Quasi-In Situ Localized Corrosion of an Additively Manufactured FeCo Alloy in 5 Wt Pct NaCl Solution

Equiatomic FeCo alloys are important materials with potential applications in electrical motors. They have low coercivity in combination with high permeability, high Curie temperature and high saturation magnetisation. During the cooling of a FeCo alloy, the A2 (body-centred cubic) structure changes to an ordered B2 structure below 730 °C.^[1,2] The ordered B2 phase is formed by the rearrangement of atoms on the lattice sites with a 0.2 pct volume change, but no change in the chemical composition. Due to the formation of the ordered B2 phase, FeCo alloys are brittle. Thus, the conventional processing (casting, rolling) of FeCo alloys is challenging. Alternatively, FeCo alloys can be processed by melt spinning or chemical vapour deposition methods. However, these processing methods cannot produce machine components in bulk form. Additive manufacturing methods such as laser powder bed fusion (LPBF) and laser engineering net shaping (LENS) can be applied to manufacture FeCo alloy parts in their final shape due to their high cooling rate (> 10³ K/s).^[3‐6] The high cooling rates suppress the formation of the ordered phase or reduce the order parameter of the ordered phase^[7] due to less time available for the rearrangement of atoms leading to fine grain sizes. Both of these factors contribute positively to an increase in the ductility of FeCo alloys. In previous studies, FeCo alloy powder and FeCo1.5 V alloy powder have successfully been consolidated to densities > 99 pct using additive manufacturing techniques.^[4,6] Moreover, the additively manufactured FeCo samples showed a high yield strength of 600 MPa, ultimate tensile strength of 700 MPa and high ductility of 35 pct with a weak crystallographic texture.^[6]

Most of the previous studies^[4‐6,8] on additively manufactured FeCo or FeCoV alloys have concentrated on the mechanical and magnetic properties. In previous studies on the corrosion of multicomponent FeCo-based alloys (HITPERM, Fe44Co44Zr7Cu1B4 alloy) in 0.1 M H₂SO₄ solution, a reduction in the magnetic saturation value is reported.^[9,10] Upon corrosion, the formation of nonmagnetic corrosion product layers changes the direction of the magnetic lines inside the material which leads to the loss of saturation magnetisation. Thus, it is important to study the corrosion behaviour of multicomponent FeCo-based alloy as corrosion leads to a reduction in the magnetic performance of electrical machine components.^[9,11] The higher the corrosion rate of the alloy, the thicker the corrosion product layer consisting of Fe₂O₃ and the higher the loss of magnetic saturation. In addition, the material loss due to corrosion will lead to premature failure of electrical machine components.

Induction and synchronous motors have a targeted use, for example, in the offshore oil and gas drilling industry such as in the electrical submersible pumps.^[11,12] The ferromagnetic components in electrical submersible pumps are exposed to corrosive downhole well fluids containing dissolved chloride ions, carbon dioxide and hydrogen sulphide.^[12] The performance of the different components (housing, shafts, pump stages, heads and bases) in the electrical submersible pumps is degraded upon exposure to the downhole well fluids. These components are also expected to be subjected to marine environments, which means a high NaCl concentration. It is well-known that chloride ions in saline solutions break down protective films and form pits which can lead to catastrophic failure. In the addressed application, a combined attack of corrosion and mechanical loads leading to stress corrosion cracking or corrosion fatigue is expected. As a FeCo alloy consists of two electrochemically different elements, the occurrence of galvanic corrosion is also anticipated. In literature, this is referred to as deironification, where a selective dissolution of Fe from the FeCo alloy occurs as Fe is electrochemically more active compared to Co.^[13] Thus, investigating the localised corrosion of FeCo alloys in NaCl solution becomes important. Moreover, in the current literature, there are no studies on the pitting corrosion, stress corrosion cracking or fatigue corrosion for FeCo alloys available to the best of the author’s knowledge.

