Introduction

Nowadays, the arbitrarily discharged effluent from chemical industries is causing serious pollution to water resources. To remove toxic and carcinogenic substances (e.g., organic dyes) from industrial effluent has become a big research topic to protect our living environment. In recent years, photocatalysis technology based on semiconductor materials has gained much attention in the remediation of water resources.1,2,3,–4 The unique advantage of this effluent treatment technology is that it is capable of decomposing toxic organic dyes under solar irradiation. Solar irradiation excites semiconductor photocatalysts to generate electrons (e−) and holes (h+) in their conduction band (CB) and valence band (VB), respectively. Benefiting from their reduction (electrons) and oxidation (holes) capacities, the photoexcited electrons/holes directly or indirectly act upon the pollutants to cause their decomposition. Nevertheless, only a few photoexcited electrons/holes are available for the photodegradation reactions because of their high geminate recombination. In order to improve the photocatalytic activity of semiconductors, it is indispensable in increasing the probability of photoexcited carriers for participating in the photoreactions. Therefore, various avenues have been adopted to prolong the carrier lifetime by spatially separating the e−/h+ pairs.5,6,7,–8

Coupling two semiconductors into heterojunctions is one of the important strategies to inhibit the recombination of photoproduced electrons and holes.9,10,–11 As expected, an internal electric field is created at the interface of the semiconductor–semiconductor heterojunctions. Under the action of an internal electric field, the photogenerated electrons/holes will be driven from one semiconductor to another. This is the dominant mechanism decreasing the geminate e−/h+ recombination in the heterojunction composite photocatalysts. Another intriguing method to enhance the e−/h+ spatial separation is to decorate electron traps onto the semiconductor photocatalysts.12,13,14,15,–16 The most important electron traps include graphene, carbon nanofibers/nanotubes/quantum dots and noble metal nanoparticles. These nanosized carbon and noble metal materials are particularly interesting and have received much attention due to their fascinating physical properties and potential technological applications in many fields.17,18,19,–20 Moreover, the excellent electron-trapping ability of these nanosized materials makes it easier for them to trap the photoproduced electrons, thus reducing the e−/h+ recombination probability. These nanomaterials can also act as interesting co-catalysts. For example, nanosized noble metals commonly exhibit strong visible-light absorption due to their localized surface plasmon resonance (LSPR) effect, which implies that they can efficiently utilize visible light to drive the photodegradation reactions.21,22 Development of ternary composite photocatalysts by simultaneously constructing heterojunctions and decorating nanosized noble metals is expected to be a promising strategy to achieve superior photodegradation performances.

Ti-oxide semiconductors have been received as the most important candidate photocatalysts in the application of industrial wastewater purification.7,23,24,–25 The perovskite-structured barium titanate (BaTiO3, BTO) is a typical example of Ti-oxide semiconductors, famously known for its piezoelectricity and ferroelectricity.26 In addition to the intriguing physical properties, BaTiO3 is particularly interesting due to its photodegradation capability for decomposing organic dyes.27,28 Moreover, BaTiO3 also stands out due to its high chemical/photo-corrosion resistance, good stability, abundant natural resources, low cost, environmental friendliness, and easy fabrication. These advantages suggest that BaTiO3 is an excellent candidate for potential applications in environmental purification. However, the large band gap energy of BaTiO3 (~ 3.2 eV) limits its photoactivity to the ultraviolet (UV) region, which occupies only a small fraction of solar energy.29,30 In contrast, another promising photocatalyst–covellite (CuS) has a relatively narrower band gap (~ 2.0 eV), making it photoactive in a wide solar spectrum.31,32 For both pristine BaTiO3 and CuS, their photocatalytic efficiencies are not satisfying owing to the easy e−/h+ recombination. To overcome this shortcoming, many avenues have been adopted to improve the photodegradation performances of BaTiO3 and CuS, e.g., construction of heterojunctions and decoration with noble metal nanoparticles.33,34,35,36,–37 It is noted that the CB/VB positions between BaTiO3 and CuS are staggered with each other, indicating that they can be coupled to form promising CuS/BTO heterojunctions.

