2022 Volume 63 Issue 7 Pages 1087-1096
Ti–27.5Al–13Nb is an intermetallic alloy that incorporates the orthorhombic Ti2AlNb phase (O phase) and the α2 phase, and was previously developed by the authors for high-temperature applications. The aim of the present study was to develop less expensive orthorhombic alloys. For this purpose, a portion of the cost-prohibitive Nb in this baseline Ti–27.5Al–13Nb alloy was replaced with an equivalent amount of Fe and/or Mo, yielding three new derivative alloys: Mo-replaced Ti–27.5Al–8.7Nb–1Mo, Fe-replaced Ti–27.5Al–5.5Nb–1Fe and (Mo and Fe)-replaced Ti–27.5Al–4.9Nb–1Mo–0.5Fe. Further, a minor amount of B (boron), i.e., 0.1 pct B, was added to these derivative alloys and baseline alloy. The microstructures and corresponding tensile and creep properties were examined in four cases: B-free and B-modified alloys with fully lamellar microstructures, and B-free and B-modified alloys with duplex microstructures. With the addition of 0.1 pct B, the prior B2 grain size of each ingot was drastically reduced by about one order of magnitude, from 600∼1000 µm for the B-free alloy to 100∼250 µm, and thereby a refined fully lamellar microstructure was obtained. However, this high degree of grain refinement did not markedly improve ductility at room temperature. A duplex microstructure consisting of a globular α2-phase and a lamellar microstructure led to the improved ductility and tensile strength. The addition of B exerts either a positive or negative effect on the creep properties depending on the compositions and microstructures, creep test temperature and applied stress. The Mo-replaced alloys had better creep properties than the baseline alloy, whereas creep properties of the Fe-replaced and (Mo and Fe)-replaced alloys were considerably inferior to those of the baseline alloy. Among the four alloys, the 0.1 pct B-modified Mo-replaced alloy with a duplex microstructure exhibited the highest ductility of 4.7 pct at room temperature and higher tensile strength up to 1073 K and better creep properties than the other two derivative alloys and the baseline alloy.
Fig. 10 High-temperature specific tensile strength for various high-strength metallic and composite materials.45) The present data are included for comparison.
Titanium intermetallic alloys based on the Ti2AlNb phase have attracted attention as potential materials for high temperature applications due to their higher specific strength and fracture toughness compared to γ-based alloys.1) Since the Ti2AlNb phase has an ordered orthorhombic structure, this phase is referred to as the O phase, and thus Ti2AlNb-based alloys are often called orthorhombic alloys.1,2) A typical orthorhombic alloy is the Ti–22Al–27Nb (in mol. pct) alloy,3–7) in which a high-temperature B2 phase (CsCl-type structure) is incorporated into the alloy’s structure to further improve ductility and fracture toughness. However, this (O+B2)-type orthorhombic alloy exhibits a larger primary creep strain than that of a nickel-based super-alloy such as IN718. Further, its yield stress and tensile strength drop sharply at temperatures higher than 923 K.8,9) Another disadvantage is that this alloy contains a large amount of niobium (Nb), which is costly.
We previously attempted to develop (O+B2)-type orthorhombic alloys with reduced Nb content and with improved tensile strength and creep strength at temperatures above 923 K. To achieve this goal, bcc (β) phase-stabilizing elements such as W, Mo, V and Fe were substituted for a portion of Nb in the Ti–22Al–27Nb alloy.10–15) Creep tests revealed that the W-replaced Ti–22Al–20Nb–2W alloy,10–12) the Mo-replaced Ti–22Al–11Nb–4Mo alloy13) and the (Mo and Fe)-replaced Ti–22Al–11Nb–2Mo–1Fe alloy14,15) exhibited markedly lower primary creep strain as well as markedly lower steady-state creep rates in the temperature range of 923 K to 1023 K as compared to those of other replaced alloys and a baseline Ti–22Al–27Nb alloy.15) However, the use temperatures of these alloys were limited to 1023 K.10,11,15) The use temperature of γ-TiAl has been reported as 1123 K.16) Thus, we must continue to explore the high-temperature capabilities of orthorhombic alloys comparable to γ-TiAl.
Considering that the α2 and O phases are stronger at high temperatures than the B2 phase,17) it can be expected that an alloy with (O+α2) two-phase microstructure would exhibit higher strength at high temperatures than the (O+B2)-type alloys. According to the isothermal section of the Ti–Al–Nb ternary phase diagram at 1173 K,8) an (O+α2) two-phase region exists around Ti–(26∼33) Al–(10∼17) Nb. We found in previous research that the (O+α2)-type Ti–27.5Al–13Nb alloy with a lamellar microstructure had a higher tensile strength above 973 K than the (O+B2)-type Ti–22Al–27Nb alloy.17) In comparison with the (O+B2)-type alloys,15) this alloy exhibited a lower steady-state creep rate, a longer 1 pct creep strain lifetime, and a longer fracture lifetime at 923 K.17)
One major drawback of this alloy is its high Nb content, which causes problems such as high cost and high density compared to γ-TiAl. To solve these problems, a compositional modification was attempted in the present study, in which the strong bcc (β) phase stabilizing elements such as Mo and Fe were substituted for a portion of Nb in a Ti–27.5Al–13Nb alloy, and three derivative alloys, Mo-replaced Ti–27.5Al–8.7Nb–1Mo, Fe-replaced Ti–27.5Al–5.5Nb–1Fe and (Mo and Fe)-replaced Ti–27.5Al–4.9Nb–1Mo–0.5Fe, were newly created.
