Surface and Coatings Technology 140 Ž2001. 44᎐50

Use of a sol᎐gel conversion coating for aluminum corrosion protection X.F. Yang a,U , D.E. Tallmana , V.J. Gelling a , G.P. Bierwagenb, L.S. Kasten c , J. Berg a b

a Department of Chemistry, North Dakota State Uni¨ ersity, Fargo, ND 58105-5516, USA Department of Polymers and Coatings, North Dakota State Uni¨ ersity, Fargo, ND 58105-5516, USA c Uni¨ ersity of Dayton Research Institute, 300 College Park, Dayton, OH 45469-0168, USA

Abstract In this study, the behavior of a sol᎐gel conversion coating alone and in combination with a polyurethane unicoat ŽTT-P-2756, self-priming topcoat. on Al 2024-T3 alloy was investigated under immersion in dilute Harrison’s solution w3.5 grl ŽNH 4 . 2 SO4 , 0.5 grl NaClx. The sol᎐gel coating consisted of SiO 2 and ZrO 2 with a ratio of 3.4:1. The evolution of the coating system under immersion was followed by atomic force microscopy ŽAFM., scanning electronic microscopy ŽSEM., electrochemical impedance spectroscopy ŽEIS., and X-ray photoelectron spectroscopy ŽXPS.. It was found that though pitting corrosion and degradation products on the sol᎐gel single coating surface were observed after 2 days of immersion, further pitting corrosion ceased after a few days of immersion. Under-coating blisters in the sol᎐gel plus polyurethane topcoat system were found at 4 weeks of immersion, after which no further increase in size of the blisters was observed. It is conjectured that aluminum oxide and silicon oxide may form a stable mixed oxide barrier layer at the interface after initial corrosion, which prohibits further pitting corrosion development. 䊚 2001 Elsevier Science B.V. All rights reserved. Keywords: Sol᎐gel; Aluminum; Corrosion protection

1. Introduction The use of sol᎐gels for metallic surface pretreatment is a relatively new approach in surface engineering and corrosion protection. Such efforts are motivated by the desire to find suitable replacements for the more traditional but environmentally hazardous chromate-based surface treatments. There are only a few papers published which examine sol᎐gel application in this field w1,2x. In this study, the sol᎐gel film is used as a conver-

U

Corresponding author. Tel.: q1-701-231-8385; fax: q1-701-2318831. E-mail address: [email protected] ŽX.F. Yang..

sion coating for a self-priming polyurethane coating. For this application, the durability of a thin sol᎐gel coating Ž50᎐100 nm. is of concern. Therefore, we inspected the performance of a sol᎐gel conversion coating alone and in combination with a polyurethane unicoat ŽTT-P-2756, self-priming topcoat. on Al 2024T3 alloy under immersion in dilute Harrison’s solution Ž3.5 grl NH 4 SO4 , 0.5 grl NaCl.. In this paper, we discuss the affect of the sol᎐gel conversion coating in the corrosion of Al 2024-T3. According to traditional views, polymer coatings are damaged by ‘under coating corrosion’ which initiates from weak spots and develops into blisters and filiform threads w3᎐6x. In the past, most investigations focused on the stage of corrosion where the corrosion could be

0257-8972r01r$ - see front matter 䊚 2001 Elsevier Science B.V. All rights reserved. PII: S 0 2 5 7 - 8 9 7 2 Ž 0 1 . 0 1 0 0 2 - 7

X.F. Yang et al. r Surface and Coatings Technology 140 (2001) 44᎐50

observed visually. In this work, using AFM, SEM, EIS and XPS, coating degradation and under-coating corrosion were investigated from its initial stage.

2. Experimental procedures 2.1. Materials preparation The metal substrate was Al 2024-T3 alloy. The sol᎐gel conversion coating supplied by Boeing Company ŽSeattle, WA, USA. contains SiO 2 and ZrO 2 with a ratio of 3.4:1. Before applying the sol᎐gel coating, the aluminum surface was degreased with methyl ethyl ketone, followed by emulsion degreasing. After using an alkaline cleaner, the surface was deoxidized with HNO3 ᎐HF᎐H 2 SO4 . The thickness of the sol᎐gel coating was approximately 50᎐100 nm. The top coating was a self-priming polyurethane coating whose commercial code is TT-P-2756 and it was provided by Boeing Company. Its thickness was in the range of 80᎐100 ␮m. All coated aluminum samples were provided as 2-mmthick sheets by Boeing. In order to fit into the sample stage of the atomic force microscope ŽAFM., samples were punched into circular shape with a diameter of 1.3 cm. To avoid mechanical damage during the punch, the sample was protected with a plastic film. To prepare a sample for an immersion test, the entire sample was rinsed with acetone and distilled water, the non-working area Žback surface and the edges of the sample. was sealed with ‘Torr Seal Low Vapor Pressure Resin Sealant’. The sealant was purchased from Varian Vacuum Products and it was a compound mixture of epoxy resin. The sealed sample was left overnight to cure before use. 2.2. Immersion conditions Corrosion experiments were carried out by immersing the sealed samples in dilute Harrison’s solution w3.5 grl ŽNH 4 . 2 SO4 q 0.5 grl NaClx. The system was held under constant immersion and was tested at time intervals from 3 h to 30 weeks. 2.3. Atomic force microscopy and scanning electron microscopy studies For microscopic examination, the samples were removed from the immersion solution, rinsed with distilled water and dried in air. The AFM used in this work was a Nanoscope IIIa, Digital MultiMode AFM ŽDigital Instruments, CA., which operates in either the tapping mode or the contact mode with either sharp silicon tips or silicon nitride tips. An E scanner Ž13.5 ␮m. and a J scanner Ž125 ␮m. were employed in this

45

study. The images were collected over a wide scan range, from 125 ␮m to 500 nm, depending on the size of the features present on a given sample. Only a few representative images are presented here. Scanning electron microscopy ŽSEM. studies were performed by a JSM-6300V scanning electron microscope ŽJEOL, Boston, MA, USA. equipped with an X-ray detector for energy dispersive X-ray analysis ŽEDX.. The microscope was operated at a 10y7 -torr vacuum. A 15-kV voltage was employed to image features and perform EDX analysis. 2.4. Electrochemical impedance spectroscopy (EIS) studies The electrochemical impedance system used in this study was an PC-3 ŽGamry Instruments, Inc. PA, USA.. Electrochemical measurements were carried out in dilute Harrison’s solution at the open circuit potential. The counter electrode was a platinum mesh electrode with an area of 2.5 cm2 and the area of the working electrode was 12.56 or 7 cm2 . The reference electrode was a saturated calomel electrode ŽSCE.. The frequency range of the applied AC potential was 0.01᎐5 KHz and the amplitude was 5 mV. The immersed samples were measured at different time intervals from 1 h to 4 weeks. 2.5. XPS analysis The XPS instrument used in this study was a Physical Electronics model 5700 MultiTechnique Spectrometer ŽX-ray photoelectron spectroscopy and Auger electron spectroscopy.. A monochromatic Al X-ray source Žhaving an energy of 1486.6 eV. was used for the analysis. The XPS system is equipped with a hemispherical analyzer and a 16-channel detector. A pass energy is 188 eV for survey scans and 58.7 eV for narrow scans, corresponding to analyzer resolutions of 1.88 and 0.235 eV, respectively. The pressure in the system was approximately 5 = 10y1 0 torr during analysis.

3. Results and discussions Fig. 1 shows an AFM image of a sol᎐gel coating before immersion. The coating possesses a rough surface. Since the Al substrate is rough, and the sol᎐gel film is too thin to cover the roughness of the Al substrate, the sample topography reflects the underlying roughness in the AFM images. The z-range was set accordingly. After 2 days of immersion in dilute Harrison’s solution, some blisters were observed on the surface ŽFig. 2.. The arrow in Fig. 2 points to one of the blisters. XPS analysis showed that at this time, the chemical composition of the sol᎐gel surface did not change, and

46

X.F. Yang et al. r Surface and Coatings Technology 140 (2001) 44᎐50

Fig. 1. AFM height image of the sol᎐gel coating.

especially, the aluminum content did not increase on the coating surface. This indicates that these surface blisters were not aluminum corrosion products. We conjecture that they are coating degradation products. It is possible that siloxane hydrolysis degraded the sol᎐gel by breaking Si᎐O᎐Si chains w7᎐9x. Fig. 3 shows two SEM images showing the sol᎐gel coating before immersion ŽFig. 3a. and after 4 months of immersion ŽFig. 3b.. After 4 months of immersion, the sol᎐gel coating was still mostly intact and its topography did not change significantly. Therefore, the sol᎐gel coating in dilute Harrison’s solution can last more than 4 months. Blisters were not observed in the SEM images since the blisters are surface features with a height of only a few tens of nanometers, which is in the limit of SEM resolution. Therefore, the blisters are much more difficult to observe in SEM than in AFM. When we inspected the surface visually, we found that after 3 days of immersion in dilute Harrison’s

Fig. 2. AFM height image showing the sol᎐gel coating after 2 days of immersion in dilute Harrison’s solution.

Fig. 3. SEM images of sol᎐gel coating before and after 4 months of immersion in dilute Harrison’s solution: Ža. before immersion; Žb. 4 months of immersion.

solution, several black pits were on the Al alloy surface. As the immersion study progressed, the development of most pits slowed down or stopped, and only a few large pits, with diameter over 100 ␮m, continued to grow. We examined more than 10 samples, each 13 mm in diameter. The number of large pits per sample was in the range of 0᎐3. Fig. 4 shows the results of EIS measurements. It is known that the high frequency part of the spectra reflects the capacitance of the coated metal surface and the electrolyte resistance whereas the low frequency part reflects the resistance of the coating. Since water penetration and metal corrosion change the resistance of the coating, the low frequency impedance modulus, < Z < 0.01Hz , is of particular interest. In Fig. 4, < Z < 0.01Hz of the sol᎐gel coating increased steadily during the 4 weeks of immersion. < Z < 0.01Hz was only approximately 200᎐300 ⍀ after 1 h of immersion while it approached 10 4 ⍀ after 1 week. It retained that level throughout the rest of the immersion. This indicates that the sol᎐gel coating had poor corrosion resistance in the initial immersion time and its corrosion resistance improved gradually in the solution. The above

X.F. Yang et al. r Surface and Coatings Technology 140 (2001) 44᎐50

47

Fig. 4. EIS result showing the modulus < Z < 0.01Hz change of the sol᎐gel coating during 4 weeks of immersion in dilute Harrison’s solution.

impedance results corroborate the visual observation that pit formation occurred on the aluminum surface in the first a few days immersion and was inhibited during further immersion. By repeating the immersion experiment, we found that < Z < 0.01Hz of the sol᎐gel coated alloy stayed in the 10 3 ᎐10 5-⍀ range throughout the immersion time and that the < Z < 0.01Hz increased upon immersion in four of the five sol᎐gel coated samples. This phenomenon shows that upon immersion, a protecting layer between the sol᎐gel and the metal surface may develop. For a better understanding of the sol᎐gel function on the Al substrate, the corrosion behavior of bare Al 2024-T3 alloy was also investigated in dilute Harrison’s solution. Fig. 5 shows a newly polished Al 2024-T3 surface. Metal grains were exposed upon polishing. The small spots on the grain surface may be oxide particles,

which rapidly formed on the surface after polishing. The Al surface had changed after 3 days of immersion ŽFig. 6.. Several visible pits appeared on the surface after 1 or 2 days of immersion and pitting corrosion developed rapidly. The SEM micro-graphs in Fig. 7b show the corroded Al surface after 4 months of immersion. Compared with Fig. 7a, which shows the Al surface before immersion, many large pits had formed in the Al surface. Pitting corrosion is caused by the local dissolution of metal due to the non-uniformity of the alloy. For example, copper enriched phases possess a higher potential than the surrounding Al area and the Al corrodes preferentially. The significant differences of aluminum surface with and without sol᎐gel coating after 4 months of immersion indicates that the sol᎐gel coating protected the metal surface from local dissolution. An impedance modulus graph in Fig. 8 recorded the

Fig. 5. An AFM height image of newly polished Al 2024-T3 bare alloy.

Fig. 6. An AFM height image of Al 2024-T3 alloy after immersion in dilute Harrison’s solution for 3 days.

48

X.F. Yang et al. r Surface and Coatings Technology 140 (2001) 44᎐50

Fig. 8. EIS result showing the < Z < 0.01Hz of bare Al 2024-T3 during 4 weeks of immersion in dilute Harrison’s solution.

Fig. 7. SEM images showing Al 2024-T3 alloy before and after 4 months of immersion in dilute Harrison’s solution: Ža. before immersion; Žb. after 4 months of immersion.

variation of the bare alloy sample < Z < 0.01Hz after 1 h, 1 day, 1 week, and 4 weeks immersion in the solution. < Z < 0.01Hz increased from approximately 10 3 to 10 5 ⍀ after just 1 day of immersion, which may result from the rapid formation of an oxide film on the surface. With further immersion, < Z < 0.01Hz decreased to 10 4 ⍀ after 1 week and to 10 3 ⍀ after 4 weeks. Repeated immersion experiments showed that, though the < Z < 0.01Hz initially increased upon immersion, there was a drop in < Z < 0.01Hz over the next 4 weeks in all of the samples. Similar to the sol᎐gel coated samples, < Z < 0.01Hz of bare aluminum remained in the 10 3 ᎐10 5-⍀ range throughout most of the immersion time. A comparison of the EIS results obtained from the bare Al and sol᎐gel coated Al shows that though the thin sol᎐gel coating cannot raise the impedance modulus, the sol᎐gel film does appear to produce a more stable protective layer which protects the metal surface from pitting corrosion. Furthermore, visual observation suggests that the sol᎐gel coated sample can arrest the pitting corrosion after it initiates. The performance of the sol᎐gel conversion coating plus polyurethane TT-P-2756 self-priming topcoat in the dilute Harrison’s solution was also investigated. Fig. 9 is a 25-␮m square AFM image showing the poly-

urethane coating on the sol᎐gel conversion coating before immersion. Craters scattered on the flat coating surface had a diameter of approximately 250 nm and a depth of 30᎐50 nm. They formed during the coating application process. Since the thickness of the coating Ž80᎐100 ␮m. is more than 1000 times the depth of these craters, the craters probably do not affect the coating performance by reducing the coating thickness. After 4 weeks of immersion, the roughness of the coating surface had increased due to coating degradation and under-coating blisters formed ŽFig. 10.. AFM measurements show that the largest blisters in the field of view were approximately 13 ␮m in diameter and 338 nm in height. Prolonging the immersion time did not seem to significantly increase the size or change the appearance of these blisters. For example, after 28 weeks immersion, the largest blisters were still approximately 10.5 ␮m in diameter and approximately 420 nm high.

Fig. 9. An AFM height image showing a polyurethane topcoat surface.

X.F. Yang et al. r Surface and Coatings Technology 140 (2001) 44᎐50

EDX analysis shows that aluminum content in the polyurethane coating was under 0.6 at.% throughout the 30-week immersion period. Thus, it is proposed that aluminum ions were blocked in the sol᎐gel or at the sol᎐gel and metal interface. Others have found that aluminum ions can replace silicon ions in the SiO 2 framework w10,11x. Thus, the reaction that occurred at the interface may be the replacement of silicon in sol᎐gel coating by aluminum. It was observed from the sol᎐gel single coating immersion experiment that pitting corrosion occurred at certain areas of the sol᎐gel coating in the initial immersion period. These areas may be coating defects, where the solution reached the metal surface easily. Further corrosion was inhibited even under prolonged immersion. Therefore, the fast initial formation and slow further development of under-topcoat blisters may be due to the existence of the sol᎐gel film. The proposed anti-corrosion function of the sol᎐gel layer is as follows. During immersion, penetrated water reacts with metal. This reaction occurs more rapidly within the sol᎐gel coating defect, which causes Al 3q to become concentrated in the solution around the sol᎐gel coating defect under the topcoat. This promotes the absorption of water and forms an osmotic cell between the topcoat and the sol᎐gel coating, leading to blister formation, which we observed on the topcoat in Fig. 10. This process is very similar to the formation of traditional under coating blisters w12᎐19x. In the next period, Al 3q replaces the silicon in the sol᎐gel network at the Al and sol᎐gel interface to form a mixed oxide layer and the released silicon oxide forms an insoluble layer with aluminum oxide in the sol᎐gel defect. As a consequence, the further development of the blisters was controlled. Fig. 11 is the schematic diagram of the

49

Fig. 11. Schematic diagram showing that when the solution penetrates into the sol᎐gel coating, corrosion occurs at initial immersion period, especially, pitting corrosion occurs in the coating defects. The dissolved aluminum forms a mixed oxide layer with silicon oxide in the interface of the sol᎐gel coating and aluminum surface, which prohibits the development of pitting corrosion.

process. Hydration may increase the volume of the sol᎐gel coating, which may also enhance the corrosion resistance of the sol᎐gel coating by sealing the layer.

4. Summary The sol᎐gel coating inhibits the local dissolution on the aluminum surface. The EIS measurement shows that when Al bare alloy was immersed in dilute Harrison’s solution, the impedance modulus increased initially and then deceased over the next 4 weeks. In contrast, the impedance modulus of the sol᎐gel coated Al alloy increases steadily during the first 4 weeks of immersion. Therefore, sol᎐gel coating stabilized the protecting layer on the aluminum surface.

Acknowledgements

Fig. 10. An AFM height image of a sol᎐gel conversion coating plus a polyurethane self-priming topcoat after immersion in dilute Harrison’s solution for 4 weeks; under-coating blisters observed.

We acknowledge Joseph Osborne and Kay Blohowiak from Boeing Defense & Space Group for providing samples, Kathy Iverson and Scott A. Payne at North Dakota State University for their assistance in the

50

X.F. Yang et al. r Surface and Coatings Technology 140 (2001) 44᎐50

SEM and EDX work. We also acknowledge The United States Air Force for providing financial support for this study ŽGrant No F49620-96-1-0284.. References w1x A.R. Di Giampaolo, M. Medina, R. Reyes, M. Velez, Surf. Coat. Technol. 89 Ž1997. 31᎐37. w2x A. Atik, M.A. Aegerter, J. Non-Cryst. Solids 147᎐148 Ž1992. 813᎐819 ŽNorth-Holland.. w3x H.H. Uhlig, Corrosion and Corrosion Control, John Wiley & Son Inc, New York, 1963, pp. 219᎐223. w4x E.V. Schmid, Exterior Durability of Organic Coatings, Fmjinternational Publications Ltd., England, 1988, pp. 143᎐152. w5x Z.W. Wicks Jr., F.N. Jones, S.P. Pappas, Organic Coatings: Science and Technology, vol. II: Applications, Properties, and Performance, John Wiley & Sons, Inc., New York, 1994, pp. 170᎐190. w6x E.E. Mcsweeney, in: W. Von Fischer, E.G. Bobalek ŽEds.., Organic Protective Coatings, Reinhold Publishing Corporation, New York, 1953, pp. 305᎐319.

w7x V.V. Moiseev, G.E. Zaikov, Chemical Resistance of Polymers in Aggressive Media, Plenum Press, New York, 1982. w8x H.R. Allcock, J. Macromol. Ser. 4C Ž1970. 149. w9x H.F. Mark, N.M. Bikales, C.G. Overberger, Encyclopedia of Polymer Science and Engineering, John Wiley & Sons, New York, 1988, p. 671. w10x P. Laszlo, Science 235 Ž1987. 1473᎐1477. w11x P.D. Taylor, R. Jugdaohsingh, J.J. Powell, J. Am. Chem. Soc. 119 Ž1997. 8852᎐8856. w12x W. Funke, U. Zorll, B.G.K. Murth, J. Coat. Technol. 41 Ž530. Ž1969. 210. w13x W. Funke, H. Haagen, Ind. Eng. Chem. Prod. Res. Dev. 17 Ž1978. 50. w14x W. Funke, J. Oil Col. Chem. Assoc. 62 Ž1979. 63. w15x W. Funke, Prog. Org. Coat. 9 Ž1981. 29. w16x W. Funke, J. Coat. Technol. 55, 705 Ž1983. 31. w17x W. Funke, Ind. Eng. Chem. Prod. Res. Dev. 24 Ž1985. 343. w18x W. Funke, J. Oil Col. Assoc. 68 Ž1985. 229. w19x W. Funke, Farbe Lack 93 Ž1987. 721.