Visible Light-Driven Photoelectrocatalytic Water Splitting Using Z-Scheme Ag-Decorated MoS2/RGO/NiWO4 Heterostructure.

Herein, we have successfully constructed a solid-state Z-scheme photosystem with enhanced light absorption capacity by combining the optoelectrical properties of AgNPs with those of the MoS2/RGO/NiWO4 (Ag-MRGON) heterostructure. The Ag-MRGON Z-scheme system demonstrates improved photo-electrochemical (PEC) water-splitting performance in terms of applied bias photon-to-current conversion efficiency (ABPE), which is 0.52%, and 17.3- and 4.3-times better than those of pristine MoS2 and MoS2/NiWO4 photoanodes, respectively. The application of AgNPs as an optical property enhancer and RGO as an electron mediator improved the photocurrent density of Ag-MRGON to 3.5 mA/cm2 and suppressed the charge recombination to attain the photostability of ∼2 h. Moreover, the photocurrent onset potential of the Ag-MRGON heterojunction (i.e., 0.61 VRHE) is cathodically shifted compared to those of NiWO4 (0.83 VRHE), MoS2 (0.80 VRHE), and MoS2/NiWO4 heterojunction (0.73 VRHE). The improved PEC water-splitting performance in terms of ABPE, photocurrent density, and onset potential is attributed to the facilitated charge transfer through the RGO mediator, the plasmonic effect of AgNPs, and the proper energy band alignments with the thermodynamic potentials of hydrogen and oxygen evolution.

Introduction

Green fuel generation over semiconductors via photo-electrochemical (PEC) water splitting is always among the proposed solutions for the global energy crisis and ensured environmental sustainability.[1,2] Since the first observation of PEC water splitting on a single-crystal rutile TiO2 photoanode under ultraviolet (UV) light irradiation in 1972,[3] tremendous efforts have been made for proper utilization of solar energy for H2 production,[4−6] water purification through degradation of organic pollutants,[7] and conversion of carbon dioxide into high-value products.[8,9] For effective PEC H2 production, the photoelectrode should possess band gap energy greater than the thermodynamic energy required to split water into its components (i.e., 1.23 eV). Furthermore, the conduction band (CB) edge of the photoelectrode has to lie at higher negative potential compared to the H+/H2 potential while the valence band (VB) edge must lie at higher positive potential compared to H2O/O2 potential. In general, the photoinduced processes in an n-type photoelectrode as an example involves three main steps: (i) light absorption and subsequent generation of electron and hole pairs in the CB and VB, respectively, (ii) the photogenerated electron and hole either recombine and dissipate as heat or separate. The holes generated within the space charge layer or nearby (within the hole diffusion length) migrate to the surface, while the electrons move to the back contact, and (iii) the electrons and holes participate in the redox reactions required for hydrogen and oxygen evolution.[10] The large band gaps of conventional materials, that is, TiO2, ZnO, etc., limit their absorption to the ultraviolet portion of the solar spectrum (i.e., approx. 4%), hence significantly restricting the practical utilization of the solar energy. Furthermore, the fast recombination rate of the photogenerated electron–hole pairs in the single-component photoanode reduces its solar-to-hydrogen conversion efficiency.[11] Therefore, the fabrication of a highly efficient and robust visible-light photoanode is more desirable in order to harvest a wide portion of the solar spectrum.

Lately, layered transition-metal dichalcogenide (TMDC) materials have been extensively investigated for PEC water splitting because of their physical and chemical properties desired for such an application.[12] These materials possess narrow band gaps, enabling them to capture a significant portion of the solar spectrum. In addition, they have a large surface area and high surface atoms and ultrathin layers (few or monolayers), which accelerate the diffusion of photogenerated holes to reach the reaction active sites.[13] These merits are anticipated to play a crucial role in reducing the electron/hole recombination, maximizing the light absorption capacity, and promoting the water-splitting process. MoS2 nanosheets (NSs), a member of the TMDC family, have been explored because of their tunable band gap, high specific surface, relatively high charge-carrier mobility, and dense catalytic active sites.[14,15] However, the PEC properties of layered MoS2 are limited by the ultrafast recombination of the photogenerated charge carriers and the photocorrosion driven by photogenerated holes.[13,16] This challenge can be overcome by heterojunctioning the MoS2 with a suitable semiconductor to facilitate the separation of the photogenerated electron and hole pairs, which promotes the charge transfer and thus leading to consistent PEC performance. MoS2-based conventional heterojunction-photoactive materials such as MoS2/CdS,[17] MoS2/ZnO,[18] In2Se3/MoS2,[19] and MoS2/ZnIn2S4[20] have demonstrated a reasonable PEC performance. However, the less-negative potential of the CB electrons and the less-positive potential of the valance band (VB) holes weaken the redox ability of the photogenerated e–/h+ pairs in the conventional heterojunction. Consequently, the conventional heterojunction systems cannot provide an efficient simultaneous charge separation and redox ability.[21] Hence, conventional heterojunctions should be replaced/modified with more convenient alignment, which could overcome the stated issues.

In recent years, the concept of the Z-scheme heterojunction photosystem was proposed to resolve the aforementioned issues. The structure of the Z-scheme photosystem is similar to the conventional type-II heterojunction system, but the mechanism of the charge-carrier transfer is totally different. In the Z-scheme system, the pathway of the charge-carrier transfer resembles the letter Z because the photoinduced electrons on the low-negative CB semiconductor, having low reduction capability, recombine with the photoinduced holes on the low-positive VB semiconductor having low oxidation capability. This maintains a higher redox capability of the photoinduced charges.[22−25] Thus, compared to the conventional heterojunction photosystem, the separated electron/hole pairs in the Z-scheme heterojunction photosystem exhibit powerful redox capabilities. Literature survey showed utilization of MoS2-based Z-scheme photosystems in PEC water-splitting technology,[26−28] and all of them showed enhanced performance in terms of photocurrent density, stability, and efficiency.

