E ﬀ ects of Al and Heat Treatment on the Microstructure and Hardness of Ti–Al Synthesized via In Situ Melting using LENS

: Producing alloys via in-situ laser alloying approach is becoming very easy and a topical issue in the ﬁeld of laser additive manufacturing. Several studies have emerged, accompanied by a statement by Gasper et al. (2017), who reported the economic beneﬁts and cost savings that can be related to 3D printing of the additive manufacturing (AM) structure using the said in situ synthesis approach. They studied the laser in situ reactive synthesis of producing titanium aluminide (Ti–Al) from elemental powders using the laser materials deposition (LMD) process. This paper highlights how the functional grade Ti–Al alloys can be produced in a single ﬂight using the 850-R Optomec Laser Engineering Net Shaping (LENS) system. The e ﬀ ects of the Al content in the as-built and heat-treated samples were investigated. The as-built samples were cracked, inhomogeneous, and had visible inter-layer boundaries due to the segregation of the alloying elements and di ﬀ erent heat proﬁles that are typically experienced during 3D printing. These characteristics led to wavy hardness proﬁles. Heat-treated samples were the least hard when compared to the as-built samples and had a homogenized microstructure. Samples with a duplex microstructure were, in particular, the least hard, indicating that they would have better ductility than others. The outcomes of this study show that a structure with di ﬀ erent composition can be fabricated with the LENS system. This would lead to adaptive structures being used, especially in the high temperature applications where a structure would have to withstand low and high temperatures during use.


