Solution Treatment—T6

Some research has been performed on the effect of solution heat treatments on the corrosion behaviour of SLM AlSi10Mg. In one study, Gu et al. [68] solution treated SLM produced AlSi10Mg material at 300 and 400 ◦C for 2 h under an argon atmosphere, followed by water quenching. In another study by Gu et al. [58], SLM produced AlSi10Mg material was solution treated at 450, 500, and 550 ◦C for 2 h under an argon atmosphere, followed by water quenching. In both studies, experiments were conducted in 3.5 wt.% NaCl solution.

In untreated specimens, the continuous Si network impedes contact between the underlying metal matrix and the electrolyte, restricting the transfer of Al3+ ions towards the electrolyte [58]. After heat treatment, the continuous Si network devolved into separate Si precipitates (see the previous section). Higher temperature heat treatments, coinciding with the growth of the Si particles, resulted in higher corrosion current densities and lower corrosion potentials, film resistances, and charge transfer resistances [58,68]. This was attributed to the change in the morphology of the Si phase. Larger separate Si precipitates restrict the formation of a compact oxide layer [58,68] while Si precipitating from the Al matrix leads to an increasing cathode area and a more active matrix [58]. Further explanations can be found by looking at the work of Revilla et al. [32]. SKPFM measurements showed that the Volta potential difference between the Si phase and the matrix increases with the size of the particles due to the influence of the surrounding matrix. As the Si particles grow in size with increasing temperature, a higher driving force for galvanic corrosion is obtained.

#### Artificial Aging Heat Treatments

Rubben et al. [54] studied the effect of AA on the corrosion properties of as produced SLM specimens. Open circuit potential (OCP), linear sweep voltammetry (LSV), and immersion experiments in 0.1 M NaCl showed a similar behaviour, resulting from the lack of significant changes in microstructure compared to the untreated specimens (see the previous section). This shows that performing the artificial aging without prior solution treatment is not only critical for retaining the mechanical properties [36] but also for retaining the corrosion resistance of SLM material.

#### *3.3. E*ff*ect of Si Content on Microstructure and Corrosion*

Even though several studies have been dedicated to investigating the corrosion behaviour of AM AlSi10Mg [32,33,44–61,68,70], and a few other works have studied the corrosion performance of AlSi12 alloy [33,42,62,63], very limited research explores the influence of Si on the microstructure and corrosion behaviour of these materials. A recent comparative study carried out on polished as-built AlSi7Mg0.6, AlSi10Mg, and AlSi12 demonstrated that the general appearance concerning the structure of melt pools for all of these additively manufactured specimens was approximately the same [33]. All the specimens, independently on the amount of Si in the alloy, presented a fine cellular structure within the melt pools, while at the borders coarser cells were detected. Additionally, heat-affected

zones, in which idiomorphic Si particles and a relatively more discontinuous Si network are present, were identified next to the MPBs for all of these specimens. Nevertheless, higher resolution SEM analysis revealed that even though the size of the Al cells was independent of the amount of Si in the alloy, the connectivity of the Si network was highly a ffected by the Si content. A much thinner and partially broken Si network was observed for the as-built AlSi7Mg0.6 specimen, while a thicker and much more connected Si network was seen in the as-built AlSi12 sample. By increasing the amount of Si, a greater level of connectivity is observed in the silicon network of the AM specimens. Similarly, a relatively higher level of connectivity was seen in the silicon network within the HAZ for higher amounts of Si [33].

The corrosion studies on these three AM specimens revealed that, even though the corrosion attacks seemed rather superficial for all the samples, the corrosion resistance in NaCl solution is slightly influenced by the specific alloying content: materials with higher amounts of Si (and lower Mg content) showed higher resistance against corrosion [33]. The corrosion resistance of the materials showed the following relationship: AlSi12 > AlSi10Mg > AlSi7Mg0.6. The relatively low resistance against corrosion for the AlSi7Mg0.6 specimen could be due to the higher amount of Mg present in this material compared to the others, since the presence of Mg in Al-Si alloys has repeatedly been shown to greatly influence their corrosion behaviour through the formation of Mg2Si precipitates [71–73]. Furthermore, earlier research has shown that the corrosion resistance of Al-Si alloys decreases by increasing the content of Mg [73]. Nevertheless, it is important to notice that (to the best of our knowledge) the actual formation of Mg2Si particles in additively manufactured Al-Si alloys has not ye<sup>t</sup> been confirmed, due possibly to the extremely high cooling and solidification rates resulting in a very fine (out-of-equilibrium) distribution of alloying elements with almost no time for the formation of the usual precipitates. We believe that the possible formation of Mg- and Fe-containing precipitates in these materials should be more systematically studied in future works.

