Corrosion Doctors site map Corrosion information hub: The Corrosion Doctor's Web site Corrosion engineering consultant



Site index

A to Z listing



Corrosion glossary


Famous scientists

Corrosion course

Distance Ed

Doomsday scenarios



Monitoring glossary

Photo gallery

Rare earths

Search this site

Textbook assignments

Toxic elements

Water glossary



Form of the Corrosion

by: McKay & Worthington, 1936

Examination of metals during corrosion or after it has occurred may indicate much about the cause of the corrosion and why it continues. It should include macroscopic examination of the surface with the corrosion product in place and also after the corrosion product has been removed. (reference)

This should indicate in most cases the general "form" of the corrosion ; that is, whether it has been a direct chemical action, a pitting, galvanic action, concentration cell corrosion, dezincification, or corrosion cracking or corrosion fatigue. Of course, several of these effects may take place at the same time but it is usually the case that where failure has occurred it is directly attributable to one of these forms with possibly some help from a second. Each one of these forms is accompanied by typical signs which are not difficult to read with a little experience. These signs will be described in connection with a discussion of each form. In making laboratory or service tests it is always desired to make a diagnosis as rapidly as possible. It will often be a matter of some difficulty of judgment to decide which one of the various forms which exhibit themselves in the early stages, is the one which will cause the final failure. Many of the mistakes which are made in evaluating corrosion tests are due to failure to make such judgment correctly. For instance, it is a mistake to assume one metal better than another or one condition less harmful on the basis of loss of weight, which is a good measure of direct chemical corrosion, if the final failure is to be by pitting rather than by direct chemical corrosion. This holds true for each of the forms considered. Each has its-own measure and the, correct one must be selected from the five others-to judge values correctly.

Direct Chemical

There are a large number of chemical reactions in which a pure metal may take part where the end products are non-metals. Each of these myriad reactions will produce corrosion if allowed to proceed at the surface of a metal. There is no more mystery about why and how they will proceed than there is about why and how any chemical reaction proceeds, with the exception that the space arrangements are far from being as perfect as in a simple chemical reaction like pouring the solution of one reactant into the solution of another. The engineer who is interested mainly in corrosion rate and design to resist it can well accept the theories of the theoretical chemist on the one hand and of the theoretical metallurgist on the other as to the mysteries of these reactions, and for a true understanding of the corrosion problem simply be sure that he understands how the accepted theories work together to produce the particular corrosion being considered.

When a metal is in contact with one of these many reactants with which it may combine to form a non-metal or to precipitate another metal, corrosion will proceed at a rate determined by the combination of "rate factors" described above. If none of the other "forms" is present, we then have the simplest and the most fundamental kind of corrosion. The action is relatively uniform over the whole surface of the metal and proceeds at a fairly uniform rate as the metal is corroded.

The appearance of the surface of the metal being corroded in this manner will be relatively smooth but obviously "etched." The color of the surface will be the same as that of a freshly ground surface of the same metal. This surface appearance is an indication that the corrosion is of the direct chemical type unless other signs are present as described later.

A common example of this type of corrosion is the action of a good pickling solution on any metal. These solutions besides dissolving or otherwise removing the oxide will leave the metal surface itself etched clean and smooth.
The loss in weight of the metal per unit of surface per unit of time is a good easy method of determining the rate of this type of corrosion.

The most common condition causing this type of corrosion is an acid solution in contact with the metal. Oxidizing agents in the solution may speed up the reaction without introducing other forms of corrosion if they are not present in too high relative concentration or activity. If any solid corrosion products or other surface films are formed they will interfere with the uniform action and if conditions tending to produce measurable electric currents are present they will localize the corrosion. Although, under such conditions, the direct chemical action may still go on at an undiminished rate, the local corrosion is likely to be the most serious cause of failure. Measurement of the weight loss will only serve to measure the corrosion if localization happens to be also measurable or uniform for the existing variations of time and conditions. (back)


Corrosion is often localized in well-defined areas and where such areas are relatively small as compared with the whole metal surface, they are spoken of as its. A pit is usually distinguishable by a relatively sharp or well-defined boundary around part or all of its perimeter. This may be just a sharp change in the direction of an otherwise level surface but the boundary is more usually a wall which approaches the perpendicular to the surface and in many cases it may be undercut or so irregular that the surface or intermediate parts of the wall extend out well over the bottom of the pit. Intergranular pits may extend in a honeycomb shape like the complicated caves of a limestone region.

