15.3.1 General corrosion

General corrosion is the most common form of corrosion and is also called uniform corrosion where the electrochemical reactions (in aqueous or atmospheric media) proceed uniformly over the entire exposed metal surface over a large area. As such, general corrosion represents the greatest destruction of metal on a tonnage basis. As a result, the metallic surface becomes rough and possibly frosted in appearance (Figure 15.1). However, general corrosion is less dangerous than the other corrosion forms because the life of equipment or structures can be accurately estimated from simple corrosion tests and sometimes by visual inspection.

The occurrence of general corrosion in magnesium alloys is quite common. Figure 15.2 shows an example of general corrosion for AZ31D after 1 week of immersion in NaCl solution. Using more corrosion-resistant materials or protective coatings is the most effective approach to control such corrosion.

15.3.2 Pitting corrosion

Pitting corrosion is more destructive than uniform attack is because it is difficult to predict and leads to unexpected failure. The naturally occurring passive films formed over many metallic substrates (such as aluminum, copper, stainless steel, etc.) can provide a considerable corrosion resistance over a long time in corrosive media. However, if corrosion occurs (due to material defects or changes in the surrounding environment), some pits could form. Pitting is most likely to occur in a halide ion containing media (mostly chloride) in the presence of oxygen or oxidizing agents.

Figure 15.3 shows a schematic representation of pitting corrosion. Pitting corrosion can occur at free corrosion potential of magnesium, when exposed to chloride ions in a nonoxidizing medium¹⁸ as a result of the passivity breakdown.^18,19 This is followed by the formation of an electrolytic cell (Figure 15.4) in which the secondary phase of AlMn, AlMnFe, and so on, acts as a cathode and the surrounding Mg matrix acts as an anode. Figure 15.5 is an example of pitting corrosion in a AZ91D magnesium alloy immersed for a week in corrosive 3.5 wt.% NaCl solution.

15.3.3 Crevice (under deposit) corrosion

Crevice corrosion is another destructive form of localized corrosion. It usually occurs in the areas under deposits where free access to the surrounding environment is restricted. Crevice corrosion is caused on contact of metals with metals or metals with nonmetals, for example, gaskets, couplings, and joints. It may occur also at washers, under barnacles, at sand grains, under applied protective films, and at pockets formed by threaded joints.

There is a large debate in the literature about the possibility of occurrence of crevice corrosion in magnesium alloys. Some researchers reported that crevice corrosion could not occur in magnesium alloys.^18,20 Although a form of attack that occurs at narrow gaps (“crevice”) appears similar to crevice corrosion, it is not a true crevice corrosion (Figures 15.6 and 15.7). The justification for this observation is that the corrosion is caused by the retention of moisture in the crevice which, being unable to evaporate, promotes the corrosion of metal in the narrow to recess over extended periods. In fact, this is typically called “filiform corrosion,” which is a peculiar type of crevice corrosion. As such, it is deduced that crevice corrosion could possibly occur in magnesium alloys.

In contrast, other researchers²¹ proposed a mechanism for the possible occurrence of crevice corrosion in magnesium alloys in which crevice corrosion could be initiated due to the hydrolysis reaction when oxygen does not play a major role in the corrosion mechanism. The formation of Mg hydroxide should influence the properties of the interface between the Mg and the solution in the crevice.

Based on the author’s previous studies with different magnesium alloys in chloride-containing solution,^2–17 typical filiform and crevice corrosion have been identified (Figures 15.7 and 15.8). It is believed the occurrence of crevice or filiform corrosion could be initiated under certain conditions. For example, in case of imperfect chemical conversion coating treatments using salts such as cerium, vanadium, zirconium, tin, or manganese to protect magnesium alloys. The crevice or filiform corrosion is caused by active galvanic cells that originate across the metal surface due to the formation of oxide thin-film of the salts of irregular surface morphology. The metal parts covered by oxide act as cathodes, whereas the uncovered Mg parts act as anodes.

15.3.4 Filiform corrosion

As mentioned above, filiform corrosion is a particular form of crevice corrosion. This type of corrosion typically occurs on coated surfaces when moisture or a corrosive solution penetrates through the defective coatings. It is caused by active galvanic cells across the metal surface. Its head is anodic, whereas the tail is cathodic (Figure 15.8). Lacquers and “quick-dry” paints are most susceptible to filiform corrosion. Thus, an ideal coating should exhibit low water vapor transmission characteristics and excellent adhesion. The occurrence of pitting and filiform corrosion for magnesium alloy AZ91 was identified by some researchers.²² Dexter²³ proposed a model of filiform corrosion for magnesium that is driven by oxygen concentration between the head and tail. However, the proposed model is in contradiction with the theory that magnesium corrosion is relatively insensitive to oxygen concentration differences.

