The composite oxide CeO₂-ZrO₂-Y₂O₃-La₂O₃ + La₂O₃-Al₂O₃ (CeZrYLa+LaAl) was obtained by precipitating the Ce(NO₃)₃·6H₂O, ZrOCO₃, La(NO₃)₃·6H₂O, and Y(NO₃)₃·6H₂O mixed solution and La(NO₃)₃·6H₂O and Al(NO₃)₃·9H₂O mixed solution with NH₃·H₂O precipitator, respectively. And then the precursor of Y³⁺ and La³⁺-modified ceria-zirconia and La³⁺-modified alumina were mixed together with intense stirring. The precipitate was filtered, washed, dried and then calcined at 800 °C for 3 h in a muffle furnace to obtain the support. The resulting CeZrYLa+LaAlcomposite oxide had a CeZrYLa/LaAl mass ratio of 7:3 and Ce/Zr/Y/La and La/Al mass ratios of 45:45:5:5 and 3:97.

Using the support prepared above as a carrier, Pt-Rh catalysts were synthesized using different preparation methods. The sample termed Cat1 was prepared using a conventional co-impregnation method. In this method, mixed aqueous solutions of Pt(NO₃)₂ and Rh(NO₃)₃ were co-impregnated on the support powder, which was then dried overnight at 120 °C. Cat2 was obtained by impregnating a quantity of the support with Pt(NO₃)₂ and a second quantity with Rh(NO₃)₃. After drying at 120 °C overnight, the two portions were mixed together. Cat3 was prepared in the same manner as Cat2 except adding 3.7 wt% La³⁺, 6.3 wt% Zr⁴⁺ and 3.7 wt% Ba²⁺ to the mixed powder. The Pt and Rh loadings in each catalyst support were 1.28 and 0.14 wt%, respectively. Each of the prepared powder catalysts was calcined at 550 °C for 3 h and ball-milled with the appropriate amount of deionized water to obtain a homogeneous slurry, following which the resulting slurry was coated onto honeycomb cordierite (length: 2.5 cm, volume: 2.7 cm³, Corning). The coated catalysts were dried overnight at 120 °C and calcined at 550 °C in air for 3 h to obtain the monolithic catalysts. The catalyst loading were approximately 150 g L⁻¹ and the Pt-Rh loading was 2.14 g L⁻¹ (60 g/ft³) with a 9:1 mass ratio.

The textural properties of the supports were assessed by N₂ adsorption-desorption at −196 °C, using an automated surface area and pore size analyzer (Autosorb SI, Quantachrome, USA). Samples were pretreated at 300 °C for 5 h prior to these analyses. Surface areas were determined using the BET model.

Powder X-ray diffraction (XRD) data were acquired on a Japan Science D/max-RA diffractometer using Cu K_α (λ = 0.15406 nm) radiation. The tube voltage and current were 40 kV and 100 mA, respectively and patterns were acquired over the range of 10°-80°.

X-ray photoelectron spectroscopy (XPS) data were obtained using a Kratos XSAM-800 spectrophotometer operating at 13 kV and 20 mA with Al K_α radiation (1486.6 eV) in the constant pass energy mode (20 eV pass energy). The C 1s peak at 284.8 eV was used for calibration of binding energy (BE) values. The pressure in the analytical chamber was approximately 10⁻⁹ Pa.

Hydrogen temperature-programmed reduction (H₂-TPR) analyses were performed using instrumentation made in-house and incorporating a thermal conductivity detector. Samples of 100 mg were pretreated in a quartz tubular micro-reactor under a flow of pure N₂ at 450 °C for 1 h to ensure clean surfaces and then cooled to room temperature. Reduction was carried out under a flow of 5% H₂/N₂ between room temperature and 900 °C at a heating rate of 8 °C/min.

