g-C3N4 composites with visible-near-infrared light response for efficient photocatalytic organic pollutant degradation

Journal Pre-proofs Z-scheme CdS/CQDs/g-C3N4 composites with visible-near-infrared light response for efficient photocatalytic organic pollutant degradation Shuting Feng, Tian Chen, Zhichao Liu, Jianhui Shi, Xiuping Yue, Yuzhen Li PII: DOI: Reference:

S0048-9697(19)35397-5 https://doi.org/10.1016/j.scitotenv.2019.135404 STOTEN 135404

To appear in:

Science of the Total Environment

Received Date: Revised Date: Accepted Date:

29 August 2019 3 November 2019 5 November 2019

Please cite this article as: S. Feng, T. Chen, Z. Liu, J. Shi, X. Yue, Y. Li, Z-scheme CdS/CQDs/g-C3N4 composites with visible-near-infrared light response for efficient photocatalytic organic pollutant degradation, Science of the Total Environment (2019), doi: https://doi.org/10.1016/j.scitotenv.2019.135404

This is a PDF file of an article that has undergone enhancements after acceptance, such as the addition of a cover page and metadata, and formatting for readability, but it is not yet the definitive version of record. This version will undergo additional copyediting, typesetting and review before it is published in its final form, but we are providing this version to give early visibility of the article. Please note that, during the production process, errors may be discovered which could affect the content, and all legal disclaimers that apply to the journal pertain.

© 2019 Elsevier B.V. All rights reserved.

Z-scheme CdS/CQDs/g-C3N4 composites with visible-near-infrared light response for efficient photocatalytic organic pollutant degradation Shuting Feng, Tian Chen, Zhichao Liu, Jianhui Shi*, Xiuping Yue, Yuzhen Li College of Environmental Science and Engineering, Taiyuan University of Technology, Taiyuan, P. R. China E-mail: [email protected] Tel: +86 18035182405

ABSTRACT We have successfully synthesized novel Z-scheme CdS/CQDs/g-C3N4 composites with visible-near-infrared light response for the photocatalytic degradation of rhodamine B (RhB), methylene Blue (MB) and phenol. Based on the energy band matching theory, CdS was coupled with g-C3N4 using carbon quantum dots (CQDs) as the electron mediator to form the Z-scheme heterojunctions through a simple calcination process. Compared with the single-phase and binary composites, the Z-scheme CdS/CQDs/g-C3N4 composites not only exhibited enhanced photocatalytic activity and photostability but also realized near-infrared light response. CQDs, as the electron mediator, can shuttle the electrons in the CdS/CQDs/g-C3N4 interface via the Z-scheme electron transfer pathway, which lead to improvements in charge separation and oxidizability of the composites. The Z-scheme electron transfer was verified using various techniques, including PL, EIS, EPR and transient photocurrent response. The mechanism of Z-scheme charge transfer was also proposed for the improved photocatalytic RhB degradation activity. In addition, CQDs can capture near-infrared

1

light through the upconversion fluorescence property, ameliorating the broadspectrum photocatalytic competence. Therefore, the Z-scheme heterojunction with visible-near-infrared light response was utilized to improve charge separation, oxidizability and solar energy utilization, as well as to provide new insights for the construction of CQDs-based Z-scheme composites for photocatalysis applications. Keywords: Photocatalysis; Carbon quantum dots; Z-scheme; Broad-spectrum; Organic pollutant degradation

1. Introduction Photocatalysis can be inherently “green” because it enables one to utilize solar energy to accomplish catalytic reactions and holds potential applications in environmental remediation fields (Ahmed and Haider, 2018; Khojasteh et al., 2017; Wang et al., 2018). Green photocatalysis is reflected not only in the utilization of energy but also in “green” photocatalysts. Recently, graphitic carbon nitride (g-C3N4) was introduced as a new type of metal-free photocatalyst for visible-light driven applications (Wang et al., 2009). Due to its non-toxicity, high stability and suitable redox band potential (2.7 eV), g-C3N4 attracted much research attention (Dong et al., 2018; Fu et al., 2018). Nevertheless, its photocatalytic activity was limited by its high recombination rate of photogenerated charge carriers in practical applications (Fu et al., 2018; Mamba and Mishra, 2016). In addition, the visible-light-active CdS photocatalyst also attracted increasing attention due to its narrow band gap (2.3 eV), which can broaden the visible light range to 520 nm (Cheng et al., 2018; 2

Hernandez-Gordillo et al., 2018). However, its high photogenerated carriers recombination rate and self-oxidation resulted in low photocatalytic activity and serious photocorrosion, respectively (Cheng et al., 2018; Hernandez-Gordillo et al., 2018). To simultaneously overcome the above problems of rapid charge recombination of the two photocatalysts, the construction of heterojunctions undoubtedly represents an ideal strategy. Furthermore, CdS with good electrical properties was one of the most suitable candidates satisfying the band alignment to form a heterojunction with g-C3N4. So far, constructing effective CdS/g-C3N4 heterojunctions has been explored to enhance their photocatalytic activity and photostability (Ji et al., 2019; Yin et al., 2018; Yuan et al., 2019). However, most of the CdS/g-C3N4 heterojunctions were the traditional II heterojunctions (Huanyan et al., 2017; Ji et al., 2019; Jian et al., 2019; Yin et al., 2018; Yuan et al., 2019; Zhang et al., 2016b). The electrons in higher CB position of PC I can migrate to lower CB position of PC II, and the holes in higher VB position of PC II can migrate to lower VB position of PC I, which indicated the redox ability of photogenerated carriers after the charges transfer. The Z-scheme photocatalytic systems present an appealing strategy for potential applications, such as pollutant degradation and H2 energy generation (Jiang et al., 2018; Reddy et al., 2019; Xu et al., 2018), derived from the biomimetic artificial photosynthesis, because its charge transport mechanism was similar to that of photosynthesis of green plants in nature (Jiang et al., 2018). The photogenerated electrons in the conduction band (CB) of photocatalyst II could transfer and 3

recombine with the photogenerated holes left in the valence band (VB) of photocatalyst I, and then the negative electrons and positive holes were left in the CB of photocatalyst I and the VB of photocatalyst II, respectively (Xu et al., 2018). The Z-scheme photocatalytic systems can accelerate photogenerated carrier separation and transfer (Jiang et al., 2018). By constructing the Z-scheme photocatalyst, the conduction band potential became more negative and the valence band potential became more positive, which was more thermodynamically favorable in the redox reactions of photocatalytic systems (Reddy et al., 2019). Hence, the Z-scheme photocatalysts showed the stronger redox capacity compared with the traditional II heterojunction photocatalysts (Jiang et al., 2018; Xu et al., 2018). Generally, all-solid-state Z-scheme photocatalysts were advantageous for constructing an efficient Z-scheme system compared with ionic-state Z-scheme photocatalysts (Zhou et al., 2014), as they reduced unwanted backward reactions and shortened the transfer distance of carriers at the interface. Noble metals (Au, Ag et al.) were used as the solid electron mediators, and they were anchored between photocatalyst I and photocatalyst II in Z-scheme systems (Li et al., 2016; Xiao et al., 2016; Yin et al., 2016). In the present study, we not only committed to the construction of metal-free Z-scheme systems to increase the redox capacity of photocatalysts but also need to broaden its light response range to generate more photogenerated carriers. Carbon quantum dots (CQDs), a kind of cost-effective carbon nanomaterial, consist of quasi-spherical nanoparticles with diameters of 2-10 nm (Xuan et al., 2017; Xuan et al., 2016). It was reported that CQDs exhibited many unique photo physical 4

