Safety of Lithium-Ion Cells and Batteries at Different States-of-Charge

Cells tested from manufacturer A exhibited venting for all SOCs tested from 0% to 100% and the results are highlighted in Fig. 2a and summarized in Table II. Venting was confirmed visually from the escape of gases from the top of the cells via vents in the header and the subsequent drop in the temperature of the cells. The lowest temperature at which venting occurred was 118 °C for the cells at 100% SOC and the highest temperature observed was 143 °C for the cells at 0% SOC. After venting, the cells continued to rise in temperature from the heat supplied by the tape heater as well as the self-heating occurring in cells due to exothermic reactions within the cells. Heating by the tape heater was shut off after the trigger cell went into thermal runaway or the cell voltage dropped to 0 V. Thermal runaway occurred in the cells at various onset temperatures for different SOC, except for cells at 0% SOC where no thermal runaway occurred. Post-test pictures in Supplementary Information Fig. S1 (available online at stacks.iop.org/JES/167/140547/mmedia) show the cells after heating test at 100% SOC for different manufacturers. Cell from manufacturer A catches fire and gets burnt following thermal runaway event. Cells tested at 15% SOC underwent thermal runaway at 213 °C with cells at higher SOC going into thermal runaway at lower temperatures. The onset temperature for thermal runaway for cells tested at 100% SOC was 174 °C. The severity of the hazards from these off-nominal tests were dependent on the SOC of the cells. Release of gases, electrolyte, and sustained fire were observed at higher SOC while mild thermal runaway with no fire was observed for cells tested at 15% SOC.

Zoom In Zoom Out Reset image size

Cells from manufacturer B are indistinguishable visually and even advertised as the same as cells from manufacturer A. Differences in measured capacity (1.8 Ah) and advertised capacity (3.2 Ah) (see Table I) in cells from manufacturer B hint at a low-quality product. Temperature profiles of cells from manufacturer B during heating tests at different SOC are shown in Fig. 2b. Venting temperatures for cells tested from manufacturer B varied from 96 °C at 100% SOC to 177 °C at 0% SOC. Electrolyte leakage was observed only in some cells after cell venting. Cells tested at 15% SOC and at 0% SOC did not undergo thermal runaway. Cells tested at 100%, 50%, 40%, and 30% SOC underwent thermal runaway at 191 °C, 141 °C, 177 °C, and 182 °C respectively. The onset temperature for thermal runaway did not exhibit a correlation with cell SOC. As discussed later, the voltage profile indicated that the internal protective features found in the traditional lithium-ion 18650 cell (manufacturer A) were not present in the cells from manufacturer B. This was confirmed with a destructive analysis of the cells. In terms of hazards, release of smoke was observed for all cells tested. Fire was only observed in the case of cells tested at 100% SOC. Cells tested at 100% SOC experience a fire and the contents were ejected as shown in Fig. S1. The low measured capacity of manufacturer B was likely the reason for the non-catastrophic behavior of the cells from manufacturer B.

Cells tested from manufacturer C were of the cylindrical 26650 format with rated and experimentally measured capacity of 5.0 Ah. The temperature profiles for the cells at different SOC is shown in Fig. 2c. Venting was observed for cells tested at 100% SOC at 135 °C and with decreasing SOC, the temperature at which venting occurred increased. The highest venting temperature observed was 154 °C for cells tested at 0% SOC. With venting, all the cells started to leak liquid electrolyte. Following cell venting, with subsequent rise in cell temperatures, the cells entered into thermal runaway at various temperatures as a function of SOC. The thermal runaway onset temperature was 177 °C for cells tested at 100% SOC and the onset followed an increasing trend in temperature with decreasing SOC. The highest temperature for the onset of thermal runaway was 193 °C for cells tested at 15% SOC. Cells tested at 0% SOC did not go into thermal runaway. Minimal smoke was observed for cells tested at 0% SOC, however the cells tested at higher SOC released large amount of smoke. Cells tested at 100% SOC and 50% SOC also exhibited sustained fire after thermal runaway events. The results of this set of tests for manufacturer C followed the expected trend of less hazard severity with decreasing SOC.

