How to Tell Which Seal Material You Are Holding

Why this matters

You have a joint open, a seal in your hand, and a box of rings on the truck that are all black and all the right size. Put the wrong compound in and the joint holds for the pressure check, holds for the drive home, and weeps a month later. The customer gets a second visit that you eat. This is the procedure for narrowing an unknown seal down to a compound with what you can actually do in the field, and for knowing when the honest answer is that you cannot.

The five sources below are ordered by how much certainty each one can deliver. Only the first can name a compound on its own. Everything after it eliminates candidates, and eliminating is worth doing precisely because the first source is so often unavailable.

Before you touch it: two hazards

Never burn-test an unknown seal. The old shop trick of lighting a scrap and reading the smoke is a serious exposure risk, because fluoroelastomers and PTFE release hydrogen fluoride and other toxic decomposition products when overheated, and you do not know which compound you are holding, which is the entire reason you are testing it. There is no small safe version of this test in an unventilated mechanical room.

Get the joint to zero energy before the seal comes out. Isolate the line, relieve pressure to zero and confirm it at a gauge or a fully open vent rather than by feel, drain it, and let hot systems cool before you loosen anything. Where somebody else could reopen the isolating valve while you are in the joint, lock or tag it and verify the zero-energy state first, per 29 CFR 1910.147.

If the fluid in the system is anything other than water, put on gloves selected against that fluid using Section 8 of the safety data sheet the supplier owes you under 29 CFR 1910.1200, not whatever box of disposables is on the truck. Glove material is its own compatibility problem, and the common thin nitrile glove is permeated by ketone solvents in minutes.

Source 1: the documentation

This is the only source that can name a compound, and skipping it is the only skip that costs you the answer rather than just narrowing it. Everything else in this procedure exists because this one is missing.

Look in this order: the equipment's parts list or service manual, the seal kit's own packaging, the maintenance record from the last time this joint was opened, and the supplier's catalog entry for the part number stamped on the old kit box. On assemblies with a rebuilt history, the record of the last rebuild beats the original parts list, because whoever did that rebuild may have substituted.

What you lose by skipping it: certainty. Nothing later in the list restores it. A field identification is an elimination, and an elimination that narrows five candidates to two still leaves you choosing.

Source 2: the service conditions

If you cannot name the compound, name what it has to survive. This eliminates families quickly and it uses information you already have.

Write down four things: the fluid, its temperature in service, whether the seal is static or moves, and everything else that touches it including the assembly lubricant and any cleaner used in the groove.

Then eliminate. Petroleum oil in the system rules out EPDM, butyl, and most silicones. Steam or hot water rules out standard nitrile and polyurethane. Brake fluid or a ketone rules out fluoroelastomer, which is the counterintuitive one because most techs treat that family as the upgrade for everything. Continuous outdoor exposure rules out nitrile on ozone and UV grounds.

What you lose by skipping it: the ability to reject a wrong answer that the physical tests hand you. Physical tests give you a family, and this step is what tells you whether that family could have been in this joint at all.

Source 3: markings and color

Color is a supplier convention, not a specification, and it is right often enough to be a trap. Brown and green rings are frequently fluoroelastomer, blue often marks a food-contact or metal-detectable compound, and black is used by every family, which means black tells you nothing at all.

Use color only as a tiebreaker between two candidates that survived source 2, and never to override an elimination. A green ring in a system full of brake fluid is not a fluoroelastomer, it is a green ring, and somebody has already made the mistake you are about to repeat.

Where a molded part carries a compound marking or a cure-date code on a flange or backing, that is real evidence and belongs with source 1 rather than here.

What you lose by skipping it: nothing you cannot get elsewhere. This step is fast and occasionally decisive, which is why it is on the list, but an identification that rests on color alone is a guess wearing a lab coat.

Source 4: comparative density

This is the most useful physical test in the field, and it works because two of the families are dramatically heavier than the rest.

Approximate specific gravity bands for compounded seal materials: EPDM and butyl around 1.1 to 1.3, nitrile around 1.25 to 1.4, silicone around 1.1 to 1.5, neoprene around 1.35 to 1.55, fluoroelastomer around 1.8 to 2.0, PTFE around 2.15 to 2.2.

Notice the shape of that list. The ordinary compounds all sit between about 1.1 and 1.55 and overlap each other heavily, so density will not tell nitrile from neoprene from EPDM. But fluoroelastomer and PTFE sit far above that whole cluster, so density answers one question cleanly: is this a fluoroelastomer or is it one of the ordinary compounds? That is frequently the exact question you need answered, because fluoroelastomer is the expensive upgrade and the one most likely to have been substituted downward by a previous tech.

