An Effective Way to Optimize the Functionality of Graphene-Based Nanocomposite: Use of the Colloidal Mixture of Graphene and Inorganic Nanosheets

The best electrode performance of metal oxide–graphene nanocomposite material for lithium secondary batteries can be achieved by using the colloidal mixture of layered CoO2 and graphene nanosheets as a precursor. The intervention of layered CoO2 nanosheets in-between graphene nanosheets is fairly effective in optimizing the pore and composite structures of the Co3O4–graphene nanocomposite and also in enhancing its electrochemical activity via the depression of interaction between graphene nanosheets. The resulting CoO2 nanosheet-incorporated nanocomposites show much greater discharge capacity of ~1750 mAhg−1 with better cyclability and rate characteristics than does CoO2-free Co3O4–graphene nanocomposite (~1100 mAhg−1). The huge discharge capacity of the present nanocomposite is the largest one among the reported data of cobalt oxide–graphene nanocomposite. Such a remarkable enhancement of electrode performance upon the addition of inorganic nanosheet is also observed for Mn3O4–graphene nanocomposite. The improvement of electrode performance upon the incorporation of inorganic nanosheet is attributable to an improved Li+ ion diffusion, an enhanced mixing between metal oxide and graphene, and the prevention of electrode agglomeration. The present experimental findings underscore an efficient and universal role of the colloidal mixture of graphene and redoxable metal oxide nanosheets as a precursor for improving the electrode functionality of graphene-based nanocomposites.


Results
The precursors of exfoliated CoO 2 and G-O nanosheets can form stable mixture colloidal suspensions with variable ratios of CoO 2 /G-O, since they possess very similar surface charge and hydrophilicity each other ( Supplementary Information, Fig. S1 and Table S1). The effect of NH 4 OH addition on the colloidal stability of G-O/CoO 2 mixture as well as on the pure colloidal suspensions of layered CoO 2 and G-O nanosheets is examined. Upon the addition of NH 4 OH, all the present colloidal suspensions remain unchanged without the formation of aggregated precipitates, clearly demonstrating the excellent colloidal stability of these suspensions ( Supplementary Information, Fig. S1A). The size distribution of exfoliated CoO 2 nanosheet is determined by a standard dynamic light scattering (DLS) analysis ( Supplementary  Information, Fig. S1B). Most of the exfoliated CoO 2 nanosheets possess the lateral size of several hundreds of nanometers, which is comparable with the reported lateral dimension of G-O nanosheet 28 . As illustrated in Fig. 1A, the hydrothermal treatment of Co 2+ ions and NH 4 OH dissolved in the mixture colloidal suspensions of the layered CoO 2 and G-O nanosheets makes possible the incorporation of layered CoO 2 nanosheets into the Co 3 O 4 -N-doped rG-O nanocomposite. Since both the exfoliated CoO 2 and rG-O nanosheets are negatively-charged, the precursor Co 2+ ions can be easily adsorbed on the surface of both the anionic nanosheets, which is followed by the crystal growth of Co 3 O 4 phase. The resulting Co 3 O 4 -layered CoO 2 -N-doped rG-O nanocomposites with different CoO 2 /G-O ratios (0, 0.5, 1, and 2wt%) are denoted as CCG0, CCG5, CCG10, and CCG20, respectively. The reduction of precursor G-O to N-doped rG-O during the synthesis is confirmed by C 1s and N 1s X-ray photoelectron spectroscopic (XPS) analysis ( Supplementary Information, Fig. S2 and Table S2). As presented in the powder X-ray diffraction (XRD) patterns of Fig. 1B, all of the present nanocomposite materials show typical Bragg reflections of spinel-structured Co 3 O 4 phase, indicating the formation of mixed valent cobalt (II,III) oxide phase during the hydrothermal reaction. On the basis of Scherrer equation, the size of Co 3 O 4 particle in the present materials is calculated to be 7.7, 8.2, 8.5, and 9.0 nm for CCG0, CCG5, CCG10, and CCG20, respectively, highlighting a slight increase of particle size upon the incorporation of layered CoO 2 nanosheets. The observed minute variation of the particle size of Co 3 O 4 upon the incorporation of CoO 2 nanosheets underscores the limited influence of layered CoO 2 nanosheets on the crystal growth of Co 3 O 4 nanoparticles.
