- Email: [email protected]
graphene-like C3N4
Molecular Catalysis 438 (2017) 103–112
Contents lists available at ScienceDirect
Molecular Catalysis journal homepage: www.elsevier.com/locate/mcat
Editor’s choice paper
Kinetics and mechanism of enhanced photocatalytic activity employing ZnS nanospheres/graphene-like C3 N4 Jia Yan a , Yamin Fan a , Jiabiao Lian a , Yan Zhao a , Yuanguo Xu a , Jiemin Gu a , Yanhua Song b , Hui Xu a,∗ , Huaming Li a,∗ a b
School of Chemistry and Chemical Engineering, Institute for Energy Research, Jiangsu University, Zhenjiang 212013, PR China School of Environmental and Chemical Engineering, Jiangsu University of Science and Technology, Zhenjiang 212003, PR China
a r t i c l e
i n f o
Article history: Received 18 December 2016 Received in revised form 10 May 2017 Accepted 23 May 2017 Keywords: ZnS Graphene-like C3 N4 Photocatalyst Methyl orange Tetracycline
a b s t r a c t In this work, a novel photocatalyst ZnS nanospheres/graphene-like g-C3 N4 (ZnS/GL-C3 N4 ) nanocomposite was synthesized by a simple agitation method. Two-dimensional (2D) nanomaterial GL-C3 N4 , synthesized by thermal exfoliation from g-C3 N4 , showed large surface area and manifested efficient photocatalytic activity than bulk g-C3 N4 . ZnS nanospheres were well anchored and covered on the surface of GL-C3 N4 nanosheets and a synergetic effect between the ZnS and GL-C3 N4 could highly contributed to improvement of the light adsorption capability of GL-C3 N4 as well as increasement of the separation efficiency of photon-generated e− -h+ pairs, therefore, enhancing its photocatalytic performance under the illumination of visible light. Methyl orange (MO, a kind of organic dyes which is hard to be degraded by pure C3 N4 ) and tetracycline (TC, a representative broad-spectrum colorless antibiotic agent) were chosen as the targets of pollutants for degradation in this study. The optimum photocatalytic MO degradation of ZnS/GL-C3 N4 (50%) was almost 3.48 and 12.4 times higher than that of pure ZnS and GL-C3 N4 , respectively. 91% TC was photodegraded in the presence of ZnS/GL-C3 N4 (50%) and higher than that of GL-C3 N4 . Furthermore, kinetics and possible photocatalytic mechanism of MO degradation under visible light was proposed in detail. Except for the synergetic effect, the trapping experiments and ESR spectra demonstrated that not only O2 • − and holes, but also • OH were active species playing an importing role in this system for effective degradation. Our research results open an easy pathway for developing highly efficiency photocatalyst. © 2017 Published by Elsevier B.V.
1. Introduction Energy is the material basis of human survival and the important material guarantee of economic sustainable development [1]. However, with the development of society, energy and environmental issues are becoming grave threats to the sustainable development of human society [2]. Solar energy has attracted more and more attention, due to its inexhaustible, clean, renewable properties. Semiconductor-based photocatalysis via sunlight-driven photoredox reactions to mineralize organic pollutants or to directly convert solar energy into chemical energy is regarded as a long-term solution to completely eliminate the environmental issues and energy shortage [3]. As we all know, TiO2 is widely used as semiconductor photocatalyst in environmental remediation for its high photocatalytic
∗ Corresponding authors. E-mail addresses: [email protected] (H. Xu), [email protected] (H. Li). http://dx.doi.org/10.1016/j.mcat.2017.05.023 2468-8231/© 2017 Published by Elsevier B.V.
activity, non-toxicity, cheap and readily availability, chemical stability, and other advantages. However, the traditional TiO2 has limitations on visible light application, due to its wide band gap [4]. It is urgent to seek for efficient visible light driven photocatalysts. Graphitic carbon nitride (g-C3 N4 ) is one of non-metal semiconductor, possessing high thermal and chemical durability as well as the interesting electronic properties, which have made them become a kind of significant materials in the field of photocatalysis [5,6]. It has been reported that the band gap energy of g-C3 N4 is 2.7 eV, making it be a good photocatalyst for use of visible light [7]. However, the photocatalytic efficiency of g-C3 N4 is limited because of the high recombination rate of e− -h+ pairs [5]. On the other side, the special electronic structure of g-C3 N4 make it be an excellent candidate for all sorts of functional materials to be coupled with, such as TiO2 /g-C3 N4 [8], ZnO/g-C3 N4 [9], MoS2 /gC3 N4 [10], rGO/g-C3 N4 [11], AgX/g-C3 N4 [7], which can decrease the high recombination rate of e− -h+ pairs and thus improve the photocatalytic performance [12,13].
104
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
It is well known that ZnS has some advantages as a good photocatalyst with eco-friendliness, excellent stability, and low cost [14]. It can rapidly generate electrons and holes by photoexcitation and has higher negative reduction potentials of the excited electrons. However, the photocatalytic efficiency of ZnS is limited for its confined surface area, low transference ability of the photoexcited carriers [14], and the most important inherent defect-wide band gap (about 3.45 eV) which imposes restrictions on its photocatalytic application in the visible-light region [15]. As a result, coupling g-C3 N4 with ZnS together to construct a g-C3 N4 /single metal sulfide photocatalyst [12] is a favorable way to enhance photocatalytic performance in the visible light region [16–19]. Unfortunately, the recombination rate of e− -h+ pairs of g-C3 N4 /ZnS nanocomposite is still high due to the own flaws of bulk g-C3 N4 . It has been reported that reduce the thickness of g-C3 N4 [1], especially g-C3 N4 monolayer structure has high electron and hole transfer property. Therefore, using the g-C3 N4 monolayer or few layer structures to form the nanocomposite is a way to get the high photocatalytic activity. So we expect to achieve much thinner g-C3 N4 and uniformly ZnS morphology, in order to synthesize the g-C3 N4 /ZnS nanocomposites with high photocatalytic activity. According to reports, few layer or graphene-like g-C3 N4 (GL-C3 N4 ) [20] synthesized by thermal exfoliation from g-C3 N4 , possessing 2D thin-layer structure with 6–9 atomic thickness (2–3 nm), large specific surface area, enhanced photocurrent response and electron transport ability, manifested efficient photocatalytic activity under visible light [21], and could be used as a photo-electrochemical sensor [1]. As we investigated, most of the photocatalysts took rhodamine B (RhB) as the target of pollutants which was easier to be degraded compared with methyl orange (MO) and yet the common colorless antibiotics degradation has not been studied. Tetracycline (TC), as a representative colorless broad-spectrum antibiotic agent, would cause multiple negative influences of the hierarchical system by inducing proliferation of bacterial drug resistance. Therefore, it is very important for the removal of TC and it was chosen as the degradation target to evaluate the photocatalytic activity of the photocatalyst. Meanwhile, the photocatalytic degradation mechanism has not been clearly clarified. In this work, ZnS nanospheres/GL-C3 N4 (ZnS/GL-C3 N4 ) nanocomposites were synthesized by depositing ZnS nanospheres onto the GL-C3 N4 . We demonstrated that GL-C3 N4 provided large surface area for the uniform distribution of ZnS nanospheres leading to the improved adsorption capability ZnS nanospheres with high surface also enhanced the contact areas in nanocomposits. The interaction between ZnS and GL-C3 N4 allowing the effective charge transfer and promoting the photo-generated e− -h+ pairs separation, which would highly make contribution to the improvement of the pollutants degradation. MO and TC were chosen as the targets of pollutants in this system for one was hard to be degraded by C3 N4 and another was a representative broad-spectrum colorless antibiotic agent, respectively. The photocatalytic efficiency of MO degradation over ZnS/GL-C3 N4 was conducted under the illumination of visible light, which showed almost 3.48 and 12.4 times higher than that of pure ZnS and GL-C3 N4 , respectively. 91% TC was photodegraded in the presence of ZnS/GL-C3 N4 (50%) higher than that of GL-C3 N4 . Photoluminescence (PL) spectra clearly showed that the introduction of ZnS decreased the recombination rate of photo-generated e− -h+ pairs, which improved the photocatalytic efficiency of GL-C3 N4 . GL-C3 N4 has lager band gap energy (2.85 eV) and more negative CB level compared with g-C3 N4 . It was favorable for energy band alignment construction with ZnS nanospheres which can serve as the acceptor of the photo-generated electrons coming from GL-C3 N4 . It was also demonstrated that both O2 •− , • OH and holes were the active species in photocatalytic degradation under the
illumination of visible light. Moreover, a possible mechanism was also proposed in detail. 2. Experimental section 2.1. Preparation of photocatalysts The materials were analytical grade (99%), and used as received without any purification. GL-C3 N4 were synthesized according to the reported procedure [1]. ZnS nanosphere was synthesized by using a simple hydrothermal method: 1 mmol ZnAc2 ·2H2 O and 1.5 g PVP were dispersed in 20 mL deionized water by mechanical stirring for 6 h, then TAA (1 mmol) was dropped into the mixture to active a clear solution. Then the clear solution was then transferred to a 20 mL Teflon-lined stainless steel autoclave and kept in an oven at 180 ◦ C for 6 h. At last the product was collected by centrifugation, washed with distilled water and absolute ethanol repeatedly, and then dried to get the sediment. The ZnS/GL-C3 N4 nanocomposite photocatalysts were synthesized by mechanical agitation method. In a typical synthesis, 0.005 g ZnS and 0.095 g GL-C3 N4 were dispersed in 50 mL distilled water by ultra-sonication for 60 min, and then the mixture was mechanical stirring for 48 h. At last, the product was collected by centrifugation, washed with deionized water and ethanol repeatedly, and then dried at 60 ◦ C. The as prepared sample was named as ZnS/GLC3 N4 (5%). Several samples with different weight percent of ZnS (10 wt%, 20 wt%, 40 wt% and 50 wt%) were synthesized by using the similar procedure, and labeled as ZnS/GL-C3 N4 (10%), ZnS/GL-C3 N4 (20%), ZnS/GL-C3 N4 (40%), ZnS/GL-C3 N4 (50%), respectively. 2.2. Characterization of photocatalyst The crystal structure of ZnS/GL-C3 N4 nanocomposites were investigated by power X-ray diffraction (XRD) measurements which were performed on Bruker D8 diffractometer using Cu K␣ radiation ( = 1.5418 À). The chemical state of elements was analyzed by X-ray photoelectron spectroscopy (XPS) on an ESCA Lab MKII X-ray photo-electron spectrometer using Mg K␣ line source. The microstructure and grain morphology were researched by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). SEM was taken with a FESEM MO del JEOL JSM-7001F and TEM was used a JEOL-JEM-2010 (JEOL, Japan) operated at 200 kV. The chemical composition was investigated by X-ray energy dispersion spectrum (EDS) performed at an acceleration voltage of 10 kV. UV–vis diffuse reflection spectra (DRS) were performed on UV-2450 UV–vis system (Shimadzu, Japan) with the reflectance standard material of BaSO4 . Photocurrents and electrochemical impedance spectroscopy (EIS) were recorded on CHI 660B electro-chemical workstation (Chenhua Instrument Company, China). PL experiments were conducted under a 325 nm light excitation . X-band Electron Spin Resonance (ESR) spectra were operated on a JES FA200 spectrometer at ambient temperature. 2.3. Photocatalytic degradation Photocatalytic properties of photocatalysts were estimated by the degradation of MO and TC under the illumination of 300W Xe lamp with 400 nm cut-off filter in custom-made photochemical reactor at 30 ◦ C under constant stirring. In every experiment, 0.0500 g photocatalyst powder was added in 50 mL of MO solution (10 mg L−1 ) or TC (20 mg L−1 ). The reaction mixture was continuously aerated by a pump to provide O2 . Prior to illumination, the mixed liquor was magnetically stirred in dark for 30 min to ensure an adsorption-desorption equilibrium between photocatalyst and pollutant. During every interval, 4.0 mL solution suspension would be taken out to separate the photocatalyst by
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 1. XRD patterns of ZnS/GL-C3 N4 composites.
