1 Introduction

Today, with the exhaustion of fossil energy represented by petroleum and the increasing serious environmental pollution, there is an urgent demand for sustainable renewable energies [1,2,3,4,5,6,7]. Moreover, the advances in technology also require smaller and lighter devices with high power [8] such as supercapacitors (SC) and lithium-ion batteries (LIBs) [9,10,11,12,13]. LIBs play important roles in electronic devices because of their high energy density, long cycle life, and environmental friendliness. But the performances of LIBs such as theoretical capacity, charge–discharge voltage, and cycle stability are affected by the inherent characteristics of electrode materials dramatically [3, 14, 15]. At present, some materials including carbon-based materials [16], silicon-based materials [17], niobium-based materials (Mo3Nb14O44 and V3Nb17O50) [18, 19], and vanadium-based materials (Na2Ca(VO3)4 and VPO5) [20, 21] are popularly used as anode materials for LIBs. Graphite, the commonly used commercial anode material, only has a theoretical capacity of 372 mAh g−1 [22, 23]. Therefore, seeking novel anode materials with higher specific capacity, higher charge–discharge rate, and cycle stability is necessary to meet future market demands [24].

Transition metal selenides are attracting more and more attentions because of their conversion reaction mechanism [25,26,27,28,29] and relatively high storage capacity of lithium. Among them, CoSe2 has been paid special attention due to the rich content of Co and its low price, high specific capacity, and safety [30,31,32,33]. However, CoSe2 suffers from a large volume change during the charge–discharge process that leads to the rapid decay of specific capacity and poor cycle stability. Therefore, researchers developed several strategies to overcome the poor cycle stability problem of CoxSey by preparing CoxSey microspheres or nanosheets, or coating CoxSey with carbon material and introducing carbon base [34,35,36,37,38,39,40,41]. However, their cycle stabilities are still not satisfactory.

It is reported that the design of multilayer templates for anode materials could provide more active sites, which is conducive to the higher specific capacity [42, 43]. Therefore, two-dimensional (2D) materials have been widely explored and used in many fields [44,45,46,47,48,49]. MXenes, discovered in 2011 [50], are a relatively new member of 2D material family, including a class of transition metal carbonitrides. The 2D layer-structured MXenes are of great interest and are considered as the most promising energy storage material because of their unique physical and chemical properties, such as metal conductivity, hydrophilic surface, large and adjustable layer spacing, and excellent mechanical properties [51,52,53]. MXenes have a general formula Mn+1XnTx, where M is transition metal (e.g. Ti, Mo, V, Nb, and Ta), X represents C or N, and T is surface functional group (e.g. OH, F, Cl, or O) that are formed by violent reactions of transition metals with water or fluoride ions [54,55,56]. Up to now, more than 40 MXenes have been synthesized successfully [57], which are applied in the fields of energy storage [58], electromagnetic shielding [59, 60], catalysis [61], and sensors [62]. Among them, Ti3C2Tx is one of the most studied MXene, exhibiting excellent cycle stability and higher theoretical lithium ion storage capacity (up to 447.8 mAh g−1) [63,64,65,66].

Therefore, novel CoSe2/Ti3C2Tx composite with high specific capacity and excellent cycle stability can be expected if the advantages of both CoSe2 and Ti3C2Tx are combined. In this study, Ti3C2Tx was firstly prepared by hydrofluoric acid corrosion method, and then CoSe2 particles were grown on Ti3C2Tx via hydrothermal method. Ti3C2Tx, as a buffer layer of volume change, could effectively alleviate the volume change and aggregation of CoSe2 particles by the formation of strong Ti-Se-Co bonds. Meanwhile, Ti3C2Tx could provide more relevant electronic channels in the composite. The fabricated CoSe2/Ti3C2Tx composites were found to possess excellent electrochemical performances.

