Aluminum Cast Alloys Based on the Al–Ce–Si–Mg System: An Influence of Silicon on Crystallization, Phase Composition, and Tensile Properties

The Al5Ce3Si0.5Mg and Al5Ce0.5Mg (wt. pct) alloys were cast in a pilot scale in-house foundry using a resistance furnace with a steel crucible coated with a mica wash under argon cover gas. The charge makeup with a weight of 15 kg comprised of 99.9 wt. pct purity elemental ingredients. After melting, a liquid alloy was held at 780 °C to 800 °C for 0.5 h under intense mixing, followed by measurements and adjustments of the chemical composition. Next, the dross/oxide was removed from the melt surface and the crucible content was poured at 730 °C into a water cooled steel mold with internal dimensions of approximately 25 × 152 × 305 mm (1 × 6 × 12 inch), having an internal surface covered with boron nitride. Following solidification, the cerium concentration was determined by the Inductively Coupled Plasma (ICP) method in accordance with MCLM F-23C and BAERD-GEN-018E specifications. The contents of remaining elements were measured by the Optical Emission Spectroscopy (OES). Detailed chemical compositions are listed in Table I.

The solidification characteristics of alloys were assessed using the Computer-Aided Cooling Curves Thermal Analysis (CCTA) associated with measurements by a Universal Metallurgical Simulator and Analyzer (UMSA).^[10] During thermal analysis, controlled heating to 780 °C was performed at a rate of 0.4 °C/s, followed by isothermal holding at 780 °C for 5 min and natural cooling to 50 °C at a rate of 0.2 °C/s. Metallographic samples were prepared using the conventional procedures, starting from grinding to polishing. No etching was applied and as-polished surfaces were examined with an optical microscope Olympus BX534 and scanning electron microscope FEI Nova NanoSEM650. The transmission electron microscopy was conducted using FEI’s Tecnai Osiris TEM equipped with X-FEG gun at 200 keV. Specimens for TEM were prepared by grinding to a thickness of 100 µm, cutting discs with a diameter of 3 mm and ion milling. The STEM mode with bright field (BF) and high angle annular dark field (HAADF) detectors were employed in a combination with energy dispersive spectrometry (EDS). The EDS with the ESPRIT software, applying a deconvolution of overlapping peaks was used for mapping the elemental distribution. EDS acquisition was performed with all spectrum images having a size of 256 × 256 pixels and 2048 energy channels. Each spectrum image used a pixel size of 34 pm and a 0.01 keV/channel energy dispersion. A single frame spectrum image was acquired with a per pixel dwell time of 5 ms. The computational thermodynamic analysis of phase composition under equilibrium and non-equilibrium (Scheil) cooling conditions was conducted using the FactSage software with the FTlite database. The phase composition was measured using PANalytical X’Pert Pro X-ray diffractometer in the powder diffraction mode, equipped with copper sealed tube source radiation, λ_CuKα = 1.5418 Å. The data were collected with the PANalytical Data Collector software. The analysis was performed in Match! software using the crystallography open database (COD). The room temperature uniaxial tensile test was conducted in accordance to the ASTM E8 standard using rectangular samples machined from as-cast plates using the Instron tensile test machine (Model 5969, 50 KN) at the strain rate of 0.4 mm/min.

A thermodynamic description of the ternary Al–Ce–Si phase diagram available in the literature^[11,12] was developed using the CALPHAD method incorporating the limited experimental data. The computed isothermal section at 500 °C is shown in Figure 1(a). According to this diagram, two three-phase equilibria exist in the Al-rich corner: (1) (Al) + Al₁₁Ce₃ + τ₁ and (2) (Al) + τ₁ + τ₂, where τ₁ is Ce(Si_1-xAl_x)₂ being an extension of the binary phase CeSi₂ into the ternary region with x = 0 to about 1 at high temperatures and x = 0.1–0.9 at 500 °C, and τ₂ is AlCeSi₂.

