Layered Pr-dodecyl sulfate mesophases as precursors of Pr₂O₂SO₄ having a large oxygen-storage capacity

Abstract

Surfactant-assisted synthesis of the oxysulfate Pr₂O₂SO₄ with a large oxygen-storage capacity was studied with the aim of increasing the rate of redox cycles at lower temperatures. From an aqueous solution of nitrate, Pr-based surfactant mesophases templated by dodecyl sulfate anion (DS = C₁₂H₂₅OSO₃⁻) were synthesized using ammonia or urea as precipitants. The precipitated mesophases, Pr-DS-NH₃ and Pr-DS-N₂H₄CO, exhibited ordered layered structures with interlayer spacings of 2.63 and 3.68 nm, respectively, suggesting alternative stacking of a Pr hydroxide layer and a DS layer. The different interlayer spacing can be explained by the fact that Pr-DS-NH₃ with the composition Pr₂(OH)₅(DS)·2.5H₂O contains monolayers of DS, whereas Pr-DS-N₂H₄CO contains bilayers of DS. As was revealed by EXAFS and FT-IR results, the sulfate head group of each DS is strongly bound to uncoordinated Pr³⁺, so that heating of Pr-DS-NH₃ could directly yield single-phase Pr₂O₂SO₄ at a low temperature (500 °C). In contrast to the macropores (≥12 nm in size, 8 m² g⁻¹) of Pr₂O₂SO₄ prepared by calcining Pr₂(SO₄)₃, Pr-DS-NH₃ after calcination showed pores of 2–3 nm in size and a larger surface area (28 m² g⁻¹). The microstructure is very effective in accelerating the solid–gas reaction in oxygen release as well as storage processes, which enables the present system to work at temperatures ≤600 °C, compared to ≥700 °C required for Pr₂O₂SO₄ prepared from Pr₂(SO₄)₃.

---

Introduction

Oxygen-storage materials, which can store a large amount of gaseous oxygen under an oxidizing atmosphere and release it under a reducing atmosphere, have extensively been utilized in automotive emission-control catalysts for regulating oxygen partial pressure during deviations from the desired stoichiometric exhaust gas mixtures.^1–3 Thanks to this component, the three-way catalysts can achieve the efficient conversion of noxious pollutants. Although CeO₂–ZrO₂ has conventionally been used,^4–8 the storage capacity due to the redox between Ce⁴⁺ and Ce³⁺ cannot exceed 0.25 mol-O₂ per mol-Ce. Thus, materials with a larger storage capacity have become of significant interest. We have previously found a novel oxygen-storage mechanism based on the redox of sulfur in lanthanide (Ln) oxysulfates/oxysulfides,^9,10 which is the first successful example that uses a nonmetallic element as a redox site in place of metallic cations. The following reaction between S⁶⁺ and S²⁻ attains the capacity of 2 mol-O₂ mol⁻¹, which is 8 times more than that of CeO₂–ZrO₂.

Ln₂O₂SO₄ ⇄ Ln₂O₂S + 2O₂

Such oxygen-storage/release cycles between oxysulfate and oxysulfide are characteristic of the lanthanides. Ln oxysulfates (Ln = La, Pr, Nd and Sm) can be prepared from the corresponding sulfates and are thermostable up to 1000 °C. However, the primary drawback is that a higher operation temperature (ca. 800 °C) is necessary, whereas CeO₂–ZrO₂ can work even below 400 °C. Our study has therefore been focused on structural modification to achieve a lower redox temperature. Among Ln oxysulfates, the Pr system has the lowest possible operational temperature,¹⁰ probably because the redox between Pr³⁺ and Pr⁴⁺ at the surface can promote the redox of sulfur. Impregnation of noble metals, which activate reducing agents as well as oxygen, can further accelerate the redox cycles.⁹ These modifications can reduce the temperature to less than 700 °C, but further improvement may be achieved by microstructural modification, because Pr oxysulfate obtained from the sulfate has a low surface area (<10 m² g⁻¹). The creation of pores to increase the specific surface area would thus be essential to reduce the redox temperature.

