Photocatalysis has been regarded as a promising strategy for environmental remediation because it is a clean process and has high degradation ability as well as low energy consumption [1, 2]. As one of the most important photocatalysts, TiO2 has attracted considerable attention since the first report by Fujishima et al.
. To date, various TiO2 nanostructures with different crystalline phases have been developed and used to photodegrade organic pollutants [4-6]. However, the wide band gap of TiO2 (3.2 eV) makes it difficult for activation by visible light, which leads to inefficient utilization of sunlight , and moreover, the low quantum efficiency in photocatalytic reactions is not satisfactory for practical applications . Therefore, it is highly desirable to overcome these inherent shortcomings of TiO2 to enhance its photocatalytic activity.
Coupling TiO2 with co-catalysts to construct hybrid photocatalysts is an effective approach to achieve this goal. Compared to single-component photocatalysts, hybrid photocatalysts combine individual advantages and show better photocatalytic performance [9, 10]. Furthermore, the presence of heterojunctions between host and guest materials changes the band bending at the interfaces and provides a driving force to separate photogenerated electrons and holes [11, 12]. Therefore, combining these aspects, i.e., morphology, phase structure, and heterojunctions, may offer an ideal way to fabricate stable and efficient photocatalysts.
Transition metal dichalcogenides (TMDs) have been a focus of research for a long time owing to their extraordinary optical and electrical properties [13, 14]. A typical TMD, WS2, has been extensively studied for multiple applications in solar cells, electrocatalysis, photocatalysis, and bioimaging labels [15, 16]. WS2 has a typical graphene-like layered structure and a tunable bandgap [17, 18]. The bandgap of WS2 is about 1.35 eV, which is close to the bandgap value for the optimum utilization of solar radiant energy . More importantly, the conduction-band minimum of WS2 is higher than that of TiO2, indicating that it is a possible co-catalyst candidate to photosensitize TiO2 in visible light. Guo et al.  reported that high photocatalytic activity and low recombination of photogenerated carriers were achieved by the deposition of nanosized WS2 on the surface of mesoporous TiO2. Liu et al.  synthesized nanoscale WS2/TiO2 composites and predicted that they may have potential applications in visible-light photocatalysis. More recently, Qi et al.  prepared WS2/TiO2 heterostructures and confirmed their good photoresponse activity and structural robustness. However, these studies only focused on the nanosize effect and ignored the crystal surface effect of the TiO2 substrate. Moreover, the TiO2 substrates used in these WS2/TiO2 composites were three-dimensional nanospheres or nanoparticles, which led to loose and non-uniform contact between WS2 and TiO2 and reduced the separation efficiency of photogenerated carriers. Hence, it is highly desirable to develop new WS2/TiO2 composites with higher photocatalytic activity and stability.
In this work, we chose 2D TiO2 nanosheets as the substrate to fabricate 2D-2D TiO2 nanosheets/layered WS2 (2D-2D TNS/WS2) heterojunctions via a facile hydrothermal method. The structure and morphology of the TNS/WS2 heterojunctions were systematically characterized. As expected, the layered WS2 were tightly coated on the surface of TiO2 nanosheet with a plate-on-plate stacked structure and connected to each other by W=O chemical bonds. The introduction of layered WS2 improved the light absorption properties of the TNS/WS2 heterojunctions and the heterogeneous interface enhanced the separation efficiency of the photo-induced carriers. As a result, the TNS/WS2 heterojunctions showed higher photocatalytic activity than pure TiO2 nanosheets and layered WS2. The photodegradation mechanism and degradation process were also discussed.
