Effect of Different Types of Glass Powders on the Corrosion and Wear Resistance of Peo Coatings Produced on 6061 Aluminum Alloy

Prior to the PEO treatment, samples 4 × 1 cm were cut from 6061 AA plate (nominal composition 1% Mg; 0.0.6% Si; 0.25% Cu; 0.1% Mn; 0.1% Fe; 0,15% Cr; Al bal.)and then polished using conventional metallographic procedures that included a grinding step with abrasive papers (580, 800, 1200, and 4000 grit) and a finishing step with cloths and diamond suspensions (6 μm and 1 μm). Following the polishing, samples were ultrasonically degreased in ethanol for 5 min. An electrolyte bath, carbon-steel mesh as well as cooling and stirring systems were included in the PEO equipment. During the PEO process, a continuous stirring of the electrolyte was applied to prevent concentration and temperature gradients. Aluminium substrate was used as anode, while a carbon-steel mesh was used as cathode. TDK-Lambda 400 V/8 A DC power supply was used for coating production, working in direct (D) and pulsed (P) current modes under a constant current density of 0.65 A/\({\text{cm}}^{2}\), 200 Hz frequency and 50% duty cycle for 3 min. In order to perform the PEO process, a base electrolyte (B) and three types of glass were used LCD, BS and SL glass, as indicated in Table 1. High-purity reactants; sodium metasilicate (\({\text{Na}}_{2}{\text{SiO}}_{3}\)), and sodium hydroxide (\(\text{NaOH}\)) were used as base electrolyte. The chemical composition of glass additives is shown in Table 2. The LCD (alkali-free glass), utilized in liquid crystal and OLED displays, came in sizeable sheets from NEG (Nippon Electric Glass Co., Ltd., Ōtsu, Japan). BS originated from a glass cullet (Kimble/Kontes, Vineland, NJ) and was primarily designed for pharmaceutical use. SL was derived from crushed glass containers, furnished in fine powder form by SASIL SpA (Biella, Italy). Notably, all these glasses were sourced from recycled glass materials. Before use, all the employed glasses were reduced into fine powder by dry ball milling (Mixer Mill MM 301, Retsch GmbH) in order to obtain glass particles with coparable size distribution for all the analyzed glasses. In detail the dimension of the obtained glass particles after the ball milling was of micrometric size, varying between 1 and 10 µm. The choiche of the dimensions of the particles was perfoemd in order to assure incorporation into the coating. Further charcterization and studies on all three types of glass can be found in other publications [35, 36, 42].

A Cambridge Stereoscan 440 scanning electron microscope (SEM), equipped with a Philips PV9800 energy dispersive X-ray spectrometer (EDS) was utilized to analyze the surface and cross-section of samples. When analyzing the cross-section, samples were cut, mounted in epoxy resin, and carefully ground and polished following standard metallographic preparation. The composition analysis was performed on the surface of the samples using EDS. Samples were sputtered with a gold film prior to the SEM analysis. All image analysis was performed using the ImageJ software package [43].

Phase analysis of coatings was studied by a Bruker® X-ray diffractometer (D8 Advance, Karlsruhe, Germany), operating at 40 kV and 40 mA (2θ range between 20° and 90° with a step size of 0.05 and counting time 5 s). All the phases in the coatings were identified using High Score Plus software.

A stylus profilometer (Hommelwerke T2000, 5 μm tip radius) was used to evaluate surface roughness in terms of Ra and Rq, according to ISO 4287 standard (measuring length Lt = 4 mm and sampling length Lc = 0.8 mm). Scratch tests (Revetest, CSM Instruments) were carried out to characterize the practical adhesion of PEO layers. Progressive scratch tests (10 mm length) were performed with a Rockwell diamond indenter (200 μm tip radius) by applying a load from 1 to 30 N and a 10 mm min⁻¹ linear speed.

Vickers microhardness tests were conducted on polished cross sections to evaluate coatings' mechanical properties. The microhardness was assessed with a Vickers Leitz Wetzlar micro-hardness tester using a normal load of 100 g.

Electrochemical impedance spectroscopy (EIS) measurements were carried out at room temperature (~ 22 °C) in a naturally aerated 3.5 wt% NaCl aqueous solution, according to the literature [44, 45]. Electrolyte 3.5% of sodium chloride was employed from one side to simulate aggressive envirnment containing chlorides that are the ions that create major problems of corrosion on aluminum alloys. Measurements were conducted after allowing the sample to stabilize at the open circuit potential (OCP) for 30 min. A standard three-electrode cell arrangement was used with the sample acting as the working electrode (1 cm²), platinum as the counter electrode, and a saturated calomel reference electrode (SCE); all potentials are given with respect to the SCE. Electrochemical measurements were performed using potentiostat GAMRY Interface 1010E. A sinusoidal perturbation of 10 mV over the frequency range of 100 kHz to 0.1 Hz was applied. Impedance spectra were fitted to an appropriate electrical equivalent circuit (EEC) using Gamry Echem Analyst software. All the EIS tests were performed on the PEO-coated and uncoated specimens in triplicate to ensure repeatability.

Local electrochemical properties were determined by scanning electrochemical microscopy (SECM) in 1 mM NaCl solution at room temperature. Before SECM testing, PEO-coated samples were embedded in epoxy resin. The Biologic M470 electrochemical workstation in dc-SECM mode was used for the measurements, and the scanned area of the sample was 300 × 300 µm with a step size of 10 µm. An ultramicroelectrode (UME) consisting of a 10 µm diameter Pt wire enclosed in glass was used as the working electrode, a Pt plate as the counter electrode, and an Ag/AgCl electrode as the reference electrode. The Pt UME was positioned at a constant height of 20 µm above the sample surface and scanned the area at a scan rate of 50 µm s⁻¹. Measurements were performed after 5 h of OCP stabilization. During the SECM testing, the tip current was recorded at each step over the area under investigation.

Wear behaviour of PEO coatings was investigated by dry sliding wear tests in ambient conditions against a bearing steel. Thanks to their high strength-to-weight, Al alloys are often used for structural light weighting. Beyond the automotive sector, typical industrial applications of Al alloys such as AA6061 regard the packaging industry, where lightweighting of components allows higher productivity. In this field, aluminium parts are often involved in non-conformal line sliding contacts against a steel counterpart [46, 47]. Based on the above a slider-on-cylinder tribometer (block-on-ring contact geometry, ASTM G77) [48] was adopted: PEO-coated stationary blocks (6 × 6 × 70 mm³) slid against a rotating cylinder made of the AISI 52100 bearing steel (40 mm diameter, 62 HRC hardness and surface roughness Ra = 0.1 µm). For all the dry sliding tests, total distance and linear speed were set at 1000 m and 0.3 m s⁻¹, respectively. Normal load ranged from 5 to 30 N, depending on the PEO layer wear resistance. Preliminary tests were conducted at 5 N for all samples, then normal load was increased every other test until the PEO layer was completely worn out. This procedure led to the identification of the load for coating failure for each PEO layer. The same configuration and testing conditions were also adopted for assessing the wear behavior of the uncoated AA6061 alloy that acted as a benchmark. Maximum contact pressures reported in Table 3 were evaluated by Hertzian theory [49] by considering elastic modulus and Poisson’s ratio of the predominant phase in the PEO layer, being SiO₂ for layers obtained with direct current mode and both γ-Al2O3 and γ-AlO(OH) for pulsed ones as identified by XRD analyses.

During wear tests, normal and friction forces, as well as vertical displacement, were continuously monitored by the aid of load cells and a linear variable differential displacement transducer (LVDT). Therefore, friction coefficient and system wear (i.e. wear of both slider and cylinder) were dynamically recorded for the whole test duration. After tests, worn surfaces were characterized by 3D-digital (Hirox KH 7700) and SEM (Zeiss EVO 50, operating in low vacuum mode) microscopy, while maximum wear depth was measured by stylus profilometry on wear scars of sliders and cylinders.

SEM/EDS analysis were performed to determine the surface morphology and chemical composition of the prepared PEO coatings. Figure 1 shows the surfaces of the PEO coatings observed by SEM prepared in direct current mode, and Fig. 2 reports the surfaces of the PEO coatings prepared in pulsed current mode. All the micrographs indicate the presence of the pores and the pancake structure. The obtained microstructure is characteristic of PEO coatings on aluminium alloy [50, 51]. Porosity is formed by the melting of the oxide layer and the ejection of gas bubbles through discharge channels, and the discharge is created preferentially in relatively thin or faulty places, according to the literature [2, 52]. Pores of significantly larger sizes, on the order of 10 μm, are present in the samples prepared in direct current mode compared to those prepared in pulsed current mode, where the pores are 1–2 μm in size. Moreover, the distribution of pores in the PEO samples prepared by the pulsed mode is more homogeneous than in those prepared by the direct mode, which is in accordance with the literature where it is reported that coatings prepared by the direct current mode have limitations in controlling uniformity [53].

EDS analysis was performed on the surface of the PEO coatings, both for samples prepared in direct current mode and for samples prepared in pulsed current mode. The results are reported in Table 4. The main elements of the coatings are Al, coming from the substrate, and Si and Na, coming from the electrolyte solution. Compared to other samples, samples added with LCD glass particles show presence of calcium compunds. As calcium oxide is a component of LCD glass, this is to be expected. A certain concentration of Ca and Mg can also be found in samples where SL glass was added, confirming the presence of glass particles in the PEO coatings. Compared to the base coatings, no additional elements were detected in PEO samples with BS glass, but an increase in the concentration of the Si element can be seen (5.68% for samples prepared by direct current mode and 0.86% for samples prepared by pulsed current mode). In our previous research, Pezzato et al. [19], there was also an increase in the concentration of Si, which was evidence of the incorporation of BS glass into PEO coatings.

The results of cross-section SEM observation are shown in the micrographs in Fig. 3 for PEO coatings prepared by direct current mode and in Fig. 4 for coatings prepared by pulsed mode. Micrographs labeled A represent the base coatings, while B, C, and D show samples containing glass additives, LCD, BS, and SL glass, respectively. The black zone on the right side of the cross-section image indicates epoxy resin used for mounting samples.–The thickness of PEO coatings, measured by image analysis of SEM cross-section micrographs, is reported in Table 5. Direct current coatings are significantly thicker, approximately 100 μm, compared to pulsed coatings, which are approximately 10 μm in thickness. Moreover, glass particles did not affect thickness of the coatings. Instead, they were incorporated into the empty zones of the base coatings in both current regimes, which is consistent with our previous work [19]. The incorporation of glass particles is clearly visible on these micrographs, and it is marked by white circles, whereas the empty zones on the base coating are marked by red arrows. The melting point of glass additives (between 1000 and 1500 °C) is significantly lower than the working temperature of the PEO process, which is in the range of 1800–2370 °C [54]. The lower melting point of the glass additives may explain the inclusion of glass additives in the base coatings so that the glass quickly melts and resolidifies, forming an amorphous phase that partly seals the pores, as stated in our previous work [19] but also in the work of Fattah-alhosseini et al. [55]. They also pointed out that particles having a high melting point, such as CeO₂ (2400 °C), SiC (2730 °C), and Si₃N₄ (1900 °C) were mostly incorporated inertly, disregarding their sizes.

Surface roughness parameters (Ra, Rq) are reported in Table 5. Considering that the base AA6061 alloy was polished before PEO and it was characterised by Ra and Rq equal to 0.4 ± 0.1 µm, all coatings resulted in a significant roughness increase. It is indeed well-known that PEO coatings are characterized by a high surface roughness due to the presence of pores and pancake structures, as also discussed previously. Overall, coatings obtained in pulsed mode were characterised by lower roughness than coatings processed with direct current. The result is supported by the analyses of the coating surface discussed in Figs. 1 and 2, that evidenced pores approximately one order of magnitude smaller in pulsed samples than in direct ones, with the smaller pores of pulsed samples being also more homogeneously distributed than in direct ones.

Generally, the addition of glass particles in the electrolyte induced an increase in the surface roughness. The effect of particle addition is more evident in samples obtained in pulsed mode, apart from BS-P glass that did not increase surface roughness if compared to the PEO layer without glass particles.

The microhardness \({\text{HV}}_{\text{0,05}/30}\) data are reported in Table 5. It is noticeable that the addition of glass particles significantly increases the microhardness of the coatings compared to the base coatings. The effect is more evident in samples with LCD glass particles. Microhardness is about 50% higher than in samples with SL glass and about 20% higher than in coatings with BS glass. This is also in agreement with the mechanical properties of the glasses, with the LCD glass being the one with the highest hardness. Considering this, it can be stated that the glass particles strongly influence the hardness of the obtained coatings.

Scratch tests carried out on PEO layers simulate the penetration of a singular asperity, allowing to assess the value of critical load related to the onset of coating failure (Lc2) and total coating breakthrough (Lc3). Samples obtained with pulsed mode, regardless of the type of electrolyte utilized in the PEO treatment, revealed comparable and very low values of both LC2 (approx. 2 N) and LC3 (approx. 4 N). On the other hand, samples processed with direct current revealed, with the exception of BS glass, quite higher critical loads, ranging from 8 to 25 N for Lc2 and from 12 to over 30 N for Lc3. As hardness is comparable between pulsed and direct samples, this result is most likely related to the lower coating thickness obtained by operating in pulsed mode. While the addition of glass particles did not affect the practical adhesion of the relatively thin pulsed samples, in the case of direct samples LCD glass particles, which also had the most beneficial influence on microhardness (Table 5), glass particles also succeeded in doubling the critical loads of the PEO coatings. On the other hand, the other types of glass additives were not beneficial. A similar ranking among PEO coatings was also evidenced by dry sling wear tests, discussed in the following sections.

The XRD diffraction patterns of PEO-coated samples in the presence and absence of glass particles are illustrated in Fig. 5 for coatings prepared by direct current mode and in Fig. 6 for coatings prepared by pulsed current mode. Strong diffraction peaks of the Al (ICSD 98-015-0692) coming from the substrate were identified in all the samples. XRD spectra show that the phase composition of the PEO coatings varies depending on the current regime used. When the direct current mode is used, mostly SiO₂ (ICSD 98-018-1307) is formed and traces of γ-Al₂O₃ (ICSD 98-017-3014) can be found. Otherwise, SiO₂ is not found in the coatings prepared by the pulsed current mode, but only the diffraction peaks of the aluminum compounds can be seen, γ-Al₂O₃ (ICSD 98-017-3014) and γ-AlO(OH) (ICSD 98-002-7865). This particular behavior between direct and pulsed current mode, when a silicate base solution is used, is in accordance with the fact that in direct current mode more strong discharge phenomena that persist longer over the processed surface occur in comparison to pulsed current mode. Due to this fact the electrolyte remain trapped into the discherge channel for longer time in the direct current mode, thus permitting the formation of higher amount of silcon oxide wich derives from the electrolyte. Diffraction peaks of glass additives were not found in the XRD spectra, suggesting that are located in the amorphous part of the spectrum, as already reported in the literature [19]. Comparing the results of the XRD analysis with the results of the EDS (Table 4) permit clearly to evidence the absence in the XRD spectra of Ca and Mg compounds, that come only from the glass, and are probably present in the coating only as an amorphous phase. Also, Na compounds are not present in the XRD spectra. Na compounds come both from the glass and from the electrolyte, the part from the glass is present probably as amorphous phase whereas the part coming from the electrolyte is below the detection limit of the XRD.

The corrosion resistance of PEO coatings was studied using EIS. Comparison between samples with and without glass additives in PEO coatings was made during two different current modes of PEO, pulsed and direct. In many cases, this electrochemical technique has been used to analyze the corrosion resistance of PEO coatings in a 3.5% NaCl electrolyte which simulates a corrosive chloride environment [56]. The experimental results are shown in the form of Nyquist plots, Fig. 7 and Bode plots, Fig. 8. The Bode plot representation is presented in the form of Bode Modulus plots and Bode Phase plots. In the figures, indication (a) refers to samples prepared by direct mode, while indication (b) refers to samples prepared by pulsed mode.

To determine quantitative electrochemical parameters of the phenomenon taking place during exposure to the electrolyte, EIS data are commonly analysed by fitting it to an EEC [57]. In EEC, different electrical elements are used to represent the charge's path through the system components.

The EIS data were fitted using three different EEC models, all shown in Fig. 9. In the EECs, a constant phase element (CPE) instead of a capacitance was used, since often the coating does not behave as a pure capacitance because the electrical properties of the material are not constant through the thickness. Then, the measured phase angle at high frequency is not equal to − 90°, which can be seen in the EIS results shown in the form of Bode Phase plots in Fig. 8, where it is noticeable that none of the curves has a value of -90° that would indicate an ideal capacitor [58]. The impedance of a CPE can be expressed by the following equation:j is the imaginary unit, Y is the constant phase element, a is an empirical exponent, and ω is the angular frequency of the sine wave defined as:where f is the frequency. The empirical exponent a represents the physical meaning of the CPE and can assume the value of 1 in the case of a perfect capacitor, and 0 in the case of a perfect resistor. Deviations of a from these values indicates the non-ideality of the system. If the value of the exponent a is approximately 1, Y is similar to a pure capacitor and can be considered:where C is the capacitance, \({\varepsilon }_{0}\) is the permittivity of vacuum, ε is the dielectric constant, A is the effective area, and d is the thickness of the layer [59] [NO_PRINTED_FORM] [60].

Accordingly, the aluminium substrate is fitted with a simple EEC (Fig. 9a), consisting of one resistive element (\({R}_{o}\)) added in parallel to one capacitive element (\({CPE}_{o}\)) representing an oxide layer on the bare aluminium alloy.

PEO coatings usually consist of two layers, an outer porous layer and an inner barrier layer [61]. The Phase Bode plot for the PEO-coated samples with and without glass additives, Fig. 8, demonstrates two-time constants which represent the resistance offered by the outer porous layer and inner barrier layer, respectively. The high-frequency semi-circle demonstrates the outer layer properties, whereas the low-frequency semi-circle demonstrates the inner barrier layer properties [62]. On the other hand, bare aluminium alloy has only one time constant, indicating that corrosion protection is given only by the oxide layer.

Therefore, the EECs that can best fit the experimental data, shown in Fig. 9b, contain \({R}_{p}\) and \({CPE}_{p}\) related to the external porous layer and \({R}_{b}\) and \({CPE}_{b}\) that represent the polarisation resistance and constant phase element of the inner barrier layer, respectively. R_s represents the resistance of the electrolyte, and it is presented in all the EEC. The value of R_s depends primarily on the electrochemical cell's geometry and the test solution's conductivity. The values obtained from the fitted spectra were consistently small, between 4.1 and 32.8 Ω cm² (see Tables 6 and 7).

Different behaviour of the impedance curves at the low-frequency regions was observed, in the sample without the glass additives prepared by Pulsed DC mode. Therefore this sample is fitted with the EEC shown in Fig. 9c. The EEC is the same as the one before, with an additional Warburg element added in series with \({R}_{b}\). This means that the corrosion process is mostly diffusion controlled, as evidenced by the increase of the Nyquist curve at low frequencies, Fig. 7b. That behaviour is caused by the empty pores in PEO coating without the glass additives, which may allow corrosive electrolytes to diffuse into them, and it is in good agreement with the SEM cross-section micrograph (Fig. 4a).

Results of the fitting of the experimental data from the EIS test with the EECs are reported in Table 6 for the samples prepared by direct current mode and Table 7 for the samples prepared by pulsed mode. The fitting results of the bare aluminium substrate are shown in both tables for comparison.

The quality of the fittings was evaluated based on the chi-squared values and the strong agreement between experimental plots, in form of circles, and fitting plots in form of lines. Chi-squared values is the square of the standard deviation between the original data and the calculated spectrum. The lower the values, the better the fitting [60]. The values were in the range 4.9 × 10⁻³ – 6.0 × 10⁻⁵ indicating good agreement between the experimental data and equivalent circuit fits (see Tables 6 and 7).

Based on the data in the Tables 6 and 7, it is apparent that the corrosion resistance is improved with glass particles added to the PEO samples, that are showing the highest \({R}_{p}\) and \({R}_{b}\) values. Results are consistent with the literature, which suggests that pore sealing in PEO coatings improves corrosion resistance [63].

Generally, the internal barrier layer, in all PEO-treated samples, provides corrosion resistance, due to its higher value of polarization resistance than the outer porous layer, which is in accordance with literature data for PEO coatings [62]. Additionally, a lower value of Y indicates more compact layer [62, 64], and all PEO coatings with glass additives have lower values of Y compared to coatings without glass additives. A lower value of Y is present in both layers of the PEO coating, the barrier and the porous one. This result agreed well with the SEM observation of cross-sections, Figs. 3 and 4, where pore sealing is clearly visible in the samples with glass additives.

In comparison with the other coatings, LCD-P stands out as the sample with the highest corrosion resistance, with an \({R}_{b}\) value of 1.2 × \({10}^{7}\) Ω \({\text{cm}}^{2}\), (two order of magnitude, 9900% of improvement in comparison with the untreated sample) which is one order of magnitude higher than the next best coating.

Moreover, pulsed samples with glass particles possess \({R}_{b}\) values ranging from 3.4 × \({10}^{6}\) Ω \({\text{cm}}^{2}\) for the BS-P sample (2733% of improvement in comparison to the untreated AA) to 1.2 × \({10}^{7}\) Ω \({\text{cm}}^{2}\) for the LCD-P sample, up to two orders of magnitude higher than those prepared by direct mode, whose values range from 6.1 × \({10}^{5}\) Ω \({\text{cm}}^{2}\) for the SL-D sample (408% of improvement in comparison to the untreated) to 1.7 × \({10}^{6}\) Ω \({\text{cm}}^{2}\) for the BS-D sample (1316% of improvement in comparison to the untread). Compared to direct, pulsed samples have a higher corrosion resistance due to their more homogenous PEO coating structures and smaller pores [28]. In this way, the pores are more likely to be filled with glass particles, reducing the possibility of free spaces within the coating.

Summarizing, considering the corrosion properties of the coated samples and comparing the value of R_b of the samples with PEO coatings to the value of R_o of the uncoated sample the following percentage of improvement of the corrosion resistance can be calculated: B-D 250%; LCD-D 1233%; BS-D 1316%; SL-D 408%; B-P -51%; LCD-P 9900%; BS-P 2733%; SL-P 5983%. Considering these results resulted clear that all the coatings containing glass particles produce remarkable improvement in the corrosion properties and that generally the layers produced in pulsed mode behave better that the ones obtained in DC mode. Considering the different types of glasses LCD glass seem the most promising in improving the corrosion performances of PEO coatings.

The distinct corrosion behaviors observed in PEO coatings with various glass additives align with the concentration of alkaline elements within the glass's chemical composition. Best outcomes were observed in LCD samples, characterized by alkali-free glass, while results with BS and SL glass varied based on the current regime of the PEO process. In corrosive conditions, alkaline elements have the potential to leach out of the glass structure, creating vacancies subsequently filled by water molecules. The principle of preferential leaching relies on the thermodynamic and kinetic stability of distinct glass modifiers. At lower temperatures and for ions with the same charge, the diffusion of larger ions becomes energetically unfavorable, while smaller ions can move more easily through the glass network. Typically, double-charged ions (alkaline earth elements) tend to exhibit lower diffusivity within the glass network compared to single-charged ions (alkaline elements). This difference primarily arises due to the pronounced influence of highly intense local electric fields acting on the double-charged ions.

Moreover, the substitution of high-radius cations like \({\text{K}}^{+}\) from the bulk creates larger voids in the glass network compared to the substitution of smaller cations like \({\text{Na}}^{+}\), thereby facilitating the penetration of water molecules into deeper regions. While all alkaline silicate glass are prone to degradation, their stability, from a thermodynamic perspective, tends to increase in the following order: \({\text{K}}_{2}{\text{SiO}}_{3}\) < \({\text{Na}}_{2}{\text{SiO}}_{3}\) < \({\text{Li}}_{2}{\text{SiO}}_{3}\)  [40]. The high corrosion resistance obtained in the sample with the addition of LCD glass can also be related to the fact that this glass is a “shorter” glass in comparison to BS and SL glasses thank to the absence of alkali. Due to this fact, during the PEO treatment, the LCD glass can soften more than the other thus better filling the pores and producing a higher barrier effect against the external environment.

The most promising samples according to the EIS results, LCD-D and LCD-P, were further analysed by the SECM method in 1 mM NaCl. When comparing the current maps (Fig. 10) recorded over the surface of the measured samples, differences can be seen. For the LCD-D sample (Fig. 10a), the current distribution is more or less uniform over the entire surface, and only positive currents related to the dissolution of the coating were detected. The situation is somewhat different for sample LCD-P, where the maximum positive current is higher than for LCD-D. Although higher current is indicating more intense dissolution at these sites, it should be noted that this maximum is strictly localised and the rest of the sample shows positive current values comparable to those of the direct curret counterpart. On the other hand, negative current values were observed in the LCD-P sample, indicating the presence of cathodes on the examined area. This could be directly attributed to the detection of base aluminium alloys due to the lower thickness of the LCD-P coating (11 vs. 70 µm). Localised electrochemical investigation of the samples showed a different behaviour than the EIS investigation, which can be attributed to the fact that the samples were exposed to the aqueous solution for a longer time and solution was able to penetrate the base material through the thinner LCD-P coating.

As regards block-on-ring dry sliding wear tests, samples were tested at increasing normal load up to coating failure, occurred at different loads as reported in Table 5. Pulsed layers, whose thickness was significantly lower than direct ones, survived at a maximum of 10 N load while direct samples were tested up to 30 N load.

Results of wear tests, in terms of steady-state coefficient of friction (COF) and maximum wear depth measured on sliders, are reported in Figs. 11 and 12. Wear depth of the steel counterpart was also taken into consideration but, regardless the applied normal load or the type of PEO layer, it was always non-detectable. Steady-state COF was evaluated, starting from dynamic acquisition during test, as average value after the run-in distance, set at 300 m. It is worth recalling that COF is the result of two different contributions: adhesive and abrasive. PEO layers, by reason of the increase in both hardness and roughness, usually raise the abrasive contribution and, therefore, by comparison to the untreated alloy, they induce an increase in the COF. However, by focusing on Fig. 11, in the present study this condition occurs only at high normal loads for PEO direct layers. Otherwise, by considering also standard deviations, COF of untreated alloy was higher or at least comparable to the PEO-treated one. By referring to the analyses of the wear scars in Figs. 13, 14 and 15, discussed in the following, this instance can be explained considering: (1) the severe adhesion wear experienced by the untreated alloy that enhanced the adhesion component of COF; (2) instability at the asperities of the PEO due to microcracking phenomena and the establishment of a protective oxidative tribo-layer between contact surfaces that limited the rise of the abrasive component of COF. Generally, the addition of glass particles induced an increase of the COF of the PEO layers, because of the increase of both surface roughness and hardness (see Table 5), with the only exception of LCD-P samples. On the other hand, all PEO layers induced a significant increase in the wear resistance, by reducing up to one order of magnitude the wear depth if compared to the untreated alloy. Among the glass particles used as additives, LCD ones maximized wear resistance of both pulsed and direct samples. By contrast, the addition of SL and BS glass in the PEO layer did not exert a beneficial effect in the wear resistance. As concerns load for coating failure (see Table 5), LCD samples showed the greatest one in both direct and pulsed configuration. In particular, in case of direct samples, LCD-D was the only layer to survive up to a 30 N load, also improving the wear resistance of PEO layers without additives (B-D). For pulsed samples, instead, LCD-P survived up to 10 N load with a wear resistance comparable to the base PEO (B-P samples). The improved tribological properties obtained with the addition of the LCD glass can be on the one hand related to the best sealing of the pores obtained with this glass (due to the fact that it is “shorter” in comparison to the other glasses, as previously discussed), and on the other hand to the higher mechanical properties (Table 5) conferred by this glass to the PEO coating.

Figures 13, 14 and 15 compares wear scars morphology of samples after 5 N tests. From Fig. 13 an overall view of wear scars is proposed for all tested samples, observed with a 3D digital optical microscope, while high-magnification detailed SEM analyses are reported in Figs. 14 and 15 for a representative set of samples, being: untreated alloy, base PEO, PEO with additive exhibiting the best and the worst tribological performance. For SEM analyses, back scattered electrons (BSE) images are proposed for the characterization at the edges of wear scars, to underline the formation and distribution of oxide-based layers, while secondary electrons (SE) images, enhancing morphological features, are proposed for the center of wear scars. Wear scars morphology in Fig. 13 show that the untreated alloy suffered severe adhesion phenomena, as shown also by the images in Fig. 14a, b. Plastic deformation dominates wear scars of untreated alloy, characterized by the presence of long and continuous grooves due to ploughing (highlighted by dashed green circles in Fig. 14b). On the other hand, scars of PEO layers who survived the tests are characterized by a tribolayer that appears reddish in the micrographs of Fig. 13 and bright grey in Figs. 14 and 15 (BSE images). The tribolayer is the results of the tribological contact with the steel counter face and consist of Fe-based oxides, as already observed in a previous work [65] and confirmed by EDS analyses. Such oxide-based tribolayer has a protective function, therefore it leads to a mild wear regime, as also demonstrated by the low wear depth measured on sliders (Fig. 12). However, it is worth recalling that wear depth of the steel counterpart was always non detectable. SEM analyses in Figs. 14 and 15 evidenced that the protective oxide layer was thicker and more continuous in case of LCD samples, obtained with both direct and pulsed configuration (Figs. 14e and 15c), supporting the negligible wear depth discussed in Fig. 12. The high magnification analyses of scar center also showed that LCD-P suffered of a minor micro-cracking phenomenon (highlighted by orange arrows in Fig. 15d), while no micro cracks are evidenced in LCD-D. Conversely, protective layers formed on the other samples were less dense and compact, especially in case of BS ones (Figs. 14g and 15e), and the detailed analyses of scars morphology evidenced the presence of several discontinuities in the protective tribolayer, thus justifying the worse wear resistance of B and BS samples than LCD ones. Such discontinuities can be synthetized as a more marked microcracking (underlined by orange arrows in Fig. 15f), eventually leading to fragmentation of the oxide layer, evidenced by the large and compacted fragments (indicated by yellow arrows in Fig. 14d, h), and the formation of debris retained in the worn surface (highlighted by the dashed line orange circle in Fig. 15b).