At present, a large amount of research has been conducted on the key influencing factors of thermal runaway (TR) such as battery chemical system and state of charge (SOC), but the core mechanism of accelerated thermal runaway in batteries is still unclear. Existing research mostly suggests that internal short circuit (ISC) is the main cause of TR, where the electrode contacts to form a current path, releasing energy to heat up and trigger accelerated TR.
However, the method of determining ISC based on voltage drop in TR testing has limitations. Multiple studies on 18650 batteries have shown that in 2.2 Ah battery testing, the voltage first drops to around 1 V due to positive electrode phase transition, and then returns to zero through SEI film decomposition regeneration and separator shrinkage fluctuations.
The voltage drop at 114 ℃ in the ARC test of a 2.9 Ah battery using the “HWS” protocol was attributed to separator shrinkage; The ARC test of a 2 Ah battery also yielded similar conclusions. It is worth noting that these studies all show that after a voltage drop (often seen as an ISC signal), the battery does not immediately enter the accelerated thermal runaway stage.
More importantly, cylindrical batteries such as 18650 and 21700 are equipped with safety devices such as CID, thermal fuses, PTC, which seriously interfere with ISC determination. Research shows that when a prismatic battery is heated to 180 ℃, the CID circuit disconnects; The voltage drop temperature of 18650 battery in high SOC state does not match the expected value. In addition, positive electrode phase transition, SEI reaction, and separator shrinkage also cause voltage drop. All of these indicate that relying solely on voltage drop to determine ISC has limitations and is prone to misjudgment when safety devices intervene.
In addition, the melting of the separator is a key factor in ISC. The melting points of traditional PE and PP separators are 135 ℃ and 165 ℃ respectively, and their thermal stability is poor; The three-layer composite PP | PE | PP separator improves safety by melting and blocking pores in the PE layer and maintaining high-temperature structural stability in the PP layer; Ceramic coated separators rely on inorganic coatings to suppress high-temperature shrinkage. However, existing research variables are mixed, and although the effects of four separators on thermal runaway of 85 mAh soft pack batteries were compared, they have not been validated in 18650 batteries.
1.test advantage
This study breaks through with three innovations:
① Customized single variable 18650 experimental battery, eliminating interference factors;
② Constructing a mechanical integrity testing system for separators that integrates thermal diffusion and jelly roll compactness analysis;
③ Using a three electrode structure with lithium iron phosphate reference electrode, compare the presence and absence of CID batteries, and accurately analyze the correlation mechanism between voltage dip and internal short circuit.
Lithium ion batteries are widely used due to their high energy density advantage, but safety hazards restrict their large-scale development. Internal short circuit is a key factor in causing thermal runaway, but existing research has shortcomings in determining internal short circuit and exploring the mechanism of separator action. Therefore, this study proposes three innovative designs to improve the relevant theories.
2 Materials and Methods
2.1 Battery Preparation and Characteristics
This study used 1.8 Ah 18650 batteries and conducted thermal abuse testing in the laboratory. The positive electrode of the battery is made of 92% NMC622, and the negative electrode is made of 93.2% graphite.
Inject 6.5 g of LP30 electrolyte (1 mol/L LiPF6/EC: DMC=1:1) into the battery in a dew point -50 ℃ drying room. The SEI film is formed by cycling at a C/5 rate for 1.5 hours at 10 ℃ and then discharging to 2 V; Before testing, charge with C/5 constant current to 4.2 V (constant voltage<C/20) until fully charged.
To monitor potential changes, LFP with good high-temperature stability was selected as the reference electrode to prepare a three electrode battery. Apply 50 mg LFP slurry onto the aluminum grid, embed it into a jelly roll after insulation treatment, and remove lithium at a C/5 rate for 90 minutes before testing. The 1C/1D cycle test (Cell 172) showed stable performance within 124 cycles and did not interfere with thermal abuse testing.
Simultaneously prepare comparative batteries with and without CID, and conduct DSC melting point tests (5 ℃/min, 25-250 ℃) on three types of separators: 16 μ m PE, 20 μ m PP | PE | PP, and 20 μ m PP ceramics.
