Bias Instability, Radiation Effects, and Low-Frequency Noise in Semiconductor Devices
Luo, Xuyi
0000-0002-1478-9431
:
2024-05-22
Abstract
As microelectronic devices have become more integrated and compact over time, ensuring their reliability has become a critical challenge. This dissertation investigates the bias instability, radiation effects, and low-frequency 1/f noise (LFN) in semiconductor devices, specifically focusing on Ge pMOS FinFETs, GaAs pseudomorphic high-electron-mobility transistors (PHEMTs), and Si bipolar junction transistors (BJTs). In Ge pMOS FinFETs with high-K dielectrics, positive interface-trap generation is the dominant defect responsible for negative-bias-temperature stress (NBTS). The LFN measurements indicate that the defect energy distributions before and after NBTS are increasing toward midgap in these devices. Newly created and/or activated border traps after NBTS related to oxygen vacancies and their complexes with hydrogen. In contrast, GaAs PHEMTs demonstrate robust device performance under electrical stress. First-principles calculations and comparisons with previous work suggest that OAs impurity centers, other oxygen-related defects, isolated AsGa antisites, and dopant-based DX centers may contribute significantly to the LFN in these devices. We also evaluate the degradation and the nature of radiation-induced defects before and after Si ion irradiation of n-p-n Si BJTs through both deep-level transient spectroscopy (DLTS) and LFN measurements. The DLTS measurements identify three prominent classic defect levels in the bulk Si that are introduced by irradiation in the base-collector junction of these transistors. Additionally, a combination of contributions from oxygen vacancies and hydrogen complexes in the oxide that overlies the base-emitter junction is inferred from the temperature-dependent LFN measurements. Our work contributes to a broader understanding of the degradation mechanisms and reliability issues in semiconductor devices across different materials and device architectures. For future research, LFN measurement remains a valuable tool for enhancing our understanding of defect densities and energy distributions in microelectronic devices.