Coordination-driven self-assembly: construction of a Fe₃O₄–graphene hybrid 3D framework and its long cycle lifetime for lithium-ion batteries

Abstract

In this study, we have demonstrated coordination-driven self-assembly between Fe₃O₄ and graphene sheets under a hydrothermal condition for the simple in situ synthesis of a 3D Fe₃O₄–graphene hybrid architecture (Fe₃O₄/G). Fine hierarchical Fe₃O₄ spheres were homogeneously dispersed and embedded in an interconnected mesoporous framework of graphene sheets. It can be noted that the critical concentration of GO assembly decreased dramatically during the self-assembly of Fe₃O₄, indicating the coordination-driven self-assembly between Fe₃O₄ and GO. When evaluated as an anode material for LIBs, the Fe₃O₄/G hybrid framework demonstrates a high reversible capacity of 1164 mA h g⁻¹ over 500 cycles at a current density of 500 mA g⁻¹ and a remarkable rate capability. The superior electrochemical performance is attributed to a strong adhesion force and synergistic effect between Fe₃O₄ and graphene sheets as well as formation of a 3D interconnected hybrid framework, which offers enhanced kinetics towards electrochemical reactions with lithium ions and provides space for alleviating the huge volume expansion that occurs during charge–discharge cycles.

---

Introduction

One of the factors that limit the energy density of lithium ion batteries (LIBs) is the carbon-based anode materials like graphite, which have a low specific capacity of 372 mA h g⁻¹. Many efforts have been made toward developing new electrode materials with high energy densities to satisfy the ever-growing demand for high performance LIBs.^1–6 Currently, transition metal oxide (CuO, Co₂O₃, NiO, and Fe₃O₄) has attracted increasing attention because of many advantages such as safety, nontoxicity and high theoretical capacities.^7–11 Among them, Fe₃O₄ is regarded as a promising anode material to replace graphite electrodes due to its low cost, eco-friendliness, natural abundance, and relatively high electronic conductivity compared with other metal oxides.^12–15 Fe₃O₄ electrodes are based on the “conversion reaction” in the charge–discharge cycles, resulting in a large volume expansion (>200%) and poor capacity retention, which greatly hinder the practical utilization of Fe₃O₄ anodes.^5,16–18

Downsizing particles to the nanoscale with special morphologies is an effective way to improve the electrochemical properties because of the high surface areas, large pore volumes and short diffusion lengths for both Li ions and electrons. In particular, the hierarchical micro-structure constructed by nanoparticles has aroused tremendous interest due to the combined advantages of nano- and micro-materials.^19–21 Another feasible approach is to form hybrid materials by introducing conducting carbon matrices to absorb the volume changes and improve the electronic conductivity of the electrode.^22–24 Graphene, a two-dimensional single layer sheet of hexagonal carbon, has emerged as one of the most appealing carbon matrices for metal oxide particles because of its excellent electrical conductivity, high surface area, and good chemical and thermal stability. Graphene oxide (GO), which has abundant functional groups and exhibits strong adhesion to metal ions, is often used in situ for the preparation of nanoparticles/graphene composites.^25,26 Recent works have revealed that 2D graphene oxide assembles into 3D architectures under hydrothermal conditions if the concentration of GO is at least 1 mg mL⁻¹.²⁷ Using a similar method, a Fe₃O₄–graphene hybrid framework can also be prepared by introducing Fe₃O₄ nanocrystals or Fe³⁺ into a GO suspension (GO concentration is up to 1.5 mg mL⁻¹). With this in mind, numerous GO-based Fe₃O₄ host materials have been successfully synthesized and show enhanced properties.^28–32 However, most of the studies focused on the assembly of graphene oxide. There are still very few reports on the simultaneous co-assembly process of Fe₃O₄ nanocrystals and graphene oxide.

In this study, we prepared a three-dimensional graphene framework with a uniform distribution of hierarchical Fe₃O₄ spheres via a one-pot solvothermal method. Accompanied with the formation and assembly of Fe₃O₄ crystals, GO is also reduced and assembled into a 3D network at the same time. The two assembly processes reinforced and interacted with each other through coordination-driven self-assembly to form the integral Fe₃O₄–graphene hybrid framework (Fe₃O₄/G). The critical feature of this method is the usage of an ethylene glycol–water system and sodium oleate. The as-formed 3D hybrid framework has the advantage of the two building blocks and exhibits an outstanding reversible capacity and excellent rate performance.

Materials and methods

Materials

Sodium oleate (96%, Aladdin), iron(III) chloride (FeCl₃·6H₂O, chemically pure, Sinopharm Chemical Reagent), ethylene glycol (chemically pure, Sinopharm Chemical Reagent), iron(III) sulfate hydrate, (Fe₂(SO₄)₃·xH₂O, chemically pure, Sinopharm Chemical Reagent), graphene oxide solution (2 mg mL⁻¹, Sinocarbon Graphene Marketing Center, Chinese Academy of Science), and ethanol (chemically pure, Sinopharm Chemical Reagent) were used as received without further purification.

