Comparison of the Electrochemical Behaviour and Self-sealing of Zirconium Conversion Coatings Applied on Aluminium Alloys of series 1xxx to 7xxx

The following aluminium alloys were used as substrates:

• (1)  
  1050A-O^a/EN AW^b-1050A/EN AW-Al99.5/UNS^c-A91050 with composition: Al 99.50 wt.%, Fe 0–0.4 wt.%, Si 0–0.25 wt.%, Zn 0–0.07 wt.%, Mn 0–0.05 wt.%, Ti 0–0.05 wt.%, Cu 0–0.05 wt.%, Mg 0–0.05 wt.%. It was produced by Impol 2000 d.d. (Slovenska Bistrica, Slovenia) in the form of a 1.5 mm thick sheet. Samples used in experiments were cut in a square shape with dimensions 3 cm × 3 cm, unless stated otherwise.
• (2)  
  2024-T3/EN AW-2024/EN AW-AlCu4Mg1/UNS-A92024 alloy with composition: Cu 4.35 wt.%, Mg 1.33 wt.%, Fe 0.25 wt.%, Mn 0.53 wt.%, Zn 0.1 wt.%, other 0.07 wt.%, Al remainder. It was produced by Kaiser Aluminum in the form of 1 mm thick sheet.
• (3)  
  3005-O/EN AW-3005/EN AW-AlMn1Mg0.5/UNS-A93005 alloy with composition: Mn 1–1.5 wt.%, Mg 0.2–0.6 wt.%, Fe 0–0.7 wt.%, Si 0–0.6 wt.%, Cu 0–0.3, Zn 0–0.25 wt.%, Cr 0–0.1 wt.%, Ti 0–0.1 wt.%, residuals 0.15 wt.%. It was produced by Impol 2000 d.d. (Slovenska Bistrica, Slovenia) in the form of 70 micrometres thick foil.
• (4)  
  A356.0-T6/EN AC^d 42100/EN AC-AlSi7Mg0.3/UNS-A13560 alloy with composition: Si 6.5–7.5 wt.%, Fe 0.15 wt.%, Cu 0.03 wt.%, Mn 0.10 wt.%, Mg 0.30–0.45 wt.%, Zn 0.07 wt.%, Ti 0.18 wt.%. It was produced by TALUM d.d. (Kidričevo, Slovenia). Samples used in experiments were cut in a rectangular shape with dimensions 2.5 cm × 4 cm × 0.3 cm.
• (5)  
  380.0-T6/EN AC 46000/EN AC-AlSi9Cu3(Fe)/UNS-A03800 alloy with composition: Si 8.0–11.0 wt.%, Cu 2.0–4.0 wt.%, Fe 0.6–1.1 wt.%, Zn 1.2 wt.%, Mn 0.55 wt.%, Mg 0.15–0.55 wt.%, Cr 0.15 wt.%, Ni 0.55 wt.%, Pb 0.35 wt.%, Sn 0.25 wt.%, Ti 0.20 wt.%. It was produced by TALUM d.d. (Kidričevo, Slovenia). Samples used in experiments were cut in a rectangular shape with dimensions 2.5 cm × 4 cm × 0.3 cm.
• (6)  
  5754-H18/EN AW-5754/EN AW-AlMg3/UNS-A95754 alloy with composition: Mg 2.6–3.6 wt.%, Mn 0–0.5 wt.%, Si 0–0.4 wt.%, Fe 0–0.4 wt.%, Cr 0–0.3 wt.%, Zn 0–0.2 wt.%, Ti 0–0.15 wt.%, Cu 0–0.1 wt.%. It was produced by Impol 2000 d.d. (Slovenska Bistrica, Slovenia) in the form of a 1.5 mm thick sheet.
• (7)  
  7075-T6/EN AW-7075/EN AW-AlZn5,5MgCu/UNS-A97075 alloy with composition: Zn 5.81 wt.%, Mg 2.55 wt.%, Fe 0.21 wt.%, Si 0.08 wt.%, Cu 1.67 wt.%, Al remainder. It was produced by Kaiser Aluminum in the form of a 1 mm thick sheet.

The following chemicals were used for experiments: NaCl (for analysis, Fischer Scientific, Leicestershire), H₂ZrF₆ (50 wt.% in water, Sigma-Aldrich, Saint Louis, USA), NH₄HCO₃ (reagent grade, Sigma-Aldrich, Steinheim, Germany) and Metaclean® (alkaline cleaner, Amity International, Barnsley, United Kingdom). Metaclean® is composed of: sodium silicate (5–15 wt.%), sodium metasilicate (5–15 wt.%), tetrasodium E.D.T.A (1–5 wt.%), amino-tris (metylenphosphonic) acid (1–5 wt.%), benzotriazol (0.1–0.3 wt.%), non-ionic surfactant (5–15 wt.%). 5 wt.% of Metaclean® solution has pH = 12.5. Milli-Q Direct water with a resistivity of 18.2 MΩ·cm at 25 °C (Merck, Darmstadt, Germany) was used for rinsing and for solution preparation.

pH was measured using a pH meter 827 pH-lab equipped with Unitrode/Pt1000 (Metrohm AG, Herisau, Switzerland). Teflon beakers (V = 1 l) were used for conversion baths, while glass beakers (V = 600 ml) were used for alkaline cleaner and rinsing baths. Sample holders were made of Teflon and conversion coating solutions were stored in polyethylene bottles.

