Boron Environments in B-Doped and (B, N)-Codoped TiO2

ARTICLE pubs.acs.org/JPCC

Boron Environments in B-Doped and (B, N)-Codoped TiO2 Photocatalysts: A Combined Solid-State NMR and Theoretical Calculation Study Ningdong Feng,† Anmin Zheng,† Qiang Wang,† Pingping Ren,† Xiuzhi Gao,† Shang-Bin Liu,‡ Zhurui Shen,§ Tiehong Chen,§ and Feng Deng†,* †

State Key Laboratory of Magnetic Resonance and Atomic and Molecular Physics, Wuhan Center for Magnetic Resonance, Wuhan Institute of Physics and Mathematics, Chinese Academy of Science, Wuhan 430071, China ‡ Institute of Atomic and Molecular Sciences, Academia Sinica, P.O. Box 23-166, Taipei 10617, Taiwan § Key Laboratory of Functional Polymer Materials of MOE, Department of Materials Chemistry, College of Chemistry and College of Physics, Nankai University, Tianjin 300071, China

bS Supporting Information ABSTRACT: The structures and local environments of boron species in B-doped and (B, N)-codoped TiO2 photocatalysts have been investigated by solid-state 11B NMR spectroscopy in conjunction with density functional theory (DFT) calculations. Up to seven different boron sites were identified in the B-doped anatase TiO2, which may be classified into three categories, including interstitial, bulk BO3/2 polymer, and surface boron species, and has been supported by results obtained from FT-IR and XPS spectroscopy as well as from DFT calculations. Two types of interstitial borons, namely the tricoordinated (T*)- and pseudotetrahedral-coordinated (Q*) borons, were observed in addition to the two types of bulk BO3/2 polymer and three types of surface B, in good agreement with experimental data. Further density of state analyses revealed that, compared to undoped TiO2, the T* species in boron-doped TiO2 are solely responsible for the observed increase in energy band gap, whereas the presence of Q* species tend to lead to a decrease in band gap and hence are more favorable for the absorption in the visible-light region. In comparison with B- and N-doped TiO2, (B, N)-codoped TiO2 tends to exhibit a much higher visible-light photocatalytic activity for the oxidation of rhodamine B. Accordingly, a photochemical mechanism of the (B, N)-codoped TiO2 under visible-light irradiation is proposed.

1. INTRODUCTION Owing to its high efficiency, nontoxicity, and low cost, titanium dioxide (TiO2) with anatase structure has been recognized as one of the most promising photocatalysts in various fields, including energy storage and removal of organic pollutants.1-3 However, because of the wide band gap (3.2 eV) possessed by the anatase TiO2, ultraviolet (UV) radiation is a prerequisite to facilitate the desirable electron-hole separation during the catalytic reaction. In view of the fact that an overwhelming majority of the solar radiation on earth is within the visible-light (vis) region and that UV light only contributes ca. 5% to the natural-light spectrum; thus, the improvement of the visible-light absorption of anatase TiO2 has become one of the most demanding tasks in photochemical research and development. To unravel this problem, one of the most common and promising approaches is to dope the anatase TiO2 by nonmetallic elements such as C, B, N, P, and F, which has been shown effective in narrowing the band gap, thus extending the absorption band to the visible-light region.1-5 It has been shown6 that by doping TiO2 with boron and nickel oxides, a significant improvement in visible-light absorption and r 2011 American Chemical Society

photocatalytic activity may be achieved. As such, considerable research attention has been focused on the structures of boron species in B-doped TiO2, aiming to reveal their correlations with the observed photocatalytic properties. Thus far, the majority of the aforementioned research was carried out with X-ray photoelectron spectroscopy (XPS); however, the assignments of XPS signals largely remain controversial. For instance, some reports ascribed the peak at 190.5-191.8 eV, which resembles the typical binding energy of TiB2 (187.0 eV), to oxygen (O) atom-substituted B sites in the anatase TiO2,6-10 whereas others attributed the signal at 191.0-192.0 eV to B weaving into the interstitial sites of the lattice.11-13 In this context, 11B (I = 3/2) solid-state nuclear magnetic resonance (SSNMR) spectroscopy represents a direct and powerful technique in providing structural information on the local environments of B-containing substances, such as zeolites, glasses, and minerals.14-18 For example, Raftery and co-workers17 Received: August 24, 2010 Revised: December 8, 2010 Published: January 20, 2011 2709

dx.doi.org/10.1021/jp108008a | J. Phys. Chem. C 2011, 115, 2709–2719

The Journal of Physical Chemistry C reported the sole presence of tetrahedral boron (BO4) species in boron- and fluorine (B, F)-codoped TiO2. Nonetheless, limited information can be obtained on B-doped TiO2 by means of conventional 11B SSNMR techniques due to severe overlap of the quadrupolar 11B NMR resonances even if the experiments were performed under magic-angle-spinning (MAS) conditions. Multiple-quantum (MQ) MAS NMR19-21 has been shown to be a unique technique for effective removal of the second-order quadrupolar interactions in solids containing nuclei with I > 1/2, leading to high-resolution spectra consisting of only isotropic peaks. However, the application of such a technique in B-doped TiO2 materials is still limited by both the low B content possessed by typical B-doped TiO2 samples and the low conversion efficiency from MQ to single-quantum (SQ) coherences. The objective of this work is to gain new insights toward the structure and catalytic activity during photochemical oxidation of rhodamine B (Rh B) near the visible-light region. To improve the sensitivity of the 11B NMR spectroscopy, we incorporated fast amplitude-modulated (FAM) radiofrequency (RF) pulse trains into the MQ MAS sequence, namely the so-called 3QzFAM MAS NMR technique,22-24 to investigate the detailed chemical environments of boron in B-doped TiO2 materials. In addition, the effects of the nonmetallic dopants on the visible-light activity of TiO2 were also studied by density functional theory (DFT) calculations. In particular, the geometries and density of states (DOSs) of B-, N-, S-, and P-doped TiO2 were calculated to correlate their optical properties with photocatalytic activities.25-30 Moreover, by incorporating the results obtained from DFT calculations with those obtained from 11B SSNMR experiments, detailed structures and optical properties of B-doped and (B, N)-codoped TiO2 were also elucidated.

