2026-07-19

Lithium Use, Storage, and Fire Risks

Table of Contents

  1. Introduction and Scope
  2. How Lithium Batteries Burn — the chemistry that defeats ordinary firefighting
  3. Risk Assessment: Causes of Lithium Battery Fires (ordered by danger)
  4. Deep Discharge, Internal Shorts, and the Recharge Danger
  5. Storing Energy in the Home — the passive fuel‑load problem
  6. Safe Storage Practices (temperature, SOC, enclosures, ventilation, separation, BMS, humidity)
  7. Charging and Discharging Discipline
  8. Fire Suppression and Emergency Response (corrected)
  9. Regulatory Standards and Codes
  10. Best Practices Summary
  11. Lithium Batteries as Latent Fuel
  12. References

1. Introduction and Scope

Lithium‑ion and lithium‑metal batteries power everything from smartphones to electric vehicles and grid‑scale energy storage. Their high energy density is exactly what makes them useful — and exactly what makes them hazardous. A charged lithium cell is both a container of stored electrochemical energy and, in most consumer chemistries, a self‑oxidizing fuel package: when it fails it can release its own oxygen, flammable gases, and enough heat to propagate the failure to neighboring cells and to surrounding combustibles.

This guide covers:

  • the mechanism of thermal runaway and why it is so hard to stop (§2);
  • the causes of lithium battery fires, ranked by danger, with the storage and use practices that mitigate each (§3);
  • the special, often‑misunderstood danger of deep discharge and the recharge that follows (§4);
  • the home‑scale problem — what it means when a household accumulates many batteries, even healthy ones, and a fire starts from something else (§5);
  • storage, charging, suppression, and code basics (§6–§9);
  • a consolidated best‑practices list (§10);

Scope is residential / hobbyist / small‑commercial. Utility‑scale BESS design is referenced for context only.

2. How Lithium Batteries Burn — the chemistry that defeats ordinary firefighting

2.1 Thermal runaway, in one paragraph

Thermal runaway is a self‑accelerating exothermic reaction inside a cell. Once a cell's internal temperature climbs past a series of thresholds, each step releases heat that drives the next: SEI breakdown (~80–120 °C) → separator melt and internal short (~130–150 °C) → cathode decomposition that releases oxygen (~190–240 °C for layered oxides) → electrolyte decomposition and violent gas venting → full runaway with peak cell temperatures that can exceed 650 °C (NMC) and vented‑gas temperatures above 1100 °C. LFP cells run far cooler (≈330–410 °C) and release far less energy.

2.2 "It provides its own oxygen" — fact‑checked and reframed

A common, slightly imprecise statement is that "lithium provides its own oxygen, so you can't smother the fire." The underlying point is correct and important, but the mechanism is usually misstated.

  • Lithium‑ion cells contain little or no lithium metal. The oxidizer that sustains a Li‑ion fire is not metallic lithium. It is oxygen released by the cathode as it decomposes. Charged layered‑oxide cathodes (NMC, NCA, LCO) undergo a phase transition (layered → spinel → rock‑salt) that liberates O₂ (and reactive oxygen species O₂⁻, O⁻).
  • The cell contains both its own oxidizer and its own fuel, creating a self‑sustaining combustion system. The oxygen released from the decomposing cathode reacts exothermically with the primary fuels: the carbonaceous anode (lithiated graphite) and the organic carbonate electrolyte (ethylene carbonate). Hou et al. (2021) quantified this for NMC811 cells: ~41 % of the released oxygen reacts with the ethylene‑carbonate electrolyte at the cathode interface (≈16 % of total heat), and ~59 % migrates to the lithiated anode where it generates ≈65 % of total heat — bringing the cell to its maximum destructive temperature.
  • Chemical crosstalk loops drive the runaway. Jo et al. (2024) identified a self‑amplifying C₂H₄ ↔ O₂ crosstalk loop where ethylene from the anode travels to the cathode and accelerates O₂ release, which returns to the anode to generate more ethylene. Crucially, they also discovered that carbon dioxide (\(CO_2\))—normally considered an extinguishing agent—actively participates in this loop when trapped in an enclosed cell, reacting exothermically with active lithium at the anode to form lithium carbonate (\(Li_2CO_3\)), which further accelerates the runaway process.
  • NMC runaway is self‑sustaining and effectively unstoppable once started. Sallard et al. (2025) report that NMC811 thermal runaway "could not be stopped even with liquid nitrogen (−196 °C)" once initiated. This is the literal meaning of "self‑oxidizing": removing external oxygen or cooling the outside does not interrupt the reaction, because the cell carries both fuel and oxidizer.
  • Chemistry matters a great deal. LFP (LiFePO₄) has a polyanionic olivine structure that is structurally stable to ~500 °C even when fully charged and releases very little oxygen; its runaway is far less violent (smoke rather than sparks and flame, <20 % mass loss vs >60 % for NMC). The safety ranking from Sallard et al. is LFP >> sodium‑ion >> NMC811.

