Microstructure, Mechanical Properties, and Fatigue Resistance of an Al-Mg-Sc-Zr Alloy Fabricated by Wire Arc Additive Manufacturing

Abstract

Al-Mg alloy wire modified by Sc and Zr additions was used to prepare a high-strength, non-heat-treated Al-Mg alloy component by wire arc additive manufacturing (WAAM) technology in the present work, and the microstructure, mechanical properties, fatigue resistance, as well as their anisotropies of the deposited Al-Mg-Sc-Zr alloy component were studied. The results show that the microstructure of the as-deposited alloy is composed of fine equiaxed grains with an average grain size of around 8 μm, and nanosized Al₃(Sc, Zr) particles (~5 nm) are also evident. The tensile properties and fatigue resistance of the deposited alloy showed significant anisotropy, and the performance of the traveling direction is always better than that of the deposition direction. The ultimate strength, yield strength, elongation, and critical fatigue life (cycles) of the as-deposited alloy along the traveling direction (0° direction) are 362 ± 7 MPa, 244 ± 3 MPa and 24.8 ± 0.3%, and 1.72 × 10⁵, respectively. The presence of weak bonding areas and high tensile (positive) residual stress between the deposition layers deteriorate the tensile properties and critical fatigue life of the sample along the deposition direction.

1. Introduction

As a lightweight material, aluminum alloys have been widely used in rail transit, aerospace, and other fields due to their many advantages such as low density, high specific strength, and high electrical and thermal conductivity [1]. With the continuous improvement of the industry’s requirements for economy and efficiency, new technologies for manufacturing large-sized and high-strength aluminum alloy parts have become an inevitable trend [2].

Wire arc additive manufacturing (WAAM) technology is a manufacturing technology that uses an arc as a heat source to melt metal wires, and through program control, it stacks the materials layer by layer to form three-dimensional solid metal components. It has the advantages of low production cost, high deposition efficiency, and high material utilization rate [3,4]. Recently, the WAAM of aluminum alloys has received widespread attention, especially for the aluminum alloys of Al-Cu (2xxx series) [5], Al-Si (4xxx series) [6], Al-Mg (5xxx series) [7], Al-Mg-Si (6xxx series) [8], and Al-Zn-Mg (7xxx series) [9]. As a non-heat-treatable strengthening alloy, an Al-Mg alloy cannot be improved in its mechanical properties through a traditional aging treatment [10]. Therefore, how to further improve the mechanical properties of an Al-Mg alloy is the key to preparing the high mechanical performance of non-heat-treated Al-Mg alloy additive components.

The addition of microalloying elements is an important way to improve the mechanical properties of aluminum alloys [11]. Compared with other elements, Sc microalloying has the most significant strengthening effect [12,13]. This is because the addition of Sc not only significantly refines the solidification grain structure of the alloy, but it also forms the Al₃Sc phase. Since the lattice parameters of the Al₃Sc phase are very close to those of the Al matrix, the presence of these Al₃Sc particles results in a significant precipitation strengthening effect and greatly improves the mechanical properties of the Al alloy. In addition, simultaneously adding Sc and Zr into aluminum alloys can form the Al₃(Sc, Zr) phase, which has the advantage of higher thermal stability and a strengthening effect compared with the Al₃Sc phase [14]. Meanwhile, the addition of Zr can reduce the content of Sc, thereby achieving a cost reduction effect [15,16].

The modification of Sc and Zr has also attracted widespread attention in the additive manufacturing of aluminum alloy components [17,18,19,20]. Agrawal et al. [18] prepared an Al-Cu-Sc-Zr alloy by the approach of laser powder bed fusion (LPBF), and the mechanical properties of the alloy were greatly improved due to the modification of Sc and Zr. Guo et al. [19] prepared an Al-Zn-Mg-Cu-Sc-Zr alloy by WAAM technology. The primary Al₃(Sc, Zr) phase plays a role in heterogeneous nucleation during the solidification process, promoting the formation of equiaxed crystals and thus refining grains. The mechanical properties of this alloy after heat treatment exceeded 600 MPa. Hou et al. [20] compared the effect of controlling interlayer temperature and continuous printing on the porosity, microstructure, and mechanical properties of an Al-Mg-Sc-Zr alloy prepared by WAAM. They found that the YS and UTS of the component of controlling interlayer temperature at 100 °C are higher than those of the component of continuous printing, but the porosity of the component of continuous printing is lower.

