Anatase - American Chemical Society

Letter www.acsami.org

Anatase (101)-like Structural Model Revealed for Metastable Rutile TiO2(011) Surface Meiling Xu,†,‡ Sen Shao,† Bo Gao,† Jian Lv,§ Quan Li,§,† Yanchao Wang,*,† Hui Wang,*,† Lijun Zhang,*,§,† and Yanming Ma†,⊥ †

State Key Lab of Superhard Materials, Department of Physics, §College of Materials Science and Engineering and Key Laboratory of Automobile Materials of MOE, and ⊥International Center of Future Science,Jilin University, Changchun 130012, China ‡ Beijing Computational Science Research Center, Beijing 100084, China ABSTRACT: Titanium dioxide has been widely used as an efficient transition metal oxide photocatalyst. However, its photocatalytic activity is limited to the ultraviolet spectrum range due to the large bandgap beyond 3 eV. Efforts to reduce the bandgap to achieve a broader spectrum range of light absorption have been successfully attempted via the experimental synthesis of dopant-free metastable surface structures of rutile-type TiO2 (011) 2 × 1. This new surface phase possesses a reduced bandgap of ∼2.1 eV, showing great potential for an excellent photocatalyst covering a wide range of visible light. There is a need to establish the atomistic structure of this metastable surface to understand the physical cause for the bandgap reduction and to improve the future design of photocatalysts. Here, we report computational investigations in an effort to unravel this surface structure via swarm structure-searching simulations. The established structure adopts the anatase (101)-like structure model, where the topmost 2-fold O atoms form a quasi-hexagonal surface pattern and bond with the unsaturated 5-fold and 4-fold Ti atoms in the next layer. The predicted anatase (101)-like surface model can naturally explain the experimental observation of the STM images, the electronic bandgap, and the oxidation state of Ti4+. Dangling bonds on the anatase (101)-like surface are abundant making it a superior photocatalyst. First-principles molecular dynamics simulations have supported the high photocatalytic activity by showing that water and formic acid molecules dissociate spontaneously on the anatase (101)-like surface. KEYWORDS: titanium dioxide, surface reconstruction, anatase, reduced bandgap, dopant-free, CALYPSO

■

INTRODUCTION Since the discovery of the Honda−Fujishima effect on n-type TiO2 electrodes under ultraviolet light in 1972,1 TiO2 has attracted a great deal of research interest to make use of its remarkable photocatalytic properties. TiO2 has been widely applied in the splitting and purifying of water, removal of organic pollutants, and air self-cleaning/disinfecting of coatings, as well as in photoelectrochemical cells.2−8 However, the actual application of TiO2 has been limited to the ultraviolet spectrum range because of its large bandgap beyond 3 eV.9 Narrowing the bandgap to achieve the adsorption of light in a broader spectrum range has been a persistent challenge. Many attempts, such as noble metal loading,10 composite semiconducting,11,12 metal ion doping,13 nonmetal elemental (N, C, and S) doping,14−16 and metal and nonmetal elemental codoping,17−19 have been made to tune the bandgap from a variety of aspects to extend the spectral response to the solar spectrum. Among these works, doping efforts have been demonstrated to be the most efficient method,20 where impurity states are introduced into the band gap, enabling the absorption of visible solar energy. However, the use of dopants would inevitably introduce charge traps that act as © 2017 American Chemical Society

recombination centers for the photogenerated charge carriers. It is therefore highly desirable to establish an alternative strategy to reduce the bandgap without the introduction of any dopants. Surface atoms are usually not saturated and have many dangling bonds, truncated from the chemically stoichiometric bulk.21 Dangling bonds can cause high surface activities to stabilize the reconstructed surface phases,2,22 deviating largely from the bulk structures. These surface phases can introduce electronic states into the bandgap, causing significantly reduced bandgaps for long-tail optical absorption and red-shifted optical emission.2,21 For example, the reconstructed anatase (001) and (101) surfaces have reduced bandgaps of ∼2.323 and 2.024 eV, respectively, showing visible light activity in the photooxidation of formic acid,23 whereas the reconstructed rutile (011) surface possesses a band gap of ∼2.1 eV.25 These experimental efforts demonstrated that the electronic and optical properties of TiO2 Received: December 22, 2016 Accepted: February 23, 2017 Published: February 23, 2017 7891

DOI: 10.1021/acsami.6b16449 ACS Appl. Mater. Interfaces 2017, 9, 7891−7896

Letter

ACS Applied Materials & Interfaces 2/3

H are needed to passivate the Ti dangling bonds and the O dangling bonds, respectively.48 For adatoms, four multiplications of the unit cell are considered (namely, TiO2, Ti2O4, Ti3O6, and Ti4O8). The 2000 structures we searched are ranked according to the calculated enthalpy. Once we have the low-energy structures at hand, finer parameters are used for calculations of surface energies and electronic properties. To obtain accurate relative surface energies, a symmetric slab containing 11 TiO2 layers and the vacuum layer of 18 Å was used for structural relaxation without any constraints and for electronic structural calculations. The surface energy that determines the stability of a surface is defined as

can be efficiently improved by self-structured modification via surface reconstruction. Because of the scientific and technical importance, reconstructed surfaces of TiO 2 have been intensively investigated in recent decades.26−36 The studies on rutile (011)-2 × 1 surface reconstruction via the examination of titanyl,37 microfaceting missing-row,38 and brookite (001)-like structural models39,40 have found that the brookite (001)-like structure is the energetically most stable one, whose bandgap, however, is similar to that of the bulk rutile phase. Recently, a metastable two-dimensional (2D) phase formed at the rutile (011) surface with a markedly reduced bandgap of ∼2.1 eV has been synthesized,25 showing great potential as an excellent photocatalyst covering a wide range of visible light. The 2D phase can be ascribed as a (2 × 1) surface unit cell, and the bright spots observed in the scanning tunneling microscope (STM) images are arranged in a (distorted) hexagonal arrangement. The atomic structure of this 2D phase has attracted much attention to understand the physical cause for the bandgap reduction and to better design the optimal photocatalyst.41,42 The titanyl model was first considered as the possible candidate for this reconstructed 2D surface structure.41 However, its simulated STM image cannot account for the experimental observation. There is a recent report on the prediction of an interesting Ti4O4-2 × 1 surface structure model42 that is able to explain the experimental STM data. This structure model breaks the ideal stoichiometry of TiO2 and gives rise to the charge state of Ti2+. Experimentally, the charge states of the Ti atoms in the metastable 2D (011) surface structure have been measured to be Ti4+ by the X-ray photoemission spectroscopy technique.25 In this work, we performed extensive structure searching studies for the metastable rutile (011) surface reconstruction via the swarm-intelligence based CALYPSO code in combination with first-principles total energy calculations. In our structure searches, a novel anatase (101)-like structure is established, with which the experimental observations from STM images, the electronic bandgap, and the oxidation state of Ti4+ are well explained. We further demonstrated, by performing first-principles molecular dynamics simulations, that the predicted anatase (101)-like surface has superior photocatalytic reactivity by efficiently dissociating water and formic acid molecules once adsorbed.

■

γ=

1 (Etot − Eref − N

∑ niμi )

(1)

where Etot and Eref are the total energy of the surface under consideration and of the reference cleaved surface, respectively; ni and μi are the number of excess atoms of the cleaved surface and the chemical potential for each species, respectively; and N is equal to (m × n) for an (m × n) surface cell and serves as a normalization factor. The chemical potentials must satisfy the following constraints: (1) μTi 1 ≤ μbulk Ti , (ii) μO ≤ 2 μO2 , and (3) μTi + 2 μO = ETiO2, where ETiO2 is the internal energy of the bulk rutile phase. Equation 1 can thus be written as γ=

1 [Etot − Eref − nTiE TiO2 − μO(nO − 2n Ti)] N

(2)

