Broadband Permittivity Characterization of a Substrate Material Using Deep Neural Network Trained With Full-Wave Simulations

It is crucial to know the permittivity of dielectric materials used in radio frequency (RF) components and devices because their operation loss characteristics are significantly affected by permittivity. In this study, we propose a characterization technique based on deep neural network (DNN). The latter was trained using data obtained from full-wave electromagnetic simulation software. With DNN with more than 95% testing accuracy, measured complex transmission coefficient material under test (MUT) assigned as an input model, MUT retrieved at output. proposed validated measuring FR-4 epoxy resin substrates different thicknesses. results model showed good agreement each other, error less 1.2% for relative value over broad range 1 – 10 GHz. We also compared those conventional analytical solutions highlight effectiveness method.