- Email: [email protected]
Fabrication of visible-light-responsive titanium dioxide layer on titanium using anodic oxidization in nitric acid
Applied Surface Science 270 (2013) 513–518
Contents lists available at SciVerse ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Fabrication of visible-light-responsive titanium dioxide layer on titanium using anodic oxidization in nitric acid Naofumi Ohtsu a,∗ , Hirotaka Kanno a , Shinji Komiya a , Yoshiteru Mizukoshi b , Naoya Masahashi b a b
Instrumental Analysis Center, Kitami Institute of Technology, Kitami 090-8507, Japan Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan
a r t i c l e
i n f o
Article history: Received 23 October 2012 Received in revised form 12 January 2013 Accepted 13 January 2013 Available online 23 January 2013 Keywords: Photocatalysis Titanium dioxide Nitrogen-doping Anodic oxidation Visible-light response
a b s t r a c t A visible-light-responsive titanium dioxide (TiO2 ) layer was fabricated by anodizing a Ti plate in aqueous nitric acid (HNO3 ) solutions, followed by annealing at 723 K for 5 h in air. The predominant structure of the oxide layer was TiO2 with an anatase structure that contained ca. 1 at% of incorporated nitrogen. The crystallite size of the anatase TiO2 was enlarged as anodizing voltage increased, but the atomic ratio of the incorporated nitrogen was almost constant. The TiO2 layers could degrade methylene blue (MB) solution under an UV light at 370 nm, exhibiting photocatalytic activity. The activity increased monotonically as the crystallite size of the oxide increased. The TiO2 layers with their crystallite size exceeding a specific value showed excellent response under visible light at 420, 450, and 505 nm. Overall, the results of this study showed that anodic oxidization of Ti in HNO3 solution is a promising technique for fabrication of a visible-light-responsive TiO2 layer. © 2013 Published by Elsevier B.V.
1. Introduction Titanium dioxide (TiO2 ) is one of the most prominent photocatalytic oxides owing to its high photocatalytic activity, low cost, nontoxicity, and environmentally friendly nature [1]. These characteristics have resulted in TiO2 attracting widespread attention for use in environmental purification media [2,3] and antibacterial coatings [4–7], which have been utilized as air purifiers and in biomedical products. However, TiO2 is a semiconductor with wide band gaps of 3.0 eV in its rutile phase and 3.2 eV in its anatase phase; accordingly, it only acts as a photocatalyst under ultraviolet (UV) light. Natural solar light is the most powerful light source, whereas UV light accounts for only a small portion of the total energy in light. As a result, the band gap must be adjusted to the visible-light region to utilize solar light effectively for activation of TiO2 photocatalysis instead of an artificial light source. In 1986, Sato demonstrated that a visible-light response appeared in response to doping TiO2 with NOx [8]. Asahi et al. also reported that the band gap of TiO2 decreased in response to substitution of a portion of oxygen for nitrogen, resulting in the compound acting as a photocatalyst under visible light [9]. Based on these findings, researchers have employed various methods to incorporate nitrogen into TiO2 . These methods have been
∗ Corresponding author. Tel.: +81 157269563; fax: +81 157269563. E-mail address: [email protected] (N. Ohtsu). 0169-4332/$ – see front matter © 2013 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.apsusc.2013.01.071
based on methods such as sol-gel synthesis [8,10–12], heating in a nitrogen-containing atmosphere [13,14,16,17] and oxidation of titanium nitride [15], or on physical processes such as ion implantation [18,19] and magnetron sputtering deposition [20,21]. The products fabricated by these methods have often shown somewhat different photocatalytic activities. It is believed that these differences originated from the location of nitrogen doping (N-doping) in the TiO2 structure; however, no general consensus has yet been reached regarding the location of nitrogen involved in photocalatytic activity. Alternative fabrication techniques might produce a N-doped TiO2 with novel photocatalytic properties [22]. Anodic oxidation is an industrial technique developed for production of a metallic oxide layer on a metallic substrate. This technique has recently also been shown to be useful for synthesis of TiO2 coatings on titanium (Ti) substrate. For example, Nakahira et al. fabricated a TiO2 coating with an anatase-type structure by anodizing Ti substrate in H3 PO4 electrolyte [23]. Similarly, Onoda et al. reported that a photocatalytic TiO2 coating with photocatalytic activity could be prepared by anodizing pre-nitrided Ti substrate in a mixed electrolyte composed of H2 SO4 , H3 PO4 , and H2 O2 [24]. The process of growing an anodic oxide layer can be explained using basic electrochemistry principles. The oxide growth is initiated by the penetration of oxygen and the production of a cation, the transport is explained by electric field-supported hopping. The electric field also attracts the anions contained in the electrolyte, which then react with the surface of the substrate. Therefore, anodic oxidization is advantageous because the oxide coating strongly adheres to the substrate and anions can be
514
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
easily incorporated into the coating by adding target anions in an electrolyte. By focusing on the latter advantage, anodic oxidation can be used to easily synthesize anion doped-TiO2 coatings via a single-step process. Mizukoshi et al. reported that sulfur-doped TiO2 coatings having rutile-type structures could be synthesized by anodizing Ti substrate in a highly concentrated sulfuric acid electrolyte [25]. The efficient photocatalytic activity of such coatings could bleach methylene blue (MB) aqueous solution under visible-light illumination at wavelengths greater than 413 nm. Based on these findings, it seems feasible that anodic oxidization in nitric acid (HNO3 ) can produce N-doped TiO2 coatings with visiblelight response. However, HNO3 acts as an oxidizing agent, thereby leading to dielectric breakdown of the TiO2 layer during anodizing. Nakahira et al. reported that, when employing 0.25-M HNO3 as an electrolyte, precipitates were produced during anodizing in exchange for degradation of the Ti plate owing to the dielectric breakdown [26]. Subsequent analysis of these precipitates revealed that they consisted of N-doped TiO2 . These findings indicate that suppression of the oxidizability of nitrate may enable the synthesis of a N-doped TiO2 layer as an alternative to TiO2 powder. In the present study, we attempted to fabricate a N-doped TiO2 coating by anodizing a Ti substrate in HNO3 with a concentration not exceeding 0.1 M. In addition, we investigated the photocatalytic activity under UV and visible-light regions from 370 nm to 505 nm. The effects of nitrogen doping and crystal structure on the photocatalytic properties were explored though careful analysis using X-ray diffractometry (XRD) and X-ray photoelectron spectroscopy (XPS).
