Reverse micellar based synthesis of ultrafine MgO nanoparticles (8–10 nm): Characterization and catalytic properties

Journal of Colloid and Interface Science 353 (2011) 137–142

Contents lists available at ScienceDirect

Journal of Colloid and Interface Science www.elsevier.com/locate/jcis

Reverse micellar based synthesis of ultrafine MgO nanoparticles (8–10 nm): Characterization and catalytic properties Aparna Ganguly a,c, Phong Trinh b, K.V. Ramanujachary b, Tokeer Ahmad c, Amos Mugweru b,⇑, Ashok K. Ganguli a,⇑⇑ a b c

Department of Chemistry, Indian Institute of Technology, New Delhi 110 016, India Department of Chemistry and Biochemistry, Rowan University, Glasboro, NJ 08028, USA Department of Chemistry, Jamia Millia Islamia, New Delhi 110 025, India

a r t i c l e

i n f o

Article history: Received 7 July 2010 Accepted 14 September 2010 Available online 19 September 2010 Keywords: Magnesium oxide Reverse micellar synthesis Surface area Catalytic properties

a b s t r a c t Anisotropic nanostructures of magnesium oxalate dihydrate were synthesized using cationic surfactant based microemulsion method. The cationic surfactant plays an important role in forming the anisotropic structures. The oxalate nanostructures acts as an excellent precursor for the synthesis of fine magnesium oxide nanoparticles (10 nm). Both the precursor and the oxide were characterized by using PXRD, IR, surface area and HRTEM. The surface area of these surfactant free oxide nanoparticles was found to be 108 m2/g. The catalytic activity of this basic oxide was examined for the Claisen–Schmidt condensation reaction and was found to be comparable to the best reported for the conventionally prepared MgO. Chalcone formation was found to increase with time as observed using gas chromatography–mass spectrometry (GC–MS). The reusability of the catalyst was checked by using the same catalyst twice which showed a reduced percentage (50% compared to first cycle) conversion. Ó 2010 Elsevier Inc. All rights reserved.

1. Introduction Nanocrystalline oxides are of immense interest to scientists due to their enriched surface chemistry and high surface area. Ultrafine nanoparticles of alkaline earth metal oxides, namely MgO and CaO have been extensively used as catalysts [1] and is also important for the pharmaceutical industry, waste remediation, paint, refractory, removal of toxic gases [2] owing to the defect structure and high surface area. Conventional method for MgO synthesis is via the thermal decomposition of various magnesium salts [3,4]. However, the resulting MgO particles inevitably possess relatively large and non-uniform particle sizes and low specific surface area, which is not preferable for the aforementioned applications. For instance, in the synthesis of MgO nanoparticles using the polyol method the particles exhibit a wide size distribution from 20 to 100 nm [5]. The other method reported employs expensive Mg precursor and involve multiple steps to get high quality nanosheets [6]. Thus the synthesis of pure nanocrystalline and uniform particles of MgO itself is of great significance. There have been efforts to synthesize MgO nanoparticles with small particle size and high surface area. There have also been reports in the literature on ⇑ Corresponding author. Fax: + 856 256 4478. ⇑⇑ Corresponding author. Fax: + 91 11 26854715. E-mail addresses: [email protected] (A. Mugweru), [email protected]. ernet.in (A.K. Ganguli). 0021-9797/$ - see front matter Ó 2010 Elsevier Inc. All rights reserved. doi:10.1016/j.jcis.2010.09.041

synthesis of MgO nanoparticles using surfactants. Nanoplates of 10–20 nm thickness and 100 nm length were synthesized in presence of non-ionic PEG-400 [7]. Non-ionic surfactants like Brij-56 and TX-100 gave particles with low surface area of 22 m2/g and 65 m2/g with crystallite size 16 and 18 nm respectively [8]. However in all these syntheses the precursor along with the surfactant is heated to get the oxide which probably leads to the decrease in surface area. Instead we resort to a reverse micellar method where the surfactant aggregate acts as a removable nanostructured template for the formation of the anisotropic precursor and is completely washed off to obtain the product. Reverse micelles (microemulsions) have been used to synthesize a variety of materials with varied shape and size [9–13]. The size of the reverse micelles (aqueous droplets), dispersed homogeneously through out the microemulsion are in nanoregime and thus these are used as nanoreactors to synthesize nanomaterials. The morphology of the product may be varied through proper choice of the surfactant and a number of parameters which include concentration of water, surfactant, and nature of non-polar phase (oil). Variation of any of these parameters changes the size and shape of the aggregates formed depending on which the morphology of the oxide nanoparticle can be designed. Applications related to alkaline earth oxides such as catalysis is considered to be highly size dependent as a small variation in particle size as well as size distribution can result in a change in its activity. Thus by judiciously choosing the surfactants and the Wo parameter of surfactant aggregates (like micelles

