Introduction

With the abuse of antibiotics, the random disposal of drug waste and the wanton discharge of drug wastewater, the purification of wastewater containing antibiotics has become a problem of human concern [1]. Semiconductors-based photocatalysis is considered as a promising and effective technology to remove organic contaminants in the aqueous solution because of its simple equipment and easy control without secondary pollution [2,3,4,5,6,7,8,9,10,11]. Up to the present, various metal oxide semiconductors have been developed as photocatalysts, such as TiO2, ZnO, SnO2, Bi2O3, BiVO4 and ZnSnO3 [12,13,14,15,16]. Among them, ZnSnO3 has attracted great interest in the photocatalytic field due to its high electron mobility, fast electrical conductivity and strong negative reduction potential of excited electrons [17, 18]. However, like most wide-band-gap semiconductors, ZnSnO3 cannot be excited to generate photogenerated carriers in the visible region, which leads to practical limitation [19].

Considerable research efforts have been devoted to expanding the photocatalytic activity of ZnSnO3 in the range of visible light, which include doping elements to broaden the visible light absorbance [20] and carbon-based material sensitization to enhance the visible-light-response range [21]. Besides, coupling the ZnSnO3 with the visible-light-driven semiconductors to form heterojunctions is also an advisable strategy [22]. Recently, graphitic carbon nitride (g-C3N4) has been attracted much attention because of its relatively narrow band gap, low cost, high chemical stability and innocuity [23,24,25,26,27]. In recent years, many reports of employing g-C3N4 to improve the photocatalytic performance of wide-band-gap semiconductors have been published. For example, Tan et al. modified the TiO2 with g-C3N4, which expands the light absorption of the photocatalysts and exhibits the favorable photocatalytic H2 evolution efficiency under visible light [28]. Tang’s group synthesized g-C3N4/ZnO and displayed the excellent visible light photodegradable efficiency by degrading 94% of Rhodamine B (RhB) in two hours, which is mainly caused by the heterostructure between g-C3N4 and ZnO; the photoexcited electrons in CB of g-C3N4 can be transferred into the CB of ZnO, thus effectively separating electrons and holes [29]. Zhang et al. fabricated visible-light-response Bi2O2CO3/g-C3N4 heterojunction photocatalysts which showed greatly improved activity for removing the three model pollutants of CH3CHO, RhB and Cr (VI) [30]. Concerning the research on g-C3N4/ZnSnO3 composite, Chen’s group has utilized zero-dimensional (0D) ZnSnO3 nanoparticles to coat on the surface of two-dimensional (2D) g-C3N4 nanosheets forming the g-C3N4/ZnSnO3 heterojunction to effectively enhance its photocatalytic activity under visible light irradiation [31]. However, with respect to the contact area at the interface of the heterojunction, the reactive sites of this point contact formed by 0D/2D heterojunction are often less than that of the face-contact heterojunction [32]. Thus, in order to further enhance the photocatalytic activity, the construction of g-C3N4/ZnSnO3 heterojunction with face-contact is more reasonable.

As is known, to design and develop the core–shell structure could provide more channels for charge transport and more surface reactive sites. For example, Hou et al. prepared BiVO4@TiO2 core–shell hybrid mesoporous nanofibers for photocatalytic hydrogen production, exhibiting the superior photocatalytic H2 production efficiency benefiting from the high specific surface area with increased reaction sites [33]. Meanwhile, Liu’s group prepared the Cd2Ge2O6/CdS core–shell structure photocatalysts, which significantly improve the charge migration capability due to the large-area contact of core–shell structure [34]. In addition, the construction of core–shell structure can also improve the stability of the core material by covering with a stable shell material. An et al. prepared the core–shell Ag2CO3@g-C3N4 photocatalyst, aiming at the prevention of the photo-corrosion of Ag2CO3 through the outer g-C3N4 shell in the composite [35]. Similarly, Jia et al. reported hybrid core–shell microgels with Cu2O nanocubes as the core and thermosensitive PNIPAM as the shell, which presents excellent colloidal stability [36]. Consequently, systematic design and controllable prepared g-C3N4/ZnSnO3 with core–shell nanostructure is expected to achieve the high active visible-light-driven heterojunction photocatalyst.

