Various renewable and clean energy sources, such as solar and wind, are growing rapidly due to the imminent depletion of non-renewable fossil fuel resources and the increasingly serious environmental pollution problems [1, 2]. However, electric energy generated by renewable energy sources cannot be directly transmitted to the grid due to its intermittent characteristics [3]. It is urgent to develop an energy storage device with high capacity. SIBs are considered to be the promising electric energy storage (EES) systems in the field of energy storage power stations due to their rich sodium resources, low cost and electrochemical behavior similar to LIBs [4-6]. The key to commercialization of SIBs is to find electrode materials with low cost and high specific capacity [7]. Nowadays, many SIBs anode materials have been developed, including hard carbon, transition metal oxides and intermetallic compounds [4]. However, the drawbacks of these materials are noticeable. For example, the drawback of hard carbon lies in low electronic conductivity and poor initial coulombic efficiency [8-10]. The transition metal oxides and intermetallic suffer massive volume changes during the reaction with the Na⁺ ions, and some of them are expensive or toxic [11-16]. Therefore, it is imperative to develop new types of SIBs anode materials with high specific capacity, low cost, simple synthesis process and ecological friendliness.

Metal borates have attracted more and more attention as a promising anode for SIBs in the past few years owing to their low atomic weight and high electronegative features [ 17-19]. To date, metal borates have been used as anode materials for sodium-ion batteries, but its sodium storage mechanism is still unclear [20-22]. Recently, Tian et al. evaluated the electrochemical performance of Fe₃BO₆ as a SIBs anode material and further explored its sodium storage mechanism [23]. Based on the ex-situ XRD analysis results, they reported that Fe₃BO₆ irreversibly transformed into crystalline Fe₂O₃ and then into crystalline Fe metal during the initial discharge process. In the subsequent charge and discharge process, crystalline Fe metal and crystalline Fe₂O₃ alternately appear. Namely, the sodium storage mechanism of Fe₃BO₆ is a conversion reaction between crystalline Fe₂O₃ (Fe³⁺) and crystalline Fe metal (Fe⁰). Wang et al. proposed the reversible conversion reaction mechanism of Zn₃B₂O₆ (crystalline) with sodium ion via air-insulated ex-situ characterizations: Zn₃B₂O₆ (amorphous) + 6Na⁺ + 6e⁻ ⇋ 3Zn (amorphous) + B₂O₃∙3Na₂O [18]. According to the ex-situ characterization results, Zn₃B₂O₆ was reversibly transformed into amorphous Zn metal during the first discharge, and amorphous Zn₃B₂O₆ was formed in the subsequent charging process. Although some metal borate materials have been studied, there is no consensus on their sodium storage mechanism. Therefore, it is necessary to determine the sodium storage mechanism of metal borate and develop new metal borate materials.

In this work, a low-cost and environmentally friendly FeBO₃ was prepared and evaluated as anode for SIBs for the first time. Electrochemical test results show that the FeBO₃ electrode has an initial capacity of 684 mAh/g (at 0.4 A/g) and a reversible capacity of 328 mAh/g. Besides, the phase evolution of FeBO₃ during charging and discharging process was revealed by in-situ XRD analysis. And the changes in the chemical composition of FeBO₃ in the sodiation/desodiation process were monitored by an air-insulated ex-situ XPS. Based on the in-situ XRD and ex-situ analysis, a new metal borate sodium storage mechanism was proposed.

As shown in Fig. 1a, all peaks of the XRD pattern of the sintered product are well indexed with the hexagonal crystal FeBO₃ (JCPDS No. 76-0701), which is composed of FeO₆ octahedrons and BO₃ plane triangles (Fig. 1b), indicating that a pure phase material was synthesized. The chemical composition and element valence of the FeBO₃ material were measured by XPS. It can be proved from the survey spectrum that Fe, B and O elements exist in FeBO₃ material (Fig. 1c). XPS spectrum for Fe 2 p consists of two peaks at 711.4 eV and 725.3 eV, and a satellite peak at 719.8 eV, which is characteristic of Fe³⁺ (Fig. 1d) [24, 25]. As shown in Fig. 1e, there is a strong peak at 191.4 eV, which is the characteristic peak of B 1s [26, 27]. As for the O 1s peak, it can be divided into two peaks at 530.5 and 531.5 eV, which are attributed to Fe-O and B-O of FeBO₃, respectively (Fig. 1f) [28, 29].

