Integrated FDTD and solid-state device simulation

Efficient Modeling of PIN Diode Switches Employing Time-Domain Electromagnetic-Physics-Based Simulators

“…on and off. Therefore, for small-signal modeling, it is not required to directly couple the semiconductor physics with Maxwell's equations during simulation, which would not be the case for nonlinear largesignal modeling [6]. Instead, in this preliminary work, the Extended-FDTD approach is used to model PIN diodes with a relatively good accuracy.…”

“…In our GaAs PIN diode case, m n and m p are 450 and 8000 cm 2 /Vs. On the other hand, under reverse bias, another approximation that is used instead of running a driftdiffusion simulator is to assume a constant depletion capacitance [9]: (6) Here e is the dielectric constant of the depletion layer material. In Fig.…”

“…Accordingly, it is indispensable to present analysis of PIN switches based on a coupled full-wave simulator. The possibility of achieving this type of modeling is addressed by global circuit modeling as has been demonstrated in [3]- [6]. Here, we present a fast electromagnetic (EM) simulator for efficient modeling and optimization of low-loss and high-isolation mm-wave PIN diode switches.…”

Coupling of Solid-State and Electromagnetic Equations for Simulation of Wireless Packaged Geometries

“…Various solid-state models exist for these types of devices [1] and have been previously integrated with electromagnetic simulators (EMS), such as FDTD [2,3,4]. However, most of the prior implantations of this coupling have been limited in application due to the fact that the application of the CFL conditions [1] and stability requirements for the two models yields disparate time steps often with several orders of magnitude difference [2,3]. The active device simulator (DS) has the smallest time and space stepping.…”

A global finite difference time domain analysis of a silicon nonlinear transmission line

“…Furthermore, in contrast to other standard NLTL's, an optimized inhomogenous doping profile was designed to achieve a faster compression of the falling edge of the output signal. For a better and faster large-signal modelling of the diodes, the device response is extracted from the geometry and doping profile itself applying the finite difference method to the drift-diffusion model as it was previously done in [9] and [10].…”