Active sites-enriched hierarchical MoS2 nanotubes: highly active and stable architecture for boosting hydrogen evolution and lithium storage

Energy Research Institute (ERI@N), Inter Technological University, Research Techn Singapore Division of Physics and Applied Physics, Sciences, Nanyang Technological University E-mail: Zexiang@ntu.edu.sg Energy Research Institute (ERI@N), Nanya Drive, 637553, Singapore. E-mail: LIJY@ntu Jiangsu Key Laboratory of Advanced Laser and Electronic Engineering, Jiangsu Norm E-mail: yanghao@jsnu.edu.cn † Electronic supplementary informa 10.1039/c6ta02034h ‡ Jin Wang and Jilei Liu contributed equa Cite this: J. Mater. Chem. A, 2016, 4, 7565


Introduction
Since the discovery of carbon nanotubes in 1991, intensive efforts have been motivated to fabricate various types of nanotubes due to their unique physical/chemical properties and expansive applications in a variety of elds. [1][2][3] Similar to carbon, fullerene-like and tubular transition-metal dichalcogenides (TMDs) have been increasingly investigated both theoretically and experimentally. [4][5][6][7][8][9] MoS 2 , as a typical layered metal dichalcogenide, has been widely recognized as a universal material for energy applications. Novel applications of tubular structured MoS 2 in catalysis, energy storage, lubricants, electronics and optoelectronics have driven the development of synthesis strategies that enable control of its structure and morphology. 10-28 MoS 2 nanotubes were rstly synthesized by Tenne et al. by a gas reaction of MoO 3 and H 2 S under a reducing atmosphere at elevated temperatures (800-950 C). 7, 29 Chen and co-workers also reported the synthesis of open-ended MoS 2 nanotubes by a gas-solid reaction. 11 However, these approaches for the fabrication of MoS 2 nanotubes always require complicated procedures with high temperatures or dangerous gases (H 2 and H 2 S). Besides, extensive research has proved that hollow nanotubes with abundant active edge sites and high specic surface area can achieve excellent performance in many applications. [30][31][32][33] Recently, a variety of solution-phase strategies have also been adopted to synthesize tubular MoS 2 structures with high surface area. For example, Zhang et al. reported an anion-exchange method for the synthesis of hierarchical tubular MoS 2 structures, which exhibit high activity and stability for photoelectrocatalytic and electrocatalytic hydrogen-evolution reaction (HER). 30 Wang et al. synthesized 3D assembled tubes constructed by single-layered MoS 2 nanosheets through a solvothermal method. 31 The assembled tubular architecture exhibits excellent electrochemical performance as an anode material of lithium-ion batteries (LIBs), and high activity in HDS catalysis. The excellent properties are probably attributed to the porous tubular architecture, which can shorten the transport path of lithium ions, and provide sufficient exposed active sites for reactions. These important results represent that it is a signicant topic to tune the properties of MoS 2 nanotubes by controlling their microstructures. Despite these advances in the fabrication of tubular nanostructures, it is still a considerable challenge to obtain MoS 2 nanotubes with sufficient exposed active sites, high specic surface area and robust stability by a facile method.
Herein, we develop a facile template-assisted approach for the construction of hierarchical MoS 2 hollow nanotubes implanted with exceptional active sites. The synthesized hierarchical nanotubes are constructed byMoS 2 hollow nanotubes decorated with MoS 2 nanosheets, which are self-supported on exible three-dimensional graphene foam (3DGF). We denote it as MT@MS/GF. The obtained hierarchical MT@MS/GF exhibits excellent electrocatalytic activity and high stability with low overpotential in an acidic electrolyte for the electrochemical HER. In addition, when evaluated as an anode material for LIBs, the binder-free hierarchical MoS 2 electrode manifests high specic capacity and excellent rate capability as well as extremely long-term cycle stability.

