Seeding-induced construction of N-doped TiO₂-bronze@g-C₃N₄ two-dimensional binary nanojunctions with enhanced photocatalytic activity

Abstract

Nitrogen-doped TiO₂-bronze@g-C₃N₄ (TiO₂ (B)@g-C₃N₄) two-dimensional binary heterojunctions were constructed based on seeding-induced growth through a microwave-assisted solvothermal process and subsequent thermal treatment in a vacuum. The morphology of the TiO₂ (B) nanosheets could be controlled by tuning the concentration of the Ti precursor, which determined the enhanced photoelectron activity. The optimal photocatalytic activity for the degradation of methyl orange (MO) under low-intensity visible-light illumination was obtained at a TiO₂ (B)/g-C₃N₄ molar ratio of 1 : 1, which was 12.7 and 7.9 times higher than that of pure g-C₃N₄ and P25, respectively. The photocatalytic activity was further enhanced by about 7.7% after in situ N-doping. The improvement in photocatalytic activity of N-doped TiO₂ (B)@g-C₃N₄ hetero-nanojunctions was attributable to the strong absorption in the visible-light region and better separation of photogenerated electron–hole pairs at the nanojunction interface, a result due to the large contact area between N-doped TiO₂ (B) and g-C₃N₄ nanosheets. We have explained the photocatalytic degradation of MO molecules largely in terms of the direct oxidation by the photogenerated holes and partly by the contribution of the superoxide radicals.

---

1. Introduction

Semiconductor photocatalysis has attracted much attention due to its potential environmental remediation in the degradation of organic pollutants in wastewater and the atmosphere.^1,2 Recently, graphite-like carbon nitride (g-C₃N₄), which has a delocalized conjugated π structure, has attracted substantial attention in visible-light-driven photocatalytic H₂ evolution and environmental purification.³ Nevertheless, rapid recombination of the photogenerated electrons and holes is a main issue limiting the practical applications of g-C₃N₄. In the past years, a number of reliable and facile strategies have been developed to fabricate modified g-C₃N₄-based photocatalysts with unique photoelectronchemical properties and photocatalytic performance, for example, protonation,⁴ forming porous structures,⁵ doping by nonmetal or metal elements,^6–8 and constructing heterostructures.^9–11 Among these, coupling g-C₃N₄ with other appropriate semiconductors to construct heterostructures is an effective pathway to improve the separation of photogenerated charge carriers, resulting in an enhanced photoelectron chemical and photocatalytic performance. Recently, many heterojunction-based g-C₃N₄ materials, such as g-C₃N₄/CeO₂,¹² g-C₃N₄/TiO₂,¹³ g-C₃N₄/Bi₂S₃,¹⁴ and g-C₃N₄/ZrO₂,¹⁵ have been reported. These materials exhibits significantly enhanced photoelectronchemical and photocatalytic activities as compared to pure g-C₃N₄. Therefore, it is envisioned that coupling g-C₃N₄ with other semiconductors might be a feasible approach to improve the visible-light harvesting efficiency and inhibit the recombination of photoexcited electron–hole pairs.

As a typical semiconductor photocatalyst, TiO₂ has been most widely investigated due to its good photocatalytic activity, low cost, and non-toxicity.^3,16 The studies on the coupling of TiO₂ with other visible-light-sensitized semiconductors have demonstrated to be an interesting way to improve its photocatalytic capability. Anatase and rutile TiO₂ have been widely studied due to their good photocatalytic performance.^17–19 TiO₂-bronze (denoted as TiO₂ (B)) is a new form of TiO₂ with a lamellar structure,²⁰ having a large surface area, which can shorten the migration time of charge carries and suppress their recombination.²¹ Since g-C₃N₄ and TiO₂ (B) have closely related structures, charge transfer is easy through the interface of the heterojunction,^22,23 while multiple functionality may be promoted. To the best of our knowledge, however, there are no reports on the coupling of g-C₃N₄ with TiO₂ (B) for the photocatalytic removal of organic pollutants.

In this study, an in situ N-doped TiO₂ (B)/g-C₃N₄ two-dimensional (2D) nanojunction was fabricated via a facile microwave-assisted solvothermal route combined with a thermal treatment. The 2D structure based on the concept of seeding-induced growth. This confined the TiO₂ (B) growth on g-C₃N₄ by “seeding” its surface with microscopic nucleation sites for crystal growth in the dilute regime to reduce homogeneous nucleation. In this way a large surface area was created, which increased the contact area for adsorption of organic molecules. The photodegradation efficiency of methyl orange (MO) over the as-prepared photocatalysts was measured under visible-light irradiation. It was found that the hybrid photocatalyst decomposed MO more efficiently than the TiO₂ (B) or g-C₃N₄ sample. From these results we propose a new photodegradation mechanism. We believe that the novel N-TiO₂ (B)@g-C₃N₄ heterostructured photocatalyst is a good candidate to be applied in environmental remediation.

2. Experimental

2.1 Photocatalyst preparation

2.1.1 g-C₃N₄ synthesis. The g-C₃N₄ powders were synthesized via the thermal polycondensation of urea according to the procedures described previously.²⁴ In a typical synthesis, urea (10.0 g) in a crucible was heated at a heating rate of 10 °C min⁻¹ in a muffle furnace from room temperature (RT) to 550 °C and kept at thus temperature for 4 h. The resulting yellow product was collected for further use.