Moreover, due to the relatively high connectivity of the Si network in the as-built AM AlSi12 material, no micro-cracks were formed after corrosion; while for the case of AlSi7Mg0.6 and AlSi10Mg several micro-cracks were seen at heat-a ffected-zones next to the MPBs [33]. This selective cracking seems to be caused by the combination of several factors: the relatively larger disruption of the Si network in the HAZ, the selective dissolution of the Al matrix around the MPBs/HAZ due to the presence of a corrosive medium, and the existence of residual internal stresses from the MAM process [33].

Cabrini et al. also conducted a comparative study using AlSi7Mg0.6 and AlSi10Mg to investigate the combined e ffect of Si content and heat treatment on the susceptibility of these materials to corrosion attack [46]. They demonstrated that in the case of as-built (untreated) specimens, higher Si content in the material resulted in a lower penetration depth of the corrosion attack. As mentioned above, this could be due to the higher level of connectivity achieved in the Si network for higher amounts of Si in the alloy [33]. On the other hand, after heat treatments, the Si network (partially) breaks into isolated Si particles, resulting in greater penetration depth of the corrosion attacks for samples with the highest Si content. For higher amounts of Si, larger Si particles will be formed after heat treatment. These isolated Si particles present a more cathodic potential than the Al matrix, being the main cause for the localized corrosion in these materials (due to galvanic coupling e ffect). For larger Si content and Si particles formed after heat treatment, larger will be the extent of this corrosion attack, which will then not be contained/stopped by the (partially) broken Si network.

#### *3.4. E*ff*ect of Surface Roughness on Corrosion*

In order to study the e ffect of surface roughness on the corrosion behaviour of additively manufactured Al-Si alloys, several works compare the corrosion performance of as-built AM specimens with that of polished or ground samples [44,50,53,60,70]. In most of these cases, it was shown that the corrosion resistance in NaCl solution [44,50,53,60], as well as the low cycle corrosion fatigue life span [70], is improved after mechanical polishing. The reduced corrosion resistance of as-produced

SLM samples was ascribed to the excessive number of cavities and other surface defects generated during the SLM process at the external sample surface. These surface defects will induce localized corrosion in the form of pits, which could then be considered as crack initiation sites. The stimulated crack initiation and propagation of the unpolished samples thus resulted in a relatively accelerated corrosion fatigue failure [70]. Nevertheless, while studying the e ffect of surface finishing on the corrosion properties of AlSi10Mg prepared by direct laser sintering, Fathi et al. [60] concluded that during the initial stage of immersion in NaCl solution, SLS-prepared samples had a higher corrosion resistance than ground specimens. Ground specimens were characterised by a high selective attack mainly at the heat-a ffected zones next to the MPBs, while only a minor attack on the surface of as produced samples was seen. The authors believe that this is caused by a less protective passive film on the ground samples. However, for long immersion times, a change in corrosion behaviour was seen. It was reported that the ground SLS samples now had the highest corrosion resistance. This might indicate that during the initial stage of immersion in a corrosive medium, the internal microstructural features (such as MPBs), which are more pronounced for polished/ground specimens, play a more dominant role in the corrosion process.

An improved surface quality of AM samples is, however, not always possible to obtain by post-printing operations, like mechanical polishing/grinding due to the complex shape of the final printed parts. Therefore, methods such as shot peening and sandblasting (among others) are sometimes applied in order to improve the surface quality of the as-built parts. Some works have been dedicated to studying the e ffect of such methods on the corrosion behaviour of these materials. The results demonstrate that shot-peened and sandblasted materials present a slightly higher resistance against corrosion than as-built parts due to the reduction of surface roughness and superficial defects [50,60]. However, a much better corrosion performance is obtained always for the polished or ground samples. Fathi et al. [60] propose that it is of the utmost importance to perform a post-grinding operation, in order to substantially improve the corrosion properties of the printed parts. Cabrini et al. [48] further showed an improvement of the corrosion resistance of as-produced parts by bright dipping in a phosphoric/nitric acid bath, attributed to the removal of the oxide film formed during printing. Moreover, they noticed that the corrosion resistance worsened after a bright dipping of mechanically polished parts, caused by the silicon enrichment at the surface during etching.

