Effect of the Preparation Method on the Physicochemical Properties and the CO Oxidation Performance of Nanostructured CeO2/TiO2 Oxides

Ceria-based mixed oxides have been widely studied in catalysis due to their unique surface and redox properties, with implications in numerous energyand environmental-related applications. In this regard, the rational design of ceria-based composites by means of advanced synthetic routes has gained particular attention. In the present work, ceria–titania composites were synthesized by four different methods (precipitation, hydrothermal in one and two steps, Stöber) and their effect on the physicochemical characteristics and the CO oxidation performance was investigated. A thorough characterization study, including N2 adsorption-desorption, X-ray diffraction (XRD), scanning electron microscopy with energy dispersive X-ray spectroscopy (SEM/EDS), transmission electron microscopy (TEM) and H2 temperature-programmed reduction (H2-TPR) was performed. Ceria–titania samples prepared by the Stöber method, exhibited the optimum CO oxidation performance, followed by samples prepared by the hydrothermal method in one step, whereas the precipitation method led to almost inactive oxides. CeO2/TiO2 samples synthesized by the Stöber method display a rod-like morphology of ceria nanoparticles with a uniform distribution of TiO2, leading to enhanced reducibility and oxygen storage capacity (OSC). A linear relationship was disclosed among the catalytic performance of the samples prepared by different methods and the abundance of reducible oxygen species.


Introduction
Ceria (CeO 2 ) has attracted significant attention in heterogeneous catalysis due to its enhanced properties such as high oxygen storage capacity (OSC) and thermal stability, [1][2][3]. The facile shift between two oxidation states (Ce 3+ /Ce 4+ ), followed by the generation of surface/structural defects, e.g., oxygen vacancies, is responsible for its improved redox behaviour [1,[4][5][6][7]. More importantly, the development of ceria particles in the nanoscale is of great significance as a decrease in the particle size can lead to distinct physicochemical properties as compared to their bulk counterparts, such as high specific surface area, tunable pore size, abundant defects and adjustable surface chemistry [8][9][10][11]. CeO 2 /TiO 2 mixed oxides with Ce/Ti atomic ratio of 4 were prepared by various methods, namely, precipitation, hydrothermal in one and two steps, and Stöber. This specific ratio was dictated from our previous studies on ceria-based mixed oxides [18][19][20].
Ceria-titania by precipitation (CeO 2 /TiO 2 -P) was prepared as follows: initially, 10.4 mmol Ce(NO 3 ) 3 ·6H 2 O and 2.6 mmol TiOSO 4 were dissolved in double deionized water (0.50 M) in order to achieve complete dissolution of the precursors. Then, NH 3 (25 vol.%) was added at room temperature (RT) to the solution under constant stirring until pH reached the value of 10. After 3 h of stirring the resulting formulation was centrifuged and washed with double deionized water and EtOH. The precipitate was treated at 90 • C for 12 h and 500 • C for 2 h under air flow (heating rate 5 • C min −1 ) [33].
Ceria-titania by one-step hydrothermal (CeO 2 /TiO 2 -H1) was prepared as follows: initially, 3.0 mol NaOH was added in 75 mL of water. An aqueous solution of 175 mL containing 20 mmol of Ce(NO 3 ) 3 ·6H 2 O and 5.0 mmol of TiOSO 4 (0.14 M) was then added under vigorous stirring in the above solution until a milky slurry was formed. The mixture was left for 1 h of additional stirring and the final slurry was placed in a Teflon vial and aged for 24 h at 90 • C. The materials were thoroughly washed with double deionized water until pH 7, followed by EtOH washing as to avoid nanoparticles agglomeration. Finally, the precipitate was dried for 12 h at 90 • C and calcined at 500 • C for 2 h under air (heating rate 5 • C min −1 ) [36].
