Effects of Ni and Co on the Corrosion Resistance of Al-Si-Cu-Zn-Fe Alloys in NaCl Solution

: The corrosion behavior of Fe-containing directionally solidiﬁed (DS) and centrifugally cast (CC) Al-Si-Cu-Zn alloys with either Co or Ni additions has been investigated. Electrochemical and immersion corrosion methods were used to investigate the corrosion behavior in 0.6 M NaCl after short (1-h) and long (30-day) exposure periods. The employed solidiﬁcation methods allowed the production of samples with a wide range of secondary dendrite arm spacing (SDAS) while preserving Si and Fe-containing phases. The 0.5 wt.% Ni and Co additions led to the growth of the AlFeSi(Ni) and AlFeSi(Co) phases, but no binary AlNi nor AlCo intermetallic particles have been generated. Potentiodynamic polarization studies at early exposure revealed an increase in the corrosion potential as the Ni was added for either fast or slow solidiﬁed samples. The electrochemical impedance spectroscopy at early exposure demonstrated that the Ni-modiﬁed alloy, on the other hand, was associated with smaller charge transfer resistances, indicating a reduction in the corrosion resistance after a short elapsed time into the electrolyte. However, the 30-day immersion tests revealed much lower corrosion rate of the Ni-modiﬁed alloy than the other alloys, while the corrosion rates of the Co-modiﬁed and non-modiﬁed alloys were similar. In the Ni-containing alloy, a decreased corrosion rate under a long-term corrosion process was attributed to the formation of a thick and dense alumina layer, effectively protecting the surface under such conditions. This work contributes to better knowledge of the corrosion behavior of Ni- and Co-corrected Al industrial scrap compositions. different Directional solidiﬁcation (DS) and centrifugal casting (CC). The to gain insight into the inﬂuence of chemical composition (Ni, Co additions), and dendritic microstructure scale on the corrosion behavior of both as-cast specimen conditions for samples exposed to the 0.6 M NaCl solution. Weight loss immersion tests and electrochemical methods were utilized to assess information on the corrosion behavior at early (1-h) and long (30-day) exposures to 0.6 M NaCl solution, in addition to microstructure characterization for determining SDAS and intermetallics characterizing the


Introduction
Addition of alloying elements is an effective manner to improve and control properties of Al alloys. If maintained in a solid solution, some alloying elements increase the corrosion resistance [1][2][3]. This is because the precipitates may act as cathodes for the preferential corrosion of the α-Al matrix [4].
In Al alloys, impurities such as Fe and Si establish cathodic phases regarding the α-Al matrix, whose shape, size, and distribution affect and harm the corrosion properties of the α-Al phase [5,6]. Local loss in passivity is reported to occur in the vicinity of the second phases, resulting in localized corrosion along the interfaces. It has been reported that the structural characteristics of the film and the severity of the corrosion attack may be influenced by (i) the chemical composition of the exposed alloys, (ii) the presence and distribution of micro-defects, macro-defects, and second phases, and (iii) the electrolyte composition [5][6][7].
Al and its alloys exposed to aggressive environments, especially those containing chloride ions (Cl -), are sensitive to pitting corrosion [8]. The effects of adding Ni to Al-Si alloys immersed in the NaCl solution have been evaluated in some previous studies [9][10][11][12][13]. Hypereutectic Al-Si-Fe-Cu-Zn alloys containing either Ni or Ni/Cr were evaluated and

Materials and Methods
Three alloys were considered: (i) Al-7%Si-0.6%Fe-0.35%Cu-0.25%Zn, (ii) Al-7%Si-0.6%Fe-0.35%Cu-0.25%Zn-0.5%Co, and (iii) Al-7%Si-0.6%Fe-0.35%Cu-0.25%Zn-0.5%Ni. They were manufactured from commercial pure elements (>99.7%), and characterized and tested in the "as-cast" conditions. The Co and Ni levels were chosen based on the suitable contents for the correction of Fe-contaminated Al alloys, as stated in the literature [15,18,19]. Two different processing methods were employed to induce distinct dendritic microstructure refinements: Directional solidification (slow) and copper mold centrifugal casting (fast solidification). Therefore, SDAS was adopted as a control parameter so that the size of the dendritic network could be translated.