Another drawback of this alloy is its very limited tensile ductility at room temperature (RT), especially when its microstructure is modified to a fully lamellar microstructure.17) For this alloy to be considered sufficiently reliable for use, some measures must be taken to improve its ductility.
It is known that a reduction in grain size improves many mechanical properties, such as the strength and ductility of metallic materials.6) As one way of reducing the grain size of Ti alloys, the addition of a minor amount of boron (B), typically 0.1 mass pct, to these alloys has been receiving a great deal of attention recently.18–39) Tamirisakandala et al. showed that the prior β-grain sizes of as-cast Ti–6Al–4V and Ti–6Al–2Sn–4Zr–2Mo–0.1Si are reduced by about an order of magnitude from 1200 to 200 µm by the addition of 0.1 pct B.18) Since then, it has been reported that grain-refined B-modified Ti–6Al–4V,18–34) Ti–6Al–2.75Sn–4Zr–0.4Mo–0.45Si (Ti-1100),35) β-type Ti,36–38) Ti–22Al–11Nb–2Mo–1Fe39) and γ-TiAl16) exhibited significant improvement in strength, stiffness, fatigue resistance and fracture toughness. By referring to those positive results, we tried to add a minor amount of B (0.1 mass pct) to the three derivative alloys and the baseline Ti–27.5Al–13Nb alloy.
Our previous study revealed17) that the cracks inducing fracture nucleated and grew along the O/α2 interfaces. Based on this observation, we conjecture that a fine duplex microstructure has a beneficial effect on the ductility, because the lengths of the O/α2 interfaces in a fine microstructure are short and thus the nucleation and growth of voids in such interfaces are difficult. Therefore, as another possible way to improve ductility, we tried to modify the microstructures of these four alloys, either with or without B, into fine duplex microstructures consisting of globular α2 phases and lamellar microstructures in the grains by applying thermo-mechanical treatment, in which hot deformation of the alloys is performed in the high-temperature (B2+α2) two-phase region, followed by annealing in the same (B2+α2) two-phase region.
Thus, to achieve the goal of improving ductility, two types of microstructure-refining methods, an addition of a minor amount of B and thermo-mechanical treatment to produce a duplex microstructure, were applied separately or concurrently to these four alloys. Since creep is one of the important properties for high-temperature applications, their creep properties were evaluated at 1023 K under 250 MPa stress.
The present authors have found previously that the (Mo and Fe)-replaced (O+B2)-type Ti–22Al–11Nb–2Mo–1Fe alloy exhibited enhanced high-temperature tensile and creep strength compared to those of the Ti–22Al–27Nb alloy.14,39) By reference to those positive results, we selected Mo and Fe as replacing elements to reduce the Nb content in the baseline alloy. The actual replacement was performed in the following manner.
Table 1 shows the critical concentrations of various β phase stabilizers in titanium necessary to retain a 100 pct bcc phase at RT,40) indicating that the strength of the β phase stabilizing ability was increased in the order Nb < Ta < V < W < Cr < Mo < Fe. We can expect from this order that the addition of small amounts of Fe and Mo would replace a considerable amount of Nb in the baseline Ti–27.5Al–13Nb alloy. Since the critical concentrations of Nb and Fe are 22.5 pct and 3.0 pct, respectively, the β phase stabilizing ability of Fe is 7.5 (i.e., 22.5 divided by 3.0 is 7.5) times greater than of Nb. In other words, the Nb equivalency of Fe is 7.5. Therefore, one atomic pct addition of Fe corresponds to the replacement of the 7.5 atomic pct of Nb. Similarly, since a critical concentration of Mo is 5.3 pct, one atomic pct addition of Mo corresponds to the replacement of 4.25 (22.5 divided by 5.3 is 4.25) atomic pct of Nb. The diffusivity of Fe in Ti is very high,41) which will result in the coarsening of the microstructure. Therefore, the amount of Fe is intentionally restricted to less than 1 pct.
Among various possible Nb-reduced derivative alloys, three alloys, Mo-replaced Ti–27.5Al–8.7Nb–1Mo (abbreviated as Mo-replaced alloy hereafter), Fe-replaced Ti–27.5Al–5.5Nb–1Fe (abbreviated as Fe-replaced alloy), and (Mo and Fe)-replaced Ti–27.5Al–4.9Nb–1Mo–0.5Fe (abbreviated as (MoFe)-replaced alloy), were selected as derivative alloys.
