Cobalt–nickel layered double hydroxide on hollow Co3S4/CuS derived from ZIF-67 for efficient overall water splitting

Electrochemical water splitting is one of the most efficient strategies to generate clean energy H2. Herein, a novel hollow Co3S4/CuS@CoNi-LDH nanocomposite was designed. This hollow Co3S4/CuS@CoNi-LDH achieved OER overpotential of 220 mV, HER overpotential of 136 mV, respectively, at the current density of 10 mA cm−2. Notably, the hollow Co3S4/CuS@CoNi-LDH serving as both anode and cathode is assembled into a two-electrode water splitting device, and the potential of no more than 1.74 V is required to achieve the overall water splitting efficiency of 10 mA cm−2. This work provides a promising bifunctional catalyst for overall electrochemical water splitting. GRAPHICAL ABSTRACT


Introduction
Due to the shortage of energy supply and environmental pressures, it is urgent to develop a new type of clean renewable energy [1][2][3]. Hydrogen (H 2 ), as a typical new, clean, efficient and renewable energy with high calorific value, has aroused extensive attention in the preparation process [4][5][6][7]. The electrochemical methods of water electrolysis offer a sustainable solution for the production of high-purity H 2 , and the efficient electrocatalyst is required to facilitate the sluggish kinetics of the reaction [8][9][10][11][12][13]. At present, Pt/C, Ir and Ru precious metal oxide catalysts are ahead of other electrocatalysts in the hydrogen evolution reaction (HER) and oxygen evolution reaction (OER), respectively [14][15][16][17][18]. However, the high price and low reserve prevent them from being widely used [19,20]. Therefore, more and more researchers have made tremendous research efforts to prepare an inexpensive and highly active catalyst [21][22][23][24].
Metal-organic framework (MOF) catalysts are constructed by coordination between organic ligands and metal ions/clusters [25]. Materials derived from MOFs generally have the advantages of high electrochemical activity and large specific surface area, which can be used in the design of high-catalytically active electrochemical catalysts [26][27][28][29]. For instance, Huo et al. designed a type of hierarchical Fe-Ni LDH/MOF electrocatalyst through in-situ pseudomorphic transformation from heterometallic MOFs [30]. Also, Li et al. reported an electrocatalyst with Co@Ir core-shell nanoparticles derived from MOF [31]. However, the low efficiency of OER limits its application in electrocatalytic water splitting [32]. Therefore, it is of great significance to develop a transition metal-based catalyst with higher efficiency for both HER and OER in alkaline media. The combination of MOF derivatives and layered double hydroxide (LDH) with high OER activity can meet this requirement [33].
Herein, we synthesized a hollow Co 3 S 4 /CuS via using ZIF-67 as a template and precursor, and then the cobalt-nickel layered double hydroxide (CoNi-LDH) was in situ grown on the hollow Co 3 S 4 /CuS to obtain the dual-functional catalyst for efficient overall water splitting. Compared with monometallic sulfides, bimetallic sulfides have higher electrical conductivity, and the electronic synergy of metal ions increases their reactivity. In addition, the Co 3 S 4 /CuS material after cation exchange shows that it is rougher, which is more conducive to the growth of nanosheets. It can be seen from the SEM that more CoNi-LDH grown on Co 3 S 4 /CuS than that supported on Co 3 S 4 . The interface interaction generated is stronger, and the more active sites are generated, so the catalytic activity can be significantly improved. Thus, in Co 3 S 4 /CuS@CoNi-LDH, the Co 3 S 4 /CuS shows high catalytical activity on HER, while the CoNi-LDH exhibits excellent catalytic performance on OER. The hollow Co 3 S 4 /CuS@CoNi-LDH achieved OER overpotential of 220 mV, HER overpotential of 136 mV, respectively, at the current density of 10 mA cm −2 . Moreover, the Co 3 S 4 /CuS@CoNi-LDH showed outstanding electrochemical stability in 50 h. Finally, the hollow Co 3 S 4 /CuS@CoNi-LDH catalysts serving as both anode and cathode are assembled into a two-electrode water splitting device, which delivered a current density of 10 mA cm −2 with the cell voltage no more than 1.74 V.

