Effect of intermetallic compound on the corrosion behaviour of resistance spot welding joints between 5182 aluminium alloy and galvanized DP780 dual-phase steel

Interfacial intermetallic compound (IMC) layer is critical during realizing the galvanic corrosion mechanism and strength degradation of aluminium/steel resistance spot welding (RSW) joints. The effect of IMC layer on the corrosion behaviour of RSW joint between 5182 aluminium alloy and galvanized DP780 dual-phase steel was investigated by immersion corrosion method and electrochemical method. Results demonstrated localized corrosion around Fe-rich phase particles on aluminium nugget and preferential corrosion at the interface front near IMC layer. The IMC layer had the highest open circuit potential value of −0.569 V and low corrosion current density among the investigated nugget and substrates. The potential difference (0.184 V) between the cathodic IMC layer and large area of anodic aluminium nugget was responsible for the preferential initiation of localized corrosion at the interface front near the IMC layer.


Introduction
Lightweight aluminium alloy and advanced high-strength steel has been more and more attractive for mixedmaterial body structures in the automotive industry, to satisfy the fuel economy improvement and carbon emission reduction. The usage of those dissimilar materials creates a challenge in establishing reliable joining techniques, due to the huge different physical and metallurgical properties [1,2]. During the various developing joining techniques, resistance spot welding (RSW) is still an attractive solution due to its flexibility, robustness and high efficiency [3]. However, the electrochemical corrosion resistance is significantly different between aluminium alloy and steel. Aluminium/steel weld joints are prone to galvanic corrosion, and thus accelerate the corrosion process of the joints [4]. The corrosion problem directly affects joint strength and fatigue life [5]. Therefore, the corrosion performance of aluminum/steel RSW joints must be investigated.
For the corrosion of aluminium/steel RSW joints, Pan et al [6] found that severe pitting corrosion attacked on the coupled aluminium zone, while less corrosion developed on the coupled steel region under the protectiveness of zinc coating. Maddela et al [7] found that the corrosion resistance of aluminium/steel RSW joints is comparable to or even better than that of self-pierce riveted (SPR) joints. They demonstrated that Zn coatings are more susceptible to galvanic corrosion than aluminium alloy and bare steel [7]. The effect of the galvanized coating layer on enhancing corrosion resistance of aluminium/steel joints were also reported by Lim et al [8] and Wang et al [9]. Joo et al [10] showed that the reduction of residual nugget diameter after corrosion is responsible for the reduction in shear strength of AA6061-T6 aluminium alloy/galvanized steel RSW joints. Structural adhesives dramatically reduced galvanic corrosion between aluminium alloy and steel substrates [7,10,11]. Joo et al [12] further indicated that the long-term mechanical reliability of the aluminium/steel RSW Any further distribution of this work must maintain attribution to the author(s) and the title of the work, journal citation and DOI. joints in a corrosive environment is strongly dependent on the bond strength. Including the metal substrates and the Zn coatings, the interfacial intermetallic compound (IMC) layer are essential for bonding between aluminium and steel substrate in RSW joint [2,7,10]. The IMC layer tend to form at the interface of dissimilar metals via atomic diffusion during welding [13][14][15]. The thickness of the IMC layer strongly affects the strength and fracture modes of aluminium/steel RSW joints. Some researchers extensively introduced the concept of critical IMC thickness [16][17][18]. Spontenously, the role of IMC layer during the corrosion of aluminium/steel RSW joint is interesting.
However, there are few studies on the effect of IMC layer on corrosion behaviour of aluminium/steel RSW joints. In other welding methods, nevertheless, several studies have reported the effect of Fe-Al IMC on the corrosion behaviour. For example, Sravanthi et al [19] showed that higher the thickness of IMC layer, more severe corrosion of 6061 aluminium/mild steel weld joints. Shi et al [20] found that the corrosion initiated at the steel substrate near IMC layer in 1060 aluminium/galvanized steel weld-brazing joint. IMC layer even increased dissolution of the adjacent weld seam material [20]. Huang et al [21] yet drew a different conclusion, that the corrosion of 5A06 aluminium alloy/ST04Z galvanized steel laser brazed joint started at local interface defects between IMC and aluminium substrate. Ma et al [22] found that the potential difference between the IMC layer and the 304 stainless steel led to IMC corrosion in friction welding joint. Lei et al [4] also found that the IMC layer in the AA6022-T4/Galvanized DC03 cold metal transfer (CMT) joint would corrode. They indicated that the corrosion of IMC layer resulted in a nearly 38% decrease in weld strength of the joint [4]. All the previous researches indicate unique corrosion behavior at/near interfacial IMC layer in aluminum/steel joints, but controversial results have been observed in different weld joints.
Up to now, no specific study on the role of IMC layer has been investigated during the corrosion of aluminium/steel RSW joints. Particularly, the intrinsic corrosion property of IMC layer has not been successfully realized by general corrosion test. In the present research, 5182 aluminum alloy to DP780 dualphase steel RSW joints was chosen to examine the effect of IMC layer. The RSW joints were prepared by using a novel method with Multi-Ring Domed (MRD) electrode [23]. The corrosion behaviour at different crosssectional locations in the joints was investigated by immersion test. Furthermore, the IMC layer without residual nugget substrate were particularly prepared and ensured from tensile fractured samples. The electrochemical corrosion behaviours of the IMC layer and the substrates were successfully investigated. Thereafter, the effect of IMC layer on the corrosion of cross-sectional samples were discussed based on the measured electrochemical properties.

Materials and methods
2.1. Materials and sample preparation 1.2 mm-thick 5182 aluminium alloy and 1.5 mm-thick DP780 dual-phase steel were chosen to prepare dissimilar RWS joints. The DP780 dual-phase steel was coated by hot dipped galvanized. Chemical compositions (in wt%) of these materials were measured using x-ray photoelectron spectroscopy (XPS) and listed in table 1.
Aluminium alloy sheet and steel sheet were resistance spot welded together after figure 1 using a 220 kVA medium frequency direct current (MFDC) pedestal type resistance spot welder at 1000 Hz. The same electrodes (MRD electrode [23]) were used on both sides of aluminium alloy and steel sheets. The cooling rate of water was controlled at 2 gallons per minute. The RSW process includes a preheating and a welding stage. In the preheating stage, the current sloped up to 6 kA and held constant for 40 ms, and the welding stage lasted for 1090 ms at 11 kA welding current. A constant holding force of 4000 N was applied on both welding stages, and the squeeze time was 1500 ms. The welded tensile shear specimens were put in an oven to bake 30 min at 175°C.

Corrosion tests
Immersion corrosion test and electrochemical corrosion test were conducted. The aluminium/steel RSW joints were cut across the weld centre for immersion corrosion test. Cross-section of the cut specimen was mounted by using resin, ground with sand paper to 2000# and ultrasonic cleaned. The specimens were then placed in 5 wt% NaCl solution for 2 h, 6 h, 12 h, 24 h and 48 h at 25 ± 1°C, respectively. After each immersion period, the samples with surface corrosion products were carefully rinsed by deionized water to clear the NaCl. Different regions of the RSW joints was focus for electrochemical corrosion, namely 5182 base metal, aluminium weld nugget, IMC, DP780 base metal, heat affect zone (HAZ) of steel, and galvanized coating of DP780. Figure 2 shows the cross-section locations of the regions. The samples at those regions were prepared from fracture pieces, after quasi-static tensile tests of the RSW joints on a uniaxial testing machine (SUST CMT5205) at a displacement rate of 3 mm min −1 . Particularly, the IMC specimens were prepared by slightly grinding the weld fracture surface on the steel side and the IMC were ensured by the energy dispersive spectrometer (EDS) measurement (figure 3).
Electrochemical tests, including open circuit potential (OCP), electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization, were conducted in 5 wt% NaCl solution using a three-electrode system   and PARSTST 2273 Electrochemical System at room temperature. An Ag/AgCl(sat. KCl) electrode was used as the reference electrode and a pure Pt-mesh was used as counter electrode. The samples at different regions of the RSW joints were used as work electrode. The OCP was measured for 1800 s. The EIS test was carried out in a frequency range from 10 mHz to 100 kHz with an amplitude of 10 mV. The potentiodynamic polarization started from the potential of 0.25 V negative than the OCP toward positive direction at a scan speed of 0.5 mV s −1 . The tested area of base metal of 5182, base metal of DP780, and galvanized coating of DP780 specimens was 1 cm 2 . The tested area of aluminium weld nugget, IMC, and steel HAZ specimens was 0.25 cm 2 due to the 7.2 ∼ 7.6 mm weld nugget diameter [24].
Optical microscope (OM, Zeiss Axiocam ERc 5 s) was used to characterize the macroscopic cross-section of spot welds and corrosion morphology. The cross-sections were chemically etched with 10% NaOH solution for 5182 aluminium alloy to observe metallographic microstructure, as well as 2% Nital solution for DP780 steel. The IMC thickness were measured using OM. Scanning electron microscope (SEM, Hitachi S-3400N) was used to characterize the microstructure of IMC, second-phase particle in aluminium alloy and corrosion morphologies. EDS was used for elemental measurements in second-phase particle and corrosion products, etc Electron probe x-ray micro-analyser (EPMA, Hitachi JXA-8530F) was employed to examine element distribution around interfacial IMC layer.

Results and discussion
