CHEMTECH
February 1999
CHEMTECH 1999, 29(2), 23-29.
Copyright © 1999 by the American Chemical Society.



ENABLING SCIENCE

Lanthanide triflates as unique Lewis acids

These water-stable materials are effective catalysts for a wide range of organic transformations carried out in aqueous media.

Wenhua Xie
Yafei Jin
Peng George Wang

E nvironmental concerns about today's chemical research and industry are ever-increasing. The challenge for a sustainable environment calls for clean reaction processes that avoid the use of hazardous and harmful organic solvents (1). Water--the most abundant and renewable resource on the planet--is no doubt the most desirable solvent in this respect. After being ignored for decades, water has recently been recognized again as a solvent for organic reactions (2-4).

Lanthanide trifluoromethanesulfonates (triflates) are unique Lewis acids that are currently of great research interest. Unlike common Lewis acids that decompose readily in the presence of water, lanthanide triflates are stable in water and function well in aqueous media (5). They effect numerous reactions in aqueous solutions or simply in water: Diels-Alder reaction (6), aldol reaction (7-12), Michael addition (13), and allylation of carbonyl and imine compounds (14, 15). In this article, we focus on the recent work in our laboratory, describing synthetic applications of lanthanide triflate-catalyzed organic reactions in aqueous media.

Aza-Diels-Alder reactions in water
Aza-Diels-Alder (aza-DA) reactions hold enormous potential for alkaloid and heterocycle syntheses. Simple iminium salts, generated in situ under Mannich-like conditions, have been reported to react with dienes in water to give the corresponding aza-DA adducts (16). This reaction conveniently combines three reactive components (an aldehyde, an imine salt, and a diene) in aqueous solution to generate nitrogen-containing heterocyclic structures. However, the reaction has been limited to the use of either formaldehyde or activated aldehydes, such as glyoxylates, as the aldehyde component. Less reactive aldehydes, such as acetaldehyde, show much lower reactivity. Dienes other than cyclopentadiene also give poor yields.

We discovered that aza-DA reactions could be effectively catalyzed by lanthanide triflates in water (17). The catalytic process greatly reduces the limitations mentioned earlier. Aldehydes other than formaldehyde and glyoxylate work well to provide the DA adducts in good yields. The reactions of various dienes other than cyclopentadiene also give excellent yields (Figure 1). Moreover, the lanthanide triflates can be recovered easily and reused after the product is extracted with methylene dichloride. In fact, the lanthanide(III) catalysts described in this article are reusable in most cases.

Figure 1 thumbnail Figure 1.

Figure 2 thumbnail Figure 2.

We expect that the lanthanide-catalyzed aza-DA reactions will be used in organic synthesis. We have successfully applied this method to synthesizing aza-sugars and their analogues, which are potential inhibitors of glycoprocessing enzymes (18) (Figure 2). For example, by starting from a D-mannose derivative, the reaction of glyceraldehyde acetonide 1 (boldface numbers appear on Figure 2) with benzylamine hydrochloride and cyclopentadiene in water proceeded readily in the presence of lanthanide triflates to afford three adducts; 4 was the main product. The corresponding reaction of another chiral aldehyde 6, prepared via diazotization of D-glucosamine hydrochloride, stereoselectively yielded 7 as the major product. As is often observed, the radii of the lanthanide ions played an important role in the catalytic process. The medium-size lanthanides gave better results than smaller or larger ones. The diastereoselectivities can be rationalized by proposed lanthanide-coordinated intermediates. Products 4 and 7 were subsequently converted into aza-sugar analogues after several steps.

The use of aqueous media allows the use of unprotected carbohydrates or their derivatives as one of the components in the aza-DA reaction. Application of the lanthanide-promoted aqueous aza-DA reaction in carbohydrate chemistry promises to provide a novel entry to numerous carbohydrate-derived nitrogen-containing heterocycles.

