Introduction

Rechargeable lithium ion batteries (LIBs) have been regarded as the most popular electrochemical energy storage devices used in hybrid electrical vehicles, electrical vehicles, and handheld electronic equipment [1,2,3]. Lots of transition metal oxides, such as SnO2 [4, 5], Co3O4 [6, 7], MnO2 [8, 9], NiO [10, 11], and Fe3O4 [12, 13], have been used as anodes in LIBs to date. Iron oxides, as an alternative of graphite, have been recognized as promising candidates due to their high specific capacities (above 1000 mAh/g), abundant resources, environmental benignity, and low cost. However, the serious volume expansion of magnetic iron oxides during the lithiation–delithiation processes is a fatal defect for application in LIBs. A series of associated problems such as the cracking of active materials, the collapse of electrode structure, and electrode polarization [14,15,16,17,18] result in an extremely bad cycle stability and rate performance. Resolving such problems would thus be beneficial.

Nowadays, coating a layer of stable electrochemical materials is considered to be effective for stabilizing the electrochemical properties of iron oxides in the lithium ions insertion–extraction processes, such as Fe3O4@porous carbon [14], Fe3O4/carbon composites [15], mesoporous Fe3O4@C submicrospheres [16], and porous Fe3O4@C hollow spheres [19]. Zhang et al. [20] enhanced the cycling performance by coating a layer of TiO2 to form Fe2O3@TiO2 core–shell (C–S) nanorods, and the results showed high and stable delivered capacities of approximately 860 mAh/g at a high current density of 1000 mA/g. Porous TiO2 was chosen as the shell, which benefits electrolyte infiltration, restricts excessive volume changes, and prevents collapse and dissolution into the electrolyte of Fe2O3 during the discharge–charge processes [20]. The design of yolk–shell (Y–S) structure has also proved to be effective for improving the cycling stability of iron oxides [21]. For example, to increase the conductivity of materials, a unique architecture with Y–S Fe2O3⊙C composites was synthesized by the classic Stöber method, which can prevent the reunion of active materials and support an interior space to alleviate the volume expansion and rupture during discharge–charge experiments. The materials could be found to possess a good cycling stability and provide an excellent rate performance [21]. To relieve the capacity fading of bare metal oxides, other metal oxides with Y–S structures, such as Cr2O3@TiO2 Y–S octahedrons [22], SnO2@C Y–S nanospheres [23], MnOx/C Y–S nanorods [24], and Co3O4@TiO2@C Y–S spheres [25], have also been synthesized using a variety of methods to achieve the good electrochemical properties.

In this work, we reported a simple method to synthesize Fe3O4@TiO2@C Y–S spheres for alleviating the capacity attenuation of bare Fe3O4 spheres as anodes in LIBs. The materials were physically characterized to investigate the structure, synthesis process, crystal structure, components, and element valence. When tested as anodes of LIBs, the Fe3O4@TiO2@C Y–S spheres show the excellent cycling stability and good rate performance. At 200 mA/g, the reversible capacity of approximately 450 mAh/g is much higher than that of bare Fe3O4 spheres (160.2 mAh/g) after 100 cycles. Therefore, constructing Y–S structures is an acceptable strategy to promote the electrochemical performance of transition metal oxides when applied in LIBs.

Experimental

Synthesis of Fe3O4@TiO2@C Y–S Spheres

Fe3O4 spheres were synthesized using a method reported in Ref. [12]. First, 0.64 g of sodium dodecyl benzene sulfonate was added to 60 mL of ethylene glycol to form a clear solution, and 2.16 g of FeCl3·6H2O and 3.72 g of sodium acetate were then mixed in the above solution under vigorous stirring for several hours. The obtained solution was transferred into a 100-mL Teflon-lined autoclave, and the oven temperature was increased to 190 °C for 15 h. Finally, a brown product was collected by centrifugation and washed with ethanol/water five times.

The synthesis method of Fe3O4@SiO2@TiO2 C–S spheres was similar to that reported in Ref. [26]. The as-synthesized Fe3O4 spheres were dispersed in a mixed solution of 150 mL ethanol/water (V/V = 4:1) using ultrasonication and vigorous agitation. 1 mL of ammonia and 1 mL of tetraethylorthosilane (TEOS) were successively and slowly added to the above solution to form Fe3O4@SiO2 C–S spheres. 0.1 g of Fe3O4@SiO2 C–S spheres was then dispersed in an ethanol (120 mL)/water (0.5 mL) solution containing hydroxypropyl cellulose (HPC) by repeated ultrasonication and continuous vigorous stirring for 10 h. Finally, tetra-n-butyl titanate (TBOT, 0.6 mL) ethanol solution was slowly added into the above mixed solution to form Fe3O4@SiO2@TiO2 C–S spheres.

