Controlled Synthesis of Heterostructured SnO 2 -CuO Composite Hollow Microspheres as Efﬁcient Cu-Based Catalysts for the Rochow Reaction

: In this work, we report the design and synthesis of a series of heterostructured SnO 2 -CuO hollow microspherical catalysts (H-SnO 2 ( x )-CuO, x is the weight ratio of Sn/Cu) for the Rochow reaction. The microspherical catalysts with nanosheets and nanoparticles as building blocks were prepared by a facile one-pot hydrothermal method coupled with calcination. When tested for the Rochow reaction, the prepared H-SnO 2 (0.2)-CuO composite exhibited higher dimethyldichlorosilane selectivity (88.2%) and Si conversion (36.7%) than the solid CuO, hollow CuO and other H-SnO 2 ( x )-CuO microspherical samples, because in the former there is a stronger synergistic interaction between CuO and SnO 2 .


Introduction
In material synthesis, how to obtain materials with well-defined structures/shapes is one of the major goals, as each of their application is dependent on their physicochemical attributes and morphologies. Recently, nanostructures with various shapes including nanowires [1], nanotubes [2], nanorods [3] and hollow nanospheres [4] have been reported. Among them, hollow nanostructures are of particular interest owing to their unique structural features such as high surface area and fast diffusion capability, which are of importance to many applications such as catalysis [5,6], sensors [7], pollution treatment [8] and energy conversion [9]. For example, it was reported that CuO and NiO hollow nanospheres showed high catalytic activity and excellent stability for the hydrogenation of 4-nitrophenol [10], while a three-dimensional (3D) graphene network encapsulating SnO 2 hollow spheres exhibited outstanding electrochemical properties as an anode material for lithium-ion batteries [11]. However, most of these reported hollow structures are made up of only single oxide component, limiting their wider application. In principle, hollow heterostructures containing two or more kinds of metal oxide components with specific morphologies will have more abundant structural features such as various synergistic interactions among different components, thus potentially having more substantial applications [12,13]. However, the controlled synthesis of these multi-component complex systems is still highly more challenging.
One of the applications for multi-Cu component containing system is for the Rochow reaction. This reaction is used to synthesize methylchlorosilanes, which are important intermediates in the organosilane industry [14]. As shown in Scheme 1, this reaction often yields various products, in which (CH 3 ) 2 SiCl 2 (M2) is the most needed, as it is widely used for the manufacture of organosilicon products such as silicon rubber and silicon oil. Therefore, to enhance the selectivity and yield of M2, high-performance Cu-based catalysts are always desired. In the past decades, various catalyst systems have been reported [15][16][17]. In our previous work, many CuO catalysts with specific morphologies and structures have also been prepared and tested for the Rochow reaction. For instance, we observed that the flower-like CuO microspheres displayed a higher M2 selectivity and Si conversion than the commercial CuO catalyst with irregular morphology [18]; the leaf-like CuO nanostructure showed enhanced catalytic properties due to its largely exposed {001} plane that could facilitate formation of more Cu 3 Si active phase during the reaction [19]; most recently, we demonstrated that the rambutan-like CeO 2 -CuO hollow microspheres exhibited superior catalytic performances to their solid counterpart and pure CuO, because of the more intimate contact of CuO with Si powder and the synergistic effect between CeO 2 and CuO in the former [20]. These results indicate that the M2 selectivity and yield could be further improved by rationally controlling the morphology, structure and composition of Cu-based catalysts. Furthermore, it has been recognized that SnO 2 is an outstanding promoter for the Rochow reaction [21,22]. In this sense, SnO 2 -CuO micro-nano composites with hollow heterostructures are most likely to be good catalyst candidates too. However, to our knowledge, so far, such a promising catalyst system has not been explored for the Rochow reaction. (1) Herein, we report the preparation of heterostructured SnO 2 -CuO composite hollow microspheres with nanosheets and nanoparticles as building blocks and with various Sn/Cu ratios using a facile bottom-up strategy. When employed as catalysts for the Rochow reaction, the SnO 2 -CuO hollow heterostructures showed a higher selectivity of M2 and Si conversion than the solid CuO, hollow CuO and commercial catalyst, particularly at the optimal Sn/Cu ratio. Moreover, the structure-dependent catalytic attributes and the reaction mechanism are also studied.