Ceria-titania by two-step hydrothermal synthesis (CeO 2 /TiO 2 -H2) was prepared as follows: bare ceria nanorods (CeO 2 NRs) were initially prepared by the hydrothermal method, as mentioned above in the one-step hydrothermal method, with the variation of utilizing 25 mmol Ce(NO 3 ) 3 ·6H 2 O instead of 20 mmol and the absence of TiOSO 4 . Then, CeO 2 /TiO 2 nanorods were prepared by a wet-chemical method. Typically, 0.15 g of bare ceria nanorods was dispersed into 200 mL of double deionized water under vigorous stirring and 0.035 g TiOSO 4 was dissolved into 1.1 mL of double deionized water and added into the aforementioned suspension at room temperature. The mixture was left for additional stirring for 3 h and aged at room temperature for 2-3 h. The precipitate was dried for 12 h at 90 • C, followed by calcination at 500 • C for 2 h under air (heating rate 5 • C min −1 ) [40].
Ceria-titania by Stöber method (CeO 2 /TiO 2 -S) was prepared as follows: bare ceria nanorods were initially synthesized by the hydrothermal method as previously described. 0.25 g of bare ceria nanorods were dispersed in 334 mL of absolute ethanol and mixed with 1 mL concentrated ammonia solution (28 wt.%) under ultrasonication for 20 min. Then, 0.125 mL TBOT was added dropwise and the solution was left stirring at 45 • C for 24 h. The resulting precipitate was recovered by centrifugation and washed with double deionized water and ethanol. The precipitate was dried for 12 h at 90 • C, followed by calcination at 500 • C for 2 h under air (heating rate 5 • C min −1 ) [29].

Materials Characterization
The textural features of the as-prepared samples were evaluated by N 2 adsorption-desorption using a Nova 2200e Quantachrome flow apparatus (Quantachrome Instruments, Florida, USA). Surface area (m 2 g −1 ) was calculated using the Brunauer-Emmett-Teller (BET) method applied in the 0.05-0.30 range of relative pressures [60,61]. At the highest relative pressure, the specific pore volume (cm 3 g −1 ) was obtained. The Barrett-Joyner-Halenda (BJH) method was applied to determine the average pore size diameter (dp, nm). Prior to measurement all samples were outgassed at 250 • C for 5 h under vacuum.
Crystallinity of the samples was investigated by powder X-ray diffraction (XRD) in a Rigaku diffractometer (model RINT 2000) (Rigaku, Tokyo, Japan) using Cu Kα radiation (λ Ka = 0.154 nm, 40 kV/80 mA) in the range of 10 • to 80 • with a step of 0.05 • min −1 . The primary crystallite size of the samples was calculated by applying the Scherrer Equation (1): where λ: wavelength of the X-ray in nm, K: Scherrer constant, B: line broadening, θ: Bragg angle. By using the Scherrer Equation (1), on the most intense diffraction peak, i.e., for CeO 2 (2θ: 28.5 • ), the primary particle size of this crystal phase was determined. Morphological characterization of the samples was determined by scanning electron microscopy (SEM, JSM-6390LV, JEOL Ltd., Tokyo, Japan) working at 20 keV, coupled with an energy dispersive X-ray spectrometry (EDS) system. The samples in the form of powders were sputtered with Au to create a coating with a thickness of 10 nm, approximately. Furthermore, the samples were imaged by transmission electron microscopy (TEM) using a JEM-2100 instrument (JEOL, Tokyo, Japan) equipped with LaB 6 filament, working at 200 kV. TEM specimens were prepared by deposition of dispersed powder samples by ultrasonication.
The redox properties of as-prepared samples were assessed by temperature-programmed reduction with hydrogen as reducing agent (H 2 -TPR). In a typical experiment, 100 mg of the sample was placed in a quartz fixed-bed reactor heated up to 800 • C with a rate of 20 K min −1 . Before TPR experiments, samples were treated up to 500 • C for 30 min under 20 vol.% O 2 in He (20 cm 3 min −1 flow), then cooled to 200 • C under the same gas mixture and finally to room temperature under pure He flow. A known quantity of CuO standard sample was used to quantify the H 2 uptake from the samples [62,63]. H 2 uptake (mmol H 2 g −1 ), corresponding to the H 2 consumed by reducible oxygen species, is calculated by the integration of TPR peaks area. Oxygen Storage Capacity (OSC), in µmol O 2 g −1 , is determined as a half of this quantity [64].

