INDUSTRY NEWS

Physicists hope to launch space technique as cleaning breakthrough

DUARTE, CA—Can a technique developed originally for space propulsion be relaunched as an economical and environmentally safe advancement for cleaning wafers in the submicron era? In their lab located in this small town near Pasadena—home to NASA's Jet Propulsion Laboratory—Julius Perel and John Mahoney are exploring that possibility.

The two atomic physicists are the founding partners of Phrasor Scientific, an R&D firm specializing in the development of industrial applications for a technology known as electrohydrodynamics. Adapting the technology for semiconductor manufacturing, the partners have developed a "macrocluster impact" technique that shoots beams of a charged water-based solvent at a wafer. Traveling at supersonic speed, the atomized clusters generate shock waves that break the bonds holding contaminants to the substrate. Because the macroclusters in the beam measure 0.01 to 0.1 µm, the physicists assert that the method can even make quick work of that chipmaking nemesis, the killer particle measuring less than 0.1 µm.

Waiting for liftoff: Julius Perel, top left, and John Mahoney examine their wafer-cleaning invention. In bottom photo Mahoney displays the working model with its capillary, neutralizer, and nozzles.

According to Perel and Mahoney, an electrohydrodynamics-based cleaning tool has several advantages over cleaning methods that use either a laser or cryogenic jet sprays. The method can be used in a vacuum; it contains no moving parts; it requires only microliters of solvent; and it minimizes the need for expensive diagnostic tools. The size of the macrocluster may also enable the beams to remove contaminants trapped in submicron trenches. Because it's an in situ application, the technology eliminates the need "to lift the wafer out, clean it, and put it back in the process tool," Perel points out. For these reasons he and Mahoney believe that a surface-cleaning tool incorporating the technology has the potential to greatly reduce IC manufacturing costs.

The U.S. Air Force and NASA developed electrohydrodynamics in the 1960s with TRW and Hughes as a method of manipulating and rotating satellites in space, Perel says. After the government lost interest in the technology, Perel and Mahoney began to explore it, eventually adapting electrohydrodynamics for a variety of applications after they formed their partnership in 1974. Industrial uses for the technology include its use in amorphous alloys, fine powders, silicon coatings, and mass spectrometry for biomedical applications. Phrasor has taken part in more than 20 government programs with agencies such as NASA and the National Institutes of Health. The firm has also worked on joint projects with companies such as Hewlett-Packard. Now the two partners have set their sights on the semiconductor industry.

Perel patiently describes the physics of the electrohydrodynamic technology while sitting at a long wooden table surrounded on three sides by eight-foot-high shelves crammed with reference books and scientific tomes. A giant periodic table is prominently enshrined on the wall above one of the bookshelves. "We use the periodic table the way that stockbrokers use the ticker tape," the physicist jokes over the din of the machine shop in the rear of Phrasor's 4000-sq-ft laboratory.

Asked what prodded their interest in wafer cleaning, Perel replies that he and Mahoney had been using the technology to bombard surfaces and noticed that when they completed the process, "the contaminants were missing." They spent about two years exploring ways to adapt the technology to surface-cleaning requirements. "We were reading about what's going on in the wafer process industry," Mahoney recalls. "It looked like a natural."

In Phrasor's machine shop Perel and Mahoney developed a device that transports a solution—pure glycerol and water/methanol, for example—to the tip of a silica glass capillary centered inside a counterelectrode. A cluster, Perel notes, is "a group of molecules held together by van der Waals forces;" the clusters that Phrasor makes are "smaller than the wavelength of light." The cluster beam source operates in a vacuum of 10^-5 torr. Subjected to a positive voltage greater than 10 kV, the metal sheath enclosing the silica capillary creates an intense electric field at the tip. The charged liquid forms a beam of energetic macroclusters when the electrostatic stress exceeds the solution's surface tension, Perel explains. At the same moment, positively charged clusters speed away from the tip.

"The voltage applied to the needle is the same voltage that accelerates the clusters and shoots them at high energy," Perel continues. "When they're fired at a surface, they knock off everything on the surface at velocities greater than the speed of sound." A neutralizer injects low-energy electrons into the cluster beam to keep charges from accumulating on the area of the substrate hit by the macrocluster beam.

Phrasor's device uses a single cluster emitter to clean a circular area measuring approximately two inches in diameter in less than two minutes. Cleaning a wafer measuring 200 mm or larger would require a compact source consisting of one or more linear arrays with each array containing four to eight macrocluster emitters. The tool "would have several nozzles like a rake that would scan across the surface, sort of like a sprinkler system," Perel says. Mahoney speculates that a fully developed system could clean a wafer in 30 to 60 seconds.

The results of particle-removal studies conducted in their lab using surfaces seeded with calibrated 1-µm particles have been verified with before-and-after SEM photographs. Additional tests showed that the process also removed SiO^₂ particles on a silicon surface as well as organic films usually too stubborn for banishment by ultrasonic cleaning, points out Mahoney, who as company vice president is in charge of the development project. A SIMS analysis by an outside laboratory confirmed the film removal results, he adds. The outside analysis also verified that the technique did not deposit additional particulates.

Perel and Mahoney have had preliminary discussions about developing a production tool with a few semiconductor equipment manufacturers whose names they prefer to keep under wraps. The partners want to conduct further tests but they need more advanced equipment to test the patent-pending cleaning process, specifically a proper vacuum chamber in which to run an experimental prototype through its paces.

Given their belief in the potential of the technology, both men are eager to prove its capabilities. "What we need is the money," laments Perel. Adds Mahoney: "If someone wants to build a prototype in their lab, that's fine. What we want is a [good] system to test this process, because it cleans like a charm. The size of the cluster is the same order of size as the contaminant. Size matches size beautifully."

The Phrasor veep mentions the huge dollops of cash that chipmakers spend on analytical tools and wonders aloud how a vacuum-compatible system using the company's process could benefit the industry. "The semiconductor industry spends a lot of money on diagnostic equipment. It seems to me that if you get this in situ cleaning system, you could save a lot of money."

Mahoney knows what even a small fraction of that sum could do for Phrasor's development process. "I wish we had one-billionth of the amount that the semiconductor industry spends on diagnostic equipment."