Manganese oxides in the bulk or supported are very important components in catalytic systems for the oxidation of carbon monoxide, benzene, methane, and other hydrocarbons, because of their high activity, durability, and low cost [16–19]. Many preparation methods and skills have been used to obtain active supported catalysts. Generally, Mn3O4 is synthesized by heating salts of manganese at about 1000 °C in air , co-precipitation , sol-gel methods , hydrothermal , solvothermal  microwave irradiation , and ultrasonic irradiation .
These preparation methods may affect the properties of the support which in turn may result in strong Pd nanoparticles support interactions, that strongly enhance the catalytic properties of the catalyst. So, It is very important to study the effects of support preparation method on the catalytic properties of Pd/Mn3O4 catalysts.
In the present work, we prepared Mn3O4 through simple chemical precipitation method at room temperature followed by hydrothermal treatment at 150°C. The structure-activity relationship of the Pd/ Mn3O4 materials are discussed in the light of a preliminary characterization of the physicochemical properties of the catalysts by surface area, X-ray diffraction (XRD), transmission electron microscopy (TEM), temperature programmed reduction (TPR). Catalytic activities of Pd/Mn3O4 catalysts were investigated by CO oxidation reaction under different reaction conditions.
2.1- Catalyst preparation
All chemicals were supplied from Sigma-Aldrich and used without further purification. Two methods were used to prepare Manganese oxides according to the following steps: A solution of Manganese (II) nitrate nonahydrate was prepared by dissolving 5 g of the salt in 50 ml deionized water. During stirring, drops of 10 N NaOH solution was added till dark brown precipitate was formed and the pH of the resulting solution becomes 12. At this step of preparation, the resulting brown gel was divided into two parts, the first part (represented as Mn3O4) of the brown gel was stirred overnight, and the pH was kept at 12 by adding drops of 10 N NaOH if required. Then, the gel was filtered, washed with deionized water dried at 80 oC under vacuum. The second part of the brown gel (represented as Mn3O4 H) was stirred for 1h (at pH 12) and transferred into in a 150 ml Teflon-sealed autoclave and heated up (heating rate 5 °C /min.) to 150°C and maintained at 150°C for 24 h. Finally, the resulting material was recovered by filtration, rinsed by deionized water several times, and dried at 80 oC under vacuum. The formation of Mn3O4 compound during the preparation in the alkaline aqueous solution may be formulated by the following reactions:
Mn2++ 2OH− → Mn(OH)2, (1)
Mn(OH)2 → MnO + H2O, (2)
3 MnO +1/2O2 → Mn3O4. (3)
Manganese (II) ions are firstly reduced in alkaline media to Mn(OH)2 and then decomposed into MnO (see (1) and (2)). Finally, MnO is oxidized to Mn3O4 by the atmospheric oxygen (see (3)).
Pd nanoparticles supported Mn3O4 was prepared [27, 28] according to the following steps: 2g of dried Mn3O4 was suspended in 50 ml of distilled water in a glass beaker to which required quantity of Palladium(II) nitrate solution was added drop wise. The mixture was stirred for 3h after that, 200μl of hydrazine hydrate was added as a reducing agent during stirring of the mixture. Then, the mixture was placed inside a conventional microwave. The microwave oven operated at full power (1000 W), 2.45 GHz, in 30-s cycles (on for 10 s, off and stirring for 20 s) for a total reaction time of 90 s. Then, the sample left to dry overnight under vacuum at 80 oC.
2.2- Catalyst characterization
The actual content of Pd in the prepared catalysts was determined by inductive coupled plasma (ICP). The analysis was carried out with Varian (Vista-MPX) Charge Coupled Device (CCD) simultaneous ICP-OES instrument. The analysis was performed using ca. 2.5 mg of sample dissolved in 5 ml of aqua regia, made up to 100 ml with deionized water.
2.2.2. X-ray diffraction
X-ray powder diffraction patterns of samples were determined using an X’Pert Philips Materials Research Diffractometer. The patterns were run with copper radiation (Cu Kα, λ = 1.5405 Å) with the second monochromator at 45 kV and 40 mA with a scanning speed of 2° in 2θ/min.
