Contents
Introduction
Experimental
Materials Synthesis of carbon spheres templates Synthesis of CSs@Pd@ZrO2 precursor Synthesis of Pd@HS-ZrO2 hollow spheres and Pd/ZrO2 catalysts Characterizations Catalytic performance measurements
Results and discussion
Conclusions
Acknowledgements
Electronic supplementary information
References
1 Introduction
Supported noble metal catalysts play a core role in the discovery of new technologies for sustainable development and energy conversion [1]. They are widely used in energy conversion, waste gas purification, wastewater treatment, and other vital fields. In order to improve the catalytic activity, highly dispersed noble metal nanoparticles are usually needed. However, due to the high surface energy and thermodynamically unstable state of nanoparticles, they are prone to sintering or agglomeration during high temperature service, which eventually leads to passivation and deactivation of the noble metal [2–3]. In recent decades, various methods have been proposed to solve the sintering problem of supported noble metal catalysts, which can be briefly divided into two categories [4–6]. The first method is to enhance the interfacial interactions between the noble metal particle and the support, by which the nanoparticle is firmly anchored on the support [7–9]. The most classical method is to construct strong metal and support interaction (SMSI) [10–11]. For example, in the work of Liu et al. [12], the SMSI effect was constructed on Au and TiOx by melamine induction, and Au nanoparticles were stabilized on the support even after calcination at 800 °C, thus improving the activity and stability of the catalyst. The second one is to confine the noble metal particles in porous supports, which acts as a physical barrier to inhibit the migration growth of noble metals, such as mesoporous silicon [13], zeolite [14], and hollow materials [15]. In Yue’s work [16], Pd@mesoporous silica core–shell nanocatalysts with good thermal stability were synthesized by sol-gel method, and the prepared catalysts showed good resistance to sintering. For the dry reforming of the methane reaction, high activity and favorable stability exhibited at 750 °C. In the course of Tian’s work [17], zeolites were used to encapsulate the noble metal Pt, in which Pt nanoparticles could be uniformly distributed in the zeolite support, and the prepared catalysts had high activity and thermal stability in the hydrodeoxygenation reaction. However, it is necessary to note that in this way, the active sites on the surface of the noble metal are easily covered by the support, thus reducing the catalytic activity to some extent. In neutral, the confinement of porous material is the most effective method to improve the thermal stability of noble metal-based catalysts at present.
As mentioned above, the immobilization of noble metal nanoparticles in nanoporous materials is an effective method to improve their thermal stability. However, the porous materials chosen as supports need to meet two basic conditions. Firstly, they must have good thermal stability and no sintering at the service temperature of the catalysts. Secondly, they must have rich porous structure capable of immobilizing noble metal nanoparticles. Based on these requirements, the nano-hollow sphere is undoubtedly the ideal choice, which has following advantages: 1) high specific surface area which can provide more active sites; 2) good permeability and porosity which are keys to ensuring adequate contact of reactants with the active sites; and 3) excellent thermal stability which ensures the catalyst to cope with harsh reaction environments [1]. For example, Yang et al. [18] prepared the Au-Fe2O3@ZrO2 yolk–shell structure, in which the Au-Fe3O4 heterostructure was integrated into ZrO2 hollow spheres. Due to the confined ZrO2 hollow spheres, the Au nanoparticles did not sinter yet even after the 900 °C high-temperature calcination. As we know, the choice of materials becomes particularly important to ensure that the hollow structure remains intact at high temperatures. ZrO2 has a high melting point and a high corrosion resistivity, which is an ideal candidate as the catalytic support.
Pd-based catalysts are always a hot topic of the catalytic research. With respect to the catalytic activity of the noble metal Pd, researchers hold different views [19–22]. In this work, we designed a three-step method to prepare Pd@HS-ZrO2 with a hierarchical pore structure, by employing carbon spheres (CSs) as sacrificial templates. The Pd nanoparticles were confined in the mesopores of the hollow sphere shell by layer-by-layer self-assembly of the ZrO2 precursor. The good thermal stability of the support allows Pd to maintain a highly dispersed state after the high-temperature calcination. Pd@HS-ZrO2 samples manifest superior catalytic activity and thermal stability than the counterpart Pd/ZrO2 catalysts, and it also demonstrates the advantages of employing ZrO2 hollow spheres as domain-limiting support and their potential value in the energy environment.
