Introduction
To meet the increasing needs for social development, more and more new energy materials emerge at the historic moment [
1–
3], such as solar cell materials, hydrogen storage materials, solid oxide battery materials, etc. With the rapid development of portable electronic devices and hybrid electric vehicle industry, the performances of lithium-ion batteries are increasingly improved, such as energy density, power density, cycle stability and other requirements [
4,
5]. Therefore, it is an irresistible trend to further improve the performances of lithium-ion batteries [
6–
8].
Silicon(Si)-based anode materials have received more and more attention in the development of lithium-ion batteries [
9], because of their high theoretical capacity of nearly 4200 mAh/g [
10,
11]. Another advantage over other materials is that silicon (Si)-based anode materials have a lower discharge potential (<0.5 V versus Li/Li
+ at room temperature) [
12]. Due to the remarkable volume change in the lithiation/delithiation processes (>300%), the unstable cycling behavior of silicon has hindered its practical application [
13,
14].
To ameliorate the volume expansion issue of silicon-based anode materials, silicon/carbon composites [
15], especially porous carbon [
16], are considered to be an effective material for improving the storage performances of lithium [
17]. Porous carbon materials are often doped with nitrogen to enhance their adsorption capacity [
18]. Most nitrogen-doped carbon materials are dominated by carbon with nitrogen serving as interstitial doping atoms in the carbon structure [
19–
21]. The addition of nitrogen atoms does not change the arrangement of carbon atoms [
22], but the amount of nitrogen in such nitrogen-doped carbon materials is limited [
23]. Graphite-like carbon nitride (g-C
3N
4) is a network of interlocking carbon and nitrogen atoms, and is the most nitrogen-rich carbon and nitrogen material, which has superior electrical, optical, structural and physicochemical properties [
24–
26]. Hence it is a worth-studying direction for high-performance lithium anode to composite silicon with g-C
3N
4. To prepare silicon nanoparticles, the silane gas chemical vapor deposition (CVD) method is generally used [
27]. However, by using two or more substances for gas phase reaction at a high temperature, this method is expensive and it is difficult to achieve large-scale preparation. Therefore, the practical application of silicon-based materials as anode of lithium batteries is still a huge challenge.
On the contrary, the magnesium thermal reduction method as an effective method to prepare silicon nanostructures is more economical [
28]. However, as the reduction process is an exothermic reaction, this brings new challenges, such as the agglomeration of silicon particles and the side reactions in the magnesium thermal reduction process [
29]. To overcome these problems by introducing g-C
3N
4 into silicon material, in this paper, a stober method was used to prepare g-C
3N
4/SiO
2 precursor first, and the SiO
2 in the composites was successfully reduced to Si by magnesium thermal reduction for forming porous sheet-like structure of g-C
3N
4/Si nanocomposites.The excellent performances of g-C
3N
4/Si nanocomposites result from the merits that g-C
3N
4 acts as a good buffer layer in nanocomposites, which effectively inhibits the volume expansion of silicon nanoparticles, accelerates the transfer of lithium-ions and electrons, and provides abundant lithium storage sites. Such materials provide a new direction for the development of silicon-based anode lithium-ion batteries.
Experimental
Chemicals and reagents
The melamine (C3N3(NH2)3, Aladdin, 99%) and magnesium power (Mg) were purchased from Shanghai Lingfeng Chemical Reagent Company. The tetraethylorthosilicate (TEOS, 28.4%) and hydrofluoric acid (HF, 40%) were bought from China National Pharmaceut. Ethanol (CH3CH2OH, 99.7%) and the ammonium-solution (NH3·H2O, 28%) was produced by Shanghai Suyi Chemical Reagent Company. All chemical reagents in the experiment were of analytical grade and used without further purification.
Preparation of g-C3N4
An appropriate amount of melamine was weighed and calcined to obtain g-C3N4 in a muffle furnace at 550°C for 3 h, with a heating rate of 2°C/min. The yellow products were collected when the reaction temperature was cooled to 25°C. 1.0 g of g-C3N4 was dispersed in 30 mL hydrochloric acid solution (10 mol/L) and vigorously stirred for 1 h for obtaining protonized g-C3N4. The product was washed several times with deionized water and ethanol and subsequently dried in a vacuum oven at 80°C for 12 h.
