Electronic Structure of Semiconducting and Metallic Tubes in TiO2

Apr 9, 2013 - School of Physics and Complex & Adaptive Systems Laboratory, University ... the metallic CNT/TiO2 system due to efficient charge separat...
0 downloads 0 Views 4MB Size
Letter pubs.acs.org/JPCL

Electronic Structure of Semiconducting and Metallic Tubes in TiO2/ Carbon Nanotube Heterojunctions: Density Functional Theory Calculations Run Long* School of Physics and Complex & Adaptive Systems Laboratory, University College Dublin, Ireland ABSTRACT: The electronic structure of the TiO2(110) surface interfaced with both a semiconducting and metallic carbon nanotube (CNT) was investigated by density functional theory. Our simulations rationalized visible light photocatalytic activity of CNT/TiO2 hybrid materials higher than that under ultraviolent irradiation and showed that the photoactivity of a semiconducting CNT decorating TiO2 is better than that of the metallic CNT/TiO2 system due to efficient charge separation across the interface. This suggests that semiconducting CNT/TiO2 could be a potential photovoltaic material. In contrast, strong interaction between a metallic CNT and TiO2 leads to large charge transfer. Such charge transfer reduces the built-in potential, in turn resulting in inefficient charge separation. Functionalizing the metallic CNT with a small platinum cluster can increase the built-in potential and drive charge separation. These observations indicate that the CNT/TiO2 interface can be a potential photovoltaic material by a metal cluster decorating a CNT despite a real tube being composed of the mixture of metallic and semiconducting CNTs. SECTION: Surfaces, Interfaces, Porous Materials, and Catalysis

T

Recently, carbon nanotube/titania (CNT/TiO2) hybrid systems have been receiving intense attention to address energy and environmental problems including photovoltaics and photocatalysts17,22−26 because CNTs have excellent mechanical properties and a large specific area (>150 m2 g−1). Especially its excellent electronic properties provide continuous (semi-infinite) electronic states in the conduction band (CB) for donating the transferring electrons from the nth von Hove singularity to the semiconductor. Liang and coworkers reported that CNTs are effective titania photosensitizers, leading to great photoactivity enhancement under visible light illumination with a slight improvement under UV irradiation.22 This concept agrees with recent experiments that showed that photoexcited electrons are able to transfer either from CNTs to TiO227 or from graphene to ZnS under visible light irradiation.28 Both composites show great visible light photocatalytic activity. However, the mechanisms of photocatalysis enhancement in the CNT/TiO2 composites remain open. A few other experiments claimed that the enhancement of the photocatalytic properties of that CNT/TiO2 systems is a result of a high-energy photon exciting an electron from the valence band (VB) to the CB of TiO2 under UV light irradiation.29,30 To understand the mechanisms of photocatalysis under UV irradiation and visible light irradiation, one should find how much the TiO2 band gap narrows with incorporation of CNTs into the TiO2 lattice. Furthermore, the charge transfer and separation affect greatly the photocatalytic

iO2 has been widely used as a promising photocatalyst and photovoltaic material for hydrogen generation via water splitting and decomposing pollution due to its high photocatalytic activity, resistance to photocorrosion, cost effectiveness, and nontoxicity.1−3 Unfortunately, the range of optical absorption of TiO2 is limited to the ultraviolet (UV) region due to its large intrinsic band gap (3.0 eV for rutile and 3.2 eV for anatase), which occupied only ∼5% of the whole solar photon spectrum. Narrowing the band gap to the visible light range provides a straightforward way for achieving higher photovoltaic and photocatalytic activity because light absorption and the consequential photoexcitation of electron−hole pairs occur when the energy of the incident photons matches or exceeds the band gap. Realization of band gap in the visible light region results in harvesting longer solar wavelengths and, all other factors remaining the same, generating greater fluxes of photoexcited electrons. Photogenerated electrons can either be channeled to create electricity directly in photovoltaic solar cells or else be used to drive chemical reactions. Doping4 or codoping5,6 with impurity atoms is an effective way to reduce the band gap. However, dopant-mediated electron−hole pair recombination in the lattice can be a significant problem, restricting practical application. Furthermore, it is hard to achieve a high concentration of dopants in experiments, or the synthesized materials possess low stability against photocorrosion.7,8 As an alternative approach, composites of TiO2 with other materials can help to overcome these shortcomings. Examples include the molecule chromophore,9 conjugated carbon polymers,10 semiconductor quantum dots,11,12 noble metal particles,13,14 and nanoscale carbon materials.15−21 © 2013 American Chemical Society

Received: March 17, 2013 Accepted: April 7, 2013 Published: April 9, 2013 1340

dx.doi.org/10.1021/jz400589v | J. Phys. Chem. Lett. 2013, 4, 1340−1346

The Journal of Physical Chemistry Letters

Letter

gradient approximation (GGA)33 was adopted for the exchange−correlation potential. The electron wave function was expanded in a plane wave cutoff energy of 400 eV. We carried out DFT calculations with a Monkhorst−Pack34 4 kpoint in the Brillouin zone (including the Gamma point) for geometry optimization, while a much denser 32 k-point (including the Gamma point) for electronic structure calculations was used with the DFT+U approach35 due to the presence of 3d electrons of Ti ions. The geometry optimization stopped until the residual forces were below 0.01 eV/Å. The onsite U = 6.0 eV and J = 0.5 eV parameters were applied to the Ti3d electrons.36,37 The interaction energies between the TiO2 surface and the CNTs (10,0) and (9,0) were found to be 0.38 and 0.25 eV, respectively, indicating both hybrids form a stable interface, Figure 1. It is hard to distinguish that the coupling between the

ability. At the same time, they play key roles in determining the heterojunction efficiency in the context of photovoltaic solar cells. Recent experiments show that the CNT/TiO2 interface stabilizes charge separation, reduces charge recombination,30 and improves the photocurrent.17,23,24 In a conventional way, charge should take place rapid recombination in a metal. Because a real CNT is composed of the mixture of metallic and semiconducting CNTs, either the metallic or semiconducting CNT dominates the components that will affect the charge separation efficiency. From a dynamics viewpoint, the photocatalytic activity of CNT/TiO2 will decrease if the metallic CNT is the main component in a real CNT. However, the real photocatalytic activity of a photocatalyst is influenced by many factors, such as the sample arrangement, component ratio, and lifetime of electron/hole carriers as well. In this Letter, we perform density functional theory (DFT) calculations to answer the above two questions via characterizing the interfacial electronic structures and electron transfer of the CNT for both semiconducting and metallic tubes interfaced with rutile a TiO2(110) surface. The calculated density of states (DOS) shows that the band gap of TiO2 has only a tiny reduction, while many CNT states are localized within the gap. These findings rationalize that improvement of photoactivity was observed under visible light irradiation, while there was a slight improvement under UV irradiation.22 Under visible light irradiation, an electron was excited on the nth von Hove singularity of the CNT from the VB to CB. Then, the photogenerated electrons was transferred to the TiO2 CB, the gap serving as high photoactivity. In contrast, the intrinsic electron transition between the VB and CB of TiO2 remains dominant under UV irradiation, which is responsible for the slight enhancement of the photoactivity. However, CNTs are able to optimize the TiO2/CNT surface morphology that stabilizes charge separation and reduces charge recombination, which rationalizes obvious enhancement of photoactivity under UV irradiation by some other experiments.17,29,30 Due to effective charge separation across the semiconducting CNT/ TiO2 interface, the semiconducting CNT is better than the metallic CNT as a photosensitizer to obtain high enhancement of the visible light photoactivity. The semiconducting CNT/ TiO2 interface can form a potentially excellent photovoltaic solar cell due to a small interaction leading to a slight amount charge transfer. In contrast, the metallic CNT/TiO2 heterojunction does not due to significant charge transfer resulting in a small built-in potential. Functionalizing the metallic CNTs with a platinum cluster can increase the built-in potential and drive efficient charge separation. Therefore, a high-efficiency CNT/TiO2 photovoltaic heterojunction can be obtained via decorating a real CNT with a suitable metal cluster. For the tubes, a (1 × 1 × 3) CNT (10,0) and (9,0) tubes are used to represent typical ∼1 nm semiconducting and metallic CNTs, respectively. To construct the periodic interface, we choose a very large (6 × 2) stoichiometric rutile TiO2(110) surface slab containing 96 O atoms and 48 Ti atoms with 3 bottom layers fixed at the bulk position, matching the CNT length of 12.8 Å in its axial direction. This only gives rise to minor axial strain, leading to a 1.5% lattice mismatch. The vacuum depth is as large as 25 Å for both models; which is sufficient to separate the interaction between periodic images. The DFT calculations are carried out by using the Vienna ab initio simulation package (VASP) code,31,32 implementing the projector augmented wave (PAW) pseudopotentials.31 The Perdew−Burke−Ernzerhof parametrization of the generalized

Figure 1. Optimized interface structures of (a) semiconducting (10,0) CNT/TiO2(110) and (b) metallic (9,0) CNT/TiO2(110). The corresponding binding energies are 0.38 and 0.25 eV, respectively. The gray, red, and cyan balls represent C, O, and Ti atoms, respectively.

(10,0) CNT and TiO2 is stronger than that for (9,0) CNT/ TiO2 due to the larger contact area of the (10,0) CNT with TiO2 than that for (9,0) CNT/TiO2. However, the extent of charge transfer can reflect the coupling strength, namely, the stronger coupling and the more charge transfer. The calculated Bader charge38 shows that a significant charge transfer of almost 0.25 electron from metallic the (9,0) CNT to the TiO2 surface. Alternatively, a small amount, 0.1 electron, was found to be transferred from the semiconducting (10,0) CNT to the TiO2 substrate. The Bader charge analysis indicated that the metallic CNT(9,0) interacts strongly with TiO2 compared to the semiconducting CNT(10,0)/TiO2 interface, leading to substantial charge transfer and hole doping on the CNTs. To understand the origin of such a charge transfer at the interface, work functions for the TiO2(110) surface, (9,0) CNT, and (10,0) CNT were calculated by aligning the Fermi level to the vacuum level. The obtained values were 6.71, 4.29, and 4.51 eV, respectively, which agree well with previous studies.37,39,40 The difference of 2.42 over 2.20 eV can rationalize a much larger charge transfer for the metallic CNT/TiO2 interface than the 1341

dx.doi.org/10.1021/jz400589v | J. Phys. Chem. Lett. 2013, 4, 1340−1346

The Journal of Physical Chemistry Letters

Letter

Figure 2. DOS for a (a) semiconducting (10,0) CNT, (b) metallic (9,0) CNT, (c) CNT(10,0)/TiO2(110) surface, and (d) CNT(9,0)/TiO2(110) surface (right panel). The DOS of TiO2 and the CNT are plotted in black and red lines, respectively. The DOS of CNT was amplified three times for the eyes. The Fermi level is set to zero. The DOS shows CNT states and dominates the states above the top of the VB. Interband electron transition from the VB to the CB of the CNT; after, this photogenerated electron is able to transfer to the bottom of the CB of TiO2, which requires small phonon energy and extends the optical absorption edge to the visible light region.

on the TiO2, serve as the obvious enhancement of photoactivity under UV irradiation.17,29,30 However, the electron is capable of exciting from the first or second van Hove singularity of either the semiconducting or metallic CNT to its CB under visible light irradiation, and then, this photogenerated electron is injected into the CB of TiO2. Once this happens, the positively charged CNT can remove an electron from the VB of TiO2, leaving a hole. The oxidation and redox reactions can take place in TiO2 and CNTs, respectively. This finding can rationalize the significant enhancement of visible light photoactivity.22,27 Here, both semiconducting and metallic CNTs act as photosensitizers whose mechanism is different from that of the C-doped TiO2,43 in which the electron is directly excited from the C impurity states into the TiO2 CB, leading to an optical absorption edge extending to the visible light region. Taking into account CNT decorating of the TiO2 would not induce large lattice distortion while achieving large effective gap narrowing, which is better than introduction of a C dopant into the TiO2 lattice via the doping approach. Comparing the semiconducting and metallic CNTs, the semiconducting CNTdecorated TiO2 should have high visible light photocatalytic activity due to low electron/hole recombination rate. Therefore, a semiconducting CNT interfaced with TiO2 is the preferable choice to achieve high photoactivity using a CNTdecorated TiO2 semiconductor. In principle, the injection of a photogenerated electron can be transferred nonadiabatically and adiabatically. According to the Fermi golden rule, the nonadiabatic electron transfer is expected to be highly efficient due to a large DOS, similar to the case of the graphene/TiO2 system.36 Panels c and d of Figure 2 show that CB states of CNTs and TiO2 hybridized significantly, which allows for strong wave function overlap,

semiconducting CNT/TiO2 interface. In addition, it should be noted that dipole correction might artificially favor charge transfer. Therefore, repeating calculations of the Bader charge in the presence and absence of dipole correction illuminates that the interface charge transfer changes are tiny. In order to understand that the photocatalytic activity of CNT/TiO2 composites can be significantly enhanced under visible light irradiation and increase slightly under UV irradiation,22 the DOSs of individual CNTs and TiO2 before and after formation of the hybrid interface are calculated, Figure 2. The band gap measured from the O2p orbitals at the valence band maximum (VBM) to the Ti3d orbitals at the conduction band minimum (CBM) of TiO2 is 2.0 eV (not shown here), smaller than the experimental value of 3.0 eV but improved a lot compared to DFT results.41,42 Panels a and b of Figure 2 show the DOSs of an individual semiconducting (10,0) CNT and the metallic (9,0) CNT with 0.78 and 0 eV band gaps, respectively. Upon formation of a hybrid interface, the electronic structures of TiO2 and the CNTs almost remain the same because donor−acceptor coupling is not strong enough in the absence of covalent bonding (Figure 2c and d) for semiconducting and metallic CNTs interfaced with TiO2, respectively. The intrinsic band gap of TiO2 in both cases has a tiny change and implies that the electron transition from O2p at the VB to Ti3d at the CB is the dominant process, which serves as a slight increase of the photoactivity under UV irradiation.22 In addition, the CNT/TiO2 composites’ morphology differs in different experiments. Some of the CNT/TiO2 hybrid materials have optimal arrangements, and the CNTs are able to improve photoexcited charge separation, transfer, and charge collection. The electrons can be shuttled freely along the conducting network of the CNT bundle. The longer-lived holes 1342

dx.doi.org/10.1021/jz400589v | J. Phys. Chem. Lett. 2013, 4, 1340−1346

The Journal of Physical Chemistry Letters

Letter

driving an efficient adiabatic electron transfer. For real-time electron-transfer dynamics, the motions of nuclei affect significantly the nonadiabatic coupling and, in turn, affect either the nonadiabatic or adiabatic mechanism, which dominates the electron-transfer event.44,45 In addition to charge injection, efficient charge separation plays a key role in affecting the photovoltaic solar heterojunction performance. The DOS shown in Figure 2c and the data shown in Table 1 demonstrate Table 1. Energies Parameters (in eV) of the Interfaced CNT/TiO2(110) Heterojunctions for the Semiconducting (10,0), Metallic (9,0), and Pt4-(9,0) Tubesa

CBM offset VBM offset KS gap

CNT(10,0)/ TiO2(110)

CNT(9,0)/ TiO2(110)

Pt4−CNT(9,0)/ TiO2(110)

0.65

0.09

0.26

1.9

2.02

2.16

0.25

n/a

n/a

a

CBM (VBM) offsets are the energy differences between the effective TiO2 CBM (VBM) and CNT(10,0) CBM (VBM) for the semiconducting tube, while it refers to the effective TiO2 CBM (VBM) relative to the Fermi level for the metallic (9,0) tube.

that a large built-in potential can drive charge carriers, resulting in effective separation, indicating that a semiconducting CNT/ TiO2 is capable of forming effective solar cells. Figure 2d shows that, however, the metallic CNT/TiO2 system seems to be ineffective due to a small charge separation driving force. To clarify the charge transfer and separation processes, we calculated the three-dimensional charge density difference by subtracting the electronic charge of a hybrid CNT/TiO2 composite from that of the individual CNT and TiO2(110) surface, as shown in Figure 3. Figure 3a shows that charge redistribution mostly occurs at the semiconducting (10,0) CNT/TiO2 interface region, while there is almost no charge transfer on the tube farther from the interface. Upon formation of an interface, the electron is able to excite from the CNT ground state to its excited state and then is transferred to the TiO2 CB, leading to electron−hole pair separation. In contrast, for a metallic CNT/TiO2 interface, charge redistribution can be seen in the whole metallic (9,0) CNT beside the interface due to a strong donor−acceptor interaction, leading to a large amount of charge transfer. Figure 3c plots the planar averaged charge density difference along the direction perpendicular to the TiO2(110) surface, which can provide quantitative results of charge transfer and redistribution. The vertical solid line is the position of the top layer of the TiO2 surface. The positive values represent electron accumulation, and negative values indicate electron depletion. This plot further confirms that efficient charge transfer occurs in both cases despite the fact that the amount in the metallic (9,0) CNT/TiO2 interface is greater than that for the semiconducting (10,0) CNT/TiO2 interface. Because the photogenerated electron will take place in rapid recombination in the metallic CNT due to the lack of a band gap, one cannot safely obtain the result that effective charge separation can be taken place across the metallic CNT/ TiO2 interface, although it does in a metallic graphene/TiO2 system.36 The electrostatic potential (EP) difference can further explain how the presence of either semiconducting or metallic CNTs would affect the charge transfer and separation process. Panels a and b of Figure 4 shows the three-dimensional

Figure 3. Charge density differences for a (a) (10,0) CNT/TiO2(110) surface and (b) (9,0) CNT/TiO2(110) surface. The yellow region represents charge accumulation, and the cyan region indicates charge depletion; the isosurface value is 0.0006 e/ Å3. (c) Planar averaged charge density difference for the (10,0) CNT/TiO2(110) surface (blue) and (9,0) CNT/TiO2(110) surface (red) as a function of position in the z-direction. The vertical line indicates the location of the top layer of the TiO2(110) surface. The plots show that charge transfer occurs significantly between the metallic (9,0) CNT and the TiO2 surface over that of semiconducting the (10,0) CNT/TiO2(110) interface.

potential, subtracted by an individual TiO2(110) surface and CNT components, for both the semiconducting and metallic CNT cases. Contrary to the case of the (10,0) semiconducting CNT where there was a minor change of potential, a strong potential drop was observed on the metallic (9,0) CNT. The planar average of the EP difference profile confirmed this observation, Figure 4c. The metallic CNT induces this attractive region for negative charges on the TiO2. In principle, large charge transfer will reduce the built-in potential; in turn, a small built-in potential cannot effectively drive the electron− hole pair separation from the interface region. For the metallic (9,0) CNT/TiO2 interface, the calculated built-in potential of 0.09 eV is too small to drive efficient charge separation, Table 1. In experiments, however, CNTs have a mixed distribution of semiconducting and metallic tubes, and it is likely that the metallic tubes interact strongly with TiO2 because of the strong electrostatic interaction. Functionalized metallic CNTs with metal clusters might result in upward movement of the Fermi level of the CNTs, which will increase the offset between the CBM of CNTs and TiO2 and drive efficient electron−hole separation. We use a Pt4 cluster to modify the electronic structure of the CNT/TiO2 interface. The most stable configuration is shown in Figure 5a, in which each of three 1343

dx.doi.org/10.1021/jz400589v | J. Phys. Chem. Lett. 2013, 4, 1340−1346

The Journal of Physical Chemistry Letters

Letter

Figure 5. (a) Optimized Pt4−CNT(9,0)/TiO2(110) structure and (b) the DOS of Pt4−CNT (red) and TiO2 (black). The DOS shows that pyramid Pt4 covalently binding to a metallic CNT leads to the CBM of the composites’ upward movement. Compared to CNT(9,0)/ TiO2(110), Figure 2d, the built-in potential increases. Correspondingly, efficient charge separation is enhancement. The gray, red, yellow, and cyan balls represent C, O, Pt, and Ti atoms, respectively.

Figure 4. Induced change in the EP (in atomic units) by the interface formation, ΔV(r) = V(r) − VCNT(r) − VTiO2(r), at the fixed ionic positions near the interface for the (a) semiconducting (10,0) CNT case and (b) metallic (9,0) CNT case. The potential plots indicate a large change of the EP in the metallic (9,0) CNT/TiO2 interface over that of the semiconducting(10,0) CNT/TiO2 interface. (c) Planar averaged electrostatic potential difference for the (10,0) CNT/ TiO2(110) surface (blue) and (9,0) CNT/TiO2(110) surface (red) as a function of position in the z-direction.

and CB. Then, this photogenerated electron injected into the CB of TiO2 can effectively reduce the transition energy, leading to a significant enhancement of phototactivity. Due to efficient charge separation across the semiconducting CNT/TiO2 relative to that for the metallic CNT/TiO2 interface, the former is better than the latter as photocatalysts and photovoltaic materials. A significant charge transfer occurs from the metallic CNT to the TiO2, resulting in a small built-in potential that could be unfavorable for exciton dissociation. When functionalizing the metallic CNT with a Pt4 cluster, the built-in potential increases, and efficient charge separation is attainable. These observations indicate that in a photovoltaic heterojunction based on a mixed CNT distribution, the CNT/ TiO2 composites are efficient to enhance the photocurrent via suitable metal-cluster-decorated CNTs.

bottom Pt atoms in pyramid Pt4 binds to the metallic (9,0) CNT via C−Pt−C covalent bonds. The simulation cell employs a very large 30 Å vacuum depth. Formation of C−Pt bonds changes local bonding characteristics, which increases the electrostatic repulsion interaction between the CNT and TiO2. The calculated DOS, Figure 5b, shows that the built-in potential increased to 0.26 eV, Table 1. This value can compete with the exciton binding energy of a single carbon nanotube.46 Considering that only one Pt4 cluster is used in the simulation cell and the density is very low, the built-in potential should increase largely with a higher density of metal clusters. Therefore, one can choose a suitable metal cluster to decorate CNTs to achieve high efficiency of photovoltaic solar cells. In summary, we performed ab initio DFT calculations to explore electronic properties and charge transfer at the interface formed between CNTs and a rutile TiO2(110) surface in the context of visible light photocatalyst and photovoltaic cells. Incorporation of either a semiconducting or metallic CNT into the TiO2 lattice only induces a tiny band gap narrowing while introduction of some carbon states within the gap. The small reduction of the band gap can explain the slight enhancement of photocatalytic activity under UV irradiation. Under visible light irradiation, however, electron can be excited from the first or second von Hove singularity of the CNT between the VB



AUTHOR INFORMATION

Corresponding Author

*E-mail: [email protected]. Notes

The authors declare no competing financial interest.



ACKNOWLEDGMENTS R.L. thanks the Science Foundation Ireland (SFI) SIRG Program (Grant Number 11/SIRG/E2172), UCD Seed Funding SF614, and the Irish Center for High End Computing for the provision of computational resources.



REFERENCES

(1) Fujishima, A.; Honda, K. Electrochemical Photolysis of Water at a Semiconductor Electrode. Nature 1972, 238, 37−38. (2) Zhang, J.; Xu, Q.; Feng, Z.; Li, M.; Li, C. Importance of the Relationship between Surface Phases and Photocatalytic Activity of TiO2. Angew. Chem., Int. Ed. 2008, 47, 1766−1769. 1344

dx.doi.org/10.1021/jz400589v | J. Phys. Chem. Lett. 2013, 4, 1340−1346

The Journal of Physical Chemistry Letters

Letter

(3) Ni, M.; Leung, M. K. H.; Leung, D. Y. C.; Sumathy, K. A Review and Recent Developments in Photocatalytic Water-Splitting Using for Hydrogen Production. Renewable Sustainable Energy Rev. 2007, 11, 401−425. (4) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. VisibleLight Photocatalysis in Nitrogen-Doped Titanium Oxides. Science 2001, 293, 269−271. (5) Wang, Y.; Wang, Y.; Meng, Y.; Ding, H.; Shan, Y.; Zhao, X.; Tang, X. A Highly Efficient Visible-Light-Activated Photocatalyst Based on Bismuth- and Sulfur-Codoped TiO2. J. Phys. Chem. C 2008, 112, 6620−6626. (6) Gai, Y.; Li, J.; Li, S.-S.; Xia, J.-B.; Wei, S.-H. Design of NarrowGap TiO2: A Passivated Codoping Approach for Enhanced Photoelectrochemical Activity. Phys. Rev. Lett. 2009, 102, 036402. (7) Graciani, J.; Alvarez, L. J.; Rodriguez, J. A.; Sanz, J. F. N Doping of Rutile TiO2 (110) Surface. A Theoretical DFT Study. J. Phys. Chem. C 2008, 112, 2624−2631. (8) Batzill, M.; Morales, E. H.; Diebold, U. Influence of Nitrogen Doping on the Defect Formation and Surface Properties of TiO2 Rutile and Anatase. Phys. Rev. Lett. 2006, 96, 026103. (9) Park, J.; Yi, J.; Tachikawa, T.; Majima, T.; Choi, W. GuanidiniumEnhanced Production of Hydrogen on Nafion-Coated Dye/TiO2 under Visible Light. J. Phys. Chem. Lett. 2010, 1, 1351−1355. (10) Wang, Y.; Li, L.; Yang, K.; Samuelson, L. A.; Kumar, J. Nanocrystalline TiO2-Catalyzed Solid-State Polymerization of Diacetylene in the Visible Region. J. Am. Chem. Soc. 2007, 129, 7238−7239. (11) Liu, L.; Hensel, J.; Fitzmorris, R. C.; Li, Y.; Zhang, J. Z. Preparation and Photoelectrochemical Properties of CdSe/TiO2 Hybrid Mesoporous Structures. J. Phys. Chem. Lett. 2009, 1, 155−160. (12) Wang, D.; Zhao, H.; Wu, N.; El Khakani, M. A.; Ma, D. Tuning the Charge-Transfer Property of PbS-Quantum Dot/TiO2-Nanobelt Nanohybrids via Quantum Confinement. J. Phys. Chem. Lett. 2010, 1, 1030−1035. (13) Zhang, N.; Liu, S.; Fu, X.; Xu, Y.-J. Synthesis of M@TiO2 (M = Au, Pd, Pt) Core−Shell Nanocomposites with Tunable Photoreactivity. J. Phys. Chem. C 2011, 115, 9136−9145. (14) Zhang, N.; Liu, S.; Xu, Y.-J. Recent Progress on Metal Core@ Semiconductor Shell Nanocomposites As a Promising Type of Photocatalyst. Nanoscale 2012, 4, 2227−2238. (15) Ng, Y. H.; Lightcap, I. V.; Goodwin, K.; Matsumura, M.; Kamat, P. V. To What Extent Do Graphene Scaffolds Improve the Photovoltaic and Photocatalytic Response of TiO2 Nanostructured Films? J. Phys. Chem. Lett. 2010, 1, 2222−2227. (16) Zhang, L.; Diao, S.; Nie, Y.; Yan, K.; Liu, N.; Dai, B.; Xie, Q.; Reina, A.; Kong, J.; Liu, Z. Photocatalytic Patterning and Modification of Graphene. J. Am. Chem. Soc. 2011, 133, 2706−2713. (17) Kongkanand, A.; Martínez Domínguez, R.; Kamat, P. V. Single Wall Carbon Nanotube Scaffolds for Photoelectrochemical Solar Cells. Capture and Transport of Photogenerated Electrons. Nano Lett. 2007, 7, 676−680. (18) Zhang, Y.; Tang, Z.-R.; Fu, X.; Xu, Y.-J. Engineering the Unique 2D Mat of Graphene to Achieve Graphene−TiO2 Nanocomposite for Photocatalytic Selective Transformation: What Advantage does Graphene Have over Its Forebear Carbon Nanotube? ACS Nano 2011, 5, 7426−7435. (19) Zhang, Y.; Tang, Z.-R.; Fu, X.; Xu, Y.-J. TiO2−Graphene Nanocomposites for Gas-Phase Photocatalytic Degradation of Volatile Aromatic Pollutant: Is TiO2−Graphene Truly Different from Other TiO2−Carbon Composite Materials? ACS Nano 2010, 4, 7303−7314. (20) Zhang, N.; Zhang, Y.; Xu, Y.-J. Recent Progress on GrapheneBased Photocatalysts: Current Status and Future Perspectives. Nanoscale 2012, 4, 5792−5813. (21) Zhang, N.; Zhang, Y.; Yang, M.-Q.; Tang, Z.-R.; Xu, Y.-J. A Critical and Benchmark Comparison on Graphene−, Carbon Nanotube−, and Fullerene−Semiconductor Nanocomposites As Visible Light Photocatalysts for Selective Oxidation. J. Catal. 2013, 299, 210−221. (22) Liang, Y. T.; Vijayan, B. K.; Lyandres, O.; Gray, K. A.; Hersam, M. C. Effect of Dimensionality on the Photocatalytic Behavior of

Carbon−Titania Nanosheet Composites: Charge Transfer at Nanomaterial Interfaces. J. Phys. Chem. Lett. 2012, 3, 1760−1765. (23) Chen, J.; Li, B.; Zheng, J.; Zhao, J.; Zhu, Z. Role of Carbon Nanotubes in Dye-Sensitized TiO2-Based Solar Cells. J. Phys. Chem. C 2012, 116, 14848−14856. (24) Jang, S.-R.; Vittal, R.; Kim, K.-J. Incorporation of Functionalized Single-Wall Carbon Nanotubes in Dye-Sensitized TiO2 Solar Cells. Langmuir 2004, 20, 9807−9810. (25) Chang, M.; Wu, L.; Li, X.; Xu, W. Effect of MWCNT Inclusion in TiO2 Nanowire Array Film on the Photoelectrochemical Performance. J. Mater. Sci. Technol. 2012, 28, 594−598. (26) Kim, S.; Jang, S.-R.; Vittal, R.; Lee, J.; Kim, K.-J. Rutile TiO2Modified Multi-Wall Carbon Nanotubes in TiO2 Film Electrodes for Dye-Sensitized Solar Cells. J. Appl. Electrochem. 2006, 36, 1433−1439. (27) Wang, W.; Serp, P.; Kalck, P.; Faria, J. L. Visible light Photodegradation of Phenol on MWNT−TiO2 Composite Catalysts Prepared by a Modified Sol−Gel Method. J. Mol. Catal. A: Chem. 2005, 235, 194−199. (28) Zhang, Y.; Zhang, N.; Tang, Z.-R.; Xu, Y.-J. Graphene Transforms Wide Band Gap ZnS to a Visible Light Photocatalyst. The New Role of Graphene as a Macromolecular Photosensitizer. ACS Nano 2012, 6, 9777−9789. (29) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemann, D. W. Environmental Applications of Semiconductor Photocatalysis. Chem. Rev. 1995, 95, 69−96. (30) Yao, Y.; Li, G.; Ciston, S.; Lueptow, R. M.; Gray, K. A. Photoreactive TiO2/Carbon Nanotube Composites: Synthesis and Reactivity. Environ. Sci. Technol. 2008, 42, 4952−4957. (31) Blöchl, P. E. Projector Augmented-Wave Method. Phys. Rev. B 1994, 50, 17953−17979. (32) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B 1996, 54, 11169−11186. (33) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77, 3865−3868. (34) Monkhorst, H. J.; Pack, J. D. Special Points for Brillouin-Zone Integrations. Phys. Rev. B 1976, 13, 5188−5192. (35) Dudarev, S. L.; Botton, G. A.; Savrasov, S. Y.; Humphreys, C. J.; Sutton, A. P. Electron-Energy-Loss Spectra and the Structural Stability of Nickel Oxide: An LSDA+U Study. Phys. Rev. B 1998, 57, 1505− 1509. (36) Long, R.; English, N. J.; Prezhdo, O. V. Photo-induced Charge Separation across the Graphene−TiO2 Interface Is Faster than Energy Losses: A Time-Domain Ab Initio Analysis. J. Am. Chem. Soc. 2012, 134, 14238−14248. (37) Du, A.; Ng, Y. H.; Bell, N. J.; Zhu, Z.; Amal, R.; Smith, S. C. Hybrid Graphene/Titania Nanocomposite: Interface Charge Transfer, Hole Doping, and Sensitization for Visible Light Response. J. Phys. Chem. Lett. 2011, 2, 894−899. (38) Sanville, E.; Kenny, S. D.; Smith, R.; Henkelman, G. Improved Grid-Based Algorithm for Bader Charge Allocation. J. Comput. Chem. 2007, 28, 899−908. (39) Su, W. S.; Leung, T. C.; Chan, C. T. Work Function of SingleWalled and Multiwalled Carbon Nanotubes: First-Principles Study. Phys. Rev. B 2007, 76, 235413. (40) Shiraishi, M.; Ata, M. Work Function of Carbon Nanotubes. Carbon 2001, 39, 1913−1917. (41) Long, R.; English, N. J. Energetic and Electronic Properties of P Doping at the Rutile TiO2 (110) Surface from First Principles. J. Phys. Chem. C 2009, 113, 9423−9430. (42) Long, R.; English, N. J.; Dai, Y. First-Principles Study of S Doping at the Rutile TiO2 (110) Surface. J. Phys. Chem. C 2009, 113, 17464−17470. (43) Yang, K.; Dai, Y.; Huang, B.; Whangbo, M.-H. Density Functional Characterization of the Visible-Light Absorption in Substitutional C-Anion- and C-Cation-Doped TiO2. J. Phys. Chem. C 2009, 113, 2624−2629. 1345

dx.doi.org/10.1021/jz400589v | J. Phys. Chem. Lett. 2013, 4, 1340−1346

The Journal of Physical Chemistry Letters

Letter

(44) Craig, C. F.; Duncan, W. R.; Prezhdo, O. V. Trajectory Surface Hopping in the Time-Dependent Kohn-Sham Approach for ElectronNuclear Dynamics. Phys. Rev. Lett. 2005, 95, 163001. (45) Jaeger, H. M.; Fischer, S.; Prezhdo, O. V. Decoherence-Induced Surface Hopping. J. Chem. Phys. 2012, 137, 22A545/1−22A545/14. (46) Maultzsch, J.; Pomraenke, R.; Reich, S.; Chang, E.; Prezzi, D.; Ruini, A.; Molinari, E.; Strano, M. S.; Thomsen, C.; Lienau, C. Exciton Binding Energies in Carbon Nanotubes from Two-Photon Photoluminescence. Phys. Rev. B 2005, 72, 241402.

1346

dx.doi.org/10.1021/jz400589v | J. Phys. Chem. Lett. 2013, 4, 1340−1346