Key Factor Study for Amphiphilic Block Copolymer‐Templated Mesoporous SnO2 Thin Film Synthesis: Influence of Solvent and Catalyst

As a crucial material in the field of energy storage, SnO2 thin films are widely applied in daily life and have been in the focus of scientific research. Compared to the planar counterpart, mesoporous SnO2 thin films with high specific surface area possess more attractive physical and chemical properties. In the present work, a novel amphiphilic block copolymer‐assisted sol–gel chemistry is utilized for the synthesis of porous tin oxide (SnO2). Two key factors for the sol–gel stock solution preparation, the solvent category and the catalyst content, are systematically varied to tune the thin film morphologies. A calcination process is performed to remove the polymer template at 500 °C in ambient conditions. The surface morphology and the buried inner structure are probed with scanning electron microscope and grazing‐incidence small‐angle X‐ray scattering. Crystallinity is characterized by X‐ray diffraction. The multi‐dimensional characterization results suggest that cassiterite SnO2 with spherical, cylindrical, and vesicular pore structures are obtained. The variation of the film morphology is governed by the preferential affinity of the utilized solvent mixture and the hydrogen bond interaction between the employed cycloether and H2O molecules in the solution.

mesoporous SnO 2 thin films will be a crucial factor for realworld application.
Typically, the following requirements are supposed to be met during the preparation of metal oxide thin films with block copolymer-assisted sol-gel chemistry. A good solvent for both polymer blocks is required for dissolving the block copolymer and a poor solvent for one of the blocks is required for stimulating the phase separation of the sol-gel system. To avoid the occurrence of macroscopic phase separation in the sol-gel solution, a good miscibility between good and poor solvent is necessary. In order to integrate the functional inorganic part into the micro-phase-separated block polymer network, a hydrogen bond interaction between the precursor molecules and a specific segment of the block copolymer template is expected. [48] According to previous studies, the factors that affect the microstructure of the block copolymer-templated metal oxide thin films include the reaction time, [49] the component content, [50][51][52] the surface conditions of the substrate, [53] the operational environment, [54] and the way of removing the polymer template. [55] In the present study, the amphiphilic diblock copolymer polystyrene-block-poly(ethylene oxide) (PS-b-PEO) is utilized as template and SnCl 4 is utilized as precursor to fabricate mesoporous SnO 2 thin films. A morphology control of the SnO 2 thin films is realized by changing two key factors in the sol-gel stock solution: the content of hydrochloric acid catalyst and the category of the organic solvent (tetrahydrofuran (THF) or 1,4-dioxane). As a poor solvent to the PS blocks, the preferential affinity of the solvent to the different polymer blocks can be significantly changed with the addition of hydrochloric acid. Through the complexation effect between SnCl 4 and the solvent molecules (THF or 1,4-dioxane), the SnCl 4 precursor can be temporarily frozen-in before adding hydrochloric acid as catalyst. This leads to the sustained release and a controlled hydrolysis of SnCl 4 . As a consequence, the formation of big clusters due to the rapid hydrolysis of SnCl 4 is significantly inhibited and a mesoscale structure can be formed. Due to the synergistic effect of hydrochloric acid and the organic solvent, different mesopore structures such as spherical, cylindrical, and vesical structures are successfully obtained. The thin film morphologies in the local and large area are probed with scanning electron microscopy (SEM) and grazing-incidence small angle X-ray scattering (GISAXS). By correlating different characterization results with the thin film preparation method, the thin film morphology evolution is understood to originate from the preferential affinity of the utilized solvent mixture and the hydrogen bond interaction between the cycloether and H 2 O molecules.

Result and Discussion
The complexation between SnCl 4 and the solvent molecules (THF or 1,4-dioxane) is schematically shown in Figure 1a. The existence of these complexes prevents the intense hydrolysis of SnCl 4 from the addition of hydrochloric acid, which accordingly restricts the aggregation of the hydrolyzed (HO) X SnCl 4−X (0 ≤ x ≤ 4) species in the sol-gel solution. Through hydrogen bond interaction, the (HO) X SnCl 4−X species are specially confined into the hydrophilic PEO domains of the phase separated system (Figure 1b). The common phase-separated structures, such as spheres, cylinders, lamella, and vesicles, can be formed by the PS-b-PEO polymer template and the hydrolyzed products in the sol-gel system are demonstrated in Figure 1c.

