*2.5. Characterization*

The surface morphologies of active materials were examined by field-emission gun scanning electron microscopes (SEM) (Hitachi, Tokyo, Japan and JEOL, Tokyo, Japan). Their microstructures and lattice fringes were examined by a high-resolution transmission electron microscope (HRTEM) (JEOL, Tokyo, Japan). The elemental mappings were obtained by the energy-dispersive X-ray spectroscopy (EDS). The chemical compositions of active materials were examined by the X-ray photoelectron spectrometer (XPS) (Thermo VG-Scientific, Waltham, MA, USA). According to the binding energies of photoelectrons emitted from the surface, the chemical state of each element could be ascertained. The vibrational modes of molecules identified from the absorption characteristics by Raman spectroscopies (Horiba Jobin Yvon, Paris, France) and infrared (Bruker, Billerica, MA, USA) ranging from 400 cm<sup>−</sup><sup>1</sup> to 2000 cm<sup>−</sup><sup>1</sup> and 400 cm<sup>−</sup><sup>1</sup> to 4000 cm<sup>−</sup>1, respectively, were used to determine the chemical compositions, bond configurations, and molecular structures of the active materials.

#### *2.6. Electrochemical Measurements*

The electrochemical properties were characterized by CV, galvanostatic charge/discharge (GCD), and electrochemical impedance spectroscopy (EIS), using a potentiostat/galvanostat (CH Instruments, Austin, TX, USA) as the analyzer. When the measurements on the electrodes coated with active materials were performed, a 5 M LiCl solution was employed as the electrolyte, and a three-electrode configuration consisting of a platinum (Pt) wire as the auxiliary electrode and silver chloride (Ag/AgCl) reference electrode was adopted.

The capacitance characteristics of the electrodes could be determined by the areas inside the CV curves obtained at different scan rates and the symmetry of the GCD curves obtained by different current densities. By Equations (1) and (2), the gravimetric specific capacitances CCV and CC-DC of individual electrodes were calculated from the CV and GCD curves, respectively [60]:

$$\mathbf{C}\_{\rm CV} = \mathbf{k} \frac{\int \mathbf{i}}{\mathbf{m} \cdot \mathbf{s}} \tag{1}$$

$$\mathbf{C}\_{\text{C-DC}} = \mathbf{k} \frac{\mathbf{i} \cdot \Delta \mathbf{t}}{\Delta \mathbf{V} \cdot \mathbf{m}} \tag{2}$$

$$\mathcal{E}\_{\rm EL}(\rm Wh/kg) = (\frac{1}{4} \times \mathcal{C}\_{\rm CV} \times \rm V^2) / 3.6\tag{3}$$

$$\text{EL}(\text{Wh/kg}) = (\frac{1}{4} \times \text{C}\_{\text{C-DC}} \times \Delta \text{V}^2) / 3.6\tag{4}$$

$$P\_{\rm EL} \left( \text{W} / \text{kg} \right) = \text{E}\_{\rm EL} / \left( \Delta \text{t} \right) \tag{5}$$

where k is the electrode constant (usually 2 for a single electrode and 4 for a couple of electrodes), i is the discharging current, - i is the integral area of a CV curve, m is the mass of electrode active materials, s is the scan rate (100 mV·s<sup>−</sup><sup>1</sup> in this work), Δt is the discharging time, and ΔV is the potential window subtracting the initial potential drop. In this study, the potential windows for individual electrodes were −1.9 V to 1.0 V. After substituting CCV into Equation (3) and CC-DC into Equation (4), the energy densities of the electrodes (EEL) could be calculated. By further substituting EEL into Equation (5), the power densities of the electrodes (PEL) were obtained [60]. On the other hand, the electronic and ionic transports across the interface of active material in the electrodes were investigated by EIS. The frequency range for EIS was 10−<sup>2</sup> Hz to 10<sup>5</sup> Hz. The AC amplitude was set as 10 mV between two electrodes.

