Catalyst for change

With growing environmental pressure to cut hazardous waste and solvent use, searching for greener and more efficient catalysts is a major quest. Dr Sarah Houlton reviews some of the recent successes

With growing environmental pressure to cut hazardous waste and solvent use, searching for greener and more efficient catalysts is a major quest. Dr Sarah Houlton reviews some of the recent successes

Catalysts have become an essential tool in the manufacture of APIs and intermediates. Whether metal-based or enzymes, or even the new breed of organic catalysts, they will not only enable reactions to take place that would not otherwise be possible, but a good catalyst will also make the reaction faster, cheaper, more selective and even cleaner. Much effort is put in by both academia and industry into making catalysts that are more effective and more cost-effective than existing ones.

Metal catalysts are typically based on palladium, rhodium or ruthenium, but this raises the issue of the potential for metal residues to remain in the product. Residues levels are strictly controlled by the regulatory authorities but if the expensive, toxic metals were to be avoided in the first place, then it would not be an issue. As a result, various groups have been looking at carrying out important reactions such as cross-couplings and hydrogenations with other metals, notably iron. This does not have the toxicity issues of some of the other transition metals used in catalysts, and is even actively permitted by the FDA in coating ingredients designed to give tablets a pearlescent finish.

Numerous different groups are working on reactions that use iron as a catalyst. For example, Robert Morris at the University of Toronto in Canada has developed an iron-based catalyst system that can be used for the asymmetric hydrogenation of polar bonds (Scheme 1).1 The catalysts used in these reactions are normally based on rhodium or ruthenium. The group looked at both acetophenone transfer hydrogenation and hydrogenation with hydrogen gas, using iron(II) catalysts. These catalysts have a tetradentate diiminodiphosphine ligand, and are used as a tetrafluoroborate complex.

The complex was tested as a catalyst for the hydrogenation of acetophenone with hydrogen, and under 25 atm of the gas at 50°C in the presence of potassium t-butoxide; a 40% conversion to 1-phenylethanol was observed, in 27% ee. Although these figures are modest, this is claimed to be the first time a well-defined iron catalyst has been found for the asymmetric hydrogenation of ketones. However, it did not work at all for transfer hydrogenation catalysis.

Two further catalysts were prepared - one a carbonyl complex made by treating the original catalyst with carbon monoxide, and the other an isocyanide complex, made with t-butyl isocyanate. While these failed to catalyse hydrogen hydrogenation, they worked fairly well in the solvent transfer hydrogenation of ketones, aldehydes and imines. Indeed, the group reported that turnover frequency of the carbonyl complex was, at 907 an hour, reasonably competitive with the results seen with the best ruthenium complexes.

Matthias Beller's group at the University of Rostock in Germany has been looking at using iron catalysts in epoxidation reactions.2 Epoxides are important building blocks in the synthesis of pharma intermediates and APIs, as they can be used to introduce a range of different functionalities. They had previously found that iron(III) chloride hexahydrate, in combination with pyridine-2,6 carboxylic acid (H2pydic) and pyrrolidine, can be used to epoxidise both mono and di-substituted aromatic olefins, with hydrogen peroxide as the oxidant. However, when the reaction was attempted with aliphatic olefins and highly substituted aromatic olefins, the process lost both reactivity and sensitivity.

The key to improving the reaction lay in finding an amine that was more effective than pyrrolidine as a co-ligand. The group found that benzylamine derivatives worked much better; however, H2pydic appeared to be essential for success as replacing it with alternatives such as picolinic acid or quinaldic acid killed all reactivity. Other iron sources were similarly unsuccessful. Using the new benzylamine-based system, mono, di and tri-substituted aromatic olefins were all effectively epoxidised, as were internal di-and tri-substituted and functionalised aliphatic olefins. Indeed, inactive aliphatic olefins can be oxidised in up to 96% yield, and they are now working on finding better protocols for tetrasubstituted aromatic and mono-substituted terminal aliphatic olefins.

Meanwhile, scientists at the Indian Institute of Technology Madras in India have developed a copper catalysed oxidation of aldehydes to carboxylic acids that is performed under mild conditions in an aqueous system.3 These reactions are commonly carried out using chromium or manganese-based oxidants, which are not ideal from an environmental standpoint, and produce substantial quantities of undesirable waste. Other reagents have been introduced over the years, but these tend to require forcing conditions, such as a strongly acidic reaction mixture, high temperatures, or large amounts of expensive or hazardous oxidising agents.