To determine the phase composition for the specific alloy chemistry used in this work, including small quantities of minor phases formed after solidification, the FactSage commercial software was used. Additional purpose of using the FactSage calculations was to estimate melting temperatures required during the alloy synthesis. FactSage uses the thermodynamics calculations to predict the solidification behavior of the alloys under equilibrium and Scheil cooling conditions. Since during casting and solidification experiments the cooling rate is high so that the diffusion does not happen uniformly in the solid material, which results in segregation what is an equivalent to the Scheil cooling condition in FactSage calculations. Under Scheil cooling conditions that assumes no diffusion in the solid but complete mixing in the liquid, as a result of solute rejection to the remaining liquid, the liquid metal saturates, causing decreasing the solidus temperature. Therefore, the solidification ends at lower temperature in comparison to the equilibrium condition. The resulting microsegregation profiles and some phases predicted under Scheil cooling condition may not exist under equilibrium assumptions.

The FactSage calculated projection of the Al–Ce–Si diagram in Al rich corner at 500 °C is shown in Figure 1(b) and projection of the liquidus temperatures is shown in Figure 1(c). According to the FactSage calculations, the Al5Ce3Si0.5Mg alloy with the specific chemistry listed in Table I should reach the solidus at 547.75 °C. Phases predicted in the Al5Ce3Si0.5Mg alloy by FactSage from a Scheil solidification calculation with all available solution phases and pure solid phases enabled are listed in Table II. It is seen that the vast majority of alloy volume of 93.4 vol. pct represents aluminum solid solution with 0.51 at. pct (0.53 wt. pct) Si and 0.30 at. pct (0.27 wt. pct) Mg. A content of 0.0019 at. pct Ce indicates practically a lack of solid-state solubility of Ce in Al. The minority phases represent 5.59 vol. pct of AlCeSi₂ along with 0.73 vol. pct Mg₂Si and 0.25 vol. pct of Si.

The objective of X-ray diffraction was to determine the bulk phase composition of alloys, to verify a possible presence of metastable phases and to determine the eventual change in Al lattice parameter due to alloying. The X-ray diffraction pattern of the Al5Ce0.5Mg alloy, measured using a Cu radiation with λ_CuKα = 1.5418 Å, is shown in Figure 2(a), where Al and Al₁₁Ce₃ phases are marked according to 96-431-3218 and 04-002-7472 cards, accordingly. Additions of Si led to a replacement of the Al₁₁Ce₃ phase by aluminum cerium silicate AlCeSi₂ with peaks marked according to the 99-900-001 card in Figure 2(b). The diffraction peaks seem also to indicate the Al₂CeSi₂ phase. A presence of expected Mg₂Si phase is difficult to definitely confirm since according to the 96-153-7741 card, key Mg₂Si peaks coincide with AlCeSi₂ peaks and peaks at 2-theta 47.433 deg and 72.901 deg are the only unique indicators. The change in the Al peak intensity at 2-theta of 65.17 and 77.50 deg, corresponding accordingly to (220) and (311) Al in both alloys is most likely caused by the preferred grain orientation (crystallographic texture) formed during solidification, and different orientation of samples used for X-ray diffraction.

A comparison of X-ray diffraction patterns from both alloys examined revealed a shift of Al peaks in Al5Ce3Si0.5Mg towards lower 2-theta values than in Al5Ce0.5Mg. The key observation is that a presence of 0.5 wt. pct Mg in the Al5Ce0.5Mg alloy led to an increase in lattice constant as compared to pure Al. An addition of Si in the Al5Ce3Si0.5Mg alloy, however, reduced the lattice constant back to the level similar to pure Al.

