Influence of Surface Properties of Nanostructured Ceria-Based Catalysts on Their Stability Performance

As the poor cycling stability of CeO2 catalysts has become the major obstacle for applications of diesel particulate filters (DPF), it is necessary to investigate how to reduce their structural and compositional changes during soot oxidation. In this study, different ratios of Samarium (Sm) were doped into the lattice of CeO2 nanoparticles to improve the catalytic performance as well as surface properties. The stability was investigated by recycling the catalyst, mixing it with soot again, and repeating the thermogravimetric analysis (TGA) tests seven times. Consistent observations were expected for more cycles. It was found that doping 5%, 10%, and 20% samarium into the CeO2 lattice can improve the catalyst stability but at the cost of losing some activity. While the catalyst became more stable with the increasing Sm doping, the 10% Sm-doped catalyst showed the best compromise between stability and activity. Ce3+ and Oα were found to play important roles in controlling catalytic soot oxidation activity. These two species were directly related to oxygen vacancies and oxygen storage capacity of the catalyst. Sm-doped catalysts showed a minimized decrease in the Ce3+ and Oα content when the fresh and spent catalysts were compared.


Introduction
Ceria has already been widely studied in soot oxidation because of its excellent redox properties and oxygen storage capacity (OSC). Previous studies have shown that metaldoped ceria can further improve the activity of ceria-based catalysts [1]. In a previous study that investigated the activities of iron-doped ceria catalysts at different doping ratios, we showed 5% iron doping prepared with the solution combustion synthesis (SCS) method performed best due to its highly reactive Fe-O-Ce sites [2].
It was reported that the active sites, crystal structures, and surface properties of ceriabased catalysts could change during the reactions, resulting in poor stability of ceria-based catalysts [3]. All studies acknowledged that pure CeO 2 shows poor stability for soot oxidation, which has become a major obstacle for the wide-ranging applications of ceriabased catalysts [3][4][5][6]. Liang et al. [3] have studied the thermal stability of CeO 2 for soot oxidation by aging the catalyst at 800 • C for 20 h in air. They found that the temperature corresponding to the maximum rate for CO 2 production with an aged catalyst increased by about 60 • C compared to a fresh catalyst because of the highly reduced surface area and sintering. Aneggi et al. [4] observed that an aging CeO 2 catalyst at 750 • C for 12 h in air decreased its activity for soot oxidation as it lost oxygen storage capacity. Liu et al. [5] investigated the deactivation of CeO 2 for soot oxidation through isothermal conditions at 300 and 350 • C; they found that deactivation occurred at isothermal conditions and became more severe at higher temperature. The main reported reason for deactivation is inefficient O − 2 formation. Corro et al. [6] tested the stability of CeO 2 towards soot oxidation during six cycles and found a slow and continuous deactivation of CeO 2 . catalyst preparation method. Both reported increases in combustion temperatures after aging but are not conclusive regarding actual catalyst stability. Anantharaman et al. [34] studied the effect of Sm content on soot-oxidation reactivity for the Sm-doped ceria catalyst prepared by the EDTA-citrate method but did not investigate catalyst stability; they found that 10% Sm doping performed best on a fresh catalyst calcined at 600 • C for 5 h.
The present study aims to develop an optimum ratio of Sm doping into the CeO 2 catalyst to improve catalyst stability for soot oxidation. Therefore, different ratios of Sm doping were investigated along with its influence on catalyst surface properties and activity. Characterizations were performed to understand surface and crystal properties of catalyst. Seven repeating reaction cycles were used to study the activity and stability of the Sm-doped CeO 2 catalyst.

Catalysts' Activity and Stability Tests
Thermogravimetric analysis (TGA) was used to test the catalyst's activity for soot oxidation through a TA Instrument Q500 apparatus (TA Instruments, New Castle, DE, USA). Printex-U carbon black was used as the model of soot. Catalyst and soot particles were weighted at a ratio of 9:1 to make sure soot can be fully oxidized, after a number of trials. They were then mixed by grinding them for 10 min to obtain a tight-contact condition [35]. For the first cycle of the TGA test, a weighted amount of a 40 mg sample was pretreated at 150 • C under 60 mL/min nitrogen for 30 min to remove water. Then, the sample was heated up to 600 • C under 40 mL/min air with a heating rate of 10 • C/min. The thermogravimetric curves were obtained by continuously recording the mass change, along with increased temperature. Soot oxidation in the absence of catalysts was performed for the comparison purpose. Note that from our past observations, heating the carbon-catalyst mixture to 600 • C is sufficient to achieve the soot conversion rate of 100% for ceria catalysts, while the pure carbon sample is completely oxidized after the temperature reaches 700 • C [2]. The catalyst's activity was evaluated by T50. T50 is the combustion temperature, identified as the temperature when 50% of soot is oxidized [36][37][38][39].
