Transistor compact model (TCM) is the key bridge between process technology and circuit design. Typically, TCM is desired to capture the nonlinear device electronic characteristics and their high-order derivatives. However, for the novel devices in advanced and future technologies, establishing TCM based on analytical equations and extracting model parameters becomes tedious. The model fitting capability for device outputs’ high-order derivatives is also limited. These drawbacks hinder fast and accurate device to circuit evaluation cycles. We develop a TCM based on multigradient neural network (MNN) using computational graph in the PyTorch framework. This MNN model is able to simultaneously capture the transistor dc/ac characteristics, such as ${I}-{V}/{Q}-{V}$ , their derivatives ( ${G}-{V}/{C}-{V}$ ), and higher order derivatives accurately. Moreover, the model architecture can be widely adapted to various device types. Based on this model scheme, software is developed to enable the automated model generation for standard SPICE simulation. Finally, the model and software are validated for novel gate-all-around (GAA) Si cold source field-effect transistors (CSFET), and 19-stage ring oscillator and two-stage operational amplifier circuit simulations have also been demonstrated. This work reduces the cycle of novel device compact model creation and circuit benchmark simulation from months or weeks to hours. In addition, it enables more precise circuit simulation for analog and RF circuits, and it provides a rapid solution for early stage design technology cooptimization (DTCO).
4 February 2022 in IEEE Transactions on Electron Devices
The complementary field-effect transistor (CFET) with stacked N-type FET (NFET) and P-type FET (PFET) is an attractive approach to shrink the footprint of multiple devices at circuit level and increase transistor density. Compared with traditional device structure, the unique geometry of CFET brings very different parasitics. In this work, we take the inverter as an exemplar CFET circuit building block and provide analytical compact models of calculating parasitic capacitance, which is critical to enable fast and accurate simulations at circuit level. The validity of the models is calibrated and verified with 3-D numerical simulations. A comparison between the CFET inverter with its same ground-rule gate-all-around FET (GAAFET) counterpart reveals the critical CFET design parameters and demonstrates performance advantage of CFET architecture.
Hafnium dioxide is a promising ferroelectric (FE) material for achieving high-density nonvolatile memory and neuromorphic computing due to its compatibility with the mainstream integrated circuit technology and the surprisingly enhanced ferroelectricity by reduced thickness. The FE switching dynamics is essential to the device performance, but the complexity of atomic structure causes unknown of various FE switching paths and domain-wall configurations. Here, we demonstrate that its low-barrier paths and domain walls can be comprehensively found and understood from a perspective of topological symmetry. By discussing pseudochirality and equivalent transformation relations in a crystal with first principles and lattice modes, we classify and analyze 4 low-barrier FE switching paths and 93 irreducible topology domain-wall configurations in . The anisotropic switching mechanism is inferred based on the investigation for 12 types of domain walls. This methodology is expected to be generally applicable to displacive ferroelectrics with low unit-cell point-group symmetries and lay a foundation for mechanism study of the switching dynamics.
As the architecture of logic devices is evolving towards gate-all-around (GAA) structure, research efforts on advanced transistors are increasingly desired. In order to rapidly perform accurate compact modeling for these ultra-scaled transistors with the capability to cover dimensional variations, neural networks are considered. In this paper, a compact model generation methodology based on artificial neural network (ANN) is developed for GAA nanosheet FETs (NSFETs) at advanced technology nodes. The DC and AC characteristics of GAA NSFETs with various physical gate lengths (Lg), nanosheet widths (Wsh) and thicknesses (Tsh), as well as different gate voltages (Vgs) and drain voltages (Vds) are obtained through TCAD simulations. Subsequently, a high-precision ANN model architecture is evaluated. A systematical study on the impacts of ANN size, activation function, learning rate, and epoch (the times of complete pass through the entire training dataset) on the accuracy of ANN models is conducted, and a shallow neural network configuration for generating optimal ANN models is proposed. The results clearly show that the optimized ANN model can reproduce the DC and AC characteristics of NSFETs very accurately with a fitting error (MSE) of 0.01.
5 February 2024 in IEEE Transactions on Electron Devices
In this article, we evaluate various crystal orientation configurations for silicon ultrascaled nanosheet FETs (NSFETs) to explore the optimum combination of wafer surface and channel orientations. The increasingly prominent physical phenomena, including quantum confinement and quasi-ballistic transport, are captured physically by advanced simulation methodology. The carrier density profile and transport-related parameters exhibit a strong dependence on the crystallographic orientations of the channel and the wafer directions. The quantum confinement effects are least pronounced in the n-type NSFETs with a (100) surface, and the p-type NSFETs with a (110) surface. The $ \langle \text{100} \rangle$ and $ \langle \text{111} \rangle$ channel orientations demonstrate the largest ballistic injection velocity for electrons and holes, respectively. Uniaxial stress technology, as an efficient performance booster, may further improve the transport properties, but the enhancement may saturate at high stress level. In the ballistic limit, we predict that the homo-oriented CMOS inverter of $\text{(}\text{110}\text{)}/\langle \text{111} \rangle$ configuration has a more balanced N/P current, and the hetero-oriented CMOS inverter, comprising a 5 ${\times}$ 10 nm n-type NSFET of $\text{(}\text{100}\text{)}/\langle \text{100} \rangle$ and a 5 ${\times}$ 10 nm p-type NSFET of $\text{(}\text{110}\text{)}/\langle \text{111} \rangle$ , will yield a speed enhancement exceeding 20% when compared with the prevailing industry standard of the $\text{(}\text{100}\text{)}/\langle \text{110} \rangle$ configuration.
27 February 2023 in IEEE Transactions on Electron Devices
Threshold voltage ( ${V_{t}}$ ) engineering is critical and challenging in the advanced logic devices. Inducing interface dipoles by the incorporation of dopants into the gate dielectrics is an emerging scheme to modulate the threshold voltage. However, the impact of those dipoles on gate-stack performance is of great concern. In this work, we present the first-principles calculations of dielectric constant and tunneling current of high- ${k}$ gate-stack with lanthanum (La) dopants. It is found that incorporation of La improves the local dielectric constant as well as the equivalent oxide thickness (EOT), but degrades the gate leakage current. The calculated results are consistent with the reported experimental data. Furthermore, by analyzing the transmission coefficients and band alignment across the gate-stack, a theoretical explanation for the change of gate leakage current is provided.