Bio-inspired Z-scheme g-C₃N₄/Ag₂CrO₄ for efficient visible-light photocatalytic hydrogen generation

Abstract

Due to low charge separation efficiency and poor stability, it is usually difficult for single-component photocatalysts such as graphitic carbon nitride (g-C₃N₄) and silver chromate (Ag₂CrO₄) to fulfill photocatalytic hydrogen production efficiently. Z-scheme charge transport mechanism that mimics the photosynthesis in nature is an effective way to solve the above problems. Inspired by photosynthesis, we report Ag₂CrO₄ nanoparticles-decorated g-C₃N₄ nanosheet as an efficient photocatalyst for hydrogen evolution reaction (HER) with methanol as sacrificial agent. The formation of Z-scheme g-C₃N₄/Ag₂CrO₄ nanosheets photocatalysts could inhibit the recombination of photogenerated electron-hole pairs, promote the generation of hydrogen by photosplitting of water. The experiment results indicate that g-C₃N₄/Ag₂CrO₄ nanocomposites present enhanced photocatalytic activity and stability in the H₂ evolution of water splitting. And the nanocomposites g-C₃N₄/Ag₂CrO₄(23.1%) show the 14 times HER efficiency compared to that of bare g-C₃N₄.

Introduction

With fossil fuel reserves dwindling every day, there is an urgent need for clean and sustainable alternative energy sources. Hydrogen energy is an attractive alternative resource to fossil fuels due to its high energy density and environmentally friendly characteristics^1–6. Many methods are used to produce hydrogen, which can be divided into two major categories based on the required raw materials and processes: reforming processes and splitting water processes. The former process is relatively mature and uses widely. However, it has some disadvantages such as high cost, large energy consumption, low efficiency, and complicated equipments and processes. Splitting water processes contains electrolytic water splitting, photocatalytic water splitting, bio-photocatalytic water splitting and thermochemical water splitting. Water electrolysis technology is relatively mature, with simple equipment, no pollution, and high product purity but large energy consumption; water biophotolysis process Bio-photocatalytic water splitting requires harsh reaction environments but has poor stability. There is no problem in the feasibility and high efficiency for thermochemical water splitting technology, but further research is needed to reduce costs and achieving efficient recycling. Among them, photocatalytic water splitting is a more efficient pathway to produce hydrogen gas and also does not generate any undesirable byproducts^7–17. Since the discovery of the Honda-Fujishima effect in 1972¹⁸, photocatalytic hydrogen production through water splitting has become a promising technology for utilizing renewable solar energy^19–24. Development of efficient photocatalytic systems for hydrogen evolution via photoinduced water splitting is an active field of energy research. However, due to narrow absorption range, poor stability, low charge-separation efficiency and weak redox ability, it is often difficult for a single-component photocatalyst to fulfill this requirement. The heterostructure composed of two semiconductor catalysts can increase the catalytic efficiency^25–29. Photosynthesis as a nature heterostructure system is widely studied due to its fast and efficient photocatalytic reaction.

In natural photosynthesis process, the Z-scheme photoreaction system is an important part of the plant photosynthesis, which involves two photochemical reactions and a series of intermediate enzymatic redox reactions. The electron transfer process is shown in Fig. 1a. After a series of reactions, the photogenerated electrons with high reducing ability and holes with high oxidizing ability are left in the LUMO of PS I and the HOMO of PS II, respectively. The transfer process of electrons in the figure constitutes a shape of letter Z, so it is called a Z-scheme^30–32. This system was first proposed by Bard in 1979 after studying the photosynthesis of plants³³. The Z-scheme photocatalytic system shows excellent redox ability in due to the different band gap of two semiconductor with broadening light absorption and the special charge transfer path ensuring high separation efficiency of photo-generated charges. Considering these unique advantages of the natural Z-scheme photocatalytic system, the artificial Z-scheme photocatalytic systems have been developed and used widely in photocatalytic field^25,34–38. The artificial heterogeneous Z-scheme photocatalytic systems, mimicking the natural photosynthesis process, overcome the drawbacks of single-component photocatalysts and satisfy those aforementioned requirements.

Figure 1.

(a) Natural photosynthesis (the electrons in HOMO of PS II are excited to its LUMO under solar light; Then, the photogenerated electrons in LUMO of PS II are transferred to HOMO of PS I through the electron mediator. Further, the electrons in HOMO of PS I are excited to its LUMO. The transfer process of electrons in the figure constitutes a shape of letter Z, so it is called a Z-scheme) and (b) an artificial analogy composed of an organic semiconductor g-C₃N₄ (equivalent to PS I) and silver chromate (equivalent to PS II) (the electrons in HOMO of Ag₂CrO₄ are excited to its LUMO under visible light irradiation; Then, the electrons in HOMO of g-C₃N₄ are excited to its LUMO under visible light irradiation; the excited electrons of Ag₂CrO₄ and the holes of g-C₃N₄ recombine; as a result, the migration route of electrons is a Z-scheme).

