Facile preparation of mesoporous Cu–Sn solid solutions as active catalysts for CO oxidation

Abstract

With a facile co-precipitation method, a series of high surface area mesoporous Cu_xSn_1−xO_y solid solution catalysts have been synthesized and applied to CO oxidation. Compared with individual SnO₂ and CuO, the activity of these catalysts is remarkably improved. The highest activity is achieved on Cu_0.5Sn_0.5O_y, a catalyst with a Cu/Sn molar ratio of 0.5/0.5 and a Caramel-Treats-like morphology. It is revealed by XRD, SEM-EDX mapping and HR-TEM results that Cu²⁺ cations have been incorporated into the crystal lattice of rutile SnO₂ to form a uniform solid solution structure. As testified by N₂ adsorption–desorption and SEM results, these Cu_xSn_1−xO_y catalysts contain well-defined mesopores and possess high surface areas and improved pore volumes, which are favourable for the dispersion of the active sites, the diffusion of the reactants and the easy interaction between the reactants and the catalyst surface. Moreover, H₂-TPR and XPS results demonstrate that more active and loosely bounded oxygen species have been formed on the surface of these catalysts. It is believed that these are the predominant reasons leading to the superior CO oxidation activity over the Cu_xSn_1−xO_y catalysts. Notably, these Cu_xSn_1−xO_y catalysts are also resistant to water vapour deactivation, indicating they have the potential to be used in real exhaust control processes.

---

Introduction

CO emitted from either mobile or stationary sources is one of the major pollutants for air. Catalytic oxidation has been proved to be the most feasible solution to eliminate CO pollution. For room temperature CO oxidation, Hopcalite, a Cu–Mn^1,2 mixed oxide catalyst invented in 1919 by Johns Hopkins University and California University together, is still the cheapest and most active catalyst but suffers from easy water deactivation. In addition, for more stringent conditions, for example, CO in exhausts from vehicles and industrial plants, it is not practical to use the Hopcalite catalyst due to thermal deactivation. Although supported precious metal catalysts such as Pd/Al₂O₃,³ Pt/TiO₂,⁴ Au/CeO₂,⁵ Pd/SnO₂,⁶ Pt/SnO₂,⁷ Au/TiO₂ (ref. 8) etc. have been reported to be active, stable and water resistant for CO oxidation, and can also stand more harsh conditions, it is well known that precious metals have limited sources and availability, which significantly increases the cost for these catalysts and restricts their application. Therefore, there is still strong motivation for people to search for cheaper non-noble metal catalysts with competent performance as substitutes.

Because of its relatively lower price, rich sources, and high activity, CuO-based catalysts have attracted much attention for CO oxidation and been intensively investigated. Over past several decades, besides the above-mentioned Hopcalite catalyst, CuO supported on various metal oxides, such as Al₂O₃,⁹ Fe₂O₃,¹⁰ TiO₂,^11–13 CeO₂,^14–20 have been studied for CO oxidation. It is concluded some of the catalysts are active, stable and have the potential to be used in real emission control processes.

SnO₂ is an n-type semiconductor, which has been widely used as gas sensing materials.^21,22 Because of its facile surface and lattice oxygen and high thermal stability (melting point, 1630 °C), the catalytic properties of SnO₂ have gotten more and more attention. Over past five years, a series of systematical work has been performed by our group to understand the catalytic chemistry of SnO₂-based materials.^23–29 It has been found that polycrystalline SnO₂ modified by Fe,³⁰ Ce,²⁵ Ta,²³ Cr,^31,32 etc. is active for CO and CH₄ oxidation. In addition, SnO₂ nano-rod with preferentially exposed (110) facets displays the catalytic behavior of precious metals.²⁴ Most importantly, we also find that SnO₂ is one of the few oxides with water resistance.²⁸ Although Cu-modified SnO₂ materials have been extensively studied as gas sensors,^33,34 electrodes,^35,36 ceramics³⁷ and photo catalysts,³⁸ the combination of Cu and Sn oxides as catalysts for pollution control catalysts have been rarely profiled. For CH₄ oxidation, both Zhu et al.³⁹ and Cui et al.⁴⁰ found that nano-sized Cu–Sn mixed oxides prepared by co-precipitation method showed evidently improved surface area as well as activity in comparison with individual CuO and SnO₂. For CO oxidation, Fuller et al.⁴¹ reported in 1973 that a SnO₂–CuO gel catalyst with a Cu/Sn atomic ratio of 0.53/1, which was prepared by co-precipitation method, possessed much higher activity than Hopcalite catalyst. On this catalyst, the steady state temperature for 50% CO conversion is 65 °C, which is 50 °C lower than that on Hopcalite catalyst. It is concluded by Fuller and co-workers that this SnO₂–CuO gel catalyst could be used for automobile exhaust purification. However, Wang et al.¹¹ found that although an 8% CuO/SnO₂ catalyst prepared with impregnation method was very active for CO oxidation, but showed still lower activity than Hopcalite catalyst, on which 100% CO conversion occurred at 130 °C.

To clarify this discrepancy and aim to develop an active and stable catalyst with water resistance for practical use, a series of high surface area and mesoporous Cu_xSn_1−xO_y catalysts have been synthesized by a simple co-precipitation method in this work without using any template. It is found that all the Cu_xSn_1−xO_y catalysts show obviously improved surface areas as well as activity than the individual CuO and SnO₂. The highest CO oxidation activity is obtained on a catalyst with an equal Cu/Sn molar ratio. The reaction results are elucidated and discussed in detail along with the characterization results provided by N₂ adsorption–desorption, SEM, TEM, XPS, XRD and H₂-TPR techniques.