properties in photocatalysis, such as photostability, efficient electrical conductivity and up-conversion of PL. These properties not only endowed CQDs with the abilities required in promising photosensitizers, such as capturing longer wavelength light (NIR light) to broaden the optical absorption range (Wang et al., 2018; Zhou et al., 2016), but also served as an excellent electronic buffer to accept or provide electrons for the contiguous semiconductors (Ma et al., 2016; Wang et al., 2018). Furthermore, CQDs easily combined with another semiconductors owing to π-conjugated structure and oxygen-containing groups at the surface (Qin and Zeng, 2017; Wang et al., 2017b), such as CdS/CQDs (Liu et al., 2013), CQDs/g-C3N4 (Xuan et al., 2016) and CQDs/TiO2 (Miao et al., 2016). Jo et al. introduced RGO into CdS/g-C3N4 and synthesized Z-scheme CdS/RGO/g-C3N4 composites, which showed excellent photocatalytic activity for H2 generation and atrazine degradation (Jo and Selvam 2017). It was because RGO is a highly favorable conductive material, similar to CQDs, that acted as a mediator of electron migration in the CdS/RGO/g-C3N4 composites. Therefore, it was speculated that the design of CQDs-mediated Z-scheme photocatalytic systems could be feasible and could result in the improved oxidizability. In addition, CQDs converted near-infrared light to visible light, which could be absorbed by g-C3N4 and CdS to launch the ensuing photocatalytic reactions. In light of solving the imperfections of g-C3N4, the Z-scheme CdS/CQDs/g-C3N4 composites were designed and fabricated in this work. The photocatalytic activities of the four composites (g-C3N4, CdS/g-C3N4, CQDs/g-C3N4 and the CdS/CQDs/g-C3N4) were compared with respected to photocatalytic degradation of RhB, MB and phenol 5

under visible or NIR light irradiation, and the photostability of the Z-scheme CdS/CQDs/g-C3N4 composites was evaluated through photocatalytic degradation of RhB under visible light irradiation. In addition, the issue of the photocorrosion of CdS was also investigated. Based upon experimental results and characterization analysis, the degradation mechanism of the Z-scheme CdS/CQDs/g-C3N4 composite for organic pollutants was presented, and the roles of Z-scheme photocatalytic systems and CQDs were also systematically analyzed.

2. Experimental section 2.1. Synthesis of carbon quantum dots CQDs were obtained via the modified alkali-auxiliary sonication method (Li et al., 2012). Typically, 9.0 g of glucose was uniformly dissolved in 50 mL deionized water, and 50 mL NaOH solution (1 mol/L) was added to achieve complete mixing. The mixed and colorless solution was transferred to ultrasonic treatment for 120 min. Then, the yellowish-brown solution was neutralized to pH = 7 with HCl. After that, further dialysis treatment with a MWCO 3000 dialysis bag was carried out for purification. Finally, the obtained CQDs were stored in aqueous solution with concentration of 16.03 g/L. 2.2. Synthesis of CdS hollow microspheres The CdS microspheres were obtained by a modified hydrothermal procedure (Liu et al., 2013). Typically, 4.271 g of Cd(NO3)2.4H2O was added to 200 mL of DI 6

water solution containing 3.161 g of thiourea and 2.553 g of glutathione, and then the above mixed solution was continuously stirred for 1 h. Subsequently, the mixtures were heated at 250 °C for 3 h in a 50 mL Teflon-lined stainless steel autoclave. Finally, the obtained precipitates were centrifuged and then washed several times with ultrapure water, then dried at 80 °C in an oven overnight. 2.3. Synthesis of Z-scheme CdS/CQDs/g-C3N4 composites g-C3N4 was obtained via a modified high-temperature thermal polymerization method (El-Kader et al., 2012). Melamine (4 g) was placed into a covered crucible and heated to 550 °C with a rapid heating rate of 55 °C/min, which was maintained for 4 h. The obtained conglomeration was cooled to room temperature naturally. Finally, the yellow bulk product was collected and marked as CN. In the typical synthesis of the CdS/CQDs/g-C3N4 composites, different amounts of CdS hollow microspheres and a calculated amount of CQDs solution was added into 60 mL of DI water. This solution was sonicated for 30 min and stirred for 60 min, g-C3N4 powders (0.4 g) were dispersed in the obtained solution, and the mixed solution was vigorously stirred for 9 h at room temperature and dried overnight at 80 °C. After that, the mixture was transferred into a covered crucible and heated at 150 °C for 1 h in a muffle furnace. A series of CdS/CQDs/g-C3N4 composites with different CdS contents and the optimum content (0.5 wt%) of CQDs were marked as YS5CCN, where Y was equal to 0.1 wt%, 0.25 wt%, 0.5 wt%, and 0.75 wt%. Therefore, the composites were labeled as 1S5CCN, 2S5CCN, 5S5CCN and 7S5CCN, respectively. The CdS/g-C3N4 composite with optimum content (0.25 wt%) of CdS 7

and the CQDs/g-C3N4 composite with optimum content (0.5 wt%) of CQDs were also obtained by a similar method and marked as 2SCN and 5CCN, respectively. The study on the effect of CQDs content in the preparation of Z-scheme SCCN composites was carried out in the supplementary experiments. 2.4. Characterization X-ray diffraction (XRD) patterns were acquired by a Rigaku-Dmax 2500 diffractometer with Cu/Kα radiation (Ka=1.54059 Å). Transmission electron microscope (TEM) images were acquired by a field-emission transmission electron microscope (JEM 2010, 200 kV). X-ray photoelectron spectroscopy (XPS) analysis was performed with an ECSALAB 250 spectrometer. UV-Vis diffuse reflectance spectroscopy (DRS) was performed with a Shimadzu UV-2450 spectrophotometer using BaSO4 as the reference. Fourier transform infrared (FT-IR) spectra was collected on a VERTEX 70 spectrophotometer. Photoluminescence (PL) spectra was obtained via a Hitachi F-4500 spectrophotometer using a 150 W xenon lamp as the excitation source. The photocurrent tests and electrochemical impedance spectra (EIS) were performed with a CHI-660B electrochemical system (China) equipped with a conventional three-electrode electrochemical cell. A 300 W Xe lamp fitted with a 420 nm and 820 nm cut-off filter were used for the light irradiation. The electron spin resonance (EPR) acquired with a Bruker Emxplus-10/12 spectrometer was used to detect the radical intermediates. The concentration of leached Cd2+ in the RhB solution was detected by atomic absorption spectroscopy (AAS, Optima 7000 DV) at a wavelength of 228.8 nm. 8

2.5. Photocatalytic evaluation The photocatalytic performances of the obtained composites were evaluated by the photodegradation of RhB, MB and phenol under visible or NIR light irradiation. An external 300 W Xe lamp fitted with 420 nm and 820 nm cut-off filters served as the light source and was placed 20 cm away from the liquid surface. Composite (0.1 g) was dispersed in a 100 mL water solution containing 10 mg L-1 RhB, MB or phenol and magnetically stirred in the absence of light for 30 min to reach the adsorption-desorption equilibrium prior to the photocatalytic reaction. During the period of visible or NIR light irradiation, 5 mL suspensions were quickly taken from the reactor at the selected time intervals, and the photocatalysts were completely removed by centrifugation. The concentrations of RhB and MB solutions were determined by UV-Vis spectrophotometer (UV-2450, Shi-madzu) at maximum absorbance wavelength of 554 nm and 664 nm, respectively. The concentration of phenol was analyzed on an Aligent 1260 high-performance liquid chromatography (HPLC) (Palo Alto, CA, USA) with a Poroshell 120 EC-C18 column and detection wavelength at 280 nm.

3. Results and discussion 3.1. Phase structure and composition The XRD patterns of the synthesized samples were displayed in Fig. 1. As shown in Fig. 1a, the synthesized g-C3N4 displayed distinct diffraction peaks at 13.1° and 9

