1 Introduction

Nowadays, the aluminium/steel joining occurs typically in automotive industry in case of car bodies, for example energy absorption elements are frequently made from aluminium/steel structure. The structural elements should be joined and these joints can be made by mechanical joining methods like self-piercing riveting [1,2,3] or clinching [4,5,6]; furthermore, the welding [7,8,9] can be used for this aim, too. In case of self-piercing riveting the tensile-shear or cross-tension static tests show appropriate joint strength quality [1, 10], but the rivet used makes the procedure more expensive. Clinching process is not expensive, since no additives are needed, but the joint strength is characteristically weaker. In some cases, the mechanical joining methods are supplemented by adhesive bonding [11], which is a relatively expensive technology; the welding can be more cost effective. In case of welding, basically the resistance spot welding (RSW) and the ultrasonic welding (UW) can be applied for hybrid spot joining [12,13,14]. In recent years special process variants appeared to improve the joint properties, like resistance element welding (REW) [15], metallic bump assisted RSW (MBaRSW) [16], (high power) ultrasonic welding ((HP)USW) [17], ultrasonic, and resistance spot welding combination [18]. Micro-RSW (MRSW) [19] and vaporizing foil actuator welding (WFAW) [20] were also investigated as prospective solutions for aluminium/steel joining. The special joining methods require additional materials and/or process steps, or the process cycle is longer in comparison with the original RSW. Table 1 summarises different joining processes used for aluminium/steel joining and their main characteristics.

Table 1 Joining technologies for aluminium/steel spot joining
Full size table

Some investigation [8, 10] highlighted that the RSW is just rarely used for aluminium/steel sheet joining because the presence of brittle intermetallic compound (IMC) and necessary to remove the oxide layer from the aluminium part before welding. Furthermore, the welding task is challenging because of the absolutely different material properties (melting point, thermal conductivity, electrical resistance, strength). The joint strength can be improved with removed oxide layer, but the oxide removing during manufacturing cannot be efficient and cheap enough [36]. The forming IMC layer is very brittle, and the joint properties are influenced by IMC thickness and containing phase type. A thin IMC layer (thinner than 10 µm) has no significant effect on the strength and ductility of the joints, but the thick IMC layer can be harmful applying laser welding [37]. In case of RSW, different values of the optimal IMC layer thickness were concluded in the joint interface. Zhang et al. recommended maximum 5 µm [38]; Chen et al. reported 3 µm [29]; and Miyamoto determined it as 2 µm [39]. For this aim, it is generally necessary to keep the process temperature and duration as low and short as possible, because the formation of the IMC layers requires atomic diffusion. Several studies were performed to identify the significant IMC phases and their effects on the joint properties. Sundman et al. [40] reported that the possible IMCs at room temperature are Al2Fe, Al5Fe2, and Al3Fe4. Kouadri-David et al. [41] analysed the IMC regions and found that the critical region of an IMC is the aluminium side, which contains more Al but there is a significant Fe fraction, too. This is because the solubility of Al in Fe is greater than the solubility of Fe in Al, and the Al-rich IMC phases are brittle, while the Fe-rich phases are relatively ductile. Therefore, the joint properties can be improved if the formations of Al rich brittle IMCs are minimised. Zhang et al. [38] found that the main IMC phases are Fe2Al5 on the steel side and FeAl3 or Fe4Al13 on the aluminium side, where the previous one is thicker than the last one. The morphology and thickness of the IMC layer depend on the process parameters and on the distance from the centre of the joint; the thickness is generally higher at the joint centre and lower on the boundary [30]. Fe2Al5 was identified as the dominant phase in the IMC layer formed at the joint interface. The Fe2Al5 grew; it evolved from randomly oriented fine grains to coarse grains with preferred direction, perpendicular to the aluminium/steel interface. Numerical simulation was performed; it was found that the majority of Fe2Al5 growth occurred when the temperature between the sheets was over 900 ℃. The growth of Fe2Al5 reduces the joint strength and can cause interfacial fracture. Cracks within Fe2Al5 mainly consisted of primary cracks propagating through the interface and secondary cracks approaching the steel side, while the primary cracks became the dominant ones with increased IMC thickness [42].

