NMR Spectroscopic Analysis of the Local Structure of Porous-Type

May 16, 2017 - Alumina is classified as an intermediate oxide that cannot form a glass and is industrially used as a multifunctional material. Intensi...
0 downloads 0 Views 3MB Size
Download PDF
Article pubs.acs.org/JPCC

NMR Spectroscopic Analysis of the Local Structure of Porous-Type Amorphous Alumina Prepared by Anodization Hideki Hashimoto,*,† Koji Yazawa,‡ Hidetaka Asoh,† and Sachiko Ono† †

Department of Applied Chemistry, School of Advanced Engineering, Kogakuin University, 2665-1 Nakano, Hachioji, Tokyo 192-0015, Japan ‡ JEOL RESONANCE Inc., 3-1-2 Musashino, Akishima, Tokyo 196-8558, Japan S Supporting Information *

ABSTRACT: Alumina is classified as an intermediate oxide that cannot form a glass and is industrially used as a multifunctional material. Intensive studies on the atomistic structure of amorphous alumina have been conducted because a fundamental understanding of its structure can be meaningful to the design of new materials and devices. Here we focused on anodic alumina as a model material for clarifying the atomistic structure of amorphous alumina and prepared anion-free and anion-incorporated poroustype amorphous alumina by anodization using chromic acid electrolyte and typical electrolytes (e.g., sulfuric acid, oxalic acid, and phosphoric acid), respectively. The local structure around aluminum atoms in the anodic alumina was investigated by nuclear magnetic resonance spectroscopy. We found that the structure of anodic amorphous alumina comprises AlO4, AlO5, and AlO6 units with predominant fractions of AlO5. We also observed that the fraction of each unit was 37.7, 54.3, and 8.0%, with an average coordination number (NAl−O) of 4.70 in the anion-free sample. The results of 1H−27Al cross-polarization with magic angle spinning measurements showed that the prevalence of AlO6 units decreased with the elimination of physisorbed water by heat treatment. We also suggest that the average NAl−O is affected by the depth, content, and/or species of incorporated anions and that the average NAl−O increases with increasing anion content in the case of samples prepared using sulfuric acid electrolyte. The results are quite meaningful because they provide precise information about the local structure of anodic amorphous alumina; previous reports on this subject have been inconsistent and controversial. We expect that the present results will accelerate structural analysis research on the atomistic structures of amorphous alumina.

■

INTRODUCTION Alumina has various polymorphs and is used as an industrially important multifunctional material. α-Alumina is widely used as a single-crystalline substrate for various electrical devices, and γalumina is used as a heterogeneous catalyst support in modern industry. The crystal structures of α-alumina and transition alumina such as γ-alumina have been clarified,1 and the surface structure of transition alumina has been intensively investigated both experimentally and in simulations from the viewpoint of catalytic properties.2−5 Amorphous alumina thin films are used as functional materials such as supports for catalytic nanoparticles, gate microelectronic devices, protective films, and surface coating material for lithium-ion battery electrodes because of their unique chemical durability and mechanical strength.6−8 However, the atomistic structure of amorphous alumina remains unclear. Amorphous alumina has also attracted much attention from researchers involved in glass science. The preparation of glasses that contain large amounts of alumina is intrinsically difficult because alumina is considered to be an intermediate oxide according to Sun’s glass formation criteria.9 Recently, a glass © XXXX American Chemical Society

with a large amount of alumina, 54Al2O3-46Ta2O5, was synthesized using the containerless aerodynamic-levitation technique.10 This glass is expected to be used as a substrate for electric devices, windows in buildings and cars, and cover glasses for smart-phones because it exhibits a remarkably high elastic modulus, high hardness, and excellent transparency. Clarifying the atomistic structure of amorphous alumina could provide answers to the fundamental question in glass science: What is the intrinsic structural character of an intermediate oxide such as alumina? Precise structural information on amorphous alumina could also provide new guidelines for preparing glasses with large amounts of alumina. In this study, we focus on anodic porous-type alumina as a target material for the structural analysis of amorphous alumina. Anodic porous-type amorphous alumina films are formed on aluminum substrates by applying voltage or flowing current to an aluminum anode in an acidic or basic electrolyte. Anion, Received: April 18, 2017 Revised: May 15, 2017 Published: May 16, 2017 A

DOI: 10.1021/acs.jpcc.7b03629 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

of aluminum atoms around physisorbed water in detail. We also investigated the influence of anion incorporation on the local structure of anodic alumina by comparing the previously reported local structures24−27 with a central focus on anion-free alumina sample prepared using H2CrO4 electrolyte.

