*2.1. Characterizations of Bare CeO<sup>2</sup> and Grey (Modified) CeO<sup>2</sup>*

The first difference between the bare CeO<sup>2</sup> and the CeO<sup>2</sup> exposed to the solar irradiation (for 3 h) in H<sup>2</sup> flow at room temperature, is the change of the powder color. Figure 1 reports two photos of the synthetized materials: un-modified CeO<sup>2</sup> appears yellow in color, while modified CeO<sup>2</sup> is grey. This latter sample is coded as "grey CeO2". Interestingly, we have noted that only the contemporaneous treatment with solar irradiation and H<sup>2</sup> flow could obtain the grey CeO2, whereas each single treatment alone did not alter the structural and chemical properties of bare CeO2. *2.1. Characterizations of Bare CeO<sup>2</sup> and Grey (Modified) CeO<sup>2</sup>* The first difference between the bare CeO<sup>2</sup> and the CeO<sup>2</sup> exposed to the solar irradiation (for 3 h) in H<sup>2</sup> flow at room temperature, is the change of the powder color. Figure 1 reports two photos of the synthetized materials: un-modified CeO<sup>2</sup> appears yellow in color, while modified CeO<sup>2</sup> is grey. This latter sample is coded as "grey CeO2". Interestingly, we have noted that only the contemporaneous treatment with solar irradiation and H<sup>2</sup> flow could obtain the grey CeO2, whereas

each single treatment alone did not alter the structural and chemical properties of bare CeO2.

**Figure 1.** Photo of the as-synthetized powders. **Figure 1.** Photo of the as-synthetized powders.

The differences in the physico-chemical properties of bare CeO<sup>2</sup> and grey CeO<sup>2</sup> were illustrated in Figure 2, where the XRD patterns (Figure 2a), the Raman (Figure 2b), and the FTIR (Figures 2c-2d) spectra are reported. The differences in the physico-chemical properties of bare CeO<sup>2</sup> and grey CeO<sup>2</sup> were illustrated in Figure 2, where the XRD patterns (Figure 2a), the Raman (Figure 2b), and the FTIR (Figure 2c,d) spectra are reported.

Both the samples exhibited the typical XRD pattern of ceria in the fluorite crystalline phase (Figure 2a), with the reflections at 2θ values of 28.6 (1 1 1), 33. 1 (2 0 0), 47.4 (2 2 0), and 56.4 (3 1 1) [34]. No substantial variation was detected in the grey CeO<sup>2</sup> compared to bare oxide, apart from a slight intensity decrease and a difference in the average crystallite size: 6.8 ± 0.8 nm for bare CeO<sup>2</sup> respect to 11.3 ± 1.1 nm for grey CeO2, calculated using the Scherrer equation on the main diffraction formation of CeO2-xdefects.

bands at 1071 and 839 cm-1

CeO<sup>2</sup> was blue-shifted by 5 cm-1

Specifically, for bare CeO2, the high intense band at 1385 cm-1

can be reasonably related to the more defective surface of grey CeO2.

observed in the grey CeO<sup>2</sup> can be reasonable connected to a decrease in crystallinity due to the

vibration of the cubic fluorite structure [36]. The position of this peak is influenced by the distortion of the Ce-O bonds [32]. Consequently, the treatment of grey CeO<sup>2</sup> led to a more defective structure with the modification of the cubic structure of CeO2, resulting in the Raman shift. However, in the as-synthesized bare CeO2, an imperfect crystalline stoichiometry was detected, being the small shoulder at about 600 cm-1 (more intense in the bare CeO2), ascribed to Frenkel-type oxygen vacancies [37]. Other differences can be seen in the FTIR spectra (Figure 2c). The bands at about and 1620 cm-1 are attributed to the stretching and the bending of the O-H groups of residual water molecules respectively, whereas the group of bands in the range 1600-500 cm-1 are related to the presence of carbonates due to the interaction of the atmospheric carbon dioxide with ceria [38]. From the zoomed spectra illustrated in Figure 2d, it is possible to note the formation of different carbonate species.

whereas the bands at 1190 and 1120 cm-1 are related to the bridged carbonate species. Finally, the

intense bands at 1012 and 872 cm-1 can be also be assigned for this sample to the presence of hydrocarbonates, whereas there is no evidence of the formation of monodentate carbonates. It is clear

band assigned to the monodentate carbonate was broader and shifted at 1395 cm-1

indicated the formation of hydrocarbonates [39–41]. In the grey CeO2, the

Interestingly, analyzing the Raman spectra of the samples (Figure 2b), the peak at 461 cm-1 of the

in the grey CeO2. The Raman peak at 461 cm-1

identifies the F2g skeletal

is due to monodentate carbonates,

, whereas the low

**Figure 2.** (**a**) XRD patterns, (**b**) Raman spectra, (**c**) FTIR spectra of the synthetized samples, and (**d**) FTIR zoom of the "carbonate" zone.

Both the samples exhibited the typical XRD pattern of ceria in the fluorite crystalline phase (Figure 2a), with the reflections at 2θ values of 28.6 (1 1 1), 33. 1 (2 0 0), 47.4 (2 2 0), and 56.4 (3 1 1) [34]. No substantial variation was detected in the grey CeO<sup>2</sup> compared to bare oxide, apart from a slight intensity decrease and a difference in the average crystallite size: 6.8 ± 0.8 nm for bare CeO<sup>2</sup> respect to 11.3 ± 1.1 nm for grey CeO2, calculated using the Scherrer equation on the main diffraction peak of ceria 2θ = 28.6 (1 1 1). This size enhancement of grey CeO<sup>2</sup> was related to the occurrence of the formation of defects inside the crystalline structure of CeO<sup>2</sup> caused by the solar irradiation in the H<sup>2</sup> stream. Furthermore, in accordance with the literature data [26,35], the intensity diminution observed in the grey CeO<sup>2</sup> can be reasonable connected to a decrease in crystallinity due to the formation of CeO2-x defects.