Anatase TiO2 nanosheets with exposed (001) and (101) facets were prepared by a facile hydrothermal process as previously reported . The synthesis process of TNS/WS2 heterojunctions is described below. Typically, 2.5 mmol sodium tungstate (Na2WO4·2H2O) was dissolved in 80 mL deionized water under magnetic stirring and the pH value of the solution was adjusted to 3 by dropping 2 mol/L hydrochloric acid. Then 7.5 mmol L-cysteine was added into the mixture, which was stirred until it became transparent again. Subsequently, a certain amount of TiO2 nanosheets was added to the above solution under vigorous stirring. The solution was transferred to a Teﬂon-lined stainless steel autoclave and reacted at 200 ℃ for 24 h. The black products were collected, washed, and dried in a vacuum oven at 60 ℃ for 12 h. The TNT/WND heterojunctions with different W:Ti molar ratios (0.20, 0.10, 0.07, and 0.05) were prepared by adjusting the amount of TiO2 nanosheets and were named TNS/WS2-0.20, TNS/WS2-0.10, TNS/WS2-0.07, and TNS/WS2-0.05. For comparison, pure WS2 was synthesized under the same conditions without adding TiO2 nanosheets. A TiO2 hollow microsphere/layered layered WS2 (THS/WS2) composite was synthesized under the same conditions by adding TiO2 hollow microspheres instead of TiO2 nanosheets.
X-ray diffraction (XRD) patterns were recorded using a Bruker AXS D8 Focus X-ray diffractometer (Germany) with Cu Kα radiation. X-ray photoelectron spectroscopy (XPS) was conducted using a PerkinElmer PHI 1600 ESCA X-ray photoelectron spectroscope with monochromatic Mg Kα radiation and binding energies normalized to the C 1s peak at 284.6 eV. Scanning electron microscopy (SEM) and energy-dispersive X-ray (EDX) analysis were performed on a Hitachi S-4800 instrument. Transmission electron microscopy (TEM) was performed with a JEM-2100F (JEOL, Japan) instrument. Raman spectra were gathered from a confocal microscope-based Raman spectrometer (Renishaw InVia) in ambient air environment with an excitation laser line of 532 nm. The Brunauer-Emmett-Teller (BET) surface area and Barrett-Joiner-Halenda (BJH) pore-size distribution measurements were obtained using a NOVA-2000 volumetric gas sorption instrument (Quantachrome, USA). UV-Vis diffuse reflectance spectra (DRS) of the samples were obtained using a Shimadzu UV-2550 spectrophotometer using BaSO4 as a reference. Photoluminescence (PL) spectra were obtained with a Fluorolog3 spectrofluorometer (Horiba Jobin Yvon) with an excitation wavelength of 325 nm. High-performance liquid chromatography-tandem mass spectrometry (HPLC-MS/MS) was performed on Bruker Daltonics micrOTOF-QII.
In the photocatalytic test, the light source was a 300-W Xe lamp with a 420 nm cut-off filter (to remove all incoming wavelengths shorter than 420 nm) placed 15 cm above the liquid surface. 20 mg of photocatalyst was dispersed in 100 mL of a rhodamine B (RhB) solution (20 mg/L). Prior to light irradiation, the suspensions were vigorously stirred for 60 min in the dark to establish an adsorption-desorption equilibrium. During the photocatalytic reaction, at given intervals, 5 mL samples were collected to test the concentration of RhB. The RhB concentration was calculated by UV-vis spectroscopy (Perkin Elmer, lambda 35).
Electrochemical tests were carried out on a three-electrode electrochemical workstation (CHI 660E, Chenhua, China) equipped with an FTO electrode deposited with the samples as a photoanode, a Pt wire as counter electrode, and Hg/Hg2Cl2/saturated KCl (SCE) as the reference electrode. In a typical experiment, 0.05 mol/L Na2SO4 (40 mL, pH = 6.8) purged with N2 was used as the electrolyte. For the fabrication of the photoanode, 10 mg of the solid sample was homogeneously mixed with 2 mL of absolute ethanol by using a vortex oscillator. The obtained sample was deposited onto the as-washed FTO glass with a controlled area of 1 cm2 by dip-coating to form a film electrode and then dried in air. Impedance-potential scanning (I-P) was carried out in the potential scanning range of ‒1.5‒1.0 V with an increase potential of 0.05 V and a frequency of 1000 Hz.