which comes from electrolyte, -free and anion-incorporated alumina are formed depending on the electrolyte species. Anodic porous alumina has a long industrial history and has attracted much attention from nanotechnological fields,11 whereas the atomistic structure of anodic alumina is not sufficiently understood. Anodic porous alumina is considered to be a suitable model material for clarifying the atomistic structure of amorphous alumina, for the following two reasons: (i) the film thickness, nanostructure, incorporated anion species, and the depth and/or the amount of anion incorporation into the cell wall can be easily controlled by changing the electrolytic conditions,11−15 and (ii) relatively large amount of solid-state amorphous alumina samples can be easily obtained by anodization and subsequent detachment from aluminum substrate. The final goal of this study is to clarify the atomistic structure of anodic alumina and the intrinsic structure of amorphous alumina. Amorphous structural analyses of anodic alumina started in 1970, when Takahashi et al. used X-ray fluorescence spectroscopy (XRF) to characterize anodic alumina.16 Researchers have since further characterized anodic alumina using electron-probe microanalysis (EPMA),17 X-ray diffraction (XRD),18−21 extended X-ray absorption fine structure (EXAFS) spectroscopy,22,23 nuclear magnetic resonance (NMR) spectroscopy,24−29 electron diffraction (ED),30 XRF,31 neutron diffraction (ND),19,20 reverse Monte Carlo (RMC) modeling techniques,20 scanning transmission electron microscopy (STEM) combined with electron energy loss spectroscopy (EELS),32 and molecular dynamics (MD) simulations.33,34 These reports except for Yakovleva et al.21 studied porous-type amorphous alumina. Although local structural analyses have been conducted using XRF, EPMA, EXAFS, and NMR, the results and interpretations concerning, for example, coordination number (NAl−O), the number of oxygen atoms around a given aluminum atom, are different among the reports and controversial. Such ambiguity should be solved for future advanced structural analysis of this system. The middle-range ordering analyses for some anionincorporated anodic alumina have been conducted primarily through diffractometry; however, the results do not agree with the reported local structure. Attempts to clarify the middlerange ordering of prototypical amorphous alumina by MD simulations have been limited by a lack of precise experimental data related to the local structure. Furthermore, experimental diffraction data are only available for anodic alumina containing sulfate ions; no experimental diffraction data for anion-free prototypical alumina are available. The fractions of AlO4, AlO5, and AlO6 polyhedral units, and their connectivity in anodic porous alumina have not yet been clarified. In this paper, we attempt to obtain precise experimental local structural data for anodic porous alumina as the first step toward revealing the atomistic structure of amorphous alumina. An anion-free prototypical alumina sample was prepared using chromic acid (H2CrO4) as the electrolyte. Anionincorporated alumina samples were prepared using typical electrolytes of sulfuric acid (H2SO4), oxalic acid ((COOH)2), and phosphoric acid (H3PO4). The local structures of the prepared samples were investigated by 27Al magic angle spinning (MAS) NMR, 27Al triple-quantum MAS (3QMAS) NMR, and 1H−27Al cross−polarization (CP) MAS NMR experiments. By analyzing the obtained spectra, we calculated the fraction of oxygen-coordinated polyhedra (AlOx) and the average NAl−O and investigated the coordination environment

■

EXPERIMENTAL SECTION Sample Preparation. A high-purity (99.99%) aluminum sheet was immersed in 1.25 mol dm−3 NaOH for 20 s at 60 °C, washed with tap water, immersed in 3.9 mol dm−3 HNO3 for 60 s, and finally washed with deionized water to remove surface native oxides and contaminants. Anodization was performed under a constant-current condition of 100 A m−2 in 1.5 mol dm−3 H2SO4 for 1 h at 20 °C. Constant-voltages of 25, 40, 80, and 120 V were applied in 0.3 mol dm−3 H2SO4 for 3 h at 20 °C, 0.3 mol dm−3 (COOH)2 for 3 h at 30 °C, 0.3 mol dm−3 H2CrO4 for 5 h at 40 °C, and 0.4 mol dm−3 H3PO4 for 5 h at 25 °C to obtain ∼50-μm-thick film. The anodization conditions are summarized in Table 1. Table 1. Anodization Conditions Used to Prepare Alumina Filmsa concentration/ mol dm−3

temperature/°C

CC or CV

time/h

H2SO4

1.5

20

1

H2SO4 (COOH)2 H2CrO4 H3PO4

0.3 0.3 0.3 0.4

20 30 40 25

CC at 100 A m−2 CV at 25 V CV at 40 V CV at 80 V CV at 120 V

electrolyte

3 3 5 5

a

CC and CV represent constant current and constant voltage, respectively.

After anodization, the samples were carefully washed with deionized water to remove residual electrolyte. To detach the anodic alumina films prepared using H2SO4, (COOH)2, and H2CrO4 electrolytes from the aluminum substrates, anodic polarization was applied in a mixed solution of 1:4 vol % of perchloric acid (60%) and ethanol (99.5%) for 1 min at voltages of 10 V higher than the formation voltages.35 For the sample prepared using H3PO4, the film was detached by using the two-layer anodization reported by Yanagishita et al.36 The anodization under a constant voltage of 120 V in 18 mol dm−3 H2SO4 was conducted for 40 min at 20 °C to form easily soluble alumina layer at the interface between the upper film formed in H3PO4 and the aluminum substrate. Subsequent immersion in 2 wt % H3PO4 for 15 min at 30 °C selectively dissolved the film generated by anodization in H2SO4, detaching the upper film formed in H3PO4 from the aluminum substrate. The detached films were carefully washed with deionized water, dried at room temperature, and crushed into powder by an agate mortar. The scanning electron microscopy (SEM) images of the detached films were obtained using a JEOL JSM-6701F microscope to characterize the microstructure of the films. To remove physisorbed water, the powder samples were heated at 300 °C for 4 h at an heating rate of 5 °C min−1 and were subsequently cooled in the furnace. Photographs of the detached alumina films and powder samples are shown in Figure S1 in the Supporting Information. 27 Al NMR Measurements. 27Al MAS NMR measurements of as-prepared and heated samples were performed on a JEOL JNM-ECZ600R (14.1 T) spectrometer at a 27Al Larmor B

DOI: 10.1021/acs.jpcc.7b03629 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

mol dm−3 H2SO4 electrolyte, as shown in the inset of Figure 1. The electrolytic behaviors are characteristic of the formation of anodic porous alumina. Figure 2a,b show typical surface and cross-sectional SEM images, respectively, of the as-anodized sample prepared using

frequency of 156.34 MHz. The samples were packed in zirconia rotors and spun at 20 kHz using a 3.2 mm HXMAS probe. The 27 Al chemical shift δiso in parts per million (ppm) was referenced to an external 1 mol dm−3 AlCl3 solution (−0.1 ppm). The NAl−O was analyzed using 27Al single-pulse MAS spectra and 27Al 3QMAS spectra. Structural analysis of nearadsorbed water was carried out using by 1H−27Al CP MAS spectra. 27Al single-pulse MAS NMR spectra were recorded using a 30° pulse (0.7 μs) and a relaxation delay of 1 s, and 1000 free induction decays (FIDs) were accumulated. 27Al 3QMAS NMR spectra were obtained using three pulse sequence employing a z-filter. The triple quantum excitation and conversion pulse lengths were optimized to 3.0 and 1.0 μs, respectively, and the length of the selective pulse was 13.0 μs. A series of 128 t1 slices were collected using 48 FIDs with a recycle delay of 1 s for each slice. The 1H−27Al CPMAS NMR spectra were obtained with a contact time of 1 ms and a recycle delay of 2 s, and 16 000 FIDs were accumulated. The 27Al single pulse MAS spectra were decomposed into the three components and the fitting values, the averaged isotropic chemical shift ( δiso), the width of the Gaussian distribution of δiso (ΔCS), and the averaged quadrupolar coupling constant ( CQ ) were determined using the “Dmfit” program37 applying a simple Czjzek model. The average NAl−O was determined by the following equation: NAl − O = ∑ NAN , where N and AN represent the number of oxygen atoms around a given aluminum atom and relative ratio of the corresponding peak area, respectively.

