Optical inspection of nanoscale structures using a novel machine learning based synthetic image generation algorithm

Sanyogita Purandare, Jinlong Zhu, Renjie Zhou, Gabriel Popescu, Alexander Gerhard Schwing, Lynford L Goddard

Research output: Contribution to journalArticle

Abstract

In this paper, we present a novel interpretable machine learning technique that uses unique physical insights about noisy optical images and a few training samples to classify nanoscale defects in noisy optical images of a semiconductor wafer. Using this technique, we not only detected both parallel bridge defects and previously undetectable perpendicular bridge defects in a 9-nm node wafer using visible light microscopy [Proc. SPIE 9424, 942416 (2015)], but we also accurately classified their shapes and estimated their sizes. Detection and classification of nanoscale defects in optical images is a challenging task. The quality of images is affected by diffraction and noise. Machine learning techniques can reduce noise and recognize patterns using a large training set. However, for detecting a rare “killer” defect, acquisition of a sufficient training set of high quality experimental images can be prohibitively expensive. In addition, there are technical challenges involved in using electromagnetic simulations and optimization of the machine learning algorithm. This paper proposes solutions to address each of the aforementioned challenges.

Original languageEnglish (US)
Pages (from-to)17743-17762
Number of pages20
JournalOptics Express
Volume27
Issue number13
DOIs
StatePublished - Jan 1 2019

Fingerprint

machine learning
inspection
defects
education
wafers
acquisition
electromagnetism
microscopy
optimization
diffraction
simulation

ASJC Scopus subject areas

  • Atomic and Molecular Physics, and Optics

Cite this

Optical inspection of nanoscale structures using a novel machine learning based synthetic image generation algorithm. / Purandare, Sanyogita; Zhu, Jinlong; Zhou, Renjie; Popescu, Gabriel; Schwing, Alexander Gerhard; Goddard, Lynford L.

In: Optics Express, Vol. 27, No. 13, 01.01.2019, p. 17743-17762.

Research output: Contribution to journalArticle

@article{cd2193b6e76946998bce66f684ead3fc,
title = "Optical inspection of nanoscale structures using a novel machine learning based synthetic image generation algorithm",
abstract = "In this paper, we present a novel interpretable machine learning technique that uses unique physical insights about noisy optical images and a few training samples to classify nanoscale defects in noisy optical images of a semiconductor wafer. Using this technique, we not only detected both parallel bridge defects and previously undetectable perpendicular bridge defects in a 9-nm node wafer using visible light microscopy [Proc. SPIE 9424, 942416 (2015)], but we also accurately classified their shapes and estimated their sizes. Detection and classification of nanoscale defects in optical images is a challenging task. The quality of images is affected by diffraction and noise. Machine learning techniques can reduce noise and recognize patterns using a large training set. However, for detecting a rare “killer” defect, acquisition of a sufficient training set of high quality experimental images can be prohibitively expensive. In addition, there are technical challenges involved in using electromagnetic simulations and optimization of the machine learning algorithm. This paper proposes solutions to address each of the aforementioned challenges.",
author = "Sanyogita Purandare and Jinlong Zhu and Renjie Zhou and Gabriel Popescu and Schwing, {Alexander Gerhard} and Goddard, {Lynford L}",
year = "2019",
month = "1",
day = "1",
doi = "10.1364/OE.27.017743",
language = "English (US)",
volume = "27",
pages = "17743--17762",
journal = "Optics Express",
issn = "1094-4087",
publisher = "The Optical Society",
number = "13",

}

TY - JOUR

T1 - Optical inspection of nanoscale structures using a novel machine learning based synthetic image generation algorithm

AU - Purandare, Sanyogita

AU - Zhu, Jinlong

AU - Zhou, Renjie

AU - Popescu, Gabriel

AU - Schwing, Alexander Gerhard

AU - Goddard, Lynford L

PY - 2019/1/1

Y1 - 2019/1/1

N2 - In this paper, we present a novel interpretable machine learning technique that uses unique physical insights about noisy optical images and a few training samples to classify nanoscale defects in noisy optical images of a semiconductor wafer. Using this technique, we not only detected both parallel bridge defects and previously undetectable perpendicular bridge defects in a 9-nm node wafer using visible light microscopy [Proc. SPIE 9424, 942416 (2015)], but we also accurately classified their shapes and estimated their sizes. Detection and classification of nanoscale defects in optical images is a challenging task. The quality of images is affected by diffraction and noise. Machine learning techniques can reduce noise and recognize patterns using a large training set. However, for detecting a rare “killer” defect, acquisition of a sufficient training set of high quality experimental images can be prohibitively expensive. In addition, there are technical challenges involved in using electromagnetic simulations and optimization of the machine learning algorithm. This paper proposes solutions to address each of the aforementioned challenges.

AB - In this paper, we present a novel interpretable machine learning technique that uses unique physical insights about noisy optical images and a few training samples to classify nanoscale defects in noisy optical images of a semiconductor wafer. Using this technique, we not only detected both parallel bridge defects and previously undetectable perpendicular bridge defects in a 9-nm node wafer using visible light microscopy [Proc. SPIE 9424, 942416 (2015)], but we also accurately classified their shapes and estimated their sizes. Detection and classification of nanoscale defects in optical images is a challenging task. The quality of images is affected by diffraction and noise. Machine learning techniques can reduce noise and recognize patterns using a large training set. However, for detecting a rare “killer” defect, acquisition of a sufficient training set of high quality experimental images can be prohibitively expensive. In addition, there are technical challenges involved in using electromagnetic simulations and optimization of the machine learning algorithm. This paper proposes solutions to address each of the aforementioned challenges.

UR - http://www.scopus.com/inward/record.url?scp=85067982406&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=85067982406&partnerID=8YFLogxK

U2 - 10.1364/OE.27.017743

DO - 10.1364/OE.27.017743

M3 - Article

C2 - 31252730

AN - SCOPUS:85067982406

VL - 27

SP - 17743

EP - 17762

JO - Optics Express

JF - Optics Express

SN - 1094-4087

IS - 13

ER -