Near-IR Thickness Measurement

We have a coating (perhaps wet) that is sufficiently clear to near-IR light so that specific wavelengths of the the light are absorbed by the coating. The substrate might be clear so the IR absorption can be measured in transmission - which requires a sensor on the opposite side of the substrate. Or if the substrate scatters, the device can be single-sided. If the substrate is clear and reasonably transparent to the specific IR wavelength then a back-scattering plate behind the substrate can also be used for those who want a single-sided systems. The thickness of the coating is calculated from a linear dependence on the absorption of specific wavelengths - assuming that there is no absoroption by the substrate. Because it can be hard to find a single wavelength that distinguishes coating from substrate, the system will generally measure the signal at a number of wavelengths (using a rotating filter) and from calibration experiemnts will find the best combination of signals to give a reliable result.

In the app only a single measurement band is used and substrate and coating each have a single absorption band for simplicity.

Enter the thickness, absorption wavelength (μm) and absorption strength of coating and substrate and the measurement wavelength. The app gives an idea of the accuracy of the resulting measurement by calculating Err_1%T. This is the error in the value of the measured thickness coming from a 1% error in the %T value. The logic behind this is discussed below.

As you play with the values you quickly find the limits of the method. The key lies in the Err_1%T. If, at the measurement wavelength, the %T of the coating is somewhere between 10% and 90% then and if the %T of the substrate is more than 90% then the thickness error from a 1% variation in measured %T (±0.5%) is not too bad. At both high and low %T the error is large, and the more the absorption of the substrate overlaps the absorption of the coating, the higher the error. That's why the measurement wavelength is so important. Normally you would expect to use the absorption maximum of the coating as the sweet spot for measurement, but if the coating is too thick (giving a very low %T) or if there is too much absorption by the substrate, moving away from the maximum can improve things.

Of course, in a real system things are more complicated, but the app gives you a good idea of the trade-offs involved in Near-IR measurements. These are:

• For very thin coatings there isn't enough absorption to give a good signal.
• For very thick coatings there may be too much absorption.
• The coating and the substrate should ideally have very different absorption bands. Water has a strong peak at 1.94μm and PE has peaks at 1.72 and 2.32μm - which explains why Near-IR measurements of water-based coatings on PE are so popular.
• A simple low-cost system might be unable to find a set of measurement wavelengths available to a more sophisticated system.
• The thickness generally has to be calibrated, though multi-wavelength systems reduce the need for calibration.
• The mix of absorption peaks of solvent and solute will complicate measurements and a set-up ideal for a wet coating may be useless for a dry coating.

Summary

• Accuracy: Broad areas of high accuracy with obvious areas of low accuracy.
• Size: Medium
• Safety: Safe
• Scanability: Easy to scan, and to use in spot mode
• Cost: Medium
• Wet/Dry: Works on both, but requires different settings
• Big limitations: Needs big difference in Near-IR absorption - harder to find for solvent coatings, easy to find for aqueous