1. Introduction
The demand for energy is rising steadily as the world's population grows and living standards improve. Hydrogen is considered a clean, plentiful, and secure energy source to address this need [
1]. The tremendous energy output of hydrogen combustion, which is far higher than that of gasoline or any other fossil fuel, makes it a better and more efficient alternate fuel. As no toxic byproducts are produced during hydrogen combustion it is also considered ecologically safe [
2]. However, carbon dioxide is usually produced during the steam reforming of hydrocarbons and coal for hydrogen production. To avoid producing greenhouse gases, finding workable alternatives is essential. One viable solution to the present energy and environmental dilemma is using solar energy to produce hydrogen from water on the surface of a catalyst [
3,
4]. Semiconductor photocatalysts have been utilized widely for the photolysis of water, since their first use on the surface of TiO
2[
5]. To maximize the use of solar power, various attempts have been made to find renewable, efficient photocatalysts with an excellent visible light response [
6,
7].
The carbon nitride (
g-C
3N
4) graphitic material has been used as a C-related and a potential candidate with characteristics of metal-free photocatalyst in hydrogen evolution and organic degradation due to its suitable band gap (ca. 2.7 eV) [
8]. Additionally, the electronic structure of the triazine units in g-C
3N
4 forms conjugated graphitic planes, which are very stable and responsive to visible light. However, its photocatalytic performance is severely impacted by both the negligible or no absorption under the visible portion of light irradiation (beyond 460 nm) & the fast recombination for the photo-induced charge carriers species [
9]. To enhance the catalytic performance and promote the separation of photo-generated holes and electrons, another semiconductor is usually coupled with
g-C
3N
4, such as
g-C
3N
4/CdS,
g-C
3N
4/TiO
2,
g-C
3N
4/MoO
3,
g-C
3N
4/BiVO
4, and
g-C
3N
4/InVO
4[
10,
11,
12,
13,
14]. However, the complex preparation process and catalyst deterioration over a few cycles make it harder to use on a broader industrial scale.
Among other alternatives, solid sulfide solutions such as ZnIn
2S
4, CdIn
2S
4, ZnCdS, and Mn
xCd
1-xS have been used in photocatalytic hydrogen production because of their appropriate band gap, high visible light response, and tunable structure [
15,
16,
17,
18]. The easily tunable band structure and superior reducing ability of ZnCdS mean it stands out among the solid sulfide solutions [
19]. However, it does have several drawbacks, including inadequate photo-generated carrier transmission efficiency, low solar energy consumption, and rapid electron-hole pair recombination, which severely restricts its photocatalytic efficacy [
20,
21]. The charge recombination efficiency and energy output can be improved by combining ZnCdS with another photocatalyst, offering more active sites and reaction sites to promote oxidation and reduction processes.
Herein, we demonstrated a simple strategy to fabricate g-C3N4 with ZnCdS to form a type II heterojunction. The conjugated graphitic planes of g-C3N4 nanosheets provided a large surface area for ZnCdS, which acted efficiently to use the charge carrier and enhance H2 production.
3. Results
The morphologies of the as-prepared samples were determined by transmission electron microscope (TEM). As can be seen in
Figure 1 and
Figure S1,
g-C
3N
4 exhibited planer nanosheet structure. ZnCdS were small irregular shaped nanoparticles distributed onto the nanosheets (yellow circles in
Figure 1a). The high-resolution TEM image shows clear lattice dispersing of
g-C
3N
4 and ZnCdS, with a value of 0.321 nm and 0.332 nm corresponding to a (002) plane distance (
Figure 1b).
