Lithium-ion batteries (LIBs) have been regarded as one of the most promising power sources in energy conversion and storage, due to their long cycle life and high energy density [1, 2]. However, the most widely used anode material, graphite, only possesses a limited capacity of 372 mAh/g [3, 4], which will not satisfy the requirements needed to power and develop next-generation electronic devices. Hence, it is one of the main challenges to design anode materials with high capacity, long cycling perfor-mance and offer a suitable voltage platform [5, 6].

To date, many materials such as carbon, metals, oxides, chalcogenides have been used as potential anode materials for LIBs [7-10]. Many nanomaterials have also attracted more attention due to their great mechanical, electrical properties and synergistic effects [11-15]. Among these materials, the metallic antimony (Sb) and Sb-based materials (oxides, chalcogenides and alloys) are of particular focus with the literature [16-20]. Prikhodchenko and co-workers designed an antimony sulfide coated reduced graphene oxide (rGO) that exceeded 720 mAh/g after 50 cycles at a current density of 250 mA/g [21]. Additionally, Park's group fabricated SnSb-CNT nanocomposite by reductive precipitation of metal chloride salts within a CNT suspension, which possessed a reversible capacity of 480 mAh/g after 50 cycles of the LIBs [22]. As one of the typical group V VI compounds, antimony selenide (Sb₂Se₃) is a p-type semiconductor with direct band-gap, and has widely been applied within optics, thermoelec-tric, electrochemical energy storage to name a few [23-25]. Regarding anode material for LIBs, 1 mol Sb₂Se₃ could accommo-date 12 mol Li⁺ during the lithiation/delithiation reaction, exhibit-ing a theoretical specific capacity of 670 mAh/g for LIBs. However, Sb₂Se₃ suffers some significant hindrances, such as aggregation, pulverization, loss of electrical contact due to the dramatic volume expansion during the lithiation/delithiation process. To solve the aforementioned problems, introduction of a carbon matrix allows for the Sb₂Se₃ to offer good performance, which can not only release the stress induced by the volume expansion of materials, but also alleviate agglomeration [26]. Although these promising results have already been obtained, further improvement can still be exhibited within the electrochemical performance.

In this work, we have reported Sb₂Se₃@C nanofibers (Sb₂Se₃@CNFs), synthesized by electrospinning method, as an anode for LIBs. One-dimensional (1D) nanofibers offer the advantage of continuous ion/electron channels, short diffusion distance and better contact between materials and current collectors, leading to an improved rate performance [27]. Additionally, these carbon nanofibers can prevent the huge volume changes during charge/discharge process as they provide a buffer layer, which enhances their overall cycling life [26]. Furthermore, the annealing temperatures were also optimized. Hence, when used as the anode for LIBs, the Sb₂Se₃@CNF annealed at 600 ℃ (Sb₂Se₃@CNF-600) could deliver remarkably high reversible capacities of 625 mAh/g at 100 mA/g after 100 cycles and 437 mAh/g at a large current density of 1.0 A/g after 500 cycles.

The synthesis process of the Sb₂Se₃@CNFs is as follows: 0.4936 g sodium antimonite (NaSbO₃ 3H₂O, Tianjin Heowns Biochemical Technology Co., Ltd., China) and 0.45 g polyacryloni-trile (PAN, Mw = 150, 000, Sigma-Aldrich Co., Ltd., USA) were dissolved in 4.5 g N, N-dimethylformamide (DMF) solvent at room temperature with vigorous stirring for 2 h. Then, 0.3948 g Se powder (Aladdin Industrial Corporation, Ltd., USA) was added into the solution and kept vigorous stirring for 12 h. The as-prepared precursor solution was transferred into a 10-mL syringe connected to a stainless steel needle. A high voltage power supply was used to provide a 12 kV high voltage for the resulting precursor. The distance was about 12 cm between the needle and the collector, and the feed speed rate was set up at 0.4 mL/h. After that, the prepared nanofibers were dried at 60 ℃ for 12 h in vacuum oven. Afterward, the fibers were annealed in a tube furnace under Ar atmosphere at 500 ℃, 600 ℃, 700 ℃ for 2 h with a heating rate of 3 ℃/min, respectively. These samples were marked as Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600 and Sb₂Se₃@CNF-700, respec-tively. The produced nanofibers were washed by distilled water and ethanol for several times after dipping in the acetic acid solution for 24 h. Finally, the Sb₂Se₃@CNFs were collected after drying overnight in the air. And the N-doped C nanofibers (NCNFs) were prepared according the same process except without sodium antimonite and selenium powder.

The X-ray diffraction (XRD, Rigaku D/MAX-2500) was utilized to determine the phase composition and purity of the as-prepared composites. The morphology and structure of Sb₂Se₃@CNFs were analyzed by scanning electron microscopy (SEM, Hitachi S4800) and transmission electron microscopy (TEM, JEOL 2010). To confirm Sb₂Se₃ content in the samples, the thermogravimetric analysis (TGA, Setaram 018124) of the samples was carried out in air atmosphere from 30 ℃ to 800 ℃ with a heating rate of 10 ℃/min. The surface area and pore width distribution of the sample were performed by Brunauer–Emmett–Teller (BET, Micro-meritics ASAP 2460) with the adsorption of N₂. X-ray photoelectron spectroscopy (XPS, Thermo Scientific Escalab 250Xi, USA) was used to research the element composition and chemical bonds of the samples.

Electrochemical performance was investigated by using CR-2025 coin-type cells. The tested coin cells were assembled in a pure argon gas glove box with the contents of H₂O and O₂ less than 0.5 ppm. And the working electrodes were prepared as following: Active material (80 wt%), acetylene carbon black (10 wt%) and carboxylmethyl cellulose (CMC, 10 wt%) were mixed together and dispersed in a mixed solvent of ethanol and distilled water, in order to form a uniform slurry. Subsequently, the mixed slurry was coated on copper foil and dried at 60 ℃ for 24 h. Generally, the average active material mass loading was controlled about 1.0 mg/cm². The counter electrode was Li metal. The electrolyte was 1.0 mol/L LiPF₆ dissolving in a 1:1 volumetric mixture of ethylene carbonate (EC) and dimethyl carbonate (DMC). Galvano-static charge–discharge testing was carried on a Neware battery testing system in a voltage window of 0.01–3 V. Cyclic voltam-metry measurements (CV) and electrochemical impedance spec-troscopy (EIS) were performed by using a CHI 660e electrochemical workstation.

The crystal structures of Sb₂Se₃@CNFs annealed at the different temperatures were characterized by XRD (Fig. 1a). It is obvious that all diffraction peaks of Sb₂Se₃@CNF-600 correspond well to the orthorhombic Sb₂Se₃ phase with the space group Pbnm (JCPDS No.15-0086) [18, 28]. The 2θ values of the main peaks are located at 15.02°, 16.88°, 23.9°, 27.38°, 28.20°, 31.16°, 32.22°, 33.10°, 34.06°, 35.68°, 38.8°, 41.3°, 45.06° and 51.88° which corresponded to the (020), (120), (130), (230), (221), (301), (311), (240), (321), (141), (250), (501) and (061) lattice planes of orthorhombic Sb₂Se₃, respectively. On the contrary, Sb₂Se₃@CNF-500 possesses other crystal phases. Among them, the diffraction peaks at 29.74°, 43.64° are associated with the (101), (102) lattice planes of Se (JCPDS No. 06-0362). Such lattice planes could be due to the separation of the reactants during the electrospinning process, and thus do not react well at 500 ℃, resulting in the existence of Se. The Sb₂Se₃@CNF-700 has no obvious diffraction peak for Sb₂Se₃, but only possesses a broad carbon peak (JCPDS No. 03-0401). This might be ascribed to the least Sb₂Se₃ content of Sb₂Se₃@CNF-700. And the crystallite size of Sb₂Se₃ can be calculated with the Scherer equation of D = kλ/(βcosθ), where θ is the Bragg angle in degrees; D is the size of crystallites; k is the shape factor of 0.89; λ is the X-ray wavelength of 1.54 Å and β is the fwhm of the reflection peaks [29]. The calculated average crystallite sizes are 26.81 and 22.08 nm for Sb₂Se₃@CNF-500 and Sb₂Se₃@CNF-600, respectively. The grain size is one of the important factors for lithium storage: the smaller the particle size benefits the mobility of Li ions [30]. Hence, the grain size of Sb₂Se₃@CNF-600 is better than Sb₂Se₃@CNF-500.

