Macroporous Gel with a Permeable Reaction Platform for Catalytic Flow Synthesis.

We mimic a living system wherein target molecules permeate through capillary and cells for chemical transformation. A monolithic porous gel (MPG) was easily prepared by copolymerization of gel matrix, tertiary amine, and cross-linking monomer in one-step synthesis. Interconnected capillaries existed in the MPG, enabling flow application with high permeability. Because the capillaries were constituted of polymer gel, Pd(0)-loaded MPG provided another permeable pathway to substrates in a gel network, contributing to its much high turnover number after 30 days of use, compared with that of Pd(0)-loaded inorganic supports. Interestingly, the gel network size of the MPG influenced the catalytic frequency. Diffusivities of the substrates and product in the gel networks increased with increasing network sizes in relation to catalytic activities. The MPG strategy provides a universal reactor design in conjunction with a practical process and precisely controlled reaction platform.

Introduction

Process innovations for chemical syntheses impact growth in chemical industries. Highly efficient chemical manufacture has motivated the development of batch and continuous-flow systems. Flow syntheses have several advantages over batch syntheses, in terms of productivity, heat and mixing efficiency, reproducibility, and operability.[1−4] Moreover, large-scale production is easily achieved by increasing the size or number of flow reactors.[5,6] Chemical manufacture requires a high-quality synthesis with excellent system operability. In this regard, the development of highly efficient flow reactors with a sophisticated synthetic design is desirable.

Designs for chemical syntheses and practical devices could result in advanced flow reactors. Taking inspiration from an evolved, natural, and efficient flow reactor, we mimic a living system where flow syntheses is processed smoothly through capillaries.[7] Nowadays, many flow reactors that possess capillary arrangements have been proposed for chemical syntheses.[8−12] Although this approach has succeeded and attracted inquisitive attention in the industrial field, the fabrication of integrated capillaries requires complicated and cost-consuming micromachining techniques. A monolithic porous material, in which capillaries are integrated three-dimensionally, has been proposed.[13] A single piece of porous material can be prepared in any desired shape in a one-step synthesis, and easily scaled up. Additionally, its porous property and chemical functionality are also precisely controllable.[14] Fréchet and Svec have developed many monolithic porous polymers and demonstrated their feasibilities in flow applications.[15−32] In the living system, substrates can be transported through not only the capillaries but also cells comprising the wall of the capillaries. The cells provide a reaction platform in their internal space. Thus, in the living system, the substrates can fully permeate throughout a body. If the wall of capillary in the monolithic porous polymer is also permeable to target molecules, its internal space could be provided as a reaction platform in a fashion similar to that of the living system. Such a hierarchical permeability (through capillaries and cells)[33] facilitates flow operation and enlarges a space available to a chemical transformation within the limited volume of the flow reactor. However, the hierarchical permeability of the target molecules have not been considered in a flow synthesis yet.

We previously developed a poly(N-isopropylacrylamide) (PNIPAm)-based monolithic porous gel (MPG) as a flow reactor for a Pd-catalyzed Suzuki coupling reaction.[34] Capillaries existed in the MPG as interspatial pores, contributing to successful application in the flow system. A reaction platform was provided in the gel network by the loading of Pd(0), affording a Pd(0)-loaded MPG (Pd/MPG). The Pd/MPG should contain two different pathway scales for the substrates and product (Figure ). First, the feed solution containing the substrates flows continuously through capillaries in the Pd/MPG. Next, the substrates easily diffuse into the walls of the capillaries that are composed of a gel. Molecular diffusivities into a gel network are well maintained owing to the flexibility of the polymer chain. Therefore, if catalysts are immobilized in the gel network, the substrates can easily access the catalysts and are transformed smoothly, giving a desired product. Finally, the product diffuses out of the gel network. This MPG strategy would be a key to fabricate the hierarchically permeable structure in a catalytic flow reactor.

Figure 1

Conceptual illustration of a flow reactor using monolithic porous gel (MPG) for the catalytic reaction.

Herein, we investigated the performance of flow reactors using Pd/MPGs with different gel network sizes. A set of Pd/MPGs with different gel network sizes was prepared. The Pd/MPGs were applied to a flow Suzuki coupling reaction as a model system to demonstrate their advantages over other types of inorganic supports (porous glass membrane, silica particles, or carbon). Using Pd/MPGs with different gel network sizes, the influence of the gel network on molecular transport under flow conditions was also investigated.

