Solutions

 

The latest version of Basler's solar inspection system enables you to detect not only microcracks, but also other types of defects, for example foreign-body inclusions. A finely optimized Classifier allows you to classify any suspect areas in accordance with various mathematical criteria to distinguish between different types of defects such as microcracks or foreign-body inclusions. The system is therefore able to sort the defective wafers according to the type of defects.

The Classifier used in this system has been successfully installed together with other inspection systems from Basler and, over the years, developed into an extremely efficient tool. Special mathematical criteria have now been applied to optimize the Classifier and enable it to distinguish between the different types of defects encountered in wafers. Another tool aided Basler application engineers in their efforts to optimize the Classifier: the Basler OpClass Tool. Using this instrument, it is possible to train the Classifier with images of various types of defects. It determines which criteria are necessary for optimum detection of the different types, supplying a file with all relevant data that can easily be uploaded into the inspection system on the production line, either by remote access or in situ.

During the past months, the system was installed at the number of customers’ premises in order to optimize performance even further. For this purpose, defective wafers were systematically check-tested and the orientation of wafers passing through the system varied. The test results were subsequently analyzed, and the procedure was repeated with different development versions of the software to produce comparable results.

System performance is determined by three essential criteria: reproducibility, detection rate and pseudo defect rate.

Until now, results have shown a minor pseudo defect rate of 0.25%, an excellent level of reproducibility exceeding 80%, and a very good detection rate of over 70%. Countertests conducted on wafers rejected by the system revealed a distinct correlation of more than 50% between percentages of waste and inspection results. This means that it is possible to reduce the percentage of waste in a production line using the Basler inspection system.