ADVANCED MATERIALS & PROCESSES | MARCH 2025 14 (a) (b) prepared in a similar manner. Optical characterization of the as-polished surfaces was performed with a Leica light microscope with up to 100 images captured automatically at 200× using Leica’s imaging software. The images were processed with a script that converts them to 8-bit files and uses shadow imaging to distinguish between the solid material and voids under grayscale; a threshold is applied, and the statistical data can be generated on all the observed voids. More information on the image analysis technique can be found in previous work[8]. Limitation in void size detectable by image analysis was <5 µm. Arguably, higher resolution optical images could be taken of the sample surfaces but would be more time consuming to focus and image across large cross-sections. Magnifications of 200× were chosen as the best compromise between resolution and imaging time. Comparisons of the number of voids detected by XCT and measured from image analysis of optical micro- scopy (OM) for reference and ruptured conditions is shown in Fig. 4; the frequency of voids shown was binned and then normalized by a factor of 100 for each data set to the maximum void frequency to be able to better compare the XCT data to that from OM. The maximum number of voids observed in the ruptured T3 and C3 samples from XCT was 11 and 196, respectively, whereas OM detected a maximum of 99 and 40 for the T3 and C3 ruptured samples, respectively. The T3 samples, being closest to the Ar inlet showed the least number of voids, as can be seen in Fig. 4a. The XCT resolved a small number of voids in the T3 reference material, <100 voids, with minor increases in number after creep testing in the ruptured sample. Optical image analysis from the reference material showed the presence of many small voids <25 µm in diameter that could not be detected by XCT. Moreover, the number of small voids increased after creep testing, most likely attributed to creep induced voids such as creep cavitation. The C3 sample, fabricated on the side of the build plate farthest from the Ar inlet and expected to have a higher number of redeposited spatter particles, exhibited a significantly higher number of voids, almost two times, in the reference sample, as analyzed by the XCT data and shown in Fig. 4b. After creep testing, there was an increase in magnitude in the number of larger voids 60-100 µm observed in the ruptured sample, and a decrease in the smaller voids <50 µm. Optical image analysis of the C3 reference and ruptured specimens showed a high concentration of smaller voids <50 µm and with no observation of larger voids. Since the OM data shown in Fig. 4 is reliant upon a single cross-section of the sample, it is understandable that detecting high frequencies of larger voids is unlikely to observed. XCT reconstructed 3D renderings of the creep specimens with voids and compressed cross-sections across a few millimeters near the fracture surface are shown in Fig. 5. 3D data from the XCT was used to register the approximate location, within 100 µm, of the reference as compared to ruptured condition. Registering the two conditions allows for tracking of the voids that likely led to fracture. Although some information is lost at the fracture surface where the two ends of the sample separated, locating likely suspects of failure is still reasonable. Note the smaller cross-section sizes in the ruptured conditions (Figs. 5b and 5d), a result of necking in the gage section. The T3 sample exhibited more ductility, as shown in Fig. 3a, and therefore, had a smaller cross-section after rupture than that of the C3 sample, as shown in Fig. 5d. As can be seen in Fig. 5a, two larger voids can be observed in the reference condition, near where the fracture surface is expected. After rupture, a significantly larger void is observed in the same spot as indicated by the arrows in Fig. 5b; additionally, other larger voids are also observed in the ruptured condition in Fig. 5b, signifying that many small voids <75 µm, either from the LPBF process or induced by creep cavitation, also coalesce. Similar behavior was observed be- tween the C3 reference and ruptured conditions in Fig. 5c and 5d, respectively, Fig. 4 — Quantification of voids detected from XCT and measured using image analysis of optical images for the (a) T3 and (b) C3 specimens before and after rupture. The frequency of voids was binned and then normalized by a factor of 100 for all data sets to better compare the magnitudes between XCT and OM.
RkJQdWJsaXNoZXIy MTYyMzk3NQ==