1 Introduction

Ti6Al4V Grade 5 (Ti64) is the most widely used titanium alloy and is well established in additive manufacturing, especially Laser Powder Bed Fusion (L-PBF). Due to the alloy’s high strength-to-weight ratio and excellent corrosion resistance, it finds applications in medical implants and automotive, but most importantly in aerospace—wherever the properties of the material outplay the costs and processing.

Especially in aerospace applications, such as gas turbine engines and adjected components, the strength of Ti64 at elevated temperature is of high interest since its density is nearly half of the otherwise used high-temperature resistant Ni-Superalloys [1, 2]. For classically wrought, forged, and machined Ti64, the maximal service temperature is limited by the poor oxidation resistance at temperatures higher than 400 °C. An increased oxygen intake leads to the formation of α case, loss in ductility and fatigue strength [1, 2]. Therefore, the MMPDS standard suggests its use up to 400 °C [3], as do Lütjering and Williams [2].

However, metals processed by L‑PBF show a different fine-grained, often anisotropic, and microstructure compared to their wrought, forged, and machined counterparts due to build-process direction and high cooling rates [4,5,6,7,8].

He et al. [9] show the influence of anisotropy on the creep behavior at 450 and 600 °C and tensile strength at room temperature after a thermal exposure of 200 h at 500 and 600 °C for furnace annealed α + β specimens and compare it to forged samples for the same thermal exposure. It shows a high anisotropic effect of temperature on ductility and creep, but with strength values higher than forged samples. Viespoli et al. [10] investigate creep [450, 500, and 600 °C] and fatigue behavior at 600 °C for as-build martensitic samples. It shows a high decrement in high cycle fatigue performance, however, creep strength(unclear) on the level of forged billets without any definition of their heat treatment and microstructure. The work of Muiruri et al. [11] investigates the impact toughness of Ti64 Grade 23 from −150 to 250 °C and shows a transition from brittle to ductile behavior, highly influenced by heat treatment applied. Zhao et al. [12] performed selected tensile tests for as-build not stress-relieved samples where the specimens were cut out of bulk material and no further information on the print direction of the specimens was tested. While very brittle, it shows an increase of ductility up to 350 °C and a decrease afterward, reaching the same elongation at break at RT and 400 °C.

Additional creep results for materials processed by L‑PBF can be found for Ni-Superalloys [13, 14]. Due to their complex microstructure, cast materials are superior and the same anisotropy as with Ti-Alloys can be observed. Rao et al. [15] investigated the hot impression creep behavior of AlSi10Mg showing a three times increase in the indentation at 100 °C for solution heat treatment, but there is no comparison to classically produced alloy.

The scope of this contribution is to investigate the tensile strength of Ti64 within the range of allowed operating temperature from RT to 500 °C. This is done for two different heat treatments and microstructural conditions, martensitic α’ and α + β respectively, to investigate possible advantages of the higher strength of α’ at high temperatures, since near α alloys can be used up to 600 °C [1, 2].

2 Methodology

Specimens were built on an EOS Eosint M280 L‑PBF machine equipped with a 400 W Ng:YAg Laser. Process parameters can be found in Table 1.

TABLE 1 L‑PBF process parameters
Full size table

Only virgin powder from a single production batch was used, whose verified properties [16] are shown in Table 2, all within Grade 5 limits according to ASTM B265. This is done to ensure that changes in powder properties within the powder reuse do not affect the results [17, 18].

TABLE 2 Chemical composition, particle size distribution, and morphology of the powder used
Full size table

All specimens underwent post-process stress relief heat treatment (2 h at 650 °C) in an Argon-flooded Lynntherm furnace. Additional furnace annealing (2 h at 800 °C) was carried out in a vacuum furnace to compare properties with stress relief condition. Related microstructures are shown in Fig. 1 and can be found in the previous work of Meier et al. [16], where the same powder batch was investigated. Stress relief led to partially decomposed martensitic microstructure with a small β phase (Fig. 1a), while furnace annealing achieved stabilized and coarser α laths within the β phase (Fig. 1b; [18]).

Fig. 1
figure 1

Microstructure after a stress relief and b furnace annealing [18]

Full size image

Tensile tests at elevated temperature were carried out on a Gleeble 3800 test machine with ohmic heating system. Samples were clamped using conical copper holders to provide high conductivity and temperature is controlled by at least two S‑Type thermocouples welded to the middle (C1) and end (C2 and C3) of the gauge length (Fig. 2).

Fig. 2
figure 2

Cylindrical specimen design for a Gleeble b Zwick Roell, and c Instron test rig with clamping mechanically turned to M10

Full size image

Table 3 shows the matrix of experiment used in this study. For the tensile tests, the samples were heated at 5 K/s and then hold for 1 min before the deformation. Tensile test were controlled by the stroke with a displacement rate of 1 mm/min.