As illustrated in the field emission-scanning electron microscopy (FE-SEM) images of Fig. 2A, all the present CCG nanocomposites commonly exhibit porous morphology formed by the house-of-cards-type stacking of nanosheet crystallites, indicating the formation of many mesopores. Such a mesoporous stacking structure is commonly observed for the self-assembled nanocomposite materials synthesized by the restacking of 2D nanosheets with 0D nanoparticles 29 . The nanoscale hybridization of cobalt oxide and graphene nanosheets is confirmed by energy dispersive X-ray spectroscopy (EDS)-elemental mapping analysis ( Supplementary Information, Fig. S3), showing the uniform distribution of cobalt, oxygen, and carbon in entire parts of the nanocomposite materials.  Fig. 2B-(c), the distance between two consecutive fringes is determined to be ~0. 29 Supplementary Information, Fig. S4). This result confirms that, like the N-doped rG-O nanosheet, the layered CoO 2 nanosheets can play a role of support for the anchoring of Co 3 O 4 nanoparticles. During the anchored growth of Co 3 O 4 phase, the layered CoO 2 nanosheets remains intact without any notable damage. This result provides clear evidence for the high stability of layered CoO 2 nanosheets against the hydrothermal synthesis.
The chemical bonding nature of rG-O and cobalt oxide components in the present nanocomposites is examined with micro-Raman spectroscopy. As illustrated in Fig. 3A, all the present nanocomposites show two intense Raman features D and G in high wavenumber region of >1000 cm −1 , characteristic of graphene species, confirming the incorporation of graphene nanosheets in these materials 30 . In contrast to N-undoped G-O and rG-O nanosheets, all the CCG nanocomposites as well as N-doped graphene demonstrate a distinct shoulder peak D', indicating N-doping for graphene component 30 . This peak D' originates from a significant perturbation of the carbon sp 2 network of graphene upon the incorporation of nitrogen element. The peak 2D reflecting the degree of the structural disorder and stacking of graphene is observed at ~2700 cm −1 for all the present materials. As the content of layered CoO 2 nanosheets increases, this 2D peak shows a slight red-shift with the increase of spectral weight, clearly demonstrating the decreased numbers of stacked graphene layers upon the intervention of layered CoO 2 nanosheets in-between the graphene nanosheets 31 (Fig. 3B). This spectral variation provides strong evidence for the weakening of π -π interaction between the graphene nanosheets upon the incorporation of layered CoO 2 nanosheets, leading to the prevention of the irreversible restacking or agglomeration of graphene nanosheets during electrochemical Li + insertion/extraction. Such a depression of the restacking or aggregation of graphene nanosheets would be beneficial in enhancing the pore structure of the present CCG nanocomposites. In the low wavenumber region, typical Raman features of spinel Co 3 O 4 phase are discernible commonly for all the CCG nanocomposites, confirming the formation of Co 3 O 4 crystals in these materials. The incorporation of layered CoO 2 nanosheets as well as the formation of Co 3 O 4 particles in the present CCG nanocomposites is cross-confirmed by Co K-edge X-ray absorption near-edge structure (XANES) spectroscopy, see Fig. 3C. As can be seen clearly from the expanded views of edge jump region, a slight but distinct blue shift of edge position is clearly observed after the incorporation of layered CoO 2 nanosheets, highlighting an increase of average Co oxidation state caused by the increase of CoO 2 content. This result confirms the HR-TEM results showing the presence of CoO 2 nanosheets in the present CCG nanocomposites. The successful incorporation of tetravalent CoO 2 nanosheets in the present materials is further evidenced by the Co 2p XPS result ( Supplementary Information,  Fig. S5), in which the tetravalent Co 4+ ions are identified and the concentration of Co 4+ ion increases As plotted in Fig. 4A, N 2 adsorption− desorption isotherm measurements clearly demonstrate the porous nature of the present CCG nanocomposites. All of the present materials display a significant N 2 adsorption at low pressure region of pp o −1 < 0.4, reflecting the existence of micropores in these materials. A distinct hysteresis commonly occurs at high pressure region of pp o −1 > 0.45 for all the present CCG nanocomposites. The observed isotherm behavior corresponds to Brunauer-Deming-Deming-Teller (BDDT) type-IV shape and