Formation of Phases in Reactively Sintered TiAl3 Alloy.

This work highlights new results on the synthesis of the TiAl3 intermetallic phase using self-propagating high-temperature synthesis. This method is considered a promising sintering route for intermetallic compounds. It was found that the reactions proceed in two stages. Below the melting point of aluminum, the Ti2Al5 phase forms at 450 °C after long annealing times by a direct solid-state reaction between the aluminum and titanium, and is converted consequently to TiAl3. This is a completely new finding; until now, many authors have believed in the preferential formation of the TiAl3 phase. The second stage, the self-propagating strongly exothermic reaction, proceeds above the melting point of aluminum. It leads to the formation of the TiAl3 phase accompanied by Ti2Al5 and Ti3Al phases. The reaction mechanism was shown in the form of chemical equations, which were supported by calculating Gibbs energy. Reaction temperatures (Tonset, Tmaximum, and Toffset) were determined after induction heating thanks to recording by an optical pyrometer. This finding provides completely new opportunities for the determination of activation energy at heating rates, in which common calorimeters are not able to detect a response or even measure. Now, the whole procedure will become accessible.


Introduction
Titanium aluminides belong to a group of modern materials that could replace nickel-based alloys in high-temperature applications such as in the aerospace industry. They offer low density, good high-temperature creep strength, stiffness, high melting points, and oxidation resistance [1,2]. However, their wider applications are hampered by low-temperature ductility [1]. They have been applied practically as exhaust valves, turbine superchargers, and low-temperature turbine blades in a GEnx TM engine, which is the first commercial aircraft engine using Ti-Al alloy [1,3].
Equilibrium phases occurring in Ti-Al systems are Ti 3 Al-compounded with hexagonal close-packed superlattice structures and TiAl and TiAl 3 intermetallic compounds with tetragonal structures [3,4]. Further, TiAl 2 and Ti 2 Al 5 phases are found on the aluminum-rich side of the diagram [3]. As can be seen, titanium-rich phases (Ti 3 Al and TiAl) exist over a large range of compositions, meanwhile the TiAl 3 phase forms as a line compound only at [3]. The occurrence of phases depends on chemical temperature of the melting point of aluminum. This means that the formation of phases took place on the interface between liquid aluminum and solid titanium. Only the heating rate of 19 • C/min caused the reaction to start at 662 • C, close to the melting point of aluminum (660 • C) and peritectic temperatures (665 • C) in diagram Ti-Al [17]. Therefore, this would suggest that TiAl 3 phases formed after the peritectic reaction resulted in the presence of a liquid phase during heating at this heating rate. This claim was confirmed by studying the microstructure and XRD, as will be shown below.
XRD, as will be shown below.
Further, it can be seen that the exothermic peak is less noticeable as the heating rate increased (Figure 1). Phases formed during a very short time at heating rates 59-102 °C/min because the observed peaks were narrower. Various heating rates also allowed for the calculation of the activation energy of phases formation according to the Kissinger equation Equation (1) [18]. Its value was determined to be only 17 kJ/mol meaning there was a fast formation of phases and reaction between titanium and aluminum. This value was probably the sum of all reactions occurring during the observed exothermic peak. A similar result was observed in work [11] where the ignition energy was calculated.
C RT where T (K) is the temperature of the exothermic peak, β (K/min) is the heating rate, E (kJ/mol) is the activation energy, R (kJ/(K⋅mol)) is the universal gas constant, and C is the constant.  Further, it can be seen that the exothermic peak is less noticeable as the heating rate increased ( Figure 1). Phases formed during a very short time at heating rates 59-102 • C/min because the observed peaks were narrower. Various heating rates also allowed for the calculation of the activation energy of phases formation according to the Kissinger equation Equation (1) [18]. Its value was determined to be only 17 kJ/mol meaning there was a fast formation of phases and reaction between titanium and aluminum. This value was probably the sum of all reactions occurring during the observed exothermic peak. A similar result was observed in work [11] where the ignition energy was calculated. ln Molecules 2020, 25, 1912 4 of 13 where T (K) is the temperature of the exothermic peak, β (K/min) is the heating rate, E (kJ/mol) is the activation energy, R (kJ/(K·mol)) is the universal gas constant, and C is the constant.