The stability tests were performed through several soot oxidation cycles using the same catalyst. One cycle is defined by a step of mixing soot and catalyst particles, followed by a step of soot oxidation in the TGA. Once the first cycle soot-oxidation reaction was completed, the remaining catalyst was collected and remixed with soot particles again under a tight-contact condition with the same weighting ratio of 9:1. The TGA test was run again with the newly mixed sample. This procedure was repeated for several cycles. Note that after each cycle, it was inevitable that some amount of the catalyst is lost (mostly during the mixing step where several particles stick to the mortar and pestle). This means that for a given amount of fresh catalyst there will be a maximum number of cycles that can be investigated; for example, with our typical 36 mg of fresh catalyst (with 4 mg of carbon), it was hardly possible to go beyond seven cycles.

Characterization of the Catalysts
An X-ray powder diffraction (XRD) analysis was carried out using an X-ray powder diffractometer (German Bruker D4 (40 kV, 30 mA), Bruker, Billerica, MI, USA with a position-sensitive detector and CuKα radiation). The XRD patterns were recorded from 5 • to 85 • in steps of 0.01 • . The XRD patterns were analyzed based on the Powder Data File database (International Centre of Diffraction Data, Newtown Square, PA, USA, accessed on: 1 September 2021). Specific surface areas were evaluated by N 2 adsorption-desorption isotherms on a Beishide 3H-2000PS2 static-volumetric method analyzer, BEISHIDE Instrument Technology, Beijing, China. The Brunauer-Emmett-Teller (BET) method was used to calculate the catalysts' surface areas. The catalysts' morphologies and microstructures were characterized by a Field-emission scanning electron microscope (FE-SEM, Zeiss MERLIN with Gemini-II column, Zeiss, Oberkochen, Germany). The X-ray photoelectron spectroscopy (XPS) analysis was conducted on the ESCALab220i-XL electron spectrometer (VG Scientific Ltd., East Grinstead, UK) with a 300 W AlKα X-ray source to study the oxidation states and oxygen species on the catalysts' surfaces. Binding energy was calibrated by standard C1s peaks at 284.8 eV. XPS peaks was analyzed using CasaXPS software, Version 2.3.24, Clearwater, FL, USA. The Raman spectra of catalysts were measured on a Renishaw inVia micro laser Raman spectrometer (Renishaw plc, Wotton-under-Edge, UK) with a 4 mW Ar + laser source (λ ex = 532 nm) and a cooled CCD detector at room temperature to distinguish chemical structures. The scanning range was 200-800cm −1 with a 60 s acquisition time.

Fresh and Spent Catalysts' Characterizations
The crystalline structures of catalysts were studied by XRD. Figure 1 presents the XRD pattern of fresh and spent Sm-doped CeO 2 catalysts, as well as pure CeO 2 for comparison. All XRD patterns show similar main diffractions peaks, which can be attributed to (111), (200), (220), (311), (222), (400), (331), and (420) planes. These peaks refer to typical cubic fluorite structures of pure CeO 2 [40]. No peaks referring to Sm 2 O 3 were found, even for 20% SDC, meaning that no individual Sm 2 O 3 -crystal structure was detected, indicating that the Sm forms a solid solution in the ceria lattice. For fresh catalysts, when increasing the Sm-doping ratio, the characteristic peaks shift to the lower 2θ diffraction angles, further suggesting that Sm is incorporated into the CeO 2 lattice and forms a solid solution. The unit-lattice parameters were calculated by Bragg's law using the strongest peak (111), and the crystallite size was calculated by the Scherrer equation. With an increase in Sm-doping, the lattice parameter (as shown in Table 1) tends to increase because of the larger Sm (1.07 Å) substituting smaller ceria (0.97 Å) in the lattice, thus resulting in lattice expansion [41]. an increase in crystallinity after several cycles of reactions on the catalyst. For all sp catalysts, Sm0.1Ce0.9 notably shows the lowest crystallite size.  The BET surface areas for Sm-doped ceria catalysts are listed in Table 1. For the f Sm-doped catalysts, except Sm0.1Ce0.9, they all have a surface area of 41-42 m 2 g −1 , w are lower than that of fresh pure CeO2 (45.5 m 2 g −1 ). The fresh Sm0.1Ce0.9 catalyst shows highest BET surface area of 48 m 2 g −1 . When comparing fresh and spent catalysts, the surface area of pure CeO2 decreases the most (from 45.5 to 40.9 m 2 g −1 ). Sm-doped C catalysts also show a slight decline in the BET surface area (2-4% loss), regardless o doping ratio. Interestingly, except for Sm0.1Ce0.9, all other spent catalysts, including p ceria, have similar surface areas at around 40-41 m 2 g −1 . There seems to be something ticular for Sm0.1Ce0.9, which has the highest surface area for both fresh and spent catal Note that this catalyst also stands out, having the lowest crystallite size after severa cles. Since the catalysts for XRD and BET characterization were made from diffe batches, it indicates that Sm0.1Ce0.9 presents interesting properties, which cannot be tributed to experimental error. As seen in Table 1, the comparison of fresh and spent catalysts shows that the lattice parameters of spent catalysts slightly increase compared to those of fresh catalysts with the same doping ratio. Although those increases are very small, the trend is clear that the change in lattice parameters decreases as the amount of Sm increases: for pure ceria, 5%, 10%, and 20% Sm; the differences in lattice parameters between fresh and spent catalysts are 0.013, 0.006, 0.004 and 0.003 Å, respectively. While for fresh catalysts, the crystallite size clearly increases as the amount of Sm increases (from 12.3 nm for pure ceria to 13.6 nm for Sm 0.2 Ce 0.8 ); such a clear trend was not observed for the spent catalysts where the crystallite size seems to converge for the size range of between 13.5 and 14.5 nm. The spent catalysts always show higher crystallite sizes than the fresh ones, indicating that there is an increase in crystallinity after several cycles of reactions on the catalyst. For all spent catalysts, Sm 0.1 Ce 0.9 notably shows the lowest crystallite size.
The BET surface areas for Sm-doped ceria catalysts are listed in Table 1. For the fresh Sm-doped catalysts, except Sm 0.1 Ce 0.9 , they all have a surface area of 41-42 m 2 g −1 , which are lower than that of fresh pure CeO 2 (45.5 m 2 g −1 ). The fresh Sm 0.1 Ce 0.9 catalyst shows the highest BET surface area of 48 m 2 g −1 . When comparing fresh and spent catalysts, the BET surface area of pure CeO 2 decreases the most (from 45.5 to 40.9 m 2 g −1 ). Sm-doped CeO 2 catalysts also show a slight decline in the BET surface area (2-4% loss), regardless of the doping ratio. Interestingly, except for Sm 0.1 Ce 0.9 , all other spent catalysts, including pure ceria, have similar surface areas at around 40-41 m 2 g −1 . There seems to be something particular for Sm 0.1 Ce 0.9 , which has the highest surface area for both fresh and spent catalysts. Note that this catalyst also stands out, having the lowest crystallite size after several cycles. Since the catalysts for XRD and BET characterization were made from different batches, it indicates that Sm 0.1 Ce 0.9 presents interesting properties, which cannot be attributed to experimental error.
