Local Surface Plasmonic Resonance, Surface-Enhanced Raman Scattering, Photoluminescence, and Photocatalytic Activity of Hydrothermal Titanate Nanotubes Coated with Ag Nanoparticles

Department of Natural Sciences Education, Saigon University, 273 An Duong Vuong, Ward 3, District 5, Ho Chi Minh City 700000, Vietnam Department of Environmental Sciences, Saigon University, 273 An Duong Vuong, Ward 3, District 5, Ho Chi Minh City 700000, Vietnam Department of Physics, College of Natural Sciences, Campus II, Can Tho University, 3/2 Street, Can Tho City, Vietnam International Training Institute for Materials Science (ITIMS), Hanoi University of Science and Technology, Hanoi 100000, Vietnam The Research Laboratories of Saigon High-Tech-Park, Lot I3, N2 Street, Saigon Hi-Tech-Park, District 9, Ho Chi Minh City 700000, Vietnam Department of Electronics and Telecommunication, Saigon University, 273 An Duong Vuong, Ward 3, District 5, 700000 Ho Chi Minh City, Vietnam


Introduction
Recently, scientists worldwide have paid much attention to plasmonic photocatalysis [1][2][3]. The term "plasmonic photocatalysis" could be understood as oxide semiconductors such as ZnO, TiO 2 , and WO 3 coated with noble metals that enhance their photocatalytic activity due to the local surface plasmon resonance (LSPR) effect, which can signifi-cantly improve visible light absorption. Amongst plasmonic photocatalytic semiconductors, TiO 2 coated with noble metal nanoparticles (NPs) has been widely studied. Many reports on Ag NP coating on TiO 2 showed visible photocatalytic enhancement in both hydrogen production [4][5][6] and organic dye degradation [7]. Several other works dealt with an anatase TiO 2 nanotube (NT), a highly specific surface area material, which was obtained by annealing the hydrothermal titanate nanotube at~400°C or higher [8][9][10][11][12]. These works used XRD characterization to confirm the pure anatase phase of TiO 2 NTs. However, by Raman scattering investigation, our recent work showed that the anatase phase in the annealed titanate NTs was from TiO 2 NPs that were formed by thermal-induced interruption due to the lowthermal stability of titanate NTs, and the titanate NT itself remained in the monoclinic phase [13]. It is worth noting that calcination treatment gives rise to enhanced photocatalytic activity due to the heterojunction formation of TNTs and TiO 2 NPs.
In this work, we fabricated titanate NTs (TNTs) and Ag NPs coating TNT (TNT/Ag) samples and studied the photodegradation of methylene blue dye under visible light exposure. Surface morphology and crystal structure of synthesized samples were characterized by transmission electron microscopy (TEM), energy-dispersive X-ray spectroscopy (EDX), X-ray diffractometry (XRD), and Raman spectroscopy. In order to obtain insights into the enhanced photocatalytic activity, optical properties of synthesized samples were investigated by UV-vis diffuse reflectance spectroscopy (DRS) and photoluminescence (PL) emission spectroscopy, in which bandgap and charge separation of synthesized samples were identified.

Synthesis of TNT.
The hydrothermal synthesis of TNT in sodium hydroxyl solution was reported previously [14,15]. Typically, 0.84 g TiO 2 was magnetically stirred in an 80 ml 10 M NaOH solution for 30 min. Then, the resulting solution was transferred to teflon autoclave for hydrothermal processing at 150°C for 24 h. After the hydrothermal process, the autoclave was naturally cooled down to room temperature. The resulting wet powder was collected and immersed in a 0.1 M HCl solution for 30 min and then washed with bidistilled water until the pH reached 7. Finally, the powder was put in an oven at 100°C for hours until completely dry.
2.3. Synthesis of TNT/Ag Nanocomposite. In order to attach Ag NPs onto the TNTs, a reduction process to form Ag from Ag + was carried out, as described previously [16]. Typically, firstly, 0.5 g TNTs were dispersed into 30 ml bidistilled water. Secondly, the resultant solution was, drop by drop, added to 0.1 M AgNO 3 with continuous magnetic stirring until a ratio of Ag to Ti of 0.05 was reached. Thirdly, the powder was collected by centrifugation with a velocity of 5000 rounds per second, and then, NaBH 4 was added as the reduction agent. Finally, the powder was washed, filtered, and dried at 90°C, and the final product was denoted as TNT/Ag.

