2021 Volume 62 Issue 8 Pages 1071-1078
Wire + Arc Additive Manufacturing (WAAM) is an advanced manufacturing technology by inexpensive Gas Tungsten Arc Welding (GTAW) technology. Key microstructural features of the as-built WAAM alloy include large columnar β grains, grain boundary α colonies, and Heat Affected Zone (HAZ) banding, which generally leads to low ductility and anisotropy. In this study, Ti–6Al–xV (x = 0, 2, 4) alloys were prepared by WAAM, the effects of vanadium content on the microstructure, tensile properties and impact toughness were investigated. Irregular-shaped, plate-like features without columnar grains and HAZ banding were obtained in Ti–6Al alloy. Columnar grains were observed in Ti–6Al–2V alloy, and the grain size was further enlarged to more than ten millimeters by 4 mass% vanadium addition. With the increasing of vanadium content, a monotonic increase in yield strength and ultimate tensile strength can be observed, while the fracture strain and impact toughness changed in the opposite trend. Ti–6Al and Ti–6Al–2V alloy exhibited better matching of strength, ductility and impact toughness compared with Ti–6Al–4V alloy.
Titanium and its alloys have been widely used in aerospace and marine fields due to its low density, high specific strength, excellent corrosion resistance weldability. Ti–6Al–4V is the most widely used titanium alloy as α + β type alloy, owing to its excellent comprehensive properties.1–3) Additive manufacturing technology has been a concern for the modern titanium industry because it provides an alternative to manufacture complicated components in near net shape.4) Among the additive manufacturing techniques, the GTAW-based wire arc additive manufacture has the most potential for building large-scale titanium alloy parts in the marine field because of its high deposition rate, low equipment investment, and low running cost.5,6) The relatively slow cooling rate of 10∼20 K/s during the β/α transformation in WAAM as-built components produced relatively large grain size and low residual stress that offers a balance between strength and ductility.
Numerous investigations have been reported concerning the microstructure and mechanical properties of Ti–6Al–4V manufactured by WAAM. For a Ti–6Al–4V component produced by WAAM, the microstructures were complex, often varying spatially within the deposition. Wang et al.7) reported that the macrostructure of WAAMed Ti–6Al–4V alloy was characterized by coarse epitaxial columnar prior-β grains growing from the substrate but fine lamellar microstructure within the prior-β grains. Baufeld et al.8) stated that the top region of the WAAM Ti–6Al–4V alloy deposit displayed very fine lamellae α, in contrast, the microstructure at the bottom region consisted of much thicker lamellae α.
In general, additively manufactured Ti–6Al–4V by WAAM has equal strength compared to wrought counterparts,9) but its plasticity and toughness need to be substantially improved for structural applications. The inadequate ductility and toughness are caused by undesired acicular-α or martensites in columnar prior-β grains because of their low capacity to tolerate crack initiation and propagation.10) Zhang et al.11) demonstrated that the difference in crack growth rate and pattern in the substrate and WAAM was attributed to the different microstructure characteristics.
A lot of literature has attempted to increase acicular-α lath size and improve ductility. Post heat-treatments enhance the ductility while reducing the yield and ultimate tensile strengths. Bermingham et al.12) performed systematic heat treatment on Ti–6Al–4V produced by WAAM and found that a short heat treatment at 653 K for 2 h improved the ductility by over 30% while not coarsening the grain size. Wang et al.13) found the α laths were coarsened after heat treatment 850°C/2 h/AC, while the strength was decreased and elongation was increased. Xu et al.14) reported that high globularization fraction of α phase in microstructure of Ti-17 alloy could decrease impact toughness.
Another efficient method to further improve the mechanical properties of WAAMed alloy is modifying the alloy composition. Vanadium is an β phase stabilizer, and its segregation in the β phase or martensite may introduce more deformation mechanisms and different properties.15) It is interesting to study the influence of vanadium additions on the microstructure and properties. Collins et al.16) pointed out that when the vanadium content increased, there was a substantial decrease in the average width of the α-laths in graded Ti–xV binary alloys produced by laser engineered net-shaping (LENS™) process. Banerjee et al.17) found the thickness of the grain boundary decreased with increasing vanadium content, for vanadium contents <2 at%, martensitic α plates or ribs decorated the inter-lath region of primary Widmanstätten α laths. Liang et al.18) explored effects of vanadium content (from 0%wt–7%wt) on microstructure and mechanical properties of hot rolled TiZr based alloy. The results showed that the phase composition ranged as α′ → α′ + (α′′ + β) → α′′ + β → β + (α′′) → β, yield strength and elastic modulus showed a trend from decline to rise with the increasing of vanadium content, while the elongation showed the reverse tendency.
