Dopant-induced shape evolution of polyhedral magnetite nanocrystals and their morphology/component-dependent high-rate electrochemical performance for lithium-ion batteries

Abstract

Monodisperse Mn_xFe_3−xO₄ (x = 0, 0.3, 0.6) polyhedrons enclosed by {100}/{111} facets with different area ratios were synthesized through the thermolysis of Fe(acac)₃ and Mn(acac)₂ by effectively tuning the Mn/Fe ratio to mediate the adsorption properties of oleic acid (OA) on crystal surfaces after annealing treatment in N₂, and studied as high rate (≥1 A g⁻¹) anode materials for lithium ion batteries (LIBs). The electrochemical results show that Mn_0.6Fe_2.4O₄ octahedra possess the best rate cycling performance compared to that of Mn_0.3Fe_2.7O₄ cuboctahedra and Fe₃O₄ cubes, characterised by a 500th discharge capacity of 803.5 mA h g⁻¹ at 1 A g⁻¹ and a rate capability of 661.5 mA h g⁻¹ when cycled at 4 A g⁻¹, as a result of high electrochemical activity of {111} facets with the highest Fe atom surface density. The present results prove that the substitution of Fe by Mn in the spinel-type anode materials can result in better cycle stability and it would be helpful for the further understanding of Fe₃O₄ based anode materials and provide a simple and practical route to design high rate anode materials for lithium-ion batteries.

---

1. Introduction

As one of the most promising next-generation redox-based anode materials for LIBs, Fe₃O₄ has been extensively investigated over the last several years due to its high theoretical capacity (926 mA h g⁻¹), ecological friendliness and abundant resource, compared with the current commercial graphite anodes (theoretical capacity of 372 mA h g⁻¹).^1,2 However, its wide application in LIBs is still limited by its poor cycle stability and low rate capacity as a result of low electric conductivity, drastic volume change, easy agglomeration of Fe₃O₄ particles, etc. during the conversion reactions.³ To circumvent these problems, various approaches have been adopted to improve the electrochemical performance of Fe₃O₄, such as element doping and special structure preparation.^4,5 Thus, compounds with partial substitution of Fe in Fe₃O₄, such as ZnFe₂O₄,⁶ NiFe₂O₄,⁷ and CoFe₂O₄,⁸ have been suggested as alternative anode materials to Fe₃O₄. However, to the best of our knowledge, Fe₃O₄ based nanostructures with excellent long cycling stability (over 100 cycles) were rarely reported, let alone their high rate performance (≥1 A g⁻¹). In addition, the electrochemical performance of Mn doped Fe₃O₄ has also been rarely reported although MnO exhibited greater reversibility than ZnO and better capacity retention was achieved when the amount of Mn increases within the Zn_1−xMn_xFe₂O₄.⁹ On the other hand, many studies have demonstrated that the surface structure of redox-based electrode materials has a great influence on their electrochemical performance. For example, Xiaoling Xiao have reported that the morphology of Co₃O₄ nanocrystals (NCs) significantly affect its electrochemical performances with a weight sequence as octahedron > truncated octahedron > cube because {111} facets are more beneficial to Li⁺ transport than {100} facets.¹⁰ Similarly, Kunfeng Chen et al. believe that the electroactivity sequence of the Cu₂O crystal planes is {100} > {111} > {110}, due to the fact that the Cu atom density of the {100} surfaces is the highest among the {100}, {111} and {110} facets.¹¹ Therefore, as a typical redox-based anode material, it is interesting to investigate whether and how the exposed facets of Fe₃O₄ based anode materials affect the electrochemical performance. However, as far as we know, there are no reports on the electrochemical performance of different shaped Mn-doped Fe₃O₄ polyhedrons for LIBs. One possible reason is that, till recently, the shape controlled synthesis of Fe₃O₄ based NCs is rather difficult though various methods based on sol–gel, co-precipitation and solvothermal processes have been developed.¹² Among them, thermal decomposition of organometallic precursors have been proven effectual for the synthesis of Fe₃O₄ based NCs with regular morphologies using various solvents and surfactants through complicated control of experimental conditions.^13,14 Nevertheless, the shape-controlled synthesis of sub-micron sized Fe₃O₄ based NCs is not yet reached due to the restrictions from the condition-sensitive property of thermal decomposition method. If the shape-guiding processes in these NCs can be controlled, it could be possible to program the system to yield the final NCs with desired shape and size without a boring materials choice and experimental design.

In the work reported here, monodisperse Mn_xFe_3−xO₄ (x = 0, 0.3, 0.6) polyhedrons enclosed by different area proportions of {100}/{111} facets were synthesized by utilizing Mn²⁺ metal ions as structure-directing agents during the thermal decomposition process, in which Fe(acac)₃ and Mn(acac)₂ were employed as the precursor, benzyl ether as reaction media, and OA as surfactant. In the meantime, the binding characteristics of OA molecules with different metal ions play significant roles in the shape determination of M_xFe_3−xO₄ (M = Mn²⁺, Ni²⁺, Zn²⁺, Co²⁺) NCs. Besides, the high-rate electrochemical performance of Mn_xFe_3−xO₄ polyhedrons with different facets was investigated as anode materials for lithium ion battery. Some new insights into the morphology/component-electrochemical properties relationship of Mn doped Fe₃O₄ anodes have been achieved, which would be helpful for the further understanding of Fe₃O₄ based anode materials and the future development of high-performance electrodes.

