Rechargeable lithium-ion batteries (LIBs) have been widely used in electric vehicles, portable electronics, and stationary energy storage due to their outstanding advantages such as long lifespan, high energy density and environmental friendliness [1-8]. Generally, graphite has been used for the anodes in commercial LIBs because of its structural stability during the cycling process [9]. However, the theoretical specific capacity of graphite is only 372 mAh/g, making it difficult for graphite anodes to meet the increasing demands for energy storage in various electronic devices [10]. Therefore, the development of LIB anodes with a high specific capacity, and excellent cycling stability is urgently necessary [11-13].

Transition metal oxides have been regarded as promising anode materials due to their relatively low environmental impact, low cost, enhanced safety and high theoretical capacity. In particular, Fe₂O₃ displays many advantages for use in LIB anodes including a high theoretical capacity (1060 mAh/g), low cost, and abundant reserves. However, its low conductivity, volume change, and severe pulverization lead to the loss of battery performance during long-term cycling process [14]. Recently, it was found that the electrochemical performance of the Fe₂O₃ anode can be greatly improved when it is coated with a carbon material to form the core-shell structure. For example, Zheng et al. prepared a core-shell Fe₂O₃@carbon material via sol-gel coating followed by the carbonization process that demonstrated excellent LIB performance (1142 mAh/g at 0.2 A/g after 100 cycles) [15]. Gao et al. wrapped Fe₂O₃ with carbon to effectively alleviate the volume changes during long-term electrochemical reactions. When used as the anode in LIBs, the wrapped Fe₂O₃ exhibited a stable capacity of 711.2 mAh/g at a current density of 0.5 A/g for 400 cycles [16]. While these strategies facilitate electron transport and alleviate the volume expansion of Fe₂O₃, the obtained Fe₂O₃-based anodes still exhibited capacity fading during long-term cycling. Moreover, the excessive internal space between the carbon and Fe₂O₃ greatly decreased the packing density of active materials, giving rise to a low volumetric energy density of the anode [17]. Thus, it is important to fabricate a core-shell Fe₂O₃@carbon material with an appropriate cavity size and suitable thickness of the carbon layer. Recently, metal-organic frameworks (MOFs) have been considered to be good precursors for the fabrication of the structure-tailored N-doped carbon (N-C) materials [18]. In particular, due to its facile synthesis, low cost and tunable size, ZIF-8 has been used as a precursor to prepare a wide range of N-C materials [19, 20].

Herein, we developed a facile method for the preparation of the core-shell Fe₂O₃@N-C anode. First, we fabricate ZIF-8 on the surface of cube-like Fe₂O₃. Then, the precursor was carbonized first at 700 ℃ for 2 h under Ar atmosphere and then at 350 ℃ for 2 h under air. Benefitting from a moderate-size cavity between the N-C layer and Fe₂O₃, the core-shell Fe₂O₃@N-C showed long-term cycle life (803.6 mAh/g at 1.0 A/g after 1100 cycles).

The fabrication of the Fe₂O₃@N-C is illustrated in Scheme 1. Briefly, cube-like Fe₂O₃ particles were first prepared according to the method described in a previous report [21]. Then, the Fe₂O₃nanocubes were directly coated with ZIF-8 using an in-situ growth method, and the obtained product was named Fe₂O₃@ZIF-8-0.08. At last, Fe₂O₃@ZIF-8-0.08 was annealed under Ar atmosphere at 700 ℃ for 2 h and then annealed under air at 350 ℃ for 2 h. The as-synthesized material was named Fe₂O₃@N-C. The experimental details are provided in the experimental section.

Fe₂O₃ was characterized by field emission scanning electron microscopy (FESEM) and powder X-ray diffraction (PXRD). As shown in Fig. 1A, Fe₂O₃ exhibits a uniform cube-like structure. PXRD results confirm that the as-prepared Fe₂O₃ is hematite (α-Fe₂O₃) (Fig. S1A in Supporting information). Compared to the original Fe₂O₃, the surface of Fe₂O₃@ZIF-8-0.08 was rough, which can be attributed to the formation of ZIF-8 on the surface of the Fe₂O₃nanocubes (Figs. 1B and C). In addition, the PXRD patterns of Fe₂O₃@ZIF-8-0.08 match the combination of the patterns of α-Fe₂O₃ and ZIF-8 quite well (Fig. S1B in Supporting information). The core-shell structure of Fe₂O₃@ZIF-8-0.08 was further confirmed by transmission electron microscopy (TEM). It is observed from the obtained SEM image (Fig. S2 in Supporting information) that the ZIF-8 nanoparticles with an average size of 8 nm are densely coated on the Fe₂O₃ surface and form a ZIF-8 shell. Moreover, transmission electron microscope-energy dispersive X-ray spectroscopy (TEM-EDX) reveals that Zn and N elements are uniformly distributed in the shell of Fe₂O₃@ZIF-8-0.08, while the Fe element is distributed in the core (Fig. S3 in Supporting information). Furthermore, we also measured the porosity and surface area of Fe₂O₃ and Fe₂O₃@ZIF-8-0.08. It is found that the Fe₂O₃ displays a typical type Ⅳ isotherm with a surface area of 29.96 m²/g and a pore size distribution centered at 1.48 nm. The Fe₂O₃@ZIF-8-0.08 displays a typical type Ⅰ isotherm with a surface area of 432.24 m²/g and a pore size distribution centered at 1.1 nm (Figs. S4A-D in Supporting information). These results indicate that the ZIF-8 shell had been successfully coated on the surface of Fe₂O₃.

The morphology of the Fe₂O₃@N-C was first investigated by FESEM (Fig. 1D). It is observed that the nanoparticles are perfectly maintained and uniformly distributed. After the annealing treatment, the original morphology of the ZIF-8 shell was destroyed, and porous carbon layers were formed. TEM images show that a thin carbon shell was uniformly coated on the inner Fe₂O₃ particles (Fig. 1E). It is important to note that Fe₂O₃ exhibits a sphere-like structure instead of a cube-like structure, with a cavity created between the carbon layer and the Fe₂O₃ particles (Fig. 1E). To elucidate the origin of the structural change, we compared the average diameters of the Fe₂O₃ particles before and after the 700 ℃ treatment. As shown in Fig. S5 (Supporting information), the sizes of the initial Fe₂O₃ particles are mainly distributed around 556 nm, which is larger than that of Fe₂O₃-700 (526 nm). The average diameter of the Fe₂O₃ particles is 528 nm, also larger than that of the Fe₂O₃-700 (500 nm). This decrease in the particle size may be attributed to the condensation of the loose microstructure of Fe₂O₃during the pyrolysis process. Furthermore, the EDS mappings results (Figs. 1F-K) show the uniform distribution of Fe and O elements inside, and of the C, N and Zn elements outside, confirming that the sample after the annealing consists of the N-C shell and the Fe₂O₃ core.

XRD measurements were carried out to study the structure and composition of the Fe₂O₃@N-C. As shown in Fig. 2A, the hematite phase appears after the pyrolysis process. Carbon diffraction peaks were not found in the XRD pattern, indicating the presence of an amorphous carbon structure [22]. In addition, the Raman spectrum of the Fe₂O₃@N-C (Fig. 2B) shows a pronounced wide peak at approximately 1500 cm⁻¹ that corresponds to amorphous carbon, further confirming the amorphous structure of the carbon shell [23]. The surface chemical composition of the Fe₂O₃@N-C was studied by X-ray photoelectron spectroscopy (XPS). As shown in Fig. 2C, typical Fe, Zn, C, O and N signals were detected. In the C 1s high-resolution spectrum (Fig. 2D), four distinct peaks located at 284.5, 284.8, 285.5 and 288.5 eV correspond to C=C, C-C, C-O and C-N, respectively [24, 25]. The Fe 2p high-resolution XPS spectrum (Fig. 2E) is deconvoluted into five peaks, with the peaks at 712.1 eV and 725.1 eV assigned to Fe 2p_3/2 and Fe 2p_1/2 of Fe³⁺, the peaks at 710.3 eV and 723.4 eV assigned to Fe 2p_3/2 and Fe 2p_1/2 of Fe²⁺, and the satellite peaks, located at 719.0 eV and 732.7 eV, respectively [14, 26, 27]. The XPS spectrum of Zn 2p is fitted by two peaks (Fig. 2F), in which the Zn 2p_3/2 signal centered at 1021.5 eV is assigned to the Zn–O bond and the peak at 1044.6 eV assigned to Zn 2p_1/2 [28]. We also fitted the O 1s spectrum, as shown in Fig. S6A (Supporting information), and the peaks at approximately 530.0 eV are assigned to O²⁻, while the peaks at 531.0 and 532.2 eV are attributed to C-O and Zn–O, respectively [29]. In the N 1s high-resolution spectrum (Fig. S6B in Supporting information), the three distinct peaks located at 398.4, 399.8 and 401.2 eV correspond to pyridinic N, pyridonic N and graphitic N, respectively [25]. According to the above analysis, an Fe₂O₃ core encapsulated with an N-doped amorphous carbon hollow shell structure was successfully prepared.

The electrochemical properties of the Fe₂O₃@N-C were first studied by cyclic voltammetry (CV) in the voltage range of 0.01-3.00 V (vs. Li⁺/Li) at a scan rate of 0.1 mV/s (Fig. 3A). The sharp peak observed in the first sweep cycle at 0.61 V can be attributed to the formation of a solid electrolyte interphase (SEI) film and the reduction of Fe³⁺ to Fe⁰, and the anodic peak at 1.65 V is related to the oxidation of Fe⁰ to Fe³⁺ [30-32]. In the subsequent 2^nd and 3^rd cycles, the cathodic and anodic peaks shift to 0.94 V and 1.71 V, respectively, implying the improved electrical contact between the electrolyte and electrodes and irreversible phase transformation [33, 34]. The well-overlapped CV curves in the 2^nd and 3^rd cycles indicate good electrochemical reversibility. The charge-discharge profiles of the Fe₂O₃@N-C electrode at different cycles under a current density of 1.0 A/g are shown in Fig. 3B. In the first cycle, the Fe₂O₃@N-C shows a voltage plateau at 0.7 V that is related to the reduction of Fe₂O₃ and the formation of the SEI layer. Interestingly, the subsequent 2^nd, 3^rd and 100^th charge-discharge profiles almost have a similar voltage plateau at approximately 1.0 V that originates from either textural modifications or the drastic lithium-driven structural change [35]. Then, the voltage plateau of 300^th drops to about 0.85 V mainly attributing to the alteration in the polarization arising from the SEI film. The Fe₂O₃@N-C electrode delivers an initial discharge-charge capacity of 1371.6/806.3 mAh/g with an initial coulombic efficiency (CE) of 58.4%. Then, the discharge capacity gradually increases to 1000 mAh/g after 300 cycles with high CEs of approximately 99.5%, demonstrating its long-term cycling reversibility. This phenomenon can be related to the continuous activation of Fe₂O₃, along with long-term cycling stability [36, 37]. Furthermore, the long-term cycling stability tests were also carried out at a current density of 1.0 A/g in order to evaluate the performance of the electrodes in LIBs. As shown in Fig. 3C, the Fe₂O₃@N-C anode retains a reversible capacity of 803.6 mAh/g after 1100 cycles. Comparison to the reports shows that the cycling performance of Fe₂O₃@N-C is better than those of some reported ferric oxide-based anode materials (Table S1 in Supporting information). The pure Fe₂O₃ anode shows a significant capacity decay (only 274.5 mAh/g after 300 cycles), further confirming the importance of the carbon shell and cavity for cycle stability. The rate capability of the Fe₂O₃@N-C electrode was investigated in order to examine the suitability of this electrode for practical applications. As shown in Fig. 3D, the Fe₂O₃@N-C electrode exhibits the average capacities of 951.7, 838.3, 768.9, 708.7, 627.3 and 324.9 mAh/g at the current densities of 0.1, 0.2, 0.5, 1.0, 2.0 and 5.0 A/g, respectively. Subsequently, a higher capacity of 1035.6 mAh/g can be obtained when the current density returns to 0.1 A/g, demonstrating its high rate capability (Fig. 3D). By contrast, the Fe₂O₃-700 electrode exhibits lower capability than that of the Fe₂O₃@N-C electrode, which exhibits the average specific capacities of 825.1, 717.4, 638.4, 561.2, 488.6 and 342.3 mAh/g at the current densities of 0.1, 0.2, 0.5, 1.0, 2.0 and 5.0 A/g, respectively, and returns to the capacity of 840 mAh/g at 0.1 A/g.

To further understand the improved performance, electrochemical impedance spectroscopy (EIS) was performed for the Fe₂O₃-700, Fe₂O₃-@N-C-0.08, Fe₂O₃-@N-C, and Fe₂O₃-@N-C samples after 300 cycles with the obtained EIS results shown in Fig. 3E and the equivalent circuit presented in the inset. An examination of Fig. 3E shows that all EIS plots display a semicircle in the range from high to medium frequency and a line inclined at approximately 45° at low frequencies. R_e is the electronic resistance of active materials and the basic parameter for the characterization of the transport resistance in materials. The semicircle is due to two contributions, namely the charge transfer resistance (R_ct) of the electrolyte-electrode interface and the solid electrolyte interface resistance (R_SEI). The low-frequency line at 45° line corresponds to the Warburg impedance (Z_W) that is related to the Li⁺ diffusion within the cathode materials and R_e is the ohmic resistance [38, 39]. The values of these parameters are presented in Table S2 (Supporting information). It is observed from an examination of the data presented in Table S2 that the R_ct for Fe₂O₃-700 is 105.9 Ω and that for Fe₂O₃-@N-C is 83.4 Ω, suggesting that the carbon coating can accelerate charge transfer during the test. The R_ct of the Fe₂O₃@N-C electrode after 300 cycles (36.9 Ω) is much smaller than that prior to cycling, which ensured the superior performance of this electrode during long-term cycling. The R_SEI for Fe₂O₃@N-C (16.60 Ω) is much smaller than that after 300 cycles (70.28 Ω), which is attributed to the formation of a thicker SEI film.

To understand the chemical reaction kinetics of Fe₂O₃@N-C during the charge-discharge process, CV tests at different scan rates were conducted (Fig. 4A). Based on the power-low relationship (eq. 1), the mathematical relationship between the scan rates (v) and the peak current (i) is

where the parameter b can be calculated from the linear plot slop of log(v) versus log(|i|) (eq. 2) [40, 41]. Generally, if the b value is 0.5, the electrode exhibits a typical diffusion-controlled process. The b-value is 1, indicating an ideal surface capacitive-controlled kinetics process [41]. In our case, the b values are 0.757 and 0.821 at cathodic and anodic peaks, respectively, suggesting a capacitive-controlled mixed diffusion behavior (Fig. 4B). Then, the current is divided into the capacitive controlled and diffusion controlled contributions according to the following equation:

where k₁v and k₂v^1/2 represent the capacitive and diffusion contributions, respectively. We can obtain a series of k₁ (the linear plot slope) and k₂ (intercept) at certain scan rates. Fig. 4C shows the obtained CV curves at 0.8 mV/s where the red represents the capacitive-controlled region and the green represents the diffusion-controlled region. The capacitive-controlled contribution is 87.4% of the overall charge stored at the 0.8 mV/s. The capacity contribution at the other four scan rates was also calculated. As shown in Fig. 4D, the proportions of the capacity contribution are 65.8%, 70.9% and 79.7% at 0.1, 0.2, 0.4 mV/s, indicating a mainly capacitive-controlled process.

To understand the effects of annealing under air and Zn element on the LIB performance, Fe₂O₃@N-C-0.04, Fe₂O₃@N-C-0.06, Fe₂O₃@N-C-0.08, Fe₂O₃@N-C-0.12 and Fe₂O₃@N-C-0.24 were prepared by changing the amount of Zn(NO₃)₂.6H₂O and 2-methylimidazole, followed by annealing only under Ar at 700℃. Fig. S7A (Supporting information) shows the cycling stability of Fe₂O₃@N-C-0.04, Fe₂O₃@N-C-0.06, Fe₂O₃@N-C-0.08, Fe₂O₃@N-C-0.12 and Fe₂O₃@N-C-0.24 at a current density of 1.0 A/g. After 200 cycles, the specific capacity values of Fe₂O₃@N-C-0.04 and Fe₂O₃@N-C-0.06 were 464.2 mAh/g and 329.7 mAh/g, respectively. Similar to Fe₂O₃-700, these electrodes show a significant capacity decay. Fe₂O₃@N-C-0.08 exhibits a reversible capacity of 583.6 mAh/g at the first cycle and 705.3 mAh/g after 200 cycles without obvious capacity decay. The capacity curve of Fe₂O₃@N-C-0.12 is similar to Fe₂O₃@N-C-0.08, but the capacity is lower than Fe₂O₃@N-C-0.08. It is interesting to note that the specific capacity of Fe₂O₃@N-C-0.24 is only 208.6 mAh/g at the first cycle and 213.6 mAh/g after 185 cycles. Thus, it is reasonable to conclude that the significant capacity decay of Fe₂O₃@N-C-0.04 and Fe₂O₃@N-C-0.06 is mainly due to the insufficient carbon layers derived from ZIF-8. By contrast, excessive carbon layers of Fe₂O₃@N-C-0.24 hinder Li-ion diffusion, resulting in a lower specific capacity [42]. In addition to changing the amount of the precursor, annealing under air is another effective approach for controlling the thickness of the carbon layer. Therefore, we anneal Fe₂O₃@N-C-0.08 at 350 ℃ in the air to further reduce the thickness of the coated carbon layer. Fe₂O₃@N-C exhibits a reversible capacity of 942.8 mAh/g after 200 cycles, which is 273.8 mAh/g higher than that of Fe₂O₃-@N-C-0.08. It is important to note that the capacity of Fe₂O₃@N-C-0.24 only is 208.6 mAh/g at 1.0 A/g which is quite similar to that of the N-doped ZIF-8-derived carbon reported by Tai et al. [43]. However, after Fe₂O₃@N-C was treated in the air for 2 h, ZnO was formed. To investigate the effect of ZnO on the performance, we synthesized N-doped carbon by annealing ZIF-8 first under Ar atmosphere and then in air. Impressively, the specific capacity of N-C was 383.3 mAh/g after 200 cycles (Fig. S7B in Supporting information) which is much higher than that of Fe₂O₃@N-C-0.24.

In summary, the Fe₂O₃@N-C electrode was successfully fabricated by coating ZIF-8 on the surface of cube-like Fe₂O₃ followed by the carbonization at 700 ℃ for 2 h under Ar and then 350 ℃ for 2 h under air. A void was created between the Fe₂O₃ and N-C carbon shell during the carbonization process that alleviates the volume expansion of Fe₂O₃ during Li intercalation. As a result, the as-prepared Fe₂O₃@N-C anode possesses a high specific capacity (1064 mAh/g at 0.1 A/g), and stable cycle life (803.6 mAh/g at 1.0 A/g after 1100 cycles). Moreover, we found that carbonization under air enables the creation of new active species and reduces the thickness of the carbon shell, promoting LIB performance. This work provides a new method for the design and fabrication of core-shell electrodes for a variety of applications.

The authors declare that they have no known competing financial interests or personal relationships that could have appeared to influence the work reported in this paper.

The authors are grateful to the Program of Introducing Talents of Discipline to Universities (No. B18030), the National Natural Science Foundation of China (Nos. 91856124 and 21531005), the Natural Science Foundation of Tianjin City (No. 19JCZDJC37200), the National Postdoctoral Program for Innovative Talents (No. BX20190157) and the Postdoctoral Research Foundation of China (No. 2019M660979).

Supplementary material associated with this article can be found, in the online version, at doi:10.1016/j.cclet.2021.08.013