Plasmonic photocatalyst Au/g-C₃N₄/NiFe₂O₄ nanocomposites for enhanced visible-light-driven photocatalytic hydrogen evolution

Abstract

A Au/g-C₃N₄/NiFe₂O₄ nanocomposite was successfully prepared and characterized, and it exhibited a significant visible-light-driven photoactivity for hydrogen production. The results suggested that NiFe₂O₄ was closely coupled with flake-like g-C₃N₄ and Au nanoparticles were successfully loaded onto the surface of an optimal g-C₃N₄/NiFe₂O₄ nanocomposite. Subsequently, the photocatalytic hydrogen evolution of the as-prepared sample was investigated and optimized, indicating that the optimal Au/g-C₃N₄/NiFe₂O₄ ternary nanocomposite with AuNP loading of 1.0 wt% showed the highest hydrogen generation rate of 1.607 mmol g⁻¹ h⁻¹, which was 30.9 times and 28.7 times higher than the values of pure NiFe₂O₄ and pure g-C₃N₄, respectively, whereas the optimized g-C₃N₄/NiFe₂O₄ binary nanocomposite with NiFe₂O₄ coupled at 49.4% exhibited a hydrogen generation rate of 0.351 mmol g⁻¹ h⁻¹. Interestingly, the photocatalytic hydrogen production mechanism was also tentatively proposed as promoted charge carrier transfer and the strong surface plasmon resonance (SPR) effect of AuNPs via PL, EIS and photocurrent measurements was verified.

---

Introduction

Since the global demand for clean energy has increased continuously, developing alternative strategies for clean energy sustainability has obtained great interest. Hydrogen is regarded as a promising fuel to solve energy and environmental problems all around the world. As one of the promising technologies, photocatalytic hydrogen evolution from water plays a significant role in hydrogen production strategies for hydrogen-based energy systems.^1–4 Nevertheless, highly efficient charge carrier separation is a key element for solar-energy conversion by semiconductor-based photocatalysts,^5,6 there are still many challenges to develop suitable photocatalysts that use visible-light effectively. Typically, in practical applications, lots of used photocatalysts can be quickly and completely recycled from water, as if they are not, they may lead to secondary pollution.⁷ On the other hand, recycling also plays an important role in evaluating the ideal photocatalyst. But this is a challenge for separating the photocatalyst with its good dispersion in the solution. To address this problem, some advanced magnetic semiconductor photocatalyst can effectively overcome the hard separation via a magnetic field.^8–10 Recent investigation of magnetic semiconductor photocatalyst has attracted great attention, such as MFe₂O₄ (M = Cu, Zn, Fe, Co, Ni) result from easy of recycling.^11–17 Among this well-known magnetic semiconductors, nickel ferrite (NiFe₂O₄) has been investigated, one of which is due to its potential electric and magnetic performance, another one is due to its superiority properties of high chemical and structure stability and wider visible region absorption (E_g = ∼2.19 eV).^18,19 Unfortunately, pure nickel ferrite exhibits lower photocatalytic activity owing to fast recombination of photoelectron–hole pairs, even for the Fenton reaction.²⁰ Several strategies have been reported to improve photocatalytic activity of NiFe₂O₄, co-doped with lanthanide elements or combined with other semiconductor to overcome the issue effectively and meaningfully.^21–23 Thus, fabricating NiFe₂O₄-based novel photocatalyst to enhance the photocatalytic efficiency would be an encouraging topic. As we all know, the organic semiconductor of graphitic carbon nitride (g-C₃N₄) with a narrow band gap (E_g = ∼2.7 eV) and the long-range π–π conjugation of g-C₃N₄ endows it with unique performance, such as splitting water for hydrogen evolution,^24,25 degrading organic pollutant in environmental water^26,27 and removing NO in air^28,29 under visible light irradiation. Additionally, it also exhibits the superiority performance of high thermal and chemical stability, photoelectrochemical and optical over other inorganic semiconductor counterparts.^30,31 It suggests that the g-C₃N₄ has a potential prospect in the photocatalytic field. Therefore, the g-C₃N₄/NiFe₂O₄ nanocomposite would be an encouraging suggestion. Indeed, although the g-C₃N₄/NiFe₂O₄ nanocomposites have been reported that are effective in Methylene Blue (MB) degradation,²⁰ the photocatalytic hydrogen evolution of g-C₃N₄/NiFe₂O₄ nanocomposites have not been investigated to date, and only a few documents focused on photocatalytic hydrogen evolution over the NiFe₂O₄-based photocatalysts, so we are interested in this work.

In order to enhance the absorption under visible light over the g-C₃N₄/NiFe₂O₄ nanocomposites, it is well known that Au,³² Ag,³³ Pt,³⁴ Pd,³⁵ and Cu³⁶ of metal nanoparticles can effectively harvest the wide range of sunlight containing UV light and visible light due to its plasmonic effect. The gold nanoparticles (AuNPs) exhibit the surface plasmon resonance (SPR) that could strongly absorb visible light.³⁷ Additionally, the size, shape and surrounding environment of AuNPs have effect on the SPR of AuNPs.³⁸ With this effect, the AuNPs take on electron sinks, leading to fast separation of charge carriers. Thus, the Au/g-C₃N₄/NiFe₂O₄ nanocomposites would be a feasible and efficient photocatalyst for evaluating photocatalytic activity properly under visible light irradiation.

Herein, a new type of the ternary hybrid plasmonic photocatalyst Au/g-C₃N₄/NiFe₂O₄ nanocomposite has been successfully prepared to investigate photocatalytic hydrogen production under visible light irradiation. At first, NiFe₂O₄ semiconductor was prepared by sol–gel method. Then, the g-C₃N₄/NiFe₂O₄ nanocomposites were fabricated by a facile calcination method. And the optimal binary nanocomposite with different mass fraction of AuNPs has been synthesized by photodeposition method. At the same time, the textural, structural, optical properties and photocatalytic performance of nanocomposites were discussed in detail. Especially, the obtained optimal ternary hybrid plasmonic photocatalyst also exhibits highly stability, durability and photocatalytic activity. Finally, the possible mechanism of ternary plasmonic photocatalyst for enhanced photocatalytic efficiency was also investigated based on the experimental results. Furthermore, to the best of our knowledge, no research on the photocatalytic hydrogen evolution of Au/g-C₃N₄/NiFe₂O₄ nanocomposites has been investigated.

Experimental

Materials and reagents

Fe(NO₃)₃·9H₂O (99.9%), Ni(NO₃)₂·6H₂O (99.9%), melamine (99.9%), triethanolamine (99.9%), and methanol (99.8%) was obtained from Aladdin Co. Ltd, and used as received without any further purification. Besides, HAuCl₄·xH₂O (99.9%) was purchased from J&K Chemical Ltd. Deionized water was obtained from Millipore Milli-Q system (18.2 MΩ cm resistivity), which was used to prepare all aqueous solution throughout all experimental.

The preparation of pure g-C₃N₄

The pure g-C₃N₄ was synthesized by calcination melamine at 550 °C for 4 h with the heating rate of 10 °C min⁻¹ directly.

The synthesis of pure NiFe₂O₄

The pure NiFe₂O₄ was prepared by a sol–gel method. Typically, 1 mole of Ni(NO₃)₂·6H₂O and 2 mole of Fe(NO₃)₃·9H₂O were dissolved in 100 mL deionized water (solution A). 3 mole of citric acid was dissolved in 100 mL deionized water (solution B) and the metal (Ni²⁺ + Fe³⁺)/citric acid molar ratio was 1. Then solution A was dropwise added into solution B under vigorous magnetic stirring. The mixed aqueous solution reacted for 1 h at 60 °C and dried at 90 °C, the obtained powder calcination at a temperature of 400 °C for 4 h with a heating rate of 10 °C min⁻¹.

