Four impedance matching methods for capacitatively coupled plasma source with transmission line and lumped parameter circuit.

Capacitively coupled plasmas (CCPs) have diverse applications in fields such as plasma etching and thin-film deposition. To maximize the efficiency of plasma generation and minimize the reflection of power during transmission, an impedance matching network (IMN) and design method for capacitively coupling plasmas play a crucial role. The coupling and nonlinear interactions of the transmission line, the lumped parameter circuit and the plasma load pose great challenges to this design^[1-5]. In this study, they are described by the Lax-Wendroff method, Kirchhoff Voltage Law (KVL), and the Particle-In-Cell/Monte Carlo Collision (PIC/MCC) method respectively. Four distinct methods are utilized to design optimal impedance matching for this intricate discharge system: (1) The impedance matching design of the plasma source with two ideal transmission lines and an L-type matching network is realized by iterative calculation. Only a few iterations are required to attain a reflection coefficient below 0.01. (2) Impedance matching is achieved by employing a dichotomy approach to adjust the frequency of the RF power supply. (3) For single-frequency CCPs, two adjustable capacitors are adjusted simultaneously by the gradient descent method to minimize the reflection coefficient. (4) Machine learning techniques are leveraged to train on existing simulation data, enabling the identification of optimal circuit parameters.