Direct Observation of the Outermost Surfaces of Mesoporous Silica

Feb 9, 2017 - AFM images of nonetched 2-dimensional hexagonal mesoporous silica thin films show that the shape of the silica layer on the surface of t...
0 downloads 11 Views 3MB Size
Article pubs.acs.org/Langmuir

Direct Observation of the Outermost Surfaces of Mesoporous Silica Thin Films by High Resolution Ultralow Voltage Scanning Electron Microscopy Maho Kobayashi,† Kyoka Susuki,† Haruo Otsuji,‡ Yusuke Sakuda,§ Shunsuke Asahina,§ Naoki Kikuchi,§ Toshiyuki Kanazawa,§ Yoshiyuki Kuroda,∥ Hiroaki Wada,† Atsushi Shimojima,† and Kazuyuki Kuroda*,†,‡ †

Department of Applied Chemistry, Faculty of Science and Engineering, Waseda University, 3-4-1 Ohkubo, Shinjuku-ku, Tokyo 169-8555, Japan ‡ Kagami Memorial Research Institute for Materials Science and Technology, Waseda University, 2-8-26 Nishiwaseda, Shinjuku-ku, Tokyo 169-0051, Japan § JEOL Ltd. SM Business Unit, 3-1-2 Musashino, Akishima-shi, Tokyo 196-8558, Japan ∥ Waseda Institute for Advanced Study, 1-6-1 Nishiwaseda, Shinjuku-ku, Tokyo 169-8050, Japan S Supporting Information *

ABSTRACT: The properties of the outermost surfaces of mesoporous silica thin films are critical in determining their functions. Obtaining information on the presence or absence of silica layers on the film surfaces and on the degree of mesopore opening is essential for applications of surface mesopores. In this study, the outermost surfaces of mesoporous silica thin films with 3-dimensional orthorhombic and 2-dimensional hexagonal structures were observed using ultralow voltage high resolution scanning electron microscopy (HR-SEM) with decelerating optics. SEM images of the surfaces before and after etching with NH4F were taken at various landing voltages. Comparing the images taken under different conditions indicated that the outermost surfaces of the nonetched mesoporous silica thin films are coated with a thin layer of silica. The images taken at an ultralow landing voltage (i.e., 80 V) showed that the presence or absence of surface silica layers depends on whether the film was etched with an aqueous solution of NH4F. The mesostructures of both the etched and nonetched films were visible in images taken at a conventional landing voltage (2 kV); hence, the ultralow landing voltage was more suitable for analyzing the outermost surfaces. The SEM observations provided detailed information about the surfaces of mesoporous silica thin films, such as the degree of pore opening and their homogeneities. AFM images of nonetched 2-dimensional hexagonal mesoporous silica thin films show that the shape of the silica layer on the surface of the films reflects the curvature of the top surface of the cylindrical mesochannels. SEM images taken at various landing voltages are discussed, with respect to the electron penetration range at each voltage. This study increases our understanding of the surfaces of mesoporous silica thin films, which may lead to potential applications utilizing the periodically arranged mesopores on these surfaces.



INTRODUCTION Mesoporous silica thin films have attracted increasing interest because of their transparency, regular structure, and morphology.1,2 They are usually prepared via the formation of mesostructured silica−surfactant composites.3 Various characteristics of mesoporous silica thin films, such as mesostructure, porosity, and film thickness, can be controlled by altering various factors of the synthesis,3−8 including the kind of surfactants, composition of the precursor solutions, and film formation conditions. Most of the previous studies on mesoporous silica thin films have focused on the entire mesopores of the films. In contrast, we herein focus on the outermost surfaces of mesoporous silica thin films because the accessibility of mesopores to incoming species is important in © 2017 American Chemical Society

some applications, and the unique structural features of their 2dimensional (2D) mesostructured surfaces are useful. Several previous studies on mesoporous silica thin films prepared by evaporation-induced self-assembly (EISA) have schematically noted the presence of thin, nonporous, silica layers on the surfaces of the films on the basis of generally accepted formation mechanisms of mesostructured silica.2 The presence of these surface silica layers has been suggested by cross-sectional transmission electron microscopy (TEM) observations9−12 and X-ray reflectivity (XRR) measureReceived: December 16, 2016 Revised: February 8, 2017 Published: February 9, 2017 2148

