Application of machine vision technology for focused laser effect accuracy improvement in microelectronic structures research and semiconductor microprocessing

Machine vision is a field of artificial intelligence that deals with processing images and videos using special algorithms. This allows devices to analyze visual information. Machine vision helps with tasks such as pattern recognition, image segmentation, object detection, and tracking. In microscopy, machine vision plays an important role, particularly in laser scanning microscopy (LSM). Laser scanning microscopy, laser scribing, and laser correction of the topology of semiconductor crystals are important technological processes in the production, control and adjustment of semiconductor crystals both on debug samples and as part of a test batch on a wafer. Laser exposure allows not only mechanically separate crystals, but also carry out more delicate and minimally invasive actions, in particular, adjustments to the values of thin-film resistors or burning of bridges necessary to adjust the circuit and disable unused blocks of the crystal. The paper lists the main parameters of the positioning system as part of the laser scanning setup, their impact on the quality of scanning and exposure to focused radiation at control points. The principles of machine vision algorithms when working with the image of the scanned object and the results of testing in the task of automated laser burning of bridges on a semiconductor wafer are described. Since the number of bridges can reach tens of thousands, and positioning systems have significant errors, machine vision allows you to correct the area and accuracy of laser exposure at any stage of scanning, which significantly improves the quality and efficiency of the process.