Crystallinity of the Calcined Thin Films
For a specific chemical reaction, the equilibrium constant K is a measure of the ratio of the concentrations of products to the concentrations of reactants, which can be calculated by its corresponding standard Gibbs free energy (ΔG 0 ) with R is the gas constant with a value of 8.314 J K −1 mol −1 , T is the temperature of the reaction, ΔG 0 is the Gibbs free energy Figure 1. a). Complexation between solvent molecules and SnCl 4 precursor; b) hydrogen-bond interaction between the hydrolyzed product of SnCl 4 and the PEO block; c) common phase-separated structures formed by amphiphilic diblock copolymers, PS blocks, PEO blocks, and hydrolyzed (HO) X SnCl 4−X (0 ≤ x ≤ 4) species are shown in blue, red, and white colors, respectively. change per mole of reaction for unmixed reactants and products at standard conditions (i.e., 298 K, 100 kPa, 1 m of each reactant and product). A study about the temperature dependence of the equilibrium constant K showed that the ln K value for reaction (2) was positive after the temperature reached 100 °C, which favors the formation of the SnO 2 product at equilibrium [58] SnCl l 2H O l SnO 4HCl g With a custom-built 2D small-angle X-ray scattering setup, Smarsly and co-authors demonstrated the pronounced crystallization of SnO 2 synthesized with poly(ethylene-co-butylene)block-poly(ethylene oxide) template to occur at around 300-350 °C. [49] In order to determine the crystallinity of the calcined SnO 2 thin films prepared in the present work, X-ray diffraction (XRD) measurement is performed (Figure 2 where λ denotes the wavelength of the X-rays, β and θ refer to the full-width at half-maximum and Bragg angle of the peaks, respectively. The average crystal sizes of the SnO 2 thin films are calculated from the (110), (101), and (211) peaks of the XRD patterns and the corresponding results are listed in Table 1.
It is observed that the grain sizes of all SnO 2 thin films are around 5 nm and independent of the utilized solvent categories and hydrochloric acid content added into the sol-gel solution. The grain sizes listed in Table 1 are likely determined by the high calcination temperature, which is in accordance with the experimental result of previous work. [38] In addition, the thin films acquired with THF solvent exhibit higher XRD peak intensities than that of the 1,4-dioxane counterpart. This difference can be assigned to the relatively larger film thickness as indicated in the cross-section SEM images in Figure S2 in the Supporting Information. The different boiling points of the utilized solvents likely account for the thickness difference. Compared with the low boiling point THF system, the slower evaporation rate of the 1,4-dioxane solvent during the spin coating slackened the vitrification process of the thin film. As a consequence, more solution is lost at the same centripetal force level during the spin-coating process, which in turn decreased the film thicknesses. The small grain sizes in the SnO 2 thin films will give rise to unique physical and chemical properties being of great potential in various applications. All thin films prepared with THF solvent exhibit randomly distributed mesoporous structures. No significant structure variation is demonstrated with progressive hydrochloric acid addition. The weak order of the mesoporous structures can be further revealed by the corresponding blurred ring-like FFT patterns of the SEM images. In comparison with the THF system, the SnO 2 thin films prepared with 1,4-dioxane exhibit more intriguing microstructures, such as cylinders and vesicles. Moreover, the FFT pattern of the thin film prepared with 50 µL hydrochloric acid demonstrates conspicuous hexagonal symmetry patterns (Figure 3e), which suggests the existence of hexagonally ordered mesoporous structures of the thin film. In addition, a porous structure similar to Figure 3a is obtained when the hydrochloric acid amount is increased to 100 µL ( Figure 3f). However, the relative blurred FFT patterns suggest the smearing of the hexagonally ordered mesopore arrangement. With further hydrochloric acid addition, a novel hybrid nanostructure consisting of spheres and cylinders is obtained (Figure 3g). The FFT pattern in Figure 3g reveals the random structure arrangement of the thin film. When the hydrochloric acid content utilized for sol-gel solution preparation is increased to 200 µL, a coexisting nanostructure of cylinders and vesicles is obtained (Figure 3h). The cross-section images shown in Figure S2 in the Supporting Information demonstrate that the thickness of all thin films is around 100 nm. To quantitatively extract contributions from different spatial frequencies, a power spectral density (PSD) function for

Film Surface Morphology
