About the automatic measurement of the dislocation density obtained by R-ECCI

J Gallet (1) , M Perez (1) , S Dubail (2) , Thibaut Chaise (3, 4) , Thierry Douillard (1) , C Langlois (1) , S Cazottes (1)

1 MATEIS - Matériaux, ingénierie et science [Villeurbanne]
2 Axon Square SAS
3 LaMCoS - Laboratoire de Mécanique des Contacts et des Structures [Villeurbanne]
4 LMI - Laboratoire des Multimatériaux et Interfaces

A proof of concept of a new method for automatic characterization of the dislocation density from scanning electron microscopy images is presented. A series of backscattered electron images are acquired while the sample is rotated. For each pixel of the region of interest, the variation of the grey-level intensity as a function of the rotation angle, called the intensity profile, is calculated. This profile can be used to determine the nature of each pixel (dislocation, matrix or noise), such that an automatic dislocation density can be determined within the region of interest. The method is well adapted for dislocation densities ranging from 1012 to 1014 m−2. The simulation of a volume containing dislocations enabled the determination of the maximum and minimum densities attainable as well as the theoretical and experimental measurement errors related to the projection of this volume on a two-dimensional image. The theoretical measurement error due to the projection of dislocation on a surface, is 3% for low dislocation densities (1012m−2) and 20% for higher dislocation densities (1014 m−2). Experimentally, measurement errors are limited by image analysis conditions, which leads to total measurement errors of 15% for 1012m−2 and 34% error for 1014 m−2. These uncertainties were obtained considering a given analyzed depth value, that could not be experimentally verified. This uncertainty on the depth value leads to large errors bars in the final measurement, which can reach an order of magnitude.

Keywords: Density dislocation; ECCI; Clustering

Domains: Mechanics of materials [physics.class-ph]; Signal and Image processing

J Gallet, M Perez, S Dubail, Thibaut Chaise, Thierry Douillard, et al.. About the automatic measurement of the dislocation density obtained by R-ECCI. Materials Characterization, inPress, ⟨10.1016/j.matchar.2022.112358⟩. ⟨hal-03797492⟩