Catalytic activity of maghemite supported palladium catalyst in nitrobenzene hydrogenation

A maghemite supported palladium catalyst was prepared and tested in nitrobenzene hydrogenation. The catalyst support was made by a newly developed combined technique, where sonochemical treatment and combustion have been used. As a first step, maghemite nanoparticles were synthesized. Iron(II) citrate was treated in polyethylene glycol by high-intensity ultrasound cavitation to get a homogeneous dispersion, then the product was combusted. The produced powder contained maghemite nanoparticles with 21.8 nm average particle size. In the second step of catalyst preparation, the magnetic nanoparticles were dispersed in the ethanolic solution of palladium(II) nitrate. The necessary energy for the reduction of Pd2+ ions was achieved in the “hot spots” by acoustic cavitation, thus catalytically active palladium was formed. The prepared maghemite supported Pd catalyst have been tested in nitrobenzene hydrogenation at three different temperatures (283 K, 293 K and 303 K) and constant pressure (20 bar). At 293 K and 303 K, the conversion and selectivity of nitrobenzene was above 99% and 96%, respectively. However, the selectivity was only 73% at 273 K because the intermediate species (azoxybenzene and nitrosobenzene) have not been transformed to aniline. All in all, the prepared catalyst is successfully applied in nitrobenzene hydrogenation and easily separable from the reaction media.


Introduction
Several different complex catalysts have been successfully applied in the hydrogenation of nitro groups, such as carbon (C), silica (SiO 2 ) or alumina (Al 2 O 3 ) supported Pd, Pt, Ru, Rh, Ni, Fe or bimetallic systems [1][2][3][4][5][6][7][8][9][10][11][12][13]. The easy handling and 1 3 separability are very important properties for the catalysts. These can be improved by introducing magnetic features (e.g. magnetic catalyst supports), which will allow the easy and efficient removal of the catalysts after the reactions. For this reason, magnetic systems have been widely used in various applications. Magnetite (Fe 3 O 4 )/ silica composite catalyst was used for esterification of palmitic acid with methanol [14]. Pd, Pr-Cu and Pr 6 O 11 decorated Fe 3 O 4 /SiO 2 catalyzed the reduction of 2,4-dinitrophenylhydrazine, 4-nitrophenol and chromium(VI) ions, Mizoroki-Heck coupling reaction, and the catalytic ozonation of acetochlor [15][16][17]. Magnetite/ carbon support was applied in Suzuki-Miyaura cross-coupling of 4-iodotoluene and phenylboronic acid and aniline synthesis by palladium [18,19]. By magnetite/ alumina supported Pd catalyst the hydrogenation of nitrate in water and 4-nitrophenol can be achieved [20,21]. Magnetic iron oxides can be combined with different layered double hydroxides (Fe 3 O 4 -LDH), complex magnesium silicates (Fe 3 O 4 -sepiolite) and hydroxyapatite (γ-Fe 2 O 3 -HAP) to use as a support for Pd and these catalytic systems can be applied to catalyze the Heck reaction between iodobenzene and styrene, and the reduction of nitroarenes and nitrobenzene [22][23][24]. Magnetite itself is also a promising catalyst support as it was proved by the applicability of Ag/Fe 3 O 4 , Ag-Ni/Fe 3 O 4 , Pd/Fe 3 O 4 and Rh/Fe 3 O 4 systems in the synthesis of 3,4-dihydropyrimidinones. 2,4-dihydropyrano [2,3-c]pyrazoles, and the hydrogenation of soybean oil and nitroarenes [25][26][27][28]. The two main components of the catalysts mentioned above are the support and the catalytically active metal. The catalysts are prepared through several steps, including the activation of metal, within which metal (e.g. palladium ions) ions or their complex ions are reduced to the catalytically active form (e.g. Pd 0 ). In the case of Pd ions, the activation (reduction) can be done on the supports in aqueous solution by molecular hydrogen (6 atm, 75 °C) or by using NaBH 4 in ethanol but the ethylene glycol is also efficient [21,24,29].
In our work, a simplified reduction step was applied during the catalyst production (palladium(II) nitrate to Pd 0 ) by applying alcohol and acoustic cavitation. The high energy of the ultrasonic treatment in liquids generates acoustic cavitation, which leads to the formation of micro vapor-bubbles. The collapse of the formed bubbles leads to "hot spots" where intense local heating (~ 5000 K), high pressure (~ 1000 atm), enormous heating and cooling rates (> 109 K/s) and liquid jet streams (~ 400 km/h) appear in a small volume [30]. The energy in the "hot spots" can cover the needs of the reduction of metal ions to metals in the presence of a reducing agent [31][32][33][34][35][36]. By using of ultrasonic cavitation, palladium nanoparticles were deposited on the surface of maghemite in methanol phase. Owing to the magnetic properties of the maghemite, this is a remarkable catalyst support in liquid phase hydrogenation because the catalyst easily separated from the reaction media by magnetic field.

