Graphene oxide-multiwalled carbon nanotubes composite as an anode for lithium ion batteries

Graphene oxide (GO) was obtained from natural graphite by modified Hummers method [20] as it was described before [21]. Composite electrodes with typically 9 % of MWNTs were prepared by the direct mixing of GO and MWNTs in N-methylpyrrolidone (NMP) as it was described previously [22]. In details, 60 mL of GO suspension in water with a concentration of 1 mg/mL were centrifuged at 15000 rpm for 70 min. The solution over the precipitate was removed. Preserved precipitate was added to 6 mg of MWNTs (Sigma Aldrich, diameter 110 nm to 170 nm, length 5 μm to 9 μm, purity 90 %) suspended in 3 mL of NMP. The suspension was stirred and heated up to 150 °C until a dense slurry was obtained. The slurry was mixed, dispersed on a 12 mm copper foil discs and then vacuum dried at 120 °C overnight. The mass of electrode was approximately 1 mg. Some of such prepared electrodes were then reduced by the photothermal method [17]. This process was carried out in air atmosphere by focusing a camera photoflash lamp on the composite GO/MWNTs electrode. Distance between the flash and electrode was ~ 1 cm. Photothermal process was carried out 10 times to ensure complete reduction.

The structure of obtained electrodes was characterized by X-ray powder diffraction (XRD) with a Siemens D500 diffractometer with a CuKα radiation (λ = 1.542 Å) source. The morphology of the materials was examined by scanning electron microscopy (SEM) Zeiss EVO 40. The wavelength-dispersive X-ray spectroscopy (WDS) was performed using Quanta FEG 250 (FEI) SEM equipped with low energy X-ray spectrometer LEXS (EDAX) to analyze samples chemical composition.

The galvanostatic charging/discharging and cyclic voltammetry experiments were carried out using a two-electrode Swagelok® type electrochemical cell with a lithium foil as a counter and reference electrodes. An LP30 (1M LiPF₆ in ethylene carbonate: dimethyl carbonate 1:1) electrolyte (Merck) was used. All the electrodes, prepared exactly as described above, were tested without using any additional binders or conductive additives. The assembly process was carried out in argon atmosphere in a glove box with humidity less than 1 ppm. The galvanostatic charge/discharge measurements were carried out in a voltage range of 0.0 V to 3.0 V with a current densities of 10 mAh·g⁻¹, 30 mAh·g⁻¹, 50 mAh·g^-1, and 100 mAh·g⁻¹. Current density and capacity values were calculated for the mass of GO taken for electrode preparation. The cyclic voltammetry measurements were performed in the same voltage range like galvanostatic experiments, with a scan rate of 0.02 mV·s⁻¹. All electrochemical experiments were done using a VMP3 modular 16 channel potentiostat/galvanostat (Bio-Logic).

Fig. 1 illustrates XRD patterns of electrode materials, as well as their components: graphite oxide and MWNTs. In the structure of MWNTs pattern sharp peak at the angle 2θ of 25.96° can be found, which can be indexed as the diffraction pattern of its (0 0 2) crystal face [23]. High angle diffraction peaks observed at Bragg angels of 43.4°, 53.8° and 77.3° can be matched to crystal faces (1 0 0), (0 0 4) and (0 0 6), respectively. Graphite oxide pattern exhibits peaks at the 2θ angle of 6.2° and 42.7° which could be identified as the graphite oxide (0 0 1) and (1 0 0) crystal faces, respectively [24]. At the GO/MWNTs composite electrode no clear diffraction peaks are seen. This suggests presence of high amount of disordered carbon planes which come from exfoliated GO and thus it does not exhibit the (0 0 1) diffraction peak of graphite oxide because of its exfoliation. Only a small diffraction peak coming from the (0 0 2) MWNTs crystal face is slightly seen. Besides, the reduction process has not changed the XRD pattern noticeably.