3.1. Microstructure analysis Figure 4 exhibits the cross-sectional metallographic microstructure of the RSW joint. Many second phase particles present in 5182 aluminium alloy ( figure 4(b)). The particles consisted of 66.45 wt% Al, 26 wt% Fe, 4.08 wt% Mn, 1.8 wt% Mg , 1.5 wt% Cu and 0.17 wt% Si by EDS, which were probably Al 6 (Mn,Fe) phase [25]. The centre of aluminium weld nugget was composed of columnar grain and grain boundary particles (figure 4(c)). EDS results indicated that the grain boundary consisted of 88.48 wt% Al, 6.24 wt% Fe, 0.59 wt% Mn, 4.56 wt% Mg and 0.13 wt% Si. The composition at internal grain is 94.58 wt% Al and 5.42 wt% Mg. Figure 5 shows the elements distribution at the interface of the weld, figure 5(c) indicates the aggregation of Fe element at the grain boundary in the aluminium nugget. The Mg element was more concentrated inside the grain than at the grain boundary ( figure 5(d)), and both Mn and Si were distributed in very small amounts in the aluminium nugget   [26][27][28]. The thickness distribution of IMC layers at the aluminium/steel interface is shown in figure 6, the thickness of IMC layers at the centre and periphery of the weld nugget were 4.5 μm and 2 μm, respectively. Figure 4(e) exhibits the microstructure of DP780 dual-phase steel, consisting of ferrite with diffusely distributed martensite. And martensite was characterized at the HAZ of DP780 steel sheet ( figure 4(f)). Figure 4(g) shows the uniform galvanized coating on steel surface.   Figure 7 shows the corrosion morphology of the cross-sectional RSW samples after immersion for 12 h and 48 h, respectively. A large amount of corrosion product generated at the steel edges, and significant corrosion occurred at the aluminium nugget, as shown in figures 7(a) and (b). In the aluminium alloy substrate, corrosion around the second phase particles Al 6 (Mn,Fe) is obverse (figures 7(c) and (d)). Such localized corrosion around the particles might be due to the electrochemical micro-galvanic corrosion, where the potential of Fe-rich phase particles was more noble than that of aluminium substrate, resulting in the accelerated corrosion of the nearby substrate [4,29]. In the aluminium nugget region, similar localized corrosion around the small particles was evidenced due to the micro-galvanic corrosion effect at the boundary of the columnar grains, as shown in figures 7(e) and (f). The corrosion of the galvanized coating was significant and a large amount of flocculated corrosion products covered on the steel substrate (figures 7(g) and (h)). EDS results indicated that the flocculated products consisted of 48.58 wt% Zn, 34.46 wt% O and 16.96 wt% Cl. Slight corrosion was observed on the HAZ of the steel (figures 7(i) and (j)). Figure 8 shows the SEM corrosion morphologies at the interfacial IMC layer of nugget region in the RSW joint. Within short immersion of 2 h, obvious corrosion is present along interface front of aluminium nugget close to IMC layer, while slight corrosion in aluminium nugget at regions relatively far from the IMC layer ( figure 8(b)). As prolonging the immersion period, the corrosion at the interface front increased more significant than that in aluminium nugget far from the interface (figures 8(c)-(e)). After long time corrosion of 48 h, dramatic corrosion developed on both interface front and aluminium nugget, and all surface of aluminium nugget was damaged ( figure 8(f)). The steel nugget also corroded with increasing immersion time. Interestingly, the interfacial IMC layer seemed not corrode in the whole immersion process. Figure 9 compared of the localized corrosion morphologies at the periphery and the centre of the nugget interface regions. Similar localized corrosion preferentially initiated at the interface front and around the second phase particle in both periphery and centre region of aluminium nugget (figures 9(a) and (c)). Also severe corrosion in aluminium nugget and no significant corrosion at interfacial IMC layer have been evidence at both periphery and centre of the weld zone after immersion of 48 h (figures 9(b) and (d)). As the thickness of IMC layers at the periphery and centre of the weld nugget were 2 μm and 4.5 μm (figure 6), respectively, it can be inferred that the thickness of the IMC layer had no significant effect on the corrosion behaviour at the nugget interface region.

