Complex and subtle, the motions of reagents undergoing stereospecific reactions tend to follow a kind of molecular choreography charted in three dimensions. But sometimes the dance occurs on or near a surface.
Chiral surface chemistry hasn't garnered the widespread attention of its all-solution-phase (homogeneous) counterpart. But if some of the homogeneous catalytic processes used for preparing enantiopure substances--for example, in the fine chemicals or pharmaceuticals sectors--were replaced with heterogeneous processes, the result could be cost savings and significant improvements in process efficiency.
For example, using solid, insoluble catalysts or catalysts fixed to solids would ease the laborious task of separating soluble catalysts from liquid products. The change would also make catalyst recycling more straightforward and could avoid problems of metal contamination associated with solution-phase systems. In addition, surface catalysis offers engineering flexibility and does away with the need to use some expensive materials. Yet despite its promise, chiral surface chemistry still hasn't made its industrial debut.
"In terms of performance and scope of applications, heterogeneous catalysis lags far behind homogeneous catalysis," says Yongkui Sun, director of catalysis and reaction discovery and development at Merck Research Laboratories, Rahway, N.J. Part of the reason is that heterogeneous systems--especially the ones that drive reactions enantioselectively--are relatively few in number and can be quite complex, and they are useful for only a limited number of reactants, Sun explains.
Yet some chiral surface systems boast impressive statistics. One of the most commonly studied examples is hydrogenation of α-ketoesters, such as ethyl pyruvate, on the surface of alumina-supported platinum catalysts that have been treated with a chiral alkaloid modifier--typically cinchonidine. Referred to as the Orito reaction for the Japanese researcher who first described it more than 20 years ago, the catalytic reduction can yield (R)-ethyl lactate with an enantiomeric excess of more than 95% (C&EN, March 25, 2002, page 43).
"The cinchonidine system is a shining star," Sun comments. The reaction has been demonstrated at the kilogram scale, but it's not thoroughly understood. "Several research groups are trying to figure out why it works so well so that we can design even better catalyst systems--ones that work in a broader range of applications," he adds.
Just a few years ago, researchers were investigating the conditions under which cinchonidine-based hydrogenation reactions run most efficiently. More recently, the focus has shifted from studying kinetics to trying to elucidate detailed reaction mechanisms. Researchers are using experimental and computational methods to study bonding geometries, molecular configurations, and interactions between reactants, surfaces, and chiral modifiers.
ONE SCENARIO explaining the reaction's enantioselectivity is based on reduction of a hydrogen-bonded complex that forms between the chiral alkaloid and the pyruvate compound. In this model, which was proposed by the research group of Alfons Baiker, a chemistry professor at the Swiss Federal Institute of Technology (ETH), in Zurich, the adsorbed alkaloid and pyruvate are hydrogen bonded between the pyruvate's carbonyl group in the ketone function (not the ester) and the nitrogen atom in cinchonidine's quinuclidine group.
In polar solvents such as acetic acid, the hydrogen-bonded complex forms between the pyruvate and a protonated form of the alkaloid. In apolar solvents such as toluene, the complex forms between the alkaloid and a half-hydrogenated state of the pyruvate, according to the ETH model. The adduct attaches via π-bonding interactions to the platinum surface through cinchonidine's fused-ring quinoline group. Hydrogenation then follows enantioselectively from the platinum side of the surface-bound species.
Backing up the hydrogen-bonding basis of the model, Baiker and coworkers Norberto Bonalumi and Thomas Bürgi recently probed interactions between the chiral modifier and a cyclic ketoester, ketopantolactone, using a reflection infrared spectroscopy technique combined with a concentration-modulation procedure. By switching the flow of reactant and modifier solutions through a reaction cell periodically and employing phase-sensitive detection methods, the group was able to isolate periodically varying signals originating from the catalytic solid-liquid interface.
Based on characteristics of the measured spectral features and various control experiments, Baiker and coworkers conclude that a broad vibrational band located in the frequency range in which an N+...H...O hydrogen bond is expected to absorb is indeed the signature of hydrogen bonding between the ketoester and cinchonidine [J. Am. Chem. Soc., 125, 13342 (2003)]. The ETH group also showed that using a cinchonidine derivative in which the ketoester's access to the quinuclidine nitrogen atom is hindered by the presence of a methyl group prevents the hydrogen-bonded complex from forming.
