Excessive Fe Contamination in Secondary Al Alloys: Microstructure, Porosity, and Corrosion Behaviour

Abstract

The characterisation of aluminium casting alloys with iron concentrations exceeding current standards is essential, as upcycling has recently become a significant concern in achieving a more circular economy. Secondary aluminium casting alloys often exhibit insufficient mechanical properties for load-bearing automotive applications due to contamination with iron, mainly due to alloy mixing or remnants from end-of-life products during downcycling. This trend is anticipated to soon lead to a surplus of scrap. This study aims to fully understand the microstructural changes, intermetallic phase morphologies, and defect formation in AlSiMg alloy highly contaminated with Fe that exists in Al scraps and is detrimental for upcycling purposes. The investigation examined the AlSi7Mg0.3 alloy with Fe concentrations ranging from 0.1 to 3.8 wt.% Fe, employing thermodynamic simulations, hardness testing, quantitative image analysis, and corrosion tests. Among these alloys, the AlSi7Mg0.3-3.8Fe, containing the highest level of contamination, exhibited the most complex microstructure. This microstructure is characterised by the presence of two distinct Fe-rich intermetallic phases with diverse shapes and sizes: petal-like α′-Al₈Fe₂Si, long and thick β-Al_4.5FeSi plaques, and very thin β-Al_4.5FeSi needles. The significant growth in these phases with higher Fe concentration resulted in increases in hardness (15 HBW), porosity (1.39%), and corrosion rate (approximately 12 times).

1. Introduction

The presence of iron (Fe) in aluminium (Al) alloys, while common, can significantly degrade their properties and hinder the recyclability of these materials. The presence of Fe in Al alloys is found in some commercial alloys such as AlSi7Mg0.3 (A356), AlSi9Cu3 (A380), AlFe4, for easing de-moulding, increasing strength and thermal conductivity, etc. Additionally, using Fe has even been explored as an alloying element in the development of heat sinks via additive manufacturing [1,2,3,4]. However, the presence of unwanted Fe in secondary Al alloys (SAAs), typically derived from shredded end-of-life automobile parts (known as twitch feedstock), poses a challenge. The excessive Fe content is often addressed by diluting it with costly and environmentally harmful primary Al during the production of these secondary materials [5]. Thus, mitigating the detrimental effects of Fe has become critical to achieving a more sustainable, circular economy for Al products [6,7]. Moreover, the presence of Fe is widely recognised as harmful, prompting numerous studies [8,9,10,11,12,13,14] focused on the effects of contamination on Al alloys’ microstructure and mechanical performance. Several other works [15,16,17,18] have tried to reverse, remove, and neutralise the impact of this contamination on commercial secondary Al alloys, which is essential for minimising downcycling and promoting upcycling of the SAAs. However, the downcycling of these alloys, coupled with quality bottleneck challenges, threatens to create a surplus of scrap. By 2030, this surplus could reach 5.4 million tons, with up to 3.6 million tons of Al scrap becoming unusable by 2040 if new, more effective recycling methods are not developed [15,19,20]. Consequently, it has become increasingly clear that the circularity of Al, particularly regarding the recyclability of casting alloys, is at risk [19,21].

The contamination of SAAs by Fe has a significant impact on ductility, with effects varying depending on the concentration of Fe. Generally, as Fe concentration increases beyond standard thresholds, it can lead to ductility reduction caused by the formation of brittle intermetallic phases and other structural changes, such as increased porosity [22]. For example, in the A380 with an excess of only 0.5 wt.% Fe of maximum concentration allowed by the EN 1706 standard [23], the elongation at break (A%) decreases to 0.1% [10]. Fe-rich intermetallic particles exhibit greater susceptibility to fracture under tensile stress than other eutectic or intermetallic particles and the Al matrix [21]. These particles serve as preferential paths for macrocrack propagation and frequently act as initiation sites for microcracks.

Numerous studies have investigated the impact of varying Fe concentrations on the properties of AlSi7Mg0.3 alloy. Karabulut et al. [24] explored Fe levels up to 0.65%, observing decreases in yield strength (YS), ultimate tensile strength (UTS), and A% alongside more brittle fracture surfaces, indicative of semi-brittle behaviour. Zhao et al. [25] systematically examined Fe-rich intermetallic compounds, noting a transition in the dominant phase from the π-AlMgFeSi phase to the β-AlFeSi phase with increasing Fe concentration. Kuchariková et al. [13] investigated Fe concentration up to 0.66 wt.%, finding that higher Fe levels promoted the formation of thicker interconnected Fe plate-like phases, increasing porosity. Tunçay et al. [12] analysed Fe amounts ranging from 0.2 to 1.2 wt.%, observing reductions in UTS and A% with increasing Fe concentration. Notably, β-AlFeSi intermetallic had a more pronounced negative effect on UTS than the α-AlFeSi phase, notably when these particles exhibited needle-like shapes with sharp edges.

Recognising this critical relationship between Fe and Si in Al alloys, Taylor et al. [22] introduced Equation (1), which calculates the critical Fe concentration threshold value in an Al-Si alloy. This threshold is critical, as surpassing it can lead to a substantial loss of ductility in the final castings. Thus, commercial alloys tend to remain well below this limit.

Fe_crit ≈ 0.075·[wt.% Si] − 0.05

(1)

In the case of the AlSi7 alloy, the Fe critical percentage is approximately 0.48 wt.%, which explains why most scientific studies avoid investigating higher contamination levels. Only a few studies have been conducted on heavily contaminated alloys, particularly those exceeding the critical level. As a result, it seems uncertain about the challenges and potential for upcycling in controlling the morphologies of intermetallic phases in Al-7Si-Fe alloys once the Fe concentration surpasses the established critical limit.

To investigate the upcycling potential of heavily contaminated Al-Si alloys, it is crucial to have a thorough understanding of the phases and morphologies that develop when the Fe content exceeds the critical threshold. This study systematically characterises the microstructure of the AlSi7Mg0.3 alloy at elevated Fe concentrations above 1 wt.%, which is higher than the typical levels found in commercially available alloys, compared with the low contamination alloys. The focus is on the morphological and size evolution of the intermetallic phases and their impact on porosity. Additionally, this research aims to conduct a more detailed evaluation of the class of intermetallic that form with increasing Fe content, with hopes of identifying critical points for optimising these materials. This will facilitate the future incorporation of highly iron-contaminated alloys in the Al foundry sector.

2. Materials and Methods

The initial phase of this investigation involved simulating incremental contamination of AlSi7Mg0.3 alloys with steel scraps, leading to excessive concentration of Fe. Microstructures were observed via optical and scanning electron microscopy (SEM). Phases were identified using Kikuchi patterns via electron backscatter diffraction (EBSD) and X-ray diffraction (XRD). Additionally, a quantitative analysis of images was conducted for alloys with high Fe levels, distinguishing the differences in intermetallic shape and sizes. The evolution of intermetallic phases is further discussed and correlated with alterations in porosity and hardness observed in the alloys and thermodynamic cooling simulations. The corrosion behaviour of these alloys was evaluated to determine the effect of Fe-rich intermetallic phases on degradation mechanisms.

2.1. Alloy Production

The production of the different alloys consisted of initially creating a master alloy with the highest Fe concentration possible. This master alloy was subsequently added to a new melt of AlSi7Mg0.3 in controlled portions to achieve the desired concentrations.

To produce the master alloy, several 304 stainless steel bars were placed in approximately 4 kg of the AlSi7Mg0.3, which was melted at 800 °C for about 20 h. The resulting melt, highly Fe-concentrated, was then cast into an ingot.

Different alloys with varying Fe contents were produced by incrementally adding controlled portions of the master alloy to a new AlSi7Mg0.3 bath, which was melted at 720 °C. This melting process was conducted in an Ar-protective atmosphere to prevent excessive hydrogen pickup and oxidation, eliminating the need for degassing. After each addition of the master alloy, samples were cast after waiting 15 min for homogenisation. The samples were made by pouring the melt into steel cups for further analysis, including a reduced pressure test (RPT). For these tests, two samples were cast simultaneously into cups around 100 g; one sample solidified in a normal atmosphere, while the other was placed in a vacuum chamber to solidify under 80 mbar for 5 min. The chemical compositions of the alloys, analysed using a SPECTROMAXx spark optical emission spectrometry analyser (SPECTRO, Kleve, Germany), are presented in

Table 1. Alloys chemical composition in wt.%.

Sample	Alloy	Si	Mg	Fe	Mn	Cr	Ti	Ni	Zn	Al
Master		6.41	0.27	6.41	0.06	0.06	0.09	0.04	0.03	Bal.
1	AlSi7Mg0.3-0.1Fe	7.44	0.34	0.14	-	-	0.11	0.01	-	Bal.
2	AlSi7Mg0.3-0.4Fe	7.49	0.33	0.36	-	0.01	0.11	0.01	-	Bal.
3	AlSi7Mg0.3-1.3Fe	7.32	0.31	1.30	0.01	0.01	0.11	0.02	-	Bal.
4	AlSi7Mg0.3-2.1Fe	7.00	0.29	2.07	0.02	0.02	0.12	0.02	0.01	Bal.
5	AlSi7Mg0.3-3.8Fe	6.74	0.26	3.80	0.04	0.04	0.12	0.03	0.03	Bal.

.

RPT was conducted to assess the density differences between samples solidified under normal atmospheric conditions and those solidified under vacuum, as the latter tends to reveal defects. Following density measurements, the density index (DI) was calculated using Equation (2):

$$\mathit{DI}={\displaystyle \frac{{\rho }_1-{\rho }_2}{{\rho }_1}}\times 100$$

(2)

where ${\rho }_1$ is the density of a sample cast at 1 atm, and ${\rho }_2$ is the density of a sample cast at the reduced pressure. Additionally, other factors were also studied, including the bifilm index (BI) calculated using Equation (3). This index represents the total estimated length of bifilms, a typical defect of Al castings caused by the entrapment of double oxide films divided by the area of the sectioned surface of the reduced pressure test samples.

$$\mathit{BI}={\displaystyle \frac{\sum \mathit{maximum} \mathit{length} \mathit{of} \mathit{pores}}{\mathit{area} \mathit{sectioned} \mathit{surface}}}$$

(3)

2.2. Thermodynamic Simulations

Thermodynamic simulations were conducted using the CALPHAD methodology, employing FactSage Education 8.3 software [26] with the FTlite Database. The Scheil–Gulliver cooling calculation within the Equilib programme was utilised, employing the Al-Si-Fe composition alloys mentioned above.

2.3. Materials Characterisation

Cross-sections of the samples were prepared using conventional metallographic techniques. This process involved grinding the samples to 1000 mesh SiC sandpapers and polishing with 6 and 1 μm diamond paste, concluding in a final polishing step using 0.3 μm colloidal silica suspension. Microstructural characterisation was conducted using an optical microscope (OM), specifically the DM6 M (Leica, Wetzlar, Germany) using the LAS X software 5.2.1 (Leica, Wetzlar, Germany), as well as a scanning electron microscope (SEM), using back-scattered electron (BSE) image mode, the FEI Quanta 400 FEG (Thermo Fisher Scientific, Hillsboro, OR, USA). Energy-dispersive X-ray spectroscopy (EDS) was employed for localised chemical composition analysis, utilising the EDAX Genesis X4M (Oxford Instrument, Abingdon, Oxfordshire, UK).

XRD analysis was performed utilising an AXS D8 Discover diffractometer (Bruker, Karlsruhe, Germany). The X-ray source utilised CuKα radiation (1.54060 Å) generated at 40 kV and 40 mA served. Diffractograms were collected between 20° and 90°, with a step size of 0.02° and an integration time of 1 s. The Bruker AXS DIFFRAC.EVA V4.2.2 software was utilised for crystal phase identification based on the ICDD database. Phase identification was complimented using EBSD TSL-EDAX detector unit (EDAX Inc. (Ametek, Mahwah, NJ, USA), mainly for index Kikuchi patterns with the EDAX-TSL Delphi version 3.2 software.

Quantitative analysis of images, including porosity, secondary dendritic arm spacing (SDAS), intermetallic phase size, and porosity, was performed using FIJI 1.54 [27] with BAR [28]. Hardness testing was performed with 20 indentations for each alloy and/or phase. The NP EN ISO 6506-1:2015 standard [29] was followed for Brinell hardness testing using the EMCO-TEST DuraVision 20G5 macro hardness tester (ZwickRoell, Ulm, Germany) with a 2.5 mm diameter ball indenter, and a test force of 612.9 N was carried out to evaluate the overall alloy’s hardness. Vickers microhardness tests were conducted using the EMCO-TEST DuraScan G5 equipment (ZwickRoell, Ulm, Germany), following ISO 6507-1:2023 [30] with a test load of 0.09807 N to evaluate the hardness of each individual phase.

2.4. Corrosion Test

The corrosion resistance of the specimens was evaluated by immersing them in a 3.5 wt% NaCl solution for 12 h, following the standard ASTM G1-03(2017)e1 [31]. Before testing, the surfaces to be exposed were ground and polished using 100 to 1000-grit SiC abrasive paper. The samples were then placed in a 95 wt% ethanol solution and subjected to ultrasonic cleaning for 5 min. The specimens were submerged in a concentrated nitric acid (70%) solution for 1 min, followed by an additional 5 min of ultrasonic cleaning in ethanol to remove any corrosion products. The mass loss was measured by weighing the alloy before and after immersion using an analytical balance with a precision of 0.0001 g. All experiments were conducted at a constant temperature of 25 °C. The corrosion rate (CR) was calculated using Equation (4).

$$\mathit{CR}={\displaystyle \frac{K\times W}{A\times T\times D}}$$

(4)

In the equation, K represents a constant that allows the calculation of the CR in millimetres per year, equal to 8.76 × 10⁴. The variables are defined as follows: W is the mass loss in grams, A is the exposed area in square centimetres (cm²), T is the time exposed in hours, and D is the density of the alloys in grams per cubic centimetre (g/cm³).

3. Results

3.1. Solidification Simulation

The Scheil–Gulliver cooling simulation illustrated in Figure 1 predicts the precipitating phases in each sample and delineates the solidification sequence. As the Fe concentration increases, the onset temperature of the first phase solidification increases considerably when the concentration exceeds 1.3%. For alloys with lower Fe concentration (AlSi7Mg0.3-0.1Fe and AlSi7Mg0.3-0.4Fe), as well as AlSi7Mg0.3-1.3Fe, solidification of α-Al occurs first, occurring around 613 °C and 602 °C, respectively. However, AlSi7Mg0.3-1.3Fe exhibits a distinct solidification sequence, with the β-Al_4.5FeSi formation phase preceding the binary eutectic reaction, L → α-Al + Si.

The two alloys containing the highest Fe concentration exhibit distinct solidification pathways. In the case of the AlSi7Mg0.3-2.1Fe alloy, solidification starts at 625 °C with the precipitation of β-Al_4.5FeSi rather than α-Al. Subsequently, α-Al formation begins at 605 °C, a temperature similar to that observed in the other alloys. With an increase in Fe concentration to 3.8 wt.%, solidification initiates at 669 °C, forming a different phase: α′-Al₈FeSi. This phase formation stops at 617 °C, and β-Al_4.5FeSi begins to precipitate, followed immediately by α-Al solidification at around 611 °C.

In the meantime, the temperature at which the liquid phase disappears is always 577 °C for all alloys, and it is attributed to the eutectic reaction, resulting in the formation of α-Al + Si + β-Al_4.5FeSi from the liquid phase.

Figure 1f depicts the solidified phase fraction as a function of the Fe concentration. A pronounced phase transition is evident, shifting from the conventional alloy phase composition to a significant increase in intermetallic phases, reaching approximately 15% for the alloy containing 3.8 wt.% Fe. It is essential to highlight that these simulations were limited to the Al-Fe-Si system due to constraints within the database. Consequently, other intermetallic phases influenced by the presence of Mg have not been explored and may potentially emerge in the microstructure.

