Effect of Copper and Nickel Content on the Corrosion Mechanisms in Ferritic Matrix Gray Cast Irons under Simulated Marine Environments

Abstract

This article investigated the influence of copper (Cu) and nickel (Ni) on atmospheric corrosion in gray cast iron under simulated marine conditions. The goal was to compare the effect of Cu and Ni addition in castings with weathering steels. Selected alloys were cast, cut, prepared, and heat-treated for microstructure homogenization. Accelerated corrosion tests were conducted using a salt spray chamber. Corroded samples were analyzed for corrosion thickness and deposits using scanning electron microscopy, X-ray diffraction, and electrochemical techniques. The results indicate that alloying elements significantly affect corrosion processes. In the long-term, Cu had a greater impact on the corrosion mechanisms than Ni. Both Cu and Ni exhibited similar effects on the corrosion mechanisms in gray cast iron and weathering steels. In the initial and final stages, the behavior was comparable to that of weathering steels, but in the intermediate stage, it differed from the literature, suggesting the presence of an additional mechanism between these stages.

1. Introduction

Corrosion is a problem that affects materials, especially metals. Corrosive processes increase maintenance costs, decrease the strength of metallic structures, and harm the quality of life of communities. According to a study by NACE International, corrosion accounts for about 3.4% of the global gross domestic product [1]. Corrosion processes occur in different types of atmospheres, with industrial, rural, and marine atmospheres being the most notable. The marine atmosphere is one of the most corrosive due to high levels of humidity and salinity.

According to the literature, the addition of Cu and Ni to ferrous alloys favors corrosion resistance [2], especially in weathering steels. Weathering steels are a family of steels designed to self-protect against corrosion by generating an adherent and compact layer of corrosion products that slows down the diffusion process of the elements that cause corrosion, delaying the corrosion process. Although this family of steels has been widely studied, the mechanisms governing these processes are not yet fully understood.

Within the family of ferrous metallic materials such as weathering steels are castings. Castings are iron alloys with a high-carbon content, and gray cast iron is a low-cost product. According to a census conducted in 2019, approximately 70% of the cast ferrous alloys in the world are gray castings [3]. For this reason, it is interesting to consider that the addition of Cu and Ni to these alloys may exhibit a similar behavior to that of weathering steels.

The main difference between castings and steels lies in their microstructure. This variation originates from an increase in the carbon and silicon content. In Figure 1, two images are presented: on the right, a gray iron with flake graphite, and on the left, a structural steel. Although both materials have a similar ferrite and pearlite matrix, the casting contains graphite flakes that affect its corrosion behavior. The presence of graphite in castings can alter the corrosion rate in two ways. First, graphite conducts electricity, which accelerates the corrosion process; on the other hand, according to LaQue [4], the orientation of the graphite flakes influences the corrosion process, potentially slowing it down.

The corrosion experienced by gray cast iron differs from the corrosion suffered by steels, which are affected by generalized corrosion. In the case of cast iron, corrosion initiates in areas where graphite is present [5]. Melchers [6,7] compiled bibliographic information on the atmospheric corrosion of cast iron, particularly the behavior of gray cast iron exposed to different environments in Panama such as the Panama Canal, Gutanlake, and Kure shore. In these locations, the material was exposed to the air near the coasts and submerged under different types of water. In his research, it is possible to appreciate that the behavior of gray castings is similar among self-weathering steels, that is, they possess a bimodal behavior [8]. The literature contains numerous studies on the corrosion of gray cast iron submerged in water or buried underground, but few regarding its behavior in atmospheric conditions. These investigations are based on real experimental data or laboratory data [6,7,9], where the gray cast iron is submerged in a solution or found in pipes. All of these studies refer to the presence of a biofilm that can reduce the corrosion rate.

Steels and castings mainly differ by the carbon content present in their alloy, modifying their microstructure. Despite this, the matrix that constitutes them, considering appropriate heat treatments, is the same for both alloys. Therefore, their behavior toward corrosion should be similar, tending to form the same corrosion products.

