Polymorphic phase study on nitrogen-doped TiO₂ nanoparticles: effect on oxygen site occupancy, dye sensitized solar cells efficiency and hydrogen production

Abstract

In this work we show that phase formation and oxygen substitution can be controlled by the source of nitrogen used during the synthesis of TiO₂ nanoparticles. By performing a thorough study on the structure of the nanoparticles, the use of NH₄⁺ or NO₃⁻ was found to influence not only the N-doping level but also the formation of the polymorphic phase. Structural and microstructural refinement obtained by XRD pattern and data processing performed by the Rietveld refinement revealed that TiO₂ obtained with HNO₃ presents ca. 98% of anatase and ca. 2% for rutile. Meanwhile TiO₂ nanoparticles synthesized with NH₄F and NH₄Cl presents a single anatase phase with ca. 7.0% and 4.4% of nitrogen substitutional oxygen sites, respectively. The local structure of N-doped TiO₂ around the Ti atoms was investigated by X-ray absorption spectroscopy. The XANES spectra show that N-doped TiO₂ possesses a characteristic pre-edge of a single anatase structure. The coordination number decreased and the shrinking Ti–O bond distances are due to the N-doping in the TiO₂ structure. The most efficient dye sensitized solar cell and the higher hydrogen production was obtained from the TiO₂/NH₄Cl, which was obtained as a single anatase phase with intermediary concentration of N substitutional oxygen sites.

---

Introduction

Within the last two decades many research groups have devoted their efforts to the development of efficient and clean ways to produce and store clean energy.^1–7 Solar cells and hydrogen production by water splitting, come to light as very promising alternatives to efficiently contribute to the world energetic matrix. In 1972, Fujishima and Honda presented to the world the possibility of hydrogen production by water splitting⁷ and in the last 40 years numerous research has been conducted in order to boost the hydrogen generation efficiency and reduce the externally applied potential.

About 25 years has passed since the work reported by O'Regan and Grätzel in 1991,⁸ in this mean time many research groups have focused their efforts to the study of new electrolytes,^9–11 sensitizers^12–14 and semiconductors,^2,15 resulting in efficiencies of ca. 12%.¹² All this effort has made dye sensitized solar cells (DSSC) a promising candidate to efficiently contribute to the world energy matrix. In DSSC, in order to generate high photocurrent, upon photoexcitation the charge transfer from the highest occupied molecular orbital (HOMO) of the dye to the conduction band of TiO₂ must be faster than charge recombination between the photoexcited electron and the hole in the dye. Also the reduction of the oxidized lowest unoccupied molecular orbital (LUMO) by redox species in the electrolyte must be faster than the charge recombination reaction between the oxidized dye and injected photoexcited electrons in the semiconductor. Although different materials such as p-NiO, p-CuInSe₂, ZnO and SnO₂ have been studied, until now, TiO₂ photoanodes have resulted in the most efficient devices, allowing the fastest charge transfer rates.^16–19 In fact, due to the wide range of application, TiO₂ is the most extensively studied metal oxide. In order to obtain TiO₂ nanostructures presenting ideal properties for different applications, many works have been conducted on the effect of dopants, structural defects, sample preparation and thermal treatment on the TiO₂ properties. The results have shown that TiO₂ properties are directly related to the crystalline phase, crystal size, chemical structure and doping level.^20–24 For both photocatalytic hydrogen production and DSSC application, the anatase phase results in more efficient systems,^25–27 due to improved catalytic activity and electron mobility; therefore the control of phase formation during the synthesis and the solid–solid phase transformation during thermal treatment is a concern. A major drawback is the wide band gap of TiO₂, presenting absorption mode only within the ultraviolet range. In order to overcome this issue, band gap narrowing has been obtained by doping TiO₂ with metals such as Cu, Co, Ni, Mo, Fe, Ru, Au and Ag.^21,28–30 Although resulting in a wider absorption spectrum, studies have shown that metal impurities result in thermal instability. Another approach been explored is the doping by non-metallic elements.^22,31–33 Among the studied elements, nitrogen has been widely explored and many reports have shown different ways to dope TiO₂ with nitrogen and the effects on photocatalytic properties resulting in materials presenting high thermal stability, low carrier-recombination rates and higher ability to absorb light in the visible range.^34–39

Although many works have been published on this subject, a wide experimental study on the effects of different nitrogen sources on the structural and microstructural properties of TiO₂, therefore on its optical and electronic properties is warranted to contribute to the a better control of this material. Herein, we synthesized N-doped TiO₂ nanoparticles obtained using three different nitrogen sources, HNO₃, NH₄Cl and NH₄F. The materials were characterized in detail by UV-Vis spectroscopy, XRD, Rietveld refinement, XANES, EXAFS and have been applied to assemble dye sensitized solar cells and to produce hydrogen by photolysis.

Experimental part

Synthesis of N-doped TiO₂ nanoparticles

Three samples of TiO₂ nanoparticles were obtained: (i) through a well-known synthetic route using HNO₃ labelled as TiO₂/HNO₃,⁴⁰ (ii) using NH₄F labelled as TiO₂/NH₄F and (iii) using NH₄Cl labelled as TiO₂/NH₄Cl. The synthetic approach used to obtain TiO₂/NH₄F and TiO₂/NH₄Cl is also based on Grätzel work where 0.05 M of acetic acid was added to 0.05 M of titanium isopropoxide at room temperature. The solution was stirred for 15 min and poured into 72.5 mL of deionized water. The mixture was maintained under constant stirring for one hour at room temperature, to complete the hydrolysis. 0.025 M of nitric acid 63% (NH₄F or NH₄Cl) was added. The mixture was stirred for 8 h at 80 °C. After adding distilled water to obtain 95 mL, the solution was kept in autoclave and heated at 230 °C for 12 h. Finally, all samples were rinsed with anhydrous ethanol, to remove residual H⁺ and water.

Microscopy analysis

The morphology of the synthesized TiO₂ nanoparticles was obtained by Transmission Electron Microscopy (TEM) performed using a Libra Zeiss 120. For TEM, a small amount of sample was dispersed in isopropyl alcohol using a 450 W ultrasound horn and then depositing them onto a carbon-coated copper grid.

UV-Vis diffuse reflectance spectroscopy

Diffuse reflectance was performed using a Shimadzu UV-2450PC spectrophotometer with integrating sphere ISR-2200, at room temperature.

Specific surface area

N₂ adsorption–desorption isotherms were obtained at liquid nitrogen boiling point, using a micrometrics TriStar II equipped with krypton accessory. The samples were previously degassed at 140 °C under vacuum, during 6 hours. The specific surface areas were determined by BET multipoint method.

