Chemical origins of β-Ti stabilization via B₄C additions in metastable β-Ti alloys and composites

Abstract

The aerospace industry relies on Ti alloys owing to their strength-to-weight ratio and corrosion resistance. In metastable β-Ti alloys, slow cooling from the β-transus leads to partial transformation into coarse α laths, which is detrimental to the mechanical properties. A refinement and decrease of α laths has been previously achieved in β-Ti alloys with B₄C additions. In these materials, the ductility of β-Ti is preserved, and the TiB and TiC particles promote strengthening. However, the mechanism of the β-Ti stabilization remains unclear. Using atom probe tomography, we propose that Mo enrichment in the α-phase limits its growth by reducing the influx of Al from the β-phase. The complex chemical environment near eutectic TiB is enriched in Al and TiO, promoting heterogeneous nucleation of fine α-phase. Increased TiO concentration is observed with the introduction of B₄C. A fundamental understanding of the α-refinement mechanism in Ti-alloys is critical for aerospace applications demanding high performance and reliability.

Introduction

Ti alloys are essential for aerospace applications due to their high specific strength, corrosion resistance, and thermal stability¹. Increasing demands for ductile and yet strong materials for next-generation aircraft engines are often met via microstructure control of Ti alloys via casting and thermo-mechanical processing². The high ductility of the body-centred cubic β-phase makes β-Ti alloys combined with reinforcement particles an attractive choice. These might include carbides (TiC), borides (TiB), and the hexagonally close-packed α-phase³. The metastable nature of β-phase often results in a β→α phase transformation during typical air or furnace cooling in industrial applications^4,5. Faster cooling rates retain the β-phase but are generally avoided as they are known to introduce undesirable residual stress and/or promote the formation of the athermal ω-phase^6,7. A common strategy to control α-phase formation during slow cooling is to add β stabilizing elements. TiB and TiC have also been shown to be suitable for control of the β→α transformation, in addition to increasing hardness and yield strength properties^8,9.

Beta 21 S is a metastable β-Ti alloy used in the plug-and-nozzle segment of aircraft engines⁵. When reinforced with TiB and TiC and used as a metal matrix composite, it has been demonstrated to deliver more attractive combinations of mechanical properties than the base alloy. A compressive yield strength of 1200 MPa and ductility of > 20% were achieved with a combined volume fraction of 20% of TiB and TiC as discontinuous reinforcements by some of the current authors^10,11. The most surprising observation was the concomitant decrease and refinement of α, i.e., β stabilization with increasing volume fraction of reinforcement particles. It was hypothesized that the consumption of Ti by these particles led to lower Ti availability in the matrix and a subsequent increase in the β-phase stability. Al, an α-stabilizing element, was thought to be rejected by the growing TiB, which caused them to act as heterogeneous nucleation sites for fine α-phase, as has also been observed for other β-Ti alloys and in Ti-6Al-4 V^12,13. These hypotheses, however, were only verified indirectly, i.e., via the crystallographic orientation relationship between the individual phases. An advanced understanding of the unknown complex local chemical environment in these microstructures and the fundamental mechanism needs to be established before this can be systematically harnessed. Here, we provide an in-depth analysis of the chemical environment across various interphase boundaries in Beta 21 S with additions of B₄C. Using atom probe tomography (APT), we reveal the mechanism of β stabilization following the formation of reinforcement particles, where we focus on TiB, the primary reinforcement phase. A better understanding of these phase transformations in β-Ti matrix composites will enable their applications in critical segments of next-generation aircraft. Their high specific strength is important to enable lighter aeroengines, leading to increased fuel efficiency.

