One-pot synthesis of manganese oxide/graphene composites via a plasma-enhanced electrochemical exfoliation process for supercapacitors

Highly ordered pyrolytic graphite 99.99%, potassium permanganate (KMnO₄), and potassium hydroxide (KOH) were purchased from Sigma-Aldrich. Deionized (DI) water was used as the solvent for all experiments. All chemicals and reagents were used directly as received without further purification.

Our experimental setup is schematically illustrated in figure 1. 150 ml of KOH (3 %) and 0.0237 g of KMnO₄ (1 mM) at a pH of approximately 12 were used as the electrolyte solution. The electrolyte solution was ultra-sonicated for 10 min to gain a homogeneous solution. A highly oriented pyrolytic graphite rod with a diameter of 6 mm and a length of 150 mm was used as the cathode calibrated above the solution and a 25 × 10 mm Pt plate used as the anode electrode was submerged deeply into the solution. The cathode tail was sharpened to 1 mm in diameter. Two electrodes were coupled under the maximum voltage of 70 V from a DC bias (110 V-3 A, Kikusui Electronics Corporation, Japan). The current density was stabilized in the range of 0.4–0.6 A by adjusting the distance of the cathode and the solution interface with a mechanical scissor lift. The reaction beaker placed in an ultrasonic cleaner operating at the frequency of 50/60 kHz and a power of 280 W. The ultrasonic bath was assisted by a cooling water-flow system to avoid any sudden rise in the surrounding temperature due to a long period of plasma activation.

Zoom In Zoom Out Reset image size

After 60 min, the product was collected by vacuum filtration through a polyvinylidene fluoride (PVDF) membrane with a pore size of 0.2 µm after being washed with DI water and ethanol to reach a neutral pH of 7. The final material was then dried at 80 °C for 24 h and stored in a drying box at room temperature with a weight of about 40 mg. The samples were denoted as G, GM1, GM5, and GM10 for the initial KMnO₄ precursor solution concentrations of 0, 1, 5, and 10 mM, respectively. A similar preparation process was conducted with two Pt electrodes in 5 mM of KMnO₄ solution to synthesize manganese oxide individually, and is denoted as sample M.

SEM images were recorded on a Hitachi S-4800 field emission SEM. SEM samples were prepared by drop-drying from a water/ethanol suspension of the composite material on a Si wafer. TEM (Hitachi 9500, 300 kV) was conducted using specimens dispersed in ethanol and then dropped onto a copper microgrid coated with a holey carbon film, followed by the evaporation of the ethanol.

The crystallographic structure of the samples was determined by a D2 XRD system (Equinox 5000) equipped with a Cu Kα radiation and a Ni filter (λ = 0.1542 nm).

Raman spectra were recorded by a LabRam HR Evolution with a laser wavelength of 532 nm. XPS measurements were performed on an S-Probe TM2803 spectrometer and fitted with OriginLab Pro 9.0 software. Quantitative analysis of the XPS spectra was carried out using Shirley background subtraction before performing a least-square-error fit with a mixture of Gaussian and Lorentzian line shapes.

Electrochemical measurements of the prepared samples were recorded via cyclic voltammetry (CV), galvanostatic charge-discharge (GCD), and electrochemical impedance spectroscopy (EIS) on a Metrohm Autolab PGSTAT 302 N with a standard three-electrode cell system composed of a working electrode, a Ag/AgCl reference electrode (in 3 M of KCl solution), and a Pt counter electrode utilizing 1 M of Na₂SO₄ aqueous solution as the electrolyte. A homogeneous slurry was prepared by mixing the active material sample, acetylene black, and PVDF at the weight ratio of 80:10:10 in a small amount of N-methyl-2-pyrrolidone solvent as a viscous paste. The working electrode was prepared by coating 20 µl of as-prepared slurry onto a Ni foam (0.25 × 0.25 cm), followed by drying in a vacuum heater at 60 °C for 2 h. The loading mass of the active materials was about 0.64 mg cm^–2. CV tests were carried out in the range of –0.6 to 0.4 V (vs. Ag/AgCl) at different scan rates of 10, 20, 40, 60, 80, and 100 m V^–1. GCD measurements were performed at different current densities (1.0, 2.0, 3.0, 4.0, and 5.0 A g⁻¹) in the potential range of −0.6–0.4 V. EIS measurements were implemented in the frequency range from 100–0.1 Hz at an open circuit potential with an AC perturbation of 10 mV. The specific capacitance C_s (F g^–1) of the active materials were calculated from CV and GCD curves using C_s = (∫I.dt)/(mΔV) and C_s = (IΔt)/(mΔV), respectively where I is the discharge current (A), Δt is the discharge time (s), ΔV is the potential window (V), and m is the weight of the electroactive materials in the working electrodes (g).

In a previous work, we synthesized molybdenum disulfide nanoparticles from a redox process of [MoS₄]²⁻anions, which contains both molybdenum and sulfur elements, with KOH and (NH₄)₂SO₄ as the electrolyte solution [24]. Thus, at first, a similar hypothesis was proposed to synthesize manganese oxide because potassium permanganate precursor also carries two vital elements in the desired products. After the first-step reduction, permanganate anion is reduced into manganate, which theoretically will continue to be reduced into manganese oxides. However, unlike the MoS₂ with high chemical resistivity, the MnO_x is unstable in the basic environment of KOH as an electrolyte solution. The manganese oxides tentatively recombine with oxygen gas from either the reaction's product or the surrounding atmosphere to reform manganate ions. This equilibrium prevents the formation of MnO_x nanoparticles.

$$\begin{align*}2MnO_4^ - \,\mathop \to \limits^{}\, \,MnO_4^{2 - } + Mn{O_x} + \frac{{4 - x}}{2}{O_2}\notag\\MnO_4^{2 - }\,\mathop \leftrightarrow \limits^{KOH} \,Mn{O_x} + \frac{{4 - x}}{2}{O_2} + \,2{e^ - }.\end{align*} \tag{ 1 }$$

The case here is much different from thiomolybdate salt since there was no chemical loop in the formation of the MoS₂ nanoparticles. Acidic solvents such as H₂SO₄, HCl, or HNO₃ cannot be used here as they could react with manganese oxide to form the respective dissolvable salts in the final solution. Our only option, therefore, is to dilute the strong basic electrolyte into a weaker one. However, if the electrolyte is diluted too much, the electrical conductivity of the solution becomes too low, which weakens the electrical field, and then deters the discharge plasma process. Ultraviolet-visible spectroscopy was utilized in order to keep track of the manganate anions. Unless there is a decline in the concentration of K₂MnO₄ after PE³P, a smaller amount of MnO_x particles is formed. Based on the observed results, the optimized concentration of KOH was 3% after many attempts from 10%, as in previous work, down to 1%, the point when no plasma occurs.

At the cathode, not just MnO_x particles are formed, but graphene is peeled off from the graphite electrode as well. The mechanism of the graphite exfoliation is due to two key factors: the formation of intercalated hydrogen bubbles and the electromagnetic force of negatively co-charged graphitic layers at the tip head (figure 1).

Being simultaneously co-synthesized, MnO_x particles tend to attach to graphene surface simply by a van der Waals force. There is no chemical bonding between these two components.

The compounds synthesized from the redox PE³P were found to contain graphene and a type of manganese oxide. MnO_x, as we assumed it while the variable x could be any number in the range of 0–3 is found to be MnO₂ based on the XRD, Raman and XPS results. Based on the data in JCPDS 18–0802, manganese oxide has the characteristic peaks of δ-MnO₂ with (001), (002), (110), and (119) planes indexed at 12.3°, 24.8°, 37.1°, and 65.1° respectively with the average size of 12.4 nm calculated by a Scherrer equation (figures 2(a) and (b)). These peaks are identical to the inter-atomic distance of the edge-sharing MnO₆ octahedron unit in the crystal structure with a basal spacing of 0.72 nm. The XRD patterns also reveals an intensive narrow peak at 26.6° corresponding to the (002) basal reflection of graphite with aninterlayer spacing of 0.34 nm (figures 2(c) and (d)) [25]. In term of Raman spectra, graphene has three characteristic peaks at 1352, 1578, and 2695 cm^–1 for the D, G, and 2D bands, respectively. The D band represents the disordered structure of graphene, and the G band corresponds to the stretching sp² bonds of the carbon atoms. Thus, the low intensity of the D band and high one of the G band suggests that the prepared sample has a low defect and the large in-plane sp² domains of C–C bonding. The ratio of I_D/I_G remained unchanged in both samples suggesting that the addition of the MnO₂ precursor was insignificant in the formation of the graphene itself.

Zoom In Zoom Out Reset image size

The chemical configuration of MnO₂ was confirmed by the analysis of Raman spectra with three identical peaks located at 490, 579, and 635 cm^–1 (figure 3). In comparison to Zhang et al work [26], manganese oxide peaks shifting to the higher values is expected due to the presence of graphene as the dominant material in the composite. The 490 cm^–1 peak is considered a characteristic feature of δ-MnO₂. The 579 cm^–1 band is attributed to the Mn—O lattice vibrations while the appearance of the 635 cm^–1 band represents an intermediate formation of manganese oxide during the decomposition of permanganate [27, 28].

Zoom In Zoom Out Reset image size

The analysis of the XPS data showed the presence of carbon, manganese, and oxygen elements in the composite sample. The peaks at around 642.5 and 653.5 eV are attributed to the doublet Mn^IV 2p_3/2 and 2p_1/2, respectively (figure 4(b)). Compared to Yang et al work, these peaks are slightly shifted positively [29]. The chemical shift in the XPS spectra is possibly due to the change in the binding energy of the core electrons corresponding to the local oxidation state of the main element. Therefore, we suggest that besides the strong interaction of MnO₂ with the graphene substrate, the high electronegativity of oxygen doped continuously to Mn^I, Mn^II, and Mn^III in the manganate equilibrium resulted in the slip in electron density, leading to a small rise in the binding energy of manganese as in Mn^IVO₂. Figure 4(c) illustrates the deconvolution of C 1 s spectrum consisting of C = C (284.5 eV) and C–C (285.3 eV) as characteristics of sp² and sp³ carbon hybridization in graphitic materials [25]. Some oxygen defects were detected in C–O/C = O at 286.3 eV, which belongs to the oxygenated groups such as hydroxyl, carbonyl, and epoxy in a graphene oxide structure [30]. In spite of the insignificant amount of GO that could not be determined from the XRD data, we hypothesize that edged carbon atoms in a few graphene sheets were likely burnt up to the closest oxidative stages due to the abundance of activated oxygen generated in the reduction of permanganate and the surrounding air under the ultra-high temporary temperature at the plasma point.

Zoom In Zoom Out Reset image size

Graphene nanoplatelets denoted as the G sample, with no manganese oxide, attained a wide lateral size with about ten layers observed as shown in figures 5(b) and (f). Not only the size of the MnO₂ particles increases but also thedistribution of its density thickens on the graphene sheets when the initial concentration of KMnO₄ increases from 1.5%–10% respectively in the GM1, GM5, and GM10 samples (figures 5(c)–(e)). This behavior is similar to that in our previous work on the synthesis of a MoS₂/graphene composite. The higher concentration of the precursor causes a higher amount of nanoparticles created at the same time, which then results in bulk aggregation due to the similarity in their chemistry. In TEM images of sample GM1, MnO₂ is observed with an average size of 10–20 nm decorated on the surface of multi-layer graphene nanoplatelets (figures 5(f)–(i)).

Zoom In Zoom Out Reset image size

Figure 6 shows the CV curves for the G, GM1, GM5, GM10, and M samples at varied scan rates within a similar potential range of −0.6 to 0.4 V. The CV curve of G (figure 6(a)) is approximately rectangular and symmetric in shape, even at a high scan rate, implying an excellent capacitive behavior and fast charge-discharge process. This could be attributed to the great access of the electrolyte to the electroactive surface of the electrode materials. The combination of graphene with MnO₂ in the composite samples resulted in the deviation of their CV curves from the rectangular shape (figures 6(b)–(d)). However, no redox peaks are observed, as in the case of the MnO₂ sample (figure 6(e)), indicating pseudocapacitive characteristics. At the same scan rate, all composite samples have a much higher internal area than the neat G (graphene) and M (MnO₂) samples, indicating a considerable enhancement of the capacitance of the composite electrodes. The possible origins of this synergistic effect can be attributed to (i) an increase in the electrical conductivity due to the presence of graphene; (ii) the larger surface area of MnO₂ nanoparticles anchored on graphene; and (iii) the promotion of electron transport between the electrode and electrolyte solution. This phenomenon has also been observed previously in MnO₂-carbon composites as well as other hybrid materials [31–33]. As shown in figure 6(f), the specific capacitance calculated at any scan rate is strongly dependent on the composition of the composite samples, and decreases in the order of GM1 > GM5 ≈ GM10 > M > G. At the scan rate of 10 mV s^–1, the C_s of GM1 is 320 F g⁻¹, while that of the G, GM5, GM10, and M samples are 34, 193, 198, and 120 F g⁻¹, respectively.

Zoom In Zoom Out Reset image size

Figure 7(a) shows the galvanostatic charge-discharge profiles of all samples at the same current density of 1 A g⁻¹. The discharge time of the samples follows the order of G < M < GM10 < GM5 < GM1, indicating that the C_s values of all the composite electrodes are higher than those of the neat graphene and MnO₂ ones. The GM1 sample exhibits the highest C_s value, harmonizing with the results of the CV tests. Here, the composite materials obtained using a higher concentration of KMnO₄ exhibit lower capacitance, probably due to the larger size of MnO₂ particles formed on the graphene support.

Zoom In Zoom Out Reset image size

The GCD profiles of the GM1 sample at different current densities are presented in figure 7(b), which demonstrates the pseudocapacitance nature of the composite materials. Also, the IR drop in the discharge curves considerably increases with current density, suggesting a relatively large internal impedance of the composite electrodes. Therefore, to reduce the internal impedance and enhance the capacitive performance, a higher content of conductive additives should be used. Figure 7(c) shows the C_s values of the GM1 sample calculated from its discharge process and plotted versus the current density. The C_s values of the GM1 sample are 286.5 and 89.5 F g⁻¹ at current densities of 1 and 5 A g⁻¹, respectively. The outstanding electrochemical performance of GM1 as compared to the other composite materials is possibly attributed to its high electrical conductivity and optimal coverage of MnO₂ particles on the graphene support. The C_s value of GM1 is reasonably high compared to various MnO₂-carbon composite systems prepared by different methods, as shown in table 1.

As presented in figure 7(d), the Nyquist plots of all the electrode materials exhibit a linear form in the low-frequency region, indicating typical capacitive behavior. The slope of those linear parts of all the composite electrodes is higher than that of the bare MnO₂, indicating faster kinetics and that a more reversible Faradic reaction occurred on their surface. The equivalent series resistance (R_s) of the electrode material is determined by the intercept of its Nyquist plot with the real axis in the high-frequency region. This is the combination of the ionic resistance of the electrolyte, the internal resistance of the electrode material, and the contact resistance of the electrode material with the current collectors. It was found that the R_s values of all the composites slightly rose with the increase of MnO₂ content. In other words, the R_s values of all the composites are lower than that of the pristine MnO₂, indicating the decisive role of graphene in enhancing the electrical conductivity of the composite. Whereas, the diameter of the semicircle increases with MnO₂ content, indicating a rise in the charge transfer resistance (R_ct) in composite electrodes. It is likely because of the presence of more MnO₂ particles which inhibit the access and exit of electrolytes during the intercalation/deintercalation process. The Nyquist plots were fitted with an equivalent electrical circuit using Nova software, and the details of the circuit parameters are given in table 2.