Progress in Iron Oxides Based Nanostructures for Applications in Energy Storage

The demand for green and efficient energy storage devices in daily life is constantly rising, which is caused by the global environment and energy problems. Lithium-ion batteries (LIBs), an important kind of energy storage devices, are attracting much attention. Graphite is used as LIBs anode, however, its theoretical capacity is low, so it is necessary to develop LIBs anode with higher capacity. Application strategies and research progresses of novel iron oxides and their composites as LIBs anode in recent years are summarized in this review. Herein we enumerate several typical synthesis methods to obtain a variety of iron oxides based nanostructures, such as gas phase deposition, co-precipitation, electrochemical method, etc. For characterization of the iron oxides based nanostructures, especially the in-situ X-ray diffraction and 57Fe Mössbauer spectroscopy are elaborated. Furthermore, the electrochemical applications of iron oxides based nanostructures and their composites are discussed and summarized. Graphic Abstract

where, M represent Ni, Cu, Fe, Co, etc. During the lithium insertion process, these oxides are reduced by lithium, and the composite consisting of metallic clusters dispersed in a matrix of amorphous Li 2 O is formed [30,31].
Among TMOs, iron oxides based anodes are one kind of excellent candidates with great potential in LIBs since they possess such advantages, e.g., abundance, low cost and nontoxicity [32,33]. However, as similar as other TMOs, iron oxides serving as LIBs anode have two critical issues. One is the large irreversible capacity, caused ( 16:138 by decomposition of electrolyte and formation of solid electrolyte interface (SEI) layer in the 1st discharge process. Additionally, the formation of Fe and Li 2 O is thermodynamically feasible, and the extraction of lithium ion (Li + ) from Li 2 O is thermodynamically instable [34], since a part of Li + cannot be extracted from Li 2 O formed in the 1st discharge process. This also results in partial irreversible capacity. Another issue is their low cycling stability mainly resulted from a large volume variation and severe aggregation of Fe in insertion/deinsertion of Li + , leading to pulverization of electrodes and rapid decay in capacity [3]. For solution to these problems, much effort is focusing on overcoming these issues and some highly effective approaches are proposed. To the best of our knowledge, one highly-effective strategy is nanostructuring of iron oxides [3,35]. For some unique nanostructures, the strain and volume variation resulted from insertion/deinsertion of Li + will be inhibited at large extent, also the Li + can be diffused in electrodes easily, leading to significantly improved electrochemical performance of the anode [11]. Furthermore, carbonaceous materials, e.g., carbon fibers (CFs), carbon nanotubes (CNTs), graphene, and pyrolyzed carbon, etc. are introduced for compositing with iron oxides [36][37][38]. The volume variation of composite electrodes in charge/discharge can be buffered by these carbonaceous materials with unique structures, hence the electronic contact and the cycling stability of iron oxides nanostructures are increased. In this review, the recently developed strategies and important research updates on the iron oxides (Fe 1−x O, Fe 2 O 3 , Fe 3 O 4 ) based nanostructures with applications in LIBs and supercapacitors are elaborated and summarized. Specifically, we concentrated on the iron oxides based nanostructure's synthesis and design, as well as their electrochemical performance.

Wustite (Fe 1−x O)
Wustite (Fe 1−x O) is a non-stoichiometric compound with 0 < x < 0.0464 [11]. It has the rocksalt cubic structure, and its lattice constant is ~ 4.330 Å [11]. Fe 1−x O, compared with Fe 2 O 3 and Fe 3 O 4 , has less applications in energy storage since its relatively low specific capacity and metastable phase below 843.15 K which tends to decompose into Fe and Fe 3 O 4 . However, Fe 1−x O, a highly promising anode for LIBs, has a higher electrical conductivity than those of Fe 2 O 3 and Fe 3 O 4 .

Synthesis and Characterization
As far as we know, the compositing of TMOs with highperformance coverage has great potential for high-performance LIBs anode [7]. In LIBs, the electrochemical mechanism of wustite anode is described as the following equation [8].
FeO/C composites were synthesized using a facile method by Gao et al. [31]. In their synthesis, α-Fe 2 O 3 particles with size of 30-120 nm were mixed with acetylene black (AB) of different percentages, and a uniform mixture was obtained by ball milling. Then, the mixture of α-Fe 2 O 3 and AB was carbonthermally reduced at 800 °C for 10 h in N 2 atmosphere to obtain uniform FeO/C composites. The FeO/C composite possesses much higher cycling stability than that of Fe 2 O 3 /AB mixture. When the content of AB is 50 wt.%, the capacity of FeO/C composite is 511 mA h g −1 , higher than 396 mA h g −1 of Fe 2 O 3 /AB. Besides, the capacity retention after 50 cycles is > 96%, obviously 70-80% higher than that of Fe 2 O 3 /AB. It is believable that the superior electrochemical performance of FeO/C composite should be resulted from its higher electrical conductivity, resulted from strengthened connection of the FeO and AB particles after carbonthermal reduction.
Afterward, in 2016, Jung et al. [11] prepared a potassium (K)-FeO/graphene composite as LIBs anode based on K-doped FeO nanoparticles by thermal diffusion of K into Fe 2 O 3 /graphene using polyol reduction. Rhombohedral Fe 2 O 3 crystals were transformed into FeO crystals (face-centered cubic, FCC), showing a broad d-spacing (5.2 Å) of (111) crystal planes, by calcination of K-doped Fe 2 O 3 /graphene. Comparing with previously studied Fe 2 O 3 /graphene composite [11], the K-FeO/graphene showed a discharge capacity of 1776 mA h g −1 with high cycling stability during 50 cycles at a current density of 100 mA g −1 , whereas Fe 2 O 3 /graphene delivered a discharge capacity of 1569 mA h g −1 . Even at a high current density of 18.56 A g −1 , the capacity of K-FeO/graphene remained at 851 mA h g −1 after 800 cycles. This difference is much larger after the electrodes are cycled longer at a high current density of 18.56 A g −1 . As shown in Fig. 1, compared with the Fe 2 O 3 /graphene, the K-FeO/ graphene anode has unique crystal structure and reaction mechanism. The high discharge capacity of K-FeO/ graphene indicates that specific capacity by storage of additional Li + should be contributed from the vacancies and broad d-spacing within Wustite lattices through potassium diffusion into Fe 2 O 3 lattices.