The Sb₂Se₃ contents in Sb₂Se₃@CNFs were carried out by TGA (Fig. 1b). The gradual decrease at the start of the curves (i.e., below 200 ℃) can be observed, due to the evaporation of water [31]. Upon the oxidation of Sb₂Se₃ to Sb₂O₄ and formation of SeO₂ and CO₂, the weights of the Sb₂Se₃@CNFs rapidly decreased over the tempera-ture range of 350–600 ℃. The oxidation reaction is: Sb₂Se₃ + 5O₂ = Sb₂O₄ + 3SeO₂↑ [7, 28]. From the result of TGA, the Sb₂Se₃ content of Sb₂Se₃@CNF-500/600/700 can be calculated as 33.43%, 21.64% and 18.24%, respectively. It is clear that the Sb₂Se₃ content of Sb₂Se₃@CNF-500/600/700 became less with the higher annealing temperatures.

To obtain the information about the elemental valences and compositions, the as-prepared products were characterized further by XPS, and results are depicted in Figs. 1c–f and Fig. S1 (Supporting information). The whole XPS spectra (Fig. 1c) reveals the existence of Sb, Se, C and N elements in all the samples tested. The high-resolution peaks of the Sb 3d in Fig. 1d can be divided into three significant peaks of 530.76 eV, 532.66 eV and 539.96 eV, respectively. Moreover, the peaks of 530.76 eV and 539.96 eV correspond to Sb 3d_5/2 and Sb 3d_3/2, respectively, indicating the presence of Sb³⁺ in the Sb₂Se₃@CNF-600. It should be noted that the peak of 532.66 eV is present due to a small account of surface oxidation of the Sb₂Se₃, relating to the O 1s [31, 32]. Presented in Fig. 1e, are peaks positioned at 55.79 eV and 58.98 eV that are related to Se 3d_3/2 and Se-O bond, meaning the presence of Se^2- in the Sb₂Se₃@CNF-600 nanofibers is evident [32, 33]. The high-resolution peaks of Sb 3d and Se 3d of Sb₂Se₃@CNF-500 and Sb₂Se₃@CNF-700 were displayed in Figs. S1a-e, exhibiting similar spectra to the Sb₂Se₃@CNF-600. Furthermore, the peaks of the C 1s in the Sb₂Se₃@C-500, Sb₂Se₃@C-600 and Sb₂Se₃@C-700 could be resolved into three binding energies, shown in Figs. S1g-i. The peak located at 284.77 eV, 286.36 eV and 288.88 eV correspond to the C—C, C—O and C＝O bond, respectively [7, 34]. In addition, the high-resolution N 1s spectra of Sb₂Se₃@CNF-600 in Fig. 1f can be attributed to pyridinic-N, pyrrolic-N and oxidic-N, which appears at ~398.26 eV, 400.05 eV and 403.23 eV, respectively. While the N 1s spectra of Sb₂Se₃@CNF-500 and Sb₂Se₃@CNF-700 are demonstrated in Figs. S1c and f, with the relative content of each nitrogen type being shown in Table S1 (Supporting information). It is proved that N-doped carbon materials have a higher capacity. First, because the doped N can offer more electrons to the π-conjugated system of carbon, enhance the electronic properties of neighboring carbon atoms, then strengthen the electronic conductivity of carbon nanofibers [35]. Secondly, the electronega-tivity of nitrogen is higher than carbon and then produces more binding sites for Li ions [36]. And that the carbon atoms around the nitrogen atoms have become more electronegative to adsorb more Li ions in these areas. On the other hand, pyridinic-N and pyrrolic-N create more open channels and defects in carbon due to N substitution for more Li⁺ insertion [37-39]. The high content of pyridinic-N and pyrrolic-N will therefore benefits the Li⁺ storage properties of the Sb₂Se₃@CNF. The overall content of pyridinic-N and pyrrolic-N within the Sb₂Se₃@CNF-600 is the highest, which can help Li⁺ storage in sample. Similarly, with the results of XRD and TGA, the characterization of XPS also indicates that 600 ℃ is the most appropriate annealing temperature.