2. Experimental section

2.1 Chemicals

Iron(III) acetylacetonate (Fe(acac)₃, 98%), manganese(II) acetylacetonate (Mn(acac)₂, 98%), nickel(II) acetylacetonate (Ni(acac)₂, 98%), zinc acetylacetonate (Zn(acac)₂, 98%), cobalt acetylacetonate (Co(acac)₂, 98%) were purchased from Aladdin Industrial Corporation, ethanol from Yasheng chemical Co., LTD and cyclohexane (C₆H₁₂, 99.5+%) from Shenbo chemical Co., LTD. Oleic acid (C₁₈H₃₄O₂, 90%) and benzyl ether (C₁₄H₁₄O, 98+%) were supplied by Alfa Aesar. All chemicals were used without further purification.

2.2 Shape-controlled synthesis of Mn-doped Fe₃O₄ NCs

Uniform-sized Mn_xFe_3−xO₄ (x = 0, 0.3, 0.6) NCs were synthesized in the presence of OA using a modified procedure described by Kim et al.¹⁵ In a typical synthesis of Mn_0.6Fe_2.4O₄ octahedrons, 1.6 mmol iron(III) acetylacetonate, 0.4 mmol manganese(II) acetylacetonate and 4 mmol OA were added to 10 ml benzyl ether. The mixture solution was then degassed at room temperature for 1 h and then heated to 260 °C at the rate of 15 °C min⁻¹ under vigorous stirring. The reaction mixture was maintained at this temperature for 30 min in a flow of argon. After that, the reaction solution was cooled to room temperature and 20 ml ethanol was added to the solution. The mixture was then centrifuged at 2500 rpm for 5 min to precipitate the synthetic products, which were then separated out and washed three times using 10 ml cyclohexane and finally re-dispersed in ethanol. Similar procedure and operations were also followed for the other synthetic experiments, which were carried out with different molar ratio of metal acetylacetonate precursors (the total amount of metal acetylacetonate precursors was fixed at 2 mmol), so as to explore the dopant caused changes in the particle's morphology. The resulting powder was finally annealed at 450 °C for 3 h at a heating rate of 2 °C min⁻¹ in N₂ atmosphere for further electrochemical characterization.

2.3 Characterization

The as-prepared products were investigated by X-ray diffractometer (SmartLab, 40 kV/30 mA, λ = 1.5418 Å/Cu Kα) for identification of crystal phases and grain size estimation. Their morphology and microstructure were characterized using field emission scanning electron microscope (FESEM, Hitachi S-4800 and JEOL JSM-7600F) and Raman/IR spectrometer (NEXUS670). The specific surface area was measured using the Brunauer–Emmett–Teller (BET) method (Micromeritics ASAP 2020 surface area and porosity analyzer) based on adsorption–desorption of nitrogen.

2.4 Electrochemical test

Electrochemical performances of the as-prepared Mn_xFe_3−xO₄ NCs were tested with the two-electrode CR2032 coin cells assembled in an argon-filled glove box. The working electrodes were prepared by coating a mixture of the Mn_xFe_3−xO₄ NCs (70 wt%), carbon black (15 wt%) and PVDF (15 wt%) on a copper foil, followed by a vacuum drying at 120 °C for 10 h. The copper foils coated with active materials were cut into circular sheets with a diameter of 1.3 cm, and the mass load of active materials was determined to be about 1.2 mg cm⁻² using a microbalance. Metallic Li sheets were used both as counter and reference electrodes and a polypropylene film (Celgard 2400) was used as a separator. The electrolyte was made of 1 M LiPF₆ dissolved in the mixture of diethyl carbonate (DEC) and ethylene carbonate (EC) (1 : 1 by volume). After assembled in the glove box at ≤1 ppm of moisture and oxygen, the cells were tested on a NEWARE BTS-5V 50 mA computer-controlled battery test station at different current densities within a voltage range from 0.01 to 3 V at 25 °C. Electrochemical impedance measurements were conducted over the 0.1 Hz to 100 kHz frequency range with a perturbation amplitude of 5 mV on a Princeton 2273 electrochemical system.

3. Results and discussion

In Fig. 1 are shown the FESEM images of Mn_xFe_3−xO₄ (x = 0, 0.3, 0.6) NCs, which were synthesized after 30 min reactions mediated by using 4 mmol surfactant OA with different amount of manganese(II) acetylacetonate and ferric acetylacetonate. It can be seen that only 90 nm-sized cubic NCs (Fig. 1A) were obtained without Mn²⁺ dopant, while when the amount of manganese(II) acetylacetonate and ferric acetylacetonate was arranged at 0.2 and 1.8 mmol, cuboctahedral NCs (Fig. 1B) with an average dimension of 94 nm were produced. After further increasing the amount of manganese(II) acetylacetonate to 0.4 mmol, 100 nm-sized octahedrons (Fig. 1C) were observed. This result clearly indicates that the Mn²⁺ dopant can significantly affect the growth of Fe₃O₄ NCs in the aspects of morphology. Higher amount of Mn²⁺ dopant can favour the development of {111} facets, leading to the formation of cuboctahedrons and octahedrons. It would practically provide an effective means to tailor the shapes of Mn_xFe_3−xO₄ NCs.

Fig. 1 FESEM images of (A) Fe₃O₄ cubes without Mn²⁺, (B) Mn_0.3Fe_2.7O₄ cuboctahedrons with 10 mol% of Mn dopant, (C) Mn_0.6Fe_2.4O₄ octahedrons with 20 mol% of Mn dopant.

Fig. 2A shows the X-ray diffraction diagrams of the as-synthesized products obtained with different amount of Mn dopant, which are in a good agreement with that of the inverse spinel Fe₃O₄ with a face-centered cubic structure (JCPDS no. 19-0629). The diffraction peaks with 2θ at 30.0°, 35.5°, 37.0°, 43.0°, 53.0°, 57.0° and 62.6° may be well ascribed to the crystallographic planes (220), (311), (222), (400), (422), (511) and (440) of normal magnetite Fe₃O₄, respectively. No impurity phases can be identified. The divalent Mn²⁺ (r_i = 96 pm) ions tend to occupy the tetrahedral sites because of larger size compared to smaller Fe³⁺ ions (r_i = 69 pm), which occupy the octahedral sites. The lattice parameter value of the compounds was calculated by using Jade software. The lattice parameter, a = 8.3915 Å, calculated for Fe₃O₄ cubes is close to the reported values (JCPDS no. 19-0629). On doping Mn²⁺, the lattice parameter increases slightly from 8.3915 Å to 8.3994 Å and 8.4002 Å as the amount of Mn²⁺ was increased to 10 mol% and 20 mol%, respectively, because of the slightly bigger size of Mn²⁺ ions. In addition, taking advantage of the diffraction-lines fitting results, the crystal size of Mn_xFe_3−xO₄ nanoparticles was also calculated as 89.7, 95.4, 102.4 nm by Scherrer's equation, which are quite close to the value measured on the FESEM images. The Raman measurements in the range 200–800 cm⁻¹ of the Mn_xFe_3−xO₄ polyhedrons have been made at room temperature and the results are shown in Fig. 2B. For Fe₃O₄ nanocubes, five Raman active modes are identified as: T_2g(1) = 206.6, E_g = 347.0, T_2g(2) = 493.5, T_2g(3) = 578.8, A_1g = 675.3 cm⁻¹ according to group theory assignment and consistent with the previous reports.²² Usually, the motion of oxygen at tetrahedral sites in the spinel compounds give bands above 600 cm⁻¹ that are of the A_1g type. The A_1g modes for Mn_xFe_3−xO₄ correspond to motion of oxygen in the MO₄ (M = Mn, Fe) tetrahedra. Upon doping, the single band splits into two at ∼671.9 and 721.1 cm⁻¹ due to the environmental change of the tetrahedrons, and therefore it confirms the substitution of Fe by Mn in Fe₃O₄.^9,16 On the other hand, FTIR spectroscopy is also an important technique to identify the stretching and bending vibrations of different materials. IR spectra of the Mn_xFe_3−xO₄ polyhedrons were recorded in the range of 1000–400 cm⁻¹, as shown in Fig. 2C. In general, stretching vibration due to metal cations present at the tetrahedral site shows a band around 600 cm⁻¹, while the cation present at the octahedral site shows a band around 400 cm⁻¹, which varies slightly depending on the strength of metal–oxygen bond.^9,17 Fe₃O₄ nanocubes show a broad band at 551.7 cm⁻¹ characteristic of Fe–O vibration. When Mn²⁺ ions are introduced into Fe₃O₄ polyhedrons with amounts of 10 and 20 mmol%, the bands shift to higher wavenumbers of 561.8 and 575.4 cm⁻¹, respectively, which may be attributed to stronger Mn–O bond. Besides, to verify the composition of all the phases, elemental analyses of Mn doped Fe₃O₄ were carried out by Inductively Coupled Plasma (ICP). The ICP results of the samples are shown in Table 1, which indicate x = 0, 0.32, and 0.61 for Mn_xFe_3−xO₄ cubes, cuboctahedrons, and octahedrons, respectively, and are close to the theoretical values.

Fig. 2 XRD patterns (A), Raman spectra (B) and infrared spectrum (C) of the Mn_xFe_3−xO₄ polyhedrons.

Table 1 Lattice parameter, crystallite size, surface area and ICP results of various compositions

Compound	Lattice parameter (Å)	Crystal size (nm)	Surface area (m^2 g^−1)	Mn (mg L^−1)	Fe (mg L^−1)
Fe3O4	8.3915	89.7	9.96	0	142.5
Mn0.3Fe2.7O4	8.3994	95.4	9.85	14.2	130.2
Mn0.6Fe2.4O4	8.4002	102.4	9.61	32.0	168.4

---

It is well known that the crystal shape is determined by the realistic growth rate ratio of different directions in a synthesis, which could be changed by selective adsorption of surfactants on the facets.^18,19 In recent years, ion doping showed attractive effect on the shape and size of nanostructures. Typically, it was found that the different adsorption properties to different metal ions of the different planes of α-Fe₂O₃ would lead to the formation of α-Fe₂O₃ nanostructures with different exposed surfaces.²⁰ Therefore, take above progresses into the consideration of different cohesive energies between metal ions and carboxylate group,²¹ it can be deduced that the changed adsorption characteristic of OA molecules on Mn_xFe_3−xO₄ facets with the adjustment of Mn²⁺ doping content may be the key factor for the formation of different shaped Mn_xFe_3−xO₄ NCs. The above mechanism was partly confirmed by a set of experiments using the standard procedure, except that different doping ion species were used. The result shows that the Zn_0.6Fe_2.4O₄ cubes (Fig. 3A), Ni_0.6Fe_2.4O₄ truncated cubes (Fig. 3B) and Co_0.6Fe_2.4O₄ cuboctahedrons (Fig. 3C) can be obtained by using 0.4 mmol zinc acetylacetonate, nickel acetylacetonate and cobalt acetylacetonate, respectively with fixed 1.6 mmol ferric acetylacetonate as precursors, suggesting that the substitution of Fe by transition metal ions is responsible for the formation of different shaped M_xFe_3−xO₄ NCs (M = Mn²⁺, Zn²⁺, Ni²⁺, Co²⁺).

