Gregg Bartlett, senior vice president, process technology and R&D, GlobalFoundries“We’re in the midst of a technological revolution tied to social networking and connectivity,” said Gregg Bartlett, senior vice president, process technology and R&D at GlobalFoundries. “Between 2009 and 2015, the industry is going to see a 50× increase in mobile data communication due to the increasing popularity of portable devices.”
A consequence to this boom in growth, Bartlett told the overflow crowd at yesterday’s keynote, “The Centrality of Silicon: How Semiconductors are Driving the Next Wave of Innovation,” is there will be a messy area where tablets, smart phones and netbooks will be competing with the traditional laptop and PC segments. Bartlett isn’t sure what the ultimate victor in this saturated market will be, but he says consumers will vote with their dollars.
Semiconductors have played a key role in the development of smaller devices, thanks to scaling, and will continue to be a dominant factor in the design of future products, Bartlett said. To meet consumer expectations, designers must turn to more advanced technologies, but that doesn’t come cheaply. “There’s a nearly exponential cost increase in bringing advanced performance to the marketplace,” Bartlett said. As a result, design starts have slowed over the past few years and this raises the stakes to a make-or-break point for the devices that do make it to market. R&D costs are rising, he pointed out, but so is revenue from advanced technology.
One of the most significant innovations that have taken place in CMOS fabrication since the inception of silicon VLSI is high-k/metal gate (HKMG) transistor technology, Bartlett said. “This is a truly revolutionary technology that enables continuous scaling benefits, and we’re confident in our ability to bring it to market, but the challenges associated with a new approach to building transistors can’t be ignored.” This is where collaboration becomes paramount to HKMG’s success.
Cost-effective 3-D stacking will also enable significant improvements in overall semiconductor performance, Bartlett added.
“The industry needs cost-effective EUV manufacturing, and failure here is not an option,” Bartlett said.
“Our strategy is to move past the pre-production tool step and straight to purchasing a production-level tool for installation in Fab 8 — our new leading-edge fab currently under construction in upstate New York,” Bartlett said. “We are planning to install this tool in the second half of 2012 so we can immediately begin the development work to enable volume production by the 2014/2015 timeframe. It is our collaborative approach to R&D that has put us in a position to make such a move — a move that will accelerate the charge to volume production for the entire industry.”
GlobalFoundries will be able to take what it learned from bringing immersion lithography to high-volume production and apply it to ramping EUV lithography to high volume, according to Bartlett. “From our perspective, we see immersion lithography getting us through the 22/20 nm node, but not without some serious cost challenges and added complexity,” he said. “We need another solution, and in our view EUV is the most promising candidate.”
Bartlett called on the industry to rethink the traditional foundry model. He emphasized a global manufacturing network instead of geographical concentrations in the Asia-Pacific region. He noted how GlobalFoundries has fabs in the United States, Germany and Singapore.
Collaboration is also going to be a huge factor in the future of semiconductors, and Bartlett broke down the combined efforts into three aspects:
- Early stage — education, research, technology development and implementation and the materials and equipment community.
- Design — IP/EDA, concurrent ecosystem solution development and leveraging shared learnings from early adopters.
- Manufacturing — fab alignment with key partners and OSATs.
He cited the IBM Alliance — working jointly with IBM, Toshiba, Samsung and others — as an example of what collaboration will look like in the future. “These organizations will share the capital expenses,” Bartlett said. “But more importantly, they will share intellectual capital.
— Arthur Patterson, SEMICON West Daily News