Through hydrothermal carbonization treatment, a layer of carbon was coated to improve the conductivity of the as-prepared samples. First, the Fe3O4@SiO2@TiO2 C–S spheres were dispersed in 60 mL of glucose (300 mg) aqueous solution under ultrasonication. Second, the mixture was transferred into a 100-mL Teflon-lined autoclave and then heated at 180 °C for 6 h in an oven to obtain the Fe3O4@SiO2@TiO2@C C–S spheres. Third, the sample was subsequently annealed at 500 °C for 4 h to further carbonize in an argon atmosphere at a heating rate of 2 °C/min. Finally, the sacrificial SiO2 layer was removed by immersing in 60 mL of LiOH aqueous solution (0.06 mol/L) to obtain the final product (Fe3O4@TiO2@C Y–S spheres) at 90 °C.

Material Characterization

Scanning electron microscopy (SEM, S4800, HITACHI, Japan) and transmission electron microscopy (TEM, Tecnai G2 F20, FEI Company, USA) were employed to observe the morphologies and sizes of samples. Energy-dispersive X-ray spectrometer (EDS, HITACHI, Japan) was used to further investigate the elements and their distribution in the materials. X-ray powder diffractometer (XRD, Bruker AXS, D8-Focus) was employed to determine the crystal structure of samples by using Cu-Kα irradiation in a range of 10° to 80°. The Raman spectra (DXR Microscope, Thermo Electron Corporation, USA) were collected using Raman microscope to characterize the carbon atom and crystal structure of Fe3O4@TiO2@C Y–S spheres.

Electrochemical Measurements

CR2032 coin-type cells were assembled in an Ar-filled glove box to measure the electrochemical properties of samples. The working electrodes were composed of active materials, acetylene black, and polyvinylidene fluoride at a mass ratio of 7:2:1. The loading mass weight density of active materials is approximately 1.0–1.1 mg/cm2. 1.0 mol/L LiPF6 in a mixed solvent of ethylene carbonate and diethyl carbonate (DEC) (V/V = 1:1) served as the electrolyte, and circular metal Li foils were used as counter electrodes in the CR2032 coin-type cells. The discharge–charge measurements of cells were conducted on a Land battery test system in a potential range of 3.0 V to 0.01 V versus Li/Li+. Cyclic voltammetry (CV) tests were conducted on a CHI660D electrochemical workstation at different scan rates of 0.10 mV/s and 0.50 mV/s between 3.0 and 0.01 V.

Results and Discussion

Figure 1a shows a schematic diagram of the stepwise synthesis of Fe3O4@TiO2@C Y–S spheres, as described in the experimental section. First, Fe3O4 spheres with various sizes were synthesized using a simple hydrothermal method. SEM image shows that the diameters of large Fe3O4 spheres are about 200–300 nm, while those of smaller Fe3O4 spheres are only dozens of nanometers (Fig. 1b). Second, a thin SiO2 layer was coated on the surface of Fe3O4 spheres to form Fe3O4@SiO2 C–S spheres with a very smooth surface by the hydrolysis of TEOS in an alkaline aqueous solution (Fig. 1c). Third, the outer TiO2 shell was coated on the surface of the SiO2 layer to obtain Fe3O4@SiO2@TiO2 C–S spheres by the hydrolysis of TBOT. SEM image shows that the obtained material became much coarser (Fig. 1d). Fourth, the Fe3O4@SiO2@TiO2@C C–S spheres were formed by the hydrolysis of glucose, followed by calcination at 500 °C. Finally, the middle SiO2 layer as sacrificial template was removed by immersing in a LiOH aqueous solution and the Fe3O4@TiO2@C Y–S spheres were obtained. Figure 1e shows that the final product has an uneven size distribution ranging from dozens of nanometers to 800 nm. It is also of note that many spheres (Fe3O4@SiO2, Fe3O4@SiO2@TiO2, and Fe3O4@TiO2@C spheres) were reunited after coating SiO2 and TiO2 layers.

