Towards More Efficient, Greener Syntheses Through Flow Chemistry

Towards More Efficient, Greener Syntheses through Flow Chemistry

The MIT Faculty has made this article openly available. Please share how this access benefits you. Your story matters.

As Published https://doi.org/10.1002/tcr.201600139

Publisher Wiley Blackwell

Version Author's final manuscript

Citable link http://hdl.handle.net/1721.1/114543

Detailed Terms http://creativecommons.org/licenses/by-nc-sa/4.0/ PERSONAL ACCOUNT

Alternatively, entire flow systems with specialized attachments Towards More Efficient, can be purchased from various commercial suppliers.[4b]

Greener Syntheses Through Flow Chemistry

Justin A.M. Lummiss,[a] Peter D. Morse,[a] Rachel L. Beingessner,[a] and Timothy F. Jamison*[a]

Abstract: Technological advances have an important role in the design of greener synthetic processes. In this Personal Account, we describe a wide range of thermal, photochemical, catalytic, and biphasic chemical transformations examined by our group. Each of these demonstrate how the merits of a continuous flow synthesis platform can align with some of the goals put forth by the Twelve Principles of Green Chemistry. In particular, we illustrate the potential for improved reaction efficiency in terms of atom economy, product yield and reaction rates, the ability to design synthetic process with chemical and solvent waste reduction in mind as well as highlight the benefits of the real- time monitoring capabilities in flow for highly controlled synthetic output. Figure 1. (a) Comparison of the major unit operations in batch relative to flow. (b) Additional tools utilized in flow reactions including a (left) mass flow controller (MFC), which is used to regulate gas flow and (right) check valves (CV), which are used to prevent back flow. 1. Introduction

[1] First outlined by Anastas and Warner in 1998, the Twelve There are several benefits of flow chemistry compared to Principles of Green Chemistry provide a valuable framework for traditional batch synthesis.[3e, 5] For example, the high surface evaluating the efficiency and sustainability of a given chemical area to volume ratio of a flow reactor enables superior mixing of [2] transformation or process. On the molecular level, these biphasic reactions, such as those between gases and liquids, principles call for the design of reactions that are more atom which can significantly improve product yields.[5b] This same economical, minimize the use of hazardous reagents, and utilize attribute also permits precise temperature control, which can renewable feedstocks. The design of reaction protocols with increase reaction efficiency (e.g. reaction rate, product yield) energy efficiency in mind, and employing real-time reaction while minimizing energy consumption.[3d] Photochemical analysis in order to minimize waste generation are also among transformations can similarly benefit from improved yields, the recommendations. Technological advances, such as decreased reaction time-scales, and reduced catalyst loadings, continuous flow synthesis, have an important role to play in due to the highly efficient irradiation that results from having a [3] advancing these goals. short path length.[5d, 6]

Continuous flow reactions on the meso- to macro-scale level Another distinguishing feature of continuous flow is that the are typically carried out within commercially available small amount of product generated is determined by the length of time diameter tubing (inner diameter 0.01-0.080”) fabricated from the entire flow regime is operated, given defined flow rates and either polymer (e.g. perfluoroalkoxy alkanes, Tefzel, reactor volumes. This is in contrast to batch, where the polyetheretherketone) or metal (e.g. stainless steel, copper) maximum quantity of product produced per reaction, is [4] materials. A range of Swagelok and HPLC fittings enable the predetermined by the quantity of starting material. It is this very relatively facile assembly of bespoke flow set-ups using some or nature of continuous production that enables small volume flow all of the components showcased in Figure 1a and b. reactors to produce comparably large quantities of product. The reduced reactor volume in flow also enables more efficient engineering controls and also inherently reduces safety risks by

[a] Dr. J. A. M. Lummiss, Dr. P. D. Morse, Dr. R. L. Beingessner, avoiding large accumulations of potentially hazardous Professor T. F. Jamison intermediates at any given point in time. Department of Chemistry Massachusetts Institute of Technology 77 Massachusetts Avenue, Cambridge, MA, 02139, USA E-mail: [email protected]

PERSONAL ACCOUNT

A major research focus of our group lies in leveraging continuous flow platforms to improve reaction outcomes and Tim Jamison was born in San Jose, CA, and enable transformations that are difficult, or simply not feasible to grew up in neighboring Los Gatos, CA. He perform in batch.[7] In this Personal Account, we highlight several received his undergraduate education at UC of these reactions that also align with the goals outlined by the Berkeley, where he conducted research in Twelve Principles of Green Chemistry.[1] The examples the laboratory of Prof. Henry Rapoport for described are diverse in nature and include a number of thermal, nearly three years. He was then a Fulbright photochemical, catalytic and biphasic (e.g. gas-liquid) reactions. Scholar with Prof. Steven A. Benner at the While they have not been scrutinized with regards to their overall ETH Zurich, and thereafter he undertook his greenness, our aim is simply to illustrate how our goals for Ph.D. studies at Harvard University with improved reaction efficiency in terms atom economy, product Prof. Stuart L. Schreiber. He then moved to yields and reaction rates, offer opportunities within the context of the laboratory of Prof. Eric N. Jacobsen at green chemistry. We also aim to demonstrate the feasibility of Harvard University, where he was a Damon chemical and solvent waste reduction within a flow synthesis Runyon-Walter Winchell postdoctoral fellow. In 1999, he began his independent career at MIT, where he currently holds the positions of R. R. design as well as highlight the merits of real-time monitoring for Taylor Professor and Head of the Chemistry Department. maintaining the integrity of larger-scale production. Overall, our hope is that the reader considers the prospects of continuous flow as a stepping stone towards greener, more efficient syntheses. 2. Flow-Enabled Greener, More Efficient Syntheses Justin Lummiss received his Ph.D. in 2015 from the University of Ottawa, for work with 2.1. Thermal Reactions - Improved Product Yields and Deryn Fogg on mechanistic organometallic Reaction Rates, Efficient Heat Transfer Processes chemistry in olefin metathesis. He is presently an NSERC postdoctoral fellow A central pillar toward greener chemistry is the design of with Timothy Jamison at MIT. increasingly efficient syntheses. While reducing the volume of waste generated by reactions is the primary means for improving E-factors, additional aspects such as increasing the product yield, decreasing the reaction time, carrying out more atom economical transformations and minimizing energy loss all impact the overall efficiency of a synthesis.

Peter Morse studied Structural Biology and One of the well-known benefits of continuous flow is the Chemistry at the University of Connecticut ability to safely heat a reaction mixture well beyond the boiling (B.S.). In 2010, he joined the lab of David point, by regulating the pressure of the system with a back Nicewicz as a graduate student at the pressure regulator shown in Figure 1a.[4b, 4c] While high University of North Carolina at Chapel Hill. temperature reactions in batch are well-established and diverse His work there focused on the development in the literature, the use of smaller reactor volumes in flow of new synthetic methods in the burgeoning reduces the risks associated with reactor failure and also field of photoredox catalysis. After obtaining facilitates reactor containment. The latter not only has safety his Ph.D. in 2015, Peter is now a implications, but also enables the insulation and recapturing of postdoctoral researcher in the lab of Tim lost energy more easily than in a larger batch process.[3d] Overall, Jamison. His current research focuses on from a manufacturing perspective, the opportunity to increase developing new methods to synthesize throughput and decrease residence time by accessing forcing bioactive molecules in flow. conditions (e.g. higher temperatures and pressures)[8] in an effective manner, has significant impacts for volume-time-output and the overall process efficiency.[9] Rachel Beingessner obtained her Ph.D. at the University of Ottawa in 2007 and then As an illustration of this concept, in 2010 we reported the batch joined the National Institute for and continuous flow synthesis of β-amino alcohols, a Nanotechnology – National Research functionality found in a number of active pharmaceutical Council Canada for one year of postdoctoral ingredients (APIs), which can be formed via the aminolysis of training prior to transitioning to a staff epoxide substrates.[10] As shown in Figures 2a and 2b, the ring research position. In 2015, she joined the opening of 2-(phenoxymethyl)oxirane (1) with tert-butylamine (2), Chemistry Department at MIT where she resulted in complete conversion and an 82% in situ yield of currently works as a Research Scientist. amine 3 under directly comparable flow and microwave batch

conditions (150 °C for 30 min). However, by taking advantage of the ability to safely increase the reaction temperature from

PERSONAL ACCOUNT

150 °C to 195 °C in flow, we were able to achieve an order of In addition to process intensification,[8] another significant magnitude reduction in the time scale of the reaction (3 min vs. attribute of a flow reactor is the high surface area to volume ratio 30 min), while maintaining a comparable yield and conversion that results from using small diameter tubing. This enables rapid (Figure 2b). Process intensification[8] in flow also improved the and highly controlled heat transfer processes, which in turn sluggish reaction between 1,4-dihydronaphthalene oxide 4 and avoids the need for lengthy, energy-costly temperature hindered secondary amine, indoline 5, from 39% at 150 °C to equilibration periods typical in batch. This is particularly 71% at 245 °C within the same 30 min time frame (Figure 3a). advantageous for multistep syntheses that require very distinct Rate enhancements were likewise observed for the preparation and extreme differences in temperature, such as our reported of intermediate 9 in the synthesis of the indacaterol active synthesis of ortho-substituted phenols, which involved the pharmaceutical ingredient (API),[11] as well as the synthesis of aerobic oxidation of in situ generated Grignard reagents (Figure metaprolol 12 (Figure 3b).[12] Both compounds were prepared in 4).[13] In this case, initial benzyne formation from 15 and good yields with residence times of only 15 min and 15 sec, nucleophilic addition required high temperatures (80-120 °C) in respectively. Reactor II, whereas the subsequent oxidation of the Grignard reagent with the renewable air feedstock in Reactor III, necessitated chilled conditions (-25 °C) to generate the ortho- substituted phenols 16-23. Rapid cooling was accomplished by simply flowing the mixture through a precooling coil for a residence time of only 1.2 min.

Figure 2. Reaction of 2-(phenoxymethyl)oxirane (1) with tert-butylamine (2) under (a) microwave batch conditions at 150 °C and (b) flow conditions.[10]

Figure 4. Synthesis of ortho-functionalized phenols using an integrated three- step continuous flow system with three unique temperature controlled reaction zones. Total residence time for the synthesis of 16 - 23 was 14 min.[13]

Efficient heat transfer is equally important during the quenching of a reaction that could potentially lead to an exothermic process. In this regard, the ability to continuously Figure 3. Process intensification in flow. Aminolysis reaction of secondary quench small volumes in-line or by dropwise collection of the [14] amine (4) 1,4-dihydronapthalene oxide (5). Reactions were run in EtOH at 1 M product stream into an appropriate quenching solution, concentration in epoxide (b) Aminolysis reactions applied to the synthesis of enables a controlled and convenient process, particularly where complex targets 9 and 12. The syntheses were performed in 9:1 NMP/H2O collection occurs over prolonged periods. In 2012 for example, and EtOH, respectively.[10] we reported the oxidation of a variety of aldehydes and alcohols

PERSONAL ACCOUNT

using bleach, a phase transfer catalyst and a MeOH/EtOAc previously established batch photochemical reactions to solvent mixture.[14a] In the case of 5-nitro-2-furaldehyde, a total mesoscale continuous-flow, in order to demonstrate the orders of 11.0 g of methyl 5-nitro-2-furoate was generated in 75 min, of magnitude increase in reaction efficiency that is possible which corresponds to 211 g of product per day. By conducting within this reaction class.[18] As a specific example, the oxidative the reaction in flow, small volumes of bleach were continuously generation of iminium ions from 24 was carried out using a flow and safely quenched at room temperature, thereby avoiding the reactor constructed from PFA tubing and commercial blue LEDs need for more energy-intensive cryogenic temperatures.[15] (Table 1). The reactive iminium 25 generated in situ was immediately trapped with a nucleophile. Under batch conditions, 2.2. Photochemical Reactions – Increased Efficiency and formation of the α-functionalized amine 26 in a 95% yield from Minimization of Solvent Use N-aryl tetrahydroisoquinoline (24), required 1.0 mol % of the

