|dc.description.abstract||Powder-bed fusion (PBF) additive manufacturing has emerged as an innovative and new way of fabricating highly complex geometrical parts. However, applying electron beam powder bed fusion (EB-PBF), one of the two major PBF processes, to the difficult-to-weld Co-29Cr-10Ni-7W alloy result in extensive hot cracking along the columnar grains boundaries. Current literature is not clear as to whether solidification or liquation cracking is the initiation mechanism and how these two contribute to crack growth during EB-PBF of difficulty-to-weld alloys. The major aim of this work is to understand the solidification, how cracks develop and the contributions of liquation and hot tear to crack growth during EB-PBF.
To understand solidification sequence and phase transition temperatures of the alloy, differential scanning calorimetry (DSC) experiments were conducted. This is followed by a series of EB-PBF experiments using an Arcam A1 machine. EB-PBF parameters were selected from the preliminary work to ensure samples to be fully dense. For the main study, EB-PBF specimens were made with one focused and one defocused beam while other parameters kept unchanged. This is to provide conditions of two different track shapes for studying grain growth. To study the solidified microstructure (including cracks), optical and scanning electron microscopy (SEM) were used. To reveal the crack network, EB-PBF samples were progressively ground off layer-by-layer starting from the top surface. Numerical simulation of a single scan using ANSYS-APDL software was carried out to illustrate how the scan direction may relate to directions of thermal stresses during track solidification. Finally, single track experiments by having a single track in the last layer were conducted to provide samples so that how crack initiated may be understood.
DSC plots have shown and confirmed the wide freezing range of about 180oC, thus indicating the alloy being susceptible to hot cracking. The two different focus offset values have resulted in the shape of melt pool/track being varied significantly and thus the grain growth direction changed considerably for a significant portion of the track. It has been found that as the beam became more focused, the cross-section of the melt track became more U-shaped. The lateral heat transfer dominance in the top region then resulted in the growth of dendrites, epitaxial and in columnar form, horizontally. The layer thickness of these horizontal dendrites is up to 80 micrometers and is significant considering the additive layer thickness being 70 micrometers. Horizontal grain growth has been found to effectively stop crack propagation. Thus, further modification on the EB-PBF parameters has successfully resulted in a novel grain structure that no cracks are observed.
Examination of surface morphologies shows that cracks appear on the samples’ top surface have hot tearing characteristics while inter-granular cracks indicate liquation cracking features. On examining the crack network, firstly, it has been observed that surface cracks tend to orientate normal to the scan direction which changes after each layer. Computer simulation has demonstrated the dominant tension parallel to scan direction. Secondly, the relationship between the orientated cracks on the surface and crack networks beneath the top layer has been shown to be the result of scan direction change after each layer. From this, how crack networks have grown and how liquation has assisted this growth have been demonstrated.
Single track experiments have been found to be useful to reveal crack initiation. Examination of the single-track samples has shown that cracks in single tracks originated from outside of the track. It can be suggested that at one point of time during the building process of multi-layers, hot cracking initiated from liquation cracking. On the other hand, as identified in this research, the cracking propagation is a combined event of liquation cracking, liquid-backfill healing, and hot tearing, which is the overall mechanism of crack initiation and growth specific to EB-PBF. The dominance of high angle misorientation grain boundaries to assist the crack growth will also be demonstrated. The approximate length of liquation that has been observed has suggested a maximum temperature of reached 2 x 10e6 K/m during the current EB-PBF. This value is close to those predicted by simulations stated in the literature.||en_NZ