Figure 2. SEM images of the as-prepared sample prepared using 1.5 mol dm−3 H2SO4 electrolyte: (a) surface and (b) cross-sectional images.

1.5 mol dm−3 H2SO4 electrolyte, which was used as a representative sample. The sample contained straight pore channels, which are distinctive characteristics of anodic porous alumina, indicating a honeycomb-like cell structure. Among the as-prepared samples, the as-anodized sample prepared using 1.5 mol dm−3 H2SO4 electrolyte exhibited the smallest pore size (∼15 nm) and the smallest interpore distance (∼30 nm), as estimated from the images in Figure 2b, respectively, because of its lowest formation voltage of 15 V. The cell dimensions of anodic porous alumina, such as the interpore distance, barrier layer thickness, and the pore diameter, are known to be proportional to the formation voltage.11,15 This structural tunability has attracted broad interest from researchers in nanotechnology-related fields (see the review of Lee and Park for details of the cell structure of anodic porous alumina).11 The depth and the amount of anion incorporation into the cell walls are also known to be highly dependent on the anodization electrolytes employed.12−14 When aluminum is anodized in H2CrO4 electrolyte, anion-free prototypical alumina can be obtained. By contrast, anodic alumina formed in the H2SO4 electrolyte contains SO42− ions in the cell wall with an incorporated region of 100%, i.e., without an anion-free region. The anodic alumina formed in (COOH)2 or H3PO4 electrolyte has a duplex structure of anion-free inner layers and C2O42−- or PO43−-containing outer layers. The schematic drawings of the anion-incorporation in the cell structure are shown in Figure S2 in the Supporting Information.13 The depth of anion incorporation and the amount of anions incorporated are both known to increase in the electrolyte order H2CrO4 < H3PO4 < (COOH)2 < H2SO4. We successfully prepared five typical samples with different cell structures, including anionfree prototypical alumina as a standard for comparison. 27 Al Single-Pulse MAS NMR Results. Figure 3a shows 27 Al single-pulse MAS NMR spectra normalized with the total peak area for the sample prepared using H2CrO4, H3PO4, (COOH)2, 0.3 mol dm−3 H2SO4, and 1.5 mol dm−3 H2SO4 electrolytes. The as-prepared and heated samples are represented by solid and dotted lines, respectively. Heating was conducted to remove the physisorbed water. Each spectrum consists of three broad peaks positioned at ∼66, ∼

■

RESULTS AND DISCUSSION Sample Preparation. Figure 1 shows the voltage−time curve of the sample prepared using 1.5 mol dm−3 H2SO4 and

Figure 1. Current−time curves of the samples prepared using 0.3 mol dm−3 H2SO4, (COOH)2, H2CrO4, and H3PO4 electrolytes. The inset shows the voltage−time curve for the sample prepared using 1.5 mol dm−3 H2SO4 electrolyte.

current−time curves of the other samples. The current values were nearly constant at ∼50, ∼ 70, ∼ 100, and ∼100 A m−2 under applied constant voltages of 80, 120, 40, and 25 V using H2CrO4, H3PO4, (COOH) 2, and 0.3 mol dm−3 H2SO4 electrolytes, respectively. The voltage was nearly constant at 15 V under an applied constant current of 100 A m−2 using 1.5 C

DOI: 10.1021/acs.jpcc.7b03629 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C

low temperature of ∼300 °C.41,42 To confirm the possibility that the decreasing [6]Al by heating is caused by the elimination of physisorbed water, we carried out 1H−27Al CPMAS measurements for both as-prepared and heated samples. This measurement can selectively detect the coordination environment around aluminum atoms existing near hydrogen atoms. That is, in the case of anodic alumina films, the CPMAS spectra shown in Figure 3b mainly provide information about aluminum atoms near hydrogen atoms of physisorbed water, i.e., the coordination environment of aluminum atoms on the outermost surface of the cell wall; by contrast, the single-pulse spectra shown in Figure 3a provide averaged information about the coordination environment of aluminum atoms. Figure 3b shows 1H−27Al CPMAS spectra normalized by the total peak area. In the spectra for all as-prepared samples, [6]Al was detected as the most prominent peak and the spectral profile was similar to each other. This result differs from that of singlepulse spectra, where the most prominent peak was that of [5]Al. A majority of [6]Al exists near surface-physisorbed water, i.e., at the outermost surface of the cell wall. The fact that [6]Al existing on the outermost surface of the cell wall decreased after heating provides important information about the local structure of the cell wall: (i) oxygen atoms of water molecules coordinate to [4]Al and/or [5]Al to form [6]Al in the as-prepared samples; (ii) the formed [6]Al near the surface decreases along with elimination of physisorbed water during heating. Because oxygen atoms of water molecules are known to coordinate to [4] Al in aluminum phosphate molecular sieves to form [6]Al,43 a similar phenomenon might occur even in the anodic amorphous alumina samples. The 1H−27Al CPMAS spectral signals were still detected for the heated samples even though we expected them to have almost no physisorbed water, the appearance of these signals presumably indicates that a tiny amount of residual physisorbed water still remained, water molecules in air readsorbed onto the samples, and/or a tiny amount of hydroxyl groups inside the samples were detected. The 1H−27Al CPMAS spectra of the samples prepared using H3PO4 and H2CrO4 electrolytes showed a low signal-to-noise ratio, which can be explained by the relatively small surface areas of these two samples compared to those of other samples: the surface area should be in the order of the samples prepared by H2SO4 > (COOH)2 > H2CrO4 > H3PO4 electrolytes considering their interpore distance and porosity. High applied voltages of 80 V for the H2CrO4 electrolyte and 120 V for the H3PO4 electrolyte resulted in large interpore distance and pore sizes of the resultant oxide films, which, in turn, led to a small surface area of the samples and small amount of physisorbed water. Comparison with Previous Reports. Kobayashi et al. collected NMR spectra of anodic films prepared using H2SO4 and (COOH)2 electrolytes and reported the existence of [5]Al in the samples and differing ratios of [4]Al, [5]Al, and [6]Al between the two samples in 1989.24 In the same year, Farnan et al. collected NMR spectra of films prepared using various electrolytes and reported the films prepared using H2CrO4 electrolyte were composed of 100% [6]Al, whereas those prepared using H3PO4 electrolyte were composed of [4]Al and [5] Al and those prepared using (COOH) 2 and H 2 SO 4 electrolytes were composed of [4]Al, [5]Al, and [6]Al.25,26 Some of the results of Farnan et al. differ from those obtained in the present work. The sample prepared using the H2 CrO4 electrolyte by Farnan et al. might have been a crystalline aluminum hydroxide because the reported NMR spectrum