According to literature and our previous investigations [36, 43], under quasi-static loading conditions the strength properties of aluminium/steel RSW joints can be comparable with aluminium/aluminium joints. Table 2 shows tensile-shear test results performed on different aluminium/steel combinations compared with the tensile-shear strengths of aluminium/aluminium similar joint.

Table 2 Results of tensile-shear tests in case of aluminium/steel and aluminium/aluminium RSW joints
Full size table

The table shows that there is no significant difference between similar aluminium and hybrid aluminium/steel joints based on the results of tensile-shear strength. However, the typical loading condition of a car body is the cyclic loading, and the aluminium/steel hybrid joints can show different behaviours under high cycle fatigue. In case of quasi-static loading, the thickness of IMC layer has influence on the mechanical behaviour of the joint; consequently under cyclic loading condition, the effect of IMC can be important, too. The IMC properties were investigated widely and connected to the quasi-static-type mechanical test results and their failure modes. The behaviour of aluminium/steel hybrid joints under cyclic loading is rarely investigated [27, 28, 31, 48]. The results have shown that the behaviour of aluminium/steel RSW joints under cyclic loading condition was better than aluminium/aluminium joints [27, 28], applying special electrode for the welding process. At the same time, most of the previous investigations were carried out using different aluminium and steel plate thicknesses and those were not covered the fatigue limit range of the S–N curves in none of the cases. The latter is also important because steels can be characterised by a definite fatigue limit, whereas aluminium alloys cannot. Furthermore, the comparisons were completed for a minority of the investigations; they were not covered the full steel/steel, aluminium/steel, and aluminium/aluminium variations. Therefore, the basic aims of our research (in medium term) are to compare aluminium/steel joining technologies and to optimise the technologies based on different criteria. The direct objective of the research and this article (in short term) is to give more and quantitative information about the aluminium/steel RSW joint properties under cyclic loading condition (high cycle fatigue).

2 Materials and technologies

DP600 steel was chosen as the steel part; it is frequently used in the automotive industry and has relatively low strength between dual-phase (DP) steels. DP steels contain hard martensite islands embedded in a ferrite matrix with a dispersed distribution. The steel side was combined with 5754-H22 aluminium alloy (strain hardened) as the first combination and with 6082-T6 aluminium alloy (heat treated) as the second combination. The 5754-H22 aluminium sheet is one of the commonly used material qualities, mainly due to its good formability and associated appropriate strength properties. Its main alloying element is Mg, its strength increased by forming, and then softened to a quarter of hardness. The weldability of this alloy is good among the aluminium alloys. The 6082-T6 base material is one of the most widely used heat-treatable aluminium alloys; it has lower formability and higher strength. It contains mainly Mg and Si alloying elements which are causing ageing. This alloy is more sensitive for welding, softening in the heat-affected zone, and hot cracking can easily happen because welding. For the experiments both similar (steel/steel, aluminium alloy/aluminium alloy) and hybrid (steel/aluminium alloy) joints were produced. Each base material had 1 mm thickness for better comparability.

Tables 3 and 4 show the chemical compositions; furthermore, Table 5 summarises the most important mechanical properties (tensile strength (Rm), yield strength (Rp0.2), and elongation (A50)) of the base materials according to quality certificates.

Table 3 Chemical compositions of the applied aluminium alloys, based on quality certificates, weight %
Full size table
Table 4 Chemical composition of the applied DP steel, based on quality certificate, weight %
Full size table
Table 5 Basic mechanical properties of base materials, based on quality certificates
Full size table

A TECNA 8007-type welding machine (AC, 50 Hz) with TE 550-type control system was used for welding; the welding force was ensured by a pneumatic cylinder. The weld control mode was constant current mode. The electrode material was CuCrZr; according to RWMA, it is in the class 2. The welding electrodes were used with a 5-mm spherical pin diameter. The radius of the spherical pin was R = 50 mm, according to the recommendation of the literature [51, 52]. The same geometry was used for the upper and lower electrodes (see Fig. 1).

Fig. 1
figure 1

The geometry of the upper and lower welding electrodes

Full size image

The used parameter combinations are previously optimised based on tensile-shear test results, Table 6 summarises the chosen parameters and resulted weld nugget/joint diameters. These parameters gave the thinnest IMC layers, lower heat input causes lack of joining, and higher heat input results thicker IMC. In case of 5754-H22/DP600 joints, 10% higher current was needed than 6082-T6/DP600 joints.

Table 6 Parameter combinations for RSW and resulted weld nugget/joint diameters
Full size table

3 Testing circumstances

The high cycle fatigue (HCF) tests were carried out according to the guidelines of the relevant standard [53]. The geometry and the dimensions of the test specimens for all base material combinations with 1 mm sheet thickness can be seen in Fig. 2. The specimen parts were cut from 1000 mm × 2000 mm dimension sheets into 100 × 30 mm dimensions strips. The investigated RSW joint located at the centre of the 30 mm overlapped area.

Fig. 2
figure 2

HCF test sample geometry and dimensions

Full size image

HCF tests were carried out by MTS 322.41 electro-hydraulic universal material testing equipment and MTS FlexTest 40 controller. Sinusoidal loading wave form was used, the applied load ratio (Fmin/Fmax) was R = 0.1 with f = 30 Hz frequency. During the HCF tests in all cases, several load levels were applied at room temperature and in laboratory air. Considering that the tests were evaluated according to [54, 55], the load levels were chosen as described in that document. Displacement limit detector was not used, so the cycles counted until complete failure. Figure 3 shows the test specimen gripping in the test system, using both a schematic and a real image.

Fig. 3
figure 3

HCF test specimen gripping in the test system and a specimen after the fatigue test

Full size image

4 Results and discussions

The number of cycles to failure values was recorded after HCF tests. The results of fatigue tests were plotted by load range (∆L) to a number of cycles to failure (N) using a logarithmic scale for the number of cycles in ∆L-N curve. Figure 4 shows the ∆L-N curves of DP600 and 5754-H22 similar and hybrid joint combinations, where the arrows indicate surviving specimens. The straight lines were determined using the least squares method (lifetime section) and calculating the mean values (endurance limit); therefore, they belong to the 50% probability. The figure clearly shows that the DP600 similar joints have the best result as it is predicted; the endurance limit was 1282.5 N. The 5754-H22 similar joints result the worst value; the endurance limit was 648 N, which is nearly the half of the steel. The 5754-H22/DP600 hybrid joints show significantly better result than the aluminium alloy, with 939 N endurance limit value.

Fig. 4
figure 4

∆L-N curves for DP600/DP600, 5754-H22/DP600, and 5754-H22/5754-H22 joints

Full size image

Figure 5 shows the result of HCF tests in case of DP600 and 6082-T6 similar and hybrid joints. The nature and trend of the results are the same as those shown in Fig. 4 for DP600 and 5754-H22 materials. In this case the endurance limit of hybrid joints is 625.5 N, which is significantly lower than DP600 steel joints (1282.5 N); furthermore, the 6082-T6 similar joints gave the worst result, the endurance limit was 513 N. The difference between the steel joints and the hybrid joints was larger than for the 5754-H22 alloy; in other words, the difference between the hybrid joint and the aluminium joint was smaller than for the 5754-H22 alloy.

Fig. 5
figure 5

∆L-N curves for DP600/DP600, 6082-T6/DP600, and 6082-T6/6082-T6 joints

Full size image

Figure 6 shows all HCF test results compared with literature data [27]. The high cycle fatigue resistance of 5754-H22/DP600 hybrid joints are significantly better than 6082-T6/DP600 hybrid joints and competitive with the result from the literature. According to literature data [27], the endurance limit is approximately same in case of 6022-T4/IF steel hybrid and 6022-T4/6022-T4 similar joints, and shows slight difference between our result (6082-T6/DP600, 5754-H22/5754-H22). It should be noted that in [27], article fatigue limit values were not specified and wider (38 mm instead of 30 mm) and thicker (1.2 mm aluminium alloy and 3 mm IF steel instead of 1 mm) specimens were used.

Fig. 6
figure 6

All high cycle fatigue results compared with literature data [27]

Full size image

Table 7 summarises the characteristics of the ΔL-N curves belonging to 50% probability, using

Table 7 Characteristics of the ΔL-N curves belonging to 50% probability
Full size table
$$\Delta L=A*ln\left(N\right)+B$$
(1)

Basquin-type [56] equation. The data in the literature are consistent with our test results, and it can be concluded that the 5754-H22/DP600 joints show a better fatigue limit than the 5754-H22/DP500 joints [23]. It should be noted that the 5754-H22/DP500 joints were made with REW process (Q235 steel rivet in aluminium side) and thicker steel plate (1.5 mm instead of 1 mm).

Failure modes after HCF showed different and not typical types in some cases. Tables 8, 9, 10, and 11 summarise these types in the failure mode column. In some cases, the base material fracture was combined with shearing. In these cases, it can be assumed that the crack started in the upper side of the spot from the heat-affected zone, then growing to the base material, but it cannot propagate fully because shearing happens in the joint line. The other failure mode in case of hybrid joint is the clear shearing, which is not typical in case of steel/steel or aluminium/aluminium joints. The 5754-H22/5754-H22, 6082-T6/6082-T6, and DP600/DP600 joints showed the same failure mode, base material fracture. Figure 7 shows the macrostructure of failure modes in all cases.

Table 8 Load range, number of cycles to failure values, and failure modes in case of DP600/DP600 similar joint specimens
Full size table
Table 9 Load range, number of cycles to failure values, and failure modes in case of 5754-H22/5754-H22 similar joint specimens
Full size table
Table 10 Load range, number of cycles to failure values, and failure modes in case of 6082-T6/6082-T6 similar joint specimens
Full size table
Table 11 Load range, number of cycles to failure values, and failure modes in case of 6082-T6/DP600 hybrid joint specimens
Full size table
Fig. 7
figure 7

Failure modes of RSW joints: a full plugging (aluminium/steel), b base material partial fracture + shearing (aluminium/steel), c base material fracture (aluminium/steel), d shearing (aluminium/steel), e base material fracture (steel/steel), and f base metal fracture (aluminium/aluminium)

Full size image

Tables 8, 9, and 10 show the numerical data of the results and failure modes for all specimens in case of DP600/DP600, 5754-H22/5754-H22, and 6082-T6/6082-T6 joints, respectively.

Tables 11 and 12 show the numerical results of HCF tests in case of hybrid joints including the failure modes.

Table 12 Load range, number of cycles to failure values, and failure modes in case of 5754-H22/DP600 hybrid joint specimens
Full size table

In hybrid joint cases, the failure modes were different. Partial or full plugging happens in case of low number of cycles, however, mostly base material fracture with or without shearing, and shearing happens in higher number of cycles.

After HCF tests selected cross-sections of resistance spot welded joints were examined, focusing on different failure modes. Based on the observed failure modes (see Table 11 and 12), cross-sections of sheared/base material fracture and plugged fracture types were selected. In case of sheared/base material fracture type, the IMC layer was identified on the steel part, and cannot be found on the aluminium part, in line with previous experiences [41]. Thus, in this case, the IMC layer thickness can be measured and it assumed as the full IMC thickness. In case of plugged and base material fracture types, the IMC thicknesses were easily measured between the two base materials. Figures 8 and 9 show the result of the measurements on 5754-H22/DP600 and 6082-T6/ DP600 joints, respectively. In the background cross-section of plugged type can be seen.

Fig. 8
figure 8

IMC thicknesses in the joint line in case of 5754-H22/DP600 plugged and sheared specimens

Full size image
Fig. 9
figure 9

IMC thicknesses in the joint line in case of 6082-T6/DP600 plugged and sheared specimens

Full size image

In case of 5754-H22/DP600 plugged failure mode, the IMC layer thickness is much lower in the whole joint line than the sheared one; the IMC thicknesses are between 0.85 and 1.32 µm. The thickness of the sheared type is between 1.1 and 2.26 µm. It means that the IMC thickness of plugged type is bigger and more equal than in case of the sheared type. The 5754-H22/DP600 plugged cross-section shows that the outer part of the joint was sheared, and the bigger inner part is plugged. By the IMC layer thickness measurement, it can be clearly identified that the IMC thickness in the inner part is bigger like in the outer part. The IMC thickness on the plugged part is more than 1.1 µm, just a small increasing happens (1.32 µm) in the centre of the joint. In this specimen the thinner IMC shows shearing.

In case of 6082-T6/DP600, the plugged joint has significantly bigger IMC layer thicknesses in the whole joint line, and the sheared one shows thinner IMC. The IMC thickness of plugged type was between 1.2 and 2.92 µm, and the thickness of IMC in case of sheared one was between 0.55 and 1.89 µm. The IMC thicknesses show big differences through the joint line in both cases. The 6082-T6/DP600 plugged cross-section shows a little bit smaller plug diameter, and the outer part is also sheared here. The IMC thickness of the sheared part is less than 1.5 µm. In the plugged part, the IMC thickness is increasing from 1.5 to 2.92 µm. In this specimen the thinner IMC shows shearing, too, like in case of 5754/DP600 plugged case, but the IMC thickness is bigger.

5 Conclusions

Based on the executed and evaluated investigations and their results, the following conclusions can be drawn.

  • The applied joining technologies gave comparable HCF test results on similar and hybrid RSW joints of DP600, 5754-H22, and 6082-T6 base materials.

  • The endurance limit of 5754-H22/DP600 hybrid joints was higher (939 N) than 5754-H22/5754-H22 similar joints (648 N), but lower than DP600/DP600 similar joints (1285.5 N). The endurance limit of 6082-T6/DP600 hybrid joints was higher (625.5 N) than 6082-T6/6082-T6 similar joints (513 N), but lower than DP600/DP600 similar joints (1285.5 N). The tendency was similar to that for 5754-H22 aluminium alloy, but the difference between aluminium and steel similar joints was larger (772.5 N vs. 637.5 N), while the difference between aluminium and hybrid joints was smaller (112.5 N vs. 291 N). No relevant, directly comparable HCF test data were found in the literature.

  • Only base material fracture was experienced for all similar welded joints; however, different failure modes were observed for hybrid joints. Base material fracture and shearing after partial base material fracture were the typical failure modes in case of 5754-H22/DP600 and 6082-T6/DP600 hybrid joints, respectively. Furthermore, in some cases, full and partial plugging as a failure modes were appeared for hybrid joints.

  • IMC layer thicknesses were measured in case of plugging, as well as of partial base material fracture and shearing failure modes. Thicker IMC layer was measured in case of sheared 5754-H22/DP600 joint than the plugged one, and the IMC thickness was bigger in joint centre than the outer part. In addition, smaller differences were measured between the inner and outer parts of plugged joint. Opposite results can be detected for 6082-T6/DP600 joint. The IMC layer thickness was thinner in case of sheared one, and thicker in case of plugged one. Furthermore, there are significant differences in IMC thicknesses between the inner and outer parts of the joint.

  • In case of plugged failure mode, it is clearly visible that the plug diameter is smaller than the IMC layer diameter. The 5754-H22/DP600-plugged specimens show well-definable IMC thickness where the plug exists (1.1 µm); however, for 6082-T6/DP600, this could not be determined perfectly.

  • In order to compare hybrid joints, further investigations are needed:

    • joints made with other technologies (e.g. clinching) should be performed;

    • HCF tests should be executed and evaluated;

    • fatigue crack growth (FCG) test should be prepared, executed, and evaluated;

    • both different technologies and cyclic behaviours should be compared and optimal application areas should be specified.