Magnetic Fe₃O₄ nanoparticles supporting Macmillan with controlled shell structure as an efficient and reusable catalyst for asymmetric reaction

Abstract

Magnetic nanoparticles grafted with chiral polymer brushes offer an effective way to bridge the gap between heterogeneous and homogeneous catalysis. Herein, we synthesized a series of chiral catalysts immobilized on magnetic Fe₃O₄ nanoparticles (MNPs) with a controlled shell structure by an efficient method. This involves the synthesis of “living” Fe₃O₄ nanoparticles with surface-bound vinyl groups and their subsequent grafting with chiral polymer brushes by the simple RAFT coupling reaction (i.e., chiral polymers with a dithioester end group couple with vinyl groups on the MNPs in the presence of a free radical initiator). The well-defined characteristics of the resulting MNPs allowed the first systematic study of the effects of various structural parameters (the chiral polymer molecular weight and grafting density) on their catalytic performance, which is of great importance for rationally designing more advanced chiral catalyst loading. In particular, the resulting Fe₃O₄-supported catalysts can effectively catalyze the asymmetric Diels–Alder reaction and can be recycled five times by simple magnetic separation while maintaining activity and selectivity.

---

1. Introduction

• Chiral imidazolidinone is a well-known enantioselective catalyst for many different chemical transformations such as Diels–Alder,^1,2 Michael addition,³ 1,3-dipolar addition,⁴ and Friedel–Crafts alkylation reactions.^5,6 However, the practical application of asymmetric organocatalysis is generally hindered due to the requirement for high catalyst loading and the difficulty in separating the catalyst from the product. There have been many examples of solid-supported Macmillan catalysts, aiming to maintain its high selectivity but still allow for recovery and reuse.^7–22 Previously reported approaches have been outlined in a review by Kristensen and Hansen which focused mainly on catalyst immobilization onto pre-formed solids,²³ including poly(ethylene glycol) (PEG) supports,²⁴ mesocellular foams,²⁵ sulfonated polystyrene and iron oxide nanoparticles.²⁶ However, the catalysts are often limited to immobilizing small molecules onto solids which are completely fixed on the substrate.²⁷ The disadvantage is lower effectiveness in terms of both activity and enantioselectivity than the corresponding homogeneous catalytic systems.^28–31

Core–shell particles (hairy particles), especially magnetic nanoparticles with iron oxide cores and polymer shells, have attracted increasing attention because of their unique magnetic responsiveness, good dispersivity and available surface functionalization.³² The polymer brushes on the surface of the particles are a dynamic system possessing a certain degree of mobility, so hairy magnetic particle-supported catalysts can combine the advantages of both homogeneous catalysts (high activity and high selectivity) and heterogeneous catalysts (easy separation). However, hairy ferromagnetic nanoparticle-supported chiral catalysts have never been reported until now. In addition, one of the major challenges in the synthesis of hairy particles is how to design polymer brushes with the desired structure and provide the material with enhanced performance. Recently, our group has reported the successful synthesis of a series of hairy particle-supported chiral catalysts by RAFT precipitation polymerization in combination with surface RAFT polymerization,^33,34 which proved to be a promising synthetic method in biocatalytic applications. However, some important structural parameters of the resulting hairy particles (e.g., the grafting density, molecular weight and distribution of the chiral polymer brushes) were difficult to control efficiently due to the complicated nature of the surface RAFT polymerization, which makes it difficult to obtain further insight into the detailed structure–property relationships of these hairy particles and thus influence their further development.

Many efficient coupling approaches have been developed for this purpose, such as copper-catalyzed Huisgen 1,3-dipolar cycloaddition of alkynes and azides³⁵ and thiol–ene chemistry.^36,37 Recently, Zhang et al. have prepared a hydrophilic molecularly imprinted polymer (MIP) microsphere by RAFT precipitation polymerization (RAFTPP) in combination with a surface RAFT coupling reaction, which provided an efficient method to control the polymer brushes on the surface of the MIP.³⁸ To our knowledge, this approach has not been widely used, and it has never been reported in the synthesis of functionalized MNPs.

Herein, we prepared a hairy Fe₃O₄ nanoparticle-supported Macmillan with a controlled shell structure by a facile and highly efficient approach, which showed excellent catalytic capacity and was recyclable and easily separated from the product. It involved the first synthesis of “living” MNPs with surface-bound vinyl groups and chiral polymers with dithioester groups, and their subsequent grafting with chiral polymer brushes by the simple coupling reaction of the macro-CTAs (CTA = chain transfer agent) with vinyl groups on the “living” MNPs. As illustrated in this work, a series of MNPs with controlled shell structures (MNP 1–5) were designed and synthesized (Scheme 1). In order to further clarify the effects of the shell structures on the catalytic performance, we have synthesized a MNP-supported Macmillan with no controlled shell structure by surface radical copolymerization (MNP-6). Some quantitative information, including the molecular weight and polydispersity of the grafted polymer brushes, and the catalyst loading, was characterized in detail. Furthermore, the obtained catalysts were used in a Diels–Alder (DA) reaction. The catalytic activity, asymmetric selectivity and recyclability were further studied.

Scheme 1 Synthesis of Fe₃O₄ nanoparticle-supported chiral catalysts by RAFT coupling reaction and surface radical copolymerization.

2. Experimental

2.1 Materials

N,N-Dimethylformamide (DMF, Jiangtian Chemicals, China) was purified by distillation under vacuum. Cumyl dithiobenzoate (CDB),³⁹ p-[(1-methyl-2,2-dimethyl-5-oxo-4-imidazolidinyl)methyl]phenyl methacrylate (M)⁷ and Fe₃O₄ nanoparticles⁴⁰ were prepared following the literature procedure. (3-Mercaptopropyl)trimethoxysilane (MPTMS, Aladdin) and CF₃COOH (TFA, Aladdin) were used as received, and all the other chemicals were used as received.

2.2 Measurements

¹H NMR spectra were recorded on a Bruker AV-400 NMR spectrometer. FT-IR spectra were recorded on a Nicolet NEXUS Fourier transform infrared spectrometer using KBr pellets. The morphologies of the prepared MNPs were studied by field emission scanning electron microscopy (FESEM, FEI, NovaNano SEM450) and transmission electron microscopy (TEM, JEOL-2010, 200 kV). The molecular weights of the polymers were measured by gel permeation chromatography (GPC) using polyethylene oxide (PEO) as a standard, DMF as a mobile phase and a refractive index (RI) detector. Two Shodex LF-404 columns were conditioned at 35 °C and a flow rate of 0.3 mL min⁻¹. High-performance liquid chromatography (HPLC) analysis was carried out on Agilent TM 1100 HPLC equipment using a DAICEL CHIRALPAK OJ-H chromatographic column.

2.3 Synthesis of magnetite nanoparticles with surface-bound vinyl groups

The magnetite NPs with surface-bound vinyl groups were synthesized according to the following procedure: 2 g (8.63 mmol) Fe₃O₄ was mechanically dispersed in 200 mL toluene and 10.45 mL (43.15 mmol) 3-(trimethoxysilyl)propyl methacrylate was added, together with 0.1 wt% butylated hydroxytoluene (BHT) as inhibitor according to a procedure previously reported in the literature.³⁸ The reaction mixture was kept at 25 °C for 24 h under nitrogen atmosphere with vigorous mechanical stirring. Then, the solvent was removed by decantation and the obtained MNPs-MA (where MA = methacrylate) were washed twice with ethanol and methylene chloride and dried under vacuum for 48 h. The MNPs with surface-bound vinyl groups were found to have been successfully prepared with a yield of 95%.

2.4 Macro-CTA synthesis

A series of macro-CTAs with different molecular weight (Scheme 1) were prepared following a typical procedure: M (1.13–3.39 g, 10–30 mmol), CDB (30.4 mg, 0.11 mmol), azobisisobutyronitrile (AIBN) (3.3 mg, 0.02 mmol), and DMF (5 mL) were added into a two-neck round-bottom flask (25 mL) successively. After being degassed with five freeze–pump–thaw cycles, the flask was sealed and immersed in a thermostated oil bath at 75 °C and stirred for 24 h. The supernatant solutions were precipitated in ether, filtered, and washed with a large amount of ether and dried at 30 °C under vacuum for 48 h, and the macro-CTAs were obtained.

