2023 Volume 64 Issue 4 Pages 925-930
Cu and Ni impurities in Mg alloys are deleterious contaminants that reduce the corrosion resistance of the alloy. Mg2Cu and Mg2Ni precipitates can cause significant anodic dissolution of the Mg matrix, owing to their potential difference. Suppression of these phases can prevent the deterioration of corrosion resistance. The neutralization of these impurities through the formation of Mg–Zn intermetallic phases has been studied, because the atomic radii of Cu and Ni are similar to that of Zn. As a result, the MgZn2 phase may precipitate during the rapid cooling that occurs during the solidification of the Mg–6 mass% Zn alloy, and introduce substitutional impurity atoms in the crystal lattice. Mg(Zn, Cu)2 and Mg(Zn, Ni)2 phases can be formed instead of Mg2Cu and Mg2Ni, in the presence of both of Zn and these impurities. The microstructures and corrosion properties of the Mg–6 mass% Zn alloy with various Cu or Ni concentrations are investigated in this work. The Cu and Ni impurities are concentrated into MgZn2 phase in the Mg–6 mass% Zn alloy without Mg2Cu or Mg2Ni formation when the concentrations of these impurities are within acceptable limits. Consequently, the corrosion rate of the Mg–6 mass% Zn alloy with 1.4 mass% Cu or 0.25 mass% Ni is almost the same as that of the alloy without Cu and/or Ni contaminations.
Mg alloys have many advantages for use as lightweight structural components. Therefore, the demand for Mg metal and alloys has increased notably in the past few decades. Consequently, recycling Mg from post-consumer scrap including Fe, Al, Cu, and other industrial metallic materials, is desirable for the sustainable use of Mg resources. Several sorting techniques for mixed metal scrap have been reported to recover the used Mg components.1–3) However, the sorted Mg scrap often contains other metals and contaminants. Fe, Cu, Ni, and Co are especially harmful impurity elements that degrade the corrosion resistance of Mg alloys.4) These trace impurities form the second-phase particles, and consequently trigger the anodic dissolution of Mg matrix through microgalvanic corrosion. The solubility of Fe in a pure Mg melt is 0.018 mass% at 650°C.5) Hanawalt et al. reported that Fe contamination >50 ppm increased the corrosion rate of pure Mg.6) Therefore, Fe removal by the addition of Mn is commercially practiced to prevent the degradation of the corrosion resistance of Mg alloys.4) On the other hand, the solubility of Ni and Cu in the Mg melt is much higher than that of Fe; for example, 32 mass% Ni and 70 mass% Cu, in a binary system, dissolve into the Mg melt at 650°C.7) These impurity elements form intermetallic phases of Mg2Ni or Mg2Cu particles in the Mg alloys,4) and are dispersed around the grain boundaries of the Mg matrix, acting as strong cathodic sites during corrosion. This result leads to the anodic dissolution of the Mg matrix, and the corrosion resistance of the Mg alloys with these deleterious elements is extremely poor. Therefore, commercial Mg alloys restrict the concentrations of these impurities, for example, to less than 0.03 mass% for Cu and 0.002 mass% for Ni in the AZ91D alloy.
Mg–6 mass% Zn alloys such as ZK60 are used in castings, extruded bars, and rolled sheets because of their superior mechanical properties, when compared to that of the AZ31 alloy. These system alloys are hardened by solution treatment and the subsequent aging.8) The phase diagram of the Mg–Zn system shows that the Mg–6 mass% Zn alloy is composed of α-Mg matrix and MgZn intermetallic phases.9,10) However, because of rapid cooling and the subsequent aging, a metastable phase of MgZn2 may be precipitated from the supersaturated solid solution.11) Moreover, the MgZn2 phase is formed by the decomposition of Mg7Zn3, also known as Mg51Zn20, the eutectoid phase, during cooling.12) This microstructural evolution produces Mg–Zn alloys with excellent mechanical properties.
Previously, the authors had reported that the corrosion resistance of the Mg–6 mass% Zn alloy with trace amounts of Cu was almost same as that of the Mg–6 mass% Zn alloy without impurities.13) This is because the Cu impurity elements are concentrated in the MgZn2 particles on the grain boundaries of the α-Mg matrix, instead of precipitating as Mg2Cu. Zn and Cu are neighboring elements in the periodic table, with a very small difference in their atomic radii (<5%). In addition, the MgZn2 phase is known as the Laves phase, and Zn in MgZn2 phase cannot be completely replaced with Cu as MgCu2, owing to the difference in valency.14,15) This may result in the partially substituted Cu element dissolving into the MgZn2 phase. The atomic radius of Ni is also similar to the atomic radii of these elements and therefore, Ni can also dissolve into MgZn2 phase. Thus, as described above, the addition of Zn to the Mg alloy might prevent the corrosion degradation caused by the Cu and Ni impurities.