In the present work, we have constructed CuS/BTO heterojunction photocatalysts by incorporating smaller-sized CuS nanoparticles onto larger-sized BaTiO3 polyhedral particles. The effect of the CuS content on the photodegradation performance of the CuS/BTO composites was investigated and the optimal composite photocatalyst was determined. Then, the optimal CuS/BTO composite was decorated with Ag nanoparticles to obtain ternary CuS/Ag/BTO composite photocatalysts. Rhodamine B (RhB), known as a typical organic dye present in industry-produced wastewater, was used as the target pollutant to evaluate the photodegradation capabilities of the as-prepared composite photocatalysts under simulated-sunlight irradiation.

Experimental

Synthesis of BaTiO3 Polyhedral Particles and CuS Nanoparticles

Analytical-grade chemical reagents, derived from chemical reagent corporations in China, were used in the present experiments. A hydrothermal reaction process was employed to synthesize BaTiO3 polyhedral particles. First, three types of solutions, separately denoted as A, B and C, were prepared. Solution A was obtained by dissolving 1.0454 g (4 mmol) of Ba(NO3)2 in 20 mL deionized water, solution B was prepared by dissolving 2.4688 g (44 mmol) of KOH in 20 mL deionized water, and solution C was derived by adding 10 mL of oleic acid and 1.0892 g (3.2 mmol) of tetra-n-butyl titanate (C16H36O4Ti) to 20 mL n-butanol. Every step mentioned above was accompanied by 30 min of magnetic stirring to ensure that a uniform solution was obtained. Next, solutions C and B were successively dropped into solution A with 30 min of magnetic stirring. Then, the resultant mixture solution was sealed in a 100-mL Teflon-lined autoclave and subjected to the hydrothermal reaction (200°C, 24 h). Finally, the precipitate produced in the autoclave was collected as BaTiO3 particles by centrifugation. To remove impurity phases and ions, the product was rinsed with 10% (V/V) dilute nitric acid (one time), deionized water (three times) and anhydrous alcohol (three times), followed by drying in an electrothermal constant-temperature dry oven (60°C, 12 h).

CuS nanoparticles were obtained via a coprecipitation synthesis method. Stoichiometric amounts of Cu(NO3)2·3H2O (0.4832 g) and Na2S·9H2O (0.4802 g) were dissolved in 30 mL and 20 mL ethylene glycol with magnetic stirring for 30 min, respectively. The obtained Na2S solution was slowly added into the Cu(NO3)2 solution, which was then magnetically stirred for 30 min at room temperature (20°C). The precipitated product was collected as CuS nanoparticles by centrifugation. A rinsing process using deionized water (three times) and anhydrous ethanol (three times) was required to remove impurity ions from the product.

Construction of CuS/BTO Composites

To prepare the 5%CuS/BTO composite (CuS occupies a mass fraction of 5% in the composite), 0.3 g of as-prepared BaTiO3 particles was loaded in 30 mL ethylene glycol (ultrasonic dispersion for 30 min), which was then dissolved with 0.0399 g of Cu(NO3)2·3H2O (magnetic stirring for 30 min). The Na2S solution, which was obtained by dissolving 0.0397 g of Na2S·9H2O in 20 mL ethylene glycol with 30 min magnetic stirring, was slowly dropped into the above suspension. The resultant mixture was magnetically stirred at room temperature for 30 min, during which CuS nanoparticles were crystallized and assembled on the surface of BaTiO3 particles. After washing and drying, the final 5%CuS/BTO composite was obtained. Based on the same route, the 3%CuS/BTO and 8%CuS/BTO composites were also prepared by loading different amounts of BaTiO3 particles in the precursor solution.

