Exploration of SiGe HBTs for power amplifiers in the 200 GHz to 500 GHz frequency range

通信、成像和传感等领域对毫米波和亚毫米波应用的需求不断增加，对相应的电子系统在带宽、输出功率和能源效率方面提出了重大挑战。在这种系统中，一个重要的部件是功率放大器（PA），它也是能量耗散的主要贡献者。根据目前的趋势，未来更高的带宽将通过移动到更高的载波频率（比目前的5G系统）来实现，其中包含数百个pa用于波束转向和大量多输入多输出信号的有源天线相控阵。这种高度集成的高频（HF）系统只能通过硅锗（SiGe） BiCMOS技术实现。因此，探索现有SiGe异质结双极晶体管（hbt）的性能和对未来性能的要求，以实现具有足够输出功率和高功率和频谱效率的高频PAs是非常有意义的。迄今为止，很少注意200千兆赫及以上的频率，这些频率与处理器对处理器通信中的高数据速率传输、微蜂窝和飞蜂窝环境、安全检查以及生物探测和医学成像等有关。在这种频率下的一个主要挑战是产生足够的输出功率，同时实现高效率。因此，拟议的项目旨在(i)探索在200…500年运行的SiGe HBT PAs的最大可实现功率密度基于实验数据和ITRS/IRDS预测的GHz范围和与设备物理相关的性能限制；（ii）对输出功率、效率、增益、带宽和线性度之间权衡的研究；(3)基于混合模式数值模拟（TCAD）和预制电路的精确大信号紧凑HBT建模和PA运行过程中的非线性瞬态和物理效应的详细分析；（iv）单独评估在非线性大信号运行时PAs产生的谐波功率，例如，模型验证。在三个明确定义的工作包中详细介绍了所提出的工作，包括导电角PAs及其构建模块的设计、制造和实验表征。到目前为止，由于缺乏对HBT操作限制的深入了解和相应的精确模型，在高频放大器设计中实现高输出功率和效率一直受到保守方法的限制。现有的障碍将通过将晶体管工作推向其物理极限，并通过仔细的PA优化，考虑所有可能的器件布局选项，基于几何可扩展的HBT建模和重要的分布式效应，如热和衬底耦合，来克服。

The continuously increasing demand for mm- and sub-mm-wave applications in, e.g., communications, imaging and sensing poses significant challenges for the corresponding electronic systems in terms of bandwidth, output power, and energy efficiency. An important component in such systems is the power amplifier (PA), which is also a major contributor of energy dissipation. According to present trends, higher bandwidth in future will be achieved by moving to much higher carrier frequencies (than those in present 5G systems) with active antenna phased arrays containing hundreds of PAs for beam steering and massive multiple-input-multiple-output signals. Such highly-integrated high-frequency (HF) systems can only be realized with silicon-germanium (SiGe) BiCMOS technology. Hence, the exploration of both the capability of existing and the requirements for future performance of SiGe heterojunction bipolar transistors (HBTs) with respect to the realization of HF PAs with sufficient output power and high power and spectral efficiency is of great interest. So far, little attention has been paid to frequencies at and beyond 200 GHz, which are relevant, e.g., for high data-rate transmission in processor-to-processor communications, pico- and femto-cell environments, and security screening as well as for biological probing and medical imaging. A major challenge at such frequencies is the generation of sufficient output power, while achieving at the same time also high efficiency. The proposed project therefore aims at (i) the exploration of the maximum achievable power density of SiGe HBT PAs operating in the 200...500 GHz range and the performance limitations related to device physics based on experimental data and ITRS/IRDS predictions; (ii) the investigation of the trade-offs between output power, efficiency, gain, bandwidth, and linearity; (iii) accurate large-signal compact HBT modeling and detailed analysis of non-linear transients and physical effects during PA operation based on mixed-mode numerical device simulation (TCAD) and fabricated circuits; (iv) separate assessment of the harmonic power generated in PAs during nonlinear large-signal operation for, e.g., model verification. The proposed work, as detailed in three well-defined work packages, includes the design, fabrication and experimental characterization of conduction-angle PAs and their building blocks. So far, achieving high output power and efficiency in HF PA designs has been constrained, among others, by conservative approaches due to the lack of deeper insight into HBT operation limits and the corresponding accurate models. The existing barriers will be overcome by pushing transistor operation to its physical limits and by careful PA optimization, considering all possible device layout options based on geometry scalable HBT modeling and important distributed effects such as thermal and substrate coupling.