Introduction

Interface chemistry has become a relevant field of study in recent years1, 2, 3, 4, 5, 6 because catalytic reactions constantly take place on the surface of catalysts. Therefore, surface states, such as the packing mode of surface atoms, the exposed facets and surface charges, have an important role in determining the feasibility and duration of the reaction. With rapid advances in chemical science, there are many nanomaterials available with well-defined sizes, shapes, crystal facets, structures and compositions.7, 8, 9, 10, 11, 12, 13, 14 There have been increasing efforts to design and control the interface chemistry between different types of components. This is because two-phase interfaces could render hybrid junctions with rich redox reactions, which are beneficial for enhancing the catalytic performance. However, four requirements must be satisfied simultaneously for successful formation of an effective interface: (1) the components should be in full contact; (2) there should be a strong synergistic effect between the two components; (3) to optimize the catalytic properties, the interface areas should be as numerous as possible (a core@shell superstructure seems a good choice); and (4) the richer the redox relationship between the two components is, the better the catalytic properties that they will show. Following the above consideration, the pair formed by the oxidative CeO2 and the reductive Cu2O seems ideal, and their hybrid nanostructures are thus desired to obtain highly active interfaces.15, 16

Unfortunately, few works have been reported on the successful hybridization of CeO2 and Cu2O components via the classic two-step process. The main reason for the lack of successful hybridization is that the preformed Cu2O cores cannot withstand the conditions during the coating of CeO2 in the following step (Cu2O could be severely etched by ammonium, which is indispensable during the CeO2 coating process). However, recently, an important one-step synthetic strategy called ‘clean synthesis’17, 18, 19, 20, 21, 22, 23 inspired us to solve this problem. Herein, ‘clean’ means not only free of toxic raw materials, especially organic compounds, but also low-cost and low-temperature aqueous synthesis, a surface free of residues and a greatly simplified post-treatment process, which are extremely valuable for practical applications. In this strategy, Ce(OH)3 and unsaturated oxides, such as WO2.72, are the most reported reductant precursors due to their strong reducing capability, whereas noble metal ions, such as Ag(NH3)2+ and PtCl42−, serve as the oxidant precursors.18, 21, 22, 23 These clean synthesis routes are mainly based on the redox principle: an auto-catalytic redox reaction followed by an spontaneous self-assembly process. During the reaction, mutual inhibition between different components generates final products with a uniform structure in the nanometer scale, uniform metal dispersion, tunable metal particle size and narrow metal particle size distribution. Owing to their clean surfaces, the unique nanostructures and their strongly coupled two-phase interface exhibited excellent catalytic activities and high-temperature stabilities compared with those samples prepared with classic two-step processes.

Our group has continuously focused on the fabrication of CeO2-based, multi-component, hybrid nanostructures.24, 25, 26, 27 Notice that in alkaline solution, Cu(OH)2 has a strong oxidation ability, which has been successfully used as Fehling’s solution to detect reducing sugar. In addition, in our previous study, it is found that Ce(OH)3 could serve as a reducer in aqueous solution. In this work, following the above consideration, a two-phase interface reaction has been triggered only in water without the use of any organic components.

Experimental procedure

Synthesis of 155 nm Cu2O@CeO2 cubes

20 ml of H2O was bubbled with Ar gas for 5 min to remove O2. Then, 1 ml of a 0.1 M Ce(NO3)3 aqueous solution was added, followed by 0.5 ml of 1 M NaOH. Finally, 1 ml of 0.1 M Cu(NO3)2 was added at 40 °C. The solution was protected by Ar gas. The products were purified by centrifugation at 8000 r.p.m. and then dispersed in 20 ml of water. The size-separation process was carried out by centrifuging at 3500–4000 r.p.m. for 5 min; this was repeated several times to separate the free CeO2 nanoparticles (NPs) from the solution.

Large-scale synthesis

The large-scale synthesis was similar to the above procedure but with larger volumes (200 times) of the solutions and using the raw material.

Synthesis of CeO2–Cu2O branches

First, 20 ml of H2O was bubbled with Ar for 5 min to remove O2. Then, 1 ml of 0.1 M Ce(NO3)3 was added, followed by 0.5 ml of 1 M NaOH. Finally, 0.5 ml of Cu(NO3)2 was added. The solution was protected by Ar and kept at 40 °C for 1 h.