#### **3. Results and Discussion**

The surface morphologies of graphite oxide, G, NG, and x-NGM composites at different magnifications were examined by SEM. The ultrathin sheet-like structures of graphite oxide, G, and NG were observed, as displayed in Figure 1. There are fine needle structures in x-NGM, which grow on the surface of NG, as shown in Figure 1d–f. The changes in the content of Mn were discovered to impact the morphologies and structures of x-NGM composites significantly. When the concentration of KMnO4 used during the preparation was higher to increase the content of Mn, a larger dimension of the needle structure resulted, as shown in Figure 1e,f. When the concentration of KMnO4 was 26.7 mM to fabricate 3-NGM, the intertwined MnO2 nanowires formed. When the KMnO4 concentration increased to 35.6 mM and 44.5 mM to obtain 4-NGM and 5-NGM, the MnO2 nanorods were observed, as displayed in Figure 1g,h. For further confirmation, elemental mappings were examined. Figure 2 demonstrates the presence of the C, O, N and Mn elements in 2-NGM and their even distributions. It verifies the successful preparation of the x-NGM composites as well. The sparser distribution of the N element was attributed to the relatively lower proportion of the N content in the composite.

Figure 3 displays the TEM micrographs and microstructures of NG, 2-NGM, and 3-NGM. The semitransparent membranous structure can be observed, as shown in Figure 3a,c,e, demonstrating the existence of NG in the composites. The lattice fringe spacing value of 0.34 nm corresponds to the d-spacing of G crystalline plane, indicating the successful formations of G and NG by the hydrothermal method [61]. Lots of fine needle structures appear on the NG surfaces of 2-NGM and 3-NGM, as shown in Figure 3c,e, respectively. The needle structure in 2-NGM is larger than that in 3-NGM. By the high-resolution atomic images in Figure 3d,f, the MnO2 in the x-NGM composites is discovered to be a two-phase mixture. Namely, the co-existence of γ-MnO2 and α-MnO2. Another two lattice fringe spacing values of 0.212 nm and 0.239 nm correspond to the (200) plane of γ-MnO2 and (211) plane of α-MnO2, respectively [29], which also contribute to confirming the presence of MnO2 and the successful preparation of the x-NGM composites.

**Figure 1.** SEM micrographs of (**a**) graphite oxide, (**b**) G, (**c**) NG, (**d**) 1-NGM, (**e**) 2-NGM, (**f**) 3-NGM, (**g**) 4-NGM, and (**h**) 5-NGM.

The functional group types on the surface of a material can be identified by the FTIR technique. Figure 4 shows the FTIR spectra of graphite oxide, G, NG, and x-NGM composites. For graphite oxide, the main absorption peaks are at 1084 cm<sup>−</sup><sup>1</sup> and 1218 cm<sup>−</sup>1, which are ascribed to the C-O stretchings. For G and NG, the two peaks at 1401 cm<sup>−</sup><sup>1</sup> and 1565 cm<sup>−</sup><sup>1</sup> are attributed to the O-H bending and C=C stretching, respectively. Graphite oxide, G, and NG all show the peak at 1720 cm<sup>−</sup>1, which is ascribed to the C=O stretching. Another peak for G at 1214 cm<sup>−</sup><sup>1</sup> can be assigned to the C-O stretching [62–65]. For the N-containing composites (NG and x-NGM), the two main absorption peaks are at 1195 cm<sup>−</sup><sup>1</sup> and 1565 cm<sup>−</sup>1, which are attributed to the C-N and C=N/C=C stretchings, respectively [66,67]. For the x-NGM composites, the two peaks at 438 cm<sup>−</sup><sup>1</sup> and 560 cm<sup>−</sup><sup>1</sup> can be ascribed to the Mn-O stretching [64–67]. One more peak for 2-NGM to 5-NGM is observed at 749 cm<sup>−</sup>1, which can also be attributed to the Mn-O stretching, and not be found in the composite without or with too less content of Mn (such as 1-NGM). The FTIR results mentioned above also contribute to confirming the successful preparation of the G, NG, and x-NGM composites.

**Figure 2.** (**a**) SEM and (**b**) EDS layered images of 2-NGM. Elemental mappings of 2-NGM: (**c**) C, (**d**) O, (**e**) Mn, and (**f**) N.

**Figure 3.** TEM micrographs and microstructures of (**<sup>a</sup>**,**b**) NG, (**<sup>c</sup>**,**d**) 2-NGM, and (**<sup>e</sup>**,**f**) 3-NGM.

Raman spectroscopy is a widely used technique to examine the structures and electronic properties of G and its derivatives. Figure 5 shows the Raman spectra of graphite oxide, G, NG, and x-NGM composites, ranging from 400 cm<sup>−</sup><sup>1</sup> to 2000 cm<sup>−</sup>1. The two feature peaks at 1329 cm<sup>−</sup><sup>1</sup> and 1590 cm<sup>−</sup><sup>1</sup> are D and G bands, respectively [52,68,69]. The D band is due to the presence of disorders in sp2-hybridized carbon systems. It can be used to estimate the defect level and content of impurity in the G sheets. The G band is derived from the stretching of sp2-hybridized carbon-carbon bonds and highly sensitive to strain effects in the sp<sup>2</sup> system within the G sheets. Furthermore, the intensity ratio of D and G bands, ID/IG, can be considered as a measure of the relative concentration of local defects or interferences, i.e., can be used to estimate the extent of sp<sup>3</sup> graphite oxide converting to sp<sup>2</sup> G [68,69]. Thus, an increment of the ID/IG value implies an increase in the number of defects. From Figure 5a, the ID/IG value of graphite oxide obtained before the hydrothermal process is calculated to be 1.73. However, those of the G, NG, and x-NGM composites drop to in between 1.55 to 1.69 after the hydrothermal process. It can be then deduced from the reduced ID/IG values that the hydrothermal process could remove oxygen-containing functional groups and reduce graphite oxide to G successfully. This again helps to ascertain the presence of G in NG and x-NGM composites. Since the

incident light wavelength of the Raman spectrometer influences excitation efficiency, the wavelength of 532 nm regarded as comparatively of less negative impact on enhancing the characteristic peak of Mn-O bonds was chosen to excite the x-NGM composites. However, as shown in Figure 5a, it still unavoidably resulted in weaker scattering intensity and thereby less distinct Mn-O characteristic peaks. After magnification, the peak at around 560 cm<sup>−</sup><sup>1</sup> attributed to the stretching vibration of Mn-O bonds [52] can be more clearly seen in Figure 5b, again confirming that the hydrothermal preparation of x-NGM composites was successful.

**Figure 4.** FTIR spectra of graphite oxide, G, NG, and x-NGM composites.

**Figure 5.** (**a**) Raman spectra of graphite oxide, G, NG, and x-NGM composites; (**b**) enlarged Raman spectra of x-NGM composites.

Figure 6 shows the XPS spectra of graphite oxide, G, NG, and x-NGM composites. The composition of a composite and chemical states of elements can be investigated by XPS. Figure 6a displays the C 1s spectra. For graphite oxide and G, they have a strong energy peak centered at approximately 283.8 eV, which is assigned to the C=C bonds (sp2-hybridized carbon atoms). Another weak peak with higher binding energy at 285.6 eV can be assigned to the C-O bonds (oxygenated carbon atoms) [62,63,70]. There is a weaker peak for graphite oxide at 287.4 eV, which is assigned to the C=O bonds. G shows an even weaker peak at 288.6 eV, which is assigned to the O-C=O bonds [52,62,63,70]. For NG and x-NGM composites, they also have energy peaks ascribed to the C=C bonds (symbol of the presence of G), centered at approximately 284.0 eV. When the containing of Mn in x-NGM to diminish the C=C bonds and sp2-hybridized carbon atoms, the peak intensity is found to be weakened. Another two weaker peaks of NG and x-NGM, centered at approximately 285.3 eV and 286.6 eV, can be assigned to the C=N and C-N bonds, respectively [62,63,69], which stand for another evidence for the existence of N element in the composites.