They had already established that copper(I) chloride could be used to catalyse the oxidation of benzylic and allylic primary alcohols to the corresponding acids, using anhydrous t-butyl peroxide in decane, with the reaction carried out in acetonitrile at room temperature. They have now found that aldehydes can be oxidised efficiently by aqueous t-butyl peroxide, with the only catalyst required in the reaction being copper(I) chloride, without the need for ligands or additives. Running the reaction without the copper salt gave no reaction, showing that it is indeed acting as the catalyst. The reaction again takes place in acetonitrile at room temperature, and numerous aromatic, vinylic and aliphatic acids were successfully synthesised in this way.

Bruce Lipshutz and his group at the University of California, Santa Barbara in the US have developed a novel supported catalyst based on copper and nickel that can be used to catalyse a number of cross-coupling reactions that would normally need group 10 and 11 metals such as palladium.4 Cross-coupling processes, like the Heck, Suzuki and Sonogashira reactions, are extremely important in the assembly of pharmaceutically important molecules. Because the catalyst is supported, it makes the work up simpler, and the catalyst can be recycled, which minimises the quantity of metal waste that is produced.

The new catalyst is charcoal-based, and they believe it is the first multi-metal on carbon catalyst. Copper and nickel are loaded onto the charcoal support in a 1:4 ratio. This is done by combining 2% copper(II) nitrate, 8% nickel(II) nitrate, activated charcoal and water in one pot. The mixture is subjected to a brief ultra-sonication to distribute and impregnate the salts into the charcoal. The catalyst is then isolated by distilling off the water and drying under vacuum.

The catalyst has several applications, including reactions that typically require two different metals to be present for them to proceed, such as Sonogashira and Stille couplings, or those that require two different metals sequentially. For example, it can be used to facilitate Suzuki-Miyaura couplings of aryl bromides and chlorides with boronic acids, and anilines and diarylamines can be formed from the corresponding halides and primary or secondary alkyl or arylamines by ligating the supported catalyst with diphenylphosphino-ferrocene. Other reactions that it successfully promotes include asymmetric conjugate reductions of activated olefins, asymmetric conjugated reductions of activated olefins, and azide-alkyne "click" type reactions.

Water as solvent

Another way of making a process greener is to remove the solvent entirely, and just use water instead. Lipshutz's group has been looking at the possibility of using water as a solvent in a variety of different cross-coupling reactions, with some success. The group has developed a version of the Heck reaction which uses micelles to facilitate the reaction in water.5 This is the first example of a Heck coupling that has been carried out in water as the only solvent, and at room temperature - previous examples had required considerable heating.

They surveyed a number of commercially available amphiphiles to make the micelles, and found that the vitamin E-based PTS (PEG600-yl a-tocopherol sebacate) worked well. It forms nanometre sized particles in water, which can accommodate the lipophilic reactions starting materials. They started with a pre-formed solution of 15% PTS in water, and added 2 mol% of the Johnson Matthey ferrocene-based palladium catalyst (dtbpf)PdCl2 and three equivalents of triethylamine, plus one equivalent of the aryl iodide coupling partner and two of the acrylate. Heck couplings generally took place at room temperature, with reaction times between one and 24 hours (Scheme 2).

The product was isolated either by a standard extraction or filtering through silica gel. Aryl bromides also work, but they are slower; a search for an additive to speed up the reaction was not very successful, but simply warming the reaction to 50°C had the desired effect.

As an example, they coupled 4-chloro-2-nitrobromobenzene with ethyl acrylate, giving a cinnamate that is an intermediate on the route to a Pfizer Cox-2 inhibitor.

Suzuki reactions can also be carried out in a similar way.6 These diaryl couplings between an aryl halide and an aryl boronic acid are typically carried out in the presence of water, and typically organic solvents are also required. Again, 2 mol% of PTS in water worked well, along with the same palladium catalyst, and triethylamine was the most effective base tried - potassium phosphate worked but as it did not form an emulsion it made the reaction's progress difficult to monitor. Even sterically hindered aryl bromides worked, but they were unable to couple a diortho substituted bromide and a similarly substituted aryl boronic acid. And lipophilic aryl bromides also reacted successfully. Chlorides reacted less readily than iodides and bromides, but they did manage to couple 4-chlorobenzonitrile with the electron deficient 2,4-difluorophenyl-boronic acid; previously cross couplings of chlorides in water required co-solvents. The group is now investigating the possibility of coupling heteroaromatic biaryls.

