In 2022, researchers refueled a first-generation Toyota Mirai at a public hydrogen station in Europe, sampling the fuel at the nozzle as it flowed in. After driving less than 30 km, they drew a second sample from a low-pressure point near the car's fuel tanks. Most contaminants matched. Water didn't: 4.4 µmol/mol from the vehicle versus 1.27 µmol/mol at the nozzle, more than three times as much.
The team listed humidity already inside the tank as one possible explanation, alongside their own sampling setup, and called for more study. Those tanks are composite overwrapped pressure vessels, and the result is a good reminder that COPV cleaning is rarely as finished as it looks.
COPVs carry helium on launch vehicles, pressurant on satellites, breathing air for firefighters, and now hydrogen for cars and trucks. They offer a higher strength-to-weight ratio than all-metal tanks. They are also far less forgiving of contamination, chemistry, water, and careless handling.
COPV cleaning is the removal of particles, residue, and moisture from a composite overwrapped pressure vessel's interior, verified against a written cleanliness requirement, without harming the liner or the overwrap.
Here is where it tends to go wrong.
Table of Contents
Key Takeaways
- A COPV combines very different materials, and COPV cleaning has to protect the liner and the composite at the same time.
- Several of the most consequential failures are invisible: residue no one measured, water no one proved was gone, or damage that left no mark.
- "Oxygen clean" on a certificate means very little without numbers behind it.
- Hydrostatic testing can undo good cleaning if drying and verification don't come after it.
- Hydrogen fuel cell vehicles have put COPV cleanliness inside a fuel quality equation measured in parts per million.
A COPV Is Two Materials Sharing One Job
A steel cylinder is one material doing everything. A COPV splits the work. A liner keeps the gas in, and a composite overwrap, usually carbon fiber in epoxy, carries some or most of the pressure load depending on the design.
The FAA's breakdown of composite pressure vessel types shows how much that varies. Type II vessels have a load-bearing metal cylinder wrapped only around the straight sidewall. Type III vessels are fully wrapped over a metal liner, usually aluminum, that still shares the load. Type IV vessels, the kind in hydrogen cars, use a polymer liner and let the carbon fiber carry nearly all of it.
Type V vessels eliminate the conventional liner entirely. Even the type numbers aren't used consistently between NASA documents and the hydrogen industry, so the vessel's actual construction should drive the COPV cleaning plan, not the label on the drawing.
That variety is the first thing that makes COPV cleaning different. Aluminum corrodes when water sits against it. Polymer can swell in the wrong solvent and has tighter temperature limits than metal. And every design adds composite material whose handling and chemical limits must be respected. A process built for steel cylinders treats all of this as the same problem. It isn't.
Why COPV Cleaning Is a Different Problem
Contamination inside a pressure vessel doesn't stay inside the pressure vessel. It rides the gas downstream into valves, regulators, filters, and engines.
In oxygen service, the consequences are severe. NASA's White Sands Test Facility warns that hardware not properly cleaned for oxygen may be susceptible to ignition and burning. Oil, grease, fibers, and dust supply fuel. Particles add a second threat. Carried by fast-moving gas, they strike valve seats and elbows and turn speed into heat. NASA's oxygen compatibility guidance identifies flow velocities typically above about 100 ft/s as one condition that can support particle impact ignition when particles and an impact point are also present.
Even a helium or nitrogen COPV may carry stringent cleanliness requirements if its contamination can reach sensitive downstream hardware. That requirement should come from the system's cleanliness allocation and hazard analysis, not the name of the gas in the bottle.
Outside oxygen service, the damage is slower but just as real. Particles score valve seats, foul regulators, and clog thruster injectors. A propulsion system is only as clean as the vessel feeding it, which is why COPV cleaning requirements for flight hardware are often stricter than anything a commercial cylinder ever sees.
The 2016 AMOS-6 Falcon 9 loss, traced to oxygen trapped between a helium COPV's liner and overwrap, wasn't a dirty-tank failure. It did show that oxygen compatibility depends on the complete COPV configuration, including liner buckling, trapped oxygen, and fill conditions, not surface cleanliness alone.
Six Ways a "Clean" COPV Isn't Actually Clean
Most COPV cleaning problems don't announce themselves. They show up later, downstream, at the worst possible time.
