By Svante Nordänger, Björn Tidbeck and Johan Tidblad, Swerea KIMAB, Sweden
Accelerated corrosion test methods have been used for a long time to predict performance of coating systems in real applications. These methods have also been used for qualifying individual systems, and for benchmarking different systems in particular applications.
Cyclic accelerated test methods, where preferably humidity and temperature vary in a specified way during the test period, have been used for roughly 80 years. Even older are test methods which have constant temperature, humidity and in some cases continuous addition of salt spray, for instance testing in neutral salt spray (NSS)1. Different accelerated methods have been developed over the years depending on which branch of the industry, or company they were used, but they all struggle against a common problem; they do not show comparable results to that achieved during long field exposure.
New coating systems, including new pretreatments, have been developed for different purposes, eg, environmental, health and/or cost, but also ease of maintenance and corrosion protection properties. When it comes to corrosion protection by paint systems, it may be that the development is connected to an accelerated method which may not correspond to realistic outdoor conditions, i.e. the correlation factor is low. This can mean that although a paint system fulfills the specified requirement based on accelerated testing, it can fail quickly under real conditions.
This article describes results from accelerated corrosion testing of paint systems, including pretreatments, compared to field testing for 18, 30 and 43 months in a marine atmosphere. Most of the paint systems and pretreatments tested are classified as relatively new, and more or less “environmentally friendly”, but more traditional systems were also studied. In total, 48 different combinations of steel, pretreatments and paint systems were tested, including wet and powder paints from five international suppliers, and passivating chemicals from three different suppliers. Traditional grit blasting was also included in the test matrix, and in some cases in combination with passivating layers.
The accelerated test methods used in this study (all including chloride addition) were:
The purpose of adding salt solution/spray was to simulate the marine environment and also generally increase the rate of corrosion.
In 2007, a joint project group in Sweden, led by Swerea KIMAB and Swerea IVF, started a project titled Heavy corrosion protection. The main aim was to study the correlation between three standardized accelerated corrosion test methods and long-term field exposure testing in marine atmosphere of different coating. Furthermore, the project group wanted to find out the correlation between short and long field testing times. A further aim was to see if chemical pretreatment can make the total coating system more robust.
SAMPLE PREPARATION, TESTING AND EVALUATION
Cold and hot rolled steel sheets (100 x 150 x 3 mm, from the same batch) with specified coating systems, were used in the study. As pretreatment, five different passivation systems (all free from phosphates, chromium6+ and heavy metals), or grit blasting, were used, including in some cases blasting in combination with passivation, and in one case zinc/manganese phosphating. The passivation chemicals were based on zirconium (one supplier) or silane (two types from each of the suppliers).
In the test matrix, paint systems with 1, 2 and 3 layers, excluding the potential passivation layer, were tested. These included powder coatings, waterborne paints as well as solvent borne paints with low VOC (Volatile Organic Compounds) identified ass HS (high solid) systems, and in a few cases, more traditional solvent borne paints were used as a part of the paint system.
A typical 3 layer solvent borne system used for bridges in a corrosivity class C4 environment was used as a reference system – a zinc rich epoxy primer, micaceous iron oxide pigmented epoxy intermediate layer and an aliphatic PUR topcoat.
All test panels were degreased with an alkaline solution and dried before pretreatment according to the test matrix. Degreasing, pretreatment and painting were performed by Swerea IVF in Gothenburg. In total 3 panels for each type of exposure environment and coating system were produced. After painting, the back and the edges of the panels were protected with epoxy mastic.
The coated test panels were conditioned for at least 1 month after painting and before corrosion testing, a 1 mm wide scribe down to the steel substrate was made using a scratch machine, in order to avoid residues of (zinc) primer on the blasted surfaces, which could lead to the risk of undesired anodic corrosion (figure 1)
The preparation of samples after painting, as well as accelerated and field corrosion testing including evaluation, was performed by Swerea KIMAB except for SCAB testing which was performed by Swerea IVF. The test station at Bohus Malmön (Kvarnvik), on the Swedish west coast, characterized as a marine site (corrosivity class C4-C5M), was used for field corrosion testing (figure 2). This site has a yearly average temperature slightly below 10 C and an average relative humidity around 80%. However, the temperature during summer can reach +25-30 C and in winter between -10 to -15 C. Chloride deposition varies during the year and also annually, but rough average values from 2012 are between 50 and 500 mg/m2/ day, however, peak values can be up to about 1,500 mg/m2, for a day or more.
In order to compare results from the field testing to those from accelerated testing (“ISONEW”, “VICT” and “SCAB”), the different coating systems were ranked according to weighted visible results, based on creep length from the scratch (according to SS 184219), where 1 mm was set to 1 score point, with an addition of score points for visible surface degradation, eg, blistering, flaking, cracking etc (according to ISO 4628, part 2-5) to give a particular rust level. The addition of the surface degradation/rust level corresponded to between 0-50 score points to the mm creep from the scratch. The total score (corrosion value, k) was then expressed in score points or mm. Creep is defined in SS 184219 as the maximum corrosion creep distance at right angles to the scratch. If there is filiform corrosion, see Figure 3, this creep value is reduced by half. After calculating the corrosion value, the ranking of the systems was then simple. The system with the lowest score was ranked as number 1, the system with the second lowest score as number 2 etc.
