Enhancing the Photocatalytic Activity and Luminescent Properties of Rare-Earth-Doped CeO₂ Nanoparticles

Abstract

Our study is focused on the structural and morphological characteristics, optical behaviour and photocatalytic properties of undoped and 5 at% Eu³⁺-, Gd³⁺- and Y³⁺-doped CeO₂ nanoparticles prepared by a green hybrid sol-gel combustion method. Several techniques such as X-ray diffraction powder (XRD), Transmission Electron Microscopy (TEM), UV-Vis spectroscopy, Photoluminescence spectroscopy (PL) and Electron Paramagnetic Resonance (EPR) have been used to investigate the obtained samples. Moreover, the correlation between the characteristics and properties has been studied. The nanoparticles observed by TEM exhibit a pseudo-spherical shape, except for Y³⁺-doped CeO₂, which presents an acicular shape. The average size of undoped and rare-earth-doped CeO₂ nanoparticles is below 10 nm, in good agreement with the calculations performed based on XRD analyses. From UV-Vis analyses it has been deduced that with doping the band gap energy decreases, which shows that additional levels are introduced by doping into the CeO₂ band gap. The EPR spectra evidence similar behaviour for all doped samples. The photocatalytic activity was evaluated by the degradation of rhodamine B (RhB) under UV light irradiation. The photodegradation mechanism has been studied in depth based on the formation of electron-hole pairs, and to evidence the reactive oxygen species, ESR coupled with spin-trapping experiments was performed. In the case of Y-doped CeO₂ nanoparticles, the generation of both •OOH and •${\mathrm{O}}_2^{-}$ radicals involved in RhB photodegradation was highlighted.

1. Introduction

Industrial progress has improved living standards, driving the development of human society. Therefore, some industrial pollutants are generated in wastewater effluents, such as heavy metals, radioactive nuclides, and organics [1]. Unfortunately, great amounts of wastewater are released into the environment, causing significant environmental pollution. Organic pollutants such as phenols, organic dyes and antibiotics are very toxic. These pollutants can be accumulated into aquatic organisms and from there along the food chain to the human body [2]. Thus, the discarding of organic pollutants represents one of the most important environmental problems. In recent years, many technologies have been employed in environmental remediation, including adsorption [3], biodegradation [4] and chemical oxidation [5]. The disadvantages of the above methods consist of the use of solvents and the generation of secondary metabolites which are very difficult to remove. One of the cost-effective methods that had been developed for the removal of organic contaminants from wastewater is Advanced Oxidation Processes (AOPs). AOPs using semiconductor photocatalysts have been proven to be a promising technique for water disinfection due to its high efficiency, acceptable price, high stability and the capability to degrade organic pollutants and mineralization to CO₂ and H₂O as the final products [6].

The efficiency of the photocatalytic process depends on the operational parameters that dominate the adsorption and photodegradation of dye molecules. Among these parameters, the most important are the initial pH of the dye solution, the concentration of the photocatalyst and dye, the reaction temperature and the dopants content [6,7]. Tirumala et al. show that controlling the incident light intensity and wavelength is a mode for controlling and tuning the activity and selectivity of the catalyst [8]. Also, in a doped semiconductor photocatalyst, the positively or negatively charged microenvironment around trapped holes or electrons influences the ability of photocatalyst to generate reactive oxygen species (ROS) involved in reactions of photo-oxidation and photo-reduction [9,10].

