Performance Limits of SiGe Power Heterojunction Bipolar Transistors for Future Wireless System-on-a-Chip (SoC)

University Of Wisconsin-Madison, Madison WI

Investigators

Abstract

0323717 Ma The rapid advancement in SiGe HBT technology allowed unprecedented integration of high frequency microwave and millimeter wave circuits on Si substrates. While the integration of SiGe HBTs with Si-based CMOS VLSI technology is promising for future wireless system-on-a-chip (SoC), the on-chip integration of high-frequency Si-based microwave transmitters is still inaccessible, owing to the lack of high performance SiGe high-power HBTs up to date. Therefore, research on high-frequency microwave high-power SiGe HBTs to surmount the existing challenges has become necessary to make these devices suitable for SoC integration. The specific objectives of the proposed research plan are to advance SiGe HBT manufacturing technology, to explore innovative SiGe HBT design techniques, and to advance fundamental understanding of SiGe HBT high-power operations mechanisms, with the goal of substantially extending the power amplification frequency of SiGe HBTs. The objectives also include the development of physics-based numerical algorithms that can be used to simulate large-signal power characteristics based on the vertical heterostructure and layout of SiGe HBTs. A new electrical engineering course will be offered through tight integration of the proposed research activities and course instruction. The following approaches are proposed to accomplish the research plan. Submicron scaling, with novel collector design and innovative processing techniques for parasitics reduction, will be employed to dramatically enhance the maximum oscillation frequency of large emitter area power SiGe HBTs. A novel heat transfer-balanced HBT layout structure will be primarily investigated by accurate modeling of heat transfer and parasitics within large-area SiGe HBTs. The SiGe HBT power operation mechanisms will be exposed by separating the thermal effects and parasitic effects on power gain degradation for both small-signal and large-signal operations. Physics-based numerical algorithms will be developed by investigating the connections between SiGe HBT structure and model parameters with the consideration of partitioned thermal effects on a single power SiGe HBT subcell. Through demonstration of SiGe HBT processing, characterization and modeling, enhanced learning effectiveness will be achieved in the new electrical engineering course. The studies of SiGe power HBT performance limits have substantial impact on microwave solid-state devices and a variety of wireless systems that are operated in the microwave frequencies spanning L- to Ka-band. The operation of high-frequency power SiGe HBTs will yield the potential of replacement of a considerable portion of III-V active devices in many critical circuits and systems. Significantly increased integration level, reliability and cost saving will be achieved with SiGe HBTs in these circuits and systems. The development of physics-based algorithms will facilitate SiGe power HBT development, reduce R&D cost, and provide modeling verification support for power amplifier circuit design. The ultimate advancement of SiGe HBT technology may create a breakthrough on completely Si-based high-frequency systems and would represent a significant contribution to the microwave and solid-state device communities. Furthermore, the breakthrough may eventually pave the way to realizing future SoC wireless communications. The proposed integration of research and education will significantly enhance students learning effectiveness and the new course will be a significant contribution to the curriculum infrastructure of electrical engineering at University of Wisconsin-Madison. The dissemination of research results through the construction of an instructive course homepage will benefit a broader section of the society. The budget of the proposal will be primary used to support graduate and undergraduate students with emphasis on supporting underrepresented minority and woman students. The proposed research will be conducted at Wisconsin Center for Applied Microelectronics (WCAM), University of Wisconsin-Madison. Wafers for iterative processing will be grown by Lawrence Semiconductor Research Laboratory (LSRL).