1 Introduction

Lithium-ion battery (LIBs) is widely used in portable electronic products, new energy electrical vehicles, aerospace, and other fields. It is necessary to improve the properties of batteries, including energy density, power density, cycle stability, etc., to meet the growing demands of new applications (Dunn et al. 2011; Huang et al. 2023). Graphite, as a commercial anode of LIBs, has a theoretical capacity of 372 mAh g−1 that is difficult to improve due to the stoichiometry limit of LiC6 of graphite (Kubota et al. 2020). Therefore, a new anode material needs to be developed to improve the capacity and rate performance of lithium-ion storage.

Graphitized carbon nitride (g-C3N4) is a typical layered two-dimensional (2D) material. Because of its structural similarities to graphite, layered g-C3N4 can be regarded as graphene with a high nitrogen content and rich planar pores, capable of supplying rich adsorption and intercalation sites (Dong et al. 2017; Xu et al. 2017a), which is garnering the interest of battery researchers. Previous studies have shown that g-C3N4 has a high lithium storage capacity of up to 524 mAh g−1 (vs. 372 mAh g−1 of graphite) (Hankel et al. 2015; Jorge et al. 2014; Veith et al. 2013). The low conductivity of g-C3N4 as a semiconductor, however, restricts the electron transfer rate, resulting in a low real capacity (Shi et al. 2016). Furthermore, g-C3N4 has high adsorption energy for Li+, as shown by density functional theory (DFT) calculations. This means that Li+ faces a significant diffusion barrier of around 2.4 eV between each adsorption site in the g-C3N4 layer (Wu et al. 2013). In addition, significant structural distortion after Li intercalation results in a rapid capacity decline in cycling. Therefore, improving the electron transport rate of g-C3N4, reducing the adsorption energy barrier of Li+ in g-C3N4, and improving the structural stability of g-C3N4 are the keys to improving the electrochemical performance of g-C3N4.

Coupling the conductive amorphous carbon with g-C3N4 to form a composite can effectively alleviate the above issues, making g-C3N4 a feasible strategy for high-performance LIBs anode (Zhu et al. 2019). However, excellent dispersion of g-C3N4 in carbon materials is difficult to achieve. Previous studies have used ball milling or chemical synthesis methods to achieve the combination of conductive carbon skeleton and g-C3N4. However, it is easy to destroy the structure of g-C3N4 or consume too many chemicals in the composite process (Weng et al. 2019). Our previous studies have shown that the chemicals bond between supramolecules (g-C3N4 precursors) and biomass molecules could be used to achieve uniform dispersion of carbon precursors and C3N4. As a result, the amorphous carbon and g-C3N4 can form covalent bonds during carbonization process. This will facilitate the formation of a new g-C3N4-based carbon material (Subramaniyam et al. 2017). As a result, the key is figuring out how to distribute the two precursors evenly.

The high-value utilization of industrial lignin is low at present (Jian et al. 2022; Zhu et al. 2023), and the high carbon content and rich functional phenolic structure give lignin the prerequisite for the carbon electrode precursor and lignin could be easily regulated microstructure (Kizzire et al. 2021; Wen et al. 2022; Zhang et al. 2024). Lignin-derived carbon is expected to be a sustainable and low-cost electrode material (Shen et al. 2019; Zhang et al. 2022). In this work, the homogeneous mixing of lignin and melamine was achieved by a spray drying method, and then the covalently bonded C3N4/LC-x material was prepared by a one-step carbonization strategy. The uniform dispersion of g-C3N4 in amorphous carbon could improve the electrical conductivity and reduce the diffusion energy barrier of lithium ions. A high proportion of edge nitrogen and a low proportion of graphite nitrogen could regulate the nitrogen-containing groups in g-C3N4 and improve the structural stability of g-C3N4. Therefore, compared with the g-C3N4 anode, the C3N4/LC-x electrode has better electrochemical behavior, including higher reversible capacity, better rate performance, and cycling stability.

2 Experimental section

First, the melamine was dissolved in 1 L of deionized water, heated at 90 ℃ and stirred until completely dissolved. Then 6.5 g corncob lignin and 2 mL ammonia were added and continued to stir for 6 h. The mass ratio of lignin to melamine was 1:1, 1:2, 1:3, and 1:10, respectively. The precursor was obtained by spray drying the obtained solution. The precursor was recorded as MA/AL-x, and x was 1, 2, 3, and 10 corresponding to the mass ratio of lignin to melamine 1:1, 1:2, 1:3, and 1:10, respectively. Then the precursor was carbonized in a tube furnace at 550 ℃ for 2 h under N2 atmosphere to obtain C3N4/LC-x. x was 1, 2, 3, and 10 corresponding to the mass ratio of lignin to melamine 1:1, 1:2, 1:3, and 1:10, respectively (Fig. 1).