According to the literature, low- and medium-alloy steels as well as castings show a variation in behavior when exposed to marine atmospheres, observed in corroded thickness versus exposure time curves, where a short- and long-term corrosion mechanism is demarcated, determined by the presence and absence of oxygen, named the aerobic and anaerobic stages, respectively [9]. This behavior was termed bimodal [8]. Given that castings and steels possess a microstructure whose matrix is similar, it is expected that the resulting corrosion products will be similar. During the corrosion process, a layer of iron oxides and oxyhydroxides forms on the surface of the substrate. These species develop based on natural wetting and drying cycles. The basic unit of the oxyhydroxides corresponds to octahedrons of Fe(O,OH)₆, which are structured in various ways, leading to different allotropic forms known as: goethite (α-FeOOH), lepidocrocite (γ-FeOOH), akaganeite (β-FeOOH), feroxyhyte (δ-FeOOH), and regarding oxides, there is hematite (α-Fe₂O₃), maghemite (γ-Fe₂O₃), magnetite (Fe₃O₄), and ferrihydrite (Fe₅HO₈∙4H₂O) [10,11,12,13,14,15,16,17]. Goethite (α-FeOOH) is an oxyhydroxide that forms at low corrosion speeds due to its structure of octahedrons that join, forming compact layers. Unlike other oxyhydroxides like lepidocrocite (γ-FeOOH), the planes of goethite are not only joined by hydrogen bridges but also by octahedrons, which provide greater resistance to corrosion by preventing the permeability of oxygen and water toward the substrate. Goethite (α-FeOOH) is an oxyhydroxide commonly used to protect steel surfaces against corrosion [18]. To achieve a high-quality protective layer, it is important to reduce the grain size of goethite as much as possible.

Several factors influence the formation of the protective layer from goethite (α-FeOOH) including temperature and the presence of other alloying elements. One study revealed that the incorporation of small amounts of nickel (Ni) and copper (Cu) into the steel alloy could significantly increase the density and uniformity of the goethite protective layer [19,20], which in turn improves the corrosion resistance of the material. In this way, the formation of the protective layer can be optimized to effectively protect the steel surface against corrosion in various environments.

The aim of this research was to evaluate the effect of adding Cu and Ni on the corrosion process mechanisms in castings in simulated marine atmospheres. Four alloys with different percentages of copper and nickel were studied, which were heat-treated to obtain a ferrite matrix in all pieces. The alloys were then characterized by optical microscopy and polarization curves. Subsequently, the samples were corroded through accelerated corrosion testing and the corrosion products obtained were analyzed using various characterization techniques such as scanning electron microscopy and X-ray diffraction.

The main conclusion of this work is phenomenological. It is possible to appreciate that there are three mechanisms in the corrosion processes. As occurs in weathering steels, it is estimated that the presence of bacteria plays a fundamental role in the short-term, however, in the medium-term, there is an inflection point in the corroded thickness vs. time curves that requires further studies to determine the mechanism at this stage. It is estimated that the corroded thickness reaches a critical point, causing the outer layer (or part of it) to detach, thereby exposing the graphite once again, which would accelerate the corrosion process.

2. Materials and Methods

2.1. Materials

The experimental procedure began with the selection of the alloys. For this purpose, it was determined to melt four alloys, with different contents of copper and nickel starting from a main alloy. A hypereutectic gray cast iron with flake graphite was melted and ingots of 100 × 120 × 200 mm³ were fabricated using an induction furnace and gray cast iron scrap in a Chilean company called Fundinox Chile S.A (Santiago, Chile). Granulated graphite was used to adjust the carbon content, and ferrosilicon was employed to adjust the silicon (Si) content and promote the development of graphite during solidification. A hypereutectic gray cast iron was chosen based on numerical simulations that determined the carbon content required to achieve a 5% graphite content in the solidified alloy, considering fixed silicon (Si) and phosphorus (P) contents of 3% and 0.01%, respectively [21]. This graphite content is approximately 150% greater than that formed in eutectic alloys. The criterion for increasing the graphite content to 5% is also because it does not significantly increase the pouring temperature and accelerates the carbon diffusion process (by reducing the spacing between flake graphite) for microstructural homogenization, a heat treatment that will be described later in this section. Lower pouring temperatures and reduced times for heat treatments reduce the furnace’s energy consumption.

The pouring temperature was approximately 1335 °C, and the casting was conducted in a sand mold bonded with phenolic resin. To prevent sand adhesion to the pieces, the mold was painted with zircon paint. The same base melt was used, to which electrolytic copper (Cu) and nickel (Ni) were added to achieve the desired combinations. The chemical compositions of the obtained alloys were determined by optical emission spectrometry using the SpectroMaXx (Spectro Analytical Instruments GmbH, Kleve, Germany) model LMX10 equipment and are shown in

Table 1. Nominal chemical composition of the samples ¹.

