Robotic Nd:YAG Fiber Laser Welding of Ti-6Al-4V Alloy

: In the present study, Ti6Al4V titanium alloy plates were joined using a robotic ﬁber laser welding method. The laser welding process was carried out at two different welding speeds. Effects of different heat input conditions on the microstructure and mechanical properties of robotic ﬁber laser welded joints were investigated. Some grain coarsening was observed in the microstructure of weld metal in samples joined using high heat input, compared to those using low heat input, and volume rates of primary α structures increased in the weld metal. The microstructure of weld metal in samples joined using low heat input was made of basket-weave or acicular α ' grains and primary β grains in grain boundaries. Tensile and yield strength of samples joined using low heat input were higher than for those joined using high heat input, but their ductility was lower.


Introduction
Ti6Al4V is the most used and preferred titanium alloy. Ti6Al4V is an α + β group alloy in which aluminum is the α stabilizer and vanadium is the β stabilizer [1,2]. Due to its low density, heat resistance, high mechanical strength, corrosion resistance, and biocompatibility properties, the Ti6Al4V alloy has a wide range of uses [3,4]. The Ti6Al4V alloy is used in jet engines, spacecraft, missiles, the auto industry, pressure vessels, orthodontics, medical implants and surgical tools [5]. TIG (Tungsten Inert Gas), friction welding, friction stir welding, plasma arc, electron beam and laser beam welding methods can be used for welding the Ti6Al4V alloy [3]. When welded joints of titanium alloys are conducted using high heat input, a high level of distortion and contamination risk are involved. Temperatures of over 500 • C in welded joints of Ti6Al4V lead to absorption of deleterious gases such as oxygen, nitrogen and hydrogen in the welding seam, leading to lower mechanical strength [6]. Slower cooling due to high heat input in the welding of titanium and its alloys results in grain coarsening and porosity formation in the welding seam [7]. Grain coarsening and porosities cause lower mechanical strength, especially in welded joints. In welding titanium and its alloys, a narrow welding seam and a microstructure made of fine grains should be achieved by lowering the heat input transfer to the weld zone, and increasing the cooling rate. On the other hand, very fast cooling conditions lead to weld metal microstructures that are completely martensitic in nature, and thus have a lower toughness [8].
Laser welding differs from traditional welding methods on account of its lower heat input, higher condensation energy, faster welding speed, narrower weld zone, its ability to produce joints with deeper penetration, its higher mechanical strength, lower distortion, and its ability to weld without the use of additional metal [9][10][11][12][13][14][15][16]. Fiber laser welding machines have been used for decades. They work at between 100 W average power and 6 kW peak power. The most important advantage of fiber lasers is the easiness of laser control, due to the wave length of the laser beam. Energy is easily transferred through fiber optical cables. The fiber laser welding method, on the other hand, has advantages of lower beam distortion rate, flexible beam distribution, lower maintenance cost, higher efficiency and high quality welded joints [17]. In contrast to traditional welding methods, laser welding provides control over heat input, making it possible to achieve the desired metallurgical properties and mechanical strength levels in welded joints of the Ti6Al4V alloy. In the present study, Ti6A14V plates were joined using a robotic Nd:YAG (neodymium-doped yttrium aluminum garnet) fiber laser. The effects of welding speed on the microstructure and mechanical properties of laser welded Ti6A14V alloy were investigated.

Material and Welding Process
In the present study, Ti6Al4V alloy, an alloy preferred in many areas of industry, was used. Its chemical composition is given in Table 1. Experimental material was prepared with dimensions of 100 mm × 330 mm × 4 mm for the welding process. Experimental samples were joined in a flat position using a FANUC ROBOT R-2000 iB 210F Nd:YAG fiber laser welding machine with 4 kW power (Fanuc, Yamanashi Prefecture, Japan), without using any additional metal ( Figure 1). The welding parameters are given in Table 2. They work at between 100 W average power and 6 kW peak power. The most important advantage of fiber lasers is the easiness of laser control, due to the wave length of the laser beam. Energy is easily transferred through fiber optical cables. The fiber laser welding method, on the other hand, has advantages of lower beam distortion rate, flexible beam distribution, lower maintenance cost, higher efficiency and high quality welded joints [17]. In contrast to traditional welding methods, laser welding provides control over heat input, making it possible to achieve the desired metallurgical properties and mechanical strength levels in welded joints of the Ti6Al4V alloy. In the present study, Ti6A14V plates were joined using a robotic Nd:YAG (neodymium-doped yttrium aluminum garnet) fiber laser. The effects of welding speed on the microstructure and mechanical properties of laser welded Ti6A14V alloy were investigated.