IUPAC H2-type hysteresis loop, suggesting the presence of open slit-shaped capillaries with very wide bodies and narrow short necks. The incorporation of layered CoO 2 nanosheets enhances the adsorption of N 2 molecule in the low pressure region and also the total amount of N 2 molecules adsorbed, underscoring the remarkable increase of micropore volume and surface area. According to the calculation of surface area using the Brunauer-Emmett-Teller (BET) equation, the surface area of the present nanocomposite is estimated to be 32 m 2 g −1 for CCG0, 64 m 2 g −1 for CCG5, 97 m 2 g −1 for CCG10, and 82 m 2 g −1 for CCG20, respectively. This result demonstrates that the surface areas of the present CCG nanocomposites become greater with increasing the content of CoO 2 nanosheets upto the composition of CCG10. However, the further addition of CoO 2 nanosheets leads to the depression of surface area. The observed lowering of the surface area of the CCG20 nanocomposite is attributable to too high content of CoO 2 nanosheet, which is much heavier than the graphene. That is, the increase of the sample mass caused by the addition of heavy CoO 2 nanosheets outweighs the accompanying  optimization of the pore structure of the nanocomposite. This result clearly demonstrates that, even at a small concentration of CoO 2 nanosheets, the incorporation of layered CoO 2 nanosheets is fairly useful in expanding the surface area of Co 3 O 4 -graphene nanocomposite. As evidenced by the micro-Raman spectroscopy ( Fig. 3B), the incorporation of CoO 2 nanosheets is effective in depressing the π -π interaction between rG-O nanosheets and also in preventing the formation of tightly packing structure of graphene. Taking into account the fact that the severe self-restacking of graphene nanosheets leads to the remarkable decrease of surface area, the observed increase of the surface area of CCG nanocomposites upon the incorporation of CoO 2 nanosheets can be attributed to the depressed interaction between the graphene nanosheets. In fact, such a prominent increase of the surface area of restacked graphene nanosheets upon the incorporation of inorganic nanosheet is also observed for other cases like Pt-layered titanategraphene nanocomposite, clarifying the effectiveness of the nanosheet addition in enhancing the porosity of graphene-based nanocomposite 9 . The calculation of pore size based on Barrett-Joyner-Halenda (BJH) method (Fig. 4B) clearly demonstrates that all the present materials possess uniform-sized mesopores with an average diameter of ~3.3-3.4 nm, which are formed by the house-of-cards-type stacking structure of nanosheet crystallites. The BJH analysis reveals the increase of pore volume upon the incorporation of layered CoO 2 nanosheets, highlighting the positive effect of inorganic nanosheet in enhancing the porosity of graphene-based nanocomposites. The present results of N 2 adsorption-desorption isotherm analysis clearly demonstrate that the incorporation of CoO 2 nanosheets is quite powerful in increasing the surface area and pore volume of the restacked graphene nanosheets via the depression of π -π interaction between the graphene nanosheets.
The present CCG nanocomposites are applied as anode materials for lithium ion batteries. Fig. 5A shows the representative cyclic voltammogram (CV) curves of the CCG10 electrode, which are collected at a scan rate of 0.5 mV s −1 in the voltage range of 0.01-3 V vs. Li/Li + . In the first cycle, an irreversible reduction peak appears at ~0.7 V, which originates from the degradation of electrolyte caused by the formation of polymer/gel-like film around the electrode particles 32 . In the second cycle, there are two cathodic peaks at 1.41 and 0.9 V, which are ascribed to the reduction of Co 3 O 4 to Co caused by the lithiation of Co 3 O 4 . The lithiation voltage of the second cycle is shifted to higher value than that of the first cycle, indicating the improved kinetics of the CCG10 nanocomposite 12 . Meanwhile, two anodic peaks at 1.46 and 2.15 V are attributable to the oxidation of Co element to Co 3 O 4 , which corresponds to the delithiation process. These redox peaks can be regarded as evidence for the electrochemical reaction of Co 3 O 4 and Li 12 . Although very small quantity of layered CoO 2 nanosheet makes it difficult to directly detect a redox peak corresponding to reaction between layered CoO 2 and Li in the present CV data, the lithium insertion/extraction reactions in the present nanocomposites can be described by the following equations.

Co
Of prime importance is that there is no significant difference in the CV data of the CCG10 nanocomposite for the 2nd and 3rd cycles, highlighting the good reversibility of this material. Fig. 5B shows the galvanostatic discharge-charge curves at a current density of 200 mA g −1 in the range of 0.01-3 V vs. Li/Li + . All the present nanocomposites exhibit promising electrode performance with the huge initial discharge capacity of 1505, 1963, 2262, and 1797 mAh g −1 for CCG0, CCG5, CCG10, and CCG20, respectively. Although notable capacity fading occurs at the second cycle due to the formation of solid-electrolyte-interphase (SEI) layer 32 , the discharge capacity of CCG nanocomposites becomes greater with proceeding the cycle. Such an increase of discharge capacity is frequently observed for porous nanostructured materials, which is related to the formation of stable diffusion paths of Li + ions during the repeated electrochemical cycling 12,[33][34][35] . Among the present nanocomposites, the CCG10 nanocomposite with the largest surface area exhibits the most prominent enhancement of discharge capacity during the cycle. After the 20th cycle, the discharge capacities of CCG nanocomposites are stabilized to ~1230 mAh g −1 for CCG0, ~1500 mAh g −1 for CCG5, ~1750 mAh g −1 for CCG10, and ~1530 mAh g −1 for CCG20, highlighting the promising electrode performance of the present nanocomposites with huge discharge capacity and good cyclability. To the best of our knowledge, the observed discharge capacity of the CCG10 nanocomposite is the largest reversible capacity of Co 3 O 4 -based materials ever-reported ( Supplementary Information, Table S3). Since all of the components in the present nanocomposite including N-doped rG-O nanosheet are electrochemically active 36 4) The graphene nanosheets suffers from a strong tendency to form tightly packed structure due to the strong π -π interaction between sp 2 carbon arrays. The incorporation of CoO 2 nanosheets induces the formation of more open stacking porous structure providing more active sites for Li + ions insertion 9 .
All the present CCG nanocomposites display high coulombic efficiency of >98%, reflecting the highly stable and reversible insertion/extraction of lithium ions. As shown in the potential profiles of the CCG nanocomposites ( Supplementary Information, Fig. S6), all the present materials show nearly identical potential profiles, indicating the retention of the original electrochemical properties of Co 3 O 4 -graphene nanocomposite upon the incorporation of CoO 2 nanosheets. Based on the present results of electrochemical measurements, it can be concluded that the incorporation of layered CoO 2 nanosheets leads to the remarkable improvement of the electrode performance of Co 3 O 4 -rG-O (i.e. CCG0) nanocomposite.
As presented in Fig. 5C, the remarkable improvement of electrode performance upon the incorporation of layered CoO 2 nanosheet is more distinct for higher current density condition. The CoO 2 -incorporated CCG10 nanocomposite exhibits larger reversible capacities for all the current densities applied than does the CoO 2 -free CCG0 material; the CCG10 nanocomposite shows the discharge capacities of 1684, 1656, 1562, 1288, 1002 and 817 mAh g −1 at the current density of 100, 200, 400, 800, 1600 and 3200 mA g −1 , respectively. However, the CoO 2 -free CCG0 electrode delivers much smaller discharge capacities of 1232, 1207, 1044, 949, 642 and 428 mAh g −1 at the same current densities, respectively. While the discharge capacity of the CCG10 nanocomposite at 100 mA g −1 is larger by 136% than that of the CCG0 material, the CCG10 material shows even twice larger discharge capacity compared with the CCG0 one at a higher current density of 3200 mA g −1 . This finding provides clear evidence for the improvement of rate performance upon the incorporation of CoO 2 nanosheets, highlighting the improvement of charge transport property.
