Oxygen Gateway Effect of CeO2/La2O2SO4 Composite Oxygen Storage Materials.

A synergistic enhancement in oxygen release/storage performance was achieved with composites formed between CeO2 as an oxygen gateway and La2O2SO4 as an oxygen reservoir. CeO2 smoothly transfers oxygen atoms between La2O2SO4 and the gas phase, whereas La2O2SO4 stores a large amount of oxygen. The composite materials exhibited enhanced anaerobic CO oxidation and reversible oxygen storage in the presence of impregnated Pt catalysts as compared to their individual constituents (Pt/CeO2 and Pt/La2O2SO4). In situ X-ray diffraction and Raman experiments demonstrated that CeO2 significantly accelerated the redox reaction between La2O2SO4 (S6+) and La2O2S (S2-), while preserving its structure. The reaction between CO and CeO2/18O-labeled La2O2SO4 composites suggested that CO mainly reacted with the lattice oxygen atoms of CeO2, and the resulting oxygen vacancies were subsequently filled with oxygen atoms supplied by La2O2SO4. This oxygen gateway effect of CeO2 greatly enhanced the oxygen release/storage rates of La2O2SO4, while maintaining the high oxygen storage capacity, which is an advanced feature of oxysulfate materials. The synergistic effect is mostly pronounced when the two different oxygen storage materials are in intimate contact to form a three-phase boundary.

Introduction

Oxygen storage materials are important in current automotive emission control catalysts.[1−7] These materials function as oxygen storage or releasing materials in autoexhausts to achieve the ideal air-to-fuel ratio required for complete conversion of noxious pollutants, including NO, CO, and hydrocarbons over noble metal catalysts (Pt, Rh, and/or Pd). Cerium-based binary oxides have been most widely used for this purpose and CeO2–ZrO2 is most widely used in practical applications.[8−20] The redox reaction between Ce4+ and Ce3+ enables fast oxygen transfer but limits the total oxygen storage capacity (OSC) to less than 0.25 mol-O2 mol–1. On the other hand, we reported an 8-fold increase in OSC values (2 mol-O2 mol–1) using lanthanum oxysulfates (La2O2SO4), which utilize sulfur as the redox center instead of metallic cations according to the following reaction:[21−27] La2O2SO4 ↔ La2O2S + 2O2. Their potential use has been studied in other applications such as water–gas shift reaction catalysts,[28] solid oxide fuel cells,[29] and batteries.[30] Even though oxysulfates exhibit the largest reported OSC values, they have drawbacks owing to their lower oxygen release rates and consequently higher operation temperatures are required (≥600 °C) than those needed for CeO2–ZrO2 (300–400 °C). To overcome this, much research efforts have been directed toward the microstructural and chemical modifications of materials by means of methods such as soft-chemical synthesis using a surfactant-templating method,[23] impregnation of noble metals such as Pt and Pd,[25] complete replacement of La by Pr (Pr2O2SO4),[24] and partial replacement of La by Ce ((La1–Ce)2O2SO4).[26] However, further modifications are still needed to address the fundamental issue for enabling lower operation temperatures, which may extend possible applications of the oxysulfate materials.

Recently, a new type of oxygen storage material, composites comprising CeO2 and other metal oxides, has been proposed.[31−36] A typical example is CeO2-grafted Fe2O3, which comprises several CeO2 grains intimately bound on the surface of Fe2O3 via Ce–O–Fe interfacial linkage.[36] Owing to its local structure, oxygen release and storage rates superior to those of CeO2 and an OSC greater than that of Fe2O3 could be simultaneously achieved. CeO2 plays the role of an oxygen gateway by accelerating the conversion between O2 and oxide ions and the transfer of oxide ions to/from Fe2O3 acting as an oxygen reservoir. CeO2 is therefore considered as an efficient catalyst for enhancing oxygen releasing and storage properties in other materials. This concept can be extended for designing new composite oxygen storage materials consisting of CeO2 (as an oxygen gateway) and other oxygen reservoirs. La2O2SO4 can be a good candidate for this purpose as it can serve as an oxygen reservoir owing to its higher OSC and lower oxygen release/storage rates than those of CeO2.

