Which materials survive atomic oxygen in low Earth orbit?

At a glance: erosion yield and density

What is atomic oxygen in low Earth orbit and why does it matter for spacecraft?

Atomic oxygen is single oxygen atoms, and in low Earth orbit it is the main reason exposed plastics lose thickness. General engineering knowledge, not in our data: sunlight splits ordinary oxygen molecules high in the atmosphere, and the atoms that result are very reactive. A spacecraft hits them at orbital speed, with enough impact energy to break the bonds in most polymers, and the yield depends on that energy and on the flux. The attack is mostly line-of-sight, so faces that point into the direction of travel suffer most, but atoms scatter and can undercut a coating at a gap or pinhole.

Our data shows what this does to a common spacecraft film. On MISSE 2, Kapton HN eroded at 2.8e-24 cm3/atom. On LDEF (the Long Duration Exposure Facility, a NASA satellite brought back after about 6 years), aluminized Kapton blankets of unstated grade took a similar fluence, 8.32e21 atoms/cm2. Most of the outermost 0.076 mm layer was lost, but the thinner layers beneath, protected by their aluminum coating, survived and the blanket stayed functional. The Kapton record notes that LDEF showed a cover is enough to shield a susceptible material, and the Space Station blanket design puts a glass-fibre Beta cloth cover on top.

How is atomic oxygen erosion yield measured, and what do the units mean?

Erosion yield is the volume of material lost for each atom that hits the surface, found by weighing a sample before and after exposure. The mass lost is divided by the density, the exposed area and the fluence, as the Teflon AF record describes. The unit is cm3/atom, and multiplying a yield by a fluence gives the depth lost. A value ending in e-24 means ten to the minus twenty-four, and one ending in e-25 is ten times smaller. Every chart starts at zero, so the fluoropolymer bars look very short; the gap from PTFE to POM is 64.4 times, more than an order of magnitude.

Worked example, MISSE 2 tray only. Multiply the Kapton HN yield by 8.43e21 atoms/cm2 and the product is about a quarter of a millimetre, roughly 240 micrometres, lost in 3.95 years. That tray's fluence is roughly a quarter of its total per year, but this is not an orbit-wide figure. A bare 0.076 mm Kapton layer at the LDEF fluence of 8.32e21 atoms/cm2 should therefore be eroded away entirely, which matches LDEF losing most of that layer, with the aluminum coating beneath it stopping further loss.

Thickness lookup at the same full MISSE 2 fluence: FEP loses about 17 micrometres and PTFE about 12. A 25 µm FEP film would lose about two thirds of its thickness, a 50 µm film about a third and a 125 µm film about a seventh. Scale by mission length only for a similar orbit and facing, and test before you rely on it.

Conditions differ between sources. Most values come from the MISSE 2 polymer tray: 3.95 years, 5000 to 6700 equivalent sun hours (a count of ultraviolet dose) and a base plate cycling nominally between -30 and 40 °C. The fluence was taken from Kapton H witness samples, reference films flown beside the test samples, which read 3.00e-24 cm3/atom; the other records do not all repeat it. Ground beams and plasmas are a different exposure: the Kapton HN record names the ASTM E2089 ground method but enters no ground value. General engineering knowledge: yields can change as a film ages under ultraviolet light and radiation, so a MISSE 2 yield is a starting estimate.

Units can differ too. The polystyrene record holds a laser-beam ground yield in g/atom, mass per atom, next to its flight yield in cm3/atom, and the two cannot be compared without the density, so we chart only the flight value. A greater-than sign marks a lower bound, as for PMMA, PEI, PBI and ADC. NASA also published predicted yields; the ETFE record says its prediction was about 45 percent above the measured value, so predictions are a cross-check, not a measurement.

Which polymers resist atomic oxygen best?

The fully fluorinated polymers resist atomic oxygen best: PTFE 1.4e-25 cm3/atom, then PFA, Teflon AF 1601 and FEP, each in the chart. The word Teflon is a brand name used for more than one fluoropolymer (general engineering knowledge), so name the exact polymer.

