Get 20M+ Full-Text Papers For Less Than $1.50/day. Start a 14-Day Trial for You or Your Team.

Learn More →

Oscillatory flow reactors for synthetic chemistry applications

Oscillatory flow reactors for synthetic chemistry applications Oscillatory flow reactors (OFRs) superimpose an oscillatory flow to the net movement through a flow reactor. OFRs have been engineered to enable improved mixing, excellent heat- and mass transfer and good plug flow character under a broad range of operating conditions. Such features render these reactors appealing, since they are suitable for reactions that require long residence times, improved mass transfer (such as in biphasic liquid-liquid systems) or to homogeneously suspend solid particles. Various OFR configurations, offering specific features, have been developed over the past two decades, with significant progress still being made. This review outlines the principles and recent advances in OFR technology and overviews the synthetic applications of OFRs for liquid-liquid and solid-liquid biphasic systems. . . . . . Keywords Oscillatory flow Multiphasic reactions Solid handling Liquid-liquid reactions Process intensification Plug flow reactors Introduction reactors, whilst maintaining plug flow behavior, intense mixing, and excellent heat transfer. Importantly, oscillation- Continuous flow technology has attracted the attention of promoted mixing is independent of the net flow rate, opening chemists in recent years, specifically in scaling up hazardous the window of operation to lower flow rates (longer residence chemistries, photochemistry, electrochemistry and many times), which would otherwise impart poor hydrodynamic others [1–4]. The use of various auxiliary and automation properties. Another consequence of the ability to operate at technologies has strengthened the potential of flow reactors low flow rates is the allowable decrease in reactor volume, [5–7] for a wide range of applications [8–11]. Despite the enabling compact setups. In addition, the quality of mixing benefits of continuous flow technology, flow reactors face can be ensured during scale-up when OFR geometric ratios some inherent problems, particularly when dealing with mul- and fluidic characteristics are maintained [13]. tiphasic systems. Rational engineering has, however, begun to A small number of reviews concerning OFRs do already tackle these issues, by developing a variety of reactor config- exist, but these have focused specifically on crystallization urations with specific features aimed at processing challeng- [14], biological processes [15] or engineering developments ing reaction media on both small and large scale [12]. [16, 17]. Here, we give an overview of the technological back- One particularly promising development in reactor technol- ground and potential of this reactor type, aimed towards the ogy is that of oscillatory flow reactors (OFRs), whereby a synthetically-oriented flow chemist. symmetrical oscillation is superimposed to the net flow Since the efficiency of flow reactors has already been well through the reactor. This has been found to offer a solution explored for homogeneous reaction media, only a few reports to multiple problems encountered by conventional flow examine these monophasic reactions in OFRs [18–24]. Furthermore, despite numerous studies concerning gas-liquid mass transfer in OFRs [25–27], few articles have been dedi- * C. Oliver Kappe cated to synthetic applications with gases, likely due to their oliver.kappe@uni-graz.at compressibility, which can dampen oscillations [28]. On the other hand, the high mass-transfer and continuous agitation Institute of Chemistry, University of Graz, NAWI Graz, capability of these reactors have already begun to see signifi- Heinrichstrasse 28, 8010 Graz, Austria cant uptake for biphasic liquid-liquid and solid-liquid reac- Center for Continuous Flow Synthesis and Processing (CC FLOW), tions. Consequently, this review will focus on these two clas- Research Center Pharmaceutical, Engineering (RCPE), Inffeldgasse 13, 8010 Graz, Austria ses of biphasic systems. First, the general principles and 476 J Flow Chem (2020) 10:475–490 advantages of OFRs will be presented for each, followed by turbulent flow regime (Fig. 1b). However, this can be obtained an overview of experimental reports. at much lower Re in an OFR, due to the additional oscillatory turbulence, quantified by the oscillatory Reynolds number (Re )[14, 29]. This enables a drastic reduction of the reactor length. The plug flow behavior related to vortex cycles can be Oscillatory flow reactor principles achieved at velocity ratios (ψ) between 2 and 10, depending and equipment on the baffle geometry. Usually, Re and St values larger than o r 100 and 0.5, respectively, are required to fully realize the Oscillation can be applied to a flow reactor using various benefits of OFRs.[14]. oscillatory devices, such as piston- bellow-, diaphragm-, Residence time distribution (RTD), the distribution of time syringe- or peristaltic pumps. The symmetrical oscillation spent by the fluid within a flow reactor, is an important mea- generates vortices, leading to improved radial mixing, whilst sure of the reactor’s plug flow character. A narrow distribution aiming to maintain plug flow character (minimal axial mixing, decreases the probability of side reactions or incomplete reac- Fig. 1a). In this way, mixing is decoupled from the net flow tion to occur (due to portions of material being present in the and depends only on the oscillation conditions and the OFR reactor for too long, or not long enough). RTD is related to the configuration. Therefore, reactions requiring long residence axial dispersion coefficient, which is inversely proportional to times can be performed efficiently without requiring large the Peclet or Bodenstein number (Table 1)[30]. RTD exper- reactor volumes. Accordingly, the numerous issues facing iments are usually performed with a colored tracer and the large volume tubular flow reactors can be obviated (e.g., high dimensionless RTD parameter, E(θ), is plotted as a function pressure drop, large footprint and poor residence time of the normalized time, θ (Fig. 1c). When plug flow is distribution). achieved, a narrow symmetric curve is centered at θ =1 OFRs can be characterized by different dimensionless (Fig. 1c, green curve). Any asymmetrical tendencies or cen- numbers (summarized in Table 1), where the maximum oscil- tering at another θ suggests axial dispersion (Fig. 1c,red latory velocity is considered. Standard flow reactors require a curve). net flow Reynolds number (Re ) above 4000 to achieve a Axial dispersion is less sensitive to oscillatory conditions at high net flow rates, since laminar flow is minimized. Conversely, experiments with low Re (generally with long residence times) require careful optimization of the oscillatory radial mixing axial mixing amplitude and frequency to obtain a narrow RTD. Oscillation amplitude has a larger impact than frequency on the axial dispersion. At low Re numbers, a higher oscillation ampli- tude decreases the axial dispersion, thanks to the creation of vortices. After reaching a critical Re value, the opposite trend appears: the axial dispersion begins to increase, due to over propagation of vortices [31–34]. The optimal Re value de- Laminar flow Turbulent flow pends on the net flow rate, such that the velocity ratio (ψ)still (Low Re) (High Re) provides good plug flow character. This value must therefore be optimized for each individual reactor system. The most common way of generating vortices is the addi- tion of baffles within the OFR. Various baffle geometries exist and offer different degrees of mixing (Fig. 2). At each colli- sion with an obstacle, the void behind the baffle creates a backward movement of the fluid and a short reverse flow appears. These short-lived vortices at each baffle intersection are called eddies (Fig. 3a). In continuous oscillatory baffled reactors (COBRs), oscillations lead to the generation and dis- appearance of eddies and therefore to homogeneous radial mixing. Fig. 1 Summary of mixing and flow regimes in flow reactors: a Radial The complex flow behavior in COBRs with different baffle versus axial mixing in a tube. b Laminar versus turbulent flow regimes. c geometries has been rigorously studied experimentally and Depiction of a reactor’s plug flow character, estimated using tracer theoretically [35–43]. However, it is still difficult to clearly injection (blue) experiments, where good plug flow character has low quantify the impact of the geometry on reactor performance, axial dispersion (green), compared to poor plug flow character, with significant axial dispersion (red) since it depends on multiple parameters. It appears, though, J Flow Chem (2020) 10:475–490 477 Table 1 Dimensionless numbers for characterizing oscillatory flow reactors Dimensionless numbers Formula Interpretation υρD Net flow Reynolds number Re ¼ Measure of the mixing intensity in net flow xo2πfρD Oscillatory Reynolds number Re ¼ Measure of the mixing intensity in oscillatory flow Reo Velocity ratio ψ ¼ Re Measure of the dominance of the oscillatory flow over the net flow Strouhal number St ¼ Degree of eddy propagation in a COBR (Fig. 3a) 4πxqffiffiffiffiffiffi Dean number De ¼ Re Likelihood of Dean vortices in a coiled baffleless OFR (Fig. 3b) 2Rc υL Peclet or Bodenstein number Bo ¼ Pe ¼ Degree of axial back-mixing a −3 −1 μ = viscosity (Pa.s), ρ =density (kg m ), υ = superficial fluid velocity (m s ), D = tube diameter (m), f = frequency of oscillation (Hz), x = center-to- −2 −1 peak amplitude (m), R = radius of curvature (m), D = axial dispersion coefficient, L = tube length (m), h = heat transfer coefficient (W m K ), k = c a t −1 −1 thermal conductivity (W m K ) that helical baffles (Fig. 2c-e) have a wide plug flow operating popularity of split-and-recombine type mixers in microfluidic range, because of additional spiraling character to the net flow applications [45, 46]. [36, 39, 40]. Disc and doughnut baffles (Fig. 2i) are also Efficient multiphasic mass transfer is another strong asset proposed to achieve efficient mixing at low oscillatory flow of OFRs. Studies, principally focused on gas-liquid systems in [35]. flow, have shown that the mass transfer coefficients for gas- The second method for vortex generation is the use of coiled liquid and liquid-liquid systems are generally higher than in reactors instead of straight tubes [44]. A secondary flow, named conventional flow reactors, and remain high for a wider range Dean flow, appears due to the centripetal force arising at the of conditions, especially at low flow rates [27, 47]. Stronger curvature of the tube. The fluid near the inner wall is pushed oscillation (high Re ) reduces the droplet or bubble size (until outwards, towards the center of the tube, while an opposing a minimal value), which increases the interfacial area. current appears from the outside to the center of the coil. This Frequency and amplitude contribute with the same weight to superimposed flow is therefore seen as a pair of counter-rotating this enhancement [27]. vortices (Fig. 3b). In coiled baffleless oscillatory flow coil reac- Similarly, heat transfer is enhanced in OFRs compared to tors, these Dean vortices are formed and collapsed continuously conventional tubes, especially at low flow rates, thanks to during oscillations, reducing stagnation zones at their center and improved radial mixing. This enhancement depends strongly improving radial mixing [29]. This behavior can be quantified by on Re ,but noton the St value. The heat transfer coefficient o r the Dean number (De,Table 1). increases with the oscillatory conditions and reaches a plateau A third oscillatory mixing principle is the combination with at Re ≈ 1300 [48–50]. When considering heat transfer, oscil- a split-and-recombine mixing structure (Fig. 3c). By applying lation frequency has a higher impact than amplitude. A recent oscillation, each of these elements will be effectively revisited study showed that helical COBRs (Fig. 2c-e) weremoreeffi- numerous times by the fluid, multiplying its effect. cient than central axial COBRs (Fig. 2b), followed by single Furthermore, these split channels have a smaller diameter, so orifice COBRs (Fig. 2g) in terms of thermal performance a higher velocity (thus, higher Re ) can be reached with a (considering both heat transfer and pressure drop) [51]. given oscillatory amplitude. This principle has received sur- prisingly little attention thus far, particularly considering the Fig. 3 Vortices generated in oscillatory flow reactors leading to improved Fig. 2 Configurations of oscillatory baffled reactors. (a) integral baffles radial mixing during oscillations. (a) Eddies in a continuous oscillatory with periodic constrictions than can be sharp or smooth (b) central axial baffled reactor (COBR) (b) Dean vortices in a baffleless OFR. (c) baffles (c) round-edged helical baffles (d) sharp-edged helical baffles (e) Oscillation-enhanced split-and-recombine mixing. The red arrows indi- sharp-edged helical baffles with a central rod (f) wire wool baffles (g) cate net flow, whilst blue arrows illustrate the mixing instilled by single-orifice baffles (h) multi-orifice baffles (i) disc and doughnut baffles oscillations 478 J Flow Chem (2020) 10:475–490 Another advantage of OFRs is their ease of scale-up. By recently that other biphasic chemical reactions have been re- keeping the dynamic parameters (Re , Re and St ) and geo- ported using OFRs. n o r metric features of the reactor (baffle spacing and open area) The implementation of biphasic reactions in COBRs was constant, similar axial dispersion and mixing behavior can be pioneered by the group of Harvey. In 2001, in cooperation obtained with changing reactor volumes [13, 52]. Finally, with a sterol producer, they evaluated the batch-to-flow tran- OFRs demonstrate better performance than stirred tank reac- sition of steryl ester saponification in a meso-COBR [59]. A tors at similar power dissipation values. Power dissipation has pair of stainless-steel pistons provided oscillations within the been estimated using simulations based on two main models, 1.3 L flow reactor. Thanks to a flow conversion model, based namely the quasi-steady flow and the eddy acoustic model on fluid mechanics and reaction kinetics, they could lower the [53, 54]. reaction temperature and improve the sterol yield compared to An important factor to consider when dealing with oscilla- a typical batch procedure. The quality of mixing and mass tions is their dampening with distance from the oscillatory transfer was high, enabling an efficient biphasic reaction be- pump. The formation of air bubbles within the reactor also tween the steryl esters and ethanol. The optimal conditions has a dampening effect that could be problematic for mass found with this “tanks-in-series” modelling approach led to transfer or for the handling of solids. Pulsations are potentially better selectivity for the desired sterol, despite the complex unsuitable for reactions involving gases since they are com- mixture at the beginning. The lab-scale tests resulted in a −1 pressible. Consequently, it recommended to ensure that the production of 5 kg h , while a pilot-scale installation in the −1 OFR is airtight and to consider the strength of the oscillation production site led to generation of sterols at about 20 kg h . at the reactor outlet, prior to experimentation [52]. These results were expanded upon a year later when the Nevertheless, OFRs have mostly been assessed for reac- same group proposed a methodology for designing tions that are inefficient in conventional flow reactors. This production-scale OFRs [13]. By using Re and Re values as o n explains why numerous articles relying on biphasic media and the starting point for the design, different parameters could be solid handling have been reported. In the following sections, determined, such as the power density and COBR geometry. these applications will be discussed in detail. Two experimental cases were then assessed. The first of these cases was dedicated to the addition of an end-group, such as anhydride, to a polymer, forming a succinimide moiety [13]. The aim of this study was to obtain Biphasic applications of oscillatory flow −1 a productivity of 2 t h by operating at 200 °C with a residence reactors time of 4 h. Such a goal was seen as impractical with a simple tubular reactor because it would require a very long reactor Liquid-liquid systems: bulk chemical applications length and high power density. These requirements were, how- ever, achieved experimentally with a compact industrial COBR Experiments examining liquid-liquid systems in OFRs ap- placed horizontally. The second case concerned the saponifica- peared at roughly the same time as their oscillatory batch tion of steryl ester [13]. Based on the previous study [59], a reactor equivalent. In the late 1990s, Ni et al. studied the −1 productivity of 50 t day at 85 °C and with a residence of correlation of polymer particle size with droplet size during 20 min was required. A much smaller COBR configuration batch methylmethacrylate polymerization [55–58]. In the than in the first case was used, because this reaction involved meantime, investigations into biodiesel production in OFRs a shorter residence time and less viscous reagents. took place over the course of about 15 years (Fig. 4). OFRs are This reactor design methodology was also used in 2008 by good candidates for this reaction, which requires high mass Mohd et al., showing that COBRs are more powerful than transfer due to the immiscibility and difference in viscosity standard plug flow reactors (PFRs) for biodiesel production between triglycerides and methanol. It is only relatively [60]. Their simulation showed that, for a fixed reactor diame- ter, COBRs give higher conversion at reduced residence time than PFRs and require a lower power density. Harvey et al. also investigated biodiesel production [61]. Their preliminary study demonstrated the potential of OFRs for this goal [62]. Then, in 2007, they showed that high oscil- lation frequencies were required to produce biodiesel that ful- fills EU standards. Small differences between batch oscillato- ry baffled reactors and standard batch reactors were demon- strated. However, operating under flow conditions gave better quality biodiesel, in terms of triglyceride and diglyceride con- Fig. 4 General procedure for biodiesel production in a COBR centrations, than with a batch oscillatory baffled reactor. J Flow Chem (2020) 10:475–490 479 In 2011, Phan et al. investigated the potential of continuous Corning reactor benefited from a wider operating window screening for biodiesel production with newly designed (higher pressure and temperature) compared to the NiTech COBRs that allow plug flow at low Re [63]. Trials were first COBR, hence speeding up the reaction further. However, re- conducted at low flow rates in a meso-COBR with central action equilibrium was never achieved in the limited residence sharp-edged baffles (Fig. 2b). Long experiments proved the time allowed by this reactor. Furthermore, clogging issues production stability over time, with small deviations that did occurred after long runs with high levels of free fatty acid not depend on the oscillatory frequency. However, mixing content. On the contrary, the reaction equilibrium could be was not sufficient and so homogeneity could not be rapidly obtained even at lower temperatures in the NiTech established. COBR, which also showed higher flexibility in the choice of The baffles were replaced with helical round wire baffles the residence time since mixing was independent from the net (Fig. 2c), but insufficient mixing led to stratification of the two flow. phases and large fluctuations in the biodiesel yield. Therefore, A comparison between two OFRs was also performed by new baffle designs were developed to increase shear stress, the same group for the esterification of fatty acids from waste resulting in the use of axially oriented sharp-edged helical cooking oil. In this instance, the transesterification was done baffles with or without a central rod (Fig. 2d–e). Slug flows with glycerol to produce glycerides, which have applications were less prevalent with this baffle geometry, but flow in the cosmetics and pharmaceutical fields [71]. Contrary to channeling was still observed in the absence of the central the previous report [70], the two OFRs: a COBR and a rod. Steady-state screening was then implemented with differ- baffleless OFR (“pulsed helix reactor”), were made of stain- ent methanol:triglyceride ratios by using the oscillatory reac- less steel to support higher temperatures and pressures. A tor at low Re . Then, dynamic screening (continuous variation pulsation dampener and a back pressure regulator (BPR) of one parameter [64–66]) was assessed and results between were employed at the reactor outlets to prevent cavitation the two screening modes were found to be in agreement. or solvent vaporization at these elevated temperatures. The same year, Harvey et al. also tested three meso- Higher conversions were obtained with these reactors than COBRs with different baffle geometries for biodiesel produc- with reactors operated at lower temperatures (a NiTech glass tion [67]. Depending on the configuration of the baffles, COBR and a batch reactor). The COBR showed better per- steady state was achieved after different startup times:1.5, formance than the baffleless OFR, which the authors attrib- 2.5, and 4.0 residence times for the integral, wire wool, and uted to the generation of smaller droplets (and higher inter- helical baffle designs (Fig. 2a, f, c), respectively. In each case, facial area) due to the baffles. However, higher conversion this startup time was shorter than that required for a single was obtained in the baffleless OFR when the acid catalyst continuous stirred tank reactor. Integral baffles gave the was changed. Different experimental parameters were stud- highest content of biodiesel at high Re . Moreover, stable ied and led to higher selectivity than those previously report- steady state was only established at high oscillatory frequen- ed in the literature. cies, especially for integral baffles, otherwise mixing was in- In 2015, Lobry et al. studied liquid-liquid dispersion in a sufficient to avoid stratification between the immiscible COBR for polymerization applications [72]. A first reactor phases. The startup time decreased with increasing Re num- configuration was developed for the liquid-liquid dispersion ber - likely due to better mixing efficiency. The decrease in study. It appeared that the dispersion properties were not af- yield at longer residence times was related to saponification of fected by the net flow rate, enabling dispersion-independent the biodiesel. A short residence time (5–10 min) was thus control of the residence time. Moreover, oscillation ampli- required to achieve optimal biodiesel yield. tudes and frequencies were the main parameters responsible The same group has also recently published two other stud- for droplet breakage. A correlation between the droplet diam- ies discussing homogeneous catalysis for biodiesel production eter and Re was determined. Conditions for obtaining stable from rapesee d oil: employing either 4- suspensions were then successfully applied to the polymeri- dodecylbenzenesulfonic acid [68], or NaOCH [69] as the zation application. catalyst. A design of experiments (DoE) methodology was A second reactor configuration was then designed for the applied to determine the optimal conditions, using a meso- polymerization of poly(vinyl acetate) (PVA) and was COBR with integral baffles (Fig. 2a). subdivided into different sections dedicated to: the introduc- In 2014, Mazubert et al. focused on the acid-catalyzed es- tion of the reagents, liquid–liquid dispersion and the reaction terification of fatty acids from waste cooking oil into methyl step. A PVA mass conversion of 30% was obtained but esters, followed by transesterification to produce biodiesel degassing during sampling led to encrusting on the COBR [70]. Two fluidic reactors, namely a microstructured walls. The polymer particles synthesized were spherical and Corning reactor and a NiTech COBR, were compared to a smooth, with a dispersion depending on the initial droplet batch reactor. Significant levels of conversion were achieved suspension homogeneity - directly related to the oscillation more rapidly for the two flow reactors than the batch one. The conditions. 480 J Flow Chem (2020) 10:475–490 Liquid-liquid systems: fine chemical applications chloropyridine (1)and N-boc-2-pyrrole boronic acid (2) (Fig. 5c) showed that this screening setup gave similar yields Turning focus towards the synthesis of small molecules, to small scale batch reactors. This was in contrast to standard Jensen et al. developed a multi-phase flow strategy for screen- flow reactors, where low yields were obtained due to poor ing biphasic catalytic reactions on lab scale (Fig. 5)[73]. This mass transfer at such low flow rates. The threshold flow rate, strategy was based on a previous report by Günther et al. Q , which ensured that the aqueous segment completely th which demonstrated the semi-continuous preparation of gold passed through the organic droplet during each half oscilla- nanorods [74, 75]. By combining oscillatory flow with an tion, was calculated for different amplitude oscillations. inert gas phase, this setup aimed to overcome mass transfer A second Suzuki–Miyaura reaction between 4-chloro-3- limitations and provide interphase mixing independent of the methylanisole (4) and 2,6-difluorophenyl boronic acid (5)em- flow rate. The gaseous phase provided internal circulation, phasized the other advantage of OFRs compared to classical whilst oscillations ensured a “back and forth” movement of flow reactors: in this case, the reduction of the reactor length the aqueous slug within the organic droplet (Fig. 5a). by afactorof254 (Fig. 5d). The third model reaction, a Three palladium-catalyzed cross-coupling reactions were Buchwald–Hartwig amination of 4-chloroanisole (7)with an- assessed. First, the Suzuki–Miyaura coupling between 2- iline (8)(Fig. 5e), showed the same reactivity trend across the three different reactor types, confirming the effectiveness of this strategy. The same strategy was employed by this group for other applications which required decoupling between mixing and residence time. In the nanoparticles field, the nucleation and growth of II-VI and III-V quantum dots (QDs) were per- formed at various high temperatures in a similar OFR and studied by in-line absorbance measurements [76]. A biphasic ligand-exchange reaction of CdSe QDs was also performed with the single droplet OFR. The in-line UV analysis enabled determination of different exchange pathways [77]. Moreover, the strategy was easily implemented in automated platforms and integrated with purification and analysis mod- ules, allowing high throughput screening with small volumes of various reactions for medicinal chemistry or photocatalysis applications [78–80]. In 2016, Roberge, Macchi et al. focused on the use of coiled baffleless OFRs for biphasic reactions instead of COBRs [81]. The hydrolysis of 4-nitrophenyl acetate (10) was used as a model reaction (Fig. 6). A BPR was put at the end of the reactor to avoid cavitation during oscillations, followed by a cyclone system for depressurization [82]. A 7- fold increase in mass transfer was calculated when Re was increased to 5000 by maximizing the oscillation amplitude. This improved mass transfer led to a higher conversion. A dispersed flow regime was predicted at high Re ,witha Fig. 5 Three-phase oscillatory flow strategy for small scale biphasic reaction screening. (a) Illustration of triphasic flow regime, showing oscillations that mix the aqueous slug back and forth within the organic slug (b) Schematic of the reactor setup used for this small scale screening platform (c)Suzuki–Miyaura coupling between N-boc-2-pyrrole boronic acid and 2-chloropyridine (d) Suzuki–Miyaura reaction between 2,6- difluorophenyl boronic acid and 4-chloro-3-methylanisole (e) Buchwald–Hartwig amination of 4-chloroanisole by aniline Fig. 6 Hydrolysis of 4-nitrophenyl acetate in coiled baffleless OFR J Flow Chem (2020) 10:475–490 481 droplet diameter small enough to limit interactions with the COBR enabled efficient extraction of 14 but also the recycling reactor walls. In this case, oscillation frequency was found to of the oxidant, improving the atom efficiency of the setup. have a more significant impact than amplitude on the mass Over the past decade, Bjørsvik et al. have demonstrated a transfer coefficient. range of chemistries in a system termed a multi-jet oscillating A comparison between these reactors and microreactor disk (MJOD) reactor. This has a similar baffle structure to a plates in terms of mass transfer performance and energy dis- COBR with multi-orifice baffles (Fig. 2h), but these baffles are sipation rate showed that OFRs are advantageous for slow attached to a central piston, which oscillates, agitating the entire reactions while microreactor plates are efficient for fast reac- length of the tubular reactor (Fig. 8a)[84, 85]. As with other tions [81]. These microreactors do appear to outperform OFRs OFRs, the amplitude and frequency of this oscillation can be when a sufficient droplet flow regime is not obtained in the varied, in this case with an amplitude between 0 and 25 mm OFR. For intermediate kinetic profiles, the authors proposed a and a frequency of 0–5 Hz. Amongst the chemistry examples hybrid use of both reactors because of their similar perfor- are numerous monophasic reactions [22–24], but also several mance. These results were integrated into a microreactor tool- studies concerning liquid-liquid biphasic reactions. box for choosing the most suitable fluidic reactor by looking The first of these examples was the phase transfer- at the reaction rate and the phases of the medium. catalyzed allylation of phenol 15 using allyl bromide 16 in In 2017, Loponov et al. developed an integrated continuous combination with tetrabutylammonium bromide (TBAB, process for a biphasic reaction and in-line separation (Fig. 7). Fig. 8b). In this case, the batch yield of 75% in a 20 min This comprised of the electrochemical generation of an oxi- reaction time could be matched. Interestingly, the intense dant, cross-flow membrane emulsification, then reaction in a mixing bestowed by this reactor also facilitated the reaction COBR with smooth periodic constrictions (Fig. 2a)[83]. The in the absence of a phase transfer catalyst. In this case, a higher dihydroxylation of styrene (13) by electrochemically generat- reaction concentration and an extended residence time of ed ammonium peroxodisulfate [(NH ) S O ] was chosen as 55 min resulted in a 60% yield [85]. 4 2 2 8 the model reaction. First, the use of a dispersion cell in batch A later study demonstrated the oxidation of alcohols 18 by proved the benefits of emulsions for overcoming mass transfer 1,3-dichloro-5,5-dimethylhydantoin 19, catalyzed by limitations and drastically reducing the reaction time. TEMPO (Fig. 8c). This reaction utilized a biphasic solvent The transposition to flow with a COBR was demonstrated system consisting of CH CN and a saturated NaHCO solu- 3 3 and the unstable emulsion obtained at high oscillatory fre- tion, where intense mixing was provided by the oscillatory quencies and amplitudes provided similar results compared piston. Excellent yields were achieved in this transformation to batch experiments. However, the addition of continuous for a range of benzylic and other alcohols [86]. membrane emulsification was needed to obtain a plug flow Further oxidation of the aldehyde products 20 was then regime in the COBR. This setup unfortunately led to coales- achieved using more forcing conditions (Fig. 8d). The more cence of styrene for long runs, emphasizing the need for fur- reactive 1,3-diiodo-5,5-dimethylhydantoin 21 was required, ther optimization of this method. A second COBR was placed combined with NH OH, longer residence time and higher in series and was equipped with an inlet for ethyl acetate. This temperature. A range of nitriles 22 were obtained in good to strategy aimed to improve the extraction of product 14,since it excellent yields using this protocol in flow. This second oxi- acts as a surfactant to stabilize droplets of by-products. The dative transformation involves the attack of NH OH on the high oscillatory frequency and amplitude favored mixing of carbonyl and, under certain conditions (high NH OH concen- the two immiscible phases, hence also improving the extrac- tration), alternative diaziridine or hydrazine derived products tion. The continuous separator module at the end of the second could be obtained, but with relatively low selectivity [86]. Fig. 7 Three-stage emulsification, oxidation and separation, allowing oxidation of styrene using an electrochemically generated and recyclable oxidant in a COBR 482 J Flow Chem (2020) 10:475–490 also settle in a flow reactor, depending on the properties of the reactor, the solid and the reaction medium. Moreover, feeding solids from pumps is another difficult task. Working at tem- peratures higher than the melting point of the solid can occa- sionally provide a solution, but is by no means a universal approach and requires an intensive optimization process [88, 89]. To overcome such issues, various continuous flow strate- gies have been adopted [90]. Packed-bed reactors are often viewed as a good solution for heterogeneous catalysis, but they suffer from uncontrolled fluid dynamics (including channeling), heat transfer limitations and high pressure drop (which complicates scale-up) [91–94]. Other approaches use active mixing to suspend solids without fouling. Continuous stirred tank reactors [95–97] and agitated cells [98, 99]have been shown to achieve such mixing, as well as the use of sonication [100, 101]. However, these reactors possess a poor RTD, especially when the vessel size is large. The addition of a screw within the fluidic reactor has also recently received attention but this setup does not support gas formation or liquid settling and is only suitable for fast reactions [102–105]. Finally, suppression of fouling can often be achieved by generating a slug flow regime, using an immisci- ble carrier phase which has minimal interactions with the re- actor walls. Scaling up such processes is challenging, though, since larger tube diameters limit the formation of these Taylor flow regimes [106, 107]. In this context, OFRs have emerged as an effective tool for handling solids. Oscillations afford enhanced mixing, which can prevent clogging and solid deposition or sedimentation on the reactor walls. Besides, the configuration of such reactors enables facile scale-up and meso-OFRs have already seen use for solid-liquid reactions. Finally, in contrast to other fluidic reactors, it is possible to improve the efficiency of the reaction within OFRs by simply optimizing the frequency and ampli- Fig. 8 Biphasic applications of a MJOD reactor. (a) Schematic diagram tude of the oscillation. of the MJOD reactor including details of disk structure. (Reprinted with permission from Ref [85]. Copyright (2019) Royal Society of Chemistry) Frequency and amplitude of the oscillation have been dem- (b) Phase transfer catalyzed allylation of phenol (c) Oxidation of alcohols onstrated to have a significant effect on solid suspension. to their corresponding aldehydes (d) Further oxidation of aldehyde prod- Nikačević et al. have distinguished different flow patterns in ucts in the presence of NH OH to yield nitriles a COBR, depending on the oscillatory velocity and, therefore, the power input of the COBR [108]. The oscillatory velocity Solid-liquid systems: general principles must be high enough to overcome sedimentation, otherwise, particles move very slowly at the lowest part of the reactor The use of solids in organic chemistry is very attractive, since (Fig. 9a). Solid suspension can then be either non-uniform their separation from the reaction medium is easy to achieve, (Fig. 9b) when the settling still occurs, or uniform (Fig. 9c) reducing the cost and time needed for downstream purifica- at high oscillatory velocity. tion. Solids can be either a catalyst, reagent or product, which Recent studies on the axial dispersion of solids in OFRs explains their occurrence in many processes for active phar- have demonstrated that solids and liquids do not experience maceutical ingredient (API) production, biomass valorization the same degree of axial dispersion, so the oscillation param- and other applications. Despite the advantages of continuous eters must be tuned accordingly. flow technology, handling solids remains very challenging In 2017, Reis et al. determined the most favorable geo- due to the use of narrow reactor channels, which are suscep- metrical parameters of a COBR for homogeneously suspending solid particles and minimizing the solid axial tible to clogging [87]. Without high flow velocities, solids can J Flow Chem (2020) 10:475–490 483 Fig. 9 Effect of the oscillatory velocity on the solid flow regimes in an oscillatory flow reactor (a) settling of solids (b) non-uniform suspension of solids (c)uniform suspension of solids dispersion [109]. Optical imaging of suspended polyvinyl Solid-liquid systems: synthetic applications chloride (PVC) particles was implemented to determine the RTD of various COBRs using a DoE approach. The Studies of solid-liquid chemical reactions in oscillatory batch results showed that solids experience a longer residence reactors first appeared around 20 years ago, where stable sus- time than liquids due to the strong vortices generated. pensions of TiO particles were achieved for the photochem- Moreover, COBRs with integral smooth periodic constric- ical oxidation of methylene blue and salicylic acid [114–116]. tion baffles (Fig. 2a) require lower oscillation amplitudes The first use of OFRs for solid-liquid systems appeared a to fully suspend solid particles than COBRs with the decade later and was focused on the use of COBRs packed equivalent sharp edged periodic constriction baffles. with a heterogeneous catalyst. More recently, the suspension Finally, the cross-sectional area of the tube was found to of solid particles in a slurry flow has also been demonstrated be the most dominant parameter for controlling back in OFRs. mixing of the solid particles. In 2013, Harvey et al. proposed a protocol for the heterog- The same year, Kramer et al. studied the effects of enous acid-catalyzed esterification of carboxylic acids with oscillation frequency and amplitude on the axial disper- methanol in a meso-COBR [117]. The catalyst consisted of sion of liquids and solids in a COBR [110]. Methylene a suspension of SBA-15 silica derivatized with propylsulfonic blue was the liquid tracer and was followed by absorbance acid (PrSO H-SBA-15). Its turnover frequency (TOF) was measurements, while melamine (the solid tracer) was demonstrated to be higher in continuous flow than stirred monitored using focused beam reflectance measurements. batch conditions. Water poisoning was responsible for the The results indicated that the minimum axial dispersion catalyst deactivation, but this phenomenon was less pro- for solids was obtained at low oscillation amplitudes, but nounced in flow, thanks to the plug flow formed in the the frequency has less impact on this parameter. This meso-COBR, which was proposed to decrease the number study also suggested that velocity ratio, ψ,alone is not of water molecules accessing the resin surface. It was also enough to identify optimal operating conditions. The proven that oscillations were required to avoid sedimentation same conclusions were reached two years later by Cruz of the particles, but the reaction was mixing independent. et al. for planar oscillatory flow crystallizers [111]. Nagy The same group published another study relating to the use et al. also confirmed these observations and found that of a more efficient catalyst (Fig. 10)[118]. This time, the oscillatory pumps afford a larger solid dispersion due to reactor was packed with the Amberlyst 70-SO H resin, which back mixing, compared with net flow pumps. However, is composed of a copolymer of divinylbenzene and styrene, this does not detract from their essential character for functionalized with sulfonic acid. The TOF calculated for this preventing particles from settling and clogging [112]. catalyst was higher than that previously reported for PrSO H- More recently, Rielly et al. used a dual backlit imaging SBA-15. Water deactivation was not observed with the new technique to simultaneously assess the RTD of the liquid catalyst, which showed good stability over time. The recovery and solid phases in a meso-COBR with integral smooth of the resin after water-spiking was facilitated by the water- periodic constriction baffles (Fig. 2a)[113]. This non- resistance of Amberlyst 70-SO H and the continuous flow of invasive technique was not subjected to errors that could reagents. result from using two different measurement techniques Three screening modes were tested in order to find the for the tracers (Procion Red HE-7B dye and polystyrene optimal reaction parameters for the reaction between hexanoic particles). A range of Re values was measured for solid- acid (23) and methanol. These were: multi-steady states liquid plug flow and it was shown that the difference in (waiting for steady state to be reached between varying each axial dispersion experienced by solids and liquids is less parameter), dynamic (continuous variation of one parameter) pronounced at higher oscillation frequency. and multi-dimensional (dynamically varying two parameters From all of these studies, it appears that a compromise at the same time). With the first mode, the highest yield between efficient particle suspension and narrow RTD (min- (95.4%) of methyl hexanoate (24) was obtained at 20 min imal axial dispersion) has to be found during the optimization residence time with a methanol:carboxylic acid ratio of 30:1. of heterogeneous reactions in OFRs. Similar results were achieved between the multi-steady state 484 J Flow Chem (2020) 10:475–490 (27) and 95.3% of methyl acetate (28) were obtained at 50 °C, proving the two-stage process to be more efficient. The first reactor for the esterification was packed with the basic Amberlyst A26-OH resin and connected to the second one, con- taining the acidic Amberlyst 70-SO H resin for the acetylation. Better yields of 28 (99.1% in 25 min residence time) and 27 (76.5% in 35 min residence time) were obtained, because less methanol was required for this process (6 equivalents instead of Fig. 10 Heterogenous acid-catalyzed esterification of hexanoic acid with 30) and the first reaction was more efficient with the basic resin. methanol with a meso-COBR Finally, around 76% of glycerol content was valorized into 27, considered as an additive to the biodiesel fuel produced during and the dynamic modes, but the dynamic mode decreased the the transesterification step. required screening time by about 16%. Finally, in the multi- In a related transformation, Yunus et al. performed the dimensional mode, the highest yield of 16 (98.2%) was semi-quantitative conversion of palm fatty acid distillate achieved with a shorter residence time of 14 min and lower (PFAD) into biodiesel with a new glucose acid catalyst methanol:carboxylic acid ratio of 21:1. [120]. This catalyst was characterized by various techniques One year later, Harvey et al. proposed another protocol before being used as a suspension in the meso-COBR. which combines biodiesel production from triglycerides with Different parameters, such as catalyst loading, temperature, the valorization of the glycerol by-product [119]. This study oscillation frequency, residence time and methanol:PFAD ra- employed a meso-COBR packed with different Amberlyst tio were optimized. The produced biodiesel showed similar resin catalysts, in a process which could be carried out in properties to petroleum and diesel fuel standards. one or two stages (Fig. 11). The transesterification of triacetin In 2017, a miniaturized COBR was manufactured by (25) was assessed with two resins: the first was basic stereolithography-based 3D printing for the preparation of sta- (Amberlyst A26-OH) and the second acidic (Amberlyst 70- ble silver nanoparticles (Ag NPs) (Fig. 12)[121]. This cheap, SO H). The basic catalyst led to a higher reaction rate and high-quality and fast manufacturing enabled the construction TOF than the acidic one. However, this basic catalyst was of a small scale COBR resistant to high pressures and inert to water-intolerant and was irreversibly deactivated, even in the various solvents. For the sake of offering a cheap and easy continuous flow of fresh reagents. setup, a syringe pump was used to generate oscillation. This Then, the reaction between glycerol and acetone (26)was was equipped with a syringe filled with an immiscible silicon studied in the presence of methanol. The single-stage process oil to prevent back mixing of the reagents. By estimating the consisted of a stream containing methanol, triacetin and acetone dimensionless RTD E(θ), thanks to absorbance measurements in a 30:1:4 ratio. The catalyst used was the acidic Amberlyst 70- with a methylene blue tracer, the RTD within the mini COBR SO H resin. After 30 min residence time, only 48.5% of solketal was proven to be narrower than in a similar tubular reactor. The experimental results of E(θ) fit well with an axial disper- sion model. The mixing quality was improved by increasing the fre- quency and the amplitude of the oscillation since it led to a higher Peclet number. The velocity ratio, ψ, was high due to the relatively low net flow rate. The mixing efficiency was advantageously exploited for the continuous flow synthesis of Ag NPs. AgNO in acetonitrile was used as the silver Fig. 11 Concomitant formation of the biodiesel methyl acetate and solketal from glycerol valorization in a packed meso-COBR during a two-stage process Fig. 12 Synthesis of silver nanoparticles in a miniaturized COBR J Flow Chem (2020) 10:475–490 485 source, NaBH as the reductant and polyvinylpyrrolidone was found to be 1.6 Hz, since higher frequency was proposed (29) as the stabilizer. TEM and in-line UV-VIS analyses to produce smaller product particles, which could not be ef- showed that Ag NPs produced in the mini COBR were more fectively recovered by filtration at the reactor outlet [123]. A uniform and stable than those synthesized with a standard separate study on the use of this iodinating agent 21 was tubular reactor. Reactor fouling was also less pronounced originally attempted using a slurry feed, but clogging prob- with the mini COBR. A low oscillation amplitude and fre- lems were observed, despite the oscillatory flow [124]. quency were, however, required to ensure NP Accordingly, H SO was added to the aqueous phase in order 2 4 monodispersity. to fully dissolve the starting material and operate under homo- In 2018, Zheng et al. used an oscillatory generator and a geneous conditions. meso-COBR packed with an acid-cation exchange resin In a further example of solid processing using this reactor, a (NKC-9) as a heterogeneous catalyst for the acetylation of novel method for reduction of olefins and aromatic nitro camphene 30 (Fig. 13)[122]. groups was described (Fig. 14c). This method uses NaBH The residence time distribution of this reactor was proven as a reductant, in combination with stoichiometric CoSO . to be close to that of an ideal PFR. The time needed to reach The authors state that a solid (proposed to be the catalytically equilibrium was shorter under flow than batch conditions be- active species, Co B) is formed, alongside H gas evolution, 2 2 cause of improved heat- and mass transfer. A similar trend justifying the use of an OFR. A high oscillation frequency of was observed for the yield and selectivity for isobornyl acetate 3 Hz enabled gram-scale processing, where excellent yields (32). The specific surface area and pore volume of the catalyst were obtained in a short residence time of 3 min [125]. after 30 h of processing in a COBR were similar to those of the In 2020, Roberge et al. implemented a coiled baffleless fresh catalyst. Moreover, the catalyst particles were less prone OFR, combined with a bellows-based oscillatory pump, to to the formation of a by-product layer at their surface in a handle solids [126]. Two reactions involving the formation COBR than in a round-bottom flask, although high oscillatory of a solid were assessed. The first consisted of a reaction amplitudes could induce the formation of cracks. between cyclohexylamine (38) and glyoxal (39), forming the Bjørsvik et al. have also been studied various solid-liquid insoluble diimine product 40 (Fig. 15a). A batch experiment reactions in the previously introduced MJOD reactor (Fig. 8a). was conducted first, to characterize the reaction. When the These include reactions where solids are fed into the reactor as a slurry, but also cases with solid reaction products being formed in the reactor itself. One of the studies discussed in the initial report of this reactor demonstrated a Hofmann rear- rangement, where the benzamide substrate 33 was fed as a slurry in 1,4-dioxane (Fig. 14a)[85]. The desired aniline prod- uct 34 could be isolated in quantitative yield after 15 min residence time at 80 °C. A reagent which has been the focus of two studies from the same group is 1,3-diiodo-5,5-dimethylhydantoin 21, whose preparation and onward reaction have both been studied using this reactor. The preparation of 21 was achieved by reacting 5,5-dimethylhydantoin 35 with ICl in a basic biphasic medi- um (Fig. 14b). In this case, the optimal oscillation frequency Fig. 14 Solid-liquid applications of a MJOD reactor. (a) Hofmann rearrangement using a slurry feed of starting material 33 (b) Preparation of iodinating reagent 21, formed as a solid during the reaction (c)NaBH reduction of nitro groups using an in-situ generated heterogeneous Fig. 13 Heterogenous acid-catalyzed acetylation of camphene in a meso- catalyst COBR 486 J Flow Chem (2020) 10:475–490 Another recent application of solid handling in OFRs is at the interface between photochemistry, chemical engineering and organic chemistry. The use of heterogeneous catalysts in photochemistry has attracted attention as a replacement for expensive and toxic rare metal homogeneous catalysts, which are difficult to recycle [128–130]. Coupled with continuous flow technology, heterogeneous photochemistry becomes more reliable, safer and greener [131]. Therefore, using OFRs for heterogeneous photocatalysis is a perfect match. A study by Kappe, Pieber et al. demonstrated the handling of solids in a baffleless OFR for C-N coupling reactions cat- alyzed by a carbon nitride semiconductor photocatalyst (CN- OA-m) coupled with NiBr .3H O(Fig. 16a)[132]. The OFR 2 2 (HANU reactor, Creaflow) consists of a series of cubic static mixing elements, providing split-and-recombine mixing (Fig. 3c). This facilitated a stable suspension in combination with an oscillatory diaphragm pulsator unit. In this case, it was essential to include a pulsation dampener and BPR to prevent suction of air or cavitation during backward pulsation. A series of RTD measurements allowed quantification of the plug flow character, using the corresponding Fig. 15 Solid formation reactions in a coiled baffleless OFR (a) Bodenstein numbers for each set of oscillation parameters. precipitation reaction for the synthesis of diimine 40 (b)Liquid−liquid The amplitude and frequency of the pulsation were opti- phase transfer catalyzed reaction for the synthesis of gem- dichlorocyclopropane 44, with solid by-product formation mized to ensure reaction efficiency, while maintaining a stable suspension of photocatalyst. At low frequencies and amplitudes, quantitative yields of 47 were obtained reaction was transposed to flow, pulsations were required to from the model reaction between 4-bromobenzoate (45) avoid clogging. A yield between 39 and 48% was obtained, and pyrrolidine (46). It was shown that the photocatalyst depending on the reagent concentrations. Using these condi- could be recycled over ten cycles without significant tions, a long run of 5 h was performed without any clogging. changes in the yield. The stability, robustness and scal- Lowering the oscillatory frequency and amplitude, to mini- ability of the protocol were also proven by performing a −1 mize the energy dissipation rate, had no detrimental impact long runof4.5hand isolating>12 g(2.67gh )of on the handling of the solid. aniline 47.Finally,anintermediate(49) in the synthesis The second reaction was a liquid-liquid reaction between of the API tetracaine was synthesized with a productivity −1 dichlorocarbene, formed by the deprotonation of chloroform of 1.12 g h after a slight reoptimization of the reaction (42) with NaOH, and cyclohexene (41)toform7,7- conditions (Fig. 16b). dichlorobicyclo[4.1.0]-heptane (44) and solid NaCl (Fig. The most recent example reported for slurry handling in an 15b). This reaction requires a phase transfer catalyst, OFR also considers a photochemical transformation performed diethylmethylamine (43), to transfer the formed in the commercial HANU reactor, reported by Debrouwer et al. dichlorocarbene between the aqueous and organic phases. [133]. Experimental and computational characterization of the This reaction had already been reported under continuous oscillatory flow regime in this type of OFR were established flow conditions in a packed bed reactor, but had a narrow and showed that the implementation of high oscillation ampli- operating window due to problems with clogging [127]. tudes increases axial mixing. Similar trends to those determined Using the previously optimized conditions, clogging rapidly by the Kappe group for the RTD under different oscillatory occurred because of a side-reaction leading to gas formation conditions were also obtained [132]. The adaptation of a hetero- 2 3 and oscillation dampening. geneous nickel/photoredox catalyzed C(sp )-C(sp ) cross- Characterization of the flow conditions showed that in- electrophile coupling to flow was assessed. Coupling an aryl creasing the oscillation frequency led to a broadening of the bromide (45) with an alkyl bromide (50) provided the product RTD. However, Dean secondary flows (Table 1)weregener- 51 (Fig. 17). Despite the excellent suspension properties of the ated, which minimized solid stagnation and deposition. The OFR, sedimentation of the solid Na CO base occured in the 2 3 particle size distribution was larger under flow than batch supply tubing to the pump. Decreasing the particle size by wet conditions. When the oscillation frequency was increased, milling overcame this issue, allowing stable processing, with a −1 smaller and less uniform particles were obtained. high productivity of 51 (0.87gh ). J Flow Chem (2020) 10:475–490 487 Fig. 16 C-N coupling reaction photocatalyzed by a carbon nitride semiconductor (CN-OA-m) and NiBr .3H O catalyst in a microstructured OFR (a) 2 2 optimized conditions for cross-coupling of 4-bromobenzoate 45 and pyrrolidine 46 as a model reaction (b) synthesis of tetracaine precursor 49 Conclusions compromise between efficacious particle suspension and minimal axial dispersion has to be found during reaction Since the addition of an oscillatory flow to a net flow im- optimization. proves heat- and mass transfer, OFRs have already been used From this overview, it is clear that OFRs broaden the possi- to overcome issues encountered by conventional flow reac- bilities to tackle chemical challenges under continuous flow con- tors. Liquid-liquid biphasic reactions that need high mass ditions. With this in mind, OFRs are certain to become a key transfer have been effectively implemented under oscillatory element in the flow chemist’s toolbox, particularly for slow or flow conditions. Biodiesel production in particular, which in- multiphasic reactions. The performance of OFRs can be weaker volves immiscible phases and long residence times, has been at high net flow rates or outside of the droplet flow regimes for studied in-depth for over 20 years. Rational design has en- liquid-liquid reactions. Moreover, the compressibility of gases hanced the assets of such reactors towards optimal baffle appears to discourage their use for such applications, because structures for large scale operation. The decoupling of mixing of oscillation dampening. Nevertheless, the diversity of baffle and residence time has also been recently exploited for other geometries renders these reactors suitable for different purposes, reactions in liquid-liquid systems, for example in small scale while oscillation conditions represent novel process parameters reaction screening. to consider and optimize. In this way, OFRs enable a smooth In addition, the appealing profile of these reactors for transition for a wider range of chemistries from batch to flow, and solid-liquid reactions has first been tested in packed meso- can facilitate scale-up based on maintaining the system’s hydro- COBRs. A deeper understanding of the effect of oscillations dynamic characteristics. It is anticipated that new OFR configu- on the handling of solids, thanks to recent studies concerning rations will be developed in the coming years, to tackle the han- axial dispersion, has enabled scientists to efficiently perform dling of dense solids and high solid loadings, which remain reactions with suspended solid particles. Nevertheless, a challenging for flow processing. 2 3 Fig. 17 Heterogeneous photocatalytic C(sp )-C(sp ) cross-electrophile coupling of aryl bromide 45 with alkyl bromide 50 in a microstructured OFR 488 J Flow Chem (2020) 10:475–490 Acknowledgements Open access funding provided by University of 17. Ni X, Mackley MR, Harvey AP, Stonestreet P, Baird MHI, Rama Graz. Rao N v. (2003). Chem Eng Res Des 81:373–338 18. Ricardo C, Ni X (2009). Org Process Res Dev 13:1080–1087 Code availability Not applicable. 19. McDonough JR, Phan AN, Harvey AP (2018). Chem Eng and Process Process Intensif 129:51–62 20. Mohd Rasdi FR, Phan AN, Harvey AP (2013). Chem Eng J 222: Funding information The CC FLOW Project (Austrian Research 282–291 Promotion Agency FFG No. 862766) is funded through the Austrian 21. McDonough JR, Phan AN, Reay DA, Harvey AP (2016). Chem COMET Program by the Austrian Federal Ministry of Transport, Eng Process Process Intensif 110:201–213 Innovation and Technology (BMVIT), the Austrian Federal Ministry of 22. Sleveland D, Bjørsvik HR (2012). Org Process Res Dev 16:1121– Science, Research and Economy (BMWFW), and by the State of Styria (Styrian Funding Agency SFG). P.B. acknowledges support by the 23. Bjørsvik HR, Liguori L (2014). Org Process Res Dev 18:1509– University of Liege through the Erasmus+ funding program. 24. Drageset A, Elumalai V, Bjørsvik HR (2018). React Chem Eng 3: Data availability Not applicable. 550–558 25. Lucas MS, Reis NM, Li Puma G (2016). Chem Eng J 296:335– Compliance with ethical standards 26. Navarro-Fuentes F, Keane M, Ni X (2019). Org Process Res Dev 23:38–44 Conflicts of interest/competing interests Not applicable. 27. Graça CAL, Lima RB, Pereira MFR, Silva AMT, Ferreira A (2020). Chem Eng J 389:124412 Open Access This article is licensed under a Creative Commons 28. Spaccini R, Liguori L, Punta C, Bjørsvik HR (2012). Attribution 4.0 International License, which permits use, sharing, adap- ChemSusChem 5:261–265 tation, distribution and reproduction in any medium or format, as long as 29. McDonough JR, Murta S, Law R, Harvey AP (2019). Chem Eng J you give appropriate credit to the original author(s) and the source, pro- 358:643–657 vide a link to the Creative Commons licence, and indicate if changes were 30. Slavnić DS, Živković L v., Bjelić A v., Bugarski BM, Nikačević made. The images or other third party material in this article are included NM (2017). J Chem Technol Biotechnol 92:2178–2188 in the article's Creative Commons licence, unless indicated otherwise in a 31. Smith KB, Mackley MR (2006). Chem Eng Res Des 84:1001– credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by 32. Phan AN, Harvey A (2010). Chem Eng J 159:212–219 statutory regulation or exceeds the permitted use, you will need to obtain 33. Zheng M, Mackley M (2008). Chem Eng Sci 63:1788–1799 permission directly from the copyright holder. To view a copy of this 34. Reis N, Vicente AA, Teixeira JA (2010). Chem Eng Process licence, visit http://creativecommons.org/licenses/by/4.0/. Process Intensif 49:793–803 35. Ahmed SMR, Phan AN, Harvey AP (2017). Chem Eng Technol 40:907–914 36. Mazubert A, Fletcher DF, Poux M, Aubin J (2016). Chem Eng References Process Process Intensif 108:78–92 37. McDonough JR, Ahmed SMR, Phan AN, Harvey AP (2017). Chem Eng Sci 171:160–178 1. Rehm TH (2019). ChemPhotoChem 3:1–21 38. Millor GJ, Lee YC, Ni X (2019). Comp Chem Eng 124:14–27 2. Atobe M, Tateno H, Matsumura Y (2018). Chem Rev 118:4541– 39. Nogueira X, Taylor BJ, Gomez H, Colominas I, Mackley MR (2013). Comp Chem Eng 49:1–17 3. Cambié D, Bottecchia C, Straathof NJW, Hessel V, Noël T 40. Phan AN, Harvey AP (2012). Chem Eng J 180:229–236 (2016). Chem Rev 116:10276–11034 41. McDonough JR, Oates MF, Law R, Harvey AP (2019). Chem 4. Gutmann B, Cantillo D, Kappe CO (2015). Angew Chem - Int Ed Eng J 361:508–518 54:6688–6728 42. Phan AN, Harvey A, Lavender J (2011). Chem Eng Process 5. Bezerra MA, Lemos VA, de Oliveira DM, Novaes CG, Barreto Process Intensif 50:254–263 JA, Alves JPS, Cerqueira UMF da M, Santos QO dos, Araújo SA 43. Reis N, Harvey AP, Mackley MR, Vicente AA, Teixeira JA (2020). Microchem J 155:10473 (2005). Chem Eng Res Des 83:357–371 6. Baumann M (2018). Org Biomol Chem 16:5946–5954 44. González Niño C, Kapur N, King MF, de Boer G, Blacker AJ, 7. Britton J, Raston CL (2017). Chem Soc Rev 46:1250–1271 Bourne R, Thompson H (2019). Int J Comput Fluid Dyn 33:317– 8. Trojanowicz M (2020). Molecules 25:1434 9. di Filippo M, Bracken C, Baumann M (2020). Molecules 25:356 45. Reintjens RWEG, Thathagar M (2011) Oscillating flow 10. Hommes A, Heeres HJ, Yue J (2019). ChemCatChem 11:4671– minireactor. US patent 10,118,149 B2 (Jan 13, 2011) 46. Jähnisch K, Hessel V, Löwe H, Baerns M (2004). Angew Chem - 11. Sui J, Yan J, Liu D, Wang K, Luo G (2019). Small 1902828:1–23 Int Ed 43:406–446 12. Berton M, de Souza JM, Abdiaj I, McQuade DT, Snead DR 47. Mielke E, Roberge DM, Macchi A (2016). J Flow Chem 6:279– (2020). J Flow Chem 10:73–92 13. Stonestreet P, Harvey AP (2002). Chem Eng Res and Des 80:31– 48. Reis N, Pereira RN, Vicente AA, Teixeira JA (2008). Ind Eng Chem Res 47:7190–7201 14. McGlone T, Briggs NEB, Clark CA, Brown CJ, Sefcik J, Florence 49. Law R, Ahmed S, Tang N, Phan A, Harvey A (2018). Chem Eng AJ (2015). Org Process Res Dev 19:1186–1202 Process Process Intensif 125:133–138 15. Abbott MSR, Harvey AP, Valente Perez G, Theodorou MK 50. González-Juárez D, Herrero-Martín R, Solano JP (2018). Appl (2013). Interface Focus 3:20120036 Therm Eng 141:494–502 16. McDonough JR, Phan AN, Harvey AP (2015). Chem Eng J 265: 51. Solano JP, Herrero R, Espín S, Phan AN, Harvey AP (2012). 110–121 Chem Eng Res Des 90:732–742 J Flow Chem (2020) 10:475–490 489 52. Ahmed SMR, Phan AN, Harvey AP (2018). Chem Eng Process 85. Liguori L, Bjørsvik HR (2011). Org Process Res Dev 15:997– Process Intensif 130:229–239 1009 53. Jimeno G, Lee YC, Ni X (2018). Chem Eng Process Process 86. Drageset A, Frøystein NÅ, Törnroos KW, Bjørsvik HR (2019). Intensif 134:153–162 React Chem Eng 4:41–51 54. Avila M, Fletcher DF, Poux M, Xuereb C, Aubin J (2020). Chem 87. Schoenitz M, Grundemann L, Augustin W, Scholl S (2015). Eng Sci 212:115310 Chem Commun 51:8213–8228 55. Ni X, Zhang Y, Mustafa I (1998). Chem Eng Sci 53:2903–2919 88. Kashani S, Sullivan RJ, Andersen M, Newman SG (2018). Green 56. Ni X, Zhang Y, Mustafa I (1999). Chem Eng Sci 54:841–850 Chem 20:1748–1753 89. Heider PL, Born SC, Basak S, Benyahia B, Lakerveld R, Zhang 57. Ni X, Johnstone JC, Symes KC, Grey BD, Bennett DC (2001). AIChE 47:1746–1757 H, Hogan R, Buchbinder L, Wolfe A, Mascia S, Evans JMB, Jamison TF, Jensen KF (2014). Org Process Res Dev 18:402–409 58. Ni X, Murray KR, Zhang Y, Bennett D, Howes T (2002). Powder 90. Hartman RL (2012). Org Process Res Dev 16:870–887 Technol 124:281–286 59. Harvey AP, Mackley MR, Stonestreet P (2001). Ind Eng Chem 91. Wang Z, Gérardy R, Gauron G, Damblon C, Monbaliu JCM (2019). React Chem Eng 4:17–26 Res 40:5371–5377 60. Mohd GATI, Resul MFMG, Yunus R, Yaw TCS (2008). Sci 92. Colella M, Carlucci C, Luisi R (2018). Top Curr Chem 376:46 Technol 3:138–145 93. Munirathinam R, Huskens J, Verboom W (2015). Adv Synth 61. Zheng M, Skelton RL, Mackley MR (2007). Process Saf Environ Catal 357:1093–1123 Prot 85:365–371 94. Rodríguez-Escrich C, Pericàs MA (2018). Chem Rec 18:1–20 62. Harvey AP, Mackley MR, Seliger T (2003). J Chem Technol 95. Pomberger A, Mo Y, Nandiwale KY, Schultz VL, Duvadie R, Biotechnol 78:338–341 Robinson RI, Altinoglu EI, Jensen KF (2019). Org Process Res 63. Phan AN, Harvey AP, Rawcliffe M (2011). Fuel Process Technol Dev 23:2699–2706 92:1560–1567 96. Mo Y, Jensen KF (2016). React Chem Eng 1:501–507 64. Sagmeister P, Poms J, Williams JD, Kappe CO (2020). React 97. Chapman MR, Kwan MHT, King G, Jolley KE, Hussain M, Chem Eng 5:677–684 Hussain S, Salama IE, González Nino C, Thompson LA, 65. Hone CA, Holmes N, Akien GR, Bourne RA, Muller FL (2017). Bayana ME, Clayton AD, Nguyen BN, Turner NJ, Kapur N, React Chem Eng 2:103–108 Blacker AJ (2017). Org Process Res Dev 21:1294–1301 66. Moore JS, Jensen KF (2014). Angew Chemie - Int Ed 53:470–473 98. Filipponi P, Gioiello A, Baxendale IR (2016). Org Process Res 67. Phan AN, Harvey AP, Eze V (2012). Chem Eng Technol 35: Dev 20:371–375 99. Browne DL, Deadman BJ, Ashe R, Baxendale IR, Ley SV (2011). 1214–1220 68. Al-Saadi LS, Alegría A, Eze VC, Harvey AP (2019). Chem Eng Org Process Res Dev 15:693–697 Technol 42:539–548 100. Falß S, Tomaiuolo G, Perazzo A, Hodgson P, Yaseneva P, Zakrzewski J, Guido S, Lapkin A, Woodward R, Meadows RE 69. Al-Saadi LS, Eze VC, Harvey AP (2020). Biofuels. https://doi. org/10.1080/17597269.2020.1727688 (2016). Org Process Res Dev 20:558–567 101. Noël T, Naber JR, Hartman RL, McMullen JP, Jensen KF, 70. Mazubert A, Aubin J, Elgue S, Poux M (2014). Green Process Buchwald SL (2011). Chem Sci 2:287–290 Synth 3:419–429 71. Mazubert A, Crockatt M, Poux M, Aubin J, Roelands M (2015). 102. Cao Q, Howard JL, Crawford DE, James SL, Browne DL (2018). Green Chem 20:4443–4447 Chem Eng Technol 38:2161–2169 72. Lobry E, Lasuye T, Gourdon C, Xuereb C (2015). Chem Eng J 103. Isoni V, Mendoza K, Lim E, Teoh SK (2017). Org Process Res Dev 21:992–1002 259:505–518 73. Abolhasani M, Bruno NC, Jensen KF (2015). Chem Commun 51: 104. Sharma BM, Atapalkar RS, Kulkarni AA (2019). Green Chem 21: 8916–8919 5639–5646 105. Crawford DE, Miskimmin CK, Cahir J, James SL (2017). Chem 74. Abolhasani M, Oskooei A, Kumacheva E, Günther A (2014). Lab Chip 14:2309–2318 Commun 53:13067–13070 106. Bannock JH, Krishnadasan SH, Nightingale AM, Yau CP, Khaw 75. Gielen F, Van Vliet L, Koprowski BT, Devenish SRA, K, Burkitt D, Halls JJM, Heeney M, de Mello JC (2013). Adv Fischlechner M, Edel JB, Niu X, deMello AJ, Hollfelder, F Funct Mater 23:2123–2129 (2013). Anal Chem 85:4761–4769 107. Poe SL, Cummings MA, Haaf MP, McQuade DT (2006). Angew 76. Abolhasani M, Coley CW, Xie L, Chen O, Bawendi MG, Jensen KF (2015). Chem Mater 27:6131–6138 Chem - Int Ed 45:1544–1548 108. Slavnić D, Bugarski B, Nikačević N (2019). Chem Eng Process 77. Shen Y, Abolhasani M, Chen Y, Xie L, Yang L, Coley CW, Process Intensif 135:108–119 Bawendi MG, Jensen KF (2017). Angew Chem - Int Ed 56: 16333–16337 109. Ejim LN, Yerdelen S, McGlone T, Onyemelukwe I, Johnston B, Florence AJ, Reis NM (2017). Chem Eng J 308:669–682 78. Coley CW, Abolhasani M, Lin H, Jensen KF (2017). Angew Chem - Int Ed 56:9847-9850 110. Kacker R, Regensburg SI, Kramer HJM (2017). Chem Eng J 317: 79. Hwang YJ, Coley CW, Abolhasani M, Marzinzik AL, Koch G, 413–423 111. Cruz P, Silva C, Rocha F, Ferreira A (2019). AIChE 65:e16683 Spanka C, Lehmann H, Jensen KF (2017). Chem Commun 53: 6649–6652 112. Oliva JA, Pal K, Barton A, Firth P, Nagy ZK (2018). Chem Eng J 351:498–505 80. Hsieh HW, Coley CW, Baumgartner LM, Jensen KF, Robinson RI (2018). Org Process Res Dev 22:542–550 113. Onyemelukwe II, Nagy ZK, Rielly CD (2020). Chem Eng J 382: 81. Mongeon SS, Roberge DM, Bittel M, Elsner P, Macchi A (2016). 122862 Org Process Res Dev 20:733–741 114. Fabiyi ME, Skelton RL (1999) J Photochem Photobiol. A 129:17– 82. Rafiee SE, Sadeghiazad MM (2017). Appl Therm Eng 114:300– 24 327 115. Fabiyi ME, Skelton RL (2000). Trans IChemE 78:399–404 83. Loponov KN, Deadman BJ, Zhu J, Rielly C, Holdich RG, Hii KK, 116. Gao P, Han Ching W, Herrmann M, Kwong Chan C, Yue PL Hellgardt K (2017). Chem Eng J 329:220–230 (2003). Chem Eng Sci 58:1013–1020 84. Liguori L, Bjørsvik HR (2007) Continuous flow reactor. World 117. Eze VC, Phan AN, Pirez C, Harvey AP, Lee AF, Wilson K (2013). patent WO 2007/058544 A1, Nov 17, 2006 Catal Sci Tech 3:2373–2237 490 J Flow Chem (2020) 10:475–490 118. Eze VC, Fisher JC, Phan AN, Harvey AP (2017). Chem Eng J 127. von Keutz T, Cantillo D, Kappe CO (2019). J Flow Chem 9:27–34 322:205–214 128. Gisbertz S, Pieber B (2020). ChemPhotoChem 4:1–21 119. Eze VC, Harvey AP (2018). Chem Eng J 347:41–51 129. Riente P, Noël T (2019). Catal Sci Technol 9:5186–5232 120. Kefas HM, Yunus R, Rashid U, Taufiq-Yap YH (2019). J Environ 130. Savateev A, Ghosh I, König B, Antonietti M (2018). Angew Chem Eng 7:102993 Chem - Int Ed 57:15936–15947 121. Okafor O, Weilhard A, Fernandes JA, Karjalainen E, Goodridge 131. Williams JD, Kappe CO (2020). Curr Opin Green Sustain R, Sans V (2017). React Chem Eng 2:129–136 Chem 25:100351 122. Zheng H, Yan Z, Chu S, Chen J (2018). Chem Eng Process 132. Rosso C, Gisbertz S, Williams JD, Gemoets H, Debrouwer W, Process Intensif 134:1–8 Pieber B, Kappe CO (2020). React Chem Eng 5:595–604 123. Ferreri M, Drageset A, Gambarotti C, Bjørsvik HR (2016). React 133. Debrouwer W, Kimpe W, Dangreau R, Huvaere K, Gemoets Chem Eng 1:379–386 HPL, Mottaghi M, Kuhn S, Van Aken K (2020). Org Process 124. Drageset A, Bjørsvik HR (2016). React Chem Eng 1:436–444 Res Dev. https://doi.org/10.1021/acs.oprd.0c00150 125. Lundevall FJ, Elumalai V, Drageset A, Totland C, Bjørsvik HR (2018). Eur J Org Chem 2018:3416–3425 126. Doyle BJ, Gutmann B, Bittel M, Hubler T, MacChi A, Roberge Publisher’snote Springer Nature remains neutral with regard to jurisdic- DM (2020). Ind Eng Chem Res 59:4007–4019 tional claims in published maps and institutional affiliations. http://www.deepdyve.com/assets/images/DeepDyve-Logo-lg.png Journal of Flow Chemistry Springer Journals

Oscillatory flow reactors for synthetic chemistry applications

Loading next page...
 
/lp/springer-journals/oscillatory-flow-reactors-for-synthetic-chemistry-applications-S2hXCL7iJO
Publisher
Springer Journals
Copyright
Copyright © The Author(s) 2020
ISSN
2062-249X
eISSN
2063-0212
DOI
10.1007/s41981-020-00105-6
Publisher site
See Article on Publisher Site

Abstract

Oscillatory flow reactors (OFRs) superimpose an oscillatory flow to the net movement through a flow reactor. OFRs have been engineered to enable improved mixing, excellent heat- and mass transfer and good plug flow character under a broad range of operating conditions. Such features render these reactors appealing, since they are suitable for reactions that require long residence times, improved mass transfer (such as in biphasic liquid-liquid systems) or to homogeneously suspend solid particles. Various OFR configurations, offering specific features, have been developed over the past two decades, with significant progress still being made. This review outlines the principles and recent advances in OFR technology and overviews the synthetic applications of OFRs for liquid-liquid and solid-liquid biphasic systems. . . . . . Keywords Oscillatory flow Multiphasic reactions Solid handling Liquid-liquid reactions Process intensification Plug flow reactors Introduction reactors, whilst maintaining plug flow behavior, intense mixing, and excellent heat transfer. Importantly, oscillation- Continuous flow technology has attracted the attention of promoted mixing is independent of the net flow rate, opening chemists in recent years, specifically in scaling up hazardous the window of operation to lower flow rates (longer residence chemistries, photochemistry, electrochemistry and many times), which would otherwise impart poor hydrodynamic others [1–4]. The use of various auxiliary and automation properties. Another consequence of the ability to operate at technologies has strengthened the potential of flow reactors low flow rates is the allowable decrease in reactor volume, [5–7] for a wide range of applications [8–11]. Despite the enabling compact setups. In addition, the quality of mixing benefits of continuous flow technology, flow reactors face can be ensured during scale-up when OFR geometric ratios some inherent problems, particularly when dealing with mul- and fluidic characteristics are maintained [13]. tiphasic systems. Rational engineering has, however, begun to A small number of reviews concerning OFRs do already tackle these issues, by developing a variety of reactor config- exist, but these have focused specifically on crystallization urations with specific features aimed at processing challeng- [14], biological processes [15] or engineering developments ing reaction media on both small and large scale [12]. [16, 17]. Here, we give an overview of the technological back- One particularly promising development in reactor technol- ground and potential of this reactor type, aimed towards the ogy is that of oscillatory flow reactors (OFRs), whereby a synthetically-oriented flow chemist. symmetrical oscillation is superimposed to the net flow Since the efficiency of flow reactors has already been well through the reactor. This has been found to offer a solution explored for homogeneous reaction media, only a few reports to multiple problems encountered by conventional flow examine these monophasic reactions in OFRs [18–24]. Furthermore, despite numerous studies concerning gas-liquid mass transfer in OFRs [25–27], few articles have been dedi- * C. Oliver Kappe cated to synthetic applications with gases, likely due to their oliver.kappe@uni-graz.at compressibility, which can dampen oscillations [28]. On the other hand, the high mass-transfer and continuous agitation Institute of Chemistry, University of Graz, NAWI Graz, capability of these reactors have already begun to see signifi- Heinrichstrasse 28, 8010 Graz, Austria cant uptake for biphasic liquid-liquid and solid-liquid reac- Center for Continuous Flow Synthesis and Processing (CC FLOW), tions. Consequently, this review will focus on these two clas- Research Center Pharmaceutical, Engineering (RCPE), Inffeldgasse 13, 8010 Graz, Austria ses of biphasic systems. First, the general principles and 476 J Flow Chem (2020) 10:475–490 advantages of OFRs will be presented for each, followed by turbulent flow regime (Fig. 1b). However, this can be obtained an overview of experimental reports. at much lower Re in an OFR, due to the additional oscillatory turbulence, quantified by the oscillatory Reynolds number (Re )[14, 29]. This enables a drastic reduction of the reactor length. The plug flow behavior related to vortex cycles can be Oscillatory flow reactor principles achieved at velocity ratios (ψ) between 2 and 10, depending and equipment on the baffle geometry. Usually, Re and St values larger than o r 100 and 0.5, respectively, are required to fully realize the Oscillation can be applied to a flow reactor using various benefits of OFRs.[14]. oscillatory devices, such as piston- bellow-, diaphragm-, Residence time distribution (RTD), the distribution of time syringe- or peristaltic pumps. The symmetrical oscillation spent by the fluid within a flow reactor, is an important mea- generates vortices, leading to improved radial mixing, whilst sure of the reactor’s plug flow character. A narrow distribution aiming to maintain plug flow character (minimal axial mixing, decreases the probability of side reactions or incomplete reac- Fig. 1a). In this way, mixing is decoupled from the net flow tion to occur (due to portions of material being present in the and depends only on the oscillation conditions and the OFR reactor for too long, or not long enough). RTD is related to the configuration. Therefore, reactions requiring long residence axial dispersion coefficient, which is inversely proportional to times can be performed efficiently without requiring large the Peclet or Bodenstein number (Table 1)[30]. RTD exper- reactor volumes. Accordingly, the numerous issues facing iments are usually performed with a colored tracer and the large volume tubular flow reactors can be obviated (e.g., high dimensionless RTD parameter, E(θ), is plotted as a function pressure drop, large footprint and poor residence time of the normalized time, θ (Fig. 1c). When plug flow is distribution). achieved, a narrow symmetric curve is centered at θ =1 OFRs can be characterized by different dimensionless (Fig. 1c, green curve). Any asymmetrical tendencies or cen- numbers (summarized in Table 1), where the maximum oscil- tering at another θ suggests axial dispersion (Fig. 1c,red latory velocity is considered. Standard flow reactors require a curve). net flow Reynolds number (Re ) above 4000 to achieve a Axial dispersion is less sensitive to oscillatory conditions at high net flow rates, since laminar flow is minimized. Conversely, experiments with low Re (generally with long residence times) require careful optimization of the oscillatory radial mixing axial mixing amplitude and frequency to obtain a narrow RTD. Oscillation amplitude has a larger impact than frequency on the axial dispersion. At low Re numbers, a higher oscillation ampli- tude decreases the axial dispersion, thanks to the creation of vortices. After reaching a critical Re value, the opposite trend appears: the axial dispersion begins to increase, due to over propagation of vortices [31–34]. The optimal Re value de- Laminar flow Turbulent flow pends on the net flow rate, such that the velocity ratio (ψ)still (Low Re) (High Re) provides good plug flow character. This value must therefore be optimized for each individual reactor system. The most common way of generating vortices is the addi- tion of baffles within the OFR. Various baffle geometries exist and offer different degrees of mixing (Fig. 2). At each colli- sion with an obstacle, the void behind the baffle creates a backward movement of the fluid and a short reverse flow appears. These short-lived vortices at each baffle intersection are called eddies (Fig. 3a). In continuous oscillatory baffled reactors (COBRs), oscillations lead to the generation and dis- appearance of eddies and therefore to homogeneous radial mixing. Fig. 1 Summary of mixing and flow regimes in flow reactors: a Radial The complex flow behavior in COBRs with different baffle versus axial mixing in a tube. b Laminar versus turbulent flow regimes. c geometries has been rigorously studied experimentally and Depiction of a reactor’s plug flow character, estimated using tracer theoretically [35–43]. However, it is still difficult to clearly injection (blue) experiments, where good plug flow character has low quantify the impact of the geometry on reactor performance, axial dispersion (green), compared to poor plug flow character, with significant axial dispersion (red) since it depends on multiple parameters. It appears, though, J Flow Chem (2020) 10:475–490 477 Table 1 Dimensionless numbers for characterizing oscillatory flow reactors Dimensionless numbers Formula Interpretation υρD Net flow Reynolds number Re ¼ Measure of the mixing intensity in net flow xo2πfρD Oscillatory Reynolds number Re ¼ Measure of the mixing intensity in oscillatory flow Reo Velocity ratio ψ ¼ Re Measure of the dominance of the oscillatory flow over the net flow Strouhal number St ¼ Degree of eddy propagation in a COBR (Fig. 3a) 4πxqffiffiffiffiffiffi Dean number De ¼ Re Likelihood of Dean vortices in a coiled baffleless OFR (Fig. 3b) 2Rc υL Peclet or Bodenstein number Bo ¼ Pe ¼ Degree of axial back-mixing a −3 −1 μ = viscosity (Pa.s), ρ =density (kg m ), υ = superficial fluid velocity (m s ), D = tube diameter (m), f = frequency of oscillation (Hz), x = center-to- −2 −1 peak amplitude (m), R = radius of curvature (m), D = axial dispersion coefficient, L = tube length (m), h = heat transfer coefficient (W m K ), k = c a t −1 −1 thermal conductivity (W m K ) that helical baffles (Fig. 2c-e) have a wide plug flow operating popularity of split-and-recombine type mixers in microfluidic range, because of additional spiraling character to the net flow applications [45, 46]. [36, 39, 40]. Disc and doughnut baffles (Fig. 2i) are also Efficient multiphasic mass transfer is another strong asset proposed to achieve efficient mixing at low oscillatory flow of OFRs. Studies, principally focused on gas-liquid systems in [35]. flow, have shown that the mass transfer coefficients for gas- The second method for vortex generation is the use of coiled liquid and liquid-liquid systems are generally higher than in reactors instead of straight tubes [44]. A secondary flow, named conventional flow reactors, and remain high for a wider range Dean flow, appears due to the centripetal force arising at the of conditions, especially at low flow rates [27, 47]. Stronger curvature of the tube. The fluid near the inner wall is pushed oscillation (high Re ) reduces the droplet or bubble size (until outwards, towards the center of the tube, while an opposing a minimal value), which increases the interfacial area. current appears from the outside to the center of the coil. This Frequency and amplitude contribute with the same weight to superimposed flow is therefore seen as a pair of counter-rotating this enhancement [27]. vortices (Fig. 3b). In coiled baffleless oscillatory flow coil reac- Similarly, heat transfer is enhanced in OFRs compared to tors, these Dean vortices are formed and collapsed continuously conventional tubes, especially at low flow rates, thanks to during oscillations, reducing stagnation zones at their center and improved radial mixing. This enhancement depends strongly improving radial mixing [29]. This behavior can be quantified by on Re ,but noton the St value. The heat transfer coefficient o r the Dean number (De,Table 1). increases with the oscillatory conditions and reaches a plateau A third oscillatory mixing principle is the combination with at Re ≈ 1300 [48–50]. When considering heat transfer, oscil- a split-and-recombine mixing structure (Fig. 3c). By applying lation frequency has a higher impact than amplitude. A recent oscillation, each of these elements will be effectively revisited study showed that helical COBRs (Fig. 2c-e) weremoreeffi- numerous times by the fluid, multiplying its effect. cient than central axial COBRs (Fig. 2b), followed by single Furthermore, these split channels have a smaller diameter, so orifice COBRs (Fig. 2g) in terms of thermal performance a higher velocity (thus, higher Re ) can be reached with a (considering both heat transfer and pressure drop) [51]. given oscillatory amplitude. This principle has received sur- prisingly little attention thus far, particularly considering the Fig. 3 Vortices generated in oscillatory flow reactors leading to improved Fig. 2 Configurations of oscillatory baffled reactors. (a) integral baffles radial mixing during oscillations. (a) Eddies in a continuous oscillatory with periodic constrictions than can be sharp or smooth (b) central axial baffled reactor (COBR) (b) Dean vortices in a baffleless OFR. (c) baffles (c) round-edged helical baffles (d) sharp-edged helical baffles (e) Oscillation-enhanced split-and-recombine mixing. The red arrows indi- sharp-edged helical baffles with a central rod (f) wire wool baffles (g) cate net flow, whilst blue arrows illustrate the mixing instilled by single-orifice baffles (h) multi-orifice baffles (i) disc and doughnut baffles oscillations 478 J Flow Chem (2020) 10:475–490 Another advantage of OFRs is their ease of scale-up. By recently that other biphasic chemical reactions have been re- keeping the dynamic parameters (Re , Re and St ) and geo- ported using OFRs. n o r metric features of the reactor (baffle spacing and open area) The implementation of biphasic reactions in COBRs was constant, similar axial dispersion and mixing behavior can be pioneered by the group of Harvey. In 2001, in cooperation obtained with changing reactor volumes [13, 52]. Finally, with a sterol producer, they evaluated the batch-to-flow tran- OFRs demonstrate better performance than stirred tank reac- sition of steryl ester saponification in a meso-COBR [59]. A tors at similar power dissipation values. Power dissipation has pair of stainless-steel pistons provided oscillations within the been estimated using simulations based on two main models, 1.3 L flow reactor. Thanks to a flow conversion model, based namely the quasi-steady flow and the eddy acoustic model on fluid mechanics and reaction kinetics, they could lower the [53, 54]. reaction temperature and improve the sterol yield compared to An important factor to consider when dealing with oscilla- a typical batch procedure. The quality of mixing and mass tions is their dampening with distance from the oscillatory transfer was high, enabling an efficient biphasic reaction be- pump. The formation of air bubbles within the reactor also tween the steryl esters and ethanol. The optimal conditions has a dampening effect that could be problematic for mass found with this “tanks-in-series” modelling approach led to transfer or for the handling of solids. Pulsations are potentially better selectivity for the desired sterol, despite the complex unsuitable for reactions involving gases since they are com- mixture at the beginning. The lab-scale tests resulted in a −1 pressible. Consequently, it recommended to ensure that the production of 5 kg h , while a pilot-scale installation in the −1 OFR is airtight and to consider the strength of the oscillation production site led to generation of sterols at about 20 kg h . at the reactor outlet, prior to experimentation [52]. These results were expanded upon a year later when the Nevertheless, OFRs have mostly been assessed for reac- same group proposed a methodology for designing tions that are inefficient in conventional flow reactors. This production-scale OFRs [13]. By using Re and Re values as o n explains why numerous articles relying on biphasic media and the starting point for the design, different parameters could be solid handling have been reported. In the following sections, determined, such as the power density and COBR geometry. these applications will be discussed in detail. Two experimental cases were then assessed. The first of these cases was dedicated to the addition of an end-group, such as anhydride, to a polymer, forming a succinimide moiety [13]. The aim of this study was to obtain Biphasic applications of oscillatory flow −1 a productivity of 2 t h by operating at 200 °C with a residence reactors time of 4 h. Such a goal was seen as impractical with a simple tubular reactor because it would require a very long reactor Liquid-liquid systems: bulk chemical applications length and high power density. These requirements were, how- ever, achieved experimentally with a compact industrial COBR Experiments examining liquid-liquid systems in OFRs ap- placed horizontally. The second case concerned the saponifica- peared at roughly the same time as their oscillatory batch tion of steryl ester [13]. Based on the previous study [59], a reactor equivalent. In the late 1990s, Ni et al. studied the −1 productivity of 50 t day at 85 °C and with a residence of correlation of polymer particle size with droplet size during 20 min was required. A much smaller COBR configuration batch methylmethacrylate polymerization [55–58]. In the than in the first case was used, because this reaction involved meantime, investigations into biodiesel production in OFRs a shorter residence time and less viscous reagents. took place over the course of about 15 years (Fig. 4). OFRs are This reactor design methodology was also used in 2008 by good candidates for this reaction, which requires high mass Mohd et al., showing that COBRs are more powerful than transfer due to the immiscibility and difference in viscosity standard plug flow reactors (PFRs) for biodiesel