1 Introduction

The rapid growing demand for electric vehicles and cheap large-scale energy grids inspire the enthusiasm to explore the next-generation energy storage systems beyond lithium-ion batteries [1,2,3]. Lithium-sulfur batteries (LSBs) promise significant advantages with respect to high theoretical capacity (1672 mAh g−1), prominent energy density (2600 Wh kg−1), low cost of sulfur and environmental friendliness [4, 5]. However, the serious shuttle effect caused by soluble intermediate lithium polysulfides (LiPSs) dissolved in liquid electrolytes, the insulating nature of sulfur and lithium sulfide (Li2S), and the sluggish sulfur redox kinetics restrict the actual performance of Li–S batteries [6, 7].

In the past decade, most studies have focused on to improve the adsorptive ability to tackle the shuttle effect, including physical confinement (e.g., porous carbon [8] and MOF [9, 10]), and chemical interactions (e.g., metal oxides [11, 12], metal sulfides [13], and metal carbides [14, 33, 34]. In recent years, 2D/2D heterostructures constituted via the face-to-face interaction of two substances are gaining popularity due to short electrons/ions diffusion distance and extra electrons transport pathways during reactions [35]. By reducing the thickness of bulk materials to ultrathin layers, the charge transport pathway will be minimized, thereby enabling high catalytic activity [36]. But until now, it is still a challenge to synthesize MXene-based 2D ultrathin heterostructured electrocatalysts, and the MXene-TMNs heterostructure has not reported so far.

Here, an in-situ and controllable nitridation method for synthesizing a hollow spherical structure consists of ultrathin 2D Ti3C2Tx-TiN heterostructures (MX-TiN) was reported for the first time. The 2D heterostructures are close contacted between Ti3C2Tx MXene and TiN with a high-quantity interface region by the created bonds to motivate high-density electron flow, achieving an overall enhancement in electrical conductivity, which can provide low resistance, high electrochemical reaction kinetics. The TiN (001) surface shows superior electronic structure that exhibits favorable adsorptive ability and high catalytic activity. Moreover, the ultrathin layered heterostructure was controlled to several nanometers in thickness, which shortens the electrons diffusion distance from catalysts to reactive sulfur species and enlarges the surface area with abundant adsorptive and catalytic sites. According to DFT calculations, the LiPSs are strongly immobilized on the surface of MX-TiN by chemical bonds, and the reaction kinetics are greatly enhanced by reducing the conversion barrier from liquid Li2S4 to solid Li2S2/Li2S. From an overall perspective, the hollow spherical structure can accomplish high sulfur loading, suppress spontaneously restack of 2D structure, tightly trap the LiPSs inside the spheres, and accommodate large volume swelling during discharging. Benefiting from these advantages, the S/MX-TiN cathode displays remarkable initial capacity, high rate capability and outstanding long-term cyclability. More importantly, even with high sulfur loading and lean electrolytes, this advanced cathode achieves an impressive areal capacity with high Coulombic efficiency (CE).

2 Experimental

2.1 Synthesis of Ti3C2Tx MXene Suspension

Ti3C2Tx MXene nanosheets were prepared via an improved etching method. In detail, 1.0 g Ti3AlC2 MAX phase was added into a mixed solution of LiF/HCl and followed by stirring at 40 °C for 24 h. MXene suspension was obtained after water washing, ultrasonic processing, and centrifugation. It is worth noting that the MXene sheets here are micron-sized and required further ultrasonic treatment for 6 h to obtain the MXene nanosheets. Finally, the concentration of MXene nanosheets suspension was measured at ~ 5 mg mL−1. Pure MXene powder can be obtained by centrifugation after adding HCl and freeze-drying.

2.2 Synthesis of Melamine–Formaldehyde Resin (MF) Spheres

Typically, 3.65 mL formaldehyde solution (37 wt% in water) was added into 80 mL deionized (DI) water and heated to 80 °C, then 1.0 g melamine was dissolved in the mixed solution. After stirring for 10 min, 0.1 mL formic acid was added to start polymerization and continued stirring at 80 °C for 1 h. Finally, the white precipitate was washed with DI water three times to obtain MF spheres.