Characterization of H-SnO 2 -CuO Precursor Particle
The XRD pattern of the collected precursor before calcination shows that the diffraction peaks match well with those of monoclinic basic cupric carbonate (Cu 2 (OH) 2 CO 3 ) (JCPDS No. 01-072-0075) (Figure 1a). The formation of Cu 2 (OH) 2 CO 3 should be the result of the chemical reaction between Cu(CH 3 COO) 2 ·H 2 O and Na 2 CO 3 solution, which is presented as follows: The low-magnification FESEM image shows that the Cu 2 (OH) 2 CO 3 precursor is in the shape of microspheres and all particles are uniform with an average size of ca. 5 µm (Figure 1b). From a typical crashed microsphere (Inset of Figure 1b), it could be clearly observed that the precursor particle possesses a hollow structure. The high-magnification FESEM image further demonstrates that some disorder nanosheets are stretched out onto the microspherical surface ( Figure 1c). The TG curve of this uncalcined sample shows a weight loss of about 27.6% from 200 to 350 • C in air, matching well with 27.9%, the theoretical value for the decomposition process from Cu 2 (OH) 2 CO 3 to CuO (Figure 1d).  ). However, no obvious peak matching with crystalline Sn species could be found, probably due to its small amount or small size. Additionally, no traces of other impurities such as Cu 2 O or Cu 2 (OH) 2 CO 3 can be observed. As we know, particle size often impacts catalytic activity [23]. Hence, the CuO nanocrystal sizes in these materials are calculated from the diffraction peaks at 2θ values of 35.5 • using the Scherrer formula, the results are listed in Table 1. It is clearly seen that the average grain size of CuO in H-SnO 2 -CuO sample is 22.9 nm, which is smaller than that in H-CuO (42.1 nm), indicating that the CuO particle size could be reduced through introducing Sn. Moreover, the BET surface area and pore volume of various samples is given in Table 1. It can be seen that all H-SnO 2 (x)-CuO composites have similar surface areas and the pore volumes but which are much larger than those of S-CuO.

Characterization of H-SnO 2 -CuO
The measured low-magnification FESEM image shows that H-SnO 2 -CuO possesses similar morphology to its precursor particle, which is also in the shape of hollow spheres with an average size of ca. 5 µm (Figure 3a,b). As revealed by the magnified images shown in Figure 3c,d, the CuO shell is constructed by interconnected disordered nanoparticles and nanosheets. The size of the primary nanoparticles is typically about 20 nm. A nearly 91% of morphological yield of this type can be achieved in our work. The SEM-EDS mapping of H-SnO 2 -CuO further reveals that the elements of Cu, O and Sn are homogeneously distributed throughout the entire particle ( Figure 3e) and the weight ratio of Sn/Cu is measured to be about 0.19, which is close to the result of ICP-OES and the feeding ratio of the two precursors.   The XPS spectra of Cu 2p, Sn 3d and O 1s were shown in Figure 4 and the presence of Cu, Sn, O and C elements are confirmed by XPS analysis of H-SnO 2 -CuO ( Figure 4a). From the XPS spectra of Cu 2p (Figure 4b), it can be seen that the major peaks at 933.5 and 953.6 eV can be assigned to Cu 2p 3/2 and Cu 2p 1/2 of CuO, respectively [24], which are accompanied with a satellite peak at ca. 943.8 eV. Comparing with that of H-CuO sample, Cu 2p 3/2 peak has a slight shift and it has decreased by 0.2 eV for H-SnO 2 -CuO. Figure 4c shows the Sn 3d XPS spectrum of the H-SnO 2 -CuO sample. The binding energies of Sn 3d 5/2 and Sn 3d 3/2 are located at 495.0 and 486.6 eV, respectively, indicating that the Sn species exists in the form of Sn 4+ [25]. The surface Cu/Sn ratios determined by XPS and the bulk Cu/Sn ratios given by ICP in all the H-SnO 2 (x)-CuO samples are both summarized in Table 1.    Table 1, which shows that adjustment of the ratio of metal precursors can control the final product composition.