2.2 Mechanical integrity test of separator
Make jelly rolls of three separators and NMC622 graphite, and place them in a drying oven. During testing, no electrolyte or shell is added. Each jelly roll is connected in series with a 680 Ω resistor (R ₁, R ₂, R ∝) and in parallel with a 0.5V constant voltage power supply. The power supply voltage, resistor voltage, and temperature data are recorded using EC Lab. When the separator contracts and the positive and negative electrodes come into contact, the resistance voltage reaches 0.5V, indicating the loss of mechanical integrity. Based on this, the total current is calculated using a formula.
2.3 Thermal abuse test
(1) ARC adiabatic test
ARC equipment using thermal hazard technology (THT) simulates the worst-case scenario through HWS protocol. During the process of thermal abuse, the increase in battery internal pressure will activate CID (disconnect circuit), causing voltage fluctuations; When ISC occurs, a soft short circuit can cause voltage fluctuations, while a hard short circuit can cause the voltage to return to zero. To filter out fluctuations and accurately read voltage, a 12 k Ω resistor is connected in parallel across the battery – this resistor does not significantly discharge (Ucell/R ≈ 0.35 mA, equivalent to C/5143 rate), but can pull the potential to zero when CID is disconnected or ISC occurs.
(2) Heat test
The heating process is: keep at 40 ℃ for 20 minutes, then heat up at a rate of 1.5 ℃/min. Record temperature and voltage during the testing process, and observe the TR process through video.
3 Experimental Results
3.1 Voltage drop mechanism during thermal runaway
This study analyzes the essence of voltage drop through three methods: three electrode testing, dual voltage monitoring, and no CID comparison.
(1) Potential variation of three electrode battery
The ARC test of the three electrode battery (PE separator) shows that at 100% SOC, the NMC622-LFP potential (Ewe) is 0.79 V, the graphite LFP potential (Ece) is -3.34 V, and the battery voltage (Ecell=Ewe Ece) is 4.12 V. Within 0-1529 minutes, Ewe changes by less than 0.65%, Ece increases by 3.41%, and the battery voltage decreases by 2.89%, indicating that the main cause of voltage changes during this stage is the decomposition and regeneration of the negative SEI film.
At 1529 min (approximately 95 ℃), Ewe and Ecell decreased synchronously, while Ece showed no significant change. Due to the temperature being lower than the melting point of the PE separator (128.9 ℃) and NMC622 not decomposing, it is speculated that the voltage drop was caused by the activation of CID – the negative electrode produced gas, which increased the internal pressure and triggered the disconnection of the positive electrode connection, resulting in the failure of positive electrode potential monitoring.
At 1816 minutes (approximately 143 ℃), the battery temperature dropped sharply by 4 ℃ , and the safety valve opened for exhaust, causing synchronous fluctuations in Ewe and Ece; 1871 min, the top cover insulator melted, causing a short circuit between the negative electrode and the top cover, resulting in a further decrease in voltage; Finally, Ece returns to zero (point H), at which point the PP ceramic separator melts, causing a short circuit between the negative electrode and the LFP reference electrode. At 224 ℃, the contact heat is out of control and accelerates, and the temperature rises to 466 ℃.
(2) Dual voltage monitoring verification
To further verify the impact of CID, the “positive top negative” voltage (Eccell1) and the “positive ear negative” voltage (Eccell2) were simultaneously monitored on the battery (Cell 151). The results showed that Ecell1 decreased at 107 ℃ (consistent with the CID activation temperature of 95 ℃ in the three electrode test) and lost signal before accelerating TR; Eccell2 decreased at 194 ℃ and triggered accelerated TR 3 minutes later (219 ℃). This result confirms that the early voltage drop (Eccell1) of standard 18650 batteries is due to CID activation, not ISC.
(3) ISC-TR correlation without CID battery
The ARC test of the PE separator battery (Cell 109) without CID shows that the exhaust temperature is 140 ℃, which is close to that of the battery with CID (143 ℃), indicating that CID does not affect thermal response; But the voltage drop was delayed to 215 ℃ (95 ℃ for CID batteries), and within 2.52 seconds after the voltage drop, the temperature sharply increased, triggering accelerated TR, with the highest temperature reaching 287 ℃ (ambient temperature 483 ℃, it is speculated that the thermocouple on the battery surface fell off in TR).
The test results of PP ceramic separator without CID battery are similar: the exhaust temperature is 167 ℃ (higher than 143 ℃ of PE separator), and the acceleration TR is triggered within 3.78 seconds after the voltage drops, further confirming that large-scale ISC is the direct cause of accelerated TR, and the interval between the two is only a few seconds.