In the present study, novel composites consisting of two oxygen storage materials CeO2 and La2O2SO4 are prepared using the wet-impregnation method to examine the oxygen gateway effect of CeO2. Oxygen release/storage performance is evaluated using anaerobic CO oxidation techniques under cycled feed stream conditions at various temperatures. Because the composite materials exhibit prominent synergistic effects in terms of the oxygen release/storage rate and capacity, the relationship between the structure and properties of the materials was studied using in situ X-ray diffraction (XRD) and in situ Raman spectroscopy under the performed reaction conditions. A possible mechanism for the oxygen gateway effect of CeO2 is discussed on the basis of the results obtained with anaerobic CO oxidation using isotope-labeled La2O2SO4.

Results and Discussion

Structure of CeO2/La2O2SO4 Composites

The XRD patterns of as-prepared CeO2/La2O2SO4 composites exhibited much less intense peaks of CeO2 even with the greatest loading of CeO2 (40 wt %) as compared to those obtained for reference samples consisting of physical mixtures of CeO2 and La2O2SO4 (Figure S1). Therefore, the phases present in the composite materials were analyzed using Raman spectroscopy. Figure shows the Raman spectra of La2O2SO4 and CeO2/La2O2SO4 (20 wt % CeO2) in the region of four fundamental vibration modes corresponding to the SO42– unit: the nondegenerate symmetric stretching mode (ν1, 990 cm–1), the symmetric bending mode of SO4 (ν2, 420 cm–1), the triply degenerate asymmetric stretching mode (ν3, 1060–1180 cm–1), and the triply degenerate asymmetric bending mode of SO4 (ν4, 595–655 cm–1).[37−39] Other bands observed in the 250–450 cm–1 region are assigned to the La–O fundamental modes.[40] These bands are observed in the CeO2/La2O2SO4 composite, and a strong band appears at approximately 460 cm–1, which is assigned to the characteristic F2g mode of CeO2 corresponding to the symmetrical stretching mode of the CeO8 vibrational unit in the cubic fluorite structure.[41,42] The Ce K edge extended X-ray absorption fine structure (EXAFS) (Figure S2 and Table S1) analysis suggested that Ce in the as-prepared CeO2/La2O2SO4 composites was present in a local environment similar to that in CeO2; however, particles sizes of CeO2 were smaller in the composite than those in the individual sample. The structural characterization results can be rationalized by considering the presence of highly dispersed CeO2 in the composite samples. The Brunauer–Emmett–Teller (BET) surface area of the as-prepared CeO2/La2O2SO4 composites was ∼10 m2 g–1 regardless of CeO2 loading (5–40 wt %).

Figure 1

Raman spectra of as-prepared (a) La2O2SO4 and (b) CeO2/La2O2SO4 (20 wt % CeO2).