2 Experimental

2.1 Material synthesis

Firstly, 2 g of Ti3AlC2 (98%, Shanghai Macklin Biochemical Co., Ltd.) was slowly added to 20 ml of 40% HF acid solution, which was stirred at room temperature for 24 h. After centrifugation of the above solution at 5000 rpm for 10 min, the supernatant was poured out. Deionized water was added to the precipitation left in the centrifuge tube, which was shaken by hand to mix the precipitation and deionized water evenly, followed by centrifugation for several times to remove the HF solution. Then, the liquid was pumped, filtered, and vacuum dried to get Ti3C2Tx powder. After that, suitable amount of Ti3C2Tx powder, Co(NO3)2·6H2O (98.5%, Xilong Scientific Co., Ltd.), Se powder (analytically pure, Tianjin Kemiou Chemical Reagent Co., Ltd.) were mixed in an ethanol aqueous solution and stirred for 1 h. The molar ratios of CoSe2/Ti3C2Tx were designed to be 1:2.5, 1:2, 1:1.5, and the resulted CoSe2/Ti3C2Tx composites were denoted by sample CT1, CT2, and CT3, respectively. The solution was then poured into a hydrothermal autoclave that was held at 200 °C for 16 h. After being cooled to room temperature naturally, the solution was taken out, filtered with suction, and vacuum dried for 2 h. The CoSe2/Ti3C2Tx composites were finally prepared. Figure 1 illustrates the typical preparation process of CoSe2/Ti3C2Tx composite.

Fig. 1
figure 1

The sketch diagram for the typical preparation process of CoSe2/Ti3C2Tx composite

Full size image

2.2 Material characterization

The raw materials used and prepared composites were characterized by X-ray powder diffractometer (XRD, D8-Advance, Germany), scanning electron microscopy (SEM, S4800, Japan), Brunauer-Emmet-Teller method (BET, ASAP 2460, USA), transmission electron microscopy (TEM, FEI Talos F200X, USA) equipped with energy-dispersive X-ray spectrometer (EDS), and X-ray photoelectron spectroscopy (XPS, Thermo Escalab 250Xi, USA).

2.3 Half‑cell measurement

The electrochemical properties of CoSe2/Ti3C2Tx composites (CT1-CT3) were tested by assembling them into 2016 coin cells. Firstly, 80 wt% CoSe2/Ti3C2Tx composite (acting as active material), 10 wt% conductive carbon black, and 10 wt% polyvinylidene fluoride (PVDF, adhesive) were mixed. N-Methylpyrrolidone (NMP) was then added and stirred for 12 h to form a slurry. The slurry was coated onto a copper foil, which was dried in a vacuum drying oven at 120 °C for 12 h. After that, the copper foil loaded with active material was cut into circular electrodes with a diameter of 10 mm. Argon (99.99%)-filled glove box (O2 < 0.1 ppm, H2O < 0.1 ppm) was employed to assemble the coin cells. Lithium sheets were used as reference electrodes, and polyethylene/polypropylene film (PE/PP, Celgard 2325) was applied as the separator. A commercial electrolyte with 1 M LiPF6 dissolved in a mixture of ethyl carbonate, diethyl carbonate, and dimethyl carbonate (1:1:1 by volume) was used. Constant current charge–discharge test was performed on a battery performance tester (Neware CT-4000, China). Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) measurements were conducted by an electrochemical workstation (Zahner Zronach, Germany). All electrochemical tests were performed at a constant temperature of 25 °C.

3 Results and discussion

Figure 2 shows the XRD patterns of the raw material of Ti3AlC2 powder, synthesized Ti3C2Tx powder, and CoSe2/Ti3C2Tx composite (CT2). It can be seen that after etching, (104) diffraction peak of Ti3AlC2 phase (JCPDS # 52–0875) disappears, indicating that Ti3AlC2 phase is successfully changed into Ti3C2Tx phase [67, 68]. Moreover, the (002) peak shifts from 9.7 to 7.1°, hinting that the layer spacing increases in Ti3C2Tx phase. From the XRD patterns of CT2, it can be seen that CoSe2 phase (JCPDS # 53–0449) appears obviously in addition to the original Ti3C2Tx phase, which reveals the formation of CoSe2/Ti3C2Tx composite.

Fig. 2
figure 2

XRD patterns of Ti3AlC2, Ti3C2Tx, and CT2

Full size image

Figure 3 demonstrates the SEM images of Ti3C2Tx, CT1, CT2, and CT3. It can be seen that Ti3C2Tx exhibits an accordion-like structure (Fig. 3a), indicating that the Al layer in Ti3AlC2 has been removed. Figure 3b–d illustrate that CoSe2 particles are successfully grown on the surface of Ti3C2Tx after hydrothermal reaction. But only in CT2 there are an appropriate amount of CoSe2 particles in the Ti3C2Tx substrate (Fig. 3c), while few CoSe2 particles are formed in CT1 (Fig. 3b) and too much CoSe2 particles with an accumulation nature are formed in CT3 (Fig. 3d).

