Shedding Light on Aging of N-Doped Titania ... - ACS Publications

Jul 17, 2015 - Faculty of Materials Science, Lomonosov Moscow State University, Lenin. Hills, Moscow 119991, Russia. ⊥. Kurnakov Institute of Genera...
0 downloads 0 Views 4MB Size
Article pubs.acs.org/JPCC

Shedding Light on Aging of N‑Doped Titania Photocatalysts Alexey Tarasov,*,†,§ Anton Minnekhanov,‡ German Trusov,§ Elizaveta Konstantinova,‡ Alexandr Zyubin,† Tatiana Zyubina,† Alexey Sadovnikov,⊥ Yury Dobrovolsky,† and Eugene Goodilin*,§,∥,⊥

J. Phys. Chem. C 2015.119:18663-18670. Downloaded from pubs.acs.org by RENSSELAER POLYTECHNIC INST on 09/20/18. For personal use only.



Laboratory of Solid State Ionics, Institute of Problems of Chemical Physics, Russian Academy of Sciences, Academician Semenov avenue 1, Chernogolovka 142432, Russia ‡ Department of Physics, §Department of Chemistry, and ∥Faculty of Materials Science, Lomonosov Moscow State University, Lenin Hills, Moscow 119991, Russia ⊥ Kurnakov Institute of General and Inorganic Chemistry, Leninsky prospect 31, Moscow 119991, Russia S Supporting Information *

ABSTRACT: A detailed analysis of nitrogen dopant behavior in nanostructured microspheres of the TiO2 photocatalyst obtained by the thermally assisted reactions in aqueous sprays method has been performed for the first time using electron paramagnetic resonance, X-ray photoelectron spectroscopy, and UV−vis spectroscopy and is supported by theoretical simulation of possible defect structures. The nitrogen species were found to undergo the N• to N− transformation during sample storage under different conditions, with an activation energy of about 0.45 eV. Three main possible evolution pathways for the dopant state were identified and discussed. It was established that the most probable transformation consists of migration of an oxygen vacancy site to an interstitial nitrogen atom followed by the formation of a nonparamagnetic substitution nitrogen center. Possible diffusion routes of oxygen vacancy and corresponding energy barriers were estimated and found to be in agreement with experimental observations.



thermodynamic stability of titania phase modifications,16 and to determine the formation of nitrogen species in the titania lattice17 and the dopant’s influence on optical18 and functional properties of doped titania.19 In most studies, theoretical research focused on the changes of the semiconductor’s electron structure caused by doping,20 the role of dopant in photoinduced charge carrier recombination processes,21 and the influence of impurities on the stability of the lattice defects,22 but only a few of them were dedicated to explore storage stability and operation durability of such materials under different conditions, which is of vital importance for future practical applications. Spadavecchia23 has studied the aging (evolution of optical, paramagnetic, and photocatalytical properties over time) of Ndoped titania samples, prepared by a hydrothermal treatment

INTRODUCTION Nowadays, titanium dioxide attracts deep interest as a promising nanomaterial for photovoltaics, water photosplitting, and air purification.1−4 Unfortunately, widely expected practical applications of TiO2-based photocatalysts are restricted by the requirement of operation using artificial UV light sources. Sunlight is a natural inexpensive source of energy for photocatalysis, but only about 3% of the solar spectrum can be absorbed by pristine TiO2 due to its large band gap (e.g., 3.2 eV for anatase). Since Asahi5 has shown the N-doping of titania to be a facile approach to overcome this drawback, many strategies have been developed to obtain doped titania, including ion implantation, magnetron sputtering, plasma and laser treatment, mechanoactivation,6−10 hydrothermal and solvothermal treatments, sol−gel routes, TiN oxidation, and high-temperature sintering under an ammonia atmosphere.11−15 A number of studies were devoted to investigate the features of the doping procedure, such as the nitrogen influence on the © 2015 American Chemical Society

Received: March 22, 2015 Revised: June 17, 2015 Published: July 17, 2015 18663

DOI: 10.1021/acs.jpcc.5b02760 J. Phys. Chem. C 2015, 119, 18663−18670

Article

The Journal of Physical Chemistry C

Figure 1. (a) Typical EPR spectrum of TON samples in the dark at room temperature. Long-term measurements of storage sample TON1-S. (b) N• peak intensity from EPR spectra recorded in the dark. (c) UV−vis diffusion reflectance spectra in comparison with the spectrum of pristine TiO2.



