Flexible Films as Anode Materials Based on rGO and TiO2/MnO2 in Li-Ion Batteries Free of Non-Active Agents

Recently, to meet the growing demand for stable and flexible batteries, anodes in the form of thin films have drawn the attention of researchers. It is clear that mass production of such batteries would bring the worldwide distribution of flexible devices and wearable electronics closer. Currently, electrodes are deposited on a flexible substrate and consist of conductive and binding agents that increase the volume/weight of the electrode. Here, we propose free-standing and non-active-material-free thin films based on reduced graphene oxide (rGO), titanium dioxide (TiO2) and manganese dioxide (MnO2) as working electrodes in lithium-ion half-cells prepared via the vacuum-assisted filtration method. The electrochemical performance of the assembled half-cells exhibited good cyclic stability and a reversible capacity at lower current densities. The addition of TiO2 and MnO2 improved the capacity of the rGO film, while rGO itself provided a stable rate performance. rGO/TiO2/MnO2 film showed the highest discharge capacity (483 mAh/g at 50 mA/g). In addition, all assembled cells displayed excellent repeatability and reversibility in cyclic voltammetry measurements and good lithium-ion diffusion through the electrolyte, SEI layer and the active material itself.


Introduction
Due to the constant utilization of fossil fuels and their negative impact on the environment, researchers around the world are persistently searching for energy storage systems that would not emit greenhouse gases nor pollute the environment [1]. In addition, there is an ongoing demand for high-tech and multifunctional portable products, for example, wearable electronics, bendable smartphones and displays. Because of this, energy storage systems in these devices should have matching mechanical properties, allowing them to tolerate deformation while maintaining their electrochemical functions [2].
Lithium-ion batteries (LIBs) are the most prominent power sources for electronic devices because of their high energy density, power density, mechanical stability and rate capability [3]. The cathode materials are usually made of LiCoO 2 or LiMn 2 O 4 . The cathode, being the source of lithium, determines the average voltage of the battery and its capacity. The cobalt-based electrodes are great for their high volumetric capacity, low self-discharge, high discharge voltage and good cycling performance, but they are quite expensive and have low thermal stability. The manganese-based cathodes are suitable due to their low cost, but nevertheless, they also have some limitations, including the tendency of manganese to dissolve into the electrolyte, leading to poor cycling stability of the positive electrode [4]. The role of the anode is to store (during charging) and release (during discharging) lithium ions originating from the cathode. Currently, the most commercially used anode material is graphite, with a maximum theoretical capacity of 372 mAh/g [5]. Hence, many researchers developed new anode materials with higher capacities, such as transition-metal oxides [6,7] or other carbon-based materials, such as graphene, carbon nanotubes [8,9] and their composites [10,11]. Graphene and its derivatives have many distinctive properties, such as a large specific surface area (2630 m 2 /g) [12], a high thermal conductivity (3000-5000 W/mK) [13] and a high carrier mobility at room temperature (~10,000 cm 2 /Vs) [14]. The graphene derivatives graphene oxide (GO) and reduced graphene oxide (rGO), which are terminated with oxygen groups that allow for the tuning of the properties of these materials, are known for their low-cost and mass production [15,16]. Due to their remarkable properties, graphene-based composites are excellent candidates for anode materials for lithium-ion batteries and a lot of research is being conducted to improve their electrochemical performance. As an alternative to full cells, the electrochemical nature of Li-ion batteries can be examined in a half-cell configuration, where metallic lithium acts as both the reference and the counter electrode (the cathode), while a carbon material acts as the working electrode (the anode). Due to a stable reference potential, as well as a very large reservoir of capacity, the reactions at the working electrode are limited by the capacity of the counter electrode [17]. For instance, Joshi et al. [18] fabricated a half-cell where an electrosprayed composite made of graphene sheets decorated with pillar-like ZnFe 2 O 4 nanoparticles acted as the working electrode/the anode. This anode material showed a high capacity, good rate capability and long-term cycling stability in comparison to cells free of graphene. The specific capacity at the current density of 100 mA/g was approximately 1100 mAh/g and stayed at that level after 30 cycles. Dong et al. [19] prepared a graphene-WS 2 multilayer film via waterevaporation-induced assembly and chemical reduction. Due to improved contact between the graphene and WS 2 , the material exhibited a high specific capacity, an effective cycling performance and a high rate capability. Moreover, the fast lithium-ion migration was attributed to the microcavities inside the material. Fang et al. [20] synthesized CoO/rGO film on Ti foil via the electrodeposition method. The composite showed an initial capacity of 1290 mAh/g under a current density of 100 mA/g and 766 mAh/g under 1000 mA/g.