2.2.3. Textural properties
The adsorption isotherms and the specific surface area (SBET) of the various catalysts were determined from nitrogen adsorption studies conducted at -196 °C using a Quantachrome Autosorb-1-C system. The samples were previously degassed at 100 °C for 2 h.
2.2.4. H2-TPR profiles
The redox properties and redox reversibility of the samples were determined by the temperature programmed reduction (H2-TPR) using Quantachrom Nova Sorbimetric system. During the experiment, 5% H2/He gas mixture with a flow rate of 36 mL min-1 was used for the (H2-TPR). The temperature of the system was increased from room temperature to 800oC at heating rate of 10oC/min. During the experiment, the amount of consumed H2 during the reduction was measured based on the analysis using mass spectrometry.
2.2.5. TEM images
Transmission electron microscope (TEM) images and the particle size were obtained using a Jeol JEM-1230 operated at 120 kV. For TEM images the sample powder was dispersed in methanol by using ultrasonic radiation for 10 min, and a drop of the suspension was placed onto the carbon-coated copper grids.
2.2.6. XPS measurements
XPS measurements were carried out on a SPECS surface analysis systems operating at a base pressure of 4 x 10-10 mbar. A dual anode non-monochromatic Mg-Kα (1253.6 eV) X-ray source was used to irradiate the sample surface with 13.5 kV, 150 W of X-ray power. The sample was positioned at a take-off angle of 90° between the sample surface and the direction of photoelectrons detected by the analyzer. XPS data were acquired utilizing a PHOIBOS 150 MCD-9 hemispherical energy analyzer operating in Fixed Analyzer Transmission (FAT) mode. The analyzer was set to the Medium Area (MA) lens mode with a rectangular entrance slit of 7 x 20 mm2. The high-resolution scans of the elements were recorded at pass energy of 20 eV, energy step of 50 meV and dwell time of 300 msec. As the standard practice in XPS studies, the adventitious carbon C1s at 285 eV corresponding to C-C bond has been used for all XPS spectra as binding energy reference for charge correction. The spectrometer was calibrated to the position of the 3d5/2 line of sputter-cleaned silver sample with binding energy 368.26 eV. For Mg-Kα excitation, the energy resolution or FWHM (Full Width at Half Maximum) at low pass energies for the Ag 3d5/2 level was 0.85 eV.
The deconvolution of the XP peaks and the evaluation of the spectra were performed using the CasaXPS software package version 2.3.16. The Shirley method and the Gaussian Lorentzian (GL) function were used for the background subtraction and peak fitting procedures.
2.3- CO oxidation experiment
The catalytic oxidation of CO by oxygen was carried out as follows: 0.05 g of the catalyst was charged in quartz tube reactor with an inner diameter of 6 mm. The reactor was placed inside a GSL-1100X-NT programmable tube furnace reactor. The tube furnace temperature was adjusted to increase from room temperature with a heating rate 10 oC/min. to the required temperature. The sample temperature was measured by a thermocouple placed near the sample. The reactant gas used was a mixture of 4 wt% CO and 20 wt% O2 in He. The gas mixture was set to flow over the sample at a rate of 50 cc/min controlled via Precision Mass Flow Controller (ZF-MFC-1X) digital flow meters, corresponding to a space velocity of 15,000 ml h−1gcat−1. The conversion of CO to CO2 was monitored using an infrared gas analyzer (MGA 3000 multi gas analyzer, ADC). All the catalytic activities were measured after a heat treatment of the catalyst at 110 oC in the pure He gas for 15 min to remove moisture and adsorbed impurities.
Results and discussion
3.1. ICP analysis
The actual composition ratios of Pd nanoparticles in all catalysts prepared by different methods were determined by ICP-OES technique. The results of ICP analysis are listed in Table 1, the results indicated that, the determined amount of Pd nanoparticles are very close to calculated amounts. This clearly indicates that most of the Pd ions in the aqueous solution had been deposited into or on the surface of Mn3O4 support.