Catalytic Activity Studies
Catalytic measurements were carried out in a quartz fixed-bed tubular microreactor (i.d.: 12.95 mm) under ambient pressure, loaded with 0.10 g of catalyst.
The reaction gas mixture consisted of 0.2 vol.% of CO and 1 vol.% O 2 both balanced with He. A total feed stream of 80 cm 3 min −1 was used, corresponding to a gas hourly space velocity (GHSV) of 40,000 h −1 .
Before catalytic tests, all samples were pretreated under 20 vol.% O 2 in He with a 20 cm 3 min −1 flow and heated to 480 • C for 30 min with a rate of 10 K min −1 . Consequently, samples were cooled to 25 • C at the same rate and purged with He in order to remove any physisorbed species. Catalytic performance evaluation was carried out up to 500 • C for each sample. CO and CO 2 were monitored by a Gas Chromatograph with two channels with independent thermal conductivity detectors, injectors and capillary columns, i.e., Molecular Sieve 5X, PoraPlot Q.
The CO conversion (X CO ) was calculated from the difference in CO concentration among inlet and outlet gas streams: where [CO] in and [CO] out are the CO concentration in the inlet and outlet gas streams, respectively. Specific reaction rates of CO conversion (in mol g −1 s −1 or mol m −2 s −1 ) were estimated under differential reaction conditions (X CO < 15%, T = 200 • C, W/F = 0.075 g s cm −3 ).
F: total flow rate, V m : gas molar volume; all at standard ambient temperature and pressure conditions (298 K, 1 bar), m cat : catalyst's mass, S BET : specific surface area.

Textural/Structural Characterization
In Table 1, the main textural and structural properties of bare CeO 2 and CeO 2 /TiO 2 samples are presented. Ceria-titania prepared by the one-step hydrothermal method (CeO 2 /TiO 2 -H1) demonstrates the highest value in BET surface area (100.9 m 2 g −1 ), followed by bare CeO 2 NRs (73.9 m 2 g −1 ), CeO 2 /TiO 2 -S (72.0 m 2 g −1 ), CeO 2 /TiO 2 -H2 (63.2 m 2 g −1 ) and CeO 2 /TiO 2 -P (41.5 m 2 g −1 ). An increase in the surface area is observed for the CeO 2 /TiO 2 -H1 sample, as compared to bare CeO 2 NRs. In contrast, the BET surface area is decreased in the CeO 2 /TiO 2 -S, CeO 2 /TiO 2 -H2 and CeO 2 /TiO 2 -P samples. Apparently, the preparation procedure notably affects the textural/structural characteristics and in turn the surface area, as further discussed below. The BJH pore size distribution (PSD) of as-prepared samples is displayed in Figure 1a and the respective adsorption-desorption isotherms are shown in Figure 1b. In all samples, the pore size maxima are over 3 nm, signifying the existence of mesopores that can be further verified by the presence of type IV adsorption-desorption isotherms ( Figure 1b). The CeO 2 /TiO 2 -S sample exhibits the highest pore size (32.0 nm), followed by CeO 2 NRs (24.2 nm), CeO 2 /TiO 2 -H2 (24.1 nm), CeO 2 /TiO 2 -H1 (15.0 nm) and CeO 2 /TiO 2 -P (9.3 nm) and it also shows a broader PSD than the rest samples. increase in the surface area is observed for the CeO2/TiO2-H1 sample, as compared to bare CeO2 NRs. In contrast, the BET surface area is decreased in the CeO2/TiO2-S, CeO2/TiO2-H2 and CeO2/TiO2-P samples. Apparently, the preparation procedure notably affects the textural/structural characteristics and in turn the surface area, as further discussed below. The BJH pore size distribution (PSD) of as-prepared samples is displayed in Figure 1a and the respective adsorption-desorption isotherms are shown in Figure 1b. In all samples, the pore size maxima are over 3 nm, signifying the existence of mesopores that can be further verified by the presence of type IV adsorption-desorption isotherms ( Figure 1b). The CeO2/TiO2-S sample exhibits the highest pore size (32.0 nm), followed by CeO2 NRs (24.2 nm), CeO2/TiO2-H2 (24.1 nm), CeO2/TiO2-H1 (15.0 nm) and CeO2/TiO2-P (9.3 nm) and it also shows a broader PSD than the rest samples.    [65][66][67]. The titanium oxide phase is hardly detected only in the CeO2/TiO2-S sample, with a small peak at 2θ value ~25°, which corresponds to the anatase (TiO2) phase [68][69][70][71]. The difficulty in detecting the TiO2 phase in the XRD patterns is probably related to the low TiO2 loading in combination with its high dispersion. By applying the Scherrer Equation (1), the primary crystallite size of ceria was calculated (Table 1) within the approximation degree induced by the method. Bare ceria shows a crystallite size of 13 nm, with the CeO2/TiO2-H1, CeO2/TiO2-S, CeO2/TiO2-P and CeO2/TiO2-H2 samples exhibiting crystallite sizes of 9, 12, 12 and 14 nm, respectively. Apart from the sample prepared by the two-step hydrothermal method (CeO2/TiO2-H2), all the other ceria-titania samples exhibit a small reduction in the crystallite size of ceria. The small differences in ceria crystallite size and morphology upon the introduction of different synthetic protocols could be ascribed to the impact of various parameters (extent of calcination, interaction between heteroatoms, etc.) on the growth rate of ceria nanoparticles [72][73][74].   [65][66][67]. The titanium oxide phase is hardly detected only in the CeO 2 /TiO 2 -S sample, with a small peak at 2θ value~25 • , which corresponds to the anatase (TiO 2 ) phase [68][69][70][71]. The difficulty in detecting the TiO 2 phase in the XRD patterns is probably related to the low TiO 2 loading in combination with its high dispersion. By applying the Scherrer Equation (1), the primary crystallite size of ceria was calculated (Table 1) within the approximation degree induced by the method. Bare ceria shows a crystallite size of 13 nm, with the CeO 2 /TiO 2 -H1, CeO 2 /TiO 2 -S, CeO 2 /TiO 2 -P and CeO 2 /TiO 2 -H2 samples exhibiting crystallite sizes of 9, 12, 12 and 14 nm, respectively. Apart from the sample prepared by the two-step hydrothermal method (CeO 2 /TiO 2 -H2), all the other ceria-titania samples exhibit a small reduction in the crystallite size of ceria. The small differences in ceria crystallite size and morphology upon the introduction of different synthetic protocols could be ascribed to the impact of various parameters (extent of calcination, interaction between heteroatoms, etc.) on the growth rate of ceria nanoparticles [72][73][74].

Morphological Characterization (TEM, SEM-EDS)
The morphological characteristics of the ceria-titania nanostructures were investigated by transmission electron microscopy analysis. In Figure 3, the TEM images of bare ceria nanorods and of CeO 2 /TiO 2 mixed oxides are depicted. A bare ceria sample ( Figure 3a) displays a rod-like morphology with 50-200 nm in length. CeO 2 /TiO 2 -P ( Figure 3b) exhibits irregular shapes, while CeO 2 /TiO 2 mixed oxides synthesized by the hydrothermal method (one or two steps) also show the rod-like morphology (Figure 3c-f). In particular, the CeO 2 /TiO 2 -H1 sample exhibits smaller-sized nanorods (25-100 nm), as compared to bare CeO 2 NRs, while no separated TiO 2 particles are detected. Although the CeO 2 /TiO 2 -H2 sample also exhibits the rod-like morphology, distinct TiO 2 nanoparticles in contact with nanorods are clearly observed (Figure 3f). In Figure 3g, the CeO 2 /TiO 2 -S sample is depicted, showing a rod-like morphology (50-200 nm) without any separated TiO 2 particles being observed. The HRTEM (high-resolution transmission electron microscopy) results of CeO 2 /TiO 2 -S sample are displayed in Figure 3h. The lattice spacing of 0.36 nm is ascribed to the TiO 2 (101) crystal plane, and the lattice spacings of 0.31 and 0.19 nm are ascribed to the CeO 2 (111) and (220) crystal planes, respectively [5,8,17,29,32,33,75,76]. Therefore, it can be argued that, in the case of CeO 2 /TiO 2 -S sample, a uniform distribution of TiO 2 around ceria nanorods can be obtained.

