Self-assembled Fe₃O₄ nanoparticle-doped TiO₂ nanorod superparticles with highly enhanced lithium storage properties

Abstract

Rational design of nanostructured anode materials is of importance for promoting the performance of lithium ion batteries (LIBs). Here the Fe₃O₄ nanoparticles (NPs) were controllably inserted into the matrix of TiO₂ nanorods (NRs) to obtain doped-superparticles (SPs) by a facile colloidal self-assembly route and subsequent calcination. To justify the effects of the doping, the lithium storage performances of the doped-SPs were evaluated as anode materials for LIBs. The results indicated that even a slight doping of Fe₃O₄ NPs can effectively enhance the properties of the anode materials compared with raw TiO₂ NR SPs. Additionally, the Fe₃O₄/(TiO₂)₇₀ SPs showed an optimal performance in term of specific capacity, rate capacity, and cycling stability, whose reversible capacity maintained around 550 mA h g⁻¹ at a current density of 1000 mA g⁻¹ after 400 cycles. The highly enhanced lithium storage of the Fe₃O₄ NP doped-TiO₂ NR SPs can substantially be attributed to the synergism of the doped-superstructure, in which the individual merits of the Fe₃O₄ NPs and TiO₂ NRs are fully played out. This work not only demonstrates the significant effects of Fe₃O₄ NP doping for TiO₂ NR SP anode materials in LIBs, but also opens an avenue for the self-assembly synthesis of doped-SPs with extended applications.

---

Introduction

In recent years, fossil fuels such as coal, oil and natural gas have become depleted along with the rapid development of human society. Development of renewable energy is an important way to solve the energy crisis. Amongst various renewable energy devices, lithium-ion batteries (LIBs) have attracted keen research interest owing to their high energy, long life, good security and environmentally friendliness in the field of energy conversion and storage.¹ As a potential electrode material candidate for LIBs, nanocrystals (NCs) exhibit several advantages over their bulk counterparts, such as higher specific capabilities, shorter ionic diffusion pathways, and higher utilization efficiency of active materials. However, the NCs generally suffer from low energy density and poor cycling life because of high surface-to-volume ratios and excessive generation of a solid electrolyte interphase (SEI) between the electrolyte and active materials.² Fortunately, the colloidal NCs can be used as building blocks to construct superparticles (SPs) through a bottom-up self-assembly route.^3–6 The resulting NC SPs can not only shorten the lithium ion diffusion path due to the existence of pores, but also provide abundant interstitial space to tolerate volume expansion of the active materials. More importantly, the well-defined spherical geometry favors the formation of stable SEI interfaces on the outer surface.⁷

Recently, Murray and co-workers extended the concept of substitutional atomic doping to NC superlattices,⁸ in which co-assembly of two NC components with carefully adjusted size and concentration ratios led to a doped self-assembled superstructure. Similar to the atomic doping effect, this NC doping concept has the potential to substantially enhance the physical properties of NC assemblies such as electrical conductivity, as has been demonstrated in previous work.^8,9 It is therefore very appealing to test the feasibility of this NC doping concept in improving lithium storage performance. To this end, the composition of the SPs can be easily adjusted by modulating the species and proportions of the building blocks, obtaining doped-SPs. The doped-SPs possess a number of structural advantages. On one hand, the SPs preserve the intrinsic properties of individual NC building blocks. On the other hand, the NC dopants can effectively improve the intrinsic properties of the NC matrix in the SPs, achieving the synergistic effect. In this regard, the rational design of the doped-SP anode materials will provide great opportunities for promoting the electrochemical performance of LIBs.

As an arresting anode material for LIBs, TiO₂ NCs possess many advantages, such as high security, high working voltage, chemical stability, environmentally friendly and abundance on Earth.^10,11 Especially, one-dimensional TiO₂ NCs, such as nanorods (NRs), nanowires (NWs) and nanotubes (NTs), exhibit unique advantages due to the efficient electron transportation along the axial direction.¹² However, the inherent drawbacks including the poor electronic conductivity, the slow lithium ion diffusion rate and the low theoretical capacity of 167.5 mA h g⁻¹ hinder the practical lithium storage application of TiO₂ NCs.^10–12 In view of this situation, a variety of strategies, such as carbon^13–24 or graphene^25–31-coating, loading of high capacity active materials,^32–46 fabrication of hollow structures,^47–50 design of specific shapes,^51,52 defect engineering^53–55 and exposed facet control⁵⁶ have been used to improve the lithium storage properties of TiO₂ NCs. Compared with TiO₂ NCs, Fe₃O₄ NCs exhibit a much higher theoretical capacity of ∼1000 mA h g⁻¹. However, Fe₃O₄ NCs still suffer from poor cycle performance due to the tremendous volume change during the charge/discharge process.^57–59 Obviously, combining TiO₂ and Fe₃O₄ NCs to form a composite superstructure could achieve complementary advantages between stability and capacity. However, to the best of our knowledge, little work has been related to the self-assembled synthesis of the Fe₃O₄ NP doped-TiO₂ NR SPs with enhanced electrochemical lithium storage properties. The strong synergies between TiO₂ and Fe₃O₄ NC building blocks in the doped-SPs are likely to provide new opportunities for the enhancement of lithium storage performances.

Inspired by the NC doping concept,^8,9 herein we design a Fe₃O₄ NP doped-TiO₂ NR SP system which exhibits a significant synergistic effect in lithium storage. We achieved the facile synthesis of the Fe₃O₄ NP doped-TiO₂ NR SPs through a bottom-up self-assembly route in a microemulsion system and subsequent calcination. In addition to possessing the general structural features of the NC SPs, the Fe₃O₄ NP doped-TiO₂ NR SPs can also be enhanced in terms of theoretical capacity and conductivity, achieving the synergy of the two components in complement to each other. These factors have a positive impact on the lithium storage performances.

Experimental

Materials

Oleic acid (OA, 90%), 1-octadecene (ODE, 90%) and dodecyltrimethylammonium bromide (DTAB, 98%) were obtained from Sigma-Aldrich. Titanium butoxide (TBOT, 98.0%) and iron chloride hexahydrate (FeCl₃·6H₂O, 99%) were obtained from Aladdin. Sodium oleate (NaOA, >97%) was purchased from TCI. Sodium fluoride (NaF, ≥98.0%) was purchased from Sinopharm. Iron oleate (Fe(OA)₃) was synthesized by reaction of NaOA and FeCl₃·6H₂O according to the method described in the literature.⁶⁰ All chemicals were used as received without further treatment.

