Synergistic Interaction of Clusters of Iron Oxide Nanoparticles and Reduced Graphene Oxide for High Supercapacitor Performance

Supercapacitors have been recognized as one of the more promising energy storage devices, with great potential use in portable electronics and hybrid vehicles. In this study, a composite made of clusters of iron oxide (Fe3O4-γFe2O3) nanoparticles and reduced graphene oxide (rGO) has been developed through a simple one-step solvothermal synthesis method for a high-performance supercapacitor electrode. Electrochemical assessment via cyclic voltammetry, galvanostatic charge–discharge experiments, and electrochemical impedance spectroscopy (EIS) revealed that the Fe3O4-γFe2O3/rGO nanocomposite showed much higher specific capacitance than either rGO or bare clusters of Fe3O4-γFe2O3 nanoparticles. In particular, specific capacitance values of 100 F g−1, 250 F g−1, and 528 F g−1 were obtained for the clusters of iron oxide nanoparticles, rGO, and the hybrid nanostructure, respectively. The enhancement of the electrochemical performance of the composite material may be attributed to the synergistic interaction between the layers of graphene oxide and the clusters of iron oxide nanoparticles. The intimate contact between the two phases eliminates the interface, thus enabling facile electron transport, which is key to attaining high specific capacitance and, consequently, enhanced charge–discharge time. Performance evaluation in consecutive cycles has demonstrated that the composite material retains 110% of its initial capacitance after 3000 cycles, making it a promising candidate for supercapacitors.


Materials Synthesis
The synthesis of clusters of iron oxide nanoparticles was conducted via the solvothermal method using a previously reported synthetic protocol [32]. Typically, 2.5 mmol (0.678 g) of iron(III) chloride hexahydrate were dissolved in 20 mL of ethylene glycol, and mechanically stirred to the formation of a clear solution, followed by the addition of sodium acetate (1.8 g) and polyethylene glycol (PEG 6000) (8.5 wt. %). The mixture was stirred vigorously for 30 min and then sealed in a stainless steel autoclave (100 mL). The autoclave was heated up to 185 • C (heating rate = 5 • C min −1 ) and kept at this temperature for 8 h under stirring (1500 rpm). Finally, the mixture was allowed to cool down to room temperature and the black powder was collected using a magnet. The resulting nanostructure was washed several times with Milli-Q water, and dried at 60 • C overnight.
Reduced graphene oxide-iron oxide (Fe 3 O 4 -γFe 2 O 3 /rGO) hybrid nanocomposites were synthesized following a previously reported method with some modifications [33]. Briefly, 0.05 g of GO sheets were dispersed in 20 mL of an EG:DEG mixture (EG:DEG = 1:19 vol.%) under ultrasonication. Subsequently, 1 g of sodium acrylate, 0.678 g of iron chloride hexahydrate, and 1.8 g of sodium acetate were added into the suspension, followed by the addition of 1.7 g (8.5 wt. %) of PEG. The resulting mixture was vigorously stirred until a homogeneous dispersion was obtained followed by the solvothermal treatment in a stainless steel autoclave using the same conditions as described for the preparation of clusters of iron oxide nanoparticles. In this case, sodium acrylate acts as a stabilizer preventing the aggregation and sodium acetate assists in the reduction of FeCl 3 to Fe 3 O 4 by altering the alkalinity [33]. The as-prepared composite was collected using a magnet and washed several times with ethanol and water, and then dried at 60 • C overnight. A control sample (Fe 3 O 4 -γFe 2 O 3 /rGO-C) was prepared using 0.015 g of GO sheets and following the same experimental procedure. Similarly, pure rGO was synthesized and used as a reference.

Materials Characterization
X-ray diffraction (XRD) patterns were collected using a Siemens D-5000 powder X-ray diffractometer operated in Bragg Brentano geometry using Ni-filtered Cu Kα radiation (λ = 0.1541 nm). Data were recorded in the 2θ range 10-80 • with an angular step size of 0.026 • and a counting time of 1 s per step. The collected data were refined using the Le Bail method by means of the software Rietica [34,35]. Raman spectra were collected from powder samples onto a glass slide as substrate, with a Renishaw in Via Reflex Raman microscope (Renishaw, Gloucestershire, UK). Experiments were conducted at room temperature using 532 and 633 nm excitation wavelengths. Field emission scanning electron microscopy (FESEM) was conducted on a JEOL JSM-6700 F. Samples for transmission electron microscopy (TEM) studies were prepared by dropping a diluted suspension of the samples onto ultra-thin carbon-coated copper grids. Imaging was performed on a JEOL JEM 1010 instrument operated at 100 kV and equipped with a CCD camera (JEOL, Tokyo, Japan).

Electrochemical Tests
The electrochemical measurements were conducted in a standard three-electrode cell configuration using a PalmSens4 ® potentiostat. The electrochemical cell was filled with 25 mL of a 3 M KOH solution prepared with Milli-Q water. A Pt wire spiral and a Ag/AgCl (3.0 M KCl) were used as the counter and reference electrode, respectively. The working electrodes were prepared as described elsewhere [36]. In total, 0.3 mg of the sample was dispersed in 24 µL of ethanol and then 6 µL of the prepared solution was drop-cast onto a carbon gas diffusion layer (1 cm × 1 cm; Sigracet 39BB) four times followed by the addition of 3 µL of 5 wt. % perfluorinated Nafion TM resin solution, which acts as a binder. Cyclic voltammetries, galvanostatic charge-discharge scanning experiments at different scan rates, and cyclic stability tests at a 50 mV s −1 scan rate (3000 cycles) were conducted in the −0.1 to +0.35 V potential window. Electrochemical impedance spectroscopy (EIS) data were obtained at the respective open circuit potential, from 50 kHz down to 0.1 Hz, taking 10 frequencies per decade, with ±5 mV sinusoidal voltage excitation.

Sample Characterization
The clusters herein studied were obtained through a solvothermal method. The formation mechanism proceeds via a two-stage growth process, with nucleation of the primary nanocrystals followed by uniform aggregation into larger secondary structures [37], whose size is primarily determined by the polyethylene glycol concentration in the reaction medium. Figure 1 and Figure S1 (see Supplementary Materials) include the TEM and SEM images of these clusters of iron oxide nanoparticles and of the iron oxide/reduced Nanomaterials 2022, 12, 2695 4 of 14 graphene oxide hybrid. TEM micrographs confirm the formation of iron oxide nanoparticles grouped in clusters of controlled size as corroborated by the size distribution analysis, which evidenced the formation of the clusters with an average diameter of 70 ± 19 nm (Gaussian fit, inset in Figure 1b). Complementarily, SEM images reveal the formation of the individual clusters of iron oxide nanoparticles in between the rGO layers. whose size is primarily determined by the polyethylene glycol c tion medium. Figures 1 and S1 (see Supplementary Materials) in images of these clusters of iron oxide nanoparticles and of the phene oxide hybrid. TEM micrographs confirm the formation of grouped in clusters of controlled size as corroborated by the si which evidenced the formation of the clusters with an average (Gaussian fit, inset in Figure 1b). Complementarily, SEM images the individual clusters of iron oxide nanoparticles in between the The XRD results of the samples are compiled in Figure 2a. T displays a broad peak at 2θ = 25° that corresponds to the ( Fe3O4-γFe2O3 nanoparticles are composed of crystallites of smal structure corresponds to either magnetite (Fe3O4), maghemite (γ them, as both iron oxides present the same spinel structure (on w or Fe 3+ cations and vacancies are arranged, respectively). Similarly along with the broad band between 20° and 30° associated with hybrid structure. Indeed, the intensity of the latter is quite low, w highly disordered structure because of oxidation and the low c light on the nature of the iron oxide species in the samples, we c analysis using Raman spectroscopy (Figure 2b). This technique e different vibrations of the crystalline lattice due to different cation and can therefore differentiate the two magnetic iron oxide phas of the clusters of Fe3O4/γ-Fe2O3 nanoparticles displays four main sociated to the A1g vibration mode of the magnetite (at 668 cm −1 centered at 350 cm −1 , 500 cm −1 , and 700 cm −1 , corresponding to th of maghemite, respectively [40][41][42]. Other characteristic features rial, such as the black coloration of the powders and their strong magnetic fields, were also observed. In the case of the rGO-ba The XRD results of the samples are compiled in Figure 2a. The diffractogram of rGO displays a broad peak at 2θ = 25 • that corresponds to the (002) plane. Clusters of Fe 3 O 4 -γFe 2 O 3 nanoparticles are composed of crystallites of small sizes whose crystalline structure corresponds to either magnetite (Fe 3 O 4 ), maghemite (γ-Fe 2 O 3 ), or a mixture of them, as both iron oxides present the same spinel structure (on which Fe 3+ or Fe 2+ cations or Fe 3+ cations and vacancies are arranged, respectively). Similarly, these diffraction peaks along with the broad band between 20 • and 30 • associated with rGO are detected in the hybrid structure. Indeed, the intensity of the latter is quite low, which might be due to its highly disordered structure because of oxidation and the low content. In order to shed light on the nature of the iron oxide species in the samples, we completed the structural analysis using Raman spectroscopy ( Figure 2b). This technique enables us to register the different vibrations of the crystalline lattice due to different cationic arrangements [38,39], and can therefore differentiate the two magnetic iron oxide phases. The Raman spectrum of the clusters of Fe 3 O 4 /γ-Fe 2 O 3 nanoparticles displays four main bands, which can be associated to the A 1g vibration mode of the magnetite (at 668 cm −1 ) and three broad bands centered at 350 cm −1 , 500 cm −1 , and 700 cm −1 , corresponding to the T 2g , E g , and A 1g modes of maghemite, respectively [40][41][42]. Other characteristic features of the synthesized material, such as the black coloration of the powders and their strong interaction with external magnetic fields, were also observed. In the case of the rGO-based hybrids, the Raman spectrum displays the two characteristic bands of rGO, i.e., D and G at 1348 cm −1 and 1594 cm −1 , respectively, along with the characteristic intensity band ca. 680 cm −1 associated with the A 1g modes of the Fe 3 O 4 (668 cm −1 ) and γ-Fe 2 O 3 (700 cm −1 ) magnetic phases. Nanomaterials 2022, 12, x FOR PEER REVIEW 5 of 14

Electrochemical Measurements
The electrochemical performance of the samples was investigated by cyclic voltammetry, galvanostatic charge-discharge scanning experiments at different scan rates, and electrochemical impedance spectroscopy. Figure 3 shows the cyclic voltammetries obtained from −0.1 V to +0.35 V in 3 M KOH at different scan rates ranging from 2 mV s −1 to 50 mV s −1 . The specific capacitance (Cs) (F g −1 ) is calculated from the cyclic voltammetry curves using Equation (1) [36], in which Δ corresponds to the difference in the peak oxidation and reduction currents in amperes (A), m is the mass loading in grams (g), and is the scan rate, in V s −1 .
The values obtained for a scan rate of 2 mV s −1 scan rate were 100 F g −1 , 250 F g −1 , and 528 F g −1 for the clusters of Fe3O4-γFe2O3 nanoparticles, the rGO sheets, and the Fe3O4-γFe2O3/rGO hybrid structure, respectively ( Figure 3). Chen et al. reported a specific capacitance of 262.1 F g −1 for Fe3O4/rGO composites at the same scan rate and using 1 M Na2SO3 as the electrolyte [43], and Qi et al. reported a value of 350.6 F g −1 for Fe3O4/rGO composites at a scan rate of 1 mV s −1 and using 6 M KOH as the electrolyte [44]. On the other hand, Sheng et al. reported a graphene/Fe3O4 nanocomposite that exhibits a specific capacitance of 268 F·g −1 at 2 mV·s −1 using 1 M Na2SO4 as the electrolyte [25], and the fabrication of magnetite (Fe3O4)-decorated carbon nanotubes with a specific capacitance of 145.4 F g −1 at 2 mV·s −1 using 0.5 M Na2SO4 as electrolyte was reported by Nawwar et al. [11]. Additional data are included in Table S1 for comparative purposes. The energy density is 14.85 W h kg −1 at a power density of 1116.5 W kg −1 , which is in line with previously reported results [15,45]. The clusters of iron oxide nanoparticles show the lowest specific capacitance of the three materials under study, which could be ascribed to the low conductivity of the iron oxide phases and/or poor electrolyte access to the surface of the particles. These aspects are improved in the case of rGO sheets, as it can be seen from the comparison of this material with the clusters of iron oxide nanoparticles. In the case of the hybrid structure,

Electrochemical Measurements
The electrochemical performance of the samples was investigated by cyclic voltammetry, galvanostatic charge-discharge scanning experiments at different scan rates, and electrochemical impedance spectroscopy. Figure 3 shows the cyclic voltammetries obtained from −0.1 V to +0.35 V in 3 M KOH at different scan rates ranging from 2 mV s −1 to 50 mV s −1 . The specific capacitance (C s ) (F g −1 ) is calculated from the cyclic voltammetry curves using Equation (1) [36], in which ∆I corresponds to the difference in the peak oxidation and reduction currents in amperes (A), m is the mass loading in grams (g), and υ is the scan rate, in V s −1 .
The values obtained for a scan rate of 2 mV s −1 scan rate were 100 F g −1 , 250 F g −1 , and 528 F g −1 for the clusters of Fe 3 O 4 -γFe 2 O 3 nanoparticles, the rGO sheets, and the Fe 3 O 4 -γFe 2 O 3 /rGO hybrid structure, respectively ( Figure 3). Chen et al. reported a specific capacitance of 262.1 F g −1 for Fe 3 O 4 /rGO composites at the same scan rate and using 1 M Na 2 SO 3 as the electrolyte [43], and Qi et al. reported a value of 350.6 F g −1 for Fe 3 O 4 /rGO composites at a scan rate of 1 mV s −1 and using 6 M KOH as the electrolyte [44]. On the other hand, Sheng et al. reported a graphene/Fe 3 O 4 nanocomposite that exhibits a specific capacitance of 268 F·g −1 at 2 mV·s −1 using 1 M Na 2 SO 4 as the electrolyte [25], and the fabrication of magnetite (Fe 3 O 4 )-decorated carbon nanotubes with a specific capacitance of 145.4 F g −1 at 2 mV·s −1 using 0.5 M Na 2 SO 4 as electrolyte was reported by Nawwar et al. [11]. Additional data are included in Table S1 for comparative purposes. The energy density is 14.85 W h kg −1 at a power density of 1116.5 W kg −1 , which is in line with previously reported results [15,45]. The clusters of iron oxide nanoparticles show the lowest specific capacitance of the three materials under study, which could be ascribed to the low conductivity of the iron oxide phases and/or poor electrolyte access to the surface of the particles. These aspects are improved in the case of rGO sheets, as it can be seen from the comparison of this material with the clusters of iron oxide nanoparticles. In the case of the hybrid structure, in which the clusters of iron oxide nanoparticles are interlayered in between the conductive rGO sheets, the specific surface area is expected to increase, thus increasing the electrolyte permeation through the rGO layers [31]. Figure 3d summarizes the specific capacitance values as a function of the scan rate. For a uniformly accessible surface, the capacitance should not depend on the scan rate, since the variations in the scan rates should result in equal magnitude variations of the difference in the peak oxidation and reduction currents given in Equation (1) [46]. However, the data presented in Figure 3d point to an unevenly accessible surface. The specific capacitance decreases exponentially with the increase in the scan rate, which points towards a decrease in the active surface. This can arise from the fact that, while at lower scan rates the electric field is able to penetrate the whole electrode structure, for higher scan rates only the outer part of the electrode material participates in the charging/discharging process [47]. in which the clusters of iron oxide nanoparticles are interlayered in between the conductive rGO sheets, the specific surface area is expected to increase, thus increasing the electrolyte permeation through the rGO layers [31]. Figure 3d summarizes the specific capacitance values as a function of the scan rate. For a uniformly accessible surface, the capacitance should not depend on the scan rate, since the variations in the scan rates should result in equal magnitude variations of the difference in the peak oxidation and reduction currents given in Equation (1) [46]. However, the data presented in Figure 3d point to an unevenly accessible surface. The specific capacitance decreases exponentially with the increase in the scan rate, which points towards a decrease in the active surface. This can arise from the fact that, while at lower scan rates the electric field is able to penetrate the whole electrode structure, for higher scan rates only the outer part of the electrode material participates in the charging/discharging process [47]. To confirm this hypothesis, EIS measurements were carried out on three selected systems: bare clusters of Fe3O4-γFe2O3 nanoparticles, the control sample Fe3O4-γFe2O3/rGO-C, and the Fe3O4-γFe2O3/rGO nanocomposite material. Figure 4 includes representative EIS Nyquist plots obtained for the systems tested, as well as the equivalent circuit employed for the data modeling. The Nyquist impedance plot of Fe3O4/γ-Fe2O3 clearly shows two domains in the low-frequency region, a capacitive arc section between 10 and 1 Hz, and a straight line with a 45° slope at the lowest frequencies. Moreover, the inset in Figure 4a, which corresponds to the high-frequency limit, clearly shows a depleted capacitive arc starting at about a 45° slope. This shape is characteristic of porous electrodes with de Levie type impedance behavior [48]. Figure 4b compiles the Nyquist impedance plot of the Fe3O4-γFe2O3/rGO hybrid material. Although the shape of the diagram is very different from that of the clusters of iron oxide nanoparticles, the same conclusions can be extracted if overlapping in the time constants in the high and medium frequency ranges is considered. The depletion of the high capacitive arc remains close to 45°. The porous structure for Fe3O4-based systems was described first for magnetite scales in heat exchangers [49], To confirm this hypothesis, EIS measurements were carried out on three selected systems: bare clusters of Fe 3 O 4 -γFe 2 O 3 nanoparticles, the control sample Fe 3 O 4 -γFe 2 O 3 /rGO-C, and the Fe 3 O 4 -γFe 2 O 3 /rGO nanocomposite material. Figure 4 includes representative EIS Nyquist plots obtained for the systems tested, as well as the equivalent circuit employed for the data modeling. The Nyquist impedance plot of Fe 3 O 4 /γ-Fe 2 O 3 clearly shows two domains in the low-frequency region, a capacitive arc section between 10 and 1 Hz, and a straight line with a 45 • slope at the lowest frequencies. Moreover, the inset in Figure 4a, which corresponds to the high-frequency limit, clearly shows a depleted capacitive arc starting at about a 45 • slope. This shape is characteristic of porous electrodes with de Levie type impedance behavior [48]. Figure 4b compiles the Nyquist impedance plot of the Fe 3 O 4 -γFe 2 O 3 /rGO hybrid material. Although the shape of the diagram is very different from that of the clusters of iron oxide nanoparticles, the same conclusions can be extracted if overlapping in the time constants in the high and medium frequency ranges is considered. The depletion of the high capacitive arc remains close to 45 • . The porous structure for Fe 3 O 4 -based systems was described first for magnetite scales in heat exchangers [49], and later applied to other systems such as particulate zinc-rich coatings [50], conversion coatings for Li-ion batteries [51], or aged passive layers in concrete [52]. and later applied to other systems such as particulate zinc-rich coatings [50], conversion coatings for Li-ion batteries [51], or aged passive layers in concrete [52].
The low-frequency feature, i.e., the 45° tilt in the Nyquist plot of Figure 4a, is characteristic of a planar diffusion. The nanostructured electrodes behave in this way as flat elec trodes due to the overlapping of the concentration profiles [53]. A relevant parameter for this research is the effect of rGO on the electrical conductivity of the nanostructured layer of particles. The EIS data can provide such information via a suitable electrical equivalen circuit that accounts for the physical phenomena above discussed.  The low-frequency feature, i.e., the 45 • tilt in the Nyquist plot of Figure 4a, is characteristic of a planar diffusion. The nanostructured electrodes behave in this way as flat electrodes due to the overlapping of the concentration profiles [53]. A relevant parameter for this research is the effect of rGO on the electrical conductivity of the nanostructured layer of particles. The EIS data can provide such information via a suitable electrical equivalent circuit that accounts for the physical phenomena above discussed. Figure 4c includes a scheme of the electrical equivalent circuit employed to model the experimental EIS data. R m accounts for the resistivity (electronic conduction) of the solid phase, i.e., the clusters of iron oxide nanoparticles or the nanocomposite material, and R s corresponds to the resistivity (ionic conduction) associated with the electrolyte filling the layer pores. Z 1 and Z 2 account for the interfacial impedances at the pore wall and bottom, respectively, and Z d accounts for the impedance associated with ionic diffusion to and from the electrode in the conditions of the quiescent solution employed.
The impedance functions Z 1 , Z 2 , and Z d are defined as follows: tan h jωτ d (5) in which R 1 and R 2 correspond to the charge transfer resistances at the pore wall and bottom, respectively; C 1 and C 2 are the corresponding parallel double layer capacitances; R d represents the diffusion resistance; τ d is the time constant; j = √ −1, and ω = 2πf, in which f represents the frequency. α 2 accounts for the Cole-Cole-type dispersion of the R 2 C 2 time constant, which is associated to heterogeneities at the pores bottom. The R 1 C 1 time constant in Equation (3) requires α 1 to improve the fitting.
Equations (2) and (3) illustrate the remarkable difference between the clusters of iron oxide nanoparticles and the nanocomposite material. The former shows conductive behavior (Equation (2)), possibly ascribed to the redox transformation between Fe 3 O 4 and γFe 2 O 3 (2Fe 3 O 4 + 2OH − ↔ 3γFe 2 O 3 + H 2 O + 2e − ) [54]. In contrast, the hybrid exhibits blocking interfacial behavior (Equation (3)). The diffusion impedance defined in Equation (5) probably relates to the flow of OH − species involved in the redox process and is not present in the hybrid structure, which is consistent with the blocking character of the pore walls that hinder the ionic flow.
The selected electrical equivalent model is able to reproduce accurately the experimental data, as shown in Figure 4a,b, which enables an in-depth analysis of the EIS spectra. The parameters compiled in Table 1 show that the only relevant difference between the bare Fe 3 O 4 -γFe 2 O 3 nanostructure and Fe 3 O 4 -γFe 2 O 3 /rGO-C seems to be, as expected, the electronic conductivity of the materials, which increases by more than two orders of magnitude upon the introduction of rGO (R m decreases from 11,200 to 23 Ω cm). This increase is accompanied by minor changes in the porosity of the pore network, as R s increased only from 16 to 24 kΩ cm. The active surface increased by one order of magnitude, as indicated by the change in C 1 from 7.1 to 63.1 mF cm −3 . Concerning the low-frequency part of the spectra, the diffusion process from the electrolyte to the solid material, or vice versa, seems to slow down in the presence of rGO (τ d increases from 0.23 to 17.7 s and R d from 38.9 to 247.4 Ω cm 2 ). This could be ascribed to the increased diffusion length due to the faster charge transfer associated with the higher active surface and higher conductivity of the Fe 3 O 4 -γFe 2 O 3 /rGO-C. Considering the thickness of the diffusion layer (δ), as δ = √ τD and diffusivity D = 10 −5 cm 2 s −1 , this parameter increases from 15 to 130 µm due to the presence of rGO. Table 1. Fitting parameter values obtained for the three systems investigated using the equivalent circuit depicted in Figure 4c. The thickness L of the nanostructured systems was fixed at 3.5 µm. R 0 corresponds to the resistance of the electrolyte outside the sample, i.e., between the reference electrode and the surface of the sample, which depends on the position of the reference electrode with respect to the sample. In the case of the composite material, although the Nyquist impedance plot is different from that of either the clusters of iron oxide nanoparticles or the control sample Fe 3 O 4 -γFe 2 O 3 /rGO-C, the electrical equivalent model is very similar, indicating an analogous structure of the composite layer. The latter exhibits a higher C 1 value associated to its blocking interfacial behavior. It also displays lower R s and higher R m values. In addition, the pores bottom is well differentiated from the pores wall, as typical double layer capacitance values in conducting interfaces, i.e., in the µF cm −2 range, are obtained for C 2 . It is also noticeable that the impedance spectra do not change significantly after cycling (data not shown), especially for the Fe 3 O 4 -γFe 2 O 3 /rGO hybrid nanocomposite, which is consistent with the performance upon prolonged cycling, as we will discuss later. It is important to note that the model applied in this study to interpret the impedance data (a transmission line model compatible with the material's porous nature) is, to the best of our knowledge, new in the field, and presents a number of advantages: (i) it allows for accessing the resistivity of the active material (via R m ) and its porosity (via R s ); (ii) the kinetic information comes from R 1 and R 2 , differentiating between charge transfer at the pore walls and at the pore bottom, respectively; and (iii) R d accounts for the ionic flux from or towards the solution.
The capacitive performance was also studied by galvanostatic charging-discharging at different current densities in the potential window from −0.1 to 0.35 V ( Figure 5). As observed from Figure 5, the time for charging and discharging decreases as the current density increases. At higher current densities, the accessibility to the electrode porosity decreases because ions from the electrolyte do not have enough time to reach the pores bottom [55]. The charging-discharging time for the Fe 3 O 4 -γFe 2 O 3 /rGO nanocomposite material is much higher than that observed for both the clusters of iron oxide nanoparticles and rGO electrodes, which is in line with the higher capacitance value of the hybrid structure. As discussed above, the intimate interaction between the rGO substrate and the clusters of Fe 3 O 4 -γFe 2 O 3 nanoparticles blocks the interface and enables facile electron transport, which is key to high specific capacitance and, consequently, a high chargedischarge time [28].
The long-term cycling stability of the Fe 3 O 4 -γFe 2 O 3 /rGO nanocomposite material was investigated for up to 3000 cycles on cyclic voltammetry at a scan rate of 50 mV s −1 in the potential window from −0.1 V to +0.35 V. The current evolution and coulombic efficiency as a function of the cycle number are summarized in Figure 6. It can be seen that the current (and hence the capacitance) increases with cycling. That increase is better visualized in the capacitance retention plot, which reaches 110% after 3000 cycles (Figure 6b). The growth of the clusters of iron oxide nanoparticles in between the rGO sheets avoids the aggregation of the oxidic phase, which, along with the improved dispersion of the clusters of Fe 3 O 4 -γFe 2 O 3 nanoparticles owing to their interaction with the rGO substrate, results in the observed performance enhancement. Moreover, the large contact area between the active material and the electrolyte increases the efficiency of charge transport, thus also contributing to the capacitance improvement [56,57]. Although additional studies are needed in order to confirm these hypotheses, the obtained results highlight the poten-tial of the fabricated hybrid nanocomposite to be employed as an electrode material for supercapacitor applications. The long-term cycling stability of the Fe3O4-γFe2O3/rGO nanocomposite material was investigated for up to 3000 cycles on cyclic voltammetry at a scan rate of 50 mV s −1 in the potential window from −0.1 V to +0.35 V. The current evolution and coulombic efficiency as a function of the cycle number are summarized in Figure 6. It can be seen that the current (and hence the capacitance) increases with cycling. That increase is better visualized in the capacitance retention plot, which reaches 110% after 3000 cycles (Figure 6b). The growth of the clusters of iron oxide nanoparticles in between the rGO sheets avoids the aggregation of the oxidic phase, which, along with the improved dispersion of the clusters of Fe3O4-γFe2O3 nanoparticles owing to their interaction with the rGO substrate, results in the observed performance enhancement. Moreover, the large contact area between the active material and the electrolyte increases the efficiency of charge transport, thus also contributing to the capacitance improvement [56,57]. Although additional studies are needed in order to confirm these hypotheses, the obtained results highlight the potential of the fabricated hybrid nanocomposite to be employed as an electrode material for supercapacitor applications.

Conclusions
In this work, a composite material containing clusters of Fe3O4-γFe2O3 nanoparticles and reduced graphene oxide was developed through a simple one-step solvothermal synthesis method as a potential candidate for supercapacitors. The electrochemical study revealed that this hybrid structure, in which the clusters of iron oxide nanoparticles are in-

Conclusions
In this work, a composite material containing clusters of Fe 3 O 4 -γFe 2 O 3 nanoparticles and reduced graphene oxide was developed through a simple one-step solvothermal synthesis method as a potential candidate for supercapacitors. The electrochemical study revealed that this hybrid structure, in which the clusters of iron oxide nanoparticles are interlayered in between the conductive rGO sheets, possesses much higher specific capacitance than the individual elements, used as reference. Specifically, 100 F g −1 , 250 F g −1 , and 528 F g −1 were obtained for the clusters of iron oxide nanoparticles, the rGO itself, and the composite material, respectively, at a scan rate of 2 mV s −1 using 3 M KOH as the electrolyte. In the case of the hybrid structure, the electrolyte permeation through the rGO layers increases, and consequently, so too does the capacitance. A state-of-the-art model was used to interpret the EIS measurements, enabling us to understand the reasons for the improved behavior, which was ascribed to the fact that the intimate interaction between the rGO substrate and the clusters of Fe 3 O 4 -γFe 2 O 3 nanoparticles blocks the interface and enables facile electron transport. The evaluation of the performance of the nanocomposite material in subsequent cycles revealed that the current (and hence the capacitance) increases with the number of cycles, reaching 110% capacitance retention after 3000 cycles. The interaction of the clusters of iron oxide nanoparticles with the rGO substrate ensures a homogeneous dispersion of the oxidic phase embedded in between the rGO sheets, thus resulting in the observed capacitance enhancement. Although more detailed studies are needed in order to confirm these premises, the obtained insights pave the way towards the design of improved supercapacitors with superior performance.