The growing complexity of manufacturing chips, displays and solar cells together with higher investment costs of advanced production facilities are changing service and support requirements.
The service business is still driven by the need to keep systems up and running: identifying mechanical problems, adjusting hardware, replacing parts, etc. However, today’s semiconductor fabs are also under increasing pressure to accelerate and maximize yield, reduce cost and improve productivity. Active fault detection and excursion control on every module is becoming increasingly vital to deliver factory output requirements.
In this blog post, I discuss Applied Global Services’ (AGS) service model which plays a vital role in meeting today’s rigorous manufacturing requirements, helping to move the industry forward.
Leading edge semiconductor fabs require increasingly sophisticated service support techniques to manufacture complex new architectures like FinFET and 3D NAND structures. These designs involve new precision materials, more process steps, much narrower process windows, and many more interacting variables. However, legacy fabs are also under new pressure to manufacture a broader variety of devices such as analog and power chips, image sensors, and MEMS needed to support the move to enhanced mobility, the Internet of Things (IoT), and increased automobile safety standards – to name a few. Meeting performance and yield goals for these devices requires state-of-the-art service technologies, advanced service expertise with deep tool knowledge, and an unprecedented level of collaboration and trust between the supplier and the customer.
Increasingly, fab managers rely on advanced technology to detect, classify, diagnose, control and predict various failure modes that can impact on-wafer results – and also impact cost and productivity. Today’s advanced capabilities monitor and fine-tune processes for optimum performance, and help customers move to more predictive operations. These technology-enabled process controls may include data mining analysis, active fault detection and classification (FDC), run-to-run control and statistical process control technologies.
These monitoring systems generate an enormous amount of process- and equipment-related data, driving the need for sophisticated analytics and data mining software to rapidly distill data and turn it into actionable information. This practical intelligence makes it possible to shorten development time, minimize process variations for higher yield and diagnosis root-cause for accurate corrective action. By turning this veritable ‘flood’ of data into usable information, critical correlations can be determined among the data and move semiconductor manufacturing towards a more predictive analytics. This approach minimizes unscheduled tool downtime for specific, tracked issues; saving cost and improving productivity.
To help customers achieve such productivity and yield improvements, service technologies and solutions have evolved significantly. For example, in the case of CMP scratches, Applied leveraged advanced analytics to predict and prevent scratch events, driving significant yield improvements. And with new techniques in large scale metrology data mining, Applied can help process engineers streamline their work. By reducing the time for some investigative tasks by as much as 80%, data mining allows engineers to pinpoint problems more quickly and focus on developing and deploying solutions.
These innovations are opening doors to Applied’s new service model that extends beyond traditional equipment maintenance practices. Service levels today and going forward must address increased semiconductor process complexity by taking advantage of the information gained through these new methodologies to better serve customers.