Fig. 3 FESEM images of (A) Zn_0.6Fe_2.4O₄ cubes, (B) Ni_0.6Fe_2.4O₄ truncated cubes, (C) Co_0.6Fe_2.4O₄ cuboctahedrons.

In order to evaluate the relationship between crystal planes/component and high-rate electrochemical properties of Mn_xFe_3−xO₄ polyhedrons, these polyhedrons were tested as anode materials for LIBs. Before that, the specific surface areas of the three kinds of Mn_xFe_3−xO₄ polyhedrons have been measured by the Brunauer–Emmett–Teller (BET) method and are shown in Fig. 4. The measured specific surface areas for cubes, cuboctahedrons, and octahedrons are 9.96, 9.85 and 9.61 m² g⁻¹, respectively.

Fig. 4 N₂ adsorption–desorption isotherms of Fe₃O₄ cubes, Mn_0.3Fe_2.7O₄ cuboctahedrons and Mn_0.6Fe_2.4O₄ octahedrons, whose specific surface areas are 9.96, 9.85 and 9.61 m² g⁻¹, respectively.

Fig. 5A displays the 1st, 2nd, 100th, 300th and 500th discharge/charge voltage profiles of Mn_0.6Fe_2.4O₄ octahedrons at a current density of 1 A g⁻¹ in the voltage range of 0.01–3.0 V versus Li/Li⁺. It can be seen that in the first cycle, Mn_0.6Fe_2.4O₄ octahedrons exhibit an initial discharge capacity of 1169.6 mA h g⁻¹, equivalent to a consumption of 10.1 moles of Li per mole, and a charge capacity of 900.6 mA h g⁻¹ corresponding to 7.8 moles of Li, with an initial coulombic efficiency of 77.0%. The irreversible capacity of 269.0 mA h g⁻¹, as the difference between discharge and charge capacities, is observed for the first cycle and it may be attributed to the irreversible processes such as the formation of amorphous Li₂O, solid electrolyte interphase (SEI) layer and the decomposition of electrolyte, which are quite common to the systems involved in those LIB's redox reactions.^22–25 After the first discharge–charge process, the reversible discharge capacity shows an initial decrease from 907.6 mA h g⁻¹ (7.8 moles of Li) at the 2nd cycle to 837.6 mA h g⁻¹ (7.2 moles of Li) at the 100th cycle and 836.8 mA h g⁻¹ (7.2 moles of Li) at the 300th cycle, but then a slight decrease to 803.5 mA h g⁻¹ (6.9 moles of Li) after the 500th cycle, which show a high available capacity and cycle stability for the Mn_0.6Fe_2.4O₄ octahedrons at a high current density of 1 A g⁻¹.

Fig. 5 Discharge–charge curves and differential capacity profiles for Mn_0.6Fe_2.4O₄ octahedrons (A and B), Mn_0.3Fe_2.7O₄ cuboctahedrons (C and D) and Fe₃O₄ cubes (E and F) in a voltage range from 0.01 to 3.0 V at 1 A g⁻¹.

To better understand the charge–discharge behavior of the Mn_0.6Fe_2.4O₄ octahedrons, a differential capacity dQ/dV vs. voltage plot for each cycle is derived from its normal discharge–charge curve and shown in Fig. 5B. It can be seen clearly from the inset diagram that the initial discharge process is characterized by one strong reduction peak at ∼0.85 V, which can be ascribed to the reductive reaction Mn_0.6Fe_2.4O₄ + Li⁺ → Mn⁰ + Fe⁰ + Li₂O and, possibly, along with the irreversible reactions of electrolyte to form a solid electrolyte interface (SEI) layer.^26,27 For the charge process, in contrast, two diffusive oxidation peaks may be found at ∼1.62 V and 1.81 V, corresponding to the oxidation reactions of Mn⁰, Fe⁰ to Mn²⁺, Fe²⁺ and Fe³⁺, respectively during the anodic process.^28,29 As to the 2nd cycle, the initial discharge peak appear to gradually split into three consequent peaks: two small peaks of them appearing at ∼0.84 and 1.25 V, and the third strong peak occurring at ∼1.01 V, suggesting a nonuniform electrochemical reaction occurring on the anode as a possible result of incomplete electrochemically activation of Mn_0.6Fe_2.4O₄ octahedrons after the first cycle. Interestingly, the discharge peaks show a continuous intensity increase and slight voltage drop during the following 500 cycles, and a high discharge voltage of 0.87 V can be retained after 500 cycles. Obviously, this voltage stability and intensity increase of discharge peak for the Mn_0.6Fe_2.4O₄ octahedrons should demonstrate a superior cycle stability of the anode materials during the repeated lithiation/delithiation process, which in nature should be originated from its initial composition and morphology features.

In comparison with Mn_0.6Fe_2.4O₄ octahedrons, as shown in Fig. 5C and E, Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes exhibit lower initial discharge and charge capacities of 1092.0/785.6 mA h g⁻¹ (9.4/6.8 moles of Li) and 910.5/595.8 mA h g⁻¹ (7.9/5.2 moles of Li), respectively, with the corresponding coulombic efficiencies of 71.9% and 65.4%. The improvement of initial capacity and coulombic efficiency of Mn_0.6Fe_2.4O₄ octahedrons may be ascribed to the enhanced electrochemical reactivity, and therefore, utilization of Mn_0.6Fe_2.4O₄.³⁰ Moreover, both of Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes show a conspicuous deterioration of discharge–charge capacity after the initial cycle. In particular, a serious discharge capacity decline of 19.1% and 27.9% can be observed from the second to 500th cycles for the Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes, respectively, indicating a low charge–discharge reversibility of Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes anodes. In addition, as shown in Fig. 5D and F, the main discharge peaks in the first cycle for Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes appear at 0.81 V and 0.73 V, respectively, which are both lower than that of Mn_0.6Fe_2.4O₄ octahedrons and therefore reveal their poor reaction kinetics. Thus, it is believed that such relatively poor electrochemical performances of Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes should be closely associated with their component and morphological details. And this discharge–charge behavior would be worsened as the current density is further increased. Furthermore, the discharge peaks of Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes both exhibit more serious voltage drop and intensity reduction as the cycle test proceeds, which suggests some cycle degradations with the Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes as anode materials while in discharge/charge at 1 A g⁻¹. So, in general, it is believed that such electrochemical performances of Mn_0.6Fe_2.4O₄ NCs, in terms of capacity performance and cycling stability, should be associated with their component and exposed crystal planes.