3.2. X-ray diffraction
The crystalline structures of all prepared catalysts were characterized by X-ray diffraction (XRD) analysis, the results are presented in Fig. 1. Careful inspection of Fig. 1 reveals that, pure Mn3O4 prepared by different methods exhibit identical reflection peaks in the XRD pattern (Fig. 1a, b). these reflection peaks could be perfectly indexed to tetragonal Mn3O4 (Hausmannite), moreover, the lattice constants, a=b= 5.758 Å and c = 9.446 Å, were also in good agreement with those of pure Mn3O4 (PDF card no. 01-089-4837). These results also suggested that, the preparation method had no effect on the crystalline structure of tetragonal Mn3O4. On the other hand, samples loaded with Pd nanoparticles exhibited diffraction peaks similar to pure support. This may indicate that the addition of Pd nanoparticles was not accompanied with any crystalline changes of manganese oxide tetragonal phase. Furthermore, no distinct metallic or ionic Pd reflections were observed in the patterns of samples (Fig. 1c-d), which may due to high dispersion of Pd nanoparticles in/on the support surface or its size is small to be detected by XRD technique. No diffraction peaks of any other phases of manganese oxides were detected, indicating the purity tetragonal Mn3O4 prepared.
3.3. surface area measurements
N2 adsorption–desorption at -196 oC technique was used to study the textural of the catalysts. The adsorption -desorption isotherms measured at -196 oC over some selected samples are displayed in Fig.2a. Moreover, the values of specific surface areas (SBET m2/g), mean pore radius (Ǻ) and total pore volume (cc/g) are listed in Table 1. Inspection of the adsorption-desorption isotherms shown in Fig.1a indicate that, all the samples exhibited type II adsorption isotherm in the IUPAC classifications, with a H3 type hysteresis loop, indicate their mesoporous structure nature and formation of slit-like mesopores [6, 29]. The isotherms of these catalysts have a closed hysteresis loops and desorption branches joints the adsorption ones at relative low pressure of about 0.15-0.25. The pore size distribution curves (Fig. 1b) of pure supports exhibit one narrow peak centered at around 12 Ǻ. Noticeably, after loading 2or 4 wt% of Pd, the pore size increases with the rise of Pd content. The pore radius increased to around 19 Ǻ in 4wt% Pd samples prepared by different methods. Moreover, limited variations in the total pore volume values even after loading of Mn3O4 with Pd nanoparticles (Table 1).
The specific surface areas of the samples are listed in Table 1. The results showed that, no significant difference in the specific surface areas of pure Mn3O4 prepared by various methods. On the other hand, the values of specific surface areas of Pd supported manganese oxide samples (prepared by different methods) are found to decrease by a moderate amount after loading of Pd nanoparticles. The surface area was found to decrease continuously as the Pd content increase. The observed decrease of surface area and pore volume with the increasing Pd loading, may be attributed to the partial blockage of the pores by Pd nanoparticles or the fine distribution of Pd nanoparticles into the pores and/or on the surface of the support. Similar effect of Pd loading on the specific surface area of the different supports was previously observed [30, 31]. Zou et al. , found that, addition of a small amount of Pd nanoparticles into MnOx-CeO2 catalyst brings about a little decrease in the specific surface area.
3.4. H2-TPR profiles
In order to study the reducibility of Mn3O4 and the effect of Pd nanoparticles deposition in different samples, the H2-TPR test was performed. Figure 3 shows comparison between H2-TPR profiles of Mn3O4 before and after 4wt% Pd deposition prepared by various methods. As it can be seen in the results presented in Fig. 3a, b, the reduction profiles of pure support show almost one narrow reduction peak at 620 and 632 oC for Mn3O4 and Mn3O4 H (sample subjected to hydrothermal treatment) respectively. The high temperature reduction may indicate that, pure Mn3O4 or Mn3O4 H are well crystalline, and the nature of oxygen is uniform in both samples. Stobbe et al.  reported that, the reducibility of Mn3O4 by hydrogen is affected by preparation method and sample crystallinity. The results also suggest that, the lattice oxygen of Mn3O4 should have relatively higher activity than Mn3O4 H . Deposition of 4wt% Pd nanoparticles on Mn3O4 and Mn3O4 H surfaces resulted in the appearance of new low temperature reduction peaks (as seen in Fig. 3a, b). Compared to Pure supports which exhibit high temperature reduction peak, 4wt% Pd supported samples exhibit two lower temperature peaks [34-36]. The first lower temperature reduction peak is attributed to the reduction of Mn4+ to Mn3+ at around 180 or 205 oC for 4Pd Mn3O4 and 4pd Mn3O4H respectively. The other reduction peak which is attributed to the reduction of Mn3+ to Mn2+ appears at around 362 and 410 oC for 4Pd Mn3O4 and 4pd Mn3O4H respectively. On the other hand, the sharp main reduction peak of pure supports at higher temperatures are shifted to lower temperatures (432, 570 oC) with very low intensities as seen in Fig. 3b.
