Enhanced Oxygen Storage Capacity of Porous CeO₂ by Rare Earth Doping

Abstract

CeO₂ is an important rare earth (RE) oxide and has served as a typical oxygen storage material in practical applications. In the present study, the oxygen storage capacity (OSC) of CeO₂ was enhanced by doping with other rare earth ions (RE, RE = Yb, Y, Sm and La). A series of Undoped and RE–doped CeO₂ with different doping levels were synthesized using a solvothermal method following a subsequent calcination process, in which just Ce(NO₃)₃∙6H₂O, RE(NO₃)₃∙nH₂O, ethylene glycol and water were used as raw materials. Surprisingly, the Undoped CeO₂ was proved to be a porous material with a multilayered special morphology without any additional templates in this work. The lattice parameters of CeO₂ were refined by the least–squares method with highly pure NaCl as the internal standard for peak position calibrations, and the solubility limits of RE ions into CeO₂ were determined; the amounts of reducible–reoxidizable Ceⁿ⁺ ions were estimated by fitting the Ce 3d core–levels XPS spectra; the non–stoichiometric oxygen vacancy (V_O) defects of CeO₂ were analyzed qualitatively and quantitatively by O 1s XPS fitting and Raman scattering; and the OSC was quantified by the amount of H₂ consumption per gram of CeO₂ based on hydrogen temperature programmed reduction (H₂–TPR) measurements. The maximum [OSC] of CeO₂ appeared at 5 mol.% Yb–, 4 mol.% Y–, 4 mol.% Sm– and 7 mol.% La–doping with the values of 0.444, 0.387, 0.352 and 0.380 mmol H₂/g by an increase of 93.04, 68.26, 53.04 and 65.22%. Moreover, the dominant factor for promoting the OSC of RE–doped CeO₂ was analyzed.

1. Introduction

Rare earth (RE), known as “Industrial vitamin”, “Industrial monosodium glutamate” and “Mother of new material”, has irreplaceable excellent magnetic, optical, and electrical properties, playing a huge role in improving product performance, increasing product variety and improving production efficiency. Although the amount is small, it can greatly optimize the properties of materials. In view of its large effect and low dosage, RE has become an important national strategic resource in improving product structure, increasing technological content, and promoting industry technological progress, and is broadly utilized in many fields, such as metallurgy, military, petrochemical, glass ceramics, agriculture and new materials, and so on [1,2,3]. Cerium (Ce) is the most abundant RE element in the crust of Earth, which has good redox performance, so that its oxide (cerium oxide, CeO₂) shows excellent oxygen transport capacity and oxygen storage/release capacity. Moreover, CeO₂ has the advantages of low toxicity and reusability, so it has attracted great attention in the detection of food biological and chemical substances, catalysis and fuel cell fields [4,5].

Oxygen storage materials are binary or multicomponent composite oxides, in which a CeO₂ and CeO₂–based solid solution are the main components. CeO₂ is a significant N–type semiconductor material with high electrical conductivity, an excellent oxygen storage/release capacity and strong redox activity. Moreover, CeO₂–based oxygen storage materials are one of the key materials in a three–way catalyst for automobile exhaust purification [6,7], as well as in water–gas–shift [8,9,10], ethanol steam reforming [11,12] and hydrocarbon reforming [13,14,15]. In an oxygen–rich environment, CeO₂ can capture ambient oxygen into its own lattice, and release these stored oxygen quickly when the oxygen content of the reaction system is reduced. Because of this, CeO₂–based oxygen storage materials can even determine the performance and service life of a catalyst [16]. Especially in the heterogeneous catalytic reactions, they can regulate the fluctuation of the oxygen content in the reaction system through their own oxygen storage and oxygen release characteristics, which can always maintain the best catalytic effect. This ability of CeO₂–based composite oxides to store and release oxygen is called its oxygen storage capacity (OSC). The oxygen evolution and absorption equilibrium reaction can be described by Reaction (1) [17,18]:

$$\mathrm{C}\mathrm{e}{\mathrm{O}}_2\underset{\mathrm{O}\mathrm{x}\mathrm{y}\mathrm{g}\mathrm{e}\mathrm{n} \mathrm{s}\mathrm{t}\mathrm{o}\mathrm{r}\mathrm{a}\mathrm{g}\mathrm{e}}{\overset{\mathrm{O}\mathrm{x}\mathrm{y}\mathrm{g}\mathrm{e}\mathrm{n} \mathrm{r}\mathrm{e}\mathrm{l}\mathrm{e}\mathrm{a}\mathrm{s}\mathrm{e}}{\rightleftarrows }}\mathrm{C}{\mathrm{e}}_{1-x}^{4+}\mathrm{C}{\mathrm{e}}_x^{3+}{\mathrm{O}}_{2-x/2}{V}_{x/2}+x/4{\mathrm{O}}_2$$

