Journal of Colloid and Interface Science 437 (2015) 1–9

Contents lists available at ScienceDirect

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

Carboxylic acid effects on the size and catalytic activity of magnetite nanoparticles Hassan Hosseini-Monfared ⇑, Fatemeh Parchegani, Sohaila Alavi Department of Chemistry, University of Zanjan, 45195-313 Zanjan, Islamic Republic of Iran

a r t i c l e

i n f o

Article history: Received 9 December 2013 Accepted 26 August 2014 Available online 16 September 2014 Keywords: Magnetite nanoparticles Carboxylic acid Epoxidation Hydrogen peroxide

a b s t r a c t Magnetite nanoparticles (Fe3O4-NPs) were successfully synthesized in diethylene glycol in the presence of carboxylic acids. They were characterized using XRD, SEM and FTIR. Carboxylic acid plays a critical role in determining the morphology, particle size and size distribution of the resulting particles. The results show that as-prepared magnetite nanoparticles are monodisperse and highly crystalline. The nanoparticles can be easily dispersed in aqueous media and other polar solvents due to coated by a layer of hydrophilic polyol and carboxylic acid ligands in situ. Easily prepared Fe3O4-NPs have been shown to be an active, recyclable, and highly selective catalyst for the epoxidation of cyclic olefins with aqueous 30% H2O2. Ó 2014 Elsevier Inc. All rights reserved.

1. Introduction Stable and nontoxic magnetic iron oxide nanoparticles (NPs) have been applied for detection and separation of proteins [1], to improve the sensitivity of magnetic resonance imaging [2], immunoassay [3], drug [4] and gene delivery [5,6]. Moreover, the use of nanoparticles attracted a lot of attention in the field of catalysis due to their high surface area [7]. In addition, the magnetic nature of, for example, magnetite particles offers the opportunity for the effective and fast recycling of such materials [8,9]. Nanoparticle catalysis in liquid phase has been used since 19th century by applying Pt-NPs to decompose hydrogen peroxide [10,11]. The catalytic properties of the NPs can be modified by particles size. The most obvious size-dependence relationship results from the change in the percentage of surface atoms which are responsible for the catalytic properties when changing the diameter. Growth mechanisms of various Fe3O4 nanostructures have been reported [12] and it is possible to prepare magnetic nanocrystals with controlled size, shape, and uniformity [13]. Although there is a general interest to transform a successful homogeneous catalyst into a heterogeneous one [14], this strategy is not necessary for nanoparticle catalysts. Unsupported nanoparticles are free in solution, thereby such as homogeneous catalysts they show high activity and selectivity because of their high surface area [15]. A homogenous catalyst usually shows more activity and

⇑ Corresponding author. Fax: +98 241 2283203. E-mail address: [email protected] (H. Hosseini-Monfared). http://dx.doi.org/10.1016/j.jcis.2014.08.056 0021-9797/Ó 2014 Elsevier Inc. All rights reserved.

selectivity than a heterogeneous one because of access substrate [16]. In principle, the same condition is provided for the NP catalyst in a solvent and therefore nanocatalysis is a bridge between homogeneous and heterogeneous catalysis [17]. The size of the particle and the structure of the surface play the dominant role in determining the high efficiency of an NP system. Nanostructured compounds may encompass the advantages of both homogeneous and heterogeneous catalysis and offer unique activity with high selectivity in catalysis. Interaction of unprotected small particles will, however lead to agglomeration or aggregation from the cohesive surface energy [17]. As a result of their colloidal instability, many nanoparticles need to be stabilized via additional (capping) agents which provide a steric, electrostatic or electrosteric particle stabilization [15]. Till now, the investigation of ‘‘free’’ nanoparticles as catalysts has been rare [18,19]. Clearly, the development of ‘‘free’’ nanoparticles with tunable catalytic activity is of great significance for the synthesis of hydrocabons. Among the variety of methods for synthesizing magnetic nanoparticles [20–23], the polyol method has been paid more attention [24]. The polyols in this method often serve as reducing agent and stabilizer to prevent interparticle aggregating. In order to stabilize and prevent further the aggregation phenomenon, one can also use different carboxylic acids in conjunction with polyol as stabilizer. The carboxyl group in carboxylic acid can provide coordination to the nanoparticles and thereby the NPs are stabilized. It is well-known that sodium citrate can attach to Au and provides a negatively charged surface for Au nanoparticle stabilization [25,26]. Furthermore, sodium citrate and cetyltrimethylammonium