Synthesis of TiO₂ NRs

TiO₂ NRs were synthesized by a modified literature method.⁶¹ In a typical synthesis, 48 mL (151 mmol) of OA, 6 mL (18 mmol) of TBOT and 20 mg (0.5 mmol) of NaF were loaded in a 100 mL flask. The above mixture was degassed at 120 °C for 30 min, back-filled with a nitrogen atmosphere, and kept at 270 °C for 180 min, and then cooled to room temperature. The TiO₂ NRs were isolated by the addition of ethanol and centrifugation. The precipitates were then washed twice. Finally, the precipitated TiO₂ NRs were redispersed in hexane to form stable colloidal solution with a concentration of ∼40 mg mL⁻¹.

Synthesis of Fe₃O₄ NPs

Fe₃O₄ NPs were synthesized by a modified literature method.⁶⁰ In a typical synthesis, 36 g (40 mmol) of Fe(OA)₃, 8.6 g (30 mmol) of OA and 140 g (554 mmol) ODE were loaded into a 500 mL flask. The above mixture was degassed at 120 °C for 60 min, back-filled with a nitrogen atmosphere, and kept at 320 °C for 60 min, and then cooled to room temperature. The Fe₃O₄ NPs were isolated by the addition of isopropanol/ethanol and centrifugation. The precipitates were then washed twice. Finally, the precipitated Fe₃O₄ NPs were redispersed in hexane to form stable colloidal solutions with a concentration of ∼40 mg mL⁻¹.

Synthesis of Fe₃O₄ NP doped-TiO₂ NR SPs

In a typical synthesis, 2.0 g (6.5 mmol) of DTAB were dissolved in 20 mL (1.11 mol) of deionized water in a 250 mL flask, the premixed 1.82 mL of a hexane solution of TiO₂ NRs (0.9 mmol) and 0.182 mL of a hexane solution of Fe₃O₄ NPs (0.03 mmol) were added to the flask. The reactants in the flask were mechanically stirred for 30 min at 800 rpm, and then heated to 60 °C, and kept mechanically stirred for 60 min at 400 rpm. Afterward, the mixture was cooled to room temperature. The precipitates were collected by centrifugation, washed using deionized water/ethanol three times and vacuum dried for 12 h at 60 °C. Sequentially, the dried solids were calcinated at 500 °C for 2 h under a nitrogen atmosphere, with a heating rate of 2 °C min⁻¹. Eventually, the resulting Fe₃O₄ NP doped-TiO₂ NR SPs was denoted as Fe₃O₄/(TiO₂)₇₀ according to the number ratio between the two particles (see ESI† for the calculation procedure). In addition, the resulting products with different volume ratios between Fe₃O₄ and TiO₂ solutions were denoted as Fe₃O₄/(TiO₂)₇, Fe₃O₄/(TiO₂)₃₅ and Fe₃O₄/(TiO₂)₁₄₀, respectively. The TiO₂ SPs was synthesized using the same procedure, except that 2.0 mL of a hexane solution of TiO₂ NRs (1.0 mmol) was added.

Characterization

Transmission electron microscopy (TEM) was performed on a Tecnai G2 20 TWIN microscope under an acceleration voltage of 200 kV. High resolution transmission electron microscopy (HRTEM), scanning transmission electron microscopy (STEM) and elemental mappings were performed on a Tecnai G2 F20 S-Twin under an acceleration voltage of 200 kV. The X-ray powder diffraction (XRD) patterns of the product were recorded using an X'pert PRO diffractometer. The operating voltage and current used were 40 kV and 40 mA, respectively. Field emission scanning electron microscopy (FESEM) images and energy dispersive spectrum (EDS) of the product were taken with a Zeiss Ultra-55 microscope operated at an acceleration voltage of 5.00 kV. The Leica EM TIC 3X Ion-Beam Cutter was used to produce samples for cross-sectional imaging and EDS analyses. The acceleration voltage and gun current were 6 kV and 2.2 mA, respectively. N₂ adsorption–desorption isotherms, pore size distributions and Brunauer–Emmett–Teller (BET) specific surface area were analyzed on a Quantachrome QuadraSorb SI adsorption apparatus. The pore size distribution curves were determined by using the Barrett–Joyner–Halenda (BJH) mode. Thermogravimetric analysis (TGA) measurements were carried out on a PerkinElmer Pyris 1 thermogravimetric analyzer. Raman spectra were measured with a HORIBA LabRAM HR800 laser Raman spectrometer with the excitation wavelength of 514 nm.

Electrochemical test

The battery performance measuring was carried out using 2016-type coin cells assembled in an argon-filled glovebox, with Fe₃O₄ NP doped-TiO₂ NR SPs as the working electrode. The electrolyte was a 1 mol L⁻¹ LiPF₆ solution in a mixture of ethylene carbonate, dimethyl carbonate, and ethyl methyl carbonate (1 : 1 : 1 in volume), and a polypropylene film (Celgard-2300) was used as the separator. The working electrodes were prepared by a slurry coating procedure. The slurry was prepared as follows: TiO₂/Fe₃O₄ SPs, acetylene black, and polyvinylidene fluoride (PVDF) at a weight ratio of 7 : 2 : 1 were mixed using N-methyl-2-pyrrolidinone (NMP) as a solvent. The slurry was coated on a copper foil, which acted as a current collector. The electrodes were dried at 90 °C for 4 h in air, and then at 90 °C under vacuum for another 12 h. Galvanostatic charge/discharge tests were carried out on a battery testing system (NEWARE, China), which was cycled between 0.01 V and 3.00 V (vs. Li⁺/Li) with different current densities. Cyclic voltammetry (CV) at a scan rate of 0.1 mV s⁻¹ between 0.1 mV and 3.00 V (vs. Li⁺/Li) and electrochemical impedance spectroscopy (EIS) in a frequency range of 100 kHz to 0.01 Hz were carried out on an electrochemical workstation (AUTOLAB PGSTAT204, Metrohm).

