Isothermal Section of the Ti-Si-B System at 1250 °C in the Ti-TiSi2-TiB2 Region

aInstituto de Ciência e Tecnologia, Universidade Federal de Alfenas – UNIFAL-MG, Rod. José Aurélio Vilela, 11.999, Cidade Universitária, CEP 37715-400, Poços de Caldas, MG, Brasil bEscola de Engenharia de Lorena – EEL, Universidade de São Paulo – USP, Est. Municipal do Campinho, s/n, CEP 12602-810, Lorena, SP, Brasil cUniversidade Federal de Itajubá – UNIFEI, Av. BPS, 1303, Pinheirinho, CEP 37500-903, Itajubá, MG, Brasil dNúcleo de Pesquisa, Centro Universitário de Volta Redonda – UniFoa, Campus Três Poços, Av. Paulo Erlei Alves Abrantes, 1325, Três Poços, CEP 27240-560, Volta Redonda, RJ, Brasil


Introduction
Initial information on phase equilibria in the Ti-Si-B system was provided by Meschter 1 and Maex et al. 2 from thermodynamic calculations, using data from the binary phases of the Ti-B and Ti-Si systems. Figure 1a,b shows the isothermal sections calculated by these authors.
From experimental results, Ramos et al. 3,4 showed the existence of a stable ternary phase with Ti 6 Si 2 B stoichiometry in this system. Additionally, the phase relations involving this phase at 1250 °C and its formation reaction from the liquid state were also showed. Yang et al. 5 carried out a thermodynamic calculation of the Ti-Si-B system, proposing a liquidus projection and an isothermal section at 1250 °C from selected experiments and the results from Ramos et al. 3,4 . The experiments carried out by Yang et al. 5 involved the production by arc-melting, annealing at 1250 °C for 100h and microstructural characterization of four samples of composition Ti-21Si-10B, Ti-19Si-6B, Ti-16.5Si-3.5B and Ti-14Si-1B. All these compositions are located in the Ti-rich region of the Ti-Si-B system, specifically in the βTi+Ti 5 Si 3 +Ti 6 Si 2 B three-phase field of this system at 1250 °C.
In order to extend the phase stability knowledge of this system, this work evaluated the phase relations in the Ti-TiSi 2 -TiB 2 region of the Ti-Si-B system at 1250 °C and the results were compared with those from Yang et al. 5 .

Experimental Procedure
Several ingots (~ 8 g) of different compositions were produced from high-purity commercially available materials: titanium (99.8 % min.), silicon (99.999 % min.) and boron (99.5 % min.). Table 1 shows the nominal compositions of the alloys used in this work. The samples were prepared via arc-melting with a non-consumable tungsten electrode on a water-cooled hearth under high purity argon atmosphere gettered by titanium. The ingots were melted five times in an effort to produce homogeneous alloys. Then, parts of the as-cast ingots were heat-treated at 1250 °C for up to 240 h under argon in quartz tubes. In order to confirm the stability of the Ti 3 B 4 and support some proposed phase relations which could not be experimentally proved, a Ti 42.89 B 57.11 (Ti 3 B 4 stoichiometry) alloy was arc-melted and heat-treated at 1250 °C for 30 days and at 1500 °C for 55 h.
All the samples were characterized by scanning electron microscopy (SEM) and selected alloys were evaluated via microanalysis wavelength dispersive spectrometry (WDS).
The SEM/BSE images were acquired in a LEO 1450VP SEM equipment. The WDS measurements were carried out in a model 440 Stereoscan/Leica Microscope at 15 kV/10 nA using PET and LSM 200 crystals to quantify the contents of Ti/Si and B, respectively. Pure elements were used as standards, and the results were obtained using a φρz correction procedure. At least three measurements for each phase were carried out. silicides present in Table 1 The microstructural characteristics of the alloys which allowed the establishment of the isothermal section shown in Figure 2 are presented below.
Alloy #40 (Ti 65 Si 32.5 B 2.5 ) presented Ti ss , Ti 5 Si 3 and Ti 6 Si 2 B phases in the as-cast as well as in the heat-treated samples. Figure 3a shows a SEM/BSE micrograph of this alloy in the heat-treated condition where Ti ss and Ti 6 Si 2 B are minor phases in a Ti 5 Si 3 matrix. The WDS results shown in Table 2 indicate a low solubility of Si and B in Ti ss as well as of B in Ti 5 Si 3 . The cracks present in the Ti 5 Si 3 phase (Figure 3a) are formed during cooling in the solid state, associated with the high anisotropy of thermal expansion of this phase 7 .
Alloy # 55 (Ti 70 Si 7.5 B 22.5 ) presented Ti ss , TiB and Ti 6 Si 2 B phases in the as-cast as well as in the heat-treated samples. Figure 3b shows a SEM/BSE micrograph of the heattreated alloy where all the phases are present in significant amounts. The WDS data ( Table 2) shows once again the low solubility of Si and B in Ti ss as well as of Si in TiB. In addition, because the compositions of the Ti ss and Ti 6 Si 2 B phases in the alloys #40 and #55 are approximately the same Figure 2 shows the partial isothermal section at 1250 °C of the Ti-Si-B system determined in this work. Data for the binaries Ti-Si and Ti-B systems were taken from 6 and includes the following phases: Ti ss , Ti 5 Si 3 , Ti 5 Si 4 , TiSi, TiSi 2 , TiB, Ti 3 B 4 and TiB 2 . There exists only one ternary phase, Ti 6 Si 2 B, which was discovered by Ramos et al. 4 and is isotypic with Ni 6 Si 2 B.

Results and Discussion
The WDS measurements data can be roughly evaluated via the measured values of B and Si in the borides and   of TiB 2 and Ti ss during heat-treatment. Figure 3d shows a SEM/BSE micrograph of the heat-treated alloy. The WDS analysis shows very low solubility of B in the Ti 5 Si 3 phase as well as of Si in the TiB phase. As can be concluded from the previous results, at 1250 °C the ternary Ti 6 Si 2 B phase equilibrates with Ti ss , TiB, and Ti 5 Si 3 phases through narrow two-phase fields.
Alloy # 46 (Ti 55 Si 40 B 5 ) presented Ti 5 Si 3 , Ti 5 Si 4 , TiSi, TiSi 2 and TiB 2 in the as-cast and Ti 5 Si 3 Ti 5 Si 4 e TiB 2 in the heat-treated sample, indicating the dissolution of TiSi and TiSi 2 during the heat-treatment. Figure 4a shows a SEM/BSE micrograph of this alloy after heat-treatment. The WDS results show the low solubility of B in both Ti 5 Si 3 and Ti 5 Si 4 as well as of Si in TiB 2 .
Alloy #77 (Ti 50 Si 45 B 5 ) presented Ti 5 Si 4 , TiSi, TiSi 2 and TiB 2 in the as-cast and Ti 5 Si 4 , TiSi and TiB 2 in the heattreated sample, showing the dissolution of the TiSi 2 phase during the heat-treatment. Figure 4b shows a SEM/BSE micrograph of this alloy after heat-treatment where all the phases are present in significant amount. The WDS analysis from the Ti 5 Si 4 and TiSi phases has shown a very low B solubility in these phases as well as of Si in TiB 2 . (Table 2), the Ti ss +Ti 6 Si 2 B two-phase region should be quite narrow at 1250 °C.
The calculated results of Yang et al. 5 have shown some solubility of Si in the Ti ss at 1250 °C. However, considering the low solubility of Si and B in Ti ss , the low solubility of B and Si in the silicides and borides respectively, as well as the difficult to accurately determine the concentration of boron, we have not included any solubility data in Figure 2.
Alloy #14 (Ti 66.7 Si 22.2 B 11.1 ) presents the nominal composition of the stoichiometric Ti 6 Si 2 B phase. Five phases were observed in the as-cast sample: TiB 2 , TiB, Ti 6 Si 2 B, Ti 5 Si 3 and Ti ss . After heat-treatment at 1250 °C, the microstructure formed was essentially constituted of the Ti 6 Si 2 B, with minor amount of Ti ss and TiB, as shown in Figure 3c. These results suggested that the Ti 6 Si 2 B should be near stoichiometry at 1250 °C. Contrasting to the behavior observed for the samples with high volume fraction of Ti 5 Si 3 phase, no crack was noticed in the Ti 6 Si 2 B phase, likely due to its lower thermal expansion anisotropy 7 .
Alloy # 69 (Ti 60 Si 25 B 15 ) presented Ti ss , Ti 5 Si 3 , TiB, TiB 2 and Ti 6 Si 2 B in the as-cast and TiB, Ti 5 Si 3 and Ti 6 Si 2 B in the heat-treated sample, showing the complete dissolution   of as-cast samples with composition in the 50 to 67 at.% B range. Both XRD and SEM results showed an increase in the quantity of Ti 3 B 4 phase due to annealing, indicating the stability of the Ti 3 B 4 phase. Based on the results of Spear et al. 8 and our own results presented above, it is assumed that the Ti 3 B 4 is also stable at 1250 °C. Thus, the existence of the TiB+Ti 5 Si 3 +Ti 3 B 4 and Ti 3 B 4 +Ti 5 Si 3 +TiB 2 three-phase regions is proposed to be consistent with the results of alloys #69 and #46 shown previously. Furthermore, the proposed isothermal section is in agreement with those calculated by Yang et al. 5 .

