Introduction

Lithium-ion batteries (LIBs) have been proved to be the main power source for portable electronic devices and hybrid electric vehicles due to their advantages of high energy density, long cycling life and environmental benignity [1,2,3]. Nevertheless, the theoretical capacity of commercialized graphite anode is just as low as 372 mAh g–1, which leads to the fact that the presently available LIBs are unable to further satisfy the requirements of large-scale energy storage [4, 5]. Consequently, considerable efforts have been devoted to develop the innovative anode materials for high-performance LIBs. Nowadays, transition metal sulfides, such as Ni3S2, NiS, NiS2, CoS2 and Co9S8 [6,7,8,9], have been regarded as possible candidates and attracted tremendous attention due to their multistep reversible redox reactions, which provide high theoretical capacity by making full use of all oxidation states of transition metals to achieve storage conversion reactions mechanism. Among them, Ni3S2 with the merits of superior theoretical capacity (446 mAh g–1), abundant resources and low cost shows great potentiality as a favorable anode of LIBs [10, 11]. Nevertheless, both its inferior electronic conductivity and severe volume change during the cycling all lead to extremely poor rate performance and fast capacity fading. The common strategies for solving the above drawbacks include coating with carbon layer, designing various Ni3S2 nanostructures, anchoring onto a conductive substrate, constructing metal sulfide composites and so on [12, 13]. These strategies can enhance lithium storage capability of electrode materials. For instance, Li et al. had successfully fabricated a heterostructure of NiO nanosheet array grown on β-NiS@Ni3S2 framework by a simple hydrothermal synthesis. The prepared NiO@β-NiS@Ni3S2 composite exhibits excellent lithium storage as the anode material of LIB [14]. Yang et al. reported a novel heterogeneous structure of NiO/Ni3S2 nanoflake composite with a carbon nanofiber (CNF) membrane. The binderless NiO/Ni3S2/carbon nanofiber electrode shows the excellent lithium storage capability [15]. Although the advancement in the LIB anodes of Ni3S2-based materials had been achieved, investigating the novel electrodes with superior capacity and cycle capability remains an immense challenge of high-performance LIBs.

In recent years, carbon nanowalls (CNWs) are an emerging material that is vertically aligned and assembled from carbon nanosheets. The preferable conductivity and three-dimensional (3D) porous structure of carbon nanowalls not provide the inherent advantages of two-dimensional carbon nanosheet, prevent the agglomeration of nanoparticles, but enhance the permeability of electrolyte and thus shorten the distance of Li-ion diffusion, especially in an abundant mesoporous structure [16, 17]. Consequently, in situ combining transition metal sulfide and carbon nanowalls to prepare the composite structure were confirmed to be an efficacious approach to optimize the rate capability and cycle stability of electrodes.

We, herein, have elaborated an in situ encapsulation of Ni3S2 nanoparticles into carbon nanowalls with 3D porous structures through a cation-adsorption approach and a hydrothermal synthesis. In this synthesis process, cation exchange resin as a carbon source can sufficiently absorb nickel ions. The catalytic nickel ions are used to catalyze the growth of carbon nanowalls during the heat treatment process. Equally important, the alkaline KHCO3 acts as a pore-forming agent not only determines the contents of Ni3S2 and CNWs, but also enables the cation exchange resin to form large surface area and abundant mesoporous structure. Therefore, the adsorbed nickel ions and the pore-forming agent of KHCO3 both play important roles in the formation of Ni3S2/CNWs. The 3D-interconnected porous carbon nanowalls, as an electrically conductive and buffered matrix, enhance the electrochemical activity and structural stability of the Ni3S2, and effectually prevent the pulverization and aggregation of Ni3S2 nanoparticles. The Ni3S2/CNWs-1 composite shows remarkable electrochemical performance, which is ascribed to the synergistic combination of well distributed Ni3S2 nanoparticles and conductive carbon nanowalls. In this contribution, our ingenious approach not only highlights an effective strategy to stabilize metal sulfides but also paves the way for the further improvement of lithium-ion battery composite electrode materials.

Experimental section

Synthesis of Ni/CNWs precursor and CNWs

All related reagents were used without further purification and purchased from Macklin. The Ni/CNWs precursor was synthesized by a cation adsorption approach. Typically, 0.04 mol of nickel acetate is dissolved in 400 mL of deionized water, followed by the addition of 20 g of cation exchange resin and magnetic stirring for 8 h. Thereafter, the resin was dried at 80 °C overnight. The dried product was pulverized into powder by a pulverizer, then transferred to a solution with 60 g KHCO3 in 300 mL ethanol and heated to 90 °C to remove moisture. The obtained product was heated to 850 °C in a nitrogen atmosphere and kept for 1 h. After complete cooling, the Ni/CNWs precursor was cleaned to neutral repeatedly with ethanol and deionized water and then dried overnight in a vacuum drying oven. In addition, the CNWs sample was obtained by adding Ni/CNWs precursor in 3 M hydrochloric acid solution and stirring strongly for 10 h to completely remove impurities and nickel elements.

Preparation of Ni3S2/CNWs and Ni3S2

Ni3S2/CNWs was done by a hydrothermal process. In general, 0.25 g of Ni/CNWs precursor and 16 mmol of thiourea were added to 80 ml of deionized water. After stirring well, the solution was transferred to an autoclave followed by heating at 180 °C for 16 h. The cooled product was filtered several times with ethanol and deionized water and then dried in a vacuum-drying oven overnight. Eventually, in order to increase the crystallinity, the sample was heated to 500 °C under a nitrogen atmosphere and kept for 2 h to obtain the Ni3S2/carbon nanowalls (designated as Ni3S2/CNWs-1).

As a comparison, Ni3S2/CNWs-2 was prepared by the same synthesis method, except increasing the amount of KHCO3 to 120 g. Additionally, Ni3S2 sample was prepared by the above method except that 24 mmol of nickel acetate was used instead of 0.25 g Ni/CNWs precursor.

Physical characterization

The crystal structures were studied by X-ray diffraction (XRD, Rigaku Corporation SmartLab Studio II) with a Cu Kα radiation. The morphologies and microstructures were determined using a scanning electron microscope (SEM, Hitachi SU8220) and a transmission electron microscope (TEM, FEI Titan ETEM G2) with high-performance X-ray energy-dispersive spectrometer (EDS). Thermogravimetric (TGA, NETZSCH-Gerätebau GmbH Company) analysis was carried out from room temperature to 1000 °C in the air at a heating rate of 10 °C min–1 to determine the Ni3S2 content in Ni3S2/CNWs composite. Raman analyses were performed on a Raman spectrometer (Horiba Jobin Yvon Company) from 600 to 4000 cm–1 with a 532 nm laser. The elemental compositions and valence state were characterized by X-ray photoelectron spectroscopy (XPS, Thermo Fisher Scientific Corporation Escalab 250Xi) with a monochromatic Al Kα radiation. The ASAP-2460 surface area analyzer (Micromeritics Instrument Corporation) uses Brunauer–Emmett–Teller (BET) method to obtain specific surface area.

Electrochemical measurements

With the aim of testing the electrochemical properties of the samples, the CR2025 button-type half-cells were assembled in an argon-filled glove box (Mikrouna company). The active material, carbon black and poly(vinylidene fluoride) (PVDF) were mixed with N-methyl-2-pyrrolidinone (NMP) at a mass ratio of 8:1:1 to form a uniform slurry to prepare the working electrodes. The prepared slurry was coated on a Cu foil and dried at 70 °C overnight. The separator, current collector and counter electrode were Celgard 2400, copper foil and Li metal foil, respectively. The electrolyte was a mixture of 1 M LiPF6 dissolved in ethylene carbonate (EC)/dimethyl carbonate (DMC) (v/v = 1:1). For button-type full cell assembly, the commercial LiFePO4 cathode (GEELY Corp.) was used instead of the metallic lithium. The LiFePO4 cathode was prepared under the same procedure as the preparation of Ni3S2/CNWs-1 anode. For the purpose of matching the capacity of cathode and anode electrodes, the capacity of anode electrodes is slightly excessive compared with that of cathode electrodes, and the capacity of Ni3S2/CNWs-1 anode to LiFePO4 cathode was about 1.2:1. Before the preparation of the full cell, the pre-lithiation of Ni3S2/CNWs-1 anodes was carried out in five charge and discharge cycles. A Neware battery testing system performs charge and discharge measurements. Cyclic voltammetry (CV) curves were obtained in the potential range of 0.01–3.00 V at 0.1 mV s−1 and electrochemical impedance spectroscopy (EIS) was conducted in the frequency range of 105−0.01 Hz using a Zennium IM6 electrochemical workstation.

