MicrostructureThe OM and SEM images of Mg–9Gd–3Y–2Zn–0.5Zr alloy are presented in Fig. 1. The OM results reveal that the sample consists of equiaxed and columnar crystals within an α-Mg matrix. The backscattered electron morphology distinctly shows a second-phase microstructure in the alloy, with brighter regions indicating enrichment of heavier elements arranged in a banded pattern on the sample surface. From Fig. 1b, the proportion of the second phase is calculated to be about 16%. Figure 1c shows the grain size distribution of the samples with an average grain size of 30–40 μm.Fig. 1: Microstructures of Mg–9Gd-3Y–2Zn–0.5Zr alloy.a Optical images, b SEM images, and c grain size.Elemental compositions of the α-Mg matrix phase and the secondary phase were detected using EDS, as shown in Fig. 2. It is evident that besides magnesium, the α-Mg matrix phase contains 8.45 wt% of Gd. Gd and Mg both possess a close-packed hexagonal crystal structure. As a result, Gd demonstrates a remarkable solubility in Mg, reaching a maximum of 23.5 wt% at 819 K. This indicates a substantial solid solution strengthening effect19. The results in Fig. 2b indicate the occurrence of element segregation of Y, Gd, and Zn in the second phase.Fig. 2: EDS spectra of two organizations in Mg–9Gd–3Y–2Zn–0.5Zr alloy.a α-Mg matrix and b second phase.XRD spectrum of the samples is presented in Fig. 3, revealing the presence of α-Mg matrix and Mg12(Y, Gd)Zn. The low content of Zr element in the alloy resulted in its undetectable presence in the XRD pattern. The second phases play a vital role in refining the grain size of magnesium alloys during the deformation process, as they can interact with dislocations to hinder basal dislocation slip, thus serving as crucial strengthening phases in magnesium alloys.Fig. 3: XRD spectrum of Mg–9Gd-3Y–2Zn–0.5Zr alloy.Mainly composed of Mg and Mg12 (Gd, Y) Zn.Corrosion morphologyIn Fig. 4, the corrosion macro morphology of the test specimen after different durations of exposure in a simulated coastal storage environment is presented. After 284 h, the surface of the specimen accumulated a significant amount of white corrosion products, which were distributed in a fragmented manner. As the test duration increased, the specimen surface was entirely corroded, and the corrosion products increased and thickened.Fig. 4: Macroscopic morphology of corrosion after different periods of simulated coastal storage environment test.a 284 h, b 568 h, c 852 h, and d 1136 h.After simulating coastal storage environmental tests, the surface corrosion products of samples were studied using SEM. The results are shown in Fig. 5, and Table 1 lists the chemical compositions at different locations analyzed by EDS. Following the 284-h test, a blocky corrosion product layer covered the surface of the samples, with cracks and localized damage on the blocky corrosion product layer revealing granular corrosion products beneath. The corrosion products were primarily composed of C, O, Mg, and Gd, with minimal amounts of Cl detected. After 568 h of testing, the blocky corrosion product layer on the sample surface exhibited severe damage, with needle-like corrosion products appearing in localized areas. EDS measurements indicated a higher concentration of chlorine in the needle-like corrosion products, which are commonly observed in outdoor atmospheric exposure and laboratory immersion studies and are typically attributed to residual MgCl220,21. Subsequently, the damaged areas on the sample surface increased, with needle-like corrosion products containing high chlorine content found at each point of damage.Fig. 5: SEM morphology of the surface after different periods of simulated coastal storage environment test.a 284 h, b 568 h, c 852 h, and d 1136 h.Table 1 Chemical composition of corrosion products on the sample surface(wt%)The micrograph and elemental distribution of samples after exposure to a simulated coastal storage environment for varying durations are illustrated in Fig. 6. As seen in Fig. 6a, after 284 h of testing, the thickness of the corrosion product layer is 29 μm, exhibiting overall compactness with numerous cracks near the substrate, potentially serving as pathways for corrosive media penetration into the substrate. Subsequently, the corrosion product layer increases to 54 μm after 568 h of testing. Here, the region near to the substrate is relatively dense but contains longitudinal cracks, while the layer farther from the substrate is more porous. Over time, the corrosion