Background

Due to its low lithium intercalation potential as well as excellent cycling performance, graphite has been widely adopted as a commercial anode for lithium-ion batteries (LIBs) [1]. Nevertheless, the theoretical capacity of graphite is only 372 mAh g−1, which cannot fulfill the ever-growing demands for high-performance batteries. Therefore, the development of next-generation anode materials with a larger specific capacity is necessary [2, 3].

Due to a large theoretical capacity of 1965 mAh g−1 and a low electrochemical potential, SiO2 is considered as a potential alternative to traditional carbonaceous anode materials. Furthermore, environmental friendliness, low cost, and natural abundance make SiO2 a commercial viable electrode material for LIBs. However, its practical application in LIB is commonly hampered by its poor electronic conductivity as well as a drastic volume variation upon charge-discharge process, resulting in particle pulverization and electrode deterioration with cycling [4,5,6].

One of the effective approaches to overcome these issues is to design SiO2-based composites by confining SiO2 particles inside conductive and flexible matrixes [7, 8]. In our previous study, Cu/carbon was introduced into the SiO2 composite as a dispersive matrix due to its good conductivity and effective buffering of the volume change of SiO2 [9]. It was shown by Yu et al. [10] that coating the SiO2 surface with carbon could be an efficient method to enhance its electrochemical performance, because such coating not only improves conductivity of the system but also accommodates the volume changes of the active material upon cycling.

Considering that the contact between SiO2@C particles is not good enough and the SiO2@C particles tend to agglomerate during charge/discharge [11] in this work, we report an effective and easy method to synthesize a core-shell SiO2@C anchored on MWNT via a sol-gel and pyrolysis route. In this composite, a carbon layer is homogeneously coated on the SiO2 particles, significantly improving the electronic conductivity of the system. Furthermore, formation of the 3D electron transportation pathways by a uniform dispersion of MWNT in the composite leads to outstanding electrochemical performance of the composite as an anode material for LIBs.

Methods

Nine cubic centimeter of tetraethyl orthosilicate (TEOS) ((C2H5O)4Si ≥ 99.5%) and 9 cm3 HCl (0.1 mol dm−3) were dispersed in ethanol (16 cm3) and stirred for 30 min. Meanwhile, 4 g citric acid (C6H8O7 · H2O ≥ 99.5%) and 2.2 cm3 ethylene glycol (C2H6O2 ≥ 99%) were dissolved in deionized water (10 cm3), and then 1.9 g MWNT dispersion (9 wt%, MWNT aqueous dispersion, Timesnano, Chengdu) (mass ratio of Si and MWNT = 6.6:1) was added into this solution with gentle stirring for 30 min. The two resulting solutions were thoroughly mixed and transferred into an evaporating dish and dried at 55 °C for 10 h. The resulting product was heated under Ar atmosphere for 1 h at 1100 °C to obtain SiO2@C/MWNT composite. A reference SiO2@C composite without MWNT was obtained following the same preparation route.

The crystal structure of the samples was characterized by X-ray diffraction (XRD D8 Discover, Bruker) employing Cu Kα radiation. Raman spectra were conducted with Ar-ion laser of 532 nm using the Via Reflex Raman imaging microscope system. The structure and morphology of the SiO2@C/MWNT composites were studied using scanning electron microscopy (SEM, Hitachi S-4800) and transmission electron microscopy (TEM, JEOL 2100), respectively. Surface elemental analysis was conducted by an energy-dispersive X-ray spectroscopy (EDX) attached to the TEM apparatus. The content of amorphous SiO2 in SiO2@C/MWNT composite was estimated by using a thermogravimetric analyzer (STD Q-600) under N2 flow (30 ml min−1).

