Rare earth magnesium alloys, particularly the WE43A grade, have gained significant attention in aerospace and automotive industries due to their excellent specific strength, specific stiffness, and high-temperature performance. These alloys are widely used in helicopter frames, aircraft wings, compressor casings, and other structural components. The application of such advanced materials is demonstrated in the following component, which is a typical magnesium alloy precision casting used in aircraft engines.
Although rare earth magnesium alloys possess outstanding mechanical properties, their manufacturing process often introduces various casting defects, such as porosity, inclusions, and segregation. These defects inevitably degrade the mechanical performance and corrosion resistance of the material. In fact, statistics show that about 90% of magnesium alloy components are produced by casting, and even a small amount of defects can lead to a significant reduction in mechanical properties. Therefore, a comprehensive understanding of how casting defects affect the material behavior is essential for improving the reliability and service life of magnesium alloy components. In this study, I systematically investigated the effects of three typical casting defects—porosity, inclusions, and segregation—on the microstructure, tensile properties, fatigue behavior, and corrosion resistance of WE43A rare earth magnesium alloy.
1. Experimental Materials and Methods
The material used in this investigation was a commercial WE43A magnesium alloy. Its nominal chemical composition is listed in Table 1. To obtain specimens with different levels of porosity, inclusions, and segregation, the casting process parameters were deliberately adjusted. Additionally, specimens with naturally occurring defects were extracted from production castings that had been inspected by X-ray radiography. The defect levels were rated according to the specifications for magnesium alloy castings. Five levels of porosity (denoted as S1–S5), five levels of inclusions (J1–J5), and eight levels of segregation (PX1–PX8) were prepared. A defect-free reference material (W) was also included.
| Y | Nd | Zr | Gd | Mg |
|---|---|---|---|---|
| 3.7–4.3 | 2.0–2.5 | 0.40–1.0 | 0.4–0.6 | Balance |
Tensile specimens were cylindrical bars with a diameter of 5.00 mm and a gauge length of 25 mm. The tensile tests were performed on an MTS universal testing machine at a strain rate of 0.00025 s⁻¹, following the HB5143 standard. Four-point bending fatigue specimens were machined into plates of 7.4 mm thickness, 21.0 mm width, and 82.2 mm length. Fatigue tests were conducted using a servo-hydraulic testing machine at a frequency of 20 Hz and a mean stress of 100 MPa. The stress amplitude was increased stepwise by 10 MPa every 500,000 cycles until failure, allowing the determination of fatigue limit and fatigue life at different stress levels. The loading configuration is shown schematically in the test fixture. After mechanical testing, the fracture surfaces were examined using scanning electron microscopy (SEM) and energy-dispersive X-ray spectroscopy (EDS). The area fraction of casting defects on the fracture surfaces was quantified using ImageJ software. For metallographic observation, specimens were ground, polished, and etched with a 10% citric acid solution. X-ray diffraction (XRD) was used to identify the phases. Corrosion resistance was evaluated by electrochemical polarization tests in a 3.5 wt.% NaCl solution using a three-electrode cell, as well as by immersion tests for 12 h followed by SEM observation.
2. Porosity Defects
2.1 Characterization of Porosity
Porosity is one of the most common casting defects in magnesium alloys. In the present study, the porosity appeared as irregular black voids, often elongated or ellipsoidal, predominantly located at grain boundaries. Such a morphology suggests that porosity formed due to insufficient liquid feeding during solidification, resulting in voids between dendrites. The fracture surfaces of tensile and fatigue specimens exhibited dark, rough regions corresponding to these pores. With increasing porosity level, the area fraction of porosity on the fracture surface increased dramatically, from approximately 0.8% at level S1 to about 30% at level S5. The measured defect areas are summarized in Table 2.