According to the constraints of the chemical potential, the range of μo 1 1 is set as 2 μO ≥ μO ≥ 2 (E TiO2 − μTi ). 2 The underlying structure relaxations, surface energy, and electronic calculations are performed with the plane-wave pseudopotential method as implemented in the VASP code.49 The Perdew−Burke− Ernzerhof generalized gradient approximation50 is chosen for the exchange-correlation functional, and a Hubbard U term (U = 4.2 eV) is considered to address the self-interaction error of the generalized gradient approximation.51 The electron−ion interaction is described by the projector augmented-wave potentials,52 with 3p63d34s1 and 2s22p4 configurations treated as valence electrons for Ti and O, respectively. A kinetic cutoff energy of 500 eV and Monkhorst−Pack k meshes with a grid spacing of 2π × 0.04 Å−1 are then adopted to give well converged total energies (∼1 meV/atom). The ionic positions are fully relaxed until the residual force acting on each ion is less than 0.02 eV/Å. In the electronic structure computation, the GGA + U (U = 4.2 eV) method51 is used to take the good account on the energy band gap. We also perform first principle Born−Oppenheimer molecular dynamics (MD) simulations as implemented in the CP2K-QUICKSTEP program.53,54 The electronic structures are calculated using density functional theory based on the mixed Gaussian plane wave approach. The local-density approximation to the exact exchange and correlation functional and Goedecker−Teter−Hutter pseudopotentials55 are employed. The auxiliary plane wave basis cutoff for the electron density is set to 600 Ry. The MD simulations are carried out with a time step of 0.5 fs, and the total simulation time of a single MD run is 20 ps. The temperature is controlled at 300 K with the Nosé− Hoover chain thermostat. The anatase (101)-like surface structure is modeled as a periodic slab with four TiO2 layers (∼9.5 Å thick), where the vacuum separation between slabs is ∼16 Å, and contains 32 TiO2 units in a 2 × 1 supercell with a surface area of 9.2 × 10.9 Å2. The adsorption energies are calculated according to the equation Eads = Eslab + Emol − Etotal, where Eslab, Emol, and Etotal represent the energies of the anatase (101)-like slab, the free molecule and the whole adsorption system, respectively.

COMPUTATIONAL METHODS

We perform a systematic surface structure searching study on the rutile (011) surface of TiO2 via the swarm-intelligence based CALYPSO (Crystal structure AnaLYsis by Particle Swarm Optimization) method and its same-name code, which has been widely applied to predict stable and metastable structures of extended crystals, low-dimensional materials, reconstructed surfaces, and nanoclusters.43−47 Our surface structure slab model typically consists of three regions: the bulk region and the unreconstructed and reconstructed surface regions (see ref 45 for details), where only the surface regions are structurally optimized. We used an asymmetric slab of 5 TiO2 layers and a vacuum layer of ∼10 Å for performing large-scale simulations of reconstructed surface structures searches using CALYPSO. This choice relies on a series of tests on the number of TiO2 layers used and the height of the vacuum layer chosen, giving a good converged surface energies (∼1.7 meV/ atom). Equally important, this set of parameters is computationally feasible for large-scale simulations of CALYPSO structure searches to ensure exploration of large configuration space that is necessary for a good structure searching simulation. During the structure search, to remove the detrimental effects of the surface dangling bonds, 4/3H and

■

RESULTS AND DISCUSSION Our structure searches independently identified the energetically stable brookite (001)-like structure39,40 and titanyl structure37 previously proposed. In addition, we predict a 7892


Letter

ACS Applied Materials & Interfaces novel metastable structure unseen in other works, which is depicted in Figure 1. The structure is slightly corrugated, with a

Figure 1. Predicted anatase (101)-like reconstructed surface structure in (a) the side view and (b) the top view. (c) Simulated (left panel) and experimental (right panel) STM images. The Ti and O atoms on the surface and subsurface oxide layer are shown in the ball and frame modes, respectively. Atoms in the gray color represent Ti atoms. Atoms in red, yellow, and green colors represent O atoms in the topmost, middle, and downmost positions of the surface oxide layer, respectively.

Figure 2. Structures of (a) the anatase (101) surface, (b) the unconstructed rutile (011) surface, and (c) the brookite (001)-like surface. The Ti and O atoms in the surface layer and in the sublayers are shown in the ball and frame modes, respectively. Atoms in gray and red colors represent Ti and O atoms, respectively.

characteristic sawtooth profile along the [100] direction (see Figure 1a). In this structure, the topmost 2-fold O atoms (O2c, in the red color) of the surface oxide layer form a quasihexagonal surface pattern, as depicted in Figure 1b (solid line), and they are bonded to two types of unsaturated Ti atoms, having 5-fold and 4-fold (denoted Ti5c and Ti4c in Figure 1a, b, respectively) coordination. The middle and downmost O atoms of the surface oxide layer are 2-fold (in the yellow color) and 3fold (O3c, in the green color), respectively. The surface oxide layer has a perfect stoichiometric form of TiO2 containing 4 Ti and 8 O atoms in a 2 × 1 supercell. The coordinates of the Ti and O atoms in the surface oxide layer are listed in Table 1. We find a large similarity of the newly predicted surface structure in the surface oxide layer to the ideal anatase (101) (Figure 2a) from the side view, which is the lowest-energy

