Pd nanoparticles encaged in nanoporous

Reactive & Functional Polymers 71 (2011) 756–765

Contents lists available at ScienceDirect

Reactive & Functional Polymers journal homepage: www.elsevier.com/locate/react

Pd nanoparticles encaged in nanoporous interpenetrating polymer networks: A robust recyclable catalyst for Heck reactions Kan Zhan a, Huanhuan You a, Wenyu Liu a, Jie Lu a, Ping Lu a, Jian Dong a,b,⇑ a b

School of Chemistry and Chemical Engineering, Shaoxing University, Shaoxing, Zhejiang 312000, China State Key Laboratory of Coordination Chemistry, Nanjing University, Nanjing, Jiangsu 210093, China

a r t i c l e

i n f o

Article history: Received 20 January 2011 Received in revised form 14 April 2011 Accepted 22 April 2011 Available online 29 April 2011 Keywords: Green chemistry Heck reaction Palladium Interpenetrating polymer network Nanoparticle

a b s t r a c t A novel type of Heck reaction catalyst composed of hydrophilic interpenetrating polymer networks (IPNs) and palladium (Pd) nanoparticles was prepared by simultaneous crosslinking of polyvinyl alcohol and polyacrylamide. The mesh sizes of the IPNs are one order of magnitude smaller than the average sizes of the uniformly dispersed Pd nanoparticles, which functions well to stabilize Pd nanoparticles and prevent aggregation. The Pd particles in the IPNs can catalyze Heck coupling reactions with high activities and be recycled over 20 times, providing more sustainability than any other polymer-stabilized Pd catalyst reported in the literature. The confined reactions inside the IPN nanopores in neat water may provide a green route for C–C coupling reactions. Ó 2011 Elsevier Ltd. All rights reserved.

1. Introduction Metal nanoparticles dispersed in polymers are attractive catalysts for a variety of reactions. They are usually in the form of dispersions in colloidal systems, membranes or porous resins [1–10]. In particular, colloidal palladium (Pd) nanoparticles protected by hydrophilic polymers, such as polyvinyl pyrrolidone (PVP) and polyvinyl alcohol (PVA), can demonstrate excellent catalytic activity and selectivity in comparison with general Pd catalysts [1,2,4,7]. Likewise, Pd nanoparticles encapsulated in dendrimers [5a,11–13], star-shaped block polymers [14], hyper-branched polymers [15], and crosslinked polymers [16,17] have been proposed as catalysts for C–C bond coupling reactions. Such Pd catalysts for C–C bond coupling reactions are extremely important in modern organic synthesis; however, additional studies have shown that Pd nanoparticles used in Heck coupling reactions could only be reused a limited number of times [5–9,14–29a]. In particular, when open polymer structures (e.g., colloidal dispersions, dendrimers or porous resins) are used as support materials [9,11–17], Pd leaching into the reaction products could become a severe problem in the pharmaceutical industry. An important goal in the applications of Heck reactions is the development of a catalyst for ‘‘permanently’’ sustainable and environmentally benign ⇑ Corresponding authors. Address: School of Chemistry and Chemical Engineering, Shaoxing University, Shaoxing, Zhejiang 312000, China. Tel.: +86 575 88342511; fax: +86 575 88341521 (Jian Dong). E-mail address: [email protected] (J. Dong). 1381-5148/$ - see front matter Ó 2011 Elsevier Ltd. All rights reserved. doi:10.1016/j.reactfunctpolym.2011.04.007

reaction conditions with negligible metal leaching during its use. Therefore, we have focused on designing ‘‘heterogeneous systems’’ by confining the Pd particles to interpenetrating polymer networks (IPNs) with nanometer-scale meshes and encaging the C–C bond coupling reactions inside the IPNs. Such a catalytic system differs from the conventional polymer supported metal systems found in the literature [1–29]. An interpenetrating polymer network (IPN) is composed of two or more polymer networks, which are interlaced on a molecular scale [30]. The first advantage of the IPNs is the attainment of a combination of properties of two component polymers, resulting in the amphiphilicity required for different solvent conditions. Because there is no covalent bonding between two component networks, each network may retain its own property while the proportion of each network can be adjusted independently. Interpenetration of two networks may also result in better resistance to harsh reaction conditions (such as high temperature and continuous processing) due to higher mechanical strengths and stability than the individual homopolymer network or conventional crosslinked resins. The second advantage of the presence of two polymer networks is that the metal particles may be effectively prevented from coalescencing during the reactions; however, to the best of our knowledge, encaging cross-coupling reactions inside IPN structures has not been reported. We show here for the first time that metal-IPN catalysts (referred to as metal@IPN) are easily separated from the reaction medium, resulting in little contamination to the products, unlike conventional colloidal metal particles used in homogeneous catalysis. Therefore, the metal@IPN

757

K. Zhan et al. / Reactive & Functional Polymers 71 (2011) 756–765

2.6 nm

d = 32.9

10.8 nm

nm

0 4.

OH

OH

OH

OH

OH

OH

O

O

OH

OH

OH

OH

OH

O

O

OH

OH

n OH

OH

OH

H2N O

H2N NH2

NH2 O

O O

O

H2N

O H2N

O

H2N

H2N

H2N NH2

O

O

O

O

NH2 H2N

H2N

HN NH2

O O

O

O

NH2 HN

n O

O

O

NH2

O H2N

CH2 H2N O

NH2 H2N

NH2 O

O

O

Scheme 1. Interpenetrating polymer networks comprised of crosslinked PVA (blue lines) and PAM (red lines) with Pd nanoparticles (solid spheres). The mesh sizes for the linear distance between consecutive crosslinks and average size of Pd particles shown here correspond to those of Pd@IPN with VA:AM = 0.826:1 in water listed in Table 1. Structures of the crosslinked PVA and PAM are shown in a magnified view in the bottom. (For interpretation of the references to colour in this figure legend, the reader is referred to the web version of this article.)

systems can be tailored as reusable, environmentally friendly and mechanically strong catalysts for rigorous reaction conditions with desired affinities in different reaction media. The confinement of the cross-coupling reactions inside the network pores can be achieved by controlling the mesh sizes and contents of the component polymers (see Scheme 1 for a diagram of an IPN with dispersed metal particles). The importance of our study can be seen from several perspectives. Heterogenization of a homogeneous catalyst (or vice versa, homogenization of a heterogeneous catalyst) is an important approach for taking full advantage of both heterogeneous and homogenous catalysis. It is known that when employing homogeneous catalysis by Pd2+ salt, Heck coupling reactions suffer from poor reusability and require the removal of residual Pd, which can be a complex process. In pharmaceutical applications, contamination with metallic species in the final products is always closely monitored. In a recent study, crosslinked polyacrylamide was used as a Pd metal support to catalyze Heck reactions; however, the catalyst could only be reused six times [31a]. Another recent study proposed that the use of a charged polymer complex stabilized

Pd nanoparticles for Heck reactions in water [31b]. We show here that the Pd nanoparticles in nanoporous IPN systems can serve as green catalysts with higher activity and better recyclability than the Pd nanoparticles in charged polymer complexes [31b].