To examine the long-term stability of the CoO 2 -incorporated nanocomposite, the extended electrochemical cycling test is carried out for the CCG10 nanocomposite with high current density of 1000 mA g −1 and in the voltage window of 0.01-3 V vs. Li/Li + . As plotted in Figs. 5D, E, this material shows an outstanding cyclability and rate capability with the maintenance of large specific capacity of ~1150 mAh g −1 upto the 500th cycle. For the entire cycle, the coulombic efficiency of the CCG10 nanocomposite is well-maintained to ~99%, verifying the high electrochemical stability of this material. The observed beneficial effect of the incorporation of CoO 2 nanosheets on the discharge capacity of the nanocomposite is attributable to the expansion of surface area and the change of pore structure, resulting in the additional storage of Li + ions in interfacial site of the CoO 2 nanosheet-scaffolded nanocomposite. The incorporation of layered CoO 2 nanosheets also induces an enhanced nanoscale mixing between Co 3 O 4 particles and rG-O nanosheets, which is responsible for the excellent cyclability and rate characteristics of the present CoO 2 -incorporated CCG nanocomposites.
To better understand the origin of the beneficial effect of CoO 2 addition, the transport property of the present nanocomposites is investigated with electrochemical impedance spectroscopy (EIS). As plotted in Fig. 6, all the present nanocomposites demonstrate partially overlapping semicircles reflecting the charge transfer resistance (R ct ) at high-to-medium frequencies and a line corresponding to Warburg impedance at low frequencies. The incorporation of layered CoO 2 nanosheets gives rise to a significant reduction in the diameter of semicircle, indicating the decrease of R ct . Among the present nanocomposites, the CCG10 material displays the smallest diameter of the semicircle, indicating its most efficient transport property. A further increase of CoO 2 content to the CCG20 nanocomposite degrades the electron transport property, which is attributable to the decrease of highly conductive graphene content. The relative order of R ct is in good agreement with the relative electrode performances of the present nanocomposites, underscoring the main role of the improvement of transport property in enhancing the electrode performance upon the incorporation of CoO 2 nanosheets. Such a variation of transport properties is further confirmed by the change of the line slope in the low frequency region. The slope of this line for the present nanocomposites becomes steeper in the order of CCG0 < CCG20 < CCG5 < CCG10, reflecting the improvement of transport property caused by the promoted nanoscale mixing of Co 3 O 4 nanoparticles and graphene nanosheets upon the incorporation of CoO 2 nanosheets.
The effects of electrochemical cycling on the crystal structure and morphology of the present nanocomposites are examined with powder XRD and FE-SEM analyses. As demonstrated in Fig. 7A, the extended electrochemical cycling induces a decrease of the particle size of the CCG10 nanocomposite whereas the CoO 2 -free CCG0 nanocomposite shows a significant aggregation of electrode particles. Such an aggregation of electrode particles is negligible for the CoO 2 -incorporated CCG10 nanocomposite, confirming the beneficial role of CoO 2 nanosheets in the maintenance of the open structure of nanocomposite. Since Co 3 O 4 experiences severe volume change during lithiation-delithiation process, the depression of particle agglomeration is surely advantageous in enhancing the electrode performance of the nanocomposite. In addition, the electrochemical cycling induces an amorphization of both the CCG0 and CCG10 nanocomposites, see Figs. 7B,C. Such a formation of disordered structure during the electrochemical cycling causes the significant enhancement of Li + diffusion via the provision of more diffusion paths 39 . The incorporation of layered CoO 2 nanosheets does not induce any significant change in the XRD data of the cycled derivative, indicating negligible effect on the structural stability of the nanocomposite. The present finding strongly suggests that the beneficial role of CoO 2 addition mainly originates from the improvement of the morphological stability of composite structure rather than the change of crystal structure. Even though the incorporated CoO 2 nanosheets is transformed into cobalt oxide particles during the electrochemical cycling, an intimate mixing between cobalt oxide and rG-O in the CoO 2 -incorporated CCG10 nanocomposite provides improved diffusion paths for Li + ions as well as a strong electronic coupling between cobalt oxide and graphene nanosheeets. Such an advantageous effect of the incorporation of metal oxide nanosheet is further evidenced from MnO 2 nanosheet-incorporated Mn 3 O 4 -N-doped rG-O nanocomposite. The obtained Mn 3 O 4 -layered MnO 2 -N-doped rG-O displays the typical XRD patterns of Mn 3 O 4 phase and porous morphology formed by the house-of-cards-type stacking of sheet-like crystallites, as observed for the present CCG nanocomposites ( Supplementary Information, Fig. S7). This material shows the discharge capacity of 1383 and 900 mAh g −1 for the 1st and 2nd cycles, respectively. The discharge capacity of this material becomes increasing with proceeding the cycling, leading to the huge discharge capacity of 1250 mAh g −1 at the 50th cycle. Even at high current density of 2000 mA g −1 , the material can still deliver a high capacity of ~700 mAh g −1 ( Supplementary Information, Fig. S7). The electrode performance of the present nanocomposite is the best one among the reported data of Mn 3 O 4 -graphene nanocomposites ( Supplementary  Information, Table S4), highlighting the excellent electrode activity of the present material. It is worthwhile to mention that, in comparison with the cobalt oxide-based CCG nanocomposites, the Mn 3 O 4layered MnO 2 -N-doped rG-O nanocomposite is much more promising electrode material because of the low price and low toxicity of Mn elements. This result underscores the usefulness of the present synthetic strategy for exploring novel efficient electrode materials highly suitable for practical use.

Discussion
In the present study, we are successful in developing a very efficient method to improve the electrode performance of graphene-based nanocomposite materials using the colloidal mixture of layered metal oxide and graphene nanosheet as a precursor. In fact, there are considerable numbers of reports about the cobalt oxide-graphene nanocomposite electrode materials for lithium secondary batteries 12,35,[40][41][42][43] . In one instance, Wang et al. reports the large discharge capacity of 1200 mAh g −1 for the Co 3 O 4 -graphene film, which is greater than the data of other previous reports 12,40 . In comparison with these data, the present CCG nanocomposites show much larger discharge capacity of ~1550-1750 mAh g −1 with excellent cyclability and good rate characteristics, which outperforms all of the previously reported Co 3 O 4 -based electrode materials ( Supplementary Information, Table S3). Similarly the electrode performance of the Mn 3 O 4 -layered MnO 2 -N-doped rG-O nanocomposite is superior to all the reported data of Mn 3 O 4graphene nanocomposite ( Supplementary Information, Table S4). This result provides strong evidence for the unique merit of metal oxide nanosheets as an additive for graphene-based nanocomposite electrode materials. The observed dramatic enhancement of the electrode performance upon the incorporation of layered metal oxide nanosheet is attributable to the increase of surface area, the enhancement of the nanoscale mixing of components, the improvement of electrical transport properties, and the enhancement of morphological stability upon the incorporation of layered metal oxide nanosheets. The present study obviously verifies a beneficial and universal role of exfoliated metal oxide nanosheets in optimizing the electrode performance of graphene-based nanocomposite. As mentioned in the introduction section, the graphene-based nanocomposites boast versatile applications such as electrodes for secondary batteries, supercapacitors, fuel cells, and solar cells, photocatalysts, redox catalysts, nanobio materials, structural materials, and so on [4][5][6][7][8]11,15,[40][41][42][43][44][45][46] . For most applications of these graphene-containing materials, the homogeneous blending between graphene nanosheets and hybridized functional materials, and the optimization of porous structure are commonly important in enhancing their functionalities. Taking into account the fact that the present synthetic strategy is readily applicable for many types of rG-O-based nanocomposites, the incorporation of exfoliated inorganic nanosheets can provide a powerful methodology to optimize the diverse functionalities of graphene-containing nanocomposites through the effective deterioration of the tight packing structure of graphene nanosheets. Our current research project is the use of the various colloidal mixtures of graphene and inorganic nanosheets for the exploration of novel graphene-based functional materials with various applicabilities for solar cells, photocatalyst, supercapacitors, and so on.