Experimental section
Synthesis of elongated core/shell Ni 3 S 2 @MoS 2 coaxial nanobers on Ni/graphene foam The growth of Ni/graphene foam (Ni/GF) was achieved by our reported method. 23 MoS 2 /Ni 3 S 2 @MoS 2 nanobers were prepared using a facile one-step hydrothermal process. 34 In a typical experiment, a mixed solution was prepared by dissolving 150 mg of thiourea (NH 2 CSNH 2 ), 75 mg of sodium molybdate (Na 2 MoO 4 ) and 0.2 g of polyvinylpyrrolidone (PVP) in 30 mL of distilled water. Then, the resulting solution was transferred into a Teon-lined stainless steel autoclave. A piece of Ni/GF with an area of 8 cm 2 was immersed into the reaction solution. The autoclave was then sealed and the hydrothermal reaction was conducted at 200 C for 12 h. Aer the autoclave was cooled down to room temperature, the samples were rinsed with DI water for several times and then dried in an electric oven at 60 C for 12 h. The achieved samples were thermally decomposed in a tube furnace at 400 C for 2 h under a H 2 /Ar (5 : 95 v/v) atmosphere with a heating rate of 10 C min À1 . For comparison, the samples were prepared with the same reactant concentration at 200 C for 8 h to prepare Ni 3 S 2 @MoS 2 nano-bers on graphene foam.

Synthesis of hierarchical MoS 2 nanotubes on graphene foam
The obtained Ni 3 S 2 -MoS 2 hybrid nanobers supported on Ni/ GF were immersed in a mixed solution of FeCl 3 (1 M) and HCl (0.5 M) for 48 h. Meanwhile, Ni 3 S 2 was oxidized to NiS 2 . In order to remove the residual NiS 2 , the obtained samples were hydrothermally treated in 5 M HCl solution at 90 C for 4 h, followed by washing with DI water until neutral pH, and drying at 60 C in an electric oven. Thus hierarchical MoS 2 nanotubes directly supported on 3DGF were obtained. Aer hydrothermal treatment, MoS 2 /Ni 3 S 2 @MoS 2 nanobers supported on Ni/GF transformed into MoS 2 hollow nanotubes decorated with MoS 2 nanosheets on 3DGF, which is denoted as MT@MS/GF. Meanwhile, Ni 3 S 2 @MoS 2 nanobers supported on Ni/GF transformed into tubular MoS 2 nanostructures on 3DGF, which is denoted as MT/G. The weight of MoS 2 in the 3D architectures is calculated by weighting 3DGF and the nal hierarchical composites aer hydrothermal treatment.

Characterization
The X-ray powder diffraction (XRD) pattern of each sample was recorded on a Bruker D8 Advance powder X-ray diffractometer using Cu Ka radiation (l ¼ 0.15406 nm). Field emission scanning electron microscopy (FESEM, model JSM-7600F, JEOL Ltd, Tokyo, Japan) was used to characterize the morphologies of the synthesized samples. Transmission electron microscopy (TEM) images were taken using a JOEL JEM 2100F microscope. Raman spectroscopy was recorded using a Renishaw Raman microscope with a 2.33 eV (532 nm) excitation laser. The Si peak at 520 cm À1 was used as the reference to calibrate the wave number. The XPS measurements were performed with a VG ESCALAB 220i-XL system using a monochromatic Al Ka1 source (1486.6 eV). All XPS spectra were obtained in the constant pass energy (CPA) mode. The pass energy of the analyser was set to 10 eV to have high measurement accuracy. The binding energy scale was calibrated with pure Au, Ag and Cu by setting Au 4f 7/2 , Ag 3d 5/2 and Cu 2p 3/2 at binding energies of 84.0, 368.3 and 932.7 eV, respectively. Surface area analysis was conducted using the Brunauer-Emmett-Teller (BET) theory (Micromeritics, ASAP 2020).