X-ray photoelectron spectroscopy

XPS experiments were carried out at beamline SXS of the Brazilian Synchrotron Light Laboratory. The operating pressure in the ultrahigh vacuum chamber (UHV) during the analysis was 1 × 10⁻⁹ Pa. The XPS spectra were collected using incident photon energy of 1840 eV. Energy steps were of 20 eV with 0.1 eV step energy and 200 ms per point acquisition time. The component of the C 1s peak of adventitious carbon was fixed at 285 eV to set the binding energy scale.

Crystalline structure

X-ray powder diffraction (XRD) patterns were obtained using a Siemens D5000 diffractometer with Cu-Kα (λ = 1.5418 Å) in a 2θ range from 10 to 90° with a step size of 0.05° and time of 1 s per step. Rietveld refinements of the structure were carried out for all samples. The structural refinements were performed from X-ray diffraction patterns for the scale factor, atomic positions, anisotropic temperature factors and patterns parameters (peak widths, cell dimensions, zero point, background point interpolation, etc.) were also varied. The instrumental resolution function of the diffractometer was obtained from well-crystallized standard LaB₆ and taken into account in separated input files. Rietveld refinements were performed using the atomic position set and the space group of the anatase structure I4₁/amd, no. 141. The unit cell is defined by the lattice vectors a and c and contains two TiO₂ units with Ti ions at 4b Wyckoff positions (0,1/4,3/8), (0,3/4,5/8) and O ions at 8e Wyckoff positions (0,1/4,u), (0,3/4,1/4 + u), (1/2,1/4,−u + 1/2) and (1/2,3/4,1/4 − u) and rutile structure P4₂/mnm, no. 136 the unit cell is defined by the lattice vectors a and c and contains two TiO₂ unit with Ti ions at 2a Wyckoff positions (0,0,0), (1/2,1/2,1/2) and O ions at 4f Wyckoff positions ±(u,u,0) and ±(u + 1/2,1/2 − u,1/2). The unit-cell parameters have been determined using X-ray diffraction and Rietveld refinement by means of the software FullProf. The pseudo-Voigt profile function of Thompson, Cox and Hastings^41,42 was used with asymmetry correction at low angle. Corrections to the preferred orientation were performed using the modified March's function. The anisotropic size broadening effects, related to the coherence volume of diffraction, were simulated using a model of spherical harmonics.^41,43,44

Electron distribution density

The electron density on a point (x; y; z) of the crystallite cell with volume V was calculated by Fourier series using the structural factors F(hkl) obtained from Rietveld refinement.⁴⁵

X-ray absorption spectroscopy (XAS) and data analysis

XAS spectra were measured on the K-edges of titanium (4966 eV) in transmission mode at room temperature using a channel-cut Si(111) crystal and three ionization chambers in the Brazilian Synchrotron Light Laboratory (LNLS, Campinas, Brazil) at the XAFS1 beamline. To calibrate the measurements, a Ti foil was used as reference sample. The powder samples were diluted with boron nitride and mounted on a tape. X-ray absorption near edge structure (XANES) data was measured from 80 eV below to 100 eV above the main absorption edge with energy step of 0.4 eV around the edge with 2 s of acquisition per step. XAS data were analysed in accordance with the standard procedure for data reduction, using ATHENA and ARTEMIS code from the IFEFFIT software package.⁴⁶ The EXAFS signal χ(k) was extracted and Fourier transformed (FT) using a Kaiser–Bessel window with a k-range from 2.2 to 11.2 Å⁻¹.

Dye-sensitized solar cells assembly

The TiO₂ pastes was screenprinted on the transparent conductive substrate (fluorine-doped tin oxide – FTO) previously soaked in 40 mM TiCl₄ aqueous solution at 80 °C for 30 minutes. The substrate was heated on a hot plate at 125 °C for 20 min and at 450 °C for 30 min in a tubular oven. The mesoporous TiO₂ electrode was immersed in 0.5 mM cis-bis(isothiocyanato)bis(2,2′-bipyridyl-4,4′-dicarboxylato)-ruthenium(II)N-719 solution of acetonitrile/tertbutyl alcohol (1 : 1 v/v) and kept at room temperature 24 h. The counter-electrodes were prepared by coating the FTO surface with a 30 μL of 1 mM hexachloroplatinic acid and heated at 400 °C. The mediator, responsible for the regeneration of the dye was placed in between the dye sensitized photoanode and the counter-electrode. The device was sealed using a polymeric film of low melting temperature (Meltonix). The electrolyte was a 0.6 M BMII, 0.03 M I₂, 0.10 M guanidinium thiocyanate and 0.5 M 4-tertbutylpyridine in a mixture of acetonitrile and valeronitrile.

Impedance spectra of DSSCs

Impedance spectra of DSSCs were recorded at open circuit potential, under 100 mW cm⁻² bias illumination over a frequency range of 100 kHz to 10 mHz at signal amplitude of 10 mV using Autolab, PGSTAT100. Circuit fitting was complied by NOVA software.

Characterization of the solar cells

The performance of the DSSCs was evaluated by current versus potential measurements, carried out using a 300 W Xenon arc lamp and an AM1.5 filter. The power of the simulated light was calibrated to 100 mW cm⁻² and recorded by a picoamperimeter Keithley, model 2400.

Hydrogen production by photolysis

10 mg of the photocatalyst: TiO₂/HNO₃, TiO₂/NH₄F or TiO₂/NH₄Cl was dispersed in a mixture of 7.5 mL of water and 2.5 mL of ethanol, kept in ultrasonic bath during 20 minutes and purged in argon. The samples were irradiated in a quartz cell by using a 240 W Xe lamp. The temperature was maintained constant at 25 °C using a cooling system. Hydrogen evolution was analysed with a gas chromatograph shimadzu 2014 equipped with a molecular column Siever 5a in a TCD detector using argon career. 50 μL aliquots were collected every 30 minutes and analysed.

Results and discussion

The size and morphology of the TiO₂ NPs were investigated by TEM. Fig. 1 shows the images of the nanoparticles as synthesized (before thermal treatment). One can observe that all samples present the same shape and nearly the same size and size distribution of around 20 nm in diameter. Hence, the morphological properties are found not dependent on the source of nitrogen used during the synthesis.

Fig. 1 (top) TEM images of the N-doped TiO₂ NPs prepared using different sources of nitrogen. (bottom) SEM images of the mesoporous films after thermal treatment at 500 °C.

The samples present similar energy band gaps with absorption edge at ca. 400 nm, towards the ultraviolet range (Fig. 2). The inset shows that nanoparticles obtained using NH₄Cl and NH₄F present a weak absorption band from ca. 400 to ca. 500 nm. Nitrogen have unpaired electrons which can easily bind to Ti⁴⁺ through acid–base reaction.^36,37 The absorption band observed from ca. 400 to ca. 500 nm is related to the interaction of mixed wave functions of nitrogen 2p with O 2p states, resulting in intermediary energy levels slightly above the top of the O 2p valence band of TiO₂.^38–40 Once the cations NH₄⁺ are hydrolyzed to NH₃ and H₃O⁺, the NH₃ seems to result in a slight higher doping level than NO₃⁻. By considering the motion of NH₃ and NO₃⁻ within TiO₂; NH₃ has lager mobility once NO₃⁻ undergoes electrostatic interaction with the TiO₂ lattice. The expected larger diffusion of NH₃ would result in a larger degree of doping, after thermal treatment. Studies on bulk diffusion of N-doped TiO₂ by NH₃ at high temperatures have suggested the presence of substitutional and interstitial nitrogen.^38,47 In addition, EPR studies has shown that calcination of NO³⁻ results in NO₂ species, meanwhile calcination of NH₃ results in NO.⁴⁸ Smaller molecules shall diffuse easily through the TiO₂ nanoparticles. The SEM images show that after thermal treatment, size and size distribution of the nanoparticles were maintained in the mesoporous film; however TiO₂/NH₄F seems to result in a more homogenous film.

Fig. 2 UV-Vis ​absorption spectra of N-doped TiO₂ NPs synthesized using different sources of nitrogen, after thermal treatment at 500 °C.

Considering the formation of intermediary energy levels upon doping, a redshift of the absorption edge is expected along with an increase in doping. In this work the redshift was only observed for TiO₂/NH₄F, which also presented a slight more intense absorption band from 400 to 500 nm than TiO₂/NH₄Cl. Previous reports have shown that improved photocatalytic activity have been obtained by co-doping TiO₂ with both nitrogen and fluorine. The improved activity was attributed to a synergetic effect of the co-doping where fluorine atoms improved visible-light absorption.^49,50 Based on these previous results, for TiO₂/NH₄F, we should account for the presence of fluorine atoms, however for refining the XRD data (discussed in next section) a good fitting was obtained by considering only nitrogen substitutional.

Fig. 3 shows the nitrogen adsorption–desorption isotherms of TiO₂/HNO₃, TiO₂/NH₄F and TiO₂/NH₄Cl as-synthesized and after thermal treatment at 450 °C. All samples show adsorption isotherms with a hysteresis type IV, characteristic of mesoporous material. The samples also present a small contribution of isotherms of the type I, characteristic of micropores. All of the isotherms present low nitrogen adsorption at P/P₀ smaller than 0.4 values, and high nitrogen volume adsorption at high values of P/P₀, above 0.8. After calcination the most significant change was observed from TiO₂/NH₄F, where a higher decrease on amount of nitrogen desorption, results from a decrease in surface area.^51,52

Fig. 3 N₂ adsorption–desorption isotherms of the N-doped TiO₂ NPs prepared using different sources of nitrogen. (left) As-synthesized and (right) thermally treated at 500 °C.

Fig. 4 shows the N 1s XPS spectra of TiO₂/HNO₃, TiO₂/NH₄F and TiO₂/NH₄Cl. One can observe a broad N 1s peak with maximum at 400 eV in the XPS spectra of N-doped samples. These values fit well with the binding energy for N-doped TiO₂ reported in the literature. Xiabo et al. showed that the binding energy peak of N 1s, related to N-doped TiO₂ is a broad peak ranging from 397.4 eV to 403.7 eV. The discussion on interstitial and substitutional doping is rather difficult and the XPS assignment for interstitial or substitutional sites in N-doped TiO₂ has been the subject of a constant debate; as stated by Asahi et al., it is subject of controversial hypothesis.⁵³ XPS peaks within 399–400 eV have been attributed to Ti–O–N and Ti–N–O,^54–56 to NH₃ adsorbed on the TiO₂ surface^56–58 and to either interstitial or substitutional nitrogen.^59–63 By refining the XRD data, the best fitting was obtained by considering substitutional nitrogen. For the sake of comparison, we have used the best possible fitting with statistically improved agreement factors, to evaluate the presence of nitrogen in the TiO₂ lattice. Therefore from the XPS and the XRD analyses it is reasonable to suggest that interstitial sites are preferentially formed at the surface while in the bulk substitutional nitrogen is observed, albeit very small.

Fig. 4 N 1s XPS spectra of TiO₂/HNO₃, TiO₂/NH₄F and TiO₂/NH₄Cl.

Fig. 5 shows the XRD patterns and Rietveld refinement fitting results for all samples. TiO₂/HNO₃ presents diffraction peaks corresponding to the anatase phase (JCPDS # 84-1286) and small additional peaks related to the rutile phase (JCPDS # 76-0649). The percentages of anatase and rutile phases from TiO₂/HNO₃ were 98.21 ± 0.70 and 1.97% ± 0.01, respectively. In both anatase and rutile structures of TiO₂, each Ti₄⁺ is surrounded by six species O₂⁻ resulting in an octahedral structure. The difference between the two polymorphic phases is related to different orthorhombic distortions of each octahedron and by the assembling of the octahedral resulting in the crystalline structure. For TiO₂/HNO₃, structural and microstructural refinement revealed that the phase of anatase was a tetragonal structure with an I4₁/amd space group, with lattice parameters: a = b = 3.7856 Å, c = 9.5068 Å, α = β = γ = 90°, an average grain size of 16.40 ± 1.53 nm and micro-strain of 15.22 ± 0.0039 (%, 10⁻⁴). In addition, the secondary phase associated to rutile was found a tetragonal structure with a P4₂/mnm space group, with lattice parameters: a = b = 4.5982 Å, c = 2.9579 Å, α = β = γ = 90°, an average grain size of 48.82 ± 0.08 nm and a micro-strain (%, 10⁻⁴) of 5.61 ± 0.0031. This difference in microstrain from anatase to rutile is mainly related to the dimension of the crystals.

Fig. 5 XRD patterns and Rietveld refinement of TiO₂/HNO₃, TiO₂/NH₄F and TiO₂/NH₄Cl. Structures of the anatase phase of TiO₂.

As observed by XPS, TiO₂/HNO₃ is also doped by nitrogen. In fact, using the approach we have used to synthesize TiO₂, nitrogen doping is process intrinsic to the synthesis. We considered TiO₂/HNO₃ as a standard sample to compare how other nitrogen sources would affect the structure and the optical properties of the materials. For N-doped TiO₂ the refinement proceeded by allowing a simultaneous presence of N and O at the 8e Wyckoff position. The atomic coordinates of N atoms were constrained to be the same as that of O. In order to study the influence of nitrogen incorporation on the sites of oxygen, the occupancies factors were let to vary freely. On the basis of chemical analysis, the rest of the positions were fixed to the theoretical occupancies factors. The occupancy factors are normalized in such a way as to represent the number of atoms in a formula unit.