27.4°, which are assigned to the in-plane structural packing motif (100) with a distance of d=0.676 nm and the interlayer-stacking structure of the conjugated aromatic system (002) with a distance of d=0.325 nm, respectively (Lu et al., 2017; Qin and Zeng, 2017; Zhao et al., 2018). The diffraction pattern of the synthesized CdS showed peaks at 26.5°, 43.9° and 51.8° corresponding to (111), (220) and (311), and was indexed a hexagonal wurtzite structure (JCPDF Card No. 65-3414), which possesses broader absorption in the visible light region and shows higher photocatalytic activity (Ran et al., 2011). The diffraction peak intensities of g-C3N4 slightly decreased with the addition of CdS, indicating that CdS has been merely deposited onto the surface of g-C3N4. The peaks of CdS in the 2SCN composite were not obvious because of the low content (<0.25%) and the low crystallinity of CdS. When CQDs were introduced, the (002) peak intensities of g-C3N4 markedly increased due to π-π stacking interactions of CQDs and g-C3N4 (Ping et al., 2018). The crystalline phase structure of g-C3N4 was unchanged in the 5CCN composite, but only a slight motion occurred for the (002) diffraction peak toward lower angle value, from 27.4° to 27.25° (Fig. 1b), which is ascribed to the intense interaction between g-C3N4 and CQDs and results in the lattice distance augmentation (Fang et al., 2016). Meanwhile, the characteristic peaks of CQDs could not be observed due to the low content of CQDs and the amorphous structure. In addition, as the doping mass of CdS changed from 0.05% to 0.75% (Fig.S1), the intensity of the (002) peak of g-C3N4 in the SCCN composites increased because the predominant (002) diffraction peak of g-C3N4 at 27.4° merged with the CdS peak at 26.5° (Li et al., 2015) and then 10

decreased because its stacking structure was influenced by the higher content of CdS. Furthermore, it is interesting to note that the (002) peak assigned to g-C3N4 slightly shifts from 27.25° to 27.65° with increasing CdS content, which means that the interlayer distance decreased after modification of CdS based on 5CCN. This may be because CdS adjusted the layer distance, suggesting intimate contact between CdS, CQDs and g-C3N4, and identified the formation of the Z-scheme composites (Jo and Selvam, 2017). In addition, CQDs possibly change the mode of CdS loading on g-C3N4, which may be because CQDs were anchored between CdS and g-C3N4 in the Z-scheme systems. Furthermore, no other impure peaks can be observed, revealing that all of the samples are of high purity.

Fig. 1. XRD patterns of the pure g-C3N4(CN), pure CdS, CQDs(0.5%)/g-C3N4 (5CCN), CdS(0.25%)/g-C3N4 (2SCN) and CdS(0.25%)/CQDs(0.5%)/g-C3N4 (2S5CCN) composites (a), together with the magnified parts of CN, 5CCN, 2SCN and 2S5CCN samples (b).

Fig. 2 described the FT-IR spectrograms of g-C3N4, CdS and the related composites. In Fig. 2a, for the pure CdS, the characteristic peaks at 3395 and 1630 cm-1 belonged to the surface-absorbed water molecules, and those centered at 1383 and 1090 cm-1 can be attributed to the Cd-S bands (Fu et al., 2013). The Cd-S bands were also observed in the 2SCN composite, suggesting that CdS had been formed in the 2SCN composite. Additionally, the Cd-S bands of the 2S5CCN composite were weaker than those of 2SCN, which was probably because CQDs were coated on the surface of CdS. The characteristic peaks in the range of 1244 cm-1 to 1637 cm-1 correspond to the CN heterocycle structure of g-C3N4 (Fu et al., 2013) ， and the 11

vibrations of the 2S5CCN composite were far stronger than those of 2SCN because CQDs provide many active sites and active groups, resulting in intimate contact between CQDs, CdS and g-C3N4. By comparing g-C3N4, 5CCN and 2S5CCN in Fig. 2b, the pure g-C3N4 presented four characteristic peaks at 3430 cm-1, 3261 cm-1, 1244-1637 cm-1, and 809 cm-1. The broad peaks which appeared at approximately 3430 cm-1 and 3261 cm-1 can be assigned to the O-H bonds and N-H vibration, which were caused by adsorbed water molecules on the surface of g-C3N4 and uncondensed amino groups, respectively (Zhang et al., 2016a). The characteristic peak which appeared at 809 cm-1 was assigned to the breathing mode of s-triazine rings of g-C3N4 and the characteristic peaks centered at 1244, 1317, 1413, 1568 and 1637 cm-1 corresponded to the typical stretching modes of the CN heterocycle structure (Zhang et al., 2016a). For 5CCN and 2S5CCN, except for the typical characteristic peaks of g-C3N4, the peak at 3070 cm-1 was assigned to the C-H bending vibration of CQDs (Gao et al., 2015), indicating that CQDs were successfully introduced into the 5CCN and 2S5CCN composite photocatalysts. Surprisingly, the vibration peaks at 1637 cm-1 increased with the addition of CQDs into the 5CCN, which was due to the conjugate vibrations between g-C3N4 and CQDs, and the peaks were also related to the stretching modes of C=C that contributed to the photoluminescence of CQDs based on π-π* transitions (Wang et al., 2014). The doped CQDs and CdS did not damage the integral structure of g-C3N4. Fig. 2. FT-IR spectra of the pure CdS, CdS(0.25%)/g-C3N4 (2SCN) and CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) composites(a); pure g-C3N4(CN), CQDs(0.5%)/g-C3N4 (5CCN) and 2S5CCN samples (b). 12

The XPS spectra of the pure g-C3N4 and the 2S5CCN composite were displayed in Fig.3. In the 2S5CCN composite, four elements including C, N, O and Cd were investigated in the survey spectrum (Fig.3a), but S was not obviously observed due to the weak photoelectron peak intensity of XPS. In detail, the 2S5CCN composite showed similar C and N peaks compared with the pure g-C3N4, which were revealed from the high-resolution C 1s and N 1s spectra in Fig.3b-3c. The C 1s spectrum can be fitted into three peaks at 284.6, 286.1 and 288.1 eV. The peaks at 284.6 and 288.1 eV

belonged

to

the

C-C

bonds

and

N-C=N

of

g-C3N4,

respectively

(Asadzadeh-Khaneghah et al., 2019; Wang et al., 2017c). However, the peak at 286.1 eV was not observed in the pure g-C3N4, and the peak can be assigned to hydroxyl groups originated from CQDs (Liu et al., 2013). Importantly, the ratio of C-C bonds to N-C=N changed from 0.109 to 0.624 after adding CQDs (Table.S1), demonstrating the presence of CQDs in the 2S5CCN composite. The high resolution N 1s spectrum (Fig. 3c) can be deconvoluted into three peaks: the peaks at 398.6 eV and 400.4 eV can be ascribed to the C=N-C units and N-(C)3 units of g-C3N4, respectively (Asadzadeh-Khaneghah et al., 2019; Wang et al., 2017a). Compared to the pure g-C3N4, the peak at 404.8 eV was assigned to the π excitations of CQDs (Liu et al., 2015). With regards to Cd 3d and S 2p spectra (Fig.3d-3e), the peaks of Cd 3d were observed at 404.9 eV (3d5/2) and 411.7 eV (3d3/2), and the peak of S 2p was observed at 161.4 eV, indicated that CdS existed in the 2S5CCN composite and that the Cd and S elements were mainly Cd2+ and S2-, respectively (Ping et al., 2018). The Cd 3d5/2 and S 2p binding energies in the 2S5CCN composite were shifted to lower values 13

compared with that of pure CdS (Fig.S2), which suggested the possibility of electronic interactions between CdS and the other components (Jo and Selvam, 2017), which was conducive to the formation of the Z-scheme heterojunctions.

Fig.3. XPS spectra of CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) (a) survey; (b) C 1s; (c) N 1s; (d) Cd 3d; (e) S 2p.

3.2. Morphology structure Fig.4 displayed the TEM images of the synthesized samples. For g-C3N4 (Fig. 4a), a layered structure with interlayer stacking was observed. The CdS exhibited hollow microspheres with an external diameter of 28 nm and inner diameter of 6 nm, as well as significant aggregation (Fig. 4b-c). For the 2S5CCN composite, CdS hollow microspheres were deposited on dense dark spot regions of CQDs, CdS was coated with CQDs, and then the CdS/CQDs was attached on the surface of g-C3N4 (Fig. 4d-4e). Even after the intensive ultrasonication treatment in ethanol for the preparation of TEM samples, the majority of CdS/CQDs was still anchored on the surface of g-C3N4, and no free CdS and CQDs were discovered, which revealed that CdS and CQDs were attached via electronic interaction rather than a simple physical mixture (Fu et al., 2013). These results suggested that the formation of CdS/CQDs/g-C3N4 interfacial junctions occurred steadily, which may arise from π-π stacking interactions between CdS/CQDs and g-C3N4 (Bhunia and Jana, 2014). HRTEM images (Fig.4f) were used to further illustrate the formation of Z-scheme 2S5CCN composite. The lattice fringes with a spacing of 0.325 nm corresponded to 14

the (002) crystal plane of g-C3N4, and the lattice spacings of 0.358 nm, 0.336 nm and 0.331 nm matched the (100), (002) and (101) planes of CdS in turn, which sited intimately and overlapped with each other. This overlapped facilitated the formation of the CdS/CQDs/g-C3N4 heterojunctions and promoted the smooth charges transfer in the interface. The elemental mapping images (Fig. 4g) also confirmed the presence of C, N, Cd and S elementals in the 2S5CCN composite, the elements were well dispersed in the selected area, suggesting CdS and CQDs successfully introduced into the 2S5CCN composite. Fig. 4. TEM images of g-C3N4 (a), CdS(b-c), CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) composites (d); HRTEM images of 2S5CCN composites(e-f); Elemental mapping patterns (g) of 2S5CCN.