Partial fluorination gives partial protection. PCTFE and ETFE erode about three times slower than Kapton HN, and the ETFE record lists atomic oxygen resistance as a strength. PVDF reads 1.3e-24 cm3/atom. ECTFE is the poorest charted at 1.8e-24 cm3/atom, close to Kapton, and its record calls it a poor choice for an exposed orbital surface. Unpigmented polyvinyl fluoride is worse than Kapton HN at 3.2e-24 cm3/atom.

A pigment can matter as much as the polymer. White polyvinyl fluoride reads 1.0e-25 cm3/atom, the lowest value among well-sourced profiles, and the record credits the titanium dioxide pigment. It also warns that thin films erode faster at their edges and can peel, removing the protective ash layer, so the protection lasts only while the layer stays in place.

Low erosion is not the whole story. The PTFE record calls it soft and says ionizing radiation breaks it down at a few megarads, the PFA record sets a dose limit of 10 kGy to watch, and the Teflon AF record calls it a coating, not a structural material. General engineering knowledge: ultraviolet light and charged particles can embrittle and roughen FEP and PTFE, so aged film may erode differently from the pristine films tested.

Which polymers erode as fast as Kapton in low Earth orbit?

Most heat-resistant engineering polymers erode at about the Kapton rate. Kapton HN sits at 2.8e-24 cm3/atom, PEEK at 3.0e-24 cm3/atom, PET (Mylar) at 3.0e-24 cm3/atom, polysulfone at 2.9e-24 cm3/atom and polycarbonate at 4.3e-24 cm3/atom. Only Kapton HN is charted among the polyimides; the others are below our bar, and one carries a table conflict of the same kind as PBT, so we do not claim they match it.

Two members are lower bounds, and they differ. PEI is at least 3.3e-24 cm3/atom, already above Kapton HN, so it is worse than Kapton. PBI is at least 2.2e-24 cm3/atom, below Kapton HN, so it could be better than Kapton or far worse; the data does not say.

The aramids are the exception: Nomex erodes at 2.2 times the Kevlar para-aramid fibre rate. Our data does not say whether the Kevlar sample represents woven fabric, which has far more surface per volume than a film (general engineering knowledge), so do not read it as a durability claim for Kevlar cloth. For any polymer in this band, a cover, a metal coating or a position inside a blanket is what keeps it alive.

Which common plastics erode fastest in atomic oxygen?

Polyoxymethylene (acetal) is the fastest among well-sourced profiles: POM reaches 9.1e-24 cm3/atom, 3.3 times the Kapton HN value. PBT is a near-tie at 9.1e-24 cm3/atom, but a second NASA table prints a value ten times lower for it. The record does not resolve the conflict, so assume the worse. PMMA (acrylic) is at least 5.6e-24 cm3/atom and ADC (CR-39) at least 6.8e-24 cm3/atom, and either could be higher. A polyimide profile below our bar holds a higher value still, but it carries a similar table conflict, so we do not rank it.

Between those extremes the group is mixed. Polystyrene reads 3.7e-24 cm3/atom, nylon 6 3.5e-24 cm3/atom and cellulose acetate 5.1e-24 cm3/atom, which is 1.9 times the Polypropylene (PP) value of 2.7e-24 cm3/atom. A few fall below Kapton: nylon 66 at 1.8e-24 cm3/atom, polyurethane at 1.6e-24 cm3/atom and ABS at 1.1e-24 cm3/atom. But ABS is still 7.7 times the PTFE value, so it is not a durable exposed surface.

Two materials need a caution. Polyethylene is missing from the chart: its profile passes our bar but holds no erosion yield, since its one figure came from a search snippet. Epoxy is missing because its profile is below our bar, but that profile holds one flight value about one and a half times the Kapton HN value, so treat epoxy as Kapton-class until a better-sourced record exists.

What about metals, oxides, glass and silicones?

Our data holds no erosion yield for any metal, oxide, glass or silicone that meets our bar, so we cannot rank them against polymers on one scale. A metal does not lose volume like a plastic; it grows an oxide film, and the question is whether that film protects the metal or cracks and falls away. A pyrolytic graphite profile, below our bar, holds one flight value between the fully fluorinated polymers and PCTFE, but we do not rank it on one sample.