In this work, we have studied the preparation of porous Pr oxysulfate by applying the surfactant-assisted synthesis of layered mesophases consisting of Pr hydroxide and dodecyl sulfate (DS = C₁₂H₂₅OSO₃⁻). Since the discovery of the mesoporous silicas known as MCM-41,^11,12 various surfactant mesophases have been synthesized as precursors to yield mesoporous materials including Ln oxides.^13,14 In such a surfactant-assisted pathway, part of the surfactant headgroup is known to be incorporated into the final material after heat treatment.¹³ In the present study, we therefore expected the surfactant to act not only as a template, but also as a reagent to yield stoichiometric Pr oxysulfate by using the interaction between a DS sulfate head group and a Pr hydroxide. The structure and its evolution to the oxysulfate by thermal decomposition were studied by means of XRD, FT-IR, TG, EXAFS, SEM, and adsorption measurement. Finally, dynamic redox cycles were carried out to evaluate the effect of the porous structure on the oxygen-storage characteristics.

Experimental

Praseodymium-based dodecyl sulfate mesophases were synthesized as precursors of the oxysulfate, Pr₂O₂SO₄. Either ammonia or urea were used to prepare Pr-DS precipitates, abbreviated as Pr-DS-NH₃ and Pr-DS-N₂H₄CO, respectively. For Pr-DS-NH₃, Pr(NO₃)₃·6H₂O (8.06 g), C₁₂H₂₅OSO₃Na (SDS) (10.89 g), 28% aqueous NH₃ (37.5 mL) and water (20 mL) were mixed at 40 °C for 1 h to yield a transparent solution, and then aged with stirring at 60 °C for more than 10 h. This was followed by cooling to room temperature, and the formation of the precipitate at pH 11. For Pr-DS-N₂H₄CO, Pr(NO₃)₃·6H₂O, SDS, urea and water were mixed in a molar ratio of 1 ∶ 2 ∶ 30 ∶ 60 and stirred at 40 °C to yield a transparent solution. The solution was then heated to 80 °C, where hydrolysis of urea increased the pH and caused the precipitation of Pr-DS composites. After 20 h at 80 °C, the resulting solid was collected when the pH value of the suspension reached ca. 10.

The precipitate thus obtained was collected by centrifugal separation, washed thoroughly with ion-exchanged and distilled water, dried by evacuation at room temperature, and finally heated in air. Pr₂O₂SO₄ was also synthesized by heating commercial Pr₂(SO₄)₃·nH₂O at 900 °C for 5 h in air (Pr-SO₄). The Pr₂O₂SO₄ thus obtained was impregnated with an aqueous solution of hydrogen palladium nitrate and then calcined at 450 °C to produce Pd-loaded samples (1 wt% loading).

The crystal structure was identified by use of a powder X-ray diffractometer (XRD, Rigaku Multiflex) with monochromated CuKα radiation (30 kV, 20 mA). Energy dispersive X-ray fluorescence analysis (XRF, Horiba MESA-500W) was used to determine the S/Pr ratio. The thermal decomposition of the as-prepared samples was studied by thermogravimetry (TG, Rigaku 8120). FT-IR spectra were recorded on a Jasco FTIR-610 spectrometer using a KBr method. The microstructure was observed by scanning electron microscopy (SEM, JEOL JSM6060LV). The BET surface area and pore size distribution were calculated from N₂ adsorption isotherms measured at 77 K (Bel Japan, Belsorp).

X-Ray absorption spectra of the Pr L_III-edge were recorded on a BL-7C instrument at the Photon Factory of the High Energy Accelerator Research Organization at Tsukuba (Proposal #2004G088), with a ring energy of 2.5 GeV and a stored current of around 300–450 mA. A Si(111) double-crystal monochromator was used. The incident X-ray was focused and the higher harmonics were removed by the total reflection on a Rh–Ni composite mirror. Pr L_III-edge spectra were recorded at room temperature in transmission mode. The incident and transmitted X-rays were monitored by ionization chambers filled with N₂/He (30 : 70) and N₂ gases, with lengths of 17 and 31 cm, respectively. A sample was pressed into a disk (diameter 10 mm) after its volume was adjusted by boron nitride powder to give an appropriate absorbance at the edge energy for the XAFS measurement. The XAFS data were processed by 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 3.0–10.8 Å⁻¹ region was Fourier-transformed. Phase shifts and backscattering amplitudes were obtained from EXAFS data of Pr₂O₂SO₄ and Pr₂(SO₄)₃ for Pr–O, Pr–S, and Pr–Pr. Data analysis was performed by the multiple shell fitting in r-space (1.4 < r < 4.7 Å, 3.0 < k < 10.8 Å⁻¹).

