Visible Light-Driven H2 Production over Highly Dispersed Ruthenia on Rutile TiO2 Nanorods

1 Visible Light-Driven H2 Production over Highly Dispersed Ruthenia on Rutile TiO2 Nanorods Thuy-Duong Nguyen-Phan,1 Si Luo,1,2 Dimitriy Vovchok,1,2 Jordi Llorca,3 Jesús Graciani,4 Shawn Sallis,5 Wenqian Xu,6 Jianming Bai,7 Louis F. J. Piper,5 Dmitry E. Polyansky,1 Etsuko Fujita,1 Sanjaya D. Senanayake,1 Dario J. Stacchiola,1* and José A. Rodriguez1,2* 1 Chemistry Department, and 7 Photon Sciences Division, Brookhaven National Laboratory, Upton, NY 11973, US 2 Department of Chemistry, Stony Brook University, Stony Brook, NY 11790, US 3 Institute of Energy Technologies and Centre for Research in NanoEngineering, Universitat Politècnia de Catalunya, Diagonal 647, 08028 Barcelona, Spain 4 Department of Physical Chemistry, University of Seville, E-41012, Seville, Spain 5 Materials Science & Engineering, Department of Physics, Applied Physics and Astronomy, Binghamton University, Binghamton, NY 13902, US 6 Advanced Photon Source, Argonne National Laboratory, Argonne, Illinois 60439, US 2 ABSTRACT The immobilization of miniscule quantities of RuO2 (~ 0.1%) onto one-dimensional (1D) TiO2 nanorods (NRs) allows H2 evolution from water under the irradiation of visible light. Rod-like rutile TiO2 structures, exposing preferentially (110) surfaces, are shown to be critical for the deposition of RuO2 to enable photocatalytic activity in the visible region. This performance is rationalized based on fundamental experimental studies and theoretical calculations, demonstrating that RuO2(110) grown as 1D nanowires on rutile TiO2(110), which occurs only at extremely low loads of RuO2, leads to the formation of a heterointerface that efficiently adsorbs visible light. KEYWORDS Titanium; Ruthenium; H2 production; Water splitting; Heterojunction. 3 INTRODUCTION Hydrogen obtained from carbon-free sources is an attractive alternative clean fuel that could potentially offer solutions to global warming through the reduction of CO2 emissions. The splitting of water into H2 and O2 utilizing solar energy is regarded as one of the most promising and sustainable technologies to produce hydrogen.1-3 Although titanium dioxide (TiO2) is the most common semiconductor oxide photocatalyst,3-6 its practical application is limited because TiO2 absorbs only a small fraction of solar light and the photogenerated electrons/holes recombine rapidly.4,5 Anchoring or coupling metal or metal oxide nanoparticles (NPs) onto TiO2 has been used to either enhance its solar-conversion efficiency by retarding the charge carrier recombination or extending its response toward the visible light region stemming from band level alignment.6,7 Precious metals possessing much larger work function than TiO2,6 such as Pt, Pd, Ru and Au, have been considered as the best candidates to suppress the charge carrier recombination and facilitate the charge transfer owing to the formation of Schottky barriers at the heterogeneous nanoscale junction and the localized surface plasmon resonance.3,6,7 However, such a method of metallization is uneconomical if it demands large amounts of precious metals. Since RuO2 belongs to the class of d-band metallic conducting transition metal oxides, its electrical resistivity is two-fold higher than that of metallic ruthenium.8 The intraband transitions and localized surface plasmon resonance of RuO2 also play a vital role in its optical response and photoactivity.9-11 Among several polymorphs of TiO2, rutile is the thermodynamically most stable phase, but the moderate-temperature post treatment usually required to achieve highly crystalline rutile phase, results in large particle sizes and nonporosity.6,5 Coupling rutile TiO2 with RuO2 has been widely studied in numerous catalytic reactions due to their lattice matching and strong metal-support interaction.12-17 Seki14 claimed that RuO2/rutile-TiO2 catalyst exhibited high catalytic activity and thermal stability 4 toward HCl oxidation to Cl2 due to highly dispersed, ultrafine RuO2 crystallites, approximately a single unit-cell thickness, that cover the surface of TiO2 primary particle with strong interaction between