2. Experimental procedures A pure titanium plate (99.9%) with dimensions of 20 mm × 10 mm × 1 mm was chemically polished using a colloidal silica suspension with an average particle size of 40 nm. This plate was used as a substrate. The substrate and a platinum electrode with dimensions of 20 mm × 10 mm × 1 mm were connected to the anode and cathode, respectively, after which direct current (DC) at 200 mA was supplied for 30 min in aqueous HNO3 solutions with concentrations of 1 mM, 10 mM, and 100 mM, respectively. Under this restricted current setting, the DC voltage could not be applied beyond a specific value owing to the dielectric breakdown of the oxide layer. The maximum voltages are actually 183 V, 40 V, and 22 V for 1 mM, 10 mM, and 100 mM, respectively. After the oxidation process was completed, the specimen was ultrasonically rinsed in distilled water and then annealed at 723 K for 5 h in the air. The surface morphology of the specimen was observed by scanning electron microscopy (SEM; JCM-5000 Neo Scope, JEOL) using secondary electron image mode at an acceleration voltage of 10 kV. The cross-sectional structure of the layer was observed by conventional transmission electron microscopy (TEM; JEM-2100, JEOL). The TEM specimens were prepared using an ion slicer (EM-09100IS, JEOL). X-ray diffractometry patterns (XRD; New D8 Advance, Bruker AXS) were measured with Bragg-Brentano geometry using Cu K␣ radiation. X-ray photoelectron spectroscopy (XPS; PHI 5000 Versa Probe, Ulvac-Phi) was conducted using monochromatized Al K˛ radiation (h = 1486.6 eV) to evaluate the incorporated nitrogen. The diameter of the X-ray probe was about 100 m, and the photoelectron take-off angle (TOA) was set at 65◦ , where TOA denotes the angle between the substrate surface and the spectrometer slit. An Ar-ion gun was used to analyze the nitrogen incorporated in the inner region. The photocatalytic activity was evaluated by a methylene blue (MB) degradation test. Before starting the degradation test, UV light was applied to the specimen for 90 min at an intensity of
ca. 6 mW cm−2 to eliminate the surface contaminants, after which the specimens were immersed in 10 mg L−1 of MB solution in a polypropylene vessel for 24 h to complete the adsorption of MB molecules onto the surface. UV and visible light from LED lamps consisted of 100 LEDs with wavelengths of 370 nm, 420 nm, 450 nm, or 505 nm were used to illuminate the vessel, and their radiation intensities were adjusted to 1 mW cm−2 . The photocatalytic activity was evaluated by measuring the absorbance of MB at 664 nm using the UV–vis spectrometer (UV-2400PC, Shimadzu) every 20 min during illumination for 180 min. The degradation rate of the MB was linearly plotted against the illumination periods, and reaction rate for MB degradation (mol L−1 min−1 ) were calculated from this plot.
3. Results and discussion 3.1. Morphology, structure and chemical state of the oxide layer anodized in HNO3 solution SEM images of Ti substrate anodized in 100-mM HNO3 with applied voltages of 15, 20, and 22 V are shown in Fig. 1. The image of non-treated Ti substrate is also shown in the figure. On the surface treated with 15 V, square-shaped pores on the micrometer scale with swelling at their boundaries were observed. When the voltage was increased to 20 V, the swelling disappeared and large hollows of ca. 100 m in size appeared on the surface. A collapsed surface containing winding and crossed grooves was formed when treated with 22 V. We consider that the NO3 − in the electrolyte facilitates the oxidization of Ti, thereby inducing dielectric breakdown of the oxide layer when applying the voltage beyond a specific value [27]. The hollows on the surface would be caused by the dielectric breakdown, and the collapsed surface containing grooves might be formed by continuing the breakdown during anodizing. To confirm our assumptions, we recorded the change in electric current between Ti substrate and the Pt electrode (Fig. 2). In the case of 15 V, the electric current decreased to near zero immediately after starting the oxidation, indicating that no dielectric breakdown occurred. In contrast, at 20 V, the current decreased slightly after starting, after which current with a specific value continued to flow during treatment. It is likely that the flowing current during the treatment is due to the dielectric breakdown in the oxide. Decreased current was not observed in the case of 22 V, implying that the intense dielectric breakdown occurs continuously. These changes in the electric current are direct evidence that the distinct morphology is related to dielectric breakdown occurring in the HNO3 solution. Furthermore, it should be noted that small precipitates were often found in the electrolytes after the oxidation process was completed. The precipitates were probably TiO2 produced via dielectric breakdown of the oxide layer, which was also reported by Nakahira et al. [26]. Fig. 3 shows a cross-sectional TEM image of the surface of the anodized Ti substrate corresponding to the Fig. 1(a). The TEM image shows that about 2-m thick oxide layer containing the grooves is formed upon the collapsed surface part of a Ti substrate. This indicates that the dielectric breakdown is most likely to induce the formation of the TiO2 layer. The XRD patterns of Ti substrate anodized in 100-mM HNO3 aqueous solution with various applied voltages are shown in Fig. 4. The peak observed at 25.4◦ is attributed to TiO2 with an anatase structure, indicating that the predominant phase in the oxide is anatase-type TiO2 . The peaks attributed to Ti with an hcp structure originated from the substrate. The peak intensity corresponding to the anatase TiO2 increased with increasing voltage. A similar trend in the XRD patterns was also observed on the surface anodized in 1and 10-mM HNO3 aqueous solutions with various applied voltages.
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
515
Fig. 1. SEM images of the Ti surfaces anodized in 100-mM HNO3 aquaous solution at (a) 22 V, (b) 20 V, and (c) 15 V and (d) non-treated Ti surface.
