Let's start with the conclusion:
A square LiCoO2-graphite battery with a nominal capacity of 6.8 Ah was externally heated in an oven. This study included cycle-aged cells, non-cycled cells stored at 60°C, and non-cycled cells stored at room temperature. Both working and non-working (failed) cells were investigated.
When heated externally, all batteries experience thermal runaway, releasing smoke and gases. In approximately half of the working cells, within about 15 seconds of thermal runaway, the gases accumulated in the oven ignite, causing a gas explosion accompanied by a significant smoke release process.
Regardless of whether the battery cell has been cycled or not, it does not affect the occurrence of gas explosions, which occur at all cycle aging levels from 0 to 300 full deep cycles.
The gas was analyzed using FTIR. HF gas was detected regardless of whether a flame was present.
Another HF precursor, POF3, which releases potentially toxic gases, was also detected along with HF. The harmful gas CO was also detected. The thermal runaway temperature was approximately 190°C and showed a weak correlation with the number of cycles during the aging process. Within the test cycle range of 0 to 300 cycles, the aging stage with the fewest toxic gases was between 100 and 200 cycles.
Three failed cells were tested. One cell failed suddenly during cycling after 229 cycles, and two other cells that had not been cycled but had been stored at 60°C for 10 months also failed. The failed cells also entered thermal runaway; however, their response to the increase in thermal runaway temperature and the rate of temperature increase was significantly lower. These batteries did not exhibit sparks, combustion, or gas explosions.
1. Introduction
Compared to other types of batteries, lithium-ion batteries generate more heat, resulting in higher risks of gas emissions, explosions, and fires. These risks are far from being fully understood, but it is possible to improve system safety through research and accident analysis. The type and severity of the risks depend on the specific application and the size of the battery system. Due to the propagation of battery and module failures, the consequences of failures can increase significantly with increasing battery system size.
Lithium-ion batteries contain all three essential parts of the flame triangle: the heat source/igniter, the combustible material, and oxygen. Furthermore, once overheated, typically starting from 70°C to 120°C, lithium-ion batteries begin to swell and release gases (venting). These vented gases are flammable and toxic. If the temperature is high enough, reaching 150°C to 200°C, the battery's self-heating enters an accelerated phase, and thermal runaway (TR) may occur. The term thermal runaway initiation temperature refers to the temperature at which the exothermic reaction begins and ultimately leads to thermal runaway, while the thermal runaway temperature refers to the very rapid temperature rise that triggers it. Thermal runaway is usually accompanied by the release of large amounts of smoke and may be accompanied by battery casing rupture, combustion, or a gas explosion. Therefore, there are two important types of explosions in the thermal runaway process: battery casing explosion and gas explosion of the flammable vented gases mixed with air. Cylindrical and rigid prismatic batteries can experience high internal pressures and are therefore designed to release gases through built-in battery safety valves; however, if the venting fails, extreme pressure can build up inside the battery, causing the battery casing to explode. There are two types of such explosions: one is an explosion inside the battery, and the other is an explosion caused by the delayed ignition of a mixture of flammable gas and air that has accumulated in a closed or semi-closed casing. The consequences of a flammable gas explosion can be far more severe than those of a battery explosion.
The emitted gases can contain products of solvent evaporation and decomposition, such as CO, CO2, H2, and CH4. Besides CO, a wide variety of toxic compounds, including fluoride gases, can be released. Hydrogen fluoride (HF) has received the most attention as it is a highly toxic gas. Few published studies report the amount of HF released during the abuse of commercial lithium-ion batteries, or the amount released during electrolyte combustion. Fluorine in batteries originates from lithium salts, such as LiPF6, as well as from electrode binders, such as PVdF, electrode materials and coatings, such as fluorophosphates and AlF3 cathode coatings, and fluorinated additives such as flame retardants. Battery safety is highly complex, and a holistic perspective is crucial. For example, while the risk of thermal runaway can be reduced by introducing an AlF3 coating, the risk of toxic fluoride gas emissions and gas explosions may be increased. Therefore, overall safety is difficult to assess, depending on the size and condition of the battery, and improvements to one parameter may actually worsen overall safety.
