GlaxoSmithKline Adopts Continuous Manufacturing for Preparation of Key Intermediate

The Suzuki−Miyaura coupling reaction has become a widely used method for creating carbon–carbon bonds. The palladium-catalyzed cross coupling between organoboron compounds (boronic acids, potassium trifluoroborates, organoboranes, or boronate esters) and halides (aryl, alkyl, alkenyl, and alkynyl) or pseudohalides (i.e., triflates). This reaction is attractive because it proceeds under mild conditions, the boron compounds are generally easy to prepare or readily available, and separation of the desired product from the inorganic byproducts is readily achieved. In addition, boron compounds are less toxic than other metallic species use in carbon–carbon coupling reactions, such as organostannanes, and the reaction is fairly tolerant of functionality. In fact, the Suzuki−Miyaura coupling reaction can also be effective for the coupling of functionalized methyl boronates with functionalized methylene groups bearing dialkylamine, ether, or thioether substituents.

Unscalable conventional method

Researchers at GlaxoSmithKline (GSK) recently developed a synthetic route to a drug candidate that involved this type of Suzuki−Miyaura coupling reaction, and as a result needed multikilogram quantities of potassium bromomethyltrifluoroborate, which is the key intermediate for preparation of functionalized methyl boronates (1). “The Suzuki−Miyaura coupling reaction was an early step in the overall synthesis, and we, therefore, needed to develop a rapid, economical, and practical method for preparing potassium bromomethyltrifluoroborate on a large scale,” notes Toby Broom, a chemist with GlaxoSmithKline Research and Development.

The conventional method for the synthesis of this compound, which involves treatment of dibromomethane (CH₂Br₂) and triisopropyl borate ((iOPr)₃B)with n-butyl lithium (n-BuLi) in tetrahydrofuran (THF) to form a lithium salt intermediate followed by treatment with potassium bifluoride (KHF₂), was not suitable for large-volume manufacturing. Cryogenic temperatures are necessary to obtain high purity product, but the reaction is very exothermic and heat removal is challenging on a large scale, according to Broom. In addition, he notes that the quantity of solvent required upon scale-up is not practical.

Ideally suited for flow
The highly exothermic nature of the reaction led the GSK researchers to adopt a continuous-flow strategy for the synthesis of bromomethyltrifluoroborate (1). Continuous manufacturing is attractive for such reactions because it generally involves small reactor volumes combined with enhanced control of reaction conditions. Specifically, because materials are introduced continuously and react on contact with continuous removal, there is better control of process variables, and the risk of side reactions is lower. For this particular reaction, a flow process was also attractive because maintaining a cryogenic temperature (-78 °C) in a microreactor would be more cost-efficient, and continuous consumption of the unstable lithium intermediate would avoid its decomposition, according to Broom.

Optimizing process conditions

Initial laboratory studies in a plate reactor demonstrated the feasibility of preparing the lithium intermediate using a flow approach (1). The biggest challenge with this reaction was the side reaction of n-BuLi with (iOPr)₃B, which led ultimately to the formation of butyltrifluoroborate, a product that was very difficult to separate from the desired bromomethyltrifluoroborate. “This problem was overcome by using an excess of dibromomethane and just one equivalent of triisopropyl borate,” notes Broom.

The reaction with KHF₂ also proved to be a challenge (1). This reactant has limited solubility in water, and thus a very dilute solution was required that led to unacceptable workup volumes, according to Broom. “At this stage, we needed to find an alternate method for converting the intermediate lithium triisopropylborate salt to the desired bromomethyltrifluoroborate. We found that 20 molar aqueous hydrofluoric acid (HF) was effective for obtaining the lithium rather than the potassium salt.”

 In this reaction, however, the formation of small amounts of insoluble lithium fluoride clogged the reactor. After extensive evaluation, therefore, the researchers elected to carry out this step in a stirred reactor (1) rather than maintain an entire continuous process. Even though the quenching reaction wasn’t fully continuous, Broom emphasizes that the addition of HF in the plant would be automated, thus reducing operator exposure. An added benefit of the use of HF was its lower cost compared to that of KHF₂.

Addressing reactor plugging on a larger scale

With a successful laboratory-scale process in hand, the GSK researchers moved to the pilot plant, where a less efficient, but more economical, tube-and-shell reactor was employed. With this reactor, however, the temperature could not be maintained sufficiently low enough, and precipitates formed that blocked the reactor (1). “Fortunately, once we went back into the lab and ran the reaction using a similar type of reactor, we were able to address this problem. Use of a more dilute triisopropylborate solution, a faster flow rate, and the addition of half of the nBuLi to each of two reactors in series helped prevent the temperature form rising above acceptable levels,” Broom explains.

In the pilot plant with these modifications, the GSK researchers completed a 240 hour continuous run with 10 kilograms (kg) of lithium bromomethyltrifluoroborate obtained per day (total 100 kg). This material was then subjected to a cation exchange reaction with potassium fluoride to convert it to the potassium salt (1). Broom notes that this last conversion step was time-consuming and could potentially be achieved with a continuous flow separation method as well, but the GSK researchers did not have time to investigate this approach. “Direct use of the lithium salt in the Suzuki−Miyaura coupling reaction is also worth investigating,” he notes.

Successful hybrid strategy

One conclusion the researchers reached was that a hybrid approach involving the use of each type of process—continuous and batch—where most appropriate was very successful in achieving the goals for the project. “Combined use of continuous flow and batch-type processes was very effective at meeting our needs for the rapid, practical, and efficient large-scale manufacture of this key intermediate,” Broom concludes.

Reference

T. Broom et a.,l

Org. Process Res. Dev.,

online, DOI10.1021/op400090a, June 10, 2013.