Plasmonic silver quantum dots coupled with hierarchical TiO2 nanotube arrays photoelectrodes for efficient visible-light photoelectrocatalytic hydrogen evolution

A plasmonic Ag/TiO2 photocatalytic composite was designed by selecting Ag quantum dots (Ag QDs) to act as a surface plasmon resonance (SPR) photosensitizer for driving the visible-light driven photoelectrocatalytic hydrogen evolution. Vertically oriented hierarchical TiO2 nanotube arrays (H-TiO2-NTAs) with macroporous structure were prepared through a two-step method based on electrochemical anodization. Subsequently, Ag QDs, with tunable size (1.3-21.0 nm), could be uniformly deposited on the H-TiO2 NTAs by current pulsing approach. The unique structure of the as-obtained photoelectrodes greatly improved the photoelectric conversion efficiency. The as-obtained Ag/H-TiO2-NTAs exhibited strong visible-light absorption capability, high photocurrent density, and enhanced photoelectrocatalytic (PEC) activity toward photoelectrocatalytic hydrogen evolution under visible-light irradiation (λ > 420 nm). The enhancement in the photoelectric conversion efficiency and activity was ascribed to the synergistic effects of silver and the unique hierarchical structures of TiO2 nanotube arrays, strong SPR effect, and anti-shielding effect of ultrafine Ag QDs.


FESEM, TEM images and size distribution.
. It should be noted that the length of the TiO 2 -NTAs in this study was rationally chosen, as it has been reported that about 2.0 μ m was the maximum penetration depth of the incident light in TiO 2 -NTAs 13 . Further increasing the tube length would improve its electronic resistance, inhibiting the photo-generated transmission located in the TiO 2 -NTAs. A cross-section FESEM view of the H-TiO 2 -NTAs (Fig. 2b) indicated a uniformly non-flat, concave-like top layer, which was closely connected with the bottom TiO 2 -NTAs. The concaves in the top layer were expected to work as nanomirrors for light reflection and scattering 13 . From Fig. 2c,d, one could clearly see that silver QDs were uniformly deposited on the surface of the pore wall of H-TiO 2 -NTAs with an average size about 2.5 nm via the proposed pulse electro-deposition route. The crystal lattice of Ag and TiO 2 can be clearly seen shown in Fig. S1. The lattice space of lattice facet with Ag (111) was 0.238 nm, and the lattice spacing of 0.352 nm was assigned to the (101) facet of TiO 2 . Besides, it can also be observed that the inner diameter of the nanotubes was 43 ± 1.0 nm, and the thickness of the TiO 2 nanotubes was 11 ± 2.0 nm. Both of the values were much lower than those of the porous surface layer of TiO 2 . As shown in Fig. S2, the size of the loaded Ag nanocrystals could be tunable from about 1.3 nm to 21.0 nm by increasing the pulse electro-deposition time from 10 s to 100 s. The space of the inner nanotubes were almost totally complete occupied by the large Ag nanocrystals upon choosing 100 s electro-deposition time. Such blocking effect resulting from the over growth of Ag nanocrystals will not allow the light to penetrate into the TiO 2 nanotubes via multiple-reflection as shown in Fig. 3. It was obvious that Ag, as photosensitizer, could exhibit SPR effects when it was irradiated with visible-light (490 nm > λ > 390 nm) 19,25 . With the aid of SPR effect, hot electrons were able to transfer from Ag nanocrystals into the conduction band of TiO 2 -NTAs. Meanwhile, the wall of TiO 2 nanotubes possessed excellent ability to transfer electrons, allowing more electrons to be trapped by protons with the formation of H 2 . Thus, it was reasonable that Ag/H-TiO 2 -NTAs would exhibit enhanced photoelectrocatalytic activities in H 2 evolution owing to the SPR effect and anti-shielding effect of ultrafine Ag QDs.