product layer increases to approximately 100 μm and stabilizes. After 1136 h of testing, a significant enrichment of Cl is observed in the rust layer. In addition, it can be seen from the cross-sectional morphology of the specimen at 284 h and 568 h that the second phase seems to hinder the development of corrosion, which may be related to the galvanic effect of the second phase22.Fig. 6: Cross-section SEM images of the investigated alloys after different test times.a 284 h, b 568 h, c 852 h, and d 1136 h.The morphology of the corrosion pits, as depicted in Fig. 7, clearly illustrates the evolution of the corrosion process. After 284 h of testing, the sample surface showed numerous pits with a certain orientation, resembling the distribution characteristics of the second phase within the sample. By 562 h of testing, the sample surface is covered with large and deep corrosion pits, indicating that the early pits are still continuously developing in-depth and around them. As the experiment progresses, after 1136 h of corrosion, river-like corrosion pits can be observed, indicating severe corrosion on the sample surface.Fig. 7: Corrosion morphology of Mg–9Gd-3Y–2Zn–0.5Zr alloy after removing corrosion products.a 284 h, b 568 h, c 852 h, and d 1136 h.Electrochemical measurementsFigure 8 illustrates the polarization curves obtained from experiments in a simulated coastal storage environment at different intervals, along with the electrochemical parameters determined through Tafel fitting. After the 284-h test, the activation–passivation transition zone of the anode branch exhibited characteristics of oxide film dissolution and local corrosion23. As the experiment progresses, the anodic polarization curves in each period demonstrate typical active dissolution characteristics.Fig. 8: The potentiodynamic polarization curves of Mg–9Gd-3Y–2Zn–0.5Zr alloy after different periods of simulated coastal storage environment test.a The samples at different periods were measured in 1 wt% NaCl solution and b electrochemical parameters (mean ± SD).The self-corrosion potential (Ecorr) and corrosion current density (icorr) of the sample after different experimental periods were obtained through Tafel fitting, as shown in Fig. 8b. The Ecorr gradually shifted positively from −1.703 V to −1.685 V between 284 h and 568 h of the experiment and increased to −1.536 V after 852 h. A higher corrosion potential indicates a lower corrosion tendency as it reveals the material’s thermodynamic stability24. As the experiment progressed, the oxide layer on the samples was gradually destroyed, transitioning the sample from localized corrosion to general corrosion. A significant quantity of corrosion products can serve as a protective barrier in corrosive environments, thereby augmenting the thermodynamic stability of the specimen against corrosion. After 852 h to 1136 h of testing, a slight negative shift in Ecorr was observed, possibly attributed to fluctuations caused by the damage to the corrosion product film on the sample surface.EIS measurements were conducted on samples to delve deeper into the degradation of the protective film and the development of corrosion product films during testing, as illustrated in Fig. 9. It is evident that after 284 h of testing, the Nyquist plot of the sample exhibits high-frequency capacitive loops and low-frequency inductive loops. The capacitive loop in the high-frequency range can be attributed to the charge transfer process at the corrosion product film/electrolyte interface, while the inductive loop in the low-frequency range is associated with the relaxation process of intermediate species adsorbed on the surface during the anodic reaction25.Fig. 9: Electrochemical impedance spectra and equivalent circuit diagrams after different cycle tests in simulated coastal storage environments.a Nyquist plots, b bode plots of impedance, c bode plots of angle, and d equivalent circuits.Figure 9d illustrates the equivalent circuit models resulting from various experiments. Specifically, Fig. 9d1 corresponds to the equivalent circuit after 284 h, while Fig. 9d2 corresponds to the equivalent circuit from 568 to 1136 h. In these diagrams, Rs represents the electrolyte resistance, Rf and Rct correspond to the resistance of the corrosion product film and charge transfer resistance, respectively, and L and RL reflect the inductance and inductive impedance. Due to the non-uniformity of the surface corrosion product film, a constant phase element CPEf is employed to represent the double-layer capacitance of the corrosion product layer. The impedance (ZCPE) and capacitance (CCPE) of CPE are described according to the following formula26:$${Z}_{{CPE}}=\frac{1}{Q{(\omega \cdot i)}^{-n}}$$
