Microstructure and Mechanical Properties of Laser Welded Al–Mg–Si Alloy Joints
2019 Volume 60 Issue 2 Pages 230-236
Details
2019 Volume 60 Issue 2 Pages 230-236
In this paper, Al–Mg–Si alloy with the thickness of 2.5 mm was overlap welded by fiber laser. Optical microscopy, scanning electron microscopy, energy dispersive spectroscopy, hardness tester, and electro-hydraulic servo testing machine were used for the microstructure observation and mechanical property examination. Specimens extracted from the welded plates were tested at room temperature for the determination of the tensile and fatigue properties of the welded joints. Results showed the columnar grains were formed in the fusion boundary, while equiaxed grains were formed in the fusion zone; the lowest hardness in the fusion zone is 72.0 HV, gradually increasing from the heat-affected zone to the BM. Tensile shear strength of the welded joint is 96.0 Mpa, which is about 25% of the tensile strength of base metal. The fatigue limit at 1 × 106 conditional cycles is 22.5 MPa. The fractography showed intergranular and quasi-cleavage fracture. Bending deformation is related to the magnitude of the force. As the stress lower, the deformation become more severe.
Fig. 7 SEM images of fatigue fracture of WJ (a) rock candy pattern, (b) cleavage terrace and river pattern, (c) and (d) crack propagation pattern at different magnification.
6013 aluminum alloy is a new type of aluminum alloy developed in the early 1990s by Alcoa, one of the world’s largest aluminum producers. It is an Al–Mg–Si series high strength aluminum alloy with good plasticity and forming properties. This type of aluminum alloy adds and controls the contents of Cu and Mn, making its strength higher than that of the 6xxx series aluminum alloy, and the corrosion resistance better than that of the 7xxx series aluminum alloy. Its initial application was aimed at the automotive industry, improving fuel efficiency by reducing the weight of parts. At present, it is processed into various extrusions such as thin plates, thick plates, wire rods, rods, etc., widely used in aerospace, rail transportation, and other industrial fields.
The preferred welding processes for aluminum alloys are frequently gas metal arc welding (GMAW) and gas tungsten arc welding (GTAW), due to their comparatively easier applicability and better economy. Compared to these welding methods, laser welding has many advantages to be quite suitable for welding of aluminum alloy, such as high energy density, high welding speed and efficiency, good weld appearance, and narrow welding heat affected zone.
Many studies on laser beam welding (LBW) of aluminum alloy have been performed recently. Pakdila M. et al.1) obtained high quality joints in laser beam welding of 6056 Al-alloy plate using AlSi12 filler wire with some acceptable amount of porosity. No liquation cracking was observed and the grain boundary liquation becomes more pronounced as the amount of Si along the grain boundaries increased. J. Enz et al.2) surveyed the theoretical fundamentals of laser weldability of metals and identified relevant thermophysical parameters. By using an Yb fibre laser with a large initial beam diameter, a top-hat beam profile and a high laser power it was possible to obtain a good weld quality. No hot cracking was observed even though no filler wire was used for welding. Hekmatjou H. et al.3) employed a 700 W pulsed Nd:YAG laser to weld 5456 aluminum alloy plates. The weld penetration and tendency of liquation cracking in the heat affected zone and solidification cracking in the weld metal are investigated. Hot cracking tendency can be reduced by increasing the laser average power to 300 W and hot cracking susceptibility can be reduced with increasing pulse frequency to 20 Hz. In the study of G.L. Qin et al.,4) 1.8 mm-6013 aluminum alloy plate was welded by activating flux CO2 laser welding in order to increase the absorption of laser energy and improve the weld appearance in laser welding of aluminum alloy.
Although having poor mechanical properties, the overlap joint has the advantages of simple preparation before welding and assembly, and no need for opening groove, which is used for the connection of non-bearing parts, such as car bodies. F. Fadaeifard et al.5) study the effect of rotational speed on macro and microstructures, hardness, lap shear performance and failure mode of friction stir lap welding on AA6061-T6 aluminum alloy with 5 mm in thickness. E. Salari et al.6) carried out the research on the influence of tool geometry and rotational speed on mechanical properties and defect formation in friction stir lap welded 5456 aluminum alloy sheets. M.R. Pishevar et al.7) study the influences of friction stir welding parameters on microstructural and mechanical properties of AA5456 at different lap joint thicknesses. The fatigue behavior of resistance spot welding in aluminum 6061-T6 alloy was experimentally investigated by R.S. Florea et al.8) The work revealed that the welding process parameters have a great influence in the microstructure and fatigue life and different fatigue failure modes were observed at several load ranges and ratios for a constant frequency and three welding currents. In the study of D. Afshari et al.9) an electro-thermal-structural-coupled finite element model and x-ray diffraction residual stress measurements have been utilized to analyze distribution of residual stresses in an aluminum alloy 6061-T6 resistance spot-welded joint with 2-mm-thickness sheet. The results have a good agreement with experimental data obtained from x-ray diffraction residual stress measurements.
Aluminum alloy lap welding is a very common method in the field of automobile manufacturing. While most of the existing researches for overlap welding are focused on friction stir welding and resistance spot welding. Less researches on laser overlap welding of aluminum alloy were carried out. Compared with other welding method, the aluminum alloy has fast welding speed, narrow weld seam, small heat affected zone, and the reduced lap side seam, which can reduce the weight of the car body. The lap forms of aluminum alloy plate commonly used in the car body are overlap and double-lap. The main research of this paper is overlap form. In the present study, 6013 aluminum alloy with the thickness of 2.5 mm was overlap welded by fiber laser. The microstructure was observed and mechanical properties (hardness, tensile and fatigue) were tested and fracture behaviour of the joint were studied. The results, which were of great practical value, provides reference for further study of overlap laser welding.
6013 aluminum alloy in T6 state with a thickness of 2.5 mm was used as the base material, and the filler metal was ER5356 welding wire with a diameter of 1.2 mm, whose chemical compositions were listed in Table 1. The welding experiments were carried out by an IPG YLS-5000 fiber laser machine controlled by an KUKA industrial robot. Diagram of the experimental process and testing specimen were presented in Fig. 1(a). The plates with the size of 300 mm × 125 mm × 2.5 mm were assembled into an overlap joint with 25 mm overlap distance. Before welding, SiC paper was used to get rid of alumina coating and acetone was used to remove surface oil contaminant. The laser power is 6 KW, and the welding rate and wire feeding speed is 1.75 m/min and 4 m/min respectively. Throughout the experiments, argon gas with a flow rate of 20 L/min was applied to protect the molten zone.
Schematic diagram of (a) laser welding process, (b) testing specimens.
After welding, the cross-section of the samples were polished and etched by keller’s reagent (HF:HCl:HNO3 = 1:1.5:2.5), and observed by optical microscope of Nikon EPIPHOT300 and VHX-6000 digital microscope. Then scanning electron microscope (SEM, Hitachi S-3400N) and energy dispersive spectroscopy (EDS) were used for characterizations. Vickers hardness values were measured both from the upper and under sheet by Microhardness Tester HV-1000. A 0.981N force was applied for 15 s with the adjacent indention distance of 0.5 mm. To evaluate the mechanical strength of the joints, tensile and fatigue testing were performed by Zwick/Roell Amsler HB250 electro-hydraulic servo testing machine. The fatigue cycling pattern of tensile loading was sinusoidal wave. The loading frequency was 20 Hz and stress ratio R = 0.1. The environmental temperature was 299 K and the relative humidity was 55%. The statistic method of fatigue S-N curve followed the standard of ISO 12107:2003. And then the fracture surfaces were observed. Tensile testing samples of the welded joint (WJ) and the method were prepared according to the National Standard GB2561-88. In order to reduce the extra bending stress during the tensile process, two 2.5 mm thick guides were placed on both sides of the sample, as is shown in Fig. 1(b). In addition, the base metal (BM) were also cut into the same length and width as a reference of the WJ.
The microstructure of the welded joint is shown in Fig. 2. After a series of chemical metallurgical reactions, the temperature of molten pool metal rapidly degrades as the heat source moves away, the weld formed, and a solid phase transition occurs during the cooling process. Different microstructure changes will occur in the Fusion Zone (FZ) and Heat-affected Zone (HAZ) under the influence of welding heat source. Figure 2(a) display the FZ consisted of slender columnar crystals with Si, Cu and Mg rich precipitates at the grain boundaries near the fusion line. The reason for the formation of this columnar-dendritic microstructure in the FZ is the high solidification rates involved in power beam welding. In the HAZ near the fusion line, coarsen grains were formed because of the high degree of influence by welding heat. Moreover, the region of coarsen grains in the HAZ is very narrow, due to the lower heat input during welding. It can be seen that the fusion zone at the edge of the weld is very narrow, and the microstructure of the entire weld exhibits a typical as-cast microstructure. Nearby the fusion line is a kind of mixed alloy formed by the wire and the base metal.10) The base grain near the FZ is used as a ready-made surface, and the molten pool metal grows up from the base metal near the fusion line. The columnar grains begin to grow toward the center of the weld at the edge of the weld. The temperature gradient decreases when growing to the center of the weld, and the equiaxed crystal structure having no apparent crystal orientation is formed. As shown in Fig. 2(b), in the FZ center, equiaxed grains, formed by the high degree of supercooling, and some broken of them can be seen. Additionally, there also exits some small cellular dendritic crystals.
Microstructure of the welded joint (a) HAZ and FZ near the fusion line, (b) Fusion center, (c) Pores near the overlap gap.
For the overlap joint, there inevitably exists gap between two plates microscopically, as is shown in Fig. 2(c), where can become the passage through which the gas escaped. Due to the weld high rate solidification, the bubbles are unable to escape from the weld pool completely and stay in the weld, and eventually form pores. The main defect in the welded joint of aluminum alloy is pores. The presence of pores reduces the effective bearing area of the welded joint, resulting in stress concentration, thus reducing the strength, and the dense pores on the surface of the weld is easy to become the fatigue source of fatigue cracking.
3.2 HardnessThe microhardness obtained from two different lines are given in Fig. 3. The line 1 was 1.5 mm from the upper surface and the line 2 was 1.5 mm from the under surface. Owing to the higher cooling rates, a narrow HAZ (about 2 mm) has been formed. The hardness observed in the HAZ and the FZ is lower than that of BM (about 125 HV) owing to the undermatching strength and disappearance of strengthening effect during the thermal cycle. The laser welding has the characteristics of rapid heating and rapid cooling. First, the welding wire and the original base material are melted, and the second phase particles are dissolved into the α-Al matrix to form a supersaturated α-Al solid solution, which is rapidly cooled by the weld. The characteristic suppresses the precipitation of the second phase, and the final weld is mainly a supersaturated α-Al solid solution.11) Due to the absence of strengthening phase of the weld, the weld strength will be seriously reduced and become a weak link of the welded joint.
Microhardness profile of the welded joint.
The microhardness of the entire HAZ shows a stepwise increase. Take the right part of the Line 1 as an example, in the HAZ near the FZ (A zone), along the direction away from the weld, the hardness value increases rapidly with increasing distance. In the middles of the HAZ (B zone), the hardness value fluctuates. Although the increase trend is significantly reduced, it is still higher than the A zone. In the HAZ close to the BM (C zone), the microhardness value increases rapidly with the increase of distance, until the transition to the BM gradually become steady. The microhardness of the HAZ increases stepwise, mainly because the welding thermal cycle is different at different locations in the heat-affected zone. Therefore, there is a difference in the behavior of dissolution and aggregation of reinforcing elements such as Mg or Si or second phase strengthening particles. The peak temperature in zone A is high, and the second phase particles are largely dissolved in the matrix. The extremely fast cooling rate in this zone inhibits the precipitation of the second phase particles and solid solution strengthening. Since the main strengthening method of the 6xxx series aluminum alloy is aging strengthening, the solid solution strengthening effect is weaker than the aging strengthening, resulting in a decrease in the hardness of the area. The peak temperature and cooling rate of zone B are lower than that of zone A. The dissolution of the second phase strengthening particles is insufficient, and the quenching effect is not obvious. It can be considered that solid solution strengthening does not occur in this zone. The second phase particles mainly grow and aggregate, that is, the aging effect occurs in the region, and this overaging effect seriously slows down the increase of the microhardness value of the region. The C zone is close to the base metal and is less affected by the welding thermal cycle. It can be considered that the second phase particles do not substantially dissolve, and the tendency of aggregation growth gradually decreases, and the hardness loss becomes smaller and smaller up to the BM and reaches a stable value.
The microhardness of FZ in the lower plate (79.4 HV) is higher than that in the upper plate (72.0 HV). The reasons for it can be explained as following: As we all known, the hardness of 6xxx series aluminum is much higher than that of the 5xxx. While welding, there are many base materials melted under the molten pool, and the top of the molten pool is mainly the welding wire. Meanwhile, for the Al–Mg–Si alloys, the major strengthening element is mainly Mg and Si. However, Mg would be seriously burnt during the welding process because of its low melting point (650°C).12) In the case of high-power laser welding, the burning of Mg on the weld surface is the most serious. Two typical points were chosen to conduct the EDX analysis, as shown in Fig. 4. In the FZ center of the upper plate, the mass% of the Mg element is 1.67%, while in the FZ center of the under plate, the Mg element is 1.42%. It can be found that the position away from the surface of the weld, the loss of Mg is reduced due to the attenuation of laser beam energy. The remaining strengthening elements or the second phase are all dissolved into the matrix structure, causing a further decrease in the hardness of the weld zone.
Energy dispersive spectrometer analysis of the weld.
The distribution of the element Mg in the weld is also related to the flow state of the liquid metal in the weld pool during the welding process. Laser welding is characterized by the existence of small hole effect. During the welding process, the laser is irradiated to the front wall of the small hole to melt the material. The molten material flows behind the small holes driven by the gasification recoil force and the steam flow force to form the trailing edge of the molten pool. The molten pool is elongated in the opposite direction of the welding rate, the temperature gradient around the small holes is uneven, and the temperature gradient of the small hole front is larger than the temperature gradient of the small hole trailing edge. According to the Mrangoni effect, the surface temperature gradient will cause a surface tension gradient. The molten metal will flow from a place with low surface tension to a place with high surface tension. Therefore, the liquid metal around the small hole continuously flows to both sides of the molten pool, and the liquid metal at the leading edge of the small hole continuously flows to the trailing edge of the small hole. During the welding process, the flow of liquid metal on the surface of the molten pool and the inside constantly stirs the molten pool, causing the material at the bottom of the molten pool to be continuously carried to the upper part of the molten pool.13)
In addition, the temperature gradient of the upper plate is smaller than that of the under during the welding process, leading to the grains coarsening. All of the three factors have come together to make the hardness increases in the direction of penetration.
3.3 Tensile testThe stress-strain curves fitted by the tensile data are shown Fig. 5. Each set of tensile tests including WJ and BM were performed three times in order to reduce the error. All tensile shear specimens extracted from the joint failed in the weld seam because of its characteristic in linking up and the formation of pores. The average tensile strength of the WJ is 96.0 Mpa, which is about 25% of the BM (390.8 MPa). The max lap shear strength of 6061-T6 aluminum alloy welded by FSW was about to 130∼140 Mpa, according to the study carried by F. Fadaeifard et al.14)
Engineering stress vs engineering strain curves.
On the other hand, the results show the WJ has significant losses in elongation (0.8%) compared to BM (12.4%), owing to strain concentration in the lower strength weld zone. The fracture surface of BM is at about 50° angle to the tensile direction and the fracture position has obvious necking, which is corresponded to its elongation. It is well known that the plasticity of the 5xxx series aluminum alloy is better than that of the 6xxx series aluminum alloy. However, due to the low tensile strength of the weld metal and the existence of pores, locally generated stress concentration and large deformation, and the crack initiated and propagated before the other part of the specimen occurred deformation. The uniform deformation did not proceed completely, resulting in the elongation of the WJ being lower than that of the BM.
3.4 Fatigue testIt is generally believed that when the cycle number of steel structure up to 1 × 107, the fatigue limit obtained can be used as the fatigue strength of that material. However, non-ferrous metal materials such as aluminum alloys do not have a true horizontal section. After 1 × 107 cycles, the curve is still not completely horizontal. Some very high-cycle fatigue (1 × 109 cycles) tests15,16) show that as the tests continue to increase, the fatigue limit of the sample is still decreasing, while the reduction is not obvious. Therefore, the usual fatigue limit is based 1 × 107 cycles. In engineering, 1 × 106 is also used as the base of the fatigue limit, which is called conditioned fatigue limit. In this test, the fatigue limit is based on 1 × 106 cycles.
At present, for the welded structure, the stress amplitude σa and the number of cycles N are generally used to indicate the fatigue life curve.
| \begin{equation} \sigma_{a}{}^{k}*N = C \end{equation} | (1) |
| \begin{equation} \mathrm{k}*\log\sigma_{a} + \log N = \log C \end{equation} | (2) |
Therefore, in logarithmic coordinates log σa and log N show a linear relationship. The S–N curve, with abscissa log N and ordinate log σa, is illustrated in Fig. 6. According to eq. (2), the fitting formulae of the curves for the WJ and BM are calculated respectively:
| \begin{equation} -0.17099*\log\sigma_{a} + \log N = 2.0952 \end{equation} | (3) |
| \begin{equation} -0.13975*\log\sigma_{a} + \log N = 2.8405 \end{equation} | (4) |
The S-N curves of BM and WJ.
It can be found that the K’s absolute value of the base material (0.13975) is slightly smaller than that of the WJ (0.17099), which indicates that the fatigue life of WJ drops down faster than BM with the increase of cycle stress. The conditional fatigue limits of the base metal and the welded joints are 200 MPa and 22.5 MPa respectively, when the conditional cycles are 1 × 106. X. Xu et al.17) investigated fatigue properties of friction stir lap-welded joints of AA6061-T6 alloy. The fatigue strength at 2 × 106 cycles is 20.13 MPa and 15.65 MPa when R = 0.1 and R = 0.5.
Detailed fatigue fracture morphologies were observed by using SEM. As shown in Fig. 7(b), crack was initiated at the pore and other defects near the surface. There exists cleavage terrace and river pattern, which shows the characteristics of cleavage fracture.
SEM images of fatigue fracture of WJ (a) rock candy pattern, (b) cleavage terrace and river pattern, (c) and (d) crack propagation pattern at different magnification.
In the crack propagation stage, as is shown in Fig. 7(c) and Fig. 7(d), dimples and cleavage terrace can be seen. Secondary cracks can also be observed, which had a negative influence on fatigue properties. Dimples is the main microscopic feature of ductile fracture. Synthesize all features of the fracture, it can be concluded that it is the mixed fractures dominated by brittle mode.
As shown in Fig. 7(a), rock candy pattern was visible with a strong three-dimensional effect. The grain boundary can be obviously seen, indicating cracks extend along grain boundaries, which is the typical intergranular fracture. A.K. Lakshminarayanan18) also observed the feature in 6061 GMAW joints. This may be caused by the combined influence of coarsen grains and a higher amount of precipitate formation at the grain boundaries. The detailed reason remains further study.
3.5 Fracture mechanismFrom Fig. 5 and Fig. 6, it can be found that the WJ were not occurred bending deformation under the increasing tensile force; while, under the fatigue force, that were occurred, and the lower the amplitude stress, the more severe the bending deformation. The mechanisms of these characteristics are explained as follows.
The loading method of the WJ and the force situation near the weld seam (WS) are illustrated in Fig. 8(a) and Fig. 8(b). In the overlap joint, the surface of the two plates contacted with each other by the weld seam, thus the load transfer through the WS, and the WS was subjected to a pair of shearing forces as well as a pair of counterforces. According to the previous analysis, the BM appears to have higher strength and hardness, hence generating a pressure to the WS. Therefore, the under plate may generate an additional bending moment preferentially, resulting in bending downwards slightly, as is illustrated in Fig. 8(c). Under completely idealized condition, the upper plate generate an upwards bending force correspondently, leading to the situation demonstrated in Fig. 8(d). After then, the location of the WS near the two surfaces subjected to more intensive concentrated stress, giving rise to the crack initiation and propagation and then fracture. The centerline model with the shear force and additional bending moment is illustrated in Fig. 8(e).
The mechanisms of bending and fracture (a) loading method, (b) force situation, (c) additional bending and tensile stress, (d) deformation form ideally, (e) centerline model.
However, there exists pores in the WS, as is shown in Fig. 2(c). As the force increasing continuously, the stress concentration is being more and more seriously. When reaching critical value, many a crack initiate and then converge together, giving rise to fracture before its deformation. That’s why bending deformation was not observed during tensile test.
During fatigue test, the force was decreased before reaching the condition of the crack initiation and then increased again. In the course of the cycle of the stress, slow plastic deformation happened. As the deformation up to an extent, the stress reached the critical value of crack initiation because of stress concentration and existence of pores. And then the cracks propagate till the joint break, as is illustrated in Fig. 9(a).
Fracture mechanism of (a) high stress cycle fatigue, (b) low stress cycle fatigue.
The mechanism of fatigue fail under low stress cycle is illustrated in Fig. 9(b). When the stress is much lower, the crack cannot initiate, the WJ will continue to deform. Based on the previous analysis, the crack first appeared in the under plate, but due to the low stress and blocked by the second phase particles, the cracks initiation could not continue. As the deformation going on, the stress concentration aggravates, leading to the cracks continue to propagate. Meanwhile, critical condition of crack initiation reached on the other side, resulting in the cracks appeared on both upper and under plates.
From the above analysis, we can conclude that, during the continuous effect of force, bending deformation is related to the magnitude of the force. When the force is great, due to the existence of pores, many microcracks generate and work together to make the WJ fracture rapidly before the macroscopic deformation occur. When the force is under the critical value, the crack will not propagate and the deformation will occur, and the lower the force, the more severe the deformation.
In the present study, 6013-T6 Al-alloys were overlap welded by laser welding method. The microstructure was observed by OM, the compositions of the FZ were analyzed by EDX, mechanical testing including hardness, tensile and fatigue of the welded joints were studied, and the fracture surface were observed by SEM. The main conclusions are summarized as follows:
