Introduction

The design of a reaction space around a catalytic center is an intriguing possibility for innovations in catalysis. Owing to their design versatility, macromolecules are attractive materials for providing catalysts with unique and desirable functions. Until now, macromolecules have been used as supporting agents for metal catalysts, where the main objectives are focused on practicability in catalysis, such as the recoverability of products and the recyclability of catalysts.1, 2, 3, 4 The representative materials are insoluble heterogeneous polymer-supported catalysts typically produced using crosslinked polystyrene gel and silica gel. Unfortunately, these materials often exhibit inferior catalytic activity and substrate selectivity compared with homogeneous catalysts, owing to the reduced accessibility of the substrate to the catalytic center. Although soluble polymer-supported catalysts have also been developed to improve activity, they sometimes leach the catalysts from the supporting agents, eventually leading to inferior product recovery and catalyst recyclability. In contrast, dendrimers,5, 6, 7 amphiphilic block copolymers for micellar catalysis8, 9 and polymersomes10, 11 are examples of macromolecular scaffolds that provide well-designed reaction spaces for a unique catalytic performance. They serve to segment the reaction space isolated from the outer environment, as viewed by an enzyme, to accelerate catalysis7, 9 and realize cascade reactions.11 However, they normally require multistep synthesis and/or complex optimization in catalytic conditions.

Microgel-core star polymers12, 13, 14, 15, 16, 17, 18, 19, 20, 21, 22, 23 are a new category of macromolecule-based scaffolds for enclosing catalysts.15, 16, 17, 18, 19, 20, 21, 22, 23 The star polymer carries a unique microgel core covered by linear arms in the center of the polymer. This promising environment encouraged us to produce metal-bearing microgel-core star polymer catalysts by ruthenium-catalyzed living radical polymerization.24, 25, 26, 27, 28 Here, the living polymers (arms) were linked with a divinyl compound (1) and RuCl2(PPh3)3 in the presence of a phosphine ligand-bearing monomer (diphenylphosphinostyrene: 2; Scheme 1).15, 16, 17, 18, 19 Importantly, the ligand monomer directly encapsulates the ruthenium polymerization catalyst into a microgel core via ligand exchange during the arm-linking reaction to produce ruthenium-bearing microgel-core star polymers. A ruthenium catalyst with triphenylphosphine is transformed into a star polymer-supported ruthenium catalyst in one-pot synthesis. On the basis of this efficient synthetic procedure, we can also successfully introduce hydrophilic, amphiphilic and thermosensitive functions to ruthenium-carrying star polymers in conjunction with poly(ethylene glycol) methyl ether methacrylate (PEGMA) as an arm monomer (Scheme 1).17, 18 This allows the one-pot transformation of ‘hydrophobic polymerization catalysts’ into ‘amphiphilic and thermosensitive star polymer catalysts’. The resultant star catalyst is completely soluble in various solvents (for example, toluene, alcohol and water) owing to the presence of poly(ethylene glycol) (PEG) in the arms, whereas the microgel core comprising multiple phosphine ligands and ruthenium is hydrophobic and crosslinked. Owing to the core-reaction pocket covered by the amphiphilic and thermosensitive arms, the star polymer catalysts induce phase-transfer catalysis in water with a unique activity and stability.18

Herein, we investigate the homogeneous transfer hydrogenation of ketones29, 30, 31, 32, 33, 34, 35 catalyzed by ruthenium-bearing microgel-core star polymers with PEG arms (Ru(II)-PEG Star)17 coupled with K2CO3 in 2-propanol (Scheme 2). In this catalysis, it is expected that the unique structure of the star polymer catalysts will allow effective accessibility of a hydrophobic substrate (ketone) to the hydrophobic reaction space (core) and smooth diffusion of the resultant hydrophilic product (alcohol) from the core to the polar arm parts achieving high activity. In addition, the crosslinked core enhances the stability of catalysts, thus improving recycle efficiency. These features will be discussed in comparison with similar polymer-supported catalysts and the original RuCl2(PPh3)3.