Figure 3. NMR spectra of the samples prepared using various electrolytes: (a) 27Al single-pulse MAS and (b) 1H−27Al CPMAS spectra of the samples. The type of electrolyte is specified on the left side of each spectrum. Solid and dotted lines show the spectra of asprepared and heated samples, respectively. All spectra were normalized by the total peak area.

39, and ∼12 ppm which are assigned to four- ([4]Al), five([5]Al), and 6-fold ([6]Al) oxygen-coordinated aluminum atoms, respectively.38 Although the peak widths and positions differed slightly among the samples, all of the samples showed analogous spectral profiles and contained common components of [4]Al, [5]Al, and [6]Al. The spectra of the samples prepared using H2CrO4 and H3PO4 electrolytes mainly comprised [4]Al and [5]Al with small amounts of [6]Al. Recently, the local structure around aluminum atom for amorphous alumina thin films prepared by vapor deposition techniques has been quantitatively studied by NMR spectroscopy, which revealed that the amorphous alumina is composed of AlO4 tetrahedra, AlO5 polyhedra, and AlO6 octahedra with predominant fractions of AlO4 and AlO5 units.38−40 The spectral profiles of our samples prepared by anodization are similar to those for amorphous alumina thin films38−40 and for the glass containing large amounts of Al2O3, 54Al2O3-46Ta2O5.10 A distinct [6]Al peak was observed in the spectra of the samples prepared using (COOH)2 and H2SO4 electrolytes containing relatively large amounts of anions; the sample prepared using 1.5 mol dm−3 H2SO4 electrolyte, in particular, showed a prominent [6]Al peak, suggesting that the depth and/or the amount of anions incorporated into the cell wall are likely associated with the amount of [6]Al in the amorphous alumina structure. Influence of Physisorbed Water. After the heating process, the spectra of the samples prepared using H2CrO4 and H3PO4 electrolytes showed no clear difference; however, the intensity of the [6]Al peak clearly decreased in the samples prepared by (COOH)2 and H2SO4, in agreement with the results reported by Iijima et al.27 Here, the reduction of the spectral component of [6]Al by heating was analyzed and discussed in detail as follows. Physisorbed water on the cell wall is known to eliminate when anodic alumina films are heated at D

DOI: 10.1021/acs.jpcc.7b03629 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C showed a sharp peak, suggesting a crystalline material, and because the reported chemical shift value of 6.5 ppm is similar to that of boehmite (γ-AlOOH).44 The NMR spectra of their other samples showed broad peaks, suggesting amorphous materials. However, the resolution of their spectra is lower than that of our spectra because they used a lower magnetic field of 8.45 T (93.8 MHz for 27Al) and because the spinning sideband overlapped due to a lower rotation frequency of 5 kHz. These effects make peak deconvolution of the spectra difficult, further contributing to the differences between the results of Farnan et al. and those obtained in the present work. Despite their technical limitation, the works of Kobayashi et al. and Farnan et al. nonetheless represent pioneering studies on the local structure of anodic alumina.24−26 Iijima et al. conducted NMR spectroscopic analyses of anodic porous alumina prepared using H2SO4, (COOH)2, and H3PO4 electrolytes and proposed that the electrolyte-contacting outer layer of the cell structure is mainly composed of [6]Al, whereas the inner layer, which located inside the cell wall, is mainly composed of [4]Al and [5]Al, as shown in Figure 2 in ref 27. Along this line of thinking, the anion-incorporated outer layer is mainly composed of [6]Al and the anion-free inner layer is mainly composed of [4]Al and [5]Al. However, our interpretation differs from that of Iijima et al. We showed following two facts in this study: (i) the 1H−27Al CPMAS spectra of all of the samples, which reflect the coordination environment of aluminum atoms on the outermost surface of the cell wall, composed of not only [6]Al but also [4]Al and [5]Al; (ii) the anion-free prototypical alumina sample prepared by H2CrO4 electrolyte is composed of not only [4]Al and [5]Al but also [6]Al. Therefore, we consider that whether the distribution of local structure inside the cell can be appropriately described as the simple duplex structure proposed by Iijima et al. is uncertain. The problem is that their duplex structure is constructed without taking into account the actual cell structures such as interpore distance, the anion content, and the depth of anion incorporation into the cell wall. Additionally, we note that it is hard to discuss the distribution of local structure of anodic porous alumina using only the results of single-pulse 27Al MAS NMR spectra, which generally provides averaged information about the samples. We expect that the distribution of the local structure of anodic porous alumina could be clarified through a combination of various analytical methods such as STEM/ EELS, diffractometry using quantum beams, and computer simulations. Quantification of Coordination Number. Figure 4a shows a typical 27Al 3QMAS spectrum of the anion-free sample prepared using H2CrO4 electrolyte. The spectra of samples formed in the other electrolytes were similar to each other except for the peak area of [6]Al. The spectra of the sample prepared using 1.5 mol dm−3 H2SO4 are shown in Figure S3 in the Supporting Information as a typical example. For amorphous materials, an isotropic chemical shift (δ) and a quadrupolar coupling constant (CQ) generally elongate along the chemical sift (CS) axis and the quadrupolar-induced shift (QIS) axis, respectively. The 27Al peaks for all of the samples elongated along the CS and QIS axes, indicating that all samples showed distinctive spectra of amorphous materials despite the difference in electrolyte species. To obtain the percentages of each [n]Al sites, deconvolution of 27Al MAS spectra was performed on the basis of the “Czjzek” model, which considers the distribution of CQ and CS using the Dmfit program. The δ and CQ values estimated from the 27Al 3QMAS

