Zero-G Fire AtlasNASA Space Apps
Insights

Interpretation

What the evidence means for crews

Seven take-aways from the 24 sources. Each card shows the verbatim evidence, how confident we are, and the evidence that pushes back.
01
⚠

1g tests are not conservative

Materials can burn in orbit below the oxygen level that passes on Earth.

High confidence

Several independent real-microgravity sources (ISS, drop towers) agree.

NASA screens materials with an upward-burning test in normal gravity. With a gentle ventilation flow, microgravity flames survived at lower oxygen than that test allows — so a pass on Earth is not a guaranteed margin in space.

Evidence

“The 1g blowoff for the ¼” rod is at 16% oxygen, whereas we obtained steady spreading at 13.6% oxygen in BASS-II. Thus the material has a negative Oxygen Margin of Safety of 2.4% O2, and the normal gravity flammability limits should be de-rated by at least that much.”
“Because of the known enhancement of microgravity flammability by ventilating flows, the safety factor afforded by normal-gravity testing is now suspect.”
“These tests show the enhanced flammability in microgravity for this geometry, since, in normal gravity air, a flame self-extinguishes in the opposed flow geometry (downward flame spread).”
“Data on right shows Oxygen Margin of Safety (negative means material burns at lower O₂ compared to normal gravity!).”
““Improved methods for screening materials in terms of flammability in space environments will enable safer space missions. Present tests, performed in normal gravity, are not adequate for reduced gravity scenarios.””

Pushes back

“In upward-burning, normal-gravity tests, the 0.27-mm samples were almost always completely consumed; the 0.61-mm thickness burned on average 7.6 cm of the available 30 cm before extinguishing; and the 1.03-mm thickness would not burn. The 0.37-mm downward-burning sample was completely consumed in normal gravity. The results indicated that flammability of silicone in microgravity appears to be reduced compared to normal gravity, at least for the flow speeds tested.”
02
≋

Airflow is the control knob

Too little flow starves the flame, a little flow feeds it, too much blows it off.

High confidence

Mir, ISS (BASS, BASS-II) and Saffire all measured the same trend.

Without buoyancy, oxygen arrives only by the cabin's ventilation. Below a limiting velocity the flame dies; at a few cm/s it is most dangerous and hardest to see; above ~40 cm/s it can blow off. Cutting the fans is the station's first fire response.

Evidence

“Nominal values are Vlim <0.3 cm/s for Delrin, 0.5 cm/s for PMMA, and 0.3 to 0.5 cm/s for polyethylene: These values are lower than those obtained in prior ground testing.”
“Microgravity flames were found to be especially sensitive to air flow speed in the range 0 to 5 cm/s.”
“Flame nearing blow off extinction as the flow velocity crosses 40 cm/s.”
“The flame burned for the allotted test time of about 7 min. (420 s). When the air flow was shut off, the flame immediately extinguished. After the flame had passed, exothermic surface smolder spots burning some of the leftover fuel were visible.”
“Lack of buoyant flow in microgravity reduces oxygen transfer into the combustion zone, which leads to flame extinction when the flow velocity is less than the limiting flow velocity Vlim for the material. The ISS FGB fire-safety system was developed based on this phenomenon.”

Pushes back

“Extinguishment [7]. Extinguishment of thin charring fuels at increased oxygen cannot be assumed to occur rapidly in quiescent conditions. Saffire V-2 (cotton jersey): low flow for 35 s; flow was off for 70 seconds. Plot axes: Flow Speed (cm/s), Sample Temperature (deg C), Relative Time (s), Sample Thermocouple #8. Media: 10 X speed video, top view.”
03
CO₂

CO₂ may be weaker in microgravity than helium

The ISS extinguisher gas can radiate heat back into the fuel.

Medium confidence

Drop-tower and parabolic tests only (seconds of µg); not yet repeated on ISS.

In microgravity flames are thicker and radiation dominates. CO₂ absorbs and re-radiates that heat to the fuel, so in some drop-tower tests flames spread faster in CO₂ than in N₂ or He. Helium was about 11× more effective by mass and has no physiological effect.

Evidence

“Second; helium is much more effective on a mass basis (by about 11 times) at µg.”
“CO2 may not be as effective as an extinguishing agent at µg as it is at earth gravity in some conditions because of the differences in spread mechanisms between the two cases.”
“On the other hand, CO2 dilution case resulted in the highest flame spread rate at Ve=0cm/s and decrease monotonically with increase in flow velocity.”
“On this plot, there is no discernable boundary between flames that were extinguished by suppressant and those that were not.”

Pushes back

“Results show, first, that the substitution of carbon dioxide as the atmospheric diluent for nitrogen greatly diminishes the flammable range; the limiting oxygen concentration for combustion is increased from 21 vol% to near 35 %.”
04
☁

Gas is the main hazard, not heat

Saffire VI: temperature barely moved, but CO and CO₂ crossed crew limits.

