Final surface modification for better wear resistance of ceramic coating on cast AlSi10Mg alloy

Ceramics International 48 (2022) 37433–37447 Available online 21 September 2022 0272-8842/© 2022 The Author(s). Published by Elsevier Ltd. This is an open access article under the CC BY license (http://creativecommons.org/licenses/by/4.0/). Final surface modification for better wear resistance of ceramic coating on cast AlSi10Mg alloy Roman Gabor a , * , Tom´ aˇ s Prymus a , Ladislav Cvrˇ cek b , V´ aclav Nehasil c , Josef Hlinka d , e , Matˇ ej Buˇ ril b , Michaela Tokarˇ cíkov´ a a , Jana Seidlerov´ a a , f a Nanotechnology Centre, CEET, VSB – Technical University of Ostrava, 17. Listopadu 15/2172, 708 00, Ostrava-Poruba, Czech Republic b Department of Materials Engineering, Faculty of Mechanical Engineering, Czech Technical University in Prague, Karlovo N´ amˇ estí 293/13, 120 00, Prague 2, Czech Republic c Department of Surface and Plasma Science, Charles University, Prague, Czech Republic d Department of Materials Engineering and Recycling, Faculty of Materials and Technology, Vˇ SB-Technical University of Ostrava, 17. Listopadu 2172/15, 708 00, Ostrava-Poruba, Czech Republic e Centre for Advanced Innovation Technologies, VSB-Technical University of Ostrava, 17. Listopadu 2172/15, 708 00, Ostrava-Poruba, Czech Republic f Department of Chemistry and Physico-Chemical Processes, Faculty of Materials and Technology, Vˇ SB-Technical University of Ostrava, 17. Listopadu 2172/15, 708 00, Ostrava-Poruba, Czech Republic ARTICLE INFO Keywords: Micro-arc oxidation Design of experiment Coating Wear Al–Si alloys ABSTRACT Using the design of experiment (DOE) method and the micro-arc oxidation (MAO) technique, ceramic layers on AlSi10Mg alloy were systematically prepared to design optimal process conditions for achieving the best tribological properties of the ceramic layers. The lowest concentrations of the applied 6 g/l NaOH and 12 g/l Na 2 SiO 3 resulted in the preparation of uniform MAO layers with the lowest rated parameters R a , R z and thickness achieved under micro-arc discharge conditions at 500 V and 60 min. With the increasing thickness of the coatings, there was an increase of Si in the MAO coating. Full factorial DOE was used to optimize the tribological properties in a polyalphaolefin (PAO) environment at 80 ◦C. The most significant influence for the preparation of abrasion-resistant layers for the investigated factors was identified on the AlSi10Mg alloy by the NaOH content in the electrolyte. The friction coefficients of MAO coatings reached an average value of 0.15. Aero-lap polishing technology was applied for increased wear resistance requirements to eliminate the deficiencies of MAO coatings, leading to decrease wear track by almost double compared to polished silumin. Removal of the outer MAO layer by polishing led to a reduction in the high corrosion resistance of the MAO coating, demonstrating the influence of the outer layer not only on the tribological properties but also on the corrosion resistance of MAO coatings. 1. Introduction Al–Si alloys are one of the most important aluminium-based foundry alloys, commonly used in automotive and aerospace industries due to their excellent properties such as castability, strength-to-weight ratio and good thermal and electrical conductivity. As a result, Al–Si alloys are finding more applications in the mentioned sectors, particularly in the manufacturing of cylinders, pistons, engine blocks and brake callipers [1–3]. Due to the requirements of the producers of individual products for the long-term durability of castings, the application of Al–Si alloys is significantly limited due to low corrosion resistance and surface hardness [4]. In order to improve the abrasion and corrosion resistance of Al–Si alloys, their surfaces are modified using anodic oxidation of the substrate [5–7] or the relatively newly developed method of thermal spraying [8,9]. The widespread implementation of convex hard anodization for surface treatment of AlSi10Mg alloys is limited by the microstructure of the alloy, which is formed by α (Al) dendrites in addition to the primary α (Al) dendrites and α (Al)+Si eutectic [10]. The resulting oxide layers prepared by hard anodization do not have sufficient abrasion and corrosion resistance. Therefore