2.5 Grafting chiral polymer brushes onto MNPs by RAFT coupling chemistry

The MNPs with surface-grafted chiral polymer brushes were prepared by the coupling reaction between the macro-CTAs and MNPs with surface-bound vinyl groups in the presence of a free radical initiator. A typical procedure for the synthesis of MNPs with chiral polymer brushes was as follows: the MNPs with surface-bound vinyl groups (150 mg), macro-CTAs with different molecular weights (M_n = 3590, 18 700 and 25 600) (0.15 g), AIBN (3.3 mg, 0.02 mmol), and DMF (30 mL) were added into a two-neck round-bottom flask (50 mL) successively. After being degassed with five freeze–pump–thaw cycles, the flask was sealed and immersed in a thermostated oil bath at 75 °C and stirred for 24 h. After magnetic separation, the resulting solid products were thoroughly washed with DMF and methanol, and then dried at 30 °C under vacuum to obtain a powder. The MNPs were assigned as MNP (1–3).

Macro-CTAs (M_n = 25 600) with different weight amounts (0.3 and 0.5 g) were grafted onto the MNPs (0.15 g) following the same experimental procedure; the MNPs were assigned as MNP (4–5) (see Table 1).

Table 1 MNPs with different molecular weights and grafting densities synthesized via polymerization from Fe₃O₄@C=C

Catalysts	Weight contents of macro-CTA (g)	Mn, GPC, macro-CTA	PDI, GPC	ΔW^a (%)	Grafting method	Graft density (μmol m^−2)	Catalyst loading (mmol g^−1)
a W60–730 obtained from TGA.
MNP-1	0.15	3590	1.14	41.0	RAFT coupling	4.10	0.68
MNP-2	0.15	18 700	1.39	47.8	RAFT coupling	1.09	0.79
MNP-3	0.15	25 600	1.32	34.9	RAFT coupling	0.41	0.58
MNP-4	0.30	25 600	1.32	60.4	RAFT coupling	1.92	1.00
MNP-5	0.50	25 600	1.32	70.7	RAFT coupling	2.30	1.17
MNP-6		24 300	2.21	73.0	Surface radical copolymerization	2.73	1.20

---

2.6 Grafting chiral polymer brushes onto MNPs by surface radical copolymerization

The MNPs with grafted polymer brushes were prepared via surface radical copolymerization according to the following procedure: MNPs (0.3 g), M (3.27 g, 10.89 mmol), AIBN (3.3 mg, 0.02 mmol), and DMF (30 mL) were added into a two-neck round-bottom flask (50 mL) successively. After being degassed with five freeze–pump–thaw cycles, the flask was sealed and immersed in a thermostated oil bath at 75 °C and stirred for 24 h. After magnetic separation, the resulting solid products were thoroughly washed with DMF and methanol, and then dried at 30 °C under vacuum to obtain a powder. The MNP was assigned as MNP-6.

The addition of the radical initiator AIBN into the above polymerization systems also led to the generation of free chiral polymers, which were obtained by precipitating the supernatant solutions (after the centrifugation of the reaction mixtures) into ethyl ether, filtered, and then dried at 40 °C under vacuum for 48 h.

2.7 General procedure for the Diels–Alder catalysis reaction

The Diels–Alder reaction was carried out as follows. Reactions were all carried out at identical reagent concentrations and the concentration of MNPs was varied to maintain the same catalyst loading: the catalytically active particles were weighed into a vial (10 mol% catalyst loading) and dispersed in the appropriate solvent. TFA (11.4 μL, 0.1 mmol) was then added, followed by the aldehyde (63.6 μL, 0.48 mmol) and the solution was allowed to stir for a few minutes before cyclopentadiene (0.168 mL, 1.96 mmol) was added. An aliquot (0.1 mL) was stirred in H₂O : TFA : CHCl₃ (1 : 1 : 2) (4 mL) before being neutralized with NaHCO₃ (2 mL) and then extracted into Et₂O (25 mL). The resulting residue was purified by silica gel chromatography to afford the product.