surface of anatase.56 We therefore name it as the anatase (101)like model. In the atomic scale, the predicted anatase (101)-like model is less corrugated along the [100] direction than its unreconstructed (011) surface (Figure 2b). In the top view, the anatase (101)-like model is characterized by “rings” of 13 under-coordinated surface atoms, as depicted in Figure 1b (dashed line). The structure parameters of the anatase (101)-like model are calculated. The average interatomic O2c-Ti [O2c is the topmost O atoms in Figure 1a and O3c-Ti [O3c is the downmost O atoms in Figure 1a distances are 1.85 and 1.97 Å, respectively, which are the same as those of the ideal anatase (101) surface. However, the average interatomic distance between Ti and O, colored by yellow in Figure 1a in the middle region of the surface oxide layer, is 1.84 Å, smaller than that (1.99 Å) of the anatase (101) surface. The surface energies of the unreconstructed, titanyl,37 Ti4O42 × 1,42 and brookite (001)-like structures as a function of the O chemical potential have been calculated, as shown in Figure 3a. The brookite (001)-like structure (see Figure 2c) possesses the lowest surface energy, while the anatase (101)-like structure is metastable with a surface energy of 1.11 eV per 1 × 1 unit cell, higher than that of the brookite (001)-like structure. The physical origin for the metastability of the anatase (101)-like structure stems from the larger number of dangling bonds compared to that in the brookite (001)-like surface, arising from the unsaturated Ti atoms. It should be noted that the unsaturated Ti atoms on the surface of the brookite (001)-like structure are 5-fold, whereas there are two types of unsaturated Ti atoms on the anatase (101)-like surface, 5-fold Ti5c and 4fold Ti4c, shown in Figures 1a, b, respectively. In Figure 3a, it is

Table 1. Coordinates of the Ti and O Atoms on the Surface Oxide Layer of the Anatase (001)-like Model anatase (101)-like coordinates (Å) atom

x

y

z

Ti-1 Ti-2 Ti-3 Ti-4 O-1 O-2 O-3 O-4 O-5 O-6 O-7 O-8

7.124 0.552 5.145 2.470 5.604 3.804 2.158 3.715 0.923 6.657 8.010 8.659

4.986 0.790 2.192 3.519 0.525 2.883 5.357 2.947 2.519 3.184 0.219 0.067

33.586 33.027 33.202 33.240 33.880 34.360 33.353 31.914 33.553 33.578 32.028 34.431 7893


Letter

ACS Applied Materials & Interfaces

The photoreactivity of a crystal facet is intimately related to both its surface atomic structure and surface electronic band structure.58 For the anatase (101)-like model, on the one hand, the number of dangling bonds is large; on the other hand, the upper edge of the valence band is actually formed by the states from the surface O atoms. Therefore, the anatase (101)-like model might have excellent photocatalytic properties for applications. Below, we perform MD simulations to examine the photocatalytic activity of the anatase (101)-like structure via the adsorption of typical water and formic acid molecules on the surface. Our MD simulations show that both water and formic acid molecules dissociate spontaneously on the anatase (101)-like surface. The dissociation is closely related to the computed high adsorption energies of 1.61 and 2.06 eV for water and formic acid molecules, respectively. The resultant dissociation structures are illustrated in Figure 4. From the right panel of Figure 4a, we can see that the OH

Figure 3. (a) Computed surface energies of various structures as a function of the chemical potential of O. (b) Calculated projected density of states (PDOS) for the anatase (101)-like structure. As an example, the label “Ti-d” represents the PDOS for the d orbital of Ti atoms.

seen that the anatase (101)-like structure is energetically more favorable than the unreconstructed surface by 0.23 eV, the earlier proposed titanyl model37 by 0.31 eV and more stable than the Ti4O4-2 × 1 model42 for 0 ≤ ΔμO ≤ −4.28 eV, respectively. The STM images of the anatase (101)-like structure are simulated using the Tersoff method57 under positive sample biasing conditions with four energy bands and a fixed distance of 1 Å above the topmost O atoms. As shown in Figure 1c, the calculated STM images show clear pseudohexagonal patterns, in excellent agreement with the experimental observations.25 We further examined the calculated STM image by extending distances up to 3 Å, within which the pseudohexagonal STM pattern remains unaltered. The computed projected densities of states (PDOSs) of the predicted anatase (101)-like surface model are shown in Figure 3b. It is seen that the valence and conduction states are derived mainly from the O 2p orbitals and the Ti 3d orbitals, respectively. No midgap states are seen within the band gap region since there are no dopants introduced. The resultant band gaps for the anatase (101)-like surface structure phase is about 2.6 eV using the GGA+U method. It is noted that the band gap of the experimentally metastable TiO2 (011)-(2 × 1) reconstruction is approximately ∼2.1 eV, which agrees with our theoretical results. This supports our suggestion that the anatase (101)-like surface structure might be the best candidate structure for the metastable rutile TiO2(011)-2 × 1 reconstructed surface.