138

A. Ganguly et al. / Journal of Colloid and Interface Science 353 (2011) 137–142

or reverse micelles) we can efficiently control the interfaces, the stability of the colloidal dispersions and the particle size [9]. The understanding of interfaces and their stability has evolved tremendously due to the rapid growth of nanoscience and nanotechnology. The reverse micellar approach also allows us to maintain the chemical homogeneity along with monodispersity of the particles thus obtained. To the best of our knowledge there have been no reports on the microemulsion – based synthesis of MgO nanoparticles, its structure and catalytic property. MgO nanoparticles are widely used as a heterogeneous catalyst for a number of organic reactions because of the presence of the reactive sites on the surface viz Mg2+ site acting as a Lewis acid, O2 site as a Lewis base, lattice bound hydroxyl groups isolated hydroxyls and cationic and anionic vacancies [6]. Asymmetric Henry and Michael reaction [14], synthesis of organic carbonates [15], aldol condensation [16], Strecker reactions [17], Meerwen Pondorf Verley reduction [1], etc. are some of the well studied reactions catalyzed by MgO nanoparticles. A heterogeneous catalyst is favored to perform the condensation reactions over a homogeneous catalyst because of the ease of handling, simple work up and regenerability of the heterogeneous catalyst. We have also chosen one such condensation reaction as the model reaction for studying the catalytic activity of the nanooxide synthesized for the formation of chalcones. Chalcones are important intermediates for the essential flavonoids which along with carotene are responsible for the coloring in flowers, vegetables and herbs. They are also important owing to their antibacterial [18], antifungal, anticancer [19] and anti-inflammatory [20] properties. The following reaction demonstrates the condensation of benzaldehyde and acetophenone to form 1,3-diphenylprop-2-en1-one via the Claisen–Schmidt condensation.

and ethyl acetate (PHARCO AAPER) were used. The chalcone obtained from TCI America was used for comparison. The magnesium oxalate nanostructures were prepared by using a reverse micellar mediated route. The microemulsions were prepared using 39.6420 g of CTAB, 40.4 ml of 1-butanol, and 200 ml of isooctane in a conical flask. Then 24 ml of 0.1 M magnesium nitrate solution was added dropwise while stirring to obtain a colorless solution. Another microemulsion was prepared using an aqueous solution of ammonium oxalate, keeping the other constituents identical. The two microemulsions were mixed while stirring and kept for 24 h. The solution became cloudy and a white precipitate was obtained. The product was centrifuged and washed (thrice) with a mixture of chloroform:methanol (1:1). The oxalate product was dried overnight at 110 °C and heated at 600 °C for 6 h to obtain the MgO nanoparticles. 2.2. Characterization of material Both the precursor and catalyst were characterized by XRD, TEM and nitrogen adsorption studies. XRD patterns were obtained by using X2 Advanced Diffraction System (Scintag Inc., USA). The nitrogen adsorption isotherms and the BET surface area were recorded by using Quantachrome NOVA 2200e. The IR spectrum of the catalyst was investigated after each reaction by using DIGILAB FTS 7000 spectrophotometer. TEM studies were carried out using FEI Technai G2 20 electron microscope operated at 200 kV. HRTEM and electron diffraction studies were carried out on a JEOL 3011, 300 kV instrument. FESEM was done on a JEOL, JSM-6700F instrument. The sample was prepared by loading an ethanolic dispersion of the sample onto a copper tape followed by a Pt coating. 2.3. Claisen–Schmidt condensation reaction

