Investigation of the Susceptibility of the WE54 Magnesium-Based Alloy to Stress Corrosion Cracking

The slow strain rate test (SSRT) was carried out in accordance with ASTM G129—00 (2013) (Ref 29)—the tests were carried out on a Zwick Kappa 50DS testing machine, and the samples were made in accordance with the standard PN-EN ISO 7539-4 (Ref 30). The alloys to be tested were formed into cylindrical specimens with a gauge length of 20 mm and a diameter of 5 mm. The specimens obtained were cleaned with acetone and then subjected to the slow strain rate test in four variants:

Each of the SSRT variants was carried out at ambient temperature, at the same strain rate of = 9·10⁻⁷ s⁻¹. The corrosive environment was 0.1 M Na₂SO₄ solution. Each variant was applied to two specimens. The following parameters were obtained in the tests:

The materials to be examined were the fracture surfaces obtained as a result of the SSRT tests. The specimens were cleaned so that the corrosion products that formed on some fracture surfaces would not hinder the examination. For this purpose, they were cleaned in a CrO₃ + 10 g/l AgNO₃ solution (200 g/l) for 1÷3 min (as necessary). The specimens were then cleaned three times in an ultrasonic washer with the use of acetone, each time using new acetone. The fracture surfaces were analyzed using a Hitachi S-3400 N scanning electron microscope.

Fracture profiles were obtained by grinding one of the fracture surfaces with SiC abrasive papers with a grit size of up to 1200. The profiles were then polished using Struers diamond suspensions with a particle size of up to 1 µm. The last stage was the final polishing using 0.05 µm Al₂O₃ suspension. Due to the high susceptibility of magnesium-based alloys to corrosion in water environments, the time between the polishing and drying was as short as possible and did not exceed 10 s.

The test specimens were subjected to cathodic hydrogen charging in an 0.1 M Na₂SO₄ aqueous solution and in a 0.1 M Na₂SO₄ solution with the addition of As₂O₃ (5 mg/l). The current density was 50 mA/cm², and the hydrogen charging time was 24 h.

The hydrogen concentration in the specimens was analyzed using a LECO TCHEN600 elemental analyzer. Hydrogen concentration was analyzed for the specimens:

In the initial condition, precipitates of intermetallic phases of various shapes and diameters not exceeding 10 µm were identified in the microstructure of WE54 (Fig. 1). The precipitates on the section perpendicular to the extrusion direction were distributed evenly (Fig. 1a). On the section parallel to the extrusion direction, the characteristic clustering of precipitates into bands corresponding to the extrusion direction was observed (Fig. 1b).

Because SSC tests were carried out in a corrosive environment, the subject of the tests was the phase composition of the WE54 alloy in the delivery condition (Fig. 1c) and after exposure in the analyzed corrosive environments (Fig. 1d and e).

The stress-strain curves obtained as a result of the four SSRT variants for the WE54 specimens are shown in Fig. 2. They have a typical shape, with an elastic and plastic strain zone, but with no distinct yield strength. Thus, a yield strength R_0.2 was adopted for the specimens tested. Irrespective of the test variant, its value was very similar and was about 200 MPa. The highest tensile strength (R_m ≈ 290 MPa) was recorded for the specimens strained in air—both those that had not been exposed to the corrosive environment (Fig. 2a) (variant I) and those that had been exposed to the corrosive environment (Fig. 2b) (variant II).

A considerable decrease in strength was observed for the specimens strained during exposure to the corrosive environment. The specimens tested under cathodic polarization had the lowest strength (R_m ≈ 230 MPa) (Fig. 2d) (variant IV). The strength of the specimens tested at open circuit (Fig. 2c) (variant III) was R_m ≈ 250 MPa. The best plasticity was displayed by the specimens tested in air (ε ≈ 24%, Z ≈ 20%). A considerable decrease in plasticity was observed for the specimens strained during exposure to the corrosive environment. The values recorded for the specimens tested at open circuit (Fig. 2c) (variant III) were: elongation ε = 6.5÷8.9% (depending on the specimen) and reduction in area Z = 7.8%. The values recorded for the specimens tested under cathodic polarization (Fig. 1d) (variant IV) were: elongation ε = 6.0% and reduction in area Z = 10%. Detailed mechanical property results of the SSRT tests on the WE54 specimens are given in Table 1.