Results and discussion

The synthesis route of the Fe₃O₄ NP doped-TiO₂ NR SPs is illustrated in Fig. 1. As building blocks, the TiO₂ NRs and Fe₃O₄ NPs were synthesized by colloidal chemical synthetic procedures (Fig. 1a). Afterwards, TiO₂ NRs and Fe₃O₄ NPs can be co-assembled through hydrophobic interactions of OA in an oil-in water (O/W) microemulsion system.^3,4 After evaporation of the organic solvent, the assembled Fe₃O₄ NP doped-TiO₂ NR colloidal SPs were formed (Fig. 1b). Inside the SPs, colloidal Fe₃O₄ NPs are wrapped by a large amount of colloidal TiO₂ NRs to form the repeating units. The OA ligands coated on the colloidal NCs can be converted to carbon by calcination in an inert atmosphere. Finally, the carbon-coated Fe₃O₄ NP doped-TiO₂ NR SPs were obtained (Fig. 1c).

Fig. 1 Schematic representation of the synthesis of the Fe₃O₄ NP doped-TiO₂ NR SPs. (a) Colloidal building blocks, (b) assembled superstructures and (c) carbonized superstructures.

The microstructures of the building blocks and assembled SPs were characterized systematically. Fig. 2 shows the XRD patterns of TiO₂ NR SPs and Fe₃O₄ NP doped-TiO₂ NR SPs. The XRD patterns of all samples can be indexed to the anatase phase of TiO₂ (JCPDS no. 73-1764). Except for Fe₃O₄/(TiO₂)₇, the diffraction peaks of Fe₃O₄ in other Fe₃O₄ NP doped-TiO₂ NR SPs are not obvious, which may be due to the high dispersibility and low amount of Fe₃O₄ NPs. The relevant Fe₃O₄ diffraction peaks detected in the XRD pattern of Fe₃O₄/(TiO₂)₇ SPs can be indexed to the cubic phase of magnetite (JCPDS no. 65-3107). With the increase of the content of Fe₃O₄, the intensity of the diffraction peaks of TiO₂ gradually decreases and broadens. According to Scherrer's equation, the average crystal size of TiO₂ for TiO₂, Fe₃O₄/(TiO₂)₁₄₀, Fe₃O₄/(TiO₂)₇₀, Fe₃O₄/(TiO₂)₃₅ and Fe₃O₄/(TiO₂)₇ SPs was determined to be 21.9, 17.2, 13.4, 12.0 and 7.5 nm, respectively. Obviously, the presence of Fe₃O₄ NPs inhibits the growth of crystalline domains of TiO₂ NRs in the assembly and subsequent calcination processes. In addition, there are no diffraction peaks of carbon materials, indicating that the carbon converted by OA ligands is amorphous.

Fig. 2 XRD patterns of the TiO₂ NR SPs and Fe₃O₄ NP doped-TiO₂ NR SPs.

Fig. 3a and b show the well-defined shape of rods and spheres of TiO₂ and Fe₃O₄ NPs, respectively. The TiO₂ NRs tend to form bundles due to the high surface energy. As can be seen from the enlarged TEM image in Fig. S1,† TiO₂ NRs have an average diameter of ∼5 nm and lengths of ∼50 nm. The Fe₃O₄ NPs exhibit good monodispersity. The average diameter of the Fe₃O₄ NPs is 17 nm. Fig. 3c and d show the representative FETEM images of the Fe₃O₄/(TiO₂)₇₀ SPs. The diameter range of the SPs is from a micron to several microns. The rough surface consists of closely packed bundles of TiO₂ NRs, and Fe₃O₄ NPs are difficult to distinguish on the surface of SPs. As can be seen from the high magnification TEM image of single SPs in Fig. 3e and S2,† some isolated Fe₃O₄ NPs are wrapped with abundant TiO₂ NRs and distributed homogeneously within the SPs. The HRTEM image in Fig. S3† reveals two different lattices with a d spacing of 0.35 and 0.25 nm, which correspond to the (101) lattice plane of TiO₂ and the (311) lattice plane of Fe₃O₄, respectively. However, unlike the assembly of spherical NCs,^3–6 the ordered assembly of building blocks was not realized in this work. This may be related to the anisotropy of TiO₂ NRs and the large shape difference between TiO₂ NRs and Fe₃O₄ NPs. Despite the randomly packed structure, the STEM image and elemental mappings shown in Fig. 3f clearly confirm the uniform distribution of Ti, Fe and C elements in the Fe₃O₄/(TiO₂)₇₀ SPs, suggesting that the homogenous assembly between TiO₂ NRs and Fe₃O₄ NPs has been achieved. According to the EDS analysis on the cross-section of the Fe₃O₄/(TiO₂)₇₀ SPs (Fig. S4†), the mass ratio between TiO₂ and Fe₃O₄ in the Fe₃O₄/(TiO₂)₇₀ SPs can be estimated to be 10.7, which is close to the feed ratio.

Fig. 3 TEM images of (a) TiO₂ NRs and (b) Fe₃O₄ NPs. (c) Low- and (d) high-magnification FESEM image of Fe₃O₄/(TiO₂)₇₀ SPs. (e) High-magnification TEM image of Fe₃O₄/(TiO₂)₇₀ SPs. For clarity, the yellow circles marked the Fe₃O₄ NPs particles on the shallow surface. (f) STEM image and Ti, Fe and C elemental mappings of Fe₃O₄/(TiO₂)₇₀ SPs.

Moreover, Fig. 4 shows the representative FESEM images of the Fe₃O₄ NP doped-TiO₂ NR SPs with different Fe₃O₄/TiO₂ ratios. Except for the Fe₃O₄/(TiO₂)₇ SPs, the morphology of Fe₃O₄ NP doped-TiO₂ NR SPs did not change significantly when varying the Fe₃O₄/TiO₂ ratios, indicating that the addition of Fe₃O₄ NPs did not affect the self-assembly behaviour of the TiO₂ NRs. However, the Fe₃O₄/(TiO₂)₇ SPs display irregular shapes, including sphere-, erythrocyte-, and shell-like morphologies. In particular, the isolated assembly of Fe₃O₄ NPs appears on the surface of the Fe₃O₄/(TiO₂)₇ (see Fig. S5†), which suggested a phase separation between the TiO₂ NRs and Fe₃O₄ NPs. Obviously, the above phenomena indicate that the excessive Fe₃O₄ doping levels interfere with the assembly of TiO₂ NRs.

Fig. 4 Representative FESEM images of (a, b) TiO₂, (c, d) Fe₃O₄/(TiO₂)₁₄₀, (e, f) Fe₃O₄/(TiO₂)₃₅ and (g, h) Fe₃O₄/(TiO₂)₇ SPs.