The fabrication of g-C₃N₄/NiFe₂O₄ nanocomposites

The preparation of sample with different mass fraction of g-C₃N₄/NiFe₂O₄ nanocomposites were as follows via a facile calcination method: 0.1 g NiFe₂O₄ and different mass amount of g-C₃N₄ mixed homogeneous by grinding an agate mortar at least 30 min, then the resulting powder calcination at a temperature of 400 °C for 4 h with a heating rate of 10 °C min⁻¹. According to the TG analysis, the different mass fraction of NiFe₂O₄ in NiFe₂O₄/g-C₃N₄ was labeled as 6.2% NiFe₂O₄/g-C₃N₄ (NFCN), 14.7% NiFe₂O₄/g-C₃N₄ (2NFCN), 36.1% NiFe₂O₄/g-C₃N₄ (3NFCN), 49.4% NiFe₂O₄/g-C₃N₄ (4NFCN), 72.5% NiFe₂O₄/g-C₃N₄ (5NFCN), 75.6% NiFe₂O₄/g-C₃N₄ (6NFCN) and 91.1% NiFe₂O₄/g-C₃N₄ (7NFCN).

The preparation of Au/g-C₃N₄/NiFe₂O₄ nanocomposites

The different mass fraction of Au loaded on optimal binary nanocomposite (4NFCN) was prepared via photodeposition method as follows: for 1.0 wt% Au/4NFCN, 1 g of as-prepared 4NFCN nanocomposites was dispersed in 54 mL of HAuCl₄ (0.545 mM) aqueous solution, and then followed by adding 6 mL of methanol. After evacuating the air completely, the mixture was irradiated by a Xe lamp (300 W) for 3 h at 25 °C to reduce the Au adequately, aged overnight.⁷⁵ The obtained samples was separated through centrifugation and washed with deionized water, and then vacuum dried at 80 °C for 12 h. The other different mass fraction of AuNPs loaded on optimal g-C₃N₄/NiFe₂O₄ nanocomposite was synthesized by the same methods through only the different quantity of the HAuCl₄ solution, and the 1.0 wt% Au/NiFe₂O₄ and 1.0 wt% Au/g-C₃N₄ nanocomposites were also prepared by photodeposition methods.

Characterization

The crystallinity of as-synthesized samples were obtained by Bruker D8 Advance X-ray diffractometer (XRD, Bruker, Germany) with Cu Kα radiation (λ = 0.15406 nm). The X-ray diffraction patterns were carried out over 2θ angle range from 10° to 80° with a step size of 0.01° and a scanning rate of 8° min⁻¹. FT-IR spectra were performed with Bruker Tensor 27 spectrometer at the frequency range of 4000–500 cm⁻¹ with a resolution of 4 cm⁻¹, KBr was used as diluents. The transition electron microscope (TEM) spectrums were obtained with high-resolution transmission electron microscopy at different scales (HRTEM; TecnaiG220, FEI). X-ray photoelectron spectroscopy (XPS, Axis Ultra DLD, Kratos) was carried out with a monochromatic X-ray source (Al Kα, 15 kV, 200 W). The Brunauer–Emmett–Teller (BET) specific surface area of samples was measured by Quantachrome NOVA 1200e. UV-visible (UV-vis) spectra were performed by a UV-visible spectrophotometer (U-3010, Hitachi) and converted from reflection to absorbance by the Kubelka–Munk method, using BaSO₄ acted as a reflectance standard. Photoluminescence (PL) spectrum was obtained with a fluorescence spectrophotometer (Hitachi, F-4500) at room temperature and the excitation wavelength was 367 nm. The surface morphology and the composition of as-prepared samples were measured by scanning electron microscopy (SEM, JSM-6510A) and energy dispersive X-ray spectroscopy (EDS, JEM-2100), respectively.

Photocurrents measurements were performed on an Auto LAB electrochemical analyzer (Metrohm Co.) at a standard three-electrode system. The prepared electrode, a Pt wire and Hg/Hg₂Cl₂ were used as the working electrode, the counter electrode, and the reference electrode, respectively. A 300 W Xe-lamp via a UV cut off filter (λ > 420 nm) served as light source. A Na₂SO₄ (1.0 M) aqueous solution was used as a supporting electrolyte. A working electrode was prepared based on the reported documents.^39,40 Electrochemical impedances spectroscopy (EIS) was examined by applying an AC voltage of 10 mV amplitude at the frequency range of 0.1 Hz to 100 kHz with a CHI 660D electrochemical analyzer also in a three-electrode system with a KOH (2.0 M) aqueous solution. The prepared electrode, a Pt wire and Hg/HgO was used as the working electrode, the counter electrode, and the reference electrode, respectively.

Photocatalytic hydrogen evolution measurements

Photocatalytic procedures for hydrogen production were examined in a Pyrex reaction cell attached to a closed gas circulation and evacuation system (see Fig. S1†). In a typical liquid-phase photocatalytic experiment, 0.01 g of the target photocatalyst in 70 mL aqueous solution containing 10% of triethanolamine (TEOA) solution for hydrogen production. A high-pressure Xe lamp (300 W) via a UV cut off filter (λ > 420 nm) was chosen as light source to afford the illumination and trigger photocatalytic reactions, which was placed on the side of photoreactor. Prior to irradiation, the suspension was deaerated by evacuation to remove CO₂ and O₂ dissolved in water. Hydrogen evolution was analyzed by an online gas chromatograph (GC7900, TCD, molecular sieve 5 Å, N₂ carrier, Tian Mei, Shanghai). The as-synthesized samples were suspended in the aqueous solution with a continuously magnetic stirrer during the irradiation. The temperature of mixture was maintained at 25 °C during photocatalytic procedures.

Results and discussion

Structure and morphology of as-prepared samples

To investigate the thermal stability of the as-prepared binary nanocomposites and the actual mass ratios of NiFe₂O₄ and g-C₃N₄ in the nanocomposite, TG analysis has been performed range from 25 °C to 800 °C with a heating rate of 20 °C min⁻¹ under air conditions, and the results as shown in Fig. 1. The results suggest that the weight of pure NiFe₂O₄ presents the almost no obvious change and exhibits a good thermal stability. It has been reported in the previous work that the pure g-C₃N₄ had two weight loss regions. The first one is from 100 °C to 400 °C, the reason for this is that may be due to the absorbed H₂O and air. Another weight loss region is from 550 °C to 750 °C, which originated from the thermal decomposition of the g-C₃N₄.⁷⁰ For the g-C₃N₄/NiFe₂O₄ nanocomposites, a little weight loss generating from 100 °C to 400 °C might be adsorption of surface bound water. It can be also seen that the g-C₃N₄ contents in the binary nanocomposites become unstable while the heat temperature is above 450 °C, and the weight of the g-C₃N₄/NiFe₂O₄ nanocomposites decreased rapidly from 450 °C to 600 °C, suggesting that the thermal decomposition of the g-C₃N₄ happened in this temperature range. The decomposition temperature of g-C₃N₄ in the nanocomposites is lower than pure g-C₃N₄, which might be result from the introducing NiFe₂O₄ result in the relatively low temperature.

Fig. 1 TG thermograms for heating pure NiFe₂O₄ and g-C₃N₄/NiFe₂O₄ nanocomposites from room temperature to 800 °C at the heating rate of 20 °C min⁻¹.

Similar results have been documented in other nanocomposite.^71,72 Furthermore, the contents of NiFe₂O₄ in the nanocomposites would be calculated after heating samples, as depicted in the Fig. 1, and the NiFe₂O₄ contents in NiFe₂O₄/g-C₃N₄ nanocomposites were estimated to be 6.2%, 14.7%, 36.1%, 49.4%, 72.5%, 75.6% and 91.1% for NFCN, 2NFCN, 3NFCN, 4NFCN, 5NFCN, 6NFCN and 7NFCN, respectively.

