O.ÃƒËœ. Knudsen, E. Bardal and U. Steinsmo
Because you are not logged-in to the journal, it is now our policy to display a 'text-only' version of the paper. This version is obtained by extracting the text from the PDF or HTML file, and it is not guaranteed that the text will be a true image of the text of the paper. The text-only version is intended to act as a reference for search engines when they index the site, and it is not designed to be read by humans!
If you wish to view the human-readable version of the paper, then please Register (if you have not already done so) and Login. Registration is completely free.JCSE Volume 2 Paper 13 ifnotmode(3,'Submitted 13th September 1999, published for public review 27th October 1999') Effect of Barrier Pigments on Cathodic Disbonding Part 1: Aluminium and Glass Pigments O.�. Knudsen1, E. Bardal2 and U. Steinsmo1 1SINTEF Materials Technology, 7465 Trondheim, Norway. 2Norwegian University of Science and Technology (NTNU), Department of Machine Design and Materials Technology, 7491 Trondheim, Norway. E-Mail Adress: mailto2('ole.knudsen','matek.sintef.no') §1 Abstract The effect of aluminium and glass barrier pigments on cathodic disbonding of epoxy coatings has been studied. Over a two years period, cathodic disbonding has been measured for nine model coatings with various concentrations of aluminium or glass pigments. The binder in the model coatings was either epoxy or epoxy mastic. The test specimens were exposed in natural seawater and polarised to -1050 mV SCE with zinc anodes. The pigments had a large effect on cathodic disbonding. Coatings without barrier pigments disbonded significantly during the test period. Coatings with 10% aluminium pigments or more did not show any sign of disbonding until the end of the test. Then some samples had about 3 mm disbonding around the damage in the coating. Coatings with 10% or more glass pigments had a very long delay time before disbonding started, but disbonded severely after that. Possible mechanisms of the effects are discussed. §2 Keywords: Organic coatings, cathodic disbonding, aluminium pigments, glass flake pigments. Introduction §3 Cathodic disbonding is one of the most important mechanisms for loss of adhesion between coating and steel. The problem has lately received new attention due to the increased use of coatings in combination with cathodic protection of offshore oil production installations. Resistance against cathodic disbonding is an important property of the coating systems to be used on submerged structures. E.g. the NORSOK standard for development and operation of oil fields in the Norwegian sector use resistance against cathodic disbonding as a criterion for prequalification of coatings for this application . §4 The purpose with this paper is to examine the effect of flake shaped glass- and aluminium pigments on cathodic disbonding. Model coatings with various contents of glass and aluminium pigments have been prepared and tested for cathodic disbonding. The conditions we have used are close to the conditions found on offshore installations in the North Sea. We have found that both glass and aluminium pigments improve the resistance of the coatings against cathodic disbonding. The effect was found with two different binders which indicate that this is a general effect, independent of the binder. We have outlined two possible mechanisms for the effect. The first possibility is improved barrier properties of the coating. Cathodic disbonding depends on transport of reactants through the coating. The barrier pigments may then impede the transport, which may slow down the cathodic disbonding. The other possibility we have looked at is buffering reactions with the pigments. Leidheiser et al. have suggested that the aluminium may reduce the pH at the coating/substrate interface and thereby decrease cathodic disbonding . §5 Aluminium pigments have been widely used in engineering paints and protective coatings for many years. The aluminium pigments are shaped as flakes in different sizes. The pigments we have used are about 20 - 40 �m wide and 1 �m thick. This geometry, along with their metallic nature, is responsible for most of their properties in protective paints. The aluminium pigments improve the barrier properties of protective paints [3,4]. From a corrosion protection point of view, their ability to impede the transport of water and oxygen is the most important property. The pigments are impermeable to water and oxygen, and the diffusion is forced to go around the particles. This may increase the diffusion path significantly and thereby decrease the permeability through the paint . In addition the pigments also act as barriers to radiation, as they reflect both