centrifugation (13000 rpm, 3 min) and researched by UV–vis spectrophotometer at 463 nm and 356 nm which were the maximal absorption peak of MO and TC, respectively. The following formula was used for obtain the photocatalytic degradation efficiency (E) of MO and TC: E = (1 −
C A ) × 100% = (1 − ) × 100% C0 A0
Where C is the concentration of the solution suspension at reaction time t, and A is the corresponding values of absorbancy, C0 is the adsorption/desorption equilibrium concentration of solution suspension, and A0 is the corresponding values of absorbancy. 2.4. Photoelectrochemical measurements To survey the transition of photo-induced electrons, the photocurrents were measured in a standard three-electrode system. Working electrodes of GL-C3 N4 and ZnS/GL-C3 N4 were prepared as follows: 20 L of the dispersion (5 mg mL−1 ) was dropped onto a piece of ITO (0.5 × 1 cm2 ) denoted as GL-C3 N4 /ITO, ZnS/GLC3 N4 /ITO. A platinum wire was employed as counter electrode while a saturated Ag/AgCl electrode was served as reference electrode. A 500 W Xe xenon lamp equipment was adopted as the photosource. The electrolyte solution of photocurrent was phosphate buffered saline (PBS, 0.1 mol L−1 , pH = 7.0). The Nyquist plots was performed in a 0.1 M KCl solution containing 5 mM Fe(CN)6 3− /Fe(CN)6 4− . Sunless conditions were taken at all electrochemical measurements. 3. Results and discussion 3.1. Structure and morphology characterization Fig. 1 showed the XRD patterns of GL-C3 N4 , ZnS and ZnS/GLC3 N4 with different ZnS percentages. From those XRD patterns, it was clearly seen that the patterns of all the synthesized ZnS/GLC3 N4 samples were combined by the diffraction peaks of both ZnS and GL-C3 N4 . While the peak at 27.8◦ and 13.1◦ could be indexed to the (002) and (100) diffraction planes of GL-C3 N4 , which was consistent with the reported literature [1]. For ZnS, the main diffraction peaks at 2 = 28.6◦ , 47.6◦ , 56.5◦ and 76.9◦ corresponding to (111), (220), (311) and (331) plane, respectively [16,22–24]. All the peaks can be indexed to the cubic phase (JCPDS 65-0309). The peaks of ZnS (28.6◦ ) and GL-C3 N4 (27.8◦ ) were very close and overlap with
105
each other. On the other hand, as the increase of ZnS content in the composites, the peaks of GL-C3 N4 weaken while the peaks of ZnS got strong. XPS analyses were adopted to determine the chemical state of every element in the composites. As shown in Fig. 2A, the ZnS/GLC3 N4 (50%) were consisted of C, N, Zn, S, and O elements, while the trace of O species was due to contamination from the atmosphere. High-resolution Zn 2p, S 2p, C 1s, N 1s spectra of the composites was shown in Fig. 2B–E. The Zn 2p peaks of ZnS/GL-C3 N4 (50%) were located at 1021.8 and 1044.6 eV (Fig. 2B), which corresponded to the Zn 2p3/2 and Zn 2p1/2 binding energies, respectively, indicating that the Zn ions existed as Zn2+ according to the previous results [25]. The S 2p peak located at 161.4 and 162.6 eV (Fig. 2C) could be due to S 2p3/2 and 2p1/2 lines, respectively, which indicated the presence of S2− [14]. From Fig. 2D, it could be observed that C 1s for GL-C3 N4 and ZnS/GL-C3 N4 had a same peak at 284.6 eV related to C C or carbon contamination from the air. The C 1s peak at 288.1 and 287.7 eV were related to N C N coordination in GL-C3 N4 and ZnS/GL-C3 N4 , respectively [20,26]. The shift could be ascribed to the interaction between ZnS and GL-C3 N4 , which could influence the chemical environment of sp2 -hybridized carbon. The N 1s peak (Fig. 2E) originating from C N C coordination [26] was located at 398.8 eV for GL-C3 N4 , and it shifted to 398.3 eV for the ZnS/GL-C3 N4 composite, which also proved that there was interaction between ZnS and GL-C3 N4 . This interaction allowed for the charge transfer between ZnS and GL-C3 N4 , promoting the separation of photo-generated e− -h+ pairs, and subsequently increasing the photocatalytic efficiency of the composites. Fig. 3 showed the SEM and TEM images of ZnS, GL-C3 N4 , ZnS/GLC3 N4 (50%) nanocomposites and the EDS of ZnS/GL-C3 N4 (50%) composite. As shown in Fig. 3A and D, ZnS exhibited a uniform spherical shape with a particle size of 0.4–0.5 m in diameter. While GL-C3 N4 was a fewer layered structure from Fig. 3C, which was consistent with previous reports [1]. Fig. 3B and E were the SEM and TEM images of ZnS/GL-C3 N4 (50%), respectively. It exhibited that ZnS nanospheres were equably deposited on the surface of GLC3 N4 and this interaction allowed for the charge transfer between ZnS and GL-C3 N4 , promoting the separation of photo-generated e− -h+ pairs, increasing the photocatalytic efficiency of composites. HRTEM of ZnS/GL-C3 N4 (50%) was shown in Fig. 3F, allowing identification of the crystallographic spacing, and the spacing of d = 0.31, 0.19, 0.16 nm corresponding to the (111), (220), (331) plane of ZnS, respectively. An HRTEM image illustrated the existence of sufficient contact between ZnS and GL-C3 N4 which would be favorable for the transfer of photoexcited carriers. EDS in Fig. 3G showed that the ZnS/GL-C3 N4 (50%) photocatalyst was only composed of C, N, Zn and S, which demonstrated the purity of the composites. Meanwhile, N2 sorption-desorption isotherm (Fig. 4) was employed to determine surface area of the photocatalysts. The surface areas were estimated to be 3.30, 30.10, 46.56, 41.73 and 42.96 m2 g−1 for g-C3 N4 , GL-C3 N4 , ZnS, ZnS/GL-C3 N4 (10%) and ZnS/GL-C3 N4 (50%), respectively. Surface area of GL-C3 N4 was 10 times higher than that of bulk g-C3 N4 and ZnS nanospheres also has big surface area. Therefore, ZnS/GL-C3 N4 nanocomposites had large surface areas which benefited for the charge transfer and the improvement of the photocatalytic performance. 3.2. Optical and electronic properties UV-vis absorption spectra of ZnS, GL-C3 N4 and nanocomposites were shown in Fig. 5A. It was clearly seen that the absorption edge of the ZnS was at about 378 nm, but the basal absorption edge of ZnS/GL-C3 N4 came to visible region at about 460 nm. While in the low-energy visible light region, the absorption intensity of these nanocomposites strengthened in comparison with GL-C3 N4 . The improvement of adsorption capability could increase the use effi-
106
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 2. High-resolution XPS survey spectra of GL-C3 N4 and ZnS/GL-C3 N4 materials: (A) XPS survey spectrum of ZnS/GL-C3 N4 (50%), (B) Zn 2p, (C) S 2p, (D) C 1s, (E) N 1s.
