1 Introduction

The rapid development of science and technology strongly relies on the support of energy. During the past decades, lithium-ion batteries (LIBs) have been widely applied in multiple fields as important energy storage devices due to their advantages of high energy density, high power density as well as high security. In recent years, due to the limited reserves and uneven distribution of lithium resources, exploring new energy techniques is urgently needed. Sodium-ion batteries (SIBs) have become one of the most popular candidates by virtue of their high natural abundance and low cost of Na, as well as the similar storage mechanism to LIBs [1]. However, the larger ionic radius of Na+ (0.102 nm) compared with Li+ (0.059 nm) generally leads to slow reaction kinetics and unsatisfied electrochemical performances [2]. In this case, it is challenging but desirable to explore suitable electrode materials with high performance for SIBs.

To date, numerous potential anode materials have been investigated, such as carbonaceous materials [3], alloy-based compounds [4], transition metal oxides [5], phosphides [6] and sulfides [7]. Among them, Sb-based materials with high capacities have attracted great attention, including Sb, Sb-based alloys, Sb2O3, Sb2O4 and Sb2S3 [8]. In particular, antimony trisulfide (Sb2S3), which is based on multielectron conversion/alloy reactions, shows attractive theoretical capacity (947 mAh·g−1), great thermal stability, abundant resource, popular price and better reversibility compared to oxides [9]. However, the problem of serious volume expansion and poor conductivity still exists in Sb2S3, which may lead to active material pulverization and slow charge transport, resulting in an inferior cycle and rate performance. Designing nanostructure and combining Sb2S3 with protective substance, which can reduce stress, relieve volume effect and improve conductivity effectively, are two common optimization approaches to address the above issues [10, 11]. In addition, a new method of fabricating heterostructure composite has been put forward recently [12]. The heterostructures with different phase compositions can effectively relieve internal pressure and promote interfacial charge transfer. For instance, Zhang et al. [13] fabricated Sb2S3/MoS2 composite, which exhibited better electrochemical properties in SIBs compared with individual Sb2S3 and MoS2 nanomaterials. Jia and co-workers [14] synthesized hierarchical Sb2S3/SnS2/C and found that the heterogeneous structure led to a built-in electric field, which decreased ion diffusion resistance and promoted ion and electron transport. CoS2/Sb2S3 nanocrystals encapsulated into a dual-carbon framework was reported by Li’s group [15], and an excellent sodium storage performance was obtained by the synergistic effect of the heterostructure. Moreover, Ke’s group [16] designed Sb2S3–Co9S8/NC heterogeneous material, which exhibited excellent cycle stability as well as fast lithium-ion transfer kinetics. Dong et al. [17] developed ZnS–Sb2S3@C core–double-shell material and obtained 630 mAh·g−1 at 100 mA·g−1 after 120 cycles in SIBs due to the improved electrochemical reaction kinetics and structure stability.

Although obvious improvements in electrochemical performance have been obtained in heterogeneous structure composites, most of the reported materials are consisted of two kinds of “sulfides”, which all have huge volume change and extremely low conductivity. To further optimize the performance of Sb2S3, incorporating alloy-type metallic Sb with sulfides seems to be a wise choice. Li et al. [18] synthesized Sb@Sb2S3@C, benefiting from the robust skeleton frameworks and high electronic conductivity from amorphous carbon nanotubes and metallic Sb, the anode exhibited a high specific capacity of 1596 mAh·g−1 at 0.1 A·g−1 and stable cycling performance in LIBs. Wu and his group [19] prepared Sb/Sb2S3@CHT, which exhibited excellent potassium-ion storage performance. These reports indicate that it is worth further investigating the combination of Sb and Sb2S3 because it is promising in achieving effective enhancement in the electrochemical performance.

In addition, the introduction of conductive matrixes is also significant for Sb-based materials to alleviate the volume expansion and increase conductivity, while most of the reported matrixes are carbonaceous materials at present. The nonpolar nature and the limited defect sites of carbon may lead to weak interface bonding force between active material and carbon, which results in an unstable interface and inefficient contact. The limited theoretical capacity of carbon may also bring a loss in the specific capacity of the whole composite. On this point, MXene (transition metal carbide or nitride), a new type of two-dimensional material, can be a better choice. Owing to the ultra-high conductivity (104 S·cm−1), excellent mechanical strength and hydrophilicity, low-diffusion-energy barrier and abundant reaction sites, MXenes have been extensively investigated in energy storage devices such as alkali metal-ion batteries and supercapacitors [20]. As a typical representative, Ti3C2Tx has a high theoretical capacity of 352 mAh·g−1 as the anode in SIBs. The rich terminal groups T (O–, OH– and F–) on the surface of Ti3C2Tx can induce a strong binding affinity to metal-based materials, which is beneficial for forming robust interfaces [21,22,23,24,25,26,27,28]. However, the restacking issue of pure MXene nanosheets affects the achievable sites for reversible charge/discharge, so it shows limited capacity when used as an electrode alone. And MXenes are commonly easy to be oxidized under a violent environment.