2.3 Synthesis of MX-TiN and MX-TiO2 and Fabrication of S-Containing Composites

Firstly, MF spheres were dispersed in DI water and mixed with MXene nanosheets. Due to electrostatic interaction, MXene is wrapped on MF spheres (MF@MXene). MF@MXene composites were transferred to an alumina ceramic crucible after freeze-drying. The composites were heated to 450 °C with a rate of 5 °C min−1 for 1 h to partially remove and carbonize the MF, followed by heating to 800 °C at 5 °C min−1 for 30 min under Ar/NH3 and finally MX-TiN was obtained after cooling. The MX-TiO2 was obtained in the same process under Ar atmosphere without extra NH3.

MX-TiN, MX-TiO2, and pure MXene was separately mixed with sulfur powder at a mass ratio of 25:75. The mixture was transferred to a glass bottle and sealed with Ar gas. Then the bottles were heated to 155 °C for 15 h.

2.4 Characterizations

The morphology of as-prepared samples was investigated by scanning electron microscopy (SEM, Hitachi S-4800) and transmission electron microscopy (TEM, JEM-2100F). X-ray diffraction (XRD) patterns were collected by a D/max-2550 PC XRD (Rigaku, CuKα radiation). X-ray photoelectron spectroscopy (XPS) spectra were obtained using Escalab 250**s (Fig. 1g), the Ti, N, and C elements are uniformly distributed throughout the sphere, revealing that the TiN is evenly decorated on the MXene without aggregation. After sulfur loading, the S/MX-TiN hybrid remains the spherical shape with similar sizes and morphologies. It should be noted that there is no obvious sulfur aggregation observed on the surface of S/MX-TiN hybrid from SEM images (Fig. S10), indicating that the sulfur was completely filled in the MX-TiN. Meanwhile, the XRD of the S/MX-TiN mainly shows the characteristic peaks of sulfur (PDF#96-2244) (Fig. S11a), and the elemental map** of S/MX-TiN indicated the uniform distribution of elements Ti, C, N, and S (Fig. S12). For comparison, the MX-TiO2 with TiO2 aggregated into nanoparticles is synthesized by the same process under the Ar atmosphere without extra NH3 (Figs. S13-S15), and then the sulfur-containing composite S/MX-TiO2 is further prepared (Fig. S11b).

Figure 2a–c displays the XPS spectra of MX-TiN, which provide further information about the surface electron state and electronic structure. The high-resolution Ti 2p XPS spectra of MX-TiN show a pair of peaks (Ti 2p3/2 and Ti 2p1/2) (Fig. 2a), the signals at 455.3 and 461.3 eV can be ascribed to Ti-N configuration. The Ti-C peaks at 456.6 and 462.8 eV are derived from Ti3C2Tx MXene without nitridation, which is consistent with the obvious C-Ti peak (281.8 eV) of the high-resolution C 1s XPS spectra in Fig. 2b, proving the coexistence of Ti3C2Tx MXene, and TiN. Here, it should be noted that the Ti–O bonds in Ti 2p XPS are derived from the –O terminal groups of MXene and the inevitable surface oxidation [14, 40]. In addition, the N-Ti (396.1 eV) bond in N 1s spectra (Fig. 2c) also verifies the successful partial nitridation of MXene, another three peaks centered at 396.7, 398.4, and 400.7 eV can be assigned to pyridinic N, pyrrolic N, and graphitic N, respectively [41]. Raman spectroscopy in Fig. S16 also clearly shows the formation of MX-TiN with typical patterns of the two components of Ti3C2Tx MXene and TiN [33]. In addition, the XPS survey spectrum verifies (Fig. S17) the presence of Ti, C, and N elements, and the atomic ratio of Ti to N is 2.02. The ratio of Ti and N is 24.93 at% and 11.87 at% according to energy dispersive spectrometer (EDS) analysis (Fig. 2d), further indicating that the MXene is not fully nitridation, which is consistent with XPS analysis (Table S1).

