Dealing with Differing Coefficients of Thermal Expansion

We are using epoxy adhesives to bond steel substrates that are exposed periodically to both 150°C and as low as 0°C. Our problem is in maintaining adhesion to the two steel surfaces at such divergent temperatures because of the response of materials to thermal conditions when going through heating and cooling cycles. Do you have any suggestions?

Differences in thermal expansion coefficients between substrates and adhesives can cause big problems in many applications. In extreme cases, I have seen epoxies used to bond glass to stainless steel, and the glass has cracked on heat-curing the adhesives.

Adhesives generally have thermal expansion coefficients some 4-10 times higher than metals. This can create large stresses in the joints on temperature cycling, and often these stresses will cause debonding from the surfaces.

The solution to the problem usually involves a complete joint redesign to eliminate the stress buildup (this sounds unlikely to happen in your situation), modifying the thermal expansion coefficient of the adhesive, or making the adhesive more flexible so the stresses can be dissipated. High levels (30-60%) of, for example, a silica filler that has a very low thermal expansion coefficient will bring your adhesive much closer to the expansion coefficient of the steel. Some epoxy adhesives are highly filled and have low thermal expansion coefficients.

The other approach is to make your adhesive much more flexible by incorporating elastomers. Maximizing adhesion to the surfaces is obviously important in sealing as well, and make sure you incorporate suitable adhesion promoters into the adhesive or use a primer on the metal before bonding.