Making Fuels Synthetically

In a scene reminiscent of a television commercial, a man holds a glass of water-white liquid and extols the virtues of a process for making ultraclean diesel fuel.

But there's no soothing background music or lush green fields in the scene, and the speaker isn't an actor on a movie set. He's a scientist at a catalysis conference.

"Diesel fuel made from natural gas using the latest generation of Fischer-Tropsch catalysts is nearly as clear as this water," says Krijn P. de Jong, nodding toward the glass. De Jong, a professor of inorganic chemistry and catalysis at Utrecht University, in the Netherlands, notes that the catalytic chemistry that transforms basic feedstocks into liquid fuels has been known for decades, yet the technology has not been commercialized on a global scale.

But that situation may change someday, in part due to recent developments that have deepened understanding in a number of areas, such as the dependence of catalytic performance on catalyst particle size and structure, the effects of water, and the nature of the active catalyst.

More than 80 years ago, Franz Fischer, director of the Kaiser Wilhelm Institute for Coal Research, in Mülheim an der Ruhr, Germany, and Hans Tropsch, a research group leader, demonstrated that liquid hydrocarbons could be prepared by reacting carbon monoxide with hydrogen in the presence of solid metal catalysts including iron, cobalt, and ruthenium. The gaseous mixture of CO and H₂, now known as synthesis gas or syngas, was prepared by treating coal with oxygen and steam at high temperature.

Later referred to as Fischer-Tropsch synthesis, the process for polymerizing CO and hydrogen was patented in Germany in 1925. The method was used there during World War II to produce synthetic fuels and other petroleum products, which were relatively scarce during the war, from Germany's abundant supply of coal. The first-generation Fischer-Tropsch technology was also used in Japan at that time, although on a smaller scale than in Germany.

As de Jong recounts the history, the discovery of enormous petroleum reserves in the Middle East in the 1950s reduced industry's interest in synthetic fuels. But the technology's prominence rose quickly following the U.S. oil crises of the 1970s, which led to the development of a second generation of Fischer-Tropsch plants and catalysts.

At that time, oil embargos imposed on South Africa during that country's apartheid era led South Africa-based Sasol to commercialize Fischer-Tropsch technology for making transportation fuels from coal. Likewise, Shell built a commercial facility in Bintulu, Malaysia, for converting natural gas to diesel fuel and other products (by way of syngas) through Fischer-Tropsch technology commonly referred to as the gas-to-liquids (GTL) process.

"Now the third generation of Fischer-Tropsch technology is being developed," de Jong says. He notes that the key forces driving the current push to broaden the use of the synthesis process are today's high oil prices and demanding environmental regulations.

Although the coal- and gas-to-liquids processes are well-established and relied upon currently to produce, for example, roughly half of South Africa's transportation fuel, the techniques' popularity has grown rather slowly over the years. The slow growth is due in part to high costs associated with building and maintaining the plants and the availability until recently of relatively inexpensive crude oil.

As petroleum prices have soared, however, Fischer-Tropsch synthesis has become more competitive economically. In addition, the synthetic fuels are essentially free of sulfur and nitrogen compounds, which would allow suppliers to sidestep the demanding procedures for purifying petroleum-derived fuels to meet ever-tightening environmental laws.

Those factors and others have led a number of fuel manufacturers to embark on major plans for developing new GTL plants. For example, joint venture Sasol Chevron, in another joint venture with Qatar Petroleum, is scheduled to begin operations later this year in the Persian Gulf state of Qatar. Shell plans to open a GTL plant there in 2009. Other companies are also developing plans for commercializing GTL chemistry. Overall, within a decade or so, natural-gas-derived diesel is predicted to account for some 10% of the global supply of diesel fuel, according to some industry experts (C&EN, July 21, 2003, page 18).

Regardless of the starting material used for preparing syngas, at the heart of Fischer-Tropsch synthesis lies the metal-based catalyst upon which subsequent carbon-carbon coupling reactions occur. The two metals employed as commercial catalysts, iron and cobalt, exhibit distinct properties. For example, using iron-based catalysts leads to a product distribution with a large fraction of olefins relative to paraffins, according to Burtron H. Davis, associate director of the Center for Applied Energy Research at the University of Kentucky, Lexington. In addition, iron catalysts make a range of oxygenated products. In contrast, cobalt catalysts give mainly paraffins, including valuable waxes used in the food industry.

Price is another important difference between the metals. Cobalt is much more expensive than iron, Davis remarks, so the costly metal needs to be highly active, long lasting, and used efficiently.

A common strategy for squeezing as much performance as possible from a solid catalyst is preparing the material in the form of tiny dispersed particles to maximize the surface area and hence the number of catalytically active sites accessible to reagent molecules. But as de Jong points out, even for a particle as small as 100 nm in diameter, 99% of the atoms reside in the interior of the particle, where they cannot participate in surface catalysis. So even smaller particles are needed.

Over the years, a number of researchers have probed the effects of particle size on catalyst performance. A significant study and review in that area was conducted in the 1990s by University of California, Berkeley, chemical engineering professor Enrique Iglesia. He found that reducing the cobalt catalyst's particle size from approximately 200 nm to 10 nm led to an increase in Fischer-Tropsch catalytic activity, as expected on the basis of increased surface area. But the turnover frequency, a measure of catalytic activity per surface site, was not influenced by cobalt particle size over that size range, according to the study. That finding suggests that even smaller particles need to be tested to find an optimum size for cobalt Fischer-Tropsch catalysts.

When it comes to particles below the 10-nm size range, however, little consensus exists about their behavior. Some research groups find abrupt changes in activity. Others do not. A clue to the source of the discrepancies comes from a study by Davis' group, in which the team examined the role of steam, which is ubiquitous in these reactions, on the interactions between the metal particles and the oxide support material upon which they are dispersed.