An SEM was used to study the morphologies of the produced catalysts. Figure 2 shows the SEM images of all fresh and spent catalysts. All fresh catalysts have spongy structures with large openings, which are due to their fast combustion reaction during the SCS procedure. The main effect of Sm doping on the catalyst morphology is to increase those An SEM was used to study the morphologies of the produced catalysts. Figure 2 shows the SEM images of all fresh and spent catalysts. All fresh catalysts have spongy structures with large openings, which are due to their fast combustion reaction during the SCS procedure. The main effect of Sm doping on the catalyst morphology is to increase those openings. However, all spent catalysts lose the spongy structure and tend to agglomerate and in the end have very similar morphologies.  XPS was used to detect different oxidation states of each element. Figure 3 is the XPS spectra of all catalysts regarding their Ce3d, O1s, and Sm 3d5/2 spectra. The binding energies of XPS spectra were calibrated using the C1s peak at 284.8 eV. C1s spectra of Nanomaterials 2022, 12, 392 7 of 16 all catalysts are provided in the Supplementary Materials. While performing the Ce3d deconvolution, the FWHM (full width at half maximum) was set to a narrow range of 2-2.7 eV, and the peak area of 3d5/2 to 3d3/2 was set to 3:2. Ce3d spectra can be split into 10 peaks for detailed analysis. The peaks labeled with "u" and "v" correspond to the spin-orbit splitting of Ce 3d5/2 and Ce 3d3/2, respectively. These peaks, v, v2, v3, u, u2, and u3, are the characteristic peaks for the Ce 4+ species, and the peaks noted with v0, v1, u0, and u1 correspond to the Ce 3+ species [42][43][44][45][46]. Detailed peak positions are provided in the Supplementary Materials. It is clear that ceria exists primarily as Ce 4+ for each catalyst, with the coexistence of some Ce 3+ . Since oxygen vacancy is generated when Ce 4+ is reduced to Ce 3+ , it is crucial to calculate the Ce 3+ percentage to evaluate the generation of oxygen vacancy on the catalyst surface [37,46]. The calculations of the integrated peak areas can be used to quantitatively analyze XPS results. The ratio of Ce 3+ was calculated by dividing the peak areas of Ce 3+ by the total peak area.  The O1s spectra were curve fitted into two peaks, including lattice oxygen, O β , at a lower binding energy of 529.5 eV and the surface oxygen species, O α , at a higher binding energy of 531.7 eV [47]. The concentration of O α is critical to evaluate the oxygen storage capacity (OSC) and the potential active oxygen for soot oxidation. The ratio, O α /(O α + O β ), was calculated by dividing the peak areas of O α by the total peak areas, whose results are shown in Table 2. This table shows that for fresh catalysts, Sm doping up to 10% marginally decreases the O α ratio (from 39.6% down to 39.1%). However, the decrease in the O α ratio is more pronounced for 20% Sm doping (O α ratio of 34.2%). The results regarding the O α ratio are in agreement with that of the Ce 3+ percentage. Regarding spent catalysts, the O α ratio of pure CeO 2 drops significantly compared to that of the fresh CeO 2 catalyst (from 39.6% to 29.5%) and becomes the lowest among all spent catalysts. On the other hand, the O α percentage for 5% and 10% Sm-doped ceria catalysts decreases by less than 2 percentage points. Finally, the 20% Sm-doped catalysts do not show a decrease in O α content, although its final value is still lower than that of the other two Sm-doped ceria catalysts. Those results suggest that Sm doping considerably reduces the loss in surface oxygen after several cycles.   The peaks around 1082.6 eV in the Sm 3d5/2 spectra confirm that Sm is present as a +3-oxidation state in the catalyst, which agrees with the literature [48,49]. The surface ratio of Sm/Ce was calculated through XPS results. For fresh catalysts, the ratios are 0.059, 0.122, and 0.342, for 5%, 10%, and 20% Sm doping, respectively. Compared to the stoichiometric ratios of 0.053, 0.111, and 0.25 for 5%, 10%, and 20% Sm doping, respectively, the fresh catalysts show slightly higher values, indicating there is an enrichment of Sm on the detected surface of these catalysts. For spent catalysts, the Sm/Ce are 0.068, 0.138, and 0.355 for 5%, 10%, and 20% Sm doping, respectively. These ratios are greater in comparison with the fresh catalyst values, which is possibly due to Sm in the bulk catalysts tending to move to the surface of catalysts during the reactions. The 20% Sm doping shows the smallest change of Sm/Ce, suggesting a more stable catalyst surface.