The morphology and detailed crystal structure information of FeBO₃ were explored through advanced FE-SEM and TEM. As shown in Figs. 2a and b, the morphology of FeBO₃ is a particle with a size of 50 − 300 nm. Further, the size of FeBO₃ particles is confirmed by the TEM image (Fig. 2c). In addition, the HRTEM image exhibits a 0.35 nm d-spacing, which is attributed to the (012) plane of FeBO₃ and agrees well with the XRD result (Fig. 2d). Moreover, the elemental mapping images (Fig. 2e) reveal the uniform distribution of Fe, B and O all over the single FeBO₃ particle.

The CV curves for the FeBO₃ electrode are measured at scan rate of 0.1 mV/s between 0.01 V and 3.0 V (Fig. 3a). A sharp and intense reduction peak appeared at 0.24 V during the first cathodic scan and then shifted to 0.62 V in the following scans. During the anodic polarization process, one peak is observed at 1.32 V, and no obvious difference is noticed in the subsequent cycles. The result indicates that the FeBO₃ undergoes irreversible reaction during the first cathodic process and then keeps stable in the following cycles. The initial discharge profile shows a plateau at about 0.4 V, and the plateau shifted to 0.76 V in the subsequent cycles (Fig. 3b). This result is in agreement with the CV curves. In the case of cycle performance, the initial discharge/charge specific capacity of the FeBO₃ electrode is 684/382 mAh/g, while the capacity decays to 78/74 mAh/g in the 100^th cycle (Fig. 3c). The rate performance of the FeBO₃ electrode is shown in Fig. 3d. The FeBO₃ electrode exhibits discharge capacities of 390, 291, 198 and 107 mAh/g at current densities of 0.1, 0.4, 0.8 and 1.6 A/g, respectively. The FeBO₃ has obvious capacity decay upon cycling. So it is necessary to improve its cycle stability.

Carbon coating has proven to be very effective in improving cycle stability [30-32]. Most metal borates are synthesized in air at a temperature of > 800 ℃. Under these conditions, carbon would be oxidized to carbon dioxide and escape. If calcined in an inert gas, the pyrolysis of the carbon source would create a reducing atmosphere, which can change the valence state of the transition metal in the metal borate or even destroy the crystal structure. As a cheap and non-toxic raw material, oleic acid has a carboxyl group that can produce a strong bond with metal ions, so it is suitable as a carbon source for coating [33]. In addition, transition metals can catalyze the graphitization rate of oleic carbon during the sintering process and oleic carbon with good conductivity can be obtained even at low temperatures [34]. Therefore, oleic acid was used as carbon source for preparing carbon-coated FeBO₃ (FeBO₃@C). The structure of the FeBO₃@C material remains unchanged after heat treatment at 500 ℃ (Fig. S1 in Supporting information). As shown in Fig. 3c, the FeBO₃@C shows a higher initial discharge/charge capacities of 719/426 mAh/g and a higher retained capacities of 332/331 mAh/g in the 100^th cycle. In addition, the FeBO₃@C displays discharge capacities of 425, 349, 321 and 272 mAh/g, at current densities of 0.1, 0.4, 0.8 and 1.6 A/g, respectively (Fig. 3d). Compared with FeBO₃, the cycle performance and rate performance of FeBO₃@C composite have been significantly improved.

In order to explore the reasons for the improved electrochemical performance, the EIS spectra of FeBO₃ and FeBO₃@C electrodes were measured. As shown in Figs. S2a and b (Supporting information), the Nyquist diagrams consist of two semicircles and a diagonal line, which can be described as solid electrolyte interface layer resistance (R_SEI), charge transfer resistance (R_ct), and Weber impedance. In addition, the R_ct and R_SEI data of FeBO₃ and FeBO₃@C calculated using the equivalent circuit (Figs. S2a and b inset) are shown in Table S1 (Supporting information). The values R_SEI of FeBO₃@C are lower than that of FeBO₃ owing to the less sodium-ion consumption during the formation of the SEI process, which explains the high coulomb efficiency and less irreversible capacity of FeBO₃@C in the first cycle [35]. Comparing the R_SEI change from 50 cycles to 100 cycles, it can be seen that the R_SEI rate of the FeBO₃ electrode is much greater than that of FeBO₃@C. This indicates that more SEI is continuously generated upon the FeBO₃ electrode during the sodiation/desodiation process, resulting in low capacity and poor stability. Simultaneously, the R_ct value of the FeBO₃ electrode increased from 27.4 Ω to 23945.3 Ω, but the R_ct value of the FeBO₃@C electrode only increased from 17.3 Ω to 48.5 Ω. The results show that the carbon coating can reduce the charge transfer resistance. The improved cycle and rate performance of the FeBO₃@C electrode composite can be ascribed to the enhancement in rapid charge transfer kinetics with the effect of C.

In addition, the reason for the improvement of electrode stability was studied by observing the evolution of the surface morphology for FeBO₃ and FeBO₃@C electrodes under different conditions. Both the fresh FeBO₃ and FeBO₃@C electrodes have smooth surface morphology (Figs. S3a and d in Supporting information). However, after 50 cycles, the FeBO₃ electrode shows some cracks on the electrode surface (Fig. S3b in Supporting information), whereas the FeBO₃@C electrode still maintains a crack-free surface (Fig. S3e in Supporting information). After continuous sodiation/desodiation for 100 cycles, the FeBO₃@C electrode still shows a smooth surface, while the FeBO₃ electrode shows a lot of pits and cracks (Figs. S3 c and f in Supporting information). The carbon coating with oleic acid as the carbon source can reduce the side reaction between the electrode and the electrolyte, thereby maintaining the stability of the electrode [36].

In order to evaluate whether carbon coating affects the reaction kinetics of FeBO₃ anode materials, the galvanostatic intermittent titration technique (GITT) was used to determine the sodium ion diffusion coefficient (Fig. 4a and Fig. S4a in Supporting information). The diffusion coefficient can be calculated by Eq. 1 [37, 38]:

(1)

Where τ, m_B, M_B, V_M, S, ΔE_S and ΔE correspond to relaxation time, active material mass, molar mass, molar volume, electrode/electrolyte contact area, voltage change caused by pulse and constant current charging (discharging), respectively. As shown in Fig. 4b and Fig. S4b (Supporting information), the calculated Na ion diffusion coefficient (D_Na) of FeBO₃@C is higher than that of the FeBO₃ electrode. The reason for this result may be that carbon coating reduces the repetitive formation of the SEI film that hinders ion diffusion [39].

In addition, to explore the charge storage mechanism of FeBO₃, pseudocapacitance analysis was performed through CV measurements at different sweep rates (0.2 − 0.6 mV/s) (Fig. 4c). The peak current (i) and sweep rate (V) of the CV curves follow the Eq. 2 below:

(2)

where a, b are adjustable values [40]. Generally, b=0.5 corresponds to a diffusion control process, b=0.5–1 is a common control process of diffusion and capacitance, and b=1 corresponds to a capacitance control process [41]. The value of b can be obtained by Eq. 3 below:

(3)

where b is the slope. As shown in Fig. 4d, the calculated b values for peak 1 and peak 2 are 0.556 and 0.671, which indicates that the kinetics of the charge storage process of the FeBO₃ electrode is controlled by diffusion and capacitance. In addition, the capacitance contribution at a given scan rate can be determined according to Eq. 4:

(4)

where k₁v corresponds to the capacitance control contribution, and corresponds to the diffusion control contribution, respectively. The capacitive-controlled contribution was calculated to be 56% for FeBO₃ at 0.6 mV/s (Fig. 4e). In addition, as shown in Fig. 4f, the ratio of capacitance control gradually increases with the scan rate increases. The results show that the diffusion process of the FeBO₃ electrode has a great influence on the charge storage process.

In-situ XRD test was performed to study the structural transformation and sodium ion storage mechanism of FeBO₃ during the sodiation/desodiation process. The obtained in-situ XRD patterns and corresponding intensity contour maps are shown in Figs. 5a and b. For the pristine electrode, all peaks are indexed to FeBO₃ with the hexagonal structure, except for the background peaks (Al foil, Be and BeO). Details information of the structural evolutions during cycling were monitored via the shifts/changes of selected characteristic peaks of 26°, 34°, 47° and 55° (Fig. 5a). The intensity of FeBO₃ peaks decreased gradually with the initial discharge process. When the cell was discharged to ~0.46 V, FeBO₃ peaks were fully disappeared and no peak reappearance was observed in the following charge process. This indicates that the crystalline FeBO₃ transformed into an non-crystal (nc) phase during the first sodiation process.

Since the FeBO₃ electrode changes to the amorphous phase after the first discharge, it is difficult to determine the phase evolution during the cycle by comparing XRD patterns. Therefore, an ex-situ XPS test was performed to study the sodium ion storage mechanism of FeBO₃ at fresh and different discharge/charge state in the second cycle. The results are shown in Figs. 5c and d. In the pristine state, the Fe 2 p peaks of the electrode appear at 711.5 and 725.3 eV with the two satellite peaks appear at 718.2 and 733.3 eV, which are characteristic peaks of Fe³⁺ [24, 25, 42]. When discharge to 0.01 V and then charge to 3 V, the characteristic peaks of Fe³⁺ does not appear. However, the Fe 2 p peaks of the electrode appear at 709 eV and 722.5 eV with the shaking satellite peaks appear at 713 eV and 729 eV, which indicates that the valence state of Fe is + 2 instead of + 3 (Fig. 5d). During the second sodiation process, Fe²⁺ partially transformed into Fe⁰, while the Fe⁰ can completely return to Fe²⁺ in the desodiation process, Fe²⁺ exist at all charging and discharging states. The results show that the electrode experienced a partially reversible process in the second cycle. The fine spectrum of B 1s was shown in Fig. 5e. It can be seen that the peak of B moves in the second cycle and finally returns to the 3 V state, indicating that boron participates in the reaction and the process is reversible. Based on the above in-situ XRD and ex-situ XPS test results, it can be concluded that the possible sodium storage mechanism of FeBO₃ is:

The first cycle:

Discharge:

(5)

Charge:

(6)

The subsequent cycle:

(7)

In summary, FeBO₃ was prepared and evaluated as a SIBs anode for the first time. The FeBO₃ electrode showed a reversible capacity of 328 mAh/g at a current density of 0.4 A/g. The cycle stability and rate performance of FeBO₃ were effectively improved by carbon coating. The prepared FeBO₃@C composite had a reversible capacity of 426 mAh/g (at 0.4 A/g) and an outstanding rate capability of 272 mAh/g (1.6 A/g). The results of EIS and GITT show that carbon coating can improve electrochemical performance by reducing the generation of side reactions, avoiding the excessive formation of SEI films and improving charge transfer kinetics. Furthermore, in-situ XRD and ex-situ XPS analysis demonstrate that FeBO₃ material transforms into an amorphous phase in the first sodiation/desodiation process. It was indicated that Fe³⁺ is irreversibly transformed into Fe²⁺. During the second sodiation process, Fe²⁺ partially transformed into Fe⁰, while the Fe⁰ can completely return to Fe²⁺ in the second desodiation process. The above results show that FeBO₃ is a new potential anode material for SIBs.

Declaration of competing interest

The authors report no declarations of interest.

Acknowledgments

This work was supported by the National Natural Science Foundation of China (Nos. 21673136 and 22075173) and the Science and Technology Commission of Shanghai Municipality (No. 19DZ2271100).

Appendix A. Supplementary data

Supplementary material related to this article can be found, in the online version, at doi:https://doi.org/10.1016/j.cclet.2021.03.014.