2. Experimental 2.1. Materials Analytical grade palladium chloride, sulfuric acid, acetic acid, glutaraldehyde (25% aqueous solution), PVA, ethanol, acrylamide, potassium persulfate, N,N0 -methylene bisacrylamide, N,N-dimethyl formamide (DMF), potassium hydroxide, hydrochloric acid, triethylamine, and hexadecyltrimethylammonium bromide were purchased from Sinopharm (Shanghai). Iodobenzene, 4-iodoanisole, styrene, methyl acrylate, acrylic acid, p-fluoroiodobenzene, 2-iodothiophene, and 4,40 -diiodobiphenyl were obtained from Aladdin Reagents Company (Shanghai). All of these chemical compounds were used as received without further purification.

758


2.2. Preparation and characterization of the Pd@IPN catalysts IPNs can be prepared using two possible methods. Preparation of IPNs using the simultaneous method requires two different types of monomers (or polymers) that are mixed homogeneously and then polymerized (and/or crosslinked) independently by different initiators (and/or crosslinkers). The sequential method of IPN preparation requires first crosslinking a linear polymer to form a network structure, then a new monomer, initiator and crosslinker are swollen into the network and polymerized and crosslinked among the first polymer to form an interpenetrated network. To acquire uniform dispersion, the Pd particles were generated by chemical reduction in solution and protected from aggregation by the presence of PVA. The PVA-protected metal solutions were subsequently transformed into IPN structures by the simultaneous approach that was originally used for synthesizing metal-free IPN structures in drug control release studies [32–34] but modified here to optimize the amount of crosslinkers (i.e., to generate mesh sizes appropriate for encaging the Pd nanoparticles) and the corresponding amount of acid required for catalyzing the crosslinking reactions to obtain optimal performance in Heck coupling reactions. The PVA–Pd colloidal solutions were mixed thoroughly with an acrylamide (AM) monomer and persulfate initiator. Two different types of crosslinkers were also mixed into the PVA–Pd solution to crosslink the PVA and polyacrylamide (PAM) independently. Solid PVA (4.0 g), with a number averaged Mn = 22,100, weight averaged Mw = 70,000, degree of polydispersity D = 2.79 (as determined by gel permeation chromatography) and a degree of saponification >99.8%, were mixed with 30 mL of ethanol and 25 mL of distilled water in a round-bottom flask. The mixture was heated to 97 °C under the reflux condition for 2 h to produce a transparent solution into which 20 mL of PdCl2–HCl with a Pd concentration of 5.64 mmol/L (prepared by mixing 0.0200 g PdCl2 and 20.0 mL of 0.5 M HCl) was added. The mixture was heated under the reflux condition for 5 h in a silicone oil bath and then cooled to room temperature. This black homogeneous solution was mixed with 0.2 mL of 10% H2SO4 (catalyst), 0.6 mL of 10% acetic acid (pH controller), and 0.8 mL of 25% glutaraldehyde aqueous solution (‘‘GA’’, crosslinker) to form solution A. The PVA–Pd solution A when crosslinked has a nominal crosslinking ratio (X) of 2.2%, which is the ratio of moles of crosslinking agent to moles of polymer repeating units. Solution B was prepared by combining 7.82 g of acrylamide and 0.93 g of N,N0 -methylene bisacrylamide (‘‘MBAM,’’ crosslinker) that were then mixed with 0.15 g of potassium persulfate (initiator) and 20 mL of distilled water. The PAM homopolymer obtained from solution B has a nominal crosslinking ratio (X) of 5.5%. Solutions A and B were mixed thoroughly in different ratios, poured into petri dishes, and heated at 70 °C for 3 h to form solid IPN gels, which were washed with distilled water repeatedly overnight and dried in a 60 °C oven overnight to generate the Pd@IPN catalysts. Nanoparticles embedded in crosslinked PVA (denoted as Pd@cPVA) were also prepared from solution A directly without mixing with solution B for comparative purposes. After heating the PVA–Pd gels at 70 °C for 3 h, the Pd@cPVA catalysts were washed with distilled water repeatedly overnight and dried in a 60 °C oven overnight. The catalysts were analyzed under a JEM-1230 transmission electron microscope at an accelerating voltage of 80 kV using the cryo-ultramicrotomy technique. Nanoparticle sizes were determined by analyzing approximately 180–300 nanoparticles from several TEM images and calculating the Feret’s diameters of the particles using ImageJ software (National Institutes of Health, USA). The standard deviations of the Feret’s diameters were calculated to measure the widths of the size distributions. To determine the equilibrium swelling ratio Q, a sample of the Pd-embedded polymers was soaked in water (or DMF) at 100 °C for 1 day to swell to equilibrium and weighed in air after the surface solvent was