2 Experimental
2.1 Materials
Glucose (≥ 99.9% purity, Sinopharm Chemical Reagent Beijing Co., Ltd.), PdCl4 (Pd ≥ 58% purity, Sinopharm Chemical Reagent Beijing Co., Ltd.), NaBH4 (≥ 99% purity, Sinopharm Chemical Reagent Beijing Co., Ltd.), and Zr(SO4)2·4H2O (≥ 98% purity, Aladdin Chemical Reagent Company) were directly purchased commercially without further purification. Absolute alcohol was used as the solvent. Deionized water was self-made in our laboratory.
2.2 Synthesis of carbon spheres templates
First, 40 g glucose was added into 350 mL deionized water, stirring for 15 min, and then the solution was transferred to the stainless-steel sealed autoclave, and held at 170 °C for 7 h. Finally, the dark brown products were obtained by suction filtration and dried at 70 °C for 6 h.
2.3 Synthesis of CSs@Pd@ZrO2 precursor
Primarily, 34.667 mg PdCl4 was dissolved in 4 mL HCl (0.1 mol·L−1) and 96 mL deionized water to obtain H2PdCl6 (1 mL H2PdCl6 contains 0.2 mg Pd), stirring for 40 min at 70 °C. 150 mg prepared carbon spheres were evenly dispersed in 100 mL alcohol–water solution (the ratio between alcohol and water is 1:1). 21 mL H2PdCl6 was transferred into the solution with stirring for 15 min, and then sufficient NaBH4 (0.01 mol·L−1) was added. Subsequently, CSs@Pd was obtained after 2 h of stirring. Next, CSs@Pd was centrifuged and dissolved in 200 mL deionized water, then 240 mg Zr(SO4)2·4H2O was added into the solution. The CSs@Pd@ZrO2 precursor was obtained by stirring for 6 h. The precursor was rinsed three times alternately with deionized water and anhydrous ethanol. Finally, the black precursor was collected by suction filtration and dried at 80 °C for 7 h.
2.4 Synthesis of Pd@HS-ZrO2 hollow spheres and Pd/ZrO2 catalysts
The precursor was divided into four parts and calcined at 450, 700, 800, and 900 °C for 2 h to obtain Pd@HS-ZrO2-450, Pd@HS-ZrO2-700, Pd@HS-ZrO2-800, and Pd@HS-ZrO2-900, respectively.
For comparison, ZrO2 solid particles-supported Pd nanoparticle (Pd/ZrO2) catalysts were also synthesized. Firstly, 100 mg commercial ZrO2 nano-powder was uniformly dispersed in 100 mL deionized water. Then 25 mL H2PdCl6 (0.2 mol·L−1) was transferred into the solution by stirring for 30 min, and then a sufficient amount of NaBH4 was added. The Pd/ZrO2 catalyst precursor was obtained by stirring for 2 h. The sample was washed three times alternately with deionized water and anhydrous ethanol. Finally, divided Pd/ZrO2 catalyst into two parts and calcined at 450 and 800 °C for 2 h, eventually obtaining Pd/ZrO2-450 and Pd/ZrO2-800, respectively.
2.5 Characterizations
The phase composition of the samples was carried out on a Bruker D8-Advance diffractometer using nickel-filtered Cu Kα radiation. X-ray photoelectron spectroscopy (XPS; ESCALAB 250Xi) was utilized to characterize the surface valence states of the catalysts. The microstructure of the samples was observed by scanning electron microscopy (SEM; S-4800) and transmission electron microscopy (TEM; JEOL JEM-1010/2010). The Brunauer–Emmett–Teller (BET) surface area of the samples was determined employing A NOVA 4000 automated gas sorption system. Thermogravimetric analysis (TGA) of the samples was conducted in air with the heating rate of 10 °C·min−1 by SDT Q600 (TA Instruments). The loading of noble metals in samples was determined by inductively coupled plasma optical emission spectrometry (ICP-OES; XII ThermoFisher). The H2-temperature programmed reduction (H2-TPR) results were obtained by Autochemll2920.
2.6 Catalytic performance measurements
The CO oxidation was used to evaluate the catalytic performance. The activity and reliability of the samples were tested via a fixed-bed flow microreactor instrumented with a Shimadzu GC-2014 gas chromatograph. For all the samples, 30 mg sample was mixed with 90 mg quartz sand and then loaded into a quartz tube and fixed with quartz wool. The composition of the reaction mixture was N2:O2:CO = 79.2:20:0.8, and the reaction velocity was 60 000 mL·g−1·h−1.