Silver (Ag) is the clear failure. Atomic oxygen swiftly forms a black oxide that cracks and spalls, and the record rates this severe. On the Solar Maximum Mission, Inconel-protected silver on Teflon showed near total silver reaction and flaking after about 4 years. On LDEF, silver beneath a micrometeoroid hole in the FEP cover oxidised in rings around the hole, and oxygen roughened silvered FEP from mirror-like to diffuse. The Gold (Au) record adds that NASA's connector tutorial prohibits silver finishes because of atomic oxygen and asks for at least 50 microinches of contact gold.

Copper (Cu) forms Cu2O and CuO; a sputtered film flown on LDEF thickened from 74.2 nm to 106.7 nm, though no exposure duration or reaction yield was found. The Molybdenum disulfide (MoS2) lubricant film is oxidised quickly at the surface, which can leave a high-friction layer after launch. Aluminium is unconfirmed: thin uncoated films appeared to protect themselves with their own oxide in a ground oxygen-plasma test, but the record rates this low and has only an abstract with no numbers.

For glass, oxides and silicones we have only general engineering knowledge, not in our data. A material that is already an oxide, such as silica, cannot be oxidised further, which is why thin oxide coatings are a common barrier on polymer surfaces. Silicones are generally understood to form a silica-like surface layer, and silicone vapour from nearby parts can deposit on other surfaces and turn to silica under atomic oxygen, but we hold no silicone data. The Fused silica (SiO2 glass) record says its atomic oxygen behaviour was not looked up.

Terms used

Atomic oxygen (AO)

single oxygen atoms, as opposed to the O2 molecules we breathe. They are abundant at the altitudes where the space station and many satellites orbit, and they are chemically very reactive.

Erosion yield

the volume of material removed for each oxygen atom that strikes it, written in cm3/atom. A smaller number means a more durable surface.

Fluence

the total number of atoms that struck each square centimetre over the whole exposure, written in atoms/cm2. The MISSE 2 samples in our data saw 8.43e21 atoms/cm2 where a record states it.

MISSE 2

a NASA experiment that exposed polymer samples on the outside of the International Space Station for 3.95 years. Most yields here come from it, so they share one exposure, though not every record repeats the fluence.

Lower bound

a value printed with a greater-than sign. The thin sample eroded all the way through, so the true yield is at least the number shown and could be much higher.

Our bar

the minimum we require before an article may cite a material profile: enough sourced properties, sources and uses, plus hazard, supply and failure-mode records. A profile below the bar may hold a value, but we do not rank it.

What to check before you order

1. Which faces point into the direction of travel, and what fluence does your orbit and mission length give? Do not reuse the MISSE 2 fluence without checking.
2. Ask the supplier for a flight or ASTM E2089 ground yield on your exact grade, thickness and colour, with Kapton H as the witness, the unit (cm3/atom or g/atom) and the density printed, and any greater-than sign explained.
3. Compare yield times fluence with the thickness you can afford to lose, with margin for thin edges, which erode faster and peel.
4. Ask what coats or covers the polymer, and whether a micrometeoroid hole or pinhole would expose the layer beneath.
5. For fluoropolymers, check the radiation limit (PFA 10 kGy, PTFE a few megarads) and stiffness; for bare metals, ask whether the oxide protects, grows or cracks off.

What do we not know yet?

• No erosion yield that meets our bar is held for silicones, glasses, oxide ceramics or bare metals. The titanium, aluminium alloy and fused silica records say their atomic oxygen behaviour was not found or not looked up.
• The epoxy, Kapton H, colorless polyimide, Upilex-S, PMR polyimide resin, Zylon, polyethylene oxide, pyrolytic graphite and clear Tedlar profiles hold yields but are below our bar. The Upilex-S record also carries an unresolved conflict between two NASA tables.
• The PBT record has an unresolved conflict between two NASA tables, a factor of ten apart. The article assumes the worse value until it is resolved.
• Orbit-wide fluence by altitude and year, impact-energy dependence, ageing under ultraviolet light and radiation, and silicone-contamination effects are not held; most materials have a single MISSE 2 exposure.

Questions people ask