removed. The equilibrium swelling ratio and the polymer volume fraction in the swollen state were calculated using the dried and water-swollen (or DMF-swollen) weights measured. 2.3. Catalysis In a typical Heck coupling reaction between iodobenzene and the olefinic compound in DMF solvent, 10 mmol of iodobenzene and 30 mmol of acrylic acid were mixed with 40 mmol of triethylamine and 10 mL of DMF. The insoluble IPN catalyst (0.50 g) containing 0.00299 mmol of Pd were added to the solution, and the reactants were heated in an oil bath at 100 °C for 12 h. After the reaction, the catalysts were filtered, washed with ethanol three times and dried in the air for the next round of use. The filtrate was decolored by activated carbon, and the white product was precipitated out by adding 1 M HCl solution, separated by filtration, washed repeatedly with cold water, and recrystallized in hot water. For a coupling reaction in aqueous solution, 10 mmol of iodobenzene, 15 mmol of styrene, 5.5 ml of triethylamine (4 mmol), and 1.82 g of hexadecyltrimethylammonium bromide (5 mmol) were mixed with 10 ml of water in a 100 ml round-bottom flask. After 0.50 g of the IPN catalyst with 0.00299 mmol of Pd were added into the reactants, the mixture was heated to 100 °C in a silicone oil bath for 10 h. The precipitates containing the product (E)stilbene (trans-diphenyl ethylene) and the catalyst were isolated from the reaction solution by filtration and then washed with 1.0 ml of methanol (or DMF) to dissolve the product and separate it from the catalyst. The methanol (or DMF) solution was then mixed with 100 mL of water to precipitate the product that was then filtered, washed with water repeatedly, and dried to a constant weight of 1.63 g. A similar procedure was used for the catalyzed reaction between iodobenzene and acrylic acid. The yield from the coupling reaction between iodobenzene and styrene was 90.5%, while that from the reaction between iodobenzene and acrylic acid was 95.9%. A melting point (m.p.) of 123.6– 124.2 °C for the product (E)-stilbene (literature 123–125 °C) and a m.p. of 132.6–133.7 °C for the product (E)-cinnamic acid (literature 133–134 °C) were obtained. For a comparative study of unsupported catalysis, 1.0 mL of PdCl2 solution (5.64 mmol/L) was mixed with 10 mmol of iodobenzene, 30 mmol of acrylic acid, 40 mmol of triethylamine, and 10 mL of water (or DMF). The catalytic reaction was carried out at 100 °C in a silicone oil bath for 10 h. The product was precipitated out by adding 1 M HCl solution, separated by filtration, washed repeatedly with cold water, and recrystallized in hot water. 2.4. Instrumentation The reaction products were characterized by 1H NMR and 13C NMR using an AVANCE III 400 MHz NMR spectrometer (Bruker Biospin). FTIR spectra were recorded on a Nexus 360 FTIR spectrometer (Thermo Nicolet). The Pd metal concentration in reaction solutions was analyzed using a Prodigy ICP–AES spectrometer (Teledyne Leeman Labs). The detailed spectroscopic data and their assignments are provided in the Supplementary data. 3. Results and discussion 3.1. Morphology of Pd nanoparticles in IPN networks The Pd nanoparticles in the IPNs (referred to as Pd@IPN) are formed by a two-step process. The particles are first formed in the presence of a stabilizing polymer prior to crosslinking. Subsequent formation of the IPN meshes tightly fastens the Pd particles

759


2936 2872

(C) PVA:PAM=1.652

2959 2876

(D) PVA:PAM=0.826

(E) PVA:PAM=0.413

3000

2000

1500

833

1096 1114 1113 1113

(F) PVA:PAM=0.207

2500

1100

1659 1616 1657 1616 1657 1616 1657 1616 1657 1619

3193

2960

3203

3332

3203

3203

3346 3346

3344

3500

833

2939 2853

(B) PVA:PAM=3.304

1659

2938 2853

(A) cPVA

2960

3338 3201

3352

trate that the Pd particles (shown as black dots in Fig. 2) are well dispersed without aggregates in the IPN supports (the polymers may show multiphase domains in the TEM). As shown in Fig. 2, the majority of the nanoparticles are nonspherical and contain a large number of sharp corners, defects, and edge sites. The surface atoms at these sites are physically unstable and chemically active. The particle sizes measured from the micrographs and sorted into population percentages are shown as histograms in Fig. 3 for Pd@cPVA and three types of Pd@IPNs with different polymer repeating unit ratios. The standard deviations of the particle sizes reflect the widths of the size distributions and are listed in Table 1. The Pd particles in cPVA have an average size of d = 27.6 nm (Feret’s diameter), with a standard deviation of 12.5 nm (counted from 217 particles), whereas those in the IPNs have slightly larger average sizes ranging from 32 to 43 nm (see exact values listed in Table 1). As shown in the histograms in Fig. 3, the largest population of the Pd particles in cPVA appears in the range of 20–30 nm, whereas in the IPNs the largest populations of the Pd particles shift to larger size ranges, corresponding to an increase in the average particle sizes by 18–50%. Therefore, the introduction of a PAM network into cPVA has moderately increased the average particle sizes. The moderate increase in the average particle sizes from the crosslinked PVA to PVA–PAM IPNs are likely caused by Ostwald ripening sintering during the crosslinking of PVA and PAM chains. This phenomenon occurs by the detachment of the Pd atoms from smaller nanoparticles and their subsequent transfer to larger particles [7,35]. As a result, the larger particles grow in size while the smaller particles shrink or dissolve gradually. This occurs because of the lower average coordination numbers of the atoms at the smaller particle surface and the relative ease of their removal; therefore, the large particles become larger at the expense of the smaller particles. Mesh sizes of the crosslinked polymers can be analyzed by the swelling method [32–34,36]. The results from the swelling tests in water were first used to calculate the equilibrium swelling ratio, which is defined as:

1096

as shown in Scheme 1, which depicts an IPN with dispersed metal particles. PAM absorbs a larger amount of water and has a higher thermostability than PVA, whereas the latter is more flexible and has much better film forming properties than the former; thus, we chose these two polymers for forming IPN networks. The hydrophilicity, swelling and mechanical properties of the PVA– PAM IPN supports depend on the PVA/PAM weight ratio and the contents of the crosslinking agents present in the individual networks. During the preparation, the stabilizing PVA was crosslinked by the acetalization of a very small number of PVA hydroxyl (–OH) groups with glutaraldehyde. The non-acetalized –OH groups in PVA remained to serve as the Pd metal binding sites, which were randomly distributed in the PVA. During the simultaneous polymerization of acrylamide (AM), the monomer AM and a small amount of a divinylic crosslinker, N,N0 -methylene bisacrylamide (MBAM), were diffused into the PVA chains after metal reduction. Fig. 1 shows FTIR spectra of the Pd catalysts in the crosslinked PVA alone (cPVA) and five types of IPNs with different contents concentrations of PVA and PAM (labeled as the molar ratios of the vinyl alcohol to acrylamide repeating units VA:AM). The spectra illustrate that a strong band present near 1096 cm1 in Fig. 1A and B, due to the stretching vibration of the acetalized O–C–O group from the cPVA, gradually disappears in the IPNs with increasing content of PAM (see Fig. 1E and F). Meanwhile, a band observed near 1657 cm1 from the amide I stretching vibration from the crosslinked PAM gradually became the strongest band in the IPNs with increasing content of PAM (see Fig. 1E and F). These spectral changes verified the existence of two types of polymers in the catalysts. The PVA colloidal dispersions were stable enough to withstand the potential aggregation or precipitation of the Pd particles during the crosslinking process of PVA and the polymerization of acrylamide. Fig. 2A–D shows transmission electron micrographs of Pd in cPVA and three types of PVA–PAM IPN networks with different polymer repeating unit ratios (VA:AM ratios). Further magnified TEM micrographs are shown in Fig. 2E and F. The TEM images illus-

1000

Wavenumber (cm -1) Fig. 1. FTIR spectra of Pd catalysts in cPVA alone and in five types of IPNs containing different contents of PVA and PAM (labeled as the molar ratios of the VA to AM repeating units).