Pitted metal commonly has a speckled or irregularly dotted appearance. Pitting is usually accompanied by the formation of solid corrosion products which are usually irregular in appearance with probably the thickest sections over the pits. A dotted appearance of the corrosion product or a surface with well defined irregularities is liable to be an indication of pits underneath. When the corrosion product is cleaned off it will often have a tendency to cling in the pits even to a greater extent than would be expected from the purely mechanical difficulty of cleaning out such a confined space. An otherwise clean uniform surface may contain pits which are full of adherent corrosion product.

Pitting as a general head is allied very closely with other subjects in this outline; these subjects are indicated below and the reader should refer to them for a more complete description of the relation of each to pitting.

Pits may be formed by a protective corrosion product which partially covers the surface so that certain areas are exposed to repeated action of the corroding agent. Pits occur on the exposed areas. While this action is not unusual and is easily diagnosed, it is not one of the most common causes of pits.

The second cause in practical importance is probably the presence of inhomogeneity in the metal. If the metal contains inclusions of a more easily corroded metal, the latter may corrode galvanically or directly at a more rapid rate and so form a pit. If there are inclusions more cathodic than the body of the metal these inclusions induce pits in their neighborhood galvanically. If there are inclusions permeable to moisture or otherwise, such as to leave a pocket which can contain corroding agent different in composition from the body of the corrodent, concentration or solution cell pitting may occur.

This latter most common form of pitting may start as an inclusion or may start from exterior factors. Because it is usually rather than decreasing in intensity by its own action, likes the other forms of pitting, it is probably the most destructive. It is more fully described under its own heading and the reader is urged to study that part carefully if he wishes to really understand the pitting phenomenon.

A fourth cause of pitting is inhomogeneity existing in the metal from some other cause than inclusions. This may be uneven distribution of strains, particularly such strains as cause near or incipient parting of the metal, or dendritic or granular segregation of the constituents. While this cause may not be as widely distributed, if all corrosion is considered, as some other forms, it is of extreme importance where it does exist and in environments where it has been found to occur it should always be looked for carefully as the tendency may exist with little obvious or measurable indication.

Pitting is measured by actual micrometer measurement of the pit depth with an instrument having a properly shaped tip, or by a calibrated high power microscope focusing on the surface and then on the pit bottom. Qualitative measurement may be made by examination of properly prepared sections or by the x-ray. Weight loss is of little help and tensile tests may be somewhat misleading though they do give indications in the proper direction. This most destructive form of corrosion is possibly the most difficult of experimental evaluation.(back)

Two-Metal Galvanic Action

The first-discovered and simplest mechanical form of electric cell consists of two metals connected with each other and immersed in a single solution. The result is well known. The action of the cell depends entirely on the different chemical properties of the two metals. The term has been broadened by some physical chemists to include cells wherein the presence of different solutions affects the results. In corrosion literature, however, it has been the practice to limit the term to its original meaning which does not include solution differences and to call cells caused by solution differences concentration cells.

It should be noted that the concentration or solution cell produces the same effects, in many cases, as the two-metal galvanic cell. The two forms are interdependent and in the more detailed discussion under "Concentration or Solution Cells" their relation is described. (back)


Two-metal galvanic corrosion is rapid corrosion of a metal due to its being electrically connected to another metal in corrosive conditions. The cause is the difference between the individual tendencies of metals to be acted on by (combine with) corrosive materials. By certain arrangements this difference in tendency of two metals to corrode can be measured in volts. This voltage of the cell is the driving force behind the corrosion.

Galvanic Series

Arranging the metals according to their tendency to corrode galvanically results in the approximate order shown here. Note that this is not an "Electrochemical Series" but, due to the fundamental chemical affinities, that it has some general analogies with the "Electrochemical Series." This series shows the tendency to corrode, but actual galvanic corrosion takes place only when current flows. To allow current to flow certain conditions are necessary as follows:


The circuit must be complete through conducting metals and solutions. Metals are usually good conductors and allow easy passage of current, but serious resistance may develop in the solution and in the metal contacts. Distilled water and air are such poor conductors galvanic corrosion is unusual in them. The presence of slight impurities and salt in water, as in ordinary tap water, makes it sufficiently conducting to produce some galvanic corrosion. Stronger salt solutions, as sea water, are good conductors, and concentrated solutions of salt and chemicals, particularly strong acids and alkalies, are very good conductors, and galvanic corrosion is common in them. Current may pass short distances through damp films of corrosion products or condensed moisture.

Metal contacts in ordinary types of metal seams are good enough to allow galvanic corrosion to take place. Accumulations of dirt or corrosion products may slow up the corrosion. This has occurred in bearings. Joints may be insulated by rubber, fiber or wood so as to prevent galvanic action completely. Insulation must be absolute and not allow 'any metal path for the current.