To avoid filiform corrosion, a suitable surface preparation or treatment prior to coating should be applied. In addition, a proper inspection of coatings to ensure the absence of coatings defects, such as holidays, or holes, should be used.

15.3.5 Galvanic corrosion

Galvanic corrosion is an electrochemical action of two dissimilar metals in the presence of an electrolyte and an electron conductive path (Figure 15.9). A potential difference usually exists when two dissimilar metals are in contact. In such case, corrosion of less resistant metal is increased while corrosion for more resistant metal is reduced; whereas, more active metal acts as an anode (the one that will corrode) and more resistant metal acts as a cathode. The further apart the metals in the galvanic series are (Figure 15.10), the higher the rate of galvanic corrosion there is. Magnesium alloys have high susceptibility to galvanic corrosion due to excessive levels of heavy metal or flux contamination. Severe galvanic corrosion takes place when magnesium is in contact with less active metals such as Fe, Ni, or Cu (they have low hydrogen overpotential and act as efficient cathodes). On the contrary, metals that combine an active corrosion potential with high hydrogen overpotential such as Al, Zn, Cd, and Sn, are much less damaging, and, therefore, they are favorable as alloying elements for magnesium alloys.

There are several factors affecting the rate of galvanic corrosion: (1) the degree of potential difference between the two metals; (2) the distance effect, with the greatest attack usually being near the junction; (3) the area effect, where it is better to have large anodes and small cathodes; and (4) the corrosiveness of the surrounding media.

The author’s recent research studies on some rare-earth-containing magnesium alloys such as AZ91E, EV31A-T6, and Elektron ZE41 Mg-Zn-rare earth revealed the occurrence of galvanic corrosion between the inert phase associated with rare-earth metals and active magnesium matrix.^6,9–12,15 Figure 15.11 shows the formation for rare-earth phases in Mg matrix and the occurrence of galvanic corrosion at the interface.

15.3.6 Stress corrosion cracking

Stress corrosion cracking (SCC) is an extremely destructive type of corrosion damage in engineering equipment and structures. SCC is caused by the effects of tensile stress in corrosive environments. Stresses may be due to applied loads, residual stresses from the manufacturing process, or a combination of both.

Most of the SCC cases reported for magnesium alloys have focused on extruded or rolled Mg alloys.²⁴ Die-cast magnesium alloys are more susceptible to SCC than rapidly solidified and semisolid cast alloys. Magnesium alloys containing A1 are more susceptible to SCC in air, distilled water, and chloride-containing solutions. Two mechanisms have been proposed for SCC in Mg alloys²⁵:

1. Continuous crack propagation by anodic dissolution at the crack tip (commonly called the dissolution model), where it includes a preferential attack model, film rupture model, tunneling theory, and so on

2. Discontinuous crack propagation by a series of mechanical fractures at the crack tip (commonly called the brittle fracture models), where it involves cleavage processes and hydrogen embrittlement theory

According to the fracture morphology of the crack, SCC can be divided into two types:

1. Transgranular SCC: This is the common form in magnesium alloys where it, mostly, has HCP crystal structures that are susceptible to cleavage due to the less slip systems available. In general, the occurrence of SCC in magnesium alloys is always linked with hydrogen evolution. Such observation has been evidenced by other authors.^24,26

2. Intergranular SCC: This is a minor form of SCC in magnesium and is mainly related to localized galvanic attack of the matrix when coupled with cathodic grain-boundary precipitates.²⁶

15.3.7 Intergranular corrosion

Intergranular corrosion is an attack that occurs at the grain boundaries of a metal or alloy due to the precipitation of secondary phase. A highly magnified cross-section can show the granular structure with a clearly defined boundary that chemically differs from the grain.

There is a large debate between the researchers whether magnesium alloys suffer intergranular corrosion or not. It has been taken for granted that true intergranular corrosion could not occur in Mg alloys because the phases at the grain boundaries are almost cathodic to the grains.²⁷ Corrosion tends to be concentrated in the area adjacent to the grain boundary until eventually the grain may be undercut and fall out.^18,20 However, some researchers recently showed that intergranular corrosion can occur on some magnesium alloys such as WE43.²⁸ In another work, researchers claimed that in the early stages of immersion, a localized attack of AE81 can be formed at the grain boundaries,²¹ which can be considered intergranular (intercrystalline) corrosion. The reason for such a tendency is still unclear. Generally, magnesium alloys containing low Al concentration corrode at a faster rate than those containing Al-rich regions.

