There's no such thing as a free lunch

12/01/2001

by Ken Goldstein, Ph.D.

The savings—fewer filters, fans, chillers, coils, pumps, motors, less steel and concrete—were actually outweighed by the additional costs of the minienvironments themselves, including I/O ports, pods, cassettes and carriers, robots and automation. But the early pioneers kept experimenting and actually built some large facilities using this new technology. When they looked at the big picture, they arrived at a very different conclusion. Because the minienvironments were highly effective at isolating the product and process from nearby contaminants, their use led to higher yields. In addition, the isolation capability of the enclosures protected the process from nearby construction activities making it easier and quicker to install new tools, permitting significantly faster production ramps.

All things considered, minienvironment technology looked like a winner.

With this as prologue, a few obvious questions might be asked. Why aren't more facilities using minienvironment technology? Why isn't everyone using them? And if they are so fantastic, why is anyone even considering conventional cleanroom designs?

Zeroth Law of Economics
One reason why everyone is not using minienvironments is that they are simply inappropriate for some applications. Minienvironments that are capable of achieving ISO Class 1 probably represent severe overkill for a product or process requiring a cleanliness level of somewhere between ISO Class 7 and 9.

A second reason involves materials handling. Recall that minienvironments typically work best as small areas of increased cleanliness within a larger room of decreased cleanliness. But very few products can be brought into a room and then produced at a single workstation. Instead, the product invariably moves through a series of steps, each located at a different location from the previous one.

This immediately raises the question: How do we protect the product from ambient room contamination while moving it through the less clean space between process steps? Somehow, we must effectively contain it within a small protective enclosure.

In semiconductor plants, these are called "pods" and we actually transfer the pods containing the product around the factory floor. But the pods are effectively sealed and we cannot perform any processing while the product is contained within them. This requires two additional materials-handling transfer steps for each production step—the transfer of the product out of the pod in preparation for the actual processing and another transfer into the pod in preparation for the next move, operation. These handling steps and transfers are costly, inefficient and do not contribute to the transformation of the product from raw material to finished goods that can be shipped and sold.

But perhaps the most common reason goes back to the first lesson that should have been covered in Economics 101. The instructor would tell us that "goods are scarce" and then refer to this as the "zeroth law of economics," which is even more basic than supply, demand or utility.

The rest of us know this truism as "there ain't no such thing as a free lunch." In the case of minienvironments, this technology is capable of doing a lot of good for us. But these benefits are not without costs. We need to go into a project with eyes wide open and fully informed of the likely costs and benefits.

Production-control schemes
Production-control schemes are perhaps the biggest cost uncertaintly associated with minienvironments.

Consider the typical (manual) production-control scheme. Operators look at identification tags on production lots and make sure that they coincide with the information on the lot traveler. The lot traveler contains a complete listing of all the production steps that will be required, what specific tools are to be used and the specific recipes to be loaded for each lot at each production step. The lot traveler would also specify the sequence of the individual products (wafers) as they are loaded in cassettes and as they are loaded into the process chambers. Everything is specified in detail and there is no room for operator creativity.

Minienvironment technology uses protective enclosures; the minienvironments themselves and the pods that protect the contamination-sensitive product from the surrounding room ambient conditions.

Note that the use of these enclosures effectively obscures our view of the individual wafers or even of the lot. In other words, we can no longer read the identification number on the wafers. We certainly cannot "see" the sequence of wafers in the cassette. And although we may be able to read the label on the outside of the pod, we can't know for sure that the number on the outside of the box corresponds with the correct number of the lot inside the box.

This invariably leads us to robotics, automation and computer control of our factory floor. Robots of one sort or another will do most of our materials handling, from individual wafer transfers to intra-bay and inter-bay movement of sealed pods. Automated sensors will read the lot and wafer numbers and load the correct machine assignments and recipes. And the factory-control computers will monitor everything, including how the individual tools are functioning. When the systems work properly, the potential advantages will be enormous.

But the role of people in this scheme will be redefined. People will be responsible for programming, hardware set-ups, maintenance and occasional checking. And of course, people will be responsible for intervening when things are clearly going astray.

Production-control schemes can be a major stumbling block for companies contemplating the use of minienvironments. It represents a paradigm shift in production and the people responsible are understandably hesitant to jump right in.

At this point, I would suggest that there are clear trends at work that will inevitably lead to a known future. How we get there is an open question. It will be interesting to watch. And for those who participate, it will also be exciting, frustrating, difficult and nerve wracking. Enjoy the ride.

Dr. Ken Goldstein is a principal with Cleanroom Consultants Inc. (Scottsdale, AZ) and is a recognized expert in planning and designing of cleanrooms and ultrahigh purity systems.

In recent columns, we looked at the history of minienvironments starting with the overblown claims of up to 30 percent facility capital savings. After considering all of the new project-related costs and savings, we were forced to reject the early claims of substantial cost savings.