Synthesis

Fe₃O₄/G was prepared by a one-pot solvothermal method. In a typical synthesis process, 5 mmol of sodium oleate was dissolved in 30 mL of the mixed solution (ethylene glycol–graphene oxide solution) by vigorous stirring for 1 h, followed by adding 3 mmol of FeCl₃·6H₂O to form a uniform gray dispersion. Subsequently, the mixed dispersion was sealed in a 50 mL Teflon-lined autoclave and hydrothermally treated at 200 °C for 20 h. After cooling down to room temperature, the product was centrifuged and washed several times with ethanol and deionized water before drying at 60 °C in an oven. As a control experiment, Fe₃O₄/G hybrid with different contents of GO (defined as the Fe₃O₄/G-x, x represent the volume of GO solution) were prepared by varying the volume of GO solution using the same procedure. For comparison, we synthesized bare Fe₃O₄ microspheres without any GO using the same experimental conditions and a mixture, namely, Fe₃O₄/G mixed, was prepared by physically mixing Fe₃O₄ microspheres with graphene powder. To remove the organic ligands, all samples are calcined at 500 °C for 3 h in argon.

Materials characterization

X-ray diffraction (XRD) patterns were obtained on a Philips PW1050 X-ray powder diffractometer with Cu K_α radiation (λ = 1.5406 Å). Morphology and microstructure of the products were investigated by transmission electron microscopy (TEM, JEM-1230) at accelerating voltages of 80 kV and scanning electron microscopy (SEM, HITCHI-4800). Structure and selected area electron diffraction were measured by a high-resolution transmission electron microscope (HRTEM, TECNAI G2 F20) operating at 200 kV. Nitrogen adsorption and desorption experiments were carried out using TristarII3020 surface area and porosity analyzer at 77 K. The surface area was calculated using the Brunauer–Emmett–Teller (BET) equation. Pore size distributions were calculated by the Barrett–Joyner–Halenda (BJH) method using the desorption branch of the isotherm curves. Thermogravimetric analysis (TGA) was performed using a CHNS/O analyzer (PE 2400II, Perkin Elmer, America) in air atmosphere from room temperature to 800 °C with a heating rate of 10 °C min⁻¹. The Raman spectra were obtained on DXR SmartRaman (Thermo Fisher, America). All the Fe 2p, O 1s and C 1s XPS spectra were obtained by an Escalab 250Xi spectrometer operating at an Al K_α radiation source.

Electrochemical characterization

The electrochemical properties of the obtained F₃O₄/G, Fe₃O₄/G mixed and pure Fe₃O₄ samples were evaluated by assembling CR2032 coin type cells in an argon-filled glove box. The working electrode was prepared by forming slurry of Fe₃O₄/G, carbon black and sodium alginate in a weight ratio of 70 : 20 : 10. Deionized water was used as the solvent. The slurry was coated on copper foil disks and dried overnight in a vacuum oven at 120 °C. Coin cells were fabricated using Li foil as the counter electrode, a porous polyethene film as the separator and LiPF₆ (1 M) in ethylene carbonate/dimethyl carbonate/diethyl carbonate (EC/DMC/DEC, 1 : 1 : 1 vol%) as the electrolyte. The charge–discharge curves were measured using a Neware Cell tester (Shenzhen Neware Technology Limited Co., China) at a cut off voltage of 0.02–3 V vs. Li/Li⁺ at different rate at room temperature. The cyclic voltammetry (CV) data were collected with an Arbin electrochemical workstation at a scan rate of 0.1 mV s⁻¹ between 0.005 and 3.0 V. Electrochemical impedance spectroscopy (EIS) measurements were performed on an electrochemical workstation (CHI600E, CH Instruments, Inc.) under a frequency range between 100 kHz and 0.01 Hz with an applied voltage of 5 mV.

Results and discussion

The F₃O₄/G was prepared via a one-pot solvothermal approach based on coordination-driven self-assembly. Fig. 1 shows a plausible model of the 3D framework formation, which can be described as follows. In the first step, Fe-ions could be anchored on graphene sheets through the functional group on the GO surface due to the good miscibility of GO in the ethylene glycol–water suspension. The negatively charged functional groups on the surface of GO could absorb positively charged Fe³⁺ ions because of electrostatic attraction, which might serve as nucleation and assembly centers for Fe₃O₄. In fact, every nucleation center could gather a lot of Fe³⁺ ions and sodium oleate molecules, and more GO would be anchored around the center due to charge balance principle. The second step is a hydrothermal reaction, in which the nucleation and assembly of Fe₃O₄ occurred simultaneously with the reduction and assembly of GO. In our system, the usage of sodium oleate was very important due to its dual role as a surfactant and reaction assistant. Experiments show that GO will form a hydrogel in ethylene glycol–water systems, which strongly depend on the volume of GO solution (V_GO). If V_GO was low (e.g., 4 mL), only a black powdery material was produced after hydrothermal reduction (Fig. S1d†). Only when the V_GO was increased to 6 mL or above, mechanically stable graphene hydrogels were obtained (Fig. S1b†). However, the threshold of GO assembly decreased dramatically when the assembly of Fe₃O₄ was introduced into the system and eventually formed a 3D F₃O₄/G hybrid framework. Even if the added volume of GO was as low as 2 mL, a Fe₃O₄/G hybrid-gel could still be obtained (Fig. S1e†). This means that the mass loading of active materials in our hybrid product is higher than that of most other Fe₃O₄–graphene composites in previous reports,^33,34 which is of great importance in commercial applications. Moreover, the size of the Fe₃O₄/G gel was smaller than that of the pure GO gel at the same volume of GO solution (Fig. S1a and b†), which illustrates that the hybrid forms a more dense and stable gel in the presence of the Fe₃O₄ assembly. In addition, the existence of GO was helpful for forming uniform Fe₃O₄ hierarchical particles (Fig. S2†). Therefore, we defined this mutual interaction between Fe₃O₄ and GO as coordination-driven self-assembly. To further confirm this mutual effect, parallel experiments with different Fe sources were carried out. As shown in Fig. S1f,† the hybrid hydrogel could not be produced without the self-assembly of Fe₃O₄. Moreover, in situ nucleation and/or self-assembly can lead to a strong binding force between Fe₃O₄ and GO, which is beneficial to electronic transmission and structural integrity.

