In situ fabrication of a graphene-coated three-dimensional nickel oxide anode for high-capacity lithium-ion batteries

The high theoretical specific capacity of nickel oxide (NiO) makes it attractive as a high-efficiency electrode material for electrochemical energy storage. However, its application is limited due to its inferior electrochemical performance and complicated electrode fabrication process. Here, we developed an in situ fabrication of a graphene-coated, three-dimensional (3D) NiO–Ni structure by simple chemical vapor deposition (CVD). We synthesized NiO layers on Ni foam through a thermal oxidation process; subsequently, we grew graphene layers directly on the surface of NiO after a hydrogen-assisted reduction process. The uniform graphene coating renders high electrical conductivity, structural flexibility and high elastic modulus at atomic thickness. The graphene-coated 3D NiO–Ni structure delivered a high areal density of ∼23 mg cm−2. It also exhibits a high areal capacity of 1.2 mA h cm−2 at 0.1 mA cm−2 for its Li-ion battery performance. The high capacity is attributed to the high surface area of the 3D structure and the unique properties of the graphene layers on the NiO anode. Since the entire process is carried out in one CVD system, the fabrication of such a graphene-coated 3D NiO–Ni anode is simple and scalable for practical applications.


Introduction
The increasing demand for high-efficiency, large-scale electrochemical energy storages (e.g., electric vehicles) has led to an expansion in new developmental efforts for high energy-density lithium-ion batteries (LIBs). 1 However, commercial graphite anodes involving the conventional Li ion (Li + ) intercalation reaction have a low theoretical specic capacity (372 mA h g À1 ), which prevents them from being applied in advanced energy storage. 2 In this context, transition metal oxides have been considered promising electrode materials because of their high theoretical capacities, chemical stability and low cost. 3,4urthermore, the conversion reaction of 2yLi + + M x O y 4 xM + yLi 2 O (M denotes transition metals such as Ni, Cu, Fe, and Co) is a thermodynamically favorable reaction that facilitates large amount of electron transfers, which leads to two-to three-fold superior capacity (600-1100 mA h g À1 ). 5 Among transition metal oxides, nickel oxide (NiO) could be a strong contender due to its higher theoretical specic capacity (718 mA h g À1 ), chemical stability, environmental benignity and low cost. 6evertheless, the inherently low electrical conductivity (r > 10 15 Um at room temperature) and low specic surface area hinder NiO from achieving high LIB performance. 7Furthermore, the conversion reaction of NiO-Li 2 O poses another problem associated with the large volume change between NiO and Li 2 O during cycling. 43][24][25][26][27][28][29][30][31][32] Various methods to prepare NiO-graphene nanocomposite structures include, but are not limited to, hydrothermal synthesis, [23][24][25][26][27] core-shell spray pyrolysis, 28 nanoparticles-sheet assembly, 29,30 ultrasonication 31 and electrical wire pulse technique. 32Overall, the resultant NiO-graphene structures show specic capacity (700-1098 mA h g À1 ) superior to NiO (100-439 mA h g À1 ).However, such a high specic capacity has never been translated into increased areal capacity (capacity per footprint area), which is crucial for practical application in LIBs.The low areal capacity has been considered a critical drawback in most nanomaterial-based anodes.Nevertheless, no attention has been paid to enhancing areal capacity of NiO-graphene anode for largescale, advanced LIB.
Our previous experiment demonstrated how 3-dimensional (3D) structures could enhance the areal capacity of electrodes in LIBs. 33,34By following this concept, we fabricated a novel structure of graphene-coated 3D NiO-Ni anode through a simple two-step thermal chemical vapor deposition (CVD) method, in which graphene layers were grown directly on a NiO-Ni structure.In this structure, porous 3D Ni substrate offered high surface area to accommodate large loading of NiO; in addition, the porous structure facilitated lithium ion diffusion within NiO. 8,9,15,16 The in situ graphene growth on NiO was achieved by a simple CVD process right aer reduction of the NiO process in the same CVD chamber; [23][24][25][26][27][28][29][30][31][32] the process was effective, yet facile, to produce a highly stable graphene network throughout the 3D NiO structure.The graphene-coated 3D NiO-Ni anode delivered improved areal density ($23 mg cm À2 ) and higher areal capacity (1.2 mA h cm À2 at 0.1 mA cm À2 ) than previously reported NiO-based anodes; [8][9][10][12][13][14][15]18,19 such values are critical for practical applications. The excellet properties and novel design of the graphene coated 3D NiO-Ni anode would expand the development of large-scale LIBs.