(1)

where “V_O” represents the oxygen vacancy defects produced via the vacancy compensation mechanism. Interestingly, CeO₂ can exhibit a large deviation from stoichiometry at low oxygen partial pressure, forming nonstoichiometric oxide CeO_2−x. Even after the loss of oxygen from the lattice and the consequent formation of numerous V_O, CeO_2−x still retains a fluorite crystal structure [19,20] and captures oxygen by filling the V_O upon exposure to oxygen, accompanied by the recovery of CeO₂ [21]. Moreover, the doping of other metallic elements into the CeO₂ lattice could control their structure and physical properties [22,23,24], such as rare–earth elements [25,26,27], transition elements [28,29,30] and alkaline earth elements [31,32,33]. In spite of the successful synthesis of CeO₂–based composite oxides, most of the previous reports have focused on the investigation of catalytic performances [34,35], transport properties [36,37] and the origin of room–temperature ferromagnetism [38,39], the theoretical data about OSC were usually quite scattered, and only a few fundamental studies on the OSC of doped CeO₂ have been reported. For example, Singh [40] et al. synthesized a series of Ce_1−xM_xO_2−σ (M = Zr, Ti, Pr, Y and Fe) nanocrystallites using the hydrothermal method using melamine and diethylenetriamine as complexing agents; up to 50% Zr and Y, 40% Ti, 25% Pr and 15% Fe were substituted for Ce⁴⁺ in CeO₂, and Ce_0.85Fe_0.15O_1.85 showed a higher OSC and higher CO conversion at a lower temperature than Ce_1−xZr_xO₂. Ansari et al. [41] reported the redox properties of Fe–doped CeO₂ nanoparticles obtained by a polyol–assisted co–precipitation process, and the 10 mol.% Fe doped CeO₂ nanoparticles exhibited excellent reduction performance. Si et al. [42] prepared Ce_1−xZr_xO₂ (x = 0~0.8) powders via a mild urea hydrolysis based on the hydrothermal method, and validated a linear relationship between the lattice strain and the OSC value of CeO₂–ZrO₂ solid solutions. Therefore, the microstructure and OSC of doped CeO₂ have to be understood at a fundamental level through a series of dopants to design advanced materials.

For that, four rare earth elements (RE = Yb, Y, Sm and La) were selected as dopants to improve the OSC of CeO₂ based on the similarity–intermiscibility theory. In order to avoid the influence of other ions on the doping effect, we only used Ce(NO₃)₃∙6H₂O, RE(NO₃)₃∙nH₂O, ethylene glycol and water as raw materials. Moreover, all experimental conditions and the purity of raw materials were the same, so, the comparison of structure and properties of RE–doped CeO₂ was reliable and effective. Based on this, the influence of the dopant elements and their amounts on the non–stoichiometric V_O and OSC were investigated and discussed. Surprisingly, the undoped CeO₂ was proved to be a porous material with a multilayered morphology without any additional templates, and the effect of RE–doping on morphology of CeO₂ also was investigated.

2. Experimental Procedure

2.1. Starting Materials

Ce(NO₃)₃∙6H₂O (99.95%), Yb(NO₃)₃∙5H₂O (99.9%), Y(NO₃)₃∙6H₂O (99.9%), Sm(NO₃)₃∙6H₂O (99.9%) and La(NO₃)₃∙6H₂O (99.9%) were supplied by Aladdin Co., Ltd. (Shanghai, China). Ethylene glycol (99.5%) and ethanol (99.7%) were obtained from Chengdu Kelong Chemical Co., Ltd. (Chengdu, China). Distilled water was used in all experiments.

2.2. Synthesis of Undoped and RE–Doped CeO₂

Firstly, the desired amounts of Ce(NO₃)₃∙6H₂O and RE(NO₃)₃∙nH₂O (RE = Yb, Y, Sm and La) with different RE/(RE + Ce) (mol.%) were dissolved in a mixed solution of 25 mL ethylene glycol and 5 mL distilled water, the total amount of Ce³⁺ and RE³⁺ ions was 4.0 mmol. Then, the mixed solution was decanted into a 50 mL Teflon–lined stainless steel autoclave and sealed. Subsequently, the solvothermal process lasted for 24 h at 200 °C. After the reaction, the resulting precipitates were collected by centrifugation, and washed thrice alternately with distilled water and ethanol. At this point, the precursors synthesized by the hydrothermal process were obtained after drying in air at 80 °C for 12 h. Finally, a series of RE–doped CeO₂ powders were obtained by following calcination in air at 500 °C for 2 h. For comparison, the Undoped CeO₂ was synthesized using the same procedure, albeit in the absence of dopants RE(NO₃)₃∙nH₂O.

2.3. Characterization

The actual doping amounts of RE elements in CeO₂ were determined using an inductively coupled plasma–atomic emission spectrometer (ICP–AES, SPECTRO ARCOS EOP, Kleve, Germany). The crystallographic phases of samples were characterized by X-ray diffraction (XRD, Rigaku D/MAX 2200 PC, Rigaku, Japan) analysis using graphite monochromatized Cu Karadiation with 40 kV tube voltages and a 40 mA current. The morphologies of CeO₂ were observed by field–emission scanning electron microscopy (SEM, JEOL–7500F, Tokyo, Japan). The surface composition and binding energy of CeO₂ were determined by X-ray photoelectron spectroscopy (XPS, ESCALAB 250, Thermo Scientific, Waltham, MA, USA). The natures of surface V_O defects were identified using Raman spectroscopy (LabRAM Aramis, Horiba Jobin Yvon, Paris, France) with a He–Cd laser of 325 nm. N₂ adsorption–desorption isotherms were measured on a QuadraSorb SI (Quantachrome, Boynton Beach, FL, USA), and the specific surface areas were determined using the Brunauer–Emmett–Teller method.

2.4. Evaluation of OSC

For the Undoped and RE–doped CeO₂ samples synthesized using the hydrothermal process at 200 °C for 24 h, followed by calcination in air at 500 °C for 2 h, the hydrogen temperature programmed reduction (H₂–TPR) measurements were employed to evaluate their OSC, which was carried out on a TP–5080 instrument with a thermal conductivity detector of gas chromatography. Typically, 50 mg CeO₂ powder was pre–treated in a 5% O₂/N₂ stream at 500 °C for 1 h. After cooling down, the sample was purged with N₂ to remove the excess O₂. Then, a flow of 5% H₂/N₂ was introduced into the reactor with a flow rate of 30 mL/min, and the temperature was raised to ~650 °C with a heating rate of 10 °C/min.

3. Results and Discussion

XRD analyses were employed to identify the phase composition and crystallographic structure of the as–obtained precursors and samples. Figure 1a showed the XRD patterns of the precursors synthesized using the solvothermal process at 200 °C for 24 h before calcination. For the precursor synthesized without RE, its major phase component was CeCO₃OH (JCPDS no. 52–0352), and similar profiles were observed for these precursors synthesized with the introduction of 10 mol.% RE in the solvothermal process. Figure 1b showed the XRD patterns of Undoped and 10 mol.% RE–doped samples synthesized at 200 °C for 24 h after calcination in air at 500 °C for 2 h. All identified peaks had a good match with the standard CeO₂ pattern (cubic fluorite structure, JCPDS no. 34–0349), and the intensities of the corresponding diffraction peaks were comparable. Moreover, no impurity phases were detected, such as Yb₃O₄, Y₃O₄, Sm₃O₄ and La₃O₄. The absence of RE impurity phases could be explained as follows. The RE impurity phases in the sample might exist as highly dispersed amorphous species. Another possibility was that the RE impurity ions partially substituted the host Ce ions to form a solid solution. Compared with Undoped CeO₂, the relative diffraction intensities of 10 mol.% RE–doped CeO₂ showed no clear differences, suggesting that there was no preferential orientation or preferential crystal growth upon the incorporation of RE. In addition, compared with Undoped CeO₂, a recognizable peak shift towards lower diffraction angles for 10 mol.% RE–doped CeO₂ was observed. These findings indicate that the larger RE impurity ions partially substituted the host Ce ions to form the RE–based solid solution based on Bragg’s equation, and the cubic fluorite crystal structure of CeO₂ was maintained.

When the impurity ions were introduced into the lattice of the matrix, its lattice parameter (a) would change. So, the change in the a value could be used to determine the solubility limit of these dopants in the matrix. In this work, the a value of CeO₂ was refined by the least–squares method, in which the highly pure NaCl (≥99.999%) was selected as an internal standard to calibrate the peak position of CeO₂. Figure 2a showed the XRD patterns of Undoped CeO₂ and 10 mol.% RE–doped CeO₂ with the internal standard of NaCl. It could be found that the diffraction intensities of (111) peak from CeO₂ and (200) peak from NaCl were comparable, suggesting the feasibility of this internal standard method. Moreover, the a values of Undoped and 1~10 mol.% RE–doped CeO₂ were calculated, and the calculated a as a function of RE contents in CeO₂ were summarized in Figure 2b. From Figure 2b, the a values of all RE–doped CeO₂ were greater than that of the Undoped one (5.4117 Å). Under the same doping concentration, the variation trend of a values was as follows: a_Yb < a_Y < a_Sm < a_La, which was consistent with the sequence of their ionic radii for CN8: R_Ce (0.97 Å) < R_Yb (0.98 Å) < R_Y (1.02 Å) < R_Sm (1.08 Å) < R_La (1.16 Å) according to Shannon’s compilation [43]. The increased a values after the introduction of RE indicated that the partial host Ce⁴⁺ (0.97 Å) ions substituted by the larger RE ions and the local lattice expansion of CeO₂ crystal occurred as a result. Moreover, the a values linearly increased with increasing RE contents, reached a maximum at 5, 4, 4 and 7 mol.% for Yb, Y, Sm and La, before decreasing and maintaining a certain level for higher RE contents. This would indicated that the solubility limits of Yb, Y, Sm and La ions in CeO₂ were 5, 4, 4 and 7 mol.%.

In order to further confirm the incorporation of RE ions and their effect on the CeO₂ lattice, high–resolution electron microscopy (HR–TEM) was performed and the corresponding HR–TEM images of Undoped and 10 mol.% RE–doped CeO₂ were synthesized using the hydrothermal process at 200 °C for 24 h, followed by calcination in air at 500 °C for 2 h, as shown in Figure 3. From the HR–TEM image of Undoped CeO₂ in Figure 3a, the interplanar spacing was measured with a value of 0.3110 nm, which fitted well with the (111) plane of cubic CeO₂, proving the generation of the CeO₂ phase. After the incorporation of 10 mol.% RE (RE = Yb, Y, Sm and La), the interplanar spacings of CeO₂ in Figure 3b–e had increased to 0.3178, 0.3202, 0.3209 and 0.3231 nm, respectively. Combined with XRD analysis results in Figure 2, both the local lattice expansion and the increased interplanar spacing indicated that these large RE (R_Yb = 0.98 Å; R_Y = 1.02 Å; R_Sm = 1.08 Å; R_La = 1.16 Å) impurity ions partially substituted the host Ce ions (R_Ce = 0.97 Å), and a solid solution was formed. Importantly, the size of the RE impurity ions was consistent with the trends of interplanar spacing. In other words, the larger the size of the doped RE ion, the greater the interplanar spacing of the as–obtained RE–doped CeO₂. In addition, the practical RE contents in CeO₂ were measured by ICP–AES, and the results are shown in

Table 1. Practical contents and nominal contents of RE in CeO₂ synthesized using the hydrothermal process at 200 °C for 24 h and followed by calcination in air at 500 °C for 2 h (RE = Yb, Y, Sm and La).

RE in CeO2 (mol.%)	Yb			Y			Sm			La
* Nominal contents	2	5	9	2	4	9	2	4	9	2	7	9
* Practical RE contents	2.58	5.26	9.62	1.92	4.24	8.67	2.41	4.27	9.38	2.19	6.79	9.25

* Nominal content (mol.%): $w (\mathrm{mol}.\%)=\frac{n_{\mathrm{R}\mathrm{E}}}{n_{\mathrm{R}\mathrm{E}}+n_{\mathrm{C}\mathrm{e}}}\times 100, (n_{\mathrm{R}\mathrm{E}}+n_{\mathrm{C}\mathrm{e}}=4_{} \mathrm{mmol})$. * Practical RE contents (mol.%): The actual RE doping amounts in CeO₂ were determined using ICP–AES, where CeO₂ was dissolved in a mixed solution of HNO₃–H₂O₂.

. As observed in Table 1, it could be found that the practical RE contents in CeO₂ were close to the corresponding nominal doped one.

XPS analysis was employed to probe the surface chemical composition and various oxidation states before and after RE–doping. Figure 4a–e shows the wide–scan XPS spectra of Undoped and 4 mol.% RE–doped CeO₂ synthesized using the hydrothermal process at 200 °C for 24 h and followed by calcination in air at 500 °C for 2 h, respectively. As observed, all wide–scan XPS spectra showed the clear CeO₂ features by the signals of Ce 3d, Ce 4d and O 1s, in good agreement with those XPS patterns of Gd– [44], Y– [45] and Dy– [46] doped CeO₂. It is worth noting that the obvious C 1s peaks located at ~284.8 eV were derived from adventitious carbon to calibrate the tested samples. Moreover, the faint RE 3d or RE 4d signals can be seen in the red dotted box in Figure 4b–e, and the corresponding Yb 4d, Y 3d, Sm 3d and La 3d XPS regions were recorded, as shown in Figure 4f–i, respectively. The characteristic peaks in Figure 4f–i implied that the Yb, Y, Sm and La elements were in +3 states. It indicated that the Yb, Y, Sm and La elements had been successfully incorporated into the CeO₂ lattice with positive trivalent states (RE³⁺).

In order to understand the effect of RE–doping on Ce ions in the CeO₂ crystal, the Ce 3d XPS regions of Undoped and 4 mol.% RE–doped CeO₂ were recorded and fitted, as shown in Figure 5a–e. The Ce 3d XPS core–levels of all CeO₂ samples were fitted into eight peaks, corresponding to four pairs of spin–orbit doublets (u_1–4 and v_1–4) of Ce ions, in which the u_i and v_i bands corresponded to the contributions of Ce 3d_3/2 and Ce 3d_5/2. Moreover, the bands of u₄, u₃ and u₁ (and those for v₄, v₃, v₁) were attributed to the Ce⁴⁺ state, while u₂ and v₂ were due to the Ce³⁺ state [47]. Meanwhile, the relative concentration of Ce³⁺ ions in CeO₂, labeled as [Ce³⁺]_XPS, could be calculated by the ratio of integrated peak areas of the peak related to the Ce³⁺ species (u₂ and v₂ peaks) to that of all peaks (u_1–4 and v_1–4 peaks) in Figure 5, and the results were summarized in

Table 2. [Ce³⁺]_XPS and [V_O]_XPS of Undoped CeO₂ and 4 mol.% RE–doped CeO₂ synthesized using solvothermal method at 200 °C for 24 h followed by calcination in air at 500 °C for 2 h (RE = Yb, Y, Sm and La).

	Sample	Undoped CeO2	4 mol.% RE–Doped CeO2
Parameter			Yb	Y	Sm	La
[Ce3+]XPS (%)		6.54	13.78	12.60	10.94	9.78
[VO]XPS (%)		13.42	30.02	26.82	26.81	17.28
Specific surface area (m2/g)		96.0	89.7	98.1	112.6	104.6

. As observed, the [Ce³⁺]_XPS values of 4 mol.% Yb, Y, Sm and La–doped CeO₂ were 13.78, 12.60, 10.94 and 9.78%, respectively, higher than that of Undoped CeO₂ (6.54%), which indicates that Undoped CeO₂ itself contained a certain number of Ce³⁺ ions, and RE–doping could promote the formation of Ce³⁺ species. In other words, the amount of reducible–reoxidizable Ceⁿ⁺ (namely, Ce³⁺ ⇔ Ce⁴⁺) ions increased with the introduction of RE ions into CeO₂ lattice, indicating that RE–doping was conducive to improving the redox capacity of CeO₂.

To investigate the chemical states of O in CeO₂, the O 1s core–level XPS spectra of Undoped and 4 mol.% RE–doped CeO₂ were recorded and fitted, as shown in Figure 6a–e. For Undoped CeO₂ in Figure 6a, its O 1s XPS spectrum could be curve–fitted into three peaks, indicating the presence of three kinds of oxygen species in CeO₂. The peaks with a binding energy of ~529.8 and ~528.4 eV could be assigned to lattice oxygen of O–Ce(IV) species and O–Ce(III) species, respectively, whereas that of ~531.6 eV (yellow region peak) could be assigned to the chemisorption of oxygen or/and weakly bonded oxygen species related to V_O defects. For the O 1s spectra of RE–doped CeO₂ in Figure 6b–e, besides the above three peaks, a new curve fitting could be observed, which might be attributed to the corresponding O–RE species, namely, the O–Yb species at ~527.6 eV, O–Y species at ~528.2 eV, O–Sm species at ~528.2 eV and O–La species at ~532.9 eV. Furthermore, the relative V_O content could be estimated by the ratio of the integrated area of the peak related to the V_O defect (yellow region peak in Figure 6a–e) to that of all peaks, labeled as [V_O]_XPS, and the results were summarized in Table 2. As observed in Table 2, the calculated [V_O]_XPS values of 4 mol.% Yb, Y, Sm and La–doped CeO₂ were 30.00, 26.82, 26.81 and 17.28%, respectively, higher than that of the Undoped one (13.42%). This result indicated that RE–doping was beneficial for the V_O creation in CeO₂.