the possibility of using the relatively new technique of micro-arc oxidation (MAO) arises for the surface treatment of Al–Si alloys which has been verified in several works [11–13]. This * Corresponding author. E-mail address: [email protected] (R. Gabor). Contents lists available at ScienceDirect Ceramics International journal homepage: www.elsevier.com/locate/ceramint https://doi.org/10.1016/j.ceramint.2022.09.224 Received 21 July 2022; Received in revised form 16 September 2022; Accepted 18 September 2022 Ceramics International 48 (2022) 37433–37447 37434 technique is used to prepare a ceramic adhesion coating with a compact corrosion-resistant and abrasion-resistant inner layer and an outer porous layer under elevated stress conditions to produce a dielectric discharge [14]. Key parameters for the preparation of MAO layers include the choice of electrolyte, which significantly affects the growth mechanism of the oxide layer and its corrosion and mechanical properties [15]. The most commonly used electrolytes for the preparation of the oxide layer in Al–Si alloys include silicon-containing electrolytes [16–18]. However, the presence of eutectic and silicate ions in the electrolyte influences the resulting topography of the oxide layer due to the emergence of “craters” on the surface of the coating representing rapidly solidified melt of the oxidized substrate flowing through discharge channels to the surface [19]. This leads to deterioration of the tribological properties and an increase of the coefficient of friction leading to a significant limitation in the use of MAO coatings for Al–Si alloys and their application in the field of lubricated engine components. As a result, increased attention has been paid to the development of self-lubricating MAO coatings based on Al 2 O 3 /PTFE composite coating [20], multiphase MAO coating containing graphite [21], Si 3 N 4 /TiO 2 nanocomposite coating [22], or synthesized MoS 2 layer on the surface of MAO coating [23]. However, an optimal, economically feasible procedure is still being sought to enable the efficient preparation of MAO coating, meeting the requirements for broader Al–Si applications in terms of tribological and corrosion resistance properties. To this end, statistical analysis methods such as the systematic design of experiment (DOE) approach are used to optimize and describe the individual input factors’ influence [24]. This article presents a procedure for the preparation of abrasion and corrosion-resistant MAO coating using DOE. Main results of this work are (1) determination of the optimum process conditions of the MAO process including the importance of the main factors influencing the tribological behaviour of the coating in polyalphaolefin (PAO) at 80 ◦C, (2) description of the application procedures of Aero-lap polishing to improve the tribological properties of MAO coatings, (3) evaluation of the tribological and corrosion properties of the outer and inner layer of MAO coatings. The results indicate the potential application of the above-mentioned technique in the field of lubricated motor components made of Al–Si alloy requiring high wear and corrosion resistance. 2. Experimental 2.1. Materials and sample preparation Surfaces of AlSi10Mg foundry alloy samples (20 mm (L)*10 mm (W)*5 mm (H)) were treated by tumbling in HV 20 tube vibrator (OTEC, Germany) to homogenise them. Steel satellites (5 ×3 mm) were used as process media for 24 h, followed by 72 h of lapping in plastic grinding chips M 10. The lapping procedures were carried out under wet conditions. Finally, the alloy was digested in an acid mixture in a Milestone Ethos Up microwave digestion plant (Italy). Selected elements (Table 1) were determined by atomic emission spectroscopy with inductively coupled plasma (AES-ICP, Spectro Arcoss, Germany). 2.2. Surface modification process The MAO coating preparation was performed according to the combination based on the DOE on the semi-operational unit shown in Fig. 1. The switching power supply (DEHOR-spec. Litvínov s.r.o., Czech Republic) with a pulse unit was controlled by a constant voltage of 500 V for 60 min. The MAO process was carried out with the sample as anode and a stainless steel sheet (1.4301) of 2 mm thickness serving as a cathode. The electrolyte composition was defined by the designed DOE with temperature controlled up to 25 ◦C during the MAO process. N. 1) Mixed degreasing bath (1 M NaOH; 45 ◦C); N. 2,3) Mixed rinsing bath (distilled water; conductivity <10 μ S/cm); N. 4) Pickling bath (20 wt % HNO 3 +2 wt % HF); N. 5,6) Mixed rinsing bath (distilled water; conductivity <10 μ S/cm); N. 7) Mixed rinsing electrolyte bath with counter - electrode (electrolyte; pH ≥12). N. 8,9) Mixed rinsing bath (distilled water; conductivity <10 μ S/cm). To improve the tribological properties, a lapping technique was applied to the optimal DOE experiment using an aero-lap polishing machine (AERO LAP YT300-OE, Japan). Verification of the tribological properties improvement was carried out at 1, 2 and 3 min time, working pressure 0.7 MPa, working distance 100 mm, and rubber granules (Multicon) of size 0.5–2 mm with diamond paste as process media. 2.3. Design of experiment (DOE) The design of experiments, including the selection of factors and their limit values, was based on the empirical experience of the authors. Table 1 Chemical composition AlSi10Mg. Sample Element (wt. %) Si Fe Cu Mn Mg Ni Zn Pb Ti AlSi10Mg 10.0 0.33 0.21 0.03 0.26 0.01 0.07 0.01 0.08 Fig. 1. MAO process. R. Gabor et al. Ceramics International 48 (2022) 37433–37447 37435 Selected three main factors were tested for their influence on the resulting tribological properties in two levels (2 3 factorial design). The selected limit values of each factor (see Table 2) were applied with repetition without central points. The full factorial design, including the combination of the levels of each factor, is shown in Table 3. 2.4. Surface characterisation Surface and cross-section images of the samples were taken using a JEOL JSM-7610F Plus scanning electron microscope (JEOL, Japan) in secondary (SE) and backscattered electron (BSE) mode. The chemical composition of MAO coatings and elemental mapping were studied using a dispersive X-ray spectrometer (EDS, ULTIM MAX 65 mm 2 , Oxford Instruments, England) attached to the SEM. The surface morphology of the samples was evaluated further using correlative analysis providing data from AFM LiteScopeTM (Nenovision s.r.o., Czech Republic) and SEM allows 3D-CPEM view (Correlative Probe and Electron Microscopy), optical 3D IF-SensorR25 instrument (Bruker Alicona, Austria), profilometer Talysurf 50 (Taylor Hobson, England). Analysis of elements and their chemical states was performed by Xray photoelectron spectroscopy (XPS) in a UHV chamber with a base pressure of around 3‧10 −7 Pa. The measurement system consists of an Omicron EA 125 multi-channel analyser and a dual anode (Mg/Al) X-ray source. The primary energy source was the Al K α line (1486.6 eV). The measurement system was calibrated against the binding energy of C1s (284.5 eV). Table 2 Levels main factors. Parameters Factor Low level (−1) High level (+1) A NaOH (g/l) 6 12 B Na 2 SiO 3 (g/l) 9 18 C Frequency (Hz) 95 130 Table 3 Full factorial design. Test NaOH (g/l) Na 2 SiO 3 (g/l) Frequency (Hz) 1 6 12 130 2 9 12 130 3 6 18 130 4 9 18 130 5 6 12 95 6 9 12 95 7 6 18 95 8 9 18 95 Fig. 2. Optical images with 3D-CPEM. R. Gabor et al. Ceramics International 48 (2022) 37433–37447 37436 2.5. Tribological tests A ball-on-disk CSM THT tribometer (CSM Instruments, Switzerland) was used to determine the tribological properties. The friction pair consisted of the MAO coating on a planar sample and an Al 2 O 3 ball 6 mm in diameter. Polyalphaolefin (PAO) liquid was chosen as the test environment. PAO is a non-polar synthetic hydrocarbon liquid that is the main synthetic base component of most oils used in industrial and automotive lubricants. PAO provide superior lubrication performance over a wider operating temperature range than petroleum oils and is less volatile. Specifically, PAO Labovac 14 (Welch, Germany) was chosen as a suitable reference for tribological measurements. The manufacturer states a viscosity index of 29.0 mm 2 /s at 40 ◦C and 5.6 mm 2 /s at 100 ◦C. Since the viscosity index decreases significantly with increasing temperature, tribological tests were performed at a temperature of 80 ◦C to ideally simulate typical engine operating temperatures. Measurements were performed twice at a normal load of 2 N, a number of laps 10000, linear sliding speed 50 mm/s and a radius of 4 mm. The friction