Accelerated Aging and Life Prediction

Predicting the service life of adhesives is a risky business. The most difficult question ever put to an adhesive consultant is, How long will the adhesive joint last in service The problem is that an adhesive joint is not made up of just one element. It contains several elements, and some of them interact. In fact, in most adhesive joints at least five elements must be considered substrate A, interface A, the adhesive, interface B, and substrate B. To understand and predict the rate of degradation of each of these elements is challenging, but it can be done. The most difficult failure situations to predict are those that result from interactive effects. 

it is important to consider and evaluate the adhesive joint as a complete system, but to do so we must generally separate these elements and look at their individual mechanisms of degradation. The following discussion considers mainly the bulk adhesive component of the joint. The process summarized below is the combination of work that has been described previously in several references.2,3 

Life prediction methodology embraces all aspects of the numerous processes that could affect the function of the element—in this case the bulk adhesive. The first step is to define the function of the adhesive clearly enough for a failure criterion to be derived. This failure criterion may be an unacceptable reduction in tensile strength, time to creep failure under a given stress, reduction in modulus due to moisture ingression, increase in modulus due to oxidation, unacceptable crack depth, or a variety of other possible criteria. It is also important that the criteria be related to practical adhesive joint performance. This is where it is difficult, and one must presume, at least for this limited analysis, that the adhesive will fail via a bulk (cohesive) property. 

After the failure criterion has been defined, the various processes that could cause this failure must be analyzed. For example, an increase in modulus could occur by thermal oxidation, increased postcure crosslinking, or the loss of plasticizer. Whatever the mechanism, each possible process needs to be identified and its rate characterized separately. Only then can interactions between different mechanisms be considered for life prediction. 

Finally, the rate of change in the critical property must be measured relative to expected environments of different severity and time intervals. If measurement cannot be made at the service temperature in the time that is available (as is generally the case), then accelerated tests may be used at elevated temperature or increased frequency. However, it is extremely important that care be taken to match the accelerated test conditions to the service conditions in as realistic a way as possible. For example, if accelerated aging by elevated temperatures is being used, the temperature must not be so high as to begin a degradation mechanism that would not normally be seen in service. 

---