Nitrogen-doped TiO₂ microspheres with hierarchical micro/nanostructures and rich dual-phase junctions for enhanced photocatalytic activity

Abstract

We successfully synthesized microspherical nitrogen-doped TiO₂ with hierarchical nano/microstructures, rich anatase TiO₂–TiO₂(B) phase junctions and a reduced band gap by a facile solvothermal process, followed by a urea-based solid-state reaction. Three kinds of nitrogen species in different doping sites, which can improve the photocatalytic activity and reduce the band gap, were found in the hierarchical nitrogen-doped TiO₂ microspheres. In particular, a suitable amount of nitrogen doping was found to effectively reduce the Ti³⁺ concentration in TiO₂, thus benefiting the photocatalytic reaction by reducing the recombination centers in the photocatalyst. The combination of hierarchical microspherical morphology, the rich phase junctions and narrowed band gap gives nitrogen-doped TiO₂ microspheres remarkably improved photocatalytic activity, compared with commercial mixed-phase TiO₂ (P25) and single-phase anatase-type TiO₂. Furthermore, nitrogen-doped TiO₂ microspheres display reliable photocatalytic performance for multiple cycles. The as-prepared nitrogen-doped TiO₂ microspheres hold promise as efficient photocatalysts for various photocatalytic applications.

---

Introduction

The capture and conversion of solar energy by semiconductor-based heterogeneous photocatalysis has gained significant attention from the scientific and engineering world because of the great potential for resolving energy and environmental issues toward sustainable processes; for example, photocatalytic water splitting, photodegradation of organic pollutants and photoreduction of CO₂ to solar fuels.^1–3 The efficiency of photocatalysis strongly relies on the catalyst itself, which should be well developed and designed.^4–6 Among various photocatalysts, titanium dioxide (TiO₂) has been shown to be promising due to its effective oxidizing power, good durability and low cost.^7–10 However, pristine TiO₂ is only photoactive in UV light, due to its wide band gap (3.2 eV), which limits the efficient utilization of sunlight because the UV light only captures approximately 5% of all solar energy. The question of how to utilize the visible portion of the solar spectrum, which contains >45% of solar energy, has become a great concern for the development of TiO₂-based photocatalysts and thus, it is of great importance to reduce the band gap of TiO₂.

The band gap of semiconductor materials can be reduced using a doping strategy in which the effective electronegativities of the compositional cations and the anion are controlled.^11–17 Some metal elements such as Ag, Fe and Mo, with electronegativities larger than Ti but closer to O, have been used as dopants for TiO₂ to reduce the band gap.^12–14 In addition, non-metal element (N, S and C) doping is also regarded as an effective way to reduce the band gap of TiO₂ because these elements have an electronegativity smaller than O but close to Ti.^15–17 Non-metal element doping is more favorable because cation doping generates more recombination centers for the photo-generated charge carriers, which are harmful to photocatalytic reactions. Among the various non-metal element dopants for TiO₂, nitrogen is the most investigated dopant because it provides the most significant improvement in the visible light response of TiO₂.^18–23 Besides the band gap, the high degree of recombination of photo-generated electrons and holes is a major limiting factor that controls photocatalytic efficiency.^11,24–26 Therefore, a major challenge in heterogeneous photocatalysis is to suppress the recombination of photo-generated electron–hole pairs in the photocatalysts. The recombination of electrons and holes is closely related to the morphology/microstructure of TiO₂, which in turn is strongly affected by the preparation method.^7,8,27,28 The microstructure may also affect the charge/mass transfer within the catalyst and consequently, the photocatalytic activity. Thus, a rational design of the morphological structure of TiO₂ is also crucial to achieving high photocatalytic activity.

In addition to doping and morphological tailoring, it has been reported that the creation of phase junctions in TiO₂ (such as anatase and rutile phases) is also an effective way to improve photocatalytic activity.^29,30 Photocatalysts composed of mixed TiO₂ phases have attracted more and more attention recently, as they display much higher catalytic activity than either of the single-phase components. There are two factors that can strongly affect the charge transfer from the phase possessing the higher conduction band edge to the other phase.³⁰ The first is the interface structure, which is crucial to the charge transfer. Defects, large crystallographic discrepancies and voids at the interface may form charge traps and will hinder the charge transfer in the photocatalysts with mixed TiO₂ phases. The second factor is the mobility of the photo-generated charges, as the excited electrons migrate much slower than the holes. This factor also affects the migration of the charges to the surface. Furthermore, the pathways are not independent but are interrelated.³⁰ For example, molecular oxygen on the TiO₂ surface can capture the excited CB electrons through the formation of superoxide ions, which can subsequently transform into other active chemical species, reducing the recombination of electron–hole pairs.³⁰ Taking the TiO₂(B) and anatase phase-junctions as an example, the difference between the band edges of the two phases can promote the charge transfer from anatase TiO₂ to TiO₂(B); the well-matched phase interfaces make it possible for the photo-generated holes to migrate from anatase TiO₂ to TiO₂(B).³⁰ Both the interphase transfer of the holes and the electron capturing on the surface contribute to reducing the recombination of the photo-generated charges, and enhance photocatalytic activity.³⁰ This suggests that the interfaces of mixed TiO₂ phases play a crucial role in the achievement of high photocatalytic activity.

