When we talk about a coating, most people picture a layer of paint that makes a surface look better or keeps rust at bay. But in failure analysis and quality work, a coating is a thin engineered layer that changes how a material responds to load, temperature, moisture, and time. It sits at the boundary between the base material and the environment, and it can shift the failure mode from sudden fracture to slow, detectable degradation. For engineers who read fracture surfaces and test data, the coating is not just a skin. It is a system variable that can hide or reveal the true condition of the substrate.
This article looks at the less obvious ways coatings protect materials. We will talk about thermal barriers, galvanic isolation, stress redistribution, sealing of surface flaws, and the way a coating can change the apparent hardness or conductivity of a part. We will also look at how coatings fail, because a coating that fails in the wrong way can create a new failure mechanism instead of preventing one.
What a Coating Actually Does at the Surface
A coating is a discontinuity in the material system. It has its own modulus, thermal expansion coefficient, porosity, and adhesion strength. When you apply a 50 µm layer of zinc to a steel bracket, you are not just adding corrosion resistance. You are adding a layer that will deform differently under load, crack at a different strain, and interact with moisture in a different way than the steel below it.
In the lab, we see this when a coated sample is bent or pulled. The coating may crack before the substrate yields. That crack can then act as a stress concentrator at the interface. If the coating is brittle and well bonded, the crack can propagate into the substrate. If the coating is ductile and poorly bonded, it may delaminate and leave the substrate exposed. Neither outcome is obvious from a simple thickness measurement.
The Coating as a Thermal Barrier
One of the clearest examples is a thermal barrier coating on a turbine blade. A yttria-stabilized zirconia layer, typically 100–400 µm thick, can reduce the metal surface temperature by 100–170 °C under engine conditions. That temperature drop is enough to extend creep life by a factor of two or more in some nickel-based superalloys. The coating itself is porous, with 10–20% porosity by volume, and that porosity is intentional. It lowers the thermal conductivity of the layer and gives it some strain tolerance.
But the same porosity that helps with thermal insulation can become a path for oxygen and molten salts. If the bond coat below the zirconia oxidizes, a thermally grown oxide layer forms. That oxide grows slowly, but it grows. When it reaches 5–10 µm, the stress at the interface can be enough to spall the ceramic top coat. The failure looks like a clean delamination, but the root cause is a chemical reaction that started the day the part went into service.
Galvanic Isolation and Sacrificial Protection
A coating can protect by being either a barrier or a sacrificial anode. A barrier coating, like an epoxy on steel, works only as long as it is continuous. A pinhole or a scratch becomes a small anode surrounded by a large cathode, and the corrosion rate at that pinhole can be surprisingly high. We have measured corrosion pits under scratched epoxy coatings that were 3–5 times deeper than pits on bare steel in the same salt spray test. The coating did not fail; it concentrated the attack.
A zinc-rich coating works differently. Zinc is anodic to steel, so it corrodes preferentially. The steel is protected even at scratches and cut edges, as long as the zinc is electrically connected and there is enough zinc to supply current. The protection distance depends on the conductivity of the environment. In seawater, zinc can protect steel for several millimetres from a scratch. In freshwater, the protection distance is much smaller, often less than 1 mm. That is why a zinc coating that works well on a marine structure may not protect a freshwater pipe fitting the same way.
How Coatings Change Mechanical Test Results
If you measure the hardness of a coated part with a standard Vickers or Rockwell indenter, the coating can change the number you get. A 20 µm hard chrome layer on a soft aluminium substrate can raise the apparent hardness by 10–20 HV, depending on the load. At low loads, the indenter may not penetrate through the coating, and you are measuring the coating, not the substrate. At high loads, the coating cracks and the substrate deforms, and the result is a composite value that is hard to interpret.
The same issue appears in tensile testing. A brittle coating can crack at 0.5–1.0% strain, long before the substrate yields. Those cracks are visible on the surface, but they do not necessarily mean the part has failed. In fatigue testing, however, those early coating cracks can initiate fatigue cracks in the substrate. We have seen fatigue life drop by 30–50% when a hard, brittle coating is applied to a steel shaft, even though the coating itself was intact after the test. The cracks were the problem, not the coating loss.
