RGO-Coated TiO2 Microcones for High-Rate Lithium-Ion Batteries.

Reduced graphene oxide (RGO)-coated TiO2 microcones have been synthesized via simple anodization and cyclic voltammetry for use in lithium-ion batteries (LIBs). Microcones had a perpendicularly oriented hollow core, an anatase structure, and a high surface area, allowing higher capacity than other nanosized TiO2 structures. TiO2 has low electrical conductivity, leading to the limitation of fast charging and high capacity; however, this was improved by the application of an RGO coating in this work. As anode materials of LIB, the obtained RGO microcone showed a capacity of 157 mAh g-1 at 10C (fully charged within ∼360 s) and sustained 1000 cycles with only 0.02% capacity fading per cycle. The capacity was 1.5 times higher than that of conventional microcone. We speculated that the decrease in the charge-transfer resistance (R ct) played a crucial role in increasing the capacity with fast charging.

Introduction

With the increased use of lithium-ion batteries (LIBs) in numerous applications,[1−3] the need for improvement in safety, cycle life, capacity, and fast charging performance is also increasing significantly.[4−8] Conventional graphite anodes for LIB still pose the risk of dendrite formation through decomposition of the electrolyte during fast charging. TiO2 is one of the most attractive alternative materials owing to its appropriate operating potential window (∼1.7 vs Li/Li+) and lack of electrolyte decomposition.[9−11] The morphology or phase of TiO2 greatly influences the electrochemical properties, such as capacity, cyclability, performance, and conductivity.[12,13] For example, TiO2 microcones, which have a hollow core with a multilayered anatase structure, can be produced via facile one-step anodization and show 3 times higher capacity than TiO2 nanotubes.[7] However, although microcones show higher capacity and excellent cycle performance, they cannot overcome the low theoretical capacity and conductivity of TiO2 itself.

In this study, graphene oxide (GO), which shows both high energy density and excellent conductivity, was applied as a coating layer to increase the capacity and electrical conductivity of microcone.[14−16] Since the oxygen functional groups of GO, such as hydroxyl, carboxyl, and carbonyl, interfere with ionic conductivity, reduced graphene oxide (RGO) microcone should be prepared for the battery anode.[17−19] We found that conventional reduction methods of GO on TiO2 microcone, such as hydrothermal reduction,[20−22] cathodic reduction,[23,24] and annealing under Ar (or +H2) gas,[25,26] were not very effective since TiO2 microcone provided abundant oxygen with GO during the reduction process. In addition, the optimized 1 cycle cyclic voltammetry (CV) was the most efficient way to produce RGO microcone from oxygen-abundant environment.[27] The prepared RGO microcone as an anode material for LIB showed a specific capacity of 157 mAh g–1 at 10C, which corresponds to 1.5 times higher capacity compared to that of the previous TiO2 microcone. In addition, the battery performance retained its capacity over 1000 cycles with only 0.02% capacity fading per cycle, meaning that we can greatly improve the LIB anode of TiO2 microcone by a facile CV reduction process.

Results and Discussion

After electrophoretic deposition of GO on TiO2 microcone, it was reduced to RGO by various conventional methods: hydrothermal reduction, cathodic reduction, annealing in H2 + Ar, annealing in Ar, and electrochemical CV method. Because of oxygen-abundant environmental surface of TiO2, GO was not very effectively reduced to RGO. For example, in the case of hydrothermal route, hazardous hydrazine hydrate was used to reduce GO on TiO2 microcone in an autoclave at 95 °C for 12 h. Due to high pressure and temperature, resultant RGOs were aggregated together (see the arrow in Figure S1a). On the other hand, cathodic reduction in 0.1 M Na2SO4 at −1 V for 1 min led to locally detach microcone from the Ti substrate (see the arrow in Figure S1b), meaning that the microcone was not stably adhered on the substrate during the cathodic reduction of GO. Annealing processes in H2 + Ar and Ar were conducted at 550 °C for 1 h and 200 °C for 2 h, respectively, to reduce GO on the microcone. Although the annealing processes showed a better reduction result over the entire surface compared to the other reduction methods, some microcones had spilt ends during the annealing in H2 + Ar (see the arrow Figure S1c) and the head of the microcone was collapsed during the annealing in Ar (see the arrow Figure S1d). We found that RGO-coated microcone was successfully prepared via electrochemical CV reduction (Figure a); microcones with a diameter of approximately 3.37–6.73 μm and a height of 4.31–9.57 μm were produced on the Ti substrate, perpendicular to the substrate face. The microcones maintained their original morphology, with no indication of collapse or detachment from the Ti substrate after the electrophoretic deposition and CV processes. Notice that a range of CV reduction was a very important factor to effectively reduce GO to RGO (Figure S2). When the potential range was extended up to −1.2 V (green line) or −1.5 V (violet line), an anodic peak appeared in the backward scan, indicating that the reduced GO was oxidized to GO again during the backward scan. Therefore, a potential range of −1.0–0.8 V was selected as an appropriate potential range, showing no anodic current peak. Furthermore, to find out the most optimum ratio between RGO and TiO2 microcone, CV reduction was conducted in the potential range of 0.8 to −1.0 V (Figure S3), which indicated that only the first cycle did not exhibit anodic current peak. This means that RGO was coated thinly and evenly on the surface of TiO2 microcone without the loss of large active surface area. Therefore, the weight was not significantly changed during CV reduction due to a very tiny amount of RGO.

