CuxO-Modified Nanoporous Cu Foil as a Self-Supporting Electrode for Supercapacitor and Oxygen Evolution Reaction

Designing and modifying nanoporous metal foils to make them suitable for supercapacitor and catalysis is significant but challenging. In this work, CuxO nanoflakes have been successfully in situ grown on nanoporous Cu foil via a facile electrooxidation method. A Ga-assisted surface Ga-Cu alloying–dealloying is adopted to realize the formation of a nanoporous Cu layer on the flexible Cu foil. The following electrooxidation, at a constant potential, modifies the nanoporous Cu layer with CuxO nanoflakes. The optimum CuxO/Cu electrode (O-Cu-2h) delivers the maximum areal capacitance of 0.745 F cm−2 (410.27 F g−1) at 0.2 mA cm−2 and maintains 94.71% of the capacitance after 12,000 cycles. The supercapacitor consisted of the O-Cu-2h as the positive electrode and activated carbon as the negative electrode has an energy density of 24.20 Wh kg−1 and power density of 0.65 kW kg−1. The potential of using the electrode as oxygen evolution reaction catalysts is also investigated. The overpotential of O-Cu-2h at 10 mA cm−2 is 394 mV; however, the long-term stability still needs further improvement.


Introduction
Though dealloying has gained fruitful achievements during the past decades, the exploration of applying dealloyed materials in catalysts, sensors, and energy-storage devices never stops [1][2][3][4]. Dealloying is a top-down strategy that dissolves the active elements from an alloy and generates a nanostructure (mostly nanoporous structure) after the diffusion/reorganization of the less active elements [5,6]. Nanoporous noble metals, e.g., Au, Ag, and Cu, are among the most common materials created by dealloying. Nanoporous Au has been a prototype to study the mechanisms of dealloying, and it is widely applied as catalysts for fuel cells, methane pyrolysis, supercapacitors, actuator, and solar steam generation, etc. [4,[7][8][9][10][11]. Nanoporous Cu has also been extensively studied, due to the relatively low price and great potential applications.
The study of dealloying Cu-based alloys can be dated back to the 1980s, when Cu-Mn alloys were dealloyed to probe into the morphology and composition of dealloyed products [12]. Raney ® Cu, produced by dealloying Al 2 Cu in a NaOH solution, is a successful catalyst for methanol production from synthesis gas [13]. Similar to Raney ® Cu, most of the dealloyed Cu materials are powdery. In addition to powdery nanoporous Cu, a self-supporting nanoporous Cu film is necessary for some application scenarios, e.g., electrodes of supercapacitors and lithium-ion batteries [14][15][16]. Nanoporous metal films can be fabricated through dealloying metallic glass ribbons, which are flexible due to the amorphous structure [17][18][19]. Another approach is to create an alloy layer on the metal film, followed by surface dealloying to generate a nanoporous layer. For instance, Diao et al. used a co-sputtering and dealloying method to fabricate bi-continuous and flexible nanoporous Cu films integrated onto the Cu foil [20]. In a previous study, the Ga-Cu

Preparation of Cu x O/Cu Foils
The commercial Cu foils (50 µm thick, 99.9 wt.%, Kejing, Shenzhen, China) were cleaned with deionized water and alcohol before painting. A certain amount of Ga ingot (Macklin, 99.99 wt.%, Macklin, Shanghai, China) was placed on a hotplate at 50 • C to guarantee the liquid state. The cleaned Cu foils were fixed on another hotplate (maintained at 40 • C) during the painting process. Liquid Ga was brushed onto the Cu foil using a paintbrush, then annealed at 150 • C for 8 h in a vacuum oven to form Ga-Cu alloys via reaction-diffusion. The annealed foils were dealloyed in a 0.2 M HNO 3 solution at 40 • C, until no bubbles emerged. The dealloyed foils were preserved in a vacuum tank at room temperature. The electrooxidation process was conducted using a two-electrode system containing a dealloyed Cu foil and graphite rod. The dealloyed foils were electro-oxidized in a 1 M KOH solution using a two-electrode system at 1 V for 15 min, 30 min, 1 h, 2 h, and 5 h, respectively, to form a copper oxide layer on the foil surface. The obtained samples were named O-Cu-Xm or O-Cu-Xh, where X referred to the electrooxidation time; 'm' is short for 'minute', and 'h' is short for 'hour'. For instance, O-Cu-15m refers to the sample that was electro-oxidized for 15 min.