O

O

O +

Base Solvent

This study was carried out with an aim to synthesize nanoparticles of MgO (size <15 nm) with high surface area using a microemulsion based method which can be used as an efficient heterogeneous catalyst. The process is simple and less-expensive than the other reported methods. In this study, we show that monodisperse and uniform MgO nanoparticles of size 8–10 nm with a surface area 108 m2/g can be synthesized by the thermal decomposition of magnesium oxalate nanorods synthesized using the reverse micellar route. We find that these nanoparticles can be used as a heterogeneous catalyst for the formation of chalcone using the Claisen–Schmidt condensation. So far the reports on the catalytic activity of MgO towards Claisen–Schmidt have been studied using the commercially procured samples without the structural details. Our study provides an important insight to the synthesis and detailed characterization of ultrafine nanostructured materials (MgO) using the reverse micellar route. 2. Experimental 2.1. Materials and methods Hexadecyl (trimethyl) azanium bromide (cetyltrimethylammonium bromide, CTAB) (Amresco), 1-butanol and phenyl methanal (Fisher Scientific), 2,2,4 trimethyl pentane (Burdick and Jackson laboratories Inc.), magnesium nitrate (Alfa Aesar), ammonium oxalate (Mallinckrodt), 1-phenyl ethanone (Acros), methyl benzene

Ten milliliter toluene, 0.3 ml acetophenone and 0.2 ml benzaldehyde were injected into a 25 ml round bottom flask. The mixture was heated and stirred to the reaction temperature (160 °C) in argon atmosphere. When the reaction mixture started to boil, 0.4 ml of the mixture was diluted by 1 ml of ethyl acetate. 143 mg of the catalyst (MgO) was then added into the flask. Aliquots were taken out periodically to determine the change in concentration of reactants and products with time by using the Agilent 6890 N gas chromatograph and 5973 N mass selective detector which is equipped with Agilent 19091S-433 column GC–MS. A carrier flow of 0.9 ml/min was used, while the temperature was raised from 40 °C to 270 °C at the rate of 15 °C/min. Post reaction the catalyst was regenerated by heating the sample in flowing nitrogen. A blank experiment (i.e. in absence of nanoparticles) has been also carried out in order to compare the catalytic activity of the nanoparticles. 3. Results and discussion 3.1. Characterization of material The X-ray diffraction pattern of the oxalate precursor obtained after overnight heating at 110 °C confirms the formation of pure phase monoclinic magnesium oxalate dihydrate (JCPDS 28-625). Fig. 1 shows the cuboidal (rod shaped) oxalate precursor with an average length of 2.5 lm and diameter of 150–750 nm. It is to be noted that reverse micellar synthesis of transition metal oxalates with +2 oxidation state of the metal ion results in the formation of nanorods [11]. This can be explained by the presence of a negative surface charge over the oxalate precursor followed by an assembly of cationic surfactants over the surface leading to the restricted growth and hence the formation of anisotropic structures. The detailed mechanism of the formation of rods using

139

A. Ganguly et al. / Journal of Colloid and Interface Science 353 (2011) 137–142

a

b Fig. 1. TEM image of magnesium oxalate dihydrate.

cationic surfactant has been discussed in an earlier report [11]. The same mechanism is also expected to dominate the morphology of magnesium oxalate rods where we are using a divalent main group element instead of a divalent transition metal. Thermogravimetric studies (TG/DTA) were carried out on the oxalate precursor which showed two weight losses, one at 180 °C corresponding to the loss of water and the second at 500 °C for the decomposition of oxalate to the oxide. Based on the TG/DTA analysis, the oxalate precursor was heated at 600 °C for 6 h to obtain the oxide catalyst. The XRD pattern (Fig. 2) was indexed on the basis of cubic magnesium oxide (JCPDS 45-946). TEM shows the formation of particles with size 8–10 nm (Fig. 3a). HRTEM image shows the lattice fringes of the oxide corresponding to (2 0 0) plane (Fig. 3b).The selected area electron diffraction (SAED) pattern confirms the nanocrystalline nature of the sample (Fig. 3c) and show reflections corresponding to (2 0 0), (2 2 0) and (2 2 2) planes, supporting the presence of MgO in cubic phase. The crystallite size as obtained from the Scherrers’ formula was 11 ± 0.5 nm and matches well with the size obtained from the TEM image. The FESEM (Supplementary material Fig. S1) image shows highly agglomerated particles which obscure the details of the surface sites. The surface area of the oxide nanoparticle was found to be 108 m2/g by using nitrogen adsorption–desorption method. This is comparatively much higher relative to MgO particles synthesized in presence of surfactant by earlier workers [8].

(200)

2nm

c

(200)

200

(220) 500

220

5 1/nm

Fig. 3. (a) TEM, (b) HRTEM image and (c) electron diffraction pattern of MgO nanoparticles.