Oxygen Release/Storage of Pt/CeO2/La2O2SO4 Composites

The oxygen release/storage performance of Pt-loaded composite materials was evaluated on the basis of the CO/O2 cycle reactions and by comparing the results with those obtained with the individual constituents (Pt/CeO2 and Pt/La2O2SO4). Figure illustrates the typical gas concentration profiles obtained at the inlet and outlet of the flow reactor when the two gas feeds, 1% CO/He (10 min) and 0.5% O2/He (20 min), are alternately switched. Over Pt/CeO2 at 500 °C (Figure a), the O2-to-CO switch results in the appearance of a temporal CO2 peak along with the simultaneous consumption of CO, suggesting CO oxidation by the lattice oxygen atoms of CeO2. The concentration of CO is then gradually increased to the initial value after 10 min. Subsequent CO-to-O2 switch results in O2 uptake owing to the oxidation of partially reduced CeO2 and the appearance of another temporal CO2 peak owing to the desorption of carbonate species formed on the surface of CeO2. The cumulative amount of O2 uptake is equivalent to the amount of CO consumed in anaerobic oxidation, corresponding to an apparent OSC of 0.134 mmol-O2 g–1. The stoichiometric CO oxidation and oxygen storage can be stably cycled at temperatures of 400–700 °C. When Pt-unloaded sample was used, very small CO conversion rates and OSC values were obtained, indicating that Pt provides a number of active sites for CO oxidation and O2 dissociation and accelerates oxygen release/storage. When the same cycle reaction is performed with Pt/La2O2SO4, a smaller conversion from CO to CO2 takes place after the O2-to-CO switch (Figure b). The obtained OSC (0.030 mmol-O2 g–1) is smaller than that of Pt/CeO2, as higher temperatures (≥600 °C) are required for achieving fast oxygen release/storage cycles of La2O2SO4.[25,26] Nevertheless, a much higher OSC is obtained for the Pt/CeO2/La2O2SO4 composite (20 wt % CeO2) as shown in Figure c. The initial CO conversion obtained at the O2-to-CO switch with the composite is higher than that of Pt/CeO2 and the conversion continues for a longer time. A subsequent CO-to-O2 switch results in O2 storage because of the O2 breakthrough time of more than 5 min. Accordingly, OSC calculated from the oxygen breakthrough curve exhibits a much larger value (0.242 mmol-O2 g–1), which is a more than 8-fold greater value than that obtained for Pt/La2O2SO4. According to previous studies,[21,22] oxygen release/storage reactions for La2O2SO4 can be written as follows, and the maximum OSC should be 4.92 mmol-O2 g–1 that corresponds to 2 mol-O2 (mol-S)−1.It should be noted that the observed OSC values are significantly lower than the maximum achievable OSC because the anaerobic CO oxidation in Figure is not completed within 10 min of the CO supply.

Figure 2

Influent and effluent gas profiles measured during CO/O2 cycle reactions over (a) Pt/CeO2, (b) Pt/La2O2SO4, and (c) Pt/CeO2/La2O2SO4 (40 wt % CeO2) under cycled feed stream conditions of 0.5% O2 or 1% CO at 500 °C. W/F = 4 × 10–3 g min cm–3.

From the reaction profiles obtained at different temperatures, the OSC values (OSCobs) of Pt/CeO2/La2O2SO4 are calculated (Figure ). The weighted sum calculated from the weight fraction and OSC of individual constituents (Pt/CeO2 and Pt/La2O2SO4) is also given (denoted as OSCcalc). The OSCcalc values obtained at the lowest temperature of 400 °C might imply a larger contribution from Pt/CeO2, in contrast to the negligible contribution from Pt/La2O2SO4. However, Pt/La2O2SO4 exhibits significantly enhanced OSC values at temperatures equal to or higher than 600 °C, whereas the OSC of Pt/CeO2 is found to be less dependent on the temperature. Notably, Pt/CeO2/La2O2SO4 exhibits much higher OSCobs values than OSCcalc values (obtained at any temperature). This shows that the composite exhibits higher OSCs than those obtained with the individual constituents, clearly indicating the synergistic effect achieved by combining two oxygen storage materials exhibiting different characteristics. The difference between the OSCobs and OSCcalc values is more obvious at low reaction temperatures and high CeO2 loading. The enhancement in OSC is found to depend on the microstructure of the composites as indicated in Figure . As compared to the physical mixture of Pt/CeO2 and Pt/La2O2SO4, the two composites, prepared by impregnating Pt onto La2O2SO4 followed by Ce impregnation or by impregnating Ce onto La2O2SO4 followed by Pt impregnation, exhibit more than 2-fold greater OSC values even though the samples have the same chemical composition. This demonstrates that the extent of interface contact between CeO2 and La2O2SO4 plays a key role in enhancing oxygen release/storage performance.

Figure 3

OSCobs and OSCcalc values for Pt/CeO2/La2O2SO4 at different temperatures (5–40 wt % CeO2). OSCcalc values calculated from OSCobs values for Pt/CeO2 and Pt/La2O2SO4.

Figure 4

OSC values for CeO2/Pt/La2O2SO4 and Pt/CeO2/La2O2SO4 composites and a physical mixture at 500 °C (20 wt % CeO2).