Regarding other rGO/TiO 2 electrodes, Dong et al. [21] hydrothermally synthesized an rGO composite with TiO 2 nanorods. This material displayed a high rate capacity (up to 10 C) and a specific capacity of 347 mAh/g at 0.2 C that remained stable after 20 cycles. Tang et al. [22] prepared a TiO 2 /graphene nanocomposite by mixing and sonicating both materials. Their composite showed an initial specific capacity of 389 mAh/g at 100 mA/g of current density; however, it was not stable and the capacity gradually decreased. For an rGO/MnO 2 composite, Xu et al. [23] obtained MnO 2 /graphene films via electrophoretic deposition (EPD). Their material delivered a high specific capacity, even after 200 cycles. Despite such extended research, new materials for anodes are being developed every day to further increase the performance of flexible lithium-ion batteries without non-active components such as binding or conductive agents.
In this study, flexible, non-active-material-free thin films based on reduced graphene oxide (rGO), titanium dioxide (TiO 2 ) and manganese dioxide (MnO 2 ) have been fabricated by the vacuum-assisted filtration method (VAF). The obtained films (rGO, rGO/TiO 2, rGO/MnO 2 and rGO/TiO 2 /MnO 2 ) were binder-free (such as polyvinylidene fluoride (PVDF)) and without a conductive agent (such as carbon black). These anode materials for lithium-ion half-cells, and their performances, have been tested in great detail. More interestingly, an additional current collector such as Al or Cu foil was redundant in the configuration of the lithium-ion half-cells, which made the electrode much lighter than a typical tablet or as a slurry deposited on a collector. What is more, no toxic solvent, such as NMP, was used during electrode preparation. Therefore, in this approach, a low-cost and environmentally friendly process for the production of thin, free standing and flexible films serving as anodes in Li-ion half-cells has been designed. Additionally, these films displayed flexibility, bendability, and excellent repeatability, reversibility and rate performance as working electrodes in lithium-ion batteries.

Synthesis of Graphene Oxide (GO)
A modified Hummers' method was used to prepare graphene oxide. Briefly, 135 mL of a mixture of concentrated H 2 SO 4 and H 3 PO 4 (ratio = 8:1) was poured onto 1 g of graphite and 6 g of KMnO 4 and stirred for 24 h at 50 • C. After the reaction, the mixture was poured into a cold solution of hydrogen peroxide (1 mL per 150 mL of H 2 O), then centrifuged and washed alternately with water, hydrochloric acid (30%) and ethanol. Finally, the GO was dried in air at 60 • C for 12 h.

Synthesis of Manganese Dioxide (MnO 2 )
Manganese dioxide was synthesized by a method reported by Cao et al. [24]. Shortly, 1.58 g of KMnO 4 was dissolved in 60 mL of distilled water (DW) and, separately, 2.28 g of oxalic acid was dissolved in 100 mL of water. After both materials were dissolved, the KMnO 4 solution was slowly added to the oxalic acid solution. Next, the mixture was stirred for 2 h and then filtrated. The obtained precipitate was washed with water and dried at 90 • C for 12 h.