Dynamic oxygen-release and storage cycles for Ln₂O₂SO₄ were studied by the use of a microbalance (Rigaku 8120) connected to a dual-gas supplying system. The oxysulfate sample (ca. 10 mg) was firstly heated in a stream of N₂ up to 600 °C, where a constant weight was attained within 30 min. Then, the gas feed to the sample was switched between 5% H₂/He and 20% O₂/He, balanced by N₂ supplied at 20 cm³ min⁻¹, and the sample weight recorded at this temperature. During the measurement, N₂ was passed through the balance chamber to protect the weighing mechanism.

Results and discussion

Structure of Pr-DS mesophases

XRF analysis suggested that as-precipitated Pr-DS-NH₃ and Pr-DS-N₂H₄CO contained the template molecule in a molar ratio of S/Pr = 0.55 and 1.08, respectively. The content of sodium was confirmed to be negligible in both cases. As shown in Fig. 1, the XRD patterns of both precipitated solids gave rise to strong reflections at lower diffraction angles, which is indicative of the formation of ordered Pr mesophases templated by DS. Pr-DS-NH₃ exhibited three characteristic diffraction peaks at 2θ = 3.44, 6.84, and 10.28°, which are attributable to the (001), (002) and (003) reflections of a layered structure with a periodicity of 2.63 nm. Peaks observed at 2θ = 27.4 and 56.4° (noted by arrows in Fig. 1a) matched with the d₁₁₀ and d₂₂₀ of Pr(OH)₃,¹⁵ respectively, implying the presence of the hydroxide moiety in Pr-DS-NH₃. However, the absence of other reflections suggests a lack of periodicity along the c axis of Pr(OH)₃.

Fig. 1 XRD patterns of a) as-prepared Pr-DS-NH₃ and b) as-prepared Pr-DS-N₂H₄CO.

Precipitation using urea (Pr-DS-N₂H₄CO) also yielded a layered structure, but the periodicity (3.68 nm) was much larger (Fig. 1b). The intensity of diffraction peaks were 8 times larger than those of Pr-DS-NH₃, suggesting higher structural ordering in the mesophase composite. However, the purity of as-prepared Pr-DS-N₂H₄CO must be low due to the co-precipitation of PrCO₃OH, which would originate from the CO₂ evolved during the hydrolysis of urea. In the following experiment, Pr-DS-NH₃ was therefore used as a precursor of the oxysulfate (Pr₂O₂SO₄). The crystallinity of the Pr-DS-NH₃ was dependent on the time and temperature of the post-precipitation aging; the mesophase resulting from aging at 40 °C for 30 h exhibited the highest X-ray diffraction intensities.

Fig. 2 shows the XRD patterns of Pr-DS-NH₃ after heating in air. The disappearance of the (00l) peaks at ≤200 °C indicates that the layered structure had totally collapsed to yield a non-crystalline phase. Nevertheless, the S/Pr ratio remained almost constant at ca. 0.5 at elevated temperatures, suggesting that DS head sulfates are strongly bound to Pr. The non-crystalline phase remained up to 400 °C, but single-phase Pr₂O₂SO₄ was finally formed at 500 °C. Fig. 3 compares TG profiles of Pr-DS-NH₃ and Pr-SO₄ in a stream of 20% O₂/N₂. The hydrous phase of Pr-SO₄ (Pr₂(SO₄)₃·nH₂O) exhibited several endothermic weight losses due to dehydration (≤300 °C) and decomposition of 2 mol-sulfate mol⁻¹ to form Pr₂O₂SO₄ (800–950 °C) that was stable up to 1100 °C. By contrast, Pr-DS-NH₃ exhibited dehydration (≤100 °C), elimination of organic moieties (200–250 °C), and final dehydroxylation (300–400 °C). Another weight loss with an exothermic peak at ca. 600 °C would seem to result from the combustion of carboneous deposits originating from organic moieties. A final product of Pr-DS-NH₃ was also a single phase of Pr₂O₂SO₄, the thermal stability of which was as same as that from Pr-SO₄. These results show that Pr-DS-NH₃ can produce Pr₂O₂SO₄ at the very low temperature of 500 °C, compared to the heating above 900 °C required for the decomposition of Pr₂(SO₄)₃ to Pr₂O₂SO₄. With each weight loss and the S/Pr ratio taken into consideration, a possible chemical formula of Pr-DS-NH₃ could be as Pr₂(OH)₅(DS)·2.5H₂O. The calculated total weight loss to Pr₂O₂SO₄ should be 38%, which is in good agreement with the experimental value (36%).