Experimental Section

Reagents

Water with a conductivity of 18.2 MΩ cm (Milli-Q, Millipore Co., Bedford, MA) was used in all experiments. NIPAm (Wako Pure Chemical Industries Ltd., Osaka, Japan), N-(3-dimethylaminopropyl)methacrylamide (DMAPM, Tokyo Chemical Industry Co., Ltd., Tokyo, Japan), N,N′-methylenebisacrylamide (BIS, Tokyo Chemical Industry Co., Ltd.), and 2,2′-azobis(2-amidinopropane)dihydrochloride (AAPD, Wako Pure Chemical Industries Ltd.) were used as the gel matrix, tertiary amine ligand, cross-linker, and radical initiator, respectively. NIPAm was purified by recrystallization from benzene/n-hexane and dried in vacuo at room temperature. The polymerization inhibitor in DMAPM was removed using an activated alumina column. The porous glass membrane (mean pore size: 2000 nm, 20 mmϕ, SPG Technology Co. Ltd., Miyazaki, Japan) and silica particles (particle size: 40–63 μm, Merck Co., Darmstadt, Germany) were tertiary-aminated using a procedure similar to that previously reported.[35] K2PdCl4 (Sigma-Aldrich Co., St. Louis, MO) and NaBH4 (Tokyo Chemical Industry Co. Ltd.) were used to prepare Pd(0)-loaded supports. Pd-loaded carbon (Pd/C, Wako Pure Chemical Industries Ltd.) was also used as a referential Pd catalyst. Phenylboronic acid (1, Tokyo Chemical Industry Co., Ltd.), 4-bromobenzoic acid (2, Tokyo Chemical Industry Co., Ltd.), and Na2CO3 (Wako Pure Chemical Industries Ltd.) were used to prepare the substrate stock solution. 4-Phenylbenzoic acid (3, Tokyo Chemical Industry Co., Ltd.) was used to prepare a standard solution of the Suzuki coupling product for calibration.

Preparation and Characterization of MPGs

MPGs were prepared by the copolymerization of NIPAm, DMAPM, and cross-linking monomer BIS in water (Scheme a, described in detail in the Supporting Information).[34,36] MPGs with different gel network sizes were synthesized with a fixed feed ratio of DMAPM (10 mol %) and different feed ratios of BIS (5, 10, or 30 mol % for MPG1, 2, or 3). After washing with sufficient water and lyophilizing, the yields of the dried MPGs were >97%. Differences among gel network sizes for MPG1–3 were confirmed by evaluating the swelling properties of the MPGs in water (see Supporting Information). Tertiary-aminated porous glass membrane and silica particles were also prepared (described in detail in the Supporting Information).

Scheme 1

(a) Chemical Reaction, (b) Loading of Pd(0), (c) Suzuki Coupling Reaction, and (d) Flow Reactor

Chemical reaction for synthesis of MPG with tertiary amine.

Loading of Pd(0) on the tertiary amine in the MPG.

Suzuki coupling reaction between phenylboronic acid (1) and 4-bromobenzoic acid (2) to give 4-phenylbenzoic acid (3) under flow condition.

Flow reactor using Pd(0)-loaded support for the Suzuki coupling reaction.

Preparation of Pd(0)-Loaded Supports

Pd/MPGs were prepared by Pd(II)-ion adsorption and subsequent reduction into Pd(0) using K2PdCl4 and NaBH4 (Scheme b, described in detail in the Supporting Information).[34] Differences among gel network sizes for Pd/MPG1–3 were also confirmed in the same manner as mentioned above. The Pd/MPGs were lyophilized and their internal structures were observed using field emission scanning electron microscopy (FE-SEM, SU8000, Hitachi High-Technologies Corporation, Tokyo, Japan). Before observation, the MPG surfaces were coated with platinum (thickness: approx. 4 nm) using an auto fine coater (JFC-1600, JEOL Ltd., Tokyo, Japan). The porous properties of the Pd/MPGs were also evaluated using mercury intrusion porosimetry (AutoPoreIV9520, Micromeritics Instrument Co., Norcross, GA). Pd(0) loaded in the MPGs was observed using transmission electron microscopy (TEM, TECNAI 20, Philips FEI, Netherlands). The Pd/MPGs were then immersed in aqua regia for 24 h, and the eluent was analyzed using inductively coupled plasma atomic emission spectroscopy (ICP-AES, ICPS-8100, Shimadzu Corporation, Kyoto, Japan) to estimate the amount of Pd loaded (qPd), which was defined as followswhere VM and ν are the bulk volume of the MPGs and the volume of solution, respectively, and subscript E indicates the eluent. Pd(0)-loaded nonporous gel (Pd/nonporous gel), porous glass membrane (Pd/porous glass membrane), silica particles (Pd/silica particles), and Pd/C were also evaluated in the same manner (see Supporting Information).

