In recent years, catalytic CO oxidation has attracted considerable attention owing to its many applications such as pollution control for vehicle exhausts, trace CO removal in enclosed atmospheres, gas purification for CO2 lasers, CO gas sensors, and fuel cells [1, 2]. Catalysts employed in such a reaction are traditionally noble metals Pt, Pd, Au, and Rh supported on convenient materials such as alumina, zirconia, ceria, and titania [3, 4, 5, 6]. However, because of the high cost of noble metals and vulnerability to sulfur poisoning, increasing numbers of research studies are focusing on new catalysts containing cheap transition metals. Among them, Cu-based catalysts are found to be an excellent candidate for CO oxidation [7]. Especially, it has been reported that when CuO is supported over reducible oxides, such as CeO2, it exhibits specific activity, even superior to those of Pt-based catalysts [8].
It is well documented that the catalytic performance of CuO/CeO2 catalysts is critically dependent on their preparation methods [9]. As reported in the literature, a variety of methods exist to prepare CuO/CeO2 catalysts such as impregnation [9, 10, 11, 12], co-precipitation [11, 12], deposition-precipitation [10, 12], hydrothermal treatment [9], urea-nitrates combustion [12], and sol-gel methods [13]. Most of the above methods can be categorized into solution-based methods, whereby water-soluble molecular complexes, ranging from simple salts to organometallic complexes, are used as precursors for the catalytically active phase. Compared with the solution-based methods, the solid state impregnation (SSI) method has received relatively little attention despite its potential advantages. Solid state impregnation appears in the literature under various names such as the solvent-free method [14, 15], solid state grinding [16], and solid-liquid route [17]. It involves mechanical mixing of the precursor and support and subsequent heating to decompose the inorganic salt into the corresponding oxide. In this procedure, metal salts, especially hydrated transition metal nitrates with low melting points, will melt before decomposing, and can thus be dispersed on the surface of the support in the absence of any additional solvents. Hence, the preparation is more energy efficient and time-saving; it produces no product residues and requires no purification or recycling [18]. Li et al. [19] synthesized a series of copper cerium oxide catalysts by SSI using silica hollow spheres as support material, and investigated their catalytic performance in the CO-preferential oxidation (CO-PROX) reaction. They discovered that the catalysts prepared by SSI were more active than their counterparts prepared by conventional solution impregnation methods, and the difference was attributed to the higher content of finely dispersed CuO clusters that strongly interacted with ceria. Medina-Mendoza et al. [20] used this methodology to introduce Pt nanoparticles into the modified mesoporous Al-SBA-15. They pointed out that the Pt nanoparticles obtained by SSI showed higher activity towards naphthalene hydrogenation (HYD) than those obtained by the wet impregnation (WI) method. The higher activity was due to the higher density of edge sites, as measured by CO adsorption Fourier transform infrared (FTIR) spectroscopy. Tang and Sun et al. [21, 22, 23] studied SSI for the preparation of SBA-15 and CeO2 loading metal oxide catalysts. The catalysts showed higher activities than those prepared by traditional WI at a given metal oxide loading.
In this work, a comparative study on the physicochemical properties of catalysts prepared by SSI and WI methods was performed. A series of CuO/CeO2 catalysts were synthesized and characterized by X-ray diffraction (XRD), laser Raman spectroscopy (LRS), H2 temperature-programmed reduction (H2-TPR), X-ray photoelectron spectroscopy (XPS), in situ diffuse reflectance infrared Fourier transform spectroscopy (DRIFTS) and a CO oxidation model reaction, with the purpose to examine the influence of the preparation method on the structure, reducibility, and catalytic performance of the CuO/ CeO2 catalysts.
A certain amount of CeO2 (obtained by thermal decomposition of Ce(NO3)3·6H2O (Sinopharm Chemical Reagent Co., Shanghai, AR) at 550 °C for 5 h in air) and Cu(NO3)2·3H2O (Sinopharm Chemical Reagent Co., Shanghai, AR) were added to an agate mortar and manually ground for 0.5 h. The ground mixture was then transferred to a crucible and calcined in air at 450 °C for 4 h. The obtained samples are denoted as xCuO/CeO2-SSI, where x is the molar percentage of Cu in the mixture, based on [Cu/(Cu + Ce) × 100%]. The contents of Cu were 1 mol%, 3 mol%, and 5 mol%, respectively.
CuO/CeO2 samples were prepared by impregnating CeO2 with an aqueous solution containing a required amount of Cu(NO3)2·3H2O (Sinopharm Chemical Reagent Co., Shanghai, AR) for 2 h. The samples were heated at 100 °C in an oil bath to vaporize the water, subsequently dried at 110 °C in an oven overnight, and then calcined in air at 450 °C for 4 h. The obtained catalysts are denoted as xCuO/CeO2-WI, and the contents of copper and ceria were identical to those in the prepared xCuO/CeO2-SSI.