Fig. 1 Schematic illustration of the synergistic effect of self-assembly and the formation of Fe₃O₄/G hybrid 3D framework.

The crystallographic structure and phase purity of the F₃O₄/G, Fe₃O₄/G mixed and bare Fe₃O₄ were determined by X-ray powder diffraction (XRD), as shown in Fig. 2. It was found that all the diffraction peaks can be well indexed to the magnetic cubic structure of Fe₃O₄ (JCPDS 79-0419). The mean crystallite size of the Fe₃O₄/G and Fe₃O₄/G mixed are 18.6 and 23.6 nm, respectively, as calculated using the Scherrer equation according to the (311) peaks. In addition, a diffraction peak at ∼26° corresponding to graphene was observed in the pattern of Fe₃O₄/G mixed, which can be ascribed to the reduction and stacking of GO. Remarkably, no apparent peak can be identified at 20–30° in the Fe₃O₄/G sample, indicating that the Fe₃O₄ spheres are efficiently deposited on the surface of graphene, suppressing the stacking of graphene layers, which reflects the fact that the F₃O₄/G is composed of few-layer stack graphene sheets.

Fig. 2 XRD patterns of the Fe₃O₄/G, Fe₃O₄/G mixed and bare Fe₃O₄.

The morphologies and microstructure of the samples were characterized by SEM and TEM. As shown in Fig. 3a and b, without the presence of GO, the bare Fe₃O₄ spheres are constructed by primary nanoparticles and have a wide size distribution of 100–500 nm. Fig. 3c is a typical SEM image of the Fe₃O₄/G framework. From the image it is clear that the Fe₃O₄ spheres are uniformly wrapped by graphene sheets and each of them has a diameter of about 210 nm (Fig. S2†). The TEM image also reveals that these Fe₃O₄ particles are firmly attached to the graphene sheet, even after extended sonication was used to disperse the Fe₃O₄/G hybrid framework for TEM characterization. We believe that this perfect combination prevents the agglomeration of nanoparticles and also enables fast electron transport through the underlying graphene layers to Fe₃O₄, guaranteeing the efficient electrochemical performance. The high-resolution TEM images in Fig. 3e and f show the interfacial structure between the graphene and Fe₃O₄ spheres. The well-resolved lattice fringes with an interplane distance of 0.253 nm are caused by the (311) plane of Fe₃O₄. It should be emphasized that the volume of GO solution is a critical factor in our one-pot synthesis approach. We found that when the volume of the GO solution was 6 mL, the Fe₃O₄ spheres would be anchored by graphene sheets to form hybrid hydrogels and do not present an obstacle to the self-assembly of Fe₃O₄ (Fig. 2c and d and S3b†). A small amount of GO solution would result in ineffective wrapping around the Fe₃O₄ spheres (Fig. S3a†), while a large amount of GO solution would damage the assembled structure of the Fe₃O₄ spheres (Fig. S3c and 3d†). Therefore, we chose the Fe₃O₄/G (6 mL GO solution) sample for the following structure and properties characterization.

Fig. 3 (a) and (b) are the SEM and TEM images of bare Fe₃O₄ spheres; (c) and (d) are the SEM and TEM images of Fe₃O₄/G framework; (e) is a magnified image of a single Fe₃O₄/G composite, and (f) is a HRTEM image of Fe₃O₄/G framework.

Fig. 4a presents N₂ adsorption–desorption isotherms for Fe₃O₄/G at 77 K. The isotherm can be identified as a type-IV isotherm with an H3 hysteresis loop that is attributed to a mesoporous structure, which indicates that the Fe₃O₄ spheres have hierarchical structure. The pore size distribution for Fe₃O₄/G is shown in the inset of Fig. 4a and it demonstrates the formation of randomly distributed pores with sizes from 4 to 16 nm with dominant 4 and 8 nm pores. The isotherms of Fe₃O₄/G mixed and pure Fe₃O₄ have similar results, as shown in Fig. S4a–d.† Although the BET specific area of Fe₃O₄/G is almost the same for both Fe₃O₄/G mixed and pure Fe₃O₄, the total pore volume of Fe₃O₄/G is about two times larger than that of the two others as shown in Fig. 4b. Combined with the TEM and SEM images, we conclude the pore structures may benefit electrolyte ion diffusion to active sites with less resistance and tolerate the volume changes of Fe₃O₄/G during the charge–discharge cycles.

Fig. 4 (a) N₂ adsorption–desorption isotherms of the Fe₃O₄/G sample; (b) comparison of the BET surface areas and total pore volumes of the Fe₃O₄/G, Fe₃O₄/G mixed and bare Fe₃O₄ samples. The inset in (a) shows the pore size distribution calculated by the BJH method.