Materials and methods

70 g ingots of the Beta 21 S Ti-alloy (Ti-15Mo-3Nb-3Al-0.2Si, wt%) were manufactured in a vacuum arc remelting furnace. B₄C was added at 0.5, 1.5 and 3 wt%. All samples were heat treated in Ar atmosphere at 1000 °C for 12 h, then furnace-cooled (~ 3 °C/min) to room temperature. The chosen temperature is above the β-transus (~ 815 °C for Beta 21 S¹⁴), and the duration was selected to evaluate the microstructure stability at high-temperature exposure, as discussed in¹⁰. Industrial heat treatments are usually shorter, however, the phenomena discussed here are expected to be equally applicable. In the following, samples are referred to as B₄C_0, B₄C_0.5, B₄C_1.5, and B₄C_3, indicating the corresponding B₄C additions (0, 0.5, 1.5, and 3 wt%). For scanning electron microscopy, samples were polished following standard metallography procedures¹⁵. Backscatter electron imaging (BSE) was performed in a JEOL 7001 F microscope at 20 kV. The area fraction of α-phase was obtained from an area of 500 μm² from BSE images via ImageJ software. APT blanks were cut from the ingots and electropolished via conventional methods¹⁶. Annular milling was performed via plasma focused ion beam (PFIB) to position the interfaces near the tip apex. BSE images of the tips were taken for APT data reconstruction. Atom probe data was collected in a CAMECA Invizo 6000 with a detector efficiency of 62%, at 50 K, and dual beam lasers at 200 kHz and 400 pJ. The detection rate was 5%. 3D data reconstruction was accomplished using the software AP Suite 6.3. 1D chemical profiles were extracted from 20 × 20 × 50–150 nm³ cylinders positioned across interphase boundaries. TiB proximity histograms have lengths between 8 and 20 nm and a step size of 0.2 nm. The composition of each phase was determined based on the isosurface objects that encompass the given phase. For phases not enveloped by an isosurface, at least five cubes of varying sizes were placed within the phase, and the composition within these cubes was regarded as the composition of the selected phase.

Results

Figure 1 presents low and high-magnification BSE images. B₄C_0 in Fig. 1(a, b, c) exhibits a 39 ± 3% area fraction of Widmanstätten α-phase embedded in the β matrix, with coarser α-phase along grain boundaries. The addition of B₄C results in the formation of TiB and TiC at a 4:1 ratio through an in-situ reaction during melting, as shown in Fig. 1(d-l). The mechanisms of TiB and TiC formation during solidification, their volume fractions, and the associated mechanical properties have been discussed in previous work¹¹. For example, the combined volume fractions of TiB and TiC in B₄C_0.5, B₄C_1.5, and B₄C_3 are approximately 5%, 10%, and 20%, respectively. However, the primary phenomenon of interest here is the refinement and decrease of α-phase with increased volume fractions of TiB and TiC, and subsequent β-phase stabilization. The α-phase in B₄C_0.5 and B₄C_1.5 is significantly finer compared to the α-phase in B₄C_0, with a reduction in area fraction of α-phase observed at 29 ± 3% in B₄C_0.5 and 25 ± 4% in B₄C_1.5. There appears to be a threshold in the fraction of reinforcement particles that drastically reduces α-phase to an almost negligible fraction in B₄C_3 (Fig. 1(k, l)). In this condition, fine α-phase seems to precipitate only from a few TiB particles, as indicated by the yellow arrows.

Fig. 1.

BSE images of Beta 21 S without B₄C are shown in (a–c), while Beta 21 S with additions of B₄C are depicted in (d–f) for B₄C_0.5, (g–i) for B₄C_1.5, and (j–l) for B₄C_3. Figures (c, f, i, and l) illustrate a refinement and reduction in α-phase with an increase in the volume fraction of TiB and TiC (dark phases). Primary TiB particles are only observed in B₄C_3 (indicated by red arrows), and fine α-phase around eutectic TiB particles are indicated by yellow arrows.