Experimental procedure

Materials for polymer synthesis

PEGMA (Mn≈475, Sigma-Aldrich) was purified by column chromatography with an inhibitor remover (Sigma-Aldrich, St Louis, MO, USA) and degassed under reduced pressure before use. Methyl methacrylate (MMA; Tokyo Kasei, Tokyo, Japan; purity >99%) was dried overnight over calcium chloride and purified by double distillation from calcium hydride before use. (MMA)2-Cl (initiator) was prepared according to the literature.36 Ethylene glycol dimethacrylate (1: Sigma-Aldrich, purity >98%) was purified by distillation from calcium hydride before use. Diphenylphosphinostyrene (2), supplied by Hokko Chemical (Tokyo, Japan) (purity >99.9%), and polystyrene crosslinked with divinylbenzene, diphenylphosphinated (PPh3-Gel (3): polystyrene crosslinked with 2% divinyl benzene; 3 mmol phosphine per g resin; Sigma-Aldrich) were degassed under reduced pressure and purged with argon before use. The 2,2′-Azobis(isobutyronitrile) (Wako, Osaka, Japan; purity >98%) was used as received. RuCl2(PPh3)3 (Sigma-Aldrich, purity >97%) was used as received and handled in a glove box under a moisture- and oxygen-free argon atmosphere (H2O <1 p.p.m. and O2 <1 p.p.m.). n-Bu3N (Tokyo Kasei, purity >98%) was bubbled with argon for 15 min immediately before use. Internal standards for gas chromatography (n-octane for MMA, tetralin for 1) were dried over calcium chloride overnight and distilled twice from calcium hydride. Toluene (solvent) was purified; passing it through a purification column (Glass Contour Solvent Systems by SG Water USA, Nashua, NH, USA) before use. Hexane (Wako, dehydrated) for polymer purification was used as received. The solvents were bubbled with argon for more than 15 min immediately before use.

Materials for hydrogenation

Substrates (S1: acetophenone, Sigma-Aldrich, purity >99%; S2: p-chloroacetophenone, Wako, purity >95%; S3: 4-methoxyacetophenone, Sigma-Aldrich, purity >99%; S4: p-butylacetophenone, Sigma-Aldrich, purity >95%; S5: valerophenone, Sigma-Aldrich, purity >99%; S6: 1-indanone, Sigma-Aldrich, purity >99%; S7: cyclopentanone, Wako, purity >95%; S8: cyclohexanone, Wako, purity >99%; S9: 2-hexanone, Wako, purity >95%; S10: 2-octanone, Wako, purity >98%; S11: 2-dodecanone, Wako, purity >99%; and S12: 5-methyl-3-heptanone, TCI (Tokyo, Japan) purity >95%) were degassed under reduced pressure and purged with argon or bubbled with argon for more than 15 min before use. K2CO3 (Wako, >99.5%) and 2-propanol (Wako, dehydrated) were degassed under reduced pressure and purged with argon before use.

Characterization

The number average molecular weight (Mn) and molecular weight distribution (Mw/Mn) of the polymers were measured by size-exclusion chromatography (SEC) in N,N-dimethylformamide (DMF) containing 10 mM LiBr at 40 °C (flow rate: 1 ml min−1) on three linear-type polystyrene gel columns (Shodex KF-805L: exclusion limit=4 × 106; particle size=10 μm; pore size=5000 Å; 0.8 cm inner diameter × 30 cm; Shodex, Tokyo, Japan) that were connected to a Jasco PU-980 precision pump, a Jasco RI-930 refractive index detector and a Jasco UV-970 UV/Vis detector set at 270 nm (Jasco, Tokyo, Japan). The columns were calibrated against 10 standard poly(MMA) samples (Polymer Laboratories, Church Stretton, UK: Mn=1000–1 200 000; Mw/Mn=1.06–1.22). The 1H-nuclear magnetic resonance (NMR) spectra were recorded in CDCl3 or CD2Cl2 at 25 °C on a JEOL JNM-LA500 spectrometer operating at 500.16 MHz (JEOL, Tokyo, Japan). The 31P-NMR spectra were recorded with (C2H5O)2POH (12 p.p.m.) as an internal standard in toluene-d8 at 25 °C on a JEOL JNM-LA500 spectrometer operating at 500.16 MHz. The absolute weight-average molecular weight (Mw) of the star polymer catalysts was determined by multiangle laser light scattering coupled with SEC (SEC-MALLS) in DMF containing 10 mM LiBr at 40 °C on a Dawn E instrument (Wyatt Technology, Santa Barbara, CA, USA: GaAs laser; λ=690 nm). The refractive index increment (dn/dc) was measured in DMF at 40 °C on an Optilab DSP refractometer (Wyatt Technology; λ=690 nm, c<2.0 mg ml−1). Ultraviolet–visible (UV–Vis) spectra to determine the ruthenium contents of the star polymer catalysts were recorded in CH2ClCH2Cl at room temperature on a Shimadzu MultiSpec 1500 (Shimadzu, Kyoto, Japan). The core-bound Ru(II) content was determined by the absorbance at 475 nm and a calibration plot made for RuCl2(PPh3)3 (0.10–2.0 mM solution) at the same wavelength. The core-bound Ru(II) content was further estimated by inductively coupled plasma atomic emission spectrometry (IRIS Intrepid II XDL Radial, Thermo Fisher Scientific, Waltham, MA, USA).

Synthesis of Ru(II)-PEG Star (C1–C3)

Ru(II)-PEG Star (C1)17 was synthesized using a syringe technique under dry argon in baked glass tubes equipped with a three-way stopcock. RuCl2(PPh3)3 (0.09 mmol, 86.3 mg) was first placed in a 50-ml round-bottomed flask. Thereafter, toluene (6.45 ml), n-Bu3N (0.18 mmol, 0.45 ml of 400 mM solution in toluene), PEGMA (4.5 mmol, 1.98 ml) and (MMA)2-Cl (0.09 mmol, 0.12 ml of 773 mM solution in toluene) were sequentially added in that order to the flask at 25 °C under argon. The total volume of the mixture was thus 9 ml. Immediately after mixing, the mixture was placed in an oil bath at 80 °C. The polymerization reached over ∼90% conversion in 10 h; subsequently, MMA (0.9 mmol, 0.096 ml) and n-octane (0.024 ml) were added to the unquenched solution. The MMA conversion reached over ∼78% in 24 h, after which a solution of 1 (1.35 mmol, 0.63 ml of 2159 mM solution in toluene), 2 (0.11 mmol, 0.11 ml of 1000 mM solution in toluene), tetralin (0.09 ml) and RuCl2(PPh3)3 (86.3 mg) in toluene (3.68 ml) was added to the unquenched arm-polymer solution (SEC: Mn=34 800, Mw/Mn=1.47). After 25 h, the reaction was terminated by cooling the mixture to −78 °C. The conversions of PEGMA, MMA, 1 and 2 were 98%, 94%, 84% and 100%, respectively, as determined by 1H-NMR with an internal standard of tetralin (PEGMA, 2) and gas chromatography with n-octane (MMA) and tetralin (1) as internal standards. The yield of the star polymers was 76%, as calculated from the area ratio of the arm residue and the star polymers using SEC curves. The quenched mixture was precipitated into hexane under argon in order to remove the remaining monomers and an amine additive. The precipitate was further purified by column chromatography with silica gel (Wako Gel 200) and toluene as an eluent under argon to remove free ruthenium complexes. The eluted solutions were evaporated to yield the final products, which were subsequently dried overnight under vacuum at room temperature before analysis and catalysis. SEC-MALLS (DMF): Mw=770 000 g mol−1; 16 arms per star polymer; Rg=15.1 nm. 1H-NMR (500 MHz, CD2Cl2, 25 °C): δ 7.4–7.7 (aromatic), 4.2–4.0 (-CO2CH2CH2-), 3.9–3.4 (-OC2H4O-), 3.4–3.2 (-OCH3), 2.2–1.7 (-CH2-), 1.6–0.8 (-CCH3). UV–Vis (CH2ClCH2Cl, 25 °C, RuCl2(PPh3)3 calibration at 475 nm): 24 μmol Ru per g polymer (C1). Inductively coupled plasma atomic emission spectrometry: 27 μmol Ru per g polymer (C1). Other Ru(II)-PEG Stars (C2, C3) were also prepared by using the same procedure coupled with a different volume of 2 and were similarly characterized, as shown in Table 1.

Table 1 Characterization of Ru(II)-bearing polymer catalystsa
Full size table

Synthesis of Ru(II)-MMA Star (C4)

Ru(II)-MMA Star (C4) was synthesized by the linking reaction of poly(methyl methacrylate) (PMMA) arms (conversion of MMA=93%, 60 h, Mn=8300, Mw/Mn=1.19) with 1 and 2 using RuCl2(PPh3)3-catalyzed living radical polymerization for an 88% yield of star polymers (conversion of MMA/1/2=98/90/100%, +20 h).15, 16 SEC-MALLS (DMF): Mw=600,000; 40 arms per star polymer; Rg=11.6 nm. 1H-NMR (500 MHz, CD2Cl2, 25 °C): δ 7.4–7.7 (aromatic), 3.6–3.5 (-OCH3), 2.2–1.7 (-CH2-), 1.6–0.8 (-CCH3). UV–Vis (CH2ClCH2Cl, 25 °C, RuCl2(PPh3)3 calibration at 475 nm): 29 μmol Ru per g polymer (C4).