Experimental

Catalyst preparation

Mesoporous Cu–Sn mixed oxides with different Cu/Sn atomic ratios were prepared by an easy co-precipitation method. In detail, the desired amount of SnCl₄ (0.5 mol L⁻¹) and Cu(NO₃)₂ (0.5 mol L⁻¹) solution was mixed and stirred thoroughly at room temperature for 60 min. Then Na₂CO₃ solution (0.5 mol L⁻¹) was dripped slowly into the above solution mixture until the pH reached 8, which was followed by another 60 min's continuous stirring. Afterwards, the precipitate was vacuum-filtered and washed with distilled deionized (DDI) water until Cl⁻ free, with a total dissolved solid (TDS) less than 10 ppm. The precipitate was dried at 110 °C overnight, and then calcined at 300 °C in air atmosphere for 4 h. For comparison study, pure SnO₂ and CuO were prepared by the same method. In addition, other two CuO/SnO₂ samples, which have the same chemical composition to Cu_0.5Sn_0.5O_y, the best catalyst in this study, were prepared by impregnating suitable amount of Cu(NO₃)₂ (0.5 mol L⁻¹) solution onto calcined or un-calcined SnO₂ powder supports, which were named as IMP-300 and IMP-110, respectively.

Catalyst characterization

The scanning electron microscope (SEM) images were taken on a Hitachi S-4800 field emission scanning electron microscope. The Transmission Electron Microscope (TEM) images were taken on a Tecnai™ F30 transmission electron microscope.

Nitrogen adsorption–desorption of the samples were carried out at 77 K on ASAP2020 instrument. The specific surface areas of the samples were calculated using Brunauer–Emmett–Teller (BET) method in the relative pressure (P/P₀) range of 0.05–0.25. The pore size distributions of the samples were calculated with Barrett–Joyner–Halenda (BJH) method. The average pore sizes of the samples were obtained from the peak positions of the distribution curves. The total pore volume of each catalyst was accumulated at a relative pressure of P/P₀ = 0.99.

The powder X-ray diffraction (XRD) patterns were recorded on a Bruker AXS D8Focus diffractometer operating at 40 kV and 30 mA, with a Cu target Kα-ray irradiation (λ = 1.5405 Å). Scans were taken with a 2θ range from 10 to 90° and with a step of 5° min⁻¹. To keep the data comparable, all of the samples were tested continuously under the same condition. The mean crystallite sizes of the samples were calculated with Scherrer equation based on the three strongest peaks of SnO₂ with hkl of (110), (101) and (211). The calculated experimental error for 2θ measurement of the peaks is ±0.01°, which ensures the reliable identification of peak shift observed by solid solution formation.

Hydrogen temperature programmed reduction (H₂-TPR) experiments were carried out on FINESORB 3010C instrument in a 30 mL min⁻¹ 10% H₂/Ar gas mixture flow. Generally, 50 mg catalysts were used for the tests except that only half amount was used for CuO and SnO₂, the two pure samples. Prior to the experiments, the catalysts were re-calcined in a high purity air flow at 300 °C for another 30 min to remove any surface impurities. The temperature was increased from room temperature to 850 °C with a rate of 10 °C min⁻¹. A thermal conductivity detector (TCD) was employed to monitor the H₂ consumption. For H₂ consumption quantification, CuO (99.99%) was used as the calibration standard.

X-ray Photoelectron Spectroscopy (XPS) test was carried out on a PerkinElmer PHI1600 system using a single Mg–K-X-ray source operating at 300 W and 15 kV. The spectra were obtained at ambient temperature with an ultrahigh vacuum. The binding energies were calibrated using the C 1s peak of graphite at 284.5 eV as a reference.

Activity evaluation

The catalysts were evaluated for CO oxidation with a U-shaped quartz tube (ID = 6 mm) reactor with a down flow over 100 mg catalyst. Typically, 0.3–0.4 mm catalyst particles were used for activity evaluation. A K-type thermocouple was placed on top of the catalyst bed with the thermocouple head point touching the catalyst to monitor the reaction temperature. To measure the light-off behaviours of the catalysts, all data were collected with increasing the temperature. The volume composition of the feed gas is 1% CO, 21% O₂ and balanced by high purity N₂, with a flow rate of 40 mL min⁻¹, which corresponds to a space velocity of 24 000 mL h⁻¹ g_cat.⁻¹. The reactants and products were analysed on-line on a GC9310 gas chromatograph equipped with a TDX-01 column and a TCD detector. To get steady state kinetic data, the reaction at each temperature was stabilized at least 30 min before analysis. The flow rate of the H₂ carrier gas is 30 mL min⁻¹.

Results

Catalytic performance

CO oxidation activity on the catalysts is shown in Fig. 1. For clarification, the temperatures corresponding to 10%, 50% and 100% CO conversion (T₁₀, T₅₀ and T₁₀₀) are also listed in Table 1. As shown in Fig. 1(A), both of the pure SnO₂ (Fig. 1(A-a)) and CuO (Fig. 1(A-h)) show low activity, on which 100% CO conversion occurs at 300 and 270 °C, respectively. With the combination of Cu and Sn oxides, all of the Cu_xSn_1−xO_y catalysts exhibit obviously improved activity, as testified by the shifting of the CO conversion curves to lower temperature region. With the increasing of Cu amount, the CO oxidation activity of the catalysts increases gradually. The highest activity is obtained with Cu_0.5Sn_0.5O_y, on which the complete CO conversion is achieved at 140 °C (Fig. 1(A-f)). At T₅₀ (117 °C) of Cu_0.5Sn_0.5O_y for 50% CO conversion, no CO oxidation starts for the pure CuO and SnO₂. However, further increasing the Cu amount decreases the CO oxidation activity, indicating that too much Cu in the Cu–Sn composite oxides has negative effect on the activity.

Fig. 1 Catalytic performance of Cu_xSn_1−xO_y catalysts. (A) CO conversion versus temperature, (B) Arrhenius plots, (C) stability tests in the absence or presence of water vapour. (a) SnO₂, (b) Cu_0.15Sn_0.85O_y, (c) Cu_0.2Sn_0.8O_y, (d) Cu_0.3Sn_0.7O_y, (e) Cu_0.4Sn_0.6O_y, (f) Cu_0.5Sn_0.5O_y, (g) Cu_0.6Sn_0.4O_y, (h) CuO.

Table 1 Reaction performance for CO oxidation over Cu_xSn_1−xO_y catalysts

Catalysts	CO oxidation activity			Reaction rate^a [10^−4 mmol g^−1 s^−1]	Ea^a [kJ mol^−1]
	T10 (°C)	T50 (°C)	T100 (°C)
a Calculated from the Arrhenius plots and formula.
SnO2	190	245	300	0.27	38.4
Cu0.15Sn0.85Oy	165	210	230	1.08	30.1
Cu0.2Sn0.8Oy	120	163	200	1.52	31.9
Cu0.3Sn0.7Oy	100	138	180	3.13	30.0
Cu0.4Sn0.6Oy	100	133	170	3.15	33.2
Cu0.5Sn0.5Oy	90	117	140	5.91	34.4
Cu0.6Sn0.4Oy	100	130	160	3.07	34.7
CuO	165	210	270	0.38	36.0

---

The Arrhenius plots of all the catalysts are compared in Fig. 1(B). As listed in Table 1, the reaction rates at 100 °C of all the catalysts are also calculated from the Arrhenius plots and formula. To exclude mass and heat transfer influence, all of the CO oxidation rates used to get Arrhenius plots for rate and activation energy calculation were collected under differential condition with CO conversion below 20%. Compared with individual CuO and SnO₂, all of the Cu_xSn_1−xO_y catalysts have evidently higher rates. The rate on Cu_0.5Sn_0.5O_y, the most active catalyst in this study, is 5.91 × 10⁻⁴ mmol g⁻¹ s⁻¹, which is about 20 times that of the two pure samples. Moreover, compared with the two individual samples, the activation energies on the Cu_xSn_1−xO_y catalysts are 3–8 kJ mol⁻¹ lower. Obviously, with the combination of Cu and Sn oxides, more active reaction sites have been formed on the Cu–Sn binary oxide catalysts.

In most of the authentic exhausts either from either stationary or mobile sources, 3–10% water vapour is generally present. The stability of a catalyst in the presence of water vapour is a crucial parameter to determine its application potential. Therefore, Cu_0.5Sn_0.5O_y, the most active sample in this study, was subject to a life time test at a CO conversion around 70% in the absence or in the presence of 5% water vapour. As shown in Fig. 1(C), 100 h's constant running, either with or without water vapour, did not lead to any decrease of the CO conversion, indicating this catalyst is stable and resistant to water deactivation. These facts demonstrate that Cu_xSn_1−xO_y catalysts with suitable Cu/Sn atomic ratios have not only remarkably improved activity but also superior stability in the presence of water vapour, indicating that they have the potential to be used in some real industrial CO abatement processes.

Nitrogen adsorption–desorption measurements

To understand the reasons influencing the CO oxidation activity of the Cu_xSn_1−xO_y catalysts with different Cu/Sn molar ratios, the catalysts were subjected to nitrogen adsorption–desorption measurements. The surface areas and pore volumes of the catalysts are listed in Table 2. After calcination at 300 °C, the surface area of pure SnO₂ is 152 m² g⁻¹, which is much higher than the 4 m² g⁻¹ of pure CuO. All of the Cu_xSn_1−xO_y catalysts possess surface areas above 110 m² g⁻¹. With the increasing of the Cu/Sn molar ratio, the surface area and pore volume increases until equal amount of Cu and Sn in the sample. The highest surface area and pore volume can be obtained with Cu_0.5Sn_0.5O_y, the most active catalyst in this study, and the values are 196 m² g⁻¹ and 0.17 cm³ g⁻¹, respectively. It is noted here that although pure SnO₂ has higher surface area than most of the Cu_xSn_1−xO_y catalysts, it shows still lower CO oxidation activity. However, all of the Cu_xSn_1−xO_y catalysts have evidently improved pore volumes than the two individual samples, which could facilitate the diffusion of the reactants and products. Therefore, it is believed that the high surface areas of the Cu_xSn_1−xO_y catalysts could contribute to their CO oxidation activity but are not the major factors. The improved pore volumes could play a more critical role to enhance the CO oxidation activity of the Cu_xSn_1−xO_y catalysts. In addition, the combination of Cu and Sn oxide could affect the bulk and surface properties of the catalysts, thus inducing the formation of more active sites, as indicated by the improved reaction rates and decreased activation energies, which will be discussed in more detail in the following sections.

Table 2 Physicochemical properties of Cu_xSn_1−xO_y catalysts

Catalysts^a	SBET (m^2 g^−1)	Average pore size (nm)	Average pore volume (cm^3)	SnO2 mean crystallite size (nm)	SnO2 (110)		SnO2 (101)
					2θ (°)	d (Å)	2θ (°)	d (Å)
a Calcined at 300 °C for 4 h in air.b The mean crystallite size of CuO is 15.5 nm.
SnO2	152	2.54	0.08	3.5	26.86	3.32	33.74	2.65
Cu0.15Sn0.85Oy	119	3.02	0.10	4.3	26.66	3.34	33.69	2.65
Cu0.2Sn0.8Oy	143	3.00	0.11	3.8	26.74	3.34	33.70	2.65
Cu0.3Sn0.7Oy	162	3.01	0.14	3.2	26.70	3.34	33.67	2.66
Cu0.4Sn0.6Oy	167	3.03	0.14	2.2	26.51	3.36	33.66	2.66
Cu0.50Sn0.5Oy	196	3.37	0.17	2.1	26.52	3.36	33.60	2.66
Cu0.6Sn0.4Oy	116	4.04	0.15	3.35	26.58	—	—	2.5
CuO^b	4	38.6	0.09	—	—	—	—	—

