A compatibility chart answers a narrower question than the one people ask it. It says: for this compound, in this fluid, at this temperature, for this duration, the change in volume and hardness fell into a band that the author considered acceptable. Change any of those four and the rating may not apply.
Reading the chart well means knowing which of the four you are changing.

What the ratings actually mean
Most charts use a letter or number scale. The words behind the scale are remarkably consistent even when the symbols are not:
| Rating | Typical meaning | What it usually hides |
|---|---|---|
| A / 1 / Excellent | Volume swell under about 10 %, hardness change small | Tested at moderate temperature, often 23 °C |
| B / 2 / Good | Swell 10–20 %, some hardness loss | May be acceptable only for static seals |
| C / 3 / Fair | Swell 20–40 %, noticeable softening | Short-term or intermittent exposure only |
| D / 4 / Not recommended | Swell above 40 %, or degradation | Sometimes still used where nothing else works |
The word "excellent" is doing a lot of work. A compound rated excellent at 23 °C may swell 30 % at 80 °C in the same fluid, because swelling is strongly temperature-dependent. If your service temperature is not the chart's test temperature, the rating is an extrapolation.
The four variables the chart hides
Temperature
Swell increases with temperature, sometimes sharply. A nitrile ring in a solvent at 23 °C might take up 8 % and at 70 °C take up 25 %. The activation is not linear, so you cannot interpolate between two ratings — you need data at, or above, your service temperature.
Concentration
"Acid" covers pH 1 and pH 6. "Caustic" covers 5 % and 50 %. Charts occasionally state concentration, usually they do not. A compound that survives dilute sulphuric acid may be destroyed by the concentrated grade.
Duration
A seal in a static joint sees the fluid continuously for years. A seal in a valve that is actuated twice a day sees it intermittently. The same rating covers both, and only one of them is a long-term exposure. For continuous service, prefer ratings that are excellent at your temperature rather than good.
Static versus dynamic
A swollen ring in a static gland has nowhere to go: it fills the groove, increases contact stress, and can hydraulically lock the joint. A swollen ring in a dynamic gland is being rubbed against a surface while soft, which accelerates wear. The acceptance criteria differ, which is why some charts carry two columns.
Swell limits worth designing to
Volume swell is the most quoted number and the least useful on its own. Three measures together tell the story:
- Volume swell. Under 5 % is negligible. 5–15 % is usually workable in a static gland, provided the groove has fill margin. Above 15 % the ring may not fit the groove as designed. Above 25 % the mechanical properties are usually degraded as well.
- Hardness change. More than about 10 points either way is significant. Softening means the ring extrudes more easily; hardening means it conforms less and may crack.
- Tensile and elongation change. A loss of more than 30 % in elongation is a warning even if the swell looks acceptable, because the ring has lost the ability to stretch over a gland edge without tearing.
A ring can pass on swell and fail on elongation. That is why a one-number chart is a starting point, not an answer.
for a static seal, target volume swell under 15 % at the maximum service temperature; for a dynamic seal, under 10 %. If the only available compound gives more, change the groove fill ratio before changing the compound.
Why two suppliers disagree about the same fluid
Three legitimate reasons, and one not-so-legitimate:
- Different test conditions. ASTM D471 allows a range of temperatures and durations. A 70 h test at 23 °C and a 70 h test at 100 °C will not agree.
- Different compounds under the same name. "FKM" covers a family with different fluorine content, cure systems and fillers. A general-purpose FKM and a low-temperature FKM behave differently in the same solvent.
- Different acceptance bands. One supplier calls 20 % swell "good", another calls it "fair". The underlying data may be identical.
- Marketing. A chart that rates everything "excellent" is not a chart, it is a brochure.
When two charts disagree, do not average them. Run the test.
Running a 70-hour immersion test
The test that settles most arguments is a straightforward ASTM D471 immersion, followed by a compression set measurement on the same specimen:
- Measure the specimen: volume (by weight in air and water), hardness, and if possible tensile properties.
- Immerse in the actual fluid — the real product, not a substitute — at the maximum service temperature. 70 hours is the usual screening duration; 168 hours is better for a continuous-service joint.
- Remove, cool, wipe, and measure within the standard recovery window.
- Report volume change, hardness change, and any surface change: gloss loss, tackiness, cracking, blistering, extraction of plasticiser.
- Where the joint is critical, run a compression set on an immersed specimen. This catches the compounds that survive immersion but lose their elastic recovery afterwards, which is the failure mode that produces a leak with an intact-looking ring.
Seventy hours is not a life prediction. It is a screen that eliminates the obviously wrong compounds and ranks the plausible ones.
Field signals of a chemical problem
On a failed part, chemical attack leaves a distinct signature: the ring is soft or tacky, may be swollen well beyond the groove, often smells of the fluid, and may leave a black residue on the fingers. Hardening and fine surface cracks point at thermal ageing instead, and a ring that crumbles points at a compound that has been extracted or attacked at the surface.
The distinction matters because the remedies differ. Chemical attack means change compound. Thermal ageing means change compound or reduce temperature. Extraction means the plasticiser in the compound is soluble in the fluid, and the fix is a compound without that plasticiser — often a different cure system rather than a different polymer.
Where charts are simply not enough
Four services need real testing rather than a chart lookup:
- Oxygen and oxidisers. A hydrocarbon-based compound in oxygen service is a fire risk. Only dedicated oxygen-compatible compounds, cleaned to oxygen service standards, are acceptable.
- Drinking water. The requirement is not chemical resistance but leachables, and the standard is a certification (NSF/ANSI 61 or equivalent), not a swell number.
- Food contact. Same logic: the constraint is migration, tested to the food-contact regulation, not swell in a simulant.
- Mixed solvents and formulated products. Charts are built on single chemicals. A paint thinner is a mixture, and mixtures do not behave like their components averaged.
Frequently asked questions
Is a compatibility chart enough to select a compound?
For a low-pressure water or air joint, usually yes. For a continuous-service joint in a formulated product at temperature, no — run a 70-hour immersion at service temperature, then a compression set on the immersed specimen.
What volume swell is acceptable for an O-ring?
Under 15 % for a static seal at maximum temperature, under 10 % for a dynamic seal. Above that, check the groove fill ratio and the hardness change before deciding.
Why did an "excellent" rated ring swell in service?
Most likely the service temperature is above the chart's test temperature, the fluid is a formulated product rather than the pure chemical, or the compound is a different grade from the one tested.
Does a swollen ring still seal?
It may seal better initially, because higher contact stress, and then fail when the groove fills and the ring becomes incompressible, or when the softened compound extrudes. Swelling that fills the available groove volume is a design failure even when the seal still holds.
Can I test with a substitute fluid if I cannot get the real one?
Only for screening, and say so in the report. A substitute with a similar chemical family can rank compounds, but it cannot tell you the absolute swell, and it is exactly the difference between a similar fluid and the real one that causes field failures.