A great number of semiconductor metal oxides, such as TiO₂ [11], ZnO [12], SnO₂ [13], etc., have been used as photocatalytic materials. As one of the most popular metal oxide semiconductors, the rare earth oxide ceria (CeO₂) is extensively used in many fields due to its ability to extract, store and release oxygen. Its applicative potential based on this special feature includes domains like catalysis, photocatalysis, solid oxide fuel cells, ceramics, sunscreens, solar cells, gas sensors and biomedical areas such as biosensing, bioimaging or targeted drug therapy, etc. [14,15,16,17,18,19]. Moreover, ceria has proved to be an ideal candidate as an artificial oxygen modulator, and this is vital for many biological processes [20]. Also, nanoceria is well-known for its enhanced catalytic reactivity due to easy conversion between cerium’s tetravalent and trivalent valence surface states [21]. In ceria, the tetravalent cerium ion, Ce⁴⁺, transforms into a trivalent state, Ce³⁺, and this transformation produces a good yield of oxygen vacancies without changes in the chemical environment [22]. Additionally, it is nontoxic, abundant and absorbs light in the near UV and also marginally in the visible region, properties which can be used in heterogeneous photocatalysis [23]. Thus, the application of CeO₂ is obstructed. Many strategies have been used to enhance the catalytic/photocatalytic activity by tuning the structural and morphological characteristics [24]. One strategy to extend the utilization of visible light over CeO₂ consists of suitable doping with various metal ions: isovalent cations (Ti⁴⁺, Zr⁴⁺, Hf⁴⁺, Sn⁴⁺, etc.) or aliovalent cations (Mn²⁺, Zn²⁺, Ca²⁺, La³⁺, Eu³⁺, Sm³⁺, Y³⁺, Eu³⁺, Mn³⁺, Sc³⁺, Gd³⁺, Sm³⁺, etc.) [25]. It has been demonstrated theoretically and experimentally that the substitution of cerium ions with isovalent dopant cations induces a decrease in oxygen vacancy formation energy due to the structural distortion, whereas in the case of aliovalent dopants the generation of extra oxygen vacancies into the ceria lattice has been evidenced due to the synergism between the structural distortion and electron modifications [25,26,27]. On the other hand, the trivalent cations of rare earth metals drastically modify the reactivity structure and optical emission of ceria [21,28,29,30]. The luminescent properties of undoped or rare earth (Eu³⁺, Sm³⁺, or Tb³⁺)-doped ceria have been reported and the emission mechanism proposed by considering the transfer of the charge carrier between the Ce 4f conduction band and the O 2p valence band. Also, the broad photoluminescence emission in the visible domain is attributed to defects, including oxygen vacancies in the ceria lattice with electronic energy levels below the 4f band [31,32,33].

To date, a lot of chemical strategies have been developed to synthesize undoped or doped ceria nanoparticles, such as solid-state reactions [34], co-precipitation [35], hydrothermal techniques [36], sol-gel [37], microwave-assisted synthesis [38], ultrasonic spray pyrolysis [39], sonochemical synthesis [40], combustion method [41] and so on, proving that the physical and chemical properties are strongly dependent on the synthetic route used. In this regard, Amoresi et al. studied the correlation between a wide range of CeO₂ morphologies, such as nanorods, nanocubes, nanobeans and nanohexagons and the photocatalytic activity [42]. A similar study has been published by Ortega et al. with emphasis on the Eu-doped ceria nanostructures [43]. It has been demonstrated that enhanced photocatalytic performance is obtained by increasing the charge carrier concentration and light absorption capacity by defects induced by the dopants into CeO₂ nanostructures. Nevertheless, studies are still ongoing for a deeper understanding of the correlations between the experimental parameters, structure, morphology, luminescence properties and photocatalytic activity.

Considering the previous research works reported in the literature, this paper focuses on the effects induced by the incorporation into the ceria lattice of Eu³⁺, Gd³⁺ and Y³⁺ aliovalent cations by studying the correlations between the structural and morphological characteristics and luminescence properties and photocatalytic activity. These materials have been synthesized by a green hybrid sol-gel combustion method and characterized by using various techniques, namely X-ray Diffraction, Transmission Electron Microscopy, UV-Vis spectroscopy, Photoluminescence (PL) and Electron Paramagnetic Resonance (EPR). Finally, we attempt to provide an overview of how the selected rare earth ions influence the ceria’s structure, morphology and photocatalytic activity.

