RequestLink
MICRO
Advertiser and
Product
Information

Buyer's Guide
Buyers Guide

tom
Chip Shots blog

Greatest Hits of 2005
Greatest Hits of 2005

Featured Series
Featured Series


Web Sightings

Media Kit

Comments? Suggestions? Send us your feedback.

 

MicroMagazine.com

EDITOR'S PAGE

Pushing the productivity paradigm

The pressure to balance technology and economics continues to escalate in the chipmaking world. There is growing concern that some of the productivity challenges facing the industry could hamper the adoption of new technologies. This was a main theme of International Sematech COO John Schmitz's keynote at the Loomis Newspanel at Semicon West in July.

He listed four key challenges the industry needs to address by 2009: achieving non-product-wafer-free manufacturing (no more test or monitor wafers); providing true plug-and-play equipment (tools that can be up and running in a matter of hours, not weeks or months); shortening cycle times (room for at least a 2X improvement in even the best-run fabs); and reaching 100% die yields. Schmitz believes that these goals can be accomplished, which could lead to significantly increased output and drastically reduced work in process (WIP). As he quipped, "getting more chips out of an expensive fab is a good thing." To hit these productivity metrics, he cited the need for good fab automation with much better advanced equipment and process control (AEC/APC), improved WIP control, and sound, data-driven, information-based decision making.

One familiar entity seems to have gone a long way toward achieving Schmitz's goals well before 2009—Advanced Micro Devices. The company recently announced its "smooth transition to 90-nm manufacturing" and revenue shipments of low-power, silicon-on-insulator (SOI)-based Athlon 64 processors for notebook computers from Fab 30 in Dresden. One of the keys to this successful 90-nm ramp, AMD says, is its proprietary Automated Precision Manufacturing (APM) capability.

APM is much more than your father's fab automation, where robot arms schlepp wafers in and out of process chambers, guided vehicles trundle between tool bays along elevated monorails or guide rails on the production floor, and stockers brim with queued-up WIP. In fact, materials-handling systems play a relatively small part in the scheme.

"In these highly automated fabs, the ability to automate the decision-making process is much more important than the ability to just automate the material movement processes," explains AMD's director of APM, Thomas Sonderman. Version 2.0 of APM's sophisticated control algorithms connect APC, yield management systems, integrated production scheduling, tool performance optimization, and product performance targeting, thus "linking the entire fabrication process together and treating it holistically," according to Sonderman. "We control every lot in the fab even though we don't measure every lot."

Citing specific examples with etch, rapid thermal annealing, and copper plating equipment, Sonderman says that APM's "very tightly controlled manufacturing environment" allows fab workers to "see things that are interacting . . . to drill down to something actually going on in a tool. This is one of the things that we believe not only allowed us to rapidly increase and actually get to high yields as we introduced a technology, but also to not get lost in the world of interactions that is causing a lot of people pain at 90 nm."

Sonderman says that development of APM version 3.0 is on schedule to roll out in conjunction with the ramp of 65-nm processes at AMD Fab 36, the company's first 300-mm facility, now under construction in Dresden. What does he see as APM's biggest challenge? "Integrating the MES (manufacturing execution system) into 3.0 and then adding capabilities to internally developed products, such as recipe management and statistical process control . . . and also integrating third-party yield management systems (YMS) and production scheduling software . . . with our internal stuff."

He lauds the integration testing going on at AMD's automation labs in Austin and Dresden. "We feel good to drive the roadmap on full wafer-level control, comprehensive adaptive sampling, and fully integrated fault detection and run-to-run infrastructure, [as well as] a much more die-level-control-based architecture for YMS." Sonderman points out that such extremely dialed-in levels of control become even more important with the integration of strained SOI, metal gates and other advanced transistor materials, and ever-more-complex process interactions.

There's certainly room for skepticism about AMD's robust claims. The word "profits" doesn't appear very often in the company's historical financial reports. Whether you buy into the hype or not, the company seems to be trying to profoundly change the way chips are made.

Tom Cheyney
Editor

tom.cheyney@cancom.com


MicroHome | Search | Current Issue | MicroArchives
Buyers Guide | Media Kit

Questions/comments about MICRO Magazine? E-mail us at cheynman@gmail.com.

© 2007 Tom Cheyney
All rights reserved.