3.2. Microstructural Characterisation

3.2.1. Low Fe Alloys

The microstructures observed by OM of the two alloys with the lowest Fe concentration are displayed in Figure 2a,b. These microstructures exhibit similarities, characterised by a distinct dendritic structure comprising primary α-Al dendrites surrounded by eutectic phase (α-Al + Si). The secondary dendritic arm spacing (SDAS) shows a slight increase with the rise in Fe concentration, from an average of 51.3 (95% CI [47.3, 55.2]) to 56.8 (95% CI [51.1, 62.5]). The eutectic Si assumes a plate-like morphology due to the absence of any modification treatment, as seen as light grey needles in Figure 2c,d.

Upon SEM examination, Fe-rich intermetallic phases were more readily noticeable owing to the atomic contrast provided by the BSE imaging mode. While these phases typically manifest as acceptable, small needles in the commercial alloy, a slight increase of approximately 0.3 wt.% Fe noticeably increases both their number and size. Further magnification (as seen in Figure 2c,f) revealed the presence of other phases nestled between these needles. Elemental analysis conducted via EDS, detailed in

Table 2. EDS elemental analysis (in at.%) of the points identified in Figure 2.

Point	Si	Fe	Mg	Al
P1	1.37	-	-	98.63
P2	70.43	-	-	29.57
P3	19.52	15.85	-	64.62
P4	28.70	6.29	15.95	49.06

, clarifies the composition of these distinct phases. The needles primarily consist of the β-Al_4.5FeSi phase, while the intertwined particles constitute the π-Al₉FeMg₃Si₅ phase [25]. It was also observed that the latter phase grows from the β-Al_4.5FeSi needles, exhibiting a branched morphology suggestive of the Chinese script shapes. Additionally, the Si phase was observed to form from the intermetallic phases. Adjacent to the β needles in both alloys, a ternary eutectic comprising Al, β-Al_4.5FeSi, and Mg₂Si was also detected with latter phases with a spot-like morphology in the Al matrix, seen as black spots in Figure 2f.

3.2.2. High Fe Alloys

Figure 3a–c presents the optical observations of the AlSi7Mg0.3-(1.3–3.8 Fe) alloys. In this context, the variation in intermetallic phases (appearing in dark grey) is distinctly visible. Figure 3a,b showcase short needle intermetallic; however, the latter shows a significant increase in number. Meanwhile, Figure 3c includes irregularly shaped phases and long, thick needles. The overall microstructure reveals a disruption in the dendritic structure for the two alloys with the highest contamination level, while the eutectic Si retains its plate-like morphology. The porosity tends to increase with Fe content as well. The quantitative analysis of the changes will be presented later.

In the AlSi7Mg0.3-3.8Fe alloy, the intermetallic phases presented as needles with two distinct dimensions: thick, long needles and small, thin ones. Petal-shaped particles were also observed. Through EDS analysis (Figure 4), the phase of the long and thick needles was also confirmed to be β-Al_4.5FeSi. The petal-shaped phases exhibited a slightly different elemental composition, approximating the stoichiometry of the α′-Al₈Fe₂Si phase [2]. Notably, in the AlSi7Mg0.3-2.1Fe alloy, these petal-shaped phases were in smaller numbers and sizes, detailed in the following sub-section. It is worth highlighting that the growth of β-Al_4.5FeSi tended to originate from these petal shapes with a needle-like morphology.

Observation of the deep-etched samples, as shown in Figure 5, was conducted to analyse the 3D morphology of the particles. In Figure 5a, within the region with the observed thin needles, these structures seem to be clusters of thin plates with thicknesses below 2 µm. Figure 5b showcases examples of a thick and long needle morphology. Here, it is evident that these needles exhibit more considerable length and thickness compared to their thinner counterparts. Upon closer examination of the particles, it becomes apparent that the sides of the plates have impressions of the α-Al dendritic structure. This observation suggests these phases likely form simultaneously as the formation of α-Al dendrites. Moreover, examining the morphology of the α′-Al₈Fe₂Si petals reveals that these structures, in reality, comprise a close-packed agglomeration of several thin plates, as highlighted by a yellow ellipse in Figure 5c. These plates have grown in various directions, resulting in a complex and compacted morphology.

3.2.3. Rosettes

A distinct constituent with a round shape was observed independently of the Fe concentration, as exemplified by the particles in the AlSi7Mg0.3-1.3Fe alloy, as shown in Figure 6a. These constituents are called rosettes due to their characteristic shape [32]. In many examples, these constituents comprise an intermetallic phase encircled by a “shadow” of a ternary eutectic constituent. The SEM image in Figure 6b reveals that this particle exhibits clearly defined zones with differing compositions: a nuclear zone and an outer area. Elemental maps, as depicted in this figure, facilitate understanding the distribution of alloy elements across these distinct zones. Al tends to decrease from the matrix towards the nuclear zone of the intermetallic particle, while Mg is solely present in the “shadow” zone of the rosette. On the contrary, Fe is detected exclusively in the nucleus intermetallic spherical particle. In addition, the Si concentration is higher in the particle’s nucleus than the shadow, with the lowest concentration observed in the outer zone of the nucleus. The particle’s inner and outer zones primarily differ in their Fe/Si ratio, increasing from 0.3 to 0.5. Notably, the phase reported with the stoichiometry closest to the inner phase can be tetragonal, δ-Al₃FeSi₂ [11,30] with Al₃FeSi₃ (Al/Si = 0.9).