This micelle approach has also been applied to Sonogashira couplings of aryl bromides.7 The couplings between aryl halides and alkynes are typically catalysed by both palladium and copper, and recently much effort has been put into finding ways that use just one or other metal. These reactions are also often carried out in a mixture of water and organic solvent, but using just water is difficult for solubility reasons. Again, the vitamin E-derived PTS was used to make micelles that solubilise the reaction components. Once again, triethylamine proved the best base. The reaction worked with a variety of different coupling partners, and the best catalyst in this case turned out to be PdCl2(CH3CN)2 along with the ligand C-Phos. The reactions worked well at ambient temperature.

Another important reaction that can be carried out in water is transfer hydrogenation. Hans Adolfsson's group at Stockholm University in Sweden has been looking at a new rhodium-based catalyst that can carry out chirally selective transfer hydrogenations of ketones in water.8 Again, this uses micelles to facilitate the reaction, this time made from sodium dodecylsulfonate (SDS).

Chiral transfer hydrogenation reactions are commonly mediated by ruthenium catalysts. While there are numerous examples of such reactions in water for substrates such as aryl alkyl ketones, alkyl alkyl ketones have proved more difficult. For substrates containing an aryl group, p-CH interactions between the aryl group and the catalyst are believed to stabilise the intermediate while the hydrogen transfer occurs, but obviously this interaction is absent in alkyl alkyl ketones, and thus the enantioselectivity of the reaction is much poorer.

Their idea was to replace the normal catalyst system with a lipophilic alternative. By introducing a monosulfonated diamine into the ligand, this should enable the catalyst to be highly soluble in the SDS micelles. They assumed that this would leave the polar end where the transition metal sits close to the micelle's hydrophilic surface, and the poor solubility of long-chained alkyl alkyl ketones in water will push the ketone into the micelle containing the catalyst, with the polar carbonyl end pointing towards the surface. They thought this might orient the ketone so that the catalyst can discriminate between its two enantiotopic faces, which might increase the stereochemical induction of the reaction.

To test this, they made a diamine ligand with a dodecyl group attached. This was used along with [Cp*RhCl2]2 as a metal precursor to form a catalyst for the asymmetric transfer hydrogenation of numerous alkyl methyl ketones, in the presence of SDS and sodium formate as the hydrogen source (Scheme 3). This mostly gave good conversion to the secondary alcohol, although more sterically demanding alkyl groups such as adamantyl and t-butyl did not work so well. Reactions were tried with both the surfactant present and absent, and modest increases in ee were seen when the surfactant was included in the reaction mixture. They also tried aryl alkyl ketones, with excellent results.

Another novel reaction in water has been developed by Mizuno's group at the University of Tokyo - the oxygenation of primary amines to amides.9 While amides are extremely important in pharmaceuticals, they are generally made from activated carboxylic acid derivatives or the rearrangement of ketoxamines, which produce large amounts of chemical by-products that are often toxic.

In theory, oxygenating the a-carbon of an amine would give a cleaner and much more atom efficient reaction, but there is no literature precedent for a catalytic form of the reaction - and the few that do exist require stoichiometric quantities of oxidant, such as ruthenium tetroxide, and need the amine group to be protected.

The new procedure from Mizuno uses a supported ruthenium hydroxide catalyst (the support is alumina), in water as the only solvent, and air is the oxidant. In theory, the only by-product is water so there are no toxic waste materials, and the catalyst is simple to remove when the reaction is complete by filtration, and can be reused.

The reaction is applicable to a range of different amines. Benzylamines are oxidised efficiently, regardless of whether they possess electron donating or electron withdrawing substituents. It works on nitriles, and also non-activated linear, branched and cyclic aliphatic amines, with yields generally above 90%. The catalyst is used at a level of 5 mol%, and for benzylamines the reaction is stirred as a slurry for 24 hours. Aliphatic amines required more forcing temperatures and increased pressure, which is achieved by carrying the reaction out in a sealed Teflon vessel placed inside an autoclave.

Reactions like this with excellent atom efficiency and which produce minimal by-products can only increase in importance and popularity in future, as the need for "greener" processes becomes more apparent.

Featured Companies

Johnson Matthey (more information, website)