Cell assembly and electrochemical measurements
To test the anode performance of all synthesised materials, CR 2016 coin cells were made using a Celgard 2400 as the separator and 1 M LiPF 6 in ethylene carbonate-diethylene carbonate (EC : DEC ¼ 1 : 1) as the electrolyte. Li-metal was used as the counter and reference electrodes. The coin cells were assembled inside an argon-lled glovebox with oxygen and water contents below 1 and 0.1 ppm, respectively. The obtained MT@MS/GF and MT/GF were punched to a disk-shaped electrode with a diameter of 12 mm for the electrochemical measurements, which were directly used as working electrodes and assembled into coil cells without adding any conductive or binding materials. Galvanostatic charging and discharging tests were conducted using a battery tester (NEWARE) at different current rates. It should be noted that all the specic capacities reported in this work are based on the total mass of the composites rather than the mass of MoS 2 only. Cyclic voltammetry (CV) was performed using an electrochemical workstation (CHI 760D, Chenhua, Shanghai) from 10 mV to 3 V at a scanning rate of 0.5 mV s À1 . Electrochemical impedance spectroscopy (EIS) was also carried out with an electrochemical workstation over a frequency range from 10 6 Hz to 100 mHz at open circuit potential aer two galvanostatic charging and discharging cycles at 100 mA g À1 .

Electrocatalytic study
The obtained MT@MS/GF and MT/GF composites with an area of 1 cm 2 were directly used as working electrodes for HER performance evaluation. The electrochemical HER tests were performed in a 0.5 M H 2 SO 4 electrolyte with platinum foil and saturated calomel electrode (SCE) as counter and reference electrodes, respectively, within a three electrode set-up. The potentials reported in our work were vs. the reversible hydrogen electrode (RHE) through RHE calibration described below. Linear sweep voltammetry was conducted in 0.5 M H 2 SO 4 with a scan rate of 5 mV s À1 . The polarization curves were replotted as overpotential (h) vs. log current (log j) to obtain Tafel plots for assessing the HER kinetics of the investigated catalysts. By tting the linear portion of the Tafel plots to the Tafel equation (h ¼ b log(j) + a), the Tafel slope (b) can be obtained. All data were reported without iR compensation. In all measurements, we used SCE as the reference electrode. It was calibrated with respect to RHE. The calibration was performed in the highpurity hydrogen-saturated electrolyte with a Pt foil as the working electrode. Cyclic voltammetry was run at a scan rate of 5 mV s À1 , and the average of the two potentials at which the current crossed 0 was taken to be the thermodynamic potential for the hydrogen electrode reaction. In 0.5 M H 2 SO 4 solution, The long-term stability tests were performed by continuous linear sweep voltammetry (LSV) scans from 0.08 to 0.88 V (vs. RHE, in 0.5 M H 2 SO 4 ) at a sweep rate of 5 mV s À1 .

Results and discussion
The synthesis strategy of hierarchical MT@MS/GF is schematically depicted in Fig. 1. First, few-layer MoS 2 nanosheets decorated Ni 3 S 2 @MoS 2 coaxial nanobers on 3DGF/Ni are employed as the template, which are synthesized through a facile hydrothermal method. 34 Scanning electron microscopy (SEM) and transmission electron microscopy (TEM) images show the as-synthesized sample with an obvious core-shell structure, in which a large amount of ultra-small few-layer MoS 2 nanosheets decorate elongated Ni 3 S 2 @MoS 2 coaxial nanobers (denoted as MoS 2 /Ni 3 S 2 @MoS 2 ) ( Fig. S1a-d †). It should be noted that MoS 2 /Ni 3 S 2 @MoS 2 nanobers are self-supported on 3DGF/Ni. The as-prepared hierarchical MoS 2 /Ni 3 S 2 @MoS 2 nanobers are transformed into hierarchical MoS 2 tubular architectures by selectively removing the Ni 3 S 2 sacricial template aer a FeCl 3 -assisted etching process, while MoS 2 is chemically stable during the etching process. Meanwhile, the nickel substrate from 3DGF/Ni can be effectively etched. Then, we can obtain hierarchical MoS 2 tubular architectures directly supported on 3DGF. The detailed reaction equations are described as follows. 