In my investigation into the safety performance of energy storage systems, the thermal behavior of lithium-ion batteries under abusive conditions remains a critical area of focus. The widespread adoption of lithium-ion batteries, driven by their high energy density and long cycle life, brings to the forefront the inherent risks associated with their failure. Among various abuse scenarios, overcharging represents a significant and realistic hazard, often arising from battery management system malfunctions or cell inconsistencies. The global increase in ambient temperatures due to climate change further complicates this safety landscape, as elevated environmental conditions can accelerate internal chemical kinetics. Therefore, I undertook a systematic study to understand how ambient temperature modulates the thermal runaway characteristics of a commercial square ternary lithium-ion battery during a constant-current overcharge process. My goal was to delineate the progression of failure, identify key characteristic parameters, and quantify their sensitivity to ambient temperature changes, thereby providing foundational data for developing adaptive safety management strategies.
The core of my experimental methodology involved subjecting identical lithium-ion battery samples to a severe overcharge test under controlled ambient conditions. The battery under test was a commercially available square aluminum-cased cell with a nominal capacity of 14 Ah and a nominal voltage of 3.7 V, utilizing a LiNiCoMnO2 (NCM523) cathode and a graphite anode. All cells were initially conditioned to 100% State of Charge (SOC) using a standard charge-discharge cycle to ensure consistency. The SOC during overcharge was calculated using the formula:
$$ SOC(t) = 100\% + \frac{1}{C_n} \int_{0}^{t} I_{charge} \, dt $$
where $C_n$ is the nominal capacity (14 Ah) and $I_{charge}$ is the constant charging current.
I placed the battery inside an Accelerating Rate Calorimeter (ARC), which served as a tightly controlled environmental chamber. The ARC allowed me to precisely set and maintain the desired ambient temperature ($T_{amb}$) for each test. The battery was charged at a constant current of 1.5C (21 A) until thermal runaway occurred. I instrumented the battery with multiple thermocouples to monitor surface temperatures at strategic locations: near the top negative terminal (T1), near the top positive terminal (T2), on the top of the side surface (T3), at the center of the largest side surface (T4), and at the bottom of the side surface (T5). Voltage across the terminals and pressure inside the ARC chamber were recorded continuously. Following thermal runaway, gases were collected for compositional analysis using a multi-component gas analyzer. I repeated this procedure for three distinct ambient temperatures: 30°C, 40°C, and 50°C.
Stage-wise Analysis of Overcharge-Induced Thermal Runaway
Analyzing the data from the test at 30°C ambient temperature allowed me to deconstruct the overcharge process into three distinct stages, demarcated by key events: Vent Opening (Lvo) and Thermal Runaway Trigger (Ltr). This staging is crucial for understanding the sequence of internal failures.
Stage 1: From Charge Initiation to Vent Opening (Pre-Lvo). This stage encompasses normal charging, lithium plating, and the initial exothermic side reactions. The voltage rose from the initial 4.2 V to a peak value (e.g., 5.93 V at 30°C) due to increasing overpotential and cell impedance. The temperature increase was initially gradual, driven by Joule heating and reaction heat. A significant observation was the divergence in surface temperatures; T1 and T2 (near the tabs) were consistently higher than T3, T4, and T5, primarily due to concentrated current and Joule heating at the tab connections. As the cell approached its venting point, the Solid Electrolyte Interphase (SEI) layer on the anode began to decompose exothermally. The decomposition of lithium alkyl carbonates (e.g., from EC) can be represented as:
$$ (CH_2OCO_2Li)_2 \rightarrow Li_2CO_3 + C_2H_4 + CO_2 + \frac{1}{2}O_2 $$
Simultaneously, plated lithium metal reacts vigorously with the electrolyte solvents (EC, DMC, DEC, PC):
$$ 2Li + C_3H_4O_3 (EC) \rightarrow Li_2CO_3 + C_2H_4 $$
These reactions generate heat and gases (e.g., CO2, C2H4), increasing internal pressure. Towards the end of Stage 1, the voltage often exhibited a slight decrease or plateau, attributed to lithium consumption and cathode structure damage counteracting the rising impedance.
Stage 2: From Vent Opening to Thermal Runaway Trigger (Between Lvo and Ltr). The rupture of the current interrupt device or vent marks the beginning of this stage. The sudden release of high-pressure gas causes an immediate, sharp drop in cell voltage as internal resistance momentarily decreases. Following this event, the voltage often resumes a steep rise. The release of hot gases elevates the temperature measured at the top of the cell (T1, T2). The influx of air (or ambient atmosphere) into the cell and continued overcharge accelerate internal reactions. This stage is characterized by rapid electrolyte vaporization and the onset of internal short circuits, leading to a final, sharp voltage spike immediately before full thermal runaway.
