Skip to main navigation Skip to search Skip to main content

Accurate conjunction of yield models for fault-tolerant memory integrated circuits

  • Chunghun Ha
  • , Way Kuo
  • , Jung Yoon Hwang

Research output: Journal Publications and ReviewsRGC 21 - Publication in refereed journalpeer-review

Abstract

Critical defects, i.e., faults, inevitably occur during semiconductor fabrication, and they significantly reduce both manufacturing yield and product reliability. To decrease the effects of the defects, several fault-tolerance methods, such as the redundancy technique and the error correcting code (ECC), have been successfully applied to memory integrated circuits. In the semiconductor business, accurate estimation of yield and reliability is very important for determining the chip architecture as well as the production plan. However, a simple conjunction of previous fault-tolerant yield models tends to underestimate the manufacturing yield if several fault-tolerance techniques are employed simultaneously. This paper concentrates on developing and verifying an accurate yield model which can be applied successfully in such situations. The proposed conjunction model has been derived from the probability of remaining redundancies and the average number of defects after repairing the defects with the remaining redundancies. The validity of the conjunction yield model is verified by a Monte Carlo simulation. © 2006 IEEE.
Original languageEnglish
Article number5159401
Pages (from-to)344-350
JournalIEEE Transactions on Semiconductor Manufacturing
Volume22
Issue number3
DOIs
Publication statusPublished - 2009

UN SDGs

This output contributes to the following UN Sustainable Development Goals (SDGs)

  1. SDG 9 - Industry, Innovation, and Infrastructure
    SDG 9 Industry, Innovation, and Infrastructure

Research Keywords

  • Fault-tolerance
  • Monte Carlo simulation
  • Semiconductor manufacturing
  • Yield modeling

Fingerprint

Dive into the research topics of 'Accurate conjunction of yield models for fault-tolerant memory integrated circuits'. Together they form a unique fingerprint.

Cite this