To investigate the specific surface areas and pore size distribution of the SPs, nitrogen adsorption–desorption analysis was performed. As shown in Fig. 5a, the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs exhibit the isotherms between type IV and V combined with the H1 hysteresis loop in the relative pressure range of 0.4–0.9 and the H3 hysteresis loop in the relative pressure range of 0.9–1.0. The complex shape of the isotherm reflects the diversity of pore structures and size distribution in the SPs. The H1 hysteresis loop is often associated with porous structures generated by agglomerates of approximately spherical particles.⁶² The H3 hysteresis loop does not exhibit any limiting adsorption at the high relative pressure range of 0.9–1.0, which indicates the aggregates of rod-like particles giving rise to slit-shaped pores.⁶² The BET specific surface area of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs is 161.79 and 134.15 m² g⁻¹, respectively. The pore volume of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs is 0.527 and 0.520 cm³ g⁻¹, respectively. Unsurprisingly, the corresponding BJH pore size distribution curves show the broader pore size distributions from a few nanometers to several tens of nanometers for both Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs. Obviously, the SPs exhibited a hierarchical distribution of pore size consisting of mesopores (2–50 nm) and macropores (>50 nm). Moreover, the integration of the Fe₃O₄ NPs in Fe₃O₄/(TiO₂)₇₀ SPs partly increases the porosity between the building blocks, resulting in the enhanced specific surface area and the narrowed pore size distribution. The hierarchical porous structure of the SPs is significant for the improvement of their electrochemical properties of lithium storage due to the existence of an efficient path for the diffusion of lithium ions and electrolytes. The carbon species in the assembled SPs, derived from the native organic ligands, also plays an important role in electrochemical lithium storage. Fig. 5b shows the TGA profiles of the SPs. The carbon content of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs is 10.3 and 10.8%, respectively. Fig. 5c displays the Raman spectra of the SPs, where the I_D/I_G values of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs are 0.77 and 0.79, respectively, consistent with the features of low-crystallized carbon. The similar I_D/I_G ratios indicate that the carbon species in both the assembled SPs have a similar degree of graphitization. These nitrogen adsorption–desorption analysis, TGA and Raman results suggest that the contribution of textural properties and carbon coating to electrochemical properties of lithium storage is nearly equivalent for the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs.

Fig. 5 (a) Nitrogen adsorption–desorption isotherms and corresponding pore size distribution (inset), (b) TGA profiles in an air atmosphere, (c) Raman spectra of Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs.

The lithium storage properties of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs are assessed through a series of electrochemical measurements. Fig. 6a shows the CV curves of the Fe₃O₄/(TiO₂)₇₀ SPs. During the first scan, a distinct reduction peak at around 1.7 V and an oxidation peak at around 2.1 V are observed, which correspond to the insertion/extraction of lithium into/out of the anatase lattice, respectively, as shown in eqn (1).¹¹

TiO₂ + xLi⁺ + xe⁻ = Li_xTiO₂ (1)

Fig. 6 Cyclic voltammograms and galvanostatic charge/discharge curves of (a, b) Fe₃O₄/(TiO₂)₇₀ and (c, d) TiO₂ SPs. Nyquist plots of Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs: (e) before and (f) after cycling.

Moreover, the distinct reduction peak approaching 0 V could be attributed to lithium storage occurring in the carbon coating. The reduction and oxidation peaks are slightly shifted in the subsequent cycles, indicating the irreversible structural modifications owing to lithium insertion/extraction and other electrolyte-related surface reactions after the first cycle. In addition, two broad reduction peaks at around 1.5 and 0.6 V are observed in the first cathodic scan. The peak at 0.6 V can be assigned to the SEI formation, whereas the peak at 1.5 V may be ascribed to some irreversible side reactions with electrolytes.³⁸ Both peaks disappear in the subsequent cycles, which indicated that the electrochemical behaviour causes irreversible capacity loss. Despite this, the substantial overlap of CV curves in subsequent cycles suggests a good structural stability and reversibility during the electrochemical process. Fig. 6b shows galvanostatic charge/discharge curves at a current density of 100 mA g⁻¹ for the Fe₃O₄/(TiO₂)₇₀ SPs. In the first cycle, the voltage plateaus located at around 1.7 and 2.1 V correspond to the insertion/extraction of lithium into/out of the anatase lattice, respectively, while the two indistinct voltage plateaus located at around 1.5 and 0.6 V in the discharge curves correspond to the irreversible side reactions with electrolytes and the formation of the SEI film, respectively. These charge and discharge plateaus are in general consistent with the above CV results. The initial discharge capacity of the Fe₃O₄/(TiO₂)₇₀ SPs reaches to 1006.5 mA h g⁻¹. The charge capacity is 478.2 mA h g⁻¹ in the first charge process. Accordingly, the initial coulombic efficiency of the Fe₃O₄/(TiO₂)₇₀ SPs are 47.5%. Similar to common anode materials, the significant capacity loss is mainly caused by the formation of SEI film during the first cycle. In the subsequent cycles, the discharge plateaus located at around 1.7 V are shortened and discharge capacities reduced. However, the coulombic efficiencies have been gradually improved, which are increased up to 94.1% in the 5^th cycle. Hereafter, charge/discharge curves are gradually stabilized, indicating the excellent stability of the Fe₃O₄/(TiO₂)₇₀ SPs. It is noteworthy that the above described CV and galvanostatic charge/discharge curves did not show the distinct electrochemical behaviours of Fe₃O₄ NPs, which may be associated with the lower content and high dispersion of Fe₃O₄ NPs in the Fe₃O₄/(TiO₂)₇₀ SPs. In order to confirm this speculation, the galvanostatic charge/discharge curves of the SPs with different doping ratios are compared. As shown in Fig. S6,† the first galvanostatic charge/discharge curves of the SPs, including TiO₂, Fe₃O₄/(TiO₂)₁₄₀, Fe₃O₄/(TiO₂)₇₀ and Fe₃O₄/(TiO₂)₃₅ do not show the remarkable voltage plateaus of Fe₃O₄. However, the first galvanostatic discharge curves of the Fe₃O₄/(TiO₂)₇ display distinct voltage plateaus located at around 0.8 V, corresponding to the discharge behaviour of Fe₃O₄ NPs. This indicates that the integrated Fe₃O₄ NPs do not significantly affect the electrochemical behaviours of TiO₂ SPs over a wide range of doping levels, which may be attributed to the unique doped-assembly structure. The similar features of the CV and galvanostatic charge/discharge curves in Fig. 6c and d also reflect that the electrochemical behaviours of the TiO₂ SPs are consistent with that of the Fe₃O₄/(TiO₂)₇₀ SPs. However, the integrated area of stabilized CV curves of the Fe₃O₄/(TiO₂)₇₀ SPs is significantly larger than that of the TiO₂ SPs, and the TiO₂ SPs exhibited a much lower initial coulombic efficiency of 39.3%. In addition, the fading of the charge/discharge capacity of the TiO₂ SPs is more serious than that of the Fe₃O₄/(TiO₂)₇₀ SPs in the subsequent cycles. As both of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs are porous microspheres and have a similar surface morphology, the effect of the surface morphology on the irreversible capacity of two types of SPs should be similar. Moreover, considering the low doping amount of Fe₃O₄, we speculate that this irreversible capacity loss is mainly caused by TiO₂ and carbon. These results again confirm the importance of the integrated Fe₃O₄ NPs in improving the lithium storage properties of Fe₃O₄/(TiO₂)₇₀ SPs.