each porous SnO 2 thin film is calculated from the SEM data. As shown in Figure S3 in the Supporting Information, distinct PSD profile distributions are observed for the SnO 2 thin films prepared with different solvents and hydrochloric acid addition. Specifically, a clear broad single peak is observed in each PSD profile of the THF system. The peak positions at 0.1921, 0.1806, 0.1770, and 0.1736 nm −1 for the thin films prepared with 50, 100, 150, and 200 µL hydrochloric acid addition correspond to an average interdomain spacing of 33, 35, 36, and 36 nm, respec-tively. For the 1,4-dioxane system, the PSD profiles of the thin films prepared with 50 and 100 µL hydrochloric acid feature a sharp single peak, and the peak positions at 0.2380 and 0.2174 nm −1 correspond to an average interdomain spacing of 26 and 29 nm, respectively. In contrast, when the hydrochloric acid addition is further increased to 150 or 200 µL, multiple maxima in the PSD profile distribution are seen, which suggests the existence of polydisperse structures. The most pronounced peak is located at 0.1389 and 0.0799 nm −1 for the thin films prepared with 150 and 200 µL hydrochloric acid. It corresponds to a dominant interdomain distance of 45 and 79 nm, respectively. Compared to the PSD profiles of the thin films prepared with THF solvent, a more pronounced variation of the peak width and position is observed for the 1,4-dioxane counterpart, which suggests a significant change of the order and interdomain distances on the film surface.

Inner Film Morphology
With SEM, the local surface morphologies of the prepared SnO 2 thin films are clearly visualized. However, the porosity for the buried structures is also of great importance for practical applications. Accordingly, GISAXS measurements are performed to probe both the surface and the inner structures of the thin films within a large sample area. Figure S5 in the Supporting Information shows the GISAXS measurement geometry. The horizontally positioned thin film samples are illuminated by an incident X-ray beam with a shallow incident angle (0.4°). The transmitted and reflected beams from the thin film are denoted as direct beam (D) and specular beam (S), respectively. The maximum scattering intensity between specular beam and direct beam refers to the so-called Yoneda peak, which is given by the maximum in the Fresnel transmission function and depends on the material. [60] Figure 4a-d refers to the 2D GISAXS data of the SnO 2 thin films prepared with THF solvent and different hydrochloric acid addition content. The specular beam is blocked by a circular beam stop to prevent oversaturation of the detector. As indicated in Figure 4a-d, all of the scattering patterns show a distinct scattering signal in the Yoneda region, and the sectorial scattering patterns indicate the existence of randomly oriented nanostructures inside the entire films. [61] For a quantitative structure analysis in lateral direction, horizontal line cuts are performed for the 2D GISAXS data at the critical angle of SnO 2 . The specific position is marked with a red arrow in Figure 4a. The corresponding line cuts are plotted in Figure 4i. All line cuts exhibit multiple intensity peaks, as indicated with the black arrows. The prominent secondorder peak position of the thin film prepared with 50 µL hydrochloric acid addition is marked with a black dash line for reference. The gradual left shift of the second-order peak position from bottom to top indicates a progressive expansion of the detected feature size with hydrochloric acid addition. Moreover, the gradual smearing of the peaks in the line cuts with increasing hydrochloric acid addition indicates an enhanced disorder, which is in good agreement with the characteristics of the PSD profiles shown in Figure S3 in the Supporting Information.  -h refers to the 2D GISAXS data of the thin films prepared with 1,4-dioxane solvent. A distinct smearing of the Bragg peak is found with increasing hydrochloric acid addition, which can be more clearly revealed in the horizontal line cuts (Figure 4k). The prominent second-order peak position of the thin film prepared with 50 µL hydrochloric acid addition is marked with a black dash line for reference. Apart from the second-order peak, a distinct third-order peak as marked with the black arrow is observed in the horizontal line cut of the sample prepared with 50 µL hydrochloric acid addition, which suggests the existence of well-ordered structures in the thin film. With further hydrochloric acid addition, the scattering peaks in the higher q-region are gradually smeared out while an enhanced intensity in the lower q-region occurs simultaneously, which reflects the enhanced disorder of the small structures and increased contribution from large structures.
In order to extract more detailed information of the nanostructures, the horizontal line cuts in Figure 4i,k are modeled within the framework of the distorted-wave Born approximation (DWBA). A 1D paracrystalline lattice and a Gaussian size distribution are assumed for modeling the structure and form factors. [59] The feature sizes including radii and center-to-center distances of the nanostructures are extracted by modeling. Figure 4. 2D GISAXS data of the SnO 2 thin films prepared with a-d) THF solvent and e-h) 1,4-dioxane. The hydrochloric acid content utilized for the stock solution preparation is denoted on the lower right corner of each frame. Horizontal line cuts of the thin films prepared with i) THF solvent and k) 1,4-dioxane solvent, respectively. The measured data are plotted as solid triangles and the fit results are shown as black lines, respectively. All curves are shifted along the intensity axis for clarity of the presentation. Extracted structure radii (R) and center-to-center distances (D) given as a function of the hydrochloric acid content are the extracted feature sizes of the thin films prepared with j) THF solvent and l) 1,4-dioxane solvent, respectively. The small-sized, medium-sized, and large-sized structures are represented as dark yellow, red, and purple color, respectively.
In order to compare with the calculated pore sizes based on the GISAXS modeling, the pore size distribution profiles are extracted through a binarization algorithm with the Image J software ( Figure S4, Supporting Information) for all SEM images of Figure 3. The peak center of the profiles is determined with a Gaussian function. For the THF system, a monotonic shift of the peak center from 21 ± 3 to 25 ± 3 nm is observed, which is generally consistent with the GISAXS results (from 20 ± 4 to 30 ± 4 nm). For the 1,4-dioxane system, no distinct peak is detected for the thin film prepared with 200 µL hydrochloric acid addition, which can be attributed to the limited data points collected from the SEM images. However, it is noteworthy that more large pore sizes are detected for the thin films prepared with 150 and 200 µL hydrochloric acid, which arise from the cylindrical and vesicle structures shown in Figure 3g,h. When the hydrochloric acid addition is increased from 50 to 150 µL, the determined peak center shifts from 20 ± 3 to 36 ± 4 nm, which is in a good agreement with the GISAXS results (from 20 ± 3 to 33 ± 3 nm) as well. The good consistency of the pore sizes acquired from the GISAXS modeling and the binarization algorithm is indicative of having similar surface and inner film morphologies as probed with SEM and GISAXS, respectively.

Mechanism of Morphology Change
Considering the nonionizing properties of the PS and PEO blocks and the low dielectric constant of the solvent (THF: 7.58, 1,4-dioxane: 2.25) utilized in the present work, the repulsion force among the polymer chains provoked by ionization is negligible. [63] Accordingly, the morphology variation of the SnO 2 thin films is analyzed using the polymer-solvent interaction parameter χ [64] where V S and δ S are the molar volume and solubility parameter of the solvent, δ P is the solubility parameter of the polymer, R is the gas constant, and T is the temperature. As described in the Experimental Section, the SnCl 4 content added in all solutions is 33.8 µL, and the corresponding hydrochloric acid content is linearly increased from 50 to 200 µL. Thus, 20.36 mg water is required for the hydrolysis of 33.8 µL SnCl 4 , whereas 50 µL of 37% hydrochloric acid solution contains 37.8 mg water already. [65] Consequently, the water content in all prepared  ( ) ( ) As a poor solvent for the hydrophobic PS block of the diblock copolymer PS-b-PEO, the water content in the hydrochloric acid solution significantly affects the phase separation behavior of the sol-gel solution. Furthermore, superior miscibility of water in THF or 1,4-dioxane solvent can be achieved through hydrogen-bond interaction. [66][67][68][69][70] Accordingly, the water content in the hydrochloric acid solution is treated as solvent except for the part consumed by the complete hydrolysis of SnCl 4 .
