Coating thickness is a through-thickness dimension. A surface instrument — magnetic induction, eddy current or X-ray fluorescence — infers thickness from a physical response. A microscope measures it directly, which is why it is the reference method when a coating matters: a plated layer on a fastener, a powder coat on a housing, an anodised film on aluminium, a thermal spray on a turbine part.
The trade-off is preparation. You have to cut the part, mount it, grind, polish and often etch it so the interface between coating and substrate becomes visible. Done correctly, the measurement is unambiguous and auditable. Done carelessly, the edge rounds over and you measure something that is no longer there.
Three things decide whether the coating edge survives to the microscope:
Every objective has its own magnification, so every objective needs its own calibration. Put a stage micrometer on the stage, capture an image at the working magnification, and tell the software how many pixels correspond to a known distance. Re-do it whenever you change objectives.
The most common measurement error is not optical — it is procedural: measuring a 50× image with the calibration entered for 20×. That yields a value 2.5× too large, and it looks plausible.
| Coating thickness | Objective | Total magnification | Why |
|---|---|---|---|
| > 50 µm | 5× or 10× | 50× – 100× | Whole layer plus both interfaces in one field |
| 10 – 50 µm | 10× or 20× | 100× – 200× | Balance of field and resolution |
| 2 – 10 µm | 20× or 50× | 200× – 500× | Layer must span enough pixels to measure |
| < 2 µm | 50× or 100× | 500× – 1000× | Needs high NA and careful focus |
A practical rule: the coating should occupy at least 30 pixels in the captured image. Below that, the uncertainty of picking the two edges exceeds the thickness itself.
Measure perpendicular to the interface, not along the cut direction. Take five readings across the field and report the mean and the range — a single reading on a sprayed or electroplated layer can be unrepresentative because thickness varies across the part.
For multilayer stacks (for example nickel over copper over steel), measure each layer separately and record them in order. Ambiguity about which interface belongs to which layer is the second most common source of error.
It happens regularly, and usually the microscope is right. Typical causes:
If the dispute matters — a supplier rejection, a warranty claim — the microscope result with a documented calibration is the one that holds.
A metallographic microscope with reflected illumination and a mechanical stage covers the optical part. Add a 12 MP camera and measurement software for traceable results, plus a cut-off machine, mounting press and polisher for the preparation. A 5× to 50× brightfield set handles most coatings; darkfield helps when the interface is low-contrast, and polarised light separates anisotropic coatings such as zinc on steel.
Tell us the material, the feature you need to see and the magnification range. We will recommend a microscope and quote it.
