Impact Assessment in Safety Testing of Lithium-Ion

Transcription

1 Technology Report Impact Assessment in Safety Testing of Lithium-Ion Secondary Battery Hideki Kawai, Arata Okuyama and Yuichi Aoki ESPEC CORP. Abstract It has been reported that the Lithium-Ion secondary Battery (LIB) was ruptured, fired, or exploded while in use. The same holds true for the safety assessment of LIB. Because of that, temperature chambers need safety systems to protect from harmful explosions and ensure operator safety. In this paper, we have studied the thermal run-away process of a lithium-ion secondary battery, and effects of safety systems for a chamber in the testing for batteries. From the results of tests, we found that State Of Charge (SOC) when Current Interrupt Device of battery operates is different according to ambient temperature. Compared with the battery of full charge, the battery of the overcharge state had very large explosion pressure, and found out the possibility of the quantitative evaluation technique. It was confirmed that an temperature chamber with no safety systems for battery failure is a real risk, and safety system was required based on testing needs in each situation. This report is an edited and expanded version of the information presented at the 42nd JUSE Symposium on Reliability and Safety. 1 Introduction With their market track record, large capacities and high rate performance, LIBs could likely become the most common type of secondary battery. With the recent rise in popularity of electric vehicles and plug-in hybrid vehicles, R&D work is increasingly looking at LIBs for vehicle use, and high-capacity, long-life batteries. The worst cases of thermal runaway in LIBs can result in fire and explosion, so safety tests and performance evaluation tests such as charge/discharge cycle tests require ample safety measures. Test standards for lithium-ion batteries include performance tests in specified temperature environments that call for the use of temperature chambers, but few evaluations of temperature chamber safety features and durability have been reported.

2 Neither is much known of the processes and mechanisms of LIB thermal runaway outside of specialist technical circles. To address this lack of information, we evaluated battery behavior, and the thermal runaway process and its effects during safety testing. This report describes these findings and our findings on temperature chamber safety features. 2 Test Methods 2-1. Testing overview The conducted the tests shown in Table 1, focusing on the points below. 1. Understanding the basic process leading to LIB fire or explosion 2. Difference in effects on temperature chamber and surroundings according to presence or absence of safety mechanisms 3. Differences in effects on fire/explosion caused by overcharging and forced heating Using typical official test standards 1, 2, 3 for reference, we set our test conditions by considering the potential for battery fire or explosion, and basic items of safety testing. Table 2 shows the typical test standards we referred to Test sample batteries We tested cylindrical LIBs of type Type batteries are very widely used in applications such as laptop PC battery packs, power tools and electric vehicle batteries. Test type Table 1 Testing conducted Test items Overcharge Safety mechanism Forced heating 1 Overcharge No Yes No Battery s initial state Time-of-purchase state 2 Forced heating Yes No Yes Full charged 3 Forced heating when overcharged Yes Yes Yes Overcharged

3 Table 2 Typical test standards for LIBs Test type Typical standards Main test conditions Heating IEC 62660, JIS C C for 10 minutes; Temperature increase of 5 C/min Vibration JIS C 8712, UL to 55 Hz, 90 minutes Low pressure JIS C 8712, UL 1642, UN 11.6 kpa, 20 C Overcharge IEC C, SOC 200% External short circuit JIS C 8712, UL 1642, UN 0.1 Ω max, 55 C (UN) 25 C/55 C (JIS, UL) 3 Evaluating Temperature chambers Without Safety Mechanisms (Overcharge) 3-1. Testing overview We used temperature chambers without safety mechanisms to conduct overcharge testing for evaluating the effects of LIB fire/explosions. Table 3 shows the specifications of the batteries tested, and the overcharge test conditions. We used a charge rate of 2C, and continued charging until battery interruption or fire/explosion. Table 3 Test sample battery specifications and test conditions Test type Overcharge Norminal capacity 2,600 mah Initial charge state Semi-charged Charge current 2C (5.2 A) Test temperature Room temperature (no temperature chamber operation) Measured items Temperature, voltage, current, wall surface strain, video record Point at which interruption and temperature rise stop End condition Point at which fire or smoke emission have subsided (if applicable) Temperature chamber Espec temperature chamber 3-2. Test results Figure 1 shows the temperature chamber and battery used for the test. Figure 2 shows the measured results. About 33 minutes after charging started, the positive electrode began to emit gas, and instantaneously rising fire was observed. When charging was continued, the battery emitted a large pillar of fire and burned, ultimately exploding. The explosion generated a very loud sound, and the explosive pressure discharged sparks and smoke from door gaps. While there was no major distortion interior of the chamber, the interior became coated in soot. Wall surface strain was greatest on the door s inner wall, and wall surface had scratches and scorch marks suggesting collisions of battery components.

