PHARMECUTICALS
Figure 3 - Flow reactor design during development
for early phase work . It is reported that 1 was prepared by the addition of
CO 2 to the lithium anion 3 , which was generated from 2 at -78 ° C in 75 % yield .
We found that the reaction ’ s sensitivity to temperature and yields were much lower at larger-scale runs , with significant amounts of dark tarlike material isolated during work-up . In addition , several impurities were formed at higher (> -65 ° C ) temperatures , due to the instability of the anion . Such uncontrolled reactions on a larger scale due to lack of sufficient heat transfer could lead to the formation of reactive benzyne-type intermediates , which polymerise violently .
We wanted to develop an improved process for the production of 3 using continuous flow chemistry . Our rationale for choosing it was based on four factors : ease of performing low-temperature reactions ; high mixing of the gas-liquid phase ; excellent heat transfer capacity under flow conditions ; and consistent yield irrespective of the scale .
Initial reactor design
Our initial concept of the reactor consisted of three loops ( Figure 3 A , B and C ) where A was used to cool
Table 1 : Details of scale-up runs a mixture of 2 and TMEDA in THF to -78 ° C ( pictured ). Just after A , the base was added and B provided the necessary residence time for the anion formation after which CO 2 was added as a gas which passed through C to give the product .
The entire reactor was submerged in a dry ice acetone bath . For the initial design , the loops were made of 1.6 mm inch HDPE tubing with two HPLC pumps for the reagents . The mixing units consist of 6.3 mm ID diameter tubing with two small stir bars trapped within . The stir bars were agitated using a magnetic stir plate , providing turbulence sufficient for mixing of the reagents . The back-pressure unit at the end of the product stream was set to 0.7 bar .
The residence times for the trial runs were based on reaction monitoring in a batch mode ( HPLC analysis ). Interestingly , the addition of the base to 2 in THF is marked with a colour change from pale orange to dark red , which decolourises upon quenching . The flow rates were adjusted to achieve a base stoichiometry twice as high as the starting material .
For the first few experiments , a small
CO 2 cylinder was directly connected
to the flow reactor . The product stream was collected after the steady stage was achieved and worked up in a batch mode by quenching with 2N HCl , extraction with ethyl acetate , and telescoping it to the subsequent reaction . After several such trial reactions , the parameters were optimised and product-dependent parameters were understood .
For the scale-up batches , three stainless steel tubes , 8 mm in diameter were built , coiled and immersed into a carboy filled with dry ice acetone . Six static mixers were inserted into the tubes after both the anion formation and CO 2 quench to provide the required mixing . Each batch was performed within one day and five such batches were sequentially carried out . Table 1 summarises the details .
Conclusion
Continuous flow chemistry offers quality , innovation and safety advantages for early-phase API synthesis . Developing flow processes requires a deep understanding of chemistry . Scaling them up entails engineering expertise . Doing both efficiently requires a full-service CMO with the infrastructure to support complex projects and the knowhow to innovate . ●
Entry
Scale ( kg )
Anion Formation
Residence Time ( min )
CO 2
Quench
Purity (% AUC )
1 5.4 2.0 91.6 2 3.5 2.0 0.9 94.4
Yield Over Two Steps
3 5 3.6 1.6 97.2 88 4 4 3.6 1.6 98.2 5 4 3.6 1.6 97.8
(%)
81
91
Sripathy Venkatraman
VICE PRESIDENT , HEAD OF GLOBAL R & D OPERATIONS
CURIA k + 1 518 331 5182 J sripathy . venkatraman @ curiaglobal . com j www . curiaglobal . com
18 SPECIALITY CHEMICALS MAGAZINE ESTABLISHED 1981