There are a number of theories on how adhesives work and there is little common agreement as to which theory is the most relevant for any particular bonding case.   It is actually quite important to know the mechanism of bonding because this has an impact on the surface preparation of the adherent surfaces and the materials being attached.

It is known that joints bonded with adhesives are generally stronger in compression, shear and tension than in peeling/tearing - it is much easier to break an adhesive joint by accessing an edge and peeling it away.  It is also apparent that it is relatively difficult to ensure that an adhesive joint is in pure tension and if the tension load is of centre or is not normal to the joint there is a tendency for peeling.  The best adhesive joints are designed for shear stresses with mechanical guidance and reinforcement e.g keys, corners, shoulders etc.

There are accepted conditions which result in higher adhesive bond strengths as listed below

There are a number of adhesive theories contributing to the overall study of bonding as listed below:

This is the simplest theory and is based on the factor that, at the microscopic level all surfaces are very rough consisting of crevices, cracks and pores.  The adhesive penetrates these features and hardens such that it keys into the surfaces and forms a strong surface bond ( this is probably similar to velcro ). The adhesive thus is able to bond two surfaces together and ideally the only weakest part of the bonded joint is the adhesive strength.

This theory is based on the assumption that the adhesive "wets" the surface of the adherent surface (meaning that the adhesive applied to the adherent spreads spontaneously when the join is formed )..   This theory has resulted in adhesive materials being developed which have a lower surface tension than the adherent surfaces.   Examples supporting this theory include epoxy resins which wet steel and result in a good bond - these resins do not wet PE or PTFE and result in a poor bond.

According to this theory, in the event of intimate contact between the adhesive and the adherent, the adhesive strength arises as a result of secondary intermolecular forces at the interface.   These may include Van der Waals forces (dipole-dipole, dipole-induced dipole interactions and hydrogen bonds ).

This is a variation on the adsorption theory in that stronger chemical bonds (ionic, covalent metallic ) form across the joint interface. ref Molecular Bonds .   In this regard, introduction of molecular bonding between the adhesive and the adherent will obviously improve the adhesive bond strength.   This can be attained by reactions at the surfaces, using proper surface treatments, or by using additional coupling agents.

This theory states that an electrostatically charged double bond develops at the bond interface as a result of the interaction of the adhesive and and adherent which contributes significantly to the bond strength.

This is a controversial theory as many have doubted the actual significance of the forces involved.  While this concept may be useful to explain some specific examples of adhesion, significant doubts have been cast regarding its overall value.   These include improved adhesion strengths with lowering of temperature for a large number of adhesive system (lower temperatures should result in poorer electrostatic forces).  Also it has been identified that virtually no changes in adhesion performance result with gross variations in the electronic character of the adhesives.

When an adhesive contains an adherent solvent the adhesive can diffuse into the adherent surface (substrate) with an interchange of molecules.   The theory is is only really applicable to polymers where a movement and entanglement of long molecules can occur.

This can be viewed as a molecular interlock enabled adhesion.   For plastics, the theory includes for effects of contact time, influence of time and temperature on bonding rate, and the influences of polymer molecular weight and polymer structure.  

While the diffusion theory applies well for cases of self-adhesion or auto-adhesion, it does not fit well in providing an explanation for polymer-polymer adhesion.   High molecular weight thermoplastic polymers often display very high melt viscosity and will not diffuse easily within the time scale of most bonding operations.

This theory results non-adhesion of surfaces due to the existence of regions of low cohesive strength in the interfacial region. For most metals there is a surface layer such as a scaly oxide layer.  For a successful bond this layer is ideally removed by surface treatments before a strong adhesive bond can be achieved.  Aluminium has a strong coherent oxide layer which is suitable for bonding.