Introduction
Intermetallic titanium aluminide (Ti-Al) alloys have comparable and superior properties over nickel super-alloys, which make them attractive for industries such as aerospace, automotive, and energy. Their low density, high strength, excellent oxidation resistance [1][2][3][4], and other thermo-mechanical properties [5][6][7][8] endorse them to serve as outstanding materials for structures, or just as high temperature barrier coatings. This family of intermetallic alloys can also be used in tissue engineering [9] over grade five titanium alloy. Depending on the method and conditions of processing, a desirable composition and microstructure can be obtained [8]. It is understood that gamma (γ) and dual-phase alloy (γ + α2) microstructures of Ti-Al are necessary for structural engineering. However, these phases are also said to be crack-sensitive and have poor tensile properties at room temperature. The lack of room temperature ductility makes it difficult to process and engineer [10]. The benefits of resolving this The Optomec LENS, model 850-R, system was used to manufacture the Ti-Al cube samples. This system used IPG fiber laser with a maximum power output of 1000 W. The fiber was connected to the depositing/cladding head of the LENS and it was controlled automatically from a computer work-station that used an Optomec application software control, version 3.1.6 (Optomce, Albuquerque, NM, USA). The set-up is depicted in Figure 1.
Metals 2019, 9, x FOR PEER REVIEW 3 of 13 The Optomec LENS, model 850-R, system was used to manufacture the Ti-Al cube samples. This system used IPG fiber laser with a maximum power output of 1000 W. The fiber was connected to the depositing/cladding head of the LENS and it was controlled automatically from a computer workstation that used an Optomec application software control, version 3.1.6 (Optomce, Albuquerque, NM, USA ). The set-up is depicted in Figure 1. The Optomec control software was used to automatically control and operate the LENS deposition head, laser beam, and powder hoppers. Due to this capability, it is possible that feedstocks can be deposited and printed simultaneously. Most importantly, this capability allows for experiments on FGMs and in-situ laser alloying to be conducted. In this paper, in-situ laser alloying was studied.
The schematic shown in Figure 1 illustrates that the powder hoppers, which are identified as Al and Ti powders, were attached to the frame of the LENS platform. Before in-situ laser alloying experiments were commenced, powder flow measurements on Al were conducted. The results are reported in Table 1 against the set rotational speed (rpm). These flow measurements were obtained when the centre purge gas and carrier gas flow rates were kept constant at 25 and 5 L/min, respectively. Argon was used as both the centre purge and carrier gas. During manufacturing, the Ti powder mass flowrate was kept constant at 4.96 g/min. This allowed for the investigation on the effect of Al content on the resulting Ti-Al microstructure. The process parameters are summarized in Table 2.  The Optomec control software was used to automatically control and operate the LENS deposition head, laser beam, and powder hoppers. Due to this capability, it is possible that feedstocks can be deposited and printed simultaneously. Most importantly, this capability allows for experiments on FGMs and in-situ laser alloying to be conducted. In this paper, in-situ laser alloying was studied.
The schematic shown in Figure 1 illustrates that the powder hoppers, which are identified as Al and Ti powders, were attached to the frame of the LENS platform. Before in-situ laser alloying experiments were commenced, powder flow measurements on Al were conducted. The results are reported in Table 1 against the set rotational speed (rpm). These flow measurements were obtained when the centre purge gas and carrier gas flow rates were kept constant at 25 and 5 L/min, respectively. Argon was used as both the centre purge and carrier gas. During manufacturing, the Ti powder mass flowrate was kept constant at 4.96 g/min. This allowed for the investigation on the effect of Al content on the resulting Ti-Al microstructure. The process parameters are summarized in Table 2.  Post-manufacturing, the samples were cut and prepared for metallographic observation in the as-built and post-heat treatment states. Heat treatment was performed on the samples for stress relief, microstructural homogenization, and phase precipitation. It is well established that the as-built samples have internal residual stress and that elemental segregation during alloying would lead to a heterogeneous microstructure. The phase precipitation was attained in accordance with the binary Ti-Al phase diagram ( Figure 2). Post-manufacturing, the samples were cut and prepared for metallographic observation in the as-built and post-heat treatment states. Heat treatment was performed on the samples for stress relief, microstructural homogenization, and phase precipitation. It is well established that the as-built samples have internal residual stress and that elemental segregation during alloying would lead to a heterogeneous microstructure. The phase precipitation was attained in accordance with the binary Ti-Al phase diagram ( Figure 2). Heat treatment was conducted using a Carbolite tube furnace in an argon-rich environment. The samples were heated to a temperature of about 1200 °C at the ramping rate of 20 °C/min and held for 2 h before furnace cooling. The samples were mounted in a black resin and polished with OP-S suspension to 0.04 micron surface finish using Struers TegrsForce-5 auto/manual polisher (Cleveland, OH, USA). The samples were etched by immersing them into Krolls' reagent for few seconds.
Olympus BX51M (Feasterville-Trevose, PA, USA), which was mounted on SC30 camera, was used for microstructural visualisation, while the Joel, JSM-6010Plus/LAM scanning electron microscope (SEM) (Peabody, MA, USA), equipped with energy dispersive X-ray, was used for microstructural and chemical composition analyses. Micro-hardness was measured as HV0.5 using the Zwick/Roell (ZHVµ) Vickers' micro-hardness tester machine (Zwick Roell AG, Ulm, Germany). During indentation, pin-loading of 500 gf and a dwelling time of 10 s was used. Indents were spaced at 200 µm from each other. Measurements were taken on the left-side, middle, and right-side of the build. The average hardness was used to plot the hardness profiles and to determine the overall hardness value of the samples.

As-Built Samples
Using process parameters reported in Table 2, Ti-Al cube samples with different compositions were produced. The produced cube samples are shown in Figure 3. Heat treatment was conducted using a Carbolite tube furnace in an argon-rich environment. The samples were heated to a temperature of about 1200 • C at the ramping rate of 20 • C/min and held for 2 h before furnace cooling. The samples were mounted in a black resin and polished with OP-S suspension to 0.04 micron surface finish using Struers TegrsForce-5 auto/manual polisher (Cleveland, OH, USA). The samples were etched by immersing them into Krolls' reagent for few seconds.
Olympus BX51M (Feasterville-Trevose, PA, USA), which was mounted on SC30 camera, was used for microstructural visualisation, while the Joel, JSM-6010Plus/LAM scanning electron microscope (SEM) (Peabody, MA, USA), equipped with energy dispersive X-ray, was used for microstructural and chemical composition analyses. Micro-hardness was measured as HV0.5 using the Zwick/Roell (ZHVµ) Vickers' micro-hardness tester machine (Zwick Roell AG, Ulm, Germany). During indentation, pin-loading of 500 gf and a dwelling time of 10 s was used. Indents were spaced at 200 µm from each other. Measurements were taken on the left-side, middle, and right-side of the build. The average hardness was used to plot the hardness profiles and to determine the overall hardness value of the samples.