2.1.2 Process of “seeding” the g-C₃N₄. Titanium tetrachloride was dissolved in deionized water to form a homogeneous aqueous solution (30 μM L⁻¹). The g-C₃N₄ powders were seeded by immersion in this TiCl₄ aqueous solution at 70 °C for 60 min, followed by thoroughly washing with deionized water and drying at 80 °C for 12 h.

2.1.3 N-TiO₂ (B)@g-C₃N₄ preparation. The TiO₂ (B)@g-C₃N₄ sample was prepared using the one-step solvothermal method. 50 mg g-C₃N₄ was added into 32 mL of ethylene glycol while applying ultrasonic stirring for 30 min to completely disperse the g-C₃N₄. Meanwhile, a TiCl₄ aqueous solution was drop wise added to the above mixture. After continuously stirring for 2 h, 1.0 mL of NH₃·H₂O (25%) was added into the mixture, which was kept at 70 °C for 1 h. Then the mixture was transferred into a Teflon-lined autoclave and microwave-irradiated at 150 °C for 2 h. In the reaction process, the power of the microwave oven (XH-800C, Beijing Xianghu Co., Ltd.) was set at 1000 W. After the autoclave was cooled to RT, the precipitate was filtered, washed with ethanol to remove the organic remnants, and dried at 50 °C overnight. The fabrication with different amounts of titanium precursors was done according to the above procedures. The TiO₂ (B)@g-C₃N₄ hybrid samples were denoted as TOCN-X, in which X was the TiO₂ (B)/g-C₃N₄ molar ratio of 0.25, 0.5, 1, and 1.5, respectively. The N-TiO₂ (B)@g-C₃N₄ (denoted as NTOCN-1) was prepared by heating the synthesized TOCN-1 at 250 °C for 2 h in vacuum. For comparison purposes, single-phase TiO₂ (B) (denoted as TO) was also prepared without g-C₃N₄ addition according to the above procedures.²¹

The working electrodes were prepared as follows: 50 mg of the as-prepared photocatalyst was dispersed in 0.75 mL of ethanol to produce a slurry, which was then overlaid on a 2 cm × 4 cm fluorine-doped tin oxide (denoted as FTO) glass electrode via the doctor blading method. The electrodes were calcined at 200 °C for 10 h in air.

Urea (CH₄N₂O), ethylene glycol (C₂H₆O₂), tert-butyl alcohol (C₄H₁₀O), and triethanolamine (C₆H₁₅NO₃) were purchased from Sinopharm Chemical Reagent Beijing Co. Ltd. Ammonia solution (NH₃·H₂O) and ammonium oxalate ((NH₄)₂C₂O₄) were purchased from Beijing Chemical Works. Titanium tetrachloride (TiCl₄) was purchased from Tianjin Fuchen Chemical Reagents Factory. All chemicals were analytical grade and used as received without further purification.

2.2. Characterization

The samples were characterized using X-ray diffraction (Shimadzu XRD-7000 with Cu Kα radiation, λ = 0.15418 nm), scanning electron microscopy (Hitachi S-8020U) and transmission electron microscopy (JEOL JEM-2010, operated at 200 kV; FEI Tecnai G2 F30, operated at 300 KV). Nitrogen adsorption–desorption isotherms and pore-size distributions were measured on a Micromeritics ASAP 2020 at −196 °C. Specific surface areas and mean pore sizes were calculated according to the BET and Barrett–Joyner–Halenda (BJH) methods, respectively. The thickness of the samples was determined by atomic force microscopy (AFM, Veeco Metrology, MultiMode-V). X-ray photoelectron spectroscopic (XPS) analysis was performed on a PHI Quantera SXM X-ray photoelectron spectrometer using the Al Kα radiation. The photoluminescence (PL) spectra of g-C₃N₄ and TiO₂ (B)@g-C₃N₄ were recorded using the Edinburgh Instruments Xe900 equipped with a xenon (Xe) lamp with an excitation wavelength of 380 nm. UV-visible (UV-vis) absorption spectra of the samples were measured on a Shimadzu-2550 UV-visible spectrophotometer. Electrochemical and photoelectrochemical activities were measured in three-electrode quartz cells with an electrochemical workstation (CHI-660D, China). 0.1 mol L⁻¹ Na₂SO₄ aqueous solution was used as the electrolyte solution. A platinum electrode was used as the counter electrode, a saturated calomel electrode (SCE) was used as the reference electrodes, and g-C₃N₄ and TiO₂ (B)@g-C₃N₄ electrodes on FTO served as the working electrodes. The visible light source was a 500 W Xe lamp (CHF-XW-500 W) with a 420 nm cutoff filter. The voltage over the cell was adjusted with reference to the SCE. Photoresponses of the samples with light on and light off were measured at 0 V and the electrochemical impedance spectra (EIS) were also determined at 0 V. A sinusoidal ac signal of 5 mV was applied to the electrode in the frequency range of 1–100 KHz.

2.3. Photocatalytic evaluation

Photocatalytic activities of the samples were evaluated by measuring the decomposition of MO during the irradiation of a LED lamp (6 W, λ = 420–425 nm, about 8.0 mW cm⁻²). 100 mg of the sample was dispersed in 100 mL of the MO aqueous solution (10 mg L⁻¹). Before illumination, the mixed solution was continuously stirred in the dark for 30 min to establish the adsorption–desorption equilibrium (Fig. S1†). The concentration of the supernatant during the degradation process was measured with a UV-vis spectrometer at λ_max, being the wavelength that corresponds to the maximum adsorption of the dye solution. λ_max = 505 nm and the pH of the above solution was adjusted to 2.9–3.0 using H₂SO₄ to guarantee that λ_max did not change. To investigate the photocatalytic mechanism of the samples, the experiments of hydroxyl radical (OH˙) scavenger were carried out using 1.0 mmol tert-butyl alcohol (TBA), and hole (h⁺) scavenger tests were conducted using 1.0 mmol ammonium oxalate (AO) and triethanolamine (TEOA). N₂ purging experiments were also made to confirm the results and to identify the role of superoxide species (˙O₂⁻) in the photocatalysis.

