Effect of Zr on Microstructure and Oxidation Behavior of α and α + α2 Ti-Al-Nb Alloys
2016 Volume 57 Issue 11 Pages 1902-1907
Details
2016 Volume 57 Issue 11 Pages 1902-1907
To develop high-temperature Ti alloys, microstructure and oxidation behavior were investigated in Ti-(10,15)Al-2Nb-(1,2)Zr (at%) alloys. For reference, ternary Ti-15Al-2Zr alloy was also investigated. A single α phase exists in Ti-10Al-2Nb-(1, 2)Zr alloys; on the other hand, an α and α2 two-phase was observed in Ti-15Al-2Nb-(1, 2)Zr alloys. Fine α2 precipitates of about 10 nm were observed in the Ti-15Al alloys. The oxidation behavior of Ti-Al-Nb-Zr and Ti-Al-Zr ternary alloys was investigated to understand the effect of Zr on oxidation behavior at 1023 K. It was found that Zr improves the oxidation resistance of the α phase more effectively than Nb. TiO2 was formed in all tested alloys. The addition of Zr drastically decreased the growth rate of TiO2. It was also found that the addition of Nb improved adhesion of the oxide layer.
Ti alloys are light materials with almost half the density of Ni-based superalloys and steel. In addition to low density, Ti alloys possess good oxidation resistance and creep properties up to 873 K; hence, they are used for the blades and disks in compressors in aircraft engines in the temperature range between 673 and 873 K1). However, above 873 K, the oxidation and creep resistance of Ti alloys decrease drastically1); thus, it is important to improve the strength, creep resistance, and oxidation resistance above 873 K.
Generally, near α-Ti alloys with an hcp structure are used for high-temperature applications because of the limited slip plane. Solid solution hardening by Al, Sn and Zr and precipitation hardening by α2-Ti3Al and/or silicide are used to improve the high-temperature strength and creep resistance of the α phase.
Oxidation behavior of conventional α-Ti alloys such as IMI 834 and Ti1100 was investigated, and the oxide scales of both alloys were found to consist of porous mixed oxides with a coarse-grained rutile TiO2 growing at the outer surface2). The oxidation kinetics of these alloys is dominated by TiO2 formation2). It was also found that the oxidation of Ti-6Al-4Nb followed linear-parabolic kinetics between 973 and 1123 K3). The oxide scale formed on Ti-6Al-4V consisted of alternating layers of Al2O3 and TiO2. Al2O3 always formed at the external gas/oxide interface; on the other hand, TiO2 always formed at the oxide/substrate interface3).
Alloying is one way to improve the oxidation resistance of α-Ti alloys. For example, Nb is found to improve the oxidation resistance of α-Ti4,5). Nb impedes the growth of TiO2 in Ti-1, 2 mass% (0.5, 1 at%)4) and Ti-5, 10 at%5). There have been many reports on the effect of Nb on α2-Ti3Al and γ-TiAl5–7). However, there are few investigations of the effect of Nb on α-Ti-Al alloys.
In our previous study, we investigated the oxidation behavior and phase equilibrium of Ti-Al-Sn-Nb quaternary alloys8). We found that the mass change during isothermal oxidation tests at 1023 K was smaller in the alloy with high Nb content. We also found that Sn accelerated oxide growth. Since then, attempts have been made to design alloys with less Sn and more Nb. The phase equilibria and oxidation behavior of Ti-Al-Nb ternary alloys were also investigated to understand the effect of Nb on the α phase9). It was found that the α and α2 two-phase region was smaller than that in the calculated ternary phase diagram10). In Ti-15Al-2Nb (at%), nano-scale α2 precipitates were formed. The oxidation rate of the Ti-15Al alloy was decreased to less than 10% of its original value by the addition of 2 at% of Nb.