Apart from MoS2, NiWO4 has been investigated in different photocatalytic heterojunction systems for water splitting because of its excellent photocatalytic properties, including a narrow band gap of 2.5 eV. Because of the good alignment of the band edge positions of NiWO4 (CB = +0.25 eV and VB = +2.78 eV vs NHE)[29] with those of MoS2 (CB = −0.33 eV and VB = +1.77 eV vs NHE),[30] a Z-scheme photosystem could be constructed by heterojunctioning MoS2 with NiWO4 in the presence of a conductor. Thus, the rational design of the Z-scheme heterojunction photosystem is expected to enrich the CB of MoS2 with electrons and the VB of NiWO4 with holes provided that good contact between the two materials is established. Based on this system, efficient charge-carrier separation and strong redox capability could be achieved. Recently, reduced graphene oxide (RGO) nanosheets have been utilized as a solid mediator to shuttle electrons in the Z-scheme heterojunction photosystem.[31,32] The high conductivity and the large surface area of RGO nanosheets facilitate the charge-carrier transfer in such a multicomponent system.

In the present study, we are reporting a rationally designed Z-scheme Ag NPs-decorated MoS2/RGO/NiWO4 (Ag-MRGON) heterojunction photoanodes. These photoanodes are synthesized by the controlled hydrothermal method. The system successfully demonstrates PEC water splitting under visible light irradiation, with significant stability and efficiency. The highly improved PEC outcomes could be ascribed to the successful fabrication of the Z-scheme photosystem. The existence of RGO NSs and Ag NPs in the current Z-scheme heterojunction photosystem was essential, and they deemed to be the key factor for enhancing the photocatalytic efficiency. The present study provides a suggestive model to design an RGO-based Z-scheme heterojunction photosystem for efficient visible-light-driven water splitting.

Results and Discussion

Structural and Morphological Characterizations

The XRD patterns of the fabricated photoanodes are provided in Figure . All XRD peaks of MoS2 NSs confirm the hexagonal phase of pure MoS2. The respective patterns positioned at 14.6, 29.3, 39.8, and 58.3° are matched with (002), (004) (103), and (110) crystallographic planes (JCPDS card no. 77-1716).[33,34] For NiWO4 NPs, the diffraction peaks positioned at 15.8, 19.3, 23.9, 30.8, 36.3, 52.2, and 54.5°, which correspond, respectively, to the (010), (100), (011), (111), (002), (130), and (202) crystallographic planes of pure NiWO4, have been observed (JCPDS card 15-0755).[35] For GO nanosheets (Figure S1), the observed intense diffraction peak (2θ = 11.8°, d = 0.75 nm) corresponds to the GO (001) crystal plane. For the MoS2/NiWO4 heterojunction, all diffraction peaks can be assigned either to MoS2 or NiWO4 evincing the purity of the composite. For the Ag-MRGON heterojunction, in comparison with the XRD pattern of the MoS2/NiWO4 heterojunction, two weak diffraction peaks at 25.8 and 31.7° have appeared. These two peaks are indexed to the (002) and (100) crystal planes of RGO and Ag NPs, respectively, which demonstrate the existence of RGO and Ag NPs in this sample. It is also worth mentioning that GO was used as a promoter for the Ag (I) reduction to metallic Ag NPs through the hydrothermal treatment.[36] The appearance of a weak and broad diffraction peak for the RGO (002) plane at 25.8° with a decrease in the interlayer spacing to 0.34 nm and the disappearance of the intense diffraction peak of the GO (001) plane at 11.8° in the pattern of the Ag-MRGON heterojunction verify the reduction of GO to RGO (see Figure ). Moreover, the weak and broad nature of the (002) peak indicates the poor ordering of the NSs along the stacking direction, imposing the formation of single or few layers of RGO NSs.[37] It is also worth mentioning that the MoS2 (002) peak at 15.8° in the patterns of MoS2/NiWO4 and Ag-MRGON heterojunctions is much weaker than that in the pure MoS2 NSs. This might be because of the NiWO4 NPs that suppress the restacking of the MoS2 NSs, leading to the formation of single- or few-layer nanostructures.[33]

Figure 1

XRD patterns of MoS2 NSs, NiWO4 NPs, MoS2/NiWO4, and Ag-MRGON.

Figure a shows the morphological structure of GO in the form of several thin layers, arranged in characteristic stacking. The appearance of wrinkled and folded layers indicates the admission of hydroxyl and epoxy (oxygen-containing functional groups) during the oxidation process.[38] For pristine MoS2 (Figure b), it can be observed that the sample presents a nanosheet-like microstructure. Compared to bulk counterparts, this obtained thin 2D nanosheet structure is favorable for photocatalytic activity because it enhances the light absorption and shortens the charge diffusion distance from the interior of the photoactive material to the surface. Furthermore, the 2D nanosheet structure can increase the electron accessibility to the electrode substrate and hence lowering the interfacial charge-transfer resistance at the back contact.[39] The field emission scanning electron microscopy (FESEM) image (Figure c) of the pristine NiWO4 shows the aggregation of nanoparticles distributed on plate-like microstructures. Interestingly, the developed morphological structure of the MoS2/NiWO4 heterojunction (Figure S2) shows assembled nanosheets and nanoparticles, preserving the morphological structures of the parent components. Also, for the Ag-MRGON heterojunction (Figure d), it is easy to recognize that the composite inherited the structures of all pristine components. Moreover, it can be observed that Ag NPs are well-distributed on the composite surface. On the other hand, in consistence with XRD results, the incorporation of NiWO4 into MoS2 in both heterojunction samples (Figures S2 and 2d) prevented MoS2 NSs from aggregating or stacking with each other and thus preserving the layered structure. Thus, FESEM analysis in line with XRD confirms the successful synthesis of MoS2/NiWO4 and Ag-MRGON heterojunction. As previously reported, because of the SPR effect, decoration with noble metal nanoparticles enhances the visible light absorption at the vicinity of the metal–semiconductor interface leading to enhanced photoactivity. Thus, the combination of NSs and NPs would facilitate perfect interface separation centers and combine the excellent photocatalytic performances of both NSs and NPs simultaneously. Indeed, as expected, the developed Z-scheme Ag-MRGON heterostructure exhibits highly enhanced PEC activity (see section ).[40,41]