High confidence

Directly measured in three large-scale Saffire flights.

In a 17–19 m³ spacecraft, 1.5–3.2 kW fires raised temperature only near the fire (far-field ≈ 20–21 °C), while PMMA burns pushed CO above the 24-hour limit and CO₂ above the 1-hour limit.

Evidence

“The human hazard from a fire in a spacecraft is very similar to terrestrial fires where the heat from the fire is not the principal hazard and the smoke and gaseous products are a much greater concern.”
“The temperature only increased significantly very near to the fire itself.”
“For the PMMA tests, ambient concentrations of carbon monoxide exceeded the 24-hour carbon monoxide limit.”
05
◌

Small flames can hide, then flare up

A dim blue flame on a thick, warm fuel can burn for minutes unseen.

High confidence

ISS (BASS) long-duration observations plus Mir colour observations.

At low flow, flames turn dark blue and nearly invisible, yet thick fuels store heat and keep them alive for over 5 minutes. A sudden increase in airflow — or a nearby fuel — can make them flare, and a nitrogen jet failed to extinguish one.

Evidence

“This has implications for spacecraft fire safety since a tiny flame might be undetected for a long time.”
“While the small flame is not particularly hazardous if it remains small, the danger is that it might flare up if the air convection is suddenly increased or if the flame spreads into another fuel source.”
“During PMMA combustion, the flame is white at V > 1 cm/s, due to the presence of soot particles, and it becomes dark-blue and almost invisible at V = 0.75 to 0.5.”
“Thicker solids can store a great deal of heat even with only a small flame. One result is that extinction with the nitrogen jet was not possible for the given flow rate.”
06
⇥

Flames get pinned at edges — and cast an oxygen shadow

Thick fuels don't keep spreading; the upstream flame starves the ones behind it.

Medium confidence

Seen in Saffire IV–V (real µg), single research team; values only in plots.

On thick PMMA in concurrent flow the flame stayed pinned to the sample's leading edge and reached a limiting size. Once a flame reached that size, flames downstream were weakened or extinguished by the oxygen 'shadow' of the upstream flame.

Evidence

“Flames on thick fuels in concurrent flow did not spread and were instead pinned to the leading edge of the sample.”
“Once a flame achieved its limiting size, downstream flames were inhibited or extinguished by the oxygen “shadow” of the upstream flame.”
“Thick fuel flame growth was found to be dissimilar from that seen in NASA STD 6001 Test 1: concurrent flames established a finite size and did not grow further. This does not invalidate the test but does affect the interpretation of the test results.”
07
↷

A flame can jump a 4.5 cm gap

Pre-heated material nearby can ignite across open space.

Medium confidence

One observed event in orbit, consistent with Mir flame-shape findings.

In Saffire V a flame crossed a 4.5 cm gap to ignite a PMMA surface it had already warmed. Near-limit flames that are short and wide raise the same risk of igniting adjacent surfaces.

Evidence

“In one case, a flame was seen to jump a 4.5 cm gap to ignite a previously warmed PMMA surface.”
“Also, as V decreases, the flames become less visible but expand, increasing the probability of igniting an adjacent surface.”
“This mode might be the most dangerous one, because the probability to ignite an adjacent element is high.”

Where the papers disagree

Conflicts are shown side by side, with our reading of why.

Is microgravity safer or more dangerous?

Less flammable: reduced flammability ranges and spread rates in still µg air.

More flammable: with ventilation, materials burn at lower O₂ than in 1g.

Reading: Both are true — the deciding variable is airflow. Still air starves µg flames; ventilation flows of a few cm/s make them more dangerous than on Earth.

Does turning off the airflow reliably extinguish a fire?

Yes: flow shutoff extinguished every Mir flame in 5–20 s, Saffire I instantly.

Not always: thin fuels at high O₂ and bubbling polymers survived no-flow periods.

Reading: Shutoff works for thin fuels in air; in enriched oxygen or on hot thick fuels it may not, and a small residual flow can help cool the material.

Is CO₂ a good extinguishing gas in space?

Yes: as a diluent it raises the oxygen needed to burn from 21 % to ~35 %.

Maybe not: in µg it can radiate heat to the fuel and speed spread.

Reading: CO₂ behaves differently once radiation dominates. Helium deserves study for exploration vehicles.

Can low pressure on Earth stand in for microgravity?

Approximately: spread rate and flame size approach Saffire II values.

With caution: colour/soot differs and the physics behind the slowdown is different.

Reading: Use low-pressure tests to screen and to estimate spread rate, not to predict flame colour, soot or toxic products.

Saffire VI-2 heat release values

Peak heat release reported as 3217 W.

Average heat release in the same row is 3650 W — higher than the peak.

Reading: Likely two different measurement methods (cabin vs fuel-consumption based). The dashboard shows both values as reported and uses 3217 W for the headline.