| Porosity level | Defect area (μm²) | Area fraction (%) |
|---|---|---|
| S1 | 1,220,112 | 0.83 |
| S2 | 3,131,143 | 2.13 |
| S3 | 8,364,367 | 5.69 |
| S4 | 31,208,129 | 21.23 |
| S5 | 44,879,112 | 30.53 |
2.2 Effect on Tensile Properties
Table 3 presents the tensile properties of WE43A alloy with different porosity levels. The reference alloy (W) exhibited a tensile strength of 280 MPa and a yield strength of 180 MPa. At the lowest porosity level (S1), the tensile strength decreased only slightly to 261 MPa, indicating that a small amount of porosity has a negligible influence. However, when the porosity level exceeded S1, the strength and ductility dropped sharply. At the highest porosity level (S5), the tensile strength was only 82 MPa, a reduction of about 70%. The yield strength decreased from 180 MPa to 68 MPa, a 62% reduction. The elongation and reduction of area showed maximum decreases of approximately 41% and 54%, respectively. Interestingly, the elastic modulus remained nearly constant around 45 GPa for all porosity levels, suggesting that porosity primarily affects the load-bearing capacity rather than the atomic bonding stiffness.
| Porosity level | Tensile strength (MPa) | Yield strength (MPa) | Elongation (%) | Reduction of area (%) | Elastic modulus (GPa) |
|---|---|---|---|---|---|
| W | 280 | 180 | 4.50 | 4.50 | 45.00 |
| S1 | 261 | 182 | 4.26 | 4.21 | 44.52 |
| S2 | 201 | 148 | 2.48 | 2.52 | 44.78 |
| S3 | 194 | 138 | 2.51 | 2.31 | 45.44 |
| S4 | 161 | 135 | 2.44 | 2.03 | 44.31 |
| S5 | 82 | 68 | 2.63 | 2.79 | 45.69 |
The degradation of tensile properties can be attributed to the reduction of effective load-bearing area and the stress concentration induced by pores. The fracture surface observations revealed a predominantly brittle fracture mode, characterized by cleavage facets and quasi-cleavage features, which is typical for magnesium alloys with hexagonal close-packed structure and limited slip systems. The pores acted as crack initiation sites, leading to premature failure.
2.3 Effect on Fatigue Properties
Fatigue tests were conducted at a mean stress of 100 MPa using the step-loading method. Table 4 summarizes the fatigue life and fatigue limit data. The defect-free alloy exhibited an average fatigue life of about 2.5 million cycles and a fatigue limit of 42 MPa. Even at the lowest porosity level (S1), the fatigue life decreased by about 20% (average life 363,915 cycles), and the fatigue limit dropped to 35 MPa. This indicates that porosity is extremely detrimental to fatigue performance because pores act as initial cracks, providing sites for crack nucleation and propagation. As the porosity level increased, the fatigue life decreased further, but the rate of reduction slowed. At level S5, the average fatigue life was only 39,445 cycles, and the fatigue limit was 26 MPa. The maximum reductions in fatigue life (logarithmic scale) and fatigue limit were approximately 28% and 38%, respectively.
