by Robert P. Donovan

Ultrapure water (UPW) is a key process fluid in cleanrooms used to manufacture semiconductor (and other precision) products. UPW is potable, municipal water purified at the semiconductor site to reduce the concentration of contaminants in that water, such as dissolved solids and ions, especially the hardness components, calcium and magnesium.

Electrical resistivity is one measure of ionic contamination in water. The resistivity of municipal water is typically 1 to 2 kohm-cm; that of UPW, 18+ megohm-cm, a full 4 orders of magnitude higher. UPW typically costs 10 times as much as municipal water, say $10/1,000 gallons for UPW vs. $1/1,000 gallons for municipal water. Water costs, however, are not usually a major contributor to the costs of producing finished semiconductor chips. Even making UPW cost free—reducing the cost of UPW to zero—would not appreciably reduce the unit cost of manufacturing a chip for most US semiconductor manufacturers despite the dramatic annual dollar savings sometimes achieved by reducing water usage.^1,2 Spreading these costs over the huge number of chips manufactured results in only a miniscule UPW cost per chip. (The major chip costs are now in processing equipment, which rapidly becomes obsolete and must frequently be replaced.)

Water availability, however, at some locations may be an important factor in optimizing UPW usage. Municipalities can simply refuse to raise the water allotment of even an established manufacturer so that further expansion of production at that site can be achieved only by squeezing more performance from the existing water allotment.³

There are a number of options for doing this. First, attempt to generate more UPW from a given volume of municipal water—at some semiconductor manufacturing sites, 2 gallons of municipal water yield just 1 gallon of UPW.

Second, rinse wafers with apparatus that uses less water to produce adequately rinsed wafers—water flow in early designs of wafer rinse tanks often allowed most water to pass around the wafers rather than flow between them.⁴

Third, reuse the rinse waters discharged from the fab wet benches—many manufacturers discharge to the municipal sewer water that is of higher quality than that which the municipality delivers to them.

Pursuing the reuse option is pretty much up to the user rather than the suppliers of water treating of wafer processing equipment. I will discuss this option in a future issue.

Click here to enlarge image

Robert P. Donovan is a process engineer assigned to the Sandia National Laboratories as a contract employee by L & M Technologies Inc., Albuquerque, NM. His Sandia project work is developing technology for recycling spent rinse waters from semiconductor wet benches.

References

1. Morrow, I.B. and D.F. Turner, “Wastewater Recycling to Process in the Semiconductor Industry,” Ultrapure Water, Vol. 16, No. 3, March, 1999, pp 33-38.
2. McBride, D., “Water Conservation at Intel’s Rio Rancho Site,” Ultrapure Water, Vol. 14, No. 7, September, 1997, pp 45, 47-49.
3. Weems, J.A., “Strategies for Water Reclamation at an Advanced Wafer Fab,” pp 167-180 in Proceedings of the 16th Annual Semiconductor Pure Water and Chemicals Conference, Vol. 2, March 3-7, 1997 (Balazs Analytical Laboratory, 272 Humboldt Court, Sunnyvale, CA 94089-1315)
4. Kempka, S. N. et al, “Evaluating the Efficiency of Overflow Wet Rinsing,” Micro, Vol. 13, No. 5, May 1995, pp 41, 42, 44-46.