Enhancing High-Temperature Durability of Aluminum/Steel Joints: The Role of Ni and Cr in Substitutional Diffusion Within Intermetallic Compounds

Abstract

The automotive and aerospace industries increasingly rely on lightweight, high-strength materials to improve fuel efficiency, making the joining of dissimilar metals such as aluminum and steel both beneficial and essential. However, a major challenge in these joints is the formation of brittle intermetallic compounds (IMCs) at the interface, even when using low heat-input solid-state welding methods like friction stir welding (FSW). Furthermore, IMC growth at elevated temperatures significantly limits the service life of these joints. In this study, an intermediate layer of stainless steel was deposited on the steel surface prior to FSW with aluminum. The resulting Al–Steel joints were subjected to heat treatment at 400 °C and 550 °C to investigate IMC growth and its impact on mechanical strength, with results compared to conventional joints without the intermediate layer. The intermediate layer significantly suppressed IMC formation, leading to a smaller reduction in mechanical strength after heat treatment. Joints with the intermediate layer achieved their highest strength (350 MPa) after heat treatment at 400 °C, while conventional joints exhibited their highest strength (225 MPa) in the as-welded condition. At 550 °C, both joint types experienced a decline in strength; however, the joint with the intermediate layer retained a strength of 100 MPa, whereas the conventional joint lost its strength entirely. This study provides an in-depth analysis of the role of IMC growth in joint strength and demonstrates how the intermediate layer enhances the thermal durability and mechanical performance of Al–Steel joints, offering valuable insights for their application in high-temperature environments.

1. Introduction

In the wake of increasing environmental concerns and carbon dioxide (CO₂) pollution, manufacturers have turned to reducing fuel consumption by reducing the weight of transportation vehicles, including cars, ships, and airplanes. Approximately, by reducing the weight of a vehicle by 10%, 5.5% can be saved in fuel consumption [1,2]. Therefore, the usages of light metals such as aluminum and magnesium alongside steel and their joining became important. In aluminum to steel joints, the final piece will have a lighter weight, better corrosion resistance, and higher strength to weight ratio despite its high strength, creep resistance, and good ductility [3,4]. On the other hand, dissimilar metals welding such as aluminum to steel is very difficult due to differences in melting points, different thermal and electrical conductivities, different cooling rates, and different heat and specific heat capacities. In addition, various intermetallic compounds (IMCs) such as Fe₂Al₅ (η), FeAl₃ (θ), FeAl₂, Fe₃Al, and FeAl are formed at the aluminum–iron joint interface, which are very brittle and susceptible to corrosion and cracking [5]. Liyanage et al. [6] reported that iron-rich IMCs such as FeAl and Fe₃Al are more ductile than aluminum-rich IMCs such as Fe₂Al₅ and Fe₄Al₁₃, and therefore iron-rich IMCs have better strength in the joint interface than brittle IMCs. However, it has been found that when the IMCs layer thickness is less than 2 μm, it does not have a significant effect on the joint interface strength [7]. It is also well known that the elements of the adjoints greatly affects the type and morphology of IMCs at the joint interface [8].

Common processes for joining dissimilar materials include fusion welding processes, solid state welding processes, weld brazing, mechanical joining, and low-dilution welding processes [2,9,10]. In aluminum to steel joints by fusion welding (FW) processes, due to the high heat input during the melting process, a thick layer of IMCs is formed and greatly weakens the interface joint between dissimilar alloys, hence joining of aluminum alloys to steel by fusion welding processes is not recommended [11]. Singh et al. [12] welded Al6061 to 304l stainless steel using MIG, FSW, and friction crush welding (FCW) processes and found that the weld appearance and tensile strength of the pieces are welded by FSW and FCW processes were more desirable than MIG process. Low total heat input per weld length, dilution control, precise beam alignment, and the absence of oxygen, nitrogen and hydrogen are some of the advantages of dissimilar metal low-dilution welding processes [13]. On the other hand, equipment cost and the need for careful joint preparation are some disadvantages and limitations of beam welding processes [13,14].

Friction stir welding (FSW) process is known as an emerged and clean process due to its low heat input, less material waste, high weld quality, no need for shielding gases, no spatter, and no need for filler materials [15,16,17]. In recent years, the FSW process of aluminum to steel alloys has been used due to the reduction in IMCs layer formation, improved mechanical properties, economic efficiency, environmental friendliness and the absence of the need for consumables materials [18,19]. Due to the solid-state friction stir welding process, most of the melting and solidification problems into the pieces will be eliminated such as oxidation, porosity, hydrogen dissolution, and shrinkage [19]. Also, the IMCs thickness at aluminum–steel joint interface will be reduced due to the low heat input in the friction stir welding process. Mitsuhiro Watanabe et al. [20] determined that in the FSW of pure aluminum to low-carbon steel, peak temperature was obtained 700–780 °C for about 60 s during rotational speed 3000 rpm and speed plunge 6.7 × ${10}^{-5} \mathrm{m}/\mathrm{s}$. Tanaka et al. [21], showed that in Al1050–mild steel joint by FSW process with the parameters of a tool rotation speed of 1800 rpm and a weld travel speed of 100 mm/min, peak temperature reached 730–738 °C for 4 s. Fe₂Al₅ and FeAl₃ layer thickness was observed to be about 1.7–6 µm. While the formation of intermetallic compounds (IMCs) is necessary to ensure metallurgical bonding, it can be replaced by an amorphous layer that is extremely thin and results in higher joint strength [22].