Sealing Surface Flaws
One of the most underappreciated benefits of a coating is that it can seal surface flaws. A machined surface has microcracks, laps, and tool marks. A coating can fill those features and reduce the stress concentration at the surface. In some cases, a 10–20 µm polymer coating can increase the fatigue strength of an aluminium alloy by 10–15%, simply because it fills the surface roughness valleys and delays crack initiation.
This effect is not universal. If the coating is brittle and well bonded, it can transmit the stress concentration from the surface flaw into the coating, and the coating cracks. If the coating is ductile and well bonded, it can bridge the flaw and share the load. The key is the adhesion strength and the modulus mismatch. A coating that is too stiff will crack; a coating that is too soft will not carry enough load to help.
When Coatings Hide Problems
A coating can also hide problems. A crack in a steel weld can be covered by a layer of paint or zinc, and it will not be visible during a visual inspection. The crack can grow under the coating, and the first sign of trouble may be a sudden fracture. In one case we reviewed, a 12 mm thick steel plate in a bridge structure had a fatigue crack that grew to 8 mm under a 200 µm paint layer. The paint was intact, and the crack was only found when the paint was removed for a routine repaint. The coating had protected the crack from corrosion, but it had also protected it from detection.
This is why coating inspection is not just about thickness and adhesion. It is about looking for the signs that the coating is hiding something. Blisters, discolouration, and localised cracking are all clues. A blister in a paint film is often a sign of corrosion under the coating, and the corrosion may be more advanced than the blister suggests.
Coatings and Electrical Contact
A coating can also change the electrical behaviour of a part. An anodised aluminium surface is an insulator, even though the aluminium below is a conductor. If you are measuring resistance or continuity, the coating can give you a false reading. In electronic enclosures, a chromate conversion coating is often used because it is thin enough to allow electrical contact while still providing some corrosion protection. The thickness is typically 0.1–0.5 µm, which is 100–1000 times thinner than a typical paint layer. That thinness is a deliberate trade-off between protection and conductivity.
Practical Ways to Evaluate Coating Protection
When we evaluate a coating in the lab, we look at more than thickness. We look at adhesion, porosity, hardness, and the way the coating responds to the same environment the part will see. A simple cross-cut tape test can tell you if the coating is well bonded, but it does not tell you how the coating will behave under thermal cycling or mechanical load. For that, we use bend tests, thermal shock tests, and electrochemical impedance spectroscopy.
Electrochemical impedance spectroscopy is particularly useful for barrier coatings. It measures the impedance of the coating at different frequencies, and that impedance changes as the coating absorbs water and begins to fail. A coating that starts with an impedance of 109 ohm·cm2 may drop to 106 ohm·cm2 after 100 hours in salt water. That drop is a warning sign, even if the coating looks unchanged.
Adhesion Testing and What It Misses
Adhesion testing is often treated as a pass/fail test. A pull-off test gives a number in MPa, and if the number is above a specification, the coating is accepted. But adhesion is not a single property. It depends on the test method, the surface preparation, the coating thickness, and the environment. A coating that passes a pull-off test at 20 °C may fail at -20 °C because the polymer becomes brittle and the thermal stresses increase. A coating that passes on a clean, blasted surface may fail on a surface with a light oil film, even if the oil film is only a few nanometres thick.
We have also seen cases where a coating with high adhesion failed in service because the substrate corroded under the coating. The adhesion was high, but the coating was permeable to water and oxygen. The corrosion products grew under the coating and lifted it off. The failure was not at the original interface; it was at the new interface between the corrosion product and the coating. Adhesion testing before service would not have predicted this.
Case Example: Zinc Flake Coating on Fasteners
Zinc flake coatings are common on automotive fasteners. They are thin, typically 8–12 µm, and they provide good corrosion resistance without the hydrogen embrittlement risk of electroplated zinc. But they also change the friction behaviour of the fastener. The coefficient of friction can vary from 0.10 to 0.18 depending on the coating formulation and the lubricant. That variation changes the clamp load for a given torque. If you torque a zinc flake coated bolt to the same value as a bare steel bolt, the clamp load can be 20–30% higher or lower, depending on the friction. That is a real, measurable effect that is not obvious from the coating thickness alone.