2. Materials and Methods

2.1. Materials

Commercial ammonium cerium (IV) nitrate, (NH₄)₂Ce(NO₃)₆, (Merck, 98.5%), gadolinium nitrate hexahydrate, Gd(NO₃)₃·6H₂O (99.9%, Merck), europium acetate hydrate Eu(CH₃-COO)₃·xH₂O (99.9%, Merck), yttrium acetate hydrate Y(CH₃-COO)₃·xH₂O (99.9%, Merck), sucrose, C₁₂H₂₂O₁₁ (99.5%, Sigma-Aldrich), pectin from citrus peel, (Sigma-Aldrich), rhodamine B (RhB), C₂₈H₃₁ClN₂O₃ (analytical standard, Merck), acetonitrile, C₂H₃N (ACN, HPLC-grade, Merck) and ortho-phosphoric acid (H₃PO₄, 85%, Merck) were used without any further purification.

2.2. Nanoparticle Synthesis

The undoped and Eu³⁺-, Gd³⁺- and Y³⁺-doped CeO₂ nanoparticles were synthetized via a green hybrid sol-gel combustion method developed by Goga et al. [44] using sucrose as a poly-condensation agent and pectin as a chelation agent. Briefly, the corresponding amounts of metal salts—for the doped systems 5 at% of Eu³⁺, Gd³⁺ and Y³⁺ ions relative to CeO₂—were dissolved in MiliQ water (0.05 M) at 60 °C for an hour under vigorous magnetic stirring. Similarly, a sucrose solution (0.1 M) was prepared using milli-Q water, at 60 °C, under magnetic stirring. The two solutions were mixed in the presence of the pectin (the sucrose to pectin weight ratio was 1:5). The pH value was monitored during the synthesis and no correction was necessary to maintain an acidic medium in the solution. The pH value of the final solutions was determined to be in the range of 1.5–2. The as-obtained mixture was dried in air, on the sand bath. At 215 °C, a spontaneous combustion of the synthesis mixture was observed and a dispersed yellow powder was obtained. The evolved gases were captured and absorbed in a soda lime aqueous solution. The following chemical reaction can be considered in the formation of ceria nanoparticles:

(1 − x)Ce³⁺ + xRE³⁺ + 2NH₄⁺ + 6NO₃⁻ + 2C₁₂H₂₂O₁₁ + 26O₂→Ce_(1−x)RE_xO₂ + 24CO₂↑ + 26H₂O + 8NO₂↑

The use of non-toxic reagents such as sucrose and pectin give the process a green synthesis approach, but the reaction is not an environmentally friendly reaction due to the production of nitrogen dioxide.

2.3. Characterization Techniques

The structural characterisation of the obtained nanoparticles was undertaken through X-ray diffraction (XRD). The equipment used was an INEL Equinox 3000 diffractometer, which works in reflection mode, and it is equipped with a curved multidetector. The wavelength used was the one characteristic of Co, Kα radiation (λ = 1.7903 Å). The mean crystallite size was calculated with the Scherrer method [45]. The lattice parameter was calculated using Celref version 3 refinement software [46]. TEM images were obtained by scanning transmission electron microscopy (STEM) using a Hitachi HD2700 Electron Microscope (Hitachi, Tokyo, Japan). Absorption spectra were recorded using a JASCO V570 UV–VIS-NIR Spectrophotometer equipped with an absolute reflectivity measurement JASCO ARN-475 accessory. Luminescent characteristics of samples were evaluated based on emission spectra registered with a Wavell JASCO FP-6500 spectrofluorimeter, equipped with a PMT R928 photomultiplier (glass filter UG1—Reichmann Feinoptik, GER). Electron paramagnetic resonance (EPR) spectroscopy measurements of powder samples were carried out using a Bruker E-500 ELEXSYS spectrometer (Karlsruhe, Germany) at room temperature under identical conditions: The measurements were conducted at Xband (9.52 GHz) microwave power 2 mW, with a modulation frequency of 100 kHz. To determine the reactive oxygen species (ROS) generation, the EPR spectrometer coupled with the spin trapping probe technique was used. As the spin trapping reagent, DMPO was used. The nanoparticles (10 mg) were dispersed in DMSO (1 mL) and homogenized in an ultrasound bath (30 min) before use. DMPO of 0.2 mol/L concentration was added to the solution. The obtained solution was subjected to visible irradiation for 25 min and immediately before measurement transferred into the quartz flat cell, optimized for liquid measurements. The photodegradation of Rhodamine B (RhB) was carried out in a Laboratory-Visible-Reactor system using a 400 W halogen lamp (Osram) [47]. The catalyst (5 mg) was suspended in an aqueous solution of RhB (1.0 × 10⁻⁵ mol·L⁻¹, 10 mL) and then the mixture was stirred in the dark to achieve the adsorption equilibrium on the catalyst surface. Each degradation experiment was continuously conducted for 6 h. The mixture (3.5 mL) was withdrawn for analysis every 60 min. After separating the catalyst from the suspensions by centrifugation, the solution was analysed by the UV–Vis spectrophotometer by recording the maximum absorbance of RhB at 554 nm.