Contact Surface

Galvanic action often produces gases (hydrogen) and corrosion products which prevent contact. Rapid flow which removes these impediments or chemicals which react with them increase two-metal galvanic corrosion. (Stagnation may produce solution concentration cells which are as active as the two-metal cells.) There are many changes which take place at corroding metal surfaces when current flows. Formation of hydrogen gas at the protected surface is common, also formation of insoluble protective films on either protected or corroded surface.

Interference with passage of currents by such films or by chemical changes is called polarization. Chemical changes take place in the contact films of solution, sometimes lowering its conductivity and sometimes etimes producing independent currents which may oppose and reverse the original one. These changes usually oppose the original effect. These effects are some of the most important factors governing galvanic action but are too complicated to describe fully here.


Surfaces must be close together. Where sheets meet at an edge all galvanic corrosion will be close to the seam, except in electrolytes of the highest conductivity.

Large Cathode Area

A large area of protected metal produces more rapid corrosion on the metal corroded. Small anodes are rapidly corroded. In good conducting solutions a steel rivet in copper or nickel alloy sheet is rapidly corroded but a copper or nickel alloy rivet in steel sheet will produce little corrosion except at the contact.

High Voltage

Large differences in voltage, i.e. distance from each other in the series list, tends to produce more rapid corrosion, but this effect is liable to be masked by preceding factors. (back)

Concentration Cell or Solution Cell

This type of corrosion has come to be accepted in a somewhat limited circle of engineers and corrosion students as one of the important "forms." In the experience of the writers and their colleagues it has been of much value in diagnosing and preventing corrosion. However, it has not had such complete experimental study and description in the literature as has, for instance, the galvanic cell or the "differential aeration" (or film) cell. It therefore seems necessary to assign rather more space and emphasis to it in this work than its relation to the other "forms" of corrosion would warrant.

Electrolytic vs. direct chemical corrosion : Corrosion may take place by direct reaction between the metal and the solution in contact with it, or the corrosion reaction may separate into anodic and cathodic parts which may take place at areas separated from each other by finite distances. If the reaction takes place without separation, it is called direct chemical corrosion and if the two parts are separated it is called electrolytic corrosion. Whether it is direct chemical or electrolytic, the course of the reaction must be in conformity with the known laws of electrochemistry and thermodynamics.

The separation of the reaction into two parts happens readily because the metal is a conductor. Separation often becomes necessary to the progress of the reaction because of the effect of the corrosion reaction on the plane of contact. The compounds of the metal formed are often rather insoluble and the contact film of corroding agent may become saturated and precipitate corrosion products. Thus the solid products of corrosion often prevent free contact of the corrosive with the metal soon after corrosion begins. But for electrolytic corrosion, it is not necessary that contact be made at the same point where metal is corroded. According to the principle of the concentration or solution cell the metal goes into solution or combination at one point when the best contact with the corrosive solution is at some other.

The amount of corrosion depends upon the supply of corroding agent to the catltwilc areas, but most of the metal disintegration occurs at the anodic areas, which thus become pits. The action of the current is in the direction described because the corroding reaction must be such as to furnish plus charges to the metal, and the areas over which a greater concentration of available plus charges are present are cathodic whereas those over which lesser concentrations are present must be anodic.

In practical work there are many cases of electrolytic corrosion where it is impossible to say "This is a two-metal galvanic cell" or "This is a solution or concentration cell." Their action often takes place simultaneously and inseparably. On the other hand there are many cases where the corrosion is obviously due to one class or the other. Where a pipe line carries a corroding solution at a high rate of speed sweeping off all corrosion products and keeping the liquid film in contact with the metal at uniform concentration and the pipe line is made of metals of widely differing corrodibilities there will be ordinary two-metal galvanic corrosion at the joints. Where a metal tank lining is of the same homogeneous material and the liquid quiet and corrosion products form on the surface of the metal there will be concentration or solution cell corrosion in pits or local areas on the metal lining. If corrosive solution stands quiet in the pipe line the concentration of corroding agent and corrosion product in the film of contact will change and be different at different points and there will be a combination of twometal and solution cell corrosion.


This term is used to mean the corroding away of one constituent of any alloy leaving the others more or less in situ. This phenomenon is perhaps most common in brasses containing high percentages of zinc but the same or parallel phenomena are familiar in the corrosion of aluminum bronzes and other alloys of metals of widely different chemical affinities. It is interesting from its very striking and destructive effects. It is an illustration of the two types of concentration cell corrosion and their interaction with galvanic cells.

It usually becomes evident as an area with well-defined boundaries and within which the more noble metal becomes concentrated as compared with the original alloy. In the case of brass the zinc is often almost completely removed and copper is present almost in a pure state, but in a very weak mechanical condition. Corrosion by the formation of such areas is usually quite rapid and since the corrosion is much localized it is one of the most destructive forms of corrosion. The most favorable conditions for dezincification seem to be a good conducting solution, as salt water, and a slightly acid condition with the presence of oxygen.

