Dual-phase engineering of CFx electrode and interphase enables low-temperature Li||CFx batteries Authors: Wei Wang, Wang Qiao, Zhuo Chen, Jialei Chen, Shan Chen, Zhuang Zhao, Xuelong Liao, Wenge Song, Rong Huang, Tiantian Lu, Youzeng Li, Haonan Liu, Jiacheng Sun, Xing Xu, Zhenhua Yan, Qing Zhao, Zhongfan Liu, Jun Chen, Huan Wang Published: 2026 ABSTRACT AbstractLow-temperature Li||CFx batteries undergo performance degradation due to the insulating nature of CFx and sluggish Li+ kinetics. Herein, we present a dual-modification strategy integrating porous positive electrode architecture design with customized electrode-electrolyte interphase engineering to overcome these limitations. Our approach features: (i) plasma-enhanced chemical vapor deposition-constructed mesoporous CFx with conformal carbon coating, and (ii) electrolyte-mediated in situ formation of a sulfur-rich inorganic interphase layer on CFx surface. The engineered porous framework establishes efficient Li+ diffusion channels, while the electrolyte-derived interphase significantly enhances Li+ transport and desolvation kinetics. Moreover, the carbon coating substantially improves electronic conductivity and suppresses LiF stacking. The optimized Li||CFx system delivers capacities of 428.4 mAh g−1 at −50 °C and 2 A g−1, and 333.4 mAh g−1 at −70 °C and 0.1 A g−1. Practical 5 Ah pouch cells deliver competitive specific energies of 470 and 332 Wh kg−1 at −50 and −70 °C under 0.01 A g−1, respectively.Subject terms: Batteries, Batteries Introduction Lithium||fluorinated carbon (Li||CFx) batteries have a theoretical specific energy of 2180 Wh kg−1 (x = 1), and low annual self-discharge (<0.5%), which makes them prime candidates for mission-critical applications ranging from Arctic infantry operations (−50 °C) to space exploration (−70 °C)1–6. However, the low-temperature operation failure stems from three bottlenecks (Fig. 1a), including: (i) exponential viscosity surge and ionic conductivity drop below electrolyte freezing point that impedes bulk Li⁺ mobility; (ii) spontaneous formation of discontinuous, unevenly isolated LiF-dominated positive electrode (cathode)-electrolyte interphase (CEI) that blocks Li+ transport pathways and burdens Li⁺ desolvation; (iii) intrinsically insulated nature of CFx matrix combined with random LiF precipitation impedes Li⁺ diffusion that further worsens with temperature decrease, thus leading to a spiking Warburg impedance (W0) and possibly becomes a rate-determining step in the reaction kinetics (Supplementary Fig. 1 and Supplementary Table 1)7,8. Challenges and resolutions in low-temperature Li||CFx batteries. Schematics for a fundamental challenges in low-temperature Li||CFx batteries and b a dual-phase optimization strategy synergistically combining PECVD treatment on bulk CFx and interphase engineering for low-temperature Li||CFx batteries. Electrolyte capacity relay denotes that the additional capacity at the end of CFx discharge contributed by the electron-mediated reduction reactions of the electrolyte. Thus far, the development of low-temperature Li||CFx batteries mainly relies on electrolyte optimization, including the introduction of low viscosity/low freezing point solvents or diluents9–12, the employment of weakly solvated solvents13–15, and facilitating inert C–F bond cleavage or insulating LiF dissolution through functional additives or solvents16–20. Although limited strategies have enabled the Li||CFx battery to work at subzero temperatures, Ah-level pouch cells have not been demonstrated at below −40 °C, which restricts their applications under cryogenic conditions such as aerospace payloads and hypersonic vehicle subsystems. These approaches essentially neglect two critical aspects of the interphase in the positive electrode interior. First, it has been well documented that the CEI component affects the performance of secondary batteries21–23, but has not been systematically explored in primary Li||CFx batteries, especially at low temperatures. It can be anticipated that Li+ kinetics behavior depends on the interfacial components, which direct LiF distribution and in turn affects Li+ transport pathway. On the other hand, in order to induce Li+ diffusion in the CFx interior to avoid LiF accumulation on the surface, it is necessary to engineer the positive electrode architecture to enable concurrent electron/Li+ conduction. Therefore, it is imperative to coordinate interphase regulation and multiscale structural design to overcome these fundamental limitations in Li||CFx batteries towards low-temperature operation. Besides, one special attribute of Li primary batteries different from secondary batteries is that the electrolyte reduction activity can be harvested in combination with CFx discharge to offer extra capacity24, which is expected to be realized via screening the lowest unoccupied molecular orbital (LUMO) energy level of solvents. Following the above guideline, we present a dual-modification strategy combining positive electrode surface engineering with electrolyte-derived interphase chemistry to overcome the kinetics limitations in low-temperature Li||CFx batteries (Fig. 1b). Specifically, a scalable plasma-enhanced chemical vapor deposition (PECVD) technique is developed to architect the CFx positive electrode with mesopores and conformal carbon coating, which promotes the diffusion of electron/Li+ to increase the CFx utilization. To enable low temperature operation, we develop a hybrid electrolyte composed of methyl acetate (MA) and dimethyl sulfite (DMS), which offers an ionic conductivity of 1.16 mS cm−1 at −80 °C through viscosity-dielectric optimization. The DMS component serves dual functions: preferential adsorption on CFx forms a sulfur-rich pre-CEI that dynamically evolves into a low-valent S-dominated interphase (Li2S/Sx2−), which exhibits strong affinity and low migration energy barrier for Li+, thereby accelerating Li+ desolvation and interfacial transport. Electrochemical decomposition at discharge termination provides a supplementary capacity through electron-mediated reduction reactions of electrolytes at low potentials, leading to a sequential dual-contribution mechanism. Consequently, the designed Li||PECVD-CFx batteries with MA/DMS operate stably in a wide temperature range from −100 to 100 °C. Even at −30 and −50 °C, high discharge capacities of 530.7 and 428.4 mAh g−1 are attained, respectively, at high rates of 5 and 2 A g−1. By using lean electrolyte (E/C ratio: 1.8 g Ah−1, the ratio of electrolyte weight to cell capacity) and low N/P ratio (1.49, the ratio of negative to positive areal capacity), we achieve 5 Ah pouch cells with cell-level specific energies of 470 and 332 Wh kg−1 at −50 and −70 °C under 0.01 A g−1, respectively. Results PECVD treatment on CFx and characterizations To break the electronic conductivity bottleneck of CFx and promote Li+ diffusion kinetics in the CFx interior at subzero temperatures, a radio frequency (RF)-PECVD technology was employed to modify the commercial CFx powder, which is denoted as PECVD-CFx (Fig. 2a). The primary factor for choosing the PECVD method is that it provides high-energy active radicals from hydrocarbon precursor25,26, which can penetrate into the CFx surface, creating holes and defects that facilitate the sp2-bonded carbon growth at low temperature. In this scenario, the holes provide channels for Li+ diffusion into the interior and thus effectively avoid the LiF accumulation and enhance the utilization efficiency of fluorine in the CFx interlayer, while the conductive carbon layers promote electron transport, collectively decreasing the voltage polarization and improving rate performance at low temperatures. Morphological and structural characterization of PECVD-treated CFx and control samples. a Schematic of CFx powder treated in a home-made PECVD system with a self-designed rotator and the photo of a hundred grams of black CFx powder after PECVD treatment. TEM images (left) and corresponding magnified views of selected regions marked by red boxes (right) for b T-CFx, c P-CFx, and d PECVD-CFx. White circles in (c, d) denote mesopores; The lattice fringes and spacing of 0.34 nm in the upper right of (d) correspond to the selected white box. e Content of semi-ionic (si) C–F bond and C=C bond based on the C 1s XPS results. f The BET surface area, pore size, and pore volume. g Comparison of room-temperature electronic conductivity among CFx-based samples. Acetylene (C2H2) serves as the precursor due to its higher decomposition ability and lower nucleation barrier, which are both beneficial for carbon layer growth26,27. X-ray diffraction (XRD) analyses reveal that the PECVD-CFx can preserve the characteristic peaks of intrinsic CFx when the growth temperature is below 600 °C (Supplementary Fig. 2), indicating the optimal PECVD treatment will not destroy the crystal structure of CFx. According to the discharge capacity and voltage of Li||CFx batteries in a standard electrolyte (1 M LiPF6 in EC/DEC (1:1, v/v), denoted as EC/DEC) (Supplementary Fig. 3), we determined that the optimal procedure is set as 15 min of PECVD treatment at 580 °C. Notably, a rotary stirring apparatus was designed and integrated into the PECVD system to enable continuous agitation of the CFx powder during the PECVD process, allowing scalable production of nearly hundreds of grams of the PECVD-treated uniform black product in a single-batch synthesis (Supplementary Fig. 4 and Supplementary Table 2). Moreover, comparative control experiments were conducted to isolate the advantages of PECVD technology by processing commercial CFx powder under gas-free thermal annealing as well as Ar plasma exposure at 580 °C, which are denoted as T-CFx and P-CFx. Transmission electron microscopy (TEM) characterizations were conducted to compare the morphology of pristine CFx, T-CFx, P-CFx, and PECVD-CFx, as shown in Fig. 2b–d and Supplementary Fig. 5. The untreated CFx displays a layered structure with a smooth surface, which becomes rough after thermal treatment, largely ascribed to the partial F removal28. Once exposed to the plasma atmosphere, abundant mesopores with sizes of ~20 nm are clearly observed. Compared to the Ar-treated P-CFx, the PECVD-CFx sample is found to be coated by a thin carbon layer with distinct lattice fringes (~0.34 nm), which affirms that the PECVD technology with C2H2 as precursor not only introduces mesoporous structures, but also in situ grows conformal nanocarbon layers, collectively facilitating concurrent electron/Li+ conduction. XRD patterns reveal three characteristic peaks at 2θ = 13.4°, 26.6°, and 41.3° in all CFx samples (Supplementary Fig. 6), suggesting that PECVD treatment preserves the pristine crystal structure of CFx. Based on X-ray photoelectron spectroscopy (XPS) analyses, all treated samples show an increase in the content of semi-ionic C–F bonds (at ~289.5 eV) alongside a decrease in the percentage of covalent C–F bonds and total F content (~10%; Fig. 2e, Supplementary Fig. 7 and Supplementary Table 3), manifesting the conversion of C–F bonds due to F cleavage, consistent with the Fourier transform infrared (FTIR) result (Supplementary Fig. 8). The PECVD-CFx sample also exhibits an increase in the percentage of C=C bond, corroborating the sp2-bonded carbon growth. Taken together, the inconspicuous difference in both the C–F bond configuration and content among T-CFx, P-CFx, and PECVD-CFx suggests that the thermal treatment is the primary driver for the defluorination and C–F bond transformation. The plasma introduction is responsible for creating mesopores, which significantly improve the specific surface area, pore size and pore volume of pristine CFx (Fig. 2f, Supplementary Fig. 9 and Supplementary Table 4). When carbon radicals are present in the plasma, sp2-bonded carbon layers in situ deposit on CFx during the etching process. This can be further supported by the Raman spectra across different CFx samples, where T-CFx and PECVD-CFx exhibit a similar intensity ratio of D band to G band (ID/IG), slightly higher than that of untreated CFx but lower than that of P-CFx (Supplementary Fig. 10). This sp2-hybridized carbon layer increases the electronic conductivity of intrinsically insulated CFx from 5 × 10−12 to 9 × 10−4 S m−1, achieving an eight orders of magnitude enhancement (Fig. 2g). Kinetics study of PECVD-CFx The direct current resistance (DCR)8 and galvanostatic intermittent titration technique (GITT)29,30 measurements show the Li||CFx batteries with PECVD-CFx maintain the lowest polarization resistance (Fig. 3a and Supplementary Fig. 11) and the highest Li+ diffusion coefficient (DLi+) (Fig. 3b and Supplementary Fig. 12) during the entire discharge process, particularly in the terminal period where bulk CFx undergoes complete conversion into LiF and carbon matrix. The lowest polarization and the highest DLi+ can be ascribed to the synergistic effect of carbon layer and mesopores that facilitate electron and Li+ transport into the CFx interior, thus avoiding LiF accumulation on the exterior. Linear sweep voltammetry (LSV) and Tafel plots (Fig. 3c and Supplementary Fig. 13) demonstrate that PECVD-CFx exhibits the highest reduction potential (2.78 V vs. Li+/Li) and largest exchange current density (3.1 × 10−4 mA cm−2 for the reduction process), indicating accelerated LiF formation kinetics3,31. PECVD treatment retains the high discharge capacity of Li||CFx batteries even with a small amount of F loss, evidencing that the mesopores can boost the F utilization through transporting electron/Li+ to the interior (Supplementary Fig. 14a). Kinetics performance of PECVD-CFx. a The DCR values at different DODs. b Calculated DLi+ based on the GITT measurements. c LSV curves to obtain the Tafel slopes. Discharge voltage profiles of d Li||CFx and e Li||PECVD-CFx batteries under high positive electrode loading. f Discharge voltage profiles of Li||PECVD-CFx batteries with positive electrode mass loading of ~20 mg cm−2 under lean E/C ratio from 0.92 to 2.75 g Ah−1. g Electrochemical evaluation of pouch cell assembled with PECVD-CFx positive electrode. h W0 derived from EIS spectra at different temperatures. i Comparison of DLi+ within CFx and PECVD-CFx materials obtained from EIS measurements at different temperatures. In addition, both the discharge voltage and capacity of the cell with PECVD-CFx positive electrode are the highest, particularly at 5 A g−1, further confirming the improved kinetics (Supplementary Fig. 14b). Consequently, the Li||PECVD-CFx cell delivers a specific capacity of 583.3 mAh g−1 at 5 A g−1 with a discharge plateau of 2.07 V, and still provides capacity at a high rate up to 20 A g−1, far exceeding the cell with untreated CFx (<200 mA g−1 at 5 A g−1; Supplementary Fig. 15). The improvement is significantly revealed when increasing the positive electrode mass to 15 mg cm−2, where the PECVD-CFx is capable of working at 4 A g−1. In sharp contrast, the commercial CFx fails to operate at 1 A g−1 (Fig. 3d, e). Further reducing both the electrolyte volume and Li foil thickness, the PECVD-CFx still offers a discharge capacity of 452.0 and 668.4 mAh g−1 at 0.5 A g−1 and under lean electrolyte (E/C ratio: 0.92 g Ah−1), respectively (Fig. 3f and Supplementary Fig. 16). The improved F utilization