At Laval University, in Quebec, chemistry professor Peter McBreen uses a related vibrational spectroscopy technique to analyze interactions between chiral modifiers and reactant molecules. By recording a series of IR spectra while exposing a platinum surface to controlled amounts of a modifier and reactant, McBreen and graduate students Marc-André Laliberté and Stéphane Lavoie monitor changes in bonding configurations and deduce reaction mechanism details.
In one study, the Laval group examined bonding interactions between methyl pyruvate and 1-naphthylethylamine, a molecule that is simpler than cinchonidine yet known to be an effective chiral modifier for Orito reactions. In the absence of the modifier and at low pyruvate concentrations, the pyruvate tends to adopt a so-called enediolate geometry, in which the carbonyl groups are in a cis configuration and the molecule bonds to the surface through both carbonyl oxygen atoms, the group reports.
They find that at higher pyruvate concentrations, the molecule assumes a trans conformation, which is stabilized thermally when the modifier is present. Based on spectral analysis, McBreen and coworkers conclude that the stabilizing interaction is hydrogen bonding between the ester carbonyl of trans-methyl pyruvate and the modifier's amine group [J. Am. Chem. Soc., 125, 15756 (2003)].
THE HYDROGEN BOND location proposed by McBreen's group--the ester side of the pyruvate--differs from the more commonly held view that an H-bond forms on the ketone side of the molecule. But ketone carbonyls form stronger bonds to the surface than do ester carbonyls, McBreen explains, which leaves the ester side available to dock with the chiral modifier.
"It's like tying a bow and holding the partly tied ribbon in place with your finger and finishing the job with your other fingers," McBreen remarks. The amine-ester H-bond orients the molecule and helps hold it in place. Hydrogenation then occurs on the ketone side of the pyruvate.
According to the Laval team, stereospecificity in the reaction is controlled by rotation about the C–C bond between the ester and ketone sides of the surface-bound pyruvate. The team proposes that the molecule adopts a configuration that brings the ketone carbonyl near enough to the modifier to form another weak hydrogen bond--this one between the ketone carbonyl and the modifier's naphthalene ring. That conformation avoids the steric repulsion that would develop between the ketone's methyl group and the naphthalene ring if the C–C bond were rotated the other way.
"This energetic inequivalence provides the driving force for enantioselection," McBreen asserts. Flip the C–C bond one way, and hydrogenation yields one of the enantiomers. Flip it the other way, and the opposite face of the molecule is accessible, which leads to the other enantiomer.
Although the ETH and Laval investigations are similar, there are a number of experimental differences. The groups used different reactants and chiral modifiers, for example. And the ETH team used solution-phase reagents on alumina-coated platinum in a flow cell while the Laval group worked in high vacuum with gases and single-crystal platinum. Nonetheless, the groups report evidence for much the same surface-bonding phenomena. Specifically, both groups present experimental evidence for a hydrogen-bonded modifier-reactant complex, and both groups claim that the ketoester changes the modifier's configuration.
An alternative approach to studying interactions between chiral modifiers and reactants is computational investigation. For Merck's Sun, the alternative approach led to an alternative reaction model.
"Based on the high enantiomeric excess obtained with cinchonidine, you would expect that the modifier and the substrate [ethyl pyruvate] interact quite strongly," Sun remarks. "The substrate has to feel the chiral environment imposed on it by cinchonidine in order for the chiral information to be transferred to the product."
A BIT OF CHEMICAL intuition concerning the strength of the modifier-reactant interaction coupled with experience working with reductive amination reactions led Sun to focus on the role played by the tertiary amine in cinchonidine's quinuclidine group. Unlike the models that invoke H-bonding between cinchonidine and the pyruvate, Sun wondered whether--as in the case of reductive amination reactions--the modifier-reactant interaction could be based on nucleophilic addition.
So Sun teamed up with Kendall N. Houk, a chemistry professor at the University of California, Los Angeles, who had investigated other types of asymmetric catalysis, to examine the cinchonidine system. Using a variety of computational techniques, Houk, Sun, and UCLA graduate student Grigoriy Vayner determined that the nucleophilic addition scenario is, indeed, a theoretically sound reaction mechanism model.