Control (1)

The synthesis was extremely similar to that of 155 nm CeO2–Cu2O cubes except for using 0.7 ml of Cu(NO3)2 instead of 1 ml of Cu(NO3)2.

Control (2)

The synthesis was very similar to that of 155 nm CeO2–Cu2O cubes except for using 0.7 ml of NaOH instead of 0.5 ml of NaOH.

Synthesis of 68 nm Cu2O@CeO2 cubes

The synthesis process was similar to that of 155 nm CeO2–Cu2O cubes except the temperature was increased to 50 °C.

Synthesis of 44 nm Cu2O@CeO2 cubes

The synthesis process was similar to that of 155 nm CeO2–Cu2O cubes except the temperature was increased to 60 °C.

Synthesis of 32 nm Cu2O@CeO2 cubes

The synthesis process was similar to that of 155 nm CeO2–Cu2O cubes except the temperature was increased to 70 °C.

Characterization

X-ray diffraction (XRD) patterns of the products were collected on a Rigaku-D/max 2500 V X-ray diffractometer (Rigaku company, Tokyo, Japan) with Cu-Kα radiation (λ=1.5418 Å) and voltage and current of 40 kV and 40 mA, respectively. Transmission electron microscopic (TEM) images were obtained with a TECNAI G2 high-resolution TEM operating at 200 kV. The catalytic performances of the catalysts were monitored on-line using gas chromatography (GC9800).

Catalytic test

CO oxidation

For this reaction, 25 mg of catalyst was placed in a stainless-steel reaction tube. The experiment was performed under the flow of a reactant gas mixture (1% CO, 20% O2, balance N2) at a rate of 30 ml min−1.

3,3,5,5-Tetramethylbi-phenyl dihydrochloride oxidation

First, 100 μl of 30% H2O2 and 100 μl of 0.01 M 3,3,5,5-Tetramethylbi-phenyl dihydrochloride (freshly prepared) were added to 3 ml of a buffer solution of pH=5.5. Then, 20 μl of a NP suspension was added to the above mixture. The oxidation reaction progress was monitored at 40 °C.

Results and Discussion

The XRD analyses of the Cu2O–CeO2 hybrid samples are shown in Supplementary Figure 1. The peaks at 2θ=28.5°, 33.0°, 47.4° and 56.3° can be indexed to the (111), (200), (220) and (311) CeO2 fluorite-phase reflections (JCPDS No. 34-0394), respectively. The other two strong peaks at 2θ=36.5° and 42.4° correspond to the pure (111) and (200) Cu2O reflections (JCPDS No. 65-3288), respectively. There is an obvious broadening in the diffraction peaks of CeO2, indicating a small particle size in both samples. The particle size of CeO2, estimated using the Scherrer equation and the XRD patterns, is ~5 nm, which is in agreement with the TEM analysis. The compositions of CeO2 and Cu2O in the Cu2O–CeO2 hybrid samples and their corresponding UV-Vis absorption spectra have been shown in Supplementary Table 1 and Figure 2, respectively.

TEM images show that the as-obtained Cu2O–CeO2 hybrids are sub-155-nm (Figure 1a) uniform core@shell cubes with a rough surface (Figure 1b) when 1.0 ml of the 0.1 M Cu(NO3)2 aqueous solution (molar ratio of Cu/Ce/OH−=1/1/5) was added (see the Experimental procedure section for details). A thin CeO2 layer can be observed clearly on each Cu2O cube core. The layer is composed of hundreds of self-assembled, 5-nm CeO2 NPs, which could be easily distinguished in the high-resolution transmission electron microscopy image (Figure 1c). The lattice spacing of 0.27 nm corresponds well to the characteristic (200) planes of the fluorite phase of CeO2, and the 0.21 nm lattice spacing corresponds to the characteristic (200) planes of the Cu2O cubic phase, in agreement with XRD results. Further scanning transmission electron microscopy-energy dispersive X-ray spectroscopy analyses have also been done in order to confirm that nature of the core@shell hybrid nanostructure. With the data shown in Figures 1d–g, Cu is present only in the core position, and Ce is distributed everywhere. Here, it should be noticed that our clean and fast method can easily achieve mass production because 3.7218 g of the Cu2O@CeO2 core@shell cubes can be obtained in one pot (Figure 1b, inset). Brunauer–Emmett–Teller analysis was performed, and the data are shown in Supplementary Figure 7. The Brunauer–Emmett–Teller surface area of Cu2O–CeO2 hybrid branches is 151.7 m2 g−1, which is greater than those for Cu2O@CeO2 cubes (104.2 m2 g−1), commercial Cu2O (99.2 m2 g−1) and CeO2 (89.9 m2 g−1) NPs.

Figure 1
figure 1

TEM images of Cu2O@CeO2 cubes (a–c). The size distribution histogram (a, inset); scanning transmission electron microscopy-energy dispersive X-ray spectroscopy analyses of Cu2O@CeO2 cubes (d–g); TEM images of Cu2O–CeO2 branches (h–k). The photo (i, inset) is of the rime (crust) on branches. All of these TEM images are taken on Mo mesh. TEM, transmission electron microscope.

Full size image

The observed color changes give an indication of the whole reaction process. After the Cu(NO3)2 aqueous solution was injected, the solution immediately became blue, a characteristic color for Cu(OH)2. The redox reaction between Ce(OH)3 and Cu precursors did not occur until the reaction was heated to 40 °C. The reaction was evident by the slow color change to yellow, a typical color for Cu2O. Obviously, the reaction underwent two steps: the nucleation of the Cu precursor at the beginning, followed by the redox reaction between the Ce(OH)3 and Cu precursors.

The amount of added Cu2+ ions was found to have a strong influence on the final hybrid structures. As shown in Table 1, when only 0.7 ml of the Cu(NO3)2 (0.1 M) aqueous solution (Cu/Ce/OH−=0.7/1/5) was added, the uniform cube-like structure disappeared, and the obtained products were thus disordered (Control (1), Supplementary Figure 5). However, a new type of branch-like structure appeared, instead of the cubes and the disordered products, when we further reduced the amount of Cu(NO3)2 to 0.5 ml (Cu/Ce/OH−=0.5/1/5); similar 5 nm CeO2 NPs stick to the Cu2O components, as shown in Figures 1h and i. In Figures 1j and k, these two components can be distinguished clearly in the branches. The lattice spacings of 0.27 and 0.21 nm correspond well to the characteristic (200) CeO2 planes and (200) Cu2O planes, respectively; these peaks are also observed in cubes. To further confirm these two components, an etching experiment with HCl was performed. As a result, the Cu2O component was removed in both cubes and branches under ambient conditions, and only the 5 nm CeO2 NPs were observed in the final products (Supplementary Figures 3 and 4).

Table 1 Summary of the synthetic conditions of CeO2–Cu2O hybrids with different nanostructures
Full size table

The amount of NaOH added also significantly influenced the hybrid structure. In the Control (2) experiment, increasing the amount of NaOH (Cu/Ce/OH−=1/1/7) produced neither cubes nor branches (Supplementary Figure 6). Notice that copper hydroxide has been identified as amphoteric; therefore, it could be redissolved by reacting with excess OH− to form Cu(OH)42−:

Furthermore, successful synthesis of Cu2O@CeO2 cubes requires the Cu/Ce/OH− ratio to be kept at 1/1/5, where the Cu(OH)2 precipitation acts as the Cu precursor. In the case of Cu2O–CeO2 branches, the Cu/Ce/OH− ratio equals 0.5/1/5 and Cu(OH)42− acts as the Cu precursor. This indicates that the different species of Cu precursors should be responsible for the formation of different hybrid structures. The schematic structural formation of Cu2O–CeO2 hybrids is depicted in Scheme 1.

