Determination of the fatigue properties of Ti6Al4V (ELI) parts built by a direct metal laser sintering system with standard process parameters followed by post-processing treatments

Abstract

Design of hard tissue medical prostheses is typically based on the mechanical properties of the materials used. However, apart from non-loadbearing implants, medical prostheses in use are exposed to various forms of dynamic loading. Therefore, to eventually deliver qualified medical prostheses produced through additive manufacturing (AM), it is necessary to develop a data bank on both their static and dynamic properties. This is done here with reference to the fatigue properties of Direct Metal Laser Sintering (DMLS) of Ti6Al4V (ELI) parts, produced at the Centre for Rapid Prototyping and Manufacturing (CRPM) of Central University of Technology, Free State. The effect of flaws inherent in DMLS-produced specimens and their microstructure on the mechanical properties for as-built and high-temperature annealed specimens is investigated here.

From literature, orthotropy in DMLS Ti6Al4V (ELI) specimens, with reference to the three mutually orthogonal DMLS build directions, has been reported. It was attributed to the presence and orientation of the DMLS process-related pores within DMLS Ti6Al4V (ELI) specimens to their loading axis, the presence and direction of the residual stresses within the DMLS Ti6Al4V (ELI) specimens to their loading axis, and the direction of the prior beta (𝛽) grains of DMLS Ti6Al4V (ELI) specimens to the loading axis. These are all factors of growth of the layers of the DMLS process in relation to the three mutually orthogonal DMLS build directions. Therefore, in this project the first step was the non-destructive testing of DMLS Ti6Al4V (ELI) specimens for the presence or lack of orthotropy. It was done by the ultrasonic testing of rectangular bars of as-built Ti6Al4V (ELI) of dimensions 60 x 11 x 11 mm built to align with X-, Y- and Z-axes of the DMLS machine, where Z is the build direction and X is the direction of movement of the recoater blade. From the data obtained, the tested DMLS Ti6Al4V (ELI) specimens showed to be both homogeneous and isotropic.

The high-cycle fatigue (HCF) properties of as-built and high-temperature annealed (HTA) DMLS Ti6Al4V (ELI) machined and polished specimens built to align with the three mutually orthogonal machine axes were investigated. This was carried out by cycling the specimens under load control, in a tension–tension fatigue testing machine. From the data collected, semi-log graphs of maximum stress (S) against life (N) of the specimens produced along the respective three build directions were plotted, and the displayed endurance limits compared. The as-built specimens aligned with the Y- and Z-axes showed the same endurance limit of 485 MPa and it was slightly higher than the 450 MPa endurance limit of the specimens aligned with the X-axis. Whereas the HTA specimens aligned with the X- and Y-axes showed the same endurance limit of 450 MPa, the specimens aligned with the Z-axis showed a slightly higher endurance limit of 486 MPa. Optical and scanning electron microscopy of the fracture surfaces were used to compare and analyse the crack initiation and propagation characteristics of the specimens. These showed that the dominating crack initiators were DMLS process-related pores. The HTA DMLS Ti6Al4V (ELI) specimens were micro-CT scanned in an attempt to relate the pores in the specimens with their fatigue properties. The micro-CT pore information from suspected crack initiation pores on the surfaces of eventually fractured specimens was used to calculate stress intensity factors, which correlated well with the decreasing cycles to failure of the fatigue test specimens for all three build directions. Three representative specimens were analysed and the “killer pore” identified in each micro-CT scan and fractographs, all which were proximal to the surface of the specimen.