Synthesis of Ru(II)-Gel (C5)

In a 50-ml round-bottomed flask, RuCl2(PPh3)3 (0.24 mmol, 230 mg) in toluene (24 ml) was added to polymer-supported triphenylphosphine (3) (1.2 mmol phosphine, 0.4 g) under argon. The mixture was stirred at 80 °C for 28 h under dispersion to yield a red-brown Ru(II)-supported powder with a colorless supernatant. The obtained powder was washed three times with toluene under argon. The supernatant exhibited no UV–Vis absorption derived from RuCl2(PPh3)3, indicating quantitative immobilization of Ru(II) complexes onto 3 to give C5: 420 μmol Ru per g polymer (C5).

Synthesis of Ru(II)-Random (C6)

The 2,2′-Azobis(isobutyronitrile) (0.3 mmol, 51.5 mg) was placed in a 50-ml round-bottomed flask. Further, toluene (4.08 ml), MMA (39.8 mmol, 4.24 ml) and 2 (2.09 mmol, 2.15 ml of 975 mM toluene solution) were sequentially added to the flask at 25 °C under argon. The mixture was placed in an oil bath at 80 °C for 25 h. The reaction was terminated by cooling the mixture to −78 °C (conversion of MMA/2=99/100%). The solution was precipitated with hexane three times, and the resulting phosphine-bearing random copolymer (4) was dried under vacuum. SEC (DMF, PMMA standards): Mn=16700, Mw/Mn=2.24. 1H-NMR (500 MHz, CD2Cl2, 25 °C): δ 7.0–7.4 (aromatic), 3.6–3.4 (-OCH3), 2.2–1.7 (-CH2-, -CHPh-), 1.6–0.8 (-CCH3). 31P-NMR (500 MHz, toluene-d8, 25 °C): δ −0.9 (PPh3). The 4-bound-phosphine ligands calculated from the monomer conversion in 4: 0.457 μmol per g polymer.

In a 50-ml round-bottomed flask, a solution of RuCl2(PPh3)3 (0.23 mmol, 217 mg) in toluene (23 ml) was added to 4 (1.13 mmol of 4-bound phosphine, 2.48 g) under argon. The mixture was placed in an oil bath at 80 °C for 23 h. After the reaction was terminated by cooling the mixture to −78 °C, the product (C6) was purified by column chromatography with silica gel (Wako Gel 200) and toluene as an eluent under argon and was dried overnight under vacuum at room temperature before analysis and catalysis. SEC (DMF, PMMA standards): Mn=20700, Mw/Mn=2.94. 1H-NMR (500 MHz, CD2Cl2, 25 °C): δ 7.0–7.4 (aromatic), 3.6–3.4 (-OCH3), 2.2–1.7 (-CH2-, -CHPh-), 1.6–0.8 (-CCH3). 31P-NMR (500 MHz, toluene-d8, 25 °C): δ −0.9 (PPh3). UV–Vis (CH2ClCH2Cl, 25 °C, RuCl2(PPh3)3 calibration at 475 nm): 50 μmol per g polymer (C6).

Transfer hydrogenation of ketones catalyzed by Ru(II)-PEG Star (C1)

The typical procedure of C1-catalyzed transfer hydrogenation of a ketone was as follows: K2CO3 (1 mmol, 138 mg) was placed in a baked 50-ml round-bottomed flask equipped with a condenser and a three-way stopcock, and the flask was purged with argon. A solution of C1 (0.42 g: [core-Ru(II)]0=0.01 mmol) and acetophenone (S1: 10 mmol, 1.17 ml) in 2-propanol (10 ml) was added to the flask at 25 °C under argon. The mixture was stirred and refluxed at 100 °C. The solution was sampled at a predetermined period by using the syringe technique under argon to determine the conversion. The yield was determined by 1H-NMR analysis of the reaction solution.

