X _BP index as a predictive tool for fast-charging performance of artificial graphite anodes in lithium-ion batteries

Abstract

Artificial graphite has garnered significant interest as a high-rate anode material for lithium-ion batteries (LIBs). While previous studies suggest that expanding the graphite interlayer spacing can enhance rate capability, the direct correlation between precursor molecular structure and fast-charging performance remains insufficiently understood. Herein, the ratio of aromatic bridging carbon to peripheral aromatic carbon in the precursor—defined as X_BP—is proposed as a key structural descriptor. Its impact on interlayer spacing and high-rate lithium storage performance is systematically investigated. By selectively isolating macerals to obtain precursors with varying degrees of condensation, the reliability of X_BP is validated, and its role in constructing an ordered, micro-expanded layered structure with enhanced Li⁺ diffusion kinetics is clarified. Specifically, as X_BP increases from 0.212 to 0.486, the 5C discharge capacity improves significantly from 92 to 201 mAh g⁻¹, revealing a strong parabolic relationship. In situ XRD and GITT measurements demonstrate that BCG135 exhibits more complete phase transitions (LiC₁₈ → LiC₁₂ → LiC₆) and faster Li⁺ diffusion compared to commercial natural graphite (Gr), highlighting its superior fast-charging capability. This study not only deepens the understanding of lithium storage mechanisms in artificial graphite but also offers an effective strategy for designing next-generation carbon anodes for high-power LIBs.

---

1. Introduction

Graphitic carbon and its derived composite materials are the dominant anode materials for lithium-ion batteries (LIBs) due to their high theoretical capacity (372 mAh g⁻¹), excellent structural stability, and sustainable precursors.^1–3 However, the highly anisotropic graphite structure restricts Li⁺ intercalation to edge-oriented interlayers (non-basal planes) with a typical spacing of ≈3.35 Å.^4,5 At high current densities, lithium-ion (Li⁺) transport is hindered by the layered structure of graphite, leading to slow kinetics of Li⁺ intercalation.⁶ In addition, the sluggish Li⁺ diffusion kinetics of natural graphite under high current conditions leads to Li⁺ accumulation at the carbon layer edges, promoting lithium dendrite formation and posing serious safety risks.^7,8

Despite extensive efforts to understand lithium intercalation into graphite, achieving the rapid formation of LiC₆ under high current densities remains a major challenge.⁹ Efficient Li⁺ insertion into graphite layers is critical for boosting fast-charging capacity and enabling practical high-power full-cell applications. To address the limited mass transport in graphite anodes, several strategies such as etching, interlayer expansion, and heteroatom doping have been employed to enhance Li⁺ diffusion efficiency.^10–14 Etching involves the destruction of the graphite lamellar structure through alkali metal modification.¹⁵ In contrast, expanding the interlayer spacing offers a more favourable pathway for Li⁺ migration without severely damaging the graphite framework.¹⁶ Additionally, doping with electronegative elements such as sulfur¹⁷ or nitrogen¹⁸ modifies surface charge distribution and enhances Li⁺ adsorption, further improving ionic transport.¹⁹ While these approaches can partially improve fast-charging capability, they often introduce undesirable side effects, such as increased solid electrolyte interphase (SEI) formation, structural instability, and limited long-term cycling performance.

The formation mechanism of artificial graphite and its impact on electrical properties have become a hot topic and pose significant challenges in the field of LIBs.²⁰ The graphite layer structure comes from the high-temperature ordered transformation of carbon precursors into stacks.^21–23 By adjusting the crystalline/amorphous composition of the precursor, obtaining a wider interlayer spacing can significantly enhance the diffusion path of lithium ions.²⁴ Qiu et al. investigated the microstructural transformations from coal to graphite during high-temperature graphitization of anthracite.²⁵ Zhang et al. studied the microstructural evolution of coals from various regions that underwent natural transformation into graphite under the thermal influence of intrusive bodies.²⁶ Zeng et al. conducted a comprehensive study on the microstructural changes in carbon precursors over a wide temperature range of 1000–2800 °C, elucidating the relationship between microcrystalline structures and carbonization–graphitization temperatures.²⁷ Li et al. systematically explored the catalytic graphitization process and assessed the electrochemical behaviour of the resulting graphite when applied as an anode in lithium-ion batteries.²⁸ Additionally, the correlation between the graphite microstructure and its electrochemical properties was explored.²⁹ Despite the prevalence of techniques such as proximate analysis, elemental analysis, and Fourier transform infrared (FTIR) spectroscopy, which are widely used to characterize coal properties, they primarily offer bulk compositional information and overlook the connectivity between aromatic units, making them inadequate for explaining the transformation of coal precursors into disordered graphitic microcrystals during high-temperature treatment. Furthermore, there is a lack of clear and effective indicators to predict the interlayer spacing of graphite and its associated fast-charging performance.

In this study, a strong correlation is revealed between the fast-charging performance of graphite anodes in lithium-ion batteries and the X_BP index, defined as the ratio of aromatic bridging carbon to peripheral aromatic carbon in the carbon precursor. A strategy is proposed involving the selective separation of coal macerals to obtain polycyclic aromatic hydrocarbon precursors with varying degrees of condensation. These precursors are then employed to investigate the formation mechanism of micro-expanded layered structures and their impact on fast-charging/discharging lithium storage behaviour. The X_BP index plays a critical role in suppressing undesirable cross-linking reactions during carbonization, thereby promoting the development of an ordered micro-expanded graphite structure with optimized Li⁺ diffusion pathways. Specifically, as the X_BP increases from 0.212 (BCG125) to 0.486 (BCG135), the 5C discharge capacity increases from 92 to 201 mAh g⁻¹, exhibiting a distinct parabolic relationship between the precursor's X_BP value and fast-charging capacity across five types of bituminous coals. Furthermore, in situ XRD and GITT analyses confirm that BCG135 undergoes more complete phase transitions (LiC₁₈ → LiC₁₂ → LiC₆) and exhibits faster Li⁺ diffusion kinetics, highlighting its superior fast-charging capability compared to Gr. This work not only elucidates the lithium storage mechanisms associated with micro-expanded artificial graphite structures but also offers an effective design guideline for the development of high-performance artificial graphite anodes for LIBs.