Fig. 1 shows the XRD patterns of layered WS2, TiO2 nanosheets, and TNS/WS2 heterojunctions. For pure layered WS2, there are three diffraction peaks at 17.7°, 32.5°, and 56.9°, which are ascribed to the (002), (101), and (110) lattice planes of hexagonal-phase WS2 (JCPDS 084-1398) . The diffraction peaks of TiO2 nanosheets and corresponding heterojunctions matched the standard diffraction patterns of anatase TiO2 (JCPDS: 73-1764) . No diffraction patterns associated with WS2 are observed in the TNS/WS2 heterojunctions, indicating that WS2 coated on the surface of TiO2 nanosheets may contain too few layers and are too thin to be detected by XRD . In addition, the (101) peak experiences a chemical shift from 25.5° to 25.2° after WS2 is incorporated into TiO2, revealing the intimate contact and strong interaction between TiO2 nanosheets and layered WS2 . Furthermore, the photos of photocatalysts show that increasing the WS2 content changes the color of the TiO2 nanosheets from white to black, implying an increased absorption of light.
To further explore the properties of the hybrid nanostructure, phonon spectra of the layered WS2, TiO2 nanosheets, and TNS/WS2-0.10 were measured by Raman scattering (Fig. 2). TiO2 nanosheets were characterized by well-defined peaks at 148 (Eg(1)), 201 (Eg(2)), 398 (B1g(1)), 519 (A1g
+ B1g(2)), and 643 cm−1 (Eg(3)), which corresponded to the anatase structure . For pure WS2, the two dominant Raman scattering peaks can be observed at 377 and 404 cm−1, and are attributed to E2g1 and A1g, respectively . When the layered WS2 was deposited on the surface of TiO2 nanosheets, characteristic peaks of WS2 and TiO2 can be detected in TNS/WS2-0.10. Thus, layered WS2 was successfully introduced into the TiO2 nanosheets. Interestingly, there is an extra peak at 972 cm−1, which may be attributed to the W=O bond of tungsten trioxide . This extra peak suggests that the layered WS2 and TiO2 nanosheets are tightly combined, which leads to high reliability.
The SEM images and corresponding EDX images of layered WS2, TiO2 nanosheets, and TNS/WS2-0.10 are presented in Fig. 3. As shown in Fig. 3(a), the pure WS2 nanosheets are stacked together to form WS2 microspheres, with petal-like features observed on the surface. For TiO2 nanosheets (Fig. 3(c)), the SEM image shows well-dispersed sheets. After the introduction of WS2, the surface of TNS/WS2-0.10 (Fig. 3(e)) becomes rough, owing to the layered WS2 coating on the surface of the TiO2 nanosheets. The corresponding EDX spectrum shows that Ti, O, W, and S simultaneously exist in TNS/WS2-0.10; the atomic ratio of W:Ti is about 1:20, which indicates that the actual content of WS2 in TNS/WS2-0.10 is about 13.4 wt%.
The microstructures of layered WS2, TiO2 nanosheets, and TNS/WS2-0.10 were further investigated by TEM. As shown in Fig. 4(a), WS2 is a well-layered structure and the number of layers ranges from one to six. The TEM image of the TiO2 nanosheets indicates that the product consists of well-defined sheet-shaped structures and that the (101) slab exhibits a clear crystal lattice. To gain a clear understanding of the interface between WS2 and the TiO2 nanosheets, the morphology of the TNS/WS2-0.10 nanostructure was vividly studied by TEM. As shown in Fig. 4(c), the TEM images of TNS/WS2-0.10 present layered and sheet-shaped morphologies, which are attributed to layered WS2 and TiO2 nanosheets, respectively. The top-view (Fig. 4(d)) image reveals that WS2 forms a homogeneous coat on the surface of the TiO2 nanosheets and that the crystal lattice can hardly be distinguished, except for the uncovered part. The side views shown in Fig. 4(e) and (f) clearly show that the layered WS2 is coated closely on the (004) and (101) facets of TiO2 nanosheets and that the number of coating layers is about four. Furthermore, the EDX elemental mapping of TNS/WS2-0.10 (Fig. 4(g)) show that O dovetails well with Ti and S dovetails well with W. The intensity of W (or S) is lower than that of Ti (or O), consistent with the content of WS2, which was in the minority. The uniform color and luster of the S and W mappings imply that the layered WS2 are homogeneously wrapped around the TiO2 nanosheets.