Fatigue fracture is one of the main failure forms of aluminum alloy components; the fatigue resistance of the alloy components prepared by additive manufacturing is also the focus of attention [21,22,23]. Miao et al. [21] found that the fatigue resistance of a T6-treated Al-Cu alloy fabricated by WAAM is better than that of an as-deposited alloy, although the former one presents higher porosity than that of the latter one. This implies that both the porosity and residual stress play an important role in the fatigue behavior of the Al alloy prepared by WAAM. He et al. [22] reported that the fatigue strength of wire arc additive manufactured 4043 aluminum was improved by the elimination of porosity and Si-rich eutectic network fragmentation via employing interlayer friction stir processing (FSP). Li et al. [23] found that the fatigue life of Al-Zn-Mg-Cu-Si-Zr-Er alloys prepared by LPBF was improved by introducing post-RRA heat treatment.

As a non-heat-treatable aluminum alloy, the fatigue performance of an Al-Mg-Sc-Zr alloy prepared by additive manufacturing has also attracted much attention, and it directly affects the service life of the components. In the present work, a non-heat-treated Al-Mg-Sc-Zr alloy component was successfully prepared by wire arc additive manufacturing technology using the welding wire of an Al-Mg alloy with the addition of Sc and Zr. The aim of the present work is to comprehensively investigate the mechanical properties, fatigue resistance, and their anisotropies of the Al-Mg-Sc-Zr alloy component to evaluate the service performance of the fabricated components. The relationship between the microstructure and mechanical properties as well as fatigue resistance was also focused on. Moreover, the effect of the microstructure and residual stress on the fatigue resistance and its anisotropy was discussed.

2. Material and Methods

In the present work, an Al-Mg-Sc-Zr alloy welding wire was used as the filler metal for the WAAM process, and the wire was manufactured by Hangzhou Kunli Welding Material Co., Ltd. (Hangzhou, China).

Table 1. Chemical compositions of the Al-Mg-Sc-Zr alloy filler wire (wt.%).

Element	Mg	Sc	Zr	Fe	Cr	Si	Ti	Al
Content	6.09	0.28	0.15	0.13	0.1	0.09	0.06	Bal

shows the chemical compositions of the filler wire, which were determined by the method of inductively coupled plasma optical emission spectroscopy (ICP-OES). The WAAM process was conducted on an ARCMAN@ P1200 workstation produced by Nanjing enigma automation Co., Ltd. (Nanjing, China). For the photo and corresponding characteristics of this workstation, one could refer to our previous work [21]. The process parameters of additive manufacturing are shown in

Table 2. The corresponding parameters for the additive manufacturing process.

Welding Wire	Scanning Speed	Wire Feeding Speed	Welding Current
Al-Mg-Sc-Zr alloy	1.8 m/min	8 m/min	160 A

. The welding current is 160 A, the wire feed speed is 8 m/min, and the scanning speed is 1.8 m/min. The final macro-morphology of the component and the sampling position of the test samples are shown in Figure 1a and Figure 1b, respectively.

The conventional mechanical grinding and polishing methods were employed to prepare the metallographic samples, and the optical microstructure was observed by optical microscope (OM, Zeiss microscope Axio Acope.A1, Zeiss, Oberkochen, Germany). A scanning electron microscope (SEM, ZEISS ∑GMA, Zeiss, Oberkochen, Germany) equipped with an electron backscattering (EBSD) system and energy-dispersive spectrometer (EDS) was also employed to characterize the microstructure such as fracture morphology and grain structure, etc. The EBSD samples were prepared via an electrolytic polishing process in the solution of perchloric acid alcohol solution (perchloric acid/alcohol = 1:9). The samples for the Transmission Electron Microscope (TEM) were prepared by twinjet polishing at 20 V using a solution of 90 mL HNO₃ and 210 mL methanol cooled at −25 to −30 °C. TEM images and High-Resolution Transmission Electron Microscopy (HRTEM) images were taken by transmission electron microscopy (FEI TECNAI G2 F20 S-TWIN, FEI, Hillsboro, OR, USA).