As regards the cycling capacities of Mn_0.6Fe_2.4O₄ octahedrons, Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes, the experimental results measured at 1 A g⁻¹ are presented in Fig. 6A. It can be seen that the Mn_0.6Fe_2.4O₄ octahedrons show a slight capacity decrease during the initial 100 cycles. Moreover, a stable charge–discharge process can be observed in the extended cycles and a discharge capacity up to 803.5 mA h g⁻¹ (6.9 moles of Li) is achieved at 500th cycle, corresponding to 88.5% of the discharge capacity at the 2nd cycle, indicating an outstanding available capacity and cycling stability for the Mn_0.6Fe_2.4O₄ octahedrons anode. In contrast, Mn_0.3Fe_2.7O₄ cuboctahedrons exhibit a continuous capacity decline to 667.5 mA h g⁻¹ (5.8 moles of Li⁺) in the initial 250 cycles and finally to 637.8 mA h g⁻¹ (5.5 moles of Li⁺) after 500 cycles. Similarly, the discharge capacity of Fe₃O₄ cubes is abruptly decreased from 601.7 mA h g⁻¹ (5.2 moles of Li⁺) to 490.2 mA h g⁻¹ (4.2 moles of Li⁺) after 250 cycles and then falls to 433.8 mA h g⁻¹ (3.8 moles of Li⁺) after 500 cycles. These facts clearly demonstrate the significant effects of the component and morphology of Mn_xFe_3−xO₄ NCs on their high rate electrochemical performance.

Fig. 6 Cycle stabilities of Mn_0.6Fe_2.4O₄ octahedrons, Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes at 1 A g⁻¹ (A), Nyquist plots (B) and Randles plots (C) after 500th cycles at 1 A g⁻¹ in the frequency range from 100 kHz to 0.1 Hz.

To gain further insights into the component and morphology-dependent electrochemical characteristics of Mn_xFe_3−xO₄ NCs, impedance measurements were carried out in the frequency range from 0.1 Hz to 100 kHz at room temperature. The impedance spectra of Mn_xFe_3−xO₄ NCs with the fully charged state after the 500th cycle are shown in Fig. 6B, which consist of a depressed semicircle in the high to medium frequency region and an inclined line in the low-frequency region that is associated with the charge transfer process and diffusion of lithium ions into the electrode materials, respectively.³¹ The impedance data were analyzed with ZSimpWin software by fitting to an equivalent electrical circuit (see the insert of Fig. 6B) composed of electrolyte (R_e) and charge-transfer (R_ct) resistances, constant-phase element (CPE_i) (due to the depressed semicircle observed in the spectra), diffusional components like Warburg impedance (W_s) and the intercalation capacitance (C_int), similar to the circuit employed for other oxide electrodes.^32,33 The symbols are the experimental data whereas the continuous lines represent the fitted spectra, and the derived parameters are presented in Table 2. It can be found that the value of the charge transfer resistance (R_ct) is 106.4 Ω for Mn_0.6Fe_2.4O₄ octahedrons, which is obviously lower than that of Mn_0.3Fe_2.7O₄ cuboctahedrons (141.5 Ω) and Fe₃O₄ cubes (199.9 Ω), indicating that the Mn_0.6Fe_2.4O₄ octahedrons anode has a faster charge-transfer process than Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes anode. Thus, Mn_0.6Fe_2.4O₄ octahedrons are proved to be beneficial for improving anode electric conductivity and hence the cycle stability and rate capability of Mn_0.6Fe_2.4O₄ anodes. Besides, the lithium ion diffusion coefficient can be calculated from the formula as following:

D_Li⁺ = R²T²/(2A²n⁴F⁴C²σ_ω²) (1)

where D_Li⁺ is the diffusion coefficient (cm² s⁻¹), R is the gas constant (8.314 J mol⁻¹ K⁻¹), T is the absolute temperature (298.5 K), A is the surface area of the electrode based on the coin cells (1.33 cm²), n is the number of the electrons per molecule attending the electronic transfer reaction (1 in our case), F is the Faraday constant (96 486 C mol⁻¹), C is shuttle concentration of Li⁺ ions (7.69 × 10⁻³ mol cm⁻³), and the value of Warburg coefficient σ_ω is the slope of line fitted to the plot of Z′–ω^−1/2 (Fig. 6C).³⁴ It can be observed that the Warburg coefficients σ_ω of octahedrons, cuboctahedrons and cubes anodes after the 500th cycle are 51.2, 117.2, 160.3 Ω s^−0.5, respectively. Thus, according to eqn (1), the D_Li⁺ for Mn_0.6Fe_2.4O₄ octahedrons can be calculated as 1.7 × 10⁻¹³ cm² s⁻¹, which is about 5.2 times that of Mn_0.3Fe_2.7O₄ cuboctahedrons (3.3 × 10⁻¹⁴ cm² s⁻¹) and 9.4 times higher than that in Fe₃O₄ cubes (1.8 × 10⁻¹⁴ cm² s⁻¹), indicating the faster Li-ion diffusivity of Mn_0.6Fe_2.4O₄ anodes during electrochemical cycles. These results agree well with above analysis and confirm that the electrons and Li⁺ can transfer more effectively in the interface of active materials and electrolyte by increasing the proportion of {111} facets and Mn²⁺ doping content, thus resulting in the enhanced electrode reaction kinetics and better electrochemical performance of the Mn_0.6Fe_2.4O₄ octahedrons during the high rate charge/discharge process.

