Nanocarbon was acquired straight from shanghai naiou nano technology Co., Ltd. This kind of nanocarbon is characterized by high compact density, suitable surface area, and readily bonds with F atoms. The following is the synthetic procedure of FNC: a suitable volume of nanocarbon powder was added to a Ni reactor, and fluorination was performed using a NF₃ gas at a specific pressure and temperature range, divided into four temperature (300, 330, 360°C and 400°C) and pressure (0.15, 0.20, 0.25,0.30 MPa) intervals for the fluorination reaction. The reaction time was at least 3 h to ensure complete reaction. The F/C ratio of the fluorination product increased as the temperature and pressure during fluorination increased, resulting in FNC-0.80, FNC-0.90, FNC-1.00, and FNC-1.05. Using the weighted method, the F/C ratio was computed using the following Equation 1. In the equation, W _c denotes the mass of the carbon source utilized in the experimental setup, ΔW represents the differential variation in the mass of the carbon source, specifically measured before and subsequent to the process of fluoridation. F C = ∆ W / 19 W c / 12 (1)

Field Emission Transmission Electron Microscopy (FETEM, JEOL JEM-F200) and Field Emission Scanning Electron Microscopy (FESEM, Apreo S LoVac, Thermo Corporation) were used to investigate the morphology and microstructure of materials. An X-ray diffractometer (Miniflex 600, Rigaku) equipped with Cu Kα radiation (λ = 1.5406 Å, 15 mA, and 40 kV) was used to analyze the crystal structure of the FNC. Chemical bonding investigations were conducted via Fourier Transform Infrared Spectroscopy (Nicolet iS 50, Thermo Corporation), a Thermo Scientific K-Alpha analyzer with a monochromated Al Kα source was used for X-ray photoelectron spectroscopy (XPS) study, while a LabRAM HR Evolution (Horiba) with laser wavelengths of 325 and 532 nm, and a wavenumber range of 200–3,000 cm^–1, was used for Raman spectroscopy. Before using an Autosorb-iQ2-MP gas adsorption equipment to record N₂ adsorption-desorption isotherms, the FNC underwent degassing at a temperature of 150°C, with the process continuing until a pressure of 2 mmHg (267 Pa) was attained. To quantify its specific surface area, the Brunauer-Emmett-Teller (BET) methodology was employed.

The electrochemical testing was performed through a constant current discharge process. For the cathode electrode fabrication, a slurry mixture consisting of FNC (80 wt.%), carboxymethyl cellulose sodium (10 wt.%), and acetylene black (10 wt.%) was prepared, utilizing deionized water as the solvent. After that, this slurry was spread over aluminum foil. Subsequently, the electrode was dried in a vacuum oven at 90°C for a duration of 12 h in order to remove residual solvents, followed by cutting into 14 mm disks, achieving an active material loading of 1.25 mg cm^–2.

An evaluation of the electrochemical performance was conducted by assembling a coin cell. Lithium metal was used to create coin cells (CR2016) as the anode material, and 1.0 M LiFSI in propylene dimethoxyethane (DME) and propylene carbonate (PC) (1:1 vol) acting as the electrolyte, alongside Celgard 2500 as the separator. Coin cells (CR2025) were fabricated utilizing sodium and potassium metal as the anode material, with 1.0 M NaPF₆ in PC and ethylene carbonate (EC) (1:1 vol), and 1.0 M KPF₆ in ethylene EC and diethyl carbonate (DEC) (1:1 vol), functioning as the electrolyte. Glass microfiber filters were employed as the separator. Assembly occurred within an argon-filled glovebox (H₂O < 0.01 ppm, O₂ < 0.01 ppm). Discharge tests were conducted at 25°C utilizing the MIHW-200-160CH (Shenzhen Neware Electronics Co., China) battery testing system, with a cutoff voltage set at 1.5 V.

The cell was discharged at a rate of 1C for 1 min using the galvanostatic intermittent titration technique (GITT), and then it was allowed to relax for 120 min. In the sections that follow, the parameters pertaining to the CF _x electrode’s Li⁺ diffusion coefficient are presented.

EIS measurements were conducted utilizing a CHI760E electrochemical workstation (Chenhua Instrument Co., China), with a frequency span from 1 MHz to 0.01 Hz, with an amplitude set at 5 mV.

Figure 1 depicts the process for preparing FNC, commencing with the introduction of ANC into the reactor, followed by the infusion of NF₃ for thorough reaction., and the synthesis of FNC with varying F/C ratios was achieved by adjusting the temperature and pressure conditions.

