Forms of Corrosion
Chapter III of The Causes and Prevention of Corrosion by Alan A. Pollitt, Ernest Benn, London, 1923
If the Ionic or Electro-chemical theory of corrosion, as described in the preceding chapter, be accepted it is then easy to divide all practical cases of corrosion into three types : (1) Auto Corrosion, (2) Contact Corrosion, and (3) Externally Induced Corrosion. (reference)
The third type of corrosion is the result of the passage of a current, generated from some external source, through the metal whilst the latter is in contact with an electrolyte. If the current flows in that direction which necessitates the metal acting as anode, then corrosion results. If the current flows in the opposite direction, i.e., from the electrolyte into the metal, the latter receives protection from corrosion which may be complete provided the E.M.F. of the current is sufficiently high. The recognition of this fact has been turned to practical account in the Cumberland system of preventing corrosion, of which more will be said later.
Externally induced corrosion is not so common as the other two types, though its frequency has no doubt increased with the more widespread use of electrical power. Thus, the corrosion of the feet of the standards used to carry the overhead cables for electric trams and trains, may often be largely due to leakage from the power transmission, and cases of corrosion in various units of ships equipment have also been traced to the presence of stray fields and leakages. This type of corrosion is nothing more nor less than simple electrolysis, and the corrosion of the metal is strictly comparable with the corrosion of the metal anodes in ordinary plating processes. In the case of iron and steel embedded in the ground, the metal and the soil act as electrodes and the moisture present is decomposed by the current into hydrogen and oxygen. The latter effects the oxidation or rusting of the metal in place of the atmospheric oxygen, whose rate of diffusion through the soil or clay is often so slow that corrosion, if it were dependent on atmospheric oxygen alone, would be extremely slow. The most commonly occurring types of corrosion are Auto and Contact corrosion.
Auto corrosion is that which occurs when a metal is in contact with an electrolyte but is not at the same time in contact with any other electrical conductor, neither metallic nor non-metallic. In the academic case of chemically pure iron, corrosion would proceed simply by the exertion of the solution pressure of the metal, in conjunction with the presence of hydrogen ions and the oxygen dissolved in the electrolyte, which both depolarises the metal surface and oxidises and precipitates the primary products of solution as ferric hydrate or rust. In practical cases, auto corrosion proceeds by the galvanic action which is set up as a result of the heterogeneous structure of the metal or alloy. No commercial metal exists in which there is perfect homogeneity, there is always some characteristic of structure, some slight degree of segregation or the presence of embedded impurities which is sufficient to impart varying potentials or solution pressures to adjacent areas of the metal surface. For this reason auto electrolysis is set up by which the more electro-positive areas dissolve and, in the case of ferrous material, are eventually precipitated as rust. The pronounced heterogeneity o$\ some alloys, such, for instance, as brass, is no ' doubt largely responsible for the rapidity with which they frequently corrode, and in the case of iron and iron alloys there is a large volume of evidence to show that heterogeneity, whether induced by structure or segregation, etc., is conducive to accelerated corrosion.
Contact corrosion occurs when the metal is in contact with some other conducting material, which is also immersed either wholly or partially in the electrolyte. If this other conductor is a metal, then the corrosion of the first metal will be either accelerated or retarded, according to the electro-chemical relationship between the two metals. If the second metal is electro-positive to the first, then it will protect the latter at its own expense by itself corroding or dissolving preferentially, but if it be electro-negative to the first metal then the corrosion of this will be accelerated (or the second metal Will receive protection at the expense of the first). The practical recognition and application of this may be found in the practice of protecting boilers from corrosion by inserting slabs of the more electro-positive metal zinc and in the protective coatings of zinc which are applied to iron products by various processes. Other conditions being the same, the rate of the contact corrosion of a metal is usually greater than the rate of its auto corrosion. If the second conductor is nonmetallic in character, it may generally be assumed to be electro-negative to the metal, and contact between them will therefore result in an accelerated corrosion of the metal.