In the FSW process of Al-Fe alloys, various factors can affect IMCs layer thickness and weld strength including pin rotation speed, welding speed, strain rate and welding pressure [23,24,25]. Coelho et al. [26] reported that, IMCs will form with high shear strain. Pourali et al. [27] showed that the most desirable tensile strength was achieved at low welding speed and high rotation speed in Al1100 to St37 by FSW. Bokov et al. [28] demonstrated the effect of pin shape in AA1100 to St14 by FSW process using cylindrical pin and imperfect conical pin and determined that more heat will be generated because the contact area of the cylindrical pin with the workpiece. Naseri et al. [29] investigated the effect of welding speed and pin rotation speed in Al1050 to 316L stainless steel by FSW process and observed that among the rotation speeds of 560 and 900 rpm and welding speed of 60, 80, 100, and 125 mm/min, the best sample in terms of microstructure, mechanical properties and weld quality belongs to the sample welded with speed of 125 mm/min and a rotation speed of 900 rpm.

Beygi et al. [30] in 2021 investigated the effect of alloying elements on aluminum–steel dissimilar joints by FSW process and showed that among the alloying elements in aluminum that can affect the growth rate of aluminum–iron IMCs, silicon (Si) has the greatest effect. Also, the alloying elements of stainless steels have a significant reducing effect on the IMCs thickness during FSW process. Nickel (Ni) and chromium (Cr) delay diffusion and may contribute to the IMCs toughness through solid solution strengthening. Beygi et al. [31] investigated the effect of the buttering process of 316L stainless steel on the surface of carbon steel and connected to AA1050 by FSW and stated that the buttering process reduced the thickness of the IMCs layer and increased the tensile strength from 450 MPa to 800 MPa due to the addition of alloying elements such as nickel and chromium. Movahedi et al. [32] investigated the effect of annealing heat treatment on the growth kinetics of the IMCs layer at the joint interface of the St-12/AA5083 friction stir lap welds. Finally, it was found that during the annealing heat treatment at 350 °C/30 min, a discontinuous intermetallic layer (IM) was formed, which became thinner and continuous as the annealing time increased to 90 °C and 180 °C. On the other hand, the thickness of the IM layer (D_IM) increased as the annealing heat treatment temperature increased to 400 °C and 450 °C. Their results indicated that the IM layer at the St12/AA5083 joint interface was composed of two IMCs, Fe₂Al₅, and FeAl₃. Springer et al. [33] investigated the effects of the annealing heat treatment on the growth kinetics of the IM layers of low carbon steel/pure Al and low carbon steel/Al–5 wt.% Si in FSW butt joint. In this research, annealing heat treatment temperature was performed from 200 °C to 600 °C and a time range of 4 min to 64 min, resulted in the IMCs layer composed of Fe₂Al₅ and narrow layer of FeAl₃. Beygi et al. [34] investigated the effect of heat treatment after AA1050-St37 FSWed and found that the growth rate of IMCs layer was very low at 300 °C for 2 h. However, the IMCs layer thickness remained unchanged at longer annealing times.

The influence of an intermediate layer in inhibiting intermetallic compound (IMC) growth during the friction stir welding (FSW) of aluminum to steel and enhancing joint mechanical strength has been previously studied. However, its role and the contribution of its elements to IMC growth at high-temperature service conditions remain unclear. This study investigates the durability of Al–Steel joints at elevated temperatures with and without an intermediate layer. Joints were heat-treated at 400 °C and 550 °C for 90 min, and the IMC layer thickness and its effect on joint strength were analyzed and discussed.

2. Material and Methods

In this study, AA1050 and St37 sheets with dimensions of 5 mm × 50 mm × 150 mm and 2 mm × 50 mm × 150 mm, respectively, were welded in butt joint configuration by FSW process. According to Figure 1, to unify the thickness of the aluminum and steel sheets, two AA1050 sheets with thicknesses of 1 mm and 2 mm were used under and above the steel, respectively. The edges of AA1050 and St37 sheets were flattened and cleaned with sandpaper and alcohol, respectively, so that the joint was free of any grooves or holes and the presence of oil, grease, and any contamination was controlled. The sheets were placed side by side in a completely parallel manner using fixtures so that the weld seam was welded uniformly to the end. After the completion of the FSW process, the extra sheets were carefully separated, leaving a joint with dissimilar thicknesses of Al and St. To investigate the effect of intermediate layer on joint durability and IMCs growth, two sets of welds were performed. A-series joints were prepared by weld-deposited layer of the faying surface of St with stainless steels (SS) 316L before FSW. In contrast, B-series joints were prepared without any intermediate layer.

Table 1. Chemical composition of the materials used and mechanical properties of the base materials.