In order to further compare the electrochemical properties of the SPs, EIS measurements were performed before and after 10 charge/discharge cycles at a current density of 1000 mA g⁻¹. The Nyquist plots were shown in Fig. 6e and f. The corresponding EIS data were fitted using the respective equivalent circuit modes (Fig. S7†) and the calculated impedance parameters are listed in Table S1.† The Nyquist plots consist of a semicircle at the high frequency which is related to the ohmic resistance (R_s, the internal resistance of electrode and electrolyte), the surface layer (SEI film) resistance (R_f) and the charge transfer resistance at the electrode/electrolyte interface (R_ct), and a straight line at the low frequency, which is related to lithium ion diffusion resistance (Warburg resistance).²⁸ As shown in Fig. 6e, the much smaller semicircle diameter of Fe₃O₄/(TiO₂)₇₀ SPs relative to TiO₂ SPs indicates a lower R_ct, while the steeper line suggests a lower Warburg resistance.¹⁵ After the cycling test, as shown in Fig. 6f, the semicircle diameters in the Nyquist plots of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs were both reduced, indicating the charge transfer resistance is reduced during cycling. The reducing trend of the charge transfer resistance may be attributed to the activation and the wetting of the electrode materials with the electrolyte. Moreover, the R_s, R_f and R_ct of Fe₃O₄/(TiO₂)₇₀ are all significantly lower than those of TiO₂ SPs after cycling. It is obvious that the resistance of the doped-SPs is significantly reduced due to the higher conductivity of Fe₃O_4. Note that the line slopes of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs were both decreased (Fig. 6f), indicating that the Warburg resistance was increased after cycling. This may be attributed to the decrease in the porosity of SPs arising from structural reconstruction occurring during repeated lithiation and delithiation. Despite the decreased ion diffusion ability upon cycling, we surmise that the significantly increased electron transport could largely increase the utilization efficiency of the active materials, thereby improving the lithium storage properties.

The rate performances of the SPs with different doping levels at varying current densities are shown in Fig. 7a. With the increase of current densities from 100 to 2000 mA g⁻¹, the discharge specific capacities of all SPs start to decline, and the discharge specific capacity can be restored to the original, or even higher capacity when the current densities are switched back to 100 mA g⁻¹. Such a result shows that the SPs have excellent rate capabilities, indicating the good structural stability of the assembled superstructures. It is worth noting that the SPs with different doping levels exhibit different rate performances in terms of capacity changes. Due to the low doping level, the capacities of Fe₃O₄/(TiO₂)₁₄₀ are only slightly improved compared with those of the pristine TiO₂ SPs, especially at high current densities of 1000 and 2000 mA g⁻¹. This result is not surprising considering the low doping level of Fe₃O₄ NPs. When the doping level of Fe₃O₄ NPs is doubled, the average discharge capacities of Fe₃O₄/(TiO₂)₇₀ are 587.4, 498.9, 457.3, 394.8 and 298.6 mA h g⁻¹ at 100, 200, 500, 1000 and 2000 mA g⁻¹, respectively, with the corresponding average charge capacities of 506.6, 487.9, 445.5, 392.5, 284.1 mA h g⁻¹, indicating a significant enhancement in the lithium storage properties. Surprisingly, further increasing the doping levels by 2 and 10 times in Fe₃O₄/(TiO₂)₃₅ and Fe₃O₄/(TiO₂)₇ did not significantly increase the capacity compared with the case of Fe₃O₄/(TiO₂)₇₀, suggesting a non-monotonic variation in capacity with respect to the change of the doping levels. Notably, the Fe₃O₄/(TiO₂)₇ even deliver a much lower discharge capacity than Fe₃O₄/(TiO₂)₇₀ and Fe₃O₄/(TiO₂)₃₅ at the large current density of 2000 mA g⁻¹, indicating the adverse effects of high doping for high rate performance. This may be attributed to the irregularly shaped superstructure arising from the phase separation between TiO₂ NRs and Fe₃O₄ NPs.

Fig. 7 (a) Rate performances of various SPs with different doping levels. (b) Cycling performances and coulombic efficiencies of Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs at 100 mA g⁻¹. (c) Cycling performances of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs at 1000 mA g⁻¹. Solid and hollow symbols denote the discharge and charge capacity, respectively.

Fig. 7b shows the cycling performances of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs at 100 mA g⁻¹. While both Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs deliver low initial coulombic efficiencies, the coulombic efficiencies gradually approach nearly 100% in subsequent cycles, indicating reasonably good cycling stability of the SPs. The reversible discharge capacities of the Fe₃O₄/(TiO₂)₇₀ and TiO₂ SPs over 100 cycles are 526.0 and 277.5 mA h g⁻¹, respectively, with the corresponding charge capacities of 525.8 and 277.7 mA h g⁻¹. In addition to the significantly enhanced capacity, the Fe₃O₄/(TiO₂)₇₀ SPs exhibit a different capacity trend compared with the TiO₂ SPs. Namely, the capacities of the Fe₃O₄/(TiO₂)₇₀ SPs decrease initially during the first 10 cycles, which is followed by a gradual capacity increase until around 40 cycles. The capacity increase with cycling may be caused by the presence of Fe₃O₄ NPs. Such a capacity recovery phenomenon associated with Fe₃O₄ electrodes is generally ascribed to the reversible growth of a polymeric gel-like SEI film resulting from kinetically activated electrolyte degradation and increasing activity of Fe₃O₄ during the conversion reaction process.^58,59 Furthermore, the Fe₃O₄/(TiO₂)₇₀ SPs exhibit remarkable long-term cycling stability at a higher current density. As shown in Fig. 7c, the cell consisting of the Fe₃O₄/(TiO₂)₇₀ SPs can maintain a reversible discharge and charge capacity of around 550 mA h g⁻¹ at 1000 mA g⁻¹ after 400 cycles, which is significantly better than TiO₂ SPs tested under the same conditions.