3. Results and discussion

3.1. Structure and textural property

The crystal structures of the as-derived TOCN architectures were analyzed by the XRD technique, as illustrated in Fig. 1. In the XRD pattern of pure g-C₃N₄, the strongest XRD peak at 2θ = 27.7° could be indexed as the (002) diffraction plane (JCPDS PDF# 87-1526), which was due to the long-range interplanar stacking of the conjugated aromatic system. The typical (100) peak at 2θ = 13° corresponded to an in-plane structural packing motif. The comparison of the d-spacing values between TiO₂ (B) and the JCPDS files is listed in Table 1. The XRD peaks of pure TiO₂ were in good agreement with the metastable polymorphic phase of TiO₂ (B) (JCPDS PDF# 74-1940).²⁴ The diffraction peaks at 2θ = 14.2°, 24.9°, and 48.6° could be assigned to the (001), (110), and (020) planes of monoclinic TiO₂ (B), respectively. It should be noted that no phase transition occurred after calcination of the samples. As for the composite samples, the intensity of the (002) crystal plane reduced substantially by increasing the TiO₂ (B) amount. Furthermore, the characteristic XRD peak at 2θ = 27.8°, which overlapped with the diffraction peak of g-C₃N₄, could be indexed as the (1̄11) crystal plane of TiO₂ (B). The peak was clearly observed in the wide-scale XRD patterns of all TiO₂ (B)@g-C₃N₄ samples.

Fig. 1 XRD patterns of the g-C₃N₄, TO, and TOCN samples.

Table 1 Comparison of the d-spacing values between NTOCN-1 and the JCPDS files

	TiO2(B): JCPDS PDF# 74-1940
	(hkl)	2-Theta	d-spacing
JCPDS	001	14.186°	0.6238 nm
	110	24.979°	0.3561 nm
	1̄11	27.681°	0.322 nm
	002	28.596°	0.311 nm
	020	48.634°	0.187 nm
	Sample
NTOCN-1	001	14.2°	0.617 nm
	110	24.9°	0.345 nm
	1̄11	27.8°	0.320 nm
	002	28.62°	0.303 nm
	020	48.6°	0.190 nm

---

The morphology of the as-prepared TOCN samples was examined by field-emission SEM (FE-SEM) and TEM. The SEM image of g-C₃N₄ in Fig. 2a shows a clearly observable wrinkle 2D structure. Fig. 2b shows the morphology of the NTOCN-1 nanojunctions. It can be seen that TiO₂ (B) nanosheets are located on the surface of g-C₃N₄. From the high magnification image (inset of Fig. 2b), TiO₂ (B) nanosheets are homogeneously dispersed on the surface of the building block nanosheets. The dispersion states and structures of NTOCN-1 nanojunctions can be seen in Fig. 2c and d. TiO₂ (B) displays a good distribution on the surface of g-C₃N₄, which is consistent with the SEM observations. The areas with different colors in Fig. 2c indicate the different phases, in which the nanosheets with dark color can be assigned to TiO₂ (B), whereas the gray area is ascribed to g-C₃N₄. The lattice fringes with a crystal plane distance of 0.303 nm (Fig. 2d) were attributed to the spacing of the (002) plane of TiO₂ (B). The interfaces between TiO₂ (B) and g-C₃N₄ can be seen clearly, which is a confirmation of the formation of N-TiO₂ (B)/g-C₃N₄ heterojunctions. Fig. 3 shows the AFM pattern of the N-TiO₂ (B)/g-C₃N₄ heterojunctions. 10 mg NTOCN-1 sample were ultrasonic dispersed in 50 mL pure ethanol, and then the solution was dropped on a monocrystalline silicon wafer. The roughness of monocrystalline silicon wafer was <0.4 nm. As shown in Fig. 3, the morphology of NTOCN-1 confirms the SEM and TEM observations, which TiO₂ (B) nanosheets are homogeneously dispersed on the surface of the g-C₃N₄.

Fig. 2 Typical FE-SEM (a) image of g-C₃N₄ and SEM (b), TEM (c), and STEM (d) images of the NTOCN-1 sample.

Fig. 3 AFM pattern of the NTOCN-1 sample.

Fig. S2† depicts the absorbance spectra of g-C₃N₄, TO, TOCN-1 and NTOCN-1 composites. Fig. S2† shows a clear red shift of the absorption in TOCN-1, NTOCN-1 samples, in comparison with that of TO. The as-prepared sample show obvious light absorption in visible light region due to the absorption edge of NTOCN-1 up to 450 nm, which may due to the existence of g-C₃N₄ and nitrogen species.

Textural properties of the as-prepared samples were investigated by the nitrogen adsorption–desorption measurement. The obtained sorption isotherms and BJH pore-size distributions are shown in Fig. 4. The adsorption–desorption isotherm of g-C₃N₄ has been classified as type IV with a H3 hysteresis loop, which is characteristic of mesoporous materials with slit-shaped pores.²⁵ TiO₂ (B) and NTOCN-1 composites also displayed typical type IV isotherms but with a H4 hysteresis loop, indicative of formation of a mesoporous structure with a slit pores (narrow pores and micropores). The surface area (S_BET) of TiO₂ (B) was 342 m² g⁻¹, whereas that of pure g-C₃N₄ was 70 m² g⁻¹. The NTOCN-1 heterostructure possessed a surface area of 312 m² g⁻¹. The pore-size distribution curves clearly show that the NTOCN-1 material had two pore-size families located at 3.8 and 48.0 nm (Fig. 4b). These different pore sizes are attributed to released NH₃ and CO₂ bubbles that could act as the soft templates during the polymerization of urea²⁶ and the inter-aggregated g-C₃N₄ flakes, respectively. These results indicate that the synthesis of TiO₂ (B) on the surface of g-C₃N₄ yielded a large surface area and mesoporosity. TiO₂ (B)/g-C₃N₄ heterojunctions prepared by other molar ratios presented regular S_BET values, which were 277.35, 290.76 and 330.85 m² g⁻¹ for the 0.25 : 1, 0.5 : 1 and 1.5 : 1 molar ratio samples, respectively. As expected, the g-C₃N₄ nanosheets showed a relatively low S_BET of 70 m² g⁻¹, but TiO₂ (B) was about 342 m² g⁻¹ because of the small size of the particles. It is to be expected that the S_BET value of TiO₂ (B)@g-C₃N₄ increased gradually by increasing the TiO₂ (B) loading. But the high S_BET value did not always correspond with the high catalytic activity, because the photocatalytic activity is a result of multiple effects, such as BET area and electron–hole pair separation efficiency. Although TOCN-1.5 showed the highest S_BET value of 330.85 m² g⁻¹, excessive TiO₂ (B) nanoparticles aggregated seriously on the surface of g-C₃N₄ and restrained the separation of photoinduced electron–hole pairs, resulting in a low catalytic activity.

Fig. 4 N₂ adsorption–desorption isotherms of (a) the representative samples and (b) pore-size distribution of the NTOCN-1 sample.

To further analyze the NTOCN-1 heterojunctions and study the interaction of TiO₂ with g-C₃N₄, XPS analysis was performed and the results are shown in Fig. 5. The C 1s spectrum (Fig. 5a) of the NTOCN-1 sample displays two peaks at binding energies (BE) = 284.48 and 288.18 eV. The peak at BE = 284.48 eV is attributed to contaminated carbon and surface carbon,²⁷ whereas the peak at BE = 288.18 eV is assigned to C–(N)3 groups of g-C₃N₄.^28–32 Fig. 5b is the N 1s XPS spectrum of the TOCN-1 sample. This spectrum has been fitted to four peaks, which are attributed to the sp²-hybridized nitrogen (C–N=C) species at BE = 398.68 eV,²⁶ the N–(C)3 species at BE = 399.76 eV,^26,28,33 the quaternary N bonded to three C atoms in the aromatic cycles at BE = 401.12 eV (ref. 34) and charging effects at BE = 404.35 eV.³⁵ Taking into account that the BE of the incorporated N dopant in TiO₂ as interstitial N or O–Ti–N is also at 399.76 eV,³³ it is obvious to assume that TiO₂ is doped by N after calcination. The O 1s spectrum (Fig. 5c) of the TOCN-1 sample could be fitted to two peaks at BE = 529.78 and 531.10 eV; these peaks are due to the lattice oxygen species in TiO₂ (ref. 33) and the surface hydroxyl species,³⁶ respectively. We also made a comparison between the Ti 2p spectra of the TOCN-1 and NTOCN-1 samples. Fig. 5d shows the Ti 2p spectra of TiO₂ in the TOCN-1 and NTOCN-1 samples. The peaks at BE = 458.48 eV (Ti 2p_3/2) and 464.18 eV (Ti 2p_1/2) in the TOCN-1 sample are attributed to the Ti⁴⁺ species in the TiO₂ clusters.³⁷ Interestingly, a slight shift in BE of Ti 2p in the NTOCN-1 sample was observed as compared to that in the TOCN-1 sample, suggesting that the electron density distribution changed in the Ti atoms. The negative shift in BE of Ti 2p has been ascribed to the presence of N in the TiO₂ lattice,³⁸ because N has a N lower electronegativity than O. This leads to the partial electron transfer from N to Ti.³⁹ From these considerations we conclude that N has been successfully doped into the TiO₂ lattice.

Fig. 5 (a) C 1s, (b) N 1s, (c) O 1s, and (d) Ti 2p XPS spectra of the typical samples.

3.2. Photocatalytic performance

MO is one of the most widely used dyes with a stable azoic structure and weak adsorption on the photocatalyst surface, and therefore complete decomposition of MO is extremely difficult. For comparison purposes, Degussa P25 and the directly mixed powders of g-C₃N₄ and TiO₂ (B) with a molar ratio of 1 : 1 were also tested under the same reaction conditions. The pH value of the MO solution was adjusted to around 3 by adding a H₂SO₄ solution.