The maximum OSC obtained was determined using a flow microbalance when reducing (1.4% H2/He) and oxidizing (0.7% O2/He) gas feeds were cycled at 600 °C (Figure ). Pt/CeO2 exhibits small weight oscillations (OSC < 0.1 mmol-O2 g–1) as oxygen release/storage is limited near the surface region. Pt/La2O2SO4 exhibits larger weight changes (0.45 mmol-O2 g–1), close to 10% of the theoretical OSC estimated from stoichiometric reactions (eqs and 2), which can be attributed to the slow reaction rates obtained at this temperature. In contrast, Pt/CeO2/La2O2SO4 (20 wt % CeO2) exhibits fast and large weight changes (3.12 mmol-O2 g–1) corresponding to >70% of the stoichiometric reactions. The oxygen release and storage rates are estimated from the initial slope of the weight change shown in Figure (Table ). The oxygen release rates obtained with the composite are more than 9-fold faster than those of La2O2SO4. Similarly, the composite exhibits a faster oxygen storage rate (>6-fold) than La2O2SO4. Both samples exhibit higher oxygen storage rates than oxygen release rates. In Figure , the weight oscillation curves a and b show a slow decay with time, whereas the S/La molar ratio remained unchanged during the oxygen release/storage cycles. A possible reason for the decay is because of the carbonaceous residue removed under O2 atmosphere, which originates from an organic templating molecule (SDS).

Figure 5

Weight change observed during oxygen release/storage cycles over (a) Pt/CeO2, (b) Pt/La2O2SO4, and (c) Pt/CeO2/La2O2SO4 (20 wt % CeO2) at 600 °C under switched feed streams of 1.4% H2/He or 0.7% O2/He.

Table 1

Oxygen Release and Storage Rates Determined from Figure

	Pt/La2O2SO4	Pt/CeO2/La2O2SO4
Oxygen release (mol-O2 g–1 min–1)	0.122 × 10–4	1.10 × 10–4
Oxygen storage (mol-O2 g–1 min–1)	0.797 × 10–4	4.82 × 10–4

Structural Change during Oxygen Release/Storage

Phase changes during oxygen release obtained in a flow of 5% H2/N2 and subsequent oxygen storage in 5% O2/N2 were analyzed using in situ XRD. Figure a shows a comparison of the XRD patterns of Pt/La2O2SO4 and Pt/CeO2/La2O2SO4 (20 wt % CeO2) recorded at 5 min intervals after starting the 5% H2/N2 gas feed at 600 °C. In the case of CeO2-unloaded Pt/La2O2SO4, the La2O2SO4 phase gradually disappears within the initial 70 min, accompanied by the simultaneous formation of La2O2S. In contrast, with the Pt/CeO2/La2O2SO4 composite, faster phase transformation from La2O2SO4 to La2O2S is obtained, which is almost complete in 30 min. According to the following oxygen storage performed at the same temperature in a stream of 5% O2/N2 (Figure b), the oxidation of La2O2S in both the samples is completed in 10 min; however, the composite exhibits a faster reaction than the La2O2SO4 sample. These results clearly demonstrate the promoting effect of CeO2 not only on the reduction of La2O2SO4 but also on the oxidation of La2O2S.

Figure 6

In situ XRD patterns of Pt/La2O2SO4 and Pt/CeO2/La2O2SO4 (20 wt % CeO2) during (a) reduction in 5% H2/N2 at 600 °C and (b) subsequent reoxidation in 5% O2/N2 at 600 °C. The XRD patterns were acquired every 5 min after starting each gas feed.