DOI: 10.1021/acs.langmuir.6b04511 Langmuir 2017, 33, 2148−2156

Article

Langmuir ments13,14 of mesoporous silica thin films. Because these layers are generally recognized to be nonporous, it is necessary to remove them using an etching method in order to open up the surface mesopores, thereby allowing their use in applications such as nanopatterning14 and nanoreplication.15 Dourdain et al. fabricated a dot pattern on a silica thin film for use in highly integrated data storage devices.14 They positioned magnetic nanoparticles in open mesopores in the surface of a 3-dimensional (3D) mesoporous silica thin film after the removal of the surface silica layer by dry etching with Ar ions. We previously reported that the outermost mesostructured surfaces of 2D mesoporous silica thin films can be used as a mold for the formation of 2D striped nanopatterns of copper thin films after removal of the surface layers by wet etching.15 However, the optimal conditions for the etching processes, including the method (wet or dry etching), etching time, and etchant, could not be determined. This was because of a lack of detailed information about the outermost surfaces, such as the degree of pore opening and homogeneity. Therefore, obtaining detailed information on the surfaces of mesoporous silica thin films, such as the nature of the open or closed mesopores, degree of pore opening, their homogeneities, and macroscopic flatness of the film surfaces, is essential for optimizing the use of such surfaces. There are currently four reports13,16−18 on the direct surface observations of the structures of the outermost surfaces of mesoporous silica thin films prepared by EISA, in addition to two studies referenced above. In these studies, atomic force microscopy (AFM),16 SEM,17 AFM and SEM,18 or AFM combined with X-ray reflectivity (XRR) measurements13 were used for determining the nature of the outermost surfaces. In these studies, SEM images were taken at relative high voltage (a few kV) with17 and without18 metal sputtering on the surfaces of the specimen. However, AFM results are not so informative in determining whether the periodically structured surfaces can be ascribed to open mesopores, for the following reasons. First, the pore diameter and in-plane thickness of silica pore walls were similar, though the sensitivity of AFM is very high for the detection of height differences and in-plane resolution. Second, the probe tip could not reach the bottom of open mesopores while maintaining suitable contact for atomic force measurements. Furthermore, care must be taken when interpreting SEM images because those taken at relatively high voltages (a few kV, for example) give information on both the outermost surfaces and deeper regions of a specimen, particularly if the specimen is composed of a substance with low mean atomic number, such as silica. Moreover, XRR and cross-sectional TEM are not suitable for direct observations of the outermost surfaces. The XRR method is useful for characterizing the surface silica layers and their average thicknesses but cannot determine the homogeneity of their thickness or pore openings, nor the degree of pore opening on the film surfaces. Crosssectional TEM observations are useful in establishing the presence or absence of surface silica layers and their average thicknesses. However, projected images contain information on the whole thickness (about 100 nm) of a slice of a mesoporous silica film, implying that superimposed images are obtained. In this study, we focused on using ultralow voltage high resolution (HR)-SEM to directly observe the outermost surfaces of mesoporous silica thin films. So-called “low voltage” is conventionally accepted as the voltage lower than a few kV, and here we define “ultralow voltage” as lower than a few hundred volts. The ultralow voltage HR-SEM has been

developed recently to reduce interference in images due to charge-up19 and observe the outermost surfaces of films.20 In order to achieve high resolution in ultralow voltage observations, decelerating optics have been developed (Figure 1).20 The primary electrons are decelerated using two

