1
Flow Chemistry at the Extremes: Turning Complex Reactions into Scalable Processes

Andrew R. Bogdan

AbbVie, Inc., Drug Discovery Science and Technology, 1 North Waukegan Rd, North Chicago, IL 60064, USA

1.1 Introduction

The use of flow chemistry within the pharmaceutical industry is often used to facilitate the discovery of an active pharmaceutical ingredient (API) or to make its manufacturing route more efficient. Through the combined efforts of academia and industry, significant advances have been made in the field of flow chemistry, which in turn has led to a prevalence of this technology in high-impact settings. The benefits of running reactions in flow are well documented in a number of comprehensive reviews [1-17]. With these reviews, how flow setups can range drastically in their complexity has become obvious. Flow systems in early-stage pharmaceutical settings tend to be automation-intensive platforms, focused on reaction scouting, automated optimization, or library synthesis. For later-stage programs, however, the focus becomes designing highly efficient routes to synthesizing intermediates or final APIs. Depending on the setting, however, the same practical considerations need to be addressed before setting up a reaction in flow. Do reactions need to be run at very low or very high temperatures? Would improved heat transfer and mixing optimize yields and selectivity? If the synthetic route involves the use of a hazardous species, can flow be used to generate this material in situ? Is an alternate energy source such as light or current required? If the answer to any of these questions is yes, flow chemistry should be explored. If reactions are heterogeneous or sluggish, however, conventional batch reactors may still be preferred (Table 1.1). In this chapter, a number of examples from the pharmaceutical industry will be discussed where flow chemistry shows obvious advantages over batch techniques. These examples can likely be used as a basis for running future reactions in flow before running reactions in batch. Over time, a number of trends have emerged, and more and more often, specific types of chemistry are preferentially being run in flow on a large scale (Figure 1.1).

Table 1.1 Examples of when and when not to use flow chemistry in the pharmaceutical industry.

In situ use of hazardous intermediates Reactions mediated by alternative energy sources (photochemistry)

Slow reaction rates

Figure 1.1 Legend for flow reactor equipment.

1.2 Temperature Extremes

Flow chemistry has remained a tried and true method for running low- and high-temperature reactions ever since the field started taking off in the early 2000s. Due to the high surface-area-to-volume ratios, flow reactors have unmatched heat transfer, often times leading to high yields and clean reaction profiles. For cryogenic reactions, flow reactors are readily able to dissipate any exotherms that may be generated. As a result, reactions that run at cryogenic temperatures in batch can frequently be run at higher temperatures in flow. Flow reactors can also reach temperatures that may otherwise be unattainable in batch, thus accelerating reaction rates and resulting in chemistry that is not feasible in batch. High-temperature reactions can also be run much safer in flow as active reactor volumes are lower in comparison to batch systems. In this section, a number of examples of flow chemistry at these two temperature extremes will be discussed in the context of the pharmaceutical industry.

1.2.1 Cryogenic Flow Chemistry

To grasp the differences in cooling a batch reactor versus a flow reactor, it is easiest to first envision cooling a large round-bottom flask. The heat transfer in this instance mainly occurs at the walls of the flask, meaning that as a flask size increases, the effective cooling of the reaction changes and temperature gradients are likely being formed across the reactor. In order to ensure proper reaction control, it is necessary to have excellent mixing and to add reagents in a dropwise manner to keep temperatures from fluctuating. Flow reactors, however, have dimensions on the order of millimeters, which result in efficient cooling and limited hotspots. Being able to rapidly dissipate exotherms results in reactions that can be run much more cleanly and at potentially higher temperatures. It is also worth noting that while a 500 ml reaction in batch would likely require at least a 1 l reactor, the same-sized reaction in flow would require a far smaller reactor volume (potentially even a few milliliters). As opposed to a stir bar or impeller, flow reactors can be mixed using a number of options that vary depending on scale (Figure 1.2).

Figure 1.2 Examples of mixers for use in flow chemistry. (a) Standard T-mixer, (b) narrow-bore tubing, (c) IMM static mixer, (d) microchip mixer, or (e) Koflo static mixer.

1.2.1.1 Organolithium Chemistry in Flow

Perhaps one of the most prevalent types of flow chemistry involves the use of organolithium species such as n-butyllithium or lithium diisopropylamide [18]. While in batch, these reactions are predominately run at -78 °C or lower, for safety and selectivity reasons. For these reason, running these reactions on large scale in batch can be somewhat limiting if these concerns are not mitigated. As a result, more and more examples of organolithium-mediated flow chemistry are being described within the literature. Frequently, these examples can be classified as "flash chemistry," a term coined by the Yoshida group, where reactions take place on the order of milliseconds to seconds [19].

1.2.1.1.1 Boronate Synthesis

Flow examples using organolithiums commonly involve a rapid deprotonation/transmetalation followed by a quench with some sort of electrophile such as a boronate (Scheme 1.1). These flow processes are completed typically by the use of some form of aqueous quench in batch, leading to the isolation of the desired product. A number of examples have been described to generate aryl boronates from both academic and industry laboratories. Ley's group has used the commercially available Polar Bear reactor to prepare gram quantities of (4-chloro-2-fluoro-3-methoxyphenyl)boronic acid [20, 21]. In this instance, the o-lithiation/quench occurred at -50 °C and was capable of preparing >100 g of product within seven hours.

Scheme 1.1 Continuous flow boronate synthesis.

Similarly, Novartis has reported a flow borylation based upon the transmetalation strategy [22]. Both the metalation and boronate quench occur at -30 °C, with a total residence time in the reactor of just one second. When used in scale-up mode, this process proves to be of incredibly high throughput, generating 146 g of product in just 25 minutes. The process operates at incredibly high flow rates, which coupled with the short reaction times allows kilos of this material to be generated in short order if needed. Takeda has exemplified two lithiation-borylation sequences that are used to generate kilogram quantities of various boronates [23, 24]. As with the work by Novartis, these examples could be characterized as "flash" chemistry, as their combined residence times are on the order of seconds. In the first reported example, a Boc-protected aminobenzoxazole is borylated at 0 °C. When run ~10 hours, 1.23 kg of the final boronate is isolated. Similarly, Takeda later reported their efforts to synthesize two aryl ether boronates on a large scale using a very similar procedure. Again, the use of flow chemistry permitted these reactions to occur at -15 °C, much higher than the norm for these types of reactions, allowing 0.75-1.55 kg of final material to be prepared in less than 5 hours. In all the cases described earlier, the use of flow has shown to be beneficial over traditional batch reactions. All processes are rapidly scaled and can operate effectively at temperatures >-78 °C. Not only does this increase in temperature provide some increase in kinetics, but also in many cases the warmer temperature can boost solubility, making the reactors far more stable for long-term operation.

1.2.1.1.2 Formylation

Similar to the boronate synthesis, aldehydes are frequently synthesized via a two-step flow process using organolithium species (Scheme 1.2). Again, a lithium-halogen exchange is carried out at low temperature in flow, followed by a rapid quench with N,N-dimethylformamide to afford the final product. Large-scale flow runs using this chemistry have been described by chemists at both Merck and Takeda. At Merck, a route to synthesize kilograms of a formylated intermediate was developed [25]. In this case, a lithium salt was used as a starting material in order to...