XPS analysis.
In order to acquire in-depth fundamental information on the interaction of Ag with TiO 2 , the X-ray photoelectron spectroscopy (XPS) technique was employed to analyze the specific surface composition and elemental binding energy of H-TiO 2 -NTAs and 20 s Ag/H-TiO 2 -NTAs. Figure 4a exhibited the high-resolution spectrum of Ag 3d from the Ag modified H-TiO 2 NTAs sample. The Ag 3d 5/2 core level of 20 s Ag/H-TiO 2 -NTAs could be fitted with a single peak at a binding energy of 368.5 eV, which was attributed to the presence metallic Ag 0 . A significant positive shift of the binding energy for Ag 3d 5/2 relative to 368.0 eV of the bulk Ag was identified due to the electron transfer from Ag to oxygen vacancies of the TiO 2 . 9 Such positive shift may be attributed to the low work function of metal Ag. Moreover, a small negative shift of the Ti 2p 3/2 after loading Ag (shown in Fig. 4b) also revealed the feasibility of the electron transfer between the Ag and H-TiO 2 -NTAs.
Electrochemical testing and Photoluminescence. The interfacial properties between the electrode and the electrolyte were detected by electrochemical impedances spectroscopy (EIS) measurements. A semicircle in the Nyquist plot at high frequency represented the charge-transfer process, while the diameter of the semicircle reflected the charge-transfer resistance (Fig. 5a). It was clear that the arch for Ag/H-TiO 2 -NTAs under visible light (> 420 nm) illumination was much smaller than that for H-TiO 2 -NTAs, implying that decoration with Ag QDs could significantly enhance the electron mobility by reducing the recombination of electron-hole pairs. Via comparison of the different samples, the arch value for Ag/TiO 2 -NTAs without porous top layer was in the middle among them under visible-light irradiation. It was noted that Ag QDs exhibited an important role in the solution-action process of   electron transfer. In addition, the capacitance measurement was performed on the electrode/electrolyte according to the Mott-Schottky equation 26 , where C is the space charge capacitance in the semiconductor, N D is the electron carrier density, e is the elemental charge value, ε 0 is the permittivity of the vacuum, ε is the relative permittivity of the semiconductor, E is the applied potential, E FB is the flat band potential, T is temperature, and k is the Boltzmann constant. Figure 5b displays the Mott-Schottky plots of 1/C 2 as a function of the applied potential, from which the positive slopes (i.e., lines) were observed, suggesting n-type semiconductors. Furthermore, the plots were extrapolated to 1/C 2 = 0 to estimate the values of E FB at − 0.251 V and − 0.213 V for H-TiO 2 NTAs and Ag/H-TiO 2 -NTAs, respectively. A 38 mV increase of E FB implied steeper band bending, forming strong Schottky junction between TiO 2 and Ag, and thereby facilitating the electron transfer. In addition, the carrier density N D could also be calculated from Fig. 3b by using the following equation 13 : We noted that the capability of charge separation by Ag QDs could also be verified by the analysis of the photoluminescence (PL) spectra as shown in Fig. 3c. PL measurements were often employed to study surface processes involving electron-hole recombination of TiO 2 . Briefly, after irradiation of the photocatalysts, electron-hole pairs underwent a recombination process, and photons were then emitted, resulting in PL 18 . As shown in Fig. 5c, the peak at 424 nm (corresponding to 2.92 eV) could be attributed to self-trapped excitations located on the TiO 2 octahedral. The peaks at 446 nm, 459 nm, and 487 nm (corresponding to 2.78 eV, 2.70 eV, and 2.55 eV, respectively) were also observed, which were associated with the oxygen vacancies [27][28][29][30] . The PL intensity for 20 s Ag/H-TiO 2 -NTAs was much lower than that of H-TiO 2 -NTAs, indicating a reduced charge carrier recombination resulting from the formation of Schottky junction between TiO 2 and Ag nanocrystals 31 . It was also noted that the pure H-TiO 2 -NTAs exhibited a lower intensity of PL compared to that of the Ag loaded traditional TiO 2 nanotube arrays (20 s Ag/TiO 2 -NTAs). This suggested that the formation of hierarchical structured TiO 2 nanotubes played an important role for inhibiting the electron-hole recombination.