From the results of XPS analyses in Figure 4, Figure 5 and Figure 6 and Table 2, it could be concluded that RE elements were successfully incorporated into the CeO₂ lattice with positive trivalent states, and RE–doping could increase the amount of redox Ceⁿ⁺ (Ce³⁺/Ce⁴⁺) of CeO₂, as well as the V_O defects.

Due to its sensitivity to the V_O defect, Raman scattering was employed to investigate the structure of Undoped and RE–doped CeO₂ synthesized using the hydrothermal process at 200 °C for 24 h and followed by calcination in air at 500 °C for 2 h [48,49]. For the Undoped CeO₂ in Figure 7a, the peak at ~458 cm⁻¹ was attributed to the triply degenerate F_2g mode from the symmetric O–Ce–O stretching mode [50], while the weak peak at ~592 cm⁻¹ was assigned to the optical LO mode related to V_O defects [51,52,53]. Upon the incorporation of RE³⁺ ions into the CeO₂ lattice, the band intensity of the F_2g mode decreased, while that of the LO mode related to the V_O defect increased (Figure 7b–e). It indicated that Undoped CeO₂ itself had a certain number of V_O defects and RE–doping could favor the presence of substoichiometric CeO_2−x underscoring an increase in V_O defects, as consistent with the analysis results of O 1s core–level XPS spectra in Figure 6.

The band at ~590 cm⁻¹ in Raman spectra was known to be associated with the V_O defect and has been widely observed in substoichiometric CeO_2−x [54]. From Figure 7a, the band intensity of both the F_2g and LO modes obviously changed upon the incorporation of RE³⁺ ions into the CeO₂ lattice, which was attributed to the increased lattice distortion caused by RE–doping and hence interfered with the vibrations of CeO_2−x. It made the quantitative analysis of V_O defects difficult. For this, an alternative approach to quantitatively estimate the relative contents of V_O defects was adopted by the ratio of the integrated area of the LO mode to that of the F_2g mode from the Raman spectra. Figure 8 showed the calculated relative V_O concentrations of Undoped and 1~9 mol.% RE–doped CeO₂ synthesized by the hydrothermal process at 200 °C for 24 h and followed by calcination in air at 500 °C for 2 h. As observed, there existed a certain amount of V_O defects in Undoped CeO₂, and the calculated value was 0.67, consistent with the analysis results of the O 1s core–level XPS spectra in Figure 6. These intrinsic V_O defects might have evolved from the redox cycle of Ceⁿ⁺ in CeO₂ (Ce³⁺ ⇔ Ce⁴⁺). The relative V_O concentrations increased almost linearly with increasing RE contents, and reached maximum when the RE contents were 5, 4, 4 and 7 mol.% for Yb, Y, Sm and La–doped CeO₂, and gradually decreased above this doping level. Before this turning point, the variation trend of relative V_O concentration under the same doping concentration was as follows: Yb > Y > Sm > La, which was consistent with their electronegativity: χ_Yb (1.26) > χ_Y (1.22) > χ_Sm (1.17) > χ_Ce (1.12) > χ_La (1.11). After the RE³⁺ ions substituted the host Ce ions into the CeO₂ lattice, the bigger its electronegativity, the stronger its ability to attract the surrounding electrons to itself, and the surrounding O²⁻ anions lost electrons more easily, thus resulting in extrinsic V_O defects.