The mean crystallite sizes of anatase TiO2 were calculated using the whole powder pattern decomposition (WPPD) method based on Pawley’s algorithm (Fig. 5). The crystallite size of the anatase TiO2 increased with increasing voltage. These trends were also observed in solution with other concentrations of HNO3 . It is possible that
dielectric breakdown during anodic oxidation induces the growth of the TiO2 layer, thereby enlarging the size of the crystallite. The incorporation of nitrogen into the TiO2 layer fabricated in 100-mM HNO3 aqueous solution was examined using XPS. To obtain information from the inner region of the layer, the spectra
Fig. 2. Changes in electric current between Ti substrate and Pt electrode during anodic treatment.
516
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
Fig. 3. Cross-sectional bright field (BF) TEM image of the Ti surfaces anodized in 100-mM HNO3 aquaous solution at 22 V. Upper-inset is a high magnitude image.
were accumulated after etching the surface, ca. 300 nm, using an Ar-ion gun with an acceleration voltage of 3 kV. The nitrogen was detected in both TiO2 layers produced at the applied voltage of 15 V and 22 V, which correspond to the voltage below and at the dielectric breakdown voltage, respectively (Fig. 6). On the other hand, the nitrogen peaks were not detected from the Ti substrate before the anodization. Accordingly, the nitrogen found in the TiO2 layers is the incorporation from HNO3 aqueous solution. The observed N 1s XPS spectra are roughly decomposed into two peaks, one at approximately 399.8 eV and another at approximately 396.7 eV. The former is usually ascribed to nitrogen incorporated at a generic interstitial site, while the latter is attributed to atomic substituted nitrogen. Valentin et al. calculated the electric structure of Ndoped TiO2 using the density functional theory (DFT) method and found that both interstitial and substitutional nitrogen induce a few localized states that appear slightly above the valence band edge, accounting for the visible-light activity [22]. Furthermore,
θ Fig. 4. XRD patterns of Ti substrate anodized in 100-mM HNO3 aqueous solution with various applied voltages.
Fig. 5. Mean crystallite size of anatase TiO2 contained in the oxide layer fabricated with various applied voltages in 100-mM HNO3 solution.
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
Fig. 6. N 1s XPS spectra of the surface of anodized Ti at applied voltages of 15 V and 22 V in 100-mM HNO3 solution.
the atomic ratio of nitrogen in the TiO2 layer calculated from the spectral intensities of Ti 2p, O 1s, and N 1s XPS spectra were approximately 0.8 at%, which many researchers accept as an appropriate concentration to extend the photoactive region toward the visible-light region. Nitrogen impurities contained in commercial Ti plates are generally below 0.1 at%; therefore, the origin of nitrogen in the TiO2 layer is NO3 − in the electrolyte. 3.2. Methylene blue (MB) degradation test
visible-light illumination. On TiO2 layers anodized in the vicinity of the dielectric breakdown voltages (22 V for 100 mM, 40 V for 10 mM, and 183 V for 1 mM), a positive rate was observed in the visible-light region. Conversely, the rate was approximately zero when anodized with the voltage apart from the dielectric breakdown voltages. XPS confirmed that there were no differences in the amount and chemical state of nitrogen in the layer anodized at and below the breakdown voltage (Fig. 6). However, a significant difference in the response to visible light was found in these layers. It is likely that the photocatalytic activity in the visible-light region is not only dominated by the incorporated nitrogen but also another factor. In contrast to the rate under visible-light conditions, none of them at wavelengths below 370 nm was zero, indicating that all of the TiO2 layers have photocatalytic activity under UV light. The rate in the UV light monotonically increases with increasing applied voltage under all HNO3 concentrations. Furthermore, the rates at the dielectric breakdown voltage are almost constant, regardless of HNO3 concentration. 3.3. Relationship between photocatalytic activity and crystallinity of the TiO2 layer Here, the differences in the photocatalytic activity of the anodic TiO2 layers are explained on the basis of their crystallinity. The chemical reactions occurring on the photocatalytic TiO2 surface are triggered by the photogeneration of electron-hole pairs. The generated electrons and holes subsequently penetrate the oxide layer and react with surface species to form highly oxidizing species. Understandably, the recombination of electrons and holes during penetration should reduce the photocatalytic activity. It is generally accepted that a boundary at the interface between the two crystallites in a polycrystalline material becomes the recombination center of photogenerated electron and hole pairs because this boundary is considered to be a large crystal lattice defect.
μ
μ
μ
To evaluate the photocatalytic activity of the anodic TiO2 layer, reaction rate for MB degradation at various wavelengths were calculated. The variation in the reaction rate for MB degradation was plotted against the wavelengths of the LED light (Fig. 7). As expected, several TiO2 layers showed photocatalytic activity under
517
Fig. 7. Changes in reaction rates for MB degradation against the wavelength of LED light.
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
μ
518
fabricated layers consisted of polycrystalline TiO2 with an anatase structure and ca. 1 at% of incorporated nitrogen at a generic interstitial site and an atomic substituted site. As the applied voltage used for anodizing was increased, the crystallite size of the TiO2 was enlarged, while the atomic ratio of the incorporated nitrogen remained nearly constant. The voltage for the anodic treatment could not be applied beyond a specific value owing to the occurrence of a dielectric breakdown of the TiO2 layer, which occurred at 183 V, 40 V, and 22 V for 1 mM, 10 mM, and 100 mM, respectively. All layers could degrade methylene blue solution under an UV light of 370 nm, exhibiting photocatalytic activity that increased monotonically as the crystallite size increased. The photocatalytic activity under visible light at wavelengths of 420, 450, and 505 nm appeared on the TiO2 layer after the crystallite size exceeded a specific value. Acknowledgements
Fig. 8. Relationships between the reaction rate of MB and mean crystallite size of anatase TiO2 .
