Chip manufacturing information is rich and complex, and bringing this information together has substantial benefits. Data variety, sources, quality, and volume are some of the characteristics which make this task challenging. SmartFactory has built a unified process control (UPC) solution that brings together in one platform the different principles of process quality, specifically pertaining to monitoring product. The unified platform takes data from multiple sources that speak a different language and, for the first time, transports them into composite data to have more insights into the factory.
We chose to build the solution on the Applied E3® equipment and process control platform, which has all the core capabilities and is something most customers already use. The resulting ecosystem provides models and linkages to E3 that make it possible to move to automated decision making (AI-enabled decisions) for process quality.
The UPC derives artifacts from many pieces of equipment rather than looking at data in a silo. This is key, since there are implications for process quality coming from each of those different artifacts, and looking at them holistically provides deeper insights. The solution brings together relevant data from multiple sources, looks at it in a combined fashion and interprets what happened, what needs to be done and how. Its interpretive vision identifies an accident as well as the specific, combined effects that led to the problem, then tells you what to fix and how.
SmartFactory’s unified platform was designed with only the required resources to provide insights, carefully filtering noise that wouldn’t have an impact on process quality. Importantly, the solution does not take a cookie-cutter approach. Because the models must be reliable and relatable to the individual semiconductor factory, they are designed one-by-one. The user can input information specific to the factory, as well as identify protocol for managing situations.
Take for example, a factory where many alarms have come from the same source in a short period of time. Each is fixed in turn, but the source continues to trigger alarms. In the fast-paced environment, manufacturers can’t afford to stop production to analyze this extensively. Fixing each alarm’s perceived cause is a band aid approach to get back to production, but it clearly isn’t working. In this case, the UPC does the work without interrupting production. It looks at the history of that source over several different shifts to see that it’s a recurring problem. The model gathers and interprets data across several shifts, looking at multiple factors that may impact process quality. It integrates the artifacts, reads relationships between them, and finds the root cause of the alarms, finally proposing a remedy.
In this way, manufacturers can make business decisions faster. Additionally, the model becomes a continuous improvement tool. Once it has identified the solution to one problem, it can move on to other process quality concerns to analyze combined data toward another solution.
This connectivity via a unified platform can help you unleash the full potential of your fab by harnessing the synergy of integrated tools and data to simplify operations and empower your teams.