As depicted in Figures 2A, B, numerous spherical nano-carbon particles aggregate to form ANC with sizes ranging from 2 to 5 μm. The aggregate nanocarbon particles exhibit a honeycomb-like structure characterized by a considerable amount of loosely packed voids, which offers unique advantages in terms of physical and chemical properties. Primary ANC particles have a particle size of around 30 nm, according to the TEM picture (Figure 2C), with high disorder observed in the high-resolution view (Figure 2D). Similarly, FNC preserves the original morphology, particle sized and crystalline structure of ANC (Figures 2E–H). The preservation of the inherent structural integrity is crucial, as it facilitates the efficient transport of Li⁺ through the inner of material. Furthermore, the maintenance of the crystalline structure ensures optimal interfacial contact between the F-Li, thereby enhancing the electrochemical performance through improved charge transfer kinetics and effective utilization of active sites. Furthermore, the homogeneous distribution of fluorine and carbon is plainly visible on the FNC-1.0’s high-angle annular dark-field scanning TEM (HAADF-STEM) (Figure 2I). According to the aforementioned findings, the original morphology of ANC was not damaged during fluorination process.

Figure 3A portrays the XRD patterns of raw ANC and FNC materials. The interlayer (002) and (100) reflections cause the ANC material to show notable broad diffraction peaks at 24° and 44°, indicating of a disordered carbon structure. The reflection line in FNC around 13° is the (001) reflection of a high-level fluorinated graphene system, but the markedly reduced intensity and the dramatically broadened width indicate the loss of interlayer correlations between fluorinated carbon sheets (Peng et al., 2019). During fluorination, the peaks associated with the (001) crystal plane exhibit a shift towards a lower angle. This peak shift serves as an indicator of the progressive fluorination depth, manifesting in a gradual enlargement of the interplanar spacing. Concurrently, a reduction in the peak half-width is observed, suggesting a weakening of the interlayer interactions between the FNC layers (Sun et al., 2014). Utilizing the Bragg equation, the calculated interlayer spacings for FNC-0.80, FNC-0.90, FNC-1.00, and FNC-1.05 were determined to be 0.693, 0.703, 0.714, and 0.736 nm, respectively.

The Raman spectrum of ANC (Figure 3B) exhibits two characteristic bands: a defect-induced D band centered at 1,319 cm^–1 and a crystalline graphite G band centered at 1,603 cm^–1, indicating the presence of an amorphous structure in ANC (Xu and Buehler, 2009; Rajagopal et al., 2017). The fluorinated FNC possesses pronounced fluorescence capabilities, while the Raman spectrum reveals an absence of distinct Raman bands (Yue et al., 2021). Under the influence of fluorescence, Figure 3B reveals two relatively weak characteristic peaks. Firstly, the peak at 1,323 cm^–1 corresponds to the D band of carbon materials, the peak at 1,591 cm^–1 corresponds to the G band, with its shift attributed to the shortened C–C bond lengths.

The chemical composition of ANC and FNC was analyzed using FTIR spectroscopy, as depicted in Figure 3C, to elucidate their respective characteristics. The FTIR spectrum of ANC indicates a lack of active functional groups, with the spectrum showing minimal distinctive features. The utilization of FTIR characterization techniques enabled the investigation of the chemical bonds present in FNC. FTIR characterization was utilized to investigate the chemical bonds of FNC. The detection peaks at 1,341 and 1,222 cm^–1 signifies the stretching vibration of–CF₂/CF₃ groups and covalent C–F bonds, respectively. Furthermore, FNC-0.80, FNC-0.90, and FNC-1.00 consistently exhibit a characteristic peak at 1,160 cm^–1, which corresponds to the stretching vibration of the semi-ionic C–F bond. This observation provides substantial evidence for the notable concentration of C–F semi-ionic bonds within these materials (Sun et al., 2014; Lee et al., 2003; Guérin et al., 2004).

For the purpose of evaluating the pore structure and quantifying the specific surface area (SSA) of the samples, nitrogen adsorption-desorption isotherms were analyzed at a temperature of 77 K. The nitrogen adsorption-desorption isotherm analysis of ANC and FNC, as depicted in Supplementary Figure S1A, reveals a type II profile without a discernible hysteresis loop throughout the curve, suggesting a non-microporous nature of the materials. The micropore volume of ANC and FNC is very small (Supplementary Table S1), drawing from density functional theory (DFT) calculations, the cumulative pore volumes of micropores for FNC-0.80, FNC-0.90, FNC-1.0, and FNC-1.05 are 0.042, 0.044, 0.035, and 0.032 m³ g^–1, respectively. This further confirms that FNC are non-microporous nanoscale carbon. With the continuous increase in fluorination temperature, the SSA of the corresponding synthesized FNC gradually decreases, which is attributed to the destruction of the FNC structure during the fluorination process. Furthermore, the pore size distribution graph (Supplementary Figure S1B) of ANC and FNC reveals the absence of pores, affirming the non-porous nature of the material.

Additionally, the surface composition of FNC was ascertained by assessing the cumulative intensity of characteristic lines within the XPS spectrum (Supplementary Figure S2). The high-intensity F signal confirmed the successful fluorination of ANC. In alignment with the FTIR analysis findings, the high-resolution C 1s spectra of FNC (Figures 3D–G) validated the existence of carbon-based components with varying compositions and configurations. The deconvoluting of the high-resolution C 1s spectra for the four samples disclosed distinct peaks, with assignments at 284.78, 287.12, 289.11, 290.4, 291.74, and 292.7 eV corresponding to sp² C=C, sp³ C–C, semi-ionic C–F bonds, covalent C–F bonds, –CF₂, and–CF₃ bonds, respectively (Sun et al., 2014; Robinson et al., 2010; Nair et al., 2010). Upon a comparative analysis of the data, it becomes evident that a substantial fraction of carbon atoms is covalently bonded to fluorine atoms through sp³ hybridization. The comprehensive results pertaining to each carbon component are tabulated in Supplementary Table S2. Moreover, quantitative analysis of the high-resolution spectra of F 1s for FNC was performed, and the results are presented in Supplementary Figures S3A–D. Correspondingly, the fittings at 687.35 eV, 688.70 eV, and 690.03 eV correspond to the covalent bonding of fluorine atoms with sp²-hybridized carbon atoms, sp³-hybridized carbon atoms, and perfluorinated functional groups formed with carbon, respectively. The detailed content and results of various fluorine compositions are provided in Supplementary Table S3. Additionally, the atomic ratio of F/C obtained through XPS analysis was found to be close to the F/C ratio obtained through gravimetric analysis. It is observed that with increasing F/C ratio, the presence of conjugated components in the FNC diminishes, the relative abundance of perfluorinated functional groups steadily increases.

FNC was assembled as the cathode of a lithium metal primary battery in the form of a coin cell to test and characterize its electrochemical performance. The galvanostatic discharge profiles of FNC at 0.01C are shown in Figure 4A. The specific capacities of FNC-0.80, FNC-0.90, FNC-1.00, and FNC-1.05 are 722, 731, 765, and 798 mA h g^–1, respectively. Based on the aforementioned data that nearly all C–F bonds participate actively in the discharge process. The corresponding Ragone plot and different rate energy density graph of the prepared FNC cathodes (Figures 4B, C) both demonstrate that FNC has excellent rate capability. Distinct from other nano scaled CF _x materials, the FNC cathode exhibits superior rate capability and supports high-power output, which is attributed to its aggregated spherical structure, wherein loosely packed and dense voids provide more continuous diffusion pathways for Li⁺ transportation. At 20C, the FNC-1.00 exhibits a power density of 33,694 W kg^–1 and an energy density of 1,250 Wh kg^–1. The specific capacity is 578 mA h g^–1 at 20C, corresponding to a 75.6% retention of the specific capacity at 0.01C.

The F/C ratio exhibited a negative correlation with the discharge voltage but a positive correlation with the specific capacity. Given the higher bond energies of–CF₂ and–CF₃ groups, which manifest their inert nature during the discharge process, perfluorinated groups do not contribute to the enhancement of the material’s capacity, but rather decrease the activity of FNC. An increase in the content of perfluorinated groups can exacerbate ohmic polarization and reduce the discharge platform. In addition, high F/C ratio diminishes electrical conductivity, heightens ohmic polarization and lowers discharge voltage (Zhang et al., 2010; Dubois et al., 2012). Consequently, all reported CF _x products demonstrate distinct voltage delays following discharge. It is observed that the ohmic polarization effect of FNC is relatively weak in the initial discharge stage, and there is no significant voltage delays, contrasting with other CF _x materials (Supplementary Figure S4) (Watanabe, 1980). In the case of FNC-1.05 with a higher F/C ratio, there is almost no voltage delay phenomenon observed at low discharge rates (<0.5C). As the discharge rate increases (>1C), minor voltage delays emerge in FNC with varying F/C ratios. Even the discharge rate is escalated to 20C, FNC does not exhibit a more severe voltage delay. In addition, the high F/C ratio of FNC presents a higher discharge plateau. The median voltage of the FNC was compared against the discharge median voltages (E _1/2) for all reported nanoscale CF _x materials (Figure 4D, and detailed information presented in Supplementary Table S4), where in FNC demonstrates a higher discharge voltage (Jiang et al., 2021; Ahmad et al., 2015; Li et al., 2019; Yang et al., 2021; Liu et al., 2023; Li et al., 2021; Peng et al., 2023; Ahmad et al., 2017; Hou et al., 2022). The E _1/2 of FNC are at 3.20, 3.21, 3.13, and 2.88 V (vs Li⁺/Li) at 0.01C, which are significantly higher than that of commercial FG (2.71 V vs Li⁺/Li) (Chen N. G et al., 2023). FNC-1.0, characterized by a high F/C ratio, and demonstrates a high E _1/2 of 3.13 V. This phenomenon is attributed to the higher content of semi-ionic C–F bonds in FNC, thereby enhancing the electrochemical activity and conductivity of the material.

To further validate the rate capability of FNC, the diffusion efficiency of Li⁺ was analyzed using GITT, and the Li⁺ diffusion coefficient (D _Li ⁺) was utilized as an assessment metric (Figure 5A). If the applied current is extremely low and the relaxation time is exceedingly brief, the following equation can be used to calculate the D _Li ⁺(Deiss, 2005): D L i + = 4 Π I 0 V m SF Z i 2 d E d δ d E d τ 2   τ ≪ L 2 D L i + (2)

The chemical diffusion coefficient of the mobile species, denoted as D _Li ⁺ (cm² s^–1), is determined by considering multiple parameters. These include the applied current I ₀(A), the molar volume V _m (cm³ mol^–1), and the thickness L (cm) of the electrode material. Additionally, the Faraday constant F, the total contact area S (cm²) between the electrolyte and electrode, and the number of charge transfer Z _i in the discharge reaction of FNC are also taken into account for the precise calculation of D _Li ⁺. Utilizing the Equation 2, the D _Li ⁺ was calculated, and Figure 5B demonstrates the gradual variation of D _Li ⁺ values during the discharge process. In the initial stage of discharge, the D _Li ⁺ value gradually increases due to the generation of carbon from the discharge of CF _x . As the voltage ranges from 3.0 V to 1.5 V, the particle surface progressively becomes saturated with LiF, impeding ion transport, as reflected in the gentle increase trend of D _Li ⁺ values. Overall, compared to traditional CF _x materials, FNC possesses a plethora of highly active semi-ionic C–F bonds, which demonstrate higher D _Li ⁺ throughout the entire discharge process, effectively mitigating the voltage delay caused by ohmic polarization. This characteristic contributes to its excellent rate capability. It is noteworthy that the D _Li ⁺ value of FNC-1.00 in the 2.5 V–1.5 V range is higher than that of FNC-1.05. These results demonstrate that the average D _Li ⁺ value of FNC-1.00 at 1C is higher than that of FNC-1.05, confirms its superior power performance.

The underline mechanism of the excellent electrochemical performance of FNC materials was further analyzed and discussed. Electrochemical impedance spectroscopy (EIS) was recorded for FNC electrode (Supplementary Figure S5A–D) at different depths of discharge (DOD) at 0.1C (cut-off 1.5 V). The Nyquist plot of the FNC cathode exhibits an approximate semicircular feature in the high-frequency region and a slanted line feature in the low-frequency region. This semicircle represents the charge transfer process, and as the discharge depth increases, the decrease in the radius of the semicircle indicates a reduction in the charge transfer resistance (Zhang et al., 2009b; Jiang et al., 2017). Based on the analysis of the equivalent circuit model depicted in the inset of Figure 5C, R _b signifies the ohmic resistance, while R _ct denotes the charge transfer resistance, Z _w represents the Warburg impedance, and CPE stands for the double-layer capacitance (Zhang et al., 2009a; Qian et al., 2022). The R _b and R _ct significantly impact the discharge performance of the battery. For materials with a high F/C ratio, the high content of–CF₂/–CF₃ groups within the material, which are considered inert fluorine-containing groups, significantly impedes charge transfer (Zhong et al., 2018). These groups, due to their nonconductive nature, impede the transfer of electrons, thereby manifesting a greater charge transfer resistance. Consequently, CF _x materials with a high F/C ratio exhibit a higher interfacial transfer impedance. With the increase in discharge depth, the resistance R _b of FNC gradually decreases after fitting (Figure 5C), attributed to the enhanced conductivity resulting from carbon formation during the reaction. As the discharge progresses, the R _ct of FNC remains relatively stable with minimal variation (Figure 5D). However, compared to the initial stage of discharge, the R _ct decreases as the reaction proceeds. This trend aligns with previous studies on CF _x material (Chen P. et al., 2023). At 100% DOD, R _ct does not rise sharply, which also confirms that CF _x material can continue to react at 1.5 V.

To further validate the aforementioned analysis of the discharge mechanism of FNC, the distribution of relaxation time (DRT) was employed to analyze the EIS impedance across various DOD and timescales (Figures 5E–H, Initiating the DRT diagnosis process involves the utilization of EIS measurement across a wide frequency spectrum) (Rashid and Gupta, 2015). Initially, it is observed that the relaxation time constant τ exhibits minimal impedance response within the interval of (−5 ≤ Logτ ≤ −1), suggesting the absence of solid electrolyte interphase (SEI) formation throughout the discharge process of FNC, thereby excluding consideration of R _sei. The relaxation time constant τ falls within the interval of (0 ≤ Logτ ≤2), indicating the charge transfer process. It manifests a pronounced impedance signal during the initial discharge stage, gradually diminishing as the discharge deepens, aligning with the aforementioned analysis findings. Overall, owing to the ultra-low conductivity of CF _x material, the charge transfer impedance is predominantly influenced by conductivity.

Subsequently, we fabricated new SPBs employing the FNC as cathode material. Equation 3, proposed by Liu et al., represents the discharge mechanism of SPBs, which is similar to that of Li/CF _x primary batteries (Liu et al., 2014): C F x + x N a + + e ‐ → C + x N a F (3)

Notably, the discharge voltage of FNC exhibits a notable reduction of approximately 250 mV in SPBs compared to LPBs. This discrepancy is attributed to the difference in the standard electrode potentials between lithium (Li⁺/Li, −3.02 V vs. SHE) and sodium (Na⁺/Na, −2.71 V vs. SHE). Moreover, compared to LPBs, SPBs exhibit less prominent rate capability, which is likely attributed to the larger ionic radius of Na⁺ (1.02 Å) compared to Li⁺ (0.76 Å), which results in more difficult transport of Na⁺(Li et al., 2019). Although limited by the characteristics of SPBs, FNC retains a high discharge voltage and exhibits excellent electrochemical performance, as illustrated in Figures 6A, C. Equation 4, proposed by Zhong et al., represents the discharge mechanism of PPBs, which is similar to that of the LPBs and SPBs batteries (Yue et al., 2021): C F x + x K + + e − → C + x K F (4)

As a cathode material, FNC exhibits excellent electrochemical properties in PPBs (relevant data are shown in Figures 6B, D; Supplementary Table S4). This performance can be attributed to the high standard electrode potential of potassium (K⁺/K, −2.93 V vs. SHE). In PPBs, FNC demonstrates high discharge voltage and excellent rate capability, with maximum capacity and energy density levels approaching those of LPBs. At a rate of 20C, it maintains sufficient stability for discharge, achieving energy density and power density of 738 Wh kg^–1 and 22,363 W kg^–1, respectively. The phenomenon indicates that FNC in various alkali-metal/CF _x primary batteries does not exhibit a significant voltage lag phenomenon during the initial discharge. When the discharge rate is increased to 10C or higher, stable discharge is achieved with a discharge voltage exceeding 1.5 V. This can be attributed to the favorable conductivity of FNC, which mitigates the influence of ohmic polarization. Overall, the high discharge voltage and small ohmic polarization associated with FNC enable an outstanding cathode material for alkali-metal/CF _x primary batteries. In comparison with other reported fluorinated nanocarbon materials, FNC represents the first instance of utilization in distinct alkali-metal primary batteries, demonstrating stable discharge even at high discharge rates.

The evolution of the electrode’s structural characteristics is likewise crucial. A comparative analysis of SEM imagery prior to and post discharge of the FNC-1.0 cathode before and after discharge at 0.01C in LPBs (Supplementary Figures S6A, B) reveals a large number of small particles uniformly covering the converted FNC surface. In the previously conducted studies, similar morphological characteristics were observed on the CF _x cathode after the discharge was completed, which is attributed to the contribution of small LiF crystals (Peng et al., 2019; Hu et al., 2023). In SPBs and PPBs, dense NaF and KF are also considered to generate in the FNC surface after discharge, and the morphology of the electrodes did not change much before and after discharge. Overall, the morphology of FNC is well maintained during discharge, in alignment with previous report (Cai et al., 2023).