Abstract

The principal aim of the present research work was to compare the role of dispersoids Al2O3 (∼15 μm average particle size) and SiC (∼15 μm average particle size) with that offered by Zr- and Ni-based intermetallics (∼35–70 μm average particle size) on the hardening of cast aluminum 354 alloy (9.1% Si, 0.12% Fe, 1.8% Cu, 0.008% Mn, 0.6% Mg, and 87.6% Al) at ambient temperature. There is no observable poisoning effect on the refinement of grain size after the addition of Zr to the alloys investigated in this study. The tensile test results were examined in light of the microstructural features of the corresponding alloy samples. The contribution of the added dispersoids or Ni and Zr alloying elements on the tensile properties of the 354 alloys was determined employing ∆P plots (where P = Property, UTS, YS, or %El), using the base alloy (in the as-cast condition) as a reference point. The tensile results were supported by investigating the precipitation-hardening phases using scanning and transmission electron microscopy as well as examining the fracture surfaces of selected conditions applying field emission scanning electron microscopy. The results show that, in all cases, Al2Cu phase is the main hardening agent. The contribution of about 1.5 vol% of SiC or Al2O3 to the strength of the base alloy is higher than that offered by Zr- and Ni-based intermetallics, under the same aging treatment.

1. Introduction

Dispersoids are defined as small and numerous finely divided particles of one substance dispersed in another. Dispersoid particles are normally hard and brittle, resulting in excellent dispersion strengthening of the softer aluminum matrix [1]. According to Srinivasan and Imam [2], dispersoids are thermodynamically stable second-phase particles. Their size is in the range of 0.1 to 1.0 μm, which is made use of to reduce the alloy grain size. Fracture of these particles under loading can lead to the nucleation of cracks ahead of the main advancing crack. The effect of Cr and Mn addition and heat treatment on AlSi3Mg casting alloy was investigated by Tocci et al. [3] who found that Cr- and Mn-based dispersoid particles improve Vickers microhardness of the aluminum matrix compared to the base alloy after solution treatment and quenching, which is in accordance with the dispersion-hardening mechanism. In a similar study, Remøe et al. [4] recommended decreasing heating rates which would increase the dispersoid density.

Dispersoids were also described as particles that form during ingot preheating caused by precipitation of the transition elements Cr, Mn, or Zr in the form of Al12Mg2Cr, Al20Cu2Mn3, or Al12Mn3Si and Al3Zr particles, respectively [5]. A significant increase in mechanical properties of the tested hypoeutectic alloys was obtained by supersaturation of the Al phase with high melting point elements during HPDC processing of AlSi9Cu3(Fe) alloy as reported by Szymczak et al. [6]. A study of dispersoid particles in two Al–Mg–Si 6xxx-type aluminum alloys and their effects on the recrystallization was conducted by Hichem and Rebai [7]. Their main conclusion was that the incorporation of Si in the metastable Al3Zr type of dispersoid particles could increase the nucleation rate by reducing the volume-free energy of formation of the dispersoid particles. Kenyon et al. [8] added that changing the dispersoid volume fraction, size, and morphology has important implications for the pinning effectiveness of the dispersoids.

It is necessary to produce a microstructure containing thermally stable and coarsening-resistant dispersoids in order to enhance the mechanical properties of an aluminum alloy, in particular at high temperatures. The dispersoids or particles play an important role in resisting coarsening if the energy of their interface with the matrix, diffusivity, and solubility is low. The Al3Zr particles are resistant to dissolution and coarsening; they can also control the evolution of the grain and subgrain structure, thereby making it possible to increase strength and ductility in the precipitation-hardened T6 condition [912]. The effectiveness of the dispersoids depends on their size, spacing, and distribution. In direct-chill cast alloys, the alloying elements are highly segregated following solidification [1315].

According to Szajewski et al. [16], three different types of dislocation-precipitate interactions could occur depending on the nature of the precipitates, i.e., whether they are coherent or noncoherent. Strengthening by coherent precipitates is caused by distortion in the crystal lattice, representing barriers to dislocation movement [1720]. Several research articles [6, 8, 2123] emphasized that binary dispersoids such as Al3Zr can be less effective due to limitations encountered during conventional manufacturing processes. Improvements can be achieved through the transformation of these binary trialuminides into ternary or quaternary forms by alloying with appropriate elements.