Construction of CuS/Ag/BTO Composites

A typical ternary 5%CuS/(1%Ag/BTO) composite photocatalyst was prepared as follows. First, the 1%Ag/BTO composite was prepared by decorating Ag nanoparticles on the surface of BaTiO3 particles via a photoreduction method. Next, 0.3 g of BaTiO3 particles and 0.025 g of ammonium oxalate (AO) were loaded in 80 mL deionized water with 30 min of ultrasonic dispersion and then 60 min of magnetic stirring. Then, 0.14 mL of 0.2 mol L−1 AgNO3 aqueous solution was added to the above suspension. A 15-W low-pressure mercury lamp (emitting UV light) was used to illuminate the resultant mixture for 30 min under magnetic stirring, during which Ag nanoparticles were formed on the BaTiO3 particle surface. The product was collected as the 1%Ag/BTO composite after rinsing and drying. Then, the derived 1%Ag/BTO composite was decorated with CuS nanoparticles to form the ternary 5%CuS/(1%Ag/BTO) composite. The decoration procedure was same to that for preparing 5%CuS/BTO, as elaborated above. The preparation process for the CuS/BTO and CuS/Ag/BTO composites is schematically depicted in Fig. 1.

Fig. 1
figure 1

Schematic preparation process for CuS/BTO and CuS/Ag/BTO composite photocatalysts.

Full size image

Characterization of Samples

Various material characterization technologies including x-ray powder diffraction (XRD), transmission/scanning electron microscopy (TEM/SEM), energy-dispersive x-ray spectroscopy (EDS), ultraviolet–visible (UV–Vis) diffuse reflectance spectroscopy (DRS), x-ray photoelectron spectroscopy (XPS) and photoluminescence (PL) spectroscopy were employed to elucidate the structures, morphologies, microstructures, optical absorption properties, chemical composition, elemental valence states and PL properties of the as-prepared photocatalysts. The XRD analysis was carried out on a D8 Advance x-ray diffractometer (λCu-kα = 0.15406 nm). A field-emission transmission electron microscope (JEM-1200EX) and a scanning electron microscope (JSM-6701F) were employed for the TEM/SEM investigations. UV–Vis DRS spectra were recorded on a TU-1901 double beam spectrophotometer. XPS spectra were derived from a multi-functional x-ray photoelectron spectrometer (PHI-5702). A RF-6000 fluorescence spectrophotometer was used to record the PL spectra (excitation wavelength: 290 nm).

A three-electrode structure including a platinum foil electrode as the counter electrode and a standard calomel electrode (SCE) as the reference electrode was employed for the photocurrent response and electrochemical impedance spectroscopy (EIS) measurements on a CST 350 electrochemical workstation. 0.1 mol L−1 Na2SO4 water solution was used as the electrolyte. A 200-W xenon lamp was used as the sunlight simulator. The procedures for the working electrode preparation and photoelectrochemical measurements were discussed in detail in our previously published papers.38

Photodegradation Experiments

The target contaminant selected was 5 mg L−1 RhB water solution to study the decomposition capacities of the as-prepared photocatalysts. The photocatalyst concentration was set as Cphotocatalyst = 0.3 g L−1 by loading 0.03 g of the photocatalyst into 100 mL RhB solution. The mixture solution was put into a 200-mL photoreactor and then magnetically stirred in the dark for 30 min to examine the RhB adsorption onto the photocatalyst. After that, the photodegradation experiment was initiated by irradiating the reaction solution with simulated sunlight emitted from a 200-W xenon lamp. The distance between the front of the lamp and the reaction solution is 15 cm. The photoreactor was cooled by circulating water so as to maintain the reaction solution at room temperature (20°C). As time progressed, the change of the RhB concentration was examined by withdrawing a small amount of the reaction solution (2.5 mL) and measuring its absorbance. After removing the photocatalyst by centrifugation, the absorbance of the reaction solution was measured by using a UV–Vis spectrophotometer (measuring wavelength: 554 nm). The degradation ratio (η) of RhB can be derived according to the relationship η = (C0 − Ct)/C0 × 100%, where C0 and Ct are the initial and residual RhB concentrations, respectively.