In the current investigation, the localised corrosion behaviour of additively manufactured FeCo alloys by a quasi-in situ method was undertaken. Here, a FeCo alloy (containing 50 wt pct Co) with a U-shaped notch is immersed in an aqueous NaCl solution. The U-shaped notch is used to track the same region for pre- and post-exposure microstructural investigation. Concurrently, the surface dissolution near the U-shaped notch is monitored by scanning electron microscopy (SEM) and electron backscattering diffraction (EBSD) after pre-defined periods. In addition, open circuit potential (OCP) and potentiodynamic polarization measurements are performed to determine the corrosion potential and current density. The topography after surface dissolution is compared to the grain orientation and the deviation of the plane normal to the \(\langle {1}00\rangle\) direction and the \(\langle {11}0\rangle\) direction. In previous literature, corrosion studies are performed on non-equiatomic compositions of FeCo alloy. However, in the present investigation, the corrosion behaviour is studied for an equiatomic FeCo alloy. Thus, this study is conducted to understand the effect of the alloy composition on the corrosion behaviour. Similarly, in the previous scientific studies, the corrosion behaviour is studied in acidic solutions whereas, in the present study, the corrosion behaviour is studied in neutral solutions. Thus, the effect of the corrosion medium on the corrosion behaviour of FeCo alloy is compared.

The OCP under equilibrium conditions is displayed in Figure 1 as it is needed to define the scanning interval for the potentiodynamic polarisation measurement. The potential is observed to drop with increasing measurement time due to the achievement of equilibrium. From Figure 1 the potential at 0 s is − 0.323 V referred to the SHE. However, the potential drifts with increasing time to − 0.37 V vs. SHE.

The potentiodynamic polarization curve of a FeCo alloy in a 5 wt pct NaCl solution is presented in Figure 2. From the potentiodynamic polarization curve, various corrosion parameters are extracted which are reported in Table I. The corrosion potential (E_corr) referred to SHE is − 0.475 V and the corrosion current density (I_corr) is 0.0848 A/m². Based on the Tafel extrapolation method (shown by the red dotted lines, Figure 2), the anodic and cathodic Tafel slopes are obtained as 172 mV/decade and 944 mV/decade, respectively. I_corr is calculated using the equation correlating the current (i) to the potential (E) aswhere i_corr represents the exchange current density, E_corr is the corrosion potential, F is the Faraday constant, T is the temperature, α is the transfer coefficient (usually 0.5), and n is the number of electrons transferred.^[17] A detailed review of the Tafel extrapolation method is presented in the literature.^[17]

The area near the U-shaped notch before the onset of the immersion test is presented in the SEM image in Figure 3(a). The polished sample surface shows some pores. After 24 hours immersion, the same area is observed in Figure 3(b) (SEM image). Figure 3(c) is the zoomed-in image of Figure 3(b), Figure 3(d) is the magnified image of the black rectangular area in Figure 3(c). The sample surface topography revealed the grain structure (Figure 3(d)) after 24 hours immersion. This is due to the corrosion of the FeCo alloy at the surface (i.e., selective conversion of Fe to Fe²⁺ ions in the aqueous solution). A previous investigation proposes the selective dissolution of Fe to Fe²⁺ ions (referred to as deironification) as Fe is electrochemically more active than Co.^[13] This study also observes a reduction in the Fe content (≈ 50 to 47 wt pct) compared to the Co content. The composition of the FeCo alloy in the present investigation is similar to the previous study,^[13] but due to different processing methods, the corrosion behaviour is different to the reported literature. The EDS spectrum from Figure 3(e) presents the entire range, however, the Figure 3(e) inset image displays the presence of an oxygen peak kα1, which indicates the adsorption of oxygen on the surface. The observed reduction in the Fe content (≈ 50 to 47 wt pct) is following the literature.^[13] The presence of oxygen in the EDS spectrum is related to the formation of an oxide phase on the surface. Oxygen adsorption is due to the reduction of oxygen at the surface i.e., the conversion of dissolved oxygen to hydroxyl ions.^[18] No adsorption of chloride ions is detected on the sample surface.

The SEM image of the same area (Figure 3(c)) after the 72 hours immersion test is portrayed in Figure 4(a). The zoomed-in SEM image of Figure 4(a) is shown in Figure 4(b), representing the surface topography of the grain structure. Figure 4(b) is rotated by 180 deg compared to Figure 4(a). The EBSD map from the black rectangular area (Figure 4(a)) with an average confidence index of 0.82 is presented in Figure 4(c). The EBSD map is plotted along the normal direction (ND). Most of the grains are elongated in shape, here grains 1, 2, 3, 4, 5, 6, 7 and 8 are marked in Figure 4(c) which are analysed further on. The pit is highlighted (circle and arrow) both, in Figures 4(b) and 4(c), which is used to correlate the SEM and EBSD images after 72 hours immersion.