Phase Composition and Microstructure of Alloys Heated at Various Heating Rates
The lowest heating rate caused the microstructure to consist of only unreacted particles of titanium surrounded by TiAl 3 phases and pores (Figure 2a). The XRD analysis also revealed the presence of the Ti 2 Al 5 phase ( Figure 3). All aluminum reacted with the titanium, leading to the formation of an aluminum-rich phase. Thus, no aluminum was found ( Figure 3). Typical positions belonging to peaks are identified for all phases. Major peaks, including minor peaks, are labeled by numbers. All phases are always identified according to the major peaks, and their presence must be confirmed by minor ones. For this reason, all peaks are marked. The changes in phase composition are always labeled by circles or arrows. A higher heating rate (59 • C/min) enabled the formation of the Ti 3 Al phase around unreacted particles of titanium ( Figure 2b). This phase had to form after the reaction between titanium and the already present aluminum-rich phase, probably with TiAl 3 or Ti 2 Al 5 phases, or thanks to high temperatures, as will be shown in part 2.3. The TiAl 3 phase formed as the main phase after annealing at 59 • C/min ( Figure 2b). The Ti 2 Al 5 phase was also found, and was detected by the XRD measurements ( Figure 3). As the heating rate increased, the area of the Ti 3 Al phase was larger (Figure 2c,d). Phase composition remained identical ( Figure 3). Results from the EDS analysis are shown in Table 2, and they are added for confirmation of the presence of the Ti 3 Al phase.

Phase Composition and Microstructure of Alloys Heated at Various Heating Rates
The lowest heating rate caused the microstructure to consist of only unreacted particles of titanium surrounded by TiAl3 phases and pores (Figure 2a). The XRD analysis also revealed the presence of the Ti2Al5 phase ( Figure 3). All aluminum reacted with the titanium, leading to the formation of an aluminum-rich phase. Thus, no aluminum was found ( Figure 3). Typical positions belonging to peaks are identified for all phases. Major peaks, including minor peaks, are labeled by numbers. All phases are always identified according to the major peaks, and their presence must be confirmed by minor ones. For this reason, all peaks are marked. The changes in phase composition are always labeled by circles or arrows. A higher heating rate (59 °C/min) enabled the formation of the Ti3Al phase around unreacted particles of titanium ( Figure 2b). This phase had to form after the reaction between titanium and the already present aluminum-rich phase, probably with TiAl3 or Ti2Al5 phases, or thanks to high temperatures, as will be shown in part 2.3. The TiAl3 phase formed as the main phase after annealing at 59 °C/min ( Figure 2b). The Ti2Al5 phase was also found, and was detected by the XRD measurements ( Figure 3). As the heating rate increased, the area of the Ti3Al phase was larger (Figures 2c,d). Phase composition remained identical ( Figure 3). Results from the EDS analysis are shown in Table 2, and they are added for confirmation of the presence of the Ti3Al phase.  is much more exothermic at a low-heating rate. Significant porosity was also found, and it is a typical phenomenon for reactive sintering. All initiation temperatures were higher than the melting point of aluminum, but liquid aluminum did not fill pores sufficiently, probably due to a limited time of existence of the melt or its low fluidity. Another cause of high porosity from the obtained microstructure is connected with the different diffusivities of titanium and aluminum supporting the Kirkendall effect.

Formation of Phases in Samples Annealed at Temperatures Below Melting Point of Al
Annealing at 450 °C for 8 and 24 h did not cause the formation of any intermetallic phase ( Figures  4a,b). Only unreacted particles, aluminum, and titanium were found, and no presence of intermetallic phases was detected by the XRD analysis ( Figure 5). The Ti2Al5 phase formed after 48 h, and this one could be observed at the interface of aluminum and titanium (Figures 4c and 5). Ti2Al5 formed due to the reaction between solid aluminum and solid titanium without the need of the melt as the transport medium. Thus, the whole reaction proceeded in solid state, and the formation had to be controlled by a mutual solid-state diffusion of aluminum and titanium. This situation is described by Equation (2). The values of Gibbs energy were calculated based on works [4,15].  As can be seen, the TiAl 3 phase was found during all applied heating rates; therefore, it can be assumed that the observed exothermic peak is associated mainly with its formation. The formation of this phase is, however, accompanied by the formation of other phases, namely, the Ti 2 Al 5 and Ti 3 Al phases. The exothermic peak is thus affected by the released heat associated with the formation of these phases. It is also evident that the highest exothermic response was observed during heating at 19 • C/min when the Ti 3 Al phase was not detected (Figure 2a). The formation of aluminum rich-phases is much more exothermic at a low-heating rate. Significant porosity was also found, and it is a typical phenomenon for reactive sintering. All initiation temperatures were higher than the melting point of aluminum, but liquid aluminum did not fill pores sufficiently, probably due to a limited time of existence of the melt or its low fluidity. Another cause of high porosity from the obtained microstructure is connected with the different diffusivities of titanium and aluminum supporting the Kirkendall effect.