Application of maghemite supported palladium catalyst
Maghemite nanoparticles, as catalyst supports were synthesized by a combustion method. 3.5 g iron(III) citrate hydrate was dispersed in 20 g polyethylene glycol (PEG 400, Sigma Aldrich) by using a Hielscher Ultrasound tip homogenizer. The iron precursor containing dispersion was heated up and burned at 500 °C in a calcining furnace for two hours.
The before-synthetized maghemite was applied for catalyst preparation by using a Hielscher Ultrasound tip homogenizer (UIP1000hDT). The palladium precursor (0.125 g) was solved in 50 ml abs. ethanol, and 1.00 g maghemite was added to the solution. The ethanolic dispersion was sonicated by using the homogenizer (115 W, 19.43 kHz) for 2 min. Then, the catalyst was removed from the dispersion with a Nd magnet, washed with ethanol, and dried at 105 °C overnight.

Characterization techniques of the nanoparticles
Maghemite and palladium nanoparticles were examined by using high-resolution transmission electron microscopy (HRTEM, FEI Technai G2 electron microscope, 200 kV). The samples were prepared by dropping their aqueous suspension on 300 mesh copper grids (Ted Pella Inc.). The diameters of the nanoparticles were measured on the HRTEM images, based on the original scale bar by using the ImageJ software. X-ray diffraction (XRD) measurements were used to identify and quantify the crystalline phases, by applying a Rigaku Miniflex II diffractometer with Cu K α radiation source (30 kV, 15 mA). The palladium content was determined with a Varian 720 ES inductively coupled optical emission spectrometer (ICP-OES), by using a Merck Certipur ICP multi-element standard IV.

Catalytic tests
The catalytic hydrogenation was carried out in a Büchi Uster Picoclave reactor, in a 200 ml stainless steel vessel with heating jacket. The hydrogen pressure was 20 bar and the reactions were carried at 283 K, 293 K and 323 K. Sampling took place after the beginning of hydrogenation at 5, 10, 15, 20, 30, 60, 120, 180, and 240 min. The initial concentration of nitrobenzene was 0.125 mol dm −3 in methanol. The total amount of the solution was 150 ml and 0.2 g catalyst was used during each test. Aniline formation was followed by applying Agilent 7890A gas chromatograph coupled with Agilent 5975C Mass Selective detector. Analytical standards (aniline, nitrobenzene, nitrosobenzene, azoxybenzene, dicyclohexylamine, o-toluidine, cyclohexylamine and n-methylaniline) were provided by Dr. Ehrenstorfer and Sigma Aldrich. The efficiency of the catalytic hydrogenation was compared by calculating the conversion (X%) of nitrobenzene based on the following equation (Eq. 1):

3
The selectivity (S%) of the catalyst was calculated as follows (Eq. 2): re n aniline and n nitrobenzene are the corresponding chemical amounts of the compounds.
By assuming that the process is a first order reaction, based on the initial and measured nitrobenzene concentrations (c 0 and c k , mol/dm 3 ), the reaction rate constant (k) was calculated at different temperatures by non-linear regression (Fig. 4) according to the following (Eq. 3):