In this study, the tolerance limits of the Cu and Ni impurity elements on the corrosion resistance of Mg–6 mass% Zn alloys are investigated. These impurities dissolve into MgZn2 phase particles instead of the deleterious precipitates of Mg2Cu and Mg2Ni, causing strong cathodic sites during Mg corrosion. The neutralizing effect of the MgZn2 phase, in the Mg alloy with impurities, on the corrosion resistance and the differences in acceptable concentrations of Cu and Ni impurities are discussed.
Mg–6 mass% Zn alloys with various concentrations of Cu and Ni impurities were prepared by smelting using an electric tube furnace at 800°C in Ar atmosphere. The raw materials—pure Mg (99.9%) ball, pure Zn (99.99%) granular powder, pure Cu (99.99%) or Ni (99.99%) particles were mixed in an MgO crucible in a predetermined composition to produce a total alloy weight of 30 g. The molten alloy in the furnace was maintained for 3 hours at 800°C and stirred every 30 min to completely dissolve the raw materials. Finally, the ingot was furnace-cooled (∼200°C/h of cooling rate) and pulled out from the crucible. The size of the ingot was approximately φ 30 × 25 mm, and the vertical cross-section of the ingot near the center was prepared for the following measurements. The alloy composition of each specimen was determined through X-ray fluorescence spectroscopy (JEOL, JSX-1000S X-ray fluorescence spectrometer). The major compositions of these alloys are listed in Table 1. The microstructures were observed using optical microscope (OLYMPUS GX53 with DP74 digital camera) and a scanning electron microscope (SEM, JEOL JSM-6060LV), and the local compositions of each phase was measured by energy-dispersive X-ray (EDX) analysis (JEOL EX-54013NSX detector). X-ray diffraction (XRD) analysis was used to determine the intermetallic phase using a RIGAKU RINT-2500N X-ray diffractometer, and the phase analysis was conducted using the integrated X-ray powder diffraction software (RIGAKU PDXL2).
An immersion test in a 5 mass% NaCl aqueous solution was performed for these specimens, to measure the corrosion resistance. Hydrogen gases were collected into the measuring cylinder during the immersion test, for 4 h. The total volume of hydrogen generated during the immersion test over 4 h was measured and converted into the corrosion rate. After the immersion test, the corroded surface of each specimen was observed using an optical microscope.
The as-solidified microstructures of typical specimens are shown in Fig. 1. The average grain size of the α-Mg matrix phase was approximately 100–200 µm, regardless of the impurity composition. The second phases were dispersed around the grain boundaries of the matrix. The volume fractions of the second phases were almost constant upon increasing the Cu concentration. The second phases of all specimens consisted of MgZn and MgZn2 phases. In particular, Mg–6 mass% Zn with more than 1.4 mass% Cu alloy contained the Mg2Cu phase, as observed in the results of XRD analysis as shown in Fig. 2(a). SEM micrographs of the Mg–6 mass% Zn alloys with 1.0 mass% Cu and 2.0 mass% Cu are shown in Fig. 3. There were two types of second phases in the Mg–6 mass% Zn–1.0 mass% Cu alloy. The lamellar regions were the precipitate phases of MgZn and α-Mg eutectoid particles, and the nonlamellar phase of MgZn2 phase was observed. The lamellar region and MgZn2 phase were adjacent because the MgZn2 phase was metastable and could have been formed during the eutectoid decomposition of Mg51Zn20. The MgZn phase was stably precipitated owing to the slow cooling of Mg51Zn20.16) Typical compositions of each phase are shown in Fig. 3. From the results of the EDX analysis, it was observed that Cu was concentrated in the MgZn2 phase, and that the MgZn phase did not contain Cu. In contrast, a high Cu concentration was detected in the block-shaped second phase of the Mg–6 mass% Zn–2.0 mass% Cu alloy. When combining this with the results of the XRD analysis, it can be assumed that these particles are of the Mg2Cu phase.
As-solidified microstructures of Mg–6 mass% Zn alloys with various Cu concentrations, (a) without Cu, (b) 0.5 mass% Cu, (c) 1.0 mass% Cu, (d) 1.4 mass% Cu, (e) 1.6 mass% Cu, (f) 2.0 mass% Cu.
XRD patterns of typical specimens of Mg–6 mass% Zn alloys with and without Cu/Ni impurities, (a) Mg–6 mass% Zn alloys with/without Cu and (b) Mg–6 mass% Zn alloys with/without Ni.
SEM microstructures and typical compositions of each phase, (a) Mg–6 mass% Zn–1.0 mass% Cu alloy and (b) Mg–6 mass% Zn–2.0 mass% Cu alloy.