Figure 2

FESEM images of GO NSs (a), MoS2 NSs (b), NiWO4NPs (c), and Ag-MRGON (d). EDX spectrum of Ag-MRGON (e).

The existence of constituent elements in the Z-scheme Ag-MRGON heterojunction was affirmed by energy-dispersive X-ray spectrometry (EDX), as presented in Figure e. The EDX patterns reveal the peaks of Mo, S, W, Ni, C, O, and Ag elements at their corresponding keV values. No other peaks have been observed, which indicates the purity of the prepared sample. Moreover, EDX elemental mapping of Ag-MRGON has been performed (Figure S3b–h) to evaluate the elemental distribution through the heterojunction. It can be seen that the constituent elements of Ag-MRGON (Mo, S, Ni, W, C, O, and Ag) are almost uniformly distributed at the micron levels.

Surface Elemental Analysis

XPS analysis was carried out to investigate further the elemental composition and the chemical state of the prepared samples, as displayed in Figures and S4. The XPS survey of the Ag-MRGON sample confirmed the presence of all expected constituents (i.e., Mo, S, Ni, W, O, and Ag) at the surface of the heterojunction, as presented in Figure a. The high-resolution scanning XPS spectra of Mo 3d, S 2p, Ni 2p, W 4f, Ag 3d, C 1s, and O 1s are shown in Figure b–h. For the C 1s XPS spectrum (Figure b), the three peaks positioned at binding energies (BEs) of ∼284.8, ∼286.6, and ∼288.2 eV are attributed to C–C (non-oxygenated sp2 carbon), C–O (hydroxyl or epoxide), and C=O (carbonyl functional groups) with atomic percentages of 79.5, 16.6, and 3.9%, respectively. However, the atomic percentage of these peaks in the GO sample (Figure S4a) was 36.1% (C–C), 55.8% (C–O), and 8.1% (C=O), respectively. Thus, the XPS analysis indicates that the total oxygen content of 63.9% (C–O: 55.8% and C=O: 8.1%) in GO was reduced to 20.5% (C–O: 16.6% and C=O: 3.9%) in the Ag-MRGON sample, while the corresponding nonoxygenated sp2 carbon content of 36.1% increased to 79.5%, imposing the successful reduction of GO in the Ag-MRGON during the hydrothermal process along with the reduction of Ag (I) to metallic Ag NPs.[36] The Ni 2p high-resolution spectrum (Figure c) splits into two spin–orbit doublets at 856.3 (Ni 2p3/2) and 873.8 eV (Ni 2p1/2), along with their corresponding shake-up satellites (Sat.). The spin–orbit separation is 17.5 eV that is an agreement with the reported values for NiWO4 NPs.[42]Figure c also shows the deconvolution of the Ni 2p3/2 peak into two core-level components at 856.1 and 859.8 eV. The Ni 2p1/2 peak was also split into two core-level signals at 873.6 and 8.75.9 eV, indicating Ni2+ and Ni3+ oxidation states in NiWO4, respectively.[43] As for the W 4f spectrum, Figure d displays two major peaks at 36.5 and 38.6 eV with a spin–orbit separation of 2.1 eV. These values are, respectively, assigned to W 4f7/2 and W 4f5/2. These two peaks are attributed to the W6+ oxidation state in tungstate (WO4), confirming the synthesis of NiWO4 NPs.[44,45] Intriguingly, the Ni 2p and W 4f signals exhibit a positive shift in the range of 0.1–1.1 eV toward higher BEs in the Ag-MRGON comparing to those of the pristine NiWO4 and MoS2/NiWO4 heterojunction (Figure S4b,c). More intriguingly, compared to NiWO4, the positive shift in the BEs of Ni 2p and W 4f in the Ag-MRGON heterojunction is greater than those in the MoS2/NiWO4 heterojunction. These results suggest the decrease in the electron density of NiWO4 in the MoS2/NiWO4 and Ag-MRGON heterojunctions because of the electron transfer from NiWO4 to MoS2. The high positively shift in BEs of Ni 2p and W 4f of 1.1 eV, and thus the more decreased electron density of NiWO4 in the Ag-MRGON implies the important role of RGO as an electron mediator between NiWO4 and MoS2. Similarly, the O 1s XPS spectrum (Figure e) was also deconvoluted into three peaks designated as OI, OII, and OIII centered at 531.3, 533.1, and 533.5 eV, respectively. The peaks OI and OII are assigned to the lattice O-bonds with W and Ni in NiWO4 NPs and the adsorbed water on the sample surface, respectively. However, the third peak OIII is attributed to the C–O bond in RGO.[46,47] These peaks reveal the chemical interaction between NiWO4 and RGO in the Ag-MRGON heterojunction. On the other hand, the Mo 3d high-resolution spectrum presented in Figure f was deconvoluted into two major peaks positioned at 229.2 (Mo 3d5/2) and 232.2 eV (Mo 3d33/2), respectively. These peaks signify the existence of the Mo4+ state in MoS2. The small peak at 225.1 eV is arising from S(2s) in MoS2. For MoS2/NiWO4 heterojunction and pristine MoS2 (Figure S4d), the Mo 3d5/2 peak positioned at 229.6 and 229.8 eV while the Mo 3d3/2 peak is positioned at 232.6 and 232.8 eV. The S 2p spectrum of Ag-MRGON (Figure g) was also decomposed into two peaks at BEs of 162.1 and 163.3 eV, assigned to S 2p3/2 and S 2p1/2 of the MoS2 phase, respectively. Remarkably, in opposite to Ni 2p and W 4f, a negative shift of about −0.2 and −0.6 eV is seen in Mo 3d and S 2p of the Ag-MRGON heterojunction compared with the MoS2/NiWO4 heterojunction and pristine MoS2 (Figure S4d,e), respectively, suggesting the increase in the electron density of MoS2 and successful heterojunction formation.[31] The BE values of Mo 3d and S 2p are in close agreement with the previously published values.[48,49] Similar to Mo 3d and S 2p, the Ag 3d XPS spectrum (Figure h) was fitted into two doublet peaks positioned at 366.8 and 372.8 eV, which can be, respectively, assigned to Ag 3d5/2 and Ag 3d3/2 of metallic Ag0. These BE values are positively shifted by 0.6 eV than the literature values of metallic silver NPs[50] indicating the electron transfer from Ag NPs to MoS2 NS and NiWO4 NPs. Thus, the binding energy shifts provide clear evidence of the expected charge-carrier transfer pathway across the interface of MoS2/NiWO4 and Ag-MRGON heterojunctions. In particular, the photogenerated electrons migrate from NiWO4 to MoS2 in the MoS2/NiWO4 heterojunction, following the direct Z-scheme mechanism as well as to RGO/MoS2 in the Ag-MRGON heterojunction. The XPS results confirm the successful synthesis of heterojunctions, and it could be anticipated that there is intimate interfacial and electronic interaction between MoS2 and RGO/NiWO4 along with Ag NPs, which might facilitate the separation and transfer of the photogenerated charge carriers in the solar-induced applications.

