Plasmon-Enhanced Photocurrent using Gold Nanoparticles on a Three-Dimensional TiO2 Nanowire-Web Electrode

In this study, an anatase/rutile mixed-phase titanium dioxide (TiO2) hierarchical network deposited with Au nanoparticles (Au/TiO2 ARHN) was synthesized using a facile hydrothermal method followed by a simple calcination step. Such a unique structure was designed for improving the light harvest, charge transportation/separation, and the performance of photo-electro-chemical (PEC) cells. The properties of the as-synthesized Au/TiO2 ARHN in PEC cells were investigated by electrochemical measurements using a three-electrode system in a 1 M NaOH electrolyte. Remarkably, a 4.5-folds enhancement of the photocurrent for Au/TiO2 ARHN was observed as compared to that for TiO2 nanowire (NW), under AM1.5G solar illumination, suggesting its potential application in PEC cells. A mechanism has been proposed to explain the high photocurrent of Au/TiO2 ARHN in PEC water splitting.

Scientific RepoRts | 7:42524 | DOI: 10.1038/srep42524 electromagnetic coupling of the Au NPs, the solid support-TiO 2 NWs connected with TiO 2 threads-was synthesized by a two-step hydrothermal process. When tested in the PEC experiment, the TiO 2 ARHN and Au/ TiO 2 ARHN exhibited 1.5 times and 4.5 times higher photocurrent than TiO 2 NWs.  (Fig. 3b) shows that the thickness of TiO 2 layer is ~1 μ m and the NWs have an average diameter of 40 nm. XRD patterns of the TiO 2 NWs show predominantly rutile phase with preferential orientation of (110) (Supplementary Fig. 1). The top-view and cross-sectional SEM image of the TiO 2 ARHN (Fig. 3c,d) shows that the TiO 2 threads are bridged with the TiO 2 NWs to form a 3D hierarchical network. Nest-like porous cavities with diameters of a few hundred nanometers are clearly observed and the diameters of the threads are ~10 nm. The XRD patterns and Raman spectra show that the threads belong to the anatase phase ( Supplementary Figs 2 and 3). The top-view SEM image of the Au/TiO 2 ARHN is shown in Fig. 3e. The white dot regions in the SEM represent the Au NPs. The size distribution histograms of Au NPs show an average particle size of 15 nm (Fig. 3f).

Results and Discussion
To investigate the formation mechanism of TiO 2 ARHN, a series of experiments were performed. Firstly, we failed to obtain TiO 2 ARHN without TiCl 4 treatment. We found that TiO 2 threads cannot grow on TiO 2 NWs with a smooth surface. Secondly, in the absence of the Ti film in step 2 (in Fig. 2), only TiO 2 NWs were observed. These findings suggest that both TiCl 4 treatment and the formation of the Ti layer for the formation of TiO 2 ARHN are indispensable. Therefore, we propose that small TiO 2 seed crystals grow on the surface of TiO 2 NWs after TiCl 4 treatment, which leads to a rough surface for the growth of TiO 2 threads. Moreover, during alkali  hydrothermal process, the Ti layers can generate large amounts of Ti-containing species 37,38 as precursors that can deposit on the TiO 2 seeds and produce a network structure.
To evaluate the enhanced PEC performance of the designed Au/TiO 2 ARHN, the linear sweep voltammograms and the photocurrent-versus-time (I-t curve) of TiO 2 NW and TiO 2 ARHN with/without Au NPs were conducted under AM 1.5 G simulated solar illumination, as shown in Fig. 4. The measured photocurrent was normalized to the sample area to obtain the photocurrent density for comparison. As presented in Fig. 4a, the TiO 2 NW electrode produced a photocurrent density of 4 × 10 −5 A cm −2 at 0.23 V vs. Ag/AgCl, which is the potential often chosen as a metric to evaluate the performance of photoanodes as it corresponds to the water oxidation potential 4 . The low photocurrent density is attributed to the limit of wide band-gap characteristics of TiO 2 (3.2 eV for anatase 39 and 3.0 eV for rutile 40 ), due to which only UV light can be used in the PEC water splitting system. The photocurrent was enhanced for the TiO 2 ARHN (6 × 10 −5 A cm −2 ) when compared with TiO 2 NW, with an enhancement factor of 1.5. As expected, a significant photocurrent density enhancement was clearly observed on the Au/TiO 2 ARHN, having a photocurrent density of 1.8 × 10 −4 A cm −2 . As compared to TiO 2 NW, a photocurrent enhancement higher than 4.5 times was achieved. From Fig. 4b, all electrode represent a good reproducibility and stability as the illumination was turned on and off. Furthermore, the sharp spike in the photocurrent during the on/off illumination cycles demonstrates the predominant transport of photogenerated electrons in the designed TiO 2 structure 41 . We suggest the enhanced photocurrent of TiO 2 ARHN electrode could be attributed to better photocatalytic activity, due to increased surface area, or better light harvest efficiency, due to the hierarchical network structure. Therefore, the dye absorption/desorption experiment and UV-visible spectrum measurement were perform to verify it. Here, dye N719 was choose as an adsorbate to execute dye absorption/desorption experiment due to it could be monolayer absorbed on the surface of TiO 2 . Therefore, we can evaluate the related surface area via measuring the absorption of N719 dye which detach from TiO 2 structure. As shown in Fig. 5, there are three absorbed peak of N719 located on 310 nm, 370 nm and 505 nm, respectively 42 . It is observed that the absorption of detached N719 solution based on TiO 2 ARHN is obviously large than TiO 2 NW on entire spectrum. It represents the surface area of TiO 2 ARHN is related large than TiO 2 NW due to there are more dye absorbed on TiO 2 ARHN. Based on the result, we make a sure that the high surface area of TiO 2 ARHN is a reasonable reason which bring to a high photo activity on PEC measurement.