Practical consequence: a smothering agent (Class D dry powder, sand, a fire blanket, a clean agent) that works by excluding atmospheric oxygen is the wrong tool for a Li‑ion fire, because atmospheric oxygen is not what is sustaining it. The real objectives are (a) cooling the cell and its neighbors below the runaway threshold to stop propagation, and (b) containment and evacuation while the compromised cell burns itself out. This is why fire services now recommend large volumes of water, not dry powder (see §8).

Key takeaway: The "self‑oxygen" claim is substantively valid for layered‑oxide chemistries (NMC/NCA/LCO) — they carry their own oxidizer and cannot be smothered — but it should be stated as cathode oxygen release, not "lithium provides its own oxygen." LFP is markedly less self‑sustaining, though (importantly) its off‑gas is more toxic and more flammable per litre (see §5.4).

2.3 What thermal runaway releases

  • Heat: peak cell temperatures >650 °C (NMC); vented gas >1100 °C; LFP ≈330–410 °C.
  • Flammable gas: NMC cells generate roughly 1.7–2.4 L of gas per Ah at 100 % SOC (Li et al. 2023), and the figure can approach ~3 L/Ah under some conditions; LFP is lower (~0.56–0.78 L/Ah). Composition (≈, NMC, air): CO₂, CO, H₂, and hydrocarbons (CH₄, C₂H₄, C₂H₆, C₃H₈). UL FSRI measured ~7–8 L of effluent gas per 18650 cell, ≈36 % CO, 32 % H₂, 22 % CO₂, 10 % hydrocarbons, with a lower flammability limit of ~10.8 % and an upper limit of ~50.3 % — an exceptionally wide flammable range.
  • Toxic gas: the LiPF₆ electrolyte salt decomposes (LiPF₆ → LiF + PF₅; PF₅ + H₂O → POF₃ + 2 HF) to release hydrogen fluoride (HF) and phosphoryl fluoride (POF₃). Larsson et al. (2017) measured HF at 20–200 mg per Wh of cell capacity — i.e., on the order of 2–20 kg of HF for a 100 kWh EV pack and 20–200 kg for a 1 MWh stationary system. HF is acutely lethal at very low concentrations (AEGL‑3, 10‑min ≈ 170 ppm). This is a hazard to occupants and firefighters even when the battery is only one fuel package in a larger structure fire (§5.4).
  • Pressure: in a sealed enclosure, internal pressure can climb into the bar‑range before a burst vent opens; the resulting flammable‑gas cloud can deflagrate. Open‑air blast overpressure from packs is much smaller (UL FSRI: ~1.7 kPa for a 10‑cell equivalent, ~6.3 kPa for a 100‑cell equivalent) — but ~3.5–7 kPa is already enough to shatter windows, and ~7–15 kPa will detach a garage door. (The earlier notes' "1–5 bar" figure applies to sealed enclosure internal pressure before venting, not open‑air blast; both are real, and they are different things.)

2.4 Real‑world incident used as a reference point

The May 2024 Gateway Energy Storage facility fire in Otay Mesa, San Diego, is a useful utility‑scale illustration of these mechanisms. The 250 MW facility held roughly 14,796 NMC lithium‑ion battery packs (commonly rounded to ~15,000; the earlier notes said "cells," which is incorrect — they are packs/modules). The fire began 15 May 2024 and smoldered with periodic flare‑ups for nearly two weeks (13 days) (through ~28 May) as packs repeatedly re‑ignited. Cal Fire committed 40 firefighters and 5 engines; interior attack was impossible due to toxic fumes, so water was applied from outside supplemented by the internal sprinkler system; evacuations and a shelter‑in‑place order for the nearby Donovan Correctional Facility were issued. Root cause remained under investigation as of May 2025. The episode demonstrates stranded energy, re‑ignition, and the inadequacy of "knock‑down" against self‑oxidizing cells — at industrial scale.

3. Risk Assessment: Causes of Lithium Battery Fires (ordered by danger)

The following ranks the primary causes from highest to lowest risk, considering both the probability of triggering thermal runaway and the severity of the resulting event. This ranking incorporates the §2 chemistry: a cause that produces runaway in a charged, layered‑oxide pack (NMC/NCA/LCO) is inherently more dangerous than the same cause in an LFP pack or in a deeply discharged (low‑energy) cell, because the available fuel and self‑generated oxygen are what turn a fault into a self‑sustaining fire.

1. Recharging Damaged or Deeply Discharged Cells — Singularly or in Packs, Even with a BMS

The single most dangerous action you can take with a lithium battery. When a cell drops below ~1.5–2.0 V, the copper current collector dissolves, releasing Cu²⁺ ions that deposit as metallic dendrites throughout the cell; the cell is permanently compromised. While dormant at deep discharge the risk is relatively low — there is minimal stored energy to drive runaway. The catastrophic failure almost always occurs during or shortly after recharge, when rising voltage redeposits copper as conductive dendrites across the separator (the mechanism for a hard short) precisely while the cell is being refilled with energy (the fuel). In parallel packs, charged neighbors dump their combined energy through a single shorted cell. Never recharge a cell that has dropped below ~1.5–2.0 V. Dispose of it. This applies whether the cell is loose, in a pack, or behind a BMS — the BMS cannot undo copper dissolution. (Full analysis in §4.)