The XRD results of
g-C
3N
4, ZnCdS, and
g-C
3N
4‒ZnCdS heterojunctionsare depicted in
Figure 2. As shown in the case of pristine
g-C
3N
4, the peak at 27.9° for the (002) diffraction plane was derived from interplanar stacking peaks of conjugated aromatic systems of C
3N
4. The peak was well matched with JCPDS # of 87-1526 of
g-C
3N
4[
23].The XRD results of ZnCdS showed peaks indexed at 27.34°, 45.32°, and 53.66° corresponding to the (111), (220), and (311) planes of the cubic phase of Zinc blend related to (ICSD # 80-0020). In the
g-C
3N
4‒ZnCdS heterojunction, there were no clear diffraction peaks of ZnCdS because of its relative low levels and smaller size compared to
g-C
3N
4 [
22].
X-ray photoelectron spectroscopic (XPS) analysis was applied for determination of the elemental composition of the prepared catalyst and chemical state of particular elements. The survey spectrum of the
g-C
3N
4‒ZnCdS heterojunctions shown in
Figure S2 indicates the sample primarily comprised C, N, Zn, Cd, and S elements. To further illustrate the elemental signal, high-resolution XPS spectra are provided in
Figure 3. In the high-resolution XPS spectra of C1s shown in
Figure 3a, the peak positioned at 284.8 eV can be related to sp
2 carbon atoms (C-C and N-C=N bonding) originating from the surface exotic C in the instrument. The 2
nd peak located at 288.3 eV can be attributed to sp
3 hybridized C-bonded to nitrogen [C-(N)
3 of
g-C
3N
4]. The high resolution XPS spectrum of N1s shows a large peak centered at 398.8 eV which can be ascribed to a nitrogen atom bonded to carbon [C-N=C], while the shoulder peak at 401.1 eV can readily be ascribed to N-(C)
3 and N-H [
24,
25] (
Figure 3b).The high resolution XPS spectrum of Cd3d shows two spin-orbit components centered at 405.2 eV and 411.9 eV, which corresponded to Cd3d
5/2 and Cd3d
3/2, respectively (
Figure 3b). Similarly, the Zn2p region also showed components indexed at 1022 eV and 1045 eV, ascribed to Zn2p
3/2and Zn2p
1/2, respectively (
Figure 3c).
Figure 3d denoted the high-resolution spectra of S2p that exhibit a peak centered at 161.87 eV, that is attributed to the S
2− valent state of S in the ZnCdS segment. All the peaks are well-matched with the values reported previously for ZnCdS [
26].
The optical properties of pure
g-C
3N
4 nanosheets, ZnCdS nanoparticles, and
g-C
3N
4‒ZnCdS heterojunctions were measured with UV-vis DRS. As shown in
Figure 4a, the characteristic absorption peak of pure
g-C
3N
4 nanosheets was at about 400 nm, arising from the intrinsic band gap of
g-C
3N
4 at about 2.7 eV, that has low visible light absorption characteristic itself. On the other hand, ZnCdS showed strong absorption towards the visible region and the absorption edge extended towards 500 nm. After introducing ZnCdS nanoparticles into g-C
3N
4 nanosheets, heterojunction formation showed increased absorption intensity compared to bare
g-C
3N
4 nanosheets, and the absorption edge also moved towards the visible region.
Figure 4B displays the related
Tauc plots and all samples was examined for evaluation of the band gap energies. Band gap energy were obtained from analysis of the plot and the intercept of the tangent of the curve (αhν)² vs. (hν) on the X-axis, as previously reported [
27]. The calculated band gap energies for samples of
g-C
3N
4 nanosheets, ZnCdS nanoparticles, and
g-C
3N
4‒ZnCdS heterojunctions were determined from
Tauc plot to be 2.72 eV, 2.25 eV, and 2.60 eV, respectively (
Figure 4b). After the introduction of ZnCdS nanoparticles, the large band gap of
g-C
3N
4 nanosheets decreased, which supported its photocatalytic performance.
The photocatalytic hydrogen evolution ability of bare
g-C
3N
4 nanosheets, ZnCdS nanoparticles, and
g-C
3N
4‒ZnCdS heterojunctions was evaluated under visible light irradiation, as shown in
Figure 5a. The H
2 production rate for
g-C
3N
4 was observed at 13.5 μmol·g
-1·h
-1. In comparison, ZnCdS was 45.3 μmol·g
-1·h
-1. Compared to bare samples, the
g-C
3N
4‒ZnCdS heterojunction showed an increase in photocatalytic H
2 production (108.9 μmol·g
-1·h
-1). This was about eight times higher than
g-C
3N
4 and 2.4 times higher than ZnCdS. The increase in the H
2 production rate indicated that a heterojunction formed between the individual components, which facilitated the mobility of the charge carrier and enhanced the photocatalytic performance. Another issue to be considered in the applicability of photocatalysts is their performance in reusability. Therefore, reusability experiments were performed for
g-C
3N
4‒ZnCdS heterojunctions, and after each run, the catalyst was recovered by centrifugation, washed with water and ethanol, and reused. As displayed in
Figure 5b, the hydrogen generation rate was remarkably stable over five cycles (94% retention rate with a value decrease from 108.9 to 102.3 μmol·g
-1·h
-1), indicating the excellent stability and sustainable utilization of the photocatalyst.
In order to explore how the
g-C
3N
4‒ZnCdS heterojunction shows better photocatalytic performance compared to individual components, the electron transfer mechanism was revealed, as shown in
Figure 6. The conduction and valence band potentials of a photocatalyst can be calculated by the following equations:
Where E
VB and E
CB represent the valence and conduction band potentials, respectively. E
g is the band-gap energy, E
e denotes the energy of the free electrons on the hydrogen scale, and X represents the absolute electronegativity [
28]. By using equations 1 and 2, the conduction and valence band potentials for ZnCdS were determinated as E
VB = 1.85 eV and E
CB = -0.39 eV, respectively, while for
g-C
3N
4, they were calculated as E
VB = 1.59 eV and E
CB = -1.13 eV. Both the valence and conduction bands for ZnCdS were lower compared to
g-C
3N
4 , which facilitates the formation of type II heterojunctions. Upon visible light irradiation, both ZnCdS and
g-C
3N
4 can be excited, and then electrons from the CB of
g-C
3N
4 can be transferred into the CB of ZnCdS, which then react with H
+ for H
2 production. At the same time, the photo-induced holes of
g-C
3N
4 and ZnCdS can be used by oxidizing agents [
29,
30].
The photocatalytic conjecture was further verified by transient photocurrent response. It was recorded for
g-C
3N
4, ZnCdS, and
g-C
3N
4‒ZnCdS heterojunctions.
Figure 7A shows
I-t curves for as-synthesized electrodes film with five ON-OFF intermittent visiblelight irradiation consecutive cycles [
31]. The responses of photocurrent were appeared in all the electrodes instantly as the light was turned on, then rapidly declined to zero (nearly) as the light was off, which was reproducible and stable. With the similar conditions of irradiation, the photo-current value of the ZnCdS electrode was about twice to that of bare
g-C
3N
4, suggesting there was low-recombination and fast migration of photogenerated electron on the
g-C
3N
4 nanosheets. Additionally, after heterojunction formation between individual components,
g-C
3N
4‒ZnCdS showed a much higher photocurrent value by about 2.6 times, confirming the photogenerated electrons from the
g-C
3N
4 were attacking part in the electron transfer process and shifted to the CB of ZnCdS efficiently.
EIS is another efficient technique to observe the charge transfer efficiency and the interface reaction ability, which explains charge transfer resistance [
32].
Figure 7B shows the Nyquist plots of
g-C
3N
4, ZnCdS, and
g-C
3N
4‒ZnCdS heterojunctions. The smaller diameter implied a low impedance and fast interface charge transfer. The
g-C
3N
4‒ZnCdS heterojunction had the smallest diameter compared to bare samples, which also showed less charge transfer resistance and coincided well with photocurrent response results. Overall, our results showed that the heterojunction formation between
g-C
3N
4 and ZnCdS enables less recombination and faster photogenerated electron migration, resulting in a higher photocatalytic performance and enhanced durability.