Considering the fact that the photocatalytic activity of TiO₂ is affected by so many parameters, optimization from any one single point may not be an efficient way to maximize the photocatalytic performance of TiO₂. Thus, full-scale tailoring of the properties of TiO₂ from different aspects is highly desirable. The simultaneous nitrogen doping, morphological control and introduction of phase junctions in TiO₂ may effectively extend the photoactive region, reduce the recombination rate of electrons and holes and increase the charge/mass transfer, thereby effectively enhancing the photocatalytic performance.

Herein, we successfully synthesized a microspherical nitrogen-doped TiO₂ with hierarchical micro/nanostructures, reduced band gap and rich anatase TiO₂–TiO₂(B) phase junctions for significantly enhanced photocatalytic activity. We used a facile solvothermal approach and a subsequent solid-state reaction with urea to synthesize this nitrogen-doped TiO₂. As expected, the as-prepared nitrogen-doped TiO₂ demonstrates much higher photocatalytic activity than commercial P25. In addition, the specific microspherical structure of nitrogen doped TiO₂ helps the homogenous dispersion of the catalyst in water using electrostatic surface charges, without the need for additional capping ligands. Furthermore, the microspherical structure enables easy recycling by filtration or sedimentation, making it a highly promising photocatalyst with great potential for various photocatalytic applications.

Experimental

All chemicals were of analytical grade and were used in this study without further purification. The synthesis of hierarchical porous TiO₂ microspheres was performed by a solvothermal process. Typically, 1 mL of tetrabutyl titanate (TBT) was added dropwise to 50 mL of a polyethylene pyrrole (PVP, K-30) (0.24 g) and acetic acid solution with continuous stirring for 5 min. The obtained white suspension was transferred to a 100 mL Teflon-lined stainless-steel autoclave. After reacting at 150 °C for 24 h, the product was washed with an ethanol–water mixture several times, centrifuged at 4000 rpm, dried at 80 °C and then calcined at 400 °C for 2 h in air to obtain the porous TiO₂ microspheres. Different TiO₂ to urea ratios were used to incorporate nitrogen including 1 : 4, 1 : 8 and 1 : 16 for N–TiO₂ (1–4), N–TiO₂ (1–8) and N–TiO₂ (1–16), respectively. Taking the ratio of 1 : 4 as an example, 0.2 g of the TiO₂ microspheres were mixed with 0.8 g of urea and then heated in a covered crucible at 400 °C in air for 3 h.

The photocatalytic activities of TiO₂ and nitrogen-doped TiO₂ samples were evaluated by a model photocatalytic reaction (degradation of methyl orange, MO). A metal halide lamp (575 W, Philips) was utilized as the radiation source. In a typical process, 200 mL of a 10 mg L⁻¹ (10 ppm) MO solution were mixed with 0.1 g of photocatalyst and continuously stirred in a double-jacket reactor, which was then irradiated by UV-visible light without a filter. The reaction temperature was maintained at 25 °C by flowing cooled water. During the light irradiation, approximately 10 mL of suspension were collected every 30 min. After the removal of the photocatalyst particles from the suspension by a millipore filter with pore size of 0.45 μm (FilterBio MCE Membrane Filter), the obtained solution was analyzed with a UV-Vis spectrometer at 463 nm. The Lambert–Beer rule was applied for an absorbance band characteristic of the dye to determine its concentration. P25 (DeGussa) was a typical reference TiO₂ for the evaluation of the photocatalytic activity of the nitrogen-doped TiO₂ microspheres and other TiO₂-based photocatalysts.

Powder XRD patterns were examined on a Bruker D8 Advance X-ray diffractometer using Cu Kα radiation with a range of 10–80° (2θ) with intervals of 0.02°. Rietveld refinement of XRD patterns was obtained using the GSAS-EXPGUI software. SEM images were obtained on a JEOL S4800 instrument with an accelerating voltage of 5.0 kV. XPS measurements were performed on a Thermo ESCALAB 250 using monochromatic Al Kα radiation (1486.6 eV). All binding energies were referenced to the C 1s peak at 283.6 eV, and experimental errors were within ±0.1 eV. The morphology and microstructure of nitrogen-doped TiO₂ were investigated by TEM (JEOL 2100). Brunauer–Emmett–Teller (BET) specific surface areas and pore size distributions were obtained using adsorption data from a Micromeritics TriStar II instrument. Fourier transform infrared spectroscopy (FTIR) was performed on a Perkin-Elmer Model FTIR-100 with an MIR detector. UV-Vis DRS of samples were recorded on a JASCO V670 spectrophotometer with a Φ60 mm integrating sphere. BaSO₄ was used as a reference material. The Raman spectra were obtained in an HR800 UV Raman microspectrometer (JOBIN YVON, France). Photoluminescence (PL) spectra were obtained in PerkinElmer luminescence spectrometer.

Results and discussion

Fig. 1 shows the X-ray diffraction (XRD) patterns of the as-prepared pristine TiO₂ microspheres and the various nitrogen-doped TiO₂ microspheres with different amounts of urea in the synthesis (N–TiO₂ (1–x), where x represents the ratio of urea to TiO₂ during the synthesis). The XRD pattern of commercial P25, which is composed of anatase and rutile TiO₂ mixed phases, is also presented as a reference. The main diffraction peaks of the as-prepared TiO₂ and the nitrogen-doped TiO₂, which were both calcined at 400 °C, can be well indexed as a mixture of anatase TiO₂ and TiO₂(B) phases, as shown in Fig. 1. This suggests that anatase TiO₂–TiO₂(B) phase junctions may be present in both samples. To obtain information about the exact phase compositions, XRD patterns of the as-prepared TiO₂ and N–TiO₂ (1–8) samples were subjected to Rietveld refinements (Fig. S1a and b†). The derived weight percentages of the different phases in both samples, and their reliability factors are listed in Table S1.† The pristine TiO₂ was composed of 88 wt% anatase TiO₂ and 12 wt% TiO₂(B), whereas the N–TiO₂ (1–8) sample consisted of 82 wt% anatase TiO₂ and 18 wt% TiO₂(B). This indicates that nitrogen doping did not have a big impact on the phase structure of TiO₂.

Fig. 1 XRD patterns of P25, the pristine TiO₂ and various nitrogen-doped TiO₂ samples with different urea to TiO₂ ratios.

Raman spectra were further investigated to confirm the mixed phases of anatase and TiO₂(B) in the pristine TiO₂ and N–TiO₂ (1–8) microspheres and the results are shown in Fig. S2.† The Raman peaks at 395, 516 and 639 cm⁻¹ are typical Raman features of the anatase TiO₂ phase. Although most of the peaks of anatase TiO₂ were superimposed with the peaks of TiO₂(B), the formation of anatase TiO₂ was confirmed by the characteristic anatase peak at 395 cm⁻¹, which did not overlap with any peaks of TiO₂(B). The other peaks (except at 395 cm⁻¹) can also be indexed as the vibration modes of the TiO₂(B) phase, which was consistent with the XRD results. The appearance of characteristic peaks of TiO₂(B) at 235 cm⁻¹ indicates that the pristine TiO₂ and N–TiO₂ (1–8) microspheres were composed of mixed TiO₂ phases. As shown in Fig. S3,† the peaks of pristine TiO₂ and N–TiO₂ (1–8) in the PL spectra were around 435 nm, also suggesting the mixed anatase and TiO₂(B) phases, since the peaks of anatase and TiO₂(B) in the PL spectra were reported to be 450 and 425 nm, respectively.^31,32

The particulate morphologies of the as-prepared TiO₂ and the various nitrogen-doped TiO₂ samples were first examined by scanning electron microscopy (SEM), with representative images shown in Fig. 2. All of the samples are composed of mono-disperse spherical particles, approximately 2–3 μm in diameter; however, this special spherical structure was partially destroyed in the N–TiO₂ (1–16) sample, which was likely caused by the large amount of gas-phase products from the urea decomposition during calcination. Fig. 3 presents high magnification SEM images of the microsphere morphology in the pristine TiO₂ and N–TiO₂ (1–8) samples. Such microspheres were actually built from many belt-shaped nanostructures, as shown in Fig. 3e, the existence of which was confirmed by our previous research.³³ No obvious difference in particulate morphology between the TiO₂ and N–TiO₂ (1–8) samples was observed. This suggests that the proper amount of nitrogen doping did not greatly influence the particulate morphology of the TiO₂ samples.

Fig. 2 SEM images of the pristine TiO₂ and various nitrogen-doped TiO₂ samples with different urea to TiO₂ ratios that were calcined at 400 °C. (a) Pristine TiO₂, (b) TiO₂ to urea ratio of 1 : 4, (c) TiO₂ to urea ratio of 1 : 8 and (d) TiO₂ to urea ratio of 1 : 16.

Fig. 3 FE-SEM images of the hierarchical porous TiO₂ and N–TiO₂ (1–8): (a and b) high magnification with a few microspheres, (c and d) a single microsphere, and (e and f) magnified portions of (c and d), respectively.

The phase composition and particulate morphology of the as-prepared samples were further examined by transmission electron microscopy (TEM) and high resolution (HR)-TEM. As shown in Fig. 4a and b, the TiO₂ microsphere was assembled from many porous TiO₂ nanosheets, with length of approximately 200 nm and thickness of 30–50 nm. These leaf-like nanobelts grew from the center and combined to produce well-organized porous microspheres with a highly open structure. The HR-TEM images demonstrate the well-crystallized and highly mesoporous nature of the nanostructures, which are composed of TiO₂ nanocrystallites. Both anatase TiO₂ and TiO₂(B) phases are present in the nanocrystallites, confirmed by the fact that clear fringe planes corresponding to both phases were observed in the same leaf. For example, the (220) and (310) fringe planes of TiO₂(B) and the (103) and (202) fringe planes of anatase TiO₂ are shown in Fig. 4c, and the (311) fringe plane of TiO₂(B) and the (105) fringe plane of anatase TiO₂ are shown in Fig. 4d. Furthermore, the anatase TiO₂ phase and the TiO₂(B) phase are well connected to each other and likely form rich phase junctions inside the N–TiO₂ (1–8) sample. It has been reported that the heterojunction interface formed between anatase TiO₂ and TiO₂(B) can promote charge migration, which benefits the photocatalytic reaction.³⁰

Fig. 4 TEM images of (a) a N–TiO₂ (1–8) microsphere, (b) the nanosheets at the surface of a microsphere. (c and d) HR-TEM images of the nanosheet with two different phase structures.

