Volume 2 Paper 2

Amorphous Zirconia Films as Adhesion Promoter for Organic Coating

F. J. Rodriguez§, L. Fedrizzi*, R. Di Maggio and S. Rossi 
Dipartimento di Ingegneria dei Materiali, Universitą degli Studi di Trento, Via Mesiano, 77, 38050 Trento, Italy,  
* Dipartimento di ICMMPM, Universitą di Roma I "La Sapienza", Via Eudossiana, Roma, Italy
§ Depto. Ingenieria Metalurgica, Facultad de Quimica, U.N.A.M. 04510 Mexico D.F., Mexico,  

Abstract

The two main features of a protective coating are the adhesion and the anticorrosive protection. In order to improve both, chemical pre-treatments have been used. The use of chromate was very popular, but recently they have been highly restricted because of their toxicity, so that chromate-free pre-treatments have been developed and tested. An interesting alternative seems to be the deposition on the metallic surface of thin layers of zirconia by sol-gel process. In this study thin films of amorphous zirconia on low carbon steel sheets have been obtained by the dip-coating technique. The behaviour of the zirconia films as promoters of adhesion was evaluated measuring detachment of cross-scratched paint coated samples after Salt Fog Chamber Test. In addition, the determination of adhesion of organic coatings was performed by swelling in N-methyl pyrrolidone. According to the results, the samples pretreated by zirconia layers showed a good performance, in comparison with commercial chemical treatments (tricationic phosphate and iron phosphate). The protective properties of the zirconia films were found strongly depending on the process parameters, e.g. concentration of the precursors solutions and chelating agents, which rule the thickness and the organic residuals amount of the amorphous zirconia. In order to assess the behaviour of these sol-gel systems, preliminary tests using Electrochemical Impedance Spectroscopy were carried out too.

Keywords: Organic coatings, Zirconia film, Adhesion, Pre-treatment.

Introduction

The most important properties of a protective coating are the anticorrosive action [1] and the adhesion to the substrate. The coatings act as barriers to aggressive chemical species, resulting also active against corrosion because of the chromate or phosphate present as pigments. However the coating protection fails, if the paint does not adhere to the substrate adequately [1, 2]. At the aim of improving adhesion properties of organic coatings on steel, chemical pre-treatments have been used for a long. Among these chromate and phosphate pre-treatments showed to be quite useful. However, by law, many countries have forbidden the use of toxic pre-treatments (chromate-like), in spite of their good anticorrosive properties. Because of that, industry is deeply concerned about the development of a chemical process able to substitute chromate pre-treatments [3]. On steel it has been employed iron or zinc or tricationic (Zn, Mn, Fe) phosphate [4]. The use of phosphate has shown good results on galvanised steel. So, it improves the adhesion of the paint film and the anticorrosive protection. But phosphates are not the best option in ecological terms. Other chromate-free pre-treatments have been developed and tested. The choice of the chemical conversion treatment must be based on several criteria: economical, ecological, technical, etc. Some of them have a chromate-like behaviour, but are not economically attractive. The application of zirconium oxide as a pre-treatment on steel appears a promising compromise. On the other hand, ZrO2-CeO2 and other oxides films, put down on stainless steel by sol-gel techniques, resulted protective against dry and wet corrosion [5,6]. Sol-gel is a process to obtain films of metallic oxides in very mild conditions, but the reproducibility of thickness and the quality of them depend on several parameters affecting the precursors solution [7]. As an example the presence of complexing agent has already been shown to be important in improving the stability of solution of metal alkoxides, very moisture sensitive compounds [8-10]. In this study it was evaluated if the use of two different complexing reagents and the control of the hydrolysis allow the formation of ZrO2 films suitable as pretreatments promoting adhesion of organic coatings. The zirconia applied by sol-gel on low carbon steel as an adhesion promoter instead of chromate and phosphate looks promising, however a technological application cannot disregard the evaluation of economical advantages.

Experimental

Low carbon steel was used as substrate. 7.5cm x 10 cm samples were first degreased with organic solvents, then dipped into solutions of zirconium tetrabutoxide in anyhdrous butanol and withdrew at constant rate of 3 cm/s, obtaining thin films of amorphous zirconia. After the deposition, the film underwent thermal treatments in oven at temperature between 150-250°C, in order to densify and remove the most organic load. Repeating the deposition run, the thickness could be increased. The chelating agents employed in this study are acetic acid and acetylacetone, which have been selected based on previous study [8,9]. The hydrolysis and therefore the formation of the film was made by reacting in controlled humidity chamber (70% RH) after each dipping, except in series 7 and 8 where this step was made using boiling water. When the total thickness was obtained, the samples were treated at 150 or 250°C In Table I are reported the concentration of precursor, chelating agent, number of layers, temperature and time of the thermal treatment. On comparison, the thickness of zirconia films on glasses, prepared in the same conditions was measured by using a profilometer Dektak 3 [6]. The value was around 0.60μm, but for series 4 and 5 it increased up to about 1.0 μm. The series 9, 10 and 11 were used as reference samples. Degreased steel was coated in order to get a blank with poor adhesion between coating and substrate. Tricationic phosphate is a common chemical treatment, which highly improves the adhesion of the coating to the substrate as well as corrosion resistance. Iron phosphate has been studied because of its use on the industry due to low cost. 

Table I.

Series Alkoxide Molarity (g/mole) N° layers Chelating agent Temperature (°C) Time (min) Comment
1 0.3 3 Acetic acid 250 30 ---
2 0.3 3 Acetic acid 250 10 ---
3 0.3 3 Acetic acid 150 30 ---
4 0.3 5 Acetic acid 150 30 ---
5 0.6 3 Acetic acid 150 30 ---
6 0.3 3 Acetylacetone 250 30 ---
7 0.3 3 Acetylacetone 250 30 Treated in boiling water
8 0.3 3 Acetylacetone 150 30 Treated in boiling water
9   Degreased steel
10   Tricationic phosphate
11   Iron phosphate
 

A multilayer polyester paint coating formed by a primer and a topcoat was applied in all cases. The average dry film thickness (DFT) was about 30 μm. This kind of paint is usually used for electrical appliances [4]. A part of the coated samples were cross-scratched and tested evaluating the damage in Salt Fog Chamber (ASTM B117). In order to assess the adhesion promoted by the pre-treatment, quantitative measurements were made using a pull-off test (Sebastian IV) and measurements of the swelling of the film using N-methyl pyrrolidone [11]. The FTIR analyses were performed by a Spectrometer BioRad FTS 165, coupled with a BioRad UMA 250 device. In addition, Electrochemical Impedance Spectroscopy was used in order to monitor the increase in wet area and some other interfacial phenomena. A three-electrode cell was used, constituted by: Working electrode: Pre-treated and painted samples Counter electrode: Platinum sheet Reference electrode: Ag/AgCl (207 mV vs SHE) The electrolyte was 0.3%wt. Na2SO4, because of its characteristics of conductivity and low aggressiveness. The equipment used was a EG&G PAR 273 potentiostat with a Schlumberger Solartron Frequency Response Analyser 1255. The results were registered with M398 software and treated with the EQUIVCRT software [12].

Results and discussion

The gel structure is determined from both the concentration of the alkoxides solution, and temperature or length of the thermal treatments. The hydrolysis and condensation of alkoxides of zirconium, similarly to those of other metallic species, easily take place in the solution, leading to a gel. This is constituted of an amorphous network of metallic oxide, even if discontinuous because of the presence of organic or hydroxylic functionalities. Being the hydrolysis of the alkoxides of zirconium faster than condensation, the gel could be formed slowing down the first reaction by adding complexing agent as acetylacetone or acetic acid. Moreover higher the molarity of alkoxide in the solution, higher the thickness of the deposited layers. The strength and the skeletal density of the gel as deposited improves after thermal treatment. The densification of the gel is accompanied by noticeable shrinkage, which could also cause cracks formation. During the heating, the most of the residuals further react leading to a more interconnected metallic oxide. In fact one of the advantages of using sol-gel process is the mild treatment necessary to form a ceramic layer. Anyway, a few organic residuals, especially the residuals deriving from the complexing substances, were always present in the film, unless the temperature of crystallization of the zirconia has been reached during the heating in air [10]. The IR spectra of all samples showed the presence of not reacted organic linked to the zirconia matrix. Increasing the maximum temperature from 150 to 250 °C the reduction of organic load is remarkable in the spectra of the samples 1 and 3 of Figure 1.

 

Figure 1. Spectra IR of zirconia samples 1 and 3. 

Salt Fog Chamber (First Part) 

The results obtained in Salt Fog Chamber test are shown in Table II. The detachment near the scratch was measured by gently tearing off the disbonded area with a knife. In some cases blisters were also observed. The disbondment was not uniform along the cross scratch, so the data reported in the table are average values. In Table II also the maximum and the minimum were shown, when the range was too wide. These data evidenciated the adhesion properties of zirconia films depend on several parameters. Time and temperature of thermal treatment appear to be the most important of them. As discussed above, the heating promotes the reduction of organic load, so that the presence of a very few residuals seems to be advantageous to the adhesion. In order to maintain low the amount of organic residuals when the thickness of the films increases, the time or the temperature of the treatment have to be increased too. On the other hand, the zirconia prepared from solution containing the acetylacetone showed the worst behaviour among the sol-gel samples, even when they underwent the same thermal treatment. The stronger chelating action of acetylacetone with respect acid and its very difficult hydrolysis could account for these findings. The acetylacetonato groups bonded to zirconium prevent the condensation and the formation of an oxide network [11]. At the aim of to promote the hydrolysis, some of the samples were also hydrolysed in boiling acidic water, but the recorded improvement was too small and not sufficient to consider interesting the way. In fact that condition did not enhance the condensation accordingly, and the surface of the samples appeared dusty and not continuous, through observation with a stereoscopic microscope. Finally the microstructure of the samples obtained from solution containing the acid appeared more suitable to adhere to the organic coating. 

Table II.

Series Detachment after 7 days (mm) Detachment after 21 days (mm) Comments
1 0.3 0.7 (0.6 to 1) No blistering
2 0.4 1.2 (1 to 1.3) No blistering
3 0.5 (0.3 to 0.8) 2 (1.8 to 2.5) Small blistering near the scratch
4 0.8 (0.5 to 1) 3.5 (3 to 4) Total detachment after 21 days. There were large blisters all over the samples.
5 0.4 0.6 (0.5 to 1) It was difficult to detach
6 1.5 (1 to 2) --- After 7 days, all the paint was detached
7 0.6 (0.2 to 1) 1.8 (1 to 2) Different detachment depending on the position in the cross scratch. Blistering
8 1.1 (0.5 to 1.8) --- Disbonded area is brittle. Blistering
9 6.5 --- Blistering all over the sample
10 0 0.1 (0.1 to 0.2) It was difficult to detach
11 0.4 2.5 (2 to 3) There was blistering mainly near the cross-scratch
 

It is remarkable that the series 5 along with series 1 resulted the best in promoting adhesion. Moreover, the series 5 resulted even better than the other at longer time of analysis. The use of a more concentrated solution of alkoxide seems to be positive for a good performance of the zirconia film. At first glance this result could appear odd. Anyway the presence of a great number of residuals, in the film prepared from a concentrated solution, means they are closer than in a layer obtained from a diluted one. That configuration is likely more favourable to give condensation reactions, resulting a dense and thick metallic oxide layer. A ranking of the adhesion behaviour in Salt Spray Fog of our samples could be done. As regard the reference samples, those without pre-treatment resulted obviously not good at all. The use of iron phosphate as a pre-treatment shows good results in short times of exposure (7 days), compared with the zirconia film in series 1, 5 and 2, but it is not a good option in long service life. Since now, tricationic phosphate looks the best option .

Salt Fog Chamber Test (Second Part)

The results previous discussed brought about the further study was restricted to samples of series 1 and 5. Different sheet were coated by zirconia and another polyester paint system (also used on domestic electric appliences). Moreover, in order to prove without any doubt the effect of the heat treatment, other samples have also been prepared, the features of which are reported in Table III, along with the results of Salt Fog Chamber test. The results were even better than those expected. All the samples with sol-gel films applied and hydrolysed without the thermal treatment showed a significant detachment in short-time exposure. Besides of the effect of the new paint, it is noteworthy the better performance of the sol-gel pre-treatment with respect the iron phosphate, even though not as good as the tricationic posphate. Moreover the behaviour of these new series maintains very good also after large time of exposure in Salt Fog Spray. The results of the test of reference samples had shown once more how important is the pre-treatment of the steel in promoting adhesion. The low corrosion and the negligible blistering phenomenon observed for the samples 5" and 5 proved the reactions of condensation take place fastly in a concentrated solution. So that a thicker and stiffer layer of gel could be obtained from those solutions, which can take more advantage from the thermal treatment already at low temperature. 

Table III.

Series Solution Molarity, Tmax and time of the thermal treatment Loss of adhesion (detachment from the scratch in mm) Comments*
    7 days 13 days 20 days 27 days  
1 0.3M, 250°C, 30’ 0 0 < 0.5 < 0.5 Few blisters and corrosion in the scratch
1’ 0.3M, no heat treatment 0 5 - - Completely covered with small blister
5 0.6M, 150°C, 30’ 0 0 < 0.5 0.5 Blisters and some points of rust in the scratch
5’ 0.6M, no heat treatment 0 2 – 3 - - Completely covered with small blister
5" 0.6M, 250°C, 30’ 0 0 < 0.5 < 0.5 Few small blisters and some points of rust in the scratch
9 Degreased steel 9 – 10 - - - Large blisters near the cross-scratch
10 Tricationic phosphate 0 0 0 0 First blistering after 27 days (small blisters). Few points in the scratch
11 Iron phosphate 4 - - - Large blisters near the cross-scratch.
* The comment refers to the appearance of the samples on the 27th day or when retired.
- Means that the samples were retired because of the total detachment of the coating.

Swelling with N-methyl pyrrolidone (NMP)

Van Ooij et al [7] have proposed the measure of the adhesion of the paint coating on the base of its swelling in N-methyl pyrrolidone (NMP). In fact the swelling produces stress, which could detach the paint from substrate. This test of adhesion of a film to metal depends on the thickness of film, the nature of the coating, and especially the pre-treatment on substrate surface. The results of this test are expressed as time of total detachment of the organic film in NMP. This parameter is named NMPRT (N-methyl pyrrolidone retention time). The test is based on the hypothesis that NMP acts on organic coating, disregarding substrate and pre-treatment. Actually the presence of organic residuals in the zirconia films of this study, even if negligible, could invalidate the assumption, so that the detachment could be due both to the stresses induced by swelling and a chemical-physical interaction between NMP and the zirconia films. In order to clarify this effect FTIR analyses on unpainted samples, before and after the immersion in NMP at 60°C, have been performed. In some cases (series 1) the organic residual part in the zirconia thin film was highly removed (Figure 2) by NMP, but in the series 5, the difference were not detectable before and after immersion (Figure 3), probably due to the greater thickness of these samples. It can also evidenciated the presence of the NMP adsorbed on the films, from a signal at about 1700 cm-1 referring to C=O bonds. So, it seems that the remotion of the organic components of the zirconia thin film by the NMP affects the results of this test.

Figure 2. FTIR spectra of series 1 before (A) and after immersion in NMP (B).

 

Figure 3. FTIR spectra of series 5 before (A) and after immersion in NMP (B).

There was no total detachment of the coating on the samples pre-treated with phosphate, but after 2 hours we could appreciate the formation of blisters reaching a 4M and 7M degree (according to ASTM D 714), for the iron and tricationic phosphates, respectively. Anyway the results showed again that the zirconia films, even if are not as good as tricationic phosphate, are better than the degreased steel.

Pull-off test

The tests carried out with a pull-off device (Sebastian IV) have not been useful in this research, because of the type of failure. The failure mode was decohesive in the topcoat, and we did not see the primer. As a result we just can say that the mechanical features of the coating are poor or inferior to the substrate adhesion.

Impedance measurements

Impedance measurements were performed in order to follow the progressive loss of adhesion between the coating and the pretreatment or substrate. Because of the high quality of the organic paint, for all the samples only a capacitive behaviour could be recorded, so that the coating system do not allow us to evaluate the differences existing among the pre-treatments. At the aim of overtaking the impasse, some defects have been introduced on the organic coating by using a metallic pin. The size of the defects was almost the same in all samples (≈150 μm diameter), according to the measurements through microscopical observations. That value was also confirmed by the analyses of the first loop in the impedance diagrams (Figure 4), which allow to determine the area of the defect. Generally the whole Nyquist diagrams show two semicircles at the beginning of the experiment, and a third one after few hours (Figure 5). The fitting of the impedance results was performed by Equivcrt software [9]. In the high frequency semicircle, the simulation have yielded values, which could be associated with the properties of the coating as well as the size of the defect. Anyway in absence of a model, the interpretation of the rest of the diagram is not easy and still in progress. As an example, the semicircle with a maximum frequency at 3.9 Hz in Figure 5 can be deconvolved in two components. Provoking larger defects (300 μm) on the coated surface, the impedance diagrams became simpler, comparing same period (Figure 6). It can be assumed the presence of corrosion products occludes the defect when it is small, hiding the progress of the detachment of the coating. Nevertheless, increasing the size of defect moves the response due to the coating out of the experimental frequency window. As a a consequence the size of the defect was shown a relevant and important parameter in the corrosion processes and further efforts have to be addressed to explain both small and large defects affect impedance response. 

 

Figure 4. Detail (high frequency zone) of the Nyquist diagram of a coated zirconia samples with an artificial defect (150 μm); the whole diagram is showed in Figure 5.  

Figure 5. Nyquist diagram of a coated zirconia sample with an artificial defect (150 μm).

Figure 6. Nyquist diagram of the coated zirconia sample with an artificial defect (300 μm).

Conclusions

On the base of these results, it is easy to infer that thin amorphous zirconia films are better pre-treatment than a commercial iron phosphate, even if they are not so good as the tricationic one. Also, the thermal treatment on forming the zirconia film is important and a very strict control in the use of temperature and time of treatment is a condition in order to obtain the best performance on adhesion of the ZrO2 film. The N-methyl pyrrolidone test for the assessment of the adhesion between organic coating and substrate cannot provide indisputable results because of its interaction with the organic component still present in the studied zirconia thin film.

References

1. H. Leidheiser, Jr., W. Funke, J.Oil Colour Chem. Assoc. 68, 121 (1985).

2. F. Deflorian, L. Fedrizzi, J. Adhesion Sci. Technol. 13 (5), 629-645 (1999).

3. L.Benedetti, P.L. Bonora, M. Sassoli, Proceedings of Eurocoat 94 (Volume II), Sitges, Spain, 1994, p. 31-49.

4. Werther Neri, Introduzione alla Verniciatura delle Superfici Metalliche, 3a. Edizione, Ed. Tecniche Nuove, Milano, 1990.

5. M. Guglielmi, J. Sol-Gel Sci. Technol. 8, 443-449 (1997).

6. R. Di Maggio, S. Rossi, L. Fedrizzi, P. Scardi, Surface & Coatings Technol. 89, 292-298 (1997).

7. C.J. Brinker and G.W. Scherer, Sol-gel Science, Academic Press, San Diego, 1990.

8. D. Peter, T.S. Ertel and H. Bertagnolli, J. Sol-Gel Sci. Technol. 5, 5-14 (1995).

9. M.Chatry, M. Henry, M. In, C. Sanchez and J. Livage, J. Sol-Gel Sci. Technol. 1, 233-240 (1994).

10. R. Di Maggio, R. Campostrini, G. Guella, Chem. Mater., 10 (12) 3839-3847 (1998).

11.W.J.Van Ooij, R.A.Edwards, A.Sabata, J.Zappia, J.Adhesion Sci. Technol. 7 (8), 897-917 (1993).

12.B. Boukamp, Solid State Ionics, 20, p. 31 (1986).