The comparison of the reduction profiles of pure and Pd supported catalysts explain the role played by Pd nanoparticles in enhancing the reducibility of the support. The H2-TPR profiles comparison reveals that, the presence of Pd nanoparticles strongly promotes the reduction process of the support as reflected by the significant shift of the main reduction peak to lower temperatures. This low temperature shift may be attributed to the presence of spillover effect involving either hydrogen activated on the metal phase or mobile lattice oxygen induced by intimate metal-support interactions [37, 38]. Moreover, the temperatures shift of the main reduction peaks in Mn3O4 and Mn3O4 H due to Pd nanoparticles loading may indicate that, the reducibility of Pd/Mn3O4 is much higher than Pd/Mn3O4 H.
3.5. TEM images
The morphologies and particle sizes of pure and Pd loaded Mn3O4 and Mn3O4 H catalysts have been investigated using TEM micrographs (Fig. 4a-f). Variety of morphologies mainly including cubic, spheroid and plate shaped have been observed in pure Mn3O4 and Mn3O4 H nanoparticles. The majority observed shape is cubic shape in all samples with some smaller spheroid shaped particles adsorbed on the cube surface. The average particles size for the samples was found to be around 38 nm for all samples prepared by various methods. Moreover, Pd nanoparticles supported on Mn3O4 and Mn3O4 H cannot be distinguished in all images, suggesting that Au nanoparticles have been highly dispersed into or on the surface of Mn3O4 and Mn3O4 H supports.
3.6. X-ray photoelectron spectroscopy (XPS) analysis
Photoelectron spectroscopy was used to gain further information of the catalyst structure, in particular the chemical state of the elements and their relative abundance at catalyst surfaces. Two representative catalysts 4Pd/Mn3O4 and 4Pd/Mn3O4 H (which exhibited the most interesting results for CO oxidation) were selected for this purpose, and the core level spectra recorded included Pd 3d, Mn 2p and O 1s. The binding energies of core electrons and surface atomic ratios for catalysts are summarized and listed in Tables 2. Figure 5a compares the Pd 3d 5/2 and 3d 3/2 XPS spectra of the two samples. The investigation of Pd 3d XPS spectra of two catalysts exhibits the binding energy of Pd 3d5/2 (low-energy band) and Pd 3d3/2 (high-energy band). The low energy Pd 3d5/2 is deconvolved into doublets at 335.5 and 337.8 eV (Table 2) which are assigned to Pd0 and Pd2+ for both samples. Moreover, the peak deconvolution of the high energy band Pd 3d3/2 gives two peaks at 340.7 and 343.1 eV which are ascribed to Pd0 and Pd2+ (PdO) [39, 40]. The XPS results shows that both samples contain mixture of Pd0 and Pd2+ on Mn3O4 surface with different ratios as listed in Table 2. The % atomic concentrations of Pd0 and Pd2+ for both samples (Table 2) reveals that, 4Au/Mn3O4 contains higher surface concentrations of Pd0 and Pd2+compared to 4Au/Mn3O4H sample. Obviously, the Pd0/Pd2+ ratio (Table 2) is found to be 0.073 and 0.046 for 4Au/Mn3O4 and 4Au/Mn3O4H samples respectively. It is also clear that the binding energy Peak shift caused by the strong metal–metal oxide interaction is not observed in XPS-Pd 3d since both samples exhibited the Pd 3d peaks at almost similar binding energies (Table 2).