To design new high-temperature Ti alloys, we also focus on Zr because Zr is expected to strengthen α-Ti alloys. Furthermore, Zr is known to be a neutral element, which means that the β transus is almost flat after the addition of Zr, and hence the α phase is stabilized by the addition of Zr. Stabilizing of the α phase is very important when using Ti alloys at high temperatures. The goal of the present study was to determine whether Zr improves the oxidation resistance of α Ti alloys. Although it has been found that the oxidation resistance of TiAl is improved by the addition of Zr11), the effect of Zr on the oxidation behavior of α-Ti alloys has not yet been ascertained. Therefore, in the present study, we focused on α-Ti-10Al-2Nb alloys (at%) and α + α2 two-phase Ti-15Al-2Nb alloys to investigate the effect of adding Zr on the microstructure and the oxidation resistance of Ti-Al-Nb alloys. The oxidation behavior of Ti-Al-Nb-Zr alloys was compared with that of Ti-Al-Zr and binary Ti-Al alloys8,9) and ternary Ti-Al-Nb alloys8,9).
Twenty-gram alloy ingots of Ti-(10,15)Al-2Nb-(1,2)Zr and Ti-15Al-2Zr alloys (at%) were melted by arc melting. The nominal composition and analyzed composition using an energy dispersive x-ray spectroscopy (EDS) of the ingots are listed in Table 1. Each ingot was melted five times to ensure compositional homogeneity. The ingots were solution treated at 1273 K for 2 h and air cooled to room temperature, followed by ageing at 973 K for 24 h and air cooling to room temperature.
| Composition (at%) | Ti (at%) | Al (at%) | Nb (at%) | Zr (at%) |
|---|---|---|---|---|
| Ti-10Al-2Nb-1Zr (Ti-6Al-4Nb-2Zr, mass%) |
88.7 | 8.50 | 1.83 | 0.93 |
| Ti-10Al-2Nb-2Zr (Ti-6Al-4Nb-4Zr, mass%) |
87.9 | 8.07 | 1.95 | 2.05 |
| Ti-15Al-2Nb-1Zr (Ti-9Al-4Nb-2Zr, mass%) |
84.4 | 12.8 | 2.02 | 0.87 |
| Ti-15Al-2Nb-2Zr (Ti-9Al-4Nb-4Zr, mass%) |
83.9 | 12.8 | 1.56 | 1.77 |
| Ti-15Al-2Zr (Ti-9Al-4Zr, mass%) |
84.8 | 12.9 | - | 2.19 |
Microstructural characterizations of the aged samples were carried out using X-ray diffraction (XRD), field emission gun scanning electron microscopy (FEG-SEM, JEOL JSM 7000F or 7001F) and transmission electron microscopy (TEM, JEOL JEM 2000FX). The XRD measurement was carried out for a plate with dimensions of 8 × 10 × 1 mm at room temperature on a RINT2500 X-ray diffractometer using Cu Kα radiation, operated at 50 kV and 300 mA. Samples for SEM observation were embedded in resin and polished using polishing papers and SiO2. The specimens for TEM observation were prepared by electrolytic polishing of thin, 3-mm-diameter discs, punched from thin sections with a thickness of ～50 μm. Electrochemical polishing was carried out in an electrolyte containing 5% perchloric acid, 35% n-butyl alcohol and 60% methanol at 20 V, using a twin jet polisher. The temperature of the electrolyte was maintained at 253–263 K, using liquid nitrogen, throughout the process of electrochemical thinning.
Isothermal oxidation tests were performed for a 4 × 4 × 1 mm sample at 1023 K in laboratory air. Each sample was put in an alumina crucible, and the weight change was measured together with the weight of the crucible. The samples in alumina crucibles were removed from the furnace after 24, 48, 72, 144, 168, 192, and 208 h of exposure. The mass change of each sample was measured using a microbalance with an accuracy of ±0.0001 g. After oxidation testing, the structure of the oxide was investigated using XRD. The cross-sectional microstructures of the oxidized samples were also observed by SEM, and line analysis was performed to measure the range of composition using EDS attached to the FEG-SEM to identify oxides.
