A Novel Dynamic Measurement Method for Complex Permittivity of Microwave Plasma

This paper proposes a novel dynamic method for the equivalent complex permittivity of microwave plasma. Firstly, the samples with different complex permittivity are simulated in the open rectangular waveguide based on WR-430 at microwave frequency of 2.45 GHz to obtain their S parameters. There exists a certain correspondence between the complex permittivity and the S parameters, which we exploit to construct a backpropagation neural network (BPNN). The simulated data obtained will serve as the training set for the BPNN. Then, we establish a two-port network experimental system to measure the S parameters of microwave plasma, which are subsequently used as input values for the BPNN. Finally, the BPNN inversion enables the complex permittivity to be obtained. We validate the feasibility of this measurement approach by comparing inversion results with theoretical calculation results.