Figure 1. Decelerating optics for ultralow voltage HR-SEM observation. (a) Schematic of the HR-SEM optics under the objective lens focusing on the decelerating optics and equation of the relationship of accelerating voltage, landing voltage, and specimen bias. (b) Visualized electron penetration depth into a dense silica film at various landing voltages.

techniques (Figure 1a). The first technique uses a super hybrid lens (SHL), consisting of both magnetic and electrostatic lenses, as the objective lens. The SHL is capable of producing a small probe size, even at low impact energies, and of imaging down to 10 V using a beam deceleration mode. The second technique is the application of a negatively charged bias (VSBias) to the specimen. When using decelerating optics, it is more appropriate to use the landing energy (eV) instead of landing voltage (V) because the electrons passing through the objective lens lose energy due to the negative specimen bias. However, in this paper, we use the landing voltage (V) instead of landing energy (eV), as per convention. The accelerating voltage underneath the electron gun (VGun) is retarded by a negatively charged bias applied to specimen (VSBias), which lowers the landing voltage (VLand). The relationship between these voltages is VLand = VGun − VSBias. The landing voltage can be controlled by altering the accelerating voltage underneath the electron gun (VGun) and/or the specimen bias (VSBias). Decelerating optics give high resolution images due to the high accelerating voltage underneath the electron gun. Ultralow landing voltages (VLand) which prevent electron beams from deeply penetrating the specimen and reduce charge-up; therefore, the outermost surfaces can be proved with high S/ N ratios. One of the authors (S.A.) has used this technique to show the presence of defects and micropores that connect mesochannels of mesoporous silica powders (SBA-15).21 This finding contributes to a deeper understanding of the formation mechanism of hydrothermally prepared mesoporous silica powders. In this study, we directly observed the presence or absence of thin silica layers on the outermost surfaces of nonetched mesoporous silica thin films using ultralow voltage HR-SEM with decelerating optics (Figure 1). Typical mesoporous silica thin films were prepared by the EISA method,22,23 and they 2149

