Aerosol-Assisted Synthesis of Sn–Si Composite Oxide Microspheres with the Hollow Mesoporous Structure for Baeyer–Villiger Oxidation

Abstract

Tetravalent Sn species, such as zeolite or oxide, possess Lewis acidic properties, and thus exhibit prominent catalytic performance in several reactions when they are incorporated into the silica framework. Unfortunately, the synthesis of Sn-based zeolite (Sn–Beta) usually suffers from several drawbacks, including a long crystallization time, limited framework Sn content and complex synthesis steps. Sn-based composite oxides are favored in the industry, due to their simple synthesis steps and easy control of their pore structure, morphology and Sn content. In this work, an aerosol-assisted method is used to prepare Sn–Si composite oxide microspheres, using CTAB as template. The method is based on the formation of aerosol from a solution of Sn, Si precursors and a template (CTAB). The introduction of CTAB causes the surface tension of the atomized droplets to decrease. During the fast drying of the droplets, the Sn–Si composite oxide microspheres with a concave hollow morphology were first formed. After calcination, calibrated mesopores of 2.3 nm were also formed, with a specific surface area of 1260 m²/g and a mesopores ratio of 0.84. Sn species are incorporated in the silica network, mainly in the form of single sites. The resulting material proved to exhibit high catalytic performances in the Baeyer–Villiger oxidation of 2-adamantanone by using H₂O₂ as green oxidant, which was mainly attributed to the enhancement of the access to the catalytic tin sites through both the continuous hollow and mesopore channels, which have a 52% conversion of 2-adamantanone after 3 h of reaction. This method is simple, convenient, cheap and can be continuously produced, meaning it has broad potential for industrial application.

1. Introduction

Developments in catalysis science have often been triggered by innovation in the preparation of catalytic materials [1]. Of particular importance is the advent of mesoporous or hollow catalytic materials, serving either as catalysts or catalyst supports with interesting structural and textural features [2,3].As they exhibit intrinsic features—including low density, a high specific surface area, abundant inner void space and pores between 2–50 nm—these characteristics endow them with low mass transport resistance, short charge transport pathways and high catalytic activities.

Inspired by the above advantages, great effort has been made towards synthesis of mesoporous and hollow structures. Sol–gel methods, combined with templating strategies, are proven as a typical and powerful way to obtain such mesoporous and hollow materials with controlled composition, texture and surface functionalities [4,5]. Composite metal oxides that directly incorporate the catalytically active metal element can be prepared using such sol–gel approaches. These start from a solution of molecular precursors, which evolve into a suspension (sol) and then reticulate to form a gel via inorganic polycondensation reactions. Importantly, the introduction of sacrificial templating agents in the synthesis mixture makes it possible to generate a variety of textural features. For example, isolated sites of Sn oxides in silica catalyze important reactions like Baeyer–Villiger (B–V) oxidation [6,7] and biomass transformations [8,9]. However, conventional sol–gel procedures mostly rely on multi-step and time-consuming batch operations [10], which are inherently limited by low energy efficiency, substantial chemical waste and high cost.

Thus, it is crucial to develop a rapid continuous process for synthesizing composite metal oxides, where facile scale-up is possible. The “aerosol-assisted sol–gel process” (AASG) has been used to obtain various types of materials [11]. It involves a very limited number of preparation steps, produces material in a continuous mode, allows for a simple collection of the powder and generates low amounts of waste. The condensation of the inorganic network occurs during a short drying time with this method, with the evaporation-induced self-assembly of surfactant serving as a sacrificial template to generate a hollow and mesoporous structure [12,13].

Since the discovery of highly ordered mesoporous molecular sieves (M41S), soft templating methods have been widely used for the synthesis of mesoporous materials [14]. Soft templates, including silylated polymers [15], cationic organosilanes [16] and surfactants [17] are firstly imbedded within the precursor and then removed by calcination to free porosity. Among these templates, surfactants such as cetyltrimethylammonium bromide (CTAB) exhibit inherent flexibility and advantages, such as tunable size and adjustable functionality, and are cheaper and more readily scalable than other templates. More importantly, the combination of the soft template and aerosol method makes the metals in the obtained composite metal oxides highly dispersed [12]. For example, Geng et al. reported the preparation of a noble metal nanoparticles supported mesoporous oxide (Au/CeO₂ and Pd/Al₂O₃), on the basis of aerosol-assisted surfactant (CTAB) strategy, which presented high activity and good endurance [18]. Bai et al. synthesized Ce–Si composite metal oxides by the one-step fast aerosol process [19]. These catalysts presented excellent catalytic performance and long-term stability in the catalytic incineration of VOCs. This was due to the highly specific surface area with a highly ordered pore structure, as well as CeO₂ particles that were well dispersed on the porous surfaces.

