Photocatalytic Performance of an Anodic TiO₂ Layer Fabricated in a NH₄NO₃/Ethylene Glycol Electrolyte with Various Crystallographic Phases

Abstract

Anodizing Ti substrates in an ammonium nitrate/ethylene glycol electrolyte is an innovative process capable of fabricating nitrogen-doped photocatalytic titanium oxide (TiO₂) layer in one step. This fabricated layer comprises both rutile and anatase TiO₂ phases; however, a major phase contributing to its excellent visible-light responsive photocatalytic performance is still unknown. In the present work, the crystallographic phase of an anodic layer was controlled by exploiting a post-thermal treatment and relationship between the phase variation and photocatalytic performance was then investigated to determine the major phase contributing to this performance. Post-thermal treatment to the anodic TiO₂ layer had little influence on the surface morphology and nitrogen doping, but the crystallographic phase, more specifically the ratio of anatase to rutile phases, changed with the heating temperature. The photocatalytic activity, evaluated by methylene blue decolorization and ethylene decomposition, increased with an increase in the ratio of anatase phase, while the correlation with the rutile phase was not observed. X-ray diffraction (XRD) analysis using a grazing incidence geometry showed that the anatase phase was concentrated in the topmost surface region when compared with the rutile phase. In conclusion, the variation of the photocatalytic performance was related to the growth of the anatase TiO₂ phase in the layer, with the treatment temperature of 723 K showing the highest photocatalytic activity.

Fig. 3 N 1s XPS spectra obtained from the sample surfaces treated at 523 K, 723 K, and 923 K.

1. Introduction

Anodic oxidation is an electrochemical process commonly used in industries for producing highly adhesive oxide layers on a metallic substrate. The use of titanium (Ti) as a substrate leads to the formation of a titanium dioxide (TiO₂) layer, which can improve its surface properties such as decorativeness^1,2) and biocompatibility.^3,4) In recent years, the TiO₂ layer is used as a photocatalyst for pollution control, due to its strong oxidative capability under light illumination,^5,6) and anodic oxidation of Ti substrate has gained attention as a fabrication process of photocatalytic TiO₂. The photocatalytic properties of an intact TiO₂ layer are only activated by ultraviolet (UV) light.^7,8) Rendering the properties of TiO₂ to lower the band gap energy would permit visible-light photocatalytic activation, and this can be effectively achieved using nitrogen doping.⁹⁾ However, as a characteristic feature of the anodic process, the TiO₂ layer fabricated by anodization includes the impurities originated from the electrolyte used. Therefore, anodization of Ti substrate in an electrolyte containing nitric ions can be a suitable process for the production of visible-light responsive photocatalytic layers.

Our research group has demonstrated that a high-performance visible-light responsive nitrogen-doped TiO₂ layer can be fabricated via anodic oxidation of a Ti substrate using a solution containing nitric ions.¹⁰⁾ In the early stages of our research, we employed nitric acid (HNO₃) as the electrolyte. The layer fabricated in HNO₃ was an amorphous TiO₂ layer containing about 1 at% of the incorporated nitrogen, and the subsequent thermal treatment at 723 K could crystallize the oxide layer to anatase-type TiO₂. As expected, excellent photocatalytic activity under visible light illumination was confirmed by a methylene-blue decolorization test. However, the resultant layer was not homogenous and included deposits of granular-like TiO₂ precipitates on the surface. In addition, the deposited TiO₂ precipitates could be easily removed by wiping the surface gently. We presumed that an aqueous solution containing nitric ions induces dielectric breakdown on the Ti surface during anodization, thereby forming TiO₂ precipitates. Consequently, anodizing in an aqueous nitric solution was considered unsuitable for producing a practical photocatalytic layer.

