300 mm: Almost Ready for Auto Pilot
John Baliga, Associate Editor -- Semiconductor International, 7/1/2000
 |
This issue of Semiconductor International contains a special section on the progress of 300 mm technology (pg. 111). Karl Mautz of Semiconductor 300 details the joint venture's findings and evaluation of 300 mm tools and automation. His article is followed by reports on individual technology segments.
In the 300 mm update section, and in the articles on automation, you will find a strong emphasis on the need for automation in 300 mm processing. In fact, a 300 mm fab must be built to suit the automation, instead of the other way around. Given this observation, something I pointed out on this editorial page two years ago bears repeating: "A tools first (automation last) approach to buiding a fab would be similar to the construction of Frankenstein's monster. All the pieces are the best, but do they make the best factory system?"
This bears repeating because it is a very new concept in IC manufacturing. Giving first thought to how the manufacturing execution system (MES), data collection and analysis, automated material handling and dispatching software will interact, and then worrying about getting the tools in the door later, is completely new to this industry. The key thing to remember is that we really aren't technology-limited anymore; we are productivity-limited.
If there is a silver lining to be found in the delay of 300 mm adoption, it is that standards for automation were completed and put in place before people started building 300 mm fabs.
A relative handful of people in this industry saw the need several years ago for the kind of computer integrated manufacturing (CIM) and automation standards that would be required. Thanks to the vision, tireless effort and persistence of these fine people, we now have a bare minimum of automation and CIM standards in place for 300 mm processing. Many will not see these standards as a bare minimum set now, but when 300 mm processing kicks into high gear, it will be obvious to all.
Since the set is the minimum needed, the least equipment suppliers can do is comply with the standards fully. The least IC manufacturers can do is pay for the extra value that complying with the standards provides.
For many years, IC manufacturers have had, in various forms, the attitude that "software should be free." It is because of this attitude that software development is behind where it should be, and why many IC manufacturers have wasted millions of dollars creating and maintaining software they did not want to pay for.
To keep problems requiring human intervention to a manageable number, 300 mm fabs must practically run themselves. The CIM and automation software used in a fab will, in reality, run the fab. The need to recognize its value is being thrust on the industry. An indifferent attitude toward this software and its value can lead to a fab building good product, but not enough to stay in business. That would be a disaster.
