Elucidating the Role of Surface Ce4+ and Oxygen Vacancies of CeO2 in the Direct Synthesis of Dimethyl Carbonate from CO2 and Methanol

Cerium dioxide (CeO2) was pretreated with reduction and reoxidation under different conditions in order to elucidate the role of surface Ce4+ and oxygen vacancies in the catalytic activity for direct synthesis of dimethyl carbonate (DMC) from CO2 and methanol. The corresponding catalysts were comprehensively characterized using N2 physisorption, XRD, TEM, XPS, TPD, and CO2-FTIR. The results indicated that reduction treatment promotes the conversion of Ce4+ to Ce3+ and improves the concentration of surface oxygen vacancies, while reoxidation treatment facilitates the conversion of Ce3+ to Ce4+ and decreases the concentration of surface oxygen vacancies. The catalytic activity was linear with the number of moderate acidic/basic sites. The surface Ce4+ rather than oxygen vacancies, as Lewis acid sites, promoted the adsorption of CO2 and the formation of active bidentate carbonates. The number of moderate basic sites and the catalytic activity were positively correlated with the surface concentration of Ce4+ but negatively correlated with the surface concentration of oxygen vacancies. The surface Ce4+ and lattice oxygen were active Lewis acid and base sites respectively for CeO2 catalyst, while surface oxygen vacancy and lattice oxygen were active Lewis acid and base sites, respectively, for metal-doped CeO2 catalysts. This may result from the different natures of oxygen vacancies in CeO2 and metal-doped CeO2 catalysts.


Introduction
The emission of CO 2 into the earth's atmosphere is continuously increasing with the growing burning of fossil fuels and industrial activities. For example, about 33,890.80 million tons of CO 2 was emitted into the atmosphere in 2018. The global CO 2 concentration in the atmosphere reached 407.65 ppm in September 2019, which is about 1.2 times as high as that of the past 40 years [1]. According to the earth's CO 2 , the concentration of CO 2 further increased to 416.47 ppm on 30 May 2020 [2]. The dramatically increased CO 2 emission results in severe climate change, which brings about a series of issues in terms of ecology, economy, and the environment. Taking these facts into consideration, it is necessary to develop effective approaches to mitigate CO 2 emission.
As an important part of carbon dioxide capture, utilization, and storage (CCUS), the chemical transformation of CO 2 to fuels and value-added chemicals can not only alleviate CO 2 emission but also reduce the dependence on fossil resources, which is of great significance for the sustainable development of energy, the environment, and the economy. Dimethyl carbonate (DMC), as one of the downstream products of methanol, is considered a green chemical, which has been widely applied as an intermediate for organic synthesis, an electrolyte for lithium-ion batteries, and a raw material for the synthesis of polycarbonate [3]. Among various synthesis approaches, the direct synthesis of DMC from The detailed texture parameters are listed in Table 1. As seen in Table 1, the specific surface area and pore volume of CeO2 gradually decreased with an increase in the reduction temperature; however, the pore size of CeO2 presented the opposite trend, which suggests that the mesopore originated from the piled pore. This phenomenon may arise from the aggregation of CeO2 particles induced by high temperature [21]. In contrast, as CeO2-H550 and CeO2-H750 were reoxidized under an air atmosphere at 350 °C, 450 °C, and 550 °C, the texture parameters did not change so much.

XRD Analysis of Catalysts
XRD characterization was performed to analyze the crystal structure of CeO2, and the results are presented in Figure 2. All the samples displayed the characteristic The detailed texture parameters are listed in Table 1. As seen in Table 1, the specific surface area and pore volume of CeO 2 gradually decreased with an increase in the reduction temperature; however, the pore size of CeO 2 presented the opposite trend, which suggests that the mesopore originated from the piled pore. This phenomenon may arise from the aggregation of CeO 2 particles induced by high temperature [21]. In contrast, as CeO 2 -H550 and CeO 2 -H750 were reoxidized under an air atmosphere at 350 • C, 450 • C, and 550 • C, the texture parameters did not change so much.