1. The Solvent Cleaned Perfectly and Still Did Damage
In COPV cleaning, a solvent can strip every trace of oil from an aluminum liner and still swell a polymer liner, attack a seal, or weaken an adhesive at the boss. That is why ASTM G93 treats cleaning agent choice as a materials decision rather than handing out one approved chemical.
The ground is also shifting. 3M has exited PFAS manufacturing, ending the Novec fluids many processes were built on. NASA's own search to replace AK-225G for propulsion oxygen systems showed that a solvent's cleaning power and its oxygen compatibility are separate questions. (We dug into the options in our Novec 7100 replacement guide.)
The cleaner itself can become the contaminant, too. ASTM G120 warns that solvent left behind on oxygen hardware can raise ignition risk on its own.
2. The Surface Nobody Reached
Access is often limited to a relatively small boss or neck opening. No one is wiping a dome by hand, and a borescope only sees so much. Fluid has to reach the far dome, the shoulders, and the transition at the boss, and the process has to prove it did. A vessel that soaked for the right amount of time can still have spots the cleaning fluid barely touched.
Orientation matters, too. Cleaning a COPV vertically lets fluid drain neck-down instead of pooling against the far dome, and spray tools have to demonstrate coverage rather than assume it. An ultrasonic bath may suit a valve, but immersing a complete COPV is a decision for the vessel's design authority, not the COPV cleaning shop alone.
3. The Hydro Test Undid the Cleaning
Many composite cylinders undergo periodic pressure requalification under their DOT specification or special permit, often including a hydrostatic test. Hydro fills the vessel with water and exposes it to a test circuit that can carry its own particles and organics. Clean the vessel before hydro and ship it after, and you may be shipping the test rig's contamination.
For oxygen, high-purity, or hydrogen service, the work plan should include whatever post-test COPV cleaning, drying, and verification it takes to restore the specified cleanliness. Cleaning never replaces requalification, either. The smartest approach treats inspection, hydro, cleaning, drying, and verification as one plan, not separate purchase orders. Our guide to DOT hydrostatic testing covers the requalification side.
4. The Vessel Was Dried on a Timer
Water is the quiet problem, and drying is easy to underestimate in COPV cleaning. Residual water can corrode metal liners, contaminate high-purity gas, freeze in cryogenic service, or violate an oxygen system's moisture limit. EIGA's cleaning guidance says water must be drained and thoroughly dried before oxygen-enriched fluid goes in.
"Purged with nitrogen for two hours" isn't proof of anything. Dryness is a measurement, not a duration. The usual acceptance test monitors the dew point or moisture content of the gas leaving the vessel until it meets the specified limit, backed by validated purge coverage, vacuum drying, or engineering review where the geometry can trap liquid. Our hydrogen system drying guide explains why that number matters.
5. The Drop That Left No Mark
Composite damage can hide beneath a surface that looks untouched. A COPV cleaning job involves a lot of lifting, rotating, and fixturing, and every move is a chance for an impact. The COPV cleaning shops worth trusting support vessels only at approved hard points, and they quarantine, report, and send any impact for engineering review instead of wiping off the mark and moving on.
6. The Certificate Said "Oxygen Clean" and Nothing Else
This COPV cleaning gap is easy to miss. "Oxygen clean" and "clean per CC1246" sound definitive. Neither tells you what was measured.
A meaningful result names the standard and its revision, a particle level, a nonvolatile residue (NVR) limit with units, a dew point, and how the sample was taken. Some specifications even speak to COPVs directly: NASA notes that MSFC-SPEC-164 may be used to set internal cleanliness requirements for metal-lined COPVs. Revisions matter, too. MIL-STD-1246C was replaced by IEST-STD-CC1246, yet older contracts still invoke it. If the purchase order doesn't name the exact standard, revision, level, and units, the certificate can't answer the question that matters.
A cleanroom's ISO class describes the air in the room, not the inside of your vessel. Our post on filter cleaning validation covers the measurement side.
The Hydrogen Highway Raises the Bar
Back to that Mirai. Its tanks are Type IV COPVs, and the PEM fuel cell they feed is sensitive to trace contaminants. The study measured its samples against the 2019 edition of the ISO 14687 hydrogen fuel quality standard, which capped water at just 5 µmol/mol. It was revised in 2025, so current projects should check the newer edition.
The authors of the NPL-led study, a single-vehicle proof of concept, also point out that there is no published information on how contaminants behave inside Type IV tanks at all. The same paper cites a European survey in which 29% of station samples exceeded at least one fuel quality limit.
The takeaway isn't that tanks are dirty. It's that tank cleanliness and dryness now sit inside a fuel quality equation measured in parts per million, and nobody has much data on them. With a polymer liner, no one should assume metal-cylinder bake temperatures. Heated drying has to stay within limits set by the manufacturer or design authority, so COPV cleaning for Type IV tanks leans on controlled drying and real dew point verification. Our posts on hydrogen NVR testing and hydrogen particulate contamination cover the rest of that picture.
Your Outsourced Partner for Cleaning and Verifying COPVs
Every failure above has the same root: a gap between steps. One vendor cleans, another runs hydro, a third dries, and the paperwork says "clean" without saying how clean. Each handoff is another chance for contamination or damage.
Precision Fabricating & Cleaning closes those gaps. PFC has worked on Florida's Space Coast since 1964, supporting NASA, SpaceX, United Launch Alliance, Lockheed Martin, Boeing, and the U.S. Air Force, including work on critical projects like NASA's Crew Access Arm and Tail Service Mast Umbilicals. In 2022, ULA named PFC its Most Innovative Supplier of the Year. PFC performs COPV cleaning today, with the full process under one roof:
- ISO 9001 / AS9100 registered facilities with ISO 14644 compliant cleanrooms.
- Vertical cleaning capability for compressed gas cylinders, so vessels drain neck-down.
- Oxygen cleaning and precision cleaning to customer-specified limits under specifications including ASTM G93, CGA G-4.1, IEST-STD-CC1246, MIL-STD-1246, MSFC-SPEC-164, and KSC-C-123, with high-purity rinse water to JSC-SPEC-C-20C where contracts require it.
- In-house hydrostatic, pneumatic, and cryogenic testing, so testing and final cleaning happen in the right order, in one place.
- Filtered GN₂ purge and dew point verification that proves dryness instead of assuming it.
- Validated sampling and laboratory analysis, including NVR and particle counts, documented in numbers your quality team can use.
If your COPVs need to be cleaned, dried, and verified to a specific requirement, talk to our oxygen service team.
Frequently Asked Questions
What does "oxygen clean" mean for a COPV?
"Oxygen clean" means the vessel has been cleaned to remove contaminants that could ignite in oxygen service. On its own, the phrase isn't a specification. A meaningful requirement pairs a process guide such as ASTM G93 with measured acceptance limits, such as particle and NVR levels under IEST-STD-CC1246, plus a dew point.
How do you know a COPV is dry after cleaning?
Measure it. The vessel is purged with filtered, dry gas, usually nitrogen, while the dew point or moisture content of the gas leaving it is monitored. It counts as dry when that reading meets the specified limit and holds steady. A purge time alone doesn't prove dryness.
When does a COPV need to be recleaned?
Whenever its cleanliness can no longer be shown. Common triggers include a hydrostatic test, loss of clean closures, expired or damaged packaging, and exposure to an uncontrolled environment. An impact calls for quarantine and structural inspection first; any recleaning follows from that evaluation.
Are hydrogen fuel cell vehicle tanks COPVs?
Yes. Production hydrogen fuel cell cars such as the Toyota Mirai store fuel in Type IV COPVs with a polymer liner and carbon fiber overwrap, typically at 700 bar. Because fuel cells are sensitive to moisture and contaminants, COPV cleaning and drying for these tanks has to respect the polymer liner's temperature limits.
The Bottom Line
A COPV earns its place with a higher strength-to-weight ratio than an all-metal tank. It pays for that with less tolerance for the wrong chemistry, leftover water, rough handling, and vague paperwork. Good COPV cleaning respects both the liner and the composite, proves the vessel is dry, and puts the result in numbers. Whether the vessel is headed to a launch pad or a hydrogen station, that's the difference between clean and merely looking clean.