NANO LETTERS
Controlled Interconversion of Nanoarray of Ta Dimples and High Aspect Ratio Ta Oxide Nanotubes
2009 Vol. 9, No. 4 1350-1355
Hany A. El-Sayed and Viola I. Birss* Department of Chemistry, UniVersity of Calgary, 2500 UniVersity DriVe NW, Calgary, Alberta, Canada T2N 1N4 Received October 3, 2008; Revised Manuscript Received January 21, 2009
ABSTRACT We report the controlled formation of either high-aspect-ratio Ta2O5 nanotubes or an organized nanoarray of Ta dimples by Ta anodization in a single H2SO4 + HF solution. Dimpled Ta is the stable surface morphology in the first few seconds, followed by the growth of dense and fully vertically aligned Ta2O5 nanotubes (up to 2.5 µm long). After 2 min, the dimpled surface morphology reappears, related to the build-up of a resistive Ta fluoride surface layer.
There is significant interest in the fabrication of nanoporous and nanotubular structures, due to their wide range of applications.1 These include separation technology, ion exchange, catalysis, sensing, the isolation and purification of biological species,2 and so forth. Inorganic nanoporous structures, such as metals and metal oxides, also have promise as components in heterojunction solar cells, water photolysis and electrolysis applications, fuel cells, molecular filtration, and tissue engineering.3 Nanoporous metal oxide structures produced by the electrochemical anodization of valve metals, such as Zr, Ti, W, Nb, Al,4-8 and recently Ta,9-11 have attracted additional interest because of their potential use in controlled catalysis, waveguides, and threedimensionally (3D) arranged Bragg-stack reflectors.12 All of these valve metals have been reported to form both disordered and ordered porous oxide structures by electrochemical anodization,4-8 except for Ta, at which only disordered porous oxide layers can be formed.9-11 Compact Ta oxide films (wide band gap of 3.9 eV13) can also be readily formed electrochemically, giving good mechanical strength and abrasion resistance,14 photoactivity in the nearUV range,13 good thermal and chemical stability (Ta2O5 is stable in almost any electrolyte except HF, which dissolves the oxide),15 a high melting point, ductility,14 and biocompatibility.16 While compact Ta2O5 films are of high quality and have a well-defined thickness that is directly proportional to the applied voltage at constant solution temperature,17,18 a key goal remains to be the fabrication of thicker, porous Ta oxide films, and particularly the formation of ordered Ta2O5 nanotubes. Therefore, our focus in the present work * To whom correspondence should be addressed. E-mail:
[email protected]. Phone: (403) 220-6432. Fax: (403) 289-9488. 10.1021/nl803010v CCC: $40.75 Published on Web 02/26/2009
2009 American Chemical Society
is on the formation of organized, nanotubular oxides on Ta by electrochemical means. The first reports of the formation of thick anodic Ta2O5 films, instead of the much more easily formed compact layers, showed that small amounts of HF added to 1 M H2SO4 produce a disordered, porous oxide layer.9-11 The fluoride ions are suggested to provide mild dissolution of the electrochemically generated oxide, which is continuously reformed in the presence of the high electric field. Although it is not yet understood why pores initiate during oxide formation at metals such as Zr, Ti, W, Nb, and Al,4-8 it is widely accepted that competition between oxide growth and dissolution leads to the continuous growth of porous oxide films. The thickness of the porous, disordered Ta2O5 layer prepared in 1 M H2SO4 + 2 wt % HF solution was reported to depend on the anodization time but did not exceed 800 nm9-11 even after 4 h of anodization. Although several practical applications have been proposed for these porous Ta2O5 layers,9-11 their usefulness has been limited by their lack of order and their limited thickness. Other efforts have focused on controlling oxide film morphology, including altering the anodizing solution viscosity. In the case of metals such as Ti, it has been established recently that anodization in viscous solvents,19 such as glycerol or ethylene glycol containing small amounts of fluoride, leads to the formation of TiO2 nanotubes of significantly improved morphology than those prepared in less viscous, aqueous solutions.12 Indeed, TiO2 nanotubes with extremely smooth and homogeneous walls, as well as high aspect ratios (length to diameter) of up to 175:1, can be formed in this way.19 However, when Ta anodization was similarly explored in these types of nonaqueous solutions,20
Figure 1. Series of FESEM images after anodization of polycrystalline Ta surface for (a) 5, (b) 10, (c) 20, (d) 60, (e) 90, and (f) 120 s in 16.4 M H2SO4 + 2.9 M HF at 15 V (vs a Pt counter electrode).