2.2 Ingot preparationThe alloy ingots, both with and without 0.1 pct B, were prepared by cold crucible levitation melting. Each cylindrical ingot had a diameter of about 70 mm, a length of about 70 mm and a weight of 1.3 kg.
The phase diagram of pseudo-binary Ti–27.5Al–Nb is shown on the right side of Fig. 1.9) As can be seen, a single B2 phase region exists at temperatures above ∼1443 K and a (B2+α2) two-phase region in the temperature range of ∼1273 K to ∼1443 K. Based on these data, hot-forging of the ingots to a cuboid shape and subsequent hot bar-rolling into 11.8 mm square bars were performed at 1373 K in the (α2+B2) two-phase region.
Phase diagram of pseudo-binary Ti–27.5Al–Nb for the Ti–Al–Nb system9) and heat treatment schedule adopted in the present study.
The heat treatment to produce fully lamellar microstructures consisted of holding the forged and hot-rolled bars at 1473 K in the single B2 phase region for 3.6 ks and cooling to room temperature at 0.03 K/s, and subsequent holding at around 1223 K for 118.8 ks in the (O+α2) two-phase region to stabilize the microstructure.
To produce a duplex microstructure, the alloys forged and hot-rolled in the (B2+α2) two-phase region were held in the same (B2+α2) two-phase region at 1373 K for 3.6 ks and then slowly cooled to RT at 0.03 K/s. Finally, the bars were held at around 1223 K for 118.8 ks in the (O+α2) two-phase region to stabilize the microstructure.
2.4 Tensile and creep properties evaluationsSmooth tensile specimens 3.5 mm in diameter and 16 mm in length were machined from the hot-rolled and heat-treated bars. Tensile tests were performed in air at a strain rate of 3 × 10−4s−1 at temperatures between RT and 1073 K. The creep test was conducted in air at 1023 K under 250 MPa stress, using a dead-weight creep rupture machine. The creep specimen had a 4.0 mm diameter and a 21 mm effective gauge length.
The prior B2 grain size of as-cast alloy ingot was reduced drastically from 600∼1200 µm to 100∼250 µm, as was the case for Ti–6Al–4V.34)
Figure 2 shows OM micrographs of the B-free and 0.1 pct B-modified baseline alloys in the as-hot bar-rolled condition. In both alloys, the prior B2 phase grains were elongated, stretching along the rolling direction, which means that complete recrystallization did not occur during the hot-bar rolling process. In the B-modified alloy, the prior B2 grains were considerably smaller overall than the B-free alloy grains. Further, the aggregate of TiB originally located at prior B2 grain boundaries in a cast ingot29,39) was untangled and dispersed in one direction, forming long and narrow black bands. Incidentally, quite the same deformation morphology was obtained for the three derivative alloys.
OM micrographs of the (a) B-free and (b) 0.1 pct B-modified Ti–27.5Al–13Nb in the as-hot bar-rolled condition.
XRD profiles of these four alloys that were water-quenched from 1473 K corresponded to the B2 phase and TiB, showing that all four alloys consist of the B2 phase and a small amount of TiB at 1473 K. When those forged and hot-bar rolled alloys were held at 1423 K for 3.6 ks XRD profiles corresponding to the B2 phase, α2 phase, and TiB were observed, indicating that the equilibrium phases at this temperature for these four alloys are B2, α2 and TiB. Figure 3 shows the XRD profiles of these four B-modified alloys that were held at 1373 K for 3.6 ks, subsequently cooled slowly to about 600 K at 0.03 K/s, and then held at 1223 K for 118.8 ks for stabilization annealing. The baseline alloy (abbreviated as BASE in Fig. 3) showed primarily an O+α2 structure with a small peak of TiB, whereas peaks of the B2 phase as well as O and α2 were observed three Mo- and Fe-containing derivative alloys (abbreviated as (Mo), (Fe) and (MoFe), respectively). Accordingly, it can hardly be said that these alloys are definitively classifiable as (O+α2)-type orthorhombic alloys. The amount of B2 phase seems to be larger in the (Mo and Fe)-replaced alloy than in the Mo-replaced and Fe-replaced alloys. The phase constitutions of these four alloys after this stabilization annealing are summarized in Table 2.
XRD profiles of Ti–27.5Al–13Nb (abbreviated as Base), Ti–27.5Al–8.7Nb–1Mo (Mo), Ti–27.5Al–5.5Nb–1Fe (Fe), and Ti–27.5Al–4.9Nb–1Mo–0.5Fe (MoFe), each with 0.1 pct B, after slow cooling at 0.03 K/s from 1373 K and subsequent stabilization annealing at 1223 K for 118.8 ks.