Synthesis of graphene-coated 3D NiO-Ni electrode
Porous Ni foam, with nominal cell size of 450 mm and porosity of 85% (Alantum), was used as a pristine substrate for NiO growth.The Ni foam was cut and then inserted into a quartz tube of a thermal chemical vapor deposition (CVD) system (Atomate).For NiO growth, the substrate was rapidly heated, in an Ar (500 sccm) and O 2 (125 sccm) mixed gas environment (4 : 1 volume ratio), to a temperature of 1000 C for 2 hours.At the nal stage, the as-grown NiO on Ni foam was naturally cooled to room temperature within the tube under inert Ar gas atmosphere.Aer the growth process, the gray Ni foam was transformed into a greenish stoichiometric NiO structure.Subsequently, the CVD tube was rapidly heated to 700 C within the Ar gas environment.Once temperature was reached, the mixture of CH 4 , H 2 , and Ar gases was introduced at ow rates of 50, 100 and 500 sccm (1 : 2 : 10 volume ratio), respectively.During the graphene growth process of 1 minute, the thermally decomposed carbon from the precursor CH 4 gas was absorbed onto the reduced Ni from NiO (the NiO reduction process simultaneously occurred by H 2 gas).Consequently, a graphenecoated NiO-Ni nanocomposite structure was synthesized.Fig. 1 schematically illustrates the fabrication procedures for the graphene coated 3D NiO-Ni foam.

Structural characterization
The morphologies of the NiO-Ni foam and graphene-coated NiO-Ni foam structures were identied with a eld emission scanning electron microscope (FESEM) (JEOL, JSM-7000F).Elemental analysis for both structures was carried out using an energy dispersive spectroscope (EDS) (FEI Helios 650).For cross-sectional SEM-EDS analysis, the samples were frozen in liquid nitrogen (77 K) and then cut into two pieces.The structural property of the samples was also characterized with an Xray diffractometer (XRD) (Rigaku, Rint-2000) using Cu K-alpha radiation in the range of 10-90 (2q) with step size 0.01 and with a Raman spectroscope (Jobin-Yvon, Labram HR) using Ar + laser with l ¼ 514 nm and 0.5 mW power.

Electrochemical characterization
A CR2032 coin cell (Wellcos Ltd.) was assembled with as-grown 3D graphene-NiO-Ni as a working electrode and lithium foil as both counter and reference electrodes.No current collector or additive was incorporated into the assembled anode; this is advantageous for enhanced energy and power density of an LIB cell (for example, deadweight of conventional current collector constitutes nearly 10% of the total weight of an LIB cell 35,36 ).1.0 M LiPF 6 in ethylene carbonate-dimethylene carbonatediethylene carbonate (EC-DMC-DEC) (1 : 1 : 1 in volume) and a typical polypropylene (PP) based membrane (Separator-2400, Wellcos Ltd.) served as an electrolyte and a separator, respectively.The complete cell assembly was conducted in an argon-lled glovebox that maintained oxygen and humidity levels less than 0.5 ppm.The charge-discharge cycling behaviors of the cell were characterized with a multi-channel battery tester (MACCOR-series 4000) in galvanostatic mode (constant current).In this study, charge and discharge processes were related to the oxidation and reduction (conversion) reactions as NiO + 2Li + + 2e À 4 Ni + Li 2 O, respectively.The cells were cycled in the voltage range of 0.01-3.0V at different current densities.Cyclic voltammetry (CV) measurements for the 3D graphene-NiO-Ni and 3D NiO-Ni anode samples were conducted using a multi-channel potentiostat (Bio Logic, VMP3) in the voltage range of 0.01 to 3.0 V (vs.Li + /Li) at a scan rate of 0.1 mV s À1 .