---

The nitrogen adsorption–desorption isotherms of all the catalysts are shown in Fig. S1.† As it can be observed in Fig. S1(A),† SnO₂ and CuO exhibit a type I isotherm with the characteristics of non-porous solids. However, All the Cu_xSn_1−xO_y samples exhibit type IV nitrogen sorption isotherms with a capillary condensation step, which is the characteristic of mesoporous materials. The H2-type hysteresis loop in the relative pressure range of P/P₀ = 0.4–0.7 is also typical for mesoporous structure formed by nanoparticle assembly. Interestingly, both Cu_0.5Sn_0.5O_y and Cu_0.6Sn_0.4O_y have a second H3-type hysteresis loop in the higher pressure region of P/P₀ = 0.8–1.0. It is generally considered that H3-type hysteresis loop corresponds to slit pore formed by flake particle accumulation,⁴² which is consistent with the results of SEM in Fig. 4 that will be discussed in detail later. The pore-size distribution curves shown in Fig. S1(B)† also demonstrate the absence of any evident pores in the two individual samples, but the presence of mesopores in all of the Cu_xSn_1−xO_y samples. As listed in Table 2, the average pore sizes of the Cu_xSn_1−xO_y samples are around 3 nm, which do not change too much with the varying of the Cu/Sn molar ratios. It is particularly mentioned here that for both Cu_0.5Sn_0.5O_y and Cu_0.6Sn_0.4O_y samples, a second group of mesopores with larger pore sizes are obviously observed, as indicated by Fig. S1(B).† The pore volumes of the Cu_xSn_1−xO_y samples are larger than that of the two individual samples, which also increases with the increasing of the Cu/Sn ratio. In summary, with the combination of Cu and Sn oxides, mesoporous Cu_xSn_1−xO_y catalysts can be synthesized. The presence of mesopores is believed to be the predominant factor to enlarge the surface area and pore volume of the Cu–Sn binary catalysts, which is favourable for the diffusion of the reactants and products, and the contacting of the reactants with the active sites, thus improving the activity of the Cu_xSn_1−xO_y catalysts significantly.

XRD analysis and SEM-EDX mapping results

XRD patterns of the Cu_xSn_1−xO_y samples are shown in Fig. 2. Pure CuO shows sharp diffraction peaks of CuO phase, testifying that it crystallizes well during the calcination process, which is in agreement with its extra low surface area. In contrast, pure SnO₂ displays the typical diffraction feature of tetragonal rutile SnO₂ phase. The low intensity of the peaks is in line with its high surface area. Interestingly, no CuO phase can be observed in the XRD patterns of the Cu_xSn_1−xO_y samples even up to a Cu/Sn molar ratio of 0.5/0.5. On the diffraction profiles of these samples, only broadened tetragonal rutile SnO₂ diffraction peaks can be observed, indicating that CuO phase is either amorphous, or Cu²⁺ cations are incorporated into the crystal lattice of SnO₂ to form a solid solution structure. To clarify this, the 2θ and d values of the two strongest peaks of rutile SnO₂ phase, peak (110) and (101), in the Cu_xSn_1−xO_y samples are identified carefully and compared in Table 2. Compared with pure SnO₂, it is found that both of the two diffraction peaks shift to lower angles but the d values increase with the incorporation of CuO, indicating the expansion of the distance between the crystal facets. These changes are typical for solid solution formation, demonstrating that Cu²⁺ cations could have been doped into the crystal lattice of SnO₂ to replace part of the Sn⁴⁺ cations to form solid solution structure. According to Vegard's law⁴³ about solid solution formation between metal oxides, if one metal oxide dissolves into the lattice of another, there is a certain capacity, which will be determined by the radii and valence states of both metal cations. As reported in literature,^25,44 Cu²⁺ with a coordination number of 6 has a radius of 0.074 nm, while that of Sn⁴⁺ with the same coordination number is 0.069 nm, which are very close to each other. However, the valence state difference between the two cations is relatively big. Therefore, it is reasonable to deduce that after the amount of Cu²⁺ cations in the SnO₂ lattice reaches the capacity, excess CuO will be formed and present on the surface of Cu–Sn solid solution. Based on this, it is not difficult to understand that when the Cu/Sn ratio increases to 0.6/0.4, CuO phase can be observed for Cu_0.6Sn_0.4O_y sample. It is noted here that similar to other Cu–Sn solid solution samples, the 2θ and d values changes are also observed for Cu_0.6Sn_0.4O_y, confirming that Cu–Sn solid solution is formed first and present in its bulk structure.

Fig. 2 XRD patterns of Cu_xSn_1−xO_y catalysts. (a) SnO₂, (b) Cu_0.15Sn_0.85O_y, (c) Cu_0.2Sn_0.8O_y, (d) Cu_0.3Sn_0.7O_y, (e) Cu_0.4Sn_0.6O_y, (f) Cu_0.5Sn_0.5O_y, (g) Cu_0.6Sn_0.4O_y, (h) CuO.

To further elucidating the formation of solid solution structure between Cu and Sn oxides, Cu_0.5Sn_0.5O_y, a typical sample in this study, was thus measured by SEM-EDX mapping, with the image shown in Fig. 3. The mapping zone is labelled in the left SEM image of the figure, in which the Cu and Sn elements distribute very uniformly, as indicated by the two Cu and Sn distribution images on the right side. This confirms that homogeneous solid solution structure has been formed in the Cu–Sn mixed oxide samples.

Fig. 3 SEM-EDX mapping images of Cu_0.5Sn_0.5O_y.

In comparison with the activity evaluation results, it is reasonable to conclude that the formation of Cu–Sn solid solution is of advantage to the CO oxidation activity. The more the formed Cu–Sn solid solution in a Cu_xSn_1−xO_y sample, the higher its CO oxidation activity. However, excess CuO on the surface of the Cu–Sn solid solution is harmful to the activity, as indicated by the reaction results obtained on Cu_0.6Sn_0.4O_y.