2. Experimental

2.1. Materials preparation

The carbon precursor for artificial graphite was bituminous coal from the Halanggou Coal Mine (Shanxi, China). Commercial natural graphite comes from Jinmei Carbon Materials Science and Technology Co., Ltd in Tianjin, China. The proximate and elemental analysis were conducted according to GB/T 212-2008, and the results are presented in Table S1. Zinc chloride solutions with varying densities (1.25, 1.30, and 1.35 g cm⁻³) were prepared and calibrated using a densitometer. Subsequently, bituminous coal powder was thoroughly mixed with the 1.25 g cm⁻³ heavy liquid for initial separation. After allowing the mixture to rest for 3 minutes, the floating material was skimmed from the surface, and the sediment at the bottom was further separated using the 1.30 g cm⁻³ solution, yielding BC125, a sample with a density range of 1.25–1.30 g cm⁻³. The sample was then sequentially separated using 1.35 g cm⁻³ and 1.40 g cm⁻³ heavy liquids, with coal having a density greater than 1.40 g cm⁻³ designated as BC140. The process of maceral sorting is illustrated in Fig. S1. The yields of the different density coals are shown in Fig. S2. The separated coal was ground and treated with hydrofluoric and hydrochloric acids to remove mineral content, resulting in ultra-low ash coal. Nearly all mineral impurities, including quartz, pyrite, calcite, and clay minerals, were effectively removed from the coal.

2.2. Graphitization process

The samples were then placed in an IGBT high-temperature graphitization furnace, with graphite crucibles utilized for this purpose. The samples were heated to 2800 °C in an argon atmosphere under a slight positive pressure (20–30 kPa) for a duration of 3 hours at a heating rate of 10 °C min⁻¹. The samples were then allowed to cool naturally to room temperature and subsequently labeled. For instance, the graphitized sample obtained at 2800 °C from BC125 was labeled as BCG125.

2.3. Characterization of synthetic graphite materials

Elemental composition was determined using a CHNS elemental analyzer (Elementar UNICUBE, Germany). Fourier transform infrared (FT-IR) spectra were collected on a VERTEX80V spectrometer over the 4000–500 cm⁻¹ range. Morphology and microstructural features were examined via scanning electron microscopy (SEM, MAIA3 LMH, Tescan, Czech Republic) and transmission electron microscopy (TEM, Tecnai G2 F20, FEI, USA). Carbon skeleton structures were evaluated by solid-state ¹³C cross-polarization/magic angle spinning nuclear magnetic resonance (CP/MAS NMR, Bruker AVANCE III HD 600 MHz). Thermogravimetric (TG) and differential scanning calorimetry (DSC) analyses were conducted using a synchronous thermal analyzer (STA 449 F3, Netzsch, Germany) to evaluate thermal stability and cross-linking behavior of coal. X-ray diffraction (XRD, Bruker D8 Advance, Cu Kα, λ = 1.5406 Å) was employed to assess crystallinity and graphitization degree. Raman spectra were recorded on a VERTEX80V spectrometer (λ = 532 nm). Surface chemical composition was analyzed by X-ray photoelectron spectroscopy (XPS, ESCALAB 250Xi), with binding energy calibrated to the C 1s peak at 284.8 eV. Specific surface area and pore structure were evaluated by nitrogen adsorption/desorption using the Brunauer–Emmett–Teller (BET) method.

2.4. Electrode fabrication and electrochemical characterization

The electrochemical properties of the products were tested using CR2025 coin cells. The artificial graphite, carbon black, and polyvinylidene fluoride (PVDF) mixture was dissolved in N-methyl-2-pyrrolidone (NMP) solution in the ratio of 8 : 1 : 1 by mass, and the resulting electrode had an active material loading of approximately 1.0–1.5 mg cm⁻². The slurry was continuously coated on Cu foil using an automatic coating machine, which was subsequently subjected to 6 h of vacuum drying at 80 °C. The assembly of CR2025 coin cells was made inside an argon-filled glovebox, where the working electrode was the as-obtained material, the reference and counter electrodes were Li metal, and the separator was a polypropylene microporous film. The electrolyte was 1.0 mol L⁻¹ LiPF₆ dissolved in ethylene carbonate (EC)/diethyl carbonate (DEC)/dimethyl carbonate (DMC) (1 : 1 : 1 vol ratio), and 5% fluoroethylene carbonate (FEC) was used as electrolyte. Electrochemical performance was evaluated using a LAND CT3002A (Wuhan, China) in the voltage range of 0.01–3.0 V at current rates from 0.1C to 5C (1C = 350 mA g⁻¹). Cyclic voltammetry (CV, 0–2.0 V) and electrochemical impedance spectroscopy (EIS, 0.1 Hz–10 kHz) were conducted on a CHI660E electrochemical workstation. The galvanostatic intermittent titration technique (GITT) was carried out on CR2025 coin cells using the same battery testing system. For half-cell configurations, the voltage window was maintained at 0.01–3.0 V vs. Li⁺/Li. GITT measurements were performed at 0.1C with a 30 min current pulse followed by a 120 min rest period. The coin-type Li⁺ full cell was prepared with BCG135 as the anode and LiNi_0.8Co_0.1Mn_0.1O₂ (NCM811) as the cathode. And the weight ratio of NCM811, super P and polyvinylidene fluoride (PVDF) was 95 : 3 : 2. Besides, the charging and discharging processes of the full cell were tested under a voltage window of 2.5–4.3 V. All electrochemical tests were performed at room temperature.

2.5. In situ XRD analysis

The in situ XRD measurements of electrodes were performed with a Bruker D2 X-ray diffractometer with Cu K, radiation. To simultaneously satisfy the requirements of collecting adequate diffraction peak data and high-rate measurement, a current density of 0.2C is selected to conduct an in situ XRD test. The discharge process was caried out with a 2-theta scan range of 20–30° and a scan speed of around 2° min⁻¹.

