The Role of Gas Sensors Based on Metal Oxide Semiconductor Nanomaterials in Thermal Runaway Early Warning for Battery Safety: A Review

With the widespread adoption of lithium-ion batteries (LIBs) in electric vehicles and energy storage systems, thermal runaway (TR) remains a critical safety bottleneck. While traditional monitoring approaches relying on voltage, temperature, and pressure often suffer from signal lag or only detect irreversible damage, the monitoring of characteristic gas leakage has emerged as the most significant early warning indicator. This review systematically summarizes the application of nanostructured metal oxide semiconductor (MOS) gas sensors for TR monitoring, highlighting their advantages of nanoscale sensing properties in high sensitivity, fast response speed, and suitability for pack-level integration and summarizing the efforts made by researchers to address the inherent limitations of its practical applications, such as high operating temperature, high cross-sensitivity, signal drift, susceptibility to environmental interference, and the complexity of multigas coupling in real-world battery failure scenarios. First, the review elucidates the multistage evolution of TR and analyzes the gas generation mechanisms driven by different battery chemistries, abuse conditions, and state of charge (SOC), providing a theoretical basis for target gas selection. Subsequently, recent progress in nanoscale MOS gas sensors is critically evaluated based on specific target gases, including H 2, CO, CO 2, hydrocarbons, and early-stage electrolyte vapors. The sensors achieve ultralow detection limits down to ppb-level for H 2 and electrolyte vapors, with fast response/recovery times of ∼1–30 s. The review then comprehensively analyzes strategies to improve sensor performance from four perspectives: nanomaterial engineering, device optimization, algorithmic improvements, and sensor arrays. Finally, this review prospects the cross-sensitivity of gases in TR and other environmental disturbances in complex battery pack environments, the contradiction between the high-temperature operation and high power consumption of nanomaterials based MOS gas sensors and their integration into the battery environment, the importance of intelligent algorithm prediction for the detection of the entire TR process, and the remaining challenges, providing a prospect for the development of intelligent multiphysics fusion early-warning systems.