Interestingly, analyzing the Raman spectra of the samples (Figure 2b), the peak at 461 cm−<sup>1</sup> of the CeO<sup>2</sup> was blue-shifted by 5 cm−<sup>1</sup> in the grey CeO2. The Raman peak at 461 cm−<sup>1</sup> identifies the F2g skeletal vibration of the cubic fluorite structure [36]. The position of this peak is influenced by the distortion of the Ce-O bonds [32]. Consequently, the treatment of grey CeO<sup>2</sup> led to a more defective structure with the modification of the cubic structure of CeO2, resulting in the Raman shift. However, in the as-synthesized bare CeO2, an imperfect crystalline stoichiometry was detected, being the small shoulder at about 600 cm−<sup>1</sup> (more intense in the bare CeO2), ascribed to Frenkel-type oxygen vacancies [37]. Other differences can be seen in the FTIR spectra (Figure 2c). The bands at about and 1620 cm−<sup>1</sup> are attributed to the stretching and the bending of the O-H groups of residual water molecules respectively, whereas the group of bands in the range 1600–500 cm−<sup>1</sup> are related to the presence of carbonates due to the interaction of the atmospheric carbon dioxide with ceria [38]. From the zoomed spectra illustrated in Figure 2d, it is possible to note the formation of different carbonate species. Specifically, for bare CeO2, the high intense band at 1385 cm−<sup>1</sup> is due to monodentate carbonates, whereas the bands at 1190 and 1120 cm−<sup>1</sup> are related to the bridged carbonate species. Finally, the bands at 1071 and 839 cm−<sup>1</sup> indicated the formation of hydrocarbonates [39–41]. In the grey CeO2, the band assigned to the monodentate carbonate was broader and shifted at 1395 cm−<sup>1</sup> , whereas the low intense bands at 1012 and 872 cm−<sup>1</sup> can be also be assigned for this sample to the presence of hydrocarbonates, whereas there is no evidence of the formation of monodentate carbonates. It is clear that the surface interaction sites in the grey CeO<sup>2</sup> were changed compared to un-treated CeO2. This can be reasonably related to the more defective surface of grey CeO2. **Figure 2.** (**a**) XRD patterns, (**b**) Raman spectra, (**c**) FTIR spectra of the synthetized samples, and (**d**) FTIR zoom of the "carbonate" zone. The textural properties of the CeO<sup>2</sup> samples are displayed in the Figure 3. Both the materials displayed a N<sup>2</sup> adsorption–desorption isotherm of type III, with a H3 hysteresis loop (Figure 3a), indicating the presence of macro-meso slit-shaped pores [42]. The treatment with solar lamp in H<sup>2</sup> flow led to a decrease in the Brunauer–Emmett–Teller (BET) surface area. The grey CeO<sup>2</sup> exhibited a lower surface area (67 ± 1 m2/g) than CeO<sup>2</sup> (81 ± 1 m2/g). This decrease can be reasonably due to the agglomeration of CeO<sup>2</sup> particles caused by the irradiation treatment under solar lamp in H<sup>2</sup> flow, as

*Catalysts* **2020**, *10*, x FOR PEER REVIEW 5 of 16

The textural properties of the CeO<sup>2</sup> samples are displayed in the Figure 3. Both the materials displayed a N<sup>2</sup> adsorption–desorption isotherm of type III, with a H3 hysteresis loop (Figure 3a), indicating the presence of macro-meso slit-shaped pores [42]. The treatment with solar lamp in H<sup>2</sup> flow led to a decrease in the Brunauer–Emmett–Teller (BET) surface area. The grey CeO<sup>2</sup> exhibited a lower surface area (67 ± 1 m<sup>2</sup> /g) than CeO<sup>2</sup> (81 ± 1 m<sup>2</sup> /g). This decrease can be reasonably due to the agglomeration of CeO<sup>2</sup> particles caused by the irradiation treatment under solar lamp in H<sup>2</sup> flow, as further confirmed by the increase in mean crystalline size calculated by XRD. As a consequence, it was verified a shift towards large pores in the Barrett, Joyner and Halenda (BJH) pore size distribution curves (Figure 3b), with the mean pore size of the grey CeO<sup>2</sup> higher (58 ± 1 nm) with respect to bare CeO<sup>2</sup> (36 ± 1 nm). This size increase, verified by the grey CeO2, is strictly correlated with the peculiar treatment of this latter sample, i.e., the simultaneous utilization of the simulated solar radiation and the H<sup>2</sup> stream. As stated before, according to the work of Aslam et al. [32], the solar light alone did not caused any change in the mean size of bare CeO2; however, we used a more focused solar lamp than in ref. [32], which led to a slight heating of the sample (from room temperature to about 40 ◦C). On the contrary, the irradiation in a reductive atmosphere (H<sup>2</sup> stream) promoted the formation of numerous oxygen vacancies, a process characterized with an increase in the internal pressure inside the ceria crystalline structure with a consequent interatomic bond cleavage [43]. Reasonably, this process resulted in an agglomeration with a measurable size increase in the grey CeO<sup>2</sup> particle size. The same linear correlation between the increase in mean crystalline size, and the decrease in the BET surface area was already reported in the literature with other CeO2-based samples [44,45]. The formation of defects did not alter the morphology of the CeO<sup>2</sup> materials, that, if prepared by chemical precipitation, are usually characterized by a random stacking of particles [23,32]. further confirmed by the increase in mean crystalline size calculated by XRD. As a consequence, it was verified a shift towards large pores in the Barrett, Joyner and Halenda (BJH) pore size distribution curves (Figure 3b), with the mean pore size of the grey CeO<sup>2</sup> higher (58 ± 1 nm) with respect to bare CeO<sup>2</sup> (36 ± 1 nm). This size increase, verified by the grey CeO2, is strictly correlated with the peculiar treatment of this latter sample, i.e., the simultaneous utilization of the simulated solar radiation and the H<sup>2</sup> stream. As stated before, according to the work of Aslam et al. [32], the solar light alone did not caused any change in the mean size of bare CeO2; however, we used a more focused solar lamp than in ref. [32], which led to a slight heating of the sample (from room temperature to about 40 °C). On the contrary, the irradiation in a reductive atmosphere (H<sup>2</sup> stream) promoted the formation of numerous oxygen vacancies, a process characterized with an increase in the internal pressure inside the ceria crystalline structure with a consequent interatomic bond cleavage [43]. Reasonably, this process resulted in an agglomeration with a measurable size increase in the grey CeO2 particle size. The same linear correlation between the increase in mean crystalline size, and the decrease in the BET surface area was already reported in the literature with other CeO2 based samples [44,45]. The formation of defects did not alter the morphology of the CeO<sup>2</sup> materials, that, if prepared by chemical precipitation, are usually characterized by a random stacking of particles [23,32]. The UV-vis Diffuse Reflectance spectra of the CeO<sup>2</sup> powders are displayed in Figure 4a where the reflectance function (Kubelka–Munk function) is plotted versus the wavelength. A slight variation in the absorption features was detected for grey CeO<sup>2</sup> with a shift at lower wavelengths that results in a slightly higher optical band-gap (3.1 ± 0.3 eV) compared to bare CeO<sup>2</sup> (2.7 ± 0.3 eV) estimated by graphing the modified Kubelka–Munk function versus the eV (Figure 4b) [46]. The lower band-gap of CeO<sup>2</sup> (activation wavelength ≤ 460 nm) is suitable to exploit, together with the UV portion, a part of visible component of the solar light, whereas the grey CeO<sup>2</sup> with a higher band-gap (activation wavelength ≤ 400 nm) will be preferentially activated by the solar UV photons.

**Figure 3.** (**a**) N<sup>2</sup> adsorption–desorption isotherms of the CeO<sup>2</sup> samples; (**b**) pore size distribution curve of the analyzed samples evaluated by means of the Barrett, Joyner and Halenda (BJH) method. **Figure 3.** (**a**) N<sup>2</sup> adsorption–desorption isotherms of the CeO<sup>2</sup> samples; (**b**) pore size distribution curve of the analyzed samples evaluated by means of the Barrett, Joyner and Halenda (BJH) method.

The UV-vis Diffuse Reflectance spectra of the CeO<sup>2</sup> powders are displayed in Figure 4a where the reflectance function (Kubelka–Munk function) is plotted versus the wavelength. A slight variation in the absorption features was detected for grey CeO<sup>2</sup> with a shift at lower wavelengths that results in a slightly higher optical band-gap (3.1 ± 0.3 eV) compared to bare CeO<sup>2</sup> (2.7 ± 0.3 eV) estimated by graphing the modified Kubelka–Munk function versus the eV (Figure 4b) [46]. The lower band-gap of CeO<sup>2</sup> (activation wavelength ≤ 460 nm) is suitable to exploit, together with the UV portion, a part of visible component of the solar light, whereas the grey CeO<sup>2</sup> with a higher band-gap (activation wavelength ≤ 400 nm) will be preferentially activated by the solar UV photons. *Catalysts* **2020**, *10*, x FOR PEER REVIEW 6 of 16

**Figure 4.** (**a**) UV-vis Diffuse Reflectance spectra of CeO<sup>2</sup> powders; (**b**) estimation of the optical bandgap of the samples by means of the modified Kubelka–Munk function. **Figure 4.** (**a**) UV-vis Diffuse Reflectance spectra of CeO<sup>2</sup> powders; (**b**) estimation of the optical band-gap of the samples by means of the modified Kubelka–Munk function.

For the photocatalytic degradation of the IMI and especially for the Fenton and photo-Fenton reactions, it is fundamental that the presence of Ce3+ defects on the surface of CeO2. To establish the presence of these defect states, the XPS analysis was performed and the results are illustrated in Figure 5. In accordance with the literature data, the Ce 3d5/2 state involves the v, v', v'' and v''' component, whereas the u, u', u'' and u''' components are related to the Ce 3d3/2 state [47–50]. The v' and u' components indicate the presence of Ce3+ , whereas the peak at 916.4 eV (u''') for CeO<sup>2</sup> and at 916.8 eV for the grey CeO<sup>2</sup> are the typical fingerprint of Ce4+ [48–50]. It is clearly visible from Figure 5, as the component v' at 885.2 eV of Ce3+ is intense for grey CeO<sup>2</sup> whereas the same signal is absent in the bare CeO2. The u' signal is covered to the u and u'' components in both the samples. Furthermore, it is possible to note that, as the ratio between the v''' and u components is different and shifted of about 0.5 eV, as for the v component, compared to un-modified CeO2. This is another indication of the modification of the ceria surface sites with the higher presence of Ce3+ states in the grey CeO<sup>2</sup> [50]. The irradiation with solar lamp in H<sup>2</sup> stream thus induced the formation of CeO2-x defects on the surface of CeO2, as also confirmed by the Raman spectroscopy, with a consequent modification to the surface chemical composition of ceria, as also indirectly corroborated by the FTIR with the formation of different carbonate species in the two CeO<sup>2</sup> samples. For the photocatalytic degradation of the IMI and especially for the Fenton and photo-Fenton reactions, it is fundamental that the presence of Ce3<sup>+</sup> defects on the surface of CeO2. To establish the presence of these defect states, the XPS analysis was performed and the results are illustrated in Figure 5. In accordance with the literature data, the Ce 3d5/<sup>2</sup> state involves the v, v', v" and v"' component, whereas the u, u', u" and u"' components are related to the Ce 3d3/<sup>2</sup> state [47–50]. The v' and u' components indicate the presence of Ce3+, whereas the peak at 916.4 eV (u"') for CeO<sup>2</sup> and at 916.8 eV for the grey CeO<sup>2</sup> are the typical fingerprint of Ce4<sup>+</sup> [48–50]. It is clearly visible from Figure 5, as the component v' at 885.2 eV of Ce3<sup>+</sup> is intense for grey CeO<sup>2</sup> whereas the same signal is absent in the bare CeO2. The u' signal is covered to the u and u" components in both the samples. Furthermore, it is possible to note that, as the ratio between the v"' and u components is different and shifted of about 0.5 eV, as for the v component, compared to un-modified CeO2. This is another indication of the modification of the ceria surface sites with the higher presence of Ce3<sup>+</sup> states in the grey CeO<sup>2</sup> [50]. The irradiation with solar lamp in H<sup>2</sup> stream thus induced the formation of CeO2-x defects on the surface of CeO2, as also confirmed by the Raman spectroscopy, with a consequent modification to the surface chemical composition of ceria, as also indirectly corroborated by the FTIR with the formation of different carbonate species in the two CeO<sup>2</sup> samples. (**a**) (**b**) **Figure 4.** (**a**) UV-vis Diffuse Reflectance spectra of CeO<sup>2</sup> powders; (**b**) estimation of the optical bandgap of the samples by means of the modified Kubelka–Munk function. For the photocatalytic degradation of the IMI and especially for the Fenton and photo-Fenton reactions, it is fundamental that the presence of Ce3+ defects on the surface of CeO2. To establish the presence of these defect states, the XPS analysis was performed and the results are illustrated in Figure 5. In accordance with the literature data, the Ce 3d5/2 state involves the v, v', v'' and v''' component, whereas the u, u', u'' and u''' components are related to the Ce 3d3/2 state [47–50]. The v' and u' components indicate the presence of Ce3+ , whereas the peak at 916.4 eV (u''') for CeO<sup>2</sup> and at 916.8 eV for the grey CeO<sup>2</sup> are the typical fingerprint of Ce4+ [48–50]. It is clearly visible from Figure 5, as the component v' at 885.2 eV of Ce3+ is intense for grey CeO<sup>2</sup> whereas the same signal is absent in the bare CeO2. The u' signal is covered to the u and u'' components in both the samples. Furthermore, it is possible to note that, as the ratio between the v''' and u components is different and shifted of about 0.5 eV, as for the v component, compared to un-modified CeO2. This is another indication of the modification of the ceria surface sites with the higher presence of Ce3+ states in the grey CeO<sup>2</sup> [50]. The irradiation with solar lamp in H<sup>2</sup> stream thus induced the formation of CeO2-x defects on the surface of CeO2, as also confirmed by the Raman spectroscopy, with a consequent modification to the surface chemical composition of ceria, as also indirectly corroborated by the FTIR