| Porosity level | Average fatigue life (cycles) | Log fatigue life | Life decline (%) | Fatigue limit (MPa) | Limit decline (%) |
|---|---|---|---|---|---|
| W | 2,506,109 | 6.399 | — | 42.0 | — |
| S1 | 363,915 | 5.561 | 13.10 | 35.0 | 16.67 |
| S2 | 138,627 | 4.994 | 21.96 | 30.3 | 27.86 |
| S3 | 68,076 | 4.833 | 24.47 | 29.2 | 30.48 |
| S4 | 44,463 | 4.648 | 27.36 | 27.0 | 35.71 |
| S5 | 39,445 | 4.596 | 28.18 | 26.0 | 38.10 |
2.4 Fatigue Life Prediction Model
Based on the linear elastic fracture mechanics approach, the fatigue life can be related to the initial crack size, which in this case corresponds to the size of the porosity defect. The Paris law describes the crack growth rate as:
$$\frac{da}{dN} = c (\Delta K)^n $$
where c and n are material constants, and ΔK is the stress intensity factor range given by:
$$\Delta K = Y \Delta \sigma \sqrt{\pi a} $$
Assuming that the porosity defect can be idealized as an initial crack of equivalent diameter a₀, and using the measured defect area A, the equivalent diameter is:
$$a_0 = \sqrt{\frac{4A}{\pi}} $$
The critical crack size at fracture can be obtained from:
$$a_c = \frac{1}{\pi}\left(\frac{\Delta K_{1c}}{Y \Delta \sigma_c}\right)^2 $$
Integrating the Paris law from the initial crack size to the critical crack size gives the total fatigue life:
$$N_c = \int_{a_0}^{a_c} \frac{da}{c Y^n (\Delta \sigma)^n (\pi a)^{n/2}} $$
For a given stress range, the integration yields:
$$N_c = \frac{1}{(n-2)c Y^n (\Delta \sigma)^n \pi^{n/2}} \left( \frac{1}{a_0^{(n-2)/2}} – \frac{1}{a_c^{(n-2)/2}} \right) $$
Combining all constants into a single parameter B and treating the second term as a constant D (for fixed specimen geometry and loading conditions), we obtain:
$$N_c = B \left( \frac{1}{a_0^{(n-2)/2}} – D \right) $$
Using the experimental fatigue data and the measured equivalent diameters (listed in Table 5), the model was fitted. The material constant n was taken as 3 for magnesium alloys. The fitting result is shown in Figure 3 (not reproduced here) with an R² value of 0.8468, indicating a reasonable agreement. The resulting relationship between fatigue life and equivalent diameter is:
$$N_c = 516,588.3 \left( \frac{1}{x^{0.5}} – 0.8073 \right) $$
Furthermore, using the empirical S-N curve relationship of the form:
$$S = S_L \left[ H + A_2 \left( N + C_2 \right)^{-B_2} \right] $$
where H = 1, A₂ = 25.7, B₂ = 0.32, and C₂ = 47.3, the fatigue limit can be expressed as a function of the equivalent diameter. Combining the fatigue life model with this equation yields:
$$S_L = \frac{150}{1 + 25.7 \times \left( A_3 \, a_0^{-0.5} – C_3 \right)^{-0.32}} $$
Fitting this relation to the experimental fatigue limits gave an R² of 0.9134, demonstrating good predictive capability. This model provides a practical tool for estimating the fatigue strength of WE43A alloy components containing porosity defects of known size.
| Porosity level | Equivalent diameter (mm) |
|---|---|
| S1 | 0.512 |
| S2 | 0.685 |
| S3 | 0.987 |
| S4 | 1.253 |
| S5 | 1.571 |
3. Inclusion Defects
3.1 Characterization of Inclusions
Inclusions in the WE43A alloy were found to be mainly magnesium oxide (MgO) particles, which formed during melting and casting due to reaction with oxygen. On the fracture surfaces, inclusions appeared as black, irregularly shaped particles in the macroscopic view, while in the microstructure they were typically spherical or ellipsoidal, with sizes ranging from a few micrometers to several hundred micrometers. A typical large inclusion is shown in the SEM image with EDS analysis indicating a high oxygen peak, confirming the MgO nature. The area fraction of inclusions on the fracture surface increased from about 0.08% at level J1 to about 6% at level J5, as listed in Table 6.
| Inclusion level | Defect area (μm²) | Area fraction (%) |
|---|---|---|
| J1 | 118,079 | 0.08 |
| J2 | 350,346 | 0.24 |
| J3 | 1,411,830 | 0.96 |
| J4 | 8,669,715 | 5.60 |
| J5 | 10,255,631 | 6.97 |
3.2 Effect on Tensile Properties
The tensile test results for different inclusion levels are summarized in Table 7. Compared to the porosity defects, the influence of inclusions on the strength of WE43A alloy is relatively mild. The tensile strength decreased from 236 MPa (W) to a minimum of 214 MPa at J2, representing a maximum reduction of about 9%. The yield strength showed a similar trend, with a maximum decrease of about 8%. The elastic modulus varied slightly within ±10%. However, the ductility was more significantly affected. The elongation decreased from 3.59% to 1.41% (a 60% reduction), and the reduction of area decreased from 3.79% to 1.83% (a 51% reduction). The scatter in the strength data was relatively large, indicating that the effect of inclusions is highly dependent on their location and orientation relative to the loading direction.
| Inclusion level | Tensile strength (MPa) | Yield strength (MPa) | Elongation (%) | Reduction of area (%) | Elastic modulus (GPa) |
|---|---|---|---|---|---|
| W | 236 | 182 | 3.59 | 3.79 | 47 |
| J1 | 221 | 176 | 2.62 | 1.99 | 44 |
| J2 | 214 | 167 | 2.75 | 2.73 | 42 |
| J3 | 222 | 172 | 1.46 | 2.23 | 44 |
| J4 | 215 | 169 | 1.41 | 2.32 | 43 |
| J5 | 224 | 172 | 1.59 | 1.83 | 44 |
The fracture surface analysis revealed that the inclusions acted as crack initiation sites, and the fracture mode was predominantly transgranular cleavage. The interface between the inclusion and the matrix was often the source of microcracks, leading to the detachment of inclusions and the formation of dimple-like cavities. The presence of inclusions also promoted the formation of secondary cracks, as shown in the fractographs.
3.3 Effect on Fatigue Properties
The fatigue behavior of specimens with inclusions is summarized in Table 8. The fatigue life decreased significantly with increasing inclusion content. The average fatigue life of the defect-free material was about 4.97 million cycles, while that of the J5 specimens was only 54,441 cycles, a reduction of about 26% on the logarithmic scale. The fatigue limit also decreased from 42 MPa to 28 MPa, a 33% reduction. Unlike tensile strength, the fatigue life showed a monotonic decrease with inclusion level, because inclusions provide localized stress concentrations that are particularly harmful under cyclic loading. The fatigue fracture surfaces exhibited multiple crack origins, all associated with inclusions, and the propagation region contained characteristic fatigue striations with varying spacing, indicating unstable crack growth.
| Inclusion level | Average fatigue life (cycles) | Log fatigue life | Life decline (%) | Fatigue limit (MPa) | Limit decline (%) |
|---|---|---|---|---|---|
| W | 4,972,069 | 6.399 | — | 42.0 | — |
| J1 | 290,569 | 5.463 | 14.6 | 34.0 | 19.1 |
| J2 | 138,481 | 5.141 | 19.7 | 30.3 | 27.9 |
| J3 | 147,547 | 5.168 | 19.2 | 31.5 | 25.0 |
| J4 | 77,665 | 4.890 | 23.6 | 28.8 | 31.5 |
| J5 | 54,441 | 4.735 | 26.0 | 28.0 | 33.3 |
4. Segregation Defects
4.1 Characterization of Segregation
Segregation in WE43A alloy manifests as the non-uniform distribution of rare earth elements, primarily in the form of intermetallic compounds along grain boundaries. The optical and SEM observations revealed that the segregated regions consist of continuous network-like phases and small particle-like phases. The XRD analysis identified the phases as α-Mg, Mg₄₁Nd₅, and Mg₂₄Y₅. The network-like phase is Mg₄₁Nd₅, and the granular phase is Mg₂₄Y₅. These phases are essentially eutectic products formed during solidification, which are rich in rare earth elements. With increasing segregation level, the area fraction on the fracture surface increased from 0.69% at PX1 to 26.14% at PX8, as listed in Table 9.