Preparation of g-C3N4/Si nanocomposites
100 mg of g-C3N4 was added to a mixture solvent of 15 mL of ethanol and 5 mL of deionized water, and ultrasonically dispersed for 30 min. Thereafter, a small amount of ammonia was added to the above suspension with continuously stirring for 15 min, followed by adding 1 mL TEOS drop by drop. According to the stober method, TEOS went through hydrolyzed condensation polymerization under alkaline conditions to form SiO2 on the surface of g-C3N4 layer. To prepare g-C3N4/Si nanocomposites, the as obtained g-C3N4/SiO2 and magnesium powder were mixed with a mass ratio of 1:1. The fully ground sample was placed on a porcelain boat, and calcined in a tube furnace at 750°C for 6 h at argon atmosphere. After natural cooling to room temperature, the product was treated with 1.0 mol/L HCl solution and the remaining silica was treated with 10 wt% of HF solution. Subsequently, the yellow brown product was collected and washed with deionized water and ethanol by centrifugation. Finally, the product was transferred to a vacuum drying oven at 80°C for 12 h.
Characterizations
The compositions and phases of the samples were confirmed by using an X-ray diffractometer (XRD, D/MAX Ultima III, Rigaku, Japan) with Cu Kα radiation. Next, the morphological characteristics and crystal structure of the sample were analyzed by using a field emission scanning electron microscope (FESEM, Nova Nano SEM 230) and a transmission electron microscope (TEM, JEOL JEM-2100F). The element distribution was analyzed by using an energy dispersive X-ray spectroscopy (EDS). The surface states were determined by an X-ray photoelectron spectroscopy (XPS) on Thermo Fisher Scientific Escalab-250Xi using Al-Kα X-ray source. N2 adsorption/desorption isotherms were investigated using a micromeritics ASAP2020 instrument at 77 K, which can obtain the specific surface area and pore size distributions of the Brunauer-Emmett-Teller (BET).
Electrochemical measurement
The lithium-ion batteries were assembled in a glove box filled with argon, and the content of H2O and O2 was below 0.1 ppm. The commercial lithium sheet was used as the counter electrode (The mass loading of the active material is around 0.5 mg/cm2). The celgard 2400 micro-porous polypropylene film was used as the separator. 1 mol/L of LiPF6 was dissolved in a mixture solvent (1:1, v/v) of ethylene carbonate and dimethyl carbonate to form electrolyte. The working electrodes were fabricated with 70 wt% of g-C3N4/Si nanocomposites, 20 wt% of acetylene black, and 10 wt% of carboxymethyl cellulose in H2O. The slurry was coated on a piece of Cu foil, and dried in a vacuum oven at 80°C over night. The cyclic voltammetry (CV) testing was performed on a CHI 660 D electrochemical workstation with a voltage range of 0.01 to 3.0 V (versus Li/Li+). Galvanostatic charge-discharge curves were tested between 0.01 and 3.0 V (versus Li/Li+) at 25°C on LANDCT2001A. The electrochemical impedance spectral (EIS) measurements were conducted on an Autolab electrochemical workstation, and the frequency range was set from 100 kHz to 10−2 Hz with an AC amplitude of 5 mV.
Results and discussion
The morphologies of g-C
3N
4 and g-C
3N
4/Si nanocomposites were characterized by FESEM. As observed in Fig. 1(a), the g-C
3N
4 powder consists of particles with irregular shapes as well as layered structures. As observed in Fig. 1(b), the surface of particles is smooth. The FESEM images of g-C
3N
4/Si nanocomposites are shown in Figs. 1(c) and 1(d). The size of particles becomes smaller compared with that of g-C
3N
4 and the surface also becomes rough, which indicates that Si nanoparticles are uniformly dispersed on the surface of g-C
3N
4 layers. To further explore the structural characteristics, the g-C
3N
4/Si nanocomposites were observed under (HR)TEM. Figure 2(a) shows a TEM image of g-C
3N
4/Si. It can be seen that the g-C
3N
4/Si nanocomposites exhibit a porous sheet-like structure with typical g-C
3N
4 features. The uniformly dispersed and irregularly shaped black nanoparticles observed on the carbon nitride substrate are Si nanoparticles, indicating that the Si nanoparticles are uniformly supported on the porous g-C
3N
4 nanosheets and are tightly bonded together. The orbicular SAED pattern (inset in Fig. 2(a)) also reveals the g-C
3N
4/Si nanocomposites obtained with a polycrystalline structure. From the HRTEM image in Fig. 2(b), the lattice fringes with a lattice spacing of 0.31 nm can be clearly observed, which correspond to the (111) plane of single crystal Si [
30]. In addition, the (001) plane of g-C
3N
4 can also be observed, indicating that heterojunction interface of g-C
3N
4/Si is formed [
31]. Figure 2(c) shows the elemental mapping of g-C
3N
4/Si. It can be seen that the Si, C, and N elements are highly distributed, indicating that Si nanoparticles are uniformly dispersed and supported on the g-C
3N
4 substrate.
The XRD patterns of g-C
3N
4 and g-C
3N
4/Si are depicted in Fig. 3(a). The XRD spectrum of pure g-C
3N
4 exhibits two distinct peaks at 13.3° and 26.2°, which are indexed to be (100) and (002) planes of graphitic materials, respectively [
32]. The low intensity peak at 26.2° is a characteristic stacking reflection of conjugated aromatic systems, indicating a graphitic structure with an interlayer distance of 3.4 nm. Due to the amorphous structure of the carbon matrix, the intensity of C
3N
4 diffraction peak without Si is low. It is clearly observed that there are five strong diffraction peaks at 28.5°, 47.2°, 56.1°, 69.1°, and 76.5° in the XRD pattern of g-C
3N
4/Si nanocomposites, corresponding to (111), (220), (311), (400), and (331) planes of single crystal Si, which is identical to the standard comparison card (JCPDS No. 27-1402) [
33]. However, an impure peak at 35.98° is observed, which corresponds to (300) of another single crystal Si, which is identical to the standard comparison card (JCPDS No. 40-0932). This result indicates that SiO
2 was successfully reduced to Si [
22]. Compared with that of pure g-C
3N
4, the characteristic peaks at 13.3° and 26.2° basically disappear, which are attributed to the low crystallinity of g-C
3N
4 and the high crystallinity of silicon nanoparticles. The surface composition of g-C
3N
4/Si nanocomposites was further analyzed by XPS. The survey spectrum is displayed in Fig. 3(b). Five distinct characteristic peaks corresponding to C 1s, O 1s, N 1s, Si 2s, and Si 2p are observed, indicating the existence of Si, C, N, and O elements [
34]. Figure 3(c) shows the XPS spectrum of N 1s, with three characteristic peaks at 398.8, 400.1, and 400.8 eV, corresponding to pyridine N, pyrrole N, and graphite N, respectively. From the fitting result, the peak intensity of pyridine N and graphite N is found to be small, while the pyrrole N has a high strength peak. The XPS fitting results show that the number of graphite N in g-C
3N
4/Si is less than that of pyridine/pyrrole species. It is worth noting that different N has different electron configuration characteristics. The pyridine N and pyrrole N have a lone pair of electrons in the sp
2 orbital and p orbital, respectively. The graphite N has one electron in each of the two p orbitals. Therefore, the former can actively absorb Li
+, which is conducive to enhancing the surface capacitance, suggesting that the as prepared g-C
3N
4/Si may have a good eletrochemical property [
35]. In the high resolution XPS of Si 2p (Fig. 3(d)), two distinct characteristic peaks are mainly seen at 99.8 and 103.8 eV, respectively [
36]. The characteristic peak at 99.8 eV corresponds to the energy peak of Si-Si, proving the existence of zero valence Si atoms in the composite, which indicates that after the magnesium thermal reduction reaction, silica is successfully transformed into Si. The characteristic peak at 103.8 eV corresponds to the energy peak of O= Si= O, indicating that there exist some Si atoms in the valence state of Si
4+ [
37]. The existence of Si
4+ is the result of slight oxidation of Si NPs in unavoidable exposure to air. Combined with the XRD pattern, the above result confirms that there is a small amount of silica on the surface of Si particles [
38].
To explore the specific surface area and porosity of g-C
3N
4/Si nanocomposites, nitrogen adsorption-desorption isotherms were recorded and shown in Fig. 4. According to ISO-IUPAC classification, the BET curve of g-C
3N
4/Si nanocomposites belongs to the typical IV type (Fig. 4(a)). The BET specific surface area was estimated to be 11.782 m
2/g with a pore volume of 0.076091 cm
3/g. As can be seen from Fig. 4(b), the pore diameter is mainly distributed around 4.81 nm, and there is a certain pore structure before 4 nm, indicating the existence of micropore and mesoporous structure [
39–
41]. The big specific surface area ensures sufficient active sites and the good porosity enables efficient transmission of lithium-ions in g-C
3N
4/Si nanocomposites [
42].