with the formation of different carbonate species in the two CeO<sup>2</sup> samples.

the IMI pesticide. Three different AOPs were investigated: a) the photocatalytic oxidation (Figure 6a) **Figure 5.** XPS spectra of the CeO<sup>2</sup> samples. **Figure 5.** XPS spectra of the CeO<sup>2</sup> samples.

We have compared the (photo)catalytic activity of the synthetized sample in the degradation of

*2.2. (Photo)catalytic Activity*

*2.2. (Photo)catalytic Activity*

### *2.2. (Photo)catalytic Activity* way, crucial steps. Interestingly, in the photo-Fenton like reaction (Figure 6c), the grey CeO<sup>2</sup> displayed the best

We have compared the (photo)catalytic activity of the synthetized sample in the degradation of the IMI pesticide. Three different AOPs were investigated: a) the photocatalytic oxidation (Figure 6a) utilized as an irradiation source a solar lamp; b) the Fenton reaction (Figure 6b), adding 5 mL of H2O<sup>2</sup> (3%, 0.9M) in the reaction mixture, c) the photo-Fenton reaction (Figure 6c) utilizing both the solar lamp and the hydrogen peroxide. performance (~35% of degradation), comparing all the investigated AOPs with the two CeO<sup>2</sup> samples. The solar irradiation could boost the further formation of Ce3+ sites. An indirect confirmation is derived from the slight enhancement of the catalytic activity of bare CeO<sup>2</sup> (25%) compared to the solar photocatalytic test (20%) that can be attributed to the formation of in situ oxygen vacancies in the surface of bare CeO<sup>2</sup> which can react with the hydrogen peroxide.

*Catalysts* **2020**, *10*, x FOR PEER REVIEW 8 of 16

these tests, Figure S1). The regeneration and the further formation of Ce3+ defect centres are, in this

**Figure 6.** (**a**) Photocatalytic degradation of imidacloprid (IMI) under solar light irradiation, (**b**) Fenton-like reaction, (**c**) photo-Fenton like reaction on the CeO2-based samples, (**d**) photo-Fenton like reaction utilizing grey CeO<sup>2</sup> in different runs. **Figure 6.** (**a**) Photocatalytic degradation of imidacloprid (IMI) under solar light irradiation, (**b**) Fenton-like reaction, (**c**) photo-Fenton like reaction on the CeO<sup>2</sup> -based samples, (**d**) photo-Fenton like reaction utilizing grey CeO<sup>2</sup> in different runs.

As reported in the literature [25,32], the irradiation of ceria with photons which possess energy higher than the CeO2 band-gap can exploit the following reaction: CeO<sup>2</sup> + hν (E ≥ Eg) → Ce+3,+4O2-x + x/2 O<sup>2</sup> (9) The interaction of the highly energetic photons with the surface of CeO<sup>2</sup> leads to the loss of surface oxygen, thus allowing the formation of the Ce3+ states. The same reaction was exploited during the preparation of grey CeO2, where the formation of Ce3+ was further increased due to the reducing atmosphere. Therefore, with the grey CeO2, owing to a higher number of defective centres compared to bare ceria (as shown by XPS and Raman analyses), it is possible to reach the best performance in the degradation of IMI by the photo-Fenton-like reaction. Furthermore, the The solar photodegradation of pesticides required harder conditions in comparison to the degradation of other pollutants (for example, dyes) [51,52]. As a result, even utilizing TiO<sup>2</sup> (the most investigated photocatalyst), the degradation efficiency is not so high [7,53]. Furthermore, in accordance with our preceding work [7], and the literature data [54,55], as confirmed for all the AOPs investigated, the IMI degradation is characterized by the formation of various by-products as amine and chloro-pyridine species. The reaction mechanism involves the breaking of C–N and the N–N bonds followed by the formation of small molecules, such as chlorine dioxide, nitrogen oxides species, water and carbon dioxide [7,54,55]. The reported degradation percentage of IMI (i.e., the variation of the IMI concentration respect to the initial IMI concentration) was low even through photocatalysis [54–56], solar photo-Fenton [57], or UV-A photolysis [28]. In particular, with UV irradiation it is possible to obtain a complete photolysis of IMI after

contemporaneous presence of the hydrogen peroxide and the solar irradiation enhances the

a long time of irradiation (about 10 h) [28], whereas in our precedent work [7] with molecularly imprinted TiO<sup>2</sup> samples, it was possible to selectively photodegrade IMI even in a pesticide mixture, although the degradation efficiency did not exceed 40% with a partial mineralization of ~35% (evaluated by the Total Organic Carbon, TOC, analysis) after 3 h of UV irradiation. Kitsiou et al. [57] found that the reaction efficiency can be improved utilizing a combination of photo-Fenton under UV-A irradiation and TiO2, due to the synergism between the homogenous iron catalyst and the heterogeneous TiO<sup>2</sup> photocatalyst (~80% of degradation after 2 h of UV-A irradiation and ~60% of TOC mineralization), whereas only the solar homogenous photo-Fenton with iron reached ~50% for both degradation and removal of organic carbon after 3 h of UV-A irradiation. The most promising result was obtained by Sharma et al. [54] with a particular TiO<sup>2</sup> supported on mesoporous silica SBA-15, that allowed to achieve ~90% IMI degradation after 3 h of solar irradiation. In this contest, the obtained (photo)catalytic performances of CeO<sup>2</sup> for the degradation of IMI described in this work are in line with the results obtained with the TiO2-based materials.

Figure 6a reports the photocatalytic degradation of the synthetized powders. In the test without catalysts, (black line in Figure 6a) no substantial variations in the initial concentration of IMI was measured, as expected. On the other hand, after 3 h of solar light irradiation the bare CeO<sup>2</sup> was able to degrade around the 20% of the initial concentration of IMI, whereas the grey CeO<sup>2</sup> showed a slightly lower performance (~16%). This can be reasonably explained considering the lower surface area and/or the slightly higher band-gap of grey CeO<sup>2</sup> with respect to the un-modified CeO2.

The catalytic activity through the Fenton reaction (Figure 6b) is significantly lower compared to the photocatalytic tests (the test was carried out without irradiation). In these tests, no substantial degradation of IMI was measured in the run carried out without catalysts, but with H2O<sup>2</sup> (Figure 6b, olive line).

As explained in the Introduction (see reactions 6-8), the Fenton process requires the presence and the fast regeneration of Ce3<sup>+</sup> defect sites. For this reason, differently to the photocatalytic tests, the grey CeO<sup>2</sup> is more active than the bare CeO2. As detected by Raman and XPS measurements, the un-modified CeO<sup>2</sup> exhibited a much lower presence of defect centres with respect to grey CeO2. Conversely, despite the major presence of surface defects in the grey CeO2, the degradation percentage measured on grey CeO<sup>2</sup> after 3 h of reaction in the Fenton-like test (~10%) was lower compared to the degradation efficiency of the photocatalytic test (~16%) obtained with the same sample, which pointed to the slow regeneration of the Ce3<sup>+</sup> sites.

As reported, the Fenton-like reaction with CeO<sup>2</sup> involves the formation of peroxide species on the surface of ceria due to the complexation of H2O<sup>2</sup> with Ce3<sup>+</sup> sites [25,30,58]. These peroxide species are chemically stable and can saturate the surface of CeO2, hindering the adsorption and subsequent oxidation of the organic target contaminant [25,30,58]. Indeed, a higher concentration of H2O<sup>2</sup> (superior to 3%, i.e., 0,9 M) both in the Fenton and in the photo-Fenton like reactions led to a considerable decrease in the catalytic activity of the grey CeO<sup>2</sup> (the highest performing sample for these tests, Figure S1). The regeneration and the further formation of Ce3<sup>+</sup> defect centres are, in this way, crucial steps.

Interestingly, in the photo-Fenton like reaction (Figure 6c), the grey CeO<sup>2</sup> displayed the best performance (~35% of degradation), comparing all the investigated AOPs with the two CeO<sup>2</sup> samples. The solar irradiation could boost the further formation of Ce3<sup>+</sup> sites. An indirect confirmation is derived from the slight enhancement of the catalytic activity of bare CeO<sup>2</sup> (25%) compared to the solar photocatalytic test (20%) that can be attributed to the formation of in situ oxygen vacancies in the surface of bare CeO<sup>2</sup> which can react with the hydrogen peroxide.

As reported in the literature [25,32], the irradiation of ceria with photons which possess energy higher than the CeO<sup>2</sup> band-gap can exploit the following reaction:

$$\text{CaO}\_2 + \text{hv (E} \ge \text{E}\_\text{g}) \to \text{Ce}^{+3,+4}\text{O}\_{2\cdot x} + \text{x/2 }\text{O}\_2 \tag{9}$$

The interaction of the highly energetic photons with the surface of CeO<sup>2</sup> leads to the loss of surface oxygen, thus allowing the formation of the Ce3<sup>+</sup> states. The same reaction was exploited during the preparation of grey CeO2, where the formation of Ce3<sup>+</sup> was further increased due to the reducing atmosphere. Therefore, with the grey CeO2, owing to a higher number of defective centres compared to bare ceria (as shown by XPS and Raman analyses), it is possible to reach the best performance in the degradation of IMI by the photo-Fenton-like reaction. Furthermore, the contemporaneous presence of the hydrogen peroxide and the solar irradiation enhances the formation of hydroxyl radicals through the photolytic decomposition of H2O2, as confirmed by the experiment carried out without a catalyst (H2O<sup>2</sup> + IMI) that led to a slight variation in the initial concentration of IMI (Figure 6c, olive line). *Catalysts* **2020**, *10*, x FOR PEER REVIEW 9 of 16 formation of hydroxyl radicals through the photolytic decomposition of H2O2, as confirmed by the experiment carried out without a catalyst (H2O<sup>2</sup> + IMI) that led to a slight variation in the initial concentration of IMI (Figure 6c, olive line). It is important to highlight that, in the photocatalytic tests, the occurrence of the photon