Fig. 2 XRD patterns of a) as-prepared Pr-DS-NH₃, and after heating at b) 100 °C, c) 200 °C, d) 400 °C and e) 500 °C.

Fig. 3 TG profiles of a) Pr-SO₄ and b) Pr-DS-NH₃ measured at 10 °C min⁻¹ in 20% O₂/N₂.

Fig. 4 compares the infrared spectra of a) SDS and b) Pr-NH₃-DS in the sulfate headgroup band region of 1400–700 cm⁻¹. The sulfate headgroup with the C_3v point group showed characteristic absorption bands due to the asymmetric and symmetric S–O stretching mode (ν_as(E): 1250 cm⁻¹, ν_as(A): 1220 cm⁻¹ and ν_s(A): 1085 cm⁻¹).¹⁶ Here, the directions of the transition dipole moment for the E and A vibrational modes are orthogonal to each other. For Pr-DS-NH₃, the positions of the ν_as(E) mode (1250 cm⁻¹) were almost unchanged, but the other ν_as(A) and ν_s(A) modes were significantly shifted to lower wavelengths, 1204 and 1066 cm⁻¹, respectively. The band shift is likely associated with an electrostatic effect involving coordination of the sulfate headgroup oxygen atom to Pr³⁺. Similar shifts of these infrared bands have been reported for DS bound to Fe³⁺,¹⁷ TiO₂,¹⁸ and Al₂O₃.¹⁹ Li and Tripp¹⁸ have reported spectral changes of admicellar DS adsorbed onto a positively charged TiO₂ surface with a DS-chain axis near normal to the surface. According to their interpretation, the two A absorptions (ν_as and ν_s) shifted, because the transition dipole moments for the vibrations are near normal to the TiO₂ surface, and thus would directly be affected by coordination of the sulfate headgroups onto the positively charged site. By contrast, the change in the E vibration of the ν_as absorption would not be significant, because the transition dipole moment for the vibration in the lateral direction would not be affected by the surface charge. By applying this consideration, the Pr-DS-NH₃ phase should contain admicellar DS moieties, the headgroups of each DS being bound to Pr³⁺ located in the two-dimensional Pr hydroxide layer.

Fig. 4 FT-IR spectra of a) SDS and b) Pr-DS-NH₃.

On the basis of the chemical formula, Pr₂(OH)₅(DS)·2.5H₂O, and structural analysis, we have proposed a structural model of Pr-DS-NH₃ as shown in Fig. 5, which consists of two different layers of Pr hydroxide and DS. The Pr hydroxide moiety appears to be similar to the hexagonal form of the Pr trihydroxides in the space group P6₃/m with lattice parameters, a₀ = 0.646 nm and c₀ = 0.377 nm.¹⁵ The S/Pr ratio of 0.5 suggests that a 2 × c₀ thick Pr hydroxide layer is stacked alternately with a DS moiety. This is consistent with the fact that the XRD peaks due to (h0l) and (00l) of Pr(OH)₃ were too weak to be observed. The periodicity along (hk0) would only be preserved in the Pr-DS-NH₃ mesophase. Pr³⁺ in Pr(OH)₃ is 9-coordinate, but Pr³⁺ exposed on the surface of each Pr hydroxide layer is coordinated by a sulfate terminus, –OSO₃, instead of 3 × OH; their electrostatic interaction is supported by the shift of infrared bands ascribable to S–O stretching modes. This interaction is so strong that the S/Pr ratio of Pr-DS-NH₃ remains constant even after calcination, allowing the facile formation of Pr₂O₂SO₄ at low temperatures.