In-Situ X-ray Diffraction
By in-situ X-ray diffraction (XRD), the information of real-time structural change during the reaction process of sample, and a large amount of comparable information can be obtained in a short period. It can not only observe the structural change of sample during the synthesis process, it also can be used to detect the corresponding structural change of sample at different temperatures (2) FeO + 2Li + + 2e − ↔ Fe + Li 2 O under charge/discharge to a certain potential, which is highly useful for monitoring the actual reaction mechanism of anode/cathode in battery. In addition, our group studied precise knowledge of wustite's lattice constant that is required for the investigation of its physical and chemical properties at high temperature. The Fe 1−x O was synthesized and characterized by high temperature X-ray diffractometer (RINT2000-TTR, Rigaku Denki Co., Ltd.) with parallel beams (Fig. 2a) for measuring the specific diffraction peaks, and the relation between the composition and lattice constant of wustite at high temperature is also investigated. The synthesis process of wustite with the α-Fe (95 wt.%) and Fe 3 O 4 (99 wt%) powders as initial materials is a eutectoid reaction (Fig. 2b). This reaction can be proceeded between 843.15 and 1673.15 K at certain Pco/Pco 2 since the wustite is unstable below 843.15 K, and the reactants are 19% Fe (wt.%) and 81% Fe 3 O 4 (wt.%). Figure 2c shows the XRD pattern of initial reactants (magnetite and iron). Experiment condition is described in Table 1. After purging Helium (He) gas (25 ml min −1 ) into the furnace of XRD system for 60 min, the reactants were heated with a constant rate  240-420 of 10 °C min −1 until the temperature of samples reached 843.15 K, which were also in He atmosphere with the flow rate of 25 ml min −1 . Then, the He gas was exhausted, and CO/CO 2 gas with certain ratio (e.g., 1:1 and 1:2) was purged into the furnace. After the calibration of high temperature by melting Au flakes, the actual temperature of the sample holder could be obtained. From 843.15 K to the desired temperature (1265.28 and 1365.28 K), the heating rate was 2 °C min −1 and the XRD was employed to measure the sample for confirmation of the wustite phase. When the sample was reserved at the desired temperature and most of which was wustite phase, the diffraction angles of wustite's crystal planes were measured during a period of 240-420 min, then the temperature of sample raised for 50 °C and remained at this temperature for 60 min. After these procedures, the temperature of sample decreased to the former desired temperature and the XRD measurement for diffraction angles of crystal planes of wustite was conducted again in a period of 180-240 min.
The lattice constant can be obtained from a linear extrapolating of the apparent lattice constants to zero of this function, that is, 2 Theta = 180°. The diffraction peaks of (111), (200), (220), (311), and (222) crystal planes of wustite are indexed in Fig. 2d, e, and the XRD pattern of the sample is obtained at Pco/Pco 2 of 1:1 and 1:2, respectively. The relation between apparent lattice constant and Nelson-Riley function under different Pco/ Pco 2 and temperatures is obtained, as shown in Fig. 2f. The straight lines represent the squares fitting to the data. By means of these straight lines, the apparent lattice constants were extrapolated to the zero of Nelson-Riley function. Therefore, as shown in Fig. 2f: the results of the true lattice constants obtained at different temperatures and Pco/Pco 2 are 4.355 Å (1265.28 K, Pco/Pco 2 = 1:1), 4.346 Å (1265.28 K, Pco/Pco 2 = 1:2), 4.362 Å (1365.28 K, Pco/Pco 2 = 1:1) and 4.354 Å (1365.28 K, Pco/Pco 2 = 1:2), respectively. As shown in Fig. 2g, the lattice constant increases with the increase of x of Fe 1−x O, and the higher the temperature, the larger the lattice constant. The relation between the composition and lattice constant of wustite at high temperature can be obtained as following Eqs. (3) and (4).