Thin Film Preparation
Thin films were prepared via the vacuum-assisted filtration (VAF) method. The composition of the obtained films (rGO, rGO/TiO 2 , rGO/MnO 2 and rGO/TiO 2 /MnO 2 ) is shown in Table 1, as well as the thickness, measured via scanning electron microscopy. Briefly, the following dispersions were prepared: 5 mg of GO in 100 mL of DW, 1 mg of TiO 2 in 100 mL of DW and 1 mg of MnO 2 in 100 mL of DW. Each mixture was sonicated until a homogeneous dispersion was achieved. After that, 5 mL of the appropriate dispersion was introduced to the microfiltration set to form the desired films. The procedure was repeated about 70 times until films of the desired thickness were fabricated (see Table 1). The obtained films were dried at room temperature, peeled off of the PC membrane and thermally reduced at 400 • C for 2 h in inert gas conditions. The reduction of GO to rGO was performed to improve the film's conductivity and electrochemical responses. The structures of the produced films are presented in Figure 1.

Electrochemical Measurement
The coin cells (CR2032) were assembled inside the Ar-filled glovebox (M. Braun Inertgas-Systeme GmbH, Garching, Germany). Constructed cells were built in a half-cell configuration, where a 15 mm disc of metallic lithium was used as both the reference and counter electrodes and a 12 mm disc of each film acted as working electrodes. Reference TiO2 and MnO2 electrodes were prepared in the form of tablets with the addition of carbon black and PVDF in the ratios of 8:1:1 for TiO2 and 4.5:4.5:1 for MnO2. The increased amount of carbon black in the MnO2-based electrodes was used due to the material pulverizing after discharge/charge cycles. This procedure allowed the electrode to be buffered more effectively. The electrolyte (1 M LiPF6 solution in ethylene carbonate and diethyl carbonate) was used to wet both glass fiber and Celgard ® separators (polypropylene/polyethylene). Before the actual electrochemical test, the cells were connected to the channels for 8 h during the OCV technique. All electrochemical measurements were performed using a BioLogic VMP3 instrument (BioLogic, Seyssinet-Pariset, France) at room temperature. The cells were charged and discharged over a voltage range of 0 to 3 V vs. Li/Li + at room temperature. Cyclic voltammetry (CV) was performed in the same range at a scan rate of 0.1 mV/s. Electrochemical impedance spectroscopy (EIS) was executed in the frequency range from 100 to 10 mHz. Open circuit voltage (OCV) was turned on between techniques to stabilize the potential. Galvanostatic cycling with potential limitation (GCPL) was used to examine the specific capacity of the assembled cells during 5 cycles at the current densities: 50, 100, 200, 400, 600, 1000, 2000, 5000, 10,000 and back at 50 mA/g.

Material Characterization
Scanning electron microscopy (SEM) was performed using the SEM column of SEM/Ga-FIB FEI Helios NanoLab™ 600i dualbeam microscope (FEI Company, Hillsboro, USA). Transmission electron microscopy (TEM, Tecnai F30 with a field emission gun operating at 200 kV, Thermo Fisher Scientific, Waltham, MA, USA) with an energy

Electrochemical Measurement
The coin cells (CR2032) were assembled inside the Ar-filled glovebox (M. Braun Inertgas-Systeme GmbH, Garching, Germany). Constructed cells were built in a half-cell configuration, where a 15 mm disc of metallic lithium was used as both the reference and counter electrodes and a 12 mm disc of each film acted as working electrodes. Reference TiO 2 and MnO 2 electrodes were prepared in the form of tablets with the addition of carbon black and PVDF in the ratios of 8:1:1 for TiO 2 and 4.5:4.5:1 for MnO 2 . The increased amount of carbon black in the MnO 2 -based electrodes was used due to the material pulverizing after discharge/charge cycles. This procedure allowed the electrode to be buffered more effectively. The electrolyte (1 M LiPF 6 solution in ethylene carbonate and diethyl carbonate) was used to wet both glass fiber and Celgard ® separators (polypropylene/polyethylene). Before the actual electrochemical test, the cells were connected to the channels for 8 h during the OCV technique. All electrochemical measurements were performed using a BioLogic VMP3 instrument (BioLogic, Seyssinet-Pariset, France) at room temperature. The cells were charged and discharged over a voltage range of 0 to 3 V vs. Li/Li + at room temperature. Cyclic voltammetry (CV) was performed in the same range at a scan rate of 0.1 mV/s. Electrochemical impedance spectroscopy (EIS) was executed in the frequency range from 100 to 10 mHz. Open circuit voltage (OCV) was turned on between techniques to stabilize the potential. Galvanostatic cycling with potential limitation (GCPL) was used to examine the specific capacity of the assembled cells during 5 cycles at the current densities: 50, 100, 200, 400, 600, 1000, 2000, 5000, 10,000 and back at 50 mA/g.