2

H. Hosseini-Monfared et al. / Journal of Colloid and Interface Science 437 (2015) 1–9

bromide (CTAB) have been applied to dissolve core–shell Fe3O4/Au nanoparticles into water by making the surface hydrophilic [27]. Oleic acid and oleylamine are two common surfactants used for NP stabilization and for the control of NP size and morphology. FePt alloy NPs can be prepared by reducing FeCl2 and Pt(acac)2 by LiBEt3H in the presence of oleic acid and oleylamine [28]. Oleic acid and oleylamine are appropriate surfactants to synthesize monodisperse spherical Fe3O4 nanoparticles [20]. Water-soluble Fe3O4-NPs were synthesized by bifunctional 2,3-dimercaptosuccinic acid [29]. Yang and co-workers used Fe(acac)3/benzyl ether/1,2-hexadecandiol and oleic acid/oleylamine as surfactants to synthesiz monodisperse Fe3O4 nanocubes with controllable sizes of 6 to 30 nm [30]. Superparamagnetic Fe3O4 nanoparticles were functionalized with 11-sulfoundecanoic acid and 10-phosphono-1-decanesulfonic acid ligands to create separable solid acid catalysts for carbohydrate hydrolysis [31]. The activity of the acid-functionalized nanoparticles was higher than the traditional solid acid catalyst Amberlyst15 for the hydrolysis of starch in aqueous solution. In addition, the acid functionalization of nanoparticles may have the potential to surpass polymerization and silica functionalization by producing more active, stable, and tunable acid catalysts for application in green chemical processes. In previous works, Sun et al. [32] have reported the preparation of monodisperse magnetite nanoparticles by the high-temperature solution reduction of Fe(acac)3 (acac = acetylcetonate) by 1,2-hexadecanediol in the presence of oleic acid and oleylamine. Hou et al. [33] have synthesized magnetite nanoparticles through solvothermal reduction of Fe(acac)3 by hydrazine in the presence of ethylene glycol, oleic acid and trioctylphosphine or hexadecylamine. Caruntu and co-workers [34] have prepared magnetite nanocrystals by hydrolysis of a stoichiometric ratio of FeCl36H2O, FeCl24H2O and NaOH in diethylene glycol. Selective oxidation of hydrocarbons is of great interest in synthetic organic chemistry and chemical industry for the preparation of oxygen containing compounds [35–39]. Finding new active and selective epoxidation catalysts for those processes that require an elimination of by-products is of great interest [40]. Magnetite nanoparticles (Fe3O4-NPs) can be completely recovered by means of an external magnetic field owing to its magnetic property; therefore, it is often used as a magnetically separable catalyst [41]. Beller et al. prepared Fe2O3-NPs which were highly stable, active, and highly selective for various oxidation reactions using hydrogen peroxide [18]. The high recyclability of these Fe2O3-NPs has also been demonstrated, indicating they are well suited for continuous processes [18]. One way to attach acidic ligands is to choose ligands that have two acids groups: one that can bind to magnetite and one that will eventually serve as the catalytic site. Portet et al. [42] attached disulfonic acids to magnetite leaving one sulfonic acid group exposed to solution. However, using a sulfonic acid to bind to magnetite is not optimal due to the sulfonic acid group’s weak binding affinity toward the magnetite surface [42–44]. In this work we report a facile polyol approach (reduction of FeCl3 with diethylene glycol) to the synthesize of Fe3O4-NPs in the presence of different bifunctional and unifunctional carboxylic acids (including succinic acid, oxalic acid, benzoic acid and formic acid) as an electrostatic stabilizer and urea as an alkli source to investigate the effects of stabilizers (acids) on aggregation of NPs in diethylene glycol at 200 °C. The resultant samples were characterized in detail by different techniques. Furthermore, the catalytic performance of Fe3O4-NPs was evaluated on the selective oxidation of cyclohexene and ciscyclooctene. The second benefit of using carboxylic acid is its cocatalyst effect. In 1998, it was discovered [45] that the oxidizing power of dinuclear manganese complex as catalyst in the H2O2 oxidation can be dramatically increased if a small amount of a carboxylic acid is added to the reaction solution.

2. Experimental 2.1. Materials Ferric chloride hexahydrate (FeCl36H2O, 98% Fluka), urea (99.5% Merck), diethylene glycol (DEG) (99% Merck, a polyhydric alcohol with a boiling point of 244–245 °C), oxalic acid (99% Merck), benzoic acid (98% Fluka), formic acid (85% Fluka), succinic acid (99% Merck), citric acid (99.9% Fluka), n-octane (Merck), acetonitrile (99.9% Merck), cyclooctene (95% Merck), cyclohexene (95% Merck), H2O2 (35% Fluka) and NH3 (25% Merck) from were used as received. The exact concentration of aqueous hydrogen peroxide (concentration 8.95 mol dm3) was determined before use by titration with standard KMnO4. 2.2. Characterization The phase of the products was examined by XRD on a powder XRD diffractometer (D8ADVANCE, Bruker, Germany) using Cu Ka (k = 0.154056 nm) radiation. The morphology and microstructure of the as-prepared samples were detected on a scanning electron microscopy (SEM). Fourier transform infrared (FT-IR) spectra for magnetite nanoparticles were recorded on a Perkin–Elmer Spectrum One spectrometer after making pellets with KBr powder. The reaction products of the oxidation were determined and analyzed by an HP Agilent 6890 gas chromatograph equipped with a HP-5 capillary column (phenyl methyl siloxane 30 m  320 lm  0.25 lm) with flame ionization detector. 2.3. Preparation of magnetite nanoparticles A reported method with slight modification has been used for the synthesis of magnetite [24]. FeCl36H2O (0.81 g, 3 mmol), carboxylic acid 1 mmol (namely succinic acid 0.118 g, oxalic acid 0.090 g, benzoic acid 0.122 g, formic acid 0.037 ml (1.22 g/mL)) and urea (1.80 g, 30 mmol) were completely dissolved in diethylene glycol (30 mL) by vigorous mechanical stirring. The solution was sealed in a Teflon lined stainless steel autoclave (50 mL capacity) and then heated at 200 °C for 4 h. After cooling down to room temperature, the black magnetite were separated magnetically and washed with ethanol and deionized water for several times to eliminate organic and inorganic impurities, and then dried at 60 °C for 6 h. 2.4. General procedure for oxidation The oxidation of cyclohexene with hydrogen peroxide was performed in a 25-mL round-bottom flask placed in a thermostatic oil bath. In a typical experiment the flask was charged with the suspension of Fe3O4-NPs (1.0 mg) in 3 mL acetonitrile, n-octane as internal standard and cyclohexene (2 mmol). The oxidation reaction was started with addition of 6 mmol aqueous 30% hydrogen peroxide and the mixture was stirred vigorously with magnetic stirrer bar at 80 °C for 6 h. At appropriate intervals, aliquots were removed and analyzed by gas chromatograph. Alkene conversion and yields are based on the starting cyclohexene. The oxidation products were identified by comparison of their retention times with those of authentic samples. The conversion was determined by the following equation

%Conversion ¼

½ðAðsubstrateÞ=AðstandardÞt¼0 h  ½ðAðsubstrateÞ=AðstandardÞt¼xh ½ðAðsubstrateÞ=AðstandardÞt¼0 h ð1Þ

A = peak area in GC chromatogram.

