Immobilised drugs

Denis Geffroy and Fred Hancock discuss possible immobilisation technologies used for catalysts in the production of pharma intermediates, or for the development of heterogeneous catalysts for asymmetric chiral routes to drug production

The world of fine chemical manufacturing has been driven in the last five years by the value of synthesising asymmetric molecules for active pharmaceutical ingredients (APIs) and the need to have synthetic asymmetric manufacturing technologies to produce them.

This article concentrates on the need to have access to asymmetric manufacturing technologies and looks at this in terms of the exceptional success of heterogeneous asymmetric catalysis in providing scaleable, economical and clean processes of asymmetric molecules. Few homogeneous processes reach such high substrate/catalyst ratios and in these cases even if the economics start to look attractive the catalyst now becomes a serious impurity in the product requiring extensive post reaction purification.

The capability to immobilise these catalysts to a specifically designed, inert, porous support to allow slurry and fixed bed operation, is not trivial and requires an excellent understanding of the catalytic reaction mechanism. In the course of this work it was found that immobilisation often leads to intrinsically better catalysts, and that there is an important synergy between homogeneous and heterogeneous catalysts that can be exploited using a two stage process development:

• Stage 1: Homogeneous ligands are easy to screen and optimise. Hence the optimum process can be found quickly and the customer can potentially produce multiple kilo quantities of a desired intermediate

• Stage 2: The best homogeneous process can be taken and the catalyst immobilised for large scale manufacturing

To have the necessary ligands for Stage 1, Synetix Chiral Technologies with the help of its advisory board has been gaining access to, or designing for itself, suitable asymmetric ligands. It has a licence to use, and have its customers use, CTH-PhanePhos from Merck & Co, CTH-P-Phos, CTH-BINAM (both from Professor Chan), and a ferrocenyl bis-phosphine with planar chirality from Aventis Pharmaceuticals, called CTH-JAFAPhos. A licence from Yale allows Synetix to use bulky phosphines from Professor John Hartwig – in particular the CTC-Q-Phos (pentaphenylferrocenyl phosphine). There is a theme of phosphine chemistry that runs through the technology and allows Synetix to develop generic phosphine immobilisation strategies. The ligands are shown.

The second stage – the immobilisation of homogeneous catalysts to design standard chemical manufacturing operations via slurry and fixed bed catalytic processes is discussed here.

There are two alternate routes to immobilise a homogeneous catalyst: link the metal to the surface or link the ligand. Studies in the mid 1990s linking the metal with Graham Hutchings, Philip Page and Don Bethell at the Leverhulme Centre in Liverpool, proved that the homogeneous activity and enantioselectivity in the solid catalyst could be kept and could often see positive chemical benefits from the linked catalyst. The methodology is summarised in Figure 1.

This catalyst was successfully used for asymmetric aziridination and this has subsequently been extended to Diels-Alder, HDA and ene carbonyl. One major chemical advantage of the immobilised system for aziridination was that it altered the properties of the copper and eliminated the decomposition of the nitrene donor that had forced other users of the chemistry to use large excesses of the alkene to minimise this side reaction. The bis-oxazoline can be modelled to fit within through two supercages.

This immobilisation technique is useful because it allows the solid catalyst to be prepared in a very simple and direct way. The optimum homogeneous catalysts could be found quickly by screening and their subsequent immobilisation requires only the optimisation of the support structure. This usually involves either using a known zeolite structure or synthesising an unstructured aluminosilicate with the required average pore size diameter. Then the final process is to find the best Si/Al ratio.

The limitation of the technique is that it requires a cationic catalyst throughout the catalytic cycle, something that for many catalytic reactions is known not to be true.

Unfortunately some of these non-cationic reactions are exceptionally important, e.g. ruthenium hydrogenations and palladium catalysed C-C bonding forming reactions. For these reactions a different approach is needed to an immobilised catalyst.

Because Synetix has technical and manufacturing skills in two other areas of inorganic synthesis – the manufacture of silica supports and the use of Ti(OiPr)₄, this allowed it to take advantage of the tremendous efforts in coatings or electronic materials technology and use sol-gel methods to make inorganic-organic hybrids of silica.

The knowledge gained in the synthesis of these materials has been applied to the design and manufacture of immobilised homogeneous catalysts.

The general route is described in Figure 2, and the porous solid can be schematically represented, Figure 3.

The availability of a wide variety of silanes makes it possible to devise essentially solid phase organic syntheses to arrive at the optimum chemical surface for the immobilisation of an asymmetric ligand. When this diversity of technical possibilities is combined with the commercial cost of these silanes then a new avenue of solid phase synthesis is opened.

The salient points with regard to these materials are:

• The porous structure of the silica can be designed for the ligand/reaction requirements.

• The organic moieties attached to the surface are available for general synthetic organic chemistry and thus the most suitable linking chemistry can be devised.

• The structure is rigid and strong and hence avoids all the issues of solvent compatibility that so detrimentally affect the organic polymer supports that have been used.

• High loading of ligand can be obtained – up to 30%w/w (>3 mmolg^-1).

• The materials can be formed by granulation or extrudation into shapes for fixed bed operation.

All these capabilities came together in a commercial agreement with Rhodia ChiRex, where the Jacobsen cobalt-salen ligand has been immobilised to give the dimer orientation, which is the active catalyst site for the Hydrolytic Kinetic Resolution (HKR).

The HKR reaction can be understood as the activation of the epoxide by the Lewis acid cobalt and activation of the water molecule by the same cobalt acting as Lewis base (see Figure 4)

The immobilisation of the salen ligand can be achieved to create the dimer orientation and this proves to be the most active catalyst with a significant increase in catalyst turnover number (mole product/mole of catalyst).

Thus the optimum surface for the catalyst has been designed and in so doing this has improved the underlying catalytic reaction. In addition a material has been created that is capable of being formed into spheres by granulation and hence the next step has been taken in the process technology for making chiral epoxides – allowing a fixed bed continuous process.

Only via continuous operation can these important building blocks of asymmetric pharmaceutical intermediates be made at the manufacturing cost required to fulfil the potential market for chiral epoxides.