DOI: 10.1021/acs.langmuir.6b04511 Langmuir 2017, 33, 2148−2156

Article

Langmuir

related structural model (Figure S10), θ−2θ XRD pattern (Figure S11), SEM images (Figure 3c and Figure S12c), and TEM images (Figure S13). All the data indicate the formation of a 2D hexagonal (distorted p6mm) mesoporous silica thin film with cell parameters a = 9.4 nm and b = 13.4 nm, with the b-axis oriented parallel to the substrate. Wet Etching of Mesoporous Silica Thin Films (3MPSF_et and 2MPSF_et). The mesoporous silica thin films (3MPSF_as and 2MPSF_as) were etched by immersing the films in an aqueous ammonium fluoride solution (0.1 M) for various immersion times (0.5−3 h). After etching, the films were washed with deionized water for 5 min and then rinsed again with deionized water. Then, they were dried in a plastic box (approximately 40% RH) to form wet-etched mesoporous silica thin films. The specimens are abbreviated as 3MPSF_et and 2MPSF_et for etched films. In addition, 3MPSF_etXh and 2MPSF_etXh (X = etching time: 0.5, 2, or 3 h) are used in order to identify films etched for a specified time. The characterization results for both films confirmed that the original mesostructures and their in-plane periodicities were retained in all the etched samples including those treated for 3 h. The film thicknesses were reduced with increasing etching times. The characterization details of 3MPSF_et are given on pages 6 and 7 of the Supporting Information. The supporting data include the optical microscopy image (Figures S1b), 2D-XRD pattern (Figure S2b), θ−2θ XRD patterns (Figure S4), and SEM images (Figure 2f and Figure S7f). The characterization details of 2MPSF_et are given on pages 7 and 8 of the Supporting Information. These results include the same types of instrumental data as above (optical microscopy image (Figure S8b), 2D-XRD pattern (Figure S9b), θ−2θ XRD patterns (Figure S11), and SEM images (Figure 3f and Figure S12f)). Characterization. The surfaces of the mesoporous silica thin films before and after etching were observed using ultralow voltage HRSEM. HR-SEM experiments were performed on a JEOL JSM-7800F Prime electron microscope with beam deceleration optics. These optics allow the imaging of specimens using landing voltages as low as 300 V.21 In the present study, the specimen bias voltage (VSBias) was set at −5 kV, and the acceleration voltages underneath gun (VGun) were set at 7.00, 5.20, and 5.08 kV. The resulting landing voltages (VLand) at the specimen were 2 kV, 200 V, and 80 V, respectively. In this study, 80 V was chosen as the lowest voltage because the excitation efficiency of secondary electrons is sufficient, thus enhancing the contrast relative to studies at lower voltages such as 50 V. The working distance between the specimen and the objective lens was either 2.0 or 2.5 mm. The probe current was 8 pA. Mesoporous silica thin films before (3MPSF_as and 2MPSF_as) and after etching (3MPSF_et and 2MPSF_et) were also characterized using two types of X-ray diffraction (XRD) measurements. First, θ−2θ XRD scanning profiles were recorded using the Bragg−Brentano geometry. This was performed with a parallel beam diffractometer (Rigaku Ultima IV) using monochromated Fe Kα radiation (40 kV, 30 mA) with SS 0.1 mm, DS 0.1 mm, and RS 0.1 mm. Second, 2D-XRD was performed on a Rigaku Nanoviewer using grazing incidence smallangle X-ray scattering (0.2° incident angle) with Cu Kα radiation (40 kV, 30 mA). A Pilatus 2D X-ray detector (Dectris) was set 730 mm from the specimen; the exposure time was 5 min. Simulations of the 2D diffraction patterns of 3MPSF and 2MPSF were performed using Crystal Maker (ver. 9.2.8, Crystal Maker Software, Ltd.) and SingleCrystal (ver. 2.3.3, Crystal Maker Software, Ltd.). A focused ion beam (FIB) was used to prepare slices of the specimens for the cross-sectional TEM observations, with a JEOL JIB-4000 instrument using 10 kV Ga+ ions. Carbon was deposited on the surfaces of the films to protect them from damage due to the FIB. Cross-sectional TEM observations were performed on a JEOL JEM-2010 electron microscope with an acceleration voltage of 200 kV. AFM observations were carried out on a Nanoscope III instrument (Digital Instruments Inc.), using the tapping mode for the surfaces of 3MPSF_as, 3MPSF_et3h, 2MPSF_as, and 2MPSF_et3h. Si probes (NCHV-10 V) were purchased from Bruker Nano Inc. Optical microscopic images were obtained using an Olympus BX51-58MU microscope. To estimate the thickness of thin silica layers on the surfaces of as-made

were wet etched in order to define differences in the SEM images of the regions of the outermost surface and those underneath the surface. We compared the SEM images taken at a conventional low landing voltage of a few kV with those taken at ultralow voltages of 200 and 80 V.