Figure 4. Structures of (a) water and (b) formic acid molecules on the anatase (101)-like surface. The left and right panels are the structures before and after dissociation, respectively. The anatase (101)-like structure and adsorption molecules (water and formic acid) are shown in the frame and ball modes, respectively. Atoms in dark gray, light gray, and red colors represent the C, H, and O atoms of water and formic acid molecules, respectively.

moiety (a hydroxyl group) of the dissociated H2O exhibits the bidentate adsorption configuration, in which the hydroxyl group binds to the surface through two O−Ti4c and O−Ti5c bonds with bond lengths of 2.03 and 1.95 Å, respectively, and H binds to the nearby topmost O2c atoms. This bidentate chelating configuration of water adsorption on the Ti atoms at the sides of each O2c has not been previously seen. After accepting the hydroxyl group, the O2c atom moves slightly downward. In the same manner, the formic acid molecule spontaneously dissociates into anionic formate (HCOO−) and cationic H+ on the anatase (101)-like surface, as shown in the right panel of Figure 4b. The formate bonds to the surface Ti5c and Ti4c 7894


Letter

ACS Applied Materials & Interfaces

(9) Kavan, L.; Grätzel, M.; Gilbert, S. E.; Klemenz, C.; Scheel, H. J. Electrochemical and Photoelectrochemical Investigation of SingleCrystal Anatase. J. Am. Chem. Soc. 1996, 118 (28), 6716−6723. (10) Anpo, M. The Design and Development of Highly Reactive Titanium Oxide Photocatalysts Operating under Visible Light Irradiation. J. Catal. 2003, 216 (1−2), 505−516. (11) Chen, X.; Shen, S.; Guo, L.; Mao, S. S. Semiconductor-Based Photocatalytic Hydrogen Generation. Chem. Rev. 2010, 110 (11), 6503−6570. (12) Kumar, S. G.; Devi, L. G. Review on Modified TiO 2 Photocatalysis under UV/Visible Light: Selected Results and Related Mechanisms on Interfacial Charge Carrier Transfer Dynamics. J. Phys. Chem. A 2011, 115 (46), 13211−13241. (13) Choi, W.; Termin, A.; Hoffmann, M. R. The Role of Metal Ion Dopants in Quantum-Sized TiO2: Correlation between Photoreactivity and Charge Carrier Recombination Dynamics. J. Phys. Chem. 1994, 98 (51), 13669−13679. (14) Asahi, R. Visible-Light Photocatalysis in Nitrogen-Doped Titanium Oxides. Science 2001, 293 (5528), 269−271. (15) Wu, G.; Wang, J.; Thomas, D. F.; Chen, A. Synthesis of FDoped Flower-like TiO2 Nanostructures with High Photoelectrochemical Activity. Langmuir 2008, 24 (7), 3503−3509. (16) Cho, I. S.; Lee, C. H.; Feng, Y.; Logar, M.; Rao, P. M.; Cai, L.; Kim, D. R.; Sinclair, R.; Zheng, X. Codoping Titanium Dioxide Nanowires with Tungsten and Carbon for Enhanced Photoelectrochemical Performance. Nat. Commun. 2013, 4, 1723. (17) 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 (3), 23−26. (18) Zhang, J.; Pan, C.; Fang, P.; Wei, J.; Xiong, R. Mo + C Codoped TiO2 Using Thermal Oxidation for Enhancing Photocatalytic Activity. ACS Appl. Mater. Interfaces 2010, 2 (4), 1173−1176. (19) Liu, Y.; Shu, W.; Chen, K.; Peng, Z.; Chen, W. Enhanced Photothermocatalytic Synergetic Activity toward Gaseous Benzene for Mo+C-Codoped Titanate Nanobelts. ACS Catal. 2012, 2 (12), 2557− 2565. (20) Khan, S. U. M. Efficient Photochemical Water Splitting by a Chemically Modified N-TiO2. Science 2002, 297 (5590), 2243−2245. (21) Liu, L.; Chen, X. Titanium Dioxide Nanomaterials: SelfStructural Modifications. Chem. Rev. 2014, 114 (19), 9890−9918. (22) Martra, G. Lewis Acid and Base Sites at the Surface of Microcrystalline Anatase. Appl. Catal., A 2000, 200 (1), 275−285. (23) Ariga, H.; Taniike, T.; Morikawa, H.; Tada, M.; Min, B. K.; Watanabe, K.; Matsumoto, Y.; Ikeda, S.; Saiki, K.; Iwasawa, Y. SurfaceMediated Visible-Light Photo-Oxidation on Pure TiO2 (001). J. Am. Chem. Soc. 2009, 131 (41), 14670−14672. (24) Dette, C.; Pérez-Osorio, M. a.; Kley, C. S.; Punke, P.; Patrick, C. E.; Jacobson, P.; Giustino, F.; Jung, S. J.; Kern, K. TiO2 Anatase with a Bandgap in the Visible Region. Nano Lett. 2014, 14 (11), 6533−6538. (25) Tao, J.; Luttrell, T.; Batzill, M. A Two-Dimensional Phase of TiO2 with a Reduced Bandgap. Nat. Chem. 2011, 3 (4), 296−300. (26) Bennett, R. A.; Stone, P.; Price, N. J.; Bowker, M. Two (1 × 2) Reconstructions of TiO2(110): Surface Rearrangement and Reactivity Studied Using Elevated Temperature Scanning Tunneling Microscopy R. Phys. Rev. Lett. 1999, 82 (19), 3831−3834. (27) Blanco-Rey, M.; Abad, J.; Rogero, C.; Mendez, J.; Lopez, M. F.; Martin-Gago, J. A.; de Andres, P. L. Structure of Rutile TiO2(110)-(1 × 2): Formation of Ti2O3 Quasi-1D Metallic Chains. Phys. Rev. Lett. 2006, 96 (5), 055502. (28) Herman, G. S.; Sievers, M. R.; Gao, Y. Structure Determination of the Two-Domain (1 × 4) Anatase TiO2(001) Surface. Phys. Rev. Lett. 2000, 84 (15), 3354−3357. (29) Lazzeri, M.; Selloni, A. Stress-Driven Reconstruction of an Oxide Surface: The Anatase TiO2(001)-(1 × 4) Surface. Phys. Rev. Lett. 2001, 87 (26), 266105. (30) Li, M.; Hebenstreit, W.; Diebold, U. Morphology Change of Oxygen-restructured TiO2(110) Surfaces by UHV Annealing: Formation of a Low-temperature (1 × 2) Structure. Phys. Rev. B: Condens. Matter Mater. Phys. 2000, 61 (7), 4926−4933.