Consequently, polycrystalline TiO2 layers containing large crystallites are expected to show higher photocatalytic activity. In the present study, when the voltage applied for anodizing increased, both the reaction rate of MB degradation and the crystallite size of the TiO2 layer increased (Fig. 7 and Fig. 5, respectively). These results imply that the size of the crystallite is related to its photocatalytic activity. To better understand this relationship, we plotted the reaction rate of MB degradation against the crystallite sizes of the TiO2 layer, as shown in Fig. 8. Good correlation between the photocatalytic activity and the crystallite size was found when illuminating the specimen with UV light at 370 nm. Conversely, when illuminating the specimens with visible light at 420, 450 and 505 nm, a linear relationship was observed; however, this relationship was observed only when the crystallite size was greater than a specific value. In cases smaller than this specific crystallite size, the reaction rates were almost zero. Irie et al. suggested that, in cases in which TiO2−x Nx had lower values of x (<0.02), the band structure should have an isolated narrow band above the O 2p valence band [28]. UV light excites electrons in both the valence band and the narrow band, but visible light only excites electrons in the narrow band. As a result, visible-light illumination gives a lower quantitative efficiency (QY) value than UV-light illumination. Valentin et al. also reported that the hole generated in the O 2p band by visible-light illumination was less mobile than that generated by UV-light illumination [22]. The distinct relationship in the visible-light region observed in this study can be explained based on these previous reports. The electrons and holes photogenerated with visible light should be more affected by the crystallite boundary than those generated with UV light owing to the difference in the QY value and mobility. An anodic TiO2 layer with a crystallite size smaller than a specific value could not decompose MB under visible-light illumination owing to its low photocatalytic activity. Beyond the specific value, the photocatalytic activity exceeds the threshold activity to degrade MB under visible light; hence, the reaction rate increases linearly with increasing crystallite size. 4. Conclusions Nitrogen-doped TiO2 layers could be fabricated by anodizing a Ti plate in 1-, 10-, and 100-mM aqueous HNO3 solutions. The
The authors gratefully acknowledge Mr. Yamane and Mr. Tokuda from Kitami Institute of Technology for their help with the XRD and FE-SEM analyses. The authors also acknowledge Mr. Ishikawa of an undergraduate student of in our laboratory for his valuable input, and Mr. Nishioka of JEOL Ltd. for preparing the TEM specimen. References [1] A. Fujishima, T.N. Tao, D.A. Tryk, Journal of Photochemistry and Photobiology C1 (2000) 1–21. [2] R.W. Matthews, Journal of Physical Chemistry 91 (1987) 3328–3333. [3] Y. Ohko, K. Hashimoto, A. Fujishima, Journal of Physical Chemistry A 101 (1997) 8057–8062. [4] J.C. Yu, W. Ho, J. Lin, P.K. Wong, Environmental Science and Technology 37 (2003) 2296–2301. [5] Y. Ohko, Y. Utsumi, C. Niwa, T. Tatsumi, K. Kobayakawa, Y. Satoh, Y. Kubota, A. Fujishima, Journal of Biomedical Materials Research 58 (2001) 97–101. [6] Y. Yao, Y. Ohko, Y. Sekiguchi, A. Fujishima, Y. Kubota, Journal of Biomedical Materials Research 85B (2008) 453–460. [7] B. Kim, D. Kim, D. Cho, S. Cho, Chemosphere 52 (2003) 277–281. [8] S. Sato, Chemical Physics Letters 123 (1986) 126–128. [9] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 293 (2001) 269–271. [10] J.L. Gole, J.D. Stout, C. Burda, Y. Lou, X. Chen, Journal of Physical Chemistry B 108 (2004) 1230–1240. [11] S. Sakthivel, M. Janczarek, H. Kisch, Journal of Physical Chemistry B 108 (2004) 19384–19387. [12] S. Sato, R. Nakamura, S. Abe, Applied Catalysis B 284 (2005) 131–137. [13] H. Chen, A. Nambu, W. Wen, J. Graciani, Z. Zhong, J.C. Hanson, E. Fujita, J.A. Rodrigez, Journal of Physical Chemistry C 111 (2007) 1366–1372. [14] J. Yang, H. Bai, Q. Jiang, J. Lian, Thin Solid Films 516 (2008) 1736–1742. [15] T. Morikawa, R. Asahi, T. Ohwaki, K. Aoki, Y. Taga, Japanese Journal of Applied Physics 40 (2001) L561–L563. [16] O. Diwald, T.L. Thompson, T. Zubkov, E.G. Goralski, S.D. Walck, J.T. Yates Jr., Journal of Physical Chemistry B 108 (2004) 6004–6008. [17] M. Miyauchi, A. Ikezawa, H. Tobimatsu, H. Irie, K. Hashimoto, Physical Chemistry Chemical Physics 6 (2004) 865–870. [18] O. Diwald, T.L. Thompson, E.G. Goralski, S.D. Walck, J.T. Yates Jr., Journal of Physical Chemistry B 108 (2004) 52–57. [19] M. Batzill, E.H. Morales, U. Diebold, Physical Review Letters 96 (2006) 026103. [20] J.M. Mwabora, T. Lindgren, E. Avendano, T.F. Jaramillo, J. Lu, S.E. Lindquist, C.G. Granqvist, Journal of Physical Chemistry B 108 (2004) 20193–20198. [21] Y. Nakano, T. Morikawa, T. Ohwaki, Y. Taga, Applied Physics Letters 86 (2005) 132104. [22] C.D. Valentin, E. Finazzi, G. Pacchioni, A. Selloni, S. Livraghi, M.C. Paganini, E. Giamello, Chemical Physics 339 (2007) 44–56. [23] A. Nakahira, K. Konishi, K. Yokota, T. Honma, H. Aritai, K. Tanaka, Journal of the Ceramic Society of Japan 114 (2006) 46–50. [24] K. Onoda, S. Yoshikawa, Journal of Solid State Chemistry 180 (2007) 3425–3433. [25] Y. Mizukoshi, N. Ohtsu, S. Semboshi, N. Masahashi, Applied Catalysis B 91 (2009) 152–156. [26] A. Nakahira, K. Konishi, K. Yokota, T. Kubo, Y. Saasaki, Y. Ikuhara, Inorganic Chemistry 49 (2010) 47–51. [27] N.M. Yakovleva, L. Anicai, A.N. Yakovlev, L. Dima, E. Ya Khanina, M. Buda, E.A. Chupakhina, Thin Solid Films 416 (2002) 16–23. [28] H. Irie, Y. Watanabe, K. Hashimoto, Journal of Physical Chemistry B 170 (2003) 5483–5486.