The results of the analysis of the hydrogen concentration in the WE54 specimens after the SSRT tests showed that the specimens tested in air had the lowest hydrogen concentration (C_H = 25.0 ± 0.0 ppm). The highest hydrogen concentration was recorded for the specimens tested under cathodic polarization (C_H = 31.7 ± 0.6 ppm) (variant IV) and the specimens tested in air following exposure to the corrosive environment under cathodic polarization (C_H = 31.3 ± 0.6 ppm). The specimens tested under open-circuit conditions had a medium hydrogen concentration (C_H = 27.3 ± 0.6 ppm) (variant III).

The fracture surfaces obtained as a result of the SSRT tests were subjected to a fractographic analysis aimed at determining the influence of particular environments on the alloy’s microstructure and mechanical properties. The analysis covered fracture surfaces, specimen side surfaces, and fracture profiles. This made it possible to examine not only the changes in the microstructure on the surfaces coming into contact with the corrosive environment, but also the depth of those changes.

The WE54 specimens tested in air were characterized by significant plastic deformation in the form of reduction in area (Fig. 3a). After tests in SSC conditions, there is a lack of visible plastic deformation of the samples—changes in sample diameter are very small (Fig. 5a, 7a, and 9a). Numerous microvoids of various sizes were identified on the specimens’ side surfaces, however, not larger than 20 µm (Fig. 3b). The fracture surface had a morphology typical of ductile fracture, with no cracks (Fig. 4). The morphology of microcracks on samples after SSRT tests in a corrosive environment is different (variants II-IV) (Fig. 5b, 7b, and 9b).

A distinct reduction in area was also observed in the specimens tested in air following prior exposure to the corrosive environment under cathodic polarization (Fig. 5a) (variant II). Moreover, on their side surfaces, there were microvoids and a large number of microcracks propagating perpendicularly to the force applied during the strain test (Fig. 5). The fracture surface (Fig. 6) had a morphology typical of ductile fracture (Fig. 6b, and d); however, there were also brittle fracture areas with numerous cracks (Fig. 6c).

The reduction in area was much smaller for the specimens tested in the corrosive environment under open-circuit potential conditions (Fig. 7a) (variant III) than for the specimens tested in air. On their side surfaces, there were corrosion pits and cracks of various lengths, propagating perpendicularly to the force applied during the strain test (Fig. 7b). Numerous microcracks of various lengths were identified on a considerable part of the fracture surfaces (Fig. 8).

The specimens strained in the corrosive environment under cathodic polarization also experienced a small reduction in area (Fig. 9a) (variant IV). There were numerous cracks and small voids on their side surfaces (Fig. 9b). A great number of cracks were identified on the fracture surface. Most cracks were several micrometers long (Fig. 10c and d), but there were also many larger cracks on the fracture surface (Fig. 10b).

A detailed analysis of the nature of the fracture of a specimen tested under cathodic polarization (variant IV) revealed that there were local intergranular areas along the edges (Fig. 11). It needs to be noted that there were also transgranular cracks in those areas.

After the WE54 specimens had undergone the SSRT tests, their fracture profiles were examined (Fig. 12). In the microstructure of the specimen tested in air (variant I), there were cracks located within the precipitates of the phases present in the alloy (Fig. 13). The crack propagation direction was perpendicular to the force applied during the strain test. The microstructure also contained voids, resulting from the plastic deformation of the alloy’s matrix (Fig. 13a). The voids initiated in the areas where cracks were present or, less frequently, at the phase-matrix interphase boundaries.

The microstructure of the specimen strained in air following exposure to the corrosive environment under cathodic polarization (variant II) also had cracks in phase precipitates and voids (Fig. 14). Cracks were also identified in the matrix, by the fracture surface (Fig. 14a) and by the side surface of the specimen (Fig. 14b).

In the case of the specimens strained in the corrosive environment, both at open circuit (Fig. 15) (variant III) and under cathodic polarization (Fig. 16) (variant IV), cracks were identified in the phase precipitates present in the alloy’s microstructure (Fig. 15a and 16a). However, the number of voids that accompanied them and their dimensions were much smaller in comparison with the specimens tested in air. Numerous cracks were also identified in the matrix, by the fracture surface (Fig. 15a and 16a) and by the side surface of the specimen (Fig. 15b and 16b).