Alloy	%C	%Si	%Mn	%S	%P	%Cr	%Ni	%Cu	%Fe
0Cu-0Ni	4.22	3.300	0.550	0.010	0.005	0.170	0.030	0.010	Balance
0Cu-2Ni	4.240	3.280	0.670	0.015	0.012	0.171	2.040	0.020
1Cu-0Ni	4.250	3.150	0.660	0.016	0.016	0.188	0.030	1.030
1Cu-2Ni	4.230	3.000	0.610	0.012	0.009	0.158	2.030	1.020

¹ Wt %.

.

2.2. Methods

After the blocks were manufactured, they were cut into rectangular sections and drilled, obtaining samples with an average size of 40 mm × 30 mm × 4 mm. From the cuts made, a total of approximately 500 test coupons were obtained. Each coupon was assigned a unique code to ensure traceability. To ensure that the microstructure of the obtained pieces was homogeneous, regardless of the chemical composition of each alloy, the samples were heated from room temperature up to 1050 °C, held for 3 h, and then gradually cooled over two days to achieve a microstructure close to equilibrium. This process aimed to ensure that all alloys had the same microstructural characteristics, composed of a ferritic matrix and graphite flakes. The heat treatment was carried out in a Mo-Si resistance furnace, capable of vacuum. Then, four samples from each alloy were selected and subjected to metallographic analysis to determine the final microstructure of the process. Metallographic examinations were conducted by optical microscopy (OM) (Olympus, Tokyo, Japan) after the samples had been ground on SiC paper, polished with a polycrystalline diamond suspension, and then etched using Nital’s etchant. Subsequently, an electrochemical characterization of the heat-treated samples was performed using polarization curves with a potentiostat-galvanostat (TEK-Corr, Pasadena, TX, USA) to evaluate the influence of the alloying elements on the corrosion resistance behavior. The configuration indicated in Figure 2 was used.

For the configuration shown in Figure 2, a “mercury calomel” reference electrode (Hg/Hg₂Cl₂), a graphite counter electrode, and the working electrode as the samples to be studied were used. The test parameters were as follows: scan rate of 5 mV/s, sweeping from −900 mV to −400 mV, the exposed area of the working electrode was approximately 1 cm², and the electrolyte used corresponded to a 0.1 M NaCl solution. The oxidation process of Fe occurs according to the reaction Fe → Fe²⁺ + 2e⁻, involving two electrons. In the reduction process, the reaction is reversed. This situation was considered for the calculation of the corrosion rate. Once the experiments were conducted, the value of the anodic and cathodic slopes was determined.

After applying the heat treatments, microstructural characterization, and electrochemical characterization, all the coupons were prepared for the corrosion process in a salt spray chamber. The sanding stage was performed using a disc machine with P80 sandpapers. Subsequently, the samples were washed with water to remove the particles adhered during the sanding. Then, they were dried with alcohol and hot air to remove the remaining water. After completing the sanding stage, the dimensions and weight of the samples were measured with a caliper and an analytical scale, respectively. Three measurements were taken to determine the length, width, and thickness of each sample, and three weight measurements of each sample.

Then, an average was calculated for each of the obtained dimensions. Subsequently, for the accelerated corrosion tests, a salt fog chamber based on the ASTM B117 standard [22] was used. The fog chamber used was an Atlas SF chamber (Atlas American Corporation, Mount Prospect, IL, USA), model SF850. A washing stage with water and a drying oven with forced air were used, where one corrosion cycle equals a full day [23]. The accelerated corrosion cycle starts with a 16-h salt wetting inside the fog chamber, calibrated under the ASTM B117 standard. This was followed by a resting stage of 50 min inside the fog chamber and a subsequent 10-min washing, which was conducted with pressurized water to remove the salt content present in the samples. After this, they were placed in the drying oven using hot air at a temperature of 50 °C for a period of 2 h. Once this time had elapsed, they were put back into the fog chamber for 2 h to be moistened, and rested inside it for 1 h. Finally, they were transferred to the drying oven to be dried for another 2 h. To determine the amount of mass lost during the corrosion process, gravimetric techniques were used. To remove the corroded layer during the accelerated corrosion test, the samples were exposed to a chemical milling process under the ASTM G1 standard [24]. Initially, the samples were placed in a ceramic container, where a previously prepared solution of hydrochloric acid diluted in water with a corrosion inhibitor was poured. Once all the corroded material was removed, the samples were washed with water to eliminate the residual solution and then with alcohol to displace the water from the surface. Finally, they were dried with hot air to volatilize the remaining alcohol. After pickling, the samples were weighed, and the mass difference requiring the calculation of the corroded thicknesses was recorded. To determine the corroded thickness, it was necessary to know the total area of the sample exposed to the test. Each curve of the corroded thickness obtained represents the behavior of each alloy during the accelerated corrosion test. To obtain a point of the curve, a total of three samples were averaged for each studied cycle. The standard deviations were based on three samples per cycle for each alloy.