Figure 3

(a) XPS survey spectrum of the Ag-decorated MoS2/RGO/NiWO4 heterojunction photoanode (Ag-MRGON). Deconvoluted spectra of (b) C 1s, (c) Ni 2p, (d) W 4f, (e) O 1s, (f) Mo 3d, (g) S 2p, and Ag 3d (h).

Optical Analysis

The UV–vis DRS analysis was performed to explore the optical characteristics of the as-synthesized photoanode materials. As depicted in Figure a, both pristine MoS2 NSs and NiWO4 NPs exhibit a broad absorption band from UV to the visible region of the electromagnetic spectrum. However, coupling MoS2 and NiWO4 to form the MoS2/NiWO4 heterojunction led to enhancing the light absorption density in both UV and visible regions. Interestingly, the Ag-MRGON sample shows a strong absorption compared to all other samples, indicating that such a heterojunction could efficiently utilize a more significant portion of visible light and enhance the generation of electron–hole pairs under visible light irradiation. The increased absorption capacity of Ag-MRGON might be ascribed to the well-established SPR absorption of metallic Ag NPs.[41,51] However, no clear plasmonic peak appears in the spectrum because of the possible aggregation of the Ag NPs. On the other hand, it is worth emphasizing that RGO does not contribute to the light absorption because of the absence of the absorption edge of RGO in the UV–vis DRS spectrum of Ag-MRGON. Thus, it can be stated that RGO functions as a mediator for shuttling electrons between NiWO4 and MoS2 in the established Z-scheme photosystem. The band gaps of the photoanodes were estimated using the Tauc’s relation and Kubelka Munk function, that is, eqs and 2:[52]where A, hν, Eg, and (F(R)) show constant, incident photon energy, band gap energy, and function of reflectance (corrected absorbance), respectively. R is the reflectance of the sample, and n describes the process of optical absorption. Theoretically, n = 0.5 and 2 for indirect and direct transitions between VBs and CBs, respectively. Assuming the indirect transition between VBs and CBs, the indirect band gaps of the synthesized materials are evaluated by plotting the (F(R)hν)0.5 against hν (Figure b). The band gap values of MoS2 NSs, NiWO4 NPs, MoS2/NiWO4, and Ag-MRGON were found to be 2.18, 2.39, 2.34, and 2.25 eV, respectively. The band gap energies of the MoS2/NiWO4 and Ag-MRGON heterojunctions lie between those for parent components. This might be due to the strong synergistic interaction between MoS2 NSs and NiWO4 NPs in the heterojunctions. Ag NPs significantly improve the absorption of photons and thus utilize more solar energy to strengthen the solar energy harnessing efficiency.

Figure 4

UV–vis diffuse reflectance spectra (a) and Tauc’s plots (b) of MoS2 NSs, NiWO4 NPs, MoS2/NiWO4, and Ag-MRGON.

PEC Water-Splitting Performance

Photocurrent Density–Potential (J–V) and Transient Photocurrent Density–Time (J–T) Characteristics

The PEC water-splitting performance of the prepared MoS2 NSs, NiWO4 NPs, MoS2/NiWO4, and Ag-MRGON photoanodes was assessed by measuring the photocurrent densities, photostability, and % ABPE at given potentials against RHE. The photocurrent density–voltage (J–V) plots are provided in Figure a for all photoanodes. In the dark, insignificant current in the range between 10–8 and 10–7 A/cm2 has been observed for all electrodes evincing that no electrocatalytic oxygen evolution reactions occur and the electrodes are stable in this potential window. Under air mass (AM) 1.5G-simulated sunlight at 1.23 VRHE, NiWO4 NPs and MoS2 NSs exhibit a photocurrent density of about ∼0.13 and 0.44 mA/cm2, respectively. The observed lower photocurrent densities can be attributed to the high rate of electron–hole pair recombination in such a single photosystem and low to the capacity of light absorption. Excitingly, the MoS2/NiWO4 heterojunction photoanode shows an increased photocurrent density of 1.1 mA/cm2 at 1.23 VRHE, which is almost 8.5 and 2.5 times higher than pristine NiWO4 NPs and MoS2 NSs, respectively. More intriguingly, the photocurrent density of the Ag-MRGON heterojunction photoanode is significantly enhanced to 3.5 mA/cm2 at 1.23 VRHE, (27-times, 8-times, and 3.3-times of NiWO4, MoS2, and MoS2/NiWO4 photoanodes). These results indicate the feasibility of rationally designed photoanode system in the current study. The Z-scheme Ag-decorated heterojunction demonstrated a substantial increase in photocurrent density with significant cathodic onset potential (Von) shift for water oxidation. As can be seen in the J–V curves (Figure a), the Z-scheme Ag NP-decorated heterojunction photoanode shows a significant effect on the Von (the voltage at which dJ/dV ≥ 0.2 mA/cm2 × V[53]). Figure S5 displays the Von determination of all photoanodes, and a summary of these values is reported in Table . It is found that the Von of the pristine NiWO4 NPs and MoS2 NSs photoanodes is 0.83 and 0.80 VRHE. The MoS2/NiWO4 heterojunction photoanode exhibits the Von of 0.73 VRHE, which is ∼100 mV and 70 mV less than those for pristine NiWO4 NPs and MoS2 NSs, respectively, indicating the suppressed recombination of the photogenerated charge carriers. Moreover, the Ag-MRGON heterojunction photoanode exhibits the lowest Von of 0.61 VRHE. This value is cathodically shifted by ∼220, 190, and 120 mV from that of NiWO4, MoS2, and MoS2/NiWO4 photoanodes, respectively. The significantly enhanced photoresponsive current density and the observed cathodic shift of the onset potential in the Ag-MRGON photoanode (Table ) imply the superior charge-carrier separation and transport abilities in the Z-scheme heterojunction.[54]

Figure 5

(a) Photocurrent densities as a function of the applied potential, (b) calculated photoconversion efficiencies (% ABPE) vs potential of MoS2 NSs, NiWO4 NPs, MoS2/NiWO4, and Ag-MRGON heterojunction photoanodes, (c) transient photocurrent densities of MoS2 NSs, MoS2/NiWO4, and Ag-MRGON photoanodes, and (d) transient photocurrent density of NiWO4 NPs measured at 1.0 VRHE under chopped simulated sunlight illumination.

Table 1

Photocurrent Densities (1.23 VRHE) and Onset Potential (Von) for Pristine NiWO4 NPs, MoS2 NSs, MoS2/NiWO4 Heterojunction, and Z-Scheme Ag-Decorated MoS2/RGO/NiWO4 Heterojunction Photoanodes

	NiWO4	MoS2	MoS2/NiWO4	Ag-decorated MoS2/RGO/NiWO4
J (mA/cm2)	0.13	0.44	1.1	3.5
Von (V)	0.83	0.80	0.73	0.61

The applied bias photon-to-current efficiency (% ABPE) of all electrodes was computed from the current–potential curves presented in Figure b using eq .[55]where Jph, Vapp, and Plight represent the photocurrent density (mA cm–2), applied potential, and illumination power density (100 mW cm–2), respectively. Figure b shows % ABPE for all photoanodes as a function of the applied bias potentials. The Ag-MRGON heterojunction photoanode achieves the highest efficiency of 0.52%, which is very significant compared to NiWO4 NPs (8.02 × 10–5%), MoS2 NSs (0.03%), and MoS2/NiWO4 (0.12%) photoanodes. To further confirm a sustainable transient photocurrent performance of the prepared photoanodes, the transient photocurrent density–time (J–t) measurements were performed at 1.0 VRHE by keeping the light ON/OFF cycles at 40 s ON and 40 s Off. As exhibited in Figure c,d, the photocurrent density jumped to the highest possible value under illumination and immediately dropped after the light is turned off, indicating that the fabricated photoanodes have the fast photoelectric response ability. The Ag-MRGON photoanode shows the highest photocurrent density (about 2.9 mA/cm2) compared with NiWO4, MoS2, and MoS2/NiWO4 heterojunction photoanodes. This distinctly enhanced photocurrent density indicates that the Z-scheme Ag-MRGON heterojunction possesses the higher interfacial charge-transfer efficiency after the introduction of RGO as an electron-transfer mediator in the MoS2/NiWO4 heterojunction along with surface decoration with Ag NPs. High interfacial charge-transfer efficiency could be attributed to the significant interaction among the constituents of the heterojunction as supported by the XPS results.

PEC Stability Behavior

Figure shows the steady-state photocurrent densities of MoS2 NSs, MoS2/NiWO4 heterojunction, and Ag-MRGON photoanodes acquired at an applied potential of 1.0 VRHE under simulated light. The observed photocurrent density values were almost at a similar level of J–V curves at the same applied potential, hence validating the steady-state behavior of heterojunctions. For the MoS2 photoanode, the photocurrent declined slightly in the first 20 min, indicating that the photocorrosion of the MoS2 film existed all the time under continuous light illumination. The photogenerated holes might either oxidize water to OH– radicals, which in turn interact chemically with sulfur atoms at the surface, hence oxidizing them to sulphates or directly oxidizing sulfide ions into metallic sulfur.[56] Therefore, the photogenerated holes within the MoS2 should be utilized very fast to eliminate the photocorrosion and enhance the photostability of MoS2. Recent studies reported that the undesirable photocorrosion of a semiconductor could be suppressed by constructing a direct Z-scheme heterojunction.[57,58] Indeed, the fabrication of the MoS2/NiWO4 heterojunction photoanode shows a stable photocurrent for 45 min (i.e., much better than pristine MoS2) followed by sharp photocurrent drop. The photocurrent drop could be attributed to photocorrosion and mechanical instability of the film. This improvement in the photostability can be mainly attributed to the formation of direct Z-scheme heterojunction between MoS2 and NiWO4 (Figure S6b). The photoexcited electrons in the CB of NiWO4 react fast with the photogenerated holes in the VB of MoS2, thus preventing the sulfide ion oxidation and suppressing the photocorrosion of MoS2 surface. As a matter of course, these results indicate the vital role of the direct Z-scheme heterojunction in improving the photostability and activity of MoS2/NiWO4. Nevertheless, enhancing the charge-transfer rate at the heterojunction interface is still required for enhancing the overall PEC performance.[56,59] Notably, Ag-MRGON photoanodes demonstrated the highest photocurrent with excellent stability for a more extended period (shown in Figure ) because of the significant suppression of the photocorrosion, efficient photogenerated charge-carrier separation, and transport demonstrating the crucial role of RGO as an electron mediator and charge transporter at the interface between MoS2 and NiWO4. In a recent report by N. Lu et al.,[60] it was shown that RGO as an electron mediator between g-C3-N4 and WO3 plays a vital role in enhancing the photocurrent density and stability of the Z-scheme g-C3-N4/RGO/WO3 heterojunction system compared with that without RGO. Moreover, the light absorption attributes of the photosystem were enhanced by the well-known SPR effect of Ag NPs[41] (more details will be discussed in the PEC water-splitting mechanism section).

Figure 6

Steady-state photocurrent densities as a function of time for MoS2 NPs, MoS2/NiWO4 heterojunction, and Ag-MRGON heterojunction photoanodes measured at 1.0 VRHE.

To investigate the long-term structural and chemical stabilities of the Ag-MRGON photoanodes, XRD and XPS spectra were recorded before and after photostability tests. The XRD spectra after photostability tests (Figure S7a) exhibited unchanged XRD patterns, indicating the stability of the photoanodes. Furthermore, XPS analysis (Figure S7b–d) showed that there are no remarkable changes in the BEs of Ni 2p, Mo 3d, and Ag 3d before and after photostability tests. However, the area of the OIII component in the resolved O 1s core level after photostability tests (Figure S7e) showed little increase, which may be due to the adsorbed water on the surface of the used photoanode. Thus, the results confirm the long-term stability of the photoanodes during PEC water-splitting reactions.

Electrochemical Impedance Spectrometry Analysis

The electrochemical impedance spectrometry (EIS) measurements were conducted to explore the charge-transport behavior of the as-fabricated photoanodes, as shown in Figure . The arc diameter of the EIS spectra (Nyquist plots) reflects the interfacial charge-transfer resistance at the electrode/electrolyte interface.[55] The smaller the arc radius, the lower the charge-transfer resistance for the corresponding photoelectrode and vice versa.[61] Clearly, the radius of the EIS spectra of the MoS2/NiWO4 photoanode is smaller than that of the MoS2 NSs and NiWO4 NPs photoanodes, suggesting an efficient separation of the photogenerated charge carriers and the enhanced charge transfer. More importantly, the smallest arc radius is observed for the Ag-MRGON photoanode indicating that it exhibits the best charge-transfer efficiency, which implies that introducing RGO between MoS2 and NiWO4 and elaborating the heterojunction surface with Ag NPs create a favorable physicochemical environment for feasible charge transport at the electrode/electrolyte interface of the heterojunction. The significantly enhanced charge separation and transfer performance in the hybrids might be due to the synergistic effect between the Z-scheme heterojunction and the SPR effect of Ag NPs.

Figure 7

EIS Nyquist plots of MoS2 NSs, NiWO4 NPs, MoS2/NiWO4 heterojunction, and Ag-MRGON heterojunction photoanodes.

Proposed PEC Mechanism

Based on the presented discussion, a viable PEC water-splitting mechanism (based on the Z-scheme pathway) can be proposed. The enhanced PEC results suggest that the Ag-MRGON heterojunction generates significant photogenerated electron–hole pairs with facile charge transfer to suppress their recombination. Furthermore, the synergetic effects between morphological structures, optical properties, and chemical constituents of the designed RGO-mediated Z-scheme PEC system are significant as (i) combining the layered MoS2 and RGO NSs with NiWO4 NPs could provide abundant photoelectroactive sites because of significant specific surface area of the layered structures, facilitate fast electron migration, and enhance light-harvesting capacity,[39] (ii) on account of the SPR effect, decorating the surface of the MoS2/RGO/NiWO4 heterojunction with plasmonic Ag NPs induces the local optical field at the nearby semiconductors, thus enhancing the visible light absorption capability, charge separation, and transfer efficiency at the interfacial region,[41,62] and (iii) introducing RGO at the interface between MoS2 and NiWO4 acts as a mediator of electron migration at the Z-scheme heterojunction interface and thus prevents the corrosion of the MoS2 by the photogenerated holes.