Figure 1

(a) Scanning electron microscopy (SEM) and (b) High resolution-transmission electron microscopy (HR-TEM) images of the as-prepared RGO microcone by CV reduction and (c) X-ray diffraction (XRD) patterns of microcone and RGO microcone.

Figure b shows an high-resolution (HR)-TEM image of the as-prepared RGO microcone, clearly confirming the presence of RGO nanosheets. The dark part represents the RGO nanosheets on the microcone, which enhanced the intrinsically poor electrical conductivity of the TiO2. Figure c shows XRD patterns of pristine microcone and RGO microcone. The detected peaks for the microcone correspond to the diffraction peaks of anatase TiO2 (anatase TiO2 relating to (101), (004), (200), and (105), (JCPDS no. 01-71-1166)) and Ti peaks from the Ti substrate. The same peaks were observed for RGO microcone, demonstrating that the microcones maintained their original crystallinity after modification with RGO. Although the presence of RGO was confirmed via the HR-TEM images (Figure b), the peak for RGO at approximately 2θ = 24.5° was not observed in the XRD spectrum of RGO microcone, which could be attributed to the overlap with the peak relating to the anatase phase (101) and a small loading of RGO.

Raman spectroscopy was used to identify the presence of RGO, confirming the reduction of GO to RGO (Figure and Table ). Two major peaks were detected: the D band is related to local defects and disorders in graphitic carbonaceous materials, and the G band is related to sp2 hybridized carbon. The intensity ratio ID/IG is used as a parameter for measuring oxidation/reduction processes.[17,19,28,29] Since more structural defects are formed after the reduction process of GO, an increase in the intensity ratio of ID/IG indicates the reduction of GO to RGO. Therefore, by comparison of the ID/IG of the pristine GO and RGO reduced by the CV reduction, which were 0.94 and 1.24, respectively, the reduction of GO to RGO by CV reduction was the best method among all routes that we tested. Since RGO has excellent electrical conductivity, we expected that it could efficiently transfer electrons to the microcone, resulting in an improvement in cell performance for LIB.

Figure 2

Raman spectra as a function of experimental methods.

Table 1

Intensity Ratio of ID/IG As a Function of Experimental Method

reducing method	GO microcone	hydrothermal reduction	cathodic reduction	annealing (H2 + Ar)	annealing (Ar)	CV reduction
ID/IG	0.94	0.90	0.92	0.95	1.05	1.24

In Figure , the X-ray photoelectron spectroscopy (XPS) spectra of the GO microcone and the RGO microcone provide evidence for the reduction of GO to RGO in terms of chemical states. The C 1s spectrum shows C=C/C–C peaks at 285.15 and 285.36 eV, C–O peaks at 287.10 eV, and C=O peaks at 288.70 eV.[30] In comparison to the spectrum of GO microcone, the decreased binding energy intensities between carbon and oxygen, especially for C–O, in the spectrum of RGO microcone indicate reduction of the oxygen-containing functional groups of GO. In the O 1s spectra, the detected peaks of Ti–O (530.18 eV) and −OH (531.80 eV) clearly confirm the TiO2 microcone structure.[28] In addition, the other peak at 532.93 eV is related to C=O, further confirming that the GO was reduced to RGO. Peaks for Ti 2p were observed at 458.88 and 464.58 eV, which are in accordance with the Ti 2p3/2 and Ti 2p1/2 in TiO2, respectively,[31] demonstrating that the microcone maintained their original chemical composition after the GO reduction process.