Considering the importance of Sn-based catalysts in various organic reactions, we focus on the aerosol process for the preparation of Sn–silicate (Sn–Si) composite oxide dedicated to Baeyer–Villiger oxidation, which is an important organic synthesis for the production of valuable esters or lactones [6,20,21]. It has been reported that transition metal-based catalysts have certain catalytic activity [22,23,24,25]. Among them, frameworks incorporated with Sn(IV) active sites, such as Sn-based zeolites [26] and Sn-based composite oxides [27], present excellent catalytic activities. Our group developed an aerosol-assisted hydrothermal method to synthesize zeolites [28]. The Sn–Si composite oxide produced by the aerosol process acted as the precursor for the further crystallization of Sn–Beta zeolite [29]. It was found that the obtained Sn–Beta zeolite exhibited excellent catalytic performance in the Baeyer–Villiger oxidation reaction with ketones and H₂O₂, which was due to the fine Lewis acid sites and its hierarchical structure. Moreover, this method is more suitable for the synthesis of the heteroatom substituted zeolite, since heteroatoms can be uniformly dispersed in the aerosol precursor. However, their preparation conditions are relatively strict, including long crystallization cycles, complex preparation steps and the need to be synthesized in an environment containing F⁻, which limits their practical applications. From this perspective, Sn–Si composite oxide catalysts with both hollow and mesoporous structures were synthesized by a simple and continuous route, which provided both outer surfaces and inner surfaces for reaction sites and short diffusion paths, which may therefore demonstrate enhanced activity levels.

In this work, Sn–Si composite oxide microspheres with a hollow mesoporous structure were synthesized by a simple one-pot aerosol process, using a very common commercial surfactant cetyl trimethyl ammonium bromide (CTAB) as a template. By optimizing the synthesis conditions—such as the amount of CTAB and Sn content—and characterizing by TEM, SEM, N₂ physical adsorption and UV-Vis spectroscopy, the formation mechanism was explored. The Baeyer–Villiger oxidation of bulk 2-adamantane with H₂O₂ was used to evaluate its catalytic performance.

2. Results and Discussion

The Sn–Si composite oxide microspheres with a hollow mesoporous structure were synthesized with a simple one-pot aerosol process, using CTAB as a template (Scheme 1). Comparing with the traditional sol–gel approaches, this aerosol method presents many advantages, such as the possible formation of a solid catalyst in a continuous way without additional hydrothermal treatments, low waste production, relatively short synthesis time and the easy collection of the solid. Such advantages make it possible to consider the catalyst preparation as a sustainable procedure. Therefore, it is necessary to systematically investigate several key factors, such as the amount of CTAB and Sn content in synthesis.

2.1. Effect of the Amount of CTAB

The effect of different amounts of CTAB were investigated, with the Si/Sn ratio first fixed to 30. Figure 1 shows the low angle XRD patterns of the Sn–Si-30-y samples, where the CTAB/SiO₂ ratio (y) ranges from 0 to 0.3. It can be seen that the sample synthesized without the addition of CTAB (Sn–Si-30-0) has no diffraction peaks, indicating the absence of a regular mesoporous structure [30]. With the introduction of CTAB, a clear d₁₀₀ crystal plane (2 theta = 2°–3°) appeared, indicating the successful introduction of regular mesopores. The Sn–Si-30-0.2 shows the highest peak intensity, while the characteristic diffraction peak of d₂₀₀ crystal plane (2 theta = 4.2°) also appears, indicating that it has the highest degree of mesoporous order. The above results clearly demonstrate that the amount of CTAB has an obvious influence on the structural ordering of the samples.