Results and Discussion

Design of Ru(II)-bearing polymer catalysts

Three types of poly(PEGMA)-armed star polymers with a Ru(II)-bearing microgel core (C1–C3) were used as catalysts for the transfer hydrogenation of ketones, compared with a similar series of Ru(II)-bearing polymer catalysts (C4–C6) and RuCl2(PPh3)3 (C7). Their chemical structures and characterization are given in Figure 1 and Table 1, respectively. Ru(II)-PEG Stars labeled as C1, C2 and C3 were directly prepared by RuCl2(PPh3)3-catalyzed living radical polymerization of PEGMA, MMA, 1 and 2, according to Scheme 1.17 Only the ratio of 2 to the initiator (r2=n=[2]0/[initiator]0) was changed as follows: r2=1.25 (C1), 2.5 (C2) and 5.0 (C3), to lead to different numbers of core-bound ruthenium (NRu=19, 36, and 87) and different number ratios of core-bound 2 per core-bound ruthenium (N2/NRu=1.0, 1.8, and 2.6), respectively. The other conditions and feed ratios were uniform, such as the degree of polymerization (arm) (DP=([PEGMA]o+[MMA]o)/[initiator]o=60) and the ratio of 1 to the initiator (r1=[1]add/[initiator]o=15). NRu and N2/NRu increased nearly proportionally to r2. C1–C3 were well soluble in 2-propanol at temperatures above 31 °C (upper critical solution temperature).17 C4 (Ru(II)-MMA Star), directly synthesized by RuCl2(PPh3)3-catalyzed living radical polymerization of MMA, 1 and 2,15 is a hydrophobic PMMA arm version of C1 with almost the same ruthenium amount (NRu). Ru(II)-Gel (C5), polystyrene gel-supported Ru(II), was obtained from the immobilization of RuCl2(PPh3)3 on phosphine-bearing crosslinked styrene gel (3). This was used as a cutout mimic of the crosslinked core of star polymer catalysts; however, it was not soluble in any solvent. Ru(II)-Random (C6) was also obtained from the immobilization of RuCl2(PPh3)3 on a linear random copolymer of MMA and 2, which was a linear analog of the Ru(II) microgel core.

Figure 1
figure 1

Structures of Ru(II)-bearing polymer catalysts.

Full size image

Transfer hydrogenation of ketones

Effects of catalyst structure. To examine the effects of catalyst structure on the activity, we utilized Ru(II)-PEG Star (C1), Ru(II)-MMA Star (C4), Ru(II)-Gel (C5), Ru(II)-Random (C6) and RuCl2(PPh3)3 (C7) as catalysts for the transfer hydrogenation of acetophenone (S1) with K2CO3 in 2-propanol at 100 °C (reflux; Figure 2). The feed molar ratio of the catalyst (Ru(II)) to the substrate (S1) was set at [S1]/[Ru(II)]=1000/1.31 Owing to its high solubility in 2-propanol at temperatures over 31 °C, C1 efficiently and homogeneously catalyzed the transfer hydrogenation of acetophenone, with an 86% yield in 4 h. The final turnover frequency (TOF) was 215 (h−1). The reduction rate for C1 was faster than that for the conventional and homogenous C7. Although the hydrophobic PMMA-armed star catalyst (C4) and its gel counterpart (C5) were also effective for reduction (C2: 80% and C3: 72%), their rates were lower than those obtained with C1 and C7 owing to the lower solubility of C4 and C5 in the reaction mixture (C4: not completely soluble, C5: insoluble). In addition, the linear counterpart (C6) exhibited low catalytic activity (14% yield in 4 h). From these results, C1 was determined to be the most active among all of the catalysts, including the original ruthenium catalyst (C7).

Figure 2
figure 2

Transfer hydrogenation of acetophenone (S1) catalyzed by various Ru(II) complexes: Ru(II)-PEG Star (C1: filled squares); Ru(II)-MMA Star (C4: filled circles); Ru(II)-Gel (C5: filled triangles); Ru(II)-Random (C6: filled diamonds); and RuCl2(PPh3)3 (C7: open circles). Conditions: S1/Ru(II)/K2CO3=10/0.010/1.0 mmol in 2-propanol (10 ml) at 100 °C. MMA, methyl methacrylate; PEG, poly(ethylene glycol).

