Features of electrochemical behavior of graphene films on metal foams

ABSTRACT

Electrochemical behaviour of 4.8 ± 0.2 nm graphene films on nickel and copper foams was investigated by cyclic voltammetry (CV). The graphene films were prepared by chemical vapor deposition and characterized by electron energy loss spectroscopy, elastic peak electron spectroscopy, scanning electron microscopy and X-ray photoelectron spectroscopy. The CV of the [Ru(NH₃)₆]^+2/+3 redox reaction was performed using these substrates. The obtained results demonstrated a high continuity of the deposited graphene film and independence of the electron transfer rate on them from the metal substrate used. The rate of outer-sphere electron transfer on the graphene surface appeared to be substantially less than that on the polished glassy carbon. Partial splitting of graphene layers due to wedged action of 2D adsorption layers, formed by camphor solution, led to the increase in double-layer capacitance.

KEYWORDS:

• 3D-graphene
• copper foam
• nickel foam
• glassy carbon
• electron transfer rate
• adsorption

Introduction

Recently, high-capacity rechargeable energy storage systems have been considered a power source for hybrid electric vehicles and all kinds of power tools [1,2]. Electrochemistry of graphene-coated electrodes has been the subject of significant number of publications in recent years. However, there are very few works dealing with their electrochemistry in a strict sense of this term [3–8]. This fact is, probably, caused by sufficient experimental difficulties in the manufacturing of electrodes from graphene monolayers. Some works consider the structures containing 3–100 layers of graphene, which are placed on a conductive substrate. Such material is known to have the following advantages for electrode production over other carbon nanomaterials: absence of catalyst trace amounts, high electrochemical activity and possibility to gain colloidal solutions within the electrochemical process. Finally, this material is cheap and easy to produce. However, there are two conflicting suggestions in the literature concerning the reactivity of the graphene sheets, their edges and basal planes with respect to the electrode reactions. In most of the studies, starting with a ‘classical’ paper by Banks, Compton and others [9–11], significantly higher electron transfer rate for boundary and defect areas compared to graphene plane is reported. It is explained by significantly higher density of states near the edge of defect regions than for less defective basal planes. Nevertheless, there is a different viewpoint [12–15] based on the results of electrochemical studies.

Based on transmission electron microscopy data, multilayer graphene derivatives grown on metal foams during chemical vapor deposition (CVD) process can possess a 3D structure caused by cross-linking of the layers [16]. It is considered to be a promising material for solar cells electrodes and electrochemical sensors [17], as well as for high-performance supercapacitors [18,19]. At the same time, the actual electrochemistry of 3D-graphenes remains underexplored. In this work, cyclic voltammetry (CV) was used to study the electrochemical behaviour of 3D-graphene deposited on nickel and copper foams by CVD. Electron transfer velocity was determined from [Ru(NH₃)₆]^+2/+3 redox reaction, which is ideal for studying the outer-sphere redox systems [20]. It is widely used as a benchmark for comparing the reactivity of carbon coatings with different morphologies [12] and not sensitive to electrode surface quality. Redox reaction of [Fe(CN)₆]^3-/4- was not applied in the present study because of surface contamination [12]. The analysis of samples was performed by electron energy loss spectroscopy, elastic peak electron spectroscopy, scanning electron microscopy (SEM) and X-ray photoelectron spectroscopy.

Experimental

3D-graphene was deposited on nickel (3DGrNi) and copper (3DGrCu) foams in a CVD reactor from a mixture of argon, hydrogen and methane as reported earlier [21]. For both metals, graphene coatings were almost invisible. The thickness of the graphene films calculated from the monolayer (~3%) light absorption was estimated to be less than 30 nm. Figure 1 demonstrates the SEM images of graphene coating on both metals. Fairly smooth continuous structures without sharp edges are presented. The surface morphology is similar for both coatings and metal basis.

Figure 1. SEM images of graphene coatings on the Ni (a–c) and Cu (d–f) foams.

Spectra of the induced electron emission were measured with hemispherical energy analyser SPECS Phoibos 225 with a 2D Charge Coupled Device detector (CCD) (SPECS GmbH, Germany). X-ray photoelectron spectra (XPS) were excited by monochromatic X-rays (1486.6 eV) using a SPECS XR-50 X-ray tube and measured at the so-called ‘magic angle’ of 55° (the angle between the X-ray beam axis and the energy analyser entrance slit axis). Elastic peak electron spectra (EPES) were measured at an excitation energy of 8 keV and a scattering angle of 120°. The electron energy loss spectra in reflection geometry (REELS) were measured at an excitation energy of 0.5 keV. In this case, the axis of the entrance slot of the energy analyser coincided with the direction of the normal to the sample surface. The EPES and REELS were induced using Kimball Physics EMG 4212 electron guns with BaO cathodes. All experiments have been done in ultra-high-vacuum (5 × 10⁻¹⁰ mbar) conditions. Before analysis, the samples were annealed in vacuum at 350°C for 15 min for the rest of gases and surface contamination removal.