111

INTENSITY (A.U.)

(222)

0 20

30

40

50

60

70

3.2. Catalytic activity

2 Theta Fig. 2. XRD pattern of MgO catalyst after heating at 600 °C for 6 h.

The Claisen–Schmidt condensation is a base-catalyzed reaction where sodium hydroxide is generally used to catalyze the

A. Ganguly et al. / Journal of Colloid and Interface Science 353 (2011) 137–142

condensation reaction. In this study we have used acetophenone and benzaldehyde as the reactants and MgO nanoparticles as the catalyst. The progress of the reaction i.e. the formation of chalcone was monitored with the help of a GC–MS. A calibration curve of pure chalcone as a standard was first obtained by plotting the intensity of pure chalcone (standard) versus the concentration of chalcone. A linear calibration curve (R = 0.998) can be seen in the supporting information (Fig. S2). Fig. 4 shows a plot of the percentage of chalcone formation with time for both fresh and used catalyst (MgO). We observe 58% of chalcone formation in about 15 h. This percentage conversion decreases by about 50% after the first round of the reaction. This indicates that the catalyst becomes less effective with successive reactions. The surface area of fresh catalyst was found to be 108 m2/g and it decreased substantially to 42 m2/g post reaction (Table 1). It is important to note that the reactivity or catalytic activity of the MgO surface is dependent on four important parameters such as exposed crystal face, coordination of surface ions, electronic structure and surface defects. The catalytic activity of the oxide nanoparticle is reduced on reuse because of the blocking of the active O2 sites by the organic reactants. As a result the basicity gets reduced and this strongly affects the catalytic activity. The amount of benzaldehyde and acetophenone in the reaction mixture also decreased with time (Fig. 5) as expected during the chalcone formation. No product formation could be seen for the blank experiment (i.e. in absence of nanoparticles) even after 10 h of reaction time (Fig. S3). Initially the reactants are not taken in 1:1 ratio so as to avoid the formation of the by-product benzoic acid formed due to the aerial oxidation of benzaldehyde. Conventionally the Claisen–Schmidt reaction has been investigated in the presence of aqueous or alcoholic KOH as the catalyst. In toluene/aqueous biphasic conditions with 10% aqueous KOH, a conversion percentage of only 15% was observed after 24 h which increased to 98% in 2 h with 50% aqueous KOH [21]. Separation of

Conversion of Benzaldehyde (%)

60

Fresh catalyst

50

0.003

Concentration (mol/L)

140

Acetophenone

0.0025

0.002

Benzaldehyde

0.0015

0.001

2

4

6

8

10

12

14

16

Time (hours) Fig. 5. Profile of acetophenone and benzaldehyde with time during the Claisen–Schmidt reaction.

reactants from the catalyst becomes cumbersome for a homogeneous catalyst. Thus heterogenous catalysts are preferred for their easy work up and regenerability. When aerogel synthesized MgO is used as a catalyst (crystallite size 4 nm, with SSA: 590 m2/g), 98% conversion of benzaldehyde [22] was observed after the reaction was allowed to proceed for 12 h. For the conventionally prepared (C.P.) MgO [22] with (SSA: 250 m2/g), the percentage conversion after 15 h of reaction was found to be 60%. The commercially available polycrystalline MgO (SSA: 25 m2/g) shows a percentage conversion of somewhat less than 20%. A comparative table listing the details of surface area and the yield has been given in Table 2. Although the various magnesium oxide powder, differing in size and surface area, report better catalytic activity, so far none of the studies report to synthesize the oxide nanoparticles and study its catalytic activity towards the condensation reaction. Our method offers a simple route to synthesize high surface area and small sized magnesium oxide crystals with conversion percentage of benzaldehyde approximately 58%. This is comparable to the catalytic activity exhibited by the conventionally prepared magnesium oxide reported by Choudary et al. in their study [22]. Though both

40

Used catalyst

Table 2 Details of the MgO catalyst obtained using different methodologies and the extent of the Claisen–Schmidt reaction catalyzed.

30

20

10

0

Sample name

Preparation route

Size

Specific surface area (m2/g)

Yield (%)

NAP

Aerogel synthesized Conventionally synthesized

Very small stacks of square plates Hexagonal platelets with diameter 150 nm and thickness 10 nm Large cubic crystals

590

98a

250

60a

25

<20b

Nanosheets of diameter 50–200 nm, thickness 3–5 nm Small particles of size 10 nm

–

100b

108

58

NA

2

4

6

8

10

12

14

16

Time (hours) CM

Fig. 4. Time dependence of chalcone formation in presence of MgO using Claisen– Schmidt condensation reaction.

