*3.3. Thermal Stability of* ω*-Ti(Fe) Produced by the HPT Process*

Complementary DSC and high-temperature XRD measurements were performed for the description of the thermal stability of the deformation-induced ω-Ti(Fe) phase. The DSC curves of the HPT-deformed samples were recorded upon heating at the heating rate of 10 K/min, and they are shown in Figure 4. The HTXRD patterns (Figure 5) were measured at the same heating rate (10 K/min). For HTXRD, the sample temperatures were calibrated with the aid of the initial temperature at the beginning of the measurements (25 ◦C) and while using the temperature of the endothermic effect at 562 ◦C that corresponds to the eutectoid reaction β α-Ti + TiFe and the transformation of the intermetallic phase TiFe. The transformation of TiFe can be detected by both techniques, DSC and XRD. The reaction temperature should be the same for all of the investigated alloys due to the invariance of the eutectoid reaction. Therefore, the same temperature calibration procedure can be applied to all investigated alloys.

Prior to the DSC and HTXRD measurements, all of the samples contained a mixture of α-Ti(Fe), TiFe and ω-Ti(Fe). The denotation α-Ti(Fe) emphasizes an increased iron solubility in α-Ti, due to (i) the HPT process and (ii) the reconversion of ω-Ti(Fe) → α-Ti(Fe) upon heating. These phenomena will be discussed in detail below. The first DSC effect observed upon heating was an exothermic peak occurring at approx. 130 ◦C (Figure 4), which was accompanied by the sharpening of the originally extremely broad XRD lines from α-Ti(Fe) (Figure 5). Concurrently, the XRD lines from α-Ti(Fe) became more intense at the expense of the XRD lines from ω-Ti(Fe), which indicates the onset of the reconversion of ω-Ti(Fe) to α-Ti(Fe). At 320 ◦C, the phase transition ω-Ti(Fe) → α-Ti(Fe) proceeds tremendously. The transition is completed at temperatures that were slightly above ~350 ◦C. In contrast to HTXRD, DSC did not recognize the end of the decomposition of ω-Ti(Fe), because it is not accompanied with a noticeable thermal effect. The ω-Ti(Fe) → α-Ti(Fe) transition is a continuous process, thus the heat release is spread over a broad temperature range.

After the decomposition of ω-Ti(Fe), all of the samples exhibited a two-phase microstructure containing hexagonal α-Ti(Fe) and cubic TiFe. The appearance of an additional diffraction line (0002 of α-Ti, c.f., Figure 5) at lower diffraction angles and the presence of anisotropic (*hkl*-dependent) line broadening upon further heating indicate changes in the Fe concentration in the α-Ti(Fe) phase. The incorporation of Fe into the hexagonal α-Ti lattice mainly leads to a reduction of the lattice parameter *c*α-Ti, while the lattice parameter *a*α-Ti remains nearly unaffected [25,27]. Accordingly, the diffraction line 0002, which appears at temperatures above 400 ◦C, and that is located at a lower diffraction angle in Figure 5b, corresponds to the equilibrium α-Ti phase that only exhibits a negligible solubility for Fe [14]. On the other side, the hexagonal α-Ti(Fe) phase, which is present in all alloys after the HPT process and that is formed by the back-transformation of ω-Ti(Fe), should possess increased iron solubility.

**Figure 4.** DSC heating curves of alloys Ti-2Fe, Ti-4Fe, and Ti-10Fe measured with the heating rate of 10 K/min. The dashed lines at 130 ◦C and 320 ◦C indicate the beginning and the end of the ω back-transformation as concluded from high-temperature X-ray diffraction (HTXRD). The dashed line at 562 ◦C marks the eutectoid reaction β-(Ti,Fe) α-Ti(Fe) + TiFe. The temperatures marked by crosses indicate the β-transus temperatures (solvus temperatures of the β phase), which were determined as inflection points of the respective DSC curves.

**Figure 5.** Low-angle part of the HTXRD patterns of Ti-2Fe, Ti-4Fe, and Ti-10Fe that were originally annealed for 4000 h at 470 ◦C and subjected to HPT. The positions of diffraction lines originating from the phases α-Ti(Fe), β, ω and TiFe are indicated by markers at the top or inside the figures. The temperature axes of the HTXRD measurements were calibrated according to the DSC measurements as described in the text. The dashed lines in (**a**) indicate the transformation temperatures upon heating from Figure 4. (**b**) illustrates the change of the 0002 line position in Ti-10Fe (see inset in (**a**)) at high temperatures.

