A solution to achieve good reusability of MNPs Fe₃O₄-supported (S)-diphenylprolinoltrimethylsilyl ether catalysts in asymmetric Michael reactions

Abstract

A new supported (S)-diphenylprolinol trimethylsilyl ether (Fe₃O₄/PVP@SiO₂/ProTMS) using polyvinylpyrrolidone (PVP)-modified MNPs Fe₃O₄ as a support was prepared via one-pot surface-modification, and exhibited high yield (93–99%), excellent diastereoselectivities (syn/anti = 81–96 : 19–4) and enantioselectivities (95–98% ee) in the asymmetric Michael addition of propanal to various nitroalkenes. TGA, XRD, IR, SEM, TEM, elemental analysis and N₂ adsorption–desorption isotherm demonstrated that the adsorption of PVP onto MNPs Fe₃O₄ resulted in the spectacular change in the chemical composition, surface morphology and pore structure of Fe₃O₄/PVP@SiO₂/ProTMS. The catalyst could be easily separated from the reaction by an external magnet and reused for ten times with high yields (77–99%) and unchanged excellent stereoselectivities (97–98% ee and syn/anti = 96/4) in the Michael addition for the first time.

---

Introduction

The Jørgensen–Hayashi organocatalysts, i.e., α,α-diarylprolinoltrimethylsilyl ethers,¹ which activate aldehydes via an enamine² and α,β-unsaturated carbonyl compounds via an iminium ion,³ can be considered as one of the most famous leading actors among the versatile organocatalysts in the asymmetric reactions such as Aldol, Michael, cycloaddition, epoxidation and desymmetrization.⁴ Moreover, Jørgensen–Hayashi catalysts have been utilized in asymmetric tandem and multi-component reactions, through which large quantities of enantiopure compounds with multiple stereogenic carbon centers can be conveniently constructed.⁵ However, the usage of organocatalyst, typically 10–30 mol%, was needed to achieve desirable catalytic performance in most cases of organocatalyzed reactions. From the viewpoint of green chemistry, several efficient and economical protocols using polymer,⁶ inorganic material,⁷ ionic liquid⁸ and graphene oxide⁹ as catalyst supports were developed to achieve the reuse of expensive Jørgensen–Hayashi catalysts.

On the other hand, due to the potential advantages of inexpensive, non-toxic, chemically stable properties and re-collection by simple magnetic decantation without filtration or centrifugation, the magnetic nanoparticles (MNPs) of Fe₃O₄ had emerged recently as a new type of catalyst support in heterogeneous asymmetric organocatalysis with good catalytic performances, especially in Aldol addition.¹⁰ However, in the asymmetric Michael addition of aldehyde to nitroalkene, all the reported MNPs Fe₃O₄-supported Jørgensen–Hayashi catalysts gave the unsatisfactory reusability. In 2010, W. Wang et al. developed Fe₃O₄@SiO₂-supported Jørgensen–Hayashi catalyst, whose yield and stereoselectivity significantly decreased to 54%, syn/anti = 83 : 17 and 85% ee in the third run.¹¹ One year later, A. Pericás et al. immobilized Jørgensen–Hayashi catalyst onto azide functionalized MNPs Fe₃O₄@SiO₂. Unfortunately, the similar sharp drop in catalytic performance was found in the fourth run (57%, syn/anti = 72 : 28 and 92% ee).¹² Recently, A. Ouali et al. also reported on a poly(benzylazide)styrene-functionalized magnetic Co/C-supported (S)-α,α-diphenylprolinoltrimethylsilyl ether with a sharp drop in the activity in the 3rd run.¹³ Untill now, the good reusability of MNPs-supported Jørgensen–Hayashi catalyst could not be achieved in the asymmetric Michael addition of aldehyde to nitroalkene.

In our attempts to solve the reusability of MNPs-supported Jørgensen–Hayashi catalyst, we report the supporting of (S)-α,α-diphenylprolinoltrimethylsilyl ether onto polyvinylpyrrolidone (PVP)-modified MNPs of Fe₃O₄ through the hydrolysis of –Si(OCH₃)₃ by the facile one-pot surface-modification (Scheme 1). Due to the vital function of PVP, the as-synthesized Fe₃O₄/PVP@SiO₂/ProTMS possessed the excellent catalytic performances (93–99% yield, syn/anti = 81–96 : 19–4, 95–98% ee) and good reusability with the unchanged excellent stereoselectivities (syn/anti = 96 : 4, 98% ee) and moderate yield (82%), even in the 10^th run.

Scheme 1 The one-pot preparation of PVP-modified MNPs Fe₃O₄-supported Jørgensen–Hayashi organocatalysts.

Experimental

Characterization

¹H and ¹³C NMR spectra were conducted using a Bruker av-600 NMR instruments, in which all chemical shifts were reported down-field in ppm relative to the hydrogen and carbon resonances of TMS. FT-IR spectroscopy was performed on a Perkin-Elmer model GX spectrometer using the KBr pellet. Thermogravimetry-differential thermal analysis was carried out on a SBTQ600 thermal analyzer at a heating rate of 10 °C min⁻¹ from 40 to 800 °C using N₂ as protective gas (100 mL min⁻¹). Elemental analysis was performed using a vario Micro cube elemental analyzer instrument. The surface morphologies of the samples were observed by JSM-6510LV scanning electron microscopy and Tecnai G2 F20 transmission electron microscope, operated at 20 kV/15 mA and 200 kV respectively. X-ray powder diffraction patterns were detected on an XRD-7000 S/L instrument: Cu-Kα radiation, X-ray tube settings of 40 kV/30 mA and a step size of 2° min⁻¹ in the 10–100° (2θ) range. N₂ adsorption–desorption isotherm was carried out at 77.4 K using an Autosorb-1 apparatus (Quantachrome), in which the sample was degassed at 120 °C for 12 h before measurement, and the specific surface area and the pore diameter were calculated by the BET method and BJH model, respectively. Magnetization curve was determined by HH-15 vibrating sample magnetometer. The syn/anti ratios and % ee values of products were determined by ¹H NMR and Agilent LC-1200 HPLC using Daicel Chiralpak OD-H 4.6 mm × 25 cm column under 20 °C, 220 nm and 1.0 mL min⁻¹ conditions, respectively.