3. Results and discussion

Bituminous coal possesses a tunable molecular structure and can be utilized as a carbon precursor.^30,31Fig. 1a presents the preparation routes for artificial graphite depicting the evolution of both molecular and microcrystalline structures. HRTEM micrographs (Fig. 1b1–b3) reveal that after graphitization at 2800 °C, the randomly arranged basic structural units (BSUs) appeared in BCG125(Fig. 1b1), which consist of a number of long and ordered graphite microcrystals and a few disordered graphite crystals. Subsequent observation of Fig. 1b3 reveals that graphite lattice fringes are more ordered and denser, which indicates that the inert components during the separation of coal micro components is beneficial for the basic structural units of the graphitization process. The interlayer spacing of BCG125, BCG, and BCG135 (0.3384 nm, 0.3377 nm, and 0.3373 nm, respectively) is slightly larger than that of Gr (≈0.3354 nm). This subtle expansion, along with the structural evolution observed in Fig. 1c, collectively contributes to enhanced electrochemical performance. Specifically, BCG125 exhibits a relatively disordered carbon layer arrangement, which may provide more lithium storage sites but also introduces defects that hinder electronic transport. In contrast, BCG135 presents a highly ordered graphitic structure with improved electronic conductivity. The combination of expanded interlayer spacing and enhanced structural ordering in BCG135 synergistically facilitates Li⁺ diffusion and electron transfer, thereby resulting in the best rate performance among all samples.

Fig. 1 (a) The preparation process of artificial graphite. (b) HRTEM images of b1 BCG125, b2 BCG and b3 BCG135. (c) Structural evolution of artificial graphite under different processing conditions.

Coal macerals, identified by their reflectance and morphology, mainly include vitrinite and inertinite. Inertinite, derived from fusinized lignocellulosic tissues, appears bright white under reflected light due to its high reflectivity.^32,33 By utilizing the density variation among coal components, bituminous coal was separated into different carbon precursors through gravity sorting. BC125 represents a low-density carbon precursor enriched in vitrinite, while BC140 is a high-density carbon precursor containing more inertinite. The intermediate-density carbon precursor contains a mixture of both maceral types (Fig. 2a–c and Table S2). As shown in Fig. 2d, carbon and oxygen atoms are the primary elements in bituminous coal, with the C/H atomic ratio increasing as coal density increases. This trend suggests a gradual enhancement in aromaticity. The FTIR spectra of bituminous coal with different densities are shown in Fig. 2e and S3. The broad peak at 3430 cm⁻¹ is attributed to –OH stretching vibration. The peaks in the range of 3000–2800 cm⁻¹ belong to aliphatic methylene(–CH₂–) symmetric/asymmetric stretching vibration. The characteristic peak at 1700 cm⁻¹ is associated with the carbonyl (–COOH) stretching vibration. The peak near 1605 cm⁻¹ corresponds to the aromatic ring (C=C) stretching vibration. Moreover, the peak observed at 900–700 cm⁻¹ is related to the bending vibration of the aromatic ring (Ph-H) in the out-of-plane configuration.^34–36 Additionally, FTIR fitting patterns are provided, with the calculated results detailed in Table S3. The intensity of the Branched index decreases with increasing coal density.³⁷ To further assess surface chemistry, XPS analysis (Fig. S4) was conducted. The oxygen content of BC125 is 14.9%, much higher than the 10.1% of BC135, confirming that the low-density carbon precursor contain more oxygen functional groups, which can influence reactivity and graphitization behavior. The microcrystalline structure of these carbon precursor was investigated using XRD analysis (Fig. 2f and S5). The XRD pattern displays a broad diffraction peak at 23°, corresponding to the (002) plane of carbon. This peak can be deconvoluted into two sub-components, attributed to disordered aliphatic carbon and well-organized aromatic domains, respectively²⁶,with the fitting results summarized in Table S4.

Fig. 2 Structural characteristics of precursors: (a–c) optical micrographs of coal macerals (black: epoxy resin; red: vitrinite; green: inertinite). (d) Elemental composition and H/C atomic ratio. Fitted curves of (e) FTIR spectra, (f) XRD patterns and (g) ¹³C-CP/MAS-NMR spectra of BC125, BC and BC135. (h) Comparison of structural parameters including X_BP, branched index, aromaticity, inertinite content, volatile content and H/C atomic ratio of BC125, BC and BC135. (i) Distribution of aromatic structural units as a function of coal macerals.

As displayed in Fig. 2g and S6, solid-state 13C-CP/MAS-NMR was conducted to calculate X_BP. The solid-state ¹³C-CP/MAS NMR spectra showed six regions representing aliphatic carbon (f_al^C:14–90 ppm), protonated and aromatic carbon (f_a^H:90–129), aromatic bridging carbon (f_a^B:129–137 ppm), alkyl group-substituted aromatic carbon (f_a^S:137–150 ppm), oxygen-substituted aromatic carbon (f_a^P:150–165 ppm) and carboxyl carbon (f_a^C:165–225 ppm).³⁸ Based on the integrated peak areas, the X_BP index was calculated using the following formula:

The calculated X_BP values of BC125, BC, and BCG135 are 0.212, 0.383, and 0.486, respectively, indicating a significantly higher content of aromatic bridge carbons in BC135 compared to BC125. The precursors exhibit a progressive increase in both the aromaticity and the branched index in the order: BC125 < BC < BC135. Conversely, the volatile content and H/C atomic ratio decrease in the same order. Notably, X_BP serves as a descriptor of the degree of polyaromatic condensation, representing the density of bridge bonds between adjacent aromatic units.³⁹ As shown in Fig. 2i and Table S6, BC135 contains a significantly higher proportion of large-sized aromatic clusters than BC125 and BC, further confirming its higher aromatic condensation and structural compactness. This highly condensed aromatic structure is expected to favor the formation of a more graphitizable and fast-charging carbon framework.

As shown in Fig. 3a, the carbon yields of BC125, BC, and BC135 at 1500 °C are 57.55%, 61.25%, and 63.02%, respectively. This observation may be attributed to the cross-linking reaction between the C and O atoms, which leads to the release of volatile components such as H₂O and CO₂.^40–42 In contrast to the heavy component, the light component exhibits a more rapid weight loss between 300 °C and 600 °C due to its swift decomposition and the subsequent formation of volatiles. Conversely, the heavy component undergoes a more gradual weight loss, attributed to its higher aromatic content. These results confirm that gravity sorting effectively isolates carbon precursors with distinct molecular structures, enabling control over alkyl side chains and enrichment of polycyclic aromatic hydrocarbons. The XRD patterns of graphitized samples, including BCG125, BCG, and BCG135, obtained after graphitization at 2800 °C, are shown in Fig. 3b and summarized in Table S7. Among these samples, BCG135 exhibits a significantly larger average crystallite size (L_a) compared to BCG125, which is attributed to its higher inertinite content in the high-density fraction, facilitating greater structural ordering during high-temperature graphitization. The Raman spectra of graphitized samples from different coal densities are presented in Fig. 3c. Three characteristic peaks are identified in the Raman spectra. The D band at 1345 cm⁻¹ is attributed to disordered carbon domains and structural defects typically associated with amorphous carbon. The G band at 1560 cm⁻¹ corresponds to the in-plane vibrational mode of sp²-hybridized carbon atoms, while the sharp 2D band at 2698 cm⁻¹ serves as a fingerprint of highly ordered graphitic stacking, indicating well-developed graphitic lattices.²¹ This observation indicates that the graphitization temperature is a crucial factor influencing the degree of graphitization in artificial graphite. As shown in Fig. 3d, the inertinite content increases from 11.61% to 65.56%, accompanied by a linear rise in the X_BP value from 0.212 to 0.486. This increase in X_BP correlates positively with the graphitization degree (from 68.48% to 74.34%) and negatively with the I_D/I_G ratio (from 0.46 to 0.13), indicating a more ordered graphite structure. These results suggest that a higher X_BP value, associated with greater inertinite content, effectively suppresses cross-linking reactions at lower temperatures and facilitates the formation of highly ordered, micro-expanded layered structures during high-temperature graphitization. To further demonstrate the microstructure changes of artificial graphite from a bituminous coal carbon precursor with different densities, SEM analysis was carried out, as shown in Fig. 3e1–e3. BCG125 exhibits a loosely packed and disordered lamellar structure with visible pores and wrinkles. In contrast, BCG135, derived from the high-density fraction, displays a compact and well-aligned lamellar structure, indicative of a higher degree of graphitization.⁴³

Fig. 3 Microcrystalline and structure characteristics of BCG125, BCG, and BCG135. (a) TG and DTG curves in Ar. (b) XRD patterns. (c) Raman patterns. (d) Linearity of inertinite content, X_BP, graphitization and I_D/I_G. (e) Scanning electron microscopy images. (f) Schematic illustrations of the structural evolution.

The pore size distribution of synthetic graphite materials, as evaluated by the BJH model (Fig. S7 and Table S8), reveals a broad distribution in the range of 2.00–10.00 nm, with a prominent peak around 3 nm. Table S9 summarizes the particle size distribution parameters of the samples. Significant differences in particle size distributions can be observed among the samples. In particular, BCG135 shows the smallest particle sizes with D₅₀ = 8.05 μm, which are markedly lower than those of the other samples and Gr. Based on these findings, a graphitization mechanism for various coal components is proposed. As illustrated in Fig. 3f, the molecular configuration of coal-derived precursors with different X_BP values significantly influences the graphitization pathway and the final microstructure. For BC135, which originates from a high-X_BP inertinite-rich precursor, a higher proportion of aromatic bridging carbon facilitates the ordered stacking of basic structural units (BSUs) and accelerates the growth of lateral (L_a) and longitudinal (L_c) graphitic domains, ultimately yielding a compact and highly graphitized lamellar structure (Fig. 3f1). In contrast, the low-X_BP precursor BC125 contains more peripheral functional groups, which are susceptible to cross-linking reactions at lower temperatures. This leads to spatial dislocation within BSUs during heat treatment, impeding the formation of well-ordered graphite domains and resulting in a distorted, folded lamellar structure (Fig. 3f2). The intermediate case, BCG, derived directly from unseparated bituminous coal, presents a hybrid structure composed of both vitrinite and inertinite. As a result, its final microstructure exhibits partially ordered graphite domains alongside disordered stacking regions, reflecting a moderate degree of graphitization (Fig. 3f3).

The above physical and chemical properties demonstrate that different structures of artificial graphite can be obtained by modifying the molecular structure of coal. Furthermore, the prepared artificial graphite will be used as an electrode for LIBs in order to evaluate its electrochemical performance. As illustrated in Fig. 4a, the initial charge capacities of BCG125, BCG and BCG135 at 0.1C are 281.8, 296.7 and 355.9 mAh g⁻¹, respectively (Table S9). The discharge profiles exhibit a short plateau at 1.0–1.2 V, corresponding to the formation of a solid electrolyte interphase (SEI) layer on the graphite surface during the first cycle. The extended plateau at lower potentials is associated with the intercalation and deintercalation of Li⁺ into the graphite layers.⁴⁰ The charge capacity below 0.5 V is a critical parameter for anode materials in practical applications, as a lower delithiation voltage can contribute to a higher output potential in full-cell devices. Consequently, the low voltage lithium storage characteristics of BCG125, BCG and BCG135 were the subject of further investigation.

Fig. 4 (a) Initial GCD curves at 0.1C of graphitized samples. (b) Quantitative capacities of the samples in different voltage zones at 0.1C during discharging. (c) Rate performance. (d) CV curves of the samples at 0.1 mV s⁻¹, respectively, (e) CV curves of BCG135 for the initial 3 cycles. (f) Nyquist plots, (g) long-term cycling, and (h) typical GCD curves of the NCM811 cathode, BCG135 anode, and NCM811//BCG135 full cell at 0.1C. (i) Rate capability of the full cell. (j) Cycling performance of the full cells. (k) Photograph of a laboratory thermometer powered by a full cell.

As shown in Fig. 4b, BCG135 exhibits the highest capacity contributions in the 0.07–0.11 V and 0.11–0.21 V voltage regions, significantly surpassing BCG and BCG125. These voltage ranges correspond to the formation of LiC₁₂, indicating that BCG135 promotes a more complete lithium intercalation process, leading to superior high-rate performance. This enhancement can be attributed to an optimized conductive network and highly ordered interlayer alignment, which significantly enhance electronic transport, thereby facilitating the formation of LiC₁₂ and LiC₁₈. The rate capabilities of the samples are illustrated in Fig. 4c and S8. BCG135 demonstrates the highest specific capacity at nearly all prevailing current densities. The specific capacity of BCG135 reaches up to 367.5 mAh g⁻¹ at 0.1C, and 318.1, 285.7, and 201.6 mAh g⁻¹ at 1C, 2C, and 5C, respectively. Furthermore, when the current density is recovered to 0.1C after 60 cycles, a high reversible capacity of 374.9 mAh g⁻¹ with 102.0% capacity retention is achieved, indicating excellent rate performance. The superior electrochemical performance of BCG135 can be attributed to its highly developed graphitic structure. As demonstrated in Fig. 4d, the graphitized samples exhibit similar voltametric features, including a weak peak at 0.9–1.2 V associated with the decomposition of the electrolyte additive FEC, and a broad peak at 0.6–0.8 V arising from the formation of the solid electrolyte interphase (SEI) layer.

During the anodic scan, sharp oxidation peaks at 0.30, 0.33, and 0.40 V are observed, indicating Li⁺ extraction from LiC_x. For BCG135 (Fig. 4e), the disappearance of peaks near 0.8 and 1.0 V in subsequent cycles suggests that a stable SEI layer was formed during the initial cycle. The EIS tests were performed to elucidate the variation in the electrochemical kinetics of the material electrodes (Fig. 4f and S11), and the obtained parameters are summarized in Table S11. All Nyquist plots display a depressed semicircle in the high-to-intermediate frequency range, followed by a sloped line at low frequencies. The semicircle corresponds to the charge-transfer resistance (R_ct), while the low-frequency tail is ascribed to Warburg impedance (W₀), associated with Li⁺ diffusion through bulk electrode materials.⁴⁴Fig. 4f shows that BCG135 has a lower R_ct, confirming enhanced electrochemical kinetics. This improvement is attributed to the ordered micro-expanded layered structure, which facilitates ion diffusion by reducing transport barriers within the graphitic domains. More surprisingly, the BCG125, BCG and BCG135 anodes show the most stable cycling performance (Fig. 4g) and undergo an electrochemical activation process at the beginning of charge and discharge, which is distinguished from the gradient current cycling steps involved in the rate testing (Fig. 4c), subsequently reaching maximum reversible capacities of 322.4, 306.3, and 400.1 mAh g⁻¹. This increasing trend of capacity usually happens under high-rate conditions, which agrees well with reported work.^12,45 To further evaluate the practical application potential of BCG135 with performance, a lithium-ion Battery (LIB) full cell was assembled using commercial NCM811 as the cathode and BCG135 as the anode (Fig. 4h). As illustrated in Fig. 4h, with an operational voltage window of 2.5–4.3 V, the full cell achieved a reversible capacity of 192 mAh g⁻¹ at 0.1C. The GCD curves of the full cells (Fig. 4i) demonstrate capacities of 192 190, 169, 154, 137and 117 mAh g⁻¹ at current densities of 0.1C, 0.2C, 0.5C, 1C, 2C and 5C. As shown in Fig. 4j, the cells retained a capacity of 100 mAh g⁻¹ after 100 cycles at a current density of 1C, respectively. The full battery can easily light up a thermometer to complete laboratory temperature and humidity detection (Fig. 4k). These results highlight the promising potential of BCG with superior fast-charging performance for practical applications in LIBs.

Based on the superior lithium storage capability and enhanced electronic transport properties demonstrated in BCG135, further investigations were conducted to compare the rate performance, phase transformation behaviour, and lithium-ion diffusion kinetics between artificial graphite (BCG135) and Gr, aiming to elucidate the fundamental differences in their electrochemical mechanisms and performance advantages. Fig. 5a possesses the charge capacities of 387.5, 316.9, 277.5, 214.8, 136.4 and 39.1 mAh g⁻¹ at 0.1C, 0.2C, 0.5C, 1C, 2C and 5C. It is worth noting that the charge plateau can be well maintained when the current density is increased from 0.1C to 5C (Fig. 5b). The capacity retention comparison is further illustrated in Fig. 5c, where BCG135 outperforms Gr across all current densities. As the current density increases to 5C, BCG135 still maintains a relatively high retention, whereas Gr exhibits a drastic capacity decay. The lithium storage behaviour of different anodes is evaluated in the Li//Gr half cells. Fig. 5d exhibits several beneficial potential plateaus below 0.50 V (vs. Li⁺/Li) that are attributed to the transition of various “staged” LiC_x before the formation of LiC₆ at 0.2C. The BCG135 anode exhibits the distinguishable phase transformation of LiC_x, indicating the enhanced Li⁺ storage ability. The differential capacity (dQ/dV) curves in Fig. 5e reveal that BCG135 has lower polarization and a more stable phase transformation process compared to Gr, further confirming its enhanced lithium-ion storage dynamics and superior fast-charging capability. The lithium-ion diffusion dynamics of BCG135 and Gr anodes are further elucidated by the galvanostatic intermittent titration technique (GITT).⁴⁶ As illustrated in Fig. 5f, the diffusion coefficients of Gr are relatively low, indicating sluggish Li-ion transport kinetics. Moreover, the significant fluctuation in log D values at different lithiation stages suggests that lithium-ion diffusion in graphite is highly dependent on the Li intercalation process and structural constraints. In contrast, as shown in Fig. 5g, BCG135 demonstrates substantially enhanced lithium-ion diffusion kinetics, with diffusion coefficients consistently higher by one to two orders of magnitude compared to those of Gr. This significant enhancement can be attributed to the abundance of aromatic bridging carbon in high-X_BP carbon precursors, which promotes the formation of continuous conjugated domains during graphitization. These graphitic domains improve electronic conductivity by reducing structural defects and enhancing π-electron delocalization. Meanwhile, the increased bridging between aromatic units facilitates the development of more ordered graphitic stacking with slightly expanded interlayer spacing, effectively providing additional Li⁺ transport pathways and overcoming the kinetic limitations commonly observed in conventional graphite anodes. The relatively stable log D values across the lithiation/delithiation process suggest a more uniform Li-ion intercalation/extraction mechanism, reducing the polarization effect commonly observed in conventional graphite. The enhanced diffusion behaviour of BCG135 can be attributed to the micro-expanded interlayer spacing, which effectively increases lithium-ion diffusion and mitigates transport limitations.¹¹ The above results indicate that BCG135 exhibits high capacity, excellent rate performance and long-cycle stability, which can be attributed to its artificial graphite structure featuring a sequential stacking structure and appropriate layer spacing, significantly superior to other reported graphitic materials, including highly crystalline graphite, spherical graphite, coke-derived graphite, and anthracite-derived graphite (Fig. 5h).^46–49

Fig. 5 Comparison of electrochemical performance between Gr and BCG135. (a and b) GCD curves of Gr and BCG135 at different current densities. (c) Comparison of rate performance. (d) Voltage curves below 0.5 V vs. Li⁺/Li and (e) corresponding differential capacity profiles within 0.01–0.5 V vs. Li⁺/Li at 0.2C. (f and g) GITT curves and calculated Li⁺ diffusion coefficient Gr and BCG135. (h) Comparison of initial coulombic efficiency (ICE) and rate capacity of BCG135 and other reported graphite materials. (i and j) In situ XRD mappings in the initial discharge−charge process of Gr and BCG135.

The lithium storage mechanism of the Gr anode involves the intercalation and deintercalation of Li⁺ ions, forming a series of lithium–graphite intercalation compounds, which results in a periodic expansion and contraction of the graphitic layers.⁵⁰ To elucidate the effect of artificial graphite on lithiation behavior and structural evolution, in situ XRD measurements were performed, focusing on the (002) reflection during the first charge–discharge cycle at 0.2C. As shown in Fig. 5i, the Gr anode exhibits a rapid voltage drop to 0.5 V, followed by a distinct voltage plateau between 0.2–0.01 V, corresponding to the formation of LiC_x phases (LiC₁₂ and LiC₆) via a typical two-phase intercalation process. In contrast, Fig. 5j shows that the (002) reflection of the BCG135 anode remains nearly unchanged above 0.5 V during discharge, suggesting that lithium storage in this region mainly occurs via surface adsorption rather than intercalation. Upon further discharge to lower voltages, a significant downshift in the (002) peak is observed for BCG135, from 26.2° to 23.9°, which is slightly broader than that of Gr (26.2° to 24.2°), indicating a more substantial layer spacing expansion and a higher degree of lithiation. Furthermore, BCG135 undergoes a complete series of phase transitions—LiC₁₈ → LiC₁₂ → LiC₆—whereas Gr only retains partial LiC_12–18 phases, reflecting limited lithiation. The interlayer spacing expansion (Δd₀₀₂) was calculated to quantify the lithiation-induced volume changes. For BCG135, the d₀₀₂ spacing increases from 3.41 Å to 3.73 Å, while the Gr anode exhibits a smaller change from 3.41 Å to 3.57 Å. These results indicate that BCG135 possesses a larger interlayer spacing, allowing deeper Li⁺ intercalation and the formation of LiC₆, thereby contributing to higher capacity at high charge/discharge rates.

The structure–performance relationships between inertinite content, X_BP index, graphitization degree, I_D/I_G and 5C charging capacity were systematically analyzed, revealing that the X_BP index is closely associated with the fast-charging capability of graphite anodes in LIBs. As shown in Fig. 6a, increasing the inertinite content from 11.61% to 65.56% led to a corresponding rise in X_BP values from 0.212 (BC125) to 0.486 (BC135), indicating a strong positive correlation between inertinite content and X_BP values. Building on this, Fig. 6b and c shows that a higher X_BP value corresponds to a higher degree of graphitization, reflected by a lower I_D/I_G ratio (decreasing from 0.43 for BCG125 to 0.13 for BCG135), while the graphitization degree increases from 68.48% to 77.69%, indicating the establishment of micro-expanded layered structures. Fig. S15 and S16 demonstrate that graphite samples with higher X_BP values exhibit smaller particle sizes and reduced specific surface areas, features that enhance electronic conductivity, shorten Li⁺ diffusion pathways, and improve electrochemical kinetics under high-rate conditions. The ratio of aromatic bridging carbon to peripheral aromatic carbon in the precursor and their corresponding 5C charging capacities were also calculated to validate this relationship. Fig. 6d shows that an increase in graphitization degree from 68.48% (BCG125) to 77.69% (BCG135) results in a substantial rise in 5C charging capacity, from 92.1 to 201.4 mAh g⁻¹. Consistent with the expected benefits of improved structural order. In line with this, Fig. 6e reveals a strong parabolic relationship between X_BP and 5C charging capacity, further verifying that X_BP serves as a key structural descriptor closely associated with both graphitization degree and fast-charging performance in BCG. The suppressed performance at low X_BP values is attributed to insufficient aromatic bridging in the precursor, which hinders the formation of extended conjugated domains and ordered graphitic layers, thereby limiting Li⁺ diffusion kinetics. In addition, the precursor density follows the order BC120 < BC125 < BC130 < BC135 < BC140, while the corresponding graphitization degrees of the derived graphite follow a different trend: BCG125 < BCG120 < BCG130 < BCG140 < BCG135. This discrepancy highlights a nonlinear relationship between precursor molecular density and graphitization behavior. Moreover, no clear correlation is observed between other precursor indicators such as f_ar^C, and f_a^P, emphasizing that the X_BP value is more reliably correlated with both graphite microstructure formation and 5C charging capacity (Fig. 6f and g).

Fig. 6 Quantitative analysis for inertinite content, X_BP, graphitization degree, I_D/I_G and 5C capacity. (a) Linearity between inertinite content and X_BP of the precursor. (b) Linearity between X_BP of the precursor and graphitization degree of graphite. (c) Linearity between X_BP of the precursor and I_D/I_G of graphite. (d) Quadratic relationship between graphitization degree and 5C capacity of graphite for LIBs. (e) Quadratic relationship between X_BP of the precursor and 5C capacity of graphite for LIBs. (f) Quadratic relationship between f_ar^C of the precursor and 5C capacity of graphite for LIBs. (g) Quadratic relationship between f_a^P of the precursor and 5C capacity of graphite for LIBs. (h) Relationship diagram among inertinite content, X_BP, graphite structure and fast-charging capacity.

4. Conclusions

In summary, the X_BP index serves as an effective structural indicator for understanding the evolution of micro-expanded layered structures and the enhanced rate performance of artificial graphite anodes for LIBs. Through selective coal maceral separation, a series of precursors with varying degrees of aromatic condensation were obtained to verify the accuracy of the X_BP value and to explore the mechanisms of micro-expanded layered structure formation and fast-charging performance. Additionally, in situ XRD and GITT analyses confirm that high-X_BP samples such as BCG135 exhibit more complete phase transitions (LiC₁₈ → LiC₁₂ → LiC₆) and significantly enhanced Li⁺ diffusion kinetics, resulting in an impressive 5C reversible capacity of 201.6 mAh g⁻¹ that markedly outperforms conventional natural graphite. This work not only elucidates the structural origin of the superior fast-charging behaviour in artificial graphite but also provides a practical and quantitative descriptor for guiding the molecular-level design of high-rate carbon anode materials for LIBs.

Author contributions

Zongxu Yao: investigation; methodology, writing – original draft; Tianqi Xu: data curation; Rongmiao Zhang: formal analysis; Yutong Xie: methodology; Zhiqiang Tang: investigation; Wei Jiang: methodology; Yinshuang Guan: formal analysis; Qun Wei: investigation; Chenming Liu: resources; Yaxin Chen: investigation, writing – review & editing; Liang Dong: conceptualization, funding acquisition, writing – review & editing; Jianguo Yang: conceptualization.

Conflicts of interest

The authors declare no conflict of interest.

Data availability

The data that support the findings of this study are available from the corresponding author Prof. Dong upon reasonable request.

Supplementary information is available. See DOI: https://doi.org/10.1039/d5ta05433h.

Acknowledgements

The work was supported by the Jiangsu Key Laboratory for Clean Utilization of Carbon Resources Research Project (BM2024007), the Fundamental Research Funds for the Central Universities, the Graduate Innovation Program of China University of Mining and Technology and the Postgraduate Research & Practice Innovation Program of Jiangsu Province (KYCX23_2817 and KYCX24_2892). The authors gratefully acknowledge the Advanced Analytical and Computing Center of China University of Mining and Technology for providing technical support in sample characterization.

Notes and references

1. C. Y. Wang, T. Liu, X. G. Yang, S. H. Ge, N. V. Stanley, E. S. Rountree, Y. Leng and B. D. McCarthy, Nature, 2022, 611, 485–490.
2. Y. Tian, G. Zeng, A. Rutt, T. Shi, H. Kim, J. Wang, J. Koettgen, Y. Sun, B. Ouyang, T. Chen, Z. Lun, Z. Rong, K. Persson and G. Ceder, Chem. Rev., 2021, 121, 1623–1669.
3. J. Shi, D. Xiao, M. Ge, X. Yu, Y. Chu, X. Huang, X. Zhang, Y. Yin, X. Yang, Y. Guo, L. Gu and L. Wan, Adv. Mater., 2018, 30, 1705575.
4. Z. J. Liu, Q. Q. Guo, J. L. Shi, G. T. Zhai and L. Liu, Carbon, 2008, 46, 414–421.
5. Y. Katayama, M. Yukumoto and T. Miura, Electrochem. Solid-State Lett., 2003, 6, A96–A97.
6. T. Gao, Y. Han, D. Fraggedakis, S. Das, T. Zhou, C.-N. Yeh, S. Xu, W. C. Chueh, J. Li and M. Z. Bazant, Joule, 2021, 5, 393–414.
7. Q. Q. Liu, C. Y. Du, B. Shen, P. J. Zuo, X. Q. Cheng, Y. L. Ma, G. P. Yin and Y. Z. Gao, RSC Adv., 2016, 6, 88683–88700.
8. C. Y. Wang, C. P. Yang and Z. J. Zheng, Adv. Sci., 2022, 9, 2105213.
9. H. L. Andersen, L. Djuandhi, U. Mittal and N. Sharma, Adv. Energy Mater., 2021, 11, 2102693.
10. Q. Cheng, R. Yuge, K. Nakahara, N. Tamura and S. Miyamoto, J. Power Sources, 2015, 284, 258–263.
11. Y. H. T. Tran, D. Kang, K. An, S. H. Song, M. K. Choi, C. Kim, H. Kim and S. W. Song, Mater. Today, 2024, 81, 12–22.
12. M. Niu, L. W. Dong, J. P. Yue, Y. Q. Li, Y. Y. Dong, S. C. Cheng, S. Lv, Y. H. Zhu, Z. T. Lei, J. Y. Liang, S. Xin, C. H. Yang and Y. G. Guo, Angew. Chem., Int. Ed., 2024, 63, e202318663.
13. W. L. Cai, C. Yan, Y. X. Yao, L. Xu, R. Xu, L. L. Jiang, J. Q. Huang and Q. Zhang, Small Struct., 2020, 1, 2000010.
14. Y. L. Fu, Y. Q. Jin, J. Ma, J. H. Liu, Z. Wang, B. Wang and X. Z. Gong, Chem. Eng. J., 2022, 442, 136184.
15. Y. Wu, L. Y. Wang, Y. F. Li, Z. Y. Zhao, L. W. Yin, H. Li and Y. J. Bai, J. Phys. Chem. C, 2017, 121, 13052–13058.
16. Y. L. An, H. F. Fei, G. F. Zeng, L. J. Ci, B. J. Xi, S. L. Xiong and J. K. Feng, J. Power Sources, 2018, 378, 66–72.
17. B. Sun, Q. Zhang, H. Xiang, F. Han, W. Tang, G. M. Yuan, Y. Cong, C. L. Fan, A. Westwood and X. K. Li, Energy Storage Mater., 2020, 24, 450–457.
18. Y. P. Liu, Y. Z. Zhen, T. R. Li, F. Bettels, T. He, M. H. Peng, Y. C. Liang, F. Ding and L. Zhang, Small, 2020, 16, e2004770.
19. H. Geng, Y. Peng, L. Qu, H. Zhang and M. Wu, Adv. Energy Mater., 2020, 10, 1903030.
20. Y. Y. Liu, H. D. Shi and Z. S. Wu, Energy Environ. Sci., 2023, 16, 4834–4871.
21. L. Wang, Z. Li, C. Du, Y. Han and J. Yang, Int. J. Coal Prep. Util., 2024, 44, 19–36.
22. A. Ghosh, C. D. A. Razzino, A. Dasgupta, K. Fujisawa, L. H. S. Vieira, S. Subramanian, R. S. Costa, A. O. Lobo, O. P. Ferreira, J. Robinson, M. Terrones, H. Terrones and B. C. Viana, Carbon, 2019, 145, 175–186.
23. T. Qiu, J. Yang and X. Bai, Int. J. Miner. Metall. Mater., 2020, 27, 162–172.
24. B. Xing, C. Zhang, L. Chen, G. Huang, H. Guo, X. Liu, Y. Ruifu, C. Zhang and S. Yuan, J. China Univ. Min. Technol., 2019, 48, 1133–1142.
25. T. Qiu, J.-G. Yang and X.-J. Bai, Energy Sources, Part A, 2021, 43, 769–782.
26. S. Zhang, Q. Liu, H. Zhang, R. Ma, K. Li, Y. Wu and B. J. Teppen, Carbon, 2020, 157, 714–723.
27. H. Zeng, B. Xing, Y. Cao, B. Xu, L. Hou, H. Guo, S. Cheng, G. Huang, C. Zhang and Q. Sun, Int. J. Min. Sci. Technol., 2022, 32, 1397–1406.
28. Y. Li, X.-d. Tian, Y. Song, T. Yang, S.-j. Wu and Z.-j. Liu, New Carbon Mater., 2022, 37, 1163–1169.
29. A. Heckmann, O. Fromm, U. Rodehorst, P. Munster, M. Winter and T. Placke, Carbon, 2018, 131, 201–212.
30. H. Lu, S. Sun, L. Xiao, J. Qian, X. Ai, H. Yang, A.-H. Lu and Y. Cao, ACS Appl. Energy Mater., 2019, 2, 729–735.
31. Q. Tian, X. Li, L. Xie, F. Su, Z. Yi, L. Dong and C. Chen, New Carbon Mater., 2023, 38, 939–953.
32. I. C. C. O. Petrology, Fuel, 1998, 77, 349–358.
33. I. C. C. O. Petrology, Fuel, 2001, 80, 459–471.
34. Y. Lu, C. Zhao, X. Qi, Y. Qi, H. Li, X. Huang, L. Chen and Y.-S. Hu, Adv. Energy Mater., 2018, 8, 1800108.
35. T. Guan, G. Zhang, J. Zhao, J. Wang and K. Li, Fuel, 2019, 242, 184–194.
36. X. Tao, C. Jie, W. Yongping, H. Chuansheng, X. Panshi, G. Xicai and S. Assem, J. China Coal Soc., 2024, 1352, 1–10.
37. A. Ping, W. Xia, Y. Peng and G. Xie, Int. J. Min. Sci. Technol., 2021, 31, 233–240.
38. B. Tian, Y. Y. Qiao, Y. Y. Tian, K. C. Xie and D. W. Li, Appl. Therm. Eng., 2016, 109, 560–568.
39. Z. Huaiqing, Z. Anning, L. Zhen, Q. Jinzhou, J. Weihong and Y. Zhiyuan, J. China Univ. Min. Technol., 2023, 52, 796–812.
40. M. S. Masnadi, R. Habibi, J. Kopyscinski, J. M. Hill, X. Bi, C. J. Lim, N. Ellis and J. R. Grace, Fuel, 2014, 117, 1204–1214.
41. Z. Sun, Y. Chen, B. Xi, C. Geng, W. Guo, Q. Zhuang, X. An, J. Liu, Z. Ju and S. Xiong, Energy Storage Mater., 2022, 53, 482–491.
42. W. Huining, W. Qiang, Y. Qi, L. Youpeng, W. Jianbo, W. Yuhua, Z. Hui and B. Hongcun, Clean Coal Technol., 2025, 31, 93–102.
43. B. Xing, C. Zhang, Y. Cao, G. Huang, Q. Liu, C. Zhang, Z. Chen, G. Yi, L. Chen and J. Yu, Fuel Process. Technol., 2018, 172, 162–171.
44. Y. Xu, E. Swaans, S. Chen, S. Basak, P. P. R. M. L. Harks, B. Peng, H. W. Zandbergen, D. M. Borsa and F. M. Mulder, Nano Energy, 2017, 38, 477–485.
45. Y. B. Mu, M. S. Han, J. Y. Li, J. B. Liang and J. Yu, Carbon, 2021, 173, 477–484.
46. Z. Q. Wang, S. Yuan, X. Y. Tang, H. L. Wang, M. Zhang, F. Liu, N. Yao and Y. Ma, Energy Storage Mater., 2023, 54, 284–293.
47. J. Wang, H. Zhao, T. Hu and F. Liu, Chin. J. Eng., 2020, 42, 884–893.
48. T. Wang, Y. Wang, G. Cheng, C. Ma, X. Liu, J. Wang, W. Qiao and L. Ling, Energy Fuels, 2020, 34, 8911–8918.
49. K. Jurkiewicz, B. Liszka, P. Gancarz, S. Smykala, D. Zygadlo, P. Nokielski, T. Lamrani, E. Talik, R. Wrzalik, M. Walkowiak and J. Ilavsky, Adv. Funct. Mater., 2024, 34, 2410409.
50. H. C. Jin, S. Xin, C. H. Chuang, W. D. Li, H. Y. Wang, J. Zhu, H. Y. Xie, T. M. Zhang, Y. Y. Wan, Z. K. Qi, W. S. Yan, Y. R. Lu, T. S. Chan, X. J. Wu, J. B. Goodenough, H. X. Ji and X. F. Duan, Science, 2020, 370, 192–197.

---