ciency of visible light and consequently led to enhancement of the photocatalytic efficiency. The band-gap energies (Eg ) of ZnS and GL-C3 N4 could be estimated respectively according to the following formula Eq. (Ah)2 = h-Eg
Where h, , Eg and A were Planck constant, light frequency, band gap energy, and absorbance, respectively. From Fig. 5B, it could be seen that the Eg of the GL-C3 N4 was calculated to be 2.85 eV and the Eg of ZnS was calculated to be 3.45 eV. Fig. 6 showed the PL spectra of the GL-C3 N4 and different ZnS/GL-C3 N4 samples. Under a 325 nm light excitation, all the samples showed the broad peak at 450 nm, which was corre-
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
107
Fig. 3. SEM images: (A) ZnS, (B) ZnS/GL-C3 N4 (50%); TEM images: (C) GL-C3 N4 , (D) ZnS, (E) ZnS/GL-C3 N4 (50%); (F) HRTEM and (G) EDS of ZnS/GL-C3 N4 (50%).
sponded to the band gap of DRS [27]. The increasing percentage of ZnS decreased the photoluminescence intensity, indicating that the ZnS/GL-C3 N4 samples had lower recombination rate of photo-
generated e− -h+ pairs than that of GL-C3 N4 , providing an evidence for the effective charge transfer between ZnS nanospheres and GLC3 N4 nanosheets.
108
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 4. Nitrogen absorption-desorption isotherms of (A) g-C3 N4 , (B) GL-C3 N4 , (C) ZnS, and ZnS/GL-C3 N4 nanocomposites: (D) 10%; (E) 50%.
In order to further confirm the important role of nanocomposites in the photocatalytic reaction, five switch source loop transient photocurrent response experiment were performed. In Fig. 7, the ZnS/GL-C3 N4 nanocomposites showed higher photocurrent intensity than that of GL-C3 N4 under the light irradiation. And ZnS/GL-C3 N4 (50%) showed highest photocurrent intensity compared with other ZnS content nanocomposites, which was consistent with the photocatalytic activities. Photocurrent results confirmed that the ZnS/GL-C3 N4 had a higher efficient separation of photo-generated e− -h+ pairs under the illumination of visible
light. Above all, we confirmed that there were effective interactions between the ZnS and GL-C3 N4 , which was favorable for photogenerated electrons excited from VB to CB of GL-C3 N4 transfer to ZnS, playing an important role in the effective separation of e− -h+ pairs of GL-C3 N4 . EIS measurements were conducted to gain deep understanding into the charge transport behavior of ZnS/GL-C3 N4 (Fig. 8). It was proved that the radius of every circle in the Nyquist diagram was based on the charge transfer process at the interface of working electrode/electrolyte, and the small radius were con-
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
109
Fig. 7. Transient photocurrent response for the GL-C3 N4 and ZnS/GL-C3 N4 nanocomposites in PBS (pH = 7.0) aqueous solution under visible light irradiation.
Fig. 5. (A) UV–vis diffuse reflectance spectra of the samples; (B) Estimated band gap of ZnS and GL-C3 N4 . Fig. 8. EIS of the GL-C3 N4 and ZnS/GL-C3 N4 nanocomposites in a 0.1 M KCl solution containing 5.0 mM Fe(CN)6 3− /Fe(CN)6 4− .
the smallest radius demonstrating a high-efficiency separation of photo-generated e− -h+ pairs. Anyway, EIS results were consistent with that of the five switch source loop transient photocurrent response experiment and PL spectra. 3.3. Photocatalytic efficiency of the composites
Fig. 6. PL spectra of the GL-C3 N4 and ZnS/GL-C3 N4 composites.
formed with low resistance [28,29], which was favorable for the charge transfer. It could be seen from Fig. 8 that there was a decline in the radius when the content of ZnS was increased from 5% to 50% of the system, indicating that the introduction of ZnS could decrease charge-transfer resistance and enhance the interfacial charge transfer ability. ZnS/GL-C3 N4 (50%) showed
From the characterizations mentioned above, the nanocomposites showed improved adsorption capability, high separation efficiency. They were supposed to be efficient visible-light driven photocatalysts. Before illumination, the mixed liquor was magnetically stirred for 30 min in dark. In the actual experiment it was found that ZnS, GL-C3 N4 , and ZnS/GL-C3 N4 nanocomposites have negligible absorption toward MO in dark. Therefore, the photocatalytic activity will be enhanced only in the presence of ZnS/GL-C3 N4 catalyst with light irradiation. As shown in Fig. 9A, GL-C3 N4 and ZnS showed weak photocatalytic activity in MO degradation, while ZnS/GL-C3 N4 composites showed higher photocatalytic degradation efficiency than that of GL-C3 N4 and ZnS. From 5% to 50% component of ZnS in ZnS/GL-C3 N4 , the photocatalytic activity gradually increased, and with further increasing the ZnS concentration to 60% and 70%, respectively, the photocatalytic activity decreased. ZnS/GL-C3 N4 (50%) nanocomposite showed the highest photocatalytic activity under visible-light irradiation in 60 min (Fig. 9A). So,
110
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 10. Photocatalytic activities of the samples for the degradation of TC under visible-light irradiation.
photocatalytic activity in MO degradation than that of champion sample of the ZnS/GL-C3 N4 (50%) composites. The control experiment results confirmed that the chemical interaction between the two semiconductors in ZnS/GL-C3 N4 composites, which allowed for the charge transfer between ZnS and GL-C3 N4 , and efficient photo-generated e− -h+ pairs separation, was important for the increasement of photocatalytic efficiency. Furthermore, it was found that the photocatalytic degradation of MO over all the studied photocatalysts followed first-order reaction dynamics under the illumination of visible-light. The first order equation model was shown as follows: −ln(C/C 0 ) = kt
Fig. 9. (A) Photocatalytic activities of the samples for the degradation of MO under visible-light irradiation; (B) Kinetic fit for the degradation of MO with the samples; (C) Cycling runs of the ZnS/GL-C3 N4 (50%) sample.
an appropriate ZnS ratio in nanocomposite was critically significant for the improvement of the resultant photocatalytic performance. The results showed the ZnS import was in favor of GL-C3 N4 photoactivity promotion under visible light and the existence interaction between ZnS and GL-C3 N4 could highly enhance e− -h+ pairs separation efficiency. Form comparision, a physically blending of ZnS and GL-C3 N4 with an appropriate ratio of 1:1 was also used as one of the photocatalysts. As could be seen from Fig. 9A, the 50% physically blending of ZnS and GL-C3 N4 showed the negligible
Where C is the current concentration of MO (mg L−1 ) at reaction time t, C0 is the initial concentration of MO (mg L−1 ), k is the reaction rate constant (min−1 ). The experimental data and the fitting curves achieved according to the equation was shown in Fig. 9B. It could be seen that the reaction rate constant over ZnS/GL-C3 N4 (50%) nanocomposite was 5.4 times higher than that of GL-C3 N4 . Meanwhile, all the investigated nanocomposites showed larger value of k than GL-C3 N4 , demonstrating the fast reaction kinetics for MO degradation, suggestive a synergistic effect of components for MO degradation. The stability of the synthesized ZnS/GL-C3 N4 (50%) nanocomposite was studied by a 5-run cycling test under the same condition. In Fig. 9C, the photocatalytic efficiency of ZnS/GL-C3 N4 (50%) had just a little loss and the degradation efficiency of MO was still near 80% after 5 recycles under the illumination of visible light. The composition stability of ZnS/GL-C3 N4 (50%) was also investigated by XRD analyses. It could be seen from Fig. 1, the recycled ZnS/GL-C3 N4 (50%) nanocomposite named “50% after” had no obvious changes, indicating that the ZnS/GL-C3 N4 (50%) remained effective and recyclable ability for MO degradation. TC can cause multiple negative influences of each level of the hierarchical system by inducing proliferation of bacterial drug resistance as a representative colorless broad-spectrum antibiotic agent. Therefore, it is very important to be removed and it was chosen to evaluate the photocatalytic activity of the ZnS/GL-C3 N4 (50%) via the photocatalytic degradation under visible light irradiation. As shown in Fig. 10, the result revealed 91% TC was photodegraded in the presence of ZnS/GL-C3 N4 (50%) and exhibited 1.35 times higher than that of pure GL-C3 N4 . The results imply that ZnS/GL-C3 N4 is a kind of efficient photocatalysts which could also be applied to antibiotics pollution treatment.
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
111
Fig. 12. Photodegradation of MO over ZnS/GL-C3 N4 (50%) photocatalyst with different quenchers.
Fig. 11. ESR spectra of radical adducts trapped by DMPO: (A) DMPO-O2 • − radical species detected for the ZnS/GL-C3 N4 (50%) dispersion in methanol, (B) DMPO-• OH radical species detected for the ZnS/GL-C3 N4 (50%) dispersion in water.