Fabrication of Asymmetric Supercapacitor Devices
In the asymmetric supercapacitor device, the Cu x O/Cu foil was used as the positive electrode, and the active carbon (AC) powders deposited on the Ni foams acted as the negative electrode. The fabrication details of the negative electrode can be seen in ref. [27]. The positive and negative electrodes were positioned face-to-face and separated by cellulose paper in a 1 M KOH electrolyte.

Characterizations of Catalysts
The phase compositions of the samples were characterized by X-ray diffraction (XRD, X'pert Pro, Malvern Panalytical, Almelo, Netherlands). The microstructural features were investigated by a field emission scanning electron microscope (SEM, Zeiss Sigma 500, Birmingham, UK) and transmission electron microscope (TEM, JEM F200, JEOL, Tokyo, Japan). The surface status of the samples was analyzed by X-ray photoelectron spectroscopy (XPS, K-Alpha, Thermo Scientific, Loughborough, UK). Raman spectra analyses were performed on a HORIBA LabRAM HR Evolution (HORIBA, Kyoto, Japan), using a laser source of 488 nm. The concentration of Cu ions in the electrolyte after OER tests was determined via inductively coupled plasma-optical emission spectrometry (ICP-OES, Agilent 5110, Santa Clara, CA, USA).

Electrochemical Measurements
The electrochemical performance of the electrodes and devices was analyzed on an electrochemical workstation (CHI660E, Shanghai, China) in a 1 M KOH electrolyte. A Pt electrode, Ag/AgCl electrode, and as-synthesized electrodes were used as the counter, reference, and working electrodes, respectively, in a three-electrode system. The cyclic voltammetry (CV) was operated under various scan rates from 5 to 100 mV s −1 , between 0−0.6 V vs. Ag/AgCl. The galvanic charging and discharging (GCD) was conducted under the current densities of 0.2 to 1 mA cm −2 , between 0−0.5 V vs. Ag/AgCl. The electrochemical impedance spectroscopy (EIS) was obtained at open circuit potential in the frequency range from 0.01 Hz to 1 MHz with a 5 mV amplitude. The O-Cu-2h//AC asymmetric supercapacitor was tested via CV under various scan rates of 5, 10, 30, 50, 70, and 100 mV s −1 , between 0.6−1.6 V, as well as GCD under current densities of 2, 3, 4, 5, 6 and 7 mA cm −2 , between 0.6−1.6 V. The calculation details of areal capacitance, specific capacitance, and energy and power densities can be found in references [29,30]. To estimate the mass of Cu x O on the surface, we measured the weight of the dealloyed Cu foils before and after oxidation. To obtain accurate mass loading, at least three foils were weighted for each oxidation time.
Linear sweep voltammetry (LSV) was used to verify the possibility of using the electrooxidized foils as OER catalysts. LSV tests were carried out in an O 2 -saturated 1 M KOH electrolyte, with a scanning rate of 5 mV s −1 at the potential range of 0.1 to 1.1 V vs. RHE. All LSV curves were iR-compensated to exclude the influence of solution resistance. Electrochemical impedance spectroscopy (EIS) was conducted in the same electrolyte from 1 MHz to 0.01 Hz at 0.7 V vs. Hg/HgO, with an amplitude of 5 mV. All the potentials were converted to the reversible hydrogen electrode (RHE) reference, according to E (vs. RHE) = E (vs. Hg/HgO) + 0.0591pH + 0.098.