In contrast with other reaction proposals that are based on weak hydrogen bonding, the Merck-UCLA team proposes that the key interaction between cinchonidine and the pyruvate is a strong covalent bond joining cinchonidine's amine to the pyruvate ketone carbonyl group. The researchers explain that the nucleophilic addition step results in a surface-bound zwitterionic adduct between cinchonidine and the pyruvate. The charged species is likely to be stabilized in acidic media via H-bonding, they say. In the next step, the newly formed C–N bond undergoes hydrogenolysis with overall symmetry inversion [J. Am. Chem. Soc., 126, 199 (2004)].
Sun points out that shortly after beginning research on the nucleophilic addition model, he discovered that Robert L. Augustine, chemistry professor emeritus at Seton Hall University, South Orange, N.J., had already proposed a cinchonidine reaction mechanism based on a nucleophilic interaction between the tertiary amine and the pyruvate, but had not investigated the model further.
One of the observations of the Merck-UCLA group is that the reactant needs to be an "activated" ketone for the reaction to proceed efficiently. In the case of pyruvates, the presence of an additional oxygenated substituent, the ester group, is credited with enhancing the molecule's reactivity and stabilizing the adduct.
The deciding factors in enantioselection, according to the Merck-UCLA researchers, are differences between the stabilities and abundances of diastereomeric forms of the surface-bound zwitterionic adduct. Using certain simplifying approximations to facilitate the calculations, the team members found that there is roughly a 4-kcal-per-mole energy difference between diastereomers. Then they calculated a product ratio (enantiomeric excess, ee) based on the energy difference that they say agrees reasonably well with the mid-90 ee values obtained experimentally.
"We're doing very accurate calculations on model systems but only approximate calculations on the real thing," Houk acknowledges. "Still, the calculations we're able to do provide strong support for the idea that enantioselectivity is determined by the relative energies of platinum-bound zwitterionic adducts," he adds.
ALTHOUGH A LARGE fraction of the studies in chiral heterogeneous catalysis involve the surfaces of metals, metals aren't the only game in town. Porous materials such as zeolites and mesoporous solids (2 to 50-Å pore diameters) play a major role as catalyst supports in conventional heterogeneous catalysis. So some scientists have tried to determine whether or not the holey materials can play a special role in chiral catalysis. Just recently, a few research groups have shown that, indeed, they can.
Sir John Meurig Thomas recalls that his interest in using mesoporous silica and other porous materials in asymmetric catalysis grew out of a research project on "heterogenizing" (immobilizing) homogeneous catalysts. Thomas, who is a professor of chemistry and former director at the Royal Institution of Great Britain, in London, and a professor of materials science at the University of Cambridge, coauthored a paper a decade ago in Nature that reported a procedure for anchoring achiral homogeneous titanium epoxidation catalysts to the interior walls of porous supports. His coauthor on that paper was Thomas Maschmeyer, who is now a chemistry professor at the University of Sydney, in Australia.
"It occurred to us that if we could do the same with an organometallic chiral catalyst--meaning a catalyst in which chiral ligands are attached to a metal atom--then we could capitalize on the confinement in the pores to do asymmetric catalysis," Thomas notes. Not only would anchoring the compounds in a confined space boost enantioselectivity, but it would also enable researchers to recover the costly catalysts, in which the chiral ligands may be even more expensive than the precious metals.
The idea is that the confining dimensions of the interior of a pore should be able to restrict the possible orientations that a bulky reactant can assume as it approaches a chiral catalytic center that's attached to the pore wall. If a reactant is nudged by its surroundings into the right orientation for a stereospecific reaction, then the reaction should proceed enantioselectively.
It took a few years to prove the hypothesis, but now the effect has been demonstrated. In a recent study, Thomas, Robert Raja, Matthew D. Jones, Dewi W. Lewis, Brian F. G. Johnson, and their coworkers used a brominated trichlorosilane compound to anchor rhodium(I) diene-diamine and palladium(II) allyl-diamine catalysts to the inner walls of porous silica. The group then compared the performance of the surface-bound versions of the catalysts with that of the homogeneous forms. In separate test reactions involving asymmetric hydrogenation of phenylcinnamic acid and methyl benzoylformate, the group found that the tethered catalysts provide significantly greater enantiomeric excesses than their homogeneous counterparts [Angew. Chem. Int. Ed., 42, 4326 (2003)].