Morphological Characterization (TEM, SEM-EDS)
The morphological characteristics of the ceria-titania nanostructures were investigated by transmission electron microscopy analysis. In Figure 3, the TEM images of bare ceria nanorods and of CeO2/TiO2 mixed oxides are depicted. A bare ceria sample (Figure 3a) displays a rod-like morphology with 50-200 nm in length. CeO2/TiO2-P (Figure 3b) exhibits irregular shapes, while CeO2/TiO2 mixed oxides synthesized by the hydrothermal method (one or two steps) also show the rod-like morphology (Figure 3c-f). In particular, the CeO2/TiO2-H1 sample exhibits smaller-sized nanorods (25-100 nm), as compared to bare CeO2 NRs, while no separated TiO2 particles are detected. Although the CeO2/TiO2-H2 sample also exhibits the rod-like morphology, distinct TiO2 nanoparticles in contact with nanorods are clearly observed (Figure 3f). In Figure 3g, the CeO2/TiO2-S sample is depicted, showing a rod-like morphology (50-200 nm) without any separated TiO2 particles being observed. The HRTEM (high-resolution transmission electron microscopy) results of CeO2/TiO2-S sample are displayed in Figure 3h. The lattice spacing of 0.36 nm is ascribed to the TiO2 (101) crystal plane, and the lattice spacings of 0.31 and 0.19 nm are ascribed to the CeO2 (111) and (220) crystal planes, respectively [5,8,17,29,32,33,75,76]. Therefore, it can be argued that, in the case of CeO2/TiO2-S sample, a uniform distribution of TiO2 around ceria nanorods can be obtained. In addition, the elemental mapping of CeO 2 /TiO 2 samples was carried out by SEM/EDS analysis ( Figure 4). The SEM images of ceria-titania mixed oxides are illustrated below (Figure 4a,e,i,m), along with the corresponding elemental mapping images (Figure 4b-d,f-h,j-l,n-p). A uniform distribution of Ce, Ti, O elements in the ceria-titania mixed oxides can be observed from SEM/EDS analysis.

Redox Properties (H 2 -TPR)
For reactions proceeding via a surface reduction step such as CO oxidation, H 2 -TPR results can identify and quantify the active oxygen sites of the catalysts. Figure 5 depicts the reduction profiles of as-prepared samples in the temperature range of 100-800 • C. The samples CeO 2 /TiO 2 -S, CeO 2 /TiO 2 -H1 and CeO 2 -NRs show a peak in the low temperature region of 500-700 • C, which can be attributed to the relatively weakly bound surface oxygen species, in contrast to bulk oxygen, which is reduced at temperatures higher than ca. 750 • C. Interestingly, the CeO 2 /TiO 2 -P and CeO 2 /TiO 2 -H2 samples do not exhibit any reduction peaks in the temperature region investigated, revealing their inferior reducibility. The latter is in complete agreement with their negligible catalytic performance, as discussed later. Moreover, it is worth noticing that TiO 2 addition to CeO 2 leads to a shift of TPR peaks to higher temperatures (Table 1) as compared to bare CeO 2 NRs, implying the inhibition induced by TiO 2 incorporation on the detachment of surface oxygen species during the reduction process. On the other hand, significant differences were obtained in relation to H 2 uptake, corresponding to main TPR peaks (Table 1). CeO 2 /TiO 2 -S exhibited the highest value (1.58 mmol H 2 g −1 ) followed by CeO 2 /TiO 2 -H1 (0.68 mmol H 2 g −1 ) and CeO 2 -NRs (0.58 mmol H 2 g −1 ), resulting in OSC values of 0.79, 0.34 and 0.29 mmol O 2 g −1 , respectively. This finding clearly reveals that the TiO 2 incorporation into CeO 2 via the Stöber method, leads to an almost 3-fold increase in the population of reducible/active oxygen species, which are expected to determine the oxygen exchange kinetics and in turn, the CO oxidation process via a Mars-van Krevelen mechanism. In this point, it should be mentioned that the differences in the reducibility could be mainly attributed to the different extent of ceria-titania interactions, revealing the significance of the preparation method on the intrinsic characteristics of mixed oxides. In view of this fact, a close relationship was disclosed between the catalytic activity and OSC (see below), corroborating the above arguments. In addition, the elemental mapping of CeO2/TiO2 samples was carried out by SEM/EDS analysis (Figure 4). The SEM images of ceria-titania mixed oxides are illustrated below (Figure 4a,e,i,m), along with the corresponding elemental mapping images (Figure 4b-d,f-h,j-l,n-p). A uniform distribution of Ce, Ti, O elements in the ceria-titania mixed oxides can be observed from SEM/EDS analysis.

