|
 |
|
|
|
300mm might not cause over capacity due to lower yield
|
Thursday, April 29, 2004
The principal motivation for going to a larger wafer size is more bang for the buck; in the case of 300mm vs. 200mm wafers, 2.5 times the bang for the buck. At least that is the theoretical rationale. With the industry scrambling to add capacity and a number of new fabs coming online, this begs the question: will the move to 300mm fabs bring on a supply-driven industry downturn?
It's not that simple according to some here at the Semiconductor Equipment and Materials International's Strategic Business Conference.
There are some early adopters of 300mm that have proven that it is indeed cost effective. But for much of the industry, will be taking its first stab at 300mm production with the 90nm node. In fact, during a panel discussion Tuesday morning, panelists suggested that the problems inherent in bringing 90nm and eventually 65nm technology to production will negate some, if not much of the capacity to be gained by the latest wafer size transition.
DRAM, followed by flash, were the first markets to really to see the impact of the 300mm transition, followed by MPU makers, namely Intel Corp., noted Steve Newberry, president and COO of process tool vendor Lam Research Corp. "But a lot of productivity has been mitigated by the fact there are more wafer passes," he said. Process windows are much tighter, and with the transition from 0.13-micron processes to 90nm processes, yields have taken center stage as an issue in the conversion to larger wafers.
But it's not the size of the wafer in and of itself that's at issue; it's the number of wafer passes through process tools and the complexity of designs and materials integration that is necessary at 90nm. As a result, the transition to 300mm wafers won't necessarily resulted in the productivity gains ¨C and thereby capacity ¨C that the industry once envisioned, Newberry suggested.
Much of the 300mm capacity being put in place right now is for production at the 90nm node, agreed Phil Ware, a senior fellow with lithography tool vendor Canon Inc. In terms of cutting edge lithography, in using 193nm exposure tools, the industry is tasked with resolving features less than half the measure of the wavelength of light used to pattern them on a wafer.
This involves increasing the complexity of the litho cell process: resolution enhancement techniques such as phase shifting and higher numerical apertures that involve advanced lens technology, for example. At the 65nm node this is likely compounded further with the introduction of immersion lithography.
"You're really focusing on a lot of issues that aren't readily solved," Ware said. "I don't think it's going to be an easy transition."
These problems are felt all along the supply chain, not just at chipmakers. With process technology getting ever more complex, process recipes are more and more tailored to individual chipmakers. For a process tool component and critical subsystems supplier like Advanced Energy Industries Inc. it means breaking from the past and replicating component and subsystems design across numerous customers.
This could lead to further consolidation in this market, suggested Doug Schatz, chairman, president and CEO of Advanced Energy. Having to serve an industry full of customers that demand individual applications leaves room only for a handful of suppliers who have global channels of sales and support, he suggested.
By: DocMemory
Copyright © 2023 CST, Inc. All Rights Reserved
|
|
|
|