2. EXPERIMENTAL SECTION 2.1. Sample Preparation. Various doped TiO2 samples prepared by the conventional sol-gel method have been made. Typically, known amounts of boric acid (with molar ratio B/Ti = 1:10) and ammonia (8 mL) were dissolved in 45 mL of a methanol/ water (8:1) mixture. Subsequently, 0.025 mol of titanium(IV) isopropoxide (Ti(OiPr)4; Alfa Aesar, 95%) was added dropwise to the solution mixture at 273 K under stirring. To ensure a complete hydrolysis of Ti(OiPr)4, the solution mixture was continuously stirred for about 10 h at room temperature (295 K). After being aged for an additional 24 h, the reagent was centrifuged, dried under vacuum, and then calcined in air at 673 K for 5 h to obtain the (B, N)-doped TiO2 sample. A similar method was adopted for the preparation of the undoped, N-doped, and B-doped TiO2 except for the different varieties of precursors added. On the basis of our NMR and XPS experiments, a B/Ti molar content of 0.061 and 0.067 was found for the B-doped and the (B, N)-codoped TiO2, respectively, whereas a N/Ti ratio of ca. 0.013 was determined for both the N-doped and the (B, N)-codoped TiO2. 2.2. Characterization Methods. The crystalline structures of various photocatalysts were determined by X-ray diffraction (XRD) on a Bruker D/max2550 instrument using Cu KR radiation (40 kV, 40 kmA). All UV-vis diffuse reflectance spectroscopic (DRS) studies were carried out on a Shimadzu UV 3150 spectrophotometer using BaSO4 as the reference. All Fouriertransform infrared (FT-IR) spectra were recorded on a Nexus470 FT-IR spectrometer under the transmission scheme by using the KBr pellet technique.

ARTICLE

XPS spectra were recorded on a Kratos Axis Ultra delay line detector (DLD) spectrometer equipped with a monochromated Al KR X-ray source (hν = 1486.6 eV), hybrid (magnetic/electrostatic) optics with a multichannel plate, and DLD. All XPS spectra were recorded using an aperture slot of 300 700 μm. Survey spectra were recorded with an energy of 160 eV, as compared to the high resolution spectra (40 eV). The accuracy of the XPS binding energies (BE) is 0.1 eV. All SSNMR experiments were performed on a Varian Infinitypuls-400 spectrometer using a Chemagnetic 4 mm doubleresonance probe. A Larmor frequency of 400.13 and 128.38 MHz, and a typical π/2 pulse length of 2.4 and 2.5 μs were adopted for 1H and 11B resonance, respectively. The excitation pulse length was adjusted to π/12 for the single-pulse 11B MAS experiments with 1H decoupling (field strength ca. 100 kHz), in which a repetition time of 2 s and a total of 20000 accumulated scans were used. For the 1Hf11B CP MAS NMR experiments, the Hartmann-Hahn condition was optimized using a solid H3BO3 sample with a contact time of 1 ms and a recycle delay of 1 s. The 11B 3Qz-FAM MAS NMR spectra were recorded with the pulse sequence proposed by Vega et al.,31 in which 64 t1 increments of 10 μs in the F1 dimension were acquired (collecting 1344 scans per t1 increment) with a recycle delay of 2 s under a sample spinning rate of 15 kHz. All 11B NMR chemical shifts were referenced to that of H3BO3 (0.1 M). All simulated parameters, viz. isotropic chemical shifts (δiso), quadrupolar coupling constants (QCC), and asymmetry parameters (η) of the 11B resonances were extracted by fitting the corresponding second-order quadrupolar line shape using the Dimfit program.32 2.3. Computational Models and Methods. A 2 2 2 supercell of anatase TiO2 was used to model the dopant structures of interstitial borons in TiO2, including those at the tricoordinated (T*) and pseudotetrahedral coordinated (Q*) sites. In our calculated model, the molar ratio of B/Ti is 0.031, similar to the actual dopant concentration (0.031). During the structure optimizations, the electron correlation effects were modeled using the generalized gradient approximation (GGA) level with PBEsol combination of exchange and correlation functionals.33 PBEsol is intended to improve PBE for equilibrium properties such as bond lengths and lattice parameters.34 The ultrasoft pseudopotentials were used for the structure optimizations. In addition, a plane-wave cutoff energy of 340 eV and a 2 2 1 Monkhorst-Pack k-point grid35 were adopted to sample the Brillouin zone. During structure optimization, the unit cell parameters (cell shape and volume) and the all the O-Ti atoms around the dopant B atom (i.e., B-O-Ti) were allowed to relax to their stable positions. On the basis of optimized crystallographic structures of the interstitial borons, the spin-polarized density of state (DOS) and quadrupole coupling constant (QCC) parameters were calculated by the GGA/PBE level using a 2 2 1 Monkhorst-Pack k-point, and the plane-wave cutoff energies were set to 340 and 550 eV for the DOS and QCC predictions. The QCCs of 11B nuclei were calculated using the gauge including projectoraugmented waves (GIPAW) method;36,37 such a method can obtain reliable NMR parameters (such as shielding tensor and QCC) for inorganic and organic systems, as demonstrated by our previous works.38 All structure optimizations, DOS, and QCC calculations were performed by CASTEP 5.0 codes utilizing the parallel 16 CPU in IBM-1350 cluster facilitated by the National Center for High-performance Computing (NCHC) in Taiwan. 2710

dx.doi.org/10.1021/jp108008a |J. Phys. Chem. C 2011, 115, 2709–2719

The Journal of Physical Chemistry C

ARTICLE

Scheme 1. Schematics of Possible Boron Species, Subdivided into Tetrahedral-Coordinated (Q) and Tricoordinated (T) Boronsa

Figure 1. XRD spectra of (a) pure, (b) B-doped, (c) N-doped, and (d) (B, N)-codoped TiO2.

Table 1. Physical Properties of Undoped and Various Doped TiO2 Samples lattice parameter (Å)a crystallite size (nm)

a-axis

c-axis

undoped

17.1

3.7774

9.4742

B-doped

23.2

3.7799

9.4943

N-doped (B, N)-codoped

27.1 24.6

3.7818 3.7774

9.4949 9.4847

TiO2 sample

a

Derived based on Bragg’s Law, 2dhkl sin θ = nλ (with d101), and the formula for a tetragonal unit cell: 1/(dhkl)2 = (h2 þ k2)/a2 þ i2/c2. Here, dhkl is the lattice spacing, θ is the incident angle, n is an integer, λ denotes the wavelength of the incident wave, (a, c) are the lattice parameters, and (h, k, l) represent the Miller indices.

a

Tetrahedral-coordinated (Q) and tricoordinated (T) borons are represented by the nomenclature Qn (n = 1-4) and Tm (m = 0-3), where the superscript n denotes the number of Ti atoms attached to oxygen atoms neighboring the tetrahedrally coordinated B and m indicates the number of bridging oxygen sites connecting the trigonal boron (i.e., B-O-B).