In this work, face-contact ZnSnO3@g-C3N4 core–shell heterojunction was successfully constructed via the calcination route. This composite photocatalyst is used for degrading TC under visible light irradiation (λ > 420 nm). It not only expands the absorption of visible light but also effectively separates the electrons and holes; thus, the photocatalytic degradation performance of TC is greatly improved. Moreover, the possible degradation mechanism over ZnSnO3@g-C3N4 was investigated through active species capture and electron spin resonance (ESR) experiments.

Experimental section

Synthesis of g-C3N4

Typically, 2 g melamine was in a covered crucible heated to 550 °C maintained for 3 h with a ramp speed of 5 °C/min. After cooling to room temperature, the resulting solids were ground to obtain the g-C3N4 powder.

Synthesis of ZnSn(OH)6 cubes

ZnSn(OH)6 cubes were prepared using a typical synthetic method according to the previous report [16]. First, 0.158 g of SnCl4·5H2O was dissolved in 20 mL of deionized water and stirred at room temperature. The pH of the mixed solution was adjusted to 12.0 by adding NaOH solution, and 0.1295 g of ZnSO4·7H2O was added under stirring at 60 °C until the appearance of a white precipitate. It was then aged at 60 °C for 12 h, cooled to room temperature and washed six times with distilled water and absolute ethanol to remove residual ions. Finally, the white precipitate was dried at 60 °C to obtain ZnSn(OH)6 cubes powder.

Synthesis of ZnSnO3@g-C3N4

The synthetic process of ZnSnO3@g-C3N4 is shown in Scheme 1. Briefly, 200 mg of ZnSn(OH)6 cubes was immersed in 20 mL deionized water, and then a certain mass of cyanamide solution (50% aqueous solution) was added. Next, the mixed solution was sonicated at 60 °C for 2 h and stirred for 12 h. It was then centrifuged, dried and calcined at 580 °C for 2 h. The obtained ZnSnO3@g-C3N4 samples were expressed as ZSO@CN (X), where X refers to the mass ratio of ZnSn(OH)6 and cyanamide. The pure ZnSnO3 sample was fabricated under the same conditions in the absence of cyanamide and labeled as ZSO.

Scheme 1
scheme 1

Schematic illustration of the fabrication process over the ZSO@CN core–shell composite photocatalyst

Full size image

The characterizations and photocatalytic experiments can be listed in the Supporting Information.

Results and discussion

The morphologies of ZSO and ZSO@CN (1:3) were observed by scanning electron microscope (SEM) and transmission electron microscope (TEM). From the SEM image of ZSO (Fig. 1a), it can be seen that the ZSO possesses a cube-like morphology with smooth surfaces and the average size of around 200 nm. After introducing the CN, the morphology of ZSO@CN composite is hardly changed in comparison with pure ZSO (Fig. 1b). And the specific morphologies of the synthesized samples were further employed by the TEM measurement. As can be seen in Fig. 1c, d, the lattice spacing on the crystal we measured is 0.262 nm, which corresponds to the (110) plane of ZnSnO3 nanocubes [37]. Notably, for the ZSO@CN composite, the layer of CN is coated on the outer surface of nanocube-like ZSO material, which is approximately 4.0 nm (Fig. 1e, f) to form a distinct core–shell structure.

Figure 1
figure 1

SEM images of a ZSO and b ZSO@CN (1:3). TEM and HRTEM images of c, d ZSO and e, f ZSO@CN (1:3)

Full size image

To estimate the distribution of the elements in the ZSO@CN composite, the HAADF-SEM and elements mapping were carried out. As shown in Fig. 2a–f, the HAADF-SEM and C, N, Zn, Sn and O elements are detected and all the elements are continuous and concentrating distribution, unambiguously identifying that CN has been successfully coated on the outer surface of ZSO cubic nanostructure, although the C and N elements in the elements mapping are not obvious due to the low content of CN on the outer surface.

