Application of Nanostructured Tungsten Fabricated by Helium Plasma Irradiation for Photoinduced Decolorization of Methylene Blue (cid:3)

The dendritic nanostructure was fabricated on a surface of tungsten plate by helium plasma irradiation. The nanostructure consisting of W metal was partially oxidized to form WO 3 on exposure to air, and the resulting surface exhibited a broad photoabsorption in the range from 1 to 5 eV. We examined photoinduced reaction of methylene blue (MB) on the material. It was found that the partially and fully oxidized surface nanostructures were able to promote a decolorization of MB under photoirradiation even with the near-infrared light ( < 1 : 55 eV), whose energy is lower than the band gap of WO 3 . The reaction rate was varied with the fraction of W and WO 3 on the surface layers, that is, the partially oxidized sample promoted the reaction at a higher rate than the fully oxidized one. It is also found that the reaction rate decreased with time, which would be caused by the products accumulation on the surface and the surface oxidation. [DOI: 10.1380/ejssnt.2014.343] Keywords:


I. INTRODUCTION
Various kinds of organic pollutants spreading over our environment have become serious problem. As one of the methods for environmental cleanup, photocatalytic degradation of organic compounds has been attracted much attention because it can utilize abundant solar energy and it can mineralize the organic compounds. It is explained that photoexcited electron and hole pairs generated in a semiconductor photocatalyst have high potentials enough to cause reductive and oxidative reactions with adsorbed molecules, respectively. Since the Honda-Fujishima effect reported in 1972 [1], photocatalysis has been widely studied and then its environmental application was extensively carried out over the world [2], where most of photocatalysts such as titanium oxide (TiO 2 ) only can function with UV light. However, since the main part in the wavelength distribution of sunlight and interior illuminations are visible light, photocatalysts that can utilize visible light has been desired for more efficient use. Tungsten oxide (WO 3 ) has been reported as such photocatalyst that can utilize visible light [3][4][5].
Generally the nanosized semiconductor photocatalyst is expected to have higher activity due to the variation of band structure while the higher energy of the light is required for the photoexcitation [6,7]. Recently, it has been noticed that surface plasmon resonance (SPR) of metal nanoparticles deposited on a semiconductor material promotes some photocatalytic reactions with visible light irradiation [8][9][10][11][12][13][14], because electrons excited by SPR transfer from the metal nanoparticles into the conduction band of the attached semiconductor.
In the field of nuclear fusion, it has been reported that the dendritic nanostructure is fabricated on a tungsten (W) plate by helium (He) plasma irradiation in certain conditions [15,16]. Helium plasma generates thermal vacancies and He bubbles, conclusively the dendritic nanostructure on the W surface. In this study, we examined to fabricate a nanostructured photoactive material with the metal-semiconductor interface prepared by the He plasma technique followed by gentle oxidation, which is expected to consist of both nanosized W and nanosized WO 3 , and the nanosized WO 3 might show photoactivity by the aid of the SPR of neighboring tungsten metal nanostructure under less energy light irradiation than band gap of WO 3 .

A. Sample Preparation
The dendritic nanostructured surface was fabricated by the He plasma irradiation to a monocrystal W plate (The Nilaco Corporation, 0.1 × 8 × 8 mm 3 , 99.95% purity) in the linear type diverter plasma simulation test device NAGDIS [17,18]. The incident energy, ion fluence and flux of the helium plasma were 50-60 eV, ∼ 5 × 10 25 m −2 and ∼ 2 × 10 22 m −2 s −1 , respectively. The surface temperature of the sample in steady state was measured as 1500 K with a radiation pyrometer. After the plasma irradiation, the surface of the sample was partially oxidized on exposure to air at room temperature. This oxidation process took a long time: we prepared three samples in this manner through the oxidation for 12, 97 and 200 days, and the surface oxidation ratios (x) of these samples estimated by XPS measurements were 20, 35 and 60%, respectively. Another sample with the surface oxidation ratio of x = 100% was prepared by further heat-treated at 773 K for 2 h in air. These samples are referred to as WO 3 (x)/W.

B. Reaction Experiment under Photoirradiation of NIR
Photoinduced reactions of methylene blue (MB, C 16 H 18 N 3 SCl), an organic dye compound, were conducted to evaluate the activities of the samples in the reported manner [18]. A typical reaction experiment consisted of placing the dendritic sample and 2 ml of MB aqueous solution (10 µmol/L) into a quartz cell, subsequently exposing the sample to near infrared light using a 300 W extra high pressure Xe lamp (USHIOSPAX R300-3J) with a cut-off filter permitting λ > 800 nm. The light absorbance of the MB aqueous solution at λ = 664 nm was measured by transmission method in UV-vis spectroscopy with a spectrometer JASCO V-670. The MB concentration in each reaction time was estimated from the absorption peak intensity. We evaluated the photoinduced reaction activity from the decrease rate of the MB concentration.