3. Results and discussion

3.1 Preparation of MNPs grafted with methacrylate end groups (Fe₃O₄@C=C)

The functionalization of the magnetite nanoparticles with methacrylate polymerizable units was aimed at covalently linking the inorganic entities within the final grafting polymerization. The successful attachment of the hybrid 3-(trimethoxysilyl)propyl methacrylate on the magnetite corona was verified through FTIR spectroscopy. In Fig. 1, the FTIR spectrum of the modified magnetite is compared to that of the pure Fe₃O₄ nanoparticles. It can be remarked that in the magnetite FTIR spectrum, the most important absorption bands were those attributed to the hydroxyl groups located at the surface of the nanoparticles (ν_OH = 3394 cm⁻¹) and to the Fe–O linkages (ν_Fe–O = 558 cm⁻¹), while upon modification with 3-(trimethoxysilyl)propyl methacrylate, the advent of new absorption bands was observed (C=O; C=C); more importantly, a band attributed to the newly created –Si–O–Fe bonds (1101 cm⁻¹) was also observed. From these results, one can conclude that the magnetite nanoparticles were functionalized with trimethoxysilyl units from the silane coupling agent (Fig. 1b) and the supplied methacrylate end groups. Consequently, the 3-(trimethoxysilyl)propyl methacrylate-grafted magnetic nanoparticles became hydrophobic compared to the initial Fe₃O₄ NPs, which are hydrophilic in nature. The grafting density of the methacrylate end groups on the Fe₃O₄ surface was determined by thermogravimetric analysis (TGA) using eqn (1):⁴¹where W_60–730 is the weight loss between 60 and 730 °C corresponding to the decomposition of the ungrafted Fe₃O₄ corrected using the thermal degradation profile, and M is the molecular weight of the grafted methacrylate group. S_spec and W_Fe3O4@C=C are respectively the specific surface area (38.88 m² g⁻¹ for Fe₃O₄ nanoparticles) and the weight loss of Fe₃O₄ determined before grafting. The average surface grafting density of the methacrylate end groups on the above-obtained Fe₃O₄ particles was evaluated to be about 13.5 μmol m⁻², revealing the rather high densities of methacrylate groups on the Fe₃O₄ particles.

Fig. 1 Fourier transform infrared (FT-IR) spectra of non-grafted Fe₃O₄ (a), Fe₃O₄@C=C (b), MNPs (1–6) grafted with chiral polymer (c–h).

3.2 Preparation of MNPs grafted with chiral polymer brushes

In comparison to other methods for surface-initiated grafting polymerization, our approach has some distinct advantages, such as its synthetic flexibility, because MNPs grafted with methacrylate end groups allow the facile and controlled one-pot synthesis of MNPs with high densities of surface-bound vinyl groups, and the subsequent simple macro-CTA-mediated coupling reaction can in principle allow the chiral polymer to be easily grafted onto the MNP surfaces. In this method, the well-defined characteristics of the resulting MNPs make it possible to systematically study the effects of various structural parameters (i.e., molecular weights, and grafting densities) of the grafted chiral polymer brushes on the particles, which provide important information for rationally designing more advanced MNP-supported catalysts.

To evaluate the scope of our strategy and demonstrate its general applicability, a series of well-defined macro-CTAs with different molecular weights (M_n) were first prepared by the end-group modification of the chiral polymer via RAFT polymerization (Fig. 2). The obtained macro-CTAs were characterized by GPC, from which the number-average molecular weights (M_n) of the grafted polymer brushes on the particles were evaluated to be 3590, 18 700 and 25 600, and their polymer dispersity index (PDI) range was 1.14–1.39. The low polymer dispersity of the polymer brushes suggested that the RAFT polymerization occurred in a well-controlled manner. They were then allowed to couple with the vinyl groups of the above-obtained MNPs in the presence of a certain amount of AIBN under mild reaction conditions, resulting in a series of MNPs with different molecular weights.

Fig. 2 The synthesis of the macro-CTA.

The surface-grafting density is the key factor affecting the catalytic performance. The grafting density of the chiral polymer on the Fe₃O₄ surface was determined by TGA using eqn (2):⁴¹

where M is the molecular weight of the grafted polymer brush. S_spec and W_Fe3O4@C=C are respectively the specific surface area (38.88 m² g⁻¹ for Fe₃O₄ nanoparticles) and the weight loss of Fe₃O₄ determined before chiral polymer grafting. The average surface grafting densities of the chiral polymer brushes on the above-obtained Fe₃O₄ particles were evaluated to be about 0.41–4.1 μmol m⁻², revealing the presence of rather densely grafted polymer brushes on the Fe₃O₄ particles. It is interesting to note that the molecular weights of the macro-CTAs are the key factor influencing the grafting densities of the polymer brushes. With the growth of the molecular weights, the grafting densities decreased considerably. According to the increased weight percentages, the catalyst loading for the final Fe₃O₄-supported Macmillan can be obtained (Table 1), and the molecular weights have a negligible influence on this value.

Well-defined macro-CTAs with M_n = 25 600 were used to couple with the vinyl groups of the above-obtained MNPs in the presence of different amounts of macro-CTA under the same reaction conditions (Table 1, MNP 3–5), and the polymer grafting densities on the particle surfaces were calculated as 0.41–2.3 μmol m⁻². Different weight increases were also observed for these Fe₃O₄ particles, suggesting an increase in the grafting densities with greater macro-CTA contents. To gain more insight into the grafting reaction, the evolution of the grafting density (determined by experimental measurement) as a function of the reaction time and the macro-CTA amounts was investigated, as shown in Fig. 3. The grafting density increased with increasing reaction time and the addition of macro-CTA contents. The maximum grafting density that could be achieved was about 2.73 μmol m⁻², which is greater than the values given in the literature for the grafting of a series of polymer brushes (Table 1).²⁷ This coupling method is particularly attractive because of the elaboration of polymer shell layers with controlled molecular weight and grafting density.

Fig. 3 Grafting density as determined from the experimental results, using eqn (1), as a function of the reaction time; the inset shows the grafting density as a function of the macro-CTA amount.

In a further experiment, the 3-(trimethoxysilyl)propyl methacrylate-grafted magnetic nanoparticles were found to undergo copolymerization with the chiral monomers to obtain MNP-supported catalysts. A larger number of monomers were introduced into the copolymerization, leading to a relatively high grafting density (MNP-6). In addition, this method only led to MNPs bearing polymer brushes with poorly controlled molecular weights and distribution (24 300, 2.21) compared with our above strategy, which heavily influenced the catalytic performance of the obtained MNPs.

SEM was firstly utilized to characterize the structure of the above-obtained MNPs. In contrast to the ungrafted MNPs, the MNPs grafted with chiral polymer brushes were coated by a thick layer of film (Fig. 4a–d), suggesting the successful grafting of polymer brushes on the MNPs. TEM was performed to investigate the morphology of the magnetic nanoparticles with core–shell structures, and the results are shown in Fig. 4e and f. Moreover, the magnetite nanoparticles after modification with chiral polymer chains were better dispersed than the unfunctionalized ones.

Fig. 4 SEM images of non-grafted Fe₃O₄ (a), MNP-1 (b), MNP-2 (c) and MNP-3 (d); TEM images of representative grafted MNP-4 (e) and MNP-5 (f).

The FT-IR spectra of Fe₃O₄ grafted with chiral polymer brushes are shown in Fig. 1c–h. In Fig. 1c, the band at 1750 cm⁻¹ corresponds to the C=O stretching vibration for the polymer brush. The bands at 1506 cm⁻¹ and 1402 cm⁻¹ can be attributed to the C–O stretching vibration (amide I band), N–H deformation vibration (amide II band). The band at 580 cm⁻¹ is consistent with the Fe–O stretching vibration. All the characteristic bands in the FT-IR spectrum demonstrated that Fe₃O₄ grafted with polymer brushes had been successfully prepared. In the six modified MNPs, the intensity of the band at about 1641 cm⁻¹ (C=C) became markedly weaker than that of Fe₃O₄@C=C; in particular, for MNP-5, the characteristic band fully disappeared, providing further evidence of the efficient grafting density.

In the present study, we prepared six systems of grafted MNPs with varied molecular weights and grafting densities (see Table 1). The magnetic content could be easily controlled by adjusting the weight content of the magnetic NPs. The thermal properties and the actual magnetic content of the MNPs were measured by the TGA technique, as illustrated in Fig. 5. Compared to the MA-grafted MNPs, the TGA curve of MNP-1 showed a weight loss of 41%. The weight loss was due to the efficient grafting of the chiral polymer brushes. This also offered further evidence for the assumption that the Macmillan catalyst was successfully attached on the MNPs. The TGA curves of the MNP (2–6)-supported Macmillans are presented in Fig. 5. Their weight losses were 47.8%, 34.9%, 60.4%, 70.7% and 73.0% when the temperature increased from 60 °C to 730 °C. The TGA data was in accordance with the corresponding experimental values (Table 1).

Fig. 5 TGA of the MNPs.

