Microstructure Evaluation & Process Optimization of Ti6Al4V-xB & Tool Steel Produced by Laser Powder Bed Fusion
thesisposted on 07.10.2020, 20:28 by Yining He
Laser powder bed fusion (LPBF) is one of the most popular metallic AM processes. It produces a 3Dparts by repeated powder layer deposition followed by laser scanning. A major obstacle of LPBF investigation is the difficulty to obtain lab-scale LPBF-powder for compositions of interest. In this
regard, this work used “powder-free approach” to explore the effects of LPBF processing on microstructure and properties. Three alloy systems were investigated, each not widely used in LPBF processing, H13 tool steel, M2 tool steel, and boron-modified Ti6Al4V. Melt pools made without powder were similar in dimensions, geometry, and microstructures to melt pools made with powder. Dimensions were quantified for Ti6Al4V and H13. The balling phenomenon occurred at lower power and velocity combinations when powder was presented. The
single-track test results enabled an empirical model to be developed for laser absorptivity. Chapter 6 presented the work on LPBF process optimization for tool steel. Two steel grades, H13 and M2 steel, were investigated. For H13 steel, a sequential set of studies was performed: single laser
tracks, multi-track pads, and 3D cubes, with 40 P-V settings tested for single tracks. Tracks and pads were made by both no-powder and powder-added laser scan. For M2 steel, no-powder single-track tests were conducted, with 28 P-V settings tested. P-V process mappings of melt pool geometry were developed for H13 and M2 steel. They identify the P-V regions where keyholing, balling and undermelt occurred. The P-V mappings for H13 and M2 steel show different keyholing P-V thresholds, indicating the influence of high-boiling-temperature element on keyholing behavior. The no-powder and powder-added P-V mappings for H13 steel show that the presence of powder layer resulted in a slight shift in the balling P-V threshold.
Microstructure inhomogeneity was observed for both H13 and M2-steel melt pool: For H13 steel, a cellular-network microstructure and an isolated-whisker microstructure co-existed at different regions in the same melt pool. Solute pile-up due to solidification microsegregation was observed at the cellular boundaries, where retained austenite was present. Relative amounts of the microstructure types in each melt pool varied with P-V parameter. Consistency in microstructure inhomogeneity was observed for tracks and pads produced at the same P-V sets. Cracks were found in some pads, which appeared to propagate through the isolated whisker microstructure. Based on this, P-V windows of different cracking tendencies were predicted based on the microstructure P-V map. 3D-cubes built by P-V sets in different windows showed crack densities in line with the predictions. A microstructure evolution mechanism was proposed, which explains the microstructure inhomogeneity as a result of varied microsegregation levels at different regions in the melt pool. Based on the proposed
mechanism, melt pool microstructure was predicted by coupling DICTRA simulation with melt pool thermal profile calculation, showing good agreement to the experimental observation. For M2 steel, microstructure inhomogeneity was induced by columnar-to-equiaxed transition (CET) in the melt pool. Single-tracks for the 28 selected P-V sets showed either fully columnar, columnar+equiaxed mixed or mainly equiaxed dendritic microstructure. A CET criteria was combined with melt pool thermal profile calculation to predict P-V process map for M2-steel microstructure
(columnar/equiaxed). The predicted and experimental P-V maps showed good agreement in the P-V space.
Chapter 7 presented the work on LPBF composition evaluation for five Ti6Al4V-xB trial composition: Ti6Al4V + (0, 1, 2, 5, 10, wt.%) B. Single-tracks, multi-track pads and overlap-track pads were made on the surface of arc-melted Ti6Al4V-xB buttons. For each Ti6Al4V-xB composition, melt pools were produced by 14 P-V sets over wide P-, V-ranges, with melt pool geometry and microstructure information gathered into P-V maps. By varying wt.% B and P-V
parameter, four TiB precipitate morphologies were produced. Melt pool microhardness showed evident enhancement from arc-melted baseline for all Ti6Al4V-xB compositions. Ti6Al4V-xB with 2-5 wt% B as a promising composition range for LPBF processing. The four TiB morphologies observed at different [B%, P-V] combinations were considered as the product of different solidification mode (primary β-Ti, primary TiB and coupled eutectic). An analytical model was developed to determine the criteria of LPBF solidification mode transition as a function of P-V
parameters and Ti6Al4V-xB material properties. The calculated P-V maps for LPBF solidification mode transition showed good agreement with experimental results. This model suggested that certain P-V window leads to coupled-eutectic solidification and non-equilibrium primary-β solidification mode even for highly hyper-eutectic Ti6Al4V-5%B, in turn leading to TiB presenting continuous network and discontinuous network agglomeration of nano-scale TiB whiskers.
DepartmentMaterials Science and Engineering
- Doctor of Philosophy (PhD)