Time-domain terahertz reflectometry has long been of interest to obtain three-dimensional information about objects transparent in the 100 GHz-3 THz frequency range. The application of such techniques has been limited by the conductivity of the materials and by conventional axial resolution criteria. Specifically, I discuss work on detecting subsurface damage in carbon fiber composites and in measuring paint-layer thicknesses down to ~20 microns. We employ a typical commercial broadband time-domain source with bandwidth from ~100 GHz to ~3 THz. Yet terahertz imaging faces two important limitations. (1) The first is that in materials with a significant conductivity, the attenuation of incident terahertz electromagnetic waves is very rapid, limiting the penetration into the object. (2) The second limitation can best be illustrated by a stratified medium. Due to the Fresnel coefficients associated with the variation of the refractive index across material interfaces, one expects a reflected signal that is composed of various echoes associated with the interfaces (as well as multiple reflections which in practice are often weak). If the layers are sufficiently thin, then the echoes from successive layers will overlap, and thus multiple echoes will not be visually evident in the reflected signal.
In this talk I discuss my group’s recent work to circumvent these limitations. The approaches rely upon applying advanced signal-processing techniques to extract the maximum information from the detected signals. Specifically, I provide case studies of coatings on metals, fiber composite laminates, and an oil painting on canvas with multiple paint layers.
|