3.4. Schematic illustration of ZnS/GL-C3 N4 nanocomposites To check the active species produced pending the photocatalysis over these samples under the illumination of visible light, ESR spin-trap technique with DMPO (5,5-dimethyl-1-pirroline-Noxide) was used. Fig. 11 indicated that there was no ESR signal for the ZnS/GL-C3 N4 without visible light irradiation. In Fig. 11A, while visible light irradiation gave signals attributed to a DMPOO2 •− species detected successfully in ZnS/GL-C3 N4 dispersions in methanolic media, indicating that O2 •− reactive species was generated during the reaction under visible light irradiation. There were also • OH signals detected during the reaction (Fig. 11B). The results confirmed that both O2 •− and • OH were main active species which were consist with the trapping experiment result. To further investigate the mechanism, trapping experiments had also been carried out. The 2Na-EDTA, t-BuOH and N2 were added as the scavenger of hole, • OH and O2 •− [26,30]. As shown in Fig. 12, the addition of N2 and t-BuOH had significantly suppressed the degradation of MO over ZnS/GL-C3 N4 (50%) photocatalyst, while 2Na-EDTA had little effect. The results demonstrated that both O2 •− , • OH and holes were active species in this system.
The band gap was considered to under the possible photocatalytic mechanism of MO degradation in the presence of ZnS/GL-C3 N4 samples under visible light. In Fig. 13, ZnS nanospheres were equably distributed on the surface of GL-C3 N4 and the band gaps of ZnS and GL-C3 N4 were 3.45 eV [22] and 2.85 eV [20], respectively. So it was clear to know that GL-C3 N4 could be excited under the visible light due to the appropriate band gap and ZnS was inert owing to its wide band gap under visible-light irradiation. Subsequently, the corresponding photo-induced electrons in GL-C3 N4 could easily transport to ZnS and accumulate in the CB of ZnS. Meanwhile, the photo-generated electrons in the CB of ZnS (−0.91 eV) was more negative than E0 (O2 /O2 •− ) (−0.046 eV vs. NHE) which could reduce O2 to generate O2 •− . While the VB value of GL-C3 N4 (+1.65 eV) was less positive than E0 (• OH/OH− ) (+1.99 eV vs. NHE), which meant no • OH was oxidized from OH− by the hVB + of ZnS/GL-C3 N4 material. For this reason, we confirmed that • OH radical was derived from the part of the O2 •− reactive species and played a role in the degradation of MO [11,31]. However, the photo-generated holes on the VB of GL-C3 N4 could oxidize the MO directly. Corresponds to the trapping experiments and ESR results, the photocatalytic mechanism was proposed and we could confirm that both O2 •− , • OH and holes were the active species that led to effective MO degradation. 4. Conclusion In summary, the ZnS/GL-C3 N4 nanocomposites were synthesized by a simple agitation method. ZnS nanospheres were uniformly deposited on the surface of GL-C3 N4 with large surface area, the formed ZnS/GL-C3 N4 nanocomposites led to the improved adsorption capability owing to their synergetic effect. The enhanced MO and TC degradation efficiency proved that the coupling of ZnS nanospheres and GL-C3 N4 nanosheets was suitable for the energy band alignment, which was favorable for promoting the light response, improving the interface electrons transport, inhibiting the recombination efficiency of e− -h+ pairs, and absorbing more active species. All these synergetic effect made the as-prepared ZnS/GL-C3 N4 nanocomposites possess excellent stability, reusability and high photocatalytic performance with fast reaction kinetics rate compared with pure ZnS and GL-C3 N4 . Meanwhile, a detail photocatalytic reaction process was proposed and it was demonstrated that not only O2 •− and holes, but also • OH were active species played the important role for MO degradation in this system.
112
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 13. Proposed photocatalytic mechanism of MO degradation in the presence of ZnS/GL-C3 N4 samples under visible light.
Acknowledgments The authors genuinely appreciate the financial support of this work by the National Natural Science Foundation of China (21476097 and 21476098), the Natural Science Foundation of Jiangsu Province (BK20131207 and BK20130513). References [1] H. Xu, J. Yan, X.J. She, L. Xu, J.X. Xia, Y.G. Xu, Y.H. Song, L.Y. Huang, H.M. Li, Nanoscale 6 (2014), 9866–9866. [2] S.W. Cao, J.X. Low, J.G. Yu, M. Jaroniec, Adv. Mater. 27 (2015) 2150–2176. [3] J.S. Zhang, B. Wang, X.C. Wang, Prog. Chem. 26 (2014) 19–29. [4] T.T. Li, L.H. Zhao, Y.M. He, J. Cai, M.F. Luo, J.J. Lin, Appl. Catal. B 129 (2013) 255–263. [5] Q.J. Xiang, J.G. Yu, M. Jaroniec, J. Phys. Chem. C 115 (2011) 7355–7363. [6] M. Tahir, C.B. Cao, N. Mahmood, F.K. Butt, A. Mahmood, F. Idrees, S. Hussain, M. Tanveer, Z. Ali, I. Aslam, ACS Appl. Mater. Interfaces 6 (2014) 1258–1265. [7] H. Xu, J. Yan, Y.G. Xu, Y.H. Song, H.M. Li, J.X. Xia, C.J. Huang, H.L. Wan, Appl. Catal. B 129 (2013) 182–193. [8] J.Q. Yan, H. Wu, H. Chen, Y.X. Zhang, F.X. Zhang, S.Z. Frank Liu, Appl. Catal. B 191 (2016) 130–137. [9] Y.M. He, Y. Wang, L.H. Zhang, B.T. Teng, M.H. Fan, Appl. Catal. B 168 (2015) 1–8. [10] W.C. Peng, X.Y. Li, Catal. Commun. 49 (2014) 63–67. [11] B. Yuan, J.X. Wei, T.J. Hu, H.B. Yao, Z.H. Jiang, Z.W. Fang, Z.Y. Chu, Chin. J. Catal. 36 (2015) 1009–1016. [12] Z.W. Zhao, Y.J. Sun, F. Dong, Nanoscale 7 (2015) 15–37. [13] Y. Hou, Z.H. Wen, S.M. Cui, X.R. Guo, J.H. Chen, Adv. Mater. 25 (2013) 6291–6297. [14] D.A. Reddy, R. Ma, M.Y. Choi, T.K. Kim, Appl. Surf. Sci. 324 (2015) 725–735.
[15] E.S. Agorku, H. Mittal, B.B. Mamba, A.C. Pandey, A.K. Mishra, Int. J. Biol. Macromol. 70 (2014) 143–149. [16] F.F. Shi, L.L. Chen, C.S. Xing, D.L. Jiang, D. Li, M. Chen, RSC Adv. 4 (2014) 62223–62229. [17] Y.Q. Shi, S.H. Jiang, K.Q. Zhou, B.B. Wang, B. Wang, Z. Gui, Y. Hu, R.K.K. Yuen, RSC Adv. 4 (2014) 2609–2613. [18] J. Wang, P. Guo, Q.S. Guo, P.G. Jönsson, Z. Zhao, CrystEngComm 16 (2014) 4485–4492. [19] L.Q. Shao, D.L. Jiang, P. Xiao, L.M. Zhu, S.C. Meng, M. Chen, Appl. Catal. B 198 (2016) 200–210. [20] J. Yan, Z.G. Chen, H.Y. Ji, Z. Liu, X. Wang, Y.G. Xu, X.J. She, L.Y. Huang, L. Xu, H. Xu, H.M. Li, Chem. Eur. J. 22 (2016) 4764–4773. [21] X.D. Zhang, X. Xie, H. Wang, J.J. Zhang, B. Pan, Y. Xie, J. Am. Chem. Soc. 135 (2013) 18–21. [22] X.D. Zhang, X.J. Liu, L. Zhang, D.L. Li, S.H. Liu, J. Alloys Compd. 655 (2016) 38–43. [23] X.J. Wang, F.Q. Wan, K. Han, C.X. Chai, K. Jiang, Mater. Charact. 59 (2008) 1765–1770. [24] B. Subash, B. Krishnakumar, M. Swaminathan, M. Shanthi, Ind. Eng. Chem. Res. 53 (2014) 12953–12963. [25] L.X. Hu, F.Y. Chena, P.F. Hub, L.P. Zoua, X. Hua, J. Mol. Catal. A: Chem. 411 (2016) 203–213. [26] H.T. Ren, S.Y. Jia, Y. Wu, S.H. Wu, T.H. Zhang, X. Han, Ind. Eng. Chem. Res. 53 (2014) 17645–17653. [27] S.C. Yan, Z.S. Li, Z.G. Zou, Langmuir 25 (2009) 10397–10401. [28] H. Zhao, Y.M. Dong, P.P. Jiang, H.Y. Miao, G.L. Wang, J.J. Zhang, J. Mater. Chem. A 3 (2015) 7375–7381. [29] I. Shakir, M. Shahid, D.J. Kang, Chem. Eng. J. 225 (2013) 650–655. [30] H. Xu, H.Z. Zhao, Y.G. Xu, Z.G. Chen, L.Y. Huang, Y.P. Li, Y.H. Song, Q. Zhang, H.M. Li, Ceram. Int. 42 (2016) 1392–1398. [31] Q. Li, N. Zhang, Y. Yang, G.Z. Wang, D.H.L. Ng, Langmuir 30 (2014) 8965–8972.