X-ray photoelectron spectroscopy (XPS) analysis was employed to further characterize the product. The wide scan XPS spectrum of the Fe₃O₄/G hybrid framework demonstrates the presence of Fe, O, and C elements (Fig. S5a†). Fig. 5a–b exhibits the high-resolution XPS spectrum of Fe₃O₄/G. In Fig. 5a, the Fe 2p_1/2 and 2p_3/2 peaks located at around 711 and 724.9 eV are observed, which indicates that the presence of both the Fe²⁺ and Fe³⁺ ions in Fe₃O₄. For carbon, the C 1s XPS profiles in Fig. 2c not only shows the presence of the carbon element but also reveals the efficient reduction of GO. After solvothermal reaction, the sp² C=C bonds at 284.5 eV increased, while the epoxy C–O bonds at 286.1 eV and C=O bonds at 288.3 eV decreased, indicating an efficient reduction.³⁵ The structural and electronic properties of graphene in the Fe₃O₄/G hybrid have been investigated using laser Raman spectroscopy. For comparison, we also measured the Fe₃O₄/G mixed and bare Fe₃O₄ and the results are shown in Fig. 5c. Both the Fe₃O₄/G hybrid and Fe₃O₄/G mixed show two prominent peaks, one is the G band at about 1600 cm⁻¹, which corresponds to ordered sp² bond carbon, and the other is the D band at around 1350 cm⁻¹, which results from edges, other defects and disordered carbon in the symmetrical hexagonal graphitic lattice.³⁶ Compared to the Fe₃O₄/G mixed sample, the shift of the peaks can be found for both D and G bands in Fe₃O₄/G hybrid, indicating a significant charge transfer between the graphene nanosheets and Fe₃O₄ spheres.¹⁵ The appearance of weak D and G peaks of the bare Fe₃O₄ can be ascribed to the carbonization of ligands during calcining. The content of Fe₃O₄ in the composite was evaluated by thermogravimetric (TG) analysis, with the results shown in Fig. 5d. The weight loss of ∼0.4% up to 150 °C is attributed to the release of water adsorbed on the samples. The weight loss that occurs during the heat treatment from room temperature to 800 °C is attributed to the oxidation of graphene/carbon, and a weight gain afterwards is believed to correspond to the oxidation of Fe₃O₄ in air. As for bare Fe₃O₄, no obvious weight loss is found, indicating that the weight gain (oxidation of Fe₃O₄) is equal to the weight loss (oxidation of amorphous). Therefore, the amorphous carbon in bare Fe₃O₄ can be derived to be about 4 wt%. For the Fe₃O₄/G and Fe₃O₄/G mixed samples, based on the total weight loss of carbon materials (9.6% and 8.4%), the original weight fraction of Fe₃O₄ is calculated as 86.4% and 87.6%, respectively, and the content of graphene in Fe₃O₄/G and Fe₃O₄/G mixed are calculated to be 5.6% and 4.4%, which is approximately consistent with the theoretical calculations. TG results show that the Fe₃O₄/G hybrid framework has a relatively high mass loading of active materials compared with other Fe₃O₄–graphene composites (Table 1).

Fig. 5 XPS spectra of core-level Fe 2p (a) and C 1s (b) for Fe₃O₄/G hybrid framework; Raman spectra (c) and TGA curves (d) of Fe₃O₄/G, Fe₃O₄/G mixed and bare Fe₃O₄ samples.

Table 1 The electrochemical properties of various Fe₃O₄-based electrode materials^a

Materials	Ratio	Cycles (density)	Capacity	Retention	Ref.
a (1 C = 1 A g^−1). Ratio: the weight percentage of active material in composites. Retention: the capacity retention from the second to the end of cycles.
Fe3O4–graphene	87.6%	500 (0.5 C)	1164	99.4%	This work
		50 (1 C)	831	—
Fe3O4–graphene	67%	300 (5.2 C)	577	∼96.1%	34
Fe3O4–graphene	70%	50 (1 C)	612	∼47.1%	32
Fe3O4–graphene	40%	50 (0.05 C)	675	∼73.4%	31
Fe3O4–graphene	87.2%	300 (0.1 C)	1107	∼83.5%	28
Fe3O4–C–graphene	50.2%	100 (0.1 C)	660	∼72.1%	54
Fe3O4–graphene	87.7%	100 (0.7 C)	∼560	∼91.0%	30
Fe3O4–graphene	∼70%	30 (0.2 C)	1100	∼91.6%	12
Fe3O4–C	45.4%	100 (0.05 C)	610	∼55.5%	38
Fe3O4–C	83%	100 (1 C)	831	∼83.1%	41
Fe3O4–C	89.8%	120 (0.2 C)	1232	107%	40