In order to study the anisotropy of the tensile properties of the material, tensile tests were carried out on the samples from different directions. Tensile test samples were prepared according to the GB/T 228.1 standard [24], and the size of the samples is shown in Figure 2a. Tensile tests were performed at room temperature at a speed of 1 mm/min on a tensile testing machine, and each sample was tested 3 times to determine the average value and standard deviation. A fatigue crack growth test was also carried out on the samples taken along different directions to study the anisotropy of the fatigue properties of the as-deposited alloy component on the MTS-810 test machine (MTS, Eden Prairie, MN, USA). The size of the samples for the fatigue crack growth test is shown in Figure 2b. The stress ratio is set to R = 0.1 and the frequency is 10 Hz.

The residual stress of the as-deposited alloy component along the traveling direction (0° direction) and the deposition direction (90° direction) was measured by the X-ray diffraction method. The sampling position of the alloy components for residual stress measurement is shown in Figure 3. Five test points of the component along the 0° direction and six test points along the 90° direction were selected for the measurement, and the public test point of the 0° direction and 90° direction was identified as the center point. The intervals (along the same direction) of the measured points were all 5 mm, and there were two points along the 90° direction located on the interlayer area.

3. Results

3.1. Microstructure

Figure 4 depicts the optical and SEM microstructure of the as-deposited Al-Mg-Sc-Zr alloy. It can be found that there are few pore defects with a size of around 20 μm (Figure 4a), which is one of the most common defects in the additive manufacturing of aluminum alloys. The microstructure of the as-deposited alloy is composed of fine and equiaxed grains, and the size of the grains is significantly lower than that of the grains in the as-deposited Al-Mg alloy reported by our previous work [25]. Figure 4b,c show the corresponding SEM image, and there are numbers of the fine second phases with white contrast embedded in the Al matrix. The EDS analysis results of these second phases are summarized in

Table 3. EDS results of the as-deposited samples corresponding to Figure 4c (at. %).

Element	Al	Mg	Sc	Zr	Fe	Si
A1	84.33	5.72	5.48	4.38	0.09	-
A2	88.76	6.03	3.46	1.71	0.04	-
A3	91.33	5.91	0.17	0.29	2.30	-
A4	90.54	6.58	0.27	0.35	2.24	0.02
A5	90.32	5.38	1.10	0.15	0.23	2.82

, which indicates that there are several different types of the second phase. Some of them are mainly composed of Al, Mg, Sc, and Zr elements. It is reasonable to speculate they are the primary Al₃(Sc, Zr) phase, which is a type of effective heterogeneous nucleation site and can significantly refine the grain size of alloys during solidification [26]. The presence of the primary Al₃(Sc, Zr) phase is the main reason for the finer grains in the as-deposited Al-Mg-Sc-Zr alloy compared to in the as-deposited Al-Mg alloy. In addition, there are also some Fe-rich phases and Si-rich phases.

Figure 5 shows EBSD images of as-deposited Al-Mg-Sc-Zr alloy samples in different sampling directions (X-Z plane and X-Y plane). There is no significant difference in the microstructure between the X-Z plane and X-Y plane, and the size of their grains are both even and fine. The average grain size of the X-Z plane is 8.0 μm while the average grain size of the X-Y plane is 7.3 μm.

Figure 6 shows the TEM images of the as-deposited Al-Mg-Sc-Zr alloy prepared by WAAM. Although the corresponding Selected Area Electron Diffraction (SAED) pattern indicates that there is only the signal of the Al matrix, the nanosized second phases are evident in the grain as shown in Bright-Field (BF) TEM images (Figure 6a,b). The HRTEM images shown in Figure 6c,d indicate that the size of this second phase is around 5 nm. The morphology and size imply that these second phase should be secondary Al₃(Sc, Zr) particles, which are precipitated under the condition of a thermal cycle during the process of WAAM. These fine Al₃(Sc, Zr) particles with a sub-nano size were also found in the laser powder bed fusion (L-PBF)-fabricated Al-Mg-Sc-Mn-Zr alloy [27].