**Figure 6.** *Cont.*

**Figure 6.** XPS spectra of graphite oxide, G, NG, and x-NGM composites: (**a**) C 1s, (**b**) O 1s, (**c**) N 1s, and (**d**) Mn 2p.

Figure 6b displays the O 1s spectra. Both graphite oxide and G show an energy peak ascribed to the C=O bonds at 531.8 eV and 532.7 eV, respectively. The three energy peaks of NG at 531.6 eV, 533.3 eV, and 535.7 eV, can be assigned to the bondings of O-C, O-H, and H2O, respectively [52,62,63]. The five x-NGM composites have similar O 1s spectra. They all exhibit the three energy peaks centered at approximately 529.8 eV, 531.4 eV, and 533.5 eV, which correspond to the Mn-O, C-O, and O-H bonds, respectively [52,62,63,69]. The increase in the content of Mn is found to enhance the intensity of the 529.8 eV energy peak. Figure 6c shows the N 1s spectra. All the six NG and x-NGM composites exhibit a strong peak at 399.0 eV and two weak peaks at 400.1 eV and 401.0 eV, which can be attributed to the three types of N-containing species on the surface: pyridinic-N, pyrrolic-N, and graphitic-N (quaternary N), respectively [52,70,71]. Again, the presence of the N element in the six active materials is confirmed. The fabrication of NG from G is demonstrated to be successful as well. Figure 6d shows the high-resolution Mn 2p spectra. All the five x-NGM composites exhibit the two peaks centered at approximately 642.3 eV and 654.1 eV, which are ascribed to the Mn 2p3/2 and Mn 2p1/2 spin-orbit splitting states, respectively. The separation of spin energy between the two peaks is 11.8 eV, indicating

the oxidation state of Mn is +4. [52]. Moreover, the intensity of the two peaks is found to increase with an increased content of Mn. This not only confirms the presence of Mn in the x-NGM composites, but also demonstrates that the use of a hydrothermal method for preparation of the composites containing both NG and MnO2 was successful [52,62,63,72]. The XPS surveys have confirmed the presence of C, Mn, O, and N on the surface of the x-NGM composites.

EIS is a technique used for acquiring information of internal impedances in an electrochemical system. The electronic and ionic transports along the bulk and across the interface of active material in the electrode were thus investigated. The Nyquist plots of the 21 electrodes with different active materials are displayed in Figure 7, where the compressed semicircles at the high and medium frequency regions are related to the electronic transport resistance, a kinetic-controlled process. The line tail connecting the semicircle at the low-frequency region is associated with the ionic diffusion resistance, a thermodynamic-controlled process. The equivalent circuit for the EIS analysis is also depicted in Figure 7, which includes [73]: (1) the charge transfer impedance in the high-frequency region at the electrode/electrolyte interface (RCT), (2) the solution resistance (RS), which is the contact series resistance between the substrate and current collector, (3) Warburg impedance (W), which is the diffusion resistance of ions in the electrolyte and in relation to the slope of the line tail in the low-frequency region, (4) the electric double-layer capacitor (C1) [65,74–76]. The corresponding RCT values obtained by simulation are listed in Table 1. By contrast, G1, NG1, 1-NGM1, 2-NGM1, 3-NGM1, 4-NGM1, and 5-NGM1 are found to exhibit smaller semicircles in the high-frequency region. Their RCT values are 3.27 Ω, 2.22 Ω, 2.17 Ω, 1.28 Ω, 1.15 Ω, 8.06 Ω, and 9.29 Ω, respectively. When the mass loading of active materials on the substrate increases to 2 mg, the RCT values of G2, NG2, 1-NGM2, 2-NGM2, 3-NGM2, 4-NGM2, 5-NGM2 are 3.52 Ω, 2.98 Ω, 2.49 Ω, 1.35 Ω, 1.23 Ω, 8.43 Ω, and 9.40 Ω, respectively. When the mass loading further increases to 3 mg, the RCT values of G3, NG3, 1-NGM3, 2-NGM3, 3-NGM3, 4-NGM3, and 5-NGM3 are 12.83 Ω, 9.21 Ω, 8.42 Ω, 2.14 Ω, 1.70 Ω, 9.46 Ω, and 11.60 Ω, respectively. According to the significantly increased RCT and larger semicircle in the high-frequency region, it can be then concluded that the preferred mass loading of active materials on the PI/graphite flexible substrate is 1 mg.