Redox Properties (H2-TPR)
For reactions proceeding via a surface reduction step such as CO oxidation, H2-TPR results can identify and quantify the active oxygen sites of the catalysts. Figure 5 depicts the reduction profiles of as-prepared samples in the temperature range of 100-800 °C. The samples CeO2/TiO2-S, CeO2/TiO2-H1 and CeO2-NRs show a peak in the low temperature region of 500-700 °C, which can be attributed to the relatively weakly bound surface oxygen species, in contrast to bulk oxygen, which is reduced at temperatures higher than ca. 750 °C. Interestingly, the CeO2/TiO2-P and CeO2/TiO2-H2 samples do not exhibit any reduction peaks in the temperature region investigated, revealing their inferior reducibility. The latter is in complete agreement with their negligible catalytic performance, as discussed later. Moreover, it is worth noticing that TiO2 addition to CeO2 leads to a shift of TPR peaks to higher temperatures (Table 1) as compared to bare CeO2 NRs, implying the inhibition induced by TiO2 incorporation on the detachment of surface oxygen species during the reduction process. On the other hand, significant differences were obtained in relation to H2 uptake, corresponding to main TPR peaks (Table 1). CeO2/TiO2-S exhibited the highest value (1.58 mmol H2 g −1 ) followed by CeO2/TiO2-H1 (0.68 mmol H2 g −1 ) and CeO2-NRs (0.58 mmol H2 g −1 ), resulting in OSC values of 0.79, 0.34 and 0.29 mmol O2 g −1 , respectively. This finding clearly reveals that the TiO2 incorporation into CeO2 via the Stöber method, leads to an almost 3-fold increase in the population of reducible/active oxygen species, which are expected to determine the oxygen exchange kinetics and in turn, the CO oxidation process via a Mars-van Krevelen mechanism. In this point, it should be mentioned that the differences in the reducibility could be mainly attributed to the different extent of ceria-titania interactions, revealing the significance of the preparation method on the intrinsic characteristics of mixed oxides. In view of this fact, a close relationship was disclosed between the catalytic activity and OSC (see below), corroborating the above arguments.

Catalytic Evaluation Studies
To gain insight into the impact of the different preparation methods on the catalytic performance of the CeO2/TiO2 samples, their CO oxidation performance was investigated. Figure 6 depicts the conversion of CO with temperature for CeO2 and CeO2/TiO2 samples. For comparison purposes, a

Catalytic Evaluation Studies
To gain insight into the impact of the different preparation methods on the catalytic performance of the CeO 2 /TiO 2 samples, their CO oxidation performance was investigated. Figure 6 depicts the conversion of CO with temperature for CeO 2 and CeO 2 /TiO 2 samples. For comparison purposes, a TiO 2 commercial sample (Evonik industries) was also employed as reference material. A strong dependence between the preparation method and the nanostructure is observed, in accordance to our previous results [19]. As it becomes obvious from Figure 6, the sample prepared by the Stöber method (CeO 2 /TiO 2 -S) exhibits the best CO conversion performance among all samples tested. The bare ceria nanorods (CeO 2 -NRs) and ceria-titania prepared by the one-step hydrothermal method (CeO 2 /TiO 2 -H1) exhibit similar catalytic behaviour but, in both cases, their profiles are shifted by approximately 30 degrees to higher temperature than CeO 2 /TiO 2 -S. Notably, ceria-titania mixed oxides prepared by the precipitation and two-step hydrothermal methods demonstrate negligible CO conversion performance, namely~8% and 22% at 500 • C, respectively, implying the key role of synthesis procedure. Similarly, bare TiO 2 is practically inactive in the temperature range investigated. In order to completely comprehend the impact of preparation procedure and catalyst composition (bare or mixed oxides) on the intrinsic reactivity, the different textural properties among the as-prepared samples ( Table 2) should be taken into account. Hence, the specific activity, in terms of catalyst surface area (μmol m −2 s −1 ) and mass (μmol g −1 s −1 ) was calculated (Differential reaction conditions: XCO < 15%, T = 200 °C, GHSV = 40,000 h −1 ). Thus the relationship between CO oxidation activity and textural properties can be elucidated. Table 2 summarizes these results. Obviously, CeO2/TiO2-S demonstrates by far the optimum catalytic performance, in terms of conversion as well as of specific activity, further revealing its superiority. It should be also noted that a stable conversion performance (ca. 99% CO conversion) was attained at 400 °C in short-term (12 h) stability experiments.  In order to completely comprehend the impact of preparation procedure and catalyst composition (bare or mixed oxides) on the intrinsic reactivity, the different textural properties among the as-prepared samples ( Table 2) should be taken into account. Hence, the specific activity, in terms of catalyst surface area (µmol m −2 s −1 ) and mass (µmol g −1 s −1 ) was calculated (Differential reaction conditions: X CO < 15%, T = 200 • C, GHSV = 40,000 h −1 ). Thus the relationship between CO oxidation activity and textural properties can be elucidated. Table 2 summarizes these results. Obviously, CeO 2 /TiO 2 -S demonstrates by far the optimum catalytic performance, in terms of conversion as well as of specific activity, further revealing its superiority. It should be also noted that a stable conversion performance (ca. 99% CO conversion) was attained at 400 • C in short-term (12 h) stability experiments. The findings of the present work can be interpreted on the basis of a Mars-van Krevelen, redox-type mechanism, which mainly involves CO chemisorption towards the formation of Ce δ+ -CO species, O 2 activation on the oxygen vacancies of ceria and active oxygen species formation [77,78]. The reaction is taking place among the chemisorbed CO and adjacent active oxygen, followed by the active sites regeneration and the replenishment of oxygen vacancies via gas phase oxygen participated in sequential catalytic cycles [77,78].
The pivotal role played by the redox properties during the CO oxidation process over CeO 2 -based oxides is clearly described by the above mechanism. In particular, the high oxygen storage capacity of CeO 2 /TiO 2 -S sample (Table 1), linked to improved oxygen exchange kinetics and reducibility, can be considered responsible for its superior oxidation performance, as compared to the rest of the samples. In view of this fact, we recently showed, by means of both in situ and ex situ techniques, that the improved redox properties of CeO 2 composites of rod-like morphology, which are related to their abundance in defects and oxygen vacancies, is the decisive factor for their improved catalytic performance [18]. The above arguments concerning structure-activity relationships are clearly supported by the linear relationship among the specific reaction rate (µmol g −1 s −1 ) and the redox properties (OSC, mmol g −1 ), Figure 7. The pivotal role played by the redox properties during the CO oxidation process over CeO2based oxides is clearly described by the above mechanism. In particular, the high oxygen storage capacity of CeO2/TiO2-S sample (Table 1), linked to improved oxygen exchange kinetics and reducibility, can be considered responsible for its superior oxidation performance, as compared to the rest of the samples. In view of this fact, we recently showed, by means of both in situ and ex situ techniques, that the improved redox properties of CeO2 composites of rod-like morphology, which are related to their abundance in defects and oxygen vacancies, is the decisive factor for their improved catalytic performance [18]. The above arguments concerning structure-activity relationships are clearly supported by the linear relationship among the specific reaction rate (μmol g −1 s −1 ) and the redox properties (OSC, mmol g −1 ), Figure 7. In light of the above issues, the superiority of CeO2/TiO2-S catalyst can be ascribed to its higher population of weakly bound oxygen species, leading to enhanced oxygen storage capacity and reducibility, as identified by H2-TPR. This can be attributed to the beneficial effect of Stöber method towards obtaining highly homogenized composites of rod-like morphology, with no obvious separation between ceria and titania phases. In view of this fact, it can be argued that Stöber method could result to a synergistic ceria-titania interaction, which facilitates the formation of oxygen vacancies and the redox interplay between the interfacial sites, leading in turn to a higher population In light of the above issues, the superiority of CeO 2 /TiO 2 -S catalyst can be ascribed to its higher population of weakly bound oxygen species, leading to enhanced oxygen storage capacity and reducibility, as identified by H 2 -TPR. This can be attributed to the beneficial effect of Stöber method towards obtaining highly homogenized composites of rod-like morphology, with no obvious separation between ceria and titania phases. In view of this fact, it can be argued that Stöber method could result to a synergistic ceria-titania interaction, which facilitates the formation of oxygen vacancies and the redox interplay between the interfacial sites, leading in turn to a higher population of active oxygen species. In good agreement to this, it has been revealed by means of DFT calculations that the formation of oxygen vacancies and Ce 3+ sites is energetically favoured at the interface of ceria-titania [79].