Furthermore, the surface roughness can also be improved by changing the printing parameters. Calignano et al. [74] studied the e ffect of di fferent printing parameters on the surface roughness of AlSi10Mg DMLS produced samples. They found that the scanning speed had the greatest influence on the surface roughness, followed by the hatching distance. Fathi et al. [57] studied the corrosion properties of as-produced SLS samples with improved surface quality. The reduced surface roughness was obtained by adjusting the beam o ffset and reducing the scanning speed and hatching distance, as was suggested by Calignano et al. [74]. They reported that by changing the printing parameters, like the hatch distance, not only does the surface quality change for the better, but the degree of overlap between melt pools changes as well. Furthermore, this study showed that by reducing the hatch distance a material with periodically large and small melt pools is obtained. This overlap between melt pools was shown to play an important role in the solidification behaviour: for higher overlap, the higher the solidification rate will be. The authors claimed that in those cases not the surface roughness, but the resulting microstructure has the largest e ffect on the corrosion properties of the as-produced parts. Moreover, the material with the finest microstructure, originating from the higher degree of melt pool overlap, showed the best corrosion properties. This was attributed to the fact that the coarser the microstructures of the Al dendrites and the Si particles are, the higher the galvanic coupling and galvanic corrosion [57].

#### *3.5. Corrosion Protection*

Even though several studies have demonstrated that the special conditions during MAM have a grea<sup>t</sup> influence on the microstructure and corrosion behaviour of these materials, not many works have been dedicated to investigating the influence of their special microstructure on the mechanisms of corrosion protection. To date, only a few studies explore the impact of microstructure, heat treatments, Si content, and defects on the anodizing behaviour of Al–Si alloys prepared by selective laser melting. Anodizing of aluminium and aluminium alloys is done to improve their corrosion protection. Galvanostatic anodizing of SLM produced Al-Si alloys in H2SO4 electrolyte was studied by Revilla et al. [41,43,67,75]. They concluded that it is possible to anodize SLM produced Al–Si alloys. However, it was shown that the characteristic microstructure of SLM produced parts will have a significant impact on the voltage-time response and on the formed anodic oxide film. Furthermore, they showed a rather significantly di fferent anodizing behaviour when compared to the cast alloy of similar chemical composition. The voltage-time response of the cast alloy shows the typical steady-state growth regime, while the AM material, on the other hand, showed a continued increase in voltage until eventually a steady-state was reached. Due to the fine distribution and high connectivity of the Si network in as-built AM Al-Si alloys, the moving oxide front will be obstructed to a larger extent resulting in the formation of a thinner oxide film. Moreover, a significantly lower oxide growth rate was seen for the SLM alloy compared to the cast alloy. This lower oxide growth rate was attributed to a larger fraction of the anodic charge consumed by Si oxidation in the SLM alloy. The e ffect of the Si distribution on the anodizing behaviour of Al-Si alloys was also studied [43]. It was shown that the melt pool borders in SLM produced parts have a eutectic Al-Si structure with alternating lamellae consisting of Al and Si. This lamellae structure was attributed to the lower cooling rate in these melt pool borders compared to the centre of the melt pools. Furthermore, it was reported that in these melt pool borders the Al content was slightly lower compared to the average value in the whole sample, and the Si content slightly higher [43]. This slightly higher Si content and eutectic Al-Si lamellar structure in the melt pool borders was suggested to cause the relatively thinner oxide film formed at these melt pool borders. An XPS analysis reported that in the anodic film of the cast alloy only a superficial layer of the Si precipitates is oxidized. However, in the SLM samples, most of the Si in the anodic film is oxidized; this was suggested to cause a severe reduction of the anodizing e fficiency compared to the cast alloy. It was shown that the anodic film formed on the cast alloy has a much higher roughness than the anodic film formed on the AM specimen. This roughness di fference was attributed to the di fference in aluminium cell size between the cast alloy and the SLM alloy [43].

