Nanofiltration‐Enabled In Situ Solvent and Reagent Recycle for Sustainable Continuous‐Flow Synthesis

Abstract Solvent usage in the pharmaceutical sector accounts for as much as 90 % of the overall mass during manufacturing processes. Consequently, solvent consumption poses significant costs and environmental burdens. Continuous processing, in particular continuous‐flow reactors, have great potential for the sustainable production of pharmaceuticals but subsequent downstream processing remains challenging. Separation processes for concentrating and purifying chemicals can account for as much as 80 % of the total manufacturing costs. In this work, a nanofiltration unit was coupled to a continuous‐flow rector for in situ solvent and reagent recycling. The nanofiltration unit is straightforward to implement and simple to control during continuous operation. The hybrid process operated continuously over six weeks, recycling about 90 % of the solvent and reagent. Consequently, the E‐factor and the carbon footprint were reduced by 91 % and 19 %, respectively. Moreover, the nanofiltration unit led to a solution of the product eleven times more concentrated than the reaction mixture and increased the purity from 52.4 % to 91.5 %. The boundaries for process conditions were investigated to facilitate implementation of the methodology by the pharmaceutical sector.


Introduction
Continuous-flow reactors can provide cost savings, facile automation, ah igher level of consistency and certainty,i mproved overall safety,a nd ar educed carbonf ootprint. [1a-c] The application of polymer-supported catalysts in flow reactors have been widely researched in recent years because 1) the use of ap olymer support eliminates the need for the removal of the catalyst from the desired product, and 2) reuse is built into the process.H owever,t he downstream processing (i.e. concentration, purification,a nd isolation)i nc ontinuousp roductionr emains challenging. Recent efforts demonstrate the need for the development of integrated continuous synthesispurification processes. [2a-b] It has recently been realized that traditional downstream separation processes can accountf or as much as two thirds of the total manufacturing costs, and eventually contribute half of the industrial energy usage. [3a-b] Solvent usage in the pharmaceutical industry accounts for 60 %o ft he total energy consumption required for the production of active pharmaceutical ingredients. [4] From another perspective, solvents account for8 0-90 %o ft he overall mass during the manufacturing processes in pharmaceutical industries. [5] Continuous-flow reactors can only be truly sustainable if solvent recovery is employed. [6a-b] Solvent waste has al arge carbon footprint and consequently it is detrimental to the sustainability of the process. Organic solvent nanofiltration (OSN) is ag reen technology, [7] that has been used for solvent recovery. [8a-g] OSN can be coupled to continuous processes for in situ separations. [8g, 9] This approachi si nl ine with the "Roadmap to ar esource efficient Europe"instigated by the European Commission, seeking to eliminate waste throughd eliberate process designw ith resource efficiency and recycling in mind. [10] With profitm argins growingt hin, there is an imperative for minimizing both the cost and environmental impact through process intensification. As effective tools, continuous processing [11a,b] andm embrane separations [3b, 7, 12] have been recognized as core technologiesthat can enable green process engineering. The newest generation of nanofiltration membranes have stable performance in aggressive media, including extreme pH solutionsa nd organic solvents, [7,13] which creates new possibilities for continuous solvent recovery,p roduct purification, and concentration.F igure 1s hows the positive attributes of continuous-flow synthesis and nanofiltrationa nd how they reinforce each other.T hese technologieso ffer the opportunity towards mooth integrationw ith further unit operations.
Solventu sage in the pharmaceutical sector accounts for as much as 90 %o ft he overall mass during manufacturing processes. Consequently,s olvent consumption poses significant costs and environmental burdens.C ontinuous processing, in particularc ontinuous-flow reactors, have great potential for the sustainable production of pharmaceuticals but subsequent downstream processing remains challenging. Separation processes for concentratinga nd purifying chemicals can account for as much as 80 %o ft he total manufacturing costs. In this work, an anofiltration unit wasc oupledt oac ontinuous-flow rector for in situ solventa nd reagent recycling. The nanofiltra-tion unit is straightforward to implement and simple to control during continuous operation. Theh ybrid process operated continuously over six weeks,r ecycling about9 0% of the solvent and reagent. Consequently,t he E-factora nd the carbon footprint were reduced by 91 %a nd 19 %, respectively.M oreover,t he nanofiltration unit led to as olution of the product eleven times more concentrated than the reactionm ixture and increased the purity from 52.4 %t o9 1.5 %. The boundaries for processc onditions were investigated to facilitate implementation of the methodology by the pharmaceutical sector.
[a] T. Fodi, + C. However,a ttempts for the synergistic coupling of flow reactors with nanofiltration are scarce (Table 1a nd references therein). In this context nanofiltrationw as used for the recovery of homogeneous catalysts as well as for solvente xchange. Herein the development of an anofiltration-enabled in situ solvent and reagent recycling process to improvet he sustainability of flow reactors is presented.
Michael Addition of nitromethane to trans-chalcone was selected as am odel reaction (Scheme 1), since polymersupported organocatalysts have attracted agreat deal of attention in continuous-flow Michael Addition processes. [16] Furthermore, OSN was used for homogeneous catalystr ecycling in a Michael Addition. [17] Hodge et al. reported the first application of aw eak anion-exchange resin Amberlyst A21 as ap olymersupported catalyst to promote Michael Addition in continuousflow conditions, [1c] which inspiredt he catalyst screening dis-closed herein. The most efficient catalysts were tested in a continuous-flow packed-bed reactor.S ubsequentlyan anofiltration membrane unit was connected to the reactor to allow in situ solvent and reagent recycling, and to consequently improve the sustainability of the synthesis.

Results and Discussion
Optimization of the flow reactor Stable,l ong-term operation of continuous hybrid processes require an efficient catalyst/solvent system for the chosen Michael Addition. These atom-efficient reactions are usually promoted by acids, bases, metal salts, or organocatalysts. With application of solid-base catalysts, formation of undesirable side-products resulting from polymerization, bis-addition,a nd self-condensation could be prevented and salt formation following the neutralization of soluble bases with acids could be avoided. Amberlyst A21, ap olymer-supported weak anion-exchange resin featuring ad ialkylbenzylamine base was reported to be ac heap, commercial catalystf or Michael Addition reactions. [1c] The performance of this catalyst was tested in numerous solvents at varying temperatures (Table2). Using ethyl   [15] Catalyst retentionPEEK Heckc oupling PFR-m-CSTR DMF Pd complex [a] This workS olvent and reagentr ecyclingDuramem 150 Michael Addition PBR Acetone Trialkylamine base [b] [a] Homogeneous.
Scheme1.MichaelA ddition of nitromethanereagent( 5equiv,60mm)to chalcone substrate (1 equiv,12mm)catalyzed by polymer-supported trialkylamine base, yielding 4-nitro-1,3-diphenylbutan-1-one product.T he molecular weights( in gmol À1 )a re given in parentheses for each compound. Ar apid decline in the catalytic activity of Amberlyst A21 was observedd uring flow reactor conditions. Although 100 %c onversion was initially achievedi n9 .3 min residence time (representing 42 gL À1 h À1 space-time yield for the reactor), it decreasedb elow 50 %w ithin two days of operation. Figures 2 and 3s how the IR spectra for the catalysti nactivation process and the underlying mechanism, respectively.T he CÀNs tretching vibrations assigned to aliphatic amine end-groups of trialkylamine catalyst constituents were observed as mediumstrength bands at 1203cm À1 (Figure 2A and C). After catalyst inactivation theseb ands disappeared and the NÀOs tretching vibrations of aliphatic nitro group were observed at 1576 cm À1 (asymmetrical) and 1371 cm À1 (symmetrical)a si ntenseb ands ( Figure 2B and D). The Michael Addition reaction is promoted by the basic character of the N,N-dimethyl N-benzylamino group (pK b = 5.02;c onjugated acidp K a = 8.98) of the catalyst and the CÀH-acidic nature of the reagent (pK a = 10.2). The basic tertiary amino group deprotonates nitromethane through an equilibrium reactiont hat resultsi nt he formation of an itronate intermediate. This ambident nucleophile attacks the soft-electrophilic center of the substrate. Under mild conditions (low temperature and/ors hort residence time;l ow amount of nitromethane) this nucleophilic addition is kinetically favorable according to the hard and soft acid-base (HSAB) principle. However,a fter longer residence time, the excess of nitromethane can promote as ide-reaction. The nitronate group can attack the electrophilic a-position of the resin, which results in the formationo faprotonated dimethylamino group. This good leaving group can undergo an ucleophilic substitution on ar esin functional group ( Figure 3). In the case of Amberlyst A21, the benzyl group promotes this reaction, because it uses the p system of the benzene ring in conjugation with the p orbital that stabilizes the transition state. On the other hand, Amberlite IRA67 lacks this stabilizing conjugation for the transition state that hinders this side-reaction. The underlyingt heory supports the fact that inactivation of the Amberlite IRA67 resin occurred at 60 8C, which is 10 8Ch igher than the threshold inactivation temperature for the Amberlys-tA21 resin.
The AmberliteIRA67 catalystw as further characterized in the flow reactor with respectt ot he effect of temperature and reagent excesso nc onversion ( Figure 4). After 8h of continuous operation, stable 80 %a nd 88 %r eactionc onversion was achieved during seven days of runtimea t3 0a nd 40 8C, respectively.F ull conversion was reached at 50-70 8Cw ithin 4h of operation. However,c atalysti nactivation was observed within 2-4 days at 60-70 8C. The attained conversion also heavily depends on the amounto fr eagent employed;7 0%,9 0%,a nd 100 %c onversion was obtained at 50 8Cw ith the use of 3, 4, and 5equiv of reagent, respectively.C onsequently,5 0 8Ca nd 5equiv of reagent were selectedf or the development of the continuous hybrid process.
Since the reagent is appliedi nh igh excessand the substrate is the limiting species, the reactionc an be assumed to follow pseudo-first order kinetics. [6b] Consequently,t he total reaction rate is described in Equation (1), and the application of t ¼ 0 and C substrate ¼ C 0 substrate boundaryc onditions gives Equation (2).  where r is the total reaction rate, k' is the pseudo-first-order reaction rate coefficient (in s À1 ), and C is the concentration of substrate (in mol L À1 ). The experimental resultsg iven in Figure 5s how excellent correlation with the first-order kinetic model that validatet he assumed linear relationship between lnðC substrate Þ and t.R efer to the Supporting Information for the conversionsa td ifferent flowrates. The reactionr ate coefficient was determined at different temperatures to estimate the apparenta ctivation energy (E a )a nd pre-exponential factor (A)f or the reaction ( Figure 5B)u sing the Arrhenius equation [Equation (3)]: where R is the ideal gas constant of 8.314 JK À1 mol À1 and T is the temperaturei nK .T he apparent activation energy was found to be 65 kJ mol À1 ,w hich falls within the 19-76kJmol À1 range reported for the same Michael Addition catalyzed by cinchona-urea-basedo rganocatalysts. [18] The obtained kinetic parameters can be used for process modelling toward automation.

Membranescreening for the optimization of the separation
The nanofiltration unit coupledt ot he flow reactor has at wofold role:1 )recycle the solvent and the reagent, and simultaneously 2) concentrate the product duringc ontinuous operation. To this effect, Duramem 150, GMT-oNF-2,S olSep NF030705, andp olybenzimidazole (PBI) solvent-resistant nanofiltration membranesw ere screened at 10-40 bar pressure ( Figure 6). Solute rejections [Equation (4)] were obtained takingi nto account the retentate and permeate concentrations. Permeance [Equation (5)] showingt he volume of liquid that permeates the membrane( V permeate )p er membrane area (A m ), per unit of time (t), and per applied pressure (p)w as also measured. High  product rejection prevents undesired recycling backi nto the reactor and subsequenta ccumulation, which can lead to decreases in productivity.O nt he contrary,l ow reagent rejection is neededt o1 )minimize product contamination in the retentate streaml eadingt oh igher product purity,a nd 2) minimize waste generation by maximizing the reuse of the reagent. Duramem150 membrane operating at 40 bar pressure was selected for the continuous process because these conditions exhibited the highest rejection for the product (100 %), but still maintaining low rejection for the reagent (12.2 %).

Solute rejection
Permeance Continuous-flow reactor/nanofiltrationh ybrid process The hybrid process comprised of af low reactor and an anofiltrationu nit that were filled with acetonea tthes tart of continuous operation (Figure 7). Thef eed solution containing 9.3 mm substratea nd 5equiv of reagent in acetone was allowed to pass through the packed-bed flow reactor loaded with 8.2 go f catalyst. The flowrate was set to 1mLmin À1 allowing 100 % substrate conversion after 6.25 min residence time. Consequently,t he crude product stream comprised of 1equiv of product and 4equiv of reagent.T his stream was fed into the nanofiltration unit comprising 0.021 m 2 membrane area and 35 mL volume. Ar ecirculation pump set at 1Lmin À1 ensured homogeneous solute concentration in the membrane loop. The membrane split the crude mixture into ac oncentrated, product-rich retentate stream (10% of feed flowrate), and ar eagent-rich permeate stream (90 %o ff eed flowrate). During start-up of the continuous system the permeate stream was discarded ( Figure 7A). Once steady-state permeate concentration was reached, recycling of the permeate stream through a dynamic mixingc hamberw as commenced ( Figure 7B). To compensate the changes (i.e. concentrationa nd flowrate) induced by the permeate recycle, botht he flowrate and concentrationso ft he feed were adjusted accordingly,a nd subsequently steady-state operation was maintained. The new feed flowrate was set to 0.091 mL min À1 ,w hereas the feed concentrations were changed to 102 mm substratea nd 1.4 equiv reagent. The concentration profiles of the solutes in the retentate and permeate streamsare shown in Figure 8. The experimental data shows good correlation with the predicted curves. Refer to the Supporting Information for the detailed mathematical framework describing the process modelling. The equilibrium time is defined as the time neededt or each 98 %o ft he steady-state concentration of as olute, and it was found to be 2.5 hf or the reagent in the permeate stream. This is the thresholdt os tartr ecycling the solventa nd the reagent. Until 2.5 ht he permeate was discarded ( Figure 7A)a nd afterward it was recycled back to the flow reactor through ad ynamic mixing chamber ( Figure 7B). The product leaves the flow reactor and enters the nanofiltrationunit at 2.5 gL À1 concentration, and then gets concentrated to 27.5 gL À1 ,h ence the considerably longere quilibrium time of 25 h. The hybrid process showed stable performance over six weekso fcontinuous operation, achieving catalystt urnover number (TON) of 32, catalyst turnover frequency (TOF) of 0.032 h À1 ,p roductivity of 3.75 gL À1 h À1 and9 1.5 %p roduct purity.I ts hould be noted that the membrane performance could change and fouling could occur over time, which can be mitigated by operating well below the solubility limit of the solutes. [19] During the six The effects of various system parameters-namely conversion, product and reagent rejections, retentate/permeate flowrate ratio, and membrane loop volume/area-ons olvent consumption,p roduct purity,p roductivity,a nd equilibrium time were investigated (Figure9). Them embrane loop volume/area does not affect the solvent consumption and the product purity but its decrease results in the improvement of both productivity and equilibrium time. Although minimization of this volumeisp ractically difficult using flat-sheet membranes (laboratory scale),t he membrane modules used in industrial applications have relatively low values of 0.8-1.2 Lm À2 . [20] The increase in product rejectioni sb eneficial for all process metrics. On the contrary,t he decrease in reagent rejection enhances product purity and shortense quilibrium time, but neither solvent consumption nor productivity is affected. Similar to prod-uct rejection, the higher the conversion, the bettert he process metrics become except for the equilibrium time, which remains constant. Decreasing the retentate/permeate flowrate ratio resultsi nafavorable increasei np urity and decrease in solventc onsumption whereas the productivity and equilibrium time are not affected. Figure 9F showst hat high-purity product can be obtained by minimizing the excess of the reagent and the reagent rejection. Furthermore, the lower the excess of reagent, the less influence the reagent rejection has on the product purity.R efer to the Supporting Information for the detailed sensitivity analysis.

Sustainability assessment of the hybrid process
The effect of nanofiltration-assisted in situ solvent andr eagent recycling on the sustainabilityo ft he continuous-flow synthesis was assessed. Changing from batch to continuous processing Figure 7. Process configurationand conditionsf or the start-up (A) and continuous steady-state(B) operation of the hybrid process. The jacketed flow reactor operated at 50 8Callowing 100 %c onversion of the substrate. The nanofiltration unit was operated continuouslyat4 0bar,splitting the crude mixture into a concentrated, product-rich retentate stream, and ar eagent-rich permeate stream. Initiallyt he permeate stream was discarded (A) and once steady-state was reached, recycling of the permeate stream was commenced (B), and the feed concentrationa nd flowrate were adjusted. ChemSusChem 2017, 10,3435 -3444 www.chemsuschem.org 2017 The Authors. Publishedb yWiley-VCH Verlag GmbH &Co. KGaA, Weinheim in af low synthesis process can considerably reduce the environmental burden of chemical manufacturing. [21a-d] None-theless, the use of organic solvents was identified as one of the largestc ontributor to waste generation in flow synthesis. [21a, 22a-b] TheE -factor and carbon footprint, defined in Equation (6) and Equation (7), have pivotal roles in improving resourcee fficiencya nd waste minimization in the chemical industries. [10] E À factor ¼ kg waste generated kg isolated product ð6Þ Carbon footprint ¼ equivalent kg of CO 2 kg isolated product ð7Þ Calculation of the E-factor ignores any recycled streams and reusedr eactants or reagents as they are not considered as waste. On the other hand, it neglects the energy required for the process. Consequently,c arbon footprint, taking into account waste generation and energy consumption, was also derived to obtain ah olisticv iew of the environmental burden of the hybrid process( Figure 10). Coupling of the nanofiltration unit to the flow reactora llows the recycling of 90.9 %o fs olvent and 90 %o fr eagent, corresponding to 287 kg solvent and 0.814 kg reagent per kg of product, respectively (Figure 10 A). The solventc onsumption is the main contributor to the Efactor,a nd the nanofiltrationu nit reduces waste generation and raw material usage by 90.7 %. The application of the nanofiltration unit is sustainable, [7] and in this particularc ase it  The carbon footprintc ontributiono ft he thermostat for the jacket of the flow reactor is 2257 kg kg À1 accounting for 78 % and 97 %ofthe total carbon footprint in the absence and presence of the nanofiltrationunit, respectively (Figure 10 B). In line with the 6th principle of green chemistry, [23] such ah igh contribution calls for energy integration within am anufacturing plant. The obtained values for the E-factor demonstrate that in situ solvent and reagent recycling can transform the process from the non-green category to the green category based on the industrial classification by Sheldon. [24] The carbonf ootprint of the continuous hybrid process rapidly decreases over time and reaches ac onstantv alue of 2323 kg kg À1 within less than two days of operation, and it takes 10 ht ob ecome more environmentally benign than the benchmark process (Figure 10 C). The reduction in carbon footprint depends on the solvent used in the process (Figure10D). The potentialr eductioni n the carbon footprint of nanofiltration was assessed considering the top solvent wastes [5] generated in the pharmaceutical sector.D ependingo nt he carbon footprint potential of each solvent, [25] the implementation of an anofiltration unit could result in 17-31% overall reduction.
The proposed process configuration enables further processwindowe xpansion. First, the continuous reaction-separation methodology can be used for sought-after continuous flow multi-step organic syntheses. [2,26] Second, the pressurized system allows the reaction to be realized at elevated temperatures up to supercritical conditions which is of recent interest. [27] Conclusions Increasing demand for more sustainable manufacturing has led to an increasing interesti nt he area of continuous processing within the pharmaceutical sector.F low synthesis allows continuous production of fine chemicals in as afe, compact, and controlled environment but subsequent downstream processing remains challenging. Concentration and purification of the product is usually done in batch operation, and these separationp rocesses can account for as much as 80 %o ft he total manufacturing costs. In this work, ac ontinuous hybrid process comprising of af low reactor and as ubsequent nanofiltration Figure 10. The effectofi nsitu solvent and reagentrecycling on the E-factor (A) and the carbon footprint (B, C) of the continuous-flow synthesis.T he main contributors of the environmental burden are the solvent waste and the energy to operate the thermostat for the jacket ofthe flow reactor. The rest of the contributors have cumulative contributions lower than 0.5 %. Refer to the Supporting Information for the actual valuesf or all contributors. Basedo nt he acetone solvent recovery flowrate and the densityo fthe different solvents, the potential of nanofiltration for the carbon footprint reduction can be estimated [25] for the most commons olvent wastes [5] generated in the pharmaceutical sector (D). ChemSusChem 2017, 10,3435 -3444 www.chemsuschem.org 2017 The Authors. Publishedb yWiley-VCH Verlag GmbH &Co. KGaA, Weinheim unit for in situ solvent and reagentr ecycling has been developed. The nanofiltrationu nit is straightforwardt oi mplement and controld uring continuous operation. Ap redictive in silico tool-where the main input parameters are the membrane permeance and solute rejections-was developed to understand the boundaries of process conditions. The hybrid process was operated continuously over six weeks, leading to the recovery of about 90 %o ft he solvent and the excesso ft he reagent. Consequently,t he E-factora nd the carbon footprint were reduced by 91 %a nd 19 %, respectively.T he nanofiltration unit concentrated the product by 11 times and increased the purity from 52.4 %t o9 1.5 %. In-line coupling of nanofiltration to flow reactors in continuouso peration can significantly improvet he sustainability of flow synthesis through in situ solvent and reagent recycling, as well as improving the concentration and purity of the product. The proposed process configuration enables furtherp rocess-window expansion for continuous-flow multi-step organic syntheses.

Experimental Section
Materials and general methods Chemicals (reagent grade) and solvents (analytical grade) were purchased from Sigma-Aldrich (UK) and were used without further purification. Three commercially available OSN membranes were used:S olSep NF010206, GMT-oNF-2 and DuraMem150 (DM150) from SolSep BV,B orsig Membrane Te chnology GmbH and Evonik-MET,r espectively.2 6wt% polybenzimidazole (MW = 27 000 gmol À1 )c ontaining 1.5 wt %l ithium chloride stabilizer dissolved in N,N-dimethylacetamide (DMAc) solution was purchased from PBI Performance Products Inc. (USA). Non-woven polypropylene fabric Novatexx 2471 was sourced from Freudenberg Filtration Te chnologies (Germany). 1 Ha nd 13 CNMR spectra were recorded on aB ruker AV-400 spectrometer.H PLC measurements were carried out on aV WR HPLC system equipped with aVWR 5160 pump, aVWR 5260 autosampler and thermostat, aV WR 5430 diode array detector (DAD) and an ACE 5C18 150 4.6 mm, 5 mmc olumn. The pump flowrate was set at 1mLmin À1 and the column temperature was 25 8C. The mobile phase was a1 :1 mixture of acetonitrile and water with 0.1 %t rifluoroacetic acid. Nitromethane was quantified using aV arian CP-3800 gas chromatograph (GC) fitted with aV arian Saturn 2200 mass spectrometer (MS) and aV arian VF-5 ms capillary column (30 m 0.25 mm, 0.25 mm). The GC oven temperature was 27 8Cf or 4min, then increased at 15 8Cmin À1 to 200 8C. LCMS measurements were carried out on an Agilent 1100 HPLC equipped with gradient pump, autosampler and photodiode array (PDA) detector.Atriple quadrupole mass spectrometer with positive electrospray ionization source was employed as the MS detector.I nfrared spectra were recorded on aB ruker Alpha-T FTIR spectrometer.

Continuous-flow reactions
AS upelco stainless steel column of 10 mm 250 mm dimension was loaded with as olid base Amberlyst A21 (13.1 g) free-base form or Amberlite IRA67 (8.1 g) free-base form, respectively.P rior to use the filled column was washed with the reaction solvent (50 mL) with af lowrate of 1.0 mL min À1 to remove air and unwanted materials from the surface of the solid base. Af eed solution containing trans-chalcone (12 mm)a nd nitromethane (1-5 equiv,1 2-60 mm) was pumped through the column at 1-10 mL min À1 flowrate corresponding to 6.25-0.625 min residence time at 30-70 8Ct emperature. AL auda Alpha RA12 thermostat was used to control the temperature with AE 0.1 8Ca ccuracy in the jacketed flow reactor.T he energy consumption of the equipment was measured with aF luke 1736 power logger with resolution 10 mA and accuracy AE 0.1 %.

Membranescreening
The polybenzimidazole (PBI) membrane was prepared according to our previously reported protocol. [8f] The feed solution for the membrane screening comprised of am ixture of substrate, reagent, and product, each at 1gL À1 concentration in acetone. The pressure range for the screening was 10-40 bar (1 bar = 0.1 MPa) and the tests were carried out in duplicate in ac ross-flow nanofiltration rig. [8f] The feed solution was recirculated for 24 hf ollowed by collection of samples from the permeate and the retentate streams. The solute rejection values for the substrate and product were determined by LCMS, whereas GCMS was used for the reagent. Refer to the Supporting Information for the membrane screening process scheme.