Suzuki Coupling Reactions in the Flow System

Suzuki coupling reactions of 1 and 2 (Scheme c) were conducted using Pd/MPG1–3, Pd/porous glass membrane, Pd/silica particles, and Pd/C in the flow system. All flow operations were carried out in an incubator at 30 °C. The Pd(0)-loaded supports were packed into the homemade column with a support screen (0.45 μm Omnipore, Millipore Co., Billerica, MA), then attached to the flow apparatus (Scheme d), syringe pump (YSP-101, YMC Co., Ltd., Kyoto, Japan), and pressure gauge (KDM-30, Krone Co., Tokyo, Japan) in a thermocontrolled incubator at 30 °C. The water permeabilities of the columns were preliminarily examined to obtain permeation coefficients using Darcy’s law (kD) (described in detail in the Supporting Information). After conditioning the reactor using aqueous Na2CO3 solution, 1 (10 mmol L–1) and 2 (11 mmol L–1, 1.1 equiv) in aqueous Na2CO3 (11 mmol L–1, 1.1 equiv) solution were permeated through the reactor at a fixed flow rate for 30 days. The residence time (τ) of the substrates was around 0.5 h. The eluent was continuously collected and analyzed using a high-performance liquid chromatography (HPLC, LC-2000Plus, JASCO CO., Tokyo, Japan) system with a reverse phase column (Mightysil RP-18 GP 250-4.6, Kanto Chemical Co.) and an ultraviolet detector. Acetonitrile and water (50:50) containing 0.1 vol % trifluoroacetic acid was employed as a mobile phase. Turnover numbers (TONs) of catalysts in the flow system were estimated using the following equationDuring substrate permeation, Pd leaching into the eluent was also confirmed using ICP-AES.

Kinetics Study of the Suzuki Coupling Reaction Using Pd/MPGs in the Flow System

Pd/MPG and a support screen were packed into the homemade column, then attached to the flow apparatus. After conditioning of the reactor using aqueous Na2CO3 solution, the substrate was pumped into the reactor at various flow rates. In the period with the short residence time (τ = 0.1 h), turnover frequencies (TOFs) of Pd catalysts in the flow system were estimated from the reaction rate per amount of Pd catalysts.

Results and Discussion

Properties of Pd/MPGs

A white and monolithic structure was obtained from all monomer compositions with different BIS contents used to prepare MPG1–3. The bulk volumes of the MPGs changed with temperature, and the thermoresponsiveness was suppressed by increasing BIS content (Figure S1a). Pd/MPG1–3 was obtained by Pd(II)-ion adsorption and subsequent reduction into Pd(0).[34] The Pd/MPGs were slightly brown (Figure a). The bulk volumes and thermoresponsiveness of the Pd/MPGs were almost identical to those of the original MPGs (Figure S1b). The internal morphologies of the Pd/MPGs were observed using FE-SEM (Figures b,c, and S2). The pore sizes in the Pd/MPGs were in the order of several hundred nanometers and several micrometers (Figure S3 and Table S1). The amounts of Pd loaded in the MPGs were around 1.3–1.5 μmol cm–3 (Table ). Pd(0) loaded on the MPGs were observed using TEM, with average Pd(0) sizes of 2.0–2.4 nm in all of the Pd/MPGs (Table , Figures d and S4a). In a similar manner, the Pd/porous glass membrane, Pd/silica particles, and Pd/C were also evaluated (Figure S6, see the Supporting Information).

Figure 2

(a) Photograph of swollen Pd/MPG1 after immersion in water at 30 °C. (b, c) FE-SEM image of Pd/MPG1 after lyophilization. (d) TEM image of Pd/MPG1 before use in the Suzuki coupling reaction.

Table 1

Bulk Volumes, Pd Amount, Gel Network Size, and Pd(0) Average Size in Pd/MPG

entry	BIS content (mol %)	VM (cm3 g–1)a	qPd (μmol cm–3)b	gel network size (nm)c	Pd(0) average size (nm)d
Pd/MPG1	5	10.4	1.3	7.0	2.4 ± 0.7
Pd/MPG2	10	7.9	1.4	5.2	2.2 ± 0.7
Pd/MPG3	30	5.3	1.5	2.8	2.1 ± 0.7

Calculated in water at 30 °C.

Estimated by immersion in aqua regia and the analysis of eluent using ICP-AES.

Cited from the literature.[38]

Measured from TEM images of Pd/MPGs.

