Presenter: Mike Barrett, Director, Global OEM Sales
Executive Forum, Hall B2
16:10 CET | Wednesday, November 17, 2021
In response to global chip shortages, many semiconductor fabs have started to demand even higher throughput from the equipment on their manufacturing floors. While process timing is often constrained by physics, opportunities do exist to reduce wait time waste by optimizing the manner in which substrates are scheduled within complex tools, for example, in equipment that combines multiple operations within a single cluster.
However, scheduling wafer flows within a complex cluster tool – particularly in a high-volume, high-mix environment – presents several engineering challenges for optimizing wafer movements. Is it possible to avoid deadlocks and maximize throughput while enforcing strict process and recipe rules? When a tool’s components are operating in a degraded mode or are disabled for maintenance, can the equipment continue to perform efficiently? A tight labor market that requires OEMs to make tough decisions about where to allocate scarce expert developers compounds the challenge.
Advances in artificial intelligence (AI) and machine learning have introduced the possibility for automated solutions that can discover optimal routing in real time, replacing the effort for creating high-quality substrate schedulers manually. Such an approach would increase fab throughput while simultaneously reducing time to market and engineering effort for OEMs.
In this talk, PEER Group will compare three generations of cluster tool schedulers: Traditional, Offline optimized via machine learning, and Real-time equipment optimization using AI.