Structural characterization of NiO-Ni foam
FESEM image of porous Ni foam showed the average pore size was 150 mm and the width was approximately 40 mm (Fig. 1(a)) with smooth polycrystalline surface.Aer thermal oxidation, assynthesized porous NiO was uniformly grown throughout the Ni surface while preserving the micro-channeled structure (Fig. 2(a)). 17 4 No other peaks relevant to the impurities were iden-tied in the XRD patterns.The Raman spectra further corroborate the results from the XRD analysis (Fig. 2(d)).A typical onephonon peak at $570 cm À1 (LO mode), three two-phonon peaks at $730 cm À1 (2TO mode), $906 cm À1 (TO + LO mode) and $1090 cm À1 (2LO mode), and one strong two-magnon peak at $1490 cm À1 (2M mode) were observed; the peaks were consistent with previous results from NiO. 41 In contrast, no Raman peak for Ni indicates the lack of active vibrational Raman mode in Ni. 9 In Fig. 2(e-g), the EDS elemental mapping images represent the distribution of Ni and oxygen (O) elements on the 3D NiO-Ni.Noticeably, an average areal density of NiO only in the 3D structure is empirically measured as $23 mg cm À2 .As evident in Fig. S1, ‡ the weight ratio of 3D NiO ($51%) as an electroactive material to 3D NiO-Ni is consistent with the compositional ratio of $50% NiO.These are the highest values reported to date by the thermal oxidation process.Therefore, we could conrm that our exerted oxidation condition is more intense than other reported ones that processed the Ni-NiO structures. 8-10,12-19,28

Structural characterization of graphene coated NiO-Ni foam
Thermal CVD is more facile for direct growth of high-quality graphene on metal substrates (e.g., Cu and Ni); additionally, it allows excellent physico-chemical properties of graphene. 22,23,31n principle, graphene is not directly grown on NiO due to insolubility of carbon into NiO. 42Thus, the H 2 reduction process is required to transform NiO to Ni for graphene growth.The amount of O vacancies in the NiO structure increases with elevated temperature during the reduction process and catalyzes cleavage of the hydrogen bond (H-H); thus, the reaction produces H 2 O gas and leaves behind a Ni structure. 43Graphene growth proceeded right aer the NiO reduction process by using the same CVD system. 44The process steps are summarized as follows: (1) 3D NiO-Ni structure was annealed in Ar environment at up to 700 C. (2) CH 4 /H 2 (1 : 2 volume ratio) gas mixture was introduced into the reactor.(3) The structure was cooled to room temperature in Ar environment.In step (2), hydrocarbon (e.g., methane in this study) is thermally decomposed; subsequently, the resultant carbon atoms dissolved into Ni aer NiO is transformed into Ni by H 2 reduction process.During step (3), carbon atoms are segregated and then precipitated on the Ni surface; thus, graphene layers are grown. 45Fig. 3(a and b) show the cross-sectional FESEM images demonstrating the graphenecoated 3D NiO-Ni structure.In the images, the presence of graphene is evidenced by the characteristic rippled and wrinkled structures. 46The presence of the as-grown graphene is also conrmed by the typical D peak at 1355 cm À1 , G peak at 1581 cm À1 and 2D peak at 2706 cm À1 in the Raman spectra (Fig. 3(c)). 47I D /I G peak ratio ($0.2) and I 2D /I G ($0.5) are indicative of high-quality and multi-layered graphene.Note that peaks corresponding to NiO structure are not observed due to the screening effect caused by graphene on the graphene/NiO surface.This is evidenced by the Raman spectra with wavenumber ranging from 300 to 1000 cm À1 , where no NiO peaks are observed (Fig. S2 ‡).Furthermore, fewer graphene layers were grown on NiO by decreasing the concentration of carbon precursor gas (C 2 H 4 ) during graphene synthesis; thus, LO-mode peak for NiO at $490 cm À1 is observed in the Raman spectra (Fig. S3 ‡).The presence of graphene is further conrmed by the intense carbon peak in the EDS spectra (Fig. 3(e)).The weak carbon (C) peak for NiO-Ni (Fig. 3(d)) is presumed to be artifacts (i.e., carbon conductive tape).The areal density of graphene ($0.17 mg cm À2 ) was measured by the weight difference of the sample before and aer CVD growth.

Lithium-ion battery performance
The electrochemical properties of the 3D graphene-NiO-Ni working electrode were tested using a CR2032 coin-type halfcell.Fig. 4(a) illustrates voltage vs. areal capacity proles for the rst two cycles at a current density of 0.1 mA cm À2 .For the rst discharge of the 3D graphene-NiO-Ni anode, an extended plateau region ($1.5 mA h cm À2 ) was observed between 0.25 V and 0.5 V.The plateau region is ascribed to the formation of solid electrolyte interphase (SEI) on the surface of the 3D graphene-NiO-Ni through electrolyte decomposition; it is also due to the reduction from Ni 2+ to Ni 0 by Li + ion uptake into NiO based on NiO + 2Li + 4 Li 2 O + Ni, forming Li 2 O. 4,23 It should be noted that the voltage uctuation at the plateau region might come from the large amount of SEI formation at the grapheneelectrolyte interface.Nanostructured materials are commonly susceptible to their unstable or irreversible capacity loss that occurs in the rst cycle. 48The gradual increase of the rst charge prole up to 2.0 V shows the plateau region of 2.0-2.1 V, in which oxidation reaction from Ni 0 to Ni 2+ occurred, forming NiO. 23,3031,32 However, coulombic efficiency increases up to $98% acquired by the ratio of charge (1.16 mA h cm À2 ) to discharge (1.18 mA h cm À2 ) capacities from the second cycle.The gradual decrease in voltage prole for the discharge process was observed in 1-1.5 V; the prole is related to the reduction reaction from Ni 2+ to Ni 0 . 4,23The second discharge curve exhibited the voltage plateau region at $1.5 V; the higher voltage over the rst discharge is closely related to the large variation of NiO microstructure and texture involved in the irreversible formation of Li 2 O and the decomposition of SEI layer formed during the rst cycle. 4For the second charge, the voltage prole was similar to that of the rst charge except with a slightly higher plateau voltage range of 2.1-2.2V due to increased anodic polarization in a cell. 30The overall voltagecapacity curves showed similar proles to the previously reported NiO-based anodes. 4,15,17,23,30Our prepared 3D NiO-Ni anode before graphene growth demonstrates different voltage proles with negligibly low areal capacities (<0.01 mA h cm À2 in the inset of Fig. 4(a)) which were seemingly due to cell resistance from a high NiO weight ratio ($51% in our experiments).These results ensured that graphene grown on 3D NiO-Ni structure contributed to such improvement in LIB performance by providing efficient conducting pathways among NiO phase regions and structural buffers against structural strains induced by large volume variations of NiO during cycling.Therefore, a facile electronic transfer from bulk electrode to electroactive NiO nanomaterials was achieved 23,32,[49][50][51] while the structural integrity of 3D NiO was preserved. 23,32The cyclic voltammetry (CV) curves for the 3D graphene-NiO-Ni and 3D NiO-Ni anode samples are displayed in Fig. S4.‡ It is noted that 3D graphene-NiO-Ni demonstrates peaks corresponding to NiO (two anodic peaks at $1.7 and $2.2 V, respectively, and a cathodic peak at $1.3 V) and graphene (a cathodic peak at $0.01 V).These results are in line with the voltage vs. capacity proles of Fig. 4(a).
Moreover, Fig. 4(b) demonstrates the cycling performance of the LIB cell for the 3D graphene-NiO-Ni as a function of current density.The areal capacities are 1, 0.9, 0.7, 0.6, 0.2 and 0.1 mA h cm À2 at 0.2, 0.5, 1, 2, 5 and 10 mA cm À2 , respectively; the values are higher than those from other reported NiO anodes. 18,19The average areal capacity of the anode (0.6 mA h cm À2 ), aer running through 10 mA cm À2 , recovered to the former value of 2 mA cm À2 (higher than 99% capacity retention); this conrmed high structural integrity and rate capability of our proposed 3D graphene-NiO-Ni anode.The subsequent cycling performance at 1 mA cm À2 resulted in an average areal capacity of 0.75 mA h cm À2 , which is $140% higher than the previous reports on nanoscale NiO anodes; 18,19 thus, the resulting capacity retention ($90%) indicated excellent cell stability.Furthermore, cycling stability of the anode at 1 mA cm À2 for 100 cycles is presented in Fig. 4(c); the overall coulombic efficiency of the anode is nearly 100% aer the rst cycle (69%) (Fig. 4(d)).Based on the promising LIB performance, we conrmed that graphene served as an important electrical conducting and structural buffering agent for the NiO, which addressed pitfalls of NiO such as its insulating nature and capacity loss induced by large volume variation during cycling.Additionally, the self-supporting 3D graphene-NiO-Ni structure required no binder, current collector or conducting agent (e.g., carbon black) for anode fabrication.Such materials will act as inefficient deadweight constituents. 33Although 3D graphene-NiO-Ni anode demonstrated improved areal capacity, gravimetric specic capacity could not be acquired due to the difficulty in determining weight fraction among the three components of the anode (i.e.graphene, NiO and Ni) aer the CVD processing.Nevertheless, as mentioned earlier, no one has yet applied the current CVD approach to enhance the areal capacity of NiO and LIB efficiency with graphene.Thus, the importance of our results from the CVD approach would be realized by its implementation into practical applications for large-scale LIBs and other energy storage systems.