Based on the current situation, in this work, we integrated nanoengineering, heterogeneous structure design and stable conductive network together, to construct a kind of MXene/carbon double-confined Sb/Sb2S3 nanoheterostructure by a facile and highly effective one-step solvothermal followed by annealing process. The in situ formed nanoparticles of Sb and Sb2S3 can be homogeneously anchored on the Ti3C2Tx nanosheets covered by ultra-thin conductive carbon networks, and the grain growth of Sb-based materials can be well confined to avoid excessive particle size. The MXene with abundant functional groups not only serves as a flexible conductive matrix to buffer the volume change and accelerate the electron transfer but also improves the structure stability by establishing a solid interface through strong covalent bonding of Ti–O–Sb and S–Ti. The coated carbon network further enhances the electronic conductivity, promotes the charge transfer kinetics, buffers the volume expansion, maintains the structural integrity of the electrode and also protects Ti3C2Tx from oxidation and avoiding the parasitic reaction between Sb-based active anode and electrolyte. Furthermore, the heterogeneous interface of Sb2S3 and Sb can accelerate electron transport. The nanostructure is beneficial for both the efficient ion and electron transport and the wettability of electrolytes. In addition, the effect of the content of Sb-based materials on the morphology and performance was also investigated. Benefiting from the particular heterostructure and the synergistic effect between Sb-based heterogeneous nanostructure and carbon-coated MXene, the composite anode with the optimal Sb/Sb2S3 content delivers high charge/discharge capacities, excellent rate properties and ultra-stable cycle performances in SIBs and LIBs. The storage mechanism of the composite was revealed by ex situ X-ray diffraction (XRD). Cyclic voltammetry (CV) test proved that the pseudocapacitance behavior is dominant in the composite. This research puts forward simple and effective tactics to establish composite anode for highly efficient energy storage applications.

2 Experimental

2.1 Materials preparation

All the chemicals including LiF (99%), HCl (AR), ammonia solution (NH3·H2O, AR), SbCl3 (AR), Sb (99%), glucose (C6H12O6, AR) and thiourea (CH4N2S, ≥ 98%) were purchased from Aladdin Reagent Co., Ltd. The Ti3AlC2 (≥ 98%) powder was purchased from Foshan Xinxi Technology Co., Ltd.

2.2 Preparation of Ti3C2Tx

Firstly, 1.6 g LiF was dissolved in HCl solution (9.0 mol·L−1, 20 ml) and stirred for 10 min. Then 1 g Ti3AlC2 powder was slowly added to the mixed solution and kept stirring for 24 h at 40 °C to etch. The mixture was washed with 1 mol·L−1 HCl with a centrifuge to remove LiF, and then the mixture was washed with water until the pH of the supernatant reached 6. After that, the precipitate was added to deaerated water and sonicated for 2 h. At last, the solution was centrifuged for 60 min to separate a few-layer MXene. The concentration of the Ti3C2Tx MXene solution was 10.0 mg·ml−1.

2.3 Preparation of Sb/Sb2S3@Ti3C2Tx@C composite

75 mg SbCl3 and 5 ml Ti3C2Tx dispersion liquid were first added to 40 ml of ethylene glycol to thoroughly mix under stirring. And 0.5 ml HCl solution was added to the above solution to inhibit the hydrolysis of SbCl3. After stirring for 24 h, 50 mg glucose and 1.5 mmol thiocarbamide was added as carbon source and sulfur source and kept stirring for the other 2 h. Then the mixture was put in the Teflon-lined autoclave and kept at 160 °C for 12 h. After being washed with water by centrifuge several times, the precipitate was put in the freeze-drier to remove water. To completely carbonize the glucose and improve the crystallinity of the composite, the dried sample was calcined at 400 °C for 2 h and named as MX-Sb/Sb2S3@C-1. In addition, to investigate the effect of the load amount of Sb-based material on electrochemical performance, two other samples with different amounts of SbCl3 (37.5/112.5 mg) were also prepared and named as MX-Sb/Sb2S3@C-0.5 and MX-Sb/Sb2S3@C-1.5, respectively. For comparison, pure Sb2S3 was also synthesized under the same treatment process.

2.4 Material characterization

The phase structures of the fabricated materials were characterized by X-ray diffraction (XRD, Rigaku SmartLab) with a copper target (λKα = 0.15406 nm). The microstructure analysis was examined using field emission scanning electron microscopy (FESEM, ZEISS SUPRA55) and transmission electron microscopy (TEM, Talos F200X). Fourier transform infrared (FTIR) spectra were recorded on a spectrometer (Shimadzu FTIR-8400S) in ambient air at room temperature (25 °C). The surface element chemical valences of the materials were studied by X-ray photoelectron spectroscopy (XPS, Thermo Escalab 250XI). The specific surface areas and pore volumes of the samples were analyzed using N2 adsorption through the Brunauer–Emmett–Teller (BET) equation. Element composition was determined by inductively coupled plasma optical emission spectrometer (ICP-OES, Aglient 5110).