Because the structural change of CeO2 occurring during reduction/oxidation could not be detected with in situ XRD, in situ Raman spectroscopy was used to analyze the redox behavior of the composite. The Raman spectrum is surface-sensitive and is reflected by the structural changes of CeO2. Figure shows the Raman spectra recorded at 30 s intervals during cycled redox feed streams of 5% H2/N2 and 5% O2/N2 at 500 °C. During the reduction, bands corresponding to the ν1 (990 cm–1) and ν2 (420 cm–1) modes of [SO4] disappear (Figure a). Even though a slight weakening of the CeO2 F2g band at 440 cm–1 is observed, it remains even after the [SO4] bands disappear. As reported in our previous paper,[36] the CeO2 F2g band weakened under a H2 atmosphere at 500 °C because oxygen release from the CeO2 lattice caused an increase in the concentration of oxygen vacancies near the surface, resulting in the disordered distribution of oxygen vacancies. The F2g band is reversibly restored by the successive admission of O2, indicating the filling of oxygen vacancies with rapid O2 uptake. A similar trend is observed in the present system when reduction and oxidation take place. The oxidation step (Figure b) initially leads to the temporal appearance of a strong band at ∼380 cm–1, which cannot be assigned at this stage. However, it may be associated with the structural change of the [La2O2] unit occurring at the initial stage of oxidation from La2O2S to La2O2SO4 because a similar phenomenon is observed with CeO2-unloaded Pt/La2O2SO4. The F2g band (440 cm–1) intensifies before the ν1 band (990 cm–1) appears, indicating that the oxygen storage of CeO2 is completed before the oxidation of La2O2S begins. Thus, when loaded onto the La2O2SO4 surface, CeO2 in the composite is assumed to lose lattice oxygen atoms to some extent during oxygen release, while still preserving its structure. The oxygen deficiency should be limited near the surface regions because the Ce K edge EXAFS measurements suggest that the atomic distances and coordination numbers for Ce–O and Ce–O–Ce spheres are negligibly affected by the oxygen release occurring during anaerobic CO oxidation (Figure S2 and Table S1). The efficient effect of CeO2 on the redox reaction between La2O2SO4 and La2O2S should be associated with the oxygen transfer at the three-phase boundary among CeO2, La2O2SO4/La2O2S, and the gas phase.

Figure 7

In situ Raman spectra of Pt/CeO2/La2O2SO4 (20 wt % CeO2) recorded during (a) reduction in 5% H2/N2 at 500 °C and (b) subsequent reoxidation in 5% O2/N2 at 500 °C. The gas feed was switched from N2 to each gas at 0 s.

Possible Mechanism of Enhanced Oxygen Release/Storage

To elucidate the possible mechanism for enhanced oxygen release/storage, anaerobic CO oxidation was performed over 18O-labeled La2O2SO4, which was prepared by the reaction between La2O2S (Pt-loaded) and 18O2 at 600 °C. In a flow system, 18O2 pulse injection to a single phase La2O2S was repeated in a He stream (Figure S3). The occurrence of cumulative oxygen uptake corresponding to the stoichiometric oxidation to form La2O2SO4 was confirmed. The as-prepared 18O-labeled La2O2SO4 sample was characterized using Raman spectroscopy (Figure ). The bands corresponding to the four fundamental vibrations of the SO4 unit of La2O2SO4 broaden and shift to lower frequencies upon 18O-exchange (Figure a,b). To confirm that these changes are associated with isotopic shifts owing to the presence of SO4 units containing different numbers of 18O, the strongest singlet ν1 mode band is deconvoluted (Figure c). The band can be split into five peak components assigned to S16O4, S16O318O, S16O218O2, S16O18O3, and S18O4. By examining the integrated peak intensities obtained for each peak component, the isotopic fraction [18O]/([16O] + [18O]) of the SO4 unit was calculated to be ∼66%. Thus, the isotopic oxygen distribution in the sample can be expressed as La2(16O0.3218O0.68)2S(16O0.3418O0.66)4. Because the oxysulfate crystal structure is formed with alternately stacked layers of [La2O2]2+ and [SO4]2–, the similar isotopic fractions observed in each structure unit suggest the occurrence of a fast isotope scrambling in the solid phase. The presence of 18O in the [La2O2]2+ unit can be associated with isotopic shifts of the bands observed in the range of 250–450 cm–1, which are assigned to the La–O fundamental modes.

Figure 8

Raman spectra of (a) La2O2SO4 and (b) 18O-labeled La2O2SO4 showing four fundamental S–O vibration modes. (c) Fitting results obtained for the ν1 vibration mode of 18O-labeled La2O2SO4.