Previously, the present authors discussed the effect of addition of SiC and Al2O3 particulates on the microstructure and tensile properties (mainly quality index values as a function of thermal treatment) of Al–Si–Cu–Mg cast alloys [23]. In addition, the authors analyzed the effect of different alloying elements on the performance of this family of alloys over several publications [2428]. The subject was readdressed in the present work to highlight the difference in the effectiveness of dispersoids, mainly Al2O3, SiC and Al3Zr, compared to intermetallics offered by the addition of Zr and Ni in the amounts of 0.25 wt.% and 0.5 wt.%.

2. Experimental Procedure

Alloy 354 modified with 200 ppm strontium and grain-refined using 0.25 wt% Ti (added using Al-5%Ti-1%B master alloy) was used as the base alloy (coded alloy A). The chemical composition in wt. % of the as-received 354 alloy ingots is listed in Table 1.

2.1. Additions
(i)The addition of Al2O3 particles was carried out using a metal matrix composite (6061 alloy + 20 vol % Al2O3) having a 2-to-30 μm particle size.(ii)The addition of SiC particles was done using a (359 alloy + 20 vol % SiC) metal matrix composite having a similar particle size.(iii)Zirconium and nickel additions were made using Al-20wt% Zr and Al-20wt% Ni master alloys, respectively.

Alloy A, with a chemical composition as shown in Table 1, was melted in an electrical furnace at 780°C, using a 60 kg SiC crucible. The molten alloy was degassed using pure, dry argon injected into the melt for 20 min by means of a graphite impeller rotating at 135 rpm. Grain refining and modification of the melt were carried out using Al-5% Ti-1% B and Al-10% Sr master alloys, respectively, to obtain levels of 0.25 wt.% Ti and 200 ppm Sr in the melt. Table 2 lists the chemical composition of the alloys which were investigated. The melt was poured into a preheated ASTM B-108 permanent mold (preheated to 460°C), to prepare test bars; the dimensions of which are shown in Figure 1. Three samples for chemical analysis were also taken at the time of the casting; this was done at the beginning, in the middle, and at the end of the casting process to ascertain the exact chemical composition of each alloy.

As will be mentioned in a later subsection, heat treatment of the test bars used for tensile testing involved solution heat treating them at 495°C for 8 hours, followed by quenching in warm water at 60°C, after which artificial aging was applied. In the present work, three aging temperatures (155°, 190°, and 350°C) and different aging times, up to 1000 h were used. After aging, the test bars were allowed to cool naturally at room temperature (25°C). All heat treatments were carried out in an air-forced Lindberg Blue M electric resistance furnace. All of the samples, whether as-cast, solution heat-treated, or aged, were tested to the point of fracture using an MTS servo-hydraulic mechanical testing machine at a strain rate of 4 × 10−4 s−1.

Ten samples from each condition were tested, for a total of 80 conditions for each alloy. Three other conditions were tested for each alloy corresponding to 200 h, 600 h, and 1000 h aging times, from which would be selected one specific aging temperature (190°C); this will be explained in detail further on in the text. The values for the mechanical properties, namely, ultimate tensile strength (UTS), yield strength (YS), and percent elongation (% El), were gathered from the computerized system of the MTS machine.

Samples for metallography were sectioned from the tensile-tested bars of all the alloys studied, about 10 mm below the fracture surface; they were then individually mounted in Bakelite and subsequently polished to a fine finish using 1 μm diamond suspension. The instrument used in this study was a Hitachi-SU-8000 FESEM, equipped with a standard secondary electron detector (SE), a backscatter electron detector (BSD), and an energy dispersive X-ray spectrometer (EDS). The FESEM was operated at a voltage of 15 kV, with a maximum filament current of 3 amperes.

Transmission electron microscopy was used in order to observe and identify the strengthening precipitates in heat-treated samples and also to investigate the coherency of the precipitates with the matrix. The FEI Tecnai G2 F20 electron microscope employed was equipped with an advanced control system which permits the integration of an EDAX™ chemical analysis system, scanning transmission electron microscopy (STEM), and electron energy loss spectroscopy (EELS). The microscope was operated at an accelerating voltage of 200 kV.

3. Results and Discussion

In the present work, we are referring to the Al2O3 or SiC particulates as dispersoids in composites B and C containing these particulates since these particulates have no solubility in the matrix, maintaining at the same time uniform spacing. On the contrary, intermetallics are phases/compounds made up of two or more elements, producing a new phase with its own composition, crystal structure, and properties (Alloys D-G). Alloy A is the base alloy.

3.1. Microstructural Characterization

Figure 2(a) shows the microstructure of alloy A in the as-cast condition, revealing almost no presence of porosity. The blue arrows point to the change in the morphology of dendrites from elongated ones to a more rounded (rosetta-like) form which enhances feeding of the interdendritic regions, reducing the porosity in the grain-refined alloy as clearly displayed in Figure 2(b), a high-magnification micrograph of Figure 2(a). According to Campbell and Tiryakiog˘lu [29], based on the work of Fuoco et al. [30], at an early stage of solidification, the dendrites can move as a slurry. Later, when eutectic freezing starts, the dendrites are fully solidified and fixed (about 45% remaining liquid) so that significant mobility of the eutectic is to be expected. However, after adding Sr, no mobility of the residual liquid is observed even at the high level of 45%. This explanation, in addition to the change in the dendrite shape presented in Figure 2(b), contradicts the model presented by Argo and Gruzleski [31] which suggests that, in unmodified alloys, the eutectic is characterized by its irregular solid/liquid interface, leading to entrapping of small pockets of liquid between advancing solidification fronts, causing the formation of microporosity. Figures 2(c) and 2(d), respectively, exhibit the distribution of the added Al2O3 and SiC particulates throughout the matrix, following solutionizing at 495°C for 8 h (SHT) revealing the complete dissolution of the Al2Cu phase. Figures 3 and 4 reveal the adherence (no debonding) between the particulates and the surrounding matrix.

In alloy D, as the Al3Zr precipitate must be present before the peritectic reaction, the formation of Al3Zr from liquid is called a properitectic reaction [32, 33]. In the present work, the Zr concentration used was 0.2 wt.% and 0.4 wt.% for alloy G and alloys E-F, respectively. From Figure 5(a), the liquidus temperature should be 718°C and 780°C, respectively. The melt was maintained at 780°C, to avoid reaching the softening point of the SiC crucible. Thus, in the present case, Zr-rich phases will be precipitated in the form of intermetallics as well as dispersoids (alloys E and F), whereas in the case of alloy G, most of the Zr will be precipitated in the form of Al3Zr dispersoids.

It should be borne in mind that the present alloys contain at least 15 elements and that the solidification rate was approximately 8°C/s. These two criteria make it difficult to strictly follow the binary equilibrium diagram shown in Figure 5(a). Figure 5(b) depicts the Zr-rich phase distribution in alloy D in the as-cast condition, solidified at a very slow rate ∼0.35°C/s. In this case, the chemical composition of the dispersoid is Al-Zr-Ti as inferred from the associated EDS spectrum displayed in Figure 6(a). In the presence of Ti in the matrix, Zr has a strong affinity to react with Ti forming Al3(Zr, Ti)5 [3436] or Al3(Zr1−xTix) dispersoids [36]. Figure 7 depicts the precipitation of Al3Zr dispersoids in alloy D in the as-cast condition, in a sample prepared from a tensile test bar solidified at about 8°C/s. According to Ikhlaq et al. [34], Zr has a preferred orientation (002) along other significant diffraction planes (100), (101), (102), and (110) of Zr confirming the hexagonal crystal structure as shown in Figure 6(b).

Figure 8 displays the progress in the size and distribution of the Al2Cu precipitation hardening particles (main hardening agent) at 190°C (peak aging). It is evident from Figures 8(b) and 8(c) that with increase in the aging time, there is an increase in the size of the precipitated Al2Cu particles coupled with change in their morphology from spherical to short rods, which leads to a gradual decrease in contribution of these precipitates to both UTS and YS regardless the alloy type, reaching almost nil after 1000 h [37]. Figure 8(d) shows dislocation/precipitate interactions; note the significant increase in the dislocation density compared to that after solutionizing treatment (Figure 8(e)). Figure 8(f) presents the EDS spectrum, corresponding to the white square area in Figure 8(c), indicating reflections due to Al and Cu.

3.2. Tensile Properties

Table 3 lists the tensile parameters of alloy A in the as-cast condition. Figure 9 presents the contribution of added particulates or alloying elements on the tensile properties of the 354 alloys B through G, using alloy A in the as-cast condition as a reference point. It should be mentioned here that, in a previous study, the authors reported on the aging behavior of alloy A in the temperature rage 155°C–350°C for times up to 100 h [38].

As can be seen in Figure 9(a), when the alloys were aged at 155°C for 2 h, composite B (containing maximum 1.5 vol% Al2O3) contributed significantly to the UTS of alloy A (about 120 MPa) followed by a linear decrease up to alloy E (40 MPa) beyond which other alloys (alloys F and G) showed a slight increase in UTS, going to 56 and 60 MPa, respectively. Aluminum oxide particulates are a ceramic compound with a hexagonal crystal lattice. The oxygen anions define a hexagonal close packed structure, and the aluminum cations occupy 2/3 of the octahedral sites in the hcp lattice. It is used for its hardness and strength, about 9 on the Mohs scale of mineral hardness (just below diamond) [34]. Although SiC (composed of tetrahedra of carbon and silicon atoms with strong bonds in the crystal lattice) is a very hard and strong material [39], its contribution is slightly lesser than that of Al2O3 which may be interpreted in terms of the actual particulate volume fraction. However, both composites B and C offered a better resistance to softening, yield strength, as well as a better improvement in the alloy ductility compared to those obtained from the other alloys containing Ni or Zr. Figures 9(b) to 9(d) demonstrate the variation in the UTS, YS, and % elongation, respectively, when both particulates and alloying elements were added to alloy A. The tensile bars were exposed to aging at 190°C for times up to 1000 h.

A comparison of Figures 9(b) and 9(c) shows that the contribution level to the alloy YS is higher than that to the UTS. Although the contribution to the ductility of alloy A in Figure 9(d) followed a similar pattern as those for UTS and YS, a sharp decrease in the %El was observed for alloys F and G (1000 h). On the contrary, low aging times (2 h–100 h) resulted in a negative change in the % elongation. Some improvement can be seen for alloy G that contains 0.25% Zr + 0.25% Ni producing same levels as composites B and C due to precipitation of Al3Zr dispersoids.

Daoud and Reif [40] suggested that the hardening process is significantly affected by the presence of micro-oxides, carbides, or SiC. The precipitation takes place on the dislocation lines. As a result of the mismatch between the coefficient of thermal expansion of the matrix and the reinforcement, it is not expected to alter the precipitation sequence. Samuel et al. [41] proposed that the relative amount of hardening precipitates is also influenced by the added dispersoids. In order to achieve a significant improvement in the alloy strength, the authors recommended increasing the reinforcement particulates by an order of magnitude than those used in composites B and C. Apparently, the added amounts of Al2O3 or SiC in the present study have a marginal effect on the composite tensile properties compared to Zr and Ni addition.

As will be observed from Figure 10, the highest standard deviations with respect to UTS are exhibited by the alloys containing micro-oxides; also, the variations are more accentuated for the temperatures where the maximum strength values were observed, i.e., at 155°C and 190°C. This figure is showing the average standard deviation values for all alloys studied, where the higher standard deviation values may be observed for the properties pertaining to composites B and C. In Figure 11, the maximum standard deviations correspond to the maximum ductility values, observed at 155°C and 350°C for alloys containing micro-oxides/carbides. In both Figures 10 and 11, the tensile test values corresponding to these two temperatures reveal the highest standard deviations, particularly for composites B and C, when the seven alloys used in this study are subjected to comparison.

3.3. Fracture Behavior

Figure 12(a) represents the fracture surface of alloy A in the as-cast condition (secondary electron image) where severe cracks (arrowed blue) can be seen propagating through the fine-dimpled structure (1.3% elongation). Solutionizing at 495°C for 8 h resulted in a significant increase in the alloy ductility (6.3%) which reflected on the alloy surface fracture as displayed in Figure 12(b) where a deep dimple structure is seen predominating the fracture surface-blue arrows. Slip bands are observed covering the wall of the dimples shown in Figure 12(b). Figure 12(c) exhibits the fracture surface of alloy A aged at 190°C for 2 h revealing severe cracks (white arrows), whereas similar features as those shown in Figure 12(b) were reported covering the fracture surface when the alloy was aged at 350°C for 100 h (overaging), see arrows in Figure 12(d). In composite C, Figure 13, the presence of micro-SiC particulates were seen at the bottom of the coarse dimples (white arrows) when the alloy was solutionized (reaching 5.3% ductility). In all solutionized alloys, clear slip bands were commonly observed as shown by the blue arrows in the figure. The fracture surface of alloy D (containing 0.4%Zr) after solutionizing treatment resulted in enhancing the alloy ductility, reaching about 4.5%. While the SHT did not change the shape of the star-like particles, Figure 14(a) shows several slip lines clearly visible on the edges (black arrows) of the particle, as well as a long crack reaching the star center (white dashed arrow). The associated EDS spectrum, Figure 14(b), exhibits reflections due to Al, Si, Zr, and Ti elements suggesting that the composition of this phase is (Al,Si)2(Ti,Zr) [28].

Figure 15(a) demonstrates the fracture of alloy E in the as-cast condition. It is evident that the presence of such a significant amount of insoluble Ni-rich phases would cause marked brittleness (1.44%El). The associated EDS spectrum in Figure 15(b), corresponding to the marked spot in (a), shows strong peaks of Al, Ni, Cu, and Si with a small peak of Mg indicating a mixture of Al3CuNi and Q-Al5Cu2Mg8Si6 phases. The fracture surface of alloy F (containing about 0.4%Ni + 0.45Zr) is illustrated in Figure 16(a) showing a fractured massive Zr-rich phase particle. The black arrow points to parallel markings on the surface of the particle. The associated EDS spectrum corresponding to the area marked X in (a) reveals that this phase is possibly (Al,Si)3(Ti,Zr), depending on the Si/Zr intensity ratio [42].

Based on the EDS spectra presented in Figures 7(c), 14(b), and 16(b), the Si/Zr intensity ratios are 0, 0.5, and 1, respectively, depending on the Zr-rich phase particle size. According to Garza-Elizondo et al. [43], the (Al,Si)2(Ti,Zr) phase contains 36 wt.% Si and 38 wt.% Zr, whereas the (Al,Si)3(Ti,Zr) phase contains 11 wt.% Si and 47 wt.% Zr, reaching Al3Zr when Si is practically nil. The atomic radii of Al, Si, Ti, and Zr are listed in Table 4. The Al3Zr phase is a closed packed structure (Figure 17(a)) with an interatomic spacing of 0.441 nm. However, there is controversy regarding the particle structure which has been reported as being cubic Ll2 or tetragonal DO23, as shown in Figure 17(b) [42, 43]. It is also reported that Al3Zr has the stable DO23 structure, but for small supersaturations of the solid solution, Al3Zr precipitates with the metastable L12 structure and precipitates with the DO23 structure only appear for prolonged heat treatment and high enough supersaturations [44, 45].

It is suggested that since the Si atom has a small radius (Table 4) compared to the interatomic spacing in Figure 17(a), the Si atoms can easily diffuse into the Al3Zr cells during the solidification process. The concentration of Si in the Al3Zr phase particle is a function of the temperature and the size of the formed Al3Zr phase particle. In other words, the larger the size of the Al3Zr phase particle, the lesser is the Si concentration, as shown in Figure 16(b). Considering the atomic radius of Ti (0.145 nm) is very close to that of the Al atom (0.143 nm), this would facilitate the substitution of a certain number of Al atoms by Ti atoms.

4. Conclusions

Based on the results obtained in this study, the following conclusions may be drawn:(1)There is no observable poisoning effect on the refinement of grain size after the addition of Zr to the alloys investigated in this study.(2)The Si/Zr ratio varies depending on the size of Zr-rich phase particles.(3)The standard deviation (SD) of the UTS and %El values is higher in SiC and Al2O3 particulate-containing alloys due to their low volume fraction and tendency for segregation.(4)In order to achieve a lower SD, the volume fraction of Al2O3 or SiC particulates should be increased by an order of magnitude.(5)The results show that, in all cases, Al2Cu phase is the main hardening agent. The contribution of about 1.5 vol.% of SiC or Al2O3 particulates to the strength of the base alloy is higher than that those offered by Zr- and Ni-based intermetallics.(6)The reduction in mechanical properties due to addition of different elements is attributed principally to the increase in the percentage of intermetallic phases formed during solidification; such phase particles would act as stress concentrators, decreasing the alloy ductility.

Data Availability

Data will be available upon request.

Conflicts of Interest

There are no conflicts of interest between the authors.