Some of the pieces obtained in cycles 6 and 25 of the corrosion tests conducted were subjected to phase determination techniques using X-ray diffraction and scanning electron microscopy. These tests aimed to determine the present phases, the relative amounts of these phases, and the morphology of the corrosion products obtained.

To observe the morphology of the corrosion products, a scanning electron microscope (SEM) JEOL JSM-6010LA (JEOL Ltd., Tokyo, Japan) operating at 20 KV was used [25]. For the X-ray diffraction study, a Rigaku diffractometer (Rigaku American Corporation, The Woodlands, TX, USA) model Miniflex 600 was employed with a lamp emitting chromium radiation (Cr Kα = 2.2939 Å) in a scanning range from 15 to 60 degrees (2θ), using a scan speed of 6 degrees per minute.

The objective of the X-ray diffraction measurements was to determine the relative amounts of different oxyhydroxides. However, many of the corrosion products possessed an amorphous structure or were of sufficiently small size to not significantly contribute to the increase in X-ray intensity caused by diffraction. For this reason, a mechanical mixture of the corrosion products with 10% by weight of ZnO (a crystalline species used as an internal reference standard) was made. This species was chosen because it is not among the corrosion products formed in the studied materials.

3. Results

In this section, the results obtained from the initial characterization of the alloys, conducted using various techniques, are discussed. Following this, the corrosion curves, which were acquired through gravimetry and referenced against the ASTM G1 standard, will be analyzed. The results from the characterization of the corrosion products will be reviewed; subsequently, the information gathered will be correlated. The most relevant observations of this work will be presented.

3.1. Heat Treatments and Metallographic Analysis

Given that the samples possessed different chemical compositions and the cooling they were subjected to during manufacturing was the same, it is normal to expect that the microstructure varied among the alloys. This is evidenced in Figure 3. In this figure, the microstructures of the alloys prior to the annealing heat treatment can be observed. In the images, it is possible to appreciate graphite flakes and variable fractions of ferrite and pearlite. The discussion of the effect of Cu and Ni on the ferrite-to-pearlite ratio of the as-cast matrix was beyond the scope of this article. To evaluate the samples in the same microstructural state, a heat treatment was carried out with the objective of homogenizing the microstructure. The applied heat treatment was designed to replicate the equilibrium conditions, allowing carbon to diffuse from the cementite in pearlite toward the graphite. In Figure 4, the evolution of the microstructure can be observed, since the alloys now present a ferritic matrix with graphite flakes.

3.2. Polarization Curves

In this section, the results obtained from the measurements of the polarization curves carried out on four heat-treated alloys with different copper and nickel contents are shown. The measurements were conducted in an electrolyte composed of an aqueous solution of 0.1 M NaCl. An exposed area of 1 cm² was used for the tests. The polarization curves obtained are shown in Figure 5, according to the previously described conditions. From the polarization curves, the Tafel slopes of the anodic and cathodic branches were determined. The intersection of the slopes of the anodic and cathodic branches allowed for the determination of the corrosion current and potential. From the corrosion current data, the corrosion rates in mils per year (MPY) shown in Figure 6 can be calculated. These results will allow us to analyze and compare the corrosive performance of the alloys without being subjected to wetting and drying cycles.

3.3. Determination of Corroded Thickness

After conducting accelerated corrosion tests and employing gravimetric techniques to determine the corroded thickness as stipulated in the ASTM G1 standard, the curves of corroded thickness vs. time were obtained.

