Fine-tuning of stress-induced martensite in TRIP/TWIP Ti alloys

Fine-tuning of stress-induced martensitic (SIM) transformation was studied in Ti-Mo based β metastable alloys, showing combined Transformation Induced Plasticity (TRIP) and Twinning Induced Plasticity effects (TWIP) effects. The work aimed to clarify the transition and interaction between the two deformation mechanisms and their influences on the mechanical properties of Ti-Mo based alloys. Electron parameter design methods (Bo-Md and e/a ratio) were cross-used to increase the β phase stability from near-TRIP to near-TWIP by adding third alloying elements. SIM α″ transformation and mechanical twinning were traced by in-situ EBSD (Electron backscatter diffraction) mapping under cyclic tension. The deformation modes of β phase exhibited significant changes when shifting its metastability via chemical composition modifications. In near-TRIP conditions (dominated by the growth of SIM α″ in plastic regime), SIM α″ transformation and internal twinning of martensite were the main mechanisms to accommodate the local stress-strain conditions. In near-TWIP ones (dominated by the growth of {332}<113>β twinning in plastic regime), SIM α″ was observed only at twinning interface during strain process and disappeared after stress release.


Introduction
Since the last few years, a new family of metastable β titanium alloys with TRansformation Induced Plasticity and/or TWinning It is well-known that the activation of mechanical twinning and SIM closely depends on the stability of the β phase. Martensitic transformation (β ® α″) can be activated by applying external stress in the β-metastable alloys when the chemical stability of β phase (usually estimated by Mo-equivalent content, Mo eq wt.%) is relatively low (a rough range of 10.5% £ alloy Mo eq £ 13%) [8]. Mechanical twinning can be activated when the β stability becomes higher (roughly 12% £ alloy Mo eq £ 15%) [8]. The overlapping of the two ranges allows the simultaneous occurrence of TRIP and TWIP effects in the same alloy. Furthermore, a continuous transition of deformation mechanisms, from near-TRIP to near-TWIP, can be expected by increasing the chemical stability of a TRIP/TWIP alloy.
Zr has been found to be a neutral element (in Ti-Zr system) or weak β-stabilizing element (in Ti-Nb-Zr system) [9,10]. Previous studies reported that the nature of Zr in a Ti-M-Zr system (M = β-stabilizer alloying element) strongly depends on the β-stability and concentration of M [9]. The electron parameter (Bo-Md and e/a ratio) design methods suggest that the addition of Zr could 2 MATEC Web of Conferences 321, 11038 (2020) https://doi.org/10.1051/matecconf/202032111038 The 14 th World Conference on Titanium shift gently TRIP/TWIP alloys towards TWIP region [11]. In order to study the effect of Zr addition on the TRIP to TWIP transition in Ti-Mo system, a comparative investigation is performed between the typical TRIP/TWIP Ti-12Mo (wt. %) [1,12] alloy and two Ti-12Mo-xZr (x = 5, 10) (wt. %) alloys designed by Bo-Md method.
Electron Backscattered Diffraction (EBSD) characterizations coupled with in-situ tensile deformation are employed to study the effects of Zr addition on the deformation mechanisms. The in-situ method is used to reveal the formation, development, and reversion of martensite/twin during the tensile loading and unloading processes.

