Firstly, the Infrared spectrum of S-C3N4, ZnIn2S4 and ZSC are described in Fig. 1a. The main characteristic absorption bands of 9% ZSC and 18% ZSC are similar with those of S-C3N4. The angular peak at 812 cm− 1 is attributed to the out-of-plane bending vibration of C-H, which is related to the triazine structure in S-C3N4. And 1628 and 1244 cm− 1 are the characteristic stretching vibrations of C = C double bonds in aromatic rings (Wang et al. 2009). However, due to the ZnIn2S4 is an inorganic compound with no significant infrared characteristic peaks, herein, the absorption bands of ZSC cannot reveal the existence of ZnIn2S4. Therefore, further analysis and detection are needed.
The elemental composition and valence conditions of 18% ZSC was disclosed by X-ray photoelectron spectroscopy (XPS). As shown in the XPS survey spectrum (Fig. 2a), the main element are C, N, S, In and Zn, which explains the hybrid of ZnIn2S4 with S-C3N4 was successful. For C 1s Fig. 2b, two main signals centered at 281.0 and 284.1 eV were related to = C of graphite and C = N units in triazine skeleton, respectively (Jiang et al. 2015). The N 1s Fig. 2c (394.8, 399.9 and 403.0 eV) was fitted with N-C, N-Hx and N = C, respectively (Zhang et al. 2016). Correspondingly, 166.0 eV in Fig. 2d (S 2p) were ascribed to S 2p1/2. 449.7 and 457.2 eV in Fig. 2e (In 3d) were associated with In 3d5/2 and In 3d3/2. For Zn 2p spectra, main peaks located at 1027.2 and 1050.3 eV (Fig. 2f) corresponds to Zn 2p3/2 and 2p1/2. Noteworthy, the bonding power of each element in ZnIn2S4 were consistent with those reported in the references, which demonstrated ZnIn2S4 is hybridized with S-C3N4 successfully (Shi et al. 2015).
For better analyzing the microstructure of ZSC sample, the TEM and HRTEM were employed in Fig. 4a and 4b. As shown in Fig. 4a, the darker nanosheets can be attributed to ZnIn2S4, while the lighter part correlated to S-C3N4 (Guo et al. 2017). The HRTEM in Fig. 4b revealed the crystal-line with an planar spacing of 0.32 and 0.40 nm, corresponding to the (006) and (102) peak of ZnIn2S4 (Guo et al. 2017). In addition, the boundary conjugated by Znln2S4 and S-C3N4 can also be seen clearly from Fig. 4b. In Fig. 4c, the EDS elemental mapping clearly verifies that the element composition of ZSC are unanimous with the result of XPS, including C, N, Zn, In and S respectively. Moreover, form Fig. 4d-h, it can be seen that distribution of all the elements are homogeneous and the position of each unit conforms to the composite. All these phenomena illustrated that the combination of ZnIn2S4 and S-C3N4 was successful and the structure of ZnIn2S4 was well preserved, which played a crucial role for improving the photocatalytic performance of ZSC.
3.2. Photocatalytic performance
The photocatalytic efficiency of different ratio of ZSC heterojunctions were evaluated by using tetracycline hydrochloride (TC, 20 mg/L, 30 mL) as pharmaceutical residue and irradiating with visible light (Fig. 5a). Firstly, the blank control trial showed that TC solvent is stable under the visible light. Without photocatalyst ZSC, only 12.6% self-degradation rate can be neglected. Secondly, compared with pure S-C3N4 samples, the ZSC composites show remarkably increased photocatalytic activities to degrading the tetracycline, indicating the formation of ZSC heterojunction is an efficient solution to upgrade the photo-catalysis level of S-C3N4. The highest photocatalytic efficiency is belongs to the 18% ZSC, which can remove 86.3% TC within 120 minute, higher than 50% degradation efficiency of pure S-C3N4. Furthermore, the ln(C/C0) curve exhibits a good linear relationship (Fig. 5b), illustrating that the photoinduced degradation of TC solution catalyzed by ZSC conforms to the first-order dynamic model. In addition, the time variation of tetracycline concentration (Fig. 5c) can directly prove that TC molecules were decomposed irradiating by visible light in the presence of 18% ZSC. With the extension of light irradiation, the absorption intensity of tetracycline hydrochloride at 370 nm drastically decreased and blue shift occurred. For further survey the application prospect of ZSC, the stability of this heterojunction under the photolysis of TC was also evaluated. As illustrated in Fig. 5d, after three successive runs, the degradation efficiency of TC with 18% ZSC can still be remained above 80%. Even recycle five times, the degradability of TC was still remained at 73.4%. These experimental data show that the as-constructed 18% ZSC possesses fine stability and potential utilization for eliminating antibiotics contaminants.
