On Process Temperature in Powder-Bed Electron Beam Additive Manufacturing: Process Parameter Effects

Build part certification has been one of the primary roadblocks for effective usage and broader applications of metal additive manufacturing (AM) technologies including powder-bed electron beam additive manufacturing (EBAM). Process sensitivity to operating parameters, among others such as powder stock variations, is one major source of property scattering in EBAM parts. Thus, it is important to establish quantitative relations between the process parameters and process thermal characteristics that are closely correlated with the AM part properties. In this study, the experimental techniques, fabrications, and temperature measurements, developed in recent work (Cheng et al., 2014, "On Process Temperature in Powder-Bed Electron Beam Additive Manufacturing: Model Development and Experimental Validation," ASME J. Manuf. Sci. Eng., (in press)) were applied to investigate the process parameter effects on the thermal characteristics in EBAM with Ti-6Al-4 V powder, using the system-specific setting called “speed function (SF)” index that controls the beam speed and the beam current during a build. EBAM parts were fabricated using different levels of SF index (20–65) and examined in the part surface morphology and microstructures. In addition, process temperatures were measured by near infrared (NIR) thermography with further analysis of the temperature profiles and the melt pool size. The thermal model, also developed in recent work, was further employed for EBAM temperature predictions, and then compared with the experimental results. The major results are summarized as follows. SF index noticeably affects the thermal characteristics in EBAM, e.g., a melt pool length of 1.72 mm and 1.26 mm for SF20 and SF65, respectively, at 24.43 mm build height. SF setting also strongly affects the EBAM part quality including the surface morphology, surface roughness and part microstructures. In general, a higher SF index tends to produce parts of rougher surfaces with more pore features and large β grain columnar widths. Increasing the beam speed will reduce the peak temperatures, also reduce the melt pool sizes. Simulations conducted to evaluate the beam speed effects are in reasonable agreement compared to the experimental measurements in temperatures and melt pools sizes. However, the results of a lower SF case, SF20, show larger differences between the simulations and the experiments, about 58% for the melt pool size. Moreover, the higher the beam current, the higher the peak process temperatures, also the larger the melt pool. On the other hand, increasing the beam diameter monotonically decreases the peak temperature and the melt pool length.