To explore the structural evolution of the Fe₃O₄/(TiO₂)₇₀ SPs during cycling, ex situ FESEM measurements were performed after 400 cycles at 1000 mA g⁻¹. Before the measurements, the electrodes were immersed in acetonitrile to remove PVDF. As shown in Fig. 8a and S8,† FESEM images confirm that the spherical morphology of SPs was retained after cycling, indicating the Fe₃O₄/(TiO₂)₇₀ SPs maintain an excellent structural stability during repeated lithium insertion/extraction. The dense and roughened surface of the Fe₃O₄/(TiO₂)₇₀ SPs could be due to SEI film covering (Fig. S8†). As shown in Fig. 8b, the STEM images and the corresponding elemental mappings clearly demonstrate the uniform distribution of Ti, Fe and C elements in the Fe₃O₄/(TiO₂)₇₀ SPs after cycling. The XRD pattern in Fig. S9† confirms the crystal phase stability of anatase after cycling. The above results show the stability of the doped-SPs.

Fig. 8 (a) Ex situ FESEM image and (b) STEM image and Ti, Fe, C elemental mappings of the Fe₃O₄/(TiO₂)₇₀ SPs after cycling.

The lithium storage performance of Fe₃O₄/(TiO₂)₇₀ SPs was comparable to or higher than that of previously reported works related to TiO₂-based composite anode materials (Table S2†). As a matter of fact, compared with those TiO₂-based anode materials,^32–39 the integrated Fe₃O₄ content in our Fe₃O₄/(TiO₂)₇₀ SPs is the lowest, indicating that the excellent lithium storage performance is not directly related to the content of the doped component. The fact that Fe₃O₄/(TiO₂)₇₀ SPs display the best overall electrochemical performance among various SPs explored not only shows that doping TiO₂ SPs with Fe₃O₄ NPs is a feasible and efficient way to boost the electrochemical performance, but also suggests an appropriate doping level is of critical importance for achieving the desired performance enhancement. Although Fe₃O₄ has a higher theoretical specific capacity than TiO₂, the improved capacity in Fe₃O₄/(TiO₂)₇₀ SPs cannot be simply explained by the added capacity contributed by the integrated Fe₃O₄ NPs. According to the theoretical capacity of TiO₂, Fe₃O₄ and carbon (calculated in low-crystallized carbon⁶³), the theoretical capacity of the Fe₃O₄/(TiO₂)₇₀ SPs is calculated to be only 276.2 mA h g⁻¹ (167.5 × 82.0% + 1000 × 7.7% + 600 × 10.3%), which is much lower than the experimentally measured capacity (550 mA h g⁻¹ at 1000 mA⁻¹ after 400 cycles) and even lower than the average capacity at 2000 mA g⁻¹ (298.6 mA h g⁻¹). From the theoretical calculation, the proportion of TiO₂, Fe₃O₄ and carbon that contributes to the experimentally measured capacity (550 mA h g⁻¹) is 25.0, 14.0 and 11.2%, respectively. The sum of the theoretical proportions is less than 100% due to the presence of the extra capacity. The extra capacity over the theoretical capacity might be partially attributed to interfacial storage,^64,65 as the porous structures and high surface areas of Fe₃O₄/(TiO₂)₇₀ SPs could provide many available extra sites for lithium storage at the interfaces.⁶⁴ Nonetheless, the theoretical contribution of Fe₃O₄ dopants is limited, indicating that a synergistic effect, presumably originating from the strong interfacial interactions between TiO₂ NRs and Fe₃O₄ NPs, plays an important role in enhancing the electrochemical performance of Fe₃O₄ NP-doped TiO₂ SPs. Additionally, the non-monotonic performance variation with the change of doping levels implies that an appropriate doping level is necessary to achieve the strong synergistic effect.

The significantly enhanced capacity, excellent cycling performance and superior rate capacity of the Fe₃O₄ NP doped-TiO₂ NR SPs can be attributed to the following aspects contributed by the integrated Fe₃O₄ NPs: (1) the integrated Fe₃O₄ NPs can improve the poor conductivity of TiO₂ SPs; (2) the Fe₃O₄ NPs with a higher theoretical capacity can contribute to the capacity of SPs; and (3) the porosity and stability of the TiO₂ matrix can provide enough stable void space to buffer the volume expansion of the Fe₃O₄ NPs, improving cycling stability. More importantly, unlike previous TiO₂-based composite anode materials,^32–39 the unique architecture of our doped-SPs has the potential to maximize the contact, and therefore interactions, between Fe₃O₄ and TiO₂. Therefore, a stronger synergistic effect is expected to develop in such doped-SPs. In addition, the ligand-carbonization step at 500 °C could further strengthen the intimate contact between Fe₃O₄ and TiO₂, potentially at the atomic level, which may contribute largely to the strong synergistic effect. The strong synergistic effect further enhances the interfacial lithium storage of SPs. Although our doped-SPs possess advantageous structural features for achieving the strong synergistic effect, we should stress that revealing the exact origin is not easy, which may require various in situ characterization techniques to probe the dynamic interactions between Fe₃O₄ and TiO₂ at the atomic level. Finally, the inherent advantages of SPs may also play important roles in enhancing the electrochemical performance. These include the stable electrochemical performance due to the formation of stable SEI interfaces on the outer surface,⁷ and the close interfacial contact between the NPs and electrolytes due to hierarchical pore size distribution and large specific surface area.⁴⁷ In addition, the carbon-coating layers transformed from the OA ligand molecules on the surface of TiO₂ NRs and Fe₃O₄ NPs improve the conductivity of SPs and prevent aggregation of TiO₂ and Fe₃O₄ components during repeated cycling.⁷