2.5 Electrochemical characterizations

The anode was composed of 80 wt% as-prepared composite, 10 wt% conductive acetylene black, 10 wt% Na-carboxymethyl cellulose (CMC) binder and deionized water as the solvent. Put the above slurry onto copper foil of 12 mm diameter and dry at 110 °C overnight. Coin cells (CR2032) were assembled in a glovebox with H2O < 0.1 × 10−6 and O2 < 0.1 × 10−6. For LIBs, the metal Li was used as the counter electrode and reference electrode. Monolayer polypropylene (Celgard2500) was used as the separator. LiPF6 dissolved in a mixture of ethylene methyl carbonate (EMC), ethylene carbonate (EC) and dimethyl carbonate (DMC) with a volume ratio of 1:1:1 was used as the electrolyte. For SIBs, Whatman GF/C microfiber glass acts as the separator, and 1 mol·L−1 NaClO4 in 1:1 (weight ratio) EC/DEC with 5% fluoroethylene carbonate (FEC) additive acts as the electrolyte.

The electrochemical performance was measured by the LAND battery test system. CV tests were recorded with CHI1000 electrochemical workstation ranging from 0.01 to 3.0 V at a scan rate of 0.1 mV·s−1. Electrochemical impedance spectra (EIS) were carried out in the EnergyLab XM electrochemical workstation with an amplitude of 5 mV in the frequency range of 1 × 105–1 × 10−2 Hz.

3 Results and discussion

3.1 Morphological and structural characterization

The schematic diagram of the synthesis procedure of the carbon-coated Sb/Sb2S3@Ti3C2Tx composite is shown in Fig. 1. Sb/Sb2S3 nanoparticles were in situ formed on the surface of MXene, which can not only confine the grown of Sb and Sb2S3, but also prevent their agglomeration. In turn, the uniformly distributed Sb-based particle can prevent MXene from collapsing and restacking. During the annealing process, the surface of Sb/Sb2S3@Ti3C2Tx would be covered by the ultra-thin carbon layer, which not only further relieves the volume expansion and increases the conductivity, but also protects MXene from oxidation, and refrains the direct contact between Sb-based material and electrolyte, thus decreasing the side reaction.

Fig. 1
figure 1

Synthesis flowchart of carbon-coated Sb/Sb2S3@Ti3C2Tx composite

Full size image

The morphology characterization of the as-prepared composites was measured by SEM. After etching from the massive Ti3AlC2 (Fig. S1), the few-layer MXene Ti3C2Tx was obtained, as observed in Fig. 2a. SEM images of MX-Sb/Sb2S3@C-1 in Fig. 2b, c show that a great number of Sb/Sb2S3 nanoparticles are uniformly distributed on the surface of lamellar Ti3C2Tx, which is covered by the ultra-thin carbon. The double confining effect of MXene and carbon can buffer the volume expansion of Sb/Sb2S3 more sufficiently and limit the particle growth simultaneously, which is due to the in situ fabrication process. According to the particle size distribution in Fig. S2, the average size of the nanoparticle is ~ 25 nm. TEM was further measured to survey the microstructure of MX-Sb/Sb2S3@C-1 composite. As shown in Figs. 2d, e, S3, the MXene with transparent lamellar structure is loaded with evenly dispersed Sb-based nanoparticles, which is consistent with SEM results. The HRTEM image in Fig. 2f displays two kinds of lattice fringes with an interplanar spacing of 0.23 and 0.34 nm, respectively, responding to the (104) lattice plane of Sb and (111) lattice plane of Sb2S3, which confirms the successful synthesis of the Sb/Sb2S3 heterostructure. Energy-dispersive spectroscopy (EDS) mapping of MX-Sb/Sb2S3@C-1 in Fig. 2g reveals the uniform distribution of Ti, C, Sb and S, and it can be seen that the elements of Sb and S are mainly gathered at the position of nanoparticles. The ICP-OES was measured to investigate the content of metal elements, and the content of Sb and Ti is 19.45 wt% and 45.76 wt%, respectively, which are close to EDS results (Fig. S4). To investigate the effect of the amount of SbCl3 on the morphology of the composite, MX-Sb/Sb2S3@C-0.5 and 1.5 were also measured by SEM. As shown in Fig. S5, there is no obvious change in the overall morphology except that the nanoparticles on the MXene grew denser with the amount of SbCl3 increased. In comparison, pure Sb2S3 shows serious aggregation and the particle sizes are larger than that of the composites (Fig. S6a). The commercial Sb shown in Fig. S6b is block.

Fig. 2
figure 2

SEM images of a Ti3C2Tx and b, c MX-Sb/Sb2S3@C-1; d, e TEM images, f HRTEM image and g EDS mappings of MX-Sb/Sb2S3@C-1

Full size image

XRD was adopted to analyze the crystal structure of the as-synthesized composites. As described in Fig. 3a, after etching with LiF, the peak of (008) and (105) in Ti3AlC2 (Fig. S7a) disappeared, indicating the formation of Ti3C2Tx. In addition, the peak of (002) in Ti3C2Tx shifts to a smaller angle (~ 6.69°) than that of Ti3AlC2, and the interlayer spacing is calculated to be 1.26 nm, which indicates that the lattice spacing increased significantly after etching. The three composites with different tin contents present similar XRD patterns, manifesting the same crystal structure. In these composites, except for the (002) peak of MXene, the diffraction peaks at 28.6°, 40°, 41.9° and 51.6° are responds to the (012), (104), (110) and (202) lattice planes of Sb (PDF No. 85-1322) [29]; and the peaks at 15.6°, 17.5°, 24.8° and 35.5° are accord with the (020), (120), (130) and (240) planes of Sb2S3 (PDF No. 42-1393) [30], demonstrating the successful formation of Sb/Sb2S3@MXene@C composite. No peak of carbon was observed due to the amorphous structure. The content ratio of Sb to Sb2S3 is ~ 3:1 in MX-Sb/Sb2S3@C-1 according to the peak intensity. The Rietveld refined XRD profile of MX-Sb/Sb2S3@C-1 is shown in Fig. S8, and the content of Sb and Sb2S3 estimated from the Rietveld refinement is 64.27 wt% and 35.73 wt% (Table S1). Furthermore, the XRD patterns of pure Sb2S3 and commercial Sb were also measured, which matched the standard card very well (Fig. S7b, c).

Fig. 3
figure 3

a XRD patterns and b FTIR spectra of all prepared anodes. c Raman spectra and d N2 absorption/desorption curves of Ti3C2Tx and MX-Sb/Sb2S3@C-0.5/1/1.5

Full size image

To further explore the structure of the as-prepared composites, FTIR spectra were investigated. As displayed in Fig. 3b, the peaks that appeared at 1110 and 1460 cm−1 in Ti3C2Tx are corresponding to O–H, and the peak at 1630 cm−1 is the characteristic peak of C=O [31, 32]. Compared with pure Ti3C2Tx, the peak intensities in the composites of MX-Sb/Sb2S3@C-0.5/1/1.5 were reduced, which is related to the introduction of Sb-based material and carbon. The peak at 418 cm−1 in MX-Sb/Sb2S3@C-0.5/1/1.5 is due to the presence of Sb and Sb2S3, and the peak at 660 cm−1 is attributed to the bond of Sb–S. It is worth noting that the peaks at 565 and 710–840 cm−1 exist only in the composite, which is related to the Ti–S and Ti–O–Sb bonds [33]. The FTIR test shows that the abundant oxygen-containing functional groups in Ti3C2Tx can bond with Sb/Sb2S3 nanoparticles via strong covalent bonds, which can improve the interface stability as well as accelerate the interfacial charge transfer. And the tight coupling of carbon and Ti3C2Tx can confine the nanoparticles adequately and prevent Ti3C2Tx from oxidation.

The Raman spectrum in Fig. 3c shows two kinds of characteristic peaks in Ti3C2Tx, the peaks at 199 and 720 cm−1 rise from the A1g vibrational mode, and the peaks at 286, 410 and 618 cm−1 belong to the Eg vibrational mode [34]. In the MX-Sb/Sb2S3@C-0.5/1/1.5 composites, the peaks at 1380 and 1570 cm−1 are ascribed to the D peak (disorder degree) and the G peak (graphitization degree) of carbon on the surface of the composites. In addition, the peak at 150 cm−1 is attributed to Sb and Sb2S3, which is consistent with that of the pure Sb and Sb2S3 in Fig. S9. The specific surface areas of the electrode materials were tested from N2 adsorption and desorption curves. As demonstrated in Fig. 3d, the specific surface area of pure Ti3C2Tx is 41.2 m2·g−1, which is lower than that of MX-Sb/Sb2S3@C-0.5/1/1.5 with the surface areas of 89.3, 82.3 and 81.3 m2·g−1, respectively. The higher surface area of the composite is related to the loading of Sb/Sb2S3 nanoparticles and the coated amorphous carbon. However, with the increase in the content of SbCl3, the surface area decreased slightly, which is related to the increased particle size as presented in the SEM images. Furthermore, the surface areas of pure Sb and Sb2S3 were measured to be 0.63 and 12.5 m2·g−1 (Fig. S10). The pore size distribution of pure Ti3C2Tx is mainly concentrated in 1–10 nm, while for the composites, in addition to 1–10 nm, they also show peaks in the range of 10–20 nm (Fig. S11). Commonly, the large surface area and abundant pore distribution are beneficial for providing abundant reaction active sites and promoting electrolyte infiltration.