The intermetallic phases were observed to have other morphologies in rarer occurrences, as seen in Figure 7. The shapes vary from single-phase spherical particles (Figure 7a,b) to a plate network resembling those observed in the matrix (Figure 7c). In the EDS spot analysis (refer to

Table 3. EDS elemental analysis (in at.%) of the various zones (Z1-Z7) represented in Figure 6.

Zones	Al	Si	Fe	Mg	O	Ni	Cr
Z1	59.19	28.93	8.32	3.55	-	-	-
Z2	59.48	33.36	1.6	5.56	-	-	-
Z3	65.59	15.96	13.1	-	-	3.13	2.22
Z4	53.3	33.03	0.7	7.14	3.61	2.22	53.3
Z5	48.78	29.07	14.09	1.84	-	4.88	1.34
Z6	57.59	31.87	4.81	3.55	-	2.18	57.59
Z7	53.3	33.03	0.7	7.14	3.61	2.22	53.3

), it was observed that the rosettes exhibit a higher concentration of Ni and Cr originating from the dissolved steel bar. Due to their similar atomic radius to Fe, these elements can effectively substitute Fe atomic places within the structure [9]. The Fe-rich intermetallic phases found in these alloys encompass a broad spectrum, like β-Al_4.5(Fe, Ni, Cr)Si (Z3) or δ-Al₃(Fe, Ni, Cr)Si₂ (Z5).

The shadowed regions of the rosettes display variations in chemical composition, characterised by zones like Z4, Z5, and Z7, which are notably rich in Ni and Cr compared to other microstructural constituents and exhibit two distinct morphologies. One morphology resembles dotted constituents (Z4), similar to ternary eutectics, while the other manifests as a lamellar phase (Z2). The size limitations imposed by the resolution of the EDS analysis technique made it infeasible to study the phases present within these shadowed areas.

3.3. Phase Identification

To validate the stoichiometry of the intermetallic phases observed via SEM, XRD, and EBSD spot analyses were performed. Figure 8 presents the XRD diffractograms collected for each alloy and the patterns used for phase identification. In the AlSi7Mg0.3-0.1Fe alloy, the prominent peaks observed are attributed to Al. The peak at 38.7° may correspond to Si and π-Al₉FeMg₃Si₅ and Al. Moreover, there is a shift in the main peak from Al (220) at 65.1° in AlSi7Mg0.3-0.1Fe to Al (111) at 38.7° in AlSi7Mg0.3-0.4Fe. A comparison of these two low-Fe concentration alloys indicates an increase in the intensity of peaks that can be associated with increasing π-Al₉FeMg₃Si₅ phase. However, peaks corresponding to β-Al_4.5FeSi were not detected, likely due to a low volume fraction of this phase combined with coincidental peak patterns at similar angles. Comparing the diffractograms for the different alloys, the intensity of some peak changes can be attributed to the coarse microstructure exhibited by these alloys and the variations in the area fractions of the different phases within the regions analysed by XRD.

The β-Al_4.5FeSi phase is solely detected in alloys with higher Fe concentration. Its characteristic peaks at specific angles increase in intensity with the rise in Fe concentration. Notably, the peak with the highest intensity shifts once more for alloys containing 2.1 and 3.8 wt.% Fe, aligning with Al (200) and β-Al_4.5FeSi (21-7) at 44.7°. This occurrence can be attributed to the presence of α′-Al₈Fe₂Si, which exhibits characteristic peaks at the same angles, enhancing the overall intensity. The overlapping of characteristic peaks from various Fe-rich intermetallic phases complicates the identification of the specific corresponding phases for each peak.

Table 4. Crystalline phase parameters detected in XRD spectra.

Phase	PDF	Lattice	Space Group	Lattice Parameters
				a (Å)	b (Å)	c (Å)	Angles
Al	00-004-0787	Cubic	Fm-3m (225)	4.04940	4.04940	4.04940	α = β = γ = 90°
Si	00-027-1402	Cubic	Fd-3m (227)	5.43088	5.43088	5.43088	α = β = γ = 90°
π-Al9FeMg3Si5	01-082-7018	Hexagonal	P-62m (189)	6.64000	6.64000	7.92000	α = β = 90°, γ = 120°
β-Al4.5FeSi	01-082-0546	Monoclinic	A2/a (15)	6.16100	6.17500	20.81300	α = γ = 90°, β = 90.42°
α′-Al8Fe2Si	00-020-0030	Hexagonal	P63/mmc	12.4000	12.4000	26.100	α = β = 90°, γ = 120°

provides the lattice parameters of the phases identified through XRD analysis. The Fe-rich intermetallic phases exhibit diverse crystal structures, with lattice types ranging from hexagonal for π-Al₉FeMg₃Si₅ and α′-Al₈Fe₂Si, which often manifest shapes resembling Chinese script to monoclinic for β-Al_4.5FeSi, typically presenting needle or plaque morphologies.

Spot EBSD analyses were performed on various microstructural constituents to establish correlations between the phases identified in XRD and those observed in SEM. The Kikuchi patterns in Figure 9c,d were confirmed to correspond to the α-Al matrix and dark plaque-like phases as Si-eutectic. The π-Al₉FeMg₃Si₅ phase (Figure 9e) was identified as the phases of the white particles resembling Chinese script, while the β-Al_4.5FeSi phases (Figure 9f) were associated with needle-like structures. Additionally, Figure 9 includes schematics of the unit cells of the π-Al₉FeMg₃Si₅ and β-Al_4.5FeSi intermetallic phases. However, it is noteworthy that indexing a pattern for the petals-shaped particles that were assumed to be α′-Al₈Fe₂Si phase proved challenging during analysis. This difficulty may arise from either the absence of the correct patterns in the database or the complex morphology of the phase, which could result in overlapping plates, making the pattern too intricate for successful indexing.

