| The vigorous development of electronic and information science cannot be separated from the progress of Integrated Circuits(IC)related technologies.Following Moore’s Law,the rapid growth of modern IC technology has helped industries continue to break through traditional technical barriers and improve people’s lives in many aspects,such as Artificial Intelligence(AI),Internet of Things(Io T),and smart medical,and so on.With integration of chips increasing,reliability issues are becoming more and more serious.Among them,Time-Dependent Dielectric Breakdown(TDDB)is always one of the critical reliability issues of Metal-Oxide-Semiconductor Field-Effect Transistors(MOSFETs).When the IC process iterates to the deep submicron node,the reliability margin of ultra-thin oxide under DC TDDB is depleted,while the phenomenon of lifetime gain found in AC TDDB can preserve it.Furthermore,the reliability assessment technique of DC TDDB gradually loses its practical value and guidance because it does not match the stress patterns in digital circuits.Therefore,the current research should focus on the reliability assessment methodology,failure mechanisms,and the lifetime model of AC TDDB.The main contents and research results of this dissertation are listed below:(1)In this work,failure mechanisms of DC,AC and HCI-coupled TDDB are summarized to clarify the physical nature of defects generation under different stress patterns.After that,this work establishes a wafer-level AC TDDB characterization system based on fast measurement techniques to characterize AC TDDB reliability behaviors accurately.Further,the test methodologies and failure criteria are improved.(2)Based on the technique mentioned,this work evaluates the lifetime gain of Ultra-High Frequency(UHF)AC TDDB in MOSFETs.It is found that lifetime gain and voltage acceleration at GHz stress can significantly improve the reliability margin of devices.Further,using GHz stress to weaken the Self-Heating Effect(SHE)and Hot Carrier Degradation(HCD),a scattering-related degradation mechanism is proposed,based on the location distribution of the breakdown path,in the case of MOSFET on-state TDDB.(3)This work also explores the AC TDDB lifetime gain under different frequencies,Vlow,and duty factors.And the contribution of failure mechanisms,such as tunneling,trapping,de-trapping and recovery,is quantitatively analyzed on the time scale.Finally,a lifetime model based on defect accumulation efficiency is proposed for predicting AC TDDB at frequency domain.This dissertation has two novelties:First,it sets an automated system for TDDB characterization applicable to various stress patterns,allowing the extension of stress frequency up to GHz.Secondly,for the first time,a frequency-dependent lifetime model of AC TDDB based on defect accumulation efficiency is proposed,providing sufficient guidance for future TDDB reliability studies of new materials and devices. |