Results and Discussion
The Composition and Microstructure Analysis A previous study reports that dealloying Ga-Cu alloy in an acidic solution resulted in the formation of nanoporous Cu [21]. In that report, the Ga-Cu alloy layer was formed via annealing at 100 • C for 5 h and generated a nanoporous structure after dealloying. Here, we fabricated a nanoporous structure on the surface of Cu film using the same strategy, followed by electrooxidation in an alkaline solution, the detailed procedure of which can be seen in Figure 1. After painting the Ga and annealing, the Ga 2 Cu phase (PDF #25-0275) formed on the surface of Cu foil (Figure 2a). The back-scattered SEM image shows the grain boundaries of the annealed Ga 2 Cu layer ( Figure S1a). EDX elemental mapping shows the distribution of the Ga and Cu elements ( Figure S1b-d). The EDX spectrum indicates that the Ga atomic concentration is 76.64 at.%, with that of Cu being 23.36 at.%, indicating that Ga was excessive on the surface after 8 h annealing at 150 • C ( Figure S1e). The Ga 2 Cu layer can be fully dealloyed in a 0.2 M HNO 3 solution with only Cu retained, as confirmed by the XRD pattern in Figure 2a. An electrooxidation method was applied to generate a Cu oxide layer on the dealloyed Cu surface. It has been reported that Cu 2 O nanoneedles could be grown on the etched Cu foam via CV cycling at -0.3-0.3 V vs. Ag/AgCl in the 1 M KOH solution [27]. Further oxidation at 0-0.6 V could transform Cu 2 O to CuO [31]. In another study, a Cu 2 O/CuO hybrid layer was formed on a nanoporous Cu ribbon by anodizing at the constant current densities in the 0.5 M KOH solution [23]. Inspired by the above method, we electro-oxidized the dealloyed Cu foil at a constant potential of 1 V in a two-electrode system, expecting to create a Cu x O layer on nanoporous Cu. Extra peaks are observed in the XRD patterns after electrooxidation for 15 min, 30 min, 1 h, 2 h, and 5 h. As shown in Figure 2b, a peak that appeared at 2θ 36.5 • can be indexed to Cu 2 O (PDF #05-0667) [32]. With the electrooxidation time increasing to longer than 30 min, the relative intensity of the Cu 2 O peaks decreases. It shows almost only the CuO phase (PDF #48-1548), when the electrooxidation time is 5 h. The Raman spectra in Figure 2c demonstrate the presence of peaks at~280, 330, and 620 cm −1 , confirming the generation of CuO after electrooxidation [33,34]. Extra peaks can be found at 148 and 217 cm −1 for sample O-Cu-15m, ascribed to the characteristic Raman bands of Cu 2 O [35]. To further probe into the surface compositions of the Cu films, we conducted XPS on samples electro- ribbon by anodizing at the constant current densities in the 0.5 M KOH solution [23]. Inspired by the above method, we electro-oxidized the dealloyed Cu foil at a constant potential of 1 V in a two-electrode system, expecting to create a CuxO layer on nanoporous Cu. Extra peaks are observed in the XRD patterns after electrooxidation for 15 min, 30 min, 1 h, 2 h, and 5 h. As shown in Figure 2b, a peak that appeared at 2θ 36.5° can be indexed to Cu2O (PDF #05-0667) [32]. With the electrooxidation time increasing to longer than 30 min, the relative intensity of the Cu2O peaks decreases. It shows almost only the CuO phase (PDF #48-1548), when the electrooxidation time is 5 h. The Raman spectra in Figure 2c demonstrate the presence of peaks at ~280, 330, and 620 cm -1 , confirming the generation of CuO after electrooxidation [33,34]. Extra peaks can be found at 148 and 217 cm -1 for sample O-Cu-15m, ascribed to the characteristic Raman bands of Cu2O [35].   After dealloying in 0.2 M HNO 3 solution, nanoporous structures evolved on the Cu foil surface, as shown in Figure 3a,b. The grain boundaries can be observed on the surface of the dealloyed Cu foil, which inherits from the Ga-Gu alloys, as shown from the back-scattering SEM image ( Figure S1a). The diameter of the dealloyed Cu ligament is~200 nm, and no Ga reserves after dealloying ( Figure S3). After electrooxidation in a 1 M KOH solution at 1.0 V vs. Ag/AgCl for 15 min, two distinct microstructures can be observed on the surface: nanobelts and nanoflakes. The EDX obtained from the rectangle area in Figure S4a (Figure 3e-g). At the initial oxidation stage, the nanoneedle Cu 2 O phase formed on the surface [27]. As the oxidation time increased, the nanoneedle or nanobelt became longer but less stable. Therefore, after 2 or 5 h oxidation, the nanobelts would grow into interlinked nanoflakes to make the structure more stable; the morphology can be observed in some CuO-based materials [23,36]. The atomic ratio of Cu to O from O-Cu-1h and O-Cu-2h keeps approximately 1, indicating that the dominated phase might be CuO (Figure 3i and Figure S6). After 5 h oxidation, the atomic ratio of Cu to O decreases to 0.73, which might suggest that the Cu was oxidized to higher valence after long-term oxidation ( Figure S7). Moreover, a typical cross-section view of the oxidized foil displays that the oxidation layer is~9.5 µm (Figure 3h). The cross-section view proves that the nanoflakes grow across the entire dealloyed layer. After dealloying in 0.2 M HNO3 solution, nanoporous structures evolved on the Cu foil surface, as shown in Figure 3a,b. The grain boundaries can be observed on the surface of the dealloyed Cu foil, which inherits from the Ga-Gu alloys, as shown from the backscattering SEM image ( Figure S1a). The diameter of the dealloyed Cu ligament is ~200 nm, and no Ga reserves after dealloying ( Figure S3). After electrooxidation in a 1 M KOH solution at 1.0 V vs. Ag/AgCl for 15 min, two distinct microstructures can be observed on the surface: nanobelts and nanoflakes. The EDX obtained from the rectangle area in Figure  S4a Figure S5), respectively, both smaller than that from O-Cu-15m. The microstructures of O-Cu-1h and O-Cu-2h are almost the same, while the nanoflakes seem thinner after 5 h oxidation (Figure 3e-g). At the initial oxidation stage, the nanoneedle Cu2O phase formed on the surface [27]. As the oxidation time increased, the nanoneedle or nanobelt became longer but less stable. Therefore, after 2 or 5 h oxidation, the nanobelts would grow into interlinked nanoflakes to make the structure more stable; the morphology can be observed in some CuO-based materials [23,36]. The atomic ratio of Cu to O from O-Cu-1h and O-Cu-2h keeps approximately 1, indicating that the dominated phase might be CuO (Figures  3i and S6). After 5 h oxidation, the atomic ratio of Cu to O decreases to 0.73, which might suggest that the Cu was oxidized to higher valence after long-term oxidation ( Figure S7). Moreover, a typical cross-section view of the oxidized foil displays that the oxidation layer 36 39   (Figure 4f). The HRTEM analyses prove the formation of CuO and Cu 2 O after electrooxidation, which agrees with the XRD results. In a previous study, pure CuO could be fabricated by anodizing the dealloyed nanoporous Cu at 15 mA cm −2 and then calcining at 200 • C for 2 h [18]. This work also obtained the mixture of CuO and Cu 2 O after one-step electrooxidation at a constant potential. The XPS results prove that more CuO generated with a longer oxidation time. The electrode is nominated as Cu x O/Cu electrode in the following section.