Surface morphology and phase composition of the catalyst
The reduction of palladium ions to elemental Pd have been confirmed by XRD measurements (Fig. 1a). Reflections at 40° and 48.8° 2θ degrees were identified on the XRD pattern, which are attributed to the Pd(111) and Pd(200) phases (Fig. 1a, red line). Other reflections were also identified such as the peaks at 24.1°, 30.3°, 35.7°, 43.3°, 54°, 57.3° and 63° 2θ degrees, which are assigned to the presence of (210), (220),(100), (400), (422), (511) and (400) planes of maghemite (γ-Fe 2 O 3 ) crystalline phase. The average size of the maghemite particles was found to be 21.8 nm (Fig. 1b). Palladium nanoparticles were deposited onto the surface of the maghemite crystals. The palladium deposition onto the surface of the maghemite led to the aggregation of the magnetic particles, the size of the nanocomposite aggregates are between 70-200 nm. The palladium particles on the maghemite aggregates (1) X% = consumed n nitrobenzene initial n nitrobenzene × 100.
(2) S% = n aniline n nitrobenzene × 100 are smaller than 8 nm (the average size is 4.5 nm) (Fig. 1c). The formation of palladium nanoparticles can be explained by the reduction effect of the appearing . CH 2 R radicals. These reactive species generated by the ultrasonic treatment through the reaction of . OH radicals and ethanol [37,38].

Catalytic activity of the magnetic Pd catalyst
The prepared magnetic Pd catalyst was tested in nitrobenzene hydrogenation. The maximum conversions were reached after 80 min at 293 K and 303 K (Fig. 2a). At 283 K the reaction is slower, but the total amount of nitrobenzene was transformed to aniline. The aniline selectivity was high, 97% and 96.7% at 293 K and 303 K, respectively (Fig. 2b). The catalytic activity was tested through five cycles at 303 K and 20 bar hydrogen pressure, while the reaction time was 80 min. The catalyst was not regenerated between the cycles, only washed with methanol. The activity started to decrease from the third cycle, which indicates that the regeneration of the catalyst is necessary (Fig. 2c).
The selectivity was lower, only 73.8%, at 283 K which can be explained by the low reaction rate, and the persistence of the intermediates which are not converted to aniline (Fig. 3a). At 283 K, azoxybenzene and nitrosobenzene have been detected during the reactions, which indicates that, the hydrogenation process follows the Haber mechanism [38][39][40][41][42]. At higher temperatures the intermediates transformed to aniline (Fig. 3b and c). The catalyst was very selective towards the formation of aniline, by-products have not been detected. All in all, the prepared maghemite supported palladium catalyst at 303 K reaction temperature and 20 bar hydrogen pressure can be applied effectively for aniline synthesis.
The reaction rate constants (k) at different temperatures were calculated based on the measured nitrobenzene concentrations by using non-linear regression [43] ( Fig. 4; Table 1).    treatment and combustion have been used. This procedure leads to nanoparticles with smaller crystalline size (21.8 nm) and high adsorption capability. The catalyst is in an active form immediately after the production of the Pd/maghemite nanocomposite, as the sonochemical treatment initiated the involvement of the dispersion media in the reduction of palladium ions to elemental palladium particles (Pd 0 ). In this sense, the catalyst does not require further post-treatments, and it does not need to be reduced under a hydrogen atmosphere, therefore the catalyst preparation method is simplified. The synthesized magnetic catalyst was efficiently applied in nitrobenzene hydrogenation at 293 K and 303 K and the conversion was more than 99% in both case. The catalyst was selective towards aniline, and the selectivity was 97.0% and 96.7% at 293 K and 303 K, respectively. By-products were not detected during the reaction. All in all, a simple method has been designed for magnetic catalyst production. The achieved catalytic system is easily separable from the reaction media, thanks to its magnetic property and successfully applicable in nitrobenzene hydrogenation.