News from Europe's pharmaceutical manufacturing industry coupled with upcoming events, and exclusive articles and interviews from industry experts. WEEKLY
In the original synthetic protocol for the DuPhos ligands, 1,2-bis (phosphino) benzene was deprotonated with n-BuLi, followed by addition of the cyclic sulfate and then further n-BuLi (see Figure 4) (14). This approach gave the ligands in good yields and purities. Upon scaling up this process to produce
kilogram amounts, yields were not reproducible, and higher levels of impurities were formed. These problems were readily solved
by an inverse addition procedure, whereby n-BuLi is added to a mixture of 1, 2-bis (phosphino) benzene and the cyclic sulfate. This process is readily scaleable, giving
high purity ligand in excellent yields on multikilogram scales (16). This process also has been applied to other 1,4-diols
to give access to ethyl and isopropyl phospholane ligands, as well as the related 1,3-diols for the proprietary phosphetano
ligands, FerroTANE (Dowpharma, Midland, MI) (17).
Rhodium catalyst fabrication
Figure 5
For some asymmetric hydrogenation catalyst systems, it is acceptable to make the precatalyst in situ from a metal complex and the ligand. We have the general approach that it is better to preform and isolate the ligand–metal
catalyst complex so it can be charged to the reactor as a defined species. This approach results in more robust and reproducible
reactions, with greater quantification (important for CGMP manufacture) and no debate as to whether the precatalyst complex
has completely formed. Initially, standard established literature methods were used to make the Rh-DuPhos complexes: (1,5-cyclooctadiene)
Rh(I) acetylacetonate] is converted into the sparingly soluble Rh bis(1,5-cyclooctadiene) tetrafluoroborate complex (18),
and the ligand reacts with this intermediate to provide the precatalyst complex in solution. Addition of an antisolvent is
required to precipitate the desired product (see Figure 5). This method worked well for a range of diphosphine ligands, but
provided material in modest yields (~70%) and variable product form. Obviously, failing to isolate 30% of valuable ligand
and metal precursor is neither desirable nor economical.
A further problem with this method was that the product form was variable, from robust deep red crystals to orange-yellow
amorphous powders. Although all of these forms performed well in asymmetric hydrogenation reactions, the powdered materials
had shorter shelf life and were prone to degrade more easily than the crystalline material. Further process development led
to a scaleable process that delivers crystalline rhodium precatalysts in very high yields (91–97%) with excellent chemical
purity and substantially improved chemical and physical stability.
Figure 6
The new process involves taking [(1,5-cyclooctadiene) Rh(I) acetylacetonate] in an ethereal solvent (see Figure 6a), treating
it with an alcohol solution of strong acid, such as tetrafluoroboric acid, to give a soluble bis-solvato species (see Figure
6b), which is then reacted with an ethereal solution of the bisphosphine ligand (see Figure 6c). Shortly after the addition
of the ligand crystallization of the precatalyst complex is observed. This protocol controls the rate of nucleation at higher
temperatures through rate of ligand addition, such that granular, free-flowing precatalyst is deposited in exceptionally high
yields (19).