15.3.8 Corrosion fatigue

Corrosion fatigue is a special kind of SCC caused by a synergistic effect between cyclic stress and corrosion. Corrosion fatigue represents the most destructive attacks that can occur on metallic structures without exception. Damage from corrosion fatigue is greater than the sum of the damage from both cyclic stresses and corrosion.

Magnesium alloys AZ91E and AZ91D specimens were supplied by Magnesium Elektron, UK. The specimens used are in the form of 60 × 30 × 3 mm, were abraded to an 800 finish, degreased, washed with distilled water, and dried for 5 min in hot air.

The stannate coating used in this study was prepared from potassium stannate, K₂SnO₃·3H₂O + 10 g/l NaOH for 30 min at pH 12.9. The coated samples were tested in 3.5 wt.% NaCl solution. The solutions were prepared from laboratory grade chemicals and nanopure distilled water. A series of samples were prepared under different conditions prior to applying stannate coating treatment. The samples were subjected to different kinds of surface modifications. Each condition is specified under its related figure.

Several electrochemical techniques such as hydrogen evolution and potentiodynamic and salt spray chamber tests, have been used for evaluating the performance of the stannate coatings over magnesium substrate. Hydrogen evolution method (Figure 15.12) was used to determine the hydrogen evolution rate of stannate-coated AZ91D and AZ91E alloys in NaCl. The calculation of hydrogen evolved due to the reaction of a metal of high corrosion susceptibility such as magnesium immersed in a corrosive solution (such as chloride) is proven to be an accurate technique for evaluating the corrosion resistance of Mg and its alloys. The reaction can be expressed as follows:

$\text{Mg}+2{\text{H}}_2\text{O}\to \text{Mg}{\left(\text{OH}\right)}_2+{\text{H}}_2$

The basic principle of hydrogen evolution technique can be understood from the electrochemical Equation (15.1), where the dissolution of one Mg atom generates one H₂ gas molecule. In other words, the evolution of 1 mole of H₂ gas corresponds to the dissolution of 1 mole of Mg. Therefore, measuring the volume of H₂ evolved is equivalent to measuring the weight loss of Mg dissolved, and the measured H₂ evolution rate is equal to the weight-loss rate. Interestingly, the anodic reaction is the dissolution of Mg and the cathodic reaction is the H₂ evolution, and there is no contribution of oxygen reduction to the cathodic process which is one of the main advantages for this technique.

The evolved hydrogen gas due to the chemical reaction of AZ91D and AZ91E magnesium alloys specimens with the electrolyte 3.5 wt.% NaCl solution was collected by a funnel just above the specimen and then went into a burette and gradually displaced the test solution in the burette. In this way, the kinetics of the evolved hydrogen gas can be determined by reading the height of the test solution level in the burette.

The durability of the stannate coating systems was measured according to ASTM B117 test²⁹ in which the corrosion resistance of coated and uncoated materials can be evaluated through exposure to a continuous indirect spray of a neutral (pH 6.5-7.2) salt water solution at an elevated temperature.

Cyclic voltammetry measurements of specimens previously immersed for 7 days in 3.5 wt.% NaCl solution were accomplished at a scan rate of 0.07 mV/s using an Autolab PGSTAT 30 galvanostat/potentiostat, Metrohm.

SEM and EDS were used to examine the microstructure of the coated samples before and after the immersion in NaCl solution. SEM images of the samples that were immersed in 3.5 wt.% NaCl for 7 days, washed with nanopure deionized water, and then dried, were obtained using a digital scanning electron microscope Model JEOL JSM 5410, Oxford Instruments, Japan. Microprobe analysis was performed using energy dispersive spectrometry, EDS, Model 6587, Pentafet Link, Oxford microanalysis group, UK.

Corrosion morphology was examined with a metallographic microscope (LEICA DMR) with a Quips programming window, LEICA Imaging Systems Ltd., Cambridge, UK, to investigate the types of corrosion produced on the substrate surfaces before and after immersion in 3.5 wt.% NaCl solution.

15.4.2 The performance of stannate-based coatings

Surface examination of as-abraded and stannate-coated samples was performed using optical microscope, visual inspection, and SEM-EDS (Figures 15.13–15.17). Generally, the stannate-coated samples showed better resistance to localized corrosion (pitting, crevice, filiform, etc.) after 7 days of immersion in NaCl solution compared to the as-abraded sample that showed severe corrosion. Interestingly, stannate-coated samples, under particular conditions, showed self-healing characteristics (see Figures 15.15 and 15.16). The pitting density of as-abraded AZ91E samples was 12 pit/cm² and decreased to be 5 pit/cm² for the samples that were directly treated with stannate without any surface modification. This observation points to the important role of stannate coatings over magnesium substrate that acts as a barrier for diffusion of aggressive ions and oxygen, besides their ability to reject the chloride ions from the surface by forming a uniformly distributed magnesium (hydro)oxide layer enriched with tin oxide (Figures 15.13 and 15.16).