Results and Discussion

Figure 2a displays the XRD patterns recorded from the as-prepared CuS, BaTiO3, 5%CuS/BTO and 5%CuS/(1%Ag/BTO) photocatalysts. The XRD pattern of BaTiO3 presents the characteristic diffraction peaks at 22.2°, 31.6°, 38.9°, 45.2°, 50.9° and 56.2°, which match well with the (100), (110), (111), (200), (210) and (211) facets of the BaTiO3 cubic phase (PDF#79-2263), respectively. The XRD pattern of CuS shows the diffraction peaks agreeing perfectly with those of the hexagonal phase of CuS (PDF#06-0464). The XRD data suggest that bare BaTiO3 and CuS are crystallized into a pure cubic structure and a hexagonal structure, respectively. The XRD patterns of 5%CuS/BTO and 5%CuS/(1%Ag/BTO) are very similar to that of bare BaTiO3, where all the diffraction peaks can be indexed into the cubic BaTiO3. The diffraction peaks of CuS and Ag are not clearly detectable on the XRD patterns, which is due to their low contents in the composites. However, the apparent colors and optical absorption properties of the samples confirm the decoration of CuS and Ag onto BaTiO3. As seen from Fig. 2b, bare BaTiO3 and CuS appear in cream white and black, respectively. In contrast, the 5%CuS/BTO and 5%CuS/(1%Ag/BTO) composites present as gray, which is obviously deeper than that of bare BaTiO3. It is noted that the physical properties of materials depend highly on their light absorption characteristics, which are necessary to characterize.39,40 Figure 2c illustrates the UV–Vis DRS spectra of the samples, from which one can see that bare BaTiO3 has a weak absorption in the visible-light region, whereas a strong visible-light absorption is observed for bare CuS. The 5%CuS/BTO and 5%CuS/(1%Ag/BTO) composites display a visible-light absorption that is obviously stronger than bare BaTiO3. The differential curves of the UV–Vis DRS spectra can be employed to estimate the band gap (Eg) of BaTiO3 in the samples,41 as shown in Fig. 2d. According to the peak on the differential spectra (i.e. absorption edge), the band gap of BaTiO3 is obtained as 3.26 eV for bare BaTiO3 and 3.34 eV for 5%CuS/BTO and 5%CuS/(1%Ag/BTO). The slightly increased band gap of BaTiO3 in the composites may be ascribed to the interaction of BaTiO3 with CuS and Ag. The Tauc plot of (αhν)2 versus hν, which is transformed from the DRS spectrum of CuS, is used for determining the band gap of CuS.42 As seen from the inset in Fig. 2d, the band gap of CuS is obtained as 2.02 eV by extrapolating the linear region of the Tauc plot to the hν-axis.

Fig. 2
figure 2

XRD patterns (a), apparent colors (b), UV–Vis DRS spectra (c) and differential curves of the DRS spectra (d) of CuS, BaTiO3, 5%CuS/BTO and 5%CuS/(1%Ag/BTO). The inset in (d) shows the Tauc plot of (αhν)2 versus hν for CuS.

Full size image

SEM images were recorded from BaTiO3, CuS, 5%CuS/BTO and 5%CuS/(1%Ag/BTO) to reveal their morphologies and sizes, as illustrated in Fig. 3. It is observed from Fig. 3a that BaTiO3 is crystallized as polyhedral particles with diameter ranging from 50 nm to 110 nm. The SEM image presented in Fig. 3b reveals the synthesis of spherical CuS nanoparticles with size of 10–25 nm. The SEM image of 5%CuS/BTO (Fig. 3c) demonstrates the decoration of smaller-sized CuS nanoparticles onto larger-sized BaTiO3 polyhedra. As seen from Fig. 3d, the 5%CuS/(1%Ag/BTO) photocatalyst is composed of BaTiO3 polyhedra decorated with CuS and Ag nanoparticles.

Fig. 3
figure 3

SEM images of (a) BaTiO3, (b) CuS, (c) 5%CuS/BTO and (d) 5%CuS/(1%Ag/BTO).