In both types, auto and contact corrosion, galvanic action is the primary cause of corrosion, each process is one of electrolysis, but in auto corrosion the electrolysis is self-induced by inherent characteristics of the metal, and may therefore be distinguished as auto electrolysis, whilst in contact corrosion the electrolysis is the result of the contact between the metal and some other conductor to which it is electro-positive. Since galvanism, then, is responsible for the majority of corrosion troubles, it may be desirable to discuss this action in more detail.
From what has already been said there will be no difficulty in differentiating between galvanic and electrical action, since the latter always implies an externally generated current. The actions are alike in principle, but their results differ greatly in degree and in character. Electrical action usually tends to induce a more or less uniform corrosion over the entire surface of the metal which conies within its sphere of influence, -but galvanic action, by reason of the low E.M.F. of the local currents which are generated, and also because of the fact that it may result from the potential differences between adjacent areas (as in auto corrosion) on the same surface, more frequently results in a selective or preferential corrosion of certain portions of the metal. For instance, an electro-positive area in contact with one of lower potential will dissolve in preference to the latter, and the final result is apparent in " pitting." A good example of this is to be found in the corrosion of brass, in. which the condition generally known as " dezincification " is often produced. The nett result in these cases is the removal of the more electro-positive constituent of the alloy, i.e., the zinc, which leaves the copper as a porous and spongy residue.
Galvanic action, whether induced by the characteristics of the corroding metal itself or by contact between dissimilar metals, is caused by differences in electrical potentials, or, in other words, and for our purpose more descriptive words, by differences in solution pressures. If two metals are in contact with each other and with an electrolyte, the one with the higher solution pressure will dissolve, and is said to be electro-positive to the other metal, which has the lower solution pressure and which does not dissolve. If a piece of iron and a piece of copper are placed in contact with each other in dilute sulphuric acid, the iron dissolves and the copper remains unattacked. Replace the copper by zinc and in this case the iron is unattacked whilst the zinc dissolves. The zinc, therefore, is electro-positive to the iron and similarly the iron is electro-positive to the copper, and in the electrolyte the direction of the galvanic current will be respectively from the iron to the copper and from the zinc to the iron, that is, from the more electro-positive metal to the less electro-positive metal. It is clear, then, that the more electro-positive component of such couples functions anodically, and conversely it may be correctly inferred that whenever a metal functions cathodically it does not corrode.
Referring again for a moment to externally induced corrosion, these observations give further emphasis to the statement that for an externally generated current to cause corrosion it must flow from the metal into the electrolyte, and that in cases where it flows in the opposite direction its action tends to be, and will be, protective, if its E.M.F. is of sufficient magnitude. Such methods of electrolytic protection as the Cumberland system (Trans. Faraday Soc., 11, 1915-16) are therefore theoretically sound, as also is the practice of boiler protection with zinc slabs, whereby .a protective galvanic current is produced for which the metal of the boiler acts as cathode, receiving the current from the anodic zinc through the water. Unfortunately, however, secondary influences arise in practice which seriously reduce the efficiency of this type of theoretically sound protective measure, in some cases even reversing the intended order of things and thus accelerating corrosion.
The following table gives a list of the commoner metals arranged in order of their solution pressures or electrical potentials. Any metal in the series is electronegative to all metals which precede it and electro-positive to all which follow it :
If any two of the above metals are immersed in an electrolyte, either in contact with each other, or kept apart and connected externally instead by a wire, then the one which precedes the other in the series will dissolve or become the anode for the current generated, whilst the one appearing later in the series becomes the cathode and does not enter solution.
For the large majority of cases this may be accepted as the rule, but it must be remembered that the relationship between the solution pressures of two metals may be influenced by other factors such as the concentration, composition, and temperature of the electrolyte. It does not invariably follow, therefore, that, because under one set of conditions a certain metal exhibits a higher solution pressure than a second, this will always be the case. The relationship may be reversed (reversal of polarity), so that the metal which was originally electro-positive is now electro-negative, or the magnitude of the difference between the two solution pressures may be increased or reduced, as evidenced by an accelerated or retarded rate of corrosion or solution of the anodic metal.