XRD patterns were recorded on a Philips X’pert Pro diffractometer (Netherlands) using an Ni-filtered Cu Kα radiation (λ 0.15418 nm). The X-ray tube was operated at 40 kV and 40 mA.
H2-TPR was carried out in a quartz U-tube reactor connected to a thermal conduction detector using a H2-Ar mixture (7.3% H2 by volume) as reductant. For each measurement, 50.0 mg of sample was used. Before switching to the H2-Ar stream, the sample was pretreated in a N2 stream at 300 °C for 1 h and then cooled to ambient temperature. After that the TPR started from room temperature at a rate of 10 °C min-1.
LRS analysis was performed on a Jobin-Yvon (France-Japan) T64000 laser Raman spectrometer using an Ar+ laser beam. The Raman spectra were recorded at an excitation wavelength of 514 nm and a laser power of 300 mW.
XPS analysis was performed on a PHI5000 VersaProbe high-performance electron spectrometer (Japan), using monochromatic Al Kα radiation (1486.6 eV), operating at an accelerating voltage of 15 kV. All binding energies (BE) were referenced to the adventitious C 1s at 284.6 eV. This reference gave BE values with an accuracy of ± 0.1 eV.
In situ DRIFTS spectra were collected from 650 to 4000 cm-1 at a spectral resolution of 4 cm−1 (32 scans) on a Nicolet 5700 FTIR spectrometer (USA) equipped with a high-sensitive mercury cadmium telluride (MCT) detector cooled by liquid N2. The DRIFTS cell (Harrick) was fitted with a ZnSe window and a heating cartridge that allowed samples to be heated to 400 °C. The fine catalyst powder placed on a sample holder was carefully flattened to enhance IR reflections. The sample was pretreated with a high purified N2 stream at 400 °C for 1 h to eliminate physisorbed water and other impurities. The sample background of each target temperature was collected during the cooling process. At ambient temperature, the sample was exposed to a controlled stream of CO-Ar (10% of CO by volume) at a rate of 5.0 mL min-1 for 1 h for saturation. Desorption studies were performed by heating the adsorbed species, and the spectra were recorded at various target temperatures at a rate of 10 °C min-1 from room temperature to 400 °C by subtraction of the corresponding background reference spectrum.
The CO oxidation activities of the catalysts were measured in a flow micro-reactor with a gas composition of 1.6 vol% CO, 20.8 vol% O2, and 77.6 vol% N2 at a space velocity of 30 000 mL g-1 h-1, and 50 mg (60-80 mesh) catalyst was used for each measurement. The catalyst was pretreated in an N2 stream at 200 °C for 1 h and then cooled to room temperature, after which the mixed gases were switched on. Two columns and a thermal conductivity detector were used for analyzing the gas production—Column A featured a 13 × molecular sieve for separating O2, N2, and CO, and Column B packed with Porapak Q for separating CO2.
Fig. 1 shows the XRD patterns of the CuO/CeO2 catalysts. As observed, all samples featured diffraction peaks attributed to cubic CeO2 with a fluorite structure [PDF-ICDD 34-0394], and no typical diffraction peaks of crystalline CuO [PDF-ICDD 48-1548] was visible, indicating either the existence of well- dispersed CuO species or CuO content that was below the detection limitation of the XRD instrument [34, 24]. Compared with pure CeO2, shifts in the position of the diffraction peaks were not obvious, revealing that the incorporation of Cu2+ into the CeO2 lattice was insignificant.
The grain size and lattice parameter of the catalysts are summarized in Table 1. All samples displayed comparable grain size (~9.5 nm). Also, the lattice parameters of the samples were comparable regardless of the preparation method employed (i.e. SSI and WI).
Fig. 2 shows the LRS spectra of the catalysts. The spectra of pure CeO2 showed a main band around 463 cm−1, which is related to the F2g symmetry vibration mode of the cubic fluorite-type structure [25, 26]. Following introduction of copper species, the F2g vibration peak gradually shifted from 463 to 459 cm−1, and became weaker and broader. This phenomenon can be interpreted by the phonon confinement model [13] and the presence of oxygen vacancies [27, 28]. Additionally, the two weak bands around 270 and 590 cm−1 were attributed to the second-order transverse acoustic (2TA) and defect-induced mode (D bond), respectively [22]. The relative peak intensity ratio of ID/IF2g is related to the concentration of surface oxygen vacancies in these samples. The higher ratio indicates higher concentrations of surface oxygen vacancies [29, 30]. As summarized in Table 1, the concentration of surface oxygen vacancies of the prepared CuO/CeO2 catalysts is higher than that of pure CeO2 in the following order: CuO/CeO2-SSI > CuO/CeO2-WI > CeO2. The higher the copper content, the higher the concentration of oxygen vacancies, indicates that the introduction of copper species is beneficial to the generation of oxygen vacancies. In particular, CuO/CeO2-SSI shows a higher oxygen vacancy concentration than CuO/CeO2-WI at a given CuO content, indicating that the preparation method can influence the structure of the catalyst.