2. Overcharging Cells

Overcharging drives the cathode to unstable potentials, decomposes the electrolyte, and generates heat and gas. Unlike deep‑discharge damage (which sets a trap for later), overcharging can trigger thermal runaway directly and immediately — and per §2.2 the cathode of a charged layered‑oxide cell is exactly the configuration that releases its own oxygen and becomes self‑sustaining. Using an incorrect or failed charger, or charging at low temperature (which plates lithium metal instead of intercalating it), are the most common pathways. Overcharge above ~4.3–4.5 V/cell (chemistry‑dependent) can initiate runaway within minutes.

3. Charging Unbalanced Packs Without a BMS

In a series pack without a BMS, cell voltages drift apart over cycles. A charger that sees only total pack voltage may deliver a "correct" 12.6 V to a 3S pack while one cell sits at 4.5 V and another at 3.8 V. The overcharged cell enters runaway while the charger remains oblivious. Common in hobbyist and DIY contexts where balance charging is neglected.

4. Shorting of Individual Cells or Unmanaged Packs

A direct external short across terminals — from loose change, tools, damaged insulation, or conductive storage surfaces — can drive 50–100+ A through a single cell. Without protection circuitry the cell heats rapidly; if the internal temperature reaches ~80–120 °C the SEI breaks down and runaway becomes self‑sustaining. Bare 18650s and pouch cells without a PCM are especially vulnerable.

5. Physical Damage to Cells or Packs

Crushing, puncturing, dropping, or bending a cell can mechanically rupture the separator, creating an instantaneous internal short with no warning — unlike degradation‑driven shorts that develop over time. Damaged cells from e‑bikes, power tools, or dropped phones should be treated as hazardous and never stored indoors.

6. Shorting of BMS‑Managed Packs

A BMS can interrupt an external short in microseconds, but protection is not infallible: MOSFETs can fail closed, firmware can freeze, and very high‑current shorts can weld contacts before the circuit responds. The BMS reduces risk substantially but does not eliminate it; treating BMS‑protected packs as "safe to short" is itself a risk factor.

7. High‑Rate Discharge of Cells or Packs

Sustained discharge at or above the rated maximum generates significant internal heat. If cooling is inadequate (pack wrapped in insulation, confined space, high ambient temperature), the cell can reach the runaway threshold. More common in high‑power applications (drones, power tools, EVs) than in storage, but it can occur if a stored pack is suddenly loaded without temperature monitoring.

8. External Heat / Involvement in a Structure Fire

Cells exposed to external heat — a building fire, adjacent equipment, a hot vehicle cabin — can be driven into runaway with no electrical fault at all. A car cabin can exceed 60 °C within an hour; most cells tolerate short 60 °C excursions, but prolonged exposure causes electrolyte decomposition, gas generation, and separator shrinkage, and at ~130–150 °C the separator melts and a hard short forms. This is the same mechanism by which a household's batteries become a fuel load in a fire that started elsewhere — see §5. Fire departments explicitly warn against leaving lithium batteries in hot vehicles.

9. Storage of Cells That Degrade Over Time

The slowest‑burning fuse but the one that sets up several of the higher‑ranked risks. Cells left for months or years without voltage monitoring self‑discharge (1–3 %/month, faster if a BMS draws parasitic current). Once voltage drops below ~2.5 V, copper dissolution begins. The dormant degraded cell is relatively safe — but the moment someone attempts to recharge it, it becomes Risk #1.

Risk‑assessment note added in this revision: the ranking above is for initiating events. It does not capture the aggregate, passive risk of a household that simply owns many batteries — a fire‑load and stranded‑energy problem that exists even when every cell is healthy and undamaged. That is the subject of §5, and it should be read as a complement to (not a replacement for) the cause‑based ranking here.

4. Deep Discharge, Internal Shorts, and the Recharge Danger

This section addresses the most critical risk in §3 (#1): recharging damaged cells.

4.1 The deeply discharged single cell: dormant danger vs. recharge danger

When a Li‑ion cell drops below ~2.5 V (and especially below 0 V):

  1. SEI breakdown — the Solid Electrolyte Interphase on the anode decomposes, exposing fresh, highly reactive electrode surfaces.
  2. Copper dissolution — the anode copper current collector oxidizes, releasing Cu²⁺ into the electrolyte. This is the defining damage mechanism of deep overdischarge.
  3. Copper deposition — those ions migrate and deposit on the cathode (and elsewhere) as metallic copper dendrites and bridges.

While dormant the cell is permanently damaged but the immediate runaway risk is low — not zero, but low. A cell at ~0.5 V or 0 V has almost no stored electrochemical energy; even a hard internal short cannot drive enough self‑heating to reach the ~80–120 °C self‑sustaining threshold. The copper dendrites may already be forming micro‑shorts (high‑resistance, causing the accelerated self‑discharge), but these are not the low‑resistance "hard shorts" that produce catastrophic current.

