Investigations on improving the properties of laser beam welded thick dissimilar joints of steel and aluminum by using filler material

Lightweight designs represent a good approach not only for automotive engineering,¹ but also for shipbuilding^2,3 in order to reduce the CO₂ emissions as a result of lowering fuel consumption.^4,5 The lightweight design principle can be implemented, for example, by substituting materials with low strength with materials with higher strength or by using hybrid materials. In shipbuilding, material combinations of steel and aluminum alloys are used for various assemblies. Due to technical and economic aspects, a combination of strong steel and lightweight aluminum alloy is often of interest. For example, in yacht building the hull is made of steel and the superstructure of aluminum alloys,⁶ or in passenger ships the uppermost deck superstructure is made of aluminum and the deck floor of steel. In addition to the reduction of weight, a lowering of the center of gravity of the ship is achieved to stabilize the ship due to the intelligent arrangement of the different materials with their specific properties. Up to now, such components have been joined together by means of an explosion-welded adapter,^7,8 which manufacturing is time-consuming and costly. This adapter exhibits excellent deformation capability, but with the tradeoff of a lower transmittable force, so the adapters are used oversized.³ 

A good alternative for the production of adapters made of steel and aluminum is offered by laser beam welding, which is already established in shipbuilding for joining steel.⁹ Laser beam welding has specific advantages, such as high welding speed, low heat load, and local melting of the joint partners to counteract the challenges of fusion welding for such dissimilar joints. In addition to the different physical properties, the formation of intermetallic phases must be considered.¹⁰ In particular, the different expansion coefficients can lead to stress fields in the weld seam during the cooling.^11,12 Such stress fields may weaken the welded joint and can lead to failure without external loading. The formation of intermetallic phases, such as FeAl₂, Fe₂Al₅, and FeAl₃, is due to the limited solubility of the elements.^10,13–15 Such phases are hard and brittle and usually form in the transition zone between the weld metal and the aluminum base material.^16–18 These phases differ from those of the constituent metals in terms of lattice structure by increasing complexity. In addition to the formation of intermetallic phases, iron and aluminum combine in the form of disordered solid solutions and superstructures (ordered solid solutions).^13,14 Selected mechanical properties, such as hardness, elongation at fracture, and fracture toughness, of relevant superstructures and intermetallic phases are shown in Table I.^19–22 

As expected, the intermetallic phases are characterized by a complicated lattice structure and extremely stable non-metallic bonds, which prevent movements of dislocations and thus deformation (no elongation at break). Even the superstructures exhibit low elongation at fracture with a relatively low aluminum content. In Ref. 23, microstructure with an aluminum content of up to about A_C = 9.8 at. % is described as ductile compounds suitable for cold forming. For this reason, iron-aluminum compounds with an aluminum content of up to A_C = 6.5 at. % are used as a deep drawing sheet in car body construction.²⁴ 

Laser beam welding of this dissimilar joint was investigated not only in the thin sheet range,^25–27 but also in the thick sheet range.^28,29 Taking into account special framework conditions and welding configurations, it was possible to produce joints with comparatively high joint strength and low weld seam irregularities. A laser beam welded lap joint, in which the steel sheet is positioned on the aluminum sheet, has proven to be advantageous. The aim here is to minimize aluminum-rich phases by welding completely through the steel sheet combined with weld into the aluminum alloy. With the exception of the diffusion zone, a more iron-rich microstructure with relatively better properties results. This can be achieved by adjusting the welding depth into the aluminum material depending on the spot diameter used and the steel plate thickness.^25,26,30,31 Shear tensile tests were used to determine an optimum penetration depth range of about t_P = 0.5 mm^30,31 for the thin sheet range (∼1 mm) and an optimized penetration depth of about t_P = 1.5 mm for the thick sheet range (∼5–10 mm).^28,29 Thus, the thin-walled dissimilar joints achieved a max. shear tensile force of F_S = 5.2 kN³⁰ and the thick-walled dissimilar joints a max. shear tensile force of about F_S = 9 kN.^28,29 On the basis of fracture appearance, it could be determined that part of the aluminum base material was also broken out. Above this penetration depth range, the structural embrittlement and the increasing crack length lead to a reduction of the joint strength. The welded joint in the aluminum sheet is not only firmly bonded but also form-fitted; below this penetration depth range, there is a reduced mechanism of the form-fit force transmission and, thus, a reduced shear tensile force. The specimens fail by unbuttoning of the weld seam. The cross tension load is seen as a more critical load case since the effect of the form-fit does not come into account. The dissimilar joints are characterized by a much lower cross tensile force of about F_K = 0.44 ± 0.01 kN for thin-walled joints³² and of about F_K = 3.5 ± 1.22 kN²⁹ for the thick-walled joints at the optimized welding depth. In principle, due to the use of a thicker steel plate, a higher penetration depth and, thus, a higher joint strength can be achieved. However, the penetration depth is limited by microstructure embrittlement and the high crack occurrence so that above the optimized penetration depth the dissimilar joint fails in the weld metal.