The surface chemical states and components of the as-prepared heterojunctions were recorded by XPS. The full range XPS spectrum shown in Fig. 5(a) reveals that the dominant elements are Ti, O, W, and S in TNS/WS2-0.10. A typical high-resolution XPS spectrum of Ti 2p is shown in Fig. 5(b). Two peaks at 458.8 and 464.6 eV are ascribed to Ti 2p3/2 and Ti 2p1/2, respectively. The separation energy is close to 5.8 eV between Ti 2p1/2 and Ti 2p3/2, indicating a normal state of Ti4+ in TiO2
. Fig. 5(c) shows the XPS spectrum of W 4f, which can be fitted to three peaks. The peaks at 35.6, 37.6, and 40.6 eV are assigned to W 4f7/2, W 4f5/2, and W 5p3/2, respectively. This result is in accordance with previous reports . As for the high-resolution spectrum of S 2p (Fig. 5(d)), the fitted peaks at 161.3 and 163.4 eV are attributed to S 2p3/2 and S 2p1/2, respectively . The XPS analysis further confirms that the layered WS2 was successfully introduced into the TiO2 nanosheets.
The nitrogen adsorption-desorption isotherms and BJH pore-size distribution plots of the TiO2 nanosheets, TNS/WS2-0.10, and WS2 are shown in Fig. 6. The appearance of distinct hysteresis loops at high relative pressure suggests that these materials consist of mesopores. From the textural parameters present in the inset, we can see that the surface areas of TNS/WS2-0.10 sharply increase, which is beneficial for the dye adsorbed on the surface. The high surface areas TNS/WS2-0.10 validate their highly porous structure, which could provide a more flexible response to active sites.
The light-absorption property of the as-prepared heterojunctions was determined by UV-vis diffuse reflectance spectra and is displayed in Fig. 7(a). For comparison, the spectra of pure TiO2 nanosheets and layered WS2 are also plotted. As we can see, pure TiO2 nanosheets exhibit a sharp fundamental absorption band edge at about 380 nm, suggesting that they are excited by UV light. In contrast, the layered WS2 display strong absorption from the ultraviolet to visible-light regions as well as in the near-infrared light region. Notably, when WS2 is introduced into TiO2, the photoabsorption regions of the TNS/WS2 heterojunctions are significantly broadened and the absorption intensity is enhanced.
The band gaps, conduction band (CB), and valence band (VB) of TiO2 and WS2 were estimated by Tauc's plots and Mott-Schottky (M-S) measurements. The band gaps of the photocatalysts can be calculated using the following formula :
Where α is the absorption coefficient, A is a constant, h is Planck's constant, Eg is the band energy, and ν is the frequency of the incident light. The value of n is decided by the property of the semiconductor. Anatase TiO2 is an indirect-band-gap semiconductor and its value of n is 1/2 ; layer-like WS2 is a direct-band-gap semiconductor and its value of n is 2 [36, 37].
The corresponding calculated results for TiO2 nanosheets and layered WS2 are displayed in Fig. 7(b) and (d), where the slopes of the tangents are band gap energies. The band gaps of TiO2 and WS2 are 3.23 and 1.43 eV, respectively. Fig. 7(c) and (e) shows the CB for TiO2 and WS2. The CBs of TiO2 and WS2 are determined to be ‒0.40 and ‒0.73 eV vs. NHE, respectively. Therefore, according to the empirical formula EVB = ECB + Eg, the VBs of TiO2 and WS2 are calculated to be 2.83 and 0.7 eV vs. NHE, respectively. Therefore, type-Ⅱ band alignment is produced at the interface of the TNS/WS2 heterojunctions, which is beneficial for the separation of electron-hole pairs.
The PL spectra of TiO2, WS2, and TNS/WS2-0.10 are shown in Fig. 7(f). Clearly, the PL intensity of TNS/WS2-0.10 is lower than that of either TiO2 or WS2. These results directly confirm that TNS/WS2-0.10 possesses more efficient carrier separation, which is a result of the heterojunctions between the TiO2 nanosheets and layered WS2 suppressing the excited electron-hole pair recombination.