H₂–TPR measurements were employed to evaluate the OSC of CeO₂. Figure 9a–e illustrated the H₂–TPR profiles of Undoped and 4 mol.% RE–doped CeO₂ (RE = Yb, Y, Sm and La) synthesized using the hydrothermal process at 200 °C for 24 h and followed by calcination in air at 500 °C for 2 h. For all CeO₂ samples in Figure 9, one can clearly find a distinct H₂ reduction band from 30 to 610 °C, with the strongest H₂ reduction peak at ~510 °C; the maximum H₂ consumption occurred at 510 °C and then decreased until ~600 °C, and after that it tended to rise. The reduction band from 30 °C to ~600 °C could be attributed to the reduction in surface/subsurface lattice oxygen, which was consistent with these reported results [55,56]. Before 200 °C, the RE–doped CeO₂ in Figure 9b–e exhibited more H₂ consumption than that of the Undoped CeO₂; especially for 4 mol.% Y, Sm and La–doped CeO₂, a minima at 170 °C occurred. This indicated that the specific surface area of CeO₂ played a dominant role in its OSC at low temperatures. To prove this conjecture, we tested the specific surface areas of 4 mol.% Yb, Y, Sm and La–doped CeO₂, and the results were summarized in Table 2. The specific surface areas of 4 mol.% Y, Sm and La–doped CeO₂ were 98.1, 112.6 and 104.6 m²/g, respectively, higher than that of Undoped CeO₂ (96.0 m²/g); however, these decreased after 4 mol.% Yb–doping (89.7 m²/g). Moreover, compared to Undoped CeO₂ in Figure 9a, there appeared to be a visible shoulder from ~350 °C in the H₂–TPR profiles of RE–doped CeO₂ in Figure 9b–e, and the reduction bands of RE–doped CeO₂ at ~600 °C were far higher than the baseline. These phenomena suggested that RE–doping optimized the surface states of CeO₂, thereby enhancing its OSC.

OSC was the fundamental performance of CeO₂ and CeO₂–based oxygen storage materials; so, the quantification of OSC was the key to evaluate their oxygen storage/release property. For that, the OSC was quantified using the amount of H₂ consumption per gram of CeO₂ powders by measuring the corresponding peak areas of H₂–TPR profiles in this work. The quantified OSC (labeled as [OSC], mmol H₂/g CeO₂) from 30 °C to ~600 °C, which was the value of H₂ consumption per gram of CeO₂ powders, is shown in Figure 10. The [OSC] of Undoped CeO₂ was 0.23 mmol H₂/g, indicating that pure CeO₂ itself possessed a certain OSC, which was attributed to the unique structure of its intrinsic V_O defect or the redox cycle of Ce³⁺ ⇔ Ce⁴⁺, supported by the XPS analyses in Figure 5 and Figure 6 and Raman analyses in Figure 7 and Figure 8. For Yb–doped CeO₂, the [OSC] value reached a maximum with a doping level of 5 mol.% and decreased at a higher Yb content. Interestingly, Y–, Sm– and La–doped CeO₂ also showed similar trends, reaching the maximum H₂ consumptions with doping contents of 4, 4 and 7 mol.%, respectively. The [OSC] values of 5 mol.% Yb–, 4 mol.% Y–, 4 mol.% Sm– and 7 mol.% La–doped CeO₂ were 0.444, 0.387, 0.352 and 0.380 mmol H₂/g, with an increase of 93.04, 68.26, 53.04 and 65.22% compared with that of the Undoped one (0.230 mmol H₂/g). These findings indicate that RE–doping could effectively improve the OSC of CeO₂, combined with the H₂–TPR curves. This enhanced OSC of RE–doped CeO₂ could be explained as follows. When RE³⁺ ions were doped into the CeO₂ lattice to substitute host Ce⁴⁺ ions, more V_O defects would be generated to keep the electric neutrality of the fluorite structure, and a substoichiometric solid solution Ce_1–xRE_xO_2−σ (RE = Yb, Y, Sm and La) was formed based on RE–doping. During the H₂ reduction of H₂–TPR, H₂ reacted with a chemisorbed oxygen from the CeO₂ surface, which was fixed by intrinsic and extrinsic V_O defects on the CeO₂ surface. As the surface chemisorbed oxygen was gradually consumed, the intrinsic and extrinsic V_O defects were exposed, and the bulk lattice oxygen began to move to the CeO₂ surface for replenishment by the V_O defects. The oxygen in the bulk RE–doped CeO₂ diffused more easily to the surface to fill the V_O defects than that in the Undoped CeO₂ due to the activation effect of RE³⁺ dopants which induced oxygen mobility [57].

In order to investigate the effect of RE–doping on the morphology of CeO₂, SEM was employed. Figure 11a–e showed the SEM images of Undoped and 10 mol.% Yb, Y, Sm and La–doped CeO₂ particles synthesized using the hydrothermal process at 200 °C for 24 h and followed by calcination in air at 500 °C for 2 h, respectively. From Figure 11a, it could be seen that the morphology of the Undoped CeO₂ particle was a multilayered structure consisting of flakes, and these flakes intertwined to form an open porous structure. After the incorporation of 10 mol.% RE (RE = Yb, Y, Sm and La) into CeO₂, the multilayered morphology was still maintained, as seen in Figure 11b–e. This finding indicates that the low concentration of RE–doping had little effect on the morphology of CeO₂. Generally, CeO₂ with a porous structure or special morphology was usually synthesized by a template–based method, in which either surfactants as soft templates or other porous inorganic material as hard templates were used. Surprisingly, the porous CeO₂ with a multilayered morphology was obtained without any additional templates in this work. The abundant porous structure and highly specific surface area would undoubtedly enhance the OSC of CeO₂. Further analysis of the porous structures was conducted using an N₂ adsorption–desorption isotherm, as discussed later.

In order to further demonstrate the porous structure of CeO₂, an N₂ adsorption–desorption experiment was performed, and the N₂ adsorption–desorption isotherm of Undoped CeO₂ is shown in Figure 12a. As observed in Figure 12a, the isotherm was similar to the Langmuir IV(a) type according to the IUPAC classification, and an obvious hysteresis loop was observed in the relative pressure range of 0.4~1.0, attributable to the type H3. It suggests that Undoped CeO₂ was a mesoporous material with a disordered mesoporous structures [58], and the isotherm was consistent with that of other reported porous CeO₂ [59,60,61]. Moreover, the specific surface areas of Undoped CeO₂ and RE–doped CeO₂ with solubility limits were estimated based on the N₂ adsorption–desorption experiment using a Brunauer–Emmett–Teller method, and the results are shown in Figure 12b as a histogram. Combined with the specific surface areas of 4 mol.% RE–doped CeO₂ in Table 2, it can be found that RE–doping had a certain influence on the specific surface area of CeO₂. However, the specific surface area was not the dominant factor for promoting the OSC of RE–doped CeO₂. Among the CeO₂ samples with 4 mol.% RE–doping, 4 mol.% Sm–doped CeO₂ displayed the minimum [OSC] value of 0.352 mmol H₂/g in Figure 10; however, it possessed the maximum specific surface area of 112.6 m²/g in Table 2. Among RE–doped CeO₂ with saturation doping concentration, 5 mol.% Yb–doped CeO₂ exhibited the minimum specific surface area of 93.1 m²/g in Figure 12b; however, it possessed the maximum [OSC] value of 0.444 mmol H₂/g. Alternatively, the morphology was also not a major factor influencing the OSC of RE–doped CeO₂, which is supported by the similar multilayered morphology in Figure 11a–e. Combined with the analyses of morphology and specific surface area of Undoped and RE–doped CeO₂, one conclusion could be drawn that the enhanced OSC might be attributed to the incorporation of positive trivalent RE³⁺ ions into the CeO₂ lattice, and partially substituted the host Ce⁴⁺ ions, promoting the formation of more V_O defects and the oxidation/reduction cycle of Ce³⁺ ⇔ Ce⁴⁺. This result could be supported by the lattice parameter analysis in Figure 2, the O 1s XPS analysis in Figure 6 and the Raman spectra analysis in Figure 7.

4. Conclusions

In summary, a series of RE–substituted CeO₂ was synthesized just using Ce(NO₃)₃∙6H₂O, RE(NO₃)₃∙nH₂O (RE = Yb, Y, Sm and La), ethylene glycol and water as raw materials. The Undoped CeO₂ was proved to be a mesoporous material with a multilayered morphology; both its multilayered morphology and cubic fluorite structure could be maintained even after 10 mol.% RE introduction. The RE elements were successfully incorporated into the CeO₂ lattice with positive trivalent states. RE–doping was beneficial for the oxidation/reduction cycle of Ce³⁺ ⇔ Ce⁴⁺, as well as the creation of extrinsic V_O defects. The solubility limits of Yb, Y, Sm and La ions in CeO₂ were determined as 5, 4, 4 and 7 mol.%. After the incorporation of larger RE³⁺, the lattice expansion of the CeO₂ crystal occurred, and more V_O defects appeared, which could induce the oxygen mobility from bulk to surface, and promote its OSC. The [OSC] values were 0.444, 0.387, 0.352 and 0.380 mmol/g, much higher than that of the Undoped one (0.230 mmol/g), with an increase of 93.04, 68.26, 53.04 and 65.22%, respectively. The enhanced OSC of RE–doped CeO₂ should be attributed to the impurity–induced defects by the substitution of host Ce⁴⁺ with RE³⁺ into CeO_2, rather than the effects of its specific surface area and morphology.