Although several investigations have been conducted on the microstructure and mechanical properties of WAAMed Ti–6Al–4V alloy, research on the relationship between vanadium content and mechanical properties was relatively less. Moreover, there was still a lack of study on impact toughness parameters of WAMMed alloys. The aim of the present work, therefore, was to quantify microstructure evolution and the mechanical behavior of Ti–6Al–xV alloys produced by GTAW-based WAAM. The accuracy analysis provided a deep understanding between chemical composition, microstructure, and mechanical properties of WAAMed alloys. Secondly, the deformation mode in quasi-static and impact conditions was investigated, as it was widely known that strength and impact toughness were two of the main criteria to evaluate titanium components.
The experiment used Fronius Magic Wave 3000 welder and KUKA KR 16-2 robot. Wires with 1.2 mm diameter were used to deposit 3 straight walls with dimensions of 110 mm × 110 mm × 25 mm, which was locally shielded by 99.99% purity argon during manufacturing. Table 1 showed the chemical composition of the wires. The substrate was a Ti–6Al–4V plate that conforms to ASTM B265 specification with 25 mm in height, 110 mm in length, and 60 mm in width. The processing parameters employed were shown in Table 2.
For microstructural analyses, optical and scanning electron microscopes (SEM) were employed on mirror polished and etched samples using the Kroll’s reagent (100 ml distilled water, 5 ml nitric acid, and 2 ml HF). The phase constitution was characterized by X-Ray Diffraction (XRD) (Rigaku, SmartLab SE, Cu Ka radiation, scan step of 0.017°, counting time of 4 s). Grain orientation information of the raw material and the deformed samples was determined by SEM (JSM-7900F) equipped with electronic back-scattered diffraction (EBSD, Hikari XP) technique. Inverse Pole Figure (IPF) maps indicating the longitudinal sample direction in the crystal reference frame were acquired. The local textures were calculated from the large set of orientations by the superposition of Gaussian functions assigned to each measured orientation. The width of each Gaussian function was fixed to 10°. The results were displayed by means of characteristic pole figures (PFs).
Mechanical samples were cut from the WAAM Ti–6Al–xV walls in the direction indicated in Fig. 1. Tensile tests were performed on circular tensile samples piled-up through the plate thickness and referenced as (TD-ND) with TD the solicitation direction and ND the normal to the specimen’s thickness. An Instron tensile machine was used with a strain rate of 10−3/s. The Charpy V-notch specimens with 10 mm × 10 mm × 55 mm were cut from fracture toughness samples. Impact toughness was conducted by using the E22.452 instrumented Charpy impact testing machine.
Schematic diagram of WAAMed Ti–6Al alloy sample.
The microtexture and the local texture of the broken Charpy or tensile specimen were characterized by EBSD. The orientations were measured in a longitudinal cut of the broken specimen after adequate surface polishing. Additional EBSD acquisitions with a smaller beam step of 2 µm were run closer to the fracture surface.
In Ti–6Al alloy, the α grains exhibited somewhat irregular-shaped, plate-like features, as shown in Fig. 2(a) and (d). By contrast, WAAMed Ti–6Al–2V alloy (Fig. 2(b) and (e)) and Ti–6Al–4V alloy (Fig. 2(c) and (f)) were characterized by the columnar prior-β grains growing from the substrate to throughout the deposited layers. The macrostructure of Ti–6Al–2V and Ti–6Al–4V samples could be divided into three regions from the bottom to the top: the bottom region with fine grains, the middle region with layer band, and the top region without layer bands. The layer bands could be attributed to multiple thermal cycles;19) therefore, the top region lacks due to sufficient thermal cyclic times. The prior-β grain size of Ti–6Al–2V and Ti–6Al–4V was quite large (exceeding several millimeters in some cases), which results from the large thermal input of the WAAM process, the slow cooling rate during solidification (90–100 K s−1), and potential grain growth during subsequent layer deposition.
Optical micrograph of Ti–6Al–xV alloys: (a), (b), (c) macroscopic morphology of the Ti–6Al, Ti–6Al–2V and Ti–6Al–4V alloy, respectively; (d), (e), (f) micrograph of the Ti–6Al, Ti–6Al–2V and Ti–6Al–4V alloy, respectively.
