A fire-safe Li metal battery via smart gas management

Significance

Flammable gases generated during thermal runaway critically impact the safety of high-energy-density Li metal batteries. Here, we demonstrate that incorporating flame-retardant interfaces (FRIs) into the cathode enables a smart response to thermal abuse, significantly suppressing gas generation and altering gas composition. FRIs reduce oxygen release at the cathode and alter the gas composition at the full-cell level, which helps lower flammability and prevents mechanical rupture. Additionally, the self-heating rate of the improved cells is reduced by 4 orders of magnitude, achieving zero thermal runaway in pouch cells. This smart gas management strategy enhances both thermal safety and electrochemical stability, offering a transformative pathway to fire-safe Li metal batteries for advanced energy storage applications.

Abstract

Lithium (Li) metal batteries offer high energy density but face significant safety challenges due to gas evolution under thermal abuse conditions. At the anode, the reduction of organic carbonate-based electrolytes generates flammable gases (e.g., H₂, CH₄), while the poor thermal stability of the cathode results in the release of O₂, CO, and CO₂. The accumulation of these gases contributes to mechanical rupture, and their migration further exacerbates thermal runaway. To address these challenges, we propose a smart gas management strategy that constructs continuous flame-retardant interfaces (FRIs) by incorporating flame-retardant polymers (FRPs) into the cathode. Smart gas management is defined as the ability to suppress gas production, alter gas composition to reduce flammability, and mitigate internal pressure buildup, thereby preventing thermal runaway. The FRIs significantly enhance the thermal stability of the cathode by suppressing oxygen release and minimizing electrolyte oxidation caused by active oxygen species. Additionally, the FRP releases flame-retardant radicals that diffuse into the electrolyte, interrupting reactions that generate flammable gases at the anode. This dual-action mechanism reduces gas production and mitigates the risks associated with thermal runaway, forming the foundation of a smart gas management strategy. With this strategy, we demonstrate zero thermal runaway in a 0.58-Ah Li||NCM811 pouch cell with 100% state of charge under thermal abuse conditions. This approach is highly compatible with current manufacturing processes, offering a scalable solution for improving the safety of high-energy-density Li metal batteries. This work provides a promising pathway toward fire-safe Li metal batteries for electric vehicles and other energy storage applications.

The development of lithium (Li) metal batteries is essential to meet the growing demand for next-generation high-energy-density energy storage systems, particularly in electric vehicles and sustainable energy applications (1, 2). With the capability to achieve energy densities exceeding 400 Wh kg⁻¹, Li metal batteries utilize lightweight Li metal anodes paired with high-voltage, nickel (Ni)-rich oxide cathodes such as LiNi_0.8Mn_0.1Co_0.1O₂ (NCM811) (3). However, this configuration presents substantial safety challenges, especially thermal runaway, where the generation and migration of reactive gases critically undermine battery safety (4, 5). These challenges originate from both the anode and cathode (Fig. 1A). At the anode, the reduction of organic carbonate-based electrolytes generates flammable gases, such as H₂, CH₄, C₂H₄, C₂H₆, which significantly increase the risk of fires and explosions (6). The high reactivity of Li further exacerbates these risks (7, 8). At the cathode, the oxidation of organic electrolytes produces CO and CO₂, while the inherently poor thermal stability of Ni-rich cathodes leads to oxygen release at high states of charge (9, 10). The migration and interaction of these reactive gases intensify thermal runaway through exothermic reactions, such as between C₂H₄ and O₂, which release substantial heat (11, 12). Moreover, gas migration to the electrode surfaces further aggravates the situation by triggering highly exothermic reactions (13). For instance, the reaction between O₂ and Li exhibits a high reaction enthalpy of –1197 kJ mol^–1 (14, 15). Additionally, the accumulation of these gases causes cell swelling, mechanical rupture, and heightened risks of catastrophic failure (16–18).

Fig. 1.

Schematic illustration of the working mechanism of the smart gas management Strategy. (A) The gas-related safety challenges in Li metal batteries: at the anode, the reduction of organic electrolytes generates flammable gases (e.g., CH₄), while at the cathode, oxygen release and gas migration, exacerbated by poor thermal stability of Ni-rich cathodes, intensify thermal runaway. (B) The strategy incorporates a phosphonate-based FRP into the cathode, forming FRIs. FRIs enhance thermal stability by suppressing oxygen release and electrolyte oxidation. FRIs can smartly respond to thermal abuse by undergoing thermal decomposition to release flame-retardant radicals, which interrupt the flammable gas generation at the anode. These effects collectively reduce gas production, mitigate pressure buildup and combustion reaction, and prevent thermal runaway.