Mo	Ni	Cr	Ti	Zn	Mg	Cu	Si	Mn	C	Al	Fe
-	-	-	-	-	-	-	-	0.25–0.4	0.08	-	Bal.	St37
			0.05	0.07	0.05	0.05	0.25	0.05	-	Bal.	0.4	AA1050
2–3	10–14	16–18	-	-	-	-	1.0	2.0	0.03	-	Bal.	316l SS
Hardness (VH)				Yield Strength (Mpa)				Tensile Strength (Mpa)
120				300				370				St37
41				85				100–135				AA1050

shows the chemical composition and mechanical properties of AA1050, St37, and SS316L filler material. The friction stir welding process has been performed in the position control mode. Here, only the optimum conditions were used for the FSW process, a 950 rpm rotation speed and a 20 mm/min welding speed, with a 1.3 mm pin offset of the tool into St and 0.1 mm plunge depth, according to our previous study.

Table 2. FSW parameters and tool pin specifications.

Tool Specification		Friction Stir Welding Parameters
Material:	H13 tool steel	Tool rotation speed (ω):	950 rpm
Tool shoulder diameter (φ):	18 mm	Tool traverse speed (Vx):	20 mm/min
Tool shoulder concavity:	2.5°	Tool offset:	1.3 mm to AS
Tool pin diameter:	5 mm	Tool tilt angle (θ):	2.5°
Tool pin length:	4.7 mm	Pin plunge depth:	4.8 mm

shows the welding parameters and tool pin specifications.

In this study, H13 tool steel with a hardness of 45 HRC was used as the FSW tool pin material. Other features of the tool include a shoulder diameter of 18 mm, a pin diameter of 5 mm, a probe height of 4.7 mm and a shoulder concavity of 2.5 degrees. To investigate the post-welding heat treatment, A- and B-series samples were heat treated for 90 min at 400 °C and 550 °C subsequently cooled in the air as shown in

Table 3. Heat treatment conditions.

Pieces Name	Sample Name	Features	Time-Temperature Cycle Heat Treatment
A-series	A1	with intermediate layer	As-welded
	A2	with intermediate layer	400 °C-90 min
	A3	with intermediate layer	550 °C-90 min
B-series	B1	without intermediate layer	As-welded
	B2	without intermediate layer	400 °C-90 min
	B3	without intermediate layer	550 °C-90 min

.

Samples were mounted to investigate the microstructure of the interface as shown in Figure 2. The interface of the samples was metallographically examined under a scanning electron microscope (SEM). SEM images of the interface were obtained at different magnifications in back scatter (BS) mode to investigate the thickness of the IMCs and analysis the morphology and distribution of IMCs. EDS analysis, line scan, and elemental mapping were performed from different areas of the joint interface to determine the chemical composition of the IMCs formed in the weld areas. Microhardness tests were performed on the weld cross sections to examine the local mechanical properties of the joints.

To examine the weld strength, 5 mm width specimens were cut and then tested in a tensile testing machine at a loading rate of 2 mm/min. Figure 3 shows the specimens prepared for the tensile test. After the samples were fractured in the tensile test, SEM images were taken at different magnifications in the SE mode to examine the fracture mechanism.

3. Results and Discussion

3.1. Al-St Joint Interface Microstructure

In Figure 4, SEM images of AA1050-St37 joint interface are shown. As it is clear, no cracks or defects were formed at the Al-St joint interface of samples during FSW. In sample B₃, due to annealing heat treatment at 550 °C and the application of high input heat, IMCs layer formed at the joint interface and became so thick that the sample separated from the weld line without applying force or stress.

Figure 5 shows BS SEM images of the interface of A-series joints and B-series. As it is clear from SEM images of sample A₁ in Figure 5a–c, no IMCs layer has formed at the Al-St interface, or in other words, in the resolution of the image no IMC layer was observed. As can be observed, the thicknesses of the IMCs layer in samples A₁ are far below 0.1 microns. No defect is observed at the interfaces. Particles of St fragments are observed in the Al-matrix. These particles are white in samples A₁. Considering that the IMCs layers are often brittle and are prone to crack initiation and growth, it can be claimed that the joint interface would have desirable weld strength in samples A₁.

As it is clear from BS SEM images of sample B₁ in Figure 5d–f, a thick IMCs layer is formed at the Al-St joint interface. The thickness of the IMCs layer at the Al-St joint interface was measured between 2 and 3 microns. Defects such as cracks and discontinuities in the IMCs layer have been observed. While the presence of a thick IMCs layer between aluminum and steel reduces the strength of the joint, these discontinuities act as barriers for crack propagation in the IMC layer during loading.