only a discontinuous porous Ta2O5 layer (thickness up to 16 µm) was formed. Another recent study showed that the anodization of Ta in a stirred, highly concentrated mixture of H2SO4 (95-98%) and HF (48%) in a volumetric ratio of 9:121,22 (i.e., 16.4 M H2SO4 and 2.9 M HF) for at least 5 min at potentials of 10-20 V actually removed the surface oxide film and resulted in an entirely new surface morphology. Rather than producing a Ta2O5 film, an electropolished Ta surface resulted, containing Nano Lett., Vol. 9, No. 4, 2009
highly ordered arrays of circular and monodisperse “dimples” (shallow pores) at the nanoscale. The absence of porous Ta2O5 was attributed to very rapid Ta2O5 dissolution in this aggressive medium.23,24 Using aqueous mixtures of 9:1 (vol/ vol) concentrated H2SO4 (95-98%) and HF (48%), containing small amounts of either ethylene glycol or dimethyl sulfoxide, Ta2O5 nanotubes were recently fabricated by Ta anodization.25 The growth of Ta2O5 nanotubes in this medium was attributed to the presence of organic additives. 1351
Here, we report the controlled formation of highly ordered, high-aspect-ratio Ta2O5 nanotubes of desired length by the anodization of Ta in the 9:1 H2SO4/HF medium, which is the same solution found previously21,22 to form only the dimpled Ta surface described above. These results clearly show that organic additives are not required to grow Ta2O5 nanotubes in concentrated HF/H2SO4 solutions, as has been recently reported.25 Further, by lengthening the time of anodization in this solution, the Ta2O5 nanotubes can be converted to the organized surface array of Ta dimples reported previously,21,22 and an explanation for this conversion process is given below. We also show by short-time FESEM studies that an organized array of dimples on Ta can be formed in only a few seconds of anodization (vs the previously reported 5-10 min), thus consuming orders of magnitude less metal by dissolution in the formation of this useful nanotemplate structure. In this work, Ta foil specimens (Alfa-Aesar, 99.95%, 0.127 mm) were first carefully rinsed in acetone, 2-propanol, and Millipore water (18.2 MΩ cm resistivity). Anodization was then carried out, typically at 15 V, in a magnetically stirred, room temperature (∼23 °C) solution of concentrated H2SO4 (95-98%) and HF (48%) in a volumetric ratio of 9:1 using a conventional two-electrode system (Ta foil working electrode (WE) and Pt wire counter electrode (CE)) connected to a Solartron 1287A Potentiostat or sometimes just a Harrison 6824A power supply. The distance between the WE and CE was kept constant at approximately 1 cm. Immediately after anodization, the Ta foil samples were rinsed copiously with deionized water. Field emission scanning electron microscope (FESEM) analysis (Hitachi S-4880 FESEM, Alberta Centre for Surface Engineering and Science, Edmonton, Alberta) of the Ta surfaces was typically performed using an electron energy of 20 keV. Figure 1 shows a series of FESEM images of a Ta surface after various anodization times from 5 to 120 s at 15 V (vs a Pt counter electrode). Figure 1a demonstrates that a dimpled surface structure is obtained in only ∼5 s of Ta anodization (vs more than 5 min, as reported previously21,22). However, the surface is not completely smooth, as seen by the presence of lighter and darker (rougher areas 100+ nm in diameter) regions in Figure 1a. Dimple formation in the first ∼5 s of anodization is then seen to be followed by nanotube growth with Figure 1b showing that short Ta2O5 nanotubes (∼250 nm in length) are already well formed in patches on the surface after only 10 s of anodization. After 20 s of anodization (Figure 1c), the nanotubes are vertically aligned on the surface with uniform pore diameters of approximately 20-25 nm and an average spacing between the tubes of ∼40 nm. The tubes are arranged in very tight bundles of ∼1.5 µm length and do not appear to be crosslinked in contrast to all previous descriptions of porous Ta2O5.9-11 After 60 s of anodization (Figure 1d), the nanotubes are notably longer, but are arranged in smaller patches than observed at 20 s (Figure 1c). When Ta is anodized for 90 s (Figure 1e), still longer Ta2O5 nanotubes (2-2.5 µm) are obtained, but they appear to be sealed at the pore entrance, presumably producing hollow Ta2O5 rods. 1352
Figure 2. FESEM-determined thickness of Ta2O5 nanotube layers (from Figure 1) formed by anodization of polycrystalline Ta in 16.4 M H2SO4 + 2.9 M HF at 15 V (vs Pt counter electrode) using different anodization times.