Ru(bpy)3Cl2 photocatalyst and a 3 h reaction time. Under Photochemical processes offer unique reactivity patterns and analogous flow conditions, a 50% reduction in catalyst loading several advantages towards a more green synthesis: photons (ca. 0.5 mol %) and a 360-fold decrease in reaction time (ca. 0.5 are an innocuous renewable reagent, radical initiators can be min) afforded the same product in a comparable 89% yield. avoided, and in some cases, visible light can be sufficient to Table 1. Comparison of the photochemical generation and trapping of iminium [16] promote these reactions. In batch, however, photochemical ions beginning from 24 in a flow and batch platform.[18] transformations suffer from limited light penetration into the reaction mixture, which becomes increasingly problematic on scale as described by the Beer-Lambert law.[6a] A flow platform, alternatively, utilizes narrow diameter tubing that results in a very short path length and, as such, a significantly more efficient irradiation process. Scaling this type of reaction in flow can easily be achieved by running parallel reactors, or, if time permits, simply increasing the duration of the product collection period. While in-house reactors such as that shown in Figure 5a can be readily fashioned for UV, LED, and visible light, flow It has also been reported previously that certain photoboxes are also commercially available (Figure 5b and photochemical reactions can be performed at significantly higher 5c).[17] concentrations in flow reactors than comparable batch reactions, with little to no deleterious effect.[16d, 18-19] This improves the overall greenness of the reaction by minimizing the amount of solvent waste generated. We have similarly observed this effect in our development of a continuous flow protocol for the photochemical synthesis of the versatile, commercially-available

half-sandwich catalyst CpRu(MeCN)3PF6 (28). In previously reported batch conditions, the maximum concentration attainable was only 20 mM and the reaction required 12 h to reach full conversion (Figure 6). By transitioning to continuous flow, the reaction was not only accelerated from 0.5 d to 5 min, but the concentration of the reaction was also increased 5-fold, while still maintaining a near-quantitative 98% conversion.[20] It should be noted that 100 mM was the highest achievable concentration due to the limiting solubility of the product 28 in acetonitrile and not a result of a limitation in the photon flux.

Photochemical flow processes also have the potential to provide access to materials in a more atom economical manner than would typically be accomplished using more traditional synthetic strategies.[21] As an illustration, conventional methods for amide and peptide bond formation rely on the condensation Figure 5. Representative continuous-flow photochemical reactors: (a) in- of carboxylic acids and amines and typically require the use of house LED reactors and (b and c) commercially available UV-photobox. Image stoichiometric quantities of an activating agent. In contrast, we (a) is reprinted with permission from Tucker, J. W.; Zhang, Y.; Jamison, T. F.; Stephenson, C. R. J., Visible-Light Photoredox Catalysis in Flow. Angewandte recently showed that a photochemical rearrangement of nitrones Chemie, International Edition 2012, 51 (17), 4144-4147. Copyright 2012 John 29 in flow via oxaziridine intermediates 30, can provide facile Wiley & Sons. access to range of amides, dipeptides, tetrapeptides, as well as symmetrical diamides in reasonable yields (Figure 7, compounds 31 - 40).[22] For instance, complete photochemical In collaboration with Corey Stephenson’s group, currently at conversion of the corresponding nitrone to amide 34 was the University of Michigan, we have translated a number of

PERSONAL ACCOUNT

observed within 10 min, with no oxaziridine remaining and a 59% isolated yield of the desired amide. While the analogous batch reaction in an NMR tube did proceed with complete conversion, the oxaziridine was the major product over the desired amide (7:1).[22] Thus, the enhanced irradiation efficiency of the flow reactor is key for promoting this atom-economical, photochemical transformation.

Figure 6. Photochemical synthesis of 28 in (a) batch and (b) continuous flow.[20, 23] Figure 7. Examples of dipeptides, tetrapeptides, and symmetrical diamides accessed via the continuous-flow photochemical rearrangement of the corresponding nitrones.[22] [a] Synthesis was performed in benzene. Overall, the short path length that results from the high surface area to volume ratio in mesoscale continuous flow reactors offers enormous opportunities for increased yields, azides in situ from alkyl halides and sodium azide,[24] the need decreased time-scales, and reductions in catalyst loadings for their handling was also eliminated, thereby creating a much within photochemical reactions. In combination with the potential safer and highly efficient process compared to batch protocols. use of more concentrated solutions relative to batch, Macrocyclizations using a similar strategy were also photochemical flow reactions bode well towards the goals of demonstrated by the same group shortly thereafter.[25] In 2014, greener chemistry. we reported the continuous flow synthesis of rufinamide (43), an anticonvulsant medication, that was prepared utilizing a copper 2.3. Catalytic Reactions – Reactor Tubing as a Catalyst flow reactor to catalyze the click reaction between an in situ Source formed aryl azide and propiolamide. As shown in Figure 8, the API was obtained in 92% yield using a telescoped 3-step The development of catalytic methods and the Principles of process.[26] This convergent synthesis compares favorably with Green Chemistry[1] pursue a common endeavour; both seek to Novartis’ patented route to rufinamide, which involves isolation develop synthetic processes that maximize the yield of desired and handling of 2,6-difluorobenzyl azide and utilizes flammable products while minimizing inputs (i.e. energy, time, etc.) and and toxic 2-chloroacrylonitrile as a reagent.[27] waste generated. To this end, several groups including ours have exploited the use of metal reactor tubing, as opposed to In addition to the triazole synthesis described previously, we polymer tubing, to provide a direct source of catalyst in flow have also illustrated the use of a copper-tubing reactor (Figure reactions. In 2009 for example, Bogdan and Sach[24] reported 9) for atom economical and generally high-yielding Sonogashira the synthesis of a library of 1,4-disubstituted 1,2,3-triazoles cross-coupling, protiodecarboxylation and Ullmann coupling using a copper flow reactor to catalyze the “click” reaction reactions.[26, 28] In the case of the Sonogashira reaction (Figure between acetylenes and organic azides. By generating these 10), products 46 – 50 were generated after a 30 min residence time at 170 °C in DMF and only required tetra-n-butylammonium acetate (TBAA) as a base.[28] Although leached copper was observed from the reactor, it was easily removed using the solid- supported metal scavenger, Quadrapure Thiourea.