The Formation Mechanism of H-SnO 2 -CuO
To understand the formation mechanism of H-SnO 2 -CuO, a series of experiments with various reaction hours were performed. Figure 6a-d illustrate the morphological evolution of the particles along with the reaction processing. In the early stage of the reaction, a fast reaction among the large number of Cu 2+ ions, the carbonate ions (CO 3 2− ) and the formed OH − by hydrolysis of CO 3 2− occurred, which yields many small nanoparticles of Cu 2 (OH) 2 CO 3 with size of 20-30 nm (before 30 min, not shown). Meanwhile, Sn 4+ ions are quickly hydrolyzed to produce SnO 2 nanoparticles. After 1 h reaction, these unstable or energetic nanoparticles have spontaneously built up solid microspheres with smooth outside surfaces (Figure 6a). Compared with these small particles in the exterior regions, those in the interior regions of the microsphere have higher surface energies, thus crystallites under the outer surface are preferentially dissolved through the process of inside-out Ostwald ripening. As a result, when the reaction duration is prolonged to 4 h, the solid SnO 2 -CuO composites evolve to be a typical hollow yolk/shell structure, as displayed in Figure 6b. Additionally, it is also found that the primary nanoparticles in the outer surfaces become orderly arranged, because they are easier to be crystallized as compared to those in the internal surface. With the reaction up to 8 h, driven by minimization of surface energy, the inner core becomes smaller gradually until its complete disappearance. At the same time, the dissolved internal particles are recrystallized into tiny nanosheets, which finally develop into the "bundles of thorns" stretching out from the surface of these hollow spheres ( Figure 6c). Eventually, after the reaction of 12 h, as depicted in Figure 6d, the initial solid structure has been evolved into a typical spherical structure with coarse surface anchoring with thick and large nanosheets. According to the above observations, this unique hollow structure evolution process follows the so-called "growth-cum-assembly" mechanism, which is depicted in Figure 6e. Such a process is very similar to that observed in the synthesis of CeO 2 -CuO hollow microsphere [20].   Table 2 summarizes the catalytic performances of all the catalysts in Rochow reaction measured at 325 • C for 24 h under atmospheric pressure. S-SnO 2 seems to be inactive. S-CuO also has a poor catalytic activity, with M2 selectivity and Si conversion being 61.2% and 7.9%, respectively. H-CuO shows a slightly better performance (68.6% of M2 selectivity and 13.1% of Si conversion), verifying the advantage of the hollow structure. The obvious improvement in catalytic performance is observed after introduction of SnO 2 into CuO. For H-SnO 2 (0.05)-CuO, the M2 selectivity and Si conversion increase to 81.1% and 20.7%, respectively. When the ratio of Sn/Cu increases from 0.05 to 0.2, a drastic enhancement in M2 selectivity (88.2%) and Si conversion (36.7%) are observed for the H-SnO 2 -CuO catalyst under the same conditions. This could be attributed to the generated synergetic interaction between CuO and SnO 2 [21]. To confirm it, a mechanically mixed sample of SnO 2 and hollow CuO (SnO 2 /CuO) with the same ratio of Sn/Cu as that in H-SnO 2 -CuO was prepared, which shows only a similar catalytic result to CuO alone. This fact suggests that the effective interfacial contacts between the CuO and SnO 2 components seems to be very important to improve the catalytic performance, while a simple physical mixing could not create such interface. However, when further increasing the amount of SnO 2 , the property of catalyst obviously becomes worse again. In the case of H-SnO 2 (0.4)-CuO, the selectivity of M2 and the conversion of Si are reduced to 72.1% and 16.4%, respectively. For all the catalysts investigated, the order of catalytic activity is as follows: H-SnO 2 -CuO > H-SnO 2 (0.1)-CuO > H-SnO 2 (0.4)-CuO > H-CuO > S-CuO > S-SnO 2 , suggesting that the excess or inadequate Sn content is adverse for the Rochow reaction. At the same time, to further confirm the superiority of H-SnO 2 -CuO, a commercial catalyst Cu-Cu 2 O-CuO (C-CAT) was also employed as a comparison. Although C-CAT has a more complex composition, it shows a poorer catalytic activity, with M2 selectivity and Si conversion being 78.0% and 16.0%, respectively, indicating that our H-SnO 2 -CuO catalyst possesses much higher activity and selectivity to synthesize M2 than the commercial catalyst. These results demonstrate the importance of both the hollow structure and the SnO 2 promoter to the superior catalytic performance. In addition, byproducts such as M1, M3, M1H, M2H, LBR and HBR are also detected but without any obvious distribution difference over the various prepared catalysts.    Figure 8a. It is found that, the composition of reacted contact masses are Si and Cu and no any CuO phase is detected, because of the reaction of MeCl with the lattice oxygen from CuO that leads to the formation of Cu in the induction period, which means the time of the transformation of the catalyst into real active phase under reaction conditions [27]. Figure 8b shows the SEM image, elemental mapping images and SEM-EDX spectra of the reacted contact masses by using H-SnO 2 -CuO catalyst, which indicates the uniform distribution of the elements Si, Cu and Cl along the interface. For the unreacted regions, the surfaces of Si particles are relatively smooth and flat; while for the reacted region (marked by yellow dotted ellipse), much obvious nick on Si surface indicate that the processing of reaction includes an etching process, which is in accordance with the so-called anisotropic etching reaction mechanics [28,29]. The previous literature reports have shown that Cu x Si species produced during the reaction are the necessities of active immediate in the M2 synthesis, therefore the formation of the active Cu x Si species is the crucial step in Rochow reaction [30][31][32]. Comparing with S-CuO, H-SnO 2 -CuO has smaller primary particles and larger surface areas, thus it can facilitate adsorption of MeCl molecules on its surface more efficiently. These adsorbed MeCl molecules will be dissociated rapidly into electronegative Cl − and electropositive CH 3 + [20,33]. As a result, the subsequent next step of forming active immediate Cu x Si species will be accelerated obviously on H-SnO 2 -CuO. In addition, the presence of Sn species on the CuO surface could reduce the activation energy of the reaction due to the synergistic effect [22], thus facilitating formation of more Cu x Si species. Based on the analysis above, a possible mechanism for the Rochow reaction using H-SnO 2 -CuO is proposed. Firstly, the reactant Si is mixed with H-SnO 2 -CuO catalyst to form the contact masses. When increasing the reaction temperature, this catalyst first reacts with MeCl to form Cu atoms, then those Cu atoms gradually diffuse into Si matrix and react with the generated Si atom, finally forming the active phase Cu x Si alloy at the interface. The formative Cu x Si species there may present modest adsorption and activation energy to MeCl, which subsequently transforms into M2. Finally, more elemental copper is generated with the reaction continues and the catalyst completely deactivates when Cu in the Cu x Si alloy is consumed.