XRD Analysis of Catalysts
XRD characterization was performed to analyze the crystal structure of CeO 2 , and the results are presented in Figure 2. All the samples displayed the characteristic diffraction peaks of a cubic fluorite structure (JCPDS no. 65-5923), and the peaks at 28.5 • , 33.0 • , 47.5 • ,  As seen in Figure 2A, the intensity and width of the characteristic diffraction peaks of CeO2 did not change obviously as the reduction temperature decreased below 550 °C, indicating the similar crystal size of CeO2, CeO2-H450, and CeO2-H550. However, as the reduction temperature increased beyond 550 °C, the peak intensity was gradually enhanced and the peak width gradually narrowed, indicating the gradually increasing crystal size. The crystal size of each catalyst is presented in Table 2, which was calculated using the Scherrer equation as follows: where K is the Scherrer constant (0.89), λ is 0.15414 nm (the wavelength of Cu Kα radiation), β is the full width at halfmaximum of the CeO2(111) diffraction peak, b is the corrected full width at half-maximum of the instrument, and θ is the diffraction angle [22]. As seen in Table 2, similar crystal sizes of about 7~8 nm were observed for CeO2, CeO2-H450, and CeO2-H550. However, the crystal sizes significantly increased from 7.4 to 46.6 nm with an increase in the reduction temperature from 550 °C to 800 °C, which resulted from the aggregation of CeO2 particles [21,23,24]. In addition, the crystal sizes of CeO2-H550 remained basically about 7~8 nm after reoxidation treatment, while those of CeO2-H750 remained basically about 30 nm after reoxidation treatment. These results revealed that reoxidation treatment does not obviously affect the crystal sizes of CeO2. As seen in Figure 2A, the intensity and width of the characteristic diffraction peaks of CeO 2 did not change obviously as the reduction temperature decreased below 550 • C, indicating the similar crystal size of CeO 2 , CeO 2 -H450, and CeO 2 -H550. However, as the reduction temperature increased beyond 550 • C, the peak intensity was gradually enhanced and the peak width gradually narrowed, indicating the gradually increasing crystal size. The crystal size of each catalyst is presented in Table 2, which was calculated using the Scherrer equation as follows: D = Kλ/[(β − b)cosθ], where K is the Scherrer constant (0.89), λ is 0.15414 nm (the wavelength of Cu Kα radiation), β is the full width at half-maximum of the CeO 2 (111) diffraction peak, b is the corrected full width at half-maximum of the instrument, and θ is the diffraction angle [22]. As seen in Table 2, similar crystal sizes of about 7~8 nm were observed for CeO 2 , CeO 2 -H450, and CeO 2 -H550. However, the crystal sizes significantly increased from 7.4 to 46.6 nm with an increase in the reduction temperature from 550 • C to 800 • C, which resulted from the aggregation of CeO 2 particles [21,23,24]. In addition, the crystal sizes of CeO 2 -H550 remained basically about 7~8 nm after reoxidation treatment, while those of CeO 2 -H750 remained basically about 30 nm after reoxidation treatment. These results revealed that reoxidation treatment does not obviously affect the crystal sizes of CeO 2 .

TEM of Catalysts
TEM was used to characterize the morphology of catalysts. As seen in Figure 3, the morphology of catalysts was irregular, including being polyhedron-and rodlike. The particle size of CeO 2 , CeO 2 -H450, CeO 2 -H550, CeO 2 -H550-O350, CeO 2 -H550-O450, and CeO 2 -H550-O550 seemed almost the same. The possible reason may be that the starting material of CeO 2 was already calcinated at 550 • C for 5 h in a muffle furnace before reduction and reoxidation treatment. However, the particles significantly aggregated with increasing reduction temperature as the reduction temperature increased beyond 550 • C. In addition, reoxidation treatment did not obviously affect the particle sizes of the catalysts. This was consistent with the results of XRD.

TEM of Catalysts
TEM was used to characterize the morphology of catalysts. As seen in Figure 3, the morphology of catalysts was irregular, including being polyhedron-and rodlike. The particle size of CeO2, CeO2-H450, CeO2-H550, CeO2-H550-O350, CeO2-H550-O450, and CeO2-H550-O550 seemed almost the same. The possible reason may be that the starting material of CeO2 was already calcinated at 550 °C for 5 h in a muffle furnace before reduction and reoxidation treatment. However, the particles significantly aggregated with increasing reduction temperature as the reduction temperature increased beyond 550 °C. In addition, reoxidation treatment did not obviously affect the particle sizes of the catalysts. This was consistent with the results of XRD. HRTEM was used to further characterize the exposed facet of catalysts, as shown in Figure 4. The observed lattice fringes of about 0.31 and 0.27 nm corresponded to the (111) and (200) facets of CeO2. Obviously, the main exposed facet was (111), and the (200) facet only appeared in CeO2-H650 and CeO2-H750-O550, which was in line with our previous studies [15,19]. These results also indicated that the reduction temperature and reoxidation treatment do not apparently affect the exposed facet of catalysts. HRTEM was used to further characterize the exposed facet of catalysts, as shown in Figure 4. The observed lattice fringes of about 0.31 and 0.27 nm corresponded to the (111) and (200) facets of CeO 2 . Obviously, the main exposed facet was (111), and the (200) facet only appeared in CeO 2 -H650 and CeO 2 -H750-O550, which was in line with our previous studies [15,19]. These results also indicated that the reduction temperature and reoxidation treatment do not apparently affect the exposed facet of catalysts.

