Adhesive joints failure times

We must now consider how these two different and conflicting ideas about adhesive joint failure have come to be accepted. An engineer must be able to hold two contradictory ideas, such as Equations (15.1) and (15.2), in his mind at the same time. Unfortunately both ideas cannot be true simultaneously. 

Figure 8.3 Relationship between the applied adhesive fracture energy, Gic, and the joint failure time, tf [20].

FIGURE 27.2 T-peel sfrength values of sulfuric acid-treated styrene-butadiene rubber (SBR)/polyurethane adhesive joints as a function of the immersion time in sulfuric acid. A = adhesion failure R = cohesion failure in the rubber. (From Cepeda-Jimenez, C.M., Pastor-Bias, M.M., Ferrandiz-Gomez, T.P., and Martm-Martmez, J.M., J. Adhes., 73, 135, 2000.)... 

In essence, the durability of metal/adhesive joints is governed primarily by the combination of substrate, surface preparation, environmental exposure and choice of adhesive. As stated earlier, the choice of the two-part nitrile rubber modified epoxy system (Hughes Chem - PPG) was a fixed variable, meeting the requirement of initial joint strength and cure cycle and was not, at this time, examined as a reason for joint failure. Durability, as influenced by substrate, surface preparation, and environmental exposure were examined in this study using results obtained from accelerated exposure of single lap shear adhesive joints. 

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. 

A comparison of experimentally determined failure times for different stress levels and those predicted by the above equation for epoxy-aluminum lap shear joints aged at 60°C and 95 percent RH is presented in Table 15.3. These results indicate that the reaction rate method is satisfactory for predicting the effects of temperature and stress on the lifetime of adhesive bonds, provided that failure is cohesive within the adhesive. This, of course, should be validated by prototype testing. 

A diagram that one might use to illustrate a possible set of experimental data to represent all failure modes of an adhesive joint is presented in Fig. 15.1. When the data are closely analyzed and the extent of ultimate service life and proper safety margins are specified, the critical failure mode and time can be defined by identifying the weakest link — in this case the corrosion mechanism. If this predicted life is longer than the expected service life of the product, then the material specified for the adhesive joint can be qualified for use. 

Long-term durability of adhesively bonded joints may require resistance to a number of individual or combined degradation modes, including environmental attack, fatigue and time-dependent failures. Time-dependent failure mechanisms are often characterized nsing either a strength approach, involving creep and creep-rupture tests, or a fracture approach, in which debond rate is determined. In creep-rupture tests, adhesive joints are subjected to... 

In instrumented creep tests taken to failure, one learns not only how long specimens last but also how deformation increases throughout the creep process. For lap joints, delay times have been seen in creep tests, probably due to the increasing uniformity of the shear stress state, as predicted by the shear lag model as the creep compliance of the adhesive increases with time. In other situations, no such delay time is seen. A schematic illustration of a creep curve for an adhesive bond consisting of a butt joint bonded with a pressure sensitive foam tape is shown in Fig. 2, exhibiting classical primary, secondary and tertiary regions of creep behaviour. 

Structural adhesives such as epoxy resins can be treated as any rigid polymer and samples can be machined from cast sheets to produce test-pieces. These can then be used to measure typical tensile properties such as failure stress and strain. Using accurate exten-sometry, it is possible to characterize completely the uniaxial properties of an adhesive. The Creep of adhesive joints is especially important for structural adhesives maintained at high temperature. It is possible to determine the creep resistance of such materials by applying suitable loads at an appropriate temperature to samples of the adhesive, and to record the deformation with time. From such data, it will soon be evident if the adhesive is suitable for use or if it will cause a joint to deform with time. It is important to remember that humidity is likely to affect the properties of the adhesive, and in a long-term creep experiment, the humidity could cause premature failure. 

By measuring water uptake, the diffusion coefficient and equilibrium concentration of water for the bulk adhesive were obtained at different temperatures. A value of 37 kJ/mol was also calculated for the activation energy of diffusion. A value for the plane-strain stress intensity factor, Kic, for the bulk adhesive was obtained using compact tension specimens. Tensile butt joints were prepared from mild steel blocks bonded with the epoxy adhesive and the fracture stress determined as a function of time of exposure to water at the different temperatures. An activation energy of 32 kJ/mol was calculated for joint failure, in close agreement with that obtained for the diffusion of water. This supports the view that the processes responsible for loss of joint strength are controlled by water diffusion. It was found that joints exposed to 20°C/55% RH showed no reduction in strength, even... 

It is known that the humidity of the environment also effects the rate of loss of strength of adhesive joints and thus the time to failure, but this is not allowed for in the above treatment. The reaction rate approach has been modified by treating the terms C, a, and b in Eq. (11) as fitting parameters that account for the temperature and stress dependence of the fracture process. 

In considering the above theories, one has to admit that the most useful concept of adhesion stays within the boundaries of the moleciilar theory and the thermodynamics of interfacial phenomena. At the same time, no one theory of adhesion can predict the real adhesion between solid and polymer or adhesion j oint strength. A large number of theoretical ideas on adhesion do not refer to the phenomena of adhesion but rather to the processes of failure of adhesion joints and their description. A clear distinction between the processes of adhesion and the formation of adhesion contact and failme is a key to rmderstanding this complex set of phenomena, referred to for convenience as adhesion. 

The above comments are seen to be reinforced by observations on the failure path in joints before and after environmental attack. The locus of joint failure of adhesive joints when initially prepared is usually by cohesive fracture in the adhesive layer, or possibly in the substrate materials. However, a classic symptom of environmental attack is that, after such attack, the joints exhibit some degree of apparently interfacial failure between the substrate (or primer) and the substrate. The extent of such apparently interfacial attack increases with time of exposure to the hostile environment. In many instances environmental attack is not accompanied by gross corrosion and the substrates appear clean and in a pristine condition, whilst in other instances the substrates may be heavily corroded. However, as will be shown later, first appearances may be deceptive. For example, to determine whether the failure path is truly at the interface, or whether it is in the oxide layer, or in a boundary layer of the adhesive or primer (if present), requires the use of modern surface analytical methods one cannot rely simply upon a visual assessment. Also, the presence of corrosion on the failed surfaces does not necessarily imply that it was a key aspect in the mechanism of environmental attack. In many instances, corrosion only occurs once the intrinsic adhesion forces at the adhesive/substrate interface, or the oxide layer itself, have failed due the ingressing liquid the substrate surface is now exposed and a liquid electrolyte is present so that post-failure corrosion of the substrate may now result. 

In order to correlate the electrical resistance shift as a function of humidity test time, a theoretical model has been developed based on Ref 39. It takes into account both oxidation and cracking, two primary failure mechanisms of conductive adhesive joints, and can thus explain the experimental observations quite well. 

---