3. RESULTS AND DISCUSSION 3.1. XRD. Figure 1 shows the XRD patterns of various calcined TiO2 samples. Similar to the undoped TiO2, the N- and B-doped as well as the (B, N)-codoped samples all revealed the structure characteristic of pure anatase phase. That no detectable diffraction peaks responsible for the crystalline B2O3 were observed for the B-doped (Figure 1b) and (B, N)-codoped (Figure 1d) samples indicated a good dispersion of the doped boron in the TiO2 nanoparticles. Accordingly, the crystallite size and lattice parameters of various samples may be deduced by using Bragg’s Law (see Table 1). That the lattice parameter along the c-axis of the B-doped TiO2 is increased by ca. 0.02 Å compared to that of the pure (undoped) TiO2 is consistent with the swelling of the unit cell caused by the interstitial borons.11,39 Thus, it can be indicative that some of the doped B species in B-doped TiO2 are presented as interstitial B (namely T* or Q*; see Scheme 1). By the same token, the presence of interstitial N may also be inferred for the N-doped TiO2, likewise, the existence of interstitial B and N in the (B, N)-codoped TiO2. The presence of such interstitial B and N in various doped TiO2 photocatalysts was further confirmed by FT-IR and XPS measurements, as will be discussed below. That the crystallite sizes increase after the doping of both B and N is consistent with a recent report.40 The nonmetal doping can introduce residual charge, which may result in an increase in the number of surface O-H groups (which is evidenced by

Figure 2. FT-IR spectra of (a) B-doped and (b) (B, N)-codoped TiO2.

O1s XPS) and then an increase in the crystallite size. In contrast, although the direct evidence is still absent, Chen et al.11 proposed that the doped boron may reduce the surface energy of TiO2 nanoparticles, which may hinder the increase in the crystallite size. Probably different sample preparation procedures lead to the difference. 2711



ARTICLE

Figure 3. XPS spectra of various TiO2 samples. Left: B1s spectra of (a) B-doped and (b) (B, N)-doped TiO2. Middle: N1s spectra of (c) N-doped and (d) (B, N)-codoped TiO2. Right: Ti2p spectra of (e) pure, (f) B-doped, (g) N-doped, and (h) (B, N)-codoped TiO2.

3.2. FT-IR. The FT-IR spectra of the B-doped TiO2 (Figure 2a) mainly consists of three intense signals at 1640, 1398, and 1280 cm-1. The vibrational band at 1640 cm-1 may be unambiguously assigned to the surface-adsorbed water and hydroxyl groups.3,41 The band at ca. 1398 cm-1 may be ascribed to the presence of tricoordinated (in the form of B3þ) interstitial borons (denoted as T*; see Scheme 1), which tend to interact with ambient oxygen atoms to exhibit chemical environments similar to that of normal Ti-O-B11,42 whereas the band at 1280 cm-1 may be attributed to the stretching vibrations of the B-O bonds of ‘boroxol rings’ (i.e., T2(ring); see Scheme 1).43 In addition to these main features, three weaker shoulder peaks at 1120, 1030, and 940 cm-1 (marked with asterisks in Figure 2) were also observed, which may be ascribed to the valence vibrations of B-O in BO4 tetrahedra.44 Similar results were observed for the (B, N)-codoped TiO2 sample (Figure 2b). That no vibrational band at ca. 1200 cm-1, which may account for the B-O stretching in crystalline B2O3,7,45 was observed for both B-doped and (B, N)-codoped TiO2 indicates a high dispersion of the doped B in these photocatalysts. This observation is, of course, in excellent agreement with the XRD results (Figure 1b and 1d). In brief, the aforediscussed FT-IR results reveal that the boron species in the B-doped and (B, N)-codoped TiO2 mainly are present as tricoordinated interstitial B (T*) and ‘boroxol rings’ (T2(ring)); nevertheless, the existence of oxygen-substituted B may not be ruled out. 3.3. XPS. It has been demonstrated6-13 that the B 1s XPS spectrum may be used to identify the boron species on the surfaces of the doped TiO2. As shown in Figure 3a, the B1s XPS spectrum of the B-doped TiO2 exhibits the typical asymmetrical broad peak in the range of 190.0-193.5 eV,6-13,46 which may be deconvoluted (carried out by the XPSPEAK41 software package) into two overlapped peaks centered at 191.2 and 192.0 eV. The former peak at 191.2 eV may be ascribed to interstitial borons whereas the peak at 192.0 eV may be attributed to signal originated from BO3/2 polymer and/or surface BO3/2 species. Similar assignments could also be made for the B 1s XPS spectrum obtained from the (B, N)-codoped TiO2 (Figure 3b), from which two peaks centered at 191.5 and 192.1 eV may be inferred. Further spectral analyses revealed that the concentrations of the interstitial borons amount to ca. 47% and 54% of the total boron B-doped and the (B, N)-codoped TiO2, respectively. It is noteworthy that, upon introducing the second dopant (N) onto the B-doped TiO2, a slight blue-shift in the binding energy of interstitial B from 191.2 to

191.5 eV was observed, suggesting probably that the interstitial B and the doped N are in close spatial proximity. Nevertheless, since no peak at 190.6 eV was observed, it is indicative that B-N bonding should be absent in the (B, N)-codoped TiO2.8 The N 1s XPS spectrum of the solely N-doped TiO2 consists of two peaks at 399.5 and 401.0 eV (Figure 3c). In general, N 1s XPS peaks locate at 396-397 eV may be attributed to oxygen substituted N (i.e., Ti-N-Ti), whereas peaks appear at considerably higher binding energies (407-408 eV) to nitrites and nitrates.2 In this context, the nitrogen species present in our Ndoped TiO2 sample are thus anticipated to be in a lower oxidation state somewhere in-between the substituted N’s and nitrites/ nitrates. As such, we assign the observed signals at 399.5 and 401.0 eV to two types of interstitial N (i.e., Ti-N-O) species within the valence states of -1 and þ1, respectively.2 A similar N 1s XPS spectrum was also observed for the (B, N)-codoped TiO2 (Figure 3d), except that the signal at higher binding energy appears to blueshift by ca. 0.3 eV compared to the corresponding peak observed for the N-doped TiO2, which may be attributed to the presence of interstitial B atoms, which are in close proximity to the interstitial N atoms. In addition, the Ti 2p3/2 XPS spectrum of the pure TiO2 exhibits a peak responsible for at 458.5 eV in addition to the 2p1/2 peak centered at ca. 464.0 eV (Figure 3e). Upon doping the TiO2, a slight red-shift of the Ti 2p3/2 peak from 458.5 to 458.2 eV was observed for the solely B-doped (Figure 3f) and N-doped (Figure 3g) TiO2, which may suggest the presence of Ti3þ species in the doped samples.47 Accordingly, it can be concluded that doping TiO2 with B or N alone tends to provoke formation of trivalence Ti3þ species, in good agreement with the results reported recently.2,30 It is noteworthy that the Ti 2p3/2 XPS peak observed for the (B, N)-codoped TiO2 is very similar to that observed for the pure TiO2. This implies that the Ti3þ species are less likely to be formed in the (B, N)-codoped TiO2 compared to the solely B- or N-doped samples. Thus, it is indicative that by introducing the second dopant (N) onto the B-doped TiO2, conversion of Ti species from Ti4þ to Ti3þ may be largely hindered in the photocatalyst. 3.4. Solid-State NMR. To gain further insights into the local structure of various boron species in the B-doped and (B, N)codoped TiO2, we have performed 11B MAS NMR and 1Hf11B CP MAS NMR experiments. It is well-known that the chemical shift and line shape of 11B MAS NMR may be directly correlated 2712



ARTICLE

Figure 4. (a, c) 11B MAS- and (b, d) 1Hf11B CP MAS NMR spectra of B-doped and (B, N)-codoped TiO2.