It is important to highlight that, in the photocatalytic tests, the occurrence of the photon interaction (reaction 9, reported above) can be exploited, but it has a minor role in determining the final performance. The presence of Ce3<sup>+</sup> was usually connected in the literature [59–61] to an improvement in the photocatalytic performance, especially under visible light irradiation, due to the presence of as-formed oxygen vacancies that shift the absorption of CeO<sup>2</sup> towards the visible-light region, improving the separation of the photogenerated charge carriers. However, the mean crystalline size and consequently the active surface area of the photocatalyst contribute considerably to the overall photocatalytic activity [34,62], as in our case, where the influence of the surface area is more preponderant than the effect of defects. For this reason, in the solar photocatalytic test, the bare CeO<sup>2</sup> (BET surface area of 81 m<sup>2</sup> /g) was more active than grey CeO<sup>2</sup> (BET surface area of 67 m<sup>2</sup> /g). interaction (reaction 9, reported above) can be exploited, but it has a minor role in determining the final performance. The presence of Ce3+ was usually connected in the literature [59–61] to an improvement in the photocatalytic performance, especially under visible light irradiation, due to the presence of as-formed oxygen vacancies that shift the absorption of CeO<sup>2</sup> towards the visible-light region, improving the separation of the photogenerated charge carriers. However, the mean crystalline size and consequently the active surface area of the photocatalyst contribute considerably to the overall photocatalytic activity [34,62], as in our case, where the influence of the surface area is more preponderant than the effect of defects. For this reason, in the solar photocatalytic test, the bare CeO<sup>2</sup> (BET surface area of 81 m2/g) was more active than grey CeO<sup>2</sup> (BET surface area of 67 m2/g). Finally, to test the reusability of grey CeO2, different photo-Fenton like reaction runs were performed on the same sample. In Figure 6d, the variation in the kinetic constant (referring to a first

Finally, to test the reusability of grey CeO2, different photo-Fenton like reaction runs were performed on the same sample. In Figure 6d, the variation in the kinetic constant (referring to a first order kinetic [7] is reported with respect to the various runs. After five runs, the kinetic constant decreases from 12 ± 1·10−<sup>4</sup> min−<sup>1</sup> to 5.5 ± 0.6·10−<sup>4</sup> min−<sup>1</sup> , highlighting that the continuous redox Ce3+→Ce4<sup>+</sup> process on the surface of grey ceria led to a progressive deactivation of the catalyst, reasonably for the saturation of the surface sites with the products of IMI degradation. Nevertheless, when the same sample were pre-treated before the tests in the H<sup>2</sup> flow at room temperature for 1 h, it was possible to exploit an almost total reversibility of grey CeO<sup>2</sup> in the photo-Fenton-like reaction. In fact, the kinetic constant raised to 10 ± 1·10−<sup>4</sup> min−<sup>1</sup> in run 6 and went back to 5.0 ± 0.5·10−<sup>4</sup> min−<sup>1</sup> after the subsequent other four runs, pointing to the crucial role of H<sup>2</sup> in the restoring the Ce3<sup>+</sup> sites on grey CeO2. order kinetic [7] is reported with respect to the various runs. After five runs, the kinetic constant decreases from 12 ± 1·10−4 min−1 to 5.5 ± 0.6·10−4 min−1, highlighting that the continuous redox Ce3+→Ce4+ process on the surface of grey ceria led to a progressive deactivation of the catalyst, reasonably for the saturation of the surface sites with the products of IMI degradation. Nevertheless, when the same sample were pre-treated before the tests in the H<sup>2</sup> flow at room temperature for 1 h, it was possible to exploit an almost total reversibility of grey CeO<sup>2</sup> in the photo-Fenton-like reaction. In fact, the kinetic constant raised to 10 ± 1·10−4 min−1 in run 6 and went back to 5.0 ± 0.5·10−4 min−1 after the subsequent other four runs, pointing to the crucial role of H<sup>2</sup> in the restoring the Ce3+ sites on grey CeO2. The comparison with the commercial TiO<sup>2</sup> P25 Degussa (Figure 7) showed, as in the photocatalytic test, that bare CeO<sup>2</sup> had a comparable catalytic behaviour with respect to TiO<sup>2</sup> (Figure 7a), whereas this latter sample exhibited no substantial activity in the Fenton-like test (Figure 7b),

The comparison with the commercial TiO<sup>2</sup> P25 Degussa (Figure 7) showed, as in the photocatalytic test, that bare CeO<sup>2</sup> had a comparable catalytic behaviour with respect to TiO<sup>2</sup> (Figure 7a), whereas this latter sample exhibited no substantial activity in the Fenton-like test (Figure 7b), and a lower activity compared to CeO<sup>2</sup> and grey CeO<sup>2</sup> in the photo-Fenton test (Figure 7c). This pointed out that, both for Fenton and photo-Fenton like reactions, the CeO2-based materials are better-performing than the commercial TiO2. As reported [63,64], the bare TiO<sup>2</sup> without structural (i.e., incorporation of surface defects) or chemical (as the formation of composites with iron oxides) modifications is not able to promote Fenton-like reactions. and a lower activity compared to CeO<sup>2</sup> and grey CeO2 in the photo-Fenton test (Figure 7c). This pointed out that, both for Fenton and photo-Fenton like reactions, the CeO2-based materials are better-performing than the commercial TiO2. As reported [63,64], the bare TiO<sup>2</sup> without structural (i.e., incorporation of surface defects) or chemical (as the formation of composites with iron oxides) modifications is not able to promote Fenton-like reactions. These data demonstrate the possibility of modifying and tuning the physico-chemical properties of CeO<sup>2</sup> with simple treatments, such as the solar light irradiation in a H<sup>2</sup> stream, so to maximize the catalytic performance. It is important to highlight, finally, that the CeO2 sample simply treated with H<sup>2</sup> or irradiated with a solar lamp without an H<sup>2</sup> stream did not show substantial changes compared to bare CeO<sup>2</sup> in the degradation performance of IMI in all the AOPs investigated.

**Figure 7.** *Cont*.

*Catalysts* **2020**, *10*, x FOR PEER REVIEW 10 of 16

**Figure 7.** (**a**) Photocatalytic degradation of IMI under solar light irradiation, (**b**) Fenton-like reaction, (**c**) photo-Fenton like reaction on the analyzed samples. **Figure 7.** (**a**) Photocatalytic degradation of IMI under solar light irradiation, (**b**) Fenton-like reaction, (**c**) photo-Fenton like reaction on the analyzed samples.