Fig. 3
figure 3

SEM images of samples a Ti3C2Tx, b CT1, c CT2, and d CT3

Full size image

Figure 4 shows the N2 adsorption/desorption isotherms of CT1-CT3. The BET-specific areas of CT1-CT3 are calculated to be 19.13, 23.04, and 29.15 m2 g−1, respectively. It can be seen that the specific surface area increases with the rise of CoSe2 content added. Large specific surface area of the CoSe2/Ti3C2Tx composite is believed to favor the provision of more active sites that facilitates the storage of lithium ions [69, 70].

Fig. 4
figure 4

Nitrogen adsorption/desorption isotherms of a CT1, b CT2, and c CT3. The BET-specific areas of d CT1-CT3

Full size image

In order to further study the structural characteristics of CoSe2/Ti3C2Tx composite, TEM characterization was carried out on sample CT2 and the results are shown in Fig. 5. CoSe2 particles grown evenly on the surface of Ti3C2Tx substrate can also be observed (Fig. 5a), and the average size of CoSe2 particles is about 240 nm (Fig. 5b). As shown in the corresponding high-resolution TEM (HRTEM) image (Fig. 5c), the lattice spacing between two neighboring fringes is estimated to be 0.928 nm, which corresponds to the (002) planes of Ti3C2Tx. The lattice spacing between two neighboring fringes of a CoSe2 particle are measured to be 0.190, 0.249, and 0.259 nm, which correspond to the (211), (120), and (111) planes of CoSe2, respectively (Fig. 5d). The elemental mapping images shown in Fig. 5e–i demonstrate the uniform distribution of C, Ti, Co, and Se elements, further revealing the successful preparation of CoSe2/Ti3C2Tx composite.

Fig. 5
figure 5

Low-magnification (a) and high-magnification (b) TEM images of CT2. HRTEM images of the part of Ti3C2Tx (c) and part of a CoSe2 particle (d). The corresponding elemental mapping images (e–i)

Full size image

The interaction between CoSe2 particles and Ti3C2Tx was studied by XPS. Figure 6a shows the Ti 2P spectra of Ti3C2Tx (top) and CT2 (bottom). The peak at 455.1 eV corresponds to Ti-C bond. In addition, two peaks are observed at 455.8 and 457.2 eV, which can be attributed to Ti2+ and Ti3+, respectively [71, 72]. Ti4+ also appears at 458.9 eV due to the surface oxidation of Ti3C2Tx during the preparation process [73]. The Co 2p spectra of CoSe2 (top) and CT2 (bottom) are illustrated in Fig. 6b. The two peaks located at 778.6 eV and 781.1 eV can be ascribed to Co3+ and Co2+, respectively [74]. Figure 6c shows the Se 3d spectra of CoSe2 (top) and CT2 (bottom), and the peak at 54.9 eV is attributed to Se2−. In Fig. 6a, the Ti2+ and Ti3+ peaks of Ti3C2Tx almost disappear after the formation of CoSe2/Ti3C2Tx composite, and the signal of Ti4+ peak is greatly enhanced. While the Co2+ peak of CoSe2/Ti3C2Tx composite gets stronger in comparison with that of CoSe2 (Fig. 6b). The valency of Se does not change significantly (Fig. 6c). Moreover, it can be found from Fig. 6a, b that electron transfer occurs between Ti and Co, Ti is connected with Se then with Co, and finally Ti-Se-Co bonds are formed.