The danger really starts when you recharge. Charging provides the electrochemical driving force that sends dissolved/deposited copper back toward the anode to precipitate as metallic dendrites — this is when copper bridges grow most aggressively — while simultaneously pumping energy back into a structurally compromised cell. Now you have both the mechanism (growing hard short) and the fuel (restored electrochemical energy) for runaway. The result can be rapid, violent thermal runaway during or shortly after charging begins.

Key takeaway: The overdischarge event causes the damage; the cell is permanently unsafe. But the moment most likely to produce catastrophic runaway is when you attempt to recharge it. The copper shunts are permanent scars; you can nurse a cell back to voltage, but you can never restore its original safety margin.

4.2 BMS packs: the one‑bad‑cell problem and the parallel‑group physics

A multi‑cell pack sitting in storage for months/years has its BMS drawing a small quiescent current. One cell — from manufacturing variance, prior damage, or higher self‑discharge — degrades faster.

A critical physics constraint: parallel cells cannot diverge in voltage. Cells bonded in parallel with milliohm‑range interconnects cannot differ by more than a few millivolts — Kirchhoff's voltage law with a time constant of milliseconds to seconds. So in a months‑long slow discharge, there is no mechanism by which one cell sits at 0 V while its parallel neighbor sits at 2.5 V; the whole parallel group drifts down together. The weak cell is not at a lower voltage — it is the one pulling the group down faster.

Real failure sequence:

  • Phase 1 — slow self‑discharge (months): all cells drift down together. Copper dissolution begins in all cells once the group drops below ~1.5–2.0 V.
  • Phase 2 — internal short while deeply discharged: total stored energy is minimal; a short is current‑limited; runaway is unlikely — the pack is dead, not dangerous.
  • Phase 3 — the danger is on recharge: copper redeposits as dendrites as voltage rises; a hard short forms in the most damaged cell while the group is now at significant voltage (e.g., 3.5 V); the charged parallel neighbors dump their combined energy through the fault → localized heating → runaway → propagation to adjacent charged cells → pack‑level fire/explosion.

The parallel configuration creates an energy‑concentration problem (not a voltage‑divergence problem): the combined energy of many cells funneled through one fault. Two sub‑scenarios: (a) progressive/partial shorts — a high‑resistance dendrite short lets terminal voltage remain equalized while neighbors feed current through the partial short, creating a growing hot spot; and (b) hard shorts during recharge — the full combined energy of the group discharges through the fault. A series‑only string behaves differently: a shorted cell becomes a low‑resistance bypass that reduces pack voltage without receiving neighbor current, but the pack is now badly imbalanced and charging it risks overcharging the remaining healthy cells.

BMS failure modes that enable this: (1) BMS parasitic drain is itself the culprit and may not trigger undervoltage protection if the drop is gradual; (2) undervoltage protection fails (component/ firmware); (3) the BMS disconnects but a user bypasses it to "jump‑start" a dead pack; (4) passive balancing left active during storage accelerates drain on all cells.

4.3 Summary table — deep discharge scenarios

ScenarioDormant dangerRecharge dangerNotes
Single cell deep‑dischargedLow (minimal stored energy)HIGH (Cu dendrites + energy input)Permanently damaged; dispose, do not recharge
Parallel group, all cells deep‑discharged togetherLow (minimal group energy)HIGH (dendrites form on recharge; combined group energy through any short)Voltages stay equalized; danger is on recharge
BMS pack, series‑only, one cell deep‑dischargedLow while dormantHIGH if charged; risk of overcharging healthy cellsNo cross‑current path from neighbors
Parallel group with progressive/partial shortModerate (neighbors feed partial short → hot spot)HIGHTerminal voltages equal; internal current asymmetric

4.4 Practical takeaways for deep discharge

  1. Never recharge a cell that has dropped below ~1.5–2.0 V. Dispose of it properly.
  2. For BMS packs in long‑term storage: disconnect the BMS if possible, or every 2–3 months check and recharge to 40–60 % SOC. The BMS itself can be the enemy in prolonged storage.
  3. Parallel configurations amplify danger — a single shorted cell can be heated to runaway by its neighbors. Cell‑level fusing (wire bonds, thin‑foil fusible links) is one design mitigation.
  4. The 3‑month voltage check is a safety check, not just a capacity check — it prevents cells entering the copper‑dissolution zone.

5. Storing Energy in the Home — the passive fuel‑load problem

This is the section the earlier notes lacked, and the reason for this revision. The point is not that batteries start fires (covered in §3). The point is that the growing inventory of batteries in an ordinary home changes the character of any structure fire, even one the batteries did not cause.

5.1 The aerosol‑can analogy

Aerosol spray‑paint cans are not, in ordinary use, treated as "a fire hazard." But fire codes do not treat a large collection of them casually. NFPA 30B (Code for the Manufacture and Storage of Aerosol Products) classifies aerosols into Levels 1–3 by chemical heat of combustion (≤20, 20–30,

30 kJ/g), and FM Global Data Sheet 7‑31 parallels it. The documented hazards of a collection of aerosol cans in a fire are:

  • Fuel load: cans are pressurized to 140–180 psi (design up to 240–400 psi); a pallet represents a substantial heat‑of‑combustion load (FM Global test product ≈14 kJ/g).
  • Fireballs and projectiles: a rupturing can produces a fireball ~8 ft (2.4 m) across, and cans can rocket and trail burning liquid well beyond the fire's origin — enough that NFPA 30B requires chain‑link fencing around segregated storage to restrain rocketing cans, and notes ceiling‑only sprinklers are insufficient.
  • Projectiles: Lithium-ion cells mirror this hazard. During thermal runaway, if the internal safety vent fails, or if pressure builds too rapidly, individual cylindrical cells (e.g., 18650s) can "rocket" as high-velocity projectiles, throwing hot debris and battery parts, which spreads fire far beyond the initial failure point.
  • The pattern, not the item: a single can on a workbench is trivial; a shelf of cans is a fuel and pressure package that intensifies a fire that started elsewhere.

Lithium batteries follow the same pattern — and add three things aerosol cans do not:

  1. They carry their own oxidizer (cathode oxygen, §2.2), so they cannot be smothered.
  2. They store electrochemical energy that re‑ignites hours to days later ("stranded energy," §5.3).
  3. Their off‑gas is acutely toxic (HF, §5.4), not merely flammable.

So the aerosol analogy is apt but conservative: a home full of batteries is a fuel‑load problem at least as serious as a home full of aerosol cans, and on top of that it is self‑oxidizing, re‑igniting, and toxic.

5.2 How much energy is now in a home, and how fast it grew

The energy stored in a typical household's batteries has climbed sharply and is still climbing:

  • Micromobility (e‑bikes, e‑scooters): large Li‑ion packs (often 0.3–1+ kWh each) used by commuters and delivery workers. NYC FDNY / Public Advocate data track the fire consequence: Li‑ion battery fires in NYC rose from 30 (2019) to 267 (2023), with 18 deaths and ~150 injuries in 2023 — the deadliest year for fires in NYC in two decades. The fire count stayed elevated or slightly flat in 2024 (exactly 277 fires, with 6 deaths) — but deaths fell sharply because fires shifted outdoors (133 non‑structural fires in 2024 vs 90 in 2023) and because NYC banned non‑UL‑certified battery sales (Local Laws 39/42/2023) and stepped up enforcement. The raw frequency of battery fires did not drop; the lethality did, because of where the fires occurred.
  • Cordless power tools: now overwhelmingly Li‑ion; a workshop typically holds many packs.
  • Residential BESS (home backup batteries such as Powerwall‑class units): typically 10–15 kWh each, often installed in garages or basements.
  • EVs in attached garages: packs of 40–100+ kWh — by far the largest single battery energy most homes will ever contain.

A modest modern home can easily hold tens of kWh of Li‑ion energy across these devices — and an EV in the garage can push that past 100 kWh. Two decades ago the same home held essentially zero.

5.3 What this means in a structure fire: heat, pressure, and stranded energy

The UL Fire Safety Research Institute, IAFF, and U.S. DOE jointly studied exactly this scenario (Considerations for Fire Service Response to Residential ESS Incidents, Dec 2023): a representative 20 × 20 ft attached two‑car garage with 3 × 17 kWh NCA ESS units (51 kWh total) plus ordinary garage commodities. Key findings:

  • Heat release: a single 18650 NCA cell peaks at ~15–30 kW; a 10‑cell pack at ~50 kW; a 152‑cell module at ~300 kW; and a full unit plus commodity fire reached ~5,000 kW within ~4 minutes. Battery fire growth is non‑linear — the standard t² growth assumption used for ordinary commodity fires does not apply.
  • Deflagration: each 18650 cell generates ~7–8 L of flammable effluent gas (LFL 10.8 %, UFL 50.3 % — a very wide range). In one test ~7,200 L of battery gas accumulated before a self‑ignited deflagration. A single 30‑cell module can produce ~1–4 kPa overpressure — enough to shatter windows; a 100‑cell equivalent produced ~6.3 kPa and threw a garage door 50 ft. In the Erie, CO (April 2023) incident a hybrid EV battery explosion propelled a garage door ~30 ft and struck the incident commander's helmet.
  • Stranded energy and re‑ignition: "stranded energy" is the energy left in undamaged cells of a damaged high‑voltage battery. It causes re‑ignition minutes, hours, or even days later, so scene management is long and resource‑heavy. Best practice is often to let the battery fire burn itself out while protecting exposures rather than declaring victory at knock‑down.
  • Size‑up is blind: there are no reliable visual or thermal‑imaging indicators that confirm battery involvement during an active room‑and‑contents fire; portable gas meters did not distinguish battery fires from baseline commodity fires; and a significant explosion hazard can develop before any exterior sign is visible. Full structural PPE with SCBA is recommended before size‑up, not after.