RhB removal rate is calculated using the following equation:

$$\mathrm{R}\mathrm{e}\mathrm{m}\mathrm{o}\mathrm{v}\mathrm{a}\mathrm{l} \mathrm{r}\mathrm{a}\mathrm{t}\mathrm{e}=\frac{A_{0 }-A_t}{A_0}\times 100$$

where A₀ and A_t represent the RhB absorbance initially and at time t, respectively.

The HPLC analyses were performed on a Jasco HPLC chromatograph (Jasco International Co., Ltd., Tokyo, Japan) that was equipped with an intelligent pump, PU-980, a ternary gradient unit, LG-980-02, an intelligent column thermostat, CO-2060 Plus, an intelligent detector, UV-975, an intelligent fluorescence detector, FP-2020 Plus, and an injection valve that was equipped with a 20 µL sample loop (Rheodyne, Thermo Fischer Scientific, Waltham, MA, USA). The system was controlled, and the experimental data processed with the ChromPass software (version v1.7, Jasco International Co., Ltd., Tokyo, Japan). Separation was performed on a Gemini chromatographic column (5μ C18, 110 Å, 250 × 4.6 mm, Phenomenex, Torrance, CA, USA) at 21 °C. Isocratic elution was performed with 45% of acetonitrile with 55% of 0.1% H₃PO₄ buffer. The flow rate was 1 mL/min and the injection volume was set to 20 μL. UV detection was performed by fluorescence at the excitation and emission wavelengths set at 535 nm and 580 nm. All the experiments were performed in triplicate. The determination of rhodamine B was performed according to the adapted method described by Chiang et al. [48].

3. Results and Discussion

The structural characteristics of undoped or RE-doped CeO₂ samples were determined by X-ray diffraction measurements and the corresponding diffraction patterns are shown in Figure 1. All diffraction lines are well indexed to the face-centred cubic (fcc) fluorite structure of CeO₂ (JCDD PDF 034-0394). No peaks belonging to any other phase have been observed in the XRD patterns, indicating the high structural purity of the samples. The effects of the dopant on the lattice parameters and crystallite sizes—calculated using the Scherrer method—are presented in

Table 1. Average crystallite sizes (D) and lattice parameters (a) of undoped and doped ceria.

Sample	Crystallite Size D (nm) a	Lattice Parameter a (Å) a	Energy Band Gap Eg (eV)
Undoped CeO2	8.16	5.4105	2.98
CeO2:Eu	8.40	5.4130	2.93
CeO2:Gd	8.40	5.4129	2.88
CeO2:Y	7.79	5.4113	2.93

^a Calculates from XRD.

.