Previous reports have shown that pH has a strong influence on the anatase and rutile phase formation in TiO₂.^64–67 In very acidic media the rutile phase have been found favoured over the anatase, meanwhile at intermediary pHs a single anatase phase can be obtained. Considering the previous studies suggesting that phase transformation from anatase to rutile is more difficult as the pH increases, a smaller concentration of H⁺ resulting from the hydrolysis of NH₄⁺ when compared to H⁺ from HNO₃ is probably favoring the single anatase phase formation in TiO₂/NH₄F and TiO₂/NH₄Cl. In addition, in the conditions studied in this work we have observed that the presence both F⁻ and Cl⁻ either favours the formation of anatase phase or do not contribute significantly against it. This result is interesting once previous work in the literature shows that F⁻ and SO₄²⁻ contributes to the formation of anatase, while Cl⁻ favours the formation of the rutile phase.⁶⁸ Considering the anatase phase of TiO₂/NH₄F a tetragonal structure was found with an I4₁/amd space group, lattice parameters: a = b = 3.7889 Å, c = 9.5042 Å, α = β = γ = 90°, and an average grain size of 18.21 ± 2.44 nm and a micro-strain (%, 10⁻⁴) of 18.50 ± 0.0078. In addition, by considering the TiO₂ sample a defect free lattice, a 2.9% of Ti vacancy and a substitution of N in sites of O correspondents to 7.03%, were calculated for TiO₂/NH₄F. For the TiO₂/NH₄Cl a tetragonal structure was found with an I4₁/amd space group, lattice parameters: a = b = 3.7858 Å, c = 9.5059 Å, α = β = γ = 90°, and an average grain size of 14.19 ± 0.98 nm and a micro-strain (%, 10⁻⁴) of 10.38 ± 0.0039. In addition a 0.14% of Ti vacancies and 4.43% of nitrogen substitutions in the oxygen sites were calculated. This result corroborates the absorption spectra (Fig. 2) showing that TiO₂/NH₄F presents a slight higher doping level than TiO₂/NH₄Cl and TiO₂/HNO₃. The average size for all of the samples and the shape of the grains were calculated using the modified Scherrer equation, using a model of spherical harmonics.^41,43

Fig. 5 shows the shape of the crystallites calculated by Rietveld refinement using the XRD diffraction patterns. The image of the grains were projected on the x, y and z directions. The grains present the same growth feature on the x/y plane, however a different anisotropic growth is observed on the y/z and x/y planes. In addition a more spherical shape is expected for TiO₂/NH₄Cl.

These results are very interesting, once Fig. 2 shows that TiO₂/NH₄F presents a larger concentration of defects and that TiO₂/HNO₃ and TiO₂/NH₄Cl presents similar optical spectra. Therefore by considering the results observed in Fig. 6 we suggest relation between shape and concentration defects, where a more regular structure presents smaller defects concentration. The good agreement between the structure factors from the experimental data simulated by Rietveld refinement, allowed us to obtain a Fourier electron density map, showing a two-dimensional mapping of the atoms inserted in the unit cell (Fig. 7).

Fig. 6 Projected images of the crystallites shape calculated by Rietveld refinement using the XRD diffraction data.

Fig. 7 Fourier electron density map for TiO₂/HNO₃, TiO₂/NH₄/F and TiO₂/NH₄Cl.

Small changes in the surroundings of the sites related to oxygen and nitrogen substitutional are observed. When compared to the TiO₂/HNO₃ the TiO₂/NH₄F and the TiO₂/NH₄Cl present larger content of substitutional nitrogen. For TiO₂/NH₄F one can observe an electronic density reduction associated to the presence of substitutional nitrogen in the TiO₂ lattice. Similar behaviour is observed for TiO₂/NH₄Cl. Additionally, as earlier observed in Fig. 6, TiO₂/HNO₃ and TiO₂/NH₄Cl present similar electron density map, meanwhile TiO₂/NH₄F presents a distinct behaviour probably related to the presence of the fluorine ion.

In the anatase phase the Ti–Ti distances are larger than in the rutile, on the other hand the Ti–O distances are shorter. In the rutile phase each octahedron is surrounded by 10 neighbour octahedrons, meanwhile in anatase each octahedron is surrounded by eight neighbours. This difference in structure can result in different mass densities and electronic band structures. In order to study the local structure around Ti atoms, we have carried out the XAS experiments at Ti K-edge.

The Ti K-edge XANES spectra for TiO₂/HNO₃, TiO₂/NH₄Cl and TiO₂/NH₄F is displayed in Fig. 8. The XANES pre-edge region was used to identify the local Ti coordination of the samples. All spectra have a very similar overall shape in the pre-edge and post edge regions. The pre-edge region of the samples shows a sharp peak that can be attributed to 1s to 3d transitions and the three pre-peaks in the same energy position in 4968.9, 4971.7 and 4974.1 eV that originates from the first and second coordination shells of Ti neighbours that gives evidence of the majority presence of TiO₂ anatase-type structure.^58,69 The post-edge XANES spectra (region II) are similar to each other. However, the TiO₂/HNO₃ present slight changes in the spectral range from 4990 to 4995 eV (see arrows in Fig. 8), with peaks features indicating the presence of the rutile phase. Therefore, TiO₂/HNO₃ present a mixture of majority TiO₂ anatase phase and small amount of rutile. Meanwhile the XANES spectra of TiO₂/NH₄Cl and TiO₂/NH₄F have the same single anatase crystalline structure, which corroborates the results obtained by XRD.⁴⁸

Fig. 8 Ti K-edge XANES for TiO₂/HNO₃, TiO₂/NH₄Cl and TiO₂/NH₄F.

Fig. 9(a, c, e) and (b, d, f) shows k³-weighted Ti K-edge EXAFS spectra and the corresponding Fourier Transforms (FT), respectively, for TiO₂/HNO₃, TiO₂/NH₄Cl and TiO₂/NH₄F. The k³-weighted EXAFS spectra of TiO₂/NH₄Cl and TiO₂/NH₄F (Fig. 9c and e) are very similar to each other, but different from the TiO₂/HNO₃ (Fig. 9a), that exhibit broad peak around at 4.5 Å⁻¹. This result indicates that the local environment around Ti atoms in TiO₂/NH₄Cl and TiO₂/NH₄F seems to resemble that in anatase crystalline structure, corroborating the previous discussion.

Fig. 9 (a, c and e) k³-Weighted Ti K edge EXAFS signal and (b, d and f) the corresponding Fourier transform magnitude ​where (a and b) TiO₂/HNO₃, (c and d) TiO₂/NH₄Cl and (e and f) TiO₂/NH₄F respectively. Red lines represent the best fits obtained.

Fig. 9(b, d and f) displays Ti K-edge FTs of EXAFS spectra corresponding to the shells of the Ti–O, Ti–Ti and O–Ti–O scattering path for the TiO₂. It can be seen that the FT features of all samples are look quite similar. The fitting results are summarized in Table 1. The best fitting results of the EXAFS spectra showed that the first peak at 1.96, 1.93 and 1.93 Å for TiO₂/HNO₃, TiO₂/NH₄Cl and TiO₂/NH₄F, respectively, are due to the single-scattering path from first coordination shells of Ti–O bond by oxygen octahedron nearest neighbors around the Ti atom. Both N-doped TiO₂ NPs present shorter Ti–O bond distances than TiO₂/HNO₃ material. Table 1 show that all coordination numbers (N) for single Ti–O first shells are near to 6 that are close with theoretical value. Considering the uncertainties of EXAFS fittings, the N-doped TiO₂ samples prepared in NH₄F and NH₄Cl possess the same coordination number. The smaller bond distances of the nearest Ti–O with respect to the TiO₂/HNO₃ sample suggests the presence of nitrogen doping in the structure of TiO₂ NPs.^68,70 The peaks appearing at 3.05, 3.04 and 3.04 Å are due to Ti–Ti bond distances at second shell. The shrinking Ti–Ti second shell coordination and increasing of Debye–Waller factor is an indication of increasing in disorder in TiO₂ structure probably due to the N doping.

Table 1 EXAFS fitting parameters: ^acoordination numbers (N), ^bbond length (Å) and ^cDebye–Waller factor (σ²) for N-doped TiO₂ samples

Sample	Scattering	R^a (Å)	N^b	σ^2c (10^−3 Å^2)
TiO2/HNO3	Ti–O	1.96 ± 0.02	6.39 ± 0.16	6.30 ± 0.01
	Ti–Ti	3.05 ± 0.01	3.30 ± 0.10	4.80 ± 0.02
TiO2/NH4Cl	Ti–O	1.93 ± 0.062	5.70 ± 0.74	4.80 ± 0.01
	Ti–Ti	3.04 ± 0.076	3.74 ± 0.85	5.10 ± 0.02
TiO2/NH4F	Ti–O	1.93 ± 0.004	5.85 ± 0.76	5.10 ± 0.16
	Ti–Ti	3.04 ± 0.010	3.98 ± 0.92	6.40 ± 0.24

---

From the current versus potential curves (Fig. 10), we have obtained the short-circuit current (I_sc), open circuit voltage (V_oc), fill factor (FF) and efficiency (η) from the devices assembled with the three TiO₂ samples (Table 2). The device assembled with TiO₂/NH₄Cl presents improved FF. Considering the data obtained from XRD, Rietveld refinement, XANES and EXAFS we can relate this result to the presence of a single phase anatase and to a lower concentration of detects (0.14% of Ti vacancies in and 4.43% of nitrogen substitutions in the oxygen sites). These properties result in improved electron mobility by hoping from one nanoparticle to another, in order to reach the charge collecting electrode.

Fig. 10 Current versus potential curves from the devices assembled with TiO₂/HNO₃, TiO₂/NH₄F and TiO₂/NH₄Cl.

Table 2 Electrical parameters and the efficiency of the assembled devices

	Isc (mA)	Voc (V)	FF (%)	η (%)
TiO2/HNO3	6.4	0.70	54	2.5
TiO2/NH4F	7.9	0.70	51	2.8
TiO2/NH4Cl	7.2	0.69	60	3.0

---

The electron transport through the TiO₂ network towards the collection electrode occurs by diffusion and is limited by trapping and detrapping from defects within the nanoparticles and at and across grain boundaries, hindering charge mobility. We suggest that the presence of rutile domains in the TiO₂ results in loss of electron mobility and therefore in loss of efficiency. In addition, the nitrogen doped mesoporous structures present lower charge recombination rate at the TiO₂/electrolyte interface.

Fig. 11 shows the results obtained from hydrogen production performed by photolysis in ethanol/water solution. The lowest efficiency was obtained from TiO₂/HNO₃ presenting hydrogen production rate of 0.15 mmol g⁻¹ h⁻¹ while TiO₂/NH₄Cl present a two fold increase, resulting in hydrogen production rate of 0.3 mmol g⁻¹ h⁻¹. The literature shows many works where nitrogen doped TiO₂ and bare TiO₂ present similar photocatalytic activity under UV light but improved photocatalytic activity under visible light is obtained by N-doped TiO₂.⁷¹ Some works correlate the enhancement in photocatalytic activity on the intensity of the peaks at 400–404 eV.^71,72

Fig. 11 Graph showing hydrogen production from TiO₂/HNO₃, TiO₂/NH₄F and TiO₂/NH₄Cl.

These results corroborate the DSSC measurements, showing that the intermediary doping combined to the single anatase phase, which results in higher photocurrent and lower recombination rate favours both photolysis and solar cell efficiency.

Fig. 12 compares the Nyquist plots of the assembled DSSCs. These plots consist of three components; ohmic resistance R₀, the left arcs (R₁) (observed in the high frequency region >1 kHz) and the larger arcs (R₂) (observed in the frequency range of 100–50 Hz). According to the literature,^73,74 the arcs in the high frequency represent the redox reaction at Pt/electrolyte interface while the arcs at lower frequency are attributed to the TiO₂/electrolyte interface. Table 3 summarizes the results obtained by fitting the experimental data, using an equivalent circuit containing a combination of constant phase element (CPE) and resistance components (inset of Fig. 12); where CPE₁ and CPE₂ correspond to R₁ and R₂, respectively.⁷⁵ These impedances have an obvious effect on the charge transportation in the DSSCs.

Fig. 12 Impedance spectra of the DSSCs from TiO₂/HNO₃, TiO₂/NH₄F and TiO₂/NH₄Cl. The inset gives the equivalent circuit used to fit the impedance data; where CPE₁ and CPE₂ are the constant phase elements.

Table 3 EIS parameters of the three devices determined by fitting the data according to the equivalent circuit model (see Fig. 12)

Sample	R0 (Ω)	R1 (Ω)	R2 (Ω)	Rtotal (Ω)
TiO2/NH4Cl	11.78	20.76	80.69	113.23
TiO2/NH4F	11.39	53.20	97.91	162.50
TiO2/HNO3	12.14	55.26	103.41	170.81

---

Consequently, the cell performance is improved when the total resistance (R_total = R₀ + R₁ + R₂) is small.^73,74 One can observe that the R_total for TiO₂/NH₄Cl and TiO₂/NH₄F is smaller than for TiO₂/HNO₃. In addition, the smallest R_total is obtained from TiO₂/NH₄Cl, which should present improved charge transportation and higher photoelectrochemical performance. These results are consistent with the energy conversion efficiency presented in Table 2.

Conclusions

In this work we show that single anatase phase TiO₂ nanoparticles can be obtained by tuning the nitrogen source used during the synthesis. The use of NH₄⁺ which is hydrolyzed to NH₃ leads to the formation of a single anatase phase, meanwhile NO₃⁻ was found to result in a mixed anatase/rutile lattice. In addition, we have observed a dependence of the amount of nitrogen on the source of nitrogen, hence effecting doping level and electronic properties of the nanoparticles. Although only a small change in the optical spectra was detected, the changes observed by XRD, XANES and EXAFS shows that the efficiency of dye sensitized solar cells and hydrogen production are deeply dependent on the N-doping of the semiconductor.