Fig. 5 shows the nitrogen adsorption–desorption isotherms and the corresponding Barrett–Joyner–Halenda (BJH) pore size distributions obtained for the various samples. The specific surface areas of the samples calculated from the nitrogen adsorption–desorption isotherms in Fig. 5a are 150.3, 128.0, 116.3 and 112.5 m² g⁻¹ for pristine TiO₂, N–TiO₂ (1–4), N–TiO₂ (1–8) and N–TiO₂ (1–16), respectively. These data indicate that the reaction with urea slightly reduced the specific surface areas of the TiO₂ samples. As shown in Fig. 5b, there are two kinds of mesopores (with average pore sizes of ∼2.2 and ∼8.0 nm) in pristine TiO₂, whereas N–TiO₂ (1–16) displays a pore size distribution that is different from the other three catalysts, in which the amount of the smaller pores increased sharply, and the larger pores almost disappeared (Fig. 5b). Such a difference may be correlated with the morphological change in the microspheres in the N–TiO₂ (1–16) sample due to the overuse of urea during the synthesis.

Fig. 5 (a) Nitrogen adsorption/desorption isotherm patterns and (b) BJH pore size distribution curves of the as-prepared pristine TiO₂ and TiO₂ doped with various amounts of nitrogen.

To confirm the successful nitrogen doping of TiO₂, FTIR was conducted. The typical spectra of pristine TiO₂ and the various N–TiO₂ samples are shown in Fig. S4.† All four samples show similar FTIR spectra, in which the bands located from 600 to 850 cm⁻¹ were attributed to Ti–O stretching and Ti–O–Ti bridging stretching modes.³⁴ The band at approximately 3400 cm⁻¹ was indexed to the stretching vibration of the O–H bond and free water, whereas the band at approximately 1631 cm⁻¹ was attributed to the O–H bending vibration of chemically adsorbed water.^34,35 It was found that the incorporation of nitrogen into the oxide lattice increased the amount of hydroxyl groups and adsorbed water molecules on the surface of the catalysts. Hydroxyl groups and adsorbed water molecules could act as hole traps and produce hydroxyl radicals for the degradation of phenol molecules and dyes in water solutions.^34,35 Fig. S4b† shows the localized profile of the FTIR spectra within a range of 950 to 1550 cm⁻¹. The bands at approximately 1474, 1250 and 1080 cm⁻¹ are attributed to the vibrations of the Ti–N bond, whereas the band at 1160 cm⁻¹ is ascribed to the nitrite peak.³⁶ In addition, the band for the bending mode of the N–H bond at 1411 cm⁻¹ and the band for NO_x at approximately 1318 cm⁻¹ were also identified. No such nitrogen-related bands were observed in the pristine TiO₂.^37,38 All of the above information further confirmed the successful incorporation of different nitrogen species into TiO₂ as a consequence of being calcined with urea.

The chemical components at the surface of the various samples were analyzed by X-ray photoelectron spectra (XPS). Nitrogen doping in TiO₂ can result in various nitrogen species such as substitutional nitrogen (N_s), interstitial nitrogen (N_i) and NO_x species.³⁹ The NO_x species are most likely present on the surface of TiO₂ or trapped in the voids of the solids,⁴⁰ but the N_i and N_s species are probably present in subsurface layers.⁴¹ The N 1s XPS spectra for the N–TiO₂ (1–8) sample are shown in Fig. 6a, and the broad peak can be fitted by three peaks at 405.9, 400.1 and 398.4 eV, suggesting the presence of three independent environments for the nitrogen. The low binding energy (BE) component located at 398.5 eV is generally known as N_s and forms an O–Ti–N bond in the TiO₂ crystal lattice.^36,42 The peak at 400.1 eV can be attributed to the nitrogen in Ti–O–N bonding, which could be assigned to N_i. The peak with a BE of approximately 405.9 eV can be assigned to NO_x species, such as NO or NO₂. The ratio of N_s to N_i in the N–TiO₂ (1–8) sample is approximately 2.12 : 1, according to the area ratio of N 1s peaks as shown in Fig. 6a. Some researchers have used theoretical calculations and experimental studies to demonstrate that the N_s species can generate states above the valence band maximum (VBM). Although these states can mix with O 2p valence states to reduce the band gap of TiO₂, no band gap narrowing was obtained by calculations for the N_i species in the form of Ti–O–N, as can be detected by XPS.^39,43,44 Furthermore, Peng et al. reported that the photocatalytic activity of interstitial nitrogen-doped TiO₂ is higher than that of substitutional nitrogen-doped TiO₂.⁴⁵ It was found that the combination of N_s and N_i is beneficial for the achievement of a reduced band gap and higher photocatalytic activity. Recently, Lin et al. reported a nitrogen-doped TiO₂ with N_s and N_i species that displayed excellent photocatalytic activity, as well as a smaller band gap, although the ratio of N_s to N_i was different from the results in this study.²³ On the other hand, it was found that the enhanced visible light response by nitrogen doping may not be due solely to N_s or N_i; NO_x species can also contribute to visible light absorption because they also generate intragap states like N_s or N_i.³⁹ Therefore, high photocatalytic activity should be expected because three different nitrogen species were incorporated together in the N–TiO₂ (1–8) sample.