The pouch format cells from manufacturer D tested at SOCs ranging from 100% to 0%, vented at temperatures between 77 °C and 96 °C following an increasing trend in venting temperature with decreasing SOC. The much lower venting temperatures observed with pouch cells compared to cylindrical metal can cells is due to the lower burst pressure of pouch cell format geometries. Lower burst pressures indicate that very high temperatures are not required for sufficient gas collection that would cause an opening of the pouch. With continued rise in cell temperatures, there was a second venting for all cells at temperatures between 117 °C–124 °C except for cells tested at 100% SOC. Cells tested 100% SOC underwent thermal runaway at 113 °C. The temperature profiles are presented in Fig. 2d. Onset of thermal runaway occurred at 169 °C and 171 °C for the cells at 50% SOC and 40% SOC, respectively. Although the cells were restrained while testing, swelling was observed in all cells. Cells that did not go into thermal runaway showed electrolyte leakage from the tab area of the cell. For cells that went into thermal runaway (SOC 40% and higher), cell pouch opening was observed along with a large amount of smoke. Fire was observed on cells tested at 100% SOC only. Figure S1 depicts the hazard observed in cell tested at 100% SOC showing burning and cell opening. There was minimal smoke and no fire for cells tested at SOC 30% and lower.

Venting occurred for all the cells from manufacturer E within a narrow temperature range from 146 °C to 160 °C. The volume of gas generated and the pressure created inside the cell dictate the vent activation. Given the narrow range of temperature observed for the cells tested at the various SOCs, it is possible that volume of gas and pressure generated is a stronger function of temperature compared to the SOC. Cells at 40% SOC and higher went into thermal runaway as shown in Fig. 2e. The onset temperatures for thermal runaway were 210 °C, 224 °C and 221 °C for cells tested at 100% SOC, 50% SOC, and 40% SOC respectively. There was no fire observed for any of the cells tested under the various SOC. Lack of fire limits the hazards for cells from manufacturer E and it is evident when comparing post-test pictures of cells tested at 100% SOC in Fig. S1. Electrolyte leakage and boiling was observed for all cells along with release of gases and separator materials from the vent holes in the header area. The amount of escaped smoke and gases were directly correlated to the cell SOC with a large amount of smoke for cells at 100% SOC and minimal smoke for cells at 0% SOC. Although no trend was observed for the venting temperatures, the amount of smoke was the maximum for cells at 100% SOC and it decreased with decreasing SOC. The estimates on the amount of smoke are given in relative terms as the exact volume of smoke was not measured and the estimates were based on a visual analysis of the video recordings.

The pouch format cells from manufacturer F did not exhibit a correlation between cell SOC and venting temperature. All cells vented at temperatures between 88 °C–99 °C. Electrolyte leakage from one corner of the cell was observed for all the cells. The temperature profiles for the cells at different SOC are shown in Fig. 2f. Thermal runaway occurred in the cells tested at 100% SOC at around 132 °C with a high amount of smoke and the cell opening from the sides. Mild thermal runaway was observed in cells at 50% SOC and 40% SOC with onset temperatures around 154 °C. For cells at 30% SOC and lower, no thermal runaway was observed. The amount of smoke released from the cells were correlated with the cell SOC and found to follow an increasing trend with increasing SOC. Although no trends were observed in the initial venting temperatures, the amount of smoke and the level of hazard decreased with decreasing SOC.

The single cell battery in pouch format from manufacturer G was subjected to heat-to-vent tests like other pouch format cells and the temperature profiles at different SOC are presented in Fig. 3. These batteries resemble those from a top-tier manufacturer (not tested in this project) and are labelled as such but are available at a much lower cost compared to those from the original manufacturer. Venting temperatures in this battery did not show any correlation with the SOC. Thermal runaway occurred when the batteries were at 30% SOC and higher. The lowest onset temperature for thermal runaway was 129 °C at 100% SOC and it displayed an increasing trend with SOC, reaching a temperature of 182 °C for the batteries at 30% SOC. The onset temperatures for venting and thermal runaway as well as the maximum temperatures recorded in battery tests are summarized in Table III. In terms of hazards, the maximum temperatures observed showed a good correlation to increasing SOC. The maximum temperature recorded was 608 °C for batteries at 100% SOC and decreased with lower SOCs. Electrolyte gases and liquid were released during venting in the batteries from the tab area for all the tests with the severity directly corresponding to the SOC of the batteries. The test articles at 100% SOC also caught on fire (Fig. S1). In summary, the venting temperatures did not show a correlation to the cell SOC, but the onset temperatures for thermal runaway went to higher temperatures with decreasing SOC and the maximum temperature recorded show a trend of decreasing temperatures with decreasing SOC.