Summing up, the positive effect of the penetration depth and the negative effect of microstructure embrittlement are opposed with regard to the metallurgical and mechanical properties of laser welded dissimilar joints of steel and aluminum alloy. Up to a penetration depth of about t_P = 1.5 mm in the aluminum sheet, thick-walled dissimilar joints can be produced without pronounced microstructure embrittlement. Based on these scientific findings, there is a high demand for a method with which a higher penetration depth can be achieved without structural embrittlement.

The aim of this study is to investigate a new, innovative method for influencing the mechanical and metallurgical properties of lap-welded dissimilar joints of steel S355 J2 (t = 5 mm) and aluminum alloy AA6082-T651 (t = 10 mm). This process is carried out by using iron-rich welding powder, which is filled into a groove made in the aluminum sheet. Subsequently, the steel sheet, the welding powder, and the aluminum alloy are welded together as one joint. The use of iron-rich powder may lead to a microstructure with a low aluminum content and, thus, to better mechanical and metallurgical weld seam properties. In the course of the welding process development, the influence of the welding powder on the process behavior, the weld geometry, the chemical composition of the weld metal, the crack length, and max. cross tension force is determined by varying the energy per unit length and groove depth. For evaluation and classification purposes, the dissimilar joints are compared with reference specimens made without the use of welding powder.

The laser welding processes were performed using a fiber laser beam source HighLight FL6000-ARM (COHERENT, Inc.) with adjustable laser beam intensity at a wavelength of λ = 1070 nm. The conversion is performed using two optical fibers arranged one inside the other. The inner optical fiber has a diameter of d_C = 70 μm, and the outer optical fiber has a diameter of d_R = 180 μm. The powers of the ring beam with an output power of P_C = 2 kW and of the core beam with an output power of P_R = 4 kW can be adjusted independently of each other. The laser beam source is characterized by a beam parameter product in the core of SPP_C = 2.5 mm mrad and in the ring of SPP_R = 10 mm mrad.

The laser processing head BEO D70 (TRUMPF Laser- und Systemtechnik GmbH) with an implemented DC scanner (ILV Ingenieurbüro für Lasertechnik und Verschleißschutz) was fixed to a three-axis portal unit system. For the optics, a focusing lens with a focal length of f_f = 560 mm and a collimating lens with a focal length of f_c = 200 mm were used, resulting in a theoretical spot diameter of d_SC = 196 μm for the core and d_SR = 504 μm (outer diameter) for the ring.

All welding processes were developed for dissimilar lap joints of the steel S355 J2 (t = 5 mm) as the top sheet and aluminum alloy AA6082-T651 (t = 10 mm) as the bottom sheet. As filler material, iron-rich welding powder is used. Table II shows the chemical composition of the steel, the aluminum alloy, and the welding powder. The average size range of the powder particles is from 63 up to 150 μm. The plates to be welded have a length of 200 mm and a width of 60 mm, which also corresponds to the overlap.

The one-sided welding process was carried out using a maximum laser beam power of P_L = 6 kW and a focal position on the top of the steel sheet. In order to realize a wobble width of 2.4 mm, a linear wobbling laser beam transverse to the welding direction was utilized with a frequency of 100 Hz. The reference specimens were welded without the use of groove and welding powder. For the welding process with welding powder, the aluminum alloy sheets were provided with a rectangular groove, which extends centrally and over the entire length of the sheet. The groove geometry was varied in terms of groove depth from t_g 1.5, 3.0, and 4.5 mm with a constant groove width of 2.0 mm. Within the scope of the process development, the energy per unit length E_L as a result of the laser beam power (P_L) and welding speed (v_W) was varied to obtain a complete coverage of the groove.

In order to determine the influence of the parameters on the quality of the dissimilar joints, metallographic analyses and tensile tests were used. In this context of the metallographic analyses, cross sections of the weld seam geometry were taken to manually determine the penetration depth and the total crack length in the weld seams. The penetration depth was taken from six samples for each parameter. The cross sections were investigated using a light microscope. For the evaluation of the cracks, three cross sections were used. The measured crack length was the summation of all cracks within the weld seam, as shown in Fig. 1.

