Unloaded CeO2 and nominal 0.50, 1.00, 1.50, 2.00, 5.00, and 10.00 mol% Fe-loaded CeO2 nanoparticles were synthesized by flame spray pyrolysis (FSP). The samples were characterized to obtain structure-activity relation by X-ray diffraction (XRD), high-resolution transmission electron microscopy (HRTEM), Brunauer, Emmett, and Teller (BET) nitrogen adsorption, X-ray photoelectron spectroscopy (XPS), and UV-visible diffuse reflectance spectrophotometry (UV-vis DRS). XRD results indicated that phase structures of Fe-loaded CeO2 nanoparticles were the mixture of CeO2 and Fe2O3 phases at high iron loading concentrations. HRTEM images showed the significant change in morphology from cubic to almost-spherical shape observed at high iron loading concentration. Increased specific surface area with increasing iron content was also observed. The results from UV-visible reflectance spectra clearly showed the shift of absorption edge towards longer visible region upon loading CeO2 with iron. Photocatalytic studies showed that Fe-loaded CeO2 sample exhibited higher activity than unloaded CeO2, with optimal 2.00 mol% of iron loading concentration being the most active catalyst. Results from XPS analysis suggested that iron in the Fe3+ state might be an active species responsible for enhanced photocatalytic activities observed in this study.

1. Introduction

Organic compounds from industries are one of the major causes of water pollution [1]. Various strategies have been employed to remove these toxic compounds [2, 3]. One of the most interesting approaches is heterogeneous photocatalysis because the process is based on the use of solar energy, which is clean and abundant in nature [4, 5]. In the recent years, cerium dioxide (CeO2 or ceria) has received considerable attention because this material shows promising applications in solid oxide fuel cells [6], environmental catalysis [7, 8], redox catalysis [9], and wet catalytic oxidation of organic pollutants [10]. However, the band gap of CeO2 (3.22 eV) has limited the activation of solar energy; only UV light can be applied to generate electron-hole pairs at the beginning of photocatalytic processes. Thus, it is necessary to extend the absorbance of CeO2 into visible region and reduce the electron-hole pairs recombination [11, 12]. There are many methods to modify light absorption properties of CeO2, such as metal doping [13, 14], surface sensitization [15], and coupling with semiconductor that has smaller band gap [16]. Recently, transition metal doping/loading has been widely used to enhance the light absorption of CeO2 [17, 18]. It has been reported in many works of literature that the metal ions of Pt [19], Ag [20], Fe [21], Mn [22], Co [23], Ni [24], and Zn [25] in CeO2 could improve CeO2 photocatalytic activity towards the visible-light region. Among these metals, Fe has been considered as a candidate owing to its special Fenton reaction of iron. The Fenton process can improve the photocatalytic activity by producing the hydroxyl radicals (OH) which are very powerful oxidizer in photocatalytic process [26]. There are many methods to prepare unloaded CeO2 and Fe-doped/-loaded CeO2 nanoparticles such as sol-gel [27], sonochemical [28], homogeneous precipitation [29], hydrothermal [30], microemulsions [31], surfactant-assisted precipitation [32], and flame spray pyrolysis (FSP) methods [33].

Among them, the latter one is a promising approach because FSP can produce the nanoparticle products with particle size in the range of 1–200 nm at high production rates up to 250 g/h in one step [34]. Other advantages are the ability to dissolve the precursor directly in the fuel and the simplicity of introduction of the precursor into the hot flame zone. In addition, the process of loading/doping metal oxide with metals can easily be done by adding dopant in the precursor solution [35, 36]. In the present work, unloaded CeO2 and Fe-loaded CeO2 nanoparticles were directly synthesized by FSP method. The formic acid and oxalic acid were chosen as model organic pollutants for photocatalytic study under visible-light irradiation.

2. Experimental

2.1. Preparation of Powders

The precursor solutions for FSP consisted of cerium nitrate hexahydrate (Sigma-Aldrich, 99.99 wt%) and iron acetyl acetonate (Sigma-Aldrich, 97 wt%). The cerium precursor was dissolved in absolute ethanol (Scharlau, 98%) to obtain a 0.50 M  concentration. Amounts of Fe loading concentration were varied as 0.50, 1.00, 1.50, 2.00, 5.00, and 10.00 mol% in order to prepare Fe-loaded CeO2 samples. The precursor mixtures were fed into the center of flame by syringe pump with a rate of 5 mL/min and dispersed by 5 L/min oxygen according to the previous report [37]. Then, the liquid precursor was dispersed quickly in an upward direction by gas stream and ignited by premixed oxygen/methane flame. The gas flow rates of oxygen and methane-supporting flame were set as constant rates of 1.19 and 2.46 L/min, respectively. After evaporation and combustion of liquid precursor droplet, nanoparticle products were collected on a glass microfiber filter papers (Whatmann GF/A, 25.7 cm in diameter) with a vacuum pump controller.