At the temperature of 562 ◦C, the mixture of α-Ti, α-Ti(Fe) and TiFe transforms via an eutectoid reaction into a two-phase mixture of β-(Ti,Fe) and α-Ti (Figure 5). In the DSC measurements (Figure 4), the temperature of the eutectoid reaction was determined from its onset point. Above the eutectoid temperature, the amount of β-(Ti,Fe) continuously increases with further heating. The transus temperatures of β-(Ti,Fe) determined using DSC depends on the iron content in the respective alloy (Figure 4). A comparison of the eutectoid temperature and the β-(Ti,Fe)-transus temperatures that were measured for the heat-treated (Figure 2) and severe plastically deformed samples (Figure 4) reveals that both temperatures are slightly shifted towards lower values after the HPT process. Figure 6 shows a comparison of the transition temperatures measured by DSC before and after HPT.

**Figure 6.** Partial phase diagram of the Ti-rich corner of the binary Ti–Fe system. The circles indicate the measured phase transition temperatures (DSC) of the samples in the initial state (red) and after deformation by HPT (green).

The sequence of the phase transitions is illustrated on the temperature dependence of the integral intensities of the XRD lines measured for individual phases (Figure 7), i.e., 1011ω/1120ω, 110*TiFe*, 110β, and 1010α/0002α/1011α. For ω-Ti(Fe) and α-Ti(Fe), the sums of the integral intensities of the measured lines were considered. For convenience, the integral intensities were converted into the phase compositions by normalizing the phase composition of the respective alloy to the phase composition from Table 2. This 'external standard' method neglects the effect of the possible changes in the preferred orientation of crystallites due to the sample recrystallization and the effect of the Debye–Waller factor on the phase composition, as the corresponding factors influencing the diffracted intensities are assumed to remain constant. However, it gives a good overview of the phase transitions, as can be seen from the almost monotonous and definitely reasonable trend of the TiFe phase fraction.

The beginning of the transformation ω-Ti(Fe) → α-Ti(Fe), which was observed during the DSC measurement as an exothermic effect at 130 ◦C, results in a rapid decrease of the ω-Ti(Fe) phase fraction and in a concurrent increase of the α-Ti(Fe) phase fraction. In alloy Ti-4Fe, the DSC peak from the exothermal effect is slightly shifted towards higher temperatures. Upon further heating, ω-Ti(Fe) continuously transforms into α-Ti(Fe). Above a temperature of 400 ◦C, all of the alloys exhibit a two-phase microstructure containing the hexagonal α-Ti(Fe) and the cubic TiFe phase. The phase amount of TiFe increases with increasing Fe content in the alloy. At temperatures above 500 ◦C and below the eutectoid temperature, the supersaturated α-Ti(Fe) releases iron, which is subsequently solved in the cubic β-(Ti,Fe) phase. Thus, β-(Ti,Fe) was formed by this process already below the eutectoid temperature. At 562 ◦C, the eutectoid reaction occurs and TiFe transforms into the equilibrium phases α-Ti + β-(Ti,Fe). A further temperature increase should lead to a continuously increasing phase amount of β-(Ti,Fe) and a decreasing amount of α-Ti. This behavior was not observed during the

HTXRD measurements, which is caused by the proceeding oxidation of the samples, which stabilizes α-Ti. Therefore, the results of HTXRD measurements can be compared with the results of the DSC measurements only up to temperatures of ~650 ◦C. In the alloys Ti-2Fe, Ti-4Fe, and Ti-10Fe that were subjected to HTXRD, the CGHE analysis revealed the oxygen concentrations of 0.0374(3) wt.%, 0.0262(3) wt.%, and 0.0032(3) wt.%, respectively, which are much higher than prior to HTXRD.

**Figure 7.** Temperature dependences of the phase composition in alloys Ti-2Fe (**a**), Ti-4Fe (**b**), and Ti-10Fe (**c**) obtained from the HTXRD measurements carried out upon heating. The integral intensities were normalized to the phase compositions at the initial state of samples after HPT (Table 2). The intensities of the diffraction lines 1010, 0002, and 1011 of α-Ti and α-Ti(Fe), and 1011 and 1120 of ω-Ti(Fe) were summed up. The temperature axes of the HTXRD measurements were calibrated according to the DSC measurements.
