Mechanistic Understanding of Thermal Stability and Safety in Lithium Metal Batteries

Abstract

As lithium-ion batteries approach their theoretical capacity limits, lithium metal batteries (LMBs) have emerged as promising candidates for next-generation energy storage, offering substantially higher energy densities. However, their practical deployment remains limited by several interrelated challenges including lithium dendrite growth, parasitic side reactions, unstable solid electrolyte interphases (SEI), and poor cycling stability. While recent advances in electrolyte design, anode architecture, and interfacial engineering have significantly improved electrochemical performance, the thermal stability and safety of LMBs, particularly at the interface and electrode levels, still require extensive investigation. This review provides a comprehensive mechanistic analysis of thermal instability in LMBs, spanning material degradation, interfacial decomposition, and cell-level thermal behavior. We critically examine the roles of lithium metal, liquid- and solid-state electrolytes, and diverse cathode chemistries (e.g., layered oxides, sulfur) in triggering exothermic reaction pathways, gas evolution, and thermal runaway. The complex coupling among electrode–electrolyte interactions, interphase chemistry, electrochemo-mechanics, morphological evolution, and thermal instability across emerging LMB chemistries is highlighted. By identifying dominant thermal instability mechanisms and key knowledge gaps, this review establishes a mechanistic foundation for designing thermally resilient LMBs and outlines future directions for advancing safety in high-energy battery systems.

1. Introduction

Lithium-ion batteries (LIBs) have become universal with the advent of consumer electronics and electric vehicles and their application in power tools and energy grids. Their widespread adoption is due to favorable attributes such as low self-discharge, long cycle life, and reduced carbon footprint compared to fossil-fueled counterparts. , However, global energy demands are forecasted to grow by an average of 3.4% each year, with the contribution from electricity consumption nearing 30% of total energy consumption by 2030. , Further, with widespread usage of artificial intelligence, electricity consumption from data centers alone is expected to reach 1000 TWh in the next year. The stricter policies adopted by governments to mitigate global warming are anticipated to exacerbate the load on energy grids, with storage installations across the US reaching 9 GW in the last year. Although state-of-the-art LIBs with graphite as the anode currently offer high energy densities (up to 250 Wh kg^–1), their ability to meet growing future energy demands remains constrained by the inherent theoretical limits of their material-specific capacities. − As a result, alternative high-energy-density systems have garnered significant research attention over the past decade, with lithium metal batteries (LMBs), wherein lithium (Li) metal is used instead of graphite as the anode, emerging as a leading candidate. The higher specific energy of LMBs is primarily due to the high theoretical capacity of Li metal (3860 mAhg^–1) and its low electrode potential (−3.04 V vs SHE), thus offering a promising pathway to overcoming the limitations of LIBs. Furthermore, Li metal serves as a pivotal transition material, enabling the development of advanced battery chemistries such as lithium–sulfur (Li–S) and solid-state batteries (SSBs). These next-generation systems have the potential to achieve energy densities comparable to those of fossil fuels, offering a transformative step forward in energy storage technologies (Figure a). − While emerging chemistries based on metallic anodes, such as sodium, magnesium, and calcium, offer promising alternatives owing to their elemental abundance and, in some cases, higher theoretical energy densities, they remain in the early stages of development. In contrast, LMBs are approaching commercial viability. Given this relative technological maturity, the present review centers on a detailed analysis of the degradation pathways and thermal stability challenges specific to LMBs.

1.

(a) Comparison of energy density and heat release per unit weight between conventional Li-ion and next-generation Li-metal chemistries at the cell-level (data accessed from refs , , ). (b) Laboratory studies on next-generation Li-based batteries have predominantly focused on electrochemical performance and material/cell characterization, with limited emphasis on thermal instabilities, which are critical for practical deployment.

The development of LMBs dates back to 1980, when Moli Energy incorporated Li metal as an anode against an intercalation cathode in a cylindrical cell format. However, recurring fire incidents caused by dendrite formation led to widespread recall of these cells. Similarly, companies like NEC and Mitsui also invested heavily in Li metal-based batteries but could not meet the necessary safety standards for practical deployment, highlighting the significant challenges associated with dendrite growth and thermal runaway. − During this period, Sony pioneered the development of the first commercial LIBs, leveraging their relatively stable chemistry, which led to a shift away from further LMB research. Li metal’s high chemical reactivity, poor electrochemical performance, and limited cycle life posed significant challenges and hindered progress in LMB development for decades. However, the recent advancement in surface science and deeper mechanistic understanding of battery constituents have renewed interest in LMB, primarily driven by the promise of higher specific capacity. − For example, implementing an artificial SEI has shown promise in mitigating the corrosion of the active Li metal, thereby enhancing the cycle life. Further, 3D current collectors or electrically conductive nanoshells resulted in dense dendritic-free deposition of Li, reducing the chances of cell shorting. , Such advancements have expanded the range of chemistries compatible with Li metal, paving the way for high-energy systems, such as Li–S batteries and SSBs. In Li–S chemistry, sulfur (S), typically used as a composite with carbon black and binders, serves as the cathode material instead of conventional layered oxides, owing to its low cost and high theoretical energy density (2600 Wh/kg). Nevertheless, their electrochemical performance was limited during the early development stages due to the corrosion of electrodes, large volumetric expansion, and polysulfide shuttle, hindering practical implementation. , In contrast, SSBs employ solid electrolytes (SEs) between the electrodes, effectively replacing both liquid electrolytes (LEs) and separators by physically isolating the electrodes while enabling ionic diffusion. The development of SSBs was initially motivated by theoretical studies suggesting that the high shear modulus of SEs could suppress dendritic Li growth. Early research on SSBs focused on the development of various SEs, such as polymers, sulfides, and oxides, that exhibited high ionic conductivity (≥0.1 mS/cm) and low electronic conductivity (10^–7 mS/cm). , Despite these advantages, recent studies have discovered various mechanisms and pathways that could lead to electrochemical cell failure, ranging from Li penetration through the SE, unstable interphase growth, and void formation. , In addition, limited solid–solid contact between the electrolyte and electrode in SSBs results in higher polarization, poor cycling capability, and lower capacity retention in comparison to their liquid counterparts.

While numerous strategies, including foreign atom doping, artificial SEI, surface coating, alloy-based anodes, and interlayers, have been proposed to overcome these drawbacks, limited attention is given to the factors that influence the thermal stability of LMBs. Thermal instability in batteries occurs when the rate of internal heat generation exceeds the heat dissipation rate. For instance, a pristine cell free of defects remains thermally stable under nominal operations as the internal heat (e.g., ohmic, reaction, and entropic) is balanced by heat dissipation through the materials. However, under extreme operating conditions (e.g., thermal or mechanical abuse and high current loads), internal heat generation increases significantly beyond the rate of dissipation, driving the battery into a thermally unstable regime. Once triggered, the battery enters a self-heating phase, in which excess heat accelerates side reactions and material degradation. Because these reactions are often exothermic, they further raise the cell temperature, intensifying heat generation and triggering a feedback loop of the escalating temperature and reaction rates. When heat generation exceeds the thermal capacity of the cell components, the system undergoes thermal runaway, often culminating in catastrophic failure. The highly reactive and corrosive nature of Li metal significantly increases the risk of thermal runaway in LMBs. In comparison to LIBs, they release greater amounts of heat and reach higher temperatures during thermal runaway, driven by their higher energy density (Figure a). − The heat released and energy density of present SSBs have been calculated by considering a SE thickness of 500 μm and a cathode with 60% active material loading. However, to meet practical requirements, future SSBs are expected to adopt thinner SEs and higher cathode loadings. Accordingly, the energy density and heat release corresponding to future SSBs (Figure a) are calculated by considering a 20 μm-thick SE and 70% active material loading. As noted, this increase in the energy density is accompanied by a significant rise in heat release for LMBs. A recent study quantitatively demonstrated that thermal instability increases with energy density, highlighting the need for significantly higher cooling power to ensure safe operation. Post-mortem analyses of failed cells, incorporating both material and cell-level characterization, have identified internal short circuits, primarily caused by dendritic Li growth and subsequent separator degradation, as the dominant trigger for thermal runaway in LMBs.

Additional pathways for thermal runaway include chemical crosstalk between the electrodes and thermal degradation of LEs, with the onset temperature closely linked to the flash point of the electrolyte. , Thus, replacing conventional separators with SEs and substituting oxide cathodes with S-based cathodes has been proposed to suppress exothermic reactions and mitigate the risk of thermal runaway. − However, recent studies have reported thermal stability concerns in SSBs despite the high shear modulus and nonflammability of SEs. , These thermal runaway scenarios are attributed to the thermodynamic instability of the SE/electrode interface, gas generation, chemical crosstalk between electrodes, and short-circuiting due to dendritic Li growth through the SE. − Furthermore, the potential heat release in next-generation SSBs with higher energy densities has been estimated to be nearly double that of LIBs, underscoring an elevated risk of thermal runaway (Figure a). Alternatively, for Li–S chemistry, despite the S cathode not releasing reactive gas upon pulverization at elevated temperatures, thermal runaway was triggered by direct redox reaction between the electrodes. These findings have renewed interest in the thermal stability of LMBs, prompting extensive research efforts using a variety of techniques. These approaches span from advanced material characterization to in-depth thermal analysis at the laboratory scale, aiming to investigate the fundamental mechanisms that trigger thermal instability in LMBs − (Figure b). However, current probing methodologies offer limited insight into the true nature of thermal stability, as they typically involve harvesting electrodes from electrochemically conditioned cells and reassembling them into fresh cells to analyze thermal runaway. Some studies have examined the effect of material properties and operational parameters such as cycling history, cathode state of charge (SOC), electrolyte flash point, and decomposition temperature, on thermal stability. In addition, thermal stability can be influenced by a broad range of mechanistic aspects, including the amount of LE, rate of free radical generation, solvation strength, composition and morphology of the interphase, aging, thickness of SE, active material loading, and the rate of temperature increase under various operating conditions (e.g., temperature, pressure, and current density). This multiscale dependency, coupled with the fragile and buried nature of interphases (e.g., in SSBs), poses significant challenges to analyzing the underlying mechanisms governing their thermal behavior.

With the growing demand for high-energy-density systems and increased concerns over thermal safety, a comprehensive understanding of thermal runaway pathways is critical for the advancement of LMBs. This review presents a mechanistic analysis of thermal instability in LMBs, delineating the wide range of material-, interface-, and cell-level factors affecting thermal safety. Section examines the intrinsic thermal stability of core components, including Li metal, electrolytes, and cathodes, highlighting fundamental differences across various material classes. Section explores the thermal responses of liquid- and solid-state electrolytes in combination with Li metal and diverse cathode chemistries, including layered oxides, spinels, olivines, and S-based systems, with an emphasis on interfacial reaction mechanisms that govern electrochemical and thermal behavior. Section delves into interelectrode chemical crosstalk, the influence of electrochemically inactive materials, cycling-induced degradation, and internal short circuits, which are factors often underrepresented in previous studies. Lastly, Section identifies emerging exothermic pathways that contribute to thermal runaway and highlights advanced characterization techniques for probing interphase evolution, decomposition products, and morphological changes. While prior research has largely focused on the reactivity of Li metal, − this review expands the scope by offering a deeper understanding of the thermal instability pathways across different chemistries, such as LE-based LMBs, SSBs, and Li–S batteries. This work provides a foundation for the rational design of thermally resilient LMBs and informs the development of next-generation batteries for applications ranging from electric vehicles to aerospace and electrified aviation.

2. Material-Level Thermal Stability

Most material properties, including physical, chemical, and ionic conductivities, follow Arrhenius-type temperature dependence. As a result, material behavior can differ significantly at elevated temperatures compared to room temperature. For instance, Li metal undergoes a phase transition from solid to liquid at a relatively low temperature of 180 °C. Similarly, oxide-based cathodes become thermodynamically unstable around 200 °C, leading to the release of oxygen (O₂) gas due to lattice instability (further details are provided in the following sections). , Given that battery temperatures during thermal runaway events can exceed 200 °C, understanding the thermal stability of individual materials is critical for the initial screening and design of thermally safe batteries. Additionally, identifying the decomposition temperatures and the nature of the resulting products provides valuable insight into the onset conditions of exothermic reactions, which in turn dictate the severity and kinetics of thermal runaway at the component level. The following section reviews extensive research efforts aimed at evaluating the thermal stability of key LMB components at the material level.

2.1. Lithium Metal

The higher energy density of LMBs (40% greater than LIB) is primarily attributed to Li metal’s high specific capacity (3860 mAhg^–1) and low electrode potential (−3.04 V vs SHE). This has generated significant interest in Li metal as the negative electrode for next-generation battery chemistries. However, its high chemical reactivity introduces critical safety challenges, particularly due to its tendency to react exothermically with various battery components and their decomposition products. For instance, when Li encounters gaseous products, such as nitrogen (N₂), carbon dioxide (CO₂), and O₂, it can initiate violent reactions that release significant amounts of heat (ranging from 54 to 300 kJ mol^–1). These reactions can sharply increase local temperatures, − posing severe risks to cell stability. Gases evolved from LEs upon heating can further exacerbate this issue by reacting with Li metal, creating a feedback loop in which rising temperatures accelerate decomposition and gas generation, thereby increasing the likelihood of thermal runaway (discussed further in Section ). Additionally, due to its low melting point of 180 °C, Li metal can transition from a solid to a liquid state at elevated temperatures, enabling it to flow freely within the cell. In this liquefied state, Li can infiltrate the porous structure of the separator or bypass it entirely, potentially reaching the cathode. Such uncontrolled migration can trigger direct reactions between Li and cathode materials, leading to thermal runaway and catastrophic cell failure. The implications of this phenomenon are discussed in further detail in Section .

2.2. Cathode

For next-generation cell chemistries, cathodes with higher nominal voltages are highly desirable, as they can boost the overall energy density by increasing the operating voltage, thereby enhancing the specific energy per unit mass of the active material. Recent research has focused on developing cathode materials based on either intercalation or conversion mechanisms. − In intercalation-based cathodes, Li-ions are reversibly inserted into and extracted from interstitial sites within the host structure during charge and discharge, typically without significant alteration of the crystal lattice. However, due to the limited number of available interstitial sites, these materials generally exhibit lower charge storage capacity. In contrast, conversion-type cathodes store charge through redox reactions that transform the active material into a different compound upon discharge and revert it during charging. This process involves the formation and breaking of chemical bonds, leading to substantial structural reorganization, large volumetric changes, and increased mechanical stress. − The following sections examine the structural transformations and phase changes observed in both intercalation and conversion cathodes under elevated temperatures. We also discuss the resulting decomposition products and their chemical reactivity, highlighting their role in triggering or exacerbating thermal runaway.

2.2.1. Layered Oxide-Based Cathode

Oxide cathodes exhibit high specific capacities (480–1700 mAhcm^–3), due to their ability to accommodate a large number of Li-ions and the presence of wide diffusion channels that facilitate efficient ion transport. Based on their crystallographic structures, these cathodes are categorized into layered (R3̅m), spinel (Fd3̅m), and olivine (Pnma) types. Layered and spinel cathodes offer high-rate capability due to their multidimensional diffusion channels, with Li occupying either the tetrahedral (spinel) or octahedral (layered) sites. However, these cathodes suffer from thermodynamic instability, leading to phase transformations and accompanying O₂ release at elevated temperatures. At the atomic scale, phase transformation occurs in two stages. Upon heating, transition metals (TM) such as Ni and Co acquire sufficient energy to overcome the barrier for site exchange, migrating into the octahedral and tetrahedral sites normally occupied by Li, a process referred to as cation mixing (Figure a and b). Elevated temperatures also reduce the energetic penalty for occupying nonideal sites, further promoting TM migration and compromising the structural integrity of the lattice. The structural instability due to increased repulsion between the O of adjacent lattice planes causes a reduction of the TM and cleavage of the M–O bond (Figure a and c). Such a structural change in the cathode induces phase change from layered (R3̅m) to rocksalt (Fd3̅m + Fm3̅m) with subsequent release of O₂ (Figure a and d). − At the particle level, the evolution of the O₂ predominantly originates from the cracked edges of the cathode material. Thermal stress during heating, often caused by increased phase inhomogeneity, leads to microcracking in the cathode particles. These newly formed cracked surfaces exhibit significant lattice disordering, transforming from a layered structure to a rocksalt phase, which triggers further O₂ release (Figure a and b). Differential scanning calorimetry (DSC) profiles further reveal that the thermal stability of layered cathodes is strongly dependent on SOC, with higher SOCs associated with lower decomposition temperatures and increased heat release (Figure c). ,− This reduced thermal stability at elevated SOC is attributed to extensive lattice disorder and crack formation induced by Li-ion depletion, which accelerates O₂ release from the cathode. Furthermore, beyond a critical SOC threshold, the elevated local temperatures during thermal runaway promote the formation of metallic compounds, in contrast to the spinel or rock-salt phases typically observed at lower SOCs.