Fig. 5 A structural model of Pr-DS-NH₃.

Subtracting the thickness of a Pr hydroxide layer (2 × c₀ = ca. 0.75 nm) from the interlayer distance, 2.63 nm, gives the thickness of the DS layer, ca. 1.9 nm. Considering the molecular length of DS (ca. 2.1 nm),^20,21 the thickness of the DS layer is attributable to the alternating monolayer arrangement, in which each alkyl chain is tilted at an angle of ca. 64° to the Pr hydroxide layer. The average area per DS sulfate group in Fig. 5 can be estimated to be 0.18 nm² from the density of Pr³⁺ exposed at the surface of a Pr hydroxide slab. This value is very close to that of anhydrous SDS (0.19 nm²).²² Another possibility for the molecular packing is DS bilayers being tilted to the [hk0] direction, but the estimated tilt angle with respect to the (00l) plane of Pr hydroxide would then be too small (ca. 30°), compared to more than 55° observed for various phases of sodium dodecyl sulfate (SDS). Such a bilayer model may instead be adopted for the Pr-DS-N₂H₄CO mesophase (S/Pr = ca. 1), because the larger interlayer distance of 3.68 nm is rationalized by the sum of a 1 × c₀ thick Pr hydroxide layer (0.377 nm) and thickness of DS bilayer tilting by ca. 60° with respect to the Pr hydroxide layer (3.30 nm). It should be noted that the monolayer arrangement of DS molecules in the present Pr-DS-NH₃ system is in striking contrast to the bilayer arrangement widely observed for anhydrous and hydrous SDS. The bilayer composite should be more stable and more structurally ordered, and thus preferentially formed when the precipitation is carried out slowly, as in the case of the homogeneous precipitation method using urea (Pr-DS-N₂H₄CO). This is confirmed by very intense X-ray reflections, as shown in Fig. 1b. In contrast, the faster precipitation caused by using ammonia would yield the metastable monolayer phase (Pr-DS-NH₃), which exhibited much weaker X-ray reflections (Fig. 1a).

XAFS analysis

To verify the model structure in Fig. 5, the local environment of Pr in Pr-DS-NH₃ was investigated by EXAFS. Fig. 6 shows the k³-weighted EXAFS spectrum and the Fourier transform of the Pr L_III-edge. In the FT-EXAFS spectrum, the three peaks appeared at atomic distances of around 2.0, 3.5, and 4.0 Å. To obtain the detailed structural parameters on the basis of the model, a curve-fitting analysis of the Pr L_III-edge EXAFS spectrum was conducted in an r-range including the observed three Fourier peaks. The model structure contains two equimolar Pr sites, i.e., one is coordinated by 9 × OH (Pr–O(1) and Pr–O(2)), whereas the other is coordinated by 6 × OH and 3 × O of the sulfate termini, –OSO₃ (Pr–O(3)). Therefore, Pr–O bonds with three different lengths should be present in the first shell. In the second shell, the Pr–S and Pr–Pr bonds should be considered. According to the crystallographic data of Pr(OH)₃,¹⁵ two Pr–Pr bonds with different lengths (3.79 and 4.18 Å) are present. The FT-EXAFS spectrum was fitted by considering the contribution from the first and second shells to determine the structural parameters, including coordination numbers (N), bond distances (r), and Debye–Waller factors (σ) as summarized in Table 1. The present six-shell fitting revealed the best fit (see Fig. 6), with an R factor of 0.78. The peaks at 2.0, 3.5 and 4.0 Å can be attributed to Pr–O, Pr–S and Pr–Pr, respectively. For the first shell, two bonds of 2.52 and 2.57 Å with N = 2.9 and 4.7, respectively, can be assigned to Pr–OH (Pr–O(1) and Pr–O(2)), whereas the shorter bond of 2.38 with N = 1.7 to the Pr–O–SO₃ (Pr–O(3)). These atomic distances are in agreement with the crystallographic data of Pr(OH)₃ and Pr₂(SO₄)₃·8H₂O, respectively. The obtained coordination numbers correspond to average values for two different Pr sites in Pr-DS-NH₃ (Fig. 5), because half of the Pr atoms possess 3 × Pr–O(1) and 6 × Pr–O(2) bonds and the other half possess 3 × Pr–O(1), 3 × Pr–O(2) and 3 × Pr–O(3) bonds. For the second shell, the Pr–Pr(1) and Pr–Pr(2) coordination numbers were found to be 1.0 and 5.0, respectively. These values may also be explained by considering that half of the Pr atoms, located inside the Pr hydroxide slab, possess 2 × Pr–Pr(1) and 6 × Pr–Pr(2) bonds, whereas the other half, located at the surface, possess 1 × Pr–Pr(1) and 3 × Pr–Pr(2) bonds. Other attempts to fit the experimental data with the other structural models could not improve the fit. For instance, a model for the alternative stacking of a single Pr hydroxide layer (1 × c₀) and a DS monolayer gave rise to an R factor more than 7.