Figure 6 shows the EDS analysis of samples A₁ and B₁. Figure 6b,e show the EDS analysis of St37 and AA1050 areas, respectively. As shown in Figure 6b, this region consists of 80.91% At Fe, and as shown in Figure 6e, this area consists of 99.13% At Al. By comparing the EDS analysis of these two areas, it can be concluded that a lower atomic percent of alloying elements has diffused into the AA1050 piece. In the EDS analysis of Figure 6d, there are 72.74% At Fe and 1.87% At Al, which means that it can be stated that the analysis was prepared from St fragments. On the other hand, considering the higher amounts of alloying elements such as chromium, nickel, and manganese in this area compared to Figure 6b, it can be stated that this fragment of iron was torn off from the surface of the piece with intermediate layer due to the rotation of the pin and thrown towards the AA1050 sheet. The EDS analysis of Figure 6c was taken from the gray interface between the St fragment and AA1050 piece and as it is clear, 83.70% At Al and 12.14% At Fe and trace amounts of alloying elements are present in this area. In the line scan analysis taken across the joint interface shown in Figure 6f, a steep gradient of Al element in the right side and Fe, Cr, and Ni elements in the left side are observed. The region over which this gradient exists is very narrow compared to the ones observed in conventional joint, less than 0.1 microns in sample A₁ and about 2 microns in sample B₁. Furthermore, the gradient is much steeper in sample A₁ with respect to B₁, which is almost flat (see Figure 6m). This is strong evidence of a lower diffusion rate of Al and Fe in sample A₁ with respect to B₁, which is attributed to substitutional diffusion of Cr and Ni in sample A₁ [30]. As it is clear from Figure 6g,h elemental map images for Al (red spot) and Fe (green spot), there is no trace of the presence of aluminum and iron atoms next to each other to form IMCs. The elemental map analyses indicate that no diffusion occurred, at least at the present scale, between Al and St to form IMC layer. In Al-St dissimilar joints, alloying elements such as Ni and Cr can reduce the thickness of the IMCs layer and improve the tensile strength of the joint interface [30]. Ni and Cr retard diffusion and may contribute to the toughness of IMCs through solid solution strengthening [29]. As it is clear from the line scan of Figure 6m,n from the IMCs layer, the IMCs layer contains about 70% At Al and 30% At Fe. Therefore, the chemical composition of the IMCs layer can be estimated to be Fe₂Al₅ in sample B₁. Elemental map analyses in Figure 6o,p at the Al-St joint interface clearly show the distribution of Al and Fe.

As it is clear from SEM images of sample A₂ in Figure 7a–c, which was heat treated at 400 °C, the IMCs layer is still absent at the Al-St joint interface. As can be observed, the thicknesses of IMCs layer in samples A₂ are far below 0.1 µm. Given that the IMCs layer is brittle and will cause crack initiation and growth when shear or normal stress is applied, it can be expected that the weld strength in sample A₂ will be significantly high due to the lack of formation of the IMCs layer at the St-Al joint interface. Figure 7d–f is specific to sample B₂ which is heat treated at 400 °C for 90 min and then slowly cooled in air. As is observed from Figure 7, a slight growth and thickening of the IMC layer occurred. The thickness of the IMCs layer in B₂ sample is measured to be between 2 and 4 µm. Instead, the IMC layer at the interface changed to a continuous and flat state, meaning that the discontinuities were eliminated. This phenomenon was reported in reference [34], where the local diffusion is enhanced at the tip of protrusions in the IMC layer due to the curvature effect. The absence of discontinuities in the IMC layer makes the crack propagation easier, as these would act as crack arrestors.

Figure 8a shows SEM BS image of sample A₂. EDS analysis of the interface of the gray areas are shown in Figure 8b,c. No specific IMC could be recognized from these EDS analyses, indicating that a solid solution of Al-Fe is formed. Figure 8e–j show elemental map of Fe, Al, Ni, Cr, and Mn for Sample A₂. A sharp interface is observed in every map indicating that an IMC layer was not formed. Figure 8j shows the SEM BS image of Sample B₂ along with the line scan analysis of IMCs layer. On the other hand, according to the EDS analysis of Figure 8m,n, which is taken from the IMCs layer of St/Al joint interface, it can be stated that in this area, likewise Sample B₁ an intermetallic of Fe₂Al₅ was formed.

Figure 9 shows the interfaces of Samples A₃ and B₃, which were heat treated at 550 °C for 90 min and then slowly air cooled. As it is clear from SEM images of Sample A₃, the IMC layer has thickened to 10 µm. A similar trend is observed in St fragments in the Al-matrix. The outer regions of these fragments have been converted to a IMC layer, leaving the core as St. All these observations imply a rapid growth of the IMC layer at 550 °C, even though Ni and Cr elements are present. According to Figure 9d heat treatment for sample B₃ of conventional joint at 550 °C for 90 min led to joint failure. The high temperature of the annealing heat treatment and the absence of alloying elements caused the excessive growth of IMCs layer at the joint interface such that the joint failed during sample manipulation.

Figure 10 shows EDS analysis of Sample A₃. The chemical composition analysis of the IMCs layer is shown in Figure 10b,c. As it is clear from Figure 10b, there are 72.19% At Al and 20.26% At Fe, and in Figure 10c, there are 75.54% At Al and 18.06% At Fe. The alloying elements chromium, nickel, and manganese are also present in this layer, which mean that it can be stated that these elements have diffused into IMC as substitutional atoms. As a result, it can be assumed that the chemical composition of the IMCs layer of the A₃ joint interface is rich in aluminum (Fe₄Al₁₃). Figure 10e–i show elemental maps of various elements across the interface. The gradient of chemical compositions is obvious, which is an indicative of IMC formation. Although other elements such as Ni and Cr have diffused into the IMC layer, Figure 10j shows an EDS line scan across the interface. At the distance of 20 to 30 microns, the line scan analysis of the IMCs layer contains about 65% Al, 20% Fe, and small amounts of alloying elements. Thus, it can be concluded that an aluminum-rich IMCs has formed in this layer. Line scan analysis across the interface shows a wider interaction layer and less steep gradient of elements in this layer, which is another indicative of higher diffusion with respect to sample A₂.