Stage 3: Full Thermal Runaway and Post-Thermal Runaway (Post-Ltr). This is characterized by the catastrophic failure of the cell. The internal short circuit becomes severe, leading to an enormous, rapid heat release. Cell voltage collapses to near zero. Surface temperatures, particularly at the top, skyrocket to their maximum values (e.g., T1 exceeding 500°C) as flame and ejected materials are expelled. The pressure inside the chamber reaches its peak. The cell casing is typically breached, and the internal components are largely consumed or ejected.
Quantifying the Impact of Ambient Temperature
My comparative analysis across the three ambient temperatures revealed profound and sometimes non-intuitive influences on the thermal runaway process of the lithium-ion battery. The data is summarized in the table below, which presents key characteristic values extracted from each test.
| Characteristic Parameter | Symbol | 30°C | 40°C | 50°C |
|---|---|---|---|---|
| Time to Thermal Runaway | $t_{TR}$ | 1407 s | 1335 s | 1273 s |
| Voltage at 1153s (Stage 1) | $V_{S1}$ | 5.8 V | 6.0 V | 6.5 V |
| T4 Temp. at 1153s (Stage 1) | $T_{4,S1}$ | 58.5 °C | 74.2 °C | 87.4 °C |
| T1 Temp. at Thermal Runaway Trigger | $T_{1,trig}$ | 99.7 °C | 103.2 °C | 107.0 °C |
| T4 Temp. at Thermal Runaway Trigger | $T_{4,trig}$ | 83.9 °C | 84.4 °C | 85.4 °C |
| Max. T1 Temp. (Post-Thermal Runaway) | $T_{1,max}$ | 492.3 °C | 520.2 °C | 543.9 °C |
| Max. T4 Temp. (Post-Thermal Runaway) | $T_{4,max}$ | 188.9 °C | 158.7 °C | 149.8 °C |
| Peak Chamber Pressure | $P_{max}$ | 0.235 MPa | 0.256 MPa | 0.270 MPa |
1. Acceleration of Failure Timeline: The most direct effect was on the time to thermal runaway ($t_{TR}$). A higher ambient temperature provided a higher initial thermal energy state, accelerating all subsequent exothermic reactions. The time decreased monotonically from 1407 s at 30°C to 1273 s at 50°C. This underscores the increased risk posed by hotter operating environments for a lithium-ion battery subjected to an overcharge fault.
2. Modulation of Voltage and Temperature Behavior: During Stage 1, both the cell voltage and the side surface temperature (T4) at a fixed time (1153s) were significantly higher at higher $T_{amb}$. The voltage increase suggests accelerated side reactions and impedance growth. The temperature rise follows from the Arrhenius law, where reaction rates $k$ increase exponentially with temperature:
$$ k = A e^{-E_a/(R T)} $$
where $E_a$ is the activation energy, $R$ is the gas constant, and $T$ is the absolute temperature. Interestingly, the temperature at the point of thermal runaway trigger ($T_{4,trig}$) was remarkably similar (~84-85°C) across all ambient temperatures. This suggests that the final internal short circuit mechanism is triggered at a critical internal temperature state that is relatively independent of the starting ambient condition for this specific lithium-ion battery. In contrast, $T_{1,trig}$ showed a greater increase with $T_{amb}$, influenced more by the ambient condition and hot gas release.
3. Divergence in Post-Thermal Runaway Temperatures: A counter-intuitive finding was the behavior of the maximum temperatures after thermal runaway. While the top temperature (T1) increased with $T_{amb}$ (from 492.3°C to 543.9°C), indicating a more violent ejection event, the side center temperature (T4) *decreased* (from 188.9°C to 149.8°C). I attribute this to the dynamics of casing rupture. At higher $T_{amb}$, the reactions proceed faster and more violently, leading to a quicker and more catastrophic breach of the cell casing. This rapid ejection removes the internal heat source from the core of the cell body before significant energy can be conducted to the side surface where T4 is located, resulting in a lower measured $T_{4,max}$.
4. Pressure and Gas Generation: The peak pressure inside the ARC chamber increased with ambient temperature, confirming that the overall gas-producing reactions are more intense. Gas analysis revealed significant shifts in composition, summarized below:
| Gas Species | Volume Fraction (30°C) | Volume Fraction (40°C) | Volume Fraction (50°C) |
|---|---|---|---|
| CH4 | 17,011 ppm | 14,014 ppm | 13,677 ppm |
| CO2 | 1,286 ppm | 1,604 ppm | 1,859 ppm |
| CO | 1,136 ppm | 1,428 ppm | 1,535 ppm |
The volume fraction of CH4, primarily from the reduction of electrolyte by plated lithium and hydrogen (from binder reaction), decreased with increasing $T_{amb}$. Conversely, the fractions of CO and CO2 increased. CO2 can be generated from SEI decomposition, carbonate solvent combustion, and reactions with PF6 decomposition products (e.g., HF, POF3):
$$ ROCO_2Li + HF \rightarrow ROH + CO_2 + LiF $$
$$ \frac{5}{2}O_2 + C_3H_4O_3(EC) \rightarrow 3CO_2 + 2H_2O $$
The increase in CO and CO2 suggests a shift towards more complete decomposition/combustion pathways or different reaction dominances at elevated initial temperatures, which has direct implications for toxicity hazards during a lithium-ion battery failure.