Fig. 1
figure 1

a Schematic diagram showing stepwise synthesis of Fe3O4@TiO2@C Y–S spheres. SEM images of b Fe3O4 spheres, c Fe3O4@SiO2 C–S spheres, d Fe3O4@SiO2@TiO2 C–S spheres, and e Fe3O4@TiO2@C Y–S spheres

Full size image

The sizes, morphologies, and internal structures of samples were characterized using TEM. As evident from the TEM images shown in Fig. 2a, the Fe3O4 spheres have a size distribution ranging from several nanometers to 300 nm. After hydrolysis of TEOS, the SiO2 shell with a thickness of 20–30 nm was coated on the surface of Fe3O4 spheres to form Fe3O4@SiO2 C–S spheres. Meanwhile, a lot of small SiO2 spheres were synthesized in this process (Fig. 2b). After coating the TiO2@C shell and then removing the SiO2 layer, it can be found from TEM images (Fig. 2c, d) that Fe3O4@TiO2@C Y–S spheres have a very unambiguous internal void space, and the width between the Fe3O4 sphere and TiO2 shell is about 20 nm. Notably, a large number of smaller spheres with sizes ranging from dozens to hundreds of nanometers exist in the final product. It can be attributed to three reasons: (1) Small Fe3O4@TiO2@C Y–S spheres and TiO2@C hollow spheres were synthesized using small Fe3O4@SiO2 C–S spheres and SiO2 spheres as templates, respectively; (2) the aggregation of TiO2 nanoparticles occurred in the hydrolysis process of TBOT; and (3) the collapse of outer layer TiO2 shells resulted in the formation of smaller TiO2 nanocrystals in the alkaline thermal etching process.

Fig. 2
figure 2

TEM images of a Fe3O4 spheres, b Fe3O4@SiO2 C–S spheres, and c Fe3O4@TiO2@C Y–S spheres. d Local magnification TEM image of Fe3O4@TiO2@C Y–S spheres

Full size image

EDS analysis and elemental mapping were applied to determine the components, elements, and their distribution in Fe3O4@TiO2@C Y–S spheres. EDS analysis clearly shows that the composites contain the elements of titanium, oxide, iron, and silicon (Fig. 3a), and the silicon content is very low. Calculation of the elemental contents (Table 1) shows that the mass ratio of Fe3O4 to TiO2 is 2.33:10. The high angle annular dark field (HAADF) image shows the existence of an unambiguous void space/boundary layer between Fe3O4 sphere and TiO2@C shell. Figure 3b shows that oxygen element is distributed in the entire sphere, while iron element is only distributed in the center of the material, which further proves that the Fe3O4 sphere is located inside Fe3O4@TiO2@C Y–S sphere. In contrast, silicon element is discretely dispersed in the Fe3O4@TiO2@C Y–S sphere, and this behavior is obviously different from that of other elements and is mainly ascribed to the absorption of only a small amount of SiO2 on the surface of Fe3O4@TiO2@C Y–S spheres during the alkaline thermal etching process.

Fig. 3
figure 3

a EDS analysis image of Fe3O4@TiO2@C Y–S spheres; b HAADF image and EDS elemental mapping images of oxygen, titanium, iron, and silicon in the Fe3O4@TiO2@C Y–S spheres

Full size image
Table 1 EDS analysis result of Fe3O4@TiO2@C Y–S spheres
Full size table

The X-ray powder diffraction (XRD) patterns of Fe3O4 spheres and Fe3O4@TiO2@C Y–S spheres are shown in Fig. 4a, and curve (i) shows that the diffraction peaks of 2θ at 18.5°, 30.3°, 35.6°, 43.3°, 53.8°, 57.3°, 62.8°, 71.4°, and 74.4° are indexed to the planes (111), (220), (311), (400), (422), (511), (440), (620), and (533) of magnetic cubic phase Fe3O4 crystals (JCPDS No. 19-0629), respectively. When the Fe3O4 spheres were parceled by TiO2@C to form Fe3O4@TiO2@C Y–S spheres, the diffraction peaks (marked by the symbol ·) of Fe3O4 crystals become much weaker, as shown in curve (ii) of Fig. 4a. The 2θ at 25.36°, 37.96°, 48.16°, 55.02°, 62.26°, and 68.8° corresponds to the planes (110), (004), (200), (211), (213) and (116) of anatase TiO2 crystals. The higher crystallinity of Fe3O4@TiO2@C Y–S spheres can be ascribed to the lower carbon content and the lower alkali thermal temperature, and this is evidently different from that of Co3O4@TiO2@C Y–S spheres in Ref. [25]. The Raman spectrum of Fe3O4@TiO2@C Y–S spheres is shown in Fig. 4b. The peaks at 198.4 cm−1 and 614.9 cm−1 are derived from the anatase TiO2 crystals [27], and the peak at 415.2 cm−1 is ascribed to iron oxides [28]. Furthermore, the elemental carbon of Fe3O4@TiO2@C Y–S spheres is confirmed from the D-band (1320 cm−1) and G-band (1620 cm−1) [29, 30].