To evaluate the photocatalytic activity, photocatalytic degradation of RhB was carried out under visible-light irradiation. As illustrated in Fig. 8(a), the self-degradation of RhB is negligible without photocatalysts. The TiO2 nanosheets show low adsorption capacity and unsatisfactory activity. In contrast, despite the unsatisfactory photocatalytic activity of WS2, its adsorption capacity is higher than that of TiO2 nanosheets. The excellent absorption capacity of WS2 was due to the TNS/WS2 heterojunctions. In addition, the photocatalytic performance of TNS/WS2 heterojunctions was obviously enhanced, especially TNS/WS2-0.10, which showed the highest photocatalytic activity as almost 100% RhB was decomposed after 90 min irradiation under visible light. It is noteworthy that the TNS/WS2 composites still showed higher photocatalytic activity than pure TiO2 nanosheets under UV irradiation (Fig. S1).
The RhB photodegradation kinetics was fitted by applying a first-order model. The ‒ln(C/Co) plot is shown in Fig. 8(b) and the fitted degradation rates (k) of RhB are illustrated in Fig. 8(c). It is obvious that the TNS/WS2-0.10 has the highest degradation rate constant, which is about 9.8 and 4.6 times higher than that of TiO2 nanosheet and layered WS2, respectively. The notably facilitated photocatalytic performance of TNS/WS2 heterojunctions may be derived from the synergistic effect of the improvement of light-harvesting properties and the effective separation of photogenerated electrons and holes.
In order to confirm this conclusion, TiO2 hollow microsphere/layered layered WS2 (THS/WS2) composites were prepared by the same method for comparison. The TEM images (Fig. S2) indicate that the THS/WS2 composites are core/shell-structured. Exposed (004) and (101) facets of the TiO2 hollow microsphere are also found in THS/WS2-0.10. The surface area of THS/WS2-0.10 (Fig. S3(a)) is 72.43 m2·g−1, which is close to that of TNS/WS2-0.10 (76.10 m2·g−1). Furthermore, the light-harvesting properties of THS/WS2-0.10 (Fig. S3(b)) are greatly improved compared to the pristine TiO2 hollow microsphere. These experimental data showed that the exposed facets, surface areas, or light-harvesting properties did not primarily contribute to improving the photocatalytic activity. The charge separation efficiency (Fig. S4) of TNS/WS2-0.10 is higher that of THS/WS2-0.10. Meanwhile, the photocatalytic activity of THS/WS2-0.10 (Fig. S5(a)) is higher than that of the pristine TiO2 hollow microsphere, but is still lower than that of TNS/WS2-0.10. The reaction rate of TNS/WS2-0.10 (Fig. S5(b)) is about five times that of THS/WS2-0.10. Hence, the 2D-2D heterojunction nanostructure is key to improving the photocatalytic activity of TNS/WS2 heterojunctions.
Fig. 8(d) shows the time-dependent absorption spectra of the RhB solution under visible-light irradiation in the presence of TNS/WS2-0.10. The continuous decrease in the absorption intensity at 553 nm indicates the gradually decreasing concentration of RhB with prolonged irradiation time. Interestingly, the major absorption peak of RhB exhibits a small blue-shift from 553 to 530 nm, corresponding to the stepwise formation of a series of N-de-ethylated intermediates . With increasing irradiation time, the peak located at 530 nm continues to shift and decrease, which indicates that the RhB molecules are further decomposed into smaller molecular fragments and the structure of RhB is also destroyed in the end.
The degradation products of RhB over TNS/WS2-0.10 were further analyzed by HPLC-MS/MS and showed in Fig. 9. The peak at a retention time of 7.5 min is attributed to RhB, and the area of peak decreases with increasing irradiation time. Finally, the peak for RhB almost disappears, demonstrating that RhB is completely decomposed. The results confirm the conclusion of Fig. 8(d). As is well known, the photocatalytic degradation of RhB involves a series of complex chemical reactions, mainly including diethyl, ring-opening, and redox reactions [39, 40]. With increasing reaction time, some new peaks appear in the HPLC chromatograms. The new peaks appear at elution times of 4.4, 2.6, 1.9, and 1.6 min, and were further analyzed by MS/MS. The mass-to-charge ratios (m/z) were 415.2, 387.2, 359.1, and 331.1, respectively (Fig. S6). The difference between two successive m/z values is 28, indicating that the diethyl reaction occurs during photocatalytic degradation. After 90 min of irradiation, the intensity of the peak at 1.1 min is obviously enhanced. The m/z for this peak is 89.5, and the chemical composition may be ethylene diacid. The results show that RhB molecules are further decomposed into smaller molecular fragments by the ring-opening reaction . This conclusion could serve as a reference for the degradation of other organic contaminants over the TNS/WS2 heterojunction photocatalyst.