After careful consideration, the conclusion was reached that nucleation of the precursor and the redox process might occur simultaneously only when the reaction temperature is higher than 40 °C. Consequently, a series of control experiments were performed by changing the reaction temperature with a constant ratio of Cu/Ce/OH− at 1/1/5. The synthetic conditions are summarized in Table 1. As shown in Figures 2a–d, when injecting Cu(NO3)2 at 50 °C, a similar cube-like structure was obtained, but the cube size dramatically decreased to ~68 nm on average; increasing the injection temperature to 60 and 70 °C led to the formation of smaller sub-44 and -32-nm cubes, respectively. With the decrease in size of the hybrid particle, more branched structures were observed. Meanwhile, the corners of the cubes became smoother as the reaction proceeds. However, the dimension of CeO2 NPs remained unchanged with different sized cubes. The dependence of the size distribution histogram with reaction temperature is shown in Figure 2l, and the whole kinetics-controlled process is depicted in Figure 2m. The changes in the cube size can be explained in the following way. If the nucleation and redox process occur at nearly the same time, Cu2O cubes are well coated by CeO2 components right after nucleation and before they continue growing. A higher reaction temperature brings stronger inhibitory effects, which are accompanied by aggregation of naked CeO2 NPs that increase the CeO2 coating. That is why the size of Cu2O cubes and the overall hybrid sizes dramatically decrease. Meanwhile, the strong aggregation tendency of CeO2 NPs, shown in Figures 2a, e and i (marked by arrows), complicates the separation of the Cu2O@CeO2 cubes from the reaction system.

Figure 2
figure 2

TEM images of Cu2O@CeO2 cubes obtained via the kinetics-controlled process. Injection of Cu(NO3)2 at 50 °C (a–d), 60 °C (e–h) and 70 °C (i–l); the relationship between injection temperature and particle size (m). From the TEM images, it is easy to see that at higher temperatures, more and more branched structures of CeO2 on Cu2O cores are observed and the corners of the cubes become smoother. TEM, transmission electron microscope.

Full size image

In Figure 3a, two broad H2-temperature programmed reduction peaks observed at 390 and 760 °C for CeO2 can be attributed to the reduction of surface capping oxygen and bulk oxygen of CeO2, respectively.28, 29 Comparatively, the commercial Cu2O sample is highly active, and its first peak appears at ~243 °C. When the two composites, CeO2 and Cu2O, are hybridized, the first peak move toward a lower temperature. By careful observation, we found that the temperature of the first reduction peak of the Cu2O@CeO2 cubes is 204 °C, and the temperature of the first reduction peak of the Cu2O–CeO2 branches is 227 °C. The disappearance of the two peaks at 243 and 390 °C has confirmed the existence of synergistic effects between CeO2 and Cu2O. Furthermore, O2-temperature programmed reduction oxidation was used to examine the mobility of the lattice oxygen of oxides. It is seen in Figure 3b that commercial Cu2O exhibits extremely limited activity and that CeO2 has a broad peak appearing at 494 °C. According to previous reports, there are two types of oxygen species for oxides: the oxygen adsorbed on the surface and bulk lattice oxygen. The peak at 494 °C should be related to the chemically adsorbed oxygen adjacent to oxygen vacancies.28 No desorption peak appeared at higher temperatures, indicating that the bulk lattice oxygen is extremely stable in an inert atmosphere. In comparison, the formation of Cu2O–CeO2 hybrids decreases the starting temperature of O2 desorption and the chemically adsorbed oxygen adjacent to the oxygen vacancies. In combination with the H2-temperature programmed reduction and O2-temperature programmed reduction oxidation results, Cu2O could have an important role in activating the oxygen vacancies of CeO2 so that the two-phase interface becomes an excellent active region for the catalytic reactions.

Figure 3
figure 3

H2-TPR (a) and O2-TPD (b) profiles of Cu2O@CeO2 core@shell cubes, Cu2O–CeO2 branches, pure CeO2 and Cu2O. a.u., arbitrary unit; H2-TPR, H2-temperature programmed reduction; O2-TPD, O2-temperature programmed reduction oxidation.

Full size image

The catalytic performance of these samples was evaluated first by the model reaction of catalytic CO oxidization. Figure 4 shows the curves of CO conversion as a function of temperature from 20 to 400 °C for samples with Cu2O@CeO2 cubes (155 nm), Cu2O–CeO2 branches, commercial CeO2 (20–50 nm, Aladdin Company, Shanghai, China) and Cu2O (Aladdin Company). The temperature T100 at which CO is 100% converted follows the sequence: Cu2O@CeO2 cubes (120 °C)<Cu2O–CeO2 branches (175 °C)<commercial CeO2 (20–50 nm, 400 °C)<commercial Cu2O (limited activity in the range of testing temperatures). Obviously, Cu2O@CeO2 cubes present the highest catalytic activity among the samples. Moreover, the activity of Cu2O@CeO2 cubes is much higher than those of the most typical CuO–CeO2,30, 31 Cu2O@CeO2,15 Pt–CeO2,32 and our previously reported Pt@CeO227 and Co3O4@CeO233 systems. Moreover, at the working temperature of 350 °C, the Ce–Cu cubes and branches remain at 100% yield of CO2 continuously for up to 10 h (Supplementary Figure 9). We also tested the catalytic performance of the as-obtained Cu2O@CeO2 cubes at the working temperature of 120 °C. As shown in Supplementary Figure 10, after continuous reaction for 20 h, the sample could convert over 92% CO to CO2. This indicates that the as-prepared Ce–Cu nanocatalysts have good thermal stability and excellent catalytic activity, exhibiting the great potential to replace Pt–CeO2 catalysts.

