Random-Dopant-Induced Drain Current Variation in Nano-MOSFETs: A Three-Dimensional Self-Consistent Monte Carlo Simulation Study Using “Ab Initio” Ionized Impurity Scattering
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.
- SourceAvailable from: Mihail Nedjalkov[Show abstract] [Hide abstract]
ABSTRACT: We present a numerical study of the evolution of a wave packet in a nanoscale MOSFET featuring an ‘atomistic’ channel doping. Our two-dimensional Monte Carlo Wigner simulation results are compared against classical Boltzmann simulation results. We show that the quantum effects due to the presence of a scattering center are manifestly non-local affecting the wave propagation much farther than the geometric limit of the center. In particular the part of the channel close to the oxide interface remains blocked for transport, in contrast to the behavior predicted by classical Boltzmann propagation.Physica A: Statistical Mechanics and its Applications 01/2014; 398:194–198. · 1.68 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
- [Show abstract] [Hide abstract]
ABSTRACT: An efficient and accurate method to include surface roughness scattering from a general, realistic synthesized surface in 3D Monte Carlo simulation is presented with verification. The method is then applied to study drain current variation due to variation in surface roughness scattering in an 18nm bulk Silicon nMOSFET, highlighting substantially increased variation at low drain bias compared with electrostatic drift diffusion simulation.01/2011;