Halogen-free Phosphonate Ionic Liquids as Precursors of Abrasion Resistant Surface Layers on Az31b Magnesium Alloy

Surface coatings formed by immersion in the ionic liquids (ILs) 1,3-dimethylimidazolium methylphosphonate (LMP101), 1-ethyl-3-methylimidazolium methylphosphonate (LMP102) and 1-ethyl-3-methylimidazolium ethylphosphonate (LEP102) on magnesium alloy AZ31B at 50 °C have been studied. The purpose of increasing the temperature was to reduce the immersion time, from 14 days at room temperature, to 48 hours at 50 °C. The abrasion resistance of the coated alloy was studied by microscratching under progressively increasing load, and compared with that of the uncoated material. The order of abrasion resistance as a function of the IL is LEP102 > LMP101 > LMP102, which is in agreement with the order obtained for the coatings grown at room temperature. The maximum reduction in penetration depth with respect to the uncovered alloy, of a 44.5%, is obtained for the sample treated with the ethylphosphonate LEP102. However, this reduction is lower than that obtained when the coating is grown at room temperature. This is attributed to the increased thickness and lower adhesion of the coatings obtained at 50 °C, particularly those obtained from methylphosphonate ionic liquids. The results are discussed from SEM-EDX and profilometry.


Introduction
In order to protect the light and highly reactive magnesium alloys from surface damage failures due to corrosion and/or wear, it is necessary to apply surface protecting coatings which prevent or reduce corrosive attack and severe wear [1,2]. Room temperature ionic liquids are fluids composed of ions which are stable in the liquid state at room temperature. They present a variety of properties such as high thermal stability, a wide electrochemical window, high conductivity, negligible volatility and nonflammability, which make them useful in a variety of applications, from solvents to thermal fluids. From the materials science and surface engineering point of view, ionic liquids have shown outstanding potential as lubricants and lubricant additives  including lubrication of light alloys [14,24,[28][29][30][31][32][33][34], nanophase modifiers [35,36], electrolytes [37], corrosion inhibitors [38][39][40][41][42][43][44][45][46][47][48][49] or reagents in the formation of corrosion protective surface coatings [50][51][52].
Phosphorus-containing coatings, in particular phosphonate derivatives [1,53,54] are currently been studied as corrosion protective layers on magnesium alloys due to their biocompatibility. Phosphorus-containing ionic liquids have been extensively applied in corrosion protection of magnesium alloys [55][56][57][58][59][60][61][62][63][64]. Together with corrosion, another major cause of surface failure of magnesium alloys is their poor tribological performance. Even the best lubricants, including ionic liquids fail to reduce friction and protect magnesium against severe wear in sliding against other materials.
Phosphonate imidazolium ionic liquids have shown to be highly reactive towards copper [65]. In a previous work [66], we have shown that AZ31B magnesium alloy can be protected against abrasion by surface layers grown by long-term immersion in phosphonate imidazolium ionic liquids at room temperature. In order to reduce immersion time, we have now studied the effect of temperature on the nature and abrasion resistance of the coatings.

Growth of Coatings on Mg AZ31B.
Figures 3 and 4 show SEM micrographs, EDX spectra and element maps of the surface layers on AZ31B after treatment with LMP101 and LMP102, respectively.
The EDX spectrum in Figure 3c shows the presence of oxygen, phosphorus and magnesium. Magnesium and phosphorus element maps ( Figure 3d) are in agreement with the phosphonate coating grains and the magnesium alloy substrate at the grain boundaries. Similar microstructures have been observed for LMP102. The main differences between the new coatings obtained at 50 °C with respect to those at room temperature [66], apart from the expected increase of reaction velocity, with a reduction in reaction time from 2 weeks to 48 hours, are their larger thickness and poorer adherence, as shown by the lack of coating layer in some regions of the surface (Figure 3a).
In the case of LMP102 (Figure 4a-d), the poor adherence of the coating layer and the starting delamination process of the coating scales can be clearly observed ( Figure 4b). As the cracking of the coating layers could be due to the vacuum conditions of the SEM observation, surface topography profiles were observed, to find a discontinuous layer also under ambient conditions ( Figure 5). The regions colored in red in Figure 5 correspond to magnesium substrate, clearly visible on the edges of the wear path, but also outside it.
The thickness of the coating obtained from LEP102 ( Figure 6) is lower, as can be observed by the lower intensity P and O peaks with respect to that of Mg, as compared with the coatings obtained from the methylphosphonate ionic liquids LMP101 and LMP102 (Figure 4c). This thinner layer can be also observed in the EDX element maps, where the presence of phosphorus is not detected in some regions of the coated surface.
(c) (d)  The observed differences in the coating layers obtained from the ethylphosphonate ionic liquid LEP102 and the methylphosphonate ionic liquids LMP101 and LMP102 could be related to a lower reactivity due to the longer alkyl chain on the phosphonate anion of LEP102. The presence of intermetallic aluminum-rich precipitates in the microstructure of AZ31B can be observed in Figure 6a as white rounded precipitates and, more clearly, in the EDX element map in Figure 7. The higher adherence and lower thickness of the more continuous coating obtained form LEP102, gives rise to a lower roughness increase with respect to the base alloy than that observed for the rest of ionic liquids (Table 1).  Figure 8 shows the variation of the friction coefficients with increasing normal load under scratching for the four materials, the uncoated AZ31B alloy and the AZ31B coated with the layers obtained from each of the ionic liquids. There is a good agreement between roughness values (Table 1) and friction coefficients, with a lower resistance to sliding for the more polished surfaces.