Material Characterization
Scanning electron microscopy (SEM) was performed using the SEM column of SEM/Ga-FIB FEI Helios NanoLab™ 600i dualbeam microscope (FEI Company, Hillsboro, OR, USA). Transmission electron microscopy (TEM, Tecnai F30 with a field emission gun operating at 200 kV, Thermo Fisher Scientific, Waltham, MA, USA) with an energy dispersive X-ray spectrometer (EDX) was used to view both the morphology and the elemental distribution of the prepared films. A Renishaw InVia Raman microscope (Renishaw, New Mills Wottonunder-Edge, UK) (λ = 785 nm) was used for vibronic characterization of the samples. The phase composition analysis of the films was performed using an Aeris diffractometer (Malvern Panalytical, Malvern, UK) and Cu-Kα radiation (λ = 1.544 Å).

Transmission Electron Microscopy
A TEM analysis, shown in Figure 2a-d, was used to reveal the morphology of the prepared films. Here, the samples were prepared by dropping GO, TiO 2 and/or MnO 2 dispersions directly on the TEM grid. Additionally, Figure 2 shows digital images of the prepared films (Figure 2e-h) and their flexible behaviors are shown with a caliper set on 100 mm (Figure 2i-l). TEM images of pristine TiO 2 and MnO 2 are presented in Figure S1. GO exhibited typical wrinkled and folded morphology (Figure 2a). The GO flakes had a wide range of sizes, ranging from few dozen to a few hundred nanometers. Titanium dioxide particles deposited on GO flakes (Figure 2b), demonstrated a tendency to agglomerate into a large irregular form. Manganese dioxide particles (Figure 2c) showed a similar morphology. Figure 2d depicts GO/TiO 2 /MnO 2 with clear TiO 2 and MnO 2 agglomerates. To study the elemental composition of the produced films energy-dispersive X-ray spectroscopy (EDX) measurements were performed ( Figure 3). The results displayed that carbon, oxygen and titanium; carbon, oxygen and manganese; and carbon, oxygen, titanium and manganese are the only elements in GO/TiO 2 , GO/MnO 2 and GO/TiO 2 /MnO 2 , respectively. Carbon was detected from both GO flakes and the TEM grid.

X-ray Diffraction
XRD patterns of the obtained films are presented in Figure 5 (patterns were collected directly from the surface of the films). Additionally, the patterns of the reference TiO 2 and MnO 2 electrodes are shown in Figure S2. TiO 2 is composed of two crystalline phases: anatase and rutile. Here, the diffraction reflections at~25. 3      141, 600, 251, 002, 541 and 132 indexes of tetragonal α-MnO2 (ICSD card no. 073363), respectively. rGO displayed one clear reflection at ~25°, which came from the 002 diffraction plane. The presence of this reflection and the absence of a reflection at ~11° confirmed that GO was reduced successfully [25]. The characteristic reflection attributed to rGO was identified in the rGO/TiO2, rGO/MnO2 and rGO/TiO2/MnO2 patterns. Moreover, those films showed low-intensity reflections of TiO2 (rGO/TiO2), MnO2 (rGO/MnO2) and both TiO2 and MnO2 (rGO/TiO2/MnO2).  Figure 6 displays the Raman spectra of the rGO-based films. Every spectrum had the typical bands for carbon materials: D at 1324 cm −1 and G at 1604 cm −1 . The first band was caused by basal plane defects in the carbon structure. The second band was related to inplane stretching vibration of the C-C structure, in sp 2 systems [26]. The band at 151 cm −1 in the rGO/TiO2 spectrum was assigned to the Eg vibrational mode of TiO2 [27]. rGO/MnO2 showed a characteristic band at 646 cm −1 that corresponded to Mn-O stretching vibration [28]. The rGO/TiO2/MnO2 spectrum showed characteristic bands for rGO, TiO2 and MnO2, with slight shifts due to the additional components. Furthermore, the peak at 429 cm −1  Figure 6 displays the Raman spectra of the rGO-based films. Every spectrum had the typical bands for carbon materials: D at 1324 cm −1 and G at 1604 cm −1 . The first band was caused by basal plane defects in the carbon structure. The second band was related to in-plane stretching vibration of the C-C structure, in sp 2 systems [26]. The band at 151 cm −1 in the rGO/TiO 2 spectrum was assigned to the E g vibrational mode of TiO 2 [27]. rGO/MnO 2 showed a characteristic band at 646 cm −1 that corresponded to Mn-O stretching vibration [28]. The rGO/TiO 2 /MnO 2 spectrum showed characteristic bands for rGO, TiO 2 and MnO 2 , with slight shifts due to the additional components. Furthermore, the peak at 429 cm −1 could be attributed to a mixture of titanium and manganese oxides [29]. Additionally, the Raman modes of reference for TiO 2 and MnO 2 are presented in the Supplementary information ( Figure S3). MnO 2 showed strong peaks at~181, 283, 306, 370, 513, 582, 635 and 654 cm −1 , which confirmed the α-MnO 2 structure [30]. TiO 2 displayed peaks at~142, 395, 516 and 638 cm −1 , corresponding to the E g , B 1g , A 1g and E g modes of the anatase phase, respectively [27].

Electrochemical Measurements
To validate the electrochemical potential of the assembled films, lithium-ion batteries were constructed in a half-cell configuration. Metallic lithium acted as both the reference and counter electrodes, while rGO-based films acted as the working electrode. The cyclic performances of the rGO-based films are shown in Figure 7. Figure 7a presents the cyclic voltammograms (CV) of the rGO film, which did not exhibit any redox peaks. The black curve displaying the first cycle of charge/discharge diverges from the next cycles due to the irreversible formation of the SEI layer. The subsequent curves are overlapping each other, which means that delithiation/lithiation process was repeatable. In Figure 7b (rGO/TiO 2 ), two peaks can be observed at 2.04 and 1.75 V, which correspond to the oxidation and reduction reactions, respectively. The electrochemical reduction/oxidation (lithiation/delithiation) of TiO 2 is described below [31]: could be attributed to a mixture of titanium and manganese oxides [29]. Additionally, the Raman modes of reference for TiO2 and MnO2 are presented in the Supplementary information ( Figure S3). MnO2 showed strong peaks at ~181, 283, 306, 370, 513, 582, 635 and 654 cm −1 , which confirmed the α-MnO2 structure [30]. TiO2 displayed peaks at ~142, 395, 516 and 638 cm −1 , corresponding to the Eg, B1g, A1g and Eg modes of the anatase phase, respectively [27].

