**1. Introduction**

Titanium and titanium-base alloys are promising materials for numerous engineering applications, because they have several outstanding properties [1]. In particular, binary Ti–Fe alloys are in the focus of ongoing research as materials for biomedical applications, because they possess excellent corrosion resistance, high wear resistance, biocompatibility, and appropriate mechanical properties, for instance, a low elastic modulus in comparison with other biocompatible metallic materials [2–4]. Furthermore, the binary intermetallic phase TiFe has been presented as a possible material for solid-state hydrogen storage applications [5,6]. Some methods of severe plastic deformation (SPD), such as ball milling and high-pressure torsion (HPT), were utilized to reduce the surface oxidation and to activate the material in order to improve the hydrogen storage properties of TiFe [7,8]. Originally, the SPD methods

were utilized to generate bulk nanocrystalline materials that show, in many cases, better physical and mechanical properties than their microcrystalline counterparts [9,10]. Later, it turned out that the unique properties are also facilitated by diffusive and displacive (martensitic) phase transformations, which occur in the material during the HPT process [11–13]. However, the mechanism of the phase transformations and the influence of the initial microstructure on the phases formed after HPT are not fully understood yet.

In the unary Ti system, titanium exists in three modifications: as hexagonal α-Ti (space group (SG) *P*63/*mmc*) that is stable at low temperatures, as cubic β-Ti (SG: *Im*3*m*), which is stable at high temperatures, and as hexagonal ω-Ti (SG: *P*6/*mmm*), which is stable at high pressures. In the binary Ti–Fe system, two intermediate intermetallic phases TiFe (SG: *Pm*3*m*) and TiFe2 (SG: *P*63/*mmc*) are formed in addition to the phases that are summarized above and in addition to α-Fe/δ-Fe (SG: *Im*3*m*) and γ-Fe (SG: *Fm*3*m*). Furthermore, at least two metastable phases, which are formed during a martensitic transformation upon quenching from the bcc-type β-(Ti,Fe) solid solution, were reported in the Ti-rich part of the Ti–Fe system. The first one is a close-packed hexagonal α´-Ti phase (SG: *P*63/*mmc*) [14–17], which is observed in Ti-rich alloys (>95 wt.% Ti), the other one is the so-called athermal ω-Ti(Fe) phase that exists and in a limited compositional range between 97 wt.% and 95 wt.% Ti [18–22]. For higher Fe concentrations (and lower Ti contents), β-(Ti,Fe) is retained as a metastable phase in the alloys.

Phase transitions and respective transformation pathways that were induced by high-pressure torsion were reported for Ti-rich Ti–Fe alloys containing 1–10 wt.% Fe in different initial states (as-cast [23] and heat-treated [24–28]), and, thus, for different phase compositions prior to the HPT process. The as-cast alloys contained a mixture of α-Ti and β-(Ti,Fe) that were partially transformed into ω-Ti(Fe) during the HPT process. The heat treatments were performed almost exclusively above the eutectoid temperature of ~595 ◦C (β-(Ti,Fe) α-Ti + TiFe) [14], i.e., in the single-phase β-(Ti,Fe) or in the two-phase α-Ti + β-(Ti,Fe) regions. After annealing, the samples were quenched in water. The retained amount of the high-pressure ω-Ti(Fe) phase after HPT and the transformation pathway varied, depending on the initial phase fractions and the chemical compositions of the quenched phases [26]. Heat-treated alloys with Fe contents below 4 wt.% contained α -Ti martensite and/or metastable β-(Ti,Fe) after quenching [24–28]. During HPT, α -Ti and β-(Ti,Fe) transformed partially to ω-Ti(Fe).

The Fe content in β-(Ti,Fe) depends on the overall Fe concentration in the Ti–Fe alloys and it can vary in a relatively broad range. However, if β-(Ti,Fe) contains ~4 wt.% Fe, athermal ω-Ti(Fe) can be formed within the β-(Ti,Fe) grains after quenching [28,29]. The phase transformation β-(Ti,Fe) → ω-Ti(Fe) is promoted for an iron content about 4 wt.% Fe, because both crystal structures possess a strong orientation relationship (OR) {111}β||(0001)<sup>ω</sup> and 110β||1120<sup>ω</sup> [18,22], and because the atomic distances within the habitus planes match perfectly together at this composition. Thus, this phase transformation proceeds diffusionless by shearing the crystal structure of β-(Ti,Fe) in the HPT process [26]. For lower Fe contents (≤2 wt.% Fe), the HPT process induces an incomplete ω-Ti(Fe) phase transition [25,28], because the transition of α -Ti to ω-Ti(Fe) dominates the phase transformation process, which involves the redistribution of iron atoms between α -Ti and ω-Ti(Fe). The mass transfer impedes finally the phase transformation [26,28].

So far, the thermal stability of HPT-induced ω-Ti(Fe) was predominantly investigated in the samples, which were annealed above the eutectoid temperature (~595 ◦C) prior to the HPT process and that contained α-Ti, β-(Ti,Fe) and ω-Ti(Fe) in different ratios after the HPT treatment [27,28]. In these samples, the HPT-induced ω-Ti(Fe) phase transformed upon heating at 380 ◦C into a supersaturated hexagonal α-Ti(Fe) phase. Above 600 ◦C, α-Ti(Fe) decomposed into the equilibrium phases α-Ti and β-(Ti,Fe), whose fraction varied with the overall composition of the alloy [27].

The thermal stability of ω-Ti(Fe) that was generated by HPT in samples annealed below the eutectoid temperature, i.e., in the α-Ti + TiFe two-phase region, was not investigated in detail yet. Thus, only very little is known regarding the phase transformations in the Ti–Fe alloys with this phase composition. Still, in Reference [24], the binary alloy Ti-4Fe (4 wt.% Fe) was heat-treated

below the eutectoid temperature at 470 ◦C for 750 h and subsequently subjected to HPT. After initial annealing, this alloy contained a mixture of the α-Ti and TiFe phases, which partially transformed into ω-Ti(Fe). Thus, the HPT-deformed microstructure consisted of α-Ti, TiFe and ω-Ti(Fe). In the present work, the HPT-induced formation of ω-Ti(Fe) in initially two-phase alloys (α-Ti + TiFe) is more systematically studied and the stability of the metastable ω-Ti(Fe) phase is investigated upon heating. The transformation pathway was derived from the results of in situ high-temperature X-ray diffraction (HTXRD) and thermal analysis (TA). The experiments were complemented by pressure-dependent thermodynamic calculations that were based on the CalPhaD (calculation of Phase Diagrams) approach, which helped to understand the phase transformations in this system during deformation and heating.
