The study of the economical scale of a semiconductor plant through simulations

The semiconductor manufacturing is complex due to the re-entry phenomenon. The semiconductor fab scales are larger and larger to avoid machine hungry and block penomena. Larger scale fabs bring up machine utilizations, meanwhile, they make the automated material handling system (AMHS) complicated. This results the poor performance of the AMHS. Semiconductor people invest a large amount of capital in the AMHS control & management system to prevent from the plant size effect. This reduces the benefit of the large fabs. Therefore, what is an optimal semiconductor fab scale needs to be discussed.
The main purpose of this study intends to evaluate the benefit from the step-by-step enlarged semiconductor fab scales and find the optimal scale of the plant by simulations. The simulation results show the cycle times and WIPs decrease gradually as the fab scales increase before 6 times. The 6-time scaled fab has relatively high cycle times and WIPs, and the 7-time scaled fab system cannot be stable. This indicates an optimal fab size exists. Besides the system performance, an AMHS cost analyses would also be discussed in this study. The unit wafer material handling cost is increased when the fab size is too large.


Keyword：semiconductor wafer fabrication, AMHS, cycle time, WIP, throughput, material handling cost