Tuning Interface Bridging Between MoSe2 and Three-Dimensional Carbon Framework by Incorporation of MoC Intermediate to Boost Lithium Storage Capability

Highlights MoSe2/MoC/C multiphase boundaries boost ionic transfer kinetics. MoSe2 (5–10 nm) with rich edge sites is uniformly coated in N-doped framework. The obtained MoSe2 nanodots achieved ultralong cycle performance in LIBs and high capacity retention in full cell. Abstract Interface engineering has been widely explored to improve the electrochemical performances of composite electrodes, which governs the interface charge transfer, electron transportation, and structural stability. Herein, MoC is incorporated into MoSe2/C composite as an intermediate phase to alter the bridging between MoSe2- and nitrogen-doped three-dimensional (3D) carbon framework as MoSe2/MoC/N–C connection, which greatly improve the structural stability, electronic conductivity, and interfacial charge transfer. Moreover, the incorporation of MoC into the composites inhibits the overgrowth of MoSe2 nanosheets on the 3D carbon framework, producing much smaller MoSe2 nanodots. The obtained MoSe2 nanodots with fewer layers, rich edge sites, and heteroatom doping ensure the good kinetics to promote pseudo-capacitance contributions. Employing as anode material for lithium-ion batteries, it shows ultralong cycle life (with 90% capacity retention after 5000 cycles at 2 A g−1) and excellent rate capability. Moreover, the constructed LiFePO4//MoSe2/MoC/N–C full cell exhibits over 86% capacity retention at 2 A g−1 after 300 cycles. The results demonstrate the effectiveness of the interface engineering by incorporation of MoC as interface bridging intermediate to boost the lithium storage capability, which can be extended as a potential general strategy for the interface engineering of composite materials. Electronic supplementary material The online version of this article (10.1007/s40820-020-00511-4) contains supplementary material, which is available to authorized users.

To address aforementioned concerns, great efforts have been focused on nano-/microstructure design and carbon modification. In general, nanosized materials can effectively eliminate the effect of volume variations, providing a short ion diffusion length and offer large pseudo-capacitance, which lead to fast ionic migration and superior rate performance [16][17][18][19]. However, the large contact area between electrode materials and electrolyte may increase the side reaction and cause the dissolution of the electrode materials. Another effective strategy is making TMD and carbon composites, in which carbon serves as a soft matrix to buffer the volume variations and fast electron conductor [20][21][22][23]. Particularly, three-dimensional (3D) porous carbon skeletons are preferred because of their extra capability to tolerant the volume changes. Zhao et al. have achieved high-performance Co 3 O 4 lithium-ion battery anode by employing 3D carbon network substrate [24]. Cao and co-workers synthesized 3D porous Mo 2 C/C architecture to improve the lithium storage capacity [25]. Although the electrochemical performances are improved greatly, the long-term cycling stability is rear reported. The difficult can be attributed to the weak bonding between the active material and carbon substrates, or the mismatch of their volume variations, which degrade the structural stability of the electrode upon long-term cycling.
Interface engineering can introduce distortions, dislocations, and lattice defects, with distinguished electronic structures, which are long-range disorder and, hence, can lower the activation barrier, thus boosting reaction kinetics [5,26,27]. More recently, through interface coupling by chemical bonds, previous theoretical calculation and practical experiments both have revealed that Mo-O-C or Mo-C bonds between carbon and MoSe 2 can effectively enhance electronic conductivity and structure stability through the interface [28,29]. However, the portion of the chemical bonding between the active material and substrate is still too limited to stabilize structure. Therefore, it is highly expected to increase the chemical bonds at the interface or develop new strategy to improve the stability of composites, in particular for long-term cycling applications. Introducing intermediate material to bridge the active material and the substrate presents to be a wise choice. Comparing to MoSe 2 , MoC owns intrinsic higher electrical conductivity and chemical stability [30][31][32]. The structural stability and charge transportation will be greatly improved by the incorporation of MoC into the composite, which bridges MoSe 2 and carbon substrates by MoSe 2 -MoC-carbon connection.