760


Fig. 2. Transmission electron micrographs of unused Pd particles (A) in cPVA, (B) in the IPN network with VA:AM = 0.826:1, (C) in the IPN network with VA:AM = 1.652:1, and (D) in the IPN network with VA:AM = 0.413:1. Scale bars in (A) to (D) are 500 nm. Higher magnification transmission electron micrographs of unused Pd in the IPN network with VA:AM = 1.652:1 (E) and in the IPN network with VA:AM = 0.413:1 (F). Scale bars in (E) and (F) are 100 nm.

Q¼

1

v 2;s

ð1Þ

where v2,s is the volume fraction of the polymer in the swollen state. A theoretical molecular weight between crosslinks M c can be determined from the nominal crosslinking ratio X and the molecular weight of the repeating unit M r of PVA (M r = 44) or PAM (M r = 71) as follows:

M c ¼ M r =2X

ð2Þ

For neutral homopolymers, experimental values of M c are the same as the theoretical values. The mesh size, n, defines the linear distance between consecutive crosslinks, which indirectly indicates the diffusional space available for solute transport and can be calculated using following equation [32–34,36]:

" n¼v

1=3 2;s

Cn

2M c Mr

!#1=2 l ¼ Q 1=3 ðC n =XÞ1=2 l

ð3Þ

where Cn is the Flory characteristic ratio, and l is the carbon–carbon bond length (0.154 nm). The value of Cn is 8.3 for PVA and 8.5 for PAM in water [37]. By applying Eq. (3), a mesh size n = 3.9 nm in water was obtained for the Pd@cPVA. For the Pd@IPNs in water, mesh sizes n1 = 3.9–4.3 nm were obtained for the PVA network, while mesh sizes n2 = 2.5–2.8 nm were obtained for the PAM network (see Table 1). The mesh size values of the polymers were one order of magnitude smaller than the average sizes of the Pd particles obtained from the TEM measurements. Water is a good solvent for both PVA and PAM, whereas DMF is a poor solvent for PVA and a nonsolvent for PAM; therefore, the swelling ratios of the IPNs in

761


35

(A)

25

Population %

Population %

30

20 15 10 5 0

0

10 20 30 40 50 60 70 80 90 100 110 120

20 18 16 14 12 10 8 6 4 2 0

(B)

0

10 20 30 40 50 60 70 80 90 100 110 120

Size/nm

Size/nm 35

30

(C)

25 20 15 10

20 15 10 5

5 0

(D)

25

Population %

Population %

30

0

0

10 20 30 40 50 60 70 80 90 100 110

0

10 20 30 40 50 60 70 80 90 100 110 120

Size/nm

Size/nm

Fig. 3. Size distributions of unused Pd particles (A) in cPVA, (B) in IPN with VA:AM = 1.625:1, (C) in IPN with VA:AM = 0.826:1, and (D) in IPN with VA:AM = 0.413:1.

Table 1 Pd particle sizes (average particle size d ± standard deviation a), polymer equilibrium swelling ratios Q, mesh sizes n, and catalysis yields from Heck coupling reaction between iodobenzene and acrylic acid a. Catalyst type

PVA IPN #1 IPN #2 IPN #3 a b c

VA:AM ratio

Pure VA 1.652:1 0.826:1 0.413:1

d ± r (nm)

27.6 ± 12.5 42.5 ± 22.0 32.9 ± 10.8 36.8 ± 17.9

Q in water

2.17 2.14 2.36 3.01

n in water (nm) PVA

PAM

3.9 3.9 4.0 4.3

– 2.5 2.6 2.8

Q in DMF

1.83 1.41 1.15 1.12

n in DMF (nm) PVA

PAM

3.7 3.4 3.1 3.1

– 2.1 2.0 2.0

Catalysis yield

b,c

(%)

85 81 86 75

See reaction conditions in Section 2. The yields refer to isolated products based on iodobenzene. All products gave satisfactory 1H NMR, FTIR, and melting points which are provided in Supplementary data at the end of the manuscript.

DMF were reduced appreciably in comparison with their counterparts swollen in water (see Table 1). This led to even smaller mesh sizes of the PVA and PAM in DMF. Thus, the covalent bonding meshes in the IPN networks are small enough to encage the Pd nanoparticles but large enough for reactants to diffuse to the particle surfaces in water and DMF. The accuracy of the mesh sizes obtained from the swelling tests can be further validated by fabrication of two types of Pd@cPVA with Pd particle sizes of d = 4.5 nm ± 1.4 nm and d = 5.8 nm ± 2.4 nm (average Feret’s diameter ± standard deviation). Unfortunately, both types of Pd particles lost their contents significantly in the polymers and leached into their solutions quickly in only two consecutive runs of the catalytic tests. Evidently, this can be ascribed to their particle sizes being so close to the mesh size of cPVA (3.7 nm in DMF) (Table 1). 3.2. Recyclability of palladium nanoparticles in Heck coupling reactions Tight fastening of the Pd metals in the covalent bonding structures of the IPNs imparts robust thermal stability against direct aggregation and suppresses the loss of Pd metals to a negligible level. A comparison of the catalytic yields of the Pd@IPN catalyst (VA:AM = 0.826:1) to that of the Pd@cPVA catalyst obtained after

recycling each catalyst for Heck coupling reactions between iodobenzene and acrylic acid (see Scheme 2, Y = COOH) is shown in Fig. 4 (black and grey bars). Note that the yields refer to real isolated products based on iodoarenes. The Pd@IPN catalyst could be reused in the DMF solvent at least 21 times without a significant change in the reaction yields. This kind of long-term stability of Pd nanoparticles has rarely been referenced in the literature. The Pd nanocatalysts used in the Heck coupling reactions reported thus far could be recycled only about 5–10 times [5–9,14–29] and even fewer times when using more open polymer support structures (e.g., dendrimers or porous resins) with quick Pd leaching [9,11– 17]. Palladium salts supported on crosslinked PAM alone could be reused only 6 times [31a], and similarly, Konjac glucomannan-(a type of natural polysaccharide) supported palladium catalysts could be recycled only five times [38]. Thus, the recycling lifetime of the Pd@IPN system is extraordinary. The methods of immobilizing the Pd metal to polymer supports described in the literature often involve the use of ionic bonding or coordination bonding to phosphines, N-heterocyclic carbenes, palladacycles, or pincer ligands. The catalytic activities in the presence of such ligands are sometimes quite controversial because some of the complexes immobilized in the Pd2+ oxidation state might undergo a Pd2+/Pd4+ catalytic cycle and complicate the mechanism of the coupling reaction [28].