Contents lists available at SciVerse ScienceDirect
Applied Surface Science journal homepage: www.elsevier.com/locate/apsusc
Fabrication of visible-light-responsive titanium dioxide layer on titanium using anodic oxidization in nitric acid Naofumi Ohtsu a,∗ , Hirotaka Kanno a , Shinji Komiya a , Yoshiteru Mizukoshi b , Naoya Masahashi b a b
Instrumental Analysis Center, Kitami Institute of Technology, Kitami 090-8507, Japan Institute for Materials Research, Tohoku University, Sendai 980-8577, Japan
a r t i c l e
i n f o
Article history: Received 23 October 2012 Received in revised form 12 January 2013 Accepted 13 January 2013 Available online 23 January 2013 Keywords: Photocatalysis Titanium dioxide Nitrogen-doping Anodic oxidation Visible-light response
a b s t r a c t A visible-light-responsive titanium dioxide (TiO2 ) layer was fabricated by anodizing a Ti plate in aqueous nitric acid (HNO3 ) solutions, followed by annealing at 723 K for 5 h in air. The predominant structure of the oxide layer was TiO2 with an anatase structure that contained ca. 1 at% of incorporated nitrogen. The crystallite size of the anatase TiO2 was enlarged as anodizing voltage increased, but the atomic ratio of the incorporated nitrogen was almost constant. The TiO2 layers could degrade methylene blue (MB) solution under an UV light at 370 nm, exhibiting photocatalytic activity. The activity increased monotonically as the crystallite size of the oxide increased. The TiO2 layers with their crystallite size exceeding a specific value showed excellent response under visible light at 420, 450, and 505 nm. Overall, the results of this study showed that anodic oxidization of Ti in HNO3 solution is a promising technique for fabrication of a visible-light-responsive TiO2 layer. © 2013 Published by Elsevier B.V.
1. Introduction Titanium dioxide (TiO2 ) is one of the most prominent photocatalytic oxides owing to its high photocatalytic activity, low cost, nontoxicity, and environmentally friendly nature [1]. These characteristics have resulted in TiO2 attracting widespread attention for use in environmental purification media [2,3] and antibacterial coatings [4–7], which have been utilized as air purifiers and in biomedical products. However, TiO2 is a semiconductor with wide band gaps of 3.0 eV in its rutile phase and 3.2 eV in its anatase phase; accordingly, it only acts as a photocatalyst under ultraviolet (UV) light. Natural solar light is the most powerful light source, whereas UV light accounts for only a small portion of the total energy in light. As a result, the band gap must be adjusted to the visible-light region to utilize solar light effectively for activation of TiO2 photocatalysis instead of an artificial light source. In 1986, Sato demonstrated that a visible-light response appeared in response to doping TiO2 with NOx [8]. Asahi et al. also reported that the band gap of TiO2 decreased in response to substitution of a portion of oxygen for nitrogen, resulting in the compound acting as a photocatalyst under visible light [9]. Based on these findings, researchers have employed various methods to incorporate nitrogen into TiO2 . These methods have been
∗ Corresponding author. Tel.: +81 157269563; fax: +81 157269563. E-mail address: [email protected] (N. Ohtsu). 0169-4332/$ – see front matter © 2013 Published by Elsevier B.V. http://dx.doi.org/10.1016/j.apsusc.2013.01.071
based on methods such as sol-gel synthesis [8,10–12], heating in a nitrogen-containing atmosphere [13,14,16,17] and oxidation of titanium nitride [15], or on physical processes such as ion implantation [18,19] and magnetron sputtering deposition [20,21]. The products fabricated by these methods have often shown somewhat different photocatalytic activities. It is believed that these differences originated from the location of nitrogen doping (N-doping) in the TiO2 structure; however, no general consensus has yet been reached regarding the location of nitrogen involved in photocalatytic activity. Alternative fabrication techniques might produce a N-doped TiO2 with novel photocatalytic properties [22]. Anodic oxidation is an industrial technique developed for production of a metallic oxide layer on a metallic substrate. This technique has recently also been shown to be useful for synthesis of TiO2 coatings on titanium (Ti) substrate. For example, Nakahira et al. fabricated a TiO2 coating with an anatase-type structure by anodizing Ti substrate in H3 PO4 electrolyte [23]. Similarly, Onoda et al. reported that a photocatalytic TiO2 coating with photocatalytic activity could be prepared by anodizing pre-nitrided Ti substrate in a mixed electrolyte composed of H2 SO4 , H3 PO4 , and H2 O2 [24]. The process of growing an anodic oxide layer can be explained using basic electrochemistry principles. The oxide growth is initiated by the penetration of oxygen and the production of a cation, the transport is explained by electric field-supported hopping. The electric field also attracts the anions contained in the electrolyte, which then react with the surface of the substrate. Therefore, anodic oxidization is advantageous because the oxide coating strongly adheres to the substrate and anions can be
514
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
easily incorporated into the coating by adding target anions in an electrolyte. By focusing on the latter advantage, anodic oxidation can be used to easily synthesize anion doped-TiO2 coatings via a single-step process. Mizukoshi et al. reported that sulfur-doped TiO2 coatings having rutile-type structures could be synthesized by anodizing Ti substrate in a highly concentrated sulfuric acid electrolyte [25]. The efficient photocatalytic activity of such coatings could bleach methylene blue (MB) aqueous solution under visible-light illumination at wavelengths greater than 413 nm. Based on these findings, it seems feasible that anodic oxidization in nitric acid (HNO3 ) can produce N-doped TiO2 coatings with visiblelight response. However, HNO3 acts as an oxidizing agent, thereby leading to dielectric breakdown of the TiO2 layer during anodizing. Nakahira et al. reported that, when employing 0.25-M HNO3 as an electrolyte, precipitates were produced during anodizing in exchange for degradation of the Ti plate owing to the dielectric breakdown [26]. Subsequent analysis of these precipitates revealed that they consisted of N-doped TiO2 . These findings indicate that suppression of the oxidizability of nitrate may enable the synthesis of a N-doped TiO2 layer as an alternative to TiO2 powder. In the present study, we attempted to fabricate a N-doped TiO2 coating by anodizing a Ti substrate in HNO3 with a concentration not exceeding 0.1 M. In addition, we investigated the photocatalytic activity under UV and visible-light regions from 370 nm to 505 nm. The effects of nitrogen doping and crystal structure on the photocatalytic properties were explored though careful analysis using X-ray diffractometry (XRD) and X-ray photoelectron spectroscopy (XPS).