Herein, we reported the temperature-induced one-step incorporation of MoC as an intermediate phase to bridge MoSe 2 -and N-doped three-dimensional carbon framework, which essentially improve the interfacial structural stability of the composite. Moreover, the in situ formed MoC can effectively inhibit the overgrowth of MoSe 2 nanocrystals and constrain MoSe 2 to nanodots with few layers and rich edge sites. The obtained materials guarantee bi-continuous fast electron/ion transport and highly reversible conversion reactions and enable fast reaction kinetics and good structural stability. As expected, 3D porous MoSe 2 /MoC/N-C electrodes deliver much improved long-term cycling stability. It retains 90% of its initial capacity after 5000 cycles at 2 A g −1 . And the assembled LiFePO 4 //MoSe 2 /MoC/N-C full cell also exhibits a high reversible capacity and good cyclic stability (86% capacity retention after 300 cycles at 2 A g −1 ).

Electrode Fabrication and Electrochemical Measurement
Electrochemical performances were evaluated by using 2032-type coin half cells with metallic lithium served as the anode. Working electrodes were prepared by casting a slurry containing the active material (70 wt%), Super P (15 wt%) and sodium carboxymethyl cellulose (15 wt%) dispersed in distilled water on a clean Cu foil current collector. Each electrode in our experiment has an area of 1.1304 cm −2 , and the loading of active material for each electrode was about 1.0-1.2 mg. All cells were assembled in a glove box (Mbraun, Germany) filled with ultrahigh-purity argon. For LIB assembly, polypropylene membrane and 1 M LiPF 6 in ethylene carbonate (EC)-dimethyl carbonate (DMC)-diethyl carbonate (DEC) (1:1:1 in volume) were used as a separator and electrolyte, respectively. As for the assembly of full cell, a pre-lithiation procedure was carried on to compensate the loss of lithium during the initial cycle in half cell. To ensure the maximized material utilization and reasonably evaluate the electrochemical property of the MoSe 2 /MoC@N-C, the full cell in this work was assembled based on the capacity ratio of ≈ 1.2:1 between the LiFePO 4 cathode and MoSe 2 / MoC/N-C anode. The cyclic voltammetry (CV) measurements were taken on an electrochemical workstation (CHI660C) at a scan rate of 0.1 mV S −1 in the voltage range of 0.01-3 V (vs. Li + / Li). The electrochemical impedance spectrometry (EIS) data were collected on a ZAHNER-IM6ex electrochemical workstation (ZAHNER Co., Germany) in the frequency range of 100 kHz to 10 mHz. Land battery tester (Land CT 2001A, Wuhan, China) was employed to investigate the galvanostatic charge/discharge performances, and all tests were conducted at room temperature.

Formation Mechanism and Structure Characterization of Materials
The synthetic mechanism of the 3D porous MoSe 2 / MoC/N-C and advantages of the structure are shown in Scheme 1. As a nitrogen-rich carbon network source, polyvinyl pyrrolidone (PVP) is a nonionic surfactant that can adsorb molybdate ions to facilitate the in situ formation of Mo-C bond. When the precursor annealed at high temperature, molybdenum can react with carbon to form MoC. Selenium powder is used to form MoSe 2 . And NaCl is employed as a template to create pores due to its environmental friendliness, cheapness, and high melting point. The obtained porous structure with 3D electrically conductive pathways possesses both large surface areas and abundant mass transportation channels, guaranteeing the good contact between electrolyte and electrode interface. More importantly, as shown in Scheme 1b, different from previous reports of MoSe 2 /C composites, MoSe 2 and carbon substrate were bridged by highly conductive MoC interphase, forming a tri-phase MoC/N-C heterostructure where the MoSe 2 / MoC/N-C connection can boost fast charge kinetics and highly reversible conversion, leading to improved long-term cycling stability. This MoSe 2 /MoC/N-C connection also induced the growth of MoSe 2 to terrace-terminated mode (illustrated in Scheme S1), resulting few-layered MoSe 2 nanodots, and thus, greatly contributed to pseudo-capacitance effect and guaranteed the fast the ion diffusion kinetics, which synergistically boosts lithium storage capability. More specific characterizations of material are comprehensively presented within Figs. 1 and 2. As shown in Fig. 1a (XRD patterns), the as-prepared sample annealed under 800 °C exhibited two additional diffraction peaks located at 35.7° and 48.7° compared to MoSe 2 phase (JCPDS No. 29-0914, 2H-MoSe 2 ), which are corresponding to MoC phase (JCPDS No. 45-1015). Specifically, compared to bare MoSe 2 and MoSe 2 /C, the broadened diffraction peaks in MoSe 2 /MoC/N-C sample indicate its relatively small crystal size. And the weaker intensity of (002) peak indicates rich defects along [002] direction [28]. Unlike the crystallinity normally increase with the rising of the annealing temperature (800 °C), the intensity and crystallites of MoSe 2 in the composite decreased. The XRD patterns suggest the growth of MoSe 2 may be changed. No detection of XRD peaks for Mo metal indicates the easier formation of MoC and its stable state at 800 °C.