The Raman spectra of Sm-doped CeO 2 catalysts are depicted in Figure 4. The dominant band around 463 cm −1 corresponds to the F2g symmetric oxygen mode within a CeO 2 cubic fluorite structure, which is in agreement with the XRD results [33]. With an increase of the Sm doping ratio, the F2g band shifts to lower frequencies due to the cell expansion in the catalysts. No peaks of Sm 2 O 3 (~375 cm −1 ) were found, which reinforces that Sm-Ce forms a solid solution [27]. The weak peak around 596 cm −1 represents the surface defects, including the intrinsic oxygen vacancies caused by Ce 3+ in the lattice [35]. The peak around 554 cm −1 refers to the extrinsic oxygen vacancy created by Sm 3+ substituting Ce 4+ to maintain charge neutrality [50]. With an increase in Sm doping, the peak intensity at 554 cm −1 becomes stronger, indicating that more Sm 3+ -associated oxygen vacancies are created. The ratio, I 554 /I 596 , represents, therefore, the ratio of oxygen vacancies originating from Sm 3+ over that from Ce 3+ . The results pertaining to the I 554 /I 596 ratio are shown in Table 3. For fresh catalysts, as expected, the incorporation of Sm increases the I 554 /I 596 ratio. Note that for spent Sm 0.2 Ce 0.8 , this ratio is above 1, indicating more oxygen vacancies from Sm 3+ than from Ce 3+ , whereas it is the opposite for Sm content below 10% (values of the I 554 /I 596 ratios below 1). The Raman spectra of Sm-doped CeO2 catalysts are depicted in Figure 4. nant band around 463 cm −1 corresponds to the F2g symmetric oxygen mode wit cubic fluorite structure, which is in agreement with the XRD results [33]. With a of the Sm doping ratio, the F2g band shifts to lower frequencies due to the cell in the catalysts. No peaks of Sm2O3 (~375 cm −1 ) were found, which reinforces t forms a solid solution [27]. The weak peak around 596 cm −1 represents the surfa including the intrinsic oxygen vacancies caused by Ce 3+ in the lattice [35]. The pe 554 cm −1 refers to the extrinsic oxygen vacancy created by Sm 3+ substituting Ce tain charge neutrality [50]. With an increase in Sm doping, the peak intensity a becomes stronger, indicating that more Sm 3+ -associated oxygen vacancies are cr ratio, I554/I596, represents, therefore, the ratio of oxygen vacancies originating over that from Ce 3+ . The results pertaining to the I554/I596 ratio are shown in T fresh catalysts, as expected, the incorporation of Sm increases the I554/I596 ratio for spent Sm0.2Ce0.8, this ratio is above 1, indicating more oxygen vacancies from from Ce 3+ , whereas it is the opposite for Sm content below 10% (values of the I55 below 1).   After five cycles of reactions, the I 554 /I 596 ratio for all catalysts further increases (by about 0.1). This indicates that after several cycles, the relative amount of oxygen vacancy correlated to Sm 3+ increases (and its corollary, the relative amount of oxygen vacancy correlated to Ce 3+ decreases). The peak intensity of the Raman spectra increases in all cases after five cycles, possibly due to the increase in crystal growth, which is in agreement with the XRD results. Another reason could be the lattice distortion of the spent catalysts leading to resonance with the inlet Ar + laser source.

Activity Test
TGA experiments were performed to study the activity of the fresh Sm-doped CeO 2 catalyst for soot oxidation. Figure 5 depicts soot conversion as a function of temperature. The fresh pure CeO 2 catalyst achieves the highest activity, as indicated by the lowest T50. Increasing the Sm content in the fresh catalysts tends to increase T50, revealing a loss in activity. In fact, the 5% and 10% Sm-doping ratios for the fresh catalyst show similar effects on catalyst activity, whereas the 20% Sm-doping sample leads to the worst activity. In comparison, soot oxidation without the catalyst was conducted under the same reaction condition (Printex-U in Figure 5), and it shows very little soot oxidation below 500 • C, suggesting that the catalyst investigated here significantly increases soot oxidation.
ing to resonance with the inlet Ar laser source.

Activity Test
TGA experiments were performed to study the activity of the fresh Sm-doped CeO2 catalyst for soot oxidation. Figure 5 depicts soot conversion as a function of temperature. The fresh pure CeO2 catalyst achieves the highest activity, as indicated by the lowest T50. Increasing the Sm content in the fresh catalysts tends to increase T50, revealing a loss in activity. In fact, the 5% and 10% Sm-doping ratios for the fresh catalyst show similar effects on catalyst activity, whereas the 20% Sm-doping sample leads to the worst activity. In comparison, soot oxidation without the catalyst was conducted under the same reaction condition (Printex-U in Figure 5), and it shows very little soot oxidation below 500 °C, suggesting that the catalyst investigated here significantly increases soot oxidation.
