Raul Valin

Implementation of the Density Gradient Quantum Corrections for 3-D Simulations of Multigate Nanoscaled Transistors

An efficient implementation of the density-gradient (DG) approach for the finite element and finite difference methods and its application in drift-diffusion (D-D) simulations is described in detail. The new, second-order differential (SOD) scheme is compatible with relatively coarse grids even for large density variations thus applicable to device simulations with complex 3-D geometries. Test simulations of a 1-D metal-oxide semiconductor diode demonstrate that the DG approach discretized using our SOD scheme can be accurately calibrated against Schrödinger-Poisson calculations exhibiting lower discretization error than the previous schemes when using coarse grids and the same results for very fine meshes. 3-D test D-D simulations using the finite element method are performed on two devices: a 10 nm gate length double gate metal-oxide-semiconductor field-effect transistor (MOSFET) and a 40 nm gate length Tri-Gate fin field-effect transistor (FinFET). In 3-D D-D simulations, the SOD scheme is able to converge to physical solutions at high voltages even if the previous schemes fail when using the same mesh and equivalent conditions. The quantum corrected D-D simulations using the SOD scheme also converge with an atomistic mesh used for the 10 nm double gate MOSFET saving computational resources and can be accurately calibrated against the results from non-equilibrium Greens functions approach. Finally, the simulated ID-VG characteristics for the 40 nm gate length Tri-Gate are in an excellent agreement with experimental data.

3-D Finite Element Monte Carlo Simulations of Scaled Si SOI FinFET With Different Cross Sections

Nanoscaled Si SOI FinFETs with gate lengths of 12.8 and 10.7 nm are simulated using 3-D finite element Monte Carlo (MC) simulations with 2-D Schrodinger-based quantum corrections. These nonplanar transistors are studied for two cross sections: rectangular-like and triangular-like, and for two channel orientations: (100) and (110). The 10.7-nm gate length rectangular-like FinFET is also simulated using the 3-D nonequilibrium Greens functions (NEGF) technique and the results are compared with MC simulations. The 12.8 and 10.7 nm gate length rectangular-like FinFETs give larger drive currents per perimeter by about 33- 37% than the triangular-like shaped but are outperformed by the triangular-like ones when normalised by channel area. The devices with a (100) channel orientation deliver a larger drive current by about 11% more than their counterparts with a (110) channel when scaled to 12.8 nm and to 10.7 nm gate lengths. ID - VG characteristics obtained from the 3-D NEGF simulations show a remarkable agreement with the MC results at low drain bias. At a high drain bias, the NEGF overestimates the on-current from about VG - VT = 0.3 V because the NEGF simulations do not include the scattering with interface roughness and ionized impurities.

Two-Dimensional Monte Carlo Simulation of DGSOI MOSFET Misalignment

In this paper, we investigate the gate misalignment effects in a 10-nm double-gate silicon-on-insulator MOSFET transistor with a 2-D Monte Carlo simulator. Quantum effects, which are of special relevance in such devices, are taken into account by using the multivalley effective conduction-band-edge method. Different gate misalignment configurations have been considered to study the impact on device performance, finding a current improvement when the gate misalignment increases the source-gate overlapping. Moreover, our results show that a 20% gate misalignment can be assumed for a drain current deviation smaller than 10%. Finally, the validity of the obtained results was assessed with a set of simulations for devices that have different gate lengths, silicon thicknesses, and oxide thicknesses.

Optimisation and parallelisation of a 2D MOSFET multi-subband ensemble Monte Carlo simulator

Multi-subband ensemble Monte Carlo simulators (MSB-EMC) are essential in semiconductor device modelling in order to study confinement effects when devices are scaled to gate lengths and silicon thicknesses approaching nanometre dimensions. To consider these effects it is necessary to solve the Schrödinger equation in the confinement direction, which has a high computational cost. In this paper we propose the parallelisation and optimisation of a 2D MSB-EMC simulator. The Open Multi-Processing (OpenMP) application program interface (API) is employed for the parallelisation and the simulations are performed using a multi-core machine. Numerical results show a speed-up of 7.40 when eight cores are used with a parallel efficiency higher than 90%.

Quantum transport of a nanowire field-effect transistor with complex phonon self–energy

In this work, the impact of the real part of the phonon self-energy on the transfer characteristics of a silicon nanowire transistor is investigated. The physical effects of the real part of the self-energy are to create a broadening and a shift on the density of states. This increases the drain current in the sub–threshold region and decreases it in the above–the–threshold region. In the first region, the current is increased as a result of an increase of charge in the middle of the channel. In the second one, the electrostatic self–consistency or the enforcement of charge neutrality in the channel reduces the current because a substantial amount of electrons are under the first subband and have imaginary wave vectors. The change in the phonon–limited mobility due to the real part of self–energy is evaluated for a nanowire transistor and a nanowire in which there is not source to drain barrier. We also assess the validity of Mathiessens rule using the self–consistent NEGF simulations and the Kubo–Greenwood formalism.

Performance of the CloudStack KVM Pod Primary Storage under NFS Version 3

Currently, there are an increasing number of open-source solutions for building Clouds. The performance of the Virtual Machines running in these Clouds is a key point. The way the Virtual Machines are created in Clouds can have important effects upon their disk I/O operations. We have selected CloudStack platform to study the disk I/O performance for KVM Virtual Machines.

Study of Local Power Dissipation in Ultrascaled Silicon Nanowire FETs

The local electron power dissipation has been calculated in a field-effect nanowire transistor using a quantum transport formalism. Two different channel cross sections and optical and acoustic phonon mechanisms were considered. The phonon models used reproduce the phonon limited mobility in the cross sections studied. The power dissipation for different combinations of source, channel, and drain dimensions have been calculated. Due to the lack of complete electron energy relaxation inside the device, the Joule heat dissipation over-estimates the power dissipated in small nanotransistors. This over-estimation is larger for large cross sections due to the weaker phonon scattering. On the other hand, in narrow wires, the power dissipation inside the device can be large, therefore, mitigating against fabrication of very narrow nanowire transistors. We have also investigated the cooling of the device source region due to the mismatch of the Peltier coefficients between the source and the channel.

Impact of different electron-phonon scattering models on the electron transport in a quantum wire

The non-equilibrium Greens function (NEGF) approach has been widely used to describe quantum transport in nanostructures. In general, the Electron-phonon (e-ph) interaction is regarded to be local in space, as transport simulations using non-local e-ph interactions are very computationally intensive. As such non-local interactions are rarely used to simulate nanotransistors. In this work, non-local e-ph interactions in the self-consistent Born approximation (SCBA) are used to simulate transport through a quantum wire. The non-local model is compared with the results for local, nearest-neighbour and second nearest- neighbour interactions. It was found that for a 40nm quantum wire with a dephasing factor of D0 = 0.001 eV2 and an applied potential of 0.1V, the local model underestimates the current by 20%. Furthermore, to study electron transport in Graphene and polymers, non-local e-ph interactions must be considered.

Causal self-energies for NEGF modelling of quantum nanowires

Electron-phonon scattering is shown to increase dramatically at small nanowire cross-sections so that the transport is not ballistic. Non-ballistic dissipative device modelling requires the full complexity of the non-equilibrium Green Function (NEGF) method. The role of causality in obtaining spectral sum rule-conserving approximations to the electron-phonon self-energies is demonstrated and applications given for wrap-round gate silicon nanowire field effect transistors. Causality violations give erroneous density of states and current densities.

An e-Science infrastructure for nanoeletronic simulations based on grid and cloud technologies

We propose the development of a new e-Science infrastructure that would take the best of both grid and cloud technologies, and it would allow different research groups that perform nanoelectronic simulations to share their local clusters and create a common infrastructure accessible through a unified point of access. Therefore, more computational power can be used to perform nanoelectronic simulations, with the consequent reduction of time required to obtain the results. The integration of local clusters to share resources, through the proposed cloud management stack, will allow deploying an elastic infrastructure that will also permit to prioritize local computing tasks over shared ones. Furthermore, it will allow not only the deployment of ad-hoc virtual machines across local sites to achieve specific tasks but also to deploy virtual machines in public clouds like Amazon AWS to get additional computing resources, and even avoiding data losing by using public storage clouds like Amazon S3.