Figure 4. NMR spectra of the sample prepared using H2CrO4 electrolyte. (a) Two-dimensional 3QMAS spectrum and projections to the F1 and F2 dimensions, whose axes represent the isotropic and MAS axes of 27Al, respectively. The peaks marked with * show the spinning side bands of the main peak. CS and QIS represent the chemical shift and quadrupole-induced shift axes, respectively. (b) The total fitting curve of the 27Al single pulse spectrum corresponding to the sum of three Czjzek fitting curves: red, blue, and green dots represent [4]Al, [5]Al, and [6]Al, respectively. Black solid and dotted lines represent the experimental and total fitting results, respectively.

NMR spectra were used as initial values for the spectral fitting. The spectrum was deconvoluted into three components, and the agreement with measured data was excellent, as shown in Figure 4b. The fitted profiles and the deduced NMR parameters, including the δiso, ΔCS, and CQ values, are summarized in Figure S4 in the Supporting Information and Table 2, respectively. In all of the samples, the fraction of [5]Al was greater than 50%, the second-most predominant peak was that of [4]Al, and the fraction of [6]Al was small (4%−15%). The fraction of [5]Al in the thin films grown from the vapor phase and 54Al2O3-46Ta2O5 glass is reported to be 36−43% and 42%, respectively.38−40 Therefore, anodic porous amorphous alumina appears to have more penta-coordinated aluminum atoms than reported vapor grown amorphous alumina and aluminacontaining glass. These results imply that the local structure of amorphous alumina is dependent on preparation technique. The average NAl−O for the as-prepared samples prepared using H2CrO4, H3PO4, (COOH)2, 0.3 mol dm−3 H2SO4, and 1.5 mol dm−3 H2SO4 electrolytes were 4.70, 4.64, 4.69, 4.75, and 4.80, respectively, and the average NAl−O values of the respective heated samples were 4.65, 4.63, 4.66, 4.69, and 4.78. Although the relative ratios of [4]Al, [5]Al, and [6]Al for our samples were slightly different from those of vapor grown thin films and alumina-containing glass, the average NAl−O for our samples E

DOI: 10.1021/acs.jpcc.7b03629 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C Table 2. NMR Parameters of the Prepared Samplesa as-prepared electrolyte (interpore distance/ nm) H2CrO4 (220)

site [4]

Al Al [6] Al [4] Al [5] Al [6] Al [4] Al [5] Al [6] Al [4] Al [5] Al [6] Al [4] Al [5] Al [6] Al [5]

H3PO4 (350)

(COOH)2 (100)

0.3 M H2SO4 (60)

1.5 M H2SO4 (30)

a

after heating

δiso /ppm

ΔCS/ ppm

CQ /MHz

area/%

72.5 47.4 16.6 71.8 46.7 16.0 71.8 46.7 17.4 71.8 47.0 16.8 71.4 47.4 16.4

22.1 16.0 12.4 25.9 16.2 12.4 25.9 15.9 9.8 22.1 18.4 9.2 23.5 16.4 6.5

6.7 7.9 5.4 6.4 8.5 6.1 6.6 8.3 6.4 6.6 8.6 6.5 6.4 9.0 5.9

37.7 54.3 8.0 39.0 55.8 5.2 37.6 53.2 9.2 30.1 60.6 9.3 27.8 57.4 14.8

average NAl−O 4.70

4.66

4.72

4.79

4.87

δiso /ppm

ΔCS/ ppm

CQ /MHz

area/%

73.0 47.2 16.6 72.0 47.4 15.9 72.1 46.9 16.7 71.6 46.9 16.5 71.8 46.9 15.3

25.0 12.6 8.6 26.1 14.9 9.6 26.7 14.2 11.8 25.3 15.7 15.6 23.5 15.1 10.0

7.0 8.8 6.4 6.4 8.6 6.1 6.5 8.2 6.1 6.4 8.0 5.8 6.4 8.6 6.4

41.3 52.0 6.7 41.6 54.2 4.2 41.9 50.7 7.4 39.1 53.2 7.7 32.8 56.3 10.9

average NAl−O 4.65

4.63

4.66

4.69

4.78

NAl−O represents the number of oxygen atoms around a given aluminum atom. Interpore distance was measured by SEM observation.

