MICROWAVE ELLIPSOMETRY CHARACTERIZATION OF ISOTROPIC MATERIALS USING SIMULATED ANNEALING COMBINED WITH LEVENBERG-MARQUARDT ALGORITHM

The characterization of thick materials using microwave ellipsometry is known as non destructive technic. New microwave Ellipsometry has been developed and tested at different frequencies: in the 24-40 GHz frequency band and at 10 GHz. Several materials were characterized through this device using Levenberg-Marquardt optimization routine. Complex objective function with several local minima has been to be solved. Many recent works have shown that this routine has been very efficient in finding local minimum, but not for global minimum unless using some tricks. In this paper, we have combined two algorithms, namely simulated annealing and Levenberg-Marquardt algorithms, in order to enhance the characterization technic in finding the global optimum. Simulations were performed and concluded that the combined algorithm is more efficient than any of its basic algorithms. Experimental results on PTFE sample considering Cauchy dispersion model validated the simulation assumption.