A novel material detection method using femtosecond laser, confocal imaging and image processing enabling endpointing in fast inspection of microelectronics

Abstract

Ultrashort pulsed lasers can offer athermal, fast and precise material removal as needed for inspection, failure analysis, and reverse engineering of microelectronics. Laser parameters must be fine-tuned, based on material type, for best results. However, often, accurate information about the material composition of sample of interest is not at hand and thus the material composition must be inferred. Furthermore, endpointing of the deprocessing system becomes increasingly difficult due to the large material removal rate of laser systems. Integration of current methods, such as energy dispersive spectroscopy (EDS) and laser induced breakdown spectroscopy (LIBS), is expensive and complex. We propose a novel technique that detects material composition based on surface texture parameters derived from confocal images of a lasered area. A multilayer fully connected neural network is used, which after sufficient training can predict the material composition of the sample with a single image of the surface.