Noncontact measurement of moisture in layered dielectrics from microwave reflection spectroscopy using an inverse technique

This paper describes a microwave-based measurement mechanism, which realizes a totally noncontact assessment of the moisture contents in layered materials. An inverse technique is used to obtain the properties of layered materials from the measurement of reflection of plane electromagnetic waves. The technique is based upon a nonlinear least squares method. Although a generalized method is presented for multilayered dielectrics the experimental results are obtained, for simplicity, for a single layer backed by metal surface. The reflection coefficients measured at X -band are analyzed according to the model to match with the results from Fresnel's equations to obtain thickness and permittivity. The convergence rate depends upon the level of noise present in the measured reflection coefficients. The effect of various types of error measures upon the convergence of iteration is also discussed. The results obtained for acrylic sheets and moist medium density fiber board are excellent. Determination of an empirical dispersion relationship for moist material from this technique is also discussed. The proposed technique has proven to be very useful in modeling noisy data for reflections from a single layer of dielectric backed by a metal surface, where most of the noniterative techniques fail. The technique can be equally useful in finding the inverse of a variety of electromagnetic problems.