H. Hosseini-Monfared et al. / Journal of Colloid and Interface Science 437 (2015) 1–9

3

In order to investigate the possibility of several recycling runs for Fe3O4-NPs, the solid catalyst was separated from the reaction mixture. In the presence of a magnetic stirrer bar, Fe3O4-NPs moved onto the stirrer bar steadily and the reaction mixture turned clear within 10 s. The catalyst can be isolated by simple decantation. After washing with CH3CN and drying in air, the Fe3O4-NPs can be reused for oxidation of cyclohexene. The used and washed catalyst was transferred to a flask and used again in a fresh reaction. The catalyst was recycled four times for cyclohexene oxidation. In general, no substantial loss in the catalytic activity of the immobilized catalyst was observed compared with that of a fresh sample. 3. Results and discussion 3.1. Synthesis and characterization Magnetite nanoparticles prepared via the reaction of FeCl36H2O with diethylene glycol (DEG) at 200 °C in DEG can be easily separated completely from the reaction solution by a magnet due to magnetic nature of the sample. The excess diethylene glycol acts as both the solvent and reductant [24]. The mechanism can be summarized as follows.

diethylene glycol ¼ RCHOHCH2 OH RCHOHCH2 OH ! RCH2 CHO þ H2 O COðNH2 Þ2 þ H2 O ! NHþ4 þ OH þ HNCO Fe3þ þ 3OH ! FeðOHÞ3 2RCH2 CHO þ 2FeðOHÞ3 ! RCH2 COCOCH2 R þ 2FeðOHÞ2 þ 2H2 O FeðOHÞ2 þ 2FeðOHÞ3 ! Fe3 O4 þ 4H2 O Carboxylic acid was chosen because the carboxylate group has strong coordination affinity to FeIII ions, which favors the attachment of carboxylate groups on the surface of the magnetite nanocrystals and prevents them from aggregating into large single crystals. Considering the strong complexation with the Fe3+ ions, the carboxylate groups could anchor on the particle surface during the solvothermal reaction, and thus enhance the dispersibility of the magnetite particles. Additionally, on using the as-synthesized Fe3O4-NPs as catalyst in the olefin oxidation, hydrogen peroxide is more stable and diminishes the self-decomposition by forming peroxycarboxylic acid [46]. At the same time, as peroxycarboxylic acid is more hydrophobic than the hydrogen peroxide, it acts as a ‘‘soft’’ ligand, favoring to the oxidation of hydrocarbon [46]. In this work, polyol acts as the solvent and reducing agent, ferric salt is used as the iron source, carboxylic acid is the catalyst and urea provides the hydroxyl ion. The synthesis process is performed in a Teflon lined stainless steel autoclave that is sealed. Poly hydroxyl alcohols and carboxylic acids both are efficient stabilizers for NP. Typically, the 18.5 nm magnetite particles were synthesized with the composition of FeCl3/carboxylic acid/urea/DEG = 3:1:30:316 at 200 °C for 4 h. The synthesized sample was characterized by different techniques. Fig. 1 shows the representative XRD pattern of Fe3O4-NPs synthesized in the presence of oxalic acid. The sharp and intense peaks indicate the good crystallinity of the sample. The diffraction peaks can be well corresponded to the standard magnetite Fe3O4 (Joint Committee on Powder Diffraction Standards Card No. 880315). The average grain sizes of the sample was calculated by the Debye–Scherrer formula [47] from the (3 1 1) facet with a shape factor of 0.9 for spherical nanoparticles.

Fig. 1. X-ray diffraction pattern of Fe3O4-NPs obtained in the presence of oxalic acid (Fe3O4-NPs/oxalic); average particle size is 18.5 nm.

D¼

0:9k bcosh

ð2Þ

where ‘k’ is wave length of X-ray (0.1541 nm), ‘b’ is FWHM (full width at half maximum), ‘h’ is the diffraction angle and ‘D’ is particle diameter size. The average grain sizes are 18.5 nm in diameter for the Fe3O4-NPs/oxalic clusters formed. The average particle size of the rest of the samples is: Fe3O4/succinic 17.6 nm, Fe3O4/no acid 15.8 nm, Fe3O4/formic 15.8 nm and Fe3O4/benzoic 15.6 nm (Fig. S1). XRD is relatively insensitive to the difference between magnetite (Fe3O4) and maghemite (c-Fe2O3), and the IR spectrum has been proved to be an alternative tool [44,48]. The nanoparticles were further characterized with Fourier transform infrared spectroscopy (FTIR). Fig. 2 clearly shows the characteristic lattice vibration of magnetite and maghemite. These results are consistent with the as-synthesized nanocrystals mainly being Fe3O4 rather than c-Fe2O3. Presence of the coating agents on the particle surface result to the formation of uniform magnetite nanoparticles. The surface iron atoms of the nanoparticles were coordinated by the carboxylic acid groups of the carboxylic acids, forming a sterically stabilizing layer that prevented nanoparticle aggregation and favored the production of monodisperse samples. The excellent water-dispersibility of the particles is attributed to the free carboxylic acid groups of the stabilizing Fe3O4-NPs that is bound to the particle surface. The presence of the coating agents on the surface of the magnetite nanoparticles was also supported by the FT-IR spectroscopy analysis. Fig. 2 shows FT-IR spectroscopy of the as-prepared the Fe3O4-NPs sample. A strong absorption band at 588 cm1 is related to the vibrations of the Fe-O functional group, which shows that the phase of as-prepared particles is mainly magnetite [49]. The absorption band at 630 cm1 might be assigned to the existence of some amount of oxidized maghemite on the magnetite surface. In the Fe–O range, fine-grained, synthetic maghemite shows broad IR bands at 668, 630, 442 cm-1 [50,51]. The spectrum of magnetite has a very broad band at 588 cm-1 [50]. The broad characteristic band at 3420 cm1 could be assigned to O–H stretching vibration arising from the Fe–OH groups on nanoparticles and adsorbed succinic acid, and water [52]. The peaks around 2925 and 2851 cm1, assignable to asymmetric/symmetric vibrations of C–H in –CH2– of succinic acid [53] and intramolecular hydrogen bond derived from OH [54,55] can be obviously found. Since the same bands (2925 and 2851 cm1) are seen in the IR spectrum of the Fe3O4-NPs

4

H. Hosseini-Monfared et al. / Journal of Colloid and Interface Science 437 (2015) 1–9

Fig. 2. FT-IR spectrum of the 69-nm-sized Fe3O4-NPs synthesized in the presence of succinic acid. The bands at 1635 and 1417 cm1 are associated with carboxylate group. Typical bands assigned to the Fe–O stretching are visible at around 588 and 441 cm1. The bands at 2924 and 3420 cm1 are ascribed to the C–H (–CH2– groups of succinic acid and intramolecular hydrogen bond derived from OH) vibrating and surface O–H vibrating, respectively.