RuO2 and TiO2. Xiang et al.15 demonstrated that the redistribution or spreading of RuO2 occurred from a spherical shape (~ sub 2 nm) to epitaxial island layers along the TiO2 surfaces during the chlorination reaction, forming a unique heterostructure at the interface. Lin et al.17 elucidated that among several oxide supports including anatase TiO2, rutile TiO2, rutile SnO2, Al2O3 and SiO2, rutile TiO2 was the most appropriate support to enhance the catalytic activity of N2O reduction. Similarly, Jiao et al.16 employed thermal treatment to transform RuO2 nanoclusters (ca. 2 nm) into an epitaxial layer on the surface of both anatase and rutile TiO2, relying on the interfacial atom arrangement match between these two oxides that led to more significant exposure of RuO2 (110) facets and thus, promoting photocatalytic CO oxidation. Several key strategies are important including: choosing an appropriate support for highly dispersing metallic Ru or RuO2 NPs to achieve optimal photocatalytic performance; improve the stability as well as decrease the amount of costly precursor being utilized but most notably, aim to preserve all specific properties of each component. In spite of its promising catalytic features, very little attention has been paid so far concerning the explanation of the photocatalytic properties of the RuO2/rutile-TiO2 system. Most reported studies employed RuO2 as a cocatalyst to facilitate the gas production under UV irradiation.3,6,18-21 Recently, a first attempt toward H2 production from water under UV excitation over RuO2/anatase-TiO2 nanocomposites was conducted, where an upward band bending at the RuO2-TiO2 interface has been proposed to explain the good performance of H2 evolution.21 In addition, density functional calculations performed on RuO2/TiO2(110) heterostructures for photocatalysis application found that the interface between RuO2 and rutile TiO2 strongly affects the atomic and electronic properties of both oxides due to a strong 5 charge accumulation at the interface.22 Upward band bending at the interface could be observed when introducing oxygen vacancies at the interface and subsequently, the electron accumulation at the interface creates a strong internal electric field, leading to efficient separation of photoinduced electron-hole pairs during a photocatalytic process.22 According to their calculation, the adsorption energy is lowered by 0.34 eV and the distance between H and bridging oxygen is shortened by 0.18 Å on RuO2/TiO2(110) in comparison with that on TiO2(110) surface, therefore facilitating the splitting of water. No attempt has been made to investigate the metal-support interaction between rutile RuO2 and rutile TiO2 support as well as to explain their photocatalytic properties in targeting H2 production. We have synthesized a RuO2/TiO2 heterostructure by immobilizing minute quantities of RuO2 onto 1D TiO2 nanorods (NRs). Complementary characterization techniques have been applied to gain insight into the promotion effect of the heterostructure on the visible-light-responsive activity combined with DFT calculations of RuO2 on rutile TiO2(110). EXPERIMENTAL Synthesis of RuO2/TiO2 heterostructures All the chemicals were purchased from Aldrich and used without purification. Typically, a mixture of 7 mL of titanium n-butoxide and 7 mL of hydrochloric acid (35 wt%) was hydrothermally treated for 11 h at 105 C. The white precipitate was carefully collected and washed several times with deionized water, thrice by aqueous 0.1 M NH4OH solution, and finally with deionized water to entirely remove chlorine contaminants. After drying at 80 C overnight, the blank TiO2 rods (denoted as ‘TiO2’) were obtained after calcined the as-syn TiO2 rods in air at 200 C for 2h. Subsequently, 0.2 g of as-syn TiO2 was re-dispersed in 20 mL of deionized water at 90 C for 30 min. The impregnation of RuO2 was carried out by drop-wise addition of an 6 appropriate concentration of 5 mL aqueous RuCl3.xH2O solution at 90 C. The fine powder after evaporation was sequentially washed with deionized water, aqueous 0.1 M NH4OH solution, and deionized water to entirely remove chlorine contaminants, and collected by high-speed centrifugation (20000 rpm). The powder was dried at 80 C overnight and then calcined in air at 200 C for 2h, denoted as ‘xRuTi’ where x = 0.01, 0.1, 0.25, and 0.5 wt% of RuO2. For comparison, bulk rutile TiO2 particle was prepared by annealed commercial TiO2 P25 (Degussa) at 800 °C for 4 h and named as ‘TiO2_p’. Characterizations Synchrotron X-ray powder diffraction (SXPD) patterns were collected at beamline 17-BM- B ( = 0.72768 Å) of the Advanced Photon Source at Argonne National Laboratory. 5 mg of powder were loaded into a 0.9-mm-ID kapton capillary and two-dimensional diffraction patterns were collected by a Perkin Elmer amorphous silicon detector. The data acquisition was integrated by QXRD while the crystalline phase identification, composition and lattice parameters were subsequently analyzed by Rietveld refinement using TOPAS 4.1 program. High-resolution transmission electron microscopy (HR-TEM) and scanning transmission electron microscopy (STEM) images were recorded on Tecnai G2 F20 S-TWIN transmission electron microscope equipped with a field emission electron source. It was operated at 200 kV while the point-to-point resolution and the resolution between lines were 0.24 and 0.10 nm, respectively. The powder samples were deposited from alcohol suspensions onto holeycarbon Cu grids. Brunauer-Emmett-Teller (BET) specific surface areas were determined by N2 adsorption/desorption at 77 K using an Altimira AMI-300ip instrument. The powders were degassed at 150 C to remove all surface-adsorbed contaminants prior to measurements. High-resolution X-ray photoemission spectra (XPS) of Ti 2p, O 1s, Ru 3d and Ru 3p were recorded on a laboratory-based monochromated Al K source with a hemispherical analyzer. The core-level and ultraviolet photoelectron spectra (UPS) were measured with a constant 7 pass energy of 23.5 eV, which corresponds to an instrumental resolution of 0.51 eV obtained from analyzing both the Au 4f7/2 and Fermi edge of Au foil. Further characterization of the materials was also conducted using several techniques available at the Center for Functional Nanomaterials (CFN) at BNL. Raman spectroscopy was performed on a WiTec Alpha combination microscope at room temperature with 633 nm laser as an excitation source. UV-Vis diffuse reflectance (DRS) measurements were collected at room temperature by PerkinElmer Lambda 950 spectrometer equipped with an integrating sphere assembly. Visible-light-driven H2 production measurements The photocatalytic hydrogen evolution was conducted in a closed gas circulation and evacuation system. Powder catalyst were suspended in aqueous methanol solution (20 vol%) in a sealed quartz cell. After evacuation and Ar purging several times, the reactor was sideirradiated by a 150 W Xenon arc lamp equipped with a CuSO4 filter and 400 nm long pass filter (400 nm <  < 625 nm) under magnetic stirring. The reaction temperature was maintained at 293 K by continuous water circulation. The evolved gases were determined by gas chromatography (GC Agilent 6890N) equipped with FID and TCD detectors using Ar as the carrier gas. Detailed experiment and calculation of apparent quantum efficiency is described in Supporting Information. Computational methods We built our theoretical model based on the previously published STM images of RuOx supported on TiO2(110) surface.23 The nanostructured ruthenia on titania was modelled as follows: (i) TiO2(110) surface consisted of four O-Ti-O three-layers, keeping the two of the bottom fixed at the optimized bulk positions, allowing a vacuum region of 15 Å between repeated slabs; (ii) in order to achieve an isolated wire-like RuO2(110) nanostructure a (3x1) and (3x6) surface models of the titania support were used for optics calculation and density of 8 states respectively; (iii) a full-relaxed three atomic layer width (O-Ru-O) wire was coupled to the titania support according to Yang et al.’s report23. We performed periodic DFT calculations using the Perdew-Wang 91 (PW91) functional24 for the exchange-correlation potential. The effect of the core electrons on the valence states was represented with the projector-augmented wave (PAW) approach,25 as implemented