Published: 9 April 2021| Version 1 | DOI: 10.17632/6f8nygghjh.1
Here, we are using three different EB PBF scan strategies: linear scan (LS), ordered spot scan (OS), and random spots scan (RS) to fuse the widely used Ti-6Al-4V powder (Ti64). The change in scan strategies and therefore the thermal gradients create variation in the defects within the microstructure of the material. Such variations can be analyzed using optical microscopy (OM).
After standard metallography sample preparation, imaging was performed using an Olympus DSX510 optical microscope on 5x magnification to characterize the entire surface of each sample with sufficiently high image resolution. High magnification images of the entire surface with smaller field-of-view were first taken step-by-step and subsequently stitched together. Samples were then etched to reveal the underlying microstructure to compare to existing EBSD scans to compare the grain structure exhibited from using different techniques.
Using image processing (MIPAR) software, AM build defects were identified, quantified, and exported to a coding language for further data analysis. The code translates the provided quantitative data and generates histograms for eccentricity (roundness of defects), orientation (with respect to the positive x-axis), and caliper diameter (size), site-specifically. This allows to analysis the spatial distribution of defects, e.g., difference between defects in the center of the build vs. edge defects in each sample.
Files are not publicly available
You can contact the author to request the files
Steps to reproduce
A DSX510 optical microscope was used to image samples on 5x magnification after standard metallographic techniques were done. Kroll's reagent was used to etch the samples before being reimaged on the same settings previously used. Automatic focus and color correcting was applied on a bright field lens and the images were taken on a 3D setting to allow for uneven depth of the surface.
Analysis was done using a collection of software programs: a MIPAR recipe and MATLAB code was developed and aided with measurements of the images from ImageJ to characterize the defects within each sample.
Ohio State University, University of Tennessee Knoxville