Machine learning-based novelty detection for faulty wafer detection in semiconductor manufacturing

Dongil Kim, Pilsung Kang, Sungzoon Cho, Hyoung Joo Lee, Seungyong Doh

Research output: Contribution to journalArticle

42 Citations (Scopus)

Abstract

Since semiconductor manufacturing consists of hundreds of processes, a faulty wafer detection system, which allows for earlier detection of faulty wafers, is required. statistical process control (SPC) and virtual metrology (VM) have been used to detect faulty wafers. However, there are some limitations in that SPC requires linear, unimodal and single variable data and VM underestimates the deviations of predictors. In this paper, seven different machine learning-based novelty detection methods were employed to detect faulty wafers. The models were trained with Fault Detection and Classification (FDC) data to detect wafers having faulty metrology values. The real world semiconductor manufacturing data collected from a semiconductor fab were tested. Since the real world data have more than 150 input variables, we employed three different dimensionality reduction methods. The experimental results showed a high True Positive Rate (TPR). These results are promising enough to warrant further study.

Original languageEnglish
Pages (from-to)4075-4083
Number of pages9
JournalExpert Systems with Applications
Volume39
Issue number4
DOIs
Publication statusPublished - 2012 Mar 1
Externally publishedYes

Fingerprint

Learning systems
Statistical process control
Semiconductor materials
Fault detection

Keywords

  • Dimensionality reduction
  • Faulty wafer detection
  • Novelty detection
  • Semiconductor manufacturing
  • Virtual metrology

ASJC Scopus subject areas

  • Engineering(all)
  • Computer Science Applications
  • Artificial Intelligence

Cite this

Machine learning-based novelty detection for faulty wafer detection in semiconductor manufacturing. / Kim, Dongil; Kang, Pilsung; Cho, Sungzoon; Lee, Hyoung Joo; Doh, Seungyong.

In: Expert Systems with Applications, Vol. 39, No. 4, 01.03.2012, p. 4075-4083.

Research output: Contribution to journalArticle

Kim, Dongil ; Kang, Pilsung ; Cho, Sungzoon ; Lee, Hyoung Joo ; Doh, Seungyong. / Machine learning-based novelty detection for faulty wafer detection in semiconductor manufacturing. In: Expert Systems with Applications. 2012 ; Vol. 39, No. 4. pp. 4075-4083.
@article{75336c7230754475a2a8e84d79e60ede,
title = "Machine learning-based novelty detection for faulty wafer detection in semiconductor manufacturing",
abstract = "Since semiconductor manufacturing consists of hundreds of processes, a faulty wafer detection system, which allows for earlier detection of faulty wafers, is required. statistical process control (SPC) and virtual metrology (VM) have been used to detect faulty wafers. However, there are some limitations in that SPC requires linear, unimodal and single variable data and VM underestimates the deviations of predictors. In this paper, seven different machine learning-based novelty detection methods were employed to detect faulty wafers. The models were trained with Fault Detection and Classification (FDC) data to detect wafers having faulty metrology values. The real world semiconductor manufacturing data collected from a semiconductor fab were tested. Since the real world data have more than 150 input variables, we employed three different dimensionality reduction methods. The experimental results showed a high True Positive Rate (TPR). These results are promising enough to warrant further study.",
keywords = "Dimensionality reduction, Faulty wafer detection, Novelty detection, Semiconductor manufacturing, Virtual metrology",
author = "Dongil Kim and Pilsung Kang and Sungzoon Cho and Lee, {Hyoung Joo} and Seungyong Doh",
year = "2012",
month = "3",
day = "1",
doi = "10.1016/j.eswa.2011.09.088",
language = "English",
volume = "39",
pages = "4075--4083",
journal = "Expert Systems with Applications",
issn = "0957-4174",
publisher = "Elsevier Limited",
number = "4",

}

TY - JOUR

T1 - Machine learning-based novelty detection for faulty wafer detection in semiconductor manufacturing

AU - Kim, Dongil

AU - Kang, Pilsung

AU - Cho, Sungzoon

AU - Lee, Hyoung Joo

AU - Doh, Seungyong

PY - 2012/3/1

Y1 - 2012/3/1

N2 - Since semiconductor manufacturing consists of hundreds of processes, a faulty wafer detection system, which allows for earlier detection of faulty wafers, is required. statistical process control (SPC) and virtual metrology (VM) have been used to detect faulty wafers. However, there are some limitations in that SPC requires linear, unimodal and single variable data and VM underestimates the deviations of predictors. In this paper, seven different machine learning-based novelty detection methods were employed to detect faulty wafers. The models were trained with Fault Detection and Classification (FDC) data to detect wafers having faulty metrology values. The real world semiconductor manufacturing data collected from a semiconductor fab were tested. Since the real world data have more than 150 input variables, we employed three different dimensionality reduction methods. The experimental results showed a high True Positive Rate (TPR). These results are promising enough to warrant further study.

AB - Since semiconductor manufacturing consists of hundreds of processes, a faulty wafer detection system, which allows for earlier detection of faulty wafers, is required. statistical process control (SPC) and virtual metrology (VM) have been used to detect faulty wafers. However, there are some limitations in that SPC requires linear, unimodal and single variable data and VM underestimates the deviations of predictors. In this paper, seven different machine learning-based novelty detection methods were employed to detect faulty wafers. The models were trained with Fault Detection and Classification (FDC) data to detect wafers having faulty metrology values. The real world semiconductor manufacturing data collected from a semiconductor fab were tested. Since the real world data have more than 150 input variables, we employed three different dimensionality reduction methods. The experimental results showed a high True Positive Rate (TPR). These results are promising enough to warrant further study.

KW - Dimensionality reduction

KW - Faulty wafer detection

KW - Novelty detection

KW - Semiconductor manufacturing

KW - Virtual metrology

UR - http://www.scopus.com/inward/record.url?scp=82255179131&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=82255179131&partnerID=8YFLogxK

U2 - 10.1016/j.eswa.2011.09.088

DO - 10.1016/j.eswa.2011.09.088

M3 - Article

AN - SCOPUS:82255179131

VL - 39

SP - 4075

EP - 4083

JO - Expert Systems with Applications

JF - Expert Systems with Applications

SN - 0957-4174

IS - 4

ER -