Point defect characterization in HAADF-STEM images using multivariate statistical analysis

Michael C. Sarahan; Miaofang Chi; Daniel J. Masiel; Nigel D. Browning

doi:10.1016/j.ultramic.2010.11.033

Point defect characterization in HAADF-STEM images using multivariate statistical analysis

Michael C. Sarahan, Miaofang Chi, Daniel J. Masiel, Nigel D. Browning

Research output: Contribution to journal › Article › peer-review

11 Scopus citations

Abstract

Quantitative analysis of point defects is demonstrated through the use of multivariate statistical analysis. This analysis consists of principal component analysis for dimensional estimation and reduction, followed by independent component analysis to obtain physically meaningful, statistically independent factor images. Results from these analyses are presented in the form of factor images and scores. Factor images show characteristic intensity variations corresponding to physical structure changes, while scores relate how much those variations are present in the original data. The application of this technique is demonstrated on a set of experimental images of dislocation cores along a low-angle tilt grain boundary in strontium titanate. A relationship between chemical composition and lattice strain is highlighted in the analysis results, with picometer-scale shifts in several columns measurable from compositional changes in a separate column.

Original language	English
Pages (from-to)	251-257
Number of pages	7
Journal	Ultramicroscopy
Volume	111
Issue number	3
DOIs	https://doi.org/10.1016/j.ultramic.2010.11.033
State	Published - Feb 2011
Externally published	Yes

Funding

MCS thanks David Morgan, Quentin Ramasse, Chad Parish and Paul Kotula for helpful conversation and Bernhard Schaffer, Patricia Abellan, and the journal referees for productive comments on this manuscript. This work was supported by the United States Department of Energy , Grant no. DE-FG02-03ER46057 and by the Materials Design Institute, Los Alamos National Laboratory, LANS contract 25110-002-06, Mod 6 and by a University of California Lab Management Fees Award. SuperSTEM is funded by EPSRC. M. Chi was supported by an LLNL SEGRF fellowship during the TEM work of this paper.

Funders	Funder number
LANS	25110-002-06
Materials Design Institute
University of California Lab Management
U.S. Department of Energy	DE-FG02-03ER46057
Lawrence Livermore National Laboratory
Los Alamos National Laboratory
Engineering and Physical Sciences Research Council

Keywords

Image processing
Multivariate statistical analysis
Point defect
STEM

Access to Document

10.1016/j.ultramic.2010.11.033

Cite this

@article{0e66716628b2494e8bf0746dbbe576fe,

title = "Point defect characterization in HAADF-STEM images using multivariate statistical analysis",

abstract = "Quantitative analysis of point defects is demonstrated through the use of multivariate statistical analysis. This analysis consists of principal component analysis for dimensional estimation and reduction, followed by independent component analysis to obtain physically meaningful, statistically independent factor images. Results from these analyses are presented in the form of factor images and scores. Factor images show characteristic intensity variations corresponding to physical structure changes, while scores relate how much those variations are present in the original data. The application of this technique is demonstrated on a set of experimental images of dislocation cores along a low-angle tilt grain boundary in strontium titanate. A relationship between chemical composition and lattice strain is highlighted in the analysis results, with picometer-scale shifts in several columns measurable from compositional changes in a separate column.",

keywords = "Image processing, Multivariate statistical analysis, Point defect, STEM",

author = "Sarahan, {Michael C.} and Miaofang Chi and Masiel, {Daniel J.} and Browning, {Nigel D.}",

year = "2011",

month = feb,

doi = "10.1016/j.ultramic.2010.11.033",

language = "English",

volume = "111",

pages = "251--257",

journal = "Ultramicroscopy",

issn = "0304-3991",

publisher = "Elsevier",

number = "3",

}

TY - JOUR

T1 - Point defect characterization in HAADF-STEM images using multivariate statistical analysis

AU - Sarahan, Michael C.

AU - Chi, Miaofang

AU - Masiel, Daniel J.

AU - Browning, Nigel D.

PY - 2011/2

Y1 - 2011/2

N2 - Quantitative analysis of point defects is demonstrated through the use of multivariate statistical analysis. This analysis consists of principal component analysis for dimensional estimation and reduction, followed by independent component analysis to obtain physically meaningful, statistically independent factor images. Results from these analyses are presented in the form of factor images and scores. Factor images show characteristic intensity variations corresponding to physical structure changes, while scores relate how much those variations are present in the original data. The application of this technique is demonstrated on a set of experimental images of dislocation cores along a low-angle tilt grain boundary in strontium titanate. A relationship between chemical composition and lattice strain is highlighted in the analysis results, with picometer-scale shifts in several columns measurable from compositional changes in a separate column.

AB - Quantitative analysis of point defects is demonstrated through the use of multivariate statistical analysis. This analysis consists of principal component analysis for dimensional estimation and reduction, followed by independent component analysis to obtain physically meaningful, statistically independent factor images. Results from these analyses are presented in the form of factor images and scores. Factor images show characteristic intensity variations corresponding to physical structure changes, while scores relate how much those variations are present in the original data. The application of this technique is demonstrated on a set of experimental images of dislocation cores along a low-angle tilt grain boundary in strontium titanate. A relationship between chemical composition and lattice strain is highlighted in the analysis results, with picometer-scale shifts in several columns measurable from compositional changes in a separate column.

KW - Image processing

KW - Multivariate statistical analysis

KW - Point defect

KW - STEM

UR - http://www.scopus.com/inward/record.url?scp=78650655918&partnerID=8YFLogxK

U2 - 10.1016/j.ultramic.2010.11.033

DO - 10.1016/j.ultramic.2010.11.033

M3 - Article

AN - SCOPUS:78650655918

SN - 0304-3991

VL - 111

SP - 251

EP - 257

JO - Ultramicroscopy

JF - Ultramicroscopy

IS - 3

ER -

Point defect characterization in HAADF-STEM images using multivariate statistical analysis

Abstract

Funding

Keywords

Access to Document

Other files and links

Fingerprint

Cite this