atoms in a bidentate adsorption configuration with bond lengths of 1.93 and 2.10 Å, respectively, and the dissociated H+ is adsorbed on the adjacent O2c atoms with a bond length of 0.93 Å. This adsorption behavior is commonly formed when a formic acid molecule dissociates at the surface of TiO22,59 Unbiased swarm structure searches and density functional total energy calculations are performed to explore the atomistic structures of the experimentally reconstructed metastable rutile (011) surface with a reduced bandgap of 2.1 eV. There is excellent mutual agreement between experiments and theory in terms of the STM image, the electronic bandgap, and the charge state of the Ti atoms, supporting the anatase (101)-like model as the likely candidate structure of the experimental metastable surface. Our MD simulations certified that the predicted anatase (101)-like surface has a superior photocatalytic property by efficiently dissociating water and formic acid molecules.

■

AUTHOR INFORMATION

Corresponding Authors

*E-mail: [email protected]. *E-mail: [email protected]. *E-mail: [email protected]. ORCID

Meiling Xu: 0000-0001-6592-8975 Yanchao Wang: 0000-0003-4518-925X Lijun Zhang: 0000-0002-6438-5486 Notes

The authors declare no competing financial interest.

■

ACKNOWLEDGMENTS The authors acknowledge funding support from the National Natural Science Foundation of China (under Grants 11534003 and 11274136), the National Key Research and Development Program of China (under Grant 2016YFB0201200), the Science Challenge Project (JCKY2016212A501), and the 2012 Changjiang Scholar of Ministry of Education. Part of the calculations were performed in the high-performance computing center of Jilin University and Tianhe2-JK in the Beijing Computational Science Research Center.