EXPERIMENTAL SECTION

Materials. Tetraethoxysilane (TEOS) with >99.0% purity was purchased from Kishida Chemical Co., Ltd. Pluronic F127 (poly(ethylene oxide)106-block-poly(propylene oxide)70-block-poly(ethylene oxide)106; 12 600 g/mol) and Pluronic P123 (poly(ethylene oxide)20block-poly(propylene oxide)70-block-poly(ethylene oxide)20; 5800 g/ mol) were purchased from Sigma-Aldrich Co. Aqueous hydrochloric acid solutions (6 and 0.01 N) were purchased from Wako Pure Chem. Ind., Ltd. Ethanol (EtOH) with >99.5% purity was purchased from Junsei Chemical Co., Ltd. Ammonium fluoride (NH4F) with >97.0% purity was purchased from Kanto Chemical Co., Ltd. All chemicals were used without further purification. Si wafers (P-doped n-type (100) wafers; 4 × 4 in.) were purchased from Silicon Technology Co., Ltd. Si substrates were cut into 2 × 2 cm panels and washed first with a semiconductor cleaning solution (Semico Clean 23, Furuuchi Chemical Co., Ltd.) and then with deionized water. Preparation of 3-Dimensional Mesoporous Silica Thin Films (3MPSF_as). 3D mesoporous silica thin films composed of cage-type mesopores were prepared on a Si (100) substrate by the EISA method, as per a previous report.22 Hereafter, the films are abbreviated as 3MPSF_as. Tetraethoxysilane, ethanol, deionized water, and hydrochloric acid (6 N) were mixed together, and the mixture was stirred for 30 min at 60 °C. Then, an EtOH solution of the surfactant (Pluronic F127) was added to the mixture, and the solution was stirred for 30 min at room temperature. The final molar ratio of the precursor solution was TEOS:F127:H2O:HCl:EtOH = 1:0.004:4.9:0.15:20. The precursor solution was spin-coated on a Si substrate (2 × 2 cm) at 2000 rpm for 30 s at 25 °C and 65% RH to form a mesostructured silica thin film. The film was dried in a plastic box (approximately 40% RH) at room temperature for 1 day, then thermally treated in air at a heating rate of 1 °C/min to 500 °C, and kept at this temperature for 4 h to prepare an as-made 3D mesoporous silica thin film (3MPSF_as). The dimensions of each film sample were 2 cm × 2 cm × 250 nm. The characterization results of 3MPSF_as are given on pages 3 and 4 of the Supporting Information. These results include the optical microscopy image (Figure S1a in the Supporting Information), 2DXRD pattern (Figure S2a) and the related structural model (Figure S3), θ−2θ XRD pattern (Figure S4), TEM images (Figure S5) and the related structural model (Figure S6), and SEM images (Figure 2c and Figure S7c). All the data indicate the formation of 3D orthorhombic (Fmmm) mesoporous silica thin films with cell parameters a = 17.6 nm, b = 12.4 nm, and c = 25.0 nm, with the b-axis oriented perpendicular to the substrate. Preparation of 2-Dimensional Mesoporous Silica Thin Films (2MPSF_as). 2D mesoporous silica thin films were prepared on a Si (100) substrate by the EISA method as per a previous report.23 Hereafter, the films are abbreviated as 2MPSF_as. A homogeneous solution of tetraethoxysilane, ethanol, deionized water, and hydrochloric acid (0.01 N) was prepared. Then, an EtOH solution of the surfactant (Pluronic P123) was added to the solution, and the mixture was stirred for 3.5 h at room temperature (ca. 25 °C). The final molar ratio of the precursor solution was TEOS:P123:H2O:HCl:EtOH = 1:0.0093:3.8:1.06:8.44. The precursor solution was spin-coated on a Si substrate (2 × 2 cm) at 25 °C and 40% RH, at 500 rpm for 2 s, and then at 3000 rpm for 30 s, to form a mesostructured silica thin film. The film was dried in a plastic box (approximately 40% RH) at room temperature for 1 day, then thermally treated in air at a heating rate of 1 °C/min up to 400 °C, and then kept at this temperature for 4 h to prepare an as-made 2D mesoporous silica thin film (2MPSF_as). The dimensions of each film sample were 2 cm × 2 cm × 400 nm. The characterization results of 2MPSF_as are given on pages 4−6 in the Supporting Information. These results include the optical microscopy image (Figure S8a), 2D-XRD pattern (Figure S9a) and the 2150

DOI: 10.1021/acs.langmuir.6b04511 Langmuir 2017, 33, 2148−2156

Article

Langmuir films, cross-sectional field emission (FE)-TEM observations were performed on a JEOL JEM-2100F electron microscope at an acceleration voltage of 200 kV. Platinum was deposited on the surfaces of the films prior to carbon deposition to mark the edge of the outermost surfaces.

