(336d) Developing Around Uncertainty in the Scale up and Tech Transfer of Flow Reactions with Highly Reactive Reagents | AIChE

(336d) Developing Around Uncertainty in the Scale up and Tech Transfer of Flow Reactions with Highly Reactive Reagents


Franklin, R. - Presenter, Georgia Institute of Technology
Armiger, T., Carnegie Mellon University
Patel, P., Merck and Co. Inc.
Otte, D., Merck and Co. Inc.
Hall, J., Merck and Co. Inc.
Larson, R., Merck and Co., Inc.
Halsey, H., Merck and Co., Inc.
Rodrigues, V., Merck & Co., Inc.
Jellett, L., Merck & Co., Inc.
Kuhl, N., Merck & Co., Inc.
Corry, J. P., Merck
Chung, C., Merck & Co., Inc.
Flow reactions involving highly energetic and moisture-sensitive compounds, like organolithiums and Grignard reagents, present unique challenges during scale-up and tech transfer for GMP intermediates. Concerns about operator safety and sample stability often prevent on-site release assays to confirm reagent concentrations, and small amounts of water in the equipment train can have a large impact on reaction performance. Additionally, differences in heat transfer and mixing performance across scales can cause uncertainty in the applicability of bench-scale process characterization experiments to full-scale operations. This talk will outline the strategies used to mitigate the uncertainties presented by the scale up of a multi-reaction flow step in the synthesis of the MK-1026 ketone intermediate, a penultimate step in the commercial route towards nemtabrutinib. As shown in the figure below, a batch deprotonation of the aryl bromide (ArBr) is carried out with the addition of methyllithium. A cascade of two reactions is then carried out in a tubular flow reactor, starting with the addition of n-butyllithium, followed by biaryl ether. An overcharge or undercharge of either of the two organolithium reagents can result in unacceptable impurity levels in the isolated ketone intermediate. A strategy for calculating, monitoring, and adjusting process stream flow rates enabled robust and consistent full-scale operation while maximizing yield. Comparison of repeated experiments across scales spanning multiple orders of magnitude enabled significantly more economical process characterization experiments on the bench to be used to guide large-scale processing decisions. The result was the successful transfer of the process from the R&D labs and pilot plant to the final commercial site and demonstration of robust and consistent performance across eight full-scale batches over nine months of operation.