Figure 5(a) demonstrates the typical CV curves of g-C
3N
4/Si nanocomposites at a scan rate of 0.1 m/s between 0.01 and 3.0 V. The cathode peak at 0.13 V is mainly attributed to the alloying process of Li ions and Si particles to form lithium silicon alloy (Li
xSi). In the charging process, the two anode peaks at 0.37 V and 0.53 V correspond to the characteristic peaks of amorphous Si [
29], and the non-crystalline Li
xSi changes to the amorphous Si in the de-lithium process on the surface. With the accumulation of cycles, the intensity of the characteristic peak gradually increases, which indicates that there is an electrochemical activation process in charging and discharging, making Si electrode more active. Figure 5(b) shows the voltage specific capacity distribution of g-C
3N
4/Si electrode at the current density of 0.2 A/g within the voltage range of 0.01 to 3 V. However, the charging capacities of the 10th, 100th, and 200th cycles are higher than that of the first cycle, which could be attributed to the gradual penetration of electrolyte for completing activation process of porous g-C
3N
4/Si nanocomposites in long time charging and discharging. In the initial discharge period, the electrical conductivity continuously decreases, and a significant plateau profile at 0.1 V can be observed, which can be attributed to the SEI layer and Li layer alloying with Si [
43]. The g-C
3N
4/Si electrode has an initial discharge capacity of 1033.3 mAh/g and a charging capacity of 528.9 mAh/g. The initial coulombic efficiency is relatively low (51.19%) due to the irreversible entry of lithinum-ions into the porous g-C
3N
4/Si nanocomposites. In the continuous charge and discharge cycles, there is no obvious voltage change, which indicates that the electrochemical stability is good. As can be seen from Fig. 5(c), the g-C
3N
4/Si nanocomposites and pure Si particles were used as the anode to measure the long cyclic stability at the current density of 0.2 A/g. The initial capacity of pure silicon electrode battery is about 2800 mAh/g and the coulomb efficiency is 89%. However, the initial capacity declines sharply to 113 mAh/g after 20 cycles due to the large volume expansion. The initial capacity of g-C
3N
4/Si nanocomposites is as high as 1026 mAh/g, and the Coulomb efficiency is only 58.8% due to the loss of some irreversible capacities and the formation of SEI film. However, the capacity of g-C
3N
4/Si nanocomposites decreases quickly to 403 mAh/g after 25 cycles due to the volume expansion of Si nanoparticles. It is noteworthy that the capacity increases continuously to 621.3 mAh/g from 25 to 300 cycles, which could be attributed to the gradual penetration of electrolyte for completing activation process of porous g-C
3N
4/Si nanocomposites in charging and discharging. After 400 cycles, the capacity can still be maintained at around 587 mAh/g, and the efficiency is close to 100%, indicating that the electrode material has good stability. Figure 5(d) indicates that the g-C
3N
4/Si electrode material has a good rate performance. As can be seen from the results, the specific capacity is 548 mAh/g when the current density is 0.1 A/g. When the current density is raised to 2 A/g, the specific capacity of the battery is 200 mAh/g.
The impedance spectra of pure Si and g-C
3N
4/Si nanocomposites are displayed in Fig. 6. Both plots display a quasi-semicircle and a slant line in the high and the low frequency region, respectively. The diameter of the semicircle corresponds to the SEI resistance (
Rsei) and the charge-transfer resistance(
Rct). The oblique line is associated with the Warburg impedance (
Zw), which is related to the diffusion coefficient of Li
+ at the electrolyte-electrode interface. By comparing the semi-circular diameters of the two anode materials in the high-frequency region, it is obvious observed that the semi-circular diameters of g-C
3N
4/Si nanocomposites are smaller, that is, g-C
3N
4/Si nanocomposites have a smaller interfacial charge transfer and lithium-ion diffusion resistance, effectively improving the electrochemical performances [
44–
46]. Based on the lithium storage performances and impedance spectra, the improved lithium storage performances are believed to have benefited from the g-C
3N
4 substrate, which plays an important role in alleviating Si volume expansion in the charging and discharging process, accelerating the transmission of lithium ions and charges, providing abundant lithium storage sites, and thereby improving electrochemical performances.
Conclusions
In summary, g-C3N4 was used as the substrate to prepare the precursor of g-C3N4/SiO2 nanocomposites by using the stober method. Besides, the SiO2 in the nanocomposites was reduced to Si by magnesium thermal reduction reaction for preparingg-C3N4/Si nanocomposites. As lithium-ion battery anode material, g-C3N4/Si nanocomposites exhibited good electrochemical performances. The g-C3N4/Si nanocomposites displayed a high initial discharge capacity of 1033.3 mAh/g at 200 mA/g, and an excellent reversible capacity of 584.4 mAh/g after 200 cycles. Even at 2 A/g, it still retained a capacity of about 200 mAh/g. The improved lithium storage performances are ascribed to the positive effect of g-C3N4 substrate in alleviating Si volume expansion in the charging and discharging process, accelerating the transmission of lithium-ions and charges, and providing abundant lithium storage sites.