The stress-strain curves obtained as a result of the SSRT tests (Fig. 2), and the results of the mechanical tests (Table 1) indicated that the WE54 alloy tested was susceptible to stress corrosion cracking in a 0.1 M Na₂SO₄ solution. This susceptibility was manifested in a deterioration in the plastic and strength properties, and characteristic of stress corrosion cracking (Ref 14). A similar susceptibility to stress corrosion cracking in an environment containing Na₂SO₄ was also identified, e.g., for pure magnesium, the AZ91 and ZK60-A alloys, and Mg-Mn-Ce alloys (Ref 15-19). In order to determine the level of stress corrosion cracking susceptibility of the alloys tested, the hydrogen embrittlement index (1) was calculated for each of the specimens subjected to SSRT variants 2, 3, and 4.

where X_pow—a given parameter (e.g., ε, R_m) recorded for the specimens tested in the non-corrosive environment—air (SSRT variant I), X_n—a given parameter (e.g., ε, R_m) recorded for the specimens tested in the corrosive environment—under SSRT variant n (II, III, IV).

The hydrogen embrittlement index describes (as a percentage) the change in the properties of a material resulting from strain in the corrosive environment. The change described is compared with the value of a given property recorded for strain in a non-corrosive environment—most frequently, air. In the case of the specimens described in this study, the hydrogen embrittlement index was calculated based on changes both in the plastic properties (ε)—HEI_ε and in the strength properties (R_m)—HEI_Rm. The HEI_ε and HEI_Rm values obtained for the specimens tested are shown in Fig. 17.

The analysis of the results obtained indicated that the exposure of WE54 to a 0.1 M Na₂SO₄ solution caused a major deterioration in its mechanical properties, with the plastic properties deteriorating noticeably more severely—even by 70÷80%. The decrease in the strength properties was not so significant and reached 15÷20%.

The most severe deterioration in mechanical properties was observed for the specimens tested in the corrosive environment under cathodic polarization (SSRT variant IV). Under such conditions, a specimen does not corrode (due to cathodic protection), so the deterioration in its mechanical properties must result from other processes. The most probable cause is the presence of hydrogen, generated in a cathodic reaction, in the environment. Hydrogen, due to its short atomic radius, can penetrate the alloy’s microstructure and travel in it relatively easily, interacting with microstructural elements being so-called hydrogen traps. The hydrogen accumulated in the microstructure leads to internal stresses and/or the formation of brittle hydride phases. The combination of stresses resulting from the interaction of hydrogen on the microstructure and external mechanical loads may lead to the initiation and growth of microcracks that may have been the cause of the deterioration in mechanical properties observed for the alloy in question. Due to the poor protective properties of the passive layer forming on the surface of magnesium-based alloys, it does not prevent hydrogen effectively from penetrating the alloy’s microstructure. Moreover, due to the strain inflicted, the layer is constantly damaged, revealing the metallic substrate and facilitating the penetration of hydrogen into the alloy.

A major deterioration in mechanical properties was also observed for the specimens tested in the corrosive environment under open-circuit potential conditions (SSRT variant III) (Fig. 2). Their elongation deteriorated by approx. 55÷70%, whereas their tensile strength decreased by 4÷15%. As in the case of the specimens tested under cathodic polarization, also for the specimens tested at open circuit, the cause of the deterioration in mechanical properties may have been the presence of hydrogen in the alloy’s microstructure. However, in the case of the specimens tested at open circuit, the source of hydrogen was corrosion processes resulting from its exposure to the corrosive environment. Due to the fact that magnesium-based alloys have poor electrochemical corrosion resistance, the intensity of corrosion processes and the quantity of hydrogen thus generated may be substantial. Moreover, the corrosion pits (Fig. 14, 15, 16) forming on the surface of the alloy act as stress concentrators and areas facilitating the penetration of hydrogen into the alloy (Ref 8), thus facilitating the initiation of cracks in the alloy. Literature data inform that the worst deterioration in the mechanical properties of magnesium-based alloys is observed when they are subjected to strain in a corrosive environment under cathodic polarization, as well as under open-circuit potential conditions (Ref 20-24).

It should also be noted that the phase precipitates present in the WE54 microstructure may form galvanic microcells with the alloy’s matrix. A large number of galvanic microcells generate larger quantities of hydrogen and form corrosion pits playing role of stress concentrators and facilitating the penetration of hydrogen. Moreover, the main components of those precipitates are magnesium, yttrium, and neodymium, which, coming into contact with hydrogen, may form hydride phases affecting adversely the alloy’s properties. Literature data (Ref 25, 26) indicate that as the concentration of these elements in magnesium-based alloys increases, the concentration of hydrogen also increases. Thus, an increased volume fraction of intermetallic phase precipitates containing rare earths may contribute both to an increased quantity of hydrogen generated in corrosion processes and an increased absorption of hydrogen by the alloy.