As-Built Samples
Using process parameters reported in Table 2, Ti-Al cube samples with different compositions were produced. The produced cube samples are shown in Figure 3.  It was shown in Figure 1 earlier that during laser metal deposition, the produced sample/cube would have different heating profiles, which is a result of accumulative heating during laser 3D printing. The heat profile of the as-built cubes will always be such that it is hot at the bottom in comparison to the middle and top parts of the build. This results in the different layers having different textures or a completely heterogeneous microstructure, hence post-heat treatment is a requirement for laser 3D printed parts. In addition, due to excess heating during sample building, the resulting build is always characterized by cracks, seen in Figure 4.  Figure 4 shows, in addition to the resulting cracks and roughness, different layers and interlayer boundary regions in the as-built cube sample. The heterogeneity in the structure of the as-built sample is attributed to the segregation of alloying elements during processing and different heat profiles due to fast cooling rates experienced by every layer during layer-by-layer deposition [4].
Cracking, which is due to the brittleness of the alloy, and residual stress build in the sample is a topical issue in this research field. Most researchers working in this field are currently developing printing strategies or implementing technology capabilities which would unlock the possibility of producing a crack-free Ti-Al build [20]. The addition of a heating stage for substrate pre-heating and heat retention during printing are necessary during the manufacturing of Ti-Al samples. Rittinghaus and co-workers [4] explained that an online monitoring system, which would allow for the observation of melt-pool dynamics and the monitoring of temperature during 3D printing, is a musthave tool if inhomogeneity in the as-built samples is to be resolved and if heat-profiles of every depositing layer are to be controlled and maintained throughout sample manufacturing. Sharma et al. [20] reviewed other possibilities of producing a crack-free Ti-Al sample. It should be noted that the samples that were being studied here were not produced with any such proposed systems in place, but in the future this will be taken into consideration.

Chemical Composition and Microstructure
The microstructures and Al content (at.%) of the as-built Ti-Al samples are reported in Table 3  It was shown in Figure 1 earlier that during laser metal deposition, the produced sample/cube would have different heating profiles, which is a result of accumulative heating during laser 3D printing. The heat profile of the as-built cubes will always be such that it is hot at the bottom in comparison to the middle and top parts of the build. This results in the different layers having different textures or a completely heterogeneous microstructure, hence post-heat treatment is a requirement for laser 3D printed parts. In addition, due to excess heating during sample building, the resulting build is always characterized by cracks, seen in Figure 4. It was shown in Figure 1 earlier that during laser metal deposition, the produced sample/cube would have different heating profiles, which is a result of accumulative heating during laser 3D printing. The heat profile of the as-built cubes will always be such that it is hot at the bottom in comparison to the middle and top parts of the build. This results in the different layers having different textures or a completely heterogeneous microstructure, hence post-heat treatment is a requirement for laser 3D printed parts. In addition, due to excess heating during sample building, the resulting build is always characterized by cracks, seen in Figure 4.  Figure 4 shows, in addition to the resulting cracks and roughness, different layers and interlayer boundary regions in the as-built cube sample. The heterogeneity in the structure of the as-built sample is attributed to the segregation of alloying elements during processing and different heat profiles due to fast cooling rates experienced by every layer during layer-by-layer deposition [4].
Cracking, which is due to the brittleness of the alloy, and residual stress build in the sample is a topical issue in this research field. Most researchers working in this field are currently developing printing strategies or implementing technology capabilities which would unlock the possibility of producing a crack-free Ti-Al build [20]. The addition of a heating stage for substrate pre-heating and heat retention during printing are necessary during the manufacturing of Ti-Al samples. Rittinghaus and co-workers [4] explained that an online monitoring system, which would allow for the observation of melt-pool dynamics and the monitoring of temperature during 3D printing, is a musthave tool if inhomogeneity in the as-built samples is to be resolved and if heat-profiles of every depositing layer are to be controlled and maintained throughout sample manufacturing. Sharma et al. [20] reviewed other possibilities of producing a crack-free Ti-Al sample. It should be noted that the samples that were being studied here were not produced with any such proposed systems in place, but in the future this will be taken into consideration.