Existing strategies to address these challenges include cathode structure stabilization, electrolyte chemistry optimization, and the development of solid-state electrolytes (SSEs) (19–22). For instance, surface treatments, grain boundary engineering, and lattice substitution have demonstrated effectiveness in reducing oxygen release by stabilizing cathode materials (23–25). Flame-retardant electrolytes have shown promise in improving thermal stability by capturing thermally decomposed free radicals (26, 27). However, current studies primarily focus on specific aspects, such as oxygen behavior at room temperature or the flammability of individual electrolyte components, without considering the combined effects of gas behavior and thermal reactions under abuse conditions (28). Moreover, while SSEs offer inherent thermal advantages, issues such as interfacial reactions with Li and manufacturing scalability limit their immediate applicability (29, 30). Thus, innovative strategies that enhance the safety of liquid electrolyte-based Li metal batteries remain critically needed.

In this work, we propose an innovative smart gas management strategy to address the dual challenges of oxygen release from the cathode and flammable gas generation at the anode under thermal abuse conditions (Fig. 1B). Building upon our previous success with FRP-coated separators that demonstrated 100% (9 h) delayed thermal runaway propagation in graphite||NCM811 pouch cells (31), we now extend this flame-retardant chemistry to cathode engineering. By incorporating phosphonate-based FRP into the cathode to construct continuous flame-retardant interfaces (FRIs), we achieve smart gas management throughout the Li||NCM811 pouch cell. This strategy involves suppressing gas production, altering gas composition to reduce flammability, and mitigating internal pressure buildup, all of which work together to prevent thermal runaway. The FRIs demonstrate two critical merits. First, they significantly improve the thermal stability of the cathode by suppressing oxygen release and minimizing the chemical oxidation of the electrolyte caused by the cathode or active O₂. Second, the FRP spontaneously releases flame-retardant radicals that diffuse into the electrolyte, interrupting reactions involving thermally decomposed active radicals responsible for flammable gas generation at the anode. These dual effects reduce gas production, mitigate the risks associated with thermal runaway, and enhance overall cell safety. Notably, this approach is highly compatible with existing manufacturing processes, making it a scalable and practical solution for advancing fire-safe, high-energy-density electrochemical energy storage systems.

Results

Improved Thermal Safety of NCM811 Cathodes with FRIs.

The thermal stability of nickel-rich layered oxide cathodes, such as NCM811, is a critical factor influencing the safety of Li metal batteries. To enhance the thermal safety of NCM811, an FRP was incorporated during electrode fabrication to construct FRIs. This process involved adding a phosphonate-based FRP precursor to the NCM811 slurry, followed by ultraviolet (UV) curing for in situ polymerization, and subsequent drying. The final electrode composition contained 4.5 wt% FRP relative to the total mass of the cathode (excluding the Al current collector), as confirmed by thermogravimetric analysis (TGA, SI Appendix, Fig. S1). Due to the low FRP content (4.5 wt%), its characteristic diffraction peaks were below the detection limit of X-ray diffraction (XRD), making direct identification challenging. To verify the presence of FRP, a control sample with higher FRP loading (~9 wt%) was prepared (EFRP@NCM811). In this sample, a distinct diffraction peak at 26° corresponding to the FRP phase was observed, consistent with the pure FRP reference (SI Appendix, Fig. S2). Additionally, the phosphonate moiety in FRP was confirmed by Raman spectroscopy: The DEAP monomer exhibited a prominent Raman peak at 808 cm⁻¹, attributed to the P–O–C stretching vibration. Importantly, this signature peak remained detectable in the pristine FRI@NCM811 electrode, albeit with reduced intensity (SI Appendix, Fig. S3). The depth distribution of the FRIs was characterized using time-of-flight secondary ion mass spectrometry (TOF-SIMS). Chemical maps (Fig. 2 A and B and Movie S1) revealed that the active cathode particle surfaces (represented by NiO₂⁻, CoO₂⁻, and MnO₂⁻) and the conductive carbon/binder domains (C_x⁻) were uniformly coated with the FRIs, identified by PO₂⁻ and PO₃⁻ signals originating from the FRP. Notably, after 65 s of Cs⁺ sputtering, signals corresponding to NCM811 particles began to emerge. Furthermore, PO₂⁻ and PO₃⁻ signals persisted even after NiO₂⁻, CoO₂⁻, and MnO₂⁻ signals stabilized, indicating that the FRP infiltrated microcracks within secondary particles (Fig. 2C). As shown in SI Appendix, Fig. S4, the results of the scanning electron microscope (SEM) and corresponding energy dispersive spectroscopies mapping (EDS-Mapping) demonstrates that FRP is evenly distributed within FRI@NCM811 electrode. This dual functionality of FRIs was corroborated by SEM imaging (SI Appendix, Fig. S5), which highlighted their role in forming a protective surface layer and penetrating particle interiors to reinforce structural stability.

Fig. 2.

Distribution of FRIs in cathodes and thermal stability of FRI@NCM811. (A and B) Chemical maps illustrating the interphase composition on the surface of NCM811 secondary particles. Comparative mapping of PO₂⁻ and PO₃⁻ ions (green) derived from the FRP, and NiO₂⁻, CoO₂⁻, and MnO₂⁻ fragments (red), and C_x⁻ fragments (blue) derived from carbon conductors and binders, during 0 to 65 s (A) and 65 to 1,300 s (B) of Cs⁺ sputtering. (C) TOF-SIMS depth profiles of key secondary ion fragments collected from the FRI@NCM811 electrode, showing the depth distribution of the FRI components. (D) TGA curves of the phosphonate-based polymer, pristine NCM811, and FRI@NCM811 electrodes. (E) DSC curves of fully charged NCM811 and FRI@NCM811 in the presence of 1 M LiPF₆ in an EC/DEC/DMC electrolyte (1:1:1 by volume), highlighting the heat suppression provided by the FRI. (F) EGA-MS results showing the reduction in oxygen (O₂) release from fully charged NCM811 with the incorporation of FRI, indicating improved lattice oxygen stability and reduced thermal degradation.

Based on the TGA results (Fig. 2D), the FRP begins to decompose around 100 °C, releasing [PO]· radicals that effectively quench the H· produced during electrolyte thermal decomposition. To evaluate the thermal safety improvements provided by FRIs, differential scanning calorimetry (DSC) was conducted to assess the heat release of fully charged NCM811 in the presence of a standard electrolyte (1 M LiPF₆ in EC/DEC/DMC, 1:1:1 by volume). The pristine NCM811 cathode exhibited a sharp exothermic peak with a total heat release of 1,271 J g⁻¹, reflecting its high thermal reactivity. In contrast, the FRI-modified cathode exhibited a significantly reduced exothermic peak, with the peak temperature shifting from 202 to 247 °C and a total heat release of 689 J g⁻¹ (Fig. 2E). The second derivative of specific heat flow (d²q) showed a significant reduction in fluctuations with temperature, indicating that FRI@NCM811 undergoes a milder and more controlled exothermic reaction (SI Appendix, Fig. S6). This suggests that the FRIs significantly slows down the kinetics of the exothermic reaction. Furthermore, to validate the FRI strategy’s compatibility with ether-based electrolytes, we fabricated FRI@S/C cathode and performed DSC analysis with 1 M LiTFSI in DOL/DME (1% LiNO₃ additive). The fully charged S/C cathode exhibited four exothermic peaks at 116 °C, 206 °C, 352 °C, and 450 °C, with a total heat release of 2,357 J g⁻¹. In contrast, the fully charged FRI@S/C cathode showed three exothermic peaks at 211 °C, 326 °C, and 441 °C, with a significantly reduced total heat release of 1,454 J g⁻¹—a 38% reduction (SI Appendix, Fig. S7). The shift to higher onset temperatures and the suppression of overall heat generation demonstrate the adaptive compatibility of the FRI strategy with both carbonate- and ether-based electrolyte systems.

Oxygen evolution, strongly correlated with thermal stability, was also investigated using evolved gas analysis-mass spectrometry (EGA-MS). High-nickel cathodes like NCM811 typically undergo phase transitions (layered → spinel → rock salt) during thermal abuse, releasing lattice oxygen that subsequently oxidizes the electrolyte (10). Quantitative analysis showed that the FRI-modified NCM811 exhibited a 49% reduction in oxygen evolution compared to the pristine cathode (Fig. 2F). Additionally, CO₂ production in the 200 to 800 °C temperature range under a helium (He) atmosphere is significantly reduced for FRI-modified samples, indicating that the FRIs acted as a physical barrier, suppressing the oxidation of conductive carbon by the highly oxidative cathode (SI Appendix, Fig. S8). In addition, to further investigate the structural evolution under elevated temperatures, fully charged NCM811 and FRI@NCM811 electrodes were subjected to postthermal treatment at 300 °C for 30 min, followed by high-resolution transmission electron microscopy (HRTEM) analysis (SI Appendix, Fig. S9A). The untreated NCM811 underwent an irreversible phase transition to a rock-salt structure, characterized by a lattice spacing of 0.24 nm. In contrast, FRI@NCM811 preserved its original layered structure with a lattice spacing of 0.47 nm, along with the emergence of a crystalline FRP phase exhibiting a spacing of 0.34 nm (SI Appendix, Fig. S9B). These findings confirm that FRIs enhance the thermal stability of NCM811 by mitigating heat and oxygen release during thermal abuse.

Pouch Cell Thermal Safety with FRI-Modified Cathodes.