IR, UV and visible light . This increases the stability of the paint, and may also reduce heat radiation from the substrate. Aluminium pigments are commercially available as leafing or non-leafing grades. The leafing pigments float to the surface of the applied film, while the non-leafing pigments are evenly distributed in the film after application. The different behaviour is due to surface treatment of the pigments with different fatty acids. §6 Glass flake pigments are also used in protective coatings, but not to the same extent as aluminium flake pigments. They are produced by breaking thin glass (about 5�m thick) into small flakes, about 100 - 400 �m wide. Their relatively large size restricts them to be used in very thick coatings only. The shape and impermeability to water and oxygen suggest that glass flakes will improve the barrier properties of the coating. Glass flake pigments have received little attention in the literature. Foster et al. have described some of the basic characteristics of glass flake filled coatings . El-Sawy et al. have tested the effect of glass pigments in a scratch test and an environmental exposure test . However, the glass pigments they used gave no positive effect on the performance of the coatings. §7 The mechanism for cathodic disbonding is not known in every detail. However, it is widely accepted in the literature that high pH at the steel/coating interface is responsible for the disbonding [2,7,8]. The high pH is produced by the cathodic reaction on the steel surface under the coating. Structures in seawater protected by zinc- or aluminium anodes are polarised to between -800 and -1100 mV SCE. At these potentials the cathodic reaction can be both reduction of oxygen and hydrogen evolution: O2 + 2H2O + 4e- = 4OH- (1) 2H2O + 2e- = H2 + 2OH- (2) §8 Leidheiser have shown that above -1000 mV SCE the oxygen reduction dominates, while below this potential the hydrogen evolution dominates . However, in the cathodic areas under the coating the potential is probably well above -1000 mV SCE, due to the high resistance between the anode and the cathode. Leidheiser have therefore concluded that the oxygen reduction is the predominant reaction under the coating . §9 Ritter et al. have measured the pH under the disbonded coating and found that it could be as high as 14 . The exact consequence of the alkaline water film is not known, but three theories have received most attention. The three mechanisms are based on different locus of failure: in the iron oxide on the steel surface [2,10], in the coating [11,12] or at the interface [8,13]. §10 The transport of reactants for the cathodic reaction under the coating has also been subject to much research. Most paint films are permeable to both oxygen and water . It is therefore reasonable to assume that these reactants mainly are transported through the film. Leidheiser et al. have also shown this experimentally . For cations, the conclusions are less clear. Several researchers have concluded that the transport of cations goes laterally from the holiday under the disbonded coating [2,15]. The time dependence of the disbonding is an indication for this. The disbonded area is usually found to increase proportionally with time, i.e. the disbonded distance from the holiday in the coating is proportional with the square root of time [2,15-17]. As the disbonded area increases, the distance for transport of ions under the disbonded film to the disbonding front also increases. Therefore the resistance to the transport of cations increase and the rate of disbonding will decrease. However, there are also those who have found indications for transport of cations through the coating. Skar measured the disbonding of coatings with film thickness between 100 and 300 �m and found that the disbonding rate decreased linearly with film thickness . He also found that the transport of cations was rate limiting. He therefore suggested that the transport of cations went through the disbonded film. Jin et al. , Leidheiser  and Parks et al.  have also concluded that cations might be transported through the coating when the specimen is under cathodic polarisation. §11 Many researchers have studied the rate determining step in cathodic disbonding. Leidheiser, among others, have shown that the transport of oxygen through organic coatings is much lower than the transport of water . Oxygen may therefore for some coatings be rate limiting for cathodic disbonding. However, for most coatings researchers have found that the transport of cations are rate limiting. When cathodic disbonding is measured in different electrolytes (LiCl, NaCl, KCl, CsCl etc.), they have found that the rate of disbonding is linearly proportional to the mobility