LAUSR

Diffraction-based analytical techniques for orientation imaging microscopy (OIM) with scanning and transmission electron microscope (SEM and TEM) instruments, such as electron backscatter diffraction (EBSD), transmission Kikuchi diffraction (TKD), and precession electron diffraction-assisted automated crystal orientation mapping (PED ACOM), offer powerful capabilities for spatially resolved studies of plastic deformation structures in materials [1]. These techniques are complementary regarding the respective combinations of the field-of-view and spatial resolution attainable. EBSD can gather data from very large areas (up to mm-scale, ≤ 10 μm) with spatial resolution limited to ~100 nm for Al, while TKD offers improved spatial resolution, ~5 to 10 nm, within reduced maximum fields of view in the ~10 to 10 μm range. PED ACOM offers the highest spatial resolution, routinely ~1 to 3 nm in a field emission TEM, but is limited to analysis of localized areas in the ~10 to 10 μm range [2-4]. For deformation studies with these electron diffraction techniques, spatially resolved crystal orientation changes must be measured with high accuracy and precision [1]. Sample preparation can strongly affect the accuracy and precision attained in strain analyses [5]. Because EBSD signals originate from the top 30 to 50 nm of a sample, high-quality surface preparation is critical for accurate OIM-based strain analysis, while TKD and PED ACOM OIM require electron transparent specimens. For accurate study of deformation structures, sample preparation artifacts, e.g., contamination, lattice damage, and additional plastic deformation, have to be minimized or avoided. We have performed a comparative study of different sample preparation protocols on the deformation structures introduced to aluminum samples by controlled uniaxial compression at room temperature to obtain plastic strains of 0, 4, 6, and 15%. As a quantitative metric for deformation, the geometrically necessary dislocation (GND) density, ρGND, has been derived from local orientation measurements under the assumption of negligible elastic stress [6]. We used two software analysis packages to determine ρGND from orientation maps: Atom [7], where ρGND is derived from the dislocation density tensor [6, 8]; and HKL CHANNEL5 [Oxford Instruments], where ρGND is calculated from representations of low angle boundaries [9, 10]. Four groups of sample preparation protocols have been applied for each deformed state of the Al and have been characterized by EBSD, TKD, and PED ACOM-based OIM: