Efficient plasmonic water splitting by heteroepitaxial junction-induced faceting of gold nanoparticles on an anatase titanium(IV) oxide nanoplate array electrode

Abstract

Plasmonic photocatalysts represented by gold nanoparticle (NP)-loaded titanium(IV) oxide (Au/TiO₂) can be promising solar-to-fuel converters by virtue of their response to visible-to-near infrared light. Hitherto, Au/rutile (R)-TiO₂ has been recognized as exhibiting photocatalytic activity higher than that of Au/anatase (A)-TiO₂. Herein, we demonstrate that the high potential of A-TiO₂ as the Au NP support can be brought out through atomic level interface control. Faceting of Au NPs is induced by a heteroepitaxial junction on an A-TiO₂(001) nanoplate array (Au/A-TiO₂ NPLA). Photoexcitation towards the Au/A-TiO₂ NPLA electrode generates current for the water oxidation reaction at λ < 900 nm with a maximum efficiency of 0.39% at λ = 600 nm, which is much larger than the values reported so far for the usual electrodes. The striking activity of the Au/A-TiO₂ NPLA electrode was rationalized using a potential-dependent Fowler model. This study presented a novel approach for developing solar-driven electrodes for green and sustainable fuel production.

---

Introduction

An ultimate goal of science and technology is the construction of artificial photosynthesis systems with water splitting as a benchmark.^1,2 Photoelectrochemical (PEC) water splitting assumes an advantage over the photocatalytic process using particles because of the inhibition of the back reaction and safety due to the spatial separation of the reaction fields where hydrogen (H₂) and oxygen (O₂) are generated. The key to achieving efficient PEC water splitting is the development of highly active photoanodes for the water oxidation reaction (WOR) with large activation energy.³ In a typical n-TiO₂ photoanode, the performance is severely limited by the slow kinetics of the holes with a diffusion length of as small as ∼10 nm (ref. 4) and its response only to UV light accounting for ∼3% of the solar energy. In view of effective solar energy utilization, plasmonic photocatalysts represented by gold nanoparticle-loaded TiO₂ (Au/TiO₂) have recently attracted much interest on account of their wide spectral response from visible to near-infrared light.^5–16 Also, the problem of sluggish hole transport in n-type semiconductor photoanodes can be eliminated in the WOR using the Au/TiO₂ plasmonic electrode because water is oxidized by the holes generated in Au NPs via localized surface plasmon resonance (LSPR) excitation without requiring their diffusion. While intensive efforts have been devoted to increase the incident photon-to-current efficiency (IPCE) of the Au/TiO₂ electrode through several approaches, such as Au–Ag alloying,¹⁷ lithium intercalation into TiO₂,¹⁸ and loading promoters for the WOR,^19,20 it is also a great challenge in materials sciences to enhance the performance of the prototype Au/TiO₂ electrode itself for underpinning the development toward the practical applications in solar-to-fuel conversion. However, in the usual Au/TiO₂ plasmonic electrode systems, the incident photon-to-current efficiency (IPCE) for the WOR remains below ∼0.03% (Table S1†).^18,21–23 The IPCE is determined by the product of efficiencies in a series of physical processes including light harvesting, hot electron transfer (HET) from Au NPs to TiO₂, charge separation, and the chemical process or WOR.²⁴ All these processes can be affected by various factors including the size²⁵ and shape²⁶ of Au NPs, and the crystal form of TiO₂.²⁷ In particular, the IPCE for the WOR by the Au/TiO₂ plasmonic electrode sharply increases when the Au particle size (d_Au) is smaller than 10 nm.²⁵ In addition, recent experimental and theoretical studies have indicated that strong electronic coupling between an electron acceptor (TiO₂) and electron donors such as plasmonic metal NPs^28,29 and CdS quantum dots³⁰ through intimate contact can expedite the photoinduced interfacial electron transfer and extend the response light wavelength. Furthermore, the IPCEs of nanostructured TiO₂ electrodes such as the nanotube array (NTA) and nanowire array (NWA), which are one to two orders magnitude greater than that of the conventional mesoporous (mp)-TiO₂ electrode consisting of TiO₂ nanocrystals,^21–23 point to the importance of effective electron transport to the electron collecting electrode on the order of microns in addition to the charge separation via the space charge layer formed by the Schottky junction between Au NPs and TiO₂.³¹ Since anatase (A)-TiO₂ possesses electron mobility much larger than rutile (R)-TiO₂,³² A-TiO₂ films with large surface areas have been favourably used as the photoanodes of dye-sensitized solar cells and perovskite solar cells since their invention.^33,34 Strangely, the IPCE for the WOR of Au/A-TiO₂ photoanodes remains at a level comparable to that of Au/R-TiO₂ photoanodes.^21–23 On the other hand, in nanoparticulate Au/TiO₂ plasmonic photocatalyst systems, Au/R-TiO₂ exhibits much higher photocatalytic activity than Au/A-TiO₂ for various reactions^35–40 including the WOR.²⁷ Recently, the superiority of Au/R-TiO₂ has been found to originate from the (111)_Au//(110)_R-TiO2 heteroepitaxial (HEPI) junction-induced faceting of Au NPs facilitating the physical processes in the reaction⁴¹ and the increase in the peripheral distance of Au NPs at the interface with TiO₂ or the catalytically active sites for the WOR.²⁷ Thus, if faceted Au NPs with d_Au < 10 nm can be formed on A-TiO₂ by overcoming the general tendency to be hemispherical to minimize the surface area, the enhancement in the IPCE of the Au/A-TiO₂ plasmonic electrode for the PEC WOR is highly expected.

Here, we show that truncated octahedral (t-Oh) Au NPs with d_Au < 10 nm can be formed on the A-TiO₂ nanoplate array (NPLA) electrode with dominant (001) surfaces through a HEPI junction with the (111)_Au//(001)_A-TiO2 orientation, and this t-Oh-Au/A-TiO₂ NPLA electrode exhibits activity for the PEC WOR that is much higher than that exhibited by hemispherical (HS)-Au/A-TiO₂ NPLA and t-Oh Au/R-TiO₂ NWA ones. Moreover, the reason is discussed on the basis of experimental and theoretical simulation results.

Experimental

Materials

Fluorine-doped tin(IV) oxide film-coated glass (FTO, TEC7) and titanium(IV) fluoride (TiF₄) were purchased from Aldrich. Tetrabutyl orthotitanate (Ti(OC₄H₉)₄ > 97.0%), hydrochloric acid (HCl > 98.0%), sodium hydroxide (NaOH > 97.0%), hydrogen tetrachloroaurate(III) tetrahydrate (HAuCl₄·4H₂O > 99%), urea (CH₄N₂O > 99.0%), and sodium borohydride (NaBH₄ > 92.0%) were purchased from Kanto Chemical Co. All chemicals were used as-received without further purification.

Synthesis of the anatase TiO₂ nanoplate array

The anatase TiO₂ nanoplate array (A-TiO₂ NPLA) was synthesized by following reported procedures with slight modification.⁴² TiF₄ (124 mg) was dissolved in a solution of Ti(OBu)₄ (0.17 mL) in 6 M HCl (10 mL) and stirred at room temperature for 15 min. The solution was poured into a Teflon container (volume 25 mL), and FTO (3 cm × 2 cm, 3 pieces) was immersed in the solution. The Teflon container was sealed in a stainless-steel cylinder and heated at 423 K for 18 h. The resulting sample was washed with 0.1 M aqueous NaOH, distilled water, and acetone, and dried in vacuo at room temperature to obtain the A-TiO₂ NPLA.

Synthesis of the rutile TiO₂ nanowire array

According the literature,⁴³ the rutile TiO₂ nanowire array (A-TiO₂ NWA) was hydrothermally synthesized. A solution of Ti(OBu)₄ (0.40 mL) in 6 M HCl (10 mL) was poured into a Teflon container (volume 25 mL), and FTO (3 cm × 2 cm, 3 pieces) was immersed in the solution. The Teflon container was sealed in a stainless-steel cylinder and heated at 423 K for 8 h. The resulting sample was washed with distilled water and acetone, and dried in vacuo at room temperature. The sample was calcined at 773 K for 8 h in air to obtain the R-TiO₂ NWA.

Au nanoparticle loading

Au nanoparticles were loaded on a TiO₂ electrode by the modified deposition–precipitation method using urea as the neutralizer.⁴⁴ Urea (146 mg) was added to an aqueous solution of HAuCl₄ (2.43 mM, 10 mL). The TiO₂ electrode was immersed in the solution and heated at 353 K for 18 h. The resulting sample was washed with distilled water and acetone, and dried in vacuo at room temperature to obtain an Au³⁺-adsorbed TiO₂ electrode. The sample was calcined at 673 K for 1 h in air to obtain the Au/TiO₂ electrode. Also, Au NPs were loaded on the TiO₂ electrode by chemical reduction. The Au³⁺-adsorbed TiO₂ electrode was immersed in an ethanol solution of NaBH₄ (1 mM), and left at room temperature for 5 min. The resulting sample was washed with distilled water and acetone. The chemical reduction procedures were repeated 3 times to completely reduce the adsorbed Au³⁺ ions.

Characterization

Scanning electron microscopy (SEM) images and energy dispersive spectroscopic (EDS) elemental mapping were conducted using a Hitachi SU8230 instrument at an applied voltage of 20 kV. Raman spectra were recorded using a Raman spectrometer (JASCO FP-1000) with a green laser (532 nm) as an excitation source. Transmission electron microscopy (TEM) images were obtained using a JEOL JEM-2100F instrument at an applied voltage of 200 kV. X-ray diffraction (XRD) patterns were recorded at 40 kV and 100 mA using a Rigaku SmartLab X-ray diffractometer. The specific surface area was quantified by means of a Micromeritics automatic surface area and porosimetry analyzer (TriStar 3000, Shimadzu) via nitrogen adsorption–desorption isotherms at 77 K. The samples were measured after degassing at 423 K for 1 h under vacuum. To quantify the Au loading amount, Au NPs were dissolved in aqua regia, and the concentration was determined by inductively coupled plasma spectroscopy (iCAP7600, Shimadzu). UV-Vis-NIR absorption spectra were obtained by the diffuse reflectance method using a UV-2600 spectrometer (Shimadzu) with an integrating sphere unit (Shimadzu, ISR-2600Plus). As the reference, BaSO₄ was used for the reflectance (R_∞). The Kubelka–Munk function [F(R_∞) = (1 − R_∞)²/2R_∞] was used for expressing the relative absorption coefficient. X-ray photoelectron (XP) spectra were measured using a Kratos Axis Nova X-ray photoelectron spectrometer with applied voltage of 15 kV and current of 10 mA using Al Kα as the X-ray source. For the energy reference, the peak of C1s (284.6 eV) was used.

Photoelectrochemical measurements

PEC measurements were carried out using the standard three-electrode electrochemical cell with the structure of Au/A-TiO₂ NPLA/FTO (working electrode) | 0.1 M NaClO₄ aqueous solution | Ag/AgCl (reference electrode) | glassy carbon (counter electrode). The active area of the working electrode was 4 cm² (2 × 2 cm). The electrolyte solution was deaerated by Ar bubbling for 30 min. The working electrode was illuminated using a solar simulator (PEC-L12, Peccell Tech. Inc.) through an optical filter Y-45 (AM 1.5, 1 sun, λ_ex > 430 nm). The linear sweep voltammetry measurements were performed using a galvanostat/potentiostat (HZ-7000, Hokuto Denko). To evaluate the incident photon-to-current efficiencies (IPCEs), the working electrode on the same cells was irradiated by the monochromatic light of a xenon lamp with a monochromator (full width at half-maximum, 10 nm) (HM-5, JASCO). The IPCEs were calculated using eqn (1):

IPCE[%] = (J_ph(E)N_Ahc/IFλ) × 100 (1)

where J_ph(E) is the photocurrent at an electrode potential of E (+0.8 V vs. RHE), N_A is the Avogadro constant, I (W cm⁻²) is the light intensity, F is the Faraday constant, h is the Planck constant, and c is the speed of light. In this study, the photocurrent was measured at the rest potential in the dark.

Also, the solar-to-current efficiency (SCE) was calculated using eqn (2):

SCE[%] = ∫IPCE(λ) × I_s(λ)dλ/∫I_s(λ)dλ (2)

where I_s(λ) is the light intensity of the sunlight at the wavelength of λ.

PEC impedance spectroscopy (PECIS)

PECIS measurements were carried out using the same cells under the illumination of simulated sunlight (λ_ex > 430 nm, 1 sun). Nyquist plots were obtained from EIS using a frequency response analyzer (HZA-FRA1, Hokuto Denko) built into the galvanostat/potentiostat. The measurements were carried out by applying a 10 mV AC sinusoidal signal over the frequency range between 100 mHz and 100 kHz. Charge transfer resistance (R_ct) values were estimated by a curve fitting for Nyquist plots assuming an RC-equivalent circuit.

Finite difference time domain (FDTD) simulations

Simulation models consist of an Au hemisphere (HS-Au) and an Au truncated octahedron (t-Oh-Au) loaded on a TiO₂ slab (650 × 650 × 50 nm³), respectively. Au particle sizes were assessed on the basis of the experimental values obtained from TEM observations. Complex refractive indices taken from the literature were used to reproduce the optical responses of these models. The light scattering behaviour of the models was simulated using a total-field scattered-field (TFSF) source. An x-polarized plane wave with wavelength from 300 to 1200 nm (f = 250–1000 THz) was generated to illuminate the top of the Au NPs (the zenith of the HS NP or the top plane of the t-Oh NP) from the z-axis direction perpendicular to the TiO₂ slab. E-field monitors were placed on the xy plane (z = 0) and the xz plane (y = 0) to collect local electric field images. Considering the balance between computational cost and accuracy, mesh refinement and perfectly matched layer (PML) with symmetric and/or anti-symmetric boundary conditions were imposed on the models. All of the FDTD simulations were carried out with an Ansys Lumerical FDTD simulation program package.