mesoporosity is simply expressed as “surface silica layer” hereafter. The images of 3MPSF_et3h taken at 80 and 200 V show an ordered arrangement of mesopores (Figures 2d,e and Figures S7d,e). This indicates that the surface silica layers were partially removed by etching. These results strongly suggest that the surfaces of the as-made 3D mesoporous silica thin films are covered with thin layers of silica. When the landing voltage was relatively high (2 kV), the mesostructure was observed in both specimens before and after etching for 3 h (Figures 2c,f and Figures S7c,f), although there are slight variations in the contrast of the mesostructure between these images. Thus, the SEM images taken at 2 kV were not suitable for determining the nature of the outermost surfaces of the films because there were no significant differences between the images of the etched and nonetched samples. Furthermore, the ultralow landing voltage (80 V) was adequate for observing the outermost surfaces of the films. The SEM images of the surfaces of the mesoporous silica thin films before and after etching taken at 80 V (3MPSF_as and 3MPSF_et, respectively) for different etching times are shown in Figure S14. The number of dark spots, and therefore mesopores, on the surfaces of the 3MPSF samples increased with increasing etching times (Figure S14). This demonstrates that the surface silica layers were gradually removed by etching, showing the coverage of thin silica layers on surface mesopores. Two-Dimensional Mesoporous Silica Thin Films (2MPSF_as and 2MPSF_et3h). SEM images of 2MPSF_as taken at 80 V, 200 V, and 2 kV are shown in Figures 3a−c (see



RESULTS AND DISCUSSION Influence of Landing Voltage on the Observation of Mesoporous Silica Thin Films. Three-Dimensional Mesoporous Silica Thin Films (3MPSF_as and 3MPSF_et3h). SEM images of 3MPSF_as taken at landing voltages of 80 V, 200 V, and 2 kV are shown in Figures 2a−c (images with lower

Figure 2. Surface SEM images of 3MPSF_as and 3MPSF_et3h taken at varied voltages.

magnification of a larger area are shown in Figures S7a−c). SEM images of 3MPSF_et3h are shown in Figures 2d−f (see Figures S7d−f for lower magnification). The surface textures of 3MPSF_as appear significantly different at different landing voltages. However, those of 3MPSF_et3h were similar at different landing voltages; dark spots assignable to mesopores were observed in each, though there were differences in the contrast and the number of dark spots (the details of the assignment to mesopores are shown in the Supporting Information (page 19)). The images of the surfaces of the films before and after etching taken at the same landing voltage were compared as follows. Significant differences in the surface textures before and after etching for 3 h were observed in the images taken at landing voltages of 80 and 200 V. At 200 V, the highmagnification image of 3MPSF_as (Figure 2b) showed nonperiodic dark spots in some regions. The spots on the FFT pattern were broader and arc shaped, which does not correspond to the 3D mesostructure (inset of Figure S7b). When the ultralow landing voltage (80 V) was used, the texture appeared smooth without any dark spots (Figure 2a and Figure S7a). This suggests the presence of a nonporous silica thin layer on the surfaces of 3MPSF_as. Please note that “nonporous” described here means nonmesoporous, and the possibility of the presence of micropores cannot be excluded because triblock copolymers were used as templates. Surface silica layer without

Figure 3. Surface SEM images of 2MPSF_as and 2MPSF_et3h taken at various landing voltages.

Figures S12a−c for lower magnification images of a larger area). Images of 2MPSF_et3h are shown in Figures 3d−f (see Figures S12d−f for lower magnification images). These SEM images show a curved stripe pattern composed of dark and bright curved lines assignable to cylindrical mesopores (i.e., mesochannels) and silica walls, respectively (the details of the assignment to mesochannels are shown in the Supporting Information (page 19)). The surface textures of 2MPSF_as 2151