Figure 2
figure 2

a HAADF-SEM and b–f elemental mapping images of ZSO@CN (1:3)

Full size image

In Fig. 3, X-ray diffraction (XRD) patterns of CN, ZSO and ZSO@CN (1:3) were recorded. It is observed that there are two characteristic diffraction peaks at (002) and (100) planes in the XRD pattern of CN, which were ascribed to in-plane structural packing motifs and the interlayer accumulation of aromatic system [38]. For the porous ZSO cubes, the characteristic diffraction peaks are assigned to the (012), (110) and (116) planes, respectively [39]. The XRD pattern of as-prepared ZSO@CN (1:3) sample almost shows the typical peaks of ZSO. However, the characteristic peaks of CN shell could not be observed in the XRD pattern of ZSO@CN (1:3) sample, which is due to the shell thickness that belongs to a short-range ordered structure and the limited amount and low intensity of CN [40].

Figure 3
figure 3

XRD patterns of CN, ZSO and ZSO@CN (1:3)

Full size image

As illustrated in Fig. 4, the Fourier transform infrared (FT-IR) spectra of ZSO, pure CN and ZSO@CN (1:3) samples are adopted. From the spectra of CN, several peaks between 1230 and 1650 cm−1 are observed, which ascribe to the aromatic C–N stretching and C=N stretching vibration modes [41]. The other two peaks about 808 and 3200–3400 cm−1 are observed separately in the spectra: The 808 cm−1 can be attributed to the breathing mode of triazine unit, whereas the 3200–3400 cm−1 to the stretching vibration of N–H [42]. For pristine ZSO, the peaks located at 475 and 600 cm−1 can be attributed to the vibration modes of Zn–O and Sn–O [43]. Similar to the XRD results, the main characteristic peaks of ZSO@CN (1:3) are basically consistent with ZSO in the FT-IR spectra. In the ZSO@CN (1:3) core–shell composite photocatalyst, it is difficult to observe the main characteristic peaks of CN, which is due to the low content of CN in the ZSO@CN (1:3).

Figure 4
figure 4

FT-IR spectra of CN, ZSO and ZSO@CN (1:3)

Full size image

X-ray photoelectron spectroscopy (XPS) technique was performed for confirming the chemical valences and compositions of CN, ZSO and ZSO@CN (1:3), and the corresponding spectra are displayed in Fig. 5. Figure 5a shows two main elements (C and N) in the survey spectrum of CN and three main elements (Zn, Sn and O) in the survey spectrum of ZSO, respectively. However, there are no obvious diffraction peaks representing C and N elements in the survey spectrum of ZSO@CN, which is due to the little content of CN in ZSO@CN, but reflected the peak intensities of the C 1 s and N 1 s high-resolution XPS spectra. There are two obvious diffraction peaks located at 288.2 eV and 284.6 eV in the high-resolution XPS spectrum of C 1 s for CN (Fig. 5b), which belongs to sp2 C–C bonds and sp2-bonded carbon (N–C=N), respectively [44]. The N 1 s spectrum of CN (Fig. 5c) can be decomposed into three peaks located at 398.5, 399.2 and 400.7 eV, which are corresponded to C–N–C, N–(C)3 and N–H, respectively [29]. The Zn 2p spectrum of ZSO (Fig. 5d) should be deconvoluted by two peaks at 1021.6 and 1044.7 eV which are associated with Zn 2p3/2 and Zn 2p1/2. In the Sn 3d spectrum of ZSO (Fig. 5e), the peaks appear at 486.3 eV for Sn 3d3/2 and 494.8 eV for Sn 3d5/2, respectively. The O 1 s spectra of ZSO (Fig. 5f) centered at 530.1 and 531.4 eV are attributed to the metal oxide bonds and –OH groups, respectively [45]. From the high-resolution XPS spectra of ZSO@CN (1:3), the positive shifts of five elements are seen, which are due to the heterojunction formed successfully.

Figure 5
figure 5

a Survey XPS spectra of CN, ZSO and ZSO@CN (1:3). High-resolution XPS spectra of b C 1 s, c N 1 s, d Zn 2p, e Sn 3d and f O 1 s

Full size image

In order to study the optical properties of as-prepared photocatalysts, the UV–Vis diffuse reflectance spectra of CN, ZSO and ZSO@CN (1:3) samples are measured and displayed in Fig. 6a. It can be seen that pristine CN shows an obvious absorption edge at around 478 nm, and in the visible light range of 400–800 nm pure ZSO shows a weak absorption. When coupling ZSO core with CN shell, the absorption capacity of ZSO@CN (1:3) composite photocatalyst has been enhanced compared with pure phase ZSO, which indicated that the addition of CN could improve the utilization rate of ZSO to light, so it was beneficial to the improvement in photocatalytic performance. According to the Tauc formula, the band gap values of ZSO and CN can be calculated to be 3.25 and 2.73 eV (Fig. 6b), respectively.