Fig. 1
figure 1

Preparation scheme for C3N4/LC-x

Full size image

3 Results and discussion

3.1 Characterization of C3N4/LC-2

According to the surface morphology of the precursor (Fig. S1), the texture of MA/AL-2 was the most unconsolidated, reflecting that the ratio of precursor lignin to melamine was the most evenly dispersed. According to the SEM of Fig. 2a, b, and c, C3N4 was unconsolidated in texture, but the composite C3N4/LC-2 shows a larger particle size than LC. LC, the lignin-carbon concave sphere was filled after the composite of C3N4. Preferable composite changes in the material surface morphology. During the carbonization process, amorphous carbon could destroy the structure of the bulk C3N4. The formed small C3N4 fragments could adhere to amorphous carbon, forming C3N4/LC materials with large interlayer spacing. The combination of amorphous carbon and C3N4 can improve the diffusion kinetics of lithium ions. Comparing LC with other composite proportions of C3N4/LC-1, 2, 3, and 10, the larger the proportions of C3N4, the smaller the particle size (Fig. S2), indicating that preferable composite may generate larger intermolecular forces (π-π bond, etc.), making the structure of C3N4/LC-2 denser. In Fig. 2d, e, f, TEM images, the 2D geometry of LC was irregular, C3N4 appears as a regular spherical shape, and the composite C3N4/LC-2 indicated the carbon layer connected with the sphere, providing ion transport channels. HRTEM (Fig. 2g, h, i) indicated the difference between the layered structure after composite and the unconsolidated structure before composite, indicating that the layers of the composite form through intermolecular forces (π-π bond, etc.) (Fu et al. 2017). The SEM elemental mapping images in Fig. 2j show that a homogeneous distribution of N, O, and C elements indicates a uniform dispersion of C3N4/LC-2, suggesting that lignin and melamine achieved uniform dispersion.

Fig. 2
figure 2

SEM images of the as-prepared (a) LC, (b) C3N4, and (c) C3N4/LC-2. TEM images of (d) LC, (e) C3N4/LC-2, and (f) C3N4; HRTEM images of (g) LC, (h) C3N4/LC-2, and (i) C3N4; (j) SEM elemental mapping images of C3N4/LC-2