Fig. 4
figure 4

a XRD patterns of Fe3O4 spheres and Fe3O4@TiO2@C Y–S spheres; b Raman spectrum of Fe3O4@TiO2@C Y–S spheres

Full size image

The galvanostatic cycling performances and coulombic efficiencies curves of Fe3O4@TiO2@C Y–S spheres, Fe3O4 spheres, and TiO2@C hollow spheres at 200 mA/g are presented in Fig. 5. The first discharge capacity of Fe3O4@TiO2@C Y–S spheres is 778.1 mAh/g with a very low coulombic efficiency of 48.2%. Such a large irreversible capacity loss is mainly derived from the irreversible decomposition of electrolyte, the solid electrolyte interface (SEI) films formed in the electrode, and the trapping of lithium ions inside TiO2 and Fe3O4 lattices [12, 14,15,16,17, 19, 20]. The second discharge capacity is reduced to only 523.6 mAh/g, and the subsequently delivered capacities show a slightly declining trend. To the 18th cycle, the discharge capacity reaches the lowest value of 402.8 mAh/g and then rises slowly to around 450 mAh/g over 100 cycles. This occurs because electrolyte infiltration gradually increases in the porous materials to improve the electron transfer and Li-ion migration rate with the number of cycles [31]. The coulombic efficiencies are nearly 100% after five cycles, which indicates that a layer of stable SEI films has formed on the surface of the electrode, and the Fe3O4@TiO2@C Y–S spheres have a good electrochemical Li-inserted reversibility. In contrast, the bare Fe3O4 spheres have poor electrochemical properties (Fig. 5a). The first discharge capacity reaches 1082 mAh/g with a high coulombic efficiency of 70.7% (Fig. 5b), but the delivered capacities of the bare Fe3O4 spheres decay very quickly. After 10 cycles and 100 cycles, the discharge capacities drop to 307.1 mAh/g and 160.2 mAh/g, respectively, which are far lower than those of the composites under the same test conditions. As such, the high and stable reversible capacities of Fe3O4@TiO2@C Y–S spheres are attributed to two aspects: First, the internal void space formed by the removal of SiO2 effectively alleviates the volumetric variation in Fe3O4 spheres, and second, the TiO2@C shell effectively restrains the pulverized inner Fe3O4 spheres to relieve the electrode polarization and capacity fading, thus leading to an obvious improvement in cycling stability [14, 16, 19,20,21]. The first discharge capacity of TiO2@C hollow spheres is 646.5 mAh/g with a very low coulombic efficiency of 30.1%, but the second discharge capacity is only 343.0 mAh/g (Fig. 5). After 100 cycles, the discharge capacity is maintained at 286.5 mAh/g. The delivered capacities of TiO2@C hollow spheres are obviously much lower than those of Fe3O4@TiO2@C Y–S spheres over 100 cycles. These obtained data clearly show that the capacities of Fe3O4@TiO2@C Y–S spheres are largely related to those of the inner Fe3O4 spheres. In addition, the first coulombic efficiency of Fe3O4@TiO2@C Y–S spheres (48.2%) is lower than that of Fe3O4 spheres (70.7%) and higher than that of TiO2@C hollow spheres (30.1%), implying that the irreversible capacity of Fe3O4@TiO2@C Y–S spheres is largely derived from the outer TiO2@C layer in the first cycle.

Fig. 5
figure 5

a Galvanostatic cycling tests and b coulombic efficiencies of Fe3O4@TiO2@C Y–S spheres, Fe3O4 spheres, and TiO2@C hollow spheres at a current density of 200 mA/g

Full size image

Figure 6 illustrates the galvanostatic discharge–charge voltage curves for the 1st, 2nd, 10th, 30th, 50th, and 100th cycles of Fe3O4@TiO2@C Y–S spheres, Fe3O4 spheres, and TiO2@C hollow spheres at 200 mA/g in a voltage range from 3.0 to 0.01 V. As shown in Fig. 6a, the voltage plateaus of Fe3O4@TiO2@C Y–S spheres at about 1.75 V and 2.0 V are attributed to the insertion–extraction of  lithium ions in the anatase TiO2 layer with good crystallinity [20], while the voltage plateaus at about 0.75 V and 1.5 V correspond to the redox reactions between Fe3O4 and Fe [12, 14,15,16,17, 19]. Notably, the formation of SEI films occurs in a voltage range below 0.75 V [19, 20].

Fig. 6
figure 6

Discharge–charge voltage profiles of a Fe3O4@TiO2@C Y–S spheres, b Fe3O4 spheres, and c TiO2@C hollow spheres at different cycles

Full size image

The galvanostatic discharge–charge voltage profiles of pure Fe3O4 spheres are shown in Fig. 6b for comparison. It is evident that the same voltage plateaus at approximately 0.75 V and 1.5 V correspond to the redox reactions between Fe3O4 and Fe. Besides, the capacity fading is fast in the first five cycles, and then, it slows from the 6th cycle to the 10th cycle. The gradual increase in the space between the discharge–charge voltage plateaus suggests that the cell polarization caused by the high resistance, low ionic transport, and low electronic conductivity increases sharply with an increase in cycles during the first 10 cycles. Such poor electrochemical properties are mainly ascribed to the pulverization of active material and the crack of electrode caused by the severe volume expansion of bare Fe3O4 spheres. The pulverized Fe3O4 can lead to a loose connection between the particles and the current collector [20]. Then, the crack of electrode results in the rupture of SEI films, which forces the new electrode surface to be exposed to the electrolyte, and unstable SEI layers are then repeatedly formed and more electrolyte is consumed [12, 16]. Following this, the capacities and voltage plateaus tend to become stable.

The discharge–charge voltage profiles of TiO2@C hollow spheres are shown in Fig. 6c, which indicates that the plateaus at about 1.75 V and 2.0 V are ascribed to the redox process of anatase phase TiO2 crystals [20], and no other voltage plateaus are observed. In summary, the stable capacities and voltage plateaus indicate that the TiO2@C layer has good electrochemical properties for application in LIBs. Therefore, coating the TiO2@C layer is a good way for improving the cycling stability of the inner Fe3O4 spheres.

Figure 7a presents the rate capability of Fe3O4@TiO2@C Y–S spheres at various current densities for 20 cycles. At a low current density of 50 mA/g, the first discharge capacity of Fe3O4@TiO2@C Y–S spheres is 950 mAh/g, and then, the capacities are stabilized at about 500 mAh/g in subsequent cycles. In addition, when the constant current densities are switched to 100 mA/g, 200 mA/g, 400 mA/g, 800 mA/g, and 1600 mA/g, the discharge capacities of Fe3O4@TiO2@C Y–S spheres remain at about 460 mAh/g, 410 mAh/g, 320 mAh/g, 150 mAh/g, and 62 mAh/g, respectively. When the current density returns to 50 mA/g, the reversible capacities of Fe3O4@TiO2@C Y–S spheres still reach more than 500 mAh/g, indicating the excellent cycle stability and good reversibility.

Fig. 7
figure 7

a Rate performance and b discharge–charge voltage profiles of Fe3O4@TiO2@C Y–S spheres at various current densities

Full size image

Figure 7b shows the discharge–charge voltage profiles of Fe3O4@TiO2@C Y–S spheres at various current densities. It is evident that at low current densities (50–400 mA/g), the fading of electrochemical reversible capacities, and the space of discharge–charge voltage plateaus increase slowly with the increasing current density. However, at high current densities (400–1600 mA/g), the fading of electrochemical reversible capacities, and the space of discharge–charge voltage plateaus increase sharply with the increasing current density. This can be attributed to three factors: (1) The middle layer TiO2 is a kind of semiconductor material that affects the charge transfer and lithium ions diffusion in the interior Fe3O4 spheres; (2) the lower carbon content results in the low electron transfer and poor lithium ions diffusion; and (3) the accumulation or depletion of  lithium ions on the electrode causes the concentration polarization of batteries due to slow diffusion at high current densities.

