Explicit modelling of the double-gate MOSFET with VHDL-AMS

Abstract

International audienceThis paper presents a new compact model for the undoped, long-channel double-gate (DG) MOSFET under symmetrical operation. In particular, we propose a robust algorithm for computing the mobile charge density as an explicit function of the terminal voltages. It allows to greatly reduce the computation time without losing any accuracy. In order to validate the analytical model, we have also developed the 2D simulations of a DG MOSFET structure and performed both static and dynamic electrical simulations of the device. Comparisons with the 2D numerical simulations give evidence for the good behaviour and the accuracy of the model. Finally, we present the VHDL-AMS code of the DG MOSFET model and related simulation results

Explicit compact model for symmetric double-gate MOSFETs including solutions for small-geometry effects

“…Then, we proposed a robust algorithm for computing the mobile charge density as an explicit function of the terminal voltages. It has allowed to greatly reduce the computation time without losing any accuracy [12]. The model continuously covers all operation regions and is in excellent agreement with 2-D numerical simulations without any fitting parameters.…”

“…In our previous works [11,12], we proposed a design oriented charge-based model for undoped DG MOSFETs under symmetrical operation that aims at giving a comprehensive understanding of the device from the design strategy. By emphasizing the link between this approach and the EKV formalism derived for bulk MOSFET, we introduced useful normalizations for current and charges that in turn lead to very simple relationships among the physical quantities [11].…”

“…Let us emphasize that our algorithm of numerical inversion fully preserves the physics of (1), and therefore its validity is technology-independent [12,16]. Then, noting that the mobile charge density is twice the gate charge density (q m = À2 AE q g ) and assuming that the drift-diffusion transport model is valid, the normalized drain current i can be expressed as:…”

“…However, (1) needs to be solved numerically and this is not desirable for circuit simulation (it requires at least several iterations). To overcome this drawback, we have developed a new methodology to compute without any iteration the mobile charge density as an explicit function of bias voltages (v g and v d or v s ) [12,16] q…”

“…The paper is organized as follows. In Section 2, we briefly describe the core of the explicit long-channel model that has already been developed in our previous works [11,12]. Section 3 details the modeling of small-geometry effects such as: subthreshold swing, threshold voltage roll-off, DIBL and mobility degradation.…”