Polarization measurements
The potentiodynamic polarization curves of the different zones of aluminium/steel RSW joint are shown in figure 11. All polarization curves are characterized by active corrosion and thus Tafel extrapolation method was used to calculate the corrosion potential and current densities. The calculated values are listed in table 2. The sequence of E corr at different zones is determined as IMC > steel HAZ > base metal of DP780 > base metal of 5182 > weld nugget of aluminium > galvanized coating of DP780, indicating that the smallest and greatest thermodynamic trend of corrosion at IMC and galvanized coating of DP780 in the RSW joint, respectively [9]. Moreover, the corrosion potential sequence from the polarization curves is accordance with the OCP sequence, indicating the possibility of galvanic corrosion between those different zones. The sequence of corrosion current density i corr on different zones is determined as galvanized coating of DP780 > steel HAZ > IMC > base metal of DP780 > weld nugget of aluminium > base metal of 5182, indicating that the corrosion rate of galvanized coating of DP780 was the fastest in the joint.  Figure 12 shows the impedance diagrams of different zones of the joint. The Nyquist plot of base metal of DP780, and steel HAZ were both composed of a capacitive loop ( figure 12(b)) and one crest with one time constant from the Bode plot ( figure 12(d)), respectively. The Nyquist plot of galvanized coating of DP780 revealed one capacitive loop followed by a straight line Warburg impedance in the low frequency ( figure 12(b)). Nonetheless, the Nyquist plot of IMC, base metal and weld nugget of aluminium alloy reveal a capacitive loop and an inductive loop at lower frequencies (figures 12(a) and (b)). The low frequency inductive loop is attributed to relaxation process of the absorbed species on the electrode surface or the formation of the corrosion pits [30].  Additionally, the greater the radius of the capacitive loop, the more impedance and the smaller the corrosion current density, and hence higher corrosion resistance [29,31,32]. According to the size of the capacitive loops radius, the sequence of corrosion resistance of different zones is determined as base metal of 5182 > weld nugget of aluminium > base metal of DP780 > IMC > steel HAZ > galvanized coating of DP780.