■

REFERENCES

(1) Fujishima, A.; Honda, K. Electrochemical Photolysis of Water at a Semiconductor Electrode. Nature 1972, 238, 37−38. (2) Diebold, U. The Surface Science of Titanium Dioxide. Surf. Sci. Rep. 2003, 48 (5−8), 53−229. (3) O’Regan, B.; Grätzel, M. A Low-Cost, High-Efficiency Solar Cell Based on Dye-Sensitized Colloidal TiO2 Films. Nature 1991, 353 (6346), 737−740. (4) Bard, A. J.; Fox, M. A. Artificial Photosynthesis: Solar Splitting of Water to Hydrogen and Oxygen. Acc. Chem. Res. 1995, 28 (3), 141− 145. (5) Liu, L.; Zhao, H.; Andino, J. M.; Li, Y. Photocatalytic CO2 Reduction with H2O on TiO2 Nanocrystals: Comparison of Anatase, Rutile, and Brookite Polymorphs and Exploration of Surface Chemistry. ACS Catal. 2012, 2 (8), 1817−1828. (6) Yaghoubi, H.; Taghavinia, N.; Alamdari, E. K.; Volinsky, A. A. Nanomechanical Properties of TiO2 Granular Thin Films. ACS Appl. Mater. Interfaces 2010, 2 (9), 2629−2636. (7) Ismail, A. a.; Bahnemann, D. W. Mesoporous Titania Photocatalysts: Preparation, Characterization and Reaction Mechanisms. J. Mater. Chem. 2011, 21 (32), 11686−11707. (8) Fujishima, A.; Zhang, X.; Tryk, D. A. TiO2 Photocatalysis and Related Surface Phenomena. Surf. Sci. Rep. 2008, 63 (12), 515−582. 7895


Letter

ACS Applied Materials & Interfaces (31) Liang, Y.; Gan, S.; Chambers, S. A.; Altman, E. I. Surface Structure of Anatase TiO2(001) Reconstruction, Atomic Steps, and Domains. Phys. Rev. B: Condens. Matter Mater. Phys. 2001, 63 (23), 235402. (32) Onishi, H.; Iwasawa, Y. Dynamic Visualization of a Metal-OxideSurface/Gas-Phase Reaction: Time-Resolved Observation by Scanning Tunneling Microscopy at 800 K. Phys. Rev. Lett. 1996, 76 (5), 791− 794. (33) Park, K. T.; Pan, M. H.; Meunier, V.; Plummer, E. W. Surface Reconstructions of TiO2(110) Driven by Suboxides. Phys. Rev. Lett. 2006, 96 (22), 226105. (34) Park, K. T.; Pan, M.; Meunier, V.; Plummer, E. W. Reoxidation of TiO2(110) via Ti Interstitials and Line Defects. Phys. Rev. B: Condens. Matter Mater. Phys. 2007, 75 (24), 245415. (35) Wang, Q.; Oganov, A. R.; Zhu, Q.; Zhou, X.-F. New Reconstructions of the (110) Surface of Rutile TiO2 Predicted by an Evolutionary Method. Phys. Rev. Lett. 2014, 113 (26), 266101. (36) Pieper, H. H.; Venkataramani, K.; Torbrugge, S.; Bahr, S.; Lauritsen, J. V.; Besenbacher, F.; Kuhnle, A.; Reichling, M. Unravelling the Atomic Structure of Cross-Linked (1 × 2) TiO2(110). Phys. Chem. Chem. Phys. 2010, 12 (39), 12436−12441. (37) Beck, T.; Klust, A.; Batzill, M.; Diebold, U.; Di Valentin, C.; Selloni, A. Surface Structure of TiO2(011)-(2 × 1). Phys. Rev. Lett. 2004, 93 (3), 036104. (38) Kubo, T.; Orita, H.; Nozoye, H. Surface Structures of Rutile TiO2(011). J. Am. Chem. Soc. 2007, 129 (34), 10474−10478. (39) Gong, X.-Q.; Khorshidi, N.; Stierle, A.; Vonk, V.; Ellinger, C.; Dosch, H.; Cheng, H.; Selloni, A.; He, Y.; Dulub, O.; Diebold, U. The 2 × 1 Reconstruction of the Rutile TiO2(011) Surface: A Combined Density Functional Theory, X-Ray Diffraction, and Scanning Tunneling Microscopy Study. Surf. Sci. 2009, 603 (1), 138−144. (40) Torrelles, X.; Cabailh, G.; Lindsay, R.; Bikondoa, O.; Roy, J.; Zegenhagen, J.; Teobaldi, G.; Hofer, W. A.; Thornton, G. Geometric Structure of TiO2(011)(2 × 1). Phys. Rev. Lett. 2008, 101 (18), 185501. (41) Wang, Q.; Oganov, A. R.; Feya, O. D.; Zhu, Q.; Ma, D. The Unexpectedly Rich Reconstructions of Rutile TiO2(011)-(2 × 1) Surface and the Driving Forces behind Their Formation: An Ab Initio Evolutionary Study. Phys. Chem. Chem. Phys. 2016, 18 (29), 19549− 19556. (42) Zhou, R.; Li, D.; Qu, B.; Sun, X.; Zhang, B.; Zeng, X. C. Rutile TiO2(011)-2 × 1 Reconstructed Surfaces with Optical Absorption over the Visible Light Spectrum. ACS Appl. Mater. Interfaces 2016, 8 (40), 27403−27410. (43) Wang, Y.; Lv, J.; Zhu, L.; Ma, Y. Crystal Structure Prediction via Particle-swarm Optimization. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 82, 094116. (44) Lv, J.; Wang, Y.; Zhu, L.; Ma, Y. Predicted Novel High-Pressure Phases of Lithium. Phys. Rev. Lett. 2011, 106 (1), No. 015503, DOI: 10.1103/PhysRevLett.106.015503. (45) Lu, S.; Wang, Y.; Liu, H.; Miao, M.-S.; Ma, Y. Self-Assembled Ultrathin Nanotubes on Diamond (100) Surface. Nat. Commun. 2014, 5 (100), 3666. (46) Gao, B.; Shao, X.; Lv, J.; Wang, Y.; Ma, Y. Structure Prediction of Atoms Adsorbed on Two-Dimensional Layer Materials: Method and Applications. J. Phys. Chem. C 2015, 119 (34), 20111−20118. (47) Wang, Y.; Lv, J.; Zhu, L.; Ma, Y. CALYPSO: A Method for Crystal Structure Prediction. Comput. Phys. Commun. 2012, 183, 2063−2070. (48) Deng, H.-X.; Li, S.-S.; Li, J.; Wei, S.-H. Effect of Hydrogen Passivation on the Electronic Structure of Ionic Semiconductor Nanostructures. Phys. Rev. B: Condens. Matter Mater. Phys. 2012, 85 (19), 195328. (49) Kresse, G.; Furthmüller, J. Efficient Iterative Schemes for Ab Initio Total-Energy Calculations Using a Plane-Wave Basis Set. Phys. Rev. B: Condens. Matter Mater. Phys. 1996, 54 (16), 11169−11186. (50) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized Gradient Approximation Made Simple. Phys. Rev. Lett. 1996, 77 (18), 3865− 3868.

(51) Dudarev, S. L.; 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: Condens. Matter Mater. Phys. 1998, 57 (3), 1505−1509. (52) Kresse, G.; Joubert, D. From Ultrasoft Pseudopotentials to the Projector Augmented-Wave Method. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59 (3), 1758−1775. (53) Vandevondele, J.; Krack, M.; Mohamed, F.; Parrinello, M.; Chassaing, T.; Hutter, J. Quickstep: Fast and Accurate Density Functional Calculations Using a Mixed Gaussian and Plane Waves Approach. Comput. Phys. Commun. 2005, 167 (2), 103−128. (54) CP2K: Open Source Molecular Dynamics, Development Version; http://www.cp2k.org. (55) Vandevondele, J.; Hutter, J. Gaussian Basis Sets for Accurate Calculations on Molecular Systems in Gas and Condensed Phases. J. Chem. Phys. 2007, 127 (11), 114105. (56) Lazzeri, M.; Vittadini, A.; Selloni, A. Structure and Energetics of Stoichiometric TiO2 Anatase Surfaces. Phys. Rev. B: Condens. Matter Mater. Phys. 2001, 63 (15), 155409. (57) Tersoff, J.; Hamann, D. R. Theory of the Scanning Tunneling Microscope. Phys. Rev. B: Condens. Matter Mater. Phys. 1985, 31 (2), 805−813. (58) Pan, J.; Liu, G.; Lu, G. Q. M.; Cheng, H.-M. On the True Photoreactivity Order of {001}, {010}, and {101} Facets of Anatase TiO2 Crystals. Angew. Chem., Int. Ed. 2011, 50 (9), 2133−2137. (59) Gong, X.; Selloni, A.; Vittadini, A. Density Functional Theory Study of Formic Acid Adsorption on Anatase TiO2 (001): Geometries, Energetics, and Effects of Coverage, Hydration, and Reconstruction. J. Phys. Chem. B 2006, 110 (6), 2804−2811.

7896