Conclusion
We have fabricated a novel 3D graphene-NiO-Ni anode via a simple two-step thermal CVD method to increase its areal density up to $23 mg cm À2 .Such value provides a higher amount of active materials for LIBs which is important for largescale practical applications.The in situ graphene grown on the reduced 3D NiO-Ni exhibits wrinkled, high-quality, and multilayered structures.Such growth is a simple and highly effective method for large-scale coating of graphene onto a nanoporous electrode.While graphene layer grown on NiO is effective as an electrically conducting and structurally buffering for the nano-porous NiO, the 3D graphene-NiO-Ni anode exhibits a high rate capability with an areal capacity of 1.2 mA h cm À2 at 0.1 mA cm À2 .Our 3D graphene-NiO-Ni anode could be applied to ever-expanding development of large-scale, advanced LIBs.

Fig. 1
Fig. 1 Schematic representation of the steps to synthesize 3D graphene-NiO-Ni with the corresponding cross-sectional models and their corresponding field emission scanning electron microscope (FESEM) images: (a) structure of porous Ni foam; (b) NiO grown on the porous Ni foam by a thermal CVD oxidation process; (c) graphene grown on the NiO-Ni foam by a two-step process of H 2 -assisted reduction of NiO and subsequent graphene growth.
Fig.2(a and b) demonstrate low and high magni-cation FESEM images showing disordered sub-micron NiO nanoparticles on the surface of Ni foam.3,4,8,9Moreover, the cross-sectional images display columnar structured NiO layers grown on porous Ni foam (Fig.1(b) and its enlarged FESEM image is included in an inset of Fig.2(a)).The growth mechanism of NiO on Ni foam is dictated by the thermal diffusion and reaction of Ni2+  and O 2À ions in Ni foam according to the Kirkendall effect.[37][38][39]Thermally induced volume expansion facilitates outward diffusing of Ni + ions through grain and grain boundaries of crystalline NiO, thus forming columnar NiO structures.The X-ray diffraction (XRD) pattern conrms the evolution of NiO phase grown on Ni by thermal CVD oxidation (Fig.2(c)).The strong intense XRD peaks appeared at 37.1 , 43.2 , 62.7 , 75.3 and 79.3 , which corresponded to the crystallographic plane indices of (1 1 1), (2 0 0), (2 2 0), (3 1 1) and (2 2 2) for a cubic NiO phase, respectively (JCPDF card 47-1049).The average size of the NiO crystallites was about 27.2 nm by Scherrer equation; this value is comparable to the previously reported NiO nanoparticles in their constituent porous NiO. 40In addition, the presence of the remaining Ni phase in the 3D NiO-Ni aer thermal oxidation was manifested by intense XRD peaks observed at 44.2 , 51.6 and 76.1 corresponding to (1 1 1), (2 0 0) and (2 2 0) for Ni, respectively (JCPDF card 4-850).Note that the presence of the Ni phase has the advantage of enhancing electrical conductivity of NiO and catalytic activity that facilitates decomposition of Li 2 O and formation of the solid electrolyte interphase (SEI) layer during the charging process.

Fig. 2
Fig. 2 (a, b) Low and high magnification FESEM images demonstrating the surface morphologies of a thermally grown 3D porous NiO on Ni, respectively (the inset of (a) illustrates a cross-sectional image of 3D NiO on Ni with high magnification); (c) XRD patterns and (d) Raman spectra of the 3D NiO-Ni hybrid and pristine 3D Ni structures; (e) SEM image of 3D NiO-Ni for EDS mapping; (f, g) EDS mapping results from nickel and oxygen elements comprising 3D NiO-Ni structure, respectively.

Fig. 3
Fig. 3 Cross-sectional FESEM images of characteristic 3D graphene grown on porous NiO structure with (a) low and (b) high magnifications, respectively; (c) Raman spectra that identify graphene on NiO structure in terms of D-, G-and 2D-band characteristic peaks.EDS spectra reveal constituting elements in (d) 3D NiO-Ni and (e) 3D graphene-NiO-Ni.

Fig. 4
Fig. 4 Electrochemical performance of the 3D graphene-NiO-Ni and 3D NiO-Ni anodic materials; (a) characteristic voltage profiles of the 3D graphene-NiO-Ni and 3D NiO-Ni for the first two cycles; (b) C-rate capability of the 3D graphene-NiO-Ni anode at the six different current densities; the denoted numbers represent the applied current densities with a unit of mA cm À2 ; (c) cycling performance and (d) coulombic efficiency of the 3D graphene-NiO-Ni anode at 1 mA cm À2 for 100 cycles.