production between triglycerides and methanol. It is only relatively [60]. Their simulation showed that, for a fixed reactor diame- ter, COBRs give higher conversion at reduced residence time than PFRs and require a lower power density. Harvey et al. also investigated biodiesel production [61]. Their preliminary study demonstrated the potential of OFRs for this goal [62]. Then, in 2007, they showed that high oscil- lation frequencies were required to produce biodiesel that ful- fills EU standards. Small differences between batch oscillato- ry baffled reactors and standard batch reactors were demon- strated. However, operating under flow conditions gave better quality biodiesel, in terms of triglyceride and diglyceride con- Fig. 4 General procedure for biodiesel production in a COBR centrations, than with a batch oscillatory baffled reactor. J Flow Chem (2020) 10:475–490 479 In 2011, Phan et al. investigated the potential of continuous Corning reactor benefited from a wider operating window screening for biodiesel production with newly designed (higher pressure and temperature) compared to the NiTech COBRs that allow plug flow at low Re [63]. Trials were first COBR, hence speeding up the reaction further. However, re- conducted at low flow rates in a meso-COBR with central action equilibrium was never achieved in the limited residence sharp-edged baffles (Fig. 2b). Long experiments proved the time allowed by this reactor. Furthermore, clogging issues production stability over time, with small deviations that did occurred after long runs with high levels of free fatty acid not depend on the oscillatory frequency. However, mixing content. On the contrary, the reaction equilibrium could be was not sufficient and so homogeneity could not be rapidly obtained even at lower temperatures in the NiTech established. COBR, which also showed higher flexibility in the choice of The baffles were replaced with helical round wire baffles the residence time since mixing was independent from the net (Fig. 2c), but insufficient mixing led to stratification of the two flow. phases and large fluctuations in the biodiesel yield. Therefore, A comparison between two OFRs was also performed by new baffle designs were developed to increase shear stress, the same group for the esterification of fatty acids from waste resulting in the use of axially oriented sharp-edged helical cooking oil. In this instance, the transesterification was done baffles with or without a central rod (Fig. 2d–e). Slug flows with glycerol to produce glycerides, which have applications were less prevalent with this baffle geometry, but flow in the cosmetics and pharmaceutical fields [71]. Contrary to channeling was still observed in the absence of the central the previous report [70], the two OFRs: a COBR and a rod. Steady-state screening was then implemented with differ- baffleless OFR (“pulsed helix reactor”), were made of stain- ent methanol:triglyceride ratios by using the oscillatory reac- less steel to support higher temperatures and pressures. A tor at low Re . Then, dynamic screening (continuous variation pulsation dampener and a back pressure regulator (BPR) of one parameter [64–66]) was assessed and results between were employed at the reactor outlets to prevent cavitation the two screening modes were found to be in agreement. or solvent vaporization at these elevated temperatures. The same year, Harvey et al. also tested three meso- Higher conversions were obtained with these reactors than COBRs with different baffle geometries for biodiesel produc- with reactors operated at lower temperatures (a NiTech glass tion [67]. Depending on the configuration of the baffles, COBR and a batch reactor). The COBR showed better per- steady state was achieved after different startup times:1.5, formance than the baffleless OFR, which the authors attrib- 2.5, and 4.0 residence times for the integral, wire wool, and uted to the generation of smaller droplets (and higher inter- helical baffle designs (Fig. 2a, f, c), respectively. In each case, facial area) due to the baffles. However, higher conversion this startup time was shorter than that required for a single was obtained in the baffleless OFR when the acid catalyst continuous stirred tank reactor. Integral baffles gave the was changed. Different experimental parameters were stud- highest content of biodiesel at high Re . Moreover, stable ied and led to higher selectivity than those previously report- steady state was only established at high oscillatory frequen- ed in the literature. cies, especially for integral baffles, otherwise mixing was in- In 2015, Lobry et al. studied liquid-liquid dispersion in a sufficient to avoid stratification between the immiscible COBR for polymerization applications [72]. A first reactor phases. The startup time decreased with increasing Re num- configuration was developed for the liquid-liquid dispersion ber - likely due to better mixing efficiency. The decrease in study. It appeared that the dispersion properties were not af- yield at longer residence times was related to saponification of fected by the net flow rate, enabling dispersion-independent the biodiesel. A short residence time (5–10 min) was thus control of the residence time. Moreover, oscillation ampli- required to achieve optimal biodiesel yield. tudes and frequencies were the main parameters responsible The same group has also recently published two other stud- for droplet breakage. A correlation between the droplet diam- ies discussing homogeneous catalysis for biodiesel production eter and Re was determined. Conditions for obtaining stable from rapesee d oil: employing either 4- suspensions were then successfully applied to the polymeri- dodecylbenzenesulfonic acid [68], or NaOCH [69] as the zation application. catalyst. A design of experiments (DoE) methodology was A second reactor configuration was then designed for the applied to determine the optimal conditions, using a meso- polymerization of poly(vinyl acetate) (PVA) and was COBR with integral baffles (Fig. 2a). subdivided into different sections dedicated to: the introduc- In 2014, Mazubert et al. focused on the acid-catalyzed es- tion of the reagents, liquid–liquid dispersion and the reaction terification of fatty acids from waste cooking oil into methyl step. A PVA mass conversion of 30% was obtained but esters, followed by transesterification to produce biodiesel degassing during sampling led to encrusting on the COBR [70]. Two fluidic reactors, namely a microstructured walls. The polymer particles synthesized were spherical and Corning reactor and a NiTech COBR, were compared to a smooth, with a dispersion depending on the initial droplet batch reactor. Significant levels of conversion were achieved suspension homogeneity - directly related to the oscillation more rapidly for the two flow reactors than the batch one. The conditions. 480 J Flow Chem (2020) 10:475–490 Liquid-liquid systems: fine chemical applications chloropyridine (1)and N-boc-2-pyrrole boronic acid (2) (Fig. 5c) showed that this screening setup gave similar yields Turning focus towards the synthesis of small molecules, to small scale batch reactors. This was in contrast to standard Jensen et al. developed a multi-phase flow strategy for screen- flow reactors, where low yields were obtained due to poor ing biphasic catalytic reactions on lab scale (Fig. 5)[73]. This mass transfer at such low flow rates. The threshold flow rate, strategy was based on a previous report by Günther et al. Q , which ensured that the aqueous segment completely th which demonstrated the semi-continuous preparation of gold passed through the organic droplet during each half oscilla- nanorods [74, 75]. By combining oscillatory flow with an tion, was calculated for different amplitude oscillations. inert gas phase, this setup aimed to overcome mass transfer A second Suzuki–Miyaura reaction between 4-chloro-3- limitations and provide interphase mixing independent of the methylanisole (4) and 2,6-difluorophenyl boronic acid (5)em- flow rate. The gaseous phase provided internal circulation, phasized the other advantage of OFRs compared to classical whilst oscillations ensured a “back and forth” movement of flow reactors: in this case, the reduction of the reactor length the aqueous slug within the organic droplet (Fig. 5a). by afactorof254 (Fig. 5d). The third model reaction, a Three palladium-catalyzed cross-coupling reactions were Buchwald–Hartwig amination of 4-chloroanisole (7)with an- assessed. First, the Suzuki–Miyaura coupling between 2- iline (8)(Fig. 5e), showed the same reactivity trend across the three different reactor types, confirming the effectiveness of this strategy. The same strategy was employed by this group for other applications which required decoupling between mixing and residence time. In the nanoparticles field, the nucleation and growth of II-VI and III-V quantum dots (QDs) were per- formed at various high temperatures in a similar OFR and studied by in-line absorbance measurements [76]. A biphasic ligand-exchange reaction of CdSe QDs was also performed with the single droplet OFR. The in-line UV analysis enabled determination of different exchange pathways [77]. Moreover, the strategy was easily implemented in automated platforms and integrated with purification and analysis mod- ules, allowing high throughput screening with small volumes of various reactions for medicinal chemistry or photocatalysis applications [78–80]. In 2016, Roberge, Macchi et al. focused on the use of coiled baffleless OFRs for biphasic reactions instead of COBRs [81]. The hydrolysis of 4-nitrophenyl acetate (10) was used as a model reaction (Fig. 6). A BPR was put at the end of the reactor to avoid cavitation during oscillations, followed by a cyclone system for depressurization [82]. A 7- fold increase in mass transfer was calculated when Re was increased to 5000 by maximizing the oscillation amplitude. This improved mass transfer led to a higher conversion. A dispersed flow regime was predicted at high Re ,witha Fig. 5 Three-phase oscillatory flow strategy for small scale biphasic reaction screening. (a) Illustration of triphasic flow regime, showing oscillations that mix the aqueous slug back and forth within the organic slug (b) Schematic of the reactor setup used for this small scale screening platform (c)Suzuki–Miyaura coupling between N-boc-2-pyrrole boronic acid and 2-chloropyridine (d) Suzuki–Miyaura reaction between 2,6- difluorophenyl boronic acid and 4-chloro-3-methylanisole (e) Buchwald–Hartwig amination of 4-chloroanisole by aniline Fig. 6 Hydrolysis of 4-nitrophenyl acetate in coiled baffleless OFR J Flow Chem (2020) 10:475–490 481 droplet diameter small enough to limit interactions with the COBR enabled efficient extraction of 14 but also the recycling reactor walls. In this case, oscillation frequency was found to of the oxidant, improving the atom efficiency of the setup. have a more significant impact than amplitude on the mass Over the past decade, Bjørsvik et al. have demonstrated a transfer coefficient. range of chemistries in a system termed a multi-jet oscillating A comparison between these reactors and microreactor disk (MJOD) reactor. This has a similar baffle structure to a plates in terms of mass transfer performance and energy dis- COBR with multi-orifice baffles (Fig. 2h), but these baffles are sipation rate showed that OFRs are advantageous for slow attached to a central piston, which oscillates, agitating the entire reactions while microreactor plates are efficient for fast reac- length of the tubular reactor (Fig. 8a)[84, 85]. As with other tions [81]. These microreactors do appear to outperform OFRs OFRs, the amplitude and frequency of this oscillation can be when a sufficient droplet flow regime is not obtained in the varied, in this case with an amplitude between 0 and 25 mm OFR. For intermediate kinetic profiles, the authors proposed a and a frequency of 0–5 Hz. Amongst the chemistry examples hybrid use of both reactors because of their similar perfor- are numerous monophasic reactions [22–24], but also several mance. These results were integrated into a microreactor tool- studies concerning liquid-liquid biphasic reactions. box for choosing the most suitable fluidic reactor by looking The first of these examples was the phase transfer- at the reaction rate and the phases of the medium. catalyzed allylation of phenol 15 using allyl bromide 16 in In 2017, Loponov et al. developed an integrated continuous combination with tetrabutylammonium bromide (TBAB, process for a biphasic reaction and in-line separation (Fig. 7). Fig. 8b). In this case, the batch yield of 75% in a 20 min This comprised of the electrochemical generation of an oxi- reaction time could be matched. Interestingly, the intense dant, cross-flow membrane emulsification, then reaction in a mixing bestowed by this reactor also facilitated the reaction COBR with smooth periodic constrictions (Fig. 2a)[83]. The in the absence of a phase transfer catalyst. In this case, a higher dihydroxylation of styrene (13) by electrochemically generat- reaction concentration and an extended residence time of ed ammonium peroxodisulfate [(NH ) S O ] was chosen as 55 min resulted in a 60% yield [85]. 4 2 2 8 the model reaction. First, the use of a dispersion cell in batch A later study demonstrated the oxidation of alcohols 18 by proved the benefits of emulsions for overcoming mass transfer 1,3-dichloro-5,5-dimethylhydantoin 19, catalyzed by limitations and drastically reducing the reaction time. TEMPO (Fig. 8c). This reaction utilized a biphasic solvent The transposition to flow with a COBR was demonstrated system consisting of CH CN and a saturated NaHCO solu- 3 3 and the unstable emulsion obtained at high oscillatory fre- tion, where intense mixing was provided by the oscillatory quencies and amplitudes provided similar results compared piston. Excellent yields were achieved in this transformation to batch experiments. However, the addition of continuous for a range of benzylic and other alcohols [86]. membrane emulsification was needed to obtain a plug flow Further oxidation of the aldehyde products 20 was then regime in the COBR. This setup unfortunately led to coales- achieved using more forcing conditions (Fig. 8d). The more cence of styrene for long runs, emphasizing the need for fur- reactive 1,3-diiodo-5,5-dimethylhydantoin 21 was required, ther optimization of this method. A second COBR was placed combined with NH OH, longer residence time and higher in series and was equipped with an inlet for ethyl acetate. This temperature. A range of nitriles 22 were obtained in good to strategy aimed to improve the extraction of product 14,since it excellent yields using this protocol in flow. This second oxi- acts as a surfactant to stabilize droplets of by-products. The dative transformation involves the attack of NH OH on the high oscillatory frequency and amplitude favored mixing of carbonyl and, under certain conditions (high NH OH concen- the two immiscible phases, hence also improving the extrac- tration), alternative diaziridine or hydrazine derived products tion. The continuous separator module at the end of the second could be obtained, but with relatively low selectivity [86]. Fig. 7 Three-stage emulsification, oxidation and separation, allowing oxidation of styrene using an electrochemically generated and recyclable oxidant in a COBR 482 J Flow Chem (2020) 10:475–490 also settle in a flow reactor, depending on the properties of the reactor, the solid and the reaction medium. Moreover, feeding solids from pumps is another difficult task. Working at tem- peratures higher than the melting point of the solid can occa- sionally provide a solution, but is by no means a universal approach and requires an intensive optimization process [88, 89]. To overcome such issues, various continuous flow strate- gies have been adopted [90]. Packed-bed reactors are often viewed as a good solution for heterogeneous catalysis, but they suffer from uncontrolled fluid dynamics (including channeling), heat transfer limitations and high pressure drop (which complicates scale-up) [91–94]. Other approaches use active mixing to suspend solids without fouling. Continuous stirred tank reactors [95–97] and agitated cells [98, 99]have been shown to achieve such mixing, as well as the use of sonication [100, 101]. However, these reactors possess a poor RTD, especially when the vessel size is large. The addition of a screw within the fluidic reactor has also recently received attention but this setup does not support gas formation or liquid settling and is only suitable for fast reactions [102–105]. Finally, suppression of fouling can often be achieved by generating a slug flow regime, using an immisci- ble carrier phase which has minimal interactions with the re- actor walls. Scaling up such processes is challenging, though, since larger tube diameters limit the formation of these Taylor flow regimes [106, 107]. In this context, OFRs have emerged as an effective tool for handling solids. Oscillations afford enhanced mixing, which can prevent clogging and solid deposition or sedimentation on the reactor walls. Besides, the configuration of such reactors enables facile scale-up and meso-OFRs have already seen use for solid-liquid reactions. Finally, in contrast to other fluidic reactors, it is possible to improve the efficiency of the reaction within OFRs by simply optimizing the frequency and ampli- Fig. 8 Biphasic applications of a MJOD reactor. (a) Schematic diagram tude of the oscillation. of the MJOD reactor including details of disk structure. (Reprinted with permission from Ref [85]. Copyright (2019) Royal Society of Chemistry) Frequency and amplitude of the oscillation have been dem- (b) Phase