Voltammetry measurements (CV) were made in a three-electrode cell with P-30S Potentiostat (Elins, Russia) under the scanning speed (v) of 0.005–1.000 V/s. GC plate was an auxiliary electrode, and a saturated calomel electrode was used as a reference. All potentials are given relative to the reference electrode. Currents for cathode and anode peaks were determined by the direct extrapolation of initial values.

Solutions for CV were prepared using multiple distilled water. Sodium sulphate (99.9% Reakhim, Russia) was twice recrystallized and calcined at 500°C before experiments. [Ru(NH₃)₆]Cl₃ (99.9% Merck) and camphor (99.9% Merck) were used as delivered. To remove dissolved oxygen, solutions under investigation were blown out with argon for 1 h. Standard measurements were done with polished GC disc of 3 mm diameter. For measurements, it was pressed into a Teflon cylinder and polished with 0.3 µm abrasive [20,22].

Results and discussion

Figure 2a shows typical XPS of the graphene coating on Ni foam. Similar data were observed for the copper substrate. Copper and nickel peaks have low intensities in both cases that correspond to significant thickness of the graphene. Oxygen intensity is negligible because of the low functionalization carbon layers by oxygen-containing groups due to their cross-linking.

Figure 2. Electron spectroscopy of graphene coating on Ni substrate: a) XPS, b) EPES and c) REELS.

Elastic peak electron spectroscopy (EPES) was used to determine the thickness of the coating [22]. Figure 2b shows the typical EPES spectrum of graphene on the nickel sample. Intense elastic peak is splitting into separate peaks for carbon and nickel, respectively. Primary energy is 8 keV, while scattering angle is 120°. The scanning depth of EPES is known to be proportional to inelastic mean free pass (IMFP). The intensity of elastic peak is proportional to differential elastic cross-section. As a result, for known material composition, one can define the thickness of the coating. Elastic cross-section was taken from National Institute of Standards and Technologies (NIST) database while IMFP was taken from TPP-2 M formula [23]. The calculation from integral peaks intensities from carbon and nickel resulted in graphene coating with an average thickness of 4.8 ± 0.2 nm.

Electron energy loss spectrum was measured with primary energy of 500 eV in the energy range of 50 eV below elastic peak. Figure 2c shows that energy of π- and (π + σ)-plasmons (energy loss of 6.4 and 26 eV, respectively) of the graphene coating is close to graphite ones. However, unlike well-known graphite’s EELS, the spectrum under investigation contains a peak with 19 eV energy loss. It can be explained by the existence of areas in 3D-graphene with substantially weakened van der Waals interaction between the graphene layers.

Figure 3 depicts CV response of 3D-graphene on the nickel and copper foams to [Ru (NH₃)₆] ^+2/+3 in comparison to a polished glassy carbon (GC) electrode, registered with a scan rate of 0.01 V/s. One can clearly observe the exact match of formal potentials E₀ = (E_C + E_A)/2 (E_C and E_A – potentials of cathode and anode peaks for three measured electrodes). In this case, E₀ value is close to equilibrium potential of the redox reaction, which is 201 ÷ 206 mV for all 3DGrNi, 3DGrCu and GC. Over-voltage in the redox reaction of [Ru (NH₃)₆]^+2/+3 depends on the nature of the electrodes. Thus, such precise equality of E₀ value is likely to indicate a high enough continuity of the 3D-graphene coating on both copper and nickel foams. It is additionally supported by the absence of the significant faradic current for graphene on Ni electrode in the background electrolyte (Figure 3, curve 4). Therefore, there is no direct electric contact between the solution and the metal surface. These results are in good agreement with those of Prasai et al. [24], where authors describe high-quality anticorrosion graphene coatings deposited on Cu and Ni foils.

Figure 3. Cyclic voltammetric response of 3DGrNi (1) and 3DGrCu (2) electrodes in comparison to polished GC (3), 1 М solution Na₂SO₄ + 5 · 10⁻³ М [Ru(NH₃)₆]Cl₃, ν = 0.01 V/s. 4 – cyclic voltammetric response of 3DGrNi electrode in 1 М solution of Na₂SO₄.

Electron transfer at ν ≤ 50 mV/s corresponds to the standard criteria of reversible electrode reaction, which is controlled by a diffusion of depolarizer into the electrode surface [25]. Figure 4 shows a symmetrical shape of the curves with equal magnitudes of cathode (I_c) and anode (I_a) currents. The subplot in Figure 5 shows an almost linear dependence of I_a and I_c to ν^0.5. Potential difference (ΔEp) between peaks is about 64 ÷ 68 mV, which proves the reversible nature of electron transfer. For scan rate higher than 0.05 V/s, potential difference ΔE_p also increases (Figure 5), which indicates the change-over to mixed kinetics [25]. In this case, the point where transition of ν occurs can be used to estimate the constant of heterogeneous electron transfer k₀ that corresponds to 0.002 ÷ 0.003 cm/s ([Ru (NH₃)₆]^+2/+3) for both cases. The value of k₀ is close to the values determined earlier [26] for multilayer defectless graphene, obtained by CVD on Ni substrate and then transferred to the insulating substrate. At the same time, the value of k₀ is higher than the one calculated after oxygen surface functionalization in acetone [27]. As demonstrated earlier [28], k₀ of metal electrodes in the case of [Ru (NH₃)₆]^+2/+3 redox reaction weakly depends on the nature of the metal and corresponds to 0.51 ÷ 1.24 cm/s. This is significantly higher than the value measured for 3D-graphene that indicates a high continuity of coating. Figure 5 also shows the dependence of ΔE_r (ν) for redox reaction with GC in [Ru(NH₃)₆]^2+/+3. For GC, one can estimate k₀ > 0.07 cm/s, which is consistent with the data [26]. Thus, the k₀ value for 3D-graphene was noticeably smaller than that for GC. Therefore, one can conclude that the velocity of electron transfer on basal planes is slower than that on the edges, defect sites or the surface of the polished GC [29].

Figure 4. Cyclic voltammetric response at variable scan rates for 3DGrNi (a) and 3DGrCu (b) electrodes. Subplots show dependencies of I_а и I_c vs. ν^0.5 (1 М solution Na₂SO₄ + 5 · 10⁻³ М [Ru(NH₃)₆]Cl₃).

Figure 5. Potential difference ΔE_р at different scan rates of 3DGrNi (1) and 3DGrCu (2) electrodes and polished GC (3).

The ability of graphenes deposited on metal foams to split was estimated from the change of the cyclic voltammogram under the presence of the surfactant – camphor. Being a compound that forms 2D condensed layers on surfaces [30], it causes intense lateral interaction between molecules and supports the formation of adsorbate monolayers with low bulk concentration (10⁻²–10⁻³ M). Figure 6 shows that the charging rate of a differential double layer capacity (C) increases 1.2 ÷ 1.8 times in the presence of camphor, which corresponds to increase in C. It is important to mention that the adsorption of this surfactant has only physical nature, which is supported by the dynamics of cyclic voltammogram in the electrolyte solution. The shape of the curve is approximately the characteristic shape for electrolyte without additives (e.g. camphor).

Figure 6. Cyclic voltammogram of 3DGrCu electrode in 1 М Na₂SO₄ (1) solution with 8 · 10⁻³ М camphor (2), scan rate – 0.01 V/s.

It is known that camphor possesses a strong surface activity on electrodes from carbon nanotube bundles [30]; due to adsorption instead of a decrease, a several times growth of capacitance was fixed, which can be explained by the Rebinder effect. Thus, the energy of graphene surface is reducing, when surfactant adsorption layer is deposited on the surface of nanotubes bundle. As a result, bundles partially disintegrate, which significantly increases the nanotubes surface accesible to the electrolyte. With the transition to carbon allotrope with smaller splitting ability (e.g. nanowires), capacitance demonstrates a decrease. A similar effect was also described by Shang et al. [31]. Thus, one can observe that camphor adsorption leads to partially reversible splitting of 3D-graphene.

Conclusions

This work demonstrates the possibility of CVD synthesis of 4.8 ± 0.2 nm graphene coatings on foamed nickel and copper, forming an entire 3D network from carbon layers. Their electrochemical behaviour was found to be similar to multilayer low-defect graphene described by Prasai et al. [24]. A high continuity of the deposited graphene film and independence of the electron transfer rate from the metal substrate used were fixed. The k₀ rate of redox reaction [Ru(NH₃)₆]^+2/+3 for electrodes from them was found to be substantially less than that for polished GC. It was shown that the presence of specific surfactant in the water solution leads to partial splitting and the corresponding increase in graphene surface area exposed to electrolyte solution.

Acknowledgments

The present work was financially supported by the Russian Ministry of Science and Higher Education, Federal Assignment Nos. АААА-А21-121011990019-4 and AAAA-A19-119032690060-9.

Disclosure statement

No potential conflict of interest was reported by the authors.

Additional information

Funding

The work was supported by the Ministry of Science and Higher Education of the Russian Federation [АААА-А21-121011990019-4] and the Ministry of Science and Higher Education of the Russian Federation [AAAA-A19-119032690060-9].