Cu 2 O or CuO is an excellent candidate for supercapacitors [37][38][39]. In this work, the Cu x O layer was grown on the flexible Cu foil, which could be potentially applied as a self-standing supercapacitor electrode. The performance of the as-synthesized Cu x O/Cu as a supercapacitor electrode was evaluated in a three-electrode cell containing a 1 M KOH electrolyte. Figure 5a compares the CVs of the oxidized and dealloyed samples recorded in the potential ranging from 0-0.6 V vs. Ag/AgCl at a 5 mV s −1 scan rate. The five electro-oxidized samples show almost the same CV curve shapes, indicating a similar reaction mechanism, which has been found in some other Cu x O-based electrodes [25,36]. The CV shape is different from those with an ideal rectangular shape generated from the electric double-layer capacitance [9,27]. O-Cu-2h has the largest integrated CV area, suggesting the most significant specific capacitance. In contrast, the pristine dealloyed Cu foil shows almost no capacitance at a 5 mV s −1 scan rate. Figure 5b displays the GCD curves of the samples oxidized for different times at a charge/discharge current density of 0.4 mA cm −2 , and the curves also show a quasi-linear shape, suggesting the pseudocapacitive property [40]. It can be concluded from CV and GCD curves that O-Cu-2h has the maximum capacitance. Figure 5c presents the CV curves of O-Cu-2h recorded at scan rates ranging from 0-100 mV s −1 , indicating a positive correlation between the integrated CV area and scan rate. The CV curves of the other four samples are displayed in Figure S8, which show a similar variation trend with O-Cu-2h. The areal capacitance calculated from the CV curves is exhibited in Figure 5d Figure S9, showing the same charge/discharge features with O-Cu-2h. Figure 5f exhibits that O-Cu-2h has areal capacitance values of 0.745, 0.709, 0.691, 0.680, and 0.676 F cm −2 at 0.2 to 1.0 mA cm −2 current densities. The areal capacitance values calculated from GCD have the same performance sequence as those calculated from CV results. It is worth noting that O-Cu-5h has low capacitance, compared to O-Cu-2h. After oxidation for 5 h, almost all Cu 2 O transformed to CuO. Therefore, the pseudocapacitance that originated from the reversible reaction between Cu 2 O and CuO became less efficient. Moreover, the thicker oxide layer of O-Cu-5h may block the ionic transfer during the charge/discharge, resulting in lower capacitance. The mass loading of Cu x O was estimated via weighing the foil before and after electrooxidation, shown in Figure S10. It can be seen that the mass loading increased with longer electrooxidation times, namely increased from 1.13 to 1.82 of O-Cu-15m, and further to 4.15 mg cm −2 after 5 h oxidation. It should be mentioned that after 5 h oxidation, the oxide layer tends to be exfoliated from the foil during the operation. The specific capacitance normalized to the mass of Cu x O versus the scan rate, and current density is shown in Figure 5g Figure S11. The internal resistance of O-Cu-15m is smaller than that of O-Cu-2h, indicating the thicker oxide layer with a longer oxidation duration. In the low frequency region, the steeper slope of O-Cu-2h reveals faster ionic diffusion between the oxide layer and electrolyte. The cycling stability of O-Cu-2h was evaluated via CV cycling at 100 mV s −1 for 12,000 cycles. The areal capacitance decreased from the initial value of 0.171 to 0.162 F cm −2 after cycling, retaining 94.71% of the initial capacitance (Figure 5i). The electrode has unchanged phase compositions after cycling, and the Cu x O flakes become thicker than the initial structure ( Figure S12 and inset of Figure 5i). We compared the specific capacitance of Cu x O-based electrodes, as shown in Table S1 [25][26][27]36,38,39,[41][42][43][44][45], and the results reveal that our electrode is comparable to the pure Cu x O material but inferior, compared to the Cu x O-based composites. Further investigations might be carried out to fabricate hybrid electrodes based on our materials. is ~9.5 μm (Figure 3h). The cross-section view proves that the nanoflakes grow across the entire dealloyed layer.  (Figure 4f). The HRTEM analyses prove the formation of CuO and Cu2O after electrooxidation, which agrees with the XRD results. In a previous study, pure CuO could be fabricated by anodizing the dealloyed nanoporous Cu at 15 mA cm -2 and then calcining at 200 °C for 2 h [18]. This work also obtained the mixture of CuO and Cu2O after one-step electrooxidation at a constant potential. The XPS results prove that more CuO generated with a longer oxidation time. The electrode is nominated as CuxO/Cu electrode in the following section.  The electrochemical properties of AC under various scan rates and current densities are present in Figure S13. The CV curves of the AC electrode show no redox peaks, and the symmetrical GCD curves have the typical characteristics of electric double-layer capacitance [27]. The asymmetric supercapacitor constructed with O-Cu-2h foil and AC decorated Ni foam was investigated, in order to prove its potential in energy storage devices. Figure 6a presents the supercapacitor's CV curves at scan rates ranging from 5 to 100 mV s -1 and a potential window of 0.6-1.6 V. The areal capacitance calculated from the CV curves is shown in Figure 6b. The areal capacitance reaches 0.53 F cm -2 at the 5 mV s -1 scan rate and decreases to 0.30 F cm -2 with a 100 mV s -1 scan rate. GCD curves in Figure 6c show an IR drop at the initial state of the discharging stage, which is an indicator of the internal resistance of the device [46]. The GCD curves present a non-linear feature, suggesting the faradaic process [40]. The areal capacitance under different current densities is calculated and shown in Figure 6d. The maximum areal capacitance is obtained at 2 mA The electrochemical properties of AC under various scan rates and current densities are present in Figure S13. The CV curves of the AC electrode show no redox peaks, and the symmetrical GCD curves have the typical characteristics of electric double-layer capacitance [27]. The asymmetric supercapacitor constructed with O-Cu-2h foil and AC decorated Ni foam was investigated, in order to prove its potential in energy storage devices. Figure 6a presents the supercapacitor's CV curves at scan rates ranging from 5 to 100 mV s −1 and a potential window of 0.6-1.6 V. The areal capacitance calculated from the CV curves is shown in Figure 6b. The areal capacitance reaches 0.53 F cm −2 at the 5 mV s −1 scan rate and decreases to 0.30 F cm −2 with a 100 mV s −1 scan rate. GCD curves in Figure 6c show an IR drop at the initial state of the discharging stage, which is an indicator of the internal resistance of the device [46]. The GCD curves present a nonlinear feature, suggesting the faradaic process [40]. The areal capacitance under different current densities is calculated and shown in Figure 6d. The maximum areal capacitance is obtained at 2 mA cm −2 , reaching 0.60 F cm −2 . The areal capacitance decreases to 0.52 F cm −2 when the current density increases to 7 mA cm −2 . The energy and power Nanomaterials 2022, 12, 2121 9 of 14 densities are essential for evaluating the performance of an energy storage device. Figure 6e shows the Ragone plots (areal and mass) of the asymmetric supercapacitor calculated from the GCD curves. The energy densities of our device are in the range of 20.86 to 24.20 Wh kg −1 (0.26 to 0.30 Wh cm −2 ), corresponding to the power densities of 2.14 to 0.65 kW kg −1 (26.49 to 8.08 W cm −2 ). The energy densities are competitive with some other Cu x O-based supercapacitors (Table S2) [26,27,36,38,39,[43][44][45]. For instance, a 3D Cu 2 O@Cu nanoneedle arrays electrode had an energy density of 26.0 Wh kg −1 at power density of 1.8 kW kg −1 [27]. An all-solid-state supercapacitor using 3D nanostructured Cu x O-modified Cu foam, as an electrode showed an energy density of 25 µWh cm −2 when the power density was 3 mW cm −2 [36]. An all-solid-state asymmetric supercapacitor with Cu x O NWs NWs@CoS 2 , as the electrode delivered an energy density of 49.8 Wh kg −1 at a power density of 0.226 kW kg −1 [47]. Unfortunately, the areal capacitance decreased gradually from 0.357 to 0.227 F cm −2 , only retaining 63.59% of the initial capacitance after 10,000 cycles at 100 mV s −1 (Figure 6f). More modifications of the Cu x O/Cu foil should be explored, in order to improve the stability of the device.