Results and Discussion
3.1. Surface Morphological and Structural Properties. Figure 1 shows transmission electron micrographs of TNT and TNT/Ag samples. Figures 1(a) and 1(b) indicate that the nanotubes were formed by the hydrothermal method with a homogeneous diameter of~8 nm, and the length varied from 20 nm to 50 nm. Figures 1(c) and 1(d) clearly show TNT/Ag composites in which uniform Ag nanoparticles with a size of~3 nm were attached and well distributed on the TNT surface. EDX mapping analysis of TNT/Ag, as shown in Figure 1(e), indicates the homogeneous distribution of Ag NPs on the TNT surface. Figure 2(a) shows XRD patterns of synthesized samples. In general, XRD patterns of TNT and TNT/Ag were similar to the monoclinic phase of H 2 Ti 3 O 7 , with diffraction angles (2θ) of 9.40°, 24.18°, 28.26°, and 48.22°corresponding to directions of (001), (110), (211), and (020) [14,17,18]. The structural transformation from anatase TiO 2 to monoclinic titanate with a layer structure was reported previously [19][20][21]. Chen et al. showed that the spacing distance of the interlayer was about 0.78 nm, which was formed by conjugated edge-sharing TiO 6 octahedrons [22]. Hydrogen atoms in the H 2 Ti 3 O 7 nanotube were located between interlayers, which replaced Na atoms in the as-prepared sodium titanate nanotube through the HCl postwashing treatment [23]. Because of the weak bonding of the H atom in the nanotube, the monoclinic crystal of H 2 Ti 3 O 7 could easily be transformed back to anatase TiO 2 by simple annealing [13]. It is worth noting that after attachment of the Ag nanoparticle, the diffraction intensity of (211) direction was recognisably increased. Figure 2(b) shows Raman active modes of TNT and TNT-Ag appearing at 192, 275.5, 445.6, and 668.3 cm -1 . These active modes represent Raman characteristic vibrations of the titanate tubular structure [24][25][26], which were sharply enhanced in the TNT/Ag due to the surfaceenhanced Raman scattering (SERS) effect by loading Ag nanoparticles onto the TNTs. It can be seen that there was no anatase active Raman mode in the synthesized samples. Hence, the hydrothermal process fully transformed the anatase TiO 2 to the monoclinic titanate nanotube. The enhanced Raman scattering in TNT/Ag proposed a useful method to characterize the Raman vibration modes of lowcrystalline structures of TNT, whose Raman modes were sometimes reportedly hard to distinguish [27] because of the low crystal structure [26,28].
The chemical composition, oxidization state, and the presence of Ag nanoparticles in TNT/Ag nanocomposite were determined by utilizing XPS analysis as seen in Figure 3. Particularly, Figure 3(a) shows that the TNT/Ag sample possessed three major elements, namely, Ti, O, and Ag, corresponding Ti 2p, O 1s, and Ag 3d spectra, accordingly. Regarding XPS spectra of the Ag 3d region (Figure 3(b)), two emerging peaks were located at 367.4 eV and 373.4 eV, which are assigned to characteristic binding energy of Ag 3d 5/2 and Ag 3d 3/2 spin states [29]. As proposed in previous studies [8,30,31], these peaks with a distance around 6 eV originated from Ag nanoparticles 2 Journal of Nanomaterials     Journal of Nanomaterials (peak area) ascribed to Ti 3+ state which is related to Ti 2 O 3 structure [4,32]. Additionally, O 1s spectra had two deconvoluted peaks, one from 529.6 eV and the other from 531.3 eV attributing to Ti-O bond and surface adsorbed OH groups [33], respectively, which are shown in Figure 3(d). Importantly, there was a negative shift of around 0.6 eV and 0.4 eV in terms of Ti 2p and O 1s spectra of TNT/Ag, respectively, compared to that of pure TNT (as seen in Figure S1). This is due to the connection of Ti-O bond in the TNT structure and Ag nanoparticles on its surface through electrostatic or covalent interaction to form Ti-O-Ag networks [34].

