据接口Vishal Saxena（博伊西州立大学）摘要：这项研究利用光子学，以满足不断增长的政府，行业和消费者对数据带宽的需求，同时显着减少日益贪婪的能源足迹，伴随着互联网云的使用。光子学使用光而不是电子来执行各种功能，如信息处理和传输。虽然我们中的大多数人仍然在办公室工作，并会见朋友的人，移动的和网络设备的能力，使大规模的在线业务和社会交易的增长。互联网及其基于云的服务用于生产系统、银行、娱乐、社交、信息分发和研究。由此产生的信息积累正在推动“大数据”的兴起，以及帮助分析师发现金融趋势、预防疾病、打击犯罪和提高研究质量的强大相关工具。随着数据呈指数级增长，我们越来越多地将内容放在云中，以便随时随地轻松访问。所有这些数据传输都使用了惊人的能量。为了降低数据中心的能耗，同时将数据容量提高十倍以上，本研究将研究新型混合数据通信接口，使用光而不是电子以更高的速度处理和传输数据。由这些混合光子互连实现的数据速率的潜在爆炸性增长可能会导致几个变革性的应用，例如未来的亿级数据中心，兆比特速度局域网和大数据应用的并行计算。这些混合光子互连不仅将对半导体行业产生广泛影响，还将对美国能源可持续性和安全性产生广泛影响，因为更节能的计算系统将减少互联网云的碳足迹。此外，为了让学生为劳动力做好准备，掌握推动未来技术的必要技能，该项目包括一个强大的教育部分。互动学习方法将被用来教授电子电路，并将光子学引入集成电路设计。该项目还包括一个高中外展计划，每年为高中生举办一个为期一周的以可持续发展为主题的智能环境夏令营，致力于培养女性和少数群体在集成电路设计中的代表性。利用集成光子电路和光学互连的技术开发迄今为止主要集中在使用硅光子调制器的二进制通信上。为了使未来的光互连实现更高的数据速率和能源效率，研究人员必须重新考虑混合集成电路的范例。一个重要的技术使能器是集成光子器件的高速信号处理能力。研究方法将首先是开发一个光子设计套件与标准单元库和紧凑的模型，使大规模集成的光子器件到混合集成电路。研究人员将采用光子器件高速光域信号处理到新的电路配置，利用电子和光子元件之间的协同作用，并形成混合信号光子架构。接下来，为了利用超越二进制互连的光子学;研究人员将开发新型混合信号光子数据转换器，他们将使用该转换器来演示一种先进的调制收发器架构，该架构可扩展到每秒太比特的数据速率，并且能耗低一个数量级。研究成果将使集成电路研究人员具备新的光子学专业知识，以应对纳米级技术设计挑战，其中数据传输瓶颈限制了系统性能。光子设计套件将降低行业壁垒，以帮助促进光子学应用到集成电路中;由此产生的混合信号光子数据转换器架构将通过实现大于10 GHz的采样率，并显著降低现有互补金属的能耗，从而建立一个新的范例。氧化物？半导体（CMOS）架构。研究人员将通过为新的CMOS光子集成电路设计研究生课程开发在线教育材料，并通过国际期刊和会议广泛传播项目成果。

项目成果

Rapid Simulation of Photonic Integrated Circuits Using Verilog-A Compact Models

使用 Verilog-A 紧凑模型快速仿真光子集成电路

• DOI: 10.1109/tcsi.2020.2983303
• 发表时间: 2020
• 期刊: IEEE Transactions on Circuits and Systems I: Regular Papers
• 作者: Shawon, Md Jubayer;Saxena, Vishal
• 通讯作者: Saxena, Vishal

Mixed-Signal Neuromorphic Computing Circuits Using Hybrid CMOS-RRAM Integration

• DOI: 10.1109/tcsii.2020.3048034
• 发表时间: 2021-02
• 期刊: IEEE Transactions on Circuits and Systems II: Express Briefs
• 作者: V. Saxena

Neuromorphic computing: From devices to integrated circuits

神经形态计算：从设备到集成电路

• DOI: 10.1116/6.0000591
• 发表时间: 2021
• 期刊: Journal of Vacuum Science & Technology B
• 影响因子: 1.4
• 作者: Saxena, Vishal