Wanqiang Chen

Hole Mobility and Thermal Velocity Enhancement for Uniaxial Stress in Si up to 4 GPa

A theoretical study of the response of hole mobility and thermal velocity, both relevant for short channel devices, to [110] uniaxial stress in Si up to 4 GPa of both tension and compression has been conducted. The strained-Si bandstructure was calculated using the kmiddotp method. Effective masses, thermal velocities, and scattering rates were calculated from the bandstructure as a function of stress. Mobilities were then calculated via full band Monte Carlo simulations. Calculated mobilities match experimental and theoretical data from prior work addressing lower degrees of stress. Large increases in both carrier thermal velocities and mobilities were found. In the high-stress regime between 1 and 2 GPa, mobilities exhibit a strong superlinear dependence, and compressive stress becomes more favorable for increasing both mobilities and thermal velocities in pMOS. Improvements in both thermal velocity and mobility finally only begin to rolloff toward apparent saturation as we push the stress toward 4 GPa in these simulations

Source-side barrier effects with very high-K dielectrics in 50 nm Si MOSFETs

High permittivity (K) gate insulators are projected for sub-100 nm Si MOSFETs since direct tunneling will likely limit SiO/sub 2/ thicknesses to 1.0-1.5 nm. High-K insulators avoid tunneling, but their larger physical thicknesses introduce subtle capacitive coupling phenomena such as fringing-induced barrier lowering (FIBL) that can compromise off-state leakage. In this study, device simulation examines both on and off-state drain current with very high-K gate insulators and sidewall spacers to reveal new source-side and boundary condition effects. Asymmetric devices help to distinguish the effects. A study of stacked gate insulators demonstrates a 10% increase in drive current achieved with high-K spacers in 50 nm devices.

Schrödinger equation Monte Carlo in two dimensions for simulation of nanoscale metal-oxide-semiconductor field effect transistors

A quantum transport simulator, Schrodinger equation Monte Carlo in two dimensions (SEMC-2D), is presented that provides a rigorous yet reasonably computationally efficient quantum mechanical treatment of real scattering processes within quantum transport simulations of nanoscale metal-oxide-semiconductor field effect transistors (MOSFETs). This work represents an extension of an early version of SEMC for simulating quantum transport and scattering in quasi-one-dimensional device geometries such as encountered in conventional and quantum-cascade lasers. In many respects SEMC is simply a variation on nonequilibrium Green’s function techniques, with scattering as well as carrier injection into the simulation region treated via Monte Carlo techniques. In this regard, SEMC also represents a quantum analog of semiclassical Monte Carlo. Scattering mechanisms considered include crystal momentum randomizing acoustic and optical intra- and intervalley scattering (and intra- and intersubband scattering), and nonrand...

Schrödinger equation Monte Carlo in three dimensions for simulation of carrier transport in three-dimensional nanoscale metal oxide semiconductor field-effect transistors

A quantum transport simulator, Schrodinger equation Monte Carlo (SEMC) in three dimensions, is presented that provides a rigorous yet reasonably computationally efficient quantum mechanical treatment of real scattering processes within quantum transport simulations of nanoscale three-dimensional (3D) metal oxide semiconductor field-effect transistor (MOSFET) geometries such as quantum wire and multigate field-effect transistors. This work represents an extension of earlier versions of SEMC for simulating quantum transport and scattering in systems with relatively simpler quasi-one-dimensional and quasi-two-dimensional geometries such as quantum-cascade lasers (via SEMC in one dimension) and silicon-on-insulator or dual-gate MOSFETs (via SEMC in two dimensions), respectively. However, the limiting computational considerations can be significantly different. The SEMC approach represents a variation in nonequilibrium Green’s function techniques with scattering as well as carrier injection into the simulation...

Two-dimensional quantum mechanical simulation of electron transport in nano-scaled Si-based MOSFETs

Abstract As the transistor dimensions scale down below the 100 nm regime, the reliability of semiclassical models of transport decreases. To offer additional insight into transport phenomena in these deeply scaled devices, simulation tools that treat non-local quantum transport and confinement effects without sacrificing the realistic treatment of scattering are needed. A unique non-equilibrium Greens function approach “Schrodinger equation Monte-Carlo” (SEMC) provides a physically rigorous approach to quantum transport and phase-breaking inelastic scattering via real ( actual ) scattering processes such as optical and acoustic phonon scattering . “One-dimensional” SEMC codes previously have been applied to model transport in systems such as quantum well lasers where the potential varies only along the nominal direction of transport, although with a fully three-dimensional (3D) treatment of scattering. In this paper, the development of a version of “SEMC-2D” code for electrostatically self-consistent treatment of quantum transport within devices with, additionally, quantum confinement normal to the direction of transport, is reported along with illustrative simulation results for nano-scaled SOI MOSFETs geometries.

Quantum Transport Simulation of Carrier Capture and Transport within Tunnel Injection Lasers

Hot electron distributions within the active region of quantum well lasers lead to gain suppression, reduced quantum efficiency, and increased diffusion capacitance, greater low-frequency roll-off and high-frequency chirp. Recently, “tunnel injection lasers” have been developed to minimize electron heating within the active quantum well region by direct injection of cool electrons from the separate confinement region into the lasing subband(s) through a tunneling barrier. Tunnel injection lasers, however, also present a rich physics of transport and scattering, and a correspondingly rich set of challenges to simulation and device optimization. For example, a Golden-Rule-based analysis of the carrier injection into the active region of the ideal tunnel injection laser would suggest approximately uniform injection of electrons among the nominally degenerate ground quantum well states from the separate confinement region states. However, such an analysis ignores (via a random-phase approximation among the final states) the basic real-space transport requirement that injected carriers still must pass through the wells sequentially, coherently or otherwise, with an associated attenuation of the injected current into each subsequent well due to electron-hole recombination in the prior well. Transport among the wells then can be either thermionic, or, of theoretically increasing importance for low temperature carriers, via tunneling. Coherent resonant tunneling between wells, however, is sensitive to the potential drops between wells that split the energies of the lasing subbands and (further) localizes the electron states to individual wells. In this work such transport issues are elucidated using Schrödinger Equation Monte Carlo (SEMC) based quantum transport simulation.

Schrödinger Equation Monte Carlo-3D for simulation of nanoscale MOSFETs

A computational efficient quantum transport simulator, Schrodinger equation Monte Carlo in three dimensions (SEMC-3D), for simulating carrier quantum transport subject to scattering in 3D nanoscale MOSFETs is presented. SEMC-3D self-consistently solves (1) for the 2D-confined eigenstates across the channel as a function of position along the channel, (2) the quasi-1D quantum transport equations for injected carriers propagating through the simulation region within the each subband subject to a rigorous treatment of various intra- and inter-subband and valley scattering processes, and (3) the 3D Poisson equation. The technique, an extension of prior 1D and 2D versions of SEMC but subject to some significantly different computational considerations, is briefly described. SEMC-3D simulations of a Si omega-gate nano-scale nMOSFET are provided to illustrate the modeling capabilities and computational efficiency of SEMC-3D.