Panels exposed to accelerated testing were evaluated based on calculated creep from scratch (M) as described in EN-ISO 12944-6, Annex A, ie M=(C-W)/2, where:
C= the maximum perpendicular corrosion distance across the scratch (mm)
W=the original width of the scratch (1 mm)
RESULTS FROM FIELD TESTING AND ACCELERATED CORROSION TESTING
The average corrosion values measured/calculated from the 3 panels of each system, after 43 months of field exposure testing (k 43), are given in detail in Table I below. (These are given in rank order, (r43) starting with the best systems shown in the green columns).
In the same table, the ranking of the coating systems in the field testing at 18 and 30 months exposure, are also given in the table (r18 and r30, respectively), as are the results from the three different accelerated tests, MISONEW, MVICT and MSCAB according to above calculations and abbreviations given at the beginning of the article.
*One of the 3 layer coating systems with zinc primer, epoxy intermediate and PUR toplayer (system 26R) shows surprisingly relatively high average corrosion value in the field testing after 46 months. When looking closer at the results it can be seen that only one of the three panels contributes to that, the other two are quite unaffected. Results from earlier inspections (r18 and r30) confirm excellent results for all panels (ranked as no. 1), which mean that something unexpected has happened with one panel after 30 months.
From this investigation it is quite clear, and not surprising, that coating systems which perform best after 43 months are those with 3 layers, i.e. with a zinc primer, epoxy intermediate layer and a PUR or polysiloxane top layer (see systems 23, 19B, 20, 13 and 19A) and with a total thickness of between 140-300 µm. None of these five systems showed any defects in the coating, except in the region of the intentionally made scribe. One of these top five systems is the reference system which is regularly used for painting bridges (no. 13). It is worth mentioning that system 4 (blasted hot-rolled steel), ranked as number 6, had only two layers (epoxy primer and PUR topcoat), but had a silane treatment after blasting, showed no defects other than those at the intentionally made scribe, on the three tested panels.
None of the systems on cold rolled steel with either a silane or zirconium passivation treatment, and coating systems based on 1 or 2 layers of powder or waterborne paint, or a 1-coat high solid solvent borne paint, showed good results.
When trying to decide whether an extra passivation step contributes to better corrosion resistance on blasted hot rolled steel or not, three cases were compared:
Comparing silane and zirconium treatments on cold rolled steel, in one case, under a waterborne 2 layer system from supplier 3 (system 11 A to 11C), and in the other, under a 1 layer PUR system from supplier 1 (system 15B to 25), both gave similar results, i.e. no difference between zirconium and silane treatment with respect to corrosion protection.
Panels with waterborne 2 layer systems, or with a waterborne primer and a solvent borne topcoat panels 3B-E, 7A, 17A-B), as well as panels with one layer of a high solid solvent borne paint (panels 14A-E) , showed filiform corrosion – figure 3.
CORRELATION BETWEEN LONGER AND SHORTER EXPOSURE TIMES IN FIELD TESTING
The correlations between long and short term exposures in field testing at Bohus Malmön of the different coating systems are shown in figures 4 and 5.
For a good correlation, the coefficient, r, according to Pearson, should be close to +1. For example, if all the points in figures 4 and 5 fell on the line, the correlation would be exactly +1. The summation of the deviation of the points from the line is often calculated as R2, and the square root of it is Pearson´s correlation coefficient, r ( the nearer to +1, the better the correlation). Figure 4 shows a strong correlation between 18 months and 43 months.
Figure 5 shows an even better correlation for 30 months - the Pearson coefficient is 0.914 compared to 0.873 after 18 months. This was expected, but has now been verified.
System 26R (ranked as no. 17), a 3 layer system with zinc primer, epoxy intermediate and PUR topcoat, is an evident outlier (marked with an arrow in figure 4). After 30 months almost no corrosion at all was found on the samples, but after 43 months there was a surprisingly high average corrosion value, giving a lower ranking. However a closer study of the results identified that only one of the three panels contributed to high corrosion rate. Results from earlier inspections (r18 and r30) confirmed excellent results for all 3 panels (ranked as no. 1), which meant that something unexpected had happened with one panel after 30 months.
CORRELATION BETWEEN FIELD TESTING AND ACCELERATED TESTING
Comparing results from field testing to results from the three accelerated tests, the correlation is much weaker than the correlation described above, as shown in Fig 6, where the correlation between field testing after 43 months and accelerated testing according to ISO 11997-2 is described. There have also been other investigations carried out which confirm this weak correlation between accelerated testing and field testing (offshore).2,3
This weak correlation (r=0,568) between field testing for 43 months and ISO 11997-2 is in line with an earlier study performed by Swerea, where other accelerated corrosion testing methods were compared to field testing after 105 months2. Also in that investigation, short term exposure in field testing (19 months) gave a relatively strong correlation to long term exposure (105 months) in the same environment. However, a better correlation with field testing after 43 months was achieved with the semi accelerated testing method ,SCAB, see figure 7 where r=0,824.