EXPERIMENTAL SECTION Nitrogen-doped TiO2 powders with an average microsphere diameter of about 0.5−3 μm were synthesized by the method of thermally assisted reactions in aqueous sprays (TARAS) using urea as a nitrogen dopant precursor26 (see Supporting Information). The studied samples consist of anatase mixed with rutile in a ratio of 60/40%. In this work, we examined the influence of storage of the sample obtained almost 2 years ago at 1% of urea concentration in the initial hydrolyzing solution annealed at 1000 °C (denoted TON1-S) and two as-prepared samples, obtained at different concentrations of urea in the initial hydrolyzing solution: 0.5% (denoted TON05) and 1% (denoted TON1) with the same calcination temperature. To explore the temperature impact on storage stability, a portion of TON1 was stored in the dark at 80 °C for a total of 5 weeks (denoted as TiON1-T). All studied samples were stored and treated in powder form. EPR spectra were recorded at room temperature (unless indicated otherwise) with a standard Bruker EPR spectrometer, ELEXSYS-500 (X-band, sensitivity of about 5 × 1010 spin/G, modulation frequency = 100 kHz). Computer fitting of the spectra was obtained using the Simfonia program. To evaluate the g values of the EPR signals and spin concentration in the samples, a MgO matrix standard with Mn2+ ions and a CuCl2· H2O standard with known spin concentration were used. The

using different nitrogen precursors (NH3, triethylamine, and urea). Faster aging observed for the samples prepared using NH3 was explained by prevailing surface localization of the defects and their consequent lower stability upon the contact with O2/humidity, while organic precursors were assumed to be less volatile, therefore forming more “permanent” defective sites during the synthesis. The high-temperature calcination of nitrogen-doped titania was investigated using radioactive nuclear probes by Mihara.24 He reported relatively higher stability of substituted N impurities in comparison with interstitial nitrogen atoms. D’Arienzo25 concluded that a high-temperature treatment causes transformation of the interstitial N into a substitution form with no impact on its photocatalytic properties. In this paper, we report for the first time the results of complex analysis of nitrogen dopant behavior in TiO 2 photocatalysts during artificial aging for different times under ambient temperatures. The study of the final states of the material performed using electron paramagnetic resonance (EPR), X-ray photoelectron spectroscopy (XPS), and UV−vis techniques is certainly in good agreement with theoretical density functional theory (DFT) simulation of possible defect structure variants. Finally, the obtained results shed light on the postsynthetic stability of N-TiO2 materials and the processes proceeding in the route of its storage. 18664

DOI: 10.1021/acs.jpcc.5b02760 J. Phys. Chem. C 2015, 119, 18663−18670

Article

The Journal of Physical Chemistry C

Figure 2. Short-term measurements of TON05 and TON1 samples: (a) N• peak intensity extracted from EPR spectra recorded in the dark; (b) UV−vis diffusion reflectance spectra in comparison to the spectrum of pristine TiO2.

comparison to that of the initial species. The values of corresponding N• signals are shown in Figure 1b, and an exponential fit gives an attenuation coefficient R0(20 °C) = −0.008 day−1. An 8-fold decrease in N• signal intensity, at the first glance, can be attributed to the exhausting of N species when leaking out from the sample. This would mean a dramatic impairment of the functional properties of doped titania under visible illuminations because the N species in the TiO2 lattice are believed to be responsible for visible light harvesting.22 UV−vis diffusion reflectance spectra of TON1-S (Figure 1c), recorded initially after preparation and after 21 months of synthesis, demonstrate an absorbance shoulder specific for Ndoped titania in the visible region of 420−550 nm, contrary to pure titania. The absorption in this visible region is usually attributed to electron excitation in a conductive band from an extrinsic energy state of nitrogen in the titania lattice (N2p → Tid),21 and its relative intensity correlates with the concentration of lattice defects. Surprisingly, the intensity of the visible absorption, measured after the storage, does not undergo significant changes compared to the EPR data. In view of the fact that different nitrogen impurities are presumed to contribute to visible light absorption, the observed decrease of N• concentration is likely to be caused by some internal processes, associated with a transformation of paramagnetic N• species into nonparamagnetic forms. For a more careful examination of the observed effect, we have repeated storage experiments for a shorter period of time to corroborate the UV−vis DRS and EPR data with the XPS technique. Short-Term Storage Samples. The N• peak intensities of TON05 and TON1 samples do not show significant impairment during the 35 days of storage at ambient temperature (Figure 2a). The high intensity of the first peaks of both samples (2387 and 1974 au), as well as deviations around the average value (375 and 950 au) observed during the storage, was explained by illumination of the sample by visible light during the preparation and mounting of the probes. The results of EPR measurements, done after intentional illumination of EPR probes by a halogen lamp before EPR measurements (N• peak intensities are 2421 and 2489 au, shown in Figure 2a as red points), fully support this statement.