As for the Raman characterization, all the samples present two bands at around 240 and 280 cm −1 (Fig. 1b), which are assigned to the A g 1 and E 1 2g stretching of MoSe 2 , respectively. The A g 1 -related peak is attributed to the exclusive peak of MoSe 2 2H structure, preferred for edgeterminated mode, while the E 1 2g peak is assigned to the characteristic peaks of MoSe 2 1T structure, favored for terrace-terminated MoSe 2 [13,33,34]. Thus, 1T/2H phases of MoSe 2 were consisted in three synthesized samples. Moreover, the increased intensity ratio of A g 1 /E 1 2g from bare MoSe 2 to MoSe 2 /MoC/N-C suggests the decreased thickness and increased exposed edge sites of MoSe 2 in the interconnected 3D porous structure. The D and G bands locating at 1375 and 1590 cm −1 confirm the presence of disordered and graphitic carbon structures within interconnected 3D porous superstructure, respectively [12,35,36]. The content of carbon was calculated as 34.6% in the MoSe 2 /C composite, determined by thermal gravimetric (TG) analysis as shown in Fig. S1, and the residual product (MoO 3 ) is verified by the XRD pattern in the inset. Furthermore, the XPS spectrum shows the existence of Mo, Se, C, and N elements (Fig. S2). The Mo 3d XPS spectrum ( Fig. 1c) shows three doublets: 229.5 and 232.4 eV assigned to Mo-Se bonds, 233.3 and 235.7 eV corresponding to Mo-O-C, and 228.6 and 231.7 eV corresponding to Mo-C bonding [13,24,33,37], confirming the formation of MoC in the MoSe 2 /MoC/N-C nanocomposite, which is consistent with the XRD results. More importantly, the formation of Mo-O-C and Mo-C bonds could be demonstrated, which implies the strong electronic coupling at the interface between MoC, MoSe 2 and 3D porous carbon. The formation of Mo-O-C bonds was mainly due to the combination of Mo ion with the carbonyl group (-C=O) in the PVP. Previous experiments and theoretical calculations have shown that Mo-C or Mo-O-C bonds in MoSe 2 /C nanocomposite could facilitate the ion and electron transportation in the composite [28,29]. For the Se 3d XPS spectra, two typical peaks at 55.5 and 54.8 eV correspond to the 3d 3/2 and 3d 5/2 peaks of Se 2− in the MoSe 2 / MoC/N-C nanocomposite (Fig. 1d). The C 1s profile in Fig. 1f can be deconvoluted into three individual peaks at 285.5 eV (C-N), 284.7 eV (C-C and C=C), and 287.1 eV (C-O, C=O). The formation of C-N bonds demonstrated the nitrogen doping in the carbon substrate. Furthermore, the N 1s peaks are partially overlapped with Mo 3p peaks (at 395.1 eV) and could be deconvoluted as graphitic N at 400.9 eV, pyrrolic-N at 399.5 eV, and pyridinic-N at 398.3 eV [13,24,33,37]. The nitrogen doping could   introduce more defects and electroactive sites, resulting in good electron transportation in the composite [38].