Fe Mössbauer Spectroscopy
The 57 Fe Mössbauer spectroscopy involves properties of the nucleus, including energy level structure of the nucleus and the chemical environment in which the nucleus is located. Hence, it can be applied accordingly to study the valence of atoms, the ionicity of chemical bonds, coordination number, crystal structure, electron density and magnetic properties of sample. The 57 Fe Mössbauer spectroscopy is widely utilized in the fields of chemistry and materials. Herein, we elaborate the 57 Fe Mössbauer spectroscopy for characterizing iron oxides. The 57 Fe Mössbauer spectroscopy is used to distinguish and characterize various iron oxide phases, and to monitor the local environment of Fe atoms in crystal lattice [39,40].
The hyperfine parameters, such as isomer shift (IS), quadrupole splitting (QS), quadrupole shift (ɛ Q ) and hyperfine magnetic field (B hf ) can be obtained by analyzing the position of the spectral lines in Mössbauer spectrum [41,42]. The characteristics of the sample can be inferred from the width and asymmetry of the spectral lines. Through temperature and field dependence of the hyperfine parameters also allows deducing valuable parameters.
Aldon et al. [42] studied lithium-induced conversion reaction of Fe 1−x O using 57 Fe Mössbauer spectroscopy. The hyperfine parameters (IS and QS) are rather characteristic of FeII species in antiferromagnetic (TN = 198 K [43]) Fe 1−x O, showing a typical paramagnetic absorption at room temperature (RT). As indicated in the 57 Fe Mössbauer spectrum (Fig. 3a), there are three broadened doublets with an IS ~ 1 mm s −1 and QS ranging from ~ 0.50 to 1.50 mm s −1 , and their relative areas are corrected from α-Fe magnetic contribution. The absorption intensities are 42, 26 and 15%. A fourth doublet centered at IS ~ 0.55 mm s −1 , QS ~ 0.90 mm s −1 with a relative area of 11% is common for FeII I species, as expected in nonstoichiometric Fe 1−x O. From the FeII/FeII I ratio, the amount of vacancies is estimated as ~ 0.057 ± 0.008, closed to 0.050 by XRD characterization. Finally, the doublet located at QS ~ 1.68 mm s −1 and IS ~ 0 mm s −1 , corresponding to α-Fe, makes a contribution of ~ 6%.
Comparing with Fe 2 O 3 and Fe 3 O 4 , the specific capacity of Fe 1−x O is lowest. Besides, its synthesis method is more complicated. In most cases, the reduction reaction at high temperature is inevitable [3-6, 9, 11]. As a result, Fe 1−x O is not an ideal LIBs anode comparing with Fe 2 O 3 or Fe 3 O 4 .

Fe 2 O 3 Based Nanostructures
Among these iron oxides, especially Fe 2 O 3 , is attracting many researchers' attention due to the high theoretical capacity, which can reach 1000 mA h g −1 [44]. Additionally, Fe 2 O 3 has distinctive advantages, such as high resistance to corrosion, low production cost, environmental friendliness, nonflammability, nontoxicity and high natural availability [45]. Due to these excellent properties, Fe 2 O 3 is highly promising for applications in LIBs anode [46][47][48][49][50]. Herein, a concise overview of recent development about the synthesis, characterization and electrochemical performance of Fe 2 O 3 based nanostructures is provided.

Synthesis and Characterization
In last decade, there are enormous efforts for exploring synthetic methods of Fe 2 O 3 based nanostructures. In this section, we elaborated and summarized synthetic methods of Fe 2 O 3 based nanostructures, including gas phase deposition [51], solution based method [52], electrochemical method [53], thermal treatment [54] and other methods [55,56]. Also, we compared different synthesis methods.
Additionally, the 57 Fe Mössbauer spectroscopy characterization of Fe 2 O 3 nanostructures is described in details. Since only certain nuclei has resonance absorption, the 57 Fe Mössbauer spectroscopy is not interfered by other elements. The range of the 57 Fe Mössbauer spectroscopy affected by extranuclear environment is generally within several nanometers, so it is very suitable for characterizing nanostructure.

Gas Phase Deposition
Gas phase deposition is widely applied in synthesis of many thin films and other nanostructures, such as Fe 2 O 3 and other iron oxides based nanostructures. Chemical vapor deposition (CVD), atomic layer deposition (ALD), physical vapor deposition (PVD), electrolytic deposition and reactive sputtering are typical methods of gas phase deposition [57][58][59][60].
ALD is a unique way to synthesize high-crystallinity thin films, and a cheaper route than liquid phase deposition. In ALD process, the chemical reaction of every layer is directly accompanied with the former layer. In this way, only one layer is deposited per reaction cycle. For instance, Lin et al. [51] utilized ALD to deposit a highquality ultrathin α-Fe 2 O 3 film on TiSi 2 nanonets. The 3D self-organized nanoporous thin films were fabricated by Yang et al. [61] through CVD and integrated into a heterogeneous Fe 2 O 3 /Fe 3 C-graphene. As LIBs anode, it's rate capacity and cyclability can be greatly improved by deposition of this thin film. Cesar et al. [62] deposited thin films of silicon-doped Fe 2 O 3 dendritic nanostructures by atmospheric pressure CVD (APCVD), which produced Fe 2 O 3 photoanodes that oxidize water under visible light with unprecedented efficiency. The dendritic α-Fe 2 O 3 nanostructures showed a macroscopic surface area of 0.5 cm 2 [62]. The vertically aligned α-Fe 2 O 3 nanorods array is grown on a silicon substrate via metal-organic CVD (MOCVD) by Wu et al. [63]. What's more, Jia et al. [64] utilized a radio frequency sputtering deposition to fabricate α-Fe 2 O 3 ultrathin films.
Although gas phase deposition is capable of preparing high-quality Fe 2 O 3 based nanostructures, it also has disadvantages. For example, APCVD and MOCVD have high toxicity and flammability in the process of precursors.

