Lucky for King Midas, he wasn’t in the catalysis business. It was bad enough that the mythological king couldn’t eat because his touch morphed food into inedible gold. If he turned catalysts to gold, Midas likely would have thought that the conversion rendered them useless too—unless, of course, he was working in gold catalysis nowadays.
Gold’s long-standing reputation as a largely unreactive metal has been turned upside down over the past several years because researchers have discovered numerous reactions catalyzed by the precious metal. Eager to understand the basis of this unexpected reactivity, scientists have recently turned their efforts to elucidating the elementary steps that make up reaction mechanisms on the surfaces of gold catalysts.
Many of those steps have by now been deduced and published. Yet as researchers discuss the results, some issues continue to be debated. By uncovering details of these chemical sequences, scientists aim to gain insights that can lead to a predictive model for fine-tuning, and possibly commercializing, gold-driven catalytic processes.
“For many years gold was considered chemically inert—even uninteresting,” says Graham J. Hutchings, a chemistry professor specializing in gold catalysis at Cardiff University, in Wales. But that premise started to unravel in the late 1980s, when Masatake Haruta, now at Tokyo Metropolitan University, observed that carbon monoxide could be oxidized to carbon dioxide at low temperatures in the presence of supported gold particles measuring just a few nanometers in diameter. Around the same time, Hutchings reported that gold could also catalyze hydrochlorination of acetylene to vinyl chloride.
News of those studies soon triggered a modern-day “gold rush,” with researchers in many labs exploring gold’s potential to catalyze a range of reactions. Just within the past few years, the list of gold-catalyzed reactions has grown significantly. It now includes selective oxidation of alcohols to aldehydes and ketones, methyl glycolate synthesis from ethylene glycol, propene epoxidation, direct production of hydrogen peroxide from hydrogen and oxygen, and several other reactions—some of which are being studied for potential commercial applications (C&EN, Sept. 24, 2007, page 87, and Nat. Nanotechnol. 2010, 5, 5).
It’s not a surprise that this area of research quickly became a hot topic. “Gold is the poster child for nanocatalysis,” says D. Wayne Goodman, a chemistry professor at Texas A&M University. Supported metal catalysts of all types have long consisted of nanometer-sized crystallites of metals dispersed on oxides or other solids. The tiny crystal size ensures that a large fraction of the highly active (and typically expensive) metal sits on the crystal surface, where it can catalyze reactions, instead of hiding in the crystal interior.
“The interesting thing with gold is that in bulk form it’s a rather unreactive noble metal. But when prepared in ‘nano’ size, suddenly gold becomes very reactive,” Goodman says.
That observation puzzles researchers. “Just a few years ago, we would have thought that our understanding of metal catalysis was rather complete,” comments Claus H. Christensen, a senior scientist with Danish catalyst manufacturer Haldor Topsøe, in Lyngby. “But nanoparticulate gold challenges our understanding.”
The origins of nanosized gold’s catalytic prowess have not spontaneously presented themselves in a neat and tidy fashion as the unequivocal outcome of mechanistic studies. Yet those investigations have revealed many details of how reactions proceed on the surfaces of gold nanocrystals. For example, at the Institute of Chemical Technology at Polytechnic University of Valencia, in Spain, a team led by Avelino Corma found that nanoscale gold supported on cerium oxide is highly active in the aerobic (air) oxidation of alcohols. Other support materials lead to less active catalysts, they observed. On the basis of the oxidation kinetics of several types of alcohols, the group proposed a reaction mechanism that features three key steps.
In the initial step, an alcohol molecule in solution coordinates with a gold cation to form an alcoholate, a species in which the hydrogen of the alcohol’s OH group is replaced with gold. Then, a hydrogen atom bonded to the carbinol carbon (the alcohol center) migrates to the metal, thereby forming a gold hydride and converting the alcohol group to a carbonyl group. In the final step, oxidation of the hydride by molecular oxygen forms water and regenerates the gold ion (Chem.—Eur. J. 2008, 14, 212).
Molecular oxygen also plays a central role in the gold-catalyzed liquid-phase selective epoxidation of propene to propylene oxide, according to a recently published study by chemical engineers Harold H. Kung and Mayfair C. Kung of Northwestern University and coworkers. Guided by results of isotope-labeling experiments, the researchers concluded that the origin of the oxygen atom in the epoxide is a peroxide species formed from O2—not from the water-methanol solvent (Chem. Commun., DOI: 10.1039/c000374c).
