1. Introduction
Study and modification of photocatalytic properties of semiconductor nanostructured materials are among the most vibrant research trends in the last decade. Depending on the research problem, nanostructured photocatalysts can perform water splitting into hydrogen and oxygen, destruction of various organic pollutants, and reduce carbon dioxide to organic compounds induced by UV and visible light. Particular attention is paid to photocatalysts that absorb visible light and can be used in natural sunlight excitation conditions [
1,
2,
3,
4].
The presence of electron donors in water (and other compounds such as amines, alcohols, aldehydes, and carboxylic acids) intensifies formation of molecular hydrogen. In this regard, it is promising to use organic pollutants as electron donors, because they provide an opportunity for simultaneous hydrogen production and removal of organic pollutants contained in domestic and industrial wastewater [
5,
6].
There are various semiconductor materials used for photocatalytic reactions. Photocatalysts based on graphitic carbon nitride (g-C
3N
4) have great prospects for practical applications. They do not contain any toxic components and can be used both to produce hydrogen and decompose organic pollutants in water solutions. [
7,
8,
9,
10]. Recent studies demonstrated that the structures based on g-C
3N
4 are also effective in the reactions of decomposition of antimicrobial drugs and disinfection under the action of visible light [
11]. They can exhibit multifunctional properties such as hydrazine sensing and hydrogen production [
12]. They can be used in combination with photo-Fenton catalysts to remove persistent pollutants from water [
13].
A number of simple methods can be applied to synthesis of g-C
3N
4 photocatalysts using available reagents such as urea, thiourea, melamine, and cyanuric acid. If synthesis according to these methods is preformed, an initial bulk structure of g-C
3N
4 is formed. It further splits into small layers by physical or chemical exfoliation, which enhances photocatalytic properties of the resulting compounds [
7,
14,
15].
These compounds are capable of absorbing visible light. However, due to a charge recombination, their photocatalytic activity turns out to be low. In this regard, g-C
3N
4 materials are mostly used as components of heterostructures that also incorporate other compounds [
8,
9]. Among various types of heterojunctions, type II and Z-scheme compounds are the most important for increasing the photocatalytic efficiency [
7,
8,
9].
Development of such hybrids with other semiconductors is an approach that contributes to separation of charge carriers into different phases and increases recombination lifetimes. It also favors redox interactions of these charges with surrounding molecules.
To create such heterojunctions, it is important to consider that the positions of the band gaps of C3N4 and another semiconductor should be shifted relative to each other. For effective absorption of visible light, it is necessary for the second semiconductor to have a band gap no greater than the respective value of C3N4 (about 2.7 eV).
There are many types of semiconductors that satisfy these requirements. Among them, photocatalysts based on cadmium sulfide nanostructures are interesting due to their catalytic properties exhibited under visible light, especially in hydrogen evolution reactions [
16]. Synthesis of mixed CdZnS ‘solid solution’ semiconductors allows to control the bandgap of this material, and also considerably reduce surface photooxidation, which is a characteristic feature of CdS photocatalysts. Development of optimal synthesis methods is necessary for fabricating hybrid materials that are stable in photooxidation processes and possess good photocatalytic properties [
17].
It was previously shown that synthesis of g-C
3N
4 hybrids with CdZnS semiconductors leads to a considerable increase in photocatalytic activity of the resulting materials in reactions of both hydrogen evolution from water and decomposition of organic compounds [
18,
19].
Depending on synthesis conditions, g-C
3N
4 и CdZnS can form hybrids with type II and Z-scheme heterojunctions [
20]. Formation of Z-scheme heterojunctions results from a considerable shift of semiconductor bandgaps relative to each other. Such heterostructures can create electron-hole pairs with a high redox potential.
The work in [
19] reports synthesis of CdZnS/g-C
3N
4 heterostructures using a simple one-pot method to calcine the mixture of precursors represented by thiourea and metal salts. The resulting materials possess good photocatalytic properties. This synthesis method is of undoubtful interest. For applied research, however, it is necessary to consider alternative methods that require smaller amounts of thiourea because its decomposition produces a large amount of toxic sulfur-containing byproducts [
21].
Materials based on ZnIn
2S
4 can be considered as alternatives to cadmium-based catalysts. The bandgap of ZnIn
2S
4 (2.4 eV) is comparable to that of CdZnS (2.3–2.5 eV). This material also exhibits photocatalytic properties under visible light. The advantages of the ZnIn
2S
4 material are its high photostability and intrinsic nano- and microlayers that provide it with a high surface area [
22,
23].
Creation of g-C
3N
4 hybrid structures with ZnIn
2S
4 contributes to a more efficient charge separation and an additional increase in their photocatalytic activity. Previously, it was shown that ZnIn
2S
4/g-C
3N
4 heterostructures as well as CdZnS/g-C
3N
4 exhibit intensive photocatalytic properties in the reactions of water splitting and decomposition of organic compounds [
24,
25,
26].
At the same time, g-C3N4 and ZnIn2S4 bandgaps have a little shift relative to each other. Therefore, ZnIn2S4/g-C3N4 hybrids are predominantly characterized by type I or type II heterojunctions.
A hydrothermal and solvothermal methods are the most commonly used approaches to synthesize ZnIn2S4 photocatalysts and their hybrid structures. During the synthesis process, ZnIn2S4 nanolayers are formed on the surface of a previously synthesized semiconductor particles (for example, g-C3N4). At high concentrations of initial components, formation of a separate ZnIn2S4 phase is also possible.
At the same time, there are no known one-pot methods of calcining the initial components for synthesis of ZnIn2S4/g-C3N4 heterostructures. Similar to CdZnS/g-C3N4, it can be assumed that formation of ZnIn2S4/g-C3N4 with effective photocatalytic properties as a result of high-temperature decomposition of thiourea.
In this work, we demonstrated an affordable one-pot method for synthesis of materials based on g-C3N4 and CdZnS/g-C3N4 and ZnIn2S4/g-C3N4 heterostructures. Decomposition of rhodamine B in water solutions was used as a sample reaction for photocatalytic studies to compare the properties of all the studied substances in the same conditions.
2. Materials and Methods
Materials: Melamine (99%, Vitareactiv, Russia), thiourea (99%, Sigma Aldrich, St. Louis, MO, USA), cadmium acetate (II) dihydrate (98%, Sigma Aldrich, St. Louis, MO, USA), zinc acetate (II) dihydrate (99%, Reahim, Russia), indium acetate (III) (99%, Sigma Aldrich, St. Louis, MO, USA), rhodamine B (99%, Lenreactiv, Russia), methylene blue (99%, Lenreactiv, Russia), neonol AF 9–12 (technical grade, Nizhnekamskneftekhim, Russia), 1,4-benzoquinone (99%, Reahim, Russia), ammonium oxalate monohydrate (99%, Lenreactiv, Russia), and isopropyl alcohol (99%, Vekton, Russia).
g-C3N4 structures were obtained by thermal decomposition of melamine at a heating rate of 5 °C/min to 550 °C (2 h) and 1-h exposure at 550 °C under argon atmosphere. In order to obtain CdZnS/g-C3N4 heterostructures, 0.266 g (1 mmol) of cadmium acetate dihydrate, 0.220 (1 mmol) of zinc acetate dihydrate, 0.381 g (5 mmol) of thiourea and 1 g (8 mmol) of melamine were mixed and finely ground in a mortar. The resulting composition was then calcined similarly to melamine. ZnIn2S4/g-C3N4 heterostructures were obtained by mixing 0.11 g (0.5 mmol) of zinc acetate dihydrate, 0.292 (1 mmol) of indium acetate, 0.381 g (5 mmol) of thiourea and 1 g (8 mmol) of melamine. Samples of CdZnS and ZnIn2S4 were synthesized in similar conditions without the use of melamine.
Elemental analysis of finely ground powders of the obtained samples was carried out on a Vario EL CUBE setup and using a Bruker M4 Tornado XRF analyzer.
The surfaces of the particles were characterized using a Carl Zeiss Evo scanning electron microscope. Sample images were obtained using a secondary electron detector. The samples were studied at the probe current IProbe = 110 pA, at the accelerating voltage EHT = 6 kV.
Specific surface area studies were performed through the dye adsorption method. 40 mg of sample were dispersed in 200 mL of methylene blue solution (C = 0.02 g/L) vigorously stirred for 1 h to establish adsorption equilibrium. After that, the solution was left for 24 h in the dark for precipitation of the catalyst and then filtered through a 0.22 μm syringe filter. The concentration of the dye solution after achieving the equilibrium was determined by spectrophotometry at the 665 nm absorption maximum.
X-ray powder diffractograms were obtained on a Bruker D8 Advance diffractometer equipped with Vario attachment and Vantec linear PSD, using Cu radiation (40 kV, 40 mA) monochromated by the curved Johansson monochromator (λ Cu K
α1 1.5406 Å). Room-temperature data were collected in the reflection mode with a flat-plate sample. The samples were loaded on a silicon plate, which reduces background scattering. Patterns were recorded in the 2θ range between 3° and 95°, in 0.008° steps, with a step time of 1–5 s. Processing of the obtained data was performed using EVA software package [
27]. Full-profile analysis of diffraction data by the Rietveld method and estimation of the dimensional characteristics of nanoparticles crystallites were performed using the TOPAS software package [
28].
ICDD PDF-2 Release 2005 powder diffraction database was used for the identification of crystalline modifications.
UV-Vis absorption spectra were recorded by Perkin Elmer LAMBDA 35 spectrophotometer. Luminescence spectra were collected on Varian Cary Eclipse spectrofluorimeter. FTIR spectra were obtained on a Bruker ALPHA-T S/N 102706 Fourier spectrometer. Hydrodynamic size of nanoparticles was measured by dynamic light scattering method at Malvern Zetasizer Nano-ZS setup.
Relative quantum yields of the samples were calculated with the relation to the reference system that was the solution of quinine sulfate in 0.1 M H2SO4 (QY = 0.54) at the excitation wavelength of 350 nm. The intensity of absorption in the form of an aqueous dispersion at a given wavelength did not exceed 0.05 a.u.
For photocatalytic studies, 5 mg (typically) of each sample were preliminarily sonicated in 15 mL of distilled water for 10 min (44 kHz, 100 W). The solution of rhodamine B (0.8 mL, 0.001 mol/L) was then added to the mixture and stirred in the dark for 30 min to reach adsorption equilibrium (T = 25 °C).
Photocatalytic studies were carried out by decomposition of rhodamine B (RhB) using a 50 W LED chip (COB module) with an emission peak of 452 nm as a radiation source (FWHM = 26 nm). The photodecomposition was controlled by sampling at certain time intervals. The temperature of the reaction mixture was maintained within 25 ± 0.3 °C. The samples were centrifuged at 3000 rpm for 15 min to precipitate the photocatalyst suspension.
3. Results and Discussion
3.1. One-Pot Synthesis of g-C3N4 Structures
Synthesis of g-C
3N
4 structures was carried out by decomposition of melamine at the temperature of 550 °C. CdZnS/g-C
3N
4 and ZnIn
2S
4/g-C
3N
4 heterostructures were synthesized from melamine and the mixture of acetates of these metals in excess of thiourea as a source of sulfur (
Figure 1).
Previously, a one-stage strategy was demonstrated for synthesis CdZnS/g-C
3N
4 heterostructures from thiourea [
19]. Thiourea decomposes at high temperatures by forming g-C
3N
4 and releasing considerable amounts of volatile products, including toxic sulfur-containing compounds. In our work, melamine was the main component used for synthesis of g-C
3N
4 structures, while thiourea was used as a source of sulfur for synthesis of CdZnS and ZnIn
2S
4. Based on the previous studies [
18,
19], we selected the optimal mass ratio of the initial components (approximately 3:1 of g-C
3N
4 and CdZnS or ZnIn
2S
4, respectively) for synthesis of the heterostructures.
3.2. Composition and Structure Characteristics
Formation of C
3N
4 structures was confirmed by elemental analysis (
Table 1). An excess amount of nitrogen and hydrogen in the studied components indicates the presence of residual amine groups of melamine. A small amount of oxygen in the samples is due to the presence of adsorbed water and oxygen molecules, which take part in the photocatalytic process.
The molar ratio of metals in heterostructures differs from their amount in the initial precursors, which is probably due to determination of the elemental composition by X-ray fluorescence analysis. It can be reliably stated that the mass fraction of the g-C3N4 phase in the CdZnS/g-C3N4 and ZnIn2S4/g-C3N4 heterostructures is more than 70%, which also indicates that its molar fraction is predominant.
The morphology of the samples was investigated by scanning electron microscopy (SEM) (
Figure 2). At low magnification, formation of a lamellar g-C
3N
4 structure can be distinctly observed. The CdZnS/g-C
3N
4 and ZnIn
2S
4/g-C
3N
4 samples possess a more porous structure. Pores in the heterostructures emerge because of abundant volatile products that release during decomposition of thiourea and the salts of the respective metals. Absence of g-C
3N
4 sheets indicates a good mutual distribution of the two phases. At higher magnification, the g-C
3N
4 sample exhibits formation of large aggregates. In CdZnS/g-C
3N
4 and ZnIn
2S
4/g-C
3N
4 samples, the structure is represented by aggregates of smaller particles.
3.3. Specific Surface Area Measurements
The photocatalytic activity of materials depends significantly on their specific surface area (SSA). An increase in the photocatalytic efficiency of heterostructures can be associated not only with a more efficient charge separation, but also with an increase in the specific surface area of the materials.
The SSA values of the samples were determined by the methylene blue (MB) adsorption method described in [
29] as follows:
where N
A is the Avogadro constant (6.02‧10
23 mol
−1), S
MB is the surface coverage per molecule of MB (~1.35 nm
2), C
0 and C
E are the initial and equilibrium concentrations of MB, V is the volume of the MB solution, M
MB is the molecular mass of MB, and m
s is the weight of the sample, respectively.
The specific surface areas of the g-C3N4, CdZnS/g-C3N4 and ZnIn2S4/g-C3N4 compounds were found to be 15.3, 12.3 and 15.1 m2/g, respectively. The values obtained for all the samples are approximately equal, which is probably due to similar conditions of their synthesis. Low specific surface area values can be explained by the high-temperature synthesis method, which results in a denser fusion of different phases and a decrease in the porosity of the material.
3.4. XRD Studies
The samples of g-C
3N
4, CdZnS/g-C
3N
4, and ZnIn
2S
4/g-C
3N
4 were analyzed by X-ray powder diffraction method. The experimental diffraction pattern for pure g-C
3N
4 (
Figure 3a, black curve) shows interference peaks that are characteristic of the ordered structure of g-C
3N
4 [
7,
30]. According to the literature data for the structure of graphitic carbon nitride, a peak with an angular position of 2θ ~ 27.5⁰ corresponds to diffraction from crystallographic planes (002) and is determined by the ordering formed due to π…π stacking of trisubstituted heptazine fragments, while a less intense peak at 2θ ~ 12.9° (plane 100) characterizes the ordering of molecular structures in planes of these fragments.
Considering the heterostructure of the CdZnS/g-C
3N
4 type, the structure of the initial carbon nitride retains, and both peaks that are characteristic of carbon nitride are observed on the experimental curve (
Figure 3a, blue curve). Additional peaks observed in the diffraction pattern were analyzed using the powder database and identified as the ones representing the polycrystalline form of Zinc Cadmium Sulfide, Zn
0.5Cd
0.5S (ref. No. 01-089-2943) (
Figure 3b). In general, the observed broadening of the diffraction peaks on all the experimental curves indicates a favored formation of nanostructured materials with very small sizes of crystallites.
In the X-ray diffraction pattern of the ZnIn
2S
4/g-C
3N
4 heterostructure, the first peak (100) is absent (
Figure 3a, red curve). This may apparently indicate a substantial deterioration in ordering. At the same time, a comparative search in the PDF-2 database for other diffraction peaks on the experimental curve indicates not only the presence of a polycrystalline form of Indium Zinc Sulfide, In
2ZnS
4 (Ref.No. 03-065-2023) in this sample (
Figure 3c), but also a possible formation and presence of Zinc Indium Sulfide, Zn
3(In
2S
6) (Ref.No. 01-080-0835) (
Figure 3d).
For all the studied samples, we calculated the average sizes of both graphitic carbon nitride crystallites and other crystalline phases using the positions and shapes of the most intense diffraction peaks. All data are summarized in
Table S1. A full-profile analysis of the diffraction curves indicated that the average sizes of crystallites of all the crystalline phases in the samples did not exceed several nanometers. For the g-C
3N
4 phase crystallites, the crystallite sizes was found to be in the range of 1–10 nm and the crystallites themselves were noticeably anisometric. The sizes of crystallites in the ZnIn
2S
4/g-C
3N
4 phase were in the range of 4–11 nm, and the sizes of crystallites of CdZnS/g-C
3N
4 phase distributed in the range of 4.5–15.9 nm. The crystallites of both these phases were elongated in one direction.
3.5. FTIR and Optical Characteristics
On the FTIR spectra, all the samples exhibit absorption peaks that are characteristic of the g-C
3N
4 structures (
Figure 4). The broad band at 3000–3300 cm
−1 can be attributed to stretching vibrations of the N–H bond and residual surface amine and hydroxyl groups. Intense bands in the range of 1200–1650 cm
−1 can be attributed to the C–N bonds of the heterocyclic ring. A narrow peak at 808 cm
−1 corresponds to bending vibrations of triazine units. Similar absorption bands in the FTIR spectra of the CdZnS/g-C
3N
4 and ZnIn
2S
4/g-C
3N
4 heterostructures indicate the presence of a well-formed g-C
3N
4 phase in these samples.
All the samples can absorb light in broad UV and visible ranges (
Figure 5a). However, CdZnS/g-C
3N
4 and ZnIn
2S
4/g-C
3N
4 heterostructures demonstrate a more intensive absorbance in the visible region as compared with g-C
3N
4. The bandgap energy of direct transitions was calculated for the g-C
3N
4 structure by the Kubelka–Munk method from the Tauc plots and was found to be 2.76 eV. This value is consistent with the previous data [
19]. The CdZnS/g-C
3N
4 and ZnIn
2S
4/g-C
3N
4 samples showed the decrease in the bandgap energy to 2.59 and 2.62 eV, respectively (direct transitions) as the result of forming hybrid structures. The synthesized compounds are, therefore, capable of absorbing light more efficiently in the visible region of the spectrum. The band gaps of the individual CdZnS and ZnIn
2S
4 components were 2.43 and 2.51 eV, respectively, which agrees with the previous data [
19,
24].
The study of the luminescence spectra provided additional information on the efficiency of charge carrier separation in the synthesized semiconductor materials. The resulting structures exhibit a low-intensity luminescence. The g-C
3N
4 sample has a broad luminescence band with the maximum at 435 nm, the relative quantum yield was calculated to be 4.2% (
Figure 5b). In the heterostructures, the luminescence intensity is considerably reduced and the emission peak is shifted to the red region, while the relative quantum yield for CdZnS/g-C
3N
4 and ZnIn
2S
4/g-C
3N
4 samples was 2.2 and 1.8%, respectively. A decrease in the luminescence intensity can be explained by an increase in lifetimes of photoexcited states of charge carriers [
20,
24]. The resulting more efficient spatial separation in heterostructures favors interactions with surrounding molecules [
19,
24,
25].
3.6. Effect of Surfactants on the Photocatalytic Properties of g-C3N4
The obtained materials possess the properties of heterogeneous catalysts and their photocatalytic properties are highly dependent on the specific surface area. The simplest and the most common method of splitting of these structures into small layers (exfoliation) of graphitic C
3N
4 is an ultrasonic treatment. Various surfactants can be used to improve their dispersity. Previously, it was shown that addition of anionic or cationic surfactants to the media can improve photocatalytic properties of heterostructures based on g-C
3N
4 [
31]. It was reported that a positive effect is in this case provided by improved separation of charge carriers on the surfaces of the particles and preventing them from agglomeration during synthesis.
The effect of nonionic surfactants on the photocatalytic properties of materials based on g-C
3N
4 has not been adequately studied. As it was previously shown in our studies, ultrasonic treatment of dispersions of carbon nanostructures in the presence of a nonionic surfactant based on ethoxylated isononylphenol can lead to a significant increase in their stability and dispersity. The best effect is provided by surfactants with the oxyethylation degree
n = 8–12 [
32,
33].
In this work, we studied the effect of ultrasonic treatment of g-C3N4 structures in water in the presence of nonionic ethoxylated isononylphenol with the degree of oxyethylation n = 12 (neonol AF 9-12) on their degree of dispersion and photocatalytic properties.
It was found that the highest colloidal stability of the g-C
3N
4 dispersion in water within an hour after ultrasonic treatment is achieved with the surfactant concentration of 0.01% (
Figure 6a). According to the dynamic light scattering data, ultrasonic treatment in the presence of surfactants increased dispersity of the samples (
Figure 6b).
The photocatalytic properties of the synthesized structures were studied using the model reaction of rhodamine B (RhB) decomposition. The dye decomposition rate was monitored by the intensity of the RhB absorption peak at 554 nm (
Figure 6c). Based on these data, the kinetic curves of the decomposition reaction were plotted and the reaction rate constants were calculated according to the first order reaction rate law:
where A is the intensity of the RhB absorption peak at t time, A
0 is the intensity of RhB peak at the beginning of reaction.
The efficiency of photocatalytic decomposition of organic compounds is highly dependent on the intensity of light absorbed by a sample, which is determined by the concentration of a catalyst and reagents, the design of a reactor, and other parameters. We initially selected the optimal concentration of RhB for our studies and also determined the effective concentration of the photocatalyst in the reaction mixture.
The decomposition rate of RhB under irradiation without a catalyst was shown to be quite low (
Figure 6d). When 2 mg of the catalyst were added to the reaction mixture (0.013% of the total weight of the solution), we observed a noticeable increase in the dye decomposition reaction rate. The maximum increase in the photodecomposition rate was observed for 5 mg (0.033%) of g-C
3N
4. An increase in the catalyst concentration to 10 mg (0.067%) led to a decrease in the decomposition rate that was probably due to an intensive aggregation of g-C
3N
4 particles.
Dispersion with a nonionic surfactant demonstrated an additional enhancement of the photocatalytic properties of g-C
3N
4 (
Figure 6d). In this case, the influence of surfactants is more pronounced at higher concentrations of the catalyst in the reaction mixture. It was determined that the dispersion of 5 mg of the catalyst in the presence of surfactants provided the 12% increase in the reaction rate constant. With 10 mg of g-C
3N
4, the increase in the reaction rate constant was 54%. Nevertheless, the maximum efficiency of photocatalysis was observed with 5 mg of the added catalyst. This amount of the catalyst was, therefore, used in our further experiments.
3.7. Study of the Photocatalytic Activity of Heterostructures
The photocatalytic activity of the pure g-C3N4 structures is quite low, which can be explained by the presence of a large number of charge carrier recombination centers on their surfaces, where the photoexcitation energy may be lost. Therefore, it is necessary to spatially separate the excited electrons and holes. A possible solution is to create hybrid heterostructures with other materials.
At the next step of this work, we studied the photocatalytic properties of heterostructures based on g-C3N4 with a low content of the CdZnS and ZnIn2S4 semiconductor materials synthesized using a simple one-pot method.
The charge carrier separation efficiency was evaluated for these heterostructures from the energy diagram, which was plotted using the values of the band gap calculated for the individual components (
Figure 7).
The band edge positions of the valence and conduction bands of semiconductors can be calculated as follows [
34]:
where X is the absolute electronegativity of the semiconductor, E
e is the energy of free electrons on the hydrogen scale (4.5 eV), and E
g is the band gap energy of the semiconductor.
The X values for g-C
3N
4, CdZnS, and ZnIn
2S
4 were found to be 4.73, 5.32, and 4.86 eV, respectively [
25,
35]. The calculated values of E
CB and E
VB of the components are consistent with the data of the previous studies.
According to the diagram, the energy levels of individual semiconductors are shifted relative to each other, which allows to form heterojunction type II (
Figure 7, red arrows) or Z-scheme (
Figure 7, blue arrow).
Compounds with such bandgap properties can be used both for hydrogen production and photooxidation of organic compounds. However, the electropositivity of the photogenerated holes in both heterostructures is not sufficient for generating nonspecific OH· radicals directly from water (
Figure 7, blue dashed line). Therefore, oxidation is expected to predominantly occur via formation of O
2·
− radicals.
Experimental results show that the photocatalytic properties of heterostructures are significantly better than those of pure g-C
3N
4 (
Figure 8a). According to the calculated values, the RhB decomposition rate constant in the presence of ZnIn
2S
4/g-C
3N
4 and CdZnS/g-C
3N
4 was almost 2- and 3-fold higher, respectively, than the rate constant in presence of pure g-C
3N
4 (
Figure 8b).
A significant increase in the photocatalytic efficiency of heterostructures confirms formation of a heterojunction between two types of the semiconductor. According to the diagram, we assume formation of a heterojunction with the type-II electron migration in ZnIn
2S
4/g-C
3N
4 heterostructures (
Figure 7, red arrows). As a result of the transfer of charge carriers from one material to another in the heterostructures, the lifetime of the excited state of electrons and holes increases so as the probability of their transfer to the surface of the photocatalyst and interaction with surrounding molecules.
In the CdZnS/g-C
3N
4 heterostructures, the levels of the individual components are more substantially shifted from each other as compared with the ZnIn
2S
4/g-C
3N
4 sample. Therefore, two pathways of charge separation are possible. The first one is a type II transition of the charge carriers from one material to another that provides their free energy reduction (
Figure 7, red arrows). Another possible option is recombination of photoexcited electrons from a low-energy CdZnS with the g-C
3N
4 holes (
Figure 7, blue arrow). In this case, charge carriers with higher redox properties are involved in the photocatalytic process (Z-scheme).
The studied heterostructures have similar specific surface areas and their difference in photocatalytic properties is not related to surface effects. Close quantum yields of the samples also indicate similar separation efficiencies of their charge carriers.
Higher photocatalytic activities of the CdZnS/g-C
3N
4 samples may also be indicative of a Z-scheme heterojunction in such heterostructures. This suggestion is consistent with the results of the work reporting a similar synthesis approach to CdZnS/g-C
3N
4 heterostructures [
19]. A Type II charge separation would lead to formation of photogenerated electrons and holes with a lower redox potential as compared to the ZnIn
2S
4/g-C
3N
4 sample.
Since the conductive level of CdZnS (−0.4 eV) in our samples is more negative than the level required for formation of O2·− radical anions (−0.33 eV), radical trapping experiments cannot give a distinct prediction about a possibility of these processes. Therefore, we consider that both types of such a heterojunction are possible in the studied CdZnS/g-C3N4 heterostructures.
The respective characterization of the studied samples is summarized in
Table 2.
3.8. RhB Photocatalytic Degradation Mechanism
Interactions of photogenerated electrones and holes on the photocatalyst surface with the solvent (water) molecules lead to formation of highly active species capable of reacting with dye molecules in the reaction media [
7,
9,
36,
37]. Oxygen molecules adsorbed on the catalyst surface favor formation of O
2·
− radical anions, which are also involved in oxidation of organic compounds:
The mechanism of the RhB photodecomposition reactions was studied in radical trapping experiments. Benzoquinone, ammonium oxalate and isopropyl alcohol were used as scavengers for O
2·
−, h
+, and OH· active species [
38]. The values of the rate constants of these reactions are represented in the diagram (
Figure 9).
The results show a substantially slower degradation of RhB in the presence of benzoquinone, which indicates a direct participation of O2·− species in oxidation reactions. The decrease in the reaction rate constant was found to be 87% for CdZnS/g-C3N4 and 88% for ZnIn2S4/g-C3N4.
The reaction with ammonium oxalate as a hole acceptor does not lead to a significant change in the dye decomposition rate, which may indicate that decomposition of RhB by the holes is insignificant.
In the presence of isopropyl alcohol, the rate constant of RhB decomposition decreased moderately for both types of heterostructures, which indicates participation of OH· radicals in dye oxidation. According to the diagram (
Figure 7), the holes do not possess a sufficient positive potential to generate OH· radicals from water. Therefore, it can be assumed that OH· radicals are formed from adsorbed oxygen molecules in a series of reactions summarized in Equations (4)–(6).
Thus, we can conclude that the photocatalytic decomposition of RhB molecules in the presence of synthesized heterostructures occurs mainly through interactions of photoexcited electrons with surface oxygen molecules and formation of O2·− radicals. In general, the contribution of active species to the dye oxidation process can be described as follows: O2·− > OH· > h+.