Fig. 3 shows the H2-TPR profiles of the CuO/CeO2 catalysts prepared by SSI and WI methods. As observed, the reduction of the catalysts is significantly influenced by the preparation method and copper content. The reduction temperature for the CuO/CeO2 catalysts shifts to lower values with increasing copper contents. At a given copper loading, the CuO/ CeO2-SSI samples displayed lower reduction temperatures than the CuO/CeO2-WI samples.
Two reduction peaks were observed in all the samples. Because no diffraction peaks associated with CuO was detected from the XRD measurements, CuO (in CuO/CeO2) is believed to be either finely dispersed or present as an amorphous matrix on the catalyst surface. As reported in the literature, CuO species identified in CuO/CeO2 catalysts can be classified into four types [31, 32, 33]: (1) isolated Cu2+ ions can strongly interact with the support; (2) weakly magnetic associates consisting of several Cu2+ ions that are in close proximity with each other; (3)small two- and three-dimensional clusters featuring loose structures, with no specific and regular lattice arrangements; and (4) large three-dimensional clusters and bulk CuO phase, with properties identical to those of pure CuO powder. Accordingly, it is reasonable to ascribe peak α in the H2-TPR profiles to the overlap of the first two types ((1) and (2)) CuO species that are highly dispersed and strongly interact with the support. Peak β is assigned to type (3) CuO species that are small clusters with amorphous-like properties that are XRD-silent.
It has been reported that in CuO/CeO2 catalysts, strong interactions can easily occur at the interface between the highly dispersed copper oxide and ceria, and remarkably lower the reduction temperature of CuO [34, 35]. Therefore, TPR results indicate that the SSI method is advantageous for enhancing the interaction between copper and cerium, thereby benefiting the reduction of CuO species.
Quantitative analysis results of the H2-TPR profiles of the CuO/CeO2 catalysts are given in Table 2. As observed, for all catalysts, the actual H2 consumption during the reduction process is larger than the theoretical H2 consumption corresponding to the reduction of CuO species to Cu0. This finding indicates that concurrent reduction of the copper oxide species and surface CeO2 takes place owing to the synergistic interaction between CuO and CeO2 [36].
The elementary oxidation states and surface compositions of the samples prepared by different methods were assessed by XPS. The corresponding O 1s, Cu 2p, and Ce 3d XPS spectra are displayed in Fig. 4. As observed in Fig. 4(a), the O 1s spectra were fitted with two Gaussian peaks. The main peak O′ at 529.1 eV was attributed to lattice oxygen bonding to metal oxides, and the shoulder O′′ peak at the higher binding energy of 531.0 eV was assigned to chemisorbed oxygen, oxygen of surface adsorbed water, and C-O species [29, 37].
Fig. 4(b) shows the Cu 2p XPS spectra of the obtained samples. As observed, 3CuO/CeO2 and 5CuO/CeO2 samples show two peaks: a main Cu 2p3/2 peak at 933.9 eV and a corresponding satellite peak at 941.5 eV. The latter peak is characteristic of Cu2+ [36]. Additionally, a weak peak centered at 931.4 eV was observed, indicating the presence of reduced Cu+ species [29]. The relative content of Cu+ can be estimated by the peak area ratio of Cu+ and (Cu+ + Cu2+). As summarized in Table 3 (the peak intensity corresponding to Cu species in 1CuO/ CeO2 is too weak for proper quantification of the Cu+ content), the Cu content of the prepared samples decreased in the following order: 5CuO/CeO2-SSI > 3CuO/CeO2-SSI > 5CuO/ CeO2- WI > 3CuO/CeO2-WI, suggesting that higher amounts of surface Cu+ are generated by the SSI method. It is well recognized that the chemisorption of CO on Cu+ species is very important to the CO oxidation reaction [38]. Hence, we can expect that the catalysts prepared by SSI are relatively more suitable for CO-related reactions than those prepared by conventional WI.
The Ce 3d XPS spectra of the prepared samples were numerically fitted with eight components, and the corresponding assignments are displayed in Fig. 4(c). The two groups of spin-orbital multiplets corresponding to 3d3/2 and 3d5/2 are denoted as u and v, which are observed over a binding energy range of 875-925 eV. It is widely reported that the bands labeled u′ and v′ that are assigned to Ce3+ and Ce4+ are related to u′′′, u′′, u, v′′′, v′′, and v [39]. Furthermore, the percentage content of Ce3+ in the catalysts can be determined according to the following equation [40]:
The calculated percentages of Ce3+ are listed in Table 3. As observed, the amount of Ce3+ in all samples is higher than that of pure CeO2 reported in the literature (i.e. 6.75%) [41]. Moreover, CuO/CeO2-SSI samples feature higher Ce3+ contents than CuO/CeO2-WI samples. As reported, the existence of Ce3+ in CeO2 suggests the formation of oxygen vacancies [42]. Therefore, it can be inferred that the SSI method is prone to produce higher levels of oxygen vacancies, consistent with the LRS results.