Nanosheet

Table 1 Surface area of the catalyst. Sample

Surface area (m2/g)

Fresh catalyst Used catalyst

108 42

RMR (this study)

a b

Ref. [22]. Ref. [6].

Commercially available Wet chemical method Reverse micellar route followed by thermal decomposition

141

A. Ganguly et al. / Journal of Colloid and Interface Science 353 (2011) 137–142

Used Catalyst

Used catalyst (second reaction)

Carbonyl groups

3300

2900

2600

2200

1800

1400

1100

-1

Wavenumber (cm ) Fig. 7. FTIR spectrum of MgO before and after Claisen–Schmidt reaction.

the catalysis reaction however they tend to appear more agglomerated. The TEM and HRTEM images for the MgO sample after the catalysis reaction has been added as a supplementary image (Fig. S4). 4. Conclusions The area of nanocatalysis is expanding at a fast pace. The current study on MgO based organic catalysis emphasizes: (1) the possibility of using microemulsions. to obtain uniform nanocatalysts of magnesium oxide with reasonably high surface area and catalytic activity equivalent to the conventionally prepared MgO, (2) the presence of surface defects on the oxide nanoparticle may explain the equivalent activity of our catalyst with low surface area (108 m2/g) compared to reported catalyst with high surface area (250 m2/g). The ability to control the homogeneity and size of particles to around 10 nm by the simple microemulsion route has wider industrial implications. Acknowledgments AKG thanks DST – Nanomission, CSIR, and IITD-EPFL (DST – Govt. of India) for financial assistance. TA thank CSIR and DST, Govt. of India for financial support. AG thanks CSIR for a fellowship.

d=2.1111

2.1111

1250

Intensity (Counts)

Fresh Catalyst

Transmittance

the catalysts resemble in terms of their particle size (C.P.: 8.8 nm and R.M.R.:11 nm) they differ largely in terms of their surface area (C.P.: 250 m2/g and R.M.R.: 108 m2/g). From the rates of the catalytic activity, calculated on the basis of normalized surface area, we found that the activity of the particles is higher for the reverse micellar route (5.6  10 3 mmol m 2 h 1) than the conventionally prepared sample (1.93  10 3 mmol m 2 h 1). The higher value suggests that the MgO nanoparticles obtained using the reverse micellar method is a better catalyst even with a low surface area and this is in accordance to the discussion by Landau et al. in their study [23]. This could be attributed to the already known fact that the oxide nanoparticle inherits its morphology from the precursor and the occurrence of defects is largely dependent on its synthetic route [24]. Here, a rod shaped oxalate precursor was synthesized and decomposed to give distinct reactive sites on the surface. Structural defects like edge formation, kinks, and corners generated during the synthesis expose the positively or negatively charged unsaturated sites. The exposure of the negatively charged sites increases the basicity. The surface heterogeneity is affected by the synthetic route which plays an important role in the enhancement of the basicity of the solid [25]. It is possibly due to the presence of these defects which explains our catalyst with low surface area (108 m2/g) but equivalent activity as the catalyst with high surface area (250 m2/g) [22]. In another study, a high yield of 100% was achieved in less than 150 min where MgO nanosheets had been used as catalyst with the sheets corresponding to (1 1 1) planes. However the synthesis of these nanosheets involved the use of expensive alkoxides along with supercritical treatment and lacks convenience for bulk scale application [6]. After the Claisen–Schmidt condensation reaction, the used catalyst was again characterized. Figs. 6 and 7 show the XRD pattern, and the IR spectrum of the catalyst, respectively. In the XRD pattern (Fig. 6) the first two of the three observed peaks shifted to higher ‘d’ value by 0.03 Å. Also the surface area reduced to less than half of the original value observed for the fresh catalyst. The IR spectra of the nanoparticles (Fig. 7) show the presence of some organic moiety in the catalyst. The peak at around 1600 cm 1 is due to C@O stretch which is a result of anchoring of the organic molecules to the Mg2+ ion. The band intensity at 1600 cm 1 is observed to increase when the used catalyst was used yet again for another reaction indicating that the trapped organic molecules also increased with the number of cycles. The reduced surface area justifies the decreased activity towards the reusability of the catalyst. The used catalyst was also characterized using TEM and HRTEM studies for morphological change occurring after the catalysis reaction. The particle size and morphology remain unaffected after

Appendix A. Supplementary material

1000 750

1.4913 d=1.4913

500

Supplementary data associated with this article can be found, in the online version, at doi:10.1016/j.jcis.2010.09.041. References

2.4405 d=2.4405

250 0 10

20

30

40

50

2-Theta (°) Fig. 6. XRD of the catalyst after Claisen–Schmidt reaction.

