It's really not so hard being green...

06/01/2002

It's widely believed that protecting the environment is a costly exercise for industry, requiring expensive add-ons such as scrubbers and retreatment plants. That's why individual companies and sometimes whole industries have resisted cleaning up their discharges, lobbied Congress to temper environmental legislation, and fought legal battles to avoid having to pay fines and clean up waste sites.

Many years ago there were similar arguments against efforts to improve "quality" in industry processes, because doing so would "obviously" prove too costly. US industry leaders once turned a deaf ear to quality program pioneers like W. Edwards Deming and Joseph M. Juran. Deming found a much more receptive audience in Japan, and became an icon there, with the treasured Japanese industrial prizes for top performance in Total Quality Management (TQM) being called the Deming Awards. Eventually the huge success of Japanese manufacturing companies forced competitors to study their secrets. To their amazement, executives all over the world began to learn that "quality" isn't really a cost — it can actually save huge amounts of money. Deming charged that any process that had too much variation was out of control. TQM companies cut down the variation that leads to defective products, tracked down causes (often working with suppliers) through techniques such as fish-bone diagrams and Pareto charts, and gained control of processes through such techniques as statistical process control (SPC).

Costly and time-consuming end-of-line testing can be completely eliminated because there are essentially no more defective products coming off the line! Six sigma programs aimed at pushing far out into the tails of a standard distribution, or Gaussian curve in statistical analysis, bring defects well below the one in a million level. This makes customers very happy, and gives field service reps little to do — like the Maytag man. Not only that, but once true TQM systems are in place with an empowered work force dedicated to making sure that everything runs smoothly all the time, manufacturing costs plummet. So not only are the end-products much better, they cost less, too.

What if a competitor catches on and does the same thing? That's where continuous improvement, kaisen, comes in. No matter how good processes get, they can always be made better, so achieving TQM is a never-ending (but highly rewarding) task. To find out what is "better," you need a goal: total customer satisfaction. Any cost a company incurs that doesn't contribute directly to that goal is considered "waste." Continuous improvement is a process of finding and eliminating this waste. Inventory is waste: essentially just money sitting around doing nothing. That's why just-in-time (JIT) manufacturing has evolved, to bring inventories to their lowest feasible levels with materials and parts being delivered only as needed.

Pollution is an obvious and often costly form of waste, especially with potential environmental fines. At one time, semiconductor fabs needed as much water as a golf course; now, they need much less. Can a fab get by with less water? Contaminated gases and chemicals are pumped out of a growing number of chipmaking processes, and they must be treated or subjected to costly disposal. Can processes be designed to minimize the chemicals and gases required? Contamination-free processing entails frequent chamber cleans. Is it possible to clean chambers less frequently, and to do it in less time with minimal effluent? Is it worth a lot of R&D effort to find processes that use less water, chemicals, and gases, and minimize chamber cleans and effluent discharges?

The chipmaking industry is betting that it is. Wringing waste out of processes not only eliminates a lot of headaches in running a fab, but can also pay off handsomely on the bottom line. Sematech has groups working on cleaner processes with less effluent, or more benign by-products. The results are shared across the industry at technical conferences. Semi has its EHS division, aimed at helping member companies clean up their processing. Once cleaner processes with lower water requirements are found (e.g., wafer cleaning with supercritical CO₂), the benefits will continue. Already some fabs are boasting about using much less water and chemicals in processing. Groups from Motorola and other chip companies are reducing the amount of chamber cleaning needed, while also cutting emissions of fluorides and other compounds that act as greenhouse gases and deplete the ozone layer.

Eventually, the whole industry will learn that it isn't really so hard being green, and the payoff will be, well, more green in the bank!

Robert Haavind
Editor in Chief