TABLE 3 Test Matrix
Full size table

The specimens’ gauge length was designed according to ASTM 8 as depicted in Fig. 2a, while the clamping was adopted to the Gleeble machine. Due to comparability to ASTM E8, tests at RT, 90°, and 120 °C were performed according to the named standard, at RT on a Zwick Roell test rig (3b), and the others on an Instron 8802/8802K9832 test rig using an Instron CP107081 temperature chamber. Threaded specimens (Fig. 3c) were used. All samples are built in vertical orientation since anisotropy in tensile strength was negligible for the material and ductile anisotropy ought to be similar to the behavior found in [15] and [19].

Fig. 3
figure 3

Stress-strain diagrams for a stress relieved and b furnace annealed samples for all investigated temperatures

Full size image

3 Results and Discussion

Figure 3a, b shows representative stress-strain curves for stress-relieved (3a) and furnace annealed (3b) samples performed on the Gleeble machine.

For both heat treatments, the strength decreases with increasing temperature. In Fig. 4, the ultimate tensile and yield strength of performed tests are compared to standard values found in MMPDS material data sheets as well as tests fully according to ASTM E8 performed with the same material. Both stress-relieved and furnace annealed samples exceed the values for standard material. Out of MMPDS standard, furnace annealed bar and sheet material < 1.72 mm (AMS4928) was selected as it was found to be closest to the tested specimen.

Fig. 4
figure 4

Yield strength (YS) and ultimate tensile strength (UTS) for stress-relieved (SR) and furnace annealed (FA) specimens over the temperature and values found in MMPDS for annealed bar material

Full size image

Stress-relieved samples reach a tensile strength of around 120 MPa more than furnace annealed ones and 260 MPa higher than the values found in MMPDS data sheets. Yield strength for both heat treatments is about 80 MPa below ultimate strength and decreases faster till 400 °C. From 400 °C to 500 °C, they seem to stabilize, while the ultimate tensile strength further decreases. A similar effect can be seen in the MMPDS. Measurements according to ASTM E8 are comparable to those of the Gleeble machine.

The maximal strain reached over raising temperature can be seen in Fig. 5. Both heat treatments show a decrease in ductility for 90 °C from where elongation at break increases up to 200 °C to decrease again with rising temperature. However, these changes are within the deviation, shown in Table 4. While more brittle than furnace annealed from room temperature to 400 °C the stress-relieved specimen show increased plastic break behavior at 500 °C reaching a maximum strain of 0.169, when furnace annealed samples drop further to a maximum strain of 0.128. The elongation reached is in between the expected values for furnace annealed bar (Fig. 5, yellow line) and sheet (Fig. 5, green line) material found in MMPDS data sheets. Both show a strong increase in ductility from 400 °C on, as does max. strain for L‑PBF stress-relieved samples.

Fig. 5
figure 5

Achieved maximal strain for stress-relieved (SR) and furnace-annealed (FA) specimens over the temperature and values found in MMPDS for annealed bar and sheet material

Full size image
TABLE 4 Middle values and deviations for all tests performed
Full size table

The middle values and deviation of all results can be found in Table 4. Deviation for strength is very limited but is higher for the maximum strain reached.

Compared to the results from Zhao et al. [12], both heat treatments show a higher ductility and a smaller difference between yield and ultimate tensile strength. However, both decrease more with increasing temperature. Taking results from stress-relieved specimens, which are most similar to the ones in Zhao et al. [12], their UTS at room temperature is the same at 1200 MPa but just 858 MPa at 400° compared to 1016 MPa. Both show a loss in ductility for 400 °C, but with a nearly double elongation at break within our work (unclear: Zhao’s or your contribution), 14 to 7.8% respectively. This might be because of the unclear orientation or missing stress-relief treatment in Zhao et al. [12]’sstudy.

Because Ti64 has a heat conductivity of just 7.1 W/mK, electric resistance heating leads to an increasing temperature gradient within the specimen. Figure 6 shows the difference between the temperature measured at the thermocouples placed at the beginning and end of the gauge length before the deformation.

Fig. 6
figure 6

Difference in temperature at thermocouple C1 (middle of gauge length) and C2 (end of gauge length)

Full size image

4 Conclusions

The tensile strength reached by L‑PBF Ti6A4V Grade 5 is higher than values in standards for forged and cast materials. Specimens with a partially martensitic microstructure show higher strength than furnace annealed specimens. The ductile behavior is found to be between furnace annealed bar and sheet material with α + β microstructure, well above SAE 4999 limits [20]. While more brittle at room temperature, stress-relieved samples approximate the ductility of furnace-annealed at higher temperatures, making martensitic α’ Ti64 an interesting option at higher temperatures. In-depth studies of oxygen intake of martensitic α’ depending on temperature are advisable as the limit for high-temperature application of Ti-Alloys lies in the oxygen intake and α‑alloys are proven to withstand higher temperatures. Therefore resistance of martensitic α’ Ti64 should be further investigated, since it can be produced by L‑PBF without elaborate heat treatment.

As creep samples show anisotropy in ductility, a future investigation in this direction might be of interest for tensile and Charpy at elevated temperatures.

Due to the thermal properties, the samples at higher temperatures show large gradients in temperature over the specimen. Hence, the development of smaller size specimen standards would be of special interestas it already would be of great use for smaller L‑PBF machines and if the availability of powder is limited.