Data Sharing and the Cost of Inaction
Why is the vacuum system so critical to Fab success?
The key reason is that many critical process steps in semiconductor manufacturing, need high vacuum or very controlled ambient conditions. So the vacuum is a key process parameter, and it is always on. Pumps are very rarely switched off on process tools. The vacuum system is also Larch and pervasive. There are many thousands of dry pumps, typically hundreds of abatement tools.
In the Fab itself, you will find hundreds of Turbo Pumps, Cryo pumps, compressors, then there are thermal management systems and other ancillary systems, and then also, miles and miles of fore lines and exhaust lines.
And finally, all the systems are inter-connected, which means that they will be affecting each other and affecting the process. It’s very important that we collect data from this critical system in the Fab.
In the previous slide, I mentioned that it is critical to collect data from the SubFab but is that enough to make the SubFab a driver for fab success.
Now, a core focus for many in the semiconductor industry is smart manufacturing, but what does that mean in the context of the SufFab?
The technical definitions around for smart manufacturing. Most will describe automated actions and interactions between machines and IOT, hardware, and software. So this feeds a vision of an autonomous manufacturing environment with very little human interaction.
But the SubFab is a complex environment where people’s people, machines, data and process just interact closely. So our challenge is less about achieving automation and more about bridging the gap between automated machines and creative humans who can solve complex problems to drive continuous improvement.
Our approach, therefore, is to reduce risk and uncertainty and create value by using smart manufacturing as described, as well as lean and circular economy principles to make machines, people, data, processes, work effectively together. And we call this operational excellence.
So what are the key challenges for the sub fab and the fab? If you want to make machines, processes, data and people interact effectively to reduce risk and uncertainty. I’ve split this up in three challenges.
You need to collect all the data from the vacuum and abatement equipment. This is a lot harder than it sounds. SubFab and Fab equipment is varied and there’s a lot of different suppliers and the complexity can also vary greatly.
There, you have a first mile problem. You need to have the network infrastructure to be able to connect to the equipment, but then further upstream, you also have a connection problem. If you want to connect to different software, process, and hardware maintenance platforms, FTC, NBS, etc
You need to combine data from different sources. And of course, this is where humans come in. You also have to account for dynamic process environment. And the dynamic process environment is driven by people and what they do mainly. And finally, the people will provide the subject matter expertise, and they will be able to use the domain knowledge supply chain, which you will need to turn all that data into actionable insight.
Of course, collection, connection, and contacts is really only the start of smart manufacturing. If we want to get the value from data, then we have to be able to turn that data into actionable insight and information.