The renewable stability is an important index for practicality of the photocatalysts. Here, we investigated the photocatalytic cycle stability of TNS/WS2-0.10. As shown in Fig. 10(a), the photocatalytic activity of TNS/WS2-0.10 is effectively maintained, except for a 7% decrease from 96% to 89% after three cycles. To further understand the stability of the samples, the XRD pattern of the sample reused for three cycles was measured. The XRD pattern of TNS/WS2-0.10 in Fig. 10(b) shows no obvious change. This confirms that TNS/WS2-0.10 has relatively good stability.
In order to further determine the main active species involved in the degradation process and possible photodegradation mechanism, trapping experiments of TNS/WS2-0.10 were conducted. For this, K2Cr2O7, ammonium oxalate (AO), tert-butyl alcohol (TBA), and benzoquinone (BQ) were employed as scavengers for the reactive species of e−, h+, ·OH, and O2·−, respectively. A comparative experiment with no quencher was also performed under the same conditions. As presented in Fig. 11, in the presence of different quenchers, the corresponding photocatalytic efficiency is partly restrained. The degradation of RhB is slightly inhibited by the addition of K2Cr2O7 or TBA, implying that ·OH or e− is not the main active species responsible for the decomposition of RhB. However, when AO or BQ was added into the reactive solution, the photocatalytic activity of TNS/WS2-0.10 was remarkably lowered. The degradation efficiency of RhB decreased to 33% in the presence of BQ. Surprisingly, RhB was not entirely degraded after adding AO adding to the reactive solution. Given the above results, holes (h+) and superoxide anion radicals (O2·−) were the main active species in the reaction process, and h+ played a dominant role for RhB degradation under visible-light irradiation.
On the basis of the experimental results discussed above and the band positions of TiO2 and WS2, the plausible photocatalytic mechanism of TNS/WS2 heterojunctions was proposed and is illustrated in Fig 12. From analyses in the above section, we could conclude that the n-n heterojunction is formed and is type-Ⅱ. When the TNS/WS2 heterojunctions are exposed to visible light, WS2 can be excited and yield electron-hole pairs. Because the CB level of WS2 is more negative than that of TiO2, the photoexcited electrons in the CB of WS2 can be injected easily into the CB of TiO2 and reacted with O2 in the aqueous solution to form O2·− and thus degrade RhB. Meanwhile, because the potential of ·OH/H2O (2.32 eV vs NHE) is more positive than that of WS2 (0.70 eV vs NHE), photogenerated holes accumulate in the VB of WS2 and directly react with RhB. The large number of holes accelerate the decomposition of RhB into a non-toxic substance. This is why the hole plays a dominant role in RhB degradation, rather than O2·-. Finally, the photoinduced electrons and holes are effectively separated through the traditional heterojunction structure, so as to solve the high recombination rate of the photogenerated electron-hole of TiO2 and greatly enhance its photocatalytic activity under visible light.
In summary, 2D-2D TiO2 nanosheets/layered WS2 (TNS/WS2) heterojunctions were successfully fabricated via a simple hydrothermal reaction. Physical characterization revealed that the layered WS2 was well-coated on the surface of the TiO2 nanosheet, which resulted in the formation of a perfect heterojunction structure. There were four coating layers, and WS2 was connected with TiO2 by a W=O bond. The as-prepared heterojunction composite exhibited excellent photocatalytic activity and reusability for the degradation of RhB under visible-light irradiation. The degradation products were traced by HPLC-MS/MS and the reaction process was explored. In addition, the typical type-Ⅱ mechanism was confirmed, demonstrating that holes and superoxide anion radicals (O2·−) played specific roles in the photodegradation process. This work could provide insights toward understanding TiO2/WS2-based composites.