3.3. Optical property analyses The UV-Visible diffuse reflectance spectra of samples were depicted in Fig. 5a. The absorption edges of the pure g-C3N4 and CdS were approximately 471 nm and 590 nm, which can be assigned to the intrinsic band gaps of g-C3N4 and CdS, respectively (Fu et al., 2013). The corresponding direct band gaps values for g-C3N4, CdS, and 2S5CCN were 2.79, 2.24 and 2.68 eV, respectively (Fig. 5b). The VB and CB band energy (related to Mulliken electronegativity) of CdS can be calculated to be 1.81 eV and -0.43 eV and those of g-C3N4 were 1.6 eV and -1.19 eV, respectively. For 2SCN, 5CCN and 2S5CCN, the absorption edges exhibited redshifts and broadened the optical absorption ranges to 482 nm, 491 nm and 503 nm in turn, which was likely due to Z-scheme CdS/CQDs/g-C3N4 heterojunction ， resulting in the effective combination of CQDs, CdS and g-C3N4 (Jo and Selvam, 2017). Based on the 15

strong light harvesting capacity of CQDs, the 2S5CCN composite presented the widest visible absorption range among the samples, which was beneficial to capturing more photons and resulted in the formation of more carriers and improvement of photocatalytic

activity

(Jo

and

Selvam,

2017).

Hence,

the

Z-scheme

CdS/CQDs/g-C3N4 composite provided the potential to become a practical photocatalytic material. Fig. 5. UV–vis DRS (a) and (Ahν)2 versus hνcurve (b) of g-C3N4(CN), pure CdS, CdS(0.25%)/g-C3N4 (2SCN), CQDs(0.5%)/g-C3N4(5CCN) and CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN); up-converted photoluminescence spectra (c) of the as-synthetic CQDs.

Up-conversion photoluminescence (PL) properties of CQDs were explored under varying excitation wavelengths. As shown in Fig. 5c, the up-converted PL emission spectra exhibited excitation wavelength-dependent fluorescence behavior, and CQDs could absorb the longer-wavelength light within visible (650-750 nm) and NIR (800-1000 nm) regions and then emit the shorter-wavelength light located at 400-600 nm. This fascinating optical property may be due to the multiphoton active process and their effective interactions with each other (Hui et al., 2016; Xu et al., 2016). The up-conversion PL properties demonstrated that CQDs can act as photosensitizers to make CdS/g-C3N4 components effectively harness solar light for broad spectrum response and enhance the photocatalytic activity. Based on reports, CQDs were confirmed as both electron donors and electron acceptors (Pan et al., 2018; Wang et al., 2018). The Z-scheme charge transfer pathway was verified by PL emission spectra (Fig. 6a). All of the samples showed a similar broad emission band centered at 550-650 nm with an excitation wavelength of 467 nm. 16

The PL emission intensity for 2SCN composites was significantly weaker than that of the pure g-C3N4, indicating that the recombination of electron-hole pairs was efficiently suppressed upon the introduction of CdS. Furthermore, the intensity for the 2S5CCN composites was comparatively lower than that of 2SCN, suggesting improved charge separation after modification of CQDs, which was because CQDs, as the electrons mediators, formed the Z-scheme heterojunction (Qian et al., 2018), and guided the electron transfer at the CdS/CQDs/g-C3N4 interface. Hence, the significant quenching of the emission at 550-650 nm indicated significant suppression of electron-hole pair recombination through the Z-scheme electron transfer pathway. The interfacial charge transfer behaviors of the 2S5CCN composite were further proved by EIS Nyquist plots (Fig. 6b). As is well-known, the smaller semicircles corresponded to weaker charge resistance (RCT), signifying more efficient charge transfer (Ming et al., 2016). Note that the best electronic conductivity appears on the 2S5CCN composite, with 2SCN composite being second place, indicating that ultrafast electron transfer of the 2S5CCN composite was achieved, which is consistent with the interpretation of the PL spectra. This result supported the Z-scheme electron transfer process, expedited interfacial electron transfer, enhanced electron-hole pair separation, and thus lowered the charge transfer resistance (Qian et al., 2018). The transient photocurrent responses also provided another corroborative evidence for verifying charge transfer behaviors of the 2S5CCN photoelectrodes under visible light illumination (Fig. 6c). It was noticeable that 2S5CCN exhibited the highest photocurrent intensity compared with those of g-C3N4, 2SCN and 5CCN, 17

demonstrating that the lifetime of the photogenerated carriers of the 2S5CCN composite was effectively prolonged. This further enhanced the separation of electron-hole pairs due to the Z-scheme electron transfer (Jo and Selvam, 2017; Qian et al., 2018), which was consistent well with PL and EIS analysis. Moreover, the transient photocurrent response was also performed to further investigate charge transfer of the samples under NIR light illumination (Fig. S3). No photocurrent signal was observed over the g-C3N4 and 2SCN composites. Dramatically, the photocurrent signals showed the significant enhancement after the introduction of CQDs due to its up-conversion PL property, which was consistent with PL result (Fig. 5c). The signal intensity of 2S5CCN was the stronger than that of 5CCN due to Z-scheme electron transfer process under NIR light. Fig.6. PL spectra (a) and EIS spectra (b) of g-C3N4 (CN), CdS(0.25%)/g-C3N4(2SCN), CQDs(0.5%)/g-C3N4 (5CCN) and CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) composites; Photocurrent responses (c) of CN, 2SCN , 5CCN and 2S5CCN composites under visible light.

3.4. Photocatalytic properties The photocatalytic degradation activities of all the catalysts under visible-light illumination were assessed in Fig. 7. RhB acted as one of the target dye pollutants in the work. As seen in Fig. S4, the content of CQDs had an obvious impact on the photocatalytic activity of the SCCN composites. The photocatalytic degradation efficiency (RE) considerably increased with loading of a small amount of CQDs on SCN, and it reached the maximum value at an optimal content of 0.5 wt% CQDs. However, the more excessive CQDs enabled the degradation rate of RhB to decrease,

18

which was ascribed to generous CQDs resulting in reduction of light absorption of by the CdS component and then reduction in the generation of oxidative active species during the photocatalytic reaction (Pan et al., 2018). The RhB degradation curves of the SCCN composites with different ratios of CdS were depicted in Fig.7a. The SCCN composites showed notably enhanced photocatalytic activity compared with the pure g-C3N4 and 5CCN composites. The photocatalytic RE of the SCCN composites increased first and then decreased gradually with the loading of CdS, and it exhibited the optimal photocatalytic activity with a loading content of 0.25 wt% CdS; the RE reached 100% upon irradiation for 20 min. The 10 min rate constants (Kobs) were used to investigate the reaction kinetics of the RhB degradation (Fig. 7b). Among the SCCN composites, 2S5CCN showed the maximum Kobs of 0.143 min-1, which was approximately 3.5 and 2.6 times higher than those of the pure g-C3N4 (0.041 min-1) and 5CCN (0.055 min-1), respectively. However, the 10 min Kobs depressed with the content of CdS exceeds 0.25 wt% to reach 0.5 or 0.75 wt%. The unexpected phenomenon appeared to be a result of the shielded active reaction sites and debased optical adsorption ability (Yu et al., 2015). In addition, the visible-light photocatalytic activities of synthesized samples, including the pure g-C3N4, pure CdS, 2SCN and 5CCN binary systems and 2S5CCN composite, were also investigated in Fig. 7c. The pure CdS showed the superior adsorption ability due to its hollow microsphere structure and excellent photocatalytic activity, while the pure CQDs exhibited the negligible photocatalytic activity. For the 2SCN and 5CCN binary systems, their photocatalytic capabilities were superior to 19