of the ions in water [2,15,18,22]. Experimental §12 We have used model coatings with two different binders: an amide cured epoxy and an amine cured epoxy modified with a hydrocarbon resin (epoxy mastic). The epoxy binder was pigmented with different concentrations of aluminium flake pigments. The epoxy mastic binder was pigmented with either aluminium flake pigments or glass flake pigments. The aluminium pigments we have used are about 20 - 40 �m wide and 1 �m thick. The glass flake pigments were up to 400 �m wide and 10 �m thick. Information about composition and pigmentation of the paints is given in Table 1 and 2. §13 Cathodic disbonding for all nine model coatings was tested in a long term test. The substrate was 200 x 200 x 5 mm steel plates cut from mild steel (DIN 17100). The composition of the steel was 0.10% C, 0.10% Si, 0.52% Mn, 0.14% P, 0.014% S, 0.23% Cr and Fe to balance. They were blast cleaned with steel grit to Sa 21/2  and medium roughness (Ry = 76 � 7 �m) . The paints were applied by airless spraying in two coats. The first coat cured for about 24 hours at room temperature before the second coat was applied. After that, the samples cured for 55 days at ambient temperature and humidity before testing. Coating thickness was measured with a magnetic gauge. The accuracy of the instrument is better than 10%. The dry film thickness is given in Table 1 and 2. The samples were given a circular holiday in the middle, 6 mm in diameter, to start the cathodic disbonding process. They were exposed in continuously running natural seawater at 8-13�C. Galvanic coupling to zinc anodes gave cathodic polarisation to about -1050 mV SCE in the coating holiday. The experimental setup is shown in Figure 1A. Cathodic disbonding was measured destructively by gently lifting the disbonded paint film off the steel substrate with a scalpel. The diameter of the disbonded circle was measured in four directions and the disbonded distance was calculated from the average diameter. §14 Table 1. Amid cured epoxy: Coating composition (wt%) of wet paint and applied thickness. Epoxy 0 Epoxy 5 Epoxy 10 Epoxy 20 Amide cured epoxy resin 36.0 34.4 32.4 31.8 Extenders 26.7 25.3 24.2 16.0 Solvent 35.9 33.7 32.0 31.0 Additives 1.4 1.3 1.3 1.2 Al paste, non-leafing - 5.3 10.1 20.0 Density of wet paint (g/cm3) 1.24 1.25 1.26 1.21 PVC (%) 21.6 24.0 26.0 26.7 Vol % solids 51 51 50 47 Applied thickness (�m) Blast cleaned steel 273 � 17 277 � 12 263 � 34 353 � 40 Ground steel 205 � 4 209 � 5 203 � 4 207 � 4 §15 The amide cured epoxy was also tested in a short term test. Steel samples of 80 x 80 x 5 mm were cut from mild steel. The samples were ground with emery paper to roughness Ry = 5 �m. The paints were applied with an applicator in two coats. The first coat was allowed to cure for about 24 hours at room temperature before the second coat was applied. After application of the second coat the samples cured for 34 days at room temperature before testing. Coating thickness was measured the same way as in the long term test. The coating thickness is given in Table 1. Cathodic disbonding was tested in substitute seawater  at 25�C. The samples were polarised to -1050 mV SCE with a potentiostat. The test setup is shown in Figure 1B. Cathodic disbonding was measured the same way as in the long term test. §16 Table 2. Amine cured epoxy mastic: Coating composition (wt %) of wet paint and applied thickness. Pigmentation Aluminium Glass Mastic 0 Mastic 10A Mastic 20A Mastic 10G Mastic 20G Amine cured epoxy modified with hydrocarbon resin 38.0 37.2 36.5 38.4 38.8 Extenders 50.5 41.6 32.7 39.7 28.8 Solvent 10.5 10.3 10.0 10.6 10.7 Additives 1.0 1.0 1.0 1.3 1.7 Aluminium paste, non-leafing - 9.9 19.8 - - Glass flakes - - - 10.0 20.0 Density of wet paint (g/cm3) 1.50 1.44 1.38 1.48 1.46 PVC (%) 33 33 33 33 33 Vol.% solids 79 75 70 79 79 Applied thickness (�m) 349 � 22 377 � 18 337 � 49 354 � 45 375 � 36 §17 Three of the epoxy mastic coatings were analysed by Scanning Electron Microscopy: Mastic 0, Mastic 10A and Mastic 10G. Cross sections were cut with an emery disc, embedded in epoxy, ground and coated with carbon by vapour deposition. Images were taken with element contrast at 200x magnification. §18 Figure 1. Test setup for cathodic disbonding. A: Long term test: Blast cleaned substrate, natural seawater, 8-13�C, -1050 mV (SCE). B: Short term test: Ground substrate, substitute seawater, 25�C, -1050 mV (SCE). Results and Discussion Effect of aluminium pigments on cathodic disbonding §19 Figure 2 shows the disbonding distance in the long term test as a function of time, for the amide cured epoxy coatings with aluminium pigments. 15 samples of each model paint were prepared, and cathodic disbonding was measured during a two years period. With increasing time intervals, 1-3 samples was removed and analysed. The aluminium pigments were highly effective in improving the resistance of the coating towards cathodic disbonding. Epoxy 10 and Epoxy 20 had not disbonded at all after 18 months of exposure. After two years Epoxy 10 had some disbonding, and about 3 mm of the coating around the holiday could be removed. Epoxy 5 disbonded about 20 mm and Epoxy 0 disbonded about 65 mm in the same period of time. §20 Figure 2. Cathodic disbonding of epoxy coating with various concentrations of aluminium pigments. Coatings applied on blast cleaned steel (Sa 21/2). Experimental conditions: Natural seawater, 8-13�C, polarised with zinc anodes to -1050 mV SCE.ifscreen('(Click for a larger image)') §21 The aluminium not only affected the rate of disbonding, but also the degree to which the coating lost adhesion. For Epoxy 5, it was much harder to remove the disbonded coating, compared to Epoxy 0. This was also seen when we looked at the metal surface in a microscope after removing the disbonded coating. Under Epoxy 5 we found more "islands" where the coating still adhered to the substrate, compared to Epoxy 0, where very little of the binder was left on the steel surface. §22 Figure 3 shows the amount of coating still found on the steel surface after removing the disbonded coating on samples of Epoxy 0 and Epoxy 5. The pictures were created and analysed in an electronic image analyser. The sample in Figure 2A was coated with Epoxy 5 and exposed for 18 months. The radius of the disbonded circle is about 32 mm. The quantitative analysis showed that remainings of the coating covered about 18% of the disbonded area. The sample in Figure 2B was coated with Epoxy 0 and exposed for one year. The radius of the disbonded circle is about 120 mm. For this coating only 4% of the disbonded area was covered with remainings of the binder. §23 Figure 3. Pictures of the disbonded area after removing the disbonded coating taken with an electronic image analyser. The white spots are remainings of the binder. A: Epoxy 5 exposed for 18 months, 32mm disbonded diameter. B: Epoxy 0 exposed for one year, 120 mm disbonded diameter.ifscreen(' (Click for a larger image)') §24 The coatings with high aluminium concentrations had little or no disbonding after two years in the long term test. To get quantitative information about how much the aluminium pigments decreased cathodic disbonding, we also tested the amide cured epoxy coatings in a short term test on ground steel. In general the rate of cathodic disbonding increases when the roughness of the substrate decreases . Epoxy 10 and Epoxy 20 disbonded faster on ground steel. Figure 4 shows the rate of cathodic disbonding on ground steel as function of aluminium content. The disbonding rate for Epoxy 10 and Epoxy 20 is about the same, but more than two decades slower than for Epoxy 0. The time dependence of the disbonding is different from the parabolic kinetics that usually is found, e.g. in Figure 2. This may indicate that some other mechanisms are involved, but we have not been able to investigate this. However, the results clearly demonstrate effect of the aluminium pigments. §25 The surface roughness had a large effect on the disbonding rate. On ground steel substrate (Ry = 5 �m) Epoxy 0 disbonded 25 mm in about one day. On blast cleaned steel (Ry = 76 �m) it took about three months to get the same disbonded distance. Watts and Castle have also tested the effect of substrate roughness on the rate of cathodic disbonding . They suggested that the effect of the substrate roughness was due to a change in the diffusion path length under the coating. As the substrate roughness increased the diffusion path length between the holiday and the disbonding front would increase. They found a reasonable correlation between the disbonding rate and the actual length of the surface profile. The effect may also be due to an increase in the contact area between the coating and the substrate, when the surface roughness increases . When we measured cathodic disbonding on a substrate that had been abraded in one direction only, the disbonded distance was always larger along the abrasion lines. Jin et al. found the same . This observation favours the mechanism suggested by Watts and Castle. §26 Figure 4. Cathodic disbonding of epoxy coating with various concentrations of aluminium pigments. Coatings applied on ground steel substrate. Experimental conditions: Artificial seawater, 25�C, -1050 mV SCE.ifscreen(' (Click for a larger image)') §27 The amide cured epoxy has low commercial interest due to the high content of solvents. The epoxy mastic binder was therefore included in the test to represent