In Figure 7, a graph depicting the corroded thickness (µm) versus number of cycles for various alloys is presented. Initially, alloys containing nickel exhibited a lower corroded thickness compared to those containing copper; however, this trend reversed over a longer period. Notably, in the 1Cu-2Ni alloy, the copper influence predominated over nickel. Furthermore, the alloys were shown to corrode less than the base alloy devoid of alloying elements.

Observation revealed that all alloys featured two inflection points. The first occurred around cycle 12, termed “t_A”, aligning with the existing literature. A second inflection point, evident in all alloys, was located near cycle 25, designated as “t_B”. This second inflection point’s existence may suggest an underlying mechanism. Specifically, while grey cast irons may undergo three distinct corrosion mechanisms, weathering steels are subject to only two [26].

3.4. Scanning Electron Microscopy

A series of observations and measurements were made using an electron microscope with an energy-dispersive X-ray spectroscopy detector. It is possible to mention that the analyses were conducted in stages. In the first stage, observations on the morphology of the corrosion products from cycles 6 and 25 were made. Then, each sample underwent compositional chemical analyses through an elemental distribution analysis using EDS. The samples were etched with a diluted solution of HCl (33% HCl and 67% water).

Figure 8, Figure 9, Figure 10, Figure 11, Figure 12, Figure 13, Figure 14 and Figure 15 reveal the distribution of iron, carbon, nickel, and copper in the corroded layers of gray cast iron under different conditions and exposure times. Similarities across the figures include the presence of graphite flakes in all samples and the development of corrosion products over time. Differences were notable in the elemental distribution and corrosion product morphology based on the alloy composition and exposure duration.

The presence of nickel did not significantly affect the corrosion layer, meaning that there was no significant variation in the quantities of goethite (which is the adherent and compact phase). This is reflected in Figure 9 and Figure 13, where the internal layer is heterogeneous. In contrast, the quantities of goethite were enhanced by the presence of copper, as seen in Figure 10 and Figure 14, where the internal phases were much more compact. Although the alloys in cycle 6 should not show a significant difference in goethite, some differences were already noticeable within this testing period.

From Figure 12, Figure 13, Figure 14 and Figure 15, it is possible to appreciate a series of images of the corroded layers obtained from cycle 25 of the alloys under study. In them, the distribution of iron, carbon, nickel, and copper was observed.

From the measurements made, it was possible to estimate the thickness of the corroded layers, that is, the thickness of both the outer and inner layers. Since the outer layer was easily removed during the cycling process and subsequent handling due to sample preparation, only the inner layer was considered. For this purpose, a comparison of information from the alloys, considering only the inner layer, is presented in Figure 16.

3.5. X-ray Diffraction

In Figure 17 and Figure 18, the diffractograms of the powder samples obtained corresponding to cycles 6 and 25, respectively, are presented. It is possible to appreciate that the samples exhibited similarities between them, that is, they presented the same phases. The diffractograms revealed the presence of magnetite, goethite, akaganeite, and lepidocrocite as the main phases. The quantification of these phases, shown in Figure 19 and Figure 20 for cycles 6 and 25, indicates that the phase distribution was influenced by the alloy composition and the number of corrosion cycles. The results show that the quantities of goethite were not influenced by the nickel content, while there was a significant increase in the goethite phase when the alloy contained only copper or a combination of copper and nickel. The phase that provided the greatest resistance to corrosion was goethite, which had higher adherence to the substrate [27].

Based on the diffraction patterns, it was possible to estimate the quantities of the phase of interest, which were the following: magnetite, goethite, akaganeite, and lepidocrocite. The results obtained from cycles 6 and 25 are presented in Figure 19 and Figure 20, respectively. It is important to consider that the difference in the phases obtained corresponded to the amorphous content. The amorphous content of the samples could not be detected by X-ray diffraction as this amorphous material did not possess sufficient crystallinity to be measured by this technique.

Considering that the corrosion product of greatest interest for this study was goethite oxyhydroxide since corrosion resistance is associated with this phase, a comparative graph of cycles 6 and 25 comparing only this phase is presented in Figure 21.

4. Discussion

The alloys under study were differentiated practically by the contents of Cu and Ni. It can be appreciated that the alloy contents, in general, were similar; however, these elements distinguish them. Moreover, the obtained microstructure corresponded to practically ferrite and graphite, which allows for a direct comparison of the results.

According to the corrosion curves shown in Figure 7, it can be inferred that the copper content had a significantly greater impact on the long-term corrosion mechanisms than the nickel content. In the long-term, it was demonstrated that copper favored the presence of goethite, which is consistent with corrosion in weathering steels, while the nickel content did not significantly influence the formation of the phase with higher corrosion resistance.