The implication for a homeowner is direct: a kitchen fire, an electrical fault, or any ordinary ignition in a home that now holds tens of kWh of batteries is a different fire than it was 20 years ago. The batteries will be heated into runaway by the surrounding fire (§3 Risk #8), will add megawatt‑class heat release, will generate a flammable‑gas deflagration hazard, will re‑ignite after knock‑down, and will poison the smoke with HF — regardless of whether any battery was defective.

5.4 Toxic gas: the hazard that reaches beyond the fire room

Because the LiPF₆ electrolyte salt decomposes to HF and POF₃ (§2.3), a battery‑involved structure fire produces hydrogen fluoride in quantities that matter to occupants and firefighters far from the cell that failed. Larsson et al. (2017) measured 20–200 mg HF per Wh — roughly 2–20 kg of HF for a 100 kWh EV pack. HF is acutely lethal at very low concentrations and attacks lungs, eyes, and skin.

A counter‑intuitive but important chemistry nuance (Bugryniec et al. 2024): LFP off‑gas is more toxic per litre than NMC off‑gas — LFP releases less total gas but a higher HF concentration, and its gas has a lower flammability limit (LFP LFL ≈6.2 % vs NMC ≈7.9–9.2 %). So while LFP is the safer chemistry for self‑sustaining runaway (§2.2), a large LFP BESS in a home still poses a serious toxic‑gas and deflagration hazard if it is heated into runaway by a surrounding fire. Choosing LFP reduces the fire‑propagation risk; it does not eliminate the structure‑fire fuel‑load and toxicity problem.

5.5 What a homeowner should do about the passive fuel‑load problem

  1. Reduce the inventory in living spaces and egress paths. Don't charge or store e‑bikes/scooters in hallways, bedrooms, or behind the front door. NYC's 2024 fatality drop came largely from moving fires outdoors — the same logic applies at home.
  2. Separate fuel from fuel. Keep batteries away from combustibles, water heaters, furnaces, electrical panels, and the ignition sources that would heat them into runaway.
  3. Don't charge unattended or while sleeping/away. Charging is the highest‑risk activity (§7).
  4. Use certified devices and chargers. NYC's Local Laws 39/42/2023 banned non‑UL‑certified micromobility batteries for a reason; uncertified and "refurbished" (second‑use‑cell) packs are disproportionately represented in fires.
  5. Notify and plan. If you have a home BESS or an EV in an attached garage, consider a pre‑incident note to your local fire department and make sure your household knows the plan is evacuate and call 911, not fight the fire (§8).
  6. Install detection. Working smoke alarms are the baseline; for a garage/basement with a large BESS, a CO detector is worthwhile (CO is a major battery‑gas component), and note that conventional smoke/CO detectors may not catch the early off‑gassing phase of a failing cell — treat any unusual "sweet/chemical" odor, hissing, or white/gray smoke from a battery as an emergency.

6. Safe Storage Practices

6.1 Temperature

  • Ideal storage: 5–20 °C (41–68 °F). Acceptable: 15–25 °C. Avoid: below 0 °C or above 30 °C for extended periods.
  • Heat accelerates calendar aging: every 10 °C rise above 25 °C roughly doubles the rate of permanent capacity loss (Arrhenius behavior).
  • Never leave lithium batteries in a parked vehicle in sun — a cabin can exceed 60 °C within an hour and drive cells toward runaway (§3 Risk #8). Keep away from furnaces, water heaters, and heating equipment.
  • Sub‑freezing storage is not immediately catastrophic but is not recommended long‑term; it raises internal resistance and risks lithium plating if the cell is later charged while still cold. Allow cold‑soaked batteries to warm to room temperature before charging.

6.2 State of charge (SOC)

  • Optimal storage SOC: 40–60 %. High voltage stresses the cathode and accelerates electrolyte oxidation; deep discharge drives copper dissolution (§4). Never store fully charged or fully depleted. Check voltage every 3 months; if SOC has fallen below ~40 %, recharge to 50–60 %.

6.3 Enclosures and ventilation

  • Metal boxes (e.g., steel ammo cans) give some fire containment but must be vented — an unvented sealed metal container can become a bomb if a cell vents gas and it ignites.
  • Purpose‑built safety boxes/cabinets add fire‑resistant insulation (often rated >1000 °C), pressure‑relief vents or burst discs, gas filtration, and optional smoke detection / automatic suppression.
  • Never use airtight, non‑vented containers for bulk storage. Ventilation is critical: thermal runaway releases large volumes of flammable gas (§2.3). Sealed enclosures need engineered pressure relief that opens below the structural failure point and directs gas away from occupied areas. For BESS containers, NFPA/IFC require deflagration venting per NFPA 68 or equivalent.
  • Small quantities: a LiPo‑safe bag inside a vented metal box is a common hobbyist solution. Larger installations: follow IFC Section 320 for ventilation, exhaust, and explosion control.

6.4 Distribution and separation

  • If one cell or pack catches fire, separation prevents propagation — to other batteries and to the structure.
  • IFC 2024 Section 320 (commercial/industrial): outdoor storage areas ≤900 sq ft (83.6 m²) and ≤10 ft (3 m) height; multiple areas separated by ≥10 ft (3 m); indoor storage in rooms with 2‑hour fire barriers and automatic sprinklers.
  • Home/hobby: store in small, separated groups; use fire‑resistant dividers (ceramic fiber boards) between large packs; keep away from flammable materials and escape routes (§5.5).

6.5 Individual cells vs. BMS‑managed packs

  • Bare cells (18650, pouch) have no built‑in protection. Keep in non‑conductive holders or original packaging; insulate terminals (tape, heat‑shrink, caps) to prevent external shorts (§3 Risk #4); store at 40–60 % SOC; never mix chemistries, capacities, or ages.
  • BMS‑managed packs monitor cell voltage, balance, and provide overcharge/over‑discharge, temperature, and short‑circuit protection — but the BMS draws quiescent current. For long‑term storage, disconnect the BMS or periodically recharge to compensate (§4.4). Treat BMS protection as a secondary layer, not a primary one; still insulate terminals and avoid conductive surfaces.

6.6 Humidity and corrosion

  • Ideal RH: 45–75 % (non‑condensing). High humidity causes terminal corrosion, leakage currents, and seal degradation; very low humidity raises static risk. Store dry; use desiccant in sealed (vented) containers without blocking vents; dielectric grease on terminals for long‑term storage.

6.7 Active maintenance vs. "set and forget"

  • Do not set and forget. 3‑month cycle: check voltage, inspect for swelling/corrosion, recharge to 50–60 % if below ~40 %. Leaving batteries undisturbed for years risks over‑discharge and the recharge trap of §4.

7. Charging and Discharging Discipline

7.1 Charging (highest‑risk activity)

  • Use only the manufacturer‑specified charger with correct voltage/current rating.
  • Charge on a non‑flammable surface (concrete, ceramic tile, metal tray) away from combustibles.
  • Never charge unattended or overnight without a safety system (smoke detector, automatic cutoff). Don't charge in egress paths or while sleeping/away (§5.5).
  • Warm cold‑soaked batteries to room temperature before charging. Charging below 0 °C causes metallic lithium plating and dendrites — extremely dangerous.
  • Never bypass a BMS to "jump‑start" a dead pack (§4.4).

7.2 Discharging

  • Deep discharge below ~2.5 V/cell causes copper dissolution and internal shorts (§4).
  • High‑rate discharge generates heat; without adequate cooling it can reach runaway. Monitor pack temperature during/after heavy discharge; halt if any cell exceeds ~60 °C. Don't discharge a cold‑soaked pack at high rates (elevated internal resistance → more heat).

7.3 Self‑discharge while sitting

  • 1–3 %/month typical; BMS parasitic drain accelerates it. Unmonitored multi‑month storage drops cells below the safe threshold and sets up the §4 recharge trap.

8. Fire Suppression and Emergency Response (corrected)

This section corrects the earlier notes. LithiumFire.md and LithiumStorage.md recommended keeping a Class D extinguisher or a bucket of dry sand for Li‑ion fires. Current fire‑service guidance (FDNY, NFPA) says the opposite for rechargeable Li‑ion: use large volumes of water, not Class D, and not dry chemical.

8.1 Why the old Class‑D recommendation is wrong for Li‑ion

  • Class D extinguishers (sodium‑chloride, graphite, dry‑powder agents) are designed for combustible‑metal fires — lithium metal, magnesium, sodium, potassium — i.e., non‑ rechargeable lithium batteries. A Li‑ion (rechargeable) battery fire is not a Class D fire.
  • Per §2.2, a Li‑ion fire is self‑oxidizing (cathode oxygen release); a smothering agent that works by excluding atmospheric oxygen does not stop it. FDNY states plainly that dry‑chemical extinguishers are "ineffective" for Li‑ion fires and that "fire extinguishers do not work on lithium‑ion battery fires."
  • Sand and dry powder can trap heat against an already‑runaway cell and do nothing about propagation or re‑ignition. They are not recommended as a primary response to a Li‑ion fire.

8.2 What fire services actually recommend (FDNY Haz‑Mat 20, April 2022)

  • Use copious amounts of water from a handline; "water application should continue until conditions are dormant — that is, when no more flame, gas or smoke is being released from the battery or mobility device." Water may not prevent runaway from starting, but if it can penetrate the battery case it cools adjacent cells and reduces propagation.
  • Post‑fire: move the battery/device to a bathtub, sink, or bucket and fully submerge the cells in water. Use non‑conductive tools (e.g., a shovel with a wooden handle). Maintain a charged hoseline. Maintain full PPE with donned facepiece at all times due to re‑ignition risk (minutes, hours, or even days later) — i.e., stranded energy (§5.3).
  • For BESS / large packs, specialized agents (encapsulator agents such as F‑500, AVD — Aqueous Vermiculite Dispersion, water mist) are studied in the literature, and large‑volume water / water‑mist is the mainstay. Sprinkler systems (NFPA 13) are designed to protect the surrounding structure, not necessarily to extinguish the battery fire.

8.3 What a homeowner should actually do

  1. If a Li‑ion battery is hissing, swelling, venting white/gray smoke, or on fire: leave the area, close the door, and call 911. Do not attempt to fight a Li‑ion mobility‑device or pack fire with a portable extinguisher — it will not work and it delays evacuation.
  2. Evacuate everyone and account for pets; HF and CO in the smoke are lethal (§5.4).
  3. If safe and only for a very small, isolated cell (e.g., a single tool pack) that is just beginning to vent and is on a non‑flammable surface outdoors: large volumes of water are the agent; never assume "knock‑down" means "out" — re‑ignition is expected.
  4. Tell the 911 operator a lithium battery is involved so responders bring the right tactic (large water, long scene time, deflagration awareness) and PPE/SCBA.
  5. Do not re‑enter until the fire department explicitly clears the scene; re‑ignition can occur long after visible fire is gone.

The primary strategy for a home is containment and evacuation, supplemented by detection and separation so that a failing cell is less likely to propagate or to ignite surroundings.

9. Regulatory Standards and Codes

  • IFC 2024 Section 320 — first comprehensive code for lithium battery storage: indoor/outdoor limits, fire barriers, sprinklers, signage, ventilation, exhaust, explosion control.
  • NFPA 855 — Standard for the Installation of Stationary Energy Storage Systems (covers residential ESS as well as utility).
  • UL 9540 / 9540A — safety standards for ESS and the fire‑test method for cells/modules.
  • NFPA 30B — Code for the Manufacture and Storage of Aerosol Products (the analogy in §5.1).
  • NFPA 13 / NFPA 68 — sprinkler installation / deflagration venting.
  • EPA BESS guidelines — siting, fire suppression, emergency planning.
  • FM Global DS 7‑112 (Li‑ion manufacturing/storage) and DS 7‑31 (aerosol storage).

10. Best Practices Summary

  1. Chemistry awareness: prefer LFP where feasible — it is markedly less self‑sustaining in thermal runaway, though its off‑gas is more toxic per litre (§2.2, §5.4).
  2. Temperature: store at 5–20 °C; never freeze; never leave in a hot car; keep away from heat sources.
  3. State of charge: 40–60 %; check every 3 months; recharge if below ~40 %.
  4. Container: vented, fire‑resistant enclosure (vented metal box with pressure relief or a certified safety cabinet). Never airtight.
  5. Ventilation & separation: let gases escape; keep batteries in small, separated groups with fire‑resistant dividers; away from combustibles and escape routes.
  6. BMS: prefer BMS‑protected packs; insulate terminals either way; disconnect BMS for long‑term storage to stop parasitic drain.
  7. Charging: correct charger only; room temperature; never unattended/overnight; never charge below 0 °C; never bypass a BMS to jump‑start a dead pack.
  8. Never recharge a deeply discharged cell (below ~1.5–2.0 V) — dispose of it (§4).
  9. Maintenance: 3‑month voltage + physical inspection; dispose of any cell below safe voltage or any swollen/dented/corroded cell.
  10. Home fuel‑load (§5): keep batteries out of living spaces and egress paths; separate from ignition sources; use certified devices; install smoke + CO detection; if you have a BESS/EV, plan with your household and consider notifying your fire department.
  11. Fire response: evacuate, close the door, call 911, say "lithium battery." Do not rely on a portable extinguisher or Class D / sand. Fire services use large volumes of water; expect re‑ignition and long scene times.
  12. Compliance: for commercial storage follow IFC 2024 §320, NFPA 855, UL 9540/9540A, and local AHJ rules.

11. Latent Fuel of Lithium Batteries

A typical high-quality 18650 lithium-ion cell:

  1. Electrochemical Energy: It holds about 3 Ah at 3.6 V, which equals 10.8 Wh (or roughly 39 kJ) of electrical energy.
  2. Combustion Energy: The cell weighs roughly 45 grams. Lithium-ion cells have an average heat of combustion of roughly 12 MJ/kg. 0.045 kg * 12 MJ/kg = 0.54 MJ (or 540 kJ)
  3. The Ratio: 540 kJ / 39 kJ ~ 13.8.

This means the battery contains nearly 14 times more latent chemical energy (combustion fuel) than its electrical storage capacity. And, it still holds this energy even if it is electrically drained. Dead batteries still burn well and they still release toxic gas while burning. However, to put this in perspective, that is equivalent of 11 ml of gasoline. A typical eScooter battery pack contains 30 cells, roughly equivalent to 1/3 of a liter of gasoline.

12. References

Thermal‑runaway chemistry and mechanism (§2, §4)

Gas volume, composition, toxicity (§2.3, §5.4)

Fire‑service guidance and residential ESS fuel load (§5, §8)

Aerosol‑can analogy (§5.1)

Incidents and statistics (§2.4, §5.2)

Storage practice, temperature, SOC, BMS (§6, §7)

Codes and standards (§9)

Public‑facing consumer guidance

Posted by at No comments:
Subscribe to: Posts (Atom)