We also found that the characteristics of the anodized surface showed a drastic change when ethylene glycol was used as the electrolyte solvent instead of water.^11,12) The surface anodized in nitrate/ethylene glycol electrolyte could form a homogeneous oxide layer without any precipitate, and this oxide layer acted as a photocatalyst under both UV-light and visible light illumination after undergoing the post-thermal treatment. The layer with the post-thermal treatment contained both rutile and anatase phases of TiO₂, and showed excellent antibacterial performance under both types of illumination discussed above. Therefore, we have applied the anodization process in nitrate/ethylene glycol electrolyte as an innovative technique to form a visible light responsive nitrogen-doped TiO₂ layer in this research project.

The aim of the next stage of our research project is to improve the photocatalytic performance of the anodized TiO₂ layer, especially by identifying the phase of TiO₂ that contributes predominantly to the photocatalytic reactions. Elucidating the major phase contributing to the photocatalytic performance provides valuable clues on how to optimize the treatment settings for producing the photocatalytic layer with the highest performance. In the present study, we control the crystallographic phase of the anodic layer by varying the temperatures of the post-thermal treatment, and then investigate the relationship between the photocatalytic performance and crystallographic phases in detail. Furthermore, we study the depth distribution of the crystallographic phases using X-ray diffraction (XRD) with grazing incidence geometry. By analyzing the results, we discuss the major phase contributing to the photocatalytic performance of the TiO₂ layer fabricated via anodization in nitrate/ethylene glycol electrolyte.

2. Materials and Methods

2.1 Preparation of the anodic layer

A Ti plate (99.5%, 10 × 20 × 1 mm³) was chemically polished using a colloidal silica suspension and was then sonicated in ethanol for 10 min. The samples were prepared by anodizing the Ti plate in an electrolyte solution using a platinum plate as the counter electrode. The electrolyte solution comprised industrial grade ethylene glycol (99.5%) containing 0.1 M NH₄NO₃.¹¹⁾ A constant direct current (DC) of 200 mA was applied, and the voltages were varied accordingly. The anodization was conducted for 60 min, after which the sample was sonicated in ethanol for 10 min. The samples were then thermally treated at 473, 523, 573, 723, 823, and 923 K using a conventional electric furnace in air (FO100, Yamato Scientific, Japan), in order to control the crystallographic structure. The temperature of the furnace was gradually increased to the target temperature within 2 h, and then maintained at constant for 5 h. After the thermal treatment, the samples were left to equilibrate inside the furnace to the room temperature (ca. 298 K).

2.2 Characterization of anodic layers

The surface morphology of the samples was imaged using scanning probe microscopy (SPM; SPM-9700, Shimadzu, Japan) in the contact mode. The surface layer crystallography was determined by XRD (New D8 ADVANCE, Bruker AXS, Germany) using Bragg-Brentano geometry and grazing incidence geometry (2° and 5°) with Cu Kα radiation. Elemental analysis of nitrogen was performed using X-ray photoelectron spectroscopy (XPS; PHI 5000 Versa Probe, Ulvac-Phi, Japan) using monochromatic Al Kα radiation (1486.6 eV).

2.3 Evaluations of the photocatalytic activity

The photocatalytic activity of the samples was evaluated through two separate tests, namely, decolorization of methylene blue (MB) dye, and gas decomposition of ethylene gas. The decolorization of MB dye was conducted in an aqueous MB solution at an ambient temperature. The samples were immersed in 10 mg·L⁻¹ of MB solution in a polypropylene vessel for 24 h in a darkroom, to allow adsorption of MB on the surface. Before testing, the MB solution was replenished and then the vessel was illuminated with UV (370 nm) and visible light (420 nm) using a light emitting diode (LED) with an irradiance of 1 mW·cm⁻¹. The photocatalytic activity was evaluated by measuring the absorbance of MB at 664 nm using a UV-vis spectrometer (UV-2400PC, Shimadzu) in 20 min intervals for 180 min. The bleaching rate of MB was plotted against the duration of radiation, and the reaction rates for MB bleaching (min⁻¹·mL⁻¹) were calculated from the slope of its linear plot.¹²⁾

The gas decomposition test was carried out in an enclosed cylindrical glass vial (100 mL). The sample was placed in the cylindrical glass vial, which was then sealed with a rubber plug. Subsequently, 1 mL of ethylene (gas) was injected into the vial, after which it was placed in a darkroom for 24 h, to allow adsorption of ethylene on the surface. After 24 h, the sample was transferred into a new vial and was sealed, followed by the injection of 1 mL of ethylene (gas) into the vial. The vial was illuminated under UV light (370 nm) and visible light (420 nm) using a LED with an irradiance of 1 mW·cm⁻¹ for 24 h. The gas filled in the vial was withdrawn using a gas syringe after the illumination, while the ethylene concentration was determined using a gas chromatograph equipped with a flame ionization detector (GC-4000, GL sciences Inc, Japan). The decomposition rate of ethylene gas was calculated by the following expression:   

\begin{equation*} \eta = (c_{0} - c_{t})/c_{0}\times 100\% \end{equation*}

where η is the decomposition rate, C₀ and C_t are the initial concentration and the concentration at time (t) of ethylene, respectively.

3. Results

3.1 Crystallographic phases of the anodic layer subjected to the post thermal treatment at various temperatures

To confirm the controllability of the crystallographic phases by the post thermal treatment, XRD patterns of the sample surface with various temperatures were measured (Fig. 1). The pattern obtained from an untreated sample surface is depicted in Fig. 1 for comparison. It is noted here that the patterns were obtained using Bragg-Brentano geometry. The crystallography of TiO₂ was found to vary with the heating temperature. For the untreated sample, a very broad peak was observed at 27.4°, which was attributed to the rutile TiO₂ phase.¹²⁾ The 473 K sample had a distinct peak at 25.3°, which is characteristic of the anatase TiO₂ phase. The peak intensity of the anatase phase increased with increasing temperature until 723 K; however, the intensity decreased when heated to 923 K. In the temperatures between 473 and 573 K, a broad peak at 27.4° was observed corresponding to the rutile TiO₂ phase. There was little difference in the rutile peak intensity in the 473 to 573 K temperature range. However, at the temperature range between 723 and 923 K, the intensity of the rutile peak increased drastically and the width of the peak became narrower. In addition, the intensity of Ti substrate peak, observed at 38.5°, decreased abruptly when the temperature rose over 823 K.

Fig. 1

XRD patterns of the samples without and with the thermal treatment at 473, 523, 573, 723, 823, and 923 K.

To summarize the XRD results, a part of the oxide layer was found to crystallize gradually to anatase TiO₂ with rising temperatures until 723 K, while the crystal structure of the untreated sample was almost amorphous. In the range beyond 823 K, an abrupt enhancement of the rutile TiO₂ phase was observed, and simultaneously, the peak intensity of the Ti substrate decreased. The decreased intensity for Ti corresponded to the increase in the thickness of the oxide layer, indicating that thermal oxidation had occurred due to the diffusion of oxygen in air.^13,14) The crystallographic phase of the anodic layer, especially the ratio of anatase to rutile phases, could be controlled by varying the temperature of the post-thermal treatment, however, the rutile phase formed at a higher temperature was the oxide layer thermally formed by the diffusion of oxygen from air.

3.2 Effect of the treatment temperature on surface morphology and nitrogen doping

Changes in the specific surface area and the amount of nitrogen doping can be expected to affect the photocatalytic performance of the TiO₂ layer. To explore this, topological microstructures of the surface of the samples heated at various temperatures were investigated by SPM (Fig. 2). The SPM images illustrate that the microstructure was independent of the thermal treatment temperature. The average roughness (Ra), estimated from the images, gradually increased with increasing temperature. The untreated sample had an average roughness of 21.3 nm (Fig. 2(a)), whereas samples treated at 523, 723, and 923 K had an average surface roughness of 24.2, 27.5, and 63.0 nm, respectively (Fig. 2(b)–(d)). It was hypothesized that the increase in roughness at 923 K was due to the development of an oxide layer on the surface, caused by thermal diffusion of oxygen. However, in the range below 723 K, the roughness was almost similar, indicating the effect of topological change caused by thermal treatment on photocatalytic performance was negligible.