Results

TiO₂ films were hydrothermally grown on fluorine-doped tin oxide (FTO) substrates from a hydrochloric acid solution of Ti(OBu)₄ in the absence and the presence of TiF₄ at 423 K. X-ray diffraction (XRD) patterns of the samples synthesized without and with TiF₄ exhibited peaks at 2θ = 36.1° and 25.3° due to diffraction from the (101) planes of R-TiO₂ and A-TiO₂, respectively (Fig. 1a). The abnormally strong (101) diffraction peaks of R-TiO₂ and A-TiO₂ indicate that TiO₂ grows with high orientation with respect to the FTO surface in each sample. FTO has weak Raman signals at 634.3 cm⁻¹ and 770.6 cm⁻¹ assignable to the A_1g and B_2g modes, respectively (Fig. 1b).⁴⁵ The three signals for the sample synthesized without TiF₄ at 234.6 cm⁻¹, 445.6 cm⁻¹, and 609.6 cm⁻¹ are assigned to the B_1u + E_u, E_g, and A_1g modes of R-TiO₂, respectively.⁴⁶ In the spectrum of the sample synthesized with TiF₄, four signals are observed at 145.8 cm⁻¹, 406.0 cm⁻¹, 513.7 cm⁻¹, and 630.8 cm⁻¹ due to the E_g, B_1g, A_1g, and E_g modes of A-TiO₂, respectively.⁴⁶ Clearly, the addition of TiF₄ to the reaction solution completely changes the crystal form of TiO₂ from rutile to anatase.

Fig. 1 XRD patterns (a) and Raman spectra (b) of the samples hydrothermally synthesized without and with TiF₄. Top view (c) and side view (d) SEM images of the DP-Au/A-TiO₂ NPLA. Top view (e) and side view (f) SEM images of the DP-Au/R-TiO₂ NWA.

Au NPs were loaded on A-TiO₂ by the deposition–precipitation (DP) and chemical reduction (CR) methods. The top view (left) and side view (right) scanning electron microscopy (SEM) images of the samples show that the A-TiO₂ NPLA (Fig. 1c and d) and the R-TiO₂ nanowire array (R-TiO₂ NWA, Fig. 1e and f) are grown vertically or obliquely on the FTO films by the hydrothermal processes with and without TiF₄, respectively. The A-TiO₂ NPLA is composed of NPLs with length ∼1.4 μm and thickness ∼100 nm, while the dimensions of the R-TiO₂ NWs are ∼3.6 μm in length and ∼200 nm in diameter. In the F⁻ ion-free system, the Cl⁻ ions are adsorbed on the R-TiO₂(110) surface, and HEPI crystal growth from the FTO substrate occurs in the [001] direction to yield the R-TiO₂ NWA.⁴⁷ On the other hand, in the F⁻ ion-added system, TiF₄ with a sheet structure can be transformed into an A-TiO₂ NPL through hydrolysis.⁴² During the growth of A-TiO₂ NPLs, the 2D structure is maintained due to the lowering of the surface energy by the preferential adsorption of F⁻ ions on the (001) facets of the A-TiO₂ NPL.^48,49

In order to study the effects of the Au NP shape, the TiO₂ support, and the state of their interface on the activity of the Au/TiO₂ plasmonic electrodes for the WOR, the size of Au NPs should be controlled to be approximately equal between samples because the activity is sensitive to NP size. As shown by the transmission electron microscopy (TEM) images, Au NPs are highly dispersed on the A-TiO₂ NPLA with d_Au of 7.6 ± 1.8 nm by the DP method (Fig. 2a) and 7.9 ± 1.6 nm by the CR method (Fig. S1†). In the high resolution (HR)-TEM image of the DP-Au/A-TiO₂ NPL (Fig. 2b), the d-spacings of 1.84 Å and 2.37 Å observed in TiO₂ are in agreement with the values of the A-TiO₂(200) and (004) crystal planes, respectively. Also, the d-spacings of 2.07 Å and 2.33 Å in the Au NP are close to the values of the Au(100) and (111) crystal planes, respectively. Interestingly, faceted Au NPs with a truncated octahedral (t-Oh) shape are formed on A-TiO₂ with an atomically commensurate interface between Au(111) and A-TiO₂(001). As recently reported for the Au/R-TiO₂ NWA,²² t-Oh Au NPs with d_Au = 6.2 ± 1.2 nm are also formed on the (110) side walls of the R-TiO₂ NW by the DP method (Fig. S2†). On the other hand, most Au NPs of a CR-Au/A-TiO₂ NPL are near-hemispherical (HS) polyhedra (Fig. S1†), and the Au NP-TiO₂ interface of the CR-Au/A-TiO₂ NPL appears to be of lower quality at an atomic level than the interfaces of the DP-Au/A-TiO₂ NPL and DP-Au/R-TiO₂ NWA.

Fig. 2 TEM (a) and HR-TEM (b) images of the DP-Au/A-TiO₂ NPLA. Top view (c) and side view (d) of faceted Au NPs on A-TiO₂(001) constructed by the Wulff theorem. (e) An interface junction model with the (111)_Au//(001)_A-TiO2 orientation.

The equilibrium shape of Au NPs on the A-TiO₂(001) plane was constructed to be a t-Oh by the Wulff theorem using the surface energies of Au(111) (0.72 J m⁻²), Au(100) (0.87 J m⁻²), A-TiO₂(001) (0.90 J m⁻²), and the interface energy.^50,51 The interface energy was estimated to be 1.23 J m⁻² from the contact angle (θ ≈ 116°) using the Young equation (Fig. S3†). The top view (Fig. 2c) and side view (Fig. 2d) are very similar to those observed in the TEM image (Fig. 2a) and HR-TEM image (Fig. 2b), respectively. The 2D-rectangular lattices of Au(111) and A-TiO₂(001) have the unit dimensions of 4.995 Å × 2.884 Å and 3.784 Å × 3.784 Å, respectively (Fig. 2e). Fourfold of the d-spacing of Au(−110) crystal planes (d_Au(−110)) overlaps with threefold of d_A-TiO2(010), and threefold of d_Au(11−2) overlaps with fourfold of d_A-TiO2(100) with small mismatches of +1.6% [{(2.884 Å × 4 − 3.784 Å × 3)/3.784 Å × 3} × 100] and −1.0% [{(4.995 Å × 3 − 3.784 Å × 4)/3.784 Å × 4} × 100], respectively, which strongly suggests the formation of the domain-matching epitaxial junction. In the Au/R-TiO₂ NW system, a HEPI junction with the orientation of (111)_Au//(110)_R-TiO2 has been reported to provoke the faceting of Au NPs.^41,52 In a similar manner, the faceting Au NPs can be induced by a domain-matching HEPI junction with the (111)_Au//(001)_A-TiO2 orientation. The HEPI junction with the same orientation has recently been reported in the system consisting of Au NPs and A-TiO₂ nanosheets with dominant (001) surfaces and thickness of ∼5 nm.⁵³

X-ray photoelectron (XP) spectra of the DP-Au/A-TiO₂ NPLA, DP-Au/R-TiO₂ NWA, and CR-Au/A-TiO₂ NPLA have two signals at binding energies of 83.7 ± 0.1 eV and 87.4 ± 0.1 eV due to emission from the Au4f_7/2 and 4f_5/2 orbitals, respectively (Fig. S4†), which are somewhat smaller than the values for the bulk Au (83.9 eV and 87.6 eV).⁵⁴ These results indicate that the Au NPs are in a metallic state in every Au/TiO₂ system, and the interfacial electron transfer from TiO₂ to Au NPs takes place to form the Schottky junction in the dark.³¹