To assess the effectiveness of FRIs at the cell level, thermal stability tests were conducted using 0.58-Ah Li||NCM811 pouch cells under adiabatic conditions with accelerating rate calorimetry (ARC). The cells were cycled and fully charged to 4.3 V, a highly hazardous state-of-charge, to evaluate their behavior under thermal abuse (SI Appendix, Fig. S10). Both pristine and FRI-modified pouch cells exhibited similar self-heating onset temperatures (T₁) of approximately 120 °C (Fig. 3 A and B and Table 1). However, their thermal behavior diverged significantly beyond this point. The pristine cell underwent rapid thermal runaway, with an accelerating temperature (T₂: defined as the temperature when the heating rate is 1 °C min⁻¹) of 127 °C and a catastrophic maximum temperature (T₃) of 1,038 °C within just 13.2 mins. The maximum self-heating rates (dT/dt_max) reached an alarming value of 4.33 × 10⁴ °C min⁻¹ (Fig. 3 C and E). This explosive behavior was attributed to mechanical rupture of the pouch, allowing accumulated gas to escape and react with air, triggering severe combustion (Fig. 3G and SI Appendix, Fig. S11A). In contrast, the FRI-modified pouch cell demonstrated significantly enhanced thermal stability. The T₂ was delayed to 166 °C, and the T₃ was substantially reduced to 220 °C, occurring over an extended period of 1596.6 mins. The maximum temperature rise rate was drastically reduced to just 1.1°C min⁻¹ (Fig. 3 D and F). Posttest analysis revealed only swelling in the FRI-modified cell, with the pouch remaining intact (Fig. 3G and SI Appendix, Fig. S11B). These results highlight the critical role of FRIs in suppressing gas generation, mitigating internal pressure buildup, and preventing mechanical rupture, effectively halting thermal runaway.

Fig. 3.

Thermal safety evaluation of Li||NCM811 pouch cells using ARC. (A and B) ARC profiles and corresponding voltage variations of Li||NCM811 (A) and FRI@NCM811||Li (B) pouch cells under thermal abuse conditions. Temperature rise rate curves of Li||NCM811 (C) and Li||FRI@NCM811 (D) pouch cells, highlighting the suppression of rapid temperature increases in FRI-modified cells. (E and F) Internal pressure evolution during thermal abuse for Li||NCM811 (E) and Li||FRI@NCM811 (F) pouch cells. (G) Post-ARC optical images of Li||NCM811 and Li||FRI@NCM811 pouch cells, showing catastrophic rupture in the control cell and intact swelling in the FRI-modified cell. (H) Comparative summary of accelerating temperature (T₂) and maximum self-heating rates (dT/dt_max) for various pouch cell systems with Ni-rich cathodes and different anodes, including graphite, Si/Gr composites, and Li metal, based on data from this work and literature reports. The hollow circles and solid circles represent the control and modified cells, respectively (8, 32–40).

Table 1.

T₁, T₂, T₃, and dT/dt of Li||NCM811 and Li||FRI@NCM811 pouch cells

Pouch cell	t1 (min)*	T1 (oC)	t2 (min)	T2 (oC)	t3 (min)	T3 (oC)	dT/dtmax (oC min−1)
Li||NCM811	0	122	12.1	127	13.2	1038	4.33 × 104
Li||FRI@NCM811	0	117	1058.3	166	1596.6	220	1.1

^*The t₁ time in the table has been adjusted to start from zero.

Both of T₂ and dT/dt_max are key metrics for evaluating the thermal safety of different battery systems. To further validate the thermal stability of the FRI-modified pouch cell, T₂ and dT/dt are compared with previously reported systems featuring Ni-rich cathodes and various anodes, including graphite, Si/Gr composites, and Li metal (Fig. 3H and SI Appendix, Table S1) (8, 32–40). The comparison reveals that batteries using Li metal anodes generally exhibit lower T₂, indicating higher thermal risks. Notably, dT/dt_max reported for most pouch cells in the literature range between 10³ and 10⁵ °C min⁻¹, regardless of the anode material. However, the FRI-modified pouch cell achieved a dramatic reduction in thermal runaway rates by 3 to 5 orders of magnitude, underscoring its superior thermal safety. Kinetic analysis using the Arrhenius equation further supports these findings, quantifying the thermal runaway characteristics of the pouch cells. The activation energy (E_a) and frequency factor (A) of the self-heating reaction were significantly reduced for the FRI-modified cell compared to the control sample (SI Appendix, Fig. S12 and Table S2). This reduction indicates that the incorporation of FRIs effectively alters the thermal reaction kinetics, leading to suppressed exothermic behavior and enhanced safety.

