7. Quantum Atomistic Simulations of Nanoelectronic Devices Using QuADS

As semiconductor devices shrink into the nanoscale regime and new classes of nanodevices emerge, device performance is increasingly being dominated by the granularity in the underlying material and the quantum mechanical effects in the electronic states. At nanoscale, modeling and simulation approaches based on a continuum representation of the underlying material typically used by device engineers become invalid. On the other side, various ab initio materials science methods offer intellectual appeal, but can only model very small systems having ∼ 100 atoms. The variety of geometries, materials, and doping configurations in semiconductor devices at the nanoscale suggests that a general nanoelectronic modeling tool is needed. This paper describes our on-going efforts to develop a multiscale Quantum Atomistic Device Simulator (QuADS) to address these needs. QuADS bridges the gap (and crosses the intellectual boundary) between continuum and ab initio modeling paradigms and enable the quantum-corrected atomistic numerical modeling of non-equilibrium charge and phonon transport phenomena in realistically-sized systems containing more than 100 million atoms! QuADS is primarily being built upon extended versions of three modules: (a) Open source LAMMPS molecular dynamics code for geometry construction and modeling structural relaxations. To enhance accuracy, ab initio ABINIT tool is used for parameterization of force and polarization coefficients and model bandstructure calculations; (b) Open source NEMO 3-D tool, which employs a variety of tight-binding models (s, sp³s^∗, sp³d⁵s^∗), for the calculation of excitonic and phonon spectra and optical transition rates; and (c) A quantum-corrected (benchmarked against the non-equilibrium Green function formalism) 3-D Monte Carlo electron–phonon transport kernel. Using QuADS, nanoelectronic device designers will be able to address many challenging issues including crystal atomicity, defects, interfaces and surfaces, strain relaxation, piezoelectric and pyroelectric polarization, quantum confinement, highly-interacting and dissipative current and phonon paths, and performance in harsh environments – all on an equal footing. With the multi-million atom handling capability, the simulator creates new engineering routes for optimizing the efficiency and reliability of nanoelectronic and optoelectronic devices that were previously infeasible. Successful applications of QuADS are demonstrated by three examples: (1) Effects of internal fields in InN/GaN quantum dots; (2) Importance of second order polarization in InAs/GaAs quantum dots; and (3) Modeling unintentional single charge effects in silicon nanowire FETs. QuADS uses several novel, memory-miserly, parallel and fast algorithms, and incorporates state-of-the-art fault-tolerant software design approaches, which enables the simulator to assess the reliability of available petaflop computing platforms (TeraGrid, NCCS, NICS). A web-based online interactive version for educational purposes will soon be available on http://​www.​nanoHUB.​org