When Does Thermal Runaway Start?

A lithium-ion battery does not go from normal operation to fire in a single step. The question of when does thermal runaway start matters because the earliest failure phase often begins well before flame, smoke or a sharp temperature spike is visible on site. For operators of BESS, UPS rooms, EV charging infrastructure and battery manufacturing lines, that timing difference is the gap between a controlled response and a major incident.

In practical terms, thermal runaway starts when heat generated inside a cell begins to exceed the cell’s ability to dissipate it, triggering self-accelerating decomposition reactions. That sounds simple, but in live infrastructure the real answer is more nuanced. The start point depends on cell chemistry, state of charge, internal defects, charging conditions, mechanical damage, ambient temperature and whether early off-gassing is detected and acted on.

When does thermal runaway start in a lithium-ion battery?

Thermal runaway starts at the point where an internal fault or abuse condition causes exothermic reactions to sustain themselves. Before that threshold, a battery may be stressed, damaged or overheating, but it has not yet entered true runaway. Once the cell crosses that line, the internal reaction generates more heat than can be removed, and the failure escalates rapidly.

For most lithium-ion cells, that sequence begins with a precursor event. It may be an internal short circuit, lithium plating, separator damage, overcharge, external heating or contamination introduced during manufacturing. The cell then enters a stage of abnormal heating and gas generation. Electrolyte vapours and gases such as hydrogen can appear before ignition. That early release is operationally significant because it creates a detectable warning window.

This is where many site teams get caught out. If thermal runaway is defined only as the point of fire, response options are already limited. If it is understood as a process with an initiation stage, there is a chance to isolate, ventilate, alarm and intervene before propagation.

The stages before visible failure

A failing battery usually moves through a chain of events rather than a single catastrophic jump. First comes the initiating fault. Then comes local heating inside the cell. As internal materials degrade, the electrolyte and electrode materials begin to break down and release vapours. Pressure rises. At a more advanced stage, the separator can fail, creating more severe shorting and a much faster heat rise. Only after that do you typically see venting, smoke, fire or explosive pressure events.

The critical point for infrastructure operators is that off-gassing can occur before thermal imaging or standard fire detection provides useful warning. A room may still appear stable while a cell is already in distress. That matters in enclosed battery rooms, containerised BESS and high-density cabinets where a single failed cell can escalate into rack-level or system-level consequences.

As a result, asking when does thermal runaway start should not be limited to a temperature figure. It should also be framed as: when does detectable battery failure begin, and what signals appear before ignition?

Temperature matters, but it is not the whole answer

There is no universal temperature at which every lithium-ion battery enters thermal runaway. Published onset ranges vary widely depending on chemistry and test conditions. Some cells may begin critical decomposition at around 80 to 120 degrees Celsius internally, while more severe runaway behaviour may occur above 150 degrees Celsius or higher. Those numbers are useful for laboratory understanding, but they are less useful as the only basis for field protection.

Why? Because the internal cell temperature that matters is not always the surface temperature you can measure. A cell can develop a dangerous internal hotspot while the external casing still looks relatively benign. Thermal cameras and temperature sensors remain important, but they may lag the earliest chemical warning signs.

State of charge also shifts the risk profile. Highly charged cells generally contain more available energy and may enter runaway more violently. Cell format makes a difference too. Pouch, cylindrical and prismatic cells fail differently, and packaging at module or cabinet level changes how heat and gases accumulate.

For engineering teams, the practical takeaway is clear: temperature monitoring is necessary, but it should not be treated as the earliest or only indicator.

Common triggers that lead to runaway onset

The most common triggers fall into electrical, mechanical, thermal and manufacturing categories. Electrical abuse includes overcharge, over-discharge and current conditions outside design limits. Mechanical abuse includes crush, penetration, vibration damage or impact during transport and service. Thermal abuse includes external heating, poor ventilation and adjacent cell failures. Manufacturing defects include burrs, contamination, misalignment and latent separator weaknesses that may remain hidden until the cell has aged in service.

Ageing adds another layer. Cells do not fail in the same way at year eight as they do in week eight. Repeated cycling, high operating temperatures and poor balancing can increase internal resistance and accelerate degradation. That does not mean older systems are automatically unsafe, but it does mean detection strategy should reflect lifecycle reality rather than nameplate assumptions.

In Australian conditions, ambient heat can amplify these stresses. Containerised systems, remote installations and sites with constrained airflow may operate closer to thermal limits for longer periods. Design margins, HVAC performance and ventilation logic become part of the runaway prevention story, not just battery chemistry.

Why off-gassing is the earliest practical warning

Before a cell reaches flaming combustion, it can release a mix of gases and electrolyte vapours as internal components decompose. Hydrogen is one important marker, but it is not the only one. Electrolyte-related compounds such as DEC and DEMC can also be present. Detecting these by-products gives operators a practical early warning of abnormal battery behaviour.

That early warning window is where engineered protection becomes valuable. If a detector identifies off-gassing at the incipient stage, control actions can begin before fire detection activates. Ventilation can be triggered, charging can be stopped, isolation logic can be engaged, alarms can be escalated through SCADA, and site teams can follow a defined response procedure.

This is a materially different safety posture from waiting for heat, smoke or flame. By that stage, damage containment is the goal. Earlier in the sequence, incident prevention is still possible.

For critical infrastructure, that distinction affects more than fire risk. It also affects uptime, asset preservation, incident investigation burden and compliance confidence.

What site teams should look for

In operational settings, early signs of battery distress may include unexplained gas presence, rising internal cabinet pressure, intermittent fault alarms, unusual odours, localised temperature anomalies, swelling in some form factors, or repeated balancing irregularities. None of these signs should be interpreted in isolation. Some are subtle, and some may have non-critical explanations. But together they point to a need for layered detection and disciplined response.

This is also why generic room smoke detection is not enough for lithium-ion risk management. Smoke detection is designed for combustion products, not necessarily the earliest vented gases from a failing cell. In a high-value BESS or data centre environment, that timing gap can be costly.

An engineered detection layer designed specifically for hydrogen and electrolyte vapours closes part of that gap. It gives facility managers and integrators a signal that is more closely aligned with the chemistry of early battery failure.

What happens after thermal runaway starts

Once true thermal runaway begins, response becomes more difficult and more time-critical. Heat release accelerates, venting intensifies, and neighbouring cells may be exposed to heat and flammable gases. In tightly packed systems, propagation risk becomes the central concern. Fire suppression and emergency response planning are still essential, but neither replaces the value of detecting the problem before self-heating becomes unstoppable.

That is the reason solution design should focus on the full hazard timeline. Detection, ventilation control, system isolation, alarm routing and response procedures all need to be aligned. A detector by itself is not a strategy. Nor is a SCADA alarm without predefined cause-and-effect logic.

For asset owners and project teams, the better question is not simply when does thermal runaway start. It is whether your site can recognise the signs before that point, and whether the system is configured to act on them fast enough.

NexaGuard’s approach to intelligent early detection is built around that operational reality. In high-energy environments, the safest incident is the one that never progresses.

Battery risk is rarely about a single number on a datasheet. It is about understanding failure progression, designing for early intervention and giving operators enough time to make the right call before a local fault becomes a site event.