As already mentioned, moisture is the bane of most adhesive bonds. It is nearly impossible to keep water from a bond exposed to the outside environment [14]. Water can readily diffuse through the adhesive or the adherend, if it is permeable, as a composite might be. Moisture can also wick or travel along the interface and it can migrate via capillary action through cracks and crazes in the adhesive. Once moisture is present, it can attack the bond by [14]
reversibly altering the adhesive, e. g., plasticization swelling the adhesive and inducing concomitant stresses disrupting secondary bonds across the adherend/adhesive interface irreversibly altering the adhesive, e. g., hydrolysis, cracking, or crazing hydrating or corroding the adherend surface.
The first three of these processes are reversible to one extent or another. Provided that bond degradation has not proceeded too far, if the joint is dried out (which may be a long process), the bond can regain some of its lost strength [10,15]. There appears to be a critical water concentration, below which either no weakening occurs [1,16], or whatever weakening that does occur is reversible [15,17]. This critical water concentration is dependent on the materials used in the joint and is likely to be dependent on the temperature and stress as well. At higher moisture levels, some strength may be recovered upon drying, but at a certain point, the failure becomes near catastrophic and is beyond recovery.
Upon moisture penetration, the locus of failure almost always switches from cohesive within the adhesive to at or near the interface. Because metal oxide surfaces are polar, they attract water molecules that can disrupt any dispersive (van der Waals) bonds across the interface. This disruption can be seen thermodynamically by the work of adhesion in an inert medium, WA, which can be represented as [1]
WA = ga + gs — gas (1)
where ga and gs are the surface free energies of the adhesive and substrate, respectively, and gas is the interfacial free energy. In the presence of a liquid such as water, the work of adhesion, WAl becomes
WAl = gal + gsl — gas (2)
where gal and gsl are now the interfacial free energies of the adhesive/liquid and substrate/ liquid interfaces, respectively. In an inert environment, the work of adhesion for a bonded system will be positive indicating a stable interface, whereas in the presence of water, the work of adhesion may become negative, indicating an unstable interface that may dissociate. Table 1 shows, in fact, that moisture will displace epoxy adhesives from iron (steel), aluminum, and silicon substrates and promote disbonding [1]. In contrast, although moisture weakens epoxy/carbon fiber bonds, these remain thermodynamically stable. Industrial experience with both metal and composite joints confirms these predictions [1].
The data presented in Table 1 illustrate the potential disastrous results when relying solely on dispersive bonds across the interface between an epoxy adhesive and metals or ceramics. To illustrate this danger, demonstration specimens can be produced that exhibit good initial strength, but fall apart under their own weight when a drop of water is placed at the crack tip.
Interface |
Work of Adhesion (mJ/m2) |
Interfacial Debonding After Immersion in Water |
|
In Inert Medium |
In Water |
||
Epoxy/ferric oxide (mild steel) |
291 |
-255 |
Yes |
Epoxy/alumina |
232 |
-137 |
Yes |
Epoxy/silica |
178 |
-57 |
Yes |
Epoxy/CFRP |
88-99 |
22-44 |
No |
CFRP, carbon fiber reinforced plastic. Source: Ref. 1. |