2.2. Characterization of Nanoparticles

The phase and crystallinity of the synthesized samples were analyzed by X-ray powder diffraction (XRD; Philips X’Pert MPD; CuKα radiation). The most intense peak corresponding to (111) plane was chosen to calculate the crystallite sizes () using Scherrer equation as follow: where is a constant equal to 0.89, is the X-ray wavelength equal to 0.154 nm , is the full width at half-maximum (FWHM), and is the half-diffraction angle [38]. The chemical composition and oxidation state of material were studied by X-ray photoelectron spectroscopy (XPS) using Mg X-ray source (MgKα, Kratos Axis Ultra DLD). The binding energy of the adventitious carbon (C 1 s) line at 285 eV    was used for calibration, and the position of other peaks was corrected according to the position of the C 1 s signal. High-resolution transmission electron microscopy (HRTEM, JEOL JEM-2010) was employed to determine the morphology of prepared samples. The mean particle size and specific surface area (SSA) were investigated using the Brunauer, Emmett, and Teller (BET) nitrogen adsorption method (QuantachromeAutosorp 1 MP). The reflectance spectra of the nanoparticle powders were obtained by using UV-visible diffuse reflectance spectrophotometry (UV-vis DRS) equipped with integrating sphere detector (Shimadzu, UV-3101PC).

2.3. Photocatalytic Activity

1.00 g /L of photocatalyst suspensions was prepared in deionized water and circulated in closed system spiral photoreactor. In a typical run, carbon burn-off step was firstly carried out by illuminating the photocatalyst suspension with a UV-A lamp (Sylvania Blacklight Blue, 18 W) in order to remove any organic impurities from the photocatalyst. The photocatalytic activities were evaluated through formic acid (Sigma-Aldrich, 98 wt%) and oxalic acid (Sigma-Aldrich, 99.99 wt%) degradations under the visible irradiation for 120 min. Finally, the generated carbon dioxide (CO2) was measured using the conductivity meter (Eutech Instruments Cyberscan PC5500, μS/cm2 precision). At the end of each photocatalytic experiment, the recorded data presented the increase in conductivity value. In order to calculate amounts of generated CO2, the values were converted from conductivity at that time to the amount of carbon by the interpolated from calibration curve.

3. Results and Discussion

3.1. X-Ray Powder Diffraction (XRD)

In Figure 1, the X-ray diffraction pattern has been used in order to study the structure and phase composition of the prepared samples. It can be seen that all samples had similar diffraction patterns of cubic fluorite structure of ceria (JCPDS 340394) [39]. However, the sample with high amount of iron loading (2.00, 5.00, and 10.00 mol%) exhibited the mixed phase of CeO2 and Fe2O3 (JCPDS 330664) [40]. The XRD peaks of all samples were magnified as shown in Figure 2. It was found that CeO2 peaks shifted towards higher 2 upon increasing iron content. The calculated d-spacing, lattice parameter, unit cell volume, and average crystallite size were also decreased as shown in Table 1. These observations could be ascribed to partial substitution of Ce4+ ions (0.101 nm) by Fe3+ ions (0.064 nm) [41]. A decrease of unit cell parameters due to the substitution of larger ion by the smaller one was also found in previous reports [42, 43].

3.2. High-Resolution Transmission Electron Microscopy (HRTEM)

As seen from Figure 3(a), the unloaded CeO2 clearly showed the cubic morphology of cubic fluorite CeO2 structure. In Figure 3(b), the particles became more spherical upon loading CeO2 with iron. This change in CeO2 morphology might be due to the incorporation of iron ions in CeO2 lattice, thus affecting the particle growth and causing lattice deformation [44]. This assumption was supported by the shift of XRD peak and the changes of lattice parameters as reported in Table 1. The average particle sizes as seen from HRTEM image were about 6–8 nm. This was in good agreement with the calculated sizes obtained by using the Scherrer equation. Figure 3(c) shows the lattice fringes of 2.00 mol% Fe-loaded CeO2. The lattice planes with -spacing of 0.16 and 0.20 nm were attributed to the (311) and (220) planes of cubic fluorite CeO2, respectively, whereas the plane with -spacing of 0.24 was assigned to the (110) plane of Fe2O3. These results confirmed the presence of mixed phase between CeO2 and Fe2O3 in the nominal 2.00 mol% Fe-loaded CeO2 as found previously in the XRD patterns (Figure 1).

3.3. Nitrogen Adsorption-Desorption Isotherms

The specific surface areas (SSA) of different samples were analyzed by Brunauer-Emmett-Teller (BET) method based on the nitrogen adsorption/desorption isotherm. The mean BET diameter () was also calculated by using the following equation [45]: where is the BET-specific surface area and is the density of the CeO2 (7.32 g/mL). As shown in Table 2, an increase of surface area accompanied with a decrease of BET diameter was clearly observed upon increasing iron content. This increased surface area would be beneficial to the efficient photocatalytic performance due to high surface adsorption of organic pollutants. The calculated BET diameter was in the range of 5–7 nm which was very well in agreement with those obtained by using the Scherrer equation (Table 1).