3.3. Photocatalytic mechanism analysis
To deeply understand the activity of composite photocatalyst 18% ZSC under visible light, some experiments like UV-vis/DRS, PL emission, photo-electrochemical, ESR properties are carried out to assess the charge carrier dissociation and transfer rate of the composites S-C3N4, ZnIn2S4 and 18% ZSC.
In Fig. 6a, it is obvious that the absorbed boundary of S-C3N4 is about 475 nm, slightly higher than 460 nm of pure C3N4, because of the doping effect of S (Long et al. 2020). Compared with S-C3N4, the ZnIn2S4 as a semiconductor with UV-visible light response offers a wide absorption (300–500 nm), with an absorption boundary around 570 nm, although the intensity of absorption peak is not as high as S-C3N4 in the ultraviolet region. Notably, when incorporation of ZnIn2S4 within S-C3N4, the obtained absorption spectra exhibit all the characteristics of ZnIn2S4 and S-C3N4. The optical absorption almost covers the entire spectral range. And although the intensity of absorption in 450–800 nm is not stronger than ZnIn2S4 and S-C3N4, the absorbed boundary of composite displays a significant red-shift to 650 nm, which is benefit to improve the utilization of visible light for photocatalysts.
Generally, the fluorescence intensity is related to the re-compounding of charge carriers. Therefore, photoluminescence (PL) spectrum of ZnIn2S4, S-C3N4 and 18% ZSC were investigated and displayed in Fig. 6b. Obviously, ZnIn2S4 has a strong absorption band in a broad spectral range from 400 to 600 nm. In case of S–C3N4, a weaker emission peak centered at 475 nm is observed, which is consistent with the previous literature reported (Shi et al. 2014). For 18% ZSC, the PL peak is much weaker than ZnIn2S4 and S-C3N4, which alludes that the synergetic effect is occurred between the two semiconductors. It suggests that the recombination rate of photogenerated carries has been inhibited significantly, which is a reasonable explanation for the improvement of photocatalytic efficiency of 18% ZSC. In addition, this result is further validated by photocurrent time curves in Fig. 6c. In comparison to ZnIn2S4 and S-C3N4, the 18% ZSC heterojunctions emerge a remarkable increased photocurrent, illustrating the 18% ZSC heterojunctions can act as the effective mediators for promoting photoinduced electron-hole pairs separation and diffusion. Simultaneously, the electrochemical impedance spectroscopy (EIS) of 18% ZSC, ZnIn2S4 and S-C3N4 are delineated in Fig. 6d. The much smaller arc size of 18% ZSC discloses the resistance of photoinduced electron-hole pairs migration within the 18% ZSC electrode is much lower than ZnIn2S4 and S-C3N4 electrodes. All these above experimental results are consistent with the most efficient photodegradation effect of 18% ZSC, originated from the more effective charge carrier partition and electron transport between S-C3N4 and ZnIn2S4.