Table 2 The parameters of equivalent circuit for anodes composed of Mn_xFe_3−xO₄ polyhedrons

Anode	Re (Ω)	Rct (Ω)	CPE (μF)	n	σω (Ω s^−0.5)	DLi^+ (cm^2 s^−1)
Mn0.6Fe2.4O4	8.312	106.4	5.6	0.8	51.2	1.7 × 10^−13
Mn0.3Fe2.7O4	7.686	141.5	4.8	0.8	117.2	3.3 × 10^−14
Fe3O4	4.543	199.9	4.2	0.8	160.3	1.8 × 10^−14

---

To evaluate the electrochemical performance of the Mn_xFe_3−xO₄ NCs at higher current densities, the rate performances of Mn_0.6Fe_2.4O₄ octahedrons, Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes electrodes were studied at different rates from 1 to 4 A. As shown in Fig. 7, it can be seen that there are a series of discharge capacity drops in response to the step-wise increases of current density. As expected, Mn_0.6Fe_2.4O₄ octahedrons electrode exhibits relative high discharge capacities of 920.0 (8.0 moles of Li⁺), 840.3 (7.3 moles of Li⁺), 747.5 (6.5 moles of Li⁺), and 661.5 (5.7 moles of Li⁺) at the current densities of 1, 2, 3 and 4 A g⁻¹ after 10 cycles, respectively. Furthermore, after the deep charge/discharge for 10 cycles at 4 A g⁻¹, a high capacity of 906.6 mA h g⁻¹ (7.8 moles of Li⁺) is recovered over 10 cycles at 0.5 A g⁻¹ as a result of reduced electrochemical polarization, indicating stable kinetic features and good reversibility of Mn_0.6Fe_2.4O₄ octahedrons. Compared to the Mn_0.6Fe_2.4O₄ octahedrons, Mn_0.3Fe_2.7O₄ cuboctahedrons deliver much lower specific capacities of 768.0–423.6 mA h g⁻¹ (6.6–3.7 moles of Li⁺) at current rates of 1–4 A g⁻¹ and 735.5 mA h g⁻¹ (6.4 moles of Li⁺) after the extended 10 cycles at 0.5 A g⁻¹ as a result of inefficient electron and lithium-ion transportation in Mn_0.3Fe_2.7O₄ cuboctahedrons. As for Fe₃O₄ cubes, the reversible capacity continuously decayed with cycling, and exhibited a capacity as low as 218.9 mA h g⁻¹ (1.9 moles of Li⁺) at the current density of 4 A g⁻¹ and 512.8 mA h g⁻¹ (4.4 moles of Li⁺) after the 10 cycles at 0.5 A g⁻¹. The above data strongly prove the excellent redox reaction kinetics and superior cycling stability of Mn_0.6Fe_2.4O₄ octahedrons at high current densities.

Fig. 7 Rate performance of Mn_0.6Fe_2.4O₄ octahedrons, Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes.

To further explore the effects of component and exposed planes on the electrochemical properties of Mn_xFe_3−xO₄ NCs, the morphology change of Mn_xFe_3−xO₄ NCs during electrochemical cycles was analyzed at 1 A g⁻¹. Different from previous reports,^35,36 it can be found from Fig. 8 that the morphology of Mn_xFe_3−xO₄ NCs was significantly changed during cycling, indicating the inevitable destruction of Mn_xFe_3−xO₄ active materials. Compared with the morphology before discharge and charge (Fig. 8A), more rough surface of refined Mn_0.6Fe_2.4O₄ particles (Fig. 8B) can be observed with the repeated insertion and extraction of Li⁺ after initial 250 cycles. And after the extended 250 cycles, Mn_0.6Fe_2.4O₄ octahedrons were completely transformed into porous agglomerates of nano-sized Fe₃O₄ particles (Fig. 8C) as a result of repeated electrochemical grinding, indicating the complete electrochemical activation of Mn_0.6Fe_2.4O₄ octahedrons at the high current density of 1 A g⁻¹. In contrast, the pristine (Fig. 8D) and cycled (Fig. 8E) Mn_0.3Fe_2.7O₄ cuboctahedrons (after 250 cycles) indicate clearly that many Mn_0.3Fe_2.7O₄ cuboctahedrons could be survived after initial 250 cycles and residual unreacted Mn_0.3Fe_2.7O₄ nanoparticles can also be observed even after 500 cycles, revealing poorer reaction kinetics of Mn_0.3Fe_2.7O₄ cuboctahedrons compared to Mn_0.6Fe_2.4O₄ octahedrons at high current density. Even worse, Fe₃O₄ cubes showed distorted cubic morphology with a greatly increased dimension after 500 cycles at 1 A g⁻¹, which reveals the poor reaction kinetics of Fe₃O₄ cubes and even possible incomplete electrochemical reaction of internal Fe₃O₄ with Li⁺. These results offer an explanation for the above distinctly different electrochemical performance (especially high-rate capacity) of different shaped Mn_xFe_3−xO₄ polyhedrons, and prove that the original morphology has a significant effect on the reaction kinetics of Mn_xFe_3−xO₄ polyhedrons during high-rate cycling, which shows great potential for the improvement of high-rate electrochemical performances through control of well-defined crystal plane structure of Mn_xFe_3−xO₄ anode materials.