There are many different types of abuse tests, with external heating being common. Several types of external heating methods are applicable to lithium-ion batteries, such as heating in an oven, heating by IR radiation, heating films or other heaters, using a heating rate calorimeter (ARC) or other types of instruments in a sealed chamber. To date, much research has been conducted on new cells, but few studies have investigated the impact of aging on safety. Component performance may change during aging, but the practical requirement is to maintain a high level of battery safety throughout the battery's lifespan. Aging typically takes the form of calendar and cycle aging. To shorten testing time, the cells are often stored and cycled at elevated temperatures, such as 35-55°C; however, measurements at these temperatures are not entirely the same as those obtained when used at ambient temperatures, such as 20°C, because other decomposition reactions may occur. The aging process of lithium-ion batteries is non-linear and complex and is not yet fully understood. For example, during aging, the solid electrolyte interface (SEI) layer changes, and the SEI plays a crucial role in the early stages of thermal runaway. Studies have used calorimetry to describe the evolution of this SEI modification, and used XRD, XPS, SEM and Raman spectroscopy to analyze the surface and describe three important stages of thermal runaway.
An experiment using ARC testing to study the thermal stability of calendar-aged Sony 18650 batteries found that the aging cells start to release heat at a temperature as high as 70°C, indicating that the aging cells exhibit a higher heat release start temperature.
Another study investigated 0.75Ah non-commercial graphite/lithium cobalt oxide (LCO) lithium-ion batteries after 10 and 200 cycles, finding that thermal safety decreased after 200 cycles in a needle penetration abuse test.
Some researchers studied 2Ah graphite/LMO-NMCLi-ion 18650 batteries stored at 60°C for 36 weeks and found that the 36-week-old cells had lower exothermic reactions and thermal runaway initiation temperatures in ARC tests.
Conversely, other studies have investigated 4.6Ah graphite/LMO lithium-ion batteries stored at 55°C for 10 to 90 days and found that the onset temperature of self-heating and thermal runaway increases with the increase of aging.
Another experiment investigated the effect of a 1.5Ah graphite/LMO-NMC high-power Li-ion 18650 battery on the thermal response of cycle aging in ARC testing. It was found that the first exothermic response and the onset temperature of thermal runaway were significantly reduced, with the onset temperature as low as 30.7℃. Lithium plating was also found on the anode of the battery after 1C cycling at -10C.
A team investigated the safety of novel and cycle-aged graphite/NMC18650 cells using 1C ARC testing at 0°C to 70% State of Health (SOH). The aged cells exhibited reduced thermal safety, with autothermal initiation temperatures as low as 30°C and earlier thermal runaway. The same authors also investigated safety through nail penetration abuse testing and found that the aged cells exhibited delayed but more severe thermal runaway. Generally, low-temperature cycling of the anodic lithium plating and charging with excessively high currents increase the risk of lithium-ion batteries.
This study investigated the safety of lithium-ion cells stored at 20C and 60C, as well as cells with 100, 200, or 300 C/2 deep cycles. All cells were of the same type: a commercially available 6.8Ah graphite/LiCoO2 lithium-ion battery. Safety was assessed through abuse testing using external heating (oven) combined with FTIR gas measurements. An ARC test was performed to compare the safety assessment methods.
2. Experiment
2.1 Testing the battery cell
These batteries were all from the same batch of commercially available lithium-ion batteries, with nominal capacities of 6.8 Ah and 3.75 V, LCO cathodes, graphite anodes, polymer separators, and a prismatic shape. See Table 1 for detailed cell parameters. The batteries contain fluorine due to the presence of LiPF6 salt in the electrolyte, but other parts of the battery may also contain fluorine; see the examples in the introduction. It should be noted that other potential sources of fluorine in the cells were not analyzed in this study.
2.2 Electrical Characteristics
Four-line electrochemical impedance spectroscopy (EIS) with a frequency range of 100 kHz to 5 mHz, a 60-point logarithmic distribution, and an amplitude of 0.1 A was performed using MetrohmAutolabPGSTAT302N and MetrohmNovav1.11 software in constant current mode.
Table 1 shows the datasheet for a commercial Li-ion battery, including the cell specification sheet, TG-FTIR electrolyte analysis, and separator analysis from DSC. The battery was stored in a Faraday cage at ambient temperature, approximately 20°C. Sensor and current measurement cables were twisted and separated in opposite loops to minimize interference.
