E-Thesis 182 views 59 downloads
Resolving geometrically necessary dislocations and application in copper. / David Kerr
Swansea University Author: David Kerr
-
PDF | E-Thesis
Download (28.28MB)
Abstract
The use of Electron Back-Scatter Diffraction (EBSD) as a tool for strain measurement is not a new application. The band contrast within a captured Kikuchi pattern has been shown to decrease in regions with high plastic strain. Recently, however, highly accurate EBSD orientation maps have been used t...
Published: |
2013
|
---|---|
Institution: | Swansea University |
Degree level: | Doctoral |
Degree name: | Ph.D |
URI: | https://cronfa.swan.ac.uk/Record/cronfa42263 |
Abstract: |
The use of Electron Back-Scatter Diffraction (EBSD) as a tool for strain measurement is not a new application. The band contrast within a captured Kikuchi pattern has been shown to decrease in regions with high plastic strain. Recently, however, highly accurate EBSD orientation maps have been used to recover dislocation content by directly relating lattice curvature, measured as point-to-point misorientation differences, to the crystal structure of the material using minimization techniques to recover the dislocation content determined to be geometrically necessary. This dislocation content is termed the Geometrically- Necessary Dislocations (GND) content. High resolution EBSD techniques exist which are able to index Kikuchi patterns much more accurately than the more widely used Hough transformation; however, the Hough transformation shows advantages of low computational cost and fast data acquistion times. As a result, larger data sets are theoretically possible using a Hough transformation, however with a lower indexing accuracy. The main aims of this work are to determine the optimal settings for data acquisition by EBSD for the purposes of strain mapping, to write code to create a strain map, and to use that code and mapping settings for an investigation into the microstructural development of bent samples of copper. Three sets of samples, each receiving varying applied strains and heat-treatments were created and statistics concerning the strain around key features, such as E3 and boundaries which are important for the mechanical performance of the material, were determined. The results of the investigation into the optimal settings for the purpose of strain mapping by Hough transformation and EBSD showed that a higher Kikuchi pattern resolution can be sacrificed to decrease mapping time and increase the size of data set possible. External time constraints, such as the microscope filament life span, limit the longest mapping time possible and so must be considered where large data sets are desirable. The accuracy of indexing by band centres and band edges was shown to be similar, however with greatly differing noise profiles. A portion of the noise of indexing by band edges was shown to be normally distributed and a noise reduction algorithm to remove normally distributed noise is suggested. It is also shown that the fraction of total dislocation content recoverable as GNDs changes as the structure of GNDs changes, which makes strain analysis based on the value of the GND content calculated ineffective. Instead, analysis of the GND structure itself is shown to correlate better with the restorative state of a material. A method of (quantifying the connectedness of GND content is created and is suggested for future use. The investigation of bent samples of copper showed the effect of strain on the development of the copper microstructure after heat-treatment. Different GND structures were observed at different stages of restoration, including a secondary recovery stage after deformation is left behind mobile boundaries. Conversely, observation of the GND structure was then shown to determine the restoration stage, which aided the conclusion that the compressive end of the samples was restoring faster than the tensile side. This is believed to be because the neutral plane, at which the strain changes from tensile to compressive in nature, is located closer to the tensile end of the sample due to strain hardening effects during bending. The larger strain in the compressive region acts as a higher driving force for restoration than the tensile end. The strain around boundaries was also measured based on the boundary character. E3 boundaries were shown to shield the grain bulk from deformation under applied strain compared to non-E3 boundaries. |
---|---|
Keywords: |
Materials science. |
College: |
Faculty of Science and Engineering |