Figure 5.

11

B 3Qz-FAM MAS NMR spectra of (a) B-doped and (b) (B, N)-codoped TiO2.

to the coordination environment of doped boron. The 11B MAS NMR spectra of both the B-doped (Figure 4a) and the (B, N)codoped (Figure 4c) TiO2 samples exhibit broad, overlapping multiplets due to borons located in various structural environments. These overlapped peaks may be further resolved by using the 11B 3Qz-FAM MAS NMR technique, by which the secondorder quadrupole interactions are effectively removed, as shown in Figure 5. As a consequence of the highly resolved resonance peaks available from the 11B 3Qz-FAM MAS NMR spectra (Figure 5), the various 11B resonances corresponding to different local environments may be differentiated. Note that the 2D spectra in Figure 5 represent correlations between the ‘normal’ spectrum (i.e., the 11B MAS NMR spectra shown in Figure 4a and 4c) in the F2 axis and the ‘isotropic’ spectrum (after eliminating second-order quadrupolar coupling) in the F1 axis. Accordingly,

up to five different B sites were identified for the solely B-doped (Figure 5a) as well as the (B, N)-codoped (Figure 5b) TiO2. In addition to these five different (B1-B5) sites, the presences of two additional sites (B6 and B7) with somewhat weaker signals were also being confirmed by the 1Hf11B CP MAS NMR technique, as shown in Figure 4b and 4d. In other words, up to seven components (Bk; k = 1-7) are needed to simulate the 11B MAS NMR spectra shown in Figure 4a and 4b for the B-doped and (B, N)-codoped TiO2. Accordingly, the simulated parameters, viz., δiso, QCC, and η corresponding to various B sites may be extracted by fitting the corresponding second-order quadrupolar line shape obtained from slices of the 11B 3Qz-FAM MAS NMR spectrum using the Dimfit program;32 the results are summarized in Table 2. As illustrated in Scheme 1, two types of coordination environments may be envisaged for borons in various inorganic 2713



ARTICLE

Table 2. List of NMR Parameters Derived from 11B MAS-NMR Results for Various Boron Species in B-Doped and (B, N)Codoped TiO2 content in doped TiO2 (%) type bulk BO3/2 polymer

a

site

δiso (ppm)

QCC (MHz)

η

assignment

a

B-doped

(B, N)-codoped

B1

14.8

2.51

0.08

BO3/2 network polymer

9.8

5.3

B2

16.4

2.42

0.20

‘boroxol rings’

27.4

20.1

interstitial B

B3 B4

18.9 2.9b

2.45 1.30b

0.17 -

interstitial T* interstitial Q*

40.5 10.3

42.3 13.8

surface B

B5

4.0b

1.03b

-

normal BO4

7.4

8.5

B6

20.6

2.37

0.16

surface BO3/2

3.9

8.6

B7

14.2c

-

-

highly hydrated BO4

0.7

1.4

Extracted by spectral fitting using the Dimfit program (see text). b Derived from eqs 1 and 2. c Apparent chemical shift.

materials, namely the tricoordinated borons (Tm; m = 0-3) that normally exhibit second-order quadrupolar line shapes, and the tetrahedral-coordinated borons (Qn; n = 1-4) that are more symmetrical and tend to show Gaussian line shapes. Clearly, resonances corresponding to sites B1, B2, B3, and B6 may be assigned to tricoordinated borons whereas those corresponding to sites B4, B5, and B7 should be associated with tetrahedral-coordinated borons. The correlations between the surface protons (such as OH or adsorbed H2O) and borons in the B-doped and B, N-codoped TiO2 may also be inferred from the 1Hf11B CP MAS NMR spectra shown in Figure 4b and 4d. That only three types of 11B NMR signals (B5, B6, and B7) prevail under the cross-polarization condition indicates that the 11B resonances mainly arise from boron species on the surfaces of the doped TiO2 nanoparticles and that these borons should also be in close proximity to the surface protons. Thus, signals corresponding to B5, B6, and B7 sites should be associated with surface borons. The δiso and QCCs (or PQ) for the B5 and B6 sites may be determined by the chemical shifts of the signal corresponding to the F1 and F2 axes (designated by δF1 and δF2) observed in the 3Qz-FAM MAS NMR spectra in Figure 5 based on the following equations:48 17 10 δF1 þ δF2 ð1Þ 27 27 rffiffiffiffiffiffiffiffiffiffiffiffiffi rffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffiffi 17 η2 ðδF1 - δF2 Þ ¼ QCC 1 þ ¼ ν0 675000 3 ð2Þ δiso ¼

PQ

where PQ and ν0 denote the estimated quadrupolar coupling constant and Larmor resonance frequency, respectively. Accordingly, the δiso values of 4.0 and 20.6 ppm and QCC values of 1.03 and 2.37 MHz were determined for the B5 and B6 site, respectively. Thus, among the three surface boron sites, the B6 site may be ascribed to tricoordinated surface BO3/2 in the forms of T1 and/or T0 (see Scheme 1)49 whereas the B5 site may be assigned due to typical tetrahedral-coordinated BO4 in the forms of Q3 and/or Q4 (Scheme 1).14,15 It is noteworthy that, compared to the B5 and B6 sites, the signal corresponding to the B7 site, which exhibited a Gaussian line shape centered at 14.2 ppm, was notably enhanced after the 1Hf11B cross-polarization process. This signal, which was nearly invisible (due to low concentration) in the 11B 3Qz-FAM MAS NMR spectra in Figure 5a and 5b, is most likely associated with boron in the form of Q2 or Q1 (Scheme 1),50

such as boric acid abundant with OH groups that absorbed on the surfaces of B-doped and (B, N)-codoped TiO2 nanoparticles. In a recent theoretical study, Finazzi et al. reported30 that two types of interstitial boron sites, namely the T* and Q* sites (Scheme 1), are stable in the interstice of anatase TiO2. That four 11 B signals (B1, B2, B3, and B4) vanished in the 1Hf11B CP MAS NMR spectra (Figure 4b and 4d) implies that these boron species present in bulk B-dope and (B, N)-codoped TiO2 particles are not in close proximity to the surface protons. Thus, the Gaussian signal at B4 with δiso = 2.9 ppm should be associated with interstitial Q* boron sites coordinating with four neighboring oxygen atoms. Furthermore, on the basis of the aforediscussed FT-IR and XPS results, the intense signal at B3 should be rationally assigned to interstitial T* boron sites connecting to three neighboring oxygens. Further quantitative analyses of the 11B MAS NMR spectra (Figure 4a and 4c) indicate that ca. 50.8% of the total doped boron species reside in the interstitial B sites (i.e., T* and Q* sties corresponding to the sum of B3 and B4 contents in Table 2) in the B-doped TiO2, in good agreement with the FT-IR and XPS results (ca. 48%). Similarly, a total content of ca. 56.1% was derived for the interstitial B sites in the (B, N)-codoped TiO2 (Table 2), in close resemblance to the FT-IR and XPS results (ca. 54%). Since the measured depth of XPS is ca. 3-6 nm, this may be the reason that the concentration of interstitial B determined by XPS is slightly lower (2-3%) than that obtained by NMR. It is well-known that the doped borons in TiO2 are prone to condense and form polymerized networks such as the bulk BO3/2 polymers identified in the aforementioned FT-IR (Figure 2) and XPS (Figure 3) spectra. By referring to the chemical shifts reported for the tricoordinated B species in borates,49 we were able to assign the signal at B1 with δiso = 14.8 ppm to BO3/2 network polymer (i.e., the T3 site; see Scheme 1). By the same token, the signal at B2 with δiso =16.4 ppm may be attributed to ‘boroxol rings’, that is the T2(ring) site shown in Scheme 1. These T3 and T2(ring) sites refer to borons in bulk BO3/2 polymer, which may exist in the defects or interfaces of the anatase B-doped or (B, N)codoped TiO2. Note that we did not observe a signal arising from the substituted boron in the 11B MAS NMR spectra (Figure 4a and 4c). This is in line with the findings by Finazzi et al., who concluded30 that substitutional borons in anatase TiO2 are metastable species that tend to convert into interstitial boron easily while annealing at high temperature. By comparing the 11B MAS NMR spectra obtained from the B-doped (Figure 4a) and (B, N)-codoped TiO2 (Figure 4c), we found that the signals at B1 and B2 sites were, respectively, decreased by 4.5 and 7.3% for the (B, N)-codoped sample, 2714



ARTICLE

Figure 6. Optimized geometries of the B-doped TiO2 for (a) tricoordinated (T*) and (b) pseudotetrahedral-coordinated (Q*) interstitial borons.

indicating a decrease of bulk BO3/2 polymer in the defects of the B-doped TiO2 upon introducing nitrogen as codopant. As a consequence of such dual element doping of TiO2, defects such as oxygen and titanium vacancies, and distortion of the anatase lattice around the interstitial B, are prone to form because of the local imbalance of charge.30,51-53 As such, some of the defects are likely filled with doped nitrogens, leading to the observed decrease of bulk BO3/2 polymer in the (B, N)-codoped TiO2. Should this be true, the doped nitrogens are most likely in close spatial proximities to the interstitial B sites. By the same token, that the 11B MAS NMR signal at B6 in Figure 4c was increased by ca. 4.7% compared to spectrum in Figure 4a indicates that more surface BO3/2 species reside on the surface of the (B, N)codoped TiO2 than on its solely B-doped counterpart. Indeed, the O 1s XPS spectra (see Supporting Information, Figure S1) of undoped, B-doped, N-doped, and (B, N)-codoped TiO2 all exhibited three binding energy peaks at 529.6, 531.6, and 532.5 eV, corresponding to that of Ti-O, the hydroxyl O-H group, and B-O, respectively.11 That a notable increase in the content of the surface O-H groups was observed for the (B, N)codoped TiO2 than the solely B-doped sample indicates that the introduction of N codopant promotes adsorption of more boric acids on the surfaces of TiO2 particles. Likewise, upon introducing N as codopant, the signal at B4 was found to increase by ca. 3.5%, implying the more preferable formation of interstitial Q* species in the (B, N)-codoped TiO2 than in the B-doped TiO2, leading to adsorption in the visible-light region (vide infra). 3.5. Theoretical Studies. As demonstrated by the 11B MAS NMR experiments, the interstitial borons found in the B-doped and (B, N)-codoped TiO2 mainly consist of T* and Q* species. It was reported11,30 that the interstitial boron would lead to the reduction of Ti4þ to Ti3þ. A certain amount of Ti3þ on the surface of TiO2 can improve the photoactivity, which can act as a shallow trap for the photogenerated electron to hinder the recombination of photoexcited electrons and holes. However, it also serves as a recombination center to impair the photoactivity. To obtain new insight into the structure and catalytic activity, it is of great importance to understand the detailed structures of the interstitial boron species in the B-doped and (B, N)codoped TiO2, as discussed below by means of results obtained from theoretical calculations.

The optimized structures of interstitial borons (T* and Q*) doped in TiO2, which were optimized at the GGA/PBE level based on the supercell model, are shown in Figure 6. Upon doping the anatase TiO2 with boron, the two Ti-O bonds adjacent to the interstitial T* boron site (Figure 6a) stretched to respective bond lengths of 1.944 and 2.086 Å, much longer than those in pure anatase TiO2 (1.930 and 1.973 Å). This indicates that the presence of an interstitial T* boron site results in an increase in the structural parameters, in good agreement with XRD results (Figure 1b vs 1a), Note that the three B-O bonds centering the tricoordinated T* site have a length of 1.370-1.410 Å. As for the tetrahedral-coordinated Q* site (Figure 6b), the interstitial B atom has four neighboring oxygen atoms, and the B-O bond lengths are in the range of 1.482-1.494 Å. In this case, the two Ti-O bonds adjacent to the interstitial Q* are also found to be stretched to 2.022 and 1.984 Å (Figure 6b). Thus, it is conclusive that borons which are present at interstitial T* and/or Q* sites are tightly coordinated with the neighboring oxygen atoms. Consequently, these T* and Q* borons tend to exist in the form of B3þ to exhibit a similar chemical environment with the normal Ti-O-B. Structural parameters obtained from these optimized geometries therefore enable us to predict the QCCs of 11B resonances corresponding to the T* and the Q* sites by means of the GIAPW method in conjunction with the periodic model. Such a combined technique has been shown to obtain reliable chemical shift and QCC values for various organic and inorganic crystals.34,54,55 Accordingly, the 11B QCCs predicted for interstitial T* and Q* were 3.3 and 1.7 MHz with η values of 0.5 and 0.06, respectively. The predicted QCC values are in reasonable agreement with the experimental results (2.45 and 1.3 MHz for T* and Q* sites, respectively; see Table 2). The discrepancies (ca. 0.8 MHz) probably arise from the effects of the transition metal. Thus, it is conclusive that interstitial B (T* and Q*) sites are indeed present in the B-doped TiO2. In addition, we also optimized the structure of substitutional B and predicted its 11B QCC. The calculated QCC (5.2 MHz) is much larger than the experimentally measured QCCs (1.0-2.5 MHz), implying the absence of substitutional B in our B-doped sample. The band gap and the density of state (DOS) of pure (undoped) and B-doped TiO2 samples were also calculated by 2715



ARTICLE

Figure 8. UV-vis absorption spectra of undoped, B-doped, N-doped, and (B, N)-codoped TiO2.

Figure 7. (a) Total density of states (DOSs) for TiO2 with undoped and doped interstitial (T* and Q*) borons, and (b) the projected DOSs into the doped boron sites, calculated by pure GGA/PBE. Inter. Q* stands for pseudotetrahedral coordinated (Q*) interstitial borons. Inter. T* stands for tricoordinated (T*) interstitial borons.

Castep package. Accordingly, the band gap calculated for pure anatase TiO2 was 2.2 eV, consistent with previous theoretical studies (1.95-2.2 eV).56,57 However, compared with the experimental value of 3.2 eV, our calculated band gap is underestimated by ca. 30%. It must be noted that DFT is in principle an exact theory to reproduce and predict ground-state properties (e.g., the total energy, the atomic structure, etc.). However, DFT is not a theory to address excited state properties, such as the band plot of a solid that represents the excitation energies of electrons injected or removed from the system. Thus, DFT seems to systematically underestimate the band gap by about 30-40% in insulators and semiconductors. However, it is still a widely accepted method to depict the defect states in the electronic structure calculations, giving qualitative explanation for the experimental results.2,58,59 The DOSs of undoped, tricoordinated interstitial B (T*) and pseudotetrahedral coordinated interstitial B (Q*) doped TiO2 photocatalysts were shown in Figure 7. It can be seen that the

valence band (VB) and conduction band (CB) mainly consist of O 2p and Ti 3d states.60 When the interstitial Q* was doped into the TiO2 supercell, the band gap between VB and CB is narrowed as shown in Figure 7a. It can be seen from the projected DOS (Figure 7b) that the narrowing of the band gap of interstitial Q* doped TiO2 originates from the appearance of new states (around 0 eV, due to B 2s and B 2p orbital) introduced by the interstitial Q*, which may extend the adsorption edge to visiblelight regions. Upon doping interstitial T* into the TiO2 supercell, no such new states are present in the gap, and the impurity states are located below the VB minimum (Figure 7a). In addition to the impurity states below VB, another two states induced by interstitial T* overlapped completely with the intrinsic VB and CB of TiO2, respectively, indicating that the interstitial T* doping is not in favor of the visible-light absorption. Some researchers11,29 reported that the interstitial B tended to result in a blue-shift of the absorption spectra. We deduce that this is probably due to the sole existence of the interstitial T*. It is noteworthy that there is no Ti3þ state in the calculated band gap of both interstitial T* and interstitial Q* doped TiO2 photocatalysts. However, it is well-known that the presence of the interstitial B normally leads to formation of Ti3þ,30,51-53 which provokes formation of a new 3d energy level at ca. 0.7-1.5 eV below the CB.61,62 It is well-known that the limitation of standard DFT functional (PBE) tends to provide a largely delocalized description of the electrons around Ti3þ, which may result in the absence of the Ti3þ state in the DOSs.30,63 3.6. Visible-Light Absorption and Photocatalytic Activity. Figure 8 displays the UV-vis absorption spectra of the undoped and doped TiO2 samples. Compared to pure TiO2, the absorption curve observed for the solely B-doped TiO2 extended only slightly toward the visible-light region. This is probably because of the low content of Q* (B/Ti = 0.6%) in B-doped TiO2 sample. By comparison, the N-doped TiO2 revealed a notable red-shift relative to the B-doped sample. Upon codoping B with N, the absorption curve of the (B, N)-codoped TiO2 exhibited a significant red-shift compared to the solely B- or N-doped TiO2 samples. Accordingly, the optical features observed for the (B, N)-codoped TiO2 show that such a photocatalyst may be activated by the visible light. The photochemical activities of various samples were further evaluated by the degradation of rhodamine B (Rh B) under the irradiation of visible light (>420 nm), as shown in Figure 9. The 2716



ARTICLE

Scheme 2. Proposed Photochemical Mechanism for (B, N)Codoped TiO2 under Visible-Light Irradiation

Figure 9. Kinetic photodegradation curves for rhodamine B on undoped, B-doped, N-doped, and (B, N)-codoped TiO2 upon irradiation with visible light (λ > 420 nm). The concentration of rhodamine B was determined by monitoring the variations of the optical intensity at λ = 553 nm.

photochemical activities observed for various samples lead to a trend rather similar to that deduced from their corresponding visible-light absorptions. For examples, the photochemical activity of the B-doped TiO2 was found to plateau at a conversion efficiency of ca. 25% after 240 min irradiation, slightly better than that of the pure TiO2 (ca. 15%) but lower than that of the N-doped TiO2 (ca. 50%) by nearly two fold. It has been proved that the presence of either interstitial B or N atoms tends to promote formation of Ti3þ that may accelerate the recombination of photoinduced holes and electrons,30,64 which eventually results in respective limits to their photochemical activities compared to the pure TiO2 sample. Therefore, although the anion doping may improve the visible-light absorption significantly, it does not always result in visible-light activity. It is noteworthy that a much higher photochemical activity, corresponding to a conversion efficiency of ca. 85% after 240 min of visible-light irradiation, was observed for the (B, N)-codoped TiO2 compared to the other pure and solely Bor N-doped samples. This is consistent with previous studies,8,12 but in contrast with that reported in ref 10. Most likely, a different sample preparation procedure causes this difference. It is probably not surprising that the (B, N)-codoped TiO2 photocatalyst, which exhibits the highest visible-light absorption, also possesses the highest concentrations of surface hydroxyl groups and interstitial Q* and N species. It had been shown65 that the presence of surface hydroxyl groups is advantageous to photocatalytic activity because of their ability to mediate oxidative electron transfer. However, a high visible-light absorption and a high surface hydroxyl content do not always lead to a high photocatalytic activity in the visible-light region. The key to the enhancement of photocatalytic activity lies in how to prolong the life of the photoinduced charge-carriers (pairs of e- and hþ) in the photocatalysts. Although it is generally accepted that the codoping of boron and nitrogen may improve the photocatalytic activity in the visible-light region,8,10 the detailed mechanism is not well understood. As revealed by the aforediscussed XPS, FT-IR, NMR, and theoretical calculation results, two types of interstitial B species, such as the T*and Q* units, are present in the (B, N)-codoped TiO2 photocatalyst. Among them, the presence of the Q* unit may be favorable for the absorption of visible light. The N 1s XPS spectrum obtained for the (B, N)-codoped TiO2 (Figure 3d) also