Synthesizing pyridine and its derivatives
The pyridine family and its derivatives are important in biological functions; they also are versatile intermediates for the synthesis of natural products such as alkaloids. Considerable research has focused on synthesizing these compounds from acyclic starting materials.

In our recent exploration of lanthanides as stable Lewis acids in aqueous solutions, we found that lanthanide triflates effectively promote reactions of aldehydes with amine hydrochlorides to afford 2,3-dihydropyridiniums and pyridinium derivatives (19). The reactions proceed readily in water at room temperature. For example, hexanal reacts with benzylamine hydrochloride in aqueous solutions of lanthanide triflates to give 3,5-dibutyl-2-pentyl-N-benzyl-2,3-dihydropyridinium and 3,5-dibutyl-2-pentyl-N-benzylpyridinium in a combined yield of 82%. The counterions in the products could be OTf- or Cl-, depending on workup conditions. The dihydropyridinium then can be converted to the corresponding pyridinium by dehydrogenation in the presence of triethylamine in refluxing toluene. The latter can be debenzylated to produce the pyridine compound (Figure 3).

Figure 3 thumbnail Figure 3.

We found that other aldehydes and amine hydrochlorides react similarly. However, one of the two products could be absent, depending on the aldehyde used. For example, the reaction of propionaldehyde or butanal with benzylamine hydrochloride gave only 2,3-dihydropyridiniums; no corresponding pyridiniums were isolated. In contrast, the reactions of phenylacetaldehyde and 4-decenal gave only pyridinium products.

We believe that the reactions proceed through sequential condensations of three molecules of the Schiff base formed between the aldehyde and the amine. The pyridinium products result from oxidation of the corresponding 2,3-dihydropyridiniums (Figure 4). This method of Ln(OTf)3-catalyzed reactions in water is a novel extension of pyridine synthesis via the condensation of aldehydes and amines (the Chichibabin reaction). The Chichibabin reaction normally requires high temperatures, high pressures, or vapor-phase reaction conditions, whereas the Ln(OTf)3-catalyzed reactions proceed under mild conditions.

Figure 4 thumbnail Figure 4.

Reactions of indoles with aldehydes
Protic acids and Lewis acids promote reactions of indole with aromatic and aliphatic aldehydes and ketones. The reactions produce azafulvenium salts, which then react further with a second indole molecule to form bisindolyl-methanes (20-22). By using lanthanide triflates as catalysts, we recently realized this reaction in aqueous media (Figure 5) (23). Ethanol-water was the best solvent system for yield and product isolation. Although all the lanthanides showed good catalytic effects, dysprosium triflate gave the best result.

Figure 5 thumbnail Figure 5.

Aromatic aldehydes with either electron-donating or electron-withdrawing substituents also reacted well to give excellent yields. The reactions of ketones took longer but still produced moderate-to-excellent yields. In contrast, the corresponding reactions using trifluoroboron etherate or aluminum chloride generated several unexpected products (24, 25). In the case of the 3-substituted indoles, the most active site (C-3) was blocked, so that both ketones and aromatic aldehydes failed to react, whereas the corresponding reaction of hexanal afforded a [1+1] ethanolysis product other than the normal [2+1] adduct.

Imines, like the carbonyl compounds we have been working with, are electrophilic and can be effectively activated by Ln(OTf)3. As an extension of the above work, we recently examined the corresponding reaction of indole with a variety of imines. The reaction proceeded in a manner similar to that shown in Figure 5, producing secondary indolyl amines as well as bisindolyl-methanes as the byproducts. This is a convenient alternative for performing the conventional Mannich reaction of indole with primary amines, which usually gives a low yield because of various side reactions.

