ADVANCED MATERIALS & PROCESSES | MARCH 2025 13 be related to a number of process variations in the LPBF system. At this point, the authors are not able to confirm all the factors that led to variations in LOF void formation. Creep samples were machined from LPBF material and analyzed using x-ray computed tomography (XCT); samples were analyzed before and after creep testing to track void growth and potential new formation between samples with a low and high density of spatter particles. Optical microscopy was used to validate the XCT for smaller voids not in resolution of the instrument, and to distinguish the formation of voids at specific locations within the microstructure. SAMPLE FABRICATION Gas atomized Haynes Ni282 (UNS N07208) powder was acquired from Powder Alloy Corp. (Ohio, USA). The powder was processed with a Renishaw AM250 LPBF system (Wotton-under- Edge, UK), equipped with a 200 W, Yb-fiber laser. As shown in Fig. 2a, a four-inch-tall build was designed with a high density of material, including rods, plates, and thin walls. It is expected that samples farthest from the Ar inlet would be more susceptible to spatter particle deposition, than samples closer to the Ar inlet. Creep specimens were machined from the rods on the right and left sides of the build plate, labeled as T and C samples. Argon was flown from the inlet near the T samples, over the top of the build volume, to the outlet, closest to the C samples, as indicated by the arrow in Fig. 2a. Camera images were taken in-situ during the printing process of the powder volume as shown in Figs. 2b and 2c. Samples labeled C1 and C2, closest to the Ar outlet show a significant number of spatter particles, as indicated by the arrows in Fig. 2b, whereas the T1 and T2 samples near the Ar inlet only show signs of a few spatter particles as can be seen in Fig. 2c. More information on the Ni282 sample fabrication and preparation can be found elsewhere[13]. CREEP RESULTS The Ni282 creep specimens were subjected to a recrystallization heat treatment at 1180°C for 1 h, followed by a single step ageing treatment at 800°C for 4 h, per the recommendation following investigations of similar AM material[14]. Creep testing was performed on two T and C cylindrical threaded samples (6.36 mm in diameter gage section, 31.75 mm gage length) at 750°C, at 300 and 350 MPa, and creep curves are shown in Fig. 3. Two thermocouples were attached at the top and bottom grips to control the temperature within ±3°C. Creep curves were generated using two linear variable differential transformers (LVDT) connected to rods attached to the specimen heads. Overall, the T samples exhibited higher ductility, almost two times more than the C counterparts tested under each condition. A lower loading stress led to a longer creep life, but the ductility was consistent between the two testing conditions for both the T and C samples. CHARACTERIZATION OF VOIDS Before creep testing, the samples were examined using XCT in a Zeiss Metrotom system with an x-ray source of 200 kV. A short-scan strategy with 145 views between 0 and 197°, each view had an average 8×1 second acquisition, resulting in a scan time of 18 mins/scan. Advanced deep-learning algorithms[15] were used to reconstruct the void space for each sample, resolving any voids above ~75 µm with high confidence, although some smaller voids were still detected. Profiles from the XCT data on the T3 and C3 samples were processed to generate quantitative data on the various voids observed in the specimens, both in the reference state before testing and after rupture during creep testing. The ruptured samples were then crosssectioned after XCT, mounted in epoxy, and metallographically prepared down to a 0.05 µm finish with colloidal silica. A reference piece taken from the top of the rod before specimen machining was Fig. 2 — (a) Ni282 build investigated and camera images taken in-situ after melting mid-build of the (b) C samples closest to the Ar outlet and (c) T samples closest to the Ar inlet. Arrows in (b and c) indicate redeposited spatter particles[13]. Fig. 3 — Creep curves of T1 and C1 samples tested at 300 MPa, as compared to T3 and C3 tested at 350 MPa; testing temperature for both sets of specimens was 750°C. (a) (b) (c)
RkJQdWJsaXNoZXIy MTYyMzk3NQ==