C. Characterization of the Samples
X-ray photoelectron spectroscopy (XPS) measurements were carried out for the samples before and after the MB decomposition reactions with a ESCA-3300 (SHI-MADZU) spectrometer with 10 mA −12 V Al Kα emission as the X-ray source. The atomic ratios for carbon, oxygen and tungsten in each sample's surface were estimated from the peak areas of C1s, O1s and W4f regions. Energy correction of the spectra was performed using C1s peak (284.6 eV) originated in the contaminants.  The dendritic W materials were observed by a field emission scanning electron microscope (FE-SEM, HI-TACHI S-4300) and an ultra-high voltage transmission electron microscope (TEM, JEOL JEM-1000K RS).
X-ray absorption fine structure (XAFS) measurements of the photocatalysts were carried out at the beam line 9C at Photon Factory of High Energy Accelerator Research Organization Institute of Materials Structure Science, using a two-crystal Si(111) monochromator. The W L 3 -edge XAFS spectra were measured in transmission yield method at room temperature in atmosphere. We measured the spectrum of dendritic W materials by exfoliating with a cellophane tape. The intensity of the incident X-ray was measured with a gas-flow type 17 cm ion chamber using a mixture of N 2 and Ar at 85:15. The intensity of the transmission X-ray was measured with a gas-flow type 31 cm ion chamber using Ar gas.
Diffuse reflectance UV-visible-NIR spectra were recorded by a spectrophotometer (JASCO V-570) equipped with an integrating sphere. Figure 1(a) shows a SEM image of the dendritic WO 3 (60)/W sample. The surface of the sample was covered with a submicrometer fine structure, corresponding to the dendritic nanostructure. In a TEM image of the piece of the sample shown as Fig. 1(b), the He bubbles were clearly observed in the dendritic structure, and it was observed that the dendritic surface was not fully but partly covered with oxidized moiety, where the WO 3 /W interface existed on the surface. The WO 3 film thickness was estimated to be less than 5 nm. Figure 2 shows the Fourier transforms (FT) of the W L 3 -edge EXAFS of the WO 3 (60)/W sample and a tungsten foil. The spectrum of the WO 3 (60)/W sample was similar to that of a W foil. The large peaks observed at ∼ 2.7Å and ∼ 3.2Å correspond to the first-and secondshells, respectively, which are assignable to the W atoms. We also performed nonlinear least-square curve-fitting to the Fourier-filtered EXAFS including the first and the  second coordination shells. In the dendritic WO 3 (60)/W sample, the average distances and coordination numbers of first neighboring W shell were estimated to be 2.75Å and 8.0, respectively, while those of second neighboring W shell were estimated to be 3.18Å and 6.4, respectively. As shown in Table I, these values are approximately same to those of a W foil, indicating that the local structure of a tungsten atom is fundamentally maintained after the He plasma irradiation. In this analysis, no oxidized phase was found, which means that the oxidized moiety is limited to only a part of the surface, supporting the TEM observation ( Fig. 1(b)). Figure 3 shows the diffuse reflectance UV-visible-NIR spectra of the WO 3 (100)/W sample, the WO 3 (60)/W samples and the pristine W plate before the He plasma irradiation. The plasma-irradiated samples exhibited higher absorption at the entire range ( Fig. 3(a) and (b)), which is consistent with the black color of these samples. Such broad absorption bands would originate from the nanostructured metal W, which is expected to absorb visible and/or near infrared light by SPR on the W nanostructured surface. The WO 3 (100)/W sample showed partly less absorbance in the energy range of 2-4 eV. Although the reason was unclarified, it should be due to the structural variation with heating at 773 K in air. The aggregation of the dendritic structure by heating might reduce the absorbance due to the SPR by nanostructured W, or, the further oxidation might change the surface chemical states from W to WO 3 to reduce the absorbance around 2-4 eV.

III. RESULTS AND DISCUSSION
The results of photoinduced reactions for each sample under near infrared light irradiation are shown in Fig. 4. In the separate experiments, we have also con- firmed that the samples show photoinduced reaction activity under photoirradiation of both UV light and visible light simultaneously. As shown in Fig. 4, the photoinduced reaction rate over the WO 3 (100)/W sample is higher than that without any sample. The WO 3 species on the WO 3 (100)/W sample would be one of the active sites, which consisted with literature [19,20]. The reaction rates of the partially oxidized samples depended on the surface oxide ratio, i.e., it increased in the order 20, 35 and 60% to the maximum, followed by decreasing in 100%. Since the WO 3 (60)/W sample would have more WO 3 /W interface sites than the other samples, it is suggested that the WO 3 /W interface on the nanostruc- tured surface should be more active sites for the photoinduced reaction than the fully oxidized surface of the WO 3 (100)/W sample. Note that the energy of the irradiation light (< 1.6 eV) was less than the bandgap of WO 3 (> 2.7 eV). Thus, this photoinduced reaction should take place with definitely different mechanism from that on the semiconductor photocatalyst as mentioned above. The photoinduced reaction mechanism in the dendritic WO 3 /W sample is speculated as the following process; (i) light absorption due to SPR on the nanosized W would occur under the near infrared light irradiation, (ii) the plasmon excitation would make electrons inject into the conduction band of the surface nanosized WO 3 moiety at the WO 3 /W interface, and finally (iii) the electrons at the WO 3 surface would react with the adsorbed MB molecules. The electrons closed to the interface would contribute to the reaction more readily, because the long traveling might cause deactivation of the excited electrons. Thus, it is suggested that the WO 3 /W interface would be the most important active sites for the photoinduced reaction. Figure 5 shows the time courses of the concentration of MB under the light and in the dark with the WO 3 (60)/W sample and the WO 3 (100)/W sample. In the dark, the MB concentration very slowly decreased with time in the presence of the dendritic WO 3 (60)/W sample, where the MB molecules would be gradually adsorbed and the adsorption density would increase on the nanostructural surface of the WO 3 (60)/W sample. On the other hand, the WO 3 (100)/W sample was adsorbed by only small amount of the MB molecules, suggesting that the oxidation treatment on the dendritic sample would reduce the specific surface area or the adsorption ability. Upon light irradiation, both samples reduced the MB concentration at higher rate, confirming that photoinduced reaction would occur on the samples. Since the separate experiments elucidated that the commercially obtained WO 3 powder and the WO 3 thin film on the tungsten plate showed no light-responsiveness, it is notable that the photoinduced activity would be unique property for the partially or fully oxidized WO 3 nanostructure on the dendritic nanostructure of tungsten.
When the photoinduced reaction test was repeated on the same sample (Fig. 6), the remarkable decrease of the reaction rate in the second run was observed for the WO 3 (60)/W sample and less deactivation was observed for the WO 3 (100)/W sample. Thus, it is considered that the active sites at the WO 3 /W interface on the WO 3 (60)/W sample would be much deactivated and the active sites on the WO 3 (100)/W sample would be less deactivated. Table II summarizes the atomic ratios of carbon, oxygen and tungsten on the surfaces of the two samples before and after the reaction, which were determined by XPS measurements. The carbon before the reaction would be the surface contamination while that after the reaction might include the adsorbed organic or carbonaceous species. The ratio of carbon on the WO 3 (100)/W sample did not vary remarkably before and after the reaction. On the other hand, the ratio of carbon on the WO 3 (60)/W sample increased after the reaction, suggesting the presence of the adsorbed products originating from the MB molecules on the surface. Figure 7(a) shows the XPS spectra of the WO 3 (60)/W sample before and after the reaction. Before the reaction, the W 4f 7/2 and 4f 5/2 peaks at 31.6 and 33.6 eV and those at 35.5 and 37.6 eV originated from W(0) and W 6+ , respectively, are recorded, suggesting that W(0) and W 6+ species coexist on the surface of the WO 3 (60)/W sample. However, after the reaction, W(0) peaks at 31.6 and 33.6 eV disappear and only W 6+ peaks at 35.5 and 37.6 eV are observed. On the other hand, XPS spectra of WO 3 (100)/W sample show only W 6+ species on the surface before and after the reaction (Fig. 7(b)). As shown in Table II, almost all of the W (0) species on the surface of the WO 3 (60)/W sample change to W 6+ species after the reaction, which means the amount of WO 3 /W interface decreased on the WO 3 (60)/W sample. As shown in the SEM images, the dendritic nanostructure thickened on the WO 3 (60)/W sample after the reaction (Fig. 8(a)  and 8(b)), while those on the WO 3 (100)/W sample exhibited no significant change (Fig. 8(c) and 8(d)). From these results, it is proposed that the obvious deterioration of the WO 3 (60)/W sample would be caused by the following reasons: (i) the oxidation of the surface during the reaction would reduce the active sites at the WO 3 /W interface, and (ii) the accumulation of the reaction products on the active sites.

IV. CONCLUSIONS
We prepared the WO 3 nanolayers partially covering the surface of the unique dendritic W nanostructure by the He plasma irradiation technique followed by gentle oxidation on exposure to air for a long period, and confirmed the activity for the photoinduced reaction of methylene blue (MB) as an organic dye. The photoinduced reaction was conducted under irradiation of near infrared light whose energy is lower than band gap of WO 3 . Moreover, we found that both the WO 3 /W interface and WO 3 surface would be the active sites for the photoinduced reaction and the former was more active than the latter. The mechanism of the photoinduced reaction is proposed as follows: (i) the initial photoexcitation due to SPR on the nanosized W would occur under the near infrared light irradiation, (ii) the plasmon excitation would make electrons inject into the conduction band of the surface nanosized WO 3 moiety at the WO 3 /W interface, and (iii) the electrons at the WO 3 surface would react with the adsorbed MB molecules. At present, the active species at the WO 3 /W interface could be unfortunately deactivated by adsorption of the reaction products and the surface oxidation during the reaction.