3 Results and discussion
Fig.1 shows the schematic diagram of a three-step method synthesis process of Pd@HS-ZrO2 catalysts by employing CSs as the sacrificial template. Pd nanoparticles were loaded on the surface of CSs, the Zr ions uniformly deposited on the surface of CSs by hydrolysis of Zr(SO4)2·4H2O, and Pd@HS-ZrO2 obtained at the high-temperature calcination. Fig.2(a) shows CSs synthesized by the glucose hydrothermal process. As can be seen, CSs are uniform in size and well dispersed, with the diameter of about 200 nm. The first step is the formation of CSs@Pd particles. Through the reducing nature of NaBH4, Pd2+ is reduced to metal Pd0 particles, which are supported on the surface of CSs. The TEM images showed that the Pd particles with a diameter of about 3 nm were evenly distributed on the surface of the CSs (Fig.2(b) and S1).
The second step is the formation of the CSs@Pd@ZrO2 precursor. A previous study has illustrated that the surface of CSs is negatively charged, which is necessary to act as an ideal hard template [23]. Based on a large number of hydroxyl functional groups on the surface of the CSs, Zr ions could be uniformly deposited on the surface of CSs via the layer-by-layer self-assembly to form core–shell CSs@Pd@ZrO2 precursor particles (Fig.2(c)). In the previous study by Chen et al. [24], it obtained the vibration of Zr−OH results from the analysis of CSs@ZrO2 by Fourier transform infrared spectroscopy (FTIR), indicating that the Zr species are present on the surface of CSs in the form of Zr(OH)2.
The third step is to obtain Pd@HS-ZrO2. Due to the porous structure of the hollow sphere shell, the oxygen could adequately react with CSs by calcination. The TGA of the sample prior to calcination was conducted to deduce the appropriate calcination temperature for the removal of the spheres template. From Fig. S2, it can be seen that the weight loss can be divided into two stages. The first stage occurs in a temperature range of 10–280 °C, and the weight loss during this stage is approximately 15 wt.%, which can be attributed to the removal of absorbed water and residual solvent from the precursor. A significant weight loss of 60 wt.% occurs in the second stage from 280 to 700 °C. It can be contributed to the decomposition and oxidation of the CSs.
The morphology and microstructure of Pd@HS-ZrO2 samples were characterized in detail, as shown by the SEM images of the samples Pd@HS-ZrO2-450 (Fig.2(d)), Pd@HS-ZrO2-700 (Fig.2(e)), Pd@HS-ZrO2-800 (Fig.2(f)), and Pd@HS-ZrO2-900 (Fig. S3). It is evident that the Pd@HS-ZrO2 samples are almost identical to the morphology of the CS templates with a diameter of about 200 nm and the presence of some broken spheres proves the hollow structure. Based on the further observation by TEM (Fig.3), the shell of Pd@HS-ZrO2-450 reveals in detail that it is a porous structure with a shell thickness of approximately 10 nm (Fig. S4). Fig.3 shows the TEM-mapping of Pd@HS-ZrO2-450 hollow spheres after removing the template, and green, blue, and red areas refer to the elements Zr, O, and Pd, respectively. It could be seen that the Pd@HS-ZrO2 sample has integrated structure and good monodispersity. The distribution of the noble metals can be seen clearly, and the lattice spacing data of the noble metals were measured to be 0.243 and 0.261 nm (Fig.3), which correspond to the (1 1 1) plane of Pd0 and the (1 0 1) plane of PdO, respectively [25]. However, the lattice spacing of the noble metal measured in Pd@HS-ZrO2-800 was 0.261 nm, which corresponds to the (1 0 1) plane of PdO (Fig. S5), and also the results show that the lattice on the () plane of ZrO2 is a spacing of 0.32 nm (Fig. S5). The reason for not observing the (1 1 1) planes of Pd0 in Pd@HS-ZrO2-800 may be that Pd has been completely oxidized.
To further demonstrate the dispersion state of the noble metals, the distributions of elements in Pd@HS-ZrO2-800 were analyzed using high-angle annular dark-field scanning transmission electron microscopy (HAAD-STEM) combined with energy-dispersive spectroscopy (EDS) elemental mapping. The corresponding element mapping images are shown in Fig. S5, on which green, blue and red areas refer to the elements O, Zr, and Pd, respectively. The results show that the Pd element is uniformly distributed in ZrO2 hollow spheres, which is in good agreement with the observation of TEM images.