**Figure 7.** Nyquist plots of the 21 electrodes with Gy, NGy, and x-NGMy composites. Insets are enlargements when the mass loading is (**a**) 1 mg, (**b**) 2 mg, and (**c**) 3 mg.


**Table 1.** RCT values of the 21 electrodes with Gy, NGy, and x-NGMy composites obtained by EIS simulation.

As shown in Table 1, after N was involved in the active materials to obtain NG composites, the RCT value decreased from 3.27 Ω (G1) to 2.22 Ω (NG1). The interconnected NG component can effectively facilitate electronic transport. When N and MnO2 were simultaneously involved in the x-NGM composites, the RCT value was further reduced from 2.22 Ω (NG1) to 1.15 Ω (3-NGM1). This indicates that the co-existence of NG and MnO2 in an active material could even more improve charge transfer. The diffusion/transport properties of electrolyte ions to the electrode surface can be examined by the linear response (slope) of the Nyquist plot in the low-frequency region. An increased slope usually illustrates a lower diffusion resistance, faster ionic transport, and thereby enhanced capacitive property [74,77]. As displayed in Figure 7, compared to those of G1 and NG1 electrodes, the Nyquist plot of the 1-NGM1 electrode has a larger slope in the low-frequency region, indicating its better ionic diffusion and higher conductivity since it contains NG and MnO2 simultaneously. The impact of the Mn content on RCT was further investigated. Among the 15 electrodes with x-NGM active materials, the RCT values for 1-NGM1, 2-NGM1, 3-NGM1, 4-NGM1, and 5-NGM1 are 2.17 Ω, 1.28 Ω, 1.15 Ω, 8.06 Ω, and 9.29 Ω, respectively. The 3-NGM1 electrode exhibits the smallest RCT, indicating its best charge transfer efficiency. Its Nyquist plot in the low-frequency region is almost vertical and shows the largest slope, representing the best ionic diffusion and capacitive properties of the 3-NGM1 electrode. It can be then deduced that the MnO2 nanowires growing on the NG surface provide more effective contacts between the electrode and electrolyte ions, and charge transfer is thus improved. However, excess Mn in the 4-NGM1 and 5-NGM1 electrodes causes increased RCT values due to fewer contacts. Ion diffusion and charge transfer capacity are thereby suppressed, and a smaller slope of the Nyquist plot in the low-frequency region results. By the aforementioned results, it is confirmed that both the mass loading and content of Mn in an active material electrode affect conductivity. The best charge transfer efficiency can be obtained only when the mass loading is 1 mg and the content of Mn in x-NGM composites is optimized (x = 3).

The capacitive characteristic of an active material electrode can be evaluated by integrating the area inside a CV curve loop. Figure 8a,b shows the CV curves of the 21 electrodes with Gy, NGy, and x-NGMy composites, obtained at the scanning rate of 100 mV·s<sup>−</sup><sup>1</sup> with a fixed potential range of −2.9 V to 1.0 V. All x-NGMy composites show redox waves, indicating that the faradaic phenomena occurred during the charge/discharge process. No redox peaks are found for Gy and NGy composites. Their CV curves exhibit similar rectangular and symmetrical form, which is related to the characteristics of an electric double-layer capacitor. In Figure 8a, the CV curve loops of the 1-NGMy electrodes are the largest, indicating that they have the highest specific capacitances, followed by the NGy and then the Gy electrodes. Among them, the 1-NGM1 electrode has the largest area inside the loop, giving rise to a high specific capacitance of 416 <sup>F</sup>·g<sup>−</sup>1. Its energy and power densities obtained by calculations are 243.2 Wh·kg−<sup>1</sup> and 959.9 <sup>W</sup>·kg−1, respectively. The capacitance enhancement can be ascribed to the synergistic effect of the higher conductivity by NG and the larger specific surface area by MnO2 nanostructures. Afterward, the content of Mn was tuned by using various KMnO4 concentrations. The impacts of mass loading on the capacitive property were also studied by using 1 mg, 2 mg, and 3 mg of active materials (y = 1, 2, and 3) for each x. From the CV curve loops of the x-NGMy electrodes in Figure 8b, the 3-NGM1 electrode shows the best capacitance performance