Fig. 6a shows the visible-light-driven photocatalytic activities of the g-C₃N₄-based and P25 photocatalysts for the degradation of MO. The degradation rate constant (k) was calculated from the fitted curve (Fig. 6b) according to the equation of −ln (C/C₀) = kt, where C₀ represents the concentration at the adsorption–desorption equilibrium of the photocatalyst before illumination and C the concentration at the illumination time. The result indicated that the TiO₂ (B)@g-C₃N₄ sample exhibited a higher photocatalytic activity than the pristine g-C₃N₄ and P25 samples under visible-light irradiation, implying that the photocatalytic ability of N-TiO₂ (B)@g-C₃N₄ was closely related to the TiO₂ (B) nanosheets. The photocatalytic activity of TiO₂ (B)@g-C₃N₄ increased gradually with higher TiO₂ (B) loading and reached a maximum when the TiO₂ (B)/g-C₃N₄ molar ratio was equal to 1 : 1 (TOCN-1). Moreover, the NTOCN-1 sample showed a higher photocatalytic efficiency than TOCN-1, namely 98.6% of the MO degradation was achieved within 115 min. The apparent rate constant obtained with the NTOCN-1 sample was 0.03845 min⁻¹, which was about 1.2 times higher than that (0.03215 min⁻¹) obtained with the TOCN-1 sample and 12.7 times higher than (0.00302 min⁻¹) obtained with the g-C₃N₄ sample. The latter achieved a degradation rate of only 29.8% within 115 min. Moreover, the degradation efficiency obtained with the NTOCN-1 sample was almost 7.9 times higher than that (0.00488 min⁻¹) obtained with the P25 sample. Huang et al. studied the photoreactivity of g-C₃N₄/TiO₂ (anatase) photocatalyst prepared by a solvothermal process, and found that it was 34% higher than that of commercial P25.¹³ Yang et al. prepared C₃N₄/N-TiO₂ (mixed phases of anatase and rutile) through calcining the precursors of g-C₃N₄ and TiO₂. However, the photodegradation efficiency of representative samples was only 1.2 times higher than that of P25.⁴⁰ The enhancement of the photocatalytic activity of our N-TiO₂ (B)@g-C₃N₄ samples might be attributed to the high efficiency of the charge separation induced by the hybrid effect of g-C₃N₄ and N-doped TiO₂. In addition, the structure of the composite might also play an important role. Firstly, the mesoporous composite displayed a large surface area, which created a large number of active sites on the surface, thereby accelerating the MO degradation rate. Secondly, the unique 2D architecture of the nanojunctions, which facilitated the transport of both photoinduced electrons and holes through the internal electrostatic field in the junction region, favored also the visible light to reach the surface and to excite the g-C₃N₄. Finally, the undoped TiO₂ in the hybrid could only absorb UV light, but N-doped TiO₂ efficiently enhanced the visible-light absorption. The photoexcited electrons of g-C₃N₄ transferred easily to the conduction band (CB) of TiO₂ under visible-light irradiation, while the holes migrated from the valence band (VB) of TiO₂ to the VB of g-C₃N₄. Thus, the recombination of photogenerated electron–hole pairs could be significantly suppressed, with more electrons staying in the CB of TiO₂ and more holes remaining in the VB of g-C₃N₄. As the loading of TiO₂ (B) further increased, however, the degradation rate decreased. This might be attributed to the aggregation of excess TiO₂ species, which led to shielding of active sites on the g-C₃N₄ surface. This decreased the intensity of visible light through the channel to the g-C₃N₄, hence, the efficiency of charge separation was reduced. In order to show the heretojunction effect of the bronze phase on the photocatalytic activity of N-doped TiO₂ and to verify the significance of interface coupling for the photocatalytic activity of TiO₂ (B)@g-C₃N₄, the mechanically mixed g-C₃N₄ and TiO₂ sample with a TiO₂/g-C₃N₄ molar ratio of 1 : 1 was used for comparison purposes. The MO photodegradation rate constant (0.01495 min⁻¹) obtained over the mixed powders was much lower than that (0.03215 min⁻¹) obtained over the TOCN-1 sample. This result indicated that the enhanced visible-light-driven photocatalytic activity of TiO₂ (B)@g-C₃N₄ came from the interface coupling. Therefore, it is crucial to inhibit the recombination of electron–hole pairs and to increase the ability for separation of photogenerated charge carriers because of the application TiO₂ (B) to the surface of g-C₃N₄.

Fig. 6 (a) Photocatalytic degradation efficiency of MO over the typical samples under 420 nm LED light irradiation and (b) the pseudo-first-order rate constant k obtained over the different samples.

In order to analyze the role of specific surface area and identify the effect of N doping or purposed nanojunction, we calculated the degradation rate constant (k) divided by the S_BET. The values of k per specific surface area were 0.69 × 10⁻⁴ min⁻¹ m⁻² for the S-mixture sample, 1.05 × 10⁻⁴ min⁻¹ m⁻² for TOCN-1 and 1.23 × 10⁻⁴ min⁻¹ m⁻² for NTOCN-1, respectively. The highest k value belonged to NTOCN-1; so, it was concluded that the enhanced photocatalytic activity must be attribute to the N doping and the nanojunctions between g-C₃N₄ and TiO₂ (B).

3.3. Photochemical behavior

PL spectra of the TiO₂ (B)@g-C₃N₄ heterojunction and g-C₃N₄ samples were recorded at an excitation wavelength of 380 nm to determine the efficiency of the transfer and separation processes of photogenerated electron–hole pairs.⁴¹ The intensity of the PL emission spectra is an indicator for the recombination level of photoexcited electron–hole pairs. The stronger the PL intensity, the higher the recombination of charge carriers.^42,43 As shown in Fig. 7a, the main emission band was centered at about 460 nm for pure g-C₃N₄, which was due to the recombination process of self-trapped excitation.⁴⁴ The emission peak positions of the TOCN-1 and NTOCN-1 samples were similar to those of the g-C₃N₄ sample. However, the emission intensity of the NTOCN-1 composite sample was lower than that of the pristine g-C₃N₄ and TOCN-1 samples. This result clearly indicates that the recombination of photogenerated charge carriers was inhibited and that both the formation of the TiO₂ (B)@g-C₃N₄ heterojunctions and N-doping into the TiO₂ lattice contributed to the separation of photoinduced charge carriers.

