《材料失效分析 Materials Failure Analysis》课程教学资源（参考书籍）Steelwork Corrosion Control（Second edition）

carried out in Sheffield in the UK, on 3- in(9.5mm) plates, 66% of the cale was removed in 2 months but only an additional 10%was removed over the next 8 months (c) If the steel is cleaned manually, e.g. by scrapers and wire-brushes loose scale and rust will be removed but intact scale is merely bur- (d)If the steel is then painted, the performance will depend on a number of factors, in particular the chemical and physical state of the surface, the environment of exposure and the paint systems used Where the scale has been virtually all removed, then the surface will be covered with rust and the effects will be as discussed below(Section 313.1) On the other hand, if the scale is practically intact with little or no rusting, then the paint performance may be reasonably good. This sit ation rarely occurs in practice, but in tests carried out in Sheffield on three roups of specimens coated with similar paint systems, the results shown in Table 3.1 were obtained However. the most common situation with scaled steel is where it has been weathered and some intact scale remains with some rust. This covers a range of conditions and possible effects. Generally, there is a consider able reduction in the life of the system but a potentially disastrous situ ation can arise. After normal cleaning, with removal of loose scale, a considerable percentage of apparently intact scale may be left on the surface. However, this scale may have been undermined to a great extent by rust and, after painting, the scale carrying the paint may flake off within a few weeks In tests carried out in a coastal atmosphere on steel speci mens weathered for a period of 3 months, leaving some apparently intact millscale, wire-brushed and then painted, virtually all the paint flaked off within a few months On similar specimens where all the scale had been removed by weathering, the paint coating lost adhesion due to rust after years. Clearly, painting over steel carrying millscale is likely to cause problems and it is generally accepted that visible and identifiable scale should be removed. Before blast-cleaning facilities were as readily available as they are today, this was carried out by leaving steel sections exposed to the Table 3.1 Life of paint system on different steel surfaces Condition Average life of paint system(years) As rolled- intact millscale Weathered -rust and scale De-scaled -no rust or scale 100 C D.A. Bavliss and D. H. Deacon

carried out in Sheffield in the UK, on -in (9.5 mm) plates, 66% of the scale was removed in 2 months but only an additional 10% was removed over the next 8 months. (c) If the steel is cleaned manually, e.g. by scrapers and wire-brushes, loose scale and rust will be removed but intact scale is merely bur￾nished. (d) If the steel is then painted, the performance will depend on a number of factors, in particular the chemical and physical state of the surface, the environment of exposure and the paint systems used. Where the scale has been virtually all removed, then the surface will be covered with rust and the effects will be as discussed below (Section 3.1.3.1). On the other hand, if the scale is practically intact with little or no rusting, then the paint performance may be reasonably good. This situ￾ation rarely occurs in practice, but in tests carried out in Sheffield on three groups of specimens coated with similar paint systems, the results shown in Table 3.1 were obtained. However, the most common situation with scaled steel is where it has been weathered and some intact scale remains with some rust. This covers a range of conditions and possible effects. Generally, there is a consider￾able reduction in the life of the system but a potentially disastrous situ￾ation can arise. After normal cleaning, with removal of loose scale, a considerable percentage of apparently intact scale may be left on the surface. However, this scale may have been undermined to a great extent by rust and, after painting, the scale carrying the paint may flake off within a few weeks. In tests carried out in a coastal atmosphere on steel speci￾mens weathered for a period of 3 months, leaving some apparently intact millscale, wire-brushed and then painted, virtually all the paint flaked off within a few months. On similar specimens where all the scale had been removed by weathering, the paint coating lost adhesion due to rust after about 2 years.3 Clearly, painting over steel carrying millscale is likely to cause problems and it is generally accepted that visible and identifiable scale should be removed. Before blast-cleaning facilities were as readily available as they are today, this was carried out by leaving steel sections exposed to the 24 Steelwork corrosion control Table 3.1 Life of paint system on different steel surfaces4 Condition Average life of paint system (years) As rolled – intact millscale 8.5 Weathered – rust and scale 2.6 De-scaled – no rust or scale 10.0 © 2002 D. A. Bayliss and D. H. Deacon

atmosphere so that all the scale eventually flakes off. However, this is achieved only by allowing the scale to be undermined by rusting, so leaving a layer of rust on the steel and some pitting. This is not a sound substrate on which to apply coatings, as discussed below 3.1.3 Surface cleanliness 3.1.3. Rust Rust is the corrosion product formed when steel reacts with oxygen and water. This has been discussed in Chapter 2. The corrosion reaction is generally denoted as follows: 4Fe +2H,0+302= 2Fe,O3.H,O Although rust is primarily hydrated ferric oxide, it also contains other nds. Rusts have a wide range of composition, dep conditions under which they are formed. Typical compositions cannot, therefore, be given but analyses of a range of rusts have indicated that air formed rusts generally contain about 5% of compounds other than Fe,O3.H,O. These derive in part from the steel, which contains elements other than iron, e.g. copper, silicon and manganese, and in part from atmospheric contaminants and pollutants, mainly sulphates and chlorides although other pollutants such as ammonium salts are also generally present in rust 3.3.2 Water-soluble contaminants The main constituents of rust. ie iron oxides are not themselves the main problem in determining the performance of paint applied over rust. In fact, iron oxides are commonly used as pigments in paints. The con stituents formed by reactions between steel and pollutants such as sulphur dioxide. sometimes called 'iron salts' or 'corrosion salts,. cause most roblems Sulphur dioxide in the air reacts with moisture to form acids. Consider- able publicity is frequently given to what has been described asacid arising from such reactions. Weak sulphuric acid solutions react with the steel to form ferrous sulphate, the presence of which can be readily detected in rust. These salts tend to form in shallow pits at the steel surface and the corrosion process is such that the sulphates tend to move inwards to the anodic areas, which are likely to be in crevices, the bottom of pits, etc. The salts are also not rustcoloured, being white or light coloured. They are very difficult, if not impossible, to remove with tools C D.A. Bavliss and D. H. Deacon

atmosphere so that all the scale eventually flakes off. However, this is achieved only by allowing the scale to be undermined by rusting, so leaving a layer of rust on the steel and some pitting. This is not a sound substrate on which to apply coatings, as discussed below. 3.1.3 Surface cleanliness 3.1.3.1 Rust Rust is the corrosion product formed when steel reacts with oxygen and water. This has been discussed in Chapter 2. The corrosion reaction is generally denoted as follows: 4Fe
2H2O
3O2 2Fe2O3. H2O (rust) Although rust is primarily hydrated ferric oxide, it also contains other compounds. Rusts have a wide range of composition, depending on the conditions under which they are formed. Typical compositions cannot, therefore, be given but analyses of a range of rusts have indicated that air￾formed rusts generally contain about 5% of compounds other than Fe2O3.H2O. These derive in part from the steel, which contains elements other than iron, e.g. copper, silicon and manganese, and in part from atmospheric contaminants and pollutants, mainly sulphates and chlorides, although other pollutants such as ammonium salts are also generally present in rust. 3.1.3.2 Water-soluble contaminants The main constituents of rust, i.e. iron oxides, are not themselves the main problem in determining the performance of paint applied over rust. In fact, iron oxides are commonly used as pigments in paints. The con￾stituents formed by reactions between steel and pollutants such as sulphur dioxide, sometimes called ‘iron salts’ or ‘corrosion salts’, cause most problems. Sulphur dioxide in the air reacts with moisture to form acids. Consider￾able publicity is frequently given to what has been described as ‘acid rain’ arising from such reactions. Weak sulphuric acid solutions react with the steel to form ferrous sulphate, the presence of which can be readily detected in rust. These salts tend to form in shallow pits at the steel surface and the corrosion process is such that the sulphates tend to move inwards to the anodic areas, which are likely to be in crevices, the bottom of pits, etc. The salts are also not ‘rust’ coloured, being white or light coloured. They are very difficult, if not impossible, to remove with tools Surface preparation 25 © 2002 D. A. Bayliss and D. H. Deacon

such as scrapers and wire brushes and are often difficult to remove even with blast-cleaning. The presence of salts such as ferrous sulphate leads to rather complex reactions involving the regeneration of the sulphuric acid from which they were formed. This in turn causes further corrosion and the production of more rust(see Figure 3.2). As rust has a considerably greater volume than the steel from which it is produced, this leads to dis- ruption of the paint film applied over it by cracking, blistering and eventu- ally flaking(Figure 3.3) SULPHUR DIOKIDE cHL。RDe STEE MOISTUR日 DROCHLORIC SULPHURIC F雪 ROus FFRROUS SULPHATE cuL。R。 ACID RUST RON OXIDE Figure 3.2 The cyclic process of rusting 今 Figure 3.3 Blistering of paint resulting from the presence of soluble iron salts under the oating. C D.A. Bavliss and D. H. Deacon

such as scrapers and wire brushes and are often difficult to remove even with blast-cleaning. The presence of salts such as ferrous sulphate leads to rather complex reactions involving the regeneration of the sulphuric acid from which they were formed. This in turn causes further corrosion and the production of more rust (see Figure 3.2). As rust has a considerably greater volume than the steel from which it is produced, this leads to dis￾ruption of the paint film applied over it by cracking, blistering and eventu￾ally flaking (Figure 3.3). 26 Steelwork corrosion control Figure 3.3 Blistering of paint resulting from the presence of soluble iron salts under the coating. Figure 3.2 The cyclic process of rusting. © 2002 D. A. Bayliss and D. H. Deacon

Ferrous sulphate is the salt most commonly found in rusts formed in industrial-type atmospheres. Near the coast, chlorides are likely to be greater problem. The reactions arising from the two types of salt, sulphate and chloride, are not necessarily the same. Chlorides are hygroscopic, i.e. they absorb moisture. It has been shown in laboratory tests that whereas rusting may occur at relative humidities below 70% with sulphates, the presence of chlorides in rust can result in corrosion of the steel at relative humidities as low as 40%. Chlorides may, therefore, be a greater imme diate problem than sulphates, but all salts present under a paint film will lead to a reduction in the coatings life. ISO and bs Standards call these ferrous salts soluble iron corrosion products'. In American literature they have been called 'non-visible contamination which is particularly The rusting of steel is complex and in many ways unpredictable. An investigation of the process was carried out by the former British Iron and Steel Research Association(BISRA) and some of the results have been published. These show that the amount of rust formed on a steel surface is not necessarily related to the length of time the steel has been exposed and, perhaps even more important, the amount of sulphate in the rust also does not relate to the length of exposure In rusts sampled in January, about 8g/m" of sulphate were measured in rusts formed over a period of 2 months, i. e. from steel exposed initially in November of the previous year This rose to a figure of about 12g/m for rusts formed over a period of a year. However, for rusts sampled in the summer months much lower sul- phate contents were obtained. In July, rusts formed over a period of a year contained about 6g/m", and over 2 months contained 2.5g/m. It follows, therefore, that irrespective of the period of rust formation, the amount of sulphate is higher in winter. Consequently, painting over rusted steel is a somewhat haphazard operation because, without carrying out chemical tests on the rust, it is virtually impossible to know the extent of iron salt formation. Painting in the summer at inland sites in the Uk is likely to provide better performance from the paint coating than in winter Similarly, chloride contamination in coastal areas can depend upon pre- vailing wind direction and even the steepness or shallowness of the coast- line. Against this must be set the fact that the chloride has a higher solubility than the sulphate and therefore more is washed from the surface by rainfall. For both contaminants, the situation where rusted surfaces are subject to atmospheric pollution and are not washed by rain is the most aggressive The reduction in durability of coatings due to soluble iron corrosion products trapped beneath coatings is most obvious for surfaces exposed to severe marine environments or frequent condensation, and also for the linings of storage tanks containing aqueous liquids. This effect is less obvious on coatings subject to normal weathering but this can also depend C D.A. Bavliss and D. H. Deacon

Ferrous sulphate is the salt most commonly found in rusts formed in industrial-type atmospheres. Near the coast, chlorides are likely to be a greater problem. The reactions arising from the two types of salt, sulphate and chloride, are not necessarily the same. Chlorides are hygroscopic, i.e. they absorb moisture. It has been shown in laboratory tests5 that whereas rusting may occur at relative humidities below 70% with sulphates, the presence of chlorides in rust can result in corrosion of the steel at relative humidities as low as 40%. Chlorides may, therefore, be a greater imme￾diate problem than sulphates, but all salts present under a paint film will lead to a reduction in the coating’s life. ISO and BS Standards call these ferrous salts ‘soluble iron corrosion products’. In American literature they have been called ‘non-visible contamination’ which is particularly appropriate. The rusting of steel is complex and in many ways unpredictable. An investigation of the process was carried out by the former British Iron and Steel Research Association (BISRA) and some of the results have been published.6 These show that the amount of rust formed on a steel surface is not necessarily related to the length of time the steel has been exposed and, perhaps even more important, the amount of sulphate in the rust also does not relate to the length of exposure. In rusts sampled in January, about 8 g/m2 of sulphate were measured in rusts formed over a period of 2 months, i.e. from steel exposed initially in November of the previous year. This rose to a figure of about 12 g/m2 for rusts formed over a period of a year. However, for rusts sampled in the summer months much lower sul￾phate contents were obtained. In July, rusts formed over a period of a year contained about 6 g/m2 , and over 2 months contained 2.5 g/m2 . It follows, therefore, that irrespective of the period of rust formation, the amount of sulphate is higher in winter. Consequently, painting over rusted steel is a somewhat haphazard operation because, without carrying out chemical tests on the rust, it is virtually impossible to know the extent of iron salt formation. Painting in the summer at inland sites in the UK is likely to provide better performance from the paint coating than in winter. Similarly, chloride contamination in coastal areas can depend upon pre￾vailing wind direction and even the steepness or shallowness of the coast￾line. Against this must be set the fact that the chloride has a higher solubility than the sulphate and therefore more is washed from the surface by rainfall. For both contaminants, the situation where rusted surfaces are subject to atmospheric pollution and are not washed by rain is the most aggressive. The reduction in durability of coatings due to soluble iron corrosion products trapped beneath coatings is most obvious for surfaces exposed to severe marine environments or frequent condensation, and also for the linings of storage tanks containing aqueous liquids. This effect is less obvious on coatings subject to normal weathering but this can also depend Surface preparation 27 © 2002 D. A. Bayliss and D. H. Deacon

upon the type of coating. Some coatings are thick and have a high cohesive strength and others are thick and are very flexible; in both these cases sub- strate corrosion can be masked until it has reached an advanced state. For conventional paint systems, such as alkyds, the effect can generally be seen within a few years by the appearance of corrosion blisters in the paint- work. The corrosion blisters form from the underlying corrosion pits The fact that the presence of ferrous salts in pits, even after blast cleaning to a high visual standard, can seriously affect coating durability has been known since the work of Chandler in 1966. It is still not possible to quantify permissible levels of soluble salts for different coatings in dif- ferent environments Most of the reliable information on the effects of water-soluble salts on performance of coatings is from the marine industry and relates to coat subjected to immersion. Here the soluble contaminants are largely chlorides from seawater and therefore the detection methods used for monitoring are either specifically for chlorides or, more commonly, by measuring conductivity. The higher the quantity of dissolved salts in water. the lower the resistance. Typically levels of 5 ug/cm"of chloride or less are considered acceptable maximum levels by Jeffrey. Information on acceptable levels of soluble salts is scarce for less demanding environments. It would be advisable to assume that, for any paint system where the longest durability is required, the initial state of he steel is Rust Grade D or worse, and the coating is subjected to some degree of wetness, that a maximum level of 15 mg/m2 of soluble iron corro- sion products as measured by ISo/TR 8502-1 would be desirable However, a requirement for excessively low limits for non-aggressive environments could be very costly and probably not justified. A Working Party of the iso Committee dealing with surface preparation of steel sub strates is charged with providing guidance levels, but the general opinion hat it may be some time before this is possible The also now ample evidence from the Highways Agency experience that conventional oleo-resinous paint systems, normally with a life expectancy of 5-7 years, have lasted at least 2 or 3 times longer. One of the essential ingredients for such success is the monitoring of blast cleaned surfaces to ensure they are free from contamination All of the current methods of determining soluble contaminants are dis cussed in Chapter 9, but as a guide it can be assumed that the deeper the corrosion pitting before surface preparation, the greater the problem. This is another sound economic reason why no area, however small, of a painted structure should be allowed to deteriorate into severe corrosion before maintenance is carried out C D.A. Bavliss and D. H. Deacon

upon the type of coating. Some coatings are thick and have a high cohesive strength and others are thick and are very flexible; in both these cases sub￾strate corrosion can be masked until it has reached an advanced state. For conventional paint systems, such as alkyds, the effect can generally be seen within a few years by the appearance of corrosion blisters in the paint￾work. The corrosion blisters form from the underlying corrosion pits. The fact that the presence of ferrous salts in pits, even after blast￾cleaning to a high visual standard, can seriously affect coating durability has been known since the work of Chandler in 1966.7 It is still not possible to quantify permissible levels of soluble salts for different coatings in dif￾ferent environments. Most of the reliable information on the effects of water-soluble salts on performance of coatings is from the marine industry and relates to coat￾ings subjected to immersion. Here the soluble contaminants are largely chlorides from seawater and therefore the detection methods used for monitoring are either specifically for chlorides or, more commonly, by measuring conductivity. The higher the quantity of dissolved salts in water, the lower the resistance. Typically levels of 5 µg/cm2 of chloride or less are considered acceptable maximum levels by Jeffrey.8 Information on acceptable levels of soluble salts is scarce for less demanding environments. It would be advisable to assume that, for any paint system where the longest durability is required, the initial state of the steel is Rust Grade D9 or worse, and the coating is subjected to some degree of wetness, that a maximum level of 15mg/m2 of soluble iron corro￾sion products as measured by ISO/TR 8502-1 would be desirable. However, a requirement for excessively low limits for non-aggressive environments could be very costly and probably not justified. A Working Party of the ISO Committee dealing with surface preparation of steel sub￾strates is charged with providing guidance levels, but the general opinion is that it may be some time before this is possible. There is also now ample evidence from the Highways Agency10 experience that conventional oleo-resinous paint systems, normally with a life expectancy of 5–7 years, have lasted at least 2 or 3 times longer. One of the essential ingredients for such success is the monitoring of blast￾cleaned surfaces to ensure they are free from contamination. All of the current methods of determining soluble contaminants are dis￾cussed in Chapter 9, but as a guide it can be assumed that the deeper the corrosion pitting before surface preparation, the greater the problem. This is another sound economic reason why no area, however small, of a painted structure should be allowed to deteriorate into severe corrosion before maintenance is carried out. 28 Steelwork corrosion control © 2002 D. A. Bayliss and D. H. Deacon