a more commercially interesting binder. The disbonding distances for the epoxy mastic coatings are given in Figure 5. Pre-treatment, application, curing and exposure were similar to the long term test of the amide cured epoxy. The unpigmented epoxy mastic gave about 8 mm disbonding in four months, but then the coating started to blister randomly all over the sample. After that, cathodic disbonding spread from the blisters, and it became impossible to measure the disbonded distance from the holiday. After about one year the whole sample had disbonded. We would like to emphasise that this is not a general problem with epoxy mastics. We have tested several commercial epoxy mastics which have perform much better than this. §28 The blistering of Mastic 0 was independent of the initial coating holiday, and was even found on the back of the samples. Inside the blisters we found a highly alkaline liquid, which means that the blistering was caused by the cathodic reaction under the coating. The blistering was therefore also a kind of cathodic disbonding. It then seems that there are two different disbonding mechanisms, one mechanism associated with a mechanical damage in the coating, and another mechanism independent of any mechanical damage. This is in accordance with Type I and Type II failure described by Watts and Castle . Type I failure was associated with downward diffusion of the cation via conductive pathways in the film, corresponding to the random blistering of our coatings. Type II failure was associated with transport of the cation under the disbonded coating from a mechanical damage in the film. For the epoxy mastic coatings the aluminium pigments had a positive effect on both types of failure. §29 Figure 5. Cathodic disbonding of epoxy mastic coating with various concentrations of aluminium and glass pigments. Coatings applied on blast cleaned steel (Sa 21/2). Natural seawater, 8-13�C, -1050 mV SCE.ifscreen(' (Click for a larger image)') §30 As for the amide cured epoxies, the epoxy mastics with 10% and 20% aluminium pigments (Mastic 10 and 20) had no cathodic disbonding at all after 18 months exposure. No blistering or other signs of coating deterioration were observed either. Due to the low content of solvent, the concentration of aluminium in the dry film was lower than for the corresponding amide cured epoxies. Since the unpigmented coating performed so badly, the aluminium pigments had a tremendous effect on both random blistering and cathodic disbonding. §31 In the amide cured epoxy coatings the pigment volume concentration (PVC) varied from 22% in Epoxy 0 to 31% in Epoxy 20 (Table 1). We could therefore suspect that the effect of the aluminium just was a general effect of the increased PVC. However, for the epoxy mastic coatings the PVC was 33% for all the five paints. The effect of the pigments on cathodic disbonding can therefore not just be a general effect of the PVC. Effect of glass pigments on cathodic disbonding §32 Due to spark hazard, the classification companies have set a limit to the concentration of aluminium in marine paints to maximum 10% by weight in the dry film . Glass flake pigments were therefore also tested with the epoxy mastic, as an alternative barrier pigment. The disbonding of these coatings is also shown in Figure 5. The glass flakes inhibited the disbonding compared to the unpigmented coating, but not as much as the aluminium pigments. After 4 months some disbonding was found on Mastic 10G. For Mastic 20G disbonding was first found after a year of exposure. After 9 - 10 months disbonding started from random points all over Mastic 10G, and after one year it was impossible to measure the disbonded distance from the holiday. After 18 months about 70% of the coating had disbonded. For Mastic 20G some random initiation of disbonding was first found after 18 months. §33 The random initiation of cathodic disbonding of Mastic 10G and Mastic 20G was probably due to opening of conductive pathways through the coating, as for the blistering of Mastic 0. The reason why Mastic 10G and Mastic 20G did not develop blisters may have been a reinforcing effect of the glass flake pigments. The random initiation of disbonding for the two films was therefore also a Type I failure, in the terminology by Watts and Castle . Since this type of failure was delayed by the glass flake pigments may, the pigments may have delayed the process of opening conductive pathways through the film. Mechanism for the effect of the pigments §34 Figure 6, 7 and 8 show SEM micrographs of cross sections of Mastic 0, Mastic 10A and Mastic 10G. Immediately after application, the flake shaped aluminium and glass pigments mainly orient themselves parallel to the substrate. In Figure 7 and Figure 8 they therefore appear as needles aligned parallel with the substrate. The irregular shaped particles in all three films is the filler. The aluminium and glass flakes are assumed to improve the barrier properties of the coating by increasing the diffusion pathway through the coating . The diffusing species (oxygen, water and ions) are forced to go around the impermeable particles, and this may increase the diffusion distance several times. The use of barrier pigments can therefore be seen as an alternative to increasing the film thickness. However, the aluminium and glass flakes may also increase the diffusion through the coating. If some pigment flakes are aligned perpendicular to the substrate they may introduce effective pathways for diffusion through the coating. This is extra risky with the glass pigments since they are so large that one pigment flake may penetrate the whole film. Coatings with such pigments must therefore always be applied thicker than the diameter of the largest pigments. §35 Figure 6. SEM micrograph of cross section of an epoxy mastic coating without barrier pigments. Element contrast. ifscreen('(Click for a larger image)') §36 Figure 7. SEM micrograph of cross section of an epoxy mastic coating with 10 wt% aluminium pigments. Element contrast. ifscreen('(Click for a larger image)') §37 Figure 8. SEM micrograph of cross section of an epoxy mastic coating with 10 wt% glass flakes. Element contrast. ifscreen('(Click for a larger image)') §38 Pigments in general, and flake shaped pigments in particular, may also affect the curing of the binder. If the pigments impede diffusion, the curing agent may not be able to diffuse to the reactive sites on the binder chains around the pigment particles. This may then leave a partly cured resin around the pigment particles, which may serve as preferred routes for diffusion. If the resin does not adhere properly to the pigments small voids may be formed around the particles. This may also increase the diffusion through the coating. §39 The initiation of cathodic disbonding from random points underneath the glass pigmented films may have been caused by the pigment flakes. They were wider than the film thickness, and some of them may have been aligned perpendicular to the substrate as described above. However, the films were applied in two coats, which should reduce the chance for pigments to penetrate the film. Since the unpigmented film performed so badly, it is more likely that the random initiation was due to the binder. §40 The barrier effect of the glass and aluminium flake pigments is the most obvious explanation for their effect on cathodic disbonding. The effect of the barrier pigments can be compared to the effect of increasing the film thickness, which also has been shown to decrease cathodic disbonding [2,16,18]. The effect of the aluminium pigments can also involve another mechanism. Leidheiser et al. have suggested that the oxide on the surface of the pigments may act as a buffer . The Pourbaix diagram for aluminium shows that aluminium corrodes at pH higher than about 9. The oxide on the surface dissolves in high pH environment and can not protect the metal. The reaction removes hydroxide from the electrolyte, and this may reduce the pH at the steel/coating interface: Al2O3 + 2 OH- = 2 AlO2- + H2O (3) §41 Since cathodic disbonding has been attributed to the build up of hydroxide at the steel coating interface, this reaction may therefore decrease the rate of disbonding. Another possibility for the effect of aluminium pigments is that they act as an oxygen scavenger. The corrosion of the aluminium pigments may consume the oxygen before it reaches the steel surface. Since the aluminium is not passive, due to the high pH, the corrosion of the pigments (4) is not inhibited: 4 Al + 3 O2 + 4 OH- = 4 AlO2- + 2 H2O (4) §42 Glass is also known to be unstable in highly alkaline solutions, but this requires very high pH, and will probably not affect the pH under the paint film significantly. The fact that the glass pigments also decreased the disbonding weigh against the active pigment theory. However, the two mechanisms may work in combination. Both aluminium and glass pigments may act as barriers. The aluminium pigments was more effective than the glass pigments in decreasing cathodic disbonding, which may be due to the buffer effect in addition to the barrier effect. Both the barrier mechanism and the buffer mechanism will be investigated in detail in Part 2. Conclusions §43 Aluminium and glass barrier pigments decreased cathodic disbonding substantially for two different epoxy coatings. After exposure in natural seawater for two years, coatings with 10% aluminium pigments gave about 3 mm disbonding, while the coatings without the aluminium pigments disbonded about 65 mm. Glass flake pigments were less effective than aluminium pigments in decreasing cathodic disbonding, but still they improved the performance of the coating significantly. The aluminium- and glass pigments act as barriers to diffusion of oxygen and water through the film. However, if transport of cations under the disbonded film is rate limiting, the barrier mechanism alone can not explain the reduced disbonding rate. For the aluminium pigments, a mechanism where the aluminium acts as a buffer that reduces the pH at the steel/coating interface is possible. Acknowledgements §44 We gratefully acknowledge the following companies and institutions for supporting the project: Kv�rner Rosenborg, Aker Engineering, Statoil, Hydro, Saga, Conoco, Scana Offshore, Carboline, Jotun, International, The Norwegian Petroleum Directorate, The Norwegian Research Council. References comment(45) NORSOK, "M-CR-501, Surface Preparation and Protective Coatings", Norwegian Technology Standards Institution, Oslo (1994). H. Leidheiser, W. Wang and L. Igetoft, "The Mechanism for the Cathodic Delamination of Organic Coatings From a Metal Surface", Prog. Org. Coat., Vol: 11, No: 1, p. 19-40 (1983). C. H. Hare, "Protective Coatings - Fundamentals of Chemistry and Composition", Technology Publishing Company, Pittsburgh (1994 W. Funke, "Towards Environmentally Acceptable Corrosion Protection by Organic Coatings - Problems and Realisation", Journal of Coatings Technology, Vol: 55, No: 705, p. 31-38 (1983). H. Foster and O. Vonen, "Glassflake Filled Polyesters for Corrosion Protection", U.K. Corrosion, (1984). S. M. El-Sawy and N. A. Ghanem, "Glass-aluminium physical barrier anticorrosive paints", Journal of Oil Colour Chemist Association, Vol: 67, No: 10, p. 253 (1984) U. R. Evans, Trans. Electrochem. Soc., Vol: 55, No: p. 243 (1929). J. F. Watts and J. E. Castle, "The application of X-ray photoelectron spectroscopy to the study of polymer-to-metal adhesion. Part 2 The cathodic disbondment of epoxy coated mild steel", J. Mat. Sci., Vol: 19, p. 2259-72 (1984). H. Leidheiser Jr, "Corrosion of Painted Metals - a Review", Corrosion, Vol: 38, No: 7, p. 374-383 (1982). J. J. Ritter, J. Coat. Technol., Vol: 54, No: 695, p. 51-57 (1982) R. S. Dickie, J. S. Hammond and J. W. Holubka, Ind. Eng. Chem. Prod. Res. Dev., Vol: 20, p. 339 (1981). M. R. Horner and F. J. Boerio, "Cathodic delamination of an epoxy/polyamide coating from steel", Journal of Adhesion, Vol: 2, No: 3, p. 141-56 (1990). E. N. Kohler, "The Mechanism for Cathodic Disbondment of Protective Organic Coatings - Aqueous Displacement at Elevated pH", Corrosion, Vol: 40, No: 1, p. 5-8 (1984). N. L. Thomas, "The Barrier Properties of Paint Coatings", Progress in Organic Coatings, Vol: 19, p. 101-121 (1991). M. Stratmann, R. Feser and A. Leng, "Corrosion Protection by Organic Films", Electrochimica Acta, Vol: 39, No: 8/9, p. 1207 (1994). X. H. Jin, K. C. Tsay, A. Elbasir and J. D. Scantlebury, "Adhesion and disbonding of chlorinated rubber on mild steel", In: D. Scantlebury and M. Kendig (ed), Advances in Corrosion Protection by Organic Coatings, The Electrochemical Society, Pennington, NJ, 37-47 (1987). M. Kendig, R. Adisson and S. Jeanjackuet, In: D. Scantlebury and M. Kendig (ed), Advances in Corrosion Protection by Organic Coatings, The Electrochemical Society, Pennington, NJ, 2-4 (1989). J. I. Skar, "Cathodic Disbonding of and Transport of Charge through Paint Films", Norwegian Institute of Technology, Trondheim (1990). H. Leidheiser, J. Adhes. Sci. Tech., Vol: 1, No: 1, p. 79 (1987). J. Parks and H. Leidheiser Jr, "Ionic Migration Through Organic Coatings and Its Consequences to Corrosion", Ind. Eng. Chem. Prod. Res. Dev., Vol: 25, No: 1, p. 1-6 (1986). H. Leidheiser, D. J. Mills and W. Bilder, "Permeability of a polybutadiene coating to ions, water and oxygen", In: M. W. Kendig and H. Leidheiser (ed), Corrosion Protection by Organic Coatings, The Electrochemical Society, Pennington, 23-36 (1987). J. D. B. Sharman, J. M. Sykes and T. Handyside, "Cathodic disbonding of chlorinated rubber coatings from steel", Corros. Sci., Vol: 5, No: 8, p. 1375-83 (1993). ISO 8501, Visual assessment of surface cleanliness (1988). ISO 8503-3, Surface roughness characteristics of blast-cleaned steel substrates (1988). ASTM D1141-90, Standard Specification for Substitute Ocean Water (1990). J. F. Watts and J. E. Castle, "The application of X-ray photoelectron spectroscopy to the study of polymer-to-metal adhesion. Part 1 Polybutadiene coated mild steel", J. Mat. Sci., Vol: 18, p. 2987-3003 (1983). DNV, "Guidlines No. 8, Corrosion Protection of Ships", Det Norske Veritas, Oslo (1996).