Catalytic activity test was performed in a multiple fixed bed continuous flow micro-reactor using a gas mixture simulating the exhaust from a NGV operating under stoichiometric condition. Trials were performed from 100 to 500 °C. The simulated exhaust was composed of CH₄ (0.087 vol%), CO (0.40 vol%), NO (0.073 vol%), H₂O (10-12 gas/vol%) and CO₂ (12 vol%), with N₂ as the balance. The O₂ content was adjusted to obtain a stoichiometric mixture. The gas hourly space velocity (GHSV) was 34000 h⁻¹. The gases were regulated using mass-flow controllers before entering the reactor. The air-fuel ratio window, defined as the ratio between available oxygen and oxygen required for the conversion of CH₄, CO and NO, was assessed at 500 °C. The concentration of CH₄ was analyzed by an online gas chromatograph equipped with a flame ionization detector (FID), while CO and NO were determined online using a five-component analyzer FGA-4100 (Fofen Analytical Instrument Co., Ltd., Foshan, China) before and after the simulated exhaust gas passed through the reactor. The conversions of CH₄, CO, and NO were calculated using the following formula:

Conversion = ((C_in-C_out)/C_in) × 100% (1)

Here C_in is the component concentration in the original simulated exhaust mixture prior to the micro-reactor, and C_out is the concentration after the micro-reactor.

The BET specific surface area, pore volume, and pore size of CeZrYLa+LaAl sample calcined at 800 °C are 148.2 m²/g, 0.40 ml/g, and 5.05 nm, respectively.

Fig. 1 shows the XRD patterns of the support and catalysts. The main diffraction peaks of all samples were consistent with the characteristic peaks of tetragonal Ce_0.5Zr_0.5O₂ (PDF-ICDD38- 1436). These results are also in good agreement with the reported by ones ^{[21, 22]}, in which tetragonal phases were formed in the 5 to 60 mol% CeO₂ composition range. Diffraction peaks corresponding to CeO₂ and ZrO₂ were not detected, suggesting that the Ce and Zr were primarily present as CeO₂-ZrO₂ mixed oxides. The patterns do not contain any γ-Al₂O₃ peaks, and the absence of characteristic Al₂O₃ peaks indicates that the Al₂O₃ was either incorporated into the CeO₂-ZrO₂ lattice to form a solid solution or formed grains that were too small to be detected by XRD ^[23]. Pt and Rh phases were not observed, indicating that the noble metals were well dispersed on the support or that the Pt and Rh concentrations were too low to be detected. Furthermore, a peak attributed to monoclinic ZrO₂ (2θ = 24.2°) was present in the pattern of the Cat3 sample because of the addition of 6.3 wt% ZrO₂.

Fig. 1. (a) XRD patterns of support and catalysts; (b) XRD patterns in the 2θ range from 26° to 33°.

The main (101) diffraction peaks appear between 26° and 33° and this region has been enlarged and is presented as Fig. 1(b). The associated crystallite sizes (calculated based on Bragg’s law: 2dsinθ = kλ) and other parameters are summarized in Table 1. The peak at 29.32° was assigned to the (101) planes of tetragonal Ce_0.5Zr_0.5O₂. The peak intensities of the Cat1 and Cat2 are slightly reduced compared with the support. In contrast, the peak intensities of the Cat3 sample are significantly reduced, demonstrating that the additive cations (La³⁺, Zr⁴⁺, Ba²⁺ and especially Zr⁴⁺) were incorporated into the CeO₂-ZrO₂ lattice, thus increasing its structural defects and reducing the degree of crystallinity. Moreover, the Cat1 and Cat2 peaks were shifted to slightly higher 2θ values, while the Cat3 exhibited a more pronounced shift from 29.32° to 29.56°. These results are in good agreement with the observed decreases in the lattice parameters (Table 1). The ionic radii of Pt²⁺, Pt⁴⁺, and Rh³⁺ are 0.080, 0.063 and 0.0665 nm, all of which are smaller than those of Ce⁴⁺ (0.097 nm) and Zr⁴⁺ (0.084 nm). Thus, a decrease in the lattice parameters indicates that a small quantity of Pt²⁺, Pt⁴⁺, or Rh³⁺ ions have possibly dissolved in the CeO₂-ZrO₂ lattice. The decrease in the lattice parameters in the case of Cat3 was greater than for the Cat1 and Cat2 because of the addition of Zr⁴⁺ (6.3 wt%), since the ionic radius of Zr⁴⁺ (0.084 nm) is smaller than that of Ce⁴⁺ (0.097 nm) ^{[24, 25]}. On the contrary, lower concentrations of La³⁺ and Ba²⁺ were added compared with the Zr⁴⁺ loading, but these ions have larger ionic radii than Ce⁴⁺ and Zr⁴⁺ (0.103 and 0.135 nm, respectively), indicating that the La³⁺ and Ba²⁺ barely dissolved in the CeO₂-ZrO₂ lattice and instead may have been well dispersed over the support. The XRD results demonstrate that the insertion of noble metals and the additive Zr⁴⁺ into the ceria lattice created a higher concentration of defects, improving the O₂ mobility ^{[24, 26]}.

