INDUSTRY NEWS

Vendors waste no time responding to need for post-CMP treatment

In these days of stratospheric production costs and environmental consciousness, few semiconductor manufacturers need or want the added burden of pollution fines. So when a chipmaker in Silicon Valley began to take hits in the pocketbook for improperly discharging CMP effluent into a municipal wastewater stream, the managers grabbed for the phone.

Lucid Treatment Systems got the call. Based in Hollister, CA, the company specializes in treating CMP wastewater. The chipmaker was releasing chemically laden slurries from its SpeedFam Auriga polishers with its waste stream, says Jack McCann, Lucid's vice president of sales and marketing. "The company was paying these huge fines to Santa Clara County," he says. Ultimately, the chipmaker became the first customer to use Lucid's system, which recycles water at the point of use to reduce the volume of waste and then concentrates the remaining waste before collecting it for disposal by an outside contractor.

Founded in 1996 as CMP Technologies, Lucid changed its name in March after receiving an infusion of cash from private investors and from Semifab, now a partial owner. Semifab, a Hollister-based vendor of minienvironments and other cleanroom-related process tools, provides manufacturing expertise. Two-year-old Lucid is using the chipmaker as a beta site to test its system, which is designed to condense three drums of waste into one. "We wanted to go out and capture some real data from real customers," asserts McCann.

Steven Browne knows a lot about county regulations, post-CMP waste, and data. At Cirent Semiconductor in Orlando, FL, where Browne is engineer of ultrapure water and industrial waste treatment, the chipmaker has "a discharge permit that has very specific figures on total suspended solids" from Orange County, the local governmental jurisdiction. Allowable levels of total suspended solids are 8 ppm, Browne says. "In order for me to meet that requirement I have to treat my CMP slurry separately in order to remove the suspended solids from the waste stream."

The solution? Pall, the giant filtration and separations company, announced in April it had installed its Microza ultrafiltration slurry treatment system at the Cirent fab. The system concentrates suspended solids up to 400—500 fold at flow rates of 50 gal/min, according to the supplier. A mixture of slurries from oxide and metal polishers, the flocculated concentrate is converted to a dry cake with a filter press. Browne says the discharged solid "is classified as a nonhazardous waste" and taken to a county landfill.

As CMP use becomes more prevalent throughout the industry, two factors are driving the growth of the waste treatment market segment, points out Vivien Krygier, Pall's senior vice president of marketing. "Suspended solids are too high for local discharge regulations, and you have to remove the suspended solids. There's that, and in a number of places they're looking to recycle the water because of the tremendous amount of water being used." She adds that "the drivers for water recycling right now are Europe and the Far East" because water costs are higher and environmental regulations, particularly in Europe, are more stringent than they are in the United States. As McCann notes: "Fabs are seen as environmentally unfriendly."

Krygier and Gary Corlett, Lucid's vice president of technology, predict rapid growth in the market for treatment systems as CMP becomes an essential technology for chips with up to 13 metal layers. "I can see this growing to a $50 million to $75 million annual market as all the different fabs come on-line with CMP. All will have to eventually treat that water," Krygier says.

Corlett, Lucid's founder and the coauthor of three pending patents related to the company's treatment technology, quotes expert projections for slurry use to increase at an annual rate of 25—30% through 2002. The CMP process accounted for approximately 5% of total process water in 1997, according to Corlett. By 2000, water used in CMP processes will account for more than 30% of all the water used in the fab, he says.

With the use of copper deposition already beginning in a few pilot lines, chipmakers have even more reason to treat the water used in CMP, which is an enabling technology for the so-called dual-damascene processes. "As we drop down in geometries, CMP will come into play more," notes Browne of Cirent. "And what we will see in the not-too-far-off future is polishing copper off the surfaces of the wafer. That presents us with a whole new bag of legal ramifications." Cirent is building a 30,000-sq-ft cleanroom "for Bell Labs' people to do copper deposition and copper cleaning on the wafer surface. Part of that [processing] involves CMP."

