Rubber Seal Performance in Cold Conditions: What Buyers Should Know About Ageing

A rubber seal can look intact and still struggle to maintain contact when temperature falls or its recovery changes with age. For automotive and industrial buyers, selecting a material requires more than checking a temperature range on a datasheet.

This article explains how low temperature, compression and ageing affect elastomeric seals, and which information to request before approving a component. It draws on published research involving elastomer O-rings. The findings are not LTHJT test results and should not be used as direct lifetime predictions for automotive hoses or weatherstrips.

Why rubber seals become less responsive in cold conditions

As temperature falls through its glass-transition region, an elastomer becomes stiffer and recovers more slowly. A seal that must respond to movement or a change in compression may therefore behave differently from one held in a stable static assembly.

Glass-transition temperature, often called Tg, is useful for comparing materials. However, it is not a universal seal-failure temperature. Differential scanning calorimetry (DSC) and dynamic mechanical analysis (DMA) characterize different aspects of the transition. Reported values also depend on the test conditions and the chosen definition.

Buyer takeaway: Ask how a quoted low-temperature limit was established. A material measurement, a recovery test and a leakage test provide different information.

Compression set: useful evidence, with limits

Compression set describes how much deformation remains after a compressed specimen is released. A lower value generally indicates better recovery under the stated conditions.

The usual expression is:

CS = [(h0 − h(t)) / (h0 − h1)] × 100%

Here, h0 is the initial height, h1 is the compressed height, and h(t) is the recovered height at the specified measurement time.

The measurement time matters because rubber recovery is time-dependent. Results obtained after different recovery periods should not be compared as though they were equivalent. Likewise, research methods adapted for O-rings or DMA equipment should not be labelled standard tests without checking their procedures.

Compression set is a screening indicator. It does not establish whether a particular assembly will leak. Seal geometry, compression, mating surfaces, pressure and movement also affect performance.

Ageing changes the seal as well as the material

Heat, oxygen, the surrounding medium and mechanical loading can change an elastomer over time. Chain scission and additional crosslinking may occur together, changing stiffness, recovery and the force maintained by a compressed seal.

An unloaded test specimen does not fully represent an installed seal. Ageing tests that include the intended compression help investigate this difference.

For O-rings, compression stress relaxation is another useful measurement: it tracks the decline in sealing force over time. Component leakage tests then help relate material changes to the assembly's actual function.

For door weatherstrips, the functional criteria may instead include water ingress, wind noise, closing effort and recovery after prolonged compression. These require their own application-specific tests.

Why a hotter ageing test does not automatically predict service life

Accelerated ageing is valuable, but converting oven exposure into years of service requires a validated model.

Two issues deserve particular attention:

Changes in ageing mechanism. The processes active at high test temperatures may differ from those governing normal service. A single Arrhenius extrapolation can then give a misleading lifetime estimate.

Diffusion-limited oxidation. Oxygen can be consumed near a specimen's surface faster than it reaches the interior. The resulting material ages unevenly. An overall property measurement may hide the difference between a strongly aged surface and a less aged core.

These effects depend on the compound, specimen dimensions and test conditions. Their presence cannot be inferred from the polymer name alone. Nor does an extrapolation error always have the same direction: the reliability and uncertainty of the prediction must be assessed.

Choosing between EPDM, NBR, HNBR and FKM

Treat these names as material families rather than complete specifications. Different grades and formulations within a family can behave differently.

Avoid selecting a compound solely because one material family performed well in a particular published experiment.

A practical qualification checklist

Before requesting a quotation or approving samples, provide:

• A drawing, sample or confirmed part reference.
• Minimum temperature, normal temperature and peak exposure.
• The contacting medium and its concentration.
• Pressure, compression, movement and installation conditions.
• Required tests and measurable acceptance criteria.
• Expected quantity and maintenance or replacement requirements.

Test the finished component where practical. For a fluid seal, leakage may be the decisive criterion; for a weatherstrip, water and air sealing may be more relevant. A chosen compression-set threshold should have an application-specific justification rather than being presented as a universal failure limit.

Conclusion

Reliable seal selection combines grade-specific material evidence with testing that represents the installed component. Low-temperature recovery, sustained compression and ageing all deserve attention, while service-life claims require more support than a single accelerated test.

Contact LTHJT with your drawing or sample, operating conditions, required test criteria and quantity for an application review and quotation. Product-specific performance should be confirmed through an agreed qualification process.

Research acknowledgement and further reading

This guide is independently organized from the customer-provided translation attributed to Chapter 30, Investigations of Elastomeric Seals—Low-Temperature Performance and Ageing Behaviour. It does not reproduce the chapter's figures or claim its measurements as our own. Complete edition and publication details should be added if available; author spelling should be checked against the original chapter.