Journal of Materials Science & Technology  2019 , 35 (8): 1543-1554 https://doi.org/10.1016/j.jmst.2019.04.002

Orginal Article

Influence of welding parameters on interface evolution and mechanical properties of FSW Al/Ti lap joints

Mingrun Yuab, Hongyun Zhaob, Zhihua Jiangb, Zili Zhangb, Fei Xub, Li Zhouab*, Xiaoguo Songab

a State Key Laboratory of Advanced Welding and Joining, Harbin Institute of Technology, Harbin 150001, China
b Shandong Provincial Key Laboratory of Special Welding Technology, Harbin Institute of Technology at Weihai, Weihai 264209, China

Corresponding authors:   *Corresponding author at: State Key Laboratory of Advanced Welding and Joining,Harbin Institute of Technology, Harbin 150001, China.E-mail address: zhou.li@hit.edu.cn (L. Zhou).

Received: 2018-12-13

Revised:  2019-03-8

Accepted:  2019-03-9

Online:  2019-08-05

Copyright:  2019 Editorial board of Journal of Materials Science & Technology Copyright reserved, Editorial board of Journal of Materials Science & Technology

More

Abstract

Friction stir welding (FSW) was performed to produce Al/Ti lap joints under various welding conditions. More heat was generated when rotational rate increased or traversing rate decreased. Two types of Al/Ti interfaces - mixed interface and diffusive interface - were formed under different welding conditions. The diffusive interface was formed with low heat input, and the mixed interface was formed more heat. The grains at the mixed interface were larger than those at the diffusive interface because of the higher heat input. Moreover, the microstructure of the mixed interface had a lower texture intensity compared with that of the diffusive interface, which was attributed to the enhanced continuous dynamic recrystallization (CDRX). TiAl3 was formed at the diffusive interface. When the interface varied to the mixed interface as heat input increased, TiAl was fomed within the Al/Ti mixture following the formation of TiAl3. In addition, TiAl3 precipitates were observed in the diffusion layer. The hardness value of the mixed interface was higher than 350 HV, due to the larger amount of intermetallic compounds (IMCs). The lap shear strength reached a maximum value of 147 MPa with medium heat input and an interface that exits in a critical state between diffusive and mixed interfaces. All the specimens fractured at the interface, which was attributed to the presence of IMCs.

Keywords: FSW ; Al/Ti lap joint ; Interfacial characteristics ; Mechanical properties

0

PDF (7172KB) Metadata Metrics Related articles

Cite this article Export EndNote Ris Bibtex

Mingrun Yu, Hongyun Zhao, Zhihua Jiang, Zili Zhang, Fei Xu, Li Zhou, Xiaoguo Song. Influence of welding parameters on interface evolution and mechanical properties of FSW Al/Ti lap joints[J]. Journal of Materials Science & Technology, 2019, 35(8): 1543-1554 https://doi.org/10.1016/j.jmst.2019.04.002

1. Introduction

Aluminum (Al) alloys have low density, high specific strength, excellent corrosion resistance, and reasonable cost [1,2]. Hence, the Al alloys are likely to replace steel as the major structural material used in transportation industry to lower the greenhouse gas emission and fuel consumption [3,4]. Although offering excellent mechanical properties and corrosion resistance, titanium (Ti) alloys have limited applications because of their high cost [5]. Al/Ti hybrid structures could combine the advantages of both Al and Ti alloys, and could be widely applied in the automotive and aviation industries [2,6]. However, Al and Ti alloys have many physical and metallurgical differences, making it difficult to join them. In particular, Al/Ti dissimilar joints are weakened by the intermetallic compounds (IMCs) that form at their interfaces [7,8]. Therefore, it is critical to identify a reliable welding process to develop such Al/Ti hybrid structures.

FSW is a novel solid-state joining process, which typically generates less heat compared to the conventional welding processes [9,10]. Owing to the low heat input, intermetallic reactions are suppressed during FSW, and, FSW is thus considered a promising technique for the joining of dissimilar alloys [[11], [12], [13]]. Considering the FSW of Al and Ti alloys, the Al microstructure of the Al/Ti weld is similar to that of an Al weld, as reported by Buffa et al. [14]. Consequently, to date, researchers have primarily focused on the interfacial microstructures and mechanical properties of FSW Al/Ti joints. Chen et al. [15] and Li et al. [16] observed a lamellar structure, in which Al and Ti alloys were mixed, at the Al/Ti interface. Bang et al. [17,18] pointed out that fracture always occurs at the interface of FSW Al/Ti joints. In addition, Aonuma et al. [19,20] determined that such fractures can be attributed to TiAl3, which is formed at the interface. Furthermore, Song et al. [21] and Choi et al. [22] determined that TiAl and Ti3Al are both formed within the Al/Ti mixture. Zhang et al. [23] reported that a large quantity of Al-Ti IMCs are formed within the Al/Ti mixture. Moreover, Chen and Nakata [24] determined that the interfacial structure of such FSW Al/Ti joints varies remarkably as the welding speed increases. In addition, Wu et al. [25] observed a 100-nm thick TiAl3 layer at the interface of a FSW Al/Ti joint, without the presence of other IMCs or Al/Ti mixtures. In a study by Chen and Yazdanian [26], the lap shear strength along the FSW Al/Ti joints was extremely unpredictable, and, the average strength of the joint was 101 MPa. Recently, friction surfacing (FS) assisted FSW was developed to join Al and Ti alloys, in order to avoid the wear of pin tool [27].

Although the FSW of dissimilar Al/Ti joints has been studied, the interfacial microstructures of FSW Al/Ti joints have not been studied systematically. The investigations about Al/Ti FSW were mainly focused on the interfacial phases and the bonding strength. Some results are conflicting due to the different interfacial microstructures. The evolution of the interface with the variation in the welding parameters has not been elucidated. Furthermore, the relationship between the interfacial characteristics and mechanical properties of such joints has not been discussed in detail. In the present study, an Al/Ti lap joint was subjected to FSW using various parameters. The heat generated during the FSW process was estimated, and the welding parameters were optimized. In addition, the interface evolution, mechanical properties, and fracture behavior of the joints were investigated using various welding conditions. The relationship between the heat input, interfacial characteristics, and mechanical properties has been proposed in this study.

2. Materials and experimental procedure

The AA6061 alloy and the Ti6Al4V alloy were employed as the base materials (BMs) in this study. The as-received AA6061 and Ti6Al4V plates had thicknesses of 3 mm and 2 mm, respectively. For each plate, the thickness was measured at five locations to ensure that the Ti matrix would be stirred during FSW. Both plates were cut into 250 mm × 75 mm pieces, respectively. The microstructural characteristics of the BMs are illustrated in Fig. 1. The AA6061 plate was T6 temper 6061 alloy, consisting of both elongated and equiaxed grains with an average diameter of 15.79 μm. In the case of the Ti6Al4V alloy, the plate consisted of rolled and recrystallized grains with an average diameter of 1.95 μm. The chemical compositions of the Al and Ti alloys are shown in Table 1.

Fig. 1.   Microstructure of the BMs: (a) AA6061, (b) Ti6Al4V.

Table 1   Chemical compositions of the base materials (wt%).

Elements
AlMgSiMnTiVFe.
AA6061Bal.1.610.500.15---
Ti6Al4V6.33---Bal.4.270.20

New window

During the FSW process, the Al plate was placed on the Ti plate at on the retreating side, as shown in Fig. 2. For temperature measurements, Type-K thermocouples (NiCr-NiSi, intended for use between -200 °C and +1350 °C) were positioned at a distance of 1.6 mm from the weld center, on the advancing side between the plates. The pin tool, made from the W-25Re alloy, has a concave shoulder with a diameter of 12 mm, and a 3.1-mm long conical probe with a maximum diameter of 3 mm. The probe was inserted into the Al alloy with a shoulder plunging depth of 0.1 mm. As shown in Table 2, the Al/Ti lap joint underwent FSW at rotational rates ranging from 600 rpm to 1400 rpm, and welding speeds ranging from 60 mm/min to 140 mm/min, using a constant tool tilt angle of 3.0°.

Fig. 2.   Schematic illustration of FSW Al/Ti lap joints.

Table 2   Detailed welding parameters for FSW Al/Ti lap joints.

Rotational rate (rpm)Traversing rate (mm/min)Tilting angle (deg.)
6001003
800
1000
1200
1400
100060
80
120
140

New window

Following welding, the specimens were cut perpendicular to the welding direction for subsequent metallographic analysis and mechanical tests. The microstructural characteristics of the joints were examined via optical microscopy (OM, DSX 510), scanning electron microscopy (SEM, MERLIN Compact), energy-dispersive X-ray spectrometry (EDS, OCTANE PLUS), electron backscatter diffraction techniques (EBSD, DIGIVIEW 5), and transmission electron microscopy (TEM, JEM-2100). The metallographic specimens intended for optical observation were etched using Keller's etchant (1.0 ml HF + 1.5 ml HCl + 2.5 ml HNO3 + 95 ml H2O). In addition, specimens for EBSD were electrically polished with a solution of 10 ml HClO4 + 90 ml CH3OH at 0 °C at 40 V for the Al microstructure and 60 V for the Ti microstructure. The EBSD was performed with a step size of 0.25 μm. For each joint, three 10-mm-wide tensile specimens were tested at a crosshead speed of 1 mm/min, using a mechanical testing machine (Instron 5967). During test, the Al plate was placed at the retreating side. Following the mechanical tests, the fracture surfaces were further examined via SEM, EDS, and X-ray diffraction (XRD, D/max 2500).

3. Results and discussion

3.1. Torque and temperature

In the case of FSW, the torque is closely related to the mechanical behavior of the pin tool, which determines the heat input. In addition, the heat generated during FSW directly affects the thermal history of the weld. Therefore, the torque and temperature were both measured during the FSW of the Al/Ti lap joint to explore the relationship between the welding parameters and heat input. Fig. 3 shows the variation in the spindle torque, as a function of time, during the FSW of the Al/Ti lap joints, which was performed under various welding conditions. The variation in the spindle torque can be divided into five stages: (1) commencement, (2) plunging, (3) holding, (4) welding, and (5) withdrawal. When the traversing rate was maintained at 100 mm/min, the average spindle torque of the welding stage was determined to be 11 N m at 600 rpm. As the rotational rate increased to 1400 rpm, the average spindle torque decreased to 5 N m. However, when the rotational rate remained at 1000 rpm and the traversing rate varied, the average spindle torque remained relatively stable at 7 N m. It could be concluded that, under the welding conditions used in the present study, the spindle torque is primarily affected by the rotational rate, and the traversing rate has little influence on the spindle torque. In addition, the welding torque is determined to be significantly higher than that measured for stirring Al matrix only (1000 rpm, 100 mm/min). Hence, the Ti alloy is considered to stir in this study.

Fig. 3.   Spindle torques with various welding parameters: (a) v = 100 mm/min, (b) ω = 1000 rpm.

The heat input during FSW is closely related to the welding parameters and torque, as discussed by Yang et al. [28]. The heat generated during FSW can be expressed by a torque-based power model as follows:

where, Q is the total heat (J) generated during the welding stage, Qω is the heat (J) generated by the rotation, QF is the heat (J) generated by traveling, ω is the rotational rate (rpm), t is time (s), Mw is the welding torque (N m), v is the traversing rate (mm/min), and Fx is the force (N) generated by the pin tool in the welding direction. In addition, Firouzdor et al. [29] indicated that, under various welding parameters, the heat input owed to traveling is less than or equal to 0.05% of the heat generated by the rotation. Therefore, QF could be neglected while calculating the total heat of the FSW process [29,30]. Consequently, Eq. (1) could be simplified as:

where l is the welding length (mm).

Fig. 4 shows the thermal history of the interface under various welding conditions. As mentioned previously, the thermocouples were placed at the edge of the stir zone, on the advancing side, between the Al and Ti plates. At rotational rates of 600 rpm, 1000 rpm, and 1400 rpm, the peak temperatures were 471 °C, 520 °C, and 564 °C, respectively, when the traversing rate was maintained at 100 mm/min, as shown in Fig. 4(a). Similar thermal cycles were obtained when the rotational rate was varied under a constant traversing rate. However, as shown in Fig. 4(b), under a constant rotational rate, the peak temperature and the duration of the thermal cycle both increased when the traversing rate was decreased. The peak temperature increased from 482 °C to 542 °C when the traversing rate was decreased from 140 mm/min to 60 mm/min under a constant rotational rate of 1000 rpm. Meanwhile, the duration of the thermal cycle was significantly extended when a low traversing rate was used. Hence, it could be concluded that the rotational and traversing rates both have a significant influence on the peak temperature. Otherwise, the duration of the thermal cycle primarily relates to the traversing rate. As shown by Eq. (2), an increase in the rotational rate results in higher heat input and higher temperatures, even though the torque slightly decreases. Similarly, a low traversing rate results in larger heat accumulation, and consequently, the peak temperature increases as the traversing rate decreases. Moreover, the variation in the peak temperature with ω2/v is shown in Fig. 4(c). It can be observed that the peak temperature is directly proportional to ω2/v, which is in accordance with the investigation by Mishra and Ma [31].

Fig. 4.   Measured temperature with various welding parameters: (a) v = 100 mm/min, (b) ω = 1000 rpm, (c) variation of peak temperatures with ω2/v.

3.2. Macrostructure

Images of the FSW Al/Ti lap joints and their cross sections are shown in Fig. 5, Fig. 6, respectively. When the traversing rate was 100 mm/min, the Al and Ti alloys were satisfactorily welded via FSW, at various rotational rates ranging from 600 rpm to 1400 rpm. As shown in Fig. 5, when a rotational rate of 1000 rpm or lower is used, the Al/Ti lap joints are perfect and defect free. However, when the rotational rate increased to over 1000 rpm, flashes on the surface and hooks within the joint could be observed. When the rotational rate was maintained at 1000 rpm, flashes and hooks were also formed at welding speeds of 60 mm/min and 80 mm/min, as shown in Fig. 6. When the traversing rate was increased to 100 mm/min or higher, flashes and hooks could hardly be observed within the joints, which could be attributed to the less plasticity of the matrix due to the lower heat input. The heat input appears to have a significant influence on the macrostructure of the FSW Al/Ti lap joints. Under conditions with a high heat input, the excessively plasticized Al and Ti alloys were extruded by the pin tool, resulting in the formation of flashes and hooks.

Fig. 5.   Surface appearance and cross sections of the FSW Al/Ti lap joints at 100 mm/min.

Fig. 6.   Surface appearance and cross sections of the FSW Al/Ti lap joints at 1000 rpm.

The evolution of the interfaces of the FSW Al/Ti lap joints is illustrated in Fig. 7. Based on the SEM image, when the rotational rate was 600 rpm and the traversing rate was 100 mm/min, deformation of the Ti alloy could be hardly observed at the interface. When the rotational rate was increased to 1000 rpm or higher, the Ti alloy was significantly deformed. Hooks and Ti fragments could be observed when the rotational rate increased to 1400 rpm. When the rotational rate was maintained at 1000 rpm and the traversing rate was increased from 60 mm/min to 140 mm/min, the interface, which initially possessed hooks and Ti fragments, transformed into an interface without significant deformation and Al/Ti intermixing. The unobvious deformation of the Ti plate could be attributed to the high elasticity of the Ti matrix, which was owed to the relatively low heat input [32]. Wu et al. [33] have reported on a FSW Ti-6Al-4V alloy that exhibits superplasticity at temperatures over 650 °C. Hence, it is reasonable to consider that the plasticity is sharply enhanced as the temperature is increased. Consequently, the deformation of the Ti plate gradually occurs as the heat input increases. This is because of the increased plasticity of the Ti matrix. Based on the aforementioned interfacial macrostructure, two types of Al/Ti interfaces, namely a diffusive interface and a mixed interface, could be observed under the various welding conditions. The diffusive interface was observed in the joints that were formed under conditions involving low heat input. When the heat input increased, the diffusive interface transformed into a mixed interface, which was accompanied by the formation of Ti fragments and hooks, owing to the sufficient plasticization of the Ti alloy.

Fig. 7.   Interface evolution with different welding parameters.

3.3. Interfacial characteristics

The EBSD results that were obtained for the various interfaces are shown in Fig. 8. In the maps, the color of each grain is based on its crystal orientation, and the areas with limited data, such as the Ti fragments, are shown as black-colored blocks. In addition, the high-angle grain boundaries (HAGBs) are plotted using black-colored lines, while white-colored lines are used for the low-angle grain boundaries (LAGBs). In the case of the joint that was welded at 800 rpm and 100 mm/min, refined and equiaxed grains could be observed on the Al side of the diffusive interface. The LAGBs, which have a fraction of 11.5%, are typically well-arranged. In addition, the LAGB-to-HAGB segment transformation could be identified from the EBSD map of the Al side, suggesting that grain subdivision occurs at the interface, which is a characteristic of continuous dynamic recrystallization (CDRX). The grains are considered to undergo CDRX during FSW, resulting in grains with an average diameter of 3.27 μm. Furthermore, based on the obtained polar figures (PFs), the {001}<110> shear texture, with a maximum intensity of 7.047, was determined to be the predominant texture of both the diffusion and mixed interfaces; this correlates with the C component of the ideal shear textures of face-centered cubic (f.c.c.) metals [34]. In addition, the Ti microstructure experienced a relatively small amount of friction and stirring during FSW. Consequently, the grains experienced little deformation and were insufficiently heated on the Ti side of the diffusive interface; this resulted in a microstructure that consisted of coarse and recrystallized grains with an average diameter of 5.78 μm. Based on the PFs, it can be determined that the predominant texture is the C2 component of the ideal shear textures of hexagonal close-packed (h.c.p.) metals [35].

Fig. 8.   EBSD results of diffusive interface.

In the case of the mixed interface, the Al and Ti alloys were intermixed at the interface when the welding parameters were 1000 rpm and 60 mm/min, as shown in Fig. 9. In addition, a larger number of Ti fragments were stirred into the Al matrix, resulting in the regions with limited data within the EBSD map. When the welding parameters were varied, the LAGB fraction decreased to 4.2% because of grain growth, indicating that CDRX was promoted during FSW. Consequently, the Al grains were also highly refined and equiaxed, having an average grain diameter of 5.12 μm at the mixed interface. In addition, the average diameter of the Al grains at the mixed interface was significantly higher than that of the Al grains at the diffusive interface, which is attributed to the grain growth that under the influence of the higher heat input. Although the welding parameters were varied, based on the PFs, the predominant texture was also considered to be the C shear texture. However, the maximum texture intensity sharply decreased to 3.727. In addition, the Ti microstructure experienced remarkable friction and stirring at the mixed interface during FSW. Therefore, the grains were significantly deformed and heated on the Ti side, resulting in a microstructure that consisted of lath and recrystallized grains with an average diameter of 5.45 μm. The reduction in the grain size, compared with that of the Ti microstructure of the diffusive interface, could be attributed to CDRX, which was promoted by the heat generation and deformation that occurred during FSW. Based on the PFs, the predominant texture of this material was also considered to be the C2 shear texture. However, the intensity of the C1 texture was enhanced, which consequently promoted the total texture intensity of the Ti microstructure. In general, it appears that the interfacial characteristics are significantly influenced by the variation of welding parameters. Owing to CDRX, the welding parameters also affect the morphologies of the interfacial microstructure. Meanwhile, the texture that is formed during FSW is closely related to the material behavior; this has also been reported in previous studies [34,35]. However, based on the aforementioned experimental results, the variation in the welding parameters leads to little change in the direction of the predominant texture, which is determined by the stir and welding directions. However, the texture intensity is significantly affected by the welding parameters through heat effects.

Fig. 9.   EBSD results of mixed interface.

Fig. 10 shows the magnified SEM images and EDS mapping results of various interfaces. The detailed chemical compositions and probable phases of the points marked in yellow are listed in Table 3. Considering the SEM and EDS results, no distinguishable IMC layer or composition variation can be observed at the diffusive interface. Tiny Ti fragments, which could hardly be observed in Fig. 7, can be observed in the vicinity of the interface, as shown in the magnified SEM image. Ti diffused into the Al alloy at the interface, resulting in a diffusion layer could be depicted within the elemental maps. Consequently, the Ti content at point A is 2.02% (at.%), which is higher than the solubility of Ti in Al (0.147% (at.%)) [21]. Therefore, TiAl3, which has the lowest standard free energy of formation among the Al-Ti IMCs [8], is considered to form at the diffusive interface. The chemical composition of the Ti side matched that of the Ti BM (point B). When the heat input increased, a wider Ti-diffusion layer (point C) was observed at the mixed interface, as shown by the EDS maps. Further, the Ti alloy was sufficiently plasticized and mixed with the Al alloy at the interface, resulting in the aforementioned Al/Ti mixture. The Al/Ti intermixing resulted in the formation of a continuous IMC layer at the mixed interface. In addition, TiAl3 is considered to form within the Al/Ti mixture, as shown by the EDS data obtained for point D, which is colored dark gray in the SEM image. Furthermore, the light gray-colored part (point E) is composed of 51.81% Al (at.%) and 46.57% Ti (at.%). The Ti content of point E is considerably higher than that of point D. This indicates that TiAl is formed as the distance from the interface increases. This was also reported by Choi et al. [22]. Consequently, it could be concluded that the Al and Ti alloys are joined by both diffusion and metallurgical reactions during the FSW process.

Fig. 10.   SEM micrographs and EDS maps of different interfaces.

Table 3   EDS results of the points shown in Fig.9 (at.%).

PointsAlTiVMgPossible phase
A96.152.02-1.83Al + TiAl3
B11.2385.703.07-Ti BM
C94.993.100.891.02Al + TiAl3
D70.2628.880.860.70TiAl3
E51.8146.571.350.27TiAl

New window

In addition, TEM and selected area electron diffraction (SAED) techniques were performed to explore the detailed microstructures of the various interfaces, as shown in Fig. 11. In the case of the diffusive interface, TiAl3 precipitates could be observed within the diffusion layer (1000 rpm, 140 mm/min). This indicates that the Ti diffused into and oversaturated the Al alloy during FSW. Further, the Ti content of the oversaturated Al(Ti) solution decreased because of the natural precipitation of the IMC. Based on the SAED results, the precipitates within the diffusion layer were determined to be TiAl3. On the other hand, it is suggested that the IMC layer consists of two layers at the mixed interface (1400 rpm, 100 mm/min). The wider layer is determined to consist of TiAl3, while the thinner one is composed of TiAl. This indicates that TiAl3 was initially formed at the Al/Ti interface because of the influence of the FSW. Subsequently, TiAl is considered to form at the interface between Ti and TiAl3, via the diffusion of Al across the TiAl3 layer. This is because the diffusion rate of Al is significantly higher than that of Ti in Al-Ti IMCs [36].

Fig. 11.   TEM results of Al/Ti interfaces: (a) TEM image of the diffusion layer, (b) TEM image of the mixed interface, (c) SAED pattern of the TiAl3 in (a), (d) SAED pattern of the TiAl in (b).

3.4. Mechanical properties

The microhardness was tested perpendicularly to the Al/Ti interface, at the center of the welds. The hardness distributions of the joints are shown in Fig. 12. In the case of the BMs, the Al and Ti alloys have hardness values of 110 HV and 330 HV, respectively. In the case of the Al/Ti lap joints, the hardness of the Ti side remains at approximately 330 HV, and sharply decreases to approximately 78 HV on the Al side. The variation in the hardness of the Al alloy is due to dynamic recrystallization and the dissolution of precipitates, which occur during the FSW [37]. When the rotational rate was increased under a constant traversing rate, the average hardness of the welded Al alloy decreased from 83 HV (at 600 rpm) to 70 HV (at 1400 rpm). This could be attributed to grain growth and the dissolution of precipitates, which occur because of the higher heat input. However, the hardness of the Ti alloy remains almost unchanged under a rotational rate of 600 rpm. When the rotational rate was increased to 1400 rpm, the Al/Ti mixture and the IMCs formed at the interface cause an increase in the hardness, to 356 HV. Similarly, the hardness of the welded Al alloy also increases from 75 HV (at 60 mm/min) to 80 HV (at 140 mm/min). In the case of the joint that was welded at 60 mm/min, the hardness of the interface increased to 351 HV. In addition, since Ti fragments were stirred into the Al alloy, the hardness of the NZ beside the interface slightly increased to 90 HV. It is suggested that the heat input significantly affects the hardness of the FSW Al/Ti lap joints. Hence, to produce Al/Ti lap joints with a favorable hardness distribution, it would be effective to restrict the heat input by regulating the welding parameters.

Fig. 12.   Hardness distribution of the FSW Al/Ti lap joints with various welding parameters: (a) v = 100 mm/min, (b) ω = 1000 rpm.

Fig. 13 shows the tensile test results obtained for the FSW Al/Ti lap joints. It can be observed that the welding parameters have a significant influence on the lap shear strength of the joints. The maximum lap shear strength is achieved using welding parameters of 1000 rpm and 100 mm/min, as shown in Fig. 13(a) and (b). The variation in the lap shear strength with the welding heat is shown in Fig. 13(c). It can be observed that the maximum strength, of 147 MPa, is achieved when the welding parameters are 1000 rpm and 100 mm/min. When the heat input was further increased, the strength slightly decreased to 140 MPa (1000 rpm, 80 mm/min) at first, and then sharply declined to 114 MPa (1400 rpm, 100 mm/min). This could be attributed to the excessive quantity of IMCs that formed in the joints. In addition, considering the interface evolution shown in Fig. 7, it could be concluded that the optimal tensile shear strength could be achieved with an interface that exists in a critical state between those of the diffusive interface and mixed interface. The Al/Ti lap joints were obtained with a higher lap shear strength in this study, compared to former studies, as shown in Fig. 13(d). The lap shear strength is promoted by 45.5% at least [26].

Fig. 13.   Tensile test results of the Al/Ti lap joints obtained under different welding conditions: (a) variation of lap shear strength with rotational rate, (b) variation of lap shear strength with rotational rate, (c) variation of lap shear strength with ω2/v, (d) comparison with former studies.

3.5. Fracture analysis

Following the tensile shear tests, the profiles of the fractured specimens were observed using OM and SEM, as shown in Fig. 14. It is suggested that the hook structure has a remarkable influence on the fracture mode. In the case of the diffusive interface, the fracture developed linearly along the interface (Fig. 14(a)). However, the hook structure significantly altered the fracture path along the mixed interface (Fig. 14(b)). The cracks were considered to initiate at the root of the hook structure, and then propagated into the Ti alloy. Meanwhile, a crack also initiated within the Al/Ti mixture of the mixed interface because of the residual stress that was caused by the formation of the IMCs. The crack that initiated within the Al/Ti mixture propagated along the IMC layer and converged with the crack that was initiated at the root of the hook structure. This resulted in the small amount of Ti alloy that remained at the root of the hook structure.

Fig. 14.   Fracture profiles of the Al/Ti joints: (a) diffusive interface, (b) mixed interface.

In addition, SEM and EDS analyses were conducted on the fractured specimens, as shown in Fig. 15. It is suggested that all the specimens fractured unevenly along the Al/Ti interface. In the case of the joint that was welded at 1000 rpm and 140 mm/min, the primary crack propagated along the Al/Ti interface. The secondary cracks mainly propagated into the Al alloy. However, in the case of the mixed interface, the fracture developed irregularly within the Al/Ti mixture. A considerable quantity of Al/Ti IMCs formed at the mixed interface, which could significantly increase the residual stress [38]. In addition, the IMCs were brittle and tough, which resulted in a stress concentration during the tensile-shear testing. It could be concluded that a crack initiated at the edge of the interface, which was owed to the influence of the stress concentration and crystalline mismatches, and then propagated along the Al/Ti IMC layer, resulting in failure.

Fig. 15.   SEM image and elements distributions of different interfaces: (a) diffusive interface, (b) mixed interface.

The XRD spectra of the fracture surfaces are shown in Fig. 16. No obvious TiAl3 layer could be observed within the SEM image. However, based on the XRD spectra of the fracture surfaces, TiAl3 was considered to form at the diffusive interface, which is in accordance with the EDS results and the results above. The intensity and width of the peak at 2θ = 44.74° significantly increased. This signifies that TiAl, in addition to TiAl3, exists on the fracture surface of the mixed interface. The aforementioned experimental results indicate that the Al and Ti plates were metallurgically joined via the diffusion of the elements and the formation of the IMCs. In addition, the Al/Ti IMCs that formed at the interface signify the strength of the Al/Ti lap joints that had undergone FSW. However, the experimental results indicate that the fracture mechanisms of the diffusion interface and mixed interface differ. At the diffusive interface, the Al/Ti lap joints were joined during FSW via the diffusion of the elements and the formation of IMCs. Consequently, the cracks primarily propagated along the Al/Ti interface owing to the formation of TiAl3. In addition, owing to the secondary crack, which propagated into the Al alloy, a considerable amount of the Al alloy remained on the fracture surface at the Ti side. However, the fracture developed irregularly along the mixed interface. The crack propagated along both the TiAl3 and TiAl, and therefore, TiAl3 and TiAl could be observed on the fracture surfaces of both sides of the failed joints. It can be concluded that such joint fracture is primarily governed by the IMCs at the interface.

Fig. 16.   XRD results of the fracture surfaces of different interfaces: (a) diffusive interface, (b) mixed interface.

4. Conclusions

FSW was conducted on the lap joints of AA6061 and Ti6Al4V alloys using various parameters. The interfacial characteristics and mechanical properties of the joints were investigated. Based on the present study, the following conclusions can be drawn:

(1) Based on the welding torque and thermal history, the welding parameters have a significant influence on the heat input during the Al/Ti FSW. The torque was primarily affected by the rotational rate. In addition, the peak temperature was determined to be directly proportional to ω2/v.

(2) The Al/Ti interface varied when the welding parameters were changed. The diffusive interface was formed with low heat input, and the mixed interface was formed with more heat. Textures were observed within the interfacial microstructures of both the Al and Ti. When the heat input increased, the predominant texture remained unchanged, and, the texture intensities decreased, which could be attributed to the promotion of CDRX. The diffusive interface was formed via elemental diffusion, and a small quantity of TiAl3 was also formed at this interface. As the heat input increased, TiAl3 and TiAl were both observed at the mixed interface.

(3) The mechanical properties of the welds were closely related to their interfacial characteristics. The interfacial hardness of the mixed interfaces increased to over 350 HV because of the formation of a considerable quantity of IMCs at the mixed interface. Furthermore, the lap shear strengths of the FSW Al/Ti lap joints reached a maximum value of 147 MPa with medium heat input and an interface that exits in a critical state between diffusive and mixed interfaces. In addition, all the specimens fractured at the interface because of the formation of the IMCs.

Acknowledgement

The authors gratefully acknowledge the financial support of the project from Shandong Provincial Natural Science Foundation, China (No. ZR2016EEM43).

The authors have declared that no competing interests exist.


/