Photocatalytic Test.
We added 0.01 g catalyst (starting TiO 2 , TNT, or TNT/Ag) in a 250 ml solution of 25 ppm methylene blue (MB) for photocatalytic evaluation of MB degradation. After that, the powders were simultaneously stirred in the dark for 60 min until the adsorption reached an equilibrium state. Then, the visible photocatalytic measurement was carried out by irradiating the resulting solution under natural solar light on a sunny day. We started measuring MB degrada-tion at different times of 5, 15, 30, 60, and 90 min under sunlight irradiation. Figure 4 shows the sequence of MB absorption spectra over exposure times (Figures 4(a)-4(c)) and the degradation efficiency of MB by using its characteristic absorption peak at 664 nm (Figure 4(d)), where C and C 0 were concentrations of MB at the time of measuring and at the beginning of the experiment, respectively. It showed that the time for MB degradation to reach equilibrium was short (~5 min) in comparison to that of the TNT sample (~90 min), with MB degradation of 40% and 12%, respectively. However, after 90 min of exposure, the photodegradation efficiency of TNT and TNT/Ag, which was~45% and~50%, respectively, was not much different. Hence, photocatalytic activity of TNT/Ag was significantly enhanced in terms of time compared to the TNT sample. In order to understand the mechanism of the enhancement, in the next section, we will investigate the optical properties of the synthesized samples by DRS and PL measurements. Figure 5 shows UV-vis DRS spectra of the synthesized samples and the bandgaps determined   [35]. In general, the DRS spectra in Figure 4(a) showed improvements in visible light absorption of the TNT/Ag composite, which played an important role in the enhanced photocatalytic activity. Furthermore, local surface plasmon resonance-(LSPR-) enhanced absorption was revealed in the range of 400-600 nm, which was around the characteristic strong absorption range of Ag nanoparticles [36,37]. Hu and coworkers reported that the broad absorption peak was possibly due to the nonuniformity of Ag nanoparticles [38]. However, as shown in TEM images, Ag NPs in TNT/Ag were highly uniform with a size of 3 nm, and the low   Journal of Nanomaterials concentration of Ag NPs could be attributed to the broadening of the resonant peak. The bandgap values of the synthesized samples were estimated from the DRS spectra using the Kubelka-Munk method [14,35] shown in Figures 5(b) and 5(c), where the values were~3.28 eV and~3.18 eV for TNT and TNT/Ag, respectively. The bandgap decrease in the composite sample was attributed to the increase of the Fermi level when n-type TNT semiconductor contacted with metallic Ag nanoparticles [39]. The small "bump" signal at lower energy in Figure 4(c) was induced by the SPR effect which was observed in the reflectance spectra. Hence, by the DRS measurement, TNT/Ag showed that Ag NPs induced an LSPR effect in which the visible light absorption capacity of the sample significantly increased, which could be assigned to the enhanced photocatalytic activity. Both enhanced light harvesting and bandgap reduction in the composite sample were beneficial to photocatalytic improvement [40][41][42][43]. Another important factor for enhanced photocatalytic activity was the radiative combination rate that brought about information of the photoinduced carriers' lifetimes, which were usually characterized by the photoluminescence (PL) emission intensity. Figure 6(a) shows normalized PL emission intensities of synthesized samples, which indicated that the radiative recombination rate of photoinduced electron-hole pairs was possibly reduced in the TNT/Ag composite [40,44,45]. Typically, PL signals that appeared at 427 nm (2.90 eV), 585 nm (2.12 eV), and 647 nm (1.92 eV) were assigned to radiative recombinations of bandband excitons, conduction band electron-singly ionized oxygen vacancy (V + O ), and electron trap-valence band hole, respectively [44,46]. The lower the radiative recombination, the longer the carrier lifetime and, thus, the higher the possibility of photoinduced carriers interacting to degrade organic compounds [47]. However, an unwanted case known as nonradiative emission could happen to reduce the PL emission intensity when the coated materials cover the host surface [48]. This phenomenon was called the PL quenching effect, which many previous photocatalytic studies did not pay much attention to whether the decrease of PL emission intensity was because of the charge separation or the PL quenching effect. In order to clarify the decreasing recombination rate in the composite, a time-resolved PL measurement was carried out, as shown in Figure 6(b). The excitation wavelength of 466 nm from a 0.5 μW laser source with 5 ps pulses was used. The fitting decay curves of timeresolved PL spectra shown in the subfigure of Figure 6(b) indicated that the photoelectron lifetime of TNT/Ag of 4.26 ns was significantly longer than that of TNT, which was~1.68 ns. Hence, the recombination rate of the composite sample was about 2.75-fold longer than that of TNT; thus, photoexcited carriers would have more time in the solution vicinity to promote highly oxidative and reductive radicals to decompose the organic dye [49,50].

Conclusions
In this work, we successfully fabricated hydrothermal TNTs coated with Ag NPs by a facile two-step method. The photocatalytic activity of the TNT/Ag composite was investigated by measuring the degradation of MB. The MB degradation equilibrium state was reached only after 5 min of sunlight irradiation with a degradation efficiency of~40%. The mechanism of enhanced photocatalytic activity was revealed by DRS and PL studies, which, respectively, showed LSPRinduced visible light absorption and charge separation improvement, i.e., photoinduced carrier lifetimes of 1.68 ns for TNTs and~4.26 ns for TNT/Ag. Furthermore, the SERS effect presented in the TNT/Ag sample was useful to characterize the TNT layered monoclinic microstructure.

Data Availability
Data is available on request.

Conflicts of Interest
The authors declare that they have no conflicts of interest.