In Figure 11, the line scan analysis of Sample A₃ is taken from the Al/St joint interface. As can be seen from Figure 11, Al mole fraction increases and iron, nickel, chromium, and manganese mole fractions decrease. This layer contains about 0.65% Al, 0.25% Fe, and small amounts of alloying elements, so it can be estimated that the chemical formula of this layer is Fe₄Al₁₃.

Figure 12a shows higher magnification SEM image of the interface of Sample B₃ in which the gap between St and Al is obvious. A distinct region is observed at the interface of St whose thickness varies between 20 and 50 µm. To further clarify this region, elemental maps of Fe and Al are provided in Figure 12b and 12c, respectively. A trace line of Al is observed at the interface of St in Figure 12c, while no trace line of Fe is observed at the interface of Al. As this trace line of Al at the interface of St is an indicative of IMC layer, its presence implies that the joint has failed along IMC/Al interface, leaving the IMC layer attached to the interface of St.

To further analyze the kinetics of IMC growth in samples A and B, the St fragments in the matrix of Al of A₃ and B₃ samples were studied. Figure 13a,c show the St fragment in the matrix of Al of A₃ and B₃ samples, respectively. The particles clearly comprise two distinct regions: a white core and a gray shell. The core is St and the surrounding is the IMC layer formed due to diffusion. Figure 13b,d show EDS line scan results along the red lines passing through the fragments. The IMC shell and the core compositions are distinguished through the gradients of the composition. Accordingly, the maximum thicknesses of the IMC layer in the fragments in A₃ and B₃ samples are 5 µm and 30µm, respectively. This difference is another evidence of lower kinetics of IMCs growth in samples with intermediate layer.

Formulation for the growth of different kinds of Al-Fe IMC is obtained according to [35]:

$$x^2=kt$$

(1)

where x (m) is IMCs thickness, k is a constant that is dependent on temperature and t (s) is time. K (${\mathrm{m}}^2/\mathrm{s})$ is obtained using Equation (2):

$$\mathrm{k}=k_{0}exp\left({\displaystyle \frac{-Q}{RT}}\right)$$

(2)

where k₀ is a constant and Q is the activation energy. The coefficient of K for IMCs of A-series and B-series are calculated and presented in

Table 4. Coefficient of K for A₂, A₃, B₂, and B₃.

Sample	X (m)/IMCs Thickness	t (s)/Annealing Time	$\mathbf{K} ({\mathbf{m}}^2/\mathbf{s}$)
A2	0	5400 (s)	0
A3	10 × ${10}^{-6}$ (m)	5400 (s)	1.85 × ${10}^{-14} \left({\displaystyle \frac{{\mathrm{m}}^2}{\mathrm{s}}}\right)$
B2	3.5 × ${10}^{-6}$ (m)	5400 (s)	2.26 × ${10}^{-15} \left({\displaystyle \frac{{\mathrm{m}}^2}{\mathrm{s}}}\right)$
B3	30 × ${10}^{-6}$ (m)	5400 (s)	1.66 × ${10}^{-13} \left({\displaystyle \frac{{\mathrm{m}}^2}{\mathrm{s}}}\right)$

.

To clearly see how the alloying elements have diffused along with Fe to form the IMC layer, the ratio of these elements to Fe have been calculated and drawn across the IMC layer. In Figure 14, the ratio of iron to alloying elements is shown as a function of distance for samples A₁, A₂, and A₃. As can be seen from Figure 14, the Fe–Cr in A₁, A₂ and, A₃ has changed almost uniformly, but the Fe–Ni and Fe–Mn have changed in a declining trend. In other words, Ni and Mn have a higher diffusion rate than Cr. It can be said that Cr has the major role in hindering the growth of the IMC layer due to its lower diffusion rate. In order to identify this phenomena, Darken’s equation will be investigated. According to [36,37], in a solid solution that presents α and β as alloying elements in component i, Fick’s first law can be obtained from Equation (3) as the function of intrinsic diffusion fluxes:

$${\mathrm{j}}_{\mathrm{\alpha }}=-D_{\mathrm{\alpha }}^i{\displaystyle \frac{{\partial \mathrm{C}}_{\mathrm{\alpha }}}{\partial \mathrm{x}}}, {\mathrm{j}}_{\mathrm{\beta }}=-D_{\mathrm{\beta }}^i{\displaystyle \frac{{\partial \mathrm{C}}_{\mathrm{\beta }}}{\partial \mathrm{x}}}.$$

(3)

The flux of component i, can be written as:

j_i = −M_iC_i ∂µ_i/∂x,

(4)

where M_i denotes the mobility of component I, $C_{\mathrm{i}}$ = ${\displaystyle \frac{N_{\mathrm{i}}}{V_M}}$ is called concentration of component i, µ_i chemical potential that can be expressed in terms of the thermodynamic activity, a_i, via:

$${\mathrm{\mu }}_{\mathrm{i}}={\mathrm{\mu }}_i^0+\mathrm{RT} \ln a_{\mathrm{i}},$$

(5)

where ${\mu }_i^0$ is the standard chemical potential (at 298 K and 1 bar). The atomic mobility M_i is connected to the tracer diffusion coefficient ($D_i^{*}$) of component i:

$$D_i^{*}={\mathrm{M}}_{\mathrm{i}}\mathrm{RT}.$$

(6)