Interestingly, the FRI-modified cell exhibited an earlier onset of self-heating (T₁), occurring 5 °C lower than that of the control cell. While a lower self-heating temperature is typically considered undesirable as it suggests earlier thermal activity, in this case, it reflects a favorable phenomenon. This earlier onset is attributed to the temperature-triggered decomposition of the FRI at approximately 100 °C, which releases flame-retardant radicals ([PO]·). These radicals act proactively to suppress heat generation and gas release during the initial stages of thermal abuse, mitigating risks before the system reaches critical runaway conditions. This “smart response” effectively stabilizes the cell early in the thermal abuse process, preventing catastrophic events such as gas accumulation, mechanical rupture, or uncontrolled temperature rise. A 0.5 Ah-Li metal pouch cell with an FRI@NCM811 cathode was fabricated for nail penetration testing (SI Appendix, Fig. S13). Remarkably, the cell exhibited a maximum surface temperature of below 32 °C during penetration. Furthermore, an overcharge test was conducted on a 0.2 Ah-Li‖FRI@NCM811 pouch cell (cathode areal capacity: 2.5 mAh cm⁻²). Upon charging to 4.8 V, abrupt voltage drops were observed at 4.5 V, 4.7 V, and 4.8 V, yet the cell exhibited no significant temperature rise throughout the test (SI Appendix, Fig. S14). These results confirm the FRI's effectiveness in suppressing exothermic chain reactions under mechanical abuse and electrical abuse conditions. Consequently, the ability of the FRIs to intelligently respond at an early stage highlights the advanced safety mechanisms afforded by this strategy, making it particularly advantageous for high-energy-density Li metal batteries.

Smart Gas Management in Pouch Cells under Thermal Abuse.

Gas generation plays a critical role in driving thermal runaway in Li metal batteries. The thermal decomposition of electrolyte solvents (EC, DMC, DEC) generates a variety of transient species (H·, CH·, CH₃·, C₂H₅·) that lead to the formation of flammable hydrocarbon gases, such as CH₄, C₂H₆, C₂H₄, and C₃H₆ (40, 41). These reactions are primarily driven by free radical mechanisms. The phosphonate-based FRP acts as an effective quencher of free radicals, suppressing the chain reactions responsible for gas generation (Fig. 4A). By trapping these radicals, FRIs not only reduce the amount of flammable gases produced but also alter the gas composition to favor less combustible species, contributing to enhanced thermal safety.

Fig. 4.

Gas generation reactions and gas production analysis during ARC test. (A) Mechanism of FRI’s temperature-responsive quenching of gas-generating reactions. (B) Schematic of gas production testing during the thermal runaway process. (C) Comparison of gas production volume for NCM811 and FRI@NCM811 during thermal runaway. (D) Gas composition ratio for flammable hydrocarbons, CO₂ and O₂ in NCM811 and FRI@NCM811 pouch cells. (E) Explosive limits of flammable gases (CH₄, C₂H₆, C₂H₄, and C₃H₆) generated during the process.

To further investigate the effects of FRIs on gas behavior, we measured the total gas production and analyzed the gas composition during thermal runaway. The gas production process was evaluated using pouch cells placed in a sealed chamber filled with inert N₂ (Fig. 4B). For the pristine cell, thermal runaway caused gas accumulation to rupture the pouch, allowing the gases to diffuse into the chamber. Gas samples were extracted for gas chromatography–mass spectrometer (GC–MS) analysis. For the FRI-modified cell, mechanical rupture did not occur due to suppressed gas accumulation, so the pouch was manually punctured after cooling to facilitate gas extraction. The total gas production of the FRI-modified pouch cell was reduced by 63% compared to the pristine cell (Fig. 4C). Notably, the production of flammable hydrocarbon gases (CH₄, C₂H₆, C₂H₄, C₃H₆) decreased significantly. Gas composition analysis (Fig. 4D) revealed a shift in gas composition: In the FRI-modified cell, the proportion of nonflammable CO₂ increased to 61%, while flammable hydrocarbons decreased from 62% (in the pristine cell) to 19%. CH₄, C₂H₆, C₂H₄, and C₃H₆ concentrations were all reduced by a factor of 5 to 11. Among these gases, C₂H₄ is particularly critical due to its broadest explosive limit and high reactivity (Fig. 4E) (42, 43). The reduction in C₂H₄ content, which is primarily generated through the reduction of the electrolyte by Li metal, is especially important for improving the thermal safety of Li metal batteries (44). Interestingly, the O₂ content in the gas of FRI-modified cell increased slightly, which may result from the FRIs suppressing the reaction between O₂ and flammable gases, thereby reducing oxygen consumption. These results demonstrate that smart gas management, achieved through the introduction of FRIs, effectively minimizes gas production, alters gas composition to reduce flammability, and mitigates internal pressure buildup during thermal abuse, significantly enhancing the safety of Li metal batteries.

Electrochemical Performance of FRI-Modified Batteries.