Material Synthesis
All the reagents in this experiment were analytical grade and purchased from Sinopharm Chemical Reagent Co. Ltd., Beijing, China and used as received. The SnO 2 -CuO hollow microspheres were synthesized by a hydrothermal method. Typically, 1.2 g of Cu(CH 3 COO) 2 ·H 2 O (6 mmol) and a varied amount of SnCl 4 ·5H 2 O were dissolved in 60.0 mL of deionized water with continuous stirring at room temperature, followed by the dropwise addition of 20.0 mL of Na 2 CO 3 solution (0.3 M). The mixed solution was then transferred into a 100-mL Teflon-lined stainless steel autoclave, which was kept at 140 • C for 12 h. After cooling, the obtained precipitate was collected by filtration, washed with deionized water and ethanol several times and dried in air at 60 • C overnight. Finally, the obtained intermediate product was calcined in muffle oven at 400 • C for 3 h. The prepared samples are denoted as H-SnO 2 (x)-CuO (x = 0.05, 0.1, 0.2, 0.4), where H refers to "Hollow structure" and x refers to the weight ratios of Sn and Cu. Unless otherwise specified, H-SnO 2 -CuO in the text refers to H-SnO 2 (0.2)-CuO. For comparison, pure H-CuO and S-SnO 2 were prepared followed with similar procedures but with only Cu(CH 3 COO) 2 ·H 2 O or SnCl 4 ·5H 2 O added, respectively and the samples are denoted as H-CuO and S-SnO 2 . In addition, the synthesis procedure for S-CuO was the same as that for H-CuO except that the hydrothermal time was 1 h, where S refers to "Solid" structure.