Moreover, a grea<sup>t</sup> anisotropy was also seen during the galvanostatic anodizing of as-built AM Al-Si samples [41,75]. The voltage–time response curves of the AM material showed a di fferent response depending on the orientation of the surface that was anodized. The voltage–time response of a surface parallel to the building platform (XY surface) was shown to give a higher steady-state potential value than that of a surface with an orientation perpendicular to the building platform (XZ surface). This asymmetric anodizing behaviour was found to be independent of the Si content in the alloy. The authors proposed that this asymmetry could be related to the di fference in the size of aluminium cells in the di fferent planes [41], or di fferences in the density of melt pool borders encountered by the anodizing front [75].

A unique pore structure was seen for the SLM alloys. A branched-like pore structure was seen throughout the whole anodic film. This pore structure was attributed to the fine distribution of the silicon phase in an almost continuous network encapsulating the aluminium in small cells [41]. Revilla et al. [75] further reported that the Si content in the alloy has a significant e ffect on the pore structure of the anodic film, i.e., the higher the Si content the higher the voltage response will be. As a consequence of the higher voltage response, wider pores with a greater inter-pore distance are obtained. Furthermore, the pore density is shown to decrease with Si content.

Rubben et al. [67] studied the e ffect of several heat treatments on the anodizing behaviour of SLM produced AlSi10Mg. It was shown that anodizing behaviour depends strongly on the morphology of the Si phase, which can be highly a ffected by the heat treatment applied. For the non-heat treated and the artificially aged (at 170 ◦C for 6 h) samples, the Si phase consists out of a rather continuous network, consequently, most of the Si encountered by the anodizing front will be oxidized. For the stress

released heat-treated samples (at 250 and 300 ◦C for 2 h), the Si network is broken up into separated Si precipitates [67]. As a consequence, only a fraction of the Si phase will be oxidized. It was shown that the di fference in Si phase morphology can give rise to significantly di fferent anodizing behaviour. Furthermore, they reported that voltage–time response during anodizing varied strongly depending on the heat treatment applied. The steady-state anodizing potential was reported to increase with the anodized Si fractions [67]. Additionally, the anodizing e fficiency was shown to decrease with a higher anodized Si fraction.

Finally, Revilla et al. [75] investigated the influence of internal pores, resulting from the MAM process, on anodizing behaviour. They showed that when the moving oxide front encounters such a pore, the sides of the pores ge<sup>t</sup> anodized. Furthermore, the sides of the anodized internal pores showed cracks, probably caused by the volume expansion of anodic layers advancing in opposite directions [75]. Even though the cracks observed were relatively small (1–2 μm), they could compromise the physical integrity of the final piece. Additionally, these cracks could facilitate the access of corrosive media to the metal matrix.

Other studies are needed in the future to further understand the anodizing behaviour of these materials under other anodizing conditions such as: potentiostatic anodizing regime, di fferent electrolytes, and study the e ffect of surface roughness. Moreover, the e ffect of microstructure and microstructural defects on other surface treatments aimed at the protection of the materials against corrosion should also be investigated.

#### **4. Other Al Alloys**

Besides the Al-Si alloys discussed in previous sections, only very few other additively manufactured Al-based alloys have been considered so far for corrosion studies. The only other cases are, to the best of our knowledge, AA2024 and AA7075 [76,77]. For AA2024, the main alloying agen<sup>t</sup> added is Cu, while for AA7075 the primary alloying element is Zn. The specifications concerning the chemical composition of these alloys are given in Table 2. AA2024 is a heat-treatable alloy widely used in aircraft structures due to its high strength to weight ratio, as well as good fatigue resistance. AA7075 has high specific strength, low density, and good thermal properties, for which it is widely used in transport applications, including marine, automotive, and aviation, as well as in-mold tool manufacturing.