A float test in water does not work; everything on that list sinks. What works is a comparative weigh. Take a known ring of the same nominal size from stock, weigh both on a gram scale, and compare. Same geometry means the mass ratio is the density ratio.

What you lose by skipping it: the fluoroelastomer question, which source 2 often cannot settle because fluoroelastomer is compatible with so much that it survives most eliminations.

Source 5: feel, stretch, and tear

Handling characteristics narrow the ordinary compounds a little further. All of these are soft evidence and none of them override an elimination from source 2.

• Silicone tears easily, has a dry matte surface with a slight drag, and has poor rebound. If a ring tears when you stretch it moderately, silicone is the first candidate.
• EPDM feels dry and slightly chalky, stretches well, and snaps back cleanly.
• Nitrile feels smoother and slightly glossy, and is noticeably tougher to tear than silicone.
• Fluoroelastomer feels dense and slick, does not stretch as far, and has visibly low rebound. Held next to a nitrile ring of the same size, the difference is obvious once you have felt it a few times.

Durometer does not help. Almost every general-purpose o-ring is around 70 Shore A regardless of compound, so a hardness reading confirms the ring is a normal o-ring and nothing else.

What you lose by skipping it: a small amount of confidence between two candidates. This is the least conclusive source on the list, and it is last for that reason.

Worked identification: a black ring from a hot oil joint

A black ring comes out of a joint on an older piece of equipment. No parts list, no rebuild record, no marking on the ring.

Source 1 is unavailable. The identification will therefore be an elimination, and that gets written in the job record as such rather than as a compound name.

Source 2. The fluid is a petroleum-based oil, running at about 180 degrees F, and the seal is static in a flanged groove. Petroleum oil eliminates EPDM, butyl, and silicone outright. That leaves nitrile, hydrogenated nitrile, neoprene, and fluoroelastomer. The 180 degrees F service does not eliminate any of the four, since general-purpose nitrile is typically good to about 250 degrees F, though it is close enough to matter for life expectancy rather than for identification.

Source 3. The ring is black, which eliminates nothing. Noted and moved past.

Source 4. A known fluoroelastomer ring of the same nominal size is on the truck. Say it weighs 2.6 grams on a gram scale, and the unknown weighs 1.8 grams. Same geometry, so the mass ratio is the density ratio: 1.8 divided by 2.6 is about 0.69. Compare that against the band midpoints, nitrile near 1.3 against fluoroelastomer near 1.9, which gives 1.3 divided by 1.9, or about 0.68. The unknown lands squarely in the ordinary-compound band and nowhere near the fluoroelastomer band. Fluoroelastomer is eliminated.

Source 5. The ring stretches well, does not tear, and has good rebound. Consistent with all three survivors, and no help separating them.

The finding: nitrile, hydrogenated nitrile, or neoprene, most likely nitrile given the petroleum service at a temperature ordinary nitrile handles and that nitrile is the default compound for it. The correct action is to fit a nitrile ring rated for the temperature, record the identification as an elimination with the evidence behind it, and note the fluoroelastomer question as answered so the next tech does not repeat the weigh.

What would change this: if the weigh had come back near 2.6 grams, the ring would be a fluoroelastomer and the joint would need a fluoroelastomer replacement, because a downgrade to nitrile in a system somebody specified for the higher family will fail on whatever the original designer was worried about. And if the equipment turns out to have a rebuild record after all, that record overrides this entire chain, including the weigh.

How to verify you got this right

• The elimination logic runs one way only. Every candidate you removed has a named reason from source 2, and no candidate was removed on color or feel alone.
• The compound you fitted survives every fluid on your list from source 2, including the assembly lubricant, not just the process fluid.
• The record says what you actually know. "Nitrile family or neoprene, fluoroelastomer eliminated by comparative weigh, no documentation available" is a useful entry. "Nitrile" written as a fact you inferred is how the next person inherits your guess as data.
• If the joint is critical, one seal from the box goes into a labeled bag with the equipment record. When the identification question comes back in three years, a physical sample beats any note.

References

• Seal and o-ring manufacturer catalog data for compound specific gravity, temperature range, and color conventions
• Safety data sheet, Section 8 (exposure controls and personal protection), for glove selection against the specific fluid in the joint, per 29 CFR 1910.1200
• 29 CFR 1910.147 for lockout/tagout and verification of zero energy before opening a pressurized joint
• See related: The Elastomers and What Attacks Each One; Gasket + O-Ring Material Reference