In the case of the specimens tested in air after prior exposure to the corrosive environment under cathodic polarization (SSRT variant II), a minor deterioration in the mechanical properties was observed (Fig. 2). In the light of this fact, it can be stated that the received values indicate a minor effect of the environment on the mechanical properties of the alloy tested, or even the absence of such effect. This may result from the absence of external mechanical loads during the exposure of the specimens tested to the corrosive environment. One of the basic roles performed by external mechanical loads is the constant damaging of the passive layer forming on the alloy surface and thus continuous revealing of the metallic substrate through which hydrogen is able to penetrate deep into the alloy more easily. In the absence of external mechanical loads, the passive layer has better protective properties and is more effective in reducing the penetration of hydrogen into the alloy’s microstructure. Another major factor is hydrogen diffusivity in the crystal lattice of magnesium (Ref 7-11). The literature provides very different data in this respect; however, most frequently, authors claim it to be 10⁻¹²÷10⁻¹³ m²/s. This being a very small value, it may be very difficult for hydrogen to diffuse at long distances in a short period of time. Thus, hydrogen penetrating the alloy should accumulate in a thin layer close to the surface, with its depth depending on the exposure time. At the same time, the presence of hydrogen in the alloy should lead to a deterioration in its properties only in the areas where it is actually present, which in the case of short exposure times is only a small fraction of a specimen’s cross section and may have a negligible effect on the properties of the entire specimen. Literature data concerning the effect of pre-exposure of magnesium-based alloys on their mechanical properties both confirm its significant influence on deterioration in the alloys’ properties (Ref 7, 20) and deny such significant influence (Ref 21). This information suggests that hydrogen diffusivity in magnesium-based alloys is sufficient for hydrogen to be able to move freely in said alloys. A more important factor reducing the penetration of hydrogen into an alloy appears to be the presence of a passive layer (Ref 7, 11), whose protective properties depend crucially on the chemical composition of the alloy on which it forms and on the chemical composition of the environment to which it is exposed.

An analysis of the concentration of hydrogen present in the specimens obtained as a result of the SSRT tests revealed that the specimens tested in the corrosive environment were characterized by the highest concentration of hydrogen in their volumes. The lowest hydrogen concentration, in turn, was recorded for the non-exposed specimens strained in air. An identical correlation was observed during the analysis of the hydrogen embrittlement indices HEI_ε and HEI_Rm, and the most severe deterioration in mechanical properties was recorded for the specimens tested in the corrosive environment under cathodic polarization (Fig. 17).

The results confirm the significant role of hydrogen in the stress corrosion cracking processes occurring in the alloy studied. Hydrogen concentration was also higher in the WE54 specimens subjected to SSRT variants II and III than in the non-exposed specimens strained in air. However, hydrogen concentration did not increase proportionately to the deterioration in the specimens’ properties in the case of each of these two variants. This may result from the fact that the measurement of the actual hydrogen concentration in the specimens analyzed may be disrupted by the presence of the remains of an Mg(OH)₂ layer, forming on the alloy surface in contact with the corrosive environment, on the surface and in the material’s discontinuities (Ref 7). Moreover, the results obtained may have been affected by the prolonged, and not identical for each sample, time between the SSRT test and the hydrogen concentration analysis. Bearing in mind that the hydrogen concentrations recorded for the specimens were low, up to 50 ppm, a change in hydrogen concentration by several ppm, caused by its escape from the material or the presence of Mg(OH)₂ in the sample tested, would significantly affect the result obtained (Fig. 18).

The analysis of the fracture surfaces obtained as a result of the SSRT tests showed that the WE54 specimens tested in air (variant I) displayed ductile fracture (Fig. 3 and 4), with the characteristic voids and bridges forming due to the considerable plastic deformation. The entire fracture surfaces obtained in this test variant had a very similar morphology. No cracks were identified on the surfaces. An analysis of the side surfaces of these specimens (Fig. 3 and 5) did not reveal the presence of any cracks, either. However, voids were detected, formed in areas where phase precipitates were present as a result of the considerable plastic deformation, manifested also in the form of a clearly visible reduction in area (Fig. 3a). An analysis of the fracture profiles of the WE54 specimens obtained in the SSRT test in air (variant 1) confirmed the absence of cracks in the specimens. None were found either on the fracture surfaces (Fig. 12a, 13a and 14a) or on the specimen side surfaces (Fig. 13b, 14b and 15b). However, intermetallic phase precipitate cracks were detected in the microstructure along with the associated voids (Fig. 19).