2. Experimental procedures A pure titanium plate (99.9%) with dimensions of 20 mm × 10 mm × 1 mm was chemically polished using a colloidal silica suspension with an average particle size of 40 nm. This plate was used as a substrate. The substrate and a platinum electrode with dimensions of 20 mm × 10 mm × 1 mm were connected to the anode and cathode, respectively, after which direct current (DC) at 200 mA was supplied for 30 min in aqueous HNO3 solutions with concentrations of 1 mM, 10 mM, and 100 mM, respectively. Under this restricted current setting, the DC voltage could not be applied beyond a specific value owing to the dielectric breakdown of the oxide layer. The maximum voltages are actually 183 V, 40 V, and 22 V for 1 mM, 10 mM, and 100 mM, respectively. After the oxidation process was completed, the specimen was ultrasonically rinsed in distilled water and then annealed at 723 K for 5 h in the air. The surface morphology of the specimen was observed by scanning electron microscopy (SEM; JCM-5000 Neo Scope, JEOL) using secondary electron image mode at an acceleration voltage of 10 kV. The cross-sectional structure of the layer was observed by conventional transmission electron microscopy (TEM; JEM-2100, JEOL). The TEM specimens were prepared using an ion slicer (EM-09100IS, JEOL). X-ray diffractometry patterns (XRD; New D8 Advance, Bruker AXS) were measured with Bragg-Brentano geometry using Cu K␣ radiation. X-ray photoelectron spectroscopy (XPS; PHI 5000 Versa Probe, Ulvac-Phi) was conducted using monochromatized Al K˛ radiation (h = 1486.6 eV) to evaluate the incorporated nitrogen. The diameter of the X-ray probe was about 100 m, and the photoelectron take-off angle (TOA) was set at 65◦ , where TOA denotes the angle between the substrate surface and the spectrometer slit. An Ar-ion gun was used to analyze the nitrogen incorporated in the inner region. The photocatalytic activity was evaluated by a methylene blue (MB) degradation test. Before starting the degradation test, UV light was applied to the specimen for 90 min at an intensity of
ca. 6 mW cm−2 to eliminate the surface contaminants, after which the specimens were immersed in 10 mg L−1 of MB solution in a polypropylene vessel for 24 h to complete the adsorption of MB molecules onto the surface. UV and visible light from LED lamps consisted of 100 LEDs with wavelengths of 370 nm, 420 nm, 450 nm, or 505 nm were used to illuminate the vessel, and their radiation intensities were adjusted to 1 mW cm−2 . The photocatalytic activity was evaluated by measuring the absorbance of MB at 664 nm using the UV–vis spectrometer (UV-2400PC, Shimadzu) every 20 min during illumination for 180 min. The degradation rate of the MB was linearly plotted against the illumination periods, and reaction rate for MB degradation (mol L−1 min−1 ) were calculated from this plot.
3. Results and discussion 3.1. Morphology, structure and chemical state of the oxide layer anodized in HNO3 solution SEM images of Ti substrate anodized in 100-mM HNO3 with applied voltages of 15, 20, and 22 V are shown in Fig. 1. The image of non-treated Ti substrate is also shown in the figure. On the surface treated with 15 V, square-shaped pores on the micrometer scale with swelling at their boundaries were observed. When the voltage was increased to 20 V, the swelling disappeared and large hollows of ca. 100 m in size appeared on the surface. A collapsed surface containing winding and crossed grooves was formed when treated with 22 V. We consider that the NO3 − in the electrolyte facilitates the oxidization of Ti, thereby inducing dielectric breakdown of the oxide layer when applying the voltage beyond a specific value [27]. The hollows on the surface would be caused by the dielectric breakdown, and the collapsed surface containing grooves might be formed by continuing the breakdown during anodizing. To confirm our assumptions, we recorded the change in electric current between Ti substrate and the Pt electrode (Fig. 2). In the case of 15 V, the electric current decreased to near zero immediately after starting the oxidation, indicating that no dielectric breakdown occurred. In contrast, at 20 V, the current decreased slightly after starting, after which current with a specific value continued to flow during treatment. It is likely that the flowing current during the treatment is due to the dielectric breakdown in the oxide. Decreased current was not observed in the case of 22 V, implying that the intense dielectric breakdown occurs continuously. These changes in the electric current are direct evidence that the distinct morphology is related to dielectric breakdown occurring in the HNO3 solution. Furthermore, it should be noted that small precipitates were often found in the electrolytes after the oxidation process was completed. The precipitates were probably TiO2 produced via dielectric breakdown of the oxide layer, which was also reported by Nakahira et al. [26]. Fig. 3 shows a cross-sectional TEM image of the surface of the anodized Ti substrate corresponding to the Fig. 1(a). The TEM image shows that about 2-m thick oxide layer containing the grooves is formed upon the collapsed surface part of a Ti substrate. This indicates that the dielectric breakdown is most likely to induce the formation of the TiO2 layer. The XRD patterns of Ti substrate anodized in 100-mM HNO3 aqueous solution with various applied voltages are shown in Fig. 4. The peak observed at 25.4◦ is attributed to TiO2 with an anatase structure, indicating that the predominant phase in the oxide is anatase-type TiO2 . The peaks attributed to Ti with an hcp structure originated from the substrate. The peak intensity corresponding to the anatase TiO2 increased with increasing voltage. A similar trend in the XRD patterns was also observed on the surface anodized in 1and 10-mM HNO3 aqueous solutions with various applied voltages.
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
515
Fig. 1. SEM images of the Ti surfaces anodized in 100-mM HNO3 aquaous solution at (a) 22 V, (b) 20 V, and (c) 15 V and (d) non-treated Ti surface.