3.2. Mechanical Properties

Figure 7a shows the stress–strain curves of the as-deposited Al-Mg-Sc-Zr alloy along the traveling and deposition directions. The stress–strain curves of both the traveling direction (0° direction) and the deposition direction (90° direction) exhibit a zigzag feature, which is often referred to as the Portevin–Le Chatelier (PLC) effect [28]. For the 5xxx series Al alloys, the PLC effect is generally caused by the dynamic interaction of the moving dislocation with the diffused Mg solute atoms [29]. Meanwhile, the tensile test results indicate that the UTS (ultimate strength), YS (yield strength), and EL (elongation) of the sample along the direction of the traveling direction are 362 ± 7 MPa, 244 ± 3 MPa, and 24.8 ± 0.3%, respectively, while those of the samples along the direction of deposition direction are 274 ± 5 MPa, 213 ± 4 MPa, and 7.2 ± 0.2%, respectively. It is clear that the tensile properties of the as-deposited alloy show obvious significant anisotropy, and the comprehensive mechanical properties of the alloy along the traveling direction are better than those of the alloy along the deposition direction. Meanwhile, the ultimate strength and yield strength of the Al-Mg-Sc-Zr alloy are over 100 MPa higher than those of the Al-Mg alloy that was reported by our previous work [25], and the respective elongation is also 3.5% higher.

In addition, Figure 7b and

Table 4. Comparison of the detailed mechanical properties of the present work and reported previous works on the Al-Mg-based alloy prepared by WAAM.

Alloys	UTS (MPa)	YS (MPa)	EL (%)
Al-4.7Mg [25]	258	141	21.3
Al-4.8Mg [30]	286	142	23.0
Al-5.0Mg [31]	263	172	20.9
Al-5.2Mg [32]	257	128	20.2
Al-5.8Mg-0.8Mn [33]	314	164	34.0
Al-6.5Mg-0.7Mn-0.3Sc-0.1Ti [12]	392	221	26.5
Al-6.54Mg-0.36Sc-0.1Zr [19]	336	171	24.8
Al-6.3Mg-0.3Sc-0.1Zr [34]	335	183	22.5
Al-6.09Mg-0.28Sc-0.15Zr (this work)	362	244	24.8

show the comparison of the mechanical properties (ultimate strength and elongation after fracture) between our present Al-Mg-Sc-Zr alloy and other excellent Al-Mg-based alloys fabricated by wire arc additive manufacturing [12,19,25,30,31,32,33,34]. It shows the superiority in the combination of ultimate strength and ductility (elongation after fracture) of the Al-Mg-Sc-Zr alloy in this work. For instance, the comprehensive mechanical properties of the Al-Mg-Sc-Zr alloy in this work are better than those of the Al-Mg-Sc-Zr alloy with higher Mg content and Sc content reported by previous works [12,19], and they are over 28% higher in R_m and 19% higher in A₅ as compared to those of the Al-Mg alloy.

3.3. Fatigue Crack Growth Behavior

Figure 8a shows the relationship between the crack length (a) and fatigue cycles (N) of the samples along two different directions (0° direction and 90° direction), also known as the a-N curve. The results indicate that there are more numbers of cycles in the sample along the traveling direction (0° direction) than in the sample along the deposition direction (90° direction) under the condition of a fixed crack size. The critical crack size of these two samples is similar and is around 21 mm, and the fatigue life (1.72 × 10⁵ cycles) of the sample along the traveling direction is also similar to but slightly higher than that (1.28 × 10⁵ cycles) of the sample along the deposition direction. This implies that the fatigue resistance of the sample along the traveling direction is higher, and this result is similar to that of the as-deposited Al-Mg alloy fabricated by WAAM in our previous work [25]. Meanwhile, the fatigue life (1.72 × 10⁵ cycles) of the as-deposited Al-Mg-Sc-Zr alloy is lower than that (2.56 × 10⁵ cycles) of the as-deposited Al-Mg alloy along the same 0° direction that was reported in previous work [25]. The factors that affected the fatigue resistance of the alloy are mainly associated with microstructure; defects (i.e., pores, cracks, etc.); and residual stress [35]. Since defects such as pores and cracks are not very significant in the alloy component (Figure 4a), the fatigue resistance of the alloy in the present work may be attributed to the microstructure and residual stress. A detailed discussion is conducted in the next section.