4 3-3. Discussion In the temperature chamber without safety mechanisms, the door was subjected to a sudden pressure rise, creating gaps between door and chamber. Repeated testing might result in loads on door locks and hinges, resulting in destruction. Smoke and sparks discharged from gaps may create the potential for fire and hazards to surroundings. Figure 1 Temperature chamber (without safety mechanisms) when battery exploded Figure 2 Battery behavior leading up to fire/explosion caused by overcharging 4 Forced Heating Tests 4-1. Safety mechanisms We carried out forced heating tests on fully charged batteries, using temperature chambers for secondary battery evaluation testing containing Espec safety mechanisms. First we will describe the safety mechanisms of Espec temperature chambers. Figure 3 shows the configuration of example safety mechanisms in an environmental test chamber. Safety mechanisms include pressure relief vents for release the inner pressure of the chamber to the outside when a battery explodes, a safety door lock function that protects operators by detecting when doors are unlocked, a function that detects gas emitted from batteries, intake/exhaust dampers, and CO2 gas fire-extinguishing equipment. We used two temperature chamber models for our battery charge/discharge testing. One model had pressure relief vents, door locks and CO2 gas fire-extinguishing equipment, while the other model also had a gas detection function.

5 Pressure relief vents Figure 3 Safety mechanism examples for temperature chamber Table 4 Test conditions for forced heating tests Initial charge Full charge state Number of First test: 1 sample Second test: 4 batteries Heating condition Test atmosphere Measurement items End condition Temperature chamber Temperature increase of 5 C/min (using ribbon heater) 25 C (set by temperature chamber) Temperature, voltage, current, wall surface strain, video record, pressure relief vent and door open detection signals After fire or smoking subsided Espec temperature chamber for charge/discharge testing 4-2. Testing overview Table 4 shows the test conditions used for forced heating testing. To enable more uniform transmission of heat to batteries, batteries were heated by wrapping a ribbon heater like the one shown in Figure 4 around them. The batteries used as test samples were the same type used in the overcharge tests described in Section Test results (for single battery) Figure 6 shows the results of our forced heating testing on a single battery. The battery voltage decreased at a battery surface temperature of about 120 C. Heating the battery further resulted in the battery maintaining a higher surface temperature than the heating control temperature, starting at about 190 C. Subsequently, the battery surface temperature rose suddenly, and the battery exploded. Figure 5 shows a pressure relief vent during an explosion. No sparks or smoke were observed being discharged from the door during this testing, and door opening was not detected. The temperature rise inside the temperature chamber was greater on wall surfaces in the direction of the ceiling and positive battery electrode. Wall surface strain was greatest on wall surfaces in the direction of the positive electrode. Observations inside the chamber after testing showed that the wall surface in the direction of the positive electrode was covered in a large amount of soot and had marks indicating collisions.

