Concepts of instrumentation and signal analysis for electrochemical impedance spectroscopy of field-effect transistor-based biosensors

Field-effect transistor-based biosensors (BioFETs) offer label-free, highly sensitive, and miniaturizable electrochemical detection for medical, environmental or industrial applications. In BioFETs, electrochemical impedance spectroscopy (EIS) is used to measure the sensor's transimpedance to identify semiconductor and interfacial processes. Thus, frequency-dependent responses are converted into reliable biochemical information. However, this technology is not widely used in every-day life due to its dependence on laboratories and complex analysis pipelines. This thesis addresses these barriers by proposing a comprehensive framework covering portable instrumentation, robust onboard analytics, physical validation tools, and adaptive frequency selection. All of these components are tailored to EIS of BioFETs and are compatible with autonomous, on-site operation. The thesis presents a portable, four-channel impedance analyzer (PIA) tailored for BioFET operation, which allows for necessary operating point adjustments and performs a frequency sweep across the relevant EIS frequency range of 10 Hz to 200 kHz. It effectively bridges the gap between lab-grade measurement setups and embedded systems for point-of-care and on-site applications. When analyzing EIS spectra, a common approach is to use complex nonlinear least squares (CNLS) algorithms to fit equivalent circuit diagrams, thereby minimizing residuals in both the real and imaginary parts. To reduce its sensitivity to initial values and improve its success rate, pre-fitting and parameter normalization are introduced. Additionally, novel geometric approaches involve fitting circles or ellipses to the Nyquist plot to directly determine the charge-transfer resistance from Randles circuits' spectra. This eliminates the need for computationally intensive model fittings and enables millisecond-level computing on embedded, low-power hardware. Consequently, it enables the analysis of EIS spectra within the sensor system itself. To validate the entire signal acquisition chain from instrumentation to algorithms, a physical Randles circuit simulator that even covers diffusion elements and configurable interfacial components is implemented to provide known ground truths of the configured parameters for benchmarking. Moreover, the EIS measurement time is reduced by selecting only informative frequencies rather than performing a logarithmic frequency sweep. Therefore, an embedded implementation minimizes the number of frequencies via iterative curve analysis and circular fitting. A real-time predetermination of excitation frequencies (RTPEF), alternatively, uses multi-objective optimization to balance fit quality, duration, and number of frequencies. Demonstrations with pH-sensitive ISFETs and biofunctionalized BioFETs for salivary biomarkers illustrate the functionalities across hardware and software components. The presented cohesive approaches enable autonomous BioFET measurements in real settings and facilitates possible transfers to other biosensing modalities.