Thermal runaway of lithium-metal and lithium-ion cells has resulted in numerous fires. Often the fires are fueled by the flammable gases that are vented from the batteries during thermal runaway. In addition to those installed on the aircraft, millions of lithium batteries are shipped every year as cargo. A Class C cargo compartment is equipped to have an initial concentration of 5% Halon 1301 fire-suppressing agent, followed by a residual concentration of 3% for the remainder of a flight. These halon concentrations are effective at mitigating fires involving typical cargo; however, there is concern whether these concentrations are sufficient to handle a cargo fire involving lithium batteries and to mitigate the risks of a potential explosion of the accumulated vented battery gases.

Tests were conducted to analyze the various gases that were vented from lithium cells in thermal runaway and evaluate the risk of the buildup and ignition of the gases within an aircraft cargo environment.

Small-scale tests were carried out in a 21.7 L combustion sphere in which a gas chromatograph, non-dispersive infrared, paramagnetic analyzer, and pressure transducer were used to quantify the individual gases released from lithium batteries. Once the gas constituents were quantified, tests were performed to measure the pressure increase from combustion of these gases. Large-scale tests were then conducted in a 10.8 m3 combustion chamber, a volume comparable with that of a cargo compartment, to validate the small-scale tests and to evaluate the effect of Halon 1301 on battery vent gas combustion.

Results of the small-scale tests showed that the volume of gas emitted from cells increased with state-of-charge (SOC). Combustion of the gases showed a lower flammability limit of 10% and an upper flammability limit that varied between 35% and 45%, depending on SOC. The combustion tests also showed a maximum pressure rise of more than 70 psia at altitude or more than 100 psia at sea level.

Tests conducted at the approximately 5% halon design concentration resulted in insignificant change to the resulting pressure rise. Tests at approximately 10% concentration effectively inerted the cargo compartment such that the battery gases were unable to ignite.

The results of these tests showed that a variation in SOC affected gas volume substantially for certain cells and that a combustion event could compromise the safety of an aircraft. The Halon 1301 fire-suppression system showed minimal effectiveness against battery gases at current design concentrations of 5%.

In collaboration with Parker Hannifin Corporation, the Fire Safety Branch of the FAA conducted testing to evaluate the effects of three potential failure conditions of hydrogen proton exchange (or polymer electrolyte) membrane fuel cell stacks supplied by Nuvera Fuel Cells. The three conditions examined were a loss of coolant to the stack, short circuit, and a crossflow condition. The testing showed that the stacks were extremely robust under a variety of failure conditions and that, with proper monitoring of key variables, the failures could have been detected and flow of reactant gases stopped prior to any hazardous effects occurring.

It is recommended that any installation of a hydrogen fuel cell system ensure that reactant supply gas pressures, stack temperatures, coolant temperatures, and stack electrical load characteristics be adequately monitored and connected to system shutdown features. In addition, provisions should be made so that the surrounding environment is monitored for any temperature or hydrogen gas concentration increases.

A series of tests was conducted to determine the effect of altitude on FAA Bunsen burner testing. The standard 12-second vertical Bunsen burner test procedure from the FAA Aircraft Materials Fire Test Handbook was used for all testing, but the ambient air pressure was varied to represent altitudes ranging from sea level to 8000 feet. The first tests completed were initially with the Bunsen burner flame by itself. The mass flow rate of the methane fuel was decreased as the altitude increased to keep the flame height constant, which was expected because the mass of oxygen in the surrounding air also decreased. The flame temperature dropped slightly as altitude increased, but remained well above the 1550°F minimum. The flame was stable at all altitudes and no visual differences were noted as the air pressure decreased. Four different materials were tested as well, but only two were used across all the altitudes because they were the only materials that produced consistently long burn times, which was necessary to provide a basis for comparison. These two materials were a 1/32″ thick glass epoxy and a 1/32″ thick woven carbon fiber. The decreased ambient air pressure of the higher altitudes did not significantly affect the flame times or burn lengths of the two materials tested.

This report is a compilation of test data and results from projects conducted by the Fire Safety Branch designed to determine the hazard from and possible hazard mitigation for the bulk shipment of lithium batteries/cells as cargo on transport airplanes. Though the main focus of this report is transport on passenger aircraft (effectiveness of Halon and effect on Class C cargo compartment), general hazard data is also applicable to freighters.

Initial testing found that current packaging did little to impede thermal runaway propagation from cell to cell and package to package. Lowering the state of charge (SOC) was shown to reduce the hazard. A 30% SOC in some cell chemistries stopped cell-to-cell propagation but did not eliminate the hazard from an external fire.

Full-scale tests in a 727 aircraft demonstrated that, even in the presence of Halon 1301, cell-to-cell propagation occurred. The tests also demonstrated that the pressure and off-gassing from the cells could reduce the Halon concentration at an accelerated rate, thereby causing the loss of fire protection earlier then designed.

Recent testing showed that lithium-ion cells in thermal runaway emitted H2 gas that can be explosive even in a 5% Halon atmosphere. The ignition of the gases vented by a small number of cells in thermal runaway (the number depends on many factors, such as chemistry, design, and SOC of the cells, and size, design, and load factor of the compartment) can cause enough damage to a Class C cargo compartment as to render the Halon suppression system ineffective.

Materials for aircraft cabin interiors must meet the flammability requirements of Title 14 Code of Federal Regulations (CFR) Part 25.853. The 14 CFR 25.853 requirement includes a test for large-area materials that measures the heat release rate (HRR) during burning using a fire calorimeter originally developed at Ohio State University (OSU). In the standard 14 CFR 25 procedure, a sample is inserted into the combustion chamber of the OSU apparatus and subjected to a calibrated radiant heat flux of 35 kW/m2 and an impinging pilot flame. Room temperature air is forced through the combustion chamber and exits through the exhaust duct at the top of the apparatus where a thermopile senses the temperature of the exhaust gases. The HRR during the test is deduced from the sensible enthalpy rise of the air flowing through the combustion chamber using the temperature difference between the exhaust gases and the ambient incoming air to calculate the amount of heat released by burning after suitable calibration using a metered methane diffusion flame. Limits of 65 kW/m2 and 65 kW/m2-min for the maximum (peak HRR) and the total heat release (HR) up to 2 minutes into the test (2-min HR), respectively, are placed on large-area materials used in passenger cabins of transport category airplanes carrying more than 19 passengers.

Results from multi-laboratory studies of the same materials tested for HR and HRR according to 14 CFR 25.853 indicated that the laboratory-to-laboratory variation of the test results was relatively high. There are many factors that can contribute to poor agreement between OSU fire calorimeter results obtained in different laboratories (i.e., reproducibility), including the accuracy of the heat flux calibration, contaminated temperature sensors, thermal inertia of the apparatus and its components, and changes in the convective environment in the combustion and bypass chambers caused by airflow and airflow distribution. This study focused on the effect of the airflow through the 14 CFR 25.853 fire calorimeter on the repeatability and reproducibility of the HR and HRR by the sensible enthalpy (temperature rise) of the standard method, compared to HR and HRR measured simultaneously by the oxygen consumption method.