By impregnating Ce onto as-prepared 18O-labeled La2O2SO4, CeO2/Pt/La2O2SO4 composites were prepared to perform anaerobic CO oxidation using the pulse injection method at 600 °C. As shown in Figure , two composites prepared by using different sequences of impregnation of Ce and Pt onto La2O2SO4 exhibit similar OSC values. Therefore, this sample can be used for analyzing the anaerobic CO oxidation mechanism. Figure shows a plot of isotopic oxygen fraction in CO2 versus the number of CO pulses at 600 °C. When 18O-labeled La2O2SO4 (Pt-loaded) is used alone, the initial isotopic oxygen fraction in CO2 (37%) is close to half the aforementioned 18O fraction of the SO4 unit (66%). This is explained by considering the following reactionThe isotopic oxygen fraction in CO2 monotonically decreases with an increase in the number of CO pulses, as the oxygen scrambling between pulsed CO and the SO4 unit decreases the concentration of 18O in La2O2SO4. Notably, quite different behavior was observed with the CeO2/Pt/La2O2SO4 composites; the isotopic oxygen fraction in CO2 is initially less than 10%, which gradually increases with an increase in the number of CO pulses. This trend is more obvious at greater CeO2 loadings, although the isotopic oxygen fractions of these different samples are almost the same after 25 times of CO pulse injections. These results indicate that injected CO is oxidized by the lattice oxygen atoms of CeO2 and the resulting oxygen vacancies in CeO2 are filled with oxygen atoms diffused from the 18O-labeled La2O2SO4. With an increase in CO pulse injection numbers, the lattice oxygen of CeO2 is further replaced by oxygen atoms transferred from La2O2SO4, which is increasingly converted into La2O2S. We could not identify the presence of 18O in CeO2 from Raman spectra because the isotopic shift of the CeO2 F2g mode was small (less than 20 cm–1)[43] and the isotopic concentration in CeO2 was low.

Figure 9

Isotopic oxygen fractions in CO2 produced by CO reaction with samples containing 18O-labeled La2O2SO4 at 600 °C.

On the basis of the analysis results obtained and the enhancement in oxygen release/storage performance, possible reaction schemes for oxygen transfer at the three-phase boundary including the CeO2/La2O2SO4 interface are proposed (depicted in Figure ). In the present study, the 1 wt % Pt-supported samples were used for the oxygen release/storage experiments. Although Pt plays a key role in the activation of O2 and reducing agents (CO and H2), the present study focuses on the oxygen gateway effect of CeO2 on La2O2SO4. Under a reducing atmosphere, the surface oxygen of CeO2 (Os) is readily removed by a reducing agent, such as CO and H2, yielding an oxygen vacancy (VO) (i–ii). As the number of oxygen vacancies increases to a certain value, VO is immediately filled with oxygen atoms supplied by La2O2SO4 (iii). Because of the greater OSC values of La2O2SO4, fast oxygen release can occur (maximum value of 2 mol-O2 (mol-S)−1). As the O2 pressure in the gas phase increases, reverse oxygen transfer occurs. Thus, O2 dissociated on the CeO2 surface is smoothly transferred to oxidize La2O2S to La2O2SO4 (iv–vi). This mechanism might explain the enhancement in OSC and the reaction rate of the CeO2/La2O2SO4 composite. The role of CeO2 in the composite material is similar to that in the CeO2-grafted Fe2O3 examined in our previous study.[36] The combination of CeO2 (as an oxygen gateway) and Fe2O3 (as an oxygen reservoir) yields a noticeable synergistic effect on oxygen release and storage. Regardless of the limited OSC (<0.25 mol-O2 mol–1) that is attributed to the redox reaction between Ce3+ and Ce4+, the unique ability of CeO2 in promoting rapid oxygen transfer between another solid and gas phase can be widely extendable for producing various composite materials exhibiting oxygen release/storage phenomena. Finally, although this is beyond the scope of this paper, it is interesting to note that the precise control of spatial distribution of Pt, CeO2, and La2O2SO4 is essential for the oxygen release/storage performance because the extent of contacts between these three components influences the final redox response of the present ternary systems (Figure ). This issue should be especially important to reduce the amount of Pt loading and will be a common target for future studies.

Figure 10

Schematic for (i–iii) oxygen release and (iv–vi) oxygen storage mechanisms at the gas/CeO2/La2O2SO4 three-phase boundary. Pt is present on the entire solid surface (not shown).

Experimental Section

Sample Preparation

La2O2SO4 was synthesized via a template route as reported previously.[23] La(NO3)3 (99.9%, Mitsuwa Chemical), C12H25OSO3Na (SDS, 98.5%; Kishida Chemical), aqueous ammonia, and deionized water with a molar ratio of La/SDS/NH3/H2O = 1:2:30:60 were mixed at 40 °C for 1 h followed by stirring at 60 °C for more than 10 h. The mixture was then cooled to room temperature to obtain a precipitate at pH 11. The precipitate was collected by centrifugal separation, washed thoroughly with ion-exchanged distilled water, dried by evacuation at room temperature, and finally heated at 600 °C for 5 h in air to obtain La2O2SO4. CeO2/La2O2SO4 composites (5–40 wt % loading as CeO2) were prepared by impregnating an aqueous solution of Ce(NO3)3 (99.9%, Mitsuwa Chemical) onto as-prepared La2O2SO4 followed by calcination at 500 °C for 2 h in air. The as-calcined samples were impregnated with an aqueous solution of H2PtCl6 and then calcined at 500 °C for 5 h to produce Pt-loaded samples (1 wt % loading as Pt). Another type of composite material was prepared by means of the reverse sequence of impregnation, in which Pt was first supported on La2O2SO4 followed by impregnation of Ce. The as-prepared composite materials thus obtained are denoted as Pt/CeO2/La2O2SO4 and CeO2/Pt/La2O2SO4. In addition, a physical mixture of Pt/CeO2 and Pt/La2O2SO4 was prepared for use as a reference sample.

Isotopic oxygen (18O)-labeled La2O2SO4 was prepared by the direct reaction between La2O2S and 18O2. Initially, Pt-loaded La2O2SO4 (1 wt % Pt) was completely reduced to La2O2S in flowing H2 at 700 °C. After evacuation, the product was placed in He flow at 600 °C, and a pulse injection of 18O2 (1.63 mL) into the gas feed was repeated until oxygen uptake was complete. The isotopic oxygen concentration in the effluent gas was measured using a quadrupole mass spectrometer (Omnistar; Pfeiffer) to calculate the isotopic fraction of obtained La2O2SO4. A composite material CeO2/Pt/La2O2SO4 (20 wt % CeO2) was prepared by impregnating Ce(NO3)3 onto 18O-labeled La2O2SO4 as described above.

Characterization

The BET surface area was calculated from the N2 adsorption isotherms measured at −196 °C (Belsorp-mini; Microtrac-Bel). Energy-dispersive X-ray fluorescence analysis (MESA-500 W; Horiba) was used to determine the loading of CeO2, the S/La molar ratio, and the presence of possible impurities for as-calcined samples. The concentration of residual chlorine associated with the impregnated H2PtCl6 was below the detection limit. Powder XRD measurements were carried out with a Rigaku Multiflex diffractometer using monochromated Cu Kα radiation (40 kV, 20 mA). The structural changes observed during the reduction and reoxidation stages were measured by means of in situ XRD using a Rigaku RINT Ultima diffractometer (Cu Kα, 40 kV, 30 mA) equipped with a high-speed two-dimensional detector D/teX-25. XRD patterns of the sample placed in a stream of 5% H2 or 5% O2 balanced with N2 (100 cm3 min–1) in a temperature-controllable chamber were measured at constant intervals of 5 min with a scan rate of 40° min–1. The high-speed detector enabled rapid scanning of each XRD pattern within 60 s, which is short enough to neglect the phase changes occurring during data acquisition.

