Precious metals

The importance of chirality in the pharmaceutical industry has never been greater, with at least 80% of all new drugs in development being chiral. Dr Sarah Houlton considers the role of transition metal catalysts in chiral production

At the heart of many chemical syntheses of chiral molecules is a transition metal chiral catalyst. Frequently built around transition metals such as rhodium, palladium, iridium or titanium, the chirality of the ligands co-ordinated to the metal is used to induce chirality in the final product, and the chirality of the small amount of catalyst is amplified into a much larger volume of chiral product.The importance of metal catalysed chiral synthesis was recognised last year, with the award of the Nobel Prize for Chemistry to three of the big names in the area. Barry Sharpless's chiral epoxidations and dihydroxylations have become staples of chiral synthesis. Ryoji Noyori created the BINAP chiral diphosphine ligands for asymmetric hydrogenation that proved the inspiration for so many of the ligands to follow, and have been used in the production of numerous drug products, including the antibiotic levofloxacin and the antiinflammatory S-naproxen. And William Knowles was responsible for the first industrial scale catalytic chiral synthesis, the Monsanto synthesis of L-dopa.

The chirality of the catalyst rests on the ligands that are co-ordinated to the metal. Noyori's BINAP ligand was the inspiration for the range of modern chiral phosphine ligands. The number of different chiral ligands that are available is truly staggering, so choosing the best combination of metal and ligand is no trivial matter. But the number of different chiral phosphine ligands that are available means there are plenty of different options to find the best combination of metal and ligand for different hydrogenation substrates.

common processes

The most common processes that are carried out industrially by chiral catalysis are the synthesis of amino acids, chiral amines, and ketone reduction, usually by asymmetric hydrogenation, and carbon;carbon bond forming reactions such as Heck and Suzuki couplings. Probably the most commercially important chiral reaction is asymmetric hydrogenation. The reaction can be difficult — particularly because it generally requires elevated temperatures and pressures — but it can be an extremely powerful tool. A survey carried out by Synetix established that access to 10 ligand systems is sufficient to carry out at least 95% of all hydrogenations.

time and cost factors

Asymmetric hydrogenation is not used as widely as it might be, explained DSM Pharmaceuticals' David Ager at the FAST symposium, organised by Synetix and held in Cambridge, UK, in September 2002. He says several factors are at play here. Time-to-market constraints lead to very short development times being needed, and competing technologies may provide a quicker answer in the first instance.

Several cost factors are involved: the cost of the metal and ligand, the activity and stability of the catalyst, and the ability to recover or recycle it. Catalysts may not be readily available, perhaps because they are patent protected, and licensing them can be expensive. And there is the whole issue of reliability, whether real or merely perceived.

screening programme

One of the major problems, says Ager, is determining which catalyst or ligand system to use in a timely manner. DSM's answer is a catalyst screening programme that is able to provide rapid answers, using a high throughput methodology, which can also be used to optimise the process once suitable catalysts have been pinpointed.

A particularly useful ligand used by DSM is MonoPhos. Ager explained that this ligand, which, unusually, is monodentate rather than bidentate, has several advantages. Libraries of complex bidentate phosphine ligands are not easy to prepare.

However, MonoPhos ligands are phosphoramidites, and are easy to make in two steps. Diversity is available from both the diol and the amine part, and chirality can be incorporated in either the skeleton or the amine. It had previously proved very successful in the copper catalysed asymmetric 1,4 addition of diethyl zinc to cyclic enones, in a reaction developed by Ben Feringa at the University of Groningen, but had not been applied to asymmetric hydrogenation. MonoPhos is very easy to make — the raw material BINOL is cheap and readily available in both enantiomers.

The dimethyl version is used for most asymmetric hydrogenations, though there are numerous analogues in DSM's screen. It is generally not necessary to pre-form the catalyst either — merely add Rh(COD)2 and the ligand to the pot.

Initial work in Groningen identified dichloromethane as the best solvent for the reaction, but this is far from ideal environmentally, and work at DSM's plant in Greenville established that ethyl acetate was as good, and reactions can be performed at room temperature.

An alternative hydrogenation strategy is transfer hydrogenation. Rather than using hydrogen gas as the hydrogen donor, an alternative such as an alcohol or an acid is used. Such a process has been developed by Avecia, as shown in Scheme 1, and described at the FAST conference by the company's John Blacker.

To minimise back-reaction in the isopropanol system, the hydrogen donor is also used as the solvent, and the co-product, acetone, removed by distillation. This increases productivity for the reaction, and stabilises the optical purity of the product. The TEAF system is more difficult to run, but is more generally applicable. The catalyst in the reaction, Avecia's CATHy catalysts, use rhodium or iridium, analogous to Noyori's ruthenium catalysts that have been used in analogous reactions.