As shown in Fig. 4, these alloys show typical fully lamellar microstructures with several colonies of similarly aligned α2 phase lamellae (seen as a white contrast phase) within the prior B2 grains and a massive α2 phase at the grain boundaries.17) In Fig. 4(a), one restricted area surrounded by thick frames is called a colony. The O phase (seen as a black contrast phase) was present as thin laths between α2 phase lamellae. The fully lamellar morphology was found to vary depending on the compositions of the alloys. The most refined fully lamellar morphology, i.e., a thinner α2 phase lamella and a narrower grain boundary α2 phase, was observed in the Mo-replaced alloy (Fig. 4(b)) followed by the baseline alloy (Fig. 4(a)). The (Fe and Mo)-replaced alloy (Fig. 4(c)) and Fe-replaced alloy (Fig. 4(d)) exhibited a thicker α2 phase lamella and a massive grain boundary α2 phase. Transition metal elements such as Fe are known to diffuse more rapidly in titanium,41) which might accelerate the formation and growth of the α2 phase lamellae and the grain boundary α2 phase, resulting in a coarser lamellar morphology compared to those of Fe-free alloys when subjected to the same heat treatments. Although the details of the mechanism of such fast diffusion are still controversial, some type of interstitial mechanism is believed to be operative.42)
Optical micrographs of B-free (a) Ti–27.5Al–13Nb, (b) Ti–27.5Al–8.7Nb–1Mo, (c) Ti–27.5Al–4.9Nb–1Mo–0.5Fe and (d) Ti–27.5Al–5.5Nb–1Fe after slow cooling at 0.03 K/s from 1473 K in the single B2 phase region and subsequent stabilization annealing at 1223 K for 118.8 ks.
Figure 5 shows lower and higher magnification images of a fully lamellar microstructure taken from the 0.1 pct B-modified baseline alloy. Considerably small prior B2 grains were observed (Fig. 5(a)). These grains were almost equiaxed and their diameters averaged roughly 150 µm. Since this diameter is equal to the length of the short axis of the elongated B2 grain in a hot bar-rolled alloy (Fig. 2(b)), we can see that B2 grain growth did not occur during the holding of hot bar-rolled material in the single B2 phase region, perhaps due to the blocking effect of the TiB bands against B2 grain growth in the short axis direction. In conjunction with this B2 grain refinement, a reduction of colony size was observed, and accordingly, each α2 phase lamella was shortened significantly and widened. Hence, the aspect ratio of the length to the width of each lamella was considerably smaller than that of the B-free alloy. This widening of the α2 phase lamella in the short axis direction is the manifestation of the same volume of each α2 phase lamella between B-free and B-modified alloys when subjected to the same heat treatments. Accordingly, the shortening of an α2 phase lamella corresponds to the occurrence of its widening in the short axis direction. The higher magnification OM micrograph (Fig. 5(b)) reveals that there were long and narrow black bands aligned along the rolling direction (see also Fig. 6) and that these bands were composed of an agglomerate of TiB.29,39) The spacing between the adjoining bands was irregular, ranging from about 30 µm to about 70 µm. The higher magnification SEM micrograph of TiB-centric morphology (Fig. 5(c)) reveals that the width of each TiB was 1∼2 µm, and the aspect ratio of the width to the length was ∼10. The nature of TiB as well as its formation along the solidification path42) and the orientation relationship between TiB and surrounding α2 in the present orthorhombic alloys merit further investigation.
(a) Lower and (b) Higher magnification optical micrographs of 0.1 pct B-modified Ti–27.5Al–13Nb after slow cooling at 0.03 K/s from 1473 K in the single B2 phase region and subsequent stabilization annealing at 1223 K for 118.8 ks. Arrows indicate TiB bands. (c) Higher magnification SEM micrograph of TiB.
Optical micrographs of 0.1 pct B-modified (a) Ti–27.5Al–13Nb, (b) Ti–27.5Al–8.7Nb–1Mo, (c) Ti–27.5Al–4.9Nb–1Mo–0.5Fe and (d) Ti–27.5Al–5.5Nb–1Fe after slow cooling at 0.03 K/s from 1473 K in the single B2 phase region and subsequent stabilization annealing at 1223 K for 118.8 ks.
Figure 6 compares, at the same magnification, the fully lamellar microstructures of the 0.1 pct B-modified four alloys. The black-colored TiB bands can be seen in all of the alloys. No special difference in the width of each α2 phase lamella was observed between the baseline alloy (Fig. 6(a)) and the Mo-replaced alloy (Fig. 6(b)), and thus the aspect ratios of the α2 phase lamellae were equivalent between the two alloys. However, the (Mo and Fe)-replaced alloy (Fig. 6(c)) and Fe-replaced alloy (Fig. 6(d)) exhibited the widened α2 phase lamellae and the coarsened grain boundary α2 phase. Particularly large widening of the α2 phase lamella was observed in the Fe-replaced alloy, suggesting the occurrence of coalescence with the neighboring lamella. Thus, the morphology of the α2 phase lamella looks like an equiaxed α2 phase, which makes it difficult to distinguish the α2 phase lamella from the grain boundary α2 phase. Similar to the case of B-free alloys, faster diffusion of Fe41) would be responsible for the formation of such a coarse morphology.