Conclusions
Phase equilibria in the Ti-TiSi 2 -TiB 2 region of the Ti-Si-B system have been experimentally evaluated at 1250 °C from heat-treated alloys prepared via arc melting. In general, it was found to be difficult to equilibrate alloys located in the TiB-Ti 5 Si 3 -TiB 2 region. The ternary phase (Ti 6 Si 2 B) previously reported in the literature has been confirmed in this study. Very low solubility of Si in the borides as well as of B in the silicides has been noticed, thus, all the two-phase fields should be very narrow. The following three-phase equilibria have been observed: Ti ss +TiB+Ti 6 Si 2 B; Ti ss +Ti 6 Si 2 B+Ti 5 Si 3 ; Ti 5 Si 3 +Ti 6 Si 2 B+TiB; Ti 5 Si 3 +TiB+Ti 3 B 4 ; Ti 5 Si 3 +Ti 3 B 4 +TiB 2 ; Ti 5 Si 3 +TiB 2 +Ti 5 Si 4 ; Ti 5 Si 4 +TiB 2 +TiSi; TiSi+TiB 2 +TiSi 2 .
Alloy # 79 (Ti 40 Si 50 B 10 ) revealed the presence of Ti 5 Si 4 , TiSi, TiSi 2 and TiB 2 in the as-cast and TiSi, TiSi 2 and TiB 2 after heat-treatment, indicating the dissolution of the Ti 5 Si 4 phase. Figure 4c shows a SEM/BSE micrograph of the heattreated sample where all the phases are present in significant amounts. The WDS data shown in Table 2 indicates the low solubility of B in both TiSi and TiSi 2 phases.
Alloys # 91, 92 and 93 presented TiB, Ti 5 Si 3 and TiB 2 in both as-cast and heat-treated microstructures. These results indicate the difficulty to form the Ti 3 B 4 from the as-cast microstructures and therefore to equilibrate these alloys in the phase fields involving this phase. Another possibility was that the Ti 3 B 4 phase is not stable, in spite of its indication in the currently accepted Ti-B phase diagram. In order to check the stability of the Ti 3 B 4 phase in the Ti-B system, an alloy with composition Ti 42.89 B 57.11 (Ti 3 B 4 stoichiometry) was arc-melted and heat-treated at 1250 °C for 30 days. The as-cast microstructure was formed by the phases TiB 2 , Ti 3 B 4 , TiB and Ti ss , the volume fraction of Ti 3 B 4 being near 13%. The heat-treatment did not change the phases nor modify significantly amounts in the as-cast microstructure. A second as-cast sample of same composition was heattreated at 1500 °C for 55 h and now a significant increase of Ti 3 B 4 volume fraction (13% => 43%) was observed, even though thermodynamic equilibrium conditions were not reached. Spear et al. 8 have evaluated the stability of the Ti 3 B 4 from annealing (1690 °C to 2070 °C for ½ to 2h)