Abstract Scope |
Introduction - Wrought, high-strength aluminum alloys are used extensively in aerospace and automotive applications due to the high strength-to-weight ratio. Further improvements to certain properties, including specific modulus and dimensional stability during thermal cycling, can be achieved by incorporating discontinuous reinforcement into the Al matrix to form metal matrix composites (MMCs). Additive manufacturing (AM) of MMCs is emerging as a promising solution to meet the demands of the aerospace and automotive industries by overcoming the limitations of conventional manufacturing processes. One common problem in AM is the formation and retention of gas porosity in the final component, which lowers the ductility and strength of the material. The formation of porosity is exacerbated in processes that occur in a vacuum, such as electron beam freeform fabrication (EBF3). In this work, Al 6061 MMCs formed by reaction synthesis were produced by both EBF3 and laser powder-bed fusion (L-PBF) The size, morphology and volume fraction of gas pores within the materials was quantified and compared. CALPHAD modeling was used in conjunction with these experimental results to identify differing sources of porosity between the two AM processes.
Experimental Procedures - Powder feedstocks designed to form 0, 2, and 10 vol.% of reaction synthesis inoculant content within an Al 6061 matrix were supplied by Elementum 3D and used for AM deposition with EBF3 and L-PBF. EBF3 being a wire-fed process, powder-cored tubular wires (PCTW) using the powders as the wire core were produced. EBF3 deposits were made at NASA-LaRC using the PCTW feedstocks. All EBF3 builds were deposited on 6.4 mm (0.25 in) thick Al 6061-T6 build plates that were preheated to 100 °C. A focused and rastered electron beam was used with a beam current of 55 mA, an accelerating voltage of 30 kV, a wire feed rate of 1016 mm/min, and a travel speed of 254 mm/min. L-PBF deposition was performed at Elementum 3D on an EOS M290 laser-powder bed fusion printer with a 400 W Yb-fiber laser. Deposition was done in an inert atmosphere with a laminar flow of Ar to clear smoke or ejected matter from the build path. A preheat of 200 °C was applied to the build plate via a PID-controlled heating element built into the EOS printer. Specimens were printed directly onto the build plate using a hatch pattern of 7 mm-wide stripes that were rotated 67° after each layer. To quantify porosity, Archimedes density measurements and MicroCT measurements using a Zeiss Versa 520 Micro-CT X-ray Microscope were performed. A FEI Quanta 600i Environmental SEM was used to capture micrographs of the pores and obtain EDS measurements of the pore interiors. Finally, Thermo-Calc® software was used to perform thermodynamic calculations to understand the expected compositions of the gas phases in each AM process and to examine gas phase stability.
Results and Discussion - MicroCT measurement showed that both EBF3 specimens tended to have larger pores than the L-PBF specimens (maximum histogram bin centers of 352 μm vs 53 μm) This is likely as a consequence of EBF3 having longer solidification times than L-PBF, thus allowing more time for dissolved gaseous species to diffuse to nucleated pores. The pore diameter distributions in the EBF3 materials were not significantly different, whereas in the L-PBF materials, the pore diameters in the 2 vol.% material were statistically larger than those in the 10 vol.% material. For the EBF3 builds, a trend of decreasing percent theoretical density with increasing inoculant content was documented via Archimedes density measurements. More sites are available for pores to form with a greater amount of poorly wetted second phase particles, which appears to be the case for the EBF3 materials. For L-PBF, the density vs. inoculant content trend is reversed, with higher inoculant content materials possessing greater theoretical percent density. The cooling rates determined in L-PBF processes are much larger than those in directed energy deposition processes like EBF3. As a result, the solidification times are reduced and shrinkage stresses are greater in L-PBF. Consequently, the data for L-PBF may indicate that the percent theoretical density is more affected by the extent of solidification cracking rather than any porosity that has sufficient time to form. The bulk of the pore aspect ratio distribution lies within the range associated with 2D lenticular morphologies, which is further evidence that pores are nucleated on poorly wetted inclusions. Thermo-Calc calculations revealed that at 1 atm of pressure, the gas phase is primarily composed of H2, whereas at 10-5 torr (EBF3 vacuum pressure), the gas phase is almost 100 mass% vaporized Mg when solidification begins. These calculations suggest two different sources of porosity between the two AM processes.
Conclusions - MicroCT measurements revealed that the EBF3 process produced larger pores than L-PBF process (352 μm vs. 53 μm). Statistical analysis found that EBF3 produced similar pore diameter distributions regardless of inoculant content, while the size distribution of the pores in L-PBF builds was affected by the amount of inoculant content. Porosity was the dominant factor affecting part density in EBF3 which contrasts with solidification cracking in L-PBF. The evidence from thermodynamic calculations and comparison with experimental pore size and morphology data strongly suggests that diffusion-controlled hydrogen gas pore growth and buoyancy-controlled magnesium vapor pore growth, both nucleating on poorly wetted inclusion particles, are the most probable sources of porosity in the L-PBF and EBF3 processes, respectively. |