revealed the presence of two types of interstitial N species (TiN-O) with valence states of -1 and þ1, with the interstitial B (B3þ-O 3 3 3 Ti3þ) in their spatial proximity (vide supra). Therefore, it is anticipated that a synergy effect between the interstitial B and N species may be responsible for the high photocatalytic activity observed for the (B, N)-codoped TiO2 photocatalyst. Accordingly, we propose a reaction mechanism for the photocatalytic oxidation of (B, N)-codoped TiO2, as shown in Scheme 2. As revealed by the Ti 2p XPS results, codoping of B and N in TiO2 tends to hinder the formation of Ti3þ in the photocatalyst (Figure 3b). Thus, the Ti3þ that is generated due to the presence of interstitial B may be partly reoxidized to Ti4þ by the adjacent N as illustrated by Process I in Scheme 2. Consequently, such a process hinders the recombination of the photoinduced charge-carriers on the Ti3þ site. Similar phenomena, in which the interstitial N tends to promote reoxidation of Ti3þ (associated with oxygen vacancy) to Ti4þ (N 3 þ Ti3þ T N- þ Ti4þ),66,67 had been clearly observed in the N-doped TiO2. After the irradiation, the color center (N-) should produce a photoinduced electron, while the interstitial Q* (i.e., B3þ in Scheme 2) should act as shallow traps for electrons to prolong the life of the photoinduced charge-carriers (pairs of e- and hþ), as illustrated by Process II in Scheme 2. This process represents the key to improve the visible-light catalytic activity of the (B, N)-codoped TiO2. Finally, the pairs of e- and hþ may diffuse to the surface of the photocatalyst and continue to react with the hydroxyl groups, adsorbed water, and adsorbed oxygen, most likely forming highly active oxidants such as •OH and O2-. Meanwhile, the surface boron species may also introduce Brønsted and Lewis acid centers on the surfaces of the (B, N)-codoped TiO2. All these factors enable the (B, N)-codoped TiO2 to be a promising and efficient photocatalyst that is applicable under a wide range of light sources.

4. CONCLUSIONS We have demonstrated that by means of theoretical calculations in conjunction with a variety of different spectroscopic techniques, viz., SSNMR, FT-IR, and XPS, the detailed local structures of the dopants in B-doped, N-doped, and (B, N)codoped anatase TiO2 photocatalysts may be inferred. Two distinct types of interstitial boron species, namely pseudotetrahedral-coordinated Q* and tricoordinated T*, were identified in the B-doped and (B, N)-codoped TiO2. With aid from theoretical calculations, the detailed structures of the interstitial B (T* and Q*) and their corresponding QCC and η values comparable 2717


The Journal of Physical Chemistry C to the experimental results may be predicted. Consequently, it was found that the presence of the Q* unit tends to promote visible-light absorption. Apart from the interstitial B species, two types of bulk BO3/2 polymer as well as three types of surface B were also identified in the B-doped and the (B, N)-codoped TiO2 photocatalysts. The photochemical activities of various samples were also evaluated by the degradation of rhodamine B under the irradiation of visible light. Accordingly, it was found that (B, N)codoped TiO2 tends to exhibit a much higher visible-light photocatalytic activity than that of the solely B- or N-doped TiO2. Accordingly, a photochemical mechanism of the (B, N)-codoped TiO2 under visible-light irradiation is proposed. It is concluded that the synergistic effect of the interstitial B and N species is the key in obtaining the high photocatalytic activity.

’ ASSOCIATED CONTENT

bS

Supporting Information. Assorted O 1s XPS spectra of the pure and doped TiO2 samples. This material is available free of charge via the Internet at http://pubs.acs.org.

’ AUTHOR INFORMATION Corresponding Author

*E-mail: [email protected]. Fax: þ86-27-87199291.

’ ACKNOWLEDGMENT This work was supported by the National Natural Science Foundation of China (21073228, 20933009, and 20773159) and partly by the National Science Council in Taiwan (NSC 98-2113M-001-017-MY3). The authors are grateful to the National Center for High-performance Computing (NCHC, Taiwan) and Shanghai Supercomputer Center (SSC, China) for their support in computing facilities. ’ REFERENCES (1) Khan, S. U. M.; Al-Shahry, M.; Ingler, W. B., Jr. Science 2002, 297, 2243. (2) Di Valentin, C.; Pacchioni, G.; Selloni, A.; Livraghi, S.; Giamello, E. J. Phys. Chem. B 2005, 109, 11414. (3) Yu, J. C.; Zhang, L.; Zheng, Z.; Zhao, J. Chem. Mater. 2003, 15, 2280. (4) Yu, J. C.; Yu, J.; Ho, W.; Jiang, Z.; Zhang, L. Chem. Mater. 2002, 14, 3808. (5) Nukumizu, K.; Nunoshige, J.; Takata, T.; Kondo, J. N.; Hara, M.; Kobayashi, H.; Domen, K. Chem. Lett. 2003, 32, 196. (6) Zhao, W.; Ma, W.; Chen, C.; Zhao, J.; Shuai, Z. J. Am. Chem. Soc. 2004, 126, 4782. (7) Lu, N.; Quan, X.; Li, J. Y.; Chen, S.; Yu, H. T.; Chen, G. H. J. Phys. Chem. C 2007, 111, 11836. (8) Liu, G.; Zhao, Y.; Sun, C.; Li, F.; Lu, G.; Cheng, H. Angew. Chem., Int. Ed. 2008, 47, 4516. (9) Gopal, N. O.; Lo, H. H.; Ke, S. C. J. Am. Chem. Soc. 2008, 130, 2760. (10) In, S.; Orlov, A.; Berg, R.; Garca, F.; Jimenez, S. P.; Tikhov, M. S.; Wright, D. S.; Lambert, R. M. J. Am. Chem. Soc. 2007, 129, 13790. (11) Chen, D.; Yang, D.; Wang, Q.; Jiang, Z. Ind. Eng. Chem. Res. 2006, 45, 4110. (12) Zaleska, A.; Sobczak, J. W.; Grabowska, E.; Hupka, J. Appl. Catal. B 2008, 78, 92. (13) Zaleska, A.; Grabowska, E; Sobczak, J. W.; Gazda, M.; Hupka, J. J. Appl. Catal. B 2009, 89, 69. (14) de Ruiter, R.; Kentgens, A. P. M.; Grootendorst, J.; Jansen, J. C.; van Bekkum, H. Zeolites 1993, 13, 128.

ARTICLE

(15) Coudurier, G.; Auroux, A.; Vedrine, J. C.; Farlee, R. D.; Abrams, L.; Shannon, R. D. J. Catal. 1987, 108, 1. (16) Bray, P. J. Inorg. Chim. Acta 1999, 289, 158. (17) Reyes-Garcia, E. A.; Sun, Y. P.; Raftery, D. J. Phys. Chem. C 2007, 111, 17146. (18) Martens, R. M.; M€uller-Warmuth, W. J. Non-Cryst. Solids 2000, 265, 167. (19) Medek, A.; Harwood, J. S.; Frydman, L. J. Am. Chem. Soc. 1995, 117, 12779. (20) Wang, S. H.; Xu, Z.; Baltisberger, J. H.; Bull, L. M.; Stebbins, J. F.; Pines, A. Solid State NMR 1997, 8, 1. (21) Smith, M. E.; van Eck, E. R. H. Prog. Nucl. Magn. Reson. Spectrosc. 1999, 34, 159. (22) Frydman, L.; Harwood, J. S. J. Am. Chem. Soc. 1995, 117, 5367. (23) Goldbourt, A.; Madhu, P. K.; Vega, S. Chem. Phys. Lett. 2000, 320, 448. (24) Amoureux, J. P.; Fernandez, C.; Steuernagel, S. J. Magn. Reson. 1996, 123, 116. (25) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Science 2001, 293, 269. (26) Umebayashi, T.; Yamaki, T.; Itoh, H.; Asai, K. Appl. Phys. Lett. 2002, 81, 454. (27) Umebayashi, T.; Yamaki, T.; Yamamoto, S.; Miyashita, A.; Tanaka, S. J. Appl. Phys. 2003, 93, 5156. (28) Yang, K.; Dai, Y.; Huang, B. J. Phys. Chem. C 2007, 111, 18985. (29) Yang, K.; Dai, Y.; Huang, B. Phys. Rev. B 2007, 76, 195201. (30) Finazzi, E.; Di Valentin, C.; Pacchioni, G. J. Phys. Chem. C 2009, 113, 220. (31) Madhu, P. K.; Goldbourt, A.; Frydman, L.; Vega, S. Chem. Phys. Lett. 1999, 307, 41. (32) Massiot, D.; Fayon, M.; Capron, I.; King, S.; Le Calve’, B.; Alonso, J.-O.; Durand, B.; Bujoli, Z.; Gan, Z.; Hoatson, G. Magn. Reson. Chem. 2002, 40, 70. (33) Perdew, J. P.; Burke, K.; Ernzerhof, M. Phys. Rev. Lett. 1996, 77, 3865. (34) Perdew, J. P.; Ruzsinszky, A.; Csonka, G. I.; Vydrov, O. A.; Scuseria, G. E.; Constantin, L.; Zhou, X.; Burke, K. Phys. Rev. Lett. 2008, 100, 136406. (35) Monkhorst, H.; Pack, J. D. Phys. Rev. B 1976, 13, 5188. (36) Pickard, C. J.; Mauri, F. Phys. Rev. B 2001, 63, 245101. (37) Profeta, M.; Mauri, F.; Pickard, C. J. J. Am. Chem. Soc. 2003, 125, 541. (38) (a) Zheng, A.; Liu, S.-B.; Deng, F. J. Phys. Chem. C 2009, 113, 15018. (b) Zheng, A.; Liu, S. B.; Deng, F. J. Comput. Chem. 2009, 30, 222. (39) Khana, R.; Kima, S. W.; Kima, T. J.; Namb, C. M. Mater. Chem. Phys. 2008, 112, 167. (40) Spadavecchia, F.; Cappelletti, G.; Ardizzone, S.; Bianchi, C. L.; Cappelli, S.; Oliva, C.; Scardi, P.; Leoni, M.; Fermo, P. Appl. Catal. B: Environ. 2010, 96, 314. (41) Yu, J.; Zhao, X.; Zhao, Q. J. Chin. Ceram. Soc. 2000, 28, 245. (42) Sagawa, T.; Sueyoshi, R.; Kawaguchi, M.; Kudo, M.; Ihara, H.; Ohkubo, K. Chem. Commun. 2004, 7, 814. (43) Boulos, E. N.; Kreidl, N. J. J. Am. Ceram. Soc. 1971, 54, 368. (44) Brik, Y.; Kacimi, M.; Bozon-Verduraz, F.; Ziyad, M. J. Catal. 2002, 211, 470. (45) Bautista, F. M.; Campelo, J. M.; Garcia, A.; Luna, D.; Marinas, J. M.; Moreno, M. C.; Romero, A. A.; Navio, J. A.; Macias, M. J. Catal. 1998, 173, 333. (46) Ling, Q.; Sun, J.; Zhou, Q.; Ren, H.; Zhao, Q. Appl. Surf. Sci. 2008, 254, 6731. (47) Sathish, M.; Viswanathan, B.; Viswanath, R. P.; Gopinath, C. S. Chem. Mater. 2005, 17, 6349. (48) Engelhardt, G.; Kentgens, A. P. M.; Koller, H.; Samoson, A. Solid State Nucl. Magn. Reson. 1999, 15, 171. (49) Kroeker, S.; Stebbins, J. F. Inorg. Chem. 2001, 40, 6239. (50) Hwang, S. J.; Chen, C. Y.; Zones, S. I. J. Phys. Chem. B 2004, 108, 18535. (51) Takeda, S.; Suzuki, S.; Odaka, H.; Hosono, H. Thin Solid Films 2001, 392, 338. 2718



ARTICLE

(52) Ikeda, S.; Sugiyama, N.; Murakami, S.; Kominami, H.; Kera, Y.; Noguchi, H.; Uosaki, K.; Torimoto, T.; Ohtani, B. Phys. Chem. Chem. Phys. 2003, 5, 778. (53) Wang, J.; Tafen, D. N.; Lewis, J. P.; Hong, Z. l.; Manivannan, A.; Zhi, M. J.; Li, M.; Wu, N. Q. J. Am. Chem. Soc. 2009, 131, 12290. (54) Yates, J. R.; Dobbins, S. E.; Pickard, C. J.; Mauri, F.; Ghi, P. Y.; Harris, R. K. Phys. Chem. Chem. Phys. 2005, 7, 1402. (55) Zheng, A.; Liu, S. B.; Deng, F. J. Comput. Chem. 2009, 30, 222. (56) Yang, K.; Dai, Y.; Huang, B. J. Phys. Chem. C 2007, 111, 12086. (57) Mi, L.; Zhang, Y.; Wang, P. Chem. Phys. Lett. 2008, 458, 341. (58) Lee, J.; Park, J.; Cho, J. Appl. Phys. Lett. 2005, 87, 011904. (59) Yang, K.; Dai, Y.; Huang, B. J. Phys. Chem. C 2007, 111, 18985. (60) Asahi, R.; Taga, Y.; Mannstadt, W.; Freeman, A. Phys. Rev. B 2000, 61, 7459. (61) Grey, I. E.; Li, C.; MacRae, C. M.; Bursill, A. J. Solid State Chem. 1996, 127, 240. (62) Henderson, M. A.; Epling, W. S.; Peden, C. H. F.; Perkins, C. L. J. Phys. Chem. B 2003, 107, 534. (63) Di Valentin, C.; Pacchioni, G.; Selloni, A. Phys. Rev. Lett. 2006, 97, 166803. (64) Kurtz, R. L.; Stock-Bauer, R.; Madey, T. E. Surf. Sci. 1989, 218, 178. (65) Hoffmann, M. R.; Martin, S. T.; Choi, W.; Bahnemannt, D. W. Chem. Rev. 1995, 95, 69. (66) Ihara, T.; Miyoshi, M.; Iriyama, Y.; Matsumoto, O.; Sugihara, S. Appl. Catal. B 2003, 42, 403. (67) Livraghi, S.; Paganini, M. C.; Giamello, E.; Selloni, A.; Di Valentin, C.; Pacchioni, G. J. Am. Chem. Soc. 2006, 128, 15666.

2719


Boron Environments in B-Doped and (B, N)-Codoped TiO2

Recommend Documents