In the short-term, nickel presented a greater resistance to corrosion than copper. However, in the long-term, there was a reversal in this behavior. The results suggest that nickel only contributes by forming oxyhydroxides like lepidocrocite, which have a negative charge, hence disfavoring the interaction of the chloride ion, thereby delaying the corrosion processes and not affecting the quantities of goethite, a phase that influences corrosion resistance, as occurs in weathering steels [28,29].

Regarding the combined effect of copper and nickel, no synergistic behavior of these alloying elements is observed. It has been determined that the effect of these alloying elements acts independently.

From the obtained corrosion curves, the presence of two inflection points, which were named “t_A” and “t_B”, could be observed. This suggests the existence of an additional mechanism to the two classics described in the literature [2,6,8]. However, what this additional mechanism corresponds to has not yet been determined. It is estimated that the presence of graphite is the main variable in this process. Particularly, the graphite flakes not only accelerate or decrease the corrosion rate, as proposed in LaQue’s theory, but also affect the oxide layer formation processes, resulting in a decrease in the corrosion rate after “t_A” and an increase after “t_B”. Possibly, the layer changes its physical characteristics upon reaching “t_B”, that is, it becomes porous, or fractures expose the graphite again and accelerate the corrosion process.

The polarization curves represent an initial stage of the corrosion processes, serving to compare the results in the short-term. The results indicate that samples containing nickel have a lower corrosion resistance than those containing copper. This is consistent with what was determined in the short-term for these alloys.

From the results obtained by scanning electron microscopy, it was not possible to obtain a correct estimation of the corroded thicknesses. This was attributed to the fact that it is a point measurement, unlike gravimetric techniques. However, by observing the distribution of elements in the analyzed sectors, it can be inferred that there were two oxide layers. These were clearly differentiated given the presence of graphite in one layer (inner) and the lack of this element in the outer layer. Additionally, it was observed that the distribution of elements in the oxide layers was homogeneous, that is, there was no preferential location that can be attributed, through this technique, to the formation of any compound rich in nickel or copper. Therefore, it is estimated that the participation of these elements in the corrosion processes is not related to any oxyhydroxide, but rather, they facilitate the participation of oxyhydroxides such as goethite or lepidocrocite [30].

From the X-ray diffraction tests performed, it is possible to mention that the technique used provides relevant information. The phase of greatest interest (goethite) showed a greater participation in the corrosion processes as the cycles increased. In addition, it was observed that the nickel contents in the alloys slightly increased the goethite contents. On the other hand, the copper contents significantly increased the amount of goethite, which was consistent with what was obtained in the gravimetric curves.

5. Conclusions

• The influence of Cu and Ni on the corrosion mechanisms in gray cast irons impacts the corrosion rates in a different way from that of weathering steels.
• It has been estimated that in the initial stages of corrosion, the alloys exhibit a behavior like that reported in the literature. However, the presence of a second inflection point in the corroded thickness vs. time curves suggests the presence of a third distinct mechanism. This new mechanism has been characterized with the parameter “t_B”.
• The effect of copper and nickel on the marine corrosion process of gray iron with flake graphite has been evaluated. The addition of copper shows an increase in the goethite phase in the long-term, while adding nickel infers a behavior that contributes to the formation of oxyhydroxides like lepidocrocite, which have a negative charge. This disfavors interaction with the chloride ion, hence delaying the corrosion processes like what occurs in weathering steels.
• In samples with a combined content of copper and nickel, it is observed that the effect of the alloying elements is practically independent. That is, in the short-term, there is a decrease in corroded thickness due to nickel’s contribution to the formation of oxyhydroxides like lepidocrocite, while in the long-term, the copper content favors the appearance of the goethite phase, which increases corrosion resistance. Therefore, the combined effect does not show synergy.
• The presence of a third mechanism could indicate that the alloy’s behavior is not the same as that of weathering steels. This behavior is attributed to the graphite phase.
• Regarding the observed corrosion rates, it is possible to say that there is no superior material. This should be evaluated based on the duration of exposure to the atmosphere. If the material is exposed to an aggressive environment for a short time, the best alternative would be a nickel-rich alloy. However, if the samples will be exposed for a long time, then the best option would be an alloy with a high copper content.