Commercially pure Al was induction melted and maintained for 20 min at approximately 800 • C in a SiC crucible prior to the insertion of Si, Fe, Cu, Co and Ni elements to the molten bath. To favor alloy homogeneity, the molten alloy was kept in the induction furnace for at least 40 min. Zn was the last element to be added given its low melting point. All alloys were Ar-degassed (before pouring) for two minutes to minimize porosity. The adopted temperature and time were necessary to permit high melting temperature elements such as Si, Fe, Cu, Co, and Ni to be diluted in the molten bath.
The base alloy was the Al-7%Si-0.6%Fe-0.35%Cu-0.25%Zn. This composition was established guided by some typical scrap compositions (in wt.%) having small impurities such as iron (0.6% Fe), zinc (0.25% Zn), and copper (0.35% Cu), as stated by Gaustad et al. [20,21]. The alloy compositions were confirmed using cooling curves and Thermo-Calc thermodynamic calculations after the alloys were produced. Both procedures were used, and the findings (liquidus temperatures) were compared to determine the alloy composition confidence. There was enough agreement in the results.
Liquidus temperature is critical information for controlling alloy composition or parameterizing all trials with the same amount of superheating (5% above T L ) to minimize further influences on final microstructural comparisons.
The molten alloy was poured into a stainless-steel cylindrical mold coupled to the directional solidification furnace (Manufacturer: Fortelab Indústria de Fornos Elétricos Ltd.a-ME, São Carlos, Brazil). The internal surface of the mold was previously coated with alumina, and the water-cooled mold's base was covered with a sheet made of SAE 1020 steel. Molten temperature was controlled through radial electrical wire. The electrical cable was unplugged after the nearest thermocouple from the cooled base indicated 5% above the alloy liquidus temperature. Thermal analysis was previously used to establish the liquidus temperatures. Finally, the water-cooling system began to extract heat in a single direction. This is due to heat extraction from the water cooling system at the mold's bottom base, which promotes upward directional solidification [16,22].
A portion of the directionally solidified (DS) castings was utilized to make centrifugally cast (CC) samples. These samples were produced by centrifugal copper-mold using a Linn High Therm induction furnace, model Titancast 700 VAC (LINN High Therm GmbH, Hirschbach, Germany). Three distinct plate mold thicknesses were used in the chamber to make pieces attaining 2 mm, 3 mm, and 4 mm. The microstructures of these CC alloy samples were investigated and compared to those of the DS alloys.
The samples were chosen to focus on the variance of Fe-containing intermetallics under slow and fast solidification regimes: SDAS of approximately 21 µm for the DS samples and of approximately 5 µm for the CC samples of the three alloys. Other features within each kind of processing, such as the Si spacing, which correspondingly follows the SDAS size as shown schematically in Figure 1, were examined, and may be regarded as constant. For alloys within each processing, the intermetallic proportion remained almost constant. When comparing the differences across processing methods, a difference in the length-scale of the phases may be seen as will be discussed later. More information regarding the phase proportions within each processing can be found in a previous article [15]. scale of the phases may be seen as will be discussed later. More information regarding th phase proportions within each processing can be found in a previous article [15]. The samples were sanded, polished, and etched (0.5% HF aqueous solution) a needed to reveal the resulting microstructures for analysis [23]. An optical microscop was used to investigate the specimen's dendritic microstructures (Olympus Corporation GX41 model, Tokyo, Japan). The measuring technique described by Gündüz and Çadir [24] was considered to calculate SDAS values. Furthermore, scanning electron microscop (SEM) was employed to register greater magnification images before and after corrosio as well as compositions/morphologies of the generated phases (Philips XL30 FEG equipped with a XFlash 6|60 SDD EDS, Philips, Eindhoven, The Netherlands).
For corrosion tests, 1200# abrasive papers were used to wet grind both 21-μm-SDA and 5-μm-SDAS samples. They were cleaned with anhydrous ethanol and washed wit distilled water before testing.
The corrosion resistance of the Al-Si-Fe-Cu-Zn (-Co, and -Ni) alloys at different as cast conditions was evaluated using electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization tests in a chloride-rich solution. The working electrode were the DS and CC samples (exposed area of 0.4 cm 2 ), the counter electrode was a plati num grid, and the reference electrode was a saturated calomel electrode (SCE) using traditional three-electrode cell setup and a Gamry 600+ potentiostat (Gamry instruments Warminster, PA, USA). The electrolyte was a naturally aerated chloride-rich solution, 0. M NaCl, pH 5.5, made using demineralized water and high purity NaCl reagent (>99 per cent). The tests were conducted in triplicate at ambient temperature and exposed to th air.
The samples were exposed to the electrolyte, and the open circuit potential (OCP registered for 1 h, at the end of which stable potential value ranges were found. After 1 at open circuit, EIS tests started using a frequency of 100 kHz to 10 mHz, and a sinusoida perturbation of 10 mVrms around −20 mV vs. OCP. Non-linear behavior was found fo some Al alloys at the anodic portion of the AC perturbation when tested after early stage of immersion in sodium chloride solution [25], so an cathodic potential of −20 mV vs. OC has been reported as effective to ensure linearity, as further validated in [26][27][28]. Follow ing the EIS testing, the samples were exposed to an open circuit condition for 10 min t ensure that the corroding system could attain the potential values range prior to EIS. Th potentiodynamic polarization measurements were completed in the last phase at 1 mV/ scan rate towards the anodic direction, commencing at a potential of −300 mV below OCP and terminating at +300 mV above OCP. The samples were sanded, polished, and etched (0.5% HF aqueous solution) as needed to reveal the resulting microstructures for analysis [23]. An optical microscope was used to investigate the specimen's dendritic microstructures (Olympus Corporation, GX41 model, Tokyo, Japan). The measuring technique described by Gündüz and Çadirli [24] was considered to calculate SDAS values. Furthermore, scanning electron microscopy (SEM) was employed to register greater magnification images before and after corrosion as well as compositions/morphologies of the generated phases (Philips XL30 FEG equipped with a XFlash 6|60 SDD EDS, Philips, Eindhoven, The Netherlands).
For corrosion tests, 1200# abrasive papers were used to wet grind both 21-µm-SDAS and 5-µm-SDAS samples. They were cleaned with anhydrous ethanol and washed with distilled water before testing.
The corrosion resistance of the Al-Si-Fe-Cu-Zn (-Co, and -Ni) alloys at different ascast conditions was evaluated using electrochemical impedance spectroscopy (EIS) and potentiodynamic polarization tests in a chloride-rich solution. The working electrodes were the DS and CC samples (exposed area of 0.4 cm 2 ), the counter electrode was a platinum grid, and the reference electrode was a saturated calomel electrode (SCE) using a traditional three-electrode cell setup and a Gamry 600+ potentiostat (Gamry instruments, Warminster, PA, USA). The electrolyte was a naturally aerated chloride-rich solution, 0.6 M NaCl, pH 5.5, made using demineralized water and high purity NaCl reagent (>99 percent). The tests were conducted in triplicate at ambient temperature and exposed to the air.
The samples were exposed to the electrolyte, and the open circuit potential (OCP) registered for 1 h, at the end of which stable potential value ranges were found. After 1 h at open circuit, EIS tests started using a frequency of 100 kHz to 10 mHz, and a sinusoidal perturbation of 10 mVrms around −20 mV vs. OCP. Non-linear behavior was found for some Al alloys at the anodic portion of the AC perturbation when tested after early stages of immersion in sodium chloride solution [25], so an cathodic potential of −20 mV vs. OCP has been reported as effective to ensure linearity, as further validated in [26][27][28]. Following the EIS testing, the samples were exposed to an open circuit condition for 10 min to ensure that the corroding system could attain the potential values range prior to EIS. The potentiodynamic polarization measurements were completed in the last phase at 1 mV/s scan rate towards the anodic direction, commencing at a potential of −300 mV below OCP and terminating at +300 mV above OCP.
The corrosion potential (E corr ) was determined from polarization curves (Log i vs. E). EIS data were interpreted from the equivalent electrical circuit (ECC) approach, where the impedance of the selected ECC was fitted to those experimentally obtained, considering the frequency range between 100 kHz and 10 mHz. The Chi-square value (χ 2 ) was used to determine the EIS data's goodness-of-fit to the EEC, which is the total of the square of the differences between theoretical and experimental data.
Immersion tests of specimens having known initial weight were carried out at room temperature so that corrosion rates using the weight loss method could be determined for all alloys and conditions, as described in [29]. The surface areas of the standardized samples were exposed to a corrosive environment of a 0.6 M NaCl solution inside a container for 30 days. To avoid extraneous particles, the container was kept closed. At the completion of the test, the samples were cleaned with nitric acid (HNO 3 ) for 2 min at room temperature to remove corrosion products [29]. To examine the shape/nature of the corrosion layers that developed on the sample's surface, SEM and X-ray diffraction (XRD) analyses were carried out. The X-ray diffraction data of the CC samples were collected on a Bruker D8 Advance ECO diffractometer (Bruker, Billerica, MA, USA) using Cu Kα (λ = 1.54056 Å). Two 2-theta selections were performed, which complied with a specific 15 • -45 • range to decode either the phases of interest or oxides/hydroxides. Figure 2 displays the microstructures highlighting how the size (scale) of the dendritic arrangement varies as a function of the cooling rate (fast and slow solidification) and elemental additions (Ni or Co). Lower cooling rates related to the DS samples (from left to right micrographs) result in an increase of SDAS of approximately four times regardless of the alloy composition. These spectra of solidification rates are essential since they are comparable with those seen in industrial processes such as die-casting, permanent mold, and lost wax casting [30,31]. Furthermore, it appears that adding Co and Ni separately to the Al-7wt.%Si alloy has no effect on the dendritic microstructure length scale. Microstructures like those in Figure 2 were used to quantify the SDAS following the methods proposed by Gündüz and Çadirli [24]. Figure 3 shows SEM images of CC samples that exhibit fine Fe-based and Si phases. The growth of fine intermetallics is related with shorter distances between SDAS, which is quite expected. The DS samples showed the same secondary phases with larger sizes. Such phases are poorly dispersed within the α-Al matrix, being characterized with sharp edges. Some of the Fe-based intermetallics have a Chinese letter morphology [32]. Even with relatively high cooling rates and dendritic fineness of the CC samples, well-distributed particles with a more rounded shape could not be observed [15]. In sum, the microstructures were constituted by an α-Al dendritic matrix, Si and AlFeSi/AlFeSi(Ni)/AlFeSi(Co) phases. Si is found in solution within the α-Al matrix.

Potentiodynamic Polarization
The potentiodynamic polarization curves of the three alloys processed by DS and CC are shown in Figure 4. The polarization curves of the DS alloys, Figure 4a, were similar, characterized by a remarkable increase of the current density even at small anodic polarization. Moreover, slopes of the cathodic branches were quite similar, suggesting that the kinetics of the cathodic reaction were insensitive to the Co or Ni additions. Besides these similarities, the corrosion potential values were also comparable. However, the currents associated with the E corr , i.e., the corrosion current density (i corr ), appear to be the highest for the Ni-containing alloy (black curve), and the lowest for the non-modified alloy (green curve). The polarization curves in Figure 4a do not allow a precise determination of i corr from Tafel extrapolation but a view at E corr clearly indicates that the i corr is likely to be located at superior values between 10 −6 and 10 −5 A/cm 2 . Metals 2022, 12, x FOR PEER REVIEW 6 of 18

Potentiodynamic Polarization
The potentiodynamic polarization curves of the three alloys processed by DS and CC are shown in Figure 4. The polarization curves of the DS alloys, Figure 4a, were similar, characterized by a remarkable increase of the current density even at small anodic polarization. Moreover, slopes of the cathodic branches were quite similar, suggesting that the kinetics of the cathodic reaction were insensitive to the Co or Ni additions. Besides these similarities, the corrosion potential values were also comparable. However, the currents associated with the Ecorr, i.e., the corrosion current density (icorr), appear to be the highest for the Ni-containing alloy (black curve), and the lowest for the non-modified alloy (green curve). The polarization curves in Figure 4a do not allow a precise determination of icorr from Tafel extrapolation but a view at Ecorr clearly indicates that the icorr is likely to be located at superior values between 10 −6 and 10 −5 A/cm 2 .
For the CC processed alloys, the same overall trends found in the DS ones persisted, as seen in Figure 4b. The Ni-containing alloy, despite showing the noblest Ecorr value, displayed the highest icorr, as assessed visually in Figure 4b. Among the tested alloys, the Nicontaining alloys seem to depict the lowest resistance against anodic polarization given the faster increase of the current density for a same anodic polarization from Ecorr. The CC non-modified and Co-containing alloys presented a slightly superior resistance against anodic polarization compared to their DS processed equivalent samples. However, it is still without a clear and extended current density plateau along the anodic polarization. For the CC processed alloys, the same overall trends found in the DS ones persisted, as seen in Figure 4b. The Ni-containing alloy, despite showing the noblest E corr value, displayed the highest i corr , as assessed visually in Figure 4b. Among the tested alloys, the Ni-containing alloys seem to depict the lowest resistance against anodic polarization given the faster increase of the current density for a same anodic polarization from E corr . The CC non-modified and Co-containing alloys presented a slightly superior resistance against anodic polarization compared to their DS processed equivalent samples. However, it is still without a clear and extended current density plateau along the anodic polarization.
After potentiodynamic polarization experiments associated with a short elapsed period in the chloride solution, the surface after testing could be mapped using SEM as a function of large and fine SDAS.  After potentiodynamic polarization experiments associated with a short elapsed pe riod in the chloride solution, the surface after testing could be mapped using SEM as a function of large and fine SDAS. Figure 5 displays typical surface characteristics identified on both the CC (Figure 5a-c) and DS (Figure 5d-f) samples. Owing to their cathodic action the Si and Fe-containing phases dissolved the α-Al areas in contact with them. Moreover the Si phase and Fe-based intermetallics remained at the interdendritic zones.
Damage appears to be more aggravated for the CC samples. As the whole micro structure is refined, as well as the formed interdendritic phases, the α-Al matrix dissolu tion increased, generating a higher number of pittings, as seen in Figure 5a-c. Comparable corrosion morphologies may be seen for the various alloy compositions examined through each process.  Spectroscopy   Figures 6 and 7 show the EIS Nyquist and Bode plots. The Nyquist plots confirmed the presence of two depressed capacitive loops, as also reported by Arthanari et al. [13] for Al-Si-Ni-Cu alloys. Moreover, the Ni-containing alloy's capacitive loop diameters decreased as compared to the other alloys, indicating lower corrosion resistance. The Al-7%Si-0.6%Fe-0.35%Cu-0.25%Zn, Al-7%Si-0.6%Fe-0.35%Cu-0.25%Zn-0.5%Co alloys followed a similar pattern, although the capacitive loop sizes were larger than that related to the Ni-containing alloy samples. Damage appears to be more aggravated for the CC samples. As the whole microstructure is refined, as well as the formed interdendritic phases, the α-Al matrix dissolution increased, generating a higher number of pittings, as seen in Figure 5a-c. Comparable corrosion morphologies may be seen for the various alloy compositions examined through each process. Spectroscopy   Figures 6 and 7 show the EIS Nyquist and Bode plots. The Nyquist plots confirmed the presence of two depressed capacitive loops, as also reported by Arthanari et al. [13] for Al-Si-Ni-Cu alloys. Moreover, the Ni-containing alloy's capacitive loop diameters decreased as compared to the other alloys, indicating lower corrosion resistance. The Al-7%Si-0.6%Fe-0.35%Cu-0.25%Zn, Al-7%Si-0.6%Fe-0.35%Cu-0.25%Zn-0.5%Co alloys followed a similar pattern, although the capacitive loop sizes were larger than that related to the Ni-containing alloy samples. Two peaks occurred in the phase angle plots, as can be seen in Figure 7b, and th phase angle limits and peak areas decreased for Ni-containing alloy samples as compare to the others, indicating least capacitive behavior. Moreover, the modulus of impedanc IZI values of the Ni-containing alloy samples also decreased if compared to the other a loys, showing their inferior corrosion resistances. Two peaks occurred in the phase angle plots, as can be seen in Figure 7b, and the phase angle limits and peak areas decreased for Ni-containing alloy samples as compared to the others, indicating least capacitive behavior. Moreover, the modulus of impedance IZI values of the Ni-containing alloy samples also decreased if compared to the other alloys, showing their inferior corrosion resistances.