Electrochemical Measurements
To validate the electrochemical potential of the assembled films, lithium-ion batteries were constructed in a half-cell configuration. Metallic lithium acted as both the reference and counter electrodes, while rGO-based films acted as the working electrode. The cyclic performances of the rGO-based films are shown in Figure 7. Figure 7a presents the cyclic voltammograms (CV) of the rGO film, which did not exhibit any redox peaks. The black curve displaying the first cycle of charge/discharge diverges from the next cycles due to the irreversible formation of the SEI layer. The subsequent curves are overlapping each other, which means that delithiation/lithiation process was repeatable. In Figure 7b (rGO/TiO2), two peaks can be observed at 2.04 and 1.75 V, which correspond to the oxidation and reduction reactions, respectively. The electrochemical reduction/oxidation (lithiation/delithiation) of TiO2 is described below [31]: (1) Figure 7c shows the CV of the rGO/MnO2 film, which did not exhibit any characteristic peaks of manganese dioxide. Moreover, Figure 6d (rGO/TiO2/MnO2) presents peaks coming only from TiO2. Figure S4 shows CV plots of TiO2 and MnO2, respectively. The first cycle of the TiO2 cyclic performance ( Figure S4a) exhibited intense peaks of oxidation and reduction reactions at 2.18 and 1.57 V, respectively. Their intensity was reduced in the following cycles due to the poor cyclic stability of TiO2. Moreover, the SEI formation in the range from 0.01 to 0.56 V could be seen. Figure S4b presents the cyclic performance of MnO2. The first cycle displayed two peaks in the discharging/lithiation region: the first peak, at 0.1 V, corresponded to the reduction of Mn 4+ to Mn 0 [32], and the second peak, at 0.57 V, corresponded to the SEI layer formation. In the charging/delithiation region there was one peak, at 1.27 V, indicating the oxidation of Mn 0 to Mn 2+ . The reduction/oxidation process of manganese dioxide is described by the equation below [32]: MnO2 + 4 Li + + 4 e − ⇄ 2 Li2O + Mn (2)  Figure 8 displays the first, third and fifth cycles of the galvanostatic charge/discharge curves of each film at 50 mA/g, recorded between 0.005 and 3.0 V. The starting discharge/charge capacities were 411/388, 400/372, 402/387 and 482/470 mAh/g for rGO, rGO/TiO2, rGO/MnO2 and rGO/TiO2/MnO2, respectively. However, in the following cycles, rGO/MnO2 discharge/charge capacity exceeded that of rGO. In the subsequent cycles each material exhibited a stable discharge/charge capacity, displaying the excellent revers-  Figure 7c shows the CV of the rGO/MnO 2 film, which did not exhibit any characteristic peaks of manganese dioxide. Moreover, Figure 7d (rGO/TiO 2 /MnO 2 ) presents peaks coming only from TiO 2 . Figure S4 shows CV plots of TiO 2 and MnO 2 , respectively. The first cycle of the TiO 2 cyclic performance ( Figure S4a) exhibited intense peaks of oxidation and reduction reactions at 2.18 and 1.57 V, respectively. Their intensity was reduced in the following cycles due to the poor cyclic stability of TiO 2 . Moreover, the SEI formation in the range from 0.01 to 0.56 V could be seen. Figure S4b presents the cyclic performance of MnO 2 . The first cycle displayed two peaks in the discharging/lithiation region: the first peak, at 0.1 V, corresponded to the reduction of Mn 4+ to Mn 0 [32], and the second peak, at 0.57 V, corresponded to the SEI layer formation. In the charging/delithiation region there was one peak, at 1.27 V, indicating the oxidation of Mn 0 to Mn 2+ . The reduction/oxidation process of manganese dioxide is described by the equation below [32]: Figure 8 displays the first, third and fifth cycles of the galvanostatic charge/discharge curves of each film at 50 mA/g, recorded between 0.005 and 3.0 V. The starting discharge/charge capacities were 411/388, 400/372, 402/387 and 482/470 mAh/g for rGO, rGO/TiO 2 , rGO/MnO 2 and rGO/TiO 2 /MnO 2 , respectively. However, in the following cycles, rGO/MnO 2 discharge/charge capacity exceeded that of rGO. In the subsequent cycles each material exhibited a stable discharge/charge capacity, displaying the excellent reversibility and outstanding cycling stability of rGO-based films. The slight increase in the capacities of the films was caused by the agglomeration of TiO 2 and the small amount of MnO 2 that was also concealed by rGO. Figure S5 presents the discharge/charge profiles of pristine TiO 2 and MnO 2 . The initial discharge/charge capacity of TiO 2 was 64/52 mAh/g and 972/896 mAh/g for MnO 2 . The latter displayed better cyclic stability than TiO 2 and a much higher capacity. During charging, intercalating lithium expands any active material particles, which pulverizes the electrode. This phenomenon was extremely noticeable in MnO 2 . Therefore, despite its great initial capacity, the MnO 2 electrode required a buffer to stabilize the active material and extend the electrochemical performance. For this reason, we proposed combining MnO 2 with GO in the form of a film, so that GO (later rGO) could shield MnO 2 against overexpansion.