Fig. 6 (a) Pr L_III-edge k³-weighted EXAFS and (b) Fourier transform of Pr-DS-NH₃. The solid and dotted lines represent the experimental and fitted oscillations, respectively.

Table 1 The fitting range and fitting parameters for second coordination sphere of Pr-DS-NH₃ (six-shell fit)

Shell	Δk/Å^−1 ^a	Δr/Å ^b	N ^c (±0.2)	r/Å ^d (±0.03)	σ ^2/10^−2 Å^2 ^e (±0.02)	R factor
a Interval of k-space to r-space of FT. b r-Space interval selected to perform reverse FT into k-space. c Coordination number. d Atomic distance. e Debye–Waller factor.
Pr–O(1)	3.0–10.8	1.41–4.73	2.9	2.52	2.13	0.78
Pr–O(2)			4.7	2.57	2.13
Pr–O(3)			1.7	2.38	0.72
Pr–S			1.1	3.48	0.37
Pr–Pr			1.0	3.81	0.14
Pr–Pr(2)			5.0	4.10	1.27

---

Microstructure and oxygen-storage properties

Once the surfactant-assisted synthesis had been found to lead to the low-temperature formation of Pr₂O₂SO₄, the effect on the microstructure was studied by comparing SEM images and N₂ adsorption isotherms. The two Pr₂O₂SO₄ samples obtained from Pr-DS-NH₃ and Pr-SO₄ exhibited a marked contrast in their SEM images, as shown in Fig. 7. Pr₂O₂SO₄ formed after heating Pr-SO₄ at 900 °C was composed of large columnar crystals, with several being 100 µm long (a). Although the external morphology was very similar to that of Pr-SO₄ crystals, a number of pores and cracks were formed due to the elimination of SO₂/O₂ at the decomposition temperature, ≤850 °C (Fig. 3). Accompanied by such a microstructural change, the BET surface area increased from <1 m² g⁻¹ (Pr-SO₄) to 8 m² g⁻¹ (Pr₂O₂SO₄). In contrast, as-prepared Pr-DS-NH₃ consisted of particles (several being 10 or 100 nm in size) and their agglomerates (b). After heating at 500 °C in air, Pr₂O₂SO₄ particles smaller than Pr-DS-NH₃ were observed (c), indicating that significant elimination of organic moieties may not cause sintering, but the volumetric shrinkage or the pulverization of precursor particles. The surface area of Pr-DS-NH₃ after calcination at 500 °C was 28 m² g⁻¹. Assuming a spherical Pr₂O₂SO₄ particle, its diameter could be estimated as 40 nm from the surface area, which is roughly consistent with the SEM image (Fig. 7c).

Fig. 7 SEM photographs of a) Pr₂O₂SO₄ prepared by heating Pr-SO₄ at 900 °C, b) as-prepared Pr-DS-NH₃ and c) Pr₂O₂SO₄ prepared by heating Pr-DS-NH₃ at 500 °C.