Full size image

Large quantities of conventional crosslinked polymer-supported catalysts (insoluble type) are generally required to achieve sufficient activity, because the active catalyst sites are just located on the surface.34 Even soluble polymer-supported catalysts often show lower activity than the original non-supported catalysts, owing to the steric hindrance and/or low mobility of the polymer backbone.15, 19, 37 However, Ru(II)-PEG Star (C1) induced the reduction of S1 faster than the original RuCl2(PPh3)3 (C7) with the same and rather small amount of ruthenium ([S1]/[Ru(II)]=1000/1), even though the star catalyst has bulky and crowded PEG side chains. This acceleration is most likely a result of the unique environment around the catalytic center,7, 9 where ruthenium catalysts are enclosed in the hydrophobic microgel core covered by the amphiphilic and polar poly(PEGMA) arms. The high activity may be explained by the following possibility (Scheme 2). As 2-propanol works as a hydrogen donor in this reaction, the efficient catalytic cycle naturally requires a sufficient supply of solvent at the ruthenium center. In this case, the hydrogen source is effectively donated to the core-bound ruthenium owing to the homogeneous solubility of the star catalyst originating from the affinity between the amphiphilic PEG-based arms and 2-propanol (Scheme 2a). In addition, the hydrophobic acetophenone (S1) can easily enter the reaction space, comprising a hydrophobic Ru(II)-bearing microgel core, whereas 1-phenylethanol, the product from S1, can efficiently escape because the alcohol product favors the polar PEG-based arm area over the hydrophobic microgel core (Scheme 2b). Such a polarity difference between the core and arms of the star catalysts might effectively diffuse the substrate and the product around the microgel-core reaction space, contributing to higher catalytic activity. This effect is further supported by the results of the reverse reaction: C1-catalyzed oxidation of 1-phenylethanol via the hydrogen transfer reaction in acetone (Supplementary Figure S1), where the activity was much lower than that of RuCl2(PPh3)3 (C7). In this case, Ru(II)-PEG Star (C1) would be structurally undesirable for efficient catalysis because the alcohol substrate favors the arm part and the ketone product, as well as the hydrophobic core (reaction space).

Effects of core-bound Ru(II). To investigate the effects of the Ru number per star polymer (NRu) on the catalytic activity, Ru(II)-PEG Stars containing various Ru(II) amounts (NRu=19 (C1), 36 (C2), 87 (C3)) were used for the hydrogenation of S1 (Figure 3). Here, the total ruthenium concentration was maintained constant ([S1]/[Ru(II)]=1000/1), indicating that a higher NRu corresponds to a smaller number of star polymer molecules used for the reaction. All Ru(II)-PEG Stars efficiently reduced S1 to 1-phenylethanol at high yields (86% (C1), 88% (C2), and 81% (C3)) in 4 h. Uniquely, the rate was dependent on NRu, increasing with decreasing NRu. This tendency is explained by the following three causes: (i) The reaction rate relies on the number of stars available for the reaction. (ii) The amount of core-bound ruthenium molecules effectively contributing to the reaction is critical to the reaction rate. The distribution and location of core-phosphine ligands (ruthenium catalysts) determine the activity. Owing to the electron-donating phosphine, 2, at a small feed ratio to 1 ([initiator]/[1]/[2]=1/15/1.25), was consumed faster than 1 during the copolymerization of 1 and 2 in the arm-linking reaction for C1, although 2 was consumed at almost the same rate as 1 for C3 ([initiator]/[1]/[2]=1/15/5.0). As a result, the phosphine ligands and ruthenium catalysts in the C1 core would be mainly located on the core surface in contrast to those in the C3 core with a homogeneous distribution of phosphine ligands. Thus, the ruthenium complexes in C1 would be more accessible to the substrate than those in C3. (iii) The number ratio of core-bound ligands per core-bound ruthenium (N2/NRu) in C1–C3 increased from 1.0 to 2.6 with increasing NRu. In other words, a single Ru(II) complex in C1 is supported by one phosphine ligand anchored in the core and has two non-bound (free) triphenylphosphines, whereas a Ru(II) complex in C3 is bound by about three phosphine ligands in the core. Thus, as NRu increased, the mobility of the core-bound ruthenium would decrease and the catalytic site would be sterically hindered. These effects of NRu and N2/NRu on the catalytic activity are consistent with the oxidation of sec-alcohols catalyzed by ruthenium-bearing microgel star polymers with poly(MMA) arms.15, 19

Figure 3
figure 3

Effects of Ru(II) number (NRu) per Ru(II)-PEG Star molecule (C1 (filled squares), C2 (open circles) and C3 (filled triangles)) on the transfer hydrogenation of acetophenone (S1): S1/Ru(II)/K2CO3=10/0.010/1.0 mmol in 2-propanol (10 ml) at 100 °C. PEG, poly(ethylene glycol).

