The first and most basic point of understanding is the endpoint. Does the material continue to maintain contact pressure throughout the test, or does it fall to zero (below the detectable limit of the load cell) before the end of the test? While there is no definitive correlation from residual load force to the onset of leakage, it should be intuitive that a material that completely relaxes and loses all contact force is likely to leak in the application. Anecdotally, multiple customers have reported that the load force must drop to very close to zero for leakage to occur in their particular test apparatus. While this is good guidance, these anecdotal reports should not be taken as a definitive answer that applies in all circumstances.
Specifications are often written such that a minimum of 10% of the initial contact load force must remain for a passing result. In practice, there is nothing special about 10%. This is a semi-arbitrary value that ensures a material continues to apply some non-zero load force to the mating surfaces, with some safety factor to ensure that it does so even after all normal test variations are considered. In practice, this appears to be a conservative limit, there is nothing magical about the 10% number.
The loss of compressive load force can be broken down into three different types of phenomena, each with its own time frame. All rubber materials relax viscoelastically when initially compressed, and this loss stabilizes within the first 24 hours. That initial drop seldom has much direct impact on real-world applications. However, in the specific case of an assembly having neither a compression limiter nor solid-to-solid contact, meaning the assembly torque of the fasteners is controlled solely by compression of the seal, this will be observed as “torque fade” if the fastener torque is rechecked a day or two after assembly. In such a case, Parker recommends against retorquing the fasteners unless leakage is observed as this retorquing can easily result in damage to the seal from excessive compression.
The second set of phenomena to resolve are solvation effects that occur when an elastomer is immersed in a liquid test media. The rubber material will absorb some amount of test fluid, causing some swelling of the rubber and a small increase in load force. When this happens, the test fluid may extract liquid constituents from the rubber, resulting in shrinkage of the seal material and a loss of load force. These processes occur simultaneously, and both typically reach equilibrium within the first 72 hours. However, the net impact on the measured load force will only be noticeable if the volume change is significant.
The third set of phenomena are degradative effects caused by high temperatures and chemical reactions. These reactions are ongoing and cumulative. If the test temperature remains constant, the rate of degradation will remain constant, as well. Unfortunately, there is no way to distinguish what percent of any degradation is due to thermal effects versus chemical effects from a CSR curve. Comparing a CSR output curve to one generated from testing in an inert control fluid at the same time and temperature may allow a user to isolate thermal effects from chemical ones, but it must be known that the control fluid does not also produce chemical degradation of the rubber material.
Ultimately, it is more important to consider the slope of the curve after the initial drop than the initial drop itself. The material should stabilize to a relatively flat line, and the slope of that line reflects how the material responds to thermal and chemical aging effects. A curve with very little slope (Figure 1) is extremely stable long-term, whereas as a CSR curve that shows a steeper negative slope (Figure 2) means the material is continuing to degrade due to chemical and/or thermal effects in that fluid and at that temperature. This does not mean the material is incompatible with that environment, but the continuing losses mean the end of service life point (onset of leakage) would be expected relatively soon after the end of the test.
Figure 1: A fluorocarbon in engine oil at 150°C shows very little change after the initial relaxation response.
Figure 2: An HNBR in engine oil at 150° shows ongoing degradation after the initial relaxation response.
In summary, much knowledge about how a material will respond in an application can be gleaned from Compressive Stress Relaxation testing if one knows what to look for. Watch for part 3 of this series, where we will focus on how to use the understanding gained from CSR testing and how to incorporate it into a material specification.
For more information or assistance with your application, contact our applications team at firstname.lastname@example.org or chat online by visiting Parker O-Ring & Engineered Seals Division website.