prepared without using any carboxylic acid (Fig. S2), they can also be assigned to intramolecular hydrogen bond derived from the Fe3O4-NPs surface OH [54]. FT-IR spectrum of the magnetite particles obtained with succinic acid shows absorption bands at 1635 and 1417 cm1 associated with asymmetric and symmetric stretching of COO coordinated to the particle surface, which proves that carboxylic acids is covalently absorbed onto the surface of the Fe3O4 particles [2,56,57]. Carboxylate groups which are chelating or coordinating with each O atom to a metal atom exhibit differences between the asymmetric and symmetric stretching frequencies, which are less than the ionic value (Dm = 164 cm1 for the acetate ion) [58]. For Fe3O4-NPs Dm = 1635–1417 = 218 cm-1 suggests a monodentatde action of each succinate carboxylate group toward the iron oxide nanoparticles. In addition, the absorption band at 1636 cm1 on the spectrum referred to the vibration of the remainder H2O in the sample [59]. The size and shape of the products were examined by scanning electron microscopy (SEM). The particle size and the particle size distribution were influenced by the nature of the carboxylic acid, as presented in Fig. 3. XRD gives the crystallite size, not particle size. The grain sizes of the samples estimated from the SEM picture is larger than that obtained from XRD data. This means that, the SEM picture indicates the size of polycrystalline particles. The observation of some larger nanoparticles may be attributed to the fact that Fe3O4 nanoparticles have the tendency to agglomerate due to their high surface energy the ultrafine nanoparticles. SEM images show that all of the magnetite particles obtained have nearly spherical shape and uniform size. The synthesized Fe3O4-NPs without using any carboxylic acid show the highest particle size (373 ± 88). By using a carboxylic acid the resulting nanoparticles became more uniform and the average diameter decreased dramatically. The diameter of the spheres decreases continuously from 373 nm in order oxalic acid (177 ± 42 nm) > benzoic acid (165 ± 49 nm) > formic acid (117 ± 45 nm) > succinic acid (69 ± 17 nm). Likewise, the smallest standard deviation for Fe3O4 particle diameters resulted when succinic acid was used. Probably succinic acid least acidity (pKa 4.2 [60]) is involved in the average particles size diameter and standard deviation. However, the complex nature of the growth process makes it difficult to obtain a straightforward correlation between particle diameter and carboxylic acidity. In addition to acidity and ionic strength which are the two main environmental conditions affecting aggregation, there are also

more subtle effects on aggregation that can have a great impact on the aggregate morphology [61]. Probably, similar to short amyloidogenic peptide which has been proved electrostatic interactions defines both aggregation kinetics and final morphology of the aggregated peptide [62], electrostatics plays an important role in the process of aggregation of Fe3O4/carboxylic particles. Owing to the electrostatic repulsion effect, the magnetite particles can be easily dispersed in water, alcohol, and tetrahydrofuran to form a stable dispersion that can stand for 3 h without visible sedimentation. When a magnet (1.2 T) is applied, the magnetite particles can be separated from their dispersion rapidly in only 1 min. By adjusting the position of the magnet, the typical macroscopic chain-like structure related to the superparamagnetic behavior of the magnetic particles can be observed. Once the magnet is withdrawn, the particles can be redispersed into the water immediately by slight shaking. 3.2. Catalytic activities The oxidation of cyclohexene applying hydrogen peroxide was investigated as a typical benchmark reaction for liquid-phase oxidation (Table 1). Cyclohexene has two fragments accessible for the attack by catalytically active species (the double bond and relatively weak C–H bonds) and is a good representative to give valuable information on the nature of the oxidizing species. Controlled experiments without using H2O2 failed to produce the desired product (entry 1, Table 1). The oxidation of cyclohexene with H2O2 but without the presence of the catalyst proceeded up to less than 1% (entry 2, Table 1). As expected, presence of the catalyst and hydrogen peroxide are essential for the oxidation. Next, we chose different particle sizes of Fe3O4-NPs because of their ferrimagnetic property, hence recovery and recycling of the catalyst can be easily achieved (see below) [63]. In the presence of Fe3O4-NPs/oxalic, in which the majority of the particles are 177 ± 42 nm in size (Fig. 3), the conversion reached to 32% with excellent epoxide selectivity (entry 4, Table 1). The catalytic activity is higher relative to the corresponding Fe3O4-NPs (373 ± 88 nm), Table 1, entry 3). The activity of Fe3O4-NPs/benzoic (165 ± 49 nm) was almost the same with that of Fe3O4-NPs/oxalic (entry 5, Table 1). Interestingly, Fe3O4-NPs/formic (117 ± 45 nm) and Fe3O4-NPs/succinic (69 ± 17 nm), which consists of rather uniform particles of markedly smaller size, showed an even higher activities. Here, the conversions were as 71% and 67%, respectively, and the selectivity

H. Hosseini-Monfared et al. / Journal of Colloid and Interface Science 437 (2015) 1–9

(a) Fe3O4-NPs

(b) Fe3O4-NPs/oxalic

(c) Fe3O4-NPs/benzoic

5

average diameter 373 (± 88) nm

average diameter 177 (± 42) nm

average diameter 165 (± 49) nm

average diameter 117 (± 45) nm

(d) Fe3O4-NPs/formic

average diameter 69 (± 17) nm

(e) Fe3O4-NPs/succinic Fig. 3. SEM images and the corresponding size distribution of Fe3O4-NPs with different sizes synthesized in the presence of different carboxylic acids. The diagrams on the right side provide a size-distribution histogram of Fe3O4-NPs; the diameter average sizes of the particles, obtained by measuring more than 60 clusters for each sample. Fe3O4-NPs prepared: in the absence of carboxylic acid (a), in the presence of oxalic acid (b), benzoic acid (c), formic acid (d) and succinic acid (e).

6

H. Hosseini-Monfared et al. / Journal of Colloid and Interface Science 437 (2015) 1–9

Table 1 Fe3O4-NPs catalyzed oxidation of cyclohexene by H2O2.a

Entry

Average particle size (nm)

Catalyst

Conversion (%)

TONb

1 2 3 4 5 6 7

373 ± 88

Fe3O4-NPs None Fe3O4-NPs Fe3O4-NPs/oxalic Fe3O4-NPs/benzoic Fe3O4-NPs/formic Fe3O4-NPs/succinic

0c 0.8 23 32 28 71 67

4 106 148 130 329 310

373 ± 88 177 ± 42 165 ± 49 117 ± 45 69 ± 17

a Reaction conditions: Fe3O4-NPs (0.001 g, 4.32 lmol), cyclohexene (2 mmol), acetonitrile (3 ml), n-octane (0.1 g), reaction time (9 h), H2O2 (6 mmol), reaction temperature (80 0C). The conversions of cyclohexene are based on GC-FID analysis with an internal standard of n-octane. Epoxide selectivity >99%. b Turnover number (TON) = (mmole of products/mmole of catalyst)  100. c In the absence of H2O2.

remained >99% (entry 6 and 7, Table 1). For cyclohexene, the catalyst turnover number reached a value of up to 329 with greater than 99% chemoselectivity (entry 6, Table 1). The reason for the improved activity of Fe3O4-NPs most probably originates from the nanometer size of the iron oxide and nature of the stabilizing carboxylic acid. Comparable activity of Fe3O4-NPs/formic (117 ± 45 nm) and Fe3O4-NPs/succinic (69 ± 17 nm) in spite of their difference in particles size, suggests also the role of stabilizing carboxylic acid in the magnetite activity. In general, nanoscale heterogeneous catalysts should offer higher surface areas, low-coordinated sites, and surface vacancies, which are responsible for the higher catalytic activity [64]. Theoretically, it can be assumed that with a decrease of the particle size down to a ‘‘molecular’’ level, the nanocatalyst behaves as a homogeneous system in which the catalytic activity is not controlled by the surface area of the catalyst but governed by the concentration [65]. The most important significance of these results is that ‘‘free’’ nano-Fe3O4, but not immobilized nano-Fe3O4, is highly active, selective, and stable mainly by controlling the particle size and carboxylic acid. The excellent selectivities of the Fe3O4-NPs/carboxylic catalysts toward epoxide can be related to co-catalytic effect of the carboxylic acids. The carboxylic acids by stabilizing the Fe3O4-NPs influence on the size and morphology of the NPs and also by involving in the formation of a reactive species, probably by promotion of H2O2 to a peracid, enhance the catalytic activity and selectivity of Fe3O4-NPs. The key role of carboxylic acid on the rate and selectivity of the epoxidation of alkenes by H2O2 have been documented in the literature. The addition of carboxylic acids enhanced the rate and selectivity of alkene oxidations with H2O2 catalyzed by [(bpmen)Fe(OTf)2] (bpmen = N,N0 -dimethyl-N,N0 bis(2-pyridylmethyl)-1,2-diaminoethane) [66]. Screening of various carboxylic acids as acid additives with different substitutents with Na2WO4/H2WO4 toward alkene epoxidation, leads to the observation that the substituted/unsubstituted acetic and benzoic acids generally enhance the rate of the epoxidations of cyclooctene and 1-octene with H2O2 [67]. For the epoxidation of various olefins with hydrogen peroxide in the presence of ruthenium trichloride, presence of pyridine-2,6-dicarboxylic acid (pydic) as ligand increased the yield and chemoselectivity [68]. The role of the carboxylic acid as co-catalyst in metalloporphyrin catalytic hydrogen peroxide oxidations was studied and a metallo-acylperoxo complex suggested as the effective oxidation intermediate [69]. Small cyclic olefins such as cyclooctene and cyclohexene were converted by the Mn(II) dicarboxylate coordination polymer [Mn(l-terephthalate)(H2O)2]n (terphthalate = –O2C–C6H4–CO2–) to the corresponding epoxides with good yields (75% and 64%) and 100%

selectivity [70]. Enhancement of the activity by acetic acid in the oxidation of cyclohexane has been reported in the literatures [71,72]. In the media of acetic acid, hydrogen peroxide is more stable and diminishes the self-decomposition by forming peroxyacetic acid. Activation of H2O2 by citric acid and the possibility of formation of peroxycitric acid just like the formation of peroxyacetic acid have been reported [73]. In search of suitable reaction conditions to achieve the maximum oxidation of olefin, the effects of the reaction temperature and oxidant (moles of H2O2 per moles of cyclooctene) on the oxidation of cis-cyclooctene were studied. Similar to cyclohexene, the oxidation of cis-cyclooctene by Fe3O4-NPs/H2O2 was epoxide with >99% selectivity. The catalytic activity of Fe3O4-NPs depends on the temperature. As shown in Fig. 4, the conversion of cis-cyclooctene increased with temperature from 27 to 80 °C and a maximum conversion of 85.5% was obtained after 24 h at 80 °C. Notably, induction period of about 5 h was observed in any of the reactions. Fe3O4-NPs increased activity is only evident at prolonged reaction times. It seems that the particles of Fe3O4-NPs become well dispersed in acetonitrile after about 5 h of magnetic stirring, and then it has the highest catalytic activity in the oxidation of cyclooctene by H2O2. This assumption, however, did not confirmed by experiment. To check this possibility Fe3O4-NPs/formic in acetonitril were treated at 80 °C for 10 h by continuous stirring in order to be dispersed well, then cyclohexene/H2O2/n-octane were added and the reaction followed by GC. Very similar to the not treated Fe3O4-NPs/formic (Table 1, entry 6) the conversion was 70% after 9 h from the addition of H2O2 but up to 8 h the conversion was only 10%. The induction period is typical of active species leaching from heterogeneous pre-catalyst (slow release of homogeneously active catalysts) that requires a time to build up a concentration of the active species [70]. However, in this study in spite of the showing the induction period, Fe3O4-NPs behave such as a heterogeneous catalyst, as evidenced by the following series of experiments. 1. In the first experiment 0.0010 g Fe3O4-NPs/succinic was used for epoxidation of cyclohexene with H2O2 in 3 cm3 CH3CN. After 9 h not detectable amount of the compound had dissolved, because about 0.0010 g could be recovered by using an external magnet and washing with CH3CN. Cyclohexene conversion was 64%. 2. In the second experiment the filtrate of the first experiment was used for catalysis. After addition of new cyclohexene–H2O2 the

Fig. 4. Temperature dependence of the oxaidation of cyclooctene with hydrogen peroxide. Reaction conditions: Fe3O4-NPs 1.0 mg, 4.32 lmol, cyclooctene 2.0 mmol, CH3CN 3 mL, n-octane 0.1 g, H2O2 6 mmol.