"Some of the biggest treatment [applications] will come in where copper CMP is used," agrees Krygier. "Now you're starting to produce wastewater that is deemed hazardous. . . so if suspended solids and the need to recycle water are not drivers for this industry segment, certainly copper in the wastewater is."

Pall and Lucid are two of at least 10 vendors targeting the market for treatment solutions, according to Corlett. Companies such as EPOC, Millipore, Sumitomo Chemical, and Ebara all offer technologies such as vacuum accumulation, ultrafiltration, and flocculation. He notes that filtration provides good water recovery, but it requires the addition of neutralizing chemicals and a filter press or centrifuge to remove enough water from the sludge to make a system cost-effective. Corlett cites figures showing the semiconductor industry could save as much as $798 million over the next 10 years recycling CMP wastewater, based on an average cost of 25 cents per gallon for "pre-RO" water.

Reclaimed water can be recycled through the RO/DI system, eventually, according to the Lucid executive. At the very least, most suppliers of these systems note that the reclaimed water can be easily used as chilled water, irrigation water, or cooling water.

McCann asserts that 75% of chipmakers believe they have a water problem and desire removal of greater than 90% of the solids measuring 0.05 µm or more from their industrial waste stream. Twenty percent think they have a water problem and want 99.9% of solids of the same micron size for reuse in cooling towers. The remaining 5% have a problem with their source water, according to McCann, and want to see all particles removed so that the water can be reused in the RO/DI system.

Based on a cost model using 12 IPEC 472 polishers, Lucid touts its point-of-use technology as a huge cost-saver over a typical centralized treatment system. The system can save a chipmaker between $150,000 and $450,000 annually, according to McCann.

"Instead of playing down in the basement [with a centralized system], why not start up in the fab where the polishers are?" he asks rhetorically, adding that the company's approach eliminates the need to build another facility. "A centralized system is probably fine for a company like Intel with 30 polishers. But most companies with six polishers don't have the money to put up a building for centrally treating the waste." Lucid will begin full commercial production by September.

One firm licking its chops in anticipation of this market growth is Microbar. A Santa Clara—based equipment supplier known for its chemical delivery and waste treatment systems, Microbar "is about to close on the acquisition of a company that has a unique technology in CMP waste treatment," says an enthusiastic Bruce Juhola, president and CEO. "It's pretty exciting, because it's a patented polymer treatment technology that removes colloidal silica and dissolved metal ions and causes them to become dissolved, insoluble particles. All of that is filtered out, and the water coming off the end of the filter is clean enough to be recycled."

Told that's quite a technology, he replies: "That's why we're banging the door down. We don't want a Pall or a U.S. Filter to start chasing it."

Microbar heard about the U.S. firm in a serendipitous fashion, Juhola relates. The company "discovered the technology with some work they were doing in the mining industry four or five years ago. . . . We got wind of it through a venture capital investor located in the same city." The firm has five basic patents, Juhola says, adding that Microbar will apply for several more once the acquisition, now in the hands of the attorneys, becomes final.

The system is installed at a fab Juhola would not name. "It's been operating for 18 months and doing a great job. We've been helping this company during the acquisition process as a member of SEMI/Sematech to get the system into testing at Sandia under a CRADA," he says. A pilot system was scheduled to be delivered to the national lab in Albuquerque at the end of June.

The technology has the potential to remove copper, Juhola says. "I mean remove it all. That's another thing that has us so excited here. Right now many of the fabs are talking of combining streams off the CMP process with copper and other kinds of streams, where they can get a concentration level low enough to dump it down the drain."

Microbar's overall strategy is "to develop chemical management systems for lithography, DUV, CMP, and copper processes," Juhola continues. "That's why when we heard of this technology we went after it so aggressively. There are huge cost-of-ownership [concerns] with both CMP and copper. Not only do we want to supply the delivery system for CMP and copper, now we can take care of the waste."

As CMP processes gain in popularity, suppliers such as Lucid, Pall, and Microbar are finding they can take care of business as well. For vendors in this market segment, CMP could also stand for can't-miss proposition.

Illustration by James Schlesinger

MicroHome | Search | Current Issue | MicroArchives
Buyers Guide | Media Kit