Fig. 2

SPM images of the anodized Ti surface for (a) untreated, (b) treated at 523 K, (c) at 723 K, and (d) at 923 K.

The nitrogen concentration and its chemical states in the surface layer were investigated using XPS. The high-resolution photoelectron spectra of N 1s obtained from the surface heated at various temperatures are shown in Fig. 3. Here, the photoelectron spectra were measured from the surface after etching for ca. 10 nm using an Ar-sputtering system. The N 1s spectra could be deconvoluted into two chemical states; the peak located at 396.7 eV corresponded to atomically substituted nitrogen, and the peak located at 399.9 eV for interstitial nitrogen.¹⁵⁾ At 523 and 723 K, the peaks corresponding to both atomically substituted and interstitial nitrogen were observed. However, at higher temperature (923 K), the peak corresponding to the substituted nitrogen was absent. It has been reported that the Ti–N bond (corresponding to the substituted nitrogen) is much weaker than Ti–O bond, and consequently the substituted nitrogen can be displaced by thermally diffusing oxygen.¹⁶⁾ The atomic ratio of nitrogen in the TiO₂ layer calculated from the spectral intensities of Ti (2p), O (1s), and N (1s), was found to be 0.4 at%, and independent of the thermal treatment temperature. Previous literature has highlighted that the incorporation of nitrogen below ∼1 at% can extend the photoactivity into the visible light region.¹⁷⁾ However, in this study, notable changes in the nitrogen doping depending on the temperatures of thermal treatment were hardly observed.

Fig. 3

N 1s XPS spectra obtained from the sample surfaces treated at 523 K, 723 K, and 923 K.

3.3 Photocatalytic activity of the anodic layer with the post thermal treatment

Figure 4 shows the photocatalytic activity of the surface layer under UV and visible light illumination, evaluated by the decolorization of MB aqueous solution. It can be seen that the photocatalytic activity of the surface layer increased upon increasing the temperature until 723 K, after which the activity decreased. No differences in the specific surface area and nitrogen doping with the heating temperature were found as seen in the previous section; accordingly, the increase in the degradation rate was considered to reflect the difference in the crystallographic phases.

Fig. 4

Photocatalytic activity of the samples without and with the thermal treatment at 473, 523, 573, 723, 823, and 923 K, evaluated by the methylene blue (MB) decolorization test under 370 nm or 420 nm LED light irradiation. Each value represents the mean ± S.D.; n = 3.

To further investigate the effect of thermal treatment temperature on the photocatalytic activity, a gas phase decomposition experiment was conducted with ethylene. Figure 5 shows the decomposition rate of ethylene gas for various samples after irradiating UV and visible lights for 24 h. The decomposition rate on the surface layer increased when increasing the temperature up to 723 K and decreased when the thermal treatment was above 723 K. These results are consistent with the results for the MB test. The highest decomposition rate was obtained for the surface layer treated at 723 K for both the UV and visible light; similar to the case for the MB test (Fig. 4).

Fig. 5

Decomposition rates of ethylene gas by the photocatalytic reaction of the samples without and with treatment temperatures of 473, 523, 573, 723, 823, and 923 K. LED lights with wavelengths of 370 and 420 nm were used as light sources. Each value represents the mean ± S.D.; n = 3.

