Determination of Small-Signal Parameters and Noise Figures of MESFET's by Physics-Based Circuit Simulator Employing Monte Carlo Technique

摘要

We present a physics-based circuit simulator employing the Monte Carlo (MC) particle technique, which serves as a bridge between the small-device physics and the circuit designs. Two different geometries of GaAs-MESFET's are modeled and analyzed by the simulator. The Y-parameters of the devices are extracted from the transient currents, and then translated into the S-parameters. The cut-off frequency (f_T) is estimated from the Y-parameters. The minimum noise figure (F_(min)) is also estimated by evaluating the fluctuation in the stationary current. The device, having the n~+-region placed just at the drain side of the gate, exhibits the better performances in both f_T and F_(min). The analysis on the equivalent circuit (EC) elements reveals that its better performances are mainly due to the reduced gate-source capacitance (C_(gs)) and the increased transcon-ductance (g_(m0)), which result from the shortened effective gate length (L_g) caused by the termination of the depletion region at the gate. edge. The termination of the depletion region, however, causes the increase of the electric field, which results in the higher heat generation rate near the gate edge. It is proven that the physics-based circuit simulator developed here is fully effective to see the inside of the small-device and to model it for the millimeter-wave circuit design.