Review—Conversion Coatings Based on Zirconium and/or Titanium

The bath pH strongly affects the uniformity of the coating formed during conversion and the resulting coating properties. The pH of 4.0 was used in the vast majority of reports.^{6,36,37,42,51,52,55–57,61,62,65,68,75,85–88,94,98,108,109} although conversion was carried out at pH between 1.5 and 6.^{30,32,46,48,49,53,64,66,68–70,77,78,82,90,95,99,106,107,110} More extensive and thicker coating formation at pH 4.0 than at 2.9 was noticed for Gardabond X4707 coating on AA6060.⁹⁴ The coating formed at pH 4 showed lower cathodic current density compared to that formed at 2.9. The concentration of Zr and Ti in the coating increased with the pH of conversion bath.³⁸ Increasing the pH from 2.5 to 4.5 during the formation of Zr-based coatings on AA1050 resulted in the decrease in OCP values in regions of surface activation and Zr conversion.⁶⁴

The effect of immersion time was investigated in numerous studies to obtain the optimal coating parameters in terms of morphology, composition, thickness and corrosion resistance.^{6,32,34,35,37,50,52,55,61,63,69,73,77–79,88,90,93,94,96,102,107,110} The immersion time of metal substrates in conversion baths ranged from less than 3 minutes,^{31,38,42,46,48,51,57,59,65,67,70–72,85,86,95,97,101,106} and between 3 and 6 minutes.^{30,47,53,64,78,87,110–112} Longer immersion times, from 10 up to 25 min, were also investigated.^{36,49,50,75,80,81} It seems, however, that immersion times were greater than 10 min in baths not containing F⁻, e.g. (ZrOCl₂)⁷⁵ and ZrO(NO₃)₃.⁸¹

In general, longer immersion time produced more uniform and thicker conversion layers. To form Zr/Ti layers on Al of measurable thickness, determined by ellipsometry, an immersion time of at least 90 s was required.⁹⁰ After 600 s, the thickness of 70 nm was reached.⁹⁰ Andreatta et al. studied the effect of immersion time on morphology, composition and Volta potential of Ti/Zr-based coatings on AA6016.¹⁰⁷ After 120 s, the coatings formed mainly on IMPs and the Volta potential decreased to 280 mV (compared to > 400 mV on bare etched substrates) (Fig. 3). When the immersion time was increased up to 300 s, the coating covered the entire surface and was accompanied by a decrease of Volta potential difference to only 5 mV. Progressive coverage of the substrate surface with increasing immersion time was generally noticed.^69,73 SEM images recorded on AA6014 substrate immersed in H₂ZrF₆ bath containing Cu for various immersion times up to 120 s are presented in Fig. 4.⁸⁸ In some studies it was noticed that, although the coatings were thicker at longer immersion time, cracks appeared, probably as a result of internal stress in thick films as the layer dried.^{37,52,60,79,96}

Zoom In Zoom Out Reset image size

Khun et al. measured surface roughness as a function of immersion time during formation of TecTalis coating on AA6061⁹³ and low carbon steel.⁶³ The roughness increased with immersion time up to 60 s as a result of the nucleation, clustering and growth of zirconia nodules.⁹³ For longer immersion times, the surface roughness decreased through the improved uniformity of the oxide layer.⁹³ The increase in surface roughness of Zr/Ti-based coatings with immersion time was beneficial to improve adhesion strength of the organic coatings by physical interactions.⁶⁹

In the majority of studies the conversion process was carried out at room temperature (20–25°C), which minimizes energy costs for industrial processes.^{33,34,36,37,47,51,52,55,59,61,62,64–66,69,70,72,78,82,84,87,90,92,94,95,98,99,101,102,104,106–108,110,112–114} Elevated temperatures were also studied, i.e. between 30 and 55°C.^{6,30,42,46,48,49,53,57,63,67,75,77,79,81,93,111} The conversion bath temperature was varied in several studies to determine the optimal value in terms of morphology, composition and corrosion resistance.^{32,35,43,58,96} The optimal conditions were different for different coatings. On HDG and Galfan coated steel, 40°C was found to be optimal in terms of the Zr concentration in the coating.⁴³ For other types of Zr-based coatings deposited on cold rolled steel (CRS)^96,97 and AA1050,⁵⁸ the optimal temperature was 20°C, as it exhibited lower corrosion current density, i_corr, compared to coatings formed at 30°C and 40°C. Ti-based coatings on electrogalvanized galvanized steel samples treated at 50°C exhibited the longest time to formation of the first white rust.³⁵ Similarly, Ti-based coatings containing Mn(III) phosphate and deposited on Zn coated steel exhibited lower i_corr when prepared at 60°C than at lower temperatures.³²

In only a few studies^{35,52,87,95,99,107} was the conversion bath agitated by stirring during deposition; most baths were stagnant. Recent studies have shown, however, that convection significantly affects the properties of the coating.^99,108 For Zr-based coatings on AA6014 and CRS, thickness increased several times as a result of stirring at 400 rpm.¹⁰⁸ Convection affects the coating conversion kinetics by supplying higher concentrations of fluorides at the substrate surface as a result of accelerated mass transport. This reduces the time required for oxide dissolution and increases the available time zirconium can deposit on the activated matrix.¹⁰⁸ Moreover, the content of copper in the coating, coming from Cu²⁺ added to conversion bath to increase the number of cathodic sites at the surfaces, is enriched on the surface with stirring.

Summarizing, conversion bath parameters (composition, pH, temperature and bath agitation) strongly affect the coating properties such as coating composition, morphology, thickness and corrosion resistance, as will be discussed below. Optimal conditions are typically pH 4, time of immersion between 2 and 5 min and temperature between room temperature and 50°C. The conversion baths have usually been left stagnant during deposition.

The aims of organic polymer additives are to improve adhesion to the underlying substrate, to improve homogeneity of the coating and to establish a base for subsequent organic coatings. Early studies mainly incorporated poly(acrylic acid) (PAA)^19,41,83 and pyrrolidine-based polymer.³⁴ The choice of polymer was reported to be critical to ensure good coating performance.⁴² Polymers had more impact on the polyester paint adhesion than the cation (Ti or Zr) of the coating.⁴² A given polymer, however, is more effective in a Zr-based coating than in a Ti-based one due to the ability of Zr to act as a cross-linking agent. When treated in baths without polymers, the coatings were nonhomogeneous and exhibited worse corrosion resistance in 3% NaCl.⁴² The presence of PAA in the bath resulted in the 2D formation of polymer film preferentially on Zr-Ti oxide. Its formation was ascribed to physical and/or chemical interactions.¹⁰⁵ Defects at the surface may act as nucleation sites for polymer film formation. After nucleation, which may be favored at hydroxide sites, rapid 2D growth of polymer takes place, extending farther than the precipitated Zr-Ti oxide.

Coatings from baths with various blends of H₂ZrF₆, PAA and polyacrylic amide (PAM) were prepared to investigate their effect on the performance when subsequently coated with epoxy.⁴⁷ The addition of both PAA and PAM to Zr-based coatings improved the corrosion resistance compared to the individual polymers alone. Additives like polyvinyl phenol phosphate,⁵⁴ silane¹⁰⁴ and polypyrrole²⁹ were also added to conversion baths.

Instead of polymers, various chelating agents may be used to enrich the coating composition. Amino trimethylene phosphonic acid (ATMP) is a powerful chelating agent and was used to improve lacquer adhesion and corrosion resistance of Zr- and Ti-based coatings on AA6061.⁴⁶ Such hybrid coatings were denser; phosphonate was formed along with Ti-Zr oxides. Phytic acid is a harmless, large organic chelating molecule that forms stable metal-phytic complexes.⁶⁷ Coatings containing phytic acid exhibited reduced wettability and higher adhesion to an epoxy topcoat than Ti conversion coatings alone. Reduced adhesion loss was ascribed to the improved bonding between epoxy coating and phytic-containing coating, i.e. a chelate reaction between Fe²⁺ or Fe³⁺ and phytic acid.

Tannic acid (TA), C₇₆H₅₂O₄₆, a type of polyphenol, has been used to improve adhesion on steel and zinc since the 19^th century. Metal tannates and tannate complexes formed with metal and metal oxides improved adhesion to the substrate although the level of protection was not substantial.^115–117 TA has been recently considered as a corrosion inhibitor for various metals and alloys.¹¹⁸ Smit et al. added TA (40 g/L) to a Ti-based conversion bath instead of PAA to form coatings on AA3003.⁴⁴ Corrosion protection was found only for immersion times up to 10 h; longer immersion led to deterioration due to dissolution of the organic component. TA (2–3 g/L) was added to a mixed Ti/Zr fluoro conversion coating on AA6063^50,61 and galvanized steel.⁷⁸ A double layer structure was postulated with metal−organic complexes located primarily in the outer layer of the conversion coating, and an inner layer made mainly of Na₃Al₆, TiO₂, Al₂O₃, etc.⁶⁰

The combined action of organic and inorganic additives was studied for Ti-based coatings on HDG steel.^31,72 The organic component was mainly poly-(4-vinylphenol) but linked to the amino group of quaternary ammonium and N-methyl glucamine;³¹ the inorganic component was manganese phosphate, Mn₃(PO₄)₂. The polymeric layer acted as a surfactant resulting in the formation of organic-inorganic adlayer, which presumably started to nucleate at titanium oxide islands (as shown by Nordlien et al.).¹⁰⁵ The role of the organic phase was to coordinate the layer conversion and thus assist in achieving a more homogeneous layer.⁷² The homogeneity, morphology and corrosion resistance of the coating was satisfactory only when the inorganic and organic components were both present; the addition of Mn₃(PO₄)₂ limits the bath acidity at 2.9 and thus stabilizes the layers, as will be discussed below. Quaternary amine groups and N-methyl glucamine rich hydroxide groups are expected to react with metallic species on the HDG steel surface to form organic/inorganic complexes.³¹ Consequently, polymer adsorption on the substrate was improved.

Copper

Cupric ions can be added to the conversion bath with the aim of forming Cu deposits on the metal surface, which serve as additional cathodic sites for the reduction reaction during conversion and thus accelerate the subsequent increase in pH and promote the deposition of Ti- and Zr-hydroxides. In several studies, the commercial hexafluorozirconic acid-based baths containing some amount of copper were investigated.^{6,62,85–88,108} When properly activated, even if the alloy itself does not contain copper (e.g. in AA1050), Cu-rich particles will deposit on the surface and promote the formation of Zr-based coating.⁸⁷ The TecTalis coating containing Cu deposited on CRS was thicker compared to that without Cu (30 nm and 20 nm, respectively).⁶ Copper was enriched, even up to 50−60 wt% at some locations. The Cu deposits were randomly distributed and not restricted to any specific locations within the coating. The content of Cu in the coating significantly increased with agitation of the conversion bath, i.e. by a factor of 3 on CRS and HDG steel substrates, and by a factor of 0.5 on AA6014.¹⁰⁸

The content of Cu in the coating differed depending on the substrate. Areas with different concentrations of Cu were observed on AA6014, with islands of copper (0.2 to 2 μm) greater in size than those observed for CR and HDG steels.^85,86 Cu particles were visible after only 5 s of deposition and then grew in size and density with increasing conversion time.⁸⁸ Copper was present both in the metallic form and as CuO. The presence of CuO on the surface and inside the coating suggests that Cu deposited in the metallic form, but subsequently oxidized, either in the conversion bath or upon contact with air.⁸⁸ Nucleation is thus instantaneous, preferentially at boundaries of intermetallics and is followed by growth of Cu-rich particles in the lateral direction.

The mass gains measured by quartz microbalance on the pure metals Fe, Al and Zn in H₂ZrF₆ conversion baths with and without Cu are presented in Fig. 5.⁶ The deposition rate was always the highest for Al followed by Fe and Zn, and was highest when Cu was present in the conversion bath. At shorter formation times (60 s), the TecTalis coating contained Cu, Cu₂O and CuO.⁶³ At longer times (180 s) the content of CuO decreased, indicating that prolonged treatment promotes the formation of Cu₂O in the ZrO₂ layer. The behavior on alloys can be different as discussed below.

Zoom In Zoom Out Reset image size

The literature contains some inconsistencies regarding the influence of Cu ions in the coating bath on the corrosion performance of the conversion coating. In terms of corrosion performance, Adhikari et al.⁶ reported that TecTalis coatings deposited on steel both with and without Cu showed high impedance values for up to 120 days in 0.5 M NaCl when coated with electrodeposited paint. However, the coating containing Cu provided a significantly improved resistance to delamination.⁶ On the other hand, the addition of Cu ions added as Cu(NO₃)₂ to hexafluorotitanium baths containing silane agent worsened the corrosion protection of AA6014-T4.¹⁰⁴ Although the layer was thicker, it may have been less homogeneous and thus less protective. Lostak et al. studied the effect of Cu²⁺ and Fe³⁺ added as nitrates to Zr-based coatings on HDG steel.⁵² The addition of Cu²⁺ led to a thicker, but more inhomogeneous, Zr oxide layer compared to the addition of Fe³⁺, especially at shorter deposition times (60 s). The addition of Cu²⁺ and Fe³⁺ shifted the OCP more positive because of their oxidizing power and the generation of micro-cathodes.

Zinc and manganese

Nickel, molybdenum and phosphorus

Vanadium, cerium and chromium

Vanadium has been added to conversion baths in both the metavanadate form, as NH₄VO₃^36,37 and as NaVO₃.^59,69,70 The V-modified Ti/Zr coating exhibited superior electrochemical characteristics and improved the adhesive strength of an organic epoxy coating.^36,69 The self-repairing ability of VO₃⁻-containing Zr-based conversion coatings deposited on AA6063 was studied by Zhong et al.⁵³ X-ray photoelectron spectra (XPS) identified the formation of V⁴⁺ and V⁵⁺ compounds linked to ZrO₂. Due to the oxidative action of H₂O₂, V⁵⁺ ions can exist and convert into hydrate, which can be transferred to the part of the surface where localized attack occurs:

In this sense, V⁵⁺ ions act similar to Cr⁶⁺ ions. The repair effect was lost after 5 d immersion in NaCl solution.

Vanadium pentoxide, V₂O₅, was added to Zr-based hybrid conversion baths that also included inorganic ammonium zirconium carbonate, organic cyclic-amine-containing polymer (CAP), and ascorbic acid.¹¹⁹ Because ascorbic acid was present, V⁵⁺ could be reduced to V⁴⁺.¹¹⁹ The coating containing both CAP and V showed the best corrosion performance in the salt spray chamber. It was suggested that co-incorporation of CAP and V gave rise to more effective coating inhibition via controlled release of V, and an ability to repair scribes through the coating.

Another possibility to achieve the self-repair functionality of Zr-based conversion coating is to add cerium ions. Cerium(III) nitrate was added to conversion bath containing zirconyl nitrate ZrO(NO₃)₂, H₂O₂ and Ce(NO₃)₃ and deposited on AA2024.⁸¹ It was suggested that deposition of Ce oxide/hydroxide (CeO₂ and Ce(OH)₄ and Ce(OH)₃) occurs simultaneously with deposition of ZrO₂. This was reflected in smaller i_corr values exhibited by the Ce-containing coating even after 168 h immersion in 0.6 M NaCl. A simulated scratch cell test showed self-healing ability as the charge transfer resistance gradually increased for up to 336 h followed by repair of defects (Fig. 6). After 168 h in salt spray testing, these coatings were comparable to chromate conversion coatings.

Zoom In Zoom Out Reset image size

A conversion coating containing Cr(III) was found to exhibit self-repairing property,¹²⁰ which other researchers suggested might be due to local production of H₂O₂ that can then oxidize Cr(III) to Cr(VI) and lead to self-repair.^121,122 Based on that information, Zr- and Cr(III)-containing conversion coatings on AA2024-T3 were prepared to achieve self-repairing.⁷⁵ After 3 days exposure to ammonium sulfate and sodium chloride solution in an artificial scratch test cell, only Cr species were present at the surface, proving that these species were responsible for the self-repairing effect.

Summarizing, organic additives are beneficial to reduce cracking of the coating, to increase the coating density and improve adhesion to organic topcoat. Inorganic additives affect the deposition kinetics by providing additional cathodic reaction sites (Cu), blocking the active sites by formation of insoluble oxides (Mn, Mo) creating nucleation sites for Al-fluoride (Mn), improving barrier properties (Ni), promoting adhesion (phosphate), and providing a self-healing agent (V, Ce, Cr).

In general, the coating formation proceeds in 5 stages:¹²³ "clean-rinse-coat-rinse-dry". The procedures within individual stages vary from study to study. The "clean" stage consists of pre-treatment of the samples to properly prepare the surface for subsequent formation of the conversion layer. Pre-treatment may include both mechanical and chemical treatment, an intermediate rinsing step, followed by coating formation, rinsing and drying. Usually, all the steps are included in the procedure, but variations are possible. Mechanical pre-treatment in the form of grinding and/or polishing may precede chemical pre-treatment but chemical pre-treatment is often applied without prior mechanical treatment. In some cases, the surface was not mechanically pre-treated but cleaned with ethanol¹⁰² or acetone,^{55,65,68,98,99} rinsed and coated.

Chemical pre-treatment may include: alkaline degreasing or etching,^{34,36,37,42,43,50,52–54,57,60,61,63,66,67,82,85,88,92,93,96,97,108,114,119,124,125} alkaline degreasing or etching followed by acid desmutting or deoxidation,^{6,30,47,58,59,64,74–77,81,84,89,94,95,104–106,112,124} and acid cleaning or activation.^{32,35,44,46,69,70,77,79,107,124} Various commercial and non-commercial chemical formulations were used as alkaline and acid agents.

For rinsing, deionized water is usually used, although also tap water^{92,94,106,112} was applied to make the procedure simpler for industrial processing. Due to environmental and health concerns related to toxicity and carcinogenicity of chromates, procedures that reduce the worker exposure and discharge into the environment were introduced. These are so called "no-rinse" or "dried-in-place" chrome treatments.⁴² For these, the rinsing step is excluded between coating formation and drying so that the use of water and waste burden is minimized. Zr/Ti-based conversion coatings can be prepared in conventional or "dried-in-place" technology^41,44 or "ready-to-coat".¹¹²

Surface chemistry resulting from the "clean" step affects the subsequent "coat" step. Better Ti-based conversion coatings were obtained when acid cleaning was used instead of alkaline cleaning as a pre-treatment for AA5005.⁷⁷ The coating contained a higher concentration of Ti and formed a more continuous layer. Alkaline cleaning followed by acid pickling was more effective than alkaline cleaning as a pre-treatment for fluorotitanate or fluorozirconate coatings.¹²⁴ In contrast, the type of chemical pre-treatment was not crucial for chromate coatings.¹²⁴

De-alloying, which occurs at and around most intermetallic particles in the Al matrix during acid or alkaline cleaning, results in the formation of Cu-rich particles and clusters. The amount of Cu-rich particles was higher after acid pre-treatment.⁸⁷ During conversion, the Cu-rich particles grow in size and, at and around these Cu-rich zones, the thickness of the Zr-based coating is at least two times higher than the one obtained without Cu deposits. When the surface was thermally pre-treated, i.e. was covered by an oxide layer and not activated, the presence of Cu in the conversion bath could not stimulate the formation of Zr deposition process. Therefore, surface activation is crucial for subsequent coating deposition.

The hydroxyl fraction on the surface was reported to have an important effect on subsequent coating formation. Samples of varying hydroxyl fractions were prepared by different pre-treatments.^51,62,86,87 For lower OH⁻ fractions, less-developed conversion coatings were formed, whereas coating formation was promoted on samples with higher OH⁻ fractions.⁸⁶ This was related to the first stage of coating formation, dissolution of oxide from interaction of free fluoride with hydroxyl group and formation of Al−F complex. For samples with high hydroxyl fraction, this occurred readily, at an earlier stage of coating formation, allowing more time for the deposition stage within the timeframe of the coating formation.

Many reactions may take place when Al alloy is exposed to a conversion bath, leading to the formation of various compounds at the concentration of several at.% including Al₂O₃ and AlOF,^41,42 AlF₃ and AlOOH,⁴⁰ Al₂O₃ and AlF₃^46,59,69,89 Al₂O₃ and Na₃AlF₆^30,50 and AlPO₄.³⁰ An example of a depth profile of Zr-based coating formed on Al- alloy is presented in Fig. 7a.¹⁰⁸

Zoom In Zoom Out Reset image size

For Al alloys in the 1xxx, 3xxx, 5xxx and 6xxx series, the large cathodic particles are predominantly Fe-containing. A Zr-based conversion coating was found not to uniformly cover the surface of AA1050, but the oxide deposition was concentrated on and around IMPs containing mainly Fe and Si.^58,64 Most studies were carried out on AA6xxx series (Table III). In 6xxx alloys, the nucleation and growth of Zr/Ti oxide amorphous layer occurs on and around the most abundant intermetallic particles, α-(Fe,Mn)₃SiAl₁₂, which act as cathodic sites.^94,105,107 Deposition is initiated when the natural oxide layer is removed by the etching action of fluoride ions. The film grows laterally surrounding these cathodic sites, progressively covering the entire surface, but is always thicker at the IMPs. Full coverage of the substrate could not be achieved at a pH of 2.9.⁹⁵ When coated at pH 4.0, the coating extended over a broader area around the particles and over the surface.⁹⁵ Once the IMPs were covered with deposits, the cathodic activity was diminished.^94,105 The Volta difference between IMPs and matrix was reduced and eventually eliminated completely at longer deposition times (300 s) (Fig. 3).¹⁰⁷

In 2xxx series, the Cu-containing IMPs are the preferential sites of cathodic process where the subsequent precipitation of Zr/Ti-rich deposits initiates.¹⁰⁶ Cu-rich IMPs are porous due to dealloying of the other intermetallic elements during the deoxidation step (i.e. dissolution of Al, Mg, etc.). Near the Cu-rich sites, the Zr coating appeared thicker and extended more toward the AA2024-T3 matrix.¹¹²

In 7xxx series, the potential difference between the Cu-containing IMPs and matrix was found to be the driving force for the initiation of the cathodic reaction and subsequent deposition process.⁷⁴ The deposition varied along the particles, being intensified in their center and then decreased toward the particle peripheries. After deposition of a conversion coating, the Volta potential difference between Cu-rich IMPs and the surrounding matrix decreased, indicating reduced corrosion activity at the surface.

The various zinc- or zinc-alloy-coated steels (hot dip galvanized, electroplated, electrogalvanized) used as substrates for conversion coatings are listed in Table III. The process of zinc corrosion in Ti- or Zr-conversion baths is a combination of anodic zinc dissolution, and cathodic hydrogen evolution or oxygen reduction, which is the rate determining process.⁴⁹ During the initial stage of conversion layer formation, the primary zinc layer is etched and replaced by a thin Zr-based oxide/hydroxide layer.³⁴ Hydrogen evolution occurs and hydrated oxides of Zr, Ti and Zn form, accompanied by HF and H₂ evolution. HF can dissolve Zn and ZnO to form ZnF₂.⁴⁸ Native Zn oxide layer can also be dissolved by "free" hexafluorides or fluorides:⁵²

Zinc was identified in the conversion layer as ZnO,^32,48,52 Zn(OH)₂,^31,34,113 and ZnF₂.^48,51 Zn(OH)₂ is more protective than ZnO due to its lower electronic conductivity.¹²⁶ The amount of ZnO/Zn(OH)₂) in the interfacial layer between substrate and conversion layer was reported to be between 5 and 10 at.%.^31,34 An example of a depth profile of a Zr-based coating formed on HDG steel is presented in Fig. 7c.¹⁰⁸ ZnO was enriched at the inner interface with the base Zn metal, followed by predominant formation of Ti- and Zr-oxide/hydroxides also incorporating ZnF₂.⁴⁸

The formation of Zr-based conversion coatings affects the Volta potential difference between Zr oxide and zinc base metal, causing a cathodic shift.³⁴ Zr-based conversion coatings deposited on Zn-Al- alloy coated steel sheets were investigated by SKP and field emission scanning electron microscopy with energy dispersion X-ray spectroscopy (FE-SEM/EDS).⁵⁵ This alloy contains a primary Zn phase, an Al rich phase and a binary eutectic MgZn₂-Zn phase. The Zn-rich phases are the most noble while the Al-rich phases are less noble; both are more noble than Mg-rich phases. Upon immersion in conversion bath containing H₂ZrF₆ for short times (up to 40 s), the deposition started preferentially at the Zn-rich phases, which acted as local cathodes. Simultaneously, the Al- and Mg-rich phases underwent anodic dissolution leading to their enrichment with Zn. Upon longer immersion, i.e. when conditions of local alkalization were achieved, the coating also precipitated on the Mg-rich phases.

XPS spectra of untreated and H₂ZrF₆-treated low carbon steel substrate revealed that the untreated steel was covered by Fe₂O₃.⁶³ After formation of Zr-based layer, the intensity of the Fe signal decreased but was still present, indicating that Fe was incorporated in the ZrO₂ coating. An example of a depth profile of Zr-based coating formed on cold rolled steel is presented in Fig. 7b.¹⁰⁸ Analogous to other substrates, the anodic reaction on CRS surface is dissolution of iron¹¹⁴ and the cathodic oxygen reduction leads to conditions for precipitation of Zr hydroxide. The formation of FeF₃ was also reported.¹⁰⁸

A schematic presentation of deposition of conversion coatings on different substrates, which summarizes the discussion presented above, is shown in Fig. 8.⁸⁵ Auger depth profiles of Zr-based coatings on Al and HDG steel show some interesting similarities and contrasts.⁴¹ The treated HDG steel was richer in substrate elements, and poorer in the elements from the Zr conversion bath than the treated Al surface. In steel, the concentration of substrate elements was relatively constant throughout the coating, whereas they gradually increased toward the substrate for Al. This difference was ascribed to a more diffuse interface on Zn-coated steel than on Al, leading to greater coating laminar differentiation. This may be related to the effect of different ionic and electronic conductivities of the substrate metal oxide.⁸⁵ The increase in conductivity enhances the anodic dissolution reactions and, consequently, coating thickness, which was the greatest for HDG steel and the smallest on the Al alloy (Figs. 7 and 8). Note that this is different than what was found for pure metals, as mentioned above.

Zoom In Zoom Out Reset image size

The composition of conversion coatings is strongly dependent on the conversion bath composition, including additives, as well as the type of substrate and its surface chemistry. These effects were already discussed in detail above. Only the first issue, i.e. the effects of the primary composition of the conversion bath, the Ti and Zr contents, will be discussed.

Most commercial fluoro-based conversion baths are mixed in that they contain both Zr and Ti compounds (Table II). The resulting coatings contain a mixture of Zr and Ti oxides, as well as fluorides, phosphates, and other organic and inorganic compounds depending on the bath composition.^{90,94,95,105,106,127} When commercial baths contain trivalent chromium, the coating also contains Cr.^84,127 The Zr-based TecTalis coating was found to contain Zr and O at a ratio corresponding closely to stoichiometric ZrO₂.⁶

Among non-commercial conversion baths discussed in the literature, more variations in composition and ratio between reagents are found (Table I). Ti was present in the coating mainly as oxide/hydroxide but also as fluoride, and other compounds such as phosphate, carbonate, etc., depending on the conversion bath additives.^31,41 The source of Ti strongly affects the coating formation and its characteristics.⁷³ Coatings obtained from a TiCl₄ bath had better uniformity but the growth rate was more than two times slower than coatings obtained from H₂TiF₆ because of faster surface activation by fluoride than chloride ions. The smaller covalent and van der Waals radius of F⁻ in comparison to Cl⁻ results in a higher charge density for F⁻, which promotes faster surface activation than Cl⁻.^40,73 TiCl₄-based conversion coatings on CRS that were subsequently coated with epoxy exhibited better corrosion and adhesion performance than samples with H₂TiF₆-based conversion coatings, indicating that the former had better uniformity but required more time to be achieved.

Mixed Zr/Ti-based conversion coatings were investigated in several reports.^{29,46,48,50,54,59,61,69,70,74,76,107} Some formulations contained more Ti,^50,59,69,70 whereas others contained more Zr.^{48,60,74,76,127} Even when the ratio of Ti and Zr compounds in the bath was reported, it is difficult to find a relationship between their concentration in the bath and their content in the coating. When both compounds were present in the bath, both were also usually found in the coating^{46,48,59,69,70} but not necessarily in the same ratio as in the bath.^46,48,50 In Zr-based coatings, Zr was present mainly as ZrO₂^49,55 or as ZrO_2-x(OH)_2x⋅nH₂O.³⁸ In some studies, fluoride was detected within the coating,^{30,34,36,40,46,49,51,52,75} and in other studies it was not detected.^55,58,63

The concentration of H₂ZrF₆ and H₂TiF₆ in the conversion bath affects the Zr and Ti concentration in in the coating, but also the concentration of F in the coating. Increasing the concentration of H₂ZrF₆ from 0.001 M to 0.1 M in conversion baths resulted in decreased Zr coating content from 10 to 1.3 at.%.³⁸ Similarly, for a H₂TiF₆ conversion bath, the highest Ti content was obtained was obtained at a concentration of 0.0001 M.³⁸ The decrease in Ti or Zr concentration in the layer deposited from more concentrated baths was accompanied by increased concentration of F. These changes in coating composition are related to the lower pH of more concentrated solutions and the fact that at higher concentration the fluoride ions became too aggressive toward the substrate and thus detrimentally affect the layer formation.^38,71 Low H₂ZrF₆ and H₂TiF₆ concentrations favor the formation of a protective Zr- and Ti-layer. Optimal corrosion performance was noticed at a concentration of 100 mg/L H₂ZrF₆⁵⁸ and 0.01 M H₂ZrF₆.⁶⁷

The coating morphology when analyzed by SEM follows the morphology of underlying, pre-treated substrate, but exhibits a somewhat less rough appearance.^{47,53,64,77,83,104} The coatings have globular structure with nanometer-sized particulates.^{30,40,42,46,48,50,56,58,61,74,76,77,95,97,98,114} The deposit at intermetallic particles is thicker, compared to the rest of surface.^{88,95,105,107}

The coating may be non-uniform due to the fact that it starts at IMPs in Al alloys or different phases at the surface of steels, and then spreads across the surface.^{38,58,71,74,102,105,106} After sufficient immersion time (Figs. 3 and 4) the coating covers the surface uniformly on Al alloys,^{69,81,107,112} HDG steel,^52,68,72 and CRS.^{6,73,82,96,97} In some cases, the coatings have not covered the surface uniformly even after prolonged immersion because the coating coarsens.^89,37,60

Cracks were observed at the coating surface, especially after long immersion times. Cracks can form during exposure to the vacuum of an SEM chamber because of dehydration and shrinkage, but also can form before introduction into the SEM vacuum during exposure in dry air^{49,52,69,79,80,94,96,106,127} or during a curing step at 70°C.¹⁰⁷ Cracking was also observed in coatings deposited in conversion baths of higher temperature (40°C).⁹⁷ The addition of Ce to Zr-based coatings on AA2024 induced submicron cracks that reached the substrate surface, which may account for the coating self-healing ability.⁸¹

The majority of Zr/Ti based conversion coatings are in the category of thin coatings, between 10 nm and 50–80 nm.^{6,32–34,44,49,52,76,77,88,90,91,97,102,105,106,108,112} Coatings with thickness between 100 nm and 600 nm were reported as well,^48,81,94,127 and even in the micrometer thickness range from 1.0 μm up to 6 μm.^{30,37,43,59,61,80,119,127} Thickness is related to other parameters such as substrate, pH, time of immersion and stirring, as discussed in Conversion bath parameters and Effect of substrate on coating formation sections.

As mentioned above, the coating deposition starts at the intermetallics and progressively grows on and around them eventually covering the entire surface, so the coating is considerably thicker at the IMPs, as indicated by SEM/EDS profiles and mapping and Auger spectroscopy.^94,108 This is different from chromate coatings, which are formed by redox reaction of hexavalent chromium from the solution at the surface and are thicker at the matrix and thinner at IMPs.⁹⁴

The corrosion performance of Ti-, mixed Ti/Zr- and Zr-based coatings on aluminum alloys differs depending on the type of coating and substrate. Furthermore, additives to conversion bath can greatly affect the final performance. Ti-based coatings were investigated on Al 5005,⁷⁷ AA3003⁴⁴ and AA6014¹⁰⁴ and showed limited protection, although improved adhesion of a subsequent organic coating was observed on Al 5005.⁷⁷

Mixed Ti/Zr-based commercial coatings did not provide strong protection of Al alloys. The Gardobond coatings X4707 (Ti/Zr) were not efficient for AA6060.^94,95 Similarly, Bonderite M-NT 5200 (Zr/Ti-based) provided almost no protection for AA2024 substrate, which was ascribed to non-uniformity or defects in the coating.¹⁰⁶ Alodine 5700 (Ti/Zr based) deposited on AA2024-T3 had unacceptable performance in the ASTM B117 salt spray test (less than 200 h) and marginal performance in the GM 9540P accelerated corrosion test.⁸⁴ Alodine 4830 (Ti- or Zr-based) coatings deposited on AA6061 achieved optimal corrosion performance when deposited at pH 4.5 for 90 s.⁸⁹ However, the protection was limited during 24 h immersion in 3.5 wt% NaCl. Under the same conditions, chromate coatings remained protective.

Various mixed Ti/Zr non-commercial coatings showed a greater ability to protect Al alloys. When containing different additions such as tannic acid,^50,59,70,76 Mo,⁷⁴ Mn^61,74 and phosphates,³⁰ better performance was achieved than for Ti-based coatings. The additions of Mo and Mn were beneficial for corrosion resistance in an Na₂SO₄+NaCl solution for Zr/Ti coatings deposited on AA2024 and, especially, on AA7075 (Fig. 9a).⁷⁴ Good performance of mixed Ti/Zr-based coatings was also found on AA6063-T52^42,50,61 and AA6016.¹⁰⁷ Excellent performance was reported for Zr/Ti-coatings containing tannic acid and metavanadate deposited on AA6063.^59,70 When coated, the E_corr in aerated NaCl solution shifted more positive by 430 mV, and i_corr was reduced from 7.5 to 0.6 mA cm⁻² (Fig. 9b).⁷⁰ Accelerated testing showed superior corrosion performance of the coating compared to uncoated alloy and commercial chromate coating (DCHZ-405).⁷⁰ After 72 h, many white corrosion products were observed on uncoated AA6063, indicating severe corrosion.

Zoom In Zoom Out Reset image size

Zr-based coatings showed better performance compared to Ti coatings. For AA1050, good electrochemical results were achieved for Zr-based coatings containing PAA and PAM in that they exhibited constant capacitance during more than 50 days exposure to 3.5% NaCl.⁴⁷ When tested in a salt spray chamber for 1000 h, however, these coatings performed worse than commercial chromate coatings.⁴⁷ Zr-based coating containing phosphate deposited on AA5083 exhibited much better corrosion performance that that of commercial Cr-free treatments and close to that of commercial chromate coatings (Fig. 9c).³⁰ Both cathodic and anodic processes were retarded on the Zr coating, and the extent of the passive range was broader for the Zr conversion coating than for CCC. When tested in a salt spray chamber for 1500 h, Zr-coating performed slightly worse than commercial CCC but better than Cr-free coating.

NCP Zr/Zn conversion coatings from NAVAIR were tested on AA2024-T3, AA6061-T6 and AA7075-T6.¹¹² The coating shifted the E_corr in the positive direction by about 250 mV, suppressed anodic currents more than cathodic currents around E_corr by at least a factor of 10 ×, and shifted the pitting potential in the noble direction (Fig. 9d). The coating also provided excellent corrosion protection to AA2024-T3, AA6061-T6 and AA7075-T6 alloys during a 14-day full immersion test in Na₂SO₄ + NaCl. However, during neutral salt spray and thin-layer mist tests, the NCP provided little stand-alone protection to the Al alloys and was inferior to a TCP coating. After a 1-week beach exposure, NCP coatings on AA2024-T3, AA6061-T6 and AA7075-T6 showed greater levels of pitting, spotting and surface damage than TCP coatings. These results demonstrated that electrochemical evaluation under immersion conditions does not always reflect coating performance during accelerated degradation or environmental exposure.¹¹² It was reported recently that PreCoat A32 coating consisting of Zr(IV), Cr(III) and fluorides achieved similar performance in a 168 h salt spray test as Alodine 1200 CCC, indicating that it is a promising alternative for aeronautic and aerospace industries.¹²⁸

Self-repair ability was studied in several reports. Vanadate-containing Zr-based coatings on AA6063 promoted self-healing as indicated by EIS spectra in 3.5% NaCl.⁵³ Without Ce(NO₃)₃, Zr-based conversion coating deposited on AA2024 improved the corrosion resistance only to a limited extent. However, when Ce(NO₃)₃ was added, the long-term results were greatly improved.⁸¹ The self-healing action of cerium was reflected in a continuous increase in charge transfer resistance during immersion (Fig. 6). When exposed to salt spray test for 168 h, the Zr-coating underwent heavy attack, whereas the Zr-Ce coating exhibited only surface discoloration after 168 h exposure with only few isolated pits, which was comparable behavior to chromate coatings.⁸¹ A Zr-based coating containing Cr(III) on AA2024-T3 exhibited self-repair, as shown by EIS experiments in an artificial scratch cell.⁷⁵ The coating persisted up to 72 h salt spray chamber when corrosion products appeared.⁷⁵ Uncoated specimens showed corrosion after only 6 h of exposure.

Summarizing, some Ti/Zr and Zr-based coatings containing various additives achieved exceptional performance on aluminum alloys. For example, Ti/Zr-based coatings containing metavanadate and tannic acid on AA6063 showed better results in a salt spray chamber than CCC,^59,70 whereas a Zr-based coating containing phosphate on AA5083 exhibited results similar to CCC.³⁰ Further improvements are required for some coatings. For example, Zr-based coatings containing Mo and Mn on AA2024 and AA7075 showed excellent electrochemical results but salt spray results were still worse than CCC.⁷⁴

Ti-, Zr-based and mixed TiZr coatings were investigated on galvanized steels. Ti-based coatings on galvanized steel prepared from TiCl₄ and H₂SiF₆ exhibited much larger impedance compared to chromate coating and uncoated substrate and developed less white rust when tested in a salt spray chamber for 240 h.³⁵ Similarly, coatings based on TiCl₃ and H₂SiF₆ deposited on Zn electroplated steel were superior to chromate coatings (Fig. 10a).⁷⁹ Even after 216 h of salt spray, no signs of zinc corrosion were observed. A Ti-based coating prepared from baths containing H₃PO₄, Mn₃(PO₄)₂ and organic phase,^31,72 and Mn(NO₃)₂ and H₃PO₄³² on HDG steel resulted in about one order of magnitude smaller i_corr and a slight positive shift of E_corr in 3.5% NaCl.^31,72 When using another Ti-based coating on Zn-coated steel, zinc corrosion products appeared on only 5% of the surface, whereas uncoated HDG steel was severely corroded.³²

Zoom In Zoom Out Reset image size

Two types of mixed Ti/Zr coatings were tested on galvanized steels and showed relatively good performance.^48,54 Good results were also reported for a commercial Zr-based coating (Bonder D 6800) developed in 1990s⁴³ and recently for Zr-based coatings containing Cu or Fe on HDG steel.⁵² Zr-based coatings with different additives have been investigated. When Ni ions were present in a Zr conversion bath, the corrosion performance was not improved in comparison to the Zr-based coating without Ni (Fig. 10b).⁶⁸ However, when a Ni coating was formed on top of a Zr-based coating, superior performance was achieved. During immersion in NiSO₄ solution, a more resistive and uniform coating was formed.

Cathodic and anodic polarization curves were measured in 0.5 M NaCl for hybrid coatings with a Zr-rich inorganic matrix containing CAP organic polymer beads and V deposited on HDG steel.¹¹⁹ The curves for coated samples were generally shifted to lower current densities with little change in E_corr, indicating protection by simple blocking of the surface (Fig. 10c). When the coating also contained vanadium, the current densities were lower, proving its beneficial effect.

Summarizing, some coatings containing additives like Mn and V exhibited good performance on galvanized steels and achieved performance similar to CCCs.^35,79 There are still possibilities for further improvement, especially in terms of additives and self-healing agents, as well as chemical post-treatment.

Ti- and Zr-based coatings on steels were investigated. The corrosion resistance of Ti-based coatings deposited on CRS with various conversion parameters (time of immersion, pH, Ti concentration) were tested in 3.5 wt% NaCl.^67,73 The Ti source was H₂TiF₆ or TiCl₄. Coatings obtained from TiCl₄ achieved better uniformity and consequently better corrosion resistance. Using optimized conversion bath parameters, the R_p increased, and i_corr decreased by approximately 6–7 x.⁷³

TiCl₄-based coatings containing NiSO₄ or MoO₄ were also tested on CRS in 3.5 wt% NaCl solution (Fig. 11a).⁸² The significant improvement when Ni was added was ascribed to the formation of an outer layer of nickel, which acts as a physical barrier to inhibit the charge transfer and movement of ions. The detrimental effect of Mo was ascribed to the formation of a cracked Mo-based layer at the substrate/coating interface. Chemical post-treatment in NiSO₄ solution increased the R_p about 80 times compared to uncoated steel and 50 times compared to TiNi-coated steel. This result points to further possibilities to improve corrosion performance of coated substrates, similar to what was observed for Zn galvanized steel.⁶⁸

Zoom In Zoom Out Reset image size

Commercial Bonderite NT-1 Zr-based coatings on CRS shifted the E_corr more positive by 200 mV and reduced i_corr 2–3 x compared to uncoated steel in 3.5 wt% NaCl.^96,97 In contrast, it was reported that Bonderite NT-1 coatings when deposited on mild steel did not show any substantial improvement of corrosion characteristics compared to the uncoated substrate.⁹⁸

Results published for Zr-based non-commercial coatings are diverse: Zr-based coatings containing MnSO₄ deposited on carbon steel in 3.5 %wt NaCl were similar to the uncoated samples, but with a smaller i_corr by approximately 2 x and a slight negative shift of E_corr.⁵⁷ Better performance of Zr-coated carbon steel was achieved in Na₂SO₄.¹²⁵ At longer immersion time in the conversion bath, the resistance increased due to the formation of a more uniform ZrO₂ layer (Fig. 11b). Similar results were obtained for low carbon steel⁶³ and mild carbon steel in NaCl solution.⁶⁷

Summarizing, conversion coatings on steels achieved up to one order of magnitude smaller i_corr values and in most cases acted as barrier coatings, i.e. without changing the shape of polarization curves. Possibilities of long-term protection should be improved, presumably using additives or chemical post-treatments.

Different topcoats on Al alloys have been tested, including polyester,^41,104 acrylic,^41,107 and epoxy.^{30,47,58,64,69,70,77,93,95} A Zr-based coating on AA1050 enhanced the pull-off adhesion of an epoxy topcoat.⁵⁸ This was ascribed to increased surface roughness and surface energy, which promote wettability of the substrate and stronger adhesive bonds.⁶⁴ Dry and recovery adhesion (after 20 days immersion in 3.5 wt% NaCl) were measured by the pull-off adhesion test. Adhesion loss was calculated according to:

The dry adhesion strength of epoxy topcoat applied on a Zr-coated sample was higher than that measured on a chemically treated uncoated substrate. The adhesion loss was the lowest for the Zr-coated sample. Excellent adhesion was reflected also in high corrosion performance during 60 days immersion in 3.5 wt% NaCl.⁶⁴

Conversion coating enhances corrosion resistance and increases paint adhesion by two mechanisms: (i) changing surface chemistry, and (ii) changing topography.^64,134 Increased wettability by increased surface energy for conversion-coated samples helps produce strong adhesion bonds with the organic coating through their polar groups. Most metal oxides, which are the main constituents of conversion coatings, can produce strong chemical bonding with highly polar groups of organic topcoats, like hydroxyls. A recent study showed that the bonding between Zr-conversion coatings on different metals and amide-functionalized molecules results from different bonding mechanisms: the increased surface hydroxide density after Zr-based coating on Al and Zn promotes Bronsted interactions, whereas Mg was characterized by fast Zr deposition kinetics accompanied by hydroxide removal.¹³⁵ Chemical bonding might not be sufficient for long-term high bond strength. Increased surface roughness by the conversion coating treatment enables the liquid applied organic coating to penetrate into the pores of the surface, thereby enhancing adhesion through mechanical interlocking. Adhesion strength increased with roughness due to an increase in area available for bonding at the primer/substrate interface, the tortuosity of path crack propagation, and an increase in plastic deformation.¹³⁴ Randomly abraded samples had higher adhesion strength than aligned abraded samples.¹³⁴

Several examples of good performance of Al alloys with an organic topcoat have been reported: epoxy coated Zr-based coatings containing phosphate deposited on AA5083,³⁰ acrylic coated Zr/Ti-pre-treated AA6016,¹⁰⁷ polyester coated Ti-based silane coatings on AA6014¹⁰⁴ and epoxy coated Ti/Zr/V coating containing tannic acid on AA6063.⁶⁹ A Zr-treated (TecTalis 1800) AA6061-T6 substrate had higher resistance to moisture-induced corrosion than an untreated substrate, which was indicated by a smaller increase in roughness of the treated alloys with prolonged exposure to 99.5% relative humidity.⁹³ In situ atomic force microscope (AFM) scratching indicated that removal of an epoxy topcoat was more difficult for pre-treated substrates because of the enhanced adhesion strength.

The propagation of corrosive de-adhesion of the organic lacquer on a Zr-based coating was dependent on the coating conversion times.³⁴ For shorter conversion time, oxygen reduction was impeded and the Volta potential was close to that of galvanized steel. Thus, a driving force for cathodic delamination did not exist. With increasing conversion time, propagation was faster due to the higher defect density of the conversion layer.³⁴

The adhesion of an epoxy topcoat on Zr-based coating containing Ni and deposited on galvanized steel was investigated.⁶⁸ Bond strength was higher than for the bare sample. The addition of Ni to the Zr coating had a negative effect on the adhesion strength, but the highest adhesion strength was found when a Ni layer was deposited onto the Zr coating. This improvement was ascribed to enhanced surface roughness, which enabled mechanical interlocking between coating and topcoat.

Mainly epoxy^{56,57,65,67,73,98,99,114,125} and polyurethane⁶³ topcoats were tested on steels. A coating on CRS prepared from a TiCl₄ bath exhibited higher adhesion strength to epoxy than that prepared from a bath containing H₂TiF₆.⁷³ Both exhibited higher strength compared to epoxy coating deposited on the bare CRS substrate.

Dry and wet adhesion properties of epoxy coated commercial Zr-based coatings on carbon steel were investigated (Fig. 12).⁹⁹ Although the impedance decreased with increasing exposure time (up to 250 h), there was no second time constant in the phase angle, which would indicate a change from capacitive to predominantly resistive behavior, as observed for uncoated substrate. Both dry and wet adhesion were higher than for untreated carbon steel even after 70 days exposure to 0.1 M NaCl. Similarly, good performance was achieved by TecTalis (Zr-based and Cu-containing) coated and painted (e-coat) CR steel samples up to 120 days immersion in 0.5 M NaCl.⁶ High impedance values around 1.4 × 10¹⁰ Ω cm² were measured. Pull-off tests, cathodic disbonding tests and EIS measurements indicated that a Zr-based coating on mild steel significantly improved adhesion of an epoxy topcoat.⁵⁶ The Zr coating increased the surface roughness (from 70 to 98 nm), which improved the adhesion of an organic topcoat and cathodic disbonding resistance as measured by EIS during 192 h in 3.5 wt% NaCl.

Zoom In Zoom Out Reset image size

The additives to the conversion coating bath may be beneficial for corrosion resistance but can degrade adhesion. For example, the addition of Mn²⁺ ions improved the corrosion resistance of Zr-based coatings deposited on carbon steel, but the coating without Mn²⁺ showed the highest adhesion strength.⁵⁷ Therefore, while Mn may modify conversion coating characteristics, it cannot necessarily improve corrosion performance and adhesion in an organic coated system.

A polyurethane topcoat on TecTalis Zr-based coating lowered the cathodic delamination on low carbon steel.⁶³ Volta potential profiles of the uncoated carbon steel and epoxy-coated steel with different pre-treatment time in Zr-conversion bath are presented in Fig. 13.¹²⁵ Two potential levels characterize the SKP profiles: the lower potential on the left associated with the NaCl solution reservoir in the artificial defect and the higher potential of an intact coating on the right. When a coating delaminates from the substrate, the potential under the coating shifts toward the defect potential. The sharp change in potential marks the position of the end of the delamination zone, and the progress of cathodic delamination is reflected by the location of this potential jump with time. The average cathodic delamination rate decreased as a function of pre-treatment time. A decrease in pH resulted in a systemic decrease in the cathodic delamination rate of epoxy coating from untreated steel substrates via the decreased concentration of hydroxyl ions at the delamination front. When the steel was pre-treated using Zr-coating, this effect was not observed.

Zoom In Zoom Out Reset image size

Summarizing, increased adhesion strength and reduced cathodic delamination are the most important benefits of Ti- and Zr-based conversion coatings. Some additives to the coating bath do not necessarily improve the adhesion strength, e.g. Ni and Mn, although they increase corrosion resistance. Therefore, the addition of additives should be optimized to improve both corrosion and adhesion performance.