Methods
Materials preparation. The exfoliated layered CoO 2 nanosheets was prepared by the proton exchange of LiCoO 2 and intercalation of tetramethylammonium (TMA) ions into HCoO 2 23 . The colloidal suspension of graphene oxide (G-O) was synthesized from graphite by a modified Hummers' method, in which the concentration of KMnO 4 was reduced to 1/6 of conventional concentration 47 . The as-prepared G-O was dispersed in anhydrous ethanol with the concentration of 0.32 mg mL −1 by ultrasonication for 0.5 h. An aqueous suspension of exfoliated G-O (48 mL) was reacted with 2.4 mL 0.2 M Co(Ac) 2 and 1 mL 30% NH 4 OH aqueous solution, 1 mL H 2 O, and layered CoO 2 nanosheets (0-2 wt% to exfoliated G-O nanosheets). The mixture was stirred at 80 °C for 10 h. Then the mixture was transferred to an autoclave for hydrothermal reaction. The reaction condition was 3 h at 150 °C. In this step, G-O was reduced to N-doped rG-O. After the reaction, powdery precipitates were collected by centrifugation, washed with ethanol and distilled water, and then freeze-dried. After the completion of the reaction, only transparent supernatant solution remained. No observation of Tyndall phenomenon for the supernatant solution clearly demonstrated the absence of any precursor colloidal particles in this solution. Additionally, no formation of precipitate upon the addition of hydroxide ions confirmed the complete incorporation of Co 2+ , CoO 2 , and graphene reactants into the precipitated nanocomposite materials. On the basis of the present findings, the weight ratios of the components in these materials could be estimated from the starting ratios of the reactants. The weight ratio of Co 3 O 4 :CoO 2 :rG-O components was estimated to 3.7:0:1 for CCG0, 3.7:0.007:1 for CCG5, 3.7:0.015:1 for CCG10, and 3.7:0.03:1 for CCG20, respectively. Since the weight of G-O component was much more convenient and precise to calculate than its molar concentration, the weight ratios of CoO 2 /G-O were applied for controlling the compositions of the present nanocomposites like many other studies about the graphene-based nanocomposites.

Materials characterization.
The crystal structures of the as-prepared CCG nanocomposites and their electrochemically cycled derivatives were analyzed by powder XRD analysis (Rigaku D/Max-2000/ PC, Cu Kα radiation, 298 K) analysis. The crystal morphology of the present samples was examined by FE-SEM (JEOL JSM-6700F) and HR-TEM/SAED (Jeol JEM-2100F, an accelerating voltage of 200 kV). The spatial elemental distribution of the present materials was probed with EDS-elemental mapping analysis. XANES spectroscopic experiment was carried out at Co K-edge at the beam line 10C at the Pohang Accelerator Laboratory (PAL) in Korea. The chemical bonding nature of nitrogen species was investigated with XPS analysis (Thermo VG, UK, Al Kα ), in which a monochromated X-ray beams was used. All the XPS spectra were calibrated with a reference to the adventitious C 1s peak at 284.8 eV to rule out any possible spectral shift by the charging effect. To avoid the accumulation of charge during the measurement, all the samples were deposited on metallic copper foil. N 2 adsorption-desorption isotherms were measured at 77 K using Micromeritics ASAP 2020 analyzer to determine the surface area. Before the measurements, the degassing of the samples was carried out at 150 °C for 3 h under vacuum. Micro-Raman spectra were obtained with a JY LabRam HR spectrometer using an excitation wavelength of 514.5 nm. The zeta potentials of the pure colloidal suspensions of G-O and layered CoO 2 nanosheets, and their colloidal mixtures were measured with Malvern Zetasizer Nano ZS (Malvern, UK).
Electrochemical measurement. The CV data were collected using an IVIUM analyzer with a scanning rate of 0.5 mV s −1 and a potential range of 0.01-3.0 V (vs. Li/Li + ). The EIS data were collected in the frequency range of 0.01 Hz-100 KHz. Electrochemical measurements were carried out at room temperature using 2016 coin-type cell of 1 M LiPF 6 in an equivolume mixture of ethylene carbonate/diethyl carbonate (EC/DEC = 50:50). The working electrodes were fabricated by mixing 80 wt% active material, 10 wt% Super P, and 10 wt% polyvinylidene fluoride (PVDF) dissolved in N-methyl-2-pyrrolidinone (NMP). The composite electrodes were prepared by coating the anode slurry onto a copper foil as a current collector and drying under vacuum at 110 °C for 12 h. The test cells were assembled in an argon-filled glove box. All the galvanostatic charge-discharge tests were performed with Maccor (Series 4000) multichannel galvanostat/potentiostat in the voltage range of 0.01-3.0 V (vs. Li/Li + ) at current density of 100-3200 mA.