CuZr Metal Glass Powder as Electrocatalysts for Hydrogen and Oxygen Evolution Reactions

: For the practical application of water electrolysis, it is essential to develop cost-effective and high efﬁciency electrocatalysts for both hydrogen evolution reaction (HER) and oxygen evaluation reaction (OER). In this work, we applied CuZr metallic glass powder, after chemical dealloying treat-ment, as electrocatalysts. The as-prepared sample had both the increased speciﬁc area and optimized surface composition of an efﬁcient catalyst. During the HER and OER processes, the dealloyed CuZr sample displayed overpotential of 195 mV and 310 mV at current density of 10 mA cm − 2 , respectively. A two-electrode water splitting cell, using the as-prepared CuZr sample, exhibited high stability towards a high current density of 500 mA cm − 2 , and lower overpotential, compared to a Pt/C//IrO 2 cell, during the 10 mA cm − 2 constant current density aging test.


Introduction
The demand for renewable energy is urgent, due to the shortage of traditional energy sources, such as fossil fuels [1,2]. Various renewable energy sources have been developed, among which electrochemical water splitting is a high-efficiency and cost-effective method to produce hydrogen [3][4][5]. The water splitting reaction involves two reactions, i.e., hydrogen evolution reaction (HER) and oxygen evolution reaction (OER). As the HER and OER require high overpotentials to drive the reactions, various efficient electrocatalysts have been developed to boost the reaction efficiency [6][7][8][9][10]. Well-developed catalysts are usually costly, containing noble metals, such as Pt-based materials, Ru and Ir oxides. Thus, developing high efficiency electrocatalysts using cost-effective earth abundant elements is imperative for practical application.
Here in this work, CuZr powder was treated with different concentrations of HF solutions for the chemical dealloying process. The as-prepared CuZr samples displayed different surface morphologies and compositions. Among the CuZr samples, CuZr−3* had

Material Characterization
The morphology of the serial dealloyed samples, treated in different concentrations of HF solutions were observed by SEM. In Figure 1a of the CuZr−1 sample, the powders displayed 40−50 µm sized spherical morphology. After 0.1 M HF treatment, the surface cracked into 2−5 µm irregular flakes with 0.1−0.2 µm in between, adhering to the surface. In this case, the selective etching of CuZr by the HF solutions was minimal, and hardly met the expected effects. Figure 1b shows the CuZr−2 sample after 0.2 M HF treatment, the cracks on the surface of the spheres had almost disappeared. In contrast, the surface was covered with honeycomb-like holes form the local enlarged image of the spheres. Even through the surface area of CuZr−2 had increased, compared to that of CuZr−1, it was not obvious enough. rational design of high-efficiency metallic glass with non-noble metals is necessary for its practical application.
Here in this work, CuZr powder was treated with different concentrations of HF solutions for the chemical dealloying process. The as-prepared CuZr samples displayed different surface morphologies and compositions. Among the CuZr samples, CuZr−3* had the largest surface area with a petal-like morphology and Cu oxides covering the surface. This ensured that CuZr−3* presented the best HER/OER performance among the CuZr samples and even had more superior OER activity than IrO2. When coupled into the water electrolysis cell, the CuZr−3* cell showed good stability in the high current density and chronopotentiometry test.