Before immersing in a conversion coating bath, samples were ground on SiC papers ranging from 320- to 4000-grit using a grinding/polishing machine LaboPol-20 (Struers, Ballerup, Denmark). Immediately after grinding, samples were ultrasonicated in absolute ethanol for 5 min using an ultrasonic bath Elmasonic P (Elma, Germany) at 37 kHz and 100% power. After cleaning, samples were dried with N₂ and conserved in a desiccator until the chemical cleaning step. Samples were chemically cleaned by immersion in 5 wt.% Metaclean solution (T = 57 °C, stirring rate 150 rpm) for 3 min. Immediately afterwards, samples were first washed off with Milli-Q Direct water, then dipped in a Milli-Q Direct water bath for 45 s. The sample was then immediately dipped in a conversion coating bath (T = 25 °C, stirring rate 450 rpm). The concentration of the conversion bath was 200 ppm of H₂ZrF₆. pH was set to 4.8 with 15 wt.% NH₄HCO₃. Following the conversion bath step, the sample was immediately rinsed with Milli-Q Direct water and then dipped in a Milli-Q Direct water bath for 1 min. Samples were then dried with an N₂ stream and left to dry in air for 24 h.

Mechanically ground/polished samples are denoted MP, mechanically ground/polished and chemically pre-treated samples are denoted CP, and zirconium conversion coated CP samples are denoted ZrCC-xxxx or ZrCC-xxx.x, where xxxx and xxx.x stand for the four digits used to designate wrought and cast aluminium alloys, respectively. Note that, for AAs 1050A and A356.0, the designations ZrCC-10150A and ZrCC-A356.0 are used, respectively.

Electrochemical experiments were carried out with a Multi Autolab/M204 (Metrohm Autolab, Utrecht, Netherlands) potentiostat/galvanostat controlled by Nova 2.1 software. Measurements were conducted in a standard three-electrode flat cell with a volume of 250 ml. The sample was the working electrode pressed against an o-ring, a carbon rod was the counter electrode, and a saturated Ag/AgCl_(sat) electrode the reference electrode (E_SHE = 0.198 V). All potentials in the text refer to the Ag/AgCl_(sat) scale. The electrolyte was freshly prepared 0.5 M NaCl. The area of the working electrode was 1 cm² and the temperature of electrolyte 25 °C.

Prior to electrochemical measurements, the sample was allowed to rest at open circuit potential (OCP) for 1 h. Electrochemical measurements were conducted in the following sequence from least destructive to more destructive: (i) measurement of the OCP as a function of resting time, (ii) recording of linear polarization resistance curves, and (iii) recording of potentiodynamic polarization curves. The system was allowed to rest for 10 s before switching to the next electrochemical technique.

Linear polarization resistance curves (LPR) were run in the range –10 to 10 mV vs OCP using a 0.1 mV s⁻¹ scan rate. Polarization resistance (R_p,LPR) was determined as the slope of the fitted potential (E) vs current density (j) curve, using Nova 2.1 software.

Potentiodynamic polarization curves were recorded in the potential range starting at −250 mV vs the OCP in an anodic direction until current density reached 0.1 mA cm⁻². The scan rate was 1 mV s⁻¹.

Corrosion current density and corrosion potential were determined using Nova 2.1 software based on the Tafel extrapolation method.⁷¹ Polarization resistance was calculated from parameters obtained with the Tafel extrapolation (R_p,Tafel) method, using Nova 2.1 software based on Eq. 1 in accordance with standard ASTM G59-97.⁷²

$$\begin{eqnarray}&&{R}_{{\rm{p}},{\rm{Tafel}}}\,=\,\displaystyle \frac{{\rm{B}}}{{j}_{{\rm{corr}}}}\end{eqnarray} \tag{ 1 }$$

where B is the Stern-Geary constant.

Electrochemical impedance spectroscopy (EIS) spectra were recorded from 100 kHz to 3 mHz frequency using an AC potential amplitude of 10 mV (rms). Spectra were recorded at the OCP, every day up to 5 d of immersion.

Electrochemical measurements were repeated for each condition for at least three different samples. Representative measurements are presented in plots, and average values with standard deviation are given. Where indicated, average values are also given in plots (e.g. for ∣Z∣_3mHz and OCP).

SEM images were taken in secondary electron (SE) and back-scattered (BSE) modes, using an FEI Helios Nanolab 650 microscope, and EDS spectra were taken using an Oxford Instruments AZtec system with X-max SDD (50 mm²) detector. The analysed area was 100 nm² with an analysis depth of 100 nm (near-surface condition) at 3 kV beam energy. EDS mapping was recorded with 3 kV beam energy and resolution of 1024 lines. Prior to analysis, samples were coated with a thin carbon layer to reduce the charging effect.

ToF-SIMS ion depth profiles were made using the ToF-SIMS V spectrometer (ION TOF GmbH—Munster, Germany). The analysis chamber was operated at a pressure below ∼10⁻⁹ mbar. Depth profiles for negative ions were recorded by interlacing a pulsed 30 keV Bi⁺ primary ion source delivering 1.2 pA target current over a 100 × 100 μm² area with sputtering using a 2 keV Cs⁺ source beam delivering 523 nA target current over a 400 × 400 μm² area. The sputtering rate for ToF-SIMS depth profiling was determined by depth profiling of a thin film of Al₂O₃ of known thickness (50 nm) on Si substrate under the same conditions as those applied for other samples. The sputtering rate was 0.30 nm s⁻¹ (±0.02 nm s⁻¹).

XPS was carried out using a PHI-TFA XPS spectrometer (Physical Electronics Inc.). The vacuum during the XPS analysis was in the range of 10⁻⁹ mbar. The analysed area was 0.4 mm in diameter and the depth 3–5 nm; X-rays were provided from a monochromatic Al Kα source at a photon energy of 1486.6 eV. The XPS spectrometer was nominally operated at an energy resolution of 0.6 eV measured on the Ag 3d_5/2 peak. XPS spectra were analysed by Multipack software, version 8.0 (Physical Electronics Inc.). The position of binding energies of the experimental spectra was aligned relative to the position of C 1s peak at 284.8 eV.

The conversion coating process was followed by measuring the open circuit potential of aluminium alloys immersed in a conversion bath of 200 ppm H₂ZrF₆ (Fig. 1). Initial decay in the potential in the OCP vs time curves represents the phase of surface activation and the initiation of coating formation (Fig. 1). During this phase, natural aluminium oxide is removed by an attack by free F⁻ ions and surface is activated for coating formation.^60,61 Immediately after the removal of natural aluminium oxide, hydrogen evolution and oxygen reduction reactions start at intermetallic particles (IMPs). These reactions increase the local pH up to 8.5, causing precipitation of ZrCC.²⁹ The minimum in the OCP curve represents the point at which the rate of precipitation starts to prevail over metal dissolution.^6,16,39,40 Afterwards, the rate of conversion coating formation is greater, leading to coating growth in the lateral direction (Fig. 1). The maximum or plateau in the OCP diagram (Fig. 1, asterisk marks) was taken as the optimal (or appropriate) conversion time, i.e. the time at which the coating is fully formed. The position of optimal conversion time was checked previously by measuring electrochemical properties and analysing the surface by SEM/EDS of selected samples prepared at different conversion times.^38,40 At shorter time, the coating does not cover the surface completely and shows inadequate corrosion protection.^38,40 At longer times, the coating tends to grow further but, due to internal stress and formation of hydrogen microbubbles, it will crack and lose barrier properties (Fig. 1).^40,55,73

Zoom In Zoom Out Reset image size

The rates of activation, coating formation and optimal conversion depend on aluminium alloy microstructure and conversion bath composition. The presence of more noble areas (IMPs, grain boundaries or phases) than there are in the aluminium alloy matrix will increase the rate of conversion coating growth, leading to a shorter optimal time of conversion. At the same time, the less stable natural aluminium oxide film will be easier to remove by the action of free F⁻ ions, giving a faster rate of surface activation and, consequently, a shorter optimal conversion time. Optimal conversion times for ZrCC formation in the 200 ppm H₂ZrF₆ conversion bath, as a function of aluminium alloy substrates, are given in Table I.

AA2024 required the shortest conversion time due to the presence of a large amount of Cu containing IMPs. In contrast, AA1050A required the longest optimal conversion time due to the small number of IMPs. Optimal conversion times listed in Table I were used in the following experiments.

Potentiodynamic polarization curves of MP, CP and ZrCC aluminium alloys immersed in 0.5 M NaCl are presented in Figs. 2–3. Before measurements, samples were rested for 1 h at the OCP. Electrochemical data deduced from the potentiodynamic polarization curves are presented in Tables II, SI–SIII. The corrosion resistances of mechanically pre-treated aluminium alloys, based on electrochemical data (small j_corr, large R_p), were in the following order: 380.0 ≤ 2024 ≤ 7075 ≤ 1050A ≤ 5754 ≤ A356.0 ≤ 3005 (Figs. 2–3, Tables SI, SIII). AAs 380.0, 2024 and 7075 had corrosion current densities in range 5·10⁻⁷ A·cm⁻² to 7·10⁻⁷ A·cm⁻² and did not show any passivity. This result indicates their low corrosion resistance and high susceptibility to pitting due to the presence of a large amount of noble Cu containing IMPs in AA380.0 and AA2024 and less noble Zn containing IMPs in AA7075. The AA1050A showed high resistance to pitting corrosion due to the absence of more noble IMPs, that resulted in passivity window (ΔE = ∣E_pit − E_corr∣) of ≈0.13 V and current density of 7·10⁻⁷ A·cm⁻² (pitting potential is denoted as E_pit). Thanks to the presence of Si and Mg, AAs A356.0 and 5754 exhibited better corrosion resistance, with ΔEs of 0.07 V and 0.05 V and j_corr of 1.7·10⁻⁷ A·cm⁻² and 1.2·10⁻⁷ A·cm⁻², respectively. AA3005, which contains Mn, showed the highest corrosion resistance, with ΔE = 0.15 V and j_corr of 3.9·10⁻⁸ A·cm⁻². These results were expected and in good agreement with those published.⁷⁴

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Chemical pre-treatment differently affected the surfaces of the alloys. For most of the alloys the pre-treatment induced reduction of the passivity window and of j_corr values (Figs. 2–3, Tables SII–SIII). In addition to j_corr, current densities at 0.1 V more negative and more positive in respect to E_corr in cathodic and anodic branches (j_c and j_a, respectively) were given as well in Table SIII. In other words, chemical pre-treatment cleaned and activated the surfaces of aluminium alloys by removing the aluminium oxide layer from IMPs. The only exception was AAA356.0 which was passivated after chemical pre-treatment, exhibiting j_corr of 6.1·10⁻⁸ A·cm⁻² and ΔE of 0.05 V (Figs. 2–3, Tables SII–SIII). This was probably due to the production of a more uniform aluminium oxide passive layer and passivation of Mg-rich IMPs in the alkaline environment. For AA1050A, AA5754 and AA380.0 the j_corr remained similar as for the MP surface indicating that that the surface was not activated.

The zirconium conversion coating process improved the corrosion resistance of all aluminium alloys investigated; this is reflected in increased ΔE and reduced j_corr (Figs. 2–3, Tables II, SIII). The corrosion potential of ZrCC coated alloys was shifted towards more negative values, thus causing the broadening of ΔE in comparison with MP and CP and indicating cathodic inhibition. Breakdown of passivity at the E_pit is related to the pitting at the IMPs that are not sufficiently protected by the conversion coating; it occurred at similar or slightly more positive values as for the CP samples. Therefore, increase of the ΔE is primarily related to the negative shift in E_corr.

Protection provided by ZrCCs based on the ΔE values was in the following order: A356.0 ≥ 1050A ≥ 5754 ≥ 3005 ≥ 2024 ≥ 7075 ≥ 380.0 while, in the following order based on the j_corr values, (from the lowest to the highest): 380.0 ≤ A356.0 ≤ 2024 ≤ 7075 ≤ 3005 ≤ 5754 ≤ 1050A.

To summarize, regardless the aluminium substrate, after 1 h immersion in 0.5 M NaCl zirconium conversion coating acted as a cathodic inhibitor, judging from the shift of E_corr to negative direction; concomitantly an extended passive plateau is established on the anodic side due to progressive precipitation of Zr oxide/hydroxide which provided barrier protection. Further evaluation of corrosion protection level requires examination of the electrochemical behaviour of ZrCC during prolonged immersion in chloride solution by EIS measurements. Namely, after pre- and conversion treatment, the j_c at −0.1 V vs E_corr was still somewhat larger than j_a at 0.1 V vs E_corr (Figs. 2 and 3, Table SIII) indicating that ZrCC did not passivate the IMPs completely and that oxygen was able to reach the substrate through pores and cracks in the coating and becomes involved in cathodic reaction. Our previous studies show that the corrosion characteristics of ZrCCs, including the reduction of j_corr, j_c and j_a, improve by immersion time in the period of several days in NaCl due to progressive release of F⁻ from the coating and formation of ZrO₂·2H₂O_(s) and Al(OH)₃.^39,40 Herein, potentiodynamic measurements were made after only 1 h immersion which is still not sufficient to achieve the best protective properties of the coating and to confirm the self-sealing effect. This will be shown below by EIS measurements.

The electrochemical behaviour of ZrCCs applied on aluminium alloys was studied using EIS during five days of immersion in 0.5 M NaCl (Figs. 4, S1–S6 (available online at stacks.iop.org/JES/167/111506/mmedia) and Tables III, SIV–SV). Corrosion resistance, the self-sealing effect and active corrosion protection of zirconium conversion coatings can be estimated with the impedance modulus at low frequencies (3 mHz) which is approximately equal to total polarization resistance (│Z│_3mHz ≈ R_p).^75,76 Values of │Z│_3mHz for ZrCC applied on aluminium alloys during prolonged immersion in 0.5 M NaCl solution are presented in Figs. 4g–4i and Tables III and SIV–SV. ZrCC-3005 showed the best corrosion resistance, with values of │Z│_3mHz increasing during five days of immersion, from 2.9·10⁵ Ω·cm² to 7·10⁶ Ω·cm² (Fig. 4g). ZrCC-A356.0 also showed excellent corrosion resistance, with the value of │Z│_3mHz increasing up to three days of immersion, reaching 2.6·10⁶ Ω·cm². At longer times it decreased slowly, reaching values of 1.4·10⁶ Ω·cm², indicating that pitting had started. Values of │Z│_3mHz for ZrCC-380.0 and ZrCC-5754 increased after one day of immersion but then slowly decreased, reaching steady values of 7.6·10⁵ Ω·cm² and 1.1·10⁵ Ω·cm², respectively. This indicated behaviour similar to that of ZrCC-A356.0 but with smaller values of│Z│_3mHz. ZrCC-1050A exhibited a similar trend, with values increasing up to three days and then decreasing, reaching 2·10⁵ Ω·cm² on the fifth day of immersion (Fig. 4h).

Zoom In Zoom Out Reset image size

In contrast, values of │Z│_3mHz for ZrCC-2024 and ZrCC-7075 decreased sharply from starting values of 5.2·10⁵ Ω·cm² and 5.6·10⁵ Ω·cm² down to 2.6·10⁵ Ω·cm² and 1.6·10⁵ Ω·cm² after one day of immersion, indicating the initiation of localized pitting corrosion. The values of │Z│ continued to decrease until the end of the test (Fig. 4h).

The increasing values of │Z│_3mHz for ZrCCs during immersion in 0.5 M NaCl were related to the self-sealing effect. These results show that this effect is closely related to the chemical composition and microstructure of the aluminium alloy substrates and occurred in the following order: 3005 ≥ A356.0 ≥ 5754 ≈ 380.0 ≥ 1050A. Self-sealing did not occur for ZrCC-2024 and ZrCC-7075, probably due to the presence of large amounts of Cu- and Zn-rich IMPs, respectively. Mechanically and chemically pre-treated aluminium alloys exhibited much lower values of │Z│_3mHz during immersion (Figs. 4a–4f). The only exception was mechanically and chemically pre-treated AA3005, which showed improvement of corrosion resistance during immersion for up to 1–2 d (Figs. 4a, 4d). This trend was probably caused by the presence of Mn, which improved the barrier properties and stability of the natural aluminium oxide film during immersion.³⁹

The trend of increasing │Z│_3mHz values is accompanied by the progressive shift in OCP to more positive values for all ZrCCs coated alloys except AA2024 and AA7075 which do not show the self-sealing effect (Fig. 4i). For the best performing ZrCC-3005, this shift was about 400 mV.

To summarize, EIS results show that ZrCCs provide excellent corrosion protection of AAs 3005 and A356.0, good corrosion protection of AAs 380.0, 5754 and 1050A, but inadequate corrosion protection of 2024 and 7075 aluminium alloys.

SEM secondary electron (SE) and backscattered (BSE) images of chemically pre-treated aluminium alloys are presented in Fig. 5. The corresponding EDS analyses of the numbered spots are given in Table IV (spot 1) and SVI–SXII (spots 1–3). EDS analysis was performed with 3 kV beam energy, corresponding to 100 nm analysis depth (near-surface condition). EDS analysis was carried out to obtain the basic composition of the surface. It was not so detailed because the microstructure of these alloys is already well known (vide infra).

Zoom In Zoom Out Reset image size

The surfaces of chemically pre-treated alloys consisted, predominantly, of spontaneously formed aluminium oxide (Table IV). Depending on the substrate, other elements were also detected, ranging from 0.6 at% up to 10.8 at% at spot 1 (larger area). When IMPs were analysed, the concentration of individual elements was larger (Tables SVI–SXII), in accordance with the predominant type of IMP in alloys. Due to the presence of a relatively high amount of Fe and Si, IMPs in AA1050A are AlFe₃ and α-AlFeSi. Both of these IMPs are more noble compared to the alloy matrix.⁷⁷ The microstructures for AA3005 and AAA356.0 have been reported in more detail.^39,40

Based on chemical composition, AA2024 has two main groups of IMPs. The first group represents round-shaped S-phase (Al–Cu–Mg) IMPs that are more active electrochemically and susceptible to anodic dissolution but, due to dealloying, become local cathodes. Larger, irregularly shaped Al–Cu–Mn–Fe (Al₇Cu₂Fe, Al₂₀Cu₂Mn₃, and Al₂Cu) IMPs belong to the second group.^78–82 These IMPs are more noble compared to the AA2024-matrix and represent local cathodes in the corrosion process.⁷⁸

AA3005 and AA5754 contain the same IMPs: (MnFe)Al₆ or α-(MnFe)₃SiAl₁₂.⁸³ However, AA3005 contains larger IMPs, and a higher number of IMPs than AA5754. Further, AA3005 contains a significant amount of α-(MnFe)₃SiAl₁₂, while AA5754 contains, mainly, (MnFe)Al₆ with a much higher ratio of Fe/Mn = 7.8 than AA3005 which has an Fe/Mn ratio of 0.8.⁸³ Also, AA3005 contains more small IMPs (dispersoids) than AA5754. Dispersoids are (MnFe)Al₆ and α-Al(Fe,Mn)Si IMPs. AA5754 also contain AlFe₃, Mg₂Si and, sporadically, Al₆CuMg₄.⁸⁴ Both of these IMPs are noble relative to the alloy's matrix and act as local cathodes.

The microstructure of AAA356.0 consists of an α-Al matrix (grains), eutectic Si grain boundaries, and IMPs.^{40,78,85–92} Typical IMPs in AAA356.0 are Al–Si–Fe, Mg–Si and Al–Si–Fe–Mg IMPs (β-Al₅FeSi-needles, π-Al₈FeMg₃Si₆, Mg₂Si); if there is a higher amount of Mn in alloy composition, α–Al–Si–Fe–Mn IMPs and α-Al₁₅(FeMn)₃Si₂—Chinese script or skeleton formations will be found. Sporadically, it is possible to find Al₅Cu₂Mg₈Si₆. All these IMPs and eutectic-Si are noble to the α-Al matrix except Mg₂Si, which is less noble.

The main AA380.0 microstructure components are α-Al matrix (grains), eutectic Si grain boundaries, and IMPs (β-Al₅FeSi-needles, Al₂Cu, Al(FeCuCr) and α–Al–Si–Fe–Mn, which are formed at grain boundaries.^{85,90,93–95} It is possible to find Al₁₅(FeMn)₃Si₂—Chinese script as well if the content of Mn in the alloy is high enough. All these IMPs and eutectic-Si are noble relative to the α-Al matrix.

MgZn₂ and Mg₂Si are IMPs in AA7075 that are susceptible to anodic dissolution. Al₂Cu, Al₃Fe and Al₇Cu₂Fe can sustain high cathodic currents and S-phase (Al₂CuMg)^{78,80,81,96–99} S-phase is susceptible to anodic dissolution. Once, when Mg is removed from S-phase, it reverses its corrosion potential (behaviour) and starts acting as a local cathode.

SEM secondary electron (SE) and backscattered (BSE) images of zirconium conversion coatings are presented in Figs. 6–7, and the corresponding EDS analysis of numbered spots is given in Table IV (spot 1), SVI–SXII (spots 1–5). Images are given for as-prepared coatings (i.e. prior to immersion in NaCl) (Figs. 6–7a-a3) and, further, after one day (Figs. 6–7b-b3) and three days (Figs. 6–7c-c3) immersion in 0.5 M NaCl. In SEM SE mode, the nodular structure of conversion coatings is visible only at higher magnification. On the other hand, the formation and lateral distribution of ZrCC throughout the surface is much more visible in SEM BSE mode, as indicated by the bright areas.⁴⁰ These denote the spots where the formation of ZrCC was initiated and resulted in much thicker final ZrCC. Noble IMPs and eutectic Si grain boundaries of AAA356.0 and AA380.0 aluminium cast alloys were present beneath ZrCC (Figs. 6, 7). EDS analysis shows that as-prepared ZrCC contains ca. 13–18 at.% Zr, 7–12 at.% F, 27–44 at.% O and a minor amount of elements originating from the substrate and/or IMPs (Tables IV, SVI–SXII). The only exception was ZrCC-2024 that contained a much smaller amount of Zr (5.3 at.%), F (2.0 at.%), O (16.9 at.%) and a much higher amount of Al (72.2 at.%) than other as-prepared ZrCCs. On the other hand, it had a higher concentration of Zr (24.6 at.%), F (6.2 at.%) and O (54.6 at.%) at bright spots (Fig. 7a1). Based on these results it can be stated that ZrCC-2024, prepared after 3 min conversion time, was formed non-uniformly across the AA2024 surface and was much thicker at bright spots (IMPs) and much thinner at the matrix. These bright spots were noble, Cu containing IMPs that led to this unequal formation of ZrCC-2024. The overall result was that the mean thickness of ZrCC-2024 is much smaller than that of the other ZrCCs and, according to EIS measurements, does not provide adequate corrosion protection (Fig. 4).

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The general trend during immersion of ZrCC in 0.5 M NaCl solution was a decrease of F concentration, indicating its leaching from the conversion coating (Table IV). After three days of immersion, the concentration of Zr remained at approximately the same level, while that of O increased by 10 to 15 at.% for ZrCC-1050A, ZrCC-5754, ZrCC-3005, ZrCC-A356.0 and ZrCC-380.0. The surfaces of ZrCC-1050A, ZrCC-5754 and ZrCC-3005 showed no signs of corrosion (Fig. 6c-c2), in good agreement with EIS measurements (Fig. 4), indicating that these ZrCCs were self-sealed. For ZrCC-A356.0 and ZrCC-380.0 there was no corrosion damage after one day of immersion (Figs. 6b3, 7b); however, after three days, corrosion damage appeared in the form of cracks/pits in both alloys (Figs. 6c3, 7c). In these areas, increased concentrations of Zn were detected, indicating that the presence of Zn caused pitting (Tables SIX, SX). The extent of corrosion damage in both alloys was low although corrosion damage was less for ZrCC-A356.0 than for ZrCC-380.0, probably due to higher concentrations of Zn and Cu in the latter.

ZrCC-2024 and ZrCC-7075 exhibited different behaviour during immersion in 0.5 M NaCl. After only one day of immersion, corrosion products appeared on ZrCC-2024, especially around IMPs (Fig. 7b1). The concentration of O was increased by ca. 30 at.%, and that of Zr and F decreased relative to as-prepared coatings (Tables IV, SXI). After three days, severe corrosion damage was visible in the form of cracks (Fig. 7c1). An even higher concentration of O (61.0 at.%) and a lower concentration of Zr (3.0 at.%) and F (0.9 at.%) were detected. These results reflect the extensive coverage of the surface with Al(OH)₃ corrosion products and the dissolution of ZrCC during the immersion, indicating that ZrCC was not able to provide proper corrosion protection for AA2024.

The corrosion of ZrCC-7075 was also initiated after one day of immersion in 0.5 M NaCl (Fig. 7b2, dark spots). At these spots, the concentrations of Zn and Cu were larger, indicating the presence of CuZn-rich IMPs under the corrosion products in the form of aluminium oxide/hydroxide (Tables IV, SXII). After three days of immersion, corrosion damage was more severe in the form of a large crack and pits. At these spots, the concentrations of Cu, Zn, Fe were increased, correlating with the presence of CuZnFe IMPs (Fig. 7c2, Tables IV, SXII). At the same time, the formation of corrosion products in the form of Al(OH)₃ is indicated. The concentration of Zr was significantly decreased at this spot due to the dissolution of ZrCC and coverage with corrosion products. The general trend during immersion of ZrCC-7075 in 0.5 M NaCl was that the concentration of F decreased 5-fold, indicating leaching of fluorine, and the concentration of O increased. However, the most interesting feature is that the concentration of Zr decreased by 30% after three days of immersion, proving severe dissolution of the coating and its coverage with corrosion products. These results correlate well with EIS measurements (Fig. 4) and show that ZrCC is not able to provide proper corrosion protection to AA7075.

Chemically pre-treated aluminium alloys, as-prepared ZrCCs and ZrCCs immersed in 0.5 M NaCl solution for three days were all characterized with XPS. The XPS method, with an analysis depth of ca. 10 nm, provides information from the top-most surface of the ZrCC. The composition of the surface obtained from the XPS survey spectra is presented in Table V.

The general characteristic of all the aluminium alloys is that the surface was composed of Al, O and P. In addition, AA5754 and AA7075 contained Mg at the surface, cast AAA356.0 and AA380.0 contained Si and AA2024 contained a small amount of Cu. Phosphorus originates from the Metaclean alkaline cleaner used in chemical pre-treatment of the aluminium alloys. Aluminium and oxygen originate from the aluminium oxide/hydroxide. Metaclean alkaline cleaner contains NaSiO₃ and P-based corrosion inhibitor/surfactant. After the conversion coating process, most of the ZrCC contained 20–23 at.% Zr, 8–10 at.% F and around 1 at.% Al. ZrCC-2024 and ZrCC-7075 contained less Zr and more Al. The latter indicates that ZrCC-2024 and ZrCC-7075 coatings were thinner than those of other alloys, as was also shown by SEM/EDS (Figs. 6, 7).

After three days of immersion in NaCl solution, the concentration of F was reduced in all ZrCCs. The concentration of Zr remained high in all ZrCCs other than ZrCC-2024 and ZrCC-7075, in which it decreased to 3 at.%. At the same time, the concentrations of O and Al increased for all aluminium alloys, the amount of Al being much higher for ZrCC-2024 and ZrCC-7075. The latter two also contained a significant amount of Cl. These results indicate that ZrCC-2024 and ZrCC-7075 dissolved and were covered by large amounts of corrosion products of aluminium hydroxide/oxide. In contrast, ZrCCs applied on other aluminium alloys were very stable and did not corrode.

High energy resolution XPS spectra for elements of interest are presented in Figs. 8–14. For most alloys, two peaks appeared in the Al 2p spectrum for CP substrates at binding energies (E_b) of 72.5 eV and 74.5 eV (Figs. 8–14), corresponding to Al metal and Al(OH)₃, respectively.^100,101 For CP-1050A and CP-380.0, only one peak appeared at 74.5 eV, corresponding to Al(OH)_3. This indicates that the surfaces of CP-1050A and CP-380.0 were covered with a thicker layer of Al(OH)₃, while the hydroxide layer was thinner on other aluminium alloys, thus allowing the detection of Al metal underneath.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

As shown in Table V, Al peaks were not detected for most of the as-prepared ZrCC coated substrates, proving complete coverage with the coating. Only for ZrCC-2024 was an Al peak detected at a binding energy of 75.5 eV, suggesting that ZrCC was very thin, and could be related to aluminium fluoride compounds.

After three days of immersion, for most of the ZrCCs, the peak at 74.4–74.8 eV re-appeared, indicating the formation of Al(OH)₃ in the top layer of coating, or the formation of corrosion products. However, in the case of 2024-ZrCC, a broad peak, related to pure Al appeared at 72.8 eV.

For all chemically pre-treated aluminium alloys, the O 1s peaks were centred at 531.7–531.9 eVs, indicating the prevalence of hydroxide over oxide (Figs. 8–14). Qualitatively, three peaks for bound H₂O, Al(OH)₃ and ZrO₂·xH₂O, at binding energies at 533 eV, 532 (531.9) eV and 530.5 eV, respectively,^101,102 could be identified. After the conversion coating process, the centre of the O 1s peak shifted to E_b 530.8–531.7 eV, depending on the aluminium alloy, and showed a shoulder at 530.5 eV that corresponds to hydrated Zr oxide and oxyfluoride (ZrO₂·xH₂O/ZrF_xO_y).

In general, after immersion of ZrCC in NaCl solution, the centre of the O 1s peak shifted back to slightly higher binding energy 531.5–531.7 eV, corresponding to ZrO₂·xH₂O and Al(OH)₃, the only exception being ZrCC-2024. In the latter, the centre shifted from 531.7 eV to 530.5 eV, corresponding to aluminium oxide.

The F 1s peak was centred at E_b 684.6–685.2 eV and was present only for the as-prepared ZrCC, suggesting the presence of ZrF₄/ZrF_xO_y in the top layer (Figs. 8–14). After three days immersion, the F 1s peak disappeared, indicating leaching of fluorine from the conversion coating and transformation of ZrF₄/ZrF_xO_y to ZrO₂·xH₂O, in accordance with the O 1s peak (Figs. 8–14).

After the conversion coating process, high energy resolution Zr 3d spectra were detected for all aluminium alloys. (Figs. 8–14). The Zr 3d_5/2 and 3d_3/2 peaks for the as-prepared conversion coatings were centred at 185.0–185.6 eV and 182.7–183.3 eV respectively, depending on the aluminium alloy. This spectrum was related to a mixture of ZrF₄/ZrF_xO_y and ZrO₂·xH₂O,¹⁰³ since the coating contained a relatively high fluorine and oxygen concentration (Table V).

After immersion in NaCl solution, the peaks shifted to lower binding energies of 184.5–185.0 eV and 182.1–182.6 eV, respectively, indicating the transformation of Zr-fluoride/oxyfluoride to ZrO₂·xH₂O. These results correlate with the disappearance of the fluorine peak in the F 1s spectrum. The changes were especially pronounced for ZrCC-2024 and ZrCC-7075, with significantly reduced intensities of the Zr 3d peaks related to the dissolution of ZrCC.

To check the thickness of ZrCCs and their structure, we recorded ToF-SIMS negative ion profiles for as-prepared conditions and for after-immersion in 0.5 M NaCl solution for three days (Figs. 15–16). The sputtering rate was 0.30 nm s⁻¹ (±0.02 nm s⁻¹). The ToF-SIMS ion depth profiles, recorded on areas of 500 μm², provide a mean conversion coating thickness, determined as based on depth profiles (as described in Experimental), given in Table VI. The boundary between the outer and inner layers was determined as the depth at which the intensity of ZrO⁻ dropped to a half of its value at the layer surface. The depth at which the AlOF⁻ intensity dropped to half its maximum value was taken as marking the boundary between the inner layer and substrate.

Zoom In Zoom Out Reset image size

Zoom In Zoom Out Reset image size

The general feature of ZrCCs applied on aluminium alloys was a bi-layer structure with its outer layer having a higher concentration of ZrO⁻ (it could be ZrO₂·xH₂O/ZrO_xF_y species) and a lower concentration of AlOF⁻. Compared to the outer layer, the inner layer had a higher concentration of AlOF⁻ ions and a lower concentration of ZrO⁻ ions. The ToF-SIMS profiles of AlOF⁻ ions were accompanied by profiles for AlO⁻, MgO⁻ and SiO⁻ ions (Fig. 16). The only exception to this behaviour was ZrCC-2024, which was much thinner than other conversion coatings and had a single-layer structure. An interesting detail to note is that, for all ZrCCs, the intensity of the ZrO⁻ ions profile was higher or the same as that of the AlOF⁻ ions profile. However, in the case ZrCC-3005, the intensity of the AlOF⁻ ions profile was higher than that of the intensity of the ZrO⁻ ions profile.

After three days of immersion in 0.5 M NaCl, the general features of ZrCC-1050A, ZrCC-3005, ZrCC-A356.0, ZrCC-380.0 and ZrCC-5754, which showed adequate corrosion protection, were as follows. The profiles of ZrO⁻ ions were similar, having a maximum in the outer layer (Fig. 16). The intensities of F⁻ and AlOF⁻ ion profiles decreased significantly, while that of Cl⁻ ions increased. The latter followed the same shape as that of F⁻ ions. This suggests that Cl⁻ ions were exchanged with F⁻ ions and incorporated within the ZrCC structure. Compared to as-prepared coatings, the intensity of the AlO⁻ profile was larger in the outer layer, reaching a maximum (most pronounced for ZrCC-3005) and then forming a plateau in the inner part of ZrCC. This trend indicates that, on immersion in NaCl solution, the coatings, especially the outer layers, become enriched in aluminium oxide/hydroxide. For the ZrCC-3005 the enrichment in AlO⁻ was also pronounced at the inner layer/substrate interface, which could be the reason for the excellent performance of this particular coating. On immersion in NaCl, the AlOF⁻ profile follows the opposite trend to that of AlO⁻, and becomes strongly diminished in the outer layers, suggesting that it is gradually transformed to AlO⁻. These transformations could be one of the reasons for the improved corrosion resistance of ZrCCs. The MgO⁻ profile decreased and shifted more deeply toward the interface with the substrate. For the majority of alloys containing Si the intensity of SiO⁻ profiles shifted towards the surface, possibly indicating incorporation of Si into the coating structure.

ZrCC-2024 and ZrCC-7075 displayed different trends in their negative ion profiles after immersion in NaCl. The intensities of the ion profiles for ZrO⁻, AlOF⁻, AlO⁻ decreased significantly for both alloys and was much lower compared to other aluminium alloys. The intensity of Cl⁻ profiles increased significantly for both alloys, however in case of ZrCC-7075 was much higher compared to other profiles. These results indicate that ZrCC-2024 and ZrCC-7075 dissolved and were covered with corrosion products. For ZrCC-2024, the intensity of CuO⁻ ion profiles increased while, for ZrCC-7075, that of FeO⁻ increased. Based on these results, ZrCC-2024 and ZrCC-7075 were not stable and, thus, did not provide proper corrosion protection during immersion in 0.5 M NaCl.

Electrochemical impedance spectroscopy measurements showed that the corrosion resistance of ZrCC-3005 increased continuously over five days of immersion in 0.5 M NaCl solution. That of ZrCC-1050A and ZrCC-A356.0 increased for up to three days but, on the fourth day of immersion, the average impedance modulus at low frequency decreased, indicating that pitting had occurred. The same trend was noted for ZrCC-5754 and ZrCC-380.0, but to a lesser extent. The impedance moduli for these two alloys showed similar trends at the beginning but then decreased slowly, reaching values of ca. 1 MΩ·cm² on the fifth day of immersion. On the other hand, after 1 h immersion, ZrCC-2024 and ZrCC-7075 showed the highest impedance moduli of all ZrCCs, which then decreased significantly, indicating that pitting had already occurred and that ZrCC-2024 and ZrCC-7075 did not provide proper corrosion protection.

SEM-EDS and XPS analyses showed that, during the immersion of ZrCC-1050A, ZrCC-3005, ZrCC-A356.0, ZrCC-380.0 and ZrCC-5754 in 0.5 M NaCl fluorine leached out of the coating. The concentration of oxygen increased while that of Zr remained at approximately the same level. High-resolution XPS spectra showed that, in the top parts of the coatings, ZrF₄/ZrO_xF_y were transformed into ZrO₂·xH₂O which was accompanied by the formation of Al(OH)_3.

ToF-SIMS depth profiles of ZrCC-1050A, ZrCC-3005, ZrCC-A356.0, ZrCC-380.0 and ZrCC-5754 had similar shapes and showed similar trends during immersion in chloride solution. Fluorine was leached out from ZrCCs, but zirconium oxide was stable after three days of immersion. At the same time, the AlO⁻ profile increased, indicating an increased concentration of aluminium oxide/hydroxide in the ZrCCs. Transformation of ZrF₄/ZrO_xF_y into ZrO₂·xH₂O and incorporation of Al(OH)₃ was responsible for the self-sealing of ZrCCs that provided increased corrosion resistance during immersion in 0.5 M NaCl solution. However, ZrCC-A356.0 and ZrCC-380.0 were pitted at Zn-rich IMPs/spots during immersion in 0.5 M NaCl.

For ZrCC-2024 and ZrCC-7075, the SEM-EDS, XPS, and ToF-SIMS analyses showed that the coatings were not stable during immersion in chloride solution. Fluorine leached out but, at the same time, ZrO₂ dissolved and was covered with corrosion products of Al(OH)₃, indicating that ZrCC did not provide proper corrosion protection for these two alloys. ZrCC-2024 had a small average thickness of 12 nm, but SEM-EDS analysis showed that it was significantly thicker at Cu-rich IMPs than at the matrix. The small and unbalanced thickness of ZrCC-2024 was the main reason for the inadequate corrosion protection. We assume that, with adequate surface pretreatment that will dissolve large Cu-rich IMPs and redeposit Cu in the form of small cathodic sites, ZrCC-2024 would perform much better. In the case of ZrCC-7075, the average coating thickness (36 nm) was significantly greater than that of ZrCC-2024 (12 nm) and uniform across the substrate surface. Our opinion, based on microstructural analysis and electrochemical data, is that the main reason for the corrosion protection of ZrCC-7075 being poor was that the content of Zn in the matrix was too high (5.81 wt.%). Even for ZrCC-A356.0 and ZrCC-380.0, which contain substantially lower contents of Zn in the matrix (0.07 and 1.2 wt.%, respectively), the pitting was initiated at Zn-rich areas. Proper surface pretreatment for AA7075 before applying ZrCC could improve the performance of ZrCC significantly. Optimal pretreatment for AA7075 should remove or passivate Zn-rich IMPs at the alloy surface before the application of ZrCC.