Figure 5b shows the XPS spectra corresponding to Mn 2p3/2 and Mn 2p1/2 peaks. As it can be seen, both samples showed similar XP spectrum for Mn 2p3/2 and Mn 2p1/2. The binding energy values of Mn 2p3/2 and Mn 2p1/2 are 641.3 eV and 653.2 eV, respectively, which are characteristic for Mn3O4 [41, 42]. Moreover, it is clear that, there is a spin–orbit splitting of 11.9 eV between the binding energy values of the Mn 2p3/2 and Mn 2p1/2 for both samples, which is in good agreement with the energy splitting reported for Mn3O4 [43- 45]. The XPS results confirm the formation of Mn3O4 in both samples, which also agree well with the XRD results.
Moreover, for each sample, three kinds of oxygen species at ca. 528 eV, 529 eV and 532 eV were observed in deconvoluted the O 1s spectrum, as shown in Fig. 5c. The peak at ca. 528 eV was assigned to lattice oxygen ions (O2-) in Mn3O4 . The other two peaks at ca. 529 eV and 532 eV in both samples were attributed to surface hydroxyl groups and adsorbed water molecules respectively [47-49]. These results confirm the formation of oxygen vacancies on the surface of the catalysts which, however, are filled by water or hydroxyl groups. Similar XPS for O1s spectra were observed for many other supports [50-52].
From XPS spectra and based on the % atomic concentrations and binding energy values listed in Table 2, it is obvious that, both of Pd0 and Pd2+ species coexist on Mn3O4 surface. Moreover, the total Pd (Pd0 and Pd2+) to Mn ration is 0.010 and 0.014 for 4Au/Mn3O4 and 4Au/Mn3O4H samples respectively, which may support the idea of better dispersion of Pd0 and Pd2+ on Mn3O4 rather than Mn3O4H surface.
3.6. Catalytic oxidation of carbon monoxide.
The catalytic performance of pure and Pd supported Mn3O4 and Mn3O4 H toward catalytic oxidation of CO by O2 are shown in Fig. 5a, b, where the CO conversion percent is represented as a function of reaction temperature (light-off test). The temperatures at which the conversion percent of CO attain 50% (T50) and 100% (T100) are used as standard to compare the catalytic activities of all prepared catalysts. All the values of T50 and T100 are listed in Table 1.
As shown in the figure, pure supports Mn3O4 and Mn3O4 H do not show any catalytic activity below ~ 175 oC. When the reaction temperature exceeds 175 oC, the catalytic oxidation of CO starts and increases as a function of temperature (Fig. 5a). The values of T50 and T100 are 205 and 245 oC for Mn3O4, while for Mn3O4 H the values of T50 and T100 are 258 and 347 oC respectively (Table 1). These results indicate that manganese oxide subjected to hydrothermal treatment (Mn3O4 H) exhibits relatively lower catalytic activity toward CO oxidation than the other manganese oxide sample (Mn3O4). These values of T50 and T100 are agree with other reported values for CO oxidation reaction over pure manganese oxide catalyst .
Moreover, the catalytic performance of the supported Pd supported catalysts are notably higher than those for the corresponding parent support materials. When 2wt% of Pd is deposited on Mn3O4 and Mn3O4 H remarkable enhancement of the catalytic activities toward CO oxidation are observed and the values of T50 and T100 are shifted to lower values as listed in Table 1. Further increase in Pd content up to 4 wt.%, the CO oxidation activities greatly improved. This improvement of the catalytic activities is also accompanied with more shift of T50 and T100 to lower temperatures. Based on the values of T50 and T100 listed in Table 1 and Fig 5b, deposition of Pd nanoparticles on pure Mn3O4 and Mn3O4 H is accompanied with great enhancement in CO oxidation activity. Moreover, the enhancement of the catalytic activity is proportional to Pd content. Also, the Pure and supported Pd nanoparticles on Mn3O4 exhibit higher catalytic activities compared to Mn3O4 H samples. Obviously, among all Pd nanoparticles samples, the lowest T50 and T100 are achieved for 4Pd/ Mn3O4, which indicates that this sample exhibits the best catalytic performance in CO oxidation.