As depicted in Figure a, the Ag-MRGON follows all-solid-state RGO-mediated Z-scheme charge-transfer system rather than a conventional Z-scheme system. The VB XPS analysis of MoS2 and NiWO4 showed the VB maxima at 1.78 and 2.72 eV, respectively (Figure b,c). Consequently, the corresponding CB minima were obtained, taking into account the band gaps according to the equation:[63]where ECB is the CB potential, EVB is the VB potential, and Eg is the band gap. Therefore, ECB of MoS2 and NiWO4 is −0.4 and 0.33 eV, respectively. When the Ag-MRGON photoanode is irradiated by photons with energy greater than their band gap, the electrons of MoS2 and NiWO4 are excited from their respective VBs to CBs, leaving holes on their VBs. Electrons on the CB of NiWO4 recombine with holes on the VB of MoS2 through the electron mediator RGO and hence cause a shuttling electronic effect. This led to the electron accumulation on the CB of MoS2, followed by transfer to the counter electrode to reduce H+ to H2. Conversely, holes are accumulated on the VB of NiWO4 and oxidize water to O2, hence performing a complete PEC water-splitting process. Indeed, the PEC water-splitting activity showed to have better efficiency in the presence of an RGO electron mediator between NiWO4 and MoS2, which is evident from the PEC results with and without RGO (Figure S6b). In contrast, the direct contact of MoS2 with NiWO4 (direct Z-scheme MoS2/NiWO4 heterojunction, Figure S6b) seems to severely reduce the transfer of the photogenerated electron–hole pairs at the interfacial region of the heterojunction because of the surface relaxation and more defect sites at the heterojunction interface.[64] Thus, the introduction of a conductive medium between two semiconductors could significantly facilitate the interfacial charge transfer during the PEC water-splitting reaction. A similar role of the RGO electron mediator to improve the photocatalytic performance of the Z-scheme g-C3N4/RGO/WO3 heterojunction[60] and Z-scheme ZnIn2S/RGO/CoOx-Bi2MoO6 heterojunction[65] has been recently reported.

Figure 8

(a) Schematic diagram for the proposed mechanism of water splitting over the Z-scheme Ag-decorated MoS2/RGO/NiWO4 heterojunction, and VB XPS spectra of (b) MoS2 and (c) NiWO4.

On the other hand, the strong SPR effect of Ag NPs under visible light produces strongly enhanced local electromagnetic fields, which would accelerate the formation rate of the electrons and holes within the semiconductors and thus contributes to the photoinduced process. Upon SPR excitation, the surface electrons of Ag NPs oscillate collectively and might be excited to occupy high energy-level states above the Fermi level. Consequently, these photoexcited electrons overcome the Schottky barrier and transfer to the CBs of the nearby MoS2 and NiWO4. The Ag electrons transferred to the CB of MoS2 promote the photogeneration of electrons which moves to the counter electrode to reduce H+ to H2,[28,41] while those injected into the CB of NiWO4 might accelerate the reaction with the photogenerated holes on the VB of MoS2 through the RGO electron mediator and thus suppressing the photocorrosion of the MoS2 surface and enhancing the overall PEC activity. We believed that the enhanced PEC water splitting over the Ag-MRGON heterojunction is attributed to the synergetic effects between MoS2, NiWO4, RGO, and Ag in the Z-scheme charge-transfer system. The synergy between the components of the designed Z-scheme heterojunction leads to a spatial separation of the photogenerated carriers, which positively prohibits the electron–hole recombination, enhances the oxidation–reduction capability, and thus promotes the PEC activity.

Conclusions

In summary, a novel synergetic and efficient solar-driven Ag NPs-decorated MoS2/RGO/NiWO4 (Ag-MRGON) Z-scheme system was successfully established by injecting RGO between MoS2 and NiWO4 and decorating the heterojunction surface with metallic Ag NPs. This arrangement demonstrated superior PEC water-splitting activity to the corresponding pristine and two-component systems such as MoS2, NiWO4, and MoS2/NiWO4. While RGO in the Ag-MRGON heterojunction serves as an electron mediator for shuttling electrons between MoS2 and NiWO4, leading to the accumulation of photogenerated electrons on the CB of MoS2 with high reduction ability and holes on the VB of NiWO4 with high oxidation ability, metallic Ag NPs could enhance the heterojunction absorption in the visible light spectrum because of the local SPR effect. The Z-scheme Ag-MRGON heterojunction showed the highest photocurrent density (3.9 mA/cm2) and cathodic shift of the onset potential Von for water splitting (0.61 V) as well as the excellent photocurrent stability (>2 h) than that of pristine MoS2, NiWO4, and MoS2/NiWO4 heterojunction. The present work provides a facile recipe to fabricate RGO-injected/Ag-decorated Z-scheme heterojunctions with enhanced PEC water-splitting activity, and it could be extended to the design of a variety of RGO/Ag-based Z-scheme heterojunction systems for future solar-driven energy applications.

Experimental Section

Chemicals

Ammonium heptamolybdate tetrahydrate (NH4)6Mo7O24·4H2O, ACS reagent, 99.98% trace metals basis), nickel (II) nitrate hexahydrate [Ni(NO3)2·6H2O ≥ 99.9%], carbon disulphide anhydrous (CS2, ≥99%), sodium tungstate dihydrate (Na2WO4·2H2O, 99.995% trace metals basis), silver nitrate (AgNO3, ACS reagent, ≥99.0%), and graphite (20 μm) were obtained from Sigma-Aldrich and used without further purification.

