Subject: Cymer's Future
Date: Sun, Dec 7, 1997 17:53 EST
From: Viking1100
Message-id: <19971207225301.RAA22792@ladder02.news.aol.com>
________________________________
I have not contributed to this board for a while since I have been really busy. But here is along post to make up for it. Please read on and let me know what you think.
There has been a lot of talk about Cymer's mid to long term prospects on this board as well as SI. Talks about IBM's synchotron based X-ray or Jmar's own X-ray source (Blue light). Talks about how Lucent's I-beam lithography tool has been successfully demostrated at 0.08 micron and it will run Cymer out of business before the turn of the century. As well as talks about other companies such as TRW and Dupont, who are experimenting with other
methods (Non DUV) of photolithography which are capab;e of going beyond 0.10. I am not an expert in the field of photolithography or even consider myself "knowledgeable" in regard to the technical details of the processes. But I know enough to state with utmost confidence that DUV lithography is the only viable solution for high volume manufacturing at sub 0.25 micron, and will be the only lithography tool used in high volume manufacturing which
operates in sub 0.25 for the next 7 to 10 years.
There is a big misconception on this board as well as SI, that lithography is the only challenge in the manufacturing of IC's. People seem to think that if "XXX Inc." came up with a light source capable of etching at 0.10 micron tomorrow, Cymer would be out of business. This is a false assumption, Cymer will sell just as many laser for the next three years, REGARDLESS of whether there is big breakthrough in the field of lithography on the sub 0.10
micron level.
I do realized that at our current situation, lithography is a "bottle neck" step in manufacturing IC's, but it is only one of many steps. Lets assume "XXX Inc." introduces a light source (EUV, I Beam or X-ray) capable of etching at 0.05 micron tomorrow. Can industry use it to manufacture IC's? Of course not! Equipment suppliers such as Applied Material do not sell any CVD or PVD equipment suited for 0.05 micron process. They do not have this
equipment, nor do they know how to make it. The polishing, spinning and sputtering technology is being pushed to its limit as it is dealing with the 0.25 process. Back end testing equipment is also far from being developed.
Many people in the industry believe that the mechanical alignment of various layers and vibrational problems will be more prohibitive for advancement beyond 0.15 than the wave length of the light source. Intel is investing in its division in Israel not because they want to be closer to Saddam, but because they believe that the seismological stability of that region is necessary for the alignment and vibrational issue which they are bound to face
beyond 0.15. There many other challenges that are not even talked about, for example as chips get smaller and more compacted, they generate a lot more heat. Thermal management issues are becoming enormously big at the current 0.25. Here is quote from Gordon Moore (the originator of Moore's law):
"There are many other challenges which are a consequence, but not related to manufacturing process of faster and more powerfull devices. If you don't do anything else, the power goes up forty fold every two generation. If you start with a 10-watt device (like a Pentium) and go up forty fold, the darn think smokes"
My point is this, chip manufacturing does not follow lithography techniques. There are enormous unsolved challenges facing the semiconductor manufacturing industry beside lithography, which will take time to resolve. The moores law will not continue to the next century, due to these challenges. Also the industrial need for such powerful CPU's, ASIC, RISC and DRAMS evolves and develops over time, so even if there was a capability to manufacture 8Gb
DRAM and 4000 Mhz CPU, we would not know what to do with them. Let me remind you that DUV was conceptually proven many years ago, but there was no need for it then. Dr. Akin has been quoted many times, saying that Cymer had to wait for its time to come and the need for DUV lithography to arrive.
Well, that time has indeed arrived. KrF 248nm is currently being used for 0.25 and will be used for 0.18 shortly. The next generation ArF based DUV at 193nm will be used down to at least 0.13 and most likely down to 0.10. By the time there is a need for smaller sizes than that, we should be fairly close to 2010. Here is another quote from Gordon More
"With tweaking of lenses, robots and ultraviolet lasers, we will be able to extract two more generation of processors from our current optical techniques. After that life gets very interesting, we have three equally unattractive alternatives"
The question for the near to mid term (3 to 5 years) is whether there exists a lithography techniques which rivals DUV in 0.25 to 0.13 range? Very unlikely! The only demonstrated, proven method is I-beam. I-beam, though it seams to be very feasible and reliable, it is by far too slow and too expensive to compete with DUV on manufacturing plant. I am not sure about its actual throughput capability is,but I have read from 0.5 to up to 10 wafer per
hour. Even at 10 wafer per hour, it is still a far cry from 100 wafers/hour that Cannon is claiming for their DUV based stepper. X-ray and EUV are not even considered an option for the 0.25 to 0.13 range. DUV is indeed the only method feasible in high volume manufacturing the current generation of devices.
Another question is whether exist other DUV suppliers such as Komatsu or Lambda Physic, which are capable of gaining market chare from Cymer. I personally do not have the slightest clue as to where they are in their stages of development or the status of their manufacturing capacity. I also know very little about their claimed specs. It is conceivable that Cymer might lose some market share to competition and not grow as much as 50% per year. But
it seams very unlikely that any DUV laser manufacturer can take a big bite out of Cymer's market share. The fact of the matter is that KrF/248 nm market belongs to Cymer. Due to this reason, along with their intellectual capacity and their superior world wide support service, Cymer should remain the dominant supplier of DUV lasers. Here is a quote from a recent (December) article in Semiconductor Business News titled as "Optical lithography
continues to surprise the experts":
""Moving from one UV wavelength to another with the microlithography systems will also mean that incumbent vendors will probably have a greater chance to hold on to their customers. "Once a manufacturer has installed your (248-nm) deep UV tools, that customer is likely to remain with you in moving to argon fluoride systems," said Mitsutoshi Ogiso, general manager of the Canon Semiconductor Equipment Business Division in Tokyo.""
Once this stages of photolithography runs its course, it will be time for the next big step in lithography. It is Cymer's opinion that the next step should be EUV (soft X-ray). IBM believes in X-ray powered by synchrotron, other like other sources of power for their X-rays, and some like I-beam. I do not have the technical knowledge to have an opinion on this matter and very few people do. But regardless of what the next litho. generation will
be, DUV should be the only game in town with Cymer owning a dominant market share until 2008. After that Cymer seams to have as much chance as the next guy, meaning the risk to reward ratio is still very favorable for the extra long term.
It is my very humble opinion that Cymer is extremely undervalued at its current level, and it currently provides and extremely attractive buying opportunity for short, medium and long term. Of course I do not know anything about the extra short term (weeks), because if I did my average entry point would have been 17 instead of 28 :( It probably will not matter three years from now when CYMI is trading at 90$ after another split :)
Just for the record, I am looking for earnings of 0.28 for next quarter :))
Good luck to all the longs,
Viking
P.S. Can some one who has access to SI, please post this message over there, they might find an opinion of "ignorant" AOL/MF member worthwhile. Viking1100@aol.com
|