Since catalytic soot oxidation is a surface sensitive reaction, i.e., reaction only happens at the contact point of soot and catalyst, the crystallite size, specific surface area, and morphology can all play important roles on the catalyst activity [1,51]. From Figure 2, the morphology does not change significantly between pure ceria and Sm-doped catalysts, and thus cannot explain the difference in reactivity. Table 1 indicates that the BET surface area increases slightly when increasing the Sm content, but this would lead to higher activity at the higher Sm content, which is the opposite of what was observed; the BET surface area, thus, cannot explain the trend in activity here. Finally, Table 1 shows a modest change in the crystallite size when increasing the Sm content and without a clear trend. Crystallite size can, therefore, also be dismissed as an explanation for the activity trend for the fresh catalysts. Soot oxidation on ceria-based catalysts occurs via the so-called Mars-van Krevelen mechanism [1], where the active surface oxygen reacts with soot through the catalyst-soot contact point. Once the active oxygen is consumed, an oxygen vacancy is generated at the Since catalytic soot oxidation is a surface sensitive reaction, i.e., reaction only happens at the contact point of soot and catalyst, the crystallite size, specific surface area, and morphology can all play important roles on the catalyst activity [1,51]. From Figure 2, the morphology does not change significantly between pure ceria and Sm-doped catalysts, and thus cannot explain the difference in reactivity. Table 1 indicates that the BET surface area increases slightly when increasing the Sm content, but this would lead to higher activity at the higher Sm content, which is the opposite of what was observed; the BET surface area, thus, cannot explain the trend in activity here. Finally, Table 1 shows a modest change in the crystallite size when increasing the Sm content and without a clear trend. Crystallite size can, therefore, also be dismissed as an explanation for the activity trend for the fresh catalysts.
Soot oxidation on ceria-based catalysts occurs via the so-called Mars-van Krevelen mechanism [1], where the active surface oxygen reacts with soot through the catalyst-soot contact point. Once the active oxygen is consumed, an oxygen vacancy is generated at the surface, and bulk O 2 fills this vacancy and create another active oxygen. Therefore, factors affecting the amount of active surface oxygen and oxygen vacancies can also play important roles in determining catalyst activity. The Ce 3+ content is a good representation of oxygen-vacancy generation, which, in turn, can be a potential site for creating active surface oxygen. Here, surface oxygen and oxygen vacancies are related to the Ce 3+ content, O α percentage, and I 554 /I 596 ratio. Those are discussed next. Figure 6a shows the relation between T50 and the Ce 3+ percentage, which indicates that the higher the Ce 3+ content, the lower T50 (and thus, higher the activity). The highest Ce 3+ ratio (24.2%) corresponds to the fresh pure CeO 2 catalyst, which can be a reason for the highest activity of pure ceria. The activity results reported in Figure 5 for Sm-doped catalysts correlates very well with the Ce 3+ content shown in Table 2: 5%-and 10%-doped CeO 2 have identical Ce 3+ content and also very similar activity (e.g., same T50), whereas Sm 0.2 Ce 0.8 has a much lower Ce 3+ content, as well as much lower activity (i.e., higher T50). Figure 6b shows the relation between T50 and the O α percentage, which follows the trend of the Ce 3+ content, indicating that surface-oxygen species are also vital for catalyst activity. It has been reported that the higher O α concentration could result in superior catalyst activity, as it can evaluate the oxygen-storage capacity and potential active oxygen for the reaction. From Table 2, the O α percentage slightly decreases for 5% and 10% Sm-doped catalysts, but it considerably decreases for 20% Sm doping when comparing with pure CeO 2 ; the O α percentage data also correlates very well with the activity results.
that the higher the Ce 3+ content, the lower T50 (and thus, higher the activity). The highest Ce 3+ ratio (24.2%) corresponds to the fresh pure CeO2 catalyst, which can be a reason for the highest activity of pure ceria. The activity results reported in Figure 5 for Sm-doped catalysts correlates very well with the Ce 3+ content shown in Table 2: 5%-and 10%-doped CeO2 have identical Ce 3+ content and also very similar activity (e.g., same T50), whereas Sm0.2Ce0.8 has a much lower Ce 3+ content, as well as much lower activity (i.e., higher T50). Figure 6b shows the relation between T50 and the Oα percentage, which follows the trend of the Ce 3+ content, indicating that surface-oxygen species are also vital for catalyst activity. It has been reported that the higher O concentration could result in superior catalyst activity, as it can evaluate the oxygen-storage capacity and potential active oxygen for the reaction. From Table 2, the O percentage slightly decreases for 5% and 10% Sm-doped catalysts, but it considerably decreases for 20% Sm doping when comparing with pure CeO2; the O percentage data also correlates very well with the activity results.
