Elizabeth K. Wilson
C&EN West Coast News Bureau

Supercritical carbon dioxide has come to be known as an environmentally friendly solvent, offered up as a cure for the woes caused by industrial processes that rely on noxious apolar solvents such as acetone, methanol, or toluene. While ordinary CO₂ is apolar, benign, and plentiful, it's also a gas. But under high pressures (1,070 psi) and modest temperatures (31 C), CO₂ reaches its supercritical point and becomes dense like a liquid yet maintains its gaslike ability to flow with almost no viscosity or surface tension.

Although supercritical CO₂ currently is the focus of unprecedented interest, scientists and engineers have been exploiting its benefits for years. It's nontoxic; it evaporates instantly once the pressure is released, leaving almost no trace behind; and it's extremely inexpensive. Those features make it ideal for industrial food processing, such as coffee decaffeination.

But supercritical CO₂ has other unusual advantages that perhaps are not quite so well known. Scientists are finding that supercritical CO₂ is an intriguing medium in which to discover and develop new materials or to improve the way old materials are made.

For instance, altering the pressure within the supercritical region allows scientists unique control of the chemistry of processes involving supercritical CO₂, such as nanocrystal growth. And because the substance can infiltrate pores or nanosized canals on a surface, supercritical CO₂ turns out to be a natural in fabricating surface coatings or polymeric foams.

During the 1980s, supercritical fluid research was centered primarily around separations, "but now, there's more emphasis on aspects of materials chemistry," said James J. Watkins , one of the organizers of a symposium on the topic held at the American Institute of Chemical Engineers' (AIChE) annual meeting in Los Angeles last month.

	Watkins (left) and Blackburn deposited a nickel film onto an etched wafer (right) by continuous chemical fluid deposition.

In particular, electronic materials manufacturing is "an area in which supercritical fluids are not conventionally involved," noted Watkins, an assistant chemical engineering professor at the University of Massachusetts, Amherst. But the area has proven to be an attractive target for chemical engineering research.

Chemical vapor deposition is a technique frequently used to deposit metals for microelectronic structures. In this vacuum process, a vapor of organometallic precursor molecules passes over a surface. The precursor then reacts, usually via chemical or thermal reduction, leaving a thin metal coating on the surface and releasing gaseous by-products. But the precursors must be extremely volatile in order to get a high enough rate of mass flow in an industrial apparatus. If the rate isn't high enough--a frequent occurrence--the coatings can be nonuniform, especially within very small or intricate surface features.

Watkins has pioneered a technique similar to chemical vapor deposition, but with a twist, which he calls chemical fluid deposition. Rather than existing as a vapor, the precursor dissolves in a bath of supercritical CO₂. Because of the CO₂'s low viscosity and surface tension, the solution flows effortlessly over the surface. In addition, the concentration of the precursor is several orders of magnitude higher than that in chemical vapor deposition, which leads to a uniform coating of very complex topographies, Watkins noted.

In this manner, Watkins and his colleagues have coated surfaces with an array of metals, including platinum, palladium, rhodium, and nickel. At the AIChE meeting, Jason M. Blackburn, a graduate student in Watkins' group, reported on depositing nickel and platinum films on silicon substrates that uniformly coat features 100 nm wide by 1 m deep.

As for industrial feasibility, the cost of maintaining a vacuum environment for the vapor may balance the cost of pressurization needed in the supercritical fluid technique. "With chemical fluid deposition, you're trying to keep pressure in, whereas with chemical vapor deposition, you're trying to keep the atmosphere out," Blackburn said.

Integrated circuits, the mainstay of modern electronics, start their lengthy and solvent-intensive manufacturing process as silicon wafers coated with polymers impregnated with a radiation-sensitive compound. In some processes, selected portions of the coating are exposed to ultraviolet light, and the light-sensitive compound, known as a photoacid generator, produces an acid that reacts with the polymer, rendering it insoluble. The remaining, unexposed portion of the coating can be washed away.

The foundation polymer layers are prepared by a method known as spin-coating that works just like it sounds: the coating material--polymer, photoacid generator, and an organic solvent--is dropped onto the wafer, then spun rapidly around. Centrifugal force spreads the material out into a thin, even layer.

Liquid CO₂ may be able to replace the organic solvent used in spin-coating, according to Erik N. Hoggan, a graduate student at North Carolina State University who works with chemical engineering professor Ruben G. Carbonell , codirector of the Kenan Center for the Utilization of CO₂ in Manufacturing & Technology. Hoggan reported the development of the group's unique process at the AIChE meeting.

University of North Carolina, Chapel Hill, graduate student Devin Flower (far left) and NC State's Hoggan investigate spin-coating with liquid CO2, such as the spun polymer film.

After the creation of the film, the next steps can be performed with supercritical CO₂. But because the usual photoacid generators aren't soluble in supercritical CO₂, the group had to custom synthesize a new compound, 2-perfluorohexyl-6-nitrobenzyl tosylate, which is soluble. Fluorine-based compounds are known to dissolve well in supercritical CO₂, and consequently scientists frequently consider fluorine in their design of compounds used with the solvent.

Industrial use of liquid CO₂ to spin-coat films is still a few years away, Hoggan noted, in part because researchers still need to improve film quality. Right now, films can be produced that have a 5% variation in film height. But for commercial applications, variations below 1% are needed.

