SI CHINA     SI JAPAN
Login  |  Register          Free Newsletter Subscription
Subscribe
Views on News   


Link This | Email this | Blog This | Comments (4)


The Other 450 mm Shoe
May 6, 2008

The three companies openly pushing for 450 mm wafers are working on a plan to subsidize the equipment industry’s 450 mm development effort. The carrot will be large, perhaps approaching a billion dollars when all is said and done. Its increasingly clear that a combined stimulus package is needed to prompt the equipment vendors to start 450 mm development, a well-placed source in Austin said.

 

Intel Corp., Samsung Electronics Co., and Taiwan Semiconductor Manufacturing Co. Ltd. went public Monday with their support for a transition to 450 mm wafers beginning in 2012. Toshiba Corp. is keeping out of the spotlight, but also is backing the transition. The companies said the 450 mm effort will develop standards within the International Sematech Manufacturing Initiative (ISMI), “as it plays a critical role in coordinating industry efforts on 450 mm wafer supply, standards setting and developing equipment test bed capabilities.”

 

ISMI, which has grown to include 16 corporate members, also is a focal point for environmental impact studies, an increasingly important area for an industry awash in chemicals. Look for ISMI to launch a significantly expanded Environment, Safety, and Health (ES&H) initiative soon. With the recent decision by ES&H guru Walter Worth to retire from Sematech/ISMI, ISMI is searching for a person who can head up the expanded ES&H initiative.

 

As the Sematech lithography, interconnect, and metrology programs have largely shifted to Albany, lured there by the Empire State's 2:1 matching grants for equipment and salaries, about 50 jobs have moved from Austin to Albany in the past six months. New job creation for Sematech’s core technology R&D also has largely been in Albany. At the same time, ISMI has grown its staff in Austin. The result, employment will be roughly balanced between Austin and Albany, with about 200 people working in each location.

 

450 Green

 

In their Monday announcement, Intel, Samsung and TSMC tied the environmental and cost advantages of the 450 mm wafers together. The press release said, “through more efficient use of energy, water and other resources, bigger wafers can help diminish overall use of resources per chip. For example, the conversion from 200 mm wafers to 300 mm wafers helped reduce aggregate emissions per chip of air pollution, global warming gasses and water, and further reduction is expected with a transition to 450 mm wafers.”

 

Fewer fabs, each with state-of-the-art environmental controls, would be seen as the Green solution in a world increasingly concerned with water and energy consumption. 



Mike Goldstein, Intel Corp., holds a 450 mm wafer

So there you have it.

 

The carrot: subsidies to equipment and materials companies which quickly get on the 450 mm train.

 

The stick: risk being left out of the four biggest equipment buyers, who account for ~40% of capex spending, and growing. When leading-edge production is considered, the four proponents of 450 mm wafers account for more than half of new equipment purchases.

 

And the sweetener? The 450 mm wafer fabs are being positioned as the environmentally safer solution.

 

The opposition to 450 mm wafers has been intense and widespread. The opponents don’t buy the basic argument that moving to the larger wafer size will be more efficient this time around. Instead, they argue, take the billions needed to develop 450 mm equipment and invest that money instead in more automated 300 mm fabs, which are accessible to companies in the middle tier.

 

Indeed, AMD, Chartered, IBM, Hynix, Powerchip, STMicroelectronics, and UMC are large chip manufacturers, but they are among the companies not quite large enough to afford 450 mm fabs. What the anti-450 crowd argues is that bigger fabs certainly help Intel and the other very large companies as they compete with smaller rivals. That economic survival battle alone nearly guarantees strong opposition to 450 mm wafers. When combined with the economic pressures facing most equipment suppliers now, the 300/450 teeter-totter hangs in the balance.

 

The debate will continue to intensify this year as the industry gathers at Semicon West beginning July 14, and will be a major focus at the ISMI Symposium on Manufacturing Effectiveness/ISMI Manufacturing Week, October 20‑23. Mark your calendars.


Posted by David Lammers on May 6, 2008 | Comments (4)


May 6, 2008
In response to: The Other 450 mm Shoe
guest commented:

This is just a repeat of history of the 300 mm transition which was also backed by the big players first. Many companies including the vendors have not felt the payoff from the 300 mm transition, so they don't believe the benefit of the 450 mm. No vendor wants only a few customers, all with very different requirements. Plus there is more opposition to 450 mm this time, now that the cost difference between 200 mm and 300 mm is better understood.




May 6, 2008
In response to: The Other 450 mm Shoe
doubtful commented:

I think two points need to be made here. First, a 450 mm wafer obviously consumes more material and energy than a 300 mm wafer. For the same amount of consumption, given a choice between running a few large wafers and idling the fab the rest of the time vs running many smaller wafers with little idling, which is more reasonable (and flexible)? Second, this transition will not intercept 32 nm or even 22 nm. Beyond that, why would there be necessarily a need for silicon wafers?




May 7, 2008
In response to: The Other 450 mm Shoe
doubt? commented:

To "doubtful" points: it's about cost, material and energy per unit out, 450 mm fabs will not run a few wafers and idle - larger wafers are about being able to make high volume more cheaply whether for low mix or high mix - and not only will the intercept happen in the timeframe indicated, Silicon will be run for a long time after.




May 7, 2008
In response to: The Other 450 mm Shoe
no doubt commented:

It is sometimes estimated a 450 mm wafer has the area of two 300 mm wafers. Equivalently you produce about the same number of chips using two 300 mm wafers or one 450 mm wafer. Cost of consumables scales with wafer area not wafer units, so the cost would be the same for the two 300 mm wafers or the one 450 mm wafer. So it is mathematically clear cost per chip doesn't change (much) with wafer size!





POST A COMMENT
Display Name or Registered Users Login Here.
Please restrict submissions to less than 7,000 characters (including any HTML formatting).

Before submitting this form, please type the characters displayed above. Note the letters are case sensitive:


Advertisement

Advertisements





©2009 Reed Business Information, a division of Reed Elsevier Inc. All rights reserved.
Use of this Web site is subject to its Terms of Use | Privacy Policy
Please visit these other Reed Business sites