3.4. UV-Visible Spectroscopy

UV-vis reflectance analysis was performed by converting the obtained reflectance spectra (Figure 4(a)) to the Kubelka-Munk absorbance spectra using the Kubelka-Munk equation as follows [46]: where and are the Schuster-Kubelka-Munk absorbance and the absolute reflectance of the sample, respectively. The plot of absorbance against wavelength for the CeO2 nanoparticle powders is shown in Figure 4(b).

The spectra showed that the absorption edge shifted to longer wavelength upon increasing the iron loading concentration. Band gap energies of the obtained samples can then be determined by using the intercept of the tangent to the graph plotting between the Kubelka-Munk absorption function and photon energy () as shown in Figure 4(c) [47, 48]. The obtained band gap energies () as reported in Table 2 decreased with increasing iron loading concentration.

3.5. Photocatalytic Activity

The photocatalytic activity of unloaded and Fe-loaded CeO2 was evaluated by degradation of formic and oxalic acids. The effects of different iron loading concentrations on the photocatalytic efficiency of CeO2 nanoparticles were evaluated under visible-light irradiation for 120 min, and the results are presented in Figure 5. According to the results, the photocatalytic activities of Fe-loaded CeO2 nanoparticles were significantly higher than those of unloaded CeO2 nanoparticles. This improved photoactivity could be partially ascribed to the enhanced light absorption in visible-light region as observed from the UV-vis study in Figure 4. However, the activity was clearly dependent on the amount of iron loading. The results demonstrated that the nominal 2.00 mol% was an optimal iron concentration for photocatalytic activity of CeO2 nanoparticles in this research. On the other hand, 5.00% and 10.00 mol% iron concentrations showed poor photocatalytic activity, probably because high iron concentration tended to cover CeO2 surface, thus preventing light from contacting the CeO2 surface [49].

Another possible reason was that too high iron loading can act as the electron-hole recombination centers instead of the trapping level, resulting in a decreased photocatalytic activity [47, 50]. The kinetic data for formic and oxalic acids degradations under visible-light illumination were found to follow pseudo first-order reaction [51] as shown in Figure 6. The pseudo first-order model is explained by where is the apparent rate constant (min−1), means the initial concentration of acid, and refers to the concentration of acid at various reaction times (). The determined pseudo first-order rate constants (, min−1) are presented in Table 3. It can be seen that the loading of 2.00 mol% iron in CeO2 nanoparticles could remarkably improve the apparent rate constant up to 5 times for formic acid and 3 times for oxalic acid compared with the unloaded one.

In order to investigate the effect of surface area on the degradation activity, the surface-area-normalized degradation values against visible-light irradiation time were plotted as shown in Figure 7, and the calculated surface-area-normalized rate constants are presented in Table 3. The results clearly suggested that surface-area of the catalyst has a crucial impact on the activity of acid degradation in this study because the surface area-normalized rate constants are significantly decreased from the original values. However, other factors such as band gap energy, amount of Fe loading, sample crystallinity, and phase composition [52] could not be neglected as these could also contribute to the difference in photocatalytic activity of the catalysts being studied.

3.6. X-Ray Photoelectron Spectroscopy (XPS)

In order to characterize the valence state of cerium and iron in 2.00 mol% Fe-loaded CeO2, X-ray photoelectron spectroscopy (XPS) was carried out as shown in Figure 8.

According to Figure 8(a), the peaks at 710.6 and 723.4 eV assignable to the core level of 2p3/2 and 2p1/2, respectively, corresponded to Fe3+ species in Fe2O3 [53, 54]. No other peaks due to Fe0 and Fe2+ were found in the XPS results. From the Ce 3d XPS spectrum, the binding energies of all peaks are shown in Figure 8(b). These peaks corresponded to the three pairs of spin-orbit doublets assignable to Ce4+ valence state which were very well in agreement with the previous reports [55, 56].

4. Conclusions

Fe-loaded CeO2 nanoparticles with different iron loading concentrations have successfully been synthesized by flame spray pyrolysis (FSP). Loading CeO2 with Fe3+ resulted in a decrease of d-spacing, lattice parameter, unit cell volume, and crystallite size but an increase of BET surface area. The UV-vis absorption spectra displayed a red shift in the band edge transition upon increasing of iron loading concentration. XPS analysis showed the presence of Fe3+ species on the surface of CeO2. This could be attributed to the presence of Fe2O3 as observed from the XRD and HRTEM analyses. Increased photocatalytic activity compared with unloaded CeO2 was clearly obtained from the Fe-loading sample. It was found from this study that the nominal 2.00 mol% was an optimum iron loading concentration, giving the highest photocatalytic activity. Band gap energy and surface of the catalyst were found to be important factors affecting the photocatalytic activity observed in this study. However, other factors such as amount of Fe loading, sample crystallinity, and phase composition could not be neglected.


This work has been supported by Thailand Research Fund (TRF) through the Royal Golden Jubilee (RGJ) Ph.D. Program. The National Research University Project under Thailand’s Office of Higher Education Commission; the Materials Science Research Center, Department of Chemistry, Faculty of Science; and the Graduate School, Chiang Mai University, are greatly acknowledged.