DOI: 10.1021/acs.langmuir.6b04511 Langmuir 2017, 33, 2148−2156

Article

Langmuir

suggest that the mesopores are ellipsoid in shape. A model of a mesopore is shown in Figure 4a. The diameters of the dark

appear clearly different at different landing voltages. The SEM images of 2MPSF_et3h taken at 80 V (Figure 3d and Figure S12d) and 200 V (Figure 3e and Figure S12e) show that the number of dark curved lines decreased and that the lines appeared narrower at lower landing voltages (200 and 80 V). We compared the images of the surfaces taken at the same landing voltage, before and after etching. As with the 3MPSF samples, there were no significant differences between the SEM images taken at 2 kV before and after etching (Figure 3c,f and Figures S12c,f). Conversely, significant differences in the surfaces before and after etching for 3 h were observed in the images taken at lower landing voltages (80 V (Figure 3a,d and Figures S12a,d) and 200 V (Figure 3b,e and Figures S12b,e)). Comparison between these low voltage images shows that the mesostructure on the surface of 2MPSF_as was observed with lower contrast than that of 2MPSF_et3h. This strongly suggests that the surfaces of as-made 2D mesoporous silica thin films were covered with thin silica layers. As stated in the previous section, a landing voltage of 80 V was adequate to observe the outermost surfaces of 2D mesoporous silica thin films. The SEM images of the surfaces of mesoporous silica thin films before and after etching (2MPSF_as and 2MPSF_et) taken at 80 V for different etching times are shown in Figure S15. As the etching time increased, the number of dark curved lines, corresponding to mesochannels, also increased (Figure S15). This indicates that etching gradually removed the surface silica layers. These results show that as-made 2D mesoporous silica thin films have thin silica layers covering the surface mesochannels. Internal pores24 were clearly observed in the images of 2MPSF_et3h taken at 2 kV (Figure 3f). Interestingly, these images are very similar to those of SBA15 powder taken at 300 V reported previously.21 In that report,21 powdered SBA15 had no outer silica layers and was cleaned with an Ar ion beam to remove the carbon deposited on the specimen during observation. The absence of the outer layers makes the observation at ultralow landing voltage adequate to observe internal pores and defects. Conversely, a higher landing voltage (2 kV) was needed to observe internal pores and defects in the films because the silica layer was not completely removed after etching for 3 h. Hence, high landing voltages are valuable for observing mesoporous silica thin films. Degree of Mesopore or Mesochannel Opening along the Vertical Direction. The etching of mesoporous silica films not only removes the outermost silica layers but also affects the degree of mesopores/mesochannels opening along the vertical direction by removing the exposed silica pore walls. The SEM images of the films etched for 3 h, taken at 80 V, show the partially remaining surface silica layers, in additions to dark spots/dark curved lines of various sizes corresponding to mesopores/mesochannels. The sizes of these dark spots/dark curved lines should correspond to the degrees of mesopores/ mesochannels opening. When the degree is 50%, the mesopores/mesochannels should have half-cup/half-pipe shapes, respectively. Instrumental and measurement errors during the calculation of the openings should be taken into account.25 (1) 3MPSF_et3h. The diameter of the mesopores in the inplane direction in 3MPSF_as was ca. 8 nm, as determined from the corresponding SEM image taken at 2 kV (Figure 2c). The TEM image of the same specimen (Figure S5d) indicates that the diameter of the mesopores in the direction of d010 (perpendicular to the film) was 3.7 nm. These measurements

Figure 4. Cross-sectional models of mesopores and various degrees of mesopore opening along the vertical direction: (a) mesopore of 3MPSF_as, (b) mesopore of 3MPSF_et3h indicated by the yellow open square in Figure 2d, and (c) mesopore of 3MPSF_et3h indicated by the red open square in Figure 2d. (The value of in-plane mesopore diameter (8 nm) was determined from the SEM image (Figure 2c). The value of the vertical mesopore size (3.7 nm) was determined from the TEM image (Figure S5d), and the size corresponds to the thickness of the first mesoporous layer. The value of surface silica layer (3.3 nm) was determined from the TEM image (Figure 7a).)