The characteristics of different solvents and polymers can be found in the polymer handbook (Table S2, Supporting Information). [71] As shown in Equation (7), the solubility parameter of the solvent (Hansen solubility parameter δ H ) consists of three parts: dispersive δ d , permanent dipole-dipole interaction δ p , and hydrogen bonding forces δ h  Table S3 in the Supporting Information.
Based on the existing literature and calculated parameters (in Tables S2 and S3, Supporting Information), the polymersolvent interaction parameters (χ P-S ) of different polymersolvent pairs are determined (Table S4, Supporting Information).
The preferential affinity of a solvent for a certain block is represented by an expression of Δχ = χ PS − S − χ PEO − S . Table 4 lists the calculated Δχ values from the different polymer-solvent interaction parameters (χ P-S ). According to the Flory-Huggins theory, the polymer can be completely dissolved in the solvent over the entire composition range when the Δχ value is lower than 0.5. [72] As shown in Table 4, the calculated Δχ values for all solvent mixtures utilized in the present work are less than 0.5, which suggests that both PS blocks and PEO blocks of the PS-b-PEO template can be well dissolved in the solvent. Since the water comprised in the hydrochloric acid is a significant factor affecting the preferential affinity of the solvent to the different polymer blocks, the morphology variation of the SnO 2 thin films is discussed in detail with respect to the water content in the sol-gel solution. Figure 5 shows the different film morphologies determined by the water content and Δχ values. It is observed that the Δχ values for both, THF and 1,4-dioxane mixture, monotonously increase with the increase of the water content. A spherical pore structure is preferentially formed for lower Δχ values (0.04 ≤ Δχ ≤ 0.18). When the Δχ value reaches 0.21, a hybrid morphology consisting of spherical and cylindrical porous structures is observed. Further increasing the Δχ value to 0.25 leads to the appearance of vesicle structures. The structure transition from sphere to vesicle in terms of the Δχ values can be explained by the different stretching of the PS blocks in the sol-gel solution. The introduced water increases the surface energy between the PS chains and the surrounding solvent, which significantly inhibits the PS chain stretching in the solution. The PS chains in spherical micelles are typically more stretched than in cylinders or vesicles. Accordingly, they are usually the first aggregates to form and can be considered as the starting morphology for other aggregates. [62][63][64][65] It should be noted that the calculated Δχ value for the SnO 2 thin film prepared with 50 µL hydrochloric acid (corresponding to 17 µL  water) and 1,4-dioxane solvent is located in the Δχ value range obtained with THF system (0.04 ≤ Δχ ≤ 0.15). However, compared to the random porous structures shown in Figure 3a-d, a significantly enhanced order of the pore arrangement is observed in Figure 3e. This difference in order can be tentatively explained by a different hydrogen bond interaction type between the employed cycloether (THF or 1,4-dioxane) and H 2 O. For a cycloether/H 2 O binary mixture, the interactions between cycloether molecules can be mediated by H 2 O molecules through bifunctional donor and acceptor hydrogen bond interactions. [66,72] In contrast to the 1,4-dioxane/H 2 O mixture, the smaller size of the THF molecule has an unfavorable effect on the hydrogen bond network formed between H 2 O molecules. This leads to a significantly enhanced micro-heterogeneity in the THF/H 2 O mixture system. As a result, THF/ H 2 O clusters enriched in THF are formed. [73] The formation of THF/H 2 O clusters circumvents the unfavorable interactions between water and the hydrophobic PS chains of the block copolymer, which weakens the phase separation tendency of the block copolymer. Accordingly, compared to the hexagonally packed pore arrangement obtained with 1,4-dioxane (Figure 3e), a reduction in number and order of the pore structures is observed in the THF system (Figure 3a-d).