60

[1] H. Heidari, M. Abedini, A. Nemati, M.M. Amini, Catal. Lett. 130 (2009) 266. [2] B.Q. Xu, J.M. Wie, H.Y. Wang, K.Q. Sun, Q.M. Zhu, Catal. Today 68 (2001) 217. [3] E. Alvarado, L.M. Torres-Martinez, A.F. Fuentes, P. Quintana, Polyhedron 19 (2000) 2345. [4] Y. Ding, G. Zhang, H. Wu, N. Hai, L. Wang, Y. Qian, Chem. Mater. 13 (2001) 435. [5] C. Feldmann, S. Matschulo, S. Ahlert, J. Mater. Sci. 42 (2007) 7076. [6] K. Zhu, J. Hu, K. Christian, R. Richards, Angew. Chem. Int. Ed. 45 (2006) 7277. [7] W. Wang, X. Qiao, J. Chen, H. Li, Mater. Lett. 61 (2007) 3218. [8] F. Khairallah, A. Glisenti, J. Mol. Catal. A: Chem. 27 (2007) 4137.

142

A. Ganguly et al. / Journal of Colloid and Interface Science 353 (2011) 137–142

[9] A.K. Ganguli, A. Ganguly, S. Vaidya, Chem. Soc. Rev. 39 (2010) 474. [10] J. Eastoe, M.J. Hollamby, L. Hudson, Adv. Colloid Interface Sci. 128 (2006) 5–15. [11] R. Ranjan, S. Vaidya, P. Thaplyal, Mohd. Qamar, J. Ahmed, A.K. Ganguli, Langmuir 25 (2009) 6469. [12] M.P. Pileni, Nat. Mater. 2 (2003) 145–150. [13] S. Vaidya, P. Rastogi, S. Agarwal, S.K. Gupta, T. Ahmad, A.M. Antonelli Jr., K.V. Ramanujachary, S.E. Lofland, A.K. Ganguli, J. Phys. Chem. C 112 (2008) 12610. [14] B.M. Choudary, K.V.S. Ranganath, U. Pal, M.L. Kantam, B. Sreedhar, J. Am. Chem. Soc. 127 (2005) 13167. [15] M.L. Kantam, U. Pal, B. Sreedhar, B.M. Choudary, Adv. Synth. Catal. 349 (2007) 1671. [16] B.M. Choudary, L.T. Chakrapani, K.V.R. Kumar, M.L. Kantum, Tetrahedron 62 (2006) 9571.

[17] M.L. Kantam, K. Mahendar, B. Sreedhar, B.M. Choudary, Tetrahedron 64 (2008) 3351. [18] P.K. Stoimenov, R.L. Klinger, G.L. Marchin, K.J. Klabunde, Langmuir 18 (2002) 6679. [19] A.G. Lafuente, E. Guillamo, A. Villares, M.A. Rostagno, Jose Alfredo Martınez, Inflamm. Res. 58 (2009) 537. [20] Y. Yamamoto, R.B. Gaynor, J. Clin. Invest. 107 (2001) 135. [21] Y. Wang, J. Ye, X. Liang, Adv. Synth. Catal. 349 (2007) 1033. [22] B.M. Choudary, M.L. Kantam, K.V.S. Ranganath, K. Mahendar, B. Sreedhar, J. Am. Chem. Soc. 126 (2004) 3396. [23] R. Vidruk, M.V. Landau, M. Herskowitz, M. Talianker, N. Frage, V. Ezersky, N. Froumin, J. Catal. 263 (2009) 196. [24] S. Utamapanya, K.J. Klabunde, J.R. Schulp, Chem. Mater. 3 (1991) 175. [25] A.O. Menezes, P.S. Silva, E.P. Hernandez, L.E.P. Borges, M.A. Fraga, Langmuir 26 (2010) 3382.