Logo: to the web site of Uppsala University

Rapid and reliable antibiotic susceptibility testing (AST) is urgently required to diagnose bacterial infectious diseases and avoid antibiotic misuse, providing valuable information on the efficacy of antibiotic agents and their dosages for treatment. However, the currently employed phenotypic ASTs normally demand the growth of bacteria into colonies, which usually takes more than two days. In this thesis, silicon nanowire field-effect transistors (SiNWFETs) are emploied to realize rapid ASTs, and a novel suspended SiNW-net sensor is also developed as a potential device platform for bacterial motility detection.

The thesis first introduces SiNWFET sensors for rapid ASTs. The extracellular pH change generated by bacterial metabolism is an efficient indicator of bacterial activity, which is monitored by our SiNWFET sensors. Rapid ASTs are achieved by using SiNWFET sensors array with a total assay time of less than 30 minutes for different bacterial strains. As a follow-up, the metabolic response of E. coli under ampicillin treatment is systematically studied. When exposed to bactericidal antibiotics, the bacterial respiration rate will be accelerated, thereby enhancing the lethality of the antibiotics. This work demonstrates the capabilities of SiNWFETs for rapid ASTs and bacterial metabolism investigations.

To further improve the detection limit of SiNWFET, Schottky junction gated SiNWFET (SJGFET) is developed, in which the noisy Si channel/gate oxide interface is replaced by a PtSi/Si junction. Ultra-low low-frequency noise is demonstrated in SJGFETs fabricated on high-quality bonded silicon-on-insulator (SOI) substrate. The best achieved S_vg are 1.2 × 10⁻¹⁰ and 1.1 × 10⁻¹¹ V²μm²/Hz at 1 Hz and 10 Hz, respectively. Then, a thorough investigation of low-frequency noise (LFN) is performed using the CNF + CMF model specifically modified for SJGFET structure on SOI substrate. The observed LFN dependence on substrate voltage and channel width is mainly ascribed to the nonuniform energy distribution of interface traps.

For the purposes of bacterial mobility detection, we propose a novel SiNW-net-based nanoelectromechanical sensor, with a 30-μm suspended SiNW-net and a metal Lorentz loop stacked on top. The lock-in amplifier measurement setup is optimized to significantly reduce the system noise. During rapid thermal processing in the device fabrication, lateral boron autodoping is discovered, which happens via ambient diffusion limited by the redeposition step at the interface. This technique enables shallow junction formation and will be integrated into SiNW-net fabrication to form well-controlled piezoresistors.