The effect of the amount of CTAB on its morphology can be characterized and analyzed by TEM, as shown in Figure 2. The Sn–Si-30-0 exhibits a regular solid spherical morphology with a particle size of approximately 2 µm, as shown in Figure 2A. When a small amount of CTAB is added (Sn–Si-30-0.05, Figure 2B), the morphology of the microspheres changes from a regular solid sphere to a concave hollow structure. As the CTAB content increases (Sn–Si-30-0.1, Figure 2C), the area of the concave inside the sphere gradually increases. When the CTAB/SiO₂ ratio is 0.2 (Sn–Si-30-0.2, Figure 2D), the sample’s morphology changes to a hollow morphology with surrounding depressions, and the wall thickness of the amorphous SiO₂ framework is about 560–650 nm. Further increases in the content of CTAB (Sn–Si-30-0.3, Figure 2E), bring no further significant change to the morphology of the sample. The above results indicate that CTAB has a significant impact on the morphology of the Sn–Si composite oxide microspheres. By increasing the amount of CTAB, the morphology of the samples gradually transformed from the regular solid sphere morphology to the concave hollow structure. The formation mechanism will be discussed in detail later.

The pore structures of the Sn–Si-30-y samples were characterized by N₂ physical adsorption, as shown in Figure 3, and the texture properties are summarized in

Table 1. The textural properties of materials by N₂ physical adsorption.

Entry	Sample	Si/Sn a	SBET b (m2/g)	Vtotal c (cm3/g)	Vmeso d (cm3/g)	Pore Size e (nm)	Vmeso/Vtotal f
1	Sn–Si-30-0	30	--	--	--	--	--
2	Sn–Si-30-0.05	31	303	0.17	0.12	2.3	0.71
3	Sn–Si-30-0.1	28	678	0.40	0.32	2.3	0.80
4	Sn–Si-30-0.2	29	1236	0.76	0.63	2.4	0.84
5	Sn–Si-30-0.3	33	1565	0.92	0.75	2.4	0.81
6	Sn–Si-60-0.2	57	1260	0.73	0.61	2.3	0.84
7	Sn–Si-90-0.2	97	1279	0.71	0.59	2.2	0.83
8	Sn–Si-125-0.2	119	1284	0.71	0.59	2.2	0.82

^a Determined by ICP-OES; ^b BET surface area; ^c p/p₀ = 0.99; ^d mesopore volume, V_meso = V_total − V_micro; ^e average pore size; ^f fraction of mesopore volume.

. The sample synthesized without the addition of CTAB (Sn–Si-30-0) is a nonporous material, consistent with TEM characterization. The rest of the samples show a typical IV/H4-type reversible isotherms in the pressure region of p/p₀ > 0.4, which are characteristic of mesoporous materials (Figure 3A). Samples with a small amount of CTAB (Sn–Si-30-0.05 and Sn–Si-30-0.1) have smaller hysteresis loops and mesopore volumes (Table 1). However, with increases in the amount of CTAB, Sn–Si-30-0.2 and Sn–Si-30-0.3 samples exhibit significant hysteresis loops, and their mesopore volumes also increase to 0.63 and 0.75 cm³/g, respectively. The BJH pore size distribution (Figure 3B) of the CTAB-doped samples show a narrow mesopore centered at 2.1 nm, which may due to the thermal removal of CTAB micelles used as structure directing agent [19]. In addition, it can be seen from Table 1 that, with the increase of CTAB content, the specific surface area, total pore volume and mesopore volume of the samples all significantly increase, which is consistent with the results observed by TEM. The fraction of mesopores volume in total volume (V_meso/V_total) of Sn–Si-30-0.2 is the highest (0.84), indicating the highest number of mesopores.

The effect of the amount of CTAB on the coordination state of the Sn species can be characterized by the UV-Vis spectra, as shown in Figure 4. All samples exhibit obvious absorption peaks at 195 nm, which are attributed to the tetrahedral coordination Sn(IV) within the amorphous silica framework [31]. The presence of small contributions at higher wavelengths (325 nm) can be assigned to extra-framework SnO₂ particles [31,32]. It is known that the synthesis of Sn–Si mixed oxide proceeds via hydrolysis of the precursors, followed by condensation, with the consequent formation of Sn–O–Si and Si–O–Si bonds, which are responsible for the formation of framework Sn(IV) species. However, the high Sn content in the precursor (Si/Sn = 30) results in a small amount of Sn species failing to incorporate the framework and forming extra-framework SnO₂ particles during the spray drying or calcination process. Moreover, the broad absorption peak above 300 nm belongs to the extra-framework SnO₂ particles, and this is more prominent when the content of CTAB is high (Sn–Si-30-0.2 and Sn–Si-30-0.3), indicating the steric hindrance of the straight chain (C16) in CTAB, which hinders the entry of Sn species into adjacent silicon framework. In addition, the actual Si/Sn molar ratios of the Sn–Si-30-y samples (Table 1) is very close to those of the theoretical feed ratio, indicating the loss of the Si and Sn source is avoided during the aerosol and calcination process. The above results indicate that the introduction of CTAB had little effect on Sn content, but had an obvious effect on the coordination state of Sn species on the silica framework.