Solution Based Synthetic Method
The solution based synthetic method is common and facile to fabricate Fe 2 O 3 and other iron oxides. Fe 2 O 3 with various morphologies, such as nanoflowers [65], nanospheres [66], nanoparticles [67], nanorods [68], nanotubes [14], nanorings [69], nanobelts [70], nanoflakes [71], nanowires [72], nanofibers [73], and microboxes [54], were synthesized by solution based method, e.g., hydrothermal, solvothermal and sol-gel approaches. These methods are highly facile and available. Zhong et al. [65] used a solvothermal method to fabricate Fe 2 O 3 In the case of (c), magnetic sextet has been slightly shifted as a guide of the eye. Spectrum (d) corresponds to the end of the first charge at 0. 94 Li with no α-Fe contribution, but the blue singlet nanosized metallic ε-Fe 0 is still present. Reprinted with Permission from [42]. Copyright, Elsevier Masson SAS nanoflowers ( Fig. 4) via an ethylene glycol-mediated self-assembly process. Vayssieres et al. [52] reported the growth of porous Fe 2 O 3 nanorods array on fluorine doped tin oxide (FTO) conducting glass by a hydrothermal process. By hydrothermal growth of α-Fe 2 O 3 precursor on SnO 2 nanowire stems, a novel six-fold-symmetry branched α-Fe 2 O 3 /SnO 2 heterostructure (Fig. 5) was synthesized [74]. There is another facile and economical technique, sol-gel method, for synthesizing Fe 2 O 3 nanostructures. Woo et al. [68] utilized a sol-gel method to obtain α-Fe 2 O 3 nanorods by reaction of ubiquitous Fe 3+ in reverse micelles. The nanorods obtained by this mechanism have low dimensionality and high surface area, which can be extended to magnetite and wustite.

Electrochemical Method
Electrochemical method is utilized to synthesize Fe 2 O 3 nanostructures, e.g., the electrochemical deposition is applied in fabricating Fe 2 O 3 nanoparticles [53]. Through anodization of iron foil in ethylene glycol electrolyte solution containing deionized (DI) water and NH 4 F at a voltage of 30-60 V, α-Fe 2 O 3 nanotubes array was obtained [75]. In addition, electrochemical anodization is employed to synthesize Fe 2 O 3 nanotubes array [76]. Mao et al. [77] reported the synthesis of Fe 2 O 3 array using electrochemical deposition. In their research work, iron was deposited into AAO template channels by electrochemical deposition, then the AAO template was removed by NaOH solution, and finally the iron nanorods array was converted to Fe 2 O 3 array. The characteristic of this research is that, by changing the duration of deposition, the length of Fe 2 O 3 nanorods can be finely controlled.

Thermal Treatment
The thermal treatment for synthesizing Fe 2 O 3 involves two significant approaches, thermal oxidation and thermal pyrolysis. For example, Zhang et al. [54] prepared Fe 2 O 3 microboxes (Fig. 6) via thermally induced oxidative decomposition of prussian blue (PB) microcubes at 350-650 °C. The solid-state approach will provide a more facile way for large-scale synthesis of uniform anisotropic hollow structures in comparison with the widely used solution based method. Fe 2 O 3 with different morphologies were prepared via commanding thermal oxidation  [71] utilized heat treatment to prepare densely aligned α-Fe 2 O 3 nanoflakes array.
Thermal pyrolysis is another common method to deposit Fe 2 O 3 thin films. Duret et al. [78] applied this method to obtain mesoscopic α-Fe 2 O 3 leaflet films through ultrasonic spray pyrolysis, while the fabricated films have higher photoactivity than those fabricated by conventional spray pyrolysis techniques. Also, the synthesis of α-Fe 2 O 3 nanoflakes, nanoflowers, nanowires and nanorods array via vapor phase deposition, liquid phase deposition and thermal treatment is reported [63,71,72,79].
There are many ways to synthesize Fe 2 O 3 , but most of them are not environmentally friendly. In 2019, Bashir et al. [56] developed an eco-friendly method to obtain α-Fe 2 O 3 nanoparticles, using Persea Americana seeds extract. They used two different precursors to prepare two samples of α-Fe 2 O 3 , one sample (A) prepared from Fe(NO 3 ) 3 ·9H 2 O, and another sample (B) prepared The 57 Fe Mössbauer spectra of samples A and B recorded at 300 K (room temperature (RT)) are presented in Fig. 7. The 57 Fe Mössbauer spectroscopy is a very valuable technique for exploring the local magnetic behavior and oxidation state of iron atoms in a particular matrix [80]. Both samples revealed magnetic ordering, and displayed only single sextet indicating magnetically ordered state.  [81]. Furthermore, the values of ɛ Q is also consistent with α-Fe 2 O 3 . Both quadrupole interactions indicate Fe as Fe 3+ since the observed IS of 0.3653 mm/s and 0.3754 mm/s for the samples A and B, respectively, are typical for Fe 3+ [82]. Therefore, the negative values of quadrupole splitting indicate the weak ferromagnetic property of the samples A and B, the characteristic of pure α-Fe 2 O 3 phase.