Full size image

To confirm the successful composite of lignin carbon with C3N4 and its chemical bond properties, a variety of analytical techniques were used. In the XRD pattern of the precursor (Fig. S3a), the characteristic peaks of the composite precursor MA/AL-x were shown. In the FT-IR spectra of the precursor (Fig. S3b), the characteristic peaks of AL after composite MA/AL-x were significantly changed in comparison, the shrinkage of OH and NH at around 3500 to 3000 cm−1 and the shrinkage of C-N at 1050 cm−1 became stronger. The larger x is, the stronger the stretching vibration containing the nitrogen bond is. The benzene peak of MA/AL-2 was the strongest(Yang et al. 2007), which indicates that the combination of this ratio enhances the π-π bond and provides conditions for the strong intermolecular force of C3N4/LC-2. The result shows that the π-π bond can be strengthened by the composite of this ratio, which provides conditions for the strong intermolecular force of C3N4/LC-2. In the XRD pattern of Fig. 3a, the (100) peak of C3N4 at 12.73° was caused by the trigonal nitrogen linkage of tri-s-triazine units (Niu et al. 2012; Yang et al. 2013), (002) peak contrast LC was located at 26.96° at a larger angle, C3N4/LC-2 after composite, (002) peak shifted to a larger angle, and (101) peak did not change. Compared with C3N4, C3N4/LC-2 did not have the same high crystalline phase content, but retained the characteristics of amorphous carbon. It could indicate that C3N4 interacted with the LC, which was attributed to the fact that C-N–C bond formed between the lone pair of pyridine nitrogen in the C3N4 void and the carbon in the LC. The XRD pattern of C3N4/LC-1, 2, 3, and 10 (Fig. S4a) also shows that as the ratio of melamine in precursor became larger, the (002) peak shifted to a larger angle, which was closer to the characteristic peak of C3N4 (Fina et al. 2015; Guo et al. 2016). In the Raman spectrum of Fig. 3b, C3N4 has no Raman scattering peak, and the ID/IG value of C3N4/LC-2 was 1.15, much higher than the ID/IG value of LC 0.76, indicating more defects and disordered phases in C3N4/LC-2. Although the formation of C3N4 results in the decrease of the graphitization degree, it will enhance the surface wettability as well as provide extra defects to enhance Li+ storage (Li et al. 2016). The comprehensive results confirm its positive influence on the resultant materials. However, the Raman spectra of C3N4/LC-3, 10 (Fig. S4c) did not have obvious D and G peaks. With the proportion of melamine increased in the precursor, the carbonized sample could generate more crystalline C3N4. Strong fluorescence coverage of crystalline C3N4 could cover the D and G bands in the Raman spectrum. Excessive crystalline C3N4 was detrimental to conductivity and diffusion kinetics of lithium ions. The FT-IR spectrum (Fig. 3c) suggests that C3N4/LC-2, compared with LC, has decreased C-H shrinkage at 800 cm−1, shrinkage of benzene carbon at 1400 cm−1, stretching vibrations of C-N and C = C at 800 cm−1, and slightly stretching vibrations at about 3300 cm−1 (Zhang et al. 2018). It indicates the presence of C3N4 fragment in C3N4/LC-2. The delocalized electrons of C-N and the electron cloud of π bond on C = C were conducive to increasing the electron transfer rate and reducing the adsorption energy of lithium ions (Wang et al. 2018). In contrast, the FT-IR spectrum of C3N4/LC-3, 10 (Fig. S4b) showed the characteristic stretching vibration of crystalline C3N4, which was not conducive to improving the electron transfer rate. In (Fig. 3d) the XPS full spectrum, it can be seen that when compared C3N4/LC-2 with LC, the N content significantly increased, but C3N4/LC-3 had higher nitrogen content (Fig. S4d). The XPS C 1 s spectrum in Fig. 3e shows the chemical bonds change of C. Compared with LC, C-N and N–C = N appear in C3N4/LC-2, which was completely different from LC. The XPS C 1 s spectra of C3N4/LC-1, 3, 10 were compared with C3N4/LC-2 (Fig. S4e), and the C–C/C = C ratio of C3N4/LC-3, 10 was too low, which was not conducive to the intercalation of lithium. The N–C = N ratio of C3N4/LC-1 was too low, the increase of delocalized electrons may be less than that of C3N4/LC-2, and the increase of electron transfer rate may be insufficient (Li et al. 2015; Tahir et al. 2013). Figure 3f shows the XPS N 1 s spectrum. Due to the low nitrogen content of LC, the analysis of chemical bonds could not be fitted. C3N4/LC-2 had higher content of N-(C)3 and lower content of C-NHx. This may be because when the precursor was mixed, the dispersion of melamine in the lignin was better, so the amino group was relatively dispersed, resulting in low thermal stability of the amino group and decomposition in the pyrolysis process (Zhao et al. 2023). Comparing XPS N 1 s spectrum of C3N4/LC-1, 3, 10 with that of C3N4/LC-2 (Fig. S4f), the proportion of N-(C)3 in C3N4/LC-3, 10 was too low, and the proportion of N–C = N in C3N4/LC-1 was too low, which was not beneficial for the improvement of conductivity (He et al. 2015; Tahir et al. 2014; Yin et al. 2023).

Fig. 3
figure 3

a XRD spectra of C3N4, LC, and C3N4/LC-2; (b) Raman spectra of C3N4, LC, and C3N4/LC-2; (c) FT-IR spectra of C3N4, LC, and C3N4/LC-2; (d) XPS surface spectra of C3N4, LC, and C3N4/LC-2; high-resolution XPS spectra of the (e) C 1s and (f) N 1s regions for C3N4, LC, and C3N4/LC-2

Full size image

3.2 Electrochemical performances

Figure 4a shows the first three cycles of the C3N4 anode CV plot, with a severe irreversible reaction. The first three cycles of C3N4/LC-1, 3, 10 anode CV also had different degrees of irreversible reaction, and the first three cycles of LC anode CV were irregular in shape and not close to capacitive reaction (Fig. S5). As demonstrated in Fig. 4b, the first three cycles of the C3N4/LC-2 anode CV had a lower degree of irreversibility and a more trapezoidal shape, suggesting that the electrochemical reaction was closer to the capacitive reaction. This suggests that the uniform dispersion of C3N4 aids in enhancing the rate of electron transfer and lowering the adsorption energy of lithium ions. The first cycle of GCD curves in Fig. 4c shows a higher capacity and ICE of C3N4/LC-2, reaching 749.2 mAh g−1 and 59.9%, respectively. The reason for the low initial coulombic efficiency of C3N4/LC-2 was caused by excessive irreversible defects from C3N4. The rate and cycling performance in Fig. 4d, e also demonstrate the high-rate performance and cycling stability of C3N4/LC-2. Compared with other composite ratios of C3N4/LC-x, C3N4/LC-2 also had the best rate performance and cycling stability (Fig. S6). It was due to the electronic structure of C3N4 being properly regulated and large interlayer spacing could improve the stability of the carbon layer.

Fig. 4
figure 4

Electrochemical performances characterization. 1st, 2nd, and 3rd CV curves of (a) C3N4 (b) C3N4/LC-2 and recorded under scan rates at 0.2 mv s−1; (c) GCD curves of 1st cycle; (d) rate performances; (e) cycling performance at 1 A g−1

Full size image

Figure 5a shows the b value of the half-cell of C3N4/LC-2 anode. The b value of the anode was as high as 0.917, close to 1, indicating that the electrochemical reaction of the anode of C3N4/LC-2 was mainly due to capacitance contribution and adsorption (Xu et al. 2017b; Zhang et al. 2021). Figure 5b shows the capacitance contribution of the anode of C3N4/LC-2 at different scan rates, and the proportion of capacitance contribution was higher with the increase of scan rate, indicating that the electrochemical reaction of the anode of C3N4/LC-2 was mainly adsorption (Yin et al. 2021, 2020). In order to elucidate the electrochemical behavior of C3N4/LC-2 anode, the impedance of C3N4/LC-2 anode was measured, and the impedance data were analyzed after preparation and cycling. The Nyquist plot (Fig. 5c) consists of a semicircle at high frequency and a straight line at low frequency which were attributed to the electrolyte or solution resistance (Rs) offered at the electrode–electrolyte interface and the charge transfer resistance (Rct) in the case of the semicircle, and the Warburg lithium diffusion (Wdiff) in the case of the straight line (Zhong et al. 2022). By comparing the EIS curves of C3N4/LC-2, C3N4 and LC negative electrodes, the Rs and Wdiff (Rs = 1.6 Ω, Wdiff = 90.7 Ω) of C3N4/LC-2 negative electrodes were the lowest (Table S1), indicating that the preferable composite of C3N4/LC effectively promoted ion diffusion and reduced the adsorption energy of lithium ions. The diffusion impedance and interface impedance were reduced. The Rct of the anode of C3N4/LC-2 was also lower than that of C3N4, indicating that the uniform dispersion of C3N4 was conducive to electron transfer. The Warburg coefficient (σ, its unit is Ω/\(\sqrt{s}\)) can be obtained from the slope of the Warburg plot (Fig. 5d), and the diffusion coefficient of lithium-ions DLi+ at the interfacial regions can be estimated by the equation σ = \(\frac{RT}{{An}^{2}{F}^{2}C\sqrt{2{D}_{{Li}^{+}}}}\) (S4) (Yi et al. 2017). The relationship between the actual resistance (Z') of the three anodes and the square root of the frequency (ω−0.5) was obtained (Ding et al. 2018). The diffusion coefficient DLi+ of C3N4/LC-2 anode was the highest, 2.7*10−13, which was much higher than 1.2*10−13 of LC anode and 1.2*10−14 of C3N4 negative electrode, indicating that the preferable composite of C3N4/LC-2 and the uniform dispersion of C3N4 effectively promoted lithium-ion diffusion.

Fig. 5
figure 5

(a) Fitting slope b value of C3N4/LC-2 CV; (b) relative contributions of the capacitive and diffusion-controlled charge storage processes at different scan rates of C3N4/LC-2; (c) Nyquist plots for LC, C3N4/LC-2, and C3N4; (d) corresponding relationship between Z’ and ω.−0.5 (where ω = 2πf) of LC, C3N4/LC-2, and C3N4

Full size image

4 Conclusion

In this study, the C3N4/LC-2 composite was crafted and synthesized through a straightforward and convenient method. This synthesized material enables uniform dispersion of C3N4 in an amorphous carbon matrix, facilitates the generation of numerous nitrogen active sites, promotes electron transfer, and reduces the energy barrier for lithium-ion adsorption. C3N4/LC-2 exhibits outstanding properties, such as high reversible capacity, favorable rate performance, and extended cycle life. Moreover, it introduces a novel approach to the structural design of modified carbon-based anodes for lithium-ion batteries.