with the highest specific capacitance of 638 <sup>F</sup>·g<sup>−</sup>1. Its energy and power densities are 372.7 Wh·kg−<sup>1</sup> and 4731.1 <sup>W</sup>·kg−1, respectively. All the capacitance parameters acquired by calculations are listed in Table 2, which reveals that there was a most appropriate KMnO4 concentration, i.e., 26.72 mM when x = 3, to achieve an NGM composite with the optimum Mn content. Moreover, the increase of specific capacitance is more significant by pseudocapacitive MnO2 than NG.


**Table 2.** Capacitance parameters obtained from the CV results of the 21 electrodes with Gy, NGy, and x-NGMy composites.

By the interlaced nanowire structures of MnO2, the contact between the electrode surface and electrolyte ions and the use of active materials are both improved. There are more sites for electrochemical reactions to occur, and enhanced diffusion of ions in the electrolyte favorable for capacitive characteristic has resulted. The higher KMnO4 concentrations during the preparation process, 35.62 mM and 44.53 mM when x = 4 and 5, led to excess MnO2 in the NGM composites, which grew into nanorod structures detrimental to more contact between the electrode surface and electrolyte ions. Inferior ion diffusion and worse capacitive characteristics of the 4-NGMy and 5-NGMy electrodes are thereby caused, as shown in Figure 8b and Table 2. It can be also seen from Table 2 that the mass loading of active material on the flexible electrode is also critical. 1 mg has been demonstrated to be the most appropriate mass loading. 2 mg and 3 mg cause overloaded active materials and thus reduced specific capacitance, energy, and power densities. By plotting energy density vs. power density obtained from the CV results and calculated by Equations (3) and (5), a Ragone plot is achieved, as shown in Figure 8c, which again demonstrates the best electrochemical performance of 3-NGM1 among the 21 x-NGMy electrodes when x = 3 and the mass loading of 1 mg were used.

**Figure 8.** CV curves of the 21 electrodes with (**a**) Gy, NGy, 1-NGMy, and (**b**) x-NGMy composites. (**c**) Ragone plot obtained from the CV results.

