450mm wafers just a distraction say fab execs

TALK ABOUT MIGRATING to 450mm wafers is nothing but a "distraction" according to executives from several of the leading semiconductor foundry companies, speaking at Semicon West this week.

The economic downturn seems to have polarised chipmakers over 450-mm wafer production, with Intel, TSMC and Samsung advocating the migration as a way to innovate out of the crisis, whilst the other big fab firms - including Global Foundries, IBM, SMIC, UMC and Chartered - say the enormous R&D costs involved are just not worth the investment , or the bother, at the moment.

"People who advocate bigger wafers have run out of ideas," said Glofo's Tom Sonderman, adding that firms advocating the migration had "run out of options for making their fabs more effective."

The sentiment was echoed by executives from IBM as well as Singapore's Chartered Semiconductor Manufacturing and China's SMIC, with SMIC's Sam Wang even going a step further and proclaiming a market might even exist for refurbished 200-mm equipment capable of 90nm production or below.

Wang explained that a number of 200-mm fabs were due to shut up shop this year, but that since advanced process tools were only available for 300-mm fabs, it might make more economic sense to fit them with 200-mm tools that can do advanced processes instead of just closing them down. "It makes a lot more sense than 450-mm," Sonderman agreed.

But others on the panel weren't so keen on the idea of continuing work on 200mm, with IBM's vice president of 300-mm semiconductor operations, Daniel Armbrust, saying that although there was always a market for revamped equipment, most firm's wouldn't want to use it for work on the most advanced process nodes. "You might be able to push an 8-inch fab into 65-nm. There may be a couple of players that try that," said Amburst, but added "realistically, that's not where you will see a lot of activity."

Armbrust did agree, however, that shovelling cash towards new 450-mm equipment during one of the decade's harshest downturns was not the way to go, saying it did make more sense to continue pushing the limits of 300mm until that size had been fully exploited. "There will be a time when 450-mm makes sense," Armbrust said, but that time, he says, is not now.

Intel and co disagree, with the threesome making a lot of noise about trying to build 450-mm prototype fabs and move to 450-mm wafers by 2012.

It's hardly surprising so many companies in the industry are against the move, with the majority only just recovering from the trauma of moving up to 300mm. 300mm started coming in during the late 90's, but it's only this year that the volume of 300mm wafers being produced has started to outnumber that of 200mm wafers.

The main gripe from chip equipment makers is that they were saddled with enormous R&D costs in order to help chipmakers 'upgrade' to 300mm, but when the economic crisis hit in 2001-2002, the chipmakers stopped buying equipment, leaving equipment makers in the red.

Another reason many are sceptical about jumping to 450mm so quickly is that historically, moving to bigger wafers hasn't necessarily proved itself an advantage. AMD, for example, managed to hold its own at 200mm, capturing over 20 per cent of the market against Intel's processors built on 300mm. AMD was so successful at 200mm, the firm only bothered to trade up to 300mm in 2005."We're just getting started at 300mm," Glofo's Sonderman told us.

"We learned to do a lot more with less," he told the INQ, adding that in his opinion, companies that are advocating a move to 450mm are running the risk of "destroying the equipment industry."

That doesn't mean 450mm will never be relevant, though. "It will come in at some point," Sonderman said, but noted he didn't think it would be a realistic notion until somewhere towards 2020. "There's so much waste in the industry right now, it's a distraction," he concluded. µ