Figure 3

High-resolution XPS spectra in Ti 2p, O 1s, and C 1s.

The electrochemical performance of the RGO microcone and microcone electrodes was studied in a coin-type cell. Prior to the electrochemical evaluation, the cell exhibited stable open circuit voltages at approximately 1.8 V (vs Li/Li+). CV curves of microcone and RGO microcone were obtained in the potential range of 1.0–3.0 V (vs Li/Li+) at a scan rate of 0.1 mV s–1 at the fifth cycle (Figure a). During cathodic and anodic sweeps of both electrodes, redox peaks at approximately 1.71 and 1.9 V (vs Li/Li+) were in accordance with insertion and extraction of lithium ions into the anatase TiO2 phase, respectively, which can be described by the following equation[32]x can vary with the polymorph, morphology, and crystallinity of the prepared TiO2 electrode. Similar anodic and cathodic peaks for microcone and RGO microcone indicated that both electrodes had excellent reversibility. Moreover, almost no difference in the CV curves was observed during repeated cycles, providing further evidence of the excellent reversibility of both electrodes. Higher current densities (0.3 mA cm–2) were observed for the RGO microcone electrode compared with the microcone electrode (0.2 mA cm–2), resulting in higher charge/discharge capacities. Figure b provides the galvanostatic charge/discharge curves for the RGO microcone and microcone between 1.0 and 3.0 V at a current density of 125 mA g–1 for the first cycle, showing that both voltage plateaus are well-matched with the CV curves of Figure a. Microcone and RGO microcone deliver discharge capacities of 181 and 235 mAh g–1 with high Coulombic efficiencies of 89 and 95%, respectively.

Figure 4

Electrochemical properties of RGO microcone and microcone: (a) CV measurements; (b) galvanostatic charge/discharge curves in the potential range 1.0–3.0 V (vs Li/Li+) at a scan rate of 0.1 mV s–1; (c) specific capacity vs cycle numbers of RGO-microcone cell at 125, 250, and 650 mA g–1; (d) rate performance at various C-rates; (e) long-term cycling performance at 10C; and (f) Nyquist plots based on the electrochemical impedance spectroscopy (EIS) data.

Figure c shows the cycling stability of the RGO-microcone electrode at specific current densities of 125, 250, and 625 mA g–1 until the 100th cycle. Initial capacity loss is observed generally due to the existence of irreversible Li trapped sites and trace amounts of adsorbed water, which is common in porous materials.[6] The charge/discharge capacity of the RGO-microcone electrode decreases with increasing specific current density and continues stably with high Coulombic efficiency, showing excellent capacity retention under all conditions. Evaluation of the rate capability was carried out at rates of up to 50C (6250 mA g–1) for the microcone and RGO-microcone electrodes (Figure d). Even at an ultrafast rate of 50C, the RGO microcone could deliver a discharge capacity of 88 mAh g–1, which is significantly improved, compared with other results (note that 100 mAh g–1@13C and 97.7 mAh g–1@16C).[17,33]Figure e shows the long-term cyclability of the RGO microcone at a high C-rate of 10C. Discharge capacity of 157 mAh g–1 was maintained for 1000 cycles, with a capacity fading rate of only 0.019% per cycle, demonstrating excellent cycling stability of the RGO microcone even at a high C-rate of 10C. This result can be attributed to the structural benefits of the microcone that create a large, electrochemically active surface area. The microcones have multilayer nanofragments and a hollow core, leading to efficient Li+ diffusion[7] in addition to the enhanced electrical conductivity provided by the RGO coating.

Figure f shows Nyquist plots of both electrodes obtained by EIS measurements before charge/discharge cycles, which shows the comparison of the electrical conductivity of the microcone and RGO-microcone electrodes. An equivalent circuit in the inset was depicted to fit the spectra. Re corresponds to the electrolyte resistance, CPEdl represents a constant-phase element of a double-layer capacitance, and Warburg impedance (W) is affected by lithium-ion diffusion in the electrode.[15,33,34] The charge-transfer resistance (Rct) was calculated as 64.65 and 25.13 Ω for microcone and RGO microcone, respectively, clearly confirming that charge transfer is significantly enhanced by RGO coating.