Formation of Phases in Samples Annealed at Temperatures Below Melting Point of Al
Annealing at 450 • C for 8 and 24 h did not cause the formation of any intermetallic phase (Figure 4a,b). Only unreacted particles, aluminum, and titanium were found, and no presence of intermetallic phases was detected by the XRD analysis ( Figure 5). The Ti 2 Al 5 phase formed after 48 h, and this one could be observed at the interface of aluminum and titanium (Figures 4c and 5). Ti 2 Al 5 formed due to the reaction between solid aluminum and solid titanium without the need of the melt as the transport medium. Thus, the whole reaction proceeded in solid state, and the formation had to be controlled by a mutual solid-state diffusion of aluminum and titanium. This situation is described by Equation (2). The values of Gibbs energy were calculated based on works [4,15].   The Ti2Al5 phase was still found at the interface of aluminum and titanium after annealing at 500 °C for 8-48 h (Figures 6a-c), and its area was larger. The reaction was still in solid state, and other aluminum-rich phases formed. This phase was TiAl3, which was detected after annealing for 24 and 48 h, and only an XRD analysis proved its formation (Figure 7). It can be assumed that the TiAl3 phase  The Ti 2 Al 5 phase was still found at the interface of aluminum and titanium after annealing at 500 • C for 8-48 h (Figure 6a-c), and its area was larger. The reaction was still in solid state, and other aluminum-rich phases formed. This phase was TiAl 3 , which was detected after annealing for 24 and 48 h, and only an XRD analysis proved its formation (Figure 7). It can be assumed that the TiAl 3 phase formed thanks to the reaction between solid aluminum and an already present Ti 2 Al 5 phase, as can be seen in Equation (3). The Ti2Al5 phase was still found at the interface of aluminum and titanium after annealing at 500 °C for 8-48 h (Figures 6a-c), and its area was larger. The reaction was still in solid state, and other aluminum-rich phases formed. This phase was TiAl3, which was detected after annealing for 24 and 48 h, and only an XRD analysis proved its formation (Figure 7). It can be assumed that the TiAl3 phase formed thanks to the reaction between solid aluminum and an already present Ti2Al5 phase, as can be seen in Equation (3).   The TiAl3 phase replaced Ti2Al5 at the interface of aluminum and titanium after annealing at 600 °C for 8 h (Figure 8a). Aluminum was not detectable by XRD after annealing for 24 and 48 h ( Figure  9), and the microstructure was much more reacted (Figures 8b, c). The structure consisted of the TiAl3 phase and the rest of the unreacted aluminum (Figure 8b). The chemical composition of this area is shown in Table 3. Residual unreacted aluminum was not detected by the XRD, probably because of its low amount. Only the TiAl3 phase was found in the structure annealed for 48 h (Figure 8c). XRD diffraction shows that the titanium and Ti2Al5 phase was still present locally ( Figure 9). The TiAl 3 phase replaced Ti 2 Al 5 at the interface of aluminum and titanium after annealing at 600 • C for 8 h (Figure 8a). Aluminum was not detectable by XRD after annealing for 24 and 48 h (Figure 9), and the microstructure was much more reacted (Figure 8b,c). The structure consisted of the TiAl 3 phase and the rest of the unreacted aluminum (Figure 8b). The chemical composition of this area is shown in Table 3. Residual unreacted aluminum was not detected by the XRD, probably because of its low amount. Only the TiAl 3 phase was found in the structure annealed for 48 h (Figure 8c). XRD diffraction shows that the titanium and Ti 2 Al 5 phase was still present locally ( Figure 9).     Table 3. Chemical composition of the Al area found in the microstructure annealed at 600 • C for 24 h (EDS).