The Ce K edge EXAFS spectrum was recorded with an NW10A instrument of the Photon Factory Advanced Ring for Pulse X-rays, High Energy Accelerator Research Organization at Tsukuba using a ring energy of 6.5 GeV and a stored current of around 35–60 mA. A Si(311) double-crystal monochromator was used, and the spectra were recorded at room temperature in the transmission mode. The EXAFS data were processed using a REX 2000 program (Rigaku). The EXAFS oscillation was extracted by fitting a cubic spline function through the post-edge region. The k3-weighted EXAFS oscillation in the 3.0–14.0 Å–1 range was Fourier-transformed. Phase shifts and backscattering amplitudes of Ce–O and Ce–O–Ce were obtained from the EXAFS data of CeO2.

Raman spectra were obtained with a Jasco NRS-3100 spectrometer using a 532.1 nm laser as the excitation source. In situ Raman scattering measurements were performed with a Horiba Jobin Yvon LabRAM HR Evolution spectrometer using a 457 nm laser excitation source. The spectrometer was connected to an infrared image heating stage (MS-TPS; Yonekura) with a gas-flow system to perform high-temperature measurements under controlled gas environments. The powder catalysts were heated at 500 °C under 5% O2/N2 flow to remove adsorbed gases. This procedure was followed by flushing the samples with N2 for 30 s and subsequently reducing them in 5% H2/N2 for 10 min. At the end of the first reduction step, N2 was flushed and 5% O2/N2 was subsequently passed for 10 min. During the two sets of reduction/oxidation cycles, Raman spectra were recorded every 30 s.

Oxygen Release/Storage Measurement

The oxygen release/storage performance of the Pt-loaded samples was determined by anaerobic CO oxidation under cycled feed stream conditions at constant reaction temperatures (400–700 °C) in a dual-supply flow system. 1% CO/He and 0.5% O2/He gas feeds were alternately switched at programmed time intervals (10 min for CO and 20 min for O2). The gas feed rate (F = 50 cm3 min–1) to the sample (W = 0.20 g) was controlled at W/F = 4.0 × 10–3 g min cm–3. The concentrations of each gas component (CO, CO2, and O2) before and after the catalyst bed were recorded using a quadrupole mass spectrometer (Omnistar; Pfeiffer). Pulse-mode CO oxidation was performed over CeO2/Pt/La2O2SO4, which was prepared from 18O-labeled La2O2SO4 as described above. Measurements were performed by placing the sample (0.1 g) in a stream of He at 600 °C and injecting a certain amount of CO (0.98 mL) repeatedly into the stream just before the sample bed with monitoring of the effluent gas composition using a quadrupole mass spectrometer (Omnistar; Pfeiffer). The oxygen release/storage cycles were also examined using a microbalance (8120; Rigaku) connected to a dual-gas supply system. The sample was first heated in a He stream up to 600 °C, and the weight was stabilized within 30 min. The gas feed to the sample was then switched between 1.4% H2/He (120 min) and 0.7% O2/He (40 min), while recording the sample weight.

Conclusions

Composite materials consisting of two different oxygen storage materials CeO2 and La2O2SO4 were prepared to examine their synergistic oxygen release and storage performance using anaerobic CO oxidation in the presence of impregnated Pt catalysts. The as-prepared CeO2/La2O2SO4 composite material achieved greater OSC and higher reaction rates than those of its individual constituents in a wide temperature range (400–700 °C). In situ XRD and Raman spectroscopy experiments demonstrated that CeO2 significantly accelerated the redox reaction between La2O2SO4 and La2O2S, while preserving its structure. The reaction between CO and 18O-labeled La2O2SO4 indicated that CO mainly reacted with the lattice oxygen atoms of CeO2, and the resulting oxygen vacancies were subsequently filled with oxygen atoms supplied from the solid interface with La2O2SO4. Consequently, the synergistic effect can be rationalized by considering that CeO2 acts as an oxygen gateway and La2O2SO4 acts as an oxygen reservoir when they are in intimate contact in a composite. The former activates the dissociation of O2 to form oxide ions or the reaction between oxide ions and reducing gases and transfers oxide ions to/from La2O2SO4, whereas the latter enhances OSC.