1 Introduction

Potassium-ion batteries (PIBs) have drawn widespread attention for their abundant raw materials and cost-effectiveness (Luo et al. 2024; Qiu et al. 2024; Zhang et al. 2023). Carbon materials have been widely utilized as anode materials in PIBs, due to their tunable structure, high specific surface areas, high electrical conductivity, and environmental friendliness (Ouyang et al. 2023; Shen et al. 2023; Xu et al. 2023; Yu et al. 2023). Graphite, commonly employed as anode materials in lithium-ion batteries, presents a specific capacity of 279 mAh g−1 when used in PIBs (Yin et al. 2023, 2021). However, the narrow interlayer spacing and huge volume expansion of graphite occurred during rapid charging and discharging cycles significantly impacted the cycling stability and overall lifespan (Destiarti et al. 2023; Zhao et al. 2023a, b). Hence, it is crucial to rapidly advance carbon-based anode materials for practical applications.

Pitch-based carbon materials (Chen et al. 2023; Hu et al. 2022; Ouyang et al. 2023; Zhang et al. 2021a, b), have attracted significant attention as promising electrode materials in the field of energy storage, especially for anodes in PIBs. As an affordable pitch-based soft carbon, needle coke (NC) exhibits a unique layered configuration that facilitates rapid intercalation and deintercalation of potassium ions (Zhu et al. 2020). However, several challenges and limitations would restrict its further applications (Liu et al. 2023a, b; Liu et al. 2020). For instance, the relatively larger radius of potassium ions and the compact carbon layers spacing present difficulties in efficient ion diffusion (Zhang et al. 2021a, b). Additionally, enhancing the electrochemical characteristics remains a formidable challenge. Despite the impressive electronic conductivity of NC, it may lack sufficient porosity and defect to accommodate a substantial number of potassium ions and facilitate the interlayer ion transport. To address these challenges, two strategies can be employed to enhance the performance of the NC anode in PIBs: (1) increasing the potassium storage capacity and accelerating ion transport by expanding the interlayer spacing; and (2) introducing more defects into the material to generate additional active sites, thereby providing more options for ion insertion and improving the overall performance.

Recent studies have revealed that the introduction of oxygen atoms not only induces defects but also serves as an active site for potassium ion storage (Chen et al. 2018). Lu et al. (2018) employed a pre-oxidation strategy to modify the microstructure of pitch-based carbon, contributing to a capacity increase from 94.0 to 300.6 mAh g−1 and Coulombic efficiency from 64.2% to 88.6%. Sun et al. (2024) suggested reducing hysteresis through oxygen-driven bulk defect engineering in carbon for rapid potassium storage at low voltages, achieving a reversible capacity of 192 mAh g−1 at 1 A g−1. Zhao et al. (2020) demonstrated that introducing oxygen functional groups into needle coke can augment sodium storage sites (C = O) and the distance between carbon layers from 0.344 nm to 0.384 nm, attaining a high reversible capacity of 385 mAh g−1 at 0.05 A g−1. It was reported that increasing the number of active sites on carbon anode surfaces can attribute to a superior adsorption capacity. After oxidation treatment, oxygenated functional groups can be incorporated into carbon materials (Qiu et al. 2023). Yin et al. (2021) using fly ash carbon (FAC), as an anode material for alkali metal-ion batteries, demonstrated that the charge storage mechanism of the FAC anode was shown to be intercalation coupled with redox reactions of oxygen functional groups. Moderate oxidation during the process can boost the electrochemical performance of carbon materials, while excessive oxidation may degrade this performance. For instance, an excess of defects can damage the material and compromise its electronic transport properties. Therefore, the manipulation of oxidation conditions is pivotal for augmenting the electrochemical attributes of carbon materials. Chemical oxidation serves as a quintessential approach to introduce defects into the structural integrity of carbon materials (Wang et al. 2022; Wu et al. 2023a, b). The prevalent strong oxidizing agents employed in various applications encompass potassium permanganate (KMnO4) (Wu et al. 2023a, b), nitric acid (HNO3) (Liu et al. 2023a, b) and hydrogen peroxide (H2O2) (Morales-Ospino et al. 2024). Among these oxidizing agents, H2O2 is used as the oxidizing agent because it is relatively mild, safe, and environmentally friendly compared to HNO3 and KMnO4. To the best of our knowledge, there have been limited reports on the oxidation of needle coke using H2O2 in PIBs.

In this work, oxygenated functional groups are incorporated into the NC through a single step of liquid phase oxidation using H2O2 with various concentrations. At the optimal oxidation level, it exhibits expanded carbon interlayer spacing, and abundant oxygenated functional groups and resultant defects on the surface of the NC serve as active sites for potassium storage and provide sufficient pathways for K+ migration. When applied as anodes for PIBs, the optimal oxidized NC achieves an enhanced reversible capacity, rate and cycle performance. This work offers novel insights for the industrialization and technological advancement of PIBs.

2 Materials and methods

2.1 Preparation of the ONC

Needle coke (NC), a typical pitch-based soft carbon material, was used as the raw material. It was firstly ball milled at 360 rpm for 8 h under an Ar atmosphere and then sieved with a 300-mesh sieve (less than 48 μm). Then, the sieved NC was washed with hydrofluoric acid and hydrochloric acid to eliminate the residual ash, followed by being washed with DI water to neutral and dried overnight in vacuum dried at 60 ℃. After drying, the deashed NC were added into H2O2 solutions with different concentrations, and stirred at room temperature for 12 h. After filtration and drying at 60 ℃ for 12 h, the solid samples were obtained and denoted as ONC-x (x = 3, 5, 7, 9) based on the concentration of H2O2 (3, 5, 7, and 9 mol L−1).

2.2 Materials characterizations

The high-resolution electron microscope (HR-TEM) was performed on a JEM 2100. The field emission scanning electron microscope (FE-SEM) was conducted on Nova 450. The X-ray Diffraction (XRD) was carried out using Rigaku D/Max 2400. The X-ray photoelectron spectroscopy (XPS) was performed using a Thermo ESCALAB 250. Raman spectroscopy was with a 532 nm laser. The N2 adsorption isotherms were obtained on a Micrometrics ASAP 2020 at 77 K. Elemental analysis was conducted by an elemental analyzer (Element, Vario EL III).

2.3 Electrochemical tests

Firstly, the NC and ONCs were sieved with a 300-mesh sieve (less than 48 μm). Subsequently, the slurries were prepared by mixing 70 wt.% deashed NC or ONCs, 20 wt.% acetylene black with polyvinylidene fluoride (PVDF, 10 wt.%) binder solution in N-methylpyrrolidinone (NMP). The slurry was homogeneously applied onto carbon-coated Al foils and vacuum dried at 60 ℃ for 12 h. The 2025 half cells were assembled in an Ar-filled glovebox (O2 and H2O below 0.1 ppm) with NC as active materials, K foil as counter anodes, and glass fiber mat as separators. The electrolyte was 0.8 mol L−1 KPF6 in a 1:1 (vol./vol.) mixture of EC/DEC. This amount of electrolyte for each cell is 180 μL. And the loading of the electrode is ~ 1.0 mg cm−2. Cyclic voltammetry (CV) and electrochemical impedance spectroscopy (EIS) tests were recorded on a VMP-300 multichannel electrochemical workstation with a scan of 0.1 mV s−1 at 0.01–3.0 V and a frequency of 100 kHz to 10 mHz, respectively. Galvanostatic charge/discharge (GCD) measurement was conducted on a NEWARE battery test system (CT-4008, Shenzhen, China) with a voltage window of 0.01–3.0 V. The apparent diffusion coefficient of K+ (DK) in the anodes was measured using the galvanostatic intermittent titration technique (GITT) by applying a pulse current of 0.1 A g−1 for 30 min between rest intervals for 2 h.

3 Results and discussions

The fabrication process of ONC-x samples is shown in Fig. 1a. Typically, the pulverized and deashed NC was added to H2O2 solution with different concentrations. This mixture was stirred for a duration of 12 h to facilitate an optimal reaction between the NC and H2O2 solution. During this process, the surface and edges of ONC-x were attacked by O2 produced by H2O2 decomposition. The distance between the carbon layers experienced an increase due to the high pressure exerted by expanding gases exceeding the van der Waals interaction forces between the interlayers (Sun et al. 2022). As shown in Fig. S1, the strong peaks (002) at around 26° demonstrate that the samples have a better graphitic crystal structure. Upon magnifying the pattern in Fig. 1b, it is evident that as the degree of oxidation increases, the width of half-maximum (FWHM) progressively widens compared to that of the NC, and the diffraction peaks (002) gradually shift to lower angles. In Table S1, the structural characterizations of oxidized NC were calculated by XRD pattern. From the results of structural characterizations, the d002 of ONC-5 was larger than that of NC, presenting that liquid oxidized treatment could make the interlayer spacing expand. This suggests that oxygenated functional groups may be incorporated onto the surface of the ONC-x samples, with an increase in carbon layer spacing (Sun et al. 2022). The increasing d002 potentially facilitates the potassiation/depotassiation to a certain extent.

Fig. 1
figure 1

a The Schematic illustration of the fabrication process of the ONC-x; (b) XRD pattern, (c) Raman spectra, and (d) N2 adsorption/desorption isotherms and pore size distribution of all samples. (e) SEM, (f) HR-TEM, and (g) the lattice spacing images of ONC-5

Full size image

Raman spectroscopy serves as an effective method for the characterization of carbon materials. As shown in Fig. 1c, the G band can be assigned to the degree of graphitization of NC, and the D band represents the defects and impurities of the NC. Thus, the ratio of the D-band to the G-band intensities (AD/AG) can serve as an indicator for evaluating the defect density in the sample (Zhang et al. 2021a, b). The AD/AG values are 1.8, 2.2, 2.3, 1.6, and 1.5 for NC, ONC-3, ONC-5, ONC-7, and ONC-9, respectively. The AD/AG value of ONC-5 sample was the biggest among all the samples, indicating that the ONC-5 sample had the highest defect density. The result demonstrates that the oxidation treatment successfully introduced defects in ONC-x samples, and the defect density in ONC-5 was the highest among all the samples. Subsequently, the D band was further fitted into three bands (Fig. S2 and Table S2), demonstrating that the irregular section of the ONC-x samples was constitutive of disordered graphitic lattice from edges (D1 ~ 1351.6 cm−1) and surface (D2 ~ 1482 cm−1), as well as amorphous carbon (D3 ~ 1213 cm−1) (Table S2; Liu et al. 2022). D1 presents a higher ratio of area than that of D2 and D3, suggesting that graphene edge sites primarily contributed to the observed defects. From Table S2, it is evident that both values of AD2/AG and AD3/AG of ONC-5 sample significantly increased in comparison with the other ONC-x samples, indicating an increased defect sites in ONC-5 sample. Therefore, ONC-5 may possess more active sites for potassium ion storage, improving its potassium storage performance.

As shown in Fig. 1d and Fig. S3, as the oxidation degree increased, the specific surface area of the ONC-3 and ONC-9 increased from 5.6 m2 g−1 up to 7.9 m2 g−1, and that of the oxidized samples was greater than that of the untreated NC sample (4.4 m2 g−1). Furthermore, there was a marked increase in both micropores and mesopores in the ONC-x samples in Fig. 1d. In contrast, the number of mesopores and micropores in the ONC-9 sample diminished, with a notable presence of pores measuring between 40–60 nm. This suggests that the excessive oxidation could be attributed to the high concentration of H2O2 solution. The SEM images in Fig. 1e and Fig. S4 clearly demonstrated the flake stacked structure of the ONC-5 sample and other ONC-x samples. The microstructures of the NC and ONC-x samples were further examined through TEM analysis, as shown in Fig. 1f and Fig. S5. Figure 1g displays a distinct carbon lattice stripe in the ONC-5 sample, with a lattice spacing of 0.351 nm. However, as the H2O2 solution concentration was over 5 mol L−1, the lattice fringes gradually disappeared, indicating that the carbon layer structure was damaged.

In Fig. S6a, the XPS spectra displays two primary peaks at 284 and 531 eV, corresponding to O 1 s and C 1 s, respectively. As depicted in Fig. S6b, the peak intensity and peak area of the O 1 s spectra increase with the degree of oxidation, which indicates an augmented content of oxygen elements. Figs. S7 and S8 present the fitting curves for the C 1 s and O 1 s spectra for NC, ONC-3, ONC-5, ONC-7, and ONC-9, respectively. In Fig. S7, the C 1 s spectra for the NC can be segmented into three distinct peaks at 284.0, 285.1, and 288.7 eV, which correspond to C–C, C = O, and HO-C = O bonds, respectively (Xia et al. 2015; Zhao et al. 2023a, b). Moreover, a new peak emerged at approximately 284.5 eV can be associated with the C-O bond. As shown in Fig. S9, the O 1 s was then fitted by three peaks at 533.0, 532.0, and 530.8 eV, corresponding to HO-C = O, O-C-O/C–OH, and C = O bonds, respectively. To analyze variations in oxygen content, the ratio of the fitting area of the O 1 s peak to the XPS spectrum is presented in Table S3. Notably, the ONC-5 sample exhibits the highest content of superficial oxygenated functional groups. These findings suggest that the ONC-5 sample contained a greater number of oxygen functional groups acting as active sites, potentially leading to superior electrochemical performance compared to other oxygen samples. In Fig. S9, the elemental analysis reveals that the ONC-5 sample possessed the highest oxygen content, aligning with the XPS results.

Figure 2a–e displays the second discharge curves of NC and ONC-x anodes at 0.05 A g−1. Soft carbon, as the anode, could storage potassium ions by the mechanisms of adsorption and intercalation. In PIBs, the regions of adsorption and intercalation can be divided by the voltage during charge and discharge process (Zhang et al. 2020). In general, the adsorption reaction usually occurs in the higher voltage region, while the intercalation reaction occurs in the lower voltage region. Combining with CV curves (Fig. 3a and Fig. S10), there is a cathodic peak in the low voltage region, which corresponds to the reaction of potassium ions intercalation/deintercalation. Therefore, the intercalation/deintercalation of K-ions mainly occurs in the low voltage region (< 0.4 V), whereas above 0.4 V, the capacity is attributed to the K+ adsorption mechanism (He et al. 2023; Zhang et al. 2019; Zhong et al. 2022). Figure 2a presents the discharge curve for the NC electrode, revealing an adsorption capacity of 32.2 mAh g−1 and an intercalation capacity of 205.8 mAh g−1. After the oxidation treatment, the adsorption and intercalation capacities of the ONC-x anodes significantly enhanced compared to the original NC anode. The adsorption capacities of the ONC-3, ONC-5, ONC-7, ONC-9 anodes were 71.1, 80.2, 62.6, and 54.8 mAh g−1, respectively, while their corresponding intercalation capacities were 263.8, 277.2, 249.2, and 252.5 mAh g−1, respectively. The adsorption and intercalation capacities of the NC and other ONC-x anodes are summarized in Fig. 2f. The observed capacity variation trend indicates that electrochemical performance correlates proportionally with the concentration of the oxidizing agent when the H2O2 solution concentration remains below 5 mol L−1. However, when the H2O2 solution concentration exceeds 5 mol L−1, there is an inverse correlation between sample capacity and oxidation degree. Based on these findings, it can be postulated that the oxidation process may compromise the structural integrity of NC to some extent, which is not conducive to preserving the structural characteristics of the ONC-7 and ONC-9 anodes (Liu et al. 2022; Zhong et al. 2015).

Fig. 2
figure 2

GCD curves of the second cycle at 0.05 A g−1 during discharge process of (a) NC, (b) ONC-3, (c) ONC-5, (d) ONC-7, and (e) ONC-9. (f) Summary of the intercalation and adsorption capacities

Full size image
Fig. 3
figure 3

a CV curves of ONC-5; (b) GCD curves of ONC-5 at 0.05 A g-1; (c) rate performance and (d) cycle performance of NC and ONC-x anodes

Full size image

Figure 3a and Fig. S10 illustrate the CV curves conducted within 0–3 V, with a scan rate of 0.1 mV s−1. In Fig. 3a, two cathodic peaks in the initial CV curve of the ONC-5 anode are observed at approximately 0.64 V and 0.31 V, suggesting the formation of solid electrolyte interphase (SEI) films and decomposition of the organic electrolyte. The subsequent cycles of the ONC-5 anode closely overlap with the subsequent cycles, indicative of its high reversibility. Additionally, a redox peak around 0.01 V is evident in the CV curves of all anodes, signifying the reversible nature of the potassium intercalation/de-intercalation process. Figure 3b presents the galvanostatic charge and discharge (GCD) curves of ONC-5 anode at the current density of 0.05 A g−1. For the ONC-5 anode, the initial coulombic efficiency (ICE) is 55.7%, which is higher than the ICE (43.5%) of the NC anode in Fig. S11. As shown in Fig. S12, from the charging and discharging curves with different rates, it can be seen that the capacity of NC anode decays rapidly, while the capacity of ONC-5 anode can maintain high voltage and capacity, indicating that the oxidized NC anode has a better electrochemical performance. Figure 3c illustrates the electrochemical rate performance of all anodes at the current density from 0.05 to 2 A g−1. The electrochemical performance of ONC-5 anode shows a reversible capacity of 322.7 mAh g−1 at 0.05 A g−1. In detail, the reversible capacities for ONC-3, ONC-7, and ONC-9 anodes were recorded as 289.5, 289.4, and 260.6 mAh g−1, respectively. Furthermore, it was noted that the capacity of the ONC-5 anode could reach up to 98.9 mAh g−1 at 2 A g−1, nearly twice less than that of the NC anode. Upon returning to a current density of 0.05 A g−1, the reversible capacity of the ONC-5 anode recovered to 277.5 mAh g−1, highlighting its superior rate performance. And when the current density returned to 0.05 A g−1, the reversible capacities of ONC-3/7/9 were 266.9, 272.0, and 1.3 mAh g−1, respectively. Though the electrochemical performance across all anodes exhibited minimal variation at 0.05 A g−1, the ONC-5 anode demonstrated superior rate performance at 2 A g−1. Notably, the ONC-9 anode lost its flexibility after operating at this higher current density, hypothesizing that the ONC-9 anode may have been dislodged from the collector. As shown in Fig. 3d, the performance of ONC-5 anode exhibited excellent reversible capacity of 106.4 mAh g−1 at 1 A g−1 after 500 cycles. While the capacities reached 17.2, 36.3, 63.2, and 34.4 mAh g−1 for NC, ONC-3, ONC-7, and ONC-9, respectively. It is evident that as oxidation increases, electrochemical performance increases as H2O2 below 5 mol L−1, and decreases above 5 mol L−1. In Fig. S13 , compared with the relevant reported researches (Xiao et al. 2023; Yu et al. 2022; Ma et al. 2021; Wu et al. 2023a, b; Jian et al. 2017; Chang et al. 2023; Mohamed et al. 2023), the electrochemical performance of ONC-x anodes was higher than others. This enhancement may be attributed to the introduction of oxygen functional groups, which create more active sites on the surface of ONC-x anodes. Additionally, increasing carbon layer spacing positively impacts the electrochemical performance. Nonetheless, excessive oxidation can lead to diminished electrical conductivity (Xia et al. 2015), consequently reducing electrochemical performance.

To further investigate intercalation/de-intercalation mechanism, CV test at different scan rates, galvanostatic intermittent titration technique (GITT), and electrochemical impedance spectroscopy (EIS) were conducted. As shown in Fig. 4a and Fig. S14, the CV curves of ONC-5 anode and other anodes were tested at the scan rates ranged from 0.1 to 1.0 mV s−1. As shown in Fig. S16, the capacitive contributions of all anodes were integrated to obtain the diffusion and capacitive capacity at different scan rates. Figure 4d shows that the capacitive contributions of NC, ONC-3, ONC-5, ONC-7, and ONC-9 anodes were 14%, 37%, 59%, 72%, and 51%, respectively at 0.4 mV s−1. Among these anodes, the NC anode exhibited the lowest capacitive contribution while the ONC-5 anode demonstrated the highest. This can be attributed to the structural integrity and minimal defects of the NC anode, which leads to the intercalation of the potassium storage to contribute to a large extent to the capacity. Additionally, the ONC-5 anode contains more active sites enhancing its adsorption capacity for potassium ion storage. Following oxidation treatment, an abundance of active sites and a large specific surface area on the surface can augment the capacitive contribution ratio.

Fig. 4
figure 4

a CV curves of ONC-5 at different scan rates in the range of 0.1 and 1.2 mV s-1 and (d) the capacitive contributions of all samples at different scan rates. b GITT curves, (e) diffusion coefficient, (c) EIS spectra, and (f) the Warburg coefficient in the low-frequency region of all samples

Full size image

As shown in Fig. 4b, the Fick second law was conducted to calculate the diffusion coefficient of K+ (DK). According to the potential-diffusion coefficient curve (Fig. 4e), the DK value of the ONC-5 anode is the highest during charge/discharge process. This can be attributed to the increased interlayer spacing and surface defect density. EIS was employed to investigate the kinetics of charge carrier transport. In Fig. 4c, the diameter of the measured semicircle represents the charge transfer resistance (Rct), while the slope of the straight line in the low-frequency part may suggest the ion diffusion resistance (Fan et al. 2018). Based on the equivalent circuit, the Warburg constant (σ; Fig. 4f; Feng et al. 2022) was identified as the slope of the linear relationship between the Z’ versus ω−1/2 (Eq. [1]) in the low frequency region. The diffusion of K+ was calculated by using the Eq. (2) (Zhao et al. 2020) in Table S4.

$${{\text{Z}}}^{\mathrm{^{\prime}}}={R}_{S}+{R}_{ct}+{\sigma }^{-{}^{1}\!\left/ \!{}_{2}\right.}$$
(1)
$${D}_{K}=0.5\times {(nF)}^{-4}{C}^{-2}{\sigma }^{-2}{R}^{2}{T}^{2}{A}^{2}$$
(2)

where σ is the Warburg constant, Rs is the electrolyte resistance, and Rct refers to the charge transfer resistance at the electrolyte-anode interface in the Eq. (1). In Eq. (2), DK represents the K+ diffusion, n is the number of transfer electrons, F is Faraday’s constant, C represents the concentration, T refers the temperature, and A represents the anode area with the electrolytes. The results demonstrate that the Rct increases with the increasing oxidation level, potentially due to the oxygen functional groups on the surface of ONC-x anodes that hinder the charge transfer process. It was observed that within an optimal concentration range, the DK increases proportionally to the degree of oxidation. This suggests that the expansion of carbon layer spacing facilitates the movement of K+ between adjacent carbon layers more efficiently.

Based on the aforementioned kinetic results, the ONC-5 anode demonstrates superior kinetic properties and K+ storage performance. These outcomes can be attributed to the following three factors: (1) after oxidation reaction, the defects provide active sites for electron transfer and K+ storage, thereby enhancing electrochemical performance; (2) the increased spacing between carbon layers facilitates rapid K+ intercalation; (3) the ONC-5 anode mitigates charge transfer resistance with the moderate oxidization treatment. Potassium ions can be stored across the adjacent carbon layers within the ONC-5 anode through its edges and defects, contributing to a higher reversible capacity and K+ transfer kinetics (Fig. 5). In contrast, the NC anode features relatively parallel carbon layers, resulting in fewer storage active sites and slower electrochemical kinetics compared to the ONC-5 anode. However, the ONC-9 anode contains an excessive number of defects, which reduce the conductivity of the carbon material and result in an inferior potassium storage performance.

Fig. 5
figure 5

Schematic diagrams of fast electrical conductivity and ion migration in the NC, ONC-3, ONC-5, ONC-7, and ONC-9 anodes

Full size image

4 Conclusion

In conclusion, the direct oxidation method enables the introduction of oxygenated functional groups onto the surface of soft carbon materials through H2O2 oxidation. The oxygen produced during this process can contribute to the swelling of the carbon layer. Oxidation creates negatively charged oxygenate functional groups, which readily adsorb positively charge potassium ions. The oxidation treatment could adjust the structure of NC, thus creating defects and channels, enhancing the diffusion rate of potassium ions, and accelerating the storage rate of potassium ions. According to the structural characterizations of ONC-x anodes, these oxygenated functional groups can serve as active sites for potassium ions adsorption, the introducing defects improve the conductivity and the expansion of interlayers spacing benefits for the improvement of K+ storage. The results show that the optimal oxidized ONC-5 maintains a high reversible capacity of 322.7 mAh g−1 at 0.05 A g−1. However, excessive oxidation may compromise the NC structure, resulting in an inferior potassium storage performance. However, excessive oxidation may compromise the NC structure, resulting in an inferior potassium storage performance. This work offers a feasible approach to enhance the potassium storage performance of soft carbon anodes.