As listed in Table 3 (except for the surface Cu/Ce ratio of 1CuO/CeO2 that is too low for accurate measurements by XPS), in general, the Cu/Ce surface atomic ratios were considerably higher than the theoretical feed ratios used during synthesis (data in parentheses), indicating the enrichment of copper species on the surface of the catalysts. As discussed earlier, XRD results demonstrated that CuO was finely dispersed or present as an amorphous matrix on the surface of CeO2. Hence, the higher the Cu/Ce ratio, the higher the level of dispersed CuO. Moreover, the Cu/Ce ratio of the catalysts prepared by SSI was higher than that of the catalysts prepared by WI at a given copper loading, indicating that the SSI method is advantageous for generating surface-dispersed CuO species.
To monitor intermediates and adsorption species in a CO atmosphere, samples were analyzed by in-situ DRIFTS, and the spectra are displayed in Fig. 5. IR bands were mainly concentrated in two regions, i.e., 900-1800 and 2000-2500 cm−1. The bands at 1306 and 1581 cm-1 are ascribed to bidentate carbonate [43] and the bands at 1475 and 1359 cm−1 are attributed to monodentate carbonate [44]; the latter carbonate species are stable, and the band intensities do not change with temperature. The band at 1217 cm−1 is assigned to surface hydrogen-carbonate species. Finally, the bands at 1521 and 1386 cm−1 correspond to different modes of formate [45].
The frequency, strength, and thermal stability of the Cu+-CO bond are often used to identify the valence states and nature of the adsorption center of copper oxide [46]. As observed for 5CuO/CeO2-SSI (Fig. 5(b)), the Cu+-CO band was first detected at room temperature, and the corresponding band intensity increased gradually and reached maximum at 160 °C. For 5CuO/CeO2-WI (Fig. 5(d)), the corresponding Cu+-CO species band intensity began to increase from 75 °C before reaching a plateau between 150 and 200 °C. The area of the Cu+-CO band for both samples was calculated and the results are displayed in Fig. 6. As observed, the area of the 5CuO/CeO2-SSI band was larger than that of the 5CuO/CeO2-WI band below 175 °C, suggesting that 5CuO/CeO2-SSI featured a higher concentration of Cu+, as generated by the redox equilibrium:
Relative to 5CuO/CeO2-WI, Reaction (1) tends to easily shift to the right to form stable Cu+ species over 5CuO/CeO2-SSI, in accordance with the XPS results.
For 5CuO/CeO2-SSI, the band at 2360 cm−1, corresponding to the evolution of gaseous CO2, appeared at 30 °C, whereas in 5CuO/CeO2-WI, this band was detected at a higher temperature of 125 °C. According to the literature [27], CO2 is generated upon reaction between the incoming CO and surface active oxides species. Thus, this finding suggests that 5CuO/CeO2-SSI is more prone to reduction, as consistent with the TPR results. By further integrating the area of the CO2 band (Fig. 6), it can be seen that the total CO2 content generated over 5CuO/ CeO2-SSI is larger than that produced over 5CuO/CeO2-WI, regardless of the reaction temperature, thereby demonstrating that higher amounts of CuO can be reduced by CO over catalysts prepared by SSI as follows:
Concurrently, the concentrations of Cu+ and surface oxygen vacancies increase as well.
The catalytic performances of the CuO/CeO2 catalysts were evaluated using CO oxidation as model reaction, and the corresponding results are displayed in Fig. 7. As observed, copper-ceria catalysts are good candidates for low-temperature CO oxidation, with complete oxidation achieved below 150 °C. Moreover, the activity of the catalysts increases with increasing CuO content, indicating that the introduction of copper species is beneficial to CO oxidation. Noticeably, catalysts prepared by SSI are more active than their counterparts prepared by conventional WI, thereby suggesting that the SSI synthesis method is more favorable for obtaining highly efficient reactions. The catalytic performance of the samples decreased in the following order: CuO/CeO2-SSI > CuO/CeO2-WI > CeO2 across the range of reaction temperatures studied. The same trends in regards to the dispersion of Cu species, Cu+ content, and oxygen vacancy concentration were observed.
