Abstract:Polytetrafluoroethylene (PTFE) is widely used in microelectronic packaging, high-frequency circuits, and medical devices due to its extremely low surface energy and excellent chemical stability. However, its poor surface wettability leads to insufficient bonding performance. Plasma activation technology can effectively improve the surface energy of PTFE, but its activation effect suffers from significant aging attenuation, seriously affecting subsequent bonding consistency and reliability. To address this issue, this paper proposes an AI-based aging control method for plasma activation of PTFE. By constructing a multivariate data model of "process parameters—surface energy—aging attenuation", XGBoost and Long Short-Term Memory (LSTM) networks are introduced to predict the surface energy attenuation process after activation, and reinforcement learning is combined to achieve dynamic optimization of process parameters.