Figure 8b shows the relationship between fatigue crack growth (FCG) rate and the stress intensity factor range (i.e., da/dN − ΔK). It can be found that the FCG rate increases by increasing the stress intensity factor. Under the condition of the same $∆K$, the FCG rate of the sample along the traveling direction is significantly lower than that of the sample along the deposition direction. It is therefore shown that the fatigue life of the sample along the traveling direction is higher. The Paris model is employed to further fit the FCG rate in the stable expansion region [22]:

$$da/dN={C(∆K)}^m$$

(1)

where C and m are material constants, which are determined by the fitting results of experimental data. The parameters after fitting and fatigue threshold value (ΔK_th) are shown in

Table 5. Fitting parameters.

Sample	C	m	R	ΔKth
0° direction	2.09 × 10−7	3.12	0.957	0.790
90° direction	1.71 × 10−6	2.43	0.938	0.309

. Table 5 suggests that the coefficient R of the samples along the 0° direction and 90° direction are both above 0.9, indicating a good agreement with the experimental data. In addition, there is significant difference in the ΔK_th value of the alloy along two different directions (0° and 90° directions). The ΔK_th value of the sample along the 0° direction is 0.79 MPa·m^1/2, which is 2.5 times the ΔK_th value of the sample along the 90° direction (0.31 MPa·m^1/2). The hindrance of the FCG of the alloy along the traveling direction is significantly greater than that of the alloy along the deposition direction.

Figure 8c,d show the macro-morphologies of the fatigue samples along the traveling direction (0° direction) and the deposition direction (90° direction), respectively. It can be found that the fatigue crack path in the sample along the traveling direction (0° direction) is relatively straight and almost parallel to the notch direction, but in the sample along the deposition direction (90° direction) it is slightly offset. This also explains the fluctuation in the (da/dN − ΔK) curve in the deposition direction (90° direction) in Figure 6b. The above results show that the fatigue properties of the alloy also show obvious anisotropy, and the fatigue performance of the alloy along the traveling direction (0° direction) is better than that of the alloy along the deposition direction (90° direction).

The processes of fatigue crack failure mainly include three stages: crack initiation, crack propagation, and crack instability, and crack propagation is considered to be one of the key factors to affect fatigue life. Figure 9 shows the fatigue fracture morphology in the crack stable growth zone of two samples. Figure 9a,c depict that the fracture surface has a layered structure and is uneven, which is because the deviation in grain boundary orientation hinders the expansion of fatigue cracks, and therefore large numbers of fatigue steps are formed. The fracture surface roughness and the secondary crack numbers are also different in two samples. The fracture surface in the sample along the 0° direction is rougher and has more secondary cracks than that in the sample along 90° direction. The secondary cracks suppress the crack growth rate by consuming the energy of the crack tip [36] and therefore improve fatigue resistance. Figure 9b,d show the fatigue fringe morphology of the samples along two different directions, and the corresponding fatigue fringe widths are 0.24 μm and 0.68 μm, respectively. Under the condition of alternating stress, fatigue fringes are initiated by continuous crack passivation and sharpening modes [37]. The spacing of fatigue fringes represents the distance of crack propagation in a loading cycle. The smaller the width of the fringes, the slower the crack growth rate and the better the fatigue resistance. The fatigue fringe widths (0.24 μm) of the sample along the 0° direction suggests that the fatigue resistance of the sample along the 0° direction is better than that of the sample along the 90° direction, which is in good agreement with the results shown in Figure 8.