spots in the image taken at 80 V were determined to be ca. 7 nm (in the yellow open square) and ca. 8 nm (in the red open square) (Figure 2d). Therefore, the degrees of pore opening were calculated to be ca. 25% and ca. 50%, respectively, as shown in Figures 4b and 4c, respectively. (2) 2MPSF_et3h. The diameter of the mesochannels in the in-plane direction, determined from the SEM image of 2MPSF_as taken at 2 kV (Figure 3c), was ca. 7 nm. The diameter of the mesochannels in the direction of d 10 (perpendicular to the film) was 4.2 nm, as determined from the TEM image of the same specimen (Figure S13b). Using these dimensions, the mesochannels were determined to be elliptic cylinders. A model of a mesochannel is shown in Figure S16a. The diameters of the dark curved lines in the image taken at 80 V were determined to be ca. 4 nm (in the yellow open square) and ca. 7 nm (in the red open square) (Figure 3d). Therefore, the degrees of channel opening were calculated to be ca. 7% and ca. 50%, respectively, as shown in Figures S16b and S16c, respectively. Outermost Surfaces of 3MPSF_as and 2MPSF_as. AFM is known to be more sensitive than SEM in determining the height of a specimen. AFM images of the surfaces of 3MPSF_as, 3MPSF_et3h, 2MPSF_as, and 2MPSF_et3h are shown in Figure 5, Figure S17, and Figure 6. At a single nanometer scale, rough surface was observed on all the films. The disordered roughness on the surfaces of 3MPSF_as is not directly related to the 3D mesostructure because the average distance between bright areas is too long to be attributed to the convex shape arising from the cage-type 3D mesostructure 2152

DOI: 10.1021/acs.langmuir.6b04511 Langmuir 2017, 33, 2148−2156

Article

Langmuir

observations are consistent with those of SEM. Conversely, the brightness differed within single domains of the SEM images of 2MPSF_as and 2MPSF_et3h, exhibiting brighter areas in the center and darker areas at the edge of each domain (Figure S12). This suggests that the height of the surface varied. The AFM images of 2MPSF_as and 2MPSF_et3h show that the center of each domain is highest in center of each domain and that the surface position lowers toward the edges of the domain, forming a curved, conelike structure (Figure 6). These findings are consistent with those from the SEM images of the same specimens. Therefore, AFM is effective in determining the macroscopic flatness of the surfaces of thin films. The final topic discussed in this section is the shape of the surfaces of 3MPSF_as and 2MPSF_as. There are four possible types of surfaces as shown in Figure S18. Type A surfaces are covered with a flat thin silica layer. Type B surfaces are covered with a thin silica layer possessing an ordered roughness which reflects the mesostructure and can be traced by AFM. Type C surfaces are covered with a thin silica layer possessing disordered roughness. In type D surfaces, the surface mesopores are open, although the degree of openness may vary. The surfaces of the films are unlikely to be Type A because the AFM images do not show any mesoscopically flat surfaces. Previous reports on mesoporous silica thin films have used this type of surface as a schematic model, which we believe should be revised. Type D is possible based on of the ultralow voltage HR-SEM images, which indicate that thin silica layers exist on the surfaces of 3MPSF_as and 2MPSF_as. In the case of 3MPSF_as, a type C surface is probable because the AFM images show that the surface has disordered roughness. In the case of 2MPSF_as, type B is probable because the AFM images show that the surface has a periodic roughness, arising from the ordered mesostructure beneath. In conclusion, the combination of ultralow voltage SEM and AFM is suitable for determining the flatness and shapes of the outermost surfaces of the silica thin films on the nanometer to micrometer scale. Electron Penetration Range. Electron penetration range is an important factor for accurate interpretation of SEM images. The electron penetration range, R, has been defined in a previous study.26 R values can be approximately calculated using the Kanaya−Okayama formula (see page 23 of the Supporting Information). The values of R in our specimens are briefly described in Figure 1b. The R value is not always equal to the depth information obtained using secondary electrons (SEs) escaped from a specimen. This is because SEs escape from a region up to 10 nm below the surface of the specimen.27 When the R value is within the escape depth of the SEs (