| Segregation level | Defect area (μm²) | Area fraction (%) |
|---|---|---|
| PX1 | 1,014,301 | 0.69 |
| PX2 | 5,277,337 | 3.59 |
| PX3 | 8,320,210 | 5.66 |
| PX4 | 11,054,447 | 7.52 |
| PX5 | 15,920,112 | 10.83 |
| PX6 | 26,548,297 | 18.06 |
| PX7 | 35,838,647 | 24.38 |
| PX8 | 38,425,868 | 26.14 |
4.2 Effect on Tensile Properties
The tensile properties of specimens with different segregation levels are presented in Table 10. Up to level PX4, the tensile strength and ductility decreased only moderately. For example, the tensile strength dropped from 268 MPa to about 245 MPa (an 8% reduction), and the elongation decreased from 3.91% to 1.98% (about 50% reduction). However, when the segregation level exceeded PX4, the properties deteriorated more significantly. At PX8, the tensile strength was only 139 MPa (a 48% reduction), and the elongation was 0.75% (an 81% reduction). The reduction of area decreased by 89% at PX8. The yield strength remained relatively constant around 180 MPa for levels up to PX7, but at PX8 the specimen fractured before yielding, indicating extreme brittleness. The elastic modulus was essentially unaffected by segregation, remaining around 45–47 GPa.
| Segregation level | Tensile strength (MPa) | Yield strength (MPa) | Elongation (%) | Reduction of area (%) | Elastic modulus (GPa) |
|---|---|---|---|---|---|
| W | 268 | 184 | 3.91 | 4.08 | 45 |
| PX1 | 246 | 181 | 2.92 | 3.53 | 45 |
| PX2 | 248 | 178 | 2.80 | 3.91 | 45 |
| PX3 | 237 | 179 | 1.80 | 2.97 | 46 |
| PX4 | 245 | 176 | 1.98 | 3.65 | 46 |
| PX5 | 206 | 189 | 1.36 | 1.67 | 47 |
| PX6 | 190 | 181 | 1.27 | 1.33 | 46 |
| PX7 | 189 | 179 | 1.14 | 1.32 | 46 |
| PX8 | 139 | 139 | 0.75 | 0.45 | 43 |
The fracture surfaces of segregated specimens showed a mixture of transgranular and intergranular fracture. The brittle eutectic phases at grain boundaries provided preferential paths for crack propagation. Secondary cracks were observed around the segregated regions, indicating poor cohesion between the matrix and the brittle phases. The fracture mode was predominantly brittle cleavage, with some intergranular facets visible where segregation was severe.
4.3 Effect on Fatigue Properties
Table 11 summarizes the fatigue results for different segregation levels. The fatigue life showed a continuous decrease with increasing segregation level. The defect-free alloy had an average fatigue life of about 4.6 million cycles (log 6.66). At PX1, the life dropped to 246,017 cycles, already more than an order of magnitude lower. At PX8, the average life was only 29,337 cycles, and the logarithmic decrease was 33%. The fatigue limit declined from 42 MPa to 25.6 MPa, a reduction of 39%. The fatigue fracture surfaces revealed that the crack initiation sites were often located at segregated areas, where the brittle eutectic phases acted as crack nuclei. The crack propagation region showed cleavage facets and secondary cracks along grain boundaries, indicating a mixed fracture mechanism.
