Initial Study of Reduced Graphene Oxide Foams Modified by Mn and Bi as Capacitive Electrode Materials

Abstract

In a view of application of porous materials in wearable electronics and self-powered systems, reduced graphene oxide (rGO) foams modified by Mn or/and Bi were produced in this study to be used as electrodes for supercapacitors. The hydrothermal method and the freeze-drying processes were used for the preparation of the materials further morphologically, elementally and structurally analyzed. Based on the electrochemical characterization, Bi-modified rGO foam was found to be more a promising material for capacitive electrodes in comparison to the other prepared materials.

1. Introduction

Electronic double-layer capacitance (EDLC) and/or pseudocapacitance are widely used in capacitive electrode materials for energy storage. Carbonaceous materials such as activated carbon and graphene have a large surface area, which can lead to the accumulation of charges due to the adsorption and desorption of electrolyte ions at the electrode/electrolyte interface [1,2,3,4,5]. In theory, graphene can be an ideal material to fabricate positive electrodes for supercapacitors due to its rich electrochemistry, high electrical conductivity (5 × 10⁻³ S/cm), and large specific surface area (2630 m²/g) [6,7], which can provide a maximum specific capacitance value of 550 F/g [8]. However, the specific capacitance values of graphene and reduced graphene oxide are unfortunately reported in the literature to be just 80–264 F/g in aqueous electrolytes [9,10,11,12].

At the same time, pseudocapacitive materials can have higher capacitances and therefore higher energy densities due to the Faraday process. However, usually, such pseudocapacitive materials are transition metal oxides with low electrical conductivity [13]. Among various pseudocapacitive materials such as nickel oxide (NiO) [14] and cobalt oxide (Co₃O₄) [15], manganese-based oxides (MnO₂, Mn₃O₄) [16] were studied as electrode materials for energy storage for a long time because of their high electrochemical performance, low cost, natural abundance, and low toxicity [17,18,19,20,21,22]. However, MnO₂ powder also has a poor electrical conductivity and a low surface area, resulting in the actual specific capacitance being far below the theoretical specific capacitance of 1370 F/g [23]. Therefore, the synergetic effect between graphene materials with high electrical conductivity, providing a conductive platform to accelerate electron transport, and pseudocapacitive materials with low conductivity but a high specific capacitance can significantly enhance the electrochemical properties of an obtained composite [24,25].

On the other hand, Bi₂O₃ has not been studied as thoroughly as other metal oxides in combination with rGO for energy storage applications. Only a few research groups have reported on the preparation and investigation of rGO with Bi-based materials for energy storage [26,27]. However, most of the studies of rGO-Bi₂O₃ presented a visible plateau during the galvanostatic charge–discharge (GCD) test, which is typical for battery-type materials, together with increased specific capacity [28,29]. The same plateau in GCD was reported for composites of graphene with BiVO₄ [30].

Furthermore, the structure and properties of electrode materials for energy storage were reported in the literature to be modified in composites of graphene-family materials combined with several metals or their oxides simultaneously, i.e., Ni-Co-rGO [31] and NiO/MnO₂@nitrogen-doped graphene oxide [32]. However, the storage mechanism of the prepared materials was not always clearly explained.

In this work, we preliminarily studied the influence of Mn or/and Bi on the electrochemical properties of an rGO foam prepared by the one-step hydrothermal method, followed by freeze-drying and annealing.

2. Materials and Methods

2.1. Preparation of Graphene Oxide Solution

A light-brown graphene oxide (GO) solution (2 g per 1 L) was prepared based on the Hummers method, as reported in a previous work [33]. In short, commercially available natural flake graphite, sulfuric acid (99.99%), potassium permanganate (99%), hydrogen peroxide (30 wt% in H₂O), and hydrochloric acid (37%) were used as received from Sigma-Aldrich (Steinheim, Germany). Deionized distilled water was used to prepare all the aqueous solutions. GO was synthesized using 50 mL of 98% sulfuric acid (placed in a 250 mL chemical glass flask immersed in a water bath with ice) and 2 g of graphite powder. Then, 6 g of potassium permanganate was added to the reactant mixture, which was heated and diluted with water and later with 10 mL of 3% hydrogen peroxide. After that, the mixture was filtered with a paper filter under evacuation. The filtered-out material was mixed with distilled water and filtered again several times. The GO was separated from the reaction products using a centrifuge. The final product was a light-brown GO solution.

2.2. Preparation of Foam of Reduced Graphene Oxide Without and with Mn and Bi

To obtain the rGO foam, the GO solution was placed in an autoclave at 180 °C for 3 h (marked hereafter as rGO). After that, the black cylindric foam was removed from the autoclave and freeze-dried. Manganese (II) nitrate tetrahydrate (Mn(NO₃)₂ × 4H₂O, 98.5%, MERCKS, Darmstadt, Germany) or bismuth (III) acetate (C₆H₉BiO₆, 99%, Alfa Aesar, Karlsruhe, Germany) was used for the preparation of the corresponding solutions with the mass ratio GO:Mn or GO:Bi as 10:1. All components were mixed and magnetically stirred for 12 h at room temperature. The manganese nitrate was preliminary dried at 30 °C. In addition, for the preparation of foams with bismuth acetate, 10 mL of acetic acid (CH₃COOH, 99%, Sigma-Aldrich, Steinheim, Germany) was added to improve the solubility of the acetate. A mass ratio of 10:1:1 was used in the case of rGO-Mn-Bi. To prepare the rGO foam containing Mn (marked hereafter as rGO-Mn) or rGO with Bi (marked hereafter as rGO-Bi) or rGO with both Mn and Bi (marked hereafter as rGO-Mn-Bi), the corresponding solutions were used for the hydrothermal process for 3 h at 180 °C and freeze-dried in a similar manner.

2.3. Characterization

The morphology of the materials under study was analyzed using scanning electron microscopy (SEM, Hitachi TM4000Plus, Tokyo, Japan). Compositional analysis was conducted with an energy-dispersive spectroscopy (EDS) system (QUANTAX 75/80, Bruker, Ettlingen, Germany) under an acceleration voltage of 10 kV of a scanning electron microscope.

The X-ray diffraction (XRD) profiles of the rGO-based foams were collected at room temperature in a continuous scanning mode (step 0.02°, time 10 s) on a Rigaku SmartLab (Tokyo, Japan) diffractometer in the 2θ range from 10 to 70°.

The electrochemical measurements of all the materials were performed in a 1M Na₂SO₄ electrolyte after coating it onto a glassy carbon electrode. The cycling voltammograms (CVs) were performed with an AUTOLAB PGSTAT 302N potentiostat (Metrohm AG, Riverview, USA) in a three-electrode arrangement with the glassy carbon electrode coated with the dispersed studied material to be tested as a working electrode, a platinum wire as a counterelectrode, and a silver|silver sulfate (KCl sat.) electrode as a reference (EAg|Ag₂SO₄|KCl sat = 0.444 V vs. SCE and 0.70 V vs. NHE at 22 °C). The working electrodes were prepared as follows: a studied material was dispersed in distilled water (concentration of 2 mg/mL), and then, 10 μL of the well-homogenized dispersion was dropped on the surface of the glassy carbon electrode and left for drying at room temperature.

GCD measurements were conducted at current densities of 0.1, 0.2, 0.5, and 1 A/g, whereas CV measurements were carried out at voltage sweep rates of 20, 50, and 100 mV/s by cycling the potential of the working electrode between 0 and +1 V. The results of the CV and GCD tests were collected after the stabilization of the cycles.

Electrochemical impedance spectroscopy (EIS) characterization was performed at an open circuit voltage with 5 mV in amplitude using the PGSTAT 302 potentiostat equipped with a FRA2 module.

3. Results and Discussion

SEM images of the foams of the non-modified rGO and rGO modified with Mn, Bi, and Mn together with Bi are shown in Figure 1a–d, respectively. The rGO foam modified with Mn (Figure 1b) is seen to be the densest. At the same time, the rGO with both Mn and Bi also looks almost nontransparent and dense (Figure 1d).