In the Kubelka–Munk-transformed absorption spectra, the A-TiO₂ NPLA and R-TiO₂ NWA have the absorption edges of 390 nm and 410 nm, which are in agreement with the values for A-TiO₂ and R-TiO₂, respectively (Fig. 3a).⁵⁵ When Au NPs were loaded on TiO₂, the absorption due to the LSPR appears in the visible region. The absorption of the DP-Au/A-TiO₂ NPLA as well as the DP-Au/R-TiO₂ NWA is much stronger and broader than that of the CR-Au/A-TiO₂ NPLA. Also, the peak wavelengths (λ_p) of the DP-Au/A-TiO₂ NPLA (549 nm) and DP-Au/R-TiO₂ NWA (538 nm) are shorter than that of the CR-Au/A-TiO₂ NPLA (564 nm). The absorption spectra of CR-Au/A-TiO₂ and DP-Au/A-TiO₂ NPLAs were calculated for HS and t-Oh Au NPs placed on A-TiO₂ employed as their models, respectively, by the finite difference time domain (FDTD) method (Fig. S5†). The spectra indicate that the faceting of Au NPs engenders the enhancement of the LSPR absorption, and the λ_p blueshifts with respect to the value for HS Au NPs. The strong absorption of the DP-Au/R-TiO₂ NW with the (110) side walls has recently been attributed to the (111)_Au//(110)_R-TiO2 HEPI junction-induced faceting of Au NPs.⁴¹ The broadening of the absorption spectrum for the DP-Au/A-TiO₂ NPLA is not well reproduced by the simulations. However, the simulated spectrum of t-Oh Au/A-TiO₂ possesses several absorption peaks corresponding to different resonance modes (Fig. S6†), while only one resonance peak is observed in the spectrum of the HS-Au/A-TiO₂. Thus, the extended LSPR absorption of the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA can result from the overlapping of each resonance peak broadened by the polydispersibility of the size and shape of Au NPs, and the significant electronic interaction between Au NPs and TiO₂ through the HEPI junction.^40,56

Fig. 3 (a) Kubelka–Munk-transformed absorption spectra of the DP-Au/A-TiO₂ NPLA, CR-Au/A-TiO₂ NPLA, and DP-Au/R-TiO₂ NWA. (b) Current (J)–potential (E) curves in the dark (dotted lines) and under visible light irradiation (λ_ex > 430 nm, solid lines). The current density was calculated using the apparent electrode surface area. (c) IPCE action spectra for the WOR with the solar spectrum for comparison (E = 0.8 V vs. RHE). (d) Nyquist plots for the plasmonic electrodes under visible light irradiation (λ_ex > 430 nm).

The current (J)–potential (E) curves of three-electrode PEC cells with the structure of the working electrode | 0.1 M NaClO₄ aqueous solution | Ag/AgCl (reference electrode) | glassy carbon (GC, counter electrode) using the DP-Au/A-TiO₂ NPLA, CR-Au/A-TiO₂ NPLA, and DP-Au/R-TiO₂-NWA as the working electrodes were measured (Fig. 3b). The electrode potential (E) is shown with respect to the reversible hydrogen electrode (RHE) below. In each system, current hardly flows in the dark at +0.4 V ≤ E ≤ +1.2 V. The illumination of visible light (AM-1.5, light intensity = 100 mW cm⁻², λ > 430 nm) generates anodic current due to the WOR at E > +0.54 ± 0.02 V.²⁵ Surprisingly, the DP-Au/A-TiO₂ NPLA supplies photocurrent far exceeding that of the CR-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA, whereas Au/A-TiO₂ particles usually show plasmonic photocatalytic activity for the WOR that is much lower than the Au/R-TiO₂ ones.^25,41 The specific surface areas of the A-TiO₂ NPLA and R-TiO₂ NWA were determined to be 5.1 m² g⁻¹ and 13.6 m² g⁻¹, respectively, by the Brunauer–Emmett–Teller (BET) method. The photocurrent normalized by the actual surface area of the DP-Au/A-TiO₂ NPLA reaches as much as ∼6 times that of the DP-Au/R-TiO₂ NWA (Fig. S7†). Clearly, the HEPI junction-induced faceting of Au NPs engenders a drastic increase in the plasmonic WOR activity of Au/A-TiO₂ to far surpass even that of the Au/R-TiO₂ electrode.

The photocurrents of the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA plasmonic electrodes for the WOR were measured under the irradiation of monochromatic light with varying wavelengths (Fig. 3c). The photocurrent flows at λ < 900 nm in each electrode system, and the DP-Au/A-TiO₂ NPLA electrode affords IPCEs much greater than those of the DP-Au/R-TiO₂ NWA electrode in the whole wavelength region. The maximum IPCE of the former reaches 0.39% at λ = 600 nm, while the value of the latter is 0.18% in accordance with the reported value.²² Recently, a maximum IPCE of ∼0.06% at λ = 550 nm has been reported for the brookite (B)-TiO₂ nanorod array electrode by selectively depositing Au NPs on the lateral {210} surfaces although the interface structure and shape of Au NPs are unclear.⁵⁷ Also, each IPCE action spectrum well traces the absorption spectrum, indicating that the photocurrent mainly caused by the 5d–6sp interband excitation of Au NPs^26,58 has to be further enhanced by the LSPR excitation. The redshift in the photocurrent peak of ∼50 nm with respect to the LSPR peak (λ_p) is ascribable to the large dielectric constant of water (80.2 at 293 K)⁵⁹ since λ_p is proportional to (1 + 2ε_m)^1/2, where ε_m is the dielectric constant of the surroundings.⁶⁰ Furthermore, the broad IPCE action spectra of the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA electrodes well matches the solar spectrum to afford the solar-to-current efficiency at λ > 430 nm of 0.15% and 0.08%, respectively.

To gain insight into the origin for the high plasmonic photocatalytic activity of the DP-Au/A-TiO₂ NPLA for the WOR, PEC impedance analysis was performed for the plasmonic electrodes with an equivalent circuit used for the analysis in which charge transfer resistance (R_ct) coupled in parallel with the constant phase element (CFE) is connected in series with ohmic resistance (R_s) (Fig. 3D and Table S2†). In the Nyquist plots under the illumination of visible light (AM-1.5, light intensity = 100 mW cm⁻², λ > 430 nm), the R_ct corresponding to the diameter of the semicircle decreases in the order of CR-Au/A-TiO₂ NPLA (2.4 × 10⁵ Ω) ≫ DP-Au/R-TiO₂ NWA (9.0 × 10³ Ω) > DP-Au/A-TiO₂ NPLA (1.3 × 10³ Ω). Evidently, the rate of interfacial charge transfer between the plasmonic electrode and water increases in the order of CR-Au/A-TiO₂ NPLA ≪ DP-Au/R-TiO₂ NWA < DP-Au/A-TiO₂ NPLA.

