• PDF / 798,996 Bytes
• 3 Pages / 594 x 774 pts Page_size
• 14 Downloads / 193 Views

DOWNLOAD

REPORT

---

out and climb, the hi-g acceleration parallel to the samples' longitudinal axis reached 1.75 g while the accelerations on the other two axes were less than 0.15 g. To allow for identification of the melt interface, an initial sample position was chosen relative to the furnace so that a portion of the sample remained unmelted. Growth rates were selected between 4 mm/min and 14 mm/min because they would provide excellent dendritic structure and would allow for a sufficient sample length for analysis. For the majority of the samples, an equivalent ground-based control sample was solidified at the same growth rate (R), thermal gradient (G), and other experimental conditions, except at a normal gravity level. The advantage of processing alloys on the KC-135 is the ability to produce alternating low-gravity and high-gravity effects in the same sample. The disadvantage is the short solidification times and the inability to achieve a steady state condition. Prior to metallographic examination, each sample was mounted longitudinally, then polished using successively smaller grit paper and etched with Kallings II to reveal the melt interface. The longitudinal sample was then removed from the mount and sliced perpendicular to the growth axis at the end of the first lo-g and hi-g parabola (based on accelerometer data). This procedure was followed for all growth rate samples (ground-based control and KC-135), except those solidified at growth rates of 8 mm/min and 12 mm/min (ground-based control and KC-135). The 8 mm/min and 12 mm/min growth rate samples were again polished and etched longitudinally to reveal the melt interface. Subsequently, the longitudinal sample was glued to a graphite sheet. Using a micrometer attached to a diamond impregnated saw, at least three cross-sectional samples (slices sometimes only slightly exceeding 1 mm in thickness) were obtained within each lo-g and hi-g parabola. This included the first lo-g and first hi-g cross sections for comparison with the other growth rate samples. The cross-section slices for each growth rate (groundbased control and KC-135) were mounted, then polished and etched with Kallings II to reveal the primary dendritic structure. To insure reproducibility of data, three pictures were taken at different locations on each slice and enlarged into an 8 x 10 inch (110 x ) photomicrograph. Figure 1 shows representative cross sections from the 12 mm/min sample. Previously, the determination of primary dendrite spacings has been accomplished by the direct measurement of the characteristic distance, which is subjective since the investigator tends to select the perfectly-shaped dendrites and disregards the grain boundaries and zones of the sample with growth faults. Using a method developed by Jacobi, 3 a more objective technique was applied by counting all the primary arms within a large area including flaws in the patterns. From the number of primary dendritic trunks present within this area an average primary dendrite spacing (X 1) was calculated using the model of a square array