Think "chiral chemistry," and what comes to mind? Perhaps a multi-billion-dollar industry driven by pharmaceutical applications. Or maybe the subtleties of the complex processes that make the valuable products.

	VIEW FULL SIZE IMAGE
	SPECIFIC Endowed with a handedness that is due to the arrangement of its surface atoms, a chiral Pt(643) electrode interacts enantiospecifically with D- (top) and L-glucose (bottom) molecules in electro-oxidation reactions. COURTESY OF GARY A. ATTARD/CARDIFF UNIVERSITY

By and large, commercial methods with the requisite finesse to steer complex reactions toward enantiomerically enriched products lie in the domain of homogeneous catalysis. But enantioselective reactions aren't limited to molecules free-floating in three-dimensional solutions. Two-dimensional surfaces also have a role to play. Though some enantioselective surface reactions were discovered decades ago, chiral surface chemistry has hardly been defined--let alone explored. Researchers around the globe are just beginning to address this emerging field's fundamental questions.

In some cases, metal surfaces that are usually achiral can be endowed with a handedness by decorating them with chiral molecules. The surface-bound molecules (adsorbates) may react with one another enantiospecifically or they may mediate stereoselective reactions. In other cases, atoms in a metal surface assemble in particular patterns, sometimes in response to the presence of adsorbed molecules, thereby converting a surface with mirror symmetry into a chiral one.

An example of adsorbed molecules interacting enantiospecifically was just discovered by researchers in Denmark studying dimerization of amino acids on solid surfaces. Aarhus University physics professor Flemming Besenbacher, graduate student Angelika Kühnle, and associate professors Trolle R. Linderoth and Bjørk Hammer find that cysteine molecules dimerize readily--but they're picky about their partner's chirality.

Using scanning tunneling microscopy to follow the surface stereochemistry, the group reports that depositing L-cysteine molecules on the (110) crystal face of gold produces dimers that appear as paired spots (distorted figure-eights). The features are aligned with a 20° clockwise rotation relative to certain lattice features. D-Cysteine molecules act just like their l counterparts, Kühnle explains, except that the rotation is counterclockwise, leading to the mirror image. "And when we use a racemic mixture," she asserts, "we only see ll dimers and dd dimers. We don't observe any heterochiral pairs on the surface" [Nature, 415, 891 (2002)].

The group also notes that, as cysteine binds to gold, the molecules sweep aside some gold atoms, restructuring the surface. Besenbacher and coworkers had seen similar effects in a previous study. By depositing chiral decacyclene molecules on a copper crystal and then moving them aside with an STM tip, the researchers found that the molecules had made a chiral hole--an imprint, of sorts--in the copper surface. The group also found that under some conditions, the surface could be decorated with well-ordered chiral holes [Angew. Chem. Int. Ed., 40, 2623 (2001)].

	CHOOSEY Formed from enantiopure reagents, LL- or DD-cysteine dimers on a gold crystal are distinguishable in scanning tunneling micrographs by their alignment relative to surface features (left = LL dimers rotated clockwise, center = DD dimers). In a racemic mixture (right), only homochiral pairs are observed. AARHUS UNIVERSITY

TO UNDERSTAND cysteine's stereoselectivity, the Aarhus team turned to quantum mechanical calculations. The group concludes that cysteine's behavior is driven by optimization of the bonds between gold and sulfur, between gold and the amino group, and between pairs of carboxylic groups. "It's a system that lets you see--at the molecular level--the three contact points needed for chiral recognition," Kühnle says.

Neville V. Richardson, a physical chemistry professor at the University of St. Andrews, in Scotland, remarks that the cysteine-dimer system is "an elegant and simple model example." Homochiral pairing is demanded in this case, he says, by the shape of the molecule, excavation of a hole in the gold surface, and the relative position of the three contacts. Richardson adds that the Aarhus work "provides a key to understanding chiral recognition in more complex systems" and helps advance chiral sensing and chiral heterogeneous catalysis.

Enantiospecific heterogeneous catalytic reactions are few in number, but a couple of examples in this area of surface chemistry boast impressive statistics. "There are two amazing systems that can generate extremely high enantioselectivities," says Yongkui Sun, a group leader at Merck Research Laboratories. The reactions, which were discovered a few decades ago by Japanese researchers, are hydrogenations of - and ,-ketoesters.

Modifying alumina-supported platinum with cinchonidine, a chiral alkaloid, produces a catalyst that can convert ethyl pyruvate to (R)-ethyl lactate with an enantiomeric excess on the order of 95%. Similar selectivity is seen when Raney nickel catalysts, modified with (R,R)-tartaric acid, are used to hydrogenate methyl acetoacetate to methyl (R)-3-hydroxybutyrate.

Several research groups have endeavored to understand the origin of the reactions' stereospecificity and to elucidate their mechanisms. Sun and coworkers at Merck, focusing on the cinchonidine/Pt system, find that enantiomeric excess depends on the modifier's concentration--but in a finicky sort of way. Too little cinchonidine, and the enantioselectivity is low. Too much, and it's low again. Only in the narrow range of five to 12 surface Pt atoms per modifier molecule does the reaction yield nearly enantiopure products under mild conditions [J. Am. Chem. Soc., 121, 4920 (1999)].

It's not surprising that, if the modifier's concentration is lower than the optimum level, then further decreasing its concentration will cause product enantiopurity to drop still lower. There just isn't enough chiral medium present in that case to drive the reaction selectively, Sun explains. But why enantiopurity drops when the modifier is added in excess of the optimum concentration is puzzling.

That finding led the Merck group to propose that cinchonidine's orientation on platinum changes as surface coverage increases. "Overloading the surface with cinchonidine forces the molecules to deviate from the optimal orientation for enantioselective hydrogenation," Sun says.

AFTER DEDUCING the optimum Pt-to-modifier ratio, Sun's group probed the practical problem of holding the ratio steady throughout the reaction. It turns out that cinchonidine itself is hydrogenated and destroyed in the acetic acid solution in which the reaction is carried out.

One way to ensure an ample supply of the modifier is adding it in large excess at the start of the reaction. But doing so means that a nonoptimal Pt-to-modifier ratio is used, and the enantioselectivity suffers accordingly. Coming up with an alternative solution, the Merck group developed a procedure for delivering carefully metered aliquots of the modifier as the reaction progresses, thereby maintaining the optimum ratio and high enantioselectivity throughout the reaction. Sun notes that the group is presently studying cinchonidine-hydrogenation kinetics to further understand the nature of the interaction between the modifier molecule and the platinum surface.

Getting a close look at the way cinchonidine adsorbs on platinum--close enough to discern something about molecular orientation and surface coverage--can be a formidable experimental task. Yet scientists at the University of California, Riverside, recently reported an advance in that area.

Chemistry professor Francisco Zaera and postdoctoral associate Jun Kubota, currently at Tokyo Institute of Technology, used a surface-sensitive infrared spectroscopy technique to study the manner in which cinchonidine's adsorption geometry changes with the modifier's surface coverage. The method probes the interface between a Pt foil and a cinchonidine solution.

	MODEL SYSTEM Besenbacher (left) and Kühnle of Aarhus University.
	PHOTO BY MITCH JACOBY

The Riverside team finds that in 5 to 20% cinchonidine solutions, the modifier molecule, which is anchored to Pt through its quinoline ring, lies flat on the surface. But at higher concentrations, where overcrowding may set in, the aromatic ring is tilted with respect to the surface [J. Am. Chem. Soc., 123, 11115 (2001)]. The researchers note that the concentration range that leads to a flat-lying geometry corresponds to the optimum Pt-to-modifier ratio observed by the Merck group.

Zaera points out that, although procedures for collecting IR spectra in a reflection mode from a liquid-solid interface have been described, they're challenging to implement and typically give weak signals. The methods can discriminate between analyte molecules (cinchonidine, in this case) adsorbed on a surface and those in the liquid phase because of differences in the way those two classes of analyte molecules interact with linearly polarized light.

A new procedure developed by Zaera's group for measuring and comparing spectra recorded with light of two distinct polarizations enables the researchers to scan their specimens quickly and conveniently and produces strong IR absorption signals. The technique can be applied to adsorbate characterization in a range of liquid-solid interface studies.

By branching out to model systems, other metals, and an assortment of experimental procedures, researchers are collecting additional clues about chiral modifiers' interactions with metal surfaces. For example, Richard M. Lambert, a chemistry professor at the University of Cambridge, uses X-ray absorption studies and other techniques to conclude that quinoline--the selfsame moiety that anchors cinchonidine to Pt--lies nearly flat at 300 K and binds to the surface primarily via the molecule's aromatic -system. Raising the temperature to 360 K causes almost no change in quinoline's adsorption geometry, Lambert notes, but increasing the surface coverage forces the molecules to tilt slightly. A steeper tilt angle is observed when lepidine, a substituted quinoline, is deposited on Pt [Surf. Sci., 498, 212 (2002)].

At the University of Liverpool in England, chemistry professor Rasmita Raval has studied copper crystals modified with tartaric acid using STM and a surface electron diffraction technique. "We find that between rows of adsorbed molecules lie empty chiral channels that expose bare metal atoms," Raval says. She proposes that the nanosized spaces may constitute "the actual active enantioselective site." Molecules adsorbing at these channels would bind in a preferential orientation, forcing hydrogenation to occur at just one reactant face [Nature, 404, 376 (2000)].

	IN THE LAB UC Riverside chemists (top photo, from left) Zaera, Hansheng Guo, Lora J. Shorthouse, and Ilkeun Lee. Bottom photo: Carnegie Mellon researchers (from left) Gellman, Greg Rohrer, and Sholl.

Supported palladium catalysts are major players in surface chemistry in general, and have been examined by a number of researchers for their usefulness as chiral heterogeneous catalysts. Reviewing the topic from the perspective of fine chemicals synthesis, Hans-Ulrich Blaser of Solvias in Switzerland notes that Pd catalysts modified with cinchona alkaloids have been used to hydrogenate carbon-carbon double bonds in ,-unsaturated acid derivatives and ketones [J. Mol. Catal. A: Chem., 173, 3 (2001)]. The catalyst has been used to hydrogenate 2-ethyl-2-pentenoic acid, for example, in a reaction yielding up to 66% enantiomeric excess.

Pd catalysts modified with vinca-type alkaloids and other compounds have been prepared by chemical engineering professor Antal Tungler of Budapest University of Technology & Economics. The work has produced catalysts that lead to moderate enantiomeric excesses. Tungler notes that, unlike the cinchonidine/Pt system where reaction rates (and enantiopurity) are boosted by the modifier's presence, chiral hydrogenations using modified-Pd catalysts are slower than reactions using unmodified Pd [J. Mol. Catal. A: Chem., 173, 231 (2001)].

Back in the domain of single-crystal surface science, some researchers take advantage of naturally occurring chiral arrangements of atoms in metal surfaces to furnish well-characterized chiral specimens. Zigzag patterns and atomic steps (a nanometer-scale staircase) present, for example, on Pt(643) and Cu(643), can endow metal surfaces with chirality. By following established methods for orienting, cutting, and preparing crystalline specimens, exposing crystal faces with these features is a relatively straightforward task.

Gary A. Attard, a physical chemistry professor at Cardiff University, in Wales, compared the electro-oxidation behavior of l- and d-glucose and other sugars on chiral and achiral Pt electrodes. Among other findings, Attard demonstrated via cyclic voltammetry that chiral platinum electrodes oxidize glucose isomers enantiospecifically [J. Phys. Chem. B, 105, 3158 (2001)].

Meanwhile, at Carnegie Mellon University, chemical engineering professor Andrew J. Gellman and graduate student Joshua D. Horvath investigated desorption of chiral molecules from chiral Cu(643) and achiral Cu(111). The team finds that (R)-3-methylcyclohexanone and (R)- and (S)-propylene oxides exhibit desorption energies that depend on the chirality of the molecule-surface combination [J. Am. Chem. Soc., 124, 2384 (2002)].

"These experiments provide insight into the adsorption sites of the chiral molecules on Cu(643)," Gellman notes. Desorption-spectrum features found exclusively in the Cu(643) experiment can be attributed to molecules that bind to Cu atoms that comprise the chiral surface patterns.

Also at Carnegie Mellon, David S. Sholl, who is an assistant professor of chemical engineering, has studied the potential usefulness of single-walled carbon nanotubes as enantiospecific adsorbents. Sholl notes that the manner in which graphitic rings wind about the tube axis endows the tubes with chirality. As such, carbon nanotubes could turn out to be a useful adsorption medium, for example, in chiral chromatography.

NO SUCH LUCK, Sholl reports. Together with graduate students Timothy D. Power and Anastasios I. Skoulidas, Sholl used computational methods to compare the heats of adsorption of pairs of enantiomeric dimethylcycloalkanes on carbon nanotubes with a range of pore sizes and chiral angles. The energy differences between enantiomers turn out to be negligibly small, suggesting that carbon nanotubes would make poor enantiospecific adsorbents [J. Am. Chem. Soc., 124,1858 (2002)].

Recent progress in chiral surface chemistry hardly exposes the tip of the iceberg of this emerging field. Currently, enantioselective heterogeneous catalysis plays no role in industrial processes. According to UC Riverside's Zaera, one reason that surface chemistry sits on the sidelines there is that small changes in most catalyst and reaction parameters can cause large and unpredictable changes in activity and selectivity. "This field is begging for some good surface-science characterization to give us a clue how to begin tuning the catalysts," Zaera says. "For these systems, basic knowledge can act like an arrow to show us which path to follow."

Chemical & Engineering News
Copyright © 2002 American Chemical Society