From the comparison of the morphology of three materials (Fig. 2), it can be inferred that vanadium contents significantly changed the morphology and size of the prior-β grains. Ti–6Al–2V alloy exhibited Widmanstätten structure, the size of prior β grain in Ti–6Al–2V was several millimeters, which was equivalent to the as cast titanium alloys. The grain size of Ti–6Al–4V was further enlarged to more than ten millimeters. Macroscopic banding, corresponding to each layer height, was also observed in Ti–6Al–2V and Ti–6Al–4V samples.
For the samples of Ti–6Al–2V alloy and Ti–6Al–4V alloy, another variation in microstructure was the generation of lamellae α and acicular α′ structure due to the differences in vanadium contents. For Ti–6Al–2V alloy, lamellar α and retained β laths were arranged alternately in larger β grains, as shown in Fig. 3(a). In contrast, the acicular α′ occupied most of the proportion of Ti–6Al–4V alloys, which developed into the network structure (Fig. 3(b)).
Lamellar α, acicular α′ and prior β grain boundaries in Ti–6Al–2V alloy (a) and Ti–6Al–4V alloy (b).
Figure 4 showed X-ray diffraction patterns of Ti–6Al–xV alloys. Although WAAM fabricated Ti–6Al–xV alloys exhibited similar constituent phase, the relative fractions of the phases are still different; the data reveal that ternary alloys consisted of bcc β and hcp α solid solutions, while binary alloy exhibited single-phase. It should be pointed out that only a small amount of β-Ti phases were observed in the Ti–6Al–4V sample, which was attributed to the fact that the β-Ti phase was usually too minor or absent to be detected by X-ray.
X-ray diffraction patterns of Ti–6Al–xV alloys.
Figure 5 presented IPF maps obtained by EBSD from cross-sections perpendicular to the building direction, from locations 50 mm below the top of the Ti–6Al–xV alloy. Like the optical micrograph, irregular-shaped, plate-like features were obtained in Ti–6Al alloy; some coarse grains, which we defined as grains larger than 100 µm, could be seen as surrounded by high angle boundaries.
IPF maps (a), (b), (c) obtained by EBSD from Ti–6Al, Ti–6Al–2V and Ti–6Al–4V alloy, respectively.
For Ti–6Al–2V and Ti–6Al–4V alloy, the prior-β grain sizes were too large to obtain sufficient data to develop a full texture description. The samples present the typical basket weave microstructure with an average grain size of several micrometers. The boundary of the columnar β structure was shown clearly.
The selection of the test areas and the pole figure results were shown in Fig. 6. The epitaxial directional solidification in the molten pool caused α grain morphological texture and a crystallographic texture. The α-laths in one colony shared the same orientation and consequently formed one grain due to the low-angle boundaries between an α-lath and its neighboring laths.20) The maximum density values of Ti–6Al alloy, Ti–6Al–2V alloy and Ti–6Al–4V alloy were 17.92, 24.16, and 29.12, respectively. With the decrease of vanadium content, the texture of the α phase in the {0001} plane became weak and more random; the Ti–6Al–4V specimen presented the strongest texture.
Pole figures of α textures in Ti–6Al–xV alloys obtained from areas of 5 × 5 mm2 EBSD maps.
Another key microstructural parameter was the prior-β grain size because this limits the size of colony-α structures that can form. When Ti–6Al–4V was cooled through the β-transus, clusters of lamellar α belong to the same crystallographic variant could form colonies that grow in size until they were impeded by other a variants growing within the parent β-grain or until impeded by the β grain boundary itself. Thus large β-grains had the potential to foster large α-colonies. Since the colony size determines slip length, it was desirable to avoid the formation of large β grains during solidification.
Generally, columnar grains with lamellar α phase were often observed in WAAMed near α alloy and α + β alloy. This phenomenon was systematically described by Wang et al.7) It was pointed out that because of the low potential for developing a significant undercooling ahead of the growth front and the absence of heterogeneous nuclei in the melt, the transition from a columnar to an equiaxed grain solidification is, therefore difficult.21)
In Ti–6Al alloy, a uniform microstructure consisting of α grains with irregular-α shaped, plate-like features was observed. This may be due to that the Ti–6Al alloy does not contain β-stable element. Vanadium preferred to enriched in the β phase.22) This alloy element partitioning effect had the consequence that the α lamellar formed in the β grains upon cooling from the recrystallization annealing temperature.
Macroscopic banding, corresponding to each layer height, was also observed in vanadium containing alloys but was not seen in Ti–6Al alloy. Ho et al.19) reported that banding could be attributed to developing a transient solute boundary layer at the solidification front. Weak microsegregation, or iron and vanadium’s coring, were also observed in the banding.
Vanadium was widely used in near α, α + β, β type titanium alloys as a β phase stabilization element.23,24) In WAAMed Ti–6Al–xV alloys, the addition of vanadium significantly changed the microstructure. Columnar grains appeared and the size of prior β grains significantly enlarged with the increase of vanadium content. The addition of the vanadium element significantly reduced the phase transformation temperature. In the thermal process of WAAM, materials entered the β temperature range several times, and titanium grains were easy to grow in the temperature above the β range.
3.2 Mechanical properties 3.2.1 Tensile propertiesFigure 7 displayed the true stress–true strain curves of the alloys at room temperature, the curves presented similar shapes. The variation trend of strain could be divided into three stages: initial straight-line elastic stage, nonlinear plastic stage, and final failure stage. With the increasement of vanadium content, a monotonic increasement in yield strength and ultimate tensile strength could be observed, while the fracture strain changed in the opposite trend.
Tensile curves of Ti–6Al–xV alloy.
The tensile properties for each alloy were averaged in Table 3. Ti–6Al alloy exhibited the highest elongation and the lowest strength. Compared with Ti–6Al–4V alloy, the 0.2% yield strength of Ti–6Al alloy decreased by about 19.4% and the tensile strength decreased by about 17.1%. It was interesting that elongation of Ti–6Al alloy determined as 16.5%, which was 43% higher than that of Ti–6Al–4V alloy, probably due to the interaction between the solute vanadium atoms and dislocations.
Figure 8 further showed typical impact load-deflection curves of Ti–6Al–xV alloys. The samples’ total absorbed energy increased from 44.60 J to 65.74 J, with the vanadium content decreased from 4 mass% to 0 mass%. The impact toughness of Ti–6Al alloy was 65.74 J, which was about 12.9% and 47.4% higher than that of Ti–6Al–2V and Ti–6Al–4V alloy, respectively.
Typical load-deflection curves for Ti–6Al–xV alloys.
The total absorbed impact energy could be divided into crack initiation energy and crack propagation energy in terms of the load-deflection data’s peak load. Table 4 illustrated the variation of crack initiation and crack propagation energy of the alloy during the impact toughness test.
All the samples had a reduced crack initiation part with similar EI values, while the load at the crack propagation stage deceased much slower, like the ductile fracture. The samples possessed a relatively small fracture initiation part and occupied about 30%∼44% of the total absorbed energy. It was common that the crack was initiated without a large energy absorption, while the crack propagation behavior differed significantly from specimen to specimen. The Strength and toughness are generally mutually exclusive, Ti–6Al alloy exhibited the highest propagation energy and the lowest strength because of the low alloy concentration. Due to the combined effect of low alloy concentration and lamellar toughening, Ti–6Al–2V alloy exhibited high propagation energy. The impact absorption energy of Ti–6Al alloy was the highest, including the initiation energy and crack propagation energy. In contrast, the peak load was gradually increased with increasing vanadium content, consistent with the strength variation.
Figure 9 showed the fracture morphologies of the Charpy impact specimens. The crack zone could be classified by crack initial zone (CIZ), crack propagation zone (CPZ), shear-lip zone (SLZ), and final propagation zone (FPZ) according to the distance from the notch tip. Generally, in the crack initial zone, the large and shallow ductile-dimpled fracture or cleavage facets and tear ridges were mostly observed (Fig. 9(a)). The CPZ of each specimen showed the same characters, including dimples and voids.
Fracture morphologies of the Charpy impact specimens: (a), (d), (g) macroscopic morphology of the Ti–6Al, Ti–6Al–2V and Ti–6Al–4V alloy, respectively; (b), (e), (h) CPZ of Ti–6Al, Ti–6Al–2V and Ti–6Al–4V alloy, respectively; (c), (f), (i) CIZ of Ti–6Al, Ti–6Al–2V and Ti–6Al–4V alloy, respectively.
A large number of dimples distribute on the whole CPZ surface of each specimen, which indicates an excellent toughness, but the area fraction of large and shallow dimples decreased (Fig. 9(b)) with the increase of vanadium content, the fracture morphologies of the specimens showed more features of cleavage facets than that of the specimen. Ti–6Al specimens exhibited higher Charpy absorbed energy, which was likely related to the changes of fracture morphologies.
Figure 10 displayed the optical micrographs of the local crack propagation paths in Ti–6Al–xV alloy samples. Generally, the coarser crack front geometry indicated greater crack growth resistance and larger impact toughness. However, from the microstructure of the fractured samples’ side faces, similar crack front geometry was observed. It was implied that the difference in impact toughness was mainly attributed to the alloy composition. Moreover, in Ti–6Al alloy, deformation twins and some twin intersections were displayed.
Optical micrographs of the local crack propagation paths in Ti–6Al–xV alloy samples: (a), (c), (e) macroscopic crack propagation path of Ti–6Al, Ti–6Al–2V and Ti–6Al–4V alloy, respectively; (b), (d), (f) detailed crack propagation path of Ti–6Al, Ti–6Al–2V and Ti–6Al–4V alloy, respectively.
To further investigate the deformation mechanism, EBSD maps of the local crack propagation paths in Ti–6Al–2V and Ti–6Al–4V samples were shown in Fig. 11. Severe plastic deformation occurred in the near fracture subsurface regions. The transgranular fracture was dominant in both alloys. The primary β grain boundary had little effect on preventing crack propagation. Similar to Ti–6Al alloy, twins appeared in the deformation region of Ti–6Al–xV alloy.
EBSD maps collected along the crack path: IPF (a) map and misorientation map (b) of Ti–6Al–2V alloy, IPF map (c) and misorientation map (d) of Ti–6Al–4V alloy.
Figure 12 displayed a summary of the strength versus elongation for the WAAMed titanium alloys.25–29) Compared with the WAAMed Ti–6Al–4V alloy, the mechanical properties of the as printed Ti–6Al–xV alloys displayed a superior combination of strength and elongation. The current work has produced WAAMed titanium alloy with the highest elongation.
Tensile properties of WAAMed Ti–6Al–xV alloys compared with Ti–6Al–4V fabricated by additive manufacturing.
Toughness and strength were equally crucial for safety-critical structures;30,31) unfortunately, these properties were generally mutually exclusive. However, there was little research on the impact toughness of WAAMed titanium alloy. W.A. Grell et al.32) reported that the impact toughness of electron beam melting Ti–6Al–4V–0.11O alloy was about 50 J, which was similar to the results in this study. Ti–6Al and Ti–6Al–2V alloy exhibited better matching of strength and impact toughness than Ti–6Al–4V fabricated by various processes, as shown in Fig. 13.33)
Strength and impact toughness of Ti–6Al–xV alloys compared with forged or hot rolled Ti–6Al–4V alloy.
Generally, vanadium as a beta stable alloy element can improve the material’s strength, but the toughness will decrease. On the other hand, it was well known that Widmanstatten structures exhibited greater toughness comparing with equiaxed structure; coarsening of Widmanstatten structure was very useful to achieve higher toughness in titanium alloys. All these three alloys exhibited transgranular fracture; crack did not propagate along the crystal boundary, indicating that the lamellar structure did not play an essential role in hindering crack propagation, consistent with the instrumented impact tests. Meanwhile, the comparatively large plastic deformation of the irregular-α structure consumed impact energy during the impact process, as confirmed by fracture morphology.
To summarize, vanadium solid solution and lamellar toughening were two competitive mechanisms in controlling the impact fracture process, depending on the vanadium content. It was also anticipated that further improvements in a range of properties could be achieved by modifying the composition.
WAAMed Ti–6Al–xV alloys were prepared, and the microstructure and mechanical properties were systemically investigated.
Irregular-shaped, plate-like features were obtained in Ti–6Al alloy without columnar grains, significantly different from WAAMed Ti–6Al–4V. With the decrease of V content, the texture of the α phase in the {0001} plane became weak and more random.
Reduced vanadium content resulted in improved elongation, impact toughness, and strength reduction due to vanadium solid solution and lamellar toughening. The impact toughness of Ti–6Al alloy was 47.4% higher than that of Ti–6Al–4V alloy, the difference in impact toughness was mainly attributed to the alloy composition. Ti–6Al and Ti–6Al–2V alloy exhibited better matching of strength, ductility and impact toughness compared with Ti–6Al–4V alloy.
All these three alloys exhibited transgranular fracture, the crack did not propagate along the crystal boundary, indicating that the lamellar structure did not play an essential role in hindering crack propagation.