Full size image

The microstructure of the 5%CuS/(1%Ag/BTO) composite was further investigated by means of TEM related techniques. Figure 4a illustrates the TEM image of the 5%CuS/(1%Ag/BTO) composite, revealing the good coupling of BaTiO3 polyhedra with CuS and Ag nanoparticles. The derived selected area electron diffraction (SAED) pattern (Fig. 4b) shows clear diffraction spots or rings, suggesting that the composite has a good crystallization degree. By comparing the diffraction spots/rings with the standard diffraction data of BaTiO3, CuS and Ag, it is found that they can be perfectly indexed in terms of BaTiO3, CuS and Ag. The measured high-resolution TEM (HRTEM) images, as depicted in Fig. 4c and d, further confirm the coupling of BaTiO3 polyhedra with CuS/Ag nanoparticles and their perfect crystalline features. The BaTiO3 polyhedral particles show the characteristic (110) lattice fringes with the d-spacing of 0.283 nm. The CuS nanoparticles are recognized by their characteristic lattice fringes with 2d008 = 0.409 nm or d103 = 0.281 nm. The nanoparticle with lattice fringes of 0.237 nm, matching perfectly with the d111 spacing of Ag, is confirmed to be the Ag nanoparticle.

Fig. 4
figure 4

TEM image (a), SAED pattern (b) and HRTEM images (c,d) of 5%CuS/(1%Ag/BTO).

Full size image

Figure 5a shows the TEM-EDS spectrum of 5%CuS/(1%Ag/BTO), suggesting the existence of Ba/Ti/Cu/S/O/Ag in the composite. The extremely high intensity of the Cu peaks may be induced by additional Cu from the microgrid holder.43 EDS elemental mapping was adopted to investigate the elemental distribution in the composite. Figure 5b shows the dark-field scanning TEM (DF-STEM) image of the 5%CuS/(1%Ag/BTO) composite. The area outlined by an orange rectangle in Fig. 5b is investigated by EDS elemental mapping. As seen from the derived mapping images in Fig. 5c–h, in addition to Ba/Ti/O uniformly existing in the composite, Cu/S/Ag are also distributed throughout the composite. The elemental mapping results are indicative of the coupling of BaTiO3 polyhedra with CuS and Ag nanoparticles.

Fig. 5
figure 5

EDS spectrum (a), DF-STEM image (b) and EDS elemental mapping images (c‒h) of 5%CuS/(1%Ag/BTO).

Full size image

XPS is a useful technology that can be used for the diagnosis of element valences.44 The XPS spectra of 5%CuS/(1%Ag/BTO) including Ba-3d, Ti-2p, O-1s, Cu-2p, S-2p and Ag-3d core-level XPS spectra are collected in order to reveal the valence states of the elements. The characteristic Ba-3d XPS spectrum presented in Fig. 6a (Ba-3d5/2 → 779.2 eV, Ba-3d3/2 → 794.6 eV) and Ti-2p XPS spectrum given in Fig. 6b (Ti-2p3/2 → 458.7 eV, Ba-2p1/2 → 464.4 eV) indicate the existence of Ba2+ and Ti4+ species in the composite.33,45 Two binding-energy peaks (530.0  eV and 531.7 eV) are observed on the O-1s XPS spectrum (Fig. 6c), of which the peak at 530.0 eV originates from the crystal lattice oxygen of BaTiO3,33,45 and the peak at 531.7 eV has contributions from the chemisorbed oxygen species.46 The Cu2+ species is confirmed by the Cu-2p XPS spectrum (Fig. 6d) with characteristic Cu-2p3/2 binding energy at 932.1 eV and Cu-2p1/2 binding energy at 952.1 eV.38 On the S-2p XPS spectrum (Fig. 6e), the peaks at 160.7 eV and 162.1 eV are characterized as the S2−-2p3/2 and S2−-2p1/2 binding energies, respectively. The characteristic peaks detected on the Ag-3d XPS spectrum (Fig. 6f) suggest that the Ag species appears in the metallic state (Ag-3d5/2 → 367.6 eV, Ba-3d3/2 → 373.6 eV).33

Fig. 6
figure 6

High-resolution XPS spectra of (a) Ba-3d, (b) Ti-2p, (c) O-1s, (d) Cu-2p, (e) S-2p and (f) Ag-3d core levels for 5%CuS/(1%Ag/BTO).