3.1.5 Duplex microstructures of B-free alloysFigure 7 shows the duplex microstructures of the B-free four alloys. Both the baseline alloy (Fig. 7(a)) and Mo-replaced alloy (Fig. 7(b)) showed typical duplex microstructures composed of the globular α2 phases and the fine lamellar microstructures. It should be noted that a homogeneous distribution of globular α2 phases was not attained, but rather the α2 phases were arranged in a necklace-like band along the rolling direction. Such an arrangement was already reported by Emura et al. in their work on an (O+B2)-type orthorhombic alloy.43) They attributed the formation of a necklace-like band of α2-phases to the inhomogeneous distribution of solute Nb elements inherited from the dendritic segregation of this high melting point element in the ingot. Accordingly, there exists macroscopically a Nb-rich and a Nb-lean areas alternately in the ingot. During the hot bar-rolling of this ingot, the Nb-rich and the Nb-lean areas were stretched along the rolling direction, forming their respective band segregation, and the α2 phases were preferably precipitated at the band-shaped Nb-lean regions, resulting in the necklace-like α2 phase arrangements.43) To avoid this segregation problem, the adoption of the pre-alloyed powder metallurgy (P/M) manufacturing method seems very useful.6,43,44)
Optical micrographs of B-free (a) Ti–27.5Al–13Nb, (b) Ti–27.5Al–8.7Nb–1Mo, (c) Ti–27.5Al–4.9Nb–1Mo–0.5Fe and (d) Ti–27.5Al–5.5Nb–1Fe after hot bar-rolling and annealing at 1373 K in the (B2+α2) two-phase region and subsequent stabilization annealing at 1223 K for 118.8 ks. Arrows indicate the α2 phases forming a necklace-like arrangement.
In the (Mo and Fe)-replaced alloy (Fig. 7(c)), both the globular α2 phase in the necklace-like α2 phase bands and the lamellar microstructure between two adjacent bands exhibited coarsened morphologies. Further, the Fe-replaced alloy (Fig. 7(d)) exhibited a much coarser morphology. The diameter of the globular α2 phase in this Fe-replaced alloy was considerably larger than that of the other three alloys. Because the globular α2 phase covers the entire area, its microstructure resembled an equiaxed rather than a duplex microstructure. It thus appeared that coalescence with the neighboring globular α2 phase occurred.
3.1.6 Duplex microstructures of B-modified alloysFigure 8 shows the duplex microstructures of the 0.1 pct B-modified four alloys. It can be seen that, in general, many of the globular α2 phases formed over almost the entire area, and thus the area of the lamellar microstructure surrounded by these globular α2-phases was very limited or completely diminished. Band-shaped Nb-lean regions appeared in these B-modified alloys. Therefore, it is thought that the globular α2 phases were formed as a result of preferential precipitation at the band-shaped Nb-lean regions. As was the case for the B-modified alloys with lamellar microstructures (Figs. 5 and 6), many of the narrow bands composed of an agglomeration of TiB stretching parallel to the rolling direction were observed. We can see that each TiB is embedded by a globular α2 phase, which indicates that TiB acts as a nucleation site for globular α2 phase formation. For these two reasons, the globular α2 phase was observed to exist over almost the entire area.
Optical micrographs of 0.1 pct B-modified (a) Ti–27.5Al–13Nb, (b) Ti–27.5Al–8.7Nb–1Mo, (c) Ti–27.5Al–4.9Nb–1Mo–0.5Fe and (d) Ti–27.5Al–5.5Nb–1Fe after hot bar-rolling and annealing at 1373 K in the (B2+α2) two-phase region and subsequent stabilization annealing at 1223 K for 118.8 ks.
In the (Fe and Mo)-replaced alloy and Fe-replaced alloy, the globular α2 phases became considerably larger, especially in the latter alloy.
3.2 Tensile properties 3.2.1 B-free and B-modified alloys with lamellar microstructuresAs Table 2 shows, the B-free alloy failed prematurely before attaining proof strength when tensile tested at RT; it therefore exhibited a total elongation value of zero or close to zero. It seems that the very large prior B2 grains that an alloy with a full lamellar microstructure possesses are responsible for the extremely low elongation value. For the baseline alloy, proof strength was obtained at temperatures above 723 K, and the total elongation value of 7.6 pct was attained at 1073 K.