Results and discussion

The whole preparation route of Ni3S2/CNWs composite is shown in Fig. 1, which includes four steps: (1) Ni2+ cations were completely adsorbed to the pretreated cation exchange resin; (2) the dried resin was pulverized and then immersed in the KHCO3/ethanol mixture; (3) in the calcination process, Ni/CNWs precursor was fabricated by KHCO3 pore-forming agent; (4) the Ni nanoparticles were sulfurized into uniform Ni3S2 by a controlled hydrothermal reaction and then annealed to increase the crystallinity of Ni3S2/CNWs.

Figure 1
figure 1

Schematic diagram of the fabrication of Ni3S2/CNWs composites

Full size image

The XRD pattern of Ni3S2/CNWs-1 sample is displayed in Fig. 2a. The peaks at 21.8°, 31.2°, 37.9°, 44.5°, 49.9°,55.3°, 64.5°, 69.2°, 73.3° and 78.1° can be well indexed to the (100), (1–10), (111), (200), (210), (2–11), (220), (300), (310) and (1–31) planes of the rhombohedral Ni3S2 structure [18], and a weak peak at 26.2° is ascribed to the (002) plane of the hexagonal graphite structure, which are consistent with the XRD standard data cards of JCPDS No.73-0698 and JCPDS No.75-1621, respectively. The sharp and intense peaks signify highly crystalline of as-synthesized samples. In addition, Fig. S1 displays that all diffraction peaks of the Ni/CNWs precursor can be accurately indexed to the cubic Ni crystal of JCPDS No.87–0712, indicating that the precursor can be successfully converted to Ni3S2 by the vulcanization process. All diffraction peaks of the control sample Ni3S2 (Fig. S2) were identical to Ni3S2 in Ni3S2/CNWs samples.

Figure 2
figure 2

a XRD patterns; b, c SEM images; d low-resolution TEM images and e, f HRTEM images of Ni3S2/CNWs-1 composite (the inset showing the corresponding FFT pattern of Ni3S2); g elemental mapping results of Ni3S2/CNWs-1

Full size image

The morphology and microstructure of Ni3S2/CNWs-1 composites were investigated by FE-SEM. The Ni3S2 nanoparticles (Fig. 2b, c) are uniformly embedded in highly conductive porous carbon nanowalls. It is further observed from the low-resolution TEM image (Fig. 2d) that the Ni3S2 nanoparticles are uniformly dispersed in 3D-interconnected carbon nanowalls, which can significantly strengthen the electrical conductivity and structural stability of Ni3S2/CNWs-1. In the Ni3S2/CNWs-1 composite, the carbon nanowall is composed of 6 layers of 1.7-nm-thick carbon layer, as displayed in the higher-resolution TEM in Fig. 2e, f. The (200) crystal plane of the Ni3S2 phase and (002) face of the graphitic carbon (Fig. 2f) is corresponding to the interplanar spacing of 0.20 and 0.34 nm, respectively, which are in accordance with the XRD results. The polycrystalline nature of Ni3S2 is reflected by the corresponding fast Fourier transformation (FFT) diffraction pattern (the inset of Fig. 2f). Furthermore, the selected area electron diffraction (SAED) pattern (Fig. S3) of Ni3S2/CNWs-1 composite shows three diffraction rings, which can be assigned to the (1–10), (1–21) and (321) crystal planes of Ni3S2, respectively. The morphology of Ni/CNWs precursor (Fig. S4) is identical to that of Ni3S2/CNWs-1, which indicates that the original structure is preserved perfectly during the hydrothermal synthesis procedure. The morphology of Ni3S2/CNWs-2 (Fig. S5) is almost homologous to the Ni3S2/CNWs-1. The control sample Ni3S2 (Fig. S6) shows an irregular shape. From the SEM images (Fig. S7), it is observed that CNWs presents a crumpled and 3D-interconnected porous structure, which may be attributed to the production of carbon dioxide and water during the pyrolysis of pore-forming agent of KHCO3. The EDS element mapping of the TEM image is displayed in Fig. 2g, from which we can observe that the C, Ni and S elements are homogeneously dispersed in the Ni3S2/CNWs-1 composite. Combined with XRD and HRTEM results, Ni3S2/CNWs-1 was prepared successfully. Additionally, element analysis of Ni3S2/CNWs-1 and Ni3S2/CNWs-2 composites was investigated via employing EDS spectroscopy, and the results are shown in Fig. S8. The atomic percentages of carbon, nickel and sulfur elements in Ni3S2/CNWs-1 are 88.64, 7.07 and 4.29 at%, respectively. The molar ratio of nickel to sulfur is about 3:2 and further confirms the formation of pure Ni3S2, which also coincides with XRD result. The contents of Ni3S2 and CNWs in the Ni3S2/CNWs-1 composite was determined by TGA analysis. As shown in Fig. 3a, the initial slight weight loss below 400 °C can be attributed to the evaporation of water and the removal of residual organics, and the subsequent weight loss from 400 to 650 °C corresponds to the consumption of CNWs and the conversion of Ni3S2 to NiO. According to the 26.9 wt% of the original mass percentage remained after TGA analysis, the contents of Ni3S2 and CNWs in the Ni3S2/CNWs-1 sample are calculated to be 28.8 and 71.2 wt%, respectively, and the corresponding calculation process is described in the Supporting Information.

Figure 3
figure 3

a TGA curve of Ni3S2/CNWs-1; b Raman spectra of CNWs, Ni/CNWs and Ni3S2/CNWs-1 samples; c XPS survey spectrum and d−f high-resolution XPS spectra of C 1s, Ni 2p, and S 2p of Ni3S2/CNWs-1

Full size image

The Raman spectrum of Ni3S2/CNWs-1 (Fig. 3b) shows two representative scattering vibrational modes at 1313 and 1566 cm–1, which can be ascribed to the D band (defects) with disorder characteristics and the G band (sp2 hybridization of carbon atoms) with graphite characteristics, respectively [19, 20]. ID/IG is the intensity ratio of the D band and G band, which reflects the surface defects of the active material and the distortion of the carbon lattice. Consequently, for the Ni3S2/CNWs-1 composite, the Ni/CNWs precursor and the control sample CNWs, the ID/IG values are determined to be 1.02, 0.92 and 0.82, respectively, indicating that the addition of sulfur source induces surface defects and produces sufficient Li-ion storage sites [16, 21].

The element composition and chemical valence state of the Ni3S2/CNWs-1 were evaluated by XPS analysis. Figure 3c shows the survey spectrum of the presence of nickel, sulfur, carbon and oxygen elements. Specifically, the nickel, sulfur and carbon elements are ascribed to the Ni3S2/CNWs composite, while the oxygen element is mainly attributed to the exposure of the sample to the air. The four major peaks of high-resolution C 1s spectrum (Fig. 3d) at 284.4, 285.1, 286.8, and 290.0 eV can be assigned to C–C/C=C, C–S, C–O, and C=O/O–C=O, respectively [22, 23]. The C–S covalent bond in the carbon nanowalls further confirmed that carbon host combines with some sulfur atoms to form a defective site, which is also consistent with the results of Raman spectroscopy analysis. Deconvolution of Ni 2p spectrum (Fig. 3e) produces two major strong peaks at 856.0 and 873.5 eV, which are ascribed to Ni 2p3/2 and Ni 2p1/2, respectively, demonstrating the trivalent valence of Ni in Ni3S2. And the intense peaks demonstrate that Ni3+ is the majority. Furthermore, the two peaks of Ni 2p3/2 and Ni 2p1/2 are located at 853.2 eV and 870.3 eV, respectively, indicating that the Ni element in Ni3S2 is a divalent state, and the weak peaks at 861.4 and 878.5 eV are satellite peaks [24, 25]. The two strong peaks of S 2p spectrum (Fig. 3f) at 161.7 and 162.9 eV are ascribed to S 2p3/2 and S 2p1/2, respectively, which are attributed to the metal-sulfur (M-S) bonds of Ni3S2. The peak with binding energy at 164.6 eV is associated with the C–S–C, revealing the perfect combination of sulfur atoms with conductive carbon nanowalls. Additionally, the peak at 168.6 eV corresponds to a satellite peak [26, 27]. The BET test was used to study the specific surface area and pore size distribution of Ni3S2/CNWs-1. The N2 adsorption–desorption isotherm of Ni3S2/CNWs-1 (Fig. S9a) exhibits a type IV with hysteresis loop, which is attributed to the presence of mesoporous characteristics in the composites [28, 29]. Note that the specific surface area of Ni3S2/CNWs-1 can be obtained from BET analysis is 484 m2 g–1. Meanwhile, the mesoporous structure can be further observed from the pore size distribution (Fig. S9b). In contrast, the specific surface areas calculated by the BET method for Ni3S2/CNWs-2, CNWs and Ni3S2 were 446, 490 and 86 m2 g–1, respectively. The abundant mesopores and suitable specific surface area are beneficial for the electrolyte immersion into the active material and enhance its electrochemical performance in the process of charging and discharging.

To clarify the electrochemical superiority of Ni3S2/CNWs-1 composite, we conducted the comparative study of Ni3S2/CNWs-2, CNWs and Ni3S2 in Fig. 4. We investigated the electrochemical behaviors of electrodes by performing Cyclic voltammetry (CV) tests within the potential range of 0.01–3.00 V. Figure 4a displays the initial three consecutive CV curves of the Ni3S2/CNWs-1 electrode at 0.1 mV s−1. Two obvious peaks in the first cathodic scan at 1.31 and 1.61 V, which can be ascribed to the insertion of Li+ into the electrode material and the reduction of Ni3S2 to Ni (4Li+  + 4e− + Ni3S2 = 3Ni + 2Li2S). Also, it is worth noting that the peak at 0.66 V is generally attributed to the decomposition of organic electrolyte and the formation of solid electrolyte interface (SEI) layer. The main peak in the potential range of 0.02–0.15 V is related to lithium ions embedded in the carbon lattice, so this electrode material has lithium storage activity, which coincides with the trend of the CV curve of the CNWs electrode (Fig. S11a). It can be observed from the first anodic scan that the strong peak appearing at 1.96 V is due to the de-intercalation of Li+ and the electrochemical decomposition of Li2S followed by the oxidation of Ni to Ni3S2 (3Ni + 2Li2S = 4Li+  + 4e− + Ni3S2) [30]. In the second cathodic scan, it can be directly observed that the peak of 0.66 V vanished, while the peak of 1.61 V moved to approximately 1.76 V, which was attributed to the irreversible phase transition of Ni3S2/CNWs-1 electrode material after the first lithium ion intercalation. After the second cycle, all redox peaks perfectly overlapped for the Ni3S2/CNWs-1 and CNWs electrodes, demonstrating the excellent stability of the SEI layer. Nevertheless, from the CV curves of Ni3S2/CNWs-2 (Fig. S10a) and Ni3S2 (Fig. S12a), it was observed that all redox peaks did not overlap, suggesting the inferior reversible performances.

Figure 4
figure 4

a CV curves and b charge–discharge curves of the Ni3S2/CNWs-1 anode at 1st, 2nd and 3rd cycle; c cycle performance at 200 mA g–1; d rate performance, and e super-long cycling performance at 2 A g–1 for the Ni3S2/CNWs-1, Ni3S2/CNWs-2, CNWs and Ni3S2 anodes

Full size image

The galvanostatic charge and discharge curves of the Ni3S2/CNWs-1 anode in the first three cycles when the current density is 200 mA g−1 are displayed in Fig. 4b. For the first cycle, the high discharge and charge capacities of Ni3S2/CNWs-1 were 1616 and 1056 mAh g−1, respectively, and the corresponding coulombic efficiency was calculated to be 65.3%. From the charge and discharge curves, we can clearly observe a representative phenomenon, that is, the electrochemical reaction-driven electrolyte degradation to form a thin SEI film, which directly causes the loss of irreversible capacity. In the first discharge curve, a voltage plateau of 1.31 V is closely related to the reduction of Ni3S2 to Ni. For the first charge curve, a voltage plateau of 1.96 V is in connection with the oxidation of Ni to Ni3S2 during the conversion reaction. All charge–discharge voltage plateaus of the Ni3S2/CNWs-1, Ni3S2/CNWs-2 (Fig. S10b), CNWs (Fig. S11b) and Ni3S2 (Fig. S12b) anodes are consistent with the redox peaks of CV curves.

Figure 4c presents the cycle stability of the Ni3S2/CNWs-1 anode when the current density is 200 mA g–1. Before the initial thirty cycles, the slow decay of capacity may be caused by the irreversible insertion of lithium into ultrafine pores. Interestingly, the discharge capacity of Ni3S2/CNWs-1 anode after cycling 200 times was 906 mAh g–1 and the coulombic efficiency reached 99.6%. By contrast, the discharge capacities of Ni3S2/CNWs-2, CNWs and Ni3S2 anodes after 200 cycles at 200 mA g–1 were 585, 570 and 355 mAh g–1, respectively. These obvious differences can be explained by the following elaboration. The continuous and severe capacity decay of the Ni3S2/CNWs-2 and Ni3S2 electrodes during cycling may be ascribed to their inherent inferior conductivity and the inevitable aggregation and pulverization of Ni3S2 nanoparticles. The large electrode–electrolyte contact area of Ni3S2/CNWs-1 and CNWs anodes possess the 3D-interconnected porous structures, which can efficaciously inhibit the volume expansion of Ni3S2 in the electrochemical cycling, so these two electrodes exhibit better cyclic reversibility. For Ni3S2/CNWs-1 electrode, although Ni3S2 only possesses a theoretical capacity of 446 mAh g–1, the combination effect of reversible capacity of CNWs and Ni3S2 enhances the lithium storage performance of Ni3S2/CNWs-1 electrode.

The rate capability is also critical for evaluating the electrochemical properties of the anodes. Figure 4d displays the rate capability of Ni3S2/CNWs-1 anode when the current density changes from 0.2 to 5 A g–1. The average reversible capacities of the Ni3S2/CNWs-1 were 905, 819, 729, 643, and 576 mAh g–1 when the current densities were 0.2, 0.5, 1, 2, and 5 A g–1, respectively. It is worth noting that the reversible capacity of the Ni3S2/CNWs-1 anode is restored to 902 mAh g–1 when the current density is restored to 0.2 A g–1, which is better than Ni3S2/CNWs-2 (616 mAh g–1), CNWs (662 mAh g–1) and Ni3S2 (405 mAh g–1) anodes, further illustrating its outstanding reversibility. Figure S13a–d shows the corresponding charge–discharge profiles of the four anodes at various current densities.

The ultralong cycling performance of Ni3S2/CNWs-1 anode at 2 A g–1 is displayed in Fig. 4e. The initial discharge and charge capacities of Ni3S2/CNWs-1 anode are 1296 and 862 mAh g–1, respectively. Also, the discharge capacity of Ni3S2/CNWs-1 (666 mAh g–1) anode after 1000 cycles is higher than that of Ni3S2/CNWs-2 (407 mAh g–1), CNWs (395 mAh g–1) and Ni3S2 (192 mAh g–1) anodes, and the capacity decay rate of each cycle is only 0.016%. Moreover, the coulombic efficiency of Ni3S2/CNWs-1 anode is as high as 99.6%, showing the excellent long-life cycle stability. Accordingly, the reversible capacity and cycle stability of the Ni3S2/CNWs-1 anode material are enhanced due to the synergistic combination of Ni3S2 nanoparticles with high theoretical capacity and carbon nanowalls with 3D porous structures. In addition, the well-designed mesoporous structure facilitates rapid transmission of Li+ and electrons, so the Ni3S2/CNWs-1 anode exhibits outstanding lithium storage properties. It can be observed that the enhanced lithium storage performance of Ni3S2/CNWs-1 composite is higher than that of other previously reported nickel sulfide anode materials (Table S1).

The superior electrochemical performance of Ni3S2/CNWs-1 anode is closely connected with its unique merits, as shown below: (1) the Ni3S2 nanoparticles are in situ encapsulated in the carbon nanowalls with 3D-interconnected porous structures, which can efficiently alleviate the larger volume expansion of Ni3S2 particles during cycling. (2) The abundant mesoporous structure is produced by the pyrolysis of pore-forming agent KHCO3, which can provide a larger contact area and shorten the transmission route of electrons and Li+ to enhance the kinetics. (3) The synergistic effect between uniformly dispersed Ni3S2 nanoparticles and conductive carbon nanowalls can greatly improve reversible capacity and cycle stability. As further evidence, the SEM and TEM images of Ni3S2/CNWs-1 (Fig. S14) and Ni3S2/CNWs-2 (Fig. S15) anodes were shown after 200 cycles at 200 mA g–1. We can clearly observe that the overall morphology of the Ni3S2/CNWs-1 anode remains almost unchanged after cycling without obvious structural collapse, and the Ni3S2 nanoparticles and carbon nanowalls can also be distinctly identified, which proves the structural integrity of the Ni3S2/CNWs-1 composite.

Charge storage kinetics analysis based on CV tests was carried out at various scan rates to explore the lithium storage mechanism of Ni3S2/CNWs-1 (Fig. 5a−e). Figure 5a shows the CV curves of Ni3S2/CNWs-1 anode at the scanning rate of 0.1–1.0 mV s–1. We can observe that the CV curve is almost free of distortion. Generally, the CV curves of different scanning rates are analyzed in combination with the following equations to calculate the contribution ratio of faradic and non-faradic (pseudocapacitive) to lithium storage [19, 31, 32]:

$$i = av^{{\text{b}}}$$
(1)
$${\log}\left( i \right) = b{\log}\left( v \right) + {\log}\left( a \right)$$
(2)
$$i = k_{{1}} v + k_{{2}} v^{{{1}/{2}}}$$
(3)

where i is the peak current, v is the scanning rate, and a, b, k1, and k2 are tunable parameters. By analyzing Eqs. (1) and (2), it can be concluded that b = 1 and 0.5 correspond to the contribution of pseudo-capacitance and ion-diffusion, respectively [33, 34]. Figure 5b shows the b values of 1 (0.70) and 2 (0.81) peaks of Ni3S2/CNWs-1, indicating that the electrochemical reaction involves pseudo-capacitance contribution. Based on Eq. (3) and the analyzed k1 and k2 constants (Fig. 5c), we calculate the detailed pseudo-capacitance contribution at 0.7 mV s−1 (Fig. 5d). Specifically, when the scanning rates are 0.1, 0.3, 0.5, 0.7, and 1.0 mV s−1, the pseudo-capacitance contributions of the Ni3S2/CNWs-1 anode (Fig. 5e) are 35.2%, 43.5%, 57.9%, 65.9%, and 76.6%, respectively, which manifests that the electrochemical charge–discharge process is dominated by the pseudocapacitive lithium storage, which also indicates that favorable capacitive kinetics of the Ni3S2/CNWs-1 anode contributes to outstanding lithium storage.

Figure 5
figure 5

a CV curves of Ni3S2/CNWs-1 anode at various scanning rates; b relationship of log(i) versus log(v) of Ni3S2/CNWs-1 anode; c plots of v1/2 versus i/v1/2 at different redox states for obtaining k1 and k2 constants; d faradic (white) and non-faradic (green) contributions at 0.7 mV s−1; e the pseudocapacitive contributions of the Ni3S2/CNWs-1 at various scanning rates; f Nyquist plots of the Ni3S2/CNWs-1, Ni3S2/CNWs-2, CNWs and Ni3S2 anodes after 200 cycles at 200 mA g−1 (inset: the equivalent circuit was fitted by EIS)

Full size image

The electrochemical impedance spectroscopy (EIS) tests of Ni3S2/CNWs-1, Ni3S2/CNWs-2, CNWs and Ni3S2 anodes before (Fig. S16) and after 200 cycles (Fig. 5f) were performed to investigate the kinetic properties of electron/ion diffusion. In the Nyquist plots, the semicircle in the high-frequency region and the sloped straight line in the low-frequency region are associated with the charge-transfer resistance (Rct) and the diffusion resistance of lithium ions (Warburg impedance) [35,36,37]. The Nyquist plots of the four anodes are fitted by the equivalent circuit (the inset of Fig. 5f). As displayed in Fig. S16, for the fresh half-cell, the charge-transfer resistance of the Ni3S2/CNWs-1 (58.0 Ω) anode is lower than that of Ni3S2/CNWs-2 (131.7 Ω), CNWs (64.4 Ω) and Ni3S2 (151.9 Ω), revealing its better conductivity. Additionally, after 200 cycles, the resistance of Ni3S2/CNWs-1 anode (Fig. 5f) at 200 mA g−1 was 55.0 Ω, which was smaller than that before the cycle, and also lower than that the Ni3S2/CNWs-2 (99.5 Ω), CNWs (115.4 Ω) and Ni3S2 (76.0 Ω) anodes after the cycle, further indicating its excellent conductivity. From the Nyquist plots, we can observe that the slope of the inclined line of Ni3S2/CNWs-1 anode is larger than that of the other three comparative anodes, suggesting its higher Li+ diffusion coefficient, thus promoting the superior reaction kinetics and cycling performance in the electrochemical reaction.

The superior electrochemical performances of the aforementioned Ni3S2/CNWs-1 anode inspire us to study its practicability in full cell. According to the schematic illustration displayed in Fig. 6a, the button-type full cell is assembled by using Ni3S2/CNWs-1 anode and LiFePO4 cathode. Figure S17 displays the electrochemical performance of LiFePO4 cathode, from which we can observe that the first charge capacity is 142 mAh g–1. The charge–discharge curves of Ni3S2/CNWs-1║LiFePO4 full cell (Fig. 6b) at 100 mA g–1 under the potential window of 0.5–3.6 V. The button-type full cell delivers the charge and discharge capacities of 135 and 131 mAh g–1 in the first cycle, respectively. After cycling 200 times, the reversible capacity of full cell was 113 mAh g–1 and the corresponding capacity retention rate was 83.7% (Fig. 6c), indicating excellent cycle stability. It is attractive that the button-type Ni3S2/CNWs-1║LiFePO4 full cell (Fig. 6d) can power 33 light-emitting diodes, demonstrating its potential for practical applications.

Figure 6
figure 6

a Schematic diagram of the Li-ion full cell with Ni3S2/CNWs-1 as anode and LiFePO4 as cathode; b charge–discharge curves of Ni3S2/CNWs-1║LiFePO4 full cell at 0.1 A g–1; c cycling performance of Ni3S2/CNWs-1║LiFePO4 full cell; d The LEDs are illuminated by the as-assembled full cell

Full size image

Conclusions

In summary, an in situ encapsulation of Ni3S2 nanoparticles into carbon nanowalls with 3D porous structures have been fabricated successfully via a cation adsorption approach and a hydrothermal method. The Ni3S2/CNWs-1 anode shows good lithium storage capacity, ultra-stable cycle performance and superior rate performance for rechargeable LIBs. In addition, the assembled button-type Ni3S2/CNWs-1║LiFePO4 full cell delivers a reversible capacity of 113 mAh g–1 and a capacity retention rate of 83.7% after 200 cycles at 100 mA g–1. Such excellent lithium storage properties are in connection with the unique advantages of the Ni3S2/CNWs-1 anode. Firstly, the carbon nanowalls with 3D-interconnected porous structures can provide the larger contact area and effectively inhibit the aggregation of Ni3S2 nanoparticles. Secondly, the Ni3S2 nanoparticles possess a high theoretical capacity due to its high lithium storage activity based on the conversion mechanism. Thirdly, the synergistic effect of Ni3S2 nanoparticles and conductive graphene sheets can greatly improve cycle stability. This work demonstrates that the as-prepared Ni3S2/CNWs-1 anode is a prospective high-performance LIB material.