The SEM image of the same area (Figure 3(c)) after 168 hours immersion is presented in Figure 5(a). The zoomed-in SEM image of Figure 5(a) is shown in Figure 5(b) which shows the surface topography due to surface dissolution. Figure 5(b) is rotated by 180 deg compared to Figure 5(a). Figure 5(b) represents the same area as Figure 4(b). The EBSD map from the black rectangular area (Figure 5(a)) with an average confidence index of 0.55 is illustrated in Figure 5(c). The low confidence index value is due to local surface dissolution upon immersion. Samples after exposure to the corrosive solution are expected to show a low confidence index. However, due to the quasi-in situ nature of the experiments, this is an acceptable confidence index value for the comparison of crystal orientation in the same region. The EBSD map is plotted along with ND. White areas show unindexed pixels due to the formation of a rough surface as a result of the surface dissolution. As expected, no change in grain morphology is corroborated by no modification in grain shape after 168 hours immersion. The pit is highlighted (circle and arrow) both, in Figures 5(b) and (c), which is used to correlate the SEM and EBSD images after 168 hours immersion.

The misorientation angle distribution of the FeCo sample after 72 and 168 hours immersion time is illustrated in Figure 6(a). The presence of a high fraction of low-angle grain boundaries (LAGBs) to high-angle grain boundaries (HAGBs) with bimodal distribution is visible (Figure 6(a)). With increasing immersion time, the fraction of LAGBs and HAGBs changes, however, the change is not significant. Between 45 and 55 deg, the modification of the HAGBs is not pronounced. This is probably due to the degradation of the image quality with increasing immersion time. In general, the HAGBs are expected to be attacked more due to their relatively higher energy which may lead to a change in the LAGBs and HAGBs fraction.^[19] Grain boundaries have a disordered structure relative to the grain interior which results in higher surface energy resulting in a strong pitting corrosion attack.

The orientations of all the grains within the EBSD map after 72 hours immersion test are plotted on an IPF as displayed in Figure 6(b). The formation of \(\langle 0{11}\rangle\)||ND orientation with strong intensity and [\(\overline{1 }\)14]||ND orientations with weaker intensity are detected (Figure 6(b)). The orientations of all grains of the EBSD map after 168 hours immersion time are plotted on an IPF (Figure 6(c)). A slight increase in the intensity of \(\langle 0{11}\rangle\)||ND orientations along with [\(\overline{2 }\) 27]||ND orientations with weak intensity are noticed. A significant weight loss per surface area is observed after 24 hours immersion (Figure 6(d)). However, with increasing immersion time from 24 to 168 hours, only a minor change in weight is detected. This indicates the formation of a corrosion-inhibiting layer containing oxygen (detected by EDS, Figure 3(e)).

The SEM microstructures after 24, 72, and 168 hours immersion times are depicted in Figure 7. Figures 7(a) and (b) (magnified image of Figure 7(a)) represent the formation of pits after 24 hours. Rectangular pits (red arrows) of an average size of 0.15 ± 0.02 μm (length) and 0.09 ± 0.02 μm (width) are formed on the sample surface (Figure 7(b)). Here, 63 pits within an area of dimension of 21 μm × 14 μm are analysed for calculating the pit dimensions. In addition, the formation of pits after 72 hours immersion time is illustrated in Figures 7(c) and (d) (zoomed-in view of Figure 7(c)). Here, also the formation of both rectangular pits (red arrows) and circular pits (blue arrows) is detected. The length of the rectangular pits is increased by 60 pct (0.24 ± 0.07 μm) and the width is increased by 66 pct (0.15 ± 0.04 μm). The diameter of the circular pits is 0.06 ± 0.02 μm. The formation of pits after 168 hours immersion time is demonstrated in Figures 7(e) and (f) (magnified image of Figure 7(e)). Here, the size of the rectangular pits (red arrows) and circular pits (blue arrows) are observed to grow. The length of the rectangular pits increased by 8.3 pct (0.26 ± 0.05 μm) and the width of the pit increased by 20 pct (0.18 ± 0.04 μm). The diameter of the pits increased by 16.6 pct (0.07 ± 0.02 μm). The change in dimensions of the rectangular and square pits with increasing immersion time is represented in Figure 7(g). A sharp increase in length and width of the rectangular pits is seen up to 72 hours immersion time. Afterwards, the increase in length and width becomes slow. The circular pits are observed for immersion times of 72 and 168 hours, and a slow increase in the circular pit sizes is noticed. Both circular and rectangular pits are created due to the breakdown of the passive protective film by the attack of chloride ions.^[20] In this study, the presence of rectangular and circular pits is observed inside the grains. The evolution of the area density of pits (number of pits per unit area) with increasing immersion time is displayed (Figure 7(h)). A sharp increase in the pit density is observed after 24 hours immersion time. However, by increasing immersion time to 72 and 168 hours, the pit density remains nearly constant. A three-stage pit density growth is reported^[20] (i) nucleation of pits, (ii) rapid propagation of pits and (iii) saturation of pit density. Comparing the present results, the 24 hours immersion test is attributed to stage (ii) showing a rapid propagation of pits, whereas the 72 and 168 hours immersion tests refer to stage (iii) behaviour, where a saturation of the pit density is detected. A sigmoidal behaviour in the growth of the pit density is fitted (red curve, Figure 7(h)) which matches a previous investigation on pitting corrosion of API P110 steel.^[20] However, the equation of the sigmoidal curve in the present investigation (N = 0.4t²/65792) is different to that described in the previous investigation (N = 923(1 − exp(−0.043t²))^[20]), where N is the pit density and t is the time. This is due to the difference in the chemical composition of the samples and different corrosive solutions.