studied samples were placed in different capillary tubes, maintained in a cylindrical cavity for the measurements in the dark, or in an optical cavity for the measurements under illumination with a 100 W halogen lamp in the spectral range of Δλ = 270−900 nm (1.38−4.6 eV). UV−vis absorption spectra were obtained by means of a Lambda 35 UV/vis spectrometer (PerkinElmer, USA) operated in diffusion reflectance mode in the range of 190−1100 nm. High-resolution X-ray photoelectron spectra (XPS) were recorded using a Kratos AXIS Ultra DLD spectrometer with an incorporated Al Kα (hν = 1486.6 eV) 120 W X-ray source on sample area of 300 × 700 μm2. The binding energy was calibrated by referencing the C 1s peak to 285 eV. Due to the low conductivity of the samples, low-energy electrons moving in a spiral path in a magnetic field were used as a positive charge neutralizer. Before XPS measurements, the samples were kept in a vacuum in order to clean the surface from the physically adsorbed species. DFT calculations of N atom migration in the anatase crystal lattice of TiO2 were performed with periodical boundary conditions and plane-wave expansions using the generalized gradient approximation in the Perdew−Burke−Ernzerhof form (GGA-PBE). The Vienna ab initio simulation package was the base of all the calculations (VASP).27−32 Within the VASP modeling, a projector-augmented plane-wave (PAW) basis set was used with the corresponding pseudopotential and the PBE functional. The plane-wave cutoff energy was set to 400 eV. For simulation of nitrogen and vacancy migration in bulk anatase, (TiO2)32 and (TiO2)72 supercells were used. For the evaluation of the transition states, the nudged elastic band method implemented in VASP was applied.



RESULTS AND DISCUSSION Long-Term Aging under Ambient Conditions. A typical EPR spectrum of the N-doped titania sample is shown in Figure 1a. EPR lines detected in the samples can be ascribed to the O2− and N• paramagnetic centers.33 In order to separate a contribution from O 2• radicals, N • concentration was monitored as peak-to-peak intensity of the right-sided satellite peak, as shown in the inset. The EPR spectra of the TON1-S storage sample, recorded after 8, 12, and 21 months of synthesis, demonstrate strong depletion of the N• signal in 18665

DOI: 10.1021/acs.jpcc.5b02760 J. Phys. Chem. C 2015, 119, 18663−18670

Article

The Journal of Physical Chemistry C

Figure 3. Short-term measurements of TON1-T, stored at 80 °C: (a) N• peak intensity from EPR spectra recorded in the dark; (b) UV−vis diffusion reflectance spectra in comparison with spectra of the TON05 sample and pristine TiO2.

The visible absorption of TON05 and TON1 samples, as expected, demonstrates no changes, similarly to the stored TON1-S sample (Figure 2b). Thus, the obtained results demonstrate a strong correlation between loading of the initial nitrogen precursor (0.5 and 1 wt %), the N• peak intensity (375 and 950 au), and visible absorption (almost 2-fold for TON1 than for TON05), but no signal impairment during short-term aging at ambient temperature was found. A different situation is observed in the case of short-term aging at slightly increased temperature (80 °C). In spite of the almost permanent shoulder of visible absorption, the N• peak intensity of the TON1-T sample, shown in Figure 3a as solid black squares, has a clear trend to decrease during 35 days of experiment (Figure 3), even below the initial value of the TON05 sample. The first elevated point at 1850 au is assumed to have the artifact nature caused by sample preparation (the intentional illumination of the TON1-T sample leads to an intensity increase to 2235 au, as shown with red circles). For fitting purposes, the first point was substituted by the average intensity of the N• peak with respect to different mass of the sample (1045 au, shown as a red star). Hence, the N• peak intensity with a corrected first point demonstrates almost 5-fold impairment of the signal throughout 35 days (the attenuation coefficient R0(80 °C) = −0.1 day−1 for exponential fitting). The temperature dependence of attenuation coefficients, observed for the TON1-S and TON1-T samples (−0.008 day−1 at 20 °C and −0.1 day−1 at 80 °C), allows us to assume that the possible processes for the observed EPR signal impairment have a remarkable activation energy. Assuming the Arrhenius nature of this process, we estimated that the activation energy Ea of the N• transformation process proposed above was Ea = 0.45 eV. The XPS technique is widely used to detect the amount and the form of nitrogen in the titania lattice.34 However, an assignment of XPS peaks to certain N forms is still under debate; N 1s core level peaks at 396−397 eV are usually ascribed to substitutional nitrogen, while the peaks at ∼400 eV are attributed to nitrogen in the interstitial site.22 The XPS spectra of TON1 and TON1-T samples, shown at Figure 4, demonstrate a single peak centered at a binding energy close to the one ascribed for a nitrogen interstitial site. The peak position slightly shifts after thermal treatment to lower energy values, while the total nitrogen concentration,