Figure 6
figure 6

a UV–Vis diffuse reflectance spectra of CN, ZSO and ZSO@CN(1:3). b Band gap energies of as-prepared pure ZSO and CN

Full size image

The visible-light-driven photocatalytic performance of the prepared photocatalysts was tested by the degradation of TC solution. Before the photocatalytic experiment, in the dark for 30 min, all the samples reached the adsorption equilibrium. The photocatalytic test results (Fig. 7a) reveal that compared to individual CN (32% degradation) and ZSO (9% degradation) within the same irradiation time, the ZSO@CN core–shell composite photocatalysts owned the superior photocatalytic activity. Notably, the photocatalytic activity of the ZSO@CN core–shell structure depends on the quality of CN introduction. With the increase in CN mass content (from 1:0.5 to 1:3), the photocatalytic activity of ZSO@CN core–shell structure was enhanced, and ZSO@CN (1:3) exhibited the optimum TC degradation rate of 90.8% after 120-min irradiation. Further increasing the content of CN mass can lead to a decrease in the ZSO@CN core–shell structure. This is attributed to the excessive CN (1:5 and 1:10 in the core–shell structure when the amount of CN is higher than its optimum deposition) which could lead to agglomeration state of the CN surface under such conditions without forming an effective core–shell interface, which hinders the transfer of photogenerated electrons [46]. As shown in Fig. 7b, all prepared photocatalysts follow the pseudo-first-order rule. Based on the calculated k values, the ZSO@CN (1:3) owns the highest TC degradation k value (0.019 min−1), which is around 6.33 and 31.67 times higher than that of CN (0.003 min−1) and ZSO (0.0006 min−1), respectively. Besides, for exploring the mineralization degree of TC during the photocatalysis, total organic carbon (TOC) removal rate over the ZSO@CN under visible light illumination was tested. As exhibited in Fig.S1, the removal rate of TOC was only 0.5% without photocatalyst, while ZSO@CN (1:3) photocatalyst reduced TOC by 46.2%. The results display that the ZSO@CN composite had good mineralization efficiency for the degradation of TC. In order to further reveal the advantage of the as-prepared photocatalyst with surface contact, we carried out a comparative experiment of ZSO nanoparticles coated on the surface of CN nanosheets to form point-contact ZSO-CN based on previous report [31]. The results of TC degradation and the pseudo-first-order reaction kinetics and k values of the as-prepared photocatalysts with different contact modes for TC degradation were further investigated. As displayed in Fig. 7c, it can be seen that the degradation of TC solution of the ZSO-CN with face contact is always higher than that with point contact. From Fig. 7d, the face-contact ZSO-CN shows the higher TC degradation k value (0.019 min−1), which is about 2.71 times higher than that of point-contact ZSO-CN (0.007 min−1). It is obvious that the face-contact ZSO-CN photocatalyst displays much more contact interfaces than the point-contact ZSO-CN photocatalyst, and the contact interfaces could play the part of efficient channels for charge transfer [47], leading to the higher photocatalytic degradation performance of face-contact ZSO-CN composite photocatalyst.

