The fundamental operating principle of terahertz time-domain spectroscopy (THz-TDS) for coating thickness measurement relies on the time-of-flight (ToF) of a broadband terahertz pulse through a layered structure. When a sub-picosecond THz pulse is incident on a coated tablet, partial reflections are generated at each dielectric interface — the air/coating boundary and the coating/tablet-core boundary. The temporal separation between these echo pulses, combined with a known or measured refractive index of the coating material, yields the optical path length and hence the physical thickness.

As formalised by Hanyang University (2021), when refractive indices of the respective layers are already known, layer thicknesses can be extracted through differences in reflection times of terahertz waves alone; when they are not known, simultaneous determination of both refractive index and thickness is achievable through combined transmission and reflection time measurements. This dual-mode capability is central to pharmaceutical applications where coating materials may vary by product or batch.

System performance parameters — particularly bandwidth and dynamic range — set fundamental limits on thickness resolution. TH Bingen University of Applied Sciences (2020) quantitatively investigated how dynamic ranges from 10 to 90 dB and bandwidth variations affect the accuracy of single-layer thickness determination for coatings on metal substrates spanning 0.5 to 100 µm, establishing minimum system requirements for reliable pharmaceutical coating metrology. The WIPO patent database documents active filings across multiple jurisdictions for these measurement architectures.

For multilayer coating structures — such as functional enteric coatings over sub-coatings — ABB Corporate Research (2014) introduced a stratified system model treating each layer's light-matter interaction realistically, enabling highly accurate material parameter extraction applicable to industrial quality control of coated products.

The accuracy of thickness extraction depends critically on the precision of the time-of-flight determination, since errors in thickness propagate directly into derived optical constants. Multiple advanced signal processing approaches have been developed and validated on pharmaceutical materials.

The Total Variation (TV) method, demonstrated by Queen Mary University of London (2013), exploits multiple internal reflections ("ringing") within a sample to estimate sample thickness to sub-micron accuracy, considerably improving the reliability of optical parameter extraction. This approach is particularly powerful for thin coatings where the temporal separation of echo pulses approaches the system's temporal resolution limit.

A complementary approach from the Chinese University of Hong Kong (2019) proposed fitting a non-inflection offset exponential function to material optical properties, demonstrating that the best fit is only achieved at the correct thickness — a technique validated on lactose pellets and thin polymer films. The FDA's Process Analytical Technology (PAT) framework explicitly supports such real-time analytical methods in pharmaceutical manufacturing.

For automated extraction from THz A-scans, the University of Cambridge (2015) proposed a wavelet denoising and peak-finding algorithm for automated coating thickness quantification. The practical deployment of these methods at production scale has been validated by TeraView Ltd. and the University of Liverpool in combined TPI/OCT in-line sensing studies. The European Medicines Agency recognises non-destructive testing as a quality-by-design enabler for solid dosage forms.