A marked contrast can also be observed for the N₂ adsorption isotherm and pore size distribution (Fig. 8 and Fig. 9). The N₂ adsorption–desorption isotherm of Pr-SO₄ after heating at 900 °C looks like type III (IUPAC), which was flat up to p/p₀ = ca. 0.9 and then sharply increased near the saturation of vapor pressure, indicating the presence of macropores. This is contrast to the isotherm for Pr-DS-NH₃ after calcination at 500 °C, in which a gradual increase of adsorption with hysteresis in the range p/p₀ ≥ 0.5 suggests the contribution of textural mesopores. In both cases, negligible adsorption near p/p₀ = 0 suggests the absence of micropores. Pore size distribution calculated from the adsorption isotherms presents peak maxima at pore radii of ca. 18 nm and ≤5 nm for Pr-SO₄ and Pr-DS-NH₃, respectively. The small surface area as well as macropores in the range ≤20 nm is in accord with the SEM micrograph. The surface area and pore structure of Pr₂O₂SO₄ synthesized from Pr-DS-NH₃ were stable up to 900 °C.

Fig. 8 N₂ adsorption isotherms of Pr₂O₂SO₄ prepared from a) Pr₂(SO₄)₃ and b) Pr-DS-NH₃.

Fig. 9 Pore size distribution of Pr₂O₂SO₄ prepared by a) Pr₂(SO₄)₃ and b) Pr-DS-NH₃.

To evaluate the effect of preparation method on the oxygen-storage/release performance, the dynamic redox behavior of Pr₂O₂SO₄ was evaluated under oscillating feed stream conditions, where reducing (5% H₂/He) and oxidizing (20% O₂/He) atmospheres were cycled, by the use of a flow microbalance. Fig. 10 exhibits the change of weight during oxygen release/storage cycles at 600 °C for 1 wt% Pd/Pr₂O₂SO₄ prepared from Pr-SO₄ (Fig. 10a) and Pr-DS-NH₃ (Fig. 10b), obtained after heating at 900 °C and 600 °C, respectively. First, the sample was heated in flowing 20% O₂/He up to 600 °C, where on approaching a constant weight, the gas feed was switched to a mixture of 5% H₂ in N₂. The oxygen release to form Pr₂O₂S in for Pr-SO₄ proceeded very slowly, taking about 10 h to complete, whereas the subsequent oxygen storage was complete within only 10 min. The different rate reflects the slower reduction of Pr₂O₂SO₄, compared to the thermodynamically favorable oxidation of Pr₂O₂S. In contrast, the weight change due to oxygen release in Pr-DS-NH₃ took place very smoothly, and approached the stoichiometric value. Table 2 summarizes the rates of oxygen release/storage, which were estimated from the slope of weight change in Fig. 10. Clearly, both oxygen release and storage for Pr-DS-NH₃ was much faster; the calculated rate was more than 7 times faster than that for Pr-SO₄. Reversible oxygen release/storage cycles were obvious from stable weight changes. In accord with the surface area and pore structure, which are thermostable up to 900 °C, the oxygen release/storage properties were not affected by further heating at 600–900 °C.

Fig. 10 Oxygen release/storage cycles in a microbalance at 600 °C of 1 wt% Pd loaded Pr₂O₂SO₄ prepared from a) Pr₂(SO₄)₃ and b) Pr-DS-NH₃. Dotted lines correspond to the ideal weight change due the redox between Pr₂O₂SO₄ and Pr₂O₂S.

Table 2 Rates of oxygen release and storage for Pr₂O₂SO₄ at 600 °C

Precursor	Release/mol-O2 mol^−1 min^−1	Storage/mol-O2 mol^−1 min^−1
Pr-SO4	0.004	0.074
Pr-DS-NH3	0.033	0.530

---

The observed acceleration of oxygen release/storage can be considered to be a simple particle size effect caused by the microstructural change, but it is considerably larger than expected from the difference of BET surface area (8 m² g⁻¹ for Pr-SO₄, 28 m² g⁻¹ for Pr-DS-NH₃). This would indicate that mass transfer in the solid and/or pores are essential as well as solid/gas interface reactions. As is indicated by SEM images (Fig. 7), Pr₂O₂SO₄ derived from Pr-SO₄ is composed of large crystals, with several being 100 µm in size, in which the reaction rate should be determined by the slow diffusion of oxide ions in the solid. In this regard, the nanostructured Pr₂O₂SO₄ particles obtained by the surfactant-assisted syntheses give rise to a much shorter diffusion distance, and thus contribute to the rapid oxygen release and storage.

