Andreas Wettstein

Quantum device-simulation with the density-gradient model on unstructured grids

We describe an implementation of the density-gradient device equations which is simple and works in any dimension without imposing additional requirements on the mesh compared to classical simulations. It is therefore applicable to real-world device simulation with complex geometries. We use our implementation to determine the quantum mechanical effects for a MOS-diode, a MOSFET and a double-gated SOI MOSFET. The results are compared to those obtained by a 1D-Schrodinger-Poisson solver. We also investigate a simplified variant of the density-gradient term and show that, while it can reproduce terminal characteristics, it does not give the correct density distribution inside the device.

Random dopant fluctuation modelling with the impedance field method

We discuss an approach for the modelling of random dopant fluctuations based on the impedance field method that has been recently integrated into DESSIS. The method is easy to use and orders of magnitudes more efficient than the statistical method.

Gate Current Calculations Using Spherical Harmonic Expansion of Boltzmann Equation

A physics-based gate current model has been de- veloped based on nonequilibrium electron energy distributions obtained from the spherical harmonic expansion of the Boltz- mann equation. The model accounts for band structure effects, relevant microscopic scattering mechanisms, and electron injec- tions caused by tunneling and thermionic emission processes with parallel momentum conservation and image potential barrier lowering. Obtained distribution functions and injection currents agree well with Monte Carlo simulations and experiments. I. INTRODUCTION Hot carrier injection into the gate oxide in MOSFETs is responsible for gate leakage and oxide degradation, and it has been used in the write operation in NOR flash memories. In order to model the hot carrier injection current, accurate knowledge of the nonequilibrium electron energy distribution is required. Although the Monte Carlo (MC) method would be the most rigorous tool to study hot electron transport (1), the MC method involves large statistical noise in the tail distribution that is important in the gate current calcu- lation. This paper describes a gate current model based on the Spherical Harmonic Expansion (SHE) of the Boltzmann Transport Equation (BTE) (2), (3), (4), (5) that we have implemented in the device simulator Sentaurus Device (6). The implemented SHE model accounts for the full band structure obtained from the empirical pseudopotential method (EPM) (7) and microscopic scattering mechanisms caused by acoustic and intervalley phonons, ionized impurities, and impact ionization. The implemented gate current model covers tunneling and thermionic emission components, and it takes into account parallel momentum conservation, image potential induced barrier lowering, and scattering probability within the image force potential well (1). We validate our model by com- paring obtained distribution functions and gate currents with MC simulations and experiments, and provide a gate current simulation example for a long-channel MOSFET where the hot electron injection is the dominant gate current mechanism.

Investigation of the Statistical Variability of Static Noise Margins of SRAM Cells Using the Statistical Impedance Field Method

The statistical variability of the static noise margin of a six-transistor bulk complementary metal-oxide-semiconductor static random access memory (SRAM) cell due to random doping fluctuations (RDFs) is investigated via 3-D technology computer-aided design simulations. The SRAM cell is created through 3-D process simulations of the entire cell as a single structure. The process flow is based on a typical 32-nm technology. The effects of RDFs on the cell performance are investigated using the highly efficient statistical impedance field method.

Simulations of ultrathin, ultrashort double-gated MOSFETs with the density gradient transport model

We report the results of numerical simulation of nanoscale SOI structures under highly non-equilibrium conditions with the Density Gradient model. The simulations have been carried out with the general purpose device simulator DESSIS. We show that 2D quantum mechanical effects are important for the structures under investigation. We demonstrate that our implementation of the DG model is robust and enables efficient simulation far from equilibrium, for both the drift-diffusion and hydrodynamic transport model.

Integration of the Density Gradient Model into a General Purpose Device Simulator

A generalized Density Gradient model has been implemented into the device simulator Dessis [DESSIS 7.0 reference manual (2001). ISE Integrated Systems Engineering AG, Balgriststrasse 102, CH-8008

The influence of localized states on gate tunnel currents-modeling and simulation

In the numerical simulation of ultra-small MOSFETs with oxide thicknesses in the range 2 to 4 nm, gate leakage currents have to be modeled on a sound physical basis. The main mechanisms apart from oxide non-idealities are direct and resonant tunneling (Fowler-Nordheim tunneling at large biases). Applying a self-consistent simulation of direct tunneling using a fully analytical model, we study the impact of the confinement of carriers in the inversion channel (quasi 2D states) on the size of the direct tunnel current. This is achieved with a Poisson-Schrodinger solver integrated with the device simulator DESSIS-ISE, and by applying Bardeens perturbational method.

On negative differential resistance in hydrodynamic simulation of partially depleted SOI transistors

We show that the negative differential resistance in the I/sub d/-V/sub ds/ characteristics observed in hydrodynamic transport simulations of partially depleted silicon-on-insulator MOSFETs disappears if the nonlocality of tunneling effects are properly accounted for in the recombination-generation process.

Simulation of DGSOI MOSFETs with a Schrodinger-Poisson based mobility model

Ultra-thin DGSOI transistors are considered as one of the most promising devices for future VLSI. Besides expected improvements in the sub-threshold behavior, a theoretical enhancement of the channel mobility was found by some authors. Here, we apply a quantum-mechanical mobility model, based on an integrated Schrodinger/Poisson solver, to double-gate SOI MOSFETs with a range of silicon slab thickness t/sub Si/ and buried-oxide thickness t/sub box/. The main finding is that the theoretical enhancement of effective mobility and on-current at t/sub Si/ /spl ap/10 nm is bound to comparable thicknesses of buried and front oxides. If t/sub box/ /spl ap/100/spl times/t/sub ox/, as e.g. in the case of SIMOX wafers, the volume-inversion related increase of the mobility completely vanishes.

Simulations of quantum transport in HEMT using density gradient model

In this paper, quantum transport simulations for AlGaAs/InGaAs HEMT devices based on the density gradient model are presented. It is shows that size quantization effects have a pronounced influence on the electrical characteristics.