To identify the phase, X-ray diffraction analysis was performed at room temperature; the X-ray diffraction patterns are shown in Fig. 1. The diffraction patterns were based on the α phase in all tested alloys. α2-Ti3Al is the ordered structure of the α phase, and can be distinguished from the α phase by the existence of peaks at 26° and 30°, however, neither peak was observed. Moreover, there are no obvious diffraction peaks from the β phase, which often forms near α Ti alloys.
X-ray diffraction patterns of (a) Ti-10Al-2Nb-1Zr, (b) Ti-10Al-2Nb-1Zr, (c) Ti-1Nb-1Z5Al-1Zr, (d) Ti-1Nb-1Z5Al-2Zr, and (e) Ti-15Al-2Zr after heat treatment at 1273 K for 3 h following by air cooling.
Since microstructures with clear contrast were not found by SEM observation, TEM observation was performed to investigate whether the α2 phase forms within the α phase. In Fig. 2(a) and (b), the bright-field images of Ti-10Al-2Nb-1Zr and Ti-10Al-2Nb-2Zr represent a single phase, and contrast resulting from any precipitates was not observed. The α grain size of Ti-10Al-2Nb-1Zr was several tens of μm and elongated grains with a width of several μm and length of several tens of μm were observed in Ti-10Al-2Nb-2Zr. The microstructure of Ti-10Al-2Nb-2Zr is similar to the typical lamellar structure with thin β phase obtained in α Ti alloys. Although the existence of the β phase was not clear in this study, it is possible that a thin β phase surrounded the α phase. Diffraction patterns of these alloys showed a single α phase; superlattice spots from an α2 phase were not observed. In the binary phase diagram as shown in Fig. 312), the composition of Ti-10Al at 1273 K is in the α phase. The present result indicates that the α phase is stable even with the addition of 2 at% Nb and 2 at% Zr. On the other hand, the diffraction patterns of Ti-15Al-2Nb-2Zr and Ti-15Al-2Zr clearly revealed the superlattice spots from the α2 phase as shown in Fig. 4(a) and (b). Dark-field images of these alloys taken using the superlattice spots indicated by the white circles clearly showed fine α2 particles smaller than 10 nm. It is known that an α and α2 two-phase region exists in Ti-15Al-2Zr13), and the present result is comparable with that of the previous study. Furthermore, the present study found that α2 precipitates form as a result of the addition of 2 at% of Nb.
Two bright-field images and diffraction patterns of (a) Ti-10Al-2Nb-1Zr and (b) Ti-10Al-2Nb-2Zr.
A phase diagram of binary Ti-Al alloys12).
Two dark-field images and diffraction patterns of (a) Ti-15Al-2Nb-2Zr and (b) Ti-15Al-2Zr.
Isothermal oxidation tests were performed on the quaternary and ternary alloys at 1023 K. The weight change is shown in Fig. 5 together with binary alloys Ti-10Al and Ti-15Al8), ternary alloy Ti-10Al-2Nb8), and a commercial alloy Ti-6242 [Ti-11Al-0.8Sn-2.3Zr-1.1Mo (at%)]8).
Weight change during isothermal oxidation tests at 1023 K of (a) Ti-10Al-2Nb-(1,2)Zr, Ti-10Al-2Nb and Ti-6242 and (b) Ti-15Al-2Nb-(1,2)Zr, Ti-15Al-2Nb, Ti-15Al-2Zr and Ti-6242.