Materials preparation
The preparation of the hollow Co 3 S 4 /CuS@CoNi-LDH is schematically demonstrated in Figure S1. First, the ZIF-67 nanoparticles were synthesized using cobalt ions and 2-methylimidazole. In TAA solution at 120°C, ZIF-67 formed hollow Co 3 S 4 polyhedra. Then, some Co cations in Co 3 S 4 were replaced with Cu cations by the cation exchange method to obtain Co 3 S 4 /CuS. Finally, the cobalt-nickel layered double hydroxide nanosheets were in situ grown on the Co 3 S 4 /CuS to obtain Co 3 S 4 /CuS@CoNi-LDH.

Results and discussion
As shown in Figure 1(a), the as-synthesized ZIF-67 nanoparticle is a clear rhombic dodecahedron shape [34], and the nanoparticle size is 289 nm. After the sulfuration, Co 3 S 4 still maintains the same shape as ZIF-67 ( Figure  1b), and the size is 275 nm. The interior of Co 3 S 4 changes from solid to hollow is revealed by transmission electron microscopy (TEM) in the inset of Figure 1(b). The hollow structure of Co 3 S 4 is also known as the Kirkendall effect [35] due to the difference in diffusivity of cobalt and sulfide ions. Then, using the cation exchange method, Co cations are replaced partly by Cu cations to obtain Co 3 S 4 /CuS heterostructure in Figure 1(c), and the morphology still maintained the hollow structure (inset of Figure 1c). The Co 3 S 4 /CuS heterostructure has a slightly rougher surface compared to the smooth surface of ZIF-67, which is more conducive to the growth of the CoNi-LDH nanosheets. Finally, the CoNi-LDH nanosheets are grown on the surface of the Co 3 S 4 /CuS, and it can be seen from Figure 1(d) that the interconnected nanosheets are wrapped on the surface of Co 3 S 4 /CuS, and the thickness of CoNi-LDH nanosheets is 11 nm. The resulting Co 3 S 4 /CuS@CoNi-LDH still maintains a hollow structure (inset of Figure 1d). The three regions in Figure  1(e) represent the lattice planes of CuS, Co 3 S 4 and CoNi-LDH, respectively. The lattice spacings are 0.305, 0.550 and 0.256 nm, which is corresponding to the (102), (111) and (110) lattice planes of CuS, Co 3 S 4 and CoNi-LDH, respectively [36], confirming the coexistence of CuS, Co 3 S 4 and CoNi-LDH. The selected area electron diffraction (SAED) can also be indexed to the crystal planes of CuS, Co 3 S 4 and CoNi-LDH ( Figure 1f). Figure 1(g) is the high-angle annular dark field scanning transmission electron microscopy (HAADF-STEM) and elements mapping of S, Co, Ni and Cu of a Co 3 S 4 /CuS@CoNi-LDH, and the uniform distribution of S, Co, Ni and Cu elements is in the Co 3 S 4 /CuS@CoNi-LDH, especially the presence of Ni element proves that CoNi-LDH grows in the entire hollow structure of Co 3 S 4 /CuS heterostructure. The morphology and elements mapping of Co 3 S 4 , Co 3 S 4 /CuS and Co 3 S 4 @CoNi-LDH are shown in Figures  S2-S4. The SEM and TEM images of the Co 3 S 4 @CoNi-LDH are shown in Figure S5-S7.