The group also found that attaching the chiral catalysts to nonporous silica reduced the enantioselectivity as compared with the same compounds attached inside the silica pores. "It shows that the magic of it all is confinement. The restricted approach of the reactant to the catalyst is the root cause of the enhancement in enantioselectivity," Thomas asserts.
Going a step further to assess the importance of confinement, Thomas, Raja, David E. W. Vaughan, and coworkers compared the performance of catalysts that were fixed to silica samples with pore sizes ranging from 38 to 250 Å. They found that the catalysts' ability to steer reactions enantiospecifically diminished systematically with increasing pore size [J. Am. Chem. Soc., 125, 14982 (2003)]. The larger the pores, the less effective the support is in constraining the reactants' orientation.
ONE SHORTCOMING of the anchoring technique is the large effort required to prepare the catalysts. So rather than anchoring the metal compounds to silica covalently by way of the brominated trichlorosilane tether, the team has adapted a noncovalent immobilization method that uses surface-bound triflate (CF3SO3-) counterions to secure the cationic catalyst compounds in place.
Raja, who is a senior research associate in the chemistry department at Cambridge, points out that the noncovalent method is quick, convenient, and industrially relevant. Catalysts prepared via the newer method using nonordered mesoporous silica are robust and effective but roughly one-eighth the cost of earlier versions of the covalently tethered catalysts, he says. Underscoring chemical manufacturers' interest in the chiral catalysis developments, Raja notes that Bayer, which provided financial support for the research, has filed several international patents to protect the work.
At Cardiff University, in Wales, the research group of chemistry professor Graham J. Hutchings also has immobilized homogeneous chiral catalysts on the interior walls of pores and demonstrated that confining the catalysts improves enantioselectivity in aziridination reactions. Aziridines are three-membered heterocycles that contain nitrogen. Hutchings points out that chiral aziridines are important intermediates to drug precursors because they contain an asymmetric carbon center bonded to nitrogen.
To carry out aziridination of alkenes, the Cardiff group used electrostatic immobilization techniques to modify zeolite HY with copper-bis(oxazoline) complexes pioneered by Harvard University chemistry professor David A. Evans. Then they compared the solution-phase and surface-bound forms of the catalysts and found that the heterogeneous version provides greater enantioselectivity than its homogeneous analog.
Further investigation of the Cu-bis(oxazoline) catalysts led to a curious result: The Cardiff chemists, including graduate students John Gullick, Sophia Taylor, and others, and their coworkers at other institutions, were surprised to find that enantioselectivity in the aziridination reaction increases with conversion of the alkene. The result was observed in homogeneous and heterogeneous reactions. The group determined that the unusual behavior is caused by the reaction product--aziridines--and sulfonamide by-products reacting with the catalyst's active center [Chem. Commun., 2003, 2808].
"The work shows that the catalyst's active site is much more complex than was originally envisaged," Hutchings remarks. "More important, we see that the aziridine product is not necessarily stable under reaction conditions and can be reacted to form the opposite enantiomer," he adds.
In related work, Hutchings and coworkers Neil A. Caplan, Frederick E. Hancock, and Philip C. Bulman Page recently demonstrated that catalysts consisting of CuH zeolite Y modified with bis(oxazoline) are very effective in facilitating asymmetric reactions of carbonyl and imino-ene compounds. The group points out that the catalysts can be recovered readily and reused without degrading the catalytic performance. The work will be published in a forthcoming issue of Angewandte Chemie International Edition.
Crossing over to the other commonly studied chiral surface-catalysis system, Hutchings' group studied the cinchonidine-based Orito reaction--but did so with a twist. The group prepared supported platinum catalysts that were premodified with cinchonidine and then carried out enantioselective hydrogenations of methyl pyruvate using gas-phase reactants. Typically, the reaction occurs at the solid-liquid interface. But the investigation by Hutchings and coworkers shows that enantioselective reactions can be promoted at the gas-solid interface as well [Chem. Commun., 2003, 1926]. Doing away with the solvent simplifies the reaction and is likely to open the reaction to new mechanistic probes.
Enantioselective heterogeneous catalysis has not yet made a big splash in industrial chemistry. Researchers are fascinated by the molecular subtleties that drive asymmetric conversions. And chemical manufacturers are keen to exploit the potential benefits offered by the catalytic systems. As mechanistic studies continue to reveal additional details of the systems' inner workings, asymmetric surface chemistry moves toward large-scale application.