To visualize the different microstructures of the weld metal, the cross sections were etched using nital with a nitric acid concentration of 3 %. Furthermore, EDX mappings were performed to determine the chemical composition of the microstructures.

The tensile tests were performed with a testing speed of 10 mm/min using a tensile testing machine (MTS Landmark, MTS Systems Corporation) (Fig. 2). The cross tensile test specimens have a length of 25 mm and a width of 60 mm. To clamp the specimens in the fixture, tapped holes were drilled orthogonally to the weld at a respective distance of 20 mm from the center of the weld. The holes were drilled on both the upper and the lower plates. A clamping device self-developed was used for the tests. It was important to ensure that the screws did not extend beyond the joint plane so as not to influence the max. cross tension force. Beside the max. cross tension force, the type and the position of the fracture is documented.

For the investigation of the new method, dissimilar joints without welding powder (reference specimens) and dissimilar joints with welding powder under variation of the energy per unit length are produced in the first step. To determine the influence of the welding powder, comparable penetration depths t_P of about 1.5, 3.0, and 4.5 mm must be achieved in both test series. Such penetration depths correspond to the groove depths investigated. Figure 3 shows the energy per unit length required for the different welding depths.

For both penetration depths, an almost linear correlation with determined energy per unit lengths can be observed. In this context, an increasing energy per unit length leads to increasing penetration depth. When welding dissimilar joints using welding powder, significantly higher energy per unit length is required for comparable penetration depths compared to the reference specimens. However, similar energies per unit length³³ are used for laser beam welding of similar joints of aluminum alloys and ferritic steel materials. The reason for higher energy per unit length required can be seen in the increase in the surface fraction due to the groove. Thereby, with increased surface fraction and thus increased heat conduction, the energy loss during laser beam welding increases. Such an energy loss is also influenced by the groove depth so that for lower penetration depth there is a smaller difference with regard to the energy per unit length. The extent to which the increased energy per unit length affects the weld geometry of dissimilar joints by means of welding powder will be investigated among other things in Sec. IV A.

In this section, the weld seam geometries of dissimilar joints produced by using welding powder as well as of the reference specimens are investigated depending on the penetration depth and the energy per unit length (Fig. 4).

With regard to the weld seam geometry, no striking differences can be observed between the two test series. Basically, the penetration geometry in the aluminum alloy is characterized by an increased penetration depth at the outer weld area compared to the weld center. This can be attributed to the dwell time at the turning points during laser beam wobbling transverse to the welding direction. Furthermore, the qualitatively determined roughness of the joints surface by means of welding powder is lower compared to the surface of the reference specimens. This qualitative determination was carried out along the weld metal based on the cross sections. Such a homogeneous weld top can be attributed to a stable process. The inhomogeneity may be caused by an increasing aluminum content and the associated reduction in melt viscosity. In addition, it is noticeable for the dissimilar joints by means of welding powder that the weld concavity increases with increasing groove depth. Such a relationship can be attributed to the filling density of the welding powder of 2.6 g/cm³. In this context, the volume of the filled groove decreases as the powder melts due to the loss of spacing between the particles and, thus, promotes the formation of the weld concavity. In percentage terms, the weld concavity remains approximately the same in relation to the groove depth.

Due to the chemical composition influences, the microstructure and, thus, also the weld quality in terms of imperfections and joint strengths, the chemical composition of the dissimilar joints is examined below, taking into account the use of welding powder and the penetration depth, cf. Fig. 5. In addition, a cross section is shown as an example for a penetration depth of t_P = 3 mm in each case.

As expected, the aluminum content in the weld metal of the reference specimens increases significantly with increasing welding depth because a higher amount of aluminum is melted. An almost linear relationship can be seen between the welding depth and the aluminum content. This also seems plausible, since the penetration depth also increases linearly with increasing energy per unit length, cf. Fig. 3. In contrast, the aluminum content of dissimilar joints welded by means of welding powder increases marginally with increasing penetration depth, since considerably less aluminum is melted. The difference in aluminum content between the reference specimens and the dissimilar joints produced by means of welding powder increases with increasing penetration depth. In general, a lower aluminum content can be observed in the upper region of the weld metal compared to the lower region, which can be attributed to the melt pool dynamics and the short melting time. In this case, the molten aluminum amount is not given enough time to distribute itself homogeneously over the entire weld metal area. Although a pronounced proportion of vertical molten pool flow is formed in this partial penetration compared with full penetration,^34,35 this is not sufficient for homogeneous distribution of the aluminum when the melting time is taken into account. The difference in the aluminum content in the upper and lower ranges of the dissimilar joints welded by means of welding powder is ΔA_C = 2.7 % (A_C from 1.5 to 4.2 wt. %) for a penetration depth of t_P = 1.5 mm, ΔA_C = 2.9 % (A_C from 4.1 to 7.0 wt. %) for a penetration depth of t_P = 3.0 mm and ΔA_C = 3.3 % (A_C from 5.0 to 8.3 wt. %) for a penetration depth of t_P = 4.5 mm. In this context, the difference increases with increasing welding depth due to the extended distance between the lower and upper measuring ranges and the generally low melting time.