The X-ray diffraction (XRD) measurement was performed on the Co 3 S 4 , Co 3 S 4 /CuS and Co 3 S 4 @CoNi-LDH and Co 3 S 4 /CuS@CoNi-LDH. As well seen, the XRD pattern of the synthesized ZIF-67 agrees well with the simulated pattern (CCDC deposition No.671073), confirming that the ZIF-67 MOFs have been synthesized successfully with high purity in Figure S8. The representative diffraction peaks located at 16.1°, 31.3°, 47.2°and 55.1°in the XRD pattern of Co 3 S 4 in Figure  2(a) are indexed to (111), (311), (422) and (440) planes, respectively [37]. After cation exchange, a diffraction peak of CuS appears at 29.3°, corresponding to the (102) plane [38]. In the Co 3 S 4 /CuS@CoNi-LDH, the diffraction peak at 31.3°of Co 3 S 4 is split into 29.3°and 32.9°c orresponding to the (102) and (006) planes of CuS, respectively. When the CoNi-LDH is grown, the diffraction peaks appear at 11.6°and 60.9°, which is corresponding to the (003) and (110) planes of CoNi-LDH [39]. The XRD results state the successful synthesis of Co 3 S 4 /CuS@CoNi-LDH.
To further demonstrate the heterostructure in Co 3 S 4 / CuS@CoNi-LDH, the Fourier transform infrared (FTIR) spectra of Co 3 S 4 , Co 3 S 4 /CuS, Co 3 S 4 @CoNi-LDH and Co 3 S 4 /CuS@CoNi-LDH were obtained, as shown in Figure 2(b). In Co 3 S 4 , the peak at 1092 cm −1 is indexed to the vibration of the Co-S bond [40]. Compared to Co 3 S 4 , Co 3 S 4 /CuS and Co 3 S 4 /CuS@CoNi-LDH show an obvious blue shift from 1092 cm −1 to 1149 cm −1 . This shift of the peak position in Co 3 S 4 /CuS and Co 3 S 4 /CuS@CoNi-LDH is due to the introduction of Cu, where part of S in Co 3 S 4 is bound to Cu in the interface of Co 3 S 4 and CuS, indicating a strong electronic interaction between Co 3 S 4 and CuS. In Co 3 S 4 /CuS and Co 3 S 4 /CuS@CoNi-LDH, the peak position of Cu-S bond has a clear red shift from 467 cm −1 to 430 cm −1 compared with the Cu-S bond of CuS in the previous literature [41], which suggests that the Co 3 S 4 /CuS heterojunction results in a strong interaction between Co 3 S 4 and CuS. For Co 3 S 4 @CoNi-LDH and Co 3 S 4 /CuS@CoNi-LDH, two peaks at 490 and 559 cm −1 are assigned to Ni-O and Co-O bond [42,43], respectively. Similarly, Raman spectra of Co 3 S 4 , Co 3 S 4 /CuS, Co 3 S 4 @CoNi-LDH and Co 3 S 4 /CuS@CoNi-LDH are shown in Figure 3(c). In Co 3 S 4 , the peaks at 498 cm −1 is corresponding to the F 2g of vibration mode of Co 3 S 4 [44]. Interestingly, the peaks have obvious red shifts at Co 3 S 4 /CuS and Co 3 S 4 /CuS@CoNi-LDH. In Co 3 S 4 /CuS, the peak shifts from 498 cm −1 to 463 cm −1 indicating that the heterojunction is formed between Co 3 S 4 and CuS, which is consistent with the FTIR result. Similar to Co 3 S 4 /CuS, the Co-S peak in Co 3 S 4 /CuS@CoNi-LDH has a red shift from 498 cm −1 to 423 cm −1 . In Co 3 S 4 /CuS and Co 3 S 4 /CuS@CoNi-LDH, the peak at 251 cm −1 is assigned to Cu-S, and the peak has an obvious red shift from 266 cm −1 to 251 cm −1 compared to the previous literature [45]. The peaks at 627 and 1518 cm −1 are attributed to the Co-O stretching and two-magnon (2M) mode of the Ni-O [46,47], respectively, in CoNi-LDH.