APT reconstructions of the materials are shown in Figs. 2, 3, 4 and 5. Atom maps are presented in Figs. 2a, 3, 4 and 5a, 3D reconstructions in Figs. 2b, 3, 4 and 5b, the 1D concentration profiles across interphase boundaries in Figs. 2c, 3, 4 and 5c, and proximity histograms in Figs. 3d, 4 and 5d, respectively. In B₄C_0 (Fig. 2), only α- and β-phases are present. Al and O, found as the complex TiO ion here¹⁷ are well-known α-stabilizing elements¹⁸. β-stabilizing elements Mo and Nb, are identified in the β-phase⁷. Si segregation is observed at the α/β interphase boundary¹³.

Fig. 2.

APT of B₄C_0 showing α-phase and β-phase. (a) Atom maps, (b) APT reconstruction with isosurfaces highlighting β-phase (red), (c) 1D concentration plots and deviation (shaded) across the yellow cylinder.

Fig. 3.

APT of B₄C_0.5 showing α-phase, β-phase and TiB. (a) Atom maps, (b) APT reconstruction with isosurfaces highlighting β (red) and TiB (blue), (c) 1D concentration plots and deviation (shaded) across the yellow cylinder, and (d) α-TiB proximity histogram.

Fig. 4.

APT of B₄C_1.5 showing α-phase, β-phase and TiB. (a) Atom maps, (b) APT reconstruction with isosurfaces highlighting β (red) and TiB (blue), (c) 1D concentration plots and deviation (shaded) across the yellow cylinder, and (d) α-TiB proximity histogram.

Fig. 5.

APT of B₄C_3 showing α-phase, β-phase and TiB. (a) Atom maps, (b) APT reconstruction with isosurfaces highlighting α (turquoise) and TiB (blue), (c) 1D concentration plots and deviation (shaded) across the yellow cylinder, and (d) (α + β)-TiB proximity histogram. Refinement of α-phase is observed, which becomes plate-shaped particles.

Figures 3, 4 and 5 show regions containing α, β, and TiB. Figure 6 shows the chemical composition variation for all phases across the samples, it can be seen that there is a slight reduction in Ti in α- and β-phases after additions of B₄C, from a maximum of 88.7 at% in α-phase and 86.5 at% in β-phase in B₄C_0 to 86.8 at% in α-phase and 83.1 at% in β-phase in B₄C_3. Additionally, α-phase becomes enriched in TiO by 1.7 at% and β-phase in Mo by 1.6 at%. Low concentrations of C and B are found in these phases. TiB is slightly off-stoichiometry with 40 at% of B and low concentrations of Mo and Nb. Si and P segregate to the α/TiB interface¹³. In B₄C_1.5 (Fig. 4), the concentration of Ti in α-phase is lower than B₄C_0.5, while Al and TiO are higher. The concentration of TiO is twice as high when compared to B₄C_0. The composition of TiB remains stable in B₄C_0.5 and B₄C_1.5, with slightly higher concentrations of minor elements compared to TiB in B₄C_0.5.

Fig. 6.

Chemical composition (at%) of various phases detected by APT in Figs. 2, 3, 4 and 5.

In B₄C_3 (Fig. 5), a 5.3% volume fraction of thin plates of α-phase is identified. This α-phase fraction, however, is likely lower in the bulk of the material, away from the TiB particles (Fig. 1l). The trend of reduction of Ti in α- and β-phase with the increase in B₄C additions is confirmed. α- and β-phases in B₄C_3 have 2.1 at% and 3.4 at% less Ti than B₄C_0, respectively. This changes the Mo equivalent that determines the stability of β-phase according to the contents of β-stabilizing elements¹⁹. In addition, β-phase becomes enriched in α-stabilizing elements, such as Al and TiO, with double the concentration of the latter (1.6 at%) in comparison to β-phase in B₄C_1.5. On the other hand, α-phase shows the highest enrichment in Mo or roughly three times more than in B₄C_0. Notably, TiB also has a distinct chemical profile, with only 10 at% B and much higher concentrations of Ti, Al, Mo, and TiO than in B₄C_0.5 and B₄C_1.5.