This offers some explanation for many curious anomalies which contact with corrosion troubles is sure to present sooner or later, and at once points out the foolishness and futility of attempting to form a generally applicable estimate of the relative corrodibilities of various metals, say of cast iron and steel, for instance. To be of any practical value such comparisons will have to be made for each specific set of conditions, and the choice of either cast iron or of a particular kind of steel must be made, not from a general consideration as to which material is generally regarded as least corrodible, but from a consideration as to which metal has been proved by experience or by definite experiment to be the most suitable for use under those particular conditions to which it will be exposed during service. There have been innumerable assertions by manufacturers of various kinds of ferrous materials as to their excellences in rust-resisting properties, but the laboratory tests or field tests upon which these statements are based vary so widely and represent such a limited range of practical conditions that the intelligent consumer pays but little attention to them nowadays, and chooses his material, as far as possible, from considerations in line with the above.
To continue with the discussion of galvanic action, it has been pointed out that chemical heterogeneity, whether the result of segregation, presence of impurities, or peculiar structure, is conducive to galvanic electrolysis or auto corrosion. It is also known that physical heterogeneity has a like effect. Strained portions of a piece of metal exhibit different potentials to the unstrained portions. Experimental proof of this may readily be obtained by taking a bar of metal, say iron or steel, cutting it into two halves, and then straining one of the pieces in some way, either by tension or torsion, etc. If the two pieces are now placed apart in an electrolyte and connected externally by wires through a galvanometer, it will be seen from the deflection of the latter that a current is flowing, thus demonstrating that the two pieces of metal, originally alike, now have different solution pressures. There is no general rule by which we may predict whether the strained metal will be electropositive or electro-negative to the unstrained metal, this relationship being largely dependent on other local influences. The modification in the potential produced by straining is always of very small magnitude.
Two characteristic and frequently occurring practical illustrations of the effect of straining a metal are to be found in the rapid corrosion which may be often observed to take place on the metal immediately surrounding the rivet holes which have been punched in steel plates and on the bends on angle irons, etc. In each case the metal in these areas has been subjected to severe strain, and where the conditions are such that the strained metal behaves electro-positively to the general bulk of the metal, these portions corrode more rapidly.
An ingenious and convincing method of demonstrating the galvanic mechanism of auto corrosion was perfected by Cushman and Walker (J. Amer. Chem. Soc., 1907, 25, 1257, and J. Iron and Steel Inst., 1909, 1, 33), who introduced the reagent now known as ferroxyl.
When a strip of iron corrodes by auto corrosion there will be, at the positive areas or points, a concentration of ferrous ions, whilst at the negative points, where no solution of the metal occurs, there will be a concentration of hydroxyl ions. If a trace of phenolphthalein is present in the electrolyte it will indicate the electronegative or cathodic areas by the development of a pink colour, and if, in addition, there is also present a little potassium ferricyanide, this will show up the electropositive or anodic areas by the production of a blue colour, due to the formation of Turnbull's blue by interreaction between the ferrous ions and the potassium ferricyanide. Cushman and Walker prepared the ferroxyl reagent by stiffening the electrolyte with agar-agar or gelatine, so that it sets to a jelly on cooling. The result is that the colours produced are much more permanent and less affected by external influences.
If a piece of iron is properly mounted in this reagent, the colours which develop are definite proof that the solution pressure of the metal varies at different points on its surface, and that its corrosion proceeds, therefore, by galvanic electrolytic action. Fig. 5 shows three needles mounted in ferroxyl. The strongly-defined blue areas and the somewhat hazy pink areas indicate the non-uniformity of the metal, which was taken direct from the packet in which it was purchased and simply cleaned in aclohol before being mounted. Fig. 6 is a further illustration in which the metallic objects consist of two needles and an ordinary nail. The reproductions of the preparations are not all that could be desired, but the fact that there are three strong colours, the yellow of the ferroxyl mount, the blue, and the pink, does not facilitate faithful photographic reproduction.