Substituting Equations (4) and (6) in Equation (3) relations between the intrinsic and the tracer diffusion coefficients is obtained as the following:

$$D_{\mathrm{\alpha }}^i=D_{\mathrm{\alpha }}^{*}{\displaystyle \frac{{\partial \mathrm{L}\mathrm{n}}_{a_{\mathrm{\alpha }}}}{{\partial \mathrm{L}\mathrm{n}}_{N_{\mathrm{\alpha }}}}}, D_{\mathrm{\beta }}^i=D_{\mathrm{\beta }}^{*}{\displaystyle \frac{{\partial \mathrm{L}\mathrm{n}}_{a_{\mathrm{\beta }}}}{{\partial \mathrm{L}\mathrm{n}}_{N_{\mathrm{\beta }}}}},$$

(7)

${\displaystyle \frac{{\partial Ln}_{a_{\alpha }}}{{\partial Ln}_{N_{\alpha }}}}$ is denoted as the thermodynamic factor and is known with Φ. So, from thermodynamic formula it can be expressed Φ as:

$$\Phi ={\displaystyle \frac{{\partial \mathrm{L}\mathrm{n}}_{a_{\mathrm{i}}}}{{\partial \mathrm{L}\mathrm{n}}_{N_{\mathrm{i}}}}}=1+{\displaystyle \frac{{\partial \mathrm{L}\mathrm{n}}_{{\gamma }_{\mathrm{i}}}}{{\partial \mathrm{L}\mathrm{n}}_{N_{\mathrm{i}}}}}={\displaystyle \frac{N_{\mathrm{\alpha }}{N}_{\mathrm{\beta }}}{RT}}{\displaystyle \frac{d^{2}G}{d{N}_i^2}}$$

(8)

γ_i ≡${\displaystyle \frac{a_i}{N_i}}$ the coefficient of thermodynamic activity of species i. In addition, as a consequence of the Gibbs–Duhem relation, there is only one thermodynamic factor for a binary alloy:

$$\Phi ={\displaystyle \frac{{\partial \mathrm{L}\mathrm{n}}_{a_{\mathrm{\alpha }}}}{{\partial \mathrm{L}\mathrm{n}}_{N_{\mathrm{\alpha }}}}}={\displaystyle \frac{{\partial \mathrm{L}\mathrm{n}}_{a_{\mathrm{\beta }}}}{{\partial \mathrm{L}\mathrm{n}}_{N_{\mathrm{\beta }}}}}$$

(9)

So, in samples A₁, A₂, and A₃ in which Cr, Ni, and Mn exist as alloying elements by the intermediate layer of 316L SS, the diffusion coefficient for Cr, Ni, and Mn is calculated according to Equation (7):

$$D_{Cr}^{Fe}={{\mathrm{D}}^{*}}_{\mathrm{Cr}} \left\{1+{\displaystyle \frac{\mathrm{d}\ln {\mathrm{\gamma }}_{\mathrm{C}\mathrm{r}}}{\mathrm{d}\ln {\mathrm{x}}_{\mathrm{C}\mathrm{r}}}}\right\}={\mathrm{M}}_{\mathrm{Cr}}\mathrm{RT}{\displaystyle \frac{{\partial \mathrm{L}\mathrm{n}}_{a_{\mathrm{C}\mathrm{r}}}}{{\partial \mathrm{L}\mathrm{n}}_{N_{\mathrm{C}\mathrm{r}}}}},$$

(10)

$$D_{Ni}^{Fe}={{\mathrm{D}}^{*}}_{\mathrm{Ni}} \left\{1+{\displaystyle \frac{\mathrm{d}\ln {\mathrm{\gamma }}_{\mathrm{N}\mathrm{i}}}{\mathrm{d}\ln {\mathrm{x}}_{\mathrm{N}\mathrm{i}}}}\right\}={\mathrm{M}}_{\mathrm{Ni}}\mathrm{RT}{\displaystyle \frac{{\partial \mathrm{L}\mathrm{n}}_{a_{\mathrm{N}\mathrm{i}}}}{{\partial \mathrm{L}\mathrm{n}}_{N_{\mathrm{N}\mathrm{i}}}}}.$$

(11)

$$D_{Mn}^{Fe}={{\mathrm{D}}^{*}}_{\mathrm{Mn}} \left\{1+{\displaystyle \frac{\mathrm{d}\ln {\mathrm{\gamma }}_{\mathrm{M}\mathrm{n}}}{\mathrm{d}\ln {\mathrm{x}}_{\mathrm{M}\mathrm{n}}}}\right\}={\mathrm{M}}_{\mathrm{Mn}}\mathrm{RT}{\displaystyle \frac{{\partial \mathrm{L}\mathrm{n}}_{a_{\mathrm{M}\mathrm{n}}}}{{\partial \mathrm{L}\mathrm{n}}_{N_{\mathrm{M}\mathrm{n}}}}}.$$

(12)

The atoms’ mobility M_i is obtained by experiments. Reference [38] suggests that it can be expressed as follows:

$${\mathrm{M}}_{\mathrm{i}}=\exp ({\displaystyle \frac{RTLn M_i^0}{RT}})\exp ({\displaystyle \frac{-Q_i}{RT}}){\displaystyle \frac{1}{RT}}{}^{\mathrm{mg}}\Omega$$

