High temperature oxidation and mechanical behavior of β21s and Ti6242S Ti-based alloys

Aircraft industry always looks for higher in-service temperatures and lighter structures. With a high specific strength, Ti-based alloys are good candidates for such applications. However, when exposed to oxidizing environments at high temperatures, they undergo large oxygen dissolution while forming an oxide scale, which can greatly affect their mechanical properties. Then, evaluating the oxidation resistance and mechanical behavior of such alloys is essential. In this aim, long term oxidation tests were performed under laboratory air between 500 °C and 625 °C on two Ti-based alloys: β21s, exhibiting a fully β microstructure supposed to dissolve lower amount of oxygen and nitrogen, and Ti6242S, with an α/β microstructure. The oxidized samples were characterized using XRD, Raman spectroscopy, SEM-EDS and micro-durometer. As for the mechanical behavior, tensile tests were performed at room temperature on not aged and on oxidized samples. While larger mass variations were obtained at 500 and 560 °C and up to 997 h at 625 °C for β21s, its mass variations became lower than those of Ti6242S for longer durations at 625 °C. Nevertheless, β21s exhibited thicker micro-hardness affected depths and underwent larger mechanical property modifications compared to Ti6242S.


Introduction
For many years, aircraft industry has been looking for higher in-service temperatures and lighter structures, needs that become even more marked today because of environmental norms. Ti-based alloys are good materials for such applications, because of their high specific strength [1,2]. Nevertheless, they undergo large oxygen (and nitrogen) dissolution in the same time as they form an oxide scale when they are exposed to oxidizing environments at high temperatures [3,4]. This oxygen diffusion, because it can greatly affect their mechanical properties [5][6][7], limits the in-service temperature of such materials. It is then essential to evaluate the oxidation resistance and mechanical behavior of oxidized Ti-based alloys. To do so, long term oxidation tests were performed under laboratory air between 500 °C and 625 °C on two Ti-based alloys: β21s and Ti6242S alloys. Few oxidized samples were analysed to characterize the formed oxide scale and the oxygen-rich metallic layer underneath. Others were mechanically tested and compared to not aged samples in order to evaluate the impact of environment on the resistance to tensile stresses.

Materials
Two Ti-based alloys were studied: Ti6242S and β21s in the form of sheet material. Energy dispersive X-ray spectroscopy (EDS) was used to determine the alloy composition in heavy elements while instrumental gas analysis (IGA) and glow discharge mass spectrometry (GDMS) analyses were realized to measure light element contents by Evans analytical Group SAS (Tournefeuille, France). Obtained compositions are reported Table 1. The alloys contained similar oxygen amounts (about 1000 ppm). The largest differences were noticed for C and N ; β21s alloy was richer in nitrogen.
Ti6242S alloy was hot rolled below β transus and duplex annealed as per AMS4919 (899°C/30min air cooling + 788°C/15 min air cooling) leading to an α/β globular microstructure, see Figure 1.a. β21s alloy was also hot rolled. However, as no solution annealing and ageing heat treatment was applied to precipitate the α-phase, the alloy exhibited a fully β microstructure, see  Two rolled sheets were available, 3.97 mm and 1.85 mm thick for Ti6242S and β21s respectively. 19.7 x 19.7 mm squared specimens were cut from these sheets, and their surface was ground using P240 SiC paper before a cleaning in an ultrasonic bath of acetone and then of ethanol for the oxidation tests.
Tensile test specimens (flat specimens, 12.5 mm in width) were cut out from the sheets, 3.97 mm and 1.85 mm thick for Ti6242S and β21s respectively. No polishing or grinding step was made so that the specimen surface remained in the as delivered condition. Prior to ageing and tensile test, specimens were cleaned in an ultrasonic bath of acetone and then of ethanol to be oxidized.

Oxidation tests
Samples were placed on ceramic bricks in Nabertherm N60/85HA and N120/85HA furnaces. Both alloys were oxidized 2958 h at 500, 560 and 650 °C under laboratory air. A Sartorius Genius (ME215-P) balance with an accuracy of 20 µg was used for mass measurements before and after oxidation. Each recorded mass was the average of two weightings.

