When used on turbine parts, TBC thicknesses are typically ranging from 300-600 µm, on combustor parts, thicknesses are usually higher, they can reach 1mm and even >1.5mm. There is also for such coatings a difference in bondcoat manufacturing process, turbine parts using VPS or HVOF bondcoat while combustor parts have APS bondcoat.
The thickness is defined by trying to get the best compromise between the bondcoat temperature (requiring thicker TBC) and the TBC surface temperature (requiring thinner TBC). Since the cooling efficiency of turbine parts is in general higher than the one of combustor parts, this ‘optimal’ compromise is achieved with a thinner TBC.
The hot gas temperature is usually a given input, with always a wish to have it as high as possible such to increase turbine efficiency (with some limitations like emissions).
Then designers have to find the best compromise in keeping both metal (substrate and bondcoat) below an upper temperature limit (which can depend on mechanical loading) and at the same time keeping the TBC temperature below another upper temperature limit. These temperatures are usually calculated in steady state conditions at full load. This can sometimes underestimate the true maximal materials temperature since some parts can reach higher temperatures at part load (the cooling efficiency decreases with lower air mass flow).
To achieve such a compromise, designers will optimize, the cooling air flow, design turbulators on the cooling air side such too increase the cooling air heat transfer coefficient, use film cooling (by designing cooling holes such that the cooling air will flow along the coating outer surface and provide a hot gas temperature 'shielding layer' for the TBC) and adjust coating thickness.
So at the end what are relevant in having a design that is acceptable are the materials temperatures. Hot gas temperature is an input which influences what kind of design is possible. The hot gas temperature is usually given by CFD models, it is neither the static nor the toal air tempertaure (no corrections for pressure or velocity).
-Cross-sections is an effective way but destructive
-Blue light scanning is quite accurate but requires scanning before and after coatings
-Eddy-Current is quite standard, but it is substrate/BC dependent, so ideally it should be calibrated on the substrate before coating
-Thermography is another way but it is affected by the coating thermal diffusivity. Some people claims it is possible to measure both. I have some doubts that the measurement accuracy allows to determine both with high accuracy. So I would consider a sort of calibration is necessary.
-Coating weight is another possibility but it provides an average value
Also one shall not forget that each method measures a different thickness due to how it is impacted by the local thickness variation. This will in most cases bring an offset in the measured values.
Usually thick coatings are preferred to gas turbine engines. The optimum thickness may range between 350 to 500 um. more thickness has higher thermal gradient which leads to spallation of coatings.