---

In order to characterize the electrochemical properties of the samples, half-cells with the lithium plate as the counter electrode were fabricated. Cyclic voltammograms (CVs) of the Fe₃O₄/G and bare Fe₃O₄ were measured between 0.005 and 3 V at a scan rate of 0.1 mV s⁻¹. As seen in Fig. 6a and Fig. S6a,† it is obvious that both the samples demonstrate similar well defined redox peaks signifying the lithiation and delithiation of Fe₃O₄. In the first scan, the cathodic peak appearing at 0.52 V is attributed to the reduction of Fe³⁺ and Fe²⁺ to Fe⁰ (Fe₃O₄ + 8Li → 3Fe + 4Li₂O) and the formation of SEI layer. In the first anodic process, two peaks were observed at 1.65 and 1.85 V in Fe₃O₄/G, attributed to the oxidation of Fe⁰ to Fe²⁺ and Fe³⁺, respectively.^37,38 It is clear that the subsequent cycles are quite different from the first one. Both the anodic and cathodic peaks are positively shifted, which is attributed to structural modification after the first cycle due to lithium insertion and extraction.³⁹ It can be noted that the cathodic peaks at 0.75 V and the anodic peaks located at about 1.70 and 1.96 V for the second or fifth cycle of the Fe₃O₄/G electrode almost overlap, indicating the high reversibility of electrode.^40,41 The charge–discharge curves of the Fe₃O₄/G hybrid and bare Fe₃O₄ electrodes for the 1^st, 2^nd, 100^th, 200^th, and 500^th cycles are illustrated in Fig. 6b and Fig. S6b.† As shown in Fig. 6b, the first discharge and charge capacities of Fe₃O₄/G are measured to be 1573 and 1113 mA h g⁻¹, respectively, resulting in an initial Coulombic efficiency of 72.4%. The bare Fe₃O₄ exhibits a similar electrochemical performance at the first cycle, with the first discharge and charge capacities of 1216 and 924 mA h g⁻¹, producing a Coulombic efficiency of 75.9% (Fig. S6b†). We note that the first discharge capacity is much higher than that of the theoretical value for Fe₃O₄ and graphene; moreover, the reasons for exceeding theoretical capacities can be ascribed to the formation of a solid electrolyte interphase resulting from electrolyte degradation at low voltage and additional interfacial/active site Li storage mechanism due to nano-scale and mesoporous structures, as well as the synergistic effect between the graphene and the Fe₃O₄ particles.^42,43 The initial Coulombic efficiency of bare Fe₃O₄ is slightly higher than that of Fe₃O₄/G, which may be ascribed to the lower surface area of bare Fe₃O₄ (Fig. 4b). The irreversible capacity loss can be attributed to the inevitable formation of a surface SEI layer accompanying the electrolyte decomposition during discharge. The long plateau at 0.7 V in the initial discharge curve corresponds to the reversible reduction from Fe³⁺ (or Fe²⁺) to Fe⁰ [Li_xFe₃O₄ + (8 − x)Li⁺ + (8 − x)e⁻ → 4Li₂O + 3Fe] and the irreversible formation of a solid electrolyte interphase film. From the second cycle, this plateau increases to a slightly higher voltage and becomes much shorter, indicating that different lithiation reactions take place between the first and the following cycles and the existence of irreversible reactions. With cycling, no obvious changes to either charge or discharge profiles are observed, indicating a highly reversible electrode reaction process and a stable electrode structure characteristic. In the case of the bare Fe₃O₄, although the first discharge curve is very similar to that of the Fe₃O₄/G, the specific capacity of the bare Fe₃O₄ electrode decreases rapidly due to volume expansion in the subsequent charge–discharge cycles, as shown in Fig. S6b.†

Fig. 6 (a) Cyclic voltammogram curves of Fe₃O₄/G hybrid at a scan rate of 0.1 mV s⁻¹; (b) charge and discharge profiles of the Fe₃O₄/G hybrid at a current density of 500 mA g⁻¹ between 0.02 and 3 V; (c) cycling performance and Coulombic efficiency of Fe₃O₄/G, Fe₃O₄/G mixed and bare Fe₃O₄ samples. Charge–discharge curves of Fe₃O₄/G anode for: (d) stage A; (e) stage B.

Fig. 6c shows the long-term cycling performance and the corresponding Coulombic efficiency of the Fe₃O₄/G hybrid framework and bare Fe₃O₄ electrodes in a potential window of 0.02–3 V vs. Li⁺/Li at a current density of 500 mA g⁻¹. In order to clearly demonstrate the superior electrochemical performance and synergistic effect in Fe₃O₄/G, we prepared the Fe₃O₄/G mixed sample by physically mixing graphene with Fe₃O₄ spheres to offer comparisons with the Fe₃O₄/G electrode. In our study, the capacity is calculated based on the total mass of the composite, including Fe₃O₄ and graphene sheets. As can be seen, the Fe₃O₄/G electrode exhibits higher capacity and good cyclic retention compared to Fe₃O₄/G mixed and bare Fe₃O₄. The initial discharge capacities are 1537, 1247 and 1216 mA h g⁻¹ for Fe₃O₄/G, Fe₃O₄/G mixed and bare Fe₃O₄, respectively. After 500 cycles, the Fe₃O₄/G still delivers a discharge of 1164 mA h g⁻¹, while the capacities of Fe₃O₄/G mixed and bare Fe₃O₄ decrease dramatically to 702 and 297 mA h g⁻¹, respectively. Moreover, Fe₃O₄/G still possesses better lithium storage properties even compared with Fe₃O₄/G mixed (the same graphene content), indicating that there exists a strong interaction force between Fe₃O₄ and graphene, which is consistent with the TEM and Raman results. In addition, the Coulombic efficiency of Fe₃O₄/G increased rapidly after several cycles and steadily reached up to 99%. Moreover, it can be noted that the specific capacity undergoes a gradual rising process after the slow fading process in the first 80 cycles. Such a phenomenon of increasing capacity upon cycling has been found in Fe₃O₄ and other metal oxide anode materials.^29,44–48 To clarify the mechanism, the cycle curve of Fe₃O₄/G hybrid was divided into two different steps based on the variation of the specific capacity, as illustrated in Fig. 6c–e. Stage (A): an obvious drop in specific capacity was recorded, which may be ascribed to the cracking of some of un-anchored Fe₃O₄ particles. Stage (B): the specific capacity stabilizes in the range of 80–150 cycles and increases after 170 cycles. This may be caused by three factors as follows. First, several discharge–charge cycles are needed to form a stable SEI film, allowing the greigite crystals to percolate throughout and establish intimate contacts with the current collector.^49,50 Second, the slight increase of capacity is due to the reversible growth of a polymeric gel-like film caused by the decomposition of electrolytes at a low potential, which enables mechanical cohesion and delivers excess capacity through a so-called “pseudo-capacity-type” mechanism. Last but not least, an anomalous monotonic increase in the discharge capacity was observed from 884 (102^th) and 1015 (201^th) to 1118 mA h g⁻¹ (260^th) and is followed by a nearly constant capacity at 1160 mA h g⁻¹ till the 500^th cycle. The voltage profiles are shown in Fig. 6e. The discharge curves overlap basically at ∼0.8 V, which indicates the sustained excellent reversibility of the traditional conversion mechanism in Fe₃O₄/G anode. However, the curve's decay below 0.8 V slows down and the voltage hysteresis becomes narrower as confirmed by the closer discharge–charge voltage profiles with cycling, which results from faster kinetics and an improved energy efficiency.^51,52 The electrochemical properties of the Fe₃O₄/G hybrid framework prepared in the present study and those of the iron oxide materials reported in the literature are summarized in Table 1. It can be seen that the Fe₃O₄/G hybrid electrode is of high capacity and excellent cycling stability, as well as a high ratio of active materials in composites. The reasons for high specific capacity and good cycling stability of the Fe₃O₄/G hybrid can be ascribed to the following factors: (1) the hierarchical structure and uniform size distribution of Fe₃O₄ spheres can alleviate stress due to volume expansion in the discharge–charge process, (2) the incorporation of graphene not only enhances the conductivity of electrode, but also partially accommodates the volume change during cycling, (3) the strong adhesive force of Fe₃O₄ to graphene empowers fast electron transport through graphene to enhance the electrochemical performance. (4) The 3D hybrid framework constructed by hierarchical Fe₃O₄ spheres and flexible graphene sheets possesses a large number of pores and large electrolyte/electrode interface area, which suggests that it has more channels for Li-ion diffusion and reversible lithium storage site and (5) the Fe₃O₄ crystals are uniformly anchored among the network of graphene, which inhibits pulverization and maintains the integrity of electrode.