SEM and TEM images of the catalysts

The morphologies of the catalysts were characterized by SEM techniques, with the images shown in Fig. 4. As shown in Fig. 4(a), the image of pure SnO₂ exhibits superfine powder structure, which is in agreement with its high surface area and low crystallinity. In comparison, pure CuO consists of uniform long particles with an average size of 100 nm, as shown in Fig. 1(f). While the image of the Sn rich sample, Cu_0.2Sn_0.8O_y, still displays the structure of superfine powder as the individual SnO₂, Cu_0.5Sn_0.5O_y and Cu_0.6Sn_0.4O_y are obviously comprised of neat and ordered Caramel-Treats-like layers with mesoporous structure, as shown in Fig. 4(c and d), which is in good agreement with the N₂ adsorption–desorption results. For easy comparison, an SEM image of Cu_0.5Sn_0.5O_y with a lower magnification is also shown in Fig. S2.† The average layer thickness of Cu_0.5Sn_0.5O_y is about 100 nm, while that of Cu_0.6Sn_0.4O_y is only half, indicating the Cu/Sn ratio influences the layer thickness and sample morphology. For comparison purpose, a Cu rich sample, Cu_0.9Sn_0.1O_y, was particularly prepared and measured by SEM, with the image shown in Fig. 4(e). It is not surprise to see that this sample shows similar morphology to pure CuO, but consisting of uniform particles with much smaller size around 20 nm. It is believed that the mesoporous structure of Cu_0.5Sn_0.5O_y sample provides a large surface area and sufficient exposure of the active sites to the reactants, hence facilitating its CO oxidation activity remarkably.

Fig. 4 SEM images of Cu_xSn_1−xO_y catalysts. (a) SnO₂, (b) Cu_0.2Sn_0.8O_y, (c) Cu_0.5Sn_0.5O_y, (d) Cu_0.6Sn_0.4O_y, (e) Cu_0.9Sn_0.1O_y, (f) CuO.

The ultrafine structures of the typical samples were further investigated with HR-TEM method, with the results shown in Fig. 5. Pure CuO is comprised of irregular big particles, which shows exposed (−1 1 1) plane with a d-spacing of 0.25 nm. Pure SnO₂ consists of uniform fine round particles with an average size of 4.2 nm, which is slightly higher than its crystallite size measured by XRD. In addition, the exposed (110) and (101) planes with d-spacing of 0.32 and 0.26 nm for the pure SnO₂ sample are clearly observed, which is in good agreement with the values measured by XRD. In contrast, with the incorporation of Cu²⁺ cations into the crystal matrix of rutile SnO₂, the particle size of Cu_0.5Sn_0.5O_y becomes smaller. Moreover, the d-spacings of the exposed (110) and (101) planes of rutile SnO₂ phase in this sample expand to 0.34 and 0.27 nm, respectively. As mentioned above, a Cu²⁺ cation with a coordination number of 6 has a radius of 0.073 nm, which is higher than that of a Sn⁴⁺ cation with the same coordination number. The introduction of the bigger Cu²⁺ cations to replace part of the Sn⁴⁺ increase obviously the d values of the planes.

Fig. 5 HR-TEM images of Cu_xSn_1−xO_y catalysts. (a and b) SnO₂, (c and d) Cu_0.5Sn_0.5O_y, (e and f) CuO.

In summary, Cu_0.5Sn_0.5O_y sample exhibits the distinct crystallite planes corresponding to tetragonal rutile SnO₂ phase. No CuO facets are observed although it contains a large amount of CuO. Together with the d-spacing change, HR-TEM provides strong and extra evidence to prove the formation of Cu–Sn solid solution structure for the Cu_xSn_1−xO_y samples with suitable Cu/Sn atomic ratios.

H₂-TPR analysis on the redox behaviour of the catalysts

H₂-TPR experiments were performed to understand the redox behaviours of the Cu_xSn_1−xO_y catalysts, with the profiles shown in Fig. 6 and the H₂ consumption amount in Table S1.† Pure SnO₂ shows a major reduction peak at 588 °C, which can be assigned to the reduction of SnO₂ to metallic Sn.^24–26,31 In addition, a small shoulder peak at 255 °C is observed, which is due to the reduction of a small amount of surface deficient oxygen species of SnO₂.⁴⁵ The H₂-TPR profile of pure CuO shows two reduction peaks from 230 °C to 450 °C, which is attributed to the stepwise reduction of CuO into metallic Cu. The quantification result of O/Cu atomic ratio of 1.0/1.0 also confirms this. Interestingly, for all of the Cu_xSn_1−xO_y catalysts, three groups of reduction peaks can be distinctly observed below 250 °C, around 350 °C and above 450 °C in sequence. Compared with the reduction behaviours of the individual samples, the multiple reduction peaks below 250 °C can still be assigned to the reduction of Cu²⁺ species in the samples into metallic Cu. The H₂ consumption amount of this part is proportional to the CuO contents from Cu_0.15Sn_0.85O_y to Cu_0.6Sn_0.4O_y within the experimental error, which further confirms this assignment. In comparison, the reduction peak above 450 °C is assigned to the reduction of SnO₂ to metallic Sn, for which the H₂ consumption amount decreases with the SnO₂ contents from Cu_0.15Sn_0.85O_y to Cu_0.6Sn_0.4O_y. Since these Cu_xSn_1−xO_y catalysts are mesoporous with very high surface area, the reduction of both Sn⁴⁺ and Cu²⁺ in the solid solution structure of the samples becomes much easier, thus the reduction peaks migrate to lower temperature region.

Fig. 6 H₂-TPR profiles of Cu_xSn_1−xO_y samples. (a) SnO₂, (b) Cu_0.15Sn_0.85O_y, (c) Cu_0.2Sn_0.8O_y, (d) Cu_0.3Sn_0.7O_y, (e) Cu_0.4Sn_0.6O_y, (f) Cu_0.5Sn_0.5O_y, (g) Cu_0.6Sn_0.4O_y, (h) CuO.

What's interesting here is the presence of a small reduction peak for all of the Cu_xSn_1−xO_y solid solution samples at around 350 °C. As previously reported, the incorporation of heteroatoms into the lattice of rutile SnO₂ to form solid solution structure can induce lattice distortion.^25,31 In addition, if the dissolved cations have different valance states, charge imbalance will also take place. Both of these factors could lead to the formation of oxygen vacancies and more mobile oxygen species. The quantified H₂ consumption amount of this small peak for each catalyst is also listed in Table S1,† which increases with the increasing of the amount of Cu content in the samples until the Cu/Sn molar ratio reaches 0.5/0.5. Further increase the Cu amount in fact suppresses this oxygen species. Therefore, it is believed that this small peak particularly observed in Cu_xSn_1−xO_y catalysts is due to the reduction of loosely bonded oxygen species aroused by the dissolving of Cu²⁺ into SnO₂ lattice to form solid solution, which could be important for the CO oxidation activity of the catalysts. Indeed, the easier reduction of the lattice oxygen and the presence of this more mobile oxygen species could be another inherent reason leading to the significantly improved CO oxidation activity on the mesoporous Cu_xSn_1−xO_y catalysts.