Fig. 6 XPS spectra of (a) N 1s for N–TiO₂ (1–8), (b) O 1s for pure TiO₂ and N–TiO₂ (1–8), and (c) Ti 2p for pure TiO₂ and N–TiO₂ (1–8).

Fig. 6b shows O 1s XPS spectra of the pristine TiO₂ and N–TiO₂ (1–8) samples. Two kinds of oxygen were found to exist in the N–TiO₂ (1–8) and pristine TiO₂ samples, and the ratio of the absorbed oxygen (531.1 eV) to lattice oxygen (529.6 eV) increased after nitrogen doping, which is in good agreement with the literature.⁴⁶ In addition, the substitution of lattice oxygen by the nitrogen anion led to a decrease in the amount of lattice oxygen, which further confirmed the successful doping of nitrogen in the TiO₂ lattice of the N–TiO₂ (1–8) sample. Fig. 6c shows the Ti 2p XPS spectra of both the as-prepared pristine TiO₂ and N–TiO₂ (1–8) samples. Two main peaks were observed and assigned to Ti 2p_1/2 and Ti 2p_3/2. Two deconvoluted Ti 2p_3/2 peaks were located at approximately 458.1 and 458.7 eV in both catalysts, which correspond to Ti³⁺ and Ti⁴⁺ valences, respectively. The corresponding Ti 2p_1/2 peaks were located at approximately 463.6 and 464.5 eV, which were also assigned to Ti³⁺ and Ti⁴⁺ valences. The atomic ratios of Ti⁴⁺ to Ti³⁺ are 0.88 and 2.09 for the pure TiO₂ and N–TiO₂ (1–8) samples, respectively. The above results suggest that nitrogen doping suppressed the formation of Ti³⁺ defects. These defects hinder the photocatalytic reactions by increasing the possibility of the recombination of electron–hole pairs.⁴⁷ As a result, enhanced photocatalytic activity can be expected when N–TiO₂ (1–8) is used as a photocatalyst.

As mentioned, to increase the photocatalytic activity of TiO₂, it is crucial to enhance the visible light absorption, which can be realized through band structure engineering. Fig. 7 compares the UV-Vis-DRS of the pristine TiO₂, the nitrogen-doped TiO₂ samples and the commercial TiO₂ (P25). The as-prepared pristine TiO₂ microspheres displayed a threshold in the UV response at 400 nm, corresponding to an estimated band gap energy (E_bg) of 3.1 eV, whereas the P25 sample had a light absorption edge of 414 nm and an E_bg of 3.0 eV. With an increasing amount of nitrogen, the absorption intensity at wavelengths below 400 nm decreased continuously, indicating reduced UV activity; this is especially true for the N–TiO₂ (1–16) sample. In addition, the adsorption edge of the nitrogen-doped samples increased continuously. N–TiO₂ (1–4) had a light adsorption edge of 433 nm, corresponding to an E_bg of 2.86 eV. As shown in Fig. 7, the samples with higher nitrogen doping, i.e., N–TiO₂ (1–8) and N–TiO₂ (1–16), had a similar threshold at 470 nm, with an estimated E_bg of 2.64 eV, which further confirmed the N_s species in the N–TiO₂ (1–8) sample, as evidenced by the XPS results. It can be concluded that the N–TiO₂ (1–8) sample is a good candidate as a catalyst for the photo-degradation of dyes because of its enhanced visible light response and comparable UV activity to the TiO₂ and N–TiO₂ (1–4) samples. It was previously reported that nitrogen doping can lead to a mixing of the N 2p orbital with the O 2p orbital to form intermediate energy levels. This action would either shift the absorption edge toward the visible light region or form an isolated narrow N 2p band above the O 2p valence band (VB) to enhance the visible light response as shown in Fig. S5† (band diagram of pristine TiO₂ and N–TiO₂).^39,43 In this case, because there is no distinct peak corresponding to a doping level in the UV-Vis spectra, nitrogen doping from urea likely results in the combination of the N 2p with O 2p orbitals to yield an intermediate energy level. This promotes electronic excitation from the valence band to the intermediate energy level by the absorption of visible light.

Fig. 7 UV-Vis diffuse reflectance spectra of P25, pristine TiO₂ and various nitrogen-doped TiO₂ samples with different urea to TiO₂ ratios.

A model photocatalytic reaction (MO degradation) was selected to test the photocatalytic activity of the as-prepared nitrogen-doped TiO₂ microspheres, and for comparison, commercial anatase TiO₂ and P25 with mixed anatase and rutile phases were also tested. To exclude the masking effect of surface adsorption on the photocatalytic activity, the test was conducted after thirty minutes to allow for the establishment of the adsorption–desorption equilibrium between MO and the catalysts in the dark. As shown in Fig. 8a, the dye adsorption is typically less than 7%, and the differences in the dye adsorption ratios for the various photocatalysts in the dark are not likely to have a strong effect on the photocatalytic activity.⁷ Once the light was turned on, different responses to the relative concentration of the dye in water with irradiation time for the various photocatalysts were observed, suggesting their different photocatalytic activities. The data were also plotted in semi logarithmic form to calculate the first-order reaction rate constants (k). As shown in Fig. 8b, the k values were 0.0009, 0.0022, 0.0026, 0.0039, 0.0049 and 0.0126 min⁻¹ for anatase TiO₂, TiO₂ microspheres calcined at 600 °C, P25, TiO₂ microspheres calcined at 400 °C, N–TiO₂ (1–16) and N–TiO₂ (1–8), respectively. Among the various photocatalysts, the N–TiO₂ (1–8) displayed the highest photocatalytic activity by far. As mentioned previously, the photocatalytic activity of a catalyst is determined by many factors such as band gap, specific surface area, particulate morphology and phase compositions.^{7–10,27–30} According to the SEM observations, the as-prepared pristine TiO₂ had a microspherical morphology similar to the structure of N–TiO₂ (1–8) (both calcined at 400 °C), with comparable specific surface areas. However, the significantly enhanced catalytic activity of N–TiO₂ (1–8) over that of the pristine TiO₂ microspheres is clearly related to the nitrogen doping. As demonstrated previously, the nitrogen doping effectively extended the light absorption edge into the visible light region (400 nm to 470 nm), thus increasing photocatalytic efficiency.

Fig. 8 (a) Photocatalytic degradation of MO under UV and visible light irradiation. (b) First-order reaction rate constants (k) vs. reaction time in the presence of different photocatalysts.

According to Fig. 8, the as-prepared pristine anatase TiO₂ (calcined at 600 °C) showed much better photocatalytic activity than the commercial anatase TiO₂ (0.0022 vs. 0.0009 min⁻¹) and was almost as high as that of P25, with rich anatase–rutile phase junctions (0.0026 min⁻¹). It should be noted that the as-prepared pristine TiO₂ calcined at 600 °C, commercial P25 and anatase TiO₂ powder samples all had similar BET surface areas of approximately 60 m² g⁻¹ (Fig. S6†), and the 600 °C calcined TiO₂ microspheres were also of a single anatase phase,⁴⁸ indicating a reasonable basis for comparison of the photocatalytic activities of these three TiO₂ materials. As mentioned, the photocatalytic activity of a catalyst is closely related to its particulate morphology, which can affect the recombination of electron–hole pairs; the nanostructure facilitates the photocatalytic reactions of TiO₂.^7–10 After calcination at 600 °C, the hierarchical porous microspherical structure with some mesopores, which were also observed for the as-prepared TiO₂ sample calcined at 400 °C, was still maintained (Fig. S6 and S7a†). On the other hand, the commercial P25 and anatase TiO₂ powders only had micropores and nanoparticles with different particle sizes, as shown in Fig. S6b and S7.† Such a special morphological structure of the as-prepared TiO₂ microspheres likely improved the interfacial charge transfer and suppressed the recombination of electron–hole pairs. These beneficial effects contribute to the superior photocatalytic activity of the as-prepared TiO₂ (600 °C), compared to the commercial TiO₂ materials.

The high specific surface area (116 m² g⁻¹), rich anatase TiO₂–TiO₂(B) phase junctions, specific microspherical particulate morphology with mesopores and the hierarchical micro/nano structures, and the nitrogen doping resulted in the superior photocatalytic performance of the N–TiO₂ (1–8) catalyst. Interestingly, the photocatalytic activity of nitrogen-doped TiO₂ was found to decrease drastically upon further increasing the nitrogen content/urea amount, which is similar to the results of nitrogen-doped TiO₂ and NaTaO₃ in literature.^19,49,50 More specifically, after 180 min of irradiation, approximately 75% of the MO was degraded by the N–TiO₂ (1–8) photocatalyst, whereas the values were 42%, 36% and 21% for N–TiO₂ (1–16), TiO₂ microspheres@400 °C and P25, respectively. The possible reason for such a difference between N–TiO₂ (1–8) and N–TiO₂ (1–16) is that excessive nitrogen doping sites may act as recombination centers for photo-induced electrons and holes, leading to a significant reduction in activity.

The recombination of the charge carriers can be explored via the PL emission spectra of pristine TiO₂, N–TiO₂ (1–8) and N–TiO₂ (1–16) samples excited by 320 nm UV light as shown in Fig. S3.† Higher PL intensity means easier recombination of the photo-generated electrons and holes. The PL spectrum of pristine TiO₂ showed the strongest emission from 400 to 550 nm at room temperature, suggesting the rapid recombination of the photo-generated electrons and holes in pristine TiO₂ and then low photocatalytic activity. In contrast, the relatively low PL intensity for nitrogen-doped TiO₂ indicates that the recombination of photo-induced electrons and holes of TiO₂ can be suppressed by nitrogen doping. Moreover, N–TiO₂ (1–8) showed lower PL intensity than N–TiO₂ (1–16) since excessive nitrogen doping sites acted as recombination centers for photo-induced electrons and holes, which is in good agreement with the results of photocatalytic activity.

