Thin film and noble metal loading effects on the photocatalytic reactivity of helium-plasma-induced nanostructured tungsten oxides

Two types of samples were used in the experiment: sheet (PLANSEE, 0.2-mm thickness, 99.97% purity) and thin film. The samples were 5 × 5 mm². W thin films were deposited on a silica glass (Labo-USQ, 0.5-mm thickness) using a radio frequency (RF) magnetron sputtering apparatus. Before the deposition, the glass was cleaned using deionized water and ethanol and then was air-dried [42]. Helium plasma irradiations were conducted in the linear plasma device NAGDIS-II, whose setting details can be found elsewhere [14]. After irradiation, the nanostructured W samples (sheet and thin-film) were calcined at different temperatures in air for 30 min using an electric furnace (FO100). A layer of noble metals (Ag, Au, and Pt) was deposited using an RF magnetron sputtering apparatus. The layer thickness of noble metals was characterized as the product of sputtering time multiplied by the deposition rate on the flat plane. Since the noble metal depositions were performed after the nanostructure formation, the actual thickness of the noble metal layer on the nanostructures was different from the measured thickness. The thickness was used as an index of the amount of loading.

The photocatalytic performances of the samples were measured using MB degradation experiments under visible light. A high-pressure Xenon lamp with a blue additive color dichroic filter and a cut-off filter permitting >400 nm was used to provide visible light in a wavelength range of 400–500 nm. The MB absorption in the wavelength range of 400–500 nm was not significant, as shown in figure 1.

Zoom In Zoom Out Reset image size

MB is often used as a substrate in photocatalytic degradation. An issue when using MB to check photocatalytic performance under visible light has been reported due to its light absorption [43]. However, Komori et al confirmed that MB decomposition occurred using K-edge XANES and showed that partially oxidized nanostructured W is an excellent photocatalyst for decolorizing MB under near-infrared light (λ > 800 nm) [17, 18]. Thus, MB can be used to assess the photocatalytic activity at least under the conditions used in this study. The photocatalytic activity of the materials can be defined from the MB decomposition rate (i.e., the rate of MB concentration change). The concentration of MB was deduced through converting the absorbance, which was recorded using a UV–vis spectrometer (UV-2600). Concerning the relationship between the microstructure of the photocatalysts and their performance, the influence of the surface morphology on the photocatalytic activity was investigated using MB [42]. The formation of FN on the W sheet surface was found to increase MB decolorization three times more than on the pristine WO₃ sheet sample.

All W thin-film samples were synthesized on silica glass using RF magnetron sputtering with argon gas at room temperature. The chamber was evacuated to ∼10⁻⁴ Pa before deposition. A 100-mm diameter and 1.0-mm thick W disc (99.95% purity) was used as a sputtering target. The RF power and working pressure were 100 W and 1 Pa, respectively. After irradiation and calcination, the surface composition analysis of the nanostructured W oxide samples (sheet) was accomplished using x-ray photoelectron spectroscopy (XPS; Shimadzu ESCA-3300) using Mg Kα (1253.6 eV) as the x-ray source. XPS measurements of the nanostructured W thin-film oxide samples were performed using a different device (ESCALAB 250Xi) with Al Kα emission for the x-ray source. Transmission electron microscope (TEM) images and energy dispersive x-ray spectroscopy (EDS) analysis of the samples was acquired from the JEOL JEM-2800 TEM.

A previous study found that a uniform nanostructured layer was formed on the surface of 300-nm-thick W thin films (W_NTF), while non-uniform isolated nanostructures were formed when the film thickness was 100 nm or less [42]. This raised the question of whether W_NTF and W sheet irradiated with the He plasma (W_Nano) have similar degradation ability at the same calcination temperature; verifing whether W_NTF can achieve a similar oxidation state to W_Nano after oxidation is important. Figure 2(a) shows the XPS (ESCALAB 250Xi) spectra of W_NTF subjected to calcination at 523K compared with W_Nano calcined at 523 K (figure 2(b)). Only the W⁶⁺ species peak was found, which demonstrated that the coexistence of W⁰ and W⁶⁺ species, which is thought to be one of the key factors of photocatalytic performance [17], was not been found on the surface. Temperature dependences of the ratios of W to the sum of W trioxide and W (W/[WO₃+W] = the sum of the peak area of ${{\rm{W}}}^{0}4{{\rm{f}}}_{5/2}$ and ${{\rm{W}}}^{0}4{{\rm{f}}}_{7/2}$))/the sum of the peak area of ${{\rm{W}}}^{0}4{{\rm{f}}}_{5/2}$, ${{\rm{W}}}^{0}4{{\rm{f}}}_{7/2}$, ${{\rm{W}}}^{6+}4{{\rm{f}}}_{5/2}$, and ${{\rm{W}}}^{6+}4{{\rm{f}}}_{7/2}$) on the sheet and thin-film samples are plotted in figure 3. The ratio of W decreased both on W_NTF and W_Nano as the calcination temperature increased. As shown in figure 3, the calcination temperature difference in an analogous oxidation state was 50 K. In other words, the necessary oxidization temperature was lower for the thin-film sample than the sheet samples.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The formation of oxygen vacancies is thought to enhance charge separation and greatly improve photocatalytic activity [34–36]. Measuring the chemical state of oxygen by XPS is one of the principal means for obtaining detailed information about oxygen vacancies. In this study, a comparative analysis of the oxygen vacancies was performed between W_Nano and W_NTF using XPS at various calcination temperatures, and the characteristics of high photocatalytic performance samples were proposed. Figure 4 displays the O1s spectra of W_NTF at 373K. The O1s spectra were fitted with two peaks at 530.5 eV and 532 eV [37], which originated from lattice oxygen (O_L) and oxygen vacancies (O_V), respectively.

Zoom In Zoom Out Reset image size

Figure 5 summarizes the proportion of oxygen vacancies (O_V) to the sum of oxygen lattice (O_L) and oxygen vacancies (O_V) as a function of the calcination temperatures for W_Nano and W_NTF. The ratio of oxygen vacancies (O_V) to the sum of oxygen lattice (O_L) and oxygen vacancies (O_V) was equal to the ratio of the peak area of O_V to the sum of the peak area of O_V and O_L. The ratio of O_V decreased with increasing temperatures, especially for W_NTF.

Zoom In Zoom Out Reset image size

XPS is also a convenient quantitative spectroscopic technique for measuring the elemental composition of a surface. The atomic ratios of X/(X+W) were achieved using the peak areas and sensitive factors of Ag(3d), Au(4f), Pt(4f) and W(4f). Table 1 shows the variation of the surface coverage of noble metals (Ag, Au, and Pt) on nanostructured WO₃ with different deposition thicknesses. As described in section 2, the deposition thickness was derived from the deposition rate measured on a flat sample and was used as an index for the amount of deposition. As seen in table 1, the surface coverage did not differ as much as deposition thickness among the Pt/WO₃ samples.

Figures 6(a)–(b) exhibits high-resolution cross-sectional TEM (JEOL JEM-2800) images of 10-nm-thick Pt supported on a nanostructured WO₃ sample (Pt10/W_Nano[523]). Noble metal layers formed by magnetron sputtering devices are comprised of island structures with ∼8 nm in diameter, and the formation mechanism can be explained by the Volmer-Weber-type growth mode [44–46]. Island structures grow and coalesce as the noble metal film thickness increases. As shown by the TEM analyses (figure 6(b)), Pt nanoparticles formed island structures in this study as well. Although only one sample was observed, the particle size likely depended on deposition thickness, as suggested in previous studies [44–46]. EDS mapping was applied to obtain the distribution pattern of the elements. Figures 6(c)–(f) shows that W fiber was oxidized and Pt was successfully deposited. Based on the preceding analysis (figure 3), it can be speculated that Pt was loaded onto the surface in the mixed state of W and O.

Zoom In Zoom Out Reset image size

Figure 7 shows the MB decomposition rate for 2 hours under visible light irradiation per unit of irradiation area using the following samples: W_NTF calcined at (i) 373 K (W_NTF[373]), (ii) 423 K (W_NTF[423]), (iii) 473 K (W_NTF[473]), and (iv) 523 K (W_NTF[523]), and W_Nano samples calcined at (v) 473 K (W_Nano[473]), (vi) 523 K (W_Nano[523]), and (vii) 573 K (W_Nano[573]). It is noted in figure 7 that the irradiation surface area of the W_NTF samples (∼4 × 4 mm²) was ∼35% less than that of the W_Nano samples (5  × 5 mm²). As shown in figure 7, the optimum calcination temperature for W_NTF samples was 423 K, while W_Nano samples had a higher optimum temperature of 523 K. As has been discussed previously [18], one of the factors that manifested the photocatalytic performance was the coexistence of W and W oxides. Full oxidization did not occur on either the ${{\rm{W}}}_{\mathrm{NTF}}(423)$ or the ${{\rm{W}}}_{\mathrm{Nano}}(523)$ samples. However, the decreased photocatalytic performance of W_NTF samples calcined from 423-473 K could not be explained only by the oxidization fraction.