It is crucial to examine the anode performance at lower-to-higher current densities to comprehend the current capacity of the active material. Hence, all films were submitted to electrochemical testing at different current densities from 50 to 10,000 mA/g (Figure 9). Each electrode exhibited a more stable cyclic performance at higher current densities. For instance, at 50 mA/g the specific discharge capacity of the rGO/TiO 2 /MnO 2 film (Figure 9d) decreased from 482 to 433 mAh/g, while at 200 mA/g it decreased from 272 to 260 mAh/g. Moreover, when cycled back at a lower current density of 50 mA/g, the cell regained the specific capacity of a current density of 100 mA/g. Such satisfactory reversible performance may have been due to the great degree of graphitization of rGO. Moreover, it is worth noting that after the addition of TiO 2 nanoparticles (Figure 9b) the specific capacity of the rGO film increased at higher current densities, while the addition of MnO 2 particles (Figure 9c) elevated the specific capacity of the rGO film at lower current densities. Such a great capacity of the films was due to the highly graphitized rGO which allowed smooth lithium diffusion through the porous and slit-like morphology. Figure S6 shows the rate performance of pristine TiO 2 and MnO 2 at different current densities. TiO 2 displayed a poor specific capacity and an inferior rate performance when current densities over 50 mA/g were applied, decreasing to 0 mAh/g over 400 mA/g. However, when cycled back at 50 mA/g the capacity exceeded that of 100 mA/g, which was not the case for rGO-based films. The unusually poor performance of TiO 2 may have been attributed to the strong tendency of TiO 2 nanoparticles to agglomerate, which in turn limited the active surface where Li + could be attached. Figure S6b exhibits the rate performance of MnO 2 . This material manifested a high specific discharge/charge capacity at 50 mA/g ranging from 972/918 mAh/g to 887/868 mAh/g for the first and the fifth cycle, respectively.
Moreover, it presented an outstanding cyclic stability and its rate performance was slightly lower than that of the films. Furthermore, the cycles repeated at 50 mA/g displayed slightly lower specific capacities.  It is crucial to examine the anode performance at lower-to-higher current densities to comprehend the current capacity of the active material. Hence, all films were submitted to electrochemical testing at different current densities from 50 to 10,000 mA/g ( Figure 9). Each electrode exhibited a more stable cyclic performance at higher current densities. For instance, at 50 mA/g the specific discharge capacity of the rGO/TiO2/MnO2 film ( Figure  9d) decreased from 482 to 433 mAh/g, while at 200 mA/g it decreased from 272 to 260 mAh/g. Moreover, when cycled back at a lower current density of 50 mA/g, the cell regained the specific capacity of a current density of 100 mA/g. Such satisfactory reversible performance may have been due to the great degree of graphitization of rGO. Moreover, it is worth noting that after the addition of TiO2 nanoparticles ( Figure 9b) the specific capacity of the rGO film increased at higher current densities, while the addition of MnO2 particles (Figure 9c) elevated the specific capacity of the rGO film at lower current densities. Such a great capacity of the films was due to the highly graphitized rGO which allowed smooth lithium diffusion through the porous and slit-like morphology. Figure S6 shows the rate performance of pristine TiO2 and MnO2 at different current densities. TiO2 displayed a poor specific capacity and an inferior rate performance when current densities over 50 mA/g were applied, decreasing to 0 mAh/g over 400 mA/g. However, when cycled back at 50 mA/g the capacity exceeded that of 100 mA/g, which was not the case for rGObased films. The unusually poor performance of TiO2 may have been attributed to the strong tendency of TiO2 nanoparticles to agglomerate, which in turn limited the active surface where Li + could be attached. Figure S6b exhibits the rate performance of MnO2. This material manifested a high specific discharge/charge capacity at 50 mA/g ranging from 972/918 mAh/g to 887/868 mAh/g for the first and the fifth cycle, respectively. Moreover, it presented an outstanding cyclic stability and its rate performance was slightly lower than that of the films. Furthermore, the cycles repeated at 50 mA/g displayed slightly lower specific capacities. EIS measurements were conducted on the half-cells over a range of frequencies from 100 kHz to 10 mHz and the recorded Nyquist plots are shown in Figure 10. This technique helped to further interpret Li-ion storage exploitation. The rGO (Figure 10a), rGO/TiO2 (Figure 10b) and rGO/MnO2 (Figure 10c) plots display a single semicircle and a steep di-  EIS measurements were conducted on the half-cells over a range of frequencies from 100 kHz to 10 mHz and the recorded Nyquist plots are shown in Figure 10. This technique helped to further interpret Li-ion storage exploitation. The rGO (Figure 10a), rGO/TiO 2 ( Figure 10b) and rGO/MnO 2 (Figure 10c) plots display a single semicircle and a steep diagonal line (Warburg line). However, the rGO/TiO 2 /MnO 2 (Figure 10d) plot exhibits an additional, weaker semicircle and a gentler slope. The equivalent circuit diagrams used to calculate the EIS results and the fitting parameters are presented in Table 2. R s (solution resistance), R SEI (SEI layer resistance) and R ct (charge-transfer resistance) were established by the fitted lines shown in Figure 10. The R SEI was not present in the rGO, rGO/TiO 2 and rGO/MnO 2 samples, which indicated the smoother lithium-ion migration through the SEI layer. In comparison, for rGO/TiO 2 /MnO 2 the R SEI value (181.1 Ω) implied a slightly more resilient Li + migration through the SEI interface. Remarkably, the R ct (chargetransfer resistance) of rGO (228 Ω) was the lowest among the films which signified the highest electric conductivity. A gentler Warburg line in the rGO/TiO 2 /MnO 2 electrode suggested a larger Li + diffusion barrier (thus, slower Li-ion diffusion kinetics) compared to the former materials. Moreover, all semicircles were incomplete which implied poor electronic transport. Additionally, Figure S7 shows the Nyquist plots of pristine TiO 2 and MnO 2 . The TiO 2 electrode ( Figure S7a) behaved similarly to the rGO-based anodes. An incomplete semicircle indicated mediocre electron transfer and a steep Warburg line indicated efficient lithium-ion diffusion through the active material. Due to an increased amount of carbon black, the MnO 2 ( Figure S7b) electrode exhibited lower R ct (75.23 Ω) in contrast to 156.9 Ω of TiO 2 . Moreover, the semicircle was completed, which implies much better electron transfer due to an excess of conductive carbon black.      Figure 10.

Conclusions
In summary, the free-standing and non-active-material-free rGO, rGO/TiO 2 , rGO/MnO 2 and rGO/TiO 2 /MnO 2 films were assembled using vacuum-assisted filtration. rGO/TiO 2 /MnO 2 attained a stable reversible capacity (353 mAh/g after 50 cycles), great cyclic stability and an outstanding rate performance. The SEI layer formed during discharging/charging was thin enough to allow smooth lithium diffusion, increasing the kinetics of the electrochemical reactions. Furthermore, reduced graphene oxide acted as the conductive agent, as well as the active material, making the electrode free of the binding agent. Additionally, the free-standing nature of the films did not require the current collector and their preparation did not utilize any harmful substances. As a result, no excess mass nor volume were reported. It is important to note that the films displayed flexibility, making them promising candidates as electrodes in rechargeable batteries for flexible electronics.
Supplementary Materials: The following are available online at https://www.mdpi.com/article/ 10.3390/en14238168/s1, Figure S1: TEM images of (a,b) MnO2 and (c,d) TiO2, Figure S2: XRD diffractograms of TiO2 and MnO2, Figure S3. Raman spectra of TiO2 and MnO2, Figure S4. Cyclic voltammetry of pristine (a) TiO2 and (b) MnO2, Figure S5. Discharge-charge profiles of (a) TiO2 and (b) MnO2 at a current density of 50 mA/g, Figure S6. Rate performance of (a) TiO2 and (b) MnO2 at different current densities, Figure S7. Nyquist plots of (a) TiO2 and (b) MnO2 after discharge/charge cycles and the equivalent circuit diagrams of the cells, Table S1. Fitted results of the equivalent circuit in Figure S7.