6 Figure 4 Battery during heating test Figure 5 Pressure relief vent operation during explosion Figure 6 Forced heating test results Figure 7 Results of forced heating tests on four batteries 4-4. Test results (four batteries) Figure 7 shows the results of forced heating tests on four batteries. Battery behavior leading up to explosion was the same as the result observed during single-battery forced heating testing. Starting at a temperature of about 190 C, the batteries maintained a surface temperature greater than the heating control temperature, then the temperature suddenly rose and the batteries exploded. Instead of all four batteries exploding simultaneously, the first battery exploded, then the other batteries exploded as though in a chain reaction. After testing, a greater amount of soot coated the wall surfaces inside the temperature chamber than after single-battery testing, and parts were scattered about. Among the safety mechanisms, the pressure relief vents opened during each explosion to reduce the rising pressure, and no door opening were detected.

7 4-5. Test results (evaluation of gas detection and CO2 gas fire-extinguisher functions) We used a temperature chamber with a gas detection function to conduct a single-battery forced heating test, to evaluate the effectiveness of the chamber s gas detection function and CO2 gas fire-extinguishing equipment. We found that the equipment was unable to detect the gas emitted from the battery before it exploded. By triggering immediately after the explosion, we found the CO2 gas fire-extinguishing equipment was able to rapidly cool the battery surface temperature and the temperature within the temperature chamber. We also found that after testing less soot had been generated within the chamber than when the fire-extinguishing equipment was not triggered Discussion The reason that large temperatures, strains and soot coatings were observed on wall surfaces in the direction of the positive electrode may be that fires and explosions started from the positive electrode side, which has a safety valve (a finding reported in Section 3). By enabling pressure rises from explosions to discharge the pressure relief vent, we found that safety mechanisms were effective at reducing the load on the door lock mechanism and temperature chamber, and maintaining the safety of the surrounding area. We found that safety mechanisms detected gas, so that only a small amount of gas was generated by a single battery used in testing. CO2 gas fire-extinguishing equipment is likely effective at reducing the spread of fire after an explosion and cooling the inside of the chamber. 5 Forced Heating when Overcharged 5-1. Testing overview We observed the behavior of batteries in overcharge tests, then subjected the overcharged batteries to forced heating testing to compare how their behavior differs from the behavior of fully charged batteries. Table 5 shows the overcharge test conditions. The test conditions used for forced heating testing were the same as the conditions given in Section 4 (see Table 3) Test results: Overcharge testing Figure 8 shows the results of overcharge testing for a chamber temperature of 25 C. The battery surface temperature did not rise greatly even when the SOC exceeded 100%. But starting from an SOC of about 140%, the temperature started to rise suddenly, and

8 the voltage decreased slightly at the same time. At an SOC of about 170%, the battery cut off, then subsequently stopped because the battery surface temperature started falling. We repeated the test with chamber temperatures of 5 C and 50 C, and found that the battery surface temperature at the time of interruption was roughly the same in each case. Compared to the initial ambient temperature of 25 C, the SOC at the time of interruption was higher for 5 C and lower for 50 C. Table 5 Overcharge test conditions Discharging capacity Charge current Test atmosphere Measurement items End condition Temperature chamber 2,150 mah, 2,950 mah 1C, 2C (current value determined by battery) 5 C, 25 C, 50 C Temperature, voltage, current, video record Point at which interruption and temperature rise stopped Point at which fire or smoke subsided (if applicable) Espec temperature chamber for charge/discharge testing Figure 8 Overcharge test results (25 C / 2,150 mah / 1C) 5-3. Discussion: Overcharge testing Battery interruption caused by overcharge testing may be due to the current interruption mechanism within the battery. 1 Our testing found that the lower the ambient temperature, the higher the SOC at which interruption occurred. The main factors affecting the interruption mechanism are probably the increase in the volume of volatile gases within the battery, the expansion of these gases, and the pressure on vacant layers in the battery caused by the battery casing expanding. At low temperatures, battery interruption may have occurred at high SOCs because the volume of volatile gases and the degree of gas expansion were relatively small. Overcharge testing therefore needs to consider ambient temperature.