Figure 5(a) shows the oxidation behavior of Ti-10Al alloys. The Ti-10Al binary alloy indicated a large weight gain, but the weight gain was drastically decreased by the addition of Nb. The weight gain of Ti-10Al after 200 hours was 8 mg/cm2, which decreased to 2 mg/cm2 with the addition of 2 at% Nb. This means that the addition of Nb reduced the weight gain by a factor of four. Moreover, the addition of Zr to Ti-10Al-2Nb decreased the mass change to less than 1 mg/cm2. However, there was no visible difference between the addition of 1 at% Zr and 2 at% Zr. The total mass gain of Ti-6242 was 4 mg/cm2 after oxidation for 200 h. The mass gains of Ti-10Al-2Nb and quaternary alloys are less than that of Ti-6242, indicating that adding Nb and/or Zr improves the oxidation resistance. The Ti-15Al alloys exhibited the same tendencies, as shown in Fig. 5(b). Comparing Ti-15Al and Ti-15Al-2Nb, the addition of Nb decreased the mass gain by less than 50% of its original value. Moreover, the addition of Zr into Ti-15Al-2Nb decreased the mass gain compared with simple Nb addition. However, the weight gain of Ti-15Al-2Nb-(1,2)Zr was almost the same as that of Ti-15Al-2Zr. This indicates that the oxidation behavior was governed by Zr rather than Nb in quaternary alloys. Comparing Ti-15Al, Ti-15Al-2Nb, and Ti-15Al-2Zr, the weight gain was 6.5, 2, and 1 mg/cm2, respectively. The addition of Zr decreased the weight gain by about 85%, which was greater than the 70% reduction caused by Nb. In the case of Ti-15Al alloys, the weight gain of alloys by the addition of Nb and/or Zr was less than that of Ti-6242, indicating an improvement of oxidation resistance for the alloys with an α and α2 two-phase structure.
Comparing Ti-10Al and Ti-15Al, the weight gain of Ti-15Al was smaller than that of Ti-10Al, indicating the effect of Al. However, this difference became small when the amounts of Nb and Zr increased. The weight gains of Ti-10Al-2Nb-2Zr and Ti-15Al-2Nb-2Zr were almost the same although the weight gain after 200 h of oxidation decreased slightly in Ti-15Al-2Nb-2Zr. Further investigation is necessary to determine whether this decrease of weight gain indicates a change in the mechanism of oxidation or simply an experimental error.
It is suggested that Nb and Zr have the following effects on Ti-Al alloy:
To identify the oxide formed on the surface, X-ray diffraction analysis was carried out on the surfaces of the oxidized alloys. The X-ray diffraction patterns are shown in Fig. 6. The major peaks were assigned to rutile-TiO2 in all tested alloys. The peaks of α-Ti phase were also detected from the substrate. The peaks of other oxides such as Al2O3, Nb2O5, or ZrO2 were not detected in the X-ray diffraction patterns.
X-ray diffraction patterns of (a) Ti-10Al-2Nb-1Zr, (b) Ti-10Al-2Nb-1Zr, (c) Ti-1Nb-1Z5Al-1Zr, (d) Ti-1Nb-1Z5Al-2Zr, and (e) Ti-15Al-2Zr after isothermal oxidation testing at 1023 K.
The cross section of alloys oxidized at 1023 K for 208 h are shown together with a line analysis from the sample surface to the interior of the samples in Fig. 7 and Fig. 8. The thickness of the observed oxide layer of Ti-10Al-2Nb-1Zr, Ti-10Al-2Nb-2Zr, Ti-15Al-2Nb-1Zr and Ti-15Al-2Nb-2Zr was around 2–4 μm, showing no large difference. This is reasonable because the weight gains of these alloys were almost the same after oxidation for 208 h. However, in Ti-15Al-2Zr, the oxide layer was not observed clearly, indicating spalling of the oxide layer as shown in Fig. 8(c). The thickness of the oxide layer of Ti-10Al, Ti-10Al-2Nb, Ti-15Al, and Ti-15Al-2Nb was 40, 10, 40, and 5 μm in our previous study9). Compared with binary and ternary alloys, it is clear that growth of the oxide layer was drastically suppressed in quaternary alloys. On the other hand, the addition of Zr by itself causes the oxide layer to spall as shown in Ti-15Al-2Zr. The addition of both Nb and Zr together improved the adhesion of the oxide layer; this indicates that Nb not only reduces oxidation but also improves oxide adhesion.