Through a combination of X-ray absorption spectroscopy techniques and catalytic activity studies, Davis, Gary Jacobs, Mingsheng Luo, and their coworkers found that under typical Fischer-Tropsch reaction conditions, cobalt particles with diameters less than 10 nm are easily oxidized in the presence of steam to cobalt oxides such as CoO. The oxides, in turn, can react irreversibly with an alumina support to form cobalt-aluminate species (Appl. Cat. A 2004, 270, 65). The upshot is that under some conditions, common catalyst supports may interact too strongly with small cobalt particles and in effect, swallow the catalyst particles.

To sidestep those problems, de Jong and colleagues developed synthetic methods for dispersing catalyst particles on an inert support material: carbon nanofibers. The group has been involved with synthesis and analysis of the fibers for the past several years and reports that the material, which forms an interwoven mesh, is mechanically strong and has high surface area. The team working on the project includes G. Leendert Bezemer, formerly a graduate student working with de Jong at Utrecht and now a research scientist at Shell in Amsterdam; Herman P. C. E. Kuipers, a manager at Shell; and their coworkers at those institutions and at Delft University of Technology.

Armed with the unconventional and inert support, the researchers prepared a large number of well-defined cobalt catalysts spanning the size range from approximately 2.5 to 25 nm and then tested the materials' propensity for catalyzing Fischer-Tropsch reactions. "We see a very sharp maximum in activity at 6 nm," de Jong says, referring to a plot of activity per gram of cobalt versus particle size.

Presenting the results in an alternative way, de Jong explains that the data show that surface specific activity (the turnover frequency) does not vary with particle size as the size is reduced from 25 nm to 6 nm. But as the particle diameter is decreased further, the team observes a sharp drop in activity and a decrease in product selectivity. Those results were measured at a syngas pressure of 1 bar. At a pressure of 35 bars, the falloff point is 8 nm (J. Am. Chem. Soc. 2006, 128, 3956). According to de Jong, the study suggests that decreasing the size of today's commercial catalysts—which he estimates to have a particle size close to 20 nm—may improve the efficiency of cobalt usage by a factor of two to four.

"It's very surprising to see such strong particle size effects for such large particles," Kuipers remarks. Effects tied to metal catalyst particle size are more commonly observed with 1- to 2-nm particles, he points out. Nonetheless, the take-home message is clear. Manufacturers should strive for this critical particle size when preparing catalysts, Kuipers urges. Larger particles make inefficient use of cobalt, and smaller particles are less useful catalysts.

Precisely why surface activity falls off so sharply for particles smaller than 6 or 8 nm in diameter is still something of an open question. De Jong, Kuipers, and colleagues have considered several possibilities. For example, oxidation may degrade the catalysts. The team reasons that perhaps the problem affects only the smaller nanoparticles because they are more readily oxidized than larger ones.

That explanation is supported by thermodynamic calculations carried out by Eric van Steen, a chemical engineering professor at the University of Cape Town, in South Africa, and coworkers at Sasol. The researchers found that under common Fischer-Tropsch reaction conditions, cobalt particles smaller than 4.4 nm in diameter are likely to be oxidized (J. Phys. Chem. B 2005, 109, 3575).

Another possibility is that the smallest nanoparticles mediate unwanted reactions that form cobalt carbide surface species, thereby triggering catalyst deactivation. But de Jong says his team, using in situ X-ray absorption spectroscopy techniques, looked for signs of oxidation and carbide formation and found no evidence that those types of reactions play a role in fouling the catalysts. To the contrary, the researchers report that even the smallest nanoparticles appear to remain in a metallic state throughout the catalytic reactions.

The X-ray investigation also revealed that cobalt's coordination number changes as the particles are exposed to syngas, and this effect suggests that the catalyst surfaces undergo reconstruction and become facetted in the presence of reagents. That finding may turn out to be the key to understanding the particle size effects.

As de Jong explains, surface steps—flat regions separated in height by atomic distances—are common features on reconstructed crystal surfaces. Such steps have been fingered by several researchers as the specific surface sites that mediate chemical reactions.

For example, using computational methods, Rutger A. van Santen, a professor at Eindhoven University of Technology's Schuit Institute of Catalysis, in the Netherlands, and coworkers showed that monoatomic steps on ruthenium, which has structural and catalytic properties similar to those of cobalt, significantly lower the activation barrier to CO dissociation, a key first step in Fischer-Tropsch reactions (J. Phys. Chem. B 2003, 107, 3808). Peijun Hu, a chemistry professor at Queen's University of Belfast, in Ireland, came to much the same conclusion in a study focused on cobalt surfaces (Surf. Sci. 2004, 562, 247).

Those findings and others led de Jong and coworkers to propose that particles with 6- or 8-nm diameters are the optimum size because they are the smallest particles with domains that are large enough to stabilize the monoatomic steps and other types of surface features needed to catalyze Fischer-Tropsch reactions efficiently. Smaller particles just don't have what it takes, they say. It's a combination of CO-induced surface reconstruction and this unusual type of size dependence that makes 6- or 8-nm particles just the right size for the job.

Fischer-Tropsch chemistry has been around for decades, and yet the technology is still in its infancy, Kuipers says. As with any chemical conversion, there still is room for increasing activity and selectivity and improving catalyst stability and lifetime.

"But the catalyst itself isn't everything," Kuipers insists. He adds that it is quite important to understand the catalyst's behavior—for example, in the reactor environment—in a way that can lead to process optimization.

Having a child is only one part of parenthood, Kuipers notes by way of analogy. Understanding how a child thinks and the way in which he or she learns is critical to making good choices when it comes to raising kids, he says.

It's taken about 80 years, but Fischer-Tropsch catalysts are starting to grow up.