Other researchers have studied related gold-catalyzed reactions and proposed similar mechanisms to account for their observations. But some scientists have proposed mechanisms that feature distinctly different sets of elementary steps that don’t involve molecular oxygen, ionic gold, or solution-phase reaction steps.
For example, Harvard University chemistry professor Cynthia M. Friend, graduate student Bingjun Xu, and coworkers have proposed mechanisms for processes in which a series of aldehydes couples with methanol to yield the corresponding methyl esters. According to the team, these reactions, which run below room temperature, are promoted by the presence of atomic oxygen on gold nanocrystals and do not require liquid-phase reagents or gold ions.
To prepare the catalytic surface, the group exposes a gold single crystal, Au(111), to ozone, which causes the assembly of gold clusters, roughly 2 nm in diameter, that are decorated with a sparse layer of atomic oxygen. Then, they treat the surface with methanol, which reacts with the oxygen to form adsorbed methoxy species and water.
In a recent study (Nat. Chem. 2010, 2, 61), Friend’s group used a surface prepared in that way to follow the reactions of four aldehydes: formaldehyde, acetaldehyde, benzaldehyde, and benzeneacetaldehyde. (The molecules have the general form RC(H)=O, where R is H, CH3, phenyl, and phenyl–CH2, respectively).
On the basis of isotope-labeling experiments, mass spectrometry, and other techniques, the Harvard group proposed that for all of the aldehydes studied, nucleophilic attack by the methoxy group at the aldehyde carbonyl carbon drives esterification. The attack forms an alkoxy hemiacetal surface intermediate, CH3O–C(H)(R)–O, they say. That species eliminates a hydrogen atom to form the ester, CH3O–C(R)=O.
As Friend explains, a general conclusion drawn from multiple surface studies of alcohols is that atomic oxygen’s presence on gold selectively activates chemical bonds (such as the one in OH) via Brønsted acid-base reactions. On that basis, Friend’s team reasoned that if O-atom-mediated reactivity is exhibited by alcohols, similar reactivity should also be exhibited by other types of molecules, such as amines.
In a follow-up study, Stanford University chemical engineering professor emeritus Robert J. Madix, who now serves as a senior research fellow at Harvard, teamed up with Friend’s group to explore reactions of dimethylamine with formaldehyde on oxygen-covered gold.
By using methods and conditions similar to the ones employed in previous studies, the researchers found that dimethylamine readily loses the amine H, thereby forming adsorbed (CH3)2N. Then, as in alcohol esterification, the deprotonated amine functions as a nucleophile and attacks the electron-deficient carbon in formaldehyde. The intermediate formed by that attack eliminates hydrogen to complete the acylation reaction, which produces N,N-dimethylformamide with nearly 100% selectivity at low oxygen coverage (Angew. Chem. Int. Ed. 2010, 40, 394).
Even as dozens of studies have revealed details of gold’s knack for mediating surface reactions, several issues remain unsettled. The chemical form of oxygen on gold, for example, is still an open question. Likewise, the role of the catalyst support material remains perplexing because some studies show a clear dependence of reactivity on the type of support. Others, however, show that reactions can proceed readily in the absence of any support (C&EN, Jan. 18, page 9).
The nature of the catalytically active site is another subject of debate. Several years ago, Goodman’s group studied custom-made model catalysts and found that a certain two-atom-thick gold structure catalyzed CO oxidation with exceptional activity (C&EN, Aug. 30, 2004, page 9). Yet other researchers have proposed that the molecular-scale sweet spot for catalytic activity lies at the interface between gold and its support or at the corners of nanoparticles, where low-coordinated metal atoms are exposed (Angew. Chem. Int. Ed. 2008, 47, 4835). “It is possible that different active sites are relevant for different reactions,” Northwestern’s Harold Kung says. “We just don’t know yet.”
That sentiment is echoed by other researchers. “As scientists, we would like to know simply that this mechanism is right and that that one is wrong,” Goodman says. But, he adds playfully, “when it comes to scientific ideas, it’s much easier to propose than dispose.” It is difficult to prove that a particular mechanism is invalid, he explains, because it might be operative for certain reactions under select conditions. According to Goodman, the key questions are, do those conditions represent an important reaction scenario, and which factors exert the strongest influence on catalysis?
It could be a while before the short and simple answers are uncovered. Until then, scientists will go on searching for clues to gold’s catalytic behavior and continue to help the precious metal dispel its reputation as lackluster, chemically speaking.