Statistical Monitoring Techniques for Contamination Data

Douglas Montgomery, J. Bert Keats, John Fowler, George Runger, Geetha Rajavelu

Research output: Contribution to journalArticle

Abstract

Monitoring of particle contamination by statistically based methods is used extensively in semiconductor manufacturing. Often Shewhart-type control charts are used for this purpose. These charts suffer from some potential disadvantages in this application environment. Specifically, they are slow to detect shifts or changes in the level of particle contamination and, in some situations, the underlying statistical assumptions of the charts are inappropriate for contamination data. We suggest the use of cumulative sum and exponentially weighted moving average control charts for monitoring particle contamination. The advantages of these charts for the types of particle contamination data typically encountered in semiconductor manufacturing are discussed and illustrated with examples.

Original languageEnglish (US)
Pages (from-to)23-30
Number of pages8
JournalJournal of the IEST
Volume40
Issue number2
StatePublished - Mar 1 1997

    Fingerprint

Keywords

  • Cumulative sum (CUSUM)
  • Exponentially weighted moving average (EWMA)
  • Particle contamination
  • Statistical monitoring

ASJC Scopus subject areas

  • Environmental Engineering
  • Environmental Chemistry
  • Safety, Risk, Reliability and Quality

Cite this