To explore the surface chemical state of the composite, XPS was carried out on MX-Sb/Sb2S3@C-1. As shown in the full spectrum in Fig. S12, the characteristic peaks of Ti, Sb, S and C can be observed. The high-resolution Sb 3d spectrum of MX-Sb/Sb2S3@C-1 can be split into two peaks (Fig. 4a) including Sb 3d5/2 (529.8, 530.6 eV) and Sb 3d3/2 (539.4, 540.1 eV). The peaks at 529.8 and 539.4 eV respond to metal Sb, and the peaks at 530.6 and 540.1 eV are ascribed to Sb3+ (Sb2S3) [18]. The O 1s spectrum (531.9, 533.4 eV) demonstrates the existence of C–O–C and oxygen-containing functional groups [35]. The S 2p spectrum in Fig. 4b can be divided into three peaks. The peaks of S 2p3/2 (161.3 eV) and S 2p1/2 (162.8 eV) are responding to S2− in Sb2S3 [36]. The peak at 163.8 eV is due to the Ti–S bond between Ti3C2Tx and Sb2S3, which conforms to the FTIR results (Fig. 3b). The Ti 2p spectrum (Fig. 4c) includes (454.8, 460.6 eV), Ti–Cx (non-stoichiometric defects, x < 1, 455.8, 461.4 eV), Ti–S (457.1, 462.6 eV) and Ti–O–Sb (458.8, 464.5 eV) bonds [37]. In addition, the C 1s spectrum in Fig. 4d can be divided into four types. The peaks at 284.8 and 288.8 eV respond to C–C and O–C=O bonds [38]. The peaks at 286.6 and 281.5 eV are ascribed to C–O (C–O–Sb) and Ti–C bonds. XPS results confirm that the Sb/Sb2S3 nanoparticles can combine with the conductive substrate (MXene and carbon) via strong covalent bonds, which can not only confine the Sb/Sb2S3 nanoparticles to reduce the particles aggregation and relieve the volume expansion, and then improve the stability and integrity of the composite, but also enhance charge mobility and boost the rate performance.

Fig. 4
figure 4

XPS spectra of MX-Sb/Sb2S3@C-1: a Sb 3d, b S 2p, c Ti 2p and d C 1s

Full size image

3.2 Electrochemical characterization

The sodium-ion storage performances of the composites were examined using half batteries. CV curves of MX-Sb/Sb2S3@C-1 for the first 3 cycles under 0.1 mV·s−1 between 0.01 and 2.0 V are shown in Fig. 5a. There is an obvious difference between the curves of the first and the following cycles due to the activated process at the beginning. There are three reduction peaks at 1.2, 0.78 and 0.2 V for the first cycle. The peak at 1.2–0.78 V is ascribed to the conversion reaction of Sb2S3 as shown in Eq. (1) and the formation of the solid electrolyte films (SEI) on the surfaces of the electrode [39]. The peak below 0.2 V corresponds to the alloying reaction between Sb and Na+ (Reactions (2, 3)) as well as the insertion of Na+ into Ti3C2Tx and carbon. Accordingly, the oxidation peaks located at 0.85 and 1.61 V reflect the dealloying reaction and conversion reaction. The small oxidation peak around 2.1 V is derived from the multiple reactions of the interface between Sb-based nanoparticles and Ti3C2Tx (combined with a covalent bond like Ti–O–Sb) [28]. The curves of the second and third cycles overlapped quite well, indicating the excellent reversibility and stability of the electrode. This phenomenon also occurred in the differential capacitance curves of MX-Sb/Sb2S3@C-1 in Fig. 5b. The CV test and differential capacitance curve analysis were also carried out on pure Sb2S3 and commercial Sb as shown in Fig. S13. The shape of the curve of Sb2S3 is similar to that of MX-Sb/Sb2S3@C-1, including the conversion reaction at around 1.22 V and the following alloying reaction. For pure Sb, there is only the peak responding to the alloying reaction at 0.48 V. However, both pure Sb2S3 and pure Sb display inferior reversibility during the following cycles, which decreases the utilization of the electrodes.

$${\text{Sb}}_{2} {\text{S}}_{3} + 6{\text{Na}}^{ + } + 6{\text e}^{ - } \leftrightarrow 2{\text{Sb}} + 3{\text{Na}}_{2} {\text{S}}$$
(1)
$$2{\text{Sb}} + x{\text{Na}}^{ + } + x{\text e}^{ - } \leftrightarrow 2{\text{Na}}_{x} {\text{Sb}}$$
(2)
$$2{\text{Na}}_{x} {\text{Sb}} + \left( {3 - x} \right){\text{Na}}^{ + } + \left( {3 - x} \right){\text e}^{ - } \leftrightarrow 2{\text{Na}}_{3} {\text{Sb}}$$
(3)
Fig. 5
figure 5

a, b CV and differential capacitance curves of MX-Sb/Sb2S3@C-1 at 0.1 mV·s−1; c charge/discharge curves of MX-Sb/Sb2S3@C-1 at 0.1 A·g−1; electrochemical property of MX-Sb/Sb2S3@C-0.5/1/1.5: d cycle performance at 0.1 A·g−1; e rate property; f EIS spectra; g cycle property at 1.0 A·g−1