*2.3. Tocixity Tests Artemia salina* dehydrated cysts were employed for the acute toxicity test. *Artemia salina* nauplii can readily ingest fine particles smaller than 50 μm [65], and it is a nonselective filter-feeder organism. For these reasons, it is currently considered as a good model organism to assess in vivo nanoparticles toxicity, as previously demonstrated [66]. Low mortality percentages were evidenced after 24 and 48 hours of exposure (Table 1), at These data demonstrate the possibility of modifying and tuning the physico-chemical properties of CeO<sup>2</sup> with simple treatments, such as the solar light irradiation in a H<sup>2</sup> stream, so to maximize the catalytic performance. It is important to highlight, finally, that the CeO<sup>2</sup> sample simply treated with H<sup>2</sup> or irradiated with a solar lamp without an H<sup>2</sup> stream did not show substantial changes compared to bare CeO<sup>2</sup> in the degradation performance of IMI in all the AOPs investigated.

different concentrations of both powders (bare CeO<sup>2</sup> and grey CeO2). Statistical analysis, carried out

### by one-way ANOVA test, gave no significant values for all the immobilization percentages nor *2.3. Tocixity Tests*

treated groups after 24 and 48 hours of exposure nor between treated and control (Ctrl, i.e., without metal oxide particles) groups (p > 0.05). The percentages of immobilized nauplii are reported in the *Artemia salina* dehydrated cysts were employed for the acute toxicity test.

3.3% (24 h)

Grey CeO<sup>2</sup>

Table 1. These data pointed to the low critical toxicity of the examined powders. Furthermore, it is possible to note that the modification of CeO<sup>2</sup> led to have a lower mortality with respect to the bare CeO2. *Artemia salina* nauplii can readily ingest fine particles smaller than 50 µm [65], and it is a non-selective filter-feeder organism. For these reasons, it is currently considered as a good model organism to assess in vivo nanoparticles toxicity, as previously demonstrated [66].

Figure 8 shows *Artemia salina* nauplii treated with bare CeO<sup>2</sup> for 24 hours, 48 hours and untreated nauplii (i.e., the controls). **Table 1.** Percentages of immobilized nauplii after exposition to CeO2 (bare CeO<sup>2</sup> and grey CeO2) at three different concentrations for 24 and 48 hours. **Sample Ctrl 10-1 10-2 10-3** Bare CeO<sup>2</sup> 1.6% (24 h) 6.6% (48 h) 11.6% (24 h) 28.3% (48 h) 6.6% (24 h) 23.3% (48 h) 3.3% (24 h) 21.6% (48 h) Low mortality percentages were evidenced after 24 and 48 h of exposure (Table 1), at different concentrations of both powders (bare CeO<sup>2</sup> and grey CeO2). Statistical analysis, carried out by one-way ANOVA test, gave no significant values for all the immobilization percentages nor treated groups after 24 and 48 h of exposure nor between treated and control (Ctrl, i.e., without metal oxide particles) groups (*p* > 0.05). The percentages of immobilized nauplii are reported in the Table 1. These data pointed to the low critical toxicity of the examined powders. Furthermore, it is possible to note that the modification of CeO<sup>2</sup> led to have a lower mortality with respect to the bare CeO2.

**Sample Ctrl 10**−**<sup>1</sup> 10**−**<sup>2</sup> 10**−**<sup>3</sup>** Bare CeO<sup>2</sup> 1.6% (24 h) 6.6% (48 h) 11.6% (24 h) 28.3% (48 h) 6.6% (24 h) 23.3% (48 h) 3.3% (24 h) 21.6% (48 h) Grey CeO<sup>2</sup> 3.3% (24 h) 8.3% (48 h) 8.3 % (24 h) 23.3% (48 h) 6.6% (24 h) 18.3% (48 h) 4.0% (24 h) 15.0% (48 h)

8.3% (48 h) 23.3% (48 h) 18.3% (48 h) 15.0% (48 h) **Table 1.** Percentages of immobilized nauplii after exposition to CeO<sup>2</sup> (bare CeO<sup>2</sup> and grey CeO<sup>2</sup> ) at three different concentrations for 24 and 48 h.

8.3 % (24 h)

6.6% (24 h)

4.0% (24 h)

(**a**) (**b**) Figure 8 shows *Artemia salina* nauplii treated with bare CeO<sup>2</sup> for 24 h, 48 h and untreated nauplii (i.e., the controls).

nauplii (i.e., the controls).

CeO2.

*2.3. Tocixity Tests*

(**c**) **Figure 7.** (**a**) Photocatalytic degradation of IMI under solar light irradiation, (**b**) Fenton-like reaction,

*Artemia salina* nauplii can readily ingest fine particles smaller than 50 μm [65], and it is a nonselective filter-feeder organism. For these reasons, it is currently considered as a good model

Low mortality percentages were evidenced after 24 and 48 hours of exposure (Table 1), at different concentrations of both powders (bare CeO<sup>2</sup> and grey CeO2). Statistical analysis, carried out by one-way ANOVA test, gave no significant values for all the immobilization percentages nor treated groups after 24 and 48 hours of exposure nor between treated and control (Ctrl, i.e., without metal oxide particles) groups (p > 0.05). The percentages of immobilized nauplii are reported in the Table 1. These data pointed to the low critical toxicity of the examined powders. Furthermore, it is possible to note that the modification of CeO<sup>2</sup> led to have a lower mortality with respect to the bare

Figure 8 shows *Artemia salina* nauplii treated with bare CeO<sup>2</sup> for 24 hours, 48 hours and untreated

**Table 1.** Percentages of immobilized nauplii after exposition to CeO2 (bare CeO<sup>2</sup> and grey CeO2) at

**Sample Ctrl 10-1 10-2 10-3**

11.6% (24 h) 28.3% (48 h)

8.3 % (24 h) 23.3% (48 h)

6.6% (24 h) 23.3% (48 h)

6.6% (24 h) 18.3% (48 h)

3.3% (24 h) 21.6% (48 h)

4.0% (24 h) 15.0% (48 h)

11 of 16

*Artemia salina* dehydrated cysts were employed for the acute toxicity test.

organism to assess in vivo nanoparticles toxicity, as previously demonstrated [66].

(**c**) photo-Fenton like reaction on the analyzed samples.

three different concentrations for 24 and 48 hours.