Conclusion
In this work, the amphiphilic block copolymer-templated SnO 2 thin film synthesis is systematically investigated in terms of the solvent category and hydrochloric acid content. Two kinds of cycloether solvents, THF and 1,4-dioxane, are utilized as good solvent for dissolution of the block copolymer PS-b-PEO. Hydrochloric acid is added as a poor solvent for stimulating the phase separation of the block polymer and to be the catalyst for the hydrolysis of SnCl 4 . SEM and GISAXS measurements are performed to detect the mesoscale structure of the thin films at different length scales. A detailed comparison and analysis demonstrates that the structural transformation of the SnO 2 thin films in terms of different solvent category and hydrochloric acid addition is governed by a synergistic effect of the following two factors: First, the preferential affinity of the solvent for a certain block; second, the hydrogen bond interaction between the employed cycloether and H 2 O. By tuning the preferential affinity of the mixed solvent with different hydrochloric acid content, various structures including spheres, cylinders, and vesicles are obtained. Specifically, the spherical pore structure is preferentially formed at relative low
Sol-Gel Stock Solution Preparation: For investigating the effect of the solvent category and poor solvent addition on the thin film morphology, two different solvents (THF and 1,4-dioxane) and four different hydrochloric acid volumes (50, 100, 150, and 200 µL) were utilized for the sol-gel stock solution preparation. In order to prevent the formation of big clusters due to the vigorous hydrolysis reaction of the SnCl 4 precursor, 2.0 mL THF or 1,4-dioxane solvent were divided into smaller portions of 1.5 and 0.5 mL for the stock solution preparation. 1.5 mL THF or 1,4-dioxane was utilized for dispersing PS-b-PEO and SnCl 4 precursor, the rest 0.5 mL THF or 1,4-dioxane was assigned for diluting different volumes of hydrochloric acid (50, 100, 150, and 200 µL). First, 21.0 mg PS-b-PEO was added into 1.5 mL THF or 1,4-dioxane solvent followed by 1 h continuous stirring to fully dissolve it. Then, 33.8 µL SnCl 4 and 0.5 mL diluted hydrochloric acid solution were sequentially pipetted into the well-dissolved PS-b-PEO solution. Because of the complexation effect between THF or 1,4-dioxane solvent and SnCl 4 molecules, white crystalline solids were formed once adding SnCl 4 to the solution (Figure 1a and Figure S1, Supporting information). However, the complex crystals were moisture sensitive and rapidly vanished with the addition of diluted hydrochloric acid solution. [49,56] The corresponding component phase diagram of the samples prepared with THF or 1,4-dioxane solvent is shown in Figure 6. The as-prepared sol-gel solution was further stirred for 1 h to stabilize the hydrogen bonds formed between the hydrophilic PEO blocks and the hydrolyzed (HO) X SnCl 4−X species.
Thin Film Preparation: For the thin film deposition, the as-prepared sol-gel solutions were spin coated on pre-cleaned silicon substrates with a Delta 6 RC TT spin coater (SÜSS Micro Tec Lithography GmbH) at 2000 rpm for 60 s. For removing the polymer template, the as-spun thin films were calcined at 500 °C for 2 h with a heating rate of 1 °C min −1 in ambient atmosphere.
Thin Film Characterization: The surface morphology of the prepared SnO 2 thin films was probed with a high-resolution field-emission SEM (Zeiss Gemini NVision 40) at a working distance of 3.5 mm and an acceleration voltage of 5 kV. For the cross-section SEM measurement, the thin films were tilted at 45° respective to the electron beam for a better view. The statistical average information of the porous structures within the bulk films was detected with GISAXS at the P03 beamline of PETRA III storage ring. [57] The measurements were performed with a wavelength of 0.96 Å and at an incidence angle of 0.4°. A sample to detector distance of 4200 mm was utilized to obtain an appropriate q range. The scattering signal was recorded by a Pilatus 300K (Dectris Ltd.) detector with a pixel size of 172 µm. A python program named Directly Programmable Data Analysis Kit (DPDAK) was used for calibration and data analysis. [58] Horizontal line cuts of the 2D GISAXS data were performed at the Yoneda peak position of the SnO 2 . The crystallinity of the prepared SnO 2 thin films was investigated with a D8 ADVANCE X-ray diffractometer using an X-ray wavelength of 1.54 Å.

Supporting Information
Supporting Information is available from the Wiley Online Library or from the author.