In this study, we successfully modified microcone structures for LIB applications by coating the microcone with RGO through various methods. For the fabrication of RGO microcone, microcones were first coated with GO by electrophoretic deposition; GO was then successfully reduced to form RGO via a simple CV method, with no evidence of morphological change or collapse of the microcone. The presence of RGO was confirmed by XPS analysis and Raman spectroscopy. RGO-microcone electrodes exhibited significantly higher electrochemical performance than that of microcone electrodes. The specific capacity of RGO microcone was around 1.5 times higher than that of microcone, showing the excellent capacity retention. We speculate that these results were caused by a sharp drop in the charge-transfer resistance induced by the highly conductive RGO.

Experimental Section

Preparation of RGO–TiO2 Microcones

Before anodization, titanium foil (0.127 mm thick, 99.7% purity, Sigma-Aldrich) was ultrasonicated sequentially in acetone, ethanol, and deionized (DI) water and then dried at room temperature. H3PO4 (1 M, Sigma-Aldrich) and hydrofluoric acid (0.5 vol %, Sigma-Aldrich) were dissolved in DI water for the electrolytes. The anodization was carried out at 60 V for 45 min using a direct current (DC) power supply (N8761A, Agilent Technologies). A homemade, two-electrode Teflon cell with titanium foil as the working electrode and platinum mesh as the counter electrode was used for the anodization process.

For the deposition of GO on the microcone, 25 mg of single-layer GO (ACS Material, LLC) was dispersed in 100 mL of DI water via ultrasonication and cooled to room temperature. The electrophoretic deposition was carried out in the same cell used for anodization, using a DC power supply. GO was deposited on the positively biased microcone at 60 V for 10 min. The prepared GO-microcone specimens were rinsed with DI water. The electrochemical reduction of GO microcone to RGO microcone was carried out in 0.1 M Na2SO4 (Sigma-Aldrich) via 1 cycle CV reduction in the optimal potential range of −1.0–0.8 V (vs Ag/AgCl) using an electrochemical workstation (PGSTAT302N, Autolab Metrohm) (Figure S2). The specimens were rinsed with DI water and dried overnight at 60 °C.

Physical Characterization

A field-emission scanning electron microscope (FE-SEM, 4300S, Hitachi) equipped with an energy-dispersive X-ray electron microscope and a high resolution-transmission electron microscope (HR-TEM, JEM-2100F, JEOL) were used for analysis of the surface morphology. The chemical composition and crystallinity of the specimens were investigated via X-ray diffraction (XRD, Rigaku D/max-RB) with Cu Kα radiation (1.54056 Å). X-ray photoelectron spectroscopy (XPS, VGESCALAB 220i-XL spectrometer, Fiscons) analysis using an Al Kα X-ray source was performed to gain more detailed information about the chemical states. The carbonaceous materials were studied further using a Raman spectrometer (Lab Ram ARAMIS, HORIBA). The weight of specimens was accurately measured by a microbalance (BM-22, AND). All RGO-microcone electrodes weighed on an average 0.6 mg.

Electrochemical Measurements

The electrochemical performance of the RGO-microcone anode material for the LIB was evaluated in coin-type cells (CR-2032; Wellcos Corporation), which were assembled in an Ar-filled glovebox with H2O and O2 content below 1 ppm. The RGO microcone was used as the working electrode, and Li foil was used as the counter and reference electrode. Glass fiber (GF/A, Whatman) was placed as a separator between the two electrodes. The electrolyte was 1 M LiPF6 dissolved in ethylene carbonate and dimethyl carbonate (1:1 by volume, Sigma-Aldrich). Note that the cell did not include binder and conducting agents as the anodic microcones were grown directly on a Ti substrate. CV measurements were carried out with the electrochemical workstation. Electrochemical impedance spectroscopy (EIS) measurements were carried out with an amplitude of 5 mV between 106 and 10–1 Hz using the electrochemical workstation. Galvanostatic charge/discharge cycling was conducted at various C-rates using a battery cycler system (WBCS 3000; WonATech).