Furthermore, the UV-visible absorption spectra of the TiO 2 NW and TiO 2 ARHN with/without Au NPs are shown in Fig. 6. TiO 2 NW exhibits a stronger absorption at the wavelengths below 400 nm due to electron transitions of TiO 2 from the valence band to the conduction band. In addition, the absorption spectra of TiO 2 ARHN showed an enhanced absorption in the entire spectral range as compared with TiO 2 NW, which is attributed to the scattering effect in the ARHN structure; this also explains the increased photocurrent in the TiO 2 ARHN. With deposited Au NPs, the absorption show a significantly enhancement on visible range which is driven by the LSPR absorption. From incident photon-to-electron conversion efficiency (IPCE) measurement ( Supplementary Fig. 4), it demonstrates that such an absorption successfully boosts the PEC performance in the region from 400 nm to 700 nm.
In this work, a LSPR peak for the Au NPs with average size of 15 nm centered at around 540 nm. For Au NPs of size 10-20 nm, the absorption peak of plasmon resonance is usually located at 520-525 nm 43,44 . A redshift of 15-20 nm of the plasmon resonance peak was observed as compared to previous reports. This may be attributed to the TiO 2 changing the surrounding dielectric property of Au NPs (Au NPs well deposited and in contact withTiO 2 surface) and the enhancement of electromagnetic field of LSPR 30,45,46 . It has been reported that the high electromagnetic field of LSPR and strong coupling between Au NPs and TiO 2 will benefit the plasmon-induced charge transportation and separation, enabling SPR-enhanced photocatalysis. Typically, the LSPR-induced charge separation at the interface between the Au NPs and TiO 2 can occur by transferring the energy contained in the oscillating electrons or local plasmonic field from Au NPs to TiO 2 through direct electron transfer, also known  as hot electron injection 47,48 . Higher electromagnetic field generates more hot electron 49,50 . In order to verify our assumption, the design of TiO 2 ARHN is helpful to improve electromagnetic field as compared to NWs, the PEC measurement was performed to check the hot electron effect. Figure 7 shows the I-t curve measured under visible-light illumination. From Fig. 7, it is obvious that the photocurrent of Au/TiO 2 ARHN is two times higher than Au/TiO 2 NW. It means that the high plasmon electromagnetic field of Au/TiO 2 ARHN results in a high hot-electron current. Therefore, we confirm that the design of TiO 2 ARHN successfully provides a model for strengthening LSPR ability and demonstrates a remarkable enhancement on PEC performance.