Material and Welding Process
In the present study, Ti6Al4V alloy, an alloy preferred in many areas of industry, was used. Its chemical composition is given in Table 1. Experimental material was prepared with dimensions of 100 mm × 330 mm × 4 mm for the welding process. Experimental samples were joined in a flat position using a FANUC ROBOT R-2000 iB 210F Nd:YAG fiber laser welding machine with 4 kW power (Fanuc, Yamanashi Prefecture, Japan), without using any additional metal ( Figure 1). The welding parameters are given in Table 2.    The impact notch test, on the other hand, was carried out at room temperature using an ALŞA testing machine (Alşa, Istanbul, Turkey). Hardness measurement was performed by a GALILEO test machine (GALILEO, Antegnate, Italy) applying 200 g to the tip for 15 s. For determination of microstructural changes, an etching procedure using Kroll's solution was used. A NIKON optic microscope (5-100× magnification, Nikon, Tokio, Japan) JEOL JSM 6060 LV and JEOL JSM 7001 LV scanning electron microscope (SEM, Jeol Ltd., Tokyo, Japan) and OXFORD X-MAX 80 (energy dispersive spectrometer, EDS) equipment (Oxford Instruments, Oxfordshire, UK) were used for microstructural examinations.

Macro and Microstructural Examinations
Macro images belonging to A1 and A2 samples are shown in the Figure 2. As can be understood from the macro images, undercut, underfill and cracks were not observed in the welding seam, but cracks were observed in optical microscope image ( Figure 5b).

Mechanical Tests and Characterization of Microstructure
In order to determine mechanical properties of welded joints, four test samples were prepared for each parameter, based on the ISO 4136: 2012 standard for the tensile test, and on the ISO 9016: 2012 standard for the impact notch test. The tensile test was conducted using an INSTRON tensile test machine with 100 kN capacity (Instron, MA, USA) at a 10 mm/min crosshead speed. The impact notch test, on the other hand, was carried out at room temperature using an ALŞA testing machine (Alşa, Istanbul, Turkey). Hardness measurement was performed by a GALILEO test machine (GALILEO, Antegnate, Italy) applying 200 g to the tip for 15 seconds. For determination of microstructural changes, an etching procedure using Kroll's solution was used. A NIKON optic microscope (5-100× magnification, Nikon, Tokio, Japan) JEOL JSM 6060 LV and JEOL JSM 7001 LV scanning electron microscope (SEM, Jeol Ltd., Tokyo, Japan) and OXFORD X-MAX 80 (energy dispersive spectrometer, EDS) equipment (Oxford Instruments, Oxfordshire, United Kingdom) were used for microstructural examinations.

Macro and Microstructural Examinations
Macro images belonging to A1 and A2 samples are shown in the Figure 2. As can be understood from the macro images, undercut, underfill and cracks were not observed in the welding seam, but cracks were observed in optical microscope image ( Figure 5b). It was observed that the microstructure of the base metal was made of equiaxed α grains (light colored grain, hexagonal close packed) and β grains in grain boundaries (dark colored grain, body centered cubic) ( Figure 3). It was observed that the microstructure of the base metal was made of equiaxed α grains (light colored grain, hexagonal close packed) and β grains in grain boundaries (dark colored grain, body centered cubic) ( Figure 3). Equiaxed grain structures in the base metal increase fracture toughness and support strength [18]. The HAZ (Heat Affected Zone) and weld metal microstructure depend on no filler metal being used with the melting of base metal. It is known that when additional filler metal is used, major differences arise in the HAZ and the weld metal microstructure. Microstructure shapes with respect to the chemical composition of alloy and cooling rate. Fine grained microstructure forms in the weld metal. However, grain coarsening can be expected with use of high laser power, or with joints Equiaxed grain structures in the base metal increase fracture toughness and support strength [18]. The HAZ (Heat Affected Zone) and weld metal microstructure depend on no filler metal being used with the melting of base metal. It is known that when additional filler metal is used, major differences arise in the HAZ and the weld metal microstructure. Microstructure shapes with respect to the chemical composition of alloy and cooling rate. Fine grained microstructure forms in the weld metal. However, grain coarsening can be expected with use of high laser power, or with joints carried out at very slow welding speeds. During the solidification process, β grains display growth in the direction of the heat flow [19].
Due to the occurrence of a rapid cooling regime, a dominant martensitic microstructure exists in the weld metal of Ti6Al4V alloy joined by laser welding. In addition, with an increase of heat input, transformed beta grains with acicular alpha grains can form in the weld metal. In the HAZ and weld metal microstructure images for the study (Figure 4), a dominant martensite microstructure, cellular dendrite and columnar dendrite occurred, due to the rapid cooling related to the low heat input that is a characteristic feature of the laser welding method. Equiaxed grain structures in the base metal increase fracture toughness and support strength [18]. The HAZ (Heat Affected Zone) and weld metal microstructure depend on no filler metal being used with the melting of base metal. It is known that when additional filler metal is used, major differences arise in the HAZ and the weld metal microstructure. Microstructure shapes with respect to the chemical composition of alloy and cooling rate. Fine grained microstructure forms in the weld metal. However, grain coarsening can be expected with use of high laser power, or with joints carried out at very slow welding speeds. During the solidification process, β grains display growth in the direction of the heat flow [19].
Due to the occurrence of a rapid cooling regime, a dominant martensitic microstructure exists in the weld metal of Ti6Al4V alloy joined by laser welding. In addition, with an increase of heat input, transformed beta grains with acicular alpha grains can form in the weld metal. In the HAZ and weld metal microstructure images for the study (Figure 4), a dominant martensite microstructure, cellular dendrite and columnar dendrite occurred, due to the rapid cooling related to the low heat input that is a characteristic feature of the laser welding method.  When the microstructures of the welded joints were investigated, it was observed that the HAZ arises from fine acicular α′ martensite and primary α structures due to fast cooling ( Figure 4). These types of microstructures can be observed at cooling temperatures lower than the β transus temperature [20]. It has been reported that cooling rates higher than 410 °C/s should be achieved to obtain a fully martensitic microstructure in Ti6Al4V alloy [3,21]. Studies reported that when the base metal is heated to temperatures over the β transus temperature during welding, martensite phase transformation from the high-temperature β phase takes place, and α′ is formed with the basket weave [22]. At cooling rates slower than 410 °C/s, on the other hand, transformed α phase occurs preferentially in the primary β grain boundary. Figure 4 shows that cooling rate as a function of laser welding parameters, i.e., as a function of heat input, affected HAZ width. Because of joining using low heat input, a narrow HAZ developed. Some grain coarsening occurred due to the weld thermal When the microstructures of the welded joints were investigated, it was observed that the HAZ arises from fine acicular α martensite and primary α structures due to fast cooling (Figure 4). These types of microstructures can be observed at cooling temperatures lower than the β transus temperature [20]. It has been reported that cooling rates higher than 410 • C/s should be achieved to obtain a fully martensitic microstructure in Ti6Al4V alloy [3,21]. Studies reported that when the base metal is heated to temperatures over the β transus temperature during welding, martensite phase transformation from the high-temperature β phase takes place, and α is formed with the basket weave [22]. At cooling rates slower than 410 • C/s, on the other hand, transformed α phase occurs preferentially in the primary β grain boundary. Figure 4 shows that cooling rate as a function of laser welding parameters, i.e., as a function of heat input, affected HAZ width. Because of joining using low heat input, a narrow HAZ developed. Some grain coarsening occurred due to the weld thermal cycle in the HAZ microstructure of samples joined using high heat input. It was observed that the weld metal microstructures of samples especially joined using low heat input was composed of basket-weave, acicular α grains and primary β grains in the grain boundaries. As a result of the slower cooling rate in laser welding due to the low heat input, a diffusionless transformation of β phase into a martensitic structure takes place. In addition, local microcrack formation was observed in the weld metal of samples joined using low heat input (Figure 5b). Crack formation is always possible in welded joints like laser welding, where cooling is fast [5]. Presence of martensite, a hard and brittle phase, is among the major factors triggering the formation of microcracks in welded joints of titanium alloys and stainless steels. There may always be crack risk in the welding seam under rapid cooling conditions. Especially in laser beam welding, the occurrence of a rapid solidification rate can increase crack risk. To be able to decrease crack risk, it is suggested that samples should be fixed on the ground using clamps during welding [5]. In this study, samples were fixed on the ground using clamps during welding (Figure 1). Titanium alloys are reactive with ambient gases at high temperatures. For this reason, a shielding gas was used to protect the molten weld pool against oxidation until enough cooling had occurred. In spite of these protective conditions, crack and porosity formation were observed in the weld metal of the A2 sample. The crack occurrence in the A2 sample indicates that the molten weld pool was not protected as expected during the welding process. Setting the gas pressure incorrectly could conceivably be the cause of this result. As is well known, titanium alloys are quite resistant to crack risk. Weld cracking, however, can still be caused by contamination cracking, such as hydrogen-delayed cracking or liquation of iron-rich eutectic [3]. In this study, it should also be stated that cracks occurred in the weld metal of the A2 sample carried out locally. Less coarsened columnar β grain structures were observed in weld metal than in the HAZ (Figure 4). The volume rate of primary α structures increased in the weld metal of samples joined using high heat input. Besides, porosity formations were observed in weld metal microstructures (Figure 5a). Porosity formation is frequently encountered in laser welded titanium alloys. Gas type porosities arise especially from hydrogen, which cannot be removed from the weld pool during the solidification of the welding seam [23]. Porosity formation is expected to lower the mechanical strength of welded joints. which cannot be removed from the weld pool during the solidification of the welding seam [23]. Porosity formation is expected to lower the mechanical strength of welded joints. EDS analysis performed on welded samples showed that no major element losses occurred in the base metal, HAZ or weld metal ( Figure 6). Based on these findings, it can be stated that heat input did not have any negative effect on laser welded joints. EDS analysis performed on welded samples showed that no major element losses occurred in the base metal, HAZ or weld metal ( Figure 6). Based on these findings, it can be stated that heat input did not have any negative effect on laser welded joints. EDS analysis performed on welded samples showed that no major element losses occurred in the base metal, HAZ or weld metal ( Figure 6). Based on these findings, it can be stated that heat input did not have any negative effect on laser welded joints.

Tensile Test
Tensile testing was carried out to determine the mechanical properties of the base metal (A sample) and laser welded A1 and A2 samples. Results are given in Table 3. Results of the tensile experiment revealed that tensile strength, yield strength and percentage elongation of the base metal were higher than those of laser welded samples joined using different heat inputs. Porosities, microcracks and microstructural transformations occurring in the welding seam are among the possible reasons for lower mechanical strength of welded joints compared to the base metal. The sample joined using low heat input (A2 sample) had higher tensile and yield strength, but a lower percentage elongation compared to the sample joined using high heat input (A1 sample). The laser welded A2 sample had a tensile strength of 840 MPa and a ductility level of 8%, while the laser welded A1 sample had a tensile strength of 830 MPa and a ductility level of 8.5%. As a result of the increase in heat input transferred to the weld zone, coarsening in grain sizes caused a decrease in tensile and yield strength, even in small amounts. Lower ductility in welded α + β alloys depends on coarsened primary β grain ratio and the formation of intragranular acicular microstructures [24]. Microsegregations occurring during the solidification of welding seam are also among the factors lowering ductility. After tensile testing, fractures occurred in the weld metal of the A2 sample, while fractures in the A1 sample occurred in the HAZ. It could be expected that fractures in A2 sample would be in the HAZ and fracture in the A1 sample would be in the weld metal; but, as a result of microstructural transformations occurring in the HAZ or weld metal (mostly because of grain coarsening), the cause for these results can be thought to be localized or heterogeneous plastic strains occurring in the welded joints during tensile testing.
When SEM images of fracture surfaces of the base metal and the laser welded samples were examined after tensile testing (Figure 7), it could clearly be seen that ductile fracture, which is a typical fracture mode for titanium, was observable in the base metal and A2 samples by looking at dimples. In addition, for the A1 sample, intragranular fracture occurred with a tear ridges fracture mode. Because welded joints have lower ductility than the base metal after tensile testing, base metal experienced the ductile dimple fracture mode. There were little signs of plastic deformation in the welded A1 sample, with the fracture path appearing to be intergranular along colony boundaries, or cleavage across colonies, which indicates that the fracture mode changed from ductile to cleavage fracture mode [25]. In addition, it is thought that microstructural transformations and loading rates play a significant role in fracture mode changes. The fracture surface morphology of the welds was much rougher than that of the base metal. There were lots of dimples with different sizes and depths in the fracture surfaces of the joints in A1, and obvious tear ridges as well. No inclusion formation was observed in the fracture surface images of welded joints.

Charpy V-Notch Impact Test
The impact toughness of the base metal and the laser welded samples was determined at room temperature. The impact toughness of the base metal was higher than that of the laser welded samples (17, 13 and 14 J, respectively). The impact toughness of the sample joined using slow welding speed i.e., higher heat input, was higher than that of the sample joined using low heat input

Charpy V-Notch Impact Test
The impact toughness of the base metal and the laser welded samples was determined at room temperature. The impact toughness of the base metal was higher than that of the laser welded samples (17, 13 and 14 J, respectively). The impact toughness of the sample joined using slow welding speed i.e., higher heat input, was higher than that of the sample joined using low heat input (14 and 13 J, respectively). Other investigators who obtained similar findings also stressed that, in order to increase the impact toughness of titanium alloys, martensite structures should be transformed into acicular α or Widmanstatten α grain structures by increasing the heat input, which occurs as a result of using either a lower welding speed or a higher laser power [26]. Cooling rate has an extremely important effect on the mechanical properties of α + β alloys. It was stressed that toughness decreases are experienced at fast cooling conditions, due to the formation of finer structured α platelets, while at slower cooling rates, toughness is improved by diverting crack propagation paths with the coarsening of α platelets [27].
Fracture surfaces of the base metal and welded samples after impact testing were examined using SEM (Figure 8). The base metal was broken in ductile fracture mode, while A1 and A2 samples had fractures in the mode of intergranular fractures with either a quasi-cleavage type or tear ridges.

Microhardness
Due to the higher power intensity of laser welding, when compared to traditional welding methods, it results in low heat input and fast solidification, leading to high hardness levels in the welded joints. Distributions of hardness in welded joints is directly related to microstructure, and martensite in the microstructure is the most important factor for the hardness increase in the HAZ and weld metal of Ti6Al4V alloy. When the average hardness levels of samples joined using different welding speeds-i.e., different heat input, were examined, it was found that the hardness of the weld metal and HAZ was higher than for the base metal ( Figure 9).

Microhardness
Due to the higher power intensity of laser welding, when compared to traditional welding methods, it results in low heat input and fast solidification, leading to high hardness levels in the welded joints. Distributions of hardness in welded joints is directly related to microstructure, and martensite in the microstructure is the most important factor for the hardness increase in the HAZ and weld metal of Ti6Al4V alloy. When the average hardness levels of samples joined using different welding speeds-i.e., different heat input, were examined, it was found that the hardness of the weld metal and HAZ was higher than for the base metal ( Figure 9).
Due to the higher power intensity of laser welding, when compared to traditional welding methods, it results in low heat input and fast solidification, leading to high hardness levels in the welded joints. Distributions of hardness in welded joints is directly related to microstructure, and martensite in the microstructure is the most important factor for the hardness increase in the HAZ and weld metal of Ti6Al4V alloy. When the average hardness levels of samples joined using different welding speeds-i.e., different heat input, were examined, it was found that the hardness of the weld metal and HAZ was higher than for the base metal ( Figure 9).  The reason for this outcome has been reported to be the formation of harder and finer acicular α martensite structures in the weld metal and HAZ, compared to the structure of the α + β base metal [28,29]. Hardness level was observed to decrease as either laser power increased or welding speed decreased; i.e., as heat input increased. Investigators reported that hardness levels decreased as a result of coarsened primary β structure due to higher heat input [29]. With lower heat input, on the other hand, an increase in hardness is expected, owing to the formation of finer-grained α structures [23]. The reason for the increase of hardness in weld metal compared to base metal was reported to be compact regions of α with high-density dislocations and phase boundaries [22]. Hardness of the weld metal was determined to be higher than that of the HAZ for both welding speeds. It has been explained by some investigators that the decrease in hardness from weld metal to HAZ was caused by the decreased martensite content and Widmanstatten α structure [30]. When hardness results were examined generally, hardness levels of the HAZ and weld metal were similar. This similarity was linked to the very low heat conduction coefficient (6.7 W·m −1 ·K −1 ) of Ti6Al4V alloy [31]. Thermal energy transferred to the weld zone of Ti6A14V alloy can stay there longer before dispersing [31]. The higher hardness in the HAZ and weld metal compared to the base metal is expected to produce more strength in the joints than in the base metal. A similarity was expected between the hardness and tensile strength results. Nevertheless, such a similarity was not observed in the present study, due to the presence of porosities and microcrack formations in the weld metal and HAZ.

Conclusions
(1) The microstructure of the base metal comprised equiaxed α grains, while the weld metal structure displayed a martensitic microstructure formation because of fast cooling at both welding speeds. Slight grain coarsening was observed in the HAZ and weld metal of the sample joined using high heat input. Some porosity and local microcrack formations were observed in the weld metal microstructure. The crack occurrence in the A2 sample indicates that the molten weld pool was not protected as expected during welding process. Setting the gas pressure incorrectly may be a potential cause of this result. As is well known, titanium alloys are quite resistant to crack risk. The weld cracking, however, can still be caused by contamination cracking, such as hydrogen-delayed cracking or liquation of iron-rich eutectic.
(2) Based on tensile testing, the strength of the base metal was higher than that of the laser welded joints. Such an outcome could be due to porosity formations observed in the welding seam, as well as to possible microcracks that might have occurred in the welding seam. The tensile and yield strength of samples joined using low heat input were higher than those of samples joined using high heat input, but percentage elongation was higher in samples joined using high heat input. After tensile testing, fracturing in the A2 sample occurred in the weld metal, while fracturing in the A1 sample occurred in the HAZ. As a result of microstructural transformations occurring in the HAZ or weld metal (mostly because of grain coarsening), it is thought that localized or heterogeneous plastic strain occurring in welded joints during tensile testing may have caused these results. While fractures after tensile testing occurred in the form of ductile fractures for the base metal and A2 samples, the A1 samples exhibited intragranular fractures occurring with a tear ridges fracture mode. Because the twelded joints show less ductility than the base metal after tensile testing, the base metal exhibited a ductile dimple fracture mode. There were minor signs of plastic deformation in the welded A1 sample, with the fracture path appearing to be either intergranular along colony boundaries or cleavage across colonies, which indicates that the fracture mode changed from a ductile to a cleavage fracture mode.
(3) Impact toughness was lower in laser welded joints than in the base metal. Impact toughness of the material joined using lower welding speed i.e., with higher heat input, was higher than that of samples joined using low heat input. Among the leading causes increasing impact toughness with higher heat input is the transformation of martensite structures to acicular α or Widmanstatten α grain structures. After impact testing, the base metal was broken in a ductile fracture mode, while the A1 and A2 samples had fractures in the mode of intergranular fracture with a quasi-cleavage type.
(4) In terms of the hardness of samples joined using different heat inputs, the hardness of the weld metal and HAZ was higher than that of the base metal. The hardest zone was the weld metal. Increasing laser power or decreasing welding speed i.e., increasing heat input, both resulted in lower hardness. Coarsening of primary β structures, which is a result of increased heat input, led to lower hardness.