Table 1 Lattice parameter and crystallite sizes as determined by XRD. Sample Main phase composition Cell parameters Crystallite size (nm) a (nm) c (nm) Support Tetragonal Ce0.5Zr0.5O2 0.37360 0.53090 5.1 Cat1 Tetragonal Ce0.5Zr0.5O2 0.37222 0.52968 4.8 Cat2 Tetragonal Ce0.5Zr0.5O2 0.37255 0.53044 4.9 Cat3 Tetragonal Ce0.5Zr0.5O2 0.37086 0.52913 4.8

Table 1 Lattice parameter and crystallite sizes as determined by XRD.

XPS analyses were performed to verify the surface compositions and elemental oxidation states. The Ce 3d XPS spectra obtained from the support and catalysts are presented in Fig. 2, while Table 2 provides the BEs of the characteristic peaks for Ce atoms and the Ce³⁺/Ce ratios for all samples. In Fig. 2, the Ce 3d peaks are seen to be complex, consisting of eight components denoted as u or v. Here the u series represents the Ce 3d_3/2 contribution while the v series represents the Ce 3d_5/2 contribution. The labels u, u2, and u3 were assigned to the Ce⁴⁺ 3d_3/2 peaks and the v, v2 and v3 labels were assigned to the Ce⁴⁺ 3d_5/2 peaks ^[27]. The Ce³⁺ signal generated only two peaks, labeled u1 and v1, because of the 3d¹⁰4f¹ orbital. In all samples, the peak intensities assigned to Ce⁴⁺ were evidently stronger than those arising from Ce³⁺, demonstrating that Ce⁴⁺ was the primary chemical state. Table 2 shows that Cat1 had the lowest Ce⁴⁺ 3d_3/2 (u) and Ce⁴⁺ 3d_5/2 (v) BE values of 900.62 and 882.37 eV, indicating that the electron cloud densities associated with Ce⁴⁺ 3d_3/2 (u) and Ce⁴⁺ 3d_5/2 (v) were lower in Cat2 and Cat3, such that the Ce⁴⁺ ions were more readily reduced to Ce³⁺. In the case of the Ce³⁺ 3d_3/2 (u1) and Ce³⁺ 3d_5/2 (v1) values, Cat3 had the lowest BE values of 903.39 and 885.20 eV, respectively, indicating that the Ce³⁺ 3d_3/2 and Ce³⁺ 3d_5/2 electron cloud densities were the highest in this sample, thus the Ce³⁺ more readily lost electrons to oxidize to Ce^{4+ [28, 29]},increasing the mobility of active oxygen between Ce³⁺ and Ce⁴⁺. This reaction may have proceeded according to the following equation:

Ce₂O₃ + 1/2 O₂ → 2 CeO₂ (2)

Fig. 2. XPS spectra showing Ce 3d peaks for the support and catalysts.

The Ce³⁺/Ce ratios in these samples were determined by dividing the sum of the areas of the Ce³⁺ peaks by the sum of the areas of all cerium peaks, and the results are given in Table 2. These data show that the Cat3 had the highest Ce³⁺/Ce ratio, suggesting the presence of a significant quantity of oxygen vacancies and Ce in the Ce³⁺ state. The low valence state of the Ce species, as well as a high concentration of oxygen vacancies and a charge imbalance on the catalyst surface, which favors the redox transformation between Ce³⁺ and Ce⁴⁺, all work to increase both oxygen activation and mobility. From these results, it is evident that the addition of Zr⁴⁺ to the ceria-zirconia solid solutions created both a charge imbalance and defects on the catalyst surface, promoting the formation of oxygen vacancies and increasing the mobility of active oxygen between Ce³⁺ and Ce⁴⁺. These findings are in agreement both with the present XRD results and with data presented in the literature ^[30].

