The bond is a chemical joint, not glue
A proper rubber-to-metal bond is formed during vulcanisation. A two-coat adhesive system — a primer that keys to the prepared metal and a cover coat that co-cures with the rubber — creates a chemical bridge between the two materials. Done correctly, the bond is stronger than the rubber itself, and a destructive test tears the rubber rather than peeling the interface.
That is the standard we work to and the one you should insist on. If a failed part shows clean, shiny metal where the rubber came away, the bond was never right. If it shows a layer of rubber still stuck to the metal, the bond was sound and the rubber was simply overloaded.
The diagnostic that takes ten seconds
Look at the metal on a failed bonded part. Clean metal means bond failure — a manufacturing problem. Rubber still adhering means the bond held and the rubber tore — a design or material problem. These need completely different fixes.
Cause one: surface preparation
By far the most common cause. Metal arrives with mill scale, rust, cutting oil or a rust-preventive film. Any of those between adhesive and steel and the joint is compromised. Preparation means degreasing, then grit blasting to a clean, uniformly roughened surface, then bonding promptly before the surface re-oxidises.
The time between blasting and priming matters. A blasted steel insert left overnight in a humid shop has already started to oxidise.
Cause two: the wrong adhesive for the elastomer
Adhesive systems are matched to polymer families. A system that bonds natural rubber beautifully may perform poorly with EPDM or silicone. This is why a supplier who only ever bonds one compound sometimes struggles when a customer asks for the same part in a different material.
Cause three: cure conditions
The bond and the rubber cure together. Too little time or temperature and the bond is immature even though the part looks finished. Too much and the adhesive can degrade. Bonded parts are less forgiving of cure drift than plain moulded parts, which is one reason we keep cure parameters recorded per part rather than per press.
Cause four: design that loads the bond in peel
Even a perfect bond is far weaker in peel than in shear or compression. A design that puts a peeling load on the edge of the bond line will eventually lift it, regardless of how well it was made.
- Keep the bond line in compression or shear wherever the geometry allows.
- Avoid sharp edges at the bond termination — a small rubber fillet running onto the metal spreads the stress.
- Do not let the rubber end flush with a loaded edge; carry it slightly over.
- Where the joint must take peel, increase the bonded area rather than the rubber thickness.
What to ask a supplier
- 1How is the metal prepared, and how long between preparation and bonding?
- 2Which adhesive system, and is it matched to this specific polymer?
- 3Is bond integrity checked on every batch, or sampled?
- 4Can you show a destructive test where the failure is in the rubber, not the interface?
A supplier who cannot answer those four questions clearly is not controlling the process. We pull-check bonded parts every batch, because a mount that debonds under a running machine is not a quality problem — it is a safety one.
Have a part that fits this description?
Send the dimensions and the operating conditions. We will tell you what compound it should be in and quote it — including saying so if it is not a job for us.