In this study, we are further interested the effectiveness of density effect and size effect on the performance of photo electrochemical property. The related data and discussion were shown in Supplementary Information. In addition, the stability test and Faradaic efficiency were obtained. Under continuously illumination for 10800 seconds (equal to 3 hr), the photocurrent density was decrease from 1.8 × 10 −4 A/cm 2 to 1.5 × 10 −4 A/cm 2 in the case of Au/ TiO 2 ARHN, as shown in Fig. 8a. This photocurrent decay is similar to previous report 51 . We suggest it could be attributed to photo induced corrosion which competes with water oxidation reaction 51 . However, such a corrosion could be suppress by surface treatment of TiO 2 nanostructure or use of sacrificial reagent/catalyst for longstanding application 51 . From Fig. 8b, the calculated Faradaic efficiency exceed 90% and 85% for TiO 2 ARHN and Au/ TiO 2 ARHN, respectively. The high value of Faradaic efficiency of oxygen gas during 10 hr demonstrated that the photo generated current indeed utilized for water oxidation. Also, we could observe hydrogen gas from real picture as shown inset diagram in Fig. 8c. Therefore, we propose the electron transfer mechanism in Au/TiO 2 ARHN system as shown in Fig. 8c. Under illumination, Au NPs absorb visible light, generating the energetic hot electrons from the process of SP excitation, and injecting them into the conduction band of the adjacent TiO 2 (green arrow). Simultaneously, the UV light is absorbed by TiO 2 , producing a photo-excited electron and a hole (black arrow). The plasmon-induced electromagnetic field promotes the separation of photogenerated electrons and holes. Furthermore, as illustrated, the energy bands of anatase and rutile are different which provides a driving force to promote electron transfer from anatase to rutile (blue arrow). Finally, the electrons transferred to the cathode (Pt) react with H + ions and produce H 2 (pink arrow) whereas the holes present in the anode oxidize H 2 O and generate O 2 .  In conclusion, this work demonstrates a plasmon-induced effect on a designed 3D web architecture constructed from rutile TiO 2 NWs, anatase TiO 2 threads and Au NPs. Such a nanostructure was achieved for the first time through a simple and inexpensive hydrothermal procedure followed by calcination. The PEC performance tests, reveal that the photocurrent of Au/TiO 2 ARHN was 4.5 times greater than that of the TiO 2 NW photoelectrode. The observed optical properties and dark current measurements confirm that the excellent PEC performance of Au/TiO 2 ARHN was due to three reasons: (1) the high surface area of TiO 2 ARHN that increase the photoactive center, (2) the scattering effect in the TiO 2 ARHN and the LSPR properties of Au NPs that enhanced the light harvest, (3) the strength coupling effect between Au NPs and TiO 2 nanostructure that accelerated the charge transportation and separation. The mechanism of charge transportation in the Au/TiO 2 ARHN was proposed based on our findings. Practical use of the Au/TiO 2 ARHN was demonstrated to indicate their significant potential for use in photoelectric conversion system. Hydrothermal synthesis of TiO 2 ARHN on FTO substrates. The TiO 2 NW substrate (without calcination) was first treated with TiCl 4 solution as mentioned above. A 200-nm-thick Ti layer was sputtered on the TiCl 4 -treated TiO 2 NW using a magnetic sputter (K575X, Quorum Technologies). The substrates were then transferred to a Teflon vessel with the addition of a 5 M aqueous NaOH solution and were encapsulated in a stainless-steel autoclave. Then, the autoclave was heated at 80 °C for 30 min. After the low-temperature hydrothermal process, the substrate was rinsed with 0.1 M HNO 3 followed by deionized water, and was finally calcined at 500 °C for 30 min to obtain hierarchical nanostructures. Synthesis of Au/TiO 2 ARHN. A 5-nm Au layer was sputtered on the TiO 2 NW and ARHN using a magnetic sputter. The Au deposited TiO 2 were subsequently calcined at 500 °C for 1 h to obtain Au/TiO 2 ARHN.

Methods
Characterizations and measurements. The surface and cross-section morphologies of samples were examined by field-emission scanning electron microscopy (FE-SEM, Zeiss Ultraplus). The SigmaScan Pro 5 software was used to calculate particle size (300 particles were counted). The crystal structure was characterized by X-ray diffraction (XRD, PANalytical X'Pert Pro MRD) and Raman spectroscopy (Tokyo Instruments, INC). A UV/Vis spectrometer (Perkin Elmer/Lambda 900) was used to obtain the absorption spectra. The dye absorption/ desorption experiment was performed by desorbing a dye-sensitized TiO 2 electrode in a 0.1 M NaOH solution in 1:1 H 2 O/EtOH 53 . Subsequently, a UV-vis spectrometer was introduced to measure the absorbance of the desorbing solution. The electrochemical measurement was carried out using three-electrode system. TiO 2 electrode with or without Au NPs was the working electrode; an Ag/AgCl (3 M KCl) electrode in saturated KCl was the reference electrode; the Pt plate was used as the counter electrode. All PEC cells were examined in 1 M NaOH solution with a PARSTAT 2263 Advanced Electrochemical System under illumination by Newport solar simulator with AM 1.5 G (100 mW/cm 2 ). The incident photon-to-current conversion efficiency (IPCE) was measured with an action spectrum measurement setup (Peccell, PEC-S20).