3.3 Hairy particle-supported organocatalyst for the asymmetric DA reaction

The catalytic efficiency of the Macmillan-functionalized MNPs was investigated using the DA reaction between cyclopentadiene and cinnamaldehyde (as shown in Scheme 2). The reactions were first carried out at 10 mol% catalyst loading with different solvents at 25 °C (Table 2). The monomer (M) was used as the reference, which gave a 62% yield (with 91% enantiomeric excess (ee) for endo, 91% ee for exo) in CH₃CN/H₂O (95 : 5 v/v) for 21 h (Table 2, entry 1). When CH₃CN/H₂O was used as solvent, MNP (3–5) as catalysts showed good catalytic activity: the yields were >75%. Specifically, MNP-3 and MNP-4 reached 80% and 85% yields in the same reaction time while MNP-5 reached a 90% yield. The reaction catalyzed by the MNP-5-supported system had the best catalytic activity. The surface-grafting density was the crucial factor affecting the catalytic performance. In contrast, the reaction catalyzed by the MNP-supported system was more efficient than the homogeneous monomeric catalyst. A clear trend was observed: the catalytic activity increased with the increase of polymer grafting density (or catalyst loading). The reaction kinetics of the MNPs was also investigated to determine whether the grafting density affected the rate of reaction (Fig. 6). The kinetics showed that the catalytic efficiency was enhanced as the polymer-grafting density increased. However, the enantioselectivity showed the opposite trend: MNP-5, which had a relatively high grafting density, showed good catalytic activity but less than ideal enantioselectivities (78% ee for endo, and 76% ee for exo). MNP-6, which had a high grafting density but no controlled molecular structure, showed good catalytic activity; however, the enantioselectivities (41–75% ee for the major product) were poor.

Scheme 2 Model Diels–Alder reaction, where there are four possible products: the endo and exo products of both enantiomers, R is variable and R = H, C₃H₇, 4-PhNO_2..

Table 2 Solvent effects on the Diels–Alder reaction between cyclohexanone and cinnamaldehyde

Catalysts	Solvent	Reaction time (h)	% yield^a	Exo/endo^b	% ee^c (endo)	% ee^c (exo)
a Yield determined after chromatographic purification.b Determined by ^1H NMR spectroscopic analysis of the product.c Determined by HPLC using a chiral column.
M	CH3CN/H2O	21	62	1.47 : 1	91	91
MNP-1	CH3CN/H2O	21	60	1.20 : 1	85	87
MNP-2	CH3CN/H2O	21	75	1.12 : 1	76	83
MNP-3	CH3CN/H2O	21	80	1.22 : 1	89	86
MNP-4	CH3CN/H2O	21	85	1.22 : 1	99	43
MNP-5	CH3CN/H2O	21	90	1.29 : 1	78	76
MNP-6	CH3CN/H2O	21	89	1.22 : 1	75	41

---

Fig. 6 The kinetics of the reaction catalyzed by the MNPs in CH₃CN/H₂O (95 : 5 v/v).

Reactions were also carried out to determine whether the polymer molecular weight affected the rate of reaction. We investigated the catalytic performance of MNP (1–3). MNP-1 gave a 60% yield (with 85% ee for endo, 87% ee for exo) in CH₃CN/H₂O (95 : 5 v/v) for 21 h, and MNP-2 gave a 75% yield (with 76% ee for endo, 83% ee for exo). In contrast, MNP-3, with the biggest molecular weight, had the best catalytic activity (an 80% yield) and good selectivity (with 89% ee for endo, 86% ee for exo) in CH₃CN/H₂O for 21 h. According to the above results, higher grafting densities showed higher catalytic activity. MNP-3, with a higher molecular weight but lower grafting density, showed good catalytic activity and enantioselectivity, which demonstrated that the polymer molecular weight was a more decisive factor affecting the catalytic performance than the grafting density.

The versatility of our MNP-supported Macmillan system was demonstrated by the catalytic efficiency of MNP-3 in the Diels–Alder reaction of a range of substrates. Cyclopentadiene was reacted with a range of aldehydes (Table 3) structurally similar to the one presented in Scheme 2 but with a different R group. MNP-3 catalyzed the reactions, achieving good yields between 62% and 82% and good enantioselectivities, demonstrating the catalytic ability of the MNPs in a range of Diels–Alder reactions.

Table 3 A range of Diels–Alder reactions (Scheme 2) catalyzed by MNP-3 in CH₃CN/H₂O (95 : 5 v/v) for 21 h

R = Dienophile	% yield^a	Exo/endo^b	% ee^c (endo)	% ee^c (exo)
a Yield determined after chromatographic purification.b Determined by ^1H NMR spectroscopic analysis of the product.c Determined by HPLC using a chiral column.
H	62	1.36 : 1	90	91
C3H7	65	1.27 : 1	85	89
4-PhNO2	82	1.24 : 1	87	82

---

Recently, Hansen and co-workers immobilized the Macmillan catalyst into a lightly cross-linked nanogel system via an emulsion polymerization process, embedding the catalyst in a hydrophobic environment.⁷ However, a decrease in enantioselectivity with recycling was reported (81% to 51% ee after four cycles). The recycling potential of the MNP-3-supported catalytic system was studied using the same DA reaction between cyclopentadiene and cinnamaldehyde in CH₃CN/H₂O. This adjustment was done to keep the reaction at 10 mol% catalyst loading throughout the different cycles. The procedure was simple: the Fe₃O₄-supported Macmillan was firstly applied to the reaction, and then recovered after completion of the reaction by separating it from the reaction mixture with an external permanent magnet (Fig. 7b), and washing it with H₂O and diethyl ether alternately. It was reused directly for the next cycle without further treatment. Fig. 8 shows the recycling efficiency of the supported system. The catalyst was successfully used in 5 cycles without losing significant activity or selectivity, recovering >98% of the catalyst in each cycle.

Fig. 7 Photographs of the dispersion of MNP-3 (left) in CH₃CN/H₂O for 1 h, and (right) the separation of MNP-3 with an external magnet.

Fig. 8 Diels–Alder reaction between cyclohexanone and cinnamaldehyde by MNP-3 in CH₃CN/H₂O for 21 h in multiple cycles.

4. Conclusions

We have demonstrated for the first time the efficient synthesis of a series of well-defined MNPs with chiral polymer brushes of desired chemical structures and molecular weights by the facile RAFT coupling chemistry. The factors influencing the structural parameters (including the molecular weights, and grafting densities) of the chiral polymer brushes on the catalytic performance of the grafted MNPs were determined and the general applicability of the strategy was also demonstrated. The catalysts can also be well dispersed in the solvent, easily magnetically recovered from the reaction mixture, and reused several times without significant loss of activity and selectivity.

Acknowledgements

This work was supported by the National Natural Science Foundation of China (No. 21204019), the Post-doctoral Foundation of China (No. 2012M521398), the Post-doctoral Foundation of Henan Province, the International Cooperation Project of Henan Province (No. 134300510055), and the Youth Backbone Teacher Foundation of Henan Normal University.

Notes and references

1. S. Guizzetti, M. Benaglia and J. S. Siegel, Chem. Commun., 2012, 48, 3188–3190.
2. N. B. Pramanik, G. B. Nando and N. K. Singha, Polymer, 2015, 69, 349–356.
3. J. W. Xie, L. Yue, W. Chen, W. Du, J. Zhu, J. G. Deng and Y. C. Chen, Org. Lett., 2007, 9, 413–415.
4. X. H. Chen, Q. Wei, S. W. Luo, H. Xiao and L. Z. Gong, J. Am. Chem. Soc., 2009, 131, 13819–13825.
5. M. Bandini, M. Fagioli, P. Melchiorre, A. Melloni and A. Umani-Ronchi, Tetrahedron Lett., 2003, 44, 5843–5846.
6. M. Bandini, M. Fagioli, M. Garavelli, A. Melloni, V. Trigari and A. Umani-Ronchi, J. Org. Chem., 2004, 69, 7511–7518.
7. B. L. Moore, A. Lu, D. A. Longbottom and R. K. O'Reilly, Polym. Chem., 2013, 4, 2304–2312.
8. A. G. Hu, S. Liu and W. B. Lin, RSC. Adv., 2012, 2, 2576–2580.
9. S. Itsuno and M. M. Hassan, RSC. Adv., 2014, 4, 52023–52043.
10. M. Benaglia, G. Celentano, M. Cinquini, A. Puglisi and F. Cozzi, Adv. Synth. Catal., 2002, 344, 149–152.
11. A. Puglisi, M. Benaglia, M. Cinquini, F. Cozzi and G. Celentano, Eur. J. Org. Chem., 2004, 2004, 567–573.
12. M. S. Chauhan, P. Kumar and S. Singh, RSC. Adv., 2015, 5, 52636–52641.
13. J. Y. Shi, C. A. Wang, Z. J. Li, Q. Wang, Y. Zhang and W. Wang, Chem.–Eur. J., 2011, 17, 6206–6213.
14. C. A. Chang, Y. Zhang, J. Y. Shi and W. Wang, Chem.–Asian J., 2013, 8, 1110–1114.
15. C. O. Liang and J. M. J. Fréchet, Macromolecules, 2005, 38, 6276–6284.
16. C. S. Pecinovsky, G. D. Nicodemus and D. L. Gin, Chem. Mater., 2005, 17, 4889–4891.
17. S. A. Selkälä, J. Tois, P. M. Pihko and A. M. P. Koskinen, Adv. Synth. Catal., 2002, 344, 941–945.
18. N. Haraguchi, H. Kiyono, Y. Takemura and S. Itsuno, Chem. Commun., 2012, 48, 4011–4013.
19. T. Mitsudome, K. Nose, T. Mizugaki, K. Jitsukawa and K. Kaneda, Tetrahedron Lett., 2008, 49, 5464–5466.
20. Q. Chu, W. Zhang and D. P. Curran, Tetrahedron Lett., 2006, 47, 9287–9290.
21. Z. Lin, Z. Chen, G. Yang and C. Lu, Catal. Commun., 2013, 35, 1–5.
22. J. K. Park, P. Sreekanth and B. M. Kim, Adv. Synth. Catal., 2004, 346, 49–52.
23. T. E. Kristensen and T. Hansen, Eur. J. Org. Chem., 2010, 75, 3179–3204.
24. T. E. Kristensen, K. Vestli, M. G. Jakobsen, F. K. Hansen and T. Hansen, J. Org. Chem., 2010, 75, 1620–1629.
25. T. E. Kristensen, K. Vestli, M. G. Jakobsen, F. K. Hansen and T. Hansen, Adv. Synth. Catal., 2002, 344, 941–945.
26. Y. Zhang, L. Zhao, S. S. Lee and J. Y. Ying, Adv. Synth. Catal., 2006, 348, 2027–2032.
27. C. S. Gill, W. Long and C. W. Jones, Catal. Lett., 2009, 131, 425–431.
28. N. Haraguchi, Y. Takemura and S. Itsuno, Tetrahedron Lett., 2010, 51, 1205–1208.
29. P. Riente, J. Yadav and M. A. Pericàs, Org. Lett., 2012, 14, 3668–3671.
30. J. Adhikary, A. Datta, S. Dasgupta, A. Chakraborty, M. I. Menendez and T. Chattopadhyay, RSC. Adv., 2015, 5, 92634–92647.
31. M. B. Gawande, P. S. Branco and R. S. Varma, Chem. Soc. Rev., 2013, 8, 3371–3393.
32. B. Zhao, X. M. Jiang, D. J. Li, X. G. Jiang, T. G. O'Lenick, B. Li and C. Y. Li, J. Polym. Sci., Part A: Polym. Chem., 2008, 46, 3438–3446.
33. X. J. Li, M. Q. Chen, B. L. Yang, S. L. Zhang, X. B. Jia and Z. G. Hu, RSC. Adv., 2014, 4, 43278–43285.
34. X. J. Li, B. L. Yang, X. B. Jia, M. Q. Chen and Z. G. Hu, RSC. Adv., 2015, 5, 89149–89156.
35. A. S. Goldmann, A. Walther, L. Nebhani, R. Joso, D. Ernst, K. Loos, C. Barner-Kowollik, L. Barner and A. H. E. Muller, Macromolecules, 2009, 42, 3707–3714.
36. G. L. Li, D. Wan, K. G. Neoh and E. T. Kang, Macromolecules, 2010, 43, 10275–10282.
37. G. L. Li, L. Q. Xu, K. G. Neoh and E. T. Kang, Macromolecules, 2011, 44, 2365–2370.
38. M. Zhao, X. J. Chen, H. T. Zhang, H. S. Yan and H. Q. Zhang, Biomacromolecules, 2014, 15, 1663–1675.
39. T. P. Le, G. Moad, E. Rizzardo and S. H. Thang, PCT Int. Appl., WO 98/01478, 1998.
40. T. Q. Zeng, W. W. Chen, C. M. Cirtiu, A. Moores, G. H. Song and C. J. Li, Green Chem., 2010, 12, 570–573.
41. C. Bartholome, E. Beyou, E. Bourgeat-Lami, P. Chaumont and N. Zydowicz, Macromolecules, 2003, 36, 7946–7952.

---