The local electric field was calculated for HS-Au(d_Au = 7.9 nm)/A-TiO₂ and t-Oh-Au(d_Au = 7.6 nm)/A-TiO₂ as the models for the CR-Au/A-TiO₂ NPLA and DP-Au/A-TiO₂ NPLA, respectively, by the 3D-FDTD method (Fig. 4a and b). In each case, Au NPs were placed on the xy plane of an A-TiO₂ slab, and light was irradiated from the z direction perpendicular to the xy plane with the excited localized electromagnetic fields monitored at the xz and xy planes. In the HS-Au/A-TiO₂ system, the electric field is uniformly distributed around the periphery of the HS-Au NPs at the interface with A-TiO₂ (Fig. 4c). On the other hand, in the t-Oh-Au/A-TiO₂ system, an intense electric field is generated around the corners and edges of t-Oh-Au NPs at the interface with A-TiO₂ (Fig. 4d and Fig. S6†). The enhancement factor (EF) defined by the square of the ratio of the local electric field intensity to the electric field intensity of incident light was calculated (Table S3†). For the HS-Au/A-TiO₂ system, the maximum EF_xz of 2.5 × 10³ and EF_xy of 3.6 × 10³ are obtained at λ_p = 702 nm. In the t-Oh-Au/A-TiO₂ system, the maximum EF_xz and EF_xy value reaches 1.1 × 10⁵ at λ_p = 650 nm, which is greater than the values for the HS-Au/A-TiO₂ system by factors of 44 and 31, respectively. The calculations were further performed for t-Oh-Au(d_Au = 6.2 nm)/R-TiO₂ as the model for the DP-Au/R-TiO₂ NWA (Fig. S8a†). An intense electric field localized near the corners and edges at the interface with R-TiO₂ is induced by LSPR excitation of the faceted Au NPs (Fig. S8b†). The maximum EF_xz and EF_xy value of 1.2 × 10⁵ at λ_p = 672 nm is comparable to the values of the t-Oh-Au/A-TiO₂ system (Table S3†).

Fig. 4 Local electric fields for HS-Au/A-TiO₂ (a) and t-Oh-Au/A-TiO₂ (b) calculated by the 3D-FDTD method. Local electric field distributions in HS-Au/A-TiO₂ (c) and t-Oh-Au/A-TiO₂ (d).

Discussion

Truncated octahedral (t-Oh) Au NPs are formed on the large-area (110) side walls of R-TiO₂ NWs with high probability by the DP method through the HEPI junction with the (111)_Au//(110)_R-TiO2 orientation.⁴¹ In a similar manner, faceting of Au NPs is induced on A-TiO₂ NPLs with dominant (001) surfaces by the (111)_Au//(001)_A-TiO2 HEPI junction during the DP process with a heating temperature of 673 K. A recent in situ TEM study has also shown that t-Oh and round-shaped Au NPs are formed on A-TiO₂(001) and (101) planes at 773 K, respectively.⁶¹ On the other hand, the room-temperature CR process produces HS-like Au NPs on A-TiO₂ NPLs. Thus, preferential surface exposure of the TiO₂ facets that can match the crystal lattice of Au(111) and heating at temperatures above 673 K are necessary for the formation of faceted Au NPs on TiO₂ with a high-quality and large-area interface, and long peripheral distance of the Au NPs at the interface with TiO₂. The Fermi energy of TiO₂ (ε_F,TiO2) can be calculated using the equation of ε_F,TiO2 = ε_CBM + kT ln(N_D/N_C), where ε_CBM is the conduction band (CB) minimum energy of TiO₂, k is the Boltzmann constant, N_D is the donor density of TiO₂, and N_C is the effective density of states near the CBM of TiO₂. The ε_F,TiO2 value is estimated to be −5.1 ± 0.1 eV by substituting ε_CBM = −5.0 ± 0.1 eV (ref. 62) and N_C = 2.5 × 10¹⁹ cm⁻³(ref. 63) for the usual N_D value of 10¹⁷ cm⁻³. When the Au NP having a large work function (Φ_{W, Au} = 5.31–5.47 eV)⁵⁹ comes into direct contact with TiO₂ (Φ_W,TiO2 = 5.1 ± 0.1 eV), the interfacial electron transfer proceeds from TiO₂ to the Au NP until their Fermi energies are equal. Consequently, a Schottky junction is formed between them at the equilibrium state in the dark.³¹

The IPCE action spectra indicate that the WOR on the Au/TiO₂ plasmonic electrode in the present systems is mainly driven via the LSPR-induced hot electron transfer (HET) mechanism (Fig. 5).^22,26 When the Au NPs of Au/TiO₂ are excited by photons with energy (ħω) larger than the Schottky barrier (Δε), a number of the hot electrons are injected into the CB of TiO₂.²⁵ In the present cases, this energetic requirement is indeed fulfilled because the photocurrent flows at ħω > 1.38 eV (Fig. 3c), which is sufficiently larger than the Δε of ∼0.5 eV reported for the Au/TiO₂ system.³¹ The electrons injected into the CB of TiO₂ can be pulled away from the interface to be transported to the FTO electrode, and the holes are accumulated at the interface according to the potential gradient in the space charge layer. As a result of charge separation, the holes effectively oxidize water at the Au NP-TiO₂-liquid triphase interface.²⁶

Fig. 5 Action mechanism of the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA plasmonic electrodes for the WOR.

The photocurrent observed for the WOR (J_ph) is proportional to the efficiencies of a series of physical processes involving light harvesting (LHE), hot-electron injection (ϕ_HEI), and charge separation/transport (ϕ_CS–CT) (eqn (3)).⁴¹ Furthermore, the ϕ_HEI can be approximated using the Fowler theory (eqn (4)):³¹

J_ph ∝ LHE × ϕ_HEI × ϕ_CS–CT. (3)

ϕ_HEI(E) ≈C(ħω − Δε)²/ħω (4)

where E is the electrode potential, C is the constant, and Δε is the energy difference between the Fermi energy of the Au NP (ε_F) and the ε_CBM of TiO₂.

As indicated by the absorption spectra (Fig. 3a), the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA have LHEs significantly larger than that of the CR-Au/A-TiO₂ NPLA, which are induced by the HEPI junctions in the former systems.²⁴ Also, the ϕ_HEI in the LSPR-induced HET mechanism is proportional to the square of the absolute value of the electric field, that is, the EF.^24,64 The EF values of the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA are several tens of times larger than that of the CR-Au/A-TiO₂ NPLA. Furthermore, the high quality of the Au NP-TiO₂ interface in the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA contributes to the increase in the ϕ_HEI.⁶⁵ Therefore, the photocurrents of the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA electrodes far surpass that of the CR-Au/A-TiO₂ NPLA electrode and are ascribable to the high LHE and extraordinarily large EF values in addition to the higher quality of the interface. However, the large difference in the photocurrent between the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA electrodes cannot be explained only in terms of these factors.

Previously, the electron affinity was determined to be 4.9 eV for R-TiO₂ and 5.1 eV for A-TiO₂ by means of photoemission electron microscopy.⁶² Thus, the energy barrier for the HET from Au NPs to TiO₂ in the DP-Au/A-TiO₂ NPLA electrode is ∼0.2 eV lower than that in the DP-Au/R-TiO₂ NWA electrode regardless of the electrode potential (Fig. 5). In addition, A-TiO₂ possesses an electron mobility of 4–20 cm² V⁻¹ s⁻¹ at 300 K, which is much larger than that for R-TiO₂ (0.1 cm² V⁻¹ s⁻¹ at 300 K).³² In the DP-Au/A-TiO₂ NPLA electrode, the electrons injected into the CB of TiO₂ can be smoothly separated from the Au NPs and transported to the electron-collecting FTO electrode due to the large electron mobility of A-TiO₂.³² As a result, the hot holes accumulated in the Au NPs are effectively concentrated around the Au NP–TiO₂–solution triphase interface to be used for the WOR with the assistance of the catalytic activity,^25,66 which leads to a drastic enhancement of the photocurrent through lowering the R_ct. Consequently, the plasmonic activity of the DP-Au/A-TiO₂ NPLA, which is much superior to that of the DP-Au/R-TiO₂ NWA electrode, can stem from its higher HET efficiency and electric conductivity.

Furthermore, it is worth noting in the electrode system that Δε in eqn (2) is a function of the electrode potential (E vs. standard hydrogen electrode, SHE) or the Fermi energy of the electrode (ε_F = −e(E + 4.44) vs. under vacuum). As a result of the increase in E, the Δε increases, while the charge separation is enhanced due to the increase in potential gradient and width of the space charge layer. These considerations predict the presence of the optimum potential in the Au/TiO₂ plasmonic electrode. Actually, the maximum photocurrents are observed in the J–E curves at 0.83 V in the DP-Au/R-TiO₂ NWA electrode and 0.96 V in the DP-Au/A-TiO₂ NPLA electrode (Fig. 3b).

In a particulate photocatalytic system, Chen and Tan have recently reported that a nanohybrid plasmonic photocatalyst consisting of Pt cuboctahedra and radial WO₃ mesocrystals provides an apparent quantum efficiency greater than 7% at λ_ex = 680 nm for the overall water splitting, which is much larger than that of spherical Pt NP-loaded WO₃.⁶⁷ This remarkable activity can also arise from the LSPR-induced strong local electric field around the interface between the faceted Pt NPs and WO₃. Also, this study points to the importance of the high catalytic activities of Pt NPs for the hydrogen evolution reaction and the oxygen evolution reaction. In this case, it should be noted that both of these catalytic activities can only work simultaneously through the separation of the functions of the Pt NPs as oxidation and reduction sites due to the large absorption of the Pt cuboctahedra and the hierarchical structure of WO₃.⁶⁸

In summary, this study aims at enhancing the activity of the prototype plasmonic electrode of Au/TiO₂ for the PEC WOR without changing the composition of each component or surface modification by promoters. The t-Oh Au NPs with d_Au < 10 nm are formed on the A-TiO₂ NPLA with dominant (001) surfaces through the (111)_Au//(001)_A-TiO2 HEPI junction by the DP process. The resulting DP-Au/A-TiO₂ NPLA possesses features including the faceting of the Au NPs and a high-quality interface. The DP-Au/A-TiO₂ NPLA electrode has been found to provide an IPCE of 0.39% at λ = 600 nm for the PEC WOR, which is greater than the values reported so far for the usual Au/TiO₂ electrodes by more than one order of magnitude. The remarkable activity of the DP-Au/A-TiO₂ NPLA plasmonic electrode is ascribable to the intense local electric field generated around the edges and corners of the faceted Au NPs, the high-quality interface between Au NPs and TiO₂, and the large electron mobility in A-TiO₂, which enhance the physical and chemical processes in the plasmonic WOR. Furthermore, we shed light on the origin of the large difference in the plasmonic activity for the WOR between the DP-Au/A-TiO₂ NPLA and DP-Au/R-TiO₂ NWA, and the unique features in the photocurrent–potential curves in terms of a potential-dependent Fowler model. This marked progress in the performance of the prototype Au/TiO₂ plasmonic electrode could contribute to its practical use for solar-to-fuel energy conversion.

Author contributions

S. N., Y. M., H. S. and T. S. prepared and characterized the electrodes, and carried out PEC experiments and data analysis; M. F. performed MD and FDTD calculations; and H. T. supervised the experimental work and data analysis.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

The authors acknowledge K. Miyazaki, R. Kuma, T. Sento, and M. Shima (Nippon Shokubai Co., Ltd) for helpful discussions. This work was financially supported by JSPS KAKENHI Grant-in-Aid for Scientific Research (C) no. 21K05236, the Futaba Foundation, Nippon Sheet Glass Foundation for Materials Science and Engineering, Sumitomo Foundation, Takahashi Industrial and Economic Research Foundation, and the Kato Foundation for the Promotion of Science.

References

1. A. Fujishima and K. Honda, Electrochemical photolysis of water at a semiconductor electrode, Nature, 1972, 238, 37–38.
2. T. Hisatomi and K. Domen, Reaction systems for solar hydrogen production via water splitting with particulate semiconductor photocatalysts, Nat. Catal., 2019, 2, 387–399.
3. M. W. Kanan and D. G. Nocera, In situ formation of an oxygen-evolving catalyst in natural water containing phosphate and Co²⁺, Science, 2008, 321, 1072–1075.
4. L. Zhang, L. Sun, Z. Guan, S. Lee, Y. Li, H. D. Deng, Y. Li, N. L. Ahlborg, M. Boloor, N. A. Melosh and W. C. Chueh, Quantifying and elucidating thermally enhanced minority carrier diffusion length using radius-controlled rutile nanowires, Nano Lett., 2017, 17, 5264–5272.
5. M. Xiao, R. Jiang, F. Wang, C. Fang, J. Wang and J. C. Yu, Plasmon-enhanced chemical reactions, J. Mater. Chem. A, 2013, 1, 5790–5805.
6. K. Ueno and H. Misawa, Surface plasmon-enhanced photochemical reactions, J. Photochem. Photobiol., C, 2013, 15, 31–52.
7. X. Lang, X. Chen and J. Zhao, Heterogeneous visible light photocatalysis for selective organic transformations, Chem. Soc. Rev., 2014, 43, 473–486.
8. C. Wang and D. Astruc, Nanogold plasmonic photocatalysis for organic synthesis and clean energy conversion, Chem. Soc. Rev., 2014, 43, 7188–7216.
9. R. Jiang, B. Li, C. Fang and J. Wang, Metal/semiconductor hybrid nanostructures for plasmon-enhanced applications, Adv. Mater., 2014, 26, 5274–5309.
10. H. Cheng, K. Fuku, Y. Kuwahara, K. Mori and H. Yamashita, Harnessing single-active plasmonic nanostructures for enhanced photocatalysis under visible light, J. Mater. Chem. A, 2015, 3, 5244–5258.
11. X. Li, J. Zhu and B. Wei, Hybrid nanostructures of metal/two-dimensional nanomaterials for plasmon-enhanced applications, Chem. Soc. Rev., 2016, 45, 3145–3187.
12. Y. Zhang, S. He, W. Guo, Y. He, J. Huang, J. R. Mulcahy and W. D. Wei, Surface-plasmon-driven hot electron photochemistry, Chem. Rev., 2018, 118, 2927–2954.
13. N. Wu, Plasmonic metal–semiconductor photocatalysts and photoelectrochemical cells: a review, Nanoscale, 2018, 10, 2679–2696.
14. A. Gelle, T. Jin, L. de la Garza, G. D. Price and L. V. Besteiro, Applications of plasmon-enhanced nanocatalysis to organic transformations, Chem. Rev., 2020, 120, 986–1041.
15. S. Luo, X. Ren, H. Lin, H. Song and J. Ye, Plasmonic photothermal catalysis for solar-to-fuelconversion: current status and prospects, Chem. Sci., 2021, 12, 5701–5719.
16. M. Sayed, J. Yu, G. Liu and M. Jaroniec, Non-noble plasmonic metal-based photocatalysts, Chem. Rev., 2022, 122, 10484–10537.
17. R. S. Haider, S. Wang, Y. Gao, A. S. Malik, N. Ta, H. Li, B. Zeng, M. Dupuis, F. Fan and C. Li, Boosting photocatalytic water oxidation by surface plasmon resonance of Ag_xAu_1-x alloy nanoparticles, Nano Energy, 2021, 87, 106189.
18. H. Li, S. Wang, M. Wang, Y. Gao, J. Tang, S. Zhao, H. Chi, P. Zhang, J. Qu, F. Fan and C. Li, Enhancement of Plasmon-Induced Photoelectrocatalytic Water Oxidation over Au/TiO₂ with Lithium Intercalation, Angew. Chem., Int. Ed., 2022, 61, e202204272.
19. R. Negishi, S. Naya, H. Kobayashi and H. Tada, Gold(core)-lead(shell) nanoparticle-loaded titanium(IV) oxide prepared by underpotential photodeposition: Plasmonic water oxidation, Angew. Chem., Int. Ed., 2017, 56, 10347–10351.
20. M. Okazaki, Y. Suganami, N. Hirayama, H. Nakata, T. Oshikiri, T. Yokoi, H. Misawa and K. Maeda, Site-selective deposition of a cobalt cocatalyst onto a plasmonic Au/TiO₂ photoanode for improved water oxidation, ACS Appl. Energy Mater., 2020, 3, 5142–5146.
21. P. A. DeSario, J. J. Pietron, D. E. DeVantier, T. H. Brintlinger, R. M. Stroud and D. R. Rolison, Plasmonic enhancement of visible-light water splitting with Au–TiO₂ composite aerogels, Nanoscale, 2013, 5, 8079–8083.
22. T. Onishi, M. Teranishi, S. Naya, M. Fujishima and H. Tada, Electrocatalytic effect on the photon−to−current conversion efficiency of gold−nanoparticle−loaded titanium(IV) oxide plasmonic electrode for water oxidation, J. Phys. Chem. C, 2020, 124, 6103–6109.
23. S.-F. Hung, F.-X. Xiao, Y.-Y. Hsu, N.-T. Suen, H.-B. Yang, H. M. Chen and B. Liu, Iridium oxide-assisted plasmon-induced hot carriers: Improvement on kinetics and thermodynamics of hot carriers, Adv. Energy Mater., 2016, 6, 1501339.
24. H. Tada, Rational design for gold nanoparticle−based plasmonic catalysts and electrodes for water oxidation towards the artificial photosynthesis, Dalton Trans., 2022, 51, 3383–3393.
25. M. Teranishi, M. Wada, S. Naya and H. Tada, Size-dependence of the activity of gold nanoparticle-loaded titanium(IV) oxide plasmonic photocatalyst for water oxidation, ChemPhysChem, 2016, 17, 2813–2817.
26. Y. Nishijima, K. Ueno, Y. Kotake, K. Murakoshi, H. Inoue and H. Misawa, Near-infrared plasmon-assisted water oxidation, J. Phys. Chem. Lett., 2012, 3, 148–1252.
27. S. Wang, Y. Gao, S. Miao, T. Liu, L. Mu, R. Li, F. Fan and C. Li, Positioning the water oxidation reaction sites in plasmonic photocatalysts, J. Am. Chem. Soc., 2017, 139, 11771–11778.
28. K. Wu, J. Chen, J. R. McBride and T. Lian, Efficient hot-electron transfer by a plasmon-induced interfacial charge-transfer transition, Science, 2015, 349, 632–635.
29. J. Ma and S. Gao, Plasmon-induced electron-hole separation at the Ag/TiO₂(110) interface, ACS Nano, 2019, 13, 13658–13667.
30. M. Yoshii, H. Kobayashi and H. Tada, Sub-bandgap excitation-induced electron injection from CdSe quantum dots to TiO₂ in a directly coupled system, ChemPhysChem, 2015, 16, 1846–1851.
31. M. W. Knight, H. Sobhami, P. Nordlander and M. J. Halas, Photodetection with active optical antennas, Science, 2011, 332, 702–704.
32. M. Voïnov and J. Augustynski, Heterogeneous Photocatalysis, ed. M. Schiavello, John Wiley & Sons, Baffins Lane, 1997.
33. B. O'Regan and M. Grätzel, A low-cost, high-efficiency solar cell based on dye-sensitized colloidal TiO₂ films, Nature, 1991, 353, 737–740.
34. M. M. Lee, J. Teuscher, T. Miyasaka, T. N. Murakami and H. J. Snaith, Efficient hybrid solar cells based on meso-superstructured organometal halide perovskites, Science, 2012, 338, 643–647.
35. E. Kowalska, R. Abe and B. Ohtani,  Visible light-induced photocatalytic reaction of gold-modified titanium(IV) oxide particles: action spectrum analysis, Chem. Commun., 2009, 45, 241–243.
36. E. Kowalska, O. O. P. Mahaney and B. Ohtani, Visible-light-induced photocatalysis through surface plasmon excitation of gold on titania surfaces, Phys. Chem. Chem. Phys., 2010, 12, 2344–2355.
37. K. Kimura, S. Naya, Y. Jin-nouchi and H. Tada, TiO₂ crystal form-dependence of the Au/TiO₂ plasmon photocatalyst's activity, J. Phys. Chem. C, 2012, 116, 7111–7117.
38. S. Naya, K. Kimura and H. Tada, One-step selective aerobic oxidation of amines to imines by gold nanoparticle-loaded rutile titanium(IV) oxide plasmon photocatalyst, ACS Catal., 2013, 3, 10–13.
39. J. B. Priebe, J. Radnik, A. J. J. Lennox, M.-M. Pohl, M. Karnahl, D. Hollmann, K. Grabow, U. Bentrup, H. Junge, M. Beller and A. Bruckner, Solar hydrogen production by plasmonic Au-TiO₂ catalsyt: impact of synthesis protocol and TiO₂ phase on charge transfer efficiency and H₂ evolution rates, ACS Catal., 2015, 5, 2137–2148.
40. S. Naya, T. Niwa, T. Kume and H. Tada, Visible-light-induced electron transport from small to large nanoparticles in bimodal gold nanoparticle-loaded titanium(IV) oxide, Angew. Chem., Int. Ed., 2014, 53, 7305–7309.
41. S. Naya, A. Akita, Y. Morita, M. Fujishima and H. Tada, Crystallographic interface control of the plasmonic photocatalyst consisting of gold nanoparticles and titanium(IV) oxide, Chem. Sci., 2022, 13, 12340–12347.
42. S. Feng, J. Yang, H. Zhu, M. Liu, J. Zhang, J. Wu and J. Wan, Synthesis of single crystalline anatase TiO₂ (001) tetragonal nanosheet-array films fluorine-doped tin oxide substrate, J. Am. Ceram. Soc., 2011, 94, 310–315.
43. B. Liu and E. S. Aydil, Growth of oriented single-crystalline rutile TiO₂ nanorods on transparent conducting substrates for dye-sensitized solar cells, J. Am. Chem. Soc., 2009, 131, 3985–3990.
44. R. Zenella, S. Giorgio, C. R. Henry and C. Louis, Alternative methods for the preparation of gold nanoparticles supported on TiO₂, J. Phys. Chem. B, 2002, 106, 7634–7642.
45. A. Dieguez, A. Romano-Rodriguez, A. Vila and J. R. Morante, The complete Raman spectrum of nanometric SnO₂ particles, J. Appl. Phys., 2001, 90, 1550–1557.
46. O. Frank, M. Zukalova, B. Laskova, J. Kürti, J. Koltai and L. Kavan, Raman spectra of titanium dioxide (anatase, rutile) with identified oxygen isotopes (16, 17, 18), Phys. Chem. Chem. Phys., 2012, 14, 14567–14572.
47. A. Akita, H. Kobayashi and H. Tada, Action of chloride ions as a habit modifier in the hydrothermal crystal growth of rutile TiO₂ nanorod from SnO₂ seed crystal, Chem. Phys. Lett., 2020, 761, 138003.
48. H. G. Yang, C. H. Sun, S. Z. Qiao, J. Zou, G. Liu, S. C. Smith, H. M. Chen and G. Q. Liu, Anatase TiO₂ single crystals with a large percentage of reactive facets, Nature, 2008, 453, 638–641.
49. S. Liu, J. Yu and M. Jaroniec, Tunable photocatalytic selectivity of hollow TiO₂ microspheres composed of anatase polyhedra with exposed {001} facets, J. Am. Chem. Soc., 2010, 132, 11914–11916.
50. D. Holec, P. Dumitraschkewitz, D. Vollath and F. D. Fischer, Surface energy of Au nanoparticles depending on their size and shape, Nanomaterials, 2020, 10, 484.
51. G. Liu, H. G. Yang, J. Pan, Y. Q. Yang, G. Q. Lu and H.-M. Cheng, Titanium dioxide crystals with tailored facets, Chem. Rev., 2014, 114, 9559–9612.
52. F. Cosandey, L. Zhang and T. E. Madey, Effect of substrate temperature on the epitaxial growth of Au on TiO₂(110), Surf. Sci., 2001, 474, 1–13.
53. W. Yuan, B. Zhu, K. Fang, X.-Y. Li, T. W. Hansen, Y. Ou, H. Yang, J. B. Wagner, Y. Gao, Y. Wang and Z. Zhang, In situ manipulation of the active Au-TiO₂ interface with atomic precision during CO oxidation, Science, 2021, 371, 517–521.
54. K. Tanaka, S. Tanuma, K. Dohmae, Y. Nagoshi and A. Nisawa, X-ray photoelectron spectroscopy, ed. T. Sawada, S. Tanuma and K. Tanaka, Maruzen, Tokyo, 1998.
55. H. Kisch, Semiconductor Photocatalysis-Principles and ApplicationsWiley-VCH, Weinheim, 2015.
56. H. Hovel, S. Fritz, A. Hilger and V. Kreibig, Width of cluster plasmon resonances: Bulk dielectric functions and chemical interface damping, Phys. Rev. B: Condens. Matter Mater. Phys., 1993, 48, 18178–18188.
57. A. Naldoni, T. Montini, F. Malara, M. M. Mroz, A. Beltram, T. Virgili, C. L. Boldrini, M. Marelli, I. Romero-Ocana, J. J. Delgado, V. Dal Santo and P. Fornasiero, Hot electron collection on brookite nanorods lateral facets for plasmon-enhanced water oxidation, ACS Catal., 2017, 7, 1270–1278.
58. L. Liu, P. Li, B. Adisak, S. Ouyang, N. Umezawa, J. Ye, R. Kodiyath, T. Tanabe, G. V. Ramesh, S. Ueda and H. Abe, Gold photosensitized SrTiO₃ for visible-light water oxidation induced by Au interband transitions, J. Mater. Chem. A, 2014, 2, 9875–9882.
59. Handbook of Chemistry and Physics, ed. D. R. Lide, CRC Press, Boca Raton, 83rd edn, 2002.
60. U. Kreibig and M. Vollmer, Optical properties of metal clusters, Springer, Berlin, 1996.
61. W. Yuan, D. Zhang, Y. Ou, K. Fang, B. Zhu, H. Yang, T. W. Hansen, J. B. Wagner, Z. Zhang, Y. Gao and Y. Wang, Direct in situ TEM visualization and insight into the facet-dependent sintering behaviours of gold on TiO₂, Angew. Chem., Int. Ed., 2018, 57, 16827–16831.
62. G. Xiong, R. Shao, T. C. Droubay, A. G. Joly, K. M. Beck, S. A. Chambers and W. P. Hess, Photoemission electron microscopy of TiO₂ anatase films embedded with rutile nanocrystals, Adv. Funct. Mater., 2007, 17, 2133–2138.
63. H. Tang, K. Prasad, R. Sanjinès, P. E. Schmid and F. Lévy, Electrical and optical properties of TiO₂ anatase thin films, J. Appl. Phys., 1994, 75, 2042–2047.
64. A. O. Govorov, H. Zhang, H. V. Demir and Y. K. Gun'ko, Photogeneration of hot plasmonic electrons with metal nanocrystals: Quantum description and potential applications, Nano Today, 2014, 9, 85–101.
65. S. Naya, T. Kume, R. Akashi, M. Fujishima and H. Tada, Red−Light−Driven Water Splitting by Au(Core)−CdS(Shell) Half−Cut Nanoegg with Heteroepitaxial Junction, J. Am. Chem. Soc., 2018, 140, 1251–1254.
66. E. Thimsen, F. Le Formal, M. Grätzel and S. C. Warren, Influence of plasmonic Au nanoparticles on the photoactivity of Fe₂O₃ electrodes for water splitting, Nano Lett., 2011, 11, 35–43.
67. Q. Chen and Y. Tan, Enhanced plasmonic absorption of Pt cuboctahedra-WO₃ nanohybrids used as visible light photocatalysts for overall water splitting, Nano Res., 2023, 16, 5919–5928.
68. R. Kojima, S. Naya and H. Tada, Three-dimensional plasmonic photocatalyst consisting of faceted gold nanoparticles and radial titanium(IV) oxide heteromesocrystal, J. Phys. Chem. C, 2023, 127, 3478–3485.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: https://doi.org/10.1039/d4nr01013b

---