Conclusions

We have successfully synthesized the Fe₃O₄ NP doped-TiO₂ NR SPs via a colloidal co-assembly route followed by calcination. Electrochemical experiments clearly indicate the structural advantages of the Fe₃O₄ NP doped-TiO₂ NR SPs as anodes for LIBs. The Fe₃O₄ NP doped-TiO₂ NR SPs exhibit highly enhanced electrochemical properties such as high specific capacity, rate capacity and stable cycle performance of lithium storage. The structural characteristics of the doped-SPs enable excellent lithium storage properties due to the synergistic effect between the integrated Fe₃O₄ NPs and the porous TiO₂ NR matrix. Importantly, our studies reveal that an optimized doping level is necessary to achieve a strong synergistic effect and therefore enhanced electrochemical performance. Moreover, we have demonstrated that doping TiO₂ SPs with a small amount of Fe₃O₄ NPs can boost the lithium storage properties. The doping assembly strategy can be extended to other doped systems, such as metal/metal oxides, metal/metal sulphides and metal sulphide/metal oxides, and so on. These doped-SP systems present broad prospects for practical applications in various areas.

Conflicts of interest

There are no conflicts to declare.

Acknowledgements

A. D. acknowledges financial support from MOST (2017YFA0207303 and 2014CB845602), Key Basic Research Program of Science and Technology Commission of Shanghai Municipality (17JC1400100), Shanghai International Science and Technology Cooperation Project (15520720100), and NSFC (21733003 and 21373052). B. X. acknowledges financial support from State Key Laboratory of Molecular Engineering of Polymers at Fudan University (K2017-22).