Figure 7
figure 7

a TC degradation over as-prepared photocatalysts under visible light irradiation and b the corresponding pseudo-first-order reaction kinetics and rate constants. c TC degradation over ZSO-CN composite photocatalysts with different contact modes under visible light irradiation and d the corresponding pseudo-first-order reaction kinetics and rate constants

Full size image

The BET analysis was carried out, and the result is shown in Fig.S2. It can be seen that the specific surface area of ZSO@CN (40.3 m2 g−1) is higher than that of ZSO (35.1 m2 g−1). This result reveals that the increasing specific surface area in ZSO@CN can provide more activity sites and be beneficial to improve the photocatalytic activity. Photocurrent experiments were carried out to reveal the separation and migration properties of photoinduced electron–hole pairs. The results are shown in Fig. 8a. Because of the wide band gap of ZSO, it is hard to use visible light, and the photocurrent intensity of pristine ZSO is very low. Notably, the photocurrent intensity for pure CN is higher than that of ZSO, which is consistent with UV–Vis diffuse reflectance spectra. Evidently, the ZSO@CN (1:3) photocatalyst displayed the highest photocurrent intensity, implying that the core–shell heterostructure between CN and ZSO is beneficial to promote the separation of electrons and holes. Similarly, compared to other samples (Fig. 8b), the impedance arc radius of the ZSO@CN (1:3) photocatalyst was the smallest, which also verified that charge transfer is promoted in the core–shell heterostructure.

Figure 8
figure 8

a Transient photocurrent spectra and b EIS spectra of CN, ZSO and ZSO@CN (1:3)

Full size image

Additionally, nanosecond (ns) time-resolved fluorescence decay spectra were recorded to further investigate charge transfer and separation (Fig. 9a). The corresponding parameters are summarized in Table 1; the fluorescence peak decay of ZSO@CN (1:3) (τ1 = 1.08 ns, τ2 = 4.82 ns) was shorter than that of the CN with half-lives of τ1 = 2.05 ns and τ2 = 6.50 ns. Meanwhile, the corresponding average fluorescence lifetimes (τav) of CN and ZSO@CN (1:3) are obtained to be 5.21 and 3.22 ns, respectively. The shorter lifetime indicates that the recombination rate of photogenerated charge carriers decreases, which makes the separation efficiency of photoinduced charge higher, according to the previous reports [48]. Hence, the τav of ZSO@CN (1:3) is shorter than that of CN, which further proves that the interfacial charge transfer can promote charge separation. In order to assess the stability of ZSO@CN composite photocatalyst, the recycling experiment was performed and the result is exhibited in Fig. 9b. After four recycles, the photocatalytic degradation rate of TC remained above 83%, revealing that the ZSO@CN (1:3) composite possesses the outstanding photocatalytic stability.

Figure 9
figure 9

a Time-resolved transient PL decay spectra of CN and ZSO@CN (1:3). b Cycle stability tests of as-prepared ZSO@CN (1:3) photocatalyst

Full size image
Table 1 Fluorescence lifetime parameters summarized from the time-resolved PL spectra with fitting curves
Full size table

In order to confirm the main active species of ZSO@CN, different scavengers, such as nitrogen (N2), isopropyl (IPA) and disodium ethylenediamine tetra-acetic acid (EDTA), were added into the reaction solution for capturing superoxide radicals (·O2−), hydroxyl radicals (·OH) and holes (h+), respectively. As shown in Fig. 10a, the addition of IPA into the reaction system did not significantly inhibit the degradation rate of TC. However, after adding N2 and EDTA into the reaction system, the photocatalytic degradation rate of ZSO@CN distinctly decreased from 91 to 49% and 33%, respectively. As shown in Fig. 10b, taking ZSO@CN composite under visible light irradiation, electron spin resonance (ESR) test was carried out to further clarify the existence of ·O2− active substance in the photocatalysis. In the absence of light, there is no obvious peak of ·O2−. Under visible light irradiation, the peak value of the ·O2− crack of the ZSO@CN composite was significantly enhanced. In summary, as we all know, ·O2− and h+ are produced during the photocatalytic process and play the significant roles in the degradation of TC over ZSO@CN core–shell composite photocatalyst.

