Multivariate statistical methods for modeling and analysis of wafer probe test data

Katina R. Skinner, Douglas Montgomery, George Runger, John Fowler, Daniel R. McCarville, Teri Reed Rhoads, James D. Stanley

Research output: Contribution to journalArticle

39 Citations (Scopus)

Abstract

Probe testing following wafer fabrication can produce extremely large amounts of data, which is often used to inspect a final product to determine if the product meets specifications. This data can be further utilized in studying the effects of the wafer fabrication process on the quality or yield of the wafers. Relationships among the parameters may provide valuable process information that can improve future production. This paper compares many methods of using the probe test data to determine the cause of low yield wafers. The methods discussed include two classes of traditional multivariate statistical methods, clustering and principal component methods and regression-based methods. These traditional methods are compared to a classification and regression tree (CART) method. The results for each method are presented. CART adequately fits the data and provides a "recipe" for avoiding low yield wafers and because CART is distribution-free there are no assumptions about the distributional properties of the data. CART is strongly recommended for analyzing wafer probe data.

Original languageEnglish (US)
Pages (from-to)523-530
Number of pages8
JournalIEEE Transactions on Semiconductor Manufacturing
Volume15
Issue number4
DOIs
StatePublished - Nov 2002

Fingerprint

Statistical methods
wafers
regression analysis
probes
Fabrication
fabrication
products
Specifications
specifications
Testing
causes

Keywords

  • CART
  • Multivariate statistical methods
  • Tree regression
  • Yield analysis

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Industrial and Manufacturing Engineering
  • Electronic, Optical and Magnetic Materials
  • Physics and Astronomy (miscellaneous)
  • Condensed Matter Physics

Cite this

Multivariate statistical methods for modeling and analysis of wafer probe test data. / Skinner, Katina R.; Montgomery, Douglas; Runger, George; Fowler, John; McCarville, Daniel R.; Rhoads, Teri Reed; Stanley, James D.

In: IEEE Transactions on Semiconductor Manufacturing, Vol. 15, No. 4, 11.2002, p. 523-530.

Research output: Contribution to journalArticle

Skinner, Katina R. ; Montgomery, Douglas ; Runger, George ; Fowler, John ; McCarville, Daniel R. ; Rhoads, Teri Reed ; Stanley, James D. / Multivariate statistical methods for modeling and analysis of wafer probe test data. In: IEEE Transactions on Semiconductor Manufacturing. 2002 ; Vol. 15, No. 4. pp. 523-530.
@article{5925a82735554a5eacba097fda92cd8d,
title = "Multivariate statistical methods for modeling and analysis of wafer probe test data",
abstract = "Probe testing following wafer fabrication can produce extremely large amounts of data, which is often used to inspect a final product to determine if the product meets specifications. This data can be further utilized in studying the effects of the wafer fabrication process on the quality or yield of the wafers. Relationships among the parameters may provide valuable process information that can improve future production. This paper compares many methods of using the probe test data to determine the cause of low yield wafers. The methods discussed include two classes of traditional multivariate statistical methods, clustering and principal component methods and regression-based methods. These traditional methods are compared to a classification and regression tree (CART) method. The results for each method are presented. CART adequately fits the data and provides a {"}recipe{"} for avoiding low yield wafers and because CART is distribution-free there are no assumptions about the distributional properties of the data. CART is strongly recommended for analyzing wafer probe data.",
keywords = "CART, Multivariate statistical methods, Tree regression, Yield analysis",
author = "Skinner, {Katina R.} and Douglas Montgomery and George Runger and John Fowler and McCarville, {Daniel R.} and Rhoads, {Teri Reed} and Stanley, {James D.}",
year = "2002",
month = "11",
doi = "10.1109/TSM.2002.804901",
language = "English (US)",
volume = "15",
pages = "523--530",
journal = "IEEE Transactions on Semiconductor Manufacturing",
issn = "0894-6507",
publisher = "Institute of Electrical and Electronics Engineers Inc.",
number = "4",

}

TY - JOUR

T1 - Multivariate statistical methods for modeling and analysis of wafer probe test data

AU - Skinner, Katina R.

AU - Montgomery, Douglas

AU - Runger, George

AU - Fowler, John

AU - McCarville, Daniel R.

AU - Rhoads, Teri Reed

AU - Stanley, James D.

PY - 2002/11

Y1 - 2002/11

N2 - Probe testing following wafer fabrication can produce extremely large amounts of data, which is often used to inspect a final product to determine if the product meets specifications. This data can be further utilized in studying the effects of the wafer fabrication process on the quality or yield of the wafers. Relationships among the parameters may provide valuable process information that can improve future production. This paper compares many methods of using the probe test data to determine the cause of low yield wafers. The methods discussed include two classes of traditional multivariate statistical methods, clustering and principal component methods and regression-based methods. These traditional methods are compared to a classification and regression tree (CART) method. The results for each method are presented. CART adequately fits the data and provides a "recipe" for avoiding low yield wafers and because CART is distribution-free there are no assumptions about the distributional properties of the data. CART is strongly recommended for analyzing wafer probe data.

AB - Probe testing following wafer fabrication can produce extremely large amounts of data, which is often used to inspect a final product to determine if the product meets specifications. This data can be further utilized in studying the effects of the wafer fabrication process on the quality or yield of the wafers. Relationships among the parameters may provide valuable process information that can improve future production. This paper compares many methods of using the probe test data to determine the cause of low yield wafers. The methods discussed include two classes of traditional multivariate statistical methods, clustering and principal component methods and regression-based methods. These traditional methods are compared to a classification and regression tree (CART) method. The results for each method are presented. CART adequately fits the data and provides a "recipe" for avoiding low yield wafers and because CART is distribution-free there are no assumptions about the distributional properties of the data. CART is strongly recommended for analyzing wafer probe data.

KW - CART

KW - Multivariate statistical methods

KW - Tree regression

KW - Yield analysis

UR - http://www.scopus.com/inward/record.url?scp=0036869292&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=0036869292&partnerID=8YFLogxK

U2 - 10.1109/TSM.2002.804901

DO - 10.1109/TSM.2002.804901

M3 - Article

VL - 15

SP - 523

EP - 530

JO - IEEE Transactions on Semiconductor Manufacturing

JF - IEEE Transactions on Semiconductor Manufacturing

SN - 0894-6507

IS - 4

ER -