Notes and references

1. D. Larcher and J. M. Tarascon, Nat. Chem., 2015, 7, 19.
2. S. H. Yu, S. H. Lee, D. J. Lee, Y. E. Sung and T. Hyeon, Small, 2015, 7, 2146.
3. F. Bai, D. S. Wang, Z. Y. Huo, W. Chen, L. P. Liu, X. Liang, C. Chen, X. Wang, Q. Peng and Y. D. Li, Angew. Chem., Int. Ed., 2007, 46, 6650.
4. J. Q. Zhuang, H. M. Wu, Y. G. Yang and Y. C. Cao, Angew. Chem., Int. Ed., 2008, 47, 2208.
5. D. L. Liu, F. Zhou, C. C. Li, T. Zhang, H. H. Zhang, W. P. Cai and Y. Li, Angew. Chem., Int. Ed., 2015, 54, 9596–9600.
6. J. E. Park, D. R. Hickey, S. Jun, S. Kang, X. Hu, X. J. Chen and S. J. Park, Adv. Funct. Mater., 2016, 26, 7791–7798.
7. 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.
8. M. Cargnello, A. C. Johnston-Peck, B. T. Diroll, E. Wong, B. Datta, D. Damodhar, V. V. T. Doan-Nguyen, A. A. Herzing, C. R. Kagan and C. B. Murray, Nature, 2015, 524, 450.
9. A. G. Dong, Sci. Bull., 2015, 60, 1964.
10. D. Deng, M. G. Kim, J. Y. Lee and J. Cho, Energy Environ. Sci., 2009, 2, 818.
11. T. Song and U. Paik, J. Mater. Chem. A, 2016, 4, 14.
12. M. Ge, C. Cao, J. Huang, S. Li, Z. Chen, K. Q. Zhang, S. S. A. Deyab and Y. Lai, J. Mater. Chem. A, 2016, 4, 6772.
13. S. J. Park, H. Kim, Y. J. Kim and H. Lee, Electrochim. Acta, 2011, 56, 5355.
14. L. Gao, X. R. Li, H. Hu, G. J. Li, H. W. Liu and Y. Yu, Electrochim. Acta, 2014, 120, 231.
15. Y. F. Sun, J. B. Zhu, L. F. Bai, Q. Y. Li, X. Zhang, W. Tong and Y. Xie, Inorg. Chem. Front., 2014, 1, 58.
16. X. Bai, T. Li, Y. X. Qi, X. P. Gao, L. W. Yin, H. Li, H. L. Zhu, N. Lun and Y. J. Bai, Electrochim. Acta, 2015, 169, 241.
17. H. Geng, H. Ming, D. Ge, J. Zheng and H. Gu, Electrochim. Acta, 2015, 157, 1.
18. H. Liu, W. Li, D. Shen, D. Zhao and G. Wang, Adv. Mater. Interfaces, 2015, 2, 1500239.
19. S. Gao, Y. Yan and G. Chen, ACS Sustainable Chem. Eng., 2016, 4, 844.
20. Y. Cheng, Z. Chen, H. Wu, M. Zhu and Y. Lu, Adv. Funct. Mater., 2016, 26, 1338.
21. R. Wu, S. Shen, G. Xia, F. Zhu, C. Lastoskie and J. Zhang, ACS Appl. Mater. Interfaces, 2016, 8, 19968.
22. H. B. Huang, Y. Yang, L. H. Chen, Y. Wang, S. Z. Huang, J. W. Tao, X. T. Ma, T. Hasan, Y. Li, Y. Xu and B. L. Su, Nanoscale, 2016, 8, 10928.
23. C. Chen, B. Zhang, L. Miao, M. Yan, L. Mai, Y. Huang and X. Hu, J. Mater. Chem. A, 2016, 4, 8172.
24. L. Y. Wang, X. Bai, Y. Wu, N. Lun, Y. X. Qi and Y. J. Bai, Electrochim. Acta, 2016, 212, 155.
25. W. Li, F. Wang, Y. Liu, J. Wang, J. Yang, L. Zhang, A. A. Elzatahry, D. Al-Dahyan, Y. Xia and D. Zhao, Nano Lett., 2015, 15, 2186.
26. Z. Zhang, L. Zhang, W. Li, A. Yu and P. Wu, ACS Appl. Mater. Interfaces, 2015, 7, 10395.
27. S. Yang, C. Cao, P. Huang, L. Peng, Y. Sun, F. Wei and W. Song, J. Mater. Chem. A, 2015, 3, 8701.
28. S. Jiang, R. Wang, M. Pang, H. Wang, S. Zeng, X. Yue, Y. Ni, Y. Yu, J. Dai, S. Qiu and Z. Zhang, Electrochim. Acta, 2015, 182, 406.
29. S. Q. Guo, M. M. Zhen, L. Liu and Z. H. Yuan, Chem.–Eur. J., 2016, 22, 11943.
30. P. Zheng, T. Liu, Y. Su, L. Zhang and S. Guo, Sci. Rep., 2016, 6, 36580.
31. Y. Liu, A. A. Elzatahry, W. Luo, K. Lan, P. Zhang, J. Fan, Y. Wei, C. Wang, Y. Deng, G. Zheng, F. Zhang, Y. Tang, L. Mai and D. Zhao, Nano Energy, 2016, 25, 80.
32. S. Li, M. Wang, Y. Luo and J. Huang, ACS Appl. Mater. Interfaces, 2016, 8, 17343.
33. H. Xu, X. D. Zhu, K. N. Sun, Y. T. Liu and X. M. Xie, Adv. Mater. Interfaces, 2015, 2, 1500239.
34. H. Wang, G. Wang, S. Yuan, D. Ma, Y. Li and Y. Zhang, Nano Res., 2015, 8, 1659.
35. D. Zhao, Q. Hao and C. Xu, Electrochim. Acta, 2016, 211, 83.
36. L. Gao, S. Li, D. Huang, Y. Shen and M. Wang, Electrochim. Acta, 2015, 182, 529.
37. G. Luo, W. Liu, S. Zeng, C. Zhang, X. Yu, Y. Fang and L. Sun, Electrochim. Acta, 2015, 184, 219.
38. S. Guo, J. Liu, S. Qiu, W. Liu, Y. Wang, N. Wu, J. Guo and Z. Guo, J. Mater. Chem. A, 2015, 3, 23895.
39. R. Dai, A. Zhang, Z. Pan, A. M. Al-Enizi, A. A. Elzatahry, L. Hu and G. Zheng, Small, 2016, 12, 2792.
40. J. Y. Liao, B. De Luna and A. Manthiram, J. Mater. Chem. A, 2016, 4, 801.
41. X. Li, Y. Chen, H. Wang, H. Yao, H. Huang, Y. W. Mai, N. Hu and L. Zhou, Adv. Funct. Mater., 2016, 26, 376.
42. Z. Yi, Q. Han, P. Zan, Y. Cheng, Y. Wu and L. Wang, J. Mater. Chem. A, 2016, 4, 12850.
43. X. Wang, L. Fan, D. Gong, J. Zhu, Q. Zhang and B. Lu, Adv. Funct. Mater., 2016, 26, 1104.
44. D. Wu, R. Yang, Q. Sun, L. Shen, W. Ji, R. Shen, M. Jiang, W. Ding and L. Peng, Electrochim. Acta, 2016, 211, 832.
45. X. Zhang, H. Chen, Y. Xie and J. Guo, J. Mater. Chem. A, 2014, 2, 3912.
46. J. Guo, X. Zhang, Y. Sun, H. Liu and X. Zhang, Mater. Lett., 2017, 195, 104.
47. H. Ren, R. Yu, J. Wang, Q. Jin, M. Yang, D. Mao, D. Kisailus, H. Zhao and D. Wang, Nano Lett., 2014, 14, 6679.
48. J. Jin, S. Z. Huang, J. Shu, H. E. Wang, Y. Li, Y. Yu, L. H. Chen, B. J. Wang and B. L. Su, Nano Energy, 2015, 16, 339.
49. B. Y. Guan, L. Yu, J. Li and X. W. D. Lou, Sci. Adv., 2016, 2, e1501554.
50. H. Xie, L. Hu, F. Wu, M. Chen and L. Wu, Adv. Sci., 2016, 3, 1600162.
51. M. Yarali, E. Biçer, S. A. Gürsel and A. Yürüm, Electrochim. Acta, 2016, 220, 453.
52. W. Yu, Y. Liu, N. Cheng, B. Cai, K. K. Kondamareddy, S. Kong, S. Xu, W. Liu and X. Z. Zhao, Electrochim. Acta, 2016, 220, 398.
53. W. Zhang and D. Liu, Electrochim. Acta, 2015, 156, 53.
54. J. Zheng, G. Ji, P. Zhang, X. Cao, B. Wang, L. Yu and Z. Xu, Chem.–Eur. J., 2015, 21, 18309.
55. C. Wang, F. Wang, Y. Zhao, Y. Li, Q. Yue, Y. Liu, Y. Liu, A. A. Elzatahry, A. Al-Enizi, Y. Wu, Y. Deng and D. Zhao, Nano Res., 2016, 9, 165.
56. X. L. Cheng, M. Hu, R. Huang and J. S. Jiang, ACS Appl. Mater. Interfaces, 2014, 6, 19176.
57. Y. Wang, L. Zhang, Y. Wu, Y. Zhong, Y. Hu and X. W. D. Lou, Chem. Commun., 2015, 51, 6921.
58. W. Zhang, X. Li, J. Liang, K. Tang, Y. Zhu and Y. Qian, Nanoscale, 2016, 8, 4733.
59. Y. Wang, Q. Qu, Y. Han, T. Gao, J. Shao, Z. Zuo, W. Liu, Q. Shi and H. Zheng, J. Mater. Chem. A, 2016, 4, 10314.
60. J. Park, K. An, Y. Hwang, J. G. Park, H. J. Noh, J. Y. Kim, J. H. Park, N. M. Hwang and T. Hyeon, Nat. Mater., 2004, 3, 891.
61. Y. D. Liu, A. W. Tang, Q. Zhang and Y. D. Yin, J. Am. Chem. Soc., 2015, 137, 11327.
62. K. S. Sing, D. H. Everett, R. A. W. Haul, L. Moscou, R. A. Pierotti, J. Rouquerol and T. Siemieniewska, Pure Appl. Chem., 1985, 57, 603.
63. H. Higuchi, K. Uenae and A. Kawakami, J. Power Sources, 1997, 68, 212.
64. J. Y. Shin, D. Samuelis and J. Maier, Adv. Funct. Mater., 2011, 21, 3464.
65. Z. Xing, Z. Ju, J. Yang, H. Xu and Y. Qian, Nano Res., 2012, 5, 477.

---

Footnote

† Electronic supplementary information (ESI) available: Calculation of ratio, TEM images, HRTEM image, EDX spectrum, FESEM images, galvanostatic discharge/charge curves, equivalent circuits, impedance parameters, XRD pattern, and lithium storage performance comparison. See DOI: 10.1039/c7se00460e

---