To further expose the photocatalytic mechanism of 18% ZSC, radical trapping experiments are conducted to reveal the primary active components in photodegradation. In Fig. 7a, it can be illustrated that the degradability of TC is reduced obviously to 10.4% after adding EDTA-2Na, demonstrating that the photo-excited cavity (h+) plays a vital effect in the degradation of TC. In addition, adding BQ directly results the elimination to 50.9%, revealing that ⋅O2− also contribute greatly to photodegradation. After adding IPA, the eliminated efficiency of TC slightly decreases to 77.2%, indicating that ·OH species has slight affected on the photolysis of TC. In short, the order of contribution to degradation efficiency is h+ > ⋅O2− > ·OH. To further verify the contribution rate of ·OH to degradation, we designed the fluorescence experiment to detect ·OH, according to the principle that p-phthalic acid can react on ·OH to generate some strong fluorescent substances. In Fig. 7b, we find a weak fluorescence peak at 430 nm in absence of photocatalyst, illustrating that the content of ·OH was very small before photodegradation. After adding catalyst, as shown in Fig. 7c, although strong fluorescence absorption peaks appeared at 340 nm and 370 nm, there was no characteristic fluorescence absorption at 430 nm, possibly due to the appearance of another strong fluorescent substance in the reaction system. This is in agreement with the result of trapping test, suggesting that ·OH is not the primary active components for TC photolysis catalyzed by 18% ZSC.
Moreover, the ESR trapping experiment were used to evaluate the efficiency of superoxide anion (⋅O2−) and photo-excited cavity (h+) in the TC degradation. In Fig. 7d, the characteristic absorption of DMPO-⋅O2− components appeared under visible light cannot be detected in the dark. By contrast, as in Fig. 7e, the peaks of h+ with same intensity can be seen either under visible light or in dark. This results further confirm that photo-excited cavity is the crucial active components in the TC photocatalysis, which is in concordance with the experiment data obtained from the radical capture experiments above.
To further estimate the photocatalytic degradation mechanism of ZSC, the band gaps and VB-XPS of catalysts were tested and exposed at Fig. 8. As shown in Fig. 8a and b, the bandgap width of the ZnIn2S4 and S-C3N4 were measured individually to be 2.46 and 2.68 eV. In Fig. 8c and d, the valence band potential of S-C3N4 and ZnIn2S4 were 1.85 and 1.37 V. Based on the band gap and valence band potential obtained above, the conduction band potential of ZnIn2S4 and S-C3N4 can be calculated as -1.09 and − 0.83 V, respectively.
According to the data required before, the possible photocatalysis principle of ZSC composites is proposed in Fig. 9. With the compact hetero-interface and well-matched band structure, a local field electric filed would be established between S-C3N4 and ZnIn2S4, which leading to the remarkably accelerating the segregation of charge (e−) -cavity (h+) pairs. When irradiated through visible spectrum, the electronics on the valence band of ZnIn2S4 is motivated to its conduction band, and then transferred to the conduction band of S-C3N4. In the meantime, the h+ is transferred from the valence band of S-C3N4 to ZnIn2S4. This process conforms to the charge transfer mechanism of type II heterojunction. Such rapid segregation of charge-cavity pairs would lead to the accumulation of h+ in VB of ZnIn2S4 and form the oxidation centers. Meanwhile, the electronics would be stored in conduction band of S-C3N4 to become reduction sites. All these would significantly decrease the reassociation of electronic charge-cavity pairs, and upgrade the photocatalysis effect. Because the conduction band potential of S-C3N4 is more negative than O2/∙O2− couple (-0.33 V), the electrons of S-C3N4 can better react with O2 to form ∙O2−, which can be partially converted to ∙OH (He et al. 2022), then degrade tetracycline. In addition, the oxidation voltage of ∙OH/OH− (1.99 V) and ∙OH/H2O (2.68 V) are more positive than the VB of S-C3N4 (1.87 V) and Znln2S4 (1.37 V). So the ∙OH radicals cannot be produced directly either from OH− or H2O by oxidation reaction. Considering the continuous accumulation of h+ at the heterointerface, which can directly oxide the pollutants, as well as partially consumed ∙O2−, it can be sure that the forming amount of these radicals in the photodegradation of tetracycline is: h+> ⋅O2− > ·OH. These was unanimous with the results of actives species detection experiments.