Comparison of Monte Carlo and NEGF Simulations of Double Gate MOSFETs

Abstract

The present work compares the simulation results of the two-dimensional full band Monte Carlo simulator (MOCA) developed at the University of Illinois at Urbana-Champaign and the two-dimensional quantum simulator (NanoMOS) developed at Purdue University. Double-gate MOSFETs of three body thicknesses -t Si = 4, 3 and 2 nm-were considered in this study. For a body thickness of 4 nm, the conduction band profiles and sheet charge densities obtained from MOCA and NanoMOS almost overlap, particularly for high gate and drain-to-source biases. However, as the body thickness is reduced, quantum effects are captured more naturally in NanoMOS since MOCA only uses a simple quantum correction scheme, with an otherwise semi-classical model. However, even for thinner devices, since MOCA makes use of a detailed band structure and scattering model, high energy transport is better reproduced by the Monte Carlo procedure. A particle description is appealing for nanoscale simulation, in order to reproduce the granularity aspects of the transport. Comparisons with a quantum model based on continuum flow equations as in NanoMOS should provide valuable insight to better incorporate the quantum mechanical aspects in a practical particle-based model.

Monte Carlo simulation of nanoelectronic devices

“…Between these extreme approaches, Ensemble Monte Carlo (EMC) simulators are widely used since they present several advantages compared to full quantum approximations. A reduced computational cost, the possibility of considering a wide variety of scattering mechanisms and high accuracy for devices with silicon thicknesses as low as a few nanometers [48] are some of the advantages of such simulators. To include quantum effects in EMC codes, two are the main solutions proposed in the bibliography:…”

Quantum Ensemble Monte Carlo simulation of silicon-based nanodevices

“…It is possible to include quantum effects in nano-transistors without solving the Schrödinger equation by adding a correction term to the electrostatic potential [1,2]. Thus, the carrier density given by the full quantum solution can be reproduced.…”

MONTE-CARLO SIMULATION OF ULTRA-THIN FILM SILICON-ON-INSULATOR MOSFETs

“…A reduced computational cost, the possibility of considering a wide variety of scattering mechanisms and high accuracy for devices with silicon thicknesses as low as a few nanometers 19 are some of the advantages of such simulators. As mentioned above to include quantum effects in EMC codes, two are the main solutions proposed in the literature:…”

“…The calibration of such corrections generally needs a set of fitting parameters (e.g. carrier effective mass in the density gradient model) but the results obtained are still accurate from the transport point of view for devices with silicon thicknesses as low as a few nanometers 19 . The most commonly used approaches include the Density Gradient 20 , the Effective Potential 21 , or the Multi-Valley Effective Conduction Band-Edge method (MV-ECBE) 22 .…”