The development of More than Moore (MtM) technology continues to drive the miniaturization of semiconductor devices and the integration of various functionalities. This MtM-driven trend increasingly emphasizes the need for 3D integration techniques to achieve superior miniaturization and improved performance. However, the rapid advances in 3D integration technology have brought new challenges in the area of failure analysis (FA). Overcoming these challenges requires the use of non-destructive methods that provide high-resolution insights while being cost-efficient and time-saving. The scanning acoustic microscope (SAM) is proving to be a crucial tool for non-destructive, cost-efficient and time-critical failure analysis. In this context, the imaging resolution and penetration depth of the conventional SAM are traditionally limited by the frequency of the transducer used. In this work, a modified SAM interferometry setup is tailored for defect detection in 3D integration technology, focusing on the characterization of through silicon vias (TSVs). This approach is more independent of the transducer frequency. Special acoustic lens configurations are used for the controlled excitation of surface acoustic waves (SAWs). The interference patterns resulting from the interaction of the SAWs and bulk acoustic waves contain valuable information about the quality of the TSVs. Artificial intelligence is used to localize and classify the interference patterns obtained and can be effectively used for error analysis of the TSVs. The different SAM interference patterns on the TSVs are investigated in detail in this thesis using acoustic simulation and validated using the scanning electron microscope (SEM). Essentially, this thesis contributes to the advancement of 3D integration technology by presenting innovative approaches for non-destructive defect analysis, automatic defect localization and comprehensive characterization of TSVs. Furthermore, it focuses on deepening the understanding of the behaviour of SAWs and their role in the context of quality detection in modern semiconductor technologies.