Full size image

The photoexcited carrier separation behavior is an important factor that significantly influences the photodegradation performance of photocatalysts, which can be characterized by PL spectroscopy.47,48 As seen from Fig. 7, the PL spectra collected from all the samples show a broad PL emission peak at 380 nm, but their intensities are different. The PL emission peak is related to the recombination behavior of photoexcited electrons and holes. It is obvious that the coupling of BaTiO3 with CuS and Ag results in the decrease in the PL emission peak when compared with that of bare BaTiO3; noticeably, the PL emission peak from 5%CuS/(1%Ag/BTO) becomes the weakest. This suggests that the highest photodegradation activity will be achieved for the 5%CuS/(1%Ag/BTO) composite due to its highest e−/h+ separation efficiency.

Fig. 7
figure 7

PL spectra of BaTiO3, 3%CuS/BTO, 5%CuS/BTO, 8%CuS/BTO and 5%CuS/(1%Ag/BTO).

Full size image

Photoelectrochemical measurement is an alternative method that can be used to elaborate the separation behavior of photoexcited charges in photocatalyststs. Transient photocurrent response curves recorded from BaTiO3, 3%CuS/BTO, 5%CuS/BTO, 8%CuS/BTO and 5%CuS/(1%Ag/BTO) are presented in Fig. 8a. It is seen that the intermittent irradiation of the photocatalyststs results in the periodic appearance of photocurrent in them. The observed photocurrent for the samples appears in the following order: BaTiO3 < 3%CuS/BTO < 8%CuS/BTO < 5%CuS/BTO < 5%CuS/(1%Ag/BTO). This implies that the e−/h+ separation efficiency is enhanced for the composite photocatalysts compared to bare BaTiO3; noticeably, the 5%CuS/(1%Ag/BTO) composite has the highest e−/h+ separation efficiency. The EIS Nyquist spectra are also recorded and displayed in Fig. 8b. The diameter of the semicircle observed in the region of high frequencies is induced by the charge-transfer resistance and can reflect the photoexcited charge separation efficiency in the photocatalyststs.49 The EIS observed result is basically in agreement with that obtained by PL spectroscopy and transient photocurrent spectroscopy.

Fig. 8
figure 8

Transient photocurrent-time curves (a) and EIS Nyquist spectra (b) of the photocatalysts BaTiO3, 3%CuS/BTO, 5%CuS/BTO, 8%CuS/BTO and 5%CuS/(1%Ag/BTO).

Full size image

Figure 9a presents the adsorption and photodegradation performances of all the as-prepared samples for RhB driven by simulated sunlight. It is seen that with 30 min of adsorption in the dark, 3.5% and 46.3% of RhB is adsorbed onto bare BaTiO3 and CuS photocatalysts, respectively. The extremely large RhB adsorption for CuS may be because the ‒OH groups attached to CuS have a strong electrostatic attraction to cationic RhB molecules.50 The blank experiment in the absence of photocatalysts shows that RhB appears to be fairly stable under simulated sunlight irradiation. The CuS/BTO and 5%CuS/(1%Ag/BTO) composite phocatalysts exhibit a dye adsorption higher than that for bare BaTiO3 due to the introduction of CuS. The photodegradation behaviors of the samples are elaborated by the decrease in the RhB concentration with increasing the irradiation time. The degradation ratio η of RhB by irradiating the photocatalytic systems for 120 min (Table I) is in the order: BaTiO3 < CuS < 3%CuS/BTO < 8%CuS/BTO < 5%CuS/BTO < 5%CuS/(1%Ag/BTO). The largest and second largest photodegradation activities are observed for 5%CuS/(1%Ag/BTO) and 5%CuS/BTO, respectively. It is obvious that the integration of BaTiO3 polyhedra with CuS and Ag nanoparticles can generate excellent composite photocatalysts with enhanced photoactivities. Figure 9b shows the plots of Ln(Ct/C0) against irradiation time t, which are linearly fitted according to the equation Ln(Ct/C0) = −kappt, where kapp is known as the apparent first-order reaction rate constant.33 The derived values of kapp are presented in Table I, which can qualitatively reflect the photodegradation capacities of the samples. It is seen that the photoactivity of the 5%CuS/(1%Ag/BTO) composite is approximately 3.3 and 2.0 times that of bare BaTiO3 and CuS, respectively. Among the photocatalysts developed in our research group, the 5%CuS/(1%Ag/BTO) composite photocatalyst does not exhibit the highest photodegradation performance, but it still has an advantage in photocatalytically degrading environmental pollutants, such as good photo-corrosive resistance.