CV curves of bare Fe3O4 spheres and Fe3O4@TiO2@C Y–S spheres were measured at a lower scan rate of 0.10 mV/s in a potential range from 3.0 to 0.01 V (Fig. 8a, b). From Fig. 8a, in the first cycle, an obvious larger reduction current in the region of below 0.75 V should be mainly ascribed to the formation of SEI films and the reduction process of Fe3O4, while the oxidation peak at 1.68 V can be attributed to the oxidation process from Fe to Fe3O4 [12, 16]. The related electrochemical reaction process is presented in Eqs. (1)–(3). Subsequently, the reduction peaks firstly shift to about 1.3 V and then shift to the lower potential, while the oxidation peaks constantly shift to the higher potential. The gradually decreased peak area and the increased peak space indicate the continuously increased rupture and volume expansion of Fe3O4 spheres, and the electrode polarization,  resulting in a poor cycling stability and low reversible capacity [10,11,12,13,14]. This is well consistent with the galvanostatic discharge–charge tests of pure Fe3O4 spheres. From Fig. 8b, the electrochemical performance of Fe3O4@TiO2@C Y–S spheres obviously exceeds that of bare Fe3O4 spheres under the same test conditions. In the first cycle, a large and irreversible reduction current below 1.0 V corresponds to the formation of SEI films and the irreversible insertion of lithium ions in TiO2 and Fe3O4 crystals [12, 14,15,16, 20]. A pair of symmetrical redox CV peaks situated at 2.0 V and 1.75 V are attributed to the redox potential of TiO2 [20], while the redox CV peaks at 1.2 V and 0.75 V correspond to the redox potential of Fe3O4 [12, 14,15,16, 20].

Fig. 8
figure 8

First ten CV curves of a Fe3O4 spheres and b Fe3O4@TiO2@C Y–S spheres at a scan rate of 0.10 mV/s and c subsequent fifty times CV curves of Fe3O4@TiO2@C Y–S spheres at 0.50 mV/s in a potential range from 3.0 to 0.01 V versus Li/Li+

Full size image

The subsequent CV curves of Fe3O4@TiO2@C Y–S spheres were further measured to investigate the electrochemical kinetics at a high scan rate of 0.50 mV/s (Fig. 8c). Compared with the tested CV curves at 0.10 mV/s, the redox peaks potential of TiO2 shifts to 2.25 V and 1.75 V, while the anodic peak of Fe3O4 at 1.2 V (marked using one red circle in Fig. 8b) also shifts to 1.5 V (marked using another red circle in Fig. 8c). The wider space between the redox peaks indicates the increasing electrode polarization at a higher scan rate, which is in good agreement with the rate performance tests. In addition, there are no obvious changes in the shifts of peak areas and positions in the subsequent 50 cycles, indicating that the Fe3O4@TiO2@C Y–S spheres possess a good cycling stability as LIBs’ anodes. The results match well with the above galvanostatic discharge–charge cycle tests. The electrode reactions are described as follows:

$${\text{Fe}}_{3} {\text{O}}_{4} + 2{\text{Li}}^{ + } + 2{\text{e}}^{ - } \, \to \,{\text{Li}}_{2} {\text{Fe}}_{3} {\text{O}}_{4}$$
(1)
$${\text{Li}}_{2} {\text{Fe}}_{3} {\text{O}}_{4} + 6{\text{Li}}^{ + } + 6{\text{e}}^{ - } \to 3{\text{Fe}} + 4{\text{Li}}_{2} {\text{O}}$$
(2)
$$3{\text{Fe}}\, + 4{\text{Li}}_{2} {\text{O}}\, \to \,{\text{Fe}}_{3} {\text{O}}_{4} + 8{\text{Li}}^{ + } + 8{\text{e}}^{ - }$$
(3)
$${\text{C}} + x{\text{Li}}^{ + } + x{\text{e}}^{ - } \to {\text{Li}}_{x} {\text{C}}$$
(4)
$${\text{TiO}}_{2} + x{\text{Li}}^{ + } + x{\text{e}}^{ - } \to {\text{Li}}_{x} {\text{TiO}}_{2}$$
(5)
$${\text{Li}}_{x} {\text{TiO}}_{2} \to {\text{TiO}}_{2} + x{\text{Li}}^{ + } + x{\text{e}}^{ - }$$
(6)

Conclusions

The Stӧber method was employed to successfully synthesize Fe3O4@TiO2@C yolk–shell (Y–S) spheres for LIB anodes. The TiO2@C shell and void space effectively alleviate the severe volume expansion of Fe3O4 spheres and restrict the rupture of electrodes, which thus improves the electrochemical performance throughout the cycles. At a constant current density of 200 mA/g, the Fe3O4@TiO2@C Y–S spheres can release a high reversible capacity of approximately 450 mAh/g after 100 cycles, which is far higher than that of bare Fe3O4 spheres (160.2 mAh/g) and TiO2@C hollow spheres (286.5 mAh/g). These results also demonstrate that the design of the Y–S structure can be applied in developing new type electrode materials for LIBs.