The relation between T50 and the I554/I596 ratio is illustrated in Figure 7, which shows that T50 increases when the I554/I596 ratio increases. This implies that Sm 3+ -associated oxygen vacancy decreases the catalyst activity. From the XPS results, it is already known that higher Ce 3+ -associated oxygen vacancy plays a key role in improving catalyst activity. Therefore, both the increase in oxygen vacancy around Sm 3+ and the decrease of oxygen vacancy adjacent to Ce 3+ lead to activity decline.  The relation between T50 and the I 554 /I 596 ratio is illustrated in Figure 7, which shows that T50 increases when the I 554 /I 596 ratio increases. This implies that Sm 3+ -associated oxygen vacancy decreases the catalyst activity. From the XPS results, it is already known that higher Ce 3+ -associated oxygen vacancy plays a key role in improving catalyst activity. Therefore, both the increase in oxygen vacancy around Sm 3+ and the decrease of oxygen vacancy adjacent to Ce 3+ lead to activity decline.  Figure 8 presents the results of stability on Sm-doped CeO 2 for soot oxidation through seven cycles. This figure shows that T50 for the pure CeO 2 catalyst increases after each cycle, going from the lowest value (379 • C) with the fresh catalyst to the highest one (462 • C) after seven cycles. The 5% Sm-doped catalyst also shows a continuous increase in T50 after each cycle but at a much lower rate than pure CeO 2 . The samples with 10% and 20% Sm doping follow very similar trends (as seen in Figure 8, they are almost parallel): first, a moderate increase in T50 after three or four cycles, followed by a near plateauing of T50. The differences in T50 between the initial and plateau values are 14 • C for 10% Sm and 8 • C for 20% Sm, respectively. Those results indicate good stability for Sm doping above 10%. As a plateau is reached after the fifth cycle, it is expected that the T50 of catalysts containing more than 10% Sm doping would have a very limited change for more cycles. Considering activity and stability together, the 10% Sm-doped catalyst performs best among all catalysts, with a stability comparable to the 20% Sm but with the lowest T50 after seven cycles (412 • C). Figure 7. The relation between T50 and I554/I596. Figure 8 presents the results of stability on Sm-doped CeO2 for soot oxidation through seven cycles. This figure shows that T50 for the pure CeO2 catalyst increases after each cycle, going from the lowest value (379 °C ) with the fresh catalyst to the highest one (462 °C ) after seven cycles. The 5% Sm-doped catalyst also shows a continuous increase in T50 after each cycle but at a much lower rate than pure CeO2. The samples with 10% and 20% Sm doping follow very similar trends (as seen in Figure 8, they are almost parallel): first, a moderate increase in T50 after three or four cycles, followed by a near plateauing of T50. The differences in T50 between the initial and plateau values are 14 °C for 10% Sm and 8 °C for 20% Sm, respectively. Those results indicate good stability for Sm doping above 10%. As a plateau is reached after the fifth cycle, it is expected that the T50 of catalysts containing more than 10% Sm doping would have a very limited change for more cycles. Considering activity and stability together, the 10% Sm-doped catalyst performs best among all catalysts, with a stability comparable to the 20% Sm but with the lowest T50 after seven cycles (412 °C).  Figure 9 shows the relationship between the change of T50 (between first and seventh cycle) and the Sm-doping ratio. The spent CeO2 catalyst leads to an 83 °C change in T50, compared to the fresh one. T50 of 5% and 10% Sm-doped catalysts show smaller differences, which are about 27 °C and 18 °C, respectively. T50 for the 20% Sm-doped catalyst shows the lowest difference (13 °C ). It is clear that the T50 difference decreases with higher Sm doping, indicating that richer Sm contributes to a more stable catalyst, albeit with lower activity.  Figure 9 shows the relationship between the change of T50 (between first and seventh cycle) and the Sm-doping ratio. The spent CeO 2 catalyst leads to an 83 • C change in T50, compared to the fresh one. T50 of 5% and 10% Sm-doped catalysts show smaller differences, which are about 27 • C and 18 • C, respectively. T50 for the 20% Sm-doped catalyst shows the lowest difference (13 • C). It is clear that the T50 difference decreases with higher Sm doping, indicating that richer Sm contributes to a more stable catalyst, albeit with lower activity. Nanomaterials 2022, 12, x FOR PEER REVIEW 13 o Figure 9. Relationship between T50 change (between first and seventh cycles) and Sm-doping r tio.