Fig. 8 FESEM images of the Mn_xFe_3−xO₄ electrodes on Cu foil and after 0, 250 and 500 discharge/charge cycles. (A–C) Mn_0.6Fe_2.4O₄ octahedrons; (D–F) Mn_0.3Fe_2.7O₄ cuboctahedrons; (G–I) Fe₃O₄ cubes.

It has been found that the first lithiation reaction is very important for subsequent electrochemical performance of redox based anode materials, and therefore, it is a key problem to improve the utilization of electrode materials in the initial cycles.^37,38 In our case, the electrochemical activity of Mn_xFe_3−xO₄ polyhedrons was successfully enhanced by increasing the content of Mn and high discharge capacity was achieved for Mn_0.6Fe_2.4O₄ octahedrons at high current density of 1 A g⁻¹. As illustrated in Fig. 9, the {100} plane of Fe₃O₄ contains only 2 Fe³⁺, but the {111} plane contains 3.75 Fe³⁺, indicating that the {111} plane has more Fe³⁺ than the {100} plane.¹⁰ The {111} plane with more metal ions is believed to be beneficial for the electronic conduction and reaction of Fe³⁺ with Li⁺. Therefore, Mn_0.6Fe_2.4O₄ octahedrons with {111} facets can react with Li⁺ more easily and can be fully activated in initial cycles even under high current densities of 1 A g⁻¹. Moreover, the particles could be simultaneously fragmented to a certain extent due to the highly uneven reaction frontiers since the active sites on the nanocrystals' facets are greatly fluctuated either in spatial distribution or in chemical activity. In this case, consequently, the fragmented nanoparticles would offer larger reaction interfacial area and render a larger utilization of electrode materials during and after the initial cycle, which should be responsible for the good long cycle performance of Mn_0.6Fe_2.4O₄ octahedrons. In contrast, the poor reaction kinetics of Mn_0.3Fe_2.7O₄ cuboctahedron and Fe₃O₄ cubes would allow only a slow and even electrochemical reaction of Mn_0.3Fe_2.7O₄/Fe₃O₄ particles with Li⁺ ions and result in a lower utilization of electrode materials due to the shroud of reaction product on the unreacted materials. Besides, it is interesting to note that Mn_0.6Fe_2.4O₄ octahedrons generate significant increased capacity retention as compared to Mn_0.3Fe_2.7O₄ cuboctahedrons and Fe₃O₄ cubes, which shows no relationship to the original exposed facets as the morphology was already destroyed during cycles. Based on the previous works on the electrochemical reversibility of metal oxides, the enhanced cycle stability of Mn_0.6Fe_2.4O₄ octahedrons may be ascribed to the great reversibility of MnO within the Mn_xFe_3−xO₄ spinel.⁹ The above insights obtained will be of benefit in the design of future high-rate anode materials for lithium ion batteries.

Fig. 9 Atomic arrangements in the (001) and (111) planes of Fe₃O₄.

4. Conclusions

Mn_xFe_3−xO₄ NCs with various morphologies have been synthesized by tuning the content of Mn²⁺ ions. The changed adsorption property of OA molecules on crystal facets composed of different metal ions has been considered as the fundamental cause of a series of morphology change. Electrochemical tests showed that the Mn_0.6Fe_2.4O₄ octahedron electrode exhibited excellent high-rate performance including high discharge capacity of 803.5 mA h g⁻¹ at a current density of 1 A g⁻¹ after 500 cycles and rate capacity of 661.5 mA h g⁻¹ at 4 A g⁻¹. Electrochemical measurements displayed that the electrochemical performance of Mn_xFe_3−xO₄ NCs can be ranked as “Mn_0.6Fe_2.4O₄ octahedron > Mn_0.3Fe_2.7O₄ cuboctahedron > Fe₃O₄ cube”, indicating the significant influence of crystal plane structure of electrode materials on their electrochemical performance. The exposed active {111} facets with highest metal atom density is believed to facilitate the redox reaction between Mn_xFe_3−xO₄ and Li⁺, resulting in the high discharge–charge capacity and excellent rate capability of Mn_xFe_3−xO₄ NCs. Besides, better cycleability were achieved by substituting the Fe with Mn in the spinel anodes. This research demonstrates a facile approach for the component and morphology-controlled synthesis of Mn_xFe_3−xO₄ NCs and provides an effective route to design high-rate anode materials for lithium-ion batteries.

Acknowledgements

The authors would like to gratefully acknowledge the funds by the Priority Academic Development Program of Jiangsu Higher Education Institutions, P. R. China, the fund by the Program of Research Innovation for University Graduate Students of Jiangsu Province (No. KYLX15_0776) and the Program for Changjiang Scholars and Innovative Research Teams in Universities (PCSIRT), IRT_15R35, of P. R. China.

References

1. X. L. Cheng, J. S. Jiang, D. M. Jiang and Z. J. Zhao, J. Phys. Chem. C, 2014, 118, 12588–12598.
2. Q. Wang, L. Jiao, H. Du, Y. Wang and H. Yuan, J. Power Sources, 2014, 245, 101–106.
3. L. H. Fan, B. J. Li, D. W. Rooney, N. Q. Zhang and K. N. Sun, Chem. Commun., 2015, 51, 1597–1600.
4. X. Fu, D. Chen, M. Wang, Y. Yang, Q. Wu, J. Ma and X. Zhao, Electrochim. Acta, 2014, 116, 164–169.
5. L. Shen, H. Song, G. Yang and C. Wang, ACS Appl. Mater. Interfaces, 2015, 7, 11063–11068.
6. L. Lian, L. Hou, L. Zhou, L. Wang and C. Yuan, RSC Adv., 2014, 4, 49212–49218.
7. M. Zhang, X. W. Gao, Z. Zi, J. Dai, J. Z. Wang, S. L. Chou, C. Liang, X. Zhu, Y. Sun and H. K. Liu, Electrochim. Acta, 2014, 147, 143–150.
8. S. Li, B. Wang, J. Liu and M. Yu, Electrochim. Acta, 2014, 129, 33–39.
9. P. F. Teh, S. S. Pramana, Y. Sharma, Y. W. Ko and S. Madhavi, ACS Appl. Mater. Interfaces, 2013, 5, 5461–5467.
10. X. Xiao, X. Liu, H. Zhao, D. Chen, F. Liu, J. Xiang, Z. Hu and Y. Li, Adv. Mater., 2012, 24, 5762–5766.
11. K. Chen, S. Song and D. Xue, CrystEngComm, 2015, 17, 2110–2117.
12. D. Su, J. Horvat, P. Munroe, H. Ahn, A. R. Ranjbartoreh and G. Wang, Chem.–Eur. J., 2012, 18, 488–497.
13. X. Li, D. Liu, S. Song, X. Wang, X. Ge and H. Zhang, CrystEngComm, 2011, 13, 6017–6020.
14. H. Khurshid, W. Li, S. Chandra, M. H. Phan, G. C. Hadjipanayis, P. Mukherjee and H. Srikanth, Nanoscale, 2013, 5, 7942–7952.
15. D. Kim, N. Lee, M. Park, B. H. Kim, K. An and T. Hyeon, J. Am. Chem. Soc., 2008, 131, 454–455.
16. J. Xie, C. Yan, Y. Zhang and N. Gu, Chem. Mater., 2013, 25, 3702–3709.
17. Y. Xu, J. Sherwood, Y. Qin, R. A. Holler and Y. Bao, Nanoscale, 2015, 7, 12641–12649.
18. G. S. Gund, D. P. Dubal, D. S. Dhawale, S. S. Shinde and C. D. Lokhande, RSC Adv., 2013, 3, 24099–24107.
19. A. Bhattacharjee, M. Ahmaruzzaman and T. Sinha, RSC Adv., 2014, 4, 51418–51429.
20. W. Wu, S. Yang, J. Pan, L. Sun, J. Zhou, Z. Dai, X. Xiao, H. Zhang and C. Jiang, CrystEngComm, 2014, 16, 5566–5572.
21. T. Bala, B. L. V. Prasad, M. Sastry, M. U. Kahaly and U. V. Waghmare, J. Phys. Chem. C, 2007, 111, 6183–6190.
22. T. Wang, X. Wang, Y. Lu, Q. Xiong, X. Zhao, J. Cai, S. Huang, C. Gu and J. Tu, RSC Adv., 2014, 4, 322–330.
23. Z. Chen, J. Zhou, X. Wang, X. Liao, X. Huang and B. Shi, RSC Adv., 2016, 6, 10824–10830.
24. R. Wang, C. Xu, J. Sun, L. Gao and C. Lin, J. Mater. Chem. A, 2013, 1, 1794–1800.
25. J. Liu, Y. Zhou, F. Liu, C. Liu, J. Wang, Y. Pan and D. Xue, RSC Adv., 2012, 2, 2262–2265.
26. F. Han, L. Ma, Q. Sun, C. Lei and A. Lu, Nano Res., 2014, 7, 1706–1717.
27. T. Xia, X. Xu, J. Wang, C. Xu, F. Meng, Z. Shi, J. Lian and J. Bassat, Electrochim. Acta, 2015, 160, 114–122.
28. C. He, S. Wu, N. Zhao, C. Shi, E. Liu and J. Li, ACS Nano, 2013, 7, 4459–4469.
29. H. Geng, Q. Zhou, J. Zheng and H. Gu, RSC Adv., 2014, 4, 6430–6434.
30. H. Wang, N. Mao, J. Shi, Q. Wang, W. Yu and X. Wang, ACS Appl. Mater. Interfaces, 2015, 7, 2882–2890.
31. H. Geng, S. Li, Y. Pan, Y. Yang, J. Zheng and H. Gu, RSC Adv., 2015, 5, 52993–52997.
32. L. Li, T. Wang, L. Zhang, Z. Su, C. Wang and R. Wang, Chem.–Eur. J., 2012, 18, 11417–11422.
33. P. F. Teh, Y. Sharma, S. S. Pramana and M. Srinivasan, J. Mater. Chem., 2011, 21, 14999–15008.
34. G. Qin, Z. Fang and C. Wang, Dalton Trans., 2015, 44, 5735–5745.
35. S. Z. Huang, J. Jin, Y. Cai, Y. Li, H. Y. Tan, H. E. Wang, G. V. Tendeloo and B. L. Su, Nanoscale, 2014, 6, 6819–6827.
36. G. L. Xu, J. T. Li, L. Huang, W. Lin and S. G. Sun, Nano Energy, 2013, 2, 394–402.
37. F. Wang, R. Robert, N. A. Chernova, N. Pereira, F. Omenya, F. Badway, X. Hua, M. Ruotolo, R. Zhang, L. Wu, V. Volkov, D. Su, B. Key, M. S. Whittingham, C. P. Grey, G. G. Amatucci, Y. Zhu and J. Graetz, J. Am. Chem. Soc., 2011, 133, 18828–18836.
38. K. Chen, D. Xue and S. Komarneni, J. Power Sources, 2015, 275, 136–143.

---