Synthesis of Jørgensen–Hayashi catalyst 5

Synthesis of compound 1. To a anhydrous THF (20 mL) suspension of NaH (0.75 g, 18.8 mmol) was added in turn 30 mL of THF solution containing Boc-L-hydroxyproline (2.0 g, 8.7 mmol), 18-crown-6 (0.23 g, 0.86 mmol) and 4-vinylbenzyl chloride (90%, 3.66 g, 21.6 mmol) at room temperature. The reaction mixture was warmed to 50 °C for 12 h. The resulting mixture was added water (30 mL), extracted with cyclohexane (50 mL × 2) to remove residual 4-vinylbenzyl chloride, adjusted to pH in the range of 2–3 by saturated aqueous KHSO₄ solution and extracted with ethyl acetate (50 mL × 3). The combined organic layers were dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. The residue was purified by silica gel gradient column chromatography using petroleum ether/ethyl acetate (v/v = 4/1 to 1/2) as eluents to afford yellow oil 1 (2.5 g, 80%). ¹H NMR (600 MHz, CDCl₃, TMS) δ 1.42–1.47 (m, 9H, CH₃), 2.10–2.46 (m, 2H, NCHCH₂), 3.51–3.73 (m, 2H, NCH₂), 4.17–4.17 (m, 1H, NCH), 4.36–4.53 (m, 3H, OCH, OCH₂), 5.24 (d, J = 10.8 Hz, 1H, =CH₂), 5.74 (d, J = 17.6 Hz, 1H, =CH₂), 6.70 (dd, J = 10.9, 17.6 Hz, 1H, =CH), 7.26–7.39 (m, 4H, Ar-H), 8.90 (s, 1H, COOH) ppm; ¹³C NMR (151 MHz, CDCl₃): δ 28.2 (CH₃), 36.7 (NCHCH₂), 52.0 (NCH₂), 57.9 (NCH), 71.0, 75.9, 80.7 (OCH₂, OCH and OC(CH₃)₃), 114.0 (=CH₂), 126.3, 127.7, 127.8, 136.4, 137.2 (Ph, =CH), 153.9 (NC=O), 178.2 (C=O) ppm.

Synthesis of compound 2. The reaction mixture of 1 (1.0 g, 2.88 mmol) and K₂CO₃ (1.59 g, 11.53 mmol) in 10 mL of DMF was added CH₃I (0.49 g, 3.47 mmol) and stirred at room temperature for 1 h. Subsequently, after being quenched by water (50 mL), the resulting mixture was extracted with ethyl acetate (50 mL × 3). The combined organic layers were washed with water (50 mL × 4), dried over anhydrous Na₂SO₄ and concentrated under reduced pressure to afford the yellow oil 2 (0.99 g, 95%). ¹H NMR (600 MHz, CDCl₃, TMS) δ 1.42–1.47 (m, 9H, CH₃), 2.10–2.46 (m, 2H, NCHCH₂), 3.51–3.73 (m, 2H, NCH₂), 4.17–4.17 (m, 1H, NCH), 4.36–4.53 (m, 3H, OCH, OCH₂), 5.24 (d, J = 10.8 Hz, 1H, =CH₂), 5.74 (d, J = 17.6 Hz, 1H, =CH₂), 6.70 (dd, J = 10.9, 17.6 Hz, 1H, =CH), 7.26–7.39 (m, 4H, Ar-H), 8.90 (s, 1H, COOH) ppm; ¹³C NMR (151 MHz, CDCl₃): δ 28.2 (CH₃), 36.7 (NCHCH₂), 51.3 (NCH₂), 57.9 (NCH), 71.0, 75.9, 80.7 (OCH₂, OCH and OC(CH₃)₃), 114.0 (=CH₂), 126.3, 127.7, 127.8, 136.4, 137.2 (Ph, =CH), 153.9 (NC=O), 178.2 (C=O) ppm.

Synthesis of compound 3. A N₂-filled three-necked flask was charged with magnesium chips (1.08 g, 45.0 mmol), bromobenzene (4.71 g, 30.0 mmol) and anhydrous THF (45 mL), refluxed for 3 h and cooled to room temperature. To the obtained reaction mixture was added anhydrous THF solution containing 2 (30 mL, 3.0 g, 8.3 mmol) dropwise at 0 °C and stirred at room temperature for 3 h. After being quenched with aqueous NH₄Cl solution (50 mL), the reaction mixture was extracted with diethyl ether (100 mL × 3). The combined organic phases were dried over anhydrous Na₂SO₄ and concentrated to afford the residue, which was purified by silica gel column chromatography using petroleum ether/ethyl acetate (v/v = 50/1 to 20/1, 10/1 and 1/1) as eluents to afford the white powder 3 (0.75 g, 77%); ¹H NMR (600 MHz, CDCl₃, TMS) δ 1.38 (s, 9H, CH₃), 2.08–2.18 (m, 2H, NCHCH₂), 2.86 (s, 1H, NCH), 3.35–3.52 (m, 2H, NCH₂), 4.28–4.34 (m, 2H, OCH₂), 5.02–5.05 (m, 1H, OCH), 5.22 (d, J = 10.9 Hz, 1H, =CH₂), 5.72 (d, J = 17.6 Hz, 1H, =CH₂), 6.69 (dd, J = 10.9, 17.6 Hz, 1H, =CH), 7.16–7.49 (m, 14H, Ar-H) ppm; ¹³C NMR (151 MHz, CDCl₃): δ 28.2 (CH₃), 36.5 (NCHCH₂), 52.9 (NCH₂), 65.1 (NCH), 70.6, 76.4, 80.7 (OCH₂, OCH, OC(CH₃)₃), 81.6 (COH), 113.8 (=CH₂), 126.2, 127.1, 127.2, 127.5, 127.8, 127.9, 136.6, 137.1, 137.5 (Ph, =CH), 145.7 (NC=O) ppm.

Synthesis of compound 4. The reaction mixture of 3 (1.47 g, 3.04 mmol) and KOH (1.68 g, 0.03 mol) in the mixed MeOH/DMSO (20 mL, v/v = 2/7) was stirred at 60 °C for 6 h, added water (20 mL) and extracted with cyclohexane (20 mL × 5). The combined organic layer was dried over anhydrous Na₂SO₄ and concentrated under reduced pressure to afford white solid 4 (1.08 g, 92%). ¹H NMR (600 MHz, CDCl₃, TMS) δ 1.64–1.68 (m, 1H, NCHCH₂), 1.76–1.83 (m, 1H, NCHCH₂), 3.09 (s, 2H, NCH₂), 4.01–4.02 (m, 1H, NCH), 4.38–4.50 (m, 2H, OCH₂), 4.57 (dd, J = 9.6, 6.7 Hz, 1H, OCH), 5.23 (d, J = 10.9 Hz, 1H, =CH₂), 5.73 (d, J = 17.6 Hz, 1H, =CH₂), 6.70 (dd, J = 10.9, 17.6 Hz, 1H, =CH), 7.13–7.57 (m, 14H, Ar-H) ppm; ¹³C NMR (151 MHz, CDCl₃): δ 33.0 (NCHCH₂), 52.4 (NCH₂), 63.4 (NCH), 70.5, 76.9, 79.5 (OCH₂, OCH and OC(CH₃)₃), 113.8 (=CH₂), 125.4, 126.0, 126.3, 126.4, 126.6, 127.8, 128.0, 128.3, 136.5, 137.1, 137.9, 144.9, 147.3 (Ph, =CH) ppm.

Synthesis of compound 5. Triethylamine (0.19 g, 1.9 mmol) and trimethylsilyl trifluoromethanesulfonate (TMSOTf, 0.42 g, 1.9 mmol) were added in turn at 0 °C to anhydrous CH₂Cl₂ solution containing 4 (15 mL, 0.39 g, 1.0 mmol) under a N₂ atmosphere. After being stirred at room temperature for 3 h, the resulting mixture was quenched with water (20 mL) and extracted with CH₂Cl₂ (20 mL × 3). The combined organic phases were dried over anhydrous Na₂SO₄ and concentrated under reduced pressure. The crude product was purified by silica gel column chromatography using petroleum ether/ethyl acetate (v/v = 20/1 to 10/1) as eluents to afford yellow oil 5 (0.36 g, 78%); ¹H NMR (600 MHz, CDCl₃, TMS) δ-0.19 (s, 9H, CH₃), 1.59–1.66 (m, 2H, NCHCH₂), 1.99 (s, 1H, NH), 2.73 (dd, J = 4.9, 11.6 Hz, 1H, NCH₂), 2.90 (dd, J = 2.1, 11.6 Hz, 1H, NCH₂), 3.69–3.71 (m, 1H, NCH), 4.27–4.38 (m, 3H, OCH, OCH₂), 5.13 (d, J = 11.0 Hz, 1H, =CH₂), 5.64 (d, J = 17.6 Hz, 1H, =CH₂), 6.61 (dd, J = 10.9, 17.6 Hz, 1H, =CH), 7.10–7.39 (m, 14H, Ar-H) ppm; ¹³C NMR (151 MHz, CDCl₃): δ 2.1 (CH₃), 34.3 (NCHCH₂), 52.9 (NCH₂), 63.7 (NCH), 70.5, 79.2, 82.9 (OCH₂, OCH, OC(CH₃)₃), 113.7 (=CH₂), 126.2, 126.8, 126.9, 127.4, 127.5, 127.6, 127.8, 128.4, 136.5, 136.9, 138.1, 145.3, 146.6 (Ph, =CH) ppm.

Preparation of MNPs Fe₃O₄ and Fe₃O₄/PVP

FeCl₃·6H₂O (11.0 g, 40.7 mmol) and FeCl₂·4H₂O (4.0 g, 20.4 mmol) in water (50 mL) were mixed well at room temperature under argon. The reaction mixture was heated to 85 °C and then added aqueous ammonia (23%) dropwise to a pH of 8–9. After being aged at 85 °C for 4 h, the MNPs Fe₃O₄ were separated by magnetic decantation and washed with water to pH = 7. Subsequently, MNPs Fe₃O₄ (4.0 g) were well-dispersed in water (50 mL) and added aqueous polyvinylpyrrolidone (PVP) solution (10.0 mL, 25 g L⁻¹). After being stirred at room temperature for 12 h, the reaction mixture was added acetone (250 mL) and stirred for 10 min. The MNPs Fe₃O₄/PVP were obtained by magnetic decantation, washed with ethanol (50 mL × 2) and dried under vacuum at 55 °C for 24 h. Anal. calcd for Fe₃O₄/PVP found: C, 0.01; H, 0.23; N, 0.05.

Preparation of supported Jørgensen–Hayashi catalysts Fe₃O₄@SiO₂/ProTMS and Fe₃O₄/PVP@SiO₂/ProTMS

A mixture of 3-mercaptotrimethoxysilane (MPTMS, 0.25 mL, 1.14 mmol), AIBN (41.0 mg, 0.25 mmol) and 5 (0.23 g, 0.5 mmol) in 15 mL of CHCl₃ was stirred at 80 °C for 12 h under a N₂ atmosphere and the solvent was removed under reduced pressure. The residues containing MPTMS and ProTMS were added slowly to ethanol solution containing Fe₃O₄ or Fe₃O₄/PVP (25 mL, 125.0 mg) pretreated with aqueous ammonia (0.75 mL, 23%) and tetraethyl orthosilicate (0.13 mL, TEOS) for 2 h. After the resulting mixture was stirred at room temperature for 24 h, the pale yellow MNPs-supported Jørgensen–Hayashi Fe₃O₄@SiO₂/ProTMS (0.46 g) and Fe₃O₄/PVP@SiO₂/ProTMS (0.57 g) catalysts were separated magnetically, washed with CHCl₃ (25 mL × 3), ethanol (25 mL × 3) and dried under vacuum for 4 h. Anal. calcd for Fe₃O₄@SiO₂/ProTMS found: C, 31.51; H, 4.10; N, 1.02; S, 6.15 and Fe₃O₄/PVP@SiO₂/ProTMS found: C, 31.56; H, 3.22; N, 0.90; S, 5.42.

General procedure for the Michael reaction

A reaction mixture of (E)-nitrostyrene (40.0 mg, 0.27 mmol), Fe₃O₄/PVP@SiO₂/ProTMS (30.0 mg, 7 mol%) and CHCl₃ (2.0 mL) was stirred at 0 °C for 10 min, added propanal (94.0 mg, 1.62 mmol) by syringe and monitored by TLC until completion. The catalyst was separated using an external magnetic field, washed with CHCl₃ (2.0 mL × 3) and reused directly for the next catalytic cycle. The combined organic layers were concentrated and purified by silica gel column chromatography using petroleum ether/ethyl acetate (v/v = 5/1) as eluents to afford the Michael adduct. The diastereoselectivities (syn/anti) and enantioselectivities (% ee) of products were determined by ¹H NMR and HPLC on a chiral stationary phase (Daicel Chiralpak OD-H, AD-H, AS-H or OJ-H, 4.6 mm × 25 cm column, see ESI†), respectively.

Results and discussion

Preparation of MNPs-supported organocatalyst

As described in Scheme 2, Jørgensen–Hayashi catalyst 5 could be synthesized using Boc-L-hydroxyproline as starting material with the moderate to good yields in each step (77–95%).¹⁴ The TEM image showed that the spherical MNPs of Fe₃O₄ with the 20–30 nm diameters could be generated at a pH of 8–9 by simple co-precipitation of FeCl₃ and FeCl₂ with molar ratio of 2 in the presence of aqueous ammonia (23%). After being modified for 12 h at room temperature in aqueous PVP solution, TEM image revealed that MNPs Fe₃O₄/PVP had no significant change in the surface morphology and particle diameter. According to the content of nitrogen (0.05%) detected by elemental analysis, the loading of PVP onto the backbone of Fe₃O₄ was calculated to be very low (about 0.4%).

Scheme 2 The synthetic route to the Jørgensen–Hayashi organocatalyst 5.

As described in Scheme 1, the Jørgensen–Hayashi catalyst 5 attached to ProTMS could be immobilized by one-pot to prepare the supported organocatalysts Fe₃O₄@SiO₂/ProTMS and Fe₃O₄/PVP@SiO₂/ProTMS through the radical addition of sulfhydryl (–SH) to C=C double bond in 5 and subsequent surface-modification taking advantage of the hydrolysis of Si(OCH₃)₃ on the surface of MNPs Fe₃O₄ and Fe₃O₄/PVP. Based to the nitrogen contents by elemental analysis (1.02% and 0.90%), the loading capacities of Jørgensen–Hayashi catalyst 5 in Fe₃O₄@SiO₂/ProTMS and Fe₃O₄/PVP@SiO₂/ProTMS were calculated to be 0.73 mmol g⁻¹ and 0.64 mmol g⁻¹, respectively. Moreover, the loadings of MPTMS onto Fe₃O₄@SiO₂/ProTMS and Fe₃O₄/PVP@SiO₂/ProTMS were also determined by elemental analysis of sulphur (6.15% and 5.42%) to be 1.92 mmol g⁻¹ and 1.69 mmol g⁻¹. Owing to the partial radical addition of sulfhydryl (–SH) to C=C double bond in 5, the free MPTMS were found to be 1.19 mmol g⁻¹ and 1.05 mmol g⁻¹. Based on the loaded MPTMS and ProTMS, it was found that Fe₃O₄@SiO₂/ProTMS and Fe₃O₄/PVP@SiO₂/ProTMS possessed the same molar ratio (1.6/1) of free MPTMS to ProTMS. Furthermore, the superparamagnetic behavior of Fe₃O₄@SiO₂/ProTMS (21.5 emu g⁻¹) and Fe₃O₄/PVP@SiO₂/ProTMS (10.4 emu g⁻¹) at 298 K was evidenced by the zero coercivity and resonance of each magnetization loop in the magnetization curves (Fig. 1).

Fig. 1 Magnetization curves for Fe₃O₄/PVP (a), Fe₃O₄@SiO₂/ProTMS (b), 10^th-recycled Fe₃O₄/PVP@SiO₂/ProTMS (c) and Fe₃O₄/PVP@SiO₂/ProTMS (d).

The role of polyvinylpyrrolidone (PVP)

Poly(vinylpyrrolidone) (PVP, K29–32), amphiphilic and nonionic polymer, was widely used in science and technology, and could be adsorbed onto a broad range of different materials as stabilizing agent.¹⁵ Herein, we endeavoured to elucidate where PVP was adsorbed onto the backbone of MNPs Fe₃O₄ and what changes in chemical and physical properties were resulted from PVP.

The covalent attachment of Jørgensen–Hayashi catalyst 5 to the backbone of Fe₃O₄ MNPs is clearly corroborated by FT-IR spectroscopy (Fig. 2), which confirmed by (1) the asymmetrical, symmetrical stretching vibration and flexural vibration of Si–O–Si at 1067, 807 and 459 cm⁻¹,¹⁶ (2) the stretching vibrations of C–S and Si–C bonds at 1249 and 702 cm⁻¹,¹⁷ (3) the various C–H stretching vibrations and characteristic phenyl ring at 3056, 3025, 2930 cm⁻¹ and in the 1632–1448 cm⁻¹ range, respectively. After being modified by PVP (0.4%), there was no significant difference in IR spectra between Fe₃O₄/PVP and bare Fe₃O₄. Furthermore, all Fe₃O₄/PVP@SiO₂/ProTMS and Fe₃O₄@SiO₂/ProTMS had a strong stretching vibration of C–O bond at 1134 cm⁻¹, which implied that a large amount of –SiOCH₃ moieties existed on the surface of Fe₃O₄/PVP@SiO₂/ProTMS and Fe₃O₄@SiO₂/ProTMS.

Fig. 2 IR spectra of Fe₃O₄ MNPs (a), Fe₃O₄/PVP MNPs (b), Jørgensen–Hayashi catalyst 5 (c), Fe₃O₄/PVP@SiO₂/ProTMS (d) and Fe₃O₄@SiO₂/ProTMS (e).

The effect of PVP on the hydrolysis degree of –Si(OCH₃)₃ attached to ProTMS and MPTMS onto MNPs Fe₃O₄/PVP and Fe₃O₄ could be further monitored by TGA. From the TG and DTG curves (Fig. 3), the thermal decomposition of organic moieties in Fe₃O₄/PVP@SiO₂/ProTMS occurred in two distinct steps: the first one in the temperature 200–500 °C range and the second one in the 500–600 °C range, accompanied by a continuous endothermic peak (see ESI†). It is noteworthy that Fe₃O₄/PVP@SiO₂/ProTMS exhibited the 15.2% higher weight loss than Fe₃O₄@SiO₂/ProTMS in the temperature range of 200–600 °C, although Fe₃O₄/PVP@SiO₂/ProTMS contained a lower amount of Jørgensen–Hayashi catalyst 5. Another phenomenon was that a sharp weight loss of Fe₃O₄/PVP@SiO₂/ProTMS (30.01%) was newly emerged in the 500–600 °C range (Fig. 3c). Based on the two points mentioned above, it was demonstrated that the hydrolysis efficiency of –Si(OCH₃)₃ decreased upon the attachment of ProTMS and MPTMS onto Fe₃O₄/PVP, and resulted in the increased SiOCH₃ moiety due to the package of PVP, which was in accord with the stronger C–O stretching vibration at 1134 cm⁻¹ in IR spectra of Fe₃O₄/PVP@SiO₂/ProTMS and verified by the weakened C–O stretching vibration of 10^th-recycled Fe₃O₄/PVP@SiO₂/ProTMS. Furthermore, besides the same typical peaks of Fe₃O₄ at 30.3°, 35.6°, 43.3°, 53.8°, 57.4° and 63.0° in the XRD pattern, Fe₃O₄@SiO₂/ProTMS and Fe₃O₄/PVP@SiO₂/ProTMS also displayed the difference in the typical peak of SiO₂ in the 16–25° range (Fig. 4a and b), which was closely evidenced to the hydrolysis degree of –Si(OCH₃)₃.

Fig. 3 Thermogravimetric curves of Fe₃O₄@SiO₂/ProTMS (a), 10^th-recycled Fe₃O₄/PVP@SiO₂/ProTMS (b) and Fe₃O₄/PVP@SiO₂/ProTMS (c).

Fig. 4 Powder X-ray diffraction patterns of Fe₃O₄/PVP@SiO₂/ProTMS (a), Fe₃O₄@SiO₂/ProTMS (b), 10^th-recycled Fe₃O₄/PVP@SiO₂/ProTMS (c) and Fe₃O₄ MNPs (d).

The presence of PVP not only altered the chemical composition but also affected the morphology of MNPs-supported catalyst. The grafting of Jørgensen–Hayashi catalyst 5 onto MNPs Fe₃O₄ and Fe₃O₄/PVP led to a serious agglomeration due to the coated SiO₂ generated from the hydrolysis of –Si(OCH₃)₃ moiety. However, seen from the SEM images, Fe₃O₄/PVP@SiO₂/ProTMS exhibited the smaller even-textured and granular particles with the diameters ranging from 0.2 to 0.3 μm (Fig. 5f), compared with irregular Fe₃O₄@SiO₂/ProTMS (Fig. 5b). The further investigation by the TEM image showed the densely agminated Fe₃O₄@SiO₂/ProTMS (Fig. 5d) and nano-structured Fe₃O₄/PVP@SiO₂/ProTMS (Fig. 5h) in an accumulative state upon being coated with SiO₂ shell, respectively due to the profound and weakened hydrolysis of –Si(OCH₃)₃ in the absence or presence of PVP.

Fig. 5 The surface morphologies for TEM of MNPs Fe₃O₄ (a), SEM of Fe₃O₄@SiO₂/ProTMS (b), TEM of Fe₃O₄@SiO₂/ProTMS (c and d), TEM of MNPs Fe₃O₄/PVP (e), SEM of Fe₃O₄/PVP@SiO₂/ProTMS (f), TEM of Fe₃O₄/PVP@SiO₂/ProTMS (g and h).

The nitrogen adsorption–desorption isotherm plots of all samples, obtained at 77 K, were shown in ESI.† After being stirred at room temperature polyvinylpyrrolidone (PVP) solution (10.0 mL, 25 g L⁻¹) for 12 h, it was found that MNPs Fe₃O₄/PVP possessed the higher BET-specific surface area (65.8 m² g⁻¹), average pore diameter (11.7 nm) and pore volume (0.39 cm³ g⁻¹) than bare Fe₃O₄ (7.9 m² g⁻¹, 5.2 nm and 0.021 cm³ g⁻¹). Especially, Fe₃O₄/PVP had five characteristic cavities centered at 0.9, 1.4, 2.1, 3.4 and 6.6 nm, accompanied by some additional pores in the different sizes with the nitrogen desorption (Dv) (<1.0 × 10⁻³ mL Å⁻¹ g⁻¹) (Fig. 6c), whereas bare MNPs Fe₃O₄ displayed a broad pore size distribution (Fig. 6a). Interestingly, except for these five characteristic cavities, all the other pores in Fe₃O₄/PVP disappeared upon the immobilization of ProTMS and MPTMS through the hydrolysis of –Si(OCH₃)₃, which indicated that Jørgensen–Hayashi catalyst 5 was entrapped inside these disappeared pores. To our delight, the remained four characteristic pores of Fe₃O₄/PVP@SiO₂/ProTMS with the high Dv values (>1.0 × 10⁻³ mL Å⁻¹ g⁻¹) provided the crucial channels for reactants to access catalytic sites in catalytic process. Unfortunately, Fe₃O₄@SiO₂/ProTMS exhibited the non-uniform pore size distribution in the 1–10 nm range without the modification of PVP.

Fig. 6 Pore size distributions (PSDs) of Fe₃O₄ (a), Fe₃O₄@SiO₂/ProTMS (b), Fe₃O₄/PVP (c) and Fe₃O₄/PVP@SiO₂/ProTMS (d).

Asymmetric Michael addition

To evaluate influence of PVP on the catalytic performance and reusability of Fe₃O₄@SiO₂/ProTMS and Fe₃O₄/PVP@SiO₂/ProTMS, the asymmetric Michael addition of propanal to trans-β-nitrostyrene was selected as a model reaction.

In the Michael addition of propanal to trans-β-nitrostyrene, the corresponding syn-adducts, whose absolute stereochemistry was determined to be 2R,3S-configurations by comparing its optical rotation and HPLC spectra with literature values,¹⁸ were obtained as major stereoisomers in the high yields and excellent enantioselectivities (Table 1). To the best of our knowledge, it was found that Fe₃O₄/PVP@SiO₂/ProTMS appeared as the most efficient MNPs-supported Jørgensen–Hayashi catalyst to date in terms of catalytic activity, stereoselectivity and reusability among the ever-reported Michael additions.^11–13 Meanwhile, Fe₃O₄/PVP@SiO₂/ProTMS (99%, syn/anti = 94/6, 98% ee) presented the superior catalytic performances to Fe₃O₄@SiO₂/ProTMS (89%, syn/anti = 93/7 and 97% ee) (entry 18), which might be rationalized in terms of catalytic accessibility and steric confinement owing the above-mentioned structural texture and morphology resulted from the vital function of PVP. In the following catalytic reactions, Fe₃O₄/PVP@SiO₂/ProTMS was chosen for the further optimization of experimental conditions.

Table 1 Screening of reaction conditions of Fe₃O₄/PVP@SiO₂/ProTMS for the Michael addition of propanal to trans-β-nitrostyrene^a

Entry	Catalyst (mg)/catalyst (mol%)	Additive	Solvent	T (°C)/time (h)	Yield^b (%)	Syn/anti^c	ee^d (%)
a Reaction conditions: trans-β-nitrostyrene (40 mg, 0.27 mmol), propanal (94 mg, 1.62 mmol), additive (10 mol%), solvent (2.0 mL).b Isolated yield.c Determined by ^1H NMR.d Determined by chiral HPLC.e 20 mol% of additive.f Fe3O4@SiO2/ProTMS used.
1	30/7	—	CHCl3	0/17	99	89/11	97
2	30/7	TFA	CHCl3	0/8	Trace	n.d.	n.d.
3	30/7	PhCO2H	CHCl3	0/12	93	92/8	97
4	30/7	2-FPhCO2H	CHCl3	0/12	95	90/10	97
5	30/7	3-NO2PhCO2H	CHCl3	0/8	99	96/4	98
6	30/7	3-NO2PhCO2H^e	CHCl3	0/24	93	93/7	97
7	30/7	3-NO2PhCO2H	C6H5CH3	0/8	87	88/12	95
8	30/7	3-NO2PhCO2H	CH2Cl2	0/8	94	90/10	96
9	30/7	3-NO2PhCO2H	EtOH	0/8	Trace	n.d.	n.d.
10	60/14	—	CHCl3	0/8	98	91/9	97
11	60/14	3-NO2PhCO2H	CHCl3	0/4	98	91/9	98
12	15/3.5	—	CHCl3	0/32	76	92/8	97
13	15/3.5	3-NO2PhCO2H	CHCl3	0/22	96	89/11	98
14	30/7	3-NO2PhCO2H	CHCl3	20/5	99	87/13	96
15	30/7	3-NO2PhCO2H	CHCl3	10/7	99	90/10	97
16	30/7	3-NO2PhCO2H	CHCl3	−10/20	86	96/4	98
17	30/7	3-NO2PhCO2H	CHCl3	−20/32	82	98/2	98
18^f	26/7	3-NO2PhCO2H	CHCl3	0/8	89	93/7	97

---

Firstly, it was found that the acid strengths of Brønsted acids such as TFA (pK_a = 0.23), 3-NO₂C₆H₄CO₂H (pK_a = 2.45), 2-FC₆H₄CO₂H (pK_a = 3.27) and C₆H₄CO₂H (pK_a = 4.21) had a great effect on catalytic performances (entries 2–6). Disappointedly, TFA with the stronger acid strength only afforded a trace of Michael adduct (entry 2). Fortunately, the other weak Brønsted acids could effectively improve catalytic rate and stereoselectivity. Among them, 3-NO₂C₆H₄CO₂H (10 mol%) was proved to be the best effective co-catalyst with the excellent performances (99%, syn/anti = 96/4, 98% ee, entry 5). It was worthwhile to note that Fe₃O₄/PVP@SiO₂/ProTMS gave the less catalytic performances (93%, syn/anti = 93/7, 97% ee, entry 6) after 24 h upon doubling the dosage of 3-NO₂C₆H₅CO₂H. Next, the effects of temperature, solvent and used amount of catalyst on catalytic performance were further screened out in detail. Our attempts to further reduce catalyst loading led to the decreased yield and stereoselectivity (entries 12 and 13), and ethanol as a solvent resulted in no activity (entry 9). Furthermore, although the decrease in temperature to −20 °C led to the product with the higher stereoselectivity (syn/anti = 98/2, 98% ee, entry 17), it took place at a considerably lower rate, requiring up to 32 h for the achievement of 82% yield.

The optimized protocol was expanded to a wide variety of trans-β-nitrostyrenes bearing electron-donating (–CH₃, OCH₃) and electron-withdrawing (–X) substituents at the o, m, p-positions. From Table 2, all trans-β-nitrostyrenes bearing electron-donating and electron-withdrawing substituents afforded the excellent catalytic performances at the same level (93–99%, syn/anti = 86–92 : 14–8, 96–98% ee, entries 1–9). It was worthwhile to note that the nitrostyrenes with o, p-OCH₃ and o-CH₃ electron-donating (entries 5, 7 and 9) and o-Cl electron-with-drawing substituent (entry 4) required respective 48 h and 24 h to achieve the 93–99% yields resulted from the electron-donating effect and vicinal hindrance of the substituents. A nitroalkene bearing a heterocyclic 2-furyl substituent also afforded a high 98% yield and 95% ee but gave a moderate diastereoselectivity (syn/anti = 81/19) (entry 10). Additionally, the Michael addition of propanal to aliphatic nitroalkene gave the good catalytic performances (97%, 97% ee, 86 : 14 dr) but needed a prolonged reaction period (34 h, entry 11).

Table 2 Asymmetric Michael addition of propanal to various nitroalkenes catalyzed by Fe₃O₄/PVP@SiO₂/ProTMS^a

Entry	t (h)	Product	Yield^b (%)	Syn/anti^c	% ee^d
a Reaction conditions: nitrolefin (40 mg, 0.27 mmol), propanal (94 mg, 1.62 mmol), 3-NO2C6H4CO2H (10 mol%), CHCl3 (2.0 mL), 0 °C.b Isolated yield.c Determined by ^1H NMR.d Determined by chiral HPLC.
1	8		98	88/12	97
2	7		99	87/13	97
3	9		95	86/14	96
4	24		98	88/12	96
5	48		99	92/8	97
6	12		97	89/11	96
7	48		93	92/8	96
8	12		98	89/11	96
9	48		95	90/10	97
10	11		98	81/19	95
11	34		97	86/14	97
12	8		>99	96/4	98

---

Reusability of MNPs-supported catalyst

One of the main advantages associated with the supporting of a catalyst onto MNPs is the facile separation of a catalyst from reaction mixture with an external magnet. When the reaction was completed, an external magnet was put nearby the reaction flask to concentrate Fe₃O₄/PVP@SiO₂/ProTMS, and the solution could be easily decanted. The recovered catalyst was directly reused for the next run after being washed with CHCl₃ (2.0 mL × 3).

Fig. 7 showed the comparative reusabilities of Fe₃O₄/PVP@SiO₂/ProTMS and Fe₃O₄@SiO₂/ProTMS including the yields and stereoselectivities. From Fig. 6, Fe₃O₄/PVP@SiO₂/ProTMS displayed a better reusability than Fe₃O₄@SiO₂/ProTMS. Unfortunately, the decreasing yields and diastereoselectivity catalyzed by Fe₃O₄@SiO₂/ProTMS after the second run like the ever-reported Jørgensen–Hayashi catalysts using MNPs as a support.^11–13 On the contrary, the yields catalyzed by Fe₃O₄/PVP@SiO₂/ProTMS remained at the very high level (>94%) in the first five cycles. Even in 10^th run, the 77% yield could be achieved after prolonging the reaction time to 48 h. More excitedly, the constant excellent stereoselectivity (98% ee and 96/4 dr) could be maintained in the ten consecutive cycles. To the best of our knowledge, Fe₃O₄/PVP@SiO₂/ProTMS exhibited the most effective reusability among the reported MNPs-supported Jørgensen–Hayashi catalysts.

Fig. 7 The comparative reusability of Fe₃O₄@SiO₂/ProTMS and Fe₃O₄/PVP@SiO₂/ProTMS under the optimized reaction conditions.

To understand the reason for the decreased yield, the structural change of the 10^th-recycled Fe₃O₄/PVP@SiO₂/ProTMS was monitored by TEM, TGA, elemental analysis and N₂ adsorption–desorption isotherm. Above all, the increased carbon, hydrogen and nitrogen contents, especially the doubled nitrogen content (0.90% → 1.76%) stemmed from nitrogenous compounds such as nitroalkene, elucidated that the reactants and products were adsorbed on the porous backbone of the catalyst, which also verified by the decreased average pore size and pore volume from 15.7 nm and 0.031 cm³ g⁻¹ to 2.5 nm and 0.021 cm³ g⁻¹ respectively. Moreover, the structural change of SiO₂ attached on the core of Fe₃O₄/PVP was elucidated by XRD and TGA, in which a wide and round XRD peak of SiO₂ in the range of 16–25° (Fig. 3b) and a disappeared weight loss in the 500–600 °C range demonstrated the further hydrolysis of –SiOCH₃. Meanwhile, it was found that the 10^th-recycled Fe₃O₄/PVP@SiO₂/ProTMS maintained its original magnetic property (Fig. 1b) and no agglomeration was observed by TEM (Fig. S10, ESI†). In conclusion, it was speculated that the adsorbed impurities and further hydrolysis of –SiOCH₃ resulted in the structural changes such as pore volume and pore size distribution and was ultimately responsible for the decreased catalytic activity of Fe₃O₄/PVP@SiO₂/ProTMS.

Conclusions

In summary, we had immobilized the homogeneous Jørgensen–Hayashi catalyst of (S)-diphenylprolinoltrimethylsilyl ether onto the backbone of Fe₃O₄/PVP via facile one-pot surface-modification, and successfully applied the MNPs Fe₃O₄-supported catalyst Fe₃O₄/PVP@SiO₂/ProTMS in the asymmetric Michael addition of propanal to various nitroalkenes with excellent catalytic performances. Due to the adsorbed poly(vinylpyrrolidone) (PVP), Fe₃O₄/PVP@SiO₂/ProTMS had a great change in chemical composition, surface morphology and pore structure related to the hydrolysis degree of –Si(OCH₃)₃, which provided a suitable micro-environment to achieve the good reusability with the high yields and unchangeable excellent stereoselectivities including syn/anti and % ee for the first time.

Acknowledgements

We are grateful to the Fundamental Research Funds for the Central Universities (XDJK2015D028) and the National Science Foundation of China (621362005).

Notes and references

1. (a) M. Marigo, T. C. Wabnitz, D. Fielenbach and K. A. Jørgensen, Angew. Chem., Int. Ed., 2005, 44, 794–797; (b) J. Franzén, M. Marigo, D. Fielenbach, T. C. Wabnitz, A. Kjæsgaard and K. A. Jørgensen, J. Am. Chem. Soc., 2005, 127, 18296–18304; (c) Y. Hayashi, H. Gotoh, T. Hayashi and M. Shoji, Angew. Chem., Int. Ed., 2005, 44, 4212–4215.
2. S. Mukherjee, J. W. Yang, S. Hoffmann and B. List, Chem. Rev., 2007, 107, 5471–5569.
3. A. Erkkilä, I. Majander and P. M. Pihko, Chem. Rev., 2007, 107, 5416–5470.
4. S. Meninno and A. Lattanzi, Chem. Commun., 2013, 49, 3821–3832.
5. (a) C. M. R. Volla, I. Atodiresei and M. Rueping, Chem. Rev., 2014, 114, 2390–2431; (b) A. Z. Halimehjani, I. N. N. Namboothiri and S. E. Hooshmand, RSC Adv., 2014, 4, 51794–51829; (c) H. Pellissier, Tetrahedron, 2013, 69, 7171–7210.
6. (a) X. J. Li, B. L. Yang, X. B. Jia, M. Q. Chen and Z. G. Hu, RSC Adv., 2015, 5, 89149–89156; (b) S. Wang, P. W. Liu, W. J. Wang, Z. G. Zhang and B. G. Li, Catal. Sci. Technol., 2015, 7, 3798–3805; (c) C. Schafer, S. C. Mhadgut, N. Kugyela, M. Torok and B. Torok, Catal. Sci. Technol., 2015, 7, 716–723; (d) A. Lu, T. P. Smart, T. H. Epps, D. A. Longbottom and R. K. O'Reilly, Macromolecules, 2011, 44, 7233–7241; (e) K. Akagawa, S. Sakamoto and K. Kudo, Synlett, 2011, 6, 817–820; (f) E. Alza and M. A. Pericàs, Adv. Synth. Catal., 2009, 351, 3051–3056.
7. (a) A. A. Elmekawy, J. B. Sweeney and D. R. Brown, Catal. Sci. Technol., 2015, 7, 716–723; (b) C. L. Wu, X. Q. Long, X. K. Fu, G. W. Wang and Z. A. Mirza, RSC Adv., 2015, 5, 3168–3176; (c) M. Angeloni, O. Piermatti, F. Pizzo and L. Vaccaro, Eur. J. Org. Chem., 2014, 8, 1716–1726; (d) S. Calogero, D. Lanari, M. Orru, O. Piermatti, F. Pizzo and L. Vaccaro, J. Catal., 2011, 282, 112–119; (e) W. Zheng, C. F. Lu, G. C. Yang, Z. X. Chen and J. Q. Nie, Catal. Commun., 2015, 62, 34–38.
8. (a) Y. Kong, R. Tan, L. L. Zhao and D. H. Yin, Green Chem., 2013, 15, 2422–2433; (b) X. Zhang, W. S. Zhao, C. K. Qu, L. L. Yang and Y. C. Cui, Tetrahedron: Asymmetry, 2012, 23, 468–473.
9. R. Tan, C. Y. Li, J. Q. Luo, Y. Kong, W. G. Zheng and D. H. Yin, J. Catal., 2013, 298, 138–147.
10. (a) R. B. N. Baig, M. N. Nadagouda and R. S. Varma, Coord. Chem. Rev., 2015, 287, 137–156; (b) R. Dalpozzo, Green Chem., 2015, 17, 3671–3686; (c) J. W. Wan, L. ding, T. Wu, X. B. Ma and Q. Tang, RSC Adv., 2014, 4, 38323–38333.
11. B. G. Wang, B. C. Ma, Q. Wang and W. Wang, Adv. Synth. Catal., 2010, 352, 2923–2928.
12. P. Riente, C. Mendoza and M. A. Pericás, J. Mater. Chem., 2011, 21, 7350–7355.
13. M. Keller, A. Perrier, R. Linhardt, L. Travers, S. Wittmann, A. M. Caminade, J. P. Majoral, O. Reiser and A. Oualia, Adv. Synth. Catal., 2013, 355, 1748–1754.
14. A. B. Xia, C. Zhang, Y. P. Zhang, Y. J. Guo, X. L. Zhang, Z. B. Li and D. Q. Xu, Org. Biomol. Chem., 2015, 13, 9593–9599.
15. (a) T. J. Yoon, J. S. Kim, B. G. Kim, K. N. Yu, M. H. Cho and J. K. Lee, Angew. Chem., Int. Ed., 2005, 44, 1068–1071; (b) M. Shokouhimehr, Y. Piao, J. Kim, Y. Jang and T. Hyeon, Angew. Chem., Int. Ed., 2007, 46, 7039–7043; (c) X. X. Zheng, S. Z. Luo, L. Zhang and J. P. Cheng, Green Chem., 2009, 11, 455–458.
16. (a) A. Khalafi-Nezhad, E. S. Shahidzadeh, S. Sarikhani and F. Panahi, J. Mol. Catal. A: Chem., 2013, 379, 1–8; (b) A. Khalafi-Nezhad and F. Panahi, Green Chem., 2011, 13, 2408–2415; (c) D. D. Feng, J. H. Xu, J. W. Wan, B. Xie and X. B. Ma, Catal. Sci. Technol., 2015, 5, 2141–2148.
17. L. L. Lou, H. L. Du, Y. B. Shen, K. Yu, W. J. Yu, Q. Chen and S. X. Liu, Microporous Mesoporous Mater., 2014, 187, 94–99; Y. Du, D. D. Feng, J. W. Wan and X. B. Ma, Appl. Catal., A, 2014, 479, 49–58; W. Wang, X. B. Ma, J. W. Wan, J. Cao and Q. Tang, Dalton Trans., 2012, 41, 5715–5726.
18. (a) Y. Hayashi, H. Gotoh, T. Hayashi and M. Shoji, Angew. Chem., Int. Ed., 2005, 44, 4212; (b) Z. L. Zheng, B. L. Perkins and B. Ni, J. Am. Chem. Soc., 2010, 132, 50–51.

---

Footnote

† Electronic supplementary information (ESI) available: TGA, TEM and N₂ adsorption–desorption isotherm of catalysts, ¹H, ¹³C NMR and HPLC spectra of intermediates and products. See DOI: 10.1039/c6ra01051b

---