PERSONAL ACCOUNT

Figure 8. Copper-tube flow reactor used in the in the continuous flow [26] synthesis of rufinamide 43. Figure 10. Sonogashira cross-coupling reaction conducted in a copper-tube flow reactor.[28]

Figure 9. (a) Copper flow coil. (b) High temperature copper tube flow reactor with the top metal jacket removed (1.0 mm inner diameter tubing). (c) Heated copper tube flow reactor using a Vaportec R4 heating module. Reprinted with permission from Zhang, Y.; Jamison, T. F.; Patel, S.; Mainolfi, N., Continuous Flow Coupling and Decarboxylation Reactions Promoted by Copper Tubing. Organic Letters 2011, 13 (2), 280-283. Copyright 2011 American Chemical Society.

Figure 11. Protiodecarboxylation reactions conducted in a copper-tube flow Protiodecarboxylation reactions were also accomplished as reactor.[28] shown in Figure 11, by simply flowing the corresponding aromatic or heteroaromatic substrates 51 through copper tubing at 250 °C, without any additives, catalysts or ligands, to give Finally, Ullmann coupling products 61 – 68 presented in products 52 – 58. Whereas this reaction would pose safety Figure 12, were produced in a copper flow reactor heated to concerns in batch due to the gas evolution and use of high 150 °C without the need for additional catalysts or ligands when temperatures, the smaller amounts of gas continuously TBAA was used as a base and acetonitrile as a solvent. generated in flow and superior reaction control mitigated such Interestingly, although the conversion of 69 to 61 in flow was safety concerns. While not implemented in this example, it is 65% after a residence time of 30 min as shown in Table 2, Entry worth noting that commercially available degassing reactors 2, carrying out the analogous reaction in a microwave batch utilizing DuPont’s gas permeable Teflon AF, can remove reactor using 10 mol % of copper powder (Entry 3) or the highly gaseous by-products during flow reactions.[5b] soluble copper iodide (Entry 4), gave only 31% and 50% conversion respectively. Although further studies are warranted, the improved efficiency of the copper-tubing catalysis relative to batch copper catalysis is a significant benefit that likely stems from the enhanced mass and heat transfer offered by the flow platform, in combination with facile formation of the active catalytic species on the surface of the reactor.

PERSONAL ACCOUNT

enabling a large surface area to be exposed to the gas at any given time. Under typical flow rates used in a mesoscale reaction, friction from the sides of the reactor also create very efficient mixing, resulting in a constant supply of “gas-depleted” liquid to the liquid/gas interface. In combination with the use of higher internal pressures to aid in gas dissolution, the significantly increased interfacial contact offered in flow can dramatically improve the efficiency of gas-liquid biphasic reactions compared to traditional batch techniques. Furthermore, in situations where superstoichiometric gas is utilized, its eventual removal is straightforward using integrated vacuum degassing chambers, whereas in batch the process can be time- consuming as well as laborious. Consequently, the utilization of renewable synthetic building blocks such as molecular oxygen or carbon dioxide gas, become more attractive reagents on the quest to achieving greener chemistry.

Figure 12. Ullmann cross-coupling reactions conducted in a copper-tube flow reactor.[28]

Table 2. Comparison of batch and copper-tube flow reactors in an Ullmann coupling reaction.[28] [a] Conversions are based on 1H NMR analysis of the crude materials. [b] Reaction was run in a PFA tube.

Figure 13. Biphasic gas-liquid reactions in flow (solution blue, gas colorless). (a) distribution of gas slugs during segmented flow and (b) enhanced mixing in flow during segmented flow.

Carbon dioxide is a particularly interesting renewable Entry Method Cu Source Cu Loading Conv. (%)[a] synthetic reagent given its potential to replace toxic C1 building (mol %) blocks such as phosgene. To that end, in 2013 we demonstrated the application of carbon dioxide in the flow synthesis of [b] 1 Flow None 0 0 synthetically versatile cyclic carbonates.[29] As shown in Figure 14a, electrophilic activation of epoxide starting materials 71 by in 2 Flow 10 mL Cu tube flow Unknown 65 situ bromine, generated from N-bromosuccinimide (NBS) and reactor benzoyl peroxide (BPO), along with nucleophilic activation of carbon dioxide by DMF, enabled a range of cyclic carbonates 72 3 Microwave Cu powder 10 31 – 81 to be prepared. Although moderate yields of the products (400 W) were obtained in some cases, the reaction conditions were (Batch) relatively mild (30 min at 120 °C and a pressure of only 100 psi) compared to other published batch protocols. For example, a Microwave 4 CuI 10 50 recent approach utilizing a solid supported NHC to activate (400 W) (Batch) carbon dioxide, required more energy-intensive conditions, including significantly longer (24 - 48 h) reaction times as well as higher pressures (290 psi) to provide similar product yields [30] 2.4 Gas-Liquid Reactions: Efficient Mixing and Utilization of (Figure 14b). Renewable Gaseous Reagents. Along with cyclic carbonates, the use of carbon dioxide gas for the synthesis of asymmetric ketones 86 – 97 from One of the challenges of using gases under traditional batch organometallic reagents in flow has also been reported by our conditions is sustaining their solubility in the reaction media. In group as illustrated in Figure 15.[14b] Notably, the high interfacial continuous flow alternatively, mass flow controllers (MFCs, contact between the gas and liquid phases enhanced the Figure 1) meter small sustained doses of gases into a liquid selectivity and reactivity of the carboxylation step, and resulted stream, resulting in small “slugs” that are regularly distributed in nearly complete suppression of undesired symmetric ketones within the reactor (Figure 13a and 13b). By varying the relative typically observed in batch.[31] Controlled quenching also flow rates of the liquid and gas streams, the size and periodicity suppressed the formation of unwanted tertiary alcohol of these slugs can be modulated and controlled, thereby byproducts. Importantly, unlike typical batch protocols which use

PERSONAL ACCOUNT

high concentrations of dry ice at low temperatures,[31a, 31b] the use of a flow platform enabled greener conditions, namely room temperature and ambient pressure.

