An AI Predicted Capacitance-Based Electromagnetic Analysis Method about Metal Fill Effect for 5G/B5G Applications

An electromagnetic analysis method based on artificial intelligence predicted capacitance for 5G/B5G chip with metal fills is proposed. The influence of the metal fills is efficiently evaluated by the effective permittivity of the dielectric layer containing metal fills. Artificial intelligence is used to accurately and efficiently evaluate the parasitic capacitors about the metal fills, which enables very approximate estimation of the effective permittivity. For verification, the S-parameters of an on-chip inductor using the proposed method and direct computation are compared.