Material Characterization
X-ray diffraction (XRD) analysis was performed on a PANalytica X'Pert PRO MPD (Almelo, the Netherlands) using the Kα radiation of Cu (λ = 1.5418 Å) and the crystallite sizes were calculated using the Debye-Scherrer equation. The dimension and morphology of the samples were characterized by field-emission scanning electron microscopy (FESEM) with energy-dispersive X-ray spectrometer (EDX) (JSM-7001F, JEOL, Tokyo, Japan). X-ray photoelectron spectroscopy (XPS) analysis was carried out on a Model VG ESCALAB 250 spectrometer, Thermo Electron, London, UK, using a non-monochromatized Al-Kα X-ray source (1486.6 eV). Thermal gravimetric (TG) analysis was executed on an EXSTAR TG/DTA 6300 (Seiko Instruments, Chiba, Japan). During the measurement, the temperature range was controlled from 30 to 900 • C and the heating rate was kept 2 • C/ min of in air (200 mL/min). The specific surface area was measured using the Brunauer-Emmett-Teller (BET) method and the relative pressure range was from 0.05 to 0.3.

Catalytic Test
The catalytic performance of the samples was evaluated with a fixed-bed reactor. 10.0 g of Si powder (20−50 mesh, provided by Tangshan Sanyou Group Co., Ltd., Tangshan, China) and 0.5 g of catalysts, together with 0.05 g of zinc (Zn, A.R., Sinopharm Chemical Reagent Co., Ltd., Tangshan, China) as the promoter were mixed homogeneously to form a fresh contact mass. N 2 gas flow (25 mL/min) was used to purge the reactor system for 30 min followed by heating to 325 • C within 1 h. Afterwards, the N 2 gas was turned off and MeCl gas was fed at a flow rate of 25 mL/min to react with Si for a settled time (12-72 h). All the reactions were carried out under atmospheric pressure and at 325 • C. After stopping the reaction, the gas products were cooled into liquid using the circulator bath kept at −5 • C by a programmable thermal circulator (GDH series, Ningbo Xinzhi Biological Technology Co., Ltd., Ningbo, China). Gas chromatograph (GC-7890A, Agilent Technologies, Santa Clara, CA, USA) equipped with a KB-201 capillary column (60 m) and a thermal conductivity detector (TCD) was employed to analyze the composition of liquid products, which were mainly composed of methyltricholorsilane (CH 3 SiCl 3 , M1), dimethyldichlorosilane ((CH 3 ) 2 SiCl 2 , M2), trimethylchlorosilane ((CH 3 ) 3 SiCl, M3), methyldichlorosilane (CH 3 SiHCl 2 , M1H), dimethylchlorosilane ((CH 3 ) 2 SiHCl, M2H), low boiler (LB) and high boiler (HB). The selectivity of products was determined by calculating the peak area ratio and the conversion of Si was determined by the mass change of Si powder before and after the reaction (Formula (4)).

Conclusions
To summarize, heterostructured SnO 2 -CuO hollow microspheres have been synthesized by a simple, template-free hydrothermal method coupled with calcination. In the resultant SnO 2 -CuO composite, SnO 2 particles are mainly enriched on the surface of CuO hollow microspheres, which are assembled into by many interconnected CuO nanoparticles in the interior of the microspheres and by the outer layer of CuO nanosheets. Moreover, the obtained hollow SnO 2 -CuO composite microspheres exhibit a higher selectivity to synthesize M2 and Si conversion than that of non-hollow structure and individual counterparts for the Rochow reaction, mainly attributed to the advanced structure and the synergistic interaction between CuO and SnO 2 . The clear structure-catalytic property relation gained in this work should be useful to the design of more-efficient multi-component Cu-based catalysts for the Rochow reaction.