Fig.4 shows X-ray diffraction (XRD) patterns of different Pd@HS-ZrO2 samples calcined at 450, 700, 800 and 900 °C, separately. It can be seen that the sample calcined at 450 °C is amorphous, but a peak corresponding to the (1 0 1) plane of PdO can be found for all four samples at 33.9°. The peaks located at 30.2° is corresponding to the (1 0 1) plane of the tetragonal ZrO2 phase (PDF #79-1771). The remaining diffraction peaks correspond to other crystal planes of monoclinic ZrO2 (PDF #86-1450). It is obvious that the content of tetragonal ZrO2 decreases with the increase of the calcination temperature, which is consistent with the literature report [26]. The phase transition occurs accompanied by the growth of the grain size, which also causes the hollow sphere broken. This is consistent with the observation in Fig. S3. The variation of the valence state of Pd with different calcination temperatures was investigated by XPS, and the results are shown in Fig.5. In the XPS spectra, the 333 and 347 eV peaks are attributed to the Zr 3p3/2 region, while the 335.8 and 336.8 eV peaks to the Pd 3d5/2 region, and the 342.5 eV peak to the Pd 3d3/2 region [27–28]. According to the data, the characteristic peak centered around 335.8 eV corresponds to Pd0, while those around 336.8 and 342.5 eV correspond to PdO [29–31]. From the magnitude of the binding energy, the peak intensity of Pd becomes very weak at 800 and 900°C, indicating that Pd has been basically completely oxidized to the PdO form after calcination. It is in general agreement with the results observed by TEM that Pd0 was barely found in the 800 °C calcined sample, but both Pd0 and PdO species could be found in Pd@HS-ZrO2-450. The previous study also showed that the oxidation state of the noble metal Pd is critical to the catalytic activity of the catalyst [29].
The CO oxidation as a model reaction is a common method to evaluate the catalyst activity [32–35]. Some references and reviews about the CO oxidation are added and compared in Table S1. In this study, the catalytic activities of Pd@HS-ZrO2 and Pd/ZrO2 were evaluated using CO as a model reaction for comparison, and the results are shown in Fig.6(a). The catalytic performance of the five samples (Pd@HS-ZrO2-450, Pd@HS-ZrO2-700, Pd@HS-ZrO2-800, Pd@HS-ZrO2-900, Pd/ZrO2-450, and Pd/ZrO2-800) can be clearly compared in Fig.6(a). The ICP-OES results show the noble metal content of 4.3 wt.% in Pd@HS-ZrO2 and 3.6 wt.% in Pd/ZrO2. This is a little lower than the goal loading of 5 wt.% due to the fact that a small number of Pd particles were shed during the preparation process. In addition, the T50 (temperature at the 50% CO conversion) value was evaluated as a criterion to assess the activity of each catalyst. The following sequence shows the corresponding catalytic activity of the samples in order: Pd@HS-ZrO2-800 (106 °C) > Pd@HS-ZrO2-900 (115 °C) > Pd@HS-ZrO2-700 (120 °C) > Pd/ZrO2-800 (132 °C) > Pd@HS-ZrO2-450 (135 °C) > Pd/ZrO2-450 (145 °C). Obviously, Pd@HS-ZrO2 has better activity than that of the corresponding Pd/ZrO2 catalyst. The Pd@HS-ZrO2-450 hollow spheres can achieve the complete conversion of CO at 140 °C, but Pd/ZrO2-450 can only be achieved at 150 °C. The results are more intuitive in samples calcined at 800 °C. At 115 °C, Pd@HS-ZrO2-800 hollow spheres are able to react with CO completely, 25 °C lower than that for Pd/ZrO2-800. Adequately demonstrating the performance advantages of the hollow spheres. The reason is that, in Pd/ZrO2 without any physical limitation at the high-temperature calcination, small Pd particles would accumulate and grow larger (as shown in Figs. S6 and S7). There is no doubt that the Pd species are active centers of the catalyst. The agglomeration of Pd reduces active sites and leads to a reduction in the activity of the catalyst [36].