Fig. 9
figure 9

(a) Adsorption and photodegradation plots of RhB over all the as-prepared samples. (b) Photodegradation kinetic plots for all samples. (c) Reusability of 5%CuS/(1%Ag/BTO) for photocatalytic degradation of RhB. (d) Influence of the scavengers on the photodegradation of RhB over 5%CuS/(1%Ag/BTO).

Full size image
Table I Photodegradation ratio (η) of RhB with 180 min reaction and apparent first-order reaction rate constant (kapp) for all the as-prepared samples.
Full size table

The 5%CuS/(1%Ag/BTO) composite was repeatedly employed for the photocatalytic decomposition of RhB in order to examine its reusability. The procedure for the recycling photocatalytic experiment was described in detail in our other publications.33 As seen from Fig. 9c, the composite photocatalyst maintains a high photodegradation of RhB after four times of repetitive use (η = 80.0%), and only 5.6% of the dye degradation is lost when compared with the first photodegradation cycling (η = 85.6%). This slight decrease in the dye degradation can be ascribed to the minor loss of photocatalyst during the recycling process. The recycling experiment suggests the 5%CuS/(1%Ag/BTO) composite photocatalyst exhibits a good photocatalytic reusability, which is superior to many other photocatalysts. For example, the highly active Ag-based semiconductor photocatalysts generally exhibit a poor photocatalytic reusability due to their photo-corrosion.51

To reveal the reactive species involved in the 5%CuS/(1%Ag/BTO) photoreactive system, isopropyl alcohol (IPA), methyl alcohol (MeOH) and benzoquinone (BQ) were employed to trap the hydroxyl (·OH), photoexcited h+ and superoxide (·O2−) radicals, respectively. The radical trapping experiments were performed according to the procedure elaborated in our other publications.52 As displayed in Fig. 9d, the photodegradation percentages of RhB (120 min photocatalysis) are observed to be 38.0%, 57.1% and 69.4% on the introduction of IPA, MeOH and BQ, respectively. It is obvious that all the scavengers cause an inhibition of the dye degradation, suggesting the dependence of the dye degradation on ·OH, h+ and ·O2− species.

Mott–Schottky method based on the electrochemical measurements in the dark can be available for the estimation of the CB and VB potentials of semiconductors.6,53 The Mott–Schottky plots derived from BaTiO3 and CuS (measured frequency 5000 Hz) are displayed in Fig. 10a and b, respectively. By linearly extrapolating the Mott–Schottky plots to the abscissa axis, the flat band potential (VFB) of BaTiO3 and CuS is obtained as − 0.92 and + 0.86 V versus SCE, respectively. The flat band potential of BaTiO3 approximately equals its CB edge potential due to its n-type semiconductivity (positive slope of the Mott–Schottky plot), whereas the flat band potential of CuS approximately equals its VB edge potential due to its p-type semiconductivity (negative slope of the Mott–Schottky plot). With reference to normal hydrogen electrode (NHE), the CB/VB potentials of BaTiO3 are estimated to be − 0.27/+ 2.99 V, and the CB/VB potentials of CuS are derived to be − 0.50/+ 1.52 V (the relationship between NHE and SCE: VNHE = VSCE + 0.059 × pH(=7) + 0.242).6

Fig. 10
figure 10

Mott–Schottky plots of (a) BaTiO3 and (b) CuS.