3.4. Quantitative Image Analysis

Section 3.2.2 and Section 3.3 addressed the morphologies and identification of intermetallic phases in high iron alloys. However, a quantitative analysis of the amount and size of each phase as the Fe concentration increases is still needed. The analysed phases include β-Thin Needles (Figure 3i), α′-Petals (Figure 3f), and β-Thick Needles (Figure 3c). The results presented in Figure 10 demonstrate the changes in size (measured by the ferret size of the particles, i.e., the longest distance between any two points of the particle boundary) of the three Fe-rich intermetallic phase morphologies.

Concerning the β-Al_4.5FeSi thin needles, it is evident from the histograms that the average ferret size tends to reduce with rising Fe levels, from 33 ± 3 to 20 ± 1 µm. This may be because of the higher availability of Fe for precipitation, which enhances nucleation rates and results in larger quantities of smaller needles. On the other hand, the α′-Al₈Fe₂Si petal particles only start to appear in the AlSi7Mg0.3-2.1Fe alloy, characterised by a sparse distribution of irregularly shaped particles, as shown in the SEM images (Figure 3). When comparing the alloys with the most elevated Fe levels, AlSi7Mg0.3-2.1Fe contains a few small particles with an average ferret size of 63 ± 6 µm. While in the AlSi7Mg0.3-3.8Fe alloy, the α′-Al₈Fe2Si petals have an average size more than three times that of the previous alloys, close to 172 ± 8 µm. Moreover, the β-Al_4.5FeSi thick needles were exclusively quantified in the alloy with the highest Fe concentration since it was the only alloy that presented this intermetallic morphology. The average length of these particles surpasses any other particle type, reaching around 900 µm, with several particles exceeding 2 mm. Thus, compared with the β thin needles in the same alloy, the thick needles are more than 40 times the length on average.

Various fields were measured throughout the analysis, although the total areas analysed varied among particle types due to their frequency of occurrence differences. However, these procedures remained consistent across the three alloys. For the β-Al_4.5FeSi thin needles, approximately 5 mm² of the area was analysed due to their abundance. In contrast, a larger area of around 120 mm² was required to obtain statistically significant results for the other particle types, given their comparatively lower occurrence.

3.5. Porosity

Figure 11 shows the change in porosity as a function of Fe concentration. Low Fe alloy presents relatively low average porosity values, around 0.05% (95% CI: 0.03 to 0.07%). Above 1 wt.% Fe, the porosity increases considerably and linearly, achieving the maximum value of 1.44% (95% CI: 1.21 to 1.68%).

Examining the results of the RPT depicted in Figure 12, a notable distinction arises between the air-solidified samples and those subjected to vacuum. Unlike the former, the vacuum-treated samples do not exhibit a linear correlation between Fe concentration and porosity behaviour.

Upon comparing the ratios obtained from the RPT samples, particularly those of DI and BI, a decrease is observed in the alloy containing 1.3 wt.% Fe. This trend then progressively increases with higher Fe concentrations. Analysing the data presented in Figure 12b regarding the evolution of average pore size and pore count in the vacuum-solidified samples reveals that the AlSi7Mg0.3-1.3Fe alloy displays the smallest average pore size yet the highest pore count among all alloys. Thus, numerous new pores are formed at this Fe concentration level. The emerging porosities primarily occur due to feeding blockages caused by intermetallic phases. Larger intermetallic particles will result in more significant blockages, leading to the formation of larger pores that allow hydrogen to flow with less restriction, reducing the expansion of bifilms necessary to form spherical pores, typically observed in the RPT vacuum-solidified samples. This observation is further corroborated by Figure 12c, where the pores’ average circularity and solidity shape descriptors decrease at higher Fe contents. Thus, pores are generally less circular and present irregular concave shapes with more indentations and lobes typical of shrinkage porosities.

A notable decrease in DI is observed for the AlSi7Mg0.3-3.8Fe alloy. Despite increased porosity for air and vacuum-solidified samples, DI decreases because their density difference also reduces. Thus, for the highly contaminated alloys, the primary mechanism of porosity formation is attributed to shrinkage due to feeding deficiencies rather than the formation of gas porosity.

3.6. Hardness

In Figure 13a, the evolution of Brinell hardness of the alloy as a function of Fe concentration is presented, revealing a positive linear correlation (R = 0.94) with Fe concentration, as expected. Notably, as Fe concentration increases, the boxes in the plot elongate. Since these boxes represent the interquartile range, a more extended box signifies greater dispersion in the dataset, indicating that the material becomes more heterogeneous with the presence of Fe. This observation aligns with previous findings, where higher-contaminated alloys demonstrated considerable variation in intermetallic phase distribution and greater porosity.

Microhardness Vickers tests were conducted on these particles, with results in Figure 13b, facilitated by the large sizes of some intermetallic phases in the AlSi7Mg0.3-3.8Fe alloy. Comparing the hardness of β thick needles (767 ± 80 HV) and α′ petals (823 ± 134 HV) with the Al-Si matrix (63 ± 2 HV), a significant difference in hardness values between the microstructure constituents is evident. This substantial difference in hardness plays a crucial role in the increased likelihood of microcracks in Fe-rich intermetallic phases that then easily propagate to the matrix.