Figure 4
figure 4

CO conversion curves of Cu2O@CeO2 cubes (155 nm), Cu2O–CeO2 branches, commercial CeO2 and commercial Cu2O (T100 is the 100% conversion temperature of CO into CO2).

Full size image

The peroxidase-like activity of Ce–Cu cubes and branches was evaluated for catalysis of the peroxidase substrate 3,3,5,5-Tetramethylbi-phenyl dihydrochloride in the presence of H2O2. In this reaction, we chose a buffer solution with pH=5.5 and a reaction temperature of 40 °C. The initial oxidation rate is evaluated by monitoring the absorbance increase of the oxidized products of 3,3,5,5-Tetramethylbi-phenyl dihydrochloride at 652 nm (Figure 5). Obviously, the as-obtained Cu2O@CeO2 core@shell cubes show a much higher activity than the four given samples during the catalytic oxidation of 3,3,5,5-Tetramethylbi-phenyl dihydrochloride. The Cu2O–CeO2 branches also show better catalytic performance than the commercial Cu2O and CeO2. The previous reports have shown that the activities of transition metal oxide material-based peroxidase mimics were mostly derived from the metal ions that have a catalytic activity using H2O2 as a substrate through a mechanism similar to that of the Fenton reaction.34, 35 Self and co-workers found that cerium ions could also perform a Fenton-like reaction:35, 36

Figure 5
figure 5

3,3,5,5-Tetramethylbi-phenyl dihydrochloride oxidation curves of Cu2O@CeO2 core@shell cubes, Cu2O–CeO2 branches, pure CeO2 and Cu2O.

Full size image

It is obvious that the existence of Ce3+ is the key to triggering the reaction. The predictable result is that the higher concentration of Ce3+ ions in the reaction solution is, the faster the catalytic reaction should be. For the Cu2O–CeO2 hybrid system, the large amount of Cu+ could react first with H2O2 to form HO2−. Therefore, before reaction (1) started, the following reaction (4) occurred:

followed by

The largely increased amount of Ce3+ might be the main reason for the improved, intrinsic, enzyme mimetic activity.

In sum, we have successfully developed a new, non-organic route for the facile, fast and low-cost synthesis of environment-friendly Cu2O–CeO2 hybrids with well-controlled nanostructures. Without the addition of any organic reducing agents or surfactants, the whole reaction is a clean redox process between a Cu precursor and Ce(OH)3 in an aqueous solution. The formation of the final hybrid nanostructures depends on the initial species of the Cu precursors. The Cu(OH)2 precursor species yield uniform Cu2O cubes with a CeO2 surface coating, whereas Cu(OH)42− precursor species lead to asymmetrical Cu2O–CeO2 branches. Such an interface engineering strategy favors strong coupling between the two components of CeO2 and Cu2O. In the following CO oxidation and peroxidase-like activity tests, the Cu2O@CeO2 core@shell cubes exhibited a rather high catalytic activity; we hypothesize that the existence of highly active interfaces is the main reason for this activity. Our new synthetic process opens a new window for better design and preparation of noble metal-free nanocatalysts, which could have great applications in catalysis, gas-sensors and bio-chemistry, as well as areas related to energy and the environment.

scheme 1

The preparation process of Cu2O–CeO2 hybrids with different Cu/OH− molar ratios at 40 °C.

Supporting Information

XRD patterns of the as-obtained Cu2O–CeO2 branches and cubes; Ultraviolet–visible absorption spectra of Cu2O–CeO2 hybrids; TEM images of Cu2O–CeO2 cubes and branches after HCl etching; TEM images of the control experiment; and ICP analysis results.