It has been demonstrated that the substitution of Ce⁴⁺ ions with a larger ion radius dopant cation alters the local chemistry and induces an enlargement of the crystal lattice due to steric effects [49,50,51]. The substitution into the ceria lattice of octahedral Ce⁴⁺ ions (Ce⁴⁺ = 0.97 Å) with a larger sized dopant (Gd³⁺ = 1.053 Å; Eu³⁺ = 1.066 Å and Y³⁺ = 1.019 A [52]) induced a slight increase in the lattice parameter values. Regarding the crystallite sizes it can be noted that the small amount of dopant ions stimulates the nucleation and the growth processes of the nanocrystals. This tendency can be noticed for the crystallite sizes in the cases of Gd³⁺ and Eu³⁺. The introduction of Y³⁺ into the ceria lattice led to a diminishing of the crystallite size.

The morphological characterization of the samples has been achieved via TEM investigations—Figure 2. TEM images show the formation of pseudo-spherical nanoparticles with a high tendency to agglomerate. Analysis of these images disclosed that the particle sizes are below 10 nm, which is in good concurrence with the XRD results. This means that each particle is composed of only one crystallite. In the case of the Y-doped ceria sample, the presence of some aggregates with an acicular shape should be noticed. The changing of nanoparticles’ shapes by introducing different types of dopants indicate that the crystallisation process can be tailored.

The absorption spectrum of nanoparticles in the UV-Vis-NIR domain is affected by various factors such as optical band gap, oxygen vacancies, the presence of an impurity centre, etc. Figure 3a shows the absorbance spectra of pure and RE-doped ceria nanostructures and the spectra exhibit a strong absorbance in the UV region. In general, the absorption band below 400 nm is assigned to the transitions of charge transfer between O 2p and Ce 4f orbitals in cerium oxide [53]. Doping with RE caused a red shift in the absorbance band of ceria, indicating that 4f or 4d transition metals ions induce intermediate electronic energy levels between the O 2p and Ce 4f orbitals of cerium oxide [54]. To evaluate the energy band gap of the samples, Tauc’s relation was applied to the absorption spectra. Figure 3b displays the Tauc plots. The energy band gaps of RE-doped CeO₂ nanoparticles are smaller than those corresponding to ceria (2.98 eV)—Table 1. The band gap narrowing of CeO₂ by doping with RE ions can be explain by the formation of 4f for Gd and Eu or 4d for Y states below the conduction band edge, which extend the absorption of ceria into the visible region.

Photoluminescence spectroscopy is considered an important technique for understanding the role of the recombination of electron pairs and defects in modifying the photocatalytic activity of a material. The registered spectra for the ceria-based samples are presented in Figure 4. The literature data reported that the energy of emitted photons due to the electron transitions from the 4f band to the valence band in CeO₂ lies in the range of 3.0–3.8 eV [21]. The band centred at 364 nm, presented in all spectra, should be attributed to this transition. Also, it is reported that the emission bands located from 400 to 500 nm are attributed to some defects, such as vacancies or dislocations in CeO₂ [21,55]. Therefore, the conclusive transitions from Ce 4f level to O 2p level due to abundant defects like O vacancies and dislocations are responsible for emission peaks from the 430 nm to 494 nm domains observed in all spectra. The replacement of Ce⁴⁺ cations by RE³⁺ cations in the lattice of CeO₂ generates the formation of defects and oxygen vacancies in order to preserve the charge balance.