For the first time, Martin³⁹ reported that the robust organic semiconductor g-C₃N₄ can be integrated into a Z-scheme water splitting in a liquid system. While in this system, they used an electron acceptor/donor (NaI/IO³⁻) pair which they all absorb light to the different extent and occur back reactions. Therefore, the number of adsorbed photons by photocatalysts decreases and the effective number of photogenerated electrons and holes decrease sharply³⁰. Mimicking the natural photosynthesis process, some all solid-state Z-scheme photocatalyst system based on g-C₃N₄ with other semiconductors were obtained and used, such as WO₃, Ag₃PO₄, BiOX (X=Cl, Br, I), V₂O₅, etc^40–44. A Z-scheme system of g-C₃N₄/WO₃ was used to generate hydrogen and its separation mechanisms were also studied deeply^40,45,46. He et al.⁴¹ reported that the Z-Scheme Ag₃PO₄/g-C₃N₄ composite was used to convert CO₂ to fuel such as CO, methanol, methane and ethanol. Liu et al.⁴⁴ synthetized a ternary composite photocatalyst g-C₃N₄/Ag₃PO₄/Ag₂MoO₄ for water splitting. BiOCl and BiOI were also compounded with g-C₃N₄ as high efficiency photocatalysts^43,47. As a promising catalyst, Bi₂WO₆ was also combined with g-C₃N₄ to form a Z-scheme catalytic system⁴⁸.

Silver chromate (Ag₂CrO₄) as a narrow band gap semiconductor with band gap of 1.80 eV has unique electronic structure, crystal structure⁴⁹, excellent light sensitivity and high photocatalytic activity^50,51. Hence, in recent years, some Ag₂CrO₄-based photocatalysts have been fabricated^51–53. However, similar to other silver-based photocatalysts, the big aggregated particle size and easy photocorrosion properties seriously caused poor stability and restricted the photocatalytic performance of the Ag₂CrO₄ photocatalyst. The g-C₃N₄/Ag₂CrO₄ nanocomposites were synthesized and characterized, the dispersion and light stability of Ag₂CrO₄ were improved greatly. And their catalytic hydrogen evolution activity was evaluated when using methanol as the sacrificial electron donor under visible light. The as-obtained g-C₃N₄/Ag₂CrO₄ composites showed distinctly enhanced photo-catalytic activity than that of pure g-C₃N₄ nanosheets on the evolution of H₂ under visible light irradiation. Compared to the two individuals, the enhancement of HER efficiency was mainly ascribed to the synergetic effect between Ag₂CrO₄ and g-C₃N₄ (as shown in Fig. 1b), effectively improved photogenerated electron-hole pairs separation efficiency via the direct Z-scheme charge transfer mechanism. Notably, the photocorrosion of Ag₂CrO₄ was efficiently hindered due to this synergistic effect. Furthermore, a possible direct Z-scheme mechanism for the enhanced photocatalytic activity of the g-C₃N₄/Ag₂CrO₄ composite was also discussed based on the relative band gap positions of these two semiconductors. The existence of the Z-scheme mechanism means strong redox ability and high transfer efficiency of photogenerated electron-hole pairs. This research will broaden the studies of Ag₂CrO₄ and g-C₃N₄ photocatalysts with excellent photocorrosion inhibition ability and high photocatalytic activity under visible light.

Results and Discussion

Phase structure and morphology analysis

In order to analyze the crystal structure of the prepared g-C₃N₄, g-C₃N₄/Ag₂CrO₄ composites, and Ag₂CrO₄, the XRD patterns were obtained (as shown in Fig. 2). Fig. 2a was the pattern of g-C₃N₄, two diffraction peaks at 27.6° and 13.0° were indexed to (002) and (100) planes of hexagonal g-C₃N₄ (JCPDS card No. 87–1526), corresponding to the graphite-like stacking and the in-plane structural repeating motifs of the conjugated aromatic units of g-C₃N₄⁵⁴. From Fig. 2b–f, the patterns correspond to the samples with an increasing Ag₂CrO₄ mass ratio to g-C₃N₄. For the g-C₃N₄/Ag₂CrO₄ composites, orthorhombic phase Ag₂CrO₄ and hexagonal phase g-C₃N₄ were both observed and no other impurity peaks were found, implying that g-C₃N₄ and Ag₂CrO₄ keep pure phase and no impurities formed in g-C₃N₄/Ag₂CrO₄ composites. Moreover, the peaks of Ag₂CrO₄ in g-C₃N₄/Ag₂CrO₄ composites were becoming clearer with the increasing of Ag₂CrO₄. And at the same time the peaks of g-C₃N₄ gradually decreased. Fig. 2g was the pattern of Ag₂CrO₄. It was clear to see that all diffraction peaks of the as-prepared Ag₂CrO₄ coincided well with the orthorhombic phase of Ag₂CrO₄ (JCPDS No. 26-0952).

Figure 2.

XRD patterns of (a) g-C₃N₄, (b) g-C₃N₄/Ag₂CrO₄(9.1%), (c) g-C₃N₄/Ag₂CrO₄(16.7%), (d) g-C₃N₄/Ag₂CrO₄(23.1%), (e) g-C₃N₄/Ag₂CrO₄(28.6%), (f) g-C₃N₄/Ag₂CrO₄(33.3%) and (g) Ag₂CrO₄.

In order to observe the morphology of the bare g-C₃N₄, Ag₂CrO₄ and the as-prepared g-C₃N₄/Ag₂CrO₄ composite, TEM was performed (see Fig. 3). The pure g-C₃N₄ sample possessed a very thin 2D layer structure (shown as Fig. 3a), which meant that the g-C₃N₄ was prepared successfully. As to the pristine Ag₂CrO₄ sample (shown in Fig. 3b), besides the obvious aggregations there was no special morphological features, and the sizes of Ag₂CrO₄ particles were between 140–300 nm and asymmetrical. The TEM image of the g-C₃N₄/Ag₂CrO₄ composite with the mass weight ratio 23.1% was showed in Fig. 3c. From the TEM picture, we could see that the Ag₂CrO₄ particles (labeled by red circles) with the size below 10 nm were uniformly embedded on g-C₃N₄. The adding of graphene-like g-C₃N₄ nanosheets made the Ag₂CrO₄ particles composites in the g-C₃N₄/Ag₂CrO₄(23.1%) closely enwrapped by g-C₃N₄ and much smaller than pure Ag₂CrO₄ nanoparticles. These results indicate that g-C₃N₄ nanosheets could not only serve as a graphene-like substrate to attach Ag₂CrO₄ nanoparticles on their surface, but also inhibit Ag₂CrO₄ nanoparticles aggregating into the big particles. Therefore, g-C₃N₄/Ag₂CrO₄ composite photocatalyst with closely interconnections between g-C₃N₄ and Ag₂CrO₄ nanoparticles were successfully synthesized, which would be beneficial for the transfer of the photogenerated electron-hole pairs. In Fig. 3d, obviously, the clear fringes spacing was ca. 0.28 nm, corresponding to the (211) lattice plane of orthorhombic phase Ag₂CrO₄^55,56. The composites with other different weight ratios (9.1%, 16.7%, 28.6%, and 33.3%) were shown in Fig. S1. We could see that when the ratio of the Ag₂CrO₄ below 23.1%, small Ag₂CrO₄ nanoparticles (below 10 nm) were uniformly embedded on g-C₃N₄ (labeled by red circles in Figs 3b, S1a and S1b). While the theoretical weight ratio was increased to 28.6%, Ag₂CrO₄ began to aggregate, the size of Ag₂CrO₄ grow to above 20 nm. When the theoretical ratio was 33.3%, the aggregation was very serious just like the pure Ag₂CrO₄.

Figure 3.

TEM images of (a) g-C₃N₄, (b) Ag₂CrO₄, (c) g-C₃N₄/Ag₂CrO₄(23.1%) (Ag₂CrO₄ were labeled by red circles), (d) HRTEM of g-C₃N₄/Ag₂CrO₄(23.1%).

Based on Figs 3 and S1, a possible formation mechanism of g-C₃N₄/Ag₂CrO₄ composites is proposed. Obviously, the structure of Ag₂CrO₄ ranged from nanoparticles (below 10 nm) to sub-microparticles (140–300 nm) with the decreasing of g-C₃N₄ mass fraction, which strongly confirmed that the g-C₃N₄ sheets played vital role in the formation of the nanojunction. Many investigations showed that large specific surface area and two-dimensional structure of g-C₃N₄ could provide a large scaffold for anchoring various substrates^57,58. In addition, the surface of urea-derived graphitic g-C₃N₄ possessed positive charge with abundant alamino groups (C-NH_x), which could provide a suitable environment for attracting negative charge particles via electrostatic attraction^59,60. Accordingly, in our study, the CrO₄²⁻ ions could adsorb onto g-C₃N₄ sheets via the electrostatic force in g-C₃N₄ suspension. The anchored CrO₄²⁻ ions would form Ag₂CrO₄ nanocrystals in situ on the surface of g-C₃N₄ sheets during the reaction, then the tiny nanocrystal nucleus grows into the nanoparticles through oriented growth on the surface of g-C₃N₄ support⁶¹. Eventually, the Ag₂CrO₄ nanoparticles uniformly and tightly distribute onto the surface of g-C₃N₄ sheets (As shown in Figs 3c and S1a,b). However, with the increase of Ag₂CrO₄ mass ratio, the capacity of electrostatic attraction between g-C₃N₄ and CrO₄²⁻ ions was decreased because the effective positive charge surface of g-C₃N₄ was decreasing. When the mass fraction of Ag₂CrO₄ was more than 28.6%, the excess of CrO₄²⁻−ions could not tightly adsorb on the surface of g-C₃N₄ sheets and would grow into Ag₂CrO₄ nanoparticles freely during the reaction progress. However, these free Ag₂CrO₄ nanocrystals possessed high specific surface energy, which would assemble spontaneously and form hierarchical sub-microparticles for reducing the interfacial energy^52,62. As a result, part of Ag₂CrO₄ nanoparticles uniformly grew on the surface of g-C₃N₄ due to the strong electrostatic attraction between CrO₄²⁻− ions and C-NH_x, whereas other Ag₂CrO₄ nanoparticles were densely self-assembled and formed 3D hierarchical structures which covered the g-C₃N₄ sheets (Fig. S1c). With the mass fraction of free Ag₂CrO₄ nanoparticles further increasing, the crystal growth of nanoparticles would cause the exfoliation of the g-C₃N₄ sheets (Fig. S1d)^63,64. The Ag₂CrO₄ sub-microparticles could directly contacted with g-C₃N₄ sheets, which were shown in Fig. S1d. The pure Ag₂CrO₄ 3D hierarchical structures was obtained with the absence of g-C₃N₄. It revealed that when the particles of Ag₂CrO₄ nanoparticles were not restricted by the functional alamino groups on the surface of g-C₃N₄ sheets, they would assemble spontaneously in a random way to form 3D sub-microparticles (shown as Fig. 3c)⁴⁹. All in all, the results clearly confirmed that the mass ratio of g-C₃N₄ could be a key parameter for formation of g-C₃N₄/Ag₂CrO₄ nanocomposites. The positive charge and 2D structure of g-C₃N₄ sheets could provide a suitable environment for the growth of nanoparticles.

Component analysis

The elemental compositions of g-C₃N₄/Ag₂CrO₄(23.1%) nanocomposite were explored by EDS technique and the results were shown in Fig. 4. It could be seen that the g-C₃N₄/Ag₂CrO₄(23.1%) composite was composed of only C, N, Ag, Cr and O elements and no other element or impurity found, demonstrating that the existence of g-C₃N₄ and Ag₂CrO₄ in the as-fabricated g-C₃N₄/Ag₂CrO₄(23.1%) composite (shown as in Fig. 4a). Quantitative analysis of the g-C₃N₄/Ag₂CrO₄(23.1%) nanocomposite showed that weight percents of C, N, Ag, Cr and O elements were 29.4, 47.2, 15.0, 3.6, and 4.8%, respectively, which they were close to the theoretical percents of 30.1, 46.8, 15.0, 3.6 and 4.5%, respectively. To further investigate elemental composition and distribution uniformity, the elemental maps for g-C₃N₄/Ag₂CrO₄(23.1%) composite were displayed in Fig. 4c–g, which indicated that Ag, Cr and O elements were homogeneously distributed in the whole host of the g-C₃N₄/Ag₂CrO₄(23.1%) composite. Considering the Ag, Cr and O presented in the form of Ag₂CrO₄, these analyses demonstrated Ag₂CrO₄ tended to integrate with g-C₃N₄ nanosheets firmly and then formed hybrid structures. Except that, X-ray photoelectron spectroscopy and Fourier-transform infrared spectroscopy were also used to analyze the composition of nanomaterials (as shown in Figs S2 and S4).

Figure 4.

(a) EDS spectra for the g-C₃N₄/Ag₂CrO₄(23.1%), (b) SEM of g-C₃N₄/Ag₂CrO₄(23.1%), (c–g) EDS mapping for different elements of the g-C₃N₄/Ag₂CrO₄(23.1%) nanocomposite.

Optical properties of the g-C₃N₄/Ag₂CrO₄ composites

It was believed that remarkable absorption enhancement in the visible-light region was beneficial for improving photocatalytic activity. Optical absorption spectra of the prepared samples were provided by UV-vis DRS and the results in the range of 200–800 nm were shown in Fig. 5A. As can be seen, the absorption edge of the pure g-C₃N₄ and Ag₂CrO₄ located at approximately 454 nm and 712 nm, respectively. It was noteworthy that the individual Ag₂CrO₄ showed the photo-absorption property in the entire waveband, which suggested that an effective utilization of the solar source over Ag₂CrO₄ could be acquired. Interestingly, the absorption of composites was gradually strengthened in the visible region with increasing of Ag₂CrO₄ content. This phenomenon could be attributed to the difference between the band gap energies of bare g-C₃N₄ and Ag₂CrO₄. With more and more Ag₂CrO₄ nanoparticles with relative narrow band gap produced on the surface of g-C₃N₄, the band gap of the g-C₃N₄/Ag₂CrO₄ would decrease. Other groups also found the same trend when coupling the broader band gap semiconductors with other relatively narrow band gap semiconductors^42,65. Therefore, introducing Ag₂CrO₄ into g-C₃N₄ might be favorable for photocatalytic reaction due to the enhancing of light absorbance. These results implied that the g-C₃N₄/Ag₂CrO₄ nanocomposite had the potential to be efficient visible light-driven photocatalyst. The UV-vis DRS spectra were also applied to calculate band gap energy (E_g) using Tauc’s equation:

$$\alpha h\nu =\mathrm{A}{(h\nu -E_g)}^{n/2}$$ 1

Figure 5.

(A) UV–vis DRS (insert are the photos of samples) of (a) g-C₃N₄, (b) g-C₃N₄/Ag₂CrO₄(9.1%), (c) g-C₃N₄/Ag₂CrO₄(16.7%), (d) g-C₃N₄/Ag₂CrO₄(23.1%), (e) g-C₃N₄/Ag₂CrO₄(28.6%), (f) g-C₃N₄/Ag₂CrO₄(33.%) and (g) Ag₂CrO₄; (B) the band gap energies of (a) g-C₃N₄, (b) g-C₃N₄/Ag₂CrO₄(23.1%), and (c) Ag₂CrO₄.

In this equation, α, hν, and A were absorption coefficient, the photon energy, and proportionality constant, respectively⁶⁶. The value of n was determined by the type of optical transition of a semiconductor (n = 1 for direct transition, and n = 4 for indirect transition), and the n values of g-C₃N₄ and Ag₂CrO₄ were 1^49,67. So according to the calculation ((αhν)² = A²hν − A²E_g), the plot of (αhν)² versus photon energy (hν) for g-C₃N₄, pure Ag₂CrO₄ and the as-prepared g-C₃N₄/Ag₂CrO₄(23.1%) composites were shown as Fig. 5B. The band gap energies of g-C₃N₄ and Ag₂CrO₄ were 2.73 eV and 1.74 eV, respectively. And the wavelength thresholds of the g-C₃N₄/Ag₂CrO₄(23.1%) composite were estimated at 461 and 708 nm, corresponding to the band gaps at 2.69 and 1.75 eV, respectively ascribed to g-C₃N₄ and Ag₂CrO₄. The reduced band gaps of the g-C₃N₄/Ag₂CrO₄(23.1%) composite and more response to the visible light were caused by the loaded Ag₂CrO₄ nanoparticles, thus more efficient utilization of solar energy could be achieved, and the improved photocatalytic activity of the g-C₃N₄/Ag₂CrO₄ composite could be anticipated.

Visible light photocatalytic activities and evaluation of stability

The photocatalytic activity of all samples was evaluated by monitoring the hydrogen evolution in the presence of a sacrificial reagent (methanol) under visible light illumination (λ ≥ 420 nm). Control experiments indicated that there was no appreciable H₂ production in the absence of photocatalysts, light irradiation or H₂O. According to the standard curve: y = 165550x, R² = 0.9997 (as shown in Fig. S5), time-dependent (Fig. 6A) and average (Fig. 6B) photoinduced H₂ evolution for different photocatalysts were observed and corresponding results were listed in Table S1. As illustrated, H₂ production was not detected if pure Ag₂CrO₄ was used as a photocatalyst for the whole visible-light irradiation for 5 h; this suggested that Ag₂CrO₄ alone was not active for photocatalytic H₂ generation. Pure g-C₃N₄ showed a poor photocatalytic activity under visible light (64.3 μmol g⁻¹ h⁻¹), owing to its limited light-harvesting efficiency and fast recombination of photogenerated electron-hole pairs. However, after the two semiconductors were combined, the evolution rate of g-C₃N₄/Ag₂CrO₄ was greatly enhanced and the largest rate as high as 902.1 μmol g⁻¹ h⁻¹ was attained by using g-C₃N₄/Ag₂CrO₄(23.1%) as photocatalyst, which was 14 times that of g-C₃N₄. Compared with other g-C₃N₄ based composite photocatalysts, the H₂ production rate of g-C₃N₄/Ag₂CrO₄ was enhanced to a great extent, which could be ascribed to the effect of Ag₂CrO₄ (Table 1). In the scope of our literature survey, our efficiency of hydrogen production is very high. It is just lower compared to that of literature ref.⁶ and ref.¹⁵. Furthermore, it could also be clearly observed that the content of Ag₂CrO₄ exhibited a great influence on the H₂ production rate. With its increasing, the H₂ production rate rose firstly and then decreased. For bare g-C₃N4, the H₂ production rate was extremely low because of the deeply trapped electrons and fast recombination of electron-hole pairs; the photogenerated electrons were not transferred to catalytic sites and were unable to participate in H₂ production⁶⁸. As a compare, the pure g-C₃N₄ and g-C₃N₄/Ag₂CrO₄(23.1%) was conducted photogenerated hydrogen without Pt co-catalyst or methanol as electro donors. From Fig. 6D (the corresponding results were listed in Table S3), we could see that with methanol but not Pt, there was no observable hydrogen generation be detected using pure g-C₃N₄ as photocatalyst under visible light irradiation. And the hydrogen production efficiency of g-C₃N₄/Ag₂CrO₄(23.1%) was reduced. While without any electron donors such as methanol, no H₂ is produced either using pure g-C₃N₄ or g-C₃N₄/Ag₂CrO₄(23.1%) as a catalyst.

Figure 6.

(A) Time courses of photocatalytic H₂ (B) the average rate of H₂ of (a) g-C₃N₄, (b) g-C₃N₄/Ag₂CrO₄(9.1%), (c) g-C₃N₄/Ag₂CrO₄(16.7%), (d) g-C₃N₄/Ag₂CrO₄(23.1%), (e) g-C₃N₄/Ag₂CrO₄(28.6%), (f) g-C₃N₄/Ag₂CrO₄(33.3%) and (g) Ag₂CrO₄ under visible light irradiation; (C) Recyclability of the g-C₃N₄/Ag₂CrO₄(23.1%) in five successive experiments for the H₂ evolution under visible light irradiation; (D) Time courses of photocatalytic H₂ (insert: the average rate of H₂) of (a) g-C₃N₄ with methanol without Pt, (b) g-C₃N₄/Ag₂CrO₄(23.1%) with methanol without Pt, (c) g-C₃N₄ without methanol with Pt, and (d) g-C₃N₄/Ag₂CrO₄(23.1%) without methanol with Pt.

Table 1.

Comparison of the photocatalytic H₂ production rate reported in the literatures with Z-scheme g-C₃N₄/Ag₂CrO₄ in our work with methanol as sacrificial agent under visible light irradiation.

Sample	Efficiency (μmol h−1 g−1)	Co-catalyst (Pt)	Light source	Refrence
g-C3N4/InVO4	212	0.6%	>420 nm	ref.5
g-C3N4/NiFe-LDH	24800	No	≥420 nm	ref.6
GCN/NT NFs	8931.3	A certain amount	simulated solar light	ref.15
CdS/Au/g-C3N4	19	No	>420 nm	ref.16
Fe2(MoO4)3/g-C3N4	0.18	No	>420 nm	ref.17
Au/SnO2/g-C3N4	770	No	>400 nm	ref.21
Au/PtO/g-C3N4	16.9	No	>400 nm	18222
TiO2/g-C3N4	74.7	0.5%	>400 nm	ref.23
MoS2/g-C3N4	231	1%	>400 nm	ref.24
g-C3N4/Au/P25	259	No	simulated solar light	ref.25
Fe/P-g-C3N4	150.6	No	>400 nm	ref.26
g-C3N4/WS2	101	No	≥420 nm	ref.27
Ag2S/g-C3N4	200	No	=420 nm	ref.28
g-C3N4/TiO2	559.7	No	Full light	ref.29
g-C3N4/Ag2CrO4	902.1	0.6%	≥420 nm	Our work

To evaluate the stability and reusability of the g-C₃N₄/Ag₂CrO₄ nanocomposites, a recycling test of the g-C₃N₄/Ag₂CrO₄(23.1%) (5 h) was performed and the corresponding results were displayed in Fig. 6C and the corresponding results were listed in Table S2. The amount of H₂ produced increased steadily with an extension in the reaction time and no significant deactivation was observed after 5 cycles, indicating that the g-C₃N₄/Ag₂CrO₄ nanocomposites had high stability in the photocatalytic H₂ evolution. Furthermore, the XRD of g-C₃N₄/Ag₂CrO₄(23.1%) before and after H₂ production were shown as Fig. S6. It was seen that the peaks of the catalyst were similar, and the morphology was steady. Furthermore, there were no new peaks according to Pt appeared due to its low content. And the TEM of g-C₃N₄/Ag₂CrO₄(23.1%) after five cycles also indicated that Ag₂CrO₄ nanoparticless were still anchored on the surface of g-C₃N₄ (Fig. S7), suggesting a good stability of inherent structure for g-C₃N₄ and Ag₂CrO₄ nanoparticles. Such high stability might result from the formation of the heterostructure between g-C₃N₄ and Ag₂CrO₄.

Charge transfer properties

As shown in Fig. 6, the Z-scheme nanocomposite formed between g-C₃N₄ and Ag₂CrO₄ dramatically enhanced the photocatalytic performance under visible-light irradiation. To explore the effect of silver chromate on separation efficiency of the photogenerated electron-hole pairs on the g-C₃N₄, PL spectra for the g-C₃N₄, g-C₃N₄/Ag₂CrO₄ and Ag₂CrO₄ samples were provided in the range of 400–600 nm and the results were displayed in Fig. 7A. Since PL emission spectra derived from the recombination of free carriers, it was generally accepted that the PL spectrum with low intensity indicated efficiently separation of the charge carriers, leading to participation of more electrons and holes in the oxidation and reduction reactions. It could be observed that all of the samples had similar spectra in this range. It was clear to see that the bare g-C₃N₄ had a wide and strong peak around 450 nm. However, there was a considerable decrease in the intensity of the PL spectrum for the g-C₃N₄/Ag₂CrO₄ nanocomposites compared to that of the pure g-C₃N₄. Hence, the photogenerated electron-hole pairs could migrate easily between g-C₃N₄ and Ag₂CrO₄, due to the matching band potentials^30,69. It was noted that the g-C₃N₄/Ag₂CrO₄(23.1%) sample showed the weakest emission intensity, meaning that it had the lowest recombination rate of the photogenerated charge carriers and the highest photocatalytic activity compared with other composites. Additionally, it was worth noting that the peaks density of PL first declined, then increased along with the further increasing content of Ag₂CrO₄. That was excess Ag₂CrO₄ species increased the recombination rate of photogenerated charge carriers, leading to inferior photocatalytic efficiency, arising from the excess Ag₂CrO₄ species might agglomerate seriously leading to decrease of the combinations between counterparts of the nanocomposite and acted as a recombination center and thereby reducing the efficiency of charge separation^70–72. The time-resolved photoluminescence spectra for g-C₃N₄ and g-C₃N₄/Ag₂CrO₄(23.1%) were tested and shown in Fig. 7B. It could be clearly observed that g-C₃N₄/Ag₂CrO₄(23.1%) showed longer decay time value compared to pure g-C₃N₄ which could be ascribed to the effective charge transfer across the interface of g-C₃N₄ and Ag₂CrO₄⁷³.

Figure 7.

(A) Photoluminescence spectra of (a) g-C₃N₄, (b) g-C₃N₄/Ag₂CrO₄(9.1%), (c) g-C₃N₄/Ag₂CrO₄(16.7%), (d) g-C₃N₄/Ag₂CrO₄(23.1%), (e) g-C₃N₄/Ag₂CrO₄(28.6%), (f) g-C₃N₄/Ag₂CrO₄(33.3%) and (g) Ag₂CrO₄; (B) Time-resolved photoluminescence spectra for (a) g-C₃N₄, (b) g-C₃N₄/Ag₂CrO₄(23.1%); (C) it of (a) g-C₃N₄, (b) g-C₃N₄/Ag₂CrO₄(9.1%), (c) g-C₃N₄/Ag₂CrO₄(16.7%), (d) g-C₃N₄/Ag₂CrO₄(23.1%), (e) g-C₃N₄/Ag₂CrO₄(28.6%), (f) g-C₃N₄/Ag₂CrO₄(33.3%) and (g) Ag₂CrO₄.

Consequently, we recorded the transient photocurrent responses of the pure g-C₃N₄, g-C₃N₄/Ag₂CrO₄ nanocomposite and Ag₂CrO₄ under dark conditions and visible light irradiation. Fig. 7C showed a comparison of the photocurrent-time curves for photocatalysts, with three on-off intermittent irradiation cycles. It was widely considered as the most efficient evidence for explaining the electrons and holes separation in the composite photocatalysts^38,74,75. Generally, the corresponding relationship was recognized as follows: the higher photocurrent implied the higher electrons-holes separation efficiency, thus leading to the higher photocatalytic activity. It was clear that the photocurrent densities rapidly decreased to zero as soon as the lamp was turned off, and that the photocurrent densities maintained stable values when the lamp was turned on, indicating a rapid photocurrent response to the on-off intermittent irradiation. It could be obviously seen that the photocurrent responses of the g-C₃N₄/Ag₂CrO₄ nanocomposites were increased significantly in comparison with that of pure g-C₃N₄, indicating a higher separation and transfer efficiency of the photogenerated electron-hole pairs under visible light irradiation, and hence higher photocatalytic activity. Additionally, it was worth noting that the photocurrent density first increased, then declined along with the increasing content of Ag₂CrO₄, which was in well accordance with the above PL results. This provided further evidence to support the above PL results. Hence, the abovementioned results obviously confirmed the superior charge transfer and recombination inhibition in the g-C₃N₄/Ag₂CrO₄ composites in comparison with only g-C₃N₄, demonstrating that the introduction of Ag₂CrO₄ could effectively enhance the separation and transfer efficiency of photogenerated electron-hole pairs of g-C₃N₄, which could improve the photocatalytic performance.

Mechanism of enhanced photoactivity

As far as bare g-C₃N₄ was concerned, normally, the photogenerated electron-hole pairs quickly recombined and only a fraction of them participated in water splitting reaction, resulting in low photocatalytic activity^5,6,15,21. Based on the above analysis of the experiment and characterization results, the possible photocatalytic reaction mechanism and electron transfer processes of the g-C₃N₄/Ag₂CrO₄ nanocomposites were proposed (Fig. 8). The conduction band potential (E_CB) and valence band potential (E_VB) of the semiconductors could be calculated by the following empirical equations⁷⁶:

$$E_{CB}=X-E^e-\frac{1}{2}{E}_g$$ 2

$$E_{VB}=E_{CB}+E_g$$ 3

where X was the absolute electronegativity of the semiconductor, which is the geometric mean of the electronegativity of the constituent atoms, and the values of the X for g-C₃N₄ and Ag₂CrO₄ are 4.72 eV and 5.83 eV, respectively^52,77. E^e was the energy of free electrons on the hydrogen scale (∼4.50 eV vs NHE). Therefore, the E_CB potentials of g-C₃N₄ and Ag₂CrO₄ were −1.15 eV and +0.46 eV, and corresponding E_VB potentials could be estimated to be +1.58 eV and +2.20 eV, respectively. The values of VB for g-C₃N₄ and Ag₂CrO₄ were also obtained by XPS and the results for them were 1.60 eV and 2.20 eV respectively (shown as Fig. S3), which were similar with the results of calculation ones. Under visible light irradiation, Ag₂CrO₄ and g-C₃N₄ were both excited and yield electron-hole pairs. According to the double-charge transfer theory (as shown in Fig. 8a)^78–80, the photo-generated electrons of g-C₃N₄ would migrate to the conduction band (CB) of Ag₂CrO₄ and focused on the Pt nanoparticales, then the holes of Ag₂CrO₄ would transfer to the valence band (VB) of g-C₃N₄, because the VB and CB potentials of Ag₂CrO₄ were both lower than those of g-C₃N₄. However, if the composites followed the double charge transfer theory, the electrons in the CB of Ag₂CrO₄ could not reduce H⁺ to generate H• due to the CB potential of Ag₂CrO₄ (0.46 eV/vs. NHE) was higher than E₀ (H⁺/H₂) (0 eV/vs. NHE). Therefore, a Z-scheme mechanism was proposed based on the above analysis. As illustrated in Fig. 8b, the photo-generated electrons on the CB of Ag₂CrO₄ could easily transfer to the VB of g-C₃N₄ and reacted with the holes of g-C₃N₄, effectively inhibiting the electron-hole pairs recombination in both Ag₂CrO₄ and g-C₃N₄, prolonging the lifetime of the photogenerated electrons on the CB of g-C₃N₄ and the photogenerated holes on the VB of Ag₂CrO₄, so enhancing the interfacial charge transfer. Furthermore, Pt as co-catalyst could accept and transfer electrons and functioning as an effective hydrogen evolution promoter for g-C₃N₄. In such a way, the photogenerated electrons and holes were efficiently separated by the Z-scheme charge transfer; it could be further confirmed by the PL spectra and photocurrent analysis. Subsequently, the photogenerated electrons left at the CB of g-C₃N₄ which have more negative potential than the standard redox potential of H⁺/H₂ (0 eV vs. NHE) could reduce H⁺ to yield H• which was the source to produce hydrogen. Holes stored in the VB of Ag₂CrO₄ could directly react with sacrificial agent (methanol).

Figure 8.

Schematic representation of the charge generation, migration and hydrogen production mechanism of Z-scheme g-C₃N₄/Ag₂CrO₄ nanocomposites.

Conclusion

In summary, an efficient visible-light-driven bio-inspired Z-scheme g-C₃N₄/Ag₂CrO₄ heterostructure nanocomposite had been successfully fabricated via a facile chemical precipitation method and was applied in the photocatalytic H₂ generation. The obtained g-C₃N₄/Ag₂CrO₄ photocatalysts exhibited excellent hydrogen evolution efficiency in comparing with the individual Ag₂CrO₄ and g-C₃N₄ under visible light irradiation (λ ≥ 420 nm). The composite with the optimal mass ratio of 23.1% exhibited the highest photocatalytic activity, which could be achieved 902.1 μmol g⁻¹ h⁻¹. The enhanced photocatalytic activityis ascribed to the formation of the Z-scheme g-C₃N₄/Ag₂CrO₄ heterostructures which possessed higher separation and transfer efficiencies of the photogenerated electron-hole pairs. Moreover, owing to the firmly combination between Ag₂CrO₄ and g-C₃N₄ in heterostructures, the photocorrosion of Ag₂CrO₄ nanoparticles was strongly suppressed. This study might help to understand the mechanism of the g-C₃N₄/silver composites and provided a new insight to the design of the Z-scheme heterostructures.

Experimental Section

Preparation

All reagents (except the urea was chemical grade) were analytical grade and used without further purification. The product of bulk g-C₃N₄ was obtained by a simple calcination method⁸¹. Typically, the precursor urea (10 g) was calcined at 600 °C for 4 h with a ramp rate of 5 °C/min in a covered alumina crucible in order to prevent sublimation of urea and kept the calcinations took place in a static air atmosphere. The obtained light-yellow powder was washed and dried at 50 °C in a vacuum oven. After that the light-yellow products were milled into powder in an agate mortar for further experiments.

The g-C₃N₄/Ag₂CrO₄ composites were prepared by an in-situ chemical precipitation method under room temperature. Typically, 100 mg as-prepared g-C₃N₄ samples were immersed into 40 mL ultra-pure water and ultrasonically dispersed for 15 min. After that, 17.6 mg of K₂CrO₄ was added to the suspension under stirring at room temperature for 1 h, and then was treated using ultrasonic treatment for 1 h to ensure that CrO₄²⁻ was fully adsorbed on the surface of g-C₃N₄ sheets. Afterwards, an aqueous solution of AgNO₃ (30.8 mg in 20 mL of water) was drop wise added to the suspension under stirring. After stirring for another 2 h, the precipitate was collected by centrifugation and washed with ethanol and water for several times, and dried at 60 °C for 12 h under vacuum condition. Finally, the g-C₃N₄/Ag₂CrO₄ composite with a theoretical weight ratio of Ag₂CrO₄ to g-C₃N₄ at 30:100 was prepared and named as g-C₃N₄/Ag₂CrO₄(23.1%). The g-C₃N₄/Ag₂CrO₄ composites with different mass ratios were fabricated by changing the addition amount of AgNO₃ and K₂CrO₄ solution. The samples were marked as g-C₃N₄/Ag₂CrO₄(9.1%), g-C₃N₄/Ag₂CrO₄(16.7%), g-C₃N₄/Ag₂CrO₄(28.6%), and g-C₃N₄/Ag₂CrO₄(33.3%). The Ag₂CrO₄ was synthesized by the similar method but without g-C₃N₄.

Characterization

The phase structures were analyzed by X-ray diffraction (XRD) with a powder diffractometer (XRD-6000, Shimadzu, Japan). The topographies were observed by using transmission electron microscopy (TEM, Tecnai G² 20 S-TWIN, FEI, America). Energy dispersive spectroscopy (EDS) was obtained from a scanning electron microscopy (JSM 7500F, Japan Electron Optics Laboratory Co., Ltd., Japan) equipped with EDS attachment (INCA Energy 250, Oxford, America). Chemical compositions of the samples were also analyzed using X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi, Thermo Fisher, America). Uv-visible diffuse reflectance spectra (UV-vis DRS) were collected on a spectrophotometer (UV-3600, Shimadzu, Japan). Fourier-transform infrared (FT-IR) spectra were obtained on a FT-IR spectrophotometer (Nicolet iS10 IR, Thermo Scientific, America). Photoluminescence (PL) spectra were recorded on a fluorescence spectrometer (F-4500, Hitachi, Japan) with photomultiplier tube voltage of 400 V and scanning speed of 240 nm/min. The photocurrent experiments were performed on an electrochemical workstation (CHI 660D, Shanghai Chen Hua Instrument Co., Ltd., China) in a standard three-electrode system with Pt plate and Ag/AgCl (saturated KCl) electrode as counter electrode and reference electrode, respectively. A solar simulator illumination (CXE-350, Beijing Aodite Photoelectronic Technology Co., Ltd., China) at intensity of 100 mW cm⁻² was used. The working electrode was prepared as follows: 10 mg powder was dispersed ultrasonically in 0.5 mL of mixture solution (water: ethanol: Nafion (5%) = 230:250:20). Then, 20 μL of the resulting colloidal dispersion (20 mg/mL) was drop-cast onto a piece of ITO with a fixed area of 1 cm × 1 cm, and the electrodes were dried under ambient for 4 h.

Photocatalytic activity of H₂ evolution

Photocatalytic H₂-evolution experiments were performed in a gas-closed circulation system (Labsolar-6A, Beijing Perfectlight Technology Co., Ltd., China) with a top-irradiation quartz vessel. In a typical experiment, 50 mg of the as-prepared photocatalyst was dispersed in 100 mL mixed solution containing 75 mL water and 25 mL methanol (with 10 mM NaHCO₃). 0.6% of co-catalyst Pt was deposited onto the surface of the photocatalyst by in-site photo-deposition when a certain amount of H₂PtCl₆ solution was added. Before irradiation, the air in the system was removed by a vacuum pump. A 300 W Xe lamp (light intensity: 200 mW cm⁻², Microsolar300, Beijing Perfectlight Technology Co., Ltd., China) with a 420 nm cut off filter was used as the visible light source. The produced hydrogen was in situ detected periodically using an online gas chromatograph (GC7900, Tech-comp Shanghai Co., Ltd., China) with argon as carried gas.

Author Contributions

Y.C. and J.Z. supervised the project. Y.C. designed and carried out all experiments. B.L. and Q.Q. helped with the experiments in electrochemical test, and with the SEM measurements. H.C. helped with the TEM measurements. K.W. and Z.L. helped with the analysis of the experimental results. Y.C. wrote the manuscript with important input from all authors.

Competing Interests

The authors declare no competing interests.

Electronic supplementary material

Supplementary information accompanies this paper at 10.1038/s41598-018-34287-w.