Synthesis

Synthesis of GO Nanosheets (NSs)

GO NSs were synthesized by a modified Hummers’ method using graphite powder.[66] The details are provided in the Supporting Information.

Synthesis of MoS2 Nanosheets (NSs)

The hydrothermal synthesis of MoS2 NSs was carried out as follows: 7.4 mmol ammonium heptamolybdate tetrahydrate (precursor for molybdenum) was added and mixed in 90 mL of DI water. This solution was transferred to a Teflon-lined reactor chamber (200 mL). To this solution, 40 mmol CS2 (precursor for sulphur) was added. Nitrogen gas was bubbled in the mixture and then the Teflon-lined stainless-steel reactor was tightly sealed and heated at 200 °C for 10 h in an oven. The reaction mixture was then allowed to cool down. The formed MoS2 precipitate was successively washed with absolute ethanol to remove any traces of CS2 and with DI water to remove any residuals of soluble sulfides and eventually dried at 110 °C overnight.

Synthesis of NiWO4 Nanoparticles (NPs)

NiWO4 NPs were also synthesized using the hydrothermal method by mixing Ni (NO3)2·6H2O (25.9 mmol in 70 mL of DI water) with Na2WO4·2H2O (26.1 mmol in 20 mL of DI water). The mixture was transferred to a 200 mL Teflon-lined autoclave reactor. The steel reactor was tightly sealed and heated at 200 °C for 3 h. After cooling, the resulting solid NiWO4 was separated by centrifugation, successively rinsed with DI water to remove the unreacted reagents and the byproduct of sodium nitrate, collected, dried, and then calcined at 550 °C under a nitrogen environment for 3 h in order to increase the crystallinity.

Synthesis of MoS2/NiWO4

For the hydrothermal synthesis of the MoS2/NiWO4 heterojunction, 26.1 mmol Na2WO4·2H2O and 25.9 mmol Ni(NO3)2·6H2O were mixed together in 100 mL of DI water. Then, 25 mmol MoS2 NSs were added, the mixture was dispersed by the aid of ultrasonication and then transferred into a Teflon-lined autoclave reactor. The reactor was sealed and kept for 5 h at 200 °C in a calibrated oven. The reactor was then allowed to cool, and the resulting nanocomposite was separated by centrifugation and washed with DI water several times. Then, it was dried at 110 °C and eventually heated at 300 °C under a nitrogen environment.

Synthesis of Z-Scheme Ag-Decorated MoS2/RGO/NiWO4

To synthesize the Ag-decorated MoS2/RGO/NiWO4 heterojunction, 20 mmol MoS2 NSs and 200 mg of GO NSs were dispersed in 50 mL of DI water by the aid of ultrasonication for 2 h. Na2WO4·2H2O (20 mmol) and Ni(NO3)2·6H2O (20 mmol) solutions were added to this mixture, and the contents were again ultrasonicated for 1 h. Finally, 50 mL of AgNO3 (0.13 M) was added, and the contents were then transferred to a Teflon-lined autoclave reactor and thermally treated at 180 °C for 3 h. The product was then collected by centrifugation, washed with DI water, dried, and processed under calcination with the flow of nitrogen at 300 °C for 2 h. This sample is hereafter labeled Ag-MRGON.

Characterization

The as-synthesized MoS2 NSs, NiWO4 NPs, MoS2/NiWO4, and Ag-MRGON were characterized using different analytical techniques. The surface morphology and elemental mapping were studied using a field emission scanning electron microscope (TESCAN VELA3) coupled with an energy-dispersive X-ray spectrometer. The crystalline structures were examined using a powder X-ray diffractometer (Rigaku Ultima IV) with Cu Kα radiation (λ = 0.15406 nm). The chemical analysis of the synthesized samples was carried out using X-ray photoelectron spectroscopy (XPS) performed in a Thermo Scientific ESCALAB 250 Xi spectrometer equipped with an Al Kα (1486.6 eV) X-ray source. The base pressure inside the spectrometer was 7 × 10–11 mbar). The binding energies of the obtained XPS spectra were calibrated with respect to the adventitious C 1s peak centered at 284.6 eV. The optical properties of the samples were examined by recording the diffuse reflectance spectroscopy (DRS) spectra using a Horiba UV–vis spectrophotometer.

PEC Water-Splitting Measurements

The PEC water-splitting activities of the prepared photoelectrodes were measured in a 0.5 M Na2SO4 electrolyte with a standard three-electrode PEC cell. The as-prepared photoanodes were used as the working electrode with Ag/AgCl (reference) and Pt foil (counter electrode) in a three-electrode cell. The working electrodes were prepared as follows: initially, 5 mg of the as-synthesized material was dispersed in 3 mL of isopropanol and 3 μL of Nafion solution. The suspension was then subjected to ultrasonic agitation for 10 min. Then, 10 μL dose of the obtained suspension was drop-casted on the fluorine-doped tin oxide substrate with an active area (1 cm × 1 cm) and dried at 100 °C to yield the working electrode. The linear sweep voltammetry (LSV) measurements were conducted under AM 1.5G simulated sunlight illumination at 100 mW cm–2. The photocurrent–potential (J–V) curves were obtained in the potential range from 0 V to 1.25 VRHE at a scan rate of 0.05 mV s–1. The transient photocurrent–time (J–t) curves were recorded under chopped light irradiation (light ON/OFF cycles: 40 s) at a fixed bias potential of 1.0 VRHE. The EIS spectra were obtained under simulated solar light (AM 1.5G, 100 mW cm–2) at 0.2 VRHE in the frequency range of 100 kHz to 0.01 Hz with a small AC amplitude of 5 mV.