Bonding Silicone: Primers, Addition Cure and the Joint That Holds

Why silicone is hard to bond

Silicone has a surface energy around 20 to 24 mN/m, among the lowest of any engineering material. Adhesives need to wet a surface to bond to it, and a liquid wets a low-energy surface poorly. That is the whole problem in one sentence.

The second problem is the release agent. Silicone parts are often moulded with an internal or external release agent, and that agent is designed to prevent exactly the adhesion you are now trying to create. A part that releases beautifully from its tool will not bond to anything until the agent is removed.

The third problem is migration. Silicone compounds contain low molecular weight fractions that migrate to the surface over time. A joint that was good on day one can fail at month six because a layer of migrated oil has formed under the adhesive.

All three problems are surface problems, and all three are solved before the adhesive is opened. That is why a primer does most of the work and why changing adhesive without changing surface preparation changes nothing.

What a primer actually does

A silicone primer is not a glue. It is a dilute solution of reactive silanes and a catalyst that, once the solvent flashes off, leaves a thin film on the substrate with two useful properties: it bonds to the metal through silanol and oxide groups, and it presents a surface that the silicone can cross-link into.

That is why the primer goes on the substrate, not on the silicone. The silicone has nothing to bond to on its own surface; the primer creates the interface. Painting primer on the silicone and pressing it to bare metal gives a joint that fails at the metal.

Film thickness matters and thin is better. A primer film of a few micrometres works; a film thick enough to see as a white haze is too thick and becomes the weak layer. Wipe on, let the solvent flash, and if there is visible residue, wipe it back.

Cure time before assembly matters too. Most primers need 15 to 30 minutes at room temperature, or a few minutes at 80 C, before the adhesive goes on. Assembling immediately traps solvent under the adhesive, and that solvent becomes a void.

What the numbers look like

Peel strength to aluminium, measured on a 1 mm silicone sheet bonded with an addition-cure adhesive: as moulded, 0.18 N/mm. An IPA wipe gets to 0.42. Corona treatment 0.95. Plasma for 60 seconds 1.62. Primer alone 2.75. Primer plus plasma 3.05.

The interesting number is not the top one. It is the gap between the wipe and the primer: 0.42 against 2.75 is a factor of six and a half, from a step that costs a few seconds per part. Most failed silicone bonds are joints where somebody wiped the surface carefully and skipped the primer.

Plasma alone at 1.62 is a real improvement over wiping, and for a joint that only needs to hold a gasket in place during assembly it is enough. For a structural or a sealing joint, it is not, and primer on top of plasma at 3.05 is the best of the routes tested.

Note what the primer plus plasma figure tells you: the two treatments are not additive in a simple way. Plasma cleans and oxidises the surface, primer provides the chemical bridge, and together they give a joint that fails cohesively in the silicone rather than at the interface. That is the failure mode you want, because it means the bond is stronger than the material.

The joint that fails at month six

Heat ageing is where the difference between primed and unprimed becomes dramatic. At 150 C, a primed joint retains 96 percent of its peel strength at 250 hours, 92 at 500, 87 at 1,000 and 81 at 2,000. An unprimed joint is at 72, 51, 30 and 14 over the same intervals.

An unprimed joint that passed a peel test on day one is at half strength by 500 hours and effectively gone by 2,000. On a part that runs at temperature, that is a failure inside three months of service, and it will be diagnosed as an adhesive problem when it is a surface preparation problem.

The mechanism is oxidation of the interface. Heat and oxygen attack the boundary layer, and without the silane bridge there is nothing holding the two phases together once the weak boundary layer oxidises. With the primer, the bridge is a covalent bond and it survives far longer.

This is why a peel test on a freshly made sample is not a qualification. A joint has to be tested after ageing at the service temperature for a representative time, and the acceptance criterion has to be the aged strength, not the initial one.

Bonding to different substrates

Aluminium bonds well with a primer, and anodised aluminium bonds better than bare because the oxide layer gives the silane something to grip. Bare aluminium that has been machined and left for a week has a loose oxide that fails cohesively; abrade and prime within a few hours.

Stainless steel is harder. The passive chromium oxide layer is chemically inert, and a primer that works on aluminium may not work on stainless. Abrading with 120 grit followed by an immediate prime is the usual answer, and a test piece is essential because the results vary by alloy.

Plastics are a mixed bag. Polycarbonate and ABS bond with a primer and often need a light solvent wipe to remove mould release. Polyolefins, PP and PE, are effectively unbondable without corona or flame treatment first, because their surface energy is even lower than silicone's.

Glass bonds well and needs only a clean, dry surface plus a primer. The failure mode on glass is usually a primer that was applied too thick, so the joint fails within the primer rather than at the interface.

Table 1 — Surface route by substrate

Reading the failure

When a bonded silicone joint fails, look at both faces. If the silicone is cleanly separated from the metal with no residue on either, the failure was at the interface: the primer did not bond, or there was no primer. If there is a film of silicone left on the metal, the failure was cohesive in the silicone, which means the bond was stronger than the material and the joint was as good as it could be.

A failure that leaves a white or chalky residue on the metal is usually an oxidised primer film, which means the primer was too old, too thick, or applied over a contaminated surface. A failure that leaves a mirror-clean metal face is a release agent problem.

The third pattern is a failure that starts at an edge and runs. That is a peel failure, and it usually means the edge was not sealed: the adhesive was squeezed out at the edge and the primer was exposed. Leaving a small fillet of adhesive around the edge, rather than squeezing it all out, prevents it.

Table 2 — Failure patterns and what each one means