Material Characterization
The morphology of the serial dealloyed samples, treated in different concentrations of HF solutions were observed by SEM. In Figure 1a of the CuZr−1 sample, the powders displayed 40−50 μm sized spherical morphology. After 0.1 M HF treatment, the surface cracked into 2−5 μm irregular flakes with 0.1−0.2 μm in between, adhering to the surface. In this case, the selective etching of CuZr by the HF solutions was minimal, and hardly met the expected effects. Figure 1b shows the CuZr−2 sample after 0.2 M HF treatment, the cracks on the surface of the spheres had almost disappeared. In contrast, the surface was covered with honeycomb-like holes form the local enlarged image of the spheres. Even through the surface area of CuZr−2 had increased, compared to that of CuZr−1, it was not obvious enough.  Figure 1c shows the CuZr−3* after 0.5 M HF treatment. A petal-like morphology with vertically aligned flakes had grown on the surface. Compared with the other two samples, the surface area of CuZr−3* had increased greatly. In this condition, the electrode/electrolyte contact area and exposure of reactive sites had increased, which was expected to enhance the electrocatalytic performance of CuZr−3*. In Figure 1d, when the HF concentration further increased to 1.0 M, the surface of CuZr−4 was etched and dissolved severely in the relatively concentrated 1 M HF solution. The spherical diameter of the CuZr−4 sample was obviously reduced, compared with the first three samples, and its surface was covered with loosely pulverized particles, which could easily dissolve after contact with the electrolyte. From the EDS mapping of CuZr−3*, the atomic ratio of copper to zirconium of the CuZr sample changed from the original 1:1 to 2:1 in the semiquantitative analysis. This indicated selective dissolution of element Zr during the dealloying process, which agreed with previous reports [25,26].
The chemical composition and crystal structure were further characterized by XRD and XPS. According to the XRD patterns of the CuZr samples, as seen in Figure 2a, they could roughly be indexed into Cu, Cu2O and CuO, but with different proportions. The pristine CuZr demonstrated the typical broad peak of metallic glass and sharp peaks  Figure 1c shows the CuZr−3* after 0.5 M HF treatment. A petal-like morphology with vertically aligned flakes had grown on the surface. Compared with the other two samples, the surface area of CuZr−3* had increased greatly. In this condition, the electrode/electrolyte contact area and exposure of reactive sites had increased, which was expected to enhance the electrocatalytic performance of CuZr−3*. In Figure 1d, when the HF concentration further increased to 1.0 M, the surface of CuZr−4 was etched and dissolved severely in the relatively concentrated 1 M HF solution. The spherical diameter of the CuZr−4 sample was obviously reduced, compared with the first three samples, and its surface was covered with loosely pulverized particles, which could easily dissolve after contact with the electrolyte. From the EDS mapping of CuZr−3*, the atomic ratio of copper to zirconium of the CuZr sample changed from the original 1:1 to 2:1 in the semi-quantitative analysis. This indicated selective dissolution of element Zr during the dealloying process, which agreed with previous reports [25,26].
The chemical composition and crystal structure were further characterized by XRD and XPS. According to the XRD patterns of the CuZr samples, as seen in Figure 2a intensified Cu and Cu peaks than the pristine CuZr sample. While for Zr 3d, in Figure  2c, the intensity of the Zr peaks was more reduced than at the pristine state, demonstrating the selective etching of the Zr element. The Zr peaks of CuZr at pristine state shifted towards higher oxidation states after dealloying to compensate for the charge imbalance. Therefore, the CuZr−3* sample, with its much larger specific surface covered by copper oxides CuO/Cu2O, on exposure of the Cu surface after dealloying, was expected to show high HER/OER catalytic activity.

HER/OER Performance and Water Splitting Reaction
The electrocatalytic activity towards HER was examined by the LSV curves in Figure  3a. The as-prepared CuZr samples were compared with that of Cu foil and Pt/C. For the CuZr samples, the CuZr−3* displayed higher HER catalytic activity than the CuZr−4, CuZr−2, CuZr−1 and pristine CuZr, as expected. This could be evidenced more clearly by the overpotential values at 10 mA cm −2 , which were 242 mV, 215 mV, 195 mV and 206 mV, from CuZr−1 to CuZr−4 ( Figure 3b). These CuZr samples were clearly much more active than the plain Cu foil with η10 of 496 mV, but they were hardly comparable with the commercial catalyst Pt/C, which had η10 only of 54 mV. The superb HER performance of the CuZr samples over the Cu foil was ascribed to the synergistic effect between copper zirconium glass alloy and copper oxides in improving catalytic activity [27]. According to the XPS spectra of Cu 2p in Figure 2b, the CuZr−3* exhibited more intensified Cu 2+ and Cu 1+ peaks than the pristine CuZr sample. While for Zr 3d, in Figure 2c, the intensity of the Zr peaks was more reduced than at the pristine state, demonstrating the selective etching of the Zr element. The Zr peaks of CuZr at pristine state shifted towards higher oxidation states after dealloying to compensate for the charge imbalance. Therefore, the CuZr−3* sample, with its much larger specific surface covered by copper oxides CuO/Cu 2 O, on exposure of the Cu surface after dealloying, was expected to show high HER/OER catalytic activity.