Contents lists available at ScienceDirect
Molecular Catalysis journal homepage: www.elsevier.com/locate/mcat
Editor’s choice paper
Kinetics and mechanism of enhanced photocatalytic activity employing ZnS nanospheres/graphene-like C3 N4 Jia Yan a , Yamin Fan a , Jiabiao Lian a , Yan Zhao a , Yuanguo Xu a , Jiemin Gu a , Yanhua Song b , Hui Xu a,∗ , Huaming Li a,∗ a b
School of Chemistry and Chemical Engineering, Institute for Energy Research, Jiangsu University, Zhenjiang 212013, PR China School of Environmental and Chemical Engineering, Jiangsu University of Science and Technology, Zhenjiang 212003, PR China
a r t i c l e
i n f o
Article history: Received 18 December 2016 Received in revised form 10 May 2017 Accepted 23 May 2017 Keywords: ZnS Graphene-like C3 N4 Photocatalyst Methyl orange Tetracycline
a b s t r a c t In this work, a novel photocatalyst ZnS nanospheres/graphene-like g-C3 N4 (ZnS/GL-C3 N4 ) nanocomposite was synthesized by a simple agitation method. Two-dimensional (2D) nanomaterial GL-C3 N4 , synthesized by thermal exfoliation from g-C3 N4 , showed large surface area and manifested efficient photocatalytic activity than bulk g-C3 N4 . ZnS nanospheres were well anchored and covered on the surface of GL-C3 N4 nanosheets and a synergetic effect between the ZnS and GL-C3 N4 could highly contributed to improvement of the light adsorption capability of GL-C3 N4 as well as increasement of the separation efficiency of photon-generated e− -h+ pairs, therefore, enhancing its photocatalytic performance under the illumination of visible light. Methyl orange (MO, a kind of organic dyes which is hard to be degraded by pure C3 N4 ) and tetracycline (TC, a representative broad-spectrum colorless antibiotic agent) were chosen as the targets of pollutants for degradation in this study. The optimum photocatalytic MO degradation of ZnS/GL-C3 N4 (50%) was almost 3.48 and 12.4 times higher than that of pure ZnS and GL-C3 N4 , respectively. 91% TC was photodegraded in the presence of ZnS/GL-C3 N4 (50%) and higher than that of GL-C3 N4 . Furthermore, kinetics and possible photocatalytic mechanism of MO degradation under visible light was proposed in detail. Except for the synergetic effect, the trapping experiments and ESR spectra demonstrated that not only O2 • − and holes, but also • OH were active species playing an importing role in this system for effective degradation. Our research results open an easy pathway for developing highly efficiency photocatalyst. © 2017 Published by Elsevier B.V.
1. Introduction Energy is the material basis of human survival and the important material guarantee of economic sustainable development [1]. However, with the development of society, energy and environmental issues are becoming grave threats to the sustainable development of human society [2]. Solar energy has attracted more and more attention, due to its inexhaustible, clean, renewable properties. Semiconductor-based photocatalysis via sunlight-driven photoredox reactions to mineralize organic pollutants or to directly convert solar energy into chemical energy is regarded as a long-term solution to completely eliminate the environmental issues and energy shortage [3]. As we all know, TiO2 is widely used as semiconductor photocatalyst in environmental remediation for its high photocatalytic
∗ Corresponding authors. E-mail addresses: [email protected] (H. Xu), [email protected] (H. Li). http://dx.doi.org/10.1016/j.mcat.2017.05.023 2468-8231/© 2017 Published by Elsevier B.V.
activity, non-toxicity, cheap and readily availability, chemical stability, and other advantages. However, the traditional TiO2 has limitations on visible light application, due to its wide band gap [4]. It is urgent to seek for efficient visible light driven photocatalysts. Graphitic carbon nitride (g-C3 N4 ) is one of non-metal semiconductor, possessing high thermal and chemical durability as well as the interesting electronic properties, which have made them become a kind of significant materials in the field of photocatalysis [5,6]. It has been reported that the band gap energy of g-C3 N4 is 2.7 eV, making it be a good photocatalyst for use of visible light [7]. However, the photocatalytic efficiency of g-C3 N4 is limited because of the high recombination rate of e− -h+ pairs [5]. On the other side, the special electronic structure of g-C3 N4 make it be an excellent candidate for all sorts of functional materials to be coupled with, such as TiO2 /g-C3 N4 [8], ZnO/g-C3 N4 [9], MoS2 /gC3 N4 [10], rGO/g-C3 N4 [11], AgX/g-C3 N4 [7], which can decrease the high recombination rate of e− -h+ pairs and thus improve the photocatalytic performance [12,13].
104
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
It is well known that ZnS has some advantages as a good photocatalyst with eco-friendliness, excellent stability, and low cost [14]. It can rapidly generate electrons and holes by photoexcitation and has higher negative reduction potentials of the excited electrons. However, the photocatalytic efficiency of ZnS is limited for its confined surface area, low transference ability of the photoexcited carriers [14], and the most important inherent defect-wide band gap (about 3.45 eV) which imposes restrictions on its photocatalytic application in the visible-light region [15]. As a result, coupling g-C3 N4 with ZnS together to construct a g-C3 N4 /single metal sulfide photocatalyst [12] is a favorable way to enhance photocatalytic performance in the visible light region [16–19]. Unfortunately, the recombination rate of e− -h+ pairs of g-C3 N4 /ZnS nanocomposite is still high due to the own flaws of bulk g-C3 N4 . It has been reported that reduce the thickness of g-C3 N4 [1], especially g-C3 N4 monolayer structure has high electron and hole transfer property. Therefore, using the g-C3 N4 monolayer or few layer structures to form the nanocomposite is a way to get the high photocatalytic activity. So we expect to achieve much thinner g-C3 N4 and uniformly ZnS morphology, in order to synthesize the g-C3 N4 /ZnS nanocomposites with high photocatalytic activity. According to reports, few layer or graphene-like g-C3 N4 (GL-C3 N4 ) [20] synthesized by thermal exfoliation from g-C3 N4 , possessing 2D thin-layer structure with 6–9 atomic thickness (2–3 nm), large specific surface area, enhanced photocurrent response and electron transport ability, manifested efficient photocatalytic activity under visible light [21], and could be used as a photo-electrochemical sensor [1]. As we investigated, most of the photocatalysts took rhodamine B (RhB) as the target of pollutants which was easier to be degraded compared with methyl orange (MO) and yet the common colorless antibiotics degradation has not been studied. Tetracycline (TC), as a representative colorless broad-spectrum antibiotic agent, would cause multiple negative influences of the hierarchical system by inducing proliferation of bacterial drug resistance. Therefore, it is very important for the removal of TC and it was chosen as the degradation target to evaluate the photocatalytic activity of the photocatalyst. Meanwhile, the photocatalytic degradation mechanism has not been clearly clarified. In this work, ZnS nanospheres/GL-C3 N4 (ZnS/GL-C3 N4 ) nanocomposites were synthesized by depositing ZnS nanospheres onto the GL-C3 N4 . We demonstrated that GL-C3 N4 provided large surface area for the uniform distribution of ZnS nanospheres leading to the improved adsorption capability ZnS nanospheres with high surface also enhanced the contact areas in nanocomposits. The interaction between ZnS and GL-C3 N4 allowing the effective charge transfer and promoting the photo-generated e− -h+ pairs separation, which would highly make contribution to the improvement of the pollutants degradation. MO and TC were chosen as the targets of pollutants in this system for one was hard to be degraded by C3 N4 and another was a representative broad-spectrum colorless antibiotic agent, respectively. The photocatalytic efficiency of MO degradation over ZnS/GL-C3 N4 was conducted under the illumination of visible light, which showed almost 3.48 and 12.4 times higher than that of pure ZnS and GL-C3 N4 , respectively. 91% TC was photodegraded in the presence of ZnS/GL-C3 N4 (50%) higher than that of GL-C3 N4 . Photoluminescence (PL) spectra clearly showed that the introduction of ZnS decreased the recombination rate of photo-generated e− -h+ pairs, which improved the photocatalytic efficiency of GL-C3 N4 . GL-C3 N4 has lager band gap energy (2.85 eV) and more negative CB level compared with g-C3 N4 . It was favorable for energy band alignment construction with ZnS nanospheres which can serve as the acceptor of the photo-generated electrons coming from GL-C3 N4 . It was also demonstrated that both O2 •− , • OH and holes were the active species in photocatalytic degradation under the
illumination of visible light. Moreover, a possible mechanism was also proposed in detail. 2. Experimental section 2.1. Preparation of photocatalysts The materials were analytical grade (99%), and used as received without any purification. GL-C3 N4 were synthesized according to the reported procedure [1]. ZnS nanosphere was synthesized by using a simple hydrothermal method: 1 mmol ZnAc2 ·2H2 O and 1.5 g PVP were dispersed in 20 mL deionized water by mechanical stirring for 6 h, then TAA (1 mmol) was dropped into the mixture to active a clear solution. Then the clear solution was then transferred to a 20 mL Teflon-lined stainless steel autoclave and kept in an oven at 180 ◦ C for 6 h. At last the product was collected by centrifugation, washed with distilled water and absolute ethanol repeatedly, and then dried to get the sediment. The ZnS/GL-C3 N4 nanocomposite photocatalysts were synthesized by mechanical agitation method. In a typical synthesis, 0.005 g ZnS and 0.095 g GL-C3 N4 were dispersed in 50 mL distilled water by ultra-sonication for 60 min, and then the mixture was mechanical stirring for 48 h. At last, the product was collected by centrifugation, washed with deionized water and ethanol repeatedly, and then dried at 60 ◦ C. The as prepared sample was named as ZnS/GLC3 N4 (5%). Several samples with different weight percent of ZnS (10 wt%, 20 wt%, 40 wt% and 50 wt%) were synthesized by using the similar procedure, and labeled as ZnS/GL-C3 N4 (10%), ZnS/GL-C3 N4 (20%), ZnS/GL-C3 N4 (40%), ZnS/GL-C3 N4 (50%), respectively. 2.2. Characterization of photocatalyst The crystal structure of ZnS/GL-C3 N4 nanocomposites were investigated by power X-ray diffraction (XRD) measurements which were performed on Bruker D8 diffractometer using Cu K␣ radiation ( = 1.5418 À). The chemical state of elements was analyzed by X-ray photoelectron spectroscopy (XPS) on an ESCA Lab MKII X-ray photo-electron spectrometer using Mg K␣ line source. The microstructure and grain morphology were researched by scanning electron microscopy (SEM) and transmission electron microscopy (TEM). SEM was taken with a FESEM MO del JEOL JSM-7001F and TEM was used a JEOL-JEM-2010 (JEOL, Japan) operated at 200 kV. The chemical composition was investigated by X-ray energy dispersion spectrum (EDS) performed at an acceleration voltage of 10 kV. UV–vis diffuse reflection spectra (DRS) were performed on UV-2450 UV–vis system (Shimadzu, Japan) with the reflectance standard material of BaSO4 . Photocurrents and electrochemical impedance spectroscopy (EIS) were recorded on CHI 660B electro-chemical workstation (Chenhua Instrument Company, China). PL experiments were conducted under a 325 nm light excitation . X-band Electron Spin Resonance (ESR) spectra were operated on a JES FA200 spectrometer at ambient temperature. 2.3. Photocatalytic degradation Photocatalytic properties of photocatalysts were estimated by the degradation of MO and TC under the illumination of 300W Xe lamp with 400 nm cut-off filter in custom-made photochemical reactor at 30 ◦ C under constant stirring. In every experiment, 0.0500 g photocatalyst powder was added in 50 mL of MO solution (10 mg L−1 ) or TC (20 mg L−1 ). The reaction mixture was continuously aerated by a pump to provide O2 . Prior to illumination, the mixed liquor was magnetically stirred in dark for 30 min to ensure an adsorption-desorption equilibrium between photocatalyst and pollutant. During every interval, 4.0 mL solution suspension would be taken out to separate the photocatalyst by
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 1. XRD patterns of ZnS/GL-C3 N4 composites.