762


I

Y

Pd(0)

+ Y

Y= COOH, Ph, etc. Scheme 2. Heck coupling reaction catalyzed by Pd nanoparticles in IPN and cPVA.

By comparison, the activity of the Pd@cPVA catalyst was reduced appreciably after the 9th repetitive use, delivering a yield of only 22% (see Fig. 4). Because the amount of the Pd metal in the IPN is the same as that in the cPVA, the fact that the Pd@IPN catalyst has a lifetime twice as long as that of the cPVA can be primarily explained by the firmer confinement of the Pd nanoparticles by the covalent chains of PAM with smaller mesh sizes. The amide side chains in the PAM are not believed to be strong ligands for the Pd metals; however, the decreased mesh size of the IPN structure did not affect the reactants ability to reach the Pd particle surfaces. This is also demonstrated by the comparison of the Pd@IPN using different molar ratios of the VA and AM repeating units for the Heck coupling reaction between iodobenzene and acrylic acid. All of the catalysts demonstrated similar high yields of (E)-cinnamic acid after isolation and almost identical activities for the reaction (see Table 1). Obviously, the presence of the PAM network chains does not affect either diffusion of the reactants to the Pd particle surfaces or the yields significantly, even though the nominal crosslinking degree of the PAM network is higher than that of the PVA network. For comparative purposes, catalysis using an unsupported PdCl2 catalyst was also performed. The Pd salt could catalyze the Heck reaction for acrylic acid and iodobenzene with a high yield near 99–100%; however, the isolation and reuse of the Pd salt from the reaction product solutions containing the base (triethylamine) and by-products became very complicated. Therefore, a significant loss of the Pd metals occurred during the work-up, and the reaction solution recovered from the first run of catalysis did not have enough activity for consecutive runs to be performed.

3.3. Catalysis for Heck coupling reactions in aqueous and organic media The second feature of the Pd@IPN described here is its capability to catalyze Heck reaction in neat water without addition of a cosol120%

100%

yield

80%

60%

40%

20%

0%

1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21

recycle times Fig. 4. Comparison of the recyclability of the Pd@IPN catalyst in the coupling reaction of iodobenzene and acrylic acid in DMF solvent (solid black bars) and in water (open bars). Also shown is the recyclability of the Pd@cPVA catalyst in DMF solvent (grey bars) under the same experimental conditions.

vent. As shown in the first entry of Table 2, the coupling reaction in water between iodobenzene and styrene proceeded smoothly in the presence of the Pd@IPN catalyst and a phase transfer agent (hexadecyltrimethylammonium bromide), producing (E)-stilbene at a high yield of 90.5%. For the coupling reaction between iodobenzene and acrylic acid (entry 2, Table 2), the product (E)-cinnamic acid was obtained with a yield of 95.9%. Apparently, the highly polar and hydrophilic nature of the component polymers rendered the reactions functional in water. The ability to use a nontoxic, environmentally benign, water-only solvent is of ecological and economical importance [39,40]. The use of the phase transfer agent promotes the migration of the base from the aqueous phase to the organic reactants to neutralize the protonated Pd2+ complex species generated in the reaction, facilitating the regeneration of the active zerovalent Pd catalyst in the reaction cycle and carrying the resulting salt into the aqueous phase. The recyclability of the Pd@IPN catalyst in water is also displayed in Fig. 4 (open bars). The catalyst can be reused 13 times for the coupling reaction between iodobenzene and acrylic acid in water. This means that the recyclability of the Pd@IPN in the water solvent is very good and comparable to that reported in a recent study, which demonstrated that Pd nanoparticles immobilized on fluorous silica gel could be reused 11–15 times [40a]. Traditionally, Heck coupling reactions are conducted in polar aprotic organic solvents, such as DMF or DMSO, with a Pd(0) complex or a Pd2+ salt (which is reduced in situ to Pd(0) during the catalytic redox cycle), preferably in the presence of a strong ligand (e.g., a toxic ligand phosphine), which stabilizes the Pd(0) atoms as complexes and avoids the fast precipitation of the Pd(0) from the solution. The Pd@IPN system requires no such ligand, as the Pd nanoparticles are prevented from coalescence and are firmly stabilized by the polar IPN structures. The catalytic activity of the Pd@IPN in terms of turnover frequencies (TOFs), which are defined as mol product/mol catalyst/hour, which is calculated from the isolated yield, the amount of Pd used, and the reaction time, are about 50 times larger than those of the recently reported polyelectrolyte complex stabilized palladium nanoparticles [31b]; however, the Pd@IPN has much longer recycling performance than the latter. For applications in aqueous conditions, it is quite reasonable that several types of hydrophilic polymers other than PVA and PAM can be used to form different IPN structures and examine the polymer effect on the catalytic activity. This will form a large family of IPNs suitable for support materials. On the other hand, the choice of the polymers also depends on the optimal condition of the Heck reaction; therefore, not all hydrophilic polymers can serve as components of the IPN for the Heck coupling reaction. Non-ionizable hydrophilic polymers have an advantage that the pH of the Heck reaction can be easily controlled by an external base added; therefore, PVA and PAM are suitable materials for the Heck coupling reactions, as demonstrated in this study. The selectivity of the catalysts was also examined for a series of Heck coupling reactions in DMF solvents by using derivatives of iodoarenes and acrylic acid or styrene. The results are listed in Table 2. In all cases, the products were identified to be either exclusively an (E)-cinnamic acid derivative or an (E)-stilbene. Other types of isomeric products were not detectable by HPLC or 1H NMR analysis. As shown in Table 2, the yields of the isolated

763

K. Zhan et al. / Reactive & Functional Polymers 71 (2011) 756–765 Table 2 Results of various Heck coupling reactions catalyzed by Pd@cPVA and Pd@IPN. Entrya

1 2 3 4 5 6 7 8 9 10 a b c d

Reactants Aryl halide (mmol)

Alkene (mmol)

Iodobenzene (10) Iodobenzene (10) Bromobenzene (10) 4-Iodoanisole (10) p-Fluoroiodobenzene (4.5) 2-Iodothiophene (10) 4,40 -Diiododiphenyl (5) 4-Iodoanisole (10) 4-Iodoanisole (10) 4-Iodoanisole (10)

Styrene (15) Acrylic acid (20) Acrylic acid (20) Acrylic acid (20) Acrylic acid (20) Acrylic acid (20) Methyl acrylate (20) Styrene (20) Methyl acrylate (20) Acrylic acid (20)

Catalyst

Solvent

Yield (%)

IPNb IPNb IPNb IPNb IPNb IPNb IPNb cPVA cPVA cPVA

Water Water Water DMF DMF DMF DMF DMF DMF DMF

91 96 35 92 63 41 69 69 85 80

c,d

TOF

302 321 98 256 78 114 192 192 237 223

See Section 2 for reaction conditions. The IPN used contain VA:AM = 0.826:1. The yields refer to isolated products based on Ar–X. All products gave satisfactory 1H NMR, 13C NMR, FTIR spectra, and melting points which are provided in the Supplementary data at the end of the manuscript.