Full size image

The galvanostatic charge–discharge tests were executed on the electrode materials. Figure 5c reveals the charge–discharge curves of MX-Sb/Sb2S3@C-1 under the current density of 0.1 A·g−1. The initial discharge/charge capacities were 592/366 mAh·g−1 with an initial coulomb efficiency (ICE) of 62%. The high capacity profited from the unique structure, which provides efficient electronic transmission paths and promotes the ion diffusion. The irreversible capacity for the first cycle is ascribed to the side reaction between the electrode and electrolyte. After that, the capacity was stabilized at 355 mAh·g−1 with CE of 93% (the 3rd cycle). To investigate the effect of the content of tin on the sodium-ion storage, the cycle performances of MX-Sb/Sb2S3@C-0.5/1/1.5 composites under 0.1 A·g−1 were tested as exhibited in Fig. 5d. It can be seen that MX-Sb/Sb2S3@C-1.5 with the maximum loading of Sb/Sb2S3 shows the highest capacity of 382 mAh·g−1 for the first cycle among the three composites, but it decayed rapidly to only 204 mAh·g−1 after 200 cycles. Similarly, without the support of MXene and carbon substrate, the capacities of pure Sb and Sb2S3 descend like a cliff after around 20 cycles although they had high initial capacities. And pure Ti3C2Tx only delivers a capacity of 24 mAh·g−1 (Fig. S14a). By contrast, the other two composites demonstrate remarkable cycle stability without capacity decay after 200 cycles (177 mAh·g−1 for MX-Sb/Sb2S3@C-0.5 and 302 mAh·g−1 for MX-Sb/Sb2S3@C-1) benefitting from the double confine protection of MXene and carbon as well as the uniform distribution of nanoparticles, which can fully relieve the volume expansion of Sb-based materials and prevent aggregation. To further survey the long cycle performances, the electrodes were measured under 1 A·g−1. As displayed in Fig. 5g, MX-Sb/Sb2S3@C-1 composite displayed fairly outstanding stability with a reversible capacity of 250 mAh·g−1 even after 2500 cycles with a low attenuation rate of only 0.0228 mAh·g−1 for each cycle. It is much higher than that of MX-Sb/Sb2S3@C-0.5 with the residual capacity of 110 mAh·g−1 and MX-Sb/Sb2S3@C-1.5 of 120 mAh·g−1 after 1000 cycles. All the composites deliver a much higher capacity than pure Sb, Sb2S3 and MXene under large currents for long cycles as shown in Fig. S15. It can be concluded that not only the structure design but also the loading content affect the storage performance of the materials significantly.

Figure 5e displays the rate performances of MX-Sb/Sb2S3@C-0.5/1/1.5 from 0.1 to 5 A·g−1. MX-Sb/Sb2S3@C-1 delivers capacities of 332.4, 307.8, 278.2, 253.7, 240.1 and 225 mAh·g−1 under 0.1, 0.2, 0.5, 1.0, 1.5 and 2 A·g−1. As the current density increased to 5 A·g−1, it can still maintain the high capacity of 200.5 mAh·g−1. As the current density recovers to 0.1 A·g−1, the capacity can still return to 306.5 mAh·g−1. MX-Sb/Sb2S3@C-1 presents an obvious advantage compared with MX-Sb/Sb2S3@C-0.5 and MX-Sb/Sb2S3@C-1.5, and much better than pure Sb, Sb2S3 and Ti3C2Tx (Fig. S14b). It is also worth noting that the sodium storage property of MX-Sb/Sb2S3@C-1 is better than most reported Sb-based anode materials (Table S2). To further discuss the structure stability of MX-Sb/Sb2S3@C-1, SEM was carried out after 200 cycles. As displayed in Fig. S16, the surface of the electrode was still quite smooth after the cycle, and the nanoparticles were still firmly anchored on the surface of lamellar Ti3C2Tx, which is due to the strong interaction force between the Sb-based material, MXene and carbon by the covalence bond derive form the in situ formation process.

The electrochemical impedance spectroscopies of MX-Sb/Sb2S3@C-0.5/1/1.5 are shown in Fig. 5f, which were tested with a frequency range of 1×105–1×10−2 Hz at a perturbation AC potential of 5 mV. All the Nyquist curves contain a half circle in high-frequency region and a slash in low-frequency region. The slope of the slash represents the Na+ diffusion coefficient; the larger the slope, the faster the diffusion. The half cycle reflects the electrolyte transfer impedance (Rs) and charge transfer impedance (Rct). Table S3 shows the impedance fitting parameters of all the samples. Compared with pure Sb, Sb2S3 and Ti3C2Tx (Fig. S14c), all the composites show obvious decreased impedance, which confirms that the interfacial chemical bonding and heterostructure design can improve ion diffusion kinetics and accelerate the charge transfer.

Furthermore, EIS of MX-Sb/Sb2S3@C-1 electrode was measured at various states of charge (Fig. 6a). Rct decreased from 527 Ω (before discharging) to 233.4 Ω (discharged to 0.7 V), and it kept decreasing during the following discharging process on the whole with occasional rising, which may be due to that the formed SEI layer blocks the charge transfer (Table S4). During the charging process, the Rct decreased obviously and it turned to 80.62 Ω when charged to 3.0 V. The same tendency was also observed in the value of Rs. It can be seen that MX-Sb/Sb2S3@C-1 was activated gradually during the first lap, which helps reduce the polarization.

Fig. 6
figure 6

a Ex situ electrochemical impedance spectroscopy; b CV curves at different scan rates and (inset) lgi versus lgv plots; c capacity contribution ratio; d GITT curves; e ex situ XRD patterns of MX-Sb/Sb2S3@C-1

Full size image

CV curves of MX-Sb/Sb2S3@C-1 under different scan rates of 0.1, 0.2, 0.6, 0.8 and 1.0 mV·s−1 were tested to reveal the dynamic characteristic as shown in Fig. 6b. With the increase in the scan rate, the position of the peak shifts slightly, which is due to the polarization of the electrode. The peak current (i) and scan rate (v) conform to the exponential relation as described in Eq. (4),

$$i = av^{b}$$
(4)

where a and b are constants. The value of b can be calculated according to the functional relationship between lgi and lgv as shown in the inset in Fig. 6b. Commonly, the value of b is between 0.5 and 1.0. When the value of b approaches 0.5, the electrochemical behavior is mainly controlled by diffusion processes. When the value is close to 1.0, the surface capacitance process dominates. The values of b calculated from the oxidation peak and reduction peak are 0.85 and 0.9, manifesting that the electrochemical behavior of MX-Sb/Sb2S3@C-1 was controlled by surface pseudocapacitance. The capacitive contribution can be further quantified according to Eqs. (5, 6),

$${\text{lg}}i = {{b}}\text{lg}v + {\text{lg}}a$$
(5)
$$i\left( V \right) = i_{{{\text{cap}}}} + i_{{{\text{diff}}}} = k_{1} v + k_{2} v^{1/2}$$
(6)

where v is the scan rate, k1 and k2 are constants, icap is the current required to charge the double layer at the electrolyte interface and idiff is the current related to the slow diffusion-controlled process. The value of k1v reflects the contribution of pseudocapacitance control, and k2v1/2 represents the contribution of diffusion control. Figure 6c presents the capacitance contribution ratio of MX-Sb/Sb2S3@C-1 under different scan rates, showing that the capacitance contribution increased with the increase in the scan rate. The contribution ratio of capacitance raised to 88% under the scan rate of 1 mV·s−1, manifesting that the capacitance behavior plays an important role in sodium storage, especially under a large current density.

Moreover, galvanostatic intermittent titration technique (GITT) is a common method to evaluate the Na+ diffusion coefficient, which can visually reflect the diffusion kinetics of Na+ in the anode. According to Eq. (7),

$$D = \frac{4}{\uppi \tau }\left( {\frac{{m_{\rm B} V_{\rm M} }}{{M_{\rm B} S}}} \right)^{2} \left( {\frac{{\Delta E_{\rm s} }}{{\Delta E_{\tau } }}} \right)^{2}$$
(7)

where τ means the time of the constant current applied (30 min); mB, VM, MB and S represent the mass of active substance, molar volume, molar mass and area, respectively; ΔES and ΔEτ are the change in equilibrium potential after a current pulse and the transient potential changes throughout the pulse, respectively. Figure 6d displays GITT curve of MX-Sb/Sb2S3@C-1. It can be discovered that during the discharging/charging process, the diffusion coefficient shows a downward trend of fluctuation. While during the transient period from the discharging process to the charging process, the diffusion coefficient increased suddenly. Usually, if the diffusion coefficient during the discharging process is smaller than that in the charging procedure, it means that the process of sodium removal is easier than the process of sodium insertion. The average value of lgD of MX-Sb/Sb2S3@C-1 is calculated to be − 11. In comparison, the MX-Sb/Sb2S3@C-1 shows a higher diffusion coefficient than that of pure Sb and Sb2S3 (Fig. S17), which may due to the small size of the particles and the effect of MXene and carbon in the composite.

To in-depth understand the sodium storage mechanism of MX-Sb/Sb2S3@C-1, the ex situ XRD test was executed. As shown in Fig. 6e, all the characteristic peaks in MX-Sb/Sb2S3@C-1 can be observed under the open-circuit potential (including Sb, Sb2S3 and MXene). When the electrode was discharged to 1 V, the peak intensity of the Sb2S3 decreased greatly, indicating the conversion reaction procedure. When the anode was discharged to 0.2 V, a new peak at 33.5° appeared, which is ascribed to the (110) lattice plane of Na3Sb (PDF No. 740-1162), reflecting the alloying reaction process between Sb and Na+. And when the electrode was discharged to 0.01 V, the peak of Sb disappeared, and the peak intensity of Na3Sb increased significantly. In the meantime, the peak at 16.7° still exists, which is Na2S (PDF No. 47-0178) derived from the conversion reaction of Sb2S3 and Na+. During the charging state at 0.65 V, the peak of Na3Sb disappeared and the peak of Sb reappeared. This phenomenon confirms the reversible alloying reaction process. When the cell was charged to 3.0 V, the peak intensity of Sb decreased, meaning the partial conversion from Sb to Sb2S3. No characteristic peak of Sb2S3 was detected after charging, which may be related to the weak crystallinity of Sb2S3. Based on the above analysis, the sodium storage mechanism of MX-Sb/Sb2S3@C-1 includes the totally alloying reaction of Sb and the partially reversible conversional reaction of Sb2S3.

To fully explore the energy storage potential of the MX-Sb/Sb2S3@C-1, the lithium-ion storage performance was also investigated. The CV curves were measured from 0.01 to 3 V in LIBs. As observed in Fig. S18a, there are three obvious reduction peaks at 1.7, 1.36 and 0.64 V for the first cycle, responding to the conversion reaction of Sb2S3, the alloying reaction of Sb and the formation of SEI film [18]. The peak below 0.3 V is related to the insertion of Li+ in Ti3C2Tx and carbon. The oxidation peaks appeared at 1.13, 1.98 and 2.35 V, which presents the dealloying reaction \(2{\text{Li}}_{3} {\text{Sb}} \to 2{\text{Sb}} + 6{\text{Li}}^{ + } + 6{\text e}^{ - }\)) and the formation of Sb2S3\((2{\text{Sb}} + 3{\text{Li}}_{2} {\text{S}} \to {\text{Sb}}_{2} {\text{S}}_{3} + 6{\text{Li}}^{ + } + 6{\text e}^{ - } )\). The coincidence of curves for the following cycles manifests the perfect reversibility of the anode. The constant current charge and discharge curve of MX-Sb/Sb2S3@C-1 at 0.1 A·g−1 are shown in Fig. S18b. The initial discharge/charge capacities are as high as 735/479.5 mAh·g−1. It can still maintain 479.5 mAh·g−1 after 100 cycles at 0.1 A·g−1 (Fig. S18c) and 231 mAh·g−1 after 300 cycles at 1.0 A·g−1 (Fig. S18e). MX-Sb/Sb2S3@C-1 also shows excellent rate performance with the capacity of 464.9, 430.4, 395.7, 364.5 and 34.3 mAh·g−1 under 0.1, 0.2, 0.5, 1.0 and 1.5 A·g−1, respectively. When the current density returns to 0.1 A·g−1 after the large current charge/discharge, the electrode can recover to 452.6 mAh·g−1 (Fig. S18d). The outstanding lithium storage property of MX-Sb/Sb2S3@C-1 can be ascribed to the synergistic effect of MXene/carbon double-matrix and the heterogeneous Sb/Sb2S3 nanostructure design. The in situ preparation method enables the firm anchoring of Sb/Sb2S3 nanoparticles on the surface of MXene covered by carbon, and it can not only confine the overgrowth of Sb-based material, but also adequately buffer the volume expansion and improve the conductivity and then effectively improve the stability and kinetic reaction rate of the Sb-based materials.

4 Conclusion

The Sb/Sb2S3@Ti3C2Tx@C double-confined nanoheterostructure composites have been successfully fabricated by a facial one-step solvothermal method followed by low-temperature heat treatment. Ultra-small Sb and Sb2S3 nanoparticles are uniformly distributed on the surface of Ti3C2Tx nanosheets covered by ultra-thin carbon layers. The unique double-confined structure design can not only better accommodate the volume expansion and facilitate the rapid transport of ions and electrons, but also preserve the structural integrity due to the strong interfacial coupling (Ti–S, Ti–O–Sb). Moreover, the heterogeneous interface between Sb and Sb2S3 can further promote charge transfer. Benefits from the unique structure, MX-Sb/Sb2S3@C-1 with the optimal tin content exhibited the ultra-stable cycle performance with a reversible capacity of 250 mAh·g−1 at 1.0 A·g−1 after 2500 cycles and comparable rate property with 200.5 mAh·g−1 under 5 A·g−1 in SIBs. The composites also exhibited excellent lithium storage performances. The ex situ XRD revealed the storage mechanism of MX-Sb/Sb2S3@C-1 including the alloying reaction and conversional procedure. The pseudocapacitance test proved that the capacitive behavior is dominant in MX-Sb/Sb2S3@C-1 (1 mV·s−1, 88%). This work highlights the necessity of architectural engineering in realizing advanced anode materials.