In one test series, we measured the clamp load of M10 fasteners with three different zinc flake coatings. The torque was the same, 50 N·m. The clamp loads were 18.5 kN, 22.1 kN, and 24.7 kN. The difference was entirely due to the coating friction. The coating that gave the highest clamp load also had the lowest friction, and it was the one with the most consistent torque-tension behaviour. The coating was protecting the fastener from corrosion, but it was also changing the joint behaviour in a way that had to be accounted for in the design.
What to Look for in a Coating Failure
When a coated part fails, the first question is whether the coating caused the failure or just happened to be there. The answer is often in the fracture surface. If the crack started at a coating defect, you will see a small, localised origin with a different appearance from the rest of the fracture. If the coating delaminated, you will see a clean separation at the interface, often with corrosion products or contamination on one side. If the coating cracked first and the crack propagated into the substrate, you will see a series of parallel cracks in the coating that line up with the main fracture.
We also look at the coating thickness and uniformity. A coating that is too thick can crack under thermal stress. A coating that is too thin may not provide enough protection. A coating that is non-uniform can create local galvanic cells. All of these are measurable, and all of them can be traced back to the coating process or the service environment.
The Role of Surface Preparation
Surface preparation is the single most important factor in coating performance. A coating applied to a surface with mill scale, rust, or oil will not perform the same as a coating applied to a clean, blasted surface. The standard for steel surface preparation, ISO 8501-1, defines four grades of blast cleaning: Sa 1, Sa 2, Sa 2½, and Sa 3. The difference between Sa 2 and Sa 2½ is the amount of staining and mill scale remaining. That difference can change the coating life by a factor of two or more in a marine environment.
We have seen coating failures that were blamed on the coating, but the real cause was a surface that was not cleaned to the specified grade. The coating was fine; the preparation was not. That is why we always document the surface condition before coating, and why we take photographs of the surface at each stage. The evidence is in the surface, not just in the coating.
Frequently Asked Questions
How does a coating protect a material from corrosion if it is not a perfect barrier?
A coating can protect by being a barrier, by being a sacrificial anode, or by changing the local chemistry. A barrier coating works by slowing the transport of water and oxygen. A sacrificial coating, like zinc on steel, works by corroding in place of the steel. Some coatings also contain inhibitors that are released slowly and reduce the corrosion rate at defects. The protection is not perfect, but it can extend the service life by orders of magnitude.
Can a coating increase the fatigue life of a metal part?
Yes, in some cases. A ductile coating that fills surface flaws and shares the load can delay crack initiation and increase fatigue life by 10–15% or more. A brittle coating can do the opposite, cracking early and initiating fatigue cracks in the substrate. The effect depends on the coating modulus, adhesion, and thickness, as well as the surface condition of the substrate.
Why do some coatings fail by delamination instead of cracking?
Delamination occurs when the stress at the coating-substrate interface exceeds the adhesion strength. That stress can come from thermal expansion mismatch, corrosion product growth, or mechanical load. If the coating is ductile and the adhesion is low, delamination is more likely than cracking. If the coating is brittle and the adhesion is high, cracking is more likely. The failure mode tells you something about the coating system and the service conditions.
How can I tell if a coating is hiding a crack in the substrate?
Look for localised signs such as blisters, discolouration, or a line of cracked coating that follows a straight or slightly curved path. These can indicate a crack or corrosion under the coating. Non-destructive methods like eddy current testing or ultrasonic testing can detect cracks under coatings, but the coating thickness and conductivity can affect the results. When in doubt, remove the coating in a small area and inspect the surface directly.
Next Steps for Your Coating Investigations
If you are working with coated parts, start by documenting the coating system: the substrate, the surface preparation, the coating type, the thickness, and the cure conditions. Then think about the service environment: temperature range, moisture, mechanical load, and any chemicals present. The coating is not a separate layer; it is part of the material system, and it changes the way the system responds to stress and environment.
In a future article, we will look at how to interpret cross-sections of coated parts under a microscope, including what the interface tells you about adhesion and what the porosity tells you about the coating process. If you have a coating failure you are trying to understand, send us a note with the details. We read every message and use the questions to shape the next articles.