Electrochemical Performance
The charge/discharge cycling at the voltage window of 0.005-3.0 V (vs. Li + ) under a current density of 200 mA g −1 at RT is shown in Fig. 8a. During the initial discharge process, there is an obvious voltage platform of ~ 0.75 V, and it gradually moved to a voltage plateau of ~ 1.0 V, and remain stable in the second and fifth cycles. Meanwhile, an ambiguous plateau was observed at ~ 1.8 V in the charge process. The first discharge profile qualitatively resembles the results by Larcher et al. [   Cyclic voltammogram (CV) is used to characterize the cells with α-Fe 2 O 3 nanoflakes anode in the 0.005-3.0 V under a slow scan rate (at RT). Li metal is used as the counter and reference electrodes [20], consistent with previously reported results [85]. CV curves of α-Fe 2 O 3 hollow spheres between 5 mV and 3 V at a scan rate of 5 mV s −1 are presented in Fig. 8c. There are apparent redox current peaks, and demonstrate good reversibility of electrochemical reaction. As shown in Fig. 8d, a distinct voltage plateau can be discovered at ~ 0.75 V, consistent with CV curves. The charge-discharge voltage profiles reflect the lithium storage capacity of α-Fe 2 O 3 . The first cyclic discharge capacity and the charge capacity is 1219 mA h g −1 and 877 mA h g −1 respectively, which lead to a relatively low irreversible capacity loss of 28%. In the second cycle, the Coulombic efficiency increased quickly to 89%. Cycling performance of two samples are demonstrated in Fig. 8e. The sample (I) exhibits excellent cyclic capacity retention from the second cycle onward. After 100 cycles of charge/discharge, the reversible capacity is still as high as 710 mA h g −1 . Compared with α-Fe 2 O 3 microparticles, the unique hierarchical α-Fe 2 O 3 hollow spheres apparently have enhanced Li storage performance, and a more stable cycling capacity retention and a higher reversible capacity are realized. This superior performance can be attributed to the thin nanosheet subunits that provide rapid and efficient transport of Li + , as well as the unique hollow interior that allows the material to effectively buffer the stress generated during charge/ discharge process.
Iron oxides are cheap, abundant and environmentally compatible, and Fe 2 O 3 has excellent electrochemical performance. Some of the above studies have shown that Fe 2 O 3 based nanostructures can be an alternative anode to replace the presently used graphite in LIBs. The nanostructured Fe 2 O 3 has great potential in LIBs anode [86][87][88][89][90][91][92][93][94][95]. Recently, several studies about Fe 2 O 3 anode in asymmetric supercapacitors are reported [96][97][98][99][100]. However, low surface area and poor electrical conductivity are still two critical issues limiting the specific capacitance and power density of Fe 2 O 3 . For solutions of these problems, CNTs and CNFs are regarded as conductive matrices to load Fe 2 O 3 nanoparticles for realizing improved performance [101]. Table 3 summarizes some typical Fe 2 O 3 based nanostructures with their synthesis and electrochemical performance.

Synthesis and Characterization
Recently, nanostructure engineering is demonstrated as a highly-effective approach to obtain improved electrochemical performance of Fe 3 O 4 and Fe 3 O 4 based composites. Therefore, various nanostructures including 0D nanoparticles [112], 1D nanorods/wires [113,114], 2D nanoflakes/sheets [115,116], 3D hierarchical/porous architectures [117,118], and hybrid nanostructures of iron oxides [16] are proposed. The electrochemical performance of Fe 3 O 4 nanostructures can be optimized by rational design of their morphology, composition, porosity and surface characteristics. Solution phase synthetic method is a facile and rapid way to obtain Fe 3 O 4 based nanostructures, because of the associative advantages, such as low synthesis temperature (always below 250 °C), easy control of morphology via adjusting hydrothermal conditions (e.g., PH, density of reactant and dosage of active agent, etc.). Solution phase synthetic method includes solvothermal synthesis [119], thermolysis [120], co-precipitation [121], sol-gel process [122,123], micro-emulsion [124], etc. Simultaneously we compare the pros and cons of these methods.

Solvothermal Synthesis
Solvothermal synthesis, which reacts in a special closed reaction vessel, is one commonly used methods for synthesizing Fe 3 O 4 . In a hermetic environment, it is a facile method using aqueous solution as reaction medium at high temperature and high-pressure hermetic environment. Fe 3 O 4 nanostructures with various morphologies (0D, 1D, 2D and 3D) were synthesized applying this approach.
An et al. [ [126]. In this preparation, ETA is critical factor for compounding Fe 3 O 4 nanoparticles with high specific surface area, and Fe 3+ is gradually reduced to Fe 2+ by ETA during dissolution process, demonstrating that Fe 2+ increased as the increase of ultrasonication time. The ratio of ETA and FeCl 3 has a large impact on the nanoscale grain size and specific surface area of Fe 3 O 4 . And the results showed that the grain size of 20-40 nm is achieved with 60 mL ETA and 6 mmol FeCl 3 . When the amount of ETA is 80 mL, smaller nanoparticles (5-10 nm) are obtained. Another representative work is reported by Chen et al. [127], in which graphene nanosheets decorated with Fe 3 O 4 nanoparticles (USIO/G) were synthesized using a facile solvothermal process. For the synthesis of USIO composite decorated with reduced graphene oxide (RGO), is used FeCl 3 ·H 2 O as precursor, then NaHCO 3 and L-ascorbic acid were added to form USIO/G. In this process, L-ascorbic acid was oxidized to dehydroascorbic acid (DHAA) by some of Fe 3+ , which were reduced to Fe 2+ . Formation process of USIO/G is schematically shown in Fig. 9. The Fe 3 O 4 nanoparticles with uniform distribution, which are beneficial for electrical conductivity of graphene, mitigation of volume expansion of Fe 3 O 4 , and facilitating Fe 3 O 4 particles into the electrolyte.
Xiong et al. [128] a kind of hierarchical hollow Fe 3 O 4 (H-Fe 3 O 4 ) microspheres prepared by controlled thermal decomposition of iron alkoxide precursor. In a classical reaction, ethylene glycol (EG) serves as reduction reagent that partly reduces Fe 3+ to Fe 2+ with sodium acetate (NaAc), and polyvinylpyrrolidone (PVP) [128]. For this synthesis, PVP served as a surface stabilizer, which has important role in the formation and transformation of hollow interiors.
With the development of solvothermal synthesis, it emerging as an efficient method with the advantages of low energy consumption, little reunion and easy to control shape, etc. Chen et al. [129] synthesized poly (acrylic acid) (PAA)-entangled Fe 3 O 4 nanospheres by a facile solvothermal method. In their synthesis, the ethylenediamine is crucial to the controlling of the uniformity of nanospheres, and the PAA molecules served as carbon source that transforms into the carbon matrix by heating treatment in inert atmosphere. As shown in SEM image of the prepared C-Fe 3 O 4 nanospheres, very uniform spherical particles with a diameter of 150-200 nm are synthesized. Observed from SEM images in Fig. 10a, the nanospheres contain small irregular particles, and have a relatively rough surface. In the control experiment without ethylenediamine (EDA), the synthesized particles are much less uniform with a wider size distribution of 100-500 nm, allowing the formation of nanospheres with smaller size.

Co-precipitation
Due to its high cost-effectiveness, environmental friendliness, and facile synthesizing protocol, co-precipitation is a general approach for Fe 3 O 4 nanoparticles. Thus, in iron based rechargeable battery systems, Fe 3 O 4 nanomaterials are especially suitable for large-scale electrochemical applications to solve the energy requirement of the modern society.
Li et al. [121] proposed Fe 3 O 4 polyhedron as LIBs anode for alkaline secondary batteries by a co-precipitation. Annealing temperature makes a high effect on the physical and electrochemical performance of Fe 3 O 4 nanomaterials. The 700 °C-annealed Fe 3 O 4 exhibited a higher electrochemical performance, such as a higher specific discharge capacity of 604.2 mA h g −1 with a charging efficiency of 83.9% at 120 mA g −1 . Ooi et al. [130] demonstrated octahedral Fe 3 O 4 nanoparticles using a facile solvothermal route. Scanning electron microscope (SEM) image of Fe 3 O 4 nanoparticles is shown in Fig. 10b, which depicts that octahedral Fe 3 O 4 nanoparticles with an average length of 93 ± 18 nm were prepared by the hydrothermal method, showing a roughly Fig. 9 Schematic illustration of the formation process of USIO/G. Reprinted with Permission from [127]. Copyright, Royal Society of Chemistry Reprinted with Permission from [130]. Copyright, Wiley-VCH Gaussian size distribution. Then, the crystal structure of octahedral nanoparticles can be further evaluated by HRTEM, and the composition of the bulk sample was further characterized by XRD and X-ray photo electron spectroscopy (XPS).

Thermolysis
The thermolysis is small monodisperse magnetic nanocrystals synthesized by organic metal compounds in high boiling point solvents containing stabilizing agent. Previous Organic metal bodies include metal acetylacetone compounds, metal cupferron, or metal Carbonyl compounds, and usually choose fatty acids, oleic acid, or hexadecyl amine as a surfactant. Zhang et al. [120] reported ultrafine Fe 3 O 4 nanocrystals uniformly encapsulated in two-dimensional (2D) carbon nanonetworks through thermolysis of Fe(C 5 H 7 O 2 ) 3 precursor at 350 °C under vacuum, which named as 2D Fe 3 O 4 /C nanonetworks. This facile process using low-cost precursor proposed a green approach for preparing Fe 3 O 4 /carbon composite. Additionally, compared with the reported Fe 3 O 4 /carbon composites, the particle size of Fe 3 O 4 is controllable and a size of ∼ 3 nm can be obtained.
Benefitting from synergistic effects of carbon nanonetworks with excellent electrical conductivity and ultrafine Fe 3 O 4 particles with uniform distribution, high reversible capacity, excellent rate capability and superior cyclability at the voltage of 0.01-3.0 V (vs. Li/Li + ) are obtained. Nanoparticles with unique iron oxide (Fe 3 O 4 ) cores and zinc oxide (ZnO) shells were prepared by Jaramillo et al. [131]. Fe 3 O 4 nanoparticle synthesized through a thermolysis method using Fe(C 5 H 7 O 2 ) 3 as organic metal body presoma, triethylene glycol as surface active agent, and core-shell Fe 3 O 4 @ZnO nanoparticles were successfully synthesized using straightforward methodologies. The structural and optical properties of the materials were characterized using a combination of X-ray diffraction, electron microscopy, and light spectroscopy. Importantly, the purity of the core and shell phases in the Fe 3 O 4 @ZnO nanoparticles was confirmed by both XRD and TEM, and the ZnO shell was shown to increase the transparency of the core-shell nanoparticles relative to the single-component Fe 3 O 4 nanoparticles. Zhang et al. [132] demonstrated a high crystalline Fe 3 O 4 -graphene composite by one-step reaction of thermolysis. And they demonstrated that the attachment of iron-organic complex with graphene oxide (GO) sheets can facilely result in magnetic graphene composites via a time-dependent calcination process.