and reaction kinetics of PECVD-CFx also enable a pouch cell to deliver a specific capacity of 789.2 mAh g−1 under practical conditions (high mass loading: 19.3 mg cm−2, low N/P ratio:1.3 and lean E/C ratio: 2.75 g Ah−1) at 25 °C, as shown in Fig. 3g. The temperature-dependent electrochemical impedance spectroscopy (EIS) was performed for comparison31,32. The W0 of pristine CFx is sensitive to temperature change at subzero temperatures and increases to ~103 kΩ at −50 °C (Fig. 3h, Supplementary Fig. 17 and Supplementary Table 5). While for PECVD-CFx, the W0 shows an inconspicuous change as temperature decreases and stays at ~10 kΩ under −50 °C, which is two orders of magnitude lower than that of pristine CFx. Quantitative analysis reveals a higher DLi+ in PECVD-CFx compared to CFx (Fig. 3i and Supplementary Fig. 18)33, demonstrating that PECVD treatment can ameliorate the low-temperature Li+ transport within the bulk CFx. Benefiting from the high conductivity and rapid Li+ diffusion kinetics of PECVD-CFx, the assembled Li||CFx batteries demonstrate improved low-temperature behavior (Supplementary Fig. 19). Post-mortem analyses show that micrometer-sized cubic LiF grains accumulate on the surface of untreated CFx due to limited Li+ transport kinetics, whereas the discharge product of PECVD-CFx is mainly dominated by small LiF nanoparticles (Supplementary Fig. 20), which reduces blockage of Li⁺ transport channels. Consequently, the charge transfer resistance of PECVD-CFx remains relatively stable and is lower than that of pristine CFx (Supplementary Fig. 21), which is beneficial for improving the rate and low-temperature performance of Li||CFx batteries. Furthermore, the PECVD-CFx positive electrode after discharge exhibits a stronger LiF diffraction peak than the pristine CFx electrode (Supplementary Fig. 22), indicating more complete conversion of CFx to LiF due to the accelerated Li+ transport in the interior that enhances F utilization. Selection of low-temperature electrolytes for in situ formation of optimized CEI Built on the PECVD-CFx, we further resort to an electrolyte recipe for boosting the low-temperature performance. Beyond the regular requirements such as low freezing point, low viscosity and high ionic conductivity, the desirable electrolyte for low-temperature primary batteries should have the following characteristics: (i) high highest occupied molecular orbital (HOMO) energy level that enables tailored CEI formation on the CFx positive electrode at high potential, facilitating Li+ desolvation and transport across the interphase; (ii) low LUMO energy level that can obtain electrons at the voltage range lower than the discharge plateau of Li||CFx batteries, thus contributing to additional capacity. DMS was chosen as the primary solvent due to its low freezing point (−141 °C) and high dielectric constant (22.5) (Fig. 4a). Critically, it exhibits the strongest interaction with CFx and the highest HOMO energy level (Fig. 4b, c and Supplementary Data 1 and 2), rendering it susceptible to decompose at high open-circuit voltage (OCV) upon contact with CFx to form a homogenous, thin pre-CEI layer. Moreover, the lower LUMO energy level of DMS enables electrochemical activity at potentials below the CFx discharge plateau, thereby contributing additional capacity (Supplementary Fig. 23), which is consistent with our design concept. However, it is rather difficult for a single solvent to possess all essential requirements for low-temperature electrolytes (Supplementary Fig. 24). MA was further selected as a co-solvent to optimize electrolyte composition, leveraging its low viscosity (0.37 cP) and fast kinetics for charge transfer in Li||CFx batteries (Supplementary Fig. 25). 1 M LiBF4 serves as the salt because of its salt dissociation behavior and associated transport kinetics34. Screening electrolyte solvents and characterization of solvation structures. a Melting point, viscosity, and dielectric constant of different solvents. b Adsorption energy of CFx with different solvents. c LUMO and HOMO energy levels of different solvents. d DSC curves to obtain the freezing points. e The temperature-dependent ionic conductivity of different electrolytes. f RDF and CN of 1 M LiBF4-MA/DMS electrolyte. g 7Li NMR spectra of different electrolytes. h Interaction energy of Li+ with different solvents. i Schematic of DMS adsorption for evolving into CEI during Li||CFx discharge. In quest of the optimal electrolyte recipe, MA/DMS mixtures with the volume ratio of 7:3, 1:1, and 3:7 were prepared, all showing a freezing point below −130 °C (Fig. 4d), which enables Li||CFx batteries to work at low temperatures. When the volume ratio of MA and DMS reaches 1:1, the ionic conductivity is the highest in the temperature ranging from −80 to 25 °C, approaching 1.16 mS cm−1 at −80 °C (Fig. 4e). Meanwhile, it exhibits the lowest bulk viscosity at −40 °C and appropriate wettability (Supplementary Figs. 26 and 27). Thus, 1 M LiBF4 in MA:DMS with a volume ratio of 1:1 is formulated as the optimum electrolyte, which is denoted as MA/DMS in the following content. Molecular dynamics (MD) simulations in conjunction with corresponding radial distribution functions (RDF) and coordination numbers (CN) were performed to probe the Li+-solvation structure of the designed MA/DMS electrolyte. The numbers of Li+, BF4−, MA, and DMS in MA/DMS (Li+, PF6−, EC, and DEC in EC/DEC) are summarized in Supplementary Table 6. The resulting snapshots are depicted in Supplementary Fig. 28 (Supplementary Data 3). According to the RDF and related CN for MA/DMS, 4.88 of DMS, 0.04 of MA and 0.57 of BF4− enter the first solvation shell of Li+ with an average distance of 2.57, 3.45, and 2.17 Å, respectively (Fig. 4f). The rich DMS squeezing into Li+ solvation sheath can be experimentally supported by the 7Li nuclear magnetic resonance (NMR), where the DMS addition induces an upfield shift compared to unitary MA (Fig. 4g)35. Furthermore, with the addition of DMS, the stretching vibration of –OSO– in Raman spectra becomes stronger and undergoes blueshifts, accompanied by the conversion of coordinated MA to free MA, indicating that DMS occupies the position of MA and participates in the Li+ solvation shell (Supplementary Fig. 29)36. While for EC/DEC, the CN for EC, DEC and PF6− anion is 3.17, 0.81 and 0.24, respectively, with corresponding average distances to Li+ being 2.81, 2.75, and 2.59 Å (Supplementary Fig. 30). Apparently, DMS has a strong coordination ability and dominates the primary solvation sheath of Li+, as further evidenced by its high binding energy with Li+ (Fig. 4h and Supplementary Data 4). Combined with its strong adsorption energy and high HOMO energy level, it can be anticipated that the MA/DMS electrolyte facilitates the formation of a DMS-derived CEI interphase (Fig. 4i). CEI composition evolution during discharge To elucidate the effect of DMS on the CEI formation, XPS analyses were conducted on the CFx surface during different depths of discharging (DODs). The cutoff voltage is 1.5 V, and the discharge curve presents an abnormal capacity up to 1280 mAh g−1 with two discharge plateaus (Supplementary Fig. 31), which correspond to the formation of LiF and the decomposition of DMS. Moreover, the MA/DMS electrolyte enhances both the discharge capacity and voltage of the first plateaus, largely due to the formation of kinetically-favorable CEI on CFx. To this end, multiple characterizations were employed to elaborately analyze the interphase evolution. Prior to discharging, the S 2p XPS spectrum with MA/DMS electrolyte exhibits distinct signals related to high-valent S (–SOx, –CSO2) and low-valent S (Sx2−, S2−), where the –SOx dominates the component ascribed to the DMS decomposition (Fig. 5a)37,38. This can be further supported by the S–F peak in the F 1s spectrum, along with the observation of C–F and Li–F species originating from the CFx positive electrode and discharge product of LiF. Upon discharging, the contents of high-valent S and S–F fragments both show a decrease, whereas the low-valent S shows an inversely changing trend. This can be explained by the reduction of high-valent S into low-valent S (e.g., Li2S, Li2S2) during the initial discharging stage, which is typically recognized for its high ionic conductivity and lower interface impedance37–39. After discharging to 100 mAh g−1, the S-related species remain stable. When approaching the theoretical capacity of CFx (~800 mAh g−1), the low-valent S percentage emerges a declining trend, whereas the –CSO2 content among high-valent S ascends, which can be ascribed to the DMS decomposition (Fig. 5b, Supplementary Figs. 32 and 33). While for the commercial electrolyte, e.g., EC/DEC and 1 M LiBF4 in PC/DME (1:1, v/v; denoted as PC/DME), no pre-CEI was formed before discharging. Upon discharging, LiF crystals are observed on the CFx surface that blocks Li⁺ transport channels and reduces F utilization efficiency (Supplementary Figs. 34–37). Composition evolution of the CEI formed in the MA/DMS electrolyte. a S 2p and F 1s XPS spectra under different DODs at 0.1 A g−1. cps: counts per second. b Percentage change of high-valent S and low-valent S as discharging proceeds based on S 2p XPS result. c TOF-SIMS 3D depth images of low-valent S (S−) and high-valent S (SO3− and S2O3H−) under different DODs at 0.1 A g−1. d Depth-dependent variations in the content ratio of low-valent S to high-valent S analyzed by TOF-SIMS at different DODs. TEM images to observe the CEI formed in e MA/DMS and f EC/DEC. g Schematic diagram illustrating the CEI composition evolution during discharge in MA/DMS. Time-of-Flight secondary ion mass spectrometry (ToF-SIMS) was then carried out to track the CEI evolution process in the MA/DMS. Prior to discharging, the high-valent S species (SO3−, S2O3H−, LiSO2F−) dominate the surface (Fig. 5c and Supplementary Fig. 38a). As discharge proceeds, low-valent S fragments gradually take over and evolve into a uniform interlayer along with a decrease in high-valent S species. This transformation process was further validated through quantitative analysis of the normalized relative abundances of these components with Cs+ sputtering for 200 s (Fig. 5d and Supplementary Fig. 39). Low-valent S increases and becomes predominant at 100 mAh g−1, accompanied by a decrease in high-valent S, and maintains stable during the subsequent discharge process, consistent with XPS analyses. Meanwhile, the discharge product of LiF progressively increases and distributes among the bulk CFx with increasing DOD (Supplementary Fig. 38b), certifying that the kinetically-favorable interphase promotes Li+ transport that guides the LiF precipitation in the CFx interior. In addition, the interphase was compared through TEM imaging. As expected, the MA/DMS electrolyte enables a uniform film formation with a thickness of ~4 nm on the CFx surface before discharging and remains relatively stable in both component and thickness as DOD increases (Fig. 5e and Supplementary Fig. 40). As mentioned above, the film is rich in high-valent S, which evolves into low-valent S, with a slight expansion in the thickness. Conversely, the interphase formed in the EC/DEC electrolyte after discharge is thicker (>200 nm) with substantial and discontinuous LiF agglomerates, largely ascribed to the lack of a uniform and compact pre-CEI that guides controlled Li+ deposition (Fig. 5f and Supplementary Fig. 41). A similar phenomenon is observed in the PC/DME electrolyte, where LiF crystal accumulate on the CFx surface with thickness >100 nm (Supplementary Fig. 42). Based on the computational and experimental results, we propose the low-valent S-rich CEI formation process in the MA/DMS electrolyte, as schematically illustrated in Fig. 5g. Specifically, the DMS with a high affinity with CFx preferentially adsorbs on the positive electrode surface, forming a homogenous and thin DMS-rich liquid film, which decomposes into a high-valent S interlayer at OCV due to its high HOMO energy level. Upon discharging, these S species obtain electrons and transform into low-valent S interphase, which features high ionic conductivity for rapid Li+ transfer. Further combined with the PECVD-CFx positive electrode, the Li+ diffusion kinetics can be both accelerated from the interface to the interior of CFx, thus effectively avoiding the LiF accumulation on the surface, which is expected to boost the rate performance at low temperatures. Kinetics study of CEI formed on PECVD-CFx in MA/DMS electrolyte Li||PECVD-CFx batteries were assembled by coupling PECVD-treated CFx positive electrodes with the designed low-temperature MA/DMS electrolyte system to examine the compatibility and CEI kinetics. In situ EIS was conducted during the discharging process at different temperatures (Supplementary Fig. 43). At 25 °C, the EIS spectra show negligible change upon discharging. Once decreasing to subzero temperatures (−20 °C, −30 °C, and −50 °C), the overall impedances become smaller as compared to those before discharging. This can be ascribed to interphase evolution from high-valent S (HS-p-CEI) to low-valent S (LS-CEI) species, certifying the strong dependence of Li+ diffusion kinetics on CEI composition under harsh conditions. The CEI activation also leads to the appearance of two small semicircles in the mid-frequency region and high-frequency region, which correspond to the impedance associated with Li+-desolvation (Rct) and transport across the CEI (RCEI), respectively. The CEI remains stable before discharging to the end of the first plateau, further corroborating the stable composition and thickness of CEI formed on the PECVD-CFx surface. This can be more intuitively visualized by the corresponding distribution of relaxation times (DRT) mappings, where both Rct and RCEI after activation show inconspicuous change regardless of operating temperature (Fig. 6a and Supplementary Fig. 44). While for the second plateau, the DMS decomposition causes an increasing interfacial impedance (Supplementary Fig. 45). Conversely, the EIS spectra of PECVD-CFx in the EC/DEC electrolyte exhibit a single semicircle (Supplementary Fig. 46), indicating the absence of a distinct and functional CEI capable of guiding interfacial Li+ transport. Moreover, the impedance undergoes continuous variation, showing an initial decline before gradually increasing, as shown in the corresponding DRT mappings. The dynamic evolution is mainly caused by the accumulation of LiF crystals on the CFx surface during discharge, hindering Li+ transport and thereby restricting its low-temperature and high-rate performance. The role of CEI on Li||CFx batteries. a DRT mapping of EIS during discharge in MA/DMS electrolyte at −50 °C. Arrhenius curves to obtain the b Ea,ct and c Ea,CEI of HS-p-CEI and LS-CEI. d Adsorption energy of Li+ with representative interphase components such as Li2S, Li2SO3, and LiF. e Simulation of Li+ migration pathways from A to B in Li2S. f Migration energy barrier of Li+ across Li2S, Li2SO3 and LiF interphase. g Rct and W0 obtained from EIS spectra at different temperatures. h Comparison of self-discharge characteristics of EC/DEC and MA/DMS electrolyte in Li||PECVD-CFx systems. Temperature-dependent EIS measurements were further employed to quantitatively evaluate the ability of DMS-derived CEI in transporting Li+ at low temperatures. It is found that the HS-p-CEI lowers the Li⁺ desolvation energy barrier and enhances the low-temperature performance (Supplementary Fig. 47 and Supplementary Table 7). Subsequently, a discharging current of 0.1 A g−1 on the PECVD-CFx positive electrode was implemented to activate the HS-p-CEI into LS-CEI in MA/DMS. When the temperature stays above 0 °C, both the Rct and RCEI are insensitive to CEI evolution. But once decreasing below 0 °C, the LS-CEI conversion lowers the Rct and RCEI, and this disparity becomes more pronounced as the temperature decreases (Supplementary Fig. 48 and Supplementary Table 8). Further fitting Rct and RCEI separately, we obtain the corresponding Li+ desolvation energy (Ea,ct) and energy barrier associated with Li+ crossing the CEI (Ea,CEI), which are calculated to be 35.2 and 31.0 kJ mol−1 for LS-CEI (Fig. 6b, c), 24.1% and 30.5% lower than those with HS-p-CEI, respectively. This can be well explained by the density functional theory (DFT) calculation results that the Li2S exhibits the strongest Li⁺ affinity (−4.49 eV) and lowest migration energy barrier for Li+ (0.33 eV), thereby facilitating both Li⁺ desolvation at the interface and subsequent diffusion through the LS-CEI (Fig. 6d–f, Supplementary Fig. 49 and Supplementary Data 5 and 6). These findings underscore that the activated CEI rich in LS-inorganics determines the interfacial transport kinetics of Li+ and final low-temperature Li||CFx batteries40,41. This also substantiates that interphase engineering is critically important for the primary Li||CFx batteries, which hold equal significance to secondary LIBs. From the above temperature-dependent EIS spectra, we obtained the values of corresponding Rct and W0 for Li||CFx batteries with different electrolytes and positive electrodes. (Fig. 6g). It is found that the Li||PECVD-CFx with MA/DMS electrolyte exhibits the lowest Rct and W0, particularly under subzero temperatures, which are more than two orders of magnitude lower than those of conventional Li||CFx batteries. This evidences the synergistic effect of PECVD treatment and MA/DMS electrolyte in boosting Li+ diffusion kinetics from the electrode-electrolyte interface to the positive electrode interior in terms of low-temperature batteries. Additionally, the tailored viscosity and dielectric properties of the MA/DMS electrolyte lead to a substantial reduction in R0, supporting Li⁺ transport in the bulk electrolyte (Supplementary Fig. 50 and Supplementary Table 9). Subsequent quantitative fitting of Rct and W0 enables systematic analysis of Li+ diffusion kinetics at the electrode-electrolyte interface and within the CFx. Overall, the PECVD-CFx positive electrode paired with MA/DMS electrolyte exhibits the lowest Ea,ct (Supplementary Fig. 51) and the highest DLi+ (Supplementary Figs. 52 and 53). Quantitative self-discharge analysis reveals the storage stability of the MA/DMS electrolyte in Li||PECVD-CFx. Following a 30-day OCV after partial discharge (200 mAh g−1), MA/DMS enables nearly identical discharge profiles (ΔU = 8 mV), outperforming the commercial electrolyte of EC/DEC (ΔU = 94 mV) and PC/DME (ΔU = 17 mV) systems (Fig. 6h and Supplementary Fig. 54). This dual functionality of enhanced kinetics and storage stability supports the use of MA/DMS electrolyte in practical low-temperature CFx batteries. Low-temperature performance of Li||PECVD-CFx batteries in MA/DMS Figure 7a plots the discharge curves of Li||PECVD-CFx coin cells from 25 to −86 °C at different rates. At 0.1 A g−1, the room-temperature cell exhibits two distinct discharge plateaus. The initial discharge plateau yields a regular capacity of ~800 mAh g−1, which corresponds to the conventional CFx-to-LiF conversion, indicating that neither PECVD treatment nor CEI formation sacrifices the intrinsic capacity. A distinct plateau emerges at ~2.10 V with –CSO2 species dominant in the CEI (Supplementary Fig. 55), assigned to the DMS reductive decomposition. This electrolyte-mediated process in Li||CFx primary batteries contributes extra capacity of ~1000 mAh g−1. Overall, the Li||PECVD-CFx cells deliver capacities of 1863.6, 1523.8, 1369.0, and 1167.8 mAh g−1 at 0.1, 0.5, 1, and 5 A g−1, respectively. When the rate reaches 20 A g−1, 40 A g−1, 50 A g−1, the second plateau all disappear with capacities of 861.1, 675.0, 638.9 mAh g−1, which can be explained by the polarization at high rates that suppresses the DMS decomposition activity. A specific power of 86 kW kg−1 was achieved at 50 A g−1 based on the PECVD-CFx material (Supplementary Fig. 56), consistent with improved kinetics. At subzero temperatures, the cells maintain good discharge capacity and rate performance. For instance, at −30 °C, the capacity attains 1321.3 mAh g−1 at 0.05 A g−1 and 530.7 mAh g−1 at 5 A g−1 along with a discharge voltage of 1.92 V. At −50 °C, it remains a discharge capacity of 978.1 mAh g−1 at 0.05 A g−1 and 428.4 mAh g−1 at 2 A g−1 with an average discharge plateau at 1.86 V. As the temperature drops to −70 °C, it still performs well with a specific capacity of 802.3 mAh g−1 at 0.01 A g−1 and 333.4 mAh g−1 at 0.1 A g−1. Electrochemical performance of Li||PECVD-CFx batteries. a Discharge profiles at various rates in the temperature range from 25 to −86 °C. b Comparison of the rate performance of Li||CFx batteries at low temperatures achieved in this work with those reported studies9–19,42. The source of the literature data shown in this figure can be found in Supplementary Table 10. c Discharge profiles of Li||PECVD-CFx coin cells under practical conditions at −30 to −86 °C. d Discharge profiles of activated Li||PECVD-CFx battery at −86 °C. The periodic fluctuations of discharge curves at −86 °C originates from the fluctuations of the refrigerator. e Low-temperature performance of pouch cells under practical operating conditions. f Discharge profiles of the 5 Ah pouch cells at −50 °C and −70 °C. Inset: the photograph of the 5 Ah pouch cell. g Discharge profile of Li||PECVD-CFx coin cell over a wide temperature range of 100 to −100 °C. When the temperature is further reduced to −86 °C, the cell operates normally, providing capacities of 409.5 and 351.0 mAh g−1 at 0.005 and 0.01 A g−1, respectively, with discharge plateaus at 1.92 and 1.75 V. These results demonstrate improved performance compared with the unmodified CFx positive electrode or without LS-CEI formation in commercial electrolytes (Supplementary Figs. 57–62). The high-rate and low-temperature capability of our Li||PECVD-CFx cells is competitive with that of the reported Li||CFx cells through CFx positive electrode engineering or electrolyte optimization (Fig. 7b and Supplementary Table 10)9–19,42. To further investigate the key factors affecting the “electrolyte capacity relay” phenomenon, we evaluated the effects of discharge rate, working temperature, CFx loading, and E/C ratio on the low-potential decomposition behavior using both coin cells and pouch cells. Our results demonstrate that the contribution of this decomposition plateau becomes reduced or even disappears under high-rate, low-temperature, high-loading, or low E/C ratio conditions (Fig. 7a, Supplementary Figs. 63 and 64, Supplementary Note 1). Therefore, the additional capacity from electrolyte decomposition should be evaluated based on the actual operating conditions of the battery. By balancing the E/C ratio, working temperature, and discharge rate within a reasonable range, the specific energy of Li||CFx batteries can be further enhanced by harvesting “electrolyte capacity relay” (Supplementary Note 2). Inspired by the low-temperature and rate performance, we further evaluated the kinetics behavior of Li||PECVD-CFx batteries in MA/DMS with high mass loading of PECVD-CFx (14–19 mg cm−2), lean electrolyte (E/C ratio: 2.75 g Ah−1), and thin Li foil (100 μm, N/P ratio: 1.3–1.8). As shown in Fig. 7c, high specific capacities of 927.7, 874.1, and 619.5 mAh g−1 are achieved at −30, −50, and −70 °C at 10 mA g−1, respectively. Even at −86 °C, the cell still delivers 398.7 mAh g−1 of specific capacity at 5 mA g−1. By pre-discharging the Li||PECVD-CFx at 25 °C for transformation of HS-p-CEI to LS-CEI, the discharge capacities are further enhanced to 600.7, 523.4, and 443.2 mAh g−1 at 1, 2, and 5 mA g−1, respectively, underscoring the significance of kinetically-favorable CEI to low-temperature performance (Fig. 7d). The fluctuations originate from the sudden power outage, which also illustrates the temperature adaptivity of Li||PECVD-CFx cell. When the specific current is further increased to 1 A g−1 at −30 °C and 0.2 A g−1 at −50 °C, PECVD-CFx still achieves discharge capacities of 349.4 and 274.6 mAh g−1, respectively, demonstrating improved reaction kinetics (Supplementary Fig. 65). To evaluate practical scalability beyond coin cells, we assembled pouch cells (4 × 5.5 cm) using Li||PECVD-CFx in MA/DMS with high mass loading (17–20 mg cm−2), lean electrolyte (E/C ratio: 3.75 g Ah−1), and thin Li foil (100 μm, N/P ratio: 1.3–1.5). At a fixed rate of 10 mA g−1, the pouch cells deliver a high specific capacity of 911.0, 756.8, and 516.8 mAh g−1 at −30, −50, and −70 °C, respectively (Fig. 7e). Furthermore, a 5 Ah PECVD-CFx pouch cell (the specific capacity provided by PECVD-CFx is designed at 800 mAh g−1) was assembled with MA/DMS electrolyte under lean conditions (E/C ratio: 1.8 g Ah−1; Supplementary Fig. 66 and Supplementary Table 11). The pouch cell delivers capacities of 4.67 Ah (747 mAh g−1) and 3.41 Ah (546 mAh g−1) at −50 and −70 °C, which correspond to a CFx utilization rate up to 93.4 and 68.2%, respectively, indicating fast Li+ transport kinetics under these low-temperature conditions. Based on the total mass of the cell, the specific energy reaches 470 and 332 Wh kg−1 at −50 and −70 °C, respectively (Fig. 7f). This realizes Ah-level Li||CFx batteries operating below −40 °C. Benefiting from the high specific energy, the 5 Ah pouch cell was able to generate a current of 85.9 mA and continuously power a set of light-emitting diode arrays under −70 °C (Supplementary Fig. 67), further demonstrating the practical application potential of the customized Li||CFx battery under a cryogenic environment. Moreover, the MA/DMS electrolyte enables the Li||PECVD-CFx batteries to work across a wide temperature ranging from 100 to −100 °C (Fig. 7g and Supplementary Fig. 68). Li||CFx batteries exhibit self-heating behavior during discharge at high rates2,43, which may affect the actual operating temperature. To evaluate this self-heating effect, we monitored the surface temperature change of 5 Ah pouch cells and small-format single-layer pouch cells during discharge at different temperatures. The results show that all pouch cells in different electrolyte systems exhibit similar and inconspicuous temperature changes, which can be attributed to the relatively low discharge rate and low active material loading, generating insufficient heat to result in a temperature rise. Thus, the temperature of the environmental chamber is close to the working temperature of Li||CFx cells, which ensures reliable performance comparison (Supplementary Figs. 69 and 70). To further improve electrolyte safety and long-term storage stability for practical applications, we incorporated ethoxy(pentafluoro)cyclotriphosphazene (PFPN) and LiFSI into the MA/DMS base electrolyte, denoted as MA/DMS–PFPN/LiFSI. This formula improves high-temperature safety and storage stability while maintaining the room- and low-temperature performance of Li||CFx cells, thereby meeting the requirements of practical batteries (Supplementary Figs. 71–74 and Supplementary Note 3). Discussion We present a synergistic positive electrode/electrolyte engineering strategy that fundamentally addresses the kinetics limitations of low-temperature Li||CFx batteries. By developing a rotor-assisted PECVD technique, we synthesized bulk CFx featuring a porous architecture with thin carbon encapsulation layers, which simultaneously enhances Li+ diffusion coefficients and F utilization efficiency. Complementing this advancement, a cryogenic electrolyte based on MA/DMS is formulated and generates a DMS-dominated solvation structure, triggering preferential adsorption and rapid decomposition of DMS on the CFx surface, which thus directs the formation of a low-valent S-rich CEI on CFx. This dual-optimization strategy reduces the W0, Rct, and RCEI values associated with Li+ transport within the CFx interior and across the interface. In addition, DMS with a low LUMO energy level contributes additional capacity through sulfur reduction chemistry at low potentials following CFx discharge. Consequently, the integrated system delivers competitive specific capacities across a wide temperature span: 1863.6 (25 °C), 1321.3 (−30 °C), 978.1 (−50 °C), 802.3 (−70 °C), and 409.5 mAh g−1 (−86 °C). Practical 5 Ah pouch cells were assembled and maintain 747 mAh g−1 (4.67 Ah) at −50 °C and 546 mAh g−1 (3.41 Ah) at −70 °C, corresponding to cell-level specific energies of 470 and 332 Wh kg−1, respectively, achieving Ah-level Li||CFx pouch-cell operation at temperatures below −40 °C. Methods Materials Graphite fluoride (C/F molar ratio: 1/1, battery grade, ≥99%) was purchased from Shandong Zhongshan Photoelectric Materials Co., Ltd. Lithium tetrafluoroborate (LiBF4, 99.9%), lithium bis(fluorosulfonyl)imide (LiFSI, 99.9%) and commercial electrolyte (1 M LiPF6 in EC/DEC (1:1, v/v), denoted as EC/DEC; 1 M LiBF4 in PC/DME (1:1, v/v), denoted as PC/DME) were purchased from Dodo Chem Co., Ltd. Lithium chips (Li, 99.99%, 14 mm diameter, 500 μm thickness) and thin lithium foil (Li, 99.99%, 60 mm width with 100 μm and 85 μm thickness) were purchased from China Energy Lithium Co., Ltd., stored in an argon-filled glove box (H2O < 0.01 ppm, O2 < 0.01 ppm, 25 °C ± 2 °C), and used without any surface pre-treatment to remove the native oxide layer. Dimethyl sulfite (DMS, 99%), methyl acetate (MA, 99%), and N-methyl-2-pyrrolidone (NMP, 99.5%, water <50 ppm) were purchased from Shanghai Macklin Biochemical Co., Ltd. Ethoxy(pentafluoro)cyclotriphosphazene (PFPN, ≥98%) was purchased from Aladdin. Polyacrylonitrile copolymer (LA133, 15 ± 0.2% solid content), CR2032-type coin cell cases (20 mm diameter, 3.2 mm thickness), coin cell spring plates (15.4 mm diameter, 1.1 mm thickness) and SZQ-type four-sided film applicator were purchased from Guangdong Canrd New Energy Technology Co., Ltd. Super P carbon black (SUPER P, 99.99%), polyvinyldifluoride (PVDF, 99.99%, Mw ~ 700,000) and carbon-coated aluminum (Al) foil (>99.9%, 16 μm thickness, 180 mm width, carbon layer: 1 µm thickness) were purchased from Shenzhen Kejing Electronic Co., Ltd. All chemicals were used as received without further purification. Preparation of plasma-enhanced chemical vapor deposition (PECVD) treated CFx (PECVD-CFx) materials The synthesis of PECVD-CFx was conducted in a customized PECVD system equipped with a 6-inch quartz tube. Specifically, the commercially available CFx powder was placed in the center of the tube furnace. The system was evacuated to a base pressure of 10 Pa and cleaned with Ar gas to remove air. The radio frequency (RF) power is 150 W. 8 sccm acetylene (C2H2) serves as a precursor. The treatment temperature (500–600 °C) and time (10–25 min) play an important role in the capacity of as-received PECVD-CFx materials. In this work, the optimal PECVD treatment temperature and time is 15 min at 580 °C. For the large-scale growth, a self-designed rotator was introduced in the PECVD system, as schematically shown in Fig. 2a. Preparation of other treated CFx control samples For the synthesis of thermal-treated CFx (T-CFx), no gas or plasma was introduced, and the CFx was annealed at 580 °C under a base pressure (10 Pa) for 15 min. While for the Ar plasma-treated CFx (P-CFx), 8 sccm Ar gas was introduced and the CFx was annealed at 580 °C with RF power (150 W) for 15 min. Electrolyte preparation A series of MA/DMS electrolytes were prepared by dissolving 1 M LiBF4 in MA:DMS mixed solvents with varying volume ratios (ranging from 0 to 100 vol%). For the MA/DMS–PFPN electrolyte, 1 M LiBF4 was dissolved in MA:DMS mixed solvents at a volume ratio of 1:1, followed by adding different contents of PFPN. For the MA/DMS–PFPN/LiFSI electrolyte, 0.25 M LiFSI was further added to the aforementioned solution. All procedures were performed in an inert atmosphere in an argon-filled glove box (H2O < 0.01 ppm, O2 < 0.01 ppm, 25 °C ± 2 °C). Electrode preparation The working electrodes were prepared by mixing the CFx under different treatment conditions with Super P and PVDF in a weight ratio of 8:1:1 (for regular testing) or 9:0.5:0.5 (for high-loading testing). The mixture was then dispersed in NMP to form a slurry. The slurry was then coated onto a carbon-coated Al foil, followed by an overnight drying at 80 °C. The final electrodes were obtained by punching the foil into 10 mm diameter disks. For regular testing, the CFx loading was ~1–1.5 mg cm−2, while for the high-loading test, the CFx loading was ~14–20 mg cm−2. The preparation procedures of the CFx electrode for pouch cells were similar to the above description, except that the electrode size is 4 × 5.5 cm. For the XPS testing of the CFx electrode during the discharge process, to avoid the influence of the F element in PVDF and the C element in Super P, the CFx electrode was assembled using an F-free LA133 binder, and without adding Super P. Specifically, the CFx electrodes were prepared by mixing 98 wt.% CFx, 2 wt.% LA133 with deionized water to form a slurry, which was then coated onto Al foil and dried overnight at 100 °C. Finally, the electrodes were cut into discs with a diameter of 10 mm, with a CFx loading of ~2–3 mg cm−2. Characterizations Differential Scanning Calorimetry (DSC, METTLER TOLEDO DSC3) was employed to measure the freezing point of the mixed electrolyte with different compositions over a temperature range from −150 to 25 °C. The measurement employed a liquid nitrogen cooling system with a heating rate of 5 °C min−1. Transmission Electron Microscopy (TEM) characterization was conducted using Talos F200X G2. Scanning electron microscopy (SEM) images were acquired with an FEI Microscope (JSM-7900F). Raman spectroscopy was performed using a LabRAM HR Evolution system with an excitation wavelength of 633 nm. X-ray diffraction (XRD) patterns were collected on a Smart Lab 9 KW diffractometer equipped with a Cu-target X-ray tube (λ = 0.154 nm) operating at 150 mA and 40 kV. The 7Li NMR spectrum was collected in the NMR spectrometer (AVANCE III 400 MHz), with 1 M LiCl in D2O as the standard solution. It was injected and sealed into the capillary tube, which was further put into the NMR tube for calibration. X-ray photoelectron spectroscopy (XPS) was performed on an Axis Ultra DLD spectrometer using monochromatic Al Kα radiation (hν = 1486.7 eV) operated at a working voltage of 12 kV and a filament current of 6 mA. The prominent C 1 s peak was calibrated to 284.8 eV. Data collection and processing, including background subtraction and peak fitting, were performed using Advantage software. Time-of-Flight secondary ion mass spectrometry (ToF-SIMS) was carried out using TOF-SIMS 5 iontof, PHI NanoTOFII. A 30 keV Bi3+ beam was used for analysis over an area of 50 μm × 50 μm. Sputtering was conducted with a 1 kV Cs+ beam at 60 nA over an area of 200 μm × 200 μm, corresponding to a sputter rate of ~0.84 nm s−1 on GaN and a sputtering time of 900 s. Electrical conductivity was measured by a four-point probe method using a ST2255 resistivity tester (Suzhou Lattice Electronics Co., Ltd) under 25 MPa pressure at 25 °C. Electrochemical measurements Ionic conductivity of the electrolytes at various temperatures was determined using electrochemical impedance spectroscopy (EIS) on symmetric cells comprising two parallel Pt-plate electrodes (10 mm × 10 mm). The distance between the two electrodes was 10 mm, and the applied frequency range was from 105 to 10−2 Hz with an AC amplitude of 10 mV, using 5 points per decade. The measurements were performed near the open-circuit potential. Prior to data acquisition, the cell was allowed to rest for 10 min to ensure a quasi-equilibrium state. EIS data were analyzed with Zview software by importing raw.txt files and fitting them with an appropriate equivalent circuit. Fitted values of the circuit elements were then extracted, and the fitting quality was assessed from the fitting errors. The ionic conductivity of the electrolytes was calculated using the following equation: Where σ (S cm−1) is the ionic conductivity of the electrolyte, L (cm) is the distance between two Pt-plate electrodes. Rs (Ω) is the electrolyte resistance. A (cm2) represents the area of the Pt-plate electrode. LSV test was carried out on an electrochemical workstation (Ivium-n-Stat, Nederlanden) in a three-electrode configuration, with CFx as the working electrode (mass loading: ~1.0–1.5 mg cm−2) and Li foil as the counter electrode and reference electrode. EIS measurements were conducted on an electrochemical workstation (Ivium-n-Stat, Netherlands) at a frequency range of 105–10−2 Hz with a perturbation amplitude of 10 mV. GITT measurements were performed by discharging at 0.1 A g−1 for 10 min with data recorded every 10 s, followed by resting for an hour; this sequence was repeated until a cutoff voltage of 1.5 V was reached. The mass loading of the CFx electrode was ~1–1.5 mg cm−2. The Li+ coefficient diffusion (DLi+) was subsequently calculated using the following simplified formula based on Fick’s second law: Where I0 (A) represents the current; Vm (cm3 mol−1) and L (cm) are the molar volume and thickness of the electrode material, respectively; S (cm2) is the contact area between the electrolyte and the electrode; F (C mol−1) is the Faraday constant; and Zi is the charge transfer number during the discharge process of the Li||CFx battery. The DCR was measured using the following protocol: (i) Discharge at 0.2 A g−1 for 30 min, followed by a 30-min rest period; (ii) Discharge at 1 A g−1 for 3 min, followed by a 30-min rest period; (iii) Repeat the above two steps for 5 cycles, then discharge at 0.2 A g−1 to 1.5 V. The mass loading of the CFx electrode was ~1–1.5 mg cm−2. Calculating the DCR value is according to the following equation: Where E1 (V) is the instantaneous voltage after discharge completion at 0.2 A g−1; E2 (V) is the instantaneous voltage before discharge at 1 A g−1; I1 (A) is the applied current at 0.2 A g−1; I2 (A) is the applied current at 1 A g−1. The discharging tests of Li||CFx batteries (2032-type coin cells and pouch cells) were performed using a standard battery tester (LAND-CT2001A, Wuhan Land Electronics Co. Ltd) with a cut-off voltage of 1.5 V. CFx, Li metal, and polyethylene separator (PE, SK-AsahiKASEI, 20 μm thickness, 16 mm diameter, 41% porosity, and an average pore size of ~40 nm) were used as the positive electrode, negative electrode, and separator, respectively. For regular testing (CFx loading: ~1.0–1.5 mg cm−2), 30 μL electrolyte and 500 μm Li chip (diameter: 14 mm) were utilized to assemble cells. For high-loading coin cells and pouch cells (CFx loading: ~14–20 mg cm−2), 100 μm Li foil (N/P ratio: 1.3 ~ 1.8) and lean electrolyte (E/C ratio: 0.92–3.75 g Ah−1) were adopted. All specific discharge capacities are calculated based on the mass of the CFx. All assembled Li||CFx batteries were rested at 25 °C for 10 h to ensure sufficient electrolyte infiltration, followed by a 2 h equilibration at different test temperatures before measurements. The electrochemical performance at different temperatures was tested in a temperature-controlled test chamber (−70 to 100 °C) and a low-temperature freezer (−86 to −10 °C, with a fluctuation of ±1.5 °C). For short-duration discharge tests within low temperature (−100 to −80 °C), a customized testing device controlled by liquid nitrogen was employed. The test temperature refers to the set temperature of the environmental chamber, which operates in forced convection mode to ensure temperature uniformity and heat exchange efficiency. Assembly of 5 Ah pouch cells PECVD-CFx was utilized as the active material in the positive electrode, which features a double-sided coating, a mass loading of 15.58 mg cm−2, and dimensions of 64.7 × 6.2 cm. After calendering under 80 tons, the electrode retained a uniform thickness of 145 ± 2 μm. The negative electrode was an 85 μm Li foil with a size of 70.5 × 6.0 cm (N/P ratio: 1.49). The PE separator was employed as the separator, and 9 g of electrolyte was used (E/C ratio: 1.8 g Ah−1). Finally, the cell was assembled by winding the positive electrode, separator, and negative electrode together, followed by encapsulation in an Al-plastic film. A stack pressure of 350 kPa was applied to the pouch cells during discharge. Computational details Molecular dynamics simulations MD simulations were performed in Materials Studio to examine molecular interactions in the electrolyte systems. Amorphous simulation cells were constructed using the Amorphous Cell module by Monte Carlo packing under three-dimensional periodic boundary conditions. The initial configurations were relaxed and equilibrated with the Forcite module. A 500 ps NPT simulation at 298 K and 0.1 MPa was first performed to obtain equilibrated cell dimensions and density, followed by a 10,000 ps NVT simulation using a Nosé thermostat with a Q ratio of 0.01. Interatomic interactions were described using the COMPASS III force field44,45. Atomic partial charges were assigned according to the force-field parameters. Van der Waals interactions were evaluated using the atom-based summation method, and electrostatic interactions were treated using the particle-particle particle-mesh method. The integration time step was 1 fs, and trajectory frames were recorded every 5000 steps. Quantum chemistry calculations Frontier molecular orbital calculations were carried out using the DMol3 module in Materials Studio46,47. Geometry optimizations were performed with the BLYP exchange-correlation functional within the generalized gradient approximation (GGA) framework48,49. The double numerical plus polarization (DNP) basis set with a cutoff radius of 4.4 Å was applied, and core electrons were treated using the all-electron approach. The convergence criteria for energy, maximum force, and displacement were set to 10−5 hartree, 0.004 hartree/Å, and 0.005 Å, respectively. The self-consistent field convergence threshold was 10−6 hartree, with thermal smearing of 0.005 hartree applied to improve convergence. Li+ migration barriers across interfacial components were calculated using DFT as implemented in the Vienna Ab initio Simulation Package50,51. Electron-ion interactions were described using the projector-augmented-wave method, and exchange-correlation interactions were treated with the Perdew–Burke–Ernzerhof functional within the GGA52. The plane-wave energy cutoff was set to 520 eV. Brillouin-zone sampling was performed with a k-point spacing of 0.4 Å−1 for all periodic structures. Structural relaxations were carried out using the conjugate-gradient algorithm until the total energy and force converged to 10−5 eV and 0.02 eV/Å, respectively. A vacuum layer larger than 15 Å was introduced to minimize interactions between periodic images. Li+ migration pathways and energy barriers were determined using the climbing-image nudged elastic band method. Supplementary information Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Supplementary Data 3 Supplementary Data 4 Supplementary Data 5 Supplementary Data 6 Transparent Peer Review file Source data Source data Acknowledgements This work was supported by the Natural Science Foundation of China (22121005, J.C.), Natural Science Foundation of Tianjin (25ZXZSSS00060, H.W. and No. 24JCZDJC00540, H.W.), Haihe Laboratory of Sustainable Chemical Transformations (25HHWCSS00024, H.W.), and Fundamental Research Funds for the Central Universities (63263137, H.W.). Author contributions H.W. and W.W. conceived the original idea and designed all the experiments. W.W. and W.Q. built the PECVD instrument and modified CFx materials. W.W. conducted the measurements of Li||CFx batteries and analyzed the data with the help of Z.C., J.L.C., S.C., Z.Z., X.L., W.S., R.H., T.L., Y.L., H.L., J.S., X.X., H.W., Z.Y., and Q.Z. Z.Z. and J.S. helped to perform the theoretical calculations and simulations. W.W. and H.W. wrote the manuscript. H.W., J.C., and Z.L. co-supervised the project. All authors were involved in the discussion of the experimental results and the preparation of the manuscript. W.W. and W.Q. contributed equally to this work. Peer review Peer review information Nature Communications thanks Kan Cao, who co-reviewed with Huijing Xue, Susana Gonzalo, and the other, anonymous, reviewer(s) for their contribution to the peer review of this work. A peer review file is available. Data availability All data that support the findings of this study are present in the paper and the Supplementary Information. Additional data related to the study are available from the corresponding author upon reasonable request. Source data are provided with this paper. Competing interests The authors have patent filings to disclose. H.W. and W.W. are the inventors on the CN patent application (Z.L. 2024 1 1398561.1) regarding the preparation of positive electrode materials and electrolytes for wide-temperature range Li||CFx batteries described in this article. The remaining authors declare no competing interests. Footnotes Publisher’s note Springer Nature remains neutral with regard to jurisdictional claims in published maps and institutional affiliations. Contributor Information Zhongfan Liu, Email: zfliu@pku.edu.cn. Jun Chen, Email: chenabc@nankai.edu.cn. Huan Wang, Email: huan.wang0520@nankai.edu.cn. Supplementary information The online version contains supplementary material available at https://doi.org/10.1038/s41467-026-76513-4. References Associated Data Supplementary Materials Supplementary Information Description of Additional Supplementary Files Supplementary Data 1 Supplementary Data 2 Supplementary Data 3 Supplementary Data 4 Supplementary Data 5 Supplementary Data 6 Transparent Peer Review file Source data Data Availability Statement All data that support the findings of this study are present in the paper and the Supplementary Information. Additional data related to the study are available from the corresponding author upon reasonable request. Source data are provided with this paper.