The capacity of each battery was measured using a multi-channel Digatron battery tester or a MetrohmAutolab with a Booster20A module. Battery capacity measurements were performed using voltage limits of 2.50V and 4.20V, a current of 1.4A (approximately C/5), and a cut-off charging current of 0.05A. After the first charge, three complete discharge-charge cycles were applied. Discharge capacity was measured using the first of the three cycles before aging, and the battery capacity was determined using the discharge capacity of the third cycle after aging.
Prior to EIS measurement, the battery was fully charged as the first charge (100% SOC). The number of cycles described in this article does not include the three charge-discharge cycles used to measure battery capacity.
2.3 Aging Procedure
2.3.1 Cyclic Aging
The batteries were individually cycled using a Digatron battery tester, achieving 100% depth of discharge (DOD) between 4.20V and 2.50V. A current of 3.4A (C/2) was used for both charging and discharging, with a charge cutoff current of 0.34A (C/20). These batteries were subjected to forced convection cooling at an ambient temperature of 21°C. Each battery had a temperature sensor mounted on its largest side.
2.3.2 Temperature Aging
Fully charged batteries can be stored in a 60°C oven for 10 months, which is the maximum permissible storage temperature according to the battery manufacturer's datasheet. The cells should be stored at 20°C before and after storage at 60°C.
2.3.3 Complete Cell Aging Process
First, the battery cells were left unused in their transport boxes at room temperature (approximately 20°C) for 12 months. Second, the batteries underwent their first charge, and the capacity and impedance of each cell were measured. Third, batteries selected for cycle aging were cycled for approximately two months (300 cycles). Fourth, capacity and impedance were measured, and the cells were stored at room temperature. Fifth, some non-cycled cells were stored in an oven at 60°C for 10 months.
The external heating abuse test was conducted approximately 2 years and 4 months after the production date. Therefore, all cells have the same calendar age, but during their lifespan, some cells have been cycled and others have been stored at 60°C for a period of time (10 days out of 28 months).
2.4 External Heating Abuse Test
2.4.1 General Settings
A total of 14 external heating abuse tests were conducted. The batteries were individually heated using a BinderFED115 thermostatically controlled oven with a 115L internal volume. The batteries were centered inside the oven and mechanically secured to a brick with wire (0.8mm diameter), as shown in Figure 1. One minute after the start of the test, the oven was adjusted to its maximum heating rate, and the temperature was set to 300°C. The total test time varied due to changes in environmental conditions and the eventual gas explosion.
The oven is custom-made with four 50mm diameter vents sealed with silicone plugs and an internal fan set to maximum speed to even out internal temperatures. The vent at the back of the oven is designed to close completely. However, this is not a perfect seal; it partially deformed during abuse testing. The oven door closed normally in the first test, but was secured with tape in subsequent tests after it opened during a gas explosion. A silicone plug on the top of the oven is installed rather loosely, serving as a pressure relief vent.
Between each test, the oven was gently cleaned/washed to minimize potential interference from sources such as particulate contamination. The glass doors and windows (triple glazing) were not mechanically broken, but were severely contaminated and etched, and therefore were replaced several times to achieve acceptable video quality.
Battery voltage and temperature were measured at 1 Hz using an Agilent 34902A reed multiplexer module. Battery voltage was measured via a K-type thermocouple cable screwed into a small drilled hole (0.8 mm diameter) in the electrical connector. The surface temperature of the Li-ion battery was measured using a K-type thermocouple with attached fiberglass tape (3M, Scotch electrical tape, 19 mm wide) at up to six locations T1-T6 (see Figure 1D), where T1-T4 measures the center temperature on each side, and T5-T6 are additional central sensors on the two largest surfaces. K-type thermocouples were also used to measure ambient (outside the oven) and oven temperature, the latter measured at two locations (Figures 1A and 1B). The testing process was video-recorded by a camera placed outside the oven door. In some tests, a second camera was also used, placed approximately 2 to 7 meters away from the oven. Cell thickness (150 mm long) and cell dimensions were manually measured using calipers and recorded at the largest dimensions, appearing at the center-to-center measurement locations.
3. Results and Discussion
3.1 Aging - Capacitance Decay and Impedance
The battery in Test 12 was intended to cycle for 300 cycles, but failed after 229 cycles and could not be recharged or discharged. The batteries in Tests 13 and 14 were initially fully charged and stored at 60°C for 10 months, after which the voltage dropped below 1V. The thickness of these batteries increased from 18.5mm to 21.3mm (approximately 15%), but the cell weight remained unchanged, indicating no leakage or venting. All other cells in this study had a thickness of 18.5mm before and after cycle aging.