Chemical Composition and Microstructure
The microstructures and Al content (at.%) of the as-built Ti-Al samples are reported in Table 3 Figure 4 shows, in addition to the resulting cracks and roughness, different layers and inter-layer boundary regions in the as-built cube sample. The heterogeneity in the structure of the as-built sample is attributed to the segregation of alloying elements during processing and different heat profiles due to fast cooling rates experienced by every layer during layer-by-layer deposition [4].
Cracking, which is due to the brittleness of the alloy, and residual stress build in the sample is a topical issue in this research field. Most researchers working in this field are currently developing printing strategies or implementing technology capabilities which would unlock the possibility of producing a crack-free Ti-Al build [20]. The addition of a heating stage for substrate pre-heating and heat retention during printing are necessary during the manufacturing of Ti-Al samples. Rittinghaus and co-workers [4] explained that an online monitoring system, which would allow for the observation of melt-pool dynamics and the monitoring of temperature during 3D printing, is a must-have tool if inhomogeneity in the as-built samples is to be resolved and if heat-profiles of every depositing layer are to be controlled and maintained throughout sample manufacturing. Sharma et al. [20] reviewed other possibilities of producing a crack-free Ti-Al sample. It should be noted that the samples that were being studied here were not produced with any such proposed systems in place, but in the future this will be taken into consideration.