It is well recognized that the activity of CuO/CeO2 catalysts towards CO oxidation is mainly influenced by two factors. One factor involves the Cu+ sites for CO chemisorption that can activate CO molecules. The other factor involves the oxygen vacancies as provided by the CeO2 support that activate the oxygen species. A reaction model of CO oxidation over the CuO/CeO2 catalysts is proposed based on the current results, as shown in Scheme 1. The catalytic cycle involves: (1) reaction of adsorbed CO and CuO, and the generation of Cu+, CO2, and oxygen vacancies; (2) chemisorption of CO onto Cu+ ions to form Cu+-CO species; (3) O2 activation on the oxygen vacancies and formation of surface-activated oxygen (e.g. O2−) or lattice oxygen (e.g. O2−) species; (4) reaction between the chemisorbed CO and nearby active oxygen; and (5) the newly generated Cu+ sites and oxygen vacancies involved in successive catalytic cycles. Based on the above discussion, higher concentrations of these intermediates on the catalyst surface can enhance the activity of the catalyst, as evidenced by the trend obtained relating to the catalytic CO oxidation performance: CuO/CeO2-SSI > CuO/CeO2-WI > CeO2.
In this work, a series of CuO/CeO2 catalysts were prepared by solid state impregnation and wet impregnation methods, and the physical and chemical properties were characterized by XRD, LRS, H2-TPR, in situ DRIFTS, XPS, and a CO + O2 model reaction. The results showed that: (1) compared with catalysts prepared by the wet impregnation method, catalysts synthesized by the solid state impregnation method have the advantages of higher levels of dispersed copper species on the surface of the catalysts and stronger interactions with the support that are beneficial to the reduction of CuO species; (2) the enhanced catalytic performance of the CuO/CeO2-SSI samples is strongly related to the higher concentrations of oxygen vacancies on the surface of the support and chemisorbed CO on Cu+ sites.
近年来, 基于在机动车尾气净化、密闭环境中微量CO的消除、CO2激光器气体纯化、CO气体传感器及燃料电池等方面的广泛应用, CO氧化反应受到了越来越多的关注[1, 2]. 通常, 应用于该反应的催化剂主要是负载于常见载体(如氧化铝、氧化锆、氧化铈、氧化钛)的Pt, Pd, Au和Rh等贵金属催化剂[3, 4, 5, 6]. 但由于贵金属催化剂价格昂贵且抗硫性能差, 研究开发能够替代贵金属的非贵金属催化剂逐渐成为催化领域中的一个研究热点. 其中, 铜基催化剂凭借其良好的催化性能受到了广泛关注[7]. 研究表明, 当CuO负载到可还原载体如CeO2上时, 其催化性能甚至能与Pt基催化剂相媲美[8].
众所周知, 制备方法的不同会显著影响CuO/CeO2催化剂的反应性能[9]. 目前, 已有诸多方法用于CuO/ CeO2催化剂的制备, 如浸渍法[9, 10, 11, 12], 共沉淀法[11, 12], 沉积-沉淀法[10, 12], 水热处理法[9], 尿素-硝酸盐燃烧法[12]和溶胶-凝胶法[13]等. 这些方法都在水溶液中进行, 所用的可溶性金属前驱体从简单的无机盐到金属有机配合物, 分布极广. 比较而言, 固相浸渍法有着一定优势, 但却受到较少关注. 在文献报道中, 固相浸渍法还往往以其他名称出现, 如: 无溶剂法[14, 15], 固相研磨法[16], 固-液路线法[17]等. 该方法制备催化剂主要包括两个步骤: 前驱体与载体的充分机械混合; 加热使得前驱盐分解为相应氧化物. 在加热过程中, 金属前驱盐, 尤其是具有较低熔点的水合过渡金属硝酸盐在分解前会首先经历熔融状态, 借助载体表面的毛细凝聚等作用力, 这些熔融态物质在无溶剂的情况下也能得到较好分散. 由于整个过程中都不会涉及制备液体的纯化和再循环等环节, 从而保证了制备过程的节能、节时[18]. Li等[19]通过固相浸渍法合成了一系列负载在SiO2空心球上的铜铈催化剂, 并考察了其在CO富氢氧化反应中的催化性能. 他们发现, 固相浸渍法制备的催化剂可以产生更多与载体紧密接触的分散态CuO, 从而表现出比传统的湿浸渍法更好的催化性能. Medina-Mendoza等[20]发现, 固相浸渍法制备的Pt/Al-SBA-15催化剂相对于湿浸渍法在萘的加氢反应中也有更好的催化活性. Tang和Sun等[21, 22, 23]分别以SBA-15和CeO2为载体负载过渡金属氧化物, 考察了固相浸渍法制备催化剂在CO氧化反应中的应用. 并指出在相同负载量条件下, 固相浸渍法相比于湿浸渍法有较大优势.