The structural identify and phase composition of the series nanocomposites were investigated by XRD, and the results as show in Fig. 2 and S2.† Fig. 2 shows the XRD patterns of representative as-prepared samples. The XRD peaks at 18.4°, 30.3°, 35.7°, 37.3°, 43.4°, 53.8°, 57.4°, 62.9°, and 74.6° were ascribed to the (111), (220), (311), (222), (400), (422), (511), (440) and (533) planes of NiFe₂O₄ (JCPDS no. 10-0325). Additionally, the pure g-C₃N₄ exhibits the two diffraction peaks at 13.1° and 27.4° can be correspond to the (002) and (100) crystal planes, which was ascribed to the characteristic crystal phase of inter-layer structural packing and aromatic systems.^41,42 As can be also seen in Fig. 2, the 4NFCN exhibits that materials consist of both NiFe₂O₄ and g-C₃N₄ phases. Compared with 4NFCN, the 1.0 wt% Au/4NFCN exhibits the additional diffraction peaks of 77.6°, 64.6°, 44.4° and 38.2°, corresponding to the (311), (220), (200), and (111) crystal planes of Au (JCPDS no. 99-0056), respectively. It would be also found that the 1.0 wt% Au/NiFe₂O₄ and 1.0 wt% Au/g-C₃N₄ exhibit additional four diffraction peaks after Au loaded in comparison with NiFe₂O₄ and g-C₃N₄ respectively, indicating that the AuNPs is successfully formed. The XRD patterns of other g-C₃N₄/NiFe₂O₄ and Au/g-C₃N₄/NiFe₂O₄ nanocomposites are shown in Fig. S2.† The results suggest that crystal phase composition of the other g-C₃N₄/NiFe₂O₄ nanocomposites and Au/g-C₃N₄/NiFe₂O₄ nanocomposites are similar to the 4NFCN and 1.0 wt% Au/4NFCN, respectively. With the exception of the NFCN and 2NFCN, the XRD peaks intensity of g-C₃N₄ reduced significantly, the reason for this is that may be due to the low concentration of g-C₃N₄. Then, the representative samples of FT-IR spectra is also investigated, and the results as shown in Fig. 3. It can be noted that the strong stretching vibration in the 1200–1650 cm⁻¹ region of pure g-C₃N₄, g-C₃N₄/NiFe₂O₄ nanocomposites, 1.0 wt% Au/g-C₃N₄ and 1.0 wt% Au/4NFCN are ascribed to C–N and C=N stretching vibration, whereas the peak at 808 cm⁻¹ is assigned to the s-triazine units.^43,44 Besides, the strong absorption peaks at 417 cm⁻¹ and 608 cm⁻¹ of pure NiFe₂O₄, 1.0 wt% Au/NiFe₂O₄, 4NFCN nanocomposites and 1.0 wt% Au/4NFCN can be attributed to stretching vibrations of metal–O bonds in octahedral and Fe–O bonds in tetrahedral positions respectively.²⁰ Fig. S3† also shows the other FT-IR spectra of g-C₃N₄/NiFe₂O₄ nanocomposites, the stretching vibrations are similar to the 4NFCN.

Fig. 2 XRD patterns of the representative as-prepared samples: pure g-C₃N₄, pure NiFe₂O₄, 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄, 4NFCN and 1.0 wt% Au/4NFCN.

Fig. 3 FT-IR spectra of (a) g-C₃N₄, (b) NiFe₂O₄, (c) 1.0 wt% Au/g-C₃N₄, (d) 1.0 wt% Au/NiFe₂O₄, (e) 4NFCN, (f) 1.0 wt% Au/4NFCN.

The morphologies of representative as-prepared samples were examined by SEM. Fig. 4 shows the SEM spectra of pure NiFe₂O₄, pure g-C₃N₄, optimized g-C₃N₄/NiFe₂O₄ and optimal Au/g-C₃N₄/NiFe₂O₄ nanocomposites. It can be found that pure NiFe₂O₄ (Fig. 4a) exhibits the amorphous granule-like particles, while the pure g-C₃N₄ (Fig. 4b) shows the flake-like. It can be observed that Fig. 4c exhibits the amorphous granule-like particles of NiFe₂O₄ coupled with flake-like g-C₃N₄ successfully. Finally, the 1.0 wt% Au/4NFCN nanocomposite (Fig. 4d) reveals the presence of flake-like g-C₃N₄, as well as AuNPs and granule-like particles of NiFe₂O₄. Additionally, the chemical composition of the optimal nanocomposite was measured by EDS. Fig. 4e and Table S1† shows the elements of 1.0 wt% Au/4NFCN, exhibiting the nanocomposite consists of C, N, O, Fe, Ni and Au elements. The element mapping of 1.0 wt% Au/4NFCN was also recorded, and the results as shown in Fig. S4,† suggesting that all elements were uniformly distributed over the nanocomposite. The detailed shape, morphology and the well-defined heterojunction formation among the g-C₃N₄, NiFe₂O₄ and AuNPs phase in the 1.0 wt% Au/4NFCN hybrid nanocomposite was also investigated using TEM, and the results as shown in Fig. 5. It can be seen that pure NiFe₂O₄ exhibits the amorphous granule-like nanoscale particles (Fig. 5a). Fig. 5b shows that NiFe₂O₄ is closely coupled with g-C₃N₄, and the AuNPs is successfully loaded on the surface of g-C₃N₄/NiFe₂O₄ nanocomposites. Fig. 5c–e shows the high-resolution TEM images of NiFe₂O₄ with the lattice planes of 0.343 nm, 0.393 nm and 0.312 nm for the (440), (533) and (511) plane, respectively. It also exhibits the HR-TEM spectra of AuNPs with the lattice planes of 0.236 nm for the (111) plane, which is consistent with the previous documented.^45,46 In addition, the Fig. 5f shows the size distribution histogram of Au nanoparticles, suggesting an average size of 4.9 nm.

Fig. 4 SEM images of pure NiFe₂O₄ (a), pure g-C₃N₄ (b), 4NFCN (c) and 1.0 wt% Au/4NFCN (d) respectively; EDS of 1.0 wt% Au/4NFCN (e).

Fig. 5 TEM images of pure NiFe₂O₄ (a) and 1.0 wt% Au/4NFCN (b–e); the size distribution histogram with Gaussian-fitting curve of AuNPs in the 1.0 wt% Au/4NFCN (f).

Fig. 6 is the XPS spectra of the optimal ternary nanocomposite 1.0 wt% Au/4NFCN. The survey spectrum is shown in the Fig. S5,† confirming that the 1.0 wt% Au/4NFCN surface also consists of C, N, O, Fe, Ni and Au elements. Fig. 6a–f shows C 1s, N 1s, O 1s, Fe 2p, Au 4f, and Ni 2p XPS spectra respectively. Fig. 6a shows that the core-level XPS spectra of C 1s exhibit two peaks. The strong peaks at 284.6 eV and 288.1 eV are ascribed to the sp²-hybridized adventitious carbon and C–(N)₃ binding energy respectively.⁴⁷ Fig. 6b exhibits three peaks with 398.6, 400.2, 404.3 eV, respectively. The peak located at 398.6 eV is attributed to C=N–C and other two peaks at 400.2 eV and 404.3 eV are ascribed to the N–(C)₃ and N–H bonding,⁴⁶ respectively. Two different peaks are found for the core-level O 1s spectra (Fig. 6c), the peak at 298.4 eV owing to O²⁻ ions in the NiFe₂O₄ phase. Another peak at 531.9 eV is attributed to the presence of a water molecule or a hydroxyl group on the surface of 1.0 wt% Au/4NFCN nanocomposite.⁴⁸ The Fe 2p XPS spectrum is deconvoluted into four peaks (Fig. 6d), the peak at 709.7 eV and 723.7 eV is ascribed to the banding energy of Fe 2p_3/2 and Fe 2p_1/2. The peak at 711.8 eV confirms that Fe³⁺ ions binded with hydroxyl groups in the hematite phase, whereas the peak at 716.8 eV may be attributed to the satellite peaks.³² The banding energies in Au 4f as shown in Fig. 6e, exhibiting two major characteristic peaks at 83.1 eV and 86.8 eV, corresponding to Au 4f_7/2 and Au 4f_5/2 band respectively, which demonstrates the formation of Au⁰ valence state.⁴⁹ The XPS spectrum of Ni 2p is consistent with in previous reported document.²¹ Fig. 6f shows the banding energy values of Ni 2p_3/2 and Ni 2p_1/2 is observed at 854.3 eV and 871.9 eV, respectively, which suggests the typical banding energy of Ni²⁺. The other two peaks at banding energies of 860.8 eV and 878.9 eV is related to the satellite peaks.^20,21

Fig. 6 XPS spectra of (a) C 1s, (b) N 1s, (c) O 1s, (d) Fe 2p, (e) Au 4f, (f) Ni 2p in 1.0 wt% Au/4NFCN.

The UV-vis diffuse reflection spectra of pure NiFe₂O₄, pure g-C₃N₄, 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄, optimal g-C₃N₄/NiFe₂O₄ (4NFCN) and optimized Au/g-C₃N₄/NiFe₂O₄ (1.0 wt% Au/4NFCN) are shown in Fig. 7a, and the band gap value of that are estimated by Kubelka–Munk equation respectively in Fig. S6.^44,46† It reveals that a sharp absorption of edge rises of g-C₃N₄ at 465 nm, owing to the band gap of 2.45 eV, whereas NiFe₂O₄ has a great stronger and wider absorption in the visible light region, and the band gap is 1.73 eV. Also, it can be noted that absorption edge of 1.0 wt% Au/g-C₃N₄, optimal g-C₃N₄/NiFe₂O₄ and optimized Au/g-C₃N₄/NiFe₂O₄ nanocomposite shows red shift compared to pure g-C₃N₄. And optimal g-C₃N₄/NiFe₂O₄ and optimized Au/g-C₃N₄/NiFe₂O₄ nanocomposites also has a much wider and stronger absorption in the visible region. The optical absorption of 1.0 wt% Au/g-C₃N₄ occurs at 475 nm with the band gap of 2.17 eV. For the optimal g-C₃N₄/NiFe₂O₄, the main absorption edge at 515 nm, exhibiting the band gap of 1.67 eV. Meanwhile, it can be also observed that the optimized Au/g-C₃N₄/NiFe₂O₄ nanocomposite shows a little blue shift in comparison with optimal g-C₃N₄/NiFe₂O₄ nanocomposite, the fundamental absorption edge of optimized Au/g-C₃N₄/NiFe₂O₄ nanocomposite begin at 485 nm with a band gap of 1.96 eV. Interestingly, it can be seen that characteristic strong absorption peak at around 745 nm in pure NiFe₂O₄, 1.0 wt% Au/NiFe₂O₄, optimal g-C₃N₄/NiFe₂O₄ and optimized Au/g-C₃N₄/NiFe₂O₄. These cases are similar to the previous document about SrTiO₃/NiFe₂O₄,⁵⁰ NiFe₂O₄/TiO₂ (ref. 51) and TiO₂/SiO₂/NiFe₂O₄ (ref. 52) nanocomposites. Besides, 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄ and 1.0 wt% Au/g-C₃N₄/NiFe₂O₄ exhibit 550 nm characteristic absorption peak of SPR of AuNPs,^32,46 further confirming the presence of gold in the form of AuNPs. Fig. S7† shows the UV-vis image of other g-C₃N₄/NiFe₂O₄ nanocomposites and Au/g-C₃N₄/NiFe₂O₄ nanocomposites. The results in Fig. S7a† indicate that the absorption intensity of g-C₃N₄/NiFe₂O₄ nanocomposite in the visible region increased with increase of NiFe₂O₄ content due to the NiFe₂O₄ with the wider visible light region. Fig. S7b† exhibits that the absorption intensity increased gradually with increase of AuNPs mass fraction from 0.5 wt%, 0.7 wt% to 1.0 wt%, indicating that the greater number of AuNPs would modify the surface to enhance the photogenerated electron–hole pairs separated, so the photocatalytic activity is gradually enhanced. But, the absorption of 1.2 wt% Au/4NFCN and 1.5 wt% Au/4NFCN is lower than 1.0 wt% Au/4NFCN, the reasons for this is that may be caused by aggregation of AuNPs on the photocatalyst surface. Fig. 7b shows the Raman spectra of the pure NiFe₂O₄, pure g-C₃N₄, 4NFCN and 1.0 wt% Au/4NFCN sample. The Raman spectra of pure g-C₃N₄ are assigned to the vibration of triazine rings. The broad peak at about 557 cm⁻¹ is attributed to in-plane symmetrical stretching vibrations of heptazine heterocycles, while the band from ∼(1000–2000) cm⁻¹ is ascribed to the amorphous carbon.⁵³ The broad peak at 694 cm⁻¹, 568 cm⁻¹, 484 cm⁻¹, and 329 cm⁻¹ attributes to A_1g, F_2g(3), F_2g(2), and E_g vibration modes respectively in pure NiFe₂O₄, whereas the 1328 cm⁻¹ is ascribed to the overtones A_1g and F_2g(3) vibrations.⁵⁴ The E_g mode attributes to the symmetric bending of O with respect to Fe, and the F_2g(2) mode is originated from asymmetric stretching of O and Fe, whereas F_2g(3) mode is belonged to asymmetric bending of O. The A_1g mode is ascribed to Fe–O bonds in tetrahedral along the symmetric bending of O atoms coordination. Interestingly, these peak intensities of 1.0 wt% Au/4NFCN are greatly enhanced in comparison with that of NiFe₂O₄. The enhancement of Raman peak intensities is related to the SERS effect previously documented.^55–57 The strongly Raman enhancement in the photocatalysts might be caused by SPR effect of AuNPs. It further reveals the formation of AuNPs on the surfaces of g-C₃N₄/NiFe₂O₄ nanocomposites, which would significantly have effect electronic structure on the plasmonic photocatalytic systems.

Fig. 7 The UV-vis images (a) of pure NiFe₂O₄, pure g-C₃N₄, 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄, 4NFCN, 1.0 wt% Au/4NFCN, respectively; the Raman patterns (b) of pure NiFe₂O₄, pure g-C₃N₄, 4NFCN and 1.0 wt% Au/4NFCN.

Magnetic properties

While this photocatalyst was fabricated, its magnetic properties made us considerable concern from beginning to end. So the hysteresis loops of pure NiFe₂O₄, 4NFCN and 1.0 wt% Au/4NFCN nanocomposites are shown in Fig. 8. It can be noted that the saturation magnetization M_s of pure NiFe₂O₄, 4NFCN and 1.0 wt% Au/4NFCN is about 38.1, 19.9 and 18.9 emu g⁻¹, respectively. The results indicate that the 4NFCN with the high-content nonmagnetic g-C₃N₄ has great influence on nanocomposite magnetism, whereas the saturation magnetization M_s of 1.0 wt% Au/4NFCN slightly lower than 4NFCN. Even so, it is strong adequate to collect and recycle the photocatalyst from the solution by an external magnetic filed, as shown in the inset of Fig. 8. The results suggest that the 1.0 wt% Au/4NFCN is an excellent magnetic photocatalysts, which will play an important role in recycling of photocatalysts.

Fig. 8 The hysteresis loops of pure NiFe₂O₄, 4NFCN and 1.0 wt% Au/4NFCN nanocomposite (the inset graph shows the sample dispersed in water (left) and separated by external magnet (right)).

Photocatalytic activity of as-prepared samples

Photocatalytic hydrogen evolution activities of all the as-prepared samples were evaluated under visible light irradiation (λ > 420 nm) using triethanolamine (TEOA) as sacrificial agent, and the results were depicted in Fig. 9, S8 and Table S1.† To confirm the hydrogen evolution, a blank experiment was studied under visible light irradiation in a condition without photocatalyst. It can be seen that no hydrogen gas was produced until the presence of powder photocatalyst to aqueous TEOA solution. Fig. S8† exhibited the GC chromatogram of hydrogen production by representative photocatalysts, indicating that the photocatalytic activity of optimal ternary photocatalysts is higher than optimal binary nanocomposite or pristine one. Fig. 9a showed the hydrogen production rate of g-C₃N₄/NiFe₂O₄ nanocomposites with loading of different amounts of NiFe₂O₄. Pure NiFe₂O₄ and pure g-C₃N₄ were used for comparison. Significantly, it was noted that hydrogen production rate of pure NiFe₂O₄ and pure g-C₃N₄ were the lowest them, their photocatalytic hydrogen production rate are 0.052 mmol g⁻¹ h⁻¹ and 0.056 mmol g⁻¹ h⁻¹, respectively. And the 4NFCN sample with NiFe₂O₄ loading 49.4% showed the highest hydrogen evolution rate of 0.351 mmol g⁻¹ h⁻¹ under the optimal conditions, which was about 4.9 times higher than that of commercial P25 powder materials (0.072 mmol g⁻¹ h⁻¹).³⁸ The results suggested that g-C₃N₄/NiFe₂O₄ heterostructure generated highly efficiency compared to pure NiFe₂O₄ and g-C₃N₄ owing to the interaction of heterostructure can promote the division and prohibit the recombination of photogenerated electron–hole pairs. In addition, in order to further enhanced photocatalysis, photocatalytic hydrogen evolution rate of the different mass fraction of AuNPs-loaded on optimal g-C₃N₄/NiFe₂O₄ nanocomposites (4NFCN) was also studied, and the results as shown in Fig. 9b. It can be found that the photocatalytic hydrogen evolution rate of Au/g-C₃N₄/NiFe₂O₄ nanocomposite increased with increase of AuNPs concentration, reached an optimum amount at 1.0 wt% and then decreased with further increase of AuNPs content. The optimal Au/g-C₃N₄/NiFe₂O₄ nanocomposite (1.0 wt% Au/4NFCN) evolved the highest hydrogen rate was 1.607 mmol g⁻¹ h⁻¹. It revealed that the photocatalytic hydrogen production rate of optimal ternary nanocomposites was also much higher than the 1.0 wt% Au/g-C₃N₄ (0.553 mmol g⁻¹ h⁻¹) and 1.0 wt% Au/NiFe₂O₄ (0.135 mmol g⁻¹ h⁻¹) binary nanocomposites obviously. The results indicated that AuNPs would further promote the transfer of photoelectrons in the heterostructure and the strongly plasmonic effect of AuNPs. While further increases of the AuNPs content results in photocatalytic activity decreased, this may be due to enhanced absorbance and scattering of photons by excess AuNPs during the photocatalytic procedure.^58,59 On the other hand, apparent quantum efficiency (AQY) of NiFe₂O₄, g-C₃N₄, 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄, 4NFCN and 1.0 wt% Au/4NFCN at 420 nm was also measured, the corresponding AQY value on sample at 420 nm has been calculated to be 0.02%, 0.02%, 0.04%, 0.17%, 0.14% and 1.12%, respectively. It indicates that the 1.0 wt% Au/4NFCN has a highest AQY, which is also much higher than that on commercial P25 and g-C₃N₄ reported in the literature (for calculation details of the AQY see ESI†).^39,74

Fig. 9 Photocatalytic activity of the as-synthesized samples (0.01 g) in a suspension aqueous solution (70 mL) containing 10% (v/v) triethanolamine under visible light irradiation (λ > 420 nm): (a) hydrogen generation rate of pure NiFe₂O₄, pure g-C₃N₄ and a series of NiFe₂O₄/g-C₃N₄ nanocomposites; (b) hydrogen generation rate 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄, optimal binary nanocomposite (4NFCN) and the different mass fraction of AuNPs loaded on optimal binary nanocomposite (irradiation for 3 h).

Additionally, the BET specific surface area (S_BET) of the representative as-prepared sample was also investigated using adsorption–desorption measurements, and the results as shown in Table 1. The surface areas of pure NiFe₂O₄, pure g-C₃N₄, 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄, 4NFCN and 1.0 wt% Au/4NFCN nanocomposite were found to be 31.0, 28.2, 22.2, 20.6, 53.2 and 32.7 m² g⁻¹, respectively. The results indicated that the photocatalytic efficiency of 4NFCN was dramatically enhanced with the highest surface areas compared to pure NiFe₂O₄ and pure g-C₃N₄, but the 1.0 wt% Au/NiFe₂O₄ and 1.0 wt% Au/g-C₃N₄ with the higher photocatalytic activity as pure NiFe₂O₄ and pure g-C₃N₄ with higher surface area had a lower activity. Besides, the photocatalytic activity of optimal ternary nanocomposite (1.0 wt% Au/4NFCN) was greatly increased with the lower surface area in comparison with optimal g-C₃N₄/NiFe₂O₄ nanocomposite. It revealed that the surface areas did not play a significantly role in enhancing photocatalytic activity, and the appropriate pore size and surface area of Au/g-C₃N₄/NiFe₂O₄ nanocomposites showed highly photoactivity for hydrogen production. The reason for this was that may be generation of heterostructure that dramatically separated charge carriers and prolonged the lifetime of charge carriers.⁶⁰ Moreover, it might be attributed to the SPR of noble metal AuNPs exhibited strongly visible light absorption and the absorption edge of optimal ternary nanocomposite showed a red shift to the visible light region result in strongly absorbing visible light.³²

Table 1 The hydrogen evolution rate, S_BET, pore volume and average pore size of as-prepared samples

Sample	SBET (m^2 g^−1)	Pore volume (cm^3 g^−1)	Average pore size (nm)	H2 (mmol g^−1 h^−1)
Pure NiFe2O4	31.0	0.11	13.5	0.052
Pure g-C3N4	28.2	0.12	17.0	0.056
1.0 wt% Au/NiFe2O4	22.2	0.08	15.2	0.135
1.0 wt% Au/g-C3N4	20.6	0.07	13.5	0.553
4NFCN	53.2	0.16	11.3	0.351
1.0 wt% Au/4NFCN	32.7	0.09	11.4	1.607

---

Furthermore, to evaluate an ideal photocatalysts, the stability and recyclability of optimal nanocomposite would be another consideration and it is of great importance. Hence, the repeated experimental of photocatalytic hydrogen production using TEOA as sacrificial agent over optimal nanocomposite (1.0 wt% Au/4NFCN) was carried out in six-run cycles (18 h), and the results as shown in Fig. 10. For each run, the optimal nanocomposite was dried and then irradiated by a 300 W Xe lamp for 3 h. Significantly, it can be noted that no inactivation of photocatalyst occurs on repeated tests, which indicated that nanocomposite exhibits higher stability and reusability. It can be demonstrated that optimal nanocomposite was not photocorroded and suitable for the photocatalytic hydrogen evolution in TEOA suspension aqueous solution. In addition, it's possible that some changes may occur after repeated tests. Hence, a comparison XRD pattern of 1.0 wt% Au/4NFCN before and after cycling test was shown in Fig. S9.† However, no obvious change was seen in the XRD patterns. Although we still try our best to solve the issue, cannot explain it now. Fortunately, it has not influence on our discovering that the as-synthesized 1.0 wt% Au/4NFCN can be regarded as a stable photocatalyst.

Fig. 10 Cycling runs in photocatalytic hydrogen production in the presence of 1.0 wt% Au/4NFCN (0.05 g) nanocomposite under visible light irradiation.

Proposed photocatalytic hydrogen evolution mechanism

In the present work, the 1.0 wt% Au/4NFCN nanocomposite shows obvious photocatalytic activity in comparison with that of pure NiFe₂O₄, pure g-C₃N₄, 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄ and optimal binary nanocomposite (4NFCN). Although increased specific surface area is generally beneficial for the efficiency of photocatalysts, BET measurement has eliminated this possibility in the case of Au/g-C₃N₄/NiFe₂O₄ nanocomposites. As indicated in the previous document,^32,73 the possible mechanism of semiconductor–semiconductor–metal heterostructure has been reported. Although the photocatalytic reaction or photocatalysts are different, the principle of the high photoactivity is supposed to be same. The enhancement in performance suggests that NiFe₂O₄ and g-C₃N₄, in combination with AuNPs, have a synergistic effect owing to efficient separation of photoelectron–hole pairs⁶¹ and highly visible light absorbance from strongly surface plasmon resonance (SPR) induced by AuNPs. Moreover, the well-defined interface between AuNPs and semiconductor after the formation of heterostructure can also play a significant role in the improvement of photocatalytic activities. When the visible light irradiation on the surface of semiconductor materials, consequently valence band (VB) photoholes and conduction band (CB) photoelectrons are generated leading to the formation of a Mott–Schottky junction between AuNPs–semiconductor interfaces.³⁸ Newly formed the Fermi level of AuNPs–semiconductor is lower than the bottom of the conduction band of semiconductor,⁶² the AuNPs take on electron sinks, leading to photoelectron–hole pairs further separated and AuNPs with a more negative potential. Additionally, the charge carrier is further transferred results in Fermi energy level equilibration between the interfaces of AuNPs–semiconductor.⁴⁹ Thereby, the high electronegativity and Fermi energy level equilibration may lead to the Fermi energy level shifting closer to the conduction band of the polar semiconductor, which results in promoting the separation of photoelectron–hole pairs.⁶³ On the other hand, while AuNPs strongly absorb visible light, the electrons near the Fermi energy level are excited to the surface plasmon resonance state (generally above the conduction band minimum of the semiconductor),^64,65 migrated to the conduction band of semiconductor and then participate in the photocatalytic processes.

Therefore, the improved photocatalytic activity over Au/g-C₃N₄/NiFe₂O₄ may be owing to the separation of photoelectron–hole pairs and strongly SPR of AuNPs. Based on the above analysis, we proposed a possible photocatalytic hydrogen production mechanism for the Au/g-C₃N₄/NiFe₂O₄ hybrid plasmonic photocatalysts, and the schematic diagram is listed in Fig. 11. In order to verify the position of VB edges of NiFe₂O₄ and g-C₃N₄, the VB XPS of NiFe₂O₄ and g-C₃N₄ is also measured, and the results as shown in Fig. S10.† The results indicate that the VB edges of NiFe₂O₄ and g-C₃N₄ are determined to be 1.02 eV and 1.88 eV, respectively. Thereby, the CB edges of NiFe₂O₄ and g-C₃N₄ are determined to be −0.71 eV and −0.57 eV, respectively. Hence, the conduction band of the NiFe₂O₄ is higher than the conduction band of g-C₃N₄, so the photogenerated electrons on the conduction band of NiFe₂O₄ can be easily migrated to the conduction band of g-C₃N₄. Under the visible light irradiation, photoelectrons are excited from valence band to conduction band in NiFe₂O₄ as well as g-C₃N₄, leaving behind photoholes in the valence band of both semiconductors. The photogenerated electrons of NiFe₂O₄ in conduction band would easily migrate to the conduction band of g-C₃N₄ and then are trapped immediately by the AuNPs. The accumulated photoelectrons trapped by AuNPs participate in photocatalytic hydrogen production process. Meanwhile, the photoholes in the valence band of g-C₃N₄ would migrate to the valence band of NiFe₂O₄. Besides, the AuNPs might produce additional charge carrier owing to the surface plasmonic resonance effect. Consequently, this process would significantly accelerate the separation of photoelectron–hole pairs and effectively minimize possibility of charge carrier recombination, resulting in the highest photocatalytic activity of Au/g-C₃N₄/NiFe₂O₄ nanocomposites.

Fig. 11 Proposed mechanisms of the photocatalytic process for the hydrogen production over the visible-light-driven Au/g-C₃N₄/NiFe₂O₄ photocatalysts.

A series measurement was measured to investigate the charge carrier transformation and separation efficiency of Au/g-C₃N₄/NiFe₂O₄ nanocomposites. PL technique can be considered as an important approach to understand charge carrier separation efficiency in a photocatalyst.^66,67 So, the PL spectra of as-prepared samples have been characterized, and the results as shown in Fig. 12 and S11.† The results in Fig. 12 suggest that NiFe₂O₄ exhibits weak emission peaks range from 380 nm to 550 nm, which might be originated from excitonic PL and attributed to defects of NiFe₂O₄ and surface oxygen vacancies.⁶⁸ It can be noted that pure g-C₃N₄ exhibits the strong emission weak at around 440 nm.⁴⁶ As can be seen the PL spectra of other nanocomposites, the PL emission peak intensity is in order of 1.0 wt% Au/4NFCN < 4NFCN < 1.0 wt% Au/NiFe₂O₄ < 1.0 wt% Au/g-C₃N₄ < g-C₃N₄, indicating the charge carrier transfer between AuNPs and NiFe₂O₄/g-C₃N₄ nanocomposites would suppress the recombination of photoelectron–hole pairs effectively, which agrees the above-mentioned mechanism. The reason is why 1.0 wt% Au/4NFCN has the highest photocatalytic efficiency for hydrogen evolution.

Fig. 12 PL spectra of pure NiFe₂O₄, pure g-C₃N₄, 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄, 4NFCN and 1.0 wt% Au/4NFCN excited at 367 nm, respectively.

The charge carrier transfer in Au/NiFe₂O₄/g-C₃N₄ nanocomposites was also demonstrated by EIS and photocurrent responses measurements. The results in Fig. 13a show that the 1.0 wt% Au/4NFCN has the smallest arc size. Considering the high frequency alternating-electric filed can cause the molecular polarization, the charge transfer resistance (R_ct) in the electrolyte from the photoelectrodes to redox species would be calculated through fitting the semi-arc in the low frequency region. Generally, a smaller radius implies that the smaller charge-transfer resistance in an EIS Nyquist plot and fast separation of photogenerated electron–hole pairs.^66,67,69 From Fig. 13a, it can be found that the R_ct of the 1.0 wt% Au/4NFCN is smaller than that of all the other representative samples. And the results in Fig. 13a suggest that the fast charge separation and electron transfer of 1.0 wt% Au/4NFCN with the smaller arc size is significantly enhanced. To further demonstrate the aforementioned photocatalytic conjecture, the transition photocurrent responses were detected for the representative as-prepared photocatalysts. Fig. 13b and S12† exhibits the I–t curves for as-synthesized electrodes with three on–off cycles under visible light irradiation. It can be found that photocurrent responses arise in all the electrodes immediately while the light is turned on, and then quickly decreased when the light is off, indicating the electrodes are great stable and reproducible. The results in Fig. 13b indicate that the photocurrent under visible light irradiation is also in order of 1.0 wt% Au/4NFCN > 1.0 wt% Au/g-C₃N₄ > 4NFCN > 1.0 wt% Au/NiFe₂O₄ > g-C₃N₄ > NiFe₂O₄, and the Fig. S12† also indicates that 1.0 wt% Au/4NFCN has a highest photocurrent compared to g-C₃N₄/NiFe₂O₄ nanocomposites and other Au/g-C₃N₄/NiFe₂O₄ nanocomposites, namely that exhibits a much higher photocurrent than either binary or units photocatalysts for the 1.0 wt% Au/4NFCN, which further indicates that Au/g-C₃N₄/NiFe₂O₄ nanocomposites greatly promote the separation of photogenerated electron–hole pairs. This result agrees strongly with the EIS, PL measurements and above-mentioned mechanism.

Fig. 13 EIS changes (a) and transient photocurrent responses (b) of pure NiFe₂O₄, pure g-C₃N₄, 1.0 wt% Au/NiFe₂O₄, 1.0 wt% Au/g-C₃N₄, 4NFCN and 1.0 wt% Au/4NFCN, respectively.

Conclusions

In summary, we studied highly efficient, reusable visible-light-driven Au/g-C₃N₄/NiFe₂O₄ photocatalysts prepared by facile calcination and photodeposition methods. Optical, structure, surface, and elemental analysis verified that well-defined heterostructure formation among the NiFe₂O₄, g-C₃N₄ and AuNPs phase in the hybrid nanocomposites. In comparison with pristine g-C₃N₄ and NiFe₂O₄, the Au/g-C₃N₄/NiFe₂O₄ nanocomposites evolved H₂ generation rate is 1.607 mmol g⁻¹ h⁻¹, which shows the 28.7-fold and 30.9-fold higher photocatalytic hydrogen production rate respectively under visible light irradiation, whereas the hydrogen generation rate is even about 4.3 times higher than that of the optimal g-C₃N₄/NiFe₂O₄ nanocomposites. The superior photocatalytic properties of the ternary hybrid nanocomposites were due to (i) the high visible absorbance for charge carrier production, (ii) the electrons captured by AuNPs results in the fast separation, and (iii) the strongly SPR of AuNPs permission the generation of a lot of carriers under visible light irradiation. Furthermore, the possible mechanism of improved photocatalytic activity was also tentatively proposed on account of experimental results. Thereby, the Au/g-C₃N₄/NiFe₂O₄ hybrid ternary plasmonic photocatalysts is an excellent photocatalysts for make the best of solar energy and development of efficient strategies for hydrogen production.

Acknowledgements

The authors gratefully acknowledge generous financial support from National Natural Science Foundation of China (no. 21571064, 21371060), the Natural Science Foundation of Guangdong Province (S2013020013091) and the research fund of the Key Laboratory of Fuel Cell Technology of Guangdong Province for financial support.

Notes and references

1. J. Liu, Y. Liu, N. Y. Liu, Y. Z. Han, X. Zhang, H. Huang, Y. Lifshitz, S. T. Lee, J. Zhong and Z. H. Kang, Science, 2015, 347, 970.
2. M. A. Shannon, P. W. Bohn, M. Elimelech, J. G. Georgiadis, B. J. Marinas and A. M. Mayes, Nature, 2008, 452, 301.
3. P. Zhou, J. G. Yu and M. Jaroniec, Adv. Mater., 2014, 26, 4920.
4. R. S. Wang, X. X. Xu, Y. Zhang, Z. M. Chang, Z. C. Sun and W. F. Dong, Nanoscale, 2015, 7, 11082.
5. R. Li, F. Zhang, D. Wang, J. Yang, M. Li, J. Zhu, X. Zhou, H. Han and C. Li, Nat. Commun., 2013, 4, 1432.
6. L. Duan, F. Bozoglian, S. Mandal, B. Stewart, T. Privalov, A. Liobet and L. Sun, Nat. Chem., 2012, 4, 418.
7. P. P. Jing, J. N. Li, L. N. Pan, J. B. Wang, X. J. Sun and Q. F. Liu, J. Hazard. Mater., 2015, 284, 163.
8. Y. Isaka, S. Kato, D. C. Hong, T. Suenobu, Y. Yamada and S. Fukuzumi, J. Mater. Chem. A, 2015, 3, 12404.
9. C. Singh, A. Goyal and S. Singhal, Nanoscale, 2014, 6, 7959.
10. A. M. Balu, B. Baruwati, E. Serrano, J. Cot, J. Garcia-Martinez, R. S. Varma and R. Luque, Green Chem., 2011, 13, 2750.
11. Y. M. Ren, L. Q. Lin, J. Ma, J. Yang, J. Feng and Z. J. Fan, Appl. Catal., B, 2015, 165, 572.
12. Y. J. Yao, F. Lu, Y. P. Zhu, F. Y. Wei, X. T. Liu, C. Lian and S. B. Wang, J. Hazard. Mater., 2015, 297, 224.
13. Z. R. Zhu, F. Y. Liu, H. B. Zhang, J. F. Zhang and L. Han, RSC Adv., 2015, 5, 55499.
14. T. H. Yu, W. Y. Cheng, K. J. Chao and S. Y. Lu, Nanoscale, 2013, 5, 7356.
15. H. Song, L. P. Zhu, Y. G. Li, Z. R. Lou, M. Xiao and Z. Z. Ye, J. Mater. Chem. A, 2015, 3, 8353.
16. G. Y. He, J. J. Ding, J. G. Zhang, Q. L. Hao and H. Q. Chen, Ind. Eng. Chem. Res., 2015, 54, 2862.
17. A. L. Tiano, G. C. Papaefthymiou, C. S. Lewis, J. K. Han, C. Zhang, Q. Li, C. Y. Shi, A. M. M. Abeykoon, S. J. Z. Billinge, E. Stach, J. Thomas, K. Guerrero, P. Munayco, J. Munayco, R. B. Scorzelli, P. Burnham, A. J. Viescas and S. S. Wong, Chem. Mater., 2015, 27, 3572.
18. H. S. Kim, D. Kim, B. S. Liu, Y. Z. Hong, W. D. Shi, S. Lin and P. Li, Chem. Eng. J., 2014, 258, 301.
19. P. Sivakumar, R. Ramesh, A. Ramanand, S. Ponnusamy and C. Muthamizhchelvan, J. Alloys Compd., 2013, 563, 6.
20. H. Y. Ji, X. C. Jing, Y. G. Xu, J. Yan, H. P. Li, Y. P. Li, L. Y. Huang, Q. Zhang, H. Xu and H. M. Li, RSC Adv., 2015, 5, 57960.
21. Y. S. Fu, H. Q. Chen, X. Q. Sun and X. Wang, AIChE J., 2012, 58, 3298.
22. K. N. Harish, H. S. B. Naik, P. N. P. Kumar and R. Viswanath, ACS Sustainable Chem. Eng., 2013, 1, 1143.
23. S. S. Patil, M. S. Tamboli, V. G. Deonikar, G. G. Umarji, J. D. Ambekar, M. V. Kulkarni, S. S. Kolekar, B. B. Kale and D. R. Patil, Dalton Trans., 2015, 44, 20426.
24. S. Nayak, L. Mohapatra and K. M. Parida, J. Mater. Chem. A, 2015, 3, 18622.
25. S. Pany and K. Parida, Phys. Chem. Chem. Phys., 2015, 17, 8070.
26. J. W. Zhou, M. Zhang and Y. F. Zhu, Phys. Chem. Chem. Phys., 2014, 16, 17627.
27. J. X. Sun, Y. P. Yuan, L. G. Qiu, X. Jiang, A. J. Xie, Y. H. Shen and J. F. Zhu, Dalton Trans., 2012, 41, 6756.
28. F. Dong, Z. W. Zhao, T. Xiong, Z. L. Ni, W. D. Zhang, Y. J. Sun and W. K. Ho, ACS Appl. Mater. Interfaces, 2013, 5, 11392.
29. Z. Y. Wang, W. Guan, Y. J. Sun, F. Dong, Y. Zhou and W. K. Ho, Nanoscale, 2015, 7, 2471.
30. S. Martha, A. Nashim and K. M. Parida, J. Mater. Chem. A, 2013, 1, 7816.
31. S. W. Cao, J. X. Low, J. G. Yu and M. Jaroniec, Adv. Mater., 2015, 27, 2150.
32. R. C. Pawar, Y. Pyo, S. H. Ahn and C. S. Lee, Appl. Catal., B, 2015, 176, 654.
33. Y. Y. Bu, Z. Y. Chen and W. B. Li, Appl. Catal., B, 2014, 144, 622.
34. J. R. Ran, J. Zhang, J. G. Yu, M. Jaroniec and S. Z. Qiao, Chem. Soc. Rev., 2014, 43, 7787.
35. A. A. Melvin, K. Illath, T. Das, T. Raja, S. Bhattacharyya and C. S. Gopinath, Nanoscale, 2015, 7, 13477.
36. H. Y. Liu, T. T. Wang and H. P. Zeng, Part. Part. Syst. Charact., 2015, 32, 869.
37. J. H. Bi, Z. Y. Zhou, M. Y. Chen, S. J. Liang, Y. H. He, Z. Z. Zhang and L. Wu, Appl. Surf. Sci., 2015, 349, 292.
38. S. Samanta, S. Martha and K. M. Parida, ChemCatChem, 2014, 6, 1453.
39. J. P. Huo and H. P. Zeng, J. Mater. Chem. A, 2015, 3, 6258.
40. J. P. Huo and H. P. Zeng, J. Mater. Chem. A, 2015, 3, 17201.
41. J. Chen, S. H. Shen, P. Wu and L. J. Guo, Green Chem., 2015, 17, 509.
42. Y. Zheng, J. Liu, J. Liang, M. Jaroniec and S. Z. Qin, Energy Environ. Sci., 2012, 5, 6717.
43. J. Y. Qin, J. P. Huo, P. Y. Zhang, J. Zeng, T. T. Wang and H. P. Zeng, Nanoscale, 2016, 8, 2249.
44. S. Martha, A. Nashim and K. M. Parida, J. Mater. Chem. A, 2013, 1, 7816.
45. N. Y. Cheng, J. Q. Tian, Q. Liu, C. J. Ge, A. H. Qusti, A. M. Asiri, A. O. Al-Youbi and X. P. Sun, ACS Appl. Mater. Interfaces, 2013, 5, 6815.
46. S. Samanta, S. Martha and K. Parida, ChemCatChem, 2014, 6, 1453.
47. S. Kumar, A. Baruah, S. Tonda, B. Kumar, V. Shanker and B. Sreedhar, Nanoscale, 2014, 6, 4830.
48. S. F. Chen, W. Zhao, S. J. Zhang and W. Liu, Chem. Eng. J., 2009, 148, 263.
49. Y. F. Zhang, F. Mao, H. J. Yan, K. W. Liu, H. M. Cao, J. G. Wu and D. Q. Xiao, J. Mater. Chem. A, 2015, 3, 109.
50. P. P. Jing, J. L. Du, J. B. Wang, W. Lan, L. N. Pan, J. N. Li, J. W. Wei, D. R. Cao, X. L. Zhang, C. B. Zhao and Q. F. Liu, Nanoscale, 2015, 7, 14738.
51. H. S. Kim, D. Kim, B. S. Kwak, G. B. Han, M. H. Um and M. Kang, Chem. Eng. J., 2014, 243, 272.
52. Y. Li, X. F. Dong and J. P. Li, Particuology, 2011, 9, 475.
53. A. A. S. Nair, R. Sundara and N. Anitha, Int. J. Hydrogen Energy, 2015, 40, 3259.
54. S. Q. Liu, B. Xiao, L. R. Feng, S. S. Zhou, Z. G. Chen, C. B. Liu, F. Chen, Z. Y. Wu, N. Xu, W. C. Oh and Z. D. Meng, Carbon, 2013, 64, 197.
55. L. Tong, Z. Li, T. Zhu, H. Xu and Z. Liu, J. Phys. Chem. C, 2008, 112, 7119.
56. H. Zhang, X. F. Fan, X. Quan, S. Chen and H. T. Yu, Environ. Sci. Technol., 2011, 45, 5731.
57. W. Zhao, Y. Guo, S. M. Wang, H. He, C. Sun and S. G. Yang, Appl. Catal., B, 2015, 165, 335.
58. T. G. Xu, L. W. Zhang, H. Y. Cheng and Y. F. Zhu, Appl. Catal., B, 2011, 101, 382.
59. X. J. Bai, L. Wang and Y. F. Zhu, ACS Catal., 2012, 2, 2769.
60. Y. J. Wang, Q. S. Wang, X. Y. Zhan, F. M. Wang, M. Safdar and J. He, Nanoscale, 2013, 5, 8326.
61. J. Zeng, H. Yang, J. Y. Deng, H. J. Liu, X. Yi, L. P. Yang and B. Yi, Chem. Eng. J., 2015, 273, 519.
62. Y. Z. Chen, D. Q. Zeng, K. Zhang, A. Lu, L. S. Wang and D. L. Peng, Nanoscale, 2014, 6, 874.
63. J. H. Bi, Z. Y. Zhou, M. Y. Chen, S. J. Liang, Y. H. He, Z. Z. Zhang and L. Wu, Appl. Surf. Sci., 2015, 349, 292.
64. N. Zhang, W. Wang, P. Wang, X. T. Zhang and Z. Y. Zhao, CrystEngComm, 2013, 15, 3301.
65. L. He, L. Li, T. T. Wang, H. Gao, G. Z. Li, X. T. Wu, Z. M. Su and C. G. Wang, Dalton Trans., 2014, 43, 16981.
66. T. Song, J. P. Huo, T. Liao, J. Zeng, J. Y. Qin and H. P. Zeng, Chem. Eng. J., 2016, 287, 359.
67. H. Y. Liu, K. Y. Zhao, T. T. Wang, J. Y. Deng and H. P. Zeng, Mater. Sci. Semicond. Process., 2015, 40, 670.
68. Y. M. He, Y. Wang, L. H. Zhang, B. T. Teng and M. H. Fan, Appl. Catal., B, 2015, 168, 1.
69. Y. Zhou, Y. H. Zhu, X. L. Yang, J. F. Huang, W. Chen, X. M. Lv, C. Y. Li and C. Z. Li, RSC Adv., 2015, 5, 50454.
70. Y. J. Xu, M. Xie, S. Q. Huang, H. Xu, H. Y. Ji, J. X. Xia, Y. P. Li and H. M. Li, RSC Adv., 2015, 5, 26281.
71. P. Niu, G. Liu and H. M. Cheng, J. Phys. Chem. C, 2012, 116, 11013.
72. L. Y. Huang, H. Xu, Y. P. Li, H. M. Li, X. N. Cheng, J. X Xia, Y. G. Xu and G. B. Cai, Dalton Trans., 2013, 42, 8606.
73. K. Yang, C. Meng, L. L. Lin, X. Y. Peng, X. Chen, X. X. Wang, W. X. Dai and X. Z. Fu, Catal. Sci. Technol., 2016, 6, 829.
74. J. Mao, T. Y. Peng, X. H. Zhang, K. Li, L. Q. Ye and L. Zan, Catal. Sci. Technol., 2013, 3, 1253.
75. R. H. Liu, H. Huang, H. T. Li, Y. Liu, J. Zhang, Y. Y. Li, S. Zhang and Z. H. Kang, ACS Catal., 2014, 4, 328.

---

Footnote

† Electronic supplementary information (ESI) available: this includes the complementary figures and tables. See DOI: 10.1039/c6ra08356k

---