As depicted in Fig. S3, pure MoSe 2 agglomerates into irregular bulks about 20 μm in width. However, as expected, both MoSe 2 /C and MoSe 2 /MoC/N-C possessed a highly interconnected three-dimensional porous network with microstructure (Fig. 2). Interestingly, compared to MoSe 2 /C, MoSe 2 /MoC/N-C possesses more uniform pore distribution and thinner network thickness (Fig. 2a, b), which can be further confirmed by TEM observations. The uniform pores (about 80 nm) were created by dissolution of NaCl and the release of gases generated during the selenization. As a result, MoSe 2 /MoC/N-C composite possesses a large Brunauer-Emmett-Teller specific surface area (S BET ) of 34.9 m 2 g −1 and MoSe 2 @N-C possesses specific surface area of 21.5 m 2 g −1 , both of which are much larger than 7.9 m 2 g −1 of bare MoSe 2 (Fig. 2c). The detailed information of pore size distribution of three samples is shown in Fig.  S4. The high S BET and the porous structure are favorable to ion diffusion and structural stability upon cycling. The semitransparent TEM images further indicate the ultrathin network structure of composites (Fig. 2d-k). When we zoom in, for MoSe 2 /C sample, thick nanosheets with various sizes were attached to or embedded in the network (Fig. 2e, f), which were proved as MoSe 2 (Fig. 2g). In general, agglomeration is more likely to occur in hightemperature reactions. However, MoSe 2 crystals in MoSe 2 / MoC/N-C sample (800 °C) are smaller and more dispersed than those in MoSe 2 /N-C (600 °C). In comparison, for MoSe 2 /MoC/N-C, the interconnected 3D porous architecture (Fig. 2h) is more obvious than that of MoSe 2 /C. Moreover, a large number of small particles with size of 5-20 nm are uniformly distributed in smooth N-C network (Fig. 2i, j). These small sizes are favorable to the lithium-ion intercalation/deintercalation. Figure 2j further indicates that the thin carbon layer totally coats the small MoSe 2 /MoC nanocrystal. The N-C capsule can effectively protect nanosized MoSe 2 from adverse reactions and facilitate fast electron transport.
The high-resolution TEM (HRTEM) image (Fig. 2g, k) reveals that nanodots in MoSe 2 /MoC/N-C nanocomposite is composed of only a few layers (about 10 layers), much less than that of MoSe 2 /C. Furthermore, the layer distance of MoSe 2 /MoC/N-C (0.676 nm) is also larger than that of MoSe 2 /C (0.647 nm), which facilitate fast Li-ion transportations. Notably, the interlayer distances of 0.25 nm marked in Fig. 2k is assigned to the distance spacing of (100) plane of MoC crystal, which is in close contact with MoSe 2 crystals. And the corresponding fast Fourier transform (FFT) result (inset in Fig. 2k) revealed the (110) and (100) plane of the MoSe 2 and MoC, respectively, further confirming the formation of MoSe 2 /MoC heterostructure. Because the carbon source of MoC is from the carbonized PVP (N-C substrate), the in situ formed MoC is bonded to carbon framework, which can be confirmed by Mo-C and Mo-O-C bonding from XPS (Fig. 1c). Therefore, it can be deduced that the intermediate MoC bridges N-C framework and MoSe 2 through MoSe 2 /MoC/N-C connection, strengthening the interfacial coupling in tri-phase MoSe 2 /MoC/N-C superstructure, which enable fast charge transportation and good structural stability. And defects in MoSe 2 /MoC/N-C multiphase boundaries with distinguished electronic structures could lower the activation barrier, thus boosting reaction kinetics. In addition, around nanodots, some graphitic carbon fringes about 0.37 nm are observed, as indicated by the yellow dotted line (Fig. 2k). The graphitic carbon in the composite can provide conductive pathways for rapid electron transfer even at high current densities.
By comparing the structure of MoSe 2 /C and MoSe 2 / MoC/N-C composites (Fig. 2), it can be concluded that the in situ formed intermediate MoC could effectively inhibit the overgrowth of MoSe 2 nanocrystals and expand their layer spacing. During the heating process, partial Mo reacted with Se to form MoSe 2 first. When the reaction temperature rose to 800 °C, Mo and carbon reacted to form dispersion MoC, while MoSe 2 crystals continue to nucleate and grow, forming MoSe 2 /MoC/C heterojunction. This MoSe 2 /MoC/C connection induced the growth of MoSe 2 to terrace-terminated mode (illustrated in Scheme S1), resulting in few-layered and small-sized MoSe 2 nanodots. Furthermore, the in situ formed MoC can pin heterostructure, preventing crystals form and agglomerating. The dark-field scanning TEM and the corresponding elemental mapping images of MoSe 2 / MoC/N-C reveal that Mo, Se, O, N, and C are evenly distributed on the ultrathin carbon wall (Figs. 2l and S12). The distribution of Mo and Se concentrate on the location of nanodots, indicating the formation of MoSe 2 again, while the uniform distribution of N throughout the sample demonstrates that the 3D porous framework consists of N-doped carbon.

Electrochemical Lithium-Ion Storage Performance and Reaction Kinetics
The MoSe 2 /MoC/N-C composite was assembled to half and full cells to investigate their electrochemical performances. The cyclic voltammetry (CV) curves obtained during the initial 3 cycles (at a scanning rate of 0.1 mV s −1 ) exhibit multiple redox reactions for MoSe 2 /MoC/N-C electrode (Fig. 3a) Fig. S8c), suggesting the reversible conversion reactions of MoC [31,39]. Figure S6 presents charge/ discharge profiles of selected cycles of MoSe 2 /MoC/N-C electrode; after initial cycle, two platforms at around 1.8 and 1.2 V can be ascribed to the contribution of MoSe 2 and MoC in the composite electrode, respectively. The result is the same as peak voltages in CV curves. The initial charge and discharge capacities are 812 and 1014 mAh g −1 , respectively, corresponding to a coulombic efficiency (CE) of 80.1%, which is higher than most reported MoSe 2 electrodes [14,24,40,41]. The irreversible capacity in the initial cycle may come from the formation of a SEI film [24]. From the second cycle, all charge/discharge curves were close to overlapping, indicating the excellent electrode reversibility. Figure 3b compares the cycling performances of three samples at 0.1 A g −1 . MoSe 2 /MoC/N-C nanocomposite exhibits the highest specific capacity of around 800 mAh g −1 compared to MoSe 2 /C (650 mAh g −1 ) and bare MoSe 2 . This result demonstrates that constructing interconnected 3D porous carbon-modified nanocomposite can indeed improve the capability of storage Li ion. By the way, in the early stages of the cycling (Fig. 3b), the capacities of MoSe 2 /C and MoSe 2 /MoC/N-C rise slightly, which may be related to the activation of the electrode material. MoSe 2 /MoC/N-C shows the best rate performance (Fig. 3c). The electrode delivered average capacities of 773, 720, 680, 648, 634, 622, and 575 mAh g −1 at 0.1, 0.2, 0.5, 1.0, 1.5 2.0, and 4.0 A g −1 , respectively, and it can fully recover and even displays the higher capacity when the current density switched back to 0.1 A g −1 . However, MoSe 2 /C and pure MoSe 2 undergo apparent capacity loss at high rates in different degrees. The obtained high specific capacity and good rate capability of the MoSe 2 /MoC/N-C electrode also show superiority compared to many other carbon-modified MoSe 2 anode materials (Fig. 3e) [41][42][43][44][45][46][47], which demonstrates the advantage of fast interfacial charge transfer through MoSe 2 / MoC/N-C connection again. What makes the MoSe 2 /MoC/N-C electrode more attractive is its ultralong cycle life with the capacity retention of 90% and a high capacity of 535 mAh g −1 after 5000 cycles at 2 A g −1 (Fig. 3d). In sharp contrast, for MoSe 2 /C electrode, a capacity of 308 mAh g −1 can be maintained just 1000 cycles, and bare MoSe 2 even undergoes a sharply fading after 295 cycles.
To confirm the role of MoC in lithium storage, threedimensional porous MoC/C was prepared through a similar routine of preparing MoSe 2 /MoC/N-C. The SEM image (Fig. S8b) shows that MoC/C possesses the similar morphology to MoSe 2 /MoC/C composites (Fig. 2b). The cyclic voltammetry (CV) curves and the cycling performance of MoC/C composite as anode for LIBs were further carried out to study the lithium storage properties of MoC (Fig.  S8c, d). As a result, it exhibits the capacity close to that of MoSe 2 /MoC/N-C and good electrochemical reversibility. Meanwhile, the weight ratio between MoSe 2 and MoC was estimate as 51:2 (molar ratio of 54:5) according to Table S1, indicating that MoC possess relatively low contents in the electrodes. Therefore, the direct capacity contribution from MoC to the whole electrode is relatively limited. However, MoC plays an important role in the interfacial structure regulation, including more stable chemical binding and reduced MoSe 2 particle size, which is more correlated with the rate performance and structural stability. MoSe 2 /MoC/N-C shows a higher capacity than MoSe 2 /C mainly because its smaller MoSe 2 nanodots can provide more lithium storage sites. The MoC connected MoSe 2 and N-C network through C/MoC/MoSe 2 multiphase boundaries, enabling fast electron/ion transport and structural stability through interface, resulting in good rate performance and structural stability.
Encouraged by the excellent performance MoSe 2 / MoC/N-C electrode in half cell, we assembled a lithiumion full by pairing with the commercial LiFePO 4 cathode (Fig. 3f). As shown in Fig. 3g, the full cell displays a complete discharge plateau at about 2.3 V with a reversible capacity of about 267 mAh g −1 at 2 A g −1 . Meanwhile, its energy density was calculated as 93.9 Wh kg −1 at the high power density of 698.5 W kg −1 . It is still a great challenge to obtain stable cycling performance of the full-cell batteries based on conversion reaction electrodes because of the low coulombic efficiency and large consumption of lithium sources during the initial cycles to form SEI layers. However, the LiFePO 4 //MoSe 2 /MoC/N-C full cell exhibits good capacity retention of 86% after 300 cycles (Fig. 3h), indicating the good cyclic stability.
The reaction kinetics of MoSe 2 /MoC/N-C electrode was systematically studied by CV measurements, galvanostatic intermittent titration technique (GITT) and EIS spectroscopy. The electrochemical kinetics in the as-prepared electrodes could be considered as capacitive-dominated and diffusion-controlled from the CV curves (Figs. 4a and S9). After detailed calculations as detailed descriptions in the Supporting Information, the higher b value of the MoSe 2 / MoC/N-C electrode (0.90 vs. 0.69 of the MoSe 2 /C at the cathodic peak) reveals a faster ionic transportation which resulted in a better rate performance (Fig. 4b) [48][49][50][51]. Figure 4c shows the calculated voltage profile for the capacitive current (colored region) at the scan rate of 1.2 mV s −1 . As a result, 74% of the total capacity is calculated as the capacitive contribution for the MoSe 2 /MoC/N-C electrode, higher than 64% of the MoSe 2 /C. Furthermore, Fig. 4d compares the calculated capacitive contribution of two electrodes at various scan rates. The ratios of capacitive contribution of MoSe 2 /MoC/N-C electrode are always higher than those of MoSe 2 /C, with a maximum value of 90% at 2 mV s −1 . The large pseudo-capacitive contribution would play a critical role to be benefit for ionic transportation in the electrode [48,49], which boosting high-rate Li-ion storage capability.
The diffusion-controlled reaction kinetics of samples were further analyzed in depth by GITT (Fig. S11). As shown in Fig. 4e, f, the diffusion coefficient of MoSe 2 /MoC/N-C electrode is about 3.93e−11 to 1.55e−10 cm 2 s −1 , higher than those of MoSe 2 /C and bare MoSe 2 , suggesting that expanded layer distance, rich edge defect MoSe 2 nanocrystal, and MoSe 2 /MoC/N-C connection definitely facilitate fast Li-ion diffusion kinetics. The electrochemical impedance spectra (EIS) results ( Fig. S10 and Table S2) demonstrated that charge transfer resistances (R ct ) of MoSe 2 /MoC/N-C (43 Ω) are smaller than those of MoSe 2 /C (72 Ω) and bare MoSe 2 (557 Ω), suggesting that MoSe 2 /MoC/N-C shows the fastest interface kinetics. All above analysis suggested that the optimized interconnected 3D porous N-C framework and stronger electronic coupling at interface through MoSe 2 / MoC/N-C connection facilitate reaction kinetics.
Ex situ TEM images were employed to investigate the structural integrity of MoSe 2 /MoC/N-C electrode after certain cycles at 2 A g −1 . Compared to the ex situ TEM images of as-synthesized MoSe 2 /C ( Fig. S13A1 and A2), the MoSe 2 flakes in the MoSe 2 /C electrode (Fig. 4g or Fig. S13A3) were broken into small pieces and form aggregates after 100 cycles, which may further lead to electrode pulverization and indicate its inferior structure stability. However, at the same cycles, MoSe 2 /MoC/N-C electrode ( Fig. 4i or Fig S13B3) displayed the good preservation of interconnected 3D porous network structure, and the small nanocrystals encapsulated in the carbon could still be observed, indicating the excellent structural stability of MoSe 2 /MoC/N-C nanocomposite. Compared to the 10-layered structure in Fig. 2k (or Fig. S13B2), the HRTEM image (Fig. 4j or Fig. S13B4) demonstrates that MoSe 2 nanodots have been exfoliated into less layers (about 4 layers) and smaller size (about 5 nm). The interlayer distance also expanded to 0.751 nm after cycling. In addition, large areas of graphitized carbon were observed with an enlarged spacing (about 0.411 nm), slightly larger than the initial value of 0.370 nm for the pristine carbon. The expanded interlayer distance of MoSe 2 and carbon could be caused by the repeated insertion/extraction of Li ions [28,36]. The smaller MoSe 2 nanocrystal can further provide Li storage sites, benefiting the capacity of electrode. And the increased graphitization carbon layer enhanced the electronic conductivity. These changes of electrodes material during cycling result in climbing capacities. To further examine the reinforced structure stability of MoSe 2 /MoC/N-C, the TEM images of MoSe 2 /MoC/N-C electrode after 500 cycles are compared in Fig. S13B5-B7. Figure S13B5 confirms that the MoSe 2 /MoC/N-C electrode material retains integrated. As shown in Fig. S13B6 and B7, small shadow dots were still distributed in the carbon framework in MoSe 2 / MoC/N-C electrode, implying the products were also evenly constrained in carbon framework. These results prove that MoSe 2 /MoC/N-C connection can constrain the in situ generated MoSe 2 and facilitate the highly reversible conversion during cycling.

Conclusion
In conclusion, 3D porous MoSe 2 /MoC/N-C nanocomposite was synthesized by a temperature-induced method. The interface bridging between MoSe 2 and carbon framework was tuned by MoSe 2 /MoC/N-C connection, greatly improves the structural stability and electronic conductivity and reduces the interfacial charge resistance. Moreover, it also exhibits many features favorable for lithium-ion storage, such as MoSe 2 nanodots encapsulated into 3D connected carbon network to improve charge/ion transfer kinetics and minimize the effect of volume expansion; expanded interlayer spacing of MoSe 2 to promote lithium-ion diffusion; large electroactive surface area from N-C and rich edge defects of MoSe 2 ; and reinforced structure stability by intermediate in situ MoC. As a result, the as-prepared electrode delivered ultralong and stable cycling performance with a specific capacity of 618 mAh g −1 after 5000 cycles (90% capacity retention) at 2 A g −1 . It also shows promising potential in LiFePO 4 //MoSe 2 /MoC/N-C full cell (86% capacity retention at 2 A g −1 after 300 cycles). Inspired by the harvest of superior cell performance, in situ engineering corresponding metal carbide as interphase to connect metal dichalcogenide and carbon matrix can be developed as an innovative and effective strategy to achieve high-electrochemical-activity materials.

Supporting Information
Experimental section, supplementary scheme diagram of interface modification, TG, XRD, BET, SEM, EIS, CV, GITT and equations associated with this article are given in the online version or from the author.
Open Access This article is licensed under a Creative Commons Attribution 4.0 International License, which permits use, sharing, adaptation, distribution and reproduction in any medium or format, as long as you give appropriate credit to the original author(s) and the source, provide a link to the Creative Commons licence, and indicate if changes were made. The images or other third party material in this article are included in the article's Creative Commons licence, unless indicated otherwise in a credit line to the material. If material is not included in the article's Creative Commons licence and your intended use is not permitted by statutory regulation or exceeds the permitted use, you will need to obtain permission directly from the copyright holder. To view a copy of this licence, visit http://creat iveco mmons .org/licen ses/by/4.0/.