The mechanism is probably as follows. Due to any accidental variation ' in the corroding solution or the metal a concentration cell or a galvanic cell is formed and one area of surface is corroded more rapidly than a neighboring one. The more highly corroded area thus has over it a concentration of copper in solution greater than that over the other area, forming a metal-ion concentration cell which deposits the copper nearly at the spot where it originally dissolved. The original concentration cell continues to act, being accelerated by the formation of insoluble corrosion products. Some copper remains in solution and the copper-ion cell thus is retrieved and acts continuously adding its effect to that of the original cell to dissolve more metal at the anodic area under the corrosion product and the copper mass. The copper mass acts as a cathode in a galvanic cell which operates in the same direction. As corrosion proceeds the solution in the copper mass becomes acid enough by anodic action to keep the zinc in solution and allow it to migrate under the effect of the current out to the edge of the mass where it is carried away or precipitated due to the lowered acidity in the body of the corroding solution.


Damage to metal from exposure to cyclic, or fatigue, stresses may be intensified very markedly by exposure at the same time to a corrosive atmosphere or solution. In cyclic stresses of the nature of simple repeated bend and of a rotating beam, it is the surface layers of the metal that are most highly stressed. Corrosion likewise involves the surface layers first. The damage known as corrosion fatigue results from both influences.

Two factors in the progress of corrosion fatigue damage are well to keep in mind. One is that corrosion of the surface of metal lowers its fatigue resistance. The other is that fatigue stresses accelerate corrosive action. Consequently metal corroded will be found to have subsequently a lower fatigue limit than uncorroded, and therefore even greater damage results where corrosion and fatigue are at once present.

Corrosion in the absence of cyclic stresses is normally superficial and confined to the surface of the metal itself. Corrosion simultaneous with fatigue stresses leads to pronounced penetration as evidenced by cracks, with the result that the damage due to corrosion fatigue may be-and often is-great.

The fatigue or endurance test is ordinarily carried out in air. It has been shown by Gough and Sopwith and by Beckinsale that, strictly, such a test should be considered as corrosion fatigue, since fatigue limits in a partial vacuum or with the exclusion of air are higher than the normal air fatigue limits. Nevertheless, air in general is less damaging than aerated salt water.

The evidence suggests that oxygen is the exceptional damaging agent in corrosion fatigue of lead. Lead was found to be damaged far more when tested in air than when tested with a protective film of oil or in an acetic acid bath. In the acetic acid dissolved oxygen was used up in depolarizing hydrogen and so was kept from entering the metal.

What seems to be further evidence that oxygen is necessary is that copper has had higher fatigue resistance when tested in water than when tested in air, and highest of all when tested in vacuo. The fairly heavy corrosion product developing in the water tests is believed to have been an obstacle to the ingress of oxygen.

Deep, steep-walled cracks are a characteristic of corrosion fatigue. Presumably the part corrosion must play in producing corrosion fatigue is to start such a crack or at least to set up a condition where such a crack can be started by the fatigue stresses. Corrosion exposure of unstressed metal is much less likely to lead to premature failure than in the case of stressed metal. It is believed that the stressing damages the film of corrosion product, oxide or otherwise, and lays bare the metal to more or less unrestrained local corrosion. If the film can be conceived as being cracked open in places by the stress, then. corrosion may progress at these cracks. The metal at these cracks, being less readily accessible to oxygen, probably becomes anodic to the surrounding metal and consequently is subject to accelerated corrosion. This view of the course of premature fatigue failure fits in well with the well-known fact of the importance of the film in normal exposure.

Corrosion Cracking

The term "corrosion cracking" is used to include failures of a different nature from those.. known as corrosion fatigue failures. Whereas the latter result from simultaneous corrosion and fatigue stresses, failures of a corrosion cracking nature derive from simultaneous corrosion and static stresses. These static stresses ordinarily exist in metal from the strain of cold work. Cold-worked, high-zinc, brass, for instance, will crack more or less spontaneously in an environment that might not be damaging to annealed brass. Cold-worked Monel metal has been known to crack-handling mercury or mercury salt solutions-where annealed metal has been satisfactory.

Corrosion that produces cracking of this nature is intergranular. The damaging medium must have a high penetrating capacity. Penetration, leading undoubtedly to some chemical action on the intergranular material, destroys the tenacity and releases the internal strains. The spontaneous disintegration results.

The intergranular attack of 18-8 or of duralumin is not strictly corrosion cracking. Stress is not a necessary accompaniment. Improper heat treatment will produce an intergranular phase that is susceptible in the absence of strains to attack.