showed similar values reported for thin films (4.48−4.93) and 54Al2O3-46Ta2O5 glass (4.70), suggesting that the anodic porous alumina samples are suitable model materials for atomistic structural analyses of amorphous alumina. Previous Reports on Middle-Range Ordering. Previous attempts to clarify the middle-range ordering of anodic alumina have been reported, and these attempts have involved incorrect local structural information, as described below. Oka et al.,18 ElMashri et al.,23 and Yakovleva et al.21 investigated anodic alumina using X-ray radial distribution functions, EXAFS techniques, and X-ray pair distribution functions, respectively. All three groups of authors concluded that anodic alumina was composed of AlO4 tetrahedra and AlO6 octahedra. Lamparter and Kniep studied the anodic porous alumina prepared by H2SO4 electrolyte using a combination of XRD, ND, and RMC methods.19,20 They concluded that the obtained RMC model was composed of AlO3 (20%), AlO4 (56%), and AlO5 (22%) with an average NAl−O of 3.94. The absence of AlO5 units in the former reports and the existence of AlO3 units and the absence of AlO6 units in the latter reports are incorrect local structural information because the present study and previously reported NMR studies24−27 demonstrate that anodic porous alumina is composed of AlO4, AlO5, and AlO6 polyhedral units. Gutiérrez and Johnsson33 and Hong34 investigated the structure of amorphous alumina using MD simulations. Whether the simulation models provided by them express the actual atomistic model is difficult to determine because their simulations were not based on the precise local structural information as described above. We expect that the simulation models are refined on the basis of our precise local structural data. Influence of Anion Content on Average Coordination Number. No remarkable difference is observed in the δiso, ΔCS, and CQ values for [4]Al, [5]Al, and [6]Al among the samples, whereas the fraction of integrated peak area that reflects the NAl−O ratio changes systematically, as shown in Table 2. Here, the fraction of integrated peak area for [4]Al, [5] Al, and [6]Al was plotted against the total anion content in the samples as shown in Figure 5a. The total anion content of the

Figure 5. (a) Fractions of peak area of [4]Al (black markers), [5]Al (red markers), and [6]Al (blue markers) and (b) the average NAl−O plotted as a function of the anion content in the samples. The anion content increases in the order of the samples prepared using H2CrO4 < H3PO4 < (COOH)2 < 0.3 mol dm−3 H2SO4 < 1.5 mol dm−3 H2SO4 electrolytes. Solid and open markers represent the as-prepared and heated samples, respectively.

samples was estimated on the basis of the results of thermogravimetry/differential thermal analysis, thermogravimetry/mass spectrometry, and inductively coupled plasma mass spectrometry. These results will be published elsewhere. For the as-prepared samples, the fraction of [4]Al (solid black circles) tends to decrease, and the fractions of [6]Al (solid blue triangles) and [5]Al (solid red squares) tend to increase with increasing anion content; the same trend is observed for the heated samples (open markers). The variation of average NAl−O F

DOI: 10.1021/acs.jpcc.7b03629 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C calculated from the fraction of integrated peak area of [4]Al, [5] Al, and [6]Al is shown in Figure 5b. As a result, the average NAl−O appears to increase with increasing anion content for both the as-prepared and heated samples. A comparison of the average NAl−O of the as-prepared samples prepared using H2CrO4 and H2SO4 electrolytes indicates that the average NAl−O increases in the order of the samples prepared using H2CrO4 (anion-free) (4.70) < 0.3 mol dm−3 H2SO4 (low anion content) (4.75) < 1.5 mol dm−3 H2SO4 (high anion content) (4.87). In addition, the average NAl−O of the sample prepared using (COOH)2 (4.72) is higher than that of the sample prepared using H2CrO4 electrolyte (4.70). These results suggest that the average NAl−O of anodic amorphous alumina increases with increasing total anion content in the sample, and the depth of anion incorporation into the cell wall should affect the local structure of the samples. However, the average NAl−O of the sample prepared using H3PO4 electrolyte (4.66) was slightly lower than that of the anion-free sample prepared using H2CrO4 electrolyte (4.70). A comparison of the spectra shown in Figure 3a and Figure S5 in the Supporting Information reveals that the entire spectrum profile of the sample prepared using H3PO4 electrolyte is broader than that of the sample prepared using H2CrO4 electrolyte. These results suggest that not only the depth and/or the amount of anion incorporation into the cell wall but also the anion species affect the local structure of anodic porous alumina. Perspectives. We obtained precise information about the local structure of amorphous alumina prepared under typical electrolytic anodization conditions. The local structure of anion-free prototypical alumina prepared using H 2 CrO4 electrolyte, which is the most important basic data for future studies on the structure of amorphous alumina, is revealed for the first time in this study. We also found that the fraction of oxygen-coordinated polyhedra (AlO4, AlO5, and AlO6) is closely related to the depth and/or the amount of anion incorporation and the anion species. The results of the present study are quite meaningful for providing precise information about the local structure of anodic porous alumina, which has been inconsistent and controversial. The fact that a relatively large amount of solid-state amorphous alumina samples without substrate can be easily prepared by anodization of aluminum is a strong merit for future diffraction experiments. On the basis of the present work, we are planning to develop an atomistic structure model of amorphous alumina using a combination of diffraction experiments involving high-energy synchrotron Xray and/or neutron radiation and computer simulation methods. Such attempts might clarify the effect of depth and/or the amount of anion incorporation and anion species on the structure of anodic amorphous alumina and will lead to an intrinsic understanding of amorphous alumina as an intermediate oxide with poor glass-forming ability.

40% AlO4, and 4−15% AlO6 units. The average NAl−O for the as-prepared samples prepared using H 2 CrO 4 , H 3 PO 4 , (COOH)2, 0.3 mol dm−3 H2SO4, and 1.5 mol dm−3 H2SO4 electrolytes were 4.70, 4.64, 4.69, 4.75, and 4.80, respectively, and the average NAl−O of the heated samples were 4.65, 4.63, 4.66, 4.69, and 4.78, respectively. On the basis of the results of 27 Al single-pulse and 1H−27Al CPMAS NMR spectra of the asprepared and heated samples, the fraction of [6]Al was found to decrease, accompanied by the elimination of surface physisorbed water during heating. We also suggest that the average NAl−O is affected by the depth, content, and/or species of incorporated anions and that the average NAl−O increases with increasing anion content in the case of samples prepared using H2SO4 electrolyte. The present NMR results for anodic porous amorphous alumina are precise, and new data can be expected to provide basic information for future atomistic structural analyses of amorphous alumina.