that of the pure g-C3N4, suggesting that the appropriate loading amounts of CdS and CQDs contributed to the enhanced visible-light acquisition and electron-hole separation owing to the formation of traditional II SCN and CCN heterojunctions (Huanyan et al., 2017; Li et al., 2016). However, compared with the binary systems, the 2S5CCN composite showed the superior photocatalytic activity because the Z-scheme electron transfer pathway facilitated the electron migration and enhanced the photocatalytic activity. Notably, the CQDs, as a solid-state electron mediator between the two semiconductors, shuttled the photogenerated electrons at the CdS/CQDs/g-C3N4 interfaces. This indicated that effective charge separation had occurred, which was ascribed to the Z-scheme electron transfer between CdS and g-C3N4 via CQDs (Pan et al., 2018). The photocatalytic degradation curves of RhB were explored via the UV-Vis absorption spectra (Fig. S5). The 2S5CCN composite exhibited the complete decomposition of RhB molecules after 20 min of irradiation and the main absorption peak located at 554 nm showed a blueshift due to the de-ethylation of RhB (Asadzadeh-Khaneghah et al., 2019). To better verify and elucidate the broad spectral response of the 2S5CCN composite, the photocatalytic experiments were carried out under NIR light (λ > 820 nm) irradiation (Fig. 7d). The g-C3N4, CdS, CQDs and 2SCN composites displayed negligible RhB degradation after 120 min NIR light irradiation, while 5CCN and 2S5CCN presented the degradation rates of 38% and 62%, respectively. These results indicated that the composites containing CQDs components can utilize long 20

wavelengths light to complete the photocatalytic reaction due to the up-conversion PL property of CQDs. More importantly, the 2S5CCN composite possessed the degradation rate of RhB prior to 5CCN, which can be caused by the Z-scheme electron transfer path in the CdS/CQDs/g-C3N4 interface. Among influencing factors, areal loading, which is of great significance to the construction of photocatalytic reaction tanks in the actual wastewater treatment plants, should be considered for the practical applications. Areal loading had an obvious impact on RhB degradation. Specifically, three RhB solutions of 100 mL containing 0.1 g 2S5CCN composite were added into 100, 250 and 500 mL beakers, corresponding to area load of 0.457, 0.256 and 0.158 g m-2, respectively. Their photocatalytic RE values at different time periods were showed in Fig.S6. The photocatalytic RE values of 10 and 20 min increased first and then decreased gradually with the reduction of areal load. The maximum photocatalytic RE values reached 100% at 20 min with the areal load of 0.256 g m-2. The areal load of 0.158 g m-2 possessed comparatively less degradation efficiency, which may be because the dominance of limited photons absorption per unit area decreased the light quantum efficiency. Thus, as an important parameter, the optimized design of areal load can ensure an efficient degradation rate in photocatalytic reaction tanks. Fig.7. Rate of RhB degradation (a) and the 10 min Kobs (b) for g-C3N4(CN) and CdS/CQDs/g-C3N4(SCCN)composites under visible light illumination; Rate of RhB degradation using unit (CN, CdS, CQDs), optimized CdS(0.25%)/g-C3N4(2SCN) and CQDs(0.5%)/g-C3N4(5CCN) binary composites and CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) composites(c); Rate of RhB degradation of CN, CdS, CQDs, 2SCN, 5CCN and 2S5CCN under long wavelength ( > 820nm) irradiation (d).

21

Apart from the degradation of RhB, another target pollutant, methylene blue (MB) has also been studied. Obviously, the SCCN composites also displayed the superior photocatalytic activity for MB degradation, and the highest degradation RE (98%) of the 5S5CCN composite was achieved at 120 min of irradiation (Fig.S7). To further verify the oxidizability of the 5S5CCN composites, the optimal 2SCN and 5CCN contents were selected for comparison (Fig.8a). The degradation efficiencies of g-C3N4, CdS, 2SCN, 5CCN and 5S5CCN were 65%, 75%, 80%, 86% and 98%, respectively, after irradiation for 120 min. Compared with 2SCN and 5CCN, the 5S5CCN composites presented excellent photocatalytic activity, which was attributed to construction of the Z-scheme CdS/CQDs/g-C3N4 heterojunctions. The reaction kinetics of the MB photocatalytic degradation were quantitatively investigated (Fig.8b). It can be seen that the average rate constants(Kobs) of the 5S5CCN composite was 0.024 min-1 under visible light irradiation (120 min), approximately 2.7 times that of g-C3N4 (0.009 min-1), 2.2 times that of CdS (0.011 min-1), 1.9 times that of 2SCN (0.013 min-1) and 1.5 times that of 5CCN(0.016 min-1), respectively. Furthermore, the UV-Vis absorption spectra of the MB degradation by the 5S5CCN composite was also provided in Fig.S8. Apart from the degradation of dye contaminants, phenol acted as one of the colorless pollutants was also studied. The degradation rate and average rate constants (Kobs) of phenol under visible light irradiation were showed in Fig.8c-d, respectively. Self-degradation of phenol was negligible. The degradation efficiencies of g-C3N4, CdS, 2SCN and 5CCN were 21%, 28%, 32% and 30% after irradiation for 120 min, 22

respectively. However, for the 2S5CCN composite, the degradation rate of phenol was significantly improved, 58% of phenol was degraded after 120 min illumination owing to suppression of electron-hole pairs recombination. Moreover, the 2S5CCN composite showed the maximum Kobs datum of 0.015 min-1, approximately 2.5 times that of pure g-C3N4 (0.006 min-1), 1.9 times that of CdS (0.008 min-1), 1.5 times that of 2SCN (0.010 min-1) and 1.7 times that of 5CCN (0.009 min-1), respectively. The above results indicated that the 2S5CCN composite had also possessed the excellent photocatalytic activity for the degradation of phenol. It was noteworthy that the Kobs for degradation of RhB, MB and phenol over the 2S5CCN composite were higher than those of other studies (Table.S2), indicating that the 2S5CCN composite has excellent photocatalytic performance. Fig.8. Rate of MB degradation (a) and the Kobs (b) for g-C3N4(CN) , CdS, CdS(0.5%)/g-C3N4 (5SCN) , CQDs(0.5%)/g-C3N4 (5CCN) and CdS(0.5%)/CQDs(0.5%)/g-C3N4 (5S5CCN) composites under visible light illumination; Rate of phenol degradation (c) and the Kobs (d) for CN, CdS, 2SCN, 5CCN and 2S5CCN composites under visible light illumination.

The photostability of the Z-scheme 2S5CCN composite was investigated by repetitive time-circle RhB degradation (Fig.9a). There was no distinguishable discrepancy in photocatalytic effect after four cycles under visible light illumination, which indicated that the formation of Z-scheme heterojunctions favors migration of charges via the Z-scheme electrons transfer, thus restraining photocorrosion of CdS. Meanwhile, stability of the composites conversely validated the formation of Z-scheme heterojunction. To further investigate the photostability of the 2S5CCN composites and photocorrosion of CdS, the concentrations of leached Cd2+ in the 23

solutions of pure CdS and the 2S5CCN composites after 4 h irradiation were determined by AAS as listed in Table.S3. The results showed that the 2S5CCN composite contains a much lower concentration of leached Cd2+ and that its dissolution trend was significantly retarded compared to pure CdS; the results also verified the enhanced photostability of the Z-scheme composite against damage by CdS photocorrosion. In order to investigate the structural stability of the composites during the degradation process, we characterized the structure of 2S5CCN after photocatalytic degradation of RhB for four cycles by XRD analysis (Fig. S9). The results indicated that the structure of the used sample did not show significant changes compared with the fresh sample. Therefore, it was confirmed that the prepared Z-scheme 2S5CCN photocatalysts can work as a stable and durable photocatalyst. 3.5. Mechanism and verification of Z-scheme electron transfer To analyze the Z-scheme electron transfer in the three-component composite (CdS/CQDs/g-C3N4), the reactive radicals involved in the photocatalytic RhB degradation were detected by adding disodium ethylenediamine tetraacetate (EDTA-2Na), isopropyl alcohol (IPA) and benzoquinone (BQ) as scavengers, which can scavenge holes (h+), hydroxyl radicals (•OH) and superoxide radicals(•O2−), respectively (Fig.8d). Slight inhibition was observed for the photocatalytic RhB degradation by adding the IPA. In contrast, the presence of BQ and EDTA-2Na resulted in the obviously decreased photocatalytic activity, implying that •O2− and h+ were the main oxidative species affecting the degradation efficiency. To further confirm the Z-scheme electron transfer path during the photocatalytic 24