Zoom In Zoom Out Reset image size

Batteries from manufacturer H containing four 18650 cells in a 2P2S configuration were tested by heating a single cell in the battery. Temperature profiles of the battery at different SOC during the heating tests are shown in Fig. 3. Thermal runaway occurred in the trigger cell (cell with heater) for tests where the SOC was at 30% and higher. In the case of the battery at 30% SOC, there was no thermal runaway propagation to other cells. However, at higher SOC, it propagated to neighboring cells and all cells in the battery experienced thermal runaway. Intense smoke and sustained fire for over 5 min was observed in batteries tested at 80% SOC and 100% SOC whereas at lower SOCs, the battery test article did not catch fire. Figure S1 shows the battery tested at 100% SOC after the test completion with damage observed to the initiation cell and adjacent cells. The intensity of smoke from the test article correlated with the SOC with minimal smoke at 0% SOC. Thermal runaway onset occurred at temperatures ranging from 171 °C to 182 °C for tests between 100% SOC and 30% SOC. Maximum temperature in the test article reached 704 °C for test at 100% SOC and followed a downward trend with a maximum temperature of 232 °C recorded for test at 0% SOC.

The application of external short for cells from manufacturer B led to an immediate increase in current to its maximum value along with a drop in voltage of the cells. The temperature and voltage profiles of cells at different SOC during the external short tests are shown in Fig. 5. The maximum current observed varied from about 30 A (for cells tested at 15% SOC) to 18 A (for cells tested at 30% SOC) and did not exhibit any correlation to the cell SOC. Following this process, there is a short stabilization period observed for the cell voltage and current and then the cells start to undergo a discharge since the load is still on the cell for a total of three hours. This period of discharge is more pronounced in cells at higher SOCs and is less apparent in cells at lower SOCs. This indicates that the remaining cell capacity decreases with lower SOCs and hence it takes shorter times for cell voltages to approach 0 V. This period is marked by a continuous discharge of the cells at a high current leading to an increase in the internal temperature of the cells accompanied by mass transport limitations within the cells. The increase in internal temperature of the cells is higher for cells at high SOC (Table VI). The maximum temperature observed was 149 °C for cells at 100% SOC and the minimum temperature of 54 °C was observed for cells at 0% SOC. The onset temperature for cells going into thermal runaway is lower for cells at higher SOC and the cells at higher SOC were more likely to go into thermal runaway compared to cells tested at lower SOC owing to the differences in internal heating. Visual observations from the videos during the tests verified the extent of damage to the cells. Cells tested at higher SOC showed the most damage including cell rupture and venting with smoke and ejection of hot liquid electrolyte from the cells. Damage to the cells and batteries after the completion of external short testing at 100% SOC is shown in Fig. S2. Cells tested at SOCs of 30% and lower displayed only heating of the terminal tabs without any venting of gases or electrolyte leakage from the cells.

Zoom In Zoom Out Reset image size

Pouch cells tested from manufacturer D exhibited a decreasing trend in maximum current with decreasing SOC after the cells were externally shorted (Fig. 5). The maximum current observed was 114 A for cells tested at 100% SOC and the lowest current observed was 41 A for cells tested at 0% SOC. After the peak current and initial decrease in voltage, high rate discharge of the cells continues until the voltage drops to 0 V for all the cells. The temperatures recorded continued to increase during this period until they reached a maximum value when the cell was completely discharged. The maximum temperature in the cell was recorded at the tabs. Cell temperatures increased with increasing SOC (Table VI). The maximum temperature for the cells at 100% SOC was 132 °C and that for cells at 0% SOC was 52 °C. Cell swelling and rupture near the negative tab were observed with leakage of electrolyte and smoke coming from the tab area (Fig. S2). The positive tab material was nickel while the negative tab material was aluminum coated with nickel. In all the tests, charring and damage to the negative tab were observed. This was particularly severe in tests with high SOC (50% and 100%) resulting in melting and disconnection of the negative tabs in cells at 100% SOC, owing to its lower melting point compared to the positive terminal tabs.