In summary, we have successfully performd a surfactant-assisted synthesis of a stoichiometric layered Pr₂(OH)₅(DS)·2.5H₂O mesophase, which can yield porous Pr₂O₂SO₄ after heating. Based on the structural and compositional analysis, DS was found to act as a reactive template to yield the Pr oxysulfate. A significant advantage of this novel synthetic route is to achieve the large oxygen-storage capacity of 2 mol-O₂ mol⁻¹ at the relatively low temperature of 600 °C, which may have broad significance for various practical applications.

Acknowledgements

This study was supported by an Industrial Technology Research Grant Program in '05 from the New Energy and Industrial Technology Development Organization (NEDO) of Japan, a Grant-in-Aid for Scientific Research on Priority Area (440) from the Ministry of Education, Culture, Sports, Science and Technology (MEXT), and the Nippon Sheet Glass Foundation for Materials Science and Engineering.

References

1. K. C. Taylar, Catal. Rev. Sci. Eng., 1993, 35, 457.
2. S. Bernal, J. Kaspar and A. Trovarelli, Catal. Today, 1999, 2, 50.
3. A. Trovarelli, in Catalysis by Ceria and Related Materials, Vol. 2, ed. A. Trovarelli, Imperial College Press, London, 2002.
4. M. Ozawa, M. Kimura and A. Isogai, J. Alloys Compd., 1993, 193, 73.
5. M. Pijolat, M. Prin, M. Soustelle, O. Touret and P. Nortier, J. Chem. Soc., Faraday Trans., 1995, 91, 3941.
6. F. Zamur, A. Trovarelli, C. de Leitenburg and G. Dolcetti, J. Chem. Soc., Chem. Commun., 1995, 965.
7. G. Balducci, P. Fornasiero, R. Di Monte, J. Kaspar, S. Meriani and M. Graziani, Catal. Lett., 1995, 33, 193.
8. J. Kaspar and P. Fornasiero, J. Solid State Chem., 2003, 171, 19.
9. M. Machida, K. Kawamura and K. Ito, Chem. Commun., 2004, 662.
10. M. Machida, K. Kawamura, K. Ito and K. Ikeue, Chem. Mater., 2005, 17, 1487.
11. C. T. Kresge, M. E. Leonowicz, W. J. Roth, J. C. Vartuli and J. S. Beck, Nature, 1992, 359, 710.
12. S. Inagaki, Y. Fukushima and K. Kuroda, J. Chem. Soc., Chem. Commun., 1993, 680.
13. M. Yada, H. Kitamura, A. Ichinose, M. Machida and T. Kijima, Angew. Chem., Int. Ed., 1999, 38, 3506.
14. L. Jongen, D. Hinz, G. Meyer and K. Binnemans, Chem. Mater., 2001, 13, 2243.
15. D. F. Mullica, W. O. Milligan and G. W. Beall, J. Inorg. Nucl. Chem., 1979, 41, 525.
16. K. Nakamoto, Infrared and Raman Spectra of Inorganic and Coordination Compounds, 4th edn, Wiley, New York, 1986.
17. B. Bai, N. P. Hankins, M. J. Hey and S. W. Kingman, Ind. Eng. Chem. Res., 2004, 43, 5326.
18. H. Li and C. P. Tripp, J. Phys. Chem. B, 2004, 108, 18318.
19. R. P. Sperline, Y. Song and H. Freiser, Langmuir, 1992, 8, 2183.
20. M. Adachi Pagano, C. Forano and J. P. Besse, Chem. Commun., 2000, 91.
21. B. T. Holland, P. K. Isbester, C. F. Blanford, E. J. Munson and A. Stein, J. Am. Chem. Soc., 1997, 119, 6796.
22. L. A. Smith, R. B. Hammond, K. J. Roberts, D. Machin and G. McLeod, J. Mol. Struct., 2000, 554, 173.

---