In order to investigate the influence of welding powder and penetration depth on crack formation, the averaged crack length is plotted as a function of the penetration depth of dissimilar joints produced by using welding powder and of the reference specimens, cf. Fig. 6. In addition, the crack length of cross sections is shown as an example for a penetration depth of 3 mm in each case.

The crack length for the reference specimens increases significantly with increasing penetration depth, which can be attributed to an increased aluminum content (cf. Fig. 5) and the associated microstructure embrittlement. Such a relationship is consistent with the results in Refs. 28 and 29. At a penetration depth of t_P = 1.5 mm, the crack length is approximately C_L = 1.0 mm, which increases to an averaged crack length of C_L = 15 mm at a penetration depth of t_P = 3.0 mm and to a crack length of C_L = 25 mm at a penetration depth of t_P = 4.5 mm. Due to the relatively low aluminum content in the weld metal of the joints produced by means of welding powder, the averaged crack length is at a low level of about C_L = 1 mm. This corresponds to the crack length of the reference specimens at a penetration depth of about t_P = 1.5 mm. Consequently, the penetration depth can be increased without enhancing crack formation.

The cracks appear in the intermetallic phase area and in the weld metal for the reference specimens with a penetration depth of more than t_P = 3.0 mm, cf. Fig. 7 (right). In contrast, in the dissimilar joints produced by means of welding powder, individual small cracks form mostly in the intermetallic phase and, at a penetration depth of t_P = 4.5 mm, sporadically also in the joint plane at the edge of the dissimilar weld metal, cf. Fig. 7 (left, middle). The cracks are cold cracks or stress cracks.

By means of the cross tensile test, the influence of the welding powder on the max. cross tension force is investigated depending on the welding depth, cf. Fig. 8. Furthermore, the fracture appearances and the bending of the sheet of the aluminum alloy at a penetration depth of t_P = 4.5 mm measured by a confocal microscope are represented. The bending of the sample was determined at the bottom of the sheet of the aluminum alloy and half of this result is presented due to the symmetry.

The reference specimens show a strength level of F_C = 2.2 kN at penetration depths t_P of about 1.5 and 3.0 mm. A higher welding depth of about t_P = 4.5 mm and a strong reduction of the max. cross tension force can be observed. In some cases, these specimens fail at the clamping process in the tensile testing machine. In this case, the failure mode changes from unbuttoning of the weld seam to fracture in the weld metal. On the one hand, the microstructure embrittlement, caused by the increased aluminum content and, on the other hand, the cracks lead to such an early failure. This behavior is reflected in previous strength studies in Refs. 28 and 29. An almost unchanged fracture behavior is shown when testing the dissimilar joints produced by means of welding powder. With the exception of one specimen, the specimens fail at a penetration depth of t_P = 4.5 mm due to a button-out fracture. Up to a penetration depth of 3.0 mm, the max. cross tension force increases to a value of F_C = 7.2 kN and then decreases again. Although the dissimilar joints at a penetration depth of t_P = 3.0 and 4.5 mm differ slightly or not at all with respect to the aluminum content and thus also the microstructure as well as crack length, such a decrease in strength occurs. One reason for this behavior is the formation of small cracks, which are located in the joint plane and act as crack initiators. Another reason is the greater bending of the aluminum sheet due to decreasing axial section modulus at a welding depth of t_P = 4.5 mm compared with a welding depth of t_P = 3.0 mm. Here, at the higher weld depth, the axial section modulus is characterized more by the weld than by the aluminum alloy below, which is softer. This results in the bending of the bottom joining partner.

In shipbuilding, thick-walled joints of steel and aluminum alloys are used with the aim of reducing weight and, thus, also fuel consumption and lowering the center of gravity. Until now, the joining of such dissimilar joints has been standardized using complex and costly explosion welding. There is a high demand from shipyards for better alternatives. Due to its advantages, laser beam welding could be a good alternative for joining these dissimilar joints. However, in laser beam welding, depending on the penetration depth, unfavorable microstructures are formed in lap joints, which have an effect on the imperfections and joint strengths. Up to a penetration depth of about t_P = 1.5 mm, the joint strength increases. If the welding depth is increased further, the joint strength decreases due to microstructure embrittlement. For this reason, there is a great demand for a method with which a higher welding depth can be achieved without microstructure embrittlement.

The aim of this study is to investigate a new, innovative method for influencing the mechanical and metallurgical properties of laser beam welded dissimilar lap joints of the upper steel S355 J2 (t = 5 mm) and the lower aluminum alloy AA6082-T651 (t = 10 mm). The goal is to increase the penetration depth without microstructure embrittlement by minimizing the aluminum content of the weld metal in order to improve the properties of dissimilar joints. This is realized by using iron-rich welding powder, which is filled into a groove made in the aluminum sheet. Varying the energy per unit length and the groove depth t_g of 1.5, 3.0, and 4.5 mm, the influence of the welding powder on the process regime, the weld geometry, the chemical composition of the weld metal, the averaged crack length, and the max. cross tension force is investigated. The joints are compared with reference specimens made without the use of groove and welding powder.

In the case of laser beam welding of dissimilar joints using welding powder, significantly higher energy per unit length is required for comparable penetration depths compared with the reference specimens, due to the increase in the surface fraction caused by the groove. In this context, the dissimilar joints produced by means of welding powder exhibit qualitatively lower roughness compared with the reference specimens, which is due to the higher viscosity caused by the reduced aluminum content in the molten pool. The roughness increases for the reference specimens with increasing penetration depth. In this context, the aluminum content in the weld metal increases. In contrast, the aluminum content of dissimilar joints welded by means of welding powder increases marginally with increasing penetration depth, since considerably less aluminum is melted. For example, the aluminum content A_C of dissimilar joints with a groove depth of t_g = 3.0 mm is in the range between 4 and 7 wt. % and of reference specimens in the range between 17 and 22 wt. %. The low aluminum content in the weld metal has a beneficial effect on the formation of cracks. Regardless of the welding depth, the crack length is at a low value of C_L = 1 mm. In the reference specimens, the crack length increases linearly due to microstructure embrittlement up to a value of about C_L = 25 mm at a penetration depth of t_P = 4.5 mm. This also affects the joint strength. The highest max. cross tension force of the reference specimens is F_C = 2.2 kN at a penetration depth t_P of 1.5 and 3.0 mm. In contrast, the max. cross tension force of joints produced by means of welding powder increases up to a welding depth of t_P = 3.0 mm to a max. cross tension force of F_C = 7.2 kN. With the help of the new method, the quality of thick-walled dissimilar joints can be significantly improved, which further increases the application potential of laser beam welding in shipbuilding.

The work presented was carried out within the scope of the German joint research project FOLAMI “Laser beam welding of form fit dissimilar joints of steel and aluminum for serviceable semi-finished products in shipbuilding” (“Formschlüssiges Laserstrahlschweißen der Mischverbindung aus Stahl und Aluminium für betriebsfeste Halbzeuge im Schiffbau”), subproject “Laser beam welding of form-fit steel-aluminum dissimilar joints by means of welding depth control” (“Laserstrahlschweißen von formschlüssigen Stahl-Aluminium-Mischverbindungen mittels Einschweißtiefenregelung”). This project (Reference No. 03SX547B) was funded by the German Federal Ministry for Economic Affairs and Climate Action (BMWK), within the framework of the funding line “Next generation maritime technologies” (“Maritime Technologien der nächsten Generation”) of the German Federal Government and supervised by the Project Management Jülich (PtJ), which is gratefully acknowledged. Furthermore, the authors wish to thank the project partners and the project committee member companies as well as their representatives for supporting the project and for their good cooperation.

The authors have no conflicts to disclose.

Rabi Lahdo: Conceptualization (lead); Funding acquisition (lead); Investigation (lead); Resources (lead); Validation (lead); Writing – original draft (lead). Sarah Nothdurft: Investigation (equal); Visualization (equal); Writing – review & editing (equal). Oliver Seffer: Conceptualization (equal); Investigation (equal); Visualization (equal); Writing – review & editing (equal). Jörg Hermsdorf: Project administration (equal); Supervision (equal); Writing – review & editing (equal). Stefan Kaierle: Supervision (lead); Writing – review & editing (equal).