**Table 2.** Specification of chemical composition for AA2024 and AA7075.

Previous work on polished wrought and additively manufactured AA2024 revealed that due to the special conditions associated with the selective laser melting process (i.e., highly localized melting and subsequent rapid solidification), the microstructure of AM AA2024 was characterized by a refined particle size [76]. The traditional micrometre-sized constituent particles and the S-phase Al2CuMg present in wrought AA2024-T3 were instead replaced by nm-sized particles, which were mainly determined to be θ-phase Al2Cu. This highly refined microstructure was shown to greatly impact the corrosion behaviour of this material according to results obtained in polished samples [76]. While the anodic current was found to increase rapidly for potentials above the corrosion potential during anodic polarization of wrought AA2024-T3, the additively manufactured AA2024 materials presented a passive-like window. This could be due to the finer microstructure found in the additively manufactured material compared to the wrought specimen, which resulted in a thicker and more stable native oxide film; as well as the absence of S-phase precipitates in AM AA2024. It was also shown that the corrosion rate of the Al matrix in 0.01 M NaCl for the AM AA2024 material was about five times lower than that measured in wrought AA2024-T3. Moreover, the dissolution ratio between the alloying elements (Cu and Mg) and that of Al was about ten times higher for the AM AA2024 sample

than for the wrought AA2024-T3. As the authors stated, the dissolution of these alloying elements at the early stages of corrosion is beneficial since it reduces the number of cathodic sites that would be detrimental for subsequent localized corrosion.

The relation between microstructure and corrosion behaviour has also been studied for AA7075 materials [77]. Gharbi et al. [77] demonstrated that while coarse (up to approximately 15 μm in size) second phase Al-Cu-Fe(-Si) particles exist in the traditional wrought AA7075-T6 material, finely distributed nm-sized Mg-Zn-Cu(-Al) (ν-phase) and Mg2Si (β-phase) characterize the microstructure of additively manufactured AA7075 prepared by SLM. This microstructure was also shown to vary significantly after solutionising and subsequent artificial ageing. The existing ν-phase in as-built specimens was dissolved after solutionising, while the subsequent aging resulted in the formation of MgZn2 precipitates. Concerning their corrosion behaviour, the authors demonstrated that, depending on the post-heat-treatment, AM AA7075 presented a higher corrosion resistance compared to wrought AA7075-T6 [77]. While as-built AM AA7075 and wrought AA7075-T6 presented a highly active behaviour during anodic polarization in NaCl solution, the heat-treated specimens presented a passive-like behaviour. The immersion of the samples in a NaCl solution revealed that pits formed in as-built AM AA7075 were notably smaller compared to those formed in wrought AA7075-T6, possibly due to the fine microstructural features and the absence of large second phase particles in the AM material compared to its wrought counterpart. Moreover, while pitting occurred predominantly along the melt pool borders in the case of polished as-built AM AA7075, pits were uniformly distributed on the Al matrix for the solutionised AM material. Due to the formation of MgZn2 precipitates in the solutionised aged samples, large pits were formed along grain boundaries [77].

To enhance the metallurgical state of the AA7075 alloy, and therefore, make it more suitable for processing by additive manufacturing, the minor addition of transition elements such as Sc, Zr, Ti, B, Fe, and Ni has been explored [78]. The addition of these elements influenced the resulting microstructure of AM AA7075. Therefore, further studies should focus on investigating the effect of this microstructure and additional alloying elements on the corrosion resistance of this material.

In general, as for AM Al-Si alloys, the special conditions during additive manufacturing (i.e., highly localized melting and solidification accompanied by extremely high cooling rates) seem to promote a high refinement of the microstructure of these materials and/or the annihilation of certain precipitates/phases. This is generally beneficial in terms of corrosion resistance since it greatly reduces the possible cathodic sites and the galvanic interactions that can promote localized corrosion. Nevertheless, there is a need for further investigations of the corrosion behaviour of other AM Al-based alloys, or Al alloys prepared by MAM techniques other than selective laser melting.

#### **5. Summary and Outlook**


affected by defects and special features typical from the metal additive manufacturing process. The intrinsic anisotropy, melt pool borders, and internal pores are among those features affecting the anodization process. More research is still needed to better understand the impact of other defects such as unmolten powder and residual stresses on the anodizing behavior of these materials, as well as to further investigate the properties of the formed anodic oxides. Moreover, additional research is required to understand the anodizing behaviour of these materials under other anodizing conditions such as potentiostatic anodizing regime and different electrolytes.

**Author Contributions:** Conceptualization, R.I.R.; writing—original draft preparation, R.I.R., D.V., and T.R.; writing—review and editing, I.D.G., D.V., T.R., and R.I.R. All authors have read and agreed to the published version of the manuscript.

**Funding:** This research received no external funding.

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