7

H. Hosseini-Monfared et al. / Journal of Colloid and Interface Science 437 (2015) 1–9

reaction solution was stirred at 80 °C for 9 h. Cyclohexene conversion was less than 1%. 3. Leaching of the Fe is excluded since the oxidation reaction continues without stopping only with 3 equivalents of H2O2. It has been well documented that presence of the traces of Fe2+ or Fe3+ results in the prompt decomposition of H2O2 [74] and as a result the oxidation reaction is stopped. If the leaching of the Fe were occurred, higher equivalents of H2O2 (about 10 mmol H2O2 for 1 mmol substrate [70]) must be used to achieve the substrate oxidation. Furthermore, analysis of the reaction solution by atomic absorption spectroscopy (AAS) at the end of the reaction (24 h) and after removing the catalyst showed no free iron ions. 4. By increasing the reaction temperature no reduction in the induction period was observed [70] which also supports the heterogeneous catalytic nature of Fe3O4-NPs.

The solid Fe3O4-NPs catalyst is reusable. When the oxidation of cyclohexene was repeated four times with the same sample of catalyst, cyclohexene was oxidized at the same rate as that of the first run and the results are seen in Table 2 and Fig. 6. Importantly, note that the ferrimagnetic property of Fe3O4-NPs made the isolation and reuse of this catalyst very easy. In the presence of a magnetic stirrer bar, Fe3O4-NPs moved onto the stirrer bar steadily and the

Table 2 Recycling studies with catalyst Fe3O4-NPs/succinic in the oxidation of cyclohexene by hydrogen peroxide.a Entry

No. recycle

Conv. (%)

1 2 3 4 5

0 1 2 3 4

64 66 63 62 66

We have monitored the progress of cyclooctene epoxidation for different molar ratios of oxidant and substrate (Fig. 5). Molar ratios H2O2/cis-cyclooctene of 2, 3, 5 and 10 were considered while keeping the fixed amount of cis-cyclooctene (2.0 mmol) and catalyst Fe3O4-NPs (4.32 lmol) in 3 ml of acetonitrile at 80 °C. It was found that at a ratio of oxidant:alkene of 2 the conversion of cyclooctene was only 14%. Increasing the H2O2/cyclooctene ratio from 2 to 3 increased the conversion from 14% to 85.5%. Further increasing the H2O2/cyclooctene molar ratio up to 5 and 10, the conversion decreased to 68% and 72%, respectively. Reduction in the activity might be due to the increased hydrogen peroxide molecules interaction and increased H2O2 dismutation. So the H2O2/cyclooctene ratio of 3 was selected as an optimized ratio for further studies. These findings support slow decomposition of hydrogen peroxide by Fe3O4-NPs. This result indicates a catalase-like dismutation of H2O2 by Fe3O4-NPs. Consequently, two main reactions, namely the substrate oxidation and the H2O2 dismutation, are competing. Previously, iron(III) complexes with porphyrin/phenolate [75] and tetraaza [14] annulenes [76] ligands have been shown to be catalase models. Because of the second reaction order in hydrogen peroxide in catalase-type dismutation (two molecules of H2O2 are required to produce one molecule of O2) a decrease of the local H2O2 concentration should strongly disfavor the dismutation reaction with respect to the catalytic oxidation. An excess of oxidant, the H2O2/cyclooctene ratio of 10, is documented for epoxidation reactions in the literature [77,78]. With 2 mmol cyclooctene as the starting material, a catalyst turnover number of 396 was obtained with >99% epoxide selectivity.

Fig. 6. The kinetics of H2O2 oxidation of cyclohexene catalyzed by Fe3O4-NPs/ succinic in acetonitrile. Reaction conditions: Fe3O4-NPs/succinic 1.0 mg, cyclohexene 2.0 mmol, CH3CN 3 mL, aqueous 30% H2O2 6.0 mmol, n-octane 0.1 g, T = 80 °C, t = 9 h. Epoxide selectivity >99%.

Fig. 5. Effect of different molar ratios of oxidant/substrate. Reaction conditions: Fe3O4-NPs 1.0 mg, 4.32 lmol cyclooctene 2.0 mmol, CH3CN 3 mL, n-octane 0.1 g, 80 °C.

Fig. 7. Infrared spectra of Fe3O4-NPs/succinic (a) and recovered Fe3O4-NPs/succinic after catalysis (b); reaction conditions are as reported in Fig. 6.

a Reaction conditions: Fe3O4-NPs/succinic 1.0 mg, cyclohexene 2.0 mmol, CH3CN 3 mL, aqueous 30% H2O2 6.0 mmol, noctane 0.1 g, T = 80 °C, t = 9 h. Epoxide selectivity >99%.

8

H. Hosseini-Monfared et al. / Journal of Colloid and Interface Science 437 (2015) 1–9

Fig. 8. Comparison of the SEM images of (a) fresh and (b) recovered Fe3O4-NPs/succinic after catalysis.

