12 June 2023 in IEEE Transactions on Computer-Aided Design of Integrated Circuits and Systems
This work proposes a statistical modeling approach for the artificial neural network (ANN)-based compact model (CM). The method of retaining part of the network features of the nominal device and further finetuning the network parameters (variational neurons) is found to accurately reproduce the static variation. A mapping from process variation to network parameters is derived by combining the proposed variational neuron selection algorithm and the backward propagation of variance (BPV) method. In addition, a secondary classification of the selected variational neurons is applied to model the fabrication-induced correlation between n- and p-type devices. The neural network-based statistical modeling approach has been well implemented and verified on the GAA simulation data and the 16nm node foundry FinFET, which indicates its great potential in modeling emerging and advanced device technology.
12 June 2023 in IEEE Transactions on Electron Devices
A physics-based SPICE model for dopant segregated tunneling field-effect transistors (DS-TFETs) with gate-to-drain underlap is developed for the design technology co-optimization (DTCO) of TFETs with the foundry process. It captures all the current components in full operation regions, including band-to-band tunneling (BTBT), forward p-i-n diode, gate leakage, and also ambipolar conduction. The channel surface potential is formulated first, with which the current expressions are derived and verified with experiments calibrated TCAD simulations. Model predictions of current, conductance, and voltage transfer characteristic (VTC) are in good agreement with measurement data of hardware devices (nTFET and pTFET) as well as circuits. This model has been written in Verilog-A language and employed in the TFET circuits design.
28 June 2023 in IEEE Electron Device Letters
A generic compact modeling approach for emerging memories is developed in this work. State transitions, as well as the electric properties of each state, usually require separate efforts from a physics perspective. To accelerate memory modeling, a data-driven approach is explored with the recurrent Non-linear Auto Regressive with Exogenous inputs (NARX) network. The recurrent module captures the state transition dynamics, while a feedforward module is responsible for the non-linear states, such as high and low resistance states. The generic NARX-based model is demonstrated with high accuracy for emerging memories of different physics, such as resistive random-access-memory (ReRAM) and ferroelectric tunnel junction (FTJ). It is extendable to statistical modeling which is useful to cover memory variations. Circuit simulations are demonstrated with successful convergence, attributed to the capability of the NARX network and its compatibility with simulators.
28 May 2024 in IEEE Transactions on Electron Devices
An accurate physics-based capacitance model is developed covering full-region operations of silicon TFET (Si-TFET) with all biasing conditions. The intrinsic and parasitic capacitances are modeled by the surface-potential-based method and justified by calibrated TCAD simulations. To address the challenges of a classical physics-based model in calculating the charge partitions, an artificial neural network (ANN)-enhanced method is proposed to capture the terminal partition strategy, which features good smoothness and high accuracy. The results show a source terminal partition of the ambipolar region in the range of 90%–100%, and the drain partition of the normal region is in the range of 85%–100%, with in-between partitions of the transition regions. This ANN-enhanced physics-based model addresses the requirements of a TFET channel charge partition strategy and has been verified in circuit simulations with improved accuracy.