Table 2 lists the capacity data before and after aging. SOH is the relative remaining capacity, calculated by dividing the current C/5 discharge capacity by the initial C/5 discharge capacity. After cycling, the cells reached the following SOH values: approximately 94% (100 cycles), 91% (200 cycles), and 89% (300 cycles). The end of life (at least regarding the life of a battery in its first use) is typically defined as approximately 70%–80% SOH. The cell parameter table shows that 600 cycles > 70% SOH (see Table 1 below), therefore, the tested cells were far from being fully aged. As shown in Table 2, the batteries in Tests 1 and 4 had lower initial discharge capacities because they were cycled 3 times before the capacity measurements. However, even though the cells in Tests 1 and 4 were cycled 3 times (see the notes in Table 2 for details), they are referred to here as 0-cycle cells.
Figure 2 shows the impedance measurements of batteries aged for different cycles. The impedance curve, Figure 2A, has the typical appearance of a lithium-ion battery, including overlapping semicircles of suppressed mid- and low-frequency peaks, corresponding to the cell internal resistance and connection impedance, SEI impedance, charge transfer effect, and mass transfer impedance. The intersection of the complex impedance plot with the real axis determines the average series impedance, as shown in Figure 2A, which is also the internal resistance of this type of cell; 13.2 mΩ for a new cell, increasing to 14.4 mΩ (a 9% increase) after 300 cycles. Figure 2B is a curve of phase angle versus the logarithm of frequency. Two peaks can be found in this graph, one above 0.1 Hz and one around 2 Hz. The low-frequency peak increases with cycle aging, while the second peak more or less disappears after a few cycles. In any case, a significant difference is detected after 3 cycles, thus the phase angle provides a new perspective on the effects of aging. For aged batteries (graphite/LCO) with the same electrode chemistry, the low-frequency semicircles in the impedance diagram are attributed to electrolyte oxidation at the cathode, thus potentially indicating that the rise in peaks above 0.1 Hz is also due to cathode oxidation in this case. This is likely because the battery is charged to a relatively high 4.20V cutoff voltage, although still within the parameters specified by the battery manufacturer.
3.2 Abuse of external heating
In tests 1-11, the cells were fully charged (100% SOC) and subjected to varying numbers of aging cycles, ranging from 0 to 300 cycles. The cells used in tests 12-14 were failed cells; therefore, their SOC could not be determined. The battery in test 12 failed during cycling after 229 cycles. The cells used in tests 13-14 had been stored at 60°C for 10 months, during which time they either self-discharged or failed, resulting in one cell having an OCV less than 1V, i.e., below 0% SOC.
3.2.1 Overview of Results
Table 3 lists the external abuse test results for 14 different aged cells, including working cells and failed cells. In all tests, the rate of temperature rise increased rapidly when the temperature reached the thermal runaway temperature, and all batteries experienced thermal runaway. For tests 1-11, after thermal runaway, there were short (less than one second) and typical burning, sparking, and jetting, examples of which are shown in Figure 3. In some cases, depending on the different stages of the fire, the cell burned for a longer time and with a larger flame, as shown in Table 3. Typically, subsequent fires were smaller, as shown in Table 3, indicating that there were one or more flames present for a longer period prior. Furthermore, the term "no open flame" is used to refer to a situation where the battery or its gases were not ignited. This does not take into account situations such as an initial brief short circuit/spark. The term "gas explosion" refers to the delayed ignition of accumulated combustible gases released from the battery mixed with air in the furnace, which, in the current case, resulted in a pressure wave that forced the furnace door to open. Gas explosions are a common phenomenon in combustion science, however, are not frequently discussed in the context of lithium-ion battery fires. In this study, as shown in Table 3, all tests for working cells were conducted after neither burning nor a gas explosion. Furthermore, regarding approximately half of the working cells and all of the aged cells, the gas explosion occurred about 30 seconds after combustion, followed by a small fire or spark for 20-50 seconds. For the failed cells, as shown in tests 12–14, the results were significantly different; video analysis did not reveal any sparks, jets, or gas explosions.
The release of a significant gas occurs almost simultaneously with the arrival of the thermal runaway temperature, rapidly increasing the rate of temperature increase.