(13)

where $M_i^0$ is frequency factor, Q_i an activation enthalpy, and ^mgΩ is a factor taking into account a ferromagnetic contribution to the diffusion. According to Raoult’s law, an ideal solid solution alloy γ_i = 1 and a_i = N_i and hence Φ = 1. However, in non-ideal solutions Φ deviates from unity, which is called Henry’s law. It is γ_i > 1 for phases with negative deviations and smaller than γ_i < 1 in the opposite case. When negative deviations (γ_i > 1) occurred, Ni, Cr, and Mn create solid solutions with Fe placing into the crystal vacancies. As it is clear from Figure 14, which shows the ratio of iron to Ni, Cr, and Mn in Samples A₁, A₂, and A₃, it can be stated that at room temperature, 400 °C and 550 °C: $D_{Cr}^{Fe}<D_{Ni}^{Fe}<D_{Mn}^{Fe}$. Moreover, according to [39], the diffusion coefficient of Cr in to iron varies from 10⁻¹⁵ to 10⁻¹² m²/s under temperatures of 600–800 °C and the Ni diffusion coefficient in to Fe between 600 °C and 680 °C, which is expressed as follows:

D = D₀ exp (−Q/RT)

(14)

where D₀ is 0.00014 m²/s and Q is 58,700 cal/mole. So, Ni diffusion coefficient in to iron between mentioned temperature is 3.4 × 10⁻⁸ m²/s [40].

Additionally, as it is clear from reference [41], which investigates the evolution of the IMCs layer between Fe-15Cr/Al and Fe-35Ni/Al, it can be stated that the presence of Cr leads to formation of Fe₂Al₅ and Fe₄Al₁₃ in which the growth of Fe₂Al₅ was described by a parabolic rate law; on the other hand, in Fe-35Ni/Al just Fe₄Al₁₃ has been formed so that the thickness of the IMCs decreases with increasing temperature. In fact, the presence of Cr promotes the formation of various IMCs layers rather than Ni. This observation indirectly reflects that chromium influences diffusion behavior, which accelerates the overall interfacial reaction kinetics.

Figure 15 shows the average thickness of IMCs layer at the joint interface in each sample. As it is clear, annealing heat treatment at 550 °C will increase the thickness of IMCs layer more significantly in samples without an intermediate layer.

3.2. Mechanical Properties

The force-extension diagram is shown in Figure 16. As it is clear from Figure 16a, the highest ultimate tensile strength (UTS) belongs to sample A₂. By depositing the 316l stainless steel layer on the surface of St37 sheet, UTS and elongation increased from 239.7 (MPa) to 342.7 (MPa) and 1.6 (mm) to 6.12 (mm), respectively. On the other hand, the fracture in sample B₁ was completely brittle and without plastic deformation, while in sample A₁, the specimen was necked from the aluminum side. From the comparison of Samples A₂ and B₂ tensile test, it can also be concluded that the addition of Cr and Ni alloy elements through the 316L intermediate layer has made the joint interface more resistant, so that UTS has increased about 48% from 168.2 (Mpa) to 354.4 (Mpa). In fact, it can be stated that the deposition process of 316l stainless steel onto the St37 sheet and then joining it to AA1050 by FSW process will increase toughness. In other words, the joint will be so strong that failure will occur from the areas adjacent to the weld line and in the aluminum part. One of the main reasons for this is that Cr and Ni will reduce the thickness of the IMCs layer and on the other hand, it will improve the toughness of the IMCs layer through solid solution strengthening [30].

Figure 17 shows the UTS values of the joints in different heat treatment conditions as well as in the as-welded condition. The strength of B-series declines by heat treatment, reaching zero at 500 °C. While a slight improvement is observed in Sample A₂, the strength is decreased in Sample B₂ by heat treatment at 400 °C. As the thickness of the IMCs layer was only slightly changed in Sample B₂ with respect to Sample B₁, this reduction cannot be attributed only to thickening of the IMC layer. A more precise comparison of the IMCs layer in these two samples shows that heat treatment at 400 °C caused the nonuniformities and irregularities to disappear in the IMC layer. This makes the crack propagation easier and thus the joint becomes less tough. The irregularities in the as-welded sample causes the fracture to be hindered and thus the joints toughness is higher. In sample A, as there is no IMC layer in the as-welded sample (A₁) and no thickening occurred by heat treatment at 400 °C (A₂), the increase in the joint strength can be attributed to the release of residual stresses.

The microhardness of Samples A₁ (blue line), A₂ (orange line), A₃ (red line), B₁ (green line), B₂ (black line), and B₃ (pink line) is shown in Figure 18. As can be seen from Figure 18, micro hardness will increase in any condition at the interface of AA1050 to St37 joint by FSW process. In samples B₃, B₂, and A₃, high micro hardness was achieved due to the formation of a thick IMCs layer at the Al-St joint interface. As can be seen from Figure 15, the average thickness of the IMCs layer in Samples B_1, B₂, and A₃ is 2 µm, 3.5 µm, and 10 µm, respectively. On the other hand, the IMCs layer formed in sample B₁ and B₂ was determined to be Fe₂Al₅. The micro hardness of samples A₁ and A₂ was 165 HV and 153 HV, respectively, due to the lack of IMCs layer at the Al-St interface. As a result, it can be stated that, given the appropriate and desirable hardness achieved by samples A₁ and A₂, they will not break brittlely and suddenly when tensile stress is applied.