transfer catalyzed allylation of phenol (c) Oxidation of alcohols onstrated to have a significant effect on solid suspension. to their corresponding aldehydes (d) Further oxidation of aldehyde prod- Nikačević et al. have distinguished different flow patterns in ucts in the presence of NH OH to yield nitriles a COBR, depending on the oscillatory velocity and, therefore, the power input of the COBR [108]. The oscillatory velocity Solid-liquid systems: general principles must be high enough to overcome sedimentation, otherwise, particles move very slowly at the lowest part of the reactor The use of solids in organic chemistry is very attractive, since (Fig. 9a). Solid suspension can then be either non-uniform their separation from the reaction medium is easy to achieve, (Fig. 9b) when the settling still occurs, or uniform (Fig. 9c) reducing the cost and time needed for downstream purifica- at high oscillatory velocity. tion. Solids can be either a catalyst, reagent or product, which Recent studies on the axial dispersion of solids in OFRs explains their occurrence in many processes for active phar- have demonstrated that solids and liquids do not experience maceutical ingredient (API) production, biomass valorization the same degree of axial dispersion, so the oscillation param- and other applications. Despite the advantages of continuous eters must be tuned accordingly. flow technology, handling solids remains very challenging In 2017, Reis et al. determined the most favorable geo- due to the use of narrow reactor channels, which are suscep- metrical parameters of a COBR for homogeneously suspending solid particles and minimizing the solid axial tible to clogging [87]. Without high flow velocities, solids can J Flow Chem (2020) 10:475–490 483 Fig. 9 Effect of the oscillatory velocity on the solid flow regimes in an oscillatory flow reactor (a) settling of solids (b) non-uniform suspension of solids (c)uniform suspension of solids dispersion [109]. Optical imaging of suspended polyvinyl Solid-liquid systems: synthetic applications chloride (PVC) particles was implemented to determine the RTD of various COBRs using a DoE approach. The Studies of solid-liquid chemical reactions in oscillatory batch results showed that solids experience a longer residence reactors first appeared around 20 years ago, where stable sus- time than liquids due to the strong vortices generated. pensions of TiO particles were achieved for the photochem- Moreover, COBRs with integral smooth periodic constric- ical oxidation of methylene blue and salicylic acid [114–116]. tion baffles (Fig. 2a) require lower oscillation amplitudes The first use of OFRs for solid-liquid systems appeared a to fully suspend solid particles than COBRs with the decade later and was focused on the use of COBRs packed equivalent sharp edged periodic constriction baffles. with a heterogeneous catalyst. More recently, the suspension Finally, the cross-sectional area of the tube was found to of solid particles in a slurry flow has also been demonstrated be the most dominant parameter for controlling back in OFRs. mixing of the solid particles. In 2013, Harvey et al. proposed a protocol for the heterog- The same year, Kramer et al. studied the effects of enous acid-catalyzed esterification of carboxylic acids with oscillation frequency and amplitude on the axial disper- methanol in a meso-COBR [117]. The catalyst consisted of sion of liquids and solids in a COBR [110]. Methylene a suspension of SBA-15 silica derivatized with propylsulfonic blue was the liquid tracer and was followed by absorbance acid (PrSO H-SBA-15). Its turnover frequency (TOF) was measurements, while melamine (the solid tracer) was demonstrated to be higher in continuous flow than stirred monitored using focused beam reflectance measurements. batch conditions. Water poisoning was responsible for the The results indicated that the minimum axial dispersion catalyst deactivation, but this phenomenon was less pro- for solids was obtained at low oscillation amplitudes, but nounced in flow, thanks to the plug flow formed in the the frequency has less impact on this parameter. This meso-COBR, which was proposed to decrease the number study also suggested that velocity ratio, ψ,alone is not of water molecules accessing the resin surface. It was also enough to identify optimal operating conditions. The proven that oscillations were required to avoid sedimentation same conclusions were reached two years later by Cruz of the particles, but the reaction was mixing independent. et al. for planar oscillatory flow crystallizers [111]. Nagy The same group published another study relating to the use et al. also confirmed these observations and found that of a more efficient catalyst (Fig. 10)[118]. This time, the oscillatory pumps afford a larger solid dispersion due to reactor was packed with the Amberlyst 70-SO H resin, which back mixing, compared with net flow pumps. However, is composed of a copolymer of divinylbenzene and styrene, this does not detract from their essential character for functionalized with sulfonic acid. The TOF calculated for this preventing particles from settling and clogging [112]. catalyst was higher than that previously reported for PrSO H- More recently, Rielly et al. used a dual backlit imaging SBA-15. Water deactivation was not observed with the new technique to simultaneously assess the RTD of the liquid catalyst, which showed good stability over time. The recovery and solid phases in a meso-COBR with integral smooth of the resin after water-spiking was facilitated by the water- periodic constriction baffles (Fig. 2a)[113]. This non- resistance of Amberlyst 70-SO H and the continuous flow of invasive technique was not subjected to errors that could reagents. result from using two different measurement techniques Three screening modes were tested in order to find the for the tracers (Procion Red HE-7B dye and polystyrene optimal reaction parameters for the reaction between hexanoic particles). A range of Re values was measured for solid- acid (23) and methanol. These were: multi-steady states liquid plug flow and it was shown that the difference in (waiting for steady state to be reached between varying each axial dispersion experienced by solids and liquids is less parameter), dynamic (continuous variation of one parameter) pronounced at higher oscillation frequency. and multi-dimensional (dynamically varying two parameters From all of these studies, it appears that a compromise at the same time). With the first mode, the highest yield between efficient particle suspension and narrow RTD (min- (95.4%) of methyl hexanoate (24) was obtained at 20 min imal axial dispersion) has to be found during the optimization residence time with a methanol:carboxylic acid ratio of 30:1. of heterogeneous reactions in OFRs. Similar results were achieved between the multi-steady state 484 J Flow Chem (2020) 10:475–490 (27) and 95.3% of methyl acetate (28) were obtained at 50 °C, proving the two-stage process to be more efficient. The first reactor for the esterification was packed with the basic Amberlyst A26-OH resin and connected to the second one, con- taining the acidic Amberlyst 70-SO H resin for the acetylation. Better yields of 28 (99.1% in 25 min residence time) and 27 (76.5% in 35 min residence time) were obtained, because less methanol was required for this process (6 equivalents instead of Fig. 10 Heterogenous acid-catalyzed esterification of hexanoic acid with 30) and the first reaction was more efficient with the basic resin. methanol with a meso-COBR Finally, around 76% of glycerol content was valorized into 27, considered as an additive to the biodiesel fuel produced during and the dynamic modes, but the dynamic mode decreased the the transesterification step. required screening time by about 16%. Finally, in the multi- In a related transformation, Yunus et al. performed the dimensional mode, the highest yield of 16 (98.2%) was semi-quantitative conversion of palm fatty acid distillate achieved with a shorter residence time of 14 min and lower (PFAD) into biodiesel with a new glucose acid catalyst methanol:carboxylic acid ratio of 21:1. [120]. This catalyst was characterized by various techniques One year later, Harvey et al. proposed another protocol before being used as a suspension in the meso-COBR. which combines biodiesel production from triglycerides with Different parameters, such as catalyst loading, temperature, the valorization of the glycerol by-product [119]. This study oscillation frequency, residence time and methanol:PFAD ra- employed a meso-COBR packed with different Amberlyst tio were optimized. The produced biodiesel showed similar resin catalysts, in a process which could be carried out in properties to petroleum and diesel fuel standards. one or two stages (Fig. 11). The transesterification of triacetin In 2017, a miniaturized COBR was manufactured by (25) was assessed with two resins: the first was basic stereolithography-based 3D printing for the preparation of sta- (Amberlyst A26-OH) and the second acidic (Amberlyst 70- ble silver nanoparticles (Ag NPs) (Fig. 12)[121]. This cheap, SO H). The basic catalyst led to a higher reaction rate and high-quality and fast manufacturing enabled the construction TOF than the acidic one. However, this basic catalyst was of a small scale COBR resistant to high pressures and inert to water-intolerant and was irreversibly deactivated, even in the various solvents. For the sake of offering a cheap and easy continuous flow of fresh reagents. setup, a syringe pump was used to generate oscillation. This Then, the reaction between glycerol and acetone (26)was was equipped with a syringe filled with an immiscible silicon studied in the presence of methanol. The single-stage process oil to prevent back mixing of the reagents. By estimating the consisted of a stream containing methanol, triacetin and acetone dimensionless RTD E(θ), thanks to absorbance measurements in a 30:1:4 ratio. The catalyst used was the acidic Amberlyst 70- with a methylene blue tracer, the RTD within the mini COBR SO H resin. After 30 min residence time, only 48.5% of solketal was proven to be narrower than in a similar tubular reactor. The experimental results of E(θ) fit well with an axial disper- sion model. The mixing quality was improved by increasing the fre- quency and the amplitude of the oscillation since it led to a higher Peclet number. The velocity ratio, ψ, was high due to the relatively low net flow rate. The mixing efficiency was advantageously exploited for the continuous flow synthesis of Ag NPs. AgNO in acetonitrile was used as the silver Fig. 11 Concomitant formation of the biodiesel methyl acetate and solketal from glycerol valorization in a packed meso-COBR during a two-stage process Fig. 12 Synthesis of silver nanoparticles in a miniaturized COBR J Flow Chem (2020) 10:475–490 485 source, NaBH as the reductant and polyvinylpyrrolidone was found to be 1.6 Hz, since higher frequency was proposed (29) as the stabilizer. TEM and in-line UV-VIS analyses to produce smaller product particles, which could not be ef- showed that Ag NPs produced in the mini COBR were more fectively recovered by filtration at the reactor outlet [123]. A uniform and stable than those synthesized with a standard separate study on the use of this iodinating agent 21 was tubular reactor. Reactor fouling was also less pronounced originally attempted using a slurry feed, but clogging prob- with the mini COBR. A low oscillation amplitude and fre- lems were observed, despite the oscillatory flow [124]. quency were, however, required to ensure NP Accordingly, H SO was added to the aqueous phase in order 2 4 monodispersity. to fully dissolve the starting material and operate under homo- In 2018, Zheng et al. used an oscillatory generator and a geneous conditions. meso-COBR packed with an acid-cation exchange resin In a further example of solid processing using this reactor, a (NKC-9) as a heterogeneous catalyst for the acetylation of novel method for reduction of olefins and aromatic nitro camphene 30 (Fig. 13)[122]. groups was described (Fig. 14c). This method uses NaBH The residence time distribution of this reactor was proven as a reductant, in combination with stoichiometric CoSO . to be close to that of an ideal PFR. The time needed to reach The authors state that a solid (proposed to be the catalytically equilibrium was shorter under flow than batch conditions be- active species, Co B) is formed, alongside H gas evolution, 2 2 cause of improved heat- and mass transfer. A similar trend justifying the use of an OFR. A high oscillation frequency of was observed for the yield and selectivity for isobornyl acetate 3 Hz enabled gram-scale processing, where excellent yields (32). The specific surface area and pore volume of the catalyst were obtained in a short residence time of 3 min [125]. after 30 h of processing in a COBR were similar to those of the In 2020, Roberge et al. implemented a coiled baffleless fresh catalyst. Moreover, the catalyst particles were less prone OFR, combined with a bellows-based oscillatory pump, to to the formation of a by-product layer at their surface in a handle solids [126]. Two reactions involving the formation COBR than in a round-bottom flask, although high oscillatory of a solid were assessed. The first consisted of a reaction amplitudes could induce the formation of cracks. between cyclohexylamine (38) and glyoxal (39), forming the Bjørsvik et al. have also been studied various solid-liquid insoluble diimine product 40 (Fig. 15a). A batch experiment reactions in the previously introduced MJOD reactor (Fig. 8a). was conducted first, to characterize the reaction. When the These include reactions where solids are fed into the reactor as a slurry, but also cases with solid reaction products being formed in the reactor itself. One of the studies discussed in the initial report of this reactor demonstrated a Hofmann rear- rangement, where the benzamide substrate 33 was fed as a slurry in 1,4-dioxane (Fig. 14a)[85]. The desired aniline prod- uct 34 could be isolated in quantitative yield after 15 min residence time at 80 °C. A reagent which has been the focus of two studies from the same group is 1,3-diiodo-5,5-dimethylhydantoin 21, whose preparation and onward reaction have both been studied using this reactor. The preparation of 21 was achieved by reacting 5,5-dimethylhydantoin 35 with ICl in a basic biphasic medi- um (Fig. 14b). In this case, the optimal oscillation frequency Fig. 14 Solid-liquid applications of a MJOD reactor. (a) Hofmann rearrangement using a slurry feed of starting material 33 (b) Preparation of iodinating reagent 21, formed as a solid during the reaction (c)NaBH reduction of nitro groups using an in-situ generated heterogeneous Fig. 13 Heterogenous acid-catalyzed acetylation of camphene in a meso- catalyst COBR 486 J Flow Chem (2020) 10:475–490 Another recent application of solid handling in OFRs is at the interface between photochemistry, chemical engineering and organic chemistry. The use of heterogeneous catalysts in photochemistry has attracted attention as a replacement for expensive and toxic rare metal homogeneous catalysts, which are difficult to recycle [128–130]. Coupled with continuous flow technology, heterogeneous photochemistry becomes more reliable, safer and greener [131]. Therefore, using OFRs for heterogeneous photocatalysis is a perfect match. A study by Kappe, Pieber et al. demonstrated the handling of solids in a baffleless OFR for C-N coupling reactions cat- alyzed by a carbon nitride semiconductor photocatalyst (CN- OA-m) coupled with NiBr .3H O(Fig. 16a)[132]. The OFR 2 2 (HANU reactor, Creaflow) consists of a series of cubic static mixing elements, providing split-and-recombine mixing (Fig. 3c). This facilitated a stable suspension in combination with an oscillatory diaphragm pulsator unit. In this case, it was essential to include a pulsation dampener and BPR to prevent suction of air or cavitation during backward pulsation. A series of RTD measurements allowed quantification of the plug flow character, using the corresponding Fig. 15 Solid formation reactions in a coiled baffleless OFR (a) Bodenstein numbers for each set of oscillation parameters. precipitation reaction for the synthesis of diimine 40 (b)Liquid−liquid The amplitude and frequency of the pulsation were opti- phase transfer catalyzed reaction for the synthesis of gem- dichlorocyclopropane 44, with solid by-product formation mized to ensure reaction efficiency, while maintaining a stable suspension of photocatalyst. At low frequencies and amplitudes, quantitative yields of 47 were obtained reaction was transposed to flow, pulsations were required to from the model reaction between 4-bromobenzoate (45) avoid clogging. A yield between 39 and 48% was obtained, and pyrrolidine (46). It was shown that the photocatalyst depending on the reagent concentrations. Using these condi- could be recycled over ten cycles without significant tions, a long run of 5 h was performed without any clogging. changes in the yield. The stability, robustness and scal- Lowering the oscillatory frequency and amplitude, to mini- ability of the protocol were also proven by performing a −1 mize the energy dissipation rate, had no detrimental impact long runof4.5hand isolating>12 g(2.67gh )of on the handling of the solid. aniline 47.Finally,anintermediate(49) in the synthesis The second reaction was a liquid-liquid reaction between of the API tetracaine was synthesized with a productivity −1 dichlorocarbene, formed by the deprotonation of chloroform of 1.12 g h after a slight reoptimization of the reaction (42) with NaOH, and cyclohexene (41)toform7,7- conditions (Fig. 16b). dichlorobicyclo[4.1.0]-heptane (44) and solid