Fig. 7 (a) Photoluminescence spectra of g-C₃N₄, TOCN-1, and NTOCN-1 nanojunctions, (b) transient photocurrent density versus time of g-C₃N₄, TO, TOCN-1, and NTOCN-1 nanojunctions, and (c) EIS Nyquist plots of g-C₃N₄, TO, TOCN-1, and NTOCN-1 nanojunctions.

To further understand the photogenerated electron transfer between g-C₃N₄ and TiO₂ (B), the transient photocurrent responses of the samples that were overlaid on the FTO electrodes were measured at pulsed illumination from a Xe lamp (λ > 420 nm). From photocurrent versus time curves (Fig. 7b) of the g-C₃N₄, TO, TOCN-1, and the NTOCN-1 composite samples with a 210 s interval in an on/off intermittent irradiation cycle mode, one can observe that all samples exhibited fast and reproducible photocurrent response at each illumination pulse. For the NTOCN-1 sample, a sharp increase in photocurrent appeared once the Xe lamp irradiation was on, but when the irradiation was interrupted the photocurrent rapidly dropped to a steady-state value. The electrodes of these samples demonstrated a rapid photocurrent response when the light was switched back on again. The generated photocurrent was reproducible and stable during the three on/off intermittent irradiation cycles. It can be seen that the photocurrent generated of the NTOCN-1 sample was about 1.7 and 1.3 times higher than that generated with the g-C₃N₄ and TOCN-1 samples, respectively. It is known that the higher the photocurrent, the higher the electrons–hole separation efficiency, and hence, the higher the photocatalytic activity. The enhanced photocurrent of the N-TiO₂ (B)@g-C₃N₄ sample implied that the photoinduced electron transfer was more efficient than that of the g-C₃N₄ sample: this was beneficial for enhancing the photocatalytic activity.

The electrochemical impedance spectroscopic (EIS) technique was used to analyze the charge carrier transport in the g-C₃N₄, TO, TOCN-1, and NTOCN-1 samples. In Nyquist diagrams, the radii of the arcs are associated with the charge transfer at the interface between the electrode and electrolyte solution; a small radius corresponds to a lower charge transfer resistance.⁴⁵ As shown in Fig. 7c, the arc radius of the EIS Nyquist plot of the NTOCN-1 composite sample was the smallest of all our samples at dark and irradiation conditions, demonstrating that the N-doped TiO₂ (B)@g-C₃N₄ nanojunctions exhibited the smallest charge transfer resistance. So, substantial shuttling of charges between the electrode and electrolyte and faster interfacial charge transfer occurred at the composite interface.⁴⁶ This agrees with foregoing conclusion that the enhanced coupling was due to the formation of nanojunctions and the doping of N in TiO₂.

3.4. Photocatalytic mechanism

In order to understand the photocatalytic mechanism of the heterostructured TiO₂ (B)@g-C₃N₄ samples, the active species generated during MO degradation over TiO₂ (B)@g-C₃N₄ were identified by performing the trapping and N₂ purging experiments. In this study, t-BuOH was used as the hydroxyl radical (˙OH) scavenger,⁴⁷ and ammonium oxalate (AO) and triethanolamine were used to reduce the number of holes (h⁺).^48–52 Superoxide radicals involved in the photodegradation could be formed according to O₂ + e = ˙O₂⁻,⁵³ so, the N₂ purging experiment was employed to identify the effect of ˙O₂⁻.

Fig. 8 shows the influence of various scavengers on the photocatalytic activity of the TOCN-1.5 sample for the degradation of MO. h⁺ was the main active species generated in the current system, since the degradation efficiency of MO decreased significantly in the presence of AO and TEOA, in which the removal rate constant (k) decreased from 0.02074 to 0.00631 and 0.00545 min⁻¹, respectively. However, the ˙OH radicals were a minor active species that was auxiliary for the oxidization of MO. For that reason the photocatalytic degradation of MO over the TOCN-1.5 sample was suppressed to a small degree after the addition of TBA to the reaction system, showing a slight influence on the k value (from 0.02074 to 0.01718 min⁻¹). The N₂ purging experiment in the anoxic suspension shows that near half of photocatalytic degradation rate was depressed, while 69.58% and 73.72% of decrease in removal rate constant were found when employing AO and TEOA as hole scavengers, respectively. These results indicate that the holes play more important role than electrons in photodegradation. In our experiments, the photoinduced electrons can transfer from the CB of g-C₃N₄ to that of TiO₂, owing to the well-matched band positions between g-C₃N₄ and TiO₂ and leading to spatial separation of photo-carriers (Fig. 9). The survived electrons cannot spontaneously transfer back to the CB of g-C₃N₄, and then can be consumed by the species in the solution when there is no O₂ feeding. With sustained N₂ purging in the suspension, oxygen was almost excluded. Then the active species ˙O₂⁻ won't be produced without O₂ participation in the solution. Hence, the trigger of the photocatalytic activity can be attributed to the photoinduced holes. These results were also confirmed by hole-trapping experiments, because the photocatalytic activity dramatically decreased when using the hole scavengers. Therefore, although all the three reactive species, ˙O₂⁻, ˙OH and h⁺, are involved in the photocatalysis, the photoinduced holes should be the greatest contributor to the oxidation of MO molecules for the TiO₂ (B)@g-C₃N₄ sample.

Fig. 8 (a) Influence of various scavengers on the visible-light-driven photocatalytic activity of TOCN-1.5, and (b) apparent rate constants obtained over TOCN-1.5 for MO degradation under visible-light irradiation.

Fig. 9 Schematic illustration of the photocatalytic mechanism over the N-TiO₂ (B)@g-C₃N₄ composite sample under visible-light irradiation.

It is inferred from the above experimental results that the enhanced photocatalytic activity of the N-doped TiO₂ (B)@g-C₃N₄ sample is related to the following characteristics. Fast generation of the photogenerated carriers on g-C₃N₄ and the rapid separation and transfer of these photogenerated carriers at the interface of g-C₃N₄ and TiO₂ (B) are paramount for the enhanced visible-light-driven photocatalytic performance of the TiO₂ (B)@g-C₃N₄ sample. Since the CB potential of g-C₃N₄ is more negative than that of TiO₂, the electrons can diffuse from g-C₃N₄ to TiO₂, resulting in fast transfer of negative charges from g-C₃N₄ through the nanojunction interfaces. Additionally, N-TiO₂ (B) and g-C₃N₄ possess matched Fermi levels. After visible-light irradiation, the electron that are excited to the CB of g-C₃N₄ can rapidly move to the TiO₂ (B) nanosheets, while the holes in the VB of TiO₂ (B) can migrate to g-C₃N₄, thus restraining the recombination of e_CB⁻ and h_VB⁺ (Fig. 9). This was also supported by results of the photoelectrochemical and PL investigations.

4. Conclusions

N-doped TiO₂ (B)@g-C₃N₄ samples have been successfully prepared based on the seeding-induced growth through a facile solvothermal process. The N-TiO₂ (B)@g-C₃N₄ samples exhibited significantly enhanced visible-light-driven photocatalytic activity, about 12.7 and 7.9 times higher than the g-C₃N₄ and P25 samples, respectively, which were attributed to the effective utilization of visible light due to N-doping and high separation efficiency of the photogenerated electron–hole pairs at the heterojunction interfaces. The seeding-induced growth was important in building the 2D binary structured photocatalytic materials and this strategy is expected to be extended for depositing other components on the surface of g-C₃N₄.

Acknowledgements

This work was supported by the National Natural Science Foundation (no. 51471006, 51225402), Beijing Natural Science Foundation (no. 2151001) and Guangxi Natural Science Foundation (no. 2014GXNSFBA118039).

References

1. M. R. Hoffmann, S. T. Martin, W. Y. Choi and D. W. Bahnemann, Chem. Rev., 1995, 95, 69–96.
2. A. Mills and S. L. Hunte, J. Photochem. Photobiol., A, 1997, 108, 1–35.
3. D. Dvoranova, V. Brezova, M. Mazur and M. A. Malati, Appl. Catal., B, 2002, 37, 91–105.
4. M. Tahir, C. B. Cao, N. Mahmood, F. K. Butt, A. Mahmood, F. Idrees, S. Hussain, M. Tanveer, Z. Ali and I. Aslam, Appl. Mater. Interfaces, 2014, 6, 1258–1265.
5. K. Kailasam, J. D. Epping, A. Thomas, S. Losse and H. Junge, Energy Environ. Sci., 2011, 4, 4668–4674.
6. J. Zhang, M. Zhang, G. Zhang and X. Wang, ACS Catal., 2012, 2, 940–948.
7. X. C. Wang, X. F. Chen, A. Thomas, X. Z. Fu and M. Antonietti, Adv. Mater., 2009, 21, 1609–1612.
8. Z. Ding, X. Chen, M. Antonietti and X. Wang, ChemSusChem, 2011, 4, 274–281.
9. Y. J. Wang, R. Shi, J. Lin and Y. F. Zhu, Energy Environ. Sci., 2011, 4, 2922–2929.
10. Y. J. Wang, X. J. Bai, C. S. Pan, J. He and Y. F. Zhu, J. Mater. Chem., 2012, 22, 11568–11573.
11. C. S. Pan, J. Xu, Y. J. Wang, D. Li and Y. F. Zhu, Adv. Funct. Mater., 2012, 22, 1518–1524.
12. L. Y. Huang, Y. P. Li, H. Xu, Y. G. Xu, J. X. Xia, K. Wang, H. M. Li and X. N. Cheng, RSC Adv., 2013, 3, 22269–22279.
13. Z. A. Huang, Q. Sun, K. L. Lv, Z. H. Zhang, M. Li and B. Li, Appl. Catal., B, 2015, 164, 420–427.
14. X. S. Rong, F. X. Qiu, J. Yan, H. Zhao, X. L. Zhu and D. Y. Yang, RSC Adv., 2015, 5, 24944–24952.
15. X. X. Wang, L. H. Zhang, H. J. Lin, Q. Y. Nong, Y. Wu, T. H. Wu and Y. M. He, RSC Adv., 2014, 4, 40029–40035.
16. S. Yin, H. Yamaki, M. Komatsu, Q. W. Zhang, J. S. Wang, Q. Tang, F. Saito and T. Sato, J. Mater. Chem., 2003, 13, 2996–3001.
17. D. O. Scanlon, C. W. Dunnill, J. Buckeridge, S. A. Shevlin, A. J. Logsdail, S. M. Woodley, C. R. A. Catlow, M. J. Powell, R. G. Palgrave, I. P. Parkin, G. W. Watson, T. W. Keal, P. Sherwood, A. Walsh and A. A. Sokol, Nat. Mater., 2013, 12, 798–801.
18. J. S. Wang, H. Li, H. Y. Li, C. Zou, H. Wang and D. S. Li, ACS Appl. Mater. Interfaces, 2014, 6, 1623–1631.
19. J. Y. Liao, B. X. Lei, H. Y. Chen, D. B. Kuang and C. Y. Su, Energy Environ. Sci., 2012, 5, 5750–5757.
20. R. Marchand, L. Brohan and M. Tournoux, Mater. Res. Bull., 1980, 15, 1129–1133.
21. G. L. Xiang, T. Y. Li, J. Zhuang and X. Wang, Chem. Commun., 2010, 46, 6801–6803.
22. M. Li, X. F. Yu, S. Liang, X. N. Peng, Z. J. Yang, Y. L. Wang and Q. Q. Wang, Adv. Funct. Mater., 2011, 21, 1788–1794.
23. X. Chen, S. Shen, L. Guo and S. S. Mao, Chem. Rev., 2010, 110, 6503–6570.
24. J. H. Liu, T. K. Zhang, Z. C. Wang, G. Dawson and W. Chen, J. Mater. Chem., 2011, 21, 14398–14401.
25. K. S. W. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rouquerol and T. Siemieniewska, Pure Appl. Chem., 1985, 57, 603–619.
26. Y. W. Zhang, J. H. Liu, G. Wu and W. Chen, Nanoscale, 2012, 4, 5300–5303.
27. C. Li, C. B. Cao and H. S. Zhu, Mater. Lett., 2004, 58, 1903–1906.
28. C. Mirand, H. Mansilla, J. Yáñez, S. Obregón and G. Colóna, J. Photochem. Photobiol., A, 2013, 253, 16–21.
29. A. Thomas, A. Fischer, F. Goettmann, M. Antonietti, J. Muller, R. Schlogl and J. M. Carlsson, J. Mater. Chem., 2008, 18, 4893–4908.
30. V. N. Khabashesku, J. L. Zimmerman and J. L. Margrave, Chem. Mater., 2000, 12, 3264–3270.
31. C. Li, X. Yang, B. Yang, Y. Yan and Y. Qian, Mater. Chem. Phys., 2007, 103, 427–432.
32. Q. Guo, Y. Xie, X. Wang, S. Lv, T. Hou and X. Liu, Chem. Phys. Lett., 2003, 380, 84–87.
33. B. Chai, T. Peng, J. Mao, K. Li and L. Zan, Phys. Chem. Chem. Phys., 2012, 14, 16745–16752.
34. D. Gao, Q. Xu, J. Zhang, Z. Yang, M. Si, Z. Yan and D. Xue, Nanoscale, 2014, 6, 2577–2581.
35. S. C. Yan, Z. S. Li and Z. G. Zou, Langmuir, 2010, 26, 3894–3901.
36. G. S. Shao, T. Y. Ma, X. J. Zhang, T. Z. Ren and Z. Y. Yuan, J. Mater. Sci., 2009, 44, 6754–6763.
37. X. Zhou, F. Peng and H. Wang, Chem. Commun., 2011, 47, 10323–10325.
38. S. Zhou, Y. Liu, J. M. Li, Y. J. Wang, G. Y. Jiang, Z. Zhao, D. X. Wang, A. J. Duan, J. Liu and Y. C. Wei, Appl. Catal., B, 2014, 158, 20–29.
39. K. M. Parida, S. Pany and B. Naik, Int. J. Hydrogen Energy, 2013, 38, 3545–3553.
40. N. Yang, G. Q. Li, W. L. Wang, X. L. Yang and W. F. Zhang, J. Phys. Chem. Solids, 2011, 72, 1319–1324.
41. S. Martha, A. Nashim and K. M. Parida, J. Mater. Chem. A, 2013, 1, 7816–7824.
42. D. Jiang, L. Chen, J. Zhu, M. Chen, W. Shi and J. Xie, Dalton Trans., 2013, 15726–15734.
43. J. Cao, B. D. Luo, H. L. Lin, B. Y. Xu and S. F. Chen, Appl. Catal., B, 2012, 111–112, 288–296.
44. C. Han, L. Ge, C. Chen, Y. Li, X. Xiao, Y. Zhang and L. Guo, Appl. Catal., B, 2014, 147, 546–553.
45. Y. Hou, F. Zuo, A. Dagg and P. Feng, Angew. Chem., Int. Ed., 2013, 52, 1248–1252.
46. Y. Hou, F. Zuo, A. Dagg and P. Feng, Nano Lett., 2012, 12, 6464–6473.
47. H. Lee and W. Y. Choi, Environ. Sci. Technol., 2002, 36, 3872–3878.
48. X. C. Wang, K. Maeda, A. Thomas, K. Takanabe, G. Xin, J. M. Carlsson, K. Domen and M. Antonietti, Nat. Mater., 2009, 8, 76–80.
49. Y. Zhang, N. Zhang, Z. R. Tang and Y. J. Xu, Chem. Sci., 2012, 3, 2812–2822.
50. W. Li, D. Li, Y. Lin, P. Wang, W. Chen, X. Fu and Y. Shao, J. Phys. Chem. C, 2012, 116, 3552–3560.
51. Y. Zhang, N. Zhang, Z. R. Tang and Y. J. Xu, ACS Nano, 2012, 6, 9777–9789.
52. X. Wang, K. Maeda, X. Chen, K. Takanabe, K. Domen, Y. Hou, X. Fu and M. Antonietti, J. Am. Chem. Soc., 2009, 131, 1680–1681.
53. J. Kim, C. W. Lee and W. Choi, Environ. Sci. Technol., 2010, 44, 6849–6854.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra04710b

---