The formation of nanotube patches (Figure 1b-d), rather than a fully compact layer of nanotubes, likely arises from shrinkage effects experienced during the drying of the film after sample withdrawal from the acidic anodization solution. While the nanotubes generally exhibit excellent adhesion, some tubes are seen to be lost at regions between patches, as has been reported previously for oxide nanotubes at Ti.26 After a full 2 min of anodization, Figure 1f shows a fully dimpled surface, similar to that reported earlier for anodization times of 5 min or more.21 These results prove that the oxide nanotubes are no longer stable on the Ta surface in this solution at 15 V and also show clearly that the highly ordered dimples formed at 2 min and reported previously21,22 evolve from the nanotubes. Indeed, the typical dimple diameter in Figure 1f is ∼30 nm, while the nanotube inner diameter in the latter stages of their lifetime (Figure 1e) is ∼25 nm. Overall, it appears that anodically formed Ta2O5 nanotubes can only be stabilized on Ta in this solution at anodization times longer than ∼10 s and less than 2 min. The effect of anodization time on the FESEM-determined thickness of the Ta2O5 nanotubes at 15 V (Figure 1) is summarized in Figure 2. The nanotube length increases with increasing anodization time and reaches a maximum of ∼2.7 µm in 90 s, with the limiting film thickness at longer times ascribed to steady-state oxide growth and dissolution. The asymptotic nature of the curve may be attributed to increasing diffusion limitations of the fluoride ion into the nanotubes to the pore base as the pores lengthen. The thickness of porous anodic oxide films formed on metals such as Al and Ti has been argued previously11,19,27 to be controlled by a balance between the rate of electrochemical formation of the oxide film at the underlying metal surface and the rate of oxide dissolution, as shown in reactions 1 and 2, respectively, for Ta in H2SO4 + HF solutions. 2Ta + 5H2O f Ta2O5 + 10H+ + 10e
(1)
Ta2O5 + 10H+ + 14F- f 2[TaF7]2- + 5H2O
(2)
Reaction 1 occurs primarily at the base of the Ta2O5 pores, where a compact oxide film is thin enough for the applied field to drive ion transport through it, while reaction 2, a Nano Lett., Vol. 9, No. 4, 2009
Figure 3. Current density-time data obtained during the anodization of a Ta at 15 V for 30 s in 16.4 M H2SO4 + 2.9 M HF solution. Arrows mark the times associated with the collection of the SEM images in Figure 1.
chemical reaction, occurs at the oxide surface, both on the pore walls and the pore base. However, the rate of Ta2O5 dissolution (reaction 2) is expected to be much higher at the pore base versus the walls due to the field focusing effect.28 Thus, the key to forming high-aspect-ratio Ta2O5 nanotubes is to increase the dissolution rate of Ta2O5 at the pore base, which is electrochemically promoted, and suppress it on the pore walls (strictly a chemical process).12 In the present work, this is likely assisted by the high viscosity of the concentrated acid anodizing solution employed.15 Reaction 1 generates protons mainly at the pore base,28 and causes local, selfinduced acidification, resulting in a pH gradient along the nanotube. The high viscosity solution helps to maintain this pH gradient by decreasing the diffusion constants,11,19 thus autocatalytically enhancing Ta2O5 dissolution at the pore base and causing Ta2O5 nanotube formation. Consistent with these arguments, Ta anodization in a lower viscosity H2SO4 (1 M) solution containing similar HF concentrations did not produce Ta2O5 nanotubes.9-11 Although the ex-situ FESEM images (Figure 1) clearly show the stages of Ta2O5 nanotube growth during the first 90 s of Ta anodization at 15 V in the standard H2SO4/HF
solution, current density-time data can be used to monitor the anodization process in situ, as shown in Figure 3 for the first 30 s. Initially (
