Random-Dopant-Induced Drain Current Variation in Nano-MOSFETs: A Three-Dimensional Self-Consistent Monte Carlo Simulation Study Using “Ab Initio” Ionized Impurity Scattering

Dept. of Electron. & Electr. Eng., Glasgow Univ., Glasgow
IEEE Transactions on Electron Devices (Impact Factor: 2.06). 12/2008; DOI: 10.1109/TED.2008.2004647
Source: IEEE Xplore

ABSTRACT A comprehensive simulation study of random-dopant-induced drain current variability is presented for a series of well-scaled n -channel MOSFETs representative of the 90-, 65-, 45-, 35-, and 22-nm technology nodes. Simulations are performed at low and high drain biases using both 3-D drift diffusion (DD) and 3-D Monte Carlo (MC). The ensemble MC simulator incorporates an ldquo ab initio rdquo treatment of ionized impurity scattering through the real-space trajectories of the carriers in the Coulomb potential of the random discrete impurities. When compared with DD simulations, the MC simulations reveal a significant increase in the drain current variability as a result of additional transport variations due to position-dependent Coulomb scattering that is not captured within the DD mobility model. Such transport variations are in addition to the electrostatic variation in carrier density that is alone captured within the DD approach. Through comparison of the DD and MC results, we estimate the relative importance of electrostatic and transport-induced variability at different drain bias conditions.

  • [Show abstract] [Hide abstract]
    ABSTRACT: Remote charge scattering (RCS) has become a serious obstacle inhabiting the performance of ultrathin gate oxide MOSFETs. In this paper, we evaluate the impact of RCS by treating the real-space full Coulomb interaction between remote charges and inversion carriers. A new approach that can be simply incorporated in ensemble Monte Carlo (EMC) based simulations without any variation of the standard EMC simulator is developed. The charge-carrier (c-c) interaction model is based on a particle-mesh (PM) calculation method which resolves both the long-range and short-range Coulomb interactions by solving Poisson’s equation on a re-fined mesh. The validity of our approach is verified by three-dimensional (3-D) resistor simulations, from which the obtained doping dependence of the low-field mobility agrees well with experimental results. The proposed approach is then used to study the impact of RCS on the drive current and carrier transport properties in the channel of a 20 nm silicon (Si) nMOS FinFET with HfO2 gate stack. We find that the influence of RCS is strongly localized in the vicinity of the remote charges and exhibits a granular nature, indicating the necessity to consider the full Coulomb interaction in RCS.
    Sciece China. Information Sciences · 0.71 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: With the scaling of field-effect transistors to the nanometre scale, it is well recognised that TCAD simulations of such devices need to account for quantum mechanical confinement effects. The most widely used method to incorporate quantum effects within classical and semi-classical simulators is via density gradient quantum corrections. Here we present our methodologies for including the density gradient method within our Drift-Diffusion and Monte Carlo simulators and highlight some of the additional benefits that this provides when dealing with the charge associated with random discrete dopants. KeywordsDensity gradient-Quantum corrections-Drift-Diffusion-Monte Carlo-Simulation-MOSFETs
    Journal of Computational Electronics 9(3):187-196. · 1.01 Impact Factor
  • [Show abstract] [Hide abstract]
    ABSTRACT: We present a compact, surface-potential-based modeling approach for deeply scaled digital or radio-frequency metal-oxide-semiconductor field-effect transistor able to account for random doping fluctuations in the device channel. Random dopant fluctuations are one of the primary causes for device variability in nanometer-scale components. The present approach is based on the Green's function formulation of the device external electrical parameters (such as the output current) small-change sensitivity to distributed, space-dependent doping variations in the channel; furthermore, the methodology is used also to assess the small-signal device parameter variations within the limits of a quasi-static description. The present approach allows for an efficient circuit-level sensitivity analysis and has been applied to the PSP compact model through a Verilog-A code implemented within the ADVANCED DESIGN SYSTEM (Agilent Technologies, Santa Clara, CA, U.S.A.) simulator. Examples are provided to show that the model predictions are in good agreement with far more time-consuming simulations. Copyright © 2013 John Wiley & Sons, Ltd.
    International Journal of Numerical Modelling Electronic Networks Devices and Fields 12/2013; · 0.54 Impact Factor