The Role of Models in Semiconductor Smart Manufacturing

The topics of "smart manufacturing" and related technologies such as the Internet of Things have gained a lot of attention in a wide variety of industries over the past couple of years, and semiconductor manufacturing is no exception. In fact, because of its long-standing emphasis on equipment/process data collection and automation, our industry is exceptionally well positioned to serve as a "thought leader" for this increasingly important domain. As device sizes and process windows have shrunk continuously according to Moore's Law, the SEMI Standards that govern the interaction between the manufacturing equipment and factory information and control systems have necessarily evolved to support the insatiable need for data exhibited by the process analysis and control systems that keep a modern fab running profitably. At the heart of these standards are models that represent the capabilities and behavior of the systems in question. In the earliest communications standards, these models were implicit in the message content, but over time, they have become much more explicit and detailed (see Table 1). The culmination of this process to date is captured in the SEMI E164 "Specification for EDA (Equipment Data Acquisition) Common Metadata" standard which specifies most of the expected structure, information content, and behavior of 300mm wafer fabrication equipment (see Figure 1). Since two of the key design principles for smart manufacturing are "interoperability" and "information transparency," one could assert that any factory that leverages explicit equipment models in creating generic manufacturing applications is a long way down the path of building a " smart factory." Supporting such an assertion is the precise objective of this presentation.