The mean crystallite sizes of anatase TiO2 were calculated using the whole powder pattern decomposition (WPPD) method based on Pawley’s algorithm (Fig. 5). The crystallite size of the anatase TiO2 increased with increasing voltage. These trends were also observed in solution with other concentrations of HNO3 . It is possible that
dielectric breakdown during anodic oxidation induces the growth of the TiO2 layer, thereby enlarging the size of the crystallite. The incorporation of nitrogen into the TiO2 layer fabricated in 100-mM HNO3 aqueous solution was examined using XPS. To obtain information from the inner region of the layer, the spectra
Fig. 2. Changes in electric current between Ti substrate and Pt electrode during anodic treatment.
516
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
Fig. 3. Cross-sectional bright field (BF) TEM image of the Ti surfaces anodized in 100-mM HNO3 aquaous solution at 22 V. Upper-inset is a high magnitude image.
were accumulated after etching the surface, ca. 300 nm, using an Ar-ion gun with an acceleration voltage of 3 kV. The nitrogen was detected in both TiO2 layers produced at the applied voltage of 15 V and 22 V, which correspond to the voltage below and at the dielectric breakdown voltage, respectively (Fig. 6). On the other hand, the nitrogen peaks were not detected from the Ti substrate before the anodization. Accordingly, the nitrogen found in the TiO2 layers is the incorporation from HNO3 aqueous solution. The observed N 1s XPS spectra are roughly decomposed into two peaks, one at approximately 399.8 eV and another at approximately 396.7 eV. The former is usually ascribed to nitrogen incorporated at a generic interstitial site, while the latter is attributed to atomic substituted nitrogen. Valentin et al. calculated the electric structure of Ndoped TiO2 using the density functional theory (DFT) method and found that both interstitial and substitutional nitrogen induce a few localized states that appear slightly above the valence band edge, accounting for the visible-light activity [22]. Furthermore,
θ Fig. 4. XRD patterns of Ti substrate anodized in 100-mM HNO3 aqueous solution with various applied voltages.
Fig. 5. Mean crystallite size of anatase TiO2 contained in the oxide layer fabricated with various applied voltages in 100-mM HNO3 solution.
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
Fig. 6. N 1s XPS spectra of the surface of anodized Ti at applied voltages of 15 V and 22 V in 100-mM HNO3 solution.
the atomic ratio of nitrogen in the TiO2 layer calculated from the spectral intensities of Ti 2p, O 1s, and N 1s XPS spectra were approximately 0.8 at%, which many researchers accept as an appropriate concentration to extend the photoactive region toward the visible-light region. Nitrogen impurities contained in commercial Ti plates are generally below 0.1 at%; therefore, the origin of nitrogen in the TiO2 layer is NO3 − in the electrolyte. 3.2. Methylene blue (MB) degradation test
visible-light illumination. On TiO2 layers anodized in the vicinity of the dielectric breakdown voltages (22 V for 100 mM, 40 V for 10 mM, and 183 V for 1 mM), a positive rate was observed in the visible-light region. Conversely, the rate was approximately zero when anodized with the voltage apart from the dielectric breakdown voltages. XPS confirmed that there were no differences in the amount and chemical state of nitrogen in the layer anodized at and below the breakdown voltage (Fig. 6). However, a significant difference in the response to visible light was found in these layers. It is likely that the photocatalytic activity in the visible-light region is not only dominated by the incorporated nitrogen but also another factor. In contrast to the rate under visible-light conditions, none of them at wavelengths below 370 nm was zero, indicating that all of the TiO2 layers have photocatalytic activity under UV light. The rate in the UV light monotonically increases with increasing applied voltage under all HNO3 concentrations. Furthermore, the rates at the dielectric breakdown voltage are almost constant, regardless of HNO3 concentration. 3.3. Relationship between photocatalytic activity and crystallinity of the TiO2 layer Here, the differences in the photocatalytic activity of the anodic TiO2 layers are explained on the basis of their crystallinity. The chemical reactions occurring on the photocatalytic TiO2 surface are triggered by the photogeneration of electron-hole pairs. The generated electrons and holes subsequently penetrate the oxide layer and react with surface species to form highly oxidizing species. Understandably, the recombination of electrons and holes during penetration should reduce the photocatalytic activity. It is generally accepted that a boundary at the interface between the two crystallites in a polycrystalline material becomes the recombination center of photogenerated electron and hole pairs because this boundary is considered to be a large crystal lattice defect.
μ
μ
μ
To evaluate the photocatalytic activity of the anodic TiO2 layer, reaction rate for MB degradation at various wavelengths were calculated. The variation in the reaction rate for MB degradation was plotted against the wavelengths of the LED light (Fig. 7). As expected, several TiO2 layers showed photocatalytic activity under
517
Fig. 7. Changes in reaction rates for MB degradation against the wavelength of LED light.
N. Ohtsu et al. / Applied Surface Science 270 (2013) 513–518
μ
518
fabricated layers consisted of polycrystalline TiO2 with an anatase structure and ca. 1 at% of incorporated nitrogen at a generic interstitial site and an atomic substituted site. As the applied voltage used for anodizing was increased, the crystallite size of the TiO2 was enlarged, while the atomic ratio of the incorporated nitrogen remained nearly constant. The voltage for the anodic treatment could not be applied beyond a specific value owing to the occurrence of a dielectric breakdown of the TiO2 layer, which occurred at 183 V, 40 V, and 22 V for 1 mM, 10 mM, and 100 mM, respectively. All layers could degrade methylene blue solution under an UV light of 370 nm, exhibiting photocatalytic activity that increased monotonically as the crystallite size increased. The photocatalytic activity under visible light at wavelengths of 420, 450, and 505 nm appeared on the TiO2 layer after the crystallite size exceeded a specific value. Acknowledgements
Fig. 8. Relationships between the reaction rate of MB and mean crystallite size of anatase TiO2 .
Consequently, polycrystalline TiO2 layers containing large crystallites are expected to show higher photocatalytic activity. In the present study, when the voltage applied for anodizing increased, both the reaction rate of MB degradation and the crystallite size of the TiO2 layer increased (Fig. 7 and Fig. 5, respectively). These results imply that the size of the crystallite is related to its photocatalytic activity. To better understand this relationship, we plotted the reaction rate of MB degradation against the crystallite sizes of the TiO2 layer, as shown in Fig. 8. Good correlation between the photocatalytic activity and the crystallite size was found when illuminating the specimen with UV light at 370 nm. Conversely, when illuminating the specimens with visible light at 420, 450 and 505 nm, a linear relationship was observed; however, this relationship was observed only when the crystallite size was greater than a specific value. In cases smaller than this specific crystallite size, the reaction rates were almost zero. Irie et al. suggested that, in cases in which TiO2−x Nx had lower values of x (<0.02), the band structure should have an isolated narrow band above the O 2p valence band [28]. UV light excites electrons in both the valence band and the narrow band, but visible light only excites electrons in the narrow band. As a result, visible-light illumination gives a lower quantitative efficiency (QY) value than UV-light illumination. Valentin et al. also reported that the hole generated in the O 2p band by visible-light illumination was less mobile than that generated by UV-light illumination [22]. The distinct relationship in the visible-light region observed in this study can be explained based on these previous reports. The electrons and holes photogenerated with visible light should be more affected by the crystallite boundary than those generated with UV light owing to the difference in the QY value and mobility. An anodic TiO2 layer with a crystallite size smaller than a specific value could not decompose MB under visible-light illumination owing to its low photocatalytic activity. Beyond the specific value, the photocatalytic activity exceeds the threshold activity to degrade MB under visible light; hence, the reaction rate increases linearly with increasing crystallite size. 4. Conclusions Nitrogen-doped TiO2 layers could be fabricated by anodizing a Ti plate in 1-, 10-, and 100-mM aqueous HNO3 solutions. The
The authors gratefully acknowledge Mr. Yamane and Mr. Tokuda from Kitami Institute of Technology for their help with the XRD and FE-SEM analyses. The authors also acknowledge Mr. Ishikawa of an undergraduate student of in our laboratory for his valuable input, and Mr. Nishioka of JEOL Ltd. for preparing the TEM specimen. References [1] A. Fujishima, T.N. Tao, D.A. Tryk, Journal of Photochemistry and Photobiology C1 (2000) 1–21. [2] R.W. Matthews, Journal of Physical Chemistry 91 (1987) 3328–3333. [3] Y. Ohko, K. Hashimoto, A. Fujishima, Journal of Physical Chemistry A 101 (1997) 8057–8062. [4] J.C. Yu, W. Ho, J. Lin, P.K. Wong, Environmental Science and Technology 37 (2003) 2296–2301. [5] Y. Ohko, Y. Utsumi, C. Niwa, T. Tatsumi, K. Kobayakawa, Y. Satoh, Y. Kubota, A. Fujishima, Journal of Biomedical Materials Research 58 (2001) 97–101. [6] Y. Yao, Y. Ohko, Y. Sekiguchi, A. Fujishima, Y. Kubota, Journal of Biomedical Materials Research 85B (2008) 453–460. [7] B. Kim, D. Kim, D. Cho, S. Cho, Chemosphere 52 (2003) 277–281. [8] S. Sato, Chemical Physics Letters 123 (1986) 126–128. [9] R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki, Y. Taga, Science 293 (2001) 269–271. [10] J.L. Gole, J.D. Stout, C. Burda, Y. Lou, X. Chen, Journal of Physical Chemistry B 108 (2004) 1230–1240. [11] S. Sakthivel, M. Janczarek, H. Kisch, Journal of Physical Chemistry B 108 (2004) 19384–19387. [12] S. Sato, R. Nakamura, S. Abe, Applied Catalysis B 284 (2005) 131–137. [13] H. Chen, A. Nambu, W. Wen, J. Graciani, Z. Zhong, J.C. Hanson, E. Fujita, J.A. Rodrigez, Journal of Physical Chemistry C 111 (2007) 1366–1372. [14] J. Yang, H. Bai, Q. Jiang, J. Lian, Thin Solid Films 516 (2008) 1736–1742. [15] T. Morikawa, R. Asahi, T. Ohwaki, K. Aoki, Y. Taga, Japanese Journal of Applied Physics 40 (2001) L561–L563. [16] O. Diwald, T.L. Thompson, T. Zubkov, E.G. Goralski, S.D. Walck, J.T. Yates Jr., Journal of Physical Chemistry B 108 (2004) 6004–6008. [17] M. Miyauchi, A. Ikezawa, H. Tobimatsu, H. Irie, K. Hashimoto, Physical Chemistry Chemical Physics 6 (2004) 865–870. [18] O. Diwald, T.L. Thompson, E.G. Goralski, S.D. Walck, J.T. Yates Jr., Journal of Physical Chemistry B 108 (2004) 52–57. [19] M. Batzill, E.H. Morales, U. Diebold, Physical Review Letters 96 (2006) 026103. [20] J.M. Mwabora, T. Lindgren, E. Avendano, T.F. Jaramillo, J. Lu, S.E. Lindquist, C.G. Granqvist, Journal of Physical Chemistry B 108 (2004) 20193–20198. [21] Y. Nakano, T. Morikawa, T. Ohwaki, Y. Taga, Applied Physics Letters 86 (2005) 132104. [22] C.D. Valentin, E. Finazzi, G. Pacchioni, A. Selloni, S. Livraghi, M.C. Paganini, E. Giamello, Chemical Physics 339 (2007) 44–56. [23] A. Nakahira, K. Konishi, K. Yokota, T. Honma, H. Aritai, K. Tanaka, Journal of the Ceramic Society of Japan 114 (2006) 46–50. [24] K. Onoda, S. Yoshikawa, Journal of Solid State Chemistry 180 (2007) 3425–3433. [25] Y. Mizukoshi, N. Ohtsu, S. Semboshi, N. Masahashi, Applied Catalysis B 91 (2009) 152–156. [26] A. Nakahira, K. Konishi, K. Yokota, T. Kubo, Y. Saasaki, Y. Ikuhara, Inorganic Chemistry 49 (2010) 47–51. [27] N.M. Yakovleva, L. Anicai, A.N. Yakovlev, L. Dima, E. Ya Khanina, M. Buda, E.A. Chupakhina, Thin Solid Films 416 (2002) 16–23. [28] H. Irie, Y. Watanabe, K. Hashimoto, Journal of Physical Chemistry B 170 (2003) 5483–5486.