4. Discussion

The results of the as-deposited alloy showed significant anisotropy in tensile properties and fatigue resistance. Figure 10 depicts the tensile fracture morphology of the samples along two different directions. For the fracture morphology of the sample along the 0° direction (Figure 10a,c), there are numbers of small and dense dimples, which are a typical characteristic of a ductile fracture. On the other hand, in addition to a small number of dimples, a large area of flat platform with smooth morphology appears on the fracture surface of the sample along the 90° direction (Figure 10b,d). This is due to the poor interfacial binding strength between layers, which is identified as a weak binding area with poor interfacial binding strength [25]. The presence of these weak binding regions in the sample along the 90° direction reduces the area of bearing capacity and therefore reduces the elongation and deteriorates the yield strength of the alloy along the 90° direction, resulting in significant anisotropy of the tensile mechanical properties.

The fatigue crack growth results also indicate that the fatigue resistance of the alloy along the traveling direction (0° direction) is significantly better than that of the alloy along the deposition direction (90° direction). Since the fatigue crack growth path of the sample along the 90° direction passes through the interlayer and inner layer area, the weak banding regions between the deposition layers may also be one of the reasons for the decrement in critical fatigue life.

In addition, residual stress is one of the main factors that affects the fatigue resistance of metals and alloys [38,39,40]. Generally, the presence of residual stress in an as-deposited alloy is a typical characteristic of additive manufacturing components due to rapid solidification during the process of additive manufacturing. The residual stress of the samples along different directions (0° and 90°) is needed for consideration and analysis. Figure 11 shows the residual stress distribution along the 0° and 90°directions. It can be found that the inner layer region of the sample presents the residual compressive stress along the traveling direction (0° direction), and there is no significant difference in the stress levels that all ranged from −50 MPa to −40 MPa. On the other hand, the residual stress of the sample shows significant differences along the deposition direction (90° direction). This difference is essentially the difference in the residual stress between the intralayer region and the interlayer region. Similarly to the sample along the 0° direction, the inner layer region of the sample along the 90° direction also presents a residual compressive stress that ranged from 40 to 50 MPa. However, the interlayer region of the sample presents the residual tensile stress, and the stress level ranged from 80 MPa to 90 MPa, which is significantly higher than that of the inner layer region. Residual tensile stress in the interlayer region can lead to a higher average combination of applied stress in the deposition direction (90° direction), which has an adverse effect on fatigue resistance. Meanwhile, the residual compressive stress will reduce the average combined stress applied in the direction of travel (0° direction), and it has a positive effect on fatigue resistance. This can be supported by previous works [39,40], which found that residual compressive stress hindered the growth rate of fatigue cracks, thus extending the fatigue life of alloys. Thus, the presence of weak bonding areas and high tensile residual stress between deposition layers play the dominant role in deteriorating the critical fatigue life of the sample along the deposition direction (90° direction).

5. Conclusions

The microstructure, mechanical properties, fatigue resistance, and their anisotropies of the Al-Mg-Sc-Zr alloy prepared by WAAM were investigated. The main conclusions are summarized as follows:

(1)

The microstructure of the as-deposited Al-Mg-Sc-Zr alloy is composed of fine equiaxed grains with an average grain size of around 8 μm due to the formation of the primary Al₃(Sc, Zr) particles with a sub-micron size. Meanwhile, the nanosized Al₃(Sc, Zr) phase with a size of about 5 nm is also evident and plays an important role in strengthening the alloy.

(2)

The mechanical properties of the as-deposited alloy show significant anisotropy, and the comprehensive mechanical properties of the traveling direction are better than those of the deposition direction. The UTS, YS, and EL of the sample along the 0° direction are 362 ± 7 MPa, 244 ± 3 MPa, and 24.81 ± 0.3%, respectively, while those of the sample along the 90° direction are 274 ± 5 MPa, 213 ± 4 MPa, and 7.22 ± 0.2%. The weak bonding zones are found on the fracture surface of the sample along the 90° direction and may have a negative impact on the mechanical properties.

(3)

The critical fatigue life of the as-deposited alloy along the traveling direction (0° direction) is 1.72 × 10⁵, which is better than that of the alloy along the deposition direction (90° direction). The presence of weak bonding areas and high tensile (positive) residual stress between the deposition layers deteriorate the critical fatigue life of the sample along the deposition direction (90° direction).