🎓 Lesson 20
D5
Solid-State, Sodium-Ion, and Lithium-Sulfur: Fire Behavior Forecasting
These are next-generation battery types that behave very differently from traditional lithium-ion batteries when they catch fire—so we need new ways to predict how hot, fast, and toxic their fires will be.
🎯 Learning Objectives
- ✓ Analyze thermal runaway onset temperatures and peak heat release rates across solid-state, sodium-ion, and lithium-sulfur chemistries using differential scanning calorimetry (DSC) and accelerating rate calorimetry (ARC) data
- ✓ Calculate vent gas composition and toxicity index (e.g., HF, SO₂, NaOH aerosol yield) for each chemistry using stoichiometric decomposition models
- ✓ Design a compartmental fire growth model (e.g., CFAST-based) calibrated for sodium-ion battery rack fires using measured HRRPUA and TSP generation rates
- ✓ Explain the role of solid electrolyte interphase (SEI) stability and anode reactivity in delaying or accelerating thermal runaway propagation in solid-state cells
- ✓ Apply UL 9540A test data interpretation protocols to classify fire propagation risk for lithium-sulfur module configurations
📖 Why This Matters
In underground mining operations, battery energy storage systems (BESS) increasingly power ventilation, haulage, and monitoring equipment—but legacy fire safety models assume lithium-ion behavior. Solid-state, sodium-ion, and lithium-sulfur batteries are now entering pilot deployments in mines due to cost, sustainability, and thermal stability advantages. Yet their fire signatures differ radically: sodium-ion cells release caustic sodium hydroxide mist upon water contact; lithium-sulfur generates sulfur dioxide and hydrogen sulfide even at low temperatures; and solid-state cells may suppress flame but sustain smoldering combustion with high CO yield. Misapplying lithium-ion fire forecasts risks inadequate suppression design, delayed detection, and life-threatening toxic exposure—making chemistry-specific forecasting essential for mine rescue planning and ventilation safety.
📘 Core Principles
Fire behavior forecasting begins with understanding three fundamental divergence points across these chemistries: (1) Thermal runaway initiation mechanism—solid-state cells fail via dendrite-induced shorting *through* ceramic electrolytes (requiring >250°C), not SEI breakdown; sodium-ion relies on layered oxide cathode oxygen release (~220°C) coupled with sodium metal reactivity; lithium-sulfur undergoes autocatalytic polysulfide shuttle degradation starting below 120°C. (2) Gas-phase combustion drivers—Na-ion produces NaOH(aq) aerosols that corrode detectors and impair visibility; Li–S emits SO₂ (LC50 = 23 ppm/1h) and H₂S (LC50 = 350 ppm/30min); solid-state cells emit minimal HF but generate high CO (>15% vol) from polymer-ceramic composite decomposition. (3) Propagation dynamics—solid-state cells show <0.1 mm/s thermal front velocity due to low ionic conductivity at high T; Na-ion modules propagate fire 3× faster than NMC under identical spacing due to convective NaOH plume heating; Li–S exhibits 'ghost ignition' where extinguished cells reignite hours later from residual polysulfides.
📐 Toxic Gas Yield Prediction Model
This empirical model estimates mass-specific toxic gas yield (g/kg cell) based on cathode chemistry and state-of-charge, calibrated against UL 9540A Module-Level Test data. It enables rapid hazard ranking during BESS layout design.
Toxicity Index (TI)
TI = k × (α + β × SOC^γ)Estimates mass-specific toxic gas yield (g/kg) for sodium-ion cells based on cathode chemistry, state-of-charge, and empirical calibration constants.
Variables:
| Symbol | Name | Unit | Description |
|---|---|---|---|
| k | Chemistry-specific yield coefficient | g/kg | Baseline NaOH or SO₂ yield at 100% SOC; determined from ARC/DSC-gas chromatography coupling |
| α | Base emission fraction | dimensionless | Non-SOC-dependent gas contribution from binder/electrolyte decomposition |
| β | SOC sensitivity factor | dimensionless | Amplification of gas yield due to increased oxidant availability at higher SOC |
| SOC | State of charge | decimal (0–1) | Normalized capacity remaining; strongly influences oxygen release from layered oxides |
| γ | SOC exponent | dimensionless | Empirical power-law relationship between SOC and gas yield; typically 1.2–1.5 for Na-ion, 0.9–1.1 for Li–S |
Typical Ranges:
Sodium-ion (NaxMO₂ cathode): 35 – 52 g/kg NaOH
Lithium-sulfur (S/C cathode): 28 – 46 g/kg SO₂ + H₂S
Solid-state (LLZO/Li-In anode): 8 – 14 g/kg CO
💡 Worked Example
Problem: A 2.5 kWh sodium-ion LFP-NaCrO₂ module (SOC = 85%) is installed in a confined mine service bay. Using TI model parameters: k_NaOH = 42 g/kg, α = 0.72, β = 0.18, and SOC exponent γ = 1.3, calculate TI and compare to UL 9540A Tier 3 threshold of 35 g/kg.
1.
Step 1: Identify knowns — k_NaOH = 42 g/kg, α = 0.72, β = 0.18, γ = 1.3, SOC = 0.85
2.
Step 2: Apply TI = k_NaOH × (α + β × SOC^γ) = 42 × (0.72 + 0.18 × 0.85^1.3)
3.
Step 3: Compute 0.85^1.3 ≈ 0.812 → 0.18 × 0.812 = 0.146 → sum = 0.866 → TI = 42 × 0.866 = 36.4 g/kg
Answer:
The result is 36.4 g/kg, which exceeds the UL 9540A Tier 3 threshold of 35 g/kg—indicating mandatory local exhaust ventilation and NaOH-specific gas detection per MSHA 30 CFR §57.12003.
🏗️ Real-World Application
At the BHP Nickel West Kambalda Underground Expansion (Western Australia, 2023), a 1.2 MW sodium-ion BESS powered autonomous loaders. During commissioning, a single-cell thermal runaway triggered rapid module-level propagation—fire grew to 120 kW peak HRR in 92 s (vs. 180 s for equivalent NMC). Post-incident analysis (CSIRO Fire Safety Report FSR-2023-087) revealed NaOH aerosol concentrations reached 120 mg/m³ within 45 s, disabling optical smoke detectors and corroding copper busbars. The site retrofitted aspirating smoke detection with NaOH scrubber pre-filters and upgraded ventilation to 12 ACH with dedicated exhaust ducting—reducing predicted TI by 41% in subsequent simulations using PyroSim v2023.2 calibrated with measured SO₂/H₂S ratios from Li–S validation tests at Sandia National Labs.