process, EPR spectra of the g-C3N4, 2SCN and 2S5CCN composites were carried with DMPO as the trapping agent (Fig.9c-d). No EPR signals of DMPO–•O2- and DMPO-•OH was observed in the dark. Upon visible light irradiation, both the DMPO-•O2- and DMPO-•OH signal intensities of 2S5CCN were much stronger than those of pure g-C3N4 and 2SCN, indicating that the more •O2- and •OH radicals were generated due to Z-scheme electron transfer path (Qin et al., 2018). Moreover, the stronger DMPO-•O2- signals and relatively weaker signal of DMPO-•OH were observed for the 2S5CCN composite, implying that the •O2- radicals played important roles in the photocatalytic process, which was consistent well with the results of trapping experiment. Since the holes in the valance band of CdS (EVB = +1.81 V, vs. NHE) were incapable of oxidizing hydroxyl groups into •OH radicals (E(•OH/OH-) = +1.99 V, vs.NHE) due to the more negative VB potential (Zhang et al., 2016b), thus the observed •OH radicals were generated from the •O2- radicals in the photochemical reaction. Fig.9. Cycling photocatalytic degradation of RhB over the 2S5CCN composites(a); Effect of different scavengers on the RhB degradation over the 2S5CCN composites(b); EPR spectra of the (c) DMPO-•O2- and (d) DMPO-•OH adducts recorded with CN, 2SCN, 2S5CCN under visible light irradiation (water system =1mL, DMPO = 40.0 μL, catalyst=10 mg).

The Z-scheme charge transfer model for light-driven oxide species generation can be elucidated based on the band potentials (Fig. 10). The CQDs played multiple important roles in the system. Under visible light ( ＜ 460 nm) irradiation, the photogenerated electrons of g-C3N4 and CdS were excited simultaneously from the valence band (VB) to the conduction band (CB). In addition, under NIR light 25

irradiation, the incorporated CQDs up-converted NIR light to visible light, which achieved the succedent excitation of the CdS/g-C3N4 system to generate e--h+ pairs (Hou et al., 2015). Subsequently, the holes in the VB of CdS, which has a strong oxidizing potential (1.81 eV) compared with g-C3N4 (VB =1.6 eV), remained. The CB electrons of CdS could rapidly migrate to CQDs (electron transfer I, CdS→CQDs) due to the fact that (i) intimate contact interface between CdS/CQDs (verified by TEM analysis) shortens charge transfer distance and (ii) CQDs function as electron-accepters contribute to electron-transfer. The electrons were trapped by CQDs and recombined with the VB holes of g-C3N4 via the Z-scheme electron pathway (electron transfer II, CdS→CQDs→g-C3N4) (Jo and Selvam, 2017). Thus, the simultaneous existence of electron transfers I and II resulted in the effective separation of carries and strong redox potential, where electrons remaining in the CB of g-C3N4 can react with O2 to generate •O2− (Zhang et al., 2016b). The main active species •O2− and h+ can oxidize RhB molecules (Ji et al., 2019). Thus, the CQDs as the electron reservoirs can facilitate the efficient electrons extraction from the CB of CdS and enhance the carrier separation. Meanwhile, CQDs with the upconversion PL property could aid the Z-scheme 2S5CCN composites to broaden the light absorption range, thereby increasing the formation speed and the quantity of the electron-hole pairs. In

brief,

the

Z-scheme

CdS/CQDs/g-C3N4

heterojunction

with

visible-near-infrared light response was dedicated to the enhancement of charge separation, redox ability and adequate utilization of solar energy. 26

Fig. 10. Energy band diagram of the Z-scheme electron transfer (CdS/CQDs/g-C3N4) mechanism.

4. Conclusions Novel designed Z-scheme CdS/CQDs/g-C3N4 heterojunction composites were successfully synthesized by a simple calcination preparation. The Z-scheme electron transfer was achieved via the CdS/CQDs/g-C3N4 interface. The Z-scheme CdS/CQDs/g-C3N4 heterojunction composites showed superior photocatalytic capabilities for the degradation of RhB, MB and phenol compared with the pure g-C3N4 and the optimized binary composites (CdS/g-C3N4 and CQDs/g-C3N4). In the Z-scheme heterojunctions system, g-C3N4 and CdS coupled based on band potential matching with CQDs acting as the electron mediators, thus resulting in accelerated charge separation and improved redox ability. Meanwhile, CQDs with the upconverted PL property can convert near-infrared light to visible light and improve the solar energy utilization. Hence, this work provided a new idea for the design and construction of Z-scheme heterojunction composites with visible-near-infrared light response based on CQDs, which were expected to degrade organic pollutants in the environment.

Acknowledgements This work was supported by National Natural Science Foundation of China (51378330), China Postdoctoral Science Foundation Funded Project (2019M651083), the Key Research and Development (R&D) Project of Shanxi Province (201603D321012, 201703D321009-5) and Qualified Personnel Foundation of Taiyuan University of Technology (QPFT) (No: tyutrc-201326c). 27

References

Ahmed, S.N., Haider, W., 2018. Heterogeneous photocatalysis and its potential applications in water and wastewater treatment: a review. Nanotechnology 29, 342001. Asadzadeh-Khaneghah, S., Habibi-Yangjeh, A., Yubuta, K., 2019. Novel g-C3N4 nanosheets/CDs/BiOCl photocatalysts with exceptional activity under visible light. J. Am. Ceram. Soc. 102, 1435-1453. Bhunia, S.K., Jana, N.R., 2014. Reduced graphene oxide-silver nanoparticle composite as visible light photocatalyst for degradation of colorless endocrine disruptors. Acs Appl. Mater. Inter. 6, 20085-20092. Cheng, L., Xiang, Q.J., Liao, Y.L., Zhang, H.W., 2018. CdS-Based photocatalysts. Energ. Environ. Sci. 11, 1362-1391. Dong, J.L., Kang, J., Lee, H.J., Dong, S.C., Sang, O.K., 2018. High activity hydrogen evolution catalysis by uniquely designed amorphous/metal interface of core-shell phosphosulfide/N-doped CNTs. Adv. Eng. Mater. 8, 1702806. El-Kader, F.H., Moharram, M.A., Khafagia, M.G., Mamdouh, F., 2012. Influence of the nitrogen content on the optical properties of CNx films. Spectrochim. Acta, Part A 97, 1115-1119. Fang, S., Xia, Y., Lv, K., Li, Q., Sun, J., Li, M., 2016. Effect of carbon-dots modification on the structure and photocatalytic activity of g-C3N4. Appl. Catal. B 185, 225-232. Fu, J., Chang, B., Tian, Y., Xi, F., Dong, X., 2013. Novel C3N4-CdS composite photocatalysts with organic-inorganic heterojunctions: in situ synthesis, exceptional activity, high stability and photocatalytic mechanism. J. Mater. Chem. A 1, 3083-3090. Fu, J., Yu, J., Jiang, C., Bei, C., Fu, J., Yu, J., Jiang, C., Bei, C., 2018. g-C3N4-Based heterostructured photocatalysts. Adv. Energy, Mater. 8, 1701503. Gao, X., Lu, Y., Zhang, R., He, S., Ju, J., Liu, M., Li, L., Chen, W., 2015. One-pot synthesis of carbon nanodots for fluorescence turn-on detection of Ag+ based on the Ag+-induced enhancement of fluorescence. J. Mater. Chem.

3, 2302-2309.

