Off-gassing

What it is (precise)

When an LFP or NMC cell is driven past its thermal limit by an internal short, overcharge, over-temperature or mechanical damage, the electrolyte breaks down and internal pressure rises until the safety vent opens. The escaping vapor is not yet burning: it is a cloud of flammable and toxic gas.

Typical constituents are hydrogen (H2), carbon monoxide (CO), carbon dioxide (CO2), methane and other light hydrocarbons, plus electrolyte vapor and hydrogen fluoride (HF) from decomposition of the fluorinated lithium salt. The mixture is largely invisible and mostly odorless, which is exactly why it is dangerous inside a sealed enclosure.

Off-gassing is the first stage of the failure cascade: incipient venting precedes smoke, smoke precedes flame, and flame precedes Propagation to neighboring cells. In many large prismatic cells the first electrolyte-vapor release appears minutes, sometimes tens of minutes, before ignition, and that window defines the entire enclosure safety strategy.

Detecting the gas rather than the heat or the fire is what buys an operator time to open contactors, stop charging, start exhaust and warn responders. Design the whole protection scheme around this earliest signal, because every later detector is one that fires too late.

Gas quantity and energy scale with cell size, chemistry and state of charge. A cell vented at high SOC releases substantially more gas, at higher temperature, than the same cell near empty, because more stored electrochemical energy drives the decomposition reactions. This is why UL 9540A testing is run at 100 percent SOC, and why some hazard analyses also examine partial-SOC behavior. Note that SOC, not state of health, is the variable here: a brand-new cell and an aged one both off-gas hard when full, so a fresh install offers no reprieve.

Why it matters in a real grid-scale project

In a containerized or walk-in BESS enclosure, off-gas accumulation is first a deflagration hazard. Hydrogen and CO build up inside a sealed, energy-dense space; if the mixture climbs into its flammable range and finds a spark, the result is an explosion that can flatten the enclosure and kill responders. The 2019 McMicken incident in Arizona is the canonical case: the battery fire stayed inside one rack, but accumulated vent gas deflagrated when firefighters opened the door, seriously injuring four. The explosion, not the flame, is the headline risk you design against.

Off-gas behavior also carries direct commercial weight. Gas composition and volume measured in UL 9540A testing feed the Hazard Mitigation Analysis, the NFPA 855 permitting package, the Deflagration venting or purge sizing, the insurance position and the final Authority Having Jurisdiction sign-off. A project that cannot show credible early detection plus a quantified gas-management design can stall in permit review for months, and insurers increasingly price premiums directly off these submittals. These numbers gate whether the site gets built, energized and covered.

Typical values and standards

Vent-gas composition varies by chemistry and test condition, but published cell data typically shows hydrogen at roughly 20 to 50 percent of the mixture, CO2 around 20 to 40 percent, CO around 10 to 30 percent, with hydrocarbons filling the rest.

Hydrogen's Lower Explosive Limit in air is about 4 percent by volume, methane's about 5 percent, CO's about 12.5 percent; the blended vent gas commonly lands near 6 to 10 percent. NFPA 69 ventilation-based prevention is designed to hold the worst-case concentration below 25 percent of that LEL, the single most load-bearing number in enclosure gas design.

Chemistry matters, but not the way marketing suggests. LFP, the dominant stationary chemistry, has a higher runaway onset (typically around 200 C and above, versus roughly 150 to 210 C for NMC) and releases less total heat, yet its vent gas is proportionally richer in hydrogen, so the explosion hazard per liter can be worse even as the fire hazard is lower. Both chemistries off-gas H2 and CO in quantity. EV packs face the same physics, but the stationary problem is distinct: the gas collects inside a fixed enclosure instead of dispersing outdoors.

The standards map rewards keeping straight. UL 9540A is the test method that characterizes thermal runaway and measures vent-gas composition, volume, LEL and burning velocity at cell, module, unit and installation level. UL 9540 is the separate product safety certification for the ESS as a system; the two are not interchangeable despite the shared number.

IEC 62619 covers safety of the cells and batteries for industrial use. NFPA 855 is the installation standard mandating explosion control, delivered via NFPA 68 (deflagration venting) or NFPA 69 (prevention by ventilation, purge or inerting), with alarm practice drawing on NFPA 72.

How it shows up in specs, studies and contracts

The first place a working engineer meets off-gassing is the vendor's UL 9540A test report. Confirm it exists at the level you install (unit level for containers, not just cells), and that it reports gas volume per cell, gas composition, mixture LEL and burning velocity, and whether runaway propagated between modules. Those exact numbers size deflagration vent panels under NFPA 68 or exhaust rates under NFPA 69. A report missing usable gas data forces conservative assumptions and larger, costlier explosion-control hardware, so treat a thin or cell-only report as a commercial red flag.

Downstream, off-gas data reappears in the Hazard Mitigation Analysis filed with the AHJ, the fire-protection drawings (detector type, count and placement), the Emergency Response Plan that tells firefighters when not to open a door, and commissioning scripts that prove detector-to-action sequences actually fire. Insurance questionnaires now routinely ask which gases are sensed, at what thresholds, and what automatic response follows. If any one of these documents disagrees with the test-report numbers, expect the Authority Having Jurisdiction to catch it and send the package back for rework.

Questions worth asking every integrator: which species do the off-gas detectors sense (H2, CO, electrolyte VOCs) and at what ppm setpoints; does detection automatically open contactors and stop charging or merely raise an alarm; does it start exhaust ventilation sized to hold below 25 percent of LEL; how are sensors calibrated and replaced across a 20-year life; and does the detection chain stay powered and functional after the container loses auxiliary supply. Vague answers here separate a paper safety case from one that works at 3 a.m.

Common pitfalls

The most common error is treating LFP's fire-safety advantage as an off-gas advantage. LFP cells genuinely resist ignition, which means a venting LFP rack is more likely to fill an enclosure with unburned, hydrogen-rich gas rather than consume it in flame, arguably raising deflagration risk relative to a chemistry that ignites promptly. Explosion control is therefore not optional for LFP: NFPA 855 requires it regardless of chemistry unless large-scale fire testing earns an exemption. Choosing LFP and skipping gas management is the exact trap the standard closes.

Two more trip-wires recur in reviews. First, water sprinklers and clean-agent systems do nothing about accumulated gas: suppression and explosion control are separate systems solving separate problems, and a Hazard Mitigation Analysis that conflates them will fail review.

Second, detector engineering is easy to botch, because hydrogen rises while some electrolyte vapors are dense and sink, so sensor placement must match the sensed species, and electrochemical H2/CO sensors drift and expire, so the O&M contract needs an explicit calibration and replacement schedule or the early-warning layer quietly dies within a few years.