Chemical Composition and Microstructure
The microstructures and Al content (at.%) of the as-built Ti-Al samples are reported in Table 3 and Figure 5, respectively. Al content in the as-built sample supposedly had to increase linearly with the increase in set rpm, but the EDS measurements proved otherwise. The Al content in the as-built samples varied considerably, as reported in Table 3.  Table 3 reports that the highest value of Al is for samples E and C, followed by samples A and D, and the lowest for samples B and F. Rittinghaus et al. [4] reported an increase in the Al content in the as-built Ti-Al alloys. They credited the LMD process for this observation since, generally, the electron beam melting process achieved an Al loss in the as-built samples. The inconsistency in the Al content, as reported in Table 3, suggests that Al content could have vaporized during the build, but this cannot be confirmed because the online monitoring system was not used to monitor the temperature profiles of the samples during printing. Moreover, since the same stand-off distance was used during deposition, the effects of particle bow shock, nozzle fouling, or mass loss in general can be ruled out completely [32]. The probable cause of this observation could be in the particle size distribution, but, since the image of the plume was not captured during manufacturing for processing and interpretation, this would be an unjustified conclusion. This occurrence, however, requires an explanation and should be looked into in the future. For now, the only plausible explanation for this observation is in the dissolution kinetics (reaction mechanism) of Al into the Ti matrix [1], or the inconsistency of the system to feed a similar increment of Al for 1.2, 1.6, and 2.0 set rpm [30], as shown in Figure 5. This increment was, on average, 0.08 g/min for 1.0, 1.4, and 1.8 set rpm values.   Table 3 reports that the highest value of Al is for samples E and C, followed by samples A and D, and the lowest for samples B and F. Rittinghaus et al. [4] reported an increase in the Al content in the as-built Ti-Al alloys. They credited the LMD process for this observation since, generally, the electron beam melting process achieved an Al loss in the as-built samples. The inconsistency in the Al content, as reported in Table 3, suggests that Al content could have vaporized during the build, but this cannot be confirmed because the online monitoring system was not used to monitor the temperature profiles of the samples during printing. Moreover, since the same stand-off distance was used during deposition, the effects of particle bow shock, nozzle fouling, or mass loss in general can be ruled out completely [32]. The probable cause of this observation could be in the particle size distribution, but, since the image of the plume was not captured during manufacturing for processing and interpretation, this would be an unjustified conclusion. This occurrence, however, requires an explanation and should be looked into in the future. For now, the only plausible explanation for this observation is in the dissolution kinetics (reaction mechanism) of Al into the Ti matrix [1], or the inconsistency of the system to feed a similar increment of Al for 1.2, 1.6, and 2.0 set rpm [30], as shown in Figure 5. This increment was, on average, 0.08 g/min for 1.0, 1.4, and 1.8 set rpm values. Microstructures of the as-built samples are reported in Figure 6. Microstructures of the as-built samples are reported in Figure 6. The microstructures of the as-built samples, as reported in Figure 6, look to be heterogeneous and had visible bands/layers. Sample E shows partially melted powder particles. The samples are also characterized by fine microstructure and agglomerated lamellar or lamellar laths (samples A, B, and C), lamellae grains (samples D and F), and formed grains (sample E). Liu et al. [41] observed coarsening in the lamellar laths, while here the laths grew small and refined into grains with the increase in Al. The features in microstructures are resolved in the SEM images that are reported in Figure 7. Indeed, the lamellar laths size decreased with an increase in Al content (see samples A and C). The variation in microstructures with Al content was anticipated. The microstructure of sample E has surface porosity and grains of different size (20-60 µm) and colour. The α2-grains (white) seem to be generally growing around the γ-grains (black). It is not clear, at this magnification, if there are The microstructures of the as-built samples, as reported in Figure 6, look to be heterogeneous and had visible bands/layers. Sample E shows partially melted powder particles. The samples are also characterized by fine microstructure and agglomerated lamellar or lamellar laths (samples A, B, and C), lamellae grains (samples D and F), and formed grains (sample E). Liu et al. [41] observed coarsening in the lamellar laths, while here the laths grew small and refined into grains with the increase in Al. The features in microstructures are resolved in the SEM images that are reported in Figure 7. The microstructures of the as-built samples, as reported in Figure 6, look to be heterogeneous and had visible bands/layers. Sample E shows partially melted powder particles. The samples are also characterized by fine microstructure and agglomerated lamellar or lamellar laths (samples A, B, and C), lamellae grains (samples D and F), and formed grains (sample E). Liu et al. [41] observed coarsening in the lamellar laths, while here the laths grew small and refined into grains with the increase in Al. The features in microstructures are resolved in the SEM images that are reported in Figure 7. Indeed, the lamellar laths size decreased with an increase in Al content (see samples A and C). The variation in microstructures with Al content was anticipated. The microstructure of sample E has surface porosity and grains of different size (20-60 µm) and colour. The α2-grains (white) seem to be generally growing around the γ-grains (black). It is not clear, at this magnification, if there are Indeed, the lamellar laths size decreased with an increase in Al content (see samples A and C). The variation in microstructures with Al content was anticipated. The microstructure of sample E has surface porosity and grains of different size (20-60 µm) and colour. The α 2 -grains (white) seem to be generally growing around the γ-grains (black). It is not clear, at this magnification, if there are lamellae structures inside these grains and how they are arranged. Nevertheless, given the observed features, this alloy can be classified as near-gamma before heat treatment. Samples A-C have agglomerated, thick lamellar which formed randomly (direction of growth is not obvious). The size of these lamellar decreased and became refined from sample A (thick) to sample C (thin), while the formed grains were not fully precipitated. This alloy was classified as nearly lamellar before heat treatment. Lamellar colonies that formed inside the grains can be confirmed only for samples D and F. The lamellae grains of sample F are bigger than that of sample D but are, by inspection, less than or about 20 µm. These two alloys can be classified as a duplex before heat treatment.
Reconciling results reported in Table 3 and microstructures in Figure 6, it is noticed that samples A and D and also samples B and F had similar Al contents, therefore they should have had a similar microstructure. This was the case even though sample A had a coarse, thick lamellar laths structure, while sample D had small, refined, short lamellar laths, which grew around the observed lamellae grains structure. Samples B and F were not that different in structure from each other. Obviously, we expected samples C and E to be different from each other and the other samples. Sample F had a gamma grain structure, while sample C had a lamellar lath structure. The effect of heat treatment on the resulting microstructure of these alloys was studied and the results are reported next. Note that since samples A and D and samples B and F had similar Al contents, we only reported on a microstructure of each set post-heat treatment, but hardness values are reported for all samples.

Heat Treated Samples
Effects of heat treatment on the resulting microstructures of Ti-Al alloys in the duplex and gamma-phase alloys are extensively reported in the literature [1,4,6,10,17,30]. This section reports on the microstructures of Ti-Al alloys (across the Al content spectrum) after heat treatment at 1200 • C and cooling inside the oven.

Chemical Composition and Microstructures
It was observed that post-heat treatment samples B and E had a duplex microstructure with Al contents of 57 and 54 at.%, respectively, while sample A was nearly lamellar, with an Al content of about 53 at.% and sample C had a fully lamellar microstructure with an Al content of 49 at.%. The white spots observed on sample A were aluminium oxide (Al 2 O 3 ) ,with the bright white phase on the microstructures rich in Al. The black grains that resemble fractures in almost all samples, particularly samples B and E, were rich in Ti. These results are summarized in Figure 8. lamellae structures inside these grains and how they are arranged. Nevertheless, given the observed features, this alloy can be classified as near-gamma before heat treatment. Samples A-C have agglomerated, thick lamellar which formed randomly (direction of growth is not obvious). The size of these lamellar decreased and became refined from sample A (thick) to sample C (thin), while the formed grains were not fully precipitated. This alloy was classified as nearly lamellar before heat treatment. Lamellar colonies that formed inside the grains can be confirmed only for samples D and F. The lamellae grains of sample F are bigger than that of sample D but are, by inspection, less than or about 20 µm. These two alloys can be classified as a duplex before heat treatment.
Reconciling results reported in Table 3 and microstructures in Figure 6, it is noticed that samples A and D and also samples B and F had similar Al contents, therefore they should have had a similar microstructure. This was the case even though sample A had a coarse, thick lamellar laths structure, while sample D had small, refined, short lamellar laths, which grew around the observed lamellae grains structure. Samples B and F were not that different in structure from each other. Obviously, we expected samples C and E to be different from each other and the other samples. Sample F had a gamma grain structure, while sample C had a lamellar lath structure. The effect of heat treatment on the resulting microstructure of these alloys was studied and the results are reported next. Note that since samples A and D and samples B and F had similar Al contents, we only reported on a microstructure of each set post-heat treatment, but hardness values are reported for all samples.

Heat Treated Samples
Effects of heat treatment on the resulting microstructures of Ti-Al alloys in the duplex and gamma-phase alloys are extensively reported in the literature [1,4,6,10,17,30]. This section reports on the microstructures of Ti-Al alloys (across the Al content spectrum) after heat treatment at 1200 °C and cooling inside the oven.

Chemical Composition and Microstructures
It was observed that post-heat treatment samples B and E had a duplex microstructure with Al contents of 57 and 54 at.%, respectively, while sample A was nearly lamellar, with an Al content of about 53 at.% and sample C had a fully lamellar microstructure with an Al content of 49 at.%. The white spots observed on sample A were aluminium oxide (Al2O3) ,with the bright white phase on the microstructures rich in Al. The black grains that resemble fractures in almost all samples, particularly samples B and E, were rich in Ti. These results are summarized in Figure 8.  Summarizing the results of the as-built and heat treated samples, it can be said that the Al content increased drastically in the heat-treated samples and the microstructures of sample F (duplex) and sample A (nearly lamellar) were the same before and after heat treatment. Sample E transformed from near-gamma to duplex, sample B from nearly lamellar to duplex, sample C from nearly lamellar to duplex, and sample D from duplex to nearly lamellar. In this regard, only sample D performed better in its as-built state, while samples B, E, and F exhibited their best engineering properties post-heat treatment, given that they all had a duplex microstructure. The overall hardness results of these alloys (before and after heat treatment) are reported next.