Discussion

The role of TiO deserves comment, as all samples were oxidized during preparation. Despite having better oxidation resistance than commercially pure Ti, Beta 21 S is known to heavily oxidize between 600 and 800 °C^20,21. TiO ions correlate to the O content and are detected as an effect of laser pulsing during APT, which causes thermally assisted evaporation of multiple elements as complex ions²². There are several potential sources of oxidation during sample preparation: Raw materials may have not been sufficiently pure, O might be present in the melting chamber, and APT tip preparation via electropolishing may have caused some degree of oxidation. More TiO is observed closer to the apex of the tips (TiO atom maps in Figs. 2, 3, 4 and 5), indicating that PFIB preparation may also contribute to this phenomenon. However, the fact that TiO tends to segregate to α-phase indicates that the most likely source is from the heat treatment, despite being performed in Ar atmosphere⁷. O was adsorbed and enriched into the α-phase during the phase transformation below β-transus.

The formation of TiB makes the materials more susceptible to oxidation, as shown by the increase in TiO in α-phase from B₄C_0 to B₄C_1.5, and in β-phase from B₄C_0 to B₄C_3. Nevertheless, the application of Beta 21 S is generally limited to service temperatures up to 425 °C, minimizing possible oxidation⁷. Maximum TiO retention in α-phase is observed in B₄C_1.5, while α-phase refinement in B₄C_3 results in a decreased TiO concentration. The smaller size and lower volume fraction of α-phase reduce the surface area and diffusion opportunities for O. Additionally, β-phase and TiB are enriched in TiO in B₄C_3, further reducing the O availability for α-phase. While the maximum solubility of O in α is as high as 34 at%²³, there is no consensus on the amount of O that can be in solution in β-phase, with reports indicating concentrations between 4 and 15 at%^24–26. The proximity histogram in Fig. 5c shows that in the vicinity of TiB, the TiO concentration is as high as 30 at%. This complex chemical profile near TiB has a 3-nm area low in Ti, and an increase in Al and Mo at the interface (Fig. 5c). This region contains alternating enrichment and depletion zones, with layers of α- and β-phase not delineated by the isosurfaces used here. The enrichment of Al and TiO in TiB in B₄C_3 corroborates the observation in Fig. 1l, as the removal of α-stabilizing elements from the matrix makes these particles heterogeneous nucleation sites for fine α. Furthermore, O does not seem to negatively affect the mechanical properties of B₄C_3, which has been shown to retain ductility with a maximum compressive strain of 20.5%, and yield strength of 1205 MPa¹⁰. Similar findings were observed in the recent pioneering development of strong and ductile Ti–O–Fe alloys via additive manufacturing by Song et al.²⁶, highlighting the importance of understanding the effects of O content in Ti alloys.

The low concentration of B in TiB in B₄C_3 also deserves discussion. The hypereutectic character of C and B in Ti in B₄C_3 explains this phenomenon. As detailed in¹¹, B₄C_0.5 and B₄C_1.5 are hypoeutectic and near-eutectic, respectively, thus all TiB particles are close to their stoichiometric compositions. In B₄C_3, B was consumed during the formation of primary TiB during solidification (indicated by red arrows in Fig. 1j), and consequently, less B was available for the formation of the remaining eutectic TiB. Therefore, the selected TiB particle in B₄C_3 (Fig. 5) is eutectic. These assumptions are based on Ti-B, Ti-C, and Ti-B₄C phase diagrams available in the literature, thus, the influence of the other alloying elements, particularly Mo, was not considered^27–29. However, negligible enrichments in Mo and Al are expected in primary TiB, as they formed first from the liquid, as confirmed by thermal analysis¹¹. In addition, most of the fine α-phase in B₄C_3 seems to form around small TiB (Fig. 1k), thus, coarse TiB do not contain sufficient α-stabilizing elements to induce heterogeneous nucleation of α-phase. On the other hand, eutectic TiB particles become nucleation sites for α^30,31 as α-stabilizing substitute Ti and B atoms in the orthorhombic crystal structure of TiB. A similar mechanism has been observed in the Ti-5Mo-5 V-8Cr-3Al alloy³² and in a metal matrix composite steel with TiB₂33.