The element composition and surface chemical environment of Co 3 S 4 /CuS@CoNi-LDH was also studied via the X-ray photoelectron spectroscopy (XPS) analysis. The XPS survey spectrum (Figure 2d) shows the presence of Cu, S, Ni, and Co as the main components of Co 3 S 4 /CuS@CoNi-LDH. The valence states of Cu in Figure 2(e) and Figure S11 are +1 and +2, indicating that the Cu ions successfully exchanged the Co ions [48]. The binding energy of Cu 2p in Co 3 S 4 /CuS ( Figure S11) and Co 3 S 4 /CuS@CoNi-LDH (Figure 2e) is lower than Cu 2p of CuS in the previous literature [49], which is due to the electron transfer from Co 3 S 4 to CuS in the interface of Co 3 S 4 /CuS. In the high-resolution XPS spectrum (Figure 2f) of S 2p, the weak peaks at 161.8 and 163.2 eV are S 2− , demonstrating the formation of metal sulfides, and the high-energy peak position around 168.6 eV can be assigned to the sulfate groups [50]. Compared with   Figures 2(g) and S10, proving the existence of CoNi-LDH [51]. The high-resolution spectra of Co 2p in Figure 2(h) and S9-S11 show that the valence states of Co in all samples are +2 and +3, which are same from those in Co 3 S 4 , further indicating the formation of Co 3 S 4 [52]. It is worth noting that the binding energy of Co (798.1 and 796.8 eV) in Co 3 S 4 /CuS is higher than that in Co 3 S 4 (797.8 and 794.7 eV), suggesting that there is the electron transfer from Co 3 S 4 to CuS in the interface of Co 3 S 4 and CuS. Based on the above experimental results, the Co 3 S 4 /CuS heterostructure is formed in the Co 3 S 4 /CuS@CoNi-LDH. Elemental compositions of Catalysts are determined by XPS in Table S1.
The N 2 adsorption-desorption isotherm (Figure 2i) was carried out, and the obtained results are listed in Table S2. The specific surface area size relationship is as follows: Co 3 S 4 /CuS@CoNi-LDH (258.10m 2 g −1 ) > Co 3 S 4 /CoNi-LDH (183.71 m 2 g− 1 ) > Co 3 S 4 / CuS (165.55 m 2 g −1 ) > Co 3 S 4 (150.55 m 2 g −1 ). The specific surface area of Co 3 S 4 /CuS/CoNi-LDH and Co 3 S 4 /CoNi-LDH is slightly larger than that of Co 3 S 4 / CuS and Co 3 S 4 , indicating that the specific surface area is enlarged after loading LDH. Among them, the largest specific surface area of Co 3 S 4 /CuS@CoNi-LDH contributes to the electrochemical performance due to its more active sites and faster electrolyte diffusion. From the pore size distribution in Figure S12, indicating that they are all mesoporous structure.
The OER activity of Co 3 S 4 /CuS@CoNi-LDH was investigated in 1 M potassium hydroxide (KOH). Figure  3(a) shows the linear sweep voltammetry (LSV) curves of Co 3 S 4 /CuS@CoNi-LDH with IR compensation, which exhibits a low overpotential of 220 mV at the current density of 10 mA cm −1 . The comparative samples overpotential are shown in Table S3. Besides, Co 3 S 4 /CuS@CoNi-LDH shows better OER performance than the state-ofthe-art RuO 2 . In addition to this, it can be observed that the LDH-supported catalyst is more active than the unsupported one. It is speculated that LDH provides more OH-to participate in the OER reaction. Figure 3(b) shows the overpotential comparison of Co 3 S 4 /CuS@CoNi-LDH and the control sample. It can be seen that Co 3 S 4 /CuS@CoNi-LDH has a relatively low OER overpotential. The Co 3 S 4 /CuS@CoNi-LDH has a lower Tafel slope of 90.78 mV dec −1 in OER (Figure 3c), which is lower than that of the state-of-the-art RuO 2 catalyst, revealing the higher OER rate and favorable kinetics of it. The turnover frequency (TOF) can reflect the intrinsic activity of the catalyst, and it can be found that the catalytic activity of Co 3 S 4 /CuS@CoNi-LDH is higher than that of other samples in Figure 3(d). Impressively, the Co 3 S 4 /CuS@CoNi-LDH exhibits superior electrocatalytic performance over other previously reported CoNibased catalysts (Figure 3e).
Electrochemical impedance spectroscopy (EIS) measurements are shown in Figure 3(f). The EIS data are all fitted by an equivalent circuit (inset of Figure 3f), consisting of a resistor (R 1 ) in series with a parallel combination of a resistor (R 2 ) and a constant-phase element [53]. The lower R 2 value corresponds to the faster reaction rates. The Co 3 S 4 /CuS@CoNi-LDH in OER shows the smallest R 2 , which is consistent with the low overpotential and small Tafel slope in the above analysis. The resulting R 2 values are shown in Table S4. Figure 3(g) shows that Co 3 S 4 /CuS@CoNi-LDH maintains a current density of 10 mA cm −2 for at least 50 h, the overpotential only changed slightly by 1.0%. As a comparison, the overpotential of RuO 2 , Co 3 S 4, Co 3 S 4 /CuS, and Co 3 S 4 @CoNi-LDH ( Figure S13-S16) at the current density of 10 mA cm −2 changed by 10.9%, 1.8%, 13.6% and 1.7%, respectively. It is worthwhile noting that the Co 3 S 4 /CuS@CoNi-LDH shows outstanding catalytical stability, and the high catalytic activity is still maintained even after 50 h of continuous operation. The cyclic voltammetric (CV) scans with different rates were performed to determine the double-layer capacitance C dl , which can be used to estimate the ECSA and the corresponding intrinsic activity [54]. The CV curves were measured in N 2 -saturated 1 M KOH in Figure S17, and C dl calculated from CV is shown in Figure S18. The C dl value of Co 3 S 4 /CuS@CoNi-LDH is much higher than that other catalysts and RuO 2 , further indicating that Co 3 S 4 /CuS@CoNi-LDH can expose more catalytical active site. The durability of the Co 3 S 4 /CuS@CoNi-LDH after the OER test was confirmed by TEM and XRD ( Figures S19 and S20), and no prominent change was found.
The electrocatalytic HER performance of each material was also investigated in Figure 4. Figure 4(a) shows the LSV curves of samples with IR compensation, which shows that the Co 3 S 4 /CuS@CoNi-LDH has high HER activity. The overpotential of Co 3 S 4 /CuS@CoNi-LDH is 136 mV at a current density of 10 mA cm −1 , which is slightly higher than the commercial Pt/C and lower than other comparative sample in Figure 4(b) and Table S5. Therefore, the enhanced activity of Co 3 S 4 /CuS@CoNi-LDH is related to the heterostructure of Co 3 S 4 /CuS. To study the kinetic activity of each catalyst, the Tafel slope curve of each catalyst is plotted in Figure 4(c). The Tafel slope of Co 3 S 4 /CuS@CoNi-LDH in HER is 126.3 mV dec −1 , which is slightly larger than the commercial Pt/C, but smaller than Co 3 S 4 /CuS, Co 3 S 4 @CoNi-LDH and Co 3 S 4 , consistent with the results measured above. The TOF of Co 3 S 4 /CuS@CoNi-LDH is higher than that of other samples in Figure 4(d). Moreover, Co 3 S 4 /CuS@CoNi-LDH exhibits superior electrocatalytic HER performance over other previously reported CoNi-based catalysts in Figure 4(e). Similarly, the catalyst kinetics at the electrode interface was investigated by EIS measurements in Figure 4(f). The order of R 2 values obtained in HER indicates that the Co 3 S 4 /CuS@CoNi-LDH has a more optimized charge transfer ability than other catalysts, and the resulting R 2 values are shown in Table S6. Figure 4(g) shows that Co 3 S 4 /CuS@CoNi-LDH maintains a current density of 10 mA cm −2 for at least 50 h, and the overpotential only changed slightly by 1.8%. As a comparison, in Figures S22-S24, the potentials of Co 3 S 4 /CuS, Co 3 S 4 , and Co 3 S 4 @CoNi-LDH at 10 mA cm −2 changed by 5.0%, 3.2%, and 3.8%, respectively. After 35 h, the potential of Pt/C decreased sharply by 23.0% in Figure  S21. Therefore, the above experimental results prove the excellent stability of the Co 3 S 4 /CuS@CoNi-LDH as a water electrolysis catalyst. As shown in Figure  S25, the CV curves in N 2 -saturated 1 M KOH. The results of C dl calculated by CV are shown in Figure  S26. The results of C dl indicate the Co 3 S 4 /CuS@CoNi-LDH has higher ECSA and more catalytical active sites. The durability of the Co 3 S 4 /CuS@CoNi-LDH after the HER test was confirmed by TEM and XRD (Figures S27 and S28), and no prominent change was found. Inspired by the excellent electrocatalytic properties of HER and OER, the Co 3 S 4 /CuS@CoNi-LDH serves as anode and cathode to prepare electrolyzers for overall water splitting in an alkaline solution. As shown in Figure 5(c), the dual-functional Co 3 S 4 /CuS@CoNi-LDH and Co 3 S 4 @CoNi-LDH achieved a current density of 10 mA cm −2 at potentials of 1.74 and 1.84 V, respectively, which is lower than the overall hydrolysis potential (1.88 V) of RuO 2 ||Pt/C. The calculated potential of total hydrolysis and the actual measurement are shown in Figure 5(d). The overall hydrolysis potential calculated based on HER and OER is slightly lower than the actual measured. In addition to the high catalytic efficiency, the Co 3 S 4 /CuS@CoNi-LDH||Co 3 S 4 /CuS@CoNi-LDH was tested at 10 mA cm −2 for 50 h, only a weak potential decay was observed, and the potential can maintain 99.9%, which further proves the durability. In Figures S29 and S30, the overall water splitting stability of Co 3 S 4 @CoNi-LDH||Co 3 S 4 @CoNi-LDH and RuO 2 ||Pt/C is compared. After 50 h, the Co 3 S 4 @CoNi-LDH||Co 3 S 4 @CoNi-LDH and RuO 2 ||Pt/C remain at 89.5% and 90.0% of their original values, respectively. To realize the efficient utilization of green resources, the assembled electrolyzer is driven by solar cells. Figure 5(a) is a schematic diagram of overall hydrolysis under the driving conditions of solar panels. When the assembled cell was exposed to sunlight, bubbles were generated at both the cathode and anode of the two-electrode system in Figure 5(b). The potential of the solar panel was measured by the voltmeter in Figure S31, the solar panel produced a potential of 1.97V. And the attached video shows the overall water splitting dynamics and the bubble generation and accumulation process driven by the solar panel, proving that the Co 3 S 4 /CuS@CoNi-LDH has excellent bifunctional catalytic performance.

Conclusion
The hollow Co 3 S 4 /CuS@CoNi-LDH is prepared by using ZIF-67 as a template and precursor, and the CoNi-LDH nanosheets are uniformly grown on the surface of hollow Co 3 S 4 /CuS to form a unique heterostructure electrocatalytic catalyst. The hollow Co 3 S 4 /CuS@CoNi-LDH achieved OER overpotential of 220 mV, HER overpotential of 136 mV, respectively, at the current density of 10 mA cm −2 . The excellent OER and HER activities can be attributed to the abundant active sites from hollow structures and heterojunctions. The hollow structure can provide a larger surface area, exposing more catalytical active sites, and the heterostructure improves the structural stability, and the synergistic effect of the bimetal sulfides increases the catalytical activity. This work opens a novel avenue for designing the bifunctional catalyst for overall electrochemical water splitting.

Disclosure statement
No potential conflict of interest was reported by the author(s).