Surface Composition Analysis of Catalysts
XPS characterization was performed to analyze the chemical state of the surface Ce and O of catalysts, as shown in Figure 5. After gaussian fitting, the Ce 3D spectra could be  [16].
In the O 1s spectra, the peaks at ~529.0 eV, ~530.5 eV, and ~531.5 eV were assigned to the surface lattice oxygen (OL), oxygen vacancies (OV), and chemisorbed oxygen species   (O C ), respectively [25,26]. The surface concentration of O V was calculated based on the peak area from gaussian fitting, and the results are listed in Table 3. As seen in Table 3, the concentration of Ce 3+ gradually increased from 13.5 to 17.7% as the reduction temperature increased to 800 • C, while that of Ce 4+ gradually decreased from 86.5 to 82.3%. Moreover, the concentration of O V also increased with an increase in the reduction temperature. These results were consistent with the work reported by our group and Chen et al. that the percentage of Ce 3+ is positively correlated with that of O V [16,19]. In addition, the concentrations of Ce 3+ and O V decreased for CeO 2 -H550 and CeO 2 -H750 after reoxidation treatment.
The concentration of the surface O V of CeO 2 is closely related to the calcination atmosphere, oxygen partial pressure, and calcination temperature [27]. The formation of O V is accomplished through the diffusion of lattice oxygen to the environment, where the content of oxygen is low. Under these circumstances, an oxygen vacancy is a kind of anionic defect. When CeO 2 was exposed to the reductive H 2 atmosphere, the lattice oxygen reacted with H 2 and formed H 2 O and therefore promoted the formation of O V ; moreover, the promoting effect gradually strengthened with increasing reduction temperature. It should be noted that two surplus electrons would be retained in the crystal structure of CeO 2 accompanied with the formation of O V , and the two surplus electrons can bind with two adjacent Ce 4+ , resulting in the conversion of Ce 4+ to Ce 3+ . Moreover, the two electrons were mobile that could migrate between two adjacent Ce ions. In other words, the two quasi-free mobile electrons were trapped around the anionic defect (O V ), which could migrate between adjacent Ce ions and cause electron conduction. A material with this kind of defect is an N-type semiconductor [28]. The O atoms were reinserted back into O V and formed O L when CeO 2 -H550 and CeO 2 -H750 were reoxidized under an air atmosphere. In this process, the trapped electrons around O V were transferred to O atoms, and the conversion of Ce 3+ to Ce 4+ occurred simultaneously. Singhal et al. [29] also found that the conversion between O V and O L is reversible to some extent.
Furthermore, as shown in Figure S1, the color of CeO 2 gradually changed from yellow to brown and finally to gray-black with the increase in the reduction temperature, and the color changed back to yellow to a certain degree when CeO 2 -H550 and CeO 2 -H750 were reoxidized under an air atmosphere. This is because the trapped electrons around O V can absorb light with a certain wavelength and the absorption capacity is gradually enhanced with the increasing number of electrons or O V [21,29]. This result also indicated that reduction treatment is favorable for the formation of O V , while reoxidation treatment plays the opposite role.

Surface Acidity and Basicity Analysis of Catalysts
The surface acidity and basicity of catalysts were characterized using NH 3 -TPD and CO 2 -TPD, as shown in Figures 6 and 7, and the detailed results are listed in Table 4. The desorption peaks in the temperature region of 50~200, 200~400, and 400~600 • C in CO 2 -TPD and NH 3 TPD profiles were attributed to the weak acidic/basic, moderate acidic/basic, and strong acidic/basic sites, respectively [30]. Studies have reported that weak basic sites are mainly ascribed to surface -OH groups, which react with CO 2 and form bicarbonates [31,32]. The moderate basic sites were derived from the Ce-O L or O V -O L ion pairs, which combined with CO 2 and promoted the formation of bridged and bidentate carbonates [8,19,32], while the strong basic sites were mainly assigned to the low-coordination oxygen anions [8,31,32], which resulted in the formation of stable monodentate carbonates [33].

Surface Acidity and Basicity Analysis of Catalysts
The surface acidity and basicity of catalysts were characterized using NH3-TPD and CO2-TPD, as shown in Figures 6 and 7, and the detailed results are listed in Table 4. The desorption peaks in the temperature region of 50~200, 200~400, and 400~600 °C in CO2-TPD and NH3TPD profiles were attributed to the weak acidic/basic, moderate acidic/basic, and strong acidic/basic sites, respectively [30]. Studies have reported that weak basic sites are mainly ascribed to surface -OH groups, which react with CO2 and form bicarbonates [31,32]. The moderate basic sites were derived from the Ce-OL or OV-OL ion pairs, which combined with CO2 and promoted the formation of bridged and bidentate carbonates [8,19,32], while the strong basic sites were mainly assigned to the low-coordination oxygen anions [8,31,32], which resulted in the formation of stable monodentate carbonates [33]. Overall, the surface acidic sites with different natures all decreased with an increase in the reduction temperature and increased with an increase in the reoxidation temperature ( Figure 6 and Table 4). A similar trend was observed for the surface basic sites under various treatment conditions ( Figure 7 and Table 4). A possible reason for the change in surface acidity and basicity is discussed in Section 3.2.