| Segregation level | Average fatigue life (cycles) | Log fatigue life | Life decline (%) | Fatigue limit (MPa) | Limit decline (%) |
|---|---|---|---|---|---|
| W | 4,606,827 | 6.66 | — | 42.0 | — |
| PX1 | 246,017 | 5.39 | 19 | 33.7 | 19.8 |
| PX2 | 162,165 | 5.21 | 22 | 38.7 | 7.8 |
| PX3 | 117,128 | 5.07 | 24 | 30.9 | 26.4 |
| PX4 | 105,808 | 5.02 | 25 | 30.6 | 27.2 |
| PX5 | 97,957 | 4.99 | 25 | 30.3 | 27.9 |
| PX6 | 64,254 | 4.81 | 28 | 28.7 | 31.7 |
| PX7 | 53,028 | 4.72 | 29 | 27.9 | 33.5 |
| PX8 | 29,337 | 4.47 | 33 | 25.6 | 39.1 |
4.4 Effect on Corrosion Resistance
The corrosion resistance of WE43A alloy with different segregation levels was evaluated by immersion tests and electrochemical polarization in 3.5 wt.% NaCl solution. The immersion tests revealed that the severity of corrosion increased with segregation level. Low segregation specimens (PX1 and PX2) exhibited minor localized corrosion, while high segregation specimens (PX7 and PX8) showed extensive corrosion with cracks and thick corrosion product layers. The corrosion products were identified by EDS as mainly MgO and MgCl₂, with Mg(OH)₂ also likely present. The electrochemical polarization curves are shown in Figure 10 of the original thesis. The corrosion potential (Ecorr) shifted from −1.573 V for PX1 to −1.432 V for PX8, indicating that the overall corrosion tendency slightly decreased with increasing segregation. This can be explained by the higher noble potentials of the Mg₂₄Y₅ and Mg₄₁Nd₅ phases compared to the α-Mg matrix. However, the corrosion current density (icorr) increased significantly from 2.090 × 10⁻⁵ A/cm² for PX1 to 2.467 × 10⁻³ A/cm² for PX8, an increase of about two orders of magnitude. This indicates that the corrosion rate actually increased markedly with segregation, despite the slightly nobler Ecorr. The increased current density is attributed to the formation of micro-galvanic cells between the segregated phases and the matrix, which accelerate localized corrosion. The corrosion morphologies showed that the segregated regions acted as preferential corrosion sites, and the corrosion products formed non-protective layers, leading to continuous attack.
5. Conclusions
In this study, I have systematically characterized the effects of three common casting defects—porosity, inclusions, and segregation—on the microstructure, mechanical properties, fatigue behavior, and corrosion resistance of WE43A rare earth magnesium alloy. The main findings are as follows:
- Porosity appears as irregular black voids at grain boundaries, and its area fraction on the fracture surface increases from 0.8% to 30% with increasing porosity level. Porosity significantly reduces tensile strength (up to 70%), yield strength (up to 62%), and ductility (up to 54% reduction of area). Fatigue life and fatigue limit decrease by up to 28% and 38%, respectively. A fatigue life prediction model based on the Paris law was developed, relating the fatigue life and fatigue limit to the equivalent diameter of porosity defects. The model shows good agreement with experimental data (R² = 0.847 for life, R² = 0.913 for fatigue limit).
- Inclusions are mainly MgO particles, appearing as black irregular shapes. The inclusion area fraction increases from 0.08% to 6%. Inclusions have a relatively minor effect on tensile strength (maximum 9% reduction) but a significant effect on ductility (up to 60% reduction in elongation). Fatigue life and fatigue limit decrease by up to 26% and 33%, respectively. Inclusions act as crack initiation sites, particularly under cyclic loading.
- Segregation consists of network-like Mg₄₁Nd₅ and granular Mg₂₄Y₅ phases distributed along grain boundaries. The segregation area fraction increases from 0.69% to 26%. When the segregation level exceeds PX4, tensile strength and ductility deteriorate significantly, with maximum reductions of 48% for tensile strength, 81% for elongation, and 89% for reduction of area. Fatigue life and fatigue limit decrease by up to 33% and 39%, respectively. Corrosion potential shifts slightly positive with increasing segregation, but corrosion current density increases by two orders of magnitude, indicating a much higher corrosion rate. The corrosion products are mainly MgO and MgCl₂.
These results demonstrate that casting defects have a profound influence on the performance of WE43A magnesium alloy. The quantitative relationships established in this work provide valuable insights for predictive modeling of component integrity and for optimizing casting processes to minimize defects and improve the reliability of magnesium alloy parts in aerospace applications.