Hernandez-Gordillo, A., Acevedo-Pena, P., Bizarro, M., Rodil, S.E., Gomez, R., 2018. Photoreduction of 4-Nitrophenol in the presence of carboxylic acid using CdS nanofibers. J. Mater. Sci-Mater. El. 29, 7345-7355. Hou, J., Cheng, H., Yang, C., Takeda, O., Zhu, H., 2015. Hierarchical carbon quantum dots/hydrogenated-γ-TaON heterojunctions for broad spectrum photocatalytic performance. Nano Energy 18, 143-153. Huan,Y., Li,C., Li,G., 2017. Combination mechanism and enhanced visible-light photocatalytic activity and stability of CdS/g-C3N4 heterojunctions. J. Mater. Sci. Technol. 33, 30-38. Hui, Z., Zhao, L., Geng, F., Guo, L.H., Wan, B., Yu, Y., 2016. Carbon dots decorated graphitic carbon nitride as an efficient metal-free photocatalyst for phenol degradation. Appl. Catal. B 180, 656-662. Ji, C., Du, C., Steinkruger, J.D., Zhou, C., Yang, S., 2019. In-situ hydrothermal fabrication of CdS/g-C3N4 nanocomposites for enhanced photocatalytic water splitting. Mater. Lett. 240, 128-131. Jian, Q., Jin, Z., Wang, H., Zhang, Y., Wang, G., 2019. Photoelectron directional transfer over a g-C3N4/CdS heterojunction modulated with WP for efficient photocatalytic hydrogen evolution. Dalton T. (Cambridge, England : 2003) 48, 4341-4352. Jiang, L., Yuan, X., Zeng, G., Jie, L., Wu, Z., Hou, W., 2018. Construction of all-solid-state Z–scheme photocatalyst based on graphite carbon nitride and its enhancement to catalytic activity. Environ. Sci. Nano. 5, 110. Jo, W.K., Selvam, N.C.S., 2017. Z-scheme CdS/g-C3 N4 composites with RGO as an electron mediator for 28

efficient photocatalytic H2 production and pollutant degradation. Chem. Eng. J. 317, 913-924. Khojasteh, H., Salavati-Niasari, M., Safajou, H., Safardoust-Hojaghan, H., 2017. Facile reduction of graphene using urea in solid phase and surface modification by N-doped graphene quantum dots for adsorption of organic dyes. Diam. Relat. Mater. 79, 133-144. Li, H., Liu, R., Liu, Y., Huang, H., Yu, H., Ming, H., Lian, S., Lee, S-T., Kang, Z., 2012. Carbon quantum dots/Cu2O composites with protruding nanostructures and their highly efficient (near) infrared photocatalytic behavior. J. Mater. Chem. 22, 17470-17475. Li, H., Yu, H., Quan, X., Chen, S., Zhang, Y., 2016. Uncovering the key role of the fermi level of the electron mediator in a Z-scheme photocatalyst by detecting the charge transfer process of WO3-metal-g-C3N4 (Metal = Cu, Ag, Au). Acs Appl. Mater. Inter. 8, 2111-2119. Li, W., Chang, F., Dai, S., Yue, J., Hua, F., Hao, H., 2015. Fabrication of sulfur-doped g-C3N4/Au/CdS Z-scheme photocatalyst to improve the photocatalytic performance under visible light. Appl. Catal. B 168-169, 465-471. Li, Z., Liu, Z., Li, B., Li, D., Ge, C., Fang, Y., 2016. Novel CdS nanorods/g-C3N4 nanosheets 1-D/2-D hybrid architectures: an in situ growth route and excellent visible light photoelectrochemical performances. J. Mater. Sci-Mater. El. 27, 2904-2913. Liu, J., Liu, Y., Liu, N., Han, Y., Zhang, X., Huang, H., Lifshitz, Y., Lee, S.T., Zhong, J., Kang, Z., 2015. Metal-free efficient photocatalyst for stable visible water splitting via a two-electron pathway. Science 46, 970-974. Liu, Y., Yu, Y.X., Zhang, W.D., 2013. Carbon quantum dots-doped CdS microspheres with enhanced photocatalytic performance. J. Alloy. Compd. 569, 102-110. Lu, C., Peng, Z., Jiang, S., Xi, W., Song, S., Zhu, M., Lou, Z., Zhe, L., Liu, F., Liu, Y., 2017. Photocatalytic reduction elimination of UO22+ pollutant under visible light with metal-free sulfur doped g-C3N4 photocatalyst. Appl. Catal. B 200, 378-385. Ma, Y., Li, X., Yang, Z., Xu, S., Zhang, W., Su, Y., Hu, N., Feng, J., Zhang, Y., 2016. Morphology control and photocatalysis enhancement by in-situ hybridization of Cu2O with nitrogen-doped carbon quantum dots. Langmuir 32, 9418. Mamba, G., Mishra, A.K., 2016. Graphitic carbon nitride (g-C3N4 ) nanocomposites: A new and exciting generation of visible light driven photocatalysts for environmental pollution remediation. Appl. Catal. B 198, 347-377. Miao, R., Luo, Z., Zhong, W., Chen, S.Y., Jiang, T., Dutta, B., Nasr, Y., Zhang, Y., Suib, SL., 2016. Mesoporous TiO2 modified with carbon quantum dots as a high-performance visible light photocatalyst. Appl. Catal. B 189, 26-38. Ming, F., Hong, J., Xu, X., Wang, Z., 2016. Dandelion-like ZnS/carbon quantum dots hybrid materials with enhanced photocatalytic activity toward organic pollutants. Rsc Adv. 6, 31551-31558. Pan, J., Liu, J., Zuo, S., Khan, UA., Yu, Y., Li, B., 2018. Structure of Z-scheme CdS/CQDs/BiOCl heterojunction with enhanced photocatalytic activity for environmental pollutant elimination. Appl. Surf. Sci. 444, 177-186. Ping, W., Yue, S., Wang, F., Yu, H., 2018. Synergistic effect of electron-transfer mediator and interfacial catalytic active-site for the enhanced H2-evolution performance: A case study of CdS-Au photocatalyst. Appl. Catal. B 220, 561-569. Qian, L., Hou, Y., Yu, Z., Li, M., Li, F., Sun, L., Luo, W., Pan, G., 2018. Metal-induced Z-scheme CdS/Ag/g-C3N4 photocatalyst for enhanced hydrogen evolution under visible light: The synergy of MIP effect and electron mediator of Ag. Mol. Catal. 458, 43-51. Qin, J., Zeng, H., 2017. Photocatalysts fabricated by depositing plasmonic Ag nanoparticles on carbon quantum dots/graphitic carbon nitride for broad spectrum photocatalytic hydrogen generation. Appl. Catal. B 209, 29

161-173. Qin, Y., Li, H., Dong, H., Ma, C., Li, X., Liu, X., Liu, Y., Li, C., Yan, Y., 2018. Enhanced photocatalytic performance and stability of visible-light-driven Z-scheme CdS/Ag/g-C3N4 nanosheets photocatalyst. New J. Chem. 42, 12437-12448. Ran, J., Yu, J., Jaroniec, M., 2011. Ni(OH)2 modified CdS nanorods for highly efficient visible-light-driven photocatalytic H2 generation. Green Chem. 13, 2708-2713. Reddy, K.R., Reddy, C.V., Nadagouda, M.N., Shetti, N.P., Jaesool, S., Aminabhavi, T.M., 2019. Polymeric graphitic carbon nitride (g-C3N4)-based semiconducting nanostructured materials: Synthesis methods, properties and photocatalytic applications. J. Environ. Manage. 238, 25-40. Wang, D., Pillai, S.C., Ho, S.H., Zeng, J., Yi, L., Dionysiou, D.D., 2018. Plasmonic-based nanomaterials for environmental remediation. Appl. Catal. B 237, 721-741. Wang, F., Chen, P., Feng, Y., Xie, Z., Liu, Y., Su, Y., Zhang, Q., Wang, Y., Yao, K., Lv, W., Liu, G., 2017a. Facile synthesis of N-doped carbon dots/g-C3N4 photocatalyst with enhanced visible-light photocatalytic activity for the degradation of indomethacin. Appl. Catal. B 207, 103-113. Wang, F., Ping, C., Feng, Y., Xie, Z., Yang, L., Su, Y., Zhang, Q., Wang, Y., Yao, K., Lv, W., 2017b. Facile synthesis of N-doped carbon dots/g-C3N4 photocatalyst with enhanced visible-light photocatalytic activity for the degradation of indomethacin. Appl. Catal. B 207, 103-113. Wang, F., Wang, Y., Feng, Y., Zeng, Y., Xie, Z., Zhang, Q., Su, Y., Chen, P., Liu, Y., Yao, K., 2017c. Novel ternary photocatalyst of single atom-dispersed silver and carbon quantum dots co-loaded with ultrathin g-C3N4 for broad spectrum photocatalytic degradation of naproxen. Appl. Catal. B 221, 510-520. Wang, J., Ng, Y., Lim, YF., Ho, G., 2014. Vegetable-extracted carbon dots and their nanocomposites for enhanced photocatalytic H2 production. Rsc Adv. 4, 44117-44123. Wang, J., Tang, L., Zeng, G., Deng, Y., Dong, H., Liu, Y., Wang, L., Peng, B., Zhang, C., Chen, F., 2018. 0D/2D interface engineering of carbon quantum dots modified Bi2WO6 ultrathin nanosheets with enhanced photoactivity for full spectrum light utilization and mechanism insight. Appl. Catal. B 222, 115-123. Wang, X., Maeda, K., Chen, X., Takanabe, K., Domen, K., Hou, Y., Fu, X., Antonietti, M., 2009. Polymer semiconductors for artificial photosynthesis: hydrogen evolution by mesoporous graphitic carbon nitride with visible light. J.Am. Chem. Soc. 131, 1680-1681. Xiao, X., Wei, J., Yang, Y., Rui, X., Pan, C., Jing, S., 2016. Photoreactivity and mechanism of g-C3N4 and Ag Co-modified Bi2WO6 microsphere under visible light irradiation. Acs Sustain. Chem. Eng. 4, 3017-3023. Xu, Q., Zhang, L., Yu, J., Wageh, S., Al-Ghamdi, A.A., Jaroniec, M., 2018. Direct Z-scheme photocatalysts: Principles, synthesis, and applications. Mater. Today 21, 1042-1063. Xu, X.Y., Bao, Z.J., Zhou, G., Zeng, H.B., Hu, J.G., 2016. Enriching photoelectrons via three transition channels in amino conjugated carbon quantum dots to boost photocatalytic hydrogen generation. Acs Appl. Mater. Inter. 8, 14118-14124. Xuan, J., Li, J.G., Yang, H.M., Cao, L.L., Zhang, E.H., Liang, Z.H., 2017. Carbon quantum dots reinforced polypyrrole nanowire via electrostatic self-assembly strategy for high-performance supercapacitors. Carbon 114, 533-543. Xuan, J., Xian, L., Yang, H.M., Li, J.G., Song, X.L., Dai, H.Y., Liang, Z.H., 2016. Construction of carbon quantum dots/proton-functionalized graphitic carbon nitride nanocomposite via electrostatic self-assembly strategy and its application. Appl. Surf. Sci. 370, 514-521. Yin, C., Cui, L., Pu, T., Fang, X., Shi, H., Kang, S., Zhang, X., 2018. Facile fabrication of nano-sized hollow-CdS@g-C3N4 Core-shell spheres for efficient visible-light-driven hydrogen evolution. Appl. Surf. Sci. 456, 464-472. 30