Zoom In Zoom Out Reset image size

The oxidation state of W_NTF(473) and W_Nano (523) were quite similar (figure 3). However, the photocatalytic performance of W_NTF(473) was not the greatest. Oxidation degree and oxidation vacancies are a sensitive part of WO₃ properties. As seen in figures 3 and 5, some differences were found in the oxidation degree and oxidation vacancies of W_Nano and W_NTF. This was due to the synergy between oxidation degree and oxidation vacancies. The ratio of O_V/O_L on W_NTF decreased significantly from 423 to 473 K. Thus, the highest photocatalytic activity was attributed to the W/ WO₃ interface and appropriate oxygen vacancies. The different photocatalytic performances were obtained by different thicknesses [47, 48]. For example, the values of steady-state photocurrents under fixed illumination intensity depended on the film thickness of WO₃ films [47]. Additionally, WO₃ film with a thickness of 2μm showed the highest photocatalytic activity to degrade acid orange 7 under UV illumination [48]. The optimum thickness likely depends on fabrication methods and focused reactions. Because a layer with a thickness of 300nm is sufficient to form a uniform nanostructure [42], the effect of oxidation temperatures and the reason for the differences found in the sheet and thin-film samples were the focus of this study.

Figure 8(a) demonstrates the time evolution of MB concentrations using (i) W_Nano(523), (ii) 5-nm-thick Pt-supported W_Nano(523) (Pt5/W_Nano[523]), (iii) 10-nm-thick Pt-supported W_Nano(523) (Pt10/W_Nano[523]), (iv) 10-nm-thick Ag-supported W_Nano(523) (Ag10/W_Nano[523]), (v) 10-nm-thick Au-supported W_Nano(523) (Au10/W_Nano[523]), and (vi) 15-nm-thick Pt-supported W_Nano(523) (Pt15/W_Nano[523]) as photocatalysts. The activity of the Pt-loaded samples was higher than that of the other samples; the deposition of Pt on W_Nano(523) had a positive effect on photocatalytic activities. Additionally, the greatest photocatalytic activity was identified with 10-nm-thick Pt loading. The lower photocatalytic activity of Pt5/W_Nano(523) compared to Pt10/W_Nano(523) was the result of the shortfall of the Pt deposit. The lower photocatalytic performance of Pt15/W_Nano(523) was because the layer was so thick that surface activity sites were obscured.

Zoom In Zoom Out Reset image size

As shown in figure 8, the deposition of Pt and Au on W_Nano(523) had an active effect on photocatalytic activities, while the effect of Ag support was negative. Young et al [39] found analogous experimental results. The interpretation was based on the formation of the Schottky barrier [49] as an effective photo-induced electron trap that prevented the recombination of electron holes and prolonged electron life. Thus, the effect of noble metals (Ag, Au, and Pt) loading is attributed to the Schottky barrier. In other words, separation of photo-induced electrons and holes becomes easier with a higher Schottky barrier.

The enhancement of photocatalytic performance in ascending order was as follows: 10 nm Ag/WO₃, unsupported WO₃, 10 nm Au/WO₃, and 10 nm Pt/WO₃. The bandgap of WO₃ was about 2.4-2.8 eV, and the electron affinity was about ${3.33}_{+0.08}^{-0.15}$ eV [50]. The work function of metals [38] and the barrier heights of metals/semiconductors are shown in table 2. The barrier height was derived from the difference between the work function of the deposited metals and the electron affinity of WO₃ (${3.33}_{+0.08}^{-0.15}$ eV). The photocatalytic activity was found to be higher when the barrier height was greater.

Surface plasmonic resonance (SPR) [51] is thought to be another interpretation of the enhancement of noble metal loading on semiconductors. However, this is still a controversy, as the flow of the electron is different based on the Schottky barrier (from the semiconductor to the metal) and the SPR effect [52].