Apart from the onslaught of marketing blitzes surrounding big industry trade shows, it's rare to have two semiconductor equipment companies launch products on the same day.
It's rarer still when the tools hitting the market are competitive technologies. Monday's news that both KLA-Tencor and Applied Materials have come out with new mask inspection gear/schemes, although timed to coordinate with this week's SPIE Photomask Japan (not exactly a major event), was more than just an unusual juxtaposition of similar commercial introductions.
Applied's Aera2 platform marks the company's official introduction into the mask inspection/review space, a domain long dominated/monopolized by KLA-Tencor. The two firms have battled for market share on the wafer inspection/detection/review side of the process control/metrology space for years, but KLA-T had the photomask/reticle business (which Gartner Dataquest says was $464 million last year) more or less to itself with its STARlight and TeraScan lines. That changed a bit when NuFlare entered the arena with an inspection system awhile back, and with AMAT's foray, the playing field has become much more competitive--a development welcomed by most in the chip- and maskmaking communities.
The move by Applied was not exactly a state secret, with reports filtering out over the past few months of mask inspection tools going through their beta site paces at several customer locations. Nor has KLA-T been resting on their laurels, rolling out a whole new brood of systems called TeraFab last month, which by the collective branding moniker more than hints at the main target business--memory and logic fabs clamoring for better incoming mask qualification tools and faster, more sensitive production mask inspection suited to their own particular processing needs at 65- and 45-nm device technologies and beyond.
But KLA-T wasn't done there, introducing another mask inspection weapon in its arsenal called Wafer Plane Inspection, or WPI, a three-letter acronym that made this blogger think the "what Price Index" (or maybe the "wafer processing index"). What the technology actually does, according to the prepared statement by Harold Lehon, VP/GM of the company's reticle and photomask inspection group, is "find all defects of interest and also accurately distinguish which mask defects are likely to transfer to the printed circuit on the wafer," meaning that "customers will have a cost-effective, direct link between mask inspection and fab yield." The company claims this capability is available "for the first time in the industry" and can meet "32-nm generation defect sensitivity requirements."
If this sounds familiar in light of this week's news, it's because Applied touts the aerial-imaging-enabled Aera2 in its press release as "the first and only mask inspection system that enables users to immediately see how the pattern on the mask will appear on the wafer. The system detects defects according to their impact on the wafer, filtering out the large number of nonprinting defects that plague conventional mask inspection systems." To underscore the momentousness of its announcement, Applied peppers its PR with descriptors such as "groundbreaking," and "unmatched," as well as hyperbolic verbs like "redefines" and "revolutionizes."
Obviously, AMAT and KLA-T didn't coordinate their messaging or timing, since the two technologies/products' similar capabilities kind of cancel out the other's claims of being unique in its class. Although there are certainly differences between the two approaches, I'll leave it to the other reporters, analysts, and bloggers to sort out the technical variations, pick apart the algorithmic mumbo-jumbo, and gauge the pros and cons of each system.
One thing is clear though: With the availability of the Aera2 and TeraFab souped up with WPI, the winners here are actually the customers at the mask houses and wafer fabs. Sure, these tools won't come cheap, with their multimillion-dollar price tags, but they are big improvements over the previous generations of mask inspection systems and incorporate something that has long been at the top of the wish list for those struggling to sort out the noise of nuisance defects on masks from the clear signal of the yield-killer kind. Finally, engineers will be able to determine, promptly and accurately, which mask-related defects will actually show up on the wafer.
|
Comment by GUEST on 2008-04-30 11:08:10
Aera system stimulate as a stepper, it locates the detect sensor on the actual exposure wafer position and provide the light source as a stepper. I am not sure that these two new tools are going to be the same design idea, even it sound like it is true, but it could be different in the image process algorism & defect process speed & software interface and make the tool working fine or not. It will be good news for Aera2 if KLA-T use same idea on the new tools, morever Aera2 need more powerful functions to meet customer test. | Comment by GUEST on 2008-04-30 11:16:58
From: Gary Hsueh, CFA – Semiconductor Equipment (contact details removed for privacy/security purposes - Fabtech Admin)
Recent checks indicate that Intel (INTC, Perform, $21.20) has chosen Applied Materials' (AMAT, Outperform, $20.94) reticle inspection tool for its entire 32nm node after a long period of evaluation against KLAC, and ordered 14 systems for '08 and '09 delivery at about $20 million per tool. This represents a $300 million shift from KLAC's dominant market share in reticle inspection.
Based on our work, KLAC's ASP for reticle inspection at INTC was ~$28 million, with a gross margin of ~80% (vs. KLAC's consolidated number of 58% in the Dec quarter). Reticle inspection is a great business, but competition out of AMAT is a double blow, with actual pricing at ~$20 million. So, even if KLAC defends its reticle inspection share at other accounts, gross margin would compress from ~80% to 72%.
AMAT is bringing to bear the same competitive dynamics to TSMC. Our checks here this week in Asia suggest that TSMC will likely splinter its reticle inspection buys between AMAT (high-end), and Lasertec (low-end) vs. predominately KLAC. Making a definitive call on KLAC's share loss at TSMC may be premature, but at the very least, competition from AMAT will likely materialize as margin erosion. |
|
Blogs 