The mechanism of catalytic oxidation of CO is affected by many factors. Li et al.  reported that the mechanism of CO oxidation involves the adsorption of CO over Pd0 nanoparticles to form active CO+ species. Then the activated CO+ species reacts with lattice oxygen to give CO2 on the interface between Pd and Mn3O4 leaving O vacant on the support surface. The O vacant quickly adsorbs oxygen from the gas phase to compensate the support. Moreover, Liu et al.  found that, highly dispersed Pd nanoparticles over Pd/FeOx as well as the support ability of supplying active oxygen were the main reasons for the excellent performance of Pd/FeOx for in co-oxidation of CO and H2. The metallic Pd0 sites has been reported to act as catalytically active site for CO oxidation reaction [14, 55-56]. On the other hand, other authors proposed that the thin film of PdO oxide at the Pd0-support interface can be the catalytically active phase in CO oxidation [57, 58]. Hinojosa et al. found that, PdO adsorbs oxygen to form adsorbed species which are very active in CO oxidation .
In the present work, based on H2-TPR results, great enhancement of the reducibility of Mn3O4 after Pd nanoparticles deposition. The dispersion of both Pd0 and PdO on the support surface (as evident by XPS) may facilitate the transfer of oxygen through Mn3O4 surface. The XPS results reveals that, the total Palladium (Pd0 and Pd2+) to Mn and Pd0 to Pd2+ ratios on the surface of Mn3O4 are higher than Mn3O4 H. the better dispersion of Pd0 and Pd2+ on the Mn3O4 surface may account for the higher reducibility and better catalytic activity in CO oxidation. These results may indicate that, the formation and dispersion of Pd0 nanoparticles is very important factor for low temperature CO oxidation. Metallic Pd nanoparticles provided active sites for CO gas adsorption, the adsorbed CO reacts with active oxygen from the support to give CO2. Moreover, presence of Pd2+ made the catalyst more reducible, that is, supplied oxygen more easily to reveal higher activity for oxidation of CO at low temperature.
3.7. Kinetic study
The kinetic study of CO catalytic oxidation is carried out according to the following first- order rate equation [59-61]:
Where X is the amount of CO converted, W is the catalyst weight used (g catalyst), F is the flow rate of reactants gas mixture (50 ml/min), and k is the reaction rate constant (ml/g catalyst min.). During this experiment, all the reaction parameters such as flow rate, heating rate,…etc are kept constant, the weight of the catalyst is the only change. The kinetic study has been monitored by using four different catalyst weights. On plotting -ln(1-CO converted) versus the catalyst weights a straight line is obtained with a slop equal the reaction rate constant. The obtained values of the reaction rate constant at different temperatures are plotted using Arrhenius equation (Fig.6), the values of the apparent activation energies Ea are calculated and listed in Table 1. The computed activation energies Ea for all catalysts obtained from the Arrhenius plots were found to change from 16.96 kJ/mol to 24.03 kJ/mol (Table 1). Moreover, the activation energies Ea of all Pd deposited samples are lower than pure supports. The sample with higher catalytic performance 4Pd/ Mn3O4 exhibits the smallest value of activation energy (16.96 kJ/mol).
The stability of the prepared catalysts was tested over 4 wt.% Pd deposited on Mn3O4 and Mn3O4H at 20 and 75 oC respectively. In the stability test, the samples were kept under continuous stream flow, all the reaction conditions such as flow rate, reaction temperature, ….etc were kept constant. The results are shown in Fig. 7. The results reveal that, the selected catalysts exhibited good stability for CO oxidation for around 30 h. The results show the prepared catalysts exhibited good catalytic stability during to CO oxidation without remarkable decrease in activity.
In this work, Nanostructured Mn3O4 prepared by two methods followed by deposition of different amounts of Pd nanoparticles were successfully prepared. The results show that, when Mn3O4 was exposed to hydrothermal treatment, the catalytic performance toward CO oxidation decreased. Moreover, the performance to CO oxidation was found to increase with the rise of Pd content. H2-TPR experiment showed great enhancement in the reducibility of the support due to Pd nanoparticles deposition. Moreover, presence of high dispersion of metallic Pd and Pd2+ nanoparticles on the support surface (as evident by XPS) may be responsible for the activity of the catalyst towards CO oxidation. Moreover, the results showed good stability of the catalysts during the CO oxidation without remarkable decrease in activity. The kinetics of catalytic oxidation of CO is carried out according to the first- order rate equation.