Micro-Hardness
Hardness measurements can be used to determine certain mechanical properties of the alloy. For instance, the general relationship between strength and hardness has been studied before [11]. The hardness profiles of Ti-Al alloys before heat treatment are typically wavy and this is, in general, attributed to the heterogeneity and phase precipitates of the build [30]. Heat treatment is supposed to give a homogenized structure, leading to a "smooth" hardness profile of the build. The hardness profiles of the heat-treated samples are reported in Figure 9, while Figure 10 (bar-diagram) compares the overall hardness of the samples before and after heat treatment. In addition, it has been reported that the (α 2 ) phase is harder than the γ phase and that a two-phase (α-Ti+α 2 -Ti 3 Al) alloy is harder than pure Ti. The latter is attributed to the precipitation ability of the α 2 -Ti 3 Al in the two-phase and the solid solution strengthening of Al [30,38]. It is anticipated that the overall hardness of these alloys should be different, but similar for the alloys with same Al content. Therefore, samples A and D and samples B and F would have similar overall hardness values, respectively. Overall, we anticipated that sample E would be the least hard, followed by sample C, then samples B and F, and finally samples A and D. However, it must be noted that, in addition to the resulting microstructure and phase, the hardness profiles of the samples are affected by other factors, such as material segregation, phase precipitation, and grain-size. The hardness profile is reported in Figure 9. Summarizing the results of the as-built and heat treated samples, it can be said that the Al content increased drastically in the heat-treated samples and the microstructures of sample F (duplex) and sample A (nearly lamellar) were the same before and after heat treatment. Sample E transformed from near-gamma to duplex, sample B from nearly lamellar to duplex, sample C from nearly lamellar to duplex, and sample D from duplex to nearly lamellar. In this regard, only sample D performed better in its as-built state, while samples B, E, and F exhibited their best engineering properties postheat treatment, given that they all had a duplex microstructure. The overall hardness results of these alloys (before and after heat treatment) are reported next.

Micro-Hardness
Hardness measurements can be used to determine certain mechanical properties of the alloy. For instance, the general relationship between strength and hardness has been studied before [11]. The hardness profiles of Ti-Al alloys before heat treatment are typically wavy and this is, in general, attributed to the heterogeneity and phase precipitates of the build [30]. Heat treatment is supposed to give a homogenized structure, leading to a "smooth" hardness profile of the build. The hardness profiles of the heat-treated samples are reported in Figure 9, while Figure 10 (bar-diagram) compares the overall hardness of the samples before and after heat treatment. In addition, it has been reported that the (α2) phase is harder than the γ phase and that a two-phase (α-Ti+α2-Ti3Al) alloy is harder than pure Ti. The latter is attributed to the precipitation ability of the α2-Ti3Al in the two-phase and the solid solution strengthening of Al [30,38]. It is anticipated that the overall hardness of these alloys should be different, but similar for the alloys with same Al content. Therefore, samples A and D and samples B and F would have similar overall hardness values, respectively. Overall, we anticipated that sample E would be the least hard, followed by sample C, then samples B and F, and finally samples A and D. However, it must be noted that, in addition to the resulting microstructure and phase, the hardness profiles of the samples are affected by other factors, such as material segregation, phase precipitation, and grain-size. The hardness profile is reported in Figure 9. Hardness curves reported in Figure 9 conclude that Sample D was hardest with fewer points rising above the homogeneity line, which can be imagined to be around 620 HV. Similar observations can be made for sample C, which was also hard. Samples A and E were wavy, while samples B and F were stable. These two samples can be interpreted as homogeneous. The overall hardness values Hardness curves reported in Figure 9 conclude that Sample D was hardest with fewer points rising above the homogeneity line, which can be imagined to be around 620 HV. Similar observations can be made for sample C, which was also hard. Samples A and E were wavy, while samples B and F were stable. These two samples can be interpreted as homogeneous. The overall hardness values (HV) for the samples after heat treatment were 351 ± 25.6, 325 ± 30.75, 420 ± 22.2, 662 ± 29.6, 314 ± 25.95, and 308 ± 17.35 for samples A-F. These hardness values were compared to the as-built samples as shown in Figure 10 (bar-diagram).  Figure 10 (bar-diagram). Figure 10. Overall hardness of the alloys compared. Figure 10 indicates clearly that only the hardness value of sample D increased with heat treatment. Meanwhile, the hardness values of samples B, E, and F decreased to an equivalent value. This is acceptable since these alloys have the same microstructure (duplex) post-heat treatment. More importantly, the hardness value of sample B decreased to almost half after heat treatment, while sample C's hardness did not significantly change. The observation of hardness across the samples support the findings of Guo et al. [30], who reviewed that the Ti3Al (α2) phase is harder than the TiAl (γ) phase. The pure γ phase samples (samples A, B, E, and F) reported less hardness when compared to sample C, which was dual phase (α2 + γ). It is the presence of the α2 phase that might be contributing to the observed increase in hardness. This observation is contradicted only by sample D, which had the most hardness after heat treatment. This contradiction might be sustained by observations that two-phase (α-Ti + α2-Ti3Al) is harder than γ and pure Ti. It could be that during heat treatment the samples experienced effective Al solid solution strengthening and precipitation of the α2-phase, hence it was the hardest. This study overall corroborates observations made by Guo et al. [30]. This is sustained by sample F (pure Ti) and sample D (α + α2) before heat treatment. Overall, Figure 10 concludes that an alloy with a duplex microstructure will be the least hard, followed by nearly lamellar and then fully lamellar microstructures. This observation holds only when sample D is accepted as an outlier or ambiguous.