2.

(a) Schematic illustration of atomic-level phase transformation phenomena in oxide cathodes accompanied by O₂ release. (b) Temperature-dependent Co and Ni occupancy at octahedral Li sites in the cathode material. (Reproduced with permission from ref . Copyright 2021 John Wiley and Sons). (c) Electron energy loss (EEL) spectra showing the reduction behavior of three TM as a function of temperature. (Reproduced with permission from ref . Copyright 2015 American Chemical Society). (d) In-situ X-ray diffraction (XRD) plots depicting phase transformations during heating for various cathode compositions: LiNi_0.5Mn_0.3Co_0.2O₂ (NMC532), LiNi_0.6Mn_0.2Co_0.2O₂ (NMC622), and LiNi_0.8Mn_0.1Co_0.1O₂ (NMC811). (Reproduced with permission from ref . Copyright 2014 American Chemical Society).

3.

(a) In-situ transmission electron microscopy (TEM) images showing crack initiation and propagation in NMC622 upon heating. The inset shows the formation of the rocksalt phase at the crack edge. (Reproduced with permission from ref . Copyright 2018 Springer Nature). (b) Schematic illustrating the particle-level O₂ release mechanism in layered oxide cathodes at elevated temperatures. (c) DSC profile demonstrating the heat flow of LiCoO₂ (LCO) cathodes at different SOCs. (Reproduced with permission from ref . Copyright 2002 Elsevier). (d) MS profiles indicating O₂ release from various cathodes: NMC532, NMC622, NMC811. (Reproduced with permission from ref . Copyright 2014 American Chemical Society). (e) Phase evolution under heating for layered cathodes with varying Ni content: lithium nickel oxide (LiNiO₂, LNO), nickel–cobalt (NC), nickel–manganese (NM), nickel–aluminum (NA), and nickel–magnesium (NMg) compositions. (Adapted with permission from ref . Copyright 2025 Springer Nature). TR stands for thermal runaway. (f) Schematics illustrating the covalent nature of M–O bonds and their influence on the bond strength and thermal stability.

The onset temperature of the phase transformation for layered cathodes is also strongly influenced by the stoichiometric ratio of the respective TMs in the cathode. For instance, mass spectroscopy (MS) analysis indicates that increasing Ni content in oxide-layered cathodes decreases thermal stability and leads to higher O₂ release (Figure d). ,, This decline in stability results from the rapid migration of Ni into Li sites, driven by its smaller ionic size and higher reduction potential compared to other TMs. , Moreover, increasing the nickel content in layered cathodes lowers the SOC at which thermal runaway occurs, primarily due to enhanced Ni oxidation at a given SOC. Interestingly, the thermal stability of the doped Ni-rich cathodes is further influenced by the covalency of the metal–oxygen bond. For example, Co-doping of LiNO₂ accelerates the S1 to S2 phase transformation, shifting the onset of this transition to a lower lithiation state and thereby reducing the critical state of charge for thermal runaway from 210 mAhg^–1 to 190 mAhg^–1. This is due to the weaker strength and delocalized electron shells of the Co–O covalent bond, facilitating a higher O₂ release rate. Conversely, the stronger ionic bonds formed by Aluminum (Al) and Magnesium (Mg) with lattice oxygen suppress O₂ evolution and the S2 phase transition, thus delaying the onset of thermal instability by 20 and 25 °C, respectively, compared to the undoped cathode (Figure e and f). Additionally, dopant concentration plays a critical role in thermal behavior, as lower concentrations have a minimal impact on the onset temperature and phase evolution. Notably, although certain dopants can slow the rate of exothermic reactions, they do not eliminate the risk of thermal runaway. This underscores the need for alternative or complementary thermal mitigation strategies to enhance the safety of Ni-rich cathode systems.

2.2.2. Olivine Phosphate-Based Cathode

Olivine cathodes feature a 3D framework formed by interconnected phosphate (PO₄) tetrahedra. These materials exhibit neither phase transformation nor O₂ release up to 400 °C due to the strong covalent bonding within the tetrahedral units. Above this temperature, minor CO₂ evolution may occur due to the decomposition of carbon additives, accompanied by a structural transition to a trigonal phase. However, this transformation does not involve a compositional change, and the amount of CO₂ released is negligible compared to the O₂ evolution observed in layered oxide cathodes. ,, Nevertheless, some high-voltage olivine cathodes are observed to undergo thermal degradation at temperatures below 350 °C, especially in the delithiated state. , For instance, in the absence of electrolyte, delithiated LiMn_0.8Fe_0.2PO₄ (LMFP) exhibited an exothermic reaction with an onset temperature of 250 °C. This was due to the phase transformation of the delithiated cathode from the olivine structure to sarcopside, releasing hydroxyl (OH) radicals and H₂O, which then react with carbon additives or binders present in the cathode. While the total heat released is lower than that from layered oxide cathodes, the radicals produced can still react violently with Li metal, posing a thermal safety risk in full LMB configurations. In contrast, LiFePO₄ (LFP) maintains high thermal stability regardless of SOC, due to the absence of significant phase transformations. However, this level of stability may not extend to other olivine-based cathodes, particularly those operated at higher cutoff voltages. , Consequently, a comprehensive evaluation of the thermal behavior across various olivine systems remains essential for their safe integration into high-energy LMBs.

2.2.3. Conversion-Type Cathode

Conversion-type cathodes are typically composed of halogens or chalcogens, which are, on average, five times more abundant in Earth’s crust than transition metals. This makes them a cost-effective alternative to conventional oxide-based cathodes. Unlike intercalation-based cathodes that rely on the reversible insertion of Li ions into a stable host structure, conversion-type cathodes operate by breaking and reforming chemical bonds. This mechanism allows for the accommodation of multiple Li ions, typically two to three per halogen or chalcogen atom, resulting in significantly higher theoretical capacities. For instance, oxide- and sulfide-based conversion cathodes can achieve theoretical capacities of up to 1670 and 1675 mAhg^–1, respectively. However, oxide-based conversion cathodes face distinct technical challenges, particularly due to the coexistence of solid and gaseous phases during the conversion reaction, which complicates reaction control and stability. , As a result, this review primarily concentrates on S-based cathodes, which offer higher technological maturity and better compatibility with Li metal anodes, making them a promising option for practical implementation.

2.2.3.1. Sulfur Cathode

S cathodes offer high theoretical capacities, yet their practical implementation has been hindered by significant volumetric deformation during cycling, low intrinsic ionic conductivity, and poor stability with common electrolytes. During discharge, S is electrochemically reduced to intermediate polysulfides (Li₂S _x ), which are soluble in ether-based nonaqueous LEs. These dissolved species diffuse toward the anode, where they are further reduced to lower-order polysulfides (LiPSs) and ultimately to lithium sulfide (Li₂S). Upon reversal of the current, only the lower-order species are soluble and transported back to the cathode. This cyclic transport process, known as the polysulfide shuttle, results in active material loss, capacity fading, and poor Coulombic efficiency (CE) in Li–S batteries. , Recent studies have highlighted that LiPS can also influence the thermal behavior of the cell, as discussed in Section . While numerous strategies, such as interlayers, host materials, and chemical trapping, have been proposed to mitigate the polysulfide shuttle, relatively limited attention has been devoted to the thermal stability of S cathodes. − Emerging evidence suggests that the thermal stability and decomposition products of S cathodes play a critical role in determining the thermal safety of Li–S cells. , In particular, the degradation of S in the presence of electrolyte components has been observed to raise the cell temperature due to increased overpotential. This overpotential rise is often linked to the loss of active material through the precipitation of Li₂S on the anode surface, which inhibits efficient redox cycling. To advance the thermal safety of Li–S batteries, future studies should systematically investigate the influence of operating conditions such as SOC, voltage window, and S cathode modifications (e.g., additives and surface coatings) on the decomposition temperature, reaction products, and associated thermal stability of S cathodes.

2.3. Liquid Electrolyte

LEs are critical components of LMBs, serving as the ionic conducting medium that enables controlled Li-ion transport between the cathode and anode, thereby influencing the overall electrochemical performance. Key properties of the LE, such as diffusion coefficient, viscosity, and salt concentration, directly affect performance metrics, including overpotential, cycle life, and CE. − Additionally, the solvation strength and activation energy of the electrolyte play an important role in shaping the physicochemical characteristics of the interphase, which in turn influence the morphology of Li deposition at the anode. − To mitigate parasitic reactions between Li metal and the LE, various strategies have been developed over the past decade, including the use of electrolyte additives, ionic liquids, and localized high-concentration electrolytes. Despite these advancements, the high chemical reactivity of Li metal with LEs remains a key factor in determining the thermal stability and overall safety of the LMBs. For instance, under abuse scenarios or extreme operating conditions such as nail penetration, high current loads, or elevated ambient temperatures, LEs with low flash points and boiling points can decompose rapidly, releasing heat and generating free radicals. These radicals may further react with Li salts, intensifying heat generation and initiating a self-sustaining exothermic chain reaction (Figure a and b). , The resulting gases can then react violently with Li metal, potentially leading to a catastrophic cell failure. In addition, solvent decomposition byproducts can increase the internal pressure, posing a risk of cell rupture and mechanical failure. At higher temperatures, the accelerated degradation of the LE may also promote the formation of a solvent-derived, nonuniform, organic-rich SEI on the electrode surfaces. Gas evolution during LE breakdown can create voids and pores in the SEI, increasing its porosity. This porous SEI facilitates further parasitic reactions between the electrode and electrolyte and leads to highly uneven ion flux, resulting in nonuniform Li deposition morphologies. Such irregular deposition can act as a nucleation site for dendrite growth, raising the likelihood of localized hotspots and thermal runaway − (discussed further in Section ). In summary, the thermal stability of LEs is a critical determinant of the safety and reliability of LMBs, especially under harsh conditions and high-temperature operation.

4.

(a) DSC profile showing the heat released for a 1:1 vol % mixture of ethylene carbonate (EC) and diethylene carbonate (DEC), in the absence and presence of lithium hexafluorophosphate (LiPF₆) salt. (Reproduced with permission from ref . Copyright 2015 Elsevier). (b) Schematic of a typical electrolyte’s degradation pathway and exothermic reaction between the solvent and salt. (c) Comparison of weight loss observed on heating electrolytes that incorporate different strategies, such as concentrated, dual salt, and radical capture, relative to a standard electrolyte such as LiPF₆-EC/DEC. (d) Schematic depicting the cause for improved thermal stability of the electrolyte with the above-mentioned strategies. (e) Flame test results for the different methods. (Reproduced with permission from refs , − . Copyright 2020 Elsevier, Copyright 2018 Elsevier, Copyright 2019 American Chemical Society, Copyright 2020 John Wiley and Sons, Copyright 2020 Elsevier).

Current research employs a range of techniques, including DSC, self-extinguishing time (SET), flame tests, thermogravimetric analysis (TGA), and MS to characterize the thermal properties of LEs and identify gaseous decomposition products. ,,,− It is well established that thermal degradation of LEs typically begins with the decomposition of solvents into free radicals and gaseous products, such as methane, O₂, and hydrogen, upon heating. These highly reactive intermediates can further react exothermically with the inorganic salts present in the electrolyte, releasing heat and elevating the system temperature, which in turn accelerates further decomposition. Based on these findings, various strategies have been proposed to interrupt this self-sustaining chain reaction and improve the thermal stability of LEs (Figures c, d, and e). One of these approaches to improving thermal stability involves incorporating thermally stable additives that bind to Li salts, thereby preventing their thermal decomposition. Another strategy focuses on interrupting free-radical chain reactions to prevent solvent combustion. Functional groups in phosphate- and fluorine-based additives, such as those found in phosphoric acid or fluorinated phosphates, can effectively trap combustible free radicals generated during solvent decomposition owing to the high electronegativity of phosphorus and fluorine atoms. This suppresses the propagation of exothermic chain reactions. Additionally, solvents with inherently high flash points, such as cyclic carbonates and ionic liquids, are being explored to further mitigate the thermal degradation of LEs. ,− Such solvents typically delay the generation of gases and reactive byproducts associated with the thermal decomposition of electrolytes. However, most existing strategies address only isolated aspects of the thermal stability of LEs, without fully accounting for their broader impact on the overall battery performance. Comparative analysis of previous data sets (Figure a), examining decomposition temperature in relation to viscosity, highlights the complex interplay between thermal stability and electrochemical performance. Although ionic liquids and cyclic carbonates exhibit higher decomposition temperatures, their elevated viscosity often leads to sluggish Li-ion diffusion, which significantly limits the rate capability. Moreover, the incorporation of ionic liquids can result in unstable SEI formation, adversely affecting CE and cycle life (Figure b and c). On the other hand, low viscosity solvents such as ester and ether-based systems enhance electrochemical performance due to their higher Li-ion diffusion coefficients. However, they exhibit decomposition at 120 and 280 °C earlier than carbonate-based and ionic liquid-based electrolytes (Figure a). The generation of reactive free radicals at lower temperatures significantly compromises the overall thermal stability of the LE. A similar trade-off is observed with commonly used thermal additives, where improvements in thermal stability often come at the cost of electrochemical stability (Figure a). Therefore, future research should focus on developing strategies that simultaneously enhance both thermal stability and electrochemical performance.

5.

(a) Comparison of thermal stability and electrochemical performance of commonly studied solvents, presented in terms of decomposition temperature and viscosity (data accessed from refs , − ). (b) Comparison of cyclic stabilities for different electrolytes: Conventional (1 M LiPF₆ in EC/DEC, 1:1 vol %), phosphate-based (1 M lithium bis­(trifluoromethanesulfonyl)­imide (LiTFSI) in triethyl phosphate (TEP)), and ionic liquid-based (1 M LiTFSI in 1,1,2,2-tetrafluoroethyl 2,2,3,3-tetrafluoropropyl ether (TTE) with 1-methyl-1-propyl pyrrolidinium bis­(fluorosulfonyl)­imide, (P₁₃FSI)). (Adapted with permission from refs , , . Copyright 2020 John Wiley and Sons, 2019 American Chemical Society, and 2020 John Wiley and Sons). (c) Flame test results for the above-mentioned electrolytes. (Reproduced with permission from refs , . Copyright 2019 American Chemical Society, Copyright 2020 John Wiley and Sons).

2.4. Solid Electrolyte

In SSBs, the SE serves a dual function, replacing both the LE and the separator by providing a continuous ionic pathway for Li-ion transport between the cathode and anode while physically isolating the electrodes to prevent internal short circuits. As such, a thorough understanding of the thermophysical properties and degradation mechanisms of SEs under elevated temperatures is essential for ensuring thermal safety. This section outlines the degradation temperatures and decomposition products of widely used classes of inorganic SEs: sulfide-, halide-, and oxide-based electrolytes, owing to their favorable ionic conductivity and mechanical robustness (e.g., high shear modulus).

2.4.1. Oxide-Based Solid Electrolyte

Oxide SEs have been extensively studied due to their advantageous characteristics, including high ionic conductivity, a wide electrochemical stability window, air stability, and strong mechanical strength. These properties enable the use of Li metal as an anode without significant concerns about unstable interphase formation or dendritic Li growth at room temperature. Furthermore, oxide SEs exhibit thermal degradation only at elevated temperatures (above 800 °C), decomposing into stable byproducts (Figure a, b, and c). ,− However, the stability of the decomposition products can vary depending on the temperature and electrolyte chemistry. Despite this, the superior electrochemical performance and thermal stability of oxide SEs are largely attributed to the strong M–O bonds within the electrolyte lattice. The high bond dissociation energy of M–O bonds imparts greater resistance to oxidation and reduction reactions at both the cathode and anode interfaces. Additionally, the localized electron distribution within the oxygen functional groups promotes high ionic conductivity while maintaining electrical insulation. ,, Despite these advantages, oxide SEs can release significant amounts of heat upon decomposition due to their high formation enthalpy. The resulting byproducts, such as metal oxides, phosphates, and in some cases, O₂, CO₂ (arising from lattice oxygen instability or impurities), may react with Li metal at elevated temperatures, increasing the risk of thermal runaway at the interphase or cell level (discussed further in Section ). To mitigate the reactivity of these decomposition products, some studies have investigated the incorporation of binding elements such as niobium (Nb), Al, and lanthanum (La). , While effective in reducing reactivity, these elements can induce lattice defects, which could undermine the structural integrity of oxide electrolytes, making them prone to mechanical fractures and Li dendrite penetration. Prolonged air exposure also results in the formation of a thin layer of carbonates or hydroxides, which, upon heating, release gaseous byproducts such as CO₂ and H₂O. The local pressure differentials generated between the carbonate surface layer and the bulk electrolyte from the gas release can create mechanical stress, resulting in the pulverization of the electrolyte pellet (Figure d and e). In addition to these concerns, oxide SEs face significant manufacturing challenges as their fabrication is both energy-intensive and costly, often requiring extremely high temperatures and pressures to produce a dense, defect-free structure. Their inherent brittleness presents challenges in achieving conformal contact with the active electrode material and in maintaining a stable, intimate interface between the electrolyte and electrode throughout operation. The importance of forming and maintaining a stable, intimate interface for improved thermal stability is discussed in detail in Section . Collectively, these factors pose significant challenges to the practical implementation of oxide electrolytes in commercial SSBs.

6.

(a) In-situ XRD illustrating the phase evolution of the Li_1.5Al_0.5Ge_1.5(PO₄)₃ (LAGP) electrolyte during heating. The inscribed line tracks the evolution of the main crystalline phase. (b) Rietveld refinement quantifying impurity phase fractions formed upon heating LAGP. (Reproduced with permission from ref . Copyright 2016 Royal Society of Chemistry). (c) Phase evolution of Li₇La₃Zr₂O₁₂ (LLZO) electrolyte and the development of impurity phases at elevated temperature. (Adapted with permission from ref . Copyright 2022 Elsevier). (d) SEM (scanning electron microscopy) cross-section of an air-exposed LLZO pellet obtained via dual-beam focused ion beam/SEM (FIB/SEM). (e) Raman spectra of air-exposed lithium carbonate (Li₂CO₃) coated LLZO over different exposure durations after sintering. (Reproduced with permission from ref . Copyright 2018 Elsevier).

2.4.2. Sulfide-Based Solid Electrolyte

Among inorganic SEs, sulfide electrolytes exhibit the highest ionic conductivity (10^–4 to 10^–3 S cm^–1) and provide excellent interfacial contact, making them strong candidates to replace LEs. This high conductivity stems from the ionic nature of the Li–S bond, which facilitates the free movement of Li-ions within the sulfide lattice. Yet, their broader application is hindered by poor air stability and a narrow electrochemical stability window. At present, various methodologies have been developed to analyze the thermal behavior of sulfide SEs and their decomposition products. − In-situ SEM studies have revealed that thermal decomposition in sulfide electrolytes induces phase transformations, resulting in recrystallized phases with ionic conductivity significantly lower than that of the original bulk electrolyte. Therefore, the formation of decomposition products not only increases cell overpotential but also contributes to an increase in internal temperature. The decomposition temperature is strongly influenced by the electrolyte’s structure, with crystalline sulfide SEs exhibiting significantly higher thermal stability than their glassy counterparts. The higher thermal stability of crystalline sulfide SEs is attributed to their ordered, highly coordinated atomic structure and uniform, strong bonding, which together result in a higher energy barrier for thermal decomposition compared with glassy SEs. Additionally, air exposure significantly lowers the decomposition temperature of sulfide SEs and leads to the release of hazardous hydrogen sulfide (H₂S) gas, posing serious safety concerns (Figure a and b). , The quantity of H₂S released increases with prolonged air exposure, which is attributed to the breakage of the P-S bond in the PS₄ ^3– tetrahedral units. This degradation process involves the substitution of S atoms with oxygen from ambient moisture, which compromises the structural integrity of the electrolyte (Figure c). As a result, the thermal stability of sulfide SEs is diminished, while additional safety risks arise from the formation of reactive hydrated phases and volatile byproducts. Some studies have reported that the incorporation of TM significantly enhances both air and thermal stability, attributed to the higher bond energy of the TM-S bond, which strengthens the structural framework against decomposition. − The stronger TM-S bond (compared to the S–S bond) enhances lattice crystallinity and thermodynamic stability, making the material less prone to oxidation and thermal degradation upon air exposure. Nevertheless, TM-doped sulfide SEs often form interphase products with high chemical affinity toward Li metal, which can introduce interfacial stability issues, as further discussed in Section .

7.

(a) DSC heat flow curves comparing pristine and air-exposed Li₆PS₅Cl (LPSCl). (Reproduced with permission from ref . Copyright 2022 Royal Society of Chemistry). (b) Quantification of H₂S generation with time of exposure for different sulfide electrolytes: Li₃PS₄ (LPS), Li_9.54Si_1.74P_1.44S_11.7Cl_0.3 (LSPSCl), and Li₄SnS₄ (LSS). (Reproduced with permission from ref . Copyright 2021 John Wiley and Sons). (c) Schematic illustrating the chemical reactivity and degradation pathway of sulfide electrolytes in the presence of moisture. (d) Comparison of various SEs with respect to decomposition temperature and ionic conductivity relative to LEs (data accessed from refs , − ).

2.4.3. Halide-Based Solid Electrolyte

Halide-based SEs have garnered significant attention in recent years due to their favorable air stability and wide electrochemical window. They offer mechanical properties and ionic conductivities comparable to those of sulfide SEs, facilitating conformal interfacial contact and efficient Li-ion transport. Moreover, many halide SEs exhibit excellent thermal stability with decomposition temperatures exceeding 400 °C (Figure d). Their high oxidative stability renders them promising candidates for pairing with high-voltage cathodes in next-generation Li metal batteries. , However, halide SEs face challenges related to the phase stability at elevated temperatures. Each polymorphic phase possesses distinct ionic conductivities, and transitions between them can significantly reduce the overall conductivity of the SE. Additionally, certain halide SEs are prone to hydrolysis under humid conditions, releasing corrosive byproducts such as hydrogen chloride (HCl), particularly in chloride-based systems, at high temperatures. , The air stability of halide SEs is noted to be heavily influenced by the size of the metal cation and the relative density of the fabricated SE pellet. , For instance, Li₃InCl₆ (LIC) has a faster water absorption rate compared to Li₃YCl₆ (LYC) owing to the higher polarization of the smaller In³⁺ cation. Furthermore, increasing the relative density of the SE reduces its deliquescent nature, likely due to a decreased contact area between the halide surface and ambient moisture. Such degradation not only compromises the intrinsic thermal stability of the halide SEs but also introduces reactive decomposition products that can interact adversely with other cell components, ultimately affecting the overall cell safety. These issues are particularly critical for large-scale manufacturing, where environmental moisture levels, although controlled, may exceed glovebox standards. Current mitigation strategies include coating the halide particles with a passive layer (such as aluminum oxide (Al₂O₃)), thereby preventing their direct interaction with moist air or doping with metallic cations (such as Zn²⁺ or In³⁺) that form hydrated complexes. However, the introduction of such layers or metal elements can create intermediate decomposition products, warranting meticulous investigation of their reactivity with other battery constituents. Overall, as illustrated in Figure d, the nonlinear correlation between electrochemical and thermal stability metrics, such as ionic conductivity and thermal decomposition temperature, increases the challenge of choosing an SE that not only enables good electrochemical performance but also enhances the thermal safety of the cell.

3. Interface- and Interphase-Level Thermal Stability

As previously discussed, material-level thermal stability provides valuable insights for the preliminary screening of battery components and serves as a foundation for identifying materials with favorable thermal properties. However, thermal behavior at the cell level is strongly influenced by interfacial phenomena, particularly the formation of the SEI, which arises spontaneously at the electrode–electrolyte interface due to their inherent chemical affinity. The SEI plays a vital role in enabling selective Li-ion transport while inhibiting electron transfer between the reactive components. Notably, the SEI often possesses distinct chemical and physical properties compared with the original electrode or electrolyte materials, and its integrity directly impacts both electrochemical performance and thermal behavior. Recent studies have underscored the pivotal role of the SEI in governing thermal stability, revealing that thermal runaway, whether in the presence of LE or SEs, often initiates at temperatures below the decomposition thresholds of the individual cell components. This behavior is frequently triggered by premature degradation of the SEI, which sets off a cascade of exothermic reactions. Accordingly, this section focuses on the thermal stability of interphases and the interfacial mechanisms that contribute to thermal instability in LMBs.

3.1. Lithium Metal Battery with Liquid Electrolyte

While the electrochemical stability of interphases formed at the cathode and anode is critical to the performance of LE-based LMBs, understanding their thermal characteristics is equally important. The thermal instability of LMBs using LEs has been attributed to the high parasitic reactivity between Li metal and the electrolyte, which elevates the cell temperature and accelerates degradation under thermal abuse conditions. The interface formed at the cathode has been observed to not only pulverize active materials but also undergo exothermic reactions with the LE, thereby intensifying heat generation during thermal runaway. The following section examines potential exothermic reactions occurring at the cathode and anode/electrolyte interface, elucidating the effect of the physical and chemical properties of each of the SEIs on the thermal response of the cell.

3.1.1. Cathode/Electrolyte Interface

High-voltage cathodes, when in contact with an LE, form an interphase termed the catholyte interphase (CEI). This interphase is critical, as it impacts the stability and performance of the cell. The structure and chemical composition of the CEI are highly dependent on the reduction potential of the LE, as this dictates which electrolyte components are reduced to form the passivation layer on the cathode surface. Although the role of the interphase is to suppress further reaction between the electrode and electrolyte, recent studies have revealed that it becomes inherently unstable at elevated temperatures. Under these conditions, the interphase can decompose, releasing heat and triggering a cascade of exothermic reactions. , As the temperature increases, layered oxide cathodes release O₂, which can react exothermically with the organic components of the LE or with decomposition products of the CEI, such as Li alkoxides and other reactive intermediates. These reactions produce additional heat, initiating a feedback loop that rapidly elevates the cell temperature. As a result, various techniques are being employed to investigate the thermal behavior of the CEI under elevated temperature conditions. ,, Calorimetry experiments have shown that O₂-releasing oxide cathodes follow two distinct reaction pathways, depending on the amount of electrolyte present. ,− At elevated temperatures, when the LE content is limited relative to cathode loading, a dominant exothermic reaction is typically observed near 200 °C. This reaction is attributed to interactions between the LE, or its decomposition products (e.g., organic radicals, Li alkoxides, PF₅), and O₂ released during the cathode’s phase transformation from layered to spinel and rocksalt structures. In contrast, when excess LE is present, an additional exothermic peak appears at higher temperatures, indicating further reactions between the surplus electrolyte and decomposition products or reactive cathode surfaces (above 320 °C). This additional heat release is driven by the oxidation of remaining solvents following the decomposition of TM oxides. Post-mortem XRD analysis of cathodes exposed to excess LE reveals the presence of metallic Co. This extensive degradation promotes additional exothermic reactions with residual organic species from the electrolyte, thereby contributing to increased heat release (Figures a and b). Although the magnitude of exothermic peaks varies with the electrolyte composition, the underlying thermal reaction pathways remain consistent and are influenced by the ratio of LE quantity to cathode loading. The cathode chemistry and lattice structure also play critical roles in determining the reaction pathways. Upon exposure to the electrolyte, the cathode surface undergoes a phase transformation with the resulting surface phase determined by the underlying lattice structure of the cathode. For example, NMC cathodes form a rocksalt phase, which is prone to O₂ release during the decomposition, whereas LFP cathodes develop an iron-rich amorphous layer that does not release O₂, contributing to their inherent thermal stability (Figure c and e). , At elevated temperatures, the exothermic reaction between the cathode and LE in LFP systems occurs at approximately 275 °C and releases 251 J g^–1 of heat, which is 31% and 68% lower than that of high-Ni cathodes. This reduced heat output arises from the limited formation of oxygen reactive species during LFP thermal decomposition (Figure d and f). Instead, the observed exothermic peak may result from reactions between the LE and decomposition products, such as CO₂ and hydrogen fluoride (HF), which originate from the binder rather than the cathode itself. Although these gaseous species are reactive, they produce a less intense exothermic response compared to oxygen-driven reactions in other cathode chemistries, due to their lower reaction enthalpy.

8.

(a) Schematic illustration of degradation pathways and exothermic reactions exhibited by phase-transforming cathodes in the presence of LEs during heating. (b) DSC profiles comparing the reaction of LCO cathodes with limited and excess electrolyte. (Reproduced with permission from ref . Copyright 2002 Electrochemical Society). High-resolution transmission electron microscopy (HRTEM) images showing lattice structure changes in NMC and LFP cathodes before and after electrolyte exposure: (c) rocksalt phase in NMC. (Reproduced with permission from ref . Copyright 2014 Springer Nature) and (e) Fe-rich amorphous phase in LFP. (Reproduced with permission from ref . Copyright 2018 John Wiley and Sons). DSC profiles of different cathode materials: (d) Li_0.1Ni_0.8Co_0.15Al_0.05O₂ (NCA) and (f) LFP in the presence of electrolyte LiPF₆–EC/DMC. (Reproduced with permission from ref . Copyright 2016 American Chemical Society).

In addition, several factors, including SOC, cutoff voltage, and cycle number, significantly influence the thermal stability of the CEI. ,, Cathodes subjected to extensive charge–discharge cycling often exhibit reduced thermal stability, undergoing phase transformations at temperatures lower than those of their pristine counterparts. This decrease in stability is largely attributed to the accumulation of mechanical strain during cycling. The volumetric expansion and contraction within cathode particles induce the formation and propagation of microcracks, which widen and deepen as cycling progresses, increasing the risk of mechanical failure. The presence of these cracks exposes fresh, reactive surfaces to LE that are otherwise protected in a pristine state. This not only increases the cell overpotential but also makes the cathode susceptible to phase changes (from a layered to a disordered rocksalt or spinel phase), destabilizing the lattice structure and leading to the release of O₂ 40 °C lower than the pristine case (Figure a and b). The early release of O₂ at lower temperatures further exacerbates the situation, as it readily reacts with the LE, resulting in exothermic reactions that generate additional heat and accelerate the temperature rise (Figures a and b). Thus, cycled cathodes with compromised structural integrity present a higher risk of thermal instability than their pristine counterparts, particularly under thermal stress. Similarly, increasing the SOC of a cathode has been shown to significantly reduce its thermal stability, leading to a lower onset temperature and increased heat release (Figure c). ,, This reduced thermal stability is attributed to compositional and structural changes in the cathode, as a greater amount of Li is extracted during charging. At higher SOC, progressive delithiation occurs near the surface, forming a Li-depleted region that differs in composition from that of the bulk material. This surface region exhibits greater lattice instability, increasing the susceptibility to irreversible structural transformations, culminating in higher O₂ release and temperature rise. Specifically, the surface regions undergo a phase transformation from spinel to rocksalt at lower temperatures compared to the bulk material due to the lack of Li-ions that would maintain the integrity of the crystal structure, triggering an early release of O₂. The released O₂ subsequently reacts exothermically with the LE, generating substantial heat (Figure d).

9.

(a) Cross-sectional SEM images of fully delithiated NCA cathode (100% SOC) after 1st cycle and 100th cycle. (b) In-situ XRD profiles depicting the phase transformation temperature for the delithiated cathode after the 1st and 100th cycle. (Reproduced with permission from ref . Copyright 2023 John Wiley and Sons). (c) DSC profiles of Li_1–x Ni_0.8Mn_0.1Co_0.1O₂ (NMC811) at different SOCs with electrolyte (1.5 M LiPF₆ - EC). (Reproduced with permission from ref . Copyright 2020 American Chemical Society). (d) Schematic illustrating the correlation between the cathode SOC and both the onset temperature and magnitude of heat release associated with the cathode–electrolyte reaction.

3.1.2. Anode/Electrolyte Interface

The high reactivity of Li metal, driven by its exceptionally low electrochemical potential, causes it to react immediately upon contact with the LE. This reaction leads to the formation of an SEI layer on the anode surface, which is crucial for the long-term stability and performance of LMBs, as it passivates the surface by preventing further parasitic reactions between Li metal and the electrolyte. Typically, SEI is a heterogeneous mixture of organic and inorganic compounds, such as Li alkyl carbonates, lithium fluoride (LiF), and lithium oxide (Li₂O). Its structure, composition, and stability are strongly influenced by the reduction potential and chemical makeup of the electrolyte. In addition to affecting electrochemical performance metrics such as cycle life, CE, and rate capability, the composition and morphology of the SEI on the anode also play a critical role in determining the thermal stability of the cell. − It was reported that the thermal behavior of SEI on Li metal varied significantly when it was paired with ether-based and carbonate-based electrolytes. , In the presence of an ether-based electrolyte, the onset temperature for exothermic reactions is higher, and the total heat release is lower compared with a carbonate-based electrolyte, indicating the superior thermal stability of the ether-Li system, making it less prone to rapid temperature increases and thermal runaway. This enhanced stability is primarily attributed to the composition of the SEI that forms on the Li surface in the presence of ether electrolytes. Ether electrolytes form an inorganically rich interphase consisting of compounds like LiF, Li₂O, and other stable Li salts. These inorganic components reduce the rate of parasitic reactions between the Li metal and the electrolyte, subsequently contributing to the lower self-heating rate of the system and the delayed onset of exothermic reactions. In contrast, carbonate-based electrolytes form an SEI that is organically rich, comprising products such as lithium alkyl carbonates (ROCO₂Li). Such interphases undergo thermal decomposition at temperatures that are 112 °C lower than those of their ether-based counterparts (Figure a and b). This exposes fresh Li surfaces, enabling additional parasitic reactions with the electrolyte, which in turn increases the self-heating rate and total heat release. The structure of SEI is also noted to have a significant impact on the thermal stability of LMBs. ,,− The porous interphase formed in the presence of carbonate-based electrolytes promotes uneven current distribution during charging, thereby facilitating the growth of Li dendrites, thin, needle-like structures (Figure c). Dendritic Li has a high surface area compared to compact deposition morphology, which greatly increases its reactivity with the LE. The resulting increase in parasitic reactions elevates the internal cell temperature, accelerating the decomposition of both the SEI and the LE, and leading to greater overall heat release during thermal events (Figure d). The SEI formed in the presence of a dual-salt LE promotes homogeneous ion flux during Li plating, facilitating compact and uniform Li deposition (Figure c), with a lower surface area compared to dendritic growth. With fewer exposed surfaces, the rate of parasitic reactions between Li metal and the electrolyte is significantly reduced, leading to a lower self-heating rate and shifting of the onset temperature of the primary exothermic reaction to 264 °C, thereby reducing the risk of thermal runaway (Figure e).

10.

(a) DSC profiles comparing the heat release for two electrolytes: 1 M LiPF₆–EC/DMC (1:1 vol %) and 1 M LiTFSI-1,3-dioxolane/1,2-dimethoxyethane (DOL/DME) (1:1 vol %). (Reproduced with permission from ref . Copyright 2022 Elsevier). (b) Schematics illustrating the influence of SEI formed at the anode on Li deposition morphology, the intensity of parasitic reactions with the LE, and the resulting magnitude of heat released. (c) Cross-sectional SEM images representing the interphase structure on Li metal in electrolytes: 1 M LiPF₆-EC/DMC, 0.8 M LiTFSI-0.2 M lithium difluoro­(oxalato)­borate (LiDFOB)-0.01 M LiPF₆ in EC/DMC (Dual salt). (Reproduced with permission from ref . Copyright 2019 Elsevier). DSC profiles representing the heat released by the Li metal in different electrolytes: (d) LiPF₆-EC/DEC and (e) LiTFSI/LiPF₆-DMC/FEC. (Reproduced with permission from ref . Copyright 2023 John Wiley and Sons).

Studies have highlighted that the thermal stability of the anode interphase in LMBs degrades significantly with increasing cycle numbers. ,,, It was reported that as the number of charge–discharge cycles increases, the morphology of Li deposition on the anode surface shifts from a dense, compact structure to a porous, dendritic form (Figure a). The change in Li deposition morphology can be attributed to the mechanically induced disruption of the SEI caused by the repeated expansion and contraction of Li metal during plating and stripping. The large surface area of dendrites exposes numerous sites for parasitic reactions with the LE, leading to accelerated electrolyte decomposition, substantial heat generation, and a rapid local temperature rise near the anode. This chain of coupled thermal and chemical processes contributes to an earlier onset of thermal runaway (Figure b, c, and d). A similar decline in thermal stability is observed when the current density is increased, leading to a lower onset temperature for thermal decomposition. ,, This is attributed to the significant impact of the current density on the morphology of Li deposition. At higher current densities, a larger concentration gradient is generated near the anode surface, creating dendritic or filament deposition of Li. This results in the formation of dendritic/filament structures on the anode surface rather than uniform, compact layers. The higher surface area of Li dendrites increases parasitic reactions with the LE, raising the self-heating rate and elevating the internal battery temperature by approximately 10 °C compared with compact Li deposits, thereby accelerating thermal degradation. To mitigate excessive heat generation and prevent catastrophic failure in LMBs, it is essential to engineer robust anode interphases that not only suppress parasitic reactions but also promote dense Li deposition and accommodate electrochemical stresses arising during high-rate cycling. Although the exothermic anode-electrolyte reactions typically exhibit lower onset temperatures, the cathode–electrolyte reactions release nearly twice as much heat (Figure c and d), reflecting their higher exothermic propensity. This underscores the importance of not only stabilizing the SEI but also preventing interactions between O₂ evolved from the cathode and the LE interactions that can accelerate the cell temperature rise. In full-cell configurations, the contribution from the additional exothermic reactions and degradation pathways may exceed the heat generation from the cathode–electrolyte interface, thereby compounding the risk of thermal runaway (as discussed in Section 4.2).

11.

(a) SEM images representing the morphology of the deposited Li on the anode as a function of the number of charge/discharge cycles. (Reproduced with permission from ref . Copyright 2019 Electrochemical Society). Heat flow profiles from DRC for (b) pristine Li, (c) cycled Li in electrolyte. (Reproduced with permission from ref . Copyright 2022 John and Wiley and Sons). (d) Comparison of heat release and onset temperature of the exothermic reaction of pristine and cycled Li in the presence of LE.

3.2. Solid-State Lithium Metal Battery

Although SEs exhibit high intrinsic thermal stability due to their elevated decomposition temperatures, this stability is significantly compromised under external heating when in contact with either the cathode or the anode. Their thermal decomposition can be influenced by the thermodynamic instabilities of the interphases formed at the cathode and anode. The onset and severity of thermal runaway are dependent on the morphology and chemical properties of these interphases, which are further dependent on the chemical properties of the active materials. The following section examines potential exothermic reactions occurring at the electrode–electrolyte interface for oxide- and sulfide-based SEs, highlighting their distinct modes of thermal instability.

3.2.1. Anode/Electrolyte Interface

Oxide electrolytes, known for their high ionic conductivity and wide electrochemical stability window, form a stable interphase with Li metal at room temperature, effectively suppressing continuous side reactions between the Li metal and the electrolyte. The SEI formed is less reactive than those formed with LEs, leading to improved performance at moderate temperatures. However, the stability of this interphase degrades significantly at elevated temperatures (above 250 °C). As the temperature increases, thermal stress can weaken the structural integrity and result in thermal degradation of the SEI. Moreover, the interphase formed by many oxide-based SEs begins to degrade at temperatures above 250 °C, leading to the formation of metal oxide and phosphate byproducts along with the release of O₂ as part of the decomposition process. This O₂ release poses a significant threat to the thermal stability of the cell, as it creates a highly reactive environment near the Li metal, which readily undergoes intense exothermic reactions with O₂. These reactions generate substantial heat, further escalating the risk of thermal runaway. These reactions can trigger further decomposition of both the oxide electrolyte and SEI, causing a cascading thermal event. The rate of O₂ release and the onset temperature of these thermal decomposition reactions are dependent on the specific composition and structure of the oxide electrolyte, as well as the nature of the SEI that forms at the Li interface (Figure a). , The thermal stability of SEs such as LLZO and Li_1.3Al_0.3Ti_1.7(PO₄)₃ (LATP) when paired with Li metal varies significantly due to differences in their structural properties and interfacial behavior. The interface between LLZO and Li does not readily decompose or release O_2, with no exothermic peaks observed until around 480 °C. This high thermal stability of the LLZO interface is attributed to the electrolyte’s robust crystal structure, which maintains its integrity under thermal stress. In contrast, LATP, which has a NASICON-type structure, demonstrates lower thermal stability when in contact with Li metal. At temperatures higher than 290 °C, LATP undergoes structural degradation, leading to the release of O₂ and other decomposition products. This release triggers exothermic reactions between the Li metal and liberated O₂, resulting in rapid heat generation and subsequent thermal runaway (Figure b and d). Interestingly, LAGP shows disparate thermal characteristics under external abuse conditions compared to other oxide electrolytes owing to its unique interface chemistry against Li metal. When paired with Li metal, LAGP forms an interphase with a high electronic conductivity of ∼10^–7 S cm^–1, which is roughly 100 times greater than the typical electronically limiting SEI (∼10^–9–10^–8 S cm^–1). This increased conductivity arises from the formation of lithium–germanium (Li–Ge) alloy compounds at the interface. This undermines the interface’s insulating properties and allows for Li metal deposition within the SEI. , The electrically conductive interphase of the LAGP is prone to localized heating and undergoes uncontrolled electrochemical reactions above 320 °C, rapidly escalating into thermal runaway (Figure c and d). Studies have reported that the interfacial growth between Li metal and SEs like LAGP can induce significant mechanical stresses, particularly during cycling. These stresses often localize at the interface, leading to the formation of microcracks that propagate along grain boundaries and compromise the structural integrity of the SE. Owing to the reduced structural rigidity, the SE becomes susceptible to dendritic Li penetration, which can eventually reach the cathode interface, resulting in internal short-circuiting and subsequent cell failure. −

12.

(a) Schematic illustration comparing the influence of electrical properties (electronic and ionic conductivity) of the anode/electrolyte interface on the thermal degradation pathway for oxide SEs. The SEI formed by oxide electrolytes releases O₂ on thermal degradation. The released O₂ undergoes an exothermic reaction with Li metal. The mixed conductive interphase releases higher amounts of O_2, leading to larger heat generation. (b) In-situ SEM images depicting the thermal degradation of LATP interphase with Li during heating. (Reproduced with permission from ref . Copyright 2021 Royal Society of Chemistry). (c) Temporal images representing the exothermic reaction of LAGP/Li interphase during the heating test under an inert atmosphere. (Reproduced with permission from ref . Copyright 2020 Elsevier). ARC profiles of thermal runaway observed in (d) LATP/Li and (e) LAGP/Li samples. (Reproduced with permission from ref . Copyright 2020 Elsevier).

Due to their limited electrochemical stability window, sulfide electrolytes undergo immediate reduction upon contact with Li metal, leading to the formation of a thermodynamically unstable SEI on contact with Li metal. This SEI is typically composed of various decomposition products, including Li₂S and other Li-rich compounds, which can be highly reactive, especially at elevated temperatures. Upon heating, these decomposition products can undergo further reactions with metallic Li, leading to exothermic reactions that can release a significant amount of heat, potentially causing thermal runaway. The onset temperature and total heat release of these reactions are strongly influenced by the crystallite structure and morphology of the sulfide SE. Amorphous structures, due to their lower bond dissociation energies, generally exhibit lower onset temperatures for thermal decomposition, whereas well-ordered crystalline structures tend to demonstrate higher thermal stability. , The SEI formed by glassy (amorphous) sulfides like LPS is noted to be less stable and exhibits decomposition around 170 °C, leading to the release of gaseous species, such as sulfur gas (S₂) upon heating (Figure a, b, and c). The liberated S₂ then reacts rapidly with the Li metal surface, resulting in an intense exothermic reaction that leads to a large and sudden heat release. This is referred to as an “indirect” mechanism, as the gas generated from the SE can diffuse across the Li surface and initiate the reaction, leading to thermal runaway without requiring direct physical contact between the two components. In contrast, crystalline sulfide electrolytes like LPSCl exhibit a slower self-heating rate, with the onset of an exothermic reaction occurring above 275 °C due to their higher structural stability. In this case, the reaction between the electrolyte and Li metal is initiated only upon direct physical contact, termed a direct mechanism, resulting in localized sites where exothermic reactions originate. This localized reaction front allows gradual propagation and controlled heat release, providing better thermal stability than the rapid, widespread reactions in glassy sulfides (Figure d, e, and f). Therefore, the difference in reaction dynamics between glassy and crystalline SEs is critical to understanding their relative safety profiles in LMBs. Studies have highlighted the influence of the electrochemical nature of the SEI on the thermal stability of sulfide SE batteries. ,,, For instance, SEs like LSPS and Li₁₀GeP₂S₁₂ (LGPS) are known to form unstable interphases at the anode during cycling due to the formation of metal alloys such as Sn–Li or Li–Ge at the interface, which impart an electrically conductive nature. Such an SEI, along with compromising the electrochemical performance, initiates severe thermal runaway at lower temperatures due to the uncontrolled growth of the metal alloy phases, potentially resulting in catastrophic failure (Figure a, b, and c). Notably, a recent study revealed that increasing the operating pressure delayed the onset of thermal runaway by suppressing pressure-sensitive reduction reactions, thereby slowing interphase evolution at the Li-SE interface. Nevertheless, thermal runaway remained unavoidable, with the first exothermic event from interphase degradation occurring near 170 °C even under a 20 MPa pressure. As discussed in Section , this behavior contrasts with the material-level trends, where the incorporation of TMs has been shown to enhance thermal stability.

13.

(a), (d) Schematics illustrating the influence of physicochemical properties of sulfide electrolytes on their decomposition pathways and heat release during thermal runaway. (b) Heating test results showing the violent reaction of LPS with Li metal at a relatively low temperature of 300 °C. (Reproduced with permission from ref . Copyright 2023 John Wiley and Sons). (c) ARC profiles for pristine and cycled LPS/Li showing no significant difference in onset temperature. (Reproduced with permission from ref . Copyright 2023 American Chemical Society). (e) Heating test result of the LPSCl electrolyte against Li, indicating that thermal runaway occurs above 400 °C. (Reproduced with permission from ref . Copyright 2023 John Wiley and Sons). (f) ARC profiles of pristine and cycled Li/LPSCl interfaces, demonstrating similar thermal stability. (Reproduced with permission from ref . Copyright 2023 American Chemical Society).

14.

(a) Schematic illustration depicting the thermal runaway pathway for sulfide SEs with a mixed interphase (both electrically and ionically conductive). (b) Operando tomography results of LPSCl/Li interphase before (top) and after (below) cycling. (Reproduced with permission from ref . Copyright 2021 Springer Nature). (c) ARC results depicting the lower thermal stability of cycled LSPS/Li interphase relative to its pristine counterpart. (Reproduced with permission from ref . Copyright 2023 American Chemical Society). (d) Comparison of the thermal and electrochemical stabilities of different SEs (oxide, sulfide, and polymer) against Li metal. LE is presented for reference. The bars indicate the electrochemical window for representative SEs (data accessed from refs , , , , , ).

Halide-based SEs like sulfides form unstable interphases against the Li metal. For example, unlike LPSCl that forms an electrically nonconductive interface (composed of Li₂S, Li₃P, and phosphorus compounds), Li₃YCl₆ against Li metal undergoes continuous decomposition due to the formation of a mixed ionic and electronic conductive interphase. , This is attributed to the high decomposition energy of halide SEs (200 meV per atom) and the chemical affinity of high valence metal cations to undergo reduction when in contact with Li metal. Given the high reactivity of halide SEs with Li metal, current strategies focus on either doping with highly electronegative halogen atoms (e.g., fluorine) to inhibit the formation of metallic phases and suppress electronic conductivity, or employing a bilayer configuration wherein an SE such as LPSCl serves as an anolyte to prevent direct contact with Li metal and facilitate the formation of a kinetically stable interphase. , However, it is essential to investigate how such strategies influence thermal runaway, as the kinetic parameters and activation energies of decomposition products formed at the interface are temperature-dependent and may significantly affect the overall thermal stability.

To evaluate the stability of different SEs against Li metal, their electrochemical stability windows were plotted against their thermal onset temperatures (Figure d). Interestingly, the data reveal no direct correlation between the electrochemical stability and thermal stability of the SEI. For instance, an SE with a wide electrochemical stability window may still form an unstable SEI that decomposes exothermically at relatively low temperatures. Among the various SEs studied, sulfide-based SEs, despite their high ionic conductivity, possess the lowest chemical stability, with an average reduction potential of 1.98 V, indicating an increased reactivity with Li metal. In contrast, oxide-based SEs exhibit a lower mean reduction potential of 1.35 V and the highest thermal stability against Li. On the other hand, polymer-based SEs exhibit high electrochemical stability due to their ability to form an effective passivating SEI with Li metal. However, their thermal stability is comparatively limited as the mechanical integrity of polymers tends to degrade at elevated temperatures relative to inorganic SEs. These complex interdependencies underscore the challenges in designing SEs that are both electrochemically stable and thermally robust.

3.2.2. Cathode/Electrolyte Interface

The wide electrochemical stability window of oxide SEs allows them to form a stable SEI even when they are paired with high-voltage cathodes. This characteristic property of oxide SEs ensures that the interphase layer formed is less prone to breakdown at elevated voltages, contributing to the overall electrochemical stability and safety of the battery. This is crucial for improving energy density, as the battery can operate across a broader voltage range without risking electrolyte decomposition or uncontrolled side reactions. Additionally, upon heating, the oxidative stability of the SEI formed further contributes to safety by preventing direct reactions with O₂ released from the cathode material (such as layered oxides). This property makes oxide electrolytes particularly suitable for analyzing thermal signatures in full battery cells as it isolates the effects of cathode–electrolyte interactions from other thermal reactions within the cell. Though interphases formed by oxide SEs have a high initial thermal stability, they might not be immune to thermal degradation. The reaction kinetics governing interfacial reactions are dictated by the Arrhenius equation, which describes the exponential dependence of the reaction rate on temperature. This implies that even a seemingly stable material can undergo accelerated degradation at higher temperatures due to the increased kinetic energy of particles, leading to the breakdown of the interphase structure and potential side reactions. ,− However, the onset of such exothermic reactions occurs at temperatures above 500 °C, which is higher than those used in conventional thermal stability assessments of batteries. − This suggests that the apparent thermal stability of oxide-based batteries may underestimate the onset of critical degradation mechanisms under extreme conditions, such as those encountered during thermal runaway or abuse conditions. Sulfide electrolytes, due to their narrower electrochemical stability window with oxidation potential of 2.27 V, are prone to oxidative decomposition when paired with high-voltage cathodes, even under nominal operating conditions (e.g., low C-rate and room temperature). This low oxidative resistance causes the breakdown of sulfide SEs, which then react with O₂ released from Ni-rich cathodes, leading to the release of gaseous byproducts such as sulfur dioxide (SO₂) under electrochemical cycling. Further, during heating release of SO₂ is observed due to the reaction of O₂ released from the cathode with the thermodynamically unstable interface formed by sulfide SEs. , SO₂ has a considerably low threshold limit value (TLV). Thus, even small amounts of SO₂ in a confined battery environment can pose significant safety risks. Further, the buildup of SO₂ during cycling can increase internal pressure, potentially leading to swelling, leakage, or even thermal runaway, escalating safety risks. Meanwhile, the exothermic nature of interphase decomposition increases the heat generation within the battery, potentially accelerating the degradation of other components. Notably, the thermal behavior of interphases formed by sulfide electrolytes differs between glassy and crystalline forms. This variation arises from differences in atomic arrangement and the density of reactive sites, with glassy sulfides typically exhibiting lower onset temperatures for degradation compared to their crystalline counterparts. − Recent studies attribute the difference in thermal onset temperatures to the distinct reaction pathways followed by glassy and crystalline sulfide electrolytes. , Upon heating, glassy sulfides release S₂ gas relatively early in the degradation process as a result of their disordered, amorphous structure. The released S₂ gas then creates a reactive environment that can destabilize the cathode material, triggering an early phase transformation and an O₂ release. This released O₂ further reacts exothermically with the remaining glassy sulfide SE, amplifying heat generation within the battery. As the reaction transitions from a gas phase to a solid phase between the cathode and SE, it is termed the gas–solid reaction pathway (Figure a, b, and c). In contrast, crystalline sulfide SEs follow a different degradation pathway due to their ordered structure and stronger atomic bonds, which often include halogen atoms that contribute to improved structural stability. Due to the higher structural stability, crystalline sulfides do not release S₂ gas when heated, delaying decomposition until higher temperatures. That is, crystalline SE systems do not exhibit any exothermic peak until the interphase itself begins to degrade, typically after the O₂ is released due to thermal degradation of the cathode (Figure d, e, and f). The absence of gas-phase reactions in crystalline sulfides, combined with their slower reaction kinetics and limited reaction front, suppresses heat generation, reducing the total heat release by 1100 J g^–1 and delaying the onset of thermal runaway by approximately 100 °C compared with glassy sulfides. The earlier phase transitions observed in glassy sulfides (∼200 °C) are attributable to their amorphous structure, where the lack of long-range atomic order introduces more reactive sites and weaker bonds. This structural characteristic contributes to the lower degradation onset temperature of glassy sulfides, making them susceptible to initiating gas release and subsequent reactions with the cathode at relatively lower temperatures of 200 °C compared with crystalline SEs (Figure g and h).

15.

(a), (d) Schematics illustrating the thermal degradation pathways of glassy (LPS7) and crystalline (LPSCl) sulfide SEs in contact with phase-transforming cathodes such as NMC. Optical imaging depicting the intensity of heat release for different cathode composite pellets with (b) Li₇P₃S₁₁ (LPS7) and (e) LPSCl under rapid heating. (Reproduced with permission from ref . Copyright 2023 Royal Society of Chemistry). MS results representing the generation of O₂ and SO₂ for (c) LPS/NMC and (f) LPSCl/NMC cathode composite pellets as a function of temperature. (Reproduced with permission from ref . Copyright 2023 Royal Society of Chemistry). Ex-situ HRTEM image and the respective electron diffraction patterns after heating for the following cathode composite mixtures: (g) LPS/NMC. (Reproduced with permission from ref . Copyright 2019 Elsevier) and (h) LPSCl/NMC. (Reproduced with permission from ref . Copyright 2023 Elsevier).

Conversely, distinct thermal behavior is observed on pairing crystalline sulfide SEs with olivine cathodes, primarily due to the enhanced structural stability of olivine-type materials. Unlike layered oxide cathodes, olivine cathodes do not release O₂ during thermal decomposition, which eliminates the primary source of exothermic reactions with the SE. As a result, batteries utilizing olivine cathodes in combination with crystalline sulfide SEs do not display pronounced exothermic peaks that are typically associated with O₂ release from the cathode and subsequent reactions (Figure a, b, and c). However, even in the absence of an O₂ release, thermal degradation can still occur at the interface between the olivine cathode and the crystalline sulfide SE. This degradation often involves chemical reactions that lead to the formation of phosphates and metallic sulfides at the interface. While these reactions signify material breakdown at elevated temperatures, they proceed relatively slowly, and the reaction kinetics limit the total heat generated (Figure d, e, and f). This slower degradation process reduces the risk of rapid temperature increase that could lead to thermal runaway. It has been observed that employing lower heating rates during thermal testing can delay the onset of phase transitions in the cathode, including the depletion of Li and changes in the cathode structure. The slower heating allows for gradual diffusion processes, including the competitive lithiation of cathode particles, which can stabilize the structure temporarily and delay the decomposition processes. This effect ultimately results in a higher onset temperature of 220 °C, a lower magnitude of heat release, and a weight loss of 11% (from 250 to 360 °C). Additionally, the SOC of the cathode plays a critical role in the thermal stability of the cathode–electrolyte interface. Higher SOC levels correspond to a higher Li content extracted from the cathode, inducing lattice instability. This instability at high SOC can facilitate earlier phase transitions and trigger the early release of O₂ from layered oxide cathodes (Figure g), thereby decreasing the onset temperature for thermal degradation. , This trend holds across various types of sulfide SEs (glassy or crystalline) (Figure g), as the inherent instability of the cathode at high SOC increases its susceptibility to thermal decomposition, with stability decreasing by an average of 190 and 152 °C per unit rise in cutoff voltage.

16.

(a) Schematic illustrating thermal runaway propagation at the cathode interphase for crystalline sulfide electrolyte (LPSCl) in contact with a thermally stable electrode, such as LFP. (b) Optical image showing the mechanical stability of the LPSCl/LFP cathode composite mixture under grinding. (c) MS profiles quantifying the generation of CO₂ and SO₂ gases during external heating of the LPSCl/LFP sample. (Reproduced with permission from ref . Copyright 2022 American Chemical Society). XPS spectra representing the interphase composition in (d) LPSCl/LFP mixture after heating to 350 °C (Reproduced with permission from ref . Copyright 2022 American Chemical Society) and (e) LPSCl/NMC mixture after heating to 400 °C. (Reproduced with permission from ref . Copyright 2023 Royal Society of Chemistry). (f) TGA results comparing weight loss across different cathode/electrolyte mixtures as a function of temperature. (Reproduced with permission from ref . Copyright 2022 American Chemical Society). (g) Comparative analysis of the thermal stability of cathode interphases formed by oxide and sulfide electrolytes as a function of the cathode voltage.

Halide SEs tend to form stable interfaces with high-voltage cathodes such as LCO, NMC, and LFP under nominal operating conditions, primarily due to their high oxidation potential (wide electrochemical window) and low interfacial reaction kinetics. , However, the stability of the cathode-halide SE interphase is strongly influenced by temperature, the crystalline structure of the SE, and the nickel content of the cathode. For instance, antiperovskite type SE, such as Li₂OHCl (LOHC), underwent thermal degradation on sintering with cathode particles owing to the reactivity of hydroxyl free radicals of the SE with the decomposition products of the interphase. Whereas rocksalt-type halide SE, LIC, displays good compatibility with high-energy density cathodes at higher temperatures due to the higher strength of the In–Cl bond, forming a thermally stable interphase. Minimal impurity evolution is observed upon annealing LIC-cathode mixtures, except in the case of NMC811, which leads to byproduct formation such as InOCl and LiCl at approximately 170 °C. The increased reactivity of the LIC can be ascribed to the weakening of the In–Cl bond from the higher electronegativity of the NMC811 electrode. Furthermore, such halide SEs are noticed to absorb the O₂ released by the layered cathodes, converting the otherwise exothermic reaction into endothermic, in turn preventing overheating and subsequently improving the thermal stability. However, this O₂-trapping mechanism can also release Cl₂ gas and promote impurity phase formation at the interface, ultimately degrading the cell’s electrochemical performance. Moreover, the interfacial stability is significantly affected by both the cathode’s SOC and the reduction potential of the halide metal cation. Interestingly, calendar-aged cathodes at lower SOC exhibit greater interfacial degradation compared with those at higher SOC, a trend that contrasts with observations in conventional LIBs. This behavior is attributed to the higher Li content in low-SOC cathodes, which promotes the reduction of indium (In³⁺) ions at the interface to metallic indium, followed by its migration into the cathode. The incorporation of metallic In disrupts the cathode lattice, leading to significant volumetric expansion and eventual pulverization of the cathode particles. In contrast, the lower Li content in high-SOC cathodes suppresses In³⁺ reduction, thereby minimizing interfacial decomposition. Despite these findings, the influence of SOC-dependent degradation pathways and the reduction potential of halide metal cations on thermal stability remain poorly understood. Therefore, it is imperative that future studies systematically investigate these parameters to guide the development of thermally robust halide-based battery systems.

3.3. Lithium Metal Battery with Sulfur Cathode

Li–S chemistry has garnered extensive attention, owing to its high theoretical capacity. While earlier iterations exhibited numerous drawbacks, recent advancements have substantially improved the electrochemical performance of Li–S cells. − Despite recent advancements, thermal stability in Li–S batteries has only recently received focused attention, even though the S cathode exhibits a low decomposition temperature and operates with low flash point organic electrolytes. The following section will outline the influence of cathode and anode interfacial products on the thermal safety of Li–S batteries.

Although Li–S cells use an electrolyte and anode similar to those of LMBs, the anode interphase in Li–S cells features a passive layer of Li₂S formed by the LiPS (Figure a). While LiPS is known to increase cell overpotential, recent reports indicate its impact on the thermal characteristics of the cell. , For instance, the deposited lower-order polysulfides (i.e., Li₂S) on Li metal during charging are noticed to raise the cell temperature due to increased joule heating induced from higher interfacial resistance (Figure a and b). Heat release and temperature are at their peak when the LE is saturated with LiPS at maximum concentration. At the cathode interface, a major heat release occurred during the dissolution of S from the cathode in LE as LiPS (Figure a and b). With increased cycling, this LiPS dissolution accelerates thermal degradation, indicating that LiPS plays a critical role in the thermal stability of the cell. This reduced stability with cycling is associated with the reduction in thermal conductivity of the LiPS-saturated electrolyte compared with the pristine LE.

17.

(a) Schematic illustrating the influence of LiPS on interphase formation and heat generation in Li–S batteries. (b) Cell overpotential and heat flow during charge–discharge of an LE-based Li–S cell measured via operando microcalorimetry. (Adapted with permission from ref . Copyright 2018 Elsevier). (c) DSC profile showing the onset temperature and magnitude of the primary exothermic reaction governing thermal stability in Li–S chemistry. (d) Temperature evolution during oven testing of LE-based Li–S cells with different capacities (1 Ah and 5 Ah). (Reproduced with permission from ref . Copyright 2020 Electrochemical Society).

4. Cell-Level Thermal Stability

It is essential to investigate the thermal behavior of full cells, as the chemical affinity between battery constituents and the temperature dependence of reaction kinetics can significantly influence the thermal stability. Thermal runaway has been observed even in cells comprising thermally stable materials and electronically insulating interphases, primarily due to chemical crosstalk between the electrodes. The intensity and the rate of runaway are typically higher at the cell level due to the presence of both electrodes. This has led to the development of numerous techniques to investigate the thermal runaway of LMBs and propose different reaction pathways at the cell level. This section explores the reaction pathways that trigger thermal runaway in full-cell configurations across different LMB chemistries, offering insights into the key factors influencing their thermal stability and safety.

4.1. Lithium Metal Battery with Sulfur Cathode

Given the heightened cross-reactions between the cathode and anode in Li–S chemistry, exothermic interactions between the active materials are expected to be a major driver of thermal runaway. Unlike LMBs with oxide cathodes, where thermal runaway is primarily driven by O₂ release, thermal runaway in Li–S LMBs is triggered by direct redox reactions between the cathode and the anode (Figure c). As the temperature rises beyond the melting point of S (∼115 °C), the molten S can flow toward the separator. This situation worsens when the Li metal melts at approximately 180 °C, allowing both molten S and Li to come into direct contact. This direct interaction results in highly exothermic reactions, releasing significant amounts of heat, which further accelerates thermal degradation and heightens the risk of a catastrophic thermal runaway. The direct contact between these reactive molten phases at elevated temperatures leads to a vigorous redox reaction between the electrodes, releasing substantial amounts of heat. ,,, However, the overall heat release during this redox reaction is notably less relative to the exothermic reaction between electrodes in LMBs that use oxide-based cathodes. ,,, This phenomenon can be primarily attributed to the lower formation enthalpy of the S cathode. Furthermore, the energy released during the reaction between Li metal and S is considerably lower than that from exothermic reactions between Li metal and O₂ evolved from oxide-based cathodes, resulting in reduced heat generation intensity in Li–S systems. , Additionally, as temperature increases, the nonaqueous electrolyte in Li–S batteries begins to evaporate at relatively lower temperatures compared to other chemistries, reducing the available electrolyte for further parasitic reactions with Li metal and thereby limiting the total heat released. However, the generation of gaseous byproducts can increase the internal pressure of the cell, compromising the integrity of the cell casing and potentially leading to cell rupture and venting. Moreover, increasing cell capacity has been observed to reduce the thermal resilience of the cell (Figure d). An increase in the active material mass elevates the stored electrochemical energy in the cell, providing a larger reservoir of reactive material capable of releasing heat during a thermal runaway event. This elevated thermal energy output can rapidly accelerate the runaway process, elevating the internal temperature by 300 °C, thus significantly increasing the severity of the event. Additionally, the presence of a larger quantity of low-boiling-point LE exacerbates the situation by contributing to rapid vaporization and pressure buildup within the cell. This pressure accumulation can ultimately lead to catastrophic cell failure.

In Li–S cells, the onset of thermal runaway is significantly delayed even after events like internal short-circuiting, which is triggered by separator degradation and physical contact between active materials. This delay is attributed to the formation of a Li₂S passivation layer on the surface of the Li metal. When an internal short-circuit occurs, the presence of soluble LiPS in the LE facilitates the deposition of Li₂S onto the Li metal, which acts as an insulating barrier, preventing the direct electrical connection between the Li metal and the S cathode (Figure a, b, and c). This characteristic of the Li–S system is particularly evident during nail penetration tests. During such tests, the formation of a Li₂S coating on the inserted nail due to the reduction of higher-order polysulfides further limits the flow of current between the punctured anode and cathode (Figure d and e). By preventing a direct electrical pathway, the heat generated by the short-circuit is substantially reduced and the cell is less likely to experience rapid exothermic reactions that would drive thermal runaway. This behavior stands in stark contrast to that of LMBs with oxide-based cathodes, where nail penetration commonly triggers an immediate and intense thermal response due to the rapid formation of an electrical pathway between electrodes. The SOC of a Li–S battery also plays a nuanced role in the onset and progression of thermal runaway. Higher SOC is noticed to lower the onset temperature for self-heating but delays thermal runaway and subsequent cell failure. The reduced temperature for the initial thermal reaction at a higher SOC can be attributed to the lower stability of the highly delithiated cathode. The delay in thermal runaway at higher SOCs is attributed to the lower concentration of soluble LiPS present in the electrolyte under these conditions. As the cell charges, more S is converted to higher oxidation states, leaving fewer LiPS in the solution. LiPS is reported to play a critical role in facilitating parasitic side reactions that accelerate heat generation. With lower LiPS levels at high SOC, these reactions are less intense, allowing the cell to dissipate the initial thermal energy gradually and delaying the onset of runaway. Moreover, cycled cells exhibit a lower risk of thermal runaway due to the increased concentration of dissolved polysulfides in the electrolyte and the formation of a thicker, inorganically rich SEI on the anode. The higher amounts of polysulfides increase the viscosity of the LE, reducing the mobility of ions and limiting the rate of heat-producing reactions during thermal events, thus providing a buffer against rapid runaway conditions. The inorganically dense SEI in cycled cells also shields the Li metal from further parasitic reactions with the LE, thus slowing the rate at which exothermic reactions can proceed. Li–S chemistry also offers better resilience to overcharging than LMB systems with oxide cathodes. This resilience is due to the S cathode’s ability to tolerate deeper charge states without undergoing structural degradation or excessive heat generation. Owing to the S cathode’s ∼30% lower reaction enthalpy with Li (ΔH _f = −437 kJ mol^–1) and the absence of O₂ evolution even under overcharging, the heat generation is substantially lower than in layered oxide cathodes, which undergo phase transformations accompanied by O₂ release.

18.

(a) Schematic illustration of a typical LE-based Li–S cell. (b) Illustration of the various thermal degradation pathways and associated exothermic reactions in a full Li–S battery. (c) Time-resolved temperature and voltage profiles of a Li–S cell discharged to 100% SOC from EV-ARC. (Reproduced with permission from ref . Copyright 2021 Elsevier). (d) Schematic comparison of the underlying mechanisms driving the distinct thermal responses of Li–S and conventional LMBs during nail penetration. (e) Time-dependent voltage and temperature evolution during a nail penetration test on a Li–S cell using LiTFSI-DOL/DME. (Reproduced with permission from ref . Copyright 2022 Elsevier).

4.2. Lithium Metal Battery with Liquid Electrolyte

Thermal runaway in LMBs can be triggered by internal short-circuit as well as through the reaction of O₂ released from the cathode with Li metal, a phenomenon referred to as “chemical crosstalk” (Figure a, b, and c). The exothermic reactions and thermal runaway pathways in LMBs differ fundamentally from those in LIBs. Typically, thermal runaway progresses through three stages: external heating, self-heating, and the onset of uncontrollable exothermic reactions. For example, self-heating of LMB occurs due to anode interphase decomposition and subsequent exothermic reaction between Li metal and LE. During the self-heating stage, separator shrinkage and Li melting are observed, leading to microshort circuits and additional heat generation. In addition, localized heating can cause the pores of the separator to close, leading to an increased internal resistance. This impedes Li-ion transport, accelerates nonuniform Li deposition, and promotes the formation of additional hotspots. Finally, the thermal runaway of the cell is triggered by the reaction of O₂ from the cathode with Li metal (Figure d). A recent study systematically investigated the potential exothermic reactions in LMBs, quantifying their individual contributions to thermal runaway by varying the cathode chemistry, LE content, and applied heating power. The primary exothermic reaction was attributed to the combustion of O₂ from the cathode with Li metal, releasing 156.8 J cm^–2 of heat and catching fire in 2.6 s. Interestingly, under the same heating power, LIBs reached temperatures nearly twice as high as those of LMBs before failing, yet did not ignite and released only about half the amount of heat (at 100% SOC with NMC811 as the cathode). Moreover, upon removal of the solvent from the LMB, ignition occurred even earlier (1.6 s) and at a lower internal cell temperature, despite a reduction in total heat release to 45.92 J cm^–2. The early fire was attributed to the direct reaction of gaseous O₂ released from the cathode with Li metal, whereas in cells with LE, the O₂ is dissolved in the solvent hence delaying the reaction. This indicates that Li metal’s high reactivity with O₂ triggers thermal runaway and ignition before the LE fully combusts. The authors further examined the reactivity of Li with O₂ released from the cathode by charging cells to different SOC levels before an internal short-circuit test. The results showed a clear SOC-dependent trend in thermal runaway intensity: at 75% SOC, the cell produced only smoke after 12 s, whereas at 50% SOC, no smoke or flame was observed. In addition to SOC, the study revealed that a critical threshold of heating power is required to trigger combustion in LMBs, regardless of the shorting current. These observations highlight that while LMBs possess comparable energy densities and material compositions (e.g., cathode and electrolyte) to conventional LIBs, they exhibit markedly lower thermal stability, releasing 67% higher heat on runaway. This underscores the need for a deeper mechanistic understanding of thermal runaway initiation and the various factors that govern its severity.

19.

(a) Schematic illustrating electrode crosstalk phenomena that influence thermal runaway in LMBs at the cell level. (b) DSC profiles comparing heat generation from the cathode alone versus a cathode-anode mixture. (c) In-situ MS results representing the concentration of O₂ during heating of the cathode powder and the cathode-anode mixture. (Reproduced with permission from ref . Copyright 2018 Elsevier). (d) Comprehensive evaluation of temperature evolution and associated stages during thermal runaway in LE-based LMBs.

Studies have explored the influence of electrochemical cycling on the thermal runaway behavior of LMBs, revealing that cycled cells exhibit a lower onset temperature. This shift is primarily attributed to increased parasitic reactions between dendritically deposited Li and the LE (Figure a and b). , Interestingly, in anode-free cells, cycling led to an increase in the magnitude of heat release, yet the onset of thermal reactions was observed to be delayed. This delay is linked to the formation of an inorganically rich SEI (with higher amounts of Li₂O) on the anode in cycled cells (Figure c). The higher onset temperature could also be influenced by reduced thermal conductivity caused by a larger presence of gaseous products in cycled cells. Upon evaluating the effect of capacity and active material loading (N/P) on runaway behavior, it is noted that thermal stability is inversely related to cell capacity and active material loading. Cells with a higher N/P ratio have larger amounts of Li metal, resulting in an increased exothermic reaction with LE (Figure d and e). Meanwhile, higher-capacity cells attain elevated temperatures and release greater amounts of heat due to their larger energy storage. Studies have demonstrated that during electrical abuse, olivine cathode cells exhibit an earlier triggering temperature for thermal runaway compared to oxide-layer cathode cells. , This behavior contrasts markedly with thermally triggered runaway in oxide-cathode cells, where O₂ release from the active material leads to a lower onset temperature. Additionally, olivine cathode cells release significantly less heat than do those with layered cathodes. Although the precise mechanism underlying this disparity remains unclear, these findings highlight that thermal runaway can still occur, even in cells employing thermally stable cathodes.

20.

SEM image of the Li anode surface after (a) 25 and (b) 100 cycles. The electrolyte used here is 1 M LiPF₆-EC/DMC (1:1). Left panels depict the correlation between Li deposition morphology and the intensity of exothermic reactions with the LE at the end of each cycle. (Reproduced with permission from ref . Copyright 2019 Elsevier). (c) ARC profiles for anode-free cell comparing the heat release characteristics after the 1st cycle and the 100th cycle. (Reproduced with permission from ref . Copyright 2023 John Wiely and Sons). (d) Schematic illustrating the influence of anode mass loading on thermal degradation pathways and heat release intensity in LMBs.

4.3. Solid-State Lithium Metal Battery

A comprehensive thermal analysis of full SSBs is essential to uncover previously overlooked exothermic reactions that may play a critical role in initiating thermal runaway. One significant exothermic mechanism involves the decomposition of SEs, which can generate substantial heat at elevated temperatures, particularly when the SE is unstable against Li metal or its degradation products are highly reactive. Furthermore, dendritic Li growth on propagation through grain boundaries in the SE, not only triggers short-circuit but also generates tremendous heating in electrical connections at the short-circuit points, thus exacerbating the occurrence of thermal runaway (Figure a). , Once thermal runaway is initiated, the heat release is both greater in magnitude and faster than in LE-based counterparts. , For instance, with similar capacities, SSBs are noticed to release 2.8 times (75.7 J cm^–2) higher heat per cell area compared to carbonate-based LIBs (27.1 J cm^–2). This behavior stems from the higher reaction kinetics (1190 kJ mol^–1 vs 470 kJ mol^–1 for LIB), the inherently poor heat-dissipation properties of SEs, namely, their low thermal conductivity and specific heat capacity, which collectively accelerate internal temperature rise within the cell. Accordingly, this section outlines the key exothermic mechanisms that may govern the thermal instability in solid-state Li metal batteries.

21.

(a) Illustration of a typical SSB. (b) Possible thermal runaway pathways at the cell-level due to chemical crosstalk between cathode and Li-metal and Li penetration through SE (c) DSC heat flow profiles for a full cell composed of LCO/LLZO/Li. (Reproduced with permission from ref . Copyright 2022 Electrochemical Society). (d) Schematics of potential exothermic reactions at the cathode/electrolyte interface. (e) High-resolution TEM images of Li deposit in LiPF₆-EC/EDC (1:1) presenting the formation of Li₂O layer at the interface. (Reproduced with permission from ref . Copyright 2021 Springer Nature). (f) Comparison of DSC heat profiles for fully discharged (100% SOC) cathode composite mixture against cathode (NMC) and electrolyte (LPSCl). (Adapted with permission from ref . Copyright 2023 Elsevier). (g) Schematics illustrating thermal runaway pathways triggered by the degradation of secondary materials in the cathode composite, such as carbon black and polymer binders. (h) DSC heat curve for LCO/LLZO/Li cell representing various exothermic peaks corresponding to the decomposition of cathode, carbon black, and the polymer binder. (Reproduced with permission from ref . Copyright 2023 American Chemical Society).

As exothermic reactions are often linked to interphase instabilities, it is anticipated that having a stable SEI at both the cathode and the anode could help mitigate the risk of thermal runaway by limiting the occurrence of highly exothermic reactions and preventing the formation of new, unstable compounds during operation. However, recent studies have reported incidents of thermal runaway in SSBs that employ oxide-based SEs, which are known for forming stable interphases with Li metal and cathode materials. Interestingly, the thermal runaway was triggered at a temperature lower than the oxide SE’s interphase thermal degradation (300 °C), suggesting the involvement of additional underlying factors (Figure b and c). One potential contributor is the mechanical stress and microcracking that develop within the SE during repeated charge–discharge cycles. Even though the interphases remain chemically stable, these microcracks can act as pathways for dendritic Li growth, enabling Li metal to bridge the anode and cathode through the SE and causing internal short circuits. During an internal short-circuit, the stored electrical energy is rapidly converted to heat, significantly reducing the onset temperature for thermal runaway. Short circuits in LMBs can also be externally induced, such as during collisions or mechanical abuse, through structural failure of the cell, wherein the collapse of the SE leads to direct contact between the cathode and Li metal anode, releasing a substantial amount of heat. Moreover, structural collapse may allow atmospheric O₂ to penetrate the cell and react with Li metal, further intensifying thermal runaway. Even in physically intact cells, the most pronounced exothermic event is often driven by the reaction between the amount of O₂ released from the cathode and the Li metal anode. This highly exothermic interaction leads to rapid internal heat generation, accelerating the onset of thermal runaway. Although O₂ diffusion through the dense structure of oxide-based SEs in SSBs is typically limited, crosstalk can still occur due to the flowability of molten Li at elevated temperatures. Upon melting, Li may spread along the internal surfaces, especially under the operational pressures typical of SSBs, and encounter O₂-rich regions, such as the periphery of the SE, where O₂ released from the cathode accumulates. This results in localized exothermic reactions and a further heat release. Additionally, O₂ can diffuse through defects, pores, or cracks in the SE, formed either electrochemically during cycling or mechanically during fabrication, and react with molten Li, thereby triggering thermal runaway. , Interestingly, crosstalk behavior is notably different in sulfide-based SSBs and occurs only beyond threshold SOC values, particularly at higher SOCs. ,,, When O₂ is released from the cathode, it tends to react almost immediately with the sulfide SE, forming Li sulfates and other degradation products (Figure d and f). This instantaneous reaction consumes the released O₂ before it can diffuse toward the anode, effectively preventing crosstalk between the cathode and the Li metal anode at lower SOCs. Additionally, this exothermic reaction releases SO₂, which can significantly increase internal cell pressure and pose serious safety risks due to its high toxicity in the event of cell rupture during thermal runaway. However, at higher SOCs, a greater amount of O₂ is released from the cathode. The increased release rate can exceed the capacity of the sulfide electrolyte to react with the O₂ immediately. As a result, some excess O₂ can reach the Li metal anode, where it reacts beyond the Li melting point, resulting in crosstalk and producing a substantial exothermic release of energy. It has also been observed that the Li metal sometimes forms a passive layer of Li₂O due to contamination during foil preparation, which could reduce the intensity of the reaction with O_2, thus increasing the time required to reach thermal runaway (Figure e). As Li₂O is a semiporous solid with a decomposition temperature higher than 180 °C, it could hinder the flowability and reactivity of molten metal with O_2, thus reducing the intensity of thermal runaway. For instance, in the presence of such a passivation layer, the heat release corresponded to the reaction of only 25% of the O₂ released from the cathode with Li metal, while the remaining heat was attributed to the reaction of O₂ with carbon black, resulting in CO₂ formation (Figure g and h). Additionally, other inactive components in the cathode mixture, such as polymer binders, have been shown to participate in exothermic reactions during thermal runaway, undergoing decomposition and releasing HF gas as a byproduct. ,, As the reaction progresses, HF reacts exothermically with Li metal, generating additional heat that intensifies the temperature increase and accelerates the onset and propagation of thermal runaway. The ratio of secondary to active materials also affects the exothermic reaction pathway, with higher electrode loadings exhibiting a 33% increase in heat release. This behavior arises from the larger quantity of secondary materials in thicker electrodes that react with the oxygen evolved from the cathode during thermal decomposition, thereby reducing its interaction with molten phases (e.g., Li), whose oxidation exhibits a much higher reaction enthalpy (43 MJ kg^–1 vs 8.93 MJ kg^–1). However, the thermal degradation of secondary cathode materials can lead to excessive gas evolution, increasing internal pressure and potentially causing cell rupture. It must be further noted that while lack of lattice oxygen release in cathode systems such as LFP can delay the onset of and intensity of thermal runaway, under sufficient thermal and mechanical perturbation, a highly exothermic reaction between lithium and oxygen leading to the formation of Li₂O can still occur through direct redox interactions between the electrodes. Such reactions involving the reduction of metal oxides by reactive lithium can elevate the internal cell temperature to nearly 2500 °C, highlighting the critical need to elucidate all possible reaction pathways at the cell level for designing safer SSBs with enhanced thermal stability.

5. Discussion and Perspective on the Safety of Lithium Metal Batteries

As discussed earlier, thermal runaway not only limits battery performance but also presents serious safety risks. Due to their high energy density and chemically reactive components, LMBs exhibit more intense exothermic behavior than LIBs. This has led to a surge of recent research aimed at unraveling the underlying thermal mechanisms. Numerous strategies have been proposed to enhance the thermal resilience of LMBs, primarily through material-level improvements. However, given the complex and multidimensional nature of thermal runaway, these strategies may not fully prevent its occurrence. In this section, we identify key exothermic reactions that are critical yet underexplored in the context of thermal runaway in LMBs. Additionally, we provide an overview of the experimental techniques essential for elucidating the underlying reaction pathways that trigger runaway. This section concludes by proposing promising and implementable strategies for advancing the development of thermally resilient LMBs.

5.1. Material-Level Thermal Stability

In full-cell configurations, thermal runaway is primarily triggered by the reaction between O₂ and molten Li; consequently, recent research has focused on strategies to elevate the Li melting point and restrict its flowability at elevated temperatures. − One extensively studied approach is the 3D architecture anode, where Li is infused into an electrically conductive framework. Such architecture electrodes are noticed to enhance the uniformity of surface chemical reactions and Li flux during cell operation, thus substantially reducing the formation of dendritic Li. Studies have also explored alloy-based anodes, where the fast diffusion coefficient of metallic Li alloys promotes dense Li deposition. − The high rigidity of the carbon framework and the increased bond strength of binary alloys (due to coherence strain) could increase the melting temperature and the flowability of the anode significantly compared to those of 2D Li metal foil. However, the conductive network can form interphases in the presence of other battery constituents, the decomposition products of which can result in additional side reactions at elevated temperatures. Further, the alloy-anodes undergo extensive volumetric change during cycling, which can initiate stress-induced crack propagation when paired against SEs, thereby triggering a short-circuit. Though the actual thermal degradation pathways in such advanced electrodes are poorly understood, they are believed to delay the onset of thermal runaway but may not prevent it. An alternative approach to avoid thermal runaway can be the modulation of Li microstructure’s crystal orientation and surface energetics, as the thermophysical properties of Li metal are typically influenced by its microstructure. Though recent studies have noticed the influence of Li microstructure, grain size, and orientation on the uniformity of Li deposition and electrochemical performance, their influence on the thermal stability of the bulk Li anode is largely unexplored. − Therefore, future studies must investigate the influence of Li purity, microstructure, texture, and morphology on thermal response in LMBs.

For high-voltage cathodes, different strategies have been developed to prevent the release of O₂ and, thereby, thermal runaway as temperature increases. − Each approach faces unique challenges, ranging from manufacturing difficulties to unwanted side reactions. Among them, single-crystalline (SC) cathodes and high-entropy cathodes have garnered increased attention due to the crack-free morphology, zero volumetric strain, and their resistance to O₂ release (Figure a and c). − Though such cathodes exhibit better electrochemical stability relative to polycrystalline (PC) cathodes owing to their minimal volume change (Figure b), their higher surface area can cause unwanted interfacial stress under thermal abuse and thus must be carefully studied. Further, it has been reported that thermally stable cathodes liberate CO₂ and CO on heating due to the degradation of the surface contamination layer (carbonates) (Figure d). − Surface contamination could occur during manufacturing due to contamination of dry rooms or a higher dew point (−40 °C) compared to a lab-scale glovebox (−80 °C). Such released gas has the potential to react exothermically with LE, thereby increasing the cell temperature and contributing to runaway (Figure e). Therefore, it is important to have a passivating coating that avoids such surface contamination without reducing the ionic or electronic conductivity of the particle. Moreover, the selected coatings must exhibit high crack resistance during prolonged cycling and remain chemically inert toward secondary materials across the cathode’s electrochemical operating window. For S cathodes, current state-of-the-art efforts primarily aim to suppress polysulfide dissolution in the LE through elemental doping (e.g., C or N) or the use of conductive porous architectures, while improvements in thermal stability remain largely overlooked. − As a result, it is crucial to develop strategies that enhance their melting temperature without affecting the electrochemical stability window.

22.

(a) Cross-sectional SEM images of PC-NMC and SC-NMC after 400 cycles. (Reproduced with permission from ref . Copyright 2021 Elsevier). (b) Rietveld refinement derived lattice parameters for SC-NMC and PC-NMC as a function of temperature. (Reproduced with permission from ref . Copyright 2023 Elsevier). (c) MS profiles comparing the evolution of gases such as O₂ and CO₂ from PC-NMC and SC-NMC. (Reproduced with permission from ref . Copyright 2022 Electrochemical Society). (d) MS profiles of the CO₂ generated for the pristine NMC811 electrode and Li₂CO₃-coated electrode. (Reproduced with permission from ref . Copyright 2023 John Wiley and Sons). (e) Schematic illustrating the mechanism of liberation of CO₂ from moisture-contaminated cathode surfaces and its impact on exothermic reactions and overall heat generation within the cell.

For LEs, current strategies focus on improving thermal stability by adding thermally stable additives that suppress free radical generation or using solvents with higher decomposition temperatures. However, they are observed to affect the cell’s electrochemical performance due to their lower ionic conductivity at room temperature. While studies have shown that this can be addressed by increasing the overall internal average temperature of the cell, higher temperatures can lead to heterogeneity at the anode interphase owing to variations in local Li-ion flux, affecting the Li deposition. Another promising approach involves modulating the solvation entropy of the LE through the incorporation of multiple additives, leveraging their influence on the Gibbs free energy of dissolution. While such additives have been shown to enhance electrolyte rate capability at lower temperatures, their thermal stability remains poorly understood. The higher configurational entropy introduced by multicomponent electrolytes may result in compositional and morphological variations in the SEI at elevated temperatures, driven by the interplay between temperature and entropy effects on the solvation energy of the LE. This underscores the need for a comprehensive investigation into the thermal stability of such advanced electrolyte systems. The complex and often nonmonotonic relationship between electrochemical performance and thermal stability presents a major challenge in designing LEs that simultaneously offer a wide electrochemical window and robust thermal resilience.

Although SEs inherently exhibit high decomposition temperatures, some can recrystallize into phases with poor ionic conductivity at elevated temperatures. Consequently, recent studies have explored the use of additives to homogenize the ionic conductivity of the resulting decomposition products. − Further, as discussed in Section , the low air stability of sulfide increases the risk of degradation and cross-contamination with other materials in a large-scale production process, where the dew points are significantly higher than lab-scale gloveboxes. Passivation coatings can mitigate the reactivity of sulfide SEs (as discussed in Section ), but their influence on the thermal stability must be evaluated under practically relevant conditions. As future SSBs require thin SEs (below 30 μm), it is essential to examine their thermal degradation behavior, especially since the fabrication process often incorporates secondary materials such as dissolving solvents and polymeric binders.

5.2. Interface- and Interphase-Level Thermal Stability

5.2.1. Anode/Electrolyte Interface

The parasitic reaction between the electrolyte and Li metal is highly exothermic, elevating the internal temperature and driving the cell into a self-heating regime that initiates a self-reinforcing cycle of heat accumulation, temperature rise, and further reaction. Consequently, current state-of-the-art strategies prioritize achieving uniform and dense Li deposition to suppress such parasitic reactions and enhance thermal stability. − Though such strategies improve the electrical performance owing to the promotion of dense Li deposition, with a few even delaying the onset of thermal reaction, it is not apparent whether they prevent thermal runaway. An emerging research thrust that warrants greater attention is the influence of fundamental electrolyte properties, such as solvation strength and ion-pairing, as these factors are observed to critically govern SEI composition and Li deposition morphology. ,− For instance, weakly solvating LEs are reported to result in an inorganic-rich SEI, owing to the coordination shell of Li-ions being composed of salt anions rather than solvents. Thus, by fine-tuning these parameters, it is possible to promote homogeneous Li deposition and engineer a tailored SEI with an inorganic-rich surface, thereby suppressing parasitic reactions and improving thermal stability.

It was also observed that increasing the stack pressure reduced the heat released due to the reaction between Li metal and LE, as the applied mechanical forces promoted denser Li deposition. The resulting low-surface-area morphology limited parasitic reactions with the LE (Figure a and b). However, during cycling, the large volumetric changes experienced by Li can induce cracks in the SEI and lead to dead metal, thereby severely undermining the thermal stability of such systems. Similarly, incorporating thermal boundaries other than isothermal conditions is noticed to influence the lithiation state of electrodes during cycling. , For instance, for LIBs, it has been illustrated that a directionally small thermal gradient (around 2 °C) has a significant influence on the cycling behavior and capacity degradation. Such electrochemical variations are attributed to temperature-driven shifts in the overpotential of the electrodes. For example, when the cathode temperature is higher than the anode temperature, it induces a positive shift in the anode overpotential, thereby facilitating increased Li plating and exacerbating parasitic reactions. In addition, thermal gradients have been observed to influence the uniformity of Li-ion flux at the LE-Li interface, subsequently affecting the initial Li deposition morphology in LMBs. These thermal gradients can readily develop in LMBs under fast charging or during extreme operation at ambient conditions. Therefore, it is essential to investigate the influence of such external thermal modulations on the thermal stability of LMBs, particularly under practically relevant operating conditions. Thermal analysis of the interface formed at Li metal/SE is particularly challenging due to its buried nature and the significant dependence on electrolyte composition and structure. Although several strategies have been developed, they do not eliminate thermal runaway. , Given the early stage of such investigations, numerous implicit and explicit parameters may influence thermal runaway, underscoring the need for further study. For instance, interphase degradation and electrolyte fracturing, driven by the larger molar volume of Li deposits during extended cycling, have been reported to facilitate dendrite propagation (Figure c and d). , Due to the low melting point of Li and the application of high pressures, dendritic growth resulting from structural degradation can accelerate the onset of thermal runaway. Furthermore, these unstable interphases can facilitate Li deposition inside the interphase, creating internal stress and cracking of SE, thereby escalating dendritic Li propagation through the electrolyte and the risk of an internal short-circuit. Additionally, some sulfide SEs experience phase changes when heated, which can drastically alter their ionic conductivity and morphology. This, in turn, may lead to higher cell overpotential and contact loss, thereby raising internal cell heating. − Therefore, phase changes and decomposition products must be tracked operando during the thermal runaway process to gain insight into their effect on the thermal stability. Although oxide SEs form a stable SEI against Li metal, high-temperature sintering could introduce voids and lattice distortion into electrolyte pellets. These defects are observed to increase the electrolyte’s porosity, particularly under heating. The higher porosity can reduce SE’s mechanical integrity, increasing the susceptibility of dendrite propagation into SE and subsequently resulting in accelerated runaway (Figure e) if addressed otherwise. Besides defects, the initial particle size of the SE is noted to have a significant influence on the amount and distribution of porosity in the pressed electrolyte pellet. Though few studies have demonstrated the reduction of sintering temperatures, thereby reducing porosity level, there is limited understanding of the influence of such strategies on the thermal resilience of the formed interphase when paired against active materials. In addition, oxide SE’s poor wettability with Li metal could escalate at higher discharging rates and cycle numbers, increasing cell overpotential and driving the formation of highly heterogeneous Li deposition morphology, thus heightening the risk of thermal runaway. ,, Accordingly, future research must account for these factors and their cumulative effects on thermal stability.

23.

(a) Comparison of the Li deposition morphology and the rate of parasitic reaction in the presence of electrolyte as a function of externally applied pressure. (b) DSC profiles depicting the heat released by the deposited Li in Li/Li symmetric cell at different stack pressures: 150 and 350 kPa. (Reproduced with permission from ref . Copyright 2023 American Chemical Society). (c) Schematic of crack propagation and Li penetration through SEs at the LPS/Li interface during extended cycling. (Reproduced with permission from ref . Copyright 2021 American Chemical Society). (d) X-ray tomography reconstruction of the 3D structure of the electrolyte pellet before and after heating. (e) ARC results comparing the thermal characteristics of Li/LATP (with defect) and Li/LiPO₂F₂-LATP (defect modified) symmetric cell. (Reproduced with permission from ref . Copyright 2021 American Chemical Society).

5.2.2. Cathode/Electrolyte Interface

The state-of-the-art strategies to mitigate exothermic reactions at the CEI primarily focus on suppressing or blocking O₂ release during thermal decomposition. Degradation at the cathode/LE interface can be reduced through surface coatings or electrolyte additives that either inhibit O₂ evolution or capture released O₂, thereby preventing its exothermic reaction with the LE. However, such strategies may increase internal pressure, exacerbating cathode particle degradation during extended cycling. Since the reaction between O₂ and LE is a primary contributor to thermal degradation, future research should focus on engineering O₂-capturing pathways that convert the otherwise exothermic process into an endothermic one, thus alleviating internal heating of the cell. This approach may slow interfacial reaction kinetics and compromise electrochemical performance, necessitating a careful balance between thermal stability and cell efficiency. − Additionally, the implemented strategies may generate reactive decomposition products at elevated temperatures during heating. Therefore, it is essential to investigate the high-temperature interphase reaction kinetics associated with these mitigation schemes to fully understand their impact on thermal behavior and long-term cell stability. For the cathode/SE interphase, the underlying mechanism of thermal runaway is not yet completely understood. Sulfide SEs have been observed to undergo significant volumetric deformation during charge and discharge when paired with high-voltage cathodes during heating. This is attributed to the space-charge layer formed at the interface, driving the migration of Li-ions between the cathode and sulfide SE (Figure a). Such repetitive volumetric change could induce cracks in electrolytes, creating pathways for dendrite propagation and resulting in an internal short-circuit. Considering the substantial influence of SE cracking on thermal stability, future research should explore the integration of high molar mass polymers with enhanced ionic conductivity to buffer the large volumetric changes in these electrolytes. Such an approach could improve the ductility of the SE while preserving its high decomposition temperature. Additionally, cosintering implemented during cathode/oxide electrolyte mixture preparation introduces impurity phases at the interface owing to their mutual reactivity (Figure b). These impurity phases, observed even in olivine cathodes, are attributed to increased Li transport at elevated temperatures. Such phase changes can act as local hotspots as they increase the local overpotential, increasing internal cell temperature and contributing to the thermal runaway. Recent studies have focused on incorporating advanced cathodes, such as two-phase, zero-strain, and SC materials due to their minimal phase transformation and low volumetric strain. However, a critical consideration remains their chemical compatibility and thermal stability with other electrode constituents, including carbon binders and SE. − As an example, though LFP has better crack resilience and does not release O₂, in the presence of carbon binders and SE shows significant oxidation, drastically increasing the interfacial impedance of the cell. Studies have also revealed that microstructural heterogeneity within the cathode/SE mixture causes nonuniform reaction kinetics and uneven heat generation, resulting in a thermal wavefront that propagates from the separator to the current collector side (Figure c and d). These nonuniform responses are due to the variation of point contact and ion-conducting pathways across the cathode mixture. Nonuniform heat generation could further lead to hotspots, especially under high charge/discharge conditions, and accelerate the thermal failure of the cell. Gas generation at the cathode interface has also been identified as a major factor contributing to the rise in cell overpotential with cycling (Figure e and f). This contrasts with previous intuition, where an increase in the contact loss was linked to an increase in overpotential. Although the source of gas generation remains unclear, the accumulation of gaseous products could create pressure differentials in addition to raising the internal pressure, worsening the nonuniform cathode usage, and potentially leading to cell rupture. One approach to avoid interfacial reactions between the cathode and SE is to develop thermally resilient, high-voltage cathodes that do not release O₂ and are composed entirely of active material (100 wt % active). The lack of an ionic conductivity network in such cathodes necessitates designing the crystallographic orientation of the individual active materials to promote a uniform Li flux at the SE/cathode interface. However, due to limited understanding of the behavior of such cathodes at higher temperatures, it is crucial to investigate their possible degradation pathways and their impact on the thermal stability at the full cell level.

24.

(a) In-situ SEM images of the NMC-LPS cathode composite mixture at room and elevated temperature. (Reproduced with permission from ref . Copyright 2017 Elsevier). (b) Illustration of the formation of impurity phases in the cathode-oxide electrolyte composite during high-temperature sintering. (c) Visualization of electrode heterogeneity-induced volumetric heat generation during 1C discharge. (d) Temporal evolution of cell voltage and lithiation distribution across the cathode during 1C discharge. (Reproduced with permission from ref . Copyright 2023 Elsevier). (e) Ultrasonic imaging of cathode/electrolyte interphase degradation as a function of cell potential. (f) Linear sweep voltammetry profiles for LCO/LLZTO+PEO/Li cells. (Reproduced with permission from ref . Copyright 2022 American Chemical Society).

5.3. Cell-Level Thermal Stability

Several strategies have been proposed to improve thermal stability at both the material and component levels; however, the multidimensional nature of thermal runaway prevents any single solution from fully eliminating it at the full-cell level. This limitation stems in part from an incomplete understanding of the various degradation pathways that contribute to these instabilities. For example, the formation of electrochemically inactive Li has been known to impact the CE of the cell for some time; however, experimental approaches to quantify its accumulation have only emerged recently. − The formation of dead metal is observed to be influenced by external factors such as temperature and current density, and internal parameters such as the solvation strength of LE (Figure a). , For instance, the amount of dead metal increases with the current density and decreases with the temperature. Yet, the influence of dead metal formation on thermal runaway remains largely unexplored. Dead metal is hypothesized to exhibit exothermic reactions and degradation pathways similar to those of Li metal, exacerbating the cell thermal runaway. Similarly, the impact of calendar aging and fast charging on a cell’s thermal stability remains largely unexplored (Figure b). It has been reported that calendar aging reduces cell capacity due to increased interphase growth from the continued corrosion of Li metal by the electrolyte. Meanwhile, fast charging elevates the risk of dendritic Li growth, from rapid concentration depletion at the anode interface and highly heterogeneous electric field distribution, leading to premature cell short-circuit (Figure b). , For Li–S chemistries, the thermal safety study is in a preliminary stage, with the effect of external factors, such as temperature, pressure, and charge/discharge protocols on the thermal degradation pathway yet to be investigated. For instance, though lower capacity loss and enhanced cycle life are observed with a restrained voltage window, it is important to investigate how such modulations influence the thermophysical properties of the cell, the generation of local hotspots, and thermal runaway during extended cycling. Thus, it is important to investigate the thermal response of liquid LMBs in light of practical charge–discharge cycles and targeted application domains.

25.

(a) Quantification of dead Li metal area fraction as a function of temperature: −20, 20, and 40 °C at current densities of 1 mA cm^–2 and 2 mA cm^–2. (Reproduced with permission from ref . Copyright 2020 American Chemical Society). (b) Demonstration of thermal degradation pathways induced during faster charging and calendar aging of LMBs. (c) Comparison of overpotentials for Li/LPSCl/Li–In alloy asymmetric cells during Li plating under different stack pressures including 5 and 20 MPa. (Reproduced with permission from ref . Copyright 2023 John Wiley and Sons). (d) Illustration of localized pressure differentials driven by O₂ release and degradation mechanisms of the O₂ in SSBs. (e) SEM images of the patterned S cathode (after preparation) and Li anode (after cycling). (f) Electrochemical cycling comparison between baseline (nonpatterned) and patterned S cathodes. (Reproduced with permission from ref . Copyright 2022 John Wiley and Sons).

For SSBs, thermal runaway phenomena at the cell level have not yet been completely understood. Most studies report thermal runaway with cells consisting of pristine components, even though cycling increases the likelihood of dendritic Li formation. The Li dendrites can penetrate through SE either along the grain boundary or through the electrically conductive interphase cycling, inducing large volume changes at the anode interface and creating cracking and pulverization of SE particles. − Voids formed during the stripping (charging) process have also been shown to dramatically increase cell overpotential, leading to point contact and current constriction, which can cause cell failure depending on the stack pressure (Figure c). High overpotential at such point contacts can further contribute to thermal runaway through increased joule heating. , Conversely, higher stack pressure can not only lead to accelerated dendrite growth but also increase the prospect of cathode active material cracking, which could instigate premature gas release and thermal runaway. In addition, other external parameters must be examined in the context of thermal runaway, particularly moisture, as it has been shown to alter the morphology and composition of interphases at both the cathode and anode, leading to a degraded electrochemical performance. Moreover, the stochastic distribution of porosity within the SE can restrict the permeation of reactive gases released from the cathode, resulting in localized pressure differentials and stress concentrations that exacerbate mechanical degradation and cracking of the SE. The increased susceptibility of local crack generation can accelerate Li propagation and internal short-circuit if not otherwise addressed (Figure d).

Meanwhile, current thermal stability tests for SSBs, such as ARC and DSC, are typically conducted at atmospheric pressure, which is significantly lower than the stack pressures of 10–20 MPa required for optimal SSB performance. Although future SSBs may operate at reduced pressures through the adoption of low-strain electrodes or engineered architectures that minimize changes in the molar volume of Li during cycling, a slight pressure above atmospheric levels may still be necessary to ensure satisfactory performance. Under such elevated pressures, once the cell temperature exceeds 180 °C, Li metal can deform and flow along the circumference of the SE, potentially initiating premature short circuits and accelerating the onset of thermal runaway. Additionally, the higher pressures present at elevated temperatures can alter the morphology and chemical composition of the decomposition products formed at the electrode/electrolyte interface. Current cell-level thermal stability studies often employ a low cathode loading (1–2 mAh cm^–2) and thick SEs (200–500 μm). For SSBs to realize their theoretical capacity, high cathode loading (4–5 mAh cm^–2) and a thin SE (20 μm) must be implemented. Reducing the SE thickness can impact the thermal stability for three reasons: 1) in the case of nonuniform Li deposition, it reduces the distance for Li metal to propagate toward the cathode side, increasing the likelihood of short circuit; 2) it increases the nonhomogeneous current flux at the cathode interface, thereby raising internal heat generation due to polarization of the cathode particles under nominal operating conditions; and 3) the increased energy density exacerbates the risk of thermal runaway during sudden discharge events, such as those caused by internal short circuits. For instance, under ambient conditions, a rapid release of just 10% of the discharge energy from an SSB with an energy density of 400–500 Wh/kg can elevate the internal temperature above the melting point of Li metal, significantly accelerating the onset of thermal runaway. Therefore, future research should prioritize the thermal stability of practically-relevant SSB designs and investigate the impact of various operational parameters such as temperature, pressure, and current density on a cell’s thermal characteristics.

Current state-of-the-art strategies for Li–S focus on improving the cell’s electrochemical performance. − , As discussed in Sections and , very few studies have explored thermal runaway at the interface and cell levels. Future research, in addition to investigating the effects of additives that improve electrochemical performance, must consider the influence of operational parameters such as current density, temperature, and pressure on the cell’s thermal characteristics. It has also been demonstrated that the surface topography of the S cathode can influence the Li deposition morphology (Figure e and f). The increased surface roughness of the cathode intensifies the Li-ion flux gradient at the anode, promoting dendritic Li growth and leading to early cell failure through short-circuiting. Accordingly, correlating the higher surface roughness of S cathodes, relative to that of oxide cathodes, with their thermal stability becomes critical. Conversely, for solid-state Li–S batteries, both the electrochemical behavior and thermal stability are still in the nascent stage. As an example, the charge–discharge profile is noticed to be fundamentally different in solid-state Li–S with the single voltage plateau and lack of polysulfide shuttling compared to their liquid counterparts. − The number of distinct phases formed during the charging and discharging processes of solid-state Li–S batteries, along with their electrochemical properties, remains unclear. Therefore, a fundamental understanding of the charge transfer mechanisms is essential before evaluating the thermal stability. Another important area that requires attention is the influence of the initial amorphous content of the S cathode on the thermal stability and degradation pathways, as it has been observed to severely affect the chemo-mechanical challenges observed in Li–S SSBs.

The advancement of thermally resilient LMBs is partly constrained by the limited availability of experimental data capturing cell-level behavior during thermal runaway. Though there are a few state-of-the-art experiments for LIBs, most focus on investigating the material-level thermal stability (Figure a, b, c, and d). ,− Most of these studies utilize in situ X-ray-based techniques as they offer a nondestructive way to study the thermal behavior of the cell. For instance, using operando X-ray tomography and high-speed radiography, the thermal runaway propagation was investigated for commercial cylindrical Li-ion cells wherein graphite was the working anode and NMC was the cathode. However, these techniques are not easily transferable to LMBs due to the higher heat release rate, reactivity of Li metal with X-rays, and the buried nature of the interphase, especially for SSBs. Consequently, future research should focus on implementing new techniques, such as soft X-ray, nuclear magnetic resonance, and neutron spectroscopy, to track the behavior of cell components (Figure e and f). ,− Combining multiple in situ techniques may also be necessary to overcome the individual limitations. Accurate experimental data will enhance the theoretical understanding of different thermal degradation pathways and contribute to the development of better safety prediction models.

26.

(a) Experimental setup integrating in situ XRD and MS to track degradation and gas evolution during heating of lithiated graphite anode. (Adapted with permission from ref . Copyright 2021 Springer Nature). (b) Operando X-ray tomography of an NMC/graphite Li-ion cell during thermal runaway, with inset radiographs capturing key stages of the event. (Adapted with permission from ref . Copyright 2015 Springer Nature). (c) Illustration of an operando X-ray spectroscopic cell designed to observe electrode morphology changes during cycling. (d) Surface SEM images of different cathodes: NMC, LFP. (Reproduced with permission from ref . Copyright 2013 Springer Nature). Operando Neutron depth profile during the plating/stripping process of Li half-cell (Li/LiPF₆-EC/DMC/Cu) depicting (e) the distribution of Li as a function of depth (from Cu surface) and time, and (f) Li-ion areal concentration across the thickness of Li deposition. (Reproduced with permission from ref . Copyright 2019 American Chemical Society).

6. Conclusion and Outlook

LMBs are rapidly emerging as leading candidates for next-generation energy storage due to their high specific capacity and energy density. While these characteristics make LMBs attractive for high-performance applications, the inherent chemical reactivity of Li metal introduces significant thermal safety concerns. This review has provided a comprehensive mechanistic analysis of thermal instability across various LMB chemistries, highlighting key pathways that dictate their thermo-electrochemical response and safety. The following subsections summarize insights and future directions for improving the thermal stability and safety of different LMB systems.

6.1. Lithium Metal Battery with Liquid Electrolyte

Although the thermal response of LE-based LMBs has been extensively studied at the materials and interface levels, cell-level thermal behavior remains underexplored. While current strategies largely focus on electrolyte formulation and additives, they often overlook how these modifications affect thermo-electrochemical stability in full-cell configurations. For instance, electrolytes incorporating thermally stable additives may still initiate exothermic reactions due to unstable interphases formed during cycling. Such factors underscore the critical need for holistic approaches that link interfacial dynamics to full-cell thermal stability.

Thermal instability in LE-based LMBs often initiates at the anode interphase and propagates through O₂ release from the cathode, interfacial decomposition, and chemical crosstalk between electrodes (Figure ). Even with thermally stable electrolytes and robust SEI formation, exothermic reactions between Li metal and oxygenated species (e.g., O₂, CO₂) released from layered cathodes can trigger thermal runaway. Surface impurities and interfacial instabilities further exacerbate the thermal runaway severity. Promising mitigation strategies include minimizing parasitic reactions at the anode and suppressing reactive gas evolution at the cathode. This may be achieved by employing low-solvation, high-flash-point electrolytes that stabilize the SEI, combined with O₂-capturing cathode coatings that transform exothermic processes into endothermic ones.

27.

Qualitative Sankey diagram, illustrating the intrinsic and extrinsic factors, along with coupled mechanistic interactions among constituent materials, that govern the thermal instability and safety response of LMBs.

Beyond material-level considerations, extrinsic factors, such as C-rate, temperature, applied pressure, and calendar aging, critically influence thermal stability (Figure ). For instance, temperature and pressure affect Li deposition morphology, which in turn impacts reactivity and exothermic behavior with LEs. Calendar aging promotes interphase thickening, potentially altering thermal response. While the presence of dead Li is well-known to hinder electrochemical performance, its role in thermal instability remains insufficiently understood. Bridging these multiscale mechanistic interactions with cell-level thermal behavior remains a key area for future research.

6.2. Solid-State Lithium Metal Battery

Despite their higher decomposition temperatures, SE-based LMBs still face significant safety concerns, with recent studies revealing multiple thermal instability pathways. Interfacial instability between Li metal and SEs is a major contributor to thermal runaway in SSBs. Future research must systematically investigate the complex interplay of intrinsic factors, such as Li morphology, microstructure, purity, and crystal orientation, with extrinsic factors like temperature, pressure, and current density, which strongly dictate Li plating/stripping behavior and its influence on thermo-electrochemical response (Figure ).

On the cathode side, most current studies focus on traditional layered oxide cathodes, where release of the O₂ at high temperatures can trigger exothermic reactions with the SE interphase. However, advanced cathode architectures, while electrochemically promising, may exhibit increased thermal reactivity due to interactions with other composite components (e.g., binder, carbon additives). These reactions can lead to the release of reactive gases, further elevating the risk of thermal runaway (Figure ). Investigating dense cathode architectures composed entirely of active material along with the impact of grain structure and particle orientation may reveal new insights into improving thermal stability. Additionally, external factors such as moisture exposure and applied pressure can alter the cathode properties and should be accounted for in thermal safety assessments. Heterogeneities in cathode microstructure can induce nonuniform current and heat distributions, further complicating thermal behavior under operational extremes (e.g., fast charging, high-power discharge). To ensure practical relevance, future studies should move beyond coin and pellet cells, focusing instead on pouch and prismatic formats with thinner SEs and higher cathode loadings.

6.3. Lithium Metal Battery with Sulfur Cathode

In Li–S batteries, most research has concentrated on improvements in electrochemical performance with thermal runaway mechanisms receiving comparatively less attention. Recent findings indicate that Li–S cells can undergo thermal runaway via direct redox reactions between S cathodes and Li metal anodes, even in the absence of O₂ evolution from the cathode. Interestingly, suppressing LiPSs may negatively impact the thermal stability. For example, some studies report that Li–S cells resist thermal runaway during nail penetration or internal shorting due to the formation of a protective Li₂S layer on the nail surface.

These observations underscore the need to evaluate the thermal implications of additives such as LiNO₃ and redox mediators, which can improve cycle life but also generate gaseous byproducts and increase internal pressure. In all-solid-state Li–S batteries, efforts have largely centered on deciphering their unique electrochemical signatures compared to LE-based systems. Although notable progress has been made, the nature of intermediate or metastable phases formed during cycling and their role in the observed low CE remains insufficiently understood. These phases may substantially impact the thermal response of solid-state Li–S batteries and thus warrant detailed investigation. Furthermore, the reduced CE is partially attributed to active material loss resulting from irreversible structural changes in the S cathode. Incorporating this phase heterogeneity into thermal analyses is crucial for developing a comprehensive understanding of degradation mechanisms and thermal runaway pathways in solid-state Li–S systems.

In conclusion, achieving a mechanistic understanding of thermal stability in LMBs requires overcoming substantial challenges, particularly those associated with characterizing buried interphases and interfacial reactions under dynamic conditions. Addressing these challenges calls for the development of advanced physics-based modeling, in situ, and operando diagnostic tools capable of capturing real-time thermal, electrochemical, and morphological evolution under nominal and abuse conditions. Coupling such techniques with machine learning and data-driven analytics will accelerate pattern recognition, enhance predictive modeling, and guide the rational design of thermally robust LMBs. These efforts will pave the way for the safe deployment of next-generation LMBs in diverse applications, including electric vehicles, electric aviation, and defense systems.

Biographies

Kausthubharam is a doctoral candidate in the School of Mechanical Engineering at Purdue University. He received his M.S. in Mechanical Engineering from Seoul National University, Korea, in 2023. His research focuses on interrogating the thermo-electrochemical interactions that drive interfacial instabilities and gradients in next-generation chemistries.

Bairav S. Vishnugopi is a Research Assistant Professor in the School of Mechanical Engineering at Purdue University. His research focuses on understanding the electro-chemo-mechanical, transport, and thermal interactions in electrode architectures and interfaces for energy storage. He has coauthored more than 65 journal publications spanning different battery chemistries and applications.

Anuththara S. J. Alujjage is a doctoral candidate in the School of Mechanical Engineering at Purdue University. She received her B.S. in Mechanical Engineering from the Northern Arizona University in 2020. Her research focuses on thermo-electrochemical analytics to understand degradation and safety mechanisms in lithium-ion and next-generation batteries.

Vinay Premnath is the Director of Research with focus on energy storage safety at the Electrochemical Safety Research Institute (UL Research Institutes). Prior to joining UL Research Institutes, Vinay was a Principal Engineer at Southwest Research Institute. Vinay received his M.S. degree from the University of Minnesota, Twin Cities.

Wan Si Tang is the Director of Research at the Electrochemical Safety Research Institute (UL Research Institutes). She has 15+ years of experience in novel functional materials for energy storage and conversion and advanced battery manufacturing. Wan Si is the recipient of the 2025 ECD Jubilee Global Diversity Award (https://wansitang.wordpress.com/).

Judith A. Jeevarajan is the Vice President and Executive Director for the Electrochemical Safety Research Institute at UL Research Institutes (ULRI). With more than 29 years of experience, she specializes in battery safety for lithium-ion and sodium-ion cells and modules, recycling, thermal runaway, fire, smoke, particulate emissions, and fire suppressants.

Partha P. Mukherjee is a University Faculty Scholar and Professor of Mechanical Engineering at Purdue University, specializing in mesoscale modeling and experimental analytics of transport, chemistry, microstructure, and interface interactions in energy storage and conversion. He has published >250 journal articles on different battery chemistries and energy conversion systems (https://engineering.purdue.edu/ETSL/).

CRediT: * Kausthubharam conceptualization, data curation, formal analysis, investigation, methodology, visualization, writing - original draft, writing - review & editing; Bairav Sabarish Vishnugopi conceptualization, data curation, formal analysis, investigation, methodology, writing - original draft, writing - review & editing; Anuththara S. J. Alujjage formal analysis, investigation, methodology, writing - original draft, writing - review & editing; Vinay Premnath formal analysis, investigation, methodology, writing - review & editing; Wan Si Tang formal analysis, investigation, methodology, writing - review & editing; Judith A Jeevarajan conceptualization, formal analysis, funding acquisition, investigation, methodology, supervision, writing - review & editing; Partha P. Mukherjee conceptualization, formal analysis, funding acquisition, investigation, methodology, resources, supervision, writing - original draft, writing - review & editing.

The authors declare no competing financial interest.