HER/OER Performance and Water Splitting Reaction
The electrocatalytic activity towards HER was examined by the LSV curves in Figure 3a. The as-prepared CuZr samples were compared with that of Cu foil and Pt/C. For the CuZr samples, the CuZr−3* displayed higher HER catalytic activity than the CuZr−4, CuZr−2, CuZr−1 and pristine CuZr, as expected. This could be evidenced more clearly by the overpotential values at 10 mA cm −2 , which were 242 mV, 215 mV, 195 mV and 206 mV, from CuZr−1 to CuZr−4 ( Figure 3b). These CuZr samples were clearly much more active than the plain Cu foil with η 10 of 496 mV, but they were hardly comparable with the commercial catalyst Pt/C, which had η 10 only of 54 mV. The superb HER performance of the CuZr samples over the Cu foil was ascribed to the synergistic effect between copper zirconium glass alloy and copper oxides in improving catalytic activity [27].
The corresponding Tafel slopes are plotted in Figure 3c. The smaller the slope, the faster the reaction kinetics [28][29][30]. In accordance with the LSVs, the CuZr−3* sample displayed a Tafel slope of 92.8 mV dec −1 , lower than the 139, 101, and 95.3 mV dec −1 of CuZr−1, CuZr−2 and CuZr−4, but higher than the 60.8 mV dec −1 of Pt/C. Among the CuZr samples, the high catalytic activity and fast reaction kinetics of CuZr−3* should be beneficial, given the much larger exposure area of the reactive sites and facilitated charge/mass transfer process. This could be verified by the electrochemical surface area (ECSA) measurements in Figure 3d,e. As shown in Figure 3f, the CuZr−3* sample with a double layer capacitance C dl value of 1.12 mF cm −2 had larger ECSAs than all the other CuZr samples. The reaction kinetics could be evidenced more clearly by the EIS measurement and corresponding equivalent circuit fitting. Figure 3g shows the Nyquist plots of the CuZr samples individually measured at their η 10 values. Through the equivalent fitting in Figure 3h, CuZr−3* displayed the lowest charge transfer resistance R ct of 16 Ω, lower than the 23, 35 and 41 Ω of CuZr−1, CuZr−2 and CuZr−4, respectively. This verified the enhanced reaction kinetics of CuZr−3*, as discussed in [31]. The electrochemical stability of the CuZr−3* was further examined by measuring the LSV curves before and after the accelerated aging 1000 cycles CV test. In Figure 3i, the aged LSV curve nearly overlaps with the initial one, indicating the superior structural and electrocatalytic stability of the as prepared CuZr−3* sample. The corresponding Tafel slopes are plotted in Figure 3c. The smaller the slope, the faster the reaction kinetics [28][29][30]. In accordance with the LSVs, the CuZr−3* sample displayed a Tafel slope of 92.8 mV dec −1 , lower than the 139, 101, and 95.3 mV dec −1 of CuZr−1, CuZr−2 and CuZr−4, but higher than the 60.8 mV dec −1 of Pt/C. Among the CuZr samples, the high catalytic activity and fast reaction kinetics of CuZr−3* should be beneficial, given the much larger exposure area of the reactive sites and facilitated charge/mass transfer process. This could be verified by the electrochemical surface area (ECSA) measurements in Figure 3d,e. As shown in Figure 3f, the CuZr−3* sample with a double layer capacitance Cdl value of 1.12 mF cm −2 had larger ECSAs than all the other CuZr samples. The reaction kinetics could be evidenced more clearly by the EIS measurement and corresponding equivalent circuit fitting. Figure 3g shows the Nyquist plots of the CuZr samples individually measured at their η10 values. Through the equivalent fitting in Figure 3h, CuZr−3* displayed the lowest charge transfer resistance Rct of 16 Ω, lower than the 23, 35 and 41 Ω of CuZr−1, CuZr−2 and CuZr−4, respectively. This verified the enhanced reaction kinetics of CuZr−3*, as discussed in [31]. The electrochemical stability of the CuZr−3* was further examined by measuring the LSV curves before and after the accelerated aging 1000 cycles CV test. In Figure 3i, the aged LSV curve nearly overlaps with the initial one, indicating the superior structural and electrocatalytic stability of the as prepared CuZr−3* sample. Figure 4 shows the OER performance of the CuZr samples in comparison with Cu foil, Ni foam and IrO2. It was interesting to find that the serial CuZr displayed even better OER activity than the IrO2 from the LSV curves in Figure 4a.  Figure 4 shows the OER performance of the CuZr samples in comparison with Cu foil, Ni foam and IrO 2 . It was interesting to find that the serial CuZr displayed even better OER activity than the IrO 2 from the LSV curves in Figure 4a.
According to their overpotentials at 10 mA cm −2 (Figure 4b), the CuZr−1, CuZr−2 and CuZr−3 exhibited η 10 values of 340 mV, 350 mV and 310 mV, lower than the 374 mV of IrO 2 , 380 mV of CuZr−4, and 538 mV of Cu foil. The superb performance of CuZr−1 to CuZr−3 indicated the positive role of CuZr alloy in supporting surface Cu oxides in boosting the OER catalytic activity [7,32,33]. It was noticed that there was an oxidation peak between 1.2 and 1.5 V from the LSV curve of each sample, except that of Cu foil. This was because the oxidation of foam nickel was exposed to alkaline electrolyte when the ink did not cover its whole surface [34]. The bare Ni foam sample wass also tested for reference. In addition, the Tafel slopes and impedance spectra were used to analyze the reaction kinetics. In Figure 4c, the CuZr samples showed low Tafel slope values of 55.4 (CuZr−3*), 80.9 (CuZr−1), 82.8 (CuZr−2) and 88.6 mVdec −1 (CuZr−4), lower than that of IrO 2 , Ni foam or Cu foil.
The lowest Tafel slope of CuZr−3* verified its excellent surface morphology in increasing the reactive area and improving charge transfer. According to Figure 4d-f, the ECSA value derived from the C dl was 220 cm 2 for the CuZr−3* sample, much larger than any of the CuZr samples. The Nyquist plots of the four CuZr samples, and their fitted values of R ct , are shown in Figure 4g,h. In accordance with the Tafel slope, CuZr−3* had the lowest R ct of 8.6 Ω, lower than the 12.5 Ω of CuZr−1, 12.9 Ω of CuZr−2 and 16.7 Ω of CuZr−4. This indicated that, with identical surface composition, the higher the surface exposure area, the faster the reaction rate. Furthermore, the accelerated aging tests result is shown in Figure 4i. After 1000 CV cycles, the OER activity of the CuZr−3* sample became even better, which could be because of the partial dissolution of zirconium oxide in the alkaline solution further increasing the reactive surface area.  [7,32,33]. It was noticed that there was an oxidation peak between 1.2 and 1.5 V from the LSV curve of each sample, except that of Cu foil. This was because the oxidation of foam nickel was exposed to alkaline electrolyte when the ink did not cover its whole surface [34]. The bare Ni foam sample wass also tested for reference. In addition, the Tafel slopes and impedance spectra were used to analyze the reaction kinetics. In Figure 4c, the CuZr samples showed low Tafel slope values of 55.4 (CuZr−3*), 80.9 (CuZr−1), 82.8 (CuZr−2) and 88.6 mVdec −1 (CuZr−4), lower than that of IrO2, Ni foam or Cu foil.
The lowest Tafel slope of CuZr−3* verified its excellent surface morphology in increasing the reactive area and improving charge transfer. According to Figure 4d-f, the ECSA value derived from the Cdl was 220 cm 2 for the CuZr−3* sample, much larger than any of the CuZr samples. The Nyquist plots of the four CuZr samples, and their fitted values of Rct, are shown in Figure 4g,h. In accordance with the Tafel slope, CuZr−3* had the lowest Rct of 8.6 Ω, lower than the 12.5 Ω of CuZr−1, 12.9 Ω of CuZr−2 and 16.7 Ω of CuZr−4. This indicated that, with identical surface composition, the higher the surface exposure area, the faster the reaction rate. Furthermore, the accelerated aging tests result is shown in Figure 4i. After 1000 CV cycles, the OER activity of the CuZr−3* sample became even better, which could be because of the partial dissolution of zirconium oxide in the alkaline solution further increasing the reactive surface area. To further examine the possibility for practical application, CuZr−3* electrodes, as cathode and anode, were assembled into the two-electrode configuration (inset of Figure 5a) for water electrocatalytic tests. The CuZr−3* sample was selected, as it best displayed HER and OER performance among the CuZr samples. According to Figure 5a, the CuZr−3*//CuZr−3* cell exhibited inferior catalytic activity than the Pt/C//IrO 2 cell but was obviously superior to the CuZr−3*//Ni foam cell. At a current density of 50 mA cm −2 , the overpotential of the Pt/C//IrO 2 cell was 530 mV, slightly lower than the 580 mV of the CuZr−3* cell. It is worth noticing that the CuZr−3* could sustain an ultra-high current density of 500 mA cm −2 , which demonstrated the structural and compositional stability of the CuZr−3* sample.
Additionally, a stability test was conducted through chronopotentiometry curves at constant current density of 10 mA cm −2 , shown in Figure 5b. The superiority of the CuZr−3* was demonstrated in this condition. The CuZr−3* cell displayed overpotential of 680 mV lasting for the more than 18 h, in contrast to the 820 mV of Pt/C//IrO 2 cell, even though its initial overpotential was much lower.
CuZr−3*//CuZr−3* cell exhibited inferior catalytic activity than the Pt/C//IrO2 cell but was obviously superior to the CuZr−3*//Ni foam cell. At a current density of 50 mA cm −2 , the overpotential of the Pt/C//IrO2 cell was 530 mV, slightly lower than the 580 mV of the CuZr−3* cell. It is worth noticing that the CuZr−3* could sustain an ultra-high current density of 500 mA cm −2 , which demonstrated the structural and compositional stability of the CuZr−3* sample. Additionally, a stability test was conducted through chronopotentiometry curves at constant current density of 10 mA cm −2 , shown in Figure 5b. The superiority of the CuZr−3* was demonstrated in this condition. The CuZr−3* cell displayed overpotential of 680 mV lasting for the more than 18 h, in contrast to the 820 mV of Pt/C//IrO2 cell, even though its initial overpotential was much lower.
The raw CuZr MG powder was chemically dealloyed in a series of HF solutions, i.e., 0.1 mol L −1 , 0.2 mol L −1 , 0.5 mol L −1 and 1.0 mol L −1 , in the PTFE container with 6 mL HF solution added. To ensure sufficient reaction, the powder contained HF solution was magnetic stirred for 24 h at room temperature. Afterwards, the dealloyed CuZr powder was collected, rinsed with deionized water three times, and centrifuged at 10,000 r min −1 for 5 min. Then the CuZr powder was put in the refrigerator for 6 h and dried with a freeze dryer. The as-prepared samples were named CuZr−1, CuZr−2, CuZr−3* and CuZr−4, corresponding to the four dealloying electrolytes.

Material Characterization
The amorphous structure of the MG samples was examined by X-ray diffractometer (XRD, Rigaku Corporation, Tokyo, Japan), using Cu Kα as the radiation source at 40 kV and 40 mA. The XRD pattern was obtained by scanning from 20° to 80° at a scan speed of
The raw CuZr MG powder was chemically dealloyed in a series of HF solutions, i.e., 0.1 mol L −1 , 0.2 mol L −1 , 0.5 mol L −1 and 1.0 mol L −1 , in the PTFE container with 6 mL HF solution added. To ensure sufficient reaction, the powder contained HF solution was magnetic stirred for 24 h at room temperature. Afterwards, the dealloyed CuZr powder was collected, rinsed with deionized water three times, and centrifuged at 10,000 r min −1 for 5 min. Then the CuZr powder was put in the refrigerator for 6 h and dried with a freeze dryer. The as-prepared samples were named CuZr−1, CuZr−2, CuZr−3* and CuZr−4, corresponding to the four dealloying electrolytes.

Material Characterization
The amorphous structure of the MG samples was examined by X-ray diffractometer (XRD, Rigaku Corporation, Tokyo, Japan), using Cu K α as the radiation source at 40 kV and 40 mA. The XRD pattern was obtained by scanning from 20 • to 80 • at a scan speed of 5-degree min −1 . The morphology and element mapping were analyzed by a scanning electron microscope (SEM, Thermo Scientific, Waltham, MA, USA) and an energy dispersive spectrometer (EDS).

Electrochemical Measurement
All electrochemical tests were carried out using electrochemical workstation of Chenhua (CHI760E) at room temperature of 25 • C. Three electrode and two electrode configurations were applied for HER/OER and water splitting tests, respectively. To prepare the working electrode of HER/OER, 5 mg of the dealloyed CuZr powder and 30 µL Nafion as the binder were dissolved in the solvent containing 700 µL deionized water and 300 µL isopropyl alcohol. Then, the obtained solution was stirred and ultrasonically treated for 30 min to prepare the electrode ink. Afterwards, 40 µL of the ink was dropped on the nickel foam (0.5 cm × 1 cm) evenly, displaying mass loading of 0.38 mg cm −2 , and dried under an infrared lamp for 20 min prior to use. The Pt/C and IrO 2 electrode preparation was the same as in our previous report [15]. For the copper electrode, the copper foil was cut into an area of 0.5 cm × 1 cm and applied as working electrode for comparison. For the three-electrode cell of HER, platinum foil (1.5 cm × 1.5 cm) was the counter electrode and Hg/HgO was used as the reference electrode. The measured potential was converted to the reversible hydrogen electrode (RHE) for standard comparison via the following equation: R RHE = E Hg/HgO + 0.926 V. All the electrochemical measurements were carried out in an alkaline solution of 1 M KOH.
During the HER measurements, the electrode was pre-activated and scanned by the cyclic voltammetry (CV) method at a scan rate of 100 mV s −1 until the CV curves were stable. Afterwards, the linear sweep voltammetry (LSV) test was performed in the potential range of −0.8 V to 0.0 V at a scan rate of 5 mV s −1 . The accelerated durability test (ADT) was carried out by CV scans between −0.8 V and 0.0 V for 1000 cycles at 100 mV s −1 . For OER, the LSV and ADT measurements followed the same procedure as HER but in the voltage range of 1.2 V to 2.0 V. The electrochemical impedance spectroscopy (EIS) was measured at an overpotential of each cell at 10 mA cm −2 in the frequency range of 100 kHz to 0.1 Hz and voltage amplitude of 5 mV. For the overall water splitting reaction, the LSV test was performed from 1.2 V to 2.1 V at a scanning speed of 5 mV s −1 . The stability test was carried out by chronopotentiometry at a constant current density of 10 mA cm −2 for 18 h.

Conclusions
The powder of CuZr metallic glass was chemically dealloyed in HF solution and applied as an electrocatalyst for HER, OER and water splitting reactions. The CuZr sample, dealloyed in 0.5 mol L −1 solution for 24 h, displayed an enlarged surface area with petallike morphology and copper oxide coverage. It displayed inferior HER performance than the commonly used Pt/C but superior OER activity than the commercial IrO 2 . This work presents a simple and effective method to activate low cost CuZr MG powder for practical application in electrocatalysis.