Optical Surface Analysis Advances Defect Inspection of Optoelectronics
- Created on Monday, 01 May 2006
Production defect data leads to better yield management practices.
KLA-Tencor Corp., San Jose, California
Faced with increasing demand, manufacturers of power devices, microdisplays, and high-brightness light-emitting diodes (HB-LEDs) are focusing on tightening manufacturing process windows to reduce defects. The transparent nature of the substrates used to make many optoelectronic devices such as glass, silicon carbide, and sapphire makes manual defect inspection using optical microscopes an ambiguous and time-consuming process incapable of high-volume production. To meet the need for improved defect inspection, Optical Surface Analyzer (OSA) instruments provide automated defect inspection for optoelectronic device wafers from 2" to 300 mm in diameter.
The design of the OSA instrument involves reflecting linearly polarized light from the sample surface and detecting reflections with multiple detectors. Thousands of data points are measured in a typical scan. The signals are analyzed to provide information about defect types, sizes, and numbers. Special software algorithms allow definition and classification of custom defect types. Measured signals also can be viewed as images for engineering troubleshooting purposes. Figure 1 shows three examples of common defects from wafer handling and cleaning processes, such as stains and surface particles. These defects were found on a sapphire substrate with a GaN epi layer used for HB-LED production.
For SiC power devices, epi layer growth faces several defect hazards: films, residues, or particles on the substrate can prevent epi growth, and lead to pits circling the original defects; particles falling on the substrate during growth can cause smaller pits; and scratches on the substrate cause deformed epi surfaces that follow the scratch outlines. These defects are all easily detected by OSA technology.
In microdisplay devices such as CMOS imagers and LCoS displays, defects have the potential to create a blurry image on the final device output or shorten the life of the device. Figure 2 shows an OSA image of rubbing marks found on a polyimide over ITO coating on a glass wafer. The glass substrate contained a scratch that resulted in deposition defects. OSA instruments can be used to inspect wafers both before and after film deposition.
As process technologies mature, relying on microscope inspection for defect detection and yield enhancement is no longer adequate. Automated defect inspection techniques with OSA provide defect density and classification information for the entire wafer surface instead of a small sample. Yield enhancement and process improvement activities benefit from having a large quantity of data generated in a consistent way.