3.4. The role of the Pd nanoparticles in the coupling reaction

products from the Pd@IPN catalyst range from 92.0% (for a phenyl ring with a methoxy-substituent) to 68.8% (for a double coupling reaction) to 41% (for a relatively difficult substrate with a thiophene ring). The yields were primarily dependent on the nature of the substituent at the para-position of the ring and the substituent on the olefin. The catalytic activities are higher than those previously reported for Pd nanoparticles encapsulated in several other types of polymers [12,31b,41]. Furthermore, for the coupling reaction between bromobenzene and acrylic acid (entry 3, Table 2), the TOF (98) and yield (35%) are also much higher than the recently reported TOF and yield given by charged polymer complex stabilized Pd nanoparticles (only about 0.25 and 5%, respectively) [31b].

Fig. 5A and B shows a representative TEM image of the nanoparticles in the IPN with VA:AM = 0.826:1 after 21 cycles of catalysis in DMF solvent and the corresponding size distribution of the nanoparticles. At this stage, the Pd@IPN still showed very significant catalytic activity; however, statistical analysis of 192 particles from the TEM images showed that the average size of the particles increased to 49.7 nm while their standard deviation also increased appreciably to 33.9 nm as compared to the average particle size of 32.9 nm with a standard deviation of 10.8 nm before catalysis. If the size increase was simply caused by the leaching of small parti-

30

(B) (B)

Population %

25 20 15 10 5 0

0

10 20 30 40 50 60 70 80 90 100 110 120 130 140

Size/nm

Fig. 5. (A) Transmission electron micrograph (scale bar = 500 nm) and (B) size distribution of the used Pd particles in the IPN with VA:AM = 0.826:1. Higher magnification transmission electron micrographs of the used Pd in the IPN (C and D) (scale bars = 100 nm).

764


cles into the reaction solutions and the retention of the large particles in the IPN, a narrower size distribution would be expected. The increase in the average particle diameter and the observation of a wider distribution of the particles after the reactions, particularly concerning the appearance of particles larger than 100 nm (Fig. 5B as compared to Fig. 3C), is best explained by the Ostwald ripening effect, which involves the process of Pd dissolution from smaller particles and re-deposition to larger particle surfaces during the catalytic reaction [7,35]. In addition, a large number of the used particles formed well-defined polyhedron shapes with straight edges, as shown in the higher magnification TEM images (Fig. 5C and D). This further illustrates that the active Pd species were re-deposited during the reaction, forming the polyhedron particles. The appearance of a large population (about 27%) of the particles in the size range of 10–20 nm in Fig. 5B could be due to new particles generated in the catalytic reaction. It is reasonable to conclude that the detached Pd atoms generated in the Ostwald ripening process were involved in the catalytic redox cycle, reacted with the iodoarenes, and then re-deposited on larger particles which ultimately led to the increase in the average particle size and the distribution width, as well as the appearance of welldefined particle shapes. In other words, the nanoparticles play a catalytic role in the Heck reaction through a homogeneous mechanism involving the participation of dissolved Pd atoms, which is in accordance with recent studies [42–44]. Analysis of the Pd metal concentration of the final reaction solutions by the ICP–AES method revealed extremely low amounts of palladium leaching from the IPNs. Only 0.0506 and 0.140 lmol of Pd were found in the DMF solution phase after the first and 23rd cycles of the reaction, respectively. These corresponded to only 0.177% and 0.488% of the total amount of Pd used in the catalytic cycle. Such a level of leaching is one order of magnitude lower than those ‘‘leach-proof’’ Pd nanoparticles reported very recently for C–C coupling reactions [21a,24] and should not significantly contribute to the increase in the average particle size and distribution width of the Pd in the IPN. In the aqueous reaction, only 0.16 and 0.125 lmol of Pd were found in the solution phase after the first and 13th cycle of the reaction, respectively, which corresponded to only 1.25% and 0.98% of the total amount of Pd used in the catalytic cycle. This indicates that the Pd leaching in the water solution reactions is slightly higher than in the DMF solvent condition. In addition, we found that the trace amount of soluble Pd ions and/or atoms in the solution phase could not catalyze the reaction effectively after the Pd@IPN was removed from the reaction system and all the new reagents were added. It is reasonable to conclude that the majority of the soluble Pd ion and/or atoms, once generated inside the IPN, are re-deposited to the large nanoparticle surfaces in the reduction step of the catalytic cycle or that they precipitate quickly inside the IPN pores and form new particles. The re-deposition of palladium active species on the surfaces of support materials and palladium particles was reported in some studies [45]. It seems that the very fast and dynamic Pd metal dissolution–deposition process driven by the particle ripening effect inside the IPN pores has the remarkable feature of minimizing the amount of Pd leached into the solution phase. This means that the catalysis is essentially confined inside the IPN pores and the Pd@IPN catalysts can be recovered as a ‘‘heterogeneous’’ reactor.