Table 15.1 shows the average amount of hydrogen evolution collected per hour for the as-abraded and stannate-coated AZ91E and AZ91D alloys over 270 h of immersion in a 3.5 wt.% NaCl solution. It was evidenced that the rate of hydrogen evolution increased as a function of time. The amount of hydrogen gas collected from the as-abraded AZ91E and AZ91D samples after 270 h of immersion in NaCl is comparable to each other. The measured amount of hydrogen gas collected after stannate coatings was less than half of those measured for as-abraded samples. AZ91D showed a better corrosion resistance than AZ91E. The results indicate that stannate conversion coatings play a crucial role in the corrosion protection process of Mg substrate by forming a tin oxide-rich magnesium hydroxide layer that acts as a barrier for oxygen diffusion in metal surface, and hence, improves the corrosion resistance.^8,14–17

The corrosion rates of as-abraded and stannate-coated AZ91E and AZ91D were measured by a salt spray chamber test according to standard ASTM B117.²⁹ The corrosion rate of as-abraded AZ91E was measured to be 296 μm/year, which is higher than that measured for AZ91D (254 μm/year) and confirms that the corrosion resistance of AZ91D is higher than AZ91E, as shown in Table 15.2. It was confirmed that stannate coating improves the corrosion resistances of both alloys. The measured corrosion rates of stannate-coated AZ91E and AZ91D samples (directly coated with stannate without any surface modification) are 184 and 160 μm/year, respectively. Such finding confirms the positive effect of stannate coatings in improving the corrosion protection of magnesium alloys.^8,14–17

Localized corrosion is the most dangerous corrosion phenomenon in degradation of materials intended for long-term use in industry. Pitting is of particular concern because it may lead to premature breaching of the materials through electrochemical dissolution processes that can accelerate with time. Cyclic voltammetry technique was used to compare between the pitting corrosion resistance of as-abraded AZ91E and AZ91D after 7 days of immersion in a 3.5 wt.% NaCl solution. Cyclic voltammetry (Figure 15.18) showed a shift of about 200 μA in the cathodic current in the passive direction for AZ91E samples compared with the AZ91D specimens. However, the corrosion potential, E_Corr, and the pitting potential, E_pit, of the AZ91D samples were 80 mV nobler than AZ91E samples. These results are in agreement with the data obtained from the hydrogen gas collection, salt spray chamber test, visual inspection, and macro- and microscopic examinations.

15.4.3 Self-healing functionality of stannate coatings

Self-healing action can be defined as the ability of the coatings to repair any possible surface defects/damages that can occur at the coated substrates due to mechanical wear or chemical attack.¹⁴ The healing process usually takes place through a chemical reaction between the coating materials and the surrounding environments to form chemical products (mainly oxides) that block the defected/damaged areas from further attack. The author’s previous work on stannate coatings^8,14–17 showed that corrosion resistances of stannate-coated magnesium samples increase with increasing the immersion time in corrosive solution. This finding contradicts the known fact that corrosion resistance of a metal or alloy is decreased when immersed in a corrosive solution over a range of time. Such behavior was attributed to the conversion of the tin-oxide film formation from a less protective form (II) to a more stable tin-oxide form (VI).

The chemistry behind the self-healing functionality due to stannate coatings lies in changing the oxidation states of tin from (II) to (VI) oxide. According to the periodic table, tin, as a member of Group 4 elements, has increasing tendency to form compounds of a + 2 oxidation state. Tin(IV) is a more stable oxidation state, which means that tin(II) compounds have a strong tendency to convert into tin(IV). This explains that Sn^2 + ions in solution are good reducing agents.

Surprisingly, the author’s previous work on stannate coatings for magnesium alloys showed that the self-healing functionality is linked to the type of magnesium alloys. In other words, the chemical composition of the magnesium alloy seems to have a direct influence on the ability of stannate coatings to practice the self-healing capability. For example, stannate coatings over AZ91D showed a perfect self-healing functionality.^14,16,17. Conversely, stannate coatings lose their self-healing characteristics over rare-earth containing magnesium alloy such as EV31A-T6.¹⁵ The reason for such an unexpected behavior is still not clear. However, it appears that the presence of certain alloying elements (such as rare-earth elements) might have an adverse catalytic effect on the conversion of tin(II) compounds to tin(IV), which is essential for self-healing ability.