Conclusions
In this paper, in-situ laser alloying of Ti and Al into a binary Ti-Al was studied. The effects of increasing the Al content and heat treatment on the resulting microstructure and hardness of binary Ti-Al are reported. The results show that, by keeping the laser power and titanium flow-rate constant while increasing the Al content via rpm, different microstructures can be produced in accordance with the binary Ti-Al alloy phase diagram. The as-built samples were found to be harder, in general, than the heat-treated (1200 °C) samples. This is attributed to the difference in the resulting microstructures before and after heat treatment. This study also showed that Al content in the builds did not increase linearly with the increase in rpm as expected. Due to this observation, the hardness of the as-built samples also did not linearly decrease as expected. Post-heat treatment, the hardness of the samples decreased with the increase in Al content, except for one sample which was thought  Figure 10 indicates clearly that only the hardness value of sample D increased with heat treatment. Meanwhile, the hardness values of samples B, E, and F decreased to an equivalent value. This is acceptable since these alloys have the same microstructure (duplex) post-heat treatment. More importantly, the hardness value of sample B decreased to almost half after heat treatment, while sample C's hardness did not significantly change. The observation of hardness across the samples support the findings of Guo et al. [30], who reviewed that the Ti 3 Al (α 2 ) phase is harder than the TiAl (γ) phase. The pure γ phase samples (samples A, B, E, and F) reported less hardness when compared to sample C, which was dual phase (α 2 + γ). It is the presence of the α 2 phase that might be contributing to the observed increase in hardness. This observation is contradicted only by sample D, which had the most hardness after heat treatment. This contradiction might be sustained by observations that two-phase (α-Ti + α 2 -Ti 3 Al) is harder than γ and pure Ti. It could be that during heat treatment the samples experienced effective Al solid solution strengthening and precipitation of the α 2 -phase, hence it was the hardest. This study overall corroborates observations made by Guo et al. [30]. This is sustained by sample F (pure Ti) and sample D (α + α 2 ) before heat treatment. Overall, Figure 10 concludes that an alloy with a duplex microstructure will be the least hard, followed by nearly lamellar and then fully lamellar microstructures. This observation holds only when sample D is accepted as an outlier or ambiguous.

Conclusions
In this paper, in-situ laser alloying of Ti and Al into a binary Ti-Al was studied. The effects of increasing the Al content and heat treatment on the resulting microstructure and hardness of binary Ti-Al are reported. The results show that, by keeping the laser power and titanium flow-rate constant while increasing the Al content via rpm, different microstructures can be produced in accordance with the binary Ti-Al alloy phase diagram. The as-built samples were found to be harder, in general, than the heat-treated (1200 • C) samples. This is attributed to the difference in the resulting microstructures before and after heat treatment. This study also showed that Al content in the builds did not increase linearly with the increase in rpm as expected. Due to this observation, the hardness of the as-built samples also did not linearly decrease as expected. Post-heat treatment, the hardness of the samples decreased with the increase in Al content, except for one sample which was thought to either be an outlier or that there was precipitation of α 2 and solid solution strengthening that occurred during heat treatment.

Future Work
In future studies, we would look into the effects of niobium on these binary alloys. Niobium is a ternary alloying element that is able to improve the oxidation and creep resistance of titanium aluminides (Kothari et al., 2012). The results should be interesting, particularly for sample A, which showed signs of oxidation post-heat treatment. EDS analyses clearly resolved that the Al 2 O 3 phase/layer formed during heat treatment. Before this future study is carried out, we will look at the effects of increased Al content via rpm on the overall Al content in the builds. This will help in addressing the observations made in this study and charter a way forward for our future studies.