Figure 19 shows the SEM fracture surfaces of the samples with an intermediate layer after a tensile test. As shown in Figure 19a,c, which are the fracture surface images from the AA1050 and St37 sides, respectively, it can be stated that the fracture occurred after necking in ductile mode in Sample A₁. Dimples and the egg-shaped surface indicate that the fracture occurred after high elongation percentage. Due to the very low thickness of the IMCs layer at the interface, the fracture did not occur in the brittle mode. In Figure 19e,g, the fracture surface images of Sample A₂ from the Al and St sides are shown. Sample A₂ also broke in ductile mode after necking. Ups and downs and the presence of dimples on the fracture surface of Al and St indicate that the welding joint was made with high strength, so that the fracture and rupture occurred from the Al side. Figure 19i,k show the Sample A₃ fracture surface, heat treated at 550 °C for 90 min. As can be seen from Figure 16, Sample A₃ fractured in brittle mode. In other words, during the tensile test due to the presence of the thick IMCs layer at the Al-St interface, the sample ruptured brittlely from the weld and at low elongations. The bright, smooth, and cleavage fracture surfaces are among the characteristic features of brittle fracture surface images. It can be concluded that by heat treating the AA1050-St37 joint at 550 °C for 90 min, the IMCs layer will form at the joint interface, leading to a decrease in the strength of the weld interface.

Figure 20a,d show the fracture surfaces of the Al and St side of B₁ after the tensile test. As can be seen, a fracture occurred completely brittle and cleavage. In the tensile test of Sample B₁, shown in Figure 16, it is clear that the sample broke at low elongations. Figure 20e,h show the fracture surface images of sample B₂ heat treated at 400 °C for 90 min. As it is clear, the fracture surface in this sample is a combination of ductile and brittle fracture. The presence of dimples and ups and downs on the sides of the piece and the smooth surface and cleavage in the middle are evident in Figure 20e,f. On the other hand, by examining the tensile test of Sample B₂ in Figure 16, it can be inferred that the sample broke at high elongations but with low strength. This means that the annealing heat treatment has increased the ductility, but the formation of the IMCs layer at the joint interface has reduced the weld strength. The increase in ductility is attributed to annealing of Al. Figure 20i,l show the fracture surface images of sample B₃ heat treated at 550 °C for 90 min. In this sample, the joint was ruptured spontaneously without applying stress after the welding operation due to the thickening of the IMCs layer because of the high heat input. The fracture surface is completely brittle and cleavage on Al and St sides. No dimple is observed even at the top and bottom of the joint.

Figure 21 shows the fracture mode schematic of Samples A and B. As it is clear from Figure 21a, which relates to samples A₁ and A₂, the failure happens around the joint interface and on the aluminum side that indicates sufficient and desirable strength of the AA1050-St37 joint in these two samples. On the other hand, according to Figure 21b, Sample A₃ is ruptured from the join interface in ductile mode at the starting and ending of the S-shaped interface and brittle mode at the middle of interface. The high thickness of the IMCs layer led to a decrease in the joint interface strength when annealing heat treatment was performed at 550 °C, so that the sample broke apart from the joint interface during the applying tensile force. For samples B₁, B₂, and B₃, similar to A₃, due to the absence of Ni and Cr into the joint interface, these samples are fractured in both of brittle mode (at the middle of joint) and ductile mode (at the beginning and ending of S-shaped interface).

4. Conclusions

The thermal durability of AA1050-St37 joints produced by friction stir welding (FSW) was investigated, with a focus on the effect of an intermediate layer of SS316L deposited on the steel surface. Two types of joints—with and without the intermediate layer—were heat-treated at 400 °C and 550 °C for 90 min. The thickness of the intermetallic compounds (IMCs) and the fracture behavior of both joints were analyzed. The key findings are summarized as follows:

• IMC Suppression by Intermediate Layer: The intermediate layer, created by depositing the St37 faying surface with SS316L, significantly retarded IMC formation during FSW, resulting in an IMC thickness of less than 100 nm.
• Improved Strength at 400 °C: Annealing at 400 °C increased the strength of the joint with the intermediate layer from 325 MPa (as-welded) to 350 MPa, with no observable IMC growth. This improvement is attributed to the release of residual stresses.
• Strength Reduction in Conventional Joints: In contrast, the joint without the intermediate layer experienced a decrease in strength from 225 MPa (as-welded) to 150 MPa after annealing at 400 °C. This decline is attributed to the elimination of irregularities and protrusions in the IMC layer, which reduced its toughness.
• High-Temperature Performance at 550 °C: Heat treatment at 550 °C increased the IMC thickness in the joint with the intermediate layer to 13 µm, causing a sharp decline in strength from 350 MPa to 100 MPa. However, this strength is still significantly higher than that of the conventional joint, which lost all strength under the same conditions.
• Enhanced Thermal Durability: The use of an intermediate layer prior to FSW significantly improves the thermal durability of Al-St joints, making them safer and more reliable for high-temperature applications, even at critical temperatures above 500 °C, where conventional joints fail entirely. This is attributed to hindered diffusion rate caused by the presence of Cr and Ni in the intermediate layer.

These findings demonstrate the effectiveness of the intermediate layer in mitigating IMC growth and enhancing the thermal durability of Al-St joints, offering valuable insights for their application in high-temperature environments.