Role of geometry on properties of additively manufactured Ti-6Al-4V structures fabricated using laser based directed energy deposition

Highlights

• Geometry of the deposited structure impacted the resulting tensile strengths and elongation.

• Orientation within the structure impacted tensile strengths and elongation.

• Tensile strength decreased with increasing height above the substrate.

• Resulting mechanical properties correlated with the average prior beta grain intercepts.

Abstract

A series of Ti-6Al-4V wall structures were additively manufactured (AM) using directed energy deposition (DED) with similar processing parameters and build paths to investigate the role of geometry on the resulting as-deposited microstructure and mechanical properties. While the aggregated tensile strengths (1049 ± 37 MPa), yield strengths (936 ± 43 MPa), and elongations (18 ± 4%) were relatively consistent, a more in-depth statistical analysis revealed statistically significant relationships between the resulting mechanical properties and the orientation with respect to the build direction. Tensile samples with the long dimension parallel to the substrate exhibited a higher average tensile strength than samples with the long dimension perpendicular to the substrate. In addition, the tensile strengths from thick multi pass wall structures were significantly higher than thin single pass wall structures. Finally, the tensile strengths decreased with increasing height above the substrate within the wall structures. Most of the observed differences in mechanical behavior can be attributed to differences observed in the average prior β grain sizes and shapes that impact the amounts of boundary strengthening within the structures. In addition, qualitative differences within the microstructure were observed at different locations within individual builds and correlated with changes in tensile strength.

Introduction

Additive manufacturing (AM) encompasses a range of processes in which discrete components are fabricated in a layer by layer manner [1], [2]. In metallic material systems, AM can produce fully dense parts with highly complex geometric features from a range of different alloys and material combinations [3], [4], [5]. There has been a longstanding and continuing effort involving the AM fabrication of titanium alloy components using both directed energy deposition (DED) and powder bed fusion (PBF) AM processes [6], [7], [8]. In order for AM fabricated components to be put into service, particularly for critical applications, certification protocols need to be developed [9], [10]. A major component of this certification process involves the development of accurate material property databases.

The processing conditions during AM can vary significantly as different component geometries and sizes are produced, creating complex process-structure-property relationships unique to AM processes. The eventual certification of AM components will require a material property database with low levels of uncertainty for mechanical properties of interest, such as yield and tensile strength, elongation, and Young's modulus. Given the interest in AM fabricated Ti-6Al-4V components, there is a rather large amount of mechanical property data available in the literature. A wide range of powder bed fusion (PBF) and directed energy deposition (DED) processes have been investigated to include PBF using electron beam melting (EBM) [11], [12], [13], [14], [15], [16], [17], [18], [19], PBF with selective laser melting (SLM) [3], [14], [15], [17], [20], [21], [22], [23], [24], [25], [26], [27], [28], DED using laser with a powder feedstock [29], [30], [31], [32], [33], [34], [35], [36], DED laser with a wire feedstock [37], [38], [39], [40], and DED using a gas tungsten arc (GTA) energy source with a wire feedstock [37], [38], [41], [42], [43], [44]. The static mechanical properties for Ti-6Al-4V components fabricated using these different AM techniques as a function of the reported tensile strength and elongation values are plotted in Fig. 1(a).

As shown in Fig. 1(a), there is wide variability in the reported mechanical behavior of AM Ti-6Al-4V components with average tensile strengths ranging from 775 MPa to 1270 MPa and average elongations from 1% to 25%. This wide variability in mechanical behavior manifests itself not only between different AM processing techniques but also within the same processing techniques. For example, the laser based PBF process typically exhibited the highest tensile strengths, that averaged from 930 to 1310 MPa, and elongations that averaged 1 to 11%, while the electron beam based PBF process exhibited a wider range of properties compared to the other processes, with tensile strengths averaging from 775 to 1200 MPa and elongations averaging from 2 to 25%. While these previous studies have covered a wide range of processes, minimal information on the processing conditions has been included, and they have focused primarily on the characterization of mechanical properties from test specimens formed directly or machined from simple blocks or wall structures. Even though the processing conditions should be nominally the same, significant variability in the mechanical properties is still observed.

This considerable variability in the mechanical properties reported in the literature even extends to single processing techniques. For example, Fig. 1(b) plots the reported tensile strengths as a function of elongation from previous investigations for a range of DED processes. Although all of these results are obtained from nominally the same DED process, there is a wide variability in the reported processing conditions including the energy source, the feedstock type, and the processing parameters. DED fabricated components with a powder feedstock exhibited higher tensile strengths that averaged from 1025 to 1170 MPa [29], [30], [31], [32], [33], [34], [35] than components fabricated with a wire feedstock for both GTA and laser based processes that averaged from 870 to 1055 MPa [37], [38], [39], [40], [41], [42], [43], [44]. The processing parameters are dependent on feedstock type, and typically a powder feedstock uses lower power levels than wire feedstock to fabricate similar components. In addition, ductility also varies considerably from the DED investigations, with elongations ranging from 3 to 20%. This variability could be the result of unseen or unreported porosity from either gas entrapment or from lack-of-fusion between deposition layers which can have a detrimental impact on ductility within AM components [20], [45], [46].

Since little information is presented on the processing parameters for previous AM investigations, there is a limited understanding of the mechanisms driving this wide variability in tensile properties across different studies. Although differences would be expected for different feedstocks and processing parameters, there is wide range of values in the reported mechanical properties even within individual investigations when the processing conditions are nominally similar. This variability could be the result of differing thermal histories experienced within the component during AM fabrication leading to location dependent differences. During processing, each location within the AM component undergoes a complex thermal history which includes melting, solidification, and thermal cycling [47]. The overall lack of the consideration of processing history for mechanical properties reported in the literature, whether in terms of processing parameters or geometry, has led, at least in part, to high levels of uncertainty in the mechanical properties needed for the design of AM components.

The eventual certification of large, complex Ti-6Al-4V AM components will require a more complete understanding of the governing process-structure-property relationships. The fabrication of even simple wall structures can be complicated by changes in the build path, leading to changes in mechanical properties at different locations and orientations which cannot be easily predicted. Although previous work has identified some of these relationships [31], [37], [39], [41], [42], [43], they were not fully investigated to determine their impact on the resulting mechanical behavior. One area, in particular, which has not been widely investigated is the role of part geometry and how changes in geometry impact the resulting mechanical properties within a component.

In order to begin building a material property database for specific geometries, several wall structures with different wall thicknesses, heights, and shapes were fabricated using a laser based DED process with consistent processing parameters and build paths. The processing conditions used in this investigation incorporated high laser powers and deposition rates, allowing for fabrication of large components in a reasonable amount of time. The relatively large size of these wall structures allowed the impact of geometry on the microstructures and mechanical properties to be investigated. Static mechanical properties and the corresponding microstructural features were analyzed for two orientations in relation to the build direction and at a range of locations throughout each wall structure. A statistical analysis was performed to determine the impact of the various geometrical features on the mechanical properties. Regardless of the different wall designs and wall thicknesses, the resulting mechanical properties obtained across all geometries were relatively consistent and well above the lower limits for cast and wrought Ti-6Al-4V components. However, this investigation showed that orientation, location, and geometry displayed statistically significant dependencies on the resulting mechanical properties. Investigating wall structures fabricated under similar processing conditions but encompassing different geometries will help increase the knowledge base leading to eventual certification of AM components for critical applications.