in the Vienna ab-initio simulation package (VASP 5.3),26,27 with the valence states defined for each atom as Ti(3s,3p,3d,4s), Ru(4s,4p,4d,5s) and O(2s,2p) electrons, while the remaining electrons were kept frozen as core states. The valence electronic states are expanded in a basis of plane waves with a cutoff of 400 eV for the kinetic energy. In order to account for eventual reduction of the titania support (occupation of the Ti 3d states) a Hubbard-like U term was used, (GGA+U), according to the Dudarev et al.’s implementation,28 which makes use of an effective parameter Ueff. We took a value Ueff = 4.5 eV satisfactorily used in our previous work dealing with supported cerium oxide particles on titania.29 Calculations were performed at the  point of the Brillouin zone for the (3x6) surface model and with a 1x4x1 k-points grid for the (3x1) model. RESULTS AND DISCUSSION Visible-light-responsive H2 production over RuO2/TiO2 heterostructures Figure 1 shows the visible light-driven hydrogen production performance of xRuTi heterostructured photocatalysts as a function of RuO2 content. The time profiles in Figure 1A demonstrates the stable, constant amounts of H2 produced over all the rod-like RuO2/TiO2 heterostructured photocatalysts under the irradiation of visible light (400 nm <  < 625 nm). Bulk rutile TiO2 particle (TiO2_p sample) is completely inactive for H2 production with visible light irradiation, while TiO2 nanorods present a moderate activity, producing 33 mol of H2 gas per gram catalyst per hour. This observation is consistent with recent report that sub-10 nm rutile NPs exhibited a significant improvement in photocatalytic activity for 9 visible-light-driven water splitting compared to the bulk oxide.30 It is worth noting that the deposition of RuO2 onto TiO2 rods significantly enhances the H2 evolution efficiency regardless of RuO2 loading amount. As seen in Figure 1B, the activity trend follows a volcano shape where the mass-normalized rate of H2 production over all xRuTi heterostructure materials is significantly higher than that over TiO2 NRs with an optimal RuO2 content of 0.1 wt% (425 mol g-1 h-1). The photocatalytic performance under full UV- visible irradiation (310-625 nm) followed the same trend, as shown in Figure S1 (Supporting Information). The activities normalized to specific surface areas and apparent quantum efficiencies are summarized in Table 1. The apparent quantum efficiencies over supported RuO2 are 4- to 12-times larger than that with TiO2 NRs alone. It is observed that increasing RuO2 loading amount to 0.25 and 0.5 wt% gradually decreases the H2 evolved rate regardless of light excitation source. This remarkable result is compared to the case of Pt, which is extensively used as a co-catalyst on TiO2. As shown in Figure S2 (Supporting Information), upon the UV irradiation, the H2 evolution activity over our Ru-based heterostructured photocatalyst is comparable to the platinized TiO2 NRs prepared by in situ photodeposition method. However, Pt does not give activity at all in the visible light region. A few studies have reported visible-light-driven photocatalytic activity over RuO2/TiO2 nanocomposites.31,32 Ismail et al.31 reported that the visible-light photoactivity enhancement toward methanol oxidation was ascribed to the incorporation of a small amount of Ru4+ into the anatase lattice and the formation of some structural defects on the surface, facilitating the interfacial charge transfer of electrons/holes, whereas the addition of RuO2 suppressed the UV-photonic efficiency of TiO2. 16 discussed above. The unparallel shift of both Ti 2p and O 1s confirms the interfacial reaction between RuO2 and TiO2, forming Ti-O-Ru bonds.35 The presence of ruthenium was clearly confirmed by a broad band representative of Ru 3d5/2 peak located at 279 - 282 eV as reported in C 1s + Ru 3d core level spectra in Figure 3C whereas the 3d3/2 doublet was superimposed by C 1s feature from carbonaceous impurities and carbonate species. The broad and weak spectrum in the inset indicates the formation of mixtures of Ru0 and oxidized Run+ species (n = 2 ~ 4).8,10,11,16,21,35 Due to the intrinsic submetallic property of RuO2,8- 11,21,31,33-35 these highly dispersed ruthenium species behave as quasi metallic contact materials to enhance both the conductivity and transfer of photoinduced holes from TiO2 valence band, further facilitating the charge separation, so that the electrons freely migrate from the conduction band of TiO2, reducing the protons and/or water to generate gaseous H2. 17 Figure 3. High resolution core-level XPS spectra of TiO2 and 0.1RuTi: Ti 2p + Ru 3p, O 1s, and C 1s + Ru 3d. 18 The local geometry of the prepared materials was further studied by electron microscopy. SEM images in Figure S6A and B (Supporting Information) shows that our pure TiO2 NRs were uniform nanorods aggregated in three-dimensional (3D) microsphere-like morphology. High-resolution TEM micrograph depicted in Figure S6C (Supporting Information) reveals that the highly crystalline rods had a rather constant diameter of about 5-10 nm and the length can be larger than several hundred nm. The well-resolved lattice fringe at 3.25 Å reveals that the highly crystalline NRs were grown along the [110] direction. After depositing 0.1 wt% RuO2 onto the TiO2 nanorods, the d-spacing of 3.25 Å corresponding to the preferred exposure of {110}-type planes of highly single-crystalline rutile structure was maintained (consistent with the FFT pattern in the inset) and the surface of TiO2 rods becomes more heterogeneous (Figure 4A). It is very difficult to recognize and identify the lattice fringes of RuO2 by HR-TEM mode stemming from lattice matching between rutile-like RuO2 and rutile TiO2, ultrafine particles and small loading amounts. Herein, the high angle annular dark field (HAADF) STEM image in Figure 4B indicates that the surface of the rod is rougher than TiO2 NRs only (Figure S6D, Supporting Information) and in some specific regions, 1 ~ 2 nm nanodots (even smaller) were sparsely decorated onto primary TiO2 NRs. The small nanoparticles are most likely a small fraction of sintered RuO2. The EDX spectrum in the inset confirmed the presence of Ti, O and Ru signals. However, the microstructure of the rod surface before and after loading RuO2 in Figure S7 (Supporting Information) obviously indicates an increase in surface roughness as well as the breakage into smaller fraction of rods. The structural analysis in Figure S8 (Supporting Information) visually revealed an obvious difference in the lattice fringes of TiO2 NRs before and after loading RuO2. Combining several factors, the heteroepitaxial growth of rutile RuO2 along rutile TiO2 NRs is proposed to occur because the lattice constants are almost identical. However, if the thickness of epitaxial layers or wires is too thin, approximately few atomic or sub-nanometer level, they 19 may not be observed microscopically. Similar behavior was studied by our group and other groups.12,13,23,36 We have observed in previous fundamental studies23,36 the growth of 1D wire-like rows of RuOx extending along the <001> direction of TiO2(110) single crystals in which each RuO2 wire with an apparent width of few angstrom covered three rows of TiO2(110) rows. Elevating temperature under UHV pressure resulted in the disappearance of RuOx wires and the formation of Ru nanoparticles. The RuO2/TiO2(110) surfaces were much more reactive towards CO oxidation, ethanol photo-oxidation and water dissociation than TiO2(110) or RuO2(110) due to the enhanced charge separation at the RuO2-TiO2 interface.23,36,47 The distribution of RuO2 on TiO2 NRs is obviously different from that over bulk TiO2 particles where the presence of abundant round-shaped RuO2 nanoparticles with an average diameter of 2-3 nm was extensively found in Figure 4C. One can clearly see a well-resolved fringe spacing of 3.25 Å that is indexed to (110) crystallographic plane of rutile TiO2 and the inset FFT pattern indicates the single-crystalline nature of the support. At high coverage of RuO2 nanoclusters whose d-spacing of 2.6 Å corresponds to (101) plane on the surface of bulk-like crystalline titania (Figure 4D), it resulted in much lower performance towards light-activated H2 production over bulk supported catalysts. 20 Figure 4. (A, B) HR-TEM and HAADF-STEM micrograph of 0.1 wt% RuO2 onto TiO2 NRs, respectively; (C, D) HR-TEM images of 0.1% RuO2 loaded on bulk TiO2 particles. The insets of A and D are corresponding FFT patterns and the inset of B is EDX spectrum. We carried out DFT studies to understand the electronic band structure of the heterojunction nanocomposites which strongly affects the photocatalytic water splitting performance. In Figure 5A, we show the density of states (DOS) of the full-relaxed (3x6) surface model for the system RuO2 nanowire (RuO2w) supported on TiO2(110) that we have 21 reported previously.23 The large band below 2 eV is mainly related to O 2p levels from both oxides. The states just below Fermi level consists primarily of Ru 4d states, with small contribution from the O 2p states of the oxygen atoms of RuO2. Therefore, the states observed in the UPS spectra just below the Fermi level are related to the supported RuO2, mainly from metallic ruthenium states. The states above the Fermi level consists firstly of Ru 4d levels and above 0.2-0.3 eV the 3d states of titanium appear strongly. Therefore d-d transitions between 4d states of the supported ruthenia along with 3d states of titania support may take place in the heterostructure with an abrupt interface. The total DOS is consistent with the experimental ultraviolet photoelectron spectroscopy valence band (UPS-VB) spectra shown in Figure 5B. The UPS-VB spectra of TiO2 NRs and 0.1RuO2/TiO2 heterostructure were measured in which two features centered at 4.8 eV and 6.6 eV are associated with the typical O 2p-Ti 3d structure of TiO2. The position of the valence band maximum (VBM) for TiO2 NRs locates at 1.87 eV (estimated by linear extrapolation), which is characteristic of a VB DOS of rutile TiO2. It should be noted that the RuO2 deposition induces a blue-shift to 1.75 eV of VB edge maximum energy and a tailing feature toward the vacuum level at ~ -0.6 eV confirms the submetallic character of RuO2. Such energy shifts in VBM and XPS core levels demonstrate the band gap narrowing, band bending, and the formation of strong builtin electric field at the heterojunction at the interface that was previously reported by DFT calculation,22 resulting in efficient separation of charge carriers during the photocatalytic process. 22 Figure 5. (A) Total and partial density of states (DOS) on O, Ru and Ti atoms of RuO2/TiO2(110) system (the Fermi level EF is set to energy zero); (B) experimental UPS-VB spectra of TiO2 NRs and 0.1RuO2/TiO2 heterostructure; (C) UV-Vis diffused absorption spectra of TiO2 and xRuTi series; and (D) representative band diagram of TiO2 and RuO2/TiO2 heterostructure. The inset of (C) is the Tauc’s plot for direct band gap determination. Figure 5C shows the UV-Vis diffused absorption spectra of TiO2 NRs and xRuTi heterostructures with various RuO2 contents. The absorption onset of bare TiO2 NRs located at ca. 410 nm due to O 2p-Ti 3d transitions. The edges were slightly red-shifted toward the longer wavelength region with increasing RuO2 loading amounts, indicating the narrowing of the band gap. Consequently, the band gaps (Eg) determined by Tauc’s plot in the inset are 23 summarized in Table 1. A gradual reduction of the band gap from 3.06 eV to 2.93 eV is observed with further increasing RuO2 contents up to 0.5 wt%. It is interesting to note that strong, broad absorption bands and long tails across the whole visible regime was observed in all oxide heterostructures which might be associated with several factors, i.e. the transition between Ti 3d and Ru 4d, band transition of RuO2 itself, defect level states, and/or the localized surface plasmon resonance mediated absorption of RuO2 nanodots.10,11 Two types of band-edge transition of RuO2 have been reported: (i) O 2p- Ru 4d interband transition, resulting in an optical absorption at shorter wavelength range (up to 2.1 eV, ~ 580 nm); and (ii) d-d intraband transition between partially filled and unfilled Ru 4d orbitals or free carrier absorptions, giving an electronic band at 1.5-1.9 eV (~ 800 – 650 nm).10,48 And the proximity of surface plasmon and interband transition energies results in a damping of surface plasmon by interband electron excitations, thus reducing the surface plasmon local field.49 In the present study, it might be conclusive that the effect of Ru 4d-Ti 3d transition (or hybridized state) and O 2p-Ru 4d interband transition of RuO2 is more pronounced than the d-d intraband transition of RuO2, exhibiting significant enhancement in visible light absorption, therefore enabling RuO2/TiO2 heterostructure to efficiently produce H2 as compared to blank TiO2. Especially, an increase in visible absorption is coincident with a decrease in UV region; and the O 2p-Ti 3d direct transition in range of 200-400 nm for 0.1RuTi sample was distinct from others, possibly stemming from the stronger local electronic distortion in TiO2 by RuO2 and faster interfacial charge transfer from TiO2 to RuO2. Both band gap narrowing and enhanced absorption features in visible regime notably confirm that all rod-like RuO2/TiO2 heterostructures are highly efficient visible-light-responsive photocatalysts. On the contrary, very poor visible-light absorption behavior was observed when loading 0.1 wt% RuO2 onto bulk TiO2 particles as indicated in Figure S9 (Supporting Information). It is interesting to note that the an absorption tail towards visible light region up to 480 nm were 24 observed on rutile TiO2 NRs which did not happen with bulk rutile TiO2. Recently, Li et al.30 reported similar visible light absorption over sub-10nm rutile TiO2 NPs compared to bulklike or larger-grain-size rutile TiO2. A slightly increased absorbance in the range of 700-1000 nm is more visible than that ranging from 400 to 550 nm, accounting for the dominance of partially filled and unfilled Ru 4d electronic bands (d-d intraband transition). That contributes to extremely lower UV activity and almost no visible-light-responsive activity (not shown here) of bulk TiO2 and 0.1RuTi_p material. The photoluminescence spectroscopy (Figure S10, Supporting Information) was additionally performed to understand the efficiency of charge carrier trapping, migration, transfer and separation. PL peak intensity of xRuTi heterostructure relatively decreases with increasing RuO2 amounts, implying the suppression of electron-hole recombination rate that favors the higher photocatalytic activity. It can be seen that although the loading beyond 0.1 % exhibits lower recombination rate, worse H2 production performance was obtained possibly due to the dominance of light shielding effect and abundance of oxygen vacancy defects as mentioned above that remarkably inhibit the performance. Gathering the experimental data of Eg and VBM of 0.1RuTi nanocomposite, a possible band level alignment diagram at the interface is schematically shown in Figure 5D. The formation of a heterojunction at the RuO2-TiO2 interface, band gap narrowing, and upward band bending at the interface effectively promote the separation of photoinduced electrons and holes, leading to an increase in photocatalytic H2 evolution. 25 Scheme 1. Schematic illustration of proposed upward band bending-involved charge transfer pathway of RuO2/TiO2 heterojunction nanocomposite to produce H2 under visible light irradiation. Scheme 1 illustrates the proposed charge transfer pathway over RuO2/TiO2 heterostructures to produce H2. By this way, it favors the dynamic transfer of electrons and holes from TiO2 to RuO2 via an intimate contact due to higher work function of RuO2,19,21,35 subsequently reacting with protons, water and methanol. The photoinduced holes from VB of TiO2 NRs are irreversibly scavenged by methanol to produce various oxidation intermediates and products such as OH radical, CH2OH, HCHO, HCOOH and CO2.31,35,50 Meanwhile, the photogenerated electrons simultaneously reduce the protons and water to generate H2 gas. These reactions proceed competitively with the recombination of electron-hole pairs. Herein, RuO2 efficiently stabilizes the holes and creates the heterojunction or a Schottky barrier at the heterointerface between RuO2 and TiO2,3,6,35 hindering the charge carrier recombination and consequently, enhancing the H2 evolution. Because Ru is more electronegative than Ti (2.2 vs. 1.54), the strong accumulation of electron density at the interface creates a strong internal 32 (31) Ismail, A.A.; Robben, L.; Bahnemann, D.W. Study of the Efficiency of UV and Visible-Light Photocatalytic Oxidation of Methanol on Mesoporous RuO2-TiO2 Nanocomposites. Chem. Phys. Chem. 2011, 12, 982-991. (32) Ovoshchnikov, D.S.; Donoeva, B.G.; Golovko, V.B. Visible-Light-Driven Aerobic Oxidation of Amines to Nitriles over Hydrous Ruthenium Oxide Supported on TiO2. ACS Catal. 2015, 5, 34-38. (33) Amama, P.B.; Itoh, K.; Murabayashi, M. Effect of RuO2 Deposition on the Activity of TiO2: Photocatalytic Oxidation of Trichloroethylene in Aqueous Phase. J. Mater. Sci. 2004, 39, 4349-4351. (34) Lin, F.; Zhang, Y.; Wang, L.; Zhang, Y.; Wang, D.; Yang, M.; Yang, J.; Zhang, B.; Jiang, Z.; Li, C. Highly Efficient Photocatalytic Oxidation of Sulfur-Containing Organic Compounds and Dyes on TiO2 with Dual Cocatalysts Pt and RuO2. Appl. Catal. B: Environ. 2012, 127, 363-370. (35) Uddin, Md.T.; Nicolas, Y.; Olivier, C.; Toupance, T.; Mller, M.M.; Kleebe, H.-J.; Rachut, K.; Ziegler, J.; Llein, A.; Jaegermann, W. Preparation of RuO2/TiO2 Mesoporous Heterostructures and Rationalization of Their Enhanced Photocatalytic Properties by Band Alignment Investigations. J. Phys. Chem. C 2013, 117, 22098- 22110. (36) Kundu, S.; Vidal, A.B.; Yang, F.; Ramírez, P.J.; Senanayake, S.D.; Stacchiola, D.; Evans, J.; Liu, P.; Sanz, J.F.; Rodriguez, J.A. Special Chemical Properties of RuOx Nanowires in RuOx/TiO2(110): Dissociation of Water and Hydrogen Production. J. Phys. Chem. C 2012, 116, 4767-4773. 33 (37) Balachandran, U.; Eror, N.G. Raman Spectra of Titanium Dioxide. J. Solid State Chem. 1982, 42, 276-282. (38) Ocaña, M.; Fornés, V.; Ramos, J.V.G.; Serna, C.J., Factors Affecting the Infrared and Raman Spectra of Rutile Powders. J. Solid State Chem. 1988, 75, 364-372. (39) Parker, J.C.; Siegel, R.W. Calibration of the Raman Spectrum to the Oxygen Stoichiometry of Nanophase TiO2. Appl. Phys. Lett. 1990, 57, 943-945. (40) Santara, B.; Giri, P.K.; Imakita, K.; Fujii, M. Microscopic Origin of Lattice Contraction and Expansion in Undoped Rutile TiO2 Nanostructures. J. Phys. D: Appl. Phys. 2014, 47, 215302. (41) Badyal, J.P.S.; Lambert, R.M.; Harrison, K.; Riley, C.C.A.; Frost, J.C. Investigation of the SMSI Phenomenon with TiO2/Ru/SiO2 Model-Dispersed Catalysts. J. Catal. 1991, 129, 486-496. (42) Wendt, S.; Matthiesen, J.; Schaub, R.; Vestergaard, E. K.;Laegsgaard, E.; Besenbacher, F.; Hammer, B. Formation and Splitting of Paired Hydroxyl Groups on Reduced TiO2(110). Phys. Rev. Lett. 2006, 96, 066107. (43) Oviedo, J.; Snchez-de-Armas, R.; ngel San Miguel, M.; Sanz, J.F. Methanol and Water Dissociation on TiO2(110): The Role of Surface Oxygen. J. Phys. Chem. C 2008, 112, 17737-17740. (44) Stacchiola, D.J.; Senanayake, S.D.; Liu, P.; Rodriguez, J.A. Fundamental Studies of Well-Defined Surfaces of Mixed-MetalOxides: Special Properties of MOx/TiO2(110) {M = V, Ru, Ce, or W}. Chem. Rev. 2013, 113, 4373-4390. 34 (45) Ohkoshi, S.; Tsunobuchi, Y.; Matsuda, T.; Hashimoto, K.; Namai, A.; Hakoe, F.; Tokoro, H. External Stimulation-Controllable Heat-Storage Ceramics. Nat. Chem. 2010, 2, 539. (46) Tanaka, K.; Nasu, T.; Miyamoto, Y.; Ozaki, N.; Tanaka, S.; Nagata, T.; Hakoe, F.; Yoshikiyo, M.; Nakagawa, K.; Umeta, Y.; Imoto, K.; Tokoro, H.; Namai, A.; Ohkoshi, S. Structural Phase Transition between γ-Ti3O5 and δ-Ti3O5 by Breaking of a One-Dimensionally Conducting Pathway. Cryst. Growth Des. 2015, 15, 653- 657. (47) Kundu, S.; Vidal, A. B.; Nadeem, M. A.; Senanayake, S. D.; Idriss, H.; Liu, P.; Rodriguez, J. A.; Stacchiola, D. Ethanol Photoreaction on RuOx/Ru-Modified TiO2(110). J. Phys. Chem. C 2013, 117, 11149-11158. (48) Tan, H.; Ye, E.Y.; Fan, W.Y. Alumina-Template Synthesis of Fluorescent RuO2 Nanotubes Derived from Ru3(CO)12 Clusters. Adv. Mater. 2006, 18, 619-623. (49) Pustovit, V.N.; Shahbazyan, T.V. Finite-Size Effects in Surface-Enhanced Raman Scattering in Noble-Metal Nanoparticles: A Semiclassical Approach. J. Opt. Soc. Am. A 2006, 23, 1369-1374. (50) Schneider, J.; D. W. Bahnemann, D. W. Undesired Role of Sacrificial Reagents in Photocatalysis. J. Phys. Chem. Lett. 2013, 4, 3479-3483. 35 TABLE OF CONTENTS AND ABSTRACT GRAPHIC