Stability Tests for Sm-Doped CeO2 Catalysts
As discussed previously, the Ce 3+ content plays an important role for catalyst acti towards soot oxidation. Figure 10 illustrates the relationship between the Ce 3+ con Figure 9. Relationship between T50 change (between first and seventh cycles) and Sm-doping ratio.
As discussed previously, the Ce 3+ content plays an important role for catalyst activity towards soot oxidation. Figure 10 illustrates the relationship between the Ce 3+ content change (between first and fifth cycle) and the Sm-doping ratio. This figure indicates that the addition of Sm (even at 5%) results in a significant decrease in the change of Ce 3+ between the fresh and spent catalysts. The Ce 3+ content for pure ceria decreases from 24.2% to 15.8%, suggesting poor stability performance. In contrast, changes in the Ce 3+ content of Sm-doped ceria catalysts are below 1%, with 20% Sm doping only 0.5%. This indicates that Sm doping is beneficial for improving catalyst stability and maintaining activity. Figure 10 also shows that the relationship between the change of the O α percentage (between first and fifth cycle) and the Sm-doping ratio. It can be observed that with an increase in the doping ratio, the decrease of the O α percentage becomes smaller. The 5% and 10% Smdoped catalysts decrease by about 2%, whereas the 20% Sm catalyst does not show any change. This suggests that the higher Sm doping makes a more stable catalyst and prevents the decline of the O α percentage. The O α percentage of the fresh CeO 2 catalyst is 39.6% and yields to 29.5% after five cycles. This indicates that the OSC of the CeO 2 catalysts is not stable, and the significant decline of OSC after reacting with soot leads to poor stability of pure ceria catalysts.

Conclusions
Doping 5%, 10%, and 20% samarium into a CeO2 lattice was lution combustion synthesis (SCS) method to improve the stability ceria-based catalyst for soot oxidation. In general, at a higher Sm c the reactivity (represented by T50) nearly plateaued after five cycle the Sm content was accompanied with a decrease in activity for soo ing the effects of Sm doping in both stability and activity, the 1 showed the best compromise between the catalyst stability and act found to play an important role in controlling catalytic soot-oxi agrees with the expectation that they are directly related to the oxy ygen storage capacity of the catalyst. Sm-doped catalysts, especial Moreover, the XRD results have shown that Sm is incorporated into the CeO 2 lattice and causes lattice expansion. The increased crystallite size of the spent catalyst compared with the fresh one indicates that recrystallization has occurred. However, the larger crystallite size does not benefit the catalyst activity towards soot oxidation, as smaller crystallite size could favor oxygen diffusion and provide more contact points. The Sm-doped catalyst shows a smaller change in crystallite size after reactions, and such a difference becomes lesser with higher amounts of Sm doping, indicating that adding Sm can inhibit this recrystallization process. Therefore, adding Sm can stabilize the catalyst crystal structure and helps maintain surface properties, such as the Ce 3+ and O α percentages.

Conclusions
Doping 5%, 10%, and 20% samarium into a CeO 2 lattice was achieved using the solution combustion synthesis (SCS) method to improve the stability of the nanostructured ceria-based catalyst for soot oxidation. In general, at a higher Sm content (10% and 20%), the reactivity (represented by T50) nearly plateaued after five cycles. However, increasing the Sm content was accompanied with a decrease in activity for soot oxidation. Considering the effects of Sm doping in both stability and activity, the 10% Sm-doped catalyst showed the best compromise between the catalyst stability and activity. Ce 3+ and O α were found to play an important role in controlling catalytic soot-oxidation activity, which agrees with the expectation that they are directly related to the oxygen vacancies and oxygen storage capacity of the catalyst. Sm-doped catalysts, especially at 10% and 20% Sm, showed a minimized decrease in Ce 3+ content and O α percentage between the fresh and spent catalysts after five cycles. Good stability in maintaining the crystallite size between the fresh and spent catalysts can also contribute to the greater stability for Sm-doped ceria catalysts. However, our experimental data did not indicate a clear correlation between the crystallite size and activity for these fresh and spent catalysts. Therefore, crystallite size growth is not considered here as a primary parameter for catalyst deactivation over cycling.