There have been reports that CuO nanosheet bundles could be applied as efficient oxygen evolution reaction (OER) electrocatalysts [48]. In this study, we also investigate the OER property of the as-prepared samples. LSV curves of the pristine Cu foil, dealloyed Cu foil, and five oxidized samples are shown in Figure 7a. It can be seen that O-Cu-30m, O-Cu-1h, O-Cu-2h, and O-Cu-5h have comparable OER performance. The overpotential of O-Cu-2h and O-Cu-5h at 10 mA cm −2 is 394 mV, slightly lower than 401 and 414 mV of O-Cu-1h and O-Cu-30m, respectively. O-Cu-15m has the largest overpotential of 449 mV, smaller than dealloyed Cu and pristine Cu foil (Figure 7b). The overpotential is higher than the reported benchmark RuO 2 (388 mV) [49]. The Tafel slope of all samples was around 140 mV dec −1 , indicating the inferior catalytic kinetics (Figure 7c). Moreover, a Tafel slope higher than 120 mV dec −1 might indicate that the first step (M + OH − → MOH) of OER in an alkaline solution is the rate-determining step [50]. The long-term stability was evaluated via a potentiostat test. The i-t curve in Figure 7d shows a rapid decline in the current density from the very beginning of the test, with an 85.3% decrease before it reaches a steady state. Although the phase compositions showed almost no change after the stability test (inset of Figure 7d), the surface Cu x O aggregated together to form thick flakes ( Figure 7e). Moreover, the atomic ratio of Cu to O became 42.62 to 57.38%, appropriate to that of O-Cu-5h, indicating the continuous oxidation of the surface (Figure 7f). We noticed that the electrolyte gradually became light blue during the test. In order to observe the dissolution of Cu ions into the electrolyte, we also used a Pt plate counter electrode to conduct the i-t test. After the test, the surface of the Pt electrode clearly showed blue deposition, which could be deduced to Cu species ( Figure S14). The ICP result confirmed 8.23 mg L −1 Cu ions in the electrolyte after the 9 h long-term test ( Figure S14). In a previous study, a CuO nanowire@Co 3 O 4 composite showed an overpotential of 258 mV at a current density of 10 mA cm −2 and Tafel slope of 72 mV dec −1 [51]. A 2D CuO on a stainless-steel substrate had an overpotential of 350 mV, with a Tafel slope of 59 mV dec −1 [48]. However, our electrode presented inferior OER performance, proving that the current as-prepared Cu x O/Cu electrode is unsuitable for OER application. In some previous studies, CuO was applied as an OER electrocatalyst when embedded or grown on other supporting materials, e.g., Co 3 O 4 and graphene networks [42,51]. In the further study of the Cu x O/Cu materials in this study, the exploration of constructing some Cu x O-based hybrid might promote its application in OER.

Conclusions
In summary, a Cu x O/Cu foil was developed via a Ga-Cu surface alloying-dealloyingelectrooxidation strategy. The alloying process generated Ga2Cu alloys on the surface, and the following dealloying resulted in the formation of a nanoporous structure with a ligament size of~200 nm. Cu x O flakes were grown on the Cu ligament, with the assistance of electrooxidation. The O-Cu-2h electrode had the areal capacitance of 0.745 F cm −2 at 0.2 mA cm −2 , and 94.71% of the capacitance could be retained after 12,000 CV cycles. The O-Cu-2h//AC asymmetric supercapacitor delivered an energy density of 24.20 Wh kg −1 , with a power density of 0.65 kW kg −1 . Though the OER stability was inferior, further investigation could be explored, in order to improve the possibility of using the electrode as OER catalysts.
Supplementary Materials: The following supporting information can be downloaded at: https: //www.mdpi.com/article/10.3390/nano12122121/s1, Figure S1: The SEM images of Ga-Cu surface. Figure S2: XPS full survey O-Cu-15m and O-Cu-2h. Figure S3: A typical EDX spectrum of the dealloyed Ga 2 Cu film corresponding to Figure 2b. Figure S4: SEM and EDS results of O-Cu-15m.  Figure S11: Nyquist impedance plots of O-Cu-2h and O-Cu-15m. Figure S12: XRD pattern and surface SEM image of O-Cu-2h after 12,000 CV cycles at 100 mV s −1 . Figure S13: CV and GCD curves of AC. Figure S14: The comparison photos of Pt electrode before and after i-t test and the ICP result of the electrolyte after i-t test. Table  S1: Comparison of Cu x O-based electrode for supercapacitors. Table S2: Comparison of the other asymmetric supercapacitors and this work of energy density and power density.