**1. Introduction**

The rapid development of portable and wearable consumer electronics results in the demand for high-performance energy-storage devices [1–7]. Solid-state supercapacitors (SSCs) have thus attracted growing attention, due to their reversibility, safe operation without the use of any liquid electrolyte, longer cycling stability than batteries, and higher power and energy densities than conventional capacitors [1,2,8,9]. Further improvement in energy density and searching for economic flexible current collectors are regarded as the two major challenges when pushing the technology of SSCs forward. Lots of effort have been made to investigate electrode active materials, to achieve higher specific capacitance. Carbonaceous materials such as activated carbon, carbon fibers, carbon nanotubes, graphene (G), and graphene oxides, etc., can be used as the active materials for electrodes of SSCs due to low cost, high conductivity, and chemical stability [10–16]. However, they usually show the drawback of lower specific capacitance, as compared to manganese oxide (MnO2) [17,18]. The theoretical

specific capacitance of porous carbon materials is only 250 <sup>F</sup>·g<sup>−</sup><sup>1</sup> when the specific surface area is 1000 <sup>m</sup>2·g<sup>−</sup><sup>1</sup> [19]. Modification of pure carbon, like mixing activated carbon with conductive additives to obtain active materials with a high surface area and better contact for more efficient electrodes, was reported to improve the conductivity and capacitances of supercapacitors (SCs) [17]. Addition of heteroatom(s) or a pseudocapacitive material in G to fabricate composites was also conducive to the performance of SCs. Nitrogen (N) has a high electronegativity and lone pair of electrons, which can be in conjugation with the π electrons of G to enhance conductivity and electron transport. So, compared to pure G, N-containing G (NG) would be more favorable to the specific capacitance of SSCs.

Compared with carbonaceous materials, transition metal oxides and hydroxides (RuO2, MnO2, Ni(OH)2, Fe2O3, etc.) are deemed as more promising active materials for electrode applications to SSCs with large specific capacitances and high energy densities because the diverse oxidation states of the transition metals permit effective charge transfer [20–24]. Among them, MnO2 has attracted considerable attention due to its low cost, natural abundance, environmentally benign nature, and high theoretical specific capacitance (1370 <sup>F</sup>·g<sup>−</sup>1) [25–27]. Its pseudocapacitive characteristic can be attributed to the single electron transfer in the Mn3+/Mn4+ redox system [28]. Nevertheless, its poor conductivity and low ion diffusion constant may suppress the further progress [29,30]. The stability and potential application may be hindered by its problematic dissolution in electrolytes. To address these issues, the combination of MnO2 nanostructures with conductive carbonaceous materials to form hybrid structures has been adopted [31–39]. Some studies have focused on the development of MnO2-G composites, which took advantages of the high capacitance of MnO2 and the high conductivity of G simultaneously.

Despite the advantages of good electrochemical stability, high conductivity, and large specific surface area, G nanosheets tend to restack during the formation of solid materials due to the strong π-π interactions. The aggregation reduces the accessible surface area for adsorption and desorption of electrolyte ions, which would finally result in a small specific capacitance. The high conductivity and unique surface characteristics of the single-layer G nanosheets can be thereby lost [40]. Thus, suppression of the aggregation would greatly optimize the electrochemical properties of electrodes. One feasible method is to anchor transition metal oxides to the surface of G, working as spacers to separate adjacent G nanosheets. Transition metal oxides such as RuO2, NiO, CoOx, and MnO2 are appropriate candidates since they have been considered as extensively explored pseduocapacitor's electrode materials showing high theoretical specific capacitance [41–44]. Introducing porous MnO2 nanostructures into G nanosheets would suppress the aggregation. The specific surface area of G was then increased, and more electrical conduction pathways were provided. The relatively higher gravimetric capacitances had been demonstrated for a variety of MnO2-G composites with low mass loadings, whereas some literature highlighted the importance of fabricating composites with higher mass loadings [45–50]. The optimization of mass loading would remain an important challenge.

The asymmetric SCs consisted of a carbonaceous negative electrode and a MnO2-based positive electrode offered enlarged operation voltage windows and thus improved power energy properties, compared with symmetric SCs [51]. For example, MnO2 nanostructures grown on activated carbon by a wet chemical reaction process was used as the positive electrode, which exhibited a high specific capacitance (345.1 <sup>F</sup>·g<sup>−</sup><sup>1</sup> at 0.5 <sup>A</sup>·g<sup>−</sup>1) and excellent cycle stability. It was assembled with an activated carbon negative electrode to fabricate asymmetric SCs, which showed high energy density of 31.0 Wh·kg−<sup>1</sup> at a power density of 500.0 <sup>W</sup>·kg−<sup>1</sup> [19]. Layered δ-MnO2 on N-doped G obtained by a hydrothermal approach was used as the cathode to improve the conductivity and present a high specific capacitance of about 305 <sup>F</sup>·g<sup>−</sup><sup>1</sup> at a scan rate of 5 mV·s<sup>−</sup>1. When it was assembled with an activated carbon anode using a gel electrolyte to fabricate flexible asymmetric SSCs (ASSCs), a maximum energy density of 3.5 mWh·cm<sup>−</sup><sup>3</sup> at a power density of 0.019 W·cm<sup>−</sup><sup>3</sup> was achieved [52]. Despite the progress, most MnO2-based SSCs did not exceed the energy density of lead acid batteries. Lots of efforts have been made to ASSCs with various electrode combinations [53–55]. Consequently, it is crucial to develop new active materials for more efficient electrodes applied to SCs.

In this study, MnO2/NG composites with various contents of Mn were fabricated by a hydrothermal approach and used as the electrode active materials for flexible ASSCs. Graphite paper on polyimide (PI) was employed as the soft substrate. NG can enhance the conductivity of composites and efficiently reduce the interfacial impedance. It can also serve as a better template for inducing the growth of MnO2 nanostructures than G. By the synergistic effect of MnO2 and NG, the specific capacitance, energy, and power densities were significantly improved. The MnO2 in the composites was found to be mixed phases containing γ-MnO2 and α-MnO2. The impacts of mass loading and the content of Mn on the capacitance parameters were also explored. The 3-NGM1 electrode with the most appropriate Mn content and mass loading of active material exhibited a high specific capacitance of 258 <sup>F</sup>·g<sup>−</sup><sup>1</sup> at a current density of 1 <sup>A</sup>·g<sup>−</sup>1. By calculating the cyclic voltammetry (CV) results, it had a superior specific capacitance of 638 <sup>F</sup>·g<sup>−</sup>1. The corresponding energy and power densities were 372.7 Wh·kg−<sup>1</sup> and 4731.1 <sup>W</sup>·kg−1, respectively. The ongoing work regarding flexible ASSCs will be designed as using G as the negative electrode and a MnO2/NG composite as the positive electrode by employing a solid gel electrolyte. It is perceived that the optimized conditions of electrodes will lead to more enhanced capacitive behavior and cycle stability of the flexible ASSCs.
