Rubber-to-Metal Bonded Parts: Design & Specification Guide

How is rubber bonded to metal?

The metal decides the bond

Almost every bond failure we are asked to investigate is a surface preparation failure, not a chemistry failure. The bonding agents available today are extremely capable. What they cannot do is stick to oil, to rust, or to a coating that is itself loosely attached.

The sequence is degrease, roughen, coat, mould — promptly. Degreasing removes the drawing oils and handling residue; roughening, usually by grit blasting, gives the primer a mechanical key as well as a chemical one; the primer and cover cement go on in a controlled thickness. A prepared insert has a working life. It is not a part you blast on Monday and mould on Friday, and it is not a part anyone should pick up with bare hands — a fingerprint is an oil film in the shape of a finger.

The most expensive mistake is bonding to a plated or painted insert. The bond will be excellent — to the plating. The joint is then only as strong as the plating’s grip on the steel underneath, which is usually far weaker than the bond and was never designed to carry a structural load. Bond first, then plate the exposed metal if it needs protecting.

Which metals bond well

Every compound we mix can be bonded — natural rubber, SBR, nitrile, EPDM, neoprene, silicone, AEM and FKM. The compound is chosen for the duty, exactly as it would be on an unbonded part, and the bonding system is selected to suit it.

Where a bonded part is the answer

A bonded part exists wherever rubber has to carry a load into a structure rather than simply sit against it.

Anti-vibration mounts and engine mounts — the rubber is the spring, and the metal is how the spring is bolted to something. Suspension and pivot bushes — an outer sleeve, an inner sleeve and rubber between them, working in torsion and radial compression. Rollers and wheels, where a rubber covering is bonded to a metal core. Plungers, pistons and diaphragm assemblies, where a rubber sealing element sits on a metal stem.

And tank and pipe lining, which is the same technology at a different scale — rubber bonded to a steel vessel to protect it from abrasion or chemical attack. We supply sheet lining, and it is a bonded product.

Design so the bond is never the weak link

The design target is not a bond strength figure. It is that the bond is stronger than the rubber — so if the part is ever overloaded, the rubber tears and the interface holds. Size the rubber for the load in the ordinary way, and then make sure nothing in the geometry attacks the bond line.

Load the bond in shear or compression, never in peel. A bonded interface is enormously strong across its face and vulnerable at its edge. A load that tries to lift one corner of the rubber off the metal will start a peel, and once a peel starts it runs. Most bonded-part failures that are genuinely design failures are peel failures.

Take the rubber over the edge of the metal, or recess the metal. A bond line that ends flush at an exposed edge is both a stress concentration and a path for moisture. Corrosion creeping under the bond edge is the long-term failure mode of an otherwise perfect part, particularly on anything that lives outdoors.

Radius the metal. Sharp corners on an insert concentrate stress in the rubber next to them, which is exactly where the bond is.

And tolerance the insert. Its position in the mould sets the rubber thickness around it, so the insert’s dimensional variation becomes the part’s variation. An insert drawing with generous tolerances produces a bonded part with none.

How a bond is tested — and why the number is only half the result

Bond strength is measured by pulling the bond apart in a defined way. ASTM D429, Rubber Property — Adhesion to Rigid Substrates, is the method set most drawings cite, and it contains several procedures because a bond can be loaded in several ways.

The ISO equivalents are ISO 813, a 90° peel test on a rubber strip bonded to a single rigid plate, reported as force divided by specimen width; and ISO 814, the two-plate method, in which a rubber cylinder bonded to two parallel plates is pulled in tension and the breaking force divided by the bonded area gives an adhesion strength in MPa.

Here is the part most specifications get wrong. A bond test result is a force and a description of where the specimen broke, recorded with the standard failure-mode symbols the test standards define. The ideal result is that the break is entirely in the rubber — the bond survived and the material gave way, which is exactly what a bonded part should do in service.

So a high force with the break at the interface is a worse result than a lower force with the rubber torn. The first says the bond is the weak link and you found its limit. The second says the bond is stronger than the material it is holding, and the number you measured is the rubber’s strength, not the bond’s.

A low number can also simply mean a soft compound tore easily. Read the failure mode first, then the force. A specification that asks only for a minimum force has not specified a bond.

What we test in-house

We test tensile strength and elongation to ASTM D412, and compression set to ASTM D395 Method B, in our own lab at Nashik, and those results ship with every consignment alongside a dimensional report and a certificate of conformance.

Adhesion testing to ASTM D429, ISO 813 or ISO 814 is not on that in-house list. Where a customer’s drawing calls for a bond test, we arrange it through an external laboratory and supply the report. We would rather say that than imply a capability we would subcontract anyway.

In production we inspect every bonded part visually and by hand at the bond line, and destructive checks are taken from the batch at an agreed frequency. A bonded part cannot be reworked — there is no second chance at the interface once the part is cured — so the control is in the preparation and the first-off, not in an inspection at the end.

We hold ISO 9001:2015, certificate 09111526A, Integral Certification Ltd, valid to 21 February 2027, scope: rubber moulded components — which is what a bonded part is, since the rubber is moulded onto the insert. That is a quality-management certification, not a product approval. We are upgrading to IATF 16949, targeted by the end of 2026.

Five mistakes with bonded rubber parts

Supplying plated inserts

The bond holds to the plating, and the plating’s grip on the steel becomes the weak link. Bond first, then plate the exposed metal.

Loading the bond in peel

A bonded interface is very strong across its face and weak at its edge. Design the load into shear or compression — once a peel starts, it runs.

Specifying a minimum force and nothing else

A bond result is a force and a failure mode. A high force with the break at the interface is worse than a lower force with the rubber torn.

Ending the bond line at an exposed edge

It is a stress concentration and a moisture path at once. Corrosion creeping under the edge is what kills otherwise perfect outdoor parts.

Loose tolerances on the insert

The insert’s position in the mould sets the rubber thickness around it. Its variation becomes the finished part’s variation.