coefficient ( μ ) was calculated from the ratio of the tangential friction force and the normal force. The width of the wear track was evaluated using a digital microscope Olympus DSX1000 (Olympus Corporation, Japan). Fig. 3. SEM images of MAO coatings. Fig. 4. Surface roughness of MAO coatings: Parameters R a , R z . R. Gabor et al. Ceramics International 48 (2022) 37433–37447 37437 2.6. Adhesion tests The CSM Revetest Xpress +device (CSM Instruments, Switzerland) was used for the scratch test. The scratch tester equipped with a Rockwell diamond indenter (tip radius 200 μ m) was used. The scratch test load was set to linearly increase from 1 N to 50 N along the 5 mm scratch path and linear speed 10 mm/min. 2.7. Corrosion tests Samples were exposed to neutral salt spray using a corrosion chamber SKB 400 A-TR (Gebr. Liebisch GmbH, Germany) according to ˇ CSN EN ISO 9227 for 300 h. The concentration of the sprayed NaCl solution was 50 ±5 g/l. The average pH of the accumulated saline solution was 7.0. The samples were stored in the corrosion chamber at an inclination of 20◦±5◦. The samples were further tested in a three-electrode system (Voltalab PGZ 100, SAS OrigaLys ElectroChem, France). The sample was connected as a working electrode; the calomel electrode was used as a reflection electrode, and the carbon rod as an auxiliary electrode. Potentiodynamic polarisation tests were performed in a 3.5 wt % NaCl solution over a sample area of 0.5 cm 2 . The initial potential of the potentiodynamic measurements was set to −150 mV vs open circuit potential (OCP) after stabilisation of the corrosion equilibrium with a polarisation rate of 5 mV/s. Fig. 5. Cross section of MAO samples. R. Gabor et al. Ceramics International 48 (2022) 37433–37447 37438 2.8. Statistical analyses Data related to surface morphology for each test were compared with each other using One-way ANOVA followed by the Tukey test. The statistical analyses were carried out using Minitab® 17 statistical software. Results were presented as mean ±standard deviation (S.D.). Differences between tests were evaluated at a statistical significance p <0.01. 3. Results and discussion 3.1. Surface topography of MAO coatings Different morphology of the input ground surface of the substrates and oxide layers prepared using different process conditions were observed in Fig. 2 (Table 3) using optical imaging and correlative CPEM analysis. The substrate surfaces of each test (shown in Fig. 3) were analysed with SEM, which confirmed the presence of micro-defects of MAO coatings. The evolution of the coating (following the reaching of the dielectric discharge) is accompanied – besides the structural changes – by the formation of so-called volcanic craters (Fig. 3, Test 4). These structures are formed in the presence of micro-discharge to form melt that flows out through so-called discharge channels towards the surface, where it is rapidly cooled in the electrolyte environment. Areas with a surface microstructure exhibit a greater presence of these craters and directly affect the resulting roughness of MAO surfaces [25]. During the MAO process, gas development occurs; the gas passes through the microchannels and, together with the solidifying melt, forms a porous structure with localised microcracks (Fig. 3, Test 1). The intensity of the internal stresses released by the coating corresponds to the micro-arc discharge conditions with subsequent rapid cooling in the electrolyte environment under the individual DOE tests [26]. Surface roughness was evaluated using the parameters R a (average roughness of surface), R z (average of the absolute values of the heights and depths) and their resulting values were compared for each DOE test using one-way analysis of variance (ANOVA). The results of the relative positions of the observed groups of parameters R a , R z are shown in Fig. 4. The results confirm the statistically significant effect of the DOE parameters on the roughness of the surfaces when the null hypothesis is rejected (p <0.01), where at least one mean of the R a and R z parameters is statistically different. The statistical agreement was demonstrated for the parameters R a , R z for tests 5 and 1. In the case of the R z surface parameter, tests 6 and 3, as well as tests 7 and 6 were found to be statistically indistinguishable. The surface parameter R a was similar for tests 8 and 3. In several studies [27,28], the influence of silicate content on the resulting surface roughness was confirmed. The DOE experiment confirmed the effect of NaOH addition and the presence of silicates on the resulting roughness (Fig. 4). As both the concentration of NaOH, Na 2 SiO 3 and the conductivity of the electrolyte increases, there was a stronger micro-arc discharge observed, more melt transfer towards the surface and an increasing representation of the so-called craters on the sample surfaces, which in turn affected the resulting surface roughness. Using the lowest NaOH, Na 2 SiO 3 contents (Test 1, 5) of the selected DOE combinations, uniform surfaces with the lowest evaluated parameters R a , R z were achieved without any observed effect of the applied source frequency. The resulting quality of the surfaces, in terms of the frequency of microdefects, is crucial for the corrosion and tribological properties of the surfaces. 3.2. Cross section and chemical composition During the MAO process (60 min at 500 V), an increase in the oxide layer (Fig. 5) was observed, accompanied by a decrease in current. Different oxide layer thicknesses were obtained by combining the parameters based on the DOE design. The results presented in Fig. 6 confirm the influence of the chosen parameters on the resulting coating thickness and, thus, lead to the rejection of the null hypothesis (p <0.01) that all diameters are equal. The use of one-way Analysis of Variance (ANOVA) also confirmed that – based on the comparison of the individual tests – the averages of the 5-1, 6-3, and 8-3 tests are statistically indistinguishable (p >0.01). The smallest thicknesses were achieved at the lowest concentrations of NaOH and Na 2 SiO 3 used; the effect of the frequency used was not evident in the case of the smallest layer thickness. On the contrary, the greatest thickness was obtained when the electrolyte had the highest NaOH and Na 2 SiO 3 content and when the frequency used was 130 Hz. These results correspond with the observed roughness parameters and can be explained based on the mechanism of MAO layer formation [29]. As the ionic content of the solution increases, a more intense discharge occurs, accompanied by a linear increase in the layer thickness due to a larger volume of melt passing through the discharge channels depositing on the cooled oxide surface with an increasing proportion of craters and pores. The elemental composition of the MAO coatings shown in Table 4 was studied using EDX at 50x magnification due to the presence of the eutectic α (Al)+Si and the inhomogeneity of the coatings. The increased Si contents correspond to tests with higher Na 2 SiO 3 content in the electrolyte; the elemental distribution of the resulting Al–O–Si system (Fig. 7) confirms the presence of Si in the layer. According to Chao et al. [30], silicon is mainly incorporated in the outer layer of the coating during the MAO process from electrophoresis and diffusion of the present SiO2− 3 from the electrolyte. During the MAO process, several reactions occur at the anode, leading to the formation of a complex Al–Si–O system under plasma discharge conditions (>3500 K) [31–33]. 2OH−−2e−→ H2O+1/2O2(1) Al → Al3++3e−(2) Fig. 6. Thickness MAO coatings. Table 4 Chemical analysis of surfaces by EDX. Test wt. % O Al Si 1 41.87 ±0.34 39.45 ±0.28 18.68 ±0.23 2 39.37 ±0.37 33.08 ±0.26 27.55 ±0.27 3 45.55 ±0.35 17.42 ±0.19 37.03 ±0.28 4 40.90 ±0.37 26.60 ±0.24 32.51 ±0.28 5 39.92 ±0.34 35.14 ±0.25 24.94 ±0.24 6 38.48 ±0.38 39.09 ±0.29 22.43 ±0.25 7 42.29 ±0.38 21.09 ±0.22 36.62 ±0.30 8 38.89 ±0.35 33.22 ±0.25 27.89 ±0.25 R. Gabor et al. Ceramics International 48 (2022) 37433–37447 37439 2Al3++3O2−→ Al2O3+6e−(3) SiO2− 3−4e−→ 2SiO2+O2(4) ySiO2+xAl2O3→ySiO2.xAl2O3(5) The results of the XPS analysis of the chemical states of the crucial elements (C, Al, Si, O) present on the surface of the layers are shown in Fig. 8. All binding energies were calibrated to the main C 1s peak (284.5 eV). The presence of carbon can be explained by the presence of graphitic carbon admixture due to the width of the determined peak. The Si 2p binding energy (103.8 ±0.2 eV) corresponds to the presence of SiO 2 , which is formed during the MAO process from silicates present in the electrolyte (Eq. (4)). The Al 2p line (75.4 eV) in the spectrum indicates the presence of an oxide phase. A shift to higher values of about 1 eV compared to the tabulated value for Al 2 O 3 indicates the probable presence of the aluminosilicate (Eq. (5)) [34]. The O 1s peak (530.6 ±0.1 eV) is in the range of binding energies that correspond to the presence of an oxide phase [35]. 3.3. Tribological properties of MAO coatings in oil The decisive parameter for selecting the most suitable type of MAO layer was the evaluation of its wear. The combination of a hard and brittle MAO layer on a tough silumin base material can be sensitive to the initial point pressure during running-in. Therefore, the tests were performed in PAO oil to make the running-in phase as gentle as possible. At the same time, a small 2 N load, a standard linear sliding speed of 50 mm/s, a radius of 4 mm, and 10000 cycles (corresponds to the track 251 m) were used. Fig. 9a shows a comparison of the friction coefficients of the MAO layers (Test 1–8), where the friction counterpart was an Al 2 O 3 ball. The MAO layers showed relatively high porosity (Fig. 3) and roughness (Fig. 4). In some cases, abrasive particles formed during the running-in phase, which caused an unstable course of friction. The exception was Test 5 layer, which behaved very stably, and its coefficient of friction was also the lowest of all tested MAO layers. However, it was still higher than that of pure silumin, which served as a reference sample. In this case, however, it was a polished surface. Fern´ andezL´ opez et al. [12] reported a similar course of the friction coefficient for silumin with and without MAO treatment with a polished surface. The situation was similar in the case of the wear track width evaluation. It can be seen in Fig. 9b that the Test 5 sample had the smallest track width and slight abrasive wear of the ball (Fig. 10a). Additionally, the thickness of the Test 5 layer was the lowest of all the layers (Fig. 6), and its abrasion should manifest itself much earlier than the other layers. However, it must be noted that the track width was only slightly smaller than that of polished silumin (Fig. 9b), where adhesive friction predominated between the ball and the sample (Fig. 10c). For a precise determination, it would be necessary to compare samples with similar surface roughness. On the contrary, the Test 8 sample showed the greatest wear track, where abrasive friction prevailed between the ball and the sample (Fig. 10b). 3.4. Adhesion of MAO coating The scratch test method was used to compare the adhesion of the MAO coatings. A linearly increasing load was set for testing from 1 N to 50 N. The load was chosen so that the resulting scratch depth was Fig. 7. SEM/EDX mapping images of MAO coatings. R. Gabor et al. Ceramics International 48 (2022) 37433–37447 37440 comparable to the scratch depth in the standard scratch test, where the hard base material and load up to 100 N are used. The tests performed did not show failures in the cohesion or adhesion of the MAO coatings. This is consistent with the growth mechanism of MAO coatings, which are formed by surface oxidation and form a compact system with the base material. The porosity of the MAO coatings varied greatly, as confirmed in Fig. 3. More porous MAO coatings could have formed abrasive particles to a greater extent and increased wear. This assumption was verified in MAO coatings with the lowest and highest wear. Fig. 11 compares the scratches of the samples with the highest wear resistance (Test 5) and the lowest wear resistance (Test 8). There was no damage to the cohesion or adhesion of the MAO coatings around the scratches. Therefore, the degradation of the wear resistance of the Test 8 sample is not caused by adhesion failure, but by the Fig. 8. XPS spectrum of MAO coatings. Fig. 9. Comparison of a) the friction coefficients and b) the wear tracks of the MAO layers Test 1–8. Polished silumin was used as a reference. R. Gabor et al. Ceramics International 48 (2022) 37433–37447 37441 Fig. 10. Comparison of the wear of the Al 2 O 3 ball (left) and the wear track of sample (right) for: a) Test 5 sample, b) Test 8 sample, and c) silumin. Fig. 11. Scratch test of MAO coatings: a) Test 5 sample and b) Test 8 sample (load 1–50 N, distance 50 mm). R. Gabor et al.