Surface Acidity and Basicity Analysis of Catalysts
The surface acidity and basicity of catalysts were characterized using NH3-TPD and CO2-TPD, as shown in Figures 6 and 7, and the detailed results are listed in Table 4. The desorption peaks in the temperature region of 50~200, 200~400, and 400~600 °C in CO2-TPD and NH3TPD profiles were attributed to the weak acidic/basic, moderate acidic/basic, and strong acidic/basic sites, respectively [30]. Studies have reported that weak basic sites are mainly ascribed to surface -OH groups, which react with CO2 and form bicarbonates [31,32]. The moderate basic sites were derived from the Ce-OL or OV-OL ion pairs, which combined with CO2 and promoted the formation of bridged and bidentate carbonates [8,19,32], while the strong basic sites were mainly assigned to the low-coordination oxygen anions [8,31,32], which resulted in the formation of stable monodentate carbonates [33]. Overall, the surface acidic sites with different natures all decreased with an increase in the reduction temperature and increased with an increase in the reoxidation temperature ( Figure 6 and Table 4). A similar trend was observed for the surface basic sites under various treatment conditions ( Figure 7 and Table 4). A possible reason for the change in surface acidity and basicity is discussed in Section 3.2.  Overall, the surface acidic sites with different natures all decreased with an increase in the reduction temperature and increased with an increase in the reoxidation temperature ( Figure 6 and Table 4). A similar trend was observed for the surface basic sites under various treatment conditions (Figure 7 and Table 4). A possible reason for the change in surface acidity and basicity is discussed in Section 3.2.

Measurement of Catalytic Activity
The DMC yields of CeO 2 treated under different conditions are displayed in Figure 8. As seen in Figure 8A, the DMC yield of CeO 2 gradually decreased from 9.02 to 1.75 mmol/g as the reduction temperature increased to 800 • C. However, as seen in Figure 8B,C, the DMC yield of CeO 2 -H550 and CeO 2 -H750 gradually increased from 6.94 and 2.85 to 7.98 and 4.21 mmol/g, respectively, as the reoxidation temperature increased to 550 • C. These results indicated that reduction treatment is detrimental to the improvement of catalytic activity but reoxidation treatment is beneficial.

Measurement of Catalytic Activity
The DMC yields of CeO2 treated under different conditions are displayed in Figure  8. As seen in Figure 8A, the DMC yield of CeO2 gradually decreased from 9.02 to 1.75 mmol/g as the reduction temperature increased to 800 °C. However, as seen in Figure  8B,C, the DMC yield of CeO2-H550 and CeO2-H750 gradually increased from 6.94 and 2.85 to 7.98 and 4.21 mmol/g, respectively, as the reoxidation temperature increased to 550 °C. These results indicated that reduction treatment is detrimental to the improvement of catalytic activity but reoxidation treatment is beneficial.

The Relationship between Catalytic Activity and Surface Acidity/Basicity
Surface basicity and acidity play a significant role in determining the catalytic activity of CeO 2 -based catalysts. Wang et al. [11] found that only bidentate carbonates can react with surface -OCH 3 originating from the dissociation of methanol and form MMC, which is considered an active intermediate species for the formation of DMC, while monodentate carbonates, bicarbonates, and bridged carbonates cannot accomplish this process. Inumaru et al. [34] also found that the catalytic activity of ZrO 2 nanocrystals is closely related to the surface bidentate carbonate species. As discussed in Section 2.4, bidentate carbonates were derived from the interaction of CO 2 with moderate basic sites. In addition, surface acidic sites significantly affected the adsorption and activation of methanol. Studies have reported that methanol is adsorbed and dissociated on acid sites and forms -OCH 3 groups, which participate in the formation of MMC [7,19,35]. Xiao et al. [8] revealed that catalytic activity is positively correlated with the moderate surface acidic/basic sites of Fe-Zr composite catalysts. Taking these findings into account, it is reasonable to build a relationship between catalytic activity and moderate surface acidic/basic sites. As shown in Figure 9A,B, catalytic activity was positively correlated with the number of moderate surface acid and basic sites (determined from NH 3 -TPD and CO 2 TPD profiles) of CeO 2 catalysts.