4. Discussion

To discuss the crystallographic influence on photocatalytic activity, the XRD peak intensities corresponding to the anatase TiO₂ phase (101) and rutile TiO₂ phase (110) are plotted against the photocatalytic activity determined by the MB test (Fig. 6). For the anatase TiO₂ phase, a linear correlation was found between the activity and peak intensity of anatase (101), for both UV and visible light (Fig. 6(a)–(b)). However, there was no linear correlation between the rutile phase (110) peak intensity and photocatalytic activity (Fig. 6(c)–(d)). A similar plot was obtained for the ethylene gas decomposition rate (Fig. 7). A linear correlation between the decomposition rate and the peak intensity of anatase was found for both UV and visible light illuminations, while no correlation was found in the plot of rutile phase. The decolorization of methylene blue is catalyzed by a reduction reaction from methylene blue to leucomethylene blue,¹⁸⁾ whereas the ethylene decomposition is an oxidative reaction.¹⁹⁾ These results indicate that the major phase contributing to photocatalytic reaction in the anodic layer fabricated in NH₄NO₃/ethylene glycol electrolyte is the TiO₂ with anatase phase, and the concentration of anatase TiO₂ included in the layer relates to the photocatalytic performance.

Fig. 6

Correlation between the XRD peak intensity of TiO₂ and the photocatalytic degradation rate of the MB. The peak intensities of TiO₂ were estimated from anatase TiO₂ (101) and rutile TiO₂ (110) peaks. The photocatalytic degradation was calculated for irradiations at 370 and 420 nm. For (a) anatase vs. 370 nm, (b) anatase vs. 420 nm, and (c) rutile vs. 370 nm, and (d) rutile vs. 420 nm.

Fig. 7

Correlation between the XRD peak intensity of TiO₂ and the photocatalytic decomposition rate of the ethylene. The peak intensities of TiO₂ were estimated from anatase TiO₂ (101) and rutile TiO₂ (110) peaks. The photocatalytic decomposition was calculated for irradiance at 370 and 420 nm. For (a) anatase vs. 370 nm, (b) anatase vs. 420 nm, and (c) rutile vs. 370 nm, and (d) rutile vs. 420 nm.

To investigate the role of the anatase TiO₂ in photocatalytic reaction, high-resolution XRD patterns of the TiO₂ phases were measured using a grazing incident angle of 2° and 5° (Fig. 8(a)–(c)). Here, the intensities of the patterns were normalized against the intensity of the rutile TiO₂ phase. It was found that the intensity ratio of the anatase to rutile phase decreased with increasing incident angle (from 2° to 5°), for thermal treatment temperatures of 523, 723, and 923 K (Fig. 8(a)–(c)). This result implies that the anatase TiO₂ phase was segregated in the topmost surface. Researchers have come to the consensus that the anatase TiO₂ phase has enhanced the photocatalytic activity due to its lower recombination rate of photogenerated electrons and holes.^20–22) Furthermore, the chemical substance in the topmost surface is found to affect catalytic reactions effectively. Therefore, it was concluded that the anatase TiO₂ phase was essential for the photocatalytic activity of anodic TiO₂ layer fabricated in NH₄NO₃/ethylene glycol electrolyte.

Fig. 8

XRD patterns with grazing incidence angles obtained from the sample surfaces for (a) treated at 523 K, (b) at 723 K, and (c) at 923 K.

5. Conclusion

A visible-light-responsive photocatalytic titanium dioxide (TiO₂) layer was prepared by the anodization of a titanium substrate in an electrolyte comprising ammonium nitrate in ethylene glycol, followed by thermal treatment at various temperatures ranging from 473 K to 923 K in air. Thereafter, their characteristics and photocatalytic performances were evaluated, in order to elucidate the relationship between the photocatalytic performances and crystallographic phases. Until 723 K, the ratio of anatase to rutile TiO₂ increased with increasing temperature, but decreased at temperatures over 723 K because growth of oxide layer due to thermal diffusion became predominant effect. In spite of the variations in crystallography, changes in the specific surface area and nitrogen concentration of the anodic TiO₂ layer were trivial and did not affect the photocatalytic performance. Photocatalytic activities under both UV and visible light, evaluated by the methylene blue decolorization test and ethylene gas decomposition test, increased with increasing ratio of anatase to rutile included in the fabricated layer. XRD analysis using grazing incidence geometry revealed that the anatase phase segregated to the topmost surface, which supports the fact that the anatase phase has a dominant contribution to photocatalytic reaction. Among the temperatures used in this study, optimum catalytic activity was achieved at the thermal treatment temperature of 723 K under both UV and visible light illumination.