Fig. 6
figure 6

XPS spectra of a Ti 2p, b Co 2p, and c Se 3d in Ti3C2Tx and CT2

Full size image

Figure 7a–c demonstrate the first four cycle CV curves of Ti3C2Tx, CoSe2, and CT2 at 0.2 mV s−1, respectively. An irreversible reduction peak appears in the first cycle, which can be attributed to the generation of SEI layer. However, this peak disappears in the subsequent cycles, hinting that a stable SEI layer is already formed after the first cycle. The curves of the second, third, and fourth cycles almost overlap (Fig. 7a). The reduction peak at ~1.134 V may be due to the chemical reaction between Ti3C2Tx and Li+, while the oxidation peak at ~0.876 V can be ascribed to the generation of Ti3C2Lix. In the first CV cycle of CoSe2 (Fig. 7b), the reduction peaks at about ~0.506 and ~1.138 V probably refer to the conversion of CoSe2 to Co metal and Li2Se, respectively. In addition, strong oxidation peaks at ~2.103 and ~1.126 V appear due to the formation of CoSe2 by the reaction between Co and Li2Se. The lithium storage mechanism of CoSe2 can be described by a two-stage reversible reaction: CoSe2 + Li+ + e− ⇋ LixCoSe2 and LixCoSe2 + e− ⇋ Co + Li2Se [75]. Moreover, compared with those in the first circle, the two reaction peaks at ~ 0.506 and ~ 1.138 V in the second circle shift to ~0.653 and ~1.394 V, respectively, which may be caused by the formation of SEI layer, decomposition of electrolyte, and microstructure change.

Fig. 7
figure 7

CV curves of a Ti3C2Tx, b CoSe2, and c CT2 at 0.2 mV s−1 in the first four cycles. d CV curves of CT2 at different rates. Charge–discharge curves of the first three cycles of e Ti3C2Tx, f CoSe2, and g CT2

Full size image

The electric potentials of the reduction peak and oxidation peak in the third and fourth cycles are similar to those in the second cycle, but the response current decreases significantly, hinting a poor reversibility of CoSe2 during the lithium-delithium process caused by the large volume change and comminution effect of CoSe2 electrode. The CV curve of CT2 is an approximate combination of CoSe2 and Ti3C2Tx CV curve, indicating an improvement of the reversibility in comparison with that of CoSe2, as is shown in Fig. 7c. And the curves in the second, third, and fourth cycles of CT2 almost overlap because Ti3C2Tx buffers the volume expansion of CoSe2 particles and the formation of Ti-Se-Co bond could prevent CoSe2 from dissolving into the electrolyte.

Figure 7d illustrates the CV curves of CT2 at scanning rates of 0.4, 0.6, 0.8, and 1.0 mV s−1. It can be observed that the redox potential difference does not change significantly with the increase of scanning rate, revealing an excellent rate capability and small electrode polarization of CT2. Figure 7e–g show the charge–discharge curves of Ti3C2Tx, CoSe2, and CT2 in the first three cycles at 0.1 A g−1, respectively, indicating the lower Coulombic efficiencies of Ti3C2Tx (52.39%), CoSe2 (55.03%), and CT2 (69.24%) that are caused by the irreversible redox reactions.

Figure 8a demonstrates the rate performance of Ti3C2Tx, CoSe2, and CT2-made electrodes at current densities of 0.03, 0.06, 0.15, 0.3, and 0.6 A g−1 after 10 cycles. When the current density returns from 0.6 A g−1 to the initial 0.03 A g−1, the specific capacities of Ti3C2Tx and CT2 nearly do not decay, revealing their good reversibilities. But the specific capacity of CoSe2 increases, indicating that CoSe2 electrode underwent more activation than other samples. In addition, it can be seen that as the current density increases, the specific capacities of all samples decrease gradually. That is because at high rates, a large number of electrolyte ions will be adsorbed on the interface between electrode and electrolyte, resulting in a rapid decrease of the concentration of electrolyte ions and thereby enhancing the concentration polarization, but the polarization voltage does not contribute any specific capacity. The rate performance of CoSe2 is poor, while Ti3C2Tx exhibits excellent rate performance. Thus, by combining Ti3C2Tx with CoSe2, the CoSe2/Ti3C2Tx composite (CT2) displays high electronic conductivity and excellent performance at high current density. Moreover, CT2 has the best rate performance because of its smaller polarization during redox process. The specific capacities of CT2 at current densities of 0.03, 0.06, 0.15, 0.3, and 0.6 A g−1 are 331.95, 310.90, 280.47, 262.22, and 220.79 mAh g−1, respectively.

Figure 8b shows the cyclic stability of Ti3C2Tx, CoSe2, CT1, CT2, and CT3 after 1000 cycles at a current density of 0.3 A g−1. It can be seen obviously that the cyclic stabilities of CT1-CT3 are higher than those of CoSe2 but lower than that of Ti3C2Tx. And after 1000 cycles, the specific capacities of Ti3C2Tx, CoSe2, CT1, CT2, and CT3 are 84.06, 39.09, 156.10, 210.80, and 232.01 mAh g−1, respectively. These results indicate that Ti3C2Tx can only load a certain amount of CoSe2 particles, and excessive CoSe2 particles will aggregate and undergo large volume change and comminution effect during the charge–discharge process, which may result in the formation of new surfaces and reactivation and finally the enhancement of specific capacity. However, CoSe2 may dissolve into the electrolyte later, causing a significant reduction of specific capacity. In addition, the excellent lithium storage performances of CT1-CT3 can be attributed to the stable layered structure of Ti3C2Tx which provides additional active sites and improves the electronic conductivity. The formation of Ti-Se-Co bonds between Ti3C2Tx and CoSe2 particles also helps to alleviate the volume change of CoSe2 during the cycling process.

Figure 9 shows the microscopic topography of CT2 after 1000 cycles at a current density of 0.3 A g−1. It can be seen that the sample as a whole still keeps a lamellar structure, indicating a good structural stability of the composite. In addition, the CoSe2 particle number seems to decrease, which may be caused by the dissolution of CoSe2 particles during the cycling process.

Fig. 8
figure 8

a Rate performance of Ti3C2Tx, CoSe2, and CT2. b Cycling stability of Ti3C2Tx, CoSe2, CT1, CT2, and CT3 at 0.3 A g−1 for 1000 cycles and the Coulombic efficiency of CT2

Full size image

Electrochemical impedance spectroscopy (EIS) reveals the electrochemical behavior of Ti3C2Tx, CoSe2, and CT2 samples, as is illustrated in Fig. 10. A typical EIS curve consists of a straight line representing the low-frequency region of the interfacial charge transfer impedance and a semicircular line representing the mid-frequency region of the lithium diffusion impedance [68]. Impedance data are described by an equivalent circuit fit in the inset. The circuit consists of electrolyte resistance (Rs1), charge transfer resistance (Rs2), constant phase element (CPE), and Warburg impedance (Zw) [76]. Obviously, the semicircle diameter of CT2 is smaller than those of CoSe2 and Ti3C2Tx; therefore, CT2 can provide more active sites and reduce its resistance, which benefit the electron transfer.

Fig. 9
figure 9

SEM images of CT2 after 1000 cycles

Full size image
Fig. 10
figure 10

Nyquist plots of Ti3C2Tx, CoSe2, and CT2

Full size image

The electrochemical properties of CT2 are compared with those of other Ti3C2Tx- and CoxSey-based anodes reported, as are listed in Table 1. It can be seen that CT2 possesses higher specific capacity than several Ti3C2Tx-based anodes and higher stability than several CoxSey anodes. It is believed that the reactive anode material of CoSe2 mainly contributes to the higher specific capacity of CT2, and the layered structure of CT2 also favors the improvement of specific capacity. On the other hand, the formation of Ti-Se-Co bonds between Ti3C2Tx and CoSe2 leads to a better structural stability. Therefore, the excellent cyclical stability of Ti3C2Tx and high specific capacity of CoSe2 enable CT2 to exhibit excellent electrochemical performances.

Table 1 The comparison of electrochemical properties between the CT2 and those related materials reported
Full size table

4 Conclusions

In summary, novel CoSe2/Ti3C2Tx composites are successfully prepared with high specific capacity and excellent cycling stability. After 1000 cycles at 0.3 A g−1, a high specific capacity of 210.80 mAh g−1 is achieved. The introduction of Ti3C2Tx is believed to ease the volume expansion of CoSe2 during charge–discharge cycles, resulting in the improvement of cycling stability. Moreover, the excellent electronic conductivity of Ti3C2Tx favors the electrochemical behavior of CoSe2/Ti3C2Tx composites. Therefore, CoSe2/Ti3C2Tx composites with high specific capacity and excellent cycling stability are achieved, which may find promising applications in LIBs as anode materials.