The Relationship between Catalytic Activity and Surface Acidity/Basicity
Surface basicity and acidity play a significant role in determining the catalytic activity of CeO2-based catalysts. Wang et al. [11] found that only bidentate carbonates can react with surface -OCH3 originating from the dissociation of methanol and form MMC, which is considered an active intermediate species for the formation of DMC, while monodentate carbonates, bicarbonates, and bridged carbonates cannot accomplish this process. Inumaru et al. [34] also found that the catalytic activity of ZrO2 nanocrystals is closely related to the surface bidentate carbonate species. As discussed in Section 2.4, bidentate carbonates were derived from the interaction of CO2 with moderate basic sites. In addition, surface acidic sites significantly affected the adsorption and activation of methanol. Studies have reported that methanol is adsorbed and dissociated on acid sites and forms -OCH3 groups, which participate in the formation of MMC [7,19,35]. Xiao et al. [8] revealed that catalytic activity is positively correlated with the moderate surface acidic/basic sites of Fe-Zr composite catalysts. Taking these findings into account, it is reasonable to build a relationship between catalytic activity and moderate surface acidic/basic sites. As shown in Figure 9A,B, catalytic activity was positively correlated with the number of moderate surface acid and basic sites (determined from NH3-TPD and CO2TPD profiles) of CeO2 catalysts.

Deep Insight into the Role of Ce 4+ and OV
Marin et al. [9] reported that conversion of CO2 and methanol to DMC over a CeO2 catalyst follows a Langmuir-Hinshelwood mechanism, with the CO2 adsorption step being rate controlling. However, there are two different theories about the adsorption and activation of CO2 for CeO2-based catalysts. One is that surface coordination unsaturated Ce 4+ ions and OL serve as Lewis acid-base pairs to promote the adsorption and activation of CO2 [7,36,37]. The other is that surface OV and OL serve as Lewis acid-base pairs to promote the adsorption and activation of CO2. Especially for metal-doped CeO2 catalysts, it was found that one of the O atoms of CO2 can be inserted into OV and thus promotes the adsorption and activation of CO2, and therefore, the adsorption of CO2 improves with an increase in surface OV [12,15,19]. Obviously, the focus was that of Ce 4+ or OV, which Lewis acid sites determine the adsorption and activation of CO2 for CeO2 catalysts?
XPS results showed that reduction treatment results in a decrease in surface Ce 4+ and an increase in OV. In contrast, reoxidation treatment resulted in a decrease in surface OV and an increase in Ce 4+ . Combining XPS and CO2-TPD results, it was apparent that the

Deep Insight into the Role of Ce 4+ and O V
Marin et al. [9] reported that conversion of CO 2 and methanol to DMC over a CeO 2 catalyst follows a Langmuir-Hinshelwood mechanism, with the CO 2 adsorption step being rate controlling. However, there are two different theories about the adsorption and activation of CO 2 for CeO 2 -based catalysts. One is that surface coordination unsaturated Ce 4+ ions and O L serve as Lewis acid-base pairs to promote the adsorption and activation of CO 2 [7,36,37]. The other is that surface O V and O L serve as Lewis acid-base pairs to promote the adsorption and activation of CO 2 . Especially for metal-doped CeO 2 catalysts, it was found that one of the O atoms of CO 2 can be inserted into O V and thus promotes the adsorption and activation of CO 2 , and therefore, the adsorption of CO 2 improves with an increase in surface O V [12,15,19]. Obviously, the focus was that of Ce 4+ or O V , which Lewis acid sites determine the adsorption and activation of CO 2 for CeO 2 catalysts?
XPS results showed that reduction treatment results in a decrease in surface Ce 4+ and an increase in O V . In contrast, reoxidation treatment resulted in a decrease in surface O V and an increase in Ce 4+ . Combining XPS and CO 2 -TPD results, it was apparent that the adsorption amount of CO 2 and the concentration of Ce 4+ showed a similar trend under various treatment conditions. Moreover, combining XPS and NH 3 TPD results, we found that the adsorption amount of NH 3 and the concentration of Ce 4+ shows a similar trend under various treatment conditions. These results indicated that surface Ce 4+ serves as active Lewis acid sites to adsorb and activate CO 2 and methanol molecules.
Nevertheless, it should be noted that the particle size of CeO 2 can also affect the adsorption of CO 2 and NH 3 . For example, as the reduction temperature increased beyond 550 • C, the aggregation of CeO 2 particles occurred ( Table 2 and Figure 3), which can also result in decreased adsorption of CO 2 and NH 3 .
The profiles of the DMC yield for catalysts treated under different conditions with a change in the concentration of Ce 4+ and O V are presented in Figures 10 and 11 adsorption amount of CO2 and the concentration of Ce 4+ showed a similar trend under various treatment conditions. Moreover, combining XPS and NH3TPD results, we found that the adsorption amount of NH3 and the concentration of Ce 4+ shows a similar trend under various treatment conditions. These results indicated that surface Ce 4+ serves as active Lewis acid sites to adsorb and activate CO2 and methanol molecules. Nevertheless, it should be noted that the particle size of CeO2 can also affect the adsorption of CO2 and NH3. For example, as the reduction temperature increased beyond 550 °C, the aggregation of CeO2 particles occurred (Table 2 and Figure 3), which can also result in decreased adsorption of CO2 and NH3.
The profiles of the DMC yield for catalysts treated under different conditions with a change in the concentration of Ce 4+ and OV are presented in Figure 10 and Figure 11, respectively.   adsorption amount of CO2 and the concentration of Ce 4+ showed a similar trend under various treatment conditions. Moreover, combining XPS and NH3TPD results, we found that the adsorption amount of NH3 and the concentration of Ce 4+ shows a similar trend under various treatment conditions. These results indicated that surface Ce 4+ serves as active Lewis acid sites to adsorb and activate CO2 and methanol molecules. Nevertheless, it should be noted that the particle size of CeO2 can also affect the adsorption of CO2 and NH3. For example, as the reduction temperature increased beyond 550 °C, the aggregation of CeO2 particles occurred (Table 2 and Figure 3), which can also result in decreased adsorption of CO2 and NH3.
The profiles of the DMC yield for catalysts treated under different conditions with a change in the concentration of Ce 4+ and OV are presented in Figure 10 and Figure 11, respectively.   It was obvious that the catalytic activity increased with the increasing concentration of Ce 4+ , while it decreased with the increasing concentration of O V . Nevertheless, as discussed before, treatment conditions greatly affected the particle size of CeO 2 , as presented in Table 2 and Figure 3. Generally, a particle with a smaller size is conducive to the exposure of more active sites and thus improves the adsorption of the reactant and catalytic activity [38]. To eliminate the particle size effect, CeO 2 , CeO 2 -H450, CeO 2 -H550, CeO 2 -H550-O350, CeO 2 -H550-O450, and CeO 2 -H550-O550 catalysts (with similar crystal size of about 7~8 nm and basically an unchanged particle size) were selected to correlate the number of moderate basic sites as well as catalytic activity with the concentration of Ce 4+ and O V . As shown in Figure 12, the number of moderate basic sites as well as catalytic activity were positively correlated with the concentration of surface Ce 4+ but negatively correlated with that of O V .
It was obvious that the catalytic activity increased with the increasing concentration of Ce 4+ , while it decreased with the increasing concentration of OV. Nevertheless, as discussed before, treatment conditions greatly affected the particle size of CeO2, as presented in Table 2 and Figure 3. Generally, a particle with a smaller size is conducive to the exposure of more active sites and thus improves the adsorption of the reactant and catalytic activity [38]. To eliminate the particle size effect, CeO2, CeO2-H450, CeO2-H550, CeO2-H550-O350, CeO2-H550-O450, and CeO2-H550-O550 catalysts (with similar crystal size of about 7~8 nm and basically an unchanged particle size) were selected to correlate the number of moderate basic sites as well as catalytic activity with the concentration of Ce 4+ and OV. As shown in Figure 12, the number of moderate basic sites as well as catalytic activity were positively correlated with the concentration of surface Ce 4+ but negatively correlated with that of OV. To further confirm the role of surface Ce 4+ in the formation of active bidentate carbonates, FTIR spectra of CO2 adsorption for CeO2, CeO2-H550, and CeO2-H550-O550 were developed, and the results are displayed in Figure 13. The bands at 1215, 1393, and 1610 cm −1 were associated with bicarbonates, while the bands at 856, 1033, 1293, and 1581 cm −1 were ascribed to bidentate carbonates [11,19]. In addition, the small bands at 1242 and 1652 cm −1 were attributed to bridged carbonates, while the small bands at 1356 and 1466 cm −1 were assigned to monodentate carbonates [11,19]. Obviously, the peak intensity To further confirm the role of surface Ce 4+ in the formation of active bidentate carbonates, FTIR spectra of CO 2 adsorption for CeO 2 , CeO 2 -H550, and CeO 2 -H550-O550 were developed, and the results are displayed in Figure 13. The bands at 1215, 1393, and 1610 cm −1 were associated with bicarbonates, while the bands at 856, 1033, 1293, and 1581 cm −1 were ascribed to bidentate carbonates [11,19]. In addition, the small bands at 1242 and 1652 cm −1 were attributed to bridged carbonates, while the small bands at 1356 and 1466 cm −1 were assigned to monodentate carbonates [11,19]. Obviously, the peak intensity and area of bidentate carbonates decreased after reduction treatment and increased after reoxidation treatment. These results strongly supported the fact that the surface Ce 4+ of CeO 2 serves as active Lewis acid sites and promotes the adsorption of CO 2 and the formation of bidentate carbonates and thus catalytic activity for the synthesis of DMC from CO 2 and methanol. and area of bidentate carbonates decreased after reduction treatment and increased after reoxidation treatment. These results strongly supported the fact that the surface Ce 4+ of CeO2 serves as active Lewis acid sites and promotes the adsorption of CO2 and the formation of bidentate carbonates and thus catalytic activity for the synthesis of DMC from CO2 and methanol. Why OV played an opposite role in the adsorption of CO2 for pure CeO2 and metaldoped CeO2 catalytic systems is thought provoking. This may be because OV has a different nature in pure CeO2 and metal-doped CeO2 catalysts. As discussed in Section of 2.3, the OV formed by H2 reduction are a kind of anionic defect, and a material with this kind of defect is an N-type semiconductor [28]. The possible reason is that the quasi-free mobile electrons around OV impede the adsorption and activation of CO2, and therefore, the catalytic activity decreases with an increase in OV. This was consistent with the results reported by Tan et al. that the catalytic activity of CeO2 prepared under different atmospheres follows the order CeO2-O2 > CeO2-Ar > CeO2-air > CeO2-H2; especially, the catalytic activity of CeO2-O2 was 2.5 times as high as that of CeO2-H2 [27]. However, for metaldoped CeO2 catalysts, the OV derived from the incorporation of metal ions (for example, Ca 2+ , Zn 2+ , Mn 2+ , Ti 4+ , and Zr 4+ ) into the lattice of CeO2 is a kind of cationic defect, and a material with this kind of defect is a P-type semiconductor [39,40]. This kind of cationic defect (OV) is beneficial for the adsorption and activation of CO2, and thus, the catalytic activity increases with an increase in OV [12,13,15,16,19]. The possible influence of two types of OV on the adsorption and activation of CO2 is shown in Figure 14. Why O V played an opposite role in the adsorption of CO 2 for pure CeO 2 and metaldoped CeO 2 catalytic systems is thought provoking. This may be because O V has a different nature in pure CeO 2 and metal-doped CeO 2 catalysts. As discussed in Section 2.3, the O V formed by H 2 reduction are a kind of anionic defect, and a material with this kind of defect is an N-type semiconductor [28]. The possible reason is that the quasi-free mobile electrons around O V impede the adsorption and activation of CO 2 , and therefore, the catalytic activity decreases with an increase in O V . This was consistent with the results reported by Tan et al. that the catalytic activity of CeO 2 prepared under different atmospheres follows the order CeO 2 -O 2 > CeO 2 -Ar > CeO 2 -air > CeO 2 -H 2 ; especially, the catalytic activity of CeO 2 -O 2 was 2.5 times as high as that of CeO 2 -H 2 [27]. However, for metal-doped CeO 2 catalysts, the O V derived from the incorporation of metal ions (for example, Ca 2+ , Zn 2+ , Mn 2+ , Ti 4+ , and Zr 4+ ) into the lattice of CeO 2 is a kind of cationic defect, and a material with this kind of defect is a P-type semiconductor [39,40]. This kind of cationic defect (O V ) is beneficial for the adsorption and activation of CO 2 , and thus, the catalytic activity increases with an increase in O V [12,13,15,16,19]. The possible influence of two types of O V on the adsorption and activation of CO 2 is shown in Figure 14. It is interesting that similar phenomena have also been observed during the conversion of syngas to light olefins over ZrCeZnOx catalysts [41,42]. It was found that the incorporation of Ce 3+ into Zn-Zr oxide improves the percentage of OV and thus catalytic activity [42]; however, the catalytic activity of ZrCeZnOx decreased after reduction with H2, although the percentage of OV improved significantly [41]. This indicated that the types of OV remarkably affects the adsorption of CO and therefore catalytic activity. It is interesting that similar phenomena have also been observed during the conversion of syngas to light olefins over ZrCeZnO x catalysts [41,42]. It was found that the incorporation of Ce 3+ into Zn-Zr oxide improves the percentage of O V and thus catalytic activity [42]; however, the catalytic activity of ZrCeZnO x decreased after reduction with H 2 , although the percentage of O V improved significantly [41]. This indicated that the types of O V remarkably affects the adsorption of CO and therefore catalytic activity.