Fig. 2
figure 2

a Ti 2p, b C 1s, andc N 1s XPS profiles of MX-TiN. d Elemental compositions of MX-TiN by EDS analysis. e Nitrogen adsorption–desorption isotherms of MX-TiN, MX-TiO2, and MXene. f TGA curves of S/MXene, S/MX-TiO2, and S/MX-TiN in Ar atmosphere

Full size image

In addition, the MX-TiN possesses a high specific surface area of 213.08 m2 g−1 based on BET method thanks to the hollow structure and the ultrathin 2D heterostructures, the value of which is much larger than those of reported MXene-based particle heterostructures (Fig. 2e and Table S2). In contrast, the surface area of MX-TiO2 and freeze-dried MXene nanosheets are smaller (145.90 and 0.6 m2 g−1, respectively) due to the aggregation of TiO2 nanoparticles and severe restacking of pure MXene nanosheets. It is well-known that both the adsorption and further conversion of sulfur species take place on the surface of catalysts [42], so the large surface area is vital to provide enough places for LiPSs adsorption and Li2S nucleation. In another concern, sulfur volatilization in the sulfur loading process by melting method causes the loss of active substances, so the sulfur content is one of the most important criteria to select sulfur host materials. The large specific surface area and hollow structure not only relieve the volumetric changes during cycling but also enable the high sulfur loading. As shown in Fig. 2f, the sulfur loading in the S/MX-TiN hybrid is 73.7 wt% (viz., only 2.3 wt% sulfur loss) determined by TGA, which is larger than those of S/MX-TiO2 (71.4%) and S/MXene (45.7%).

3.2 Adsorptive and Catalytic Mechanism

The strong adsorption ability of MX-TiN to LiPSs is the key to alleviating the shuttle effect. Firstly, the first-principle DFT calculations were performed to understand this chemisorption mechanism. The heterostructure was constructed by combining the TiN (001) surface on MXene surface. Figure 3a presents the side views of optimized configurations of Li2S6 adsorbed on MX-TiN and Ti3C2Tx MXene, and the configurations of other LiPSs species at different lithiation stages (Li2S8, Li2S4, Li2S2, and Li2S) are detailed in Fig. S18. In details, Li2S6 was immobilized by Li–N bonds on the MX-TiN surface with an adsorption energy of − 2.507 eV, while the adsorption energy on pure MXene is − 1.407 eV. As shown in Fig. 3b, it can be clearly observed that the adsorption energies of MX-TiN to all types of LiPSs are much higher than that of pure MXene, implying a stronger chemical interaction is formed between LiPSs and MX-TiN from a theoretical view.

Fig. 3
figure 3

a Optimized configurations of Li2S6 on MX-TiN and Ti3C2Tx MXene. b Adsorption energies of different LiPSs species on MX-TiN and Ti3C2Tx MXene. c Optical photographs of pure Li2S4 solution, and MXene/Li2S4, MX-TiO2/Li2S4, and MX-TiN/Li2S4 dispersions at 5 min and 8 h. d UV–Vis spectra of Li2S4 solutions after exposure at different adsorbents for 8 h. e S 2p XPS profile of MX-TiN/Li2S4 after soaking for 8 h

Full size image

To further verify the strong adsorption of MX-TiN towards LiPSs, the visual adsorption tests were performed by adding 15 mg sulfur host materials (MX-TiN, MX-TiO2, and MXene) into 3 mL 10 mM Li2S4 solution. As shown in Fig. 3c, after soaking for 8 h, the solution containing MX-TiN becomes complete transparent, implying the strong adsorption ability. However, the solution containing of MXene is still slightly yellow after the same time. Ultraviolet–visible (UV–Vis) absorption spectrum was carried out to monitor the concentration differences of Li2S4 in the solution after 8 h (Fig. 3d) [43]. Notably, the order of remaining LiPSs intensity follows Blank > MXene > MX-TiO2 > MX-TiN, which trend is consistent with the visual adsorption tests. To further understand the reaction mechanism between MX-TiN and LiPSs, XPS was conducted after adsorption for 8 h. The high-resolution S 2p XPS spectra in Fig. 3e indicate the coexistence of polysulfide (LiPSs), thiosulfate ([S2O3]2−) and polythionate ([O3S2–(S)x–2–S2O3]). The formation of sulfur species besides polysulfide proves the strong interaction between LiPSs and MX-TiN, which is essential for realizing the fast-kinetic reaction of LiPSs conversion [34, 44].

The preferred sulfur host materials not only favor the adsorption of LiPSs, but also catalytically promote the redox reactions between LiPSs and Li2S2/Li2S, which can achieve efficient sulfur utilization and superior cycling performance for Li–S batteries. First, the electrocatalytic effect of MX-TiN to the conversion of LiPSs and nucleation of Li2S is theoretically confirmed by DFT calculations. The catalytic behavior is closely related to the intrinsic electronic structure properties of the heterostructures [45]. Figure 4a shows the Gibbs free-energy diagrams of the simplified reactions from S8 to Li2S on the MX-TiN and pure MXene. The reaction from S8 to Li2S6 is spontaneous and there is little difference for the conversion from Li2S6 to Li2S4. However, the reaction from Li2S4 to Li2S2/Li2S involves not only liquid–solid phase transitions but also slower solid–solid transitions, where requires an extremely high reaction free-energy, implying that this is the rete-controlling step and the origination of the sluggish sulfur redox reaction [46]. The reduction of Li2S4 and further of Li2S2 needs to obtain electrons, while the TiN (001) surface is dominated by metallic Ti-3d states, which is responsible for transmitting electrons from the high conductive MX-TiN matrix thus guaranteeing the efficient electrocatalytic activity. The reaction free-energy from Li2S4 to Li2S2 and finally to Li2S on the surface of MX-TiN is + 0.306 and + 0.371 eV, which is much lower than that of pure MXene (+ 0.395 and + 0.631 eV), indicating that MX-TiN effectively reduces the conversion barrier from Li2S4 to Li2S2/Li2S and thereby enhances the reaction kinetics. Note that the transition from Li2S4 to Li2S contributes to 75% theoretical capacity, suggesting that the effective catalytic ability is very important for the electrochemical performance in Li–S system.

Fig. 4
figure 4

a The Gibbs free-energy profiles of LiPSs conversion on MX-TiN and pure MXene. b CV curves of MXene, MX-TiO2, and MX-TiN symmetric cells with and without Li2S4 electrolyte. Potentiostatic discharge profile at 2.05 V on c MX-TiN electrode, d MX-TiO2 electrode, and e MXene electrode with Li2S8 electrolyte for evaluating the nucleation kinetics of Li2S. f First discharging/charging profiles and corresponding ex-situ Raman spectra at different potentials of S/MX-TiN cathode. g CV profiles at different scan rates of S/MX-TiN cathode and h corresponding I–v2 slope curves

Full size image

To further demonstrate the catalytic effect of MX-TiN towards the conversion of LiPSs, symmetric cell tests and Li2S precipitation tests were carried out. Symmetric cells were assembled with 50 μL 0.1 M Li2S4 as the electrolyte to characterize the electrocatalytic effect of MX-TiN, MX-TiO2 and MXene for LiPSs conversion. As shown in Fig. 4b, the CV curves of MX-TiN exhibit a pair of redox peaks at − 0.120 and 0.122 V. These peaks can be attributed to the electrochemical conversion of Li2S4. The anodic peak is the reduction of LiPSs to Li2S2/Li2S. And in the cathodic scan, the Li2S2/Li2S was oxidized to LiPSs and finally to sulfur [47]. The MX-TiN symmetric cell without Li2S4 (Blank Group) exhibits a neat-zero capacitive current. In sharp contrast, the CV curves of MX-TiO2 and MXene electrodes display obviously lower current densities and more serious hysteresis voltage. These results strongly support the significant improvement of the electrocatalytic effect of MX-TiN to LiPSs conversion. Li2S precipitation tests are shown in Figs. 4c–e and S19, the profiles can be divided into contributions from the initial reduction of Li2S8, the bottom reduction of Li2S6, as well as the deposition of Li2S. According to Faraday’s law [7], MX-TiN electrode exhibits the highest precipitation capacity of 262.80 mAh g−1, implying higher catalytic efficiency for the precipitation of Li2S compared with those of MX-TiO2 (205.94 mAh g−1) and MXene (150.89 mAh g−1) electrodes. The catalytic ability of MX-TiN-2 h (single phase TiN) was also evaluated by Li2S precipitation test (Fig. S19), the current curve lies between that of MX-TiN and pure MXene, demonstrating the superiority of MX-TiN heterostructure. As we know, the precipitation of Li2S consists of crystal nucleation and crystal growth, and Li2S crystal nucleation in this process requires a higher driving force than its growth [48]. The peak current density and current response velocity can be used to evaluate the nucleation rate of Li2S. The ultrathin Ti3C2Tx-TiN electrocatalysts shortens electrons diffusion distance, thus decreasing the transfer time from the MX-TiN to sulfur species. Therefore, MX-TiN electrode displays a higher current density (0.46 mA) and a faster current response at 1500 s than those of MX-TiO2 electrode (0.36 mA and 2460 s, respectively), indicating smaller nucleation barrier of Li2S on MX-TiN. The MXene electrode shows the lowest Li2S precipitation capacity, which can be blamed on weak adsorption ability and less nucleation sites due to its spontaneously restacking. To make things worse, the catalytic sites are inactivated as the insulating Li2S continuously covering on the surface of catalysts, and this problem is more serious under high sulfur loading. In order to prove that the ultrathin 2D MX-TiN can provide enough active sites for constant catalytic reactions, the cells were disassembled after 24,000 s potentiostatic discharge, the morphologies of deposited Li2S on the carbon fiber electrode were further observed by SEM (Fig. S20). The carbon fiber is uniform covered by a layer of Li2S on MX-TiN electrode, while the nonuniformly distributed Li2S particles can be observed on the other electrodes, which further proves the enhanced catalytic ability of MX-TiN for Li2S deposition. Meanwhile, the EDS map** images also confirm the uniform Li2S coating on carbon fiber of the MX-TiN electrode (Fig. S19b). The decomposition of Li2S was further investigated by a potentiostatic charging process (Fig. S21), the MX-TiN electrode exhibits the highest capacity, demonstrating enhanced Li2S oxidation process. The above results clearly prove the bidirectional catalytic effect of MX-TiN for the conversion between LiPSs and Li2S.

To gain insight into the electrochemical reaction process during cycling, ex-situ Raman characterization of cycling electrodes was performed. In order to avoid oxidation, the electrodes were encapsulated in glass under Ar atmosphere before Raman test (Fig. S22). Figure 4f shows the first discharge–charge profiles and the corresponding ex-situ Raman spectra of S/MX-TiN cathode at some typical potentials. However, it should be noted that the peaks of MX-TiN are too weak to be witnessed compared to the strong peaks of sulfur species [49]. At the initial stage of discharge (OCP—2.30 V), the appeared weak peak of [S2O3]2− close to the characteristic peaks of S8 demonstrates the gradual transformation of sulfur on MX-TiN [50], which is in line with the XPS results of Fig. 3e. As the discharge continued to 1.7 V, all of the other peaks disappeared and the strong peak of S2− (Li2S) appeared gradually. It is proved that the efficient conversion from sulfur to Li2S under the catalytic effect of MX-TiN, which improves the utilization of active materials. The opposite process can be observed in the charging stage, implying the high reversibility of conversion reaction.

In order to demonstrate the enhanced reaction kinetics by MX-TiN, we studied the diffusion coefficient of lithium-ion. As shown in Figs. 4g and S23a–c, CV curves of different cathodes at incremental scanning rates are performed. In the cathodic scan process, there are two reduction peaks which can be ascribed to the reduction of S8 to soluble LiPSs (peak A), and subsequent reduction of LiPSs to solid Li2S2/Li2S (peak B). In the reversible anodic scan, Li2S2/Li2S are oxidized to LiPSs, and finally to S8 (peak C). Notably, the cathode with MX-TiN displayed a higher current density and a lower reaction polarization than those of MX-TiO2 and MXene. In addition, as the scan rate increases, MX-TiN cathode maintains clear CV curves and sharp redox peaks, indicating enhanced reaction kinetics and good rate performance. However, as for S/MX-TiO2 and S/MXene cathodes, the current increase slowly and the redox peaks become blurred as the scan rate increases. The diffusion coefficient of lithium-ion (DLi+) can be calculated by the Randles–Sevcik equation, I = 2.686 × 105n1.5ADLi+Cv0.5, where I is the peak current, n is the number of charges transferred, A is the geometric area of electrode, DLi+ is diffusion coefficient of lithium-ion, C is the concentration of lithium-ion in electrolyte, and v is the scanning rate. The peak current I has a good positive correlation with the square root of scan rate v (Figs. 4g and S23b, d), and the slopes of I–v0.5 for S/MX-TiN cathode are much bigger than those of S/MX-TiO2 and S/MXene, demonstrating the faster lithium-ion diffusion. The rapid lithium-ion diffusion can promote sulfur utilization and improve high-rate performance. Furthermore, the DLi+ of S/MX-TiN cathode at redox peaks A, B, C are 1.75 × 10–7, 2.96 × 10–7, 10.14 × 10–7 cm2 s−1, respectively, well above the DLi+ values for S/MX-TiO2 and S/MXene, further indicating the fast reaction kinetics (Fig. S24). We concluded that the larger DLi+ measured in S/MX-TiN cathode was due to the excellent conductive skeletons constructed by MX-TiN, which provides the fast channels for lithium-ions transport.

Taken all theoretical calculations and experimental results together, Fig. 5 schematically illustrates the adsorptive and catalytic mechanism of different sulfur species on (a) pure MXene, (b) MX-TiO2, and (c) MX-TiN. Firstly, the hollow spherical structure of MX-TiN derived from template-removal method prevents the restacking of 2D heterostructures, enabling sufficient space to store sulfur and accommodate the large sulfur expansion. Secondly, the construction of ultrathin thin layered structure not only exhibits the large surface area but also constructs fast channels for electrons and ions transport. In principle, a larger surface area will provide more interface contacting sulfur species. Therefore, compared with MX-TiO2 and pure MXene, MX-TiN with ultrathin layered 2D structures possess the most adsorptive and catalytic sites and the shortest electron transfer distances from catalysts to sulfur species. Thirdly, the metallic TiN (001) surface makes the heterostructures exhibit better conductivity and superior electronic structure, which reduces the barriers of LiPSs conversion and Li2S nucleation and enhances the catalytic activity, thus greatly boosting the Li–S reaction kinetics. And MX-TiN shows greater interfacial interaction to LiPSs, which captures the LiPSs strongly and effectively mitigates the serious shuttle effect.

Fig. 5
figure 5

Schematic illustration of adsorptive and catalytic mechanism of LiPSs species by different structures of a pure MXene, b MX-TiO2, and c MX-TiN

Full size image

3.3 Electrochemical Performance

Coin cells were assembled with lithium metal as anode and S/MX-TiN, S/MX-TiO2, S/MXene as cathodes to evaluate their electrochemical performance. Figure 6a displays the first discharge–charge profiles at a current density of 0. C. The initial capacity of S/MX-TiN is 1229.9 mAh g−1, much higher than those of S/MX-TiO2 (1060.2 mAh g−1) and S/MXene (997.0 mAh g−1). It is worth noting that Li2S decomposition is the first step for the charging process, and the conversion of Li2S to LiPSs leads to phase transformations and requires a large overpotential as the driving force to overcome the Li2S dissociation barrier. The S/MX-TiN cathode exhibits the smallest polarization (ΔE) and the lowest overpotential for Li2S oxidation (inset of Fig. 6a), which can strongly support efficient oxidation of Li2S and high cyclic reversibility. In contrast, the larger ΔE and overpotential of S/MX-TiO2 and S/MXene lead to accumulation of Li2S, even worse the nondecomposable Li2S will passivate the electrode and decrease the utilization of sulfur [51]. Meanwhile, Fig. S25 shows similar galvanostatic discharge–charge profiles of different cathodes at various current densities from 0.2 to 5C. The S/MX-TiN delivered a stable plateau at elevated current densities due to enhanced reaction kinetics, thus leading to superior rate performance and high CE. In sharp contrast, the S/MX-TiO2 and S/MXene electrodes present much larger polarization and the vanishing plateau under high C-rates. The rate performance was measured by increasing current density every ten cycles (Fig. 6b). The S/MX-TiN cathode exhibits high capacities of 1161.1, 960.4, 878.1, 768.1, 691.7, 635.9, 593.9 mAh g−1 (average capacities) at current densities from 0.2C, 0.5C, 1C, 2C, 3C, 4C to 5C, respectively. In comparison, the S/MX-TiO2 and S/MXene electrode deliver only 477.3 and 263.8 mAh g−1 at the high current rate of 5 C.

Fig. 6
figure 6

a The initial discharge–charge profiles of S/MXene, S/MX-TiO2, and S/MX-TiN cathodes. b Rate performance and c cycling performance at 0.2C of S/MXene, S/MX-TiO2, and S/MX-TiN cathodes. d Long-term cyclability of S/MX-TiN, S/MX-TiO2, and S/MXene at 5C. e Rate capability of S/MX-TiN at a high sulfur loading of 5.15 mg cm−2 and E/S ratio of 11.61. f Cycling performance of S/MX-TiN cathode at different sulfur loading of 4.99, 7.14, and 10.16 mg cm−2, respectively. Inset: Optical photograph of a “DHU 1951 2021” shape composed of 115 LED lamps. g Comparison of sulfur loading and areal capacity of S/MX-TiN cathode with other reported MXene-based and titanium-based sulfur cathodes

Full size image

Galvanostatic cycling stability performance was evaluated at different current densities. As shown in Fig. 6c, the S/MX-TiN delivers a reversible capacity of 1028.2 mAh g−1 with a good capacity retention of 87.1% after 100 cycles at 0.2C, which is much higher than those of S/MX-TiO2 (632.9 mAh g−1, 65.6%) and S/MXene (443.9 mAh g−1, 45.9%). The voltage plateaus of S/MX-TiN cathode change slightly after 100 cycles (Fig. S26a), indicating the high-capacity reversibility. However, there is a sharp capacity decrease with the increase of polarization for S/MX-TiO2 and S/MXene cathode (Fig. S26b-c). To further explore the long cyclic stability at a high current density, S/MX-TiN was tested at 5C over 1000 cycles (Fig. 6d). The S/MX-TiN displays an initial specific capacity of 647.7 mAh g−1, and maintains a capacity of 516.9 mAh g−1 after 1000 cycles with the CE as high as 99.3%, corresponding to 79.8% capacity retention and an average capacity decay of as low as 0.022% per cycle, which is much better than that of S/MX-TiO2 and S/MXene. The capacity of pure MX-TiN was tested under the same conditions to avoid its interference, its capacity contribution to the total capacity is negligible (Fig. S27). Furthermore, according to the CV measurements (Fig. S28), MX-TiN remains electrochemical inert over a voltage range of 1.7–2.8 V.

Electrochemical impedance spectra (EIS) of different cathodes were carried out on fresh batteries. As shown in Fig. S29a, the EIS consists of a semicircle at high frequency region and an inclined line on low frequency region. The intersection between the initial part of the semicircle and the horizontal axis represents the electrolyte resistance (Re). And the diameter of the semicircle represents the charge transfer resistance (Rct), which reflects the reaction kinetics of electrolyte and electrode interface. The equivalent circuit was performed (Fig. S29b) and the simulated results are summarized in Table S3. The S/MX-TiN exhibits the lowest Rct compared to the other two cathodes, demonstrating the smaller charge-transfer barrier at the MX-TiN heterostructured interface and enhanced interfacial redox kinetics. Moreover, to explore the changes of lithium metal anode and cathode of different cells during cycling, we disassembled the cells after cycling and characterized the surface morphology of anodes and cathodes by SEM (Figs. S30–S31). Thanks to the good adsorption and efficient catalytic effect of MX-TiN towards LiPSs, there is only a small amount of speckled Li2S deposited on the surface of lithium metal after cycling. Meanwhile, the surface of S/MX-TiN cathodes remains flat after cycling due to the buffering ability of the hollow structure, while the S/MXene cathode is cracked. Therefore, we conclude that the MX-TiN alleviates the shuttle effect and catalyzes the rapid conversion of LiPSs, which accounts for the remarkable cycling performance for S/MX-TiN cathode. In sharp contrast, lots of irregular sediments appeared on the surface of lithium for the cell with S/MXene cathode after cycling, which is caused by the LiPSs shuttle through the separator and form an insulating Li2S layer on the surface of lithium metal directly. This results in a rapid decline in capacity and an increase in resistance.

To broaden the practical applications of S/MX-TiN electrode, the evaluations of high-sulfur loading with lean electrolytes should be critically considered. We first investigate the wettability of electrolyte with S/MX-TiN cathode by testing contact angle. As shown in Fig. S32, the contact angle becomes less than 5° immediately when the electrolyte contacts the electrode, and the electrolyte completely infiltrate the electrode within a few seconds, implying the excellent wettability of electrolyte and S/MX-TiN cathode. Figure 6e shows the rate performance of S/MX-TiN cathode at a high-sulfur loading of 5.15 mg cm−2 and an E/S ratio of 11.61 μL mg−1. The S/MX-TiN electrode exhibits excellent rate capacities of 1042.3, 802.9, 690.8, 467.6 mAh g−1, the corresponding areal capacities of 5.36, 4.14, 3.56, 2.41 mAh cm−2 at 0.1C, 0.2C, 0.5C, and 1C, respectively. Notably, as the discharge rate decreases from 1C to 0.2C, the capacity returns to 808.1 mAh g−1 (4.16 mAh cm−2), conferring good reversibility at a relatively high-sulfur loading. Accordingly, the discharge–charge profiles of S/MX-TiN have a stable plateau at different current densities (Fig. S33), indicating its excellent reaction kinetics even at a high-sulfur loading. Moreover, Fig. 6f displays the cycling performance of S/MX-TiN cathodes at the different sulfur loading and E/S ratio at 0.2C. Impressively, the cathode with lean electrolytes (E/S = 7.84) and 10.16 mg cm−2 sulfur loading exhibits an initial capacity of 8.27 mAh cm−2 and 81.6% capacity retention with an average CE of 97.7% after 80 stable cycles. The cathode with 4.99 mg cm−2 and a slightly higher electrolyte (E/S = 11.98) exceeds 4 mAh cm−2 (capacity of typical commercial lithium-ion batteries) even after 160 cycles. Furthermore, a fresh cell battery with S/MX-TiN cathode (sulfur loading: ~ 5 mg cm−2) can light up 115 light-emitting diode (LED) lamps with the words “DHU 1951 2021” (inset Fig. 6f). Evidently, the obtained electrochemical performance in terms of sulfur loading and areal capacity show considerable competitiveness compared with other MXene-based and titanium-based sulfur host materials (Fig. 6g) [14, 44, 4952,53,54,55,56,57,58,59,60].

4 Conclusions

In summary, we rationally designed and synthesized a novel electrocatalyst for high-performance Li–S batteries based on ultrathin 2D Ti3C2Tx-TiN heterostructure. We explained the enhanced catalytic performance from the aspects of interfacial areas, sulfur species binding energy, polysulfides conversion and Li2S nucleation barriers, and lithium-ion diffusion coefficient. By constructing the ultrathin layered heterostructures within the thickness of several nanometers, MX-TiN exposes large surface area up to 213.08 m2 g−1 with abundant adsorptive and catalytic sites, shortens the electrons transfer distances to the reactive sulfur species. The metallic TiN (001) surface exhibits excellent conductivity and superior electronic structures, favoring a greater interaction with polysulfides and reducing the conversion barriers between LiPSs and Li2S. Meanwhile, the internal hollow structure of MX-TiN provides enough room for high sulfur loading and accommodates the large volume expansion of active materials during discharging, guaranteeing the good electrical contact and stability of the electrodes. On the basis of these fundamental understandings, the S/MX-TiN cathode exhibits low polarization, high initial capacity, and long-term cycling stability with extremely low-capacity decay rate. More importantly, the cathode delivers impressive area capacity (8.27 mAh cm−2) under high sulfur loading (10.16 mg cm−2) and lean electrolytes (E/S = 7.84), which improves the feasibility for practical applications of Li–S batteries. With insights generated by this work, we expect the rational design and construct of ultrathin 2D heterostructures not only improve the performance of Li–S battery system, but also broaden the horizons of electrocatalyst design for other energy storage and conversion systems.