2007Unpublished venueRequires access

A fast semi-numerical technique for the solution of the poisson-boltzmann equation in a cylindrical nanowire

Ashok T. Ramu, Manjeri P. Anantram, Kaustav Banerjee

Open publisher page 1 citations

Abstract

Silicon nanowire (SiNW) based devices have aroused great interest since they exhibit high carrier mobilities and sub-threshold slopes close to 69 mV/decade due to good charge control. The charge in the nanowire channel is obtained by solving the Poisson equation self-consistently with the Schroedinger equation. However this is a very computationally expensive process and it is usually preferable to solve for the potential to the first order using only the Poisson equation, and then to make a quantum correction for the charge based on the potential gradient at the surface of the wire. It is the former task, that of solving the non-linear Poisson's equation within the nanowire, that we address here. The built-in non-linear ODE solver in MATLAB does not converge in most cases because the equation is highly non-linear. Commercial numerical solvers like DESSIS are very accurate but slow. Depending on the grid size, they may take several hours to run.

About this research paper

What this paper is about

Silicon nanowire (SiNW) based devices have aroused great interest since they exhibit high carrier mobilities and sub-threshold slopes close to 69 mV/decade due to good charge control. The charge in the nanowire channel is obtained by solving the Poisson equation self-consistently with the Schroedinger equation. However this is a very computationally expensive process and it is usually preferable to solve for the potential to the first order using only the Poisson equation, and then to make a quantum correction for the charge based on the potential gradient at the surface of the wire. It is the former task, that of solving the non-linear Poisson's equation within the nanowire, that we address here. The built-in non-linear ODE solver in MATLAB does not converge in most cases because the equation is highly non-linear. Commercial numerical solvers like DESSIS are very accurate but slow. Depending on the grid size, they may take several hours to run.

Why it matters

OpenAlex reports 1 citations for this work. Citation counts describe recorded attention and do not establish research quality.

Key contribution

A contribution statement is not available in the OpenAlex record.

Method / approach

Method details are not available in the OpenAlex metadata.

Main findings

Findings are not separately available in the OpenAlex metadata.

Limitations

Limitations are not available in the OpenAlex metadata.

Applications

Application details are not available in the OpenAlex metadata.

Available abstract

Silicon nanowire (SiNW) based devices have aroused great interest since they exhibit high carrier mobilities and sub-threshold slopes close to 69 mV/decade due to good charge control. The charge in the nanowire channel is obtained by solving the Poisson equation self-consistently with the Schroedinger equation. However this is a very computationally expensive process and it is usually preferable to solve for the potential to the first order using only the Poisson equation, and then to make a quantum correction for the charge based on the potential gradient at the surface of the wire. It is the former task, that of solving the non-linear Poisson's equation within the nanowire, that we address here. The built-in non-linear ODE solver in MATLAB does not converge in most cases because the equation is highly non-linear. Commercial numerical solvers like DESSIS are very accurate but slow. Depending on the grid size, they may take several hours to run.

Key concepts: Poisson's equation, Nanowire, Poisson–Boltzmann equation, Solver, Ode, Discrete Poisson equation, Charge (physics), Boltzmann equation

Related papers

Back to paper searchBrowse research topicsOriginal source
A fast semi-numerical technique for the solution of the poisson-boltzmann equation in a cylindrical nanowire — Research Paper | ScholarLens