The symmetry of charge and discharge curves can be investigated to understand capacitive behavior. Figure 9a shows the GCD curves of the 9 electrodes with Gy, NGy, and 1-NGMy composites, obtained by a fixed potential range of −1.2 V to 1.0 V under different current densities. Those of the 1-NGMy electrodes are slightly distorted from the ideal triangle shape because of the pseudocapacitive contribution from MnO2. The curvature implies that they are typical Faraday capacitance curves [78]. It is revealed that the 1-NGMy electrodes have the best capacitive characteristics, followed by the NGy electrodes, and the Gy electrodes are the worst. The 1-NGM1 electrode exhibits a longer charge and discharge time at a current density of 0.2 <sup>A</sup>·g<sup>−</sup>1, resulting in a specific capacitance of 188 <sup>F</sup>·g<sup>−</sup>1, and the corresponding energy and power densities are 63.1 Wh·kg−<sup>1</sup> and 249.2 <sup>W</sup>·kg−1, respectively. Afterward, the impacts of the Mn content and mass loading on the capacitive parameters of the electrodes were also explored. The specific capacitances can be calculated from the GCD curves by Equation (2), as listed in Table 3. Figure 9b shows the plots of specific capacitance vs. current density for the 21 electrodes. The Gy, NGy, 1-NGMy, 4-NGMy, and 5-NGMy electrodes cannot endure the current densities larger than 1 <sup>A</sup>·g<sup>−</sup>1. The 2-NGM2 electrode can endure only the current densities of 1 <sup>A</sup>·g<sup>−</sup>1,3A·g<sup>−</sup>1, and 5 <sup>A</sup>·g<sup>−</sup>1. The 3-NGM2 electrode can endure the current densities of 1 <sup>A</sup>·g<sup>−</sup>1, 3 <sup>A</sup>·g<sup>−</sup>1,5A·g<sup>−</sup>1, and 7 <sup>A</sup>·g<sup>−</sup>1. The 2-NGM3 and 3-NGM3 electrodes can endure only the current densities of 1 <sup>A</sup>·g<sup>−</sup><sup>1</sup> and 3 <sup>A</sup>·g<sup>−</sup>1, whereas the 2-NGM1 and 3-NGM1 electrodes can endure all current densities. Among the 21 electrodes, the 3-NGM1 electrode exhibits the best endurance. Its energy and power densities are 86.7 Wh·kg−<sup>1</sup> and 1100.0 <sup>W</sup>·kg−1, respectively. The above results have also confirmed the optimum conditions for the mass loading and content of Mn.

**Figure 9.** (**a**) GCD curves of the 9 electrodes with Gy, NGy, and 1-NGMy composites. (**b**) Plots of specific capacitance vs. current density for the 21 electrodes with Gy, NGy, and x-NGMy composites. (**c**) GCD curves of the 3-NGM1 electrode under the current densities of 1 <sup>A</sup>·g<sup>−</sup><sup>1</sup> to 9 <sup>A</sup>·g<sup>−</sup>1. (**d**) Ragone plot obtained from the GCD results.

Since the 3-NGM1 electrode has shown the best sustainable ability to permit its higher charging capacity, it was selected to perform further GCD investigation by different current densities, are shown in Figure 9c. The rapid intercalation/deintercalation of metallic cations in an active material reveals the redox concerning the oxidation state transitions between Mn (III) and Mn (IV). It is inferred that at higher current densities, only the external surface of an active material is involved in charge/discharge, leading to insufficient redox and relatively lower specific capacitances. The charge/discharge time decreases along with a small number of electrolyte ions occupying the active sites. By contrast, at lower current densities, more internal and external active sites are involved, attaining more complete redox reactions and higher specific capacitances [79]. The increased charge/discharge time results from most electrolyte ions being anchored to the active sites at the interface. As displayed in Table 3, when the current density applied to the 3-NGM1 electrode increases from 1 <sup>A</sup>·g<sup>−</sup><sup>1</sup> to 9 <sup>A</sup>·g<sup>−</sup>1, the specific capacitance reduces from 258 <sup>F</sup>·g<sup>−</sup><sup>1</sup> to 13 <sup>F</sup>·g<sup>−</sup>1. The massive capacitance decay implies that the rate capability of the x-NGMy electrodes still needs considerable improvement. Moreover, it is perceived from Table 3 that the mass loading of active materials on the flexible electrode is very critical. 1 mg has been proven to be the optimum. Both 2 mg and 3 mg are overloads to cause lower conductivity and inferior capacitive parameters. For the Gy, NGy, and x-NGMy electrodes, their energy and power densities calculated from the GCD results are plotted as a Ragone plot, as shown in Figure 9d. Consistent with the EIS and CV results, the synergistic effect of NG with MnO2 is demonstrated again, to promote reversible redox reactions on the pseudocapacitive materials and play grea<sup>t</sup> impacts on the capacitance characteristics of the electrodes.


**Table 3.** Capacitance parameters obtained from the GCD results of the 21 electrodes with Gy, NGy, and x-NGMy composites by different current densities.