3.7. Corrosion Behaviour

Figure 14a illustrates the mass loss of each sample as a function of immersion time. As anticipated, mass loss tends to increase over time due to corrosion. Generally, a higher Fe concentration leads to more significant mass loss. Notably, the 2.1% Fe alloy exhibits lower corrosion over extended periods than the 1.3% Fe alloy. While low Fe alloys consistently lose mass, highly contaminated alloys gradually reduce mass loss over time.

When analysing corrosion rates (Figure 14b), the low Fe alloys demonstrate relatively consistent rates, averaging 1.15 mm/year for the 0.1% Fe alloy and 3.12 mm/year for the 0.4% Fe alloy. This suggests that even a modest increase of 0.3% Fe among the low Fe alloys results in approximately a 73% increase in the corrosion rate. In contrast, the corrosion rate of the highly contaminated alloys tends to slow down after immersion times more significant than 4 h. Additionally, the difference in corrosion rates among the three alloys decreases as immersion time increases.

Figure 15 and Figure 16 illustrate the cross-section and top view of the corroded surfaces for both low and high Fe alloys, respectively. In the low Fe alloys, corrosion is predominantly in the eutectic areas, while the α-Al dendritic arms remain largely unaffected. On the contrary, in highly contaminated alloys, corrosion progresses through the eutectic zones and near the Fe-rich intermetallic phases.

However, at concentrations above 2.1% Fe, the corrosion progression appears to be influenced by thicker β-Al_4.5FeSi needle-like structures. These particles slow down the corrosion and influence the formation of corrosion products (as shown in Figure 16c), resulting in a slower mass loss rate. The corrosion products tend to be composed of Al oxides and are trapped between intricate morphologies. From the top view of the corroded surface of the alloy with 3.8% Fe (Figure 16f), it is evident that, unlike other alloys, the thick β-Al_4.5FeSi needles tend to be the last to undergo corrosion, appearing to protrude from the surface and are surrounded by corroded α-Al.

4. Discussion

The current study provides insights into the complex evolution of intermetallic phases as Fe concentration on AlSi7Mg0.3 alloy exceeds critical values. Notably, it has been demonstrated that the increase in Fe concentration leads to the formation of new phases and an increase in both size and number of intermetallic particles.

Given the tendency of recycling systems to raise contamination levels in SAA and the need to reduce reliance on primary Al for dilution, it is crucial to initiate upcycling studies focusing on Al casting alloys containing excessive amounts of Fe, up to 4%. This necessity is further underscored by the significant changes in intermetallic particle size and shape, porosity levels, and hardness in these alloys.

Upon analysing the diverse microstructures of the alloys studied in the current study, it becomes evident that the AlSi7Mg0.3-3.8Fe alloy exhibits the most complex microstructure, exhibiting significant variations in intermetallic phases characterised by substantial differences in both size and shape.

Figure 17 presents a schematic representation of the alloy, which began cooling from 720 °C, delineating the following significant steps:

• At 668 °C, the α′-Al₈Fe₂Si phase begins to precipitate and grow.
• By reaching to 617 °C, the β-Al_4.5FeSi begins to precipitate instead of α′-Al₈Fe₂Si. This phase precedes the ternary eutectic reaction, usually termed β-Al_4.5FeSi pre-eutectic [33].
• At 611 °C, the liquid transforms into the matrix phase, α-Al, with the β-Al_4.5FeSi pre-eutectic transforming into thick needles.
• Finally, at 577 °C, the final reaction occurs, with the last α-Al solidifying while the Si-eutectic and very thin β-Al_4.5FeSi needles precipitate, denoted as the β-Al_4.5FeSi eutectic. This previous reaction corresponds to a ternary eutectic reaction at 577 °C, L → α-Al + Si + β-Al_4.5FeSi, occurring at similar temperatures as observed in other investigations [34].

Owing to the substantial temperature range of approximately 50 °C between the onset and end of α′-Al₈Fe₂Si precipitation and the slow cooling process, this phase can grow into complex and sizable particles, with average lengths nearing 170 µm. Similarly, the β-Al_4.5FeSi pre-eutectic thick, elongated needles (average length around 890 µm) can form due to the considerable temperature interval of 40 °C. These needles tend to grow at the same time as the α-Al dendrites, as evidenced in deep-etched samples (Figure 5b), where the dendritic structure of Al indented the needles. In contrast, the β-Al_4.5FeSi eutectic needles are extremely thin and small (around 20 µm in average length) because this phase precipitates during the eutectic ternary reaction, which occurs significantly more rapidly. Observations from deep-etched samples reveal that these phases are more complex, with the α′-Al₈Fe₂Si and β-Al_4.5FeSi phases comprising a network of several plates.

The influence of Fe on porosity is significant, especially in highly contaminated alloys, where it was observed that porosity primarily resulted from shrinkage. As noted in bibliographies such as Taylor et al. [22] and supported by observations, this porosity is attributed to long β-Al_4.5FeSi platelets that impede the free flow of liquids during feeding to interdendritic spaces. This effect is more prominent in alloys with a solidification sequence where the formation of β-Al_4.5FeSi phase platelets occurs before the eutectic reaction. Consequently, a higher volume fraction of pre-eutectic β-Al_4.5FeSi worsens interdendritic feeding and aggravates porosity. Consequently, as Fe concentration increases, the defect volume fraction also rises, potentially reducing elongation values and diminishing these alloys’ mechanical properties and applicability.