reaction mixture turned clear within 10 s. The catalyst can be isolated by simple decantation. After washing with CH3CN and drying in air, the Fe3O4-NPs can be directly reused without any deactivation even after four rounds of selective oxidation of cyclohexene. In general, no substantial loss in the activity of the catalyst was observed compared with that of fresh sample, as shown in Fig. 6. The reused catalyst displayed consistent reactivity and selectivity. The characterization of the Fe3O4-NPs before and after reuse four times showed the same characteristics and particle size by FT-IR (Fig. 7) and scanning electron microscopy (Fig. 8), respectively. 4. Conclusions Magnetite nanoparticles have been prepared by using a modified polyol route in the presence of various carboxylic acids. The nature of the carboxylic acid plays an important role in determining the particle size and the size distribution. Easily prepared Fe3O4-NPs has been shown to be an active, stable, and highly selective catalyst for oxidations of cyclic olefins applying with 3 equiv aqueous 30% H2O2. In the oxidation of cyclooctene, products were formed in up to 396 turnovers. Activation of H2O2 by this complex raises the prospect of using this type of simple catalyst for selective organic syntheses. Because of its simple recyclability, the catalyst is well-suited for continuous processes. Acknowledgment The authors are grateful to the University of Zanjan for financial support of this study. Appendix A. Supplementary material Supplementary data associated with this article can be found, in the online version, at http://dx.doi.org/10.1016/j.jcis.2014.08.056. References [1] Y. Deng, D. Qi, C. Deng, X. Zhang, D. Zhao, J. Am. Chem. Soc. 130 (2008) 28. [2] J. Jia, J.C. Yu, X.-M. Zhu, K.M. Chan, Y.-X.J. Wang, J. Colloid Interface Sci. 379 (2012) 1. [3] A. Agrawal, T. Sathe, S.M. Nie, J. Agric. Food Chem. 55 (2007) 3778. [4] M.K. Yu, Y.Y. Jeong, J. Park, S. Park, J.W. Kim, J.J. Min, K. Kim, S. Jon, Angew. Chem. Int. Ed. 47 (2008) 5362. [5] X.L. Wang, L.Z. Zhou, Y.J. Ma, X. Li, H.C. Gu, Nano Res. 2 (2009) 365. [6] B.R. Hao, R. Xing, Z. Xu, Y. Hou, S. Gao, S. Sun, Adv. Mater. 22 (2010) 2729. [7] D. Astruc, F. Lu, J.R. Aranzaes, Angew. Chem. 117 (2005) 8062. [8] D. Lee, J. Lee, H. Lee, S. Jin, T. Hyeon, B.M. Kim, Adv. Synth. Catal. 348 (2006) 41. [9] R. Abu-Reziq, H. Alper, D. Wang, M.L. Post, J. Am. Chem. Soc. 128 (2006) 5279. [10] J.S. Bradley, in: G. Schmid (Ed.), Clusters and Colloids, VCH, Weinheim, 1994, p. 459 (Chapter 6). [11] J.A. Widegren, R.G. Finke, J. Mol. Catal. A: Chem. 198 (2003) 317. [12] C. Yang, J. Wu, Y. Hou, Chem. Commun. 47 (2011) 5130. [13] L. Fei, Z. Jing-Han, H. Yang-Long, G. Song, Chin. Phys. B 22 (2013) 107503-1. [14] A. Corm, H. Garcia, Chem. Rev. 103 (2003) 4307. [15] D. Astruc, F. Lu, J.R. Aranzaes, Angew. Chem. Int. Ed. 44 (2005) 7852. [16] A. Corma, H. Garcia, Adv. Synth. Catal. 348 (2006) 1391.