Electrochemical Impedance
The fitted equivalent circuit (EC) adopted here is illustrated in Figure 8 [33], while the fitted parameters are provided in Table 1. EC curve fitting analyses were performed for all alloy EIS data. The experimental findings correspond well with the estimated values, and the error (χ 2 ) values are in the acceptable order of 10 −2 . The EC was composed of: R s (solution resistance), R f (surface layer resistance), and R ct (charge transfer resistance), as well as constant phase elements (CPE) of the surface layer (CPE f ) and double layer (CPE dl ), with CPE being characterized by its Q and α parameters. The fitted equivalent circuit (EC) adopted here is illustrated in Figure 8 [33], while the fitted parameters are provided in Table 1. EC curve fitting analyses were performed for all alloy EIS data. The experimental findings correspond well with the estimated values, and the error (χ 2 ) values are in the acceptable order of 10 −2 . The EC was composed of: Rs (solution resistance), Rf (surface layer resistance), and Rct (charge transfer resistance), as well as constant phase elements (CPE) of the surface layer (CPEf) and double layer (CPEdl), with CPE being characterized by its Q and α parameters.  The fitted equivalent circuit (EC) adopted here is illustrated in Figure 8 [33], while the fitted parameters are provided in Table 1. EC curve fitting analyses were performed for all alloy EIS data. The experimental findings correspond well with the estimated values, and the error (χ 2 ) values are in the acceptable order of 10 −2 . The EC was composed of: Rs (solution resistance), Rf (surface layer resistance), and Rct (charge transfer resistance), as well as constant phase elements (CPE) of the surface layer (CPEf) and double layer (CPEdl), with CPE being characterized by its Q and α parameters.   As shown in Table 1, the R f values of the non-modified, Co-and Ni-containing alloys were 45.9, 30.4, and 10.0 kΩ·cm 2 for the DS samples, respectively. For the CC samples, 31.7, 31.7, and 7.6 kΩ·cm 2 were determined for the same sequence of alloys. These findings revealed that the surface layer produced on the surface during electrolyte exposure contributed to the decrease in corrosion resistance of the Ni-containing alloy. Interestingly, the DS alloys exhibited higher R f values compared to the CC processed alloys, attributed to the presence of larger Si phase and Fe-based intermetallics, decreasing the microgalvanic coupling areas between these particles and the α-Al matrices.
Charge transfer resistance (R ct ) values were found to be 100.9, 111.3, and 67.3 kΩ·cm 2 and 69.7, 62.7, and 9.4 kΩ·cm 2 for the non-modified, Co-modified, and Ni-modified alloys considering DS and CC samples, respectively. The Ni-modified alloy exhibited approximately 1.5 and 7.4 times lower R ct values compared to the alloy without modification, confirming its worst corrosion resistance. The difference in R ct between the Co-containing and the non-modified alloys was insignificant.
The EIS results (see Figure 7) agree with those from potentiodynamic polarization (see Figure 4), both pointing to the Ni deleterious role regarding the corrosion response after a short immersion period (1 h). The assumption that the corrosion will proceed at a constant rate may be misleading for alloys that develop a protective layer for long exposure time to the electrolyte. Indeed, the different corrosion behaviors at early-and long-exposure time to the corroding electrolyte are often the source of controversial conclusions about the effect of a given alloying element to the Al alloys.

Immersion Corrosion Results
Electrochemical analyses (short term) showed that the addition of Ni was detrimental to the corrosion resistance while the other two alloys showed similar results. That is, the Co addition did not affect the corrosion behavior determined by polarization and EIS. These results should be considered with caution, since they reflect the first stage of the corrosion process under the enforced parameters, as detailed in the last section. There is no electrochemical approach that can be used alone as an accelerated test for evaluating Al alloys [37]. The findings of electrochemical studies must be compared to those of traditional tests, such as weight loss experiments, for instance. This sort of test has allowed assessing the performance of Al alloys in a variety of environments and conditions. To supplement the current findings, immersion corrosion experiments were carried out for 30 days to determine the corrosion rates of all examined alloys.
In Figure 9, macroscopic images of the sample surfaces after 30 days of immersion in 0.6 M NaCl with and after the removal of the cleaning procedure, which aims examining the corrosion products, are compared. After 30 days in 0.6 M NaCl solution, the corrosion product surface layer can be seen covering the surface of all alloys. Before removal procedure, white corrosion products were less identified for the Ni-containing alloy, being concentrated in a vertical center line aligned with the sample hole. In the rest of the sample surfaces a more uniform grayish layer was observed as indicated by the arrows in Figure 9. As also observed by Kaiser et al. [9], the white products were identified as being "mushroom-type" layers. The active dissolution of the alloy resulted in Al(OH)3, which is insoluble in water, and precipitates as a white gel with a gelatinous flake aspect [37]. The gray layer related to the Ni-containing alloy in Figure 9 appears to have a nature other than the hydroxide gel. More details about this structure will be seen later.
After cleaning, the samples were weighed to be compared to the masses before immersion. Figure 10 shows the resulting corrosion rates. The weight loss of the alloy containing Ni is much lower as compared to the others, while some mass gain related to the CC samples could be observed. The formation of protective and dense layer on the surface of the Ni-containing samples may be the cause of this slight weight gain. This layer worked as a corrosion barrier, severely reducing the corrosion rate. It should be noted that a tiny portion of the corrosion products were not eliminated for the Ni-containing alloy; however, this did not affect the mass measurement results when compared to the other alloys. In other words, the corrosive rate was indeed significantly lower for the alloy with Ni, as assessed after 30 days of exposure. In their immersion tests, Kaiser et al. [9] observed an inversion of the behavior of the Al alloy containing Ni. While the alloy containing Ni had a greater corrosion rate for shorter periods (one to three days), it demonstrated a lower corrosion rate for longer elapsed times. This is consistent with what is reported in the current study when the results through the electrochemical and the 30-day immersion tests are compared to each other. As also observed by Kaiser et al. [9], the white products were identified as being "mushroom-type" layers. The active dissolution of the alloy resulted in Al(OH) 3 , which is insoluble in water, and precipitates as a white gel with a gelatinous flake aspect [37]. The gray layer related to the Ni-containing alloy in Figure 9 appears to have a nature other than the hydroxide gel. More details about this structure will be seen later.
After cleaning, the samples were weighed to be compared to the masses before immersion. Figure 10 shows the resulting corrosion rates. The weight loss of the alloy containing Ni is much lower as compared to the others, while some mass gain related to the CC samples could be observed. The formation of protective and dense layer on the surface of the Ni-containing samples may be the cause of this slight weight gain. This layer worked as a corrosion barrier, severely reducing the corrosion rate. It should be noted that a tiny portion of the corrosion products were not eliminated for the Ni-containing alloy; however, this did not affect the mass measurement results when compared to the other alloys. In other words, the corrosive rate was indeed significantly lower for the alloy with Ni, as assessed after 30 days of exposure. In their immersion tests, Kaiser et al. [9] observed an inversion of the behavior of the Al alloy containing Ni. While the alloy containing Ni had a greater corrosion rate for shorter periods (one to three days), it demonstrated a lower corrosion rate for longer elapsed times. This is consistent with what is reported in the current study when the results through the electrochemical and the 30-day immersion tests are compared to each other. Under these long-term corrosion exposure conditions, it appears that neither the addition of Co nor the scale of the dendritic microstructure has significant impacts on the corrosion rate. However, the Ni-containing alloys exhibited an unusual behavior.
XRD results were analyzed for the three rapid solidified alloy samples subjected to the 30-day immersion corrosion (see Figure 11). In the case of the alloy with Ni, the central position (white area) and periphery (gray area) were examined through separate spectra. All results confirmed the formation of the Al(OH)3. Aside from the phases that constitute the alloy structures, all spectra revealed the existence of Al2O3 oxides in the protective layers that had been formed. Formation of corrosion products often displayed an outer hydroxide layer and an inner oxide layer at the alloy's surface, with the latter being formed from further oxidation of hydroxides. Oxides are recognized to present a more dense, percolated, and more effective structure at restraining the ingress of corrosive species towards the active bare surface of the alloy, which is desirable as a protective layer, as compared to hydroxides and oxyhydroxides. Under these long-term corrosion exposure conditions, it appears that neither the addition of Co nor the scale of the dendritic microstructure has significant impacts on the corrosion rate. However, the Ni-containing alloys exhibited an unusual behavior.
XRD results were analyzed for the three rapid solidified alloy samples subjected to the 30-day immersion corrosion (see Figure 11). In the case of the alloy with Ni, the central position (white area) and periphery (gray area) were examined through separate spectra. All results confirmed the formation of the Al(OH) 3 . Aside from the phases that constitute the alloy structures, all spectra revealed the existence of Al 2 O 3 oxides in the protective layers that had been formed. Formation of corrosion products often displayed an outer hydroxide layer and an inner oxide layer at the alloy's surface, with the latter being formed from further oxidation of hydroxides. Oxides are recognized to present a more dense, percolated, and more effective structure at restraining the ingress of corrosive species towards the active bare surface of the alloy, which is desirable as a protective layer, as compared to hydroxides and oxyhydroxides.
The details of stable oxide layers (areas indicated by arrows in Figure 9) mostly composed of Al 2 O 3 on the surface for the alloy containing Ni may be seen in Figures 12 and 13. According to the results from Arthanari et al. [33], the generation of this type of structure may be associated with the reduction in the amount of hydrogen developed as the immersion period was extended for the Ni-containing alloys. This prevents aggressive ions from penetrating deeper into the system, improving corrosion resistance. It appears that Ni addition could provide reduced hydrogen volume, suggesting its positive role as a modifying element. Indeed, the cathodic reactions that generate gases play an important role in the development and integrity of protective layers on alloys [38]. Reducing the rate of hydrogen developed underneath the protective layer prevents its severe and continuous spalling, favoring thickening and conversion of corrosion products into oxides.  The details of stable oxide layers (areas indicated by arrows in Figure 9) mostly composed of Al2O3 on the surface for the alloy containing Ni may be seen in Figures 12 and  13. According to the results from Arthanari et al. [33], the generation of this type of structure may be associated with the reduction in the amount of hydrogen developed as the immersion period was extended for the Ni-containing alloys. This prevents aggressive ions from penetrating deeper into the system, improving corrosion resistance. It appears that Ni addition could provide reduced hydrogen volume, suggesting its positive role as a modifying element. Indeed, the cathodic reactions that generate gases play an important role in the development and integrity of protective layers on alloys [38]. Reducing the rate of hydrogen developed underneath the protective layer prevents its severe and continuous spalling, favoring thickening and conversion of corrosion products into oxides.
Because the secondary phases are nobler than the α-Al phase, they function as cathodes speeding up the dissolution of the α-Al matrix, as can be seen in the corroded regions highlighted in Figure 12. The α-Al dissolution is quite evident. CC and DS samples showed comparable corrosion morphologies, with selective dissolution of the α-Al phase verified, and secondary phases mostly unaffected, continuing to occupy their sites in the interdendritic regions.

Conclusions
In this study, the microstructure formation and corrosion behavior of as-cast Co and Ni containing Al-Si-based alloys were investigated, and the following findings were reached: • The microstructures of the alloys were formed by the α-Al dendritic matrix, Si and AlFeSi/AlFeSi(Ni)/AlFeSi(Co) phases. The additions of either Co or Ni were not able Because the secondary phases are nobler than the α-Al phase, they function as cathodes speeding up the dissolution of the α-Al matrix, as can be seen in the corroded regions highlighted in Figure 12. The α-Al dissolution is quite evident. CC and DS samples showed comparable corrosion morphologies, with selective dissolution of the α-Al phase verified, and secondary phases mostly unaffected, continuing to occupy their sites in the interdendritic regions.

Conclusions
In this study, the microstructure formation and corrosion behavior of as-cast Co and Ni containing Al-Si-based alloys were investigated, and the following findings were reached:

•
The microstructures of the alloys were formed by the α-Al dendritic matrix, Si and AlFeSi/AlFeSi(Ni)/AlFeSi(Co) phases. The additions of either Co or Ni were not able to change the SDAS as compared to the non-modified alloy. The fast cooling process resulted in a reduction of SDAS of approximately four times.

•
Considering the short elapsed time measurements, the currents associated with the E corr , i.e., the corrosion current density (i corr ), appear to be highest for the Ni-containing alloy and lowest for the unmodified alloy. While not applying the Taefl extrapolation, the estimated i corr values of the Ni-containing alloy were found to be higher, ranging from 10 −6 to 10 −5 A/cm 2 . Moreover, the resistance to anodic polarization was marginally higher in the CC non-modified and Co-containing alloys in comparison to the corresponding DS samples. • Due to the formation of a thick and dense alumina layer containing Ni in its inner layer, the Ni-containing alloy showed a lower corrosion rate under long exposure conditions (30 days). It is understood from the results of the present investigation that the addition of a small amount of Ni may be beneficial for longer exposure times to the saline electrolyte. Data Availability Statement: Data presented in this study are available on request from the corresponding author. Data are not publicly available because pertain to a research still in development.

Conflicts of Interest:
The authors declare no conflict of interest.