Thermal events taking place during alloys solidification were recorded using the computer-aided cooling curves thermal analysis. For a cooling curve that represents the balance between evolution of heat in the sample and the heat flow away from the sample, solidification start-up is determined by the latent heat changed by a liquid–solid transformation. The cooling curve for the Al5Ce0.5Mg alloy, plotted along with the first derivative dT/dt as a function of time, is shown in Figure 3(a) where proeutectic solidification of primary Al and the eutectic solidification Al + Al₁₁Ce₃ are the major thermal events. The similar plot for the Al5Ce3Si0.5Mg alloy is shown in Figure 3(b).The first peak of dT/dt, aligned with a change of the slope on cooling curve, is associated with beginning of solidification of the proeutectic αAl (solid solution) phase. The change in slope of the cooling curve around 1900s is associated with the start of univariant eutectic reaction L → α(Al) + AlCeSi₂ + L. The second large peak of derivative dT/dt, aligned with a beginning of the short plateau on cooling curve, around time of 2270 s, indicates the invariant quaternary eutectic reaction of L → α(Al) + Si + Al₂MgSi/Al₂Mg₂Si + Al–Si–Mg-Ce. To determine precisely temperatures of thermal events, dT/dt curves for the Al5Ce3Si0.5Mg alloy and the Al5Ce0.5Mg alloy were plotted as a function of temperature (Figure 3(c)). For the Al5Ce0.5Mg alloy the distinct peak at 648 °C (liquidus) marks the start of primary Al solidification and the second peak at 638 °C is associated with the binary eutectic transformation L → Al + Al₁₁Ce₃. A presence of Si reduced the liquidus temperature to 630 °C, then the univariant eutectic started at 618 °C and invariant eutectic transformation occurred at 539 °C, thus drastically expanding the solidification range from 10 °C to 91 °C. The broad peak at 573 °C is related to solidification of the AlCe₂Si/Al₂CeSi₂ bulky compound. The summary of thermal analysis is listed in Table III.

At a magnification of optical microscope the Al5Ce0.5Mg alloy shows hypoeutectic features, where the proeutectic aluminum is arranged into a dendritic morphology while the eutectic fills interdendritic regions (Figure 4(a)). At higher magnifications, the morphology of the eutectic intermetallic phase is revealed. In contrast to uniform lamellae developed for the binary Al–Ce alloys,^[13] the eutectic phase in the Al5Ce0.5Mg alloy resembles rather a Chinese script with a portion representing plates or rods (Figure 4(b)). For larger contents of Mg and rapid solidification Ce-rich precipitates having near-equiaxed and cellular shapes were reported.^[14]

A presence of Si led to substantial changes in the alloy microstructure. First, there is no dominant morphology of proeutectic aluminum, which was replaced by smaller dendritic islands (Figure 4(c)). The intermetallic compounds exhibit three dark-contrast morphologies; plate-like marked as ternary T eutectic, blocky phases as well as areas with dark and white contrast located in interdendritic regions marked as quaternary Q eutectic (Figure 4(d)). The volume fraction of the latter is small, as revealed earlier through thermal analysis, reaching just 7 pct.

The engineering stress–strain plots, recorded during tensile test at room temperature, are shown in Figure 11 with the sample geometry detailed in the figure inset. The assessment and comparison of the results is summarized in Figure 12. Additions of Si, caused an almost 3 times increase in the yield stress from 47.5 MPa to 136.5 MPa along with a raise in the ultimate tensile strength from 113.5 to 142 MPa (Figure 12(a)). At the same time, however, there was a significant lowering in alloy ductility with 10 times reduction of elongation from 7.7 pct to 0.9 pct, and area reduction from 11.5 to 1.5 pct (Figure 12(b)).

The rupture surfaces of tensile samples were imaged with SEM to determine the mechanisms of failure that operated at a local, i.e. microscopic scale and to correlate them with the microstructural features. In the Al5Ce0.5Mg alloy, the fracture surface exhibits mixed regions; those failed due to void coalescence but also those failed by mechanical instability of the alloy itself (Figure 13(a)). In general, in dimpled regions characteristic of ductile fracture, relatively large amount of plastic deformation is seen that took place before failure, apparently associated with the pre-eutectic aluminum dendrites. Two types of fracture can be distinguished at higher magnifications where regions of brittle fracture resemble the Al₁₁Ce₃ plates or laths arrangement in the eutectic with a separation at the Al₁₁Ce₃/Al interface, as marked in Figure 13(b).

The fracture surface of the Al5Ce3Si0.5Mg alloy exhibits highly brittle nature with very little deformation of the alloy around the fracture (Figure 13(c)). Brittle fracture includes both cleavage and intergranular fracture, where energy is absorbed through regions of small plastic deformation. Since there is the prevalence of the flat plateau regions, they likely are responsible for the relatively low elongation at fracture of 0.9 pct during tensile testing at room temperature (Figure 13(d)). For both alloy compositions, an occasional casting shrinkage porosity was seen with pores promoting failure by instability through acting as crack nucleation sites.