2.2. Effect of Sn Content

In heterogeneous catalysis, the morphology, pore structure, the number of active centers and the coordination state of the catalyst can all affect catalytic performance. It is known that framework Sn(IV) species is the catalytic active center for B–V oxidation reaction [33]. However, extra-framework SnO₂ species are also present in Sn–Si-30-y samples (Figure 4), which may impede the catalytic performance of the sample. Therefore, on the basis of the Sn–Si-x-0.2 samples, it was also necessary to optimize the range of the Si/Sn ratio.

The effect of the Sn content on the morphology of the Sn–Si-x-0.2 samples was characterized and analyzed by SEM, as shown in Figure S1. All the samples exhibit a smooth conical framework and a concave hollow spherical morphology. This indicates that the Sn content has little impact on the morphology of the Sn–Si composite oxide microspheres.

Figure S2 shows the N₂ physical adsorption isotherms and BJH pore size distribution of the Sn–Si-x-0.2 samples. The Sn–Si-x-0.2 exhibit typical IV-Type adsorption isotherms and H₄ hysteresis loops. The BJH pore size distribution shows that the pore size of Sn–Si-x-0.2 is concentrated 2.2 nm, as shown in Figure S2B. The specific surface area, pore volume and proportion of mesopores of the material do not show the obvious changes (Table 1), indicating the influence of Sn content on the pore structure and textural properties of the Sn–Si-x-0.2 samples are negligible.

Figure 5 shows the UV-Vis spectra of Sn–Si-x-0.2 series samples in order to determine the Sn coordination. All samples exhibit obvious absorption peaks at 195 nm, which are attributed to the tetrahedrally coordinated framework Sn(IV) within the amorphous silica framework [31]. However, only the sample with an Si/Sn ratio of 30 (Sn–Si-30-0.2) exhibits a significant shoulder peak at 325 nm, which can be assigned to the extra-framework SnO₂ particles [32]. Moreover, the absorption peak intensity at 195 nm of Sn–Si-60-0.2 was the highest, indicating the highest framework of Sn(IV)content. The above results show that most of the Sn species from the precursor solution can be incorporated into amorphous silica framework as framework Sn(IV) species during the aerosol spray drying process, while extra-framework SnO₂ species may also form when the Si/Sn ratio is too high in the precursor solution.

2.3. Transformation Processes and Formation Mechanism

On the basis of the above results, it is speculated that CTAB plays a crucial role in the changes from a solid microsphere to concave hollow morphology. Although the use of an aerosol-assisted template strategy to synthesize metal mixed oxide with various morphology has been reported elsewhere [11], a concave hollow morphology has not been obtained. Our previous study used this hollow mesoporous Sn–Si composite oxide microsphere (SnSi-x-y) as the precursor and then hydrothermally crystallized to obtain Sn–Beta zeolite [34]. However, the transformation processes and formation mechanism are still ambiguous. Therefore, it is essential to get a deep insight into the formation mechanism of the whole synthesis process.

The synthesis is very simple, derived from a precursor solution containing an Sn source, Si source, H₂O and CTAB, transformed into the Sn–Si composite oxide through the aerosol process and calcination. Interestingly, the as-synthesized Sn–Si composite oxide (uncalcined), obtained through the aerosol process, presented a concave hollow morphology (Figure S3) and no obvious change after calcination (Figure S1A), indicating that a concave hollow morphology was formed during the aerosol process, rather than calcination. Therefore, we focus on the whole aerosol spray drying process next. This process can be divided into three steps, as shown in Figure 6A. Firstly, the precursor solution is fed into the atomizer through a peristaltic pump for conversion to droplet form, with diameters ranging between 1 and 100 µm. Secondly, the atomized droplets enter the heating chamber via the high-speed and high-pressure carrier gas (compressed air) and undergo evaporation and solute condensation within the droplet. This step results in the solvent removal, while the desired components remain in the final product. The final product can be collected after a cyclone separator.

In the first step, the precursor solution is atomized through an air flow atomizer to form single and homogenous aerosol droplets. The driving force is the large relative speed between the flow rate of precursor solution (low) and gas carrier (high), which result in the liquid film being drawn into filaments, and then split into small droplets. The size of the droplets depends on the relative speed and the viscosity of precursor solution [35].

The obtained aerosol droplets enter the heating chamber via the flow of carrier gas and form particles by solvent evaporation. It is noted that the difference in morphology of the produced particles can be considered to be the result of an initial deformation of the droplet [13], and the structural stability of atomized droplets can be explained by Bond’s number β, the ratio of the inertial force and surface tension effects. The Bond number is given by:

$$\mathsf{\beta }=\frac{∆\rho a{d}^2}{\sigma }$$

(1)

where $∆\rho$ is the difference in the densities of the droplet and the surrounding fluid, a is the acceleration (changing gas velocity), d the droplet size and σ is the interfacial tension/surface force [36]. Spherical particles are typically generated using the aerosol spray method, because the maximum structural stability of the droplet is in a spherical form (values of β tends to 0), which is the fundamental reason for producing this shape. However, if a surfactant (CTAB) is added in the precursor solution, the surface tension of the droplets ($\sigma )$ decreases, resulting in an increase in Bond’s number β, which finally makes the droplet become unstable. Therefore, under the action of high-speed carrier gas, the droplet deforms when entering the heating chamber, forming a flat “mushroom like” shape (Figure 6B). With the help of high temperature of heating chamber, the solvent of the deformed droplets begins to vaporize. Solvent evaporation from the droplet induces capillary flow, and the capillary forces cause self-assembly condensation of the solutes into the close-packed arrays with a nearly spherical shape (Figure 6D). After the solvent in the droplet completely evaporates, the concave hollow morphology is revealed (Figure 6F). After the final calcination to remove CTAB, the concave hollow morphology of the sample was preserved, while abundant mesopores were generated within the amorphous framework.

After understanding the above process, it is easier to comprehend the sample without the addition of CTAB (Sn–Si-30). In this case, the surface tension of the droplets remains high, and the purge of the carrier gas does not deform the droplets, which still maintain their spherical shape (Figure 6C), After solvent evaporation, the atomized droplets self-assemble via condensation in order to maintain the minimum tension (Figure 6E,G).

2.4. Catalytic Evaluation

The isolated Sn active sites in the silica framework demonstrate Lewis acidity, making this material capable of catalyzing various catalytic reactions [37]. Hence, the acidity of the Sn–Si composite oxide hollow mesoporous microspheres was also detected before being utilized as catalyst. The NH₃-TPD of the Sn–Si-30-0, Sn–Si-60-0.2 and Sn–Si-30-0.2 samples are shown in Figure 7. The wide desorption peaks of NH₃ on weak acid sites are observed at ca. 150–350 °C for all samples. No additional desorption peaks at above 350 °C were observed, indicating the absence of strong acid sites [38]. The number of weak acid sites of all samples are given in Table S1, and are calculated based on the peak areas of NH₃ desorbed on weak acid sites obtained from TPD profiles. Sn–Si-30-0 was only 67 μmol NH₃/g, whereas it increased sharply to 200 μmol NH₃/g for the Sn–Si-60-0.2 sample, which can be clearly attributed to the hollow mesopore structure being exposed on more acidic Sn sites. It is noted that, although Sn–Si-30-0.2 has the highest Sn content, the acidity dropped to 182 μmol NH₃/g, due to formation of extra SnO₂ species with a less acidic framework. This is also in accordance with the UV-Vis characterization shown in Figure 5.

The Baeyer–Villiger oxidation of bulk 2-adamantanone with H₂O₂, which was catalyzed by Lewis acid, was performed in order to illustrate the advantages of the Sn–Si composite oxide with hollow mesoporous microspheres catalyst in terms of diffusion, as shown in

Table 2. Comparison of Baeyer–Villiger oxidation of 2-adamantane with H₂O₂ among varying Sn–Si composite microspheres and other Sn-based catalysts ^a.

Entry	Catalyst	Conversion (%)	Selectivity (%)	Yield (%)
1	black	0	0	0
2	Sn–Si-30-0	7	>99	7
3	Sn–Si-60-0	2	>99	2
4	Sn–Si-30-0.05	10	>99	10
5	Sn–Si-30-0.1	14	>99	14
6	Sn–Si-30-0.2	29	>99	29
7	Sn–Si-30-0.3	22	>99	22
8	Sn–Si-60-0.2	52	>99	52
9	Sn–Si-90-0.2	28	>99	28
10	Sn–Si-125-0.2	25	>99	25
11 b	Sn–SBA-15	50	92	46
12 c	Sn–MCM-41	49	91	45
13 c	Sn–MSNSs	87	97	85

^a Reaction conditions: cat., 0.05 g; 2-admantanone, 2 mmol; H₂O₂ (30%), 4 mmol; 1,4-dioxane, 10 mL; temp, 363 K; time, 180 min; ^b according to Ref. [40]; ^c according to Ref. [27].

. Note that all the catalysts present as high as >99% selectivity to the lactone, which should be due to its rigid structure, consistent with our previous reports [29,34,39]. The reaction did not occur without a catalyst being added, which indicates that the oxidation of H₂O₂ is insufficient (Table 2 Entry 1). A negligible catalytic performance was observed for the catalysts synthesized without addition of CTAB (Table 2 Entry 2–3), which result in the catalytic reaction only occurring on the outer surface of the catalyst due to their nonporous solid structure. With the introduction of CTAB and the increase of the addition amount, the conversion of 2-amantadone gradually increased (Table 2 Entry 4–6). When the molar ratio of CTAB/SiO₂ reaches 0.2 (Sn–Si-30-0.2), the conversion of 2-amantadone reaches the highest (29%) among the Sn–Si-30-y serious catalysts. This may be attributed to the sharp increase in the specific surface area, pore volume and the number of mesopores (Table 1), as well as the concave hollow morphology (Figure 2), which not only reduces mass-transfer limitation of the catalyst, but also exposes more active sites. However, further increasing the amount of CTAB (Sn–Si-30-0.3) decreases the conversion of 2-amantadone (Table 2 Entry 7). Although this catalyst also presents a hollow mesoporous structure with the largest specific surface area, excessive CTAB hinders the entry of Sn species into an adjacent silicon framework and produces obvious SnO₂ particles, resulting in a decrease in the number of active centers (Figure 4). Surprisingly, the catalytic performance can be further improved by tuning the Sn/Si ratio in the Sn–Si-x-0.2 samples (Table 2 Entry 8–10). Sn–Si-60-0.2 shows the best catalytic performance, with 52% conversion, nearly eight times higher than that of its parent (Sn–Si-30-0), which can be ascribed to the highest Sn active site (framework Sn(IV)) content (Figure 5) and the number of weak acid sites (200 μmol NH₃/g) in this catalyst. These results demonstrate that the unique hollow mesopore structure of the Sn–Si composite oxide microspheres plays a crucial role in the tremendous increases of activity, while the number of Sn active sites also make a considerable contribution.

It is known that Sn incorporated in ordered mesoporous materials (Sn-SBA-15 and Sn-MCM-41) can be used as a typical Sn-based mesoporous catalyst, which shows good catalytic activity in the Baeyer–Villiger oxidation of 2-adamantane with H₂O₂ [27,40]. A similar catalytic activity was observed when Sn-SBA-15 and Sn-MCM-41 are compared with Sn–Si-60-0.2 (Table 2 Entry 8, 11–12). However, in terms of synthesis methods, the advantages of our synthesis strategy (aerosol route) are its simplicity and the short time it takes to carry out. It is noted that mesoporous silica nanospheres containing Sn (Sn-MSNSs) exhibited much higher activity in the Baeyer–Villiger oxidation of 2-adamantane (Table 2 Entry 13), which may benefit from the uniform crater-like mesopores and strong hydrophobicity [40]. Therefore, improving the hydrophobicity of our catalyst is also expected to improve its catalytic performance, which will also be our next work.

