A MONUMENTAL research effort has yielded unprecedented gold-oxo complexes, those in which one ligand is a single oxygen atom multiply bonded to gold. These complexes, once thought improbable if not impossible to make, could provide new insights into the mechanisms of catalytic oxidation reactions. The work might also help inorganic chemists further utilize specialized ligands to coerce transition metals into doing more of the unexpected.
Synthesizing and characterizing the gold-oxo complexes involved 18 chemists from seven research groups in North America, Europe, and Asia, according to Emory University's Craig L. Hill, who directed the project. Each member of the team "contributed top expertise in 16 different confirmatory spectroscopic, physical, and chemical methods to be sure of these unusually controversial compounds," Hill told C&EN. He described the work at a Division of Inorganic Chemistry symposium during last month's American Chemical Society national meeting in Boston. The research has since been published (J. Am. Chem. Soc. 2007, 129, 11118).
Terminal metal-oxo complexes, LnM=O (where L is a ligand), are common for early and middle transition metals-those in groups 3 to 8, Hill explained. Some examples, such as molybdenum (group 6) and iron (group 8), play important roles in proteins. Other examples, such as tungsten (group 6) and vanadium (group 5), are important in research and industrial chemical production.
Single oxygen atoms are good ligands for these transition metals because oxygen is a strong electron donor and the two or more electron pairs being donated can delocalize into vacant d orbitals on the metal, he noted. But the metal-oxo group becomes progressively less stable in moving from left to right across the periodic table. By group 9, metal-oxo complexes are rare.
Inorganic chemists have speculated for some time about whether or not terminal metal-oxo complexes of the late-transition-metal elements in groups 10 to 12 might exist. Several possibilities include complexes that are transient intermediates in copper-based enzymes or on the surfaces of precious-metal oxidation catalysts. But so far, there's little to show from numerous studies, Hill said.
"For years, conventional wisdom has held that an 'oxo wall' exists along the vertical boundary between groups 8 and 9," explained Clark R. Landis, a chemical bonding specialist at the University of Wisconsin, Madison, whom C&EN asked to comment on the gold-oxo work. "Metals to the left of the wall can support terminal oxo ligands, whereas group 10 and 11 metals only ligate oxygen atoms by using them to bridge between two metal atoms," Landis said.
The presumed origin of the oxo wall is the large number of d valence electrons in late transition metals, Landis noted. These electrons are forced to reside in antibonding orbitals in the metal-oxo unit, and the strong repulsion between these lone pairs of electrons and oxygen's donated electrons is overwhelming.
"Late-transition-metal-oxo complexes are so unusual, and the concept of the oxo wall so entrenched, that acceptance by a skeptical chemistry community demands extraordinary evidence of their existence," Landis said. "Hill and coworkers have now provided just that."
PRIOR TO 2004, there was only one example of a metal-oxo complex containing a transition metal beyond group 8. This iridium-oxo complex representing group 9 was reported in 1993 by the late Nobel Laureate Geoffrey Wilkinson of Imperial College London and coworkers.
Hill's group started to make progress in this arena recently by taking advantage of the delocalized metal-based molecular orbitals in known polyoxometalate ligands such as the tungsten-oxide cluster [PW9O34]9-. These ligands are capable of accommodating some of the additional electron density and providing the needed stability for late-transition-metal-oxo complexes to form. Hill and colleagues first met success with group 10 by synthesizing a platinum-oxo complex in 2004. A year later, his group repeated the feat by preparing a palladium-oxo complex.
With these examples in hand, Hill excitedly turned to gold, which is in group 11. Hill and coworkers reacted aqueous solutions of AuCl3 and tungsten-oxide cluster ligands in the presence of O2 to synthesize two gold-oxo complexes. The complexes were isolated as salts: K15H2[Au(O)(H2O)(PW9O34)2]•25H2O and K7H2[Au(O)(H2O)(PW10O35)2(H2O)2]•27H2O.
In the first compound, the [PW9O34]9- ligands look like butterfly wings attached to either side of the molecule's core, which consists of the gold atom along with oxo and water ligands. The second compound, which has two bridging tungsten units connecting the two [PW9O34]9- wings, is more stable in aqueous solution and was prepared to make it easier to characterize the first compound.
Despite the complexity of the molecular formulas and the structures they represent, the tricky part of the work was proving that the terminal oxygen atom was indeed attached to the gold atom. Hill and his large cast of coworkers utilized X-ray crystallography, X-ray absorption spectroscopy (carried out at the Stanford Synchrotron Radiation Laboratory), neutron diffraction (carried out at Argonne National Laboratory), 17O nuclear magnetic resonance spectroscopy, elemental analysis, redox titrations, electrochemical measurements, and more to justify their claim.
Altogether, seven methods were used to confirm that gold—and not tungsten—is the central atom in the complexes, as similar reactions with Au(III) solutions in the past have always led to tungsten-only products. It turns out that the secret for obtaining the gold-oxo complexes is to prepare the ligands in situ and carefully adjust the pH and ionic strength of the aqueous solution, Hill said. The neutron diffraction work further confirmed the presence of the terminal oxo ligand by discounting the possibility of it being a hydroxyl group, Hill added.
IN ANY CASE, one piece of confirmatory data stands out for Hill: the extraordinarily short Au-O bond length of about 1.76 ??. It's the shortest reported Au-O bond and was observed in both X-ray and neutron diffraction studies. It appears clear that it's a gold-oxygen unit with considerable multiple-bond character, Hill said, as other complexes with bridging Au-O single bonds have bond lengths of 1.90-2.10 ??.
For a terminal metal-oxo complex to form, there must be vacant orbitals of appropriate orientation and symmetry to receive the donated electrons from the oxo ligand. For this reason, terminal oxo complexes are favored by early transition metals with few d electrons, metal atoms in high oxidation states, and metal atoms with few ligands, explained Paul R. Sharp, a chemistry professor at the University of Missouri, Columbia. Sharp's group has synthesized bridging metal-oxo complexes utilizing metals in groups 9 to 11.
But moving to the right in the periodic table, the oxidation state and coordination number become more important, Sharp pointed out. Wilkinson's iridium-oxo complex "nicely illustrates this point," he said. The Ir(V) oxidation state is unusually high for iridium, and the coordination number for the ligands is four, a relatively low number. "And until Hill's results on platinum, iridium was the farthest transition metal to the right of the periodic table to show a terminal oxo complex," he reiterated.
"But if we look at Hill's complexes, none of these ideas is followed," Sharp observed. "The complexes are octahedral, which means a high coordination number of six. And gold is in the common Au(III) oxidation state. The gold atom shouldn't have any vacant orbitals in which to receive the electrons!
"Yet, the short Au-O bond distance indicates that the metal center is accepting π electrons from the oxo ligand," Sharp continued. "In fact, the geometry looks very similar to that of early-transition-metal terminal oxo complexes, such as tungsten. It's as if gold is acting like tungsten, which is five periodic groups away.
"To me, this feels like a good magic performance in which the magician levitates an object," Sharp said. "I can't believe it is real, but I am unable to see how the illusion is accomplished. In the case of the gold-oxo complexes, nature is the magician, and it will take some time to sort out exactly what is going on here."
The fact that gold-, palladium-, and platinum-oxo complexes exist means that "they could well be intermediates in a host of technologies based on the workhorse noble-metal catalysts and O2, ranging from automobile catalytic converters to fuel cells to metal catalysts for greener industrial oxidations," Hill said. These compounds could provide insight into the workings of gold nanoparticles that are becoming popular for activating O2 in air. Some examples include oxidizing CO to CO2, converting alkenes into epoxides, and producing chiral organic intermediates, he noted.