The reaction has very widespread application, Blacker said, and works with halides, thio compounds, alkenes, amides, acids, esters and nitriles. It has been scaled up to 200l, giving a yield of 95% and 97% ee. The concentration of the reaction mixture, around 5% w/v, is perfectly acceptable to industry, added Blacker.

asymmetric catalysis

An analogous imine version, which transforms an imine into a chiral amine, can be run at up to 30% w/v in an irreversible process. The metal and ligands can be mixed and matched to generate a library of catalysts for substrate screening. The CATHy catalysts are both air and water stable, and both the catalysts and screening kits are available through Strem chemicals.

Asymmetric catalysis is also important in a number of processes other than hydrogenation. For example, Degussa's DeguPhos has recently been applied to the reductive amination of a-keto acids (Scheme 2). Amino acids can be made by asymmetric hydrolysis, though the standard catalyst for such reactions, DuPhos, isn't perfect, said the company's Ian Grayson. So Degussa developed a new catalyst for these reactions, called MalPhos, which works well for the synthesis of b-amino acids, where it is better and faster on the troublesome Z isomer.

Another advantage is that the raw materials for making the catalyst itself are cheaper, so the catalyst is likely to be less expensive to produce.

highly reactive

Avecia has developed a range of cheap, highly active titanium or vanadium based salen catalysts called CACHy catalysts. They are based on Jacobsen's salen technology, but are much more reactive. In the example given in Scheme 3, it is used at 0.1mol% in the cyanation of aryl aldehydes. The catalyst shows similar reactivity with alkyl aldehydes and ketones, and is applicable to the synthesis of the commercially important mandelic acid derivatives.

'It is never easy to cost a process until you run it at large scale,' said Blacker at the conference, 'but I believe this has some real potential.'

The outlay on expensive catalyst can reduced if it can be recycled. One way that this can be facilitated is by immobilisation — attaching the catalyst to some form of solid support. This means it can then simply be filtered off after the reaction, and recycled into another reaction.

Avecia has developed an immobilisation method, EnCat, which can do just this. For example, in reactions such as palladium catalysed carbonylations, our Suzuki or Heck couplings, the expensive catalyst can be filtered off, leading to purer products with less palladium residues left behind, and the recycled catalyst does not lose its activity. 'It should be possible to use it in a continuous process,' said Blacker. The process is also applicable to other metals and chiral catalysts. 'We don't know yet whether it will work in all cases,' said Blacker, 'but it certainly has potential.'

already patented

One problem with developing a new ligand is that so many of the backbones are already patented, and modifications of these can be far from straightforward. Ciba Specialty Chemicals' Ulrich Berens explained the requirements for a new backbone at the conference.

The ideal ligand backbone should be readily available, without the need for special equipment. It should be possible to tune electron density at donor atoms and the bite angle, as well as offer simple variation of donor atoms. And there should also be the option for immobilisation by involving covalent bonding.

Berens and his team looked at

benzothiophene as a potential new backbones, as it is readily functionalised by electrophiles in the three position, it is easily metallated at the two position, and it is simple to attach anchoring groups at the five position as a handle for immobilisation. Unlike many existing ligand backbones, the resulting ligands, named ButiPhane, are not C2 symmetric.

As Ciba has backed away from the pharmaceutical ingredient and intermediate sector, it has sold the patent for the finished ligands to Solvias. As Berens said: 'The benzothiophene backbone offers a unique potential to reach a new level of modularity in ligand synthesis. It offers facile changing of electronic properties and bite angles, which is important for tuning the

chiral reactivity.'

rich source

Academia is a rich source of new chiral ligands, some of which were described at the FAST conference. These include the P-Phos ligands that emerged from the research group of Albert Chan at Hong Kong Polytechnic University. These have the big advantage over BINAP that it merely requires hydrogen to be bubbled through the reaction mixture — it is not sensitive to oxygen like BINAP. It gives high ees, and very high turnover numbers.

And Steve Nolan and his group at the University of New Orleans in the US are working on carbene ligands as phosphine analogues, which are stabilised as metal complexes. As Franco Sannicol² of the University of Milan pointed out to the

conference: 'There is no universal catalyst ligand or catalyst.'

This means it is always necessary to fine-tune the catalyst and ligands used within any process. High throughput methods are extremely useful here, and it is possible to screen a whole range of metals and ligands to find the best one for any particular application.