1.6% (24 h) 6.6% (48 h)

3.3% (24 h) 8.3% (48 h)

Bare CeO<sup>2</sup>

**Figure 8.** *Artemia salina* nauplii: nauplii exposed to bare CeO<sup>2</sup> for 24 hours (**a**), nauplii exposed to bare CeO<sup>2</sup> for 48 hours (**b**), control at 24 hours (**c**), control at 48 hours (**d**). **Figure 8.** *Artemia salina* nauplii: nauplii exposed to bare CeO<sup>2</sup> for 24 h (**a**), nauplii exposed to bare CeO<sup>2</sup> for 48 h (**b**), control at 24 h (**c**), control at 48 h (**d**).

### **4. Materials and Methods 3. Materials and Methods**

### *4.1. Samples Preparation 3.1. Samples Preparation*

The bare CeO<sup>2</sup> was prepared via chemical precipitation from Ce(NO3)3·6H2O (Fluka, Buchs, Switzerland) at pH > 8 utilizing a solution of KOH (1M, Fluka, Buchs, Switzerland). The obtained slurry was maintained under stirring at 80 °C for 3 h. After digestion for 24 h, it was filtered, washed with deionized water several times, and dried at 100 °C for 12 hours. Finally, the powders were calcined in air at 450 °C for 3 h. The modified CeO<sup>2</sup> (grey CeO2) was obtained with the same synthetic procedure reported above, but exposing the powders after calcination to the light of a solar lamp for 3 h (OSRAM Vitalux 300 W, 300-2000 nm; OSRAM Opto Semiconductors GmbH, Leibniz, Regensburg Germany) in a hydrogen stream (20 cc/min) at room temperature. The bare CeO<sup>2</sup> was prepared via chemical precipitation from Ce(NO3)3·6H2O (Fluka, Buchs, Switzerland) at pH > 8 utilizing a solution of KOH (1M, Fluka, Buchs, Switzerland). The obtained slurry was maintained under stirring at 80 ◦C for 3 h. After digestion for 24 h, it was filtered, washed with deionized water several times, and dried at 100 ◦C for 12 h. Finally, the powders were calcined in air at 450 ◦C for 3 h. The modified CeO<sup>2</sup> (grey CeO2) was obtained with the same synthetic procedure reported above, but exposing the powders after calcination to the light of a solar lamp for 3 h (OSRAM Vitalux 300 W, 300–2000 nm; OSRAM Opto Semiconductors GmbH, Leibniz, Regensburg Germany) in a hydrogen stream (20 cc/min) at room temperature.

### *4.2. Samples Characterization 3.2. Samples Characterization*

*4.3. (Photo)catalytic Experiments*

X-ray powder diffraction (XRD) measures were performed with a PANalytical X'pertPro X-ray diffractometer (Malvern PANalytical, Enigma Business Park, Grovewood Road, Malvern United Kingdom), employing a Cu Kα radiation. The identification of the crystalline phases was made comparing the diffractions with those of standard materials reported in the JCPDS Data File. Raman spectra were carried out with a WITec alpha 300 confocal Raman system (WITec Wissenschaftliche Instrumente und Technologie GmbH Ulm, Germany) with an excitation source at 532 nm under the same experimental condition reported in the ref. [67]. Fourier Transform Infrared Spectroscopy (FTIR) spectra were obtained in the range 4000–400 cm−1 using a Perkin Elmer FT-IR System 2000 (Perkin-Elmer, Waltham, MA, USA). The background spectrum was carried out with KBr. The textural properties of the samples were measured by N<sup>2</sup> adsorption–desorption at −196 °C with a Micromeritics Tristar II Plus 3020 (Micromeritics Instrument Corp. Norcross, USA), out-gassing the analysed materials at 100 °C overnight. UV-Vis-Diffuse Reflectance (UV-Vis DRS) spectra were obtained in the range of 200-800 nm using a Cary 60 spectrometer (Agilent Stevens Creek Blvd. Santa Clara, United States). X-ray photoelectron spectroscopy (XPS) measurements were recorded using a K-Alpha X-ray photoelectron spectrometer (Thermo Fisher Scientific Waltham, MA USA), utilizing the C 1 peak at 284.9 eV (ascribed to adventitious carbon) as a reference. X-ray powder diffraction (XRD) measures were performed with a PANalytical X'pertPro X-ray diffractometer (Malvern PANalytical, Enigma Business Park, Grovewood Road, Malvern United Kingdom), employing a Cu Kα radiation. The identification of the crystalline phases was made comparing the diffractions with those of standard materials reported in the JCPDS Data File. Raman spectra were carried out with a WITec alpha 300 confocal Raman system (WITec Wissenschaftliche Instrumente und Technologie GmbH Ulm, Germany) with an excitation source at 532 nm under the same experimental condition reported in the ref. [67]. Fourier Transform Infrared Spectroscopy (FTIR) spectra were obtained in the range 4000–400 cm−<sup>1</sup> using a Perkin Elmer FT-IR System 2000 (Perkin-Elmer, Waltham, MA, USA). The background spectrum was carried out with KBr. The textural properties of the samples were measured by N<sup>2</sup> adsorption–desorption at <sup>−</sup><sup>196</sup> ◦C with a Micromeritics Tristar II Plus 3020 (Micromeritics Instrument Corp. Norcross, USA), out-gassing the analysed materials at 100 ◦C overnight. UV-Vis-Diffuse Reflectance (UV-Vis DRS) spectra were obtained in the range of 200–800 nm using a Cary 60 spectrometer (Agilent Stevens Creek Blvd. Santa Clara, United States). X-ray photoelectron spectroscopy (XPS) measurements were recorded using a K-Alpha X-ray photoelectron spectrometer (Thermo Fisher Scientific Waltham, MA USA), utilizing the C 1 peak at 284.9 eV (ascribed to adventitious carbon) as a reference.

stirred for 120 min in the dark so as to achieve the adsorption/desorption equilibrium. During the tests, aliquots of the suspension were withdrawn at a given time interval to measure the IMI concentration by means of Cary 60 UV–vis spectrophotometer (Agilent Stevens Creek Blvd. Santa Clara, United States). The IMI degradation was evaluated by following the absorbance peaks at 270 nm in the Lambert–Beer regime, reporting the *C/C<sup>0</sup>* ratio as a function of time *t*, where *C* is the concentration of the contaminant at the time *t*, while *C<sup>0</sup>* is the starting concentration of the pollutant.

The photocatalytic tests were performed utilizing a solar lamp (OSRAM Vitalux 300 W, 300-2000
