Revolutionizing Semiconductor Material Specifications

TL;DRAbstract

Statistical process control is being widely adopted to improve yields and profitability in semiconductor manufacturing. Controlling and reducing key process parameter variability is essential for yield improvement and cost control. Accordingly, statistically-controlled processes must be supplied with raw materials which are similarly statistically controlled. This need has been inadequately recognized by vendors and standards-writing groups. This paper discusses the importance of controlling materials variability. Examples from IC manufacturing are given. It is shown that raw material specifications cannot be based solely on engineering limits, but must include centered means, statistical control limits, and the requirement that vendors deliver their products from their own in-control manufacturing processes.

Chat with Paper

AI Agents for this Paper