硅基CMOS太赫兹低噪声放大器的增益提升理论与设计方法研究

Silicon-based CMOS Terahertz low noise amplifier with good performance of power gain and noise figure is the key to realize high sensitive silicon-based Terahertz receiver systems. As the operation frequency approaches the maximum oscillation frequency (fmax) of a transistor, the maximum available power gain will degrade significantly, which makes the design of silicon-based Terahertz amplifier extremely challenging. This project is to investigate the gain-boosting theory and gain-noise co-design methodology of silicon-based CMOS Terahertz low noise amplifier. Based on the attribute that Mason'U is an invariant across all the linear-lossless-reciprocal embedded networks, the Gain Plane Two Step gain boosting theory will be established. Further this project is to consider the noise figure into the gain boosting theory with the limited design margin of power gain, and then to propose the gain-noise co-design methodology. Under the guidance of the proposed theory, based on CMOS process a complete Terahertz LNA chip operating at 3/4 fmax will be taped out and tested, which will demonstrate the proposed theory and design methodology in this project. The objective of this research is to provide the theoretical and methodological support for the design of THz silicon-based low noise amplifier, and to lay some foundation for the realization of high sensitive silicon-based Terahertz receiver systems.

具有良好功率增益和噪声系数的硅基CMOS太赫兹低噪声放大器是实现高灵敏度太赫兹接收机系统的关键。当工作频率接近单个晶体管的最大振荡频率（fmax）时，晶体管的可利用功率增益会显著下降，使得硅基太赫兹低噪声放大器的设计十分困难。本项目拟针对硅基CMOS太赫兹低噪声放大器的功率增益和噪声系数开展理论与设计方法研究，基于线性无损互易网络下Mason’s U值不变的特点，提出Gain Plane两步功率增益提升理论；进一步，在有限的功率增益设计裕度下加入噪声系数模型，提出增益-噪声协同优化设计方法。在此基础上，通过硅基CMOS工艺流片实现工作频率为3/4 fmax的太赫兹低噪声放大器，验证本项目提出的理论与设计方法。以期通过本项目的研究工作为太赫兹频段的硅基低噪声放大器设计提供一定的理论与设计方法支持，为获得高灵敏度硅基太赫兹接收机系统奠定一定的基础。

硅基太赫兹电路在高速通信、雷达、成像、安全检测等领域具有十分广阔的应用前景。由于工作于太赫兹频段的放大器，其工作频率接近单个晶体管的fmax时，晶体管的可利用功率增益会显著下降，使得传统的射频集成电路设计方法难以适用于硅基太赫兹放大器设计，对于硅基太赫兹低噪声放大器带来了严峻的挑战。本项目针对这些问题，从两个方面开展了理论研究和实验验证：（1）硅基太赫兹低噪声放大器太赫兹放大器的功率增益提升理论。提出了Over-push增益提升理论，可以解决匹配网络损耗对增益提升带来负效果的问题；提出了针对cascode放大器的Bottom-Over-Neutralization增益提升技术，可以实现cascode放大器中的Gma和U的同时提升；提出了自顶而下的匹配网络综合算法，可以自适应的获得低损耗的匹配网络阶数和取值。（2）CMOS太赫兹低噪声放大器的噪声系数和功率增益协同优化设计方法。提出了基于增益噪声平面理论，并基于此平面提出了太赫兹低噪声放大器增益噪声协同设计方法；基于增益噪声平面理论，对Cascode太赫兹低噪声放大器噪声设计方法进行研究，通过实验验证获得性能优秀的硅基太赫兹低噪声放大器芯片。本项目的研究成果可为高灵敏度硅基太赫兹接收机系统的研究奠定一定的基础，为未来超大容量通信等应用提出技术支撑。