Benefiting from the unique structure of Fe₃O₄/G hybrid framework, the Fe₃O₄/G electrode exhibits exceptional rate performance. As shown in Fig. 7a, the electrode delivers capacities of 1280, 1240, 1145, 956, 831 and 510 mA h g⁻¹ when cycled at high rate of 100, 200, 400, 800, 1000 and 2000 mA g⁻¹, respectively, which present a superior rate performance in comparison with presently reported results. When the charge–discharge system is back to the 200 mA g⁻¹, the discharge capacities can recover. To demonstrate the synergistic effect between Fe₃O₄ and graphene, the rate capability of Fe₃O₄/G mixed is also investigated under the same condition. Compared to the Fe₃O₄/G electrode, Fe₃O₄/G mixed electrode delivers much lower specific capacities 200–1000 mA h g⁻¹ at current rates of 100–2000 mA g⁻¹ as a result of inefficient electron and lithium-ion transportation between Fe₃O₄ and graphene. As for the bare Fe₃O₄ electrode, the reversible capacity continuously decayed with cycling, and exhibited a capacitance less than 50 mA h g⁻¹ at a current of 2000 mA g⁻¹. In addition, the Fe₃O₄/G hybrid exhibits excellent rate cyclic performance, because there is no significant capacity loss at each rate test.

Fig. 7 (a) Rate performance of Fe₃O₄/G, Fe₃O₄/G mixed and bare Fe₃O₄ cycled between 0.02 and 3 V. (b) Nyquist plots at the OCP (5 mV perturbation) for the electrodes.

As an important index for insight into the electrochemical behaviour, electrochemical impedance spectroscopy (EIS) was conducted to identify the relationship between the electrochemical performance and electrode kinetics. All the coin cells are cycled by rate testing before EIS testing. The Nyquist plots for the samples are shown in Fig. 7b with a frequency range from 100 kHz to 0.01 Hz, which share the common feature of a depressed semicircle at high frequency and an inclined line at low frequency. The diameter of the semicircle for the Fe₃O₄/G hybrid framework electrode in the high-medium frequency region is much smaller than that of Fe₃O₄/G mixed or bare Fe₃O₄, indicating that the former electrodes possess the minimum charge transfer resistance (R_ct) among the three samples. This means that the rapid electron transfer during the electrochemical reaction of the composite can be ascribed to the increased contact area and strong binding force between Fe₃O₄ and graphene as well as the enhanced electrical conductivity of the overall electrode.^53,54 In the low frequency region, a more vertical straight line of Fe₃O₄/G compared to Fe₃O₄/G mixed and bare Fe₃O₄ is further evidence for the faster Li⁺ ion diffusion behavior of the Fe₃O₄/G hybrid framework electrodes. The Fe₃O₄/G hybrid framework presents satisfying electrochemical kinetics properties, leading to a remarkable improvement in cyclic stability and rate properties.

As shown by the results presented above, the Fe₃O₄/G hybrid electrode displays excellent electrochemical performance and structural stability. These outstanding properties can be attributed to several potential factors. Firstly, the 3D Fe₃O₄/G hybrid frameworks provide a large surface area and efficiently reduce the diffusion length of both electrons and lithium ions. Secondly, graphene sheets provide a highly conductive matrix for electron transfer during the lithiation and delithiation processes. Moreover, the intimate interaction between Fe₃O₄ and graphene sheets prevents the aggregation of Fe₃O₄ nanoparticles and the restacking of graphene nanosheets.