The photoluminescence spectrum of Eu-doped ceria is significantly different from the other registered spectra, showing well-defined emission bands in the 540–620 region. These bands result directly from the incorporation of the Eu³⁺ ions into the CeO₂ lattice. Depending on the excitation wavelength, possible f-f transitions of Eu³⁺ cations (⁵D₀ → ⁷F_J) take place at 550 nm (J = 0), 593 nm (J = 1), 610–630 nm (J = 2), 650 nm (J = 3), and 710 nm (J = 4) [56]. It is already known that the Ce site in the fluorite CeO₂ lattice presents Oh symmetry with eightfold oxygen coordination. Therefore, the Ce site possesses inversion symmetry and the electric dipole emission ⁵D₀ → ⁷F₂ could be forbidden and could not be observed only if the Eu³⁺ ion occupied the Ce site without changing the original inversion symmetry. When Eu³⁺ replaces Ce⁴⁺, it needs charge compensation in order to obtain a charge equilibrium. Two kinds of defects could be formed during charge compensation: * interstitial Eu³⁺ (four Eu³⁺ replacing three Ce⁴⁺) and **O²⁻ vacancies (one Eu₂O₃ replacing two CeO₂). These defects induce the symmetry distortion of the local environment around the Eu³⁺ ions. Moreover, due to the difference in ionic radius between Ce⁴⁺ and Eu³⁺, the replacement of Ce⁴⁺ ions with Eu³⁺ ions also generates the symmetry distortion of the local environment around the Eu³⁺ ions [56].

The presence in the spectrum (λ_exc = 325 nm) of ⁵D₀ → ⁷F₀ transition, around 549 nm, ⁵D₀ → ⁷F₁ transition at 592 nm and ⁵D₀ → ⁷F₂ at 606 and 612 nm, respectively, suggests that Eu³⁺ cations replace Ce⁴⁺ sites in the CeO₂ lattice and moreover is indicative of the formation of oxygen vacancies [57,58,59,60].

To evidence the presence of paramagnetic species, the EPR technique was used. The spectra were similar, and as an example the experimental spectrum of CeO₂-Y nanoparticles obtained at room temperature is illustrated in Figure 5a. The spectrum shows a signal at a higher value than the free spin value (g_e = 2.0023), and two signals at lower g values. The EPR signal at g = 2.00935 was assigned to physisorbed oxygen molecules or to paramagnetic species such as •${\mathrm{O}}_2^{-}$ generated from these oxygen molecules [61], and that centred g = 1.964 has been attributed both to quasi-free electrons and to Ce³⁺ ions [62]. Both paramagnetic impurities are due to the non-complete stoichiometry obtained during the synthesis process.

The effect of the charge separation in semiconducting oxides under irradiation can be probed by EPR. Therefore, the CeO₂-Y sample was irradiated with UV to check the modification of the spectrum (Figure 5a). Under irradiation, only the resonance signal centred at g = 2.00935 changes its intensity. This behaviour can be explained by the following: when CeO₂ is irradiated with photons owing energy higher than that of the band gap, electron–hole pairs are generated. If these pairs do not recombine, the electrons are stabilized on cations, while the holes are stabilized by oxygen anions [63]. Thus, the photoexcited electrons are stabilized by Ce⁴⁺ ions, being reduced to Ce³⁺; the increases of the resonance signal attributed to this signal are not visible because of the high value of spin orbit coupling and therefore the very fast relaxation time [64].

To illustrate the charge generation and recombination, the intensity of the EPR signal centred at g = 2.00935 was monitored under UV irradiation in three on–off cycles. The obtained kinetic is exhibited in Figure 5b. It can be observed that when the UV is turned on, the EPR signal increased faster and charge carriers stayed at the excited energy level; after turning off the light, the signal decreased, but a little slowly, which means that the recombination process is longer. The on–off cycle experiments show that the process is reversible.

The photocatalytic activity of the samples was investigated against Rhodamine B (RhB) solution under visible irradiation. Prior to the photocatalytic experiments, the samples dispersed in solution were kept for 60 min in the dark to achieve an adsorption–desorption equilibrium. The RhB removal rate was monitored by the temporal evolution of RhB absorbance. As an example, Figure 6 shows the evolution in time after dark adsorption and the visible irradiation of RhB absorbance in the presence of the CeO₂-Gd sample. It could be observed that the RhB absorbance decreases with irradiation time. The decrease in the absorption peak is due to the cleavage of the RhB chromophore ring structure, and the shift of the absorption maximum results in the N-deethylation of RhB molecules [65].

The RhB removal rate in the presence of all the samples is depicted in Figure 7. All doped samples possess enhanced adsorption capacity relative to ceria nanoparticles. Figure S2 shows the adsorption capacity of Y dopedCeO₂ sample for 420 min which proves that after 60 min, the adsorption equilibrium was reached. The enhanced dark adsorption capacity of doped samples could be due to much more active sites for the dye molecules to react. The removal rate of the CeO₂-Gd sample is almost three times higher than that corresponding to ceria.

The photocatalytic process could be described by the first order kinetic model following the equation

$$-ln\frac{A_t}{A_0^{*}}=k_i·t$$

where A_t—absorbance of RhB at t time; $A_0^{*}$—absorbance of RhB after dark adsorption; t—irradiation time; and k_i—apparent kinetic constant. The obtained plots show a linear relationship with the irradiation time (Figure 8). In

Table 2. The apparent first-order rate constant k of photocatalytic degradation and correlation coefficient, R².

Sample	K × 10−3	R2
CeO2	0.37	0.9860
CeO2:Eu	1.13	0.9441
CeO2:Gd	1.09	0.9244
CeO2:Y	1.24	0.9594

are shown the calculated rate constants and correlation coefficients corresponding to Figure 8. From the obtained k values, it can be seen that the best performance was obtained using Y-CeO₂, followed by Eu-CeO₂ > Gd-CeO₂ > CeO₂.

The photocatalytic mechanism can be explained based on the formation of electron–hole pairs. Under visible irradiation, electrons absorb energy and move from the valence band to the conduction band, resulting in the appearance of electron–hole pairs. The electrons interact with oxygen and produce superoxide radicals and the holes interact with H₂O, resulting in hydroxyl radicals. These radicals are known as radical oxygen species and are capable of degrading organic pollutants. The photocatalytic activity depends on the capacity of the photocatalyst to separate the electron–hole pairs. The most important factor is represented by the rate of electron–hole pair separation and recombination. The delay in recombination drives enhanced photocatalytic activity [24].

To evidence the reactive oxygen species, ESR coupled with spin-trapping experiments was performed. The experimental spectrum is shown in Figure 9. The obtained spectrum is a complex one, consisting of many resonance lines which are due to several species. To identify the species, a simulation of the spectrum was carried out. The simulated spectrum is also shown in Figure 9 and was obtained through a linear combination of the following spin adducts: ●DMPO–OCH₃ (a_N = 13.3 G, ${\mathrm{a}}_{\mathrm{H}}^{\mathsf{\beta }}$ = 8.3 G, ${\mathrm{a}}_{\mathrm{H}}^{\mathsf{\gamma }}$ = 1.9 G, relative concentration 60%), ●DMPO–${\mathrm{O}}_2^{-}$ (a_N = 12.9 G, ${\mathrm{a}}_{\mathrm{H}}^{\mathsf{\beta }}$ = 11.4 G, ${\mathrm{a}}_{\mathrm{H}}^{\mathsf{\gamma }}$ = 2.6 G, relative concentration 31%) and nitroxide-like radicals (a_N = 13.9 G, relative concentration 9%).

The proof of the ●OH radical existence in the system is due to the presence of ●DMPO–OCH₃ spin adducts, since these spin adducts are generated by the reaction between ●OH radicals and DMSO solvent [66]. Nitroxide-like radicals were generated by the cleavage of the NAC bond followed by the ring opening of DMPO [67].

The roles of scavengers in elucidating the active radical species involved in photodegradation are to capture and suppress the action of the species under investigation during the photocatalytic process. The scavenger experiment was performed to evaluate the active radical species that participated in the photocatalytic process. Ethylenediaminetetraacetic acid (EDTA) and vitamin C (Vit C) were used as h⁺ and •${\mathrm{O}}_2^{-}$ radical scavengers, respectively [68]. The scavenger experimental conditions were the same as those used for the RhB degradation and 1.0 mM scavenger was added to the solution. In Figure 10, all samples showed lower RhB degradation than the control experiment (with no scavenger), which attested that these radical species are involved in the photocatalytic process. The dominant active species showed the weakest photocatalytic activity. Therefore, the lowest RhB degradation was obtained with Vit C, indicating that •${\mathrm{O}}_2^{-}$ radicals are the main species involved in RhB photodegradation.

Based on the above studies, the photocatalytic mechanisms of Y-, Gd- and Eu-doped CeO₂ nanoparticles are illustrated in Figure 11. The doping of CeO₂ results in a higher concentration of oxygen vacancies and additional doping transition levels. When the doped CeO₂ is irradiated with visible light, the photogenerated electrons will move in the conduction band and the holes will be generated in the valence band. Part of the electrons will be trapped on the doping levels, limiting the recombination of the photoinduced electron and hole pairs [69]. Therefore, the electrons from the conduction band will interact with oxygen-producing superoxide radicals and the holes will interact with H₂O, generating hydroxyl radicals. Both generated radical species will participate in a redox reaction, being capable of degrading the RhB molecule. From scavenger experiments it can be found that •${\mathrm{O}}_2^{-}$ radicals are the main species required to degrade RhB dye; indeed, •${\mathrm{O}}_2^{-}$ radicals are generated by the sample, but hydroxyl radicals are the majority, which explains the not-very-high photocatalytic performances of the samples.

The HPLC method has been used as a complementary method to determine the concentration in RhB of the solutions formed after the UV photodegradation process. The obtained concentration values for all samples are summarized in

Table 3. HPLC-FD concentrations of RhB before and after photodegradation with CeO₂-doped catalysts (CeO₂:Eu; CeO₂:Gd; CeO₂:Y).

Catalyst	No Catalyst	CeO2	CeO2:Eu	CeO2:Y	CeO2:Gd
Rhodamine B [mg/L]	4.640	2.440	1.386	0.700	0.430

and the chromatograms are presented in Figure S1—Supplementary Material.

The obtained results suggest a decreasing tendency for RhB concentration after the photodegradation of the pollutant solution with CeO₂-doped catalysts (CeO₂ > CeO₂:Eu > CeO₂:Y > CeO₂:Gd), sustaining the previous results obtained using UV–Vis spectroscopy.

4. Conclusions

In summary, CeO₂ nanoparticles doped with rare earth ions (Eu³⁺, Gd³⁺ and Y³⁺) were synthetized via a green hybrid sol-gel combustion method. The nanoparticles have a face-centred cubic (fcc) fluorite structure, and the slight increase in the cell parameters sustains the replacement of Ce⁴⁺ with dopant ions. Undoped CeO₂ has a small crystallite size (around 8 nm) and by doping a slight increase was noticed, except for Y ions. The particle size evaluated by TEM is in accordance with the crystallite size evaluated by XRD, sustaining the formation of monocrystalline particles. By doping the band gap, the energy decreases, which shows that additional levels are introduced by doping into the CeO₂ band gap. The EPR spectra evidence similar behaviour for all doped samples. The intense signal at g~2.0093 was assigned to physisorbed oxygen molecules or to paramagnetic species such as •${\mathrm{O}}_2^{-}$ generated from these oxygen molecules, and a lower one at g~1.964 was due to quasi-free electrons or Ce³⁺ ions. Under UV light irradiation, the signal at g~2.0093 increases in intensity, which suggests an efficient charge separation by capturing the photogenerated holes with oxygen anions. The charge generation and recombination processes are reversible. The RhB removal rate was tested under UV light irradiation. The removal of pollutant molecules is due to two processes, adsorption and photocatalysis. The best performance was obtained using Gd-CeO₂, with the adsorption playing an important role. But the best photocatalytic activity was obtained for Y-doped CeO₂. The spin-trapping experiment revealed the generation by Y-CeO₂ nanoparticles of both •OOH and •${\mathrm{O}}_2^{-}$ radicals involved in RhB photodegradation. In our future works, studies of the photocatalytic degradation of other organic pollutants will be considered.