The corrosion surfaces of the Mg–6 mass% Zn alloy with different Cu concentrations are shown in Fig. 4. The metallic surface, in the light-grey area in Fig. 4(a), remained in the α-Mg matrix phase in the Mg–6 mass% Zn–1.0 mass% Cu alloy after 4 h of the immersion test. The second phases with MgZn and MgZn2 as the white areas were well-distributed in the light grey-area. This result indicates that the anodic dissolution of the matrix was suppressed by the surrounding second phases. In the dark-grey area, the matrix was corroded owing to the loss of the second phases. In contrast, the Mg–6 mass% Zn–2.0 mass% Cu alloy appeared to be almost completely corroded on matrix grains, as shown in Fig. 4(b). From the results of the second-phase identification, it can be assumed that the Mg2Cu phase in the Mg–6 mass% Zn–2.0 mass% Cu alloy accelerates the anodic dissolution of the Mg matrix phase. In the Mg–6 mass% Zn alloy with a high Cu concentration, a certain amount of Cu was distributed in the MgZn2 phase, while excess Cu formed an Mg2Cu phase and had a deleterious effect on the corrosion resistance of the Mg alloys. Figure 5 shows the relationship between the corrosion rate of the alloy and Cu content. The corrosion rates of the Mg–6 mass% Zn alloy with up to approximately 1.4 mass% Cu were almost the same as that of the Mg–6 mass% Zn alloy without Cu, whereas the Mg–6 mass% Zn alloy with more than 1.4 mass% Cu exhibited a high corrosion rate, which increased with increasing Cu content in the alloy. This result indicated that the acceptable concentration limit of Cu in the Mg–6 mass% Zn alloy was approximately 1.4 mass%, although it had a possibility to vary with the volume fraction of the MgZn2 phase when the alloy was subject to certain types of heat treatment. As described above, an acceptable concentration of Cu in Mg–6 mass% Zn for maintaining the corrosion resistance was determined.
Macroscopic images after immersion test for 4 hours, (a) Mg–6 mass% Zn–1.0 mass% Cu alloy and (b) Mg–6 mass% Zn–2.0 mass% Cu alloy.
Relationship between corrosion rate and Cu concentration in Mg–6 mass% Zn alloy.
The lattice parameters of the MgZn2 phase were calculated from the XRD peak of $(20\bar{2}2)$ plane. The c-axis lattice parameters of the hexagonal MgZn2 phase were almost constant (approximately 8.598 Å) with increasing Cu concentration. The relationship between the a-axis lattice parameter of the MgZn2 phase and Cu concentration is shown in Fig. 6. The lattice parameter decreased with increasing Cu concentration up to 1.4 mass% Cu, and was almost constant at exceeding 1.4 mass%. This result indicates that the solubility of Cu in the MgZn2 phase was maximized in the Mg–6 mass% Zn–1.4 mass% Cu alloy. As the result, the excess Cu were precipitated as a harmful Mg2Cu intermetallic phase.
The relationship between MgZn2 lattice parameter of a-axis and Cu concentration.
The Mg51Zn20 phase was detected, instead of the MgZn phase, in the Mg–6 mass% Zn alloys with Cu impurities, in the XRD analysis as shown in Fig. 2(b). This is because the cooling rate of the alloy is determined to precipitate which phases during eutectoid decomposition of Mg51Zn20. When the cooling rate is low, the Mg51Zn20 phase decomposes to MgZn + α-Mg.11,16) The MgZn2 phase is precipitated by the decomposition of Mg51Zn20 at a relatively rapid cooling rate. The MgZn2 phase was successfully distributed as the second phase in the Mg–6 mass% Zn alloy with Ni impurities, although there were some other precipitates. The SEM microstructures of the Mg–6 mass% Zn–0.25 mass% Ni and Mg–6 mass% Zn–0.35 mass% Ni alloys are shown in Fig. 7. Second-phase particles were dispersed on the grain boundaries of the α-Mg matrix. XRD and EDX analyses revealed that these particles were MgZn2 and Mg51Zn20 in the Mg–6 mass% Zn–0.25 mass% Ni alloy. The MgZn2 phase was adjacent to the Mg51Zn20 phase because of the metastable precipitation, as described above. Ni was also concentrated in the MgZn2 phase, similar to the Mg–6 mass% Zn alloy with less than 1.4 mass% of Cu contamination, and the Mg51Zn20 phase included no impurities of Ni, similar to the MgZn phase in the Mg–6 mass% Zn alloys with Cu. On the other hand, angular particles of Mg2Ni were independently distributed in the Mg–6 mass% Zn–0.35 mass% Ni alloy. The Ni concentration in the MgZn2 particle, as a result of the EDX analysis, was approximately 10 at% both in the Mg–6 mass% Zn–0.25 mass% Ni and Mg–6 mass% Zn–0.35 mass% Ni alloys. The excess Ni, which could not dissolve into the MgZn2 phase, precipitated as Mg2Ni in the Mg–6 mass% Zn alloys with more than 0.35 mass% Ni.
