January_AMP_Digital

A D V A N C E D M A T E R I A L S & P R O C E S S E S | J A N U A R Y 2 0 2 0 2 0 cracks consisted of a main radial crack emanating out from the pin in a singular direction with smaller cracks branching from the main. Third, it was determined that the direction used for serial sec- tioning greatly impacted the total quan- tity and dimensions of cracks observed. Longitudinal sectioning exacerbated existing cracks when advancing the pol- ishing plane. This is likely caused by re- lieving the compressive stress imparted on the glass by the housing. For these reasons, it is not recommended to sec- tion glass-to-metal seals perpendicular to the length of the pin for seal inspec- tion. It was determined from trans- verse mounts, Fig. 3(b), that preexisting cracks did exist in these seals; however their size varied significantly along the length of the seal and were not found to reach the housing material in the cases observed. Case study #3 – Porosity in ad- ditively manufactured stainless steel: Additive manufacturing (AM) enables geometric complexities and produc- tion agility not possible using tradition- al manufacturing techniques and has been demonstrated to produce metal with strength and ductility compara- ble to wrought material [18-22] . However, one of the primary challenges to the im- plementation of AM technologies is the development of qualification pathways for AM metal given internal defect pop- ulations. Much work has been done to identify the impact of porosity defects on performance. While a consensus on the impact of defects is still in debate, several researchers agree porosity can contribute to variability in mechanical performance [18-22] . 3D characterization methods have been used extensively to investigate po- rosity in AM metals, with the most com- mon technique being µCT. The resulting data can be difficult to quantitatively assess and can lead to uncertainty in measures of porosity. Here, porosity in a single AM 17-4PH stainless steel sample measuring 1 x 1 x 6 mm 3 is investigated using µCT and serial sectioning. Materi- al was manufactured by laser powder bed fusion using a ProX 200 printer. 3D reconstruction of solid material and porosity, as measured by µCT and serial sectioning, are shown in Fig. 4. Microcomputed tomography data used here, had a voxel resolution of 9 mi- crons. Over 1900 slices were acquired for the serial sectioning reconstruction at a resolution and slice thickness of 6 microns. Qualitatively, it appears there are more pores in the serial sectioning data than µCT. This suggests that either one technique is failing to detect a sig- nificant portion of porosity during ac- quisition or image processing decisions are resulting in the disparity. While differences in data acqui- sition influence segmentation ap- proach, the threshold value used for pore identification is found to play a significant role in this disparity. Here, a single threshold value on the 8-bit scale (0 to 255 grayscales) was stipu- lated to segment voxels representing pores from those representing solid material. Figure 4 presents an equiv- alent slice from (a) µCT and (b) serial sectioning at grayscale thresholds of 100, 175, and 215. In comparison, the processed serial sectioning data more accuratelyreflectstheoriginalslicedata. Figure 5 quantitatively demonstrates this by plotting the number of pores measured as a function of threshold value for µCT and serial sectioning. Two conclusions can be drawn: (1) Porosity in this AM metal is easy to identify using serial sectioning, and (2) variability in quantitative measures of porosity can be highly dependent on the acquisition method. In this study, serial section- ing produced data that was easier to interrogate because it provided great- er contrast and little blurring between the solid material and internal porosity compared to µCT. CONCLUSIONS Features, spatial locations, and morphological measures not readily attainable using 2D methods are acces- sible using 3D characterization. Here, the utility of mechanical serial section- ing using a Robo-Met.3D was demon- strated through three studies focused on (1) locating a failure in a micro-in- ductor, (2) identifying crack presence and extent in a highly constrained glass- to-metal seal, and (3) quantifying po- rosity in an AM precipitation-hardened stainless steel. Mechanical serial sectioning offers the ability to characterize large volumes at high resolution and can provide easi- er-to-interpret results compared to NDT methods such as ultrasound and µCT. Features not readily identifiable us- ing NDT methods that possess com- plex morphology and character (e.g., Fig. 4 — 3D reconstructions of solid material and pores in AM 17-4PH stainless steel from (a) µCT and (b) serial sectioning, along with representative slices at grayscale threshold values of 100, 175, and 215. Fig. 5 — Variation in pore quantity as a function of grayscale threshold value, i.e., image processing parameter, for µCT and serial sectioning data. (a) (b)