Acknowledgements

The authors are grateful for financial support from the following Brazilian agencies: Conselho Nacional de Desenvolvimento Científico e Tecnológico (Processo: 477804/2011-0, 490221/2012-2 and 472243/2013-6). Our thanks are also extended to the Brazilian Synchrotron Light Laboratory (LNLS) for use of its XAFS1 experimental facilities (proposal XAFS1-17154).

Notes and references

1. N. Tétreault and M. Grätzel, Novel Nanostructures for next Generation Dye-Sensitized Solar Cells, Energy Environ. Sci., 2012, 5, 8506.
2. M. T. Laranjo, N. C. Ricardi, L. T. Arenas, E. V. Benvenutti, M. C. de Oliveira, M. J. L. Santos and T. M. H. Costa, TiO₂ and TiO₂/SiO₂ Nanoparticles Obtained by Sol–Gel Method and Applied on Dye Sensitized Solar Cells, J. Sol-Gel Sci. Technol., 2014, 72, 273–281.
3. J. A. Fernandes, P. Migowski, Z. Fabrim, A. F. Feil, G. Rosa, S. Khan, G. J. Machado, P. F. P. Fichtner, S. R. Teixeira and M. J. L. Santos, TiO₂ Nanotubes Sensitized with CdSe via RF Magnetron Sputtering for Photoelectrochemical Applications under Visible Light Irradiation, Phys. Chem. Chem. Phys., 2014, 16, 9148–9153.
4. R. G. Sabat, M. J. L. Santos and P. Rochon, Surface Plasmon-Induced Band Gap in the Photocurrent Response of Organic Solar Cells, Int. J. Photoenergy, 2010, 2010, 5.
5. Y. Shi, R. B. M. Hill, J. H. Yum, A. Dualeh, S. Barlow, M. Grätzel, S. R. Marder and M. K. A. Nazeeruddin, High-Efficiency Panchromatic Squaraine Sensitizer for Dye-Sensitized Solar Cells, Angew. Chem., Int. Ed., 2011, 50, 6619–6621.
6. X. Fan, Z. Chu, F. Wang, C. Zhang, L. Chen, Y. Tang and D. Zou, Wire-Shaped Flexible Dye-Sensitized Solar Cells, Adv. Mater, 2008, 20, 592–595.
7. A. Fujishima and K. Honda, Nature, 1972, 238, 37–38.
8. B. O'Regan and M. Grätzel, A Low-Cost, High-Efficiency Solar Cell Based on Dye-Sensitized Colloidal TiO₂ Films, Nature, 1991, 353, 737–740.
9. N. Papageorgiou, Y. Athanassov, M. Armand, P. Bonho te, H. Pettersson, A. Azam and M. Grätzel, The Performance and Stability of Ambient Temperature Molten Salts for Solar Cell Applications, J. Electrochem. Soc., 1996, 143, 3099–3108.
10. U. Scherf, Counterion Pinning in Conjugated Polyelectrolytes for Applications in Organic Electronics, Angew. Chem., Int. Ed., 2011, 50, 5016–5017.
11. W. Zhang, J. Li, S. Jiang and Z.-S. Wang, POSS with Eight Imidazolium Iodide Arms for Efficient Solid-State Dye-Sensitized Solar Cells, Chem. Commun., 2014, 50, 1685–1687.
12. F. Gao, Y. Wang, D. Shi, J. Zhang, M. Wang, X. Jing, R. Humphry-baker, P. Wang, S. M. Zakeeruddin and M. Gra, Enhance the Optical Absorptivity of Nanocrystalline TiO Film with High Molar Extinction Coefficient Ruthenium Sensitizers for High Performance Dye-Sensitized Solar Cells, J. Am. Chem. Soc., 2008, 130, 10720–10728.
13. C.-W. Lee, H.-P. Lu, C.-M. Lan, Y.-L. Huang, Y.-R. Liang, W.-N. Yen, Y.-C. Liu, Y.-S. Lin, E. W.-G. Diau and C.-Y. Yeh, Novel Zinc Porphyrin Sensitizers for Dye-Sensitized Solar Cells: Synthesis and Spectral, Electrochemical, and Photovoltaic Properties, Chemistry, 2009, 15, 1403–1412.
14. S. Ito, H. Miura, S. Uchida, M. Takata, K. Sumioka, P. Liska, P. Comte, P. Péchy and M. Grätzel, High-Conversion-Efficiency Organic Dye-Sensitized Solar Cells with a Novel Indoline Dye, Chem. Commun., 2008, 41, 5194–5196.
15. S. Ito, M. K. Nazeeruddin, S. M. Zakeeruddin, P. Péchy, P. Comte, M. Gratzel, T. Mizuno, A. Tanaka and T. Koyanagi, Study of Dye-Sensitized Solar Cells by Scanning Electron Micrograph Observation and Thickness Optimization of Porous TiO₂ electrodes, Int. J. Photoenergy, 2009, 2009, 12–14.
16. S. A. Haque, Y. Tachibana, D. R. Klug and J. R. Durrant, Charge Recombination Kinetics in Dye-Sensitized Nanocrystalline Titanium Dioxide Films under Externally Applied Bias, J. Phys. Chem. B, 1998, 102, 1745–1749.
17. S. Ardo and G. J. Meyer, Photodriven Heterogeneous Charge Transfer with Transition-Metal Compounds Anchored to TiO₂ Semiconductor Surfaces, Chem. Soc. Rev., 2009, 38, 115–164.
18. Y. Tachibana, J.-E. Moser, M. Grätzel, D. R. Klug and J. R. Durrant, Subpicosecond Interfacial Charge Separation in Dye-Sensitized Nanocrystalline Titanium Dioxide Films, J. Phys. Chem., 1996, 100, 20056–20062.
19. T. Hannappel, B. Burfeindt, W. Storck and F. Willig, Measurement of Ultrafast Photoinduced Electron Transfer from Chemically Anchored Ru-Dye Molecules into Empty Electronic States in a Colloidal Anatase TiO₂ Film, J. Phys. Chem. B, 1997, 101, 6799–6802.
20. A. Di Paola, G. Marcì, L. Palmisano, M. Schiavello, K. Uosaki, S. Ikeda and B. Ohtani, Preparation of Polycrystalline TiO₂ Photocatalysts Impregnated with Various Transition Metal Ions: Characterization and Photocatalytic Activity for the Degradation of 4-Nitrophenol, J. Phys. Chem. B, 2002, 106, 637–645.
21. A. Fuerte, M. D. Hernández-Alonso, A. J. Maira, A. Martínez-Arias, M. Fernández-García, J. C. Conesa and J. Soria, Visible Light-Activated Nanosized Doped-TiO₂ Photocatalysts, Chem. Commun., 2001, 24, 2718–2719.