Cross-section microstructure and line analysis of the oxide layer for (a) Ti-10Al-2Nb-1Zr and (b) Ti-10Al-2Nb-2Zr.
Cross-section microstructure and line analysis of the oxide layer for (a) Ti-15Al-2Nb-1Zr, (b) Ti-15Al-2Nb-2Zr and (c) Ti-15Al-2Zr.
In Ti-10Al-2Nb-1Zr, a thin outer gray layer, a thin dark layer and a thick inner gray layer were observed as shown in Fig. 7(a). Line analysis of the thick inner gray layer gave a Ti:O ratio of about 1:2, indicating the formation of TiO2. In TiO2, large cracks were observed. The bottom diagram indicates enlarged scale line analysis. In TiO2, the concentrations of Al, Nb and Zr were less than 1 at%. The outer dark layer included a higher concentration of Al. It is assumed that titanium oxide and aluminum oxide formed together in the outer layer. The concentration of Ti increased again in the thin outer gray layer suggesting formation of TiO2. Small particles with similar contrast in the outer layer were observed in the TiO2 layer, suggesting the formation of particles of Al2O3. Therefore, oxide layers was divided into three parts: 1) gray outer layer (mainly TiO2), 2) dark outer layer (mainly Al2O3) and 3) gray inner layers (TiO2 + Al2O3 particles). The Nb and Zr content was very small, less than 1 at% in both Al2O3 and TiO2 layers as shown in the bottom diagram. The oxide layers with the same structure were observed in Ti-15Al-2Nb-2Zr as shown in Fig. 8(b). In Ti-10Al-2Nb-2Zr, an outer layer of Al2O3 was not observed, but the structure of oxide was again same as oxide in Ti-10Al-2Nb-1Zr as shown in Fig. 7(b). In Ti-15Al-2Nb-1Zr and Ti-15Al-2Zr, the oxide layer was not clearly observed because of spallation of the oxide layer; this indicates low adhesion of oxide after Zr addition as shown in Fig 8(a) and (c). These observations indicated that the formation of oxide is the same for all tested quaternary alloys; that is, TiO2 and Al2O3 formed layer by layer, although in some alloys the outer thin TiO2 was missing because of spallation.
Oxide formation in Ti alloys is well known from previous studies3,7): Oxygen diffuses into pure Ti, forming a wide oxygen diffusion zone; and TiO2 forms in the outer layer. As the Al content increases, the thickness of the oxygen diffusion zone decreases and a multilayer oxide consisting of TiO2 and Al2O3 is formed. The number of the TiO2 and Al2O3 layers increases with increasing exposure time and temperature3).
In this study, the phase constituents and microstructures are different in Ti-10Al-2Nb-(1, 2)Zr and Ti-15Al-2Nb-(1, 2)Zr alloys; that is, the α single phase exists in Ti-10Al alloys and the α2 phase forms as fine precipitates in the α phase in Ti-15Al alloys. Even with this difference in phase constituents and microstructures, the formation of oxide is similar in both alloys. The TiO2 was the main oxide with small particles of Al2O3 and a thin layer of Al2O3 or TiO2 of about 100 nm formed on this mixed layer in some cases. A composition of Nb and Zr less than 1 at% represents that solid solution of Nb and Zr to TiO2 is very small or fine Nb and Zr oxides distribute in TiO2 although Nb oxide or Zr oxide was not detected in the oxide layer in either composition analysis or in X-ray diffraction analysis. A zone that had been depleted of Ti, Al, Nb, and Zr was observed under the oxide layer as shown in Figs. 7 and 8. This indicates that outward diffusion of ions of Ti, Al, Nb, and Zr occurred.
The process of formation of oxide in α Ti alloys with/without α2 precipitates is believed to be as follows (Fig. 9): Initially, TiO2 forms preferentially on the surface of Ti alloy because of the higher activity of Ti compared to Al (stage 1)3). Then, outward diffusion of Ti4+ and Al3+ cations occurs through the TiO2, and TiO2 grows together with a small amount of Al2O3 (stage 2). At this stage, Nb5+ and Zr4+ cations also diffuse outward, but the amount of outward diffusion of Nb and Zr is small. Then, Nb5+ and Zr4+ dissolve into TiO2 or form fine Nb or Zr oxides in TiO2. After the growth of TiO2, a thin layer of Al2O3 starts to form—perhaps because the supply of Ti was delayed or the activity of Al became high at the air/TiO2 interface as shown in Ref. 3) (stage 3). However, the Al content in Ti alloys is low, and the formation of stable, thick Al2O3 is difficult. Then, TiO2 again forms on the Al2O3 (stage 4). In Ti-6Al-4V, similar oxidation behavior was observed; that is, an Al2O3 layer formed outside of the TiO2, and a TiO2 layer and Al2O3 layer formed alternately3).
Formation process of oxide in α Ti alloys.
In our study, the weight gain of ternary Ti-Al-Nb and Ti-Al-Zr alloys during the oxidation test was smaller than that of conventional Ti-6242 alloys as shown in Fig. 5. The addition of Zr to the Ti-Al-Nb alloys decreased the weight gain even more (Fig. 5(b)). Our results indicate that the addition of Zr to Ti-Al and Ti-Al-Nb alloys is a very effective way to improve oxidation resistance. Although there are not many studies on the effects of Zr on oxidation behavior of α-Ti alloys, several investigations on the effect of Zr on the oxidation behavior of pure Ti or TiAl have been reported11,14). The mass gain of Ti-Zr and Ti-Nb binary alloys was investigated and found to decrease with increasing Zr content. The suppression of oxide growth of Nb to pure Ti was more effective than Zr. It is well known that TiO2 is a non-stoichiometric compound and mainly grows by inward diffusion of oxygen through oxygen vacancies15). It is believed that the Nb5+ cation decreases the number of oxygen vacancies to maintain the electro-neutrality in TiO2 because of the higher valance of Nb compared to Ti. As a result, suppression of oxide growth occurs15). On the other hand, since the valance of Zr (Zr4+) is the same as that of Ti (Ti4+), the decrease of oxygen vacancies is considered to be smaller than in the case of Nb, resulting in less suppression of oxide growth. However, in our study, Zr addition to Ti-Al alloys was effective to suppress oxide growth. If Al3+, with a smaller valance than Ti4+, is substituted in TiO2, oxygen vacancies will be increased to maintain the electro-neutrality. However, it is well known that Al addition suppresses the formation of TiO2. It is said that if Al3+ occupies interstitial positions, oxygen vacancies will decrease16). The existence of Al in TiO2 is believed to change the electrical field in the TiO2; as a result, the behavior of Zr4+ and Nb 5+ is also changed and Zr4+ will decrease oxygen vacancy in TiO2. Although further investigation is necessary to understand the role of Zr clearly, at least it can be said that the effect of Zr addition is different in Ti-(10–15)Al alloys and in pure Ti. Zr in TiO2 suppresses the growth rate of TiO2 in Ti-(10–15)Al alloys. On the other hand, the effect of Zr on TiAl is enhancement of formation of Al2O3 and suppression of growth of TiO2, which is different mechanism for improvement of oxidation resistance in Ti-10–15Al alloys14).
(1) A single α phase was formed in Ti-10Al-2Nb-(1, 2)Zr while α and α2 phases were formed in Ti-15Al-2Nb-(1, 2)Zr; however, the formation of oxide was similar regardless of differences in the constituents and microstructures.
(2) The addition of Zr by itself greatly improved the oxidation resistance of Ti-(10–15)Al alloys. Also, the addition of Nb together with Zr improved the adhesion of the oxide layer.
This research was done under a project titled “High-temperature materials and surface technologies for efficient energy conversion”. The authors thank Dr. D-H. Ping for performing the TEM observations.