Figure 15. Synthesis of asymmetric ketones in flow.[14b]

2.5 Solvent Minimization for Waste Reduction

One of the more obvious mechanisms for designing greener synthetic processes lies in the management of waste production. While improving the atom economy of a reaction is an important aspect to this, the majority of waste generated often stems from reaction and purification solvents. Thus, one strategy to Figure 14. (a) Synthesis of cyclic carbonates from epoxides and carbon minimize waste is to conduct syntheses under highly [29] dioxide in flow. (b) Representative batch reactor synthesis of cyclic concentrated or even neat conditions. In batch however, carbonates from epoxides and carbon dioxide.[30] solvents dissipate heat,[33] which mitigates exotherm production and ensuing chemical decomposition. Fortuitously, the excellent heat transfer properties in flow described in Section 2.1, permits Overall, these two examples demonstrate the opportunity neat or highly concentrated reactions to be performed more that flow provides for the facile utilization of renewable gaseous safely.[34] The ability to achieve and maintain high pressures as feedstocks, while also improving overall reaction efficiency. well as high temperatures, also implies that hot melts produced While we have also explored the use of oxygen gas in flow as from solid reagents, starting materials or their products from a briefly mentioned in Section 2.1, tremendous opportunities exist reaction, can be flowed though the system. Added to this, is the for the application of other renewal gases including hydrogen as ability to telescope multistep sequences to avoid wasteful well as carbon monoxide as demonstrated in the literature.[32] As solvent need during isolation and purification steps. a notable side point, liquid-gaseous reactions in batch typically require investment in specialized high pressure reactors. In a A proof-of-principle demonstration of waste minimization in flow platform alternatively, no additional investment beyond the flow was the solvent-free and atom economical end-to-end standard equipment (Figure 1) is necessary. continuous flow synthesis of the antihistamine, diphenhydramine hydrochloride recently reported by our group (Figure 16). It features a nucleophilic substitution reaction followed by real-time crystallization of the API.[35] By combining the liquid chlorodiphenylmethane (98) and dimethylaminoethanol (99) in a flow reactor at 175 °C, the resulting diphenylhydramine hydrochloride (100), which has a melting point of 168 °C, was successfully flowed in the form of an ionic liquid. Direct crystallization of the API salt was then achieved through the

PERSONAL ACCOUNT

addition of a minimum amount of pre-heated (175 °C, to avoid was quenched to provide ibuprofen (101) in only 3 min and at a immediate, inline product precipitation) isopropanol anti-solvent production rate of 8.09 g h-1. (1:1, v:v), followed by cooling to 5 °C in a collection tank. In this manner, 100 was provided in 84% yield as a 13.6:1 mixture with dimethylaminoethanol hydrochloride. Compared to Rieveschl’s original batch approach[36] that is currently used on production scale, this proof-of-principle continuous flow strategy minimizes waste by avoiding the use of a base as well as a reaction solvent. Accordingly, this work serves as an excellent example of how the merits of a flow platform can be harnessed to create a more green synthetic process.

Figure 16. Flow synthesis of diphenylhydramine hydrochloride using hot melts and real-time crystallization to minimize waste production.[35]

Another well-suited illustration of the potential for waste and solvent reduction can be found in the multistep continuous flow synthesis of ibuprofen (101).[37] As presented in Figure 17, the general synthetic process consisted of 3-bond forming steps, Figure 17. Continuous flow synthesis of ibuprofen features a Friedel-Crafts one work-up and one in-line liquid-liquid separation. Notably, the acylation under neat conditions, in-line separation of product 104 in its neat initial Friedel-Crafts acylation between 102 and 103 utilized neat form and the formation of an iodine monochloride melt.[37] conditions and the reagents themselves were used to solubilize the inorganic AlCl3. After a residence time of only 1 min at 87 °C, an exothermic in-line quench of high concentrations of 2.6 Real-time Monitoring – Maintaining Optimal Conditions precipitation-prone AlCl3 was then performed using aqueous HCl. and Consistent Output Whereas a Syrris Asia pump was chemically compatible with the aggressive AlCl3, an HPLC pump was required to provide Incorporating real-time monitoring in flow chemistry enables very smooth, consistent delivery of the acidic quenching solution. tight control over chemical processes that ensure maximum Flowing through a membrane separator under high pressure efficiency and minimal waste. In batch processes, there exists a (200 psi) then facilitated the separation of the aqueous homogeneous mixture of starting materials and products that in byproducts (Al, HCl) from 104, again in the absence of any an ideal case, evolve toward complete product formation as a organic solvent. function of time (Figure 18, left). In a flow platform, the system To induce the subsequent 1,2-aryl migration step, product contains a mixture of the reaction at all stages that progresses 104 was then treated with a solution of trimethyl orthoformate in as a function of distance travelled through the system (Figure 18, DMF, followed by addition of neat iodine monochloride, which is right). The small volume of material present at any given stage commercially supplied as a liquid, solid or mixture of both of a flow process enables a very rapid response to changes in depending on the polymorph composition. Since the melting parameters, which can be used to keep the reaction within point of this reagent is low (ranges from 3.9 °C to 27.2 °C), the desired tolerances. More specifically, given that only a small need for solvent was avoided in the case of a solid, by simply volume of material is present at a given stage and time, any heating to 35 °C. A Syrris pump with a re-designed pump valve material that happens to deviate from the required tolerances, was necessary to avoid corroding by the iodine monochloride. can be easily identified using real-time monitoring and diverted After a residence time of only 1 min at 90 °C, the resulting out of the process, thereby maintaining the integrity of the larger methyl ester was then saponified and the iodine monochloride batch.

PERSONAL ACCOUNT

Additionally, the flow rate of the dilution solvent (EtOAc) and the material exiting the dilution tank was adjusted in real-time to ensure constant concentration (and stoichiometry) of the boc- protected amine. By detecting and eliminating disturbances in real-time, the outcome of the integrated process (yield, purity) was maintained over extended periods with a very high degree of reproducibility.

Figure 18. Simplified representation comparing reaction progress in batch 3. Summary and Outlook reactors versus flow reactors. As illustrated herein, flow chemistry represents an enabling technology with the ability to address a number of challenges set Many sensors have been developed for monitoring reaction forth by the Principles of Green Chemistry.[1] The benefits conditions in flow such as the temperature, pressure, and flow provided in a flow platform are broad and include enhanced rate. By coupling sensors to appropriate software, it is possible mixing within biphasic systems, ability to superheat reactions in to achieve autonomous real-time maintenance of reaction a safe manner, efficient irradiation, facile integration of catalytic [38] conditions. A range of detectors have also been used for processes, and in-line monitoring as a means to maximize [39] monitoring the progress of flow reactions including flow-IR, reaction product integrity. In harnessing these benefits to our [40] [41] [42] flow-UV, flow-Raman, and flow NMR. Additionally, advantage, we have been able to carry out numerous syntheses [43] analytic techniques such as liquid and gas chromatography that have common themes of waste reduction and improved [44] and ESI mass spectroscopy are inherently flow-techniques reaction efficiency (e.g. atom economy, product yields and and have also been interfaced to flow systems to provide real- reaction rates) relative to batch. It should be cautioned however, time reaction monitoring (Figure 19). that assessing the overall efficiency of a particular flow process invariably requires taking many detailed factors into account that have been beyond our scope.[3d, 46] One noteworthy example of such a rigorous analysis on a major manufacturing process, was recently reported by Kralisch and co-workers. Their studies indicated that converting the production of epoxidized soybean oil (a plasticizer in polyvinyl chloride) from conventional batch reactors (240,000 tonnes/yr) to a flow process, would have a meaningful impact on lowering overall global energy demands.[47]

In terms of the equipment commercially available for various unit operations in flow, many advancements have been made in recent years, such as the development of liquid-liquid separators.[48] Other tools such as the rotary evaporator or simple processes like product drying post-separation, however, have yet to be developed. Moreover, advances in pumping technology including chemical inertness, improvements in Figure 19. Representative images of common sensors and detectors that delivery rates and pressures, and the ability to better enable real-time monitoring of reactions in continuous-flow. accommodate solid formation would be of great value. As the technology advances along these lines and the flow toolbox Our collaborative efforts in the end-to-end continuous rounds out with increasingly financially accessible devices, so manufacturing of pharmaceuticals that meet USP (US too will the scope and ease of flow processes. Accordingly, its Pharmacopeial Convention) standards exemplifies the utility of application and synergy with the ideals of green chemistry are real-time monitoring.[45] For instance, several days into running expected to follow suit. the integrated synthesis of aliskiren hemifumarate (107), there was a disturbance in the volume maintained in the crystallization tank following the initial aminolysis reaction (Figure 20). While Acknowledgement the volume of solution in this tank was not critical, a constant concentration of the boc-protected amine in the subsequent We thank the Novartis-MIT Center for Continuous Manufacturing, dilution tank was essential for maintaining the stoichiometry of the Bill and Melinda Gates Foundation, and the Defense the ensuing deprotection reaction with HCl. Thus, to re-establish Advanced Research Project Agency (DARPA) for their generous the desired holding volume in the crystallization tank, the system financial support over the years. J.A.M.L thanks NSERC Canada controls adjusted the flow rate of material exiting this tank. for his postdoctoral fellowship.

PERSONAL ACCOUNT

Figure 20. Disturbance mitigation in the end-to-end synthesis of aliskiren enabled by real-time monitoring.[45b]

Keywords: Flow chemistry • Green chemistry • Real-time [4] a) F. Darvas, G. Dorman, in Flow Chemistry (Eds.: F. Darvas, G. monitoring • Process intensification • Photochemistry Dorman, V. Hessel), de Gruyter, Boston, 2014, pp. 9–58. ; b) M. Fekete, T. Glasnov, in Flow Chemistry (Eds.: F. Darvas, G. Dorman, V. Hessel), de Gruyter, Boston, 2014, pp. 95-140. ; c) J. H. Bannock, S. H. Krishnadasan, M. Heeney, J. C. de Mello, Materials Horizons 2014, 1, References 373-378. [5] a) P. Plouffe, A. Macchi, D. M. Roberge, Organic Process Research & [1] P. Anastas, J. Warner, Green Chemistry: Theory and Practice, Oxford Development 2014, 18, 1286-1294; b) M. Brzozowski, M. O'Brien, S. V. University Press, New York, 1998. Ley, A. Polyzos, Accounts of Chemical Research 2015, 48, 349-362; c) [2] a) R. A. Sheldon, Green Chemistry 2007, 9, 1273-1283; b) G. J. Ruiz- B. J. Deadman, S. G. Collins, A. R. Maguire, Chemistry - A European Mercado, A. Carvalho, H. Cabezas, ACS Sustainable Chemistry & Journal 2015, 21, 2298-2308; d) K. Gilmore, P. H. Seeberger, Chemical Engineering 2016, 4, 6208-6221; c) J. Andraos, ACS Sustainable Record 2014, 14, 410-418; e) R. Porta, M. Benaglia, A. Puglisi, Organic Chemistry & Engineering 2016, 4, 1917-1933; d) J. Auge, M.-C. Process Research & Development 2016, 20, 2-25; f) J.-i. Yoshida, Y. Scherrmann, New Journal of Chemistry 2012, 36, 1091-1098; e) D. J. C. Takahashi, A. Nagaki, Chem. Commun. 2013, 49, 9896-9904. Constable, C. Jimenez-Gonzalez, in Handbook of Green Chemistry [6] a) Z. J. Garlets, J. D. Nguyen, C. R. J. Stephenson, Israel Journal of (Ed.: P. T. Anastas), Wiley, Hoboken, 2012, (Vol. 7) pp. 35-67; f) C. Chemistry 2014, 54, 351-360; b) B. D. A. Hook, W. Dohle, P. R. Hirst, Jimenez-Gonzalez, D. J. C. Constable, C. S. Ponder, Chemical Society M. Pickworth, M. B. Berry, K. I. Booker-Milburn, Journal of Organic Reviews 2012, 41, 1485-1498; g) R. A. Sheldon, J. P. M. Sanders, Chemistry 2005, 70, 7558-7564. Catalysis Today 2015, 239, 3-6; h) M. Tobiszewski, M. Marc, A. [7] P. D. Morse, R. L. Beingessner, T. F. Jamison, Israel Journal of Galuszka, J. Namiesnik, Molecules 2015, 20, 10928-10946. Chemistry 2016, Ahead of Print. [3] a) S. V. Ley, Chemical Record 2012, 12, 378-390; b) L. Vaccaro, D. [8] a) T. Razzaq, C. O. Kappe, Chemistry - An Asian Journal 2010, 5, Lanari, A. Marrocchi, G. Strappaveccia, Green Chemistry 2014, 16, 1274-1289; b) V. Hessel, Chemical Engineering & Technology 2009, 32, 3680-3704; c) V. Hessel, D. Kralisch, N. Kockmann, T. Noel, Q. Wang, 1655-1681. ChemSusChem 2013, 6, 746-789; d) S. G. Newman, K. F. Jensen, [9] a) S. Monfette, M. Eyholzer, D. M. Roberge, D. E. Fogg, Chemistry - A Green Chemistry 2013, 15, 1456-1472; e) S. Kobayashi, Chemistry - European Journal 2010, 16, 11720-11725; b) R. Dach, J. J. Song, F. An Asian Journal 2016, 11, 425-436; f) N. E. Leadbeater, in New and Roschangar, W. Samstag, C. H. Senanayake, Organic Process Future Developments in Catalysis: Catalysis for Remediation and Research & Development 2012, 16, 1697-1706. Environmental Concerns (Ed. S. L. Suib), Elsevier, Waltham, 2013, pp. [10] M. W. Bedore, N. Zaborenko, K. F. Jensen, T. F. Jamison, Organic 19-39., 2013; g) C. Wiles, P. Watts, Green Chemistry 2012, 14, 38-54. Process Research & Development 2010, 14, 432-440.

PERSONAL ACCOUNT [11] a) B. Cuenoud, I. Bruce, R. A. Fairhurst, D. Beattie, Novartis, AG, Basel, [39] a) C. F. Carter, H. Lange, S. V. Ley, I. R. Baxendale, B. Wittkamp, J. G. Switzerland, W.O. Patent 0075114, 2000; b) O. Lohse, C. Vogel, Goode, N. L. Gaunt, Organic Process Research & Development 2010, Novartis, AG, Basel, Switzerland, W.O. Patent 076422, 2004; c) R. G. 14, 393-404; b) J. S. Moore, K. F. Jensen, Organic Process Research Sturton, A. Trifilieff, A. G. Nicholson, P. J. Barnes, Journal of & Development 2012, 16, 1409-1415; c) C. F. Carter, I. R. Baxendale, Pharmacology and Experimental Therapeutics 2008, 324, 270-275. M. O'Brien, J. B. J. Pavey, S. V. Ley, Organic & Biomolecular [12] J. K. Mehra, A. Choubey, B. K. Srivastava, R. K. Porwal, P. Gautam, Chemistry 2009, 7, 4594-4597. U.S. Patent 2005/0107635, 2005. [40] a) W. Ferstl, T. Klahn, W. Schweikert, G. Billeb, M. Schwarzer, S. [13] Z. He, T. F. Jamison, Angewandte Chemie, International Edition 2014, Loebbecke, Chemical Engineering & Technology 2007, 30, 370-378; b) 53, 3353-3357. F. Benito-Lopez, W. Verboom, M. Kakuta, J. G. E. Gardeniers, R. J. M. [14] a) A. B. Leduc, T. F. Jamison, Organic Process Research & Egberink, E. R. Oosterbroek, A. van den Berg, D. N. Reinhoudt, Chem. Development 2012, 16, 1082-1089; b) J. Wu, X. Yang, Z. He, X. Mao, T. Commun. 2005, 2857-2859. A. Hatton, T. F. Jamison, Angewandte Chemie, International Edition [41] a) T. A. Hamlin, N. E. Leadbeater, Beilstein J Org Chem 2013, 9, 1843- 2014, 53, 8416-8420. 1852; b) S. Mozharov, A. Nordon, D. Littlejohn, C. Wiles, P. Watts, P. [15] C. Jimenez-Gonzalez, D. J. C. Constable, Green chemistry and Dallin, J. M. Girkin, Journal of the American Chemical Society 2011, engineering : a practical design approach, Wiley, Hoboken, 2014. 133, 3601-3608. [16] a) N. Hoffmann, Chem. Rev. 2008, 108, 1052-1103; b) T. Bach, J. P. [42] a) V. Sans, L. Porwol, V. Dragone, L. Cronin, Chemical Science 2015, Hehn, Angewandte Chemie, International Edition 2011, 50, 1000-1045; 6, 1258-1264; b) A. M. R. Hall, J. C. Chouler, A. Codina, P. T. Gierth, J. c) A. Albini, M. Fagnoni, Handbook of Synthetic Photochemistry, Wiley, P. Lowe, U. Hintermair, Catalysis Science & Technology 2016, 6, 8406- Hoboken, 2010; d) E. E. Coyle, M. Oelgemoller, Photochemical & 8417. Photobiological Sciences 2008, 7, 1313-1322; e) C. K. Prier, D. A. [43] D. C. Fabry, E. Sugiono, M. Rueping, Reaction Chemistry & Rankic, D. W. C. MacMillan, Chem. Rev. 2013, 113, 5322-5363. Engineering 2016, 1, 129-133. [17] R. Telmesani, S. H. Park, T. Lynch-Colameta, A. B. Beeler, [44] a) S. Koster, E. Verpoorte, Lab on a Chip 2007, 7, 1394-1412; b) J. S. Angewandte Chemie, International Edition 2015, 54, 11521-11525. Mathieson, M. H. Rosnes, V. Sans, P. J. Kitson, L. Cronin, Beilstein [18] J. W. Tucker, Y. Zhang, T. F. Jamison, C. R. J. Stephenson, Journal of Nanotechnology 2013, 4, 285-291. Angewandte Chemie, International Edition 2012, 51, 4144-4147. [45] a) A. Adamo, R. L. Beingessner, M. Behnam, J. Chen, T. F. Jamison, K. [19] a) J. P. Knowles, L. D. Elliott, K. I. Booker-Milburn, Beilstein Journal of F. Jensen, J.-C. M. Monbaliu, A. S. Myerson, E. M. Revalor, D. R. Organic Chemistry 2012, 8, 2025-2052; b) While the extent to which a Snead, T. Stelzer, N. Weeranoppanant, S. Y. Wong, P. Zhang, Science reaction can be concentrated without sacrificing yield will be limited by 2016, 352, 61-67; b) S. Mascia, P. L. Heider, H. Zhang, R. Lakerveld, B. its intermolecular side reactions, the increased homogeneity of Benyahia, P. I. Barton, R. D. Braatz, C. L. Cooney, J. M. B. Evans, T. F. irradiation within flow reactors, discussed in references 16d and 19a, Jamison, K. F. Jensen, A. S. Myerson, B. L. Trout, Angewandte suggests a benefit. Chemie, International Edition 2013, 52, 12359-12363. [20] A. C. Gutierrez, T. F. Jamison, Journal of Flow Chemistry 2011, 1, 24- [46] a) C. Jimenez-Gonzalez, P. Poechlauer, Q. B. Broxterman, B.