Discussion
As can be seen, the formation of titanium aluminides is not so exothermic ( Figure 1). However, induction heating is the most suitable for the sintering of these intermetallics, as was mentioned in work [19]. The used midfrequency induction heating has the capability to rapidly heat the whole sample volume, and it is used for the preheating of powder mixtures before the PWP-SHS reaction due to the low exothermicity of titanium aluminides [19]. However, we found induction heating also to be a suitable source for TE-SHS of TiAl 3 . Work [11] also studied the ignition of TiAl 3 . When we compared the obtained heating curves, our temperature profile did not show a plateau appearing at approximately 645 • C, which corresponds to the formation of the melt. This can be caused by the heat source when work [11] used a laser beam for combustion. Laser power (approximately 60 W) was focused on only one small spot, and this spot was observed. Moreover, their studied tablet of the sample was very thin. Our samples had a height of approximately 20-30 mm. Moreover, the reactions were probably initiated very quickly after the melt occurred, so the plateau was very short, practically invisible. The microstructure of the obtained samples contained unreacted titanium, and it is known that the dissolution of the higher-melting elements (titanium) in molten aluminum is one of the major kinetic aspects for combustion of aluminides [19]. The explanation lies in the particle size of titanium. The used particle size was 44 µm, which resulted in a very fast dissolution in molten aluminum and better interfacial contact between particles. Thus, the kinetic barrier could not arise. According to heating curves, the exothermic reaction took only a few seconds, during which intermetallic phases formed. Therefore, unreacted particles of titanium were observed in microstructures (Figure 2a-d). Unreacted particles of titanium were also observed in work [11], where the reaction was even faster. In order to obtain a single TiAl 3 phase, the necessity of preheating the green bodies was also demonstrated [20]. Activation energy 17 kJ/mol was similar to the value set in work [11], but it was approximately twice smaller than the activation energy calculated for the TiAl 3 phase in work [21], suggesting that our activation energy is probably a combination of the values for the formation of more phases. This statement is supported by the results of the XRD analysis showing the presence of Ti 3 Al, Ti 2 Al 5 , and TiAl 3 phases after induction heating. The activation energy can also be affected by the surface quality of the powders. All reaction temperatures increased with an increased heating rate, which was shown in, e.g., [22]. On the basis of the obtained results, it could also be deduced that increasing the amount of aluminum caused the reaction temperatures to increase [22]. In high-aluminum alloy (TiAl63), much more melted aluminum forms during SHS reactions, which allows the dissolution of more titanium. The liquid phase, hence, acts as the transport medium for reactants, because the diffusion coefficients of metals in the liquid phase are several orders higher than in solid state. Thus, there are more places where the intensive reactions between titanium and aluminum proceed at the same time. The formation of intermetallic compounds is associated with the release of heat, which results in the shift of temperatures.
Many works claim that the first phase, which forms in the whole Ti-Al system at all compositions (Ti 3 Al, TiAl, and TiAl 3 ), is the TiAl 3 phase [8,14,21,[23][24][25][26]. It is attributed to aluminum, which is considered as the dominant diffusing component through TiAl 3 , resulting in its easier formation [21]. However, we revealed the preferential formation of other aluminum-rich phases of Ti 2 Al 5 , which was confirmed in this work. This phase was observed in our previous works in low-aluminum alloy composition, Ti 3 Al [16], and equimolar alloy composition, TiAl [15]. It is in good agreement with the thermodynamic analysis stated in work [15]. The Ti 2 Al 5 phase reacted with aluminum, and TiAl 3 could have arisen because this phase appeared at the interface of Ti 2 Al 5 /Al. Work [15] also explained how Ti 3 Al can form in this alloy composition. Its existence is associated with Gibbs energy, which is negative in the temperature range 400-900 • C. This phase, thus, could form in high-aluminum alloys at high temperatures (above 735 • C), as was observed after induction heating. At lower temperatures, the reactions lead to the formation of high-aluminum phases, which are preferred in this alloy composition; therefore, the Ti 3 Al phase was not found during lower temperature annealing experiments in solid state. According to obtained results, it is obvious that microstructures after induction heating and after annealing at temperatures are different. At the temperature below the melting point of aluminum, the solid-state diffusion is the controlling mechanism; hence, a long period of time (8-48 h) is needed for the formation of detectable amounts of Ti 2 Al 5 and TiAl 3 . During induction heating, the sample is heated quickly; hence, the solid-state reaction stage is skipped and the reactions are initiated at the solid-liquid interface.