Preparation of CeO 2
First, 3.47 g of Ce(NO 3 ) 3 ·6H 2 O was dissolved in 30 mL of deionized water. Next, 100 mL of NaOH solution (2 mol/L) was added to the solution of Ce(NO 3 ) 3 with a constantflux pump under stirring. After aging for 1 h at ambient temperatures, the slurry was filtered and washed with deionized water and ethanol several times until it became neutral. Finally, the products were dried for 12 h at 80 • C and then calcinated for 5 h at 550 • C in a muffle furnace. The resultant catalyst was labeled as CeO 2 and used as a starting material for further treatment.

Reduction Treatment of CeO 2
Typically, 0.25 g of CeO 2 was placed in a tube furnace, and then, the temperature was increased to 450, 550, 600, 650, 700, 750, and 800 • C under 10a % H 2 -N 2 atmosphere for 4 h at a heating rate of 5 • C/min. The obtained catalysts were named as CeO 2 -HT 1 , where T 1 is the reduction temperature.

Reoxidation Treatment of CeO 2 -H550 and CeO 2 -H750
Typically, 0.2 g of CeO 2 -H550 or CeO 2 -H750 was placed in a muffle furnace, and then, the temperature was increased to 350, 450 and 550 • C under an air atmosphere for 4 h at a heating rate of 2 • C/min. The obtained catalysts were named as CeO 2 -H550-OT 2 or CeO 2 -H750-OT 2 , where T 2 is the reoxidation temperature.