-S. Yang, 27. D. am Ende, J. Baird, C. Bertsch, R. E. Hannah, P. Dell'Orco, H. [21] a) A. El-Faham, F. Albericio, Chem. Rev. 2011, 111, 6557-6602; b) K. Noorman, S. Yee, R. Reintjens, A. Wells, V. Massonneau, J. Manley, E. Schwieter, J. N. Johnston, Journal of the American Chemical Society Organic Process Research & Development 2011, 15, 900-911; b) T. 2016, 138, 14160-14169. Anastas Paul, B. Zimmerman Julie, Environmental science & [22] Y. Zhang, M. L. Blackman, A. B. Leduc, T. F. Jamison, Angewandte technology 2003, 37, 94A-101A. Chemie, International Edition 2013, 52, 4251-4255. [47] D. Kralisch, I. Streckmann, D. Ott, U. Krtschil, E. Santacesaria, M. Di [23] B. M. Trost, C. M. Older, Organometallics 2002, 21, 2544-2546. Serio, V. Russo, L. De Carlo, W. Linhart, E. Christian, B. Cortese, M. H. [24] A. R. Bogdan, N. W. Sach, Advanced Synthesis & Catalysis 2009, 351, J. M. de Croon, V. Hessel, ChemSusChem 2012, 5, 300-311. 849-854. [48] a) J. G. Kralj, H. R. Sahoo, K. F. Jensen, Lab on a Chip 2007, 7, 256- [25] a) A. R. Bogdan, K. James, Chemistry - A European Journal 2010, 16, 263; b) A. Adamo, P. L. Heider, N. Weeranoppanant, K. F. Jensen, 14506-14512; b) A. R. Bogdan, K. James, Organic Letters 2011, 13, Industrial & Engineering Chemistry Research 2013, 52, 10802-10808; 4060-4063. c) M. O'Brien, P. Koos, D. L. Browne, S. V. Ley, Organic & [26] P. Zhang, M. G. Russell, T. F. Jamison, Organic Process Research & Biomolecular Chemistry 2012, 10, 7031-7036; d) M. O'Brien, D. Cooper, Development 2014, 18, 1567-1570. Synlett 2016, 27, 164-168; e) M. O'Brien, D. A. Cooper, P. Mhembere, [27] R. Portmann, Novartis, AG, Basel, Switzerland, U.S. Patent 6,156,907, Tetrahedron Letters 2016, 57, 5188-5191. 2000. [28] Y. Zhang, T. F. Jamison, S. Patel, N. Mainolfi, Organic Letters 2011, 13, 280-283. [29] J. A. Kozak, J. Wu, X. Su, F. Simeon, T. A. Hatton, T. F. Jamison, Journal of the American Chemical Society 2013, 135, 18497-18501. [30] H. Zhou, Y.-M. Wang, W.-Z. Zhang, J.-P. Qu, X.-B. Lu, Green Chemistry 2011, 13, 644-650. [31] a) H. Gilman, P. R. Van Ess, Journal of the American Chemical Society 1933, 55, 1258-1261; b) E. J. Soloski, C. Tamborski, Journal of Organometallic Chemistry 1978, 157, 373-377; c) J. C. Anderson, S. Broughton, Synthesis 2001, 2379-2380. [32] C. J. Mallia, I. R. Baxendale, Organic Process Research & Development 2016, 20, 327-360. [33] W. Sabel, Basic Techniques of Preparative Organic Chemistry, Pergamon Press, New York, 1967. [34] A. Weiler, Chem. Ind. 2013, 77, 20-23. [35] D. R. Snead, T. F. Jamison, Chemical Science 2013, 4, 2822-2827. [36] G. Rieveschl, Jr., Parke, Davis & Co., U.S. Patent 2421714, 1947. [37] D. R. Snead, T. F. Jamison, Angewandte Chemie, International Edition 2015, 54, 983-987. [38] B. J. Reizman, K. F. Jensen, Accounts of Chemical Research 2016, 49, 1786-1796.

PERSONAL ACCOUNT

Entry for the Table of Contents

PERSONAL ACCOUNT

Intersection of Green and Flow Justin A.M. Lummiss, Peter D. Morse, Chemistry. The principles of green Rachel L. Beingessner, Timothy F. chemistry identify the major Jamison* challenges in achieving efficient, sustainable, and safe chemical Page No. – Page No. transformations. Continuous flow Towards More Efficient, Greener platforms represent an enabling Synthesis Through Flow Chemistry technology for improving reaction efficiency and reducing waste production. The intersection of these fields offers exciting possibilities for improved chemical synthesis.