Although bare titania is catalytically inert in CO oxidation, in contrast to CeO 2 -NRs ( Figure 6), their combination could lead to a synergistic interaction offering highly active CeO 2 /TiO 2 mixed oxides. The latter, however, strongly depends on the synthesis procedure followed; the hydrothermal method in one or two steps does not alter (CeO 2 /TiO 2 -H1) or even worsen (CeO 2 /TiO 2 -H2) the catalytic activity as compared to bare ceria, in opposition to Stöber method. In particular, the introduction of titania into ceria nanorods by the hydrothermal method (CeO 2 /TiO 2 -H1) is clearly enhancing the textural properties (surface area), without however having an analogous effect on the catalytic performance. This can be explained by taking into account the distinctive shift of the TPR peak to higher temperature ( Figure 5, Table 1), indicating a harder partial Ce 4+ → Ce 3+ reduction and detachment of O atoms, i.e., lower reducibility and an increase in oxygen vacancy formation energy. Moreover, no significant modifications were induced by the hydrothermal method on the concentration of reducible oxygen species, as verified by the similar OSC values between CeO 2 -NRs and CeO 2 /TiO 2 -H1 samples (Table 1). On the other hand, although the Stöber method slightly inhibits surface oxygen reduction (shift of TPR peak to higher temperature), it drastically increases the population of active oxygen species, offering a ca. 3-fold increase in OSC compared to that provided by the hydrothermal method. Summarizing, the catalytic performance of CeO 2 /TiO 2 -S samples is superior to that of bare CeO 2 , demonstrating the beneficial effect of a second oxide phase in conjunction to the synthesis procedure applied. These materials could be further employed as supporting carriers of various transition metals, boosting the transition to a noble metal free catalysis [3].

Conclusions
Ceria-titania mixed oxides were prepared by the hydrothermal, Stöber and precipitation methods. The CO oxidation was used as probe reaction to reveal possible structure-property relationships. The following order, in terms of CO conversion, was obtained: CeO 2 /TiO 2 -S > CeO 2 /TiO 2 -H1 > CeO 2 -NRs > CeO 2 /TiO 2 -H2 > CeO 2 /TiO 2 -P > TiO 2 . A characterization study by various complementary techniques revealed the significant effect of the synthesis procedure on the textural, morphological and redox features. Despite the inferior textural properties of the mixed oxides synthesized by the Stöber method, they exhibit the best catalytic performance, which can be ascribed to their improved reducibility, associated with their relative abundance in loosely bound oxygen species and high oxygen storage capacity. The Stöber method leads to rod-shaped ceria nanoparticles uniformly decorated by titania, which can be considered responsible for synergistic ceria-titania interactions towards an increased population of highly reducible active sites. The findings of the present study reveal that the rational design of mixed metal oxides without noble metals in their composition could provide the base for a wide range of materials, which could be used as catalysts or supporting carriers in various environmental and energy applications.
Author Contributions: S.S., M.L., D.F., V.B. and P.K.P. contributed to materials synthesis and characterization, results interpretation and paper-writing; V.N.S. and M.K. contributed to the conception, design, results interpretation and writing of the paper. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.