The bulk volumes of the MPG1–3 were dependent on temperature, which was attributed to the phase transition of PNIPAm as a gel matrix (Figure S1a). The volume-change responsiveness to temperature was suppressed by increasing the BIS content. The increase in the cross-linker content in the PNIPAm gel could be regarded as parallel to the decrease in the gel network size.[37] The network sizes of PNIPAm gels with BIS contents of 5, 10, and 30 mol % were calculated to be 7.9, 5.2, and 2.8 nm, respectively.[38] The gel network sizes of the MPGs were indicated to be controlled by their BIS contents. The loading of Pd(0) did not affect the bulk volumes of the Pd/MPGs (Figure S1b), suggesting that the gel network sizes were conserved. The FE-SEM images showed that the Pd/MPGs had capillaries as interspatial pores (Figures b,c, and S2), which would facilitate the flow system using the Pd/MPGs. As the average sizes of Pd(0) in the MPGs were smaller or almost equal to the gel network sizes (Table ), reaction platforms were expected to be provided within the gel networks.

Permeabilities of Flow Reactors

Before application in the flow system, the water permeabilities of support materials were evaluated. When water permeated through a nonporous gel (Figure S5, see the Supporting Information) at a flow rate of 100 mL h–1, the pressure loss was tremendously high and never reached a steady state (Figure S7a). However, water permeation through MPG1 gave moderate pressure losses under the same conditions. Pressure losses at the steady state proportionally increased with increasing flow rates (Figure S7b). The kD of MPGs, porous glass membrane, and silica particles were estimated to be around 10–14–10–13 m2 (Figure ), which were comparable to those of commercially available membrane filters with pore sizes ranging from 1000 to 10 000 nm.[39] After loading of Pd(0), the kD was in almost of the same order as that of the original supports (Figures and S7c). Pd/C also showed almost the same permeability (2.4 × 10–14 m2) as that of porous glass membrane and silica particles.

Figure 3

Permeation coefficients of support materials before (open rectangle) and after (filled rectangle) loading of Pd.

The MPG1–3 exhibited their excellent permeabilities, which were attributed to the capillaries (Figure S7b). However, water could not permeate through the nonporous gel (Figure S7a). Therefore, it was difficult to apply the nonporous gel to a flow process. As the water permeabilities of the Pd(0)-loaded supports were acceptable (Figures and S7c), it was expected that Suzuki coupling reactions could be achieved in the flow system with moderate pressure losses.

Suzuki Coupling Reaction in the Flow System

Pd/MPG1–3, Pd/porous glass membrane, Pd/silica particles, and Pd/C were applied to the Suzuki coupling reaction between 1 and 2 under flow condition (Scheme c). The TONs of Pd(0)-loaded supports were estimated along with permeation of the substrate at a fixed flow rate (τ = 0.5 h) for 30 days (Figure ). In all of the catalysts, there were no aggregations of Pd(0) observed from TEM images after the Suzuki coupling reaction (Figures S4b and S6b). Remarkably, the TONs of Pd/MPG1, Pd/MPG2, and Pd/MPG3 were much high to give 2631, 2290, and 1333, respectively, compared with those of the Pd/porous glass membrane, Pd/silica particles, and Pd/C (65, 144, and 26), respectively. In addition, Pd leachings from the Pd/MPGs were below the detection limit of ICP-AES. However, Pd leachings from the Pd/porous glass membrane, Pd/silica particles, and Pd/C were detected to be 5.0, 2.9, and 1.6%, respectively.

Figure 4

TONs of Pd(0)-loaded supports and Pd leaching study for the Suzuki coupling reaction in the flow system. Reaction conditions: 1 (10 mmol L–1) and 2 (11 mmol L–1, 1.1 equiv) in aqueous Na2CO3 solutions (11 mmol L–1, 1.1 equiv), 30 °C, τ: 0.5 h, 30 days.

Surprisingly, the productivities of Pd/MPG1–3 were well maintained without any Pd leaching, whereas those of the Pd/porous glass membrane and Pd/silica particles were lost during 30 days of use due to Pd leaching (Figure ). It should be noted that the TONs of the Pd/porous glass membrane, Pd/silica particles, and Pd/C under batch condition were much higher (248, 1004, and 720, Table S2, described in detail in the Supporting Information) than of those under continuous-flow condition (65, 144, and 26). Using the porous glass membrane, silica particles, and carbon as support materials, Pd was localized on the surface. The localized Pd was only accessible to substrates and eluted easily under flow conditions. The solid–liquid contacting efficiency on the Pd/porous glass membrane was expected to be higher than that on the Pd/silica particles.[40] However, the improved efficiency resulted in a larger leaching percentage from the Pd/porous glass membrane. Using the MPGs as support materials, Pd was fully taken up into the gel networks. A moderate steric hindrance exerted by the polymer chain allowed effective retention of Pd and fast molecular transport in the gel network. Also, this MPG strategy was effective for Suzuki coupling reactions between phenylboronic acid and other aryl halides (Table S3). Even in the flow system, the Pd/MPGs conducted chemical synthesis effectively, indicating their advantages over other types of Pd(0)-loaded supports.

Effect of Gel Network Size on Pd Catalytic Activity in the Flow System

Kinetics studies of the Suzuki coupling reaction under flow conditions were conducted using Pd/MPGs with different BIS contents (Figure ). The percentage conversion into 3 increased with increasing τ of the substrates. The TOFs of the Pd/MPG1, Pd/MPG2, and Pd/MPG3 were 27.4, 16.1, and 7.8 h–1, respectively (Table ). Interestingly, the Pd/MPG1 exhibited a 3.5-fold higher TOF than that of Pd/MPG3.

Figure 5

Kinetics study on the Suzuki coupling reaction using Pd/MPGs in the flow system. Reaction conditions: 1 (10 mmol L–1) and 2 (11 mmol L–1, 1.1 equiv) in aqueous Na2CO3 solutions (11 mmol L–1, 1.1 equiv), 30 °C.

Table 2

TOFs for the Suzuki Coupling Reaction in the Flow System Using Pd/MPGsa and Diffusion Coefficients of the Substrates and Product in the Gel Network

	D in gel network (×10–10 m2 s–1)b			TOF (h–1)
entry	1	2	3
Pd/MPG1	5.8	8.0	3.5	27.4
Pd/MPG2	4.9	3.9	2.1	16.1
Pd/MPG3	1.2	3.2	1.2	7.8

Reaction conditions: 1 (10 mmol L–1) and 2 (11 mmol L–1, 1.1 equiv) in aqueous Na2CO3 solutions (11 mmol L–1, 1.1 equiv), 30 °C. BIS contents: 5 mol % for gels.

Estimated using nonporous gels with the same monomer composition as MPGs.

Using Pd/MPG1, the highest TOF was achieved for the Suzuki coupling reaction in the flow system (Table ). The surface areas of capillaries in the MPGs increased with increasing BIS contents (Table S1), supporting that molecular transport on the surface of the capillaries was not rate-limiting. The TOFs of the Pd/MPGs were likely to be dependent on the size of the gel network provided as a reaction platform.

Diffusivities of Substrates and Product in the Gel Network

Nonporous gels with different BIS contents were prepared and tested for diffusion experiments of 1–3 in the gels (Figure S8, described in detail in the Supporting Information). It was found that the diffusion coefficients (D) of 1–3 in the gels were comparable to those of small molecules in bulk liquid.[41] Moreover, the D increased with increasing the BIS contents of the nonporous gels (Table ).

In the gel network, diffusivities of substrates and product were well maintained because of the flexible polymer chain of PNIPAm gel. The D values increased with increasing gel network size (Table ) and were similar to those previously reported.[42−45] If the molecular diffusions in the gel network were rate-limiting steps in the Suzuki coupling reaction, the D values would strongly influence catalytic activities, as demonstrated (Figure and Table ). Molecular diffusion in the gel network were controlled by tuning the network size. Therefore, the design of hierarchical permeabilities in a capillary and gel network were key factors in constructing a highly efficient continuous-flow reactor for chemical synthesis.

Conclusions

A flow reactor using the monolithic porous gel, MPG, for chemical synthesis was proposed. The MPGs had capillaries as interspatial pores, and the MPG gel network sizes were controlled by tuning a feed ratio of the cross-linking monomer. The capillaries in the MPG allowed flow application. The MPGs showed excellent permeabilities comparable with those of commercially available membrane filters with pore sizes in the range 1000–10 000 nm. Owing to the moderate steric hindrance of the gel network, Pd/MPG showed significant stability and catalytic activity in a Suzuki coupling reaction under flow conditions. However, little performance was observed in the inorganic supports. The catalytic activities of the Pd/MPGs were dependent on the gel network size, which corresponded to the diffusivities of the substrates and product. The MPG provides an easy and sophisticated strategy for preparing and functionalizing a flow reactor for chemical synthesis. In conjunction with a practical process, the MPG would have a significant impact on the field of current flow reactors. The conceptual design of a hydrogel network can be extended to a reactor design for chemical synthesis in a flow system under aqueous media. The concept should be further extended to other types of flow catalytic reactions using transition-metal catalysts under flow conditions.