本文采用固相浸渍法和常规湿浸渍法制备了一系列CuO/CeO2催化剂, 同时结合X射线衍射(XRD)、氢气-程序升温还原(H2-TPR)、激光拉曼光谱(LRS)、原位漫反射红外光谱(in situ DRIFTS)、X射线光电子能谱(XPS)等表征手段和CO氧化模型反应, 对比研究了制备方法对催化剂结构、还原性质及其催化CO氧化性能的影响.
将计算量的Cu(NO3)2·3H2O (国药集团化学试剂有限公司, 分析纯)和CeO2 (由Ce(NO3)3·6H2O (国药集团化学试剂有限公司, 分析纯)空气中550 °C焙烧5 h制得)载体混合研磨0.5 h后置于坩埚中, 转入马弗炉中在流动空气下经450 °C焙烧4 h. 所得样品记为xCuO/CeO2-SSI, 其中x代表Cu/(Cu + Ce)摩尔百分比. 如5NiO/CeO2-SSI代表催化剂中含铜量为5%.
将Cu(NO3)2·3H2O水溶液浸渍到CeO2载体表面上, 搅拌2 h后在100 °C下蒸干. 所得样品经110 °C干燥过夜后放入马弗炉中在流动空气下450 °C下焙烧4 h. 与上类似, 所得样品记为xCuO/CeO2-WI, 其中铜含量与SSI相对应.
XRD在荷兰Philips X' pert Pro衍射仪上进行, 采用Cu-Kα辐射源, 入射波长λ = 0.15418 nm, Ni滤波片, 管电压40 kV, 管电流40 mA.
H2-TPR实验在自组装的TPR仪器上进行, 采用热导池检测器. 将50 mg样品装入U型石英管中, 首先用高纯N2在300 °C下吹扫1 h, 冷却至室温后, 切换成7.3% H2/Ar (体积比)混合气. 升温速率为10 °C min-1.
LRS测试在Jobin-Yvon T64000型光谱仪上进行, 采用Ar+激光束, 激发波长为514 nm, 激光功率为300 mW.
XPS测试在PHI 5000 VersaProbe型能谱(日本UlVAC-PHI公司)上测定, 采用Al Kα单色射线作激发光源(hν = 1486.6 eV), 加速电压为15 kV. 以C 1s的结合能(284.6 eV)作为校正标准, 结合能的精度为 ± 0.1 eV.
In situ DRIFTS光谱在Nicolet 5700型傅里叶变换红外光谱仪(美国, 赛默飞世科技)上进行测定, 检测器为高灵敏度的MCT (液氮温度下工作), 扫描范围为650-4000 cm-1, 扫描次数为32次, 分辨率为4 cm-1. 所用原位池配备有ZnSe窗片及加热部件可供样品加热至400 °C. 首先, 将样品粉末装入原位池, 使样品表面尽量平整以反射信 号; 其次, 在400 °C下用高纯N2吹扫样品1 h以去除样品表面吸附水及其它杂质; 然后, 在降温过程中采集每个目标温度下的样品背景, 冷却至室温后以5.0 ml min-1的速率通入CO-Ar混合气(CO的体积百分数为10.0%) 1 h使催化剂吸附饱和; 最后, 从室温开始以10 °C min-1的速率升温至400 °C, 并在设定的温度下采集扣除相应背景的红外谱图.
催化反应测试在连续流动石英微反应器中进行. 将50 mg催化剂压片后过60-80目筛. 反应中的气体空速为30000 mL g-1 h-1, 反应气的体积组成: CO 1.6, O2 20.8, N2 77.6%. 在反应前, 催化剂需用高纯N2在200 °C下吹扫1 h, 冷却至室温并切换至反应气, 活性数据在反应达到平衡后采集. 产物经两根色谱柱和一个热导池检测器(TCD)分离、检测. 柱A填充13X分子筛用于分离O2, N2和CO. 柱B填充Paropak Q用于分离CO2.
图1为CuO/CeO2催化剂的XRD谱. 可以看出, 所有样品均出现CeO2立方萤石结构[PDFICDD 34-0394]晶相衍射峰, 但并未发现晶相CuO [PDF-ICDD 48-1548]的衍射峰. 这表明CuO主要以高分散或低于XRD检测限的小颗粒形式存在于载体表面[29]. 同CeO2载体相比, 催化剂各衍射峰位置并未发现明显偏移, 说明Cu2+未掺入CeO2体相晶格.
样品晶粒尺寸和晶胞参数结果见表1. 可以发现, 所有样品的晶粒尺寸非常接近, 大致都在9.5 nm左右, 且固相浸渍法和湿浸渍法制备催化剂的晶胞参数也未表现出明显差异.
