Study on the Optimization of the Tensile Properties of an Al-Li Alloy Friction Stir-Welding T-Joint

Abstract

The softening of aluminum–lithium alloy welded joints generally leads to a reduction in mechanical properties. In this study, a piece of 2A97-T3 aluminum–lithium alloy with a thickness of 2.8 mm was selected as the test material, and the tool and process used for wire-filled stationary shoulder friction stir welding (SSFSW) were developed. By increasing the bearing area of the softening zone, an equal-strength T-joint was manufactured. Good weld formation was obtained when the rotation speed was set to 2000 rpm and the welding speed ranged from 100 to 120 mm/min. The thickness of the softening zone was controlled by adjusting the reserved gap between the shoulder and the workpiece. The softening mechanism of the weld joint was revealed. The softening was attributed to the coarsening of the main precipitated phases (T1 and θ′ phases) in the heat-affected zone (HAZ) and the dissolution of precipitated phases in the thermo-mechanically affected zone (TMAZ). Grain refinement in the nugget zone (NZ) led to a certain fine-grained strengthening effect, although the precipitated phase was almost completely dissolved. Due to the thermal effect of second-pass welding, the hardness value of the NZ and HAZ in the center of the skin further decreased, and the minimum hardness was approximately 70% that of the base material. Tensile testing results indicated that the softening effect was largely offset by the increased bearing area of the softening zone, resulting in the successful welding of high-strength Al-Li alloy T-joints with equal strength. The strength coefficient was found to be 0.977.

1. Introduction

The increasing urgency of creating lighter advanced aircraft has led to a growing emphasis on the utilization of aluminum–lithium alloys [1]. Known for its high specific strength and stiffness in thin-walled structures, this alloy represents a significant avenue for achieving overall weight reduction [2]. Its thin-walled ribbed structure primarily comprises two components, namely, the skin and the stringer, with the T-joint formed by welding them together serving as its key characteristic. However, the welding process for Al-Li alloy T-joints poses considerable challenges, resulting in low performance levels that often fail to meet usage requirements. Key issues include the poor weldability of the Al-Li alloy, the loss of lithium through burning, and porosity during the fusion welding process [3]. Additionally, severe welding residual stress and deformation, along with pronounced softening in the heat-affected and fusion zones, contribute to a reduction in the overall load-bearing capacity of the joint [4].

Friction stir welding (FSW) is a solid-phase welding technique that offers significant advantages in the welding of Al-Li alloy thin-walled ribbon-reinforced structures [5,6]. In FSW, there is no melting and solidification of weld metal, thereby avoiding the loss of Li through burning and the formation of pores [7,8]. Furthermore, due to the low welding peak temperature and the good rigidity of the welding jig employed in this technique, welding residual stress can be effectively controlled, and welding deformation can be minimized. Additionally, high automation and low pollution are inherent advantages of friction stir welding. However, the friction stir welding of T-joints presents challenges, including limited welding space and process difficulties. Softening of the joint remains unavoidable, thereby restricting this technique’s application in aerospace lightweight structures.

Recently, Zhao et al. [9] and Jesus et al. [10] utilized conventional friction stir welding to weld the outer side of the skin (the outer side of the T-joint), resulting in the formation of a traditional T-joint. This joint typically takes the form of a butt or lap joint, which may reduce the bearing area of the joint and lead to issues such as poor internal corner welding. Cui et al. [11] also pointed out that weld defects such as a lack of complete bonding and tunnels were difficult to eliminate with the through-skin configuration, a problem associated with the difficulty in ensuring insufficient material flow near the tip of the pin. To solve the problems associated with outer-wall thinning, Sun et al. [12] employed stationary shoulder friction stir welding (SSFSW) to weld the inner side of the T-joint. The thermal field of the first and second passes thermally welded the material and resulted in increased over-aging of the weld material. Yu et al. [13] also conducted T-joint SSFSW welding experiments. The highest tensile strength of the skin and the stringer reached 50.8% and 56.9% of the base material, indicating that achieving a T-joint with high strength remains challenging. The Harbin Institute of Technology conducted T-joint welding experiments using an 8 mm thick piece of 2A14 aluminum alloy, employing a process of stationary shoulder friction stir welding without wire filling [14]. This approach resulted in the production of satisfactory welded joints; however, the section thickness of the joint was reduced. The tensile strength of the joint reached 71.21% that of the base metal.

Given the aforementioned requirements and issues, the test material selected for this study was a 2.8 mm thick piece of 2A97-T3 aluminum–lithium alloy. First, a wire-filled stationary shoulder friction stir welding (SSFSW) tool was designed specifically for T-joints, and a welding process was subsequently developed. Following an explanation of the mechanisms governing weld formation and joint softening, an equal strength manufacturing scheme for T-joints with thin-walled structures of Al-Li alloy is proposed. The results demonstrate that, under optimized SSFSW parameters, the bearing area of the softening zone increased through the utilization of rounded shoulder extrusion and a controlled gap, leading to the production of equal-strength T-joints.

2. Experiments

The welded material examined in this paper is a 2.8 mm thick plate of 2A97-T3 aluminum–lithium alloy, which was provided by Aluminum Corporation of China Limited. Its chemical composition and tensile properties are presented in

Table 1. Chemical composition and mechanical properties of 2A97.

Chemical Composition									Mechanical Properties
Li	Si	Fe	Cu	Mn	Mg	Zn	Zr	Al	Rm	Rp0.2	A
1.46	0.03	0.06	3.68	0.26	0.48	0.44	0.11	Bal.	447.4 MPa	327.5 MPa	16.1%

. The wire used in the study was fabricated from the plate, with a square cross-section, and its length is parallel to the rolling direction.

In this paper, the 2A97-T3 aluminum alloy skin and stringer were incorporated into a T-joint, as depicted in Figure 1b, with square wire added to the apex of the inner corner of the T-joint, as illustrated in Figure 1a. To facilitate wire filling during welding, a press wheel, as shown in Figure 1a, was installed at the leading end of the stationary shoulder to exert pressure on the wire. During the welding process, the equipment moves to rotate the press wheel, driving the wire into the stirring area. Under the high-speed rotation of the pin, the material in the stirring area reaches a plasticized state, and recrystallization occurs, eliminating the interface between the wire, the skin, and the stringer. The non-rotating stationary shoulder compacts the plasticized material and alters the root of the T-joint from the right angle depicted in Figure 1b to the rounded angle shown in Figure 1c through the rounded R3 structure at the bottom of the shoulder. After single-side welding (1S side) is completed, the T-joint is flipped so that welding can be carried out on the other side (side 2S), resulting in the double-side welded T-joint depicted in Figure 1d. The welding direction (WD) is parallel to the rolling direction (RD) of the plate.

After the completion of welding, the appearance of the weld is observed, followed by the inspection of a typical T-joint for sectional metallographic observation. The microstructure of the defect-free T-joint was analyzed using EBSD (JSM-7800 F, JEOL Ltd., Tokyo, Japan) and TEM (JEM-2100F, JEOL Ltd., Tokyo, Japan), with the analysis encompassing grain morphology, grain size, and precipitated phases. A defect-free T-joint was selected for tensile and microhardness testing at room temperature. In the actual working process of the reinforced structure, the stringer primarily serves to strengthen rigidity, producing significantly lower force compared to the skin. Therefore, this study focuses on conducting tensile tests on the skin. A schematic diagram of the tensile specimen is illustrated in Figure 2, in which the tensile loading direction is perpendicular to the WD. The fracture location and morphology were analyzed using optical microscope (Smartzoom5) and scanning electron microscopes (JSM-7800 F, JEOL Ltd., Tokyo, Japan).

3. Results and Discussion

3.1. Macromorphology of T-Joint

The appearance of the 2A97-T3 Al-Li alloy T-joint produced using wire-filled stationary shoulder friction stir welding (SSFSW T-joint) is depicted in Figure 3. Figure 3a shows the overall appearance of the first welded side (1S side), while Figure 3b displays the appearance of the second welded side (2S side). Figure 3c–e depict the weld appearance of 2515 (rotation speed, 2500 rpm; welding speed, 150 mm/min), 2010 (rotation speed, 2000 rpm; welding speed, 100 mm/min), and 2012 (rotation speed, 2000 rpm; welding speed, 120 mm/min) joints, respectively. As indicated in Figure 3c, the 2515 T-joint exhibits poor welding quality with groove-like defects at surface, whereas the T-joints of 2010 and 2012 exhibit smooth appearances without apparent defects (as depicted in Figure 3d,e). The cross sections of the 2515, 2010, and 2012 T-joints were inspected, and the overall structures of the joints were examined via optical microscopy. The results are illustrated in Figure 4a–c. As shown in Figure 4a, there was a noticeable material shortage in the 2515 T-joint, primarily concentrated on the upper surface of the skin. Conversely, there were no defects in the 2010 and 2012 T-joints. Moreover, the junction between the skin and the stringer transitions from a right angle to a rounded structure with a smooth transition, as shown in Figure 4b and Figure 4c, respectively.

3.2. Microstructure of the T-Joint

In Figure 4c, the left side represents the first welding side (1S side), and the right side represents the second welding side (2S side). Different positions within the joint experience varying temperatures and plastic deformation induced by friction stir welding, resulting in the formation of distinct microstructures in each region. Based on the microstructural characteristics observed in different areas, the 2A97-T3 T-joint can be classified into the nugget zone (NZ), the thermo-mechanically affected zone (TMAZ), the heat-affected Zone (HAZ), and the base metal (BM). The NZ is situated at the central region of the T-joint, surrounded by the TMAZ, while the HAZ encompasses the outer region of the TMAZ. The overlapping area between the 1S and 2S sides, positioned within the central region of the NZ and subjected to double stirring, is defined as the secondary stir zone (NZ2 in Figure 4c).

The microstructures of BM, HAZ, TMAZ, NZ1, and NZ2 were analyzed using EBSD, based on the Figure 5b–f positions illustrated in Figure 5a, with the corresponding results presented in Figure 5b–f, respectively. In Figure 5b, a typical rolling structure with coarse and long grains can be observed in the BM. Similarly, in Figure 5c, the grain structure of HAZ closely resembles that of the BM, albeit with a larger grain size, attributed to the stirring by the pin and sole exposure to welding heat.

The microstructure of the TMAZ, as depicted in Figure 5d, is characterized by curved elongated grains and fine equiaxed crystals. Dislocations were absorbed by some small-angle grain boundaries in the original grains during welding, leading to an increase in the grain boundary angle and the formation of new, fine, recrystallized grains. However, due to the low temperature and strain rate in the TMAZ, only a few small-angle grain boundaries were transformed into large-angle grain boundaries. Consequently, a large number of small-angle grain boundaries persist within the grains, resulting in their enlargement. When subjected to stirring, these grains become distorted and elongated.

In Figure 5e, it can be observed that the microstructure of the NZ1 consists of fine equiaxed crystals with an average grain size of approximately 8.9 μm, which is significantly refined compared to the base material. This refinement occurs because dislocations and substructures within the original grains are continuously generated and eliminated during the welding process [15,16]. The movement of dislocations leads to polygonization, resulting in the formation of sub-grains within the grains. When subjected to stirring, adjacent sub-grains rotate and gradually absorb lattice distortions, eventually evolving into large-angle grain boundaries. This process gives rise to fine equiaxed crystals, thereby predominantly featuring large-angle grain boundaries (denoted by black lines), with fewer small-angle grain boundaries (denoted by white lines). In Figure 5f, it is evident that the grain size in NZ2 is smaller than that in NZ1, averaging about 3.9 μm. This reduction is attributed to two complete dynamic recrystallizations. Additionally, during the welding of the 2S side, the NZ1 zone near the NZ2 zone experiences greater thermal cycling, leading to the growth of some fine equiaxed crystals.

The TEM results for the 2A97-T3 Al-Li alloy SSFSW T-joint are depicted in Figure 6. Figure 6a,b show that the BM primarily comprises four strengthening phases: T1 (Al₂CuLi), θ′ (AlCu₂), δ′ (Al₃Li), and β′ (Al₃Zr) precipitates [14,15]. The main strengthening phases, namely, the θ′ and T1 precipitates, predominantly exist in the form of thin sheets or slender rods. The δ′ and β′ phases exhibit a metastable and spherical morphology. Additionally, a few coarse δ equilibrium precipitates can be observed at the BM grain boundaries [17,18]. In Figure 6c, the precipitates in HAZ mainly comprise T1 and a small amount of θ′, both of which exhibit significant coarsening. The absence of δ′ and β′ precipitates suggests that the metastable phases have been dissolved into the matrix due to welding thermal cycling. Similarly, Figure 6d illustrates that the TMAZ also contains some T1 precipitates and fewer θ′ precipitates, with the metastable phase being dissolved into the matrix. However, compared to the HAZ, the distribution density of the strengthening phase in the TMAZ is lower, indicating an increased dissolution degree of the strengthening phases. Figure 6e,f reveal that only the coarsened T1 phase is present in NZ1, with the other strengthening phases being dissolved into the matrix. Conversely, there is almost no T1 precipitate in NZ2, a fact that can be attributed to the material undergoing two welding thermal cycles, which promote the dissolution of strengthening phases. The increased dislocation density in NZ2 is due to intensified plastic deformation that occurred during the secondary stirring of the pin. Phase identifications made using the diffraction pattern are also shown in Figure 6g–j. Overall, the results indicate that the dissolution degree of the strengthening phase increases when moving from the BM to NZ, particularly with a higher welding heat input, wherein the strengthened phases in NZ2 are nearly completely dissolved into the matrix.

3.3. Microhardness of the T-Joints

The microhardness of T-joints from 2010 and 2012 was tested with a scanning step of 0.5 mm, a loading force of 0.98 N, and a loading time of 15s. The test results were represented in hardness cloud maps, as depicted in Figure 7a and Figure 7b, respectively. It can be observed in the figures that the hardness value of the stringer is lower than that of the skin, and the low hardness zone is more extensive. This result can primarily be attributed to the fact that welding on both sides subjects this area to twice the number of high heat cycles, whereas the skin experiences only one high heat cycle, leading to more noticeable softening of the stringer. Additionally, a low-hardness area is also evident in NZ1 and the HAZ in the central of the skin due to the heat effect occurring during welding on the 2S side, resulting in decreased hardness. Comparison between Figure 7a,b reveals that the 2012 T-joint exhibits higher hardness values and a smaller softening area. This can be attributed to the faster welding speed and lower heat-per-unit-length weld, leading to reduced softening [19,20]. Hence, to mitigate softening of the T-joint while ensuring welding quality, a higher welding speed should be selected.

3.4. Tensile Properties of the T-Joint

The tensile test results for the 2A97-T3 base metal and T-joints at room temperature are presented in Figure 8. A column diagram and stress–strain curves are displayed in Figure 8a and Figure 8b, respectively. One can observe that the strength and elongation of the 2012 T-joint are higher, with tensile strength and yield strength reaching about 437.2 MPa and 322.1 MPa, respectively, both exceeding that of the base metal by 97.7%. A noticeable decrease in elongation of the T-joint compared to the base material was observed, which was attributed to non-uniform microstructure across different zones of the T-joint and a significant increase in T-joint thickness. These disparities in microstructure and structure result in uneven deformation during the stretching process, leading to particularly challenging deformation in the thickening area, thereby reducing elongation. Fracture analysis, as depicted in Figure 9a, revealed that the fracture of the T-joint occurred in the base material. Figure 9b,c are low-power and high-power SEM photos of the fracture, respectively, showing an abundance of dimples, a typical feature of ductile fracture. Consequently, the stationary shoulder friction-stir-welded T-joint demonstrates a high bearing capacity.

4. Discussion

This paper explores the welding of the T-joint of a thin-wall ribbed structure using the high-strength 2A97-T3 Al-Li alloy as the welding material. A stationary shoulder friction-stir-welding (SSFSW) tool and corresponding welding process were developed for this purpose. During welding, a gap was intentionally maintained between the stationary shoulder and the skin by controlling the displacement of the former relative to the latter. The rotating pin rapidly heated the materials of the skin, stringer, and wire in the stirring area to a plasticized state. Under the extrusion force exerted by the shoulder with an R3 rounded angle, the plasticized material is squeezed into the gap on both sides, thereby increasing the bearing area of the softening zone. The hardness and tensile test results indicate that an increase in the bearing area effectively compensates for the strength loss caused by softening and the uniform reinforcement of the thin-walled Al-Li alloy T-joint.

The results indicate that the heat input during SSFSW is lower than that in conventional FSW owing to the high-speed rotation of only the pin while the stationary shoulder remains static [21,22]. In the case of thin-walled T-joints, the pin size is restricted, resulting in a relatively low heat production capacity. To achieve superior joint quality, it is essential to choose process parameters characterized by high rotation speed and low welding speed. At high rotation speeds (2000–2500 rpm), achieving satisfactory T-joint quality is typically challenging when the welding speed is high (150 mm/min), whereas defect-free T-joints can be attained at welding speeds of 100~120 mm/min. This is primarily attributed to the reduction in welding speed, which increases heat input per unit length of the weld, facilitating complete plasticization and recrystallization of the material, thereby eliminating interface imperfections [23,24].

The hardness cloud map in Figure 9 and the tensile test results in Figure 7 reveal that the hardness and strength of the 2010 T-joint are both lower than those of the 2012 T-joint [25]. It is indicated that as welding speed decreases and heat input increases, the softening of the T-joint is further exacerbated. Consequently, under conditions of ensuring welding quality, a higher welding speed should be chosen to yield a T-joint with superior performance.

As depicted in Figure 9, softening was observed in the HAZ, the TMAZ, and NZ. The HAZ underwent high thermal cycling, while the TMAZ experienced some plastic deformation [2,23]. Grain growth was evident in both the HAZ and TMAZ, as illustrated in Figure 5. Figure 6 reveals the dissolution of the metastable phase in the HAZ, along with coarsening of the equilibrium phase, and a similar dissolution of the equilibrium phase in the TMAZ [26,27]. These alterations in microstructure and precipitated phases contributed to the attenuation of fine-crystal strengthening and precipitation strengthening. Welding on both sides of the T-joint resulted in the formation of an overlapping zone (NZ2) at the joint′s center. Material in this zone was subjected to stirring multiple times, underwent significant deformation, and experienced high thermal cycling. Recrystallization occurred again, leading to further grain refinement (refer to Figure 5f), which played a role in fine-crystal strengthening. However, as depicted in Figure 6f, the precipitated phase largely dissolved, and the precipitation-strengthening effect diminished considerably. When the 2S side was being welded, grains in NZ1 adjacent to NZ2 were reheated, causing the fine equiaxed crystals to grow, resulting in a grain size approximately 2.28 times that of the NZ2 zone. According to the Hall–Petch formula, this led to a significant weakening of the fine-grained strengthening effect. Furthermore, the precipitated phase in NZ1 dissolved, contributing to noticeable softening in the region.

The softening of high-strength aluminum–lithium alloy T-joints is unavoidable [28]. To mitigate the impact of softening on the bearing capacity of the T-joint, local reinforcement is provided through structural and process improvements in welding. For thin-walled T-joints, a wire-filled friction-stir-welding tool is designed to create a small gap between the stationary shoulder and the skin, as well as the stringer of the T-joint during welding. The material in the stir area reaches a plasticized state under the extrusion of the rounded corners on the stationary shoulder, and the plasticized material is subsequently filled into the gap [29,30]. This process locally thickens the softening zone of the T-joint. As depicted in Figure 10a, the thickness of the skin exhibits a gradual increase from Sections 1 to 5 of 2012-1S. In other words, the thickness gradually increases from the base material to NZ, with a faster rate of increase observed closer to the stringer. Figure 10b provides a magnified view of the hardness of the 1S side of the 2012 T-joint. By comparing Figure 10a,b, it is evident that the thickness of sections 1 to 5 is almost proportional to the hardness values. When combined with the tensile test results presented in this paper, it becomes apparent that fractures occur in the base material area, and the difference in strength between the T-joint and the base material is minimal. This finding proves that different regions of the T-joint possess nearly identical bearing capacities. Moreover, as shown in Figure 10c, a SSFSW joint with equal cross-sectional area was tested (the stringer of the T-joint was removed); a fracture occurred in the HAZ due to the softening effect. The tensile test result further verified the importance of increasing the bearing area of the softening zone for aluminum alloy SSFSW T-joints.

Therefore, in this study, focused on the Al-Li alloy T-joint of the thin-wall structure, we increased the bearing area of the softening zone and essentially eliminated the softening effect caused by welding through the design of a wire-filled stationary shoulder friction-stir-welding tool and process control. In other words, structural strengthening offsets softening, enabling the realization of equal-strength welding for the T-joint of high-strength aluminum–lithium alloy materials.

5. Conclusions

In this study, the test material utilized was a 2.8 mm thick 2A97-T3 aluminum–lithium alloy plate. A wire-filled stationary shoulder friction-stir-welding tool and process were developed to yield a T-joint with equal strength.

(1) Under the conditions of 2000 rpm and a welding speed ranging from 100 to 120 mm/min, a well-formed and partially thickened T-joint was achieved. Most of the grains in the TMAZ were bent and elongated. Complete recrystallization was observed in the NZ. The grain size of NZ1 was approximately 8.9 mm, while that of NZ2 located within the middle of the NZ was about 3.9 mm.

(2) The main strengthening phases (the T1 phase and θ′ phase) are coarse in the HAZ and dissolved in the TMAZ, resulting in the softening of the joint. Concurrently, grain refinement is evident in the NZ, leading to a fine-grained strengthening effect. However, almost all of the strengthening phase is dissolved. Due to the thermal effect of second-pass welding, the hardness value of the NZ and HAZ in the center of the skin further decreased, and the minimum hardness was approximately 70% that of the base material.

(3) The tensile strength and yield strength of the T-joint were measured at 437.2 MPa and 322 MPa, respectively, with a strength coefficient of 0.977. Fractures within the T-joint were observed to occur within the base material area, where ductile fracture characteristics were exhibited.

(4) The bearing area of the softening zone increased, which essentially mitigated the softening induced by the welding process, thereby allowing equal-strength welding of the high-strength aluminum–lithium alloy T-joint.