Adhesion: Four Tests, and Three of Them Destroy the Part

**The bend test.** Bend the coupon through 180 degrees over a mandrel of a few times its thickness, then examine the outside of the bend under magnification. A well-adhered coating **cracks with the substrate** — it may craze, but the pieces stay attached, because the coating and the metal are failing together. A poorly adhered coating **lifts, flakes or peels away from the bend**, often in a sheet. That is a clean separation at the interface, and it means the surface was not clean. **The burnish test.** Rub a small area briskly with a smooth hard tool — the back of a knife, a burnisher — for fifteen seconds. A well-adhered coating polishes. A poorly adhered one blisters, because the friction works it loose from underneath. **The file test.** File across the edge of the coupon at 45 degrees, from the substrate towards the coating. A good coating is cut through cleanly; a bad one lifts ahead of the file. **Look at the WAY it failed.** A coating that lifts in large sheets with a shiny back was on a greasy surface. One that lifts in small flakes with a dull grey back was on an oxidised one. The fracture surface names the cause.

**Heat the part and quench it.** The coating and the substrate expand by different amounts, so a sudden temperature change puts a large shear stress directly on the interface — which is exactly what an adhesion test wants to do. Typical: 150 to 300 C for an hour, depending on the substrate, then straight into cold water. Blistering or peeling afterwards is a fail. **It is the most searching of the tests**, because it stresses the whole area at once rather than one line, and it finds marginal adhesion that a bend test passes. **It doubles as the hydrogen bake.** Heating plated high-strength steel to 190 to 220 C for several hours drives dissolved hydrogen out before it can cause delayed brittle fracture, and it must be done within hours of plating. A part that both bakes and survives the quench has answered two questions. **Do not thermal-shock a part that has been hardened**, unless the temperature is safely below its tempering temperature. The test would ruin the part in a different way. **Record which test was used** alongside the result. "Passed adhesion" means nothing without it, in the same way that a purity figure means nothing without its method.

**Contamination.** Oil, grease, a fingerprint, polishing compound, or drag-in from a previous bath. This is the great majority of failures, and the water-break test would have caught it. **Oxide.** The part was pickled, then stood in air, then plated. A clean steel surface re-oxidises visibly in minutes and invisibly in seconds. **Passivity.** Some metals — stainless steel, aluminium, titanium, chromium — carry an oxide film that re-forms faster than it can be removed. They need a specific activation step: a **Wood's nickel strike** for stainless, a zincate for aluminium. Plating them directly always fails, however clean they are. **Interrupted current.** If the current stops mid-plate and then restarts, the surface passivates in the gap and the second layer does not bond to the first. A bath that has lost power is stripped and restarted, not resumed. **Internal stress.** A highly stressed deposit — bright nickel with too much brightener, or chromium, which is always in tension — pulls itself off a marginal bond. It can peel weeks later with nothing else having changed. **Dissimilar expansion.** A thick hard coating on a substrate with a very different expansion coefficient, in service across a temperature range. The first two account for most of it, and both are preparation.

A failed coating has to come off completely before another can go on, and a partially stripped part plates worse than a bare one. **Chemical stripping** is the usual route and it is specific to the pair of metals. The requirement is a solution that dissolves the coating and does not attack the substrate — which is exactly the opposite of the plating chemistry, and is why a strip for nickel-on-steel is not a strip for nickel-on-brass. **Reverse-current stripping** makes the part the anode and dissolves the coating off electrochemically. Controllable, and it stops when the coating is gone if the solution is chosen so the substrate passivates. **Mechanical stripping** — blasting, grinding — always removes some substrate and changes the dimensions. It is the last resort on a precision part and the first choice on a rough one. **Then start from the beginning.** A stripped part goes back through mechanical preparation, degreasing and pickling, not straight into the plating tank. The stripping solution is itself contamination. **Consider whether it is worth it.** Stripping and replating a small part often costs more than the part. The decision is easier with a test coupon, because the failure is discovered on the coupon rather than after the part is assembled into something.