Figure 4. N 1s core level XPS spectra of TON1 and TON1-T samples.

calculated from XPS data, remains unchanged with respect to the overall sample weight. Thus, EPR, UV−vis DRS, and XPS data support the suggestion that the depletion of the N• signal has to be associated with the transformation into another nitrogen form rather than exhaustion out of the titania lattice. Mechanism of the N• Signal Depletion. To explain the • N annihilation phenomena, we suggest three possible routes as shown below: N2 formation via Ni• radical recombination (1), N2 formation via recombination of Ni• with Ns− (2), and Ni• reduction into Ni− (3). Ni• + Ni• → N2(i) •



(1)

Ni + Ns → N2(s)

(2)

Ni• + VO → Ns−

(3)

All the routes should be limited by at least two factors: the diffusion of corresponding species in the titania lattice and the abundance of the species in the lattice. In order to determine the most plausible route of N• degradation, a relative abundance for N(i)•, N(s)−, and VO species and corresponding migration energies were estimated using the EPR technique supported by DFT-GGA simulation. 18666

DOI: 10.1021/acs.jpcc.5b02760 J. Phys. Chem. C 2015, 119, 18663−18670

Article

The Journal of Physical Chemistry C At first, we compared processes (1) and (2). While substitutional nitrogen atoms Ns− are immobilized in the titania lattice and only interstitial Ni• diffusion can drive the processes, the probability of both the processes is defined by Ni• and Ns− relative concentrations. In accordance with the XPS measurements (Figure 4a), the concentration of substitutional nitrogen Ns− in as-prepared samples is negligibly low. Therefore, the formation of interstitial N2, similar to the samples described elsewhere,35 seems to be more feasible in the first suggested route. The competition of (1) and (3) processes depends on two parameters: relative concentration and diffusion abilities of Ni• and VO species in the lattice. Since the formation of oxygen vacancies is coupled with the formation of a paramagnetic Ti3+ center, the relative concentration of Ni• and VO species in the sample can be estimated from the intensity of their EPR signals. The EPR spectrum of the TON1 sample, measured at 115 K (to compensate short relaxation time of Ti3+ centers), is shown in Figure 5. As can be seen, the new large EPR signal is

on N and O atoms have approximately values the same as those for the system with no vacancy (deviations less than 0.1 eV). After electron transition from the vacancy to NO, the number of unpaired electrons changes from 3 to 1; electrostatic potential values on N and O atoms are shifted to 1.1 and 0.9 eV, accordingly, and the total energy of the system is lowered to 1−2 eV, depending on the VO−N(i) distance (Figure 6). However, the total energy of the system with substitutional nitrogen (Figure 6, TiO2_Ns) is ∼2 eV lower that any state with an electron transfer.

Figure 6. Changes of system energy caused by electron transfer from an oxygen vacancy to the interstitial nitrogen atom at different VO−Ni positional relationships, compared to the system with substitution of nitrogen. The VO−Ni distances are (a) 7.35, (b) 3.07, (c) 2.54, and (d) 2.45 Å.

Figure 5. EPR measurements of TON1 at 115 K.

detected with g1 = 1.996 and g2 = 1.925. According to the literature data, we can ascribe this EPR signal to the Ti3+/ oxygen vacancy paramagnetic centers.33 The concentrations of N• and Ti3+ centers, calculated by double integration of the signal, are in the ratio of 1:125. Thereby, the third route of Ni• to Ns− transformation, caused by the interaction with oxygen vacancy, is most likely to be responsible for the observed EPR signal depletion. The N• to N− transformation is described in many studies as an inherent part of the doped titania formation process36 or as a result of high-temperature treatment.25 However, to the best of our knowledge, we report for the first time a continuation of this process in the prepared N-doped titania during its longterm storage. DFT Simulations of the N• to N− Transformation. The interaction of N• and VO should include the approach by migration of at least one component, electron transfer, and proper lattice rearrangement. In order to elucidate the features of these processes, they were simulated using the DFT-GGA approach. At first, we estimated the possibility of electron transfer from a single oxygen vacancy to a nitrogen atom. Without electron transfer, the NO fragment has one unpaired electron and a vacancy has two electrons, so the system as a whole has three unpaired electrons (m = 3). In this case, electrostatic potentials

These results give us two important conclusions: (i) the electron transfer should be taken into account during calculations of species migration and (ii) an approach of N• and VO leads to nitrogen trapping by the oxygen vacancy with stable N(s) formation. The diffusion of nitrogen in a perfect anatase lattice was calculated by Xiao37 using the DFT-GGA approach. The migration barriers for three identified unequal paths were identified as EA = 1.43 eV, EB = 1.94 eV, and EC = 1.83 eV. In our calculations, the influence of the electron transfer from oxygen vacancy to interstitial nitrogen established above was found to have a significant impact on nitrogen migration barriers. The three diffusion paths and corresponding energy barriers for a N− species migration are shown in Figure 7 The migration barriers for d1−d2 and d2−d3 are close to those reported by Xiao, while the d2−d4 is significantly higher (2.1 eV) than that calculated without a vacancy influence (1.43 eV). The recalculated without vacancy d2−d4 migration barrier was found to be 1.42 eV. The reason for this difference becomes clear when considering the actual forms of nitrogen species during migration with and without an oxygen vacancy. In the absence of vacancy, nitrogen moves through the d1−d2 and d2−d3 paths with destruction of N−O bonds in the barrier 18667

DOI: 10.1021/acs.jpcc.5b02760 J. Phys. Chem. C 2015, 119, 18663−18670

Article

The Journal of Physical Chemistry C

trapping. The vacancy migration barriers in the rutile phase were calculated to be from 0.6 to 1.7 eV for different directions.38−40 In the present study, we calculated the migration barriers in anatase using a (TiO2)72 cluster. The TiO6 octahedron in anatase is distorted, giving C2v symmetry, thus it has four different vacancy positions (1, 2, 3, 4) with nearly similar relative energy (±0.1 eV), as shown in Figure 8a.

Figure 8. (a) Enlarged fragment of the supercell; the numbers in the picture represent different oxygen sites (marked in pink). (b) Oxygen diffusion channels (marked with pink circles and arrows) in the anatase lattice. The Ti atoms are the blue circles, the O atoms are the red ones. Figure 7. Different positions of the nitrogen atom in the titania lattice and corresponding transition state energies. The Ti atoms are the blue circles, the O atoms are red.

The migration barrier between the positions of 1 and 2 (0.18 eV) is significantly lower than for 1−3 and 1−4 (1.14 and 1.9 eV) paths (Figure 9). It can be seen that the oxygen atoms in the 1 and 2 positions form the TiO2Ti rhombohedra, connected in flat noncrossing ribbons (shown in Figure 8b), spread in two mutually perpendicular directions, (010) and

region (EB = 1.94 eV and EC = 1.83 eV), while it forms the O− N−O angled fragment that decreases its relative energy (EA = 1.43 eV) through the d2−d4 path. In the presence of an oxygen vacancy and if an electron transfer from the oxygen vacancy to this nitrogen atom occurs, the transferred electron gives an antibonding effect for the O−N bonds; therefore, nitrogen moves through the d2−d4 path, destructing the N−O bonds and increasing the migration barrier (2.1 eV). It should be mentioned that the lower migration barrier in anatase (d2−d4 path) corresponds to the movement along the diagonal of the TiO2Ti rhombohedron, with the shortest O−O distance in the TiO6 octahedron (2.5 Å compared to 2.8 and 3.0 Å). These rhombohedra form flat noncrossing ribbons in the lattice, and migration through the d2−d3 path causes only quasi-one-dimensional movement. Any transfer from one ribbon to another is associated with overcoming the other two barriers of about 2 eV. In the rutile form, the corresponding O−O distances are close to 2.6, 2.8, and 3.0 Å, and thus migration barriers are likely to be even higher than those in anatase. Moreover, the TiO2Ti rhombohedra in the rutile phase with the shortest O−O distance are not connected directly and do not form one-dimensional transfer ribbons with minor migration energy like in anatase. Thus, the nitrogen migration in anatase or rutile lattices through oxygen defects is associated with overcoming the barriers of about 2 eV and is much higher compared to the activation energy found for aged samples, which is as low as 0.45 eV in the experiments. Alternatively, the oxygen vacancy migration can also cause the approach of N• and an oxygen vacancy with subsequent N− species formation caused by electron transfer or nitrogen

Figure 9. Migration barriers between different possible vacancy positions in the anatase lattice. The Ti atoms are the blue circles, the O atoms are red. 18668

DOI: 10.1021/acs.jpcc.5b02760 J. Phys. Chem. C 2015, 119, 18663−18670

Article

The Journal of Physical Chemistry C

However, any storage at room temperature would guarantee preserving high enough activity of the nanostructured titania.

(100), shown in Figure 8b. In the anatase lattice, these ribbons can act as channels of high oxygen vacancy conductivity with a low activation energy (0.18 eV). To realize three-dimensional migration in such a lattice, the oxygen vacancy needs to overcome higher energy barriers (1.14 or 1.9 eV) and switch between the channels. However, the migration along ribbons is already sufficient to cause the observed phenomena of doped titania aging. It should be mentioned that, in the rutile phase, the oxygen positions correspond to a low-energy transition and do not form continuous ribbons because of different TiO6 octahedron orientation. The vacancy can only jump/swing between two neighboring positions but do not migrate through the rutile lattice. Thus, the above proposed mechanism of N• transformation can occur only in the anatase form of the samples. The migration barrier of oxygen diffusion, calculated by the DFT-GGA approach, does not differ drastically from the prediction of activation energy given from experiments of EPR signal depletion. Hence, the depletion of the N• EPR signal, observed during the aging of N-doped titania at elevated temperature, is likely caused by its interaction with an oxygen vacancy followed by paramagnetic substitutional N− formation. This conclusion corroborates the results of XPS measurements, where the position of nitrogen shifts after calcination toward the lower binding energy, which is assumed to be typical of substitutional species. Our findings are in good agreement with calculations done by Tsetseris41 using the same DFT-GGA approach with a smaller TiO2 supercell. The different rate of aging, observed by Spadavecchia23 on N-doped titania samples, obtained from different nitrogen precursors, can be explained by the different oxygen vacancy concentration in the lattice, defined by various reducing abilities of the nitrogen source. It is worth noticing that during sample aging we did not observe any significant changes in photocatalytic activity of the doped titania samples under either UV−visible or visible irradiations (see Supporting Information). This fact agrees well with the assumption that all forms of nitrogen in the titania lattice contribute to its photocatalytic performance and that particular transformation of N• into N− causes no impact on the functional properties of doped materials. Thus, it is arguable that the aging of nitrogen-doped titania at moderate temperature (up to 80 °C) does not impact its functional properties, in spite of the continuation of internal processes of defect interactions.



ASSOCIATED CONTENT

S Supporting Information *

Material morphology (SEM) and photocatalytic activity (crystal violet photodegradation). The Supporting Information is available free of charge on the ACS Publications website at DOI: 10.1021/acs.jpcc.5b02760.



AUTHOR INFORMATION

Corresponding Authors

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

The authors declare no competing financial interest.



ACKNOWLEDGMENTS This work was supported by RSF Grant No. 14-23-00218. The EPR measurements were performed using the facilities of the Collective Use Center at the Moscow State University and in frame of the program of development of M.V. Lomonosov Moscow State University.



REFERENCES

(1) O’Regan, B.; Grätzel, M. Low-cost, high-efficiency solar cell based on dye-sensitized colloidal TiO2 films. Nature 1991, 353, 737−740. (2) Selli, E.; Chiarello, G.-L.; Quartarone, E.; Mustarelli, P.; Rossetti, I.; Forni, L. A photocatalytic water splitting device for separate hydrogen and oxygen evolution. Chem. Commun. 2007, 5022−5024. (3) Fujishima, A.; Hashimoto, K.; Watanabe, T. TiO2 Photocatalysis: Fundamentals and Applications; BKC: Tokyo, 1999. (4) Ao, C. H.; Lee, S. C. Indoor air purification by photocatalyst TiO2 immobilized on an activated carbon filter installed in an air cleaner. Chem. Eng. Sci. 2005, 60 (1), 103−109. (5) Asahi, R.; Morikawa, T.; Ohwaki, T.; Aoki, K.; Taga, Y. Visiblelight photocatalysis in nitrogen-doped titanium oxides. Science 2001, 293 (5528), 269−271. (6) Ghicov, A.; Macak, J. M.; Tsuchiya, H.; Kunze, J.; Haeublein, V.; Frey, L.; Schmuki, P. Ion implantation and annealing for an efficient N-doping of TiO2 nanotubes. Nano Lett. 2006, 6, 1080−1082. (7) Mwabora, J. M.; Lindgren, T.; Avendano, E.; Jaramillo, T. F.; Lu, J.; Lindquist, S. E.; Granqvist, C. G. Structure, composition and morphology of photoelectrochemically active TiO2‑xNx thin films deposited by reactive magnetron dc sputtering. J. Phys. Chem. B 2004, 108, 20193−20198. (8) Hong, Y.; Bang, C.; Shin, D.; Uhm, H. Band gap narrowing of TiO2 by nitrogen doping in atmospheric microwave plasma. Chem. Phys. Lett. 2005, 413, 454. (9) Somekawa, S.; Kusumoto, Y.; Yang, H.; Abdulla-Al-Mamun, M.; Ahmmad, B. Fabrication, N-doping mechanism and evaluation of Ndoped TiO2 thin films based on laser ablation method. J. Sci. Res. 2010, 2 (1), 17−23. (10) Chen, S. F.; Chen, L.; Gao, S.; Cao, G. Y. The preparation of nitrogen-doped photocatalyst TiO2−XNX by ball milling. Chem. Phys. Lett. 2005, 413 (4−6), 404−409. (11) Kometani, N.; Fujita, A.; Yonezawa, Y. Synthesis of N-doped titanium oxide by hydrothermal treatment. J. Mater. Sci. 2008, 43 (7), 2492−2498. (12) Yang, G.; Jiang, Z.; Shi, H.; Xiao, T.; Yan, Z. Preparation of highly visible-light active N-doped TiO2 photocatalyst. J. Mater. Chem. 2010, 20, 5301−5309. (13) Nolan, N. T.; Synnott, D. W.; Seery, M. K.; Hinder, S. J.; Van Wassenhoven, A.; Pillai, S. C. Effect of N-doping on the photocatalytic activity of sol−gel TiO2. J. Hazard. Mater. 2012, 211−212, 88−94.



CONCLUSIONS The results of our complex postsynthesis study of N-doped TiO2 show that the storage of doped titania samples at moderate temperature is accompanied by interaction of nitrogen impurities with lattice defects. Full nitrogen concentration in the TiO2 lattice remains unchanged, although the concentration of interstitial paramagnetic N• centers in the lattice decreases with time. The transformation of N• species into a nonparamagnetic N− state was attributed to N• capture by oxygen vacancy, caused by oxygen vacancy diffusion in the anatase lattice. The (010) and (100) directions were determined as the most possible diffusion directions with migration barriers of ∼0.2 eV. The photocatalytic activity of the samples under visible irradiation does not undergo any significant changes during the aging. Thus, it is arguable that the storage of nitrogen-doped titania at moderate temperature (up to 80 °C) does not impact its functional properties, in spite of the continuation of internal processes of defect interactions. 18669

DOI: 10.1021/acs.jpcc.5b02760 J. Phys. Chem. C 2015, 119, 18663−18670

Article

The Journal of Physical Chemistry C (14) Saha, N. C.; Tompkins, H. G. Titanium nitride oxidation chemistry: An x-ray photoelectron spectroscopy study. J. Appl. Phys. 1992, 72, 3072−3079. (15) Kosowska, B.; Mozia, S.; Morawski, A. W.; Grzmil, B.; Janus, M.; Kałucki, K. The preparation of TiO2-nitrogen doped by calcination of TiO2•xH2O under ammonia atmosphere for visible light photocatalysis. Sol. Energy Mater. Sol. Cells 2005, 88 (3), 269−280. (16) Bu, X.; Zhang, G.; Zhang, C. Effect of nitrogen doping on anatase−rutile phase transformation of TiO2. Appl. Surf. Sci. 2012, 258 (20), 7997−8001. (17) Ceotto, M.; Lo Presti, L.; Cappelletti, G.; Meroni, D.; Spadavecchia, F.; Zecca, R.; Leoni, M.; Scardi, P.; Bianchi, C. L.; Ardizzone, S. About the nitrogen location in nanocrystalline N-doped TiO2: combined DFT and EXAFS approach. J. Phys. Chem. C 2012, 116, 1764−1771. (18) Lindgren, T.; Mwabora, J. M.; Avendaño, E.; Jonsson, J.; Hoel, A.; Granqvist, C.-G.; Lindquist, S.-E. Photoelectrochemical and optical properties of nitrogen doped titanium dioxide films prepared by reactive DC magnetron sputtering. J. Phys. Chem. B 2003, 107 (24), 5709−5716. (19) Ananpattarachai, J.; Kajitvichyanukul, P.; Seraphin, S. Visible light absorption ability and photocatalytic oxidation activity of various interstitial N-doped TiO2 prepared from different nitrogen dopants. J. Hazard. Mater. 2009, 168 (1), 253−261. (20) Di Valentin, C.; Pacchioni, G.; Selloni, A. Origin of the different photoactivity of N-doped anatase and rutile TiO2. Phys. Rev. B: Condens. Matter Mater. Phys. 2004, 70, 085116. (21) Barolo, G.; Livraghi, S.; Chiesa, M.; Paganini, M. C.; Giamello, E. Mechanism of the photoactivity under visible light of N-doped titanium dioxide. Charge carriers migration in irradiated N-TiO2 investigated by electron paramagnetic resonance. J. Phys. Chem. C 2012, 116, 20887−20894. (22) Di Valentin, C.; Pacchioni, G.; Selloni, A.; Livraghi, S.; Giamello, E. Characterization of paramagnetic species in N-doped TiO2 powders by EPR spectroscopy and DFT calculations. J. Phys. Chem. B 2005, 109 (23), 11414−11419. (23) Spadavecchia, F.; Ardizzone, S.; Cappelletti, G.; Oliva, C.; Cappelli, S. Time effects on the stability of the induced defects in TiO2 nanoparticles doped by different nitrogen sources. J. Nanopart. Res. 2012, 14, 1301. (24) Mihara, M.; Kumashiro, S.; Fujiwara, H.; Matsumiya, R.; Matsuta, K.; Nakashima, Y.; Zheng, Y. N.; Ogura, M.; Sumikama, T.; Nagatomo, T.; Minamisono, K.; Fukuda, M.; Izumikawa, T.; Minamisono, T. Microscopic observation of impurities in TiO2 using radioactive nuclear probes. Phys. B 2006, 376−377, 955−958. (25) D’Arienzo, M.; Scotti, R.; Wahba, L.; Battocchio, C.; Bemporad, E.; Nale, A.; Morazzoni, A. Hydrothermal N-doped TiO2: Explaining photocatalytic properties by electronicand magnetic identification of N active sites. Appl. Catal., B 2009, 93, 149−155. (26) Tarasov, A.; Trusov, G.; Minnekhanov, A.; Gil, D.; Konstantinova, E.; Goodilin, E.; Dobrovolsky, Yu. Facile preparation of nitrogen-doped nanostructured titania microspheres by a new method of Thermally Assisted Reactions in Aqueous Sprays. J. Mater. Chem. A 2014, 2, 3102−3109. (27) Perdew, J. P.; Burke, K.; Ernzerhof, M. Generalized gradient approximation made simple. Phys. Rev. Lett. 1996, 77, 3865. (28) Kresse, G.; Hafner, J. Ab initio molecular dynamics for liquid metals. Phys. Rev. B: Condens. Matter Mater. Phys. 1993, 47, 558−561. (29) Kresse, G.; Hafner, J. Ab initio molecular-dynamics simulation of the liquid-metal−amorphous-semiconductor transition in germanium. Phys. Rev. B: Condens. Matter Mater. Phys. 1994, 49, 14251. (30) Kresse, G.; Furthmüller, J. Efficiency of ab-initio total energy calculations for metals and semiconductors using a plane-wave basis set. Comput. Mater. Sci. 1996, 6, 15. (31) 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, 11169.

(32) Kresse, G.; Joubert, D. From ultrasoft pseudopotentials to the projector augmented-wave method. Phys. Rev. B: Condens. Matter Mater. Phys. 1999, 59, 1758. (33) Kokorin, A. I.; Bahnemann, D. W. Chemical Physics of Nanostructured Semicontuctors; VSP: The Netherlands, 2003. (34) Chen, X.; Burda, C. Photoelectron spectroscopic investigation of nitrogen-doped titania nanoparticles. J. Phys. Chem. B 2004, 108, 15446−15449. (35) Mi, Q.; Ping, Y.; Li, Y.; Cao, B.; Brunschwig, B. S.; Khalifah, P. G.; Galli, G. A.; Gray, H. B.; Lewis, N. S. Thermally stable N2intercalated WO3 photoanodes for water oxidation. J. Am. Chem. Soc. 2012, 134 (44), 18318−18324. (36) Di Valentin, C.; Finazzi, E.; Pacchioni, G.; Selloni, A.; Livraghi, S.; Paganini, M. C.; Giamello, E. N-doped TiO2: Theory and experiment. Chem. Phys. 2007, 339, 44−56. (37) Hu, X.; Tu, R.; Wei, J.; Pan, C.; Guo, J.; Xiao, W. Nitrogen atom diffusion into TiO2 anatase bulk via surfaces. Comput. Mater. Sci. 2014, 82, 107−113. (38) Zhu, L.; Hu, Q.-M.; Yang, R. The effect of electron localization on the electronic structure and migration barrier of oxygen vacancies in rutile. J. Phys.: Condens. Matter 2014, 26, 055602. (39) Iddir, H.; Ogut, S.; Zapol, P.; Browning, N. D. Diffusion mechanisms of native point defects in rutile TiO2: ab initio totalenergy calculations. Phys. Rev. B: Condens. Matter Mater. Phys. 2007, 75, 073203. (40) Wang, Z. W.; Shu, D. J.; Wang, M.; Ming, N. B. Strain effect on diffusion properties of oxygen vacancies in bulk and subsurface of rutile TiO2. Surf. Sci. 2012, 606, 186−191. (41) Tsetseris, L. Stability and dynamics of carbon and nitrogen dopants in anatase TiO2: A density functional theory study. Phys. Rev. B: Condens. Matter Mater. Phys. 2010, 81, 165205.

18670

DOI: 10.1021/acs.jpcc.5b02760 J. Phys. Chem. C 2015, 119, 18663−18670