Conclusion

In summary, we have demonstrated the coordination-driven self-assembly of Fe₃O₄ with graphene sheets under hydrothermal conditions for simple in situ synthesis of a 3D Fe₃O₄/G hybrid architecture. The key factors of our approach are the introduction of ethylene glycol–water system and the addition of sodium oleate, which can ensure the occurrence of self-assembly of both Fe₃O₄ and graphene oxide. The two assembly processes mutually reinforced and interacted with each other, obtaining the integral Fe₃O₄/G hybrid framework. The Fe₃O₄/G hydrogel could be formed at a very low concentration of graphene oxide due to the coordination-driven self-assembly, which indicates the high mass proportion of active materials in a composite. Such a special and unique architecture established by the inter-dispersion of Fe₃O₄ spheres provides better protection against aggregation and volume changes of the active material, guarantees more lithium-storage sites, and ensures favourable transport kinetics for both electrons and lithium ions. As a result, the Fe₃O₄/G hybrid frameworks demonstrate outstanding enhancement of durability and rate performance compared with Fe₃O₄/G mixed and bare Fe₃O₄, with a very high reversible capacity of 1164 mA h g⁻¹ even after 500 cycles at a current density of 500 mA g⁻¹ and 510 mA h g⁻¹ at a high rate of 2000 mA g⁻¹. It is believed that this approach can be readily applied to other metal oxide–GO hybrid nanostructures in order to study their synergetic and self-assembly properties.

Acknowledgements

The authors gratefully acknowledge the financial support for this study from the National Natural Science Foundation of China (no. 51272231), and the Doctoral Fund of the Ministry of Education of China (no. 20100101110039).