Discussion
Photoelectrochemical properties of the pure and Ag QDs modified H-TiO 2 -NTAs and schematic diagram of SPR charge carrier transfer mechanisms. To evaluate the enhanced PEC performance of the fabricated Ag/H-TiO 2 -NTAs under visible-light irradiation, the transient photocurrent was done to verify the separation of photo-induced charges. Chronoamperometric I-t curves were measured and recorded in Fig. 6a by irradiating the electrodes with visible-light (λ > 420 nm) at a potential of 0.7 V vs SCE. It was surprising that 20 s-Ag/H-TiO 2 -NTAs with hierarchical structures exhibited a very high photocurrent density (0.104 mA/cm −2 ), about 3 times of that (0.034 mA/cm −2 ) of the traditional Ag/ TiO 2 -NTAs. Such greatly enhanced photo-response ability could be attributed to its special structure and enhanced SPR intensity. It was also noted that the photocurrent density of Ag/H-TiO 2 -NTAs increased with the pulse deposited time of Ag nanocrystals. Prolonging the deposition time over 20 s could result in the decrease of photocurrent density because of the formation larger size of Ag nanocrystals, which could shield the multi-reflection of light in the hierarchical TiO 2 -NTAs.
As schematically presented in Fig. 6b, Ag QDs, as the sensitizer, were excited under visible-light irradiation. Hot electrons from the SPR of metal Ag, possessed enough energy to overcome the Schottky barrier between the Ag and TiO 2 -NTAs 32,33 and inject into the conduction band of the adjacent H-TiO 2 -NTAs. Meanwhile, these hot electrons would transfer along the wall of TiO 2 -NTAs, which could provide the fast transfer paths. Besides, the ultrafine size of Ag nanocrystals could also be favorable for enhancing the transfer rate of electrons owing to its short electron-migration distance between excited Ag and the wall of TiO 2 -NTAs.
With the aid of external bias voltage, the electrons could transfer from photoanode to cathode (platinum foil), and further react with the H + ions to form H 2 on the surface of Pt foil. The positive charges formed on Ag QDs possessed certain oxidation ability to participate in oxidation reaction, in which ethylene glycol was transfered to glyoxal, oxalic acid or other by-products. In addition, the optical properties of H-TiO 2 -NTAs and Ag/H-TiO 2 -NTAs samples were investigated by UV-vis diffuse reflectance spectroscopy (DRS). As shown in Fig. 6c, the absorption edge at a wavelength lower than 380 nm could be attributed to the intrinsic band-gap absorption of anatase TiO 2 . The absorption of bare H-TiO 2 -NTAs Scientific RepoRts | 5:10461 | DOi: 10.1038/srep10461 (Fig. 6c) in the visible-light region could be assigned to the hierarchical ordered nanoarrays. Such ordered TiO 2 nanoarrays can form the photonic crystals, allowing the TiO 2 films to adsorb visible-light 13 . However, such adsorbed visible light cannot be utilized to activate the pure TiO 2 for the formation of electro-hole pairs. Upon loading Ag nanocrystals, Ag/H-TiO 2 -NTAs exhibited similar multi-peaks in the visible-region and a significant enhancement of visible-light absorption due to the SPR absorption of the Ag QDs. Among all these Ag enhanced H-TiO 2 -NTAs samples, 20s Ag/H-TiO 2 -NTAs displayed the highest SPR intensity, indicating that the SPR effect of Ag was well maintained even its size was decreased to about 2.5 nm via carefully choosing the pulse electro-deposition route. This was consistent with the measured SPR wavelength of 440 nm of the same-sized Ag 22 nanocrystals dispersed in solution. And an enhanced visible-light absorption region (wavelength 400-550 nm) was obvious when the Ag NPs size was very small 34 compared with other samples (in Fig. S2). It should be pointed out that further increasing the size of Ag nanocrystals, via extending the pulse electron-deposition time over 20 s, resulted in a decrease of the SPR intensity in Fig. 6c. Such decrease could be ascribed to the light shielding effect of the Ag nanocrystals with larger grain size on the pore wall of TiO 2 nanotubes.
Size of Ag QDs and photocatalytic H 2 production. As shown in Fig. 7a, the size of Ag QDs could be tuned from 1.6 ± 0.3 nm to 20 ± 1.0 nm via choosing various pulse-deposition time. Based on the  linear fitting equation (y = − 2.9 + 0.22x,y represents the size of Ag QDs, x is the deposition time), it was noted that the size of Ag QDs could be controlled by the deposition time consistent with a linear variation. Thus, the pulse deposition method was a suitable route for the fabrication of Ag QDs with controllable size. These results were consistent with the SEM images in Fig. S2. Figure 7b shows the hydrogen generation performance of Ag QDs enhanced hierarchical TiO 2 nanotubes arrays via choosing traditional TiO 2 nanotubes as standard. The sample of 20 s Ag/H-TiO 2 -NTAs exhibited the highest H 2 evolution rate (124.5 μ mol*cm −2 *h −1 ) among all these Ag QDs modified hierarchical TiO 2 nanotube arrays. Also it indicated that the macroporous structure covering on the surface of TiO 2 nanotubes contributed lots to enhancing the photoelectrocatalytic activity of electrodes compared to the traditional TiO 2 nanotubes. The amount of hydrogen production exhibited a volcanic pattern from H-TiO 2 -NTAs to the Ag/H-TiO 2 -NTAs with increasing the size of Ag QDs (1.3 nm to 21.0 nm). The relative low H 2 production rate over 10 s-Ag/H-TiO 2 -NTAs was attributed to the lower loaded amount of Ag. The poor H 2 production rate over H-TiO 2 -NTAs with large size of Ag QDs may be attributed to the shielding effect toward light of the sizes, which was consistent with the order of light absorption capability in visible-light region as shown in Fig. 6c. It should be pointed out that the H 2 evolution rate of hierarchical TiO 2 -NTAs, upon being loaded Ag QDs (20 s deposition time), exhibited a significant increase, more than 3.45 times of that of pure hierarchical TiO 2 -NTAs. However, the H 2 production rate of Ag/TiO 2 -NTAs only exhibited an increase about 1.4 times compared to that of pure TiO 2 -NTAs. These results could further confirm that the hierarchical structure with modification of Ag QDs would enhance the photoelectrocatalytic activity for H 2 production. It might be attributed to the strong Ag-TiO 2 interaction and the excellent light-utilization efficiency of Ag QDs and H-TiO 2 -NTAs.
In summary, Ag QDs enhanced hierarchical TiO 2 -NTAs electrode was fabricated for effectively driving H 2 evolution through photoelectrocatalytic water-splitting under visible-light (λ > 420 nm) irradiation. Ag QDs with various average size (1.3-21.0 nm) could be uniformly deposited on the surface of the TiO 2 hierarchical nanotubes. Owing to the strong SPR effect and anti-shielding effect of ultrafine Ag QDs, and the unique property of hierarchical structures of TiO 2 nanotube arrays, the photoelectric conversion efficiency and activity were greatly increased. Especially, the special hierarchical TiO 2 nanotubes with macroporous layers was proved effective for enhancing the interaction between Ag QDs and TiO 2 , with the formation of high H 2 evolution activity.

Methods
Materials. Titanium sheets (0.3 mm thick, 99.5%) were purchased from Shanghai Right Titanium Industry Co., Ltd. ethylene glycol (EG), ammonium fluoride, silver nitrate, sodium nitrate, sodium sulfate, acetone and ethanol of analytical grade were obtained from Aladdin Company without further purification. All solutions were prepared using deionized (DI) water with a resistivity of 18.2 MΩ cm prepared by Millipore system.

Preparation of H-NTAs.
A pure TiO 2 nanotube arrays electrode was fabricated by an electrochemical anodic oxidation technique. Prior to anodization, the titanium sheets were rinsed in an ultrasonic bath of acetone, ethanol and distilled water for 15 min successively. The anodization was carried out using a conventional two-electrode system with the Ti sheet as the working electrode and a Pt foil (99.99%, 0.1 mm, 2 cm * 2 cm) as the counter electrode, respectively. All electrolytes consisted of 0.5 wt% NH 4 F in ethylene glycol solution and 2 vol% water. Firstly, the Ti sheet was anodized at 60 V for 1.5 h, then the anodized Ti sheet was ultrasonically removed in DI water. Subsequently, a second anodization of the same Ti sheet was performed at 30 V for 0.5 h. All the anodization processes were carried out at 30 o C. After the two-step anodization, the prepared hierarchical structure TiO 2 -NTAs samples were cleaned with DI water. All the as-anodized samples were crystallized by ambient annealing (500 o C) with a heating and cooling rate of 2 o C min −1 for 2 h. X-ray photoelectron spectroscopy (XPS) analysis was performed on a Perkin-Elmer PHI 5000C ESCA system. All binding energies were calibrated by using the contaminant carbon (C1s = 284.8 eV) as a reference.

Preparation of Ag/H-NTAs.
Electrochemical measurements. Photoelectrochemical measurements were carried out in a conventional three-electrode, single-compartment quartz cell on an electrochemical station (CHI 660D). The H-TiO 2 NTAs electrode and the TiO 2 -NTAs electrode with an active area of ca. 4 cm 2 served as working electrodes. The counter electrode and the reference electrode were a platinum sheet and saturated calomel electrode (SCE), respectively. A bias voltage of 0.7 V was utilized for driving the photo-generated electrons transfer from the working electrode to the platinum electrode. A 300 W Xe lamp with an  ultraviolet filter (λ > 420 nm) used as the visible light source was positioned 10 cm away from the photoelectrochemical cell. A 0.5 M Na 2 SO 4 aqueous solution was used as the electrolyte. Impedance measurements were performed under illumination (300 W Xe lamp) in 0. 5 M Na 2 SO 4 solution at open circuit voltage over a frequency range from 10 5 to 10 −1 Hz with an AC voltage at 50 mV. The Mott-Schottky plots were obtained at a fixed frequency of 1 KHz to determine the flat-band potential and carrier density.
Hydrogen evolution measurements. Hydrogen evolution measurements were carried out with a home-made reactor instrument which contains two houses separated by different tubular chambers made by quartz showed in Fig. 9. Such reactor can avoid the mixing of hydrogen generated on the Pt electrode and oxygen (consumed by sanctified agent) generated on the photoanode, and thus the hydrogen and oxygen gases can be separately collected. The home-made reactor device was filled with a 2 M ethylene glycol and 0.5 M Na 2 SO 4 solution. Ag/H-TiO 2 -NTAs was used as photoanode. SCE electrode was used as reference electrode. Pt foil was used as counter electrode. For testing the amount of the generated H 2 , a 0.5 mL of the gas was sampled intermittently through the septum after 1 h visible-light irradiation, and hydrogen was analyzed by gas chromatograph (GC9800 (N)), Shanghai Ke Chuang Chromatograph Instruments Co. Ltd, China, TCD, with nitrogen as a carrier gas and 5 A molecular sieve columns.