Test equipment b, data logging, oven camera, and external heating power were interrupted for approximately 9 minutes (no data recording for 9 minutes and 39 seconds). The secondary camera outside the oven, primarily used to observe the important gas emission process and assumed to be for the TR time, was still operational (powered by the laptop battery).
c is determined by an auxiliary camera located outside the oven.
Regarding all tests, video analysis shows that upon reaching the thermal runaway temperature, the battery safety valve located at the top of the cell opens and releases a large amount of smoke, rapidly filling the oven space. The released smoke is typically white or light gray in color. If the battery safety valve fails to open, for example due to a malfunction or poor design, the battery casing could explode, a dangerous situation involving the risk of ejecting casing fragments. However, this did not occur in the current set of experiments.
Working cells lost more weight and expanded more (thicker) than failed cells. The average weight loss for working cells was 22.6%, compared to 17.0% for failed cells. The thickness of working cells increased from 18.5 mm to an average of 27.2 mm (a 47% increase), while the average thickness of failed cells after external abuse was 23.8 mm. The total test time varied, as shown in Table 3, resulting in different heating times. Some trends can be observed: both weight loss and thickness increase are functions of the number of cycles (more cycles result in less weight loss and greater dimensional expansion). These effects must occur within the minimum internal test time range (75 minutes, Test 6).
3.2.2 Temperature Results
Table 4 lists the temperature results of the external heating abuse test. The thermal runaway temperature values in Table 4 are defined as the temperatures at which a rapid temperature rise occurs. For working cells, the thermal runaway temperature is easily determined, while for failed cells, particularly test 12, it is less clear. Failed cells exhibit significantly higher thermal runaway temperatures, lower rates of temperature rise, and lower peak temperatures. Failed cells with 0 cycles, stored at 60°C for 10 months with a portion of their lifespan in tests 13-14, showed high reproducibility. Although significant changes must occur at the electrode interface, failed cells may still contain substantial amounts of flammable electrolyte. No cells with different SOC levels were studied; therefore, any potential similarities between cells with low SOC levels and failed cells could be confounded in the results.
Battery surface temperature sensors T1-T6 are generally reliable when the thermal runaway temperature is reached. For all tests except Test 13 (see notes in Table 4), the thermal runaway temperature value is calculated as the average of sensors T1 and T3. Above the thermal runaway temperature, the temperature recorded by the sensors will vary significantly, and sometimes the sensors may detach from the battery due to high temperatures, battery expansion, and eventual gas explosion. Therefore, another average, Tavg2, is used to determine the maximum average battery surface temperature, the corresponding temperature increment (ΔT), and the time length (step time Δt). Tavg2 is calculated using data from all available T1-T6 sensors, as detailed in Table 4. Available sensors are defined as those that have not been lost and are in contact with the cell surface. The results given in Table 4 will naturally vary due to the different numbers and locations of available battery temperature sensors. The thermal runaway temperature value can be better defined when the distribution of the highest temperature values is wider.
Prior to thermal runaway, the battery surface temperature sensors displayed relatively similar temperature values, while significant temperature differences emerged between the sensors after thermal runaway. Figure 4 shows the battery voltage, average temperature, and temperature measurements for Test 7, one of the few tests where all six temperature sensors were available throughout the testing process. The battery surface temperature change in Figure 4 is approximately 100°C. In other tests, localized cell surface temperature changes reached as high as approximately 300°C. For this type of measurement, it is crucial to use multiple cell surface temperature sensors and appropriate verification methods to obtain reliable temperature measurements. The large temperature differences on the cell surface can be explained by the anisotropy of the heat transfer process, the rapid and substantial heat generation, and the effects of exhaust and flames from the reaction process. During thermal runaway, internal battery heating may result in the highest temperatures near the battery center. The differences between in-plane and inter-plane heat transfer within the battery are substantial (anisotropic thermal conduction).
Due to temperature changes caused by phase transitions (e.g., separator melting) and mass losses (e.g., venting, flame), the thermal properties (e.g., thermal conductivity, specific heat capacity, density) of each material differ. The tested battery had an aluminum casing with high thermal conductivity, yet the temperature variation was significant. For other types of battery casings, such as pouch cells, an even greater temperature distribution is likely.
