Owing to the unique characteristics of terahertz (THz) waves and their interaction with matter, THz imaging has significant potential in diverse fields, including life sciences, non-destructive quality control, and security. However, today’s THz technology faces major obstacles in providing the cost, compactness, and functional scope that is required to facilitate a wide-reaching use outside the laboratory. Therefore, the general aim of this thesis is to increase the societal impact of THz technology by exploring novel THz imaging modalities with integrated circuits (ICs) based on conventional silicon technolgy.
Previous research in THz ICs mostly focused on building compact integrated source and detector components for far-field imaging systems that exploit the ability of THz waves to penetrate through diverse dielectric materials. One part of this thesis joins this research and demonstrates low-cost volumetric THz imaging based on computed tomography with silicon components for the first time. To this end, a high-power 0.43 THz source in 0.13 μm silicon-germanium heterojunction bipolar transistor (SiGe HBT) technology is developed in this thesis. However, the resolution of THz far-field imaging is limited by diffraction to the millimeter range, whereas some of the most promising THz applications in life sciences, such as intraoperative imaging of cancerous tissue, are in need for a microscopic resolution to resolve material properties on the cellular level.
This thesis presents the realization of microscopic THz imaging with a silicon-based sensor system. The central contribution of this work is the development and analysis of a 128-pixel THz near-field sensor System-on-a-Chip (SoC) operating at 0.55 THz and showing a spatial resolution around 10 μm. The sensor exploits the capacitive near-field interaction between split-ring-resonator probes and imaging objects, giving an imaging contrast based on the dielectric permittivity. To simultaneously enable real-time image acquisition, high sensor sensitivity, and an integration level that is comparable to conventional consumer electronics, the full integration capabilities of a high-speed 0.13 μm SiGe bipolar CMOS (SiGe-BiCMOS) technology are exploited. In particular, the presented SoC employs cointegration of a chip-scale one-dimensional THz sensor front-end, analog signal processing, and digital circuitry for controlling the chip and external communication. The achieved results in terms of imaging speed, system cost, and integration level reach well beyond the state of the art in THz near-field imaging.
The ability to rapidly acquire THz images with micrometer-scale resolution will be of benefit to fundamental research into material properties in the THz range and lays the foundation for the exploration of THz bioimaging applications on the cellular level. In particular, the sensor enables the conduction of large-scale clinical studies on the relevance of microscopic THz imaging to ex vivo tumor margin assessment in breast cancer surgeries for the first time.