The addition of 0.1 pct B to the baseline alloy increased the total tensile elongation value from 0.1 pct to 0.8 pct and from 7.6 pct to 10.9 pct at RT and 1073 K, respectively, and this addition increased the tensile strength from 404 MPa to 581 MPa and from 392 MPa to 479 MPa at RT and 1073 K, respectively. However, the elongation value of 0.8 pct at RT is not sufficient to make it a reliable structural material. The Mo-replaced, Fe-replaced and (Mo and Fe)-replaced alloys still failed prematurely at RT. Thus, it turned out that the beneficial effect of the addition of 0.1 pct B in improving the ductility and tensile strength of these four alloys with fully lamellar microstructures was very limited, which means that the reduction of the prior B2 grain size conferred little or no ductility improvement.
3.2.2 B-free and B-modified alloys with duplex microstructuresAs shown in Fig. 9 and Table 2, the tensile strength of each alloy, either with or without B, decreased and ductility increased with increasing temperature, and each alloy was fractured before attaining peak strength, i.e., ultimate tensile strength, at temperatures up to 923 K. Each alloy passed through the peak on the tensile curves at 1073 K.
Tensile curves of B-free and 0.1 pct B-modified (a), (b) Ti–27.5Al–13Nb, (c), (d) Ti–27.5Al–8.7Nb–1Mo, (e), (f) Ti–27.5Al–4.9Nb–1Mo–0.5Fe and (g), (h) Ti–27.5Al–5.5Nb–1Fe after hot bar-rolling and annealing at 1373 K in the (B2+α2) two-phase region and subsequent stabilization annealing at 1223 K for 118.8 ks.
Among the B-free four alloys, the baseline alloy (Fig. 9(a)) and Mo-replaced alloy (Fig. 9(c)) exhibited the highest strengths of 493 MPa and 494 MPa, respectively. The tensile ductility at RT was improved considerably by applying thermo-mechanical treatment. This was particularly true for the baseline alloy, whose total elongation was increased to 4.3 pct. For the Mo-replaced alloy, total elongation was increased to 1.5 pct without exhibiting a premature failure. The deformation behavior at 1073 K after the tensile strength had passed its peak was found to be highly dependent on the alloy composition. The baseline alloy and Mo-replaced alloy, which showed uniform elongation of 14.3 pct and 8.8 pct, respectively, exhibited gradual decreases in tensile strength after passing peak strength and showed fairly large local elongations of 4.8 pct and 13.6 pct, respectively. Therefore, the total elongations for the respective alloys were 19.1 pct and 22.4 pct. In the (Mo and Fe)-replaced alloy (Fig. 9(e)) and Fe-replaced alloy (Fig. 9(g)), the 0.2 pct proof strength and tensile strength were lower than those of the baseline alloy and Mo-replaced alloy when compared at the same temperature, and post-peak tensile strength decreased gradually rather than rapidly until the final failure. These two alloys showed very large elongations of more than 80 pct, which is close to the elongation of superplastic metallic material. Based on these property data, it can be said that the best combinations of tensile strength and ductility among the B-free four alloys with duplex microstructures were achieved by the baseline alloy followed by the Mo-replaced alloy.
Due to the further refinement of the duplex microstructure by the addition of 0.1 pct B, the ductility of each B-modified alloy was largely improved compared to that of the B-free counterpart. For example, the tensile elongation of the Mo-replaced alloy (Fig. 9(d)) at RT was increased from 1.5 pct for the B-free alloy to 4.7 pct. Moreover, the 0.2 pct proof strength and the tensile strength of this Mo-replaced alloy at 1073 K were considerably higher than those of the B-modified baseline alloy (Fig. 9(b)). It is also seen that the ductilities of the (Mo and Fe)-replaced alloy and Fe-replaced alloy (Fig. 9(f) and Fig. 9(h), respectively) at RT were improved, although not remarkably, by the addition of 0.1 pct B. Therefore, it was confirmed that the addition of 0.1 pct B concurrent with thermo-mechanical treatment to transform the microstructure to a duplex microstructure is effective to improve the ductility at RT. Moreover, among the B-modified four alloys, the Mo-replaced alloy was revealed to have the highest tensile strength and ductility in the whole temperature range up to 1073 K.
Figure 10 compares the high-temperature specific tensile strengths of conventional high-temperature titanium alloys such as Ti–6Al–2Sn–4Zr–2Mo and γ-TiAl45) to those obtained in the present study. It is seen that the specific tensile strengths of Ti–27.5Al–13Nb and 0.1 pct B-modified Ti–27.5Al–8.7Nb–1Mo alloys with duplex microstructures lie nearly at the upper bound of the γ-TiAl scatter band. Therefore, these two alloys, and especially the 0.1 pct B-modified Ti–27.5Al–8.7Nb–1Mo alloy, can be regarded as promising lightweight and high-strength alloys for high-temperature applications, with tensile strength similar to that of γ-TiAl and with sufficient materials ductility.
High-temperature specific tensile strength for various high-strength metallic and composite materials.45) The present data are included for comparison.
Creep curves tested at 1023 K under 250 MPa stress are shown in Fig. 11. The Mo-replaced alloy was chosen for creep evaluation because this alloy showed the highest high-temperature tensile strength among the three derivative alloys. As to the B-free alloys, the lamellar microstructure has a superior creep property (Fig. 11(a)), that is, the lamellar microstructure showed a lower steady-state creep rate compared to the duplex microstructure (Fig. 11(b)).
Creep curves of the B-free and 0.1 pct B-modified Ti–27.5Al–13Nb alloy and Ti–27.5Al–8.7Nb–1Mo tested at 1023 K under 250 MPa stress. (a) Lamellar microstructure and (b) Duplex microstructure.
The reason for the superiority of creep characteristics recognized in the B-free alloy with a lamellar microstructure can be explained as follows.
Tang and Hagiwara previously investigated the creep behavior of the baseline alloy,17) and the (O+B2)-type Ti–22Al–20Nb–2W alloy10) at 873 K, 973 K and 1023 K under 310 MPa stress. From the values of the stress component and activation energy, the creep mechanism of these alloys under this creep test condition was regarded as the dislocation-climb.17) The creep test conditions applied to the present alloys, 1023 K/250 MPa, were very similar to those in the two cases described above. Therefore, it is reasonable to consider that the creep mechanism of the present four alloys was controlled by the dislocation-climb.
The dislocation climb is mainly affected by two factors, i.e., the diffusivity of constitutional elements and the number density of boundaries such as phase interface and grain boundary. It is expected that microstructures with many effective barrier sites to block the dislocation movement show a lower steady-state creep rate. Since we compare the creep properties of different microstructures generated in the same alloy, the diffusivity of constitutional elements was excluded from the discussion on the dislocation-climb of the creep process. Therefore, we discuss the difference of creep properties between lamellar and equiaxed microstructures in terms of the number of barrier sites in each microstructure as follows.
The lamellar microstructure has semi-coherent interfaces between adjacent two lamellae and interfaces between two adjacent colonies. These interfaces are effective barrier sites against moving dislocations and act to confine active dislocations to the individual lamella.4,46) In the duplex microstructure, the interface between the equiaxed O phase and surrounding phase can be considered as the major barrier sites against the dislocation motion, and the number of such interfaces is regarded to be smaller in this microstructure than in the lamellar microstructure.
Further, it is shown that the degree of creep deformation is influenced by the width of each lamella or by the diameter of the equiaxed phase, and the resistance to the creep deformation becomes stronger as these lengths become shorter, i.e., the slip length across the lamella or equiaxed phase becomes shorter.46,47) As shown in Fig. 4 and 7 for a lamellar and a duplex microstructure, respectively, for the present alloy and another alloy, e.g., Ti–6Al–4V,29) the width of the lamella is considerably smaller than the diameter of equiaxed phases.
These two observations, i.e., the observation that there are many barriers against the dislocation motion and the observation that the slip length is shorter in each microstructural unit such as a lamella, can explain the superiority of the creep properties observed in the lamellar microstructure over the duplex microstructure.
It was found that the addition of a minor amount of B had a different effect on the creep properties depending on the compositions and microstructures of the alloys. As shown in Fig. 11(a), the creep curve of the 0.1 pct B-modified baseline alloy with a lamellar microstructure almost completely coincided with that of the B-free alloy. Contrary to this, the 0.1 pct B-modified Mo-replaced alloy with a lamellar microstructure showed a lower steady-state creep rate than its B-free counterpart. As to the duplex microstructure, the creep properties of these baseline and Mo-replaced alloys were degraded by the addition of 0.1 pct B (Fig. 11(b)).
For comparison, creep curves of the B-free and 0.1 pct B-modified (O+B2)-type Ti–22Al–11Nb–2Mo–1Fe alloy, each either with lamellar microstructure or with duplex microstructure tested at 923 K under 200 MPa stress are shown in Fig. 12.39) In this alloy, contrary to the cases of the baseline and Mo-replaced alloys with duplex microstructures, the addition of B resulted in superior creep properties over the B-free counterparts.
Creep curves of the B-free and 0.1 pct B-modified (O+B2)-type Ti–22Al–11Nb–2Mo–1Fe alloy with lamellar and duplex microstructures tested at 923 K under 200 MPa stress.39)
To sum up, the addition of a minor amount of B exerts either a positive or negative effect on the creep properties depending on the phase constitution, compositions and microstructures of the alloys, creep test temperature and applied stress. Such different effects of B on creep properties can be explained qualitatively as follows.