Characterization of Supports and Catalysts
N 2 physisorption was performed on a Beishide 3H-2000PS2 instrument at −196 • C using N 2 as an adsorbate. The sample was outgassed at 250 • C for 4 h in vacuum to remove the physisorbed impurities before measurement. The specific surface area and pore volume were calculated using the BET and BIH methods, respectively.
XRD characterization was carried out on a Rigaku SmartLab SE diffractometer (40 kV, 40 mA) with a diffraction angle (2θ) from 10 to 80 • and a scanning rate of 10 • /min. TEM images were taken with FEI Tecnai F30 at 200 kV. Before measurement, the sample was first dispersed in ethanol, then ultrasonicated for 30 min, and finally deposited on a carbon-coated copper film.
TPD characterization was performed on Micromeritics Autochem 2920. First, 100 mg of the sample was degassed for 30 min at 250 • C with N 2 (30 mL/min) to remove the physisorbed impurities. After the temperature was cooled down to 50 • C, 15% NH 3 in He (30 mL/min) or pure CO 2 (30 mL/min) was introduced and maintained for 30 min to allow saturated adsorption of NH 3 and CO 2 . Subsequently, the physisorbed substance was removed with He (30 mL/min) at 50 • C for 30 min. Thereafter, the temperature was gradually increased to 800 • C at a heating rate of 10 • C/min with He as a carrier gas (30 mL/min), and the desorbed NH 3 and CO 2 were detected with TCD.
XPS spectra were analyzed with an ESCALAB 250Xi (Thermo Scientific, Waltham, MA, USA) photoelectron spectrometer under AlKα radiation at 300 W. The binding energies were referred to the C1s line from adventitious carbon at 284.6 eV.
CO 2 -FTIR spectra were recorded using a Bruker Tensor II instrument equipped with an MCT detector by collecting 64 scans at a resolution of 4 cm −1 . About 20 mg of the sample was loaded into the cell and then pretreated under He flow (30 mL/min) at 400 • C for 1 h to remove the physisorbed impurities. After the temperature dropped down to 140 • C (reaction temperature), pure CO 2 (30 mL/min) was introduced and maintained for 30 min to allow saturated adsorption of CO 2 . Subsequently, the physisorbed substance was removed with He (30 mL/min) for 30 min, and then, FTIR spectra were recorded.

Catalytic Evaluation
The catalytic activity of catalysts for DMC synthesis from CO 2 and methanol was investigated in a 250 mL high-pressure slurry-bed reactor equipped with a mechanical stirrer. Typically, 0.2 g of a catalyst and 30 mL of methanol were placed into the reactor, which was subsequently sealed. Next, the reactor was filled with CO 2 several times to check the tightness and remove air. Thereafter, the reactor was pressurized with CO 2 to 3 MPa at ambient temperatures, and then, the reaction temperature was increased to 140 • C and maintained for 4 h. After the reaction finished, the reactor was depressurized and cooled down to room temperature. The liquid product was analyzed with n-propanol as an internal standard substance using a gas chromatograph (FID-GC, O Hua 9160). It should be noted that no by-products were formed except DMC, and the yield of DMC was defined as follows: Y DMC (mmol/g) = n DMC (mmol)/m catalyst (g) (1)

Conclusions
(1) The concentrations of surface Ce 4+ and oxygen vacancies of CeO 2 were regulated with reduction and reoxidation treatments under different conditions. The reduction treatment promoted the conversion of Ce 4+ to Ce 3+ and improved the surface concentration of oxygen vacancies, while the reoxidation treatment favored the conversion of Ce 3+ to Ce 4+ and decreased the concentration of oxygen vacancies. (2) Catalytic activity was positively correlated with the number of moderate surface acidic/basic sites of catalysts. Moreover, the number of moderate basic sites and catalytic activity were positively correlated with the surface concentration of Ce 4+ but negatively correlated with that of oxygen vacancies. (3) The surface Ce 4+ rather than oxygen vacancies served as active Lewis acid sites, and lattice oxygen served as active Lewis base sites to adsorb and activate CO 2 , promoting the formation of active bidentate carbonates species and DMC. (4) The cationic oxygen vacancy was beneficial but the anionic oxygen vacancy was detrimental to the formation of active bidentate carbonates species and DMC, which sheds light upon the active sites for CeO 2 and metal-doped CeO 2 catalysts and provides evidence for the design of efficient catalysts.
Supplementary Materials: The following supporting information can be downloaded at https:// www.mdpi.com/article/10.3390/molecules28093785/s1, Figure S1: Change in the color of catalysts. Data Availability Statement: All the relevant data used in this study have been provided in the form of figures and tables in the published article, and all data provided in the manuscript are available to whom they may concern.