O-ring failure

Materials World magazine
29 Aug 2019

Seals often fail because the material chosen wasn’t right for the object’s application. Find out how you can test a seal. 

The simple component, an o-ring, is mostly used for sealing against liquids and gases, yet its performance is often critical for complex and sometimes life-dependent machinery such as aircraft, medical devices and power stations. There are several causes of seal degradation, but these mechanisms can combine to cloud the overall picture. In addition, the properties of elastomers are time-dependent, so the condition of an o-ring when it has been taken out of service may be different to how it appears when it arrives into the lab for analysis. But it is possible to identify the causes of seal degradation and the leaks that follow.

Chemical attack

As elastomers are polymers, they may absorb fluids, causing them to swell. Certain liquids, including particular brands of hydraulic fluid, are notorious for causing some elastomers to swell, with the fluid manufacturer advising against use for certain common o-ring materials – fluorinated hydrocarbons, polyisoprene and polychloroprene are considered of poor compatability – with some elastomers in an Eastman Chemical Company compatibility chart, published in 2014, categorised as no.

In addition to swelling, some fluids will cause the polymer chains to break, softening the material. While some elastomers work well with certain fluids, for example, ethylene propylene diene monomer is advised for use with hydraulic fluid, there are no elastomers that are immune to chemical attack.

The issue with swelling and softening by fluids is that o-rings are designed to operate in a chamber, referred to as a gland. The gland is deliberately wider than it is deep, so that a pocket of gas or liquid pushes the o-ring into the low-pressure sidewall. The sealing force is greater with higher pressure, as the gas pressure will squeeze the o-ring onto the sealing face walls. In this way, o-rings can be used to seal against very high pressures.

If the o-ring swells, it can fill the gland, ensuring there is no pressure on the side to squeeze it against the wall, and so the o-ring starts to leak. As such, an o-ring that operated perfectly well on installation may start to leak after a short time in service. The softening of the o-ring might progress to the extent that the seal itself starts to extrude through the gap in the sealing surfaces. At this stage, the mechanical movement of the sealing surfaces can cut chunks out of the seal, in a process called nibbling. Ultimately, this process can tear the seal apart, bit by bit.

Finding chemical attack signs 

O-ring swelling can be determined by measuring ring dimensions, except you would need to know for certain how large it was to begin with, as sometimes the wrong o-rings are installed, or they are made out of tolerance. To further confuse matters, once the o-ring is taken out of the swelling fluid, it will start to dry out and contract again. Therefore, the o-ring measurement is best carried out using a high-precision optical measurement, as any mechanical measurement, such as vernier calipers, will squash it, making determining the dimensions difficult.

In addition to dimensional measurements, it is important to look for signs of nibbling, which can also provide evidence of swell. Binocular microscopes are effective for this purpose, as they give good depth perception at high magnification.

Rubber hardness is also a powerful analysis tool for determining fluid attack. When fluid is absorbed into elastomers, the polymer chains have more space to move around, so the modulus of the material decreases. This property can be measured using an indenter, such as Shore or International rubber hardness degrees (IRHD), which measure the force used to press a probe into the tested material.

A decrease in hardness from the original value is a strong indicator of chemical attack. A function of the size and shape of the test sample is hardness, so to carry out a realistic test of that, a piece from a good o-ring should be cut down to the same dimensions of an example from a failed section, with the hardness taken from the same place on both. A reading taken in this way would be lower than that from a complete o-ring or a test sheet, but the identically sized samples can be directly compared.

Hardness testing is described in detail in international standards, such as ASTM D2240, but in essence a needle-shaped indenter is pressed into the surface of the o-ring and the extent of the needle penetration into the surface determines the hardness. In traditional instruments, the indenter projects through a foot and is pressed into a sample until the foot contacts the sample. Some of the indenter penetrates the sample and the rest retracts back into the instrument, with the length of the indenter forced back into the instrument, providing the hardness reading. A very hard material would give a reading of 100 Shore hardness, as the whole of the indenter would be pushed back into the instrument, instead of penetrating the sample.

Tensile testing the o-ring can be used to determine if the elastomer has been attacked by chemicals, as the fluid can cause chain scission and thus reduce the strength of the elastomer. This would require a good cross section of the o-ring to be cut and measured to convert the failure load to failure stress. O-rings that have only swollen, but have not been attacked by the fluid, would not have their tensile failure load affected, although the tensile strength would appear to decrease because the cross section has increased.

A complete o-ring can be tensile tested by looping the ring around circular bars that are held in mechanical testing machine grips. The machine then pulls the o-ring apart until it snaps, with the load versus displacement typically measured by a load cell and displacement transducers in the testing frame. Modern testing machines plot the force versus displacement curve and can be set to automatically record the maximum tensile load on the sample. If the o-ring has already been snapped, then pneumatic or wedge grips can be used to clamp the sample while it is pulled. The sample can be observed using a camera during testing to look for cracks that indicate surface damage to the o-ring.

Heat ageing

This is where oxidation of the elastomer is accelerated through temperature. This oxidation increases the cross-linking density between the elastomer chains, increasing the hardness and reducing the elongation to failure. Heat ageing can accelerate compression set, where the ring’s permanently adopts the shape of the ring groove and when the sealing face gap increases, the ring may not elastically recover to prevent a leak from occurring.

Heat-affected o-rings may not appear different, unless they are affected to the point that cracks appear on the surface. These can be seen using high magnification binoculars when twisting the o-ring into a figure of eight. Typically, heat-damaged o-rings are identified by the fact that their hardness has increased past their starting value. The problem with diagnosing the issue in this was is that if the o-rings have been heat-damaged and chemically swollen, their hardness may have stayed at its original value, as the effects of additional cross-linking and plasticisation from swelling cause simultaneous hardening and softening.

Many o-rings are used in hot conditions to seal against aggressive fluids. Fresh lubricating oil can be benign to the rings, but as the oil ages it becomes acidic and can start to attack the seals. The change in hardness resulting from oxidation can be masked by the effect of fluid attack. A degraded o-ring can appear to be as good as new, but it might be well along a path to leaking.

A tensile test can be used to see if the o-ring has suffered from heat ageing, as the elongation will decrease. If this is observed, yet the hardness appears to be the same, it is likely that chemical and thermal attack have been at work. Other mechanisms can cause the hardness to increase, while decreasing the elongation, such as o-zone attack or UV light, but these tend to attack the surface first, so observing the tensile test with a microscope may assist in spotting the surface cracks that will open up at low elongation. Unlike a metallic sample, o-rings are not mounted for inspection and are best viewed using a binocular or trinocular microscope system, as this will provide depth perception when observing the sample. Variable lighting is important, as o-rings are often black, so seeing features with a single light source can be difficult.

Compression set assessment

In terms of the effects on the o-ring of heat ageing, compression set can cause leaks to appear long before the ring completely fails due to the decrease in elongation to failure. The performance of compression set o-rings can be determined by placing them between two platens at the same distance apart as the minimum sealing face distance, within a mechanical testing machine that can measure the load. Once the force on platens reaches an equilibrium value, the top surface should be lifted until the force on the upper platern drops away.

If the distance required to lose the sealing force is less than the maximum sealing face separation, then this indicates that the o-ring would leak. In order to make the test representative, it should be carried out at the expected service temperature of the ring, as thermal expansion of the o-ring would increase the sealing force for a given sealing face separation.