Figure 5A shows the average battery surface temperature before and during the early stages of thermal runaway, with all tests synchronized with the thermal runaway temperature time. A few minutes before thermal runaway appears on some curves, the temperature drops for several minutes, particularly noticeable on the blue line (100 cycles), presumably due to a temperature drop caused by gas release, specifically the second venting process in this case. Figure 5B shows the relationship between the thermal runaway temperature of the working cell and the number of aging cycles, showing a minimum between 100 and 200 cycles. This figure also shows the relationship between post-test weight loss and thickness and the number of cycles. In Test 11, a total power interruption occurred at the test site a few seconds before thermal runaway. Regardless, the significant venting time in Test 11 can be determined using an externally operated camera. The last recorded data point points to 188°C, with approximately 5 seconds between the observed gas release and its corresponding thermal runaway; although short, the battery temperature rise was relatively rapid. The temperature heating curves of Tests 10 and 11 match very well until the power outage.
The thermal runaway temperature values presented in this paper refer to the last temperature point before the rapid heating begins, as shown in Figure 5A. The thermal runaway temperature values given here are approximately 190°C for working cells and between 201 and 205°C for failed cells. Similar temperature values have been reported previously.
If external heating stops sometime before the thermal runaway temperature, the battery can still enter a thermal runaway state depending on the battery temperature, the battery's own heating rate, and environmental conditions (such as the battery cooling rate). However, the experimental method used in this work does not employ the method of pausing the heating step. Another commonly used method in the field of lithium-ion battery safety is the ARC test of the Heat Wait Search (HWS) procedure, in which the cell is heated with a highly sensitive heater, and if heat dissipation is paused, and under adiabatic conditions, the cell's heat dissipation is detected.
The heating time in ARC is typically long, allowing sufficient time for boiling/venting and potential side reactions such as SEI and electrode material breakdown and electrolyte degradation to occur at high temperatures. If the battery were heated more quickly, this could affect the test results. Assuming the electrolyte has sufficient time to boil/vent at lower temperatures, thermal runaway will not occur at higher temperatures because there is no electrolyte electrode reaction, making thermal runaway impossible. The heating test time used in this paper is approximately 60 minutes, also allowing time for side reactions and electrolyte boiling/venting, but shorter than the ARC test method. For comparison, in the ARC test at the National University of Singapore, batteries from the same batch at 100% SOC were subjected to external abuse. If the thermal runaway temperature were determined by ARC measurement using a similar method to that used in the oven experiment, the result would be approximately 140°C. The results of the two measurements differ by approximately 50°C. Therefore, the thermal runaway temperature value for the same type of battery depends on the test method and its implications, as well as the location, number, and measurement quality of the temperature sensors used. Understanding this is crucial when comparing thermal runaway or onset temperature values from different studies.
The minimum observed thermal runaway temperature, as shown in Figure 5B, reflects the findings of Wu et al. They reported a decrease in thermal stability of the aged lithium-ion battery after similar testing (approximately <87% SOH) for 200 cycles. In this case, scanning calorimetry (DSC) analysis of the electrodes indicated that, through aging, some of the lithium content in the cathode was irreversibly transferred to the anode, forming an SEI layer obtained through reaction with the electrolyte. Regarding 300 cycles, we observed less severe reactions in terms of thermal runaway and maximum temperature, which was only slightly correlated with a decrease in energy storage capacity.
3.2.3 Combining Results with Gas Measurements
Figure 6 shows the measurements of temperature, battery voltage, and gas emissions for Test 5, a battery aged for 100 cycles. Three venting processes were detected. The first venting process released dimethyl carbonate (DMC) and ethyl acetate (EA) vapors when the battery voltage dropped to approximately 0V. The process began when the battery voltage dropped and the surface temperature was approximately 130°C. DSC measurements were performed on two commercially available separators, a single-layer PP and a three-layer shut-off separator (PP/PE/PP), as well as a separator extracted from a non-abuse battery. This temperature is very close to the first melting temperature at which the battery shuts off the separator pores. A temperature decrease was expected due to separator melting, as the process is endothermic; however, the opposite was true, with a small temperature rise clearly observed within 12 seconds in the battery surface temperature measurements. One possible explanation for the observed temperature rise is that the battery experienced an internal short circuit, resulting in heat generation; however, a short circuit should only occur if both layers of the separator lose their insulating function (melt). A second venting phase occurred 3.5 minutes before thermal runaway, also releasing ethylene carbonate (EC). During this venting phase, the battery temperature dropped significantly due to the cooling effect of the overflowing gas. The characteristics of the first and second venting phases were not seen or heard in the video and were only determined by FTIR gas measurements.