9 5-4. Test results: Forced Heating when Overcharged We conducted forced heating testing on battery that had been subjected to overcharge testing. Figure 9 shows the temperature graph leading to explosion. The temperature behavior of the batteries used in this test was nearly the same as the temperature behavior of charged batteries. Starting from a temperature of about 190 C, the battery maintained a temperature at least as high as the heating control temperature, then the temperature suddenly rose and the battery exploded. Figure 10 shows the pressure relief vent operating during an explosion. The explosions were louder than the explosions observed for the full-charge forced heating testing described in Section 3. The vent operated significantly faster and emitted a large volume of sparks. But no door releases were detected. Figure 9 Forced heating testing of overcharged batteries Figure 10 Pressure relief vent operating 6 Comparative Evaluation of Generated Pressures 6-1. Analysis model The effects observed on equipment during forced heating tests led us to suspect that the pressure generated by overcharged batteries was much greater than the pressure generated by fully charged batteries. We attempted to quantify this difference by estimating the pressure generated by fully charged and overcharged batteries from the behavior of the pressure relief vent in response to test conditions. As shown in Figure 11, we considered the vent to be a uniform beam of mass m and length L. When acted on by a force F, the vent has a rotation time of t, rotation angle of θ, rotational moment of I and rotational torque of T. Its angular speed ω and angular acceleration α are given by:

10 Rotational moment I, torque T, angular acceleration α and force F are related as follows: Figure 11 Pressure relief vent analysis model Using this relationship, Formulas (1) and (2) enable F to be expressed as: Using Formula (6) as our evaluation function, we compared the forces generated by the explosions of each test. Table 6 shows the pressure calculation results. We found that the pressures might vary by a factor of nearly 10. Our calculation does not take into account the fact that pressure starts to fall at the same time the pressure relief vent opens, or that pressure is not always applied uniformly in all directions at the same time. But even the approximations it provides indicate a clear difference. By improving measurement precision, it should be possible to make more precise estimates, making our method an effective way to easily understand the effects of explosions. Table 6 Estimates of pressures generated by full charge and overcharged batteries Battery charge state Full charge Overcharge Rotation angle (degrees) Rotation time (seconds) Pressure ratio Conclusion This report has described our observations of LIB behavior during safety testing, and the overcharge and forced heating tests we carried out to evaluate the thermal runaway process and the safety features of temperature chambers. Our main findings are listed below. 1. We observed temperature chambers without pressure relief vents experience excessive pressure on the door, and discharge fire and smoke from door gaps. We observed chambers with pressure relief vents release pressure from the vent, and have no door lock failure. 2. We observed that the use of CO2 fire-extinguishing equipment inhibited the spread of

11 fire after explosions, and was effective in cooling the atmosphere within the chamber. 3. We observed that overcharge testing caused current interruption mechanisms in batteries to operate at different SOCs according to the ambient temperature. 4. We observed that overcharged batteries generate much larger explosive pressures than fully charged batteries, and found a possible method of quantitatively evaluating the value. Our research found that temperature chambers require safety features designed for various test conditions. Future Challenges The testing described here is limited safety testing on Type cylindrical LIBs. In addition to this type of battery, there are batteries with non-cylindrical structures such as lithium polymer batteries and laminate-type batteries. There are also batteries with cells in module packs used as electric vehicle batteries. In future, researchers will need to evaluate batteries of larger capacities and sizes, and ensuring the safety of the battery test area will be a greater requirement. Tester safety performance will therefore be an important issue, and quantitative safety evaluations will be key information when considering safety. Bibliography (all sources in Japanese) 1. The Electrochemical Society of Japan s Committee of Battery Technology, Battery Handbook, Ohmsha, Technical Information Institute, Lithium ion Secondary Battery/Material Thermal Conduction Behavior/Deterioration Evaluation and Test Methods, Technical Information Institute, CMC Research, Li-Ion Secondary Battery Product Standards and Safety Testing 2011, CMC Publishing, Hideki Kawai, Arata Okuyama and Yuichi Aoki, Impact Assessment in Safety Testing of Lithium-Ion Secondary Battery, Proceedings of the 42nd JUSE Symposium on Reliability and Safety, pp. 249 to 254, 2012