Fig. 1

In Fig. 2, the cross-section SEM images of electrode materials are shown. Fig. 2a presents low magnification of non-reduced electrode. Fig. 2b and Fig. 2c show higher magnifications of nonreduced and photothermally reduced electrodes, respectively. These SEM images, as well as the XRD patterns, confirm the high degree of disorder between GO planes. As shown in the Fig. 2b, due to presence of MWNTs between GO layers, even electrodes before the reduction process exhibited porous structure. That is appropriate in the case of good electrolyte access. Such porosity could be the great advantage of our composite electrodes in comparison with the previous work [17], where the porous structure was induced by rapid outgassing of GO during the photothermal reduction process resulting in GO exfoliation. However, in our experiment, high electrode porosity is noticeable also for non-reduced electrode material. Hence, in spite of evident degassing of the sample during the reduction process, the exfoliation is highly limited because of good gas escape. Therefore the structure of reduced electrode, shown in Fig. 2c, is only slightly expanded in comparison to the non-reduced one.

Fig. 2

In order to analyze the effect of reduction on the elemental composition, WDS measurements were carried out. In Fig. 3, the WDS spectrum of GO/MWNTs non-reduced electrode is shown. There are a clearly visible intense peak of CKα line and two lower ones of NKα and OKα. The percentage quantifications of carbon, nitrogen and oxygen in the electrodes before and after the reduction processes are presented in Table 1. As it is seen, the reduction process has increased the amount of carbon by about 3 % with respect to the amount of nitrogen. Moreover, the sum of air components is still quite high after the reduction process, which suggests that such kind of reduction process is not as effective as e.g. some other complex, multi-step strategies [25].

Fig. 3

The first loop of cyclic voltammetric curves for non-reduced GO/MWNTs and photothermally reduced electrodes are shown in Fig. 4. For the nonreduced electrode, the cyclic voltammogram does not show any distinct peaks, which is characteristic of disorder carbons. Reduction process leads to appearance of the intercalation peaks. The oxidative peak maxima are located at 125 mV and 195 mV for reduced electrode. They are slightly shifted in comparison to graphite intercalation peaks. The appearance of intercalation peaks suggests partial graphitization of the electrode materials.

Fig. 4

In Fig. 5a, the first cycle charge/discharge profiles of electrodes are shown. There is clearly seen high irreversible capacity. Fig. 5b shows the cycling performances of electrodes at different current densities of 10 mA·g⁻¹, 30 mA·g⁻¹, 50 mA·g⁻¹ and 100 mA·g⁻¹. The first cycle reversible capacity of non-reduced electrode at the current density of 10 mA·g⁻¹ equals 136 mAh·g⁻¹, but such capacity rapidly decreases to 105 mAh·g⁻¹ just in 5^th cycle. Such rapid decreasing of the capacity is a typical behavior for all disordered carbons, especially rGO [11–15]. Higher current densities induce further capacity decreasing up to slightly above 30 mAh·g⁻¹ for current density of 100 mA·g⁻¹. After the photothermal reduction, capacity of the electrode in first cycle achieves 219 mA·g⁻¹, next it falls to 173 mA·g⁻¹ in 5^th cycle at current density of 10 mA·g⁻¹, and drops to 50 mA·g⁻¹ for current density of 100 mA·g⁻¹. Noticeable is, thus, an increase in capacity of more than 50 % after the reduction process. However, in a previous work [17] the authors obtained for the photothermally reduced graphene oxide paper much higher capacities of an order of 550 mA·g⁻¹. The explanation of such difference could be the much lower exfoliation of GO layers in our composite electrode during the rapid outgassing affected by reduction. It is highly probable that such diminished exfoliation is caused by the better gas escape realized by the porous electrode structure, what was also shown by SEM. That weakened the exfoliation of the electrode induced by the degassing process, what usually accompanies to reduction process. Consequently, it limits the increase of the capacity originating from higher specific surface area.

Fig. 5

The second reason of rising the capacity after the reduction process is the creation of conductive pathways in GO [26]. It can also increase capacity thanks to the shorter path lengths for the electrons and higher electrode conductivity. Postulated elsewhere [1] creation of favorable bonding sites for lithium by the functional groups, should lead, on the contrary, to decrease in the capacity after reduction. Hence, impact of this, if exists, should be much less than that of two previous effects.

Finally, this leads to the conclusion that the GO reduction process can influence its capacity mainly by the changes in specific surface area of the electrodes and GO flakes conductivity. It leads to an increase of the capacity after reduction process. The opposite, capacity-enhancing effect, induced by the presence of the functional groups, has not been observed. This leads to the conclusion that oxidative functionalities do not create here additional active sites for electrochemical reactions of lithium storage.