22. S. Sato, R. Nakamura and S. Abe, Visible-Light Sensitization of TiO₂ Photocatalysts by Wet-Method N Doping, Appl. Catal., A, 2005, 284, 131–137.
23. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga, Visible-Light Photocatalysis in Nitrogen-Doped Titanium Oxides, Science, 2001, 293, 269–271.
24. Y. Hwu, Y. D. Yao, N. F. Cheng, C. Y. Tung and H. M. Lin, X-Ray Absorption of Nanocrystal TiO₂, Nanostruct. Mater., 1997, 9, 355–358.
25. J. Zhang, P. Zhou, J. Liu and J. Yu, New Understanding of the Difference of Photocatalytic Activity among Anatase, Rutile and Brookite TiO₂, Phys. Chem. Chem. Phys., 2014, 16, 20382–20386.
26. Y. Wang, L. Li, X. Huang, Q. Li and G. Li, New Insights into Fluorinated TiO₂ (brookite, Anatase and Rutile) Nanoparticles as Efficient Photocatalytic Redox Catalysts, RSC Adv., 2015, 5, 34302–34313.
27. X. Wang, A. Kafizas, X. Li, S. J. A. Moniz, P. J. T. Reardon, J. Tang, I. P. Parkin and J. R. Durrant, Transient Absorption Spectroscopy of Anatase and Rutile: The Impact of Morphology and Phase on Photocatalytic Activity, J. Phys. Chem. C, 2015, 150423173617003.
28. A. Di Paola, G. Cufalo, M. Addamo, M. Bellardita, R. Campostrini, M. Ischia, R. Ceccato and L. Palmisano, Photocatalytic Activity of Nanocrystalline TiO₂ (brookite, Rutile and Brookite-Based) Powders Prepared by Thermohydrolysis of TiCl₄ in Aqueous Chloride Solutions, Colloids Surf., A, 2008, 317, 366–376.
29. J. Zhu, F. Chen, J. Zhang, H. Chen and M. Anpo, Fe³⁺–TiO₂ Photocatalysts Prepared by Combining Sol-Gel Method with Hydrothermal Treatment and Their Characterization, J. Photochem. Photobiol., A, 2006, 180, 196–204.
30. H. Yamashita, M. Harada, J. Misaka, M. Takeuchi, Y. Ichihashi, F. Goto, M. Ishida, T. Sasaki and M. Anpo, Application of Ion Beam Techniques for Preparation of Metal Ion-Implanted TiO₂ Thin Film Photocatalyst Available under Visible Light Irradiation: Metal Ion-Implantation and Ionized Cluster Beam Method, J. Synchrotron Radiat., 2001, 8, 569–571.
31. Y. Cong, J. Zhang, F. Chen and M. Anpo, Synthesis and Characterization of Nitrogen-Doped TiO₂ Nanophotocatalyst with High Visible Light Activity, J. Phys. Chem. C, 2007, 111, 6976–6982.
32. T. Ohno, T. Mitsui and M. Matsumura, Photocatalytic Activity of S-Doped TiO₂ Photocatalyst under Visible Light, Chem. Lett., 2003, 32, 364–365.
33. Y. Liu, X. Chen, J. Li and C. Burda, Photocatalytic Degradation of Azo Dyes by Nitrogen-Doped TiO₂ Nanocatalysts, Chemosphere, 2005, 61, 11–18.
34. S. Rodrigues, K. T. Ranjit, S. Uma, I. N. Martyanov and K. J. Klabunde, Single-Step Synthesis of a Highly Active Visible-Light Photocatalyst for Oxidation of a Common Indoor Air Pollutant: Acetaldehyde, Adv. Mater, 2005, 17, 2467–2471.
35. I. Paramasivam, H. Jha, N. Liu and P. Schmuki, A Review of Photocatalysis Using Self-Organized TiO₂ Nanotubes and Other Ordered Oxide Nanostructures, Small, 2012, 8, 3073–3103.
36. R. P. Vitiello, J. M. Macak, A. Ghicov, H. Tsuchiya, L. F. P. Dick and P. Schmuki, N-Doping of Anodic TiO₂ Nanotubes Using Heat Treatment in Ammonia, Electrochem. Commun., 2006, 8, 544–548.
37. R. Nakamura, T. Tanaka and Y. Nakato, Mechanism for Visible Light Responses in Anodic Photocurrents at N-Doped TiO₂ Film Electrodes, J. Phys. Chem. B, 2004, 108, 10617–10620.
38. O. Diwald, T. L. Thompson, T. Zubkov, S. D. Walck and J. T. Yates, Photochemical Activity of Nitrogen-Doped Rutile TiO₂(110) in Visible Light, J. Phys. Chem. B, 2004, 108, 6004–6008.
39. J. A Rodriguez, T. Jirsak, J. Dvorak, S. Sambasivan and D. Fischer, Reaction of NO₂ with Zn and ZnO: Photoemission, XANES, and Density Functional Studies on the Formation of NO₃, J. Phys. Chem. B, 2000, 104, 319–328.
40. S. Ito, T. N. Murakami, P. Comte, P. Liska, C. Grätzel, M. K. Nazeeruddin and M. Grätzel, Fabrication of Thin Film Dye Sensitized Solar Cells with Solar to Electric Power Conversion Efficiency over 10%, Thin Solid Films, 2008, 516, 4613–4619.
41. J. Rodríguez-Carvajal, Recent Advances in Magnetic Structure Determination by Neutron Powder Diffraction, Phys. B, 1993, 192, 55–69.
42. P. Thompson, D. E. Cox and J. B. Hastings, Rietveld Refinement of Debye–Scherrer Synchrotron X-Ray Data from Al₂O₃, J. Appl. Crystallogr., 1987, 20, 79–83.
43. W. A. Dollase, Correction of intensities of preferred orientation powder diffractometry: application of the March model, J. Appl. Crystallogr., 1986, 19, 267–272.
44. L. B. McCusker, R. B. Von Dreele, D. E. Cox, D. Louër and P. Scardi, Rietveld Refinement Guidelines, J. Appl. Crystallogr., 1999, 36–50.
45. J. K. Burdett, T. Hughbanks, G. J. Miller, J. W. Richardson and J. V. Smith, Structural-Electronic Relationships in Inorganic Solids: Powder Neutron Diffraction Studies of the Rutile and Anatase Polymorphs of Titanium Dioxide at 15 and 295 K, J. Am. Chem. Soc., 1987, 109, 3639–3646.
46. B. Ravel and M. Newville, ATHENA, ARTEMIS, HEPHAESTUS: Data Analysis for X-Ray Absorption Spectroscopy Using IFEFFIT, J. Synchrotron Radiat., 2005, 12, 537–541.