The use of this reagent affords also convincing proof of the correctness of the Electro-chemical or Ionic theory of corrosion. The ferroxyl is extremely sensitive and has been of considerable value in the theoretical study of corrosion. It shows very plainly that even the purest iron develops points of different potential in an electrolyte, and that, therefore, perfect homogeneity would appear to be an unattainable ideal. The more pronounced the heterogeneity of the specimen, the stronger are the colours produced, the more clearly defined are their outlines, and the greater is the speed with which they are developed.
A further interesting phenomenon which is clearly evidenced by ferroxyl is the reversal of polarity which may occur between electrodic points on a corroding iron surface. When corrosion first commences the cathodic and anodic areas are indicated by the pink and blue colours as already described. After a time, however, it may sometimes be observed that the originally blue areas are dispersing and are being replaced by pink zones, whilst the pink zones are being replaced by blue ones. The explanation of this is not difficult to find. The continued solution of an anodic area may result in the exposing of an area which is electro-negative to the original cathodic area so that the latter now commences to act anodically. Such reversals in polarity between points on the metal surface may occur frequently, and it is not assuming too much to say that the more homogeneous the metal, the greater the frequency of reversal, and hence the more uniform the corrosion. Conversely, the more pronounced the segregation or heterogeneity, the greater the tendency to corrosion in the form of pitting.
Whilst the ferroxyl reagent is of undoubted value in demonstrating the electrochemical mechanism of the corrosion of iron and the effects which segregation and local strain tend to produce, or induce, yet it would not appear to be a warrantable procedure to apply unreservedly the inferences from observations of the behaviours of metals in the reagent to actual practice. In the first place, corrosion in the reagent takes place under unique and highly specialised conditions (conditions which would never obtain in practice) which remain comparatively constant and which are but little affected by external influences. The conditions under which metal corrodes whilst in service are infinitely more complex and are constantly being aggravated or alleviated, added to, or reduced in number. The anodic portions of specimens mounted in ferroxyl may be proved to coincide with known variations in the nature of the material, such as the presence of impurity, segregation or strain, and though it may be correctly inferred that such variations facilitate electro-chemical action, it does not necessarily follow, for instance, that a strained portion of a piece of metal will act anodically in service. The reagent is so designed that it is very sensitive to extremely slight differences in electrical potentials, the effects of which may readily be masked, reversed, or rendered negligible in the presence of other corrosive factors which occur in practice.
The recognition of the " passive " or chemically inactive state of iron is one of long standing, but up to the present time no theory has been advanced which gives a thoroughly satisfactorily or complete explanation of the phenomenon. It is beyond doubt that passivity, however induced, is wholly associated with the surface film or layer of the metal, and the behaviour of passive iron may be due equally well to either a physical or a chemical change in this layer. Many instances might be quoted to illustrate the varying degrees of chemical activity which may be conferred upon a substance by alterations in its physical condition, the chemical composition and structure remaining, withal, unchanged.
A strict criticism of what is implied by the term " passive," i.e., chemical inactivity of the METAL, would necessitate its interpretation as a definite or peculiar physical condition ; for if passivity is the result of a change in the chemical condition, such, for instance, as the deposition of an oxide or gaseous film, then the activity of the underlying metal is obviously unimpaired, but is prevented from manifesting itself by the presence of a protective film or coating whose extremely low solution pressure renders it more or less permanent. Hence if passivity is due to such a protective film on the metal surface, the term cannot be rightly regarded as denoting a specific condition or property of the metal.
As generally understood, however, "passive" is the term used to indicate that condition of the metal in which the appearance of its surface is not visibly altered, but in which its chemical activity or solution pressure is reduced almost to nil. The condition is always fugitive, though the length of its duration may vary according to the method by which it is induced, and it is considerably lengthened if the passified iron is kept dry (Heathcote, J. Soc. Chem. Ind., 1907, 26, 899).