Sensitivity Analysis of Characteristic Parameters
To effectively design a battery management system (BMS) that can adapt to environmental changes, it is essential to understand not just the trend but the sensitivity of different monitoring parameters to ambient temperature. I defined a normalized sensitivity for a parameter $Y$ between two ambient temperatures as:
$$ S_{Y}^{T1 \rightarrow T2} = \frac{(Y_{T2} – Y_{T1}) / Y_{T1}}{(T_{amb, T2} – T_{amb, T1}) / T_{amb, T1}} $$
A larger absolute value of $S$ indicates higher sensitivity. Analyzing my data reveals critical insights:
- High Sensitivity Parameters: The voltage during Stage 1 ($V_{S1}$) showed high sensitivity. A 33% increase in $T_{amb}$ (from 30°C to 40°C) led to a 3.4% increase in voltage, and a further increase to 50°C caused a 12% total increase from the 30°C baseline. This makes cell voltage a potentially strong early indicator for adjusting charge algorithms in real-time based on ambient temperature.
- Moderate to Low Sensitivity Parameters: The time to thermal runaway ($t_{TR}$) and the side temperature $T_{4,S1}$ are moderately sensitive, decreasing and increasing, respectively, with $T_{amb}$. The thermal runaway trigger temperature ($T_{4,trig}$) showed very low sensitivity, reinforcing its role as a more intrinsic threshold.
- Negative Sensitivity: The post-thermal runaway side temperature $T_{4,max}$ exhibited negative sensitivity, decreasing with increasing $T_{amb}$. This is a crucial finding, as it indicates that a cooler side surface measurement after an event does not necessarily imply a less severe failure when the ambient temperature is high.
Furthermore, the rate of change of sensitivity itself is informative. For several parameters like $t_{TR}$, $T_{1,max}$, and $T_{4,max}$, the magnitude of change between 40°C and 50°C was smaller than between 30°C and 40°C for the same 10°C interval, indicating a potential nonlinear saturation effect or that the process becomes increasingly dominated by the internal reaction kinetics rather than the initial condition as temperature rises.
Conclusions and Implications for Lithium-Ion Battery Safety
My comprehensive investigation into the effect of ambient temperature on the overcharge-induced thermal runaway of a commercial lithium-ion battery yields several definitive conclusions with important practical implications:
1. Elevated ambient temperature unequivocally accelerates the entire failure process of a lithium-ion battery under overcharge, reducing the time to thermal runaway and increasing the severity as indicated by higher peak pressures and top-surface temperatures. This translates to a significantly reduced window for intervention in hotter climates or environments.
2. The thermal runaway pathway is chemically modulated by temperature. Higher initial temperatures shift gas generation profiles, reducing CH4 yields while increasing the production of toxic and asphyxiant gases like CO and CO2. This alters the secondary hazard profile following a lithium-ion battery failure.
3. Different monitoring parameters exhibit vastly different sensitivities and even directional responses to ambient temperature change. Key findings include:
– Voltage during overcharge is highly sensitive and increases with $T_{amb}$.
– The side-surface temperature at thermal runaway trigger is relatively invariant.
– The maximum side-surface temperature measured *after* thermal runaway decreases with increasing $T_{amb}$, which is counter to intuition.
The critical implication for the safety of lithium-ion battery systems is that static, fixed thresholds for voltage, temperature, or pressure alarms are suboptimal and potentially unsafe. A BMS must incorporate ambient temperature as a direct input to dynamically adjust its safety algorithms. For instance:
– When high ambient temperature is detected, the maximum allowed charging current (C-rate) should be derated.
– Voltage safety limits during charging could be made slightly more restrictive, or their rate-of-change triggers could be tightened.
– The interpretation of temperature measurements, especially from different cell locations, must be contextualized with the ambient temperature. A lower-than-expected side temperature in a hot environment could be a sign of an imminent, very violent failure rather than a benign condition.
This work underscores that ensuring the safety of lithium-ion batteries, particularly under abuse conditions like overcharge, requires a nuanced, context-aware approach. As operational environments become more varied and extreme, adaptive safety strategies rooted in a deep understanding of parameter sensitivity will be indispensable for the reliable and safe deployment of lithium-ion battery technology across all applications.