Abrasion Resistance of the New Coatings
The lowest friction coefficient is obtained for the uncoated alloy, with the most polished surface, while the highest friction coefficient is obtained for the material with the highest average roughness, covered with the layer generated from LMP102. Figure 9 compares the evolution of penetration depth values with increasing normal load for the uncoated alloy and the alloy covered with the surface layers from the three ionic liquids. AZ31B alloy shows an approximately linear increase in penetration depth up to 10N. At that point, a transition to more severe abrasion wear takes place, as observed by the slope increase in the penetration depth curve. Finally, for loads higher than 20 N, an asymptote is reached. Under low loads, up to 10 N, the layer generated from LMP101 shows the lowest abrasion resistance, while the behavior of the layer from LMP102 is similar to that of the uncoated alloy. For normal loads higher than 10 N, all coatings show a better abrasion resistance than the base alloy, in the order: LEP102 > LMP101 > LMP102 > AZ31B. Under the maximum load of 30 N the penetration depth reductions with respect to AZ31B are of a 27.6% for LMP101, a 11.5% for LMP102 and a 44.5% for LEP102. It is important to notice that only the layer from LEP102 improves abrasion resistance of AZ31B under the whole range of applied loads. Figure 10a-e shows SEM micrographs of one of the wear tracks on the material covered with the layer formed with LMP102. Figure 10a shows the severe surface damage at the tip of the wear track, under the highest normal applied load of 30 N. The coating has been eliminated, not only along the wear path, but also on extensive areas outside the wear track. Figure 10b shows the presence of wear debris particles from the surface coating. These are large platelets with a mean thickness of 6-7 μm. Figure 10c,d shows the increasing severity of the fracture and removal of the coating as normal load increases. This mechanism explains the high friction coefficient values (Figure 8). Finally, Figure 10e shows the fracture under the initial application of load at the start of the wear path, with the corresponding decreases of the size of coating grains. A similar wear mechanism is observed in the case of LMP101. These observations are in contrast with the thinner layer obtained from LEP102. A possible explanation for this could be the higher reactivity of the ionic liquids with shorter alkyl substituents.   (Figure 11c). In this case, no fracture or generation of wear debris particles form the coating layer is observed. A smother wear track surface is obtained, where the surface layer penetrates into the magnesium alloy substrate due to the applied load.
The phosphorus and magnesium element maps in Figure 12 show that, up to 10 N, the whole or large areas of the wear track remain covered by the phosphorus containing layer. From that load on, although large areas of magnesium from the substrate are present at the surface, coating particles remain within the wear track, even under the highest applied load at the end of the track. This mechanism is similar to that observed for the coating obtained at room temperature [66] and is in sharp contrast with the coating failure observed for the methylphosphonates LMP101 and LMP102 (Figure 9). This could be attributed to the better adherence of the thinner layer generated from the ethylphosphonate LEP102.

Conclusions
Alkylphosphonate ionic liquids react with the AZ31B magnesium alloy to form phosphorus-containing surface layers able to reduce the surface damage due to abrasion. At 50 °C, the reaction time is reduced from 14 days, at room temperature, to 48 hours.
The new coatings present higher friction coefficients due to the increase in surface roughness with respect to the base alloy. The highest surface roughness is obtained for the coatings generated from methylphosphonate ionic liquids.
In the case of the methylphosphonate ionic liquids, the coating layers obtained at 50 °C are 6-7 μm thick and present a very poor adherence to the substrate alloy. This is the main cause for the delamination and failure of the coating under abrasion. In this way, the surface layers obtained from methylphosphonate ionic liquids only reduce the penetration depth under loads higher than 10 N, after the base alloy has experienced a transition to more severe wear.
In contrast, the ethylphosphonate ionic liquid is less reactive, giving rise to a very thin coating, with a lower proportion of phosphorus, which shows a better adherence and absence of delamination under load. This thinner coating shows the highest abrasion resistance and reduces penetration depth in a 44.5% with respect to the uncoated alloy.
In order to reduce reaction times, electrochemical processes are currently being studied to develop protective surface coatings from phosphonate ionic liquids.