By ED SPERLING
Semiconductor equipment vendors are starting to add more sensors into their tools in an effort to improve fab uptime and wafer yield, and to reduce cost of ownership and chip failures.
Massive amounts of data gleaned from those tools is expected to provide far more detail than in the past about multiple types and sources of variation, including when and where that variation occurred and how, when and why equipment failures occur. Combined with data about device failures in the field, along with such things as design layout and verification, it’s becoming possible to create a detailed timeline of how chips were designed, manufactured and what goes wrong along the way. That, in turn, can be used to improve quality, identify potential sources of defects, and add increased efficiency into processes.