XPS analysis on the surface composition of the catalysts

To further understand the surface feature of the Cu_xSn_1−xO_y solid solution catalysts, Cu_0.5Sn_0.5O_y, the typical and most active catalyst in this study, was thus analysed by XPS technique and compared with pure SnO₂ and CuO in Fig. 7. It is noted that all the binding energies are calibrated by using a C 1s internal standard.

Fig. 7 XPS measurements of the samples. (A) O 1s, (B) Cu 2p, (C) Sn 3d.

Fig. 7(A) compares the O 1s signals of all the samples. While pure CuO displays a very symmetrical O 1s peak at 531.8 eV, both of the pure SnO₂ and Cu_0.5Sn_0.5O_y show unsymmetrical O 1s peaks at 531.6 eV and 533.3 eV, respectively. This proves the presence of multiple types of oxygen species with different chemical environments on the surface of these two catalysts, which is in agreement with the H₂-TPR results. To gain deeper understanding of the surface oxygen properties, the O 1s peaks of these two samples are deconvoluted and shown in Fig. S3.† It has been reported previously that a deconvoluted O 1s peak with higher binding energy can be assigned to loosely bounded surface oxygen species (O_ads), and the peak with lower binding energy can be assigned to surface lattice oxygen species (O_lat).²⁵ Based on this, the ratios of the O_ads/O_lat of the pure SnO₂ and Cu_0.5Sn_0.5O_y are calculated and listed in Table 3. Apparently, on the surface of Cu_0.5Sn_0.5O_y sample, 30% more loosely bounded oxygen species is present in comparison with pure SnO₂, in good agreement with the H₂-TPR results. For CO oxidation, the presence of this active surface oxygen species can obviously benefit the activity of the Cu–Sn solid solution catalysts.

Table 3 O_ads/O_lat ratio calculated from XPS measurements

Catalyst	Oads/Olat	Cu/Sn atomic ratio
SnO2	0.75	—
Cu0.5Sn0.5Oy	0.97	0.78

---

As shown in Fig. 7(B), pure CuO sample shows two peaks at 933.2 eV and 953.5 eV, which are assigned to Cu 2p_3/2 and Cu 2p_1/2 in sequence. For Cu_0.5Sn_0.5O_y sample, these two peaks shift to higher binding energy, indicating the change of the chemical environment of Cu due to solid solution formation. Both of the samples show a shake-up peak at about 940–945 eV, which is typical for Cu²⁺ according to the literature,⁴⁶ testifying that in the solid solution samples, Cu is fully oxidized.

Sn 3d spectra of pure SnO₂ and Cu_0.5Sn_0.5O_y are compared in Fig. 7(C). Two distinguished peaks at 486.5 and 495.1 eV, which are attributed to Sn 3d_3/2 and Sn 3d_5/2 in sequence, can be observed for pure SnO₂ and were attributed to Sn⁴⁺. The two same peaks are also observed for Cu_0.5Sn_0.5O_y sample, but shifting to higher binding energies, indicating the change of the chemical environment of Sn due to solid solution formation. However, the binding energy is characteristic for Sn⁴⁺, which excludes the possibility for the presence of lower valence state Sn species. The calculated surface Cu/Sn atomic ratio of Cu_0.5Sn_0.5O_y listed in Table 3 is 0.78, which is less than 1.0, the bulk ratio of the sample, indicating the surface of the sample is Sn rich. This provides extra evidence to prove that Cu²⁺ cations have been incorporated into the lattice of tetragonal rutile SnO₂ to form solid solution structure.

Discussion

Over the past decades, the preparation of nano sized mesoporous metal oxide catalysts with high surface area and various morphologies has attracted much attention and is a hot research topic. Many publications can be found on synthesizing different metal oxides with mesopores and particular morphologies. It has been reported that compared with regular nano powders, metal oxides with special morphologies have quite some advantages, such as higher surface areas, larger pore volumes and pore sizes and more facile oxygen species etc., which generally improves the catalytic performance of the resulted catalysts.^10,47 For the synthesis of mesoporous SnO₂, several methods have been reported in the literature such as sol–gel,⁴⁸ sonochemical,⁴⁹ hydrothermal,⁵⁰ microwave hydrothermal⁵¹ and template-assisted methods.⁵² With these methods, the preparation procedures are generally complicated and time consuming, and involve various templates. In this study, for the first time, a series of mesoporous Cu_xSn_1−xO_y catalysts with high surface areas have been synthesized easily by a simple co-precipitation method without using any template. As mentioned above, all of these Cu_xSn_1−xO_y catalysts possess significantly improved CO oxidation activity in comparison with the individual SnO₂ and CuO.

XRD, SEM-EDX mapping and HR-TEM results testified that Cu²⁺ cations have been dissolved into the lattice of tetragonal rutile SnO₂ with certain capacity to form solid solution structure. Due to lattice distortion and charge imbalance, more active oxygen species can be formed in these Cu–Sn solid solution catalysts, as evidence by H₂-TPR and XPS results. It is commonly accepted that CO oxidation over metal oxide catalysts generally follows Mars-van Krevelen mechanism, through which the CO is primarily oxidized by the lattice oxygen or any facile surface oxygen species of the metal oxide. In turn, the consumed oxygen species will be regenerated by the gas-phase oxygen.⁵³ Therefore, the formation of more active oxygen species is favourable for the CO oxidation activity of the catalysts, which could indeed be one of the major reasons accounting for the superior CO oxidation activity of the Cu_xSn_1−xO_y catalysts.