Chansena et al.^[13] have performed the potentiodynamic polarisation measurements of electrodeposited FeCo alloys with varying compositions (e.g., Fe₂₀Co₈₀, Fe₄₀Co₆₀, Fe₆₀Co₄₀, and Fe80Co20) in aerated deionised water and aerated acidic solution. In aerated deionised water, E_corr is − 0.482 V vs. SHE for Fe₄₀Co₆₀ and − 0.446 V vs. SHE for Fe₆₀Co₄₀.^[13] From the current investigation, in 5 wt pct NaCl solution, E_corr is − 0.475 V vs. SHE for Fe₅₀Co₅₀. So, Fe₅₀Co₅₀ alloy (in 5 wt pct NaCl solution) is electrochemically less active than the Fe₄₀Co₆₀ alloy. However, Fe₅₀Co₅₀ is electrochemically more active than Fe₆₀Co₄₀ (in aerated deionised water). The OCP of the investigated alloy is between the OCP of the alloys presented by Chansena et al.^[13] Thus, the measured OCP is following this study. The corrosion behaviour and OCP depend not only on the alloy composition but on the test solution, as well. The corrosion rate of Fe₅₀Co₅₀ in 5 wt pct NaCl solution is 0.183 mm/year (calculated from I_corr). For comparison in aerated deionised water, the corrosion rate of Fe₄₀Co₆₀ is 0.0048 mm/year, whereas, for Fe₆₀Co₄₀, the corrosion rate is 0.0051 mm/year.^[13] Thus, the addition of 5 wt pct NaCl leads to a huge increase in the corrosion rate for iron-cobalt alloys.

In an aerated acidic solution (pH 3) the corrosion rate increases to 2.4 mm/year (2.11 A/m²) for Fe₄₀Co₆₀ and 3.1 mm/year (2.76 A/m²) for Fe₆₀Co₄₀.^[13] Thus, the aerated acidic solution makes the FeCo alloys electrochemically more active, leading to a high corrosion rate. Another important observation is, that the corrosion rate increases by increasing iron content in FeCo alloys [i.e., in aerated acidic solution (pH 3) the corrosion rate of Fe₈₀Co₂₀ (7.7 mm/year, 6.67 A/m²) is higher than for Fe₂₀Co₈₀ (1.1 mm/year)].^[13] As Fe (OCP =  − 0.447 V vs. SHE) is electrochemically more active than Co (OCP =  − 0.277 V vs. SHE), an increase in the iron content makes the alloy electrochemically more active leading to a higher corrosion rate. For the current investigation, a neutral pH solution is selected leading to a corrosion rate (0.0848 A/m²) significantly lower than an aerated acidic solution. Under this investigation, the corrosion behaviour of FeCo alloys is studied in general. So, a neutral pH solution is selected to study the corrosion behaviour of the FeCo alloy.

Based on the quasi-in situ immersion test of the additively manufactured FeCo sample up to 168 hours, the following conclusions can be drawn: