Ctpc

JCC

a

ARRAA

KCZpHA

1

traiait(salJpmoi

0d

Applied Catalysis A: General 407 (2011) 163– 172

Contents lists available at SciVerse ScienceDirect

Applied Catalysis A: General

j ourna l ho me page: www.elsev ier .com/ locate /apcata

atalytic asymmetric epoxidation of unfunctionalized olefins using a series novelype of layered crystalline organic polymer–inorganic hybrid zinchosphonate–phosphate immobilized aryldiamine modified chiral salen Mn(III)omplex

ing Huang ∗, Xiangkai Fu, Qiang Miaoollege of Chemistry and Chemical Engineering Southwest University, Research Institute of Applied Chemistry Southwest University, The Key Laboratory of Applied Chemistry ofhongqing Municipality, The Key Laboratory of Eco-environments in Three Gorges Reservoir Region Ministry of Education, Chongqing 400715, China

r t i c l e i n f o

rticle history:eceived 29 May 2011eceived in revised form 19 August 2011ccepted 21 August 2011vailable online 26 August 2011

eywords:

a b s t r a c t

In this report we demonstrate the suitability of layered crystalline organic polymer–inorganic hybridmaterial ZnPS–PVPA with different content of the organic group and different inorganic phosphate ascatalyst supports for asymmetric catalysis. A series of the chiral salen Mn(III) complex immobilized ontoZnPS–PVPA modified by aryldiamine were synthesized and characterized by FT-IR, diffusion reflectionUV–vis, AAS, N2 volumetric adsorption, SEM, TEM, XRD and TG. The supported catalysts displayed superiorcatalytic activities in the asymmetric epoxidation of �-methylstyrene and indene with m-CPBA and NaIO4

hiral Mn(III) saleninc poly(styrene-phenylvinylhosphonate)-phosphateeterogeneous catalystsymmetric epoxidation

as oxidants, compared with the corresponding homogeneous catalyst (ee, >99% vs 54% and >99% vs 65%).And the heterogeneous catalysts are relatively stable and can be recycled nine times in the asymmetricepoxidation of �-methylstyrene. These results revealed that the special structure of ZnPS–PVPA playedvital impacts on the conversion and enantioselectivity. Furthermore, this novel type of catalyst can alsobe validly used in large-scale reactions with superior catalytic disposition being maintained at the samelevel, which possessed the potentiality for application in industry.

. Introduction

Owing to the great importance of the chiral compounds inhe manufacture of drugs, vitamins, fragrances, and optical mate-ial, the preparation of chiral building blocks has attracted specialttention. Epoxides of unfunctionalized olefins are very importantntermediates in the manufacture of drugs, vitamins, fragrances,nd optical materials, because they can be readily transferrednto various compounds via regioselective ring opening or func-ional transfer reactions [1–5]. Just as this, asymmetric epoxidationAE) has become a useful preparative method in organic synthe-is. Chiral Mn (salen) complexes developed by Jacobsen, Katsukind coworkers are effective catalysts for AE of olefins with highevels of enantioselectivity [6–8]. However, a major limitation ofacobsen’s catalyst is that it cannot be recycled due to the decom-osition of the catalyst and the formation of inactive dimeric �-oxo

anganese(IV) species [9] in homogeneous phase. The anchoring

f the Mn(III) salen complexes onto supports has been found toncrease the catalyst stability since the generation of the dimeric

∗ Corresponding author. Tel.: +86 2368253704; fax: +86 2368254000.E-mail address: hj41012@163.com (J. Huang).

926-860X/$ – see front matter © 2011 Elsevier B.V. All rights reserved.oi:10.1016/j.apcata.2011.08.035

© 2011 Elsevier B.V. All rights reserved.

�-oxo manganese(IV) species can be avoided. Many efforts havebeen focused on the covalent binding of manganese(III) salen com-plexes to organic polymer [10] and inorganic supports such asMCM-41 [11], zeolites [12] and clays [13]. The trend to developreusable salen–metal complexes with high efficiency and catalyticstability is increasing from the environmental concerns togetherwith economic considerations. Accordingly, methodologies for theheterogenization of homogeneous salen–metal complexes haveemerged. Among them, heterogenization of salen–metal com-plexes into/onto inorganic or inorganic–organic hybrid supportsis one of the promising strategies. For these supported salen–metalcomplexes, the host–guest interaction can be mechanical, physical,or chemical [14,15].

In the last decades, our groups have reported a series oforganic–inorganic hybrid zirconium phosphonate–phosphatesZr(HPO4)2−x(O3P–G)x·nH2O (x = 0–2, G is organic groups)as various kinds of catalysts or catalyst supports,such as solid acid catalysts zirconium sulfophenylphosphonate–phosphate Zr (HPO4)2−x(O3PC6H4SO3H)x·nH2O,

and zirconium [N, N di (phosphono-methyl)iminodiaceticacid] Zr[(O3PCH2)2N(CH2COOH)]·nH2O and their palladiumcomplexes as catalysts in hydrogenation [16–19]. Apart fromthese facts, we have also focused on the immobilization of

1 sis A:

hpopIhcu

ospshesecfdpmciwrspvth

2

2

emaasspo

Rst3

tcaNapusldsbtom

sodium carbonate saturated solution was added to neutralize themixture, and the solvent was evaporated under reduced pres-sure, filtered, washed with deionized water and dried in vacuo

64 J. Huang et al. / Applied Cataly

omogeneous chiral salen Mn(III) complexes, such as the sup-orted chiral salen Mn(III) catalysts on modified zirconiumligostyrenylphosphonate–phosphate (ZSPP) and zirconiumoly(styrene-phenylvinylphosphonate)-phosphate (ZPS–PVPA).

t was noteworthy that the heterogeneous catalysts showedigher enantioselectivity than that of the homogeneous chiralatalyst and superior reusability in asymmetric epoxidation ofnfunctionalized olefins under the same conditions [20–23].

Whereas, few investigations were explored in therganic–inorganic hybrid zinc phosphonate–phosphate as catalystupports, even less in organic polymer–inorganic hybrid zinchosphonate–phosphate used for the immobilization of chiralalen Mn(III). Moreover, some of the heterogeneous catalystsave proved to be successful in the epoxidation of olefins, whichncouraged us to pursue a new functional group to bind chiralalen Mn(III) onto ZnPS–PVPA to produce a recyclable asymmetricpoxidation catalyst. Herein, it is still of academic interest andommercial importance to develop efficient and reusable catalystor the epoxidation of non-functionalized olefins. In view of this, weocumented that a series of new type of layered crystalline organicolymer–inorganic hybrid materials ZnPS–PVPA and aryldiamineodified ZnPS–PVPA applied to immobilize the chiral salen Mn(III)

omplexes through axial coordination. And we also reported themmobilized catalysts were stable, recoverable, reusable catalysts

ith superior enantioselectivity, and can be used in large-scaleeactions with the enantioselectivity being maintained at theame level. In addition, the question as to whether or not variousroportions of organic phosphonate to inorganic phosphate,arious inorganic phosphate and different linkages contributed tohe catalytic activities and enantioselectivities were also examinedere.

. Experiment

.1. Materials and instruments

(1R,2R)-(−)-1, 2-diaminocyclohexane, chloromethyl methylther (toxic compound), �-methylstyrene, n-nonane, N-ethylmorpholine N-oxide (NMO) and m-chloroperbenzoic

cid (m-CPBA) were supplied by Alfa Aesar. Other commerciallyvailable chemicals were laboratory-grade reagents from localuppliers. Chiral salen ligand and chiral homogeneous catalystalen Mn(III) were synthesized according to the standard literaturerocedures [24], and further identified by analysis and comparisonf IR spectra with literature [25].

FT-IR spectra were recorded from KBr pellets using a BrukerFS100/S spectrophotometer (USA) and diffuse reflectance UV–vispectra of the solid samples were recorded in the spectrophotome-er with an integrating sphere using BaSO4 as standard. 1H NMR and1P NMR were performed on AV-300 NMR instrument at ambientemperature at 300 and 121 MHz, respectively. All of the chemi-al shifts were reported downfield in ppm relative to the hydrogennd phosphorus resonance of TMS and 85% H3PO4, respectively.umber- and weight-average molecular weights (Mn and Mw)nd polydispersity (Mw/Mn) were estimated by Waters1515 gelermeation chromatograph (GPC; against polystyrene standards)sing THF as an eluent (1.0 mL min−1) at 35 ◦C. X-ray photoelectronpectrum was recorded on ESCALab250 instrument. The inter-ayer spacings were obtained on DX-1000 automated X-ray poweriffractometer, using Cu K� radiation and internal silicon powdertandard with all samples. The patterns were generally measured

etween 3.00◦ and 80.00◦ with a step size of 0.02◦ min−1 and X-rayube settings of 36 kV and 20 mA. C, H and N elemental analysis wasbtained from an EATM 1112 automatic elemental analyzer instru-ent (Thermo, USA). TG analyses were performed on a SBTQ600

General 407 (2011) 163– 172

thermal analyzer (USA) with the heating rate of 20 ◦C min−1 from25 to 1000 ◦C under flowing N2 (100 mL min−1). The Mn contentsof the catalysts were determined by a TAS-986G (Pgeneral, China)atomic absorption spectroscopy. SEM were performed on KYKY-EM3200 (KYKY, China) micrograph. TEM were obtained on a TECNAI10(PHILIPS, Holland) apparatus. Nitrogen adsorption isotherms weremeasured at 77 K on a 3H-2000I (Huihaihong, China) volumet-ric adsorption analyzer with BET method. The racemic epoxideswere prepared by epoxidation of the corresponding olefins by 3-chloroperbenzoic acid in CH2Cl2 and confirmed by NMR (BrukerAV-300), and the gas chromatography (GC) was calibrated withthe samples of n-nonane, olefins and corresponding racemic epox-ides. The conversions (with n-nonane as internal standard) andthe ee values were analyzed by gas chromatography (GC) with aShimadzu GC2010 (Japan) instrument equipped using a chiral col-umn (HP19 091G-B213, 30 m × 30 m × 0.32 mm × 0.25 �m) and FIDdetector, injector 230 ◦C, detector 230 ◦C. Ultrapure N2 was the car-rier gas (rate 34 mL min−1) with carrier pressure 39.1 kPa and theinjection pore temperature was set at 230 ◦C. The column temper-ature for indene, �-methylstyrene was programmed in the rangeof 80–180 ◦C. The oven temperature program was initially startedat 80 ◦C, held for 3 min; then raised to 150 ◦C at 7 ◦C min−1 andheld for 5 min at 150 ◦C; raised to 220 ◦C at 7 ◦C min−1, and wasfinally set at 220 ◦C constant for 3 min. The total run time of theGC program for �-methylstyrene was 30 min. The appearance timeof n-nonane, �-methylstyrene, two racemic epoxides was at 5.29,9.2, 12.9, 13.0 min, respectively. The similar run program was setfor indene.

2.2. Synthesis of the support (Scheme 1)

2.2.1. Synthesis of styrene–phenylvinyl phosphonic acidcopolymer (PS–PVPA)

1-Phenylvinyl phosphonic acid (PVPA) was synthesized accord-ing to literature [26] and its structures were confirmed by 1H NMR,31P NMR and FT-IR. 1H NMR (CDCl3): 6.06 (d, 1H), 6.23 (d, 1H),7.26–7.33 (m, 3H), 7.48 (m, 2H). 31P NMR (CD3OD): 15.9. IR (KBr):2710, 2240, 1500, 1200, 1040, 950, 780, 720, 700 cm−1.

1-Phenylvinyl phosphonic acid (4 g, 21.7 mmol), styrene (20 mL,173.9 mmol), ethyl acetate (150 mL) and benzoyl peroxide (BPO,1.0 g, 4.7 mmol) were used for preparation of PS–PVPA copoly-mer as literature [22] yield 7.52 g. GPC: Mn = 38,608, m = 38, n = 8,Mw/Mn = 2.

2.2.2. Synthesis of ZnPS–PVPAPS–PVPA (1.0 g, 1 mmol), sodium dihydrogen phosphate (0.62 g,

4 mmol), zinc acetate (1.1 g, 5 mmol) and Et3N (0.68 g, 6.7 mmol)were used for the synthesis of ZnPS–PVPA according to the litera-ture [27]. IR (KBr):�max/cm−1 3059, 3028, 2923 (CH), 1686, 1493,1453, 756, 698 (–C6H5), 1027 (P O).

2.2.3. Synthesis of chloromethyl-zinc poly(styrene-phenylvinyl-phosphonate)-phosphate (ZnCMPS–PVPA)

Chloromethyl methyl ether (9.3 mL), anhydrous zinc chloride(3.32 g, 24.34 mmol) and 1a (5.0 g, 3.4 mmol) were mixed in 40 mLchloroform and stirred at 40 ◦C for 10 h. After cooling down,

to obtain 2a (5.84 g, 90.1%). 2b–2h were synthesized in compli-ance with the similar course. IR (KBr):�max/cm−1 3026, 2925 (CH),2341(O P–OH), 1650, 1542, 1510, 1493 (–C6H5), 1267 (P O), 700(C–Cl) cm−1.

J. Huang et al. / Applied Catalysis A: General 407 (2011) 163– 172 165

nthesi

2p

N0amtttdcsZi

2Z

wEAwtro

2

t

2

2

iun(Cq

Scheme 1. The sy

.2.4. Synthesis of arylaminomethyl-zinc poly(styrene-henylvinylphosphonate)-phosphate (ZnAMPS–PVPA)

Proportional amount of aryldiamines (such as: m-NH2PhNH2, p-H2PhNH2, Benzidine) was blended with 2a (1 g), Na2CO3 (1.06 g,.01 mol), CuI (0.2 g, 1 mmol) and alcohol 50 mL (the mol ratio ofryldiamine to chlorine element in ZnAMPS–PVPA is 5:1), and theixture was stirred and kept at 70 ◦C for 12 h. After the reaction,

he mixture was neutralized by dilute hydrochloric acid and thenhe solvent was vaporized under decompression. Subsequently,he product 3a was filtered and washed with deionized water andried in vacuo. 3b–3h were gained in accordance with the sameourse. Reaction yield always exceeded 90%. Correspondingly, theharp C–Cl peak (owing to –CH2Cl groups) at 700 cm−1 in thenCMPS–PVPA actually vanished or was seen as a weak band afterntroduction of aryldiamines.

.3. Synthesis grafting chiral salen Mn(III) catalyst ontonAMPS–PVPA (Scheme 2)

Chiral salen Mn(III) (4 mmol) in 10 mL of THF was added drop-ise to the solution of 3a (0.5 g) pre-swelled in THF for 30 min and

t3N (5 mmol) with stirring. Then the mixture was refluxed for 10 h.fter cooling down, the solution was neutralized and the solventas evaporated. The dark brown powder 5a was obtained by fil-

ration and washed thoroughly with CH2Cl2 and deionized waterespectively until no Mn could be detected by AAS. 5b–5h werebtained according to the same process.

.4. Synthesis of arylamine modified chiral salen Mn(III)

The homogeneous chiral salen Mn(III) were prepared accordingo similar procedure to heterogeneous catalysts (in Section 2.3).

.5. Asymmetric epoxidation

.5.1. Using m-CPBA as oxidantThe activity of the prepared catalysts were tested for the epox-

dation of unfunctionalized olefins in CH2Cl2 at −40 ◦C for 5 hsing m-CPBA/NMO as oxidant and with alkene (1 mmol), n-

onane (internal standard, 1 mmol), NMO (5 mmol), homogeneous5 mol%) or heterogeneous salen Mn(III) catalysts (5 mol%) and m-PBA (2 mmol). After reaction, Na2CO3 (2 mL, 1.0 M) was added touench the reaction.

s of the supports.

2.5.2. Using NaIO4 as oxidantFor NaIO4/imidazole system, the reaction was carried out in the

2:1 mixture of acetonitrile: water at room temperature for 2.5 h andwith alkene (1 mmol), NaIO4 (2 mmol) in the presence of 5 mol%catalysts.

2.6. The reusability of the catalyst

In a typical recirculation, the equal volume of hexane was addedto the mixture after the reaction. Subsequently, the organic phasewas separated, and the catalyst was washed with hexane anddeionized water, and dried over vacuum at 60 ◦C. The recovereddried solid catalyst was weighed and reused in the next run. Inevery run the same proportion of the substrate-to-catalyst andsolvent-to-catalyst was retained.

2.7. General procedure for large-scale asymmetric epoxidationreaction

A solution of catalyst 3b (2.5 mmol), n-nonane (50 mmol) and �-methylstyrene (50 mmol) in CH2Cl2 (150 mL) at −40 ◦C was stirredfor 30 min. Then, m-CPBA (100 mmol) was added to the solutionstep by step. After reaction, Na2CO3 (100 mL, 1.0 M) was added toquench the reaction. And the organic layer was dried over sodiumsulfate, and the catalyst was precipitated out from the solutionby adding hexane and kept for subsequent use without furtherpurification. The conversion and ee values of the epoxide weredetermined by GC.

3. Results and discussion

3.1. Characterizations of the supports and the heterogeneouschiral catalysts

3.1.1. IR spectroscopy and UV–vis spectroscopyThe most informative evidence, which confirmed the anchoring

of the chiral salen Mn(III) complex 4 to the aryldiamine mod-ified ZnPS–PVPA, was obtained by FT-IR spectra (Fig. S1). Theazomethene (C N) stretching band of the complex 4 appeared at1612 cm−1 (5 in Fig. S1). While for the supported catalysts this band

was also observed at the vicinity of 1613 cm−1. All the samples(5a–5h) and the complex 4 had shown the same band at 1638 cm−1

attributed to the vibration of imine group. The stretching vibrationat 1030 cm−1 which was assigned to characteristic vibrations of the

166 J. Huang et al. / Applied Catalysis A: General 407 (2011) 163– 172

Scheme 2. Synthetic rou

ptTwt1p3t

stcttasiM3bZ

3

istwkofh

groups are pushed out and located on the external surface of

Fig. 1. XRD of (a) the heterogeneous catalyst 5c2; (b) ZnPS–PVPA.

hosphonic acid group in the support was obviously weakened dueo the electronic structure changes for the host–guest interaction.he common prominent bands in the spectra of compounds 1a–1fere three peaks at 1145, 1089, and 986 cm−1, which are attributed

o R–PO32− phosphonate stretching vibrations. The adsorptions at

201, 1144, and 1077 cm−1 were due to the phosphonate and phos-hate stretching vibrations. Moreover, an additional band around408 cm−1 was observed for the samples, which was assigned tohe stretching vibration of N–H groups.

Diffuse reflectance UV–vis spectra (Fig. S2) indicated that thepectra of the supported catalysts displayed features similar tohose of the neat chiral salen Mn(III) complex 4. According to theomplex 4, the bands at 334 nm could be attributed to the chargeransfer transition of salen ligand. The band at 435 nm was dueo the ligand-to-metal charge transfer transition, and the bandt 510 nm was assigned to the d–d transition of Mn(III) salenystem. While for the heterogeneous catalysts, all the character-stic bands appeared in their spectra but the immobilized salen

n(III) catalysts exhibited a blue shift from 334, 435 and 510 nm to30, 427 and 503 nm, which indicated that an interaction existedetween the salen Mn(III) complex and the aryldiamine modifiednPS–PVPA.

.1.2. Thermal gravimetric analysis and powder XRDAs described in the TG curves (Fig. S3), according to 5c2, the

nitial weight loss was 3.38% below 200 ◦C. It was ascribed tourface-bound or crystalline water in this stage. In the tempera-ure range 200–850 ◦C, the organic moieties decomposed. The totaleight loss was found to be 69.32%. Obviously, catalyst 5c2 still

ept high stability lower than 200 ◦C. In general, organic reactions

f heterogeneous catalysis were carried out below 200 ◦C. There-ore, catalyst 5c2 had adequate thermal stability to be applied ineterogeneous catalytic reactions.

te of the catalysts.

As could be seen from Fig. 1, the XRD patterns of ZnPS–PVPAdisplayed a broad 0 0 1 peak (the lowest-angle diffraction peak inthe pattern), accompanied with other peaks at higher-order 0 0 npeaks at larger angles and lower intensities such as at 38.04◦.Although the intensities of all peaks decreased after immobiliza-tion of Mn(III) salen complexes, the reflections for ZnPS–PVPAand the catalyst 5c2 indicated that the mesoporous structure ofthe parent supports remained intact on modification with aryl-diamine. Simultaneously, the interlayer distance which could becalculated (via the Bragg equation, n� = 2d sin �) of the immobi-lized catalyst 5c2 (43.3 ± 0.2 A) was nearly twice as much as thatof ZnPS–PVPA (21 ± 1.3 A), owing to the chiral salen Mn(III) intro-duced in ZnPS–PVPA making the zinc layer stretched and becomingbroader. The peculiar appearance was entirely different to the mostof the results reported [20–22]. The amount of Mn(salen) anchoredonto ZnPS–PVPA is in the range of 0.66–0.78 mmol/g ascertainedby AAS based on Mn element. In view of these facts, it could beinferred that the chiral salen Mn had been attached.

3.1.3. Nitrogen adsorption–desorption isothermsThe corresponding textural parameters calculated by N2

adsorption–desorption isotherms were presented in Table 1.As described in Table 1, by means of chloromethylation and ary-

lamination, an obvious increase in BET surface area was observed(1c vs 2c vs 3c2, from 4.9 to 36.9 and to 42.5 m2/g), represented as3c2 in Scheme 1, as well as increase in the pore volume (1c vs 2cvs 3c2, from 1.3 to 18.82 and to 24.2 × 10−2 cm3/g) and in averagepore diameter (1c vs 2c vs 3c2, from 3.5 to 10.21 and to 11.39 nm). Incontrast with this phenomenon, a decrease in BET surface area (3c2vs 5c2, from 42.5 to 31.66 m2/g), in pore volume (3c2 vs 5c2, from24.2 to 5.43 × 10−2 cm3/g) and in average pore diameter (3c2 vs 5c2,from 11.39 to 1.56 nm) was observed upon immobilization of thecomplex 2 onto ZnPS–PVPA modified by arylaminomethyl (3c2). Onthe basis of this, it could be deduced that some chiral salen Mn(III)complexes were immobilized on the external surface of ZnPS–PVPAand other chiral salen Mn(III) complexes were present inside thenanopores. In other words, there were two forms of immobilizationof ligand: inner type and outer type, just as shown in Fig. 2.

In Fig. 2, it was displayed the structures of chiral salen Mn(III)immobilized onto ZnPS–PVPA with different x values. Obviously,the polystyrenyl groups are located on the external surfaces orbetween the layers of ZnPS–PVPA. If the x values are big like 1e(x = 0.5) the room between two polystyrenyl groups will be small,and in order to exclude the higher energy arrangement of pen-dant groups segregating on the interlayer most of the polystyrenyl

ZnPS–PVPA. In contrast, if x values are small like 1c (x = 0.33) and1b (x = 0.25) the space between two polystyrenyl groups will be big.And most of the polystyrenyl groups are naturally located between

J. Huang et al. / Applied Catalysis A: General 407 (2011) 163– 172 167

Table 1Physicochemical characterization data of 1c, 2c, 3c2, 5a, 5c2, 5f and 5h.

Sample Elemental analysis Surfacearea (m2/g)

Pore volum(×10−2 cm3/g)

Average porediameter (nm)

Mn content(mmol/g)

Calc. Found

1c C 56.21 55.08 4.9 1.3 3.5 –H 5.02 4.97N – –

2c C 45.16 44.48 36.9 18.82 10.21 –H 4.35 4.09N – –

3c C 59.32 58.86 42.5 24.2 11.39 –H 5.14 4.91N 6.45 6.38

5a C 61.13 59.21 39.26 8.51 0.6 0.68H 6.57 6.25N 5.76 5.62

5c2 C 61.26 60.33 31.66 5.43 1.56 0.72H 6.35 6.01N 5.68 5.54

5f C 63.28 62.35 31.47 10.16 6.46 0.75H 7.34 7.16N 6.27 6.05

5h C 63.49 62.35 40.99 15.41 7.52 0.66H 6.51 6.13N 5.78 5.62

p anc

trl

3

sinttteflo(bgbtp

opp–PO3H2 group (7) by polystyrene chain which lied on the sur-face of another contiguous zinc interlayer space. Both opp–PO3H2group (5) and opp–PO3H2 group (6) which were conjunct to oneanother were located on the interlayer surface of the same zinc

Fig. 2. The modes of the organic grou

he layers of ZnPS–PVPA. Just as these special configurations gaveise to different catalytic activities which would be discussed in theatter.

.1.4. Structure analysis of the supportsIn the hypothesized models deduced for ZnPS–PVPA (Fig. 3),

ome oxygen atoms of the hydroxyl groups or hydroxy sodiumn the segments of the inorganic phosphate groups were coordi-ated with zinc atoms, making the zinc atoms self-assemble inhe same plane, while the other oxygen atoms of the portion ofhe inorganic phosphate groups in the ZnPS–PVPA stretched overhe surface of the zinc layer. On the other side, there are sev-ral types of organic polymer phosphonate–PO3H2 (opp–PO3H2)ormed in ZnPS–PVPA: opp–PO3H2 (3) was located on the inter-ayer surface of one zinc layer, and was connected to other particlef ZnPS–PVPA by polystyrene segment, in other words, opp–PO3H23) and its one neighboring opp–PO3H2 were located not in sameut in different particle of ZnPS–PVPA. The same to opp–PO3H2

roup (10). Opp–PO3H2 group (2) and (4) were linked each othery polystyrene chain and situated on the interlayer surface of thewo adjacent zinc layers respectively. Opp–PO3H2 group (1) waserched on the interlayer surface of one zinc layer and joined to

hored (a) inner type; (b) outer type.

Fig. 3. The hypothesized layered structure of ZnPS–PVPA.

1 sis A:

lpici

3

twagasaatw

tdtait

3

3˛

ewgTcido

t(portMimeZotzogrta1ec0tta

68 J. Huang et al. / Applied Cataly

ayer, similar to opp–PO3H2 group (8) and opp–PO3H2 group (9). Soores or channels of various sizes and shapes by appropriate mod-

fication of the styrene–phenylvinylphosphonic acid copolymerhain were formed that consequentially give birth to significantmpact on the excellent catalytic activity [27].

.1.5. Analysis of surface morphologyShown in Fig. 4, SEM images of 1c indicated the diameter of

he particles of the support were in the scope of micron. And 1cas consisted irregularly of many small and big layered particles

nd the particles which were aggregates of lots of minor crystallinerains. Meanwhile, the supports were various caves, holes, porousnd channels with different shape and size. Some micropores andecondary channels would increase the surface area of the catalystnd provide enough chance for substrates to access to the catalyticctive sites. According to 5c2, the SEM took on the amorphous struc-ure which was loose, and various caves, holes, porous and channelsith different shape and size.

The TEM photography of 1c (Fig. 5) manifested the structure ofhe support was spheroid, its channels, holes and cavums could beiscerned clearly, and their sizes were about 70–80 nm. While forhe heterogeneous catalyst 5c2, the configuration of it was filiformnd loose, and its channels, holes and cavums were also existed int. Just for this, substrates would have more chance to transfer tohe internal catalytic active sites in solution.

.2. Enantioselective epoxidation of unfunctionalized olefins

.2.1. The effect of the x values in the epoxidation of-methylstyrene and indene

The catalytic activity of immobilized catalyst 5a–5h for thepoxidation of �-methylstyrene and indene was studied in CH2Cl2ith m-CPBA as oxidative systems. Jacobsen’s catalyst 4 and homo-

eneous catalyst 6b were also examined for comparable purposes.he data obtained were summarized in Table 2. The heterogeneousatalysts exhibited comparable or even higher enantioselectiv-ties than Jacobsen’s catalyst 4 and homogeneous catalyst 6bid for the asymmetric epoxidation of some unfunctionalizedlefins.

In this text, �-methylstyrene was then chosen to investigatehe heterogeneous chiral Mn(III) salen catalysts. The ee values87.2–99%) with the catalysts of 5h and 5c2 are higher as com-ared with Jacobsen’s catalyst 4 (ee, 54%). Similar results werebtained by Kim and Shin [28] and Xiang et al. [11]. Kim and Shineported that, for the asymmetric epoxidation of �-methylstyrene,he ee increased from 51 to 59% after immobilization of chiral

n(III) salen on the siliceous MCM-41 by multi-step grafting. Thencrease in enantiomeric excess may also be attributed to the

icroenvironment effect and confinement effect [20]. Here, theseffects are provided by the layered structure and mesopores ofnPS–PVPA and the balance adjustment between the hydrophobicf polystyrene parts and the hydrophilic of phosphate parts. Andhese features are different from either pure polystyrene or pureinc phosphates. Moreover, the supports with different ratios ofrganic phosphonate and inorganic phosphate or different inor-anic phosphate resource could contribute to different catalyticesults. The conversions varied from 81.5% to 87.2% and the enan-ioselectivities varied from 13.3% to 14.6% as the x values were 0.25nd 0.2 (entries 5 and 4). In contrast, the x values were from 0.4 to, accompanied with the conversions from 35.5% to 84.1% and thenantioselectivities from 19.3% to 37% (entries 7–10). Notably, theonversions as x = 0.2, 0.25 were higher than that did as x = 0.4, 0.5,

.75, 1; while the enantioselectivities x = 0.2, 0.25 were lower thanhat of the catalysts when x = 0.4, 0.5, 0.75, 1. In other words, enan-ioselectivity was relatively low and conversion was relative highs x value was little; on the contrary, enantioselectivity showed

General 407 (2011) 163– 172

high and conversion displayed low when x value was compara-tively large. The phenomenon was ascribed to the confinementeffect of the nanopores which would be discussed attentively in thelatter.

Remarkably, the supported catalyst 5c2 with 1:2 ratio of organicphosphonate to inorganic phosphate displayed excellent conver-sion and enantioselectivity (conv, >99% and ee, >99%) comparedwith the other supported catalyst in Table 2 (entries 4–5 and entries7–11). Meanwhile, heterogeneous catalyst 5h also showed higherenantioselectivity than that of Jacobsen’s catalyst 4 (ee, 87.2% vs54%) and a little lower catalytic activity than that of the supportedcatalyst 5c2 (conv, 98.5% vs >99%; ee, 87.2% vs >99%). It was denotedthat NH4

+ in the support put effects on the catalytic activity. On thebasis of these results, it could be inferred that both proper ratio oforganic phosphonate and inorganic phosphate and the pertinentinorganic phosphate resource played vital impacts on the catalyticactivity.

The relatively bulkier alkene like indene (entries 16–28) wasalso chosen to test the activity of the supported catalysts 5a–5h forthe asymmetric epoxidation. These reaction results show that theenantioselectivity and activity of the immobilized catalyst 5b forindene (ee, 89.9% yield 85.9%) (entry 20) are found to be lower thanthat of 5c2 (entry 21), the reason may be that indene is too large toaccommodate into the micropores and layers of ZnPS–PVPA. There-fore, indene may merely react with a few active sites on the externalsurface of the ZnPS–PVPA and 5b cannot afford confinement effectfor indene. Meanwhile, under the same conditions the yield of theepoxide with 5a (entry 19) are better than those of 5b (entry 20), butthey show lower ee values. One possible explanation for this resultis that the amount of chiral salen Mn(III) catalytic active centers onexternal surface for 5a is much more than that for 5b. The higher eevalues obtained for the catalyst immobilized in the nanopores thanthat on the external surface can be really attributed to the enhancedchiral induction by the confinement effect of the nanopores. Whenthe nanopore size of the support is tuned to a suitable value, thechiral catalysts in the nanopores can give higher ee value in somecases. These results strongly suggest that the confinement effect ofnanopores is able to enhance the asymmetric induction as long asthe pore size is tuned to a suitable value depending on the catalyticreaction system [29].

3.2.2. The effect of different linkers in the epoxidation of˛-methylstyrene and indene

As described in Table 2, the catalyst 5c2 showed higher conver-sions and enantioselectivities than those of the catalyst 5c1 in theasymmetric epoxidation of indene (conv%, >99 vs 80.5; ee, >99 vs94.5), owing to the high symmetry of the catalyst 5c2 which coulddecreased the steric obstacles. In addition, the supported catalyst5c3 also displayed lower activity than the catalyst 5c2 did (conv%,86.4 vs >99; ee, 92.3 vs >99), which was ascribed to the bulkierlinker benzidine making the substrates approaching the catalystdifficultly. In addition, the homogeneous catalyst 6b also showedhigher ee values than that of the Jacobsen’s catalyst 4 (ee, 90.8–54%),which indicated that the rigid linker was devoted to the increaseof ee values. In other words, the steric properties of the linkagesreally played vital impacts on the configuration of the transitionstate for the asymmetric reactions. Moreover, the ee values fur-ther increased from 90.8% to >99% after the homogeneous catalyst6b immobilized onto ZnPS–PVPA. Above all, the whole immobi-lized chiral salen Mn(III) catalysts include the support ZnPS–PVPA,the rigid linkers and chiral salen Mn altogether contributed to theincrease of ee values.

3.2.3. The effect of axial ligands: NMO and imidazoleSurprisingly, the heterogeneous catalyst 5c2 and 5h displayed

high ee values and conversions in the absence of the additive NMO

J. Huang et al. / Applied Catalysis A: General 407 (2011) 163– 172 169

Fig. 4. SEM images of 1c (1) and 5c2 (2).

and th

wTmatitcnorsWaltrmrNtsclr

aaapscl

Fig. 5. TEM photographs of 1c (A)

hich is commonly required to improve the catalytic activity(inable 2). Practically, adding the axial ligand NMO to our reactionixture did not improve the asymmetric induction but result in

dramatically reduced enantioselectivity and reactivity. In thisext, the ee values for the epoxides of �-methylstyrene typicallyncreased from 3.1% to >99% and the conversion increased from 6.6%o >99% (entry 14 vs 6) without the addition of NMO. This stood inontrast to the most literatures reported [30]. The exceptional phe-omenon originated in chiral salen Mn(III) complex immobilizedn phenoxy-modified ZPS–PVPA has been reported by our groupecently [23]. It was ascribed to organic polymer–inorganic hybridupport ZPS–PVPA and the phenoxide axial coordinating group.

hereas, this unusual phenomenon in this text was induced bynother factors. At first, the structures of the immobilized cata-ysts similar with N-oxide ligand acted as axial ligands leaded tohe unusual phenomenon. Simultaneously, additives are generallyegarded as axial ligands on the transition metal catalyst, whichake for activating the catalyst either toward oxidation or toward

eactivity with the olefin. Thus, there was a steric hindrance when-oxide ligand was added and the optimal geometric configura-

ion of the reactive intermediate salen Mn(V) O was altered. It wasteric hindrance that made olefins approach salen Mn(V) O diffi-ultly and the lower ee values were obtained. Obviously, there is aot of work to be done on mechanistic and structural aspects of oureaction and catalyst.

On the other hand, the heterogeneous catalysts 5a–5h were alsopplied in the epoxidation �-methylstyrene in the 2:1 mixture ofcetonitrile: water with NaIO4 as oxidative system. Jacobsen’s cat-lyst 4 and homogeneous catalyst 6b were tested for comparable

urposes. As described in Table 2, the supported catalysts 5a–5hhowed comparable catalytic activity to that of the homogeneousatalyst 6 and Jacobsen’s catalyst 4. The conversion and enantiose-ectivity of the catalysts 5a–5h all exceeded 95%, even >99%. It could

e heterogeneous catalyst 5c2 (B).

be deduced that the heterogeneous catalysts 5a–5h possessed effi-cient catalytic abilities whether the axial ligand imidazole existedor not. That is to say, the axial ligand imidazole made subtle impacton the catalytic activities that the conversion and enantioselectiv-ity increased a little in the presence of imidazole (entry 34 vs 35:conv%, from 98.6 to >99; ee%, >99 vs >99).

In general, the additives N-methylmorpholine N-oxide (NMO)and imidazole, which are used to improve epoxidation yields andenantioselection, bind to the Mn(III) center prior to the epoxida-tion reaction, as evidenced by the alteration of the Mn(III) parallelmode EPR signal [31]. Additives to the Mn(III) salen reaction mix-ture, such as NMO and imidazole, generally facilitate faster reactionrates, higher epoxide yields, and improved enantioselectivity. How-ever, the additives in this text played such different roles that thecatalytic activities did not increase but decrease with the additionof NMO in m-CPBA as oxidative system or slightly increased in thepresence of imidazole with NaIO4 as oxidative system. Further-more, the additives were expensive commonly and the superiorcatalytic activities were still obtained in the absence of them. Fromthe commercial viewpoint, the heterogeneous catalyst 5c2 had thepotential application in industry.

3.3. The confinement effect originated in the x values ofZnPS–PVPA

In the ZnPS–PVPA hybrid material, residue or side chain forOPP-groups of P–G and inorganic phosphate P–OH are sufficientlydifferent in size, obviously OPP-groups of P–G are quite bulky andinorganic phosphate P–OH are relatively small. These hybrid mate-

rials usually contain a random distribution of the organic groupssuch that all layers have identical stoichiometry. In such systems,the interlayer spacing (d-space) is a function of pendant group sto-ichiometry, and has a generally linear dependence on component

170 J. Huang et al. / Applied Catalysis A: General 407 (2011) 163– 172

Table 2Asymmetric epoxidation of �-methylstyrene and indene catalyzed by homogeneous and heterogeneous catalysts (5a–5h) with m-CPBA/NMOa as oxidant systems.

Entry Substrateb Catalyst Oxidant system Time (h) T (◦C) Conv % eec TOFd× 10−4 (s−1)

1 A 4 m-CPBA/NMO 5 −40 >99 54 11.112 A 6b m-CPBA/NMO 5 −40 88 86 9.783 A 6b m-CPBA 5 −40 64 90.8 7.114 A 5a m-CPBA 5 −40 87.2 14.6 9.695 A 5b m-CPBA 5 −40 81.5 13.3 9.056 A 5c2 m-CPBA 5 −40 >99 >99 11.117 A 5d m-CPBA 5 −40 73.7 31.2 8.198 A 5e m-CPBA 5 −40 76.2 37 8.479 A 5f m-CPBA 5 −40 35.5 19.3 3.94

10 A 5g m-CPBA 5 −40 84.1 27.1 9.3411 A 5h m-CPBA 5 −40 98.5 87.2 10.9412 A 5c1 m-CPBA 5 −40 98.6 >99 10.9513 A 5c3 m-CPBA 5 −40 96.8 >99 10.7514 A 5c2 m-CPBA/NMO 5 −40 6.6 3.1 0.7315 A 5sh m-CPBA/NMO 5 −40 96.7 1.2 10.7416 B 4 m-CPBA/NMO 1 0 92 65 51.1117 B 6b m-CPBA/NMO 1 0 91.2 25.5 50.6618 B 6b m-CPBA 1 0 98.7 83.7 54.8319 B 5a m-CPBA 1 0 91.7 50 50.9420 B 5b m-CPBA 1 0 85.9 89.9 47.7221 B 5c2 m-CPBA 1 0 >99 >99 55.5522 B 5d m-CPBA 1 0 74.2 78 41.2223 B 5e m-CPBA 1 0 73.3 83.7 40.7224 B 5f m-CPBA 1 0 67.9 93.6 37.7225 B 5g m-CPBA 1 0 48.1 95.8 26.7226 B 5h m-CPBA 1 0 96.1 54.6 53.3827 B 5c1 m-CPBA 1 0 80.5 94.5 44.7228 B 5c3 m-CPBA 1 0 86.4 92.3 47.9929 A 4 NaIO4/Imidazole 7 25 >99 >99 7.9430 A 6b NaIO4/Imidazole 7 25 >99 >99 7.9431 A 6b NaIO4 7 25 >99 >99 7.9432 A 5c2 NaIO4/Imidazole 7 25 >99 >99 7.9433 A 5c2 NaIO4 7 25 >99 >99 7.9434 A 5h NaIO4/Imidazole 7 25 98.6 >99 7.8335 A 5h NaIO4 7 25 >99 >99 7.94

a Reactions were carried out in CH2Cl2 (4 mL) with alkene (1 mmol), n-nonane (internal standard, 1 mmol), NMO (5 mmol), homogeneous (5 mol%) or heteroge-neous salen Mn(III) catalysts (5 mol%) and m-CPBA (2 mmol). The conversion and the ee value were determined by GC with chiral capillary columns HP19091G-B 213,30 m × 0.32 mm × 0.25 �m.

b

] × time (s−1).

mc[

wwerlttPps(et“sganfZ“(m

A, �-methylstyrene; B, indene.c (S)-form.d Turnover frequency (TOF) is calculated by the expression of [product]/[catalyst

ole fraction. Intermediate x values result in intermediate stericonstraints, resulting in d values between the two organic groups27].

Then the models of Zn (NaPO4)1−x [PS–PVPA]x·yH2O series, viz.here x = 0.2, 0.25, 0.33, 0.4, 0.5, 0.75 and 1 were deduced. Ase know, each zinc atom is bonded to six oxygen atoms, and

very three of these oxygen atoms are bonded to one phospho-us atom. As a result, the layers of the ZnPS–PVPA are formed. Aayer of ZnPS–PVPA 1e (x = 0.5) presents that in order to excludehe higher-energy arrangement of pendant groups segregating onhe interlayer, there is always a P–G (B) groups insert between two–OH (A) or reverse. That is to say, the ideal model can be sim-ly denoted as “ABABAB. . .”, and the scheme of the ideal crossection (Fig. 6) indicates that two bulky P–G (B) and four P–OHA) groups are located around one P–G group, sterically directednergy allows bulky P–G groups to get out of the layers. Similarly,he ideal model for ZnPS–PVPA 1c (x = 0.33) can also be denoted asAABAABAAB. . .”, and the scheme of the ideal cross section (Fig. 6)uggests that six small P–OH (A) groups are located around one P–Group. The d-space is bigger than that of ZnPS–PVPA 1e (x = 0.5),nd it can offer enough space for some smaller substrates coordi-ation to the active sites. Meanwhile, confinement effect generated

or steric hindrance among the substrates and active sites. For

nPS–PVPA 1b (x = 0.25) the ideal model can also be denoted asBAAABAAABAA. . .”, and the scheme of the ideal cross sectionFig. 6) suggests that the space of two neighboring P–G groups is

uch larger than the other two materials. Although there is big

Fig. 6. The cross section of the x = 0.25, x = 0.33 and x = 0.5.

room for catalytic reaction, the confinement effect and the contentof the active sites always are decreased [27].

Qualitatively, the changes in the d spacing with composition canbe rationalized in terms of interrelated factors that pertain to thesteric interactions of the bulkier styrene–phenylvinylphosphonicacid copolymer groups and the conformations of the organic

groups. The styrene–phenylvinylphosphonic acid copolymerchains are located on the external surfaces or between the layersof ZnPS–PVPA. If the x values are big like ZnPS–PVPA 1e (x = 0.5)or more the room between two styrene–phenylvinylphosphonic

J. Huang et al. / Applied Catalysis A: General 407 (2011) 163– 172 171

atoansoaisutbaacnMise

3

thts

cscaccaskAaiytcM

Table 3The recycles of catalyst 5c2 in the asymmetric epoxidation of �-methylstyrene.a

Run Time (h) Conversion (%) ee (%)b TOFc × 10−4 (s−1)

1 5 >99 >99 11.112 5 >99 >99 11.113 5 >99 >99 11.114 5 >99 >99 11.115 5 97.8 >99 10.876 5 96 >99 10.677 5 92 97 10.228 5 90 93 9.999 5 88 86.1 9.78

10 5 85.2 78.9 9.4711 5 80.1 65.4 8.9012 5 73 29.4 8.11

a Reactions were carried out at −40 ◦C in CH2 Cl2 (2 mL) with �-methylstyrene(1 mmol), n-nonane (internal standard, 1 mmol), m-CPBA (0.38 mmol), het-erogeneous salen Mn(III) catalysts (5 mol%). The conversion and the eevalue were determined by GC with chiral capillary columns HP19091G-B213,30 m × 0.32 mm × 0.25 �m.

b Same as in Table 2.c Same as in Table 2.

Table 4Large-scale asymmetric epoxidation reaction of �-methylstyrene.a

Entry Time (h) Conversion (%) eef(%) TOFg × 10−4 (S−1)

1b 5 >99 >99 11.112c 5 >99 >99 11.113d 5 >99 >99 11.114e 5 >99 >99 11.11

a Reactions were carried out at −40 ◦C in CH2 Cl2 with �-methylstyrene, n-nonane,m-CPBA, heterogeneous salen Mn(III) catalysts (5 mol%). The conversion and theee value were determined by GC with chiral capillary columns HP19091G-B213,30 m × 0.32 mm × 0.25 �m.

b The usage amounts of reagents were �-methylstyrene (1 mmol), n-nonane(1 mmol), heterogeneous catalyst 3b (0.05 mmol), m-CPBA(2 mmol), respectively.

c The usage amounts of reagents were �-methylstyrene (50 mmol), n-nonane(50 mmol), heterogeneous catalyst 3b (2.5 mmol), m-CPBA(100 mmol), respectively.

d The usage amounts of reagents were �-methylstyrene (50 mmol), n-nonane(50 mmol), heterogeneous catalyst 3b (0.5 mmol), m-CPBA(100 mmol), respectively.

e The usage amounts of reagents were �-methylstyrene (100 mmol), n-nonane

Fig. 7. The hypothesized structure of the heterogeneous catalyst.

cid copolymer groups will be small, and in order to excludehe higher-energy arrangement of pendant groups segregatingn the interlayer the most of the styrene–phenylvinylphosphoniccid copolymer groups are pushed out and located on the exter-al surface of ZnPS–PVPA. In contrast, if x values are relativelymall like ZnPS–PVPA 1c (x = 0.33) and ZnPS–PVPA 1b (x = 0.25)r less the space between two styrene–phenylvinylphosphoniccid copolymer groups will be big which was in good accord-ng with the results obtained by XRD. And more part of thetyrene–phenylvinylphosphonic acid copolymer groups are nat-rally located between the layers of ZnPS–PVPA. In conclusion,he frameworks of ZnPS–PVPA can be easily designed and assem-led to generate pores or channels of various sizes and shapes byppropriate modification of the styrene–phenylvinylphosphoniccid copolymer chain (Fig. 7). The porous hosted materials affectatalytic performance due to a cooperative interaction among theanoporous solid, immobilizing linker, and Mn–salen complex.esoporous materials are the most applicable supports for the

mmobilization of Mn–salen complexes [32]. Just as this specialtructures of ZnPS–PVPA contributed to the excellent catalyticffect.

.4. The reusability of the catalyst

The reusability of a heterogeneous catalyst is of great impor-ance from synthetic and economical points of view. Theomogeneous catalysts could not recover even one time, in con-rast, the supported catalysts 5a–5h could be filtered and reusedeveral times without significant loss of their activity.

To assess the long-term stability and reusability of the supportedhiral salen Mn(III) catalysts, ˛-methylstyrene was used as a modeubstrate, and recycling experiments were carried out with theatalyst 5c2. At the end of the each reaction, the catalyst was sep-rated by adding hexane, washed with deionized water and driedarefully before using it in the next run. Above 95% recycle of theatalyst was achieved in every run. The recovered dried solid cat-lyst was weighed and reused in the next run. In every run theame ratio of the substrate-to-catalyst and solvent-to-catalyst wasept. The filtrates were collected for determination of Mn leaching.fter using of catalyst 5c2 for twelve consecutive times, the resultsre listed in Table 3. Obviously, the yield and the enantioselectiv-ty decreased slightly after recycling for nine times and still gave

ield (88%) and enantioselectivity (86.1%). The effective separationhe chiral Mn(III) salen complexes by the solid support ZnPS–PVPAontributed to the good stability of the heterogeneous chiraln(III) salen catalyst in case that they would dimerize to inactive

(100 mmol), heterogeneous catalyst 3b (5 mmol), m-CPBA(200 mmol), respectively.f Same as in Table 3.g Same as in Table 3.

�-oxo-Mn(IV) species. The decrease of the yield can be attributedto the decomposition of the chiral Mn(III) salen complex underepoxidation conditions [33] and the loss of the hyperfine gran-ules of the heterogeneous chiral Mn(III) salen catalysts (formed inreaction due to stirring). The Mn content of the heterogeneous cat-alyst 5c2 is 0.46 mmol/g compared with the total amount (around0.72 mmol/g) when the heterogeneous catalyst recycled for 9 times.

The nature of the recovered catalyst 5c2 was followed by IR(Fig. S4). The result indicated that characteristic bands of the cat-alyst at 2954, 2864 and 1630 cm−1 disappeared or weaken afterrecycling ten times. These revealed that the active sites of salenMn(III) complex and the ZnPS–PVPA support under acid reactionconditions were partly destroyed (Fig. 8). Moreover, other effectscan be used to explain these results: (i) leaching of metal complexesfrom the materials or (ii) blocking of the pores and secondary chan-nels either by inactive Mn(IV)-oxo species believed to be generatedduring the catalytic mechanism [34] or by some other insolubledegraded product obtained by side reactions, which after severalwashing could not be removed from the materials or (iii) collapsingof some of the pillars during the catalysis experiments.

3.5. Large-scale asymmetric epoxidation reaction

We further performed different proportions of large-scale asym-metric epoxidation reactions with n-nonane and �-methylstyrene

172 J. Huang et al. / Applied Catalysis A: General 407 (2011) 163– 172

ogres

amcrdtda

4

mtbtmpTnweoeatwsitfacaep

A

SaNuH

[

[[[[[

[[[

[[[

[

[

[

[

[

[[[[[

123 (2001) 5710–5719.

Fig. 8. The theoretic changing pr

nd m-CPBA. The same catalyst loading of 5 mol% as in the experi-ental scale was used. The large-scale experiments can be facilely

arried out using the same procedure as for the experimental scaleeactions. As can be seen from the results summarized in Table 4,elightfully, the conversion and enantioselectivity maintained athe same level for the large-scale reactions under whichever con-ition that the large scale is 50 times or 100 (Fig. S5) times as muchs the experimental scale.

. Conclusions

In summary, novel types of organic polymer–inorganic hybridaterial layered crystalline ZnPS–PVPA with different content of

he organic group (x) and different inorganic phosphate haveeen synthesized and applied as catalyst supports. Thereafter,he chiral salen Mn(III) complex immobilized onto aryldiamine

odified ZnPS–PVPA through axial coordination were also pre-ared. The characterization of FT-IR, UV–vis spectra, AAS, SEM andEM for the supported catalysts indicated that the axial coordi-ation attachment of the chiral salen Mn(III) to ZnAMPS–PVPAas successful. The heterogeneous chiral Mn(III) salen catalysts

xhibited comparable or even higher enantioselectivities than thatf homogeneous catalyst for the asymmetric epoxidation of sev-ral unfunctional olefins in the absence of imidazole and NMO. Inddition, the influences of x values, linkages and the additives onhe catalytic activities of the heterogeneous Mn(III) salen catalystsere explored at length. Furthermore, the supported chiral Mn(III)

alen catalysts are relatively stable and can be recycled nine timesn the asymmetric epoxidation of �-methylstyrene. Delightfully,his organocatalyzed asymmetric epoxidation reaction can be per-ormed on a large-scale with the catalytic ability being maintainedt the same level. The excellent properties of the heterogeneousatalysts are attributed to the special structure of ZnPS–PVPA. In

word, chiral salen Mn(III) anchored on ZnPS–PVPA are stable,ffective and promising catalysts and may be provided with theotentiality for industry.

cknowledgements

This work was financially supported by National Ministry ofcience and Technology Innovation Fund for High-tech Small

nd Medium Enterprise Technology (NO. 09C26215112399) andational Ministry of Human Resources and Social Security Start-p Support Projects for Students Returned to Business, Office ofuman Resources and Social Security Issued 2009 (143).

[[

[

s of catalyst 5c2 in acid solution.

Appendix A. Supplementary data

Supplementary data associated with this article can be found, inthe online version, at doi:10.1016/j.apcata.2011.08.035.

References

[1] E.M.M. Garrigle, D.G. Gilheany, Chem. Rev. 105 (2005) 1563–1602.[2] V. Mirkhani, M. Moghadam, S. Tangestaninejad, I.M. Baltork, N. Rasouli, Inorg.

Chem. Commun. 10 (2007) 1537–1540.[3] E. Angelescu, O.D. Pavel, R. Bıˆrjega, M. Florea, R. Zavoianu, Appl. Catal. A: Gen.

341 (2008) 50–57.[4] S. Zhong, Y. Tan, Z. Fu, Q. Xie, F. Xie, X. Zhou, Z. Ye, G. Peng, D. Yin, J. Catal. 256

(2008) 154–158.[5] L. Feng, E. Urnezius, R.L. Luck, J. Org. Chem. 693 (2008) 1564–1571.[6] P. Besse, H. Veschambre, Tetrahedron 50 (1994) 8885–8927.[7] J.F. Larrow, E.N. Jacobsen, Top. Organomet. Chem. 6 (2004) 123–152.[8] T. Katsuki, Synlett (2003) 281–297.[9] K.P. Bryliakov, O.A. Kholdeeva, M.P. Vanina, et al., J. Mol. Catal. A: Chem. 178

(2002) 47–53.10] F. Minutulo, D. Pini, P. Antonella, P. Salvadori, Tetrahedron: Asymmetry 7 (1996)

2293–2302.11] S. Xiang, Y.L. Zhang, Q. Xin, C. Li, Chem. Commun. 22 (2002) 2696–2697.12] S.B. Ogunwumi, T. Bein, Chem. Commun. 9 (1997) 901–902.13] R.I. Kureshy, N.H. Khan, S.H.R. Abdi, I. Ahmad, J. Catal. 221 (2004) 234–240.14] C. Baleizão, H. Garcia, Chem. Rev. 106 (2006) 3987–4043.15] N.S. Venkataramanan, G. Kuppuraj, S. Rajagopal, Coord. Chem. Rev. 249 (2005)

1249–1268.16] X.K. Fu, W. Kang, S.Y. Wen, C.Y. Deng, Chin. J. Appl. Chem. 12 (1995) 47–49.17] R.Q. Zeng, X.K. Fu, C.B. Gong, Y. Sui, J. Mater. Sci. 41 (2006) 4771–4776.18] R.Q. Zeng, X.K. Fu, C.B. Gong, Y. Sui, X.B. Ma, X.B. Yang, J. Mol. Catal. A: Chem.

229 (2005) 1–5.19] X.B. Ma, X.K. Fu, J. Mol. Catal. A: Chem. 208 (2004) 129–133.20] R.F. Bai, X.K. Fu, H.B. Bao, W.S. Ren, Catal. Commun. 9 (2008) 1588–1594.21] W.S. Ren, X.K. Fu, H.B. Bao, R.F. Bai, P.P. Ding, B.L. Sui, Catal. Commun. 10 (2009)

788–793.22] B.W. Gong, X.K. Fu, J.X. Chen, Y.D. Li, X.C. Zou, X.B. Tu, P.P. Ding, L.P. Ma, J. Catal.

262 (2009) 9–17.23] X.C. Zou, X.K. Fu, Y.D. Li, X.B. Tu, S.D. Fu, Y.F. Luo, X.J. Wu, Adv. Synth. Catal. 352

(2010) 163–170.24] W. Zhang, J.L. Loebach, S.R. Wilson, E.N. Jacobson, J. Am. Chem. Soc. 112 (1990)

2801–2803.25] M. Palncki, P.J. Pospisil, W. Zhang, E.N. Jacobsen, J. Am. Chem. Soc. 116 (1994)

9333–9334.26] B.W. Gong, X.K. Fu, H.S. Shen, L.H. Ma, Y.J. Xia, Fine Chem. 25 (2008)

603–605.27] J. Huang, X.K. Fu, G. Wang, C. Li, X.Y. Hu, Dalton Trans. 40 (2011) 3631–3639.28] G.J. Kim, J.H. Shin, Tetrahedron Lett. 40 (1999) 6827–6830.29] C. Li, H.D. Zhang, D.M. Jiang, Q.H. Yang, Chem. Commun. 6 (2007) 547–558.30] C. Li, Catal. Rev. 46 (2004) 419–492.31] K.A. Campbell, M.R. Lashley, J.K. Wyatt, M.H. Nantz, R.D. Britt, J. Am. Chem. Soc.

32] H.D. Zhang, S. Xiang, C. Li, Chem. Commun. 9 (2005) 1209–1211.33] C.E. Song, E.J. Roh, B.M. Yu, D.Y. Chi, S.C. Kim, K.J. Lee, Chem. Commun. (2000)

615–616.34] P.M. Morn, G.J. Hutchings, Chem. Soc. Rev. 33 (2004) 108–122.