Image-pixel averaging for accurate analysis of line-edge and linewidth roughness

Atsushi Hiraiwa, Akio Nishida

Research output: Contribution to journalArticle

14 Citations (Scopus)

Abstract

Line-edge and linewidth roughness (LER and LWR) is mostly characterized using the edge position data obtained by detecting edges in scanning electron microscope (SEM) images. In order to reduce data errors caused by SEM-image noise, image pixels are usually averaged along or across pattern edges (longitudinal or crosswise averaging) before the edge detection. In the case of the longitudinal averaging, it not only reduces the image-noise-induced additional LWR, but smoothes the real LWR. Because of this smoothing effect, the power spectral density (PSD) of the LWR decreases more markedly at larger wave numbers. A distinct feature of this PSD is oscillatory structures, which depend on the number of averaged pixels. It is difficult to formulate these modifications of PSD in the cases of any LWRs under arbitrary measurement and analysis conditions. By contrast, the crosswise averaging of the image pixels reduces only the image-noise effect without affecting the real LWR, enabling to enjoy the benefits of the methodology developed so far. Accordingly, the authors propose to apply only the crosswise averaging to SEM-image pixels before the edge detection.

Original languageEnglish
Article number023010
JournalJournal of Micro/Nanolithography, MEMS, and MOEMS
Volume10
Issue number2
DOIs
Publication statusPublished - 2011

Fingerprint

Linewidth
Power spectral density
roughness
Surface roughness
Pixels
pixels
Electron microscopes
Edge detection
Scanning
edge detection
electron microscopes
scanning
smoothing
methodology

Keywords

  • Averaging
  • Image noise
  • Line-edge roughness
  • Linewidth roughness
  • Power spectrum
  • PSD

ASJC Scopus subject areas

  • Electrical and Electronic Engineering
  • Mechanical Engineering
  • Electronic, Optical and Magnetic Materials
  • Condensed Matter Physics
  • Atomic and Molecular Physics, and Optics

Cite this

Image-pixel averaging for accurate analysis of line-edge and linewidth roughness. / Hiraiwa, Atsushi; Nishida, Akio.

In: Journal of Micro/Nanolithography, MEMS, and MOEMS, Vol. 10, No. 2, 023010, 2011.

Research output: Contribution to journalArticle

@article{f4ac40a589dc4cd8bafc57c87c359218,
title = "Image-pixel averaging for accurate analysis of line-edge and linewidth roughness",
abstract = "Line-edge and linewidth roughness (LER and LWR) is mostly characterized using the edge position data obtained by detecting edges in scanning electron microscope (SEM) images. In order to reduce data errors caused by SEM-image noise, image pixels are usually averaged along or across pattern edges (longitudinal or crosswise averaging) before the edge detection. In the case of the longitudinal averaging, it not only reduces the image-noise-induced additional LWR, but smoothes the real LWR. Because of this smoothing effect, the power spectral density (PSD) of the LWR decreases more markedly at larger wave numbers. A distinct feature of this PSD is oscillatory structures, which depend on the number of averaged pixels. It is difficult to formulate these modifications of PSD in the cases of any LWRs under arbitrary measurement and analysis conditions. By contrast, the crosswise averaging of the image pixels reduces only the image-noise effect without affecting the real LWR, enabling to enjoy the benefits of the methodology developed so far. Accordingly, the authors propose to apply only the crosswise averaging to SEM-image pixels before the edge detection.",
keywords = "Averaging, Image noise, Line-edge roughness, Linewidth roughness, Power spectrum, PSD",
author = "Atsushi Hiraiwa and Akio Nishida",
year = "2011",
doi = "10.1117/1.3598169",
language = "English",
volume = "10",
journal = "Journal of Micro/ Nanolithography, MEMS, and MOEMS",
issn = "1932-5150",
publisher = "SPIE",
number = "2",

}

TY - JOUR

T1 - Image-pixel averaging for accurate analysis of line-edge and linewidth roughness

AU - Hiraiwa, Atsushi

AU - Nishida, Akio

PY - 2011

Y1 - 2011

N2 - Line-edge and linewidth roughness (LER and LWR) is mostly characterized using the edge position data obtained by detecting edges in scanning electron microscope (SEM) images. In order to reduce data errors caused by SEM-image noise, image pixels are usually averaged along or across pattern edges (longitudinal or crosswise averaging) before the edge detection. In the case of the longitudinal averaging, it not only reduces the image-noise-induced additional LWR, but smoothes the real LWR. Because of this smoothing effect, the power spectral density (PSD) of the LWR decreases more markedly at larger wave numbers. A distinct feature of this PSD is oscillatory structures, which depend on the number of averaged pixels. It is difficult to formulate these modifications of PSD in the cases of any LWRs under arbitrary measurement and analysis conditions. By contrast, the crosswise averaging of the image pixels reduces only the image-noise effect without affecting the real LWR, enabling to enjoy the benefits of the methodology developed so far. Accordingly, the authors propose to apply only the crosswise averaging to SEM-image pixels before the edge detection.

AB - Line-edge and linewidth roughness (LER and LWR) is mostly characterized using the edge position data obtained by detecting edges in scanning electron microscope (SEM) images. In order to reduce data errors caused by SEM-image noise, image pixels are usually averaged along or across pattern edges (longitudinal or crosswise averaging) before the edge detection. In the case of the longitudinal averaging, it not only reduces the image-noise-induced additional LWR, but smoothes the real LWR. Because of this smoothing effect, the power spectral density (PSD) of the LWR decreases more markedly at larger wave numbers. A distinct feature of this PSD is oscillatory structures, which depend on the number of averaged pixels. It is difficult to formulate these modifications of PSD in the cases of any LWRs under arbitrary measurement and analysis conditions. By contrast, the crosswise averaging of the image pixels reduces only the image-noise effect without affecting the real LWR, enabling to enjoy the benefits of the methodology developed so far. Accordingly, the authors propose to apply only the crosswise averaging to SEM-image pixels before the edge detection.

KW - Averaging

KW - Image noise

KW - Line-edge roughness

KW - Linewidth roughness

KW - Power spectrum

KW - PSD

UR - http://www.scopus.com/inward/record.url?scp=80455176583&partnerID=8YFLogxK

UR - http://www.scopus.com/inward/citedby.url?scp=80455176583&partnerID=8YFLogxK

U2 - 10.1117/1.3598169

DO - 10.1117/1.3598169

M3 - Article

VL - 10

JO - Journal of Micro/ Nanolithography, MEMS, and MOEMS

JF - Journal of Micro/ Nanolithography, MEMS, and MOEMS

SN - 1932-5150

IS - 2

M1 - 023010

ER -