When it comes to defects and contamination in the semiconductor manufacturing industry, most people tend to think of small, sub-nm defects at the transistor level. As important as those are, there are plenty of things that can go wrong and be seen at the macro level. Scratches, fingerprints, hot spots, spin defects, edge chips, poly haze, missing patterns, etc. are usually visible with the naked eye, perhaps aided by a green light or a microscope.

Fabs often do manual visual inspections, but it tends to be fairly random, only sampling a few wafers at a time. “You put some wafers on the screen, and you look sporadically at five, ten points on a few of the wafers,” notes Reiner Fenske, founder, CEO and president of Microtronic (Hawthorne, NY). “If you find something, typically it’s very difficult to feed that information forward. You might take a picture, but then where does that picture go?” It’s also difficult to compare defects, such as scratches, with previously seen defects. “How many scratches did you have last week? Does that scratch look like the one that you had last night?” Reiner asks.

An automated macro inspection tool – such as the newly released Microtronic EAGLEview 5, which will be running wafers at North Hall Booth #5467 at Semicon West this week — solves those problems, without requiring any recipes and quickly scanning every wafer in the cassette, noting and logging various defects. The EAGLEview 5 represents a big upgrade over the company’s previous offering. “There’s really a dramatic difference in terms of defect detection, defect resolution, defect sensitivity, and there’s no hit to throughput, so we’re still looking at 3,000 wafers a day, which is incredibly fast,” said Mike LaTorraca, Microtronic’s Chief Marketing Officer. Errol Akomer, Applications Director at Microtronic, adds that in addition to the higher resolution, it’s a much cleaner signal. “The signal-to- noise ratio is much better — there’s a 5X improvement in that as well,” he said. Internally developed software algorithms also results in less nuisance defects and increased defect detection.

With these new capabilities, LaTorraca said they’ve created a bridge between micro and macro, and manual and automated. “We can take manual microscope images and put them into the same software that runs on EagleView. We can start to integrate defect information and the actual defect images from the manual microscope world into our tool, and that gives the fab owners a much more unified approach, a better, more comprehensive view, to make better decisions,” he said.

EAGLEview 5 is equipped with advanced imaging technology, analytical software, robotics and a 4-cassette multi-size (100mm-300mm) wafer platform. EAGLEview ProcessGuard Client Software provides defect visualization, digital guard-banding, wafer randomization/slot positional analysis, together with integration with manual microscopes for fab-wide defect tracking and reporting.

Every wafer is automatically OCR read, imaged, 100% inspected and stored for any step throughout the manufacturing process providing a comprehensive, centralized record – or ‘waferbase’ – that is also compatible with the fab’s manual microscope inspection data providing a more integrated, wholistic view of both micro and macro defects.

EAGLEview 5 acts as a hub for defect management across the fab by integrating manual microscope inspection, high resolution EAGLEview wafer images. EAGLEview 5 replaces legacy manual/micro wafer inspection by automating and standardizing wafer inspection processes. Blindly sampling 5 sites on a wafer is no longer needed. The newly developed ProcessGuard microscope interface software records micro defect classifications. This coupled with on-board commonality analysis allows root cause to be determined for micro defects and breathes new life into existing microscope inspection strategies. EAGLEview was originally designed to be comparable to naked eye 1x green light inspection.   EAGLEview 5 shifts the line between a macro green light inspection and microscope inspection.

“You can put all the micro defects into our database in the same ways you did the macro, so you classify your macro defects and you classify all your micro defects,” Fenske explained. “Now you have a record of what, where, how many, and because we collect all the history of where the lot went to, which tools it went through, we can then use that information to do commonality studies to figure out which tool caused the problem. With the microscope, there hasn’t been that type of integration, so we can now take all of those legacy things everyone needs to use and actually give them a new life.”