The morphologies of the as-obtained Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600, Sb₂Se₃@CNF-700 and NCNFs were characterized by SEM, TEM and HR-TEM (Figs. 2a–f, Figs. S2 and S3 in Supporting information), respectively. The SEM images of Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600 and Sb₂Se₃@CNF-700 in Figs. 2a–c reveal the nanofiber-like morphology that displays individual and uniform nanofibers with a diameter of ~150-200 nm, which is also confirmed by the TEM images (Figs. S2d-f). As exhibited in the HR-TEM images (Figs. 2d–f), there are lattice fringes with distances of 0.312 nm and 0.283 nm for the Sb₂Se₃@CNF-500 and Sb₂Se₃@CNF-600 corresponding to the (2 1 1) plane and (2 2 1) plane of orthorhombic Sb₂Se₃ (JCPDS No. 15-0086), indicating that Sb₂Se₃ exists within the carbon nanofibers. The electron diffraction X-ray spectroscopy (EDS) spectra of Sb₂Se₃@CNF-600 in Fig. 2g demon-strate the C, N, Se, Sb elements distribution within the nanofibers. The flexible matrix consisting of 1D nanofibers offers the advantages of continuous ion/electron channels and short diffu-sion distances, and also prevents the huge volume changes during charge/discharge process acting as a buffering layer [27, 40]. Thus, the structure of Sb₂Se₃@CNFs will facilitate them to enhance both the rate performance and cycling life.

BET analysis was carried out to study the specific surface area of the Sb₂Se₃@CNFs. Fig. 3 and Table S2 (Supporting information) show the nitrogen adsorption–desorption isotherms, pore width and specific surface area of Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600 and Sb₂Se₃@CNF-700, which are about 4.62 m²/g, 6.07 m²/g and 7.58 m²/g, respectively. The more irreversible solid electrolyte interphase (SEI) is formed due to the larger specific surface area. In addition to this, higher pore width can lead to a higher irreversible Li storage and then present poor cycling performance [41]. The suitable specific surface area can make Sb₂Se₃@CNFs exhibit excellent performance for LIBs.

The electrochemical performances of the Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600 and Sb₂Se₃@CNF-700 as anode materials for LIBs were next investigated. The initial three cyclic voltammograms (CV) of the Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600 and Sb₂Se₃@CNF-700 over the voltage range of 0.001 V to 3.0 V (at a scan rate of 0.1 mV/s) are depicted in Figs. 4a–c. Upon the first cathodic scan (Fig. 4b), three peaks Sb₂Se₃@CNF-600 positioned at 1.37, 1.05 and 0.58 V are observed. The reduction peaks at 1.37 and 1.05 V could correspond to the Li⁺ intercalation into the Sb₂Se₃ host, to form Li_xSb₂Se₃ and the formation of Sb nanoparticles embedded within the Li₂Se, respectively. The peak located at 0.58 V is associated with the alloying reaction between Sb and Li⁺ [42]. Additionally, upon the first anode scan, three oxidation peaks of 1.06, 1.86 and 2.06 V are observed. The distinctive peak at 1.06 V corresponds to the Li₃Sb dealloying reaction, leading to the formation of Sb. Due to the conversion reaction between Li₂Se and Sb, in order to form Li_xSb₂Se₃, the small peak about 1.86 V was present. While the peak at 2.06 V is associated with the delithiation reaction of Li_xSb₂Se₃ and formation of Sb₂Se₃ [31]. In the following two cycles, the reduction peaks at about 1.9, 1.41 and 0.6 V were associated with the Li⁺ intercalation reaction, the reversible conversion reaction of Sb₂Se₃ and alloying reaction between Sb and Li⁺. They are different from the value of the first cycle, due to formation of the solid electrolyte interphase (SEI) film upon the first cycle. While the oxidation peaks in the subsequent two cycles are similar to that in the first cycle, presented at 1.05, 1.83 and 2.1 V. Meaning that the took place of dealloying reaction between Sb and Li⁺, the conversion reaction and Li⁺ deintercalation reaction. The electro-chemical reactions between Sb₂Se₃@CNF and lithium can be expressed as follows:

Intercalation/deintercalation:

Conversion reaction:

Alloying/dealloying reaction:

The charge/discharge profiles of Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600 and Sb₂Se₃@CNF-700 for the 1^st, 2^nd, 3^rd and 5^th cycle at a current density of 100 mA/g are displayed at in Figs. 4d–f. The charge–discharge voltage platforms of Sb₂Se₃@CNF can be consistent with the redox peaks of above CV curves. The first cycle of discharge plateaus at 1.38 V is related to the Li⁺ intercalation and 1.18 V is associated with conversion reaction of Sb₂Se₃. Moreover, the plateau at 0.87 V demonstrates the Li Sb alloying process. The initial discharge–charge capacities of the Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600 and Sb₂Se₃@CNF-700 are 1087.30/845.24, 952.32/724.21 and 783.02/589.62 mAh/g, respec-tively. Furthermore, the initial coulombic efficiency (CE) is 77.74%, 76.04% and 75.3%, respectively. The low CE may be resulted from the formation of SEI film and the irreversible reactions of electrode in the first cycle [19, 43]. According to BET analysis, specific surface area of Sb₂Se₃@CNF-700 is 7.58 m²/g, which is the largest in three samples. If the specific surface area is larger, the more solid electrolyte interphase (SEI) which is irreversible will form. So the initial coulombic efficiency (CE) of Sb₂Se₃@CNF-700 is lowest.

The cycling performance of Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600 and Sb₂Se₃@CNF-700 were shown in Fig. 5a. The discharge capacity of Sb₂Se₃@CNF-600 is ~625 mAh/g at a current density of 100 mA/g after 100 cycles, while the discharge capacities of Sb₂Se₃@CNF-500 and Sb₂Se₃@CNF-700 only remain 580 and 506 mAh/g, respectively. From the result of TGA, the Sb₂Se₃ content of Sb₂Se₃@CNF-600 can be calculated as 21.64%, and the carbon content is 78.36%. It is reported that the theoretical capacity of graphite and Sb₂Se₃ is 372 and 670 mAh/g. Hence, the theoretical capacity of Sb₂Se₃@CNF-600 based on the TGA results is calculated as 436.4 mAh/g. The 1D N-doped carbon nanofibers can interlink into a 3D conductive network, it not only enhances the electrons and Li⁺ ions diffusion kinetics, but also weakens the large volume fluctuation of Sb₂Se₃. So Sb₂Se₃@CNF-600 has the higher capacity than theoretical capacity. Fig. 5b and Fig. S4b (Supporting information) demonstrates the rate performance of Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600, Sb₂Se₃@CNF-700 and NCNFs, respectively. The average capacities of the Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600, Sb₂Se₃@CNF-700 and NCNFs, are 613/645/ 533/572, 560/638/492/429, 488/546/438/288, 460/520/420/320, 440/487/400/314 and 400/437/376/279 mAh/g at varying current densities of 0.1, 0.2, 0.5, 0.8, 1 and 2 A/g, respectively. When the current density returned to 100 mA/g, the discharge capacity of Sb₂Se₃@CNF-500, Sb₂Se₃@CNF-600, Sb₂Se₃@CNF-700 and NCNFs increased back to 600, 657, 524 and 463 mAh/g, respectively. These results indicate that the Sb₂Se₃@CNFs possesses good capacity retention upon increasing current density. However, beyond that, Sb₂Se₃@CNF-600 demonstrates an increased rate performance than the other annealing temperatures due to the high content of pyridinic-N and pyrrolic-N within the carbon matrix. Fig. 5c demonstrates that the Sb₂Se₃@CNF-600 exhibits a good long-term cycling performance, needed for futuristic electronic devices. When the current density was set to 1 A/g, the discharge capacity remained at 490 mAh/g after 500 cycles and the coulombic efficiency reached at 100%. While in Fig. S4c (Supporting information), NCFNs only remains at 386 mAh/g after 350 cycles with an applied current density of 1.0 A/g. From the SEM images (Fig. S5 in Supporting information) of the Sb₂Se₃@CNF-600 electrode after cycling, we can observe the obvious carbon fibers could be retained after cycling. It demonstrates that the Sb₂Se₃@CNF-600 electrodes have the better cycle performance and rate capability due to N-doped carbon nanofibers which can remit the huge volume changes. In addition to above results, compared with other Sb₂Se₃ and Sb₂Se₃/C composites using different synthetic methods (Table S3 in Supporting information) [18, 31, 42, 44-48], it is obvious that the Sb₂Se₃@CNF-600 have long cycling stability and remarkable rate performance when applied as a material within LIBs.

To further understand the resistance and charge transfer properties of the Sb₂Se₃@CNF electrodes, EIS measurements were carried out. The Nyquist plot reveals a depressed semicircle in the state of high frequency that is attributed to the impedance of SEI film and charge transfer resistance on the electrode/electrolyte interface, while the inclined line in low-frequency regions was related to the diffusion of Li⁺ within the materials [49]. As depicted in Fig. 5d, Sb₂Se₃@CNF-600 has a charge transfer resistance (R_ct) of 151 Ω, which is much smaller than that of Sb₂Se₃@CNF-500 (540.4 Ω) and Sb₂Se₃@CNF-700 (358.3 Ω) electrode. The smaller R_ct benefits the higher rate capability, which proves the excellent rate performance of Sb₂Se₃@CNF-600. The relationships between Z' with ω^-1/2 (ω = 2πf) in Fig. 5e shows that the slope of Sb₂Se₃@CNF-600 is the smallest than others. And the diffusion coefficients of Li⁺ in all electrodes can be calculated by the equation: D = 0.5R²T²/S²n⁴F⁴C²σ² [50]. Hence, the Li⁺ diffusion coefficients of Sb₂Se₃@CNF-500/600/700 are calculated to be 0.40×10^-16 cm²/s, 1.89×10^-16 cm²/s and 1.8×10^-16 cm²/s, respectively. It strongly manifests that the Sb₂Se₃@CNF-600 possesses excellent electrical conductivity and high Li⁺ mobility, which can enhance lithium storage performance [51, 52].

In summary, we have successfully fabricated Sb₂Se₃@CNF, via a novel electrospinning method, for utilization as anode materials within LIBs. The Sb₂Se₃@CNF-600 delivered a remarkably high discharge capacity of 625 mA h/g after 100 cycles and 490 mAh/g after 500 cycles at current density of 100 mA/g and 1.0 A/g, respectively. The N-doped carbon nanofibers contribute to the excellent electrochemical performance of the Sb₂Se₃ due to its roles within the conductive matrix and its ability to act as a buffer layer, preventing the destructive volume expansion of Sb₂Se₃. Therefore, Sb₂Se₃@CNF-600 can be regarded as a promising anode for LIBs in the future.

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.

This work was supported by the National Natural Science Foundation of China (No. 51302079) and the Natural Science Foundation of Hunan Province (No. 2017JJ1008).

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.cclet.2019.11.039.