Supercritical CO₂ may offer itself as a tool for controlling the processing and behavior of nanocrystals. These metal particles, usually between 20 and 100 in diameter, have electronic and optical properties that are different from bulk materials and have numerous potential applications in areas such as microelectronics or biological tagging. Nanocrystals are typically grown in acetone or other solvents, but graduate student Parag S. Shah, at the University of Texas, Austin, reported at the AIChE meeting his group's experiments with nanocrystals in supercritical CO₂.

Shah and chemical engineering professors Keith P. Johnston and Brian A. Korgel believe they can control a number of nanocrystal properties with the solvent--for example, altering the pressure or temperature of supercritical CO₂ changes the solvent's density, which could make it possible to precipitate nanocrystals of a specific size. Nanocrystals are usually coated with molecules that keep the particles from aggregating into a clump, and the group was able to accomplish such steric stabilization in supercritical CO₂.

The use of supercritical CO₂ in materials research is not reserved exclusively for microelectronics. Because the solvent has such unusual properties, it stands to reason it would affect surface chemistry. Consider the case of additives for copy machine toner, which are generally surface-modified silica powders. The surfaces of oxides like silica are typically quite polar, readily adsorbing atmospheric water. Consequently, the additive particles need to be coated with something that will keep them dry.

There are a number of standard chemicals used to treat the surface of metal oxides, such as hexamethyldisilazane and octadecyltrichlorosilane. But the solvents used in the process, such as cyclohexane or toluene, present problems even beyond their environmental toxicity. "Most of the solvents physisorb to the silica, even after aggressive drying," said chemical engineer James R. Combes of Xerox in Mississauga, Ontario. The result is a cake of toner that has to be ground up to return it to its powder state.

Combes reported that his group can get around many of these processing headaches with supercritical CO₂. Because depressurization instantly releases the gas, there's little or no caking, and the process is almost entirely waste-free. And CO₂'s relative purity means water can be readily removed from silica.

The success of this approach has implications for supercritical CO₂ and surface chemistry in general. The absence of normally omnipresent water, along with the solvent's pressure tunabil-ity, creates a whole new research landscape. "If you can get rid of the atmospheric water, it opens unique synthetic pathways," Combes said.

Supercritical CO₂ also can be used in investigations of a type of conductive foam. The foams are a blend of a polypyrrole--a conducting polymer--and another polymer such as polyurethane. The combination gives the foams elasticity and processibility. Their syntheses, not surprisingly, require organic solvents such as methanol. The host polymer is impregnated with an oxidant, which oxidatively polymerizes pyrrole to polypyrrole within the host.

Can Erkey and Robert A. Weiss, chemical engineering professors at the University of Connecticut, and their colleagues began looking at supercritical CO₂ as a way to circumvent the solvent problem while working with Rogers Corp. in Rogers, Conn. Because the commonly used oxidants such as FeCl₃ have extremely low solubilities in supercritical CO₂, the group developed alternatives. One of their compounds, ferric triflate, becomes relatively soluble with the addition of a small amount of ethanol.

The group was proceeding with the development of this process when Erkey and his colleagues discovered in some older literature that plain iodine was a potential pyrrole oxidant. Even better, they found iodine's vapor pressure indicated it could be soluble in supercritical CO₂ even without ethanol, Erkey said. And indeed, it was. The solubility was "a lot higher than ferric triflate--and iodine is dirt cheap," he said. They now have a greatly simplified, relatively inexpensive process for making conductive foams in supercritical CO₂.

They hope the foams might be useful as "electronic noses"--sensors that could identify chemicals through a change in conductivity.

Another useful feature of supercritical CO₂ is its ability to swell polymers, which opens their pores for easy chemical access. Of course, many liquid solvents will do that too, but driving off all residual solvent traces within a polymer's intricate chambers can be a chore. With supercritical CO₂, all it takes is a release of pressure. And its low viscosity also means it can penetrate porous materials more deeply.

These properties lend themselves to a growing body of research on polymers as drug delivery devices. Drug-laden polymers that release substances slowly, either into the body or at a specific site such as a target organ, would help ensure steady concentrations of drugs where they're needed. A natural polysaccharide known as chitin, found in crustacean shells and insect exoskeletons, and its deacylated derivative, chitosan, transport oxygen well, making them promising for biomedical uses such as drug-impregnated wound dressings.

Getting the drugs into the pores of a polymer requires swelling and good transport--and supercritical CO₂ may be one solution. Randy D. Weinstein , a chemical engineering professor at Villanova University, Villanova, Pa., and his colleagues tested the effect of supercritical CO₂ on chitin, loading the polysaccharide with acetaminophen and ketoprofen. They're still not sure whether or not the drug loading is uniform or what will happen when the CO₂ pressure is released; this is now under investigation.

The group has a broader goal for its experiments, that of using the results to develop predictive models of how different chemical structures affect a drug's solubility in supercritical CO₂ and its ability to get inside a polymer's pores. The researchers would also like to understand how changes in supercritical CO₂'s temperature and pressure affects a polymer's swelling and ability to accept drugs.

"That's the ultimate goal," Weinstein said, "to design different polymers that can have different drug loadings just by changing CO₂ properties."

[Previous Story] [Next Story]

Top

Chemical & Engineering News
Copyright © 2000 American Chemical Society