Regarding the effect of this grain refinement on the creep properties, it was reported that in a dislocation-controlled creep process, the secondary creep rate was not dependent on the prior B2/β grain size at a range from 33 to 337 µm.4) Thus, grain refinement due to B addition is not expected to detrimentally affect the creep behavior.48)
In a creep process controlled by dislocation climb, the presence of TiB effectively exerts a blocking effect against grain boundary sliding and dislocation motion, and thereby increases the creep resistance.48,49) Therefore, the enhanced creep properties observed in the Mo-replaced alloy with a lamellar microstructure (Fig. 11(a)) and in the Ti–22Al–11Nb–2Mo–1Fe alloy with both lamellar and duplex microstructures (Fig. 12) can be considered to be based on the presence of TiB in the microstructure.
However, this enhancement of the creep properties by the addition of B was diminished in the baseline alloy with a lamellar microstructure (Fig. 11(a)). Moreover, the diminishment of the creep properties by dispersing TiB was noticeable in the baseline and Mo-replaced alloys with duplex microstructures (Fig. 11(b)). These alloys with low creep resistance were tested at 1023 K. At high-temperatures such as 1023 K, the grain boundary sliding and cracking, and also the phase-interface sliding and cracking will become active.11,50) Thus, the occurrence of such sliding and cracking will lower the creep resistance and thereby increase the steady-state creep rate. Concurrently with such sliding and cracking deformation, the creep deformation by the dislocation climb mechanism seems to be operating at this temperature. In spite of this, the negative effects on the creep deformation caused by such sliding and cracking offset the positive effect of TiB. As shown in Fig. 11, the negative effect of TiB at 1023 K appeared more prominently in the duplex microstructure. Since the duplex microstructure has fewer barrier sites against dislocation motion compared to the lamellar microstructure, it is therefore reasonable to consider that, under the conditions that both the grain boundary sliding and the phase-interface sliding, and the dislocation climb mechanism are operating, degradation occurs more prominently in the microstructure with fewer barrier sites against the dislocation motion, i.e., in the duplex microstructure. In the Ti–22Al–11Nb–2Mo–1Fe alloy, such sliding and cracking deformation was not observed in the lamellar microstructures when creep tested at 923 K under 310 MPa stress,50) which may be one of the reasons for the experimental results that the creep properties were improved by the addition of B in both the lamellar and duplex microstructures.
3.3.2 Comparison of creep properties among B-modified four alloysCreep properties of the 0.1 pct B-modified baseline alloy and three derivative alloys, each having a duplex microstructure, were compared at 1023 K under 250 MPa stress, as shown in Fig. 13. Among these B-free four alloys, the Fe-containing alloys, i.e., the Fe-replaced alloy and (Mo and Fe)-replaced alloy, exhibited the inferior resistance to creep deformation. At the stage of alloy design for the derivative alloys, it was anticipated that the inclusion of Fe of 1 pct or less in the alloy composition would not exert a detrimental effect on the creep properties. Thus a new, cost-effective Fe-containing alloy with creep properties almost equal or slightly inferior to those of the base alloy will be produced. This idea was based on the fact that the 1 pct Fe-containing (O+B2)-type Ti–22Al–11Nb–2Mo–1Fe alloy exhibited creep properties superior to those of the original Ti–22Al–27Nb alloy.39,50) Contrary to our expectations, however, in the present case, the creep test results proved disappointing. In a future study, discussions of the role played by Fe in improving or degrading creep properties would be a subject of much interest, especially when Fe is added in combination with Mo.
Comparison of creep curves of the 0.1 pct B-modified baseline Ti–27.5Al–13Nb alloy and three derivative alloys, each with a duplex microstructure, tested at 1023 K under 250 MPa stress.
Looking at Figs. 11(a) and 11(b), another noteworthy point is that, under the conditions depicted, the Mo-replaced alloy exhibited creep properties in both the lamellar and duplex microstructures superior to those of the baseline alloy, which means that the substitution of Mo for a portion of Nb in the baseline alloy was effective for improving the creep properties. Further, it should be mentioned that the 0.1 pct B-modified Mo-replaced alloy with a lamellar microstructure ranked first in creep superiority among the creep curves.
Compositionally less expensive Ti–27.5Al–8.7Nb–1Mo, Ti–27.5Al–5.5Nb–1Fe and Ti–27.5Al–4.9Nb–1Mo–0.5Fe alloys were newly created by replacing a portion of Nb in the baseline Ti–27.5Al–13Nb alloy with an equivalent amount of Mo and Fe. Further, 0.1 pct B was added to these derivative alloys.
From the results of tensile and creep property evaluations, the following two key findings were obtained. First, the concurrent application of two types of microstructure-refining methods, i.e., an addition of 0.1 pct B and a thermo-mechanical treatment to produce a duplex microstructure, proved to be effective in improving the ductility and tensile strength of these three derivative alloys and a baseline alloy. Second, the 0.1 pct B-modified Mo-replaced Ti–27.5Al–8.7Nb–1Mo alloy with a duplex microstructure exhibited the highest ductility of 4.7 pct at RT and highest tensile strength at high-temperature up to 1073 K, in addition to creep properties superior to those of other derivative alloys and the baseline alloy.