Experimental
The 200g ingots with nominal compositions of Ti-12Mo-xZr (x= 0, 5, 10) were fabricated by vacuum arc-melting using pure Ti, Mo and Zr metals (purity > 99.9wt.%). Each ingot was remelted for five times and homogenized at 1373 K for 72 ks under a high vacuum of 10 -4 Pa, followed by water quenching. The as-quenched ingots were cut to a billet of 10mm in thickness 65mm in diameter. The billet was cold-rolled to 0.5mm in thickness (reduction rate of 95%). Subsize tensile samples (ASTM E8M) and insitu samples (L40mm, W2mm, T0.5mm) were cut from the as-rolled sheet along the rolling direction. The samples were solutiontreated (ST) at 1173 K for 1.8 ks under high vacuum then water quenched.
The uniaxial tensile tests were performed with a constant strain rate of 10 −3 s −1 using INSTRON 5982 machine. The in-situ samples were polished to mirror-like surface finishing. In-situ EBSD scans were performed using a field emission gun scanning electron microscope (FEG-SEM Zeiss LEO1530 with a high-speed Nordif UF-1100 EBSD camera) operating at 20kV. The indexation of EBSD raw data was performed by TSL OIM Analysis Software. The EBSD scans were performed by holding the tensile strain before loading, during loading and after unloading on a Proxima 100-Micromecha in-situ tensile test machine. The tensile properties of the 3 alloys are listed in table 1. The YS and UTS strongly increase with the Zr content, whereas the ductility reduces. The strengthening might be due to solution strengthening effect of Zr addition. The reduction of ductility could be caused by many reasons, one of them could probably be the change of deformation mechanism.   It might suggest that the "green" SIM α″ is the primary product originated from direct β ® α″ transformation [13]. However, no such relationship can be found between the "purple" α″ variant and β matrix. In contrast, a classical {130}<310> α″ twin relationship was found between the "green" and "purple" α″ variants. This suggests that the {130}<310> α″ mechanical twinning occurred in α″ SIM during the straining. Such twinning mode has already been reported by Bertrand et al. [14] in deformed Ti-25Ta-20Nb wt.% alloy. After loading to ε = 0.04 and fully unloaded, the "purple" α″ variant disappeared by transforming partially back to "green" α″ variant and partially reversed to β phase ("pink" β phase in fig. 1c). This "pink" β phase exhibits a {130} <310> α″ twin type misorientation with the β surrounding matrix ( fig. 1c). This phenomenon has been reported by Castany et al. [5] in Ti-Nb system. In the present case, one can considered that {332}<113> β twins are thus not formed directly from the β phase but are the result of the reversion of {130}<310> α″ parent twins occurring in α″ SIM during unloading.   fig. 2a zoomed in fig. 2b) was thus followed by extending the in-situ loading up to ε = 0.1. The thickening of the α″ martensite band was recorded up to about 10 μm and remained single variant ("purple" α″ variant). The crystallographic relationship analysis between this "purple" α″ and its surrounding b matrix, again does not satisfy the classical b/ α″orientation relationships. It can be noticed that the image quality of the band at ε = 0.1 (Fig. 2a) was affected due to important surface relief. In order to improve the EBSD map quality, a fine re-polishing (depth < 5 μm) was applied to the unloaded sample to remove the surface relief. The EBSD map of the same band after re-polishing (Fig. 2c) presented two variants, a "green" and a "purple" variant, in which the "purple" variant was the same variant observed under load at e= 0.1 (Fig. 2a and 2b). The crystallographic relationships among the "green" α″ variant, "purple" α″ variant and β matrix were identified as:
This analysis suggests again that, during loading, the b phase transformed into α″ martensite which immediately mechanically twinned with the {130}<310> α″ mode. This situation is very similar to the above Ti-12Mo except that the martensite did not reverse into b phase when the sample was unloaded. The observed deformation mechanism of Ti-12Mo-5Zr was then close to that of Ti-12Mo at the early stage of the plastic regime. Addition of 5wt.% Zr did not strongly affect the occurrence of the α″ SIM and its mechanical twinning. This suggests that 5wt.% Zr introduced noticeable solid solution strengthening but did not change significantly the β stability of Ti-12Mo. In contrast, the deformation mechanisms appeared to be significantly modified by 10wt.%Zr addition. Figure 3 shows in-situ EBSD maps of deformed Ti12Mo10Zr sample.
Almost all the observed major deformation bands were {332}<113> β from yielding to ε = 0.04. The {332}<113>β ("blue" variant in Fig. 3a) became the dominant deformation mechanism in the Ti-12Mo-10Zr sample. A minor volume fraction of α″ SIM ("green" variant in Fig. 3a and 3b) was observed, always connected to β twin. In order to figure out if the α″ SIM is from b matrix or {332}<113> β twin, the orientation relationship analysis has been carried out. The results showed that the α″ SIM yielded the classical β®α″ martensitic transformation relationships, whilst no classical relationships were found between the α″ SIM and the adjacent {332}<113> β twin. This suggested that the α″ SIM transformed from β matrix but nucleated at {332}<113> β twin interface. When unloaded, the α″ SIM fully transformed back to the β matrix (Fig. 3c), suggesting that its formation probably acts as a relaxation mechanism to accommodate local stress which happened usually near twin interfaces and their intersections.
The fact that the primary deformation mechanism changes from α″ SIM to {332}<113>β mechanical twinning of the β phase when the Zr amount increases, suggests that this alloying element increases the critical shear stress of SIM α″, confirming the βstabilizing character of this element in Ti-Mo alloys. This stabilization of the β phase limits the formation of α″ SIM to high-stress level volumes. This could explain the observed ductility reduction of Ti-12Mo-10Zr because of the partial suppression of relaxational effect of the α″ SIM.
According to the comparative study of the samples, two effects can be noticed of the Zr addition to the SIM in TRIP/TWIP Ti-12Mo: i. no obvious modification up to 5wt.% Zr addition; ii. significant reduction of SIM with 10wt.% Zr addition. Zr started to increase the β phase stability of Ti-12Mo when > 5wt.%, leading to the shift of the deformation mechanism from TRIP/TWIP to TWIP-dominant. Therefore, a fine-tuning of the SIM could probably be achieved by controlling Zr content in metastable Ti-Mo-Zr system.

Conclusion
Several effects of Zr additions were observed on the mechanical properties and deformation mechanisms of Ti-12Mo-xZr alloys. A noticeable solid-solution strengthening effect was obtained with both 5% and 10% Zr additions. A shift of deformation mechanism from combined TRIP/TWIP to TWIP-dominant was observed when Zr content is increased from 5% to 10%. The fine-tuning of the volume fraction of SIM can be achieved in the 5% -10% Zr range, in which Zr acts as a b-stabilizer in Ti-12Mo. Thanks to insitu tensile tests connected to EBSD mapping, the first evidence of {130}<310> α″ twins reverse transformation to {332}<113> β twin was observed in Ti-12Mo during an in-situ loading-unloading cycle. New insights about the SIM-Twinning interactions were revealed in the complex TRIP/TWIP microstructure evolution.