Sol-Gel Process
The specific method is using the metal alkoxide, metal mineral compound or a mixture of the above two substances to hydrolysis and polymerization, uniform gel gradually, then condense into a transparent gel, however, after drying and heating, finally the oxide ultrafine powders was received. Tang et al. [122] prepared nanostructured magnetite thin film by sol-gel method using inexpensive iron (II) chloride precursor. Fe 3 O 4 nanoparticles were prepared at 300 °C, however, α-Fe 2 O 3 is generated when temperature increased to 350 °C, and this result restricts its applications. Xu et al. [123] proposed magnetite nanoparticles by virtue of sol-gel process combined with annealing in vacuum at 200-400 °C using nontoxic and low-cost ferric nitrate. In their study, Fe 3 O 4 nanoparticles with various sizes can be synthesized facilely through varying the annealing temperature.

Micro-emulsion Method
Micro-emulsion is composed of two mutual miscibility of liquid mixture of thermodynamic stability and isotropy dispersion, one of these or two kinds of liquid called micro area, and fixed by interface layer of the surfactant molecules. The key factors controlling the reaction solution contain concentration, pH value, reaction time and temperature. Micro-emulsion as a rapid expansion of new technology possesses many advantages. For example, high purity and uniform particle size distribution molecular dopant was synthesized at low temperature and simple reaction process. But there are also some shortcomings, for instance, the reaction system mostly contains organic solvents, which leads to high cost, pollution of environmental health and long reaction time. The prepared Fe 3 O 4 nanoparticles have excellent catalytic performance for the synthesis of quinoxaline in different solvents. Novel core-shell magnetic Fe 3 O 4 / silica nanocomposites with triblock-copolymer grafted on their surface (Fe 3 O 4 @SiO 2 @MDN) were successfully synthesized by combining sol-gel process with seeded aqueous-phase radical copolymerization approach [133]. The Fe 3 O 4 @SiO 2 @MDN microspheres were synthesized in following three steps. Firstly, the initial magnetic Fe 3 O 4 microspheres were synthesized by a solvothermal reaction. Then a sol-gel process was utilized to prepare silica coated Fe 3 O 4 microspheres (Fe 3 O 4 @SiO 2 ), and a thin amorphous silica layer was formed on Fe 3 O 4 microspheres. Afterward, the Fe 3 O 4 @SiO 2 microspheres were modified by 3-(methacryloxypropyl) trimethoxysilane (MPS). Finally, the triblock copolymer was fabricated by aqueous phase radical copolymerization reaction among MPS, divinylbenzene (DVB) and N-Vinyl-2-pyrrolidone (NVP) on the surface of Fe 3 O 4 @SiO 2 . The magnetic Fe 3 O 4 particles with narrow size distribution have nearly spherical shape and smooth surface. Li et al. [124] reported hexagonal and triangular monodisperse Fe 3 O 4 nanosheets by a two-step microemulsion solvothermal approach, in which the uniform Fe 3 O 4 nanoparticles are prepared and then these hydrophobic nanocrystals are dispersed in a uniform microemulsion environment as "seeds" for further re-growth through a secondary solvothermal process. In the first step, nearspherical monodisperse 7-8 nm Fe 3 O 4 nanoparticles were formed through a kinetically controlled process. In the second step, the formation of anisotropic Fe 3 O 4 nanosheets is a thermodynamically controlled process and all the exposed surfaces of the triangular and hexagonal nanosheets are (111) crystal planes, which have the lowest surface energy for FCC Fe 3 O 4 .

Other Methods
Physical methods are also significant ways to prepared Fe 3 O 4 nanostructure for anode of LIBs. Several advantages, such as good crystallization, fine-tuned particle size, and high purity of products are highlighted in recent literatures. But these methods usually demand advanced and expensive equipment, result in a higher cost, poor dispensability of particles dispersion, and agglomeration of nanostructures. For instance, Du et al. [109] fabricated activated carbon (AC)-Fe 3 O 4 nanoparticles asymmetric supercapacitor, and Fe 3 O 4 nanostructure was prepared by microwave method. The precursor, FeSO 4 ·7H 2 O and NH 3 ·H 2 O mixed solution, was heated in microwave oven. The black precipitate was separated by magnet and washed repeatedly with DI water. The resulted microstructural properties of prepared nanoparticle were characterized by nitrogen adsorption (Quantachrome NOVA 2000), XRD and SEM [109]. Chen et al. [127] synthesized graphene nanosheets decorated with ultra-small Fe 3 O 4 nanoparticles (USIO/G). Seo et al. [134] reported an integrated usage of magnetic particles in microalgal downstream processes, specifically microalgal harvesting and lipid extraction through one-step aerosol spray pyrolysis and applied in microalgal harvesting and serial microalgal lipid entrapment. TEM/EDS, XPS, and FT-IR analysis suggested that the cationic and lipophilic functionalities arose from not fully decomposed PVP, due to the short residence time in the reactor. Kang et al. [135] proposed Fe 3 O 4 nanocrystals confined in mesocellular carbon foam (MSU-F-C) by a "host-guest" approach and applied it as LIBs anode. In this study, a precursor of Fe(NO 3 ) 3 ·9H 2 O is impregnated in MSU-F-C having uniform cellular pores with a diamter of ~ 30 nm, followed by heating treatment at 400 °C for 4 h in argon (Ar) atmosphere. Fe 3 O 4 nanocrystals with size of 13-27 nm were fabricated inside the pores of MSU-F-C. The existance of the carbon most likely allows the reduction of some Fe 3+ to Fe 2+ ions by a carbothermoreduction process. The physical performance and pore structure of MSU-F-C and Fe 3 O 4 -loaded composites were characterized with nitrogen sorption, and the composites have high capacities of ∼800-1000 mA h g −1 at 0.1 A g −1 (∼0.1 C rate), high rate capability and good cycling performance.