NaCl (Fig. The most recent example reported for slurry handling in an 15b). This reaction requires a phase transfer catalyst, OFR also considers a photochemical transformation performed diethylmethylamine (43), to transfer the formed in the commercial HANU reactor, reported by Debrouwer et al. dichlorocarbene between the aqueous and organic phases. [133]. Experimental and computational characterization of the This reaction had already been reported under continuous oscillatory flow regime in this type of OFR were established flow conditions in a packed bed reactor, but had a narrow and showed that the implementation of high oscillation ampli- operating window due to problems with clogging [127]. tudes increases axial mixing. Similar trends to those determined Using the previously optimized conditions, clogging rapidly by the Kappe group for the RTD under different oscillatory occurred because of a side-reaction leading to gas formation conditions were also obtained [132]. The adaptation of a hetero- 2 3 and oscillation dampening. geneous nickel/photoredox catalyzed C(sp )-C(sp ) cross- Characterization of the flow conditions showed that in- electrophile coupling to flow was assessed. Coupling an aryl creasing the oscillation frequency led to a broadening of the bromide (45) with an alkyl bromide (50) provided the product RTD. However, Dean secondary flows (Table 1)weregener- 51 (Fig. 17). Despite the excellent suspension properties of the ated, which minimized solid stagnation and deposition. The OFR, sedimentation of the solid Na CO base occured in the 2 3 particle size distribution was larger under flow than batch supply tubing to the pump. Decreasing the particle size by wet conditions. When the oscillation frequency was increased, milling overcame this issue, allowing stable processing, with a −1 smaller and less uniform particles were obtained. high productivity of 51 (0.87gh ). J Flow Chem (2020) 10:475–490 487 Fig. 16 C-N coupling reaction photocatalyzed by a carbon nitride semiconductor (CN-OA-m) and NiBr .3H O catalyst in a microstructured OFR (a) 2 2 optimized conditions for cross-coupling of 4-bromobenzoate 45 and pyrrolidine 46 as a model reaction (b) synthesis of tetracaine precursor 49 Conclusions compromise between efficacious particle suspension and minimal axial dispersion has to be found during reaction Since the addition of an oscillatory flow to a net flow im- optimization. proves heat- and mass transfer, OFRs have already been used From this overview, it is clear that OFRs broaden the possi- to overcome issues encountered by conventional flow reac- bilities to tackle chemical challenges under continuous flow con- tors. Liquid-liquid biphasic reactions that need high mass ditions. With this in mind, OFRs are certain to become a key transfer have been effectively implemented under oscillatory element in the flow chemist’s toolbox, particularly for slow or flow conditions. Biodiesel production in particular, which in- multiphasic reactions. The performance of OFRs can be weaker volves immiscible phases and long residence times, has been at high net flow rates or outside of the droplet flow regimes for studied in-depth for over 20 years. Rational design has en- liquid-liquid reactions. Moreover, the compressibility of gases hanced the assets of such reactors towards optimal baffle appears to discourage their use for such applications, because structures for large scale operation. The decoupling of mixing of oscillation dampening. Nevertheless, the diversity of baffle and residence time has also been recently exploited for other geometries renders these reactors suitable for different purposes, reactions in liquid-liquid systems, for example in small scale while oscillation conditions represent novel process parameters reaction screening. to consider and optimize. In this way, OFRs enable a smooth In addition, the appealing profile of these reactors for transition for a wider range of chemistries from batch to flow, and solid-liquid reactions has first been tested in packed meso- can facilitate scale-up based on maintaining the system’s hydro- COBRs. A deeper understanding of the effect of oscillations dynamic characteristics. It is anticipated that new OFR configu- on the handling of solids, thanks to recent studies concerning rations will be developed in the coming years, to tackle the han- axial dispersion, has enabled scientists to efficiently perform dling of dense solids and high solid loadings, which remain reactions with suspended solid particles. Nevertheless, a challenging for flow processing. 2 3 Fig. 17 Heterogeneous photocatalytic C(sp )-C(sp ) cross-electrophile coupling of aryl bromide 45 with alkyl bromide 50 in a microstructured OFR 488 J Flow Chem (2020) 10:475–490 Acknowledgements Open access funding provided by University of 17. Ni X, Mackley MR, Harvey AP, Stonestreet P, Baird MHI, Rama Graz. Rao N v. (2003). Chem Eng Res Des 81:373–338 18. Ricardo C, Ni X (2009). Org Process Res Dev 13:1080–1087 Code availability Not applicable. 19. McDonough JR, Phan AN, Harvey AP (2018). Chem Eng and Process Process Intensif 129:51–62 20. Mohd Rasdi FR, Phan AN, Harvey AP (2013). Chem Eng J 222: Funding information The CC FLOW Project (Austrian Research 282–291 Promotion Agency FFG No. 862766) is funded through the Austrian 21. McDonough JR, Phan AN, Reay DA, Harvey AP (2016). Chem COMET Program by the Austrian Federal Ministry of Transport, Eng Process Process Intensif 110:201–213 Innovation and Technology (BMVIT), the Austrian Federal Ministry of 22. Sleveland D, Bjørsvik HR (2012). Org Process Res Dev 16:1121– Science, Research and Economy (BMWFW), and by the State of Styria (Styrian Funding Agency SFG). P.B. acknowledges support by the 23. Bjørsvik HR, Liguori L (2014). Org Process Res Dev 18:1509– University of Liege through the Erasmus+ funding program. 24. Drageset A, Elumalai V, Bjørsvik HR (2018). React Chem Eng 3: Data availability Not applicable. 550–558 25. Lucas MS, Reis NM, Li Puma G (2016). Chem Eng J 296:335– Compliance with ethical standards 26. Navarro-Fuentes F, Keane M, Ni X (2019). Org Process Res Dev 23:38–44 Conflicts of interest/competing interests Not applicable. 27. Graça CAL, Lima RB, Pereira MFR, Silva AMT, Ferreira A (2020). Chem Eng J 389:124412 Open Access This article is licensed under a Creative Commons 28. Spaccini R, Liguori L, Punta C, Bjørsvik HR (2012). Attribution 4.0 International License, which permits use, sharing, adap- ChemSusChem 5:261–265 tation, distribution and reproduction in any medium or format, as long as 29. McDonough JR, Murta S, Law R, Harvey AP (2019). Chem Eng J you give appropriate credit to the original author(s) and the source, pro- 358:643–657 vide a link to the Creative Commons licence, and indicate if changes were 30. Slavnić DS, Živković L v., Bjelić A v., Bugarski BM, Nikačević made. The images or other third party material in this article are included NM (2017). J Chem Technol Biotechnol 92:2178–2188 in the article's Creative Commons licence, unless indicated otherwise in a 31. Smith KB, Mackley MR (2006). Chem Eng Res Des 84:1001– credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by 32. Phan AN, Harvey A (2010). Chem Eng J 159:212–219 statutory regulation or exceeds the permitted use, you will need to obtain 33. Zheng M, Mackley M (2008). Chem Eng Sci 63:1788–1799 permission directly from the copyright holder. To view a copy of this 34. Reis N, Vicente AA, Teixeira JA (2010). Chem Eng Process licence, visit http://creativecommons.org/licenses/by/4.0/. Process Intensif 49:793–803 35. Ahmed SMR, Phan AN, Harvey AP (2017). Chem Eng Technol 40:907–914 36. Mazubert A, Fletcher DF, Poux M, Aubin J (2016). Chem Eng References Process Process Intensif 108:78–92 37. McDonough JR, Ahmed SMR, Phan AN, Harvey AP (2017). Chem Eng Sci 171:160–178 1. Rehm TH (2019). ChemPhotoChem 3:1–21 38. Millor GJ, Lee YC, Ni X (2019). Comp Chem Eng 124:14–27 2. Atobe M, Tateno H, Matsumura Y (2018). Chem Rev 118:4541– 39. Nogueira X, Taylor BJ, Gomez H, Colominas I, Mackley MR (2013). Comp Chem Eng 49:1–17 3. Cambié D, Bottecchia C, Straathof NJW, Hessel V, Noël T 40. Phan AN, Harvey AP (2012). Chem Eng J 180:229–236 (2016). Chem Rev 116:10276–11034 41. McDonough JR, Oates MF, Law R, Harvey AP (2019). Chem 4. Gutmann B, Cantillo D, Kappe CO (2015). Angew Chem - Int Ed Eng J 361:508–518 54:6688–6728 42. Phan AN, Harvey A, Lavender J (2011). Chem Eng Process 5. Bezerra MA, Lemos VA, de Oliveira DM, Novaes CG, Barreto Process Intensif 50:254–263 JA, Alves JPS, Cerqueira UMF da M, Santos QO dos, Araújo SA 43. Reis N, Harvey AP, Mackley MR, Vicente AA, Teixeira JA (2020). Microchem J 155:10473 (2005). Chem Eng Res Des 83:357–371 6. Baumann M (2018). Org Biomol Chem 16:5946–5954 44. González Niño C, Kapur N, King MF, de Boer G, Blacker AJ, 7. Britton J, Raston CL (2017). Chem Soc Rev 46:1250–1271 Bourne R, Thompson H (2019). Int J Comput Fluid Dyn 33:317– 8. Trojanowicz M (2020). Molecules 25:1434 9. di Filippo M, Bracken C, Baumann M (2020). Molecules 25:356 45. Reintjens RWEG, Thathagar M (2011) Oscillating flow 10. Hommes A, Heeres HJ, Yue J (2019). ChemCatChem 11:4671– minireactor. US patent 10,118,149 B2 (Jan 13, 2011) 46. Jähnisch K, Hessel V, Löwe H, Baerns M (2004). Angew Chem - 11. Sui J, Yan J, Liu D, Wang K, Luo G (2019). Small 1902828:1–23 Int Ed 43:406–446 12. Berton M, de Souza JM, Abdiaj I, McQuade DT, Snead DR 47. Mielke E, Roberge DM, Macchi A (2016). J Flow Chem 6:279– (2020). J Flow Chem 10:73–92 13. Stonestreet P, Harvey AP (2002). Chem Eng Res and Des 80:31– 48. Reis N, Pereira RN, Vicente AA, Teixeira JA (2008). Ind Eng Chem Res 47:7190–7201 14. McGlone T, Briggs NEB, Clark CA, Brown CJ, Sefcik J, Florence 49. Law R, Ahmed S, Tang N, Phan A, Harvey A (2018). Chem Eng AJ (2015). Org Process Res Dev 19:1186–1202 Process Process Intensif 125:133–138 15. Abbott MSR, Harvey AP, Valente Perez G, Theodorou MK 50. González-Juárez D, Herrero-Martín R, Solano JP (2018). Appl (2013). Interface Focus 3:20120036 Therm Eng 141:494–502 16. McDonough JR, Phan AN, Harvey AP (2015). Chem Eng J 265: 51. Solano JP, Herrero R, Espín S, Phan AN, Harvey AP (2012). 110–121 Chem Eng Res Des 90:732–742 J Flow Chem (2020) 10:475–490 489 52. Ahmed SMR, Phan AN, Harvey AP (2018). Chem Eng Process 85. Liguori L, Bjørsvik HR (2011). Org Process Res Dev 15:997– Process Intensif 130:229–239 1009 53. Jimeno G, Lee YC, Ni X (2018). Chem Eng Process Process 86. Drageset A, Frøystein NÅ, Törnroos KW, Bjørsvik HR (2019). Intensif 134:153–162 React Chem Eng 4:41–51 54. Avila M, Fletcher DF, Poux M, Xuereb C, Aubin J (2020). Chem 87. Schoenitz M, Grundemann L, Augustin W, Scholl S (2015). Eng Sci 212:115310 Chem Commun 51:8213–8228 55. Ni X, Zhang Y, Mustafa I (1998). Chem Eng Sci 53:2903–2919 88. Kashani S, Sullivan RJ, Andersen M, Newman SG (2018). Green 56. Ni X, Zhang Y, Mustafa I (1999). Chem Eng Sci 54:841–850 Chem 20:1748–1753 89. Heider PL, Born SC, Basak S, Benyahia B, Lakerveld R, Zhang 57. Ni X, Johnstone JC, Symes KC, Grey BD, Bennett DC (2001). AIChE 47:1746–1757 H, Hogan R, Buchbinder L, Wolfe A, Mascia S, Evans JMB, Jamison TF, Jensen KF (2014). Org Process Res Dev 18:402–409 58. Ni X, Murray KR, Zhang Y, Bennett D, Howes T (2002). Powder 90. Hartman RL (2012). Org Process Res Dev 16:870–887 Technol 124:281–286 59. Harvey AP, Mackley MR, Stonestreet P (2001). Ind Eng Chem 91. Wang Z, Gérardy R, Gauron G, Damblon C, Monbaliu JCM (2019). React Chem Eng 4:17–26 Res 40:5371–5377 60. Mohd GATI, Resul MFMG, Yunus R, Yaw TCS (2008). Sci 92. Colella M, Carlucci C, Luisi R (2018). Top Curr Chem 376:46 Technol 3:138–145 93. Munirathinam R, Huskens J, Verboom W (2015). Adv Synth 61. Zheng M, Skelton RL, Mackley MR (2007). Process Saf Environ Catal 357:1093–1123 Prot 85:365–371 94. Rodríguez-Escrich C, Pericàs MA (2018). Chem Rec 18:1–20 62. Harvey AP, Mackley MR, Seliger T (2003). J Chem Technol 95. Pomberger A, Mo Y, Nandiwale KY, Schultz VL, Duvadie R, Biotechnol 78:338–341 Robinson RI, Altinoglu EI, Jensen KF (2019). Org Process Res 63. Phan AN, Harvey AP, Rawcliffe M (2011). Fuel Process Technol Dev 23:2699–2706 92:1560–1567 96. Mo Y, Jensen KF (2016). React Chem Eng 1:501–507 64. Sagmeister P, Poms J, Williams JD, Kappe CO (2020). React 97. Chapman MR, Kwan MHT, King G, Jolley KE, Hussain M, Chem Eng 5:677–684 Hussain S, Salama IE, González Nino C, Thompson LA, 65. Hone CA, Holmes N, Akien GR, Bourne RA, Muller FL (2017). Bayana ME, Clayton AD, Nguyen BN, Turner NJ, Kapur N, React Chem Eng 2:103–108 Blacker AJ (2017). Org Process Res Dev 21:1294–1301 66. Moore JS, Jensen KF (2014). Angew Chemie - Int Ed 53:470–473 98. Filipponi P, Gioiello A, Baxendale IR (2016). Org Process Res 67. Phan AN, Harvey AP, Eze V (2012). Chem Eng Technol 35: Dev 20:371–375 99. Browne DL, Deadman BJ, Ashe R, Baxendale IR, Ley SV (2011). 1214–1220 68. Al-Saadi LS, Alegría A, Eze VC, Harvey AP (2019). Chem Eng Org Process Res Dev 15:693–697 Technol 42:539–548 100. Falß S, Tomaiuolo G, Perazzo A, Hodgson P, Yaseneva P, Zakrzewski J, Guido S, Lapkin A, Woodward R, Meadows RE 69. Al-Saadi LS, Eze VC, Harvey AP (2020). Biofuels. https://doi. org/10.1080/17597269.2020.1727688 (2016). Org Process Res Dev 20:558–567 101. Noël T, Naber JR, Hartman RL, McMullen JP, Jensen KF, 70. Mazubert A, Aubin J, Elgue S, Poux M (2014). Green Process Buchwald SL (2011). Chem Sci 2:287–290 Synth 3:419–429 71. Mazubert A, Crockatt M, Poux M, Aubin J, Roelands M (2015). 102. Cao Q, Howard JL, Crawford DE, James SL, Browne DL (2018). Green Chem 20:4443–4447 Chem Eng Technol 38:2161–2169 72. Lobry E, Lasuye T, Gourdon C, Xuereb C (2015). Chem Eng J 103. Isoni V, Mendoza K, Lim E, Teoh SK (2017). Org Process Res Dev 21:992–1002 259:505–518 73. Abolhasani M, Bruno NC, Jensen KF (2015). Chem Commun 51: 104. Sharma BM, Atapalkar RS, Kulkarni AA (2019). Green Chem 21: 8916–8919 5639–5646 105. Crawford DE, Miskimmin CK, Cahir J, James SL (2017). Chem 74. Abolhasani M, Oskooei A, Kumacheva E, Günther A (2014). Lab Chip 14:2309–2318 Commun 53:13067–13070 106. Bannock JH, Krishnadasan SH, Nightingale AM, Yau CP, Khaw 75. Gielen F, Van Vliet L, Koprowski BT, Devenish SRA, K, Burkitt D, Halls JJM, Heeney M, de Mello JC (2013). Adv Fischlechner M, Edel JB, Niu X, deMello AJ, Hollfelder, F Funct Mater 23:2123–2129 (2013). Anal Chem 85:4761–4769 107. Poe SL, Cummings MA, Haaf MP, McQuade DT (2006). Angew 76. Abolhasani M, Coley CW, Xie L, Chen O, Bawendi MG, Jensen KF (2015). Chem Mater 27:6131–6138 Chem - Int Ed 45:1544–1548 108. Slavnić D, Bugarski B, Nikačević N (2019). Chem Eng Process 77. Shen Y, Abolhasani M, Chen Y, Xie L, Yang L, Coley CW, Process Intensif 135:108–119 Bawendi MG, Jensen KF (2017). Angew Chem - Int Ed 56: 16333–16337 109. Ejim LN, Yerdelen S, McGlone T, Onyemelukwe I, Johnston B, Florence AJ, Reis NM (2017). Chem Eng J 308:669–682 78. Coley CW, Abolhasani M, Lin H, Jensen KF (2017). Angew Chem - Int Ed 56:9847-9850 110. Kacker R, Regensburg SI, Kramer HJM (2017). Chem Eng J 317: 79. Hwang YJ, Coley CW, Abolhasani M, Marzinzik AL, Koch G, 413–423 111. Cruz P, Silva C, Rocha F, Ferreira A (2019). AIChE 65:e16683 Spanka C, Lehmann H, Jensen KF (2017). Chem Commun 53: 6649–6652 112. Oliva JA, Pal K, Barton A, Firth P, Nagy ZK (2018). Chem Eng J 351:498–505 80. Hsieh HW, Coley CW, Baumgartner LM, Jensen KF, Robinson RI (2018). Org Process Res Dev 22:542–550 113. Onyemelukwe II, Nagy ZK, Rielly CD (2020). Chem Eng J 382: 81. Mongeon SS, Roberge DM, Bittel M, Elsner P, Macchi A (2016). 122862 Org Process Res Dev 20:733–741 114. Fabiyi ME, Skelton RL (1999) J Photochem Photobiol. A 129:17– 82. Rafiee SE, Sadeghiazad MM (2017). Appl Therm Eng 114:300– 24 327 115. Fabiyi ME, Skelton RL (2000). Trans IChemE 78:399–404 83. Loponov KN, Deadman BJ, Zhu J, Rielly C, Holdich RG, Hii KK, 116. Gao P, Han Ching W, Herrmann M, Kwong Chan C, Yue PL Hellgardt K (2017). Chem Eng J 329:220–230 (2003). Chem Eng Sci 58:1013–1020 84. Liguori L, Bjørsvik HR (2007) Continuous flow reactor. World 117. Eze VC, Phan AN, Pirez C, Harvey AP, Lee AF, Wilson K (2013). patent WO 2007/058544 A1, Nov 17, 2006 Catal Sci Tech 3:2373–2237 490 J Flow Chem (2020) 10:475–490 118. Eze VC, Fisher JC, Phan AN, Harvey AP (2017). Chem Eng J 127. von Keutz T, Cantillo D, Kappe CO (2019). J Flow Chem 9:27–34 322:205–214 128. Gisbertz S, Pieber B (2020). ChemPhotoChem 4:1–21 119. Eze VC, Harvey AP (2018). Chem Eng J 347:41–51 129. Riente P, Noël T (2019). Catal Sci Technol 9:5186–5232 120. Kefas HM, Yunus R, Rashid U, Taufiq-Yap YH (2019). J Environ 130. Savateev A, Ghosh I, König B, Antonietti M (2018). Angew Chem Eng 7:102993 Chem - Int Ed 57:15936–15947 121. Okafor O, Weilhard A, Fernandes JA, Karjalainen E, Goodridge 131. Williams JD, Kappe CO (2020). Curr Opin Green Sustain R, Sans V (2017). React Chem Eng 2:129–136 Chem 25:100351 122. Zheng H, Yan Z, Chu S, Chen J (2018). Chem Eng Process 132. Rosso C, Gisbertz S, Williams JD, Gemoets H, Debrouwer W, Process Intensif 134:1–8 Pieber B, Kappe CO (2020). React Chem Eng 5:595–604 123. Ferreri M, Drageset A, Gambarotti C, Bjørsvik HR (2016). React 133. Debrouwer W, Kimpe W, Dangreau R, Huvaere K, Gemoets Chem Eng 1:379–386 HPL, Mottaghi M, Kuhn S, Van Aken K (2020). Org Process 124. Drageset A, Bjørsvik HR (2016). React Chem Eng 1:436–444 Res Dev. https://doi.org/10.1021/acs.oprd.0c00150 125. Lundevall FJ, Elumalai V, Drageset A, Totland C, Bjørsvik HR (2018). Eur J Org Chem 2018:3416–3425 126. Doyle BJ, Gutmann B, Bittel M, Hubler T, MacChi A, Roberge Publisher’snote Springer Nature remains neutral with regard to jurisdic- DM (2020). Ind Eng Chem Res 59:4007–4019 tional claims in published maps and institutional affiliations.

Journal

Journal of Flow ChemistrySpringer Journals

Published: Sep 28, 2020

There are no references for this article.