Gigahertz Scanning Acoustic Microscopy

In the ever-evolving semiconductor industry, advancing metrology techniques is crucial to meet the demands of shrinking feature sizes, increasing device complexity, and improving production yield. As transistors continue to scale down into the nanometer regime, precise, non-destructive characterization methods that can provide high-resolution, three-dimensional imaging are essential for quality control and process optimization. Traditional metrology techniques, such as optical microscopy and atomic force microscopy (AFM), face limitations in terms of resolution, depth penetration, and speed, especially as semiconductor architectures grow more intricate.

GHz ultrasound microscopy (GUM) has emerged as a promising alternative for high-resolution, non-invasive metrology in semiconductor technology. Operating at frequencies in the GHz range, GUM offers the ability to probe sub-surface features with nanometer-level precision, overcoming some of the limitations of optical diffraction and providing deeper penetration into semiconductor materials. Its acoustic nature allows it to map mechanical properties like elasticity, density, and layer thickness, providing crucial insights into the structural integrity and quality of semiconductor devices at the nanoscale. Today, its low-frequency counterpart, i.e Scanning Acoustic Microscopy (SAM), has become the reference technique for void detection and hence turned into a critical inspection technique for 3D integration and bonding applications. However, its detection limit is being challenged as the hybrid bond pitch scales down. We foresee GUM to be an obvious contender to replace SAM as sub-µm voids become killer defects.

This PhD proposal aims to explore the application of GHz ultrasound microscopy for metrology in semiconductor fabrication processes. Building on the development of integrated high frequency leaky ultrasound resonators, we propose to develop compact GHz SoDAR heads, enabled by monolithically integrated support electronics. This development will pave the way for the mass parallelization of the unique capabilities of GUM. Tackling the throughput issue of GUM, this research will further focus on developing new methodologies to enhance imaging resolution, accuracy, and speed, while addressing current challenges such as signal-to-noise ratio and sample interaction. The goal is to establish GHz ultrasound microscopy as a viable and indispensable tool for the next generation of semiconductor metrology, contributing to the advancement of nanoelectronics and ensuring reliable production in future semiconductor technologies.