The working electrodes were prepared by coating a homogeneous slurry containing 80 wt% active material, 10 wt% acetylene black (MTI, 99.5%), and 10 wt% polyvinylidene fluoride (PVDF) (Kynar, HSV900) binder dissolved in 1-methyl-2-pyrrolidinone (NMP, Sigma-Aldrich, 99.5%) onto a copper current collector by a doctor blade, and further drying at 65 °C for 12 h in a vacuum oven. The resulting SiO2@C/MWNT and SiO2@C composite electrode was punched into circular disks with a diameter of 10 mm and a mass loading of ~4 mg cm−2. The coin-type cells with high-purity lithium metal as the counter electrode were assembled in a glove box (MBraun) filled with argon (99.9995%). Galvanostatic charge and discharge tests were conducted on a multichannel battery tester (Neware, BTS-5 V5 mA) with the potential range of 0.01–2.5 V vs. Li/Li+ at various cycling rates. The Versa STAT electrochemical workstation was used to conduct cyclic voltammetry (CV) tests between 0.01 and 3 V vs. Li/Li+ at a scanning rate of 0.1 mV s−1 and electrochemical impedance spectroscopy (EIS) measurements in a frequency range from 100 kHz to 1 mHz.

Results and Discussion

The phase purity of the SiO2@C/MWNT ternary composites was confirmed by XRD. It can be seen from Fig. 1 that in contrast with SiO2@C, the SiO2@C/MWNT composite shows a typical peak of graphitic carbon at 26.1°, indicating the presence of MWNT with the structure planes (200) [12]. A weak peak around 43° corresponds to a diffusion scattering of amorphous carbon coating, while a broad diffraction peak around 21° is associated with amorphous SiO2 [13, 14]. All the above results demonstrate that as designed, SiO2@C/MWNT ternary composite was successfully obtained.

Fig. 1
figure 1

XRD patterns of SiO2@C and SiO2@C/MWNT composites

Raman spectroscopy was further performed to investigate the phase compositions in the SiO2@C/MWNT composite and the SiO2@C counterpart as shown in Fig. 2. Both samples possess double distinct peaks at 1340 and 1595 cm−1, related to the D and G bands of carbon, respectively [15]. These two vibration peaks demonstrate the low crystallinity of carbon [16]. The D band describes the defect-mediated zone-edge phonons and indicates the disordered carbon, edges, and defects, whereas the G band is a characteristic of the graphitic sheets, which according with the scattering of the E2g mode apperceived for sp2 domains [17,18,19]. It is worth to note that the ID/IG ratio for SiO2@C and SiO2@C/MWNT composites are 0.94 and 0.99, respectively. The ID/IG of SiO2@C/MWNT composites increased compared with that of SiO2@C as a result of a strong binding interaction and the increased structural defects between Si and O [20, 21].

Fig. 2
figure 2

Raman spectra of SiO2@C and SiO2@C/MWNT composites

As displayed in Fig. 3a, SEM confirms the micro/nano structure of SiO2@C/MWNT composite. The sample shows a disordered configuration with a wide size distribution. This could be considered as a verification of the amorphous structure of the material. From the TEM image (Fig. 3b), it can be seen that the MWNT-like bridges are directly connected to the SiO2@C particles, and this feature could support the structural integrity retention of the composite and favor the fast electron transfer. Meanwhile, MWNT of about 20–50 nm diameter intersperses among SiO2@C, which has an amorphous structure. The EDX element mapping (Fig. 3 (c1–c4)) indicates that the SiO2@C/MWNT composite contains homogeneously distributed O, Si, and C. One can see from Fig. 3d that an amorphous carbon layer with a thickness of about 2–7 nm is formed on the surface of SiO2. A turbostratic structure without crystalline lattice is discovered, indicating that the SiO2@C/MWNT composite has an amorphous structure. It is worth noting that MWNT is evenly distributed in the disordered matrix. A small amount of a microcrystalline structure domain could be observed in the composite, which lattice fringes with the spacing of about 0.205, 0.215, and 0.411 nm agree well with the spacing between (222), (311), and (111) of SiO2.

Fig. 3
figure 3

a SEM image and b TEM image of SiO2@C/MWNT composite. c1 EDX mapping of C (c2), O (c3), and Si (c4) elements. d HRTEM image of SiO2@C/MWNT composite

In order to verify the content of amorphous SiO2 in SiO2@C/MWNT composite, the TG and DTG data were collected and the results are shown in Fig. 4. The prominent weight loss between 550 and 730 °C, reflected in the TG curve, is related to oxidization of carbon and MWNT. Furthermore, the DTG curve shows two distinct peaks at 635 and 690 °C, which correspond to decomposition reaction of carbon layer and MWNT. Based on the positions of these two curves, the SiO2 content in the ternary composite can be estimated as ca. 77.5 wt%. Considering these data and the TG results, the mass composition of SiO2@C/MWNT could be estimated as SiO2:C :MWNT = 77.5:17: 5.5 wt%.

