MOF-Derived Ultrathin Cobalt Molybdenum Phosphide Nanosheets for Efficient Electrochemical Overall Water Splitting

The development of high-performance and cost-effective earth-abundant transition metal-based electrocatalysts is of major interest for several key energy technologies, including water splitting. Herein, we report the synthesis of ultrathin CoMoP nanosheets through a simple ion etching and phosphorization method. The obtained catalyst exhibits outstanding electrocatalytic activity and stability towards oxygen and hydrogen evolution reactions (OER and HER), with overpotentials down to 273 and 89 mV at 10 mA cm−2, respectively. The produced CoMoP nanosheets are also characterized by very small Tafel slopes, 54.9 and 69.7 mV dec−1 for OER and HER, respectively. When used as both cathode and anode electrocatalyst in the overall water splitting reaction, CoMoP-based cells require just 1.56 V to reach 10 mA cm−2 in alkaline media. This outstanding performance is attributed to the proper composition, weak crystallinity and two-dimensional nanosheet structure of the electrocatalyst.


Introduction
Hydrogen, with a high gravimetric energy density (142 MJ kg −1 ) and zero-carbon emissions, is both a key component in the chemical industry and a very appealing energy carrier for clean and sustainable energy storage and supply [1,2]. Since molecular hydrogen is not freely available in nature, it needs to be extracted from hydrogen-containing compounds. Currently, fossil fuels are the main source of H 2 , which involves the release of large amounts of carbon. The electrochemical water splitting is the main green alternative to produce H 2 , but it is seriously hampered by the high cost and insufficient durability of current electrocatalysts, based on scarce novel metals such as Pt, Ir and Ru [3][4][5][6], and the slow kinetics of the hydrogen and oxygen evolution reactions (HER, OER), which makes water electrolysis not competitive with steam reforming of natural gas or coal gasification processes [7][8][9].

Preparation of ZIF-67
ZIF-67 was produced following a previously reported procedure with some modifications [16,43]. Briefly, 0.87 g Co(NO 3 ) 2 ·6H 2 O was dissolved in 30 mL of methanol to obtain a clear solution. Subsequently, the above solution was poured into 30 mL of methanol containing 1.97 g of 2-methylimidazole under vigorous stirring. After mixing completely, the solution was incubated for 24 h at room temperature. Purple precipitates were collected by centrifugation; they were washed with methanol three times and then dried at 60 • C overnight.

Preparation of Mo-Co MOFs
One hundred and twenty milligrams of as-prepared ZIF-67 powder was ultrasonically re-dispersed in 20 mL of ethanol. This solution was poured into 100 mL of an aqueous solution containing 50 mg, 100 mg and 200 mg of ammonium molybdate under continuous magnetic stirring. The mixture was then stirred vigorously for 24 h at room temperature. Lavender precipitates were collected by centrifugation, washed with water at least three times and freeze-dried overnight.

Preparation of CoP and CoMoP
The obtained ZIF-67 and Mo-Co MOFs powders were placed in a porcelain boat within a horizontal tube furnace. In another boat, a 20× mass amount of NaH 2 PO 2 ·H 2 O was placed at the upstream side of the tube furnace. The material was then annealed at 350 • C under N 2 flow. After calcination for 2 h, the final black products were denoted as CoP and CoMoP, respectively.

Structural Characterization
Powder X-ray diffraction (XRD) was performed on a Bruker AXS D8 Advance X-ray diffractometer (Bruker, Billerica, MA, USA) with Cu-Kα radiation (λ = 1.5406 Å). Scanning electron microscopy (SEM) analysis was conducted with a Zeiss Auriga microscope (Carl Zeiss, Jena, Germany) equipped with an energy dispersive spectroscope analyses (EDS) detector operating at 20 kV. Transmission electron microscopy (TEM), High-resolution TEM (HRTEM), Annular dark-field scanning transmission electron microscope (HAADF-STEM) and electron energy loss spectroscopy (EELS) analysis were obtained using a field emission gun FEI Tecnai F20 microscope (FEI, Hillsboro, OR, USA) with a Gatan Quantum filter (Pleasanton, CA, USA) at 200 kV. X-ray photoelectron spectroscopy (XPS) measurements were conducted on a SPECS using an Al anode XR50 source at 150 W and a 150 MCD-9 detector from Phoibos (SPECS, Berlin, Germany).

Electrochemical Measurements
Electrochemical characterization was performed in a standard three-electrode system using an electrochemical workstation (CHI 760E, CH Instruments, Shanghai, China) in 1 M KOH solution (PH = 14). A graphite rod counter electrode and a Hg/HgO reference electrode were employed. Electrochemical impedance spectroscopy (EIS) was measured within a frequency from 0.01 Hz to 10 kHz at 10 mV amplitude. The initial voltage was fixed at the overpotential required to obtain a current density of 10 mA cm −2 . The electrochemically active surface area (ECSAs) was determined using the electrochemical double-layer capacitance (C dl ) obtained with cyclic voltammetry data at different scan rates (v = 20-100 mV·s −1 ). Stability was determined by CV using 3000 cycles at 100 mV·s −1 and by chronopotentiometry at 10 mA·cm −2 . Overall water splitting tests were carried out in a two-electrode system with the voltage range of 0-2.0 V at a scan rate of 5 mV·s −1 in 1.0 M KOH electrolyte.

Characterization of Electrocatalysts
CoMoP nanosheets were produced by a three-step process involving the synthesis of a Co-MOF, its etching and partial cation exchange and a final phosphorization step, as schematically illustrated in Figure 1a. First, a cobalt-based zeolitic imidazolate framework (ZIF-67), consisting of polyhedral-shaped micrometer-size particles, was produced as a self-sacrificial template (Figure 1b). The ZIF-67 was reacted with ammonium molybdate with the double role of etching the structure and partially replacing Co 3+ cations by Mo 6+ , yielding a porous wrinkled nanosheet-based material that we refer to as Co-Mo MOF ( Figure 1c). Finally, the Co-Mo MOF was annealed within a tube furnace containing NaH 2 PO 2 at 350 • C for 2 h to produce a porous phosphide with a similar wrinkled nanosheet-based morphology that we denoted as CoMoP (Figures 1d and S1). EDS analysis of CoMoP showed a Co-Mo atomic ratio of Co/Mo = 4, and a phosphorus-metal atomic ratio of P/M =2.6 (M = Co + Mo) ( Figure S2). NaH2PO2 at 350 °C for 2 h to produce a porous phosphide with a similar wrinkled nanosheet-based morphology that we denoted as CoMoP (Figures 1d and S1). EDS analysis of CoMoP showed a Co-Mo atomic ratio of Co/Mo = 4, and a phosphorus-metal atomic ratio of P/M =2.6 (M = Co + Mo) ( Figure S2).
In order to study the effect of ammonium molybdate, we replaced this chemical with an alternative Mo precursor. Using sodium molybdate as Mo source, the morphology of the final product was much more compact, consisting of partially porous cubes ( Figure  S3). Besides, EDS analysis revealed the molybdenum content of this material to be much lower (Co/Mo = 18.8) than that of CoMoP. We will refer to this material as Mo-CoP. As a reference, a Mo-free CoP was obtained by directly annealing the ZIF-67 in the presence of the phosphorous source, with no etching step ( Figure S4). The obtained material also displayed a more compact geometry than that of the CoMoP nanosheets.   Figure 2d) showed CoMoP to present a weak crystallinity, with strong middle/long-range disordered [44][45][46]. In this regard, while the XRD patterns of ZIF-67 and Na2MoO4-ZIF-67 displayed a good crystallinity (Figure 3a), the Co-Mo MOF already presented a mostly amorphous structure. After phosphorization, CoP maintained a relatively well-organized lattice, and CoMoP displayed a weak crystallographic order, consistently with HRTEM results (  In order to study the effect of ammonium molybdate, we replaced this chemical with an alternative Mo precursor. Using sodium molybdate as Mo source, the morphology of the final product was much more compact, consisting of partially porous cubes ( Figure S3). Besides, EDS analysis revealed the molybdenum content of this material to be much lower (Co/Mo = 18.8) than that of CoMoP. We will refer to this material as Mo-CoP. As a reference, a Mo-free CoP was obtained by directly annealing the ZIF-67 in the presence of the phosphorous source, with no etching step ( Figure S4). The obtained material also displayed a more compact geometry than that of the CoMoP nanosheets.  Figure 2d) showed CoMoP to present a weak crystallinity, with strong middle/long-range disordered [44][45][46]. In this regard, while the XRD patterns of ZIF-67 and Na 2 MoO 4 -ZIF-67 displayed a good crystallinity (Figure 3a), the Co-Mo MOF already presented a mostly amorphous structure. After phosphorization, CoP maintained a relatively well-organized lattice, and CoMoP displayed a weak crystallographic order, consistently with HRTEM results ( Figure 3b).

Oxygen Evolution Reaction
The OER activity of CoMoP was evaluated at room temperature using a three-electrode system in a 1.0 M KOH alkaline solution. As a reference, Mo-CoP, CoP and a commercial RuO2 catalyst were also evaluated in the same cell and reaction conditions. The LSV polarization curves displayed the CoMoP to be characterized by an outstanding OER activity, with an overpotential of only 273 mV at a current density of 10 mA cm −2 (Figures  4a and S8). This overpotential is well below that of Mo-CoP, CoP and the RuO2 electro-

Oxygen Evolution Reaction
The OER activity of CoMoP was evaluated at room temperature using a three-electrode system in a 1.0 M KOH alkaline solution. As a reference, Mo-CoP, CoP and a commercial RuO 2 catalyst were also evaluated in the same cell and reaction conditions. The LSV polarization curves displayed the CoMoP to be characterized by an outstanding OER activity, with an overpotential of only 273 mV at a current density of 10 mA cm −2 (Figures 4a and S8). This overpotential is well below that of Mo-CoP, CoP and the RuO 2 electrocatalyst tested here and outperforms that of previously reported CoP-based OER catalysts, as shown in Table S1. As displayed in Figure 4b, CoMoP was not only characterized by the lowest overpotential at 10 mA cm −2 but also provided the lowest Tafel slope, 54.9 mV dec −1 . This value was significantly below that of Mo-CoP (60.4 mV dec −1 ), CoP (71.5 dec −1 ) and RuO 2 (86.4 mV dec −1 ), which indicates CoMoP to have associated a faster OER reaction kinetics [53,54]. The charge transport/transfer ability of the electrocatalysts was evaluated by electrochemical impedance spectroscopy (EIS). An equivalent circuit model including a charge transfer resistance (Rct) and a solution resistance (Rs) during the OER process was used to fit the Nyquist plots displayed in Figure 4d [55,56]. CoMoP exhibited the smallest Rct (17.99 Ω), well below that of Mo-CoP (Rct = 27.90 Ω), CoP (Rct = 36.70 Ω) and RuO2 (45.47). These results reveal the CoMoP nanosheets to enable a faster charge transfer at the electrode/electrolyte interfaces, thus accelerating the OER electrocatalytic kinetics.
The long-term stability of CoMoP was further analyzed by CV and chronopotentiometry measurements [57,58]. Figure 4e shows how the LSV curve of CoMoP after 3000 CV cycles closely resembles that of the first cycle. The chronoamperometry test displayed CoMoP to have an outstanding long-term catalytic activity with just a 3% current density decay after 100 h of operation at 273 mV (Figure 4f). SEM images of the post-catalysts after OER testing at high current showed the ultrathin CoMoP nanosheets to partially sinter into a highly porous structure with thicker walls (Figure S10). At the same time, the EDX result showed that a loss of P occurred during the OER. These results are consistent with the reorganization of the metal phosphide into a metal (oxy)-hydroxide during the OER reaction [59,60].   (Figures S9 and 4c). CoMoP displayed larger C dl (12.6 mF cm −2 ) than Mo-CoP (8.7 mF cm −2 ), CoP (4.9 mF cm −2 ) and RuO 2 (2.3 mF cm −2 ). This result indicates that CoMoP offers a higher density of accessible electrochemical active sites, which we relate to the proper composition and nanosheet structure of CoMoP.
The charge transport/transfer ability of the electrocatalysts was evaluated by electrochemical impedance spectroscopy (EIS). An equivalent circuit model including a charge transfer resistance (R ct ) and a solution resistance (R s ) during the OER process was used to fit the Nyquist plots displayed in Figure 4d [55,56]. CoMoP exhibited the smallest R ct (17.99 Ω), well below that of Mo-CoP (R ct = 27.90 Ω), CoP (R ct = 36.70 Ω) and RuO 2 (45.47). These results reveal the CoMoP nanosheets to enable a faster charge transfer at the electrode/electrolyte interfaces, thus accelerating the OER electrocatalytic kinetics.
The long-term stability of CoMoP was further analyzed by CV and chronopotentiometry measurements [57,58]. Figure 4e shows how the LSV curve of CoMoP after 3000 CV cycles closely resembles that of the first cycle. The chronoamperometry test displayed CoMoP to have an outstanding long-term catalytic activity with just a 3% current density decay after 100 h of operation at 273 mV (Figure 4f). SEM images of the post-catalysts after OER testing at high current showed the ultrathin CoMoP nanosheets to partially sinter into a highly porous structure with thicker walls (Figure S10). At the same time, the EDX result showed that a loss of P occurred during the OER. These results are consistent with the reorganization of the metal phosphide into a metal (oxy)-hydroxide during the OER reaction [59,60].

Hydrogen Evolution Reaction
The HER performance of CoMoP was evaluated in 1.0 M KOH using a three-electrode system, and it was compared with that of Mo-CoP, CoP and a commercial Pt/C catalyst. As shown in Figures 5a and S8b, the CoMoP electrocatalyst displayed a relatively low HER overpotential of 89 mV at the current density of 10 mA cm −2 , slightly above that of Pt/C (42 mV) and well below that of Mo-CoP (154 mV), CoP (165 mV) and most previously reported phosphide-based HER electrocatalysts (Table S2). The Tafel slope of CoMoP (69.7 mV dec −1 ) was also much lower than those of Mo-CoP (83.7 mV dec −1 ), CoP (113.4 mV dec −1 ) and close to that of Pt/C (56.1 mV dec −1 ) (Figure 5b), which indicated rapid HER reaction kinetics following the Volmer-Heyrovsky mechanism [61,62]. CoMoP also displayed the smallest semicircular diameter in the Nyquist plot of the EIS data among the phosphide catalysts tested (Figure 5c), showing the lowest charge transfer resistance during catalytic processes. In terms of stability under HER conditions, Figure 5d displays how CoMoP suffered a minor change in the LSV curves after 3000 cycles. Additionally, the CA measurement showed the current density to decrease just 6% after 100 h of continuous operation under HER conditions at an overpotential 89 mV (Figure 5e). The morphology and composition of the catalyst after long-term HER are displayed in Figure S11. In this case, minor changes in structure and a moderate P loss were observed, which points to notable catalyst stability under HER.

Hydrogen Evolution Reaction
The HER performance of CoMoP was evaluated in 1.0 M KOH using a three-electrode system, and it was compared with that of Mo-CoP, CoP and a commercial Pt/C catalyst. As shown in Figures 5a and S8b, the CoMoP electrocatalyst displayed a relatively low HER overpotential of 89 mV at the current density of 10 mA cm −2 , slightly above that of Pt/C (42 mV) and well below that of Mo-CoP (154 mV), CoP (165 mV) and most previously reported phosphide-based HER electrocatalysts (Table S2). The Tafel slope of CoMoP (69.7 mV dec −1 ) was also much lower than those of Mo-CoP (83.7 mV dec −1 ), CoP (113.4 mV dec −1 ) and close to that of Pt/C (56.1 mV dec −1 ) (Figure 5b), which indicated rapid HER reaction kinetics following the Volmer-Heyrovsky mechanism [61,62]. CoMoP also displayed the smallest semicircular diameter in the Nyquist plot of the EIS data among the phosphide catalysts tested (Figure 5c), showing the lowest charge transfer resistance during catalytic processes. In terms of stability under HER conditions, Figure 5d displays how CoMoP suffered a minor change in the LSV curves after 3000 cycles. Additionally, the CA measurement showed the current density to decrease just 6% after 100 h of continuous operation under HER conditions at an overpotential 89 mV (Figure 5e). The morphology and composition of the catalyst after long-term HER are displayed in Figure  S11. In this case, minor changes in structure and a moderate P loss were observed, which points to notable catalyst stability under HER.

Overall Water Splitting
Due to the excellent OER and HER performances demonstrated by CoMoP, a twoelectrode configuration electrolyzer with CoMoP both as the positive and negative electrodes was constructed and tested for OWS in 1.0 M KOH solution (Figure 6a). As shown from the polarization curves displayed in Figure 6b, the assembled device just required a cell voltage of 1.56 V to reach a current density of 10 mA cm −2 , which is significantly below that of a cell containing Pt/C and RuO 2 electrodes (1.61 V). More importantly, after 40 h of continuous operation at 100 mA cm −2 , the CoMoP-based cell still maintained an outstanding performance, with just a 14.8% loss at high current density (Figure 6c). Thus, the as-prepared CoMoP can be considered as a highly competitive electrocatalytic catalyst for OWS compared with the previously reported OWS catalysts (Figure 6d, Table S3). Besides, its outstanding stability demonstrates its potential for large-scale hydrogen production from water splitting.

Overall Water Splitting
Due to the excellent OER and HER performances demonstrated by CoMoP, a twoelectrode configuration electrolyzer with CoMoP both as the positive and negative electrodes was constructed and tested for OWS in 1.0 M KOH solution ( Figure 6a). As shown from the polarization curves displayed in Figure 6b, the assembled device just required a cell voltage of 1.56 V to reach a current density of 10 mA cm −2 , which is significantly below that of a cell containing Pt/C and RuO2 electrodes (1.61 V). More importantly, after 40 h of continuous operation at 100 mA cm −2 , the CoMoP-based cell still maintained an outstanding performance, with just a 14.8% loss at high current density (Figure 6c). Thus, the asprepared CoMoP can be considered as a highly competitive electrocatalytic catalyst for OWS compared with the previously reported OWS catalysts (Figure 6d, Table S3). Besides, its outstanding stability demonstrates its potential for large-scale hydrogen production from water splitting.

Conclusions
In conclusion, ultrathin CoMoP nanosheets were engineered using the Co-MOF ZIF-67 as a self-sacrificial template and ammonium molybdate as a shape-defining agent and Mo source. CoMoP nanosheets exhibited outstanding performance towards HER and OER in alkaline media, which we associate with the proper transport properties and electronic energy levels provided by their composition and their porous nanosheet structure. In particular, CoMoP presented low overpotentials of 89 and 273 mV at a current density of 10 mA cm −2 for HER and OER, respectively. Furthermore, CoMoP electrocatalysts also showed excellent long-term stabilities in alkaline electrolytes, with a minor current density decrease after 100 h continuous operation. When used for OWS, a cell voltage of only 1.56 V was needed to reach a current density of 10 mA cm −2 . This work provides a suitable strategy to synthesize high-performance Co-Mo-P electrocatalysts with abundant exposed active sites and effective avenues for charge and electrolyte transport, and it can be

Conclusions
In conclusion, ultrathin CoMoP nanosheets were engineered using the Co-MOF ZIF-67 as a self-sacrificial template and ammonium molybdate as a shape-defining agent and Mo source. CoMoP nanosheets exhibited outstanding performance towards HER and OER in alkaline media, which we associate with the proper transport properties and electronic energy levels provided by their composition and their porous nanosheet structure. In particular, CoMoP presented low overpotentials of 89 and 273 mV at a current density of 10 mA cm −2 for HER and OER, respectively. Furthermore, CoMoP electrocatalysts also showed excellent long-term stabilities in alkaline electrolytes, with a minor current density decrease after 100 h continuous operation. When used for OWS, a cell voltage of only 1.56 V was needed to reach a current density of 10 mA cm −2 . This work provides a suitable strategy to synthesize high-performance Co-Mo-P electrocatalysts with abundant exposed active sites and effective avenues for charge and electrolyte transport, and it can be employed to further tune the structure and composition of other 2D nanostructures with optimized performance towards OWS and other electrocatalytic reactions.
Author Contributions: The manuscript was prepared through the contribution of all authors. A.C. guided the project and supervised the work. X.W., L.Y., C.X., R.D., R.H. and P.G. conceived and prepared the manuscript. X.W. and L.Y. produced the samples. X.W. and C.X. performed the electrochemical measurements and analyzed the results. X.H., J.A. and P.G. performed TEM, HRTEM and discussed these results. X.W. and R.H. performed XPS measurements and discussed these results. The manuscript was corrected and improved by all authors. All authors have read and agreed to the published version of the manuscript.

Conflicts of Interest:
The authors declare no conflict of interest.