Yin, W., Bai, L., Zhu, Y., Zhong, S., Zhao, L., Li, Z., Bai, S., 2016. Embedding metal in the interface of a p-n heterojunction with a stack design for superior Z-scheme photocatalytic hydrogen evolution. Appl. Surf. Sci. 8, 23133-23142. Yu, H., Chen, F., Chen, F., Wang, X., 2015. In situ self-transformation synthesis of g-C3N4-modified CdS heterostructure with enhanced photocatalytic activity. Appl. Surf. Sci. 358, 385-392. Yuan, S., Wang, Y., Zhang, Y., Zhu, W., Lv, X., Wu, Y., Zhou, Y., Chen, W., 2019. CdS nanospheres hybridized with graphitic C3N4 for effective photocatalytic hydrogen generation under visible light irradiation. Appl. Organomet. Chem. 33(2). Zhang, H., Zhao, L., Geng, F., Guo, L.H., Wan, B., Yang, Y., 2016a. Carbon dots decorated graphitic carbon nitride as an efficient metal-free photocatalyst for phenol degradation. Appl. Catal. B 180, 656-662. Zhang, L., Huang, F., Liang, C., Zhou, L., Zhang, X., Pang, Q., 2016b. Ultrasound exfoliation of g-C3N4 with assistance of cadmium ions and synthesis of CdS/g-C3N4 ultrathin nanosheets with efficient photocatalytic activity. J. Taiwan Inst. Chem. E. 60, 643-650. Zhao, Q., Yang, C., Zhao, H., Gao, G., Jiang, T., 2018. Fabrication of ternary structure photocatalyst CdS/CQD@mpg-C3N4 with enhanced photocatytic performance based on synergistic effect. Chemistryselect 3, 11824-11834. Zhou, P., Yu, J., Jaroniec, M., 2014. All-solid-state Z-scheme photocatalytic systems. Adv. Mater. 26, 4920-4935. Zhou, Y., Zhang, L., Huang, W., Kong, Q., Fan, X., Wang, M., Shi, J., 2016. N-doped graphitic carbon-incorporated g-C3N4 for remarkably enhanced photocatalytic H2 evolution under visible light. Carbon 99, 111-117.

31

Figure Captions：

Fig. 1. XRD patterns of the pure g-C3N4(CN), pure CdS, CQDs(0.5%)/g-C3N4 (5CCN), CdS(0.25%)/g-C3N4 (2SCN) and CdS(0.25%)/CQDs(0.5%)/g-C3N4 (2S5CCN) composites (a), together with the magnified parts of CN, 5CCN, 2SCN and 2S5CCN samples (b).

Fig.2. FT-IR spectra of the pure CdS, CdS(0.25%)/g-C3N4 (2SCN) and CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) composites(a); pure g-C3N4(CN), CQDs(0.5%)/g-C3N4 (5CCN) and S5CCN samples (b).

32

Fig.3. XPS spectra of CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) (a) survey; (b) C 1s; (c) N 1s; (d) Cd 3d; (e) S 2p;

33

Fig. 4. TEM images of g-C3N4 (a), CdS(b-c), CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) composites (d); HRTEM images of 2S5CCN composites(e-f); Elemental mapping patterns (g) of 2S5CCN.

34

Fig. 5. UV–vis DRS (a) and (Ahν)2 versus hνcurve (b) of g-C3N4(CN), pure CdS, CdS(0.25%)/g-C3N4 (2SCN), CQDs(0.5%)/g-C3N4 (5CCN) and CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN); up-converted photoluminescence spectra (c) of the as-synthetic CQDs.

35

Fig.6. PL spectra (a) and EIS spectra (b) of g-C3N4 (CN), CdS (0.25%)/g-C3N4(2SCN), CQDs(0.5%)/g-C3N4 (5CCN) and CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) composites; Photocurrent responses (c) of CN, 2SCN , 5CCN and 2S5CCN composites under visible light.

36

Fig.7. Rate of RhB degradation (a) and the 10 min Kobs (b) for g-C3N4(CN) and CdS/CQDs/g-C3N4(SCCN)composites under visible light illumination; Rate of RhB degradation using unit (CN, CdS, CQDs), optimized CdS(0.25%)/g-C3N4(2SCN) and CQDs(0.5%)/g-C3N4(5CCN) binary composites and CdS(0.25%)/CQDs(0.5%)/g-C3N4(2S5CCN) composites(c); Rate of RhB degradation of CN, CdS, CQDs, 2SCN, 5CCN and 2S5CCN under long wavelength ( > 820nm) irradiation (d).

37

Fig.8. Rate of MB degradation (a) and the Kobs (b) for g-C3N4(CN) , CdS, CdS(0.5%)/g-C3N4 (5SCN) , CQDs(0.5%)/g-C3N4 (5CCN) and CdS(0.5%)/CQDs(0.5%)/g-C3N4 (5S5CCN) composites under visible light illumination; Rate of phenol degradation (c) and the Kobs (d) for CN, CdS, 2SCN, 5CCN and 2S5CCN composites under visible light illumination

38

Fig.9. Cycling photocatalytic degradation of RhB over Z-scheme 2S5CCN composites(a); Effect of different scavengers on the RhB degradation over 2S5CCN composites(b); EPR spectra of the (c) DMPO-•O2− and (d) DMPO-•OH adducts recorded with CN, 2SCN, 2S5CCN under visible light irradiation (water system =1mL, DMPO = 40.0 μL, catalyst=10 mg).

39

Fig. 10. Energy band diagram of the Z-scheme electron transfer (CdS/CQDs/g-C3N4) mechanism.

40

Declaration of interests

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

☐The authors declare the following financial interests/personal relationships which may be considered as potential competing interests:

41

42

Highlights (1) A novel Z-scheme photocatalyst CdS/CQDs/g-C3N4 was synthesized. (2) The photocatalyst can response under the visible light and near infrared light. (3) The Z-scheme composites exhibited enhanced photocatalytic activity and photostability. (4) CQDs can shuttle the electrons in the composites’ interface via the Z-scheme transfer pathway. (5) CQDs can capture near-infrared light through the upconversion fluorescence property.

43