Fig. 1 shows the microstructure of the DED-processed Ti-Mo sample at the top, middle, and bottom positions under the as-fabricated and heat-treated conditions. It can be seen that the DED-processed sample possesses a mostly dense structure without a clear porosity and cracks. Fig. 1a, c shows that a needle-like martensitic α-Ti appears in the as-fabricated sample, owing to the high cooling rate of the DED system25 at approximately 103–105 K s–1. In addition, martensite α’ has an extremely small size without any preferred orientation. Fig. 1e shows that the bottom region presents a different microstructure with few α phases, which will be further determined using EBSD in the next section. Thus, it can be concluded that the as-fabricated Ti-Mo sample presents an inhomogeneous gradient structure from top to bottom (Fig. 1a, c, e). After the triple cycling heat treatment, the entire sample presents a mostly homogenous microstructural feature, which consists of a matrix phase (β-phase) and a short rod-like or equiaxed α phase. This spheroidization behaviour is attributed to the growth of the α phase during the cycling heat treatment. In addition, there still exists a slight difference between the top and bottom regions. For instance, the width of the α phase increases from 3–4 μm at the top to 5–8 μm at the bottom. It is worth noting that the rod-like α phase at the bottom presents an orientated growth, which is different from the top and middle with a random distribution shown in Fig. 1b, d.
Fig. 1 Microstructure of DED processed Ti-Mo sample at top, middle, and bottom positions: (a, c and e) as-fabricated and (b, d and f) heat-treated conditions. (α, hcp-Ti; α’, martensitic hcp-Ti).
The XRD patterns of the DED-processed Ti-Mo sample in the as-fabricated and heat-treated conditions from altitudes of P1–P8 are presented in Fig. 2a, b. For both conditions, the DED-processed Ti-Mo specimens possess mixed phases of hcp-α and bcc-β without any non-melted Mo phase and oxidation, which is similar to the equilibrium phase diagram of Ti-Mo26. Interestingly, in the case of the as-fabricated sample, the relative intensity of the β-phase peaks increases significantly with a continuous improvement starting from P3. This result indicates a graded phase constitution from α dominant at the top to β dominant at the bottom. By contrast, the heat-treated sample presents a more homogenous phase distribution from the entire scale of the part, which is attributed to the non-equilibrium to equilibrium transition behaviour in rapidly solidified Ti-based alloys during heat treatment27, 28. However, owing to the limitation of using an X-ray diffraction in a quantitative analysis, an EBSD phase analysis is described in the next section as a supplement.
Fig. 2 XRD patterns of DED processed Ti-Mo sample in a as-fabricated and b heat-treated conditions from altitude positions 1–8. (RI, region I for α phase; RII, region II for β phase)
The distribution and morphology of α-hcp Ti and β-bcc Ti in the as-fabricated Ti-Mo sample, which were taken from the top, middle, and bottom regions, are shown in Fig. 3. As shown in Fig. 3a, d, the top region (P1) presents a high content of the α phase of approximately 95% in a submicron-sized morphology. In the middle region (P4), the α phase content decreased from 95% to 75% (see Fig. 3b, d). In the meantime, the α phase shows a large-sized lath-like morphology at the micrometre scale with the β phase located between them. As shown in Fig. 3c, d, bbc-β (70%) is the main phase in the bottom region instead of the α phase at the top. This phenomenon was attributed to two factors: (1) the effect of thermal cycling on the fully melted Mo particles in the Ti matrix during the DED process29 and (2) the high cooling rate during solidification of the bottom region, which leads to the α-β transition24. The effect of the Mo element, as an effective β phase stabiliser, was amplified under non-equilibrium conditions. Thus, the β phase content in the bottom region was much higher than that in the top region.
Fig. 3 Distribution and morphology of α-hcp Ti and β-bcc Ti in DED processed Ti-Mo sample.a P1-top, b P4-middle, and c P7-bottom regions and d statistical data on the phase.
Fig. 4 shows the dual phase (α+β) distributions and morphologies of the heat-treated DED-processed sample from the corresponding positions to the as-fabricated sample (see Fig. 3). Overall, compared with the as-fabricated sample, the heat-treated sample presents a relatively uniform phase constitution, distribution, and morphology at the top, middle, and bottom regions. As summarised in Fig. 4d, the phase constitution of the entire sample showed approximately hcp-α (90%) and bcc-β (10%) phases. Moreover, it can be seen from Fig. 4a–c that no lath-like α phases appear, but nearly equiaxed grains can be observed. It should be noted that a slight decrease in β appears at the bottom region (Fig. 4c), which also possesses a certain lath α phase. This phase transition can be explained using the equilibrium phase diagram26. During the designed triple cycling heat treatment, the solution treatment step possesses a high temperature of approximately 960 ℃, which is much higher than the β-α transition temperature of 800 ℃. With the long holding time and low cooling rate of furnace cooling, a metastable β phase with non-equilibrium is transferred into an α phase with equilibrium. In addition, the importance of Mo diffusion and segregation in the phase transition between α and β cannot be ignored. The results of our previous study17 on the as-DED-processed sample indicate that Mo micro-segregation was observed at the bottom region, which led to a high β phase content. After heat treatment, the phase constitution changed to a state of equilibrium in the phase diagram (Ti-7Mo)26, which is similar to the top region in the as-fabricated sample. Thus, it can be concluded that Mo diffusion appears in the bottom region to eliminate segregation after heat treatment.