(1)
$${C}_{{CPE}}={Y}_{0}^{\frac{1}{n}}{R}^{\frac{n-1}{n}}$$
(2)
The symbol ω represents the angular frequency (rad/s), and i denotes the imaginary unit. When n equals 1, the CPE is regarded as an ideal double-layer capacitor; whereas when 0.5 < n < 1, CPE is viewed as a non-ideal capacitor. By considering the distinctions in the equivalent circuit, the polarization resistance Rp can be determined using the provided formula:$${R}_{{P}}={R}_{{f}}+\frac{1}{\frac{1}{{R}_{{ct}}}+(\frac{1}{{R}_{{L}}})}$$
(3)
The fitting parameters are shown in Table 2 (see Table 2 for details). It is evident that the value of Rf remained relatively stable during the first two experimental cycles, but increased after the third cycle with the prolongation of the experimental duration, indicating the presence of a loosely packed and defective corrosion product film within the 568-hour test period. On the other hand, the polarization resistance RP increased gradually as the experiment progressed, suggesting a decrease in the sample’s dissolution rate, which aligns with the trend observed in the corrosion current density.Table 2 Parameters for fitting EIS spectraComposition of corrosion productFigure 10 shows the composition of corrosion products in the sample. The corrosion products of the samples after simulated coastal storage environment testing primarily consist of Mg(OH)2, MgO, MgCl2, Gd2O3, and Y2O3. MgO and Mg(OH)2 are common corrosion products in magnesium alloys, while the presence of Gd2O3 and Y2O3 indicates the involvement of Gd and Y in the corrosion process of the samples. The Pilling–Bedworth (P–B) ratio is commonly used to evaluate the density of oxide films. The P–B ratio between 1 and 2 signifies a relatively dense film, which can be calculated using Eq. (3)27,28:$${R}_{\text{P}-\text{B}}=\frac{{M}_{{Oxide}}\cdot {\rho }_{{Me}}}{n{M}_{\rm{Me}}\cdot {\rho }_{{Oxide}}}$$
(4)
Fig. 10: XRD patterns of Mg–9Gd–3Y–2Zn–0.5Zr alloy after different periods of simulated coastal storage environment testing.Composition of corrosion products after testing.In the given formula, M represents the relative molecular mass of the oxide, ρ denotes the density, and n signifies the quantity of metal atoms within the oxide molecule.According to Eq. (3), the Pilling–Bedworth ratios for MgO, Gd2O3, and Y2O3 are calculated to be 0.806, 1.751, and 1.613, respectively. The Pilling-Bedworth ratios of Gd2O3 and Y2O3 being greater than 1 suggest that corrosion product films containing these oxides can effectively enhance the compactness of the films. Additionally, as the experiments progress, the diffraction peaks of Gd2O3 and Y2O3 significantly intensify, indicating an enhanced protective effect of the corrosion product layer on the substrate.Corrosion kineticsIn Fig. 11, the corrosion weight loss and calculated instantaneous corrosion rate of samples in simulated coastal storage environment are presented. It can be observed that throughout the experimental duration, there is a well-fitted power-law relationship between the corrosion weight loss and the experimental time. The power value of the fitting function is less than 1, indicating that the corrosion product layer can mitigate the corrosion of the substrate.Fig. 11: Corrosion kinetics curves of Mg–9Gd–3Y–2Zn-0.5Zr alloy under simulated coastal storage environmental experiments.The corrosion rate of the sample gradually decreases (mean ± SD).Since there are few natural exposure tests for rare earth magnesium alloys, the test results are compared with the corrosion behavior of AZ31 alloys exposed in the marine environment. After one cycle of accelerated testing, the corrosion weight loss of the specimen was 17 g/m2, which was about 1/6 of the exposure of the dynamic shipping route in the western Pacific and 1/2 of the exposure of the Xisha Islands in 1 year20,29. In addition, we also found that the n value of the corrosion kinetics of AZ31 alloys is greater than 1 in the marine atmosphere, indicating that the corrosion product layer is less protective. Our test results show that the corrosion product layer of the sample has good protection, which may be related to the lower Cl concentration in the storage environment.
Corrosion behavior of the second phase in Mg–9Gd–3Y–2Zn–0.5Zr alloy under simulated coastal storage environment