Full size image

Substrate versatility. The high solubility of Ru(II)-PEG Star (C1)17 encouraged us to apply various substrates to C1-mediated transfer hydrogenation. Table 2 summarizes the reaction time, yield and TOF for each reaction in comparison with those obtained with RuCl2(PPh3)3 (C7: parentheses).

Table 2 Ru(II)-PEG Star (C1)-catalyzed transfer hydrogenation of ketonesa
Full size table

C1 efficiently hydrogenated all ketones (S1–S12) to their corresponding alcohols. In comparison with acetophenone (S1: non-substituted), the para-substituted acetophenone derivatives (S2, S3, S4) exhibited different TOFs depending on their substituents. S2, with an electron-withdrawing substituent (Cl), was more rapidly reduced than S1, with a high yield (93%) and a high TOF (930 h−1) at 1 h. In contrast, S3 (OCH3), a ketone with an electron-donating substituent, led to a lower yield (65%) and lower TOF (81 h−1) at 8 h as compared with S1. Among S1–S4, the TOFs increased in the order of their substituents: OCH3 (S3: 81 h−1) <n-C4H9 (S4: 198 h−1) <H (S1: 215 h−1) <Cl (S2: 930 h−1). These results indicate that the turnover-limiting step is the hydride transfer from a ruthenium hydride (metal center) to the carbonyl carbon of a ketone (substrate) coordinating onto the ruthenium.38, 39, 40 The reduction of a long alkyl-aryl ketone (S5) by C1 proceeded with a relatively high yield (82%), whereas the TOF (103 h−1) was much smaller than that of S1, owing to the steric hindrance of S5. For aromatic substrates (S1–S6), the yields with C1 exhibited values similar to those of C7. These results demonstrate that the catalytic property of ruthenium bound by C1 to the substrate is the same as that of C7, with only the surroundings around the core ruthenium differing from those of C7.

For non-aromatic substrates (S7–S12), C1 showed a higher yield than C7 under the same conditions. The TOF for cyclohexanone (S8) with C1 also almost reached 1000. C1 completely induced a homogeneous reaction for non-aromatic ketones owing to the affinity between the PEG-based arms and the various substrates and products. In contrast, C7 was sometimes precipitated in the latter stage, because the non-aromatic alcohol products are often poor solvents for RuCl2(PPh3)3. Therefore, the high solubility and stability of C1 also afford higher yields and TOFs than C7.

Relative catalytic activity of Ru(II)-PEG Star and RuCl2(PPh3)3. To examine the reaction rate, the half-life periods of the substrates (time to reach 50% yield: T1/2) with PEG-Ru(II) Star (C1) were compared with those of RuCl2(PPh3)3 (C7). Figure 4 shows the relative reduction rate (R1/2(Star/Ru)=T1/2(C7)/T1/2(C1)) for aromatic (S1–S6) and non-aromatic (S7–S12) substrates. The R1/2(Star/Ru) results are almost 1.0 and over 1.0, indicating that C1 exhibited almost the same and/or superior activity to C7 depending on the substrates. The large value of R1/2(Star/Ru) tended to be more distinct for non-aromatic ketones such as cyclopentanone (S7) and 5-methyl 3-heptanone (S12; R1/2(Star/Ru)>2.0). The high activity is due to the affinity of the substrates to the core and that of the products to the arms. The respective compatibility led to the stable homogeneity of the star catalyst and the efficient diffusion cycle during the reaction, in which a substrate goes into the microgel core and the resultant product goes out from the reaction space. Therefore, the star polymer catalyst (C1) showed high substrate versatility in the transfer hydrogenation of ketones including aliphatic ones that were sometimes unfavorable for the conventional RuCl2(PPh3)3 (C7).

Figure 4
figure 4

Catalytic activity of C1–C7 evaluated for the relative reduction rate (R1/2(Star/Ru)) of various substrates (S1–S12). R1/2(Star/Ru)=T1/2(C7)/T1/2(C1). T1/2(C7 or C1): half-life periods of substrates catalyzed by C7 or C1, respectively. Conditions: substrate/Ru(II)/K2CO3=10/0.010/1.0 mmol in 2-propanol (10 ml) at 100 °C. S6: R1/2(Star/Ru) calculated from the respective yields (C1 and C7=22%) at 8 h.

Full size image

Catalyst recyclability. One of the attractive advantages of polymer-supported catalysts is catalyst recyclability and easy catalyst separation from the products.1, 2, 3, 4, 34, 35 Thus, the reusability of C1 was examined for the transfer hydrogenation of acetophenone (S1) and 2-octanone (S10; Figures 5a and b). The catalyst recycling was performed in three steps: (1) after a reaction, the solvent (2-propanol) was evaporated to yield the catalyst, K2CO3, and non-volatile organic compounds such as the unreacted substrate and resultant products; (2) the catalyst and the base were washed twice with hexane under argon to remove the non-volatiles; (3) the substrate and solvent were recharged for the next run. As shown in Figure 5a, C1 more efficiently reduced S1 for three cycles (yield (8 h): 89% (first); 80% (second); and 81% (third)) compared with C7 (yield (8 h): 88% (first); 74% (second); and 46% (third)). Furthermore, C1 carried out the reduction of S10 without any loss of activity for three cycles (Figure 5b). The solvent (hexane) exhibited no color (transparent) and no UV–Vis absorption from ruthenium complexes after washing C1 during the recycle experiments. This strongly indicates that the ruthenium complexes are steadily supported by the microgel core and do not leach from the core.18, 19 Ruthenium contamination in the products was characterized by UV–Vis spectroscopy, which is fully consistent with that by inductively coupled plasma atomic emission spectrometry.18 Thus, the almost-pure product was easily recovered from the precipitation of the reaction solution, followed by filtration and evaporation. The superior catalyst reusability and product recovery are due to the effective protection and immobilization of the ruthenium complexes by the crosslinked microgel core in C1.

Figure 5
figure 5

(a) Recycle experiments of C1 (light gray) and C7 (dark gray) for the hydrogenation of S1. (b) Recycle experiments of C1 for the hydrogenation of S10 (cycle: first (open squares), second (open circles) and third (filled triangles)). Conditions: S1 or S10/Ru(II)/K2CO3=15/0.015/1.5 mmol in 2-propanol (15 ml) at 100 °C for 8 h.

Full size image

Conclusion

We have demonstrated the transfer hydrogenation of various ketones with Ru(II)-PEG Star polymer catalysts in 2-propanol. The star catalysts were directly obtained from Ru(II)-catalyzed living radical polymerization of PEGMA and a sequential crosslinking reaction in the presence of a phosphine ligand monomer. Importantly, although the ruthenium complexes were placed in the core and shielded from the outside region, the star polymers efficiently and homogeneously reduced aromatic and non-aromatic ketones into their corresponding alcohols, in a manner superior to that of other polymer-bound catalysts and the original one. This high activity most likely arises from the unique ‘reaction space’, which consists of a ruthenium-embedded hydrophobic microgel core and amphiphilic and polar PEG-bearing arm polymers. Not only are the star polymers completely soluble in 2-propanol (solvent), but the arms and the core also exhibit a high affinity for products and substrates, respectively. The design around the catalytic site leads to high homogeneity during catalysis, independent of the substrate species, and smooth diffusion of the substrate and the resultant product around the microgel core. Furthermore, a PEG Star catalyst can be reused three times, which is better than conventional ruthenium, in addition to facile recovery of almost-pure products from the star catalyst. These reaction properties are also a result of the encapsulation and protection of the ruthenium complexes by the microgel core. Therefore, a PEG-armed ruthenium-bearing microgel star polymer catalyst provides a catalyst-enclosed reaction space that achieves high activity, versatility and catalyst recyclability in the transfer hydrogenation of ketones.

scheme 1

One-pot synthesis of Ru(II)-PEG Star polymer catalysts via Ru(II)-catalyzed living radical polymerization. PEG, poly(ethylene glycol).

scheme 2

Microgel-core reaction space of Ru(II)-PEG Star catalysts for the transfer hydrogenation of ketones: (a) supply of a hydrogen donor (2-propanol) to core-bound Ru(II) and (b) diffusion of substrates (ketones) and products (sec-alcohols) around the microgel-core reaction space. PEG, poly(ethylene glycol).