Fig. 4 Distribution and morphology of α-hcp Ti and β-bcc Ti in cycling heat treated DED processed Ti-Mo sample.a P1-top, b P4-middle, and c P7-bottom regions and d statistical phase data.
The texture analysis results of α and β Ti in the as-fabricated and heat-treated DED-processed Ti-Mo samples from the top to bottom regions are shown in Fig. 5 and 6. In the case of the as-fabricated condition, the top region presents a random orientation distribution with a grain size of approximately 5 μm and a width of 1–2 μm (see Fig. 5a). As the distance from the top surface increased, the α grain changed from a small-sized lath to large-sized acicular morphology (see Fig. 5b). The length of the α phase reaches 20 μm, which is attributed to the thermal cycling and thermal accumulation induced grain growth behaviour. Moreover, it can be seen that both the top and middle regions exhibit texture-less features. By contrast, both α and β present a clear texture in the bottom region, given the high thermal gradient near the substrate17. Owing to the Burgers orientation relationship in Ti-based alloys, where the close-packed plane of the α phase (0 0 0 1) is parallel to the (1 1 0) plane of the β phase, the highly textured primary β phase induces a high texture of the α phase. Specifically, from the point of example density, the α phase (0 0 0 1) plane at the bottom region exhibits a higher density of 51.64 than that of the top region (32.25)17.
Fig. 5 EBSD orientation maps of α-hcp Ti and β-bcc Ti in DED processed Ti-Mo sample: a P1-top, b P4-middle, and c P7-bottom regions and d scale plate.
Fig. 6 EBSD orientation maps of α-hcp Ti and β-bcc Ti in cycling heat treated DED processed Ti-Mo sample: a P1-top, b P4-middle, and c P7-bottom regions and d scale plate.
Fig. 6 shows the texture analysis results of the heat-treated sample from the top to bottom regions. Compared with the as-fabricated sample, no texture was observed for the entire sample. Moreover, some large equiaxed grains appear instead of lath and acicular grains, as indicated by the black arrows in Fig. 6a–c. Considering the phase transition from β to α during heat treatment, it can be concluded that metastable β (0 0 1) transfers to α (0 1 -1 0). The grains then grow with an obvious reduction in the texture. Given that the phase transition induced a misorientation between α (0 1 -1 0) and α (0 0 0 1), the grain equiaxed growth velocity at the bottom region is smaller than that of the middle and top regions.
Fig. 7 shows the microhardness of the DED-processed Ti-Mo samples under the as-fabricated and heat-treated conditions. In the case of the as-fabricated samples, the bottom region presents a higher hardness (392 HV) than that of the middle and top regions (approximately 280 Hv), which is attributed to the graded structure with a high content of the β phase and fine microstructure at the bottom. After heat treatment, the microhardness significantly increased from the entire sample by approximately 50%, in which the highest hardness also appears in the bottom region. Owing to the microhardness improvement appearing from the entire sample, it can be concluded that the phase transition is not the main reason. Thus, this improvement can be attributed to the spheroidization behaviour of the α phase (see Figs. 1 and 6), which results in a simultaneous improvement in the strength and ductility. The reason behind the spheroidization behaviour is discussed in the discussion section.
Fig. 7 Microhardness of DED processed samples in as-fabricated and heat-treated conditions from top, middle, and bottom regions.
The compression strain–stress curves of the DED-processed samples in the as-fabricated and heat-treated conditions taken from different altitudes are presented in Fig. 8. The compression properties and fracture behaviour are listed in Table 1. As shown in Fig. 8a, for the as-fabricated sample, the top region (P1) possesses low strength but high ductility without a crushing failure. By contrast, P1, the sample taken from the bottom region (P5 and P7) of the as-fabricated sample presents a high strength but low plasticity. Moreover, a crushing failure was observed during the compression test. Fig. 8b shows that a transition stage with a multistep failure appears between the top and bottom regions (see P3 in Fig. 8b). Therefore, it can be clearly seen that the as-fabricated sample possesses graded mechanical properties because of the graded structural features. In detail, the yield strength (YS) increases from 648 MPa to 1074 MPa from the top to bottom regions, which is attributed to the increase in high strength β from the top to bottom regions (see Fig. 3). Meanwhile, the plastic deformation and strength at the cracks show a similar tendency. It should be noted that a crushing phenomenon appears in almost the entire sample with the exception of the high-ductile top region. In the case of the heat-treated sample, the entire sample exhibited uniform mechanical properties such as strength, ductility, and fracture behaviour (see Fig. 8c, d). For instance, the difference between the maximum and minimum values of the YS of samples taken from different altitudes decreased from 39% to 3.9% after heat treatment. Only the bottom (P7) region exhibited a crushing failure. This homogenisation in mechanical properties is significantly influenced by the mitigation of the microstructure gradient after cycling post heat treatment, which is described in the discussion section.
Fig. 8 Compression strain–stress curves of DED processed samples in a and b as-fabricated and c and d heat-treated conditions.b and d partial enlarged drawing of a and c.
Sample Number Yield strength (MPa) Plastic deformation(%) Strength at cracks(MPa) Crush failure or not AF HT AF HT AF HT AF HT P1 648.4 851.5 Non Non Non Non N N P3 942.6 836.8 9.8 Non 1325.6 Non Y N P5 1041.2 887.8 11.2 Non 1498.4 Non Y N P7 1074.6 886.4 14.3 27.3 1569.8 1564.7 Y Y
Table 1. Compressive properties and fracture behaviour of DEDed Ti-Mo alloys under as-fabricated (AF) and heat-treated (HT) conditions.