Acknowledgements

The authors would like to acknowledge Mr. Yamane from the Kitami Institute of Technology for his assistance with the XRD and XPS analysis. Part of this work was supported by the Adaptable and Seamless Technology Transfer Program through Target-driven R&D, Japan Science and Technology Agency (JST).

REFERENCES

• 1)   S.V.  Gilsa,  P.  Masta,  E.  Stijnsb and  H.  Terryna: Surf. Coat. Tech. 185 (2004) 303–310.
• 2)   M.V.  Diamanti,  B.D.  Curto and  M.P.  Pedeferri: Color Res. Appl. 33 (2008) 221–228.
• 3)   H.  Ishizawa and  M.  Ogino: J. Biomed. Mater. Res. 29 (1995) 65–72.
• 4)   L.H.  Li,  Y.M.  Kong,  H.W.  Kim,  Y.W.  Kim,  S.J.  Seo and  J.Y.  Koak: Biomaterials 25 (2004) 2867–2875.
• 5)   B.  Tryba,  A.W.  Morawski and  M.  Inagaki: Appl. Catal. B 41 (2003) 427–433.
• 6)   C.  Hu,  J.C.  Yu,  Z.  Hao and  P.K.  Wong: Appl. Catal. B 46 (2003) 35–47.
• 7)   N.  Ohtsu,  S.  Komiya and  K.  Kodama: Thin Solid Films 534 (2013) 70–75.
• 8)   C.S.  Turchi and  D.F.  Ollis: J. Catal. 122 (1990) 178–192.
• 9)   S.  Sato: Chem. Phys. Lett. 123 (1986) 126–128.
• 10)   N.  Ohtsu,  H.  Kanno,  S.  Komiya,  Y.  Mizukoshi and  N.  Masahashi: Appl. Surf. Sci. 270 (2013) 513–518.
• 11)   N.  Ohtsu,  K.  Yokoi and  A.  Saito: Surf. Coat. Tech. 262 (2015) 97–102.
• 12)   N.  Ohtsu and  K.  Yokoi: Surf. Coat. Tech. 294 (2016) 109–114.
• 13)   J.G.  Yu,  H.G.  Yu,  B.  Cheng,  X.J.  Zhao,  J.C.  Yu and  W.K.  Ho: J. Phys. Chem. B 107 (2003) 13871–13879.
• 14)   J.G.  Yu,  J.F.  Xiong,  B.  Cheng and  S.W.  Liu: Appl. Catal. B 60 (2005) 211–221.
• 15)   R.  Asahi,  T.  Morikawa,  T.  Ohwaki,  K.  Aoki and  Y.  Taga: Science 293 (2001) 269–271.
• 16)   Z.P.  Wang,  W.M.  Cai,  X.T.  Hong,  X.L.  Zhao,  F.  Xu and  C.G.  Cai: Appl. Catal. B 57 (2005) 223–231.
• 17)   H.M.  Yates,  M.G.  Nolan,  D.W.  Sheel and  M.E.  Pemble: J. Photochem. Photobiol. A 179 (2006) 213–223.
• 18)   O.  David,  M.  Andrew and  L.  Katherine: Appl. Catal. B 184 (2016) 201–207.
• 19)   S.  Yamazaki,  S.  Tanaka and  H.  Tsukamoto: J. Photochem. Photobiol. A 121 (1999) 55–61.
• 20)   L.  Sopyan,  S.  Murasawa,  K.  Hashimoto and  A.  Fujishima: Chem. Lett. 23 (1994) 723–726.
• 21)   M.  Liu,  H.M.  Lia,  Y.S.  Zeng and  T.C.  Huang: Appl. Surf. Sci. 274 (2013) 117–123.
• 22)   Z.Y.  Wang,  K.L.  Lv,  G.H.  Wang,  K.J.  Deng and  D.G.  Tang: Appl. Catal. B 100 (2010) 378–385.