Aziridine synthesis
Aziridines are versatile building blocks for synthesizing various biologically important molecules (26, 27). Adding a carbene moiety to an imine is one of the simplest approaches to aziridine synthesis. Templeton and co-workers have shown that Lewis acids such as BF3·O(C2H5)2, AlCl3, and TiCl4 are efficient catalysts in the aziridination reaction of ethyl diazoacetate (EDA) with imines (28). The aziridination is effected by activating the imines followed by a nucleophilic addition of EDA. Just recently, by using lanthanide triflates, we accomplished this aziridination reaction in protic media that included the CH3CN-H2O system.

The reactions produced the corresponding aziridines with high cis selectivities. In fact, for most reactions examined, we isolated only cis aziridines. However, use of heavier lanthanides such as Er and Yb reduced the selectivity. No carbene-coupling product was detected under the experimental conditions. However, in several cases, the formation of the aziridines was accompanied by 11 (a, e, f) and 12 (a, d, e) with varied yields (Figure 6). We propose that the Ln(OTf)3-catalyzed aziridination proceeds like typical Lewis acids (Figure 7).

Figure 6 thumbnail Figure 6.

Figure 7 thumbnail Figure 7.

As we found in control experiments, the carbene insertion reaction of EDA and the ring-opening of the aziridine product can also occur to some extent in the presence of lanthanide triflates. However, these reactions do not actually interfere with the aziridination reactions because of their significantly lower reaction rates.

Polymer-supported lanthanide catalysts
Solid-supported catalysts offer many advantages in preparative procedures. Simplified workup, separation, and isolation, as well as catalyst recycling, not only lead to economical automation, but also reduce hazardous pollution to achieve environmentally friendly processes.

We recently prepared a series of polymer-supported lanthanide(III) catalysts via ion-exchange processes by using various commercially available resins (29). When using H+-form resins directly, in most cases >90% of the protons on the resins could be exchanged with the lanthanide(III) ions. The experimental results of this process with ytterbium are shown in Table 1. Other lanthanides also can be exchanged onto the resins. This process was performed conveniently in aqueous solutions of lanthanides at room temperature (Figure 8).

TO SIDEBAR: Table 1.

Figure 8 thumbnail Figure 8.

We have tested the supported lanthanides as Lewis acid catalysts in aqueous or alcoholic media in various reactions, including aldol reaction, acetalization, allylation, glycosylation, and epoxide ring-opening. Resins with high degrees of linking and large surface areas, such as Amberlyst XN-1010 and Amberlyst 15, are more suitable supports than the others tested. In most cases, the polymer-supported lanthanides demonstrated good-to-excellent catalytic effects compared with soluble triflate salts. However, in a few cases, such as the aza-DA reaction, they were less effective catalysts. Nevertheless, the polymer-supported lanthanides offer many advantages. For example, lanthanide triflates and polymer-supported lanthanides promote the glycosylation of unprotected glucosyl fluoride in methanol. Whereas it is troublesome to remove and recycle the soluble lanthanides in the former case, the separation and recycling of the supported lanthanides require only simple filtration.

Kobayashi and Nagayama recently developed a new kind of polymer-supported lanthanide catalyst by immobilizing Sc(OTf)3 onto polystyrene microcapsules (30). The microencapsulated Sc(OTf)3 exhibited an activity even higher than the monomeric form. Combining the advantages of a solid phase and the unique features of lanthanides as water-stable Lewis acids, polymer-supported lanthanide(III) compounds are promising for both research laboratories and industrial processes.

Outlook
The field of lanthanide-catalyzed organic reactions in aqueous solvents has evolved considerably over the past few years. More opportunities are waiting to be explored, and we expect to see an increasing number of organic transformations in aqueous media with the use of this technology. Moreover, because most enzymatic transformations also proceed in aqueous solutions, lanthanide catalysis has the potential to accomplish tandem combinations of chemical processes and enzymatic transformations. Such combinations are highly desirable for the development of environmentally benign processes.

Acknowledgments
We thank the National Science Foundation, the U.S. Environmental Protection Agency, and Hercules Inc. for financial support.

References

ACS Pubs ChemPort ChemCenter