Fig.6 (a) CO conversion curves of samples (Pd@HS-ZrO2-450, Pd@HS-ZrO2-700, Pd@HS-ZrO2-800, Pd@HS-ZrO2-900, Pd/ZrO2-450, and Pd/ZrO2-800). (b) Cycle test of Pd@HS-ZrO2-800. (c) Catalytic stability test of the Pd@HS-ZrO2-800 catalyst. (d) Apparent activation energies of Pd@HS-ZrO2 samples. |
In order to demonstrate the catalytic stability of Pd@HS-ZrO2, cyclic tests were carried out on Pd@HS-ZrO2-800. It is clearly observed that the curves from Cycle 1 to Cycle 5 largely overlap (Fig.6(b)). Stability tests (Fig.6(c)) were conducted at 115 °C and the activity of the Pd@HS-ZrO2-800 catalyst did not decay during 30 h of service. Subsequently, the tested samples were characterized by XRD, XPS, and EDS-mapping (Figs. S8–S10). The XRD patterns show that the diffraction peaks before and after the reaction basically do not change, and the XPS results are also basically the same as before the test. Combined with the analysis on the TEM image, the Pd@HS-ZrO2 catalyst remained intact after the test, and the EDS-mapping result clearly shows that Pd particles maintain the monodisperse state. It is fully demonstrated that the Pd@HS-ZrO2 catalyst has good catalytic stability and thermal stability. Fig.6(d) reflects the apparent activation energies of different Pd@HS-ZrO2 samples. It is observed that the values of Pd@HS-ZrO2-800 and Pd@HS-ZrO2-900 are significantly lower than those of Pd@HS-ZrO2-450 and Pd@HS-ZrO2-700. The reason is that the oxidation degrees of Pd at 800 and 900 °C are significantly larger than those at 450 and 700 °C. The lattice oxygen of PdO will participate in the reaction, and it facilitates the reaction when the Pd particle in Pd@HS-ZrO2 is in the form of the oxidation state. The other reason is that the phase transition of ZrO2 may be accompanied by an increase in the surface oxygen vacancy concentration, which improves the mobility of oxygen [26]. The difference in activation energies was further explored by H2-TPR, and the results are shown in Fig. S11. The peak in the 200–300 °C interval corresponds to the reduction of PdO [37–38]. However, the reduction peaks of PdO can only be found in the samples of Pd@HS-ZrO2-800 and Pd@HS-ZrO2-900, but not in Pd@HS-ZrO2-450 and Pd@HS-ZrO2-700. Combined with the apparent activation energy of Pd@HS-ZrO2, it can be demonstrated that the deepening of the oxidation of Pd facilitates the catalytic reaction. And the peaks in the 430–530 °C interval correspond to the interaction of the ZrO2 support and PdO [39]. According to above results, compared with those of Pd@HS-ZrO2-450, Pd@HS-ZrO2-700, and Pd@HS-ZrO2-900, the reduction peak of Pd@HS-ZrO2-800 was much lower, indicating that the Pd@HS-ZrO2-800 catalyst has better reducibility at a lower temperature. It also demonstrates why Pd@HS-ZrO2-800 has the best catalytic activity.
Combined with previous TEM, XRD, and XPS analyses, we can explore the reasons for the enhanced catalytic stability of Pd@HS-ZrO2. The TEM image clearly contrasts that in Pd@HS-ZrO2, the noble metal is still able to be in a highly dispersed state after calcination at 800 °C, as in the sample calcined at 450 °C. For the BET testing (Fig. S12) of samples, hollow spheres exhibit type IV isotherms and H2 hysteresis loops, indicating a typical mesoporous structure. The combination of BET tests and TEM images provides that the shell of hollow spheres is not closed, and there are abundant pore channels in the shell where the reaction gases can easily flow. The advantages of the hollow structure are well demonstrated by the fact that the porous ZrO2 shell layer does not mask the active sites on the Pd surface, while at the same time inhibiting the growth of Pd nanoparticles. There is a clear PdO signal peak in the XRD pattern, and the XPS pattern demonstrates a significant increase in PdO after the calcination at 800 °C, compared to 450 °C. Therefore, it can be concluded that on the premise of ensuring the highly dispersed Pd noble particles, the increased oxidation degree of Pd can significantly improve the catalytic performance of the catalyst [29]. And under the premise that the hollow sphere structure is complete, the increase of monoclinic ZrO2 is also conducive to the improvement of catalytic activity. The postulated reaction mechanisms are as follows: PdO as active sites adsorbs CO, and the adjacent oxygen vacancies in ZrO2 can promote the adsorption and activation of O2, leading to the generation of reactive oxygen species. Immediately afterward, the CO adsorbed on PdO may migrate to the interface to react with reactive oxygen species to form CO2. And the lattice oxygen of PdO also reacts with CO [26,40].
4 Conclusions
In summary, we designed a three-step method to prepare Pd@HS-ZrO2 catalysts with the hierarchical pore structure. The CSs act as the sacrificial templates and the hierarchical porous shells of ZrO2 act as “micro-containers” to encapsulate tiny noble metal Pd particles. The Pd particles remain in a highly dispersed state even after the calcination at 800 °C. The prepared Pd@HS-ZrO2 catalysts exhibit good catalytic activity and high temperature stability in the oxidation of CO. The complete conversion of CO is achieved at 115 °C with Pd@HS-ZrO2-800, which is reduced by 25 °C compared to the counterpart Pd/ZrO2-800 catalysts. The enhanced catalytic activity and thermal stability of Pd@HS-ZrO2 can be attributed to the nanoconfinement effect offered by the 10 nm wall thickness of the ZrO2 hollow spheres, which suppresses the coarsening of Pd nanoparticles (active center for catalysis).