The Mo equivalent of Beta 21 S is ~ 13, thus, 100% of metastable β is retained upon quenching from temperatures above the β transus⁷. Stable β-Ti alloys require a Mo equivalent of > 50, as exemplified by the heavily stabilized Alloy C (Ti-35 V-15Cr, wt%) with a Mo equivalent of 47.5, which was still not stable⁷. B₄C_3 is not categorized as β stable due to the presence of fine α-phase, however, the degree of β stabilization is close to Ti-alloys with much higher Mo equivalents. Despite a 26% increase in Mo in β-phase from B₄C_0 to B₄C_3 and corresponding increases of the β-phase stability, a concomitant 32% increase in Al and a four-fold increase in TiO reduce β-phase stability. Thus, a different mechanism must be active in addition to the Mo equivalent in determining the stability of β-phase in this condition.

Based on our high-volume APT data, we propose that the enrichment of Mo in α-phase must play a major role in its refinement and β-phase stabilization. Mo, a slow diffusing element, inhibits the growth of α-phase by blocking the influx of Al from β-phase. This is caused by the formation of TiB and TiC, which reduce the availability of Ti and Al in the matrix, as observed by the 20 times higher concentration of Al in the TiB in B₄C_3 (4 at%) than in B₄C_1.5 (0.2 at%). Interestingly, Nb concentration remains unchanged in all phases, highlighting that the major effect on the increase in β stability must be due to Al and Mo diffusion phenomena.

Lastly, while APT provides exceptional resolution at the sub-nanometre scale³⁴, it’s important to consider its inherent limitations. The technical challenges associated with APT are numerous and well-documented in the literature^16,35,36, but a comprehensive discussion of these issues is beyond this study’s scope. However, it’s worth noting some key artifacts: Local magnification effects can artificially expand segregation zones, particularly those aligned with the tip apex. Non-uniform evaporation due to density variations, such as those caused by borides can lead to inaccurate ion reconstruction at phase boundaries. Currently, fully eliminating or correcting for these artifacts is not possible. Nevertheless, understanding these limitations is crucial for accurate interpretation of APT data and readers should keep these factors in mind when analysing and drawing conclusions from the current APT results. We recommend future research to explore additional high resolution techniques to complement these findings.

Conclusions

In summary, furnace cooled β-Ti alloy Beta 21 S with 39% area fraction of α-Ti is refined by adding B₄C followed by the formation of reinforcement particles. The composite material with additions of 0.5 (B₄C_0.5), 1.5 (B₄C_1.5) and 3 wt% (B₄C_3) of B₄C reduce α-Ti to 29%, 25% and 5% fractions, respectively. The β-phase stabilization mechanisms associated with the phase transformations are elucidated via APT. Segregation of Si and P are observed in α/TiB boundaries. Higher TiO formation in α- and β-phase with additions of B₄C may be beneficial to the mechanical strength-ductility relationship, as shown in previous studies^10,26. In B₄C_3, eutectic TiB particles with low concentrations of B become enriched in Al, and provide nucleation sites for fine α-phase, driving its refinement/decrease. Coarse primary TiB do not seem to promote α-phase formation. Refined α-phase becomes enriched in Mo, inhibiting Al influx, and subsequent α coarsening, promoting a near stable β-phase. Control of α-phase formation via the B₄C additions can lead to stronger β-Ti alloys and composites for applications in next-generation aeroengines.

Author contributions

V. V. R: Conceptualization, Data curation, Formal analysis, Investigation, Writing - original draft, Writing - review & editing, Project administration. R. J. C.: Resources, Writing - review & editing, Funding acquisition. S. P. : Supervision, Writing - review & editing, Funding acquisition.

Data availability

The data that support the findings of this study are available from the corresponding author upon reasonable request.

Declarations

Competing interests

The authors declare no competing interests.