The maximum currents observed from manufacturer E cells during the electrical abuse via external short were independent of cell SOC. This maximum value varied from slightly above 40 A in the case of cells tested at 50% SOC to above 20 A for cells at 0% SOC. Following the initial decline in voltage and the maximum current, a steady period for cell voltage and current was observed (Fig. 5). The cell temperatures continued to increase during this period until a maximum value was reached when the cell voltages decreased to 0 V. The maximum temperature achieved during the tests decreased with decreasing SOC and is summarized in Table VI. The highest recorded temperature was 93 °C for cells tested at 100% SOC and was 37 °C for cells at 0% SOC. The high temperatures achieved during the tests led to over-heated tabs, but there were no additional cell hazards observed such as smoke, rupture, or release of liquid electrolyte (Fig. S2).

When the pouch cells from manufacturer F were externally shorted, there was an immediate drop in voltage and the current reached the maximum value. The voltage and current dropped until the cells were fully discharged and the voltage reached 0 V. Cell temperatures continued to rise, with the highest temperatures observed around the tabs. The temperature and voltage profiles at different SOC are shown in Fig. 5. The maximum current and the maximum temperatures observed in the cells at different SOC (Table VI) were not dependent on the SOC of the cells under test. The highest value of the maximum current was observed was 196 A for cells at 100% SOC while the lowest value of that was 163 A for cells at 40% SOC. The positive tabs burned off in the cells from 100% to 30% SOC due to its lower melting point compared to negative tabs preventing the cells from experiencing the short circuit load (Fig. S2). The positive tab material in the cells was copper while the negative tab material was copper coated with nickel. The positive tab area also went into flames and was charred from overheating. Although this feature prevents the full understanding of the current and temperatures observed during an external short, the burning of the tabs is a fail-safe condition.

Batteries from manufacturer G were tested by removing the battery management system (BMS). The test results observed at different SOC follows similar trends as discussed earlier with the drop in voltage and a sharp increase in current with the application of the short circuit load. Temperature and voltage profiles of the batteries at different SOC are shown in Fig. 6. The high-rate discharge of the battery proceeds with an increase in internal cell temperature until the cell is fully discharged and the voltage reached 0 V. The highest value of maximum current recorded was 39 A in the case of the battery tested at 40% SOC and the lowest value of maximum current was 21 A for those at 100% SOC (Table VI). In the case of the test at 100% SOC, thermal runaway occurs with the maximum temperature reaching about 410 °C. Sustained fire in the tab area was accompanied by thermal runaway (Fig. S2). The maximum temperature observed followed a decreasing trend with SOC, with 118 °C recorded for the batteries at 50% SOC and 27 °C for those at 0% SOC. Swelling of the pouch was observed for all tests except for those at 0% SOC.

Zoom In Zoom Out Reset image size

Batteries from manufacturer H subjected to the low impedance external short exhibited minimal hazard. The external short test was first conducted on the battery with the protective circuit board which is the BMS and it was found that the BMS provided the protection against the external short and hence there was no increase in current or temperature. Hence the tests were subsequently conducted after removing or bypassing the BMS in the battery pack. When the board was removed, the battery experienced the short circuit condition. However, the cells are fitted with an internal PTC that protects the cells from experiencing a catastrophic event. The maximum current recorded was relatively the same across all SOCs at about 44 A, except for 0% SOC where the maximum current was about 30 A. Accordingly, the temperature of the test article was below 82 °C for all the tests (Fig. 6 and Table VI) except for 0% SOC where the temperature observed was about 41 °C.