Materials and Methods
Samples were prepared from elementary powders (titanium: purity 99.5%, particle size 44 µm; and aluminum: purity 99.62%, particle size 44 µm, STREM CHEMICALS, Newburyport, MA, USA). The powders were mixed to 3 g of the mixture and homogenized manually for 5 min. The powder mixture with the chemical composition TiAl63 (in wt.%) corresponded to the TiAl 3 phase [17]. Subsequently, 3 g of the mixture was compacted into a green body by the universal loading machine, LabTest 5.250SP1-VM (Labortech, Opava, Czech Republic), at pressure 450 MPa for 5 min. The obtained samples had a diameter of 10 mm. Reactive sintering was performed in an induction furnace (Leybold-Heraeus GmbH, Cologne, Germany), where heating was observed and recorded by an optical pyrometer, Optris OPTP20-2M (Optris, Portsmouth, NH, USA), under Ar atmosphere. Various heating rates were studied, and their values were determined according to the slope of the obtained curves: 19, 59, 89, and 102 • C/min. The scheme of the used apparatus is illustrated in Figure 10. This method enabled us to perform a thermal analysis at a much higher heating rate than during a common, differential thermal analysis. In order to find if the phases formed below the melting point of aluminum, green bodies were inserted into silica ampoules and annealed for 8, 24, and 48 h at 450, 500, and 600 • C, respectively. Molecules 2020, 25, x FOR PEER REVIEW 12 of 14

Conclusions
The present results show that the reactions in the TiAl63 (in wt.%) starting powder mixture proceed in two stages. The solid-state process occurs below the melting point of aluminum, leading to the preferential formation of the Ti2Al5 phase already at 450 °C. The TiAl3 phase formed consequently by the reaction between this phase with aluminum. During rapid heating in induction furnace, this solid-state reaction stage is skipped and the self-propagating high-temperature synthesis reaction is initiated at the interface between solid titanium and molten aluminum, i.e., above the melting point of aluminum. At this stage, the TiAl3 phase forms as the main product of sintering, accompanied by the other aluminum-rich phase, Ti2Al5. The Ti3Al phase formed only around unreacted particles of titanium, and its formation was probably enabled thanks to the reaction between titanium and an already present phase, probably the TiAl3 phase, or thanks to the high temperature and shorter time of reaction. Reactive-sintered samples were ground by sandpapers with SiC abrasive particles (P80-P2500, Hermes Schleifmittel GmbH, Hamburg, Germany), and polished by suspension Eposil F (ATM GmbH, Mammelzen, Germany) mixed with hydrogen peroxide in volume ration 1:6. Kroll's reagent (5 mL HNO 3 , 10 mL HF, and 85 mL H 2 O) was used for etching. The microstructure was observed by the scanning electron microscope, TESCAN VEGA 3 LMU (Tescan, Brno, Czech Republic), equipped with the OXFORD Instrument, X-max 20 mm 2 SDD EDS analyzer (Oxford Instruments, High Wycombe, UK) for the identification of the chemical composition of the individual phases. The phase composition was studied by X-ray diffraction using the PANalytical X pert Pro software package with the PDF2 database (PANalytical, Almeo, The Netherlands).

Conclusions
The present results show that the reactions in the TiAl63 (in wt.%) starting powder mixture proceed in two stages. The solid-state process occurs below the melting point of aluminum, leading to the preferential formation of the Ti 2 Al 5 phase already at 450 • C. The TiAl 3 phase formed consequently by the reaction between this phase with aluminum. During rapid heating in induction furnace, this solid-state reaction stage is skipped and the self-propagating high-temperature synthesis reaction is initiated at the interface between solid titanium and molten aluminum, i.e., above the melting point of aluminum. At this stage, the TiAl 3 phase forms as the main product of sintering, accompanied by the other aluminum-rich phase, Ti 2 Al 5 . The Ti 3 Al phase formed only around unreacted particles of titanium, and its formation was probably enabled thanks to the reaction between titanium and an already present phase, probably the TiAl 3 phase, or thanks to the high temperature and shorter time of reaction.
Funding: This work was supported by the grant of Specific university research, grant No A1_FCHT_2020_003.

Conflicts of Interest:
The authors declare no conflict of interest.