Full size image

The possible mechanisms of the binary CuS/BTO and ternary CuS/Ag/BTO composites for enhanced photodegradation of RhB are proposed in Fig. 11. When CuS nanoparticles are decorated onto BaTiO3 polyhedra, spontaneous electron diffusion from n-type BaTiO3 to p-type CuS (hole diffusion from p-type CuS to n-type BaTiO3) occurs at the constructed CuS/BTO heterojunctions. This carrier diffusion process generates an internal electric field at the CuS/BTO interface (positive charge centers at the interface of BaTiO3 and negative charge centers at the interface of CuS), which prevents the continuous carrier diffusion. Finally, the CuS/BTO heterojunctions reach a thermal equilibrium state. During the photocatalytic process, both BaTiO3 and CuS are photoexcited to produce VB holes and CB electrons. Considering that the CB and VB potentials of CuS are respectively negative to those of BaTiO3, the photoexcited electrons will be driven by the internal electric field from the CB of CuS to the CB of BaTiO3, and conversely, the photoexcited holes from the VB of BaTiO3 to the VB of CuS, as schematically depicted in Fig. 11a. This transfer process of electrons/holes prolongs their lifetime, consequently making them participate in the photodegradation reactions more efficiently. This is the dominant mechanism resulting in the enhanced photoactivities of the CuS/BTO composites compared to bare BaTiO3 and CuS. The accumulated electrons in the CB of BaTiO3 are coupled to O2 species to form ·O2− radicals, as expected, because of their sufficiently negative potential (− 0.27 V) with respect to E0 (O2/·O2−) = − 0.13 V versus NHE.54 It is noted that only some of the VB holes in BaTiO3 can be transferred to the VB of CuS because CuS occupies only a small fraction of the CuS/BTO composites. The remaining holes in the BaTiO3 VB possess a potential (+ 2.99 V) that is more positive than E0(OH−/·OH) = +1.99 V and E0(H2O/·OH) = +2.38 V versus NHE,54 suggesting that they are able to react with OH– or H2O species to produce ·OH radicals, whereas the accumulated holes in the VB of CuS directly participate in the oxidative degradation reactions with RhB.

Fig. 11
figure 11

Possible photodegradation mechanisms of (a) binary CuS/BTO composites and (b, c) ternary CuS/Ag/BTO composites.

Full size image

The introduction of Ag nanoparticles onto the CuS/BTO composites results in the further increase in their photodegradation performances. There are two locations for the Ag nanoparticles. One is that Ag nanoparticles are located on the surface of BaTiO3 polyhedra without contacting the CuS nanoparticles, as shown in Fig. 11b. Because of their excellent electron-trapping ability, Ag nanoparticles will capture the CB electrons in BaTiO3 to make the further reduction of the e−/h+ recombination probability in BaTiO3. Ag nanoparticles can be also photoexcited under simulated-sunlight irradiation to generate free electrons, which may play a role in the photoreactions. Furthermore, the Ag-LSPR-induced electromagnetic field can stimulate the excitation of BaTiO3 to produce additional electrons and holes. Another location for Ag nanoparticles sandwiched between BaTiO3 and CuS, as depicted in Fig. 11c. In this case, the CB electrons in both BaTiO3 and CuS are trapped by Ag nanoparticles, thus promoting the e−/h+ separation. Meanwhile, both BaTiO3 and CuS can benefit from the LSPR of Ag nanoparticles to produce more e−/h+ pairs.

Conclusions

In this work, CuS/BTO and CuS/Ag/BTO composite photocatalysts have been developed. In these composites, CuS and Ag nanoparticles are anchored on the surface of BaTiO3 polyhedra. Compared with bare BaTiO3 and CuS, the as-developed composite photocatalysts obviously exhibit enhanced simulated-sunlight photodegradation towards RhB. The highest photoactivity is observed for the 5%CuS/(1%Ag/BTO) composite, which is increased to about 3.3 and 2.0 times as compared with that of bare BaTiO3 and CuS, respectively. Efficient e−/h+ separation resulting from carrier transfer process and LSPR effect of Ag nanoparticles are available for interpreting the enhanced photodegradation of RhB by the CuS/Ag/BTO composites.