B–V oxidation involves mainly two steps: (1) the addition of the hydrogen peroxide oxidant and the ketone substrate to form the “Criegee” intermediate; and (2) B–V rearrangement of the Criegee intermediate to the ester product [6]. In this case, the ketone carbonyl molecule firstly coordinates to the Sn site on the Sn–Si composite oxide microspheres and is then activated by the Lewis acid site of the Sn species, resulting in the increased electrophilicity of carbonyl carbon. Subsequently, the electrophilicity enhancement of the carbonyl carbon facilitates the nucleophilic attack by a H₂O₂ molecule to form a “Criegee” intermediate [21]. Finally, the ester is formed following the intermediate rearrangement. The unique hollow mesopore structure, combined with the high Sn active site (framework Sn(IV)) content of the Sn–Si composite oxide microspheres catalyst, gives the B–V oxidation of bulk 2-adamantanone with H₂O₂ not only a high conversion rate but also high chemoselectivity.

Another key factor for the catalyst evaluation is reusability. The Sn–Si-60-0.2 catalyst after reaction can be regenerated by continuous filtration and calcination for the next run. Unfortunately, an obvious decrease in 2-adamantanone conversion is observed after being run five times (Figure 8). A comparison of texture properties derived from N₂ physical adsorption between the fresh and regenerated catalysts (named Sn–Si-60-0.2-5) is shown in Table S2. Slight decreases in the BET specific surface, total volume and mesoporous volume are observed of the regenerated catalyst, indicating a partial collapse of the mesopore. The UV-Vis spectra of the regenerated catalyst showed the sole absorption peak intensity at 195 nm, which was attributed to a decrease in the framework Sn(IV) species. The actual Si/Sn molar ratio of the regenerated catalyst was also decreased (Table S2), which reveals a loss of the framework Sn(IV) species (Figure S5). Therefore, the loss of activity (~24%) should be due to the collapse of mesoporous structures, which results in the loss of the Sn active site (framework Sn(IV) species). Our previous study indicated that Sn–Beta zeolites show a good reusability, which may be due to their excellent hydrophobicity [29,34,39]. Thus, the reusability of Sn–Si composite microspheres in this study can be improved by increasing their hydrophobicity.

3. Materials and Methods

3.1. Materials

Tetraethyl orthosilicate (TEOS, analytically pure, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China); stannic chloride pentahydrate (SnCl₄·5H₂O, analytically pure, Kermel Chemical Reagent Co., Ltd., Tianjin, China); chlorobenzene (C₆H₅Cl, analytically pure, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China) and 1,4-dioxane (C₄H₈O₂, analytically pure, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China); hexadecyl trimethyl ammonium bromide (CTAB, Sinopharm Chemical Reagent Co., Ltd., Shanghai, China); 2-adamantanone (analytically pure, Aladdin Biochemical Technology Co., LTD, Shanghai, China); hydrochloric acid aqueous solution (HCl, 37%, Fuyu Fine Chemical Co., Ltd., Tianjin, China); hydrogen peroxide (H₂O₂, 30%, Damao Chemical Reagent Factory, Tianjin, China); deionized water (DI, lab-made). The above chemical reagents had not been further purified.

3.2. Method

The Sn–Si-x-y composite oxide microspheres—where “x” and “y” represented Si/Sn and CTAB/Si molar ratio in the initial solution, respectively—were configured according to the molar composition ratio of 1.0 SiO₂: (0.008–0.033) SnO₂: (0–0.2) CTAB: 20 H₂O: 0.15 HCl. Generally, a certain amount of tin (IV) chloride (SnCl₄·5H₂O) was dissolved in a hydrochloric acid (HCl, 2.3 g) aqueous solution (55 g), then 32 g of tetraethyl orthosilicate (TEOS) was added to yield solution A. Cetyltrimethylammonium bromide (CTAB, 11 g) was dissolved in DI water (56 g) to yield solution B. Both solutions were left stirring for at least 2 h at room temperature. After that, solutions A and B were mixed together and then stirred for another 2 h to form clear precursor solution.

The resultant transparent solution was transformed into aerosol particles by an aerosol apparatus (BL-6000Y, Shanghai BILON Instrument Co., Ltd., Shanghai, China). Droplets were formed by the atomizer at the delivery rate of 30 rpm. SnO₂–SiO₂ composite oxide powders were formed during a fast evaporation of water at 483 K. Last, these amorphous powders were collected by cyclone separator at a frequency of 60 Hz and then dried at 373 K for 3 h and calcined under air atmosphere at 773 K for 6 h, as shown in Scheme 1. For comparison, Sn–Si-x-0 sample was also synthesized according to the above steps, without adding CTAB.