centrifugation (13000 rpm, 3 min) and researched by UV–vis spectrophotometer at 463 nm and 356 nm which were the maximal absorption peak of MO and TC, respectively. The following formula was used for obtain the photocatalytic degradation efficiency (E) of MO and TC: E = (1 −
C A ) × 100% = (1 − ) × 100% C0 A0
Where C is the concentration of the solution suspension at reaction time t, and A is the corresponding values of absorbancy, C0 is the adsorption/desorption equilibrium concentration of solution suspension, and A0 is the corresponding values of absorbancy. 2.4. Photoelectrochemical measurements To survey the transition of photo-induced electrons, the photocurrents were measured in a standard three-electrode system. Working electrodes of GL-C3 N4 and ZnS/GL-C3 N4 were prepared as follows: 20 L of the dispersion (5 mg mL−1 ) was dropped onto a piece of ITO (0.5 × 1 cm2 ) denoted as GL-C3 N4 /ITO, ZnS/GLC3 N4 /ITO. A platinum wire was employed as counter electrode while a saturated Ag/AgCl electrode was served as reference electrode. A 500 W Xe xenon lamp equipment was adopted as the photosource. The electrolyte solution of photocurrent was phosphate buffered saline (PBS, 0.1 mol L−1 , pH = 7.0). The Nyquist plots was performed in a 0.1 M KCl solution containing 5 mM Fe(CN)6 3− /Fe(CN)6 4− . Sunless conditions were taken at all electrochemical measurements. 3. Results and discussion 3.1. Structure and morphology characterization Fig. 1 showed the XRD patterns of GL-C3 N4 , ZnS and ZnS/GLC3 N4 with different ZnS percentages. From those XRD patterns, it was clearly seen that the patterns of all the synthesized ZnS/GLC3 N4 samples were combined by the diffraction peaks of both ZnS and GL-C3 N4 . While the peak at 27.8◦ and 13.1◦ could be indexed to the (002) and (100) diffraction planes of GL-C3 N4 , which was consistent with the reported literature [1]. For ZnS, the main diffraction peaks at 2 = 28.6◦ , 47.6◦ , 56.5◦ and 76.9◦ corresponding to (111), (220), (311) and (331) plane, respectively [16,22–24]. All the peaks can be indexed to the cubic phase (JCPDS 65-0309). The peaks of ZnS (28.6◦ ) and GL-C3 N4 (27.8◦ ) were very close and overlap with
105
each other. On the other hand, as the increase of ZnS content in the composites, the peaks of GL-C3 N4 weaken while the peaks of ZnS got strong. XPS analyses were adopted to determine the chemical state of every element in the composites. As shown in Fig. 2A, the ZnS/GLC3 N4 (50%) were consisted of C, N, Zn, S, and O elements, while the trace of O species was due to contamination from the atmosphere. High-resolution Zn 2p, S 2p, C 1s, N 1s spectra of the composites was shown in Fig. 2B–E. The Zn 2p peaks of ZnS/GL-C3 N4 (50%) were located at 1021.8 and 1044.6 eV (Fig. 2B), which corresponded to the Zn 2p3/2 and Zn 2p1/2 binding energies, respectively, indicating that the Zn ions existed as Zn2+ according to the previous results [25]. The S 2p peak located at 161.4 and 162.6 eV (Fig. 2C) could be due to S 2p3/2 and 2p1/2 lines, respectively, which indicated the presence of S2− [14]. From Fig. 2D, it could be observed that C 1s for GL-C3 N4 and ZnS/GL-C3 N4 had a same peak at 284.6 eV related to C C or carbon contamination from the air. The C 1s peak at 288.1 and 287.7 eV were related to N C N coordination in GL-C3 N4 and ZnS/GL-C3 N4 , respectively [20,26]. The shift could be ascribed to the interaction between ZnS and GL-C3 N4 , which could influence the chemical environment of sp2 -hybridized carbon. The N 1s peak (Fig. 2E) originating from C N C coordination [26] was located at 398.8 eV for GL-C3 N4 , and it shifted to 398.3 eV for the ZnS/GL-C3 N4 composite, which also proved that there was interaction between ZnS and GL-C3 N4 . This interaction allowed for the charge transfer between ZnS and GL-C3 N4 , promoting the separation of photo-generated e− -h+ pairs, and subsequently increasing the photocatalytic efficiency of the composites. Fig. 3 showed the SEM and TEM images of ZnS, GL-C3 N4 , ZnS/GLC3 N4 (50%) nanocomposites and the EDS of ZnS/GL-C3 N4 (50%) composite. As shown in Fig. 3A and D, ZnS exhibited a uniform spherical shape with a particle size of 0.4–0.5 m in diameter. While GL-C3 N4 was a fewer layered structure from Fig. 3C, which was consistent with previous reports [1]. Fig. 3B and E were the SEM and TEM images of ZnS/GL-C3 N4 (50%), respectively. It exhibited that ZnS nanospheres were equably deposited on the surface of GLC3 N4 and this interaction allowed for the charge transfer between ZnS and GL-C3 N4 , promoting the separation of photo-generated e− -h+ pairs, increasing the photocatalytic efficiency of composites. HRTEM of ZnS/GL-C3 N4 (50%) was shown in Fig. 3F, allowing identification of the crystallographic spacing, and the spacing of d = 0.31, 0.19, 0.16 nm corresponding to the (111), (220), (331) plane of ZnS, respectively. An HRTEM image illustrated the existence of sufficient contact between ZnS and GL-C3 N4 which would be favorable for the transfer of photoexcited carriers. EDS in Fig. 3G showed that the ZnS/GL-C3 N4 (50%) photocatalyst was only composed of C, N, Zn and S, which demonstrated the purity of the composites. Meanwhile, N2 sorption-desorption isotherm (Fig. 4) was employed to determine surface area of the photocatalysts. The surface areas were estimated to be 3.30, 30.10, 46.56, 41.73 and 42.96 m2 g−1 for g-C3 N4 , GL-C3 N4 , ZnS, ZnS/GL-C3 N4 (10%) and ZnS/GL-C3 N4 (50%), respectively. Surface area of GL-C3 N4 was 10 times higher than that of bulk g-C3 N4 and ZnS nanospheres also has big surface area. Therefore, ZnS/GL-C3 N4 nanocomposites had large surface areas which benefited for the charge transfer and the improvement of the photocatalytic performance. 3.2. Optical and electronic properties UV-vis absorption spectra of ZnS, GL-C3 N4 and nanocomposites were shown in Fig. 5A. It was clearly seen that the absorption edge of the ZnS was at about 378 nm, but the basal absorption edge of ZnS/GL-C3 N4 came to visible region at about 460 nm. While in the low-energy visible light region, the absorption intensity of these nanocomposites strengthened in comparison with GL-C3 N4 . The improvement of adsorption capability could increase the use effi-
106
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 2. High-resolution XPS survey spectra of GL-C3 N4 and ZnS/GL-C3 N4 materials: (A) XPS survey spectrum of ZnS/GL-C3 N4 (50%), (B) Zn 2p, (C) S 2p, (D) C 1s, (E) N 1s.
ciency of visible light and consequently led to enhancement of the photocatalytic efficiency. The band-gap energies (Eg ) of ZnS and GL-C3 N4 could be estimated respectively according to the following formula Eq. (Ah)2 = h-Eg
Where h, , Eg and A were Planck constant, light frequency, band gap energy, and absorbance, respectively. From Fig. 5B, it could be seen that the Eg of the GL-C3 N4 was calculated to be 2.85 eV and the Eg of ZnS was calculated to be 3.45 eV. Fig. 6 showed the PL spectra of the GL-C3 N4 and different ZnS/GL-C3 N4 samples. Under a 325 nm light excitation, all the samples showed the broad peak at 450 nm, which was corre-
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
107
Fig. 3. SEM images: (A) ZnS, (B) ZnS/GL-C3 N4 (50%); TEM images: (C) GL-C3 N4 , (D) ZnS, (E) ZnS/GL-C3 N4 (50%); (F) HRTEM and (G) EDS of ZnS/GL-C3 N4 (50%).
sponded to the band gap of DRS [27]. The increasing percentage of ZnS decreased the photoluminescence intensity, indicating that the ZnS/GL-C3 N4 samples had lower recombination rate of photo-
generated e− -h+ pairs than that of GL-C3 N4 , providing an evidence for the effective charge transfer between ZnS nanospheres and GLC3 N4 nanosheets.
108
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 4. Nitrogen absorption-desorption isotherms of (A) g-C3 N4 , (B) GL-C3 N4 , (C) ZnS, and ZnS/GL-C3 N4 nanocomposites: (D) 10%; (E) 50%.
In order to further confirm the important role of nanocomposites in the photocatalytic reaction, five switch source loop transient photocurrent response experiment were performed. In Fig. 7, the ZnS/GL-C3 N4 nanocomposites showed higher photocurrent intensity than that of GL-C3 N4 under the light irradiation. And ZnS/GL-C3 N4 (50%) showed highest photocurrent intensity compared with other ZnS content nanocomposites, which was consistent with the photocatalytic activities. Photocurrent results confirmed that the ZnS/GL-C3 N4 had a higher efficient separation of photo-generated e− -h+ pairs under the illumination of visible
light. Above all, we confirmed that there were effective interactions between the ZnS and GL-C3 N4 , which was favorable for photogenerated electrons excited from VB to CB of GL-C3 N4 transfer to ZnS, playing an important role in the effective separation of e− -h+ pairs of GL-C3 N4 . EIS measurements were conducted to gain deep understanding into the charge transport behavior of ZnS/GL-C3 N4 (Fig. 8). It was proved that the radius of every circle in the Nyquist diagram was based on the charge transfer process at the interface of working electrode/electrolyte, and the small radius were con-
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
109
Fig. 7. Transient photocurrent response for the GL-C3 N4 and ZnS/GL-C3 N4 nanocomposites in PBS (pH = 7.0) aqueous solution under visible light irradiation.
Fig. 5. (A) UV–vis diffuse reflectance spectra of the samples; (B) Estimated band gap of ZnS and GL-C3 N4 . Fig. 8. EIS of the GL-C3 N4 and ZnS/GL-C3 N4 nanocomposites in a 0.1 M KCl solution containing 5.0 mM Fe(CN)6 3− /Fe(CN)6 4− .
the smallest radius demonstrating a high-efficiency separation of photo-generated e− -h+ pairs. Anyway, EIS results were consistent with that of the five switch source loop transient photocurrent response experiment and PL spectra. 3.3. Photocatalytic efficiency of the composites
Fig. 6. PL spectra of the GL-C3 N4 and ZnS/GL-C3 N4 composites.
formed with low resistance [28,29], which was favorable for the charge transfer. It could be seen from Fig. 8 that there was a decline in the radius when the content of ZnS was increased from 5% to 50% of the system, indicating that the introduction of ZnS could decrease charge-transfer resistance and enhance the interfacial charge transfer ability. ZnS/GL-C3 N4 (50%) showed
From the characterizations mentioned above, the nanocomposites showed improved adsorption capability, high separation efficiency. They were supposed to be efficient visible-light driven photocatalysts. Before illumination, the mixed liquor was magnetically stirred for 30 min in dark. In the actual experiment it was found that ZnS, GL-C3 N4 , and ZnS/GL-C3 N4 nanocomposites have negligible absorption toward MO in dark. Therefore, the photocatalytic activity will be enhanced only in the presence of ZnS/GL-C3 N4 catalyst with light irradiation. As shown in Fig. 9A, GL-C3 N4 and ZnS showed weak photocatalytic activity in MO degradation, while ZnS/GL-C3 N4 composites showed higher photocatalytic degradation efficiency than that of GL-C3 N4 and ZnS. From 5% to 50% component of ZnS in ZnS/GL-C3 N4 , the photocatalytic activity gradually increased, and with further increasing the ZnS concentration to 60% and 70%, respectively, the photocatalytic activity decreased. ZnS/GL-C3 N4 (50%) nanocomposite showed the highest photocatalytic activity under visible-light irradiation in 60 min (Fig. 9A). So,
110
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 10. Photocatalytic activities of the samples for the degradation of TC under visible-light irradiation.
photocatalytic activity in MO degradation than that of champion sample of the ZnS/GL-C3 N4 (50%) composites. The control experiment results confirmed that the chemical interaction between the two semiconductors in ZnS/GL-C3 N4 composites, which allowed for the charge transfer between ZnS and GL-C3 N4 , and efficient photo-generated e− -h+ pairs separation, was important for the increasement of photocatalytic efficiency. Furthermore, it was found that the photocatalytic degradation of MO over all the studied photocatalysts followed first-order reaction dynamics under the illumination of visible-light. The first order equation model was shown as follows: −ln(C/C 0 ) = kt
Fig. 9. (A) Photocatalytic activities of the samples for the degradation of MO under visible-light irradiation; (B) Kinetic fit for the degradation of MO with the samples; (C) Cycling runs of the ZnS/GL-C3 N4 (50%) sample.
an appropriate ZnS ratio in nanocomposite was critically significant for the improvement of the resultant photocatalytic performance. The results showed the ZnS import was in favor of GL-C3 N4 photoactivity promotion under visible light and the existence interaction between ZnS and GL-C3 N4 could highly enhance e− -h+ pairs separation efficiency. Form comparision, a physically blending of ZnS and GL-C3 N4 with an appropriate ratio of 1:1 was also used as one of the photocatalysts. As could be seen from Fig. 9A, the 50% physically blending of ZnS and GL-C3 N4 showed the negligible
Where C is the current concentration of MO (mg L−1 ) at reaction time t, C0 is the initial concentration of MO (mg L−1 ), k is the reaction rate constant (min−1 ). The experimental data and the fitting curves achieved according to the equation was shown in Fig. 9B. It could be seen that the reaction rate constant over ZnS/GL-C3 N4 (50%) nanocomposite was 5.4 times higher than that of GL-C3 N4 . Meanwhile, all the investigated nanocomposites showed larger value of k than GL-C3 N4 , demonstrating the fast reaction kinetics for MO degradation, suggestive a synergistic effect of components for MO degradation. The stability of the synthesized ZnS/GL-C3 N4 (50%) nanocomposite was studied by a 5-run cycling test under the same condition. In Fig. 9C, the photocatalytic efficiency of ZnS/GL-C3 N4 (50%) had just a little loss and the degradation efficiency of MO was still near 80% after 5 recycles under the illumination of visible light. The composition stability of ZnS/GL-C3 N4 (50%) was also investigated by XRD analyses. It could be seen from Fig. 1, the recycled ZnS/GL-C3 N4 (50%) nanocomposite named “50% after” had no obvious changes, indicating that the ZnS/GL-C3 N4 (50%) remained effective and recyclable ability for MO degradation. TC can cause multiple negative influences of each level of the hierarchical system by inducing proliferation of bacterial drug resistance as a representative colorless broad-spectrum antibiotic agent. Therefore, it is very important to be removed and it was chosen to evaluate the photocatalytic activity of the ZnS/GL-C3 N4 (50%) via the photocatalytic degradation under visible light irradiation. As shown in Fig. 10, the result revealed 91% TC was photodegraded in the presence of ZnS/GL-C3 N4 (50%) and exhibited 1.35 times higher than that of pure GL-C3 N4 . The results imply that ZnS/GL-C3 N4 is a kind of efficient photocatalysts which could also be applied to antibiotics pollution treatment.
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
111
Fig. 12. Photodegradation of MO over ZnS/GL-C3 N4 (50%) photocatalyst with different quenchers.
Fig. 11. ESR spectra of radical adducts trapped by DMPO: (A) DMPO-O2 • − radical species detected for the ZnS/GL-C3 N4 (50%) dispersion in methanol, (B) DMPO-• OH radical species detected for the ZnS/GL-C3 N4 (50%) dispersion in water.
3.4. Schematic illustration of ZnS/GL-C3 N4 nanocomposites To check the active species produced pending the photocatalysis over these samples under the illumination of visible light, ESR spin-trap technique with DMPO (5,5-dimethyl-1-pirroline-Noxide) was used. Fig. 11 indicated that there was no ESR signal for the ZnS/GL-C3 N4 without visible light irradiation. In Fig. 11A, while visible light irradiation gave signals attributed to a DMPOO2 •− species detected successfully in ZnS/GL-C3 N4 dispersions in methanolic media, indicating that O2 •− reactive species was generated during the reaction under visible light irradiation. There were also • OH signals detected during the reaction (Fig. 11B). The results confirmed that both O2 •− and • OH were main active species which were consist with the trapping experiment result. To further investigate the mechanism, trapping experiments had also been carried out. The 2Na-EDTA, t-BuOH and N2 were added as the scavenger of hole, • OH and O2 •− [26,30]. As shown in Fig. 12, the addition of N2 and t-BuOH had significantly suppressed the degradation of MO over ZnS/GL-C3 N4 (50%) photocatalyst, while 2Na-EDTA had little effect. The results demonstrated that both O2 •− , • OH and holes were active species in this system.
The band gap was considered to under the possible photocatalytic mechanism of MO degradation in the presence of ZnS/GL-C3 N4 samples under visible light. In Fig. 13, ZnS nanospheres were equably distributed on the surface of GL-C3 N4 and the band gaps of ZnS and GL-C3 N4 were 3.45 eV [22] and 2.85 eV [20], respectively. So it was clear to know that GL-C3 N4 could be excited under the visible light due to the appropriate band gap and ZnS was inert owing to its wide band gap under visible-light irradiation. Subsequently, the corresponding photo-induced electrons in GL-C3 N4 could easily transport to ZnS and accumulate in the CB of ZnS. Meanwhile, the photo-generated electrons in the CB of ZnS (−0.91 eV) was more negative than E0 (O2 /O2 •− ) (−0.046 eV vs. NHE) which could reduce O2 to generate O2 •− . While the VB value of GL-C3 N4 (+1.65 eV) was less positive than E0 (• OH/OH− ) (+1.99 eV vs. NHE), which meant no • OH was oxidized from OH− by the hVB + of ZnS/GL-C3 N4 material. For this reason, we confirmed that • OH radical was derived from the part of the O2 •− reactive species and played a role in the degradation of MO [11,31]. However, the photo-generated holes on the VB of GL-C3 N4 could oxidize the MO directly. Corresponds to the trapping experiments and ESR results, the photocatalytic mechanism was proposed and we could confirm that both O2 •− , • OH and holes were the active species that led to effective MO degradation. 4. Conclusion In summary, the ZnS/GL-C3 N4 nanocomposites were synthesized by a simple agitation method. ZnS nanospheres were uniformly deposited on the surface of GL-C3 N4 with large surface area, the formed ZnS/GL-C3 N4 nanocomposites led to the improved adsorption capability owing to their synergetic effect. The enhanced MO and TC degradation efficiency proved that the coupling of ZnS nanospheres and GL-C3 N4 nanosheets was suitable for the energy band alignment, which was favorable for promoting the light response, improving the interface electrons transport, inhibiting the recombination efficiency of e− -h+ pairs, and absorbing more active species. All these synergetic effect made the as-prepared ZnS/GL-C3 N4 nanocomposites possess excellent stability, reusability and high photocatalytic performance with fast reaction kinetics rate compared with pure ZnS and GL-C3 N4 . Meanwhile, a detail photocatalytic reaction process was proposed and it was demonstrated that not only O2 •− and holes, but also • OH were active species played the important role for MO degradation in this system.
112
J. Yan et al. / Molecular Catalysis 438 (2017) 103–112
Fig. 13. Proposed photocatalytic mechanism of MO degradation in the presence of ZnS/GL-C3 N4 samples under visible light.
Acknowledgments The authors genuinely appreciate the financial support of this work by the National Natural Science Foundation of China (21476097 and 21476098), the Natural Science Foundation of Jiangsu Province (BK20131207 and BK20130513). References [1] H. Xu, J. Yan, X.J. She, L. Xu, J.X. Xia, Y.G. Xu, Y.H. Song, L.Y. Huang, H.M. Li, Nanoscale 6 (2014), 9866–9866. [2] S.W. Cao, J.X. Low, J.G. Yu, M. Jaroniec, Adv. Mater. 27 (2015) 2150–2176. [3] J.S. Zhang, B. Wang, X.C. Wang, Prog. Chem. 26 (2014) 19–29. [4] T.T. Li, L.H. Zhao, Y.M. He, J. Cai, M.F. Luo, J.J. Lin, Appl. Catal. B 129 (2013) 255–263. [5] Q.J. Xiang, J.G. Yu, M. Jaroniec, J. Phys. Chem. C 115 (2011) 7355–7363. [6] M. Tahir, C.B. Cao, N. Mahmood, F.K. Butt, A. Mahmood, F. Idrees, S. Hussain, M. Tanveer, Z. Ali, I. Aslam, ACS Appl. Mater. Interfaces 6 (2014) 1258–1265. [7] H. Xu, J. Yan, Y.G. Xu, Y.H. Song, H.M. Li, J.X. Xia, C.J. Huang, H.L. Wan, Appl. Catal. B 129 (2013) 182–193. [8] J.Q. Yan, H. Wu, H. Chen, Y.X. Zhang, F.X. Zhang, S.Z. Frank Liu, Appl. Catal. B 191 (2016) 130–137. [9] Y.M. He, Y. Wang, L.H. Zhang, B.T. Teng, M.H. Fan, Appl. Catal. B 168 (2015) 1–8. [10] W.C. Peng, X.Y. Li, Catal. Commun. 49 (2014) 63–67. [11] B. Yuan, J.X. Wei, T.J. Hu, H.B. Yao, Z.H. Jiang, Z.W. Fang, Z.Y. Chu, Chin. J. Catal. 36 (2015) 1009–1016. [12] Z.W. Zhao, Y.J. Sun, F. Dong, Nanoscale 7 (2015) 15–37. [13] Y. Hou, Z.H. Wen, S.M. Cui, X.R. Guo, J.H. Chen, Adv. Mater. 25 (2013) 6291–6297. [14] D.A. Reddy, R. Ma, M.Y. Choi, T.K. Kim, Appl. Surf. Sci. 324 (2015) 725–735.
[15] E.S. Agorku, H. Mittal, B.B. Mamba, A.C. Pandey, A.K. Mishra, Int. J. Biol. Macromol. 70 (2014) 143–149. [16] F.F. Shi, L.L. Chen, C.S. Xing, D.L. Jiang, D. Li, M. Chen, RSC Adv. 4 (2014) 62223–62229. [17] Y.Q. Shi, S.H. Jiang, K.Q. Zhou, B.B. Wang, B. Wang, Z. Gui, Y. Hu, R.K.K. Yuen, RSC Adv. 4 (2014) 2609–2613. [18] J. Wang, P. Guo, Q.S. Guo, P.G. Jönsson, Z. Zhao, CrystEngComm 16 (2014) 4485–4492. [19] L.Q. Shao, D.L. Jiang, P. Xiao, L.M. Zhu, S.C. Meng, M. Chen, Appl. Catal. B 198 (2016) 200–210. [20] J. Yan, Z.G. Chen, H.Y. Ji, Z. Liu, X. Wang, Y.G. Xu, X.J. She, L.Y. Huang, L. Xu, H. Xu, H.M. Li, Chem. Eur. J. 22 (2016) 4764–4773. [21] X.D. Zhang, X. Xie, H. Wang, J.J. Zhang, B. Pan, Y. Xie, J. Am. Chem. Soc. 135 (2013) 18–21. [22] X.D. Zhang, X.J. Liu, L. Zhang, D.L. Li, S.H. Liu, J. Alloys Compd. 655 (2016) 38–43. [23] X.J. Wang, F.Q. Wan, K. Han, C.X. Chai, K. Jiang, Mater. Charact. 59 (2008) 1765–1770. [24] B. Subash, B. Krishnakumar, M. Swaminathan, M. Shanthi, Ind. Eng. Chem. Res. 53 (2014) 12953–12963. [25] L.X. Hu, F.Y. Chena, P.F. Hub, L.P. Zoua, X. Hua, J. Mol. Catal. A: Chem. 411 (2016) 203–213. [26] H.T. Ren, S.Y. Jia, Y. Wu, S.H. Wu, T.H. Zhang, X. Han, Ind. Eng. Chem. Res. 53 (2014) 17645–17653. [27] S.C. Yan, Z.S. Li, Z.G. Zou, Langmuir 25 (2009) 10397–10401. [28] H. Zhao, Y.M. Dong, P.P. Jiang, H.Y. Miao, G.L. Wang, J.J. Zhang, J. Mater. Chem. A 3 (2015) 7375–7381. [29] I. Shakir, M. Shahid, D.J. Kang, Chem. Eng. J. 225 (2013) 650–655. [30] H. Xu, H.Z. Zhao, Y.G. Xu, Z.G. Chen, L.Y. Huang, Y.P. Li, Y.H. Song, Q. Zhang, H.M. Li, Ceram. Int. 42 (2016) 1392–1398. [31] Q. Li, N. Zhang, Y. Yang, G.Z. Wang, D.H.L. Ng, Langmuir 30 (2014) 8965–8972.