Notes and references

1. K. Kang, Y. S. Meng, J. Bréger, C. P. Grey and G. Ceder, Science, 2006, 311, 977–980.
2. V. Augustyn, J. Come, M. A. Lowe, J. W. Kim, P.-L. Taberna, S. H. Tolbert, H. D. Abruña, P. Simon and B. Dunn, Nat. Mater., 2013, 12, 518–522.
3. B. Kang and G. Ceder, Nature, 2009, 458, 190–193.
4. M. Wu, J. E. Sabisch, X. Song, A. M. Minor, V. S. Battaglia and G. Liu, Nano Lett., 2013, 13, 5397–5402.
5. P. Poizot, S. Laruelle, S. Grugeon, L. Dupont and J.-M. Tarascon, Nature, 2000, 407, 496–499.
6. L. Zhang, H. B. Wu, Y. Yan, X. Wang and X. W. Lou, Energy Environ. Sci., 2014, 7, 3302–3306.
7. C. Zhu, D. Chao, J. Sun, I. M. Bacho, Z. Fan, C. F. Ng, X. Xia, H. Huang, H. Zhang, Z. X. Shen, G. Ding and H. J. Fan, Adv. Mater. Interfaces, 2015, 2, 1400499.
8. W. Wen, J.-M. Wu and M.-H. Cao, Nanoscale, 2014, 6, 12476–12481.
9. L. Li, H. B. Wu, L. Yu, S. Madhavi and X. W. D. Lou, Adv. Mater. Interfaces, 2014, 1, 1400050.
10. X. Li, A. Dhanabalan and C. Wang, J. Power Sources, 2011, 196, 9625–9630.
11. H. Wang, Z. Xu, H. Yi, H. Wei, Z. Guo and X. Wang, Nano Energy, 2014, 7, 86–96.
12. W. Chen, S. Li, C. Chen and L. Yan, Adv. Mater., 2011, 23, 5679–5683.
13. B. Lim, J. Jin, J. Yoo, S. Y. Han, K. Kim, S. Kang, N. Park, S. M. Lee, H. J. Kim and S. U. Son, Chem. Commun., 2014, 50, 7723–7726.
14. S. H. Choi, Y. N. Ko, K. Y. Jung and Y. C. Kang, Chemistry, 2014, 20, 11078–11083.
15. C. Ban, Z. Wu, D. T. Gillaspie, L. Chen, Y. Yan, J. L. Blackburn and A. C. Dillon, Adv. Mater., 2010, 22, E145–E149.
16. Y. Piao, H. S. Kim, Y. E. Sung and T. Hyeon, Chem. Commun., 2010, 46, 118–120.
17. S. Jin, H. Deng, D. Long, X. Liu, L. Zhan, X. Liang, W. Qiao and L. Ling, J. Power Sources, 2011, 196, 3887–3893.
18. Q. Q. Xiong, J. P. Tu, Y. Lu, J. Chen, Y. X. Yu, Y. Q. Qiao, X. L. Wang and C. D. Gu, J. Phys. Chem. C, 2012, 116, 6495–6502.
19. C. Yuan, H. B. Wu, Y. Xie and X. W. Lou, Angew. Chem., Int. Ed., 2014, 53, 1488–1504.
20. S. H. Lee, S. H. Yu, J. E. Lee, A. Jin, D. J. Lee, N. Lee, H. Jo, K. Shin, T. Y. Ahn, Y. W. Kim, H. Choe, Y. E. Sung and T. Hyeon, Nano Lett., 2013, 13, 4249–4256.
21. Y. Li, Z.-Y. Fu and B.-L. Su, Adv. Funct. Mater., 2012, 22, 4634–4667.
22. X. Fan, J. Shao, X. Xiao, L. Chen, X. Wang and S. Li, J. Mater. Chem. A, 2014, 2, 14641–14648.
23. Q. Liu, Z. F. Li, Y. Liu, H. Zhang, Y. Ren, C. J. Sun, W. Lu, Y. Zhou, L. Stanciu, E. A. Stach and J. Xie, Nat. Commun., 2015, 6, 6127.
24. F. Wu, R. Huang, D. Mu, B. Wu and S. Chen, ACS Appl. Mater. Interfaces, 2014, 6, 19254–19264.
25. R. Raccichini, A. Varzi, S. Passerini and B. Scrosati, Nat. Mater., 2014, 14, 271–279.
26. J. Yao, X. Shen, B. Wang, H. Liu and G. Wang, Electrochem. Commun., 2009, 11, 1849–1852.
27. Y. Xu, K. Sheng, C. Li and G. Shi, ACS Nano, 2010, 4, 4324–4330.
28. J. Yin, H. Shi, P. Wu, Q. Zhu, H. Wang, Y. Tang, Y. Zhou and T. Lu, New J. Chem., 2014, 38, 4036–4040.
29. Q. Zhou, Z. Zhao, Z. Wang, Y. Dong, X. Wang, Y. Gogotsiac and J. Qiu, Nanoscale, 2014, 6, 2286–2291.
30. G. Zhou, D.-W. Wang, F. Li, L. Zhang, N. Li, Z.-S. Wu, L. Wen, G. Q. Lu and H.-M. Cheng, Chem. Mater., 2010, 22, 5306–5313.
31. M. Sathish, T. Tomai and I. Honma, J. Power Sources, 2012, 217, 85–91.
32. S. Bhuvaneswari, P. M. Pratheeksha, S. Anandan, D. Rangappa, R. Gopalan and T. N. Rao, Phys. Chem. Chem. Phys., 2014, 16, 5284–5294.
33. T.-q. Wang, X.-l. Wang, Y. Lu, Q.-q. Xiong, X.-y. Zhao, J.-b. Cai, S. Huang, C.-d. Gu and J.-p. Tu, RSC Adv., 2014, 4, 322–330.
34. L. Fan, B. Li, D. W. Rooney, N. Zhang and K. Sun, Chem. Commun., 2015, 51, 1597–1600.
35. H.-K. Kim, S.-H. Park, S.-B. Yoon, C.-W. Lee, J. H. Jeong, K. C. Roh and K.-B. Kim, Chem. Mater., 2014, 26, 4838–4843.
36. R. Chen, T. Zhao, W. Wu, F. Wu, L. Li, J. Qian, R. Xu, H. Wu, H. M. Albishri, A. S. Al-Bogami, D. A. El-Hady, J. Lu and K. Amine, Nano Lett., 2014, 14, 5899–5904.
37. H.-S. Lim, B.-Y. Jung, Y.-K. Sun and K.-D. Suh, Electrochim. Acta, 2012, 75, 123–130.
38. G. Chen, M. Zhou, J. Catanach, T. Liaw, L. Fei, S. Deng and H. Luo, Nano Energy, 2014, 8, 126–132.
39. M. M. Thackeray, J. Am. Ceram. Soc., 1999, 82, 3347–3354.
40. G. Gao, S. Lu, B. Dong, Z. Zhang, Y. Zheng and S. Ding, J. Mater. Chem. A, 2015, 3, 4716–4721.
41. L. Shen, H. Song, H. Cui, X. Wen, X. Wei and C. Wang, CrystEngComm, 2013, 15, 9849–9854.
42. S. Xu, C. M. Hessel, H. Ren, R. Yu, Q. Jin, M. Yang, H. Zhao and D. Wang, Energy Environ. Sci., 2014, 7, 632.
43. J. Morales, L. Sánchez, F. Martín, F. Berry and X. Ren, J. Electrochem. Soc., 2005, 152, A1748.
44. X. Li, L. Qiao, D. Li, X. Wang, W. Xie and D. He, J. Mater. Chem. A, 2013, 1, 6400–6406.
45. P. L. Taberna, S. Mitra, P. Poizot, P. Simon and J. M. Tarascon, Nat. Mater., 2006, 5, 567–573.
46. J.-S. Do and C.-H. Weng, J. Power Sources, 2005, 146, 482–486.
47. Y. Xu, J. Guo and C. Wang, J. Mater. Chem., 2012, 22, 9562–9567.
48. Y.-Y. Hu, Z. Liu, K.-W. Nam, O. J. Borkiewicz, J. Cheng, X. Hua, M. T. Dunstan, X. Yu, K. M. Wiaderek, L.-S. Du, K. W. Chapman, P. J. Chupas, X.-Q. Yang and C. P. Grey, Nat. Mater., 2013, 12, 1130–1136.
49. G. Li, B. Zhang, F. Yu, A. A. Novakova, M. S. Krivenkov, T. Y. Kiseleva, L. Chang, J. Rao, A. O. Polyakov, G. R. Blake, R. A. de Groot and T. T. M. Palstra, Chem. Mater., 2014, 26, 5821–5829.
50. Y. Yu, C.-H. Chen, J.-L. Shui and S. Xie, Angew. Chem., Int. Ed., 2005, 44, 7085–7089.
51. V. Etacheri, R. Marom, R. Elazari, G. Salitra and D. Aurbach, Energy Environ. Sci., 2011, 4, 3243–3262.
52. F. Ma, A. Yuan and J. Xu, ACS Appl. Mater. Interfaces, 2014, 6, 18129–18138.
53. B. G. Choi, M. Yang, W. H. Hong, J. W. Choi and Y. S. Huh, ACS Nano, 2012, 6, 4020–4028.
54. S. Tao, W. Yue, M. Zhong, Z. Chen and Y. Ren, ACS Appl. Mater. Interfaces, 2014, 6, 6332–6339.

---

Footnote

† Electronic supplementary information (ESI) available. See DOI: 10.1039/c5ra04837k

---