Table 2 The binding energies of the characteristic peaks of Ce atoms and Ce3+/Ce ratios. Sample Binding energy (eV) Ce3+/Ce (%) Ce3+ 3d3/2 (u1) Ce4+ 3d3/2 (u) Ce3+ 3d5/2 (v1) Ce4+ 3d5/2 (v) Support 904.21 900.93 885.60 882.23 10.76 Cat1 904.24 900.62 886.10 882.37 10.68 Cat2 904.52 900.80 885.35 882.43 10.93 Cat3 903.39 901.05 885.20 882.48 15.94

Table 2 The binding energies of the characteristic peaks of Ce atoms and Ce3+/Ce ratios.

The surface atomic concentrations of Ce, Zr, Al, O, La, and Ba are summarized in Table 3. These data show that the absolute percentages of Ce, Zr, Al, La, and O were slightly decreased in the Cat1 and Cat2 samples compared with the original support, since the Pt and Rh species would have occupied some surface sites. In addition, the Zr, La, and Ba concentrations in the Cat3 increased very little in comparison with those in the support; the loading of La³⁺, Zr⁴⁺, and Ba²⁺ into Cat3, which occupied some surface sites, led to only slight decreases in the atomic percentages of Ce and O. These results agree with the Ce 3d peaks intensities obtained for the catalysts as presented in Fig. 2.

Table 3 Surface elemental compositions and atomic ratios of the support and catalysts as determined by XPS. Sample Surface atomic concentration (%) Ce Zr Al O La Ba Support 4.54 6.72 28.02 59.61 1.11 — Cat1 2.73 5.19 30.29 59.21 0.91 — Cat2 3.35 5.87 29.35 58.69 1.06 — Cat3 2.49 6.92 28.84 58.20 1.14 1.18

Table 3 Surface elemental compositions and atomic ratios of the support and catalysts as determined by XPS.

Rh was not detected owing to its low concentrations in the prepared samples. To further study the chemical states of Pt and Rh in these catalysts, the concentration of Rh was increased in newly prepared Pt, Rh and Pt-Rh bimetallic catalysts, using the same synthesis procedure noted above. The Pt and Rh catalysts were prepared by impregnation method (2 wt% Pt and 2 wt% Rh, labeled as Pt and Rh, respectively). The bimetallic Pt-Rh catalysts (1 wt% Pt and 1 wt% Rh) were prepared using the same co-impregnation procedure employed to produce Cat1 (labeled as C-Pt+Rh) and by physically mixing Pt and Rh catalyst powders in the same manner as for Cat2 (labeled as M-Pt+Rh). The Rh 3d and Pt 4d_5/2 peaks are shown in Fig. 3. A comparison of the spectra of M-Pt+Rh and C-Pt+Rh with the linear superposition spectra of Pt and Rh (Fig. 3) shows that the M-Pt+Rh spectrum is completely coincident with the linear superposition spectra of Pt and Rh, while the Pt 4d_5/2 peak for C-Pt+Rh is increased to 316.40 eV. These results indicate that the bimetallic Pt-Rh catalyst prepared by physically mixing Pt and Rh monometallic catalysts had a catalyst surface structure equivalent to a homogeneous mixture of Pt and Rh sites. On the contrary, there was a strong interaction between Pt and Rh in the material produced by co-impregnation, as shown by the Pt 4d_5/2 transition.

Fig. 3. XPS spectra of Rh 3d and Pt 4d5/2 peaks. Pt and Rh (a) and M-Pt+Rh and C-Pt+Rh (b) compared with the linear superposition spectrum of Pt and Rh. Solid and dotted lines represent data from the catalyst and the linear superposition spectra of Pt and Rh.

Furthermore, as is evident from Fig. 3, all samples exhibited peaks around 312.12 eV, attributed to Y 3p_1/2. Peaks at 309.50 eV due to Rh 3d_5/2 and 315.20 eV due to Pt 4d_5/2 are seen in the case of the M-Pt+Rh sample, whereas the Pt 4d_5/2 peak is shifted to 316.4 eV for the C-Pt+Rh sample. In the C-Pt+Rh and M-Pt+Rh catalyst spectra, the Rh 3d, Pt 4d_5/2 and Y 3p_1/2 peaks overlap, although the individual peaks can be obtained by peak fitting. From this fitting procedure, the peak positions and relative areas as well as the distance between the Rh 3d_5/2 and Rh 3d_3/2 peaks can be determined, and the results are presented in Fig. 4. Here the Rh 3d_5/2 and Rh 3d_3/2 BE values for C-Pt+Rh and M-Pt+Rh are seen to be 309.50 and 314.30 eV. The Y 3p_1/2 peaks of the two samples remain at the same position of about 312.12 eV. For the M-Pt+Rh catalyst, the Pt 4d_5/2 peak appears at 315.20 eV, while the same peak for C-Pt+Rh is shifted to 316.40 eV, in agreement with the results shown in Fig. 3. The surface Rh/(Rh+Pt) and Pt/(Rh+Pt) atomic ratios were calculated and are found in Table 4. The results show that the Pt/(Rh+Pt) value for C-Pt+Rh was higher than that for M-Pt+Rh, possibly because the co-impregnation sample surface was enriched in terms Pt, which blocked some Rh active sites even though the initial concentration ratio of Pt to Rh was 1:1. It is worth noting that the Pt-Rh active sites in M-Pt+Rh were not chemically activated and that the Pt and Rh exhibited close-coupled physical contact. Both Pt and Rh were thus active components in the catalyst. However, there was a strong interaction between Pt and Rh in the C-Pt+Rh, and so Pt likely migrated to the catalyst surface and blocked some Rh sites.

Fig. 4. Peak fitting of the Rh 3d, Y 3p1/2 and Pt 4d5/2 XPS spectra of the M-Pt+Rh and C-Pt+Rh samples.

Table 4 Surface atomic ratios of noble metals for the C-Pt+Rh and M-Pt+Rh samples. Sample Surface atomic ratios (%) Rh/(Rh+Pt) Pt/(Rh+Pt) C-Pt+Rh 0.15 0.85 M-Pt+Rh 0.25 0.75

Table 4 Surface atomic ratios of noble metals for the C-Pt+Rh and M-Pt+Rh samples.

The redox properties of all samples were determined by H₂-TPR and the results are summarized in Fig. 5. The support presented a main peak at 595 °C with a shoulder at the low temperature side, attributed to the reduction of the ceria-zirconia solid solution. The TPR profiles of pure CeO₂ typically show two peaks at about 500 and 850 °C, attributed to surface and bulk reduction, respectively ^[31]. In this work, the TPR profile of the composite oxide exhibited a single, broad main peak between 385 and 660 °C. This was the result of the increased mobility of bulk oxygen in ceria-zirconia solid solutions after the introduction of Zr⁴⁺ into the ceria lattice. This result was in good agreement with previous reports ^{[32, 33]}.

Fig. 5. H2-TPR profiles of support and catalysts.

The TPR profiles of the Pt-Rh/CeZrYLa+LA catalysts were obviously different from that of the support. The peak (α) below 200 °C was dominant for each of the catalysts, while the peaks (β and γ) at approximately 240 and 400 °C were weak. As reported in the Ref. ^[34], all Pt species can be reduced to Pt metal below 500 °C, with the main reduction peak appearing at 180 °C. The reduction of Rh and Pt is seen below 200 °C, especially in the case of Rh which appears on the lower temperature side ^[7]. It is evident that the peak area ratio (α/β) was far greater than the Rh/Pt (1:9) ratio of the as-prepared catalyst, suggesting that the first peak (α) was because of the reduction of Rh species, and to the reduction of overlapped Pt and Rh species, as well as surface and some subsurface Ce⁴⁺ species. This resulted from the addition of noble metals that effectively promoted the reduction ability of the ceria-zirconia support, activating H₂ and then spilling it over onto the support ^[30]. The second peaks (β) can be assigned to the presence of small Pt particles strongly interacting with the support. The third peaks (γ) were very weak at about 400 °C, because of the removal of oxygen from a small proportion of the unpromoted subsurface ceria-zirconia powders. As shown in Fig. 5, the order of the reduction temperatures of α and β peaks was Cat3 < Cat2 < Cat1. The catalysts generated by physically mixing Pt and Rh showed better reducibility than the catalyst co-impregnated with Pt and Rh. This difference may result from the interaction between Pt and Rh during co- impregnation, which modifies the nature of the active surface. The XPS results above demonstrate that there was significant interaction between the Pt and Rh in the co-impregnation Pt-Rh catalyst, and that the catalyst surface was enriched in Pt which, in turn, blocked some Rh sites. It has been reported ^[35] that Pt enriched-bimetallic Pt-Rh particles exhibit large particle sizes and a high reduction temperature. The profile of the additive-promoted Cat3 showed the lowest reduction temperature, indicating that the addition of the additives La³⁺, Ba²⁺ and Zr⁴⁺ was able to facilitate the reduction of precious metals species. This could have occurred because these additives dispersed the precious metals to form a discontinuous phase, perhaps reducing the interaction between the noble metals and the support.

The catalytic activities for CH₄ and CO oxidation and NO reduction over all catalysts under a simulated exhaust gas with a GHSV of 34000 h⁻¹ are depicted in Fig. 6. It is evident that the order of activities for CH₄, CO, and NO conversion was Cat3 ≈ Cat2 > Cat1. The conversions of CH₄ and CO are also seen to have increased continuously with temperature, while the NO conversions over Cat2 and Cat3 increased as temperature increased but varied in an irregular manner over Cat1. In the range from 150 to 390 °C, CH₄ did not react, while the NO reactivity initially increased before plateauing at about 250 °C and then decreasing with further increases in temperature. Because the CH₄ did not react, the NO only reacted with the CO. In the range from 390 to 500 °C, the NO conversion coincided with the CH₄ conversion and increased rapidly with increased temperature. It has been reported ^[15] that NO conversion can be controlled via its reaction with CO at low temperatures but proceeds by reaction with CH₄ at high temperatures, which is in agreement with our results. The main reactions associated with this process are as follows:

CO + 1/2O₂ → CO₂ (3)

CO + NO → 1/2N₂ + CO₂ (4)

CH₄ + 2O₂ → CO₂ + 2H₂O (5)

CH₄ + 4NO → 2N₂ + 2H₂O + CO₂ (6)

Fig. 6. Conversions of CH4 (a), CO (b), and NO (c) as functions of temperature over different catalysts.

In the case of Cat1, Reactions (3) and (4) are predominant from 150 to 250 °C. When the temperature increases from 250 to 390 °C, the CO + O₂ reaction (Reaction 3) proceeds more quickly than the NO + CO reaction (Reaction 4), leading to a decrease in the NO conversion over Cat1 in this range. With further increases in temperature, CH₄ takes part in the reaction and NO conversion proceeds via Reaction (6) and increases rapidly in the range from 390 to 500 °C. NO conversions over Cat2 and Cat3 were superior to that over Cat1, possibly because of Pt enrichment of the Pt-Rh active sites on Cat1, blocking the high efficiency of Rh for the reduction of NO. The XPS results might also be relevant here, in which Cat3 and Cat2 showed greater quantities of oxygen vacancies and Ce³⁺, as well as a higher mobility of active oxygen between Ce³⁺ and Ce⁴⁺ than Cat1. Thus, the conversion of NO benefited from the Ce³⁺/Ce⁴⁺ redox pair and was also related to the stability of the surface oxygen vacancies ^[36], such that the NO conversion occurred at the expense of the Ce³⁺/Ce⁴⁺ redox couple. Furthermore, both CH₄ and CO conversion over Cat2 and Cat3 were higher than over Cat1. Therefore, the catalytic activity of the catalyst made by physically mixing Pt and Rh was superior to that of the catalyst co-impregnated with Pt and Rh.

T₅₀, T₉₀, and ΔT (T₉₀-T₅₀) data are presented in Table 5. Here, the light-off temperature (T₅₀) and complete conversion temperature (T₉₀) are the temperatures at which the conversion of a given pollutant reached 50% and 90%, respectively, and ΔT is the temperature range between T₅₀ to T₉₀. Generally, T₅₀ and T₉₀ are used to evaluate the catalytic activity during exhaust gas purification. It can be seen that Cat3 had the lowest T₅₀ for CO (114 °C) and NO (149 °C), and the lowest T₉₀ for CH₄ (398 °C) and CO (179 °C). The ΔT values for Cat3 were 34 °C for CH₄ and 65 °C for CO; these values were much smaller than those of Cat2 and Cat1. The small ΔT values suggest that each pollutant reached complete conversion almost immediately after light-off, indicating that the catalyst exhibited remarkable temperature properties. Cat2 showed the lowest T₅₀ for CH₄, the lowest T₉₀ for NO and the lowest ΔT for NO, while exhibiting slightly lower catalytic activity than Cat3. Cat1 had the highest T₅₀ and T₉₀ and the largest ΔT out of all three catalysts.

Table 5 Light-off temperatures and complete conversion temperatures for CH4, NO and CO conversion over different catalysts. Sample T50 (°C) T90 (°C) △T (T90-T50) (°C) CH4 CO NO CH4 CO NO CH4 CO NO Cat1 440 153 414 >500 361 451 >60 208 37 Cat2 342 117 157 405 269 174 63 152 17 Cat3 364 114 149 398 179 192 34 65 43

Table 5 Light-off temperatures and complete conversion temperatures for CH4, NO and CO conversion over different catalysts.

CH₄, CO and NO conversions over the three catalysts at 500 °C under different O₂ concentrations are illustrated in Fig. 7. For each catalyst, all three pollutants showed the best conversion at an O₂ level near 0.22 vol%. Under oxygen-poor condition (O₂ < 0.22%), NO could be 100% converted using Cat2 and Cat3, with above 94% conversion for Cat1. CH₄ conversions were between 76% and 86% for Cat1, 67% to 92% for Cat2 and reached 100% for Cat3 under oxygen-poor condition. Cat3 showed excellent activity for CH₄ and NO conversions under rich burning conditions because of its higher concentration of oxygen vacancies and Ce³⁺/Ce⁴⁺ redox pairs ^[36]. Under oxygen-rich conditions (O₂ > 0.22%), NO conversions over all catalysts rapidly decreased with increases in the O₂ concentration. The CH₄ conversions over Cat1 and Cat2 followed the same trend as the NO conversion, while the CH₄ conversion over Cat3 decreased slightly, indicating that the CH₄ conversion over Cat3 was higher than over Cat1 or Cat2. In each O₂ working-window, CO conversions over all catalysts increased as the O₂ concentration increased, and Cat3 exhibited the highest CO conversion.

Fig. 7. Variations of the CH4, CO, and NO conversions with the O2 concentration over different catalysts at 500 °C. (a) Cat1; (b) Cat2; (c) Cat3.

From the above discussion, it can be concluded that the catalytic activity and working-window of Cat3 were superior to those of the other catalysts. In the case of thosecatalysts made by physically mixing Pt and Rh (Cat2 and Cat3), thus preventing the generation of Pt-enriched large bimetallic Pt-Rh particles, both Pt and Rh functioned as active components and each made a significant contribution to CH₄/CO/NO conversions. Conversely, the incorporation of additives (La³⁺, Zr⁴⁺, and Ba²⁺) into Cat3 meant that the La³⁺ ions could enhance the dispersion of active components and active sites, thus improving the catalytic activity ^[19]. The additives might also work to disperse the precious metals to form a discontinuous phase, reducing the interactions between the noble metals and the support, thus improving the reducibility of the catalyst, as seen in the H₂-TPR results. Moreover, the insertion of Zr⁴⁺ ions into the ceria-zirconia lattice created a higher concentration of lattice defects and thus improved the O₂-mobility ^{[24, 26]}. The presence of monoclinic ZrO₂ in Cat3 is also important; it has been reported that monoclinic ZrO₂ shows higher activity for CH₄ oxidation ^{[37, 38]}. In addition, Cat3 had the highest Ce³⁺/Ce ratio, which also assisted in the conversion of NO ^[36]. In this work, Cat3 exhibited remarkable catalytic performance, and therefore has potential applications in stoichiometric NGV exhaust purification.