Effect of Heat Treatment on Microstructure and Thermal Fatigue Properties of Al-Si-Cu-Mg Alloys

Specimen preparation and mechanical properties test

The experimental alloys were prepared in a 3 kg capacity SiC crucible by mixing commercial master alloys of Al-30%Si, Al-50%Cu, Al-10%Mn, Al-5 %Cr, Al-10%Ni, Al-10%Sr, Al-5 %Zr, Al-10%RE and pure metals, Mg, Zn and V to achieve the targeted chemistry. The chemical compositions of the alloys were shown in Table 1. The alloys were melted at 720±5 ℃ in an electrical resistance furnace under a protective atmosphere. Before the experiment, each element needs to be burden calculation with the percentage of them. After melt, the alloys also need to be refining, degassing and slagging off. Then the molten alloy should kept at 720±5 ℃ for another 20 mins before pouring into the steel mold with the size of 200×150×60 mm. The mold was preheated to a temperature of 220 ℃ during the whole casting process.

In order to confirm the appropriate temperature in the solution heat treatment, it is necessary to identify the melting temperature of the copper-rich phases. In this study, the differential thermal analysis and differential scanning calorimetry (DTA and DSC, respectively) techniques were applied to figure out the melting temperature of the alloy during the heating cycle. Figure 1 shows the DSC curve of as-cast alloys. This curve indicated that the copper-rich phases began to dissolved at 509.2℃ which was the melting point of the non-equilibrium eutectic phase. Considering the onset melting temperature of the copper-rich phases, the single grade solution temperature of the alloys should be no higher than 509.2℃. In this experiment, the experiment samples were kept at 485℃ for 4 h and then immediately quenched into warm water at 80℃. The artificial aging was performed at 180℃ for 0, 2, 4, 6, 8, 10, 12 h, respectively. After aged, the test specimens were cooled naturally and the mechanical performances were tested at room temperature. The heat treatment conditions were summarized in Table 2.

Figure 1:

DSC curve for heating process. Copper-rich phase start to melt at 509.2°C.

The micro-hardness measurements were carried out on polished samples with the HV-1000 micro-hardness tester under a load of 200 g with a holding time of 30 s. The microstructure observation was conducted on optical microscopy (OM) and the scanning electron microscopy (SEM) equipped with energy dispersive X-ray detector (EDX). The test samples were prepared according to standard metallographic procedures, which were etched using Keller’s reagent (2.5 ml HNO₃, 1.5 ml HCl, 1 ml HF, and 95 ml H₂O). To examine the effect of aging time on the mechanical properties of alloys, the tensile test was conducted by a CMT5205e universal electronic tensile testing machine with an elongation rate of 1 mm/min at room temperature. Figure 2 shows the shape and the dimensions of tensile specimens, which were prepared according to the ASTM E-8 standard.

Figure 2:

Shape and dimensions of tensile specimens (unit:mm).

Fatigue test

The fatigue test was conducted by the self-restraint thermal fatigue testing machine. During the experiment, the sample temperature was controlled by thermocouple and the counting device was applied to calculate the cycle times. In order to make the process of heating and cooling automatic, the reciprocating motion of the fatigue test specimen was dominated by computer. The fatigue test specimens were heat preservation at electric resistance furnace for 120 s and then putted into the cool water cooling for 10 s immediately. The temperature of experiment was preliminary design to 350℃ and the temperature of cooling water was about 20℃ during the thermal cycling.

The fatigue test specimen had a 3 mm deep V-notch which was prefabricated by artificial and the detailed parameter of the thermal fatigue specimen was shown in Figure 3. The purpose of V-notch was to bring an area of stress concentration where the crack would appear initially. Under the stress of thermocycling, the fatigue crack length (Optical Microscopy observations) and the plastic deformation around the V-notch were used as a standard to evaluate the effect of microstructure variation on the thermal fatigue resistance. The fatigue crack propagation path of specimen was observed by optimal microscopy after each 1000 cycles.

Figure 3:

Fatigue specimen geometry.

The microstructure and mechanical properties after heat treatment

Figure 4 shows the relationship between the mechanical properties and aging time in the experimental alloys. The tensile strength of as-cast specimen was at the level of 285 MPa. By aging treatment, the tensile strength increased from 360 MPa(aging for 2 h) to 375 MPa(aging for 6 h) and the elongation decreased from 5 % to 4.5 %, respectively. In comparison with the as-cast alloys, the micro-hardness and tensile strength reached to the maximum after aging for 6 h at the aging temperature and decreased subsequently. When the aging time was longer than 6 h, the tensile strength and elongation decreased with the coarsening of precipitation phase.

Figure 4:

Mechanical properties of Al-Si-Cu-Mg alloys under different aging time.

It has been reported that aging introduced precipitation hardening and solution treatment caused morphology changes are two kinds of methods to elevate the strength of the alloy [12]. Figure 5 shows the microstructure of the Al-Si-Cu-Mg alloys at different treatment condition. Instead of the fibrous morphology observed in the as-cast microstructure, the eutectic Si exhibited spheroidal morphology with an apparent segregation in the eutectic areas. As observed in Figure 5(c)–(d), the eutectic silicon particles were refined and spheroidized under artificial aging treatment. The formation of the finely dispersed eutectic silicon particles within the grains deplete the split action from long needle eutectic Si to the matrix [13, 14, 15] that was identified as the mechanism of mechanical properties improvement as shown in Figure 4. Additionally, from the observation of the microstructure micrographs in Figure 5, the mechanical properties of the alloys depend more on the microstructural changes such as the area fraction of eutectic silicon particles, distribution and the amount of the precipitation phase [16].

Figure 5:

Optical micrographs of Al-Si-Cu-Mg alloys: (a) as-cast state; (b) solution treated; (c) aging-treated at 180℃ for 2 h; (d) aging-treated at 180℃ for 6 h.

A general view of selected specimens in as-cast and heat-treated conditions, fractured surface after tension were shown in Figure 6. For the as-cast fracture, there were less tearing ridges on fracture surface, as shown in Figure 6(a). No evident cracks were observed on the cleavage plane, implying a typical brittle fracture characteristic. In Figure 6(b), the aging treated tensile fracture consists of equal-axis ductile voids basically and the morphology of dimples shows deep and small. There were also numerous cleavage steps distribute around the tearing ridges and most of the tearing edges were connected end to end with a mass of dimples on it [7, 17]. The higher magnification image of red circle on the fracture surface in Figure 6(b) was shown in Figure 6(c). The amount of dimples on the fracture morphologies of aging treated samples was larger than that of as-cast alloy and the fracture surface was consisted of cleavage planes and dimples on it. In the process of tensile deformation, the secondary phase particles crack into the dimple sources [18]. Moreover, the flow stress in the matrix has increased a lot and a higher level of stress was transferred to the eutectic Si area in the tensile test. In addition, cracks primarily initiated and propagated along the interfaces of eutectic Si area and the fracture of the alloy generally started at brittle phase area that has caused the fracture of the materials in Al-Si-Cu-Mg alloys as shown in Figure 6(c). Some secondary phase particles were also found in the bottom of dimples at the same time. With the stress growing, the dimples were finally split up and the protuberant tearing cracks were formed with large deformations around the dimples. Combined with the microstructure of the fracture surface, the fracture of the alloys were belonging to the mixed fracture among the intracrystalline failure and intergranular fracture caused by secondary phase [18, 19].

Figure 6:

SEM micrograph showing the tensile fracture of Al-Si-Cu-Mg alloys (a) as-cast alloy; (b-c) aging-treated alloy.

Figure 7 shows the XRD spectra of the as-cast material(1#), the solution treated specimen(2#), the aging for 2 h (3#) and 6 h(4#) at 180℃ after solution treatment. As revealed by the mechanical properties in Figure 4 and XRD patterns in Figure 7(1# and 2#), the precipitation phases has been dissolved into the aluminum matrix, such as Al₂Cu and Mg₂Si, and finally promotes the formation of supersaturated solid solution matrix during the heat treatment. The UTS and micro-hardness value has been increased significantly in Figure 4, that’s because the dissolution of precipitation phase particles lead to the morphology change of matrix and elevate the mechanical properties through the aging treatment. However, the value of the elongation decreased in comparison with those who were only treated by solution treatment. The X-ray diffraction spectra of the aged treated alloy (3# and 4#) also illustrates some secondary phases appeared again after aging treatment. Considering the micrographs in Figure 5 and the XRD patterns in Figure 7, it can obviously indicated that the maximum solution strengthening has achieved whether the Al₂Cu phase is dissolved or not, and the result also presented the variation of mechanical properties mainly attributed to the precipitation of the Al₂Cu and Mg₂Si in some extent.

Figure 7:

XRD patterns of the test specimens with different treatment condition: (1#) as-cast state; (2#) solution treated; (3#) aging-treated at 180℃ for 2 h; (4#) aging-treated at 180℃ for 6 h.

Effect of heat treatment on the fatigue properties

Figure 8 clearly presents four kinds of fatigue crack growth behaviors of different samples under 350℃, including crack initiation and extension. In the curve, the rate of fatigue crack growth has decreased obviously and the fatigue resistance improved significantly after heat treatment during the fatigue test. Moreover, the fatigue crack growth rate increased firstly and then decreased with the number of cycles growing. During the thermal cycles, the length of crack about 1# sample reached to 0.1 mm firstly after 9000 cycles and the crack propagate rate was the fastest among these samples. Meanwhile, the crack length about 2# sample reached to 0.1 mm after 11,000 cycles, indicating an obvious improvement of the fatigue resistance. Compared with 1# and 2# samples, the crack length about 3# and 4# samples were shorter and the rate of crack initiation appeared slower with minor-cycle crack extension. This indicated that the fatigue resistance of heat treated specimens was better than that of as-cast alloys. In Figure 4, the comprehensive mechanical properties of 3# and 4# specimens were better than 1# and 2# specimens and the measure results were also in good agreement with the fatigue properties in Figure 8. As portrayed in Figure 5, the grain size appeared little difference under different aging time. With the stronger recrystallization, the grain size increased slightly after aging for 6 h which resulted in the improvement of the driving force for the crack propagation and reduced the rate of fatigue crack growth [20].

Figure 8:

Thermal fatigue crack growth rate of Al-Si-Cu-Mg alloys with different treatment condition: (1#) as-cast state; (2#) solution treated; (3#) aging-treated at 180℃ for 2 h; (4#) aging-treated at 180℃ for 6 h.

Because the thermal stress around fatigue crack is mainly released through crack extension, the crack growth rate decreased after 16,000 cycles. The heat treatment promotes the period of crack initiation becomes longer than the as-cast alloys. When the length of crack reach to some extent, the stress concentration around the crack tip will loose rapidly with the decreases of crack propagation rate on the rusting crack tip [9, 10, 21].

The effect of eutectic Si particles on the fatigue crack growth

During the experiment, thermal fatigue specimens were cooling by water quenching. When the alternating stress was higher than the yield strength of material, the defect was created by the reciprocating glide of dislocation [22]. Additionally, owing to the thermal stress, the stress concentration was created and leading to the fatigue crack propagate in the weak area of the matrix [23]. Present study reveals that the longer aging time leads to the lower increments of the dislocation density. The defect density around the V-notch was higher and the crack was firstly initiated when the stress concentration reach to some extent. As mentioned above, because of primary silicon particles were hard and brittle in nature, they usually act as a significant role in the fatigue crack growth behavior and naturally developed into the prioritized regions of the fatigue crack propagation in Al-Si-Cu-Mg alloys [22, 24]. On the other hand, the cracks may come from the debonding of Si particles, and then propagate through the boundaries with α-Al phase.

Figure 9 shows the fatigue crack propagation behavior of as-cast alloy and heat treated alloy after 12,500 thermal cycles. The length of fatigue crack for the as-cast alloy reaches to 223.18 μm and the fatigue crack propagated along the eutectic Si particles is shown in Figure 9(a), but it’s only 132.58 μm for the T6 alloy as shown in Figure 9(b). Compared with the as-cast alloy, the amount of branching eutectic Si was decreased after aging treatment and turned into spheroidized finally. These changes not only decreased the negative impact on α-Al matrix but also had a function of dispersion strengthen and elevate the flexibility, reduce the route of cracking initiation, and offer more resistance by crack deflecting around Si particles. Observation of microstructure revealed that, during the heat treatment of Al-Si-Cu-Mg alloys, the averages size of eutectic Si is more refined from 20–25 μm (as-cast alloy) to 12–15 μm (heat-treated alloy). Besides, the eutectic Si is mostly in the strip morphology with the much scattering size distribution in the heat-treated alloy. The length of the strip is measured to be ~6 μm which is greatly different from the as-cast alloy. The study of the Al-Si alloy has found the interconnected networks of eutectic Si and primary α-Al acting as the main strengthening phase during the thermal fatigue test. The decreasing of sizes and aspect ratios of eutectic Si through heat treatment lead to the remarkably reduction of fatigue resistance. Griffith equation gives the relation between the intrinsic fracture stress (σ_f) on the primary silicon and the internal defects length (C).

Figure 9:

Fatigue crack propagate along the eutectic Si particles (a) as-cast alloy after 12,500 thermal cycles; (b) aging treated alloy after 12,500 thermal cycles.

γ, the fracture surface energy; E, the Young’s modulus of the particles. According to the Griffith equation, the internal defects of coarse primary Si crystals are much longer than fine Si crystals, which result in lowering the intrinsic fracture stress (σ_f). On the other hand, coarse irregular primary Si and flake-like eutectic Si present sharp edges or ends, which are stress concentration sites and preferred cracks initiation. With the eutectic particles spheroidized, the modification of eutectic Si structure can decrease or eliminate premature cracks initiation and propagation at the same time.

The effect of secondary phase on the mechanism of fatigue crack propagation

Some papers pointed out that the precipitates phases play an important role in the nucleation position and the probability of crack initiation becomes greater with the size of the secondary phase growing [23, 25]. In this experiment, because of the difference between the expansion coefficient and heat conductivity coefficient among the matrix and secondary phase, the thermocycling would easily lead to stress concentration at the interface of two-phases. After the critical stress break the interfacial bonds between the particles and the ductile matrix, voids are formed and finally turned into micro-crack with the thermal-cycle going.

Figure 10 shows the SEM images and EDS spectrums of the as-cast alloy and heat-treated samples. It can be observed in Figure 10(a) and (b), the crack was normally grown up in the weak area among the matrix and secondary phase. Observation of EDX spectrums of secondary phase, Cu and Mg were precipitated as Al₂Cu and Mg₂Si phases, but the amount and distribution appeared different. At the initial stage of crack growth, with the stripping of the secondary phase, the micropores were easily initiated and leading to the formation of micro-crack. At the same time, a large amount of crack extension power was consumed with the movement of dislocation impeded by coarse surplus-phase. The rate of crack growth is finally retarded and result in the slowdown of the crack propagate rate, as verified in Figure 8. In the fatigue crack initiation period, the fatigue crack normally exists along the primary phase and eutectic Si particles, the trace of crack propagation usually depletes by the primary phase which was precipitate distributed during the heat treatment. In the stage of fatigue crack propagation, with the solution of solute atoms(Cu, Mg) through the heat treatment, the evolution of primary phase leads to the growing of fatigue resistance. Figure 10 illustrates the fatigue cracks are mainly initiated around the coarse surplus-phase on the surface of specimen. From the observation of microstructure and fatigue crack growth curve, the fraction of surplus-phase also leads to the same effect on the fatigue crack growth rate and the probability of the crack initiation has decreased with the surplus-phase dissolved. The crack was normally propagated through the fast channel which provided the surplus-phase, and the fatigue crack growth rate was increased with the volume fraction of surplus-phase growing [24]. Moreover, the brittle surplus-phase particles on the grain boundary often fracture and stripping from the matrix under the cyclic loading and finally promote the fatigue crack formation .

Figure 10:

SEM micrographs and EDS analyses of crack initiation and propagation behavior around the secondary phase in Al-Si-Cu-Mg alloy (a) as-cast; (b) aging-treated; (c) Al₂Cu phase; (d) Mg₂Si phase.

The mechanism of fatigue crack propagation

Figure 11 illustrates the process of a typical fatigue crack propagate via weak regions where the microvoids and the loose regions initiated around the V-notch. The observation on the SEM micrographs of fatigue crack shows the effect of porosity on the fatigue life of T6 alloys. In Figure 11(a), the micropores around the loose region were appeared after the critical stress break the combination among the α-Al matrix and grain interface. The formation of microvoids was the process of vacancy segregation where the thermal stress and temperature gradient reached to some extent [21, 25]. The mounts of microvoids were growing with the number of cycles increasing. After that, the microvoids turned into micro-crack with the depth of voids growing. In Figure 11(b), some microvoids were formed around the loose region where the binding force between the ductile matrix and the second particles was broken by the critical stress. Considering the microstructure of test specimens in Figure 11 and the fatigue crack growth curve in Figure 8, this phenomenon clearly indicates the period of crack initiation were gentle compared to fatigue crack extension, that’s because enough tension-compression stress was needed to break the combination among the matrix. With the cycling thermal stress going, the micro-crack appeared with the formation of micropores around the secondary-phase particles and most of the surface porosity acting as the main crack initiation site of fatigue test samples.

Figure 11:

Cracks initiating at the microvoids and loose regions.