■

ASSOCIATED CONTENT

S Supporting Information *

The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.7b03629. Photographs of the prepared samples, two-dimensional 3QMAS spectra, and fitting results. (PDF)

■

AUTHOR INFORMATION

Corresponding Author

*Tel./Fax.: +81 426284537. E-mail: [email protected]. jp. ORCID

Hideki Hashimoto: 0000-0003-4771-1240 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS We thank Dr. Tatsuya Masuda for sample preparation and Mr. Yuki Fujita for sample preparation and SEM observations. A part of this study was financially supported by Kurita Water and Environment Foundation. We also acknowledge a Strategic Research Foundation Grant-aided Project for Private Universities matching fund subsidy from the Ministry of Education, Culture, Sports, Science, and Technology of Japan.

■

REFERENCES

(1) Levin, I.; Brandon, D. Metastable Alumina Polymorphs: Crystal Structures and Transition Sequences. J. Am. Ceram. Soc. 1998, 81, 1995−2012. (2) Sohlberg, K.; Pennycook, S. J.; Pantelides, S. T. Hydrogen and the Structure of the Transition Aluminas. J. Am. Chem. Soc. 1999, 121, 7493−7499. (3) Lizárraga, R.; Holmström, E.; Parker, S. C.; Arrouvel, C. Structural Characterization of Amorphous Alumina and its Polymorphs from First-Principles XPS and NMR Calculations. Phys. Rev. B: Condens. Matter Mater. Phys. 2011, 83, 094201. (4) Coleman, S. P.; Spearot, D. E. Atomistic Simulation and Virtual Diffraction Characterization of Stable and Metastable Alumina Surfaces. Acta Mater. 2014, 78, 354−368. (5) Wischert, R.; Florian, P.; Copéret, C.; Massiot, D.; Sautet, P. Visibility of Al Surface Sites of γ-Alumina: A Combined Computational and Experimental Point of View. J. Phys. Chem. C 2014, 118, 15292−15299.

■

CONCLUSIONS We focused on porous amorphous alumina prepared by anodization as a model material for revealing the structure of amorphous alumina and conducted local structural analyses of anodic alumina via solid-state 27Al MAS NMR spectroscopy. The five samples were prepared using typical electrolytes, 0.3 mol dm−3 H2CrO4, 0.4 mol dm−3 H3PO4, 0.3 mol dm−3 (COOH)2, 0.3 mol dm−3 H2SO4, and 1.5 mol dm−3 H2SO4, and the as-prepared and heated samples were measured using an NMR spectrometer. All of the samples were composed of AlO4, AlO5, and AlO6 polyhedral units with 50−60% AlO5, 30− G

DOI: 10.1021/acs.jpcc.7b03629 J. Phys. Chem. C XXXX, XXX, XXX−XXX

Article

The Journal of Physical Chemistry C (6) Kresse, G.; Schmid, M.; Napetschnig, E.; Shishkin, M.; Kohler, L.; Varga, P. Structure of the Ultrathin Aluminum Oxide Film on NiAl (110). Science 2005, 308, 1440−1442. (7) Stierle, A.; Renner, F.; Streitel, R.; Dosch, H.; Drube, W.; Cowie, B. C. X-ray Diffraction Study of the Ultrathin Al2O3 Layer on NiAl (110). Science 2004, 303, 1652−1656. (8) Scott, I. D.; Jung, Y. S.; Cavanagh, A. S.; Yan, Y.; Dillon, A. C.; George, S. M.; Lee, S. H. Ultrathin Coatings on Nano-LiCoO2 for Liion Vehicular Applications. Nano Lett. 2011, 11, 414−418. (9) Sun, K. H. Fundamental Condition of Glass Formation. J. Am. Ceram. Soc. 1947, 30, 277−281. (10) Rosales-Sosa, G. A.; Masuno, A.; Higo, Y.; Inoue, H.; Yanaba, Y.; Mizoguchi, T.; Umada, T.; Okamura, K.; Kato, K.; Watanabe, Y. High Elastic Moduli of a 54Al2O3-46Ta2O5 Glass Fabricated via Containerless Processing. Sci. Rep. 2015, 5, 15233. (11) Lee, W.; Park, S. J. Porous Anodic Aluminum Oxide: Anodization and Templated Synthesis of Functional Nanostructures. Chem. Rev. 2014, 114, 7487−556. (12) Thompson, G. E.; Furneaux, R. C.; Wood, G. C. Electron Microscopy of Ion Beam Thinned Porous Anodic Films Formed on Aluminium. Corros. Sci. 1978, 18, 481−498. (13) Thompson, G. E.; Wood, G. C. Porous Anodic Film Formation on Aluminium. Nature 1981, 290, 230−232. (14) Ono, S.; Masuko, N. The Duplex Structure of Cell Walls of Porous Anodic Films Formed on Aluminium. Corros. Sci. 1992, 33, 503−507. (15) Ono, S.; Masuko, N. Evaluation of Pore Diameter of Anodic Porous Films Formed on Aluminum. Surf. Coat. Technol. 2003, 169− 170, 139−142. (16) Takahashi, T.; Nagano, T.; Ikegaya, M.; Tagai, H. Studies on Coordination Numbers of Aluminum and Sulfur in Aluminum Anodic Coatings by Fluorescent X-ray Method. Kinzoku Hyomen Gijutsu 1970, 21, 612−616. (17) Takahashi, T.; Nagano, T.; Wada, K.; Ikegaya, M.; Tagai, H. Studies on Co-ordination Number of Aluminum in Aluminum Anodic Oxide Coatings. Kinzoku Hyomen Gijutsu 1971, 22, 400−403. (18) Oka, Y.; Takahashi, T.; Okada, K.; Iwai, S. Structural Analysis of Anodic Alumina Films. J. Non-Cryst. Solids 1979, 30, 349−357. (19) Kniep, R.; Lamparter, P.; Steeb, S. Structure of Anodic Oxide Coatings on Aluminum. Angew. Chem. 1989, 101, 975−977. (20) Lamparter, P.; Kniep, R. Structure of Amorphous Al2O3. Phys. B 1997, 234−236, 405−406. (21) Yakovleva, N. M.; Yakovlev, A. N.; Chupakhina, E. A. Structural Analysis of Alumina Films Produced by Two-Step Electrochemical Oxidation. Thin Solid Films 2000, 366, 37−42. (22) Saito, J.; Ikegaya, M.; Takahashi, T. Color and Film Structure of Anodic Oxide Film Formed on Aluminum. Kinzoku Hyomen Gijutsu 1976, 27, 511−520. (23) El-Mashri, S. M.; Jones, R. G.; Forty, A. J. An Electron-Yield EXAFS Study of Anodic-Oxide and Hydrated-Oxide Films on Pure Aluminium. Philos. Mag. A 1983, 48, 665−683. (24) Kobayashi, K.; Shimizu, K.; Kurima, Y. 27Al Solid-State MASNMR Spectroscopy of Amorphous Anodic Oxide Films on Aluminum. Hyomen Gijutsu 1989, 40, 486−487. (25) Farnan, I.; Dupree, R.; Forty, A. J.; Jeong, Y. S.; Thompson, G. E.; Wood, G. C. Structural Information about Aorphous Anodic Alumina from 27Al MAS NMR. Philos. Mag. Lett. 1989, 59, 189−195. (26) Farnan, I.; Dupree, R.; Jeong, Y.; Thompson, G. E.; Wood, G. C.; Forty, A. J. Structural Chemistry of Anodic Alumina. Thin Solid Films 1989, 173, 209−215. (27) Iijima, T.; Kato, S.; Ikeda, R.; Ohki, S.; Kido, G.; Tansho, M.; Shimizu, T. Structure of Duplex Oxide Layer in Porous Alumina Studied by 27Al MAS and MQMAS NMR. Chem. Lett. 2005, 34, 1286−1287. (28) Brown, I. W. M.; Bowden, M. E.; Kemmitt, T.; MacKenzie, K. J. D. Structural and Thermal Characterisation of Nanostructured Alumina Templates. Curr. Appl. Phys. 2006, 6, 557−561. (29) Kirchner, A.; MacKenzie, K. J. D.; Brown, I. W. M.; Kemmitt, T.; Bowden, M. E. Structural Characterisation of Heat-Treated Anodic

Alumina Membranes Prepared Using a Simplified Fabrication Process. J. Membr. Sci. 2007, 287, 264−270. (30) Kobayashi, M.; Niioka, Y. The Structure of Amorphous Anodic Oxide Films on Aluminum. Corros. Sci. 1990, 31, 237−242. (31) Parkhutik, V. P.; Belov, V. T.; Chernyckh, M. A. Study of Aluminium Anodization in Sulphuric and Chromic Acid Solutions-II. Oxide Morphology and Structure. Electrochim. Acta 1990, 35, 961− 966. (32) Nigo, S.; Kubota, M.; Harada, Y.; Hirayama, T.; Kato, S.; Kitazawa, H.; Kido, G. Conduction Band Caused by Oxygen Vacancies in Aluminum Oxide for Resistance Random Access Memory. J. Appl. Phys. 2012, 112, 033711. (33) Gutiérrez, G.; Johansson, B. Molecular Dynamics Study of Structural Properties of Amorphous Al2O3. Phys. Rev. B: Condens. Matter Mater. Phys. 2002, 65, 104202. (34) Hong, V. V. Molecular Dynamics Study on Structure and Properties of Liquid and Amorphous Al2O3. Phys. Rev. B: Condens. Matter Mater. Phys. 2004, 70, 134204. (35) Masuda, T.; Asoh, H.; Haraguchi, S.; Ono, S. Fabrication and Characterization of Single Phase α-Alumina Membranes with Tunable Pore Diameters. Materials 2015, 8, 1350−1368. (36) Yanagishita, T.; Masuda, H. High-Throughput Fabrication Process for Highly Ordered Through-Hole Porous Alumina Membranes Using Two-Layer Anodization. Electrochim. Acta 2015, 184, 80−85. (37) Massiot, D.; Fayon, F.; Capron, M.; King, I.; Le Calvé, S.; Alonso, B.; Durand, J.-O.; Bujoli, B.; Gan, Z.; Hoatson, G. Modelling One- and Two-Dimensional Solid-State NMR Spectra. Magn. Reson. Chem. 2002, 40, 70−76. (38) Lee, S. K.; Lee, S. B.; Park, S. Y.; Yi, Y. S.; Ahn, C. W. Structure of Amorphous Aluminum Oxide. Phys. Rev. Lett. 2009, 103, 095501. (39) Lee, S. K.; Park, S. Y.; Yi, Y. S.; Moon, J. Structure and Disorder in Amorphous Alumina Thin Films: Insights from High-Resolution Solid-State NMR. J. Phys. Chem. C 2010, 114, 13890−13894. (40) Sarou-Kanian, V.; Gleizes, A. N.; Florian, P.; Samélor, D.; Massiot, D.; Vahlas, C. Temperature-Dependent 4-, 5- and 6-Fold Coordination of Aluminum in MOCVD-Grown Amorphous Alumina Films: A Very High Field 27Al-NMR Study. J. Phys. Chem. C 2013, 117, 21965−21971. (41) Masuda, T.; Asoh, H.; Haraguchi, S.; Ono, S. Nanoporous αAlumina Membrane Prepared by Anodizing and Heat Treatment. Electrochemistry 2014, 82, 448−455. (42) Ono, S.; Nakamura, M.; Masuda, T.; Asoh, H. Fabrication of Nanoporous Crystalline Alumina Membrane by Anodization of Aluminum. Mater. Sci. Forum 2014, 783−786, 1470−1475. (43) Nakashiro, K.; Ono, Y. Characterizations of MAPO-36 Molecular Sieve by Solid-State N.M.R. Zeolites 1993, 13, 561−564. (44) O’Dell, L. A.; Savin, S. L.; Chadwick, A. V.; Smith, M. E. A 27Al MAS NMR Study of a Sol-Gel Produced Alumina: Identification of the NMR Parameters of the Theta-Al2O3 Transition Alumina Phase. Solid State Nucl. Magn. Reson. 2007, 31, 169−73.

H

DOI: 10.1021/acs.jpcc.7b03629 J. Phys. Chem. C XXXX, XXX, XXX−XXX