Reconvergent specular detection of material defects on silicon

Abstract

The ability to detect and classify material defects, such as epitaxial stacking faults, and pits in the surface region of silicon substrates is rapidly gaining importance as an additional wafer evaluation criterion. Traditionally, surface scanning inspection systems (SSIS) have been utilized to detect and quantify particulate contamination in process control applications. This study investigates the ability to detect and image these material defects using an enhanced SSIS. The imaging apparatus studied for this unique method operates concurrently with the conventional mode of operation for particle detection which relies on light scattering events. In the case of imaging material defects, a loss in the reflected light source beam intensity is measured. This technique, reconvergent specular detection (RSD), samples the reflected beam and is more commonly known as light channel detection. Stacking faults, pits, and slurry residue, common defect features on silicon, are examined in this study using the light channel detector. This work establishes that these types of defects are difficult to quantify when restricted to conventional detection methods. The light channel detection method, however, is capable of accurately imaging these defects according to size and shape. This paper highlights these results explains light channel phenomena in terms of detection theory and defect surface area. This novel imaging method offers a means of both detecting material defects and classifying them.