Figure 10
figure 10

a Photocatalytic activities of ZSO@CN for the degradation of TC in the presence of different scavengers. b DMPO spin-trapping ESR spectrum for ZSO@CN (1:3) in methanol dispersion for capturing the signals of DMPO-•O2−

Full size image

To further explore the main intermediate products during the photocatalysis over ZSO@CN (1:3) photocatalyst, the liquid chromatography–mass spectrometry (LC–MS) technique was performed and the corresponding spectrum is exhibited in Fig.S3. It is well known that the substance with a molecular weight of 445 is TC [49]. Moreover, m/z values of 410, 388, 337, 241, 201, 154 and 141 were also formed in the degradation process, which were labeled as TC2–TC8, respectively, and the possible degradation pathways are revealed in Fig. 11, through further consideration of intermediates. Based on previous reports [50, 51], three typical functional groups (phenolic group, double bond and amine group) were easily attacked by O2− and h+ [52, 53]. TC2 was formed through breaking bond reaction from TC. TC3 and TC4 were formed via dehydration reaction, losing N-methyl and ring opening reaction. The products TC5, TC6, TC7 and TC8 with smaller molecular weight were formed by ring opening reaction and the breaking bond reaction. The above reaction is actually produced by the synergistic effect of O2− and h +. Finally, most of the intermediate products would be degraded into CO2 and H2O.

Figure 11
figure 11

Possible TC degradation pathway of ZSO@CN core–shell photocatalyst

Full size image

According to the above analysis, the possible photocatalytic mechanism over ZSO@CN core–shell heterostructure photocatalyst was initially described and is discussed in Fig. 12. By comparing the energy levels of ZSO with CN, it is exactly right to find that the energy levels of ZSO and CN are well-matched overlapping band structures for forming the type II heterojunction [46]. When exposed to visible light, due to the narrower band gap of CN (2.73 eV), the photoinduced electrons and holes were produced, while for ZSO (Eg = 3.25 eV), it was difficult to produce photogenerated charges during the visible light due to its broader band gap. The suitable energy band positions of CN and ZSO were inferred and are drawn in the right of Fig. 12. Specifically, the photoexcited electrons could transfer from CB of CN (− 1.16 eV vs. NHE) to the CB of ZSO (− 0.15 eV vs. NHE) and left the photoinduced holes on the VB of CN; hence, the separation rate of photogenerated electron–hole pairs was promoted effectively. The CB potential of ZSO was more negative than O2/·O2− (− 0.046 eV vs. NHE); photogenerated electrons on the CB of ZSO can capture O2 to form ·O2− directly [54]. Owing to the VB of ·OH/OH− (2.38 eV vs. NHE) was higher CN (1.53 eV vs. NHE), therefore the h+ directly oxidizes TC in VB of CN through photocatalytic reaction. At the same time, the process can be described as follows:

$${\text{CN}} + {\text{hv}} \to {\text{CN}}\;\left( {{{\text{e}}^- } + {{\text{h}}^+ }} \right)$$
(1)
$${\text{CN}}\;\left( {{{\text{e}}^- } + {{\text{h}}^+ }} \right) + {\text{ZSO}} \to {\text{CN }}({{\text{h}}^+ }) + {\text{ZSO}}\;\left( {{{\text{e}}^- }} \right)$$
(2)
$${\text{ZSO}}\;\left( {{{\text{e}}^- }} \right) + {{\text{O}}_2} \to {^\cdot}{\text{O}}_2^-$$
(3)
$${^\cdot}{\text{O}}_2^- + {{\text{h}}^+ } + {\text{TC}} \to \;{\text{C}}{{\text{O}}_2} + {{\text{H}}_{2}}{\text{O}}$$
(4)
Figure 12
figure 12

Proposed possible photocatalytic degradation mechanism of TC over ZSO@CN core–shell heterostructure photocatalyst

Full size image

Conclusions

In summary, ZSO@CN core–shell heterojunction was successfully constructed via a simple calcination route to improve the photocatalytic performance for the degradation of TC under visible light irradiation. The results indicated that the photocatalytic activity of as-prepared face-contact core–shell ZSO@CN is higher than that of pure ZSO, CN and point-contact ZSO-CN. The reason for higher photocatalytic degradation performance of face-contact ZSO-CN composite photocatalyst is mainly ascribed to the core–shell structure between CN and ZSO, which could own much more contact interfaces than that of point-contact ZSO-CN photocatalyst that play the part of efficient channels for charge transfer, leading to the higher photocatalytic degradation performance of face-contact ZSO-CN composite photocatalyst. The capture experiments indicated that both ·O2− and h+ played significant roles in the photocatalysis. This study laid a foundation to prepare for more ZSO-containing composite photocatalysts and promote the purification of wastewater containing organic pollutants.