[17] O.T. Mefford, M.L. Vadala, J.D. Goff, M.R.J. Caroll, R. Mejia-Ariza, B.L.T.G. Caba, St. Pierre, R.C. Woodward, R.M. Davis, J.S. Riffle, Langmuir 24 (2008) 5060. [18] F. Shi, M.K. Tse, M.-M. Pohl, A. Brückner, S. Zhang, M. Beller, Angew. Chem. Int. Ed. 46 (2007) 8866. [19] H. Hosseini-Monfared, H. Meyer, C. Janiak, J. Mol. Catal. A: Chem. 372 (2013) 72. [20] S.H. Sun, H. Zeng, J. Am. Chem. Soc. 124 (2002) 8204. [21] J. Park, E. Lee, N.M. Hwang, M.S. Kang, S.C. Kim, Y. Hwang, J.G. Park, H.J. Noh, J.Y. Kini, J.H. Park, T. Hyeon, Angew. Chem. Int. Ed. 44 (2005) 2872. [22] Z. Li, B. Tan, M. Allix, A.I. Cooper, M.J. Rosseinsky, Small 4 (2008) 231. [23] X.H. Sun, C.M. Zheng, F.X. Zhang, Y.L. Yang, G.J. Wu, A.M. Yu, N.J. Guan, J. Phys. Chem. C 113 (2009) 16002. [24] C. Cheng, F. Xu, H. Gu, New J. Chem. 35 (2011) 1072. [25] G. Frens, Nat. Phys. Sci. 241 (1973) 20. [26] W.S. Sutherland, J.D. Winefordner, J. Colloid Interface Sci. 48 (1992) 129. [27] Z. Xu, Y. Hou, S. Sun, J. Am. Chem. Soc. 129 (2007) 8698. [28] S. Sun, S. Anders, T. Thomson, J.E.E. Baglin, M.F. Toney, H.F. Hamann, C.B. Murray, B.D. Terris, J. Phys. Chem. B 107 (2003) 5419. [29] Z.P. Chen, Y. Zhang, S. Zhang, J.G. Xia, J.W. Liu, K. Xu, N. Gu, Colloids Surf. A 316 (2008) 210. [30] H.T. Yang, T. Ogawa, D. Hasegawa, M. Takahashi, J. Appl. Phys. 103 (2008) 07d526. [31] M. Ikenberry, L. Peña, D. Wei, H. Wang, S.H. Bossmann, T. Wilke, D. Wang, V.R. Komreddy, D.P. Rillema, K.L. Hohn, Green Chem. 16 (2014) 836. [32] S. Sun, H. Zeng, D.B. Robinson, S. Raoux, P.M. Rice, S.X. Wang, G.X. Li, J. Am. Chem. Soc. 126 (2004) 273. [33] Y. Hou, J. Yu, S. Gao, J. Mater. Chem. 13 (2003) 1983. [34] D. Caruntu, G. Caruntu, Y. Chen, C.J. O’Connor, G. Goloverda, V.L. Kolesnichenko, Chem. Mater. 16 (2004) 5527. [35] R.A. Sheldon, J.K. Kochi, Metal Catalyzed Oxidations of Organic Compounds, Academic Press, New York, 1981. [36] T. Punniyamurthy, S. Velusamy, J. Iqbal, Chem. Rev. 105 (2005) 2329. [37] B. Zhan, B. Moden, J. Dakka, J. Santiesteban, S.C. Reyes, E. Iglesia, US 2011/7 868 201 B2. [38] G.B. Shul’pin, J. Mol. Catal. A: Chem. 189 (2002) 39. [39] E.P. Talsi, K.P. Bryliakov, Coord. Chem. Rev. 256 (2012) 1418. [40] H. Hosseini-Monfared, V. Aghapoor, M. Ghorbanloo, P. Mayer, Appl. Catal. A: Chem. 372 (2010) 209. [41] C. Huang, H. Zhang, Z. Sun, Y. Zhao, S. Chen, R. Tao, Z. Liu, J. Colloid Interface Sci. 364 (2011) 298. [42] D. Portet, B. Denizot, E. Rump, J.-J. Lejeune, P. Jallet, J. Colloid Interface Sci. 238 (2001) 37. [43] Y. Sahoo, H. Pizem, T. Fried, D. Golodnitsky, L. Burstein, C.N. Sukenik, G. Markovich, Langmuir 17 (2001) 7907. [44] R. Cornell, U. Schwertmann, The Iron Oxides: Structure, Properties, Reactions, Occurrences, and Uses, second ed., Wiley-VCH, 2003. [45] J.R. Lindsay Smith, G.B. Shul’pin, Tetrahedron Lett. 39 (1998) 4909. [46] Y. Du, Y. Xiong, J. Li, X. Yang, J. Mol. Catal. A: Chem. 298 (2009) 12. [47] C. Gnanaprakash, S. Mahadevan, T. Jayakumar, P. Kalyanasundaram, J. Philip, B. Raj, Mater. Chem. Phys. 103 (2007) 168. [48] P. Tartaj, M.D. Morales, S. Veintemillas-Verdaguer, T. Gonzalez-Carreno, C.J. Serna, J. Phys. D Appl. Phys. 36 (2003) R182. [49] R.M. Cornell, U. Schwertmann, The Iron Oxides, Wiley-VCH, Weinheim, Germany, 1996. [50] G.W. Poling, J. Electrochem. Soc. 116 (1969) 958. [51] R.M. Taylor, U. Schwertmann, Clay Min. 10 (1974) 29. [52] W. Cai, J.Q. Wan, J. Colloid Interface Sci. 305 (2007) 366. [53] X. Liang, X. Wang, J. Zhuang, Y.T. Chen, D.S. Wang, Y.D. Li, Adv. Funct. Mater. 16 (2006) 1805. [54] Z. Pei, H. Xu, Y. Zhang, J. Alloys Compd. 468 (2009) L5. [55] Navendu Goswami, P. Sen, Solid State Commun. 132 (2004) 791. [56] K.V.P.M. Shafi, A. Ulman, X.Z. Yan, N. Yang, C. Estournes, H. White, M. Rafailovich, Langmuir 17 (2001) 5093. [57] Y.T. Tao, J. Am. Chem. Soc. 115 (1993) 4350. [58] K. Nakamato, Infrared and Raman Spectra of Inorganic and Coordination Compounds, fourth ed., Wiley, New York, 1986. p 232. [59] T. Rajh, L.X. Chen, K. Lukas, T. Liu, M.C. Thurnauer, D.M. Tiede, J. Phys. Chem. B 106 (2002) 10543.

H. Hosseini-Monfared et al. / Journal of Colloid and Interface Science 437 (2015) 1–9 [60] H.C. Brown et al., in: E.A. Braude, F.C. Nachod (Eds.), Determination of Organic Structures by Physical Methods, Academic Press, New York, 1955. [61] A.H. DePace, J.S. Weissman, Nat. Struct. Biol. 9 (2002) 389. [62] A.M. Portillo1, A.V. Krasnoslobodtsev1, Y.L. Lyubchenko1, Phys. Condens Matter. 24 (2012) 164205. [63] C.O. Dálaigh, S.A. Corr, Y. GunQko, S.J. Connon, Angew. Chem. 119 (2007) 4329. [64] G. Pacchioni, Surf. Rev. Lett. 7 (2000) 277. [65] Y. Zhao, K. Aoki, Chem. Phys. Lett. 430 (2006) 117. [66] L.S. Morris, M.P. Girouard, M.H. Everhart, W.E. McClain, J.A. van Paridon, R.D. Pike, C. Goh, Inorg. Chim. Acta 413 (2014) 149. [67] P.U. Maheswari, X. Tang, R. Hage, P. Gamez, J. Reedijk, J. Mol. Catal. A: Chem. 258 (2006) 295. [68] M. Klawonn, M.K. Tse, S. Bhor, C. Döbler, M. Beller, J. Mol. Catal. A: Chem. 218 (2004) 13.

9

[69] A.M.d’A. Rocha Gonsalves, A.C. Serra, J. Mol. Catal. A: Chem. 168 (2001) 25. [70] H. Hosseini-Monfared, A. Mohajeri, A. Morsali, C. Janiak, Monatsh. Chem. 140 (2009) 1437. [71] T. Sooknoi, J. Limtrakul, Appl. Catal. A: Gen. 233 (2002) 227. [72] G.B. Shul’pin, G. S-Fink, J.R.L. Smith, Tetrahedron 55 (1999) 5345. [73] A.W. Karunditu, C.M. Carr, K. Dodd, Textile Res. J. 64 (1994) 570. [74] F.A. Cotton, G. Wilkinson, Advanced Inorganic Chemistry, fifth ed., Wiley, New York, 1988. [75] A. Takahashi, T. Kurahashi, H. Fujii, Inorg. Chem. 48 (2009) 2614. [76] R. Sustmann, H.-G. Korth, D. Kobus, J. Baute, K.-H. Seiffert, E. Verheggen, E. Bill, M. Kirsch, H. de Groot, Inorg. Chem. 46 (2007) 11416. [77] B.S. Lane, M. Vogt, V.J. DeRose, K. Burgess, J. Am. Chem. Soc. 124 (2002) 11946. [78] S. Ueno, K. Yamaguchi, K. Yoshida, K. Ebitani, K. Kaneda, Chem. Commun. (1998) 295.