| |
Well, if you are a high volume DRAM producer or just a high volume producer of a limited number of IC designs, you can afford the horrendous price of the PSM reticle sets or the off axis approach. You might get away with it if your device has a reasonable number of levels and your factory has a low number of designs to produce.
When you move into the ASIC arena of those fabs that have low volume and high mix of designs, it is far more cost effective to go the way of DUV.
Not to straddle the fence here but as a hedge against being wrong, I have a strong position in the reticle producers (DPMI and PLAB) such that I can benefit if the PSM tricks are successful across the board.
I'll wager these guys are praying that PSM goes mainstream so they can beef up their bottom lines with these extremely high margin reticles.
So yes, I do not subscribe to the 250nm all i-line process technology, especially since DUV is being used in Japan in production and IBM is doing the same thing in Manassas, Virginia. As much as I have learned never to say never in lithography over the years, "NO WAY i-line is going below 0.25u". However, if you are successful at 250nm with i-line, you are up against the wall for the next generation of device technology. Also, over the years, it has been imprudent to stretch your equipment to the edge and beyond its specifications. We all know a robust process is when you do not stretch the limits of your process or equipment and leave a little margin of comfort.
Here's the other good rub. Let's say I am wrong and 250nm i-line is successful. Most of the new systems bought will be retrofittable to DUV down the road. I know 3 major providers of i-line steppers have stated that their new DUV systems are based on the existing i-line equipment platform such that an almost seamless production transition (operator training) could be accomplished to DUV. With this said, the major beneficiary here would be CYMI for the upgraded DUV supplies. (OOPS, CYMI is already up close to 50% this month,<GG>).
Finally, if you look at the new device technology timeline you see the move to smaller design rules is happening at a much faster rate, almost on an expotential curve. No longer is a design rule lifetime measured in years, it is now a matter of months. Therefore, while you are ramping and running your 250nm process, the designers are putting together the 180nm and 130nm design libraries for prototyping. Before you know it, your i-line systems are obsolete.
Obsolescence is the order of the day in lithography and its tooling.
Remember the time delays between the 80286, 80386, 80486, Pentium, Pentium II, and now Merced with all the speed improvements in between??? They are coming out faster than we can keep track of. This year alone will have 2-3 speed enhancements while a new chip, Merced is scheduled for next year. Technology moves on and if you do not prepare yourself, you are left holding the bag.
No one is going to share all the lessons that need to be learned relative to DUV, you will just have to muddle through it yourself and go down the learning curve. You need, at a minimum, to mix and match DUV with i-line in order to prepare yourself for the inevitable. And as stated before and paraphrased here, it is a great deal easier to drive a Ferrari at 120 MPH than it is to drive a VW at that speed<GG>.
We already have seen CYMI announce that they will introduce 2 new DUV sources, system, and/or supplies this calendar year which will extend the range of the lithographic exposure systems offered by the likes of Nikon, Canon, ASMLF, and SVGL.
Just now I received an analyst update that projects a 12 month price target for ASMLF at 105. IT is presently at 75, up from its 65 level of a few days ago.
This thread is General Lithography, so a General view and approach is being taken by myself. I strongly believe in the success of those companies providing enabling technology such as Lithography, CMP, Defect Detection, and Wafer Isolation Technology. This group is performing well recently with DPMI and CYMI leading the way.
For those of us in the Litho arena, the safe harbors are DPMI, PLAB, CYMI, ASMLF and Maybe Nikon, Canon, and MASK.
Andrew |
|