Li-Ion Batteries
Due to conversion reaction of Fe 3 O 4 during charge/discharge process and other advantages, the Fe 3 O 4 is usually studied and applied as LIBs anode [143][144][145][146][147]. For TMOs, they have higher theoretical capacity (~ 500-1000 mA h g −1 ) than conventional graphite (about 372 mA h g −1 ). Furthermore, Fe 3 O 4 has superior conductivity compared with other transition metal oxides. Thus, it is well-focused by recent studies. It has been reported that composite electrodes with graphene have high performance due to their large surface area, high electrical conductivity and adaptive or flexible structure for high reliability. Qiu et al. [3] reported a kind of composite anode composed of ultra-dispersed Fe 3 O 4 nanoparticles (3-8 nm) and RGO sheet. It has excellent cyclic performance (624 mA h g −1 for up to 50 charge/discharge cycles at a current density of 0.1 A g −1 ), and good specific capability (624 and 415 mAh g −1 at 0.1 and 2.4 A g −1 , respectively) for LIBs. The obtained Fe 3 O 4 /RGO exhibited high and ultrastable Photo-Fenton activity (Fig. 11).
Pyrolyzed carbon is also a good "companion" for Fe 3 O 4 anodes. Apart from the facile protocol, porous Rapid aerosol Capacitance of 1153 F g −1 at 2 A g −1 ; retention of 96.7% after 8000 cycles [4] structure formed by pyrolysis always exhibits high specific capacity of Fe 3 O 4 composite anode. Wang et al. [12] reported hollow N-doped Fe 3 O 4 /C nanocages with hierarchical porosities by carbonizing polydopamine-coated PB nanocrystals as LIBs anode (Fig. 12). The specific capacity of N-doped Fe 3 O 4 /C nanocages is ~ 878.7 mA h g −1 after 200 cycles at a current density of 200 mA g −1 , much higher than that of N-doped Fe 3 O 4 /C derived from pure PB (merely 547 mA h g −1 ). It is also desirable to design anisotropic structure of Fe 3 O 4 nanoparticles with carbon coated layer. Zhang et al. [19] reported a kind of carbon-coated Fe 3 O 4 nanospindles derived from α-Fe 2 O 3 nanospindles with length of about 500 nm and an axis ratio of ~ 4. Following by a hydrothermal synthesis method with glucose, the obtained LIBs anode delivered a high reversible capacity of ~ 745 mA h g −1 at C/5 and ~ 600 mA h g −1 at C/2. The most impressive work towards this field is probably the mesoporous iron oxide nanoparticle clusters with carbon coating reported by Lee et al. [148]. After a few cycles, the formation of SEI greatly enhanced the stability of interface between electrode and electrolyte. Electrochemical test exhibited a high specific capacity of 970 mA h g −1 for LIBs.

Supercapacitors
Fe 3 O 4 is a highly promising candidate for supercapacitor electrode because of its relatively high electrical conductivity, fast reversible redox reaction, low cost and eco-friendly nature [149][150][151][152]. Similar to batteries, high performance supercapacitors also require two factors: large specific surface area and long-term stability. Those two features usually were achieved by building some porous structures and carbon coated layers. Fe 3 O 4 nanoparticle with a high specific surface area was synthesized by Wang et al. [126] using a bottom up approach. Ferric chloride was firstly sonicated with ethanolamine and then processed through a solvothermal reaction. The obtained active nanomaterials showed a specific surface area of 165.05 m 2 g −1 and a specific capacitance of 207.7 F g −1 at 0.4 A g −1 .
Also, highly dispersed Fe 3 O 4 nanosheets on 1D CNFs is reported by Mu et al. [125]. The Fe 3 O 4 /CNFs composites showed a higher specific capacitance than pure Fe 3 O 4 in 1 M Na 2 SO 3 . To further enlarge the specific capacitance and cycle stability, hierarchically porous carbon spheres with Fe 3 O 4 using as supercapacitors exhibited high capacitivity of 1153 F g −1 at 2 A g −1 and high specific capacitance of 514 F g −1 at 100 A g −1 . In addition, the assembled asymmetric supercapacitor with double-shelled hollow carbon spheres and Fe 3 O 4 , has excellent cycling stability (96.7% retention after 8000 cycles) and high energy density (17-45 Wh kg −1 ) at a power density of 400-8000 W kg −1 [4].

Conclusion
Iron oxides (Fe 1−x O, Fe 2 O 3 , Fe 3 O 4 ) based nanostructures have much higher specific capacities than those of commercial carbon based anodes. They are considered as highly promising candidates for LIBs anode. However, large irreversible capacity and low cycle stability are two serious problems that obstruct the application of iron oxides based nanostructures. In this review, we summarized the recent progress on novel iron oxides and their composites as LIBs anode and supercapacitor electrode. Several typical synthetic methods of various novel iron oxides based nanostructures are listed. By comparing the electrochemical performance of these various iron oxides based nanostructures, some strategies are expected to solve the problems of iron oxides based nanostructures.