SEM microstructures and typical compositions of each phase, (a) Mg–6 mass% Zn–0.25 mass% Ni alloy and (b) Mg–6 mass% Zn–0.35 mass% Ni alloy.
Figure 8 shows the results of corrosion immersion test of the Mg–6 mass% Zn alloys with various Ni concentrations. The corrosion rate of the Mg–6 mass% Zn with up to 0.25 mass% Ni concentration was constant, and it increased rapidly as the Ni content increased above 0.25 mass% Ni. Therefore, the acceptable concentration of Ni in the Mg–6 mass% Zn alloy was determined as 0.25 mass%. The relationship between the lattice parameters of MgZn2 and Ni concentrations is shown in Fig. 9. The lattice parameter of the a-axis was also decreased with increasing Ni concentration. However, it was almost constant at more than 0.25 mass% Ni.
Relationship between corrosion rate and Ni concentration in Mg–6 mass% Zn alloy.
The relationship between MgZn2 lattice parameter of a-axis and Ni concentration.
The acceptable concentration of Ni in the Mg–6 mass% Zn alloy was approximately one-fifth that of Cu; however, there were no differences in the overall distributions of MgZn2 phase, between the Cu and Ni-contaminated alloys. The concentration of Cu in the MgZn2 phase is approximately 15–25 at% measured by EDX in Mg–6 mass% Zn–1.0 and 2.0 mass% Cu alloys. This result indicates that there is a considerable difference in the solubilities of the Cu and Ni impurities in the MgZn2. Both the Cu and Ni impurities exist in the MgZn2 phase as substitutional solutes and are designated as Mg(Zn, Cu)2 and Mg(Zn, Ni)2. Zn, Cu and Ni are transition elements, although not strictly for Zn, and are neighboring elements in the periodic table with atomic radii of 134 pm, 128 pm and 124 pm, respectively. The difference in atomic radius between Zn and the impurity elements is less than 10%. This is a reasonable basis for the formation of a substitutional intermetallic phase, as per the Hume—Rothery rules.14) In addition, the solubility of Cu in the MgZn2 phase is larger than that of Ni because the difference in atomic radius from that of Zn is smaller. Thus, Cu can dissolve more easily in the MgZn2 phase than Ni. The lattice parameter of the MgZn2 phase decreases upon increasing both Cu and Ni concentrations, until the maximum acceptable concentrations are reached. The lattice parameter of a-axis in the 0.25 mass% Ni alloy (approximately 5.224 Å) is larger than that in the 1.4 mass% Cu alloy (5.125 Å), as shown in Figs. 6 and 9. This result suggests that the maximum solubility of impurity elements may not on the atomic radius alone. Even though the solubility of Ni in the MgZn2 phase is lesser than that of Cu, the value of the acceptable concentration limit is sufficiently large when compared to that of the practical Mg alloy composition, for example, ≤0.002 mass% Ni in AZ91D. There is a possibility of neutralizing the elements deleterious for the corrosion resistance of Mg alloys; in addition to Cu and Ni, Fe and Co, may be simultaneously stabilized in the Mg–6 mass% Zn alloy when the MgZn2 phase is distributed. The mechanism of formation of the Mg(Zn, Cu)2 and Mg(Zn, Ni)2 phases, instead of Mg2Cu and Mg2Ni, is uncertain because the MgZn2 is a metastable phase in Mg–6 mass% Zn alloys. Sturkey and Clark reported that the MgZn2 phase was precipitated by the eutectoid decomposition of Mg7Zn3 (or Mg51Zn20) phase with the formation of Mg solid solution below 235°C in a Mg–5 mass% Zn alloy.16) This result indicates that the fraction of MgZn2 may vary depending on the heat treatment. The behavior of impurities during solidification or cooling is not clear, and the mechanism of formation of Mg(Zn, Cu)2 and Mg(Zn, Ni)2 in the presence of Cu and Ni should be investigated in further studies.
Deleterious impurities of Cu and Ni degrading the corrosion resistance of Mg alloys were successfully neutralized by the formation of Mg(Zn, Cu)2 and Mg(Zn, Ni)2 phases instead of Mg2Cu and Mg2Ni. The corrosion rates of the Mg–6 mass% Zn alloys with certain impurity concentrations (less than the acceptable limit) were almost the same as that of the alloys without impurities. The acceptable concentration limits of these impurities in the Mg–6 mass% Zn alloy were approximately 1.4 mass% Cu and 0.25 mass% Ni, respectively. These results originated from the substitutional dissolution of impurities into the MgZn2 phase and the difference in the solubility of the Cu and Ni impurities in the MgZn2 phase.