Fig. 4
figure 4

TG (black) and DTG data (red) of SiO2@C/MWNT composite

The CR2025 coin cells were assembled to test the electrochemical performance of the SiO2@C/MWNT nanocomposite. Figure 5 shows the CV data of SiO2@C/MWNT. The CV curves present a reduction peak at approximately 0.57 V vs. Li/Li+ at the first cycle. It is related to the reduction reactions of lithium with SiO2 resulting in the side products of Li4SiO4, Li2Si2O5, and Li2O. Among these, Li2Si2O5, as reported, is active in the subsequent cycles, which enhances the electrochemical performance of the system [22], and Li2Si2O5 is reversible while the Li2O and Li4SiO4 phases are irreversible upon cycling. The increase of current in the CV curves could be related with this phenomenon. Along with this, this phenomenon could be considered as a part of electrochemical activation of the electrode upon its cycling, which is commonly observed for porous composite systems. A cathodic peak at 0–0.5 V can be observed in the initial cycle, corresponding to the alloying process of SiO2 [23]. On the other hand, the anodic peak at 0.24–0.9 V is extensive in the Li extraction part, matching well with the de-alloying process between amorphous Li-Si alloys and amorphous SiO2 [24, 25].

Fig. 5
figure 5

Cyclic voltammograms of SiO2@C/MWNT composite electrode at a scan rate of 0.1 mV s−1

Figure 6a presents the charge/discharge curves of the SiO2@C/MWNT composite anode. The composite exhibits the initial discharge capacity of about 991 mAh g−1 while a corresponding charge capacity is about 615 mAh g−1, and this results in the initial coulombic efficiency of 62%. This relatively low coulombic efficiency could mainly be due to the formation of the solid electrolyte interface (SEI) on the electrode surface during the initial charge/discharge process. The discharge capacity becomes stable after 10 cycles, and the coulombic efficiency increases to ~100%. It is found that the charging potential profile is extraordinarily steep at potentials exceeding 1.4 V, which is due to a glassy state character of SiO2 with a strong polarization [26]. As shown in Fig. 6b, the potential profiles of the SiO2@C composite are similar to the profiles of the ternary composite but they exhibit lower capacities.

Fig. 6
figure 6

Charge/discharge profiles of a SiO2@C/MWNT and b SiO2@C composite electrode at a current density of 100 mA g−1

The counterpart SiO2@C composite was tested in the same electrochemical environment. As shown in Fig. 7, the comparative cycling performance studies of the binary and ternary electrodes were evaluated at a current density of 100 mA g−1. It is obvious that the SiO2@C/MWNT sample shows remarkably enhanced cyclability than its SiO2@C counterpart. Specifically, the SiO2@C/MWNT exhibits a high specific capacity of 744 mAh g−1 at 100 mA g−1 in the second cycle and maintains a capacity of 557 mAh g−1 after 40 cycles. However, the corresponding capacity of SiO2@C retains only a capacity of about 333 mAh g−1 at the 40th cycle. The superior cycling stability of the SiO2@C/MWNT electrode could be attributed to the introduction of well-dispersed MWNT in the composite. Incorporation of MWNT with SiO2@C is designed to provide pathways for electrolyte/Li+ ingress and to accommodate the anode active mass volume expansion during cycling [27]. An outstanding rate capability of the SiO2@C/MWNT ternary electrode is illustrated in Fig. 8. One can see that after 100 cycles, the specific discharge capacity of the cell with the SiO2@C/MWNT composite cathode slightly decreases, and it exhibits a capacity of 215 mAh g−1 at a high-current density of 1000 mA g−1, presenting its enhanced electrochemical stability. In the same time, the SiO2@C composite retains a capacity of only around 95 mAh g−1 when cycled at the same current density.

Fig. 7
figure 7

Cycle performance of SiO2@C and SiO2@C/MWNT composite electrodes at a current density of 100 mA g−1

Fig. 8
figure 8

Rate performance of SiO2@C and SiO2@C/MWNT composite electrodes at a current density of 1000 mA g−1

In order to further clarify the role of MWNT networks in the ternary composite, the EIS measurements were performed and the results are shown in Fig. 9. It can be seen that for the fresh cells, the diameter of the compressed semicircle in the high-to-medium frequency range for the SiO2@C/MWNT ternary electrode corresponds to 95 Ω, which is about half of that for SiO2@C, indicating that MWNT remarkably improves the conductivity and enhances the charge transfer properties of the ternary electrode. Figure 9b shows changes of EIS upon cycling and an equivalent circuit with a series of constant phase elements (CPE) and resistances obtained from the EIS data fitting. RE reflect the bulk resistance of the electrolyte. The CPE1 and RSEI are the charge capacitance and resistance of the solid electrolyte interphase (SEI) layer, respectively. The CPE2 and RCT are related to charge-transfer, which mirrors the lithium ions intercalation into the electrode. The inclined line is generated by the Warburg impedance (ZW), which represents the lithium-diffusion process within SiO2@C/MWNT. After the initial cycle, the diameter of the semicircle remains the same at about 95 Ω but the slope of the Warburg component decreases compared with that of a fresh cell, reflecting the lithium-ion diffusion process within the electrode. Further, the resistance of the ternary electrode decreases to about 30 Ω due to the activation process. After 50 cycles, the diameter of the semicircle tends to stabilize, i.e., there is no remarkable impedance change, which evidences the stability of the ternary electrode upon cycling and its ability to be well adapted to the volume changes. These results confirm that MWNT can obviously improve the conductivity and enhance the structure stability of the SiO2@C/MWNT ternary electrode.

Fig. 9
figure 9

a EIS spectra of SiO2@C/MWNT and SiO2@C electrodes before cycling. b EIS spectra of SiO2@C/MWNT electrode upon cycling and an equivalent circuit obtained for this system

Furthermore the SiO2@C/MWNT nanocomposite electrode exhibits a good-rate capability as shown in Fig. 10. The SiO2@C/MWNT electrode delivers reversible capacities of ~710, 570, 300, 250, and 220 mAh g−1 at current densities of 100, 200, 500, 750 and 1000 mA g−1, respectively. When further, the current density was returned to 100 mA g−1, about 95% of the initial capacity could be recovered, indicating a good structural and electrochemical stability of the system. It can also be seen from Fig. 10 that the reversible capacities of SiO2@C are lower than that of SiO2@C/MWNT over a whole range of the current densities studied. It can be concluded that the MWNT component enhances the conditions for lithium-ion diffusion and the electric conductivity of the composite, favoring its rate capability.

Fig. 10
figure 10

Rate performance of SiO2@C/MWNT and SiO2@C electrodes at various current densities

Table 1 compares the performance data reported for the silicon anode for lithium-ion batteries with the results of this work. It can be seen that the SiO2@C/MWNT electrode prepared in this work exhibits an enhanced electrochemical performance compared with those reported previously. One can see that the reversible capacity and capacity retention of SiO2@C/MWNT at 40th cycles are higher than for most of other silicon electrodes reported in the literature. These results indicate that the SiO2@C/MWNT composite with a carbon containing layer structure and MWNT could be considered as a promising anode for high-performance Li-ion batteries.

Table 1 Performance comparison of SiO2 and SiO2@C electrodes for LIBs

Conclusions

The SiO2@C/MWNT ternary composite was successfully synthesized by a simple sol-gel method using low-cost citric acid and TEOS as starting materials, followed by heat treatment. Due to its unique core-shell and network structure and enhanced contact between its individual components, the resulting ternary composite cathode exhibited a remarkably enhanced electrochemical performance compared with the binary SiO2@C counterpart. Considering the simplicity and efficiency of the preparation process and outstanding electrochemical performance, the SiO2@C/MWNT composite can be considered as a promising anode material for the next generation lithium-ion batteries.