EIS measurements
In order to explain the corrosion mechanism, the equivalent circuit models for different zones are shown in figure 13. Figure 13(a) displays the equivalent circuit model for base metal of DP780 and steel HAZ. Film circuit and Warburg impedance was employed in the equivalent circuit model ( figure 13(b)) to explore the role of surface deposited production due to high corrosion rate of galvanized Zn coating. In figure 13(c), an equivalent circuit model with inductive component was used to fit the localized corrosion round the second phase particles in base metal and weld nugget of 5182 aluminium alloy. Figure 13(d) displays the equivalent circuit model for IMC. The fitting results are listed in table 3. R s refers to the solution resistance. R ct and CPE dl refer to the charge transfer resistance and the double-layer capacitance at the electrolyte/substrate surface. R f and CPE f represent the resistance and the capacity of the corrosion products film. Y is the admittance magnitude of the CPE constant and n is the deviation parameter from 0 to 1. L and R L represent the inductance and the inductance resistance, and W is Warburg impedance.
The smallest capacitive loop ( figure 12) and reaction resistance on galvanized coating of DP780 (table 3) indicate the lowest general corrosion resistance [33,34], which accord with the lowest OCP value and greatest corrosion current (figures 10 and11). The high value of the reaction resistance and the existence of inductance resistance on base metal and weld nugget of aluminium ( figure 12 and table 3) imply the high resistance capability to general corrosion but slight resistance capability to localized corrosion [34][35][36]. Such result meshes with the corrosion morphologies at base metal and weld nugget of aluminium alloy ( figure 7). The presence of inductive curve on IMC layer ( figure 12(b)) also means the corrosion and possibility of localized corrosion without considering its interaction to aluminium alloy. When the galvanic effect was involved, the IMC layer with more noble potential (figures 10 and 11) might be protected from corrosion due to cathodic polarization in the galvanic couple. It is noted that the test area of IMC layer for OCP and polarization (figure 3) is much larger than the actual cross-sectional area of IMC layer during immersion test ( figure 7). The cathodic IMC layer in tiny  area could further immune to corrosion under the protection of anodic aluminium nugget with relative huge area in the couples. This assumption is true that no obvious corrosion on corrosion-sectional IMC even after immersion of 48 h ( figure 8(f)). Moreover, the acceleration effect on aluminium nugget was restricted at the interface front due to tiny cathodic area of IMC to large anodic area of aluminium of the galvanic couple. Spontaneously, the localized corrosion preferentially initiated at the interface front ( figure 8(b) and figure 9).

Conclusions
(1)The corrosion morphology of cross-sectional RSW weld joints between 5182 aluminium alloy and DP780 steel after immersion test in 5 wt% NaCl shows, that sever localized corrosion in aluminium nugget occurred   around Fe-rich phase particles and at the interface front near IMC layer, but slight corrosion was observed on IMC layer.
(2)From the OCP and potentiodynamic polarization test, highest OCP value of −0.569 V and corrosion potential of −0.603 V on IMC layer was verified among the nugget and substrate by using the modified sample perparation method. The galvanic series toward negative potential direction is thereafter achieved as: IMC > steel HAZ > base metal of DP780 > base metal of 5182 > weld nugget of aluminium > galvanized coating of DP780, meaning significant galvanic corrosion tendency between IMC layer and aluminium nugget. With the potential difference of 0.184 V, IMC layer might be protected under cathodic polarization while the aluminium nugget might be degraded under anodic polarization in the galvanic couple.
(3)Without considering galvanic corrosion effect, EIS results show the existence of inductive loop due to localized corrosion on aluminium and IMC layer. However, in the cross-sectional weld joint during immersion test, the tiny area of cathodic IMC layer ratio to large anodic area of aluminium nugget is responsible for the slight corrosion of IMC layer, as well as for the preferential initiation of localized corrosion at the interface front near IMC layer.

Data availability statement
All data that support the findings of this study are included within the article (and any supplementary files).