Furthermore, N₂ adsorption–desorption proves that these Cu_xSn_1−xO_y catalysts possess high surface areas and pore volumes due to the presence of mesoporous structure. In general, the large specific surface area of a metal oxide catalyst favours the dispersion of its active phase, and the porous structure and larger pore volumes can facilitate the diffusion of the reactants and products, hence promoting the catalytic activity remarkably. Therefore, the mesoporous structure, improved pore volumes and high surface areas of the Cu_xSn_1−xO_y catalysts are believed to be other predominant factors leading to their improved CO oxidation activity.

To further confirm the formation of solid solution and mesoporous structure is important for the CO oxidation activity of the catalysts, other two samples having the same amount of CuO to Cu_0.5Sn_0.5O_y were prepared by impregnating Cu(NO₃)₂ solution onto the regular non-porous SnO₂ powder calcined at 300 °C or simply dried at 110 °C, which are named as IMP-300 and IMP-110, respectively. Prior to activity test, the two catalysts were also calcined at 300 °C for 4 hours. As shown by Fig. S4,† the CO oxidation activity of these two samples is obviously lower than that of Cu_0.5Sn_0.5O_y.

It is particularly noted here that Cu_0.5Sn_0.5O_y, the most active catalyst in this study, consists of neat Caramel-Treats-like layers with mesoporous structure and a thickness of ∼100 nm. It has the highest surface area and pore volumes among all of the catalysts. In addition, H₂-TPR results indicate that on its surface, the largest amount of loosely bounded oxygen species (0.6 mmol g⁻¹) has been formed. As a result, this catalyst exhibits the highest CO oxidation activity among all of the catalysts. The last but not the least, Cu_0.5Sn_0.5O_y demonstrates also potent water resistance to water deactivation, implying it has the possibility to be potentially used in some real exhaust cleaning processes.

Conclusions

In this study, a series of high surface area mesoporous Cu_xSn_1−xO_y mixed oxide catalysts with different Cu/Sn molar ratios have been successfully fabricated by a simple co-precipitation method without using any template. Compared with the individual SnO₂ and CuO, CO oxidation activity on these catalysts is remarkably improved. The highest activity is achieved on Cu_0.5Sn_0.5O_y, a catalyst with a Cu/Sn molar ratio of 0.5/0.5, on which CO can be completely oxidized at 140 °C. It is revealed by XRD, SEM-EDX mapping and HR-TEM results that Cu²⁺ cations have been incorporated into the crystal lattice of tetragonal rutile SnO₂ to form uniform solid solution structure. As testified by N₂ adsorption–desorption and SEM results, these Cu_xSn_1−xO_y catalysts contain well-defined mesopores, possess high surface areas and high pore volumes, which are favourable for the dispersion of the active sites, the diffusion of the reactants and the easy interaction between the reactants and the catalyst surface. Moreover, due to lattice distortion and charge imbalance, more active loosely bounded oxygen species has been formed on the surface of the Cu_xSn_1−xO_y catalysts, as evidenced by H₂-TPR and XPS results. It is believed that these are the predominant reasons leading to the superior CO oxidation activity for the Cu_xSn_1−xO_y catalysts. Cu_0.5Sn_0.5O_y, the most active catalyst in this study, consists of neat Caramel-Treats-like layers with mesoporous structure. It has also the highest surface area, pore volumes and the largest amount of loosely bounded surface oxygen species among all of the catalysts. As a result, it exhibits the highest CO oxidation activity. Notably, these Cu_xSn_1−xO_y catalysts show potent resistance to water vapour deactivation, indicating they have the potential to be used in real exhaust control processes.

Acknowledgements

This work is supported by the Chinese Natural Science Foundation (21263015, 21203088), the Education Department of Jiangxi Province (GJJ14205, KJLD14005) and the Natural Science Foundation of Jiangxi Province (20142BAB213013), which is greatly acknowledged by the authors.