3.3. Catalyst Characterization

Powder X-ray diffraction (XRD) patterns were recorded on a Shimadzu Rigaku D/Max 2400 diffractometer (Kyoto, Japan) with Cu–Kα radiation at 40 kV and 10 mA. The diffraction diagrams of the samples were recorded in the 2 theta range of low-angles at 0.6–5°, with steps of 0.4° and a count time of 60 s at each point. The morphology and structure of the samples were characterized by transmission electron microscope (TEM, HT7700 EXALENS, Hitachi, Tokyo, Japan) operating at 100 kV and scanning electron microscopy (SEM, Carl Zeiss Jean company, Jena, Germany) operating at 5 kV. The porous structures of all the materials were measured by N₂ physical adsorption isotherm by the ASAP 2020 fully automatic physical adsorption analyzer from the Micromeritics Instrument Corporation (Norcross, GA, USA). The samples were treated at 573 K for 3 h prior to the analysis. The size distribution was calculated from the adsorption isotherm using the Barrett–Joyner–Halenda (BJH) method. The Sn loadings were determined by inductively coupled plasma (ICP) analysis on Perkin Elmer Optima 2000DV (Waltham, MA, USA) inductively coupled plasma atomic emission spectroscopy. The coordination status of Sn was detected by the UV-550 ultraviolet spectrophotometer from the Jasco Company (Tokyo Japan), and the barium sulphate powder was used as the reference. NH₃-TPD was carried out on a AutoChem III chemisorption analyzer (Micromeritics Instrument Corporation, Norcross, GA, USA) to measure the acid proper-ties of the samples. Before the measurements, the sample was pretreated at 600 °C in a He flow for 1 h. Subsequently, the pure ammonia was adsorbed on the samples at 100 °C for 10 min. After purging in flowing He at 100 °C for 2 h, the TPD process was carried out from 100 to 600 °C at a heating rate of 10 °C min⁻¹. The desorbed NH₃ was quantified using a thermal conductivity detector (TCD, Micromeritics Instrument Corporation, Norcross, GA, USA).

3.4. Performance Evaluation

The Baeyer–Villiger oxidation of 2-adamantane and H₂O₂ was performed in a 25 mL round-bottomed flask with a water condenser. A total of 3 mmol of 2-adamantanone, 10 mL of 1,4-dioxane, 50 mg catalyst, 0.5 g chlorobenzene (GC internal standard) and 4 mmol H₂O₂ (30 wt%) were mixed and stirred in the flask. The reaction was conducted at 363 K for 3 h. After cooling, the reaction mixture and catalyst were separated by centrifugation, while the liquid phase was analyzed by gas chromatography (GC-7890, Techcomp, Shanghai, China), equipped with an SE-54 (30 m × 0.32 mm × 0.50 μm) capillary column and an FID detector. The conversion of 2-adamantane and the selectivity of lactones were calculated accordingly.

The recycling tests of the Baeyer–Villiger oxidation of 2-adamantane reaction was performed under above-mentioned conditions over the Sn–Si-60-0.2 sample. After the first reaction run, this sample was centrifuged and washed with ethanol and DI water three times, respectively. Afterwards, the obtained sample was dried at 373 K overnight and calcined at 773 K for 6 h. The recycling test was repeated 5 times in total.

4. Conclusions

In conclusion, the Sn–Si composite oxide microspheres with both a concave hollow morphology and abundant mesopore structure were successfully synthesized via a simple aerosol combined surfactant (CTAB) approach. Several key synthesis parameters, such as the amount of CTAB and the Si/Sn molar ratio, were systematically studied. It was found that the addition of CTAB has a great effect on the physicochemical properties and the formation of concave hollow morphology of the microspheres. The condition-optimized Sn–Si composite oxide microsphere (Sn–Si-60-0.2) showed remarkably enhanced catalytic performance for the Baeyer–Villiger oxidation of 2-adamantane, nearly 8 times higher than that of its parent (Sn–Si-30-0). This may mainly be due to the interesting features such as high surface area (1260 m²/g), unique hollow mesopore structure, the highest Sn active site (framework Sn(IV)) content and number of weak acid sites (200 μmol NH₃/g) in the catalyst, which not only reduced the mass-transfer limitation of the catalyst, but also exposed more acidic Sn sites. This easy, one-pot and continuous aerosol assisted method presents great potential for developing more efficient industrially relevant heterogeneous catalysts.