The advancement in the field of 3D integration circuit technology leads to new challenges for quality assessment of interconnects such as through silicon vias (TSVs) in terms of automated and time-efficient analysis. In this paper, we develop a fully automated high-efficient End-to-End Convolutional Neural Network (CNN) model, utilizing two sequentially linked CNN architectures, suitable to classify and locate thousands of TSVs as well as provide statistical information. In particular, we generate interference patterns of the TSVs conducting a unique concept of Scanning Acoustic Microscopy (SAM) imaging. Scanning Electron Microscopy (SEM) is used to validate and also disclose the characteristic pattern in the SAM C-scan images. By comparing the model with semi-automated machine learning approaches its outstanding performance is illustrated, indicating a localisation and classification accuracy of 100% and over 98%, respectively. The approach is not limited to SAM-image data and presents an important step towards zero failure management.

Nanocrystalline alloy thin films offer a variety of attractive properties, such as high hardness, strength and wear resistance. A disadvantage is the large residual stresses that result from their fabrication by deposition, and subsequent susceptibility to defects. Here, we use experimental and modelling methods to understand the impact of minority element concentration on residual stresses that emerge after deposition in a tungsten-titanium film with different titanium concentrations. We perform local residual stress measurements using micro-cantilever samples and employ machine learning for data extraction and stress prediction. The results are correlated with accompanying microstructure and elemental analysis as well as atomistic modelling. We discuss how titanium enrichment significantly affects the stress stored in the nanocrystalline thin film. These findings may be useful for designing stable nanocrystalline thin films.

The advancement in the field of 3D integration circuit technology leads to new challenges for quality assessment of interconnects such as through silicon vias (TSVs) in terms of automated and time-efficient analysis. In this paper, we develop a fully automated high-efficient End-to-End Convolutional Neural Network (CNN) model, utilizing two sequentially linked CNN architectures, suitable to classify and locate thousands of TSVs as well as provide statistical information. In particular, we generate interference patterns of the TSVs by conducting a unique concept of Scanning Acoustic Microscopy (SAM) imaging. Scanning Electron Microscopy (SEM) is used to validate and also disclose the characteristic pattern in the SAM C-scan images. By comparing the model with semi-automated machine learning approaches its outstanding performance is illustrated, indicating a localisation and classification accuracy of 100% and greater than 96%, respectively. The approach is not limited to SAM-image data and presents an important step towards zero defect strategies.

Nanocrystalline metallic alloy thin films provide a variety of interesting properties. A disadvantage concerns their inherent instability and generation of residual stresses. Here, we present a unique framework incorporating machine learning assisted experimental and modelling methods to perceive the impact of the minority element concentration on the generated residual stresses within a nanocrystalline thin film. As a candidate system we use a W-Ti alloy thin film with different Ti concentrations. We perform machine-learning assisted local residual stress measurements, correlate the results with accompanied microstructure and elemental analysis and apply density functional theory. We inquire why the experimental observed Ti enrichment can be strongly reduced at smaller concentrations and discuss how it significantly effects the stress stored in the nanocrystalline thin film. The presented perception is highly crucial to yield future design guidelines for more stable nanocrystalline thin films.