Notes and references

1. E. C. Njagi, C. Chen, H. Genuino, H. Galindo, H. Huang and S. L. Suib, Appl. Catal., B, 2010, 99, 103–110.
2. A. B. Lamb, W. C. Bray and J. C. W. Frazer, J. Ind. Eng. Chem., 1920, 12, 213–221.
3. A. S. Ivanova, E. M. Slavinskaya, R. V. Gulyaev, V. I. Zaikovskii, O. A. Stonkus, I. G. Danilova, L. M. Plyasova, I. A. Polukhina and A. I. Boronin, Appl. Catal., B, 2010, 97, 57–71.
4. N. Li, Q. Chen, L. Luo, W. Huang, M. Luo, G. Hu and J. Lu, Appl. Catal., B, 2013, 142–143, 523–532.
5. Z. Wang, L. Li, D. Han and F. Gu, Mater. Lett., 2014, 137, 188–191.
6. M. M. Gadgil, R. Sasikala and S. K. Kulshreshtha, J. Mol. Catal., 1994, 87, 297–309.
7. N. Yamaguchi, N. Kamiuchi, H. Muroyama, T. Matsui and K. Eguchi, Catal. Today, 2011, 164, 169–175.
8. P. Kast, G. Kučerová and R. J. Behm, Catal. Today, 2015, 224, 146–160.
9. M. Luo, P. Fang, M. He and Y. Xie, J. Mol. Catal. A: Chem., 2005, 239, 243–248.
10. J. Cao, Y. Wang, X. Yu, S. Wang, S. Wu and Z. Yuan, Appl. Catal., B, 2008, 79, 26–34.
11. K. Li, Y. Wang, S. Wang, B. Zhu, S. Zhang, W. Huang and S. Wu, J. Nat. Gas Chem., 2009, 18, 449–452.
12. J. Huang, Y. Kang, L. Wang, T. Yang, Y. Wang and S. Wang, Catal. Commun., 2011, 15, 41–45.
13. M. Y. Kang, H. J. Yun, S. Yu, W. Kim, N. D. Kim and J. Yi, J. Mol. Catal. A: Chem., 2013, 368–369, 72–77.
14. D. Mrabet, A. Abassi, R. Cherizol and T.-O. Do, Appl. Catal., A, 2012, 447–448, 60–66.
15. H. Zou, S. Chen, Z. Liu and W. Lin, Powder Technol., 2011, 207, 238–244.
16. G. Aguila, S. Guerrero and P. Araya, Appl. Catal., A, 2013, 462–463, 56–63.
17. M. Monte, D. Gamarra, A. López Cámara, S. B. Rasmussen, N. Gyorffy, Z. Schay, A. Martínez-Arias and J. C. Conesa, Catal. Today, 2014, 229, 104–113.
18. Z. Wu, H. Zhu, Z. Qin, H. Wang, J. Ding, L. Huang and J. Wang, Fuel, 2013, 104, 41–45.
19. S. Zeng, W. Zhang, M. Śliwa and H. Su, Int. J. Hydrogen Energy, 2013, 38, 3597–3605.
20. D. Qiao, G. Lu, Y. Guo, Y. Wang and Y. Guo, J. Rare Earths, 2010, 28, 742–746.
21. K. Xu, D. Zeng, S. Tian, S. Zhang and C. Xie, Sens. Actuators, B, 2014, 190, 585–592.
22. W. Zeng, Y. Li, B. Miao, L. Lin and Z. Wang, Sens. Actuators, B, 2014, 191, 1–8.
23. X. Han, X. Xu, W. Liu, X. Wang and R. Zhang, Solid State Sci., 2013, 20, 103–109.
24. X. Wang, L. Xiao, H. Peng, W. Liu and X. Xu, J. Mater. Chem. A, 2014, 2, 5616.
25. X. Xu, R. Zhang, X. Zeng, X. Han, Y. Li, Y. Liu and X. Wang, ChemCatChem, 2013, 5, 2025–2036.
26. J. Yu, D. Zhao, X. Xu, X. Wang and N. Zhang, ChemCatChem, 2012, 4, 1122–1132.
27. J. Liu, H. Peng, W. Liu, X. Xu, X. Wang, C. Li, W. Zhou, P. Yuan, X. Chen, W. Zhang and H. Zhan, ChemCatChem, 2014, 6, 2095–2104.
28. X. Wang, J. S. Tian, Y. H. Zheng, X. L. Xu, W. M. Liu and X. Z. Fang, ChemCatChem, 2014, 6, 1604–1611.
29. Y. Ma, X. Wang, X. You, J. Liu, J. Tian, X. Xu, H. Peng, W. Liu, C. Li, W. Zhou, P. Yuan and X. Chen, ChemCatChem, 2014, 6, 3366–3376.
30. X. Wang and Y. C. Xie, Z. Phys. Chem., 2002, 216, 919–929.
31. X. Wang and Y. Xie, Appl. Catal., B, 2001, 35, 85–94.
32. X. Wang and Y. C. Xie, New J. Chem., 2001, 25, 1621–1626.
33. S.-W. Choi, A. Katoch, J. Zhang and S. S. Kim, Sens. Actuators, B, 2013, 176, 585–591.
34. F. Shao, M. W. G. Hoffmann, J. D. Prades, R. Zamani, J. Arbiol, J. R. Morante, E. Varechkina, M. Rumyantseva, A. Gaskov, I. Giebelhaus, T. Fischer, S. Mathur and F. Hernández-Ramírez, Sens. Actuators, B, 2013, 181, 130–135.
35. S. H. Choi and Y. C. Kang, Nanoscale, 2013, 5, 4662–4668.
36. H. El-Shinawi, M. Böhm, T. Leichtweiß, K. Peppler and J. Janek, Electrochem. Commun., 2013, 36, 33–37.
37. T. S. Zhang, L. B. Kong, X. C. Song, Z. H. Du, W. Q. Xu and S. Li, Acta Mater., 2014, 62, 81–88.
38. X. Zheng, Y. Wei, L. Wei, B. Xie and M. Wei, Int. J. Hydrogen Energy, 2010, 35, 11709–11718.
39. C. J. Zhou, W. Lin, Y. X. Zhu and Y. C. Xie, Chin. J. Catal., 2003, 24, 229–232.
40. M. S. Cui, L. J. Ma, C. Jia, N. Zhao, X. W. Huang and A. Chen, Chin. J. Rare Met., 2007, 31, 520–525.
41. M. J. Fuller and M. E. Warwick, J. Chem. Soc., Chem. Commun., 1973, 210a.
42. Y. Wang, S. Luo, Z. Wang and Y. Fu, Appl. Clay Sci., 2013, 80–81, 334–339.
43. A. R. Denton and N. W. Ashcroft, Phys. Rev. A, 1991, 43, 3161–3164.
44. M. Luo, J. Ma, J. Lu, Y. Song and Y. Wang, J. Catal., 2007, 246, 52–59.
45. N. M. G. R. Sasikala and S. K. Kulshreshtha, Catal. Lett., 2001, 71, 69–73.
46. M. S. P. Francisco, V. R. Mastelaro, P. A. P. Nascente and A. O. Florentino, J. Phys. Chem. B, 2001, 105, 10515–10522.
47. G. K. Pradhan, K. H. Reddy and K. M. Parida, Catal. Today, 2014, 224, 171–179.
48. D. M. Antonelli and J. Y. Ying, Angew. Chem., Int. Ed. Engl., 1995, 34, 2014–2017.
49. L. Zhang and J. C. Yu, Chem. Commun., 2003, 2078–2079.
50. Y. Wang, H. Xu, X. Wang, X. Zhang, H. Jia, L. Zhang and J. Qiu, J. Phys. Chem. B, 2006, 110, 13835–13840.
51. L. Man, J. Zhang, J. Wang, H. Xu and B. Cao, Particuology, 2013, 11, 242–248.
52. N. Lakshminarasimhan, E. Bae and W. Choi, J. Phys. Chem. C, 2007, 111, 15244–15250.
53. V. P. Santos, M. F. R. Pereira, J. J. M. Órfão and J. L. Figueiredo, Appl. Catal., B, 2010, 99, 353–363.

---

Footnotes

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra00635j

‡ These authors contributed equally to this work.

---