The incorporation of FRIs not only significantly enhances the thermal safety of Li metal batteries but also maintains their excellent electrochemical performance. A key advantage of the smart gas management strategy is its compatibility with existing cathode preparation processes, making it highly scalable for industrial production. As illustrated in Fig. 5A, the preparation process involves adding the FRP precursor into the NCM811 slurry, followed by rapid UV curing after the slurry is coated onto Al foil. Any unconverted monomers are effectively removed during the subsequent solvent drying step. The inset of Fig. 5A displays the optical image of a long FRI@NCM811 cathode, emphasizing the feasibility of scaling up the process for large-scale applications. Previous studies have demonstrated that phosphate-based flame-retardant additives can enhance battery safety by suppressing electrolyte flammability; however, their poor reductive stability toward Li metal anodes often compromises electrochemical performance (45, 46). The strategy overcomes this limitation by confining the FRP exclusively within the cathode. By isolating the flame-retardant components from the anode, this innovative approach eliminates detrimental interactions with the highly reactive Li metal, ensuring stable cycling performance and preserving room-temperature battery functionality. The X-ray photoelectron spectroscopy (XPS) result exhibits that the cycled Li anode with LiBF₄-based carbonate electrolytes exhibits no detectable signals in P 2p spectrum, conclusively demonstrating that FRP components do not undergo crossover or deposition on Li metal during operation (SI Appendix, Fig. S15). This result aligns with our interfacial stability hypothesis and confirms the spatial confinement of FRI within the cathode interface.

Fig. 5.

Electrochemical performance of FRI@NCM811 batteries. (A) Schematic illustration of the preparation process for FRI@NCM811. The inset shows an optical image of the FRI@NCM811 cathode prepared for industrial scalability. (B) Cycling performance and (C) charge-discharge curves of the Li||FRI@NCM811 coin cell. (D) Projected energy density of Gr||FRI@NCM811, Si/C | |FRI@NCM811, Li||FRI@NCM811, and Li-free||FRI@NCM811 pouch cells with a design capacity of 10 Ah.

Electrochemical testing of the FRI@NCM811 full-cell revealed remarkable cycling stability, retaining approximately 80.0% of its initial capacity after 200 cycles at 0.5 C (Fig. 5B). The corresponding charge-discharge curves (Fig. 5C) further demonstrate a reversible capacity of 180 mAh g⁻¹ at 0.5 C, highlighting the ability of FRIs to support prolonged cycling without significant degradation. To further validate the feasibility of FRI@NCM811 under practical conditions, a Li||FRI@NCM811 pouch cell was fabricated and delivers a capacity retention of 85.5% after 25 cycles at 0.2 C (SI Appendix, Fig. S16). Following extended cycling, we conducted a series of structural and thermal analyses to assess the functional integrity of the FRI. HRTEM imaging of the 50-cycle FRI@NCM811 electrode revealed a dominant lattice spacing of 0.34 nm, consistent with FRP crystallographic features (SI Appendix, Fig. S17). This nanoscale coexistence confirms the structural integrity of the FRI at the electrode–electrolyte interface during extended cycling, with FRP domains maintaining their phase purity despite continuous electrochemical operation. Moreover, Raman spectroscopy of the cycled FRI@NCM811 electrode shows a persistent P–O–C vibrational signal at 808 cm⁻¹, providing direct spectroscopic evidence for the retention of FRP (even below 5 wt%) within the electrode matrix after prolonged cycling (SI Appendix, Fig. S18).

To assess whether the FRI strategy retains its thermal protection functionality after extended use, DSC measurements were conducted on electrodes at 100% state-of-charge (SOC) after 100 cycles. As shown in SI Appendix, Fig. S19, the cycled FRI@NCM811 exhibited a delayed exothermic onset (237 vs. 218 °C for cycled bare NCM811) and a 34% reduction in total heat release (952 J g ⁻¹ vs. 1,439 J g ⁻¹), confirming its continued effectiveness in mitigating thermal runaway risks. Notably, we introduced the Specific Heat Release per Capacity (SHRC) metric—defined as the heat released per gram of active material and per ampere-hour of capacity (J g ⁻¹ Ah⁻¹), normalized by state-of-health (SOH)—to account for capacity fade during cycling. The cycled FRI@NCM811 system showed a significantly lower SHRC of 1.3 J g ⁻¹ Ah⁻¹ compared to 1.8 J g ⁻¹ Ah⁻¹ for the cycled bare NCM811, further evidencing the FRI’s sustained thermal regulation capability under practical cycling conditions.

To investigate the influence of the FRI modification on electrochemical charge/discharge kinetics, electrochemical impedance spectroscopy (EIS) and rate-performance tests were conducted (SI Appendix, Fig. S20). The observed marginal capacity reduction and increased impedance indicate that FRI induces slightly polarization increase and retarded kinetics. However, given the profound enhancement in battery thermal safety afforded by the FRI strategy, this comprehensive multilevel safety improvement—spanning the material, cell, and system levels—justifies its implementation, despite the associated moderate kinetic penalties. Additionally, future work could aim to mitigate these kinetic issues by reducing the FRP content and reinforcing the conductive network.

The scalability and effectiveness of the smart gas management strategy were further demonstrated in larger-scale applications. Beyond Li metal anodes, the versatility of the FRI-modified cathode enables compatibility with a wide range of anode chemistries, including graphite, Si/C composites, and even Li-free designs. When paired with a Li-free anode, the specific energy of a 10-Ah Li-free||FRI@NCM811 pouch cell was projected to reach as high as 470 Wh kg⁻¹ (Fig. 5D, SI Appendix, Table S3), showcasing the potential of FRI-modified cathodes for high-energy-density applications. While the FRI strategy demonstrates exceptional effectiveness in smartly addressing thermal abuse risks and regulating heat/gas generation, it is crucial to note that the nonconductive nature of FRP polymers necessitates future studies to mitigate its adverse impact on battery rate capability.

Discussion

In summary, this work presents a smart gas management strategy to achieve fire-safe Li metal batteries by integrating continuous FRIs into the cathode. The introduction of FRIs significantly enhances the thermal stability of the cathode, as evidenced by a 49% reduction in oxygen release, a delayed peak exothermic temperature of the cathode–electrolyte system from 202 to 247 °C, and a 45% decrease in heat release. The thermal decomposition of FRIs around 100 °C (the initial stage of thermal runaway) triggers the gradual release of flame-retardant radicals, which block electrolyte decomposition and gas generation reactions early on. This dual-action mechanism results in a 63% reduction in total gas production, with the proportion of flammable hydrocarbon gases decreasing from 62 to 19%, significantly reducing the risk of thermal runaway. Furthermore, the reduction in oxygen and flammable gases also mitigates the risks associated with cross-talk reactions, which can exacerbate thermal runaway. Gas accumulation plays a critical role in thermal runaway, often leading to mechanical rupture and explosive combustion, which causes an exponential increase in the battery’s temperature rise rate. This phenomenon, with temperature rise rates reaching up to 43,300 °C min⁻¹, was experimentally validated in a 0.58-Ah pouch cell. In contrast, the FRIs reduced gas accumulation, preventing thermal runaway and limiting the maximum temperature to 220 °C, down from 1,038 °C. The temperature rise rate in the nonexplosive cells was reduced to just 1.1 °C min⁻¹. Importantly, the proposed strategy not only enhances thermal safety but also preserves the electrochemical performance of the battery. After 200 cycles, the FRI-modified cells retained 80% of their capacity and demonstrated a reversible capacity of 180 mAh g⁻¹ at 0.5 C. By addressing both thermal safety and electrochemical stability, this smart gas management strategy bridges a critical gap in the development of next-generation energy storage technologies, providing a pathway toward fire-safe, high-energy-density batteries for electric vehicles and other applications.

Materials and Methods

Preparation of NCM811 Cathodes with FRIs.

The NCM811 cathode slurry was prepared by mixing NCM811, Super P, and PVDF (polyvinylidene difluoride) in NMP (N-methyl-2-pyrrolidone) with a mass ratio of 8:1:1. The FRP precursor solution was prepared by mixing DEAP (diethyl allylphosphonate; Sigma-Aldrich Inc.) and ETPTA (ethoxylated trimethylolpropane triacrylate; Sigma-Aldrich Inc.) in a 1.3:1 volume ratio, which was then combined with a PVDF-HFP solution (55 mg μL ⁻¹ in NMP) in a 1:1 volume ratio. This precursor solution was added to the NCM811 slurry at a dosage of 0.18 μL mg⁻¹ (relative to the NCM811 mass), followed by the addition of a photoinitiator (2-hydroxy-2-methyl-1-phenyl-1-propanone; Sigma-Aldrich Inc.) at a dosage of 0.04 μL mg⁻¹ (relative to the NCM811 mass). The mixture was thoroughly stirred to prepare the FRI@NCM811 cathode slurry, which was cast onto carbon-coated aluminum foil. The coated slurry was cured under UV light (2 kW) for 15 min and dried at 80 °C for 12 h to remove NMP and unconverted monomers.

Preparation of S/C Cathodes with FRIs.

The S/C cathode (70 wt% S and 30 wt% CNT) slurry was prepared by mixing NCM811, Super P, and PVDF (polyvinylidene difluoride) in NMP (N-methyl-2-pyrrolidone) with a mass ratio of 8:1:1. The above FRI precursor solution was added with a dosage of 0.18 μL mg⁻¹ (relative to the S/C mass), followed by the addition of photoinitiator at a dosage of 0.04 μL mg⁻¹ (relative to the S/C mass).

Preparation and Thermal Analysis of the FRP Film.

The FRP precursor solution was cast onto a clean glass plate and cured under UV light (2 kW) for 15 min. The cured film was dried at 80 °C for 12 h to remove NMP and unconverted monomers, then carefully separated from the glass substrate using a blade. To assess the thermal stability of the FRP film, thermogravimetric analysis (TGA, STA 449F3 Jupiter) was performed under a nitrogen atmosphere at a scan rate of 10 °C min⁻¹.

Cell Fabrication and Electrochemical Testing.

CR2032-type coin cells were assembled in an argon-filled glove box (H₂O < 0.1 ppm, O₂ < 0.1 ppm). Each cell consisted of a metallic Li foil (thickness: 500 μm), a Celgard 2,400 separator, and either an NCM811 or FRI@NCM811 cathode. The FRI@NCM811 cathodes had an areal mass loading of 6.4 mg cm⁻², and 80 μL of an electrolyte solution (1 M LiPF₆ in EC/DEC/DMC, 1:1:1 by volume, with 2% fluoroethylene carbonate additive, Canrd Technology Co. Ltd., Code MA-EN-ET-0351) was added. Cells were activated at 0.1 C (1 C = 200 mA g ⁻¹) for two cycles before cycling at 0.5 C within a voltage range of 2.8 to 4.3 V. For DSC testing, the FRI modified cells were prepared with 80 μL of a carbonate-based electrolyte (1 M LiPF₆ in EC/DEC/DMC, 1:1:1 by volume) without additives. The NCM811 full-cells were cycled at 0.2 C for five cycles (2.8 to 4.3 V) before being charged to 4.4 V with a constant voltage hold at 4.4 V. The S/C full-cells were cycled at 0.1 C (1 C = 1,675 mA g ⁻¹) for three cycles (1.8 to 2.8 V) before being charged to 2.8 V with a constant voltage hold at 2.8 V. Pouch cells were fabricated in a dry room (dew point < –50 °C) using FRI@NCM811 or NCM811 cathodes with an areal mass loading of 15.1 mg cm⁻², 50-μm-thick metallic Li foils coated on copper, and Al₂O₃-coated separators. A commercial carbonate electrolyte (3 g Ah⁻¹) was injected into the pouch cells, which were sealed under an argon atmosphere.

Material Characterization.

Scanning electron microscopy (SEM, JEOL 8100) was used to observe the microstructure of NCM811 and FRI@NCM811 cathodes. A ToF-SIMS M6 mass spectrometer (ION-ToF GmbH, Münster, Germany) equipped with a 30 keV Bi₃⁺ primary ion gun was employed to collect the signals of NiO₂⁻, CoO₂⁻, MnO₂⁻, PO₂⁻, and PO₃⁻ ions from FRI@NCM811. A Cs⁺ ion gun (500 eV, 42 nA) was applied to sputter the surface of the sample. The Bi₃⁺ primary ion source was operated using a pulsed beam at 10 kHz (mass range: 0 to 543 Da). Negative ion spectra were calibrated by O⁻, OH⁻, C₂⁻, C₂H⁻, and Cl⁻, etc. High spatial resolution images were collected in fast imaging mode and using delayed extraction method by 256 × 256 pixels over a 50 × 50 μm² area with the highest resolution of ca. 200 nm. XPS (XPS, ESCALAB250XI, Thermo Fisher Scientific Inc.) was performed to analyze surface chemistry. Raman spectra were collected on a confocal laser Raman spectrometer (LabRAM HR Evolution).

Thermal Safety Testing.

DSC was performed on fully charged NCM811 and FRI@NCM811 samples under argon with a heating rate of 10 °C min⁻¹. Cathode powders were prepared for DSC by disassembling cycled cells, washing cathodes three times with DMC, and scraping active materials from the aluminum foil. The powders were sealed with 1 μL mg⁻¹ carbonate electrolyte (1 M LiPF₆ in EC/DEC/DMC, 1:1:1 by volume) in a high-pressure crucible with a gold-plated surface inside an argon-filled glove box. The DSC test of S/C and FRI@S/C was performed with the same method. ARC testing was performed on 0.58-Ah pouch cells using an EV⁺-ARC system. Cells were activated at 50 mA for one cycle, charged to 4.3 V, and held at a constant voltage until fully charged. The test began at 50 °C with 5 °C incremental heating steps. Self-heating was identified when the temperature rise reached 0.02 °C min⁻¹. For gas analysis, pouch cells were placed in a sealed chamber filled with inert N₂. During ARC testing, the gas production was collected, and a GC–MS analysis was performed to analyze the composition of the generated gases. Nail test and overcharge test was commissioned by authorized third parties, China Automotive Battery Research Institute Co., Ltd.

Supplementary Material

Appendix 01 (PDF)

Movie S1.

Evolution of Chemical Maps for FRI@NCM811 Cathodes During ToF-SIMS Analysis. The video illustrates the chemical map evolution of FRI@NCM811 cathodes observed in a time-of-flight secondary ion mass spectrometry (ToF-SIMS) test. The playback speed has been accelerated 20 times for clarity and visualization.

Data, Materials, and Software Availability

All study data are included in the article and/or supporting information.