The presence of Bi and Mn in the rGO foams was verified using EDS, as shown in Figure 2.

The XRD results are presented in Figure 3. A clear peak at ~24.8° can be seen for the non-modified rGO foam in Figure 3a, which implies that the graphene oxide is fully reduced after the hydrothermal process at 180 °C. At the same time, the XRD patterns of the composite foams show the coexistence of several phases. Peaks that are characteristic of α-MnO₂ (JCPDS card no. 00-053-0633) (200), (310), and (211) hkl-planes and MnO (JCPDS card no. 01-075-6876) (111), (200), and (220) planes are detected in Figure 3b for the graphene oxide that was hydrothermally reduced at 180 °C after the addition of manganese nitrate tetrahydrate (Mn(NO₃)₂ × 4H₂O). The XRD profile of the graphene oxide that was hydrothermally reduced at 180 °C after the addition of bismuth acetate presented in Figure 3c reveals strong diffraction peaks at ~27.5°, 31.6°, 32.8°, 46.3°, 47.1°, 54.3°, 55.7°, and 58.1°, assigned to the (201), (002), (220), (222), (400), (203), (421), and (402) hkl-planes of β-Bi₂O₃ (JCPDS card no. 27-0050). At the same time, a sharp peak at 2θ of ~11.7° and small peaks at ~24.2°, 36.6°, 41°, and 49.8° can be associated with Bi₂O₂CO₃ (JCPDS card no. 06–0249) as a subproduct of bismuth acetate (C₆H₉BiO₆) and graphene oxide, obtained after the hydrothermal process at 180 °C. The rGO foam modified with both Mn nitrate and Bi acetate are presented in Figure 3d, with XRD peaks related to the samples of the rGO with Mn and the rGO with Bi.

Figure 4 shows the cyclic voltammograms of the rGO-based materials coated onto a glassy carbon electrode and measured in Na₂SO₄ at different scan rates with a potential window from 0 to +1 V. Although the CV shape of the rGO-based electrode in a positive potential window can be different (smaller) from that measured in a negative potential window [34], here, the non-modified rGO presents a combination of the typical capacitive behavior of an EDLC material and the pseudocapacitance impact, which can be seen as a peak at ~+0.4 V in Figure 4a. The addition of Mn to rGO increases the CV area, keeping the pseudocapacitance contribution as shown in Figure 4b. A higher increase in the CV area is observed after the addition of Bi to rGO in Figure 4c, also presenting a strong pseudocapacitance impact both at low and high scan rates. In contrast, as seen in Figure 4d, the simultaneous addition of Mn and Bi to rGO leads to a smaller CV area in comparison to that of the rGO with Bi or even the rGO with Mn, also diminishing the pseudocapacitive contribution. The specific capacity for rGO was already reported to increase after the addition of Mn [21] due to an enlarged specific surface area, although the elevated specific surface area does not always lead to an increase in the specific capacitance [35,36,37,38]. However, there has been no report so far on the influence of Bi on the value of the specific surface area of rGO. On the other hand, the real value of the specific capacitance must be calculated from the GCD, allowing only for a qualitative analysis of the CV.

Figure 5 shows the GCD curves of the rGO-based materials at different discharge current densities. The lowest charge–discharge time is obtained for the non-modified rGO (Figure 5a), increasing with the addition of Mn (Figure 5b), Bi (Figure 5c), and particularly them both (Figure 5d). However, only the GCD curves of rGO have the obviously symmetrical triangular shape that is typical for EDLC materials, allowing us to calculate the specific capacitance using Equation (1):

$$C_{sp}={\displaystyle \frac{I\times ∆t}{m\times ∆V}}$$

(1)

where I is the applied current, m is the mass of the active electrode material, ΔV is the potential range, and Δt is the discharge time in GCD. The GCD of the rGO with Mn (Figure 4b) or Bi (Figure 4c) or both Mn and Bi (Figure 4d) must be calculated by using Equation (2) with an integral area:

$$C_{sp}={\displaystyle \frac{1}{m\times v\times (V_2-V_1)}}{\int }_{V_1}^{V_2}I\left(V\right)dV$$

(2)

where ∫V(t)dt is the integral area, t(V_max) is the initial time at which the potential has maximal value, and t(V_min) is the final time at which the potential has minimal value. The calculated specific capacitance values are found to decrease strongly with current density increase. The highest specific capacity of ~100 F/g at 1 A/g is obtained for the rGO with Bi and is almost twice as high as that of the non-modified rGO. The specific capacitances of the rGO with Mn only and rGO with both Mn and Bi are similar, reaching ~75 F/g at 1 A/g. Such not very high values, in contrast to the theoretical value of the specific capacitance of graphene, can be related to the agglomeration of rGO layers and its low chemical activity in Na₂SO₄ electrolytes, requiring further deep study.

As the behavior of the GCD of rGO with both Mn and Bi (Figure 5d) is far from the ideal triangular shape, the dominant storage mechanism in such a material can be of a capacitive or faradaic/battery-type. Therefore, the parameter b needs to be estimated from the slope of the were obtained from the CV at scan rates v of 20, 50, and 100 mV/s, and at a fixed potential chosen to be 0.4 V since the i peak can be observed in the diapason of 0.35–0.45 V. The obtained linear fits and equations describing them, typically as y = a + bx, are also presented in Figure 6a. Based on the fits, the highest parameter b value of 0.84 was obtained for the non-modified rGO, decreasing to 0.8 for the rGO with Bi and further to 0.78 for the Mn-modified rGO and 0.77 for the rGO with both Mn and Bi. Pseudocapacitive materials with 0.8 < b < 1 are considered to have predominantly capacitive storage, whereas electrodes with 0.5 < b < 0.8 have dominant faradaic (battery-type) behavior [39,40,41,42].

In addition, the EIS results are presented in Figure 6b as a Nyquist plot. The non-modified rGO, as well as the rGO with Bi, reveals similar pseudocapacitive behavior in the low and medium frequency range. At the same time, the rGO with Mn presents less vertically aligned and hence more resistive behavior. Moreover, the rGO simultaneously modified with both Mn and Bi presents the least vertically aligned variation, approaching a slope of 45°, which is typical for battery-type materials and indicates the impact of a faradaic/intercalation storage mechanism on the main pseudocapacitive storage [43].

Although energy and power densities should not be calculated from measurements in a three-electrode system, since they are not equal to a two-electrode cell of the real capacitor device, the possible energy density (E in W × h/kg) and power density (P in W/kg) are estimated here based only on the galvanostatic charge–discharge curve using the following expressions:

$$E={\displaystyle \frac{1}{2\times 3.6}}{C}_{sp}{V}^2$$

(3)

$$P={\displaystyle \frac{E}{∆t}}$$

(4)

where C_sp is the measured electrode capacitance, V is the operating voltage window (1 V), and Δt is the discharge time in hours. An energy density of ~13 W × h/kg with a power density of ~4 kW/kg are achieved in the rGO-Bi foam material and are the highest values among the other studied materials here: ~6.4 W × h/kg and ~0.6 kW/kg for the unmodified rGO, ~10 W × h/kg and ~2.1 kW/kg for the rGO with Mn only, and ~10.5 W × h/kg and ~2.8 kW/kg for the rGO modified with both Mn and Bi.

4. Conclusions

In this work, Mn, Bi, and both Mn with Bi were used to modify reduced graphene oxide foams, which were structurally and morphologically studied in comparison with a non-modified foam. The EDLC electrodes obtained by the coating of these materials on a glassy carbon electrode were also electrochemically assessed in a three-electrode system in a Na₂SO₄ electrolyte. The specific capacitance was in the range of 70–100 F/g at 1 A/g, with the highest value obtained for the rGO modified with Bi only, even though the rGO with both Mn and Bi presented not only capacitive behavior but also some faradaic/intercalation storage. Therefore, these results indicate that among these materials, the rGO with Bi is the most promising for application in wearable electronics and self-powered systems and must be studied thoroughly.