The most likely cause of their formation is stress occurring in the alloy’s microstructure during the SSRT test. This is confirmed by the direction in which the cracks were oriented, and it was perpendicular to the force applied during the strain test. This is clearly visible in Fig. 19. The voids accompanying the cracks form when fragmented phase precipitates move along with the surrounding matrix, which undergoes plastic strain. It should be noted that thus formed phase precipitate cracks have not been observed to initiate cracks in the volume of the neighboring solid solution.

The fracture surfaces of the WE54 specimens tested under the other SSRT variants (II, III, IV) (Fig. 6, 8 and 10) were characterized by the presence of both ductile fracture areas and brittle (mainly transgranular) fracture areas. In the case of the specimens tested in the corrosive environment under cathodic polarization, intergranular fracture was detected in some areas, near the edges of the fracture surfaces (Fig. 11). The literature concerning magnesium-based alloys informs that stress corrosion cracking is often indicated by the presence of transgranular brittle fracture areas on the fracture surfaces (Ref 10, 19, 21, 23, 27). Additionally, fracture surfaces of a mixed nature are often encountered with transgranular fracture areas and scarce intergranular fracture areas (Ref 7, 11). Intergranular fracture is the least frequently encountered type of fracture (Ref 28). A characteristic feature of the fracture surfaces of the specimens tested under SSRT variants II, III, and IV was the presence of secondary cracks. A particularly large number of such cracks were observed on the surfaces of the specimens tested in the corrosive environment, both at open-circuit conditions (variant III) and under cathodic polarization (variant IV). The great majority of those cracks were short, and their lengths were comparable to the size of a grain in the alloy. Moreover, it was observed that the cracks were often parallel to one another within small areas whose size was comparable to the size of a grain in the specimens tested (Fig. 20). This fact suggests that within a single grain, such cracks formed on parallel crystallographic planes. Cracks were also identified on the side surfaces of the specimens analyzed (Fig. 5, 7 and 9). They had various sizes and their propagation direction was perpendicular to the force applied during the strain test. Additionally, corrosion pits were identified on the side surfaces of the specimens tested during exposure to the corrosive environment under open-circuit conditions (Fig. 7).

An analysis of the fracture profiles of the WE54 specimens tested in air following exposure to the corrosive environment (variant 2) and in the corrosive environment under open-circuit conditions (variant 3) and under cathodic polarization (variant 4) also revealed the presence of cracks in the microstructure of the alloys tested (Fig. 14, 15 and 16). The cracks were located near all surfaces coming into contact with the corrosive environment, i.e., both by the fracture surfaces and by the side surfaces. The numbers and the sizes of the cracks were noticeably larger in the case of the specimens tested in the corrosive environment, both under open-circuit conditions and under cathodic polarization. However, the branched crack layouts resembling trees, propagating from the surface deep into the material and characteristic of hydrogen damage in, e.g., steel were not observed in any of the specimens. The reasons behind this are probably the low diffusivity of hydrogen in the crystal lattice of magnesium and the formation of a passive layer on the surface of the metal (Ref 8, 30-34). These two factors reduce the penetration of hydrogen deep into the alloy in such a short time as the duration of an SSRT test. Moreover, it was observed that the cracks did not always propagate from the surface of the specimen. Some cracks were located within the α-Mg solid solution, a small distance from the surface of the specimen or in close vicinity of the tips of other cracks (Fig. 21).

This suggests that cracks present in the microstructure of the WE54 alloy may propagate both by growth in the region directly adjacent to the crack tip, where the peak tensile stresses occur, and by the coalescence of individually propagating cracks located in hydrogen-saturated areas in the alloy. Moreover, in the microstructure of the specimens tested in the corrosive environment, there were no voids, or there were very small voids, located in crack areas (Fig. 22). This indicates that the alloy matrix surrounding them had much worse plastic properties, which may have resulted from the presence of hydrogen in those areas. Another effect of that loss of ductility by the alloys matrix is the presence of cracks in its volume, initiated by the cracking of the intermetallic phase precipitates present in the alloy (Fig. 22). No similar effect was observed for the specimens that were not exposed to a hydrogen-containing environment. Similar effects have been observed for other magnesium alloys with rare earth elements (Ref 35).