The project focuses on developing a hybrid metrology workflow that combines high‑resolution X‑ray tomography (XCT), laser‑scanning microscopy, ion‑beam surface preparation, and electron‑microscopy techniques to detect, classify, and quantify defects in advanced MEMS‑CMOS integrated structures. The primary goal is to replace costly destructive failure‑analysis methods with rapid, non‑destructive alternatives that can be applied during mass production.

X‑ray tomography is used to generate three‑dimensional images of through‑silicon vias (TSVs) and other critical interconnects. The XCT data are compared with results from partner‑provided samples, and a cross‑correlation analysis is performed to validate the imaging accuracy. The workflow is extended by incorporating the Hitachi IM5000 and Keyence VK‑X1000 laser‑scanning microscope. The Keyence system offers a 150× objective that delivers up to 28,000× magnification with a lateral resolution of 0.5 nm. Its variable illumination—coaxial, ring, and 404 nm UV laser—enables non‑destructive height and surface‑roughness measurements across large areas. The laser‑scanning principle involves rastering the beam in X‑Y, stepping in Z, and reconstructing a 3‑D surface map, which is then used to locate and characterize defects identified by XCT.

For surface preparation and defect isolation, the Hitachi ArBlade 5000 ion‑beam system is employed. Its Penning‑type ion source generates a stable argon plasma; ions are accelerated to 0–8 kV and directed at the sample surface. The sputter effect removes surface atoms without chemical alteration, allowing precise thinning of the region around a suspected defect. This preparation facilitates subsequent high‑resolution imaging.

When XCT and laser‑scanning data indicate a potential flaw, the sample is transferred to a transmission electron microscope (TEM) for confirmation. Energy‑dispersive X‑ray spectroscopy (EDX) and electron energy‑loss spectroscopy (EELS) are used to analyze elemental composition and bonding states at the defect site. These complementary techniques provide definitive evidence of defect type and origin, enabling the development of a defect taxonomy that can be fed back into the manufacturing process.

The project’s technical outcomes include a validated hybrid workflow that reduces analysis time by up to 50 % compared with traditional destructive methods, a defect classification scheme based on combined XCT, laser‑scanning, and TEM data, and a set of best‑practice guidelines for integrating these techniques into routine quality control. The high‑resolution imaging capabilities of the Keyence system and the precise ion‑beam preparation offered by Hitachi are central to achieving these results.

Collaboration is led by X‑FAB MEMS Foundry GmbH, a global leader in analog and mixed‑signal semiconductor manufacturing. The project is coordinated by two industry‑oriented partners: Cool Silicon in Dresden and CEITEC Nano in Brno. Funding is provided under the German grant code 03INT610BC, supporting the development of advanced failure‑analysis tools for the semiconductor sector. The partnership brings together expertise in foundry operations, advanced imaging, and ion‑beam processing, ensuring that the developed methods are both scientifically robust and industrially relevant. The project timeline spans several years, with initial tool demonstrations conducted before the main research phase, culminating in the integration of the hybrid workflow into X‑FAB’s production environment.