Quantum Diamond Microscopy as a Non-Destructive Method for Short Localization in High-Density 2.5D MIMCAP

Abstract

The increasing demand for compute power is pushing system scaling toward advanced 3D and backside-integrated architectures. While the wafer backside opens a dynamic design space with new opportunities to optimize power delivery in scaled systems, it also introduces significant challenges for failure analysis (FA). In particular, the application of traditional backside optical fault isolation techniques, widely used for decades, becomes increasingly limited at current leading-edge nodes. As a result, a variety of alternative fault-isolation techniques, including X-ray, thermal, magnetic, and electron-beam approaches, are actively used and developed. Among these, magnetic-field-based techniques enable non-invasive access to buried current paths and connectivity defects, owing to the ability of magnetic fields to propagate through insulators, semiconductors, and metals with minimal disturbance. Early implementations of magnetic field imaging in FA predominantly relied on scanning-probe approaches based on SQUID and GMR sensors. More recently, however, quantum diamond microscopy (QDM) based on nitrogen-vacancy (NV) color centers has gained significant attention, as it enables quantitative current imaging through vector magnetic-field measurements with full-field, parallel, camera-based readout.