As anticipated, the hardness significantly increases with the Fe concentration, reaching a maximum value of 83 ± 3 HBW 2.5/62.5 (a 22% increase from the commercial alloy). This observation aligns with the findings of Taghiabadi et al. [10] for an A380 alloy containing 2.5 wt.% Fe. The formation of a higher number of intermetallic phases, which exhibit considerably higher hardness values than the matrix, contributes to the overall increase in alloy hardness. The increased value dispersion with rising Fe concentration may also be attributed to the heterogeneity observed in the microstructures, characterised by considerably large intermetallic particles and high porosity. The excessive hardening of the alloy may significantly constrain its applicability in various contexts. Therefore, it is imperative to explore upcycling methods for these highly contaminated alloys to enhance their utility and sustainability.

An incremental mass loss was observed with an increase in the Fe concentration of the alloy and immersion time. Even a small increase in Fe resulted in a significant increase in the CR. This corrosion occurs because the Al matrix forms micro-galvanic pairs with the intermetallic and eutectic compounds, leading to pitting. As the amount of Fe-rich intermetallic compounds increases, the intensity of corrosion becomes more pronounced. Another factor influencing the corrosion resistance of the alloys is the SDAS since coarse dendritic spacing promotes the formation of large cathodic Si and intermetallic particles adjacent to the anodic Al matrix. Alloys with higher levels of Fe contamination exhibited a more irregular distribution of α-Al dendritic arms, which can further promote corrosion [35,36]. Although higher contamination levels significantly decrease corrosion resistance, the last particles to fully corrode are the β-thick needles. These needles also tend to halt the propagation of corrosion.

Through SEM imaging and EDS analysis, a significant number of spherical constituents containing intermetallic particles, consistently belonging to the Al-Fe-Si system, were observed, as also noted by Ferdian et al. [37]. It was observed that both β-Al_4.5FeSi and δ-Al₃FeSi₂ intermetallic phases can assume highly spherical shapes. Alexander et al. [32] proposed that these rosettes result from solidified droplets formed due to the unstable growth of columnar Al cells. Under favourable conditions, Al dendritic arms may split, promoting solute enrichment in the liquid between these new arms. If these two arms of the cell coalesce, highly solute-concentrated liquid droplets are trapped within the solidified dendritic arms. Then, these droplets solidify mainly into intermetallic phases due to high solute concentrations. As solidification progresses near this instability limit, the dendritic arms may cyclically split and coalesce, leaving a string of droplets and then rosettes as the growth front progresses. Understanding the mechanisms that lead to the formation of these rosettes, which are not fully known, could help manipulate their shape and transform platelets into spherical particles. This manipulation may serve as a potential upcycling method, which could reduce stress concentrations and alleviate the fragility of the alloy.

These findings can guide future research on the SAAs upcycling, emphasising that such studies cannot solely focus on the Fe concentration levels typical of other commercial alloys (up to 1.3 wt.% for A380), given the evolution of intermetallic phases and the possibility of increasing level of contamination within the current cascading recycling system.

5. Conclusions

5.1. Microstructure Variation with Fe Concentration

The current study evaluated the impact of varying Fe concentrations (0.1, 0.4, 1.3, 2.1, and 3.8 wt.%) on the microstructure of AlSi7Mg0.3 alloys. A clear understanding of the relationship between alloy composition and microstructural and intermetallic phases evolution was possible through a detailed qualitative and quantitative characterisation.

Fe concentrations above 1% lead to the extensive precipitation of large intermetallic phases. Additionally, the structure of the α-Al phase becomes increasingly irregular as the Fe percentage rises. The typical dendritic structure found in Al casting is completely lost in alloys with high contamination levels.

The AlSi7Mg0.3-3.8Fe alloy, which exhibited the highest contamination level, showed the most complex microstructure, characterised by three main intermetallic phases with distinct shapes and sizes:

• Petal-shaped α′-Al₈Fe₂Si with an average length of 170 µm.
• Thick pre-eutectic β-Al_4.5FeSi plaques with an average length of 890 µm.
• Thin eutectic β-Al_4.5FeSi needles with an average length of 20 µm.

The distinct intermetallic phase morphologies, particularly rosettes with spherical Fe-rich intermetallic particles, present opportunities for tailoring material properties to meet specific performance requirements, particularly enhancing ductility.

5.2. Porosity, Hardness, and Corrosion Variation with Fe Concentration

The porosity slightly increased with the addition of Fe, likely due to the formation of shrinkage defects caused by feeding blockages from intermetallic phase networks. The hardness of the material increased with the Fe content, reaching a maximum of 83 ± 3 HBW 2.5/62.5 at 3.8 wt.%. This increase in hardness can be attributed to extremely hard Fe-containing phases compared to the hardness of the Al-Si microstructure. However, corrosion resistance decreased at higher Fe levels due to the formation of galvanic couples between Fe-rich intermetallic phases and the Al matrix.

Properties such as wear and tribocorrosion resistance due to high concentrations of hard particles and performance at elevated temperatures affected by the intrinsic characteristics of intermetallic phases should be addressed in future studies to evaluate new potential applications. However, the initial focus must be optimising the microstructure of highly contaminated alloys by managing phase morphologies and minimising porosity. Future research should also investigate alternative alloy compositions and processing conditions to enhance mechanical properties, particularly emphasising mechanical characterisation to assess the ductility of these alloys. This is crucial for ensuring the alloys achieve their optimal performance.