It is usual, for purposes of demonstration, to passify iron by immersion in strong nitric acid, and it was in this manner that the phenomenon was originally observed. It was soon found, however, that other acids, such as chromic, iodic, and chloric acids, would induce the same condition as well as mixtures of acids and salts, or even aqueous solutions of certain salts, for instance, lead and silver nitrates, permanganate, and bichromate of potassium, provided suitable concentrations were employed. Passivity may also be induced by momentarily heating the metal in air or by exposure to certain gases, such as nitric oxide and nitric acid fumes. Iron may also be rendered passive by making it the anode during the electrolysis of an aqueous electrolyte, usually caustic soda.
The outstanding property of passive iron is, of course, its reluctance to enter into those reactions which characterise the active metal. It will not rust, for instance, and it is insoluble in acids, unless made the anode for an electrolysing current. Its rust-resisting properties have naturally attracted attention with a view to producing a permanently passive condition, but so far without success, so that the passive condition cannot yet be turned directly to any practical account, though certain passifying agents are of considerable importance in the protection of iron and steel. The other properties are only of significance in illustrating the methods by which passivity may be destroyed, etc., and in providing evidence either for or against the several theories which have been suggested to explain the nature of the passive condition.
The oldest theory is, strange to say, the one which has proved to be the most compatible with the facts which subsequent research has brought to light, and is therefore the one which has received the most generous adoption. That it is not wholly beyond criticism will soon become apparent, and although it is the generally accepted explanation, yet it is felt that an entirely satisfactory theory has still to be evolved.
The Oxide theory assumes that passivity is the result of the production of a film of metallic oxide, which covers the entire surface of the metal and prevents it from coming into contact with reagents to whose attack it is susceptible. The film may-must, in fact-be so thin that it is impossible to detect its presence even microscopically. An iron mirror may be rendered passive without incurring any estimable reduction in its reflecting power. Provided it is continuous and unbroken, the film may be so tenuous that its thickness may be regarded as being of molecular dimensions. Its composition has been the cause of much speculation, and although several alternatives have been suggested it is generally considered to consist of ferroso-ferric oxide, Fe3O4.
A second theory attributes passivity to the formation of an adherent film of gas which acts similarly to the film of metallic oxide discussed above. The gas may vary in kind according to the method used in passifying; thus it has been suggested that passivity induced by anodic polarisation in sulphuric acid is the result of a film of gaseous oxygen, whilst passivity produced by immersion in strong nitric acid is due to an oxide of nitrogen film. This is known as the Gaseous Film theory.
The third, or Physical theory, presumes an alteration in the physical or electrical condition of the surface of the metal by passifying agents, such as the production of trivalent iron. How far this assumption is justified it is difficult to say, there is no evidence to show that trivalent iron is inactive chemically ; this is certainly not the case when trivalent iron occurs in combination with other elements. If, however, passivity is really a metallic property and not the result of protection as assumed by the Oxide and Gaseous Film theories, then the Physical theory would appear to be the most rational, but since we do not know whether passivity is a metallic property or not, this theory must be accepted or rejected on its merits as revealed by a careful examination of the evidence available. Since, however, the passivities produced by alternative methods are not in every case identical in their behaviours, it may be possible that each theory is applicable to certain forms of passivity.
The evidence from which an estimate of the comparative validities of these theories may be formed may be summarised in the following manner by considering some well-established properties of passive iron. When passivity results from immersion in a liquid media the latter is invariably an oxidising agent. This provides considerable evidence in favour of the Oxide theory, but it does not necessarily reflect adversely on the other two theories. The gaseous oxygen which may be liberated on the surface of the metal would conform to the Gaseous Film theory, since this oxygen need not combine with the iron. It is known that the metal may remain in contact with oxygen, under certain conditions, for an indefinite period without any combination occurring between the two elements. We are also bound to regard the film of gas as sufficiently continuous and impenetrable to prevent contact between the iron and the liquid medium, and in the absence of such contact reaction between the metal and the oxygen will not take place. Again, contact with an oxidising liquid may be looked upon as particularly favourable to the production of a physical change in the metal, such as the formation of trivalent iron or other modification in the electrical condition of the metal.
Passive iron is readily restored to its original active condition by various chemical and mechanical means. Immersion in dilute acids destroys passivity, and the rate of destruction is accelerated by those external influences, such as elevation of temperature and mechanical agitation, which usually increase the rates of chemical reactions. Lightly scratching the surface, or rubbing, renders passive metal active, as does also the galvanic effect produced by contact, whilst in an electrolyte, with a piece of a more electro-positive metal, i.e., zinc. These observations lend support to the Oxide theory and also to the Gaseous film theory. The action of dilute acids and the acceleration of this action by the agencies referred to may be explained by the fact that the tenuous film of iron oxide is dissolved by the acids, or that the hydrogen liberated on the iron, when the latter is acting cathodically in contact with zinc, reduces the oxide film or the gaseous oxygen film, thus removing the cause of passivity, whichever this may be. Mechanical abrasion would, of course, destroy either form of protection, if not wholly, by at least so damaging it that its continuity is broken, thereby exposing the electro-positive metal beneath it to attack, when such action will be accelerated by differences in potential between the remaining portions of the film and the exposed metal. The remainder of a film of such excessive thinness as is assumed to suffice to impart passivity would, under these conditions, be almost instantaneously removed, and hence the apparent result that .scratching, etc., renders all portions of the surface active simultaneously. It has been shown, however, that the untouched areas are not actified in this way. If a piece of passive iron is scratched at one end then the immersion of the other end in some reagent shows that it is still passive, but if the level of the liquid is raised until the damaged end is submerged, then the whole piece at once exhibits activity.
It has already been pointed out that passivity is fugitive, and this fact has been quoted as evidence against the Oxide theory. Such evidence is not conclusive, however ; it is clear that moist conditions will accelerate the activation of passive iron, but even in the absence of moisture an oxide film of molecular dimensions as regards thickness may be anticipated to be extremely susceptible to such influences as slight variations in temperature, etc., which could result in the flaking off of the film in much the same way as hammer scale is known to peel away from iron articles.
Another method by which passivity may be destroyed is by heating the passive metal in a reducing atmosphere, and this may obviously be acclaimed as substantial evidence in favour of the Oxide and Gaseous Film theories, for in a reducing atmosphere it is equally easy to imagine the reduction of either a metallic oxide film, a gaseous oxygen film, or a gaseous oxide of nitrogen film. Galvanic activation may be _ explained in the same way, since when the passive iron is made cathodic, gaseous hydrogen will be liberated on its surface and may reduce the passivifying film as above. This process of destroying passivity does not appear to afford much support to the physical theory, except that it might be argued that if an oxidising medium is conducive to the production of that physical condition which confers passivity on the metal, then the antithesis of this, i.e., a reducing medium, may be expected to operate in the reverse direction, by restoring a normal physical condition associated with normal chemical activity.
Heathcote (J. Soc. Chem. Ind., 1907, 26, 899) threw considerable doubt on the tenability of the Gaseous Film theory by carefully conducted experiments, in which he showed that, contrary to previously recorded observations, passivity is not destroyed by high vacua. In his experiments the passive material was subjected to pressures of the order of 1/50 of a millimetre, and it is only reasonable to suppose that any film of gas on the surface of the metal would have been broken up by such treatment, especially as it was observed that the metal gave up occluded gases. Yet on removal from the apparatus the iron was still passive.
Further observations might be given to supplement the foregoing, but they would serve no further purpose than to emphasise that the cause of passivity has still to be definitely established, though it may ultimately be found that each theory is correct in that passivity may not necessarily be the result of one particular cause only, but may be induced by a number of alternative causes.
On the object of stray current corrosion see also: DC traction, Cathodic protection, Coating, Contour plots, Definition, Detection, Examples, External currents, Historical perspective, Impressed current, Interference, Mechanisms, Modeling. Pipeline, Potential distribution, Prevention, Stray fields and leakage, Transit systems