Toxic fluoride gas emissions from lithium-ion battery fires

Overview

The study asks how much toxic fluoride gas burning Li-ion batteries release and whether water mist changes it. Seven types of commercial cells and packs (LCO, LFP, NCA/LATP and a laptop pack) were burned over a propane burner under the Single Burning Item hood at RISE, across several states of charge. Heat release and fluoride gases were measured, with HF cross-checked by FTIR and by gas-washing bottles. In total 39 fire tests were run using a 16 kW propane burner; HRR was measured in 38 tests, FTIR in 35 and gas-washing bottles in 19.

Burning Li-ion batteries released 20 to 200 mg/Wh of HF relative to nominal energy capacity, verified by two independent methods. Extrapolated, this is 2-20 kg of HF for a 100 kWh battery system and 20-200 kg for a 1000 kWh system such as a small stationary storage installation. For comparison, the IDLH level for HF is 0.025 g/m³ (30 ppm) and the 10-minute lethal AEGL-3 is 0.0139 g/m³ (170 ppm). POF3 was measured at 15-22 mg/Wh, but only for one cell type (type A) at 0% SOC. Four repeat tests of type B LFP pouch cells at 100% SOC gave a mean FTIR-measured HF release of 166.8 mg/Wh (standard deviation 11.5 mg/Wh).

Water mist (851 g of water applied) increased the peak HF production rate by 35% but had no significant effect on the total HF released; it delayed the fire and briefly raised the HF production rate. The authors argue that toxic gas, not only heat, must be included in risk assessments for large battery installations, especially in confined spaces.

Limits: measurements come from small-scale fire tests of cells and packs and are extrapolated linearly to large systems; POF3 was measured for a single cell type at 0% SOC; and the water-mist test used a limited amount of water.

Key points

1. Burning Li-ion batteries released 20 to 200 mg/Wh of HF relative to nominal energy capacity, verified by two independent measurement methods.
2. Extrapolated, this is 2-20 kg of HF for a 100 kWh battery system and 20-200 kg for a 1000 kWh system.
3. For comparison, the IDLH level for HF is 0.025 g/m³ (30 ppm) and the 10-minute lethal AEGL-3 is 0.0139 g/m³ (170 ppm).
4. POF3 was measured at 15-22 mg/Wh, but only for one cell type (type A) at 0% SOC.
5. Water mist (851 g of water applied) increased the peak HF production rate by 35% but had no significant effect on the total HF released.
6. Four repeat tests of type B LFP pouch cells at 100% SOC gave a mean FTIR-measured HF release of 166.8 mg/Wh (standard deviation 11.5 mg/Wh).
7. In total 39 fire tests were run with a 16 kW propane burner; HRR was measured in 38 tests, FTIR in 35 and gas-washing bottles in 19.

Using it in design

• Add HF toxicity to risk assessments for enclosed battery rooms and BESS: take 20-200 mg/Wh of stored energy and compare the resulting room concentration with the IDLH (0.025 g/m³) and 10-minute AEGL-3 values.
• Knowing water mist does not reduce total HF and only delays the fire while raising peak production rate, do not justify the suppression system as an HF-reduction measure; add evacuation and ventilation controls.
• Check the HF yield input of probabilistic models against this study's mg/Wh range.

Who should read it

Engineers assessing fire and toxic-gas risk for Li-ion battery installations in enclosed spaces.

Document contents

1. Introduction
2. Results (fire tests, heat release, fluoride gas emissions, repeatability, water mist measurements)
3. Conclusions
4. Methods (batteries, fire tests, HF measurement, water mist test)

Related standards

• NFPA 855
• UL 9540A
• IEC 62619

Read the document

Publication details

Authors

Fredrik Larsson; Petra Andersson; Per Blomqvist; Bengt-Erik Mellander

Publisher

Scientific Reports (Springer Science and Business Media LLC)

Year

2017

Volume

7

Issue

1

Pages

10018

Document language

English

Source type

Scientific

Access

Open licence — downloadable

DOI

10.1038/s41598-017-09784-z

Licence

CC BY 4.0 — share with attribution

Licence evidence

www.nature.com

Citation