35 Ni 3 S 2 + 4FeCl 3 / NiS 2 + 4FeCl 2 + 2NiCl 2 (1) Detailed procedures are given in the Experimental section. The phase and morphology of the as-prepared sample were systematically investigated by XRD and SEM. The XRD pattern of the obtained sample can be assigned to the hexagonal phase MoS 2 (a ¼ b ¼ 0.316 nm, c ¼ 1.230 nm, JCPDS card no.  and GF (Fig. 2a). The GF displays a typical diffraction peak at 26.5 corresponding to the (002) reection of graphitic carbon (JCPDS card no. 75-1621). The highly crystalline structure of GF is further characterized in Fig. S2. † The SEM image reveals its tubular morphology with uniform diameters of 50-100 nm and lengths of above 2 mm (Fig. 2b). The MT@MS/GF composites were sonicated in ethanol for 10 min. The hierarchical tubular structures collapsed due to the ultrasonication effect, which were then dropped onto a TEM grid and dried under ambient conditions. The TEM image shows uniform nanotubes with inner and outer diameters of 30 and 50 nm (Fig. 2c). Close examination clearly indicates that a typical nanotube is composed of MoS 2 layers (5-8 layers) with an interlayer spacing of 0.68 nm, which are further decorated with a large amount of ultra-small MoS 2 nanosheets ( Fig. 2d and e). Partial nanotube walls show obvious broken layers with rich defects, highlighted by the dashed lines in Fig. 2d, which are probably caused by the detachment of Ni 3 S 2 nanobers during the etching treatment.
Further, the HRTEM image shows the atomic structures of the nanotube (Fig. 2e). The (001) planes are parallel to each other and together constitute the tubular structure with an interlayer distance of 0.68 nm. The enlarged HRTEM image shows the regular hexagonal lattice orientation of 2H-MoS 2 (Fig. 2f). The fast Fourier transformation (FFT) pattern in the inset reveals the  orientation along the [001] zone axis. High-angle annular dark-eld scanning transmission electron microscopy (HAADF-STEM) images clearly conrm the hollow tube-like morphologies and uniform composition distribution of the obtained nanotubes ( Fig. 2g-i). The energy dispersive X-ray spectrometry (EDS) results also conrm the composition of MoS 2 (Fig. 2j).
Raman spectroscopy was used to further investigate the structures of the MT@MS/GF and bare GF. Two obvious peaks at 1580 and 2720 cm À1 are attributed to the G and 2D bands of GF, respectively (Fig. 3a). The high 2D/G intensity ratio as well as the absence of the D band at 1350 cm À1 in the GF spectrum demonstrates the high quality of GF before MoS 2 growth. By contrast, the appearance of the D band at 1350 cm À1 aer MoS 2 growth indicates the presence of sp 3 bonding which successively corroborates a strong coupling between MoS 2 and GF. Two characteristic peaks at 380 and 405 cm À1 in the MT@MS/ GF sample are attributed to E 1 2g and A 1 g modes of MoS 2 , respectively. X-ray photoelectron spectroscopy (XPS) was employed to further probe the chemical states of S and Mo in MT@MS/GF (Fig. 3b and c). Two peaks of Mo 3d 3/2 and Mo 3d 5/2 at binding energies of 232.5 and 229.3 eV correspond to the characteristic peaks of Mo 4+ in MoS 2 . The S 2p peaks at 162 and 163.2 eV are ascribed to the characteristic peaks of S 2À in MoS 2 . 36,37 In virtue of the hollow tubular structure and ultrathin nanosheets, the MT@MS/GF sample with the hierarchical hollow structure exhibits a high Brunauer-Emmett-Teller (BET) surface area of 89.6 m 2 g À1 with abundant mesoporous channels (Fig. 3d).
The morphology of hierarchical MoS 2 hollow nanotubes on 3DGF can be optimized using the template precursor. This enables us to study the correlation between the microstructure and electrochemical performance. Herein, for comparison, we also fabricate MoS 2 hollow nanotubes on 3DGF, but without the decorating MoS 2 nanosheets (denoted as MT/GF), which are derived from core-shell Ni 3 S 2 -MoS 2 nanobers with a homogeneous heterointerface on 3DGF/Ni (denoted as Ni 3 S 2 @MoS 2 ) (Fig. S1d-f †). The TEM image of MT/GF shows that the typical nanotubes are composed of MoS 2 layers with a wall thickness of 5-10 nm ( Fig. 4a and b). The enlarged HRTEM image clearly indicates that a typical hollow nanotube is indeed composed of MoS 2 layers ($10 layers) with an interlayer spacing of 0.68 nm (Fig. 4c). Further examination shows the regular hexagonal lattice orientation of 2H-MoS 2 (Fig. 4d). The fast Fourier transformation (FFT) pattern in the inset reveals the orientation along the [001] zone axis. The XRD pattern also conrms the pure phase of MT/GF (Fig. S3 †).
Recently, MoS 2 and its composites demonstrated their intriguing performance for lithium storage and remarkable efficiency as catalysts of HER. 32,33,[38][39][40][41][42][43][44][45][46][47][48][49][50][51][52][53][54] The decent lithium storage and HER properties have been demonstrated (see a complete list in Tables S1 and S2 †). In this study, we investigate the electrocatalytic HER activities of MT@MS/GF in comparison with MT/GF. A three-electrode setup is employed to evaluate the electrocatalytic HER activities in 0.5 M H 2 SO 4 electrolyte. Note that the test is operated in a static state (without rotation) to simulate the real industrial operation. The bare GF shows negligible background activity for H 2 evolution (Fig. 5a), which is similar to other reports. 39,55,56 A reductive sweep of the MT@MS/GF sample shows a low overpotential (h) of 77 mV for HER, beyond which a sharp increase in the cathodic current starts to appear, attributed to electrocatalytic H 2 evolution (Fig. 5a). By contrast, the MT/GF catalyst exhibits inferior HER activity with lower catalytic current and larger onset overpotential of 161 mV.  In order to prove the presence of increased active sites provided by the hierarchical MT@MS/GF, the number of active sites for MT@MS/GF and MT@GF was estimated based on the voltammetry method (Fig. S4 †). 21,57,58 The related calculation method is presented in the ESI. † As shown in Table 1, the MT@MS/GF catalyst exhibits a density of active sites of 1.878 Â 10 À3 mol g À1 , which is obviously higher than that of the MT@GF catalyst (0.965 Â 10 À3 mol g À1 ). The larger number of active sites implanted in MT@MS/GF can be attributed to its hierarchical nanotubes with massive defects and the inherent reactive surfaces of nanotubes due to their curvature. It has been proved that MoS 2 morphology with a radius of curvature of nanoscale dimensions is expected to engender more exposed edge sites by imposing geometric constraints that can markedly change the thermodynamic stabilities of different surface features. 59 Another possible cause is that the hierarchical tubular structure possesses high surface area and diverse mesoporous channels, which are favourable for the fast ion diffusion and close electrode-electrolyte contact, leading to the enhanced electrocatalytic activity. The turnover frequency (TOF) for each active site of the MT@MS/GF catalyst was calculated to be 0.957 s À1 at h ¼ 300 mV, which is much higher than the TOF value of the MT@GF catalyst, indicating its better intrinsic catalytic activity. The HER kinetics of the various catalysts mentioned above was investigated through the corresponding Tafel plots (log j-h) (Fig. 5b). The linear portions in the Tafel plots were tted according to the Tafel equation (h ¼ b log|j| + a) and the Tafel slope (b) was obtained. A Tafel slope of 52 mV per decade was obtained for the MT@MS/GF sample (Fig. 5b), which is far lower than that of 94 mV per decade for the MT/GF sample (Table 1), testifying the superior HER kinetics of the MT@MS/GF sample. At high current densities, the MT@MS/GF sample also acts as a more efficient catalyst than MT/GF. The inherent HER activities of these catalysts were estimated with the exchange current density (j 0 ). The MT@MS/GF sample exhibits a j 0 of 6.4 Â 10 À2 mA cm À2 , outperforming the value of 2.5 Â 10 À2 mA cm À2 for the MT/GF sample. The outstanding kinetic metrics (low onset overpotential of 77 mV and Tafel slope of 52 mV per decade) and large j 0 (6.4 Â 10 À2 mA cm À2 ) highlight the extraordinary hydrogen evolution efficiency of this novel structured MT@MS/GF catalyst. Stability is another signicant criterion to evaluate the catalysts. The negligible difference in the J-V curves before and aer 3000 CV cycles demonstrates the superior stability of the MT@MS/GF sample in a long-term electrochemical reaction (Fig. 5c). To probe the durability of the MT@MS/GF sample in an acidic environment, continuous HER measurement at a static overpotential and long-term cycling test was conducted. The curve of time dependence of current density shows a typical serrate shape under a constant potential of À0.17 V vs. RHE (Fig. 5d), which is probably related to the alternative processes of bubble accumulation and release. For the hierarchical MT@MS/GF sample, the current density decreases gradually in the initial 0.5 h, and then increases slightly over 4 h of continuous operation. It is hypothesized that the activation process during the initial 0.5 h can allow more electrolyte to access the MT@MS/GF interfaces by capillary penetration due to the hierarchical structure, leading to the increased HER current density. Besides, XPS and Raman studies of the sample aer 4 h of operation revealed no obvious chemical state change of HER-active Mo (Fig. S5 †). In comparison, under the same measurement conditions, the MT/ GF catalyst exhibited a slow but progressive decrease in HER activity (Fig. S6 †). The extraordinary long-term durability of the tubular MT@MS/GF catalyst suggests its promise as a realistic hydrogen evolution electrode. Meanwhile, the hierarchical tubular structures of the MT@MS/GF electrode are still detectable even aer long-term HER cycling, indicating the good stability of its hierarchical tubular structure (Fig. S7 †).
The experimentally measured high catalytic HER activities (j 0 ¼ 6.4 Â 10 À2 mA cm À2 ) of the MT@MS/GF catalyst prompted us to investigate the mechanism responsible for the enhanced activity. Previous density functional theory (DFT) calculations indicated that the MoS 2 basal plane was inactive while only edge sites of nanoscale particles with coordinated unsaturated Mo atoms were active for HER, which was later proved by Jaramillo   resulted in partial cracking of the catalytically inert basal planes, leading to the exposure of additional active edge sites. 40 Herein, our MT@MS/GF sample with high specic surface area and diverse mesopores possesses cracked MoS 2 nanotubes with broken layers and massive defects can extend massive exposed active edge sites, leading to the enhancement of the HER activity. Also, a mass of ultra-small few-layered MoS 2 sheets covers the entire surfaces of nanotubes, which provides accessible catalytic centres for HER. Furthermore, the entire surfaces of nanotubes are regarded to be reactive due to the curvature and a distortion of the bond angles, leading to a preferentially maximum exposure of the catalytically active sites. Meanwhile, the graphene scaffolds may also boost the HER-activity of MT@MS due to the increased electrical conductivity, leading to fast electron transfer and high electrocatalytic efficiency towards HER. For HER in acidic media, two separate pathways (the Volmer-Tafel and the Volmer-Heyrovsky mechanism) have been proposed for reducing H + to H 2 . The Volmer-Heyrovsky mechanism involves two principal steps, referred to as the Volmer [eqn (3) The Tafel slope, an intrinsic property of electrocatalysts, could be used to probe the rate-determining step and elementary steps involved in the H 2 evolution. In a recent study, under a specic set of conditions, it has been reported that a Tafel slope of about 120, 40 or 30 mV per decade will be obtained if the Volmer, Heyrovsky or Tafel reaction is the rate-determining step, respectively. 63 Ideally, the Tafel slope is an inherent property of electrocatalytic materials, which is a useful indicator of the rate-limiting step for reactions involving electron transfer. In reality, however, the Tafel slope can be dependent on many factors other than the kinetic exponent of the electrons, such as the coverage of adsorbates and the mass transport effect in porous structures. 64 Since our Tafel slope was determined in the high coverage region, the value of 52 mV may imply a Heyrovsky rate-deterring step. Due to the complexity of the reaction mechanism, analyzing the Tafel slope is still inconclusive for identifying the HER of MoS 2 . However, the Tafel slope of 52 mV per decade obtained in this work is comparable to the value of 36-68 mV per decade of bulk Pt (shape-dependent) in 0.5 M H 2 SO 4 and close to that of defect-rich MoS 2 ultrathin nanosheets reported previously, which suggests a similar surface chemistry of our hierarchical MT@MS/GF. 40 In this study, we also investigate the electrochemical properties of MT@MS/GF as an anode material for LIBs. The CV curves of MS@MT/GF in the rst four cycles are presented in Fig. 6a, which are similar to those reported. 65,66 The shape of the CV curve and the oxidation/reduction peaks are still retained in the cycles, indicating the good stability of the anode. The hierarchical MS@MT/GF delivers initial discharge and charge capacities of 1487 and 1215 mA h g À1 , respectively, giving a coulombic efficiency of 81.7% (Fig. 6b). The coulombic efficiency rapidly reaches more than 99% aer the second cycle. The second and third discharge proles almost coincide with each other, proving the excellent cycling performance. The hierarchical MS@MT/GF exhibits excellent rate performance. At high current densities of 1 and 2 A g À1 , specic capacities of 1025 and 916 mA h g À1 can be maintained, respectively (Fig. 6c). Even when cycled at a high current density of 5 A g À1 , a specic capacity of 878 mA h g À1 can be maintained. Remarkably, these values are much higher than that of MT/GF (731 mA h g À1 ). The cycling performance is also outstanding: even aer 200 cycles, both discharge and charge capacities of the hierarchical MS@MT/GF can remain at 892 mA h g À1 at a current density of 500 mA g À1 , delivering nearly 76.5% of capacity retention (Fig. 6d). In contrast, the MT/GF electrode shows lower capacities and much faster capacity fading under the same testing conditions, and a capacity of only $727.4 mA h g À1 is retained aer 100 cycles (Fig. S8 †). A complete comparison of the electrochemical properties of these two samples, summarized in Table S3, † shows that MS@MT/GF possesses excellent electrochemical properties, including high initial coulombic efficiency, high specic capacity, excellent rate performance and stable cyclability. The remarkable lithium storage properties of the hierarchical MT@MS/GF catalyst are probably related to its unique tubular architecture constructed by MoS 2 nanosheetsdecorated nanotubes. Especially, the hierarchical tubular structure possesses high surface area and massive mesopores, which are favorable for a short diffusion distance of lithium ions and a large electrode-electrolyte contact area, leading to the enhanced rate capability. Meanwhile, the decorating ultrathin 2D nanosheets and the inner void space in the hollow tubular structure not only allow fast lithium ion diffusion, but also effectively buffer the mechanical stress caused by volume variation during the lithium intercalation/exfoliation, thus leading to the enhanced cycling stability. The hierarchical tubular structures in the MT@MS/GF electrode are still sustained with little structural deformation even aer long-term cycling, indicating their good stability (Fig. S9 †).

Conclusions
In summary, we have developed a facile method to prepare robust MoS 2 hollow nanotubes implanted with high density of active sites that exhibit excellent catalytic activity for HER. The obtained hierarchical nanotubes are constructed by MoS 2 hollow nanotubes decorated with MoS 2 nanosheets, which are self-supported on exible 3DGF. The hierarchical hollow nanotubes show enhanced lithium storage and excellent electrocatalytic activity towards HER with quite good stability, arising from preferentially exposed more catalytically active edge sites. In particular, the hierarchical tubular structure possesses high surface area and diverse mesoporous channels, which are favorable for fast ion diffusion and close electrodeelectrolyte contact, leading to the enhanced electrocatalytic activity and lithium storage. All these merits undoubtedly contribute to the enhancement of electrochemical properties, including a low onset overpotential of 77 mV, Tafel slope of 52 mV per decade and large exchange current density of 6.4 Â 10 À2 mA cm À2 . In addition, when evaluated as an anode material for LIBs, these hierarchical MoS 2 nanotubes manifest high specic capacity and excellent rate capability as well as extremely longterm cycle stability. This method can be used for the preparation of other tubular structures for future clean energy applications.