Doping Profile Optimization for Power Devices Using Topology Optimization

Katsuya Nomura*, Tsuguo Kondoh, Tsuyoshi Ishikawa, Shintaro Yamasaki, Kentaro Yaji, Kikuo Fujita

*Corresponding author for this work

Research output: Contribution to journalArticlepeer-review

Abstract

This paper proposes an optimization method with a high degree of freedom for the doping profile optimization of power devices. In this method, which is based on topology optimization, the design domain is divided into many elements, and the doping or dose densities of all of the elements are taken as the design variables to be optimized. An optimal structure is obtained by repeatedly modifying a given structure using mathematical programming. The proposed optimization method was applied to two design problems. First, the tradeoff between the breakdown voltage and ON-resistance was optimized for a p-n diode by using a function of the electric field strength as the objective function under a constraint to the ON-resistance. Second, for a junction termination extension structure, a robust optimization method considering the ion implantation process was applied to prevent the concentration of the electric field by dose fluctuation. The numerical results from these two problems demonstrate that the proposed method can be effectively used to derive an improved doping profile for power devices.

Original languageEnglish
Article number8418456
Pages (from-to)3869-3877
Number of pages9
JournalIEEE Transactions on Electron Devices
Volume65
Issue number9
DOIs
Publication statusPublished - 2018 Sept
Externally publishedYes

Keywords

  • Doping profile optimization
  • ion implantation
  • power device
  • robust optimization
  • topology optimization
  • tradeoff between breakdown voltage and ON-resistance

ASJC Scopus subject areas

  • Electronic, Optical and Magnetic Materials
  • Electrical and Electronic Engineering

Fingerprint

Dive into the research topics of 'Doping Profile Optimization for Power Devices Using Topology Optimization'. Together they form a unique fingerprint.

Cite this