图2给出样品的拉曼结果. 由图中可以看出, CeO2载体在463 cm-1呈现出一个归属为CeO2立方萤石结构F2g振动模式的散射峰[25, 26]. 在负载铜物种以后, 该振动峰变弱、变宽, 并由463 cm-1逐渐向低波数方向偏移至459 cm-1. 这可以从声子限域效应[13]、氧空位的产生[27, 28]等方面得到解释. 另外, 在270和590 cm-1处出现的两个弱峰分别归属于二级横向声子振动和缺陷诱导振动[22]. 拉曼光谱中ID/IF2g值能够反映催化剂样品中氧空位浓度的大小, 比值越大代表氧空位越多[29, 30]. 如表1所示, CuO/CeO2催化剂中氧空位浓度均高于纯CeO2并呈现如下顺序: CuO/CeO2-SSI > CuO/CeO2-WI > CeO2. 可以看出, 催化剂负载的CuO含量越高, 其氧空位浓度越高, 表明CuO的引入有利于形成更多的表面氧空位. 另外, 在相同CuO负载量条件下, CuO/CeO2-SSI氧空位浓度要高于CuO/CeO2-WI, 表明制备方法的不同会导致催化剂结构的差异.
由不同方法制备的CuO/CeO2催化剂的H2-TPR结果如图3所示. 可以发现, 催化剂的还原明显地受到活性组分含量和制备方法的影响. 随着铜负载量的增加, 还原峰向低温方向偏移. 与此同时, 当铜负载量相同时, CuO/CeO2-SSI样品的还原峰温低于CuO/CeO2-WI样品.
表2给出了催化剂H2-TPR定量分析结果, 可以发现, 这些催化剂在还原过程中的实际耗氢量明显大于其按Cu2+物种完全还原到Cu0所需的理论耗氢量, 表明CuO物种和表面CeO2发生了共同还原, 这主要是由于CuO与CeO2之间的协同作用所致[36].
我们进一步采用XPS表征来研究催化剂的表面组成和元素价态. O 1s, Cu 2p和Ce 3d的XPS谱图见图4. O1s谱图如图4(a)所示, 所有样品都在529.1 eV左右呈现出一个归属为键合到金属离子上的晶格氧主峰(O′), 并在较高结合能531.0 eV附近出现一个与表面吸附氧、吸附水和C-O物种相关的肩峰(O′′) [29, 37].
Cu 2p谱图见图4(b). 3CuO/CeO2和5CuO/CeO2样品催化剂的Cu 2p3/2的主峰都出现在933.9 eV左右, 同时在941.5 eV左右出现一个归属为Cu2+特征信号的卫星峰[36]. 此外, 在931.4 eV处也出现一个弱峰, 表明Cu+物种的存在[29]. 其中, Cu+的相对含量可以通过Cu+和(Cu+ + Cu2+)的峰面积来计算(1CuO/CeO2催化剂样品由于Cu物种信号太弱而不能对Cu+准确定量). 如表3所示, 其大小顺序如下: 5CuO/CeO2-SSI > 3CuO/CeO2-SSI > 5CuO/ CeO2-WI > 3CuO/CeO2-WI, 表明固相浸渍法有利于产生更多的Cu+物种. 众所周知, 在CO氧化反应中, Cu+物种上吸附态的CO分子对反应的进行至关重要[38]. 我们可以推断, 在有CO分子参与的反应中, 固相浸渍法制备催化剂要比传统湿浸渍法更有优势.
所有Ce 3d的XPS谱图被分为8个峰, 相关结果见图4(c). 在875-925 eV范围内出现两组自旋轨道多重峰, 分别对应于3d3/2和3d5/2, 标记为u和v. 其中, u′和v′峰归属为Ce3+, 而u′′′, u′′, u, v′′′, v′′, v和Ce4+相关[39]. 此外, 表面Ce3+含量可通过u′和v′的峰面积按照以下公式来计算[40]:
表3给出了这些样品表面Ce3+的含量. 可以发现, 所有样品中Ce3+的含量均高于文献中所报导的纯CeO2含量(Ce3+ = 6.75%)[41], 且SSI的更高. CeO2中Ce3+的存在意味着有氧空位产生[42]. 所以, 固相浸渍法制备样品倾向于生成更多氧空位, 这与LRS结果一致. 我们发现, 除1CuO/CeO2样品中铜物种含量太少不能被XPS准确定量外, 其它所有样品的表面Cu/Ce原子比都大于其理论值, 表明Cu物种在这些CuO/CeO2催化剂表面富集. 同时, 我们发现固相浸渍法制备样品中Cu/Ce比高于湿浸渍法制备样品, 结合之前XRD讨论结果, 所有样品中CuO都是以分散态或小颗粒的形式存在于载体表面,故Cu/Ce比越高, 表面高分散的CuO含量越多. 这说明固相浸渍法有利于铜物种在载体表面的分散.
为了研究CuO/CeO2催化剂在CO气氛中的反应过程及吸附物种, 我们对5CuO/CeO2样品进行了in situ DRIFTS谱表征. 从图5可以看出, 两个样品的振动峰都可以分为两个区域: 碳酸盐、甲酸盐的振动(900-1800 cm-1)和羰基的振动(2109 cm-1左右). 其中, 1306和1581 cm-1处的振动峰对应于双齿碳酸盐、1475和1359 cm-1归属于单齿碳酸盐、1045 cm-1为多齿碳酸盐振动. 这些碳酸盐稳定存在, 不受温度的影响. 1217 cm-1归属于表面碳酸氢盐的振动. 1521和1386 cm-1对应于甲酸盐的不同振动模式[45].
Cu+-CO红外振动峰的频率、强度和热稳定性等通常被用来确定铜氧化物吸附中心的价态和性质[46]. 对于5CuO/CeO2-SSI样品 (图5(b)), 我们可以发现, 从室温开始即可检测到Cu+-CO振动峰, 且峰强度随着温度的增加而增强, 到160 °C时达到最大值. 而对于5CuO/CeO2-WI (图5(d)), 虽然室温也检测到Cu+-CO的存在, 但其强度从75 °C才开始逐渐增加并在150-200 °C间保持稳定. 催化剂Cu+-CO红外振动峰面积与温度的关系见图6. 低于在175 °C时, 5CuO/CeO2-SSI催化剂Cu+-CO峰面积均高于5CuO/CeO2-WI, 而Cu+的产生主要源自于:
说明5CuO/CeO2-SSI比5CuO/CeO2-WI更有利于该反应向右进行, 从而形成更多稳定存在的Cu+物种, 也与XPS结果相符.
位于2360 cm-1处的谱峰归属为气相CO2的振动, 在5CuO/CeO2-SSI催化剂中30 °C时就已检测到该峰的存在, 而在5CuO/CeO2-WI中到125 °C才被探测到. 这些CO2主要是CO气体还原表面氧化物所生成[27], 这说明5CuO/CeO2-SSI更易被还原, 也与TPR结果一致. 进一步研究CO2峰面积与温度关系, 可以发现在整个反应温度区间中, 5CuO/CeO2-SSI的峰面积均大于5CuO/CeO2-WI. 这意味着固相浸渍法制备样品中有更多的CuO被CO还原, 遵循下列反应:
与此同时, 也导致催化剂中Cu+及氧空位浓度的提高.
我们以CO氧化为模型反应对这些样品的催化性能进行了评估, 相关结果见图7. 可以发现, 所有样品在150 °C以下就已使CO气体完全转化, 说明CuO/CeO2是良好的CO低温消除催化剂. 另外, 催化剂的活性随着铜负载量的增加而提升, 表面铜的引入有利于反应的进行. 并且在相同CuO负载量条件下, 固相浸渍法制备样品比湿浸渍法表现出更好的催化活性. 还可以发现, 这些样品的催化性能优劣顺序如下: CuO/CeO2-SSI > CuO/CeO2- WI > CeO2, 与催化剂表面铜物种分散量、Cu+及氧空位浓度一致.
众所周知, 在CO氧化反应中CuO/CeO2催化剂的反应性能主要受两个因素的影响: 一个是Cu+上CO的化学吸附, 用于活化CO分子; 另一个是CeO2载体上的氧空位, 用于活化氧物种. 基于上述讨论结果, 我们尝试性地提出了CuO/CeO2催化剂表面CO氧化反应的可能机理, 如图示1所示, 反应循环包括: (1) 表面吸附的CO与CuO反应, 并且生成Cu+, CO2和氧空位; (2) Cu+上吸附CO并形成Cu+-CO物种; (3) O2通过氧空位得以活化, 并生成表面活性氧物种(如: O2-)和晶格O2-物种; (4) 吸附态CO与邻近活化氧反应; (5) 新生成的Cu+和氧空位继续循环反应. 结合以上讨论, 我们得出反应中间体Cu+和氧空位浓度越高, 越有利于CO氧化反应进行的结论, 并可得到催化性能结果的支持: CuO/CeO2-SSI > CuO/CeO2-WI > CeO2.
采用固相浸渍法和湿浸渍法制备了一系列CuO/CeO2催化剂, 并重点研究了制备方法对催化剂物理化学性质及其在CO氧化反应中催化性能的影响. 结果表明: (1) 相比于湿浸渍法, 固相浸渍法制备的催化剂更有利于铜物种在载体表面的分散及加强其与载体的相互作用, 从而促进CuO物种的还原; (2) CuO/CeO2-SSI样品的催化性能明显提高, 这与其表面氧空位及Cu+-CO物种的浓度的提高紧密相关.
2014, Vol. 35 Issue (8): 1347-1358
PDF (786KB)
2014, Vol. 35 Issue (8): 1347-1358
PDF (786KB)