It is also evident that testing according to the SCAB protocol for 2 years has much bigger effect on the panels (higher M values, i.e. corrosion from scratch) than accelerated testing according to ISO 11997-2 for 12 weeks.
The correlation between field testing for 43 months and VICT for 6 weeks is roughly in between the correlation for 43 months and ISO 11997-2 and SCAB, but as with ISO 11997-2, the correlation for VICT to field testing is also weak. This is shown in figure 8.
DISCUSSION AND CONCLUSIONS
This investigation has shown that accelerated test methods used to determine performance have a weak correlation to field testing after 43 months in a marine atmosphere. However, when looking at the results from field testing and accelerated testing for the best ranked systems after 43 months of field testing ie the 3 layer systems, it can be seen that low corrosion after field testing (k43 ≤ 2,7 mm) corresponds to relatively low corrosion (low M values) in VICT and SCAB testing, and vice versa.
The best accelerated method tested in this study to predict actual performance is SCAB testing, which is a semi accelerated method, similar to field testing as it also varies temperature and humidity, and uses sunlight and winter conditions similar to a marine environment. The wear of coatings due to wind, dirt, air particles etc., is also similar. The main difference compared to a marine atmosphere however, is the higher amount of salt spray added in SCAB testing compared to that which panels are exposed to in field testing.
Testing according to ISO 11997-2, which is intended to be used for qualification of paint systems on offshore steel constructions, does not appear to be a good method to predict real corrosion performance in marine environment. The best method currently to predict corrosion performance is to perform field testing for a relatively short time, say 18 months, which gives a good correlation with longer term performance.
It is still a challenge to design an accelerated test method that can accurately simulate corrosion and the degradation processes that occur in field exposure, therefore it is difficult to rank the coating system performance based only on accelerated testing.
However, a possible method could be to use the accelerated VICT and SCAB tests ,to compare new coating systems with a reference as a screening tool, and if the new system has similar or better results compared to the reference sample in these two accelerated tests together, it can be classified as “promising” and be qualified for further testing in short term field tests. If the new system is much worse than the reference sample in VICT and SCAB testing, it will have a high probability of poor performance in field testing.
As mentioned above, and as expected, three layer paint systems with a zinc rich primer, an intermediate epoxy layer and a top coating of PUR or polysiloxane coating, were confirmed as the best with respect to corrosion resistance. Also in the top five systems there were three systems with waterborne zinc rich epoxies.
When it comes to passivation of the surface with zirconium or silane based treatments, the results vary. A passivation treatment does not always seem to improve the corrosion resistance with blasted hot rolled surfaces, and regarding the influence of zirconium or silane treatments on corrosion protection on cold rolled steel, no difference between the passivation methods was seen.
1. M.Ström: “A century with salt spray testing: Time for a final phasing-out by a replacement based on newly developed more capable test regimes”, Eurocorr, Pisa, Italy 2014.
2. P.Reuterswärd; J. Tidblad: “Accelerated Corrosion Test Methods – Can they be Misleading? “, JPCL Europe, pp. 18-33, December 2014.
3. O.O. Knudsen; M.Bjordal; U.Steinsmo, S.Nijjer: “Correlation between Four Accelerated Tests and Five Years of Offshore Field Testing”, PCE, pp.52-56, December 2001
The project was mainly financed by VINNOVA (the Innovation Agency of Sweden), with part financed by one of the member research consortia (MRC) within Swerea KIMAB.
The authors would like to thank Patrik Reuterswärd, CPA Consult (earlier at Swerea KIMAB), Arne Finman, Finman Färgkonsult AB and Lars Österberg (both earlier at Swerea IVF) for initiating and doing most of this study with respect to preparing samples and evaluations.
The authors also thank the suppliers of chemicals and paints; Beckers (Sherwin Williams), Candor, Chemetall, Dupont, FreiLacke, Henkel, Jotun and Tikkurila. SSAB is thanked for their contribution of hot and cold rolled steel sheets.
ABOUT THE AUTHORS
Svante Nordänger, a senior researcher at Swerea KIMAB, has a master degree in Chemical Engineering from Chalmers University of Technology, Gothenburg, Sweden. He has worked at Swerea IVF for 10 years, within the areas inorganic surface treatment and corrosion protection. The last two years he has been working with plastics in corrosive environments, organic surface treatment and corrosion protection for steel.
Björn Tidbeck, a researcher at Swerea KIMAB, has a master degree in Chemical Engineering from the Royal Institute of Technology in Stockholm. He has experience in the area of paint systems and corrosion protection within infrastructure. He is also a FROSIO inspector in the area of paint systems and coatings.
Johan Tidblad has worked with atmospheric corrosion modelling for more than 20 years and is manager for Corrosion Protection and Surface Technology at Swerea KIMAB. He is also convenor of ISO TEC 156, WG 7, Corrosion Testing.