47. R. G. Palgrave, D. J. Payne and R. G. Egdell, Nitrogen Diffusion in Doped TiO₂(110) Single Crystals: A Combined XPS and SIMS Study, J. Mater. Chem., 2009, 19, 8418–8425.
48. C. Di Valentin, E. Finazzi, G. Pacchioni, A. Selloni, S. Livraghi, M. C. Paganini and E. Giamello, N-Doped TiO₂: Theory and Experiment, Chem. Phys., 2007, 339, 44–56.
49. C. Di Valentin and G. Pacchioni, Catal. Today, 2013, 206, 12–18.
50. D. Huanga, S. Liao a, J. Liu, Z. Danga and L. Petrik, J. Photochem. Photobiol., A, 2006, 184, 282–288.
51. J. C. Yu, J. Yu, W. Ho, Z. Jiang and L. Zhang, Effects of F-Doping on the Photocatalytic Activity and Microstructures of Nanocrystalline TiO₂ Powders, Chem. Mater., 2002, 14, 3808–3816.
52. L. Lo Presti, M. Ceotto, F. Spadavecchia, G. Cappelletti, D. Meroni, R. G. Acres and S. Ardizzone, Role of the Nitrogen Source in Determining Structure and Morphology of N-Doped Nanocrystalline TiO₂, J. Phys. Chem. C, 2014, 118, 4797–4807.
53. R. Asahi, T. Morikawa, H. Irie and T. Ohwak, Nitrogen-Doped Titanium Dioxide as Visible-Light-Sensitive Photocatalyst: Designs, Developments, and Prospects, Chem. Rev., 2014, 114, 9824–9852.
54. J. Wang, D. Nyago Tafen, J. P. Lewis, Z. Hong, A. Manivannan, M. Zhi, M. Li and N. Wu, Origin of Photocatalytic Activity of Nitrogen-Doped TiO₂ Nanobelts, J. Am. Chem. Soc., 2009, 131(39), 12290–12297.
55. J. H. Xu, J. X. Li, W. L. Dai, Y. Cao, H. X. Li and K. N. Fan, Simple fabrication of twist-like helix N, S-codoped titania photocatalyst with visible-light response, Appl. Catal., B, 2008, 79, 72–80.
56. M. Sathish, B. Viswanathan, R. P. Viswanath and C. S. Gopinath, Chem. Mater., 2005, 17, 6349.
57. J. F. Moulder, W. F. Stickle, P. E. Sobol and K. D. Bomben. Handbook of X-ray photoelectron Spectroscopy, Eden Prairie, 1992.
58. R. J. J. Jansen and H. Van Bekkum, Carbon, 1995, 33, 1021.
59. B. Viswanathan and K. R. Krishanmurthy, Nitrogen Incorporation in TiO₂: Does It Make a Visible Light Photo-Active Material?, Int. J. Photoenergy, 2012, 269654.
60. J. Wang, D. N. Tafen and J. P. Lewis, et al. Origin of Photocatalytic Activity of Nitrogen-Doped TiO₂ Nanobelts, J. Am. Chem. Soc., 2009, 131(341), 2290–12297.
61. Y. Cong, J. Zhang, F. Chen, M. Anpo and D. He, Preparation, photocatalytic activity, and mechanism of nano-TiO₂ codoped with nitrogen and iron(III), J. Phys. Chem. C, 2007, 111(28), 10618–10623.
62. J. Yang, H. Bai, X. Tan and J. Lian, IR and XPS investigation of visible-light photocatalysis-Nitrogen-carbon-doped TiO₂ film, Appl. Surf. Sci., 2006, 253(4), 1988–1994.
63. D. Mitorj and N. H. Kisch, The nature of nitrogen-modified titanium dioxide photocatalysts active in visible light, Angew. Chem., Int. Ed., 2008, 47, 9975–9978.
64. R. Parra, M. S. Goes, M. S. Castro, E. Longos, P. R. Bueno and J. A. Varela, Reaction pathway to the synthesis of anatase via the chemical modification of titanium isopropoxide with acetic acid, Chem. Mater., 2008, 20, 143–150.
65. X. Chen and G. Gu, Study on synthesis of nanometer TiO₂ crystal from organic compound in liquid phase at normal pressure and low temperature, Chin. J. Inorg. Chem., 2002, 18(7), 749–752.
66. Y. Hu, H.-L. Tsai and C.-L. Huang, Effect of brookite phase on the anatase-rutile transition in titania nanoparticles, J. Eur. Ceram. Soc., 2003, 23, 691–696.
67. S. M. Abdel-Azim, A. K. Aboul-Gheit, S. M. Ahmed, D. S. El-Desouki and M. S. A. Abdel-Mottaleb, Preparation and Application of Mesoporous Nanotitania Photocatalysts Using Different Templates and pH Media, Int. J. Photoenergy, 2014, 687597, 1–11.
68. J. M. Wu, S. Hayakawa, K. Tsuru and A. Osaka, Porous Titania Films Prepared from Interactions of Titanium with Hydrogen Peroxide Solution, Scr. Mater., 2002, 46, 101–106.
69. T. Kubo and A. Nakahira, Local Structure of TiO₂-Derived Nanotubes Prepared by the Hydrothermal Process, J. Phys. Chem. C, 2008, 112, 1658–1662.
70. C. Belver, R. Bellod, S. J. Stewart, F. G. Requejo and M. Fernández-García, Nitrogen-Containing TiO₂ Photocatalysts. Part 2. Photocatalytic Behavior under Sunlight Excitation, Appl. Catal., B, 2006, 65, 309–314.
71. J. G. Yu, M. H. Zhou, B. Cheng and X. J. Zhao, J. Mol. Catal. A: Chem., 2006, 246, 176.
72. X. B. Chen and C. Burda, J. Phys. Chem. B, 2004, 108, 15446.
73. T. Hoshikawa, T. Ikebe, R. Kikuchi and K. Eguchi, Effects of electrolyte in dye-sensitized solar cells and evaluation by impedance spectroscopy, Electrochim. Acta, 2006, 51, 5286–5294.
74. X. Sun, Y. Liu, Q. Tai, B. Chen, T. Peng, N. Huang, S. Xu, T. Peng and X.-Z. Zhao, High Efficiency Dye-Sensitized Solar Cells based on a Bi-Layered Photoanode Made of TiO₂ Nanocrystallites and Microspheres with High Thermal Stability, J. Phys. Chem. C, 2012, 116(22), 11859–11866.
75. Q. Wang, J.-E. Moser and M. Gratzel, Electrochemical Impedance Spectroscopic Analysis of Dye-Sensitized Solar Cells, J. Phys. Chem. B, 2005, 109, 14945–14953.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra17225j

---