The efficiency of reusable photocatalysts is also very important for practical applications.^18,50 In this study, we tested the cycling performance of the N–TiO₂ (1–8) photocatalyst. No obvious deactivation of the N–TiO₂ (1–8) photocatalyst was observed in the cycling tests for three runs, as shown in Fig. S8,† which reveals the outstanding photocatalytic stability of the catalyst. The degradation ratios of MO after 150 min of reaction were 74.4%, 72.0% and 70.3% for the first run, second run and third run, respectively. Fig. S9† is the SEM image of the photocatalyst after three cycles, and it can be seen that the porous microspherical morphology was well maintained, which enabled the high photocatalytic activity and effective suppression in the recombination of photo-generated electron–hole pairs.

Conclusions

In summary, nitrogen-doped porous TiO₂ microsphere-based photocatalysts with hierarchical nano and microstructures have been successfully synthesized by a two-step method, which includes a solvothermal process and a urea-based solid state reaction. Yellowish N–TiO₂ exhibits rich phase junctions between anatase TiO₂ and TiO₂(B) crystalline phases, as evidenced by the XRD and TEM results. Nitrogen doping can narrow the band gap of the pristine TiO₂ and then enhance the absorption of visible light. The hierarchical structure and abundant phase junctions of the photocatalysts can promote the charge carrier separation, which enhanced the photocatalytic activity remarkably. However, excessive nitrogen doping can destroy the microspherical morphology and generate lattice defects that serve as recombination centers, which are harmful for photocatalytic reactions. As a result, N–TiO₂ (1–8) exhibits a higher photocatalytic activity than the benchmark P25, pure anatase phase TiO₂ nanoparticles and microspheres and N–TiO₂ (1–16). In addition, the N–TiO₂ (1–8) photocatalyst displays good cycling performance. We propose that the high photocatalytic activity of N–TiO₂ (1–8) arises primarily from the synergistic effects of the hierarchical microstructure, rich phase junctions and the enhanced absorption of visible light. Therefore, the nitrogen-doped TiO₂ microspheres can be promising photocatalysts for various photocatalytic applications.

Acknowledgements

The authors would like to thank the Australia Research Council for supporting the project under contract FT100100134. Dr Wei Wang also acknowledges Curtin University for a postdoctoral fellowship. The authors acknowledge the facilities, scientific and technical assistance of the Curtin University Electron Microscope Facility and Curtin X-Ray Laboratory, both of which are partially funded by the University, State and Commonwealth Governments.

Notes and references

1. S. M. A. Shibli, P. S. Arun and A. V. Raj, RSC Adv., 2015, 5, 19393.
2. L. Guo, Z. Yang, B. Zu, B. Lu and X. Dou, RSC Adv., 2016, 6, 358.
3. J. E. Benedetti, D. R. Bernardo, A. Morais, J. Bettini and A. F. Nogueira, RSC Adv., 2015, 5, 33914.
4. Z. Shen, S. Sun, W. Wang, J. Liu, Z. Liu and J. C. Yu, J. Mater. Chem. A, 2015, 3, 3285.
5. Z. Jin, Q. Zhang, S. Yuan and T. Ohno, RSC Adv., 2015, 5, 4026.
6. L. Xin and X. Liu, RSC Adv., 2015, 5, 71547.
7. H. B. Wu, H. H. Hng and X. W. Lou, Adv. Mater., 2012, 24, 2567.
8. J. B. Joo, I. Lee, M. Dahl, G. D. Moon, F. Zaera and Y. Yin, Adv. Funct. Mater., 2013, 23, 4246.
9. J. H. Pan, X. Z. Wang, Q. Huang, C. Shen, Z. Y. Koh, Q. Wang, A. Engel and D. W. Bahnemann, Adv. Funct. Mater., 2014, 24, 95.
10. H. Liu, J. B. Joo, M. Dahl, L. Fu, Z. Zeng and Y. Yin, Energy Environ. Sci., 2015, 8, 286.
11. W. Wang, M. O. Tadé and Z. P. Shao, Chem. Soc. Rev., 2015, 44, 5371.
12. S. Krejčíková, L. Matějová, K. Kočí, L. Obalová, Z. Matěj, L. Čapek and O. Šolcová, Appl. Catal., B, 2012, 111–112, 119.
13. J. G. Yu, Q. J. Xiang and M. H. Zhou, Appl. Catal., B, 2009, 90, 595.
14. J. Zhang, C. Pan, P. Fang, J. Wei and R. Xiong, ACS Appl. Mater. Interfaces, 2010, 2, 1173.
15. C. Chen, P. Li, G. Wang, Y. Yu, F. Duan, C. Chen, W. Song, Y. Qin and M. Knez, Angew. Chem., Int. Ed., 2013, 52, 9196.
16. B. Sambandam, A. Surenjan, L. Philip and T. Pradeep, ACS Sustainable Chem. Eng., 2015, 3, 1321.
17. C. McManamon, J. O'Connell, P. Delaney, S. Rasappa, J. D. Holmes and M. A. Morris, J. Mol. Catal. A: Chem., 2015, 406, 51.
18. T. Wang, X. Yan, S. Zhao, B. Lin, C. Xue, G. Yang, S. Ding, B. Yang, S. Ding, B. Yang, C. Ma, G. Yang and G. Yang, J. Mater. Chem. A, 2014, 2, 15611.
19. A. Tarasov, G. Trusov, A. Minnekhanov, D. Gil, E. Konstantinova, E. Goodilin and Y. Dobrovolsky, J. Mater. Chem. A, 2014, 2, 3102.
20. Z. Xiong and X. S. Zhao, J. Am. Chem. Soc., 2012, 134, 5754.
21. M. Xie, Y. Feng, Y. Luan, X. Fu and L. Jing, ChemPlusChem, 2014, 79, 737.
22. X. Chen and C. Burda, J. Am. Chem. Soc., 2008, 130, 5018.
23. T. Lin, C. Yang, Z. Wang, H. Yin, X. Lü, F. Huang, J. Lin, X. Xie and M. Jiang, Energy Environ. Sci., 2014, 7, 967.
24. K. Maeda and K. Domen, J. Phys. Chem. C, 2007, 111, 7851.
25. M. Kitano and M. Hara, J. Mater. Chem., 2010, 20, 627.
26. R. Abe, J. Photochem. Photobiol., C, 2010, 11, 179.
27. X. Li, P. Liu, Y. Mao, M. Xing and J. Zhang, Appl. Catal., B, 2015, 164, 352.
28. K. Selvam and M. Swaminathan, RSC Adv., 2012, 2, 2848.
29. J. Zhang, Q. Xu, Z. C. Feng, M. J. Li and C. Li, Angew. Chem., Int. Ed., 2008, 47, 1766.
30. D. Yang, H. Liu, Z. Zheng, Y. Yuan, J. Zhao, E. R. Waclawik, X. Ke and H. Zhu, J. Am. Chem. Soc., 2009, 131, 17885.
31. Z. Wang, K. Lv, G. Wang, K. Deng and D. Tang, Appl. Catal., B, 2010, 100, 378.
32. M. Bellardita, A. Di Paola, L. Palmisano, F. Parrino, G. Buscarino and R. Amadelli, Appl. Catal., B, 2011, 104, 291.
33. J. Wang, Y. Zhou and Z. P. Shao, Electrochim. Acta, 2013, 97, 386.
34. L. G. Devi and R. Kavitha, Mater. Chem. Phys., 2014, 143, 1300.
35. Y. Niu, M. Xing, J. Zhang and B. Tian, Catal. Today, 2013, 201, 159.
36. G. Yang, Z. Jiang, H. Shi, T. Xiao and Z. Yan, J. Mater. Chem., 2010, 20, 5301.
37. V. Kumar, Govind and S. Uma, J. Hazard. Mater., 2011, 189, 502.
38. S. Sakthivel, M. Janczarek and H. Kisch, J. Phys. Chem. B, 2004, 108, 19384.
39. R. Asahi and T. Morikawa, Chem. Phys., 2007, 339, 57.
40. C. Di Valentin, G. Pacchioni, A. Selloni, S. Livraghi and E. Giamello, J. Phys. Chem. B, 2005, 109, 11414.
41. C. Di Valentin, E. Finazzi, G. Pacchioni, A. Selloni, S. Livraghi, M. C. Paganini and E. Giamello, Chem. Phys., 2007, 339, 44.
42. H. U. Lee, Y. Lee, S. C. Lee, S. Y. Park, J. W. Lee, C. Lim, C. Choi, M. Choi, S. Y. Lee, Y. Oh and J. Lee, Chem. Eng. J., 2014, 254, 268.
43. R. Asahi, T. Morikawa, T. Ohwaki, K. Aoki and Y. Taga, Science, 2001, 293, 269.
44. J. Wang, D. N. Tafen, J. P. Lewis, Z. Hong, A. Manivannan, M. Zhi, M. Li and N. Wu, J. Am. Chem. Soc., 2009, 131, 12290.
45. F. Peng, L. Cai, H. Yu, H. Wang and J. Yang, J. Solid State Chem., 2008, 181, 130.
46. J. Wang, H. Li, H. Li, S. Yin and T. Sato, Solid State Sci., 2009, 11, 182.
47. A. Kasahara, K. Nukumizu, T. Takata, J. N. Kondo, M. Hara, H. Kobayashi and K. Domen, J. Phys. Chem. B, 2003, 107, 791.
48. Y. Cai, B. T. Zhao, J. Wang and Z. P. Shao, J. Power Sources, 2014, 253, 80.
49. D. Nassoko, Y. Li, H. Wang, J. Li, Y. Li and Y. Yu, J. Alloys Compd., 2012, 540, 228.
50. D. Liu, Y. Jiang and G. Gao, Chemosphere, 2011, 83, 1546.

---

Footnote

† Electronic supplementary information (ESI) available: Rietveld analyses of XRD patterns, FTIR spectra, nitrogen adsorption/desorption isotherm patterns, BJH pore size distribution curves, SEM images and multiple cycled performance. See DOI: 10.1039/c6ra02966c

---