Tensile tests
Tensile tests were performed at room temperature according to EN2002-1, using a strain rate equal to 2 mm/min up to failure.
For each alloy, three samples were oxidized before mechanical testing for 2958 h at 625 °C under laboratory air, with the squared specimens used for oxide scale and oxygen diffusion characterization. Their mechanical behavior was compared to the one of three unaged specimens of corresponding alloy. Characterization X-ray diffraction analyzes (XRD) were performed using a Seifert 3000TT apparatus with a copper anti-cathode (λ = 1.54056 Å).
Raw materials were analyzed using a theta-theta mode between 20 and 120 ° (2q) with a step of 0.04 ° and an acquisition time of 3 s per step. Oxidized sample analyses were done between 20 and 80 ° (2q) with a low incidence angle (between 2 and 8 °), a step of 0.05 ° and an acquisition time of 10 s per step.
Raman and fluorescence spectroscopy analyses were made using a Labram HR 800 spectrometer from Horiba Yvon Jobin, equipped with a confocal microscope. Spectra were recorded using a laser with a 532 nm line.
Cross-sections were made starting by a pre-coating with an epoxy resin to protect the oxide scale during the cut. The grinding and polishing sequence started with a P180 up to a P2400 SiC paper and finished using OPS.
SEM observations of the surface and cross-section of oxidized and reference samples were done with a LEO 435VP microscope using the secondary electron imaging mode (SE) or the backscattered electron imaging mode (BSE). EDS analyses were done with an IMIX system from PGT and quantification was based on real standards. To reach higher magnifications to observe the aged microstructures, a JSM-7800F FEG-SEM from JEOL was used while the fracture surface of the tensile specimens was performed with a FEI Quanta 600 microscope. Vickers micro-hardness measurements were realised using a Buehler Omnimet 2100 serie with a 50 g diamond indent. To estimate the depth of the micro-hardness affected zone, the reference value was chosen as being the average value of microhardness measurements obtained in the non-affected zone, close to the bulk (imprints exhibiting an un-explained high value in the bulk were not considered to evaluate the reference value). Then, micro-hardness affected zone was evaluated as the depth where hardness was at least 25 Hv higher than the reference value.

Oxidation tests
The net mass changes measured during long isothermal oxidation tests are presented on Figure 2. While both alloys underwent relatively small mass variations after 2958 h at 500 °C and 560 °C (inferior to 1 mg/cm 2 ), mass variations became more marked during the exposition at 625 °C. Besides, alloy ranking was reversed between 625 °C and the lower oxidizing temperatures. At 500 °C and 560 °C, β21s alloy exhibited larger mass changes compared to Ti6242S alloy (1.8 and 1.7 times larger than the one of Ti6242S at 500 and 560 °C respectively). At 625 °C, β21s maintained a higher mass change than Ti6242S alloy up to 997 h.
However, for longer durations at this temperature, the mass variation of Ti6242S alloy exceeded the one of β21s alloy, and became 1.3 times larger than the one of β21S at 2958 h. XRD, Raman spectroscopy and fluorescence spectroscopy analyses were performed to determine the nature of the oxide scale. 5 pathways of α phase without β precipitates inside may lead to a rapid oxygen diffusion in depth in β21s alloy. This α-phase network can "irrigates" the whole microstructure of β21s alloy. Nevertheless, as the overall α phase volume fraction of β21s is smaller than the one of Ti6242S, β21s dissolves a smaller mass of oxygen. This deep oxygen diffusion was then responsible for the larger decrease in elongation of the alloy when tested under tensile stresses. Indeed, Porter et al. observed a good agreement between the crack depth compared to the micro-hardness affected zone and the oxygen-rich layer measured after etching [10].

Conclusions
Two titanium-based alloys were oxidized between 500 and 625 °C for 2958 h: Ti6242S and β21s alloys. Samples oxidized at 625°C were also mechanically tested and compared to non-aged materials. Oxide scale compositions and thicknesses were close but the microstructures were very different between alloys. Ti6242S alloy had an α/β globular microstructure which did not significantly evolved. In the contrary, β21s microstructure, initially β single-phased, exhibited a basketweave microstructure within the former β grains and a wide α phase network, imprint of the previous β grain boundaries. Ti6242S alloy showed smaller mass variations than β21s alloy at 500 and 560 °C. However, it exhibited greater mass variations after 1000 h at 625 °C but its ductility was less decreased than the one of β21s alloy when tested under tensile stresses after an oxidation of 2958 h at 625 °C.
With a similar or lower quantity of oxygen dissolved, ductility of β21S was more affected by the environment. From the microhardness profiles, and from the observation of fracture surfaces and crack lengths, it seems that oxygen diffused deeper in the β21S alloy. This could be due to the continuous network of α phase at the former β grain boundaries.