4. Conclusions There is a need to improve the stability of the Pd nanoparticle in cross-coupling reactions because either leaching or severe aggregation of the nanoparticles may lead to quick deactivation [46]. One issue in the study of these Pd catalysis reactions has been establishing a reasonable balance between recyclability (like a

heterogeneous catalysis) and activity (like a classical homogeneous catalysis). We have shown here that palladium nanoparticles in properly designed hydrophilic IPN structures can achieve the goal of superior durability and can be used for Heck coupling reactions not only in organic solvents but also in neat water. The reaction involves a homogeneous mechanism and also bears the unique advantages of a heterogeneous catalysis; that is, the catalysts can be recovered easily by simple filtration and regenerated to full activity. As a result, the highly stable catalysts show negligible metal leaching and can be reused more than 20 times in DMF and 13 times in water. Such robust character bridges the gap between homogeneous and heterogeneous catalysis and combines the advantages of both types of reactions, which is quite valuable for reducing the cost of the use of noble metals, minimizing the contamination of the metals in the products, and simplifying the experimental process. The results reported in this study demonstrate an improvement in the use of Pd nanoparticles for C–C cross-coupling reactions when compared with Pd nanoparticles dispersed in other hydrophilic structures, e.g., poly(N-isopropylacrylamide)-grafted Pd nanoparticles [47], polymethacrylic acid copolymer bead-supported Pd nanoparticles [48], polyion complexes composed of polyacrylic acid and poly[4-chloromethylstyrene-co-(4-vinylbenzyl)tributylammoniumchloride] [31b], and chitosan-supported Pd particles [41a] in which low activities and a quick decrease in the catalytic activity was observed in the recycling of the catalyst. The combined benefits of the ability to use water as solvent for the coupling reactions and the increased recyclability indicate that the Pd-encaged IPNs are green sustainable catalyst reactors and are suitable for applications requiring continuous processing conditions. Acknowledgements This work was supported by the National Natural Science Foundation of China (No. 20974063), Natural Science Foundation of Zhejiang Province (No. Y4090504), State Key Laboratory of Coordination Chemistry (Nanjing University), Department of Education, Zhejiang Province (No. 20070495), Department of Science and Technology of Zhejiang Province (No. 2009R10040), Zhejiang Xinmiao Project, and Shaoxing University (No. 08LG1005). Appendix A. Supplementary material Supplementary information including NMR and FTIR spectroscopic data of the reaction products and their assignments, can be found in the online version of this manuscript. Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.reactfunctpolym.2011.04.007. References [1] L.D. Rampino, F.F. Nord, J. Am. Chem. Soc. 63 (1941) 2745–2749. [2] H. Hirai, J. Macromol. Sci. Part A 13 (1979) 633–649. [3] B. Corain, G. Schmid, N. Toshima, Metal Nanoclusters in Catalysis and Materials Science, Elsevier, Amsterdam, 2008. [4] N. Toshima, Y. Shiraishi, T. Teranishi, M. Miyake, T. Tominaga, H. Watanabe, W. Brijoux, H. Bonnemann, G. Schmid, Appl. Organometal. Chem. 15 (2001) 178– 196. [5] (a) D. Astruc, Tetrahedron: Asymmetry 21 (2010) 1041–1054; (b) D. Astruc, Nanoparticles and Catalysis, Wiley-VCH Verlag GmbH & Co., Weinheim, 2008, p. 1.; (c) D. Astruc, Inorg. Chem. 46 (2007) 1884–1894. [6] D. Astruc, F. Lu, J.R. Aranzaes, Angew. Chem. Int. Ed. 44 (2005) 7852–7872. [7] (a) P. Li, L. Wang, H. Li, Tetrahedron 61 (2005) 8633–8640; (b) Y. Li, E. Boone, M.A. El-Sayed, Langmuir 18 (2002) 4921–4925; (c) R. Narayanan, M.A. El-Sayed, J. Am. Chem. Soc. 125 (2003) 8340–8347. [8] (a) B. Corain, M. Kralik, J. Mol. Catal. A: Chem. 159 (2000) 153–162; (b) P. Centomo, M. Zecca, M. Kralik, D. Gasparovicova, K. Jerabek, P. Canton, B. Corain, J. Mol. Catal. A: Chem. 300 (2009) 48–58.

K. Zhan et al. / Reactive & Functional Polymers 71 (2011) 756–765 [9] A.M. Caporusso, P. Innocenti, L.A. Aronica, G. Vitulli, R. Gallina, A. Biffis, M. Zecca, B. Corain, J. Catal. 234 (2005) 1–7. [10] (a) Y.M.A. Yamada, Y. Uozumi, Tetrahedron 63 (2007) 8492–8498; (b) T. Teranishi, N. Toshima, J. Chem. Soc. Dalton Trans. (1994) 2967–2974. [11] R.M. Crooks, M. Zhao, L. Sun, V. Chechik, L.K. Yeung, Acc. Chem. Res. 34 (2001) 181–190. [12] L.K. Yeung, R.M. Crooks, Nano Lett. 1 (2001) 14–17. [13] (a) E.H. Rahim, F.S. Kamounah, J. Frederiksen, J.B. Christensen, Nano Lett. 1 (2001) 499–501; (b) K.R. Gopidas, J.K. Whitesell, M.A. Fox, Nano Lett. 3 (2003) 1757–1760. [14] M.A.R. Meier, M. Filali, J.-F.Gohy, U.S. Schubert, J. Mater. Chem. 16 (2006) 3001–3006. [15] R. Sablong, U. Schlotterbeck, D. Vogt, S. Mecking, Adv. Syn. Catal. 345 (2003) 333–336. [16] K. Okamoto, R. Akiyama, H. Yoshida, T. Yoshida, S. Kobayashi, J. Am. Chem. Soc. 127 (2005) 2125–2135. [17] (a) S. Pathak, M.T. Greci, R.C. Kwong, K. Mercado, G.K.S. Prakash, G.A. Olah, M.E. Thompson, Chem. Mater. 12 (2000) 1985–1989; (b) R. Abu-Reziq, D. Wang, M. Post, H. Alper, Chem. Mater. 20 (2008) 2544– 2550. [18] J.H. Ding, D.L. Gin, Chem. Mater. 12 (2000) 22–28. [19] R. Narayanan, M.A. El-Sayed, J. Catal. 234 (2005) 348–355. [20] C. Evangelisti, N. Panziera, A. D’Alessio, L. Bertinetti, M. Botavina, G. Vitulli, J. Catal. 272 (2010) 246–252. [21] (a) N. Ren, Y.-H. Yang, Y.-H. Zhang, Q.-R. Wang, Y. Tang, J. Catal. 246 (2007) 215–222; (b) Y. Wan, H. Wang, Q. Zhao, M. Klingstedt, O. Terasaki, D. Zhao, J. Am. Chem. Soc. 131 (2009) 4500–4541. [22] C.M. Crudden, M. Sateesh, R. Lewis, J. Am. Chem. Soc. 127 (2005) 10045– 10050. [23] A. Gniewek, J.J. Ziolkowski, A.M. Trzeciak, M. Zawadzki, H. Grabowska, J. Wrzyszcz, J. Catal. 254 (2008) 121–130. [24] U. Laska, C.G. Frost, G.J. Price, P.K. Plucinski, J. Catal. 268 (2009) 318–328. [25] K. Yu, W. Sommer, M. Weck, C.W. Jones, J. Catal. 226 (2004) 101–110. [26] J.-N. Young, T.-C. Chang, S.-C. Tsai, L. Yang, S.J. Yu, J. Catal. 272 (2010) 253–261. [27] M.I. Burguete, E. Garcia-Verdugo, I. Garcia-Villar, F. Gelat, P. Licence, S.V. Luis, V. Sans, J. Catal. 269 (2010) 150–160. [28] J.D. Webb, S. MacQuarrie, K. McEleney, C.M. Crudden, J. Catal. 252 (2007) 97– 109. [29] (a) I. Pryjomska-Ray, A. Gniewek, A.M. Trzeciak, J.J. Ziolkowski, W. Tylus, Top. Catal. 40 (2006) 173–184; (b) A.N. Pustam, S.D. Alexandratos, React. Funct. Polym. 70 (2010) 545–554; (c) D. Duraczynska, A. Drelinkiewicz, E.M. Serwicka, D. Rutkowska-Zbik, E. Bielanska, R. Socha, A. Bukowska, W. Bukowski, React. Funct. Polym. 70 (2010) 382–391; (d) S.E. Lyubimov, A.A. Vasil’ev, A.A. Korlyukov, M.M. Ilyin, S.A. Pisarev, V.V. Matveev, A.E. Chalykh, S.G. Zlotin, V.A. Davankov, React. Funct. Polym. 69 (2009) 755–758; (e) J. Lee, S. Bernard, X.-C. Liu, React. Funct. Polym. 69 (2009) 650–654; (f) L. Calore, G. Cavinato, P. Canton, L. Peruzzo, L. Tauro, B. Corain, React. Funct. Polym. 70 (2010) 639–646; (g) S. Song, S. Shen, X. Cui, D. Yao, D. Hu, React. Funct. Polym. 71 (2011) 512– 519;

[30]

[31]

[32] [33] [34] [35] [36] [37] [38] [39]

[40]

[41]

[42]

[43] [44] [45] [46] [47] [48]

765

(h) A. Drelinkiewicz, A. Zie˛ba, J.W. Sobczak, M. Bonarowska, Z. Karpinski, A. Waksmundzka-Gora, J. Stejskal, React. Funct. Polym. 69 (2009) 630– 642. (a) L.H. Sperling, in: J.C. Salamone (Ed.), Polymeric Materials Encyclopedia, vol. 5, CRC Press, Boca Raton, FL, 1996, p. 3292; (b) L.H. Sperling, V. Mishra, Polym. Adv. Technol. 7 (1996) 197–208; (c) L.H Sperling, in: D. Klempner, L.H. Sperling, L.A. Utracki (Eds.), Interpenetrating Polymer Networks, American Chemical Society, Washington, DC, 1994, p. 3. (a) H. Mahdavi, A. Zirakzadeh, J. Amani, React. Funct. Polym. 67 (2007) 716– 722; (b) A. Ohtaka, Y. Tamaki, Y. Igawa, K. Egami, O. Shimomura, R. Nomura, Tetrahedron 66 (2010) 5642–5646. N.A. Peppas, S.L. Wright, Macromolecules 29 (1996) 8798–8804. N.A. Peppas, S.L. Wright, Eur. J. Pharm. Biopharm. 46 (1998) 15–29. A.S. Hickey, N.A. Peppas, J. Membrane Sci. 107 (1995) 229–237. R. Narayanan, M.A. El-Sayed, J. Phys. Chem. B. 108 (2004) 8572–8580. J. Berger, M. Reist, J.M. Mayer, O. Felt, N.A. Peppas, R. Gurny, Eur. J. Pharm. Biopharm. 57 (2004) 19–34. M. Bohdanecky, V. Petrus, B. Sedlaek, Makromol. Chem. 184 (1983) 2061– 2073. P. Liu, Y. Yang, Y. Liu, X. Wang, React. Funct. Polym. 68 (2008) 384–388. (a) S. Sawoo, D. Srimani, P. Dutta, R. Lahiri, A. Sarkar, Tetrahedron 65 (2009) 4367–4374; (b) J. Gil-Molto, S. Karlstrom, C. Najera, Tetrahedron 61 (2005) 12168–12176; (c) M.S. Anson, A.R. Mirza, L. Tonks, J.M.J. Williams, Tetrahedron Lett. 40 (1999) 7147–7150; (d) L. Botella, C. Najera, Tetrahedron Lett. 45 (2004) 1833–1836; (e) H. Hagiwara, Y. Sugawara, T. Hoshi, T. Suzuki, Chem. Commun. (2005) 2942–2944. (a) M. Lamblin, L. Nassar-Hardy, J.-C. Hierso, E. Fouquet, F.-X. Felpin, Adv. Syn. Catal. 352 (2010) 33–79; (b) A. Boffi, S. Cacchi, P. Ceci, R. Cirilli, G. Fabrizi, A. Prastaro, S. Niembro, A. Shafir, A. Vallribera, ChemCatChem 3 (2011) 347–353. (a) C.M. Park, M.S. Kwon, J. Park, Synthesis (2006) 3790–3794; (b) J.J.E. Hardy, S. Hubert, D.J. Macquarrie, A.J. Wilson, Green Chem. 6 (2004) 53–56. (a) M.B. Thathagar, J.E. ten Elshof, G. Rothenberg, Angew. Chem. Int. Ed. 45 (2006) 2886–2890; (b) A.V. Gaikwad, A. Holuigue, M.B. Thathagar, J.E. ten Elshof, G. Rothenberg, Chem. Eur. J. 13 (2007) 6908–6913. C.C. Cassol, A.P. Umpierre, G. Machado, S.I. Wolke, J. Dupont, J. Am. Chem. Soc. 127 (2005) 3298–3299. S.J. Broadwater, D.T. McQuade, J. Org. Chem. 71 (2006) 2131–2134. F. Zhao, M. Shirai, Y. Ikushima, M. Arai, J. Mol. Catal. A: Chem. 180 (2002) 211– 219. C.E. Chan-Thaw, A. Villa, P. Katekomol, D. Su, A. Thomas, L. Prati, Nano Lett. 10 (2010) 537–541. G. Wei, W. Zhang, F. Wen, Y. Wang, M. Zhang, J. Phys. Chem. C 112 (2008) 10827–10832. P. Zheng, W. Zhang, J. Catal. 250 (2007) 324–330.