Density Optimisation & In-Situ Microstructure Decomposition of As-Built Selective Laser Melted Ti-6Al-4V

Abstract

Additive manufacturing of titanium provides many advantages over traditional methodologies such as the ability to produce complex parts at no additional cost. One such AM method, selective laser melting (SLM) is a powder bed fusion process that has been heavily studied with regards to several metallic materials. One such material, Ti-6Al-4V is the most widely used titanium alloy accounting for more than 50% of worldwide titanium usage due to its good corrosion, biocompatibility, and high strength properties. Prior research into SLM processing of Ti64 has highlighted three primary challenges with regards to mechanical property optimisation. Production of fully dense components has proven difficult whilst the inherent high thermal gradients produce a brittle α’ martensitic microstructure and residual stresses. These challenges are mostly alleviated through costly post process treatments. However a small number of studies have identified process parameter sets that can alleviate residual stresses and decompose the brittle microstructure to a more preferable (α+β) structure that results in superior mechanical properties. The hypothesis of this research is that optimisation of density and microstructure simultaneously in as-built samples will remove the need for post process HIP/annealing treatments. Density optimisation was achieved through a two-stage optimisation process. Firstly, the process environment was characterised through CFD and empirical methods where an increase in meltpool width was correlated to an increase in density. Then, a comprehensive statistical DOE of the laser parameters was completed where laser power and scanning velocity were identified as the most statistically significant variables. An optimal parameter set was identified where fully dense samples were produced and were proven to be repeatable over multiple build jobs. Finally, an increase in density was correlated to increased fracture strain and stress values. Following density optimisation; laser spot size, layer thickness, pre-heating temperature and hatch distance were examined and their effect on density and microstructure reported. Microscopy and XRD analysis revealed in-situ decomposition of the brittle α’ martensitic structure into a Widmanstätten (α+β) structure at high pre-heating temperatures and tight hatch distances. Decreasing laser spot size was statistically significant with reference to both density and microstructure decomposition where conflicting trends were observed. Increasing the spot size increased the density but was detrimental to decomposition. Mechanical properties in the form of microhardness and tensile properties were then examined. All microhardness values were within a small range, indicative of